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 |
|---|---|---|---|---|---|---|---|---|---|---|
PMC9563457 | 36231106 | Peng Jiao,Jinpeng Wang,Jian Yang,Xingping Wang,Zhuoma Luoreng | Bta-miR-223 Targeting the RHOB Gene in Dairy Cows Attenuates LPS-Induced Inflammatory Responses in Mammary Epithelial Cells | 06-10-2022 | bta-miR-223,RHOB,NF-κB,mastitis | Bovine mammary epithelial cells (bMECs) are part of the first line of defense against pathogens. In recent studies, bta-miR-223 has been reported to activate congenital and innate immunity against inflammatory damage during the pathogenesis of mastitis in dairy cows. The purpose of this study was to identify the regulatory mechanism of bta-miR-223 and its downstream target genes in inflammatory bMECs. A double luciferase reporter gene assay demonstrated that ras homolog family member B (RHOB) was the target gene of bta-miR-223. To further elucidate the role of bta-miR-223 in congenital immune responses, bta-miR-223 mimics (mimic/inhibitor) were transfected into bMECs stimulated with lipopolysaccharide (LPS), which activates the Toll-like receptor 4/nuclear factor-κB (TLR4/NF-κB) signaling pathway. Real-time quantitative PCR (qPCR) and Western blot were used to detect the expression of related genes and proteins, and enzyme-linked immunosorbent assay (ELISA) was used to detect secreted inflammatory factors. Results showed that bta-miR-223 expression during inflammation in bMECs reduced the secretion of inflammatory factors by targeting RHOB and deactivation of NF-κB gene activity. Silencing RHOB inhibited LPS-induced inflammatory response in bMECs. Overall, bta-miR-223 attenuated LPS-induced inflammatory response, and acted as a negative feedback regulator via targeting RHOB, providing a novel avenue for mastitis treatment. | Bta-miR-223 Targeting the RHOB Gene in Dairy Cows Attenuates LPS-Induced Inflammatory Responses in Mammary Epithelial Cells
Bovine mammary epithelial cells (bMECs) are part of the first line of defense against pathogens. In recent studies, bta-miR-223 has been reported to activate congenital and innate immunity against inflammatory damage during the pathogenesis of mastitis in dairy cows. The purpose of this study was to identify the regulatory mechanism of bta-miR-223 and its downstream target genes in inflammatory bMECs. A double luciferase reporter gene assay demonstrated that ras homolog family member B (RHOB) was the target gene of bta-miR-223. To further elucidate the role of bta-miR-223 in congenital immune responses, bta-miR-223 mimics (mimic/inhibitor) were transfected into bMECs stimulated with lipopolysaccharide (LPS), which activates the Toll-like receptor 4/nuclear factor-κB (TLR4/NF-κB) signaling pathway. Real-time quantitative PCR (qPCR) and Western blot were used to detect the expression of related genes and proteins, and enzyme-linked immunosorbent assay (ELISA) was used to detect secreted inflammatory factors. Results showed that bta-miR-223 expression during inflammation in bMECs reduced the secretion of inflammatory factors by targeting RHOB and deactivation of NF-κB gene activity. Silencing RHOB inhibited LPS-induced inflammatory response in bMECs. Overall, bta-miR-223 attenuated LPS-induced inflammatory response, and acted as a negative feedback regulator via targeting RHOB, providing a novel avenue for mastitis treatment.
Mastitis in dairy cows is an inflammatory disease caused by pathogenic microorganisms infecting the mammary tissue that has a high incidence, low cure rate, frequent recurrence and expensive treatment. Furthermore, mastitis results in decreased milk production, lower milk quality, and even premature culling of cows, causing great economic losses to the dairy industry [1]. In recent years, a substantial body of research has been published on the molecular regulation mechanism of mastitis in dairy cows. The studies demonstrated that the occurrence, development, susceptibility and resistance to mastitis are regulated by a gene network composed of several genes, among which the Toll-like receptor 4/nuclear factor-κB (TLR4/NF-κB) signaling pathway is an important regulatory pathway [2,3]. Lipopolysaccharide (LPS) is a component of the outer membrane of Gram-negative bacteria and is a well-established agonist of TLR4 on the cell membrane surface of host cells, ultimately activating the NF-κB signaling pathway [4,5]. Therefore, the use of LPS stimulation to establish animal models of inflammation has been tested in a variety of species. Stimulation of bovine mammary epithelial cells (bMECs) with LPS significantly increases the expression of inflammatory and chemokine genes, such as interleukin-8 (IL-8), interleukin-6 (IL-6) and interleukin-1β (IL-1β) [6,7]. MicroRNAs (miRNAs) are a class of endogenous non-coding RNAs that regulate intrinsic and adaptive immunity by targeting and inhibiting the 3′ untranslated regions (3′ UTRs) of mRNAs. miR-223 was first identified by qPCR in 2003 [8]. It was previously reported that bta-miR-223 could alleviate inflammation-mediated damage in bMECs by targeting CBLB to inhibit the PI3K/AKT/NF-κB pathway, thereby suppressing IL-6 expression [9]. The functional SNP in the 3′ UTR of the HMGB1 gene affects bta-miR-223 binding, thereby tempering the regulation of mastitis in dairy cows [10]. In recent years, the regulation of RHOB by miR-223 has been reported in other species [11,12,13,14,15,16]. A typical example is that, in LPS-induced A549 (lung adenocarcinoma cells), low expression of miR-223 made targeting RHOB to inhibit the NLRP3 inflammasome and TLR4/NF-κB signaling pathway less effective. This resulted in an exacerbation of lung inflammation-mediated damage [11]. Therefore, miR-223 plays an important regulatory role in the biological processes of inflammatory diseases. However, the molecular regulatory mechanisms in bovine mastitis are not yet clear. Moreover, no molecular regulatory mechanism between bta-miR-223 and RHOB has been identified in cattle at the time of publication. In this study, the molecular regulatory mechanisms of bta-miR-223 on RHOB/NF-κB were explored at the cellular level to elucidate their possible response mechanisms in LPS-treated bMECs.
The bMECs (an immortalized MAC-T cell line) were cultured in DMEM/F12 medium (Ge Healthcare Life Sciences Hyclone Laboratories, South Logan, Utah) containing 10% fetal bovine serum (FBS) (Biological Industries Israel Beit Haemek Ltdkibbutz Beit Haemek, Israel) [17]. The purity of the bMECs was assessed by immunofluorescence staining to detect the expression of Cytokeratin 18 (a marker protein for epithelial cells) when the bMECs reached 80–90% confluence based on previously reported methods [17]. Human embryonic kidney (HEK)-293T cells were recovered and cultured in high-glucose DMEM (Ge Healthcare Life Sciences Hyclone Laboratories, South Logan, Utah) containing 5% FBS (Biological Industries Israel Beit Haemek Ltdkibbutz Beit Haemek, Israel). All incubation conditions were 37 °C and 5% CO2 concentration.
The mimic of miR-223 (dsRNA oligonucleotides), mismatched negative control mimic (NC_mimic), miR-223 inhibitor (single-stranded oligonucleotides) and mismatched negative control inhibitor (NC_inhibitor), each Cy3-labeled NC mimic/inhibitor, were synthesized by RIBOBIO Co., Ltd. (Guangzhou, China). Mimics and inhibitors were used to assess the effects of overexpression and inhibition of miR-223 on its activity in HEK-293 or bMECs. Then, the fluorescence transfection efficiency was measured (Figure S1A,B). The interference sequence of RHOB was synthesized by Sangon Biotech (Shanghai) Co., Ltd. (siRNA-RHOB labeled with the fluorescent dye FAM, herein referred to as FAM NC), was used to track transfection efficiency (Figure S1C). The sequence of siRNA-RHOB and FAM-labeled NC primers are presented in Table S1.
RHOB was predicted to be a potential target gene of bta-miR-223 through comparison of the seed region of bta-miR-223 and the 3′ UTRs of potential target genes, which was conducted using TargetScan (Version 8.0). Design of forward and reverse primers for the RHOB 3′ UTR was performed using Primer Premier 5.0 (Table S1). Then, total RNA from bMECs was extracted using TRIzol Reagent (Invitrogen, Waltham, MA, USA), reverse transcribed to complementary DNA (cDNA) according to the product specification of the PrimeScriptTM RT reagent Kit with gDNA Eraser (Takara Biomedical Technology Co., Ltd., Beijing, China). The 3′ UTR fragment of RHOB was amplified according to the following reaction conditions: 20 μL total reaction volume consisting of forward/reverse primer 0.8 μL each; cDNA 1 μL; Taq PCR Master Mix 10 μL; ddH2O 7.4 μL, amplification result is in Figure S2. The 3′ UTR double-stranded cDNA fragment of RHOB was enzymatically cleaved and cloned into the XhoI-NotI site of the psiCHECKTM-2 vector (Promega Corp., Madison, WI, USA) containing the Renilla and firefly luciferase reporter genes, to form the psiCHECK-2-RHOB-wt (wild-type) dual-luciferase reporter gene recombinant vector. The psiCHECK-2-RHOB-wt recombinant plasmid was extracted with the Endo-Free plasmid Mini Kit (Omega Bio-Tek, Norcross, GA, USA). The resulting recombinant vector was verified by sequence analysis. The psiCHECK-2-RHOB-mut (mutant-type) was synthesized by Sangon Biotech (Shanghai) Co., Ltd., and the sequence was verified (Figure S3). HEK-293T cells were seeded in 24-well plates at a density of 1.0 × 105 cells/well [13]. After 24 h, the cells were transfected with the above plasmids using 2 μL of X-tremeGENE HP DNA transfection reagent (Roche, Penzberg, Germany) according to the manufactur-er’s instructions. The HEK-293 cells were co-transfected with psiCHECK-2-RHOB-wt recombinant plasmid (0.5 μg) and mimic of bta-miR-223 (also set control NC_mimic 0.5 μg, final working concentration 75 nM) [18]. After 48 h, Renilla luciferase activity (normalized to firefly luciferase activity) was assayed using a dual-luciferase reporter assay (Promega, Madison, WI) according to the manufacturer’s instructions. If the ratio of Renilla to firefly luciferase in the CHECK2 3′ UTR + mimic transfected group was more than 30% lower than that in the control group, it indicated that the miRNA interacted with the target gene 3′ UTR.
The bMECs were seeded into a 6-well plate and incubated until they reached 60–70% confluence, according to the cell transfection method we have previously reported [18]. Three replicates of each group were transfected with 5 μL of miR-223 mimic and NC_mimic, 10 μL of miR-223 inhibitor and NC_inhibitor, and 20 ng/μL of LPS was added at 42 h after transfection to stimulate inflammation in bMECs [19]. The bMECs were collected after 6 h for qPCR and Western blot analyses, and culture supernatants were collected for enzyme-linked immunosorbent assay (ELISA) analysis of inflammatory factors. Transfection of siRNA-RHOB and NC was performed as described above. The bMECs were cultured for 42 h, when 20 ng/μL of LPS was added to initiate inflammatory responses. After 6 h, cells and supernatants were extracted for qPCR and ELISA, respectively.
In this experiment, total RNA was extracted from the above-transfected bMECs using TRIzol (Invitrogen, Waltham, MA, USA) according to the manufacturer’s protocol. The concentration and purity (OD260/280 ≥ 1.8; OD260/230 ≥ 1.0) of RNA was detected using a multifunctional microplate reader (Biotek Synergy SLXA, USA). Subsequently, the degradation and contamination of RNA were checked by agarose gel electrophoresis. Briefly, 1000 ng of qualified RNA was reverse transcribed into cDNA using a PrimeScript™ RT reagent Kit with gDNA Eraser (Takara Biomedical Technology Co., Ltd., Beijing, China). For bta-miR-223 detection, the RNA was reverse transcribed into a first-strand cDNA synthesis by the stem-loop method using PrimeScript™ RT reagent Kit with gDNA Eraser (Takara Biomedical Technology Co., Ltd., Beijing, China) and specific primer (5′ GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGA CTGGGGTAT 3′). Bta-miR-223 and all mRNA qPCR primers were designed using Primer Premier 5.0 (Table S1). The qPCR reaction was performed using a 2 × M5 HiPer SYBR Premix Estaq (with Tli RNaseH) Kit (Mei5 Biotechnology Co., Ltd., Beijing, China) on a Bio-Rad CFX 96 Touch instrument (Bio-Rad, Hercules, CA, USA). The qPCR reaction mixture was 20 μL total volume, consisting of 2 × M5 HiPer SYBR Premix EsTaq (with Tli RNaseH) 10 μL, forward/reverse primers 0.8 μL each; cDNA 2 μL; ddH2O 6.4 μL. The qPCR thermocycling profile was as follows: pre-denaturation at 95 °C for 30 s, 95 °C for 5 s, 60 °C for 30 s, 45 cycles, 95 °C for 10 s, 65 °C for 5 s. Three replicate wells and 3 biological replicates were set up for each assay. After the qPCR was completed, the relative expression of genes was calculated using the 2−ΔΔCt method using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and ribosomal protein S18 (RPS18) as internal references.
The bMECs were washed twice with PBS, then digested with trypsin into 1.5 mL centrifuge tubes. The digestion was terminated by adding an equal volume of medium, followed by centrifugation and discarding the supernatant. The cells were then washed with PBS and the supernatant was discarded following centrifugation. Total protein was extracted with via whole-cell Lysis Assay (Sangon Biotech Co. Ltd., Shanghai, China) and quantified with a BCA protein assay kit (KeyGEN BioTECH, Nanjing, China). Equal amounts of total protein (40 μg) were extracted from each sample, resolved by SDS-PAGE with 12% (NF-κB and β-Actin) and 15% (RHOB), and transferred to nitrocellulose pre-membranes (GE Healthcare/Amersham, Pittsburgh, PA). The membranes were then blocked with 5% skim milk powder and then incubated overnight at 4 °C with an appropriate dilution of primary antibody. All of the primary and the secondary antibodies, including RHOB (1:200, catalog no. sc-8048), NF-κB (1:500, catalog no. AF5006), β-actin (1:500, catalog no. sc-47778), goat anti-mouse IgG-HRP (1:5000, catalog no. sc-17829), goat anti-rabbit IgG-HRP (1:5000, catalog no. sc-2030) and RHOB were purchased from Santa Cruz Biotechnology Inc. Then, goat anti-mouse IgG-HRP and goat anti-rabbit IgG-HRP were used to detect antigen bound primary antibodies. β-Actin was used as a control group for comparison between groups. Western Blot images were analyzed using Image J software.
Following transfection and subsequent incubation for 48 h, the supernatants of the bMEC cultures from each treatment group were collected, and the secretion of cytokines IL-8, IL-6 and IL-1β was detected using the relevant ELISA kits (CUSABIO BIOTECH CO., Ltd., Wuhan, China) according to the manufacturer’s instructions.
The experimental data are display as the mean ± standard error of the mean (SEM), with three replications. The significant differences between groups were tested by Student’s t-test using GraphPad Prism 9 (GraphPad Software, Inc., La Jolla, CA, USA). p < 0.05 indicates a significant difference between the treatment groups (LPS, mimic, inhibitor and siRNA-RHOB) and the control groups (control, NC_mimic, NC_inhibitor and NC).
The results of immunofluorescence staining showed that Cytokeratin 18 was positively expressed in the bMECs after subculturing and was mainly distributed in the cytoplasm (Figure 1). This observation suggests that the cultured bMECs in this study are pure and suitable for subsequent studies.
The 3′ UTR binding site of the RHOB gene by bta-miR-223 was predicted using TargetScan (Figure 2A). To verify that RHOB was a target gene of bta-miR-223, the gene’s 3′ UTR was ligated into the psiCHECK-2 vector (named RHOB-wt) and co-transfected with psiCHECK-2-RHOB-wt and bta-miR-223 mimic (negative control) into HEK-293 cells; the fluorescence efficiency was detected using a dual-luciferase reporter system. It was observed that the luciferase activity was significantly down-regulated in the co-transfected psiCHECK-2-RHOB-wt and bta-miR-223 mimic groups compared with the negative control (p < 0.001) (Figure 2B). In addition, when the 3′ UTR of the RHOB gene was mutated (RHOB-mut), the efficiency of bta-miR-223-mediated luciferase activity inhibition was found to be completely lost (p > 0.05) (Figure 2B). These results suggest that bta-miR-223 targets the 3′ UTR end of the RHOB gene, resulting in a reduction in luciferase activity.
To detect the expression of bta-miR-223 in inflammatory bMECs, the expression of bta-miR-223 after 6 h of LPS stimulation was analyzed by qPCR. It was observed that the expression of inflammatory factors IL-1β, IL-8 and IL-6 were significantly up-regulated (p < 0.01) in LPS-stimulated bMECs (Figure 3A–C). Therefore, the model of inflammatory bMECs was successfully established. Subsequently, the expression of bta-miR-223 was significantly up-regulated (p < 0.01) in the inflammatory cell model (Figure 3D), suggesting that bta-miR-223 may play some role in the inflammatory response of bMECs, which requires further validation.
To explore the function and molecular mechanism of bta-miR-223 in bMECs, the cells were transfected with miR-223 mimic, NC_mimic, miR-223 inhibitor, or NC_inhibitor. The qPCR results show that, in bMECs, bta-miR-223 mimics significantly increased the expression of mature bta-miR-223 (p < 0.0001) (Figure 4A), whereas bta-miR-223 inhibitors significantly decreased the expression of mature bta-miR-223 (p < 0.001) (Figure 4B). Next, mRNA and protein levels of RHOB and NF-κB in the bta-miR-223 mimic transfection group and inhibitor transfection group, as well as their respective control groups, were analyzed by qPCR and Western Blot, respectively. The results indicated that compared with the negative control (NC_mimic), overexpression of bta-miR-223 in bMECs significantly suppressed the mRNA and protein expression levels of RHOB and NF-κB (p < 0.05) (Figure 4C–G). Conversely, inhibition of bta-miR-223 significantly up-regulated the mRNA and protein expression (p < 0.05) (Figure 4C–G). These results suggest that bta-miR-223 inhibits the expression of RHOB and NF-κB at the transcriptional and translational levels.
To investigate whether bta-miR-223 is involved in regulating the expression of inflammatory factors in bMECs, the mimic and inhibitor of bta-miR-223 and their NCs were transfected into bMECs. The bMECs were then stimulated with LPS to produce an inflammatory response, and the mRNA expression of IL-1β, IL-8 and IL-6 and the secretion levels of inflammatory factors were detected using qPCR and ELISA, respectively. The results showed that compared with the control group (NC_mimic), the bta-miR-223 mimic group significantly inhibited the expression of IL-8 and IL-6 mRNA and protein (p < 0.05) (Figure 5A–D). Although IL-1β mRNA was not significantly inhibited (p > 0.05), protein secretion was significantly inhibited (p < 0.05) (Figure 5E,F). Meanwhile, the bta-miR-223 inhibitor group significantly up-regulated the expression of inflammatory factors IL-8, IL-6 and IL-1β at the mRNA and secretory levels compared with the control group (NC_inhibitor) (p < 0.05) (Figure 5A–F). The above results indicate that bta-miR-223 inhibits the expression of IL-8 and IL-6 transcription and secretion in bMECs, whereas IL-1β is inhibited by bta-miR-223 only at the secretion level. Taken together, bta-miR-223 is a negative regulator of inflammatory responses in bMECs.
To further confirm that bta-miR-223 inhibits the expression of inflammatory factors at the mRNA and secretion levels by regulating RHOB, an interference test of RHOB in bMECs was conducted and analyzed by qPCR and ELISA to detect inflammatory markers before and after silencing RHOB in bMECs. The results show that transfection of siRNA-RHOB resulted in a 50.1% down-regulation of RHOB gene expression in bMECs (p < 0.0001, Figure 6A). Silencing of RHOB significantly inhibited the expression of IL-8, IL-6 and IL-1β gene expression (p < 0.01, Figure 6B–D), and also significantly suppressed the secretion of the respective proteins in bMECs (p < 0.05, Figure 6E–G). Therefore, the above results indicate that silencing RHOB inhibits the release of inflammatory factors in bMECs, which is consistent with the results of the bta-miR-223 mimic group, resulting in the attenuation of the inflammatory response in bMECs.
miRNAs are important regulators of several components of mammalian immune responses. It has been demonstrated previously that miR-223 is significantly up-regulated in inflammation-induced granulocytes, dendritic cells, T cells, endothelial cells and epithelial cells [20]. One study showed that bta-miR-223 was up-regulated approximately 2.5- to 3-fold in dairy cows with mastitis [21,22]. In the present experiment, it was observed that the expression of bta-miR-223 was also up-regulated in bMECs during inflammation, which corroborates the above findings. Regulation of gene expression by miRNA occurs through specific binding to target genes containing complementary nucleotide sequences. The bovine and human miR-223 sequences are identical. The bovine and human RHOB sequences are highly homologous, however, they are not identical. Although miR-223 has been identified to regulate RHOB gene expression in humans, the molecular regulatory mechanism in LPS-induced mastitis in cows has not been examined. It has been shown that in human lung inflammation, miR-223 can alleviate the inflammation caused by the release of inflammatory factors in the lung by directly targeting the 3′ UTR of RHOB, and low expression of RHOB has been closely associated with a lack of NF-κB signaling pathway activity [11]. In the present experiments, the bta-miR-223 binding site within the 3′ UTR of the RHOB gene was predicted and verified by a dual-luciferase reporter assay. It was observed that bta-miR-223 does indeed bind to the RHOB target, which is consistent with the findings of the above-mentioned authors. RHOB is a low molecular weight GTPase [23]. In human colon epithelial cells, it was reported that miR-223 targeting of RHOB limited the spread of colon cancer cells [24]. In non-small cell lung adenocarcinoma (NSCLC) cells, miR-223 mediated RHOB knockdown and significantly inhibited RHOB protein expression, thereby suppressing the development of NSCLC [14]. In dendritic cells, miR-223 down-regulated RHOB protein expression, thereby promoting T cell differentiation [15]. In human acute lung injury (ALI), the activation of RHOB potentiates the TLR4/NF-κB signaling pathway activity, resulting in the release of inflammatory factors [11]. Interestingly, in LPS-induced inflammation in mouse macrophages, miR-223 inhibited the release of several inflammatory factors (tumor necrosis factor-α (TNF-α), IL-6 and IL-1β) associated with TLR4 activation by targeting RHOB, thereby causing macrophage inflammatory damage [16]. Meanwhile, studies have reported that RHOB is a negative regulator of the NF-κB (p65/50) signaling pathway, which has been demonstrated to act in a dose-dependent manner [25,26]. This resulted in reduced expression of inflammatory factors, thereby alleviating the inflammatory damage induced by LPS [27]. In macrophages, the expression of inflammatory factors (TNFα, IL-6 and IL-1β) can be significantly inhibited at the mRNA and protein levels by overexpression of RHOB [28]. In contrast, the silencing of RHOB in human endothelial cells inhibited the production of IL-8 and IL-6 [29]. In this experiment, inflammatory factors (IL-6, IL-8 and IL-1β) were found to be suppressed at both mRNA and protein levels when RHOB expression was knocked down, which is consistent with the above results. Therefore, RHOB is a potential novel target for the treatment of mastitis in cows. Once pathogenic bacteria have entered the breast, bMECs respond rapidly to initiate the antimicrobial response [30,31]. Interestingly, mRNA levels of all inflammatory factors increased rapidly 2–4 h after stimulation of bMECs with LPS and LTA, but only LPS stimulation resulted in sustained inflammatory factor expression [19,32] of the TLR4/NF-κB inflammatory signaling pathway [33]. It is therefore hypothesized that the NF-κB signaling pathway is only continuously activated in LPS-stimulated bMECs. According to Han et al. [9], bta-miR-223 in LTA-induced mastitis does not inhibit the precursor protein of NF-κB (p50/p65) and only inhibits phosphorylation of NF-κB. In contrast, in LPS-induced bMECs, overexpression of bta-miR-146a inhibited the precursor protein of NF-κB [18]. In addition, it has been shown that heat-inactivated LPS activates NF-κB through the TLR4/NF-κB signaling pathway [33,34]. In response to activation of the TLR signaling pathway, NF-κB binds to the promoters of inflammatory factor genes, resulting in increased mRNA and protein levels of the inflammatory cytokines TNFα, IL-6, IL-1β and IL-8 within 1–3 h of LPS activation [35,36,37]. It is well established that IL-1β and IL-8 play an important role in the recruitment and activation of immune cells via local and systemic effects to eliminate pathogens, thereby inhibiting the development of acute inflammation [38,39,40]. IL-6 is the main pro-inflammatory cytokine mediating the inflammatory response and plays an important role in the acute phase response to pathogen detection [41]. In order to investigate whether bta-miR-223 is involved in regulating the expression of inflammatory cytokines in cows, LPS-stimulated bMECs were employed as a model of inflammation. The results showed that overexpression of bta-miR-223 could down-regulate the translation of RHOB, NF-κB and the associated inflammatory factors (IL-8, IL-6 and IL-1β). Silencing RHOB also effectively suppressed the secretion of inflammatory factors, further indicating that bta-miR-223 alleviated the inflammation of bMECs caused by LPS via suppression of NF-κB and RHOB (Figure 7). Taken together, these data implicate a potential mechanism of action and the utility of bta-miR-223 in the molecular diagnosis of mastitis in dairy cows.
In this study, it was demonstrated that elevated expression of bta-miR-223 in LPS-induced bMECs suppressed the expression of the RHOB and NF-κB genes, thus resulting in decreased secretion of inflammatory mediators IL-1β, IL-8 and IL-6 (Figure 7). Furthermore, the results of inflammatory factor secretion from bMECs after silencing the RHOB gene corroborated the observations from bta-miR-223 overexpression experiments. The above results suggest that bta-miR-223 is a negative regulator of inflammation in bMECs and can alleviate the inflammatory response of bMECs by knockdown of RHOB gene expression. This study sheds light on the regulatory mechanism of bta-miR-223 on the immune response to mammary gland infection in cows at the cellular level, providing a new molecular target for the treatment of mastitis. | true | true | true |
PMC9563608 | Bianca C. Bernardo,Gunes S. Yildiz,Helen Kiriazis,Claudia A. Harmawan,Celeste M. K. Tai,Rebecca H. Ritchie,Julie R. McMullen | In Vivo Inhibition of miR-34a Modestly Limits Cardiac Enlargement and Fibrosis in a Mouse Model with Established Type 1 Diabetes-Induced Cardiomyopathy, but Does Not Improve Diastolic Function | 03-10-2022 | diabetes,microRNA,miR-34a,cardiomyopathy,fibrosis | MicroRNA 34a (miR-34a) is elevated in the heart in a setting of cardiac stress or pathology, and we previously reported that inhibition of miR-34a in vivo provided protection in a setting of pressure overload-induced pathological cardiac hypertrophy and dilated cardiomyopathy. Prior work had also shown that circulating or cardiac miR-34a was elevated in a setting of diabetes. However, the therapeutic potential of inhibiting miR-34a in vivo in the diabetic heart had not been assessed. In the current study, type 1 diabetes was induced in adult male mice with 5 daily injections of streptozotocin (STZ). At 8 weeks post-STZ, when mice had established type 1 diabetes and diastolic dysfunction, mice were administered locked nucleic acid (LNA)-antimiR-34a or saline-control with an eight-week follow-up. Cardiac function, cardiac morphology, cardiac fibrosis, capillary density and gene expression were assessed. Diabetic mice presented with high blood glucose, elevated liver and kidney weights, diastolic dysfunction, mild cardiac enlargement, cardiac fibrosis and reduced myocardial capillary density. miR-34a was elevated in the heart of diabetic mice in comparison to non-diabetic mice. Inhibition of miR-34a had no significant effect on diastolic function or atrial enlargement, but had a mild effect on preventing an elevation in cardiac enlargement, fibrosis and ventricular gene expression of B-type natriuretic peptide (BNP) and the anti-angiogenic miRNA (miR-92a). A miR-34a target, vinculin, was inversely correlated with miR-34a expression, but other miR-34a targets were unchanged. In summary, inhibition of miR-34a provided limited protection in a mouse model with established type 1 diabetes-induced cardiomyopathy and failed to improve diastolic function. Given diabetes represents a systemic disorder with numerous miRNAs dysregulated in the diabetic heart, as well as other organs, strategies targeting multiple miRNAs and/or earlier intervention is likely to be required. | In Vivo Inhibition of miR-34a Modestly Limits Cardiac Enlargement and Fibrosis in a Mouse Model with Established Type 1 Diabetes-Induced Cardiomyopathy, but Does Not Improve Diastolic Function
MicroRNA 34a (miR-34a) is elevated in the heart in a setting of cardiac stress or pathology, and we previously reported that inhibition of miR-34a in vivo provided protection in a setting of pressure overload-induced pathological cardiac hypertrophy and dilated cardiomyopathy. Prior work had also shown that circulating or cardiac miR-34a was elevated in a setting of diabetes. However, the therapeutic potential of inhibiting miR-34a in vivo in the diabetic heart had not been assessed. In the current study, type 1 diabetes was induced in adult male mice with 5 daily injections of streptozotocin (STZ). At 8 weeks post-STZ, when mice had established type 1 diabetes and diastolic dysfunction, mice were administered locked nucleic acid (LNA)-antimiR-34a or saline-control with an eight-week follow-up. Cardiac function, cardiac morphology, cardiac fibrosis, capillary density and gene expression were assessed. Diabetic mice presented with high blood glucose, elevated liver and kidney weights, diastolic dysfunction, mild cardiac enlargement, cardiac fibrosis and reduced myocardial capillary density. miR-34a was elevated in the heart of diabetic mice in comparison to non-diabetic mice. Inhibition of miR-34a had no significant effect on diastolic function or atrial enlargement, but had a mild effect on preventing an elevation in cardiac enlargement, fibrosis and ventricular gene expression of B-type natriuretic peptide (BNP) and the anti-angiogenic miRNA (miR-92a). A miR-34a target, vinculin, was inversely correlated with miR-34a expression, but other miR-34a targets were unchanged. In summary, inhibition of miR-34a provided limited protection in a mouse model with established type 1 diabetes-induced cardiomyopathy and failed to improve diastolic function. Given diabetes represents a systemic disorder with numerous miRNAs dysregulated in the diabetic heart, as well as other organs, strategies targeting multiple miRNAs and/or earlier intervention is likely to be required.
The number of people with diabetes has been rising globally and was estimated to exceed 450 million adults worldwide in 2017, representing approximately 8.4% of adults (18–99 years). By 2045, the prevalence is predicted to rise to 9.9%, representing over 690 million adults [1]. Diabetes is associated with increased mortality and morbidity, attributed largely to cardiovascular and kidney disease [2,3]. The heart undergoes adverse remodeling in a setting of diabetes and this is associated with an increased risk of heart failure and sudden death after myocardial infarction [4]. This elevated risk in diabetic patients remains even after adjusting for age, blood pressure, weight, cholesterol, and coronary artery disease [5,6,7]. The terms “diabetic cardiomyopathy” and “diabetic heart” have been used to define the cardiac dysfunction which can occur in diabetic patients in the absence of other factors including coronary artery disease and hypertension. Two key features of the diabetic heart include diastolic dysfunction and cardiac fibrosis [2,3,5,6]. There is no specific therapy for the diabetic heart, and with the exception of sodium-glucose contransporter-2 inhibitors (SGLT2i), therapies have not demonstrated positive cardiovascular outcomes (including thiazolidinediones, incretin-based therapies (glucagon-like peptide 1 and dipeptidyl peptidase-4 inhibitors)) [5,7,8,9]. There remains an unmet need to develop novel therapies for the diabetic heart. MicroRNAs (miRNAs) are short strands of RNA that are not transcribed into protein, but instead, regulate the expression of many genes by interacting with specific sites in 3′ untranslated regions of messenger transcripts to prevent protein translation and gene expression [10]. miRNAs play a crucial role in health and disease, and as such, molecular tools to manipulate miRNA activity have been developed and proved successful in preclinical disease models [11,12]. Furthermore, some have entered clinical trials in patients with hepatitis C [13], cancer, type 2 diabetes, and non-alcoholic fatty liver diseases [14,15]. Alterations in miRNA expression in cardiac tissue from diabetic and non-diabetic mice have been identified from profiling studies [16,17], and represents a new class of targets for the development of therapeutics for diabetic cardiomyopathy. Over the past decade, miRNA-34a has received considerable attention for its therapeutic potential in the heart. Expression of miR-34a is elevated in different animal models of cardiac stress [18,19,20,21,22,23], and was also shown to be elevated in the human failing heart [24,25]. We, and others, have demonstrated that therapeutic inhibition of miR-34a is beneficial in settings of heart pathology such as aging, pressure overload-induced pathological hypertrophy, acute myocardial infarction and dilated cardiomyopathy [18,19,20,21,22,23]. However, the therapeutic potential of targeting miR-34a has not yet been explored in the diabetic heart. Circulating levels of miR-34a were elevated in diabetic patients compared with controls [26], and was also shown to be upregulated in the hearts of type 2 diabetic patients with ischemic heart disease compared to the hearts of non-diabetic patients with ischemic heart disease [27]. Thus, miR-34a represents an unexamined target in a diabetic setting. The key aim of this study was to determine whether silencing miR-34a can improve diastolic function and attenuate cardiac fibrosis in the type 1 diabetic heart.
All experiments using animals were conducted in accordance with the Australian code for the care and use of animals for scientific purposes (National Health & Medical Research Council of Australia, 8th Edition, 2013). All animal procedures and care were approved by the Alfred Research Alliance Animal Ethics Committee (approval number E/1430/2014/B). Male mice on a FVB/N background were housed and bred onsite in a temperature-controlled environment under a 12 h light/dark cycle, and up to six littermates per cage. Mice had ad libitum access to standard rodent chow and water. Aged-matched (~six-week-old) littermate male mice were randomly allocated into diabetic and non-diabetic groups. Mice received five consecutive daily intraperitoneal (i.p.) injections of streptozotocin (STZ; 55 mg/kg in 0.02 mol/L citrate buffer, pH 4.5) to induce type 1 diabetes, or five consecutive daily intraperitoneal injections of citrate buffer vehicle of equivalent volume (non-diabetic group), and housed up to two littermates per cage. Mice were followed for 16 weeks, as previously described [28,29,30]. Diabetes was confirmed by measurements of blood glucose (via tail vein prick) every two to four weeks using a glucometer (blood glucose levels exceeding 28 mmol/L were considered diabetic). After eight weeks of untreated diabetes, mice were administered subcutaneously three consecutive daily doses of locked nucleic acid (LNA)-antimiR-34a (25 mg/kg per dose) or saline (150 μL). We have previously demonstrated that this dosing regime is sufficient to achieve knockdown of miRNAs in the heart for at least eight weeks [18,21,31,32]. Based on significant prior work, showing an LNA-control was not different to saline [31,33], control diabetic mice were administered saline rather than an LNA-control in the current study. In brief, the sequence of the LNA-control used in our prior miRNA studies [18,21,22] (5′-TcAtaCTatAtGaCA—3′ and 5′-TCATACTA—3′ (LNA uppercase, DNA lowercase) was checked against numerous databases and had no perfect match binding sites in the transcriptome [33]. The LNA-control was also validated in multiple assays (in vitro and in vivo) and shown not to differ from saline or untransfected/mock samples [31,33]. Three mice were administered LNA-antimiR-34a and 3 mice were administered saline at 10 weeks post diabetes induction due to an unexpected delay in the delivery of the LNA-antimiR-34a. However, this had no effect on knockdown of miR-34a gene expression (Supplementary Data Figure S1) and these mice were included in the study. After LNA oligonucleotide administration, mice were followed for a further eight weeks. A flowchart for the reporting of animal use, allocation, experimental analysis, follow up and exclusions following the CONsolidated Standards of Animal Experiment ReporTing (CONSAERT) guidelines as proposed by Weeks et al. [34] is provided in Supplementary Materials Figure S2.
Evaluation of hyperglycemia (>28 mmol/L) was performed every two to four weeks and at endpoint from blood collected from non-diabetic and diabetic mice. A tail vein prick was performed and glucose was measured using a glucometer (Accu-Chek® Performa, Roche, Basel, Switzerland) [28,29,30]. The upper detection limit of the glucometer is 33.3 mmol/L. Therefore, readings at or above this value (recorded as HI on the glucometer) were given the upper value of 33.3 mmol/L. At endpoint, blood was collected by cardiac puncture in a heparinized syringe and glycated hemoglobin (HbA1c) was assessed using a Cobas b 101 machine (Roche, Basel, Switzerland). The normal range for HbA1c levels is between 4% and 5.6%, and under 42 mmol/mol. Mice were considered diabetic if mouse HbA1c levels were >6.5% and >48 mmol/mol. The lower detection limit of the Cobas b 101 machine is 4% and 20 mmol/mol.
The miRCURY LNA™ microRNA Inhibitor (mmu-miR-34a) is LNA™-enhanced (LNA/DNA mixmer with ~50% LNA content) and contains a phosphorothioate backbone (Qiagen, Hilden, Germany). The LNA-oligonucleotide was purified and analyzed using anion-exchange high-performance liquid chromatography, desalted and lyophilized as a sodium salt, and the identity of the compound was confirmed by mass spectrometry. The LNA-oligonucleotide was resuspended in saline at 5 mg/mL, aliquoted and stored at −20 °C. The sequence for mmu-miR-34a antimiRNA is 5′A*G*C*T*A*A*G*A*C*A*C*T*G*C*C 3′ (Batch Number 630532 and Lot number 235893690), where the phosphorothioate bonds are indicated by an asterix (*).
To obtain measures of LV diastolic function, echocardiography was performed in anesthetized mice (ketamine/xylazine/atropine, KXA: 80/8/0.96 mg/kg, i.p.) at baseline, eight weeks following STZ administration (pre-LNA antimiRNA administration) and after 16 weeks of diabetes (eight weeks post LNA antimiRNA administration and before tissue harvest), using a Vevo 2100 High Frequency Ultrasound System (Visual Sonics, Toronto, ON, Canada). Once mice were anesthetized, fur was removed using a depilatory cream, wiped clean and mice then placed on a Visualsonics handling platform in a supine position. Acoustic coupling gel was placed on the chest area and images acquired using a MS550D transducer. Core temperature was monitored using a rectal probe and maintained at the physiological level (36–37 °C). Evaluation of diastolic function was performed by echocardiography using measurement of transmitral flow parameters including the early (E) and late (A) diastolic filling velocities, the E/A ratio, from an apical four chamber view with pulsed wave Doppler. Tissue Doppler imaging was also performed to obtain early (e’) and late (a’) diastolic mitral annular velocity (e’/a’ ratio) and E/e’ ratio. At completion of echocardiography, mice were administered atipamezole (0.2 mg/kg, subcutaneously) to aid with recovery post KXA anesthesia. All echocardiography imagery was acquired by a single operator and analyzed blinded offline using the Vevo Lab Software (Version 3.2.6, Visual Sonics). All data were independently validated [35].
At study endpoint, mice were euthanized by exsanguination in anesthetized mice (KXA: 100:10:1.2 mg/kg, i.p.). Tissues (heart, lung, kidney, liver, spleen) were excised whole, weighed and snap frozen in liquid nitrogen and stored at −80 °C. The atria were removed from the hearts and weighed. Hind legs were removed, submerged in 1M NaOH and incubated at 37 °C for 7–10 h to remove surrounding tissue. Tibias were cleaned and rinsed in H2O then air dried. Determination of tibia bone length was measured using a Vernier caliper.
Total RNA was isolated from frozen mouse tissues using TRI Reagent (Sigma-Aldrich, St Louis, MO, USA). Briefly, mouse tissue was homogenized in 500 μL of TRI Reagent and centrifuged to remove high molecular weight components. The supernatant containing RNA was collected and chloroform added, and samples vortexed. Following centrifugation with chloroform, the upper aqueous phase was recovered and RNA collected by isopropanol precipitation and rehydration. RNA was quantified using the Implen® NanoPhotometer NP80 (Implen, Munich, Germany).
For miRNA gene expression analysis, 50 ng of total RNA was reverse transcribed using the TaqMan® MicroRNA Reverse Transcription Kit (Life Technologies, Carlsbad, CA, USA) according to manufacturer’s recommendations. The reverse transcription protocol consisted of 30 min at 16 °C, 30 min at 42 °C, and 5 min at 85 °C. For gene expression analysis, 2 μg of total RNA was reverse transcribed using the High-Capacity cDNA Reverse Transcription Kit (Life Technologies, Carlsbad, CA, USA) according to manufacturer’s recommendations. The reverse transcription protocol consisted of 10 min at 25 °C, 120 min at 37 °C, and 5 min at 85 °C. Following cDNA synthesis, heart cDNA was diluted to a final concentration of 25 ng/uL.
MiRNA expression was assessed via RT-qPCR using an Applied Biosystems Quant Studio 7 Flex real-time PCR instrument. MiRNA expression was measured using the TaqMan® Universal Master Mix II, no UNG kit and TaqMan® MicroRNA Assays (20× primers; Life Technologies, Carlsbad, CA, USA; see Supplementary Materials, Table S1 for details) according to manufacturer’s recommendations. The PCR conditions consisted of 10 min at 95 °C; then 40 cycles of 15 s at 95 °C, and 60 s at 60 °C. Gene expression was normalized against snoU6 and data analyzed using the 2−ΔΔCt algorithm. For gene expression analysis, RT-qPCR was performed on an Applied Biosystems Quant Studio 7 Flex real-time PCR instrument. mRNA expression was measured using either the TaqMan® Fast Advanced Master Mix and TaqMan® Gene Expression Assays (Applied Biosystems, Foster City, CA, USA; 20X primer/probe mix, see Supplementary Materials, Table S2 for a list of assays) or SYBR™ Green PCR Master Mix and custom designed primers (Table S3) made by Integrated DNA Technologies (Coralville, IA, USA). The PCR conditions for TaqMan® fast assays consisted of 20 s at 95 °C, then 40 cycles of 1 s at 95 °C, and 20 s at 60 °C. The PCR conditions for SYBR™ Green Assays consisted of 2 min at 50 °C, 10 min at 95 °C, then 40 cycles of 15 s at 95 °C, and 1 min at 60 °C, and a melt curve stage of 15 s at 95 °C, 1 min at 60 °C and 15 s at 95 °C. All primers had a single peak in the melt curve and were 90–110% efficient on standard curves. Hypoxanthine phosphoribosyltransferase 1 (Hprt1) was used to standardize for cDNA concentration and all data were analyzed using the 2−ΔΔCt method of quantification.
Data was mined from a prior study in which transcriptomic data was collected from adult mouse hearts subjected to a control/sham surgery or trans-aortic constriction. Pure single cell fractions of cardiac myocytes, cardiac fibroblasts and endothelial cells were obtained from retrograde collagenase-II perfusion, followed by pre-plating and sorting techniques [36]. Here, we mined data from adult male mice following 2 weeks of sham surgery as a representation of miR-34a expression in different cell types within the adult mouse heart under basal conditions (publicly available dataset GSE66974).
Heart samples were fixed in 4% paraformaldehyde overnight and paraffin embedded for histological analysis (Gribbles Veterinary Pathology, Clayton, Australia). Cardiac collagen deposition/interstitial and perivascular fibrosis was assessed by Masson’s trichrome stain on 6 μm cross-sections (Gribbles Veterinary Pathology, Clayton, Australia). Images were acquired using a light microscope (BX43, Olympus, Center Valley, PA, USA) at 40× magnification. Fibrosis was quantified using Image-Pro Plus 7 by assessing the number of blue pixels (collagen/fibrosis). The percentage of total fibrosis was calculated by dividing the blue-stained collagen tissue by the total area of the LV (combined fibrotic area and pink/red stained healthy cardiac tissue). The percentage of perivascular fibrosis was calculated by dividing the blue-stained collagen tissue around vessels by the total area of the LV. Interstitial fibrosis was calculated by subtracting the percent of perivascular fibrosis from the percent of total fibrosis. For assessment of myocardial capillary density (angiogenesis), 4 μm cross-sections were deparaffinized in xylene and rehydrated in a graded ethanol series, rinsed in dH2O and then washed in 1X PBS. Slides were blocked in a FITC Protein Blocking Agent (Thermo Fisher, Scoresby, Australia) for 1 h at room temperature, then co-stained with Alexa Fluor 568-conjugated isolectin B4 (Invitrogen I21412, 1:10 dilution, overnight at 4 °C) and FITC-conjugated wheat germ agglutinin (WGA; Vector Labs FL1021, 1:50 dilution, 2 h at room temperature). Sections were mounted with ProLong™ Gold antifade reagent (Life Technologies, Carlsbad, CA, USA) and visualized under an Olympus BX61 fluorescence microscope at 20× magnification. For each heart, capillary density was calculated by dividing the number of capillaries by the number of cardiomyocytes. Image acquisition and analysis were performed blinded to disease and treatment group.
Statistical analyses were performed using GraphPad Prism (Version 8.1.2, San Diego, CA, USA). Results are presented as means ± SEM. Differences between groups were identified using one-way analysis of variance (ANOVA) followed by Fisher’s post hoc test or an unpaired t-test for comparing 2 groups alone, i.e., the non-diabetic and diabetic group. For blood glucose/HbA1c measurements, data were analyzed using a mixed-effects analysis with Fisher’s post hoc test, or a Kruskal–Wallis non-parametric one-way ANOVA with Dunn’s post hoc Test. For echocardiography parameters, differences between groups were identified using a two-way repeated measures ANOVA followed by Fisher’s post hoc test. A value of p < 0.05 was considered significant. All relative units are expressed as a fold change with the relevant control group normalized to 1.
Following STZ injection, but prior to administration of LNA-antimiR-34a or saline (8 weeks post-STZ. Figure 1A), mice within the diabetic groups had elevated blood glucose compared to non-diabetic mice (Figure 1B). Compared to diabetic mice administered saline, administration of LNA-antimiR-34a had no impact on blood glucose during the study (Figure 1B) or at endpoint (Figure 1C). HbA1c levels were also comparably elevated in the diabetic groups compared to non-diabetic mice (Figure 1D and Table S4). miR-34a expression was elevated in the hearts of diabetic mice administered saline compared with non-diabetic mice. In contrast, miR-34a was significantly lower in hearts of diabetic mice administered LNA-antimiR-34a compared to both diabetic mice administered saline and non-diabetic mice (Figure 1E). Specific cardiac cell types were not isolated from diabetic hearts in the current study, but on mining a publicly available dataset (GSE66974) we present data from the adult male mouse heart to show that miR-34a is expressed in cardiac myocytes, cardiac fibroblasts and endothelial cells (Figure 1F).
Diastolic dysfunction is a characteristic feature of the diabetic heart. Diastolic function was assessed in diabetic mice 8 weeks post-STZ, and prior to administration of LNA-antimiR-34a or saline (Figure 1A). Diabetic mice displayed diastolic dysfunction prior to treatment based on lower E/A and e’/a’ ratios, a recognized feature of the STZ diabetic model [37,38,39] (Figure 2A,B). At study endpoint, both ratios remained lower in diabetic mice versus non-diabetic mice, and treatment with LNA-antimiR-34a had no significant effect (Figure 2A–C, Table S5).
Diabetes was associated with an increase in normalized kidney weight, liver weight and spleen weight; consistent with a diabetic phenotype (Supplementary Materials, Table S4). These organ weights were not impacted by LNA-antimiR-34a treatment. There was no significant increase in heart weight in the diabetic model using one-way ANOVA (Table S4), but there was a small but significant increase in heart weight (unpaired t-test, p = 0.01, Figure 3A,B) and normalized heart weight when comparing the diabetic saline group to the non-diabetic group alone (unpaired t-test, p = 0.02, Figure 3C). Both of these modest increases in wet heart weight and HW/TL ratio were not apparent in the LNA-antimiR-34a treated diabetic mice compared to non-diabetic mice (unpaired t-test, p = 0.32 and p = 0.33, respectively, Figure 3B,C and Table S4). However, there were no significant differences between the two diabetic groups alone (unpaired t-test, p = 0.29 for heart weight and p = 0.36 for HW/TL, Figure 3B,C). Atrial weight was also elevated with diabetes, but inhibition of miR-34a had no impact (Table S4). Atrial natriuretic peptide (ANP) and B-type natriuretic peptide (BNP) ventricular gene expression were elevated with diabetes (Figure 3D,E), in accordance with our prior studies [37,38,39]. Inhibition of miR-34a had no impact on ANP expression, but appeared to have a modest effect on BNP ventricular gene expression. BNP was significantly elevated in the ventricles of diabetic saline mice versus non-diabetic mice (p = 0.002) but not in diabetic LNA-antimiR-34a treated mice versus non-diabetic mice (p = 0.06). There was no significant difference between the two diabetic groups (p = 0.075) (Figure 3E). There were no significant changes in gene expression of α-and β-MHC, and Serca2a (Supplementary Materials Figure S3A), or selected markers of oxidative stress, mitochondrial and inflammatory markers between diabetic mice administered saline or LNA-antimiR-34a (Supplementary Materials Figure S3B,C).
Sixteen weeks of diabetes was associated with a modest increase in total LV fibrosis and interstitial fibrosis (Figure 4A–C). LNA-antimiR-34a prevented a significant increase in both total and interstitial fibrosis, and total LV fibrosis tended to be lower in the LNA-antimiR-34a treated diabetic mice (p = 0.05) (Figure 4A–C). Perivascular fibrosis was not significantly different between groups by one-way ANOVA, but there was a mild increase based on the result from an unpaired t-test comparing the diabetic saline and non-diabetic groups alone (Figure 4D,E). Connective tissue growth factor (CTGF) gene expression was elevated in diabetic mice but not reduced by LNA-antimiR-34a treatment (Figure 4F). There were no significant changes in gene expression of collagen 1, collagen 3, matrix metalloproteinase-2 (MMP-2), or tissue inhibitors of metalloproteinases (TIMPs 1–4) (Supplementary Materials Figure S4). This may not be surprising given fibrosis in this diabetic cardiomyopathy model is mild in comparison to other cardiac models such as pressure overload due to aortic constriction (Supplementary Materials Figure S5).
A feature of the diabetic heart is reduced myocardial capillary density [5]. In the current study, capillary density was lower in hearts of diabetic mice administered saline compared to non-diabetic mice (p = 0.011, Figure 5A). Capillary density was also lower in LNA-antimiR-34a treated diabetic mice versus non-diabetic mice (p = 0.048), and not significantly different from diabetic saline mice (p = 0.378, Figure 5A). We also assessed the expression of genes known to play key roles in regulating vascular biology including vascular endothelial growth factor (VEGF [40]; target of miR-34a [41,42]) and miR-92a [42]. Vegfa and Vegfb gene expression were not significantly altered in diabetic hearts compared to non-diabetic hearts (Figure 5B). Expression of miR-92a was elevated in hearts from control diabetic mice (p = 0.004) but not LNA-antimiR-34a treated diabetic mice (p = 0.119), but there was no significant difference between the two diabetic groups (p = 0.061).
To explore potential reasons as to why inhibiting miR-34a provided limited protection in the diabetic heart, we first examined the expression of miR-34 target genes, some of which were previously shown to be regulated in settings of cardiac protection (vinculin (Vcl), Bcl2, Bcl6, cyclin D1, Notch1, Pnuts, Pofut1, Sema4b, Sirt1) [22,23]. Vcl which plays a critical role for protecting the heart against cardiomyopathy [43] tended to be greater in hearts from LNA-antimiR-34a treated diabetic mice than control diabetic mice (p = 0.08) and there was a significant inverse relationship between miR-34a expression with Vcl expression (Figure 6A). However, there was no obvious regulation of any other miR-34a target genes in the diabetic heart, besides Notch1 which was unexpectedly decreased rather than elevated (Figure 6B). Next, we assessed the expression of miR-34b and miR-34c. In models of severe cardiac pathology in which antimiR-34a provided little or no protection, we identified upregulation of miR-34b and miR-34c [18,21,22]. In the current study, there were no significant differences in miR-34b or miR-34c assessed by one way ANOVA. However, when comparing expression of miR-34b and miR-34c in the diabetic saline group versus the non-diabetic control group alone with an unpaired t-test, miR-34b and -34c were elevated (p = 0.026 and p = 0.047, respectively) (Figure 6C).
The major goal of this study was to assess whether therapeutically targeting miR-34a in the diabetic heart with established diastolic dysfunction would improve diastolic function and prevent cardiac fibrosis. To our knowledge, this is the first study to inhibit miR-34a in the diabetic heart in vivo. The key findings to come from this study are summarized in Figure 7A. Despite, elevation of miR-34a in the diabetic heart, inhibition of miR-34a had no significant impact on diastolic function, and only modestly attenuated cardiac enlargement, fibrosis, and ventricular gene expression of BNP, and an anti-angiogenic miRNA (miR-92a). One miR-34a target, vinculin, was inversely correlated with miR-34a expression in the diabetic heart, but several other miR-34a targets were unchanged. In previous work from us and other investigators, it had been shown that inhibiting miR-34a provided cardiac protection in settings of moderate cardiac pathology including aging [23], a model of moderate pressure overload-induced hypertrophy [18], acute myocardial infarction [23], and dilated cardiomyopathy [21]. However, we also showed that inhibiting miR-34a in more severe cardiac models of chronic myocardial infarction [22], severe pressure overload [18], and dilated cardiomyopathy with atrial fibrillation [21] provided little or no advantage. These severe cardiac models displayed significantly more fibrosis and systolic dysfunction than the models of moderate cardiac pathology. It was concluded that in more severe heart disease models, inhibiting miR-34a alone was inadequate to attenuate or reverse established severe pathology. The STZ model of diabetic cardiomyopathy used in the current study was considered to represent a mild-to-moderate model of cardiac pathology since it is associated with diastolic dysfunction, moderate cardiac fibrosis and modest elevation of ventricular BNP and/or ANP [37,38,39]. Thus, we hypothesized that silencing miR-34a in the diabetic heart would provide substantial benefit. A number of factors may explain why inhibition of miR-34a was unable to restore diastolic function in the diabetic heart, and only had a limited or no impact on other features including hypertrophy, fibrosis, capillary density, and cardiac gene expression. First, diabetes represents a systemic disorder impacting multiple organs/systems within the body. By contrast, in our prior studies in which inhibiting miR-34a was able to improve systolic function, we used models in which the heart was specifically targeted, i.e., surgical interventions targeting the heart alone or cardiac-specific transgenic mouse models. The systemic nature of diabetes is associated with the dysregulation of multiple miRNAs in numerous tissues and circulation, as well as inter-organ cross talk [44]. In the current study, kidney, spleen and liver weights were elevated with diabetes. Thus, targeting diabetes with miRNA-based drugs may be more challenging than targeting pathologies which are more limited to one tissue type. Second, targeting a single miRNA in the diabetic heart may be insufficient to improve heart function in a setting of established dysfunction. In our prior work, we found that miR-34b and miR-34c were elevated in addition to miR-34a in settings of substantial pathology due to severe pressure overload or chronic myocardial infarction [22,45], and that inhibiting the miR-34 family (i.e., miR-34a, -34b and -34c) attenuated pathology/dysfunction but inhibiting miR-34a alone did not [18,22]. Both miR-34b and miR-34c were elevated in hearts of diabetic mice versus non-diabetic mice, but not in hearts from LNA-antimiR-34a treated mice. Thus, any additional protection from targeting the miR-34 family in the diabetic heart is considered to be minimal. Rather, it may be necessary to target other miRNAs which play key functional roles in the diabetic heart such as miR-133a, miR-195, miR-30c, miR-200b and miR-15a/b [46,47,48]. A profiling study on LV samples from an STZ-induced diabetic mouse model showed that over 300 miRNAs were differentially regulated and over 200 were upregulated by greater than 2-fold [49]. This may explain, at least in part, why LNA-antimiR-34a treatment in the current study did not regulate several miR-34a targets genes known to contribute to cardiac protection including Bcl2, Pnuts, Sirt1, Vegfa and Vegfb. For example, of the upregulated miRNAs described in previous diabetic studies, a number including miR-195, miR-483-3p, miR-210, miR-451, miR-181c, miR-1183, miR-145, miR-199, miR-137 and miR-499 have been shown to regulate mRNAs that are recognized targets of miR-34a including Bcl2 and Sirt1 [27,50,51,52,53,54,55,56,57,58] (Figure 7B). It is therefore probably not surprising that in a setting of diabetes in which multiple miRNAs are upregulated and targeting the same mRNAs, inhibiting miR-34a alone had no significant impact on the regulation of some target genes (Figure 7B). This may also explain a prior report in the human diabetic heart, in which one miR-34a target was regulated but another two targets were not [27]. Third, we previously demonstrated that anitmiR-34 could regulate the expression of other miRNAs in addition to miR-34 by binding to miR-34 direct target genes which also play roles as transcription factors, and hence regulating other miRNAs with other targets [59]. A relevant example to the diabetic heart is that antimiR-34 differentially regulated miR-15a and miR-15b [59], which have been implicated in playing a role in the diabetic heart [48]. Finally, the timing of intervention may also be important. We chose to intervene at a time point with established cardiac dysfunction. Katare and colleagues showed that circulating and cardiac miR-34a were elevated in type 2 diabetic patients prior to any obvious systolic or diastolic dysfunction. This suggests that miR-34a is upregulated in the early stages of disease [27]. Thus, an earlier intervention may be more advantageous in attenuating diastolic dysfunction. However, in a clinical setting, an intervention would need to be commenced based on elevated circulating levels of miR-34a rather than the presence of any cardiac dysfunction. This may be challenging given that circulating levels of miR-34a can be altered by other disease states and conditions including chronic migraine [60] and non-alcoholic fatty liver disease [61]. While silencing miR-34a was unable to restore diastolic dysfunction in a setting of established dysfunction, there was evidence that inhibition of miR-34a had the potential to attenuate diabetes-induced total and interstitial cardiac fibrosis. Consistent with this result, a previous in vitro study demonstrated that a miR-34a inhibitor could suppress TGFβ-induced proliferation of rat cardiac fibroblasts [62]. We also showed that antimiR-34a attenuated fibrosis in a mouse model with dilated cardiomyopathy [21]. However, of note, other miRNAs have been implicated in regulating fibrosis in the diabetic heart including miR-133a [63], miR-200b [64], and miR-15a/b [48]. Inhibiting miR-34a also appeared to have a modest impact on preventing diabetes-induced cardiac enlargement and BNP expression, but no effect on ANP gene expression. This could be in part because two miRNAs shown to play a key role in repressing ANP mRNA in human embryonic stem cell-derived cardiomyocytes (miR-425 and miR-155; [65]), were significantly decreased in the STZ-induced diabetic mouse heart [49]. In the current study, myocardial capillary density was reduced in the diabetic heart, regardless of treatment. At the molecular level, miR-92a was significantly elevated in hearts of diabetic control mice but not diabetic antimiR-34a treated mice, but other regulators of vascular biology (Vegfa and Vegfb) were unchanged. miR-34a and miR-92a are recognized to have anti-angiogenic properties via the down-regulation of a shared target, Sirt1 [66,67,68]. MiR-34a was also shown to promote vascular smooth muscle cell calcification in the aorta of mice, and miR-34a knockout mice were protected [66]. The absence of any significant vascular protection in diabetic antimiR-34a treated mice may be related to the absence of any significant changes in Sirt1, Vegfa and Vegfb. Collectively our data and previous literature suggest a drug targeting multiple miRNAs may offer more protection. In reviewing the literature, there is limited evidence of any miRNA-based therapeutic or intervention restoring or improving cardiac function of the diabetic heart, i.e., assessment of an intervention in a setting of established diastolic dysfunction. Examples of miRNAs considered to represent therapeutic targets have included miR-133a, miR-195, miR-30c, miR-200b and miR-15a/b [46,47,48]. However, within these studies, restoration of miRNAs were undertaken in in vitro settings rather than in vivo [48,69], or for the in vivo assessments, the miRNA intervention (therapeutic/transgenic) occurred simultaneously with the pathology (i.e., a prevention study), or it was unclear whether there was established cardiac dysfunction prior to treatment/miRNA intervention [50,64,70,71]. There has been intense interest in developing miRNA-based drugs for various diseases including the diabetic heart [46,47]. To be successful in a diabetic setting with established cardiac dysfunction and fibrosis, it is likely that targeting multiple miRNAs will be required. In developing strategies targeting multiple miRNAs, the role of miRNAs in other tissue and cell types will also need to be considered. For example, a therapy targeting miR-34 was used in a clinical trial for liver cancer [72]. However, in the context of cancer, the strategy is to increase miR-34 rather than silence it. Given miR-34 and other miRNAs are ubiquitously expressed, it will be important to target organs or cell types specifically, for example with antibody or vector-based approaches [12,73]. The application of LNA-based antimiR therapies in preclinical animal models has been established in multiple disease settings [74,75,76], and recently in a setting of diabetes-induced heart disease and cardiac dysfunction [77]. Several studies have demonstrated efficacy and safe use of LNA-based antimiR therapies in human clinical trials, including miravisen (targeting the liver expressed miR-122 for the treatment of hepatitis C [13]), cobomarsen (an inhibitor of miR-155 in patients with T-cell lymphoma [78]), and MRG-110 (an inhibitor of miR-92a-3p, which is known to have a role in cardiovascular disease and wound healing [79]). However, delivery of LNA-antimiRs to the heart still poses some challenges. Further strategies to improve specificity, uptake and targeted delivery of antimiR therapies will be required [76,80,81]. One emerging technology to deliver miRNA inhibitors to heart tissue is the use of ultrasound and a microbubble-targeted delivery system [82]. This has proven to be an effective technique to deliver LNA-antmiR-23a to the heart in a mouse model of pathological cardiac hypertrophy [82]. Limitations and future work: There are some limitations to our work which could form the basis of future work. Firstly, we showed that miR-34a was elevated in hearts of diabetic mice and reduced in hearts of LNA-antimiR-34a treated diabetic mice, and that miR-34a is expressed in mouse cardiac myocytes, cardiac fibroblasts and endothelial cells, but we did not assess whether LNA-antimiR-34a reduced miR-34a in each of these cell types. Second, we did not measure the level of circulating miRNA-34a in the plasma of non-diabetic and diabetic mice. Circulating miRNAs have been proposed as biomarkers in the prognosis of cardiac disease, as they reflect the progression and severity of the disease state. A previous study demonstrated a correlation between cardiac and circulating miR-34a levels in human diabetic hearts, and suggested the elevated circulating miR-34a may be of cardiac origin [27]. Thirdly, we did not investigate the effect of miR-34a inhibition on proliferation. In cancer cells, it is well recognized that miR-34a inhibits cellular proliferation [83]. In the hearts of neonatal mice, overexpression of miR-34a limited cardiomyocyte proliferation and regeneration following cardiac injury [20]. In a diabetic setting, in vitro studies suggest that inhibition of miR-34a may have differential effects depending on cell type. Inhibition of miR-34a decreased high glucose induced cell death in cultured human adult cardiomyocytes, whereas in human cardiac progenitor cells isolated from the diabetic heart, inhibition of miR-34a reduced proliferation via a mechanism involving the miR-34a targets, cyclin D1, Bcl2 and Sirt1 [27]. While, we did not directly assess the effect of miR-34a inhibition on proliferation in the diabetic heart, the expression of cyclin D1, Bcl2 and Sirt1 were unchanged by LNA-antimiR-34a treatment. Finally, when cardiac fibrosis is mild, as observed in our model of diabetic cardiomyopathy, it will be challenging to identify significant changes in any one gene with interventions. To explore the potential antifibrotic actions of antimiR-34a or other miRNAs it would be advisable to study a model with more fibrosis.
In summary, despite LNA-antimiR-34a preventing a diabetes-induced increase in miR-34a in the mouse heart, inhibition of miR-34a was unable to improve diastolic function in a setting of established diastolic dysfunction, and only had a modest effect on attenuating diabetes-induced cardiac enlargement, cardiac fibrosis and elevation of ventricular BNP. Since the expression of a number of miRNAs are elevated in the diabetic heart and share target mRNAs with miR-34a, strategies targeting multiple miRNAs and/or earlier intervention is likely to be required to achieve significant functional benefit. | true | true | true |
|
PMC9563637 | Biting Cao,Hongfeng Wang,Jinjuan Bai,Xuan Wang,Xiaorong Li,Yanfeng Zhang,Suxin Yang,Yuke He,Xiang Yu | miR319-Regulated TCP3 Modulates Silique Development Associated with Seed Shattering in Brassicaceae | 01-10-2022 | TCP3,miR319,Arabidopsis,rapeseed,silique shattering | Seed shattering is an undesirable trait that leads to crop yield loss. Improving silique resistance to shattering is critical for grain and oil crops. In this study, we found that miR319-targeted TEOSINTE BRANCHED 1, CYCLOIDEA, and PROLIFERATING CELL NUCLEAR ANTIGEN BINDING FACTOR (TCPs) inhibited the process of post-fertilized fruits (silique) elongation and dehiscence via regulation of FRUITFULL (FUL) expression in Arabidopsis thaliana and Brassica napus. AtMIR319a activation resulted in a longer silique with thickened and lignified replum, whereas overexpression of an miR319a-resistant version of AtTCP3 (mTCP3) led to a short silique with narrow and less lignified replum. Further genetic and expressional analysis suggested that FUL acted downstream of TCP3 to negatively regulate silique development. Moreover, hyper-activation of BnTCP3.A8, a B. napus homolog of AtTCP3, in rapeseed resulted in an enhanced silique resistance to shattering due to attenuated replum development. Taken together, our findings advance our knowledge of TCP-regulated silique development and provide a potential target for genetic manipulation to reduce silique shattering in Brassica crops. | miR319-Regulated TCP3 Modulates Silique Development Associated with Seed Shattering in Brassicaceae
Seed shattering is an undesirable trait that leads to crop yield loss. Improving silique resistance to shattering is critical for grain and oil crops. In this study, we found that miR319-targeted TEOSINTE BRANCHED 1, CYCLOIDEA, and PROLIFERATING CELL NUCLEAR ANTIGEN BINDING FACTOR (TCPs) inhibited the process of post-fertilized fruits (silique) elongation and dehiscence via regulation of FRUITFULL (FUL) expression in Arabidopsis thaliana and Brassica napus. AtMIR319a activation resulted in a longer silique with thickened and lignified replum, whereas overexpression of an miR319a-resistant version of AtTCP3 (mTCP3) led to a short silique with narrow and less lignified replum. Further genetic and expressional analysis suggested that FUL acted downstream of TCP3 to negatively regulate silique development. Moreover, hyper-activation of BnTCP3.A8, a B. napus homolog of AtTCP3, in rapeseed resulted in an enhanced silique resistance to shattering due to attenuated replum development. Taken together, our findings advance our knowledge of TCP-regulated silique development and provide a potential target for genetic manipulation to reduce silique shattering in Brassica crops.
Angiosperms show various types of siliques, most of which are derived from the ovary walls and fertilized ovules. Their structures are conserved in Arabidopsis thaliana and some species of the Brassicaceae family, including cabbage, broccoli, Chinese cabbage (Brassica rapa), and rapeseed (Brassica napus) [1]. In Arabidopsis thaliana, siliques are composed of fertilized ovules and three major regions, namely the valves, repla, and valve margins. The valve margins were formed between the valves and the repla, and two types of cells at the valve margins regulated the silique opening. The separation layer is also called dehiscence zone (DZ), which is composed of 2–3 layers of thin-walled cells and separates the heavily lignified cells of the pericarp edge from the replum. Fruit dehiscence is accompanied by degradation of the thin-walled cells [2,3,4,5]. Canola seeds, which are harvested for oil, are often lost owing to premature silique shattering, particularly under adverse weather conditions, and premature dehiscence or silique shattering leads to significant crop losses [6]. Reducing silique shattering will increase the proportion of seed obtained during harvest, which is conducive to economic income. Several key genes contributing to silique development have been identified in Arabidopsis. REPLUMLESS (RPL) plays an important role in replum development. rpl mutants exhibit reduced replum width, and the strong alleles display a complete absence of outer replum [7]. SHATTERPROOF1 (SHP1) and SHP2 are necessary for the proper valve margins development. The loss of SHP1 and SHP2 function results in lacking the lignified and separation layers in siliques, which prevents silique opening [8]. INDEHISCENT (IND) and ALCATRAZ (ALC) work downstream of SHP [9,10]. The atypical bHLH protein IND is required for seed dissemination [9]. The small cells of the separation zone and the adjacent lignified cell layers were defective in the siliques of the ind mutant, and mutation of ALC resulted in the lack of the non-lignified cell layer in the separation layer. NO TRANSMITTING TRACT (NTT) encodes a zinc finger transcription factor, and loss of NTT function affects the normal transmitting-tract development [11] and replum development in Arabidopsis fruits [12]. In addition, the siliques of activation-tagged allele of NTT (ntt-3D) are indehiscence and almost lack separation and lignification layer [13]. The cells in the mesophyll tissue layer of fruitfull (ful) mutant are lignified at the later stage of fruit development and are much smaller than in the wild type [14,15]. Overall, the FUL-SHP network play important role in fruit morphology, which is evolutionally conserved in plants [16]. Brassica species are the most closely related crops to Arabidopsis, and they disperse seeds in a similar way [17,18]. Ectopic expression of the Arabidopsis AtFUL gene under control of the Cauliflower Mosaic Virus 35S promoter in Brassica juncea leads to complete shattering-resistant siliques similar to those of 35S::FUL transgenic plants [19,20]. JAGGED (JAG) is involved in maintaining the integrity of boundaries between cell groups with indeterminate or determinate fates. BnJAG-33 mutants enhance silique shatter resistance [21]. BraA.IND.a and BolC.IND.a genes control valve margin cell fate and inhibit replum formation [22]. BnIND mutations in B. napus show higher shatter resistance [23,24]. Taken together, these conserved genetic regulators of silique development can be applied in Brassica breeding for improving seed shattering resistance. The plant-specific TEOSINTE BRANCHED 1, CYCLOIDEA, and PROLIFERATING CELL NUCLEAR ANTIGEN BINDING FACTOR (TCP) family with a bHLH motif that allows DNA binding and protein–protein interactions are involved in growth-related progress, such as branching, floral organ morphogenesis and leaf development [25,26]. There are 24 members of the TCP family in Arabidopsis thaliana, and TCP2, TCP3, TCP4, TCP10, and TCP24 are the targets of miR319 [27]. miR319-regulated TCP genes function in leaf development by regulating cell division arrest [28,29]. In a previous work, we reported that an anther with four microsporangia was transformed into the one with two microsporangia in the jaw-D mutant in which the MIR319a gene is activated [30]. However, the role of miR319-regulated TCPs in fruit development remains unclear. In this study, we revealed that the siliques replum of jaw-D plants, in which the MIR319a gene was activated, was significantly enlarged, and AtTCP3 was one of the major regulators in against to silique elongation and dehiscence via increasing FUL expression. Finally, hyper-activation of BnTCP3.A8 in rapeseed increased silique resistance to shattering. These results uncovered that miR319-regulated TCPs played viral role in silique development and can be used as a genetic editing resource for seed shattering resistance in Brassica crops.
The Arabidopsis thaliana Col-0 ecotype was used as the wild type in this study. Seeds of Col-0, jaw-D mutants and transgenic Arabidopsis lines were sterilized using 70% (v/v) ethanol. plants were grown on Murashige and Skoog (MS) plates with 1% sucrose under long-day conditions (16-h light/8-h dark) and then transferred to a growth room at 22 °C under long-day conditions. The rapeseed accession K407 was acquired from Hybrid Rape Research Center of Shaanxi Province [31]. Rapeseed wild-type (accession K407) and transgenic rapeseed lines overexpressing BnTCP3.A8 were sown in a greenhouse. Then, the seedlings were transplanted into a field at the Songjiang Farm Station of the Shanghai Institute of Plant Physiology and Ecology in early September.
To generate the mAtTCP3-overexpression construct, the full-length CDS of AtTCP3 (AT1G53230) was amplified using specific primers. Then, a silent mutation was introduced into the miR319 target site of the AtTCP3 CDS to generate the mAtTCP3 construct using a Site-Directed Mutagenesis Kit (TaKaRa). The chimeric pAtTCP3::AtTCP3SRDX was constructed as described previously [32]. To generate the p35S::AtFUL construct, the full length CDS of the AtFUL gene was amplified. To generate the p35S::BnTCP3.A8 construct, the full length CDS of the Brassica napus TCP3 gene (Acc. number GSBRNA2T00114181001) was amplified and cloned into the binary vector pChimera. The Agrobacterium tumefaciens strain GV3101 pMP90RK was used for stable plant transformations. All the primers used for plasmid construction are listed in Table S1. For rapeseed hypocotyl transformations, we followed the protocol described by Liu et al. and Moloney et al. [33,34]. The p35S::mAtTCP3, pAtTCP3:: AtTCP3SRDX and p35S:: AtFUL vectors were transferred into Col-0 using the floral dip method [35]. The p35S::BnTCP3.A8 and p35S::mAtTCP3 transgenic plants were selected on MS medium containing 50 mg/mL kanamycin after Agrobacterium-mediated transformation. Positive seedlings (T0) were transplanted into soils. In the F1 generation, all transgenic plants were confirmed using PCR with gene-specific primers and maintained in a growth room at 22 °C under long-day conditions. Primers used in this study are listed in Supplementary Table S1.
Total RNA was extracted from plants using TRIzol reagent. Before performing quantitative real-time PCR, RNA was treated with DNase I (TaKaRa) to avoid DNA contamination, and then RNA was purified with the phenol chloroform. Real-time detection of all target genes was based on the method of Li et al. (2020) [36]. The quantification of the relative expression levels was performed as reported previously [37]. The relative transcript levels of each gene in Arabidopsis thaliana and B. napus were normalized to that of Actin [36] and UBC21 [38] for quantitation, respectively. Primer Premier 5 software was used to design oligonucleotide primers (http://www.premibiosoft.com (accessed on 1 September 2022)) for real-time PCR to amplify the reference and target genes. All the primers used in this study are shown in Table S1.
For toluidine blue staining, the siliques were harvested one week after opening of flower buds and fixed in FAA for more than 18 h, sectioned and stained with 0.1% toluidine blue [7,8,39]. For phloroglucinol lignin staining, the siliques at the same stage were stained with 2% phloroglucinol solution overnight. Then, the siliques were decolorized in ethanol and photographed.
Silique (stage 10–12) sections in the wild-type flowers were prepared using previously described pretreatment and hybridization methods [40]. The primers used to generate hybridization probes specific for AtTCP3 using the fragments (513 bp) in coding sequences according to the method of Wang et al. (2015) [30]. LNA (Locked Nucleic Acid) modified probes specific for AtTCP3 were synthesized and labeled with DIG at the 3′-end by TaKaRa.
Siliques (stage 17) of Arabidopsis and B. napus were harvested and fixed overnight at 25 °C in FAA, dehydrated through an ethanol series, and critical-point dried. Samples were sputter-coated with gold and viewed using a Hitachi S-2300 electron microscope. Replum width were measured in the middle of siliques photographed according to the method of Marsch-Martinez, N. et al. (2014) [12].
Mature siliques were collected for the measurement of silique shattering resistance by a random impact test (RIT) [41]. In total, 20 siliques were preserved in a mesh bag overnight after incubation at 60 °C for 60 min to equilibrate the moisture content, and then, they were placed into a cylindrical container, having a diameter of 6 cm and a height of 17 cm, which was pre-loaded with 12 steel balls with a diameter of 13 mm. On a horizontal shaker, the container was shaken at 300 rpm and then the numbers of cracked siliques were recorded after 1 min shaking intervals. A total of 10 recordings were taken. After each recording, the broken siliques were removed, and all the experiments were performed in triplicate. The silique shattering-resistance index (SRI) was calculated using the following equation: where Xi represents the number of cracked siliques in time ith, with 1 ≤ i ≤ 10.
Arabidopsis AtTCP protein sequences were downloaded from the Arabidopsis Information Resource (https://www.arabidopsis.org/ (accessed on 1 September 2022)). BnTCP3.A8 and Brp.TCP3 were selected based on their high similarity with AtTCP3. Full-length amino acid sequence multiple alignments were performed using ClustalW and GeneDoc. Unrooted phylogenetic trees were constructed from the aligned amino acid sequences using the neighbor-joining method in MEGA 6.0, and bootstrapping was carried out with 1000 iterations [42].
We calculate statistical significance using two-tailed Student’s t-test and error bars indicate standard error (SE). Values of p < 0.05 are statistically significant.
In previous studies, the development defect of jaw-D mutant plant with activation of AtMIR319a was observed in both vegetable and reproductive stages (Wang et al., 2015). However, the molecular mechanism of silique defect in jaw-D plant has not been uncovered. Using scanning electron microscopy (SEM), we found that the repla of the jaw-D silique was wider than those of the wild type in the base, middle, and tip of the mature silique (Figure 1A). Transverse sections of siliques showed that the jaw-D repla were significantly wider compared with wild-type repla (Col-0) in the middle of mature siliques (Figure 1B,C). Thus, these results indicated that activation of AtMIR319a enlarged repla in siliques. To investigate the expression changes of the five miR319-targeted TCP genes in different tissues of jaw-D mutant plants, real-time PCRs were performed to analyze the expression levels of AtTCP2, AtTCP3, AtTCP4, AtTCP10, and AtTCP24 in cauline leaves, inflorescences, flowers, and siliques. In cauline leaves and inflorescences, the expression of AtTCP2, AtTCP3, AtTCP4, AtTCP10, and AtTCP24 were consistently reduced to a similar level in jaw-D mutants compared to the wild type (Figure 2A). In contrast, their expressions were differently suppressed in flowers (stage 14–15) and siliques (stage 17). Specifically, AtTCP3 was mostly suppressed among all the five TCP genes in silique of jaw-D plants (Figure 2A). Detailed analysis of AtTCP3 expressional level in different plant tissues confirmed that this gene was most accumulated in siliques as compared to in rosette leaves, inflorescences, and flowers (Figure 2B). Given the siliques of Arabidopsis thaliana originated from gynoecium with two fused carpels, we further investigated the mRNA abundance of AtTCP3 during early flowering stage by in situ hybridization. In young flower buds of wild-type plants, AtTCP3 transcripts were highly accumulated in the developing carpel (Figure 2C). These results suggested that AtTCP3 was one of the major regulators of silique development among these five miR319 targeted TCPs. Notably, TCP4 and TCP24 were also significantly repressed in silique of jaw-D plants compared to Col-0 (Figure 2A). Knockout mutants of AtTCP3 showed no visible phenotypic alterations due to the functional redundancy of the miR319-target TCPs [43,44]. To further confirm the function of AtTCP3 in repla development, we constructed the pAtTCP3::AtTCP3SRDX transgenic line fusing TCP3 with plant-specific ERF-associated amphiphilic repression (EAR) motif repression domain (SRDX), which suppressed target genes of endogenous TCP3 and its functionally redundant TCPs. We found that the replum width of the pAtTCP3::AtTCP3SRDX siliques was similar to jaw-D siliques (Figure 1C). These results further suggested that miR319-regulated TCPs played critical role in replum development in siliques.
To further study the function of TCP3 in developing siliques, a miR319a-resistant version (p35S::mAtTCP3) of AtTCP3 was introduced into Arabidopsis. As expected, the AtTCP3 transcription level was significantly increased in the transgenic plants compared with wild-type (Figure 3A). In wild-type siliques, valve margins or DZ were differentiated as a narrow strip consisted of few cells within the separation layer and the lignified regions, both of which contributed to the active fruit-opening process [19]. Consistent with the previous studies [43], p35S:mAtTCP3 plants were smaller than wild-type plants (Figure 3B), and their siliques were much shorter (Figure 3C–E). Based on the SEM image of siliques (Stage 17), the repla of p35S:mAtTCP3 siliques were much narrower than those in wide type and jaw-D siliques (Figure 1A). Cross sections in the middle of siliques also showed that the replum cells were relatively smaller in p35S:mAtTCP3 siliques than in wild type and jaw-D. In addition, a layer of large and sparse cells were seen between the valve and the replum in the Col silique, which facilitated the dehiscence of the siliques, whereas this layer of cells was much denser in p35S::mAtTCP3 siliques (Figure 1B). These results suggested that AtMIR319a-regulated TCP3 functioned as a negative regulator of silique development. Silique wall thickness and replum development are correlated with lignin accumulation in plants [5]. To investigate whether TCP3 play a role in regulating replum lignification, phloroglucinol, a lignin-specific histological stain, was applied to siliques of wild-type, p35S::mAtTCP3, pAtTCP3::AtTCP3SRDX, and jaw-D plants for evaluating the degree of tissue lignification. In the wild-type siliques, lignin-specific signals were clearly seen in the repla and valve margin cells adjacent to the DZ throughout the siliques, while the signals were very faint in the repla and valve margin cells of p35S::mAtTCP3 siliques (Figure 3F). In contrast, the lignin-specific signals were much stronger in the repla and valve margin cells of jaw-D and pAtTCP3::AtTCP3SRDX siliques compared to those in wide type. These results indicated that AtTCP3 overexpression attenuated lignification of repla in the siliques.
Recent molecular and genetic studies in Arabidopsis have identified several crucial genes involved in the regulation of silique development [45]. To determine the relationship of TCP3 with those valve- and valve-margin-related genes, the expressional level of AtFUL, AtNTT, AtIND, AtALC, AtJAG, AtFIL, AtYAB3, and the replum-related gene AtPRL, were analyzed using real-time PCR. Transcript levels of AtFUL and AtRPL were significantly elevated in the p35S::mAtTCP3 plants as compared to wild-type (Figure 4A). In addition, transcript levels of AtFUL, AtIND, AtRPL, and AtJAG were significantly downregulated in the jaw-D mutants compared to those in the wild type (Figure 4A). In combination, these results implied that AtTCP3 acted as an upstream regulator of AtFUL to suppress the differentiation of the replum. To further understand the genetic relationship between AtTCP3 and AtFUL, the AtFUL gene were either overexpressed or suppressed in jaw-D mutant background. In FUL jaw-D plant overexpressing FUL, silique repla were narrower in width than that of jaw-D silique but still wider than the wild type, indicating that the wide replum phenotype of jaw-D mutant was partially complemented by AtFUL (Figure 1C and Figure 4B). Besides, the wavy leaf margins phenotypes observed in jaw-D were also restored by AtFUL overexpression in FUL jaw-D plants (Figure 4C). Moreover, ful jaw-D double mutants showed similar repla width with ful mutant (Figure 1C and Figure 4D). Together, these results suggest that AtFUL acted downstream of TCP genes.
Rapeseed and Arabidopsis are members of the Brassicaceae family and display similar silique morphologies. To investigate whether TCP3 is also involved in replum development of rapeseed, we identified the B. napus TCP3 gene by blast searching for the AtTCP3 protein coding sequence in the Brassica database (BRAD). The obtained locus, GSBRNA2T00114181001, which located on chromosome A08, showed the highest similarity to AtTCP3. Thus, this gene was named as BnTCP3.A8 in following studies (Figure 5A, Figure S1A,B). Further sequence alignment of this gene in the same database revealed that BnTCP3.A8 displayed a 98% similarity to B. rapa BrpTCP3 (Figure 5A). BnTCP3.A8 also contained a conserved BrpmiR319a-targeted site within the identified sequences (Figure 5B). To study the functions of BnTCP3.A8 in silique development, the BnTCP3.A8 CDS was introduced into rapeseed plants (accession K407) under the regulation of the cauliflower mosaic virus 35S promoter. The expression of BnTCP3.A8 was significantly upregulated in p35S::BnTCP3.A8 transgenic lines T2, T4 (T2 generation), and T26 (T3 generation) (Figure 6A,B). In those plants, silique lengths of the p35S::BnTCP3.A8 line were shorter than that of the wild-type (Figure 6C,D). For immature gynoecium [46], the placenta development of p35S::BnTCP3.A8 line was weaker than that of the wild-type (Figure 6E). Further, the image of SEM showed that the replum in T26 silique was much narrower than that of the wild-type (Figure 6F). We measured the replum width in the middle of siliques. Replum width of T26 silique was significantly narrower than that of the wild-type (Figure 6G). These results indicated that BnTCP3.A8 negatively regulated replum development in rapeseed, consistent with the role of AtTCP3 in Arabidopsis. Spence et al. (1996) showed that silique shattering resistance was negatively associated with the degree of lignification in the valves, repla, and valve margin cells [18]. To investigate whether the p35S::BnTCP3.A8 plants also reduced lignification in repla, the siliques of both the wild-type and transgenic lines were stained by phloroglucinol. The siliques of T26 showed weaker lignin-specific signals in the repla than those of the wild-type (Figure 6H). In addition, the siliques of T26 exhibited smaller valve margin cells adjacent to the DZs (Figure 6I). These observations indicated that the reduced lignification in the repla was responsible for the resistance of p35S::BnTCP3.A8 plants to silique shattering.
To determine whether BnTCP3.A8 is also involved in the regulation of replum- and valve-related genes as AtTCP3, we analyzed expression levels of BnFUL and BnSHP1, BnALC, BnJAG, BnFIL, and BnYAB3 genes in T26 siliques. Real-time PCR showed that BnFUL and BnSHP1 were upregulated in p35S::BnTCP3.A8 siliques, while BnALC, BnJAG, BnFIL, and BnYAB3 were downregulated (Figure 6J). These results revealed that the TCP3-mediated gene regulatory pathways were conserved in rapeseed during silique development. To define whether the deficiency in replum width of rapeseed affects silique shattering, we determined the silique shattering ratios using the modified method of Bruce et al. (2002) [47]. The siliques of the wild-type plants were opened easily by shattering treatment and released many seeds (Figure 6K); however, the siliques of T26 plants were less opened than the wild-type and released fewer seeds. The SRI (shattering-resistance index) of T26 was 0.58 at 300 rpm, which was significantly high than the SRI of wild-type (SRI = 0.48) (Figure 6L). Thus, these results suggested that overexpression of BnTCP3.A8 enhanced silique shattering resistance.
The miR319-targeted TCP genes control cell division arrest [25,27,28,48,49,50]. In general, proteins encoded by genes expressed in the replum often negatively regulate genes expressed in the valves [1]. We found that these TCP genes were involved in development of repla because repla became wider in jaw-D mutant. Among them, AtTCP3 was down-regulated mostly among these TCP genes. In situ hybridization revealed that AtTCP3 was preferentially expressed in middle region of carpel. In addition, p35S::mAtTCP3 reduced the replum width, and pTCP3::mAtTCP3SRDX increased the replum width. Taken together, these results suggested that AtTCP3 functioned in regulation of replum development. The factors that control the establishment of medio-lateral silique patterns also regulate proper shoot development and leaf formation [51,52,53,54,55]. We found that AtTCP3 positively regulated valve- and valve margin-related genes AtFUL and AtSHP1. Genetic analysis showed AtFUL acted downstream of miR319a-targeted AtTCP3 in replum deficiency. This evidence indicates that TCP3 functioned in repressing replum enlargement.
Cell wall lignification is a complex process that only occurs in higher plants, and its main function is to strengthen plant vascular bodies. Cell wall lignification affects silique dehiscence in siliques. SHP1 and SHP2 promotes valve margin lignification [8]. In this study, we found that the lignin-specific staining of the repla of Arabidopsis p35S::mAtTCP3 and p35S::BnTCP3.A8 plants was weaker as compared to wild-type, while the lignin-specific staining of the repla of jaw-D, and pAtTCP3::AtTCP3SRDX siliques were much stronger than that of the wild type. The premature cracking of siliques before or during ripening causes silique shattering and drastically reduces production. Rapeseed is widely planted in temperate regions. Total yields may reduce 20% owing to silique shattering, and in arid environments this reduction reaches 50% [56,57]. In recent years, some progress has been made in transgenic approaches using Arabidopsis genes to produce indehiscence rape [58]. Chung et al. (2013) reported that the ntt-3D mutant, an activation tagged allele of NTT, showed an enlarged replum and the fruit indehiscence in Arabidopsis [13]. Overexpression of AtFUL gene in B. juncea produces pod shatter-resistant Brassica fruit [19]. The shatter-resistant B. juncea siliques may have smaller non-lignified separation layers than B. napus siliques [59]. A similar phenotype was observed in 35S::FUL Arabidopsis plants, which showed reductions in the lignification of cells adjacent to the DZs, resulting in the formation of indehiscent siliques that did not release seeds normally [20]. The replum of the transgenic line p35S::BnTCP3.A8 displayed a narrower replum and a lower degree of lignification, compared with that of the wild type, and t p35S::BnTCP3.A8 plants exhibited higher silique shatter resistance. Thus, the deficiency in replum may affect silique shattering. Whether are the size and shape of valve margin changed by TCP3 deregulation remains unclear. Both the Arabidopsis p35S::mAtTCP3 and B. napus p35S::BnTCP3.A8 plants exhibited shorter and smaller siliques. This fact implies that the effects of TCP3 on silique development are much broader than expected. An attempt to explore the new function of TCP3 and the other miR319a-targeted gene is underway in our lab. In FUL jaw-D double mutants, silique repla were narrower in width than that of jaw-D silique but still wider than the wild-type, and ful jaw-D double mutants no significant change of repla width was observed compared to ful mutant. TCP3 positively regulated FUL expression and affected replum lignification in Arabidopsis and B. napus. Alternatively, TCP3 may inhibit replum growth in a direct or indirect path unknown. Our results provide novel insights into the mechanisms of silique dehiscence in Arabidopsis and B. napus, which can most likely be applied to other crops and lay a foundation for the development of oil crop varieties having strong shattering-resistance levels.
In Summary, this study revealed a critical role of miR319-regulated TCPs in silique development, reducing replum width and lignification via FUL-regulated pathway, which contributed to the silique shattering resistance in both Arabidopsis and B. napus. Specifically, we found that hyper-activation of BnTCP3.A8 enhanced silique resistance to shattering in rapeseed, providing a potential genetic locus for molecular breeders to improve silique shattering resistance in Brassica crops. | true | true | true |
|
PMC9563714 | Stephen Dela Ahator,Yang Liu,Jianhe Wang,Lian-Hui Zhang | The virulence factor regulator and quorum sensing regulate the type I-F CRISPR-Cas mediated horizontal gene transfer in Pseudomonas aeruginosa | 30-09-2022 | CRISPR-Cas system,quorum sensing,horizontal gene transfer,virulence factor regulator,second messenger,CRISPR adaptation,calcium | Pseudomonas aeruginosa is capable of thriving in diverse environments due to its network of regulatory components for effective response to stress factors. The survival of the bacteria is also dependent on the ability to discriminate between the acquisition of beneficial and non-beneficial genetic materials via horizontal gene transfer (HGT). Thus, bacteria have evolved the CRISPR-Cas adaptive immune system for defense against the deleterious effect of phage infection and HGT. By using the transposon mutagenesis approach, we identified the virulence factor regulator (Vfr) as a key regulator of the type I-F CRISPR-Cas system in P. aeruginosa. We showed that Vfr influences the expression of the CRISPR-Cas system through two signaling pathways in response to changes in calcium levels. Under calcium-rich conditions, Vfr indirectly regulates the CRISPR-Cas system via modulation of the AHL-QS gene expression, which could be vital for defense against phage infection at high cell density. When encountering calcium deficiency, however, Vfr can directly regulate the CRISPR-Cas system via a cAMP-dependent pathway. Furthermore, we provide evidence that mutation of vfr reduces the CRISPR-Cas spacer acquisition and interference of HGT. The results from this study add to the regulatory network of factors controlling the CRISPR-Cas system in response to abiotic factors in the environment. The findings may facilitate the design of effective and reliable phage therapies against P. aeruginosa infections, as targeting Vfr could prevent the development of the CRISPR-Cas mediated phage resistance. | The virulence factor regulator and quorum sensing regulate the type I-F CRISPR-Cas mediated horizontal gene transfer in Pseudomonas aeruginosa
Pseudomonas aeruginosa is capable of thriving in diverse environments due to its network of regulatory components for effective response to stress factors. The survival of the bacteria is also dependent on the ability to discriminate between the acquisition of beneficial and non-beneficial genetic materials via horizontal gene transfer (HGT). Thus, bacteria have evolved the CRISPR-Cas adaptive immune system for defense against the deleterious effect of phage infection and HGT. By using the transposon mutagenesis approach, we identified the virulence factor regulator (Vfr) as a key regulator of the type I-F CRISPR-Cas system in P. aeruginosa. We showed that Vfr influences the expression of the CRISPR-Cas system through two signaling pathways in response to changes in calcium levels. Under calcium-rich conditions, Vfr indirectly regulates the CRISPR-Cas system via modulation of the AHL-QS gene expression, which could be vital for defense against phage infection at high cell density. When encountering calcium deficiency, however, Vfr can directly regulate the CRISPR-Cas system via a cAMP-dependent pathway. Furthermore, we provide evidence that mutation of vfr reduces the CRISPR-Cas spacer acquisition and interference of HGT. The results from this study add to the regulatory network of factors controlling the CRISPR-Cas system in response to abiotic factors in the environment. The findings may facilitate the design of effective and reliable phage therapies against P. aeruginosa infections, as targeting Vfr could prevent the development of the CRISPR-Cas mediated phage resistance.
Pseudomonas aeruginosa is a ubiquitous opportunistic pathogen that thrives in diverse habitats and often infects immunocompromised patients, causing various forms of acute and chronic infections. P. aeruginosa is capable of transforming into a virulent pathogen upon sensing favorable changes in the host (Passador et al., 1993). This bacterium accounts for over 10% of nosocomial infections, making it a highly significant pathogen in hospital settings (Jones, 2010). It causes hard-to-treat infections due to the development of resistance mechanisms against most conventional antibiotics (Drenkard and Ausubel, 2002; Mah et al., 2003; Wolfgang et al., 2003a). As P. aeruginosa becomes increasingly antibiotic-resistant, there is an urgent need for developing novel treatments and disease prevention strategies. Over the last few years, various non-antibiotic disease control treatments have been tested, including quorum quenching (Dong et al., 2007), bacterial vaccines, and phage therapy (Hoggarth et al., 2019). The interest in phage therapy is based on the ubiquity of bacteriophages (phages), host specificity, and their ability to cause detrimental effects on host cells reminiscent of the action of antibiotics. Phages attach to the bacterial host cell via surface receptors and inject their genetic material into the host cell. By hijacking the host’s molecular building blocks and enzymes, they replicate their genetic materials and produce more progeny phages that are released by the lysis of the host cell (Doss et al., 2017). To date, over 130 phages that attack P. aeruginosa have been reported with the fully sequenced genome (Hoggarth et al., 2019), signifying a large repository of natural genetic resources to be exploited in developing practical and effective phage therapy. However, similar to the development of antibiotic resistance, bacteria evolve resistance to phages, in part due to the clustered regularly interspaced short palindromic repeats (CRISPR) and the CRISPR Associated (Cas) proteins, which are widespread in bacteria and archaea (Makarova et al., 2015). This adaptive immune system is composed of a genomic CRISPR array with short sequences known as spacers acquired for previously encountered foreign genetic materials (Jansen et al., 2002). The acquisition of the spacers results in adaptive or heritable immunity. Therefore, upon reinfection or exposure to complementary sequences, the CRISPR array is transcribed and processed into short non-coding CRISPR RNAs (crRNA), which form a complex with the Cas proteins and targets the invading complementary sequences to mediate their cleavage (Fineran and Charpentier, 2012). Given the essential roles of the CRISPR-Cas system in bacterial defense against phage infection and the deleterious effect of horizontal gene transfer (HGT), its expression and maintenance are controlled by factors in the cell in response to biotic and abiotic factors in the bacterial external environment (Westra et al., 2015; Alseth et al., 2019). So far, a few regulators such as the cAMP receptor protein (CRP), H-NS (Agari et al., 2010; Patterson et al., 2015), and LeuO (Westra et al., 2010), and environmental factors such as membrane stress (Perez-Rodriguez et al., 2011), temperature (Høyland-Kroghsbo et al., 2018), and metabolic stress (Yang et al., 2014; Patterson et al., 2015), and nutrient availability (Westra et al., 2015) has been associated with the regulation of the CRISPR-Cas system. Noticeably, the Crp can modulate the CRISPR-Cas system either as a positive or negative regulator depending on the host bacterial species. In Escherichia coli, the cAMP–Crp complex, which controls catabolite repression (Saier Jr and Ramseier, 1996), acts to inhibit the transcription expression of the CRISPR-Cas system (Yang et al., 2014), whereas in Thermus thermophiles and Pectobacterium atrosepticum the cAMP–Crp complex functions as a positive regulator for the regulation of the CRISPR-Cas system (Agari et al., 2010; Patterson et al., 2015). Several Pseudomonas species possess the type I-F CRISPR-Cas system made up of 6 cas genes flanked by two CRISPR arrays (Wiedenheft et al., 2011). The CRISPR-Cas system in Pseudomonas aeruginosa PA14 is inducible, with its expression and function dependent on biotic and abiotic factors such as temperature, microbial interactions, phage exposure, nutrient availability, and population density (Høyland-Kroghsbo et al., 2017, 2018). The Type I-F CRISPR-Cas system in P. aeruginosa is regulated by the AHL QS system, where the Cas genes are significantly upregulated at high cell density but repressed by about 50% when the lasI and rhlI genes encoding biosynthesis of AHL signals were deleted (Høyland-Kroghsbo et al., 2017). The population density-dependent regulation of the CRISPR-Cas system is particularly vital as it allows the bacteria to defend against phage infection and the detrimental effects of horizontal gene transfer at a high cell density where it is vulnerable to phage infection (Abedon, 2012; Høyland-Kroghsbo et al., 2017). In P. aeruginosa, the population density-dependent regulation of the CRISPR-Cas system is mediated particularly by the acyl-homoserine lactone (AHL) quorum sensing (QS) systems, the las, and rhl (Høyland-Kroghsbo et al., 2017). In this bacterium, the QS system is hierarchically organized, with the las on top of the hierarchy controlling the expression of the pqs, which in turn positively regulates the rhl systems (Lee and Zhang, 2015). The las system also directly regulates some rhl-controlled genes as both systems share overlapping regulon (Dekimpe and Déziel, 2009; Luo et al., 2015; Kostylev et al., 2019). In the las system, the transcriptional regulator LasR controls the expression of the autoinducer synthase LasI, which produces 3OC12HSL (N-3-oxo-dodecanoyl-L-homoserine lactone, OdDHL). The rhl system includes the autoinducer synthases RhlI, which produces the C4HSL (N-butanoyl-L-homoserine lactone). Expression of the rhlI is the under control of the transcriptional regulator RhlR. The pqs system produces the quinolone signal PQS (2-heptyl-3-hydroxy-4 (1H)-quinolone) encoded by the pqs gene cluster and regulated by MvfR (Ahator and Zhang, 2019). The regulation of the QS system in P. aeruginosa is influenced by environmental cues and cross-talk from other global regulators in the bacteria. One global regulator involves in cross-talk with the QS system is the Virulence factor regulator (Vfr) which positively regulates the LasR and RhlR. The Vfr-mediated induction of the AHL QS regulators is an essential part of the QS regulatory cascade as the autoinducer synthases LasI and RhlI depend on their cognate regulators for maximal expression (Albus et al., 1997). The P. aeruginosa Vfr is a homolog of the Escherichia coli cAMP regulatory protein (CRP). However, Vfr does not function in catabolite repression control as seen in E. coli or other bacteria (Suh et al., 2002). Under the Vfr regulon, the expression of virulence genes is either indirectly via the QS system or through direct interaction with the Vfr binding sites in the promoter region of the genes. Such virulence factors include pyocyanin, elastase, exotoxin A, protease, type IV pili, and the type III secretion system (T3SS) (West et al., 1994; Albus et al., 1997; Fuchs et al., 2010; Berry et al., 2018). The function of Vfr is reported to depend on environmental factors such as calcium availability which influences the production of cAMP in P. aeruginosa (Beatson et al., 2002; Wolfgang et al., 2003b). The second messenger cAMP is an allosteric activator of Vfr. However, in P. aeruginosa both cAMP-dependent and -independent Vfr regulation of the las QS system has been detected (Fuchs et al., 2010), which shows that Vfr may be involved in other roles in P. aeruginosa via a cAMP-independent pathway. P. aeruginosa thrives in diverse environments with varying levels of nutrients and trace elements, where they are outnumbered by phages. Given the importance of the CRISPR-Cas system for defense against phage infection and HGT, it was speculated that other endogenous and environmental factors were involved in regulating the CRISPR-Cas system. In this study, a genome-wide transposon screen was performed to identify regulators of the CRISPR-Cas system in P. aeruginosa. By exploiting the inducible property of the CRISPR-Cas system and the random insertion property of transposons, inactivation of the virulence factor regulator (Vfr) was identified to reduce cas expression. Further analysis showed that Vfr could modulate the expression of the type I-F CRISPR-Cas system via QS-dependent and -independent pathways. This regulatory cascade is vital for the CRISPR-Cas interference of HGT and the acquisition of spacers.
Supplementary Table S1, S2 list all the strains, plasmids and oligonucleotides used in this study. P. aeruginosa strain PA14 and mutants were grown at 37°C in tryptic soy broth (TSB) supplemented with 5 mM CaCl2 (Sigma) for calcium-rich media and or 5 mM EGTA (Sigma) for calcium-depleted media as indicated. Exogenous QS (AHL) molecules were added at a final concentration of 10 μM OdDHL (Sigma) + 50 μM BHL (Sigma) when necessary. For P. aeruginosa strains, carbenicillin, tetracycline, and gentamicin were added to the media at a final concentration of 300, 50, and 30 μg/ml, respectively, when needed. For E. coli strains, carbenicillin, tetracycline, and gentamicin were added to the media when necessary, at a final concentration of 200, 10, and 5 μg/ml, respectively. The plasmid pUCPT, a derivative of pUCP19 containing the oriT fragment from pK18mobsacB, which enhances transfer from E. coli Si17 to P. aeruginosa was used as the CRISPR non-targeted plasmid control, and for the construction of CRISPR-Cas targeted plasmid following the methods described previously with minor modifications (Patterson et al., 2015). The CRISPR 2 spacer1 fragment from P. aeruginosa PA14 was inserted into the HindIII/EcoRI multiple cloning sites of pUCPT to generate the CRISPR-targeted plasmid, pUCTSp1. The CRISPR-Cas targeted constructs pUCPTSp2n and pUCPTSp4n were created by inserting the CRISPR 2 spacer1 fragment with 2 and 4 nucleotide substitutions, respectively, into the HindIII/EcoRI multiple cloning sites of pUCPT using the ClonExpress MultiS One Step Cloning Kit (Vazyme Biotech). Sequences were verified by PCR and DNA sequencing using the M13F and M13R primers.
To create chromosomal in-frame deletion in P. aeruginosa strains, the upstream and downstream DNA fragments flanking the gene of interest were amplified and ligated with the EcoRI/HindIII-digested pK18mobsacB using the ClonExpress MultiS One Step Cloning Kit (Vazyme Biotech). For chromosomal integrative cas1-lacZ, the up and down DNA fragments flanking the ATG of cas1 and the lacZ gene were amplified with primers stated in Supplementary Table S2 and ligated with the EcoRI/HindIII-digested pK18mobsacB using the ClonExpress MultiS One Step Cloning Kit (Vazyme Biotech). The ligation products were transformed into E. coli DH5α and positive colonies selected by colony PCR and DNA sequencing. The correct constructs with the right fragment orientation were transformed into E. coli S17-1λ for conjugation with P. aeruginosa strains. Transconjugants were selected on minimal medium (MM) [0.2% (w/v) (NH4)2SO4 (Sigma), 0.41 mM MgSO4 (Sigma), 0.2% (w/v) mannitol (Oxoid), 40 mM K2HPO4 (Sigma), 14.7 mM KH2PO4,(Sigma) 32.9 μM FeSO4,(Sigma), 90 μM CaCl2,(Sigma), 16 μM MnCl2 (Sigma)(pH 7.2)] containing gentamicin (30 μg/ml), followed by selection of in-frame deletion mutants on MM supplemented with sucrose (Sigma) (10% w/v). Mutants were further confirmed by PCR and DNA sequencing.
The Mariner transposon, pBT20 was transferred from E. coli S17 to PA14 carrying the construct pMEPcas1-lacZ by conjugation. The resulting mating spot was scrapped and resuspended in 500 ml MM from which aliquot of serial dilution (10−3) was spread on MM agar (1.5% w/v) supplemented with Tetracycline, 50 μg/ml and X-gal [5-bromo-4-chloro-3-indoyl-D-galactopyranoside (Sigma)], 50 μg/ml. Single colonies of the transconjugants were picked onto the selection media composed of MM agar (1.5% w/v) and X-gal (50 μg/ml). Plates were incubated for 48 h and transposon mutants visually inspected for altered expression of the Pcas1-lacZ construct evident by the blue coloration of the colonies in comparison to wild-type. Colonies or transposon mutants with altered coloration were selected by colony tail PCR and DNA sequencing using primers listed in Supplementary Table S2 to map the position of transposon insertion through blast search against the P. aeruginosa UCBPP-PA14 genome.
Cells were harvested after growth in specified media to OD600 = 1.5. RNA was extracted using the RNA extraction kit according to the manufacturer’s protocol (Qiagen). The quantity and integrity of the RNA was determined by Nanodrop and gel electrophoresis. One step Qrt PCR reaction were performed using the Tiangen One-step SYBR Green kit with the Applied Biosystems QuantStudio 6 Flex RT-PCR System.
DNA promoter fragments for the EMSA probes were constructed by PCR using the indicated primers in Supplementary Table S2 and end-labeled with biotin using the biotin 3I end DNA labeling kit (Thermo Fisher Scientific) as described in the kit protocol. EMSAs were performed using the LightShift chemiluminescent EMSA kit (Thermo Fisher Scientific) according to the kit protocol. Briefly, 1 nm of DNA fragments was incubated with cAMP-Vfr or Vfr and the binding buffer containing 1 μg/μL Poly (dI.dC), 50% Glycerol, 1% NP-40 1 M KCl 100 mM MgCl2 and 200 mM EDTA supplied with the kit for 25 min at 25°C. The cAMP was added at a concentration of 20 μM (Ferrell et al., 2008; Fuchs et al., 2010). The binding products were resolved on a 6% native polyacrylamide gel in 0.5X TBE and transferred to the nylon membrane at 380 mA (~100 V) for 30 min. The membrane was crosslinked at 120 mJ/cm2 for 45–60 s followed by detection of the biotin-labeled DNA by chemiluminescence using the Tanon 5,200 imaging system.
The conjugation efficiency was performs using the method described by Patterson A.G. and colleagues (Patterson et al., 2015) with minor modification. The E. coli S17λ was used to transfer the CRISPR-targeted plasmids; pUCPTSp1, pUCPTSp2n, and pUCPTSp4n, and non-targeted plasmid, pUCPT into the P. aeruginosa strains through conjugation. Overnight cultures of E. coli and P. aeruginosa were mixed at a ratio of 1:1, washed twice and pellets resuspended in LB from which 50 μl were spotted on LB agar gently to prevent splatter. The mating spot was allowed to dry and incubated for 16 h at 37°C. The mating spot was scrapped completely and resuspended in 250 μl TSB from which serial dilutions of 10−5 were platted on TSB + EGTA and TSB + CaCl2 agar supplemented with carbenicillin (200 μg/ml). The conjugation efficiency was calculated as the ratio of transformants with the targeted plasmid compared with the transformants with the non-targeted plasmid.
The plasmid loss and spacer acquisition assay was performed using a method described by Patterson A.G. and colleagues (Patterson et al., 2015) with modifications. The non-targeted plasmid pUCPT and the CRISPR2 spacer1-targeted plasmid, pUCPTSp1 were used to test CRISPR-Cas mediated interference assay. The plasmids were transferred to P. aeruginosa by mating with E. coli S17. Selected colonies were cultured overnight in 5 ml TSB with or without 5 mM EGTA or 5 mM CaCl2 and passaged for 5 days by sub-culturing 20 μl into 5 ml of fresh media. Each passage was serially diluted and 10−6 dilutions plated on LB with or without carbenicillin to count colonies that retain the plasmid. The CRISPR-Cas targeted constructs pUCPTSp2n and pUCPTSp4n, which contain a protospacer similar to CRISPR2 spacer1 with adaptation-priming mutations of 2 and 4 nucleotide substitutions, respectively, were used for the adaptation assay. P. aeruginosa strains initially transformed and passaged with pUCPTSp2n were further transformed with pUCPTSp4n to assay for primed adaptation. The genomic DNA from the samples of the final passage was extracted for the identification of expanded CRISPR2 arrays using primers stated in Supplementary Table S2. PCR products were resolved by 3% agarose gel electrophoresis to detect expansion of the CRISPR array.
Pseudomonas aeruginosa cells were grown under specified culture conditions to appropriate time point and optical density (OD600). Briefly, 200 μl of cells was removed, pelleted and supernatants removed completely. Subsequently, 200 μl of Z buffer (8.52 g Na2HPO4, 5.5 g NaH2PO4.H2O, 0.75 g KCl and 0.246 g MgSO4.7H2O, pH 7.0), 20 μl 0.1% SDS and 20 μl chloroform was added and vortexed for 3 min. A volume of 200 μl ONPG (ortho-Nitrophenyl-β-galactoside) (Sigma), 4 mg/ml (dissolved in Z buffer) was added to the reaction mix and incubated at 37°C for a specified period of time. Finally, 600 μl of 1 M Na2CO3 was added to stop the reaction and the absorbance measured at A420nm. The blank sample was composed of 200 μl Z buffer; 200 μl ONPG and 600 μl Na2CO3. β-galactosidase activity was calculated as (1,000 × A420)/ (OD600 × Volume (ml) × time of reaction (min)) and expressed as miller units (MU) as previously described (Miller, 1972).
The vector pET-28b (+) was used for Vfr protein expression. The Vfr coding sequence was amplified using the primer pairs listed in Supplementary Table S2 and ligated with the pET-28b (+) resulting in C-terminal His-tagged fusion. The resulting construct pET-VfrHis was transformed into E. coli BL21 (DE3) and grown in LB broth supplemented with kanamycin at 37°C to OD600 = 0.5 followed by induction with isopropyl β-D-thiogalactoside (IPTG) (Invitrogen) (0.5 mM) at 16°C overnight. Bacterial pellets obtained were resuspended in ice-cold lysis buffer [50 mM NaH2PO4 (Sigma), 300 mM NaCl (Sigma), 1 mM DTT (Sigma), 10 mM imidazole (Sigma), pH 7.5] containing protease inhibitors (Complete mini, EDTA free, Roche) and lysed by sonification. Cell-free supernatants incubated with ProteinIso Ni-NTA Resin (TransGene Biotech, China) at 4°C for 2 h. Subsequently, the resins were washed 4 times with wash buffer (50 mM NaH2PO4, 300 mM NaCl, 1 mM DTT, 50 mM imidazole, pH 7.5) and the proteins eluted with the elution buffer (50 mM NaH2PO4, 300 mM NaCl, 1 mM DTT, 300 mM imidazole, pH 7.5). The protein purity was determined by SDS-PAGE analysis (Supplementary Figure S1C) and dialyzed against the PBS buffer (PBS, 5% glycerol, pH 7.4) at 4°C.
Quantification of intracellular cAMP was performed by adapting the method used by Fulcher et al. (2010). Bacteria were grown in TSB supplemented with EGTA or CaCl2 to OD600 = 1.0. Briefly, 1 ml of bacterial culture was centrifuged at 12,000 rpm for 2 min at 4°C and the pellets washed twice with 1 ml of ice-cold 0.9 M NaCl. The cells were lysed by resuspension in 100 μl of 0.1 N HCl, incubated on ice for 15 min with intermittent agitation every 5 min. The lysates were centrifuged at 12,000 rpm for 5 min at 4°C and the cell-free supernatant obtained for cAMP quantification using the cAMP enzyme immunoassay kit (Sigma-Aldrich) as per the manufacturer’s recommendation for sample acetylation. For protein determination, duplicate samples were suspended in 100 μl of ice-cold phosphate buffered saline (PBS), lysed by 3 freeze–thaw cycles and centrifuged at 1,200 rpm for 5 min. The protein concentration was determined by the Pierce BCA protein assay kit (Thermo Fisher Scientific). The intracellular cAMP values were presented as pmole per μg of total protein.
In this study, the Pseudomonas aeruginosa UCBPP_PA14, which contains the type I-F CRISPR-Cas system with six cas genes flanked by two CRISPR arrays (Figure 1A), was the wild-type (WT) strain used. Screening for the potential regulators of the CRISPR-Cas system was performed by genome-wide random transposon mutagenesis with the mariner transposon in the WT expressing the Pcas1-lacZ reporter construct. The lacZ gene was placed under the control of the cas1 promoter to identify transposon insertion sites that result in the downregulation of cas1 promoter activity. Colonies with downregulated Pcas1-lacZ expression resulting in lighter blue color to the WT parental mating strains were selected for transposon insertion mapping (Supplementary Figure S1A). Of particular interest among the transposon mutants identified from the screening was the insertion into the genes encoding the Virulence factor regulator (Vfr) (Supplementary Figure S1A). Additional bioassay of the cas1 promoter in the vfr transposon mutant showed a reduction in the activity of the promoter (Supplementary Figure S1B). Since the secondary messenger cAMP required for Vfr function is dependent on calcium availability (Dasgupta et al., 2006), the effect of Vfr on the Cas gene expression was examined in calcium-rich and calcium-depleted media. Using the in-frame deletion mutant showed that Vfr positively regulates the Cas gene expression under both calcium depleted and calcium-rich conditions (WT/∆vfr in TSB + EGTA: p < 0.0001, t = 65, df = 9; WT/∆vfr in TSB + CaCl2: p < 0.0001, t = 37, df = 9) (Figure 1B; Supplementary Figure S2C). This shows that Vfr can regulate the CRISPR-Cas system in the presence or absence of cAMP, which is not surprising as the cAMP-independent functionality of Vfr is possible in P. aeruginosa (Fuchs et al., 2010).
The Virulence factor regulator, Vfr in P. aeruginosa recognizes the nucleotide sequence “tgnga-N6tcaca” in its target promoter region as the binding site. Typically, Vfr binding sites show a high degree of variability at the left palindromic sequence portion (Fuchs et al., 2010; Figure 2A). Nucleotide sequence alignment revealed the sequence “gctca N6 tcaca” in the promoter region of cas1, which shares similarity with the conserved binding sequence of Vfr (Figure 2A) in known Vfr-regulated genes in P. aeruginosa (Figure 2A; Supplementary Figure S3; Yahr and Wolfgang, 2006; Ferrell et al., 2008; Fuchs et al., 2010). This implies that Vfr could regulate cas1 expression through protein-promoter interaction. Since Vfr regulates its target genes via interacting with specific sequences (Vfr box) in the presence of the second messenger cAMP (West et al., 1994), an EMSA analysis was used to investigate the interaction between Vfr and the nucleotide sequence of the cas1 promoter (Pcas1). The EMSA analysis showed an interaction between the Vfr and the Pcas1 DNA fragment occurred when cAMP was added to the reaction mix (Figure 2B). By mutating the putative Vfr binding sequence in Pcas1 to “atatg N6 atccc,” Vfr did not interact with the Pcas1 fragment with and without cAMP (Figure 2B). Furthermore, the Vfr interaction was examined using the DNA probe from the promoter region of the cyaB (PcyaB), a major intracellular cAMP synthase in P. aeruginosa which lacks the Vfr binding site as a negative control (Smith et al., 2004; Topal et al., 2012) and the DNA probe from the promoter region of ptxR (PptxR), a Vfr-regulated transcriptional regulator as a positive control (Ferrell et al., 2008). Vfr showed interaction with the PptxR but not PcyaB (Figure 2B). To further investigate the cas1 gene expression, the transcriptional lacZ fusion constructs composed of the promoter region of the cas1 gene with intact (Pcas1) and mutated Vfr binding site (Pcas1:VBS) were constructed and transformed into the WT and vfr mutant (∆vfr). In the WT, the expression of the cas1 promoter with the mutant Vfr binding site was reduced compared with the reporter containing the intact Vfr binding site (Pcas1/ Pcas1VBS expression in WT (+CaCl2): p < 0.0001, t = 31.47 df = 7.37; (–CaCl2): p < 0.0001, t = 43.64 df = 6.63)) (Figure 2C). In the ∆vfr, a similar effect of the Pcas1 expression from the intact Vfr binding site was observed as in the WT possessing the mutated Vfr binding site (WT (Pcas1VBS)/∆vfr (Pcas1) in (+CaCl2): p = 0.076, t = 2.95 df = 8.33; (–CaCl2): p = 0.069, t = 2.112 df = 7.79) (Figure 2C). Unexpectedly, in the vfr mutant, expression of cas1 from Pcas1:VBS construct significantly reduced compared to the expression from the Pcas1 construct under both calcium-rich and depleted conditions (∆vfr (Pcas1/Pcas1VBS) + CaCl2: p < 0.0001, t = 13.86 df = 8.221; –CaCl2: p < 0.0001, t = 9.70 df = 8.46)) (Figure 2C). These results demonstrate that Vfr and its binding site are required for activation of the cas operon and that Vfr can control activation of the cas genes via an alternative pathway. Also, the deletion of cyaB, the major cAMP synthase in P. aeruginosa under calcium depleted conditions (Wolfgang et al., 2003b; Supplementary Figure S4), resulted in a reduction in the cas gene expression in calcium depleted media but not in the calcium-rich media (∆cyaB/WT in (+EGTA): p < 0.0001, t = 36.1 df = 7; in (+CaCl2): p = 0.055, t = 2.39 df = 5.8) (Figure 3). This implies that the second messenger cAMP influences cas gene expression under calcium-depleted conditions and that the Vfr regulation of the cas gene expression under calcium-rich conditions may occur via an alternative pathway independent of cAMP.
In P. aeruginosa, Vfr regulates the las QS system both in the presence and absence of cAMP (Figure 4A; Supplementary Figure S5; Albus et al., 1997; Fuchs et al., 2010). Prompted by the regulation of the CRISPR-Cas system by the hierarchically organized AHL QS system (Lee and Zhang, 2015; Høyland-Kroghsbo et al., 2017) and the presence of the las/rhl box in the promoter region of the cas1 (Supplementary Figure S3), we hypothesized that the Vfr could indirectly regulate the CRISPR-Cas system via the AHL system in the absence of cAMP. To investigate this Vfr-QS-CRISPR regulatory cascade, the cas gene expression was examined in the double deletion lasI and rhlI mutant designated as “∆ahl” in the ∆vfr strain or WT background grown under calcium depleted (+EGTA) and calcium-rich (+CaCl2) conditions. In both calcium depleted and calcium replete media, cas1 expression was reduced in the ∆vfr∆ahl, ∆ahl, and ∆vfr strains compared to the WT (Figures 4B,C). Addition of exogenous AHL (10 μM OdDHL +50 μM BHL) to the ∆ahl strain rescued the expression of the cas1 gene in both conditions (∆ahl/∆ahl + AHL in (+EGTA): p < 0.0001, t = 24.45 df = 6.67; ∆ahl + AHL /WT in (+EGTA): p = 0.14, t = 1.62 df = 9.45), however, cas1 expression in the ∆vfr∆ahl strain was not fully rescued to the WT level when the media was supplemented with exogenous AHL (∆vfr∆ahl + AHL/WT in (+CaCl2): p = 0.004, t = 4.83 df = 10.17; (+EGTA): p < 0.0001, t = 17.75 df = 6.26) (Figure 4B). Due to the global regulon of the Vfr and the QS system, there is a possibility that other factors that are vital for complete regulation of the Vfr-QS-CRISPR regulatory cascade may be affected by the deletion of both Vfr and the AHL QS synthases. Taken together, these results show that the Vfr can regulate the CRISPR-Cas system in P. aeruginosa either directly or via the AHL QS system.
The CRISPR-Cas system facilitates the targeted degradation of invading genetic materials that share similarities with spacers located in the CRISPR array. The spacers in the CRISPR arrays are derived from the invaders and are essential for immunologic memory and defense against previously encountered foreign elements (Bhaya et al., 2011; Dy et al., 2014). Following the transcriptional control of the Vfr and the AHL QS system on the cas gene expression, the impact of the Vfr and QS on CRISPR-Cas mediated interference, spacer acquisition, and conjugation efficiency were further investigated. Using the E. coli S17 as a donor, the WT, ∆vfr, ∆ahl, and ∆cas3 strains were transformed with the CRISPR-targeted plasmid containing a spacer with the GG PAM recognized by CRISPR2 spacer 1 (pUCPTSp1) and the non-targeted plasmid (pUCPT) devoid of spacers recognized by the CRISPR-Cas system. The conjugation efficiency was calculated as the ratio of the colonies retaining the CRISPR-targeted plasmid to that of the non-targeted plasmid. In the ∆cas3 strain, which is defective in CRISPR-mediated interference and spacer acquisition (Høyland-Kroghsbo et al., 2017), the conjugation efficiency of the targeted plasmid was comparable to the non-targeted plasmid under both conditions tested (Figure 5A). The WT showed the least conjugation efficiency in comparison to the ∆vfr ((+EGTA): p < 0.0001, t = 10.75 df = 5; (+CaCl2): p < 0.0001, t = 15.52 df = 9) and ∆ahl (+CaCl2): p < 0.0001, t = 10.62 df = 9; (+EGTA): p < 0.0001, t = 21.31 df = 7) strains under both calcium-rich and calcium depleted conditions (Figure 5A), showing that in the absence of the Vfr and the AHL QS system, the CRISPR-mediated interference of plasmid transfer is reduced. Next, the impact of Vfr and AHL QS on CRISPR-mediated plasmid loss was tested by transforming the CRISPR-targeted and non-targeted plasmids into the P. aeruginosa strains followed by a 5-day successive passage in calcium-depleted and calcium-rich media. Aliquots of serial dilutions from each passage were plated on LB agar supplemented with carbenicillin and X-gal. The plasmid loss was assessed by counting the colonies that grew on the plates following overnight incubation at 37°C. Under both calcium-rich and calcium-depleted conditions, over 5 days of passage, the ∆vfr, ∆ahl, and ∆vfr∆ahl showed significantly increased retention of the CRISPR-targeted plasmid compared to the WT (Figures 5B, C). However, the retention of the non-targeted plasmid was similar in the WT, ∆vfr, ∆ahl, and ΔvfrΔahl strains (Supplementary Figure S6). The CRISPR-Cas system builds immunological memory against previously encountered genetic elements by incorporating the target sequence into the CRISPR array, which results in the expansion of the array. To investigate the impact of Vfr and the AHL QS system on the expansion of the arrays, a PCR reaction targeting the CRISPR2 array, which has a higher frequency of adaptation (Westra et al., 2015; Høyland-Kroghsbo et al., 2017) was performed. The plasmids pUCPTsp2n and pUCPTSp4n containing a protospacer similar to CRISPR2 spacer1 with adaptation-priming mutations were transformed into the P. aeruginosa strains and assayed for the expansion of the CRISPR2 locus in the colonies passaged in calcium-rich and calcium depleted media. In the absence of cas3, no expansion of the CRISPR array occurred under both conditions tested (Figure 6). Also, the WT containing the naïve plasmid with no protospacer targeted by the CRISPR-Cas system did not induce the expansion of the CRISPR array (Figure 6). In the WT there was an expansion of the arrays with the incorporation of spacers when passaged in both calcium-rich and depleted conditions. In comparison with the WT, expansion of the CRISPR array was less in the ∆vfr strain when passaged in the calcium-depleted medium compared to the calcium-rich medium (Supplementary Figure S7A). As expected, expansion of the CRISPR array was reduced in the ∆cyaB when passaged in calcium-depleted conditions compared to calcium-rich conditions (p < 0.022, t = 3.8 df = 3.7). Spacer acquisition in the ∆vfr∆ahl and ∆ahl strains was significantly reduced, however, the addition of exogenous AHLs to the ∆vfr∆ahl strain rescued spacer acquisition to the WT level (Supplementary Figure S7B). In general, these results demonstrate that the Vfr and the AHL QS system are required for CRISPR-mediated spacer acquisition and interference of HGT in P. aeruginosa.
The CRISPR-Cas immune system is a defense mechanism for most bacteria and archaea against phage infection and acquiring deleterious genetic elements via HGT. Despite the remarkable advancements in genetic engineering achieved by repurposing the CRISPR-Cas system, the molecular mechanisms by which bacteria regulate this defense mechanism are largely unknown. Identifying the regulatory factors that control the CRISPR-Cas system in bacteria will provide a platform for understanding its role in bacterial lifestyle and exploitation for controlling bacterial infections. Here, using Pseudomonas aeruginosa and the type I-F CRISPR-Cas system, the global transcriptional regulator Vfr is shown as a regulator of the CRISPR-Cas immune system. The deletion of vfr reduces the cas gene expression and attenuates CRISPR-mediated defense mechanisms such as the spacer acquisition and interfering with previously encountered genetic elements. This work provides evidence that Vfr can directly or indirectly control the transcription of the genes required for the maintenance CRISPR-Cas system function depending on the availability of calcium in the local environment. Furthermore, the Vfr regulatory cascade is shown to act with or without the second messenger cAMP in controlling the CRISPR-Cas immune system via an alternative pathway that involves the AHL QS systems, lasIR, and rhlIR. Given the previously known functions of Vfr involve interaction with the allosteric activator, cAMP, whose biosynthesis is induced by calcium limitation (Wolfgang et al., 2003b), this work highlights an alternative pathway for Vfr and the type I-F CRISPR-Cas system regulation in P. aeruginosa dependent on calcium availability. The Vfr-cAMP complex promotes direct regulation of the CRISPR-Cas system under calcium-depleted conditions whereas, under calcium-rich conditions, the AHL QS system mediates the Vfr-CRISPR-Cas regulatory cascade (Figure 4C). The ability of Vfr to regulate the QS system in calcium-rich conditions where cAMP levels are reduced is contrary to the previous reports that Vfr is functionally dependent on its allosteric regulator cAMP (Wolfgang et al., 2003b). Prompted by the AHL QS regulation of the CRISPR-Cas system in P. aeruginosa (Høyland-Kroghsbo et al., 2017), we investigated an alternative pathway under calcium-rich conditions, where Vfr regulates the expression of the CRISPR-Cas system via the AHL QS system (Figure 4, Supplementary Figure S5). In support of this regulatory cascade, Vfr controls the expression of the AHL QS genes under both calcium-rich and calcium–deplete conditions (Supplementary Figure S5). Also, double deletion of the vfr and AHL synthases (∆vfr∆ahl) significantly reduced cas gene expression (Figure 4), HGT interference (Figure 5), and spacer acquisition (Figure 6) compared to the ∆vfr and ∆ahl single mutants under calcium-rich conditions. The identification of Vfr as a regulator of the CRISPR-Cas system further expands the functional spectrum of the global regulator in P. aeruginosa. Vfr is a Crp-family transcriptional regulator and shares a similar binding site as the cAMP receptor protein (Crp) of E. coli, however, it is not functionally complementary with Crp and is not involved in carbon catabolite regulation as observed in E. coli (West et al., 1994). In P. aeruginosa, Vfr regulates quorum sensing, pyocyanin, elastase, and exotoxin A production (West et al., 1994; Albus et al., 1997). Transcriptome analysis showed that deletion of vfr results in decreased expression of over 200 genes, including those encoding the type III secretion system, type IV pilus biogenesis, and type II secretion (Wolfgang et al., 2003b). This work, therefore, extends the function of Vfr from a regulator of virulence factors (offense) to a regulator of an adaptive immune system (defense), which ensures protection of the bacterial cells while competing for survival under conditions where it is prone to phage infection and harmful effects of HGT. Acquiring genetic materials via HGT has added benefits for bacteria as it drives evolutionary adaptive traits such as antibiotic resistance, virulence, and adaptation to environmental stress conditions (Vogan and Higgs, 2011). This implies that constitutive expression of the CRISPR-Cas system will not be overall beneficial as the bacteria may lose out on the benefits of HGT. Controlled expression of the CRISPR-Cas system in the bacteria will thus allow the acquisition and incorporation of beneficial genetic elements. Also, a hyperactive CRISPR-Cas system runs the risk of autoimmunity, which can be particularly deleterious to the bacterial population (Stern et al., 2010; Høyland-Kroghsbo et al., 2017). Hence, the induction of the CRISPR-Cas system by multiple pathways observed in the Vfr-QS-CRISPR regulatory cascade in response to specific environmental factors will be more beneficial to the bacterial population. The Vfr is hierarchically above the AHL QS system and shares regulon with the las and rhl systems which make up over 20% of P. aeruginosa genes (Coggan and Wolfgang, 2012; Ahator and Zhang, 2019). It is unknown if other factors under the Vfr and QS regulon may account for the inability of exogenous AHL to rescue the Cas gene expression in the ∆vfr∆ahl strains (Figures 4B,C) but restore its ability to incorporate spacers into its CRISPR array (Figure 6B). Identifying such factors will help understand the gamut of the Vfr-QS-CRISPR regulatory network. In the cas1 promoter region, the Vfr box partially overlaps with one of the las/rhl box identified using the Prodoric database (Supplementary Figure S3; Münch et al., 2003), which may account for the reduced expression of cas1 from the construct with altered Vfr binding site (Figure 2). This partial overlap in binding sites reveals the possibility of the Vfr and the AHL regulators competing for the binding site but rules out simultaneous binding at the overlapping site. Simultaneous binding at the other distant las/rhl boxes may be possible but not yet experimentally verified. How the bacteria coordinate the Vfr-QS-CRISPR regulatory cascade may depend on the combination of bacterial metabolic requirements and response to environmental factors. Despite the overlapping regulon, the Vfr regulates pili formation (Coggan et al., 2022), which serves as phage binding sites and entry portals for nucleic acid (Craig et al., 2004; Harvey et al., 2018). Identification of Vfr as a central regulator of the CRISPR-Cas immune system might have significant implications for understanding bacterial physiology. The two mechanisms with which Vfr controls the transcriptional expression and function of the CRISPR-Cas immune system would enable the pathogen to activate the immune system against phage infections and HGT regardless of the changes in the bacterial quorum level or the local calcium concentrations, which could vary drastically under either in vivo or in vitro environmental conditions. For example, the decontrolled calcium homeostasis in the Cystic Fibrosis lung results in elevated calcium in body fluids (Broder et al., 2016). Similarly, wounding accompanies a surge in calcium concentrations from early in the post-wound period through the inflammatory and proliferative phases and the remodeling phase (Lansdown, 2002). Furthermore, the Vfr regulation of the CRISPR-Cas immune system might facilitate the design and development of effective and reliable phage therapy. Firstly, the Vfr is hierarchically above the AHL QS system in regulating the CRISPR-Cas system. Secondly, the Vfr-cAMP complex regulates factors such as type IV pili biogenesis which is not under the control of QS (Beatson et al., 2002). Aside from the roles in pathogenesis and biofilm formation, the type IV pili of P. aeruginosa is vital for transformation, conjugation, phage adsorption, and infections (50–52; Craig et al., 2004). Also, pili-mediated twitching motility increases the chances of phage-bacteria interactions due to the cell–cell aggregated movement, which creates a spatial vulnerability for phage interaction with the cells (Abedon, 2012; Alexandre, 2015). Therefore, targeting Vfr could have dual effects in safeguarding phage therapy by turning down the expression of the CRISPR-Cas phage immune system and avoiding the formation of pili, which serve as receptors and entry ports for phage particles (Craig et al., 2004). The diagrammatic representation of the Vfr-QS-CRISPR regulatory cascade under calcium-depleted and calcium-rich conditions is in Supplementary Figure S7.
The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.
SA and L-HZ designed the experiments. SA, YL, and JW conducted the experiments. SA, YL, JW, and L-HZ performed the data analysis. SA and L-HZ wrote the manuscript. All authors contributed to the article and approved the submitted version.
This work was supported by the Guangdong Forestry Science and Technology Innovation Project (2020KJCX009) and Guangdong Technological Innovation Strategy of Special Funds (grant no. 2018B020205003). YL is supported by the China Scholarship Council (CSC) (Grant No. 202008440425).
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 |
|
PMC9563724 | Koushik Mondal,Richard C. Grambergs,Rajashekhar Gangaraju,Nawajes Mandal | A Comprehensive Profiling of Cellular Sphingolipids in Mammalian Endothelial and Microglial Cells Cultured in Normal and High-Glucose Conditions | 30-09-2022 | sphingolipids,Hexosyl-ceramide,Lactosyl-ceramide,Sphingosine-1-phosphate,high-glucose,endothelial cells,microglial cells,inflammatory markers,proliferative markers | Sphingolipids (SPLs) play a diverse role in maintaining cellular homeostasis. Dysregulated SPL metabolism is associated with pathological changes in stressed and diseased cells. This study investigates differences in SPL metabolism between cultured human primary retinal endothelial (HREC) and murine microglial cells (BV2) in normal conditions (normal glucose, NG, 5 mM) and under high-glucose (HG, 25 mM)-induced stress by sphingolipidomics, immunohistochemistry, biochemical, and molecular assays. Measurable differences were observed in SPL profiles between HREC and BV2 cells. High-glucose treatment caused a >2.5-fold increase in the levels of Lactosyl-ceramide (LacCer) in HREC, but in BV2 cells, it induced Hexosyl-Ceramides (HexCer) by threefold and a significant increase in Sphingosine-1-phosphate (S1P) compared to NG. Altered SPL profiles coincided with changes in transcript levels of inflammatory and vascular permeability mediators in HREC and inflammatory mediators in BV2 cells. Differences in SPL profiles and differential responses to HG stress between endothelial and microglial cells suggest that SPL metabolism and signaling differ in mammalian cell types and, therefore, their pathological association with those cell types. | A Comprehensive Profiling of Cellular Sphingolipids in Mammalian Endothelial and Microglial Cells Cultured in Normal and High-Glucose Conditions
Sphingolipids (SPLs) play a diverse role in maintaining cellular homeostasis. Dysregulated SPL metabolism is associated with pathological changes in stressed and diseased cells. This study investigates differences in SPL metabolism between cultured human primary retinal endothelial (HREC) and murine microglial cells (BV2) in normal conditions (normal glucose, NG, 5 mM) and under high-glucose (HG, 25 mM)-induced stress by sphingolipidomics, immunohistochemistry, biochemical, and molecular assays. Measurable differences were observed in SPL profiles between HREC and BV2 cells. High-glucose treatment caused a >2.5-fold increase in the levels of Lactosyl-ceramide (LacCer) in HREC, but in BV2 cells, it induced Hexosyl-Ceramides (HexCer) by threefold and a significant increase in Sphingosine-1-phosphate (S1P) compared to NG. Altered SPL profiles coincided with changes in transcript levels of inflammatory and vascular permeability mediators in HREC and inflammatory mediators in BV2 cells. Differences in SPL profiles and differential responses to HG stress between endothelial and microglial cells suggest that SPL metabolism and signaling differ in mammalian cell types and, therefore, their pathological association with those cell types.
Sphingolipids (SPLs) are a diverse and ubiquitous family of lipids that function as structural components of the cellular membranes and have signaling roles in various cellular processes, including cell growth, cell adhesion, inflammation, and apoptosis [1,2,3,4]. Biochemically, all SPLs contain an amino-alcohol backbone and a covalently bound fatty acid forming the base structure of an SPL, the Ceramide (Cer). Cer can have numerous possible head groups and a variety of fatty acid side chains, allowing for a large array of unique species with diverse bioactivity. SPL metabolism is complex, and SPL subspecies may have unique roles in different cell types [5]. Besides being integral structural components of the biomembrane, SPL species play signaling roles. The key SPL metabolite, Cer, signals for inflammation and apoptosis. However, Cer-derived SPL, Sphingosine-1-phosphate (S1P), regulates the processes such as cell survival, proliferation, and formation of cellular junctions [2]. Thus, by altering and maintaining the balance between various bioactive species, cellular SPLs may play a major role in maintaining cellular homeostasis under normal conditions and under stresses [6,7,8]. Further, their involvement in human diseases showcases the importance of SPL metabolism and signaling in human physiology as SPL metabolic derangements are associated with several diseases, including Farber, Tay–Sachs/Sandhoff, Gaucher, Krabbe, and Niemann Pick disease [9,10]. Besides metabolic disorders, studies in the past two decades revealed SPL’s and bioactive SPL’s association with various human diseases, including neurodegenerative, inflammatory, neovascular, neoplastic, and diabetes mellitus (DM) [9,11]. Abnormal lipid metabolism (dyslipidemia) is a recognized pathophysiological factor for microvascular and macrovascular complications in DM. Alterations of SPL metabolism are also gaining recognition in the pathophysiology of DM and many other chronic inflammatory disorders [11,12,13]. SPL metabolic changes are suspected to be one of the root causes of developing diabetic complications such as diabetic retinopathy (DR) and diabetic neuropathy (DN) [14,15]. Endothelial cells maintain vascular permeability by forming a continuous monolayer covering the inner lumen of blood vessels and help regulate solute trafficking, immune responses, and angiogenesis [16]. Breakdown or leakage of the endothelial barrier is one of the pathological hallmarks of diabetic microvascular complications, as occurs in cases of DR [17]. Studies report that dysfunctional retinal vasculature and compromised vascular repair process in DM are associated with activation of a major Cer-generating enzyme, acid sphingomyelinase (aSMase), in retinal endothelial cells (HREC) [18,19]. Microglial cells, on the other hand, are resident macrophages in the central nervous system (CNS) and respond to infection and tissue injury as the primary effectors of CNS inflammatory responses [20]. These immune cells are believed to be a primary driver of neuroinflammation-related neurodegeneration [21]. Increased levels of Cer via activation of SMase in microglia activate proinflammatory factors [22]. Exposure to high-glucose conditions has been shown to stimulate proinflammatory cytokine signaling in microglial cells [23,24]. Microglia can contribute to perivascular inflammation in diseases such as DR, which exacerbates vascular pathology and neovascularization [25]. While SPL signaling is implicated in critical physiological processes governed by endothelial and microglial cells, there is still little information regarding the composition of SPLs within these distinct cell types and how they modulate their SPL profile when exposed to higher environmental glucose, such as in the cases of DM. The present study investigated the SPL profiles of human retinal endothelial cells (HREC) and murine-derived microglia (BV2 cells) in normal physiologic conditions and in high-glucose conditions and the expression and function of their metabolic enzymes and inflammatory and vascular permeability markers. Significant differences were observed in SPL profiles between these cell types at baseline and in response to high-glucose stress. Along with the changes in their profiles, we also observed changes in the expression and function of their metabolic enzymes, vascular permeability factors, and proinflammatory markers in both cell types in response to high-glucose stress. These findings could reflect innate differences in SPL metabolism and SPL mediation of cell processes in response to cellular stress between endothelial and microglial cells.
Human retinal endothelial cells (HREC) (ACBRI 181) were obtained from Cell Systems, Kirkland, WA, USA. They were cultured in 75 cm2 culture flasks with Cell Systems Complete Medium (4NO-500) (Cell Systems, Kirkland, WA, USA) and Culture Boost (4CB-500) (Cell Systems, Kirkland, WA, USA), following the manufacturer guidelines. The cultures were maintained at 37 °C in a humidified atmosphere containing 95% air and 5% CO2. Before plating the cells, the plates were treated with Attachment Factor (4ZO-201) (Cell Systems, Kirkland, WA, USA). The cells were then plated in 6-well plates with 1 × 106 cells/well (passage 6–7) following the same procedure. After 24 h of incubation, media were replaced with fresh media containing 5 mM glucose for control as normal glucose (NG) with 20 mM of glucose (NG + 20 mM D-glucose) as high glucose (HG) and cultured for 48 h to represent diabetic endothelial cells, as reported earlier [26]. In addition to normal glucose (5 mM), cells were also cultured with 20 mM of L-glucose (LG; NG + 20 mM L-glucose) or 20 mM of mannitol (Man; NG + 20 mM mannitol) served as additional controls. Subsequently, cells were harvested for sphingolipidomic and biochemical analyses. The mouse microglial cell line, BV2, was a kind gift from Professor Grace Sun, Ph.D., University of Missouri, Columbia, MO, USA. BV2 cells were routinely grown in 100 mm cell culture dishes in Dulbecco’s Modified Eagle Medium (DMEM) with 10% fetal bovine serum and antibiotics, as described previously [27]. BV2 cells were exposed to normal or high-glucose conditions described above and proceeded for sphingolipidomic, molecular, and biochemical analyses.
Following previously published procedures, sphingolipids in HREC and BV2 were quantified and analyzed in the Lipidomic Core facility at Virginia Commonwealth University, Richmond, VA, USA [7,28]. Internal standards were procured from Avanti Polar Lipids (Alabaster, AL, USA) and added to samples at a ratio of ethanol/methanol/water (7/2/1) as a cocktail of 500 pmol each. Standards for sphingoid bases and sphingoid base 1-phosphates were 17-carbon chain length analogs: C17 Sphingosine, C17 Sphinganine, C17 Sphingosine-1-Phosphate, and C17 Sphinganine-1-Phosphate. Standards for N-acyl sphingolipids were C12-fatty acid analogs: C12 Sphingomyelin, C12 Ceramide, C12 Glycosylceramide, C12 Lactosylceramide, C12 Ceramide-1-Phosphate. For LC-MS/MS analyses, a Shimadzu LC-20 AD binary pump system coupled to an SIL-20AC autoinjector and DGU20A3 degasser coupled to an ABI 4000 quadrupole/linear ion trap (QTrap) (Applied Biosystems, Foster City, CA) operating in a triple quadrupole mode was used. Q1 and Q3 were set to pass molecularly distinctive precursor and product ions (or a scan across multiple m/z in Q1 or Q3), using N2 to collisionally induce dissociations in Q2 (which was offset from Q1 by 30–120 eV); the ion source temperature set to 500 °C. Samples were collected in borosilicate tubes (13 × 100 mm), and 1 mL of CH3OH and 0.5 mL of CHCl3 were added along with a cocktail of internal standards. Then, samples were dispersed with ultra sonicator for 30 s at room temperature and incubated overnight at 48 °C. After cooling the mixture, 75 µL of 1 M KOH in CH3OH was added. To cleave potentially interfering glycerolipids, the mixture was then sonicated briefly and incubated in a shaking water bath for 2 h at 37 °C. Using 6 µL of glacial acetic acid, the extract was brought into neutral pH, then centrifuged, and the supernatant was collected and transferred into a new tube. The extract was dried using a Speed Vac. The dried residue was reconstituted in 0.5 mL of starting mobile phase solvent for LC/MS and sonicated for 15 s, preceded by centrifugation. The supernatant was collected and transferred to an autoinjector for analysis. A combination of C18 and LC-NH2 columns were used to analyze all species of sphingolipids following the methodology developed by Dr. Alfred Merill [29]. To separate complex sphingolipids and sphingoid base 1-phosphates, reverse phase LC using a Supelco 2.1 (internal diameter) ×50 mm Ascentis Express C18 column (Sigma, St. Louis, MO, USA) and a binary solvent system at a flow rate of 0.5 mL/min with column oven set at 35 °C was used. Before injection of the sample, the column was equilibrated with a solvent mixture of 95% mobile phase A1 (CH3OH/H2O/HCOOH, 58/41/1, v/v/v, with 5 mM ammonium formate) and 5% mobile phase B1 (CH3OH/HCOOH, 99/1, v/v, with 5 mM ammonium formate) for 5 min, and after sample injection (typically 40 µL), the A1/B1 ratio was maintained at 95/5 for 2.25 min, followed by a linear gradient to 100% B1 over 1.5 min, which was held at 100% B1 for 5.5 min, followed by a 0.5 min gradient return to 95/5 A1/B1. The column was re-equilibrated with 95:5 A1/B1 for 30 s before the next run. The species of Cer, HexCer, SM, and sphingoid lipids such as sphingosine (Sph), dihydro-sphingosine (Dh-Sph), S1P, and Dh-S1P were identified based on their retention time and m/z ratio and quantified as described in previous publications [28,29,30,31].
Total RNA from HREC and BV2 cell pellets was extracted using the Ambion RNA mini extraction kit following the manufacturer’s protocol (Ambion TRIzol® Plus RNA purification kit: Life Technologies, Carlsbad, CA, USA); cDNA synthesis was carried out by SuperScript™ IV First-Strand Synthesis SuperMix (Invitrogen, Carlsbad, CA, USA). Quantitative RT-PCR was performed according to our previously published procedure [4,7]. The sequence of the primers for human gene expression (for HREC) and rodent gene expression (for BV2) are provided in Supplementary Tables S1 and S2.
For immunocytochemical analysis, cells were seeded on glass coverslips and cultured overnight. The following day, cells were incubated for 3, 24, and 48 h in normal and high-glucose conditions, washed, and fixed with 4% paraformaldehyde for 20 min at room temperature. Cells were incubated with anti-LacCer antibody (CD17, Santa Cruz, SC 65253), followed by secondary antibody (Alexa Fluor 568; Invitrogen catalog # A11037), and mounted with ProLong Diamond Antifade mountant with DAPI (Invitrogen, Carlsbad, CA, USA), then examined using a Zeiss 710 confocal laser scanning microscope (Carl Zeiss, Thornwood, New York, NY, USA) Images were captured using ZEN 2012 imaging software. The images were analyzed and quantified by ImageJ software from 5–6 field/image per condition and from at least three independent experiments.
The enzymatic activities of acidic and neutral sphingomyelinase (aSMase and nSMase, respectively) were measured in HREC and BV2 using the Amplex Red Sphingomyelinase Assay Kit (Invitrogen, Carlsbad, CA, USA) following previously published protocols [4,32].
Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA), and one-way ANOVA (for more than 2 groups) or t-test (between 2 groups) were used to calculate the significant difference between the groups. Data are presented as mean ± SEM. Statistical significance was carried out on raw data and p < 0.05 was considered as significant.
Bioactive SPL signaling is involved in numerous cellular processes in various cell types. The present study characterized the baseline sphingolipid profiles of two cell lines with distinct biological roles, HREC and BV2 cells. No report was found on comprehensive baseline SPL profiles of retinal endothelial cells and microglial cells. We cultured HREC and BV2 cells with media containing 5 mM glucose for 48 h, then performed sphingolipid analysis by LC-MS/MS. HREC were found to have total SPL levels of 220 pmol/mg protein, while the relative composition of major SPL classes was 79% SM, 14% Cer, 4% LacCer, and 3% HexCer (Figure 1A,C). BV2 cells, on the other hand, were found to have total SPL levels amounting to 130 pmol/mg protein, and the relative composition of major SPL classes was 78% SM, 6% Cer, 13% HexCer, and 3% LacCer (Figure 1B,D). While both the cell lines have similar levels of SM, they significantly differ in the composition of Cer and HexCer. Further analysis revealed differences in the mole percentage of various SPL species (Cer, SM, HexCer, and LacCer) between HREC and BV2 cells (Figure 2). In HREC, short-chain species of Cer (C16:0 and C18:0) were found to be higher compared to BV2 cells (Figure 2A,B), whereas long-chain Cer species (C24:0 and C24:1) were higher in BV2 cells (Figure 2A,B). Similarly, long-chain SM levels were higher in BV2 cells, as well as C14:0 and C22:0 species (Figure 2C,D), whereas, C16:0 SM level was higher in HREC (Figure 2C,D). HexCer composition was dominated by C22:0 and C24:0 species, which accounted for 70% of HexCer in HREC (Figure 2E), while about 50% of HexCer was distributed between carbon chains C22:0 (16%) and C24:0 (35%), and the remaining 50% were distributed between C16:0 (26%) and C24:1 (21%) in BV2 (Figure 2F). Differences were also noted in short-chain LacCer composition between HREC and BV2 (C16:0: 23% vs. 20%; C18:0: 6% vs. 1%; C22:0: 10% vs. 6%) (Figure 2G,H). In BV2, 73% of LacCer species composition were long-chain, mainly C24:0 and C24:1, while in HREC, they accounted for approximately 60% of LacCer (Figure 2G,H).
After determining the composition of different SPL species in endothelial and microglial cells under normal conditions, we investigated the SPL profiles in HREC and BV2 cells exposed to high-glucose conditions (HG; NG + 20 mM D-glucose) for 48 h. HREC and BV2 cells cultured in normal glucose (5 mM D-glucose), high L-glucose (LG; NG + 20mM L-glucose), or mannitol (Man; NG + 20mM mannitol) served as osmolarity controls. HG treatment in HREC cells showed a twofold increase in LacCer levels and a significantly reduced level of SM relative to NG HREC (Figure 3A). The relative percent composition of LacCer in HG-treated cells was 13% (* p < 0.05) (Figure 3B(iv)), whereas it was within 5% to 7% in control conditions (NG, LG, and Man) (Figure 3B(i–iii)). Among the individual LacCer species, significant increases were found in C16:0 (** p < 0.01); C22:0, C24:0, and C24:1 (* p < 0.05) in HG (Figure S1A). The relative composition of SM in HG-cultured cells was 69% (Figure 3B(iv)) compared to 79%, 77%, and 78% in NG, LG, and Man, respectively (Figure 3B(i–iii)), indicating a decrease in relative SM composition under HG. No changes were observed in minor bioactive long-chain base (LCB) SPL metabolites, such as Sph, dhSph, and S1P in HG-cultured HREC cells, but there was a significant decrease in dhS1P levels (* p < 0.05) (Figure S2B). The immunocytochemical analysis further revealed accumulation of LacCer in HG-treated HREC at all time points investigated, with data reaching statistical significance at 48 h compared to NG cells (* p < 0.05, Figure 4). HG treatment in BV2 cells revealed a significant increase in the level of total HexCer (40.0 vs. 17.0 pmol/mg protein; ** p < 0.01) (Figure 5A). In addition, the relative composition of HexCer increased by approximately 2–3-fold (Figure 5B(iv)) compared to NG, LG, and Man (Figure 5B(i–iii)). Analysis of individual HexCer species showed significant increases in C14:0 (*** p < 0.001), C18:1, C18:0 (** p < 0.01), C22:0, C24:0, and C24:1 (* p < 0.05) relative to NG (Figure S1B). Increases in HexCer coincided with significant decreases in the levels of SM and LacCer in HG-treated BV2 compared to NG (* p < 0.05). The relative composition of SM in HG-treated BV2 was reduced to 50% of total SPLs compared to NG (77%), LG (73%), and Man-cultured cells (69%) (Figure 5B). HG treatment also reduced the levels of Sph and dhS1P, and there was a significant increase in S1P levels (*** p < 0.001) (Figure S2C,D). Further comparison of the bioactive SPLs and their relative compositions indicated a significant increase in the ratio of Cer:Sph (Figure 6C) in HG-treated HREC, suggesting an accumulation of Cer. However, HG in BV2 induced a significant increase in the ratios of Cer:SM and S1P:Sph, suggesting an accumulation of S1P (Figure 6B,F). The 2.5-fold increase in S1P:Sph ratio in HG-treated BV2 was not reflected in HG-treated HREC (Figure 6E,F).
VLC SPLs, a special group of SPLs, have recently been under active investigation for their recognized roles in various human diseases and their association with the function of the vascular endothelial cells [33]. In HREC cells, HG treatment caused a significant decrease in VLC Cer and VLC SM species of carbon chain lengths C28 to C30, and a significant increase in C26:1 and C26:0 VLC SM (Figure 7A,B). A similar trend in VLC LacCer was noted with an approximately 2.5-fold increase in C26:1 and C26:0 VLC LacCer, but there were significant decreases in C28:1 and C28:0 species in HG-treated HREC (Figure 7D). The elevations of C26:1 and C26:0 fatty acids of Lac-Cer and SM, along with concomitant decreases in C28 and C30 fatty acids in all SPL, suggest that the function of the enzyme, elongation of very-long-chain fatty acid-like 4 (ELOVL4) that converts C26 fatty acids to higher-chain-length fatty acids, might have been affected in HG. In line with this observation, when gene expression was measured, ELOVL4 expression was found to be significantly lower in HG-treated HREC than NG (Figure 7E). ELOVL4 expression was reduced in control cells at after 48 h compared to 3 h (* p < 0.05). ELOVL4 expression was significantly reduced by HG at three hours (* p < 0.05) compared to controls (Figure 7E) and was further suppressed at 48 h (* p < 0.05) (Figure 7E). In BV2 cells, a significant increase in C26 to C30 VLC-HexCer (Figure S3C) was noted with unaltered VLC Cer (Figure S3A), and there was a decreasing trend of VLC-LacCer (Figure S3D). We could not detect a measurable expression of Elovl4 in BV2 cells (data not shown).
We next sought to compare SMase activity between HG and control (NG, LG, and Man)-cultured HREC and BV2 cells at 3, 24, and 48 h. In HG-treated HREC, elevated activity of aSMase at 3 and 48 h was noted compared to controls (Figure 8A). However, no significant change was observed in the activity of nSMase (Figure 8C). In BV2 cells, however, increased nSMase activity was observed at 3 and 24 h (Figure 8D), whereas aSMase activity remained unaltered with high glucose (Figure 8B).
Along with alterations of the sphingolipid profiles in HG HREC, we expected changes in the expression of their metabolic genes, cellular markers of inflammation, and vascular permeability factors. Using qRT-PCR, we studied the transcript levels of relevant marker genes at 3, 24, and 48 h in HREC cultured in NG and HG. Significant increases in the expression of Serine palmitoyl transferase 1 (SPTLC1), Ceramide synthase 4 (CERS4), Sphingomyelin phosphodiesterase 1 (SMPD1), and Sphingomyelin phosphodiesterase 2 (SMPD2) genes involved in the de novo Cer synthesis and SMase pathways, respectively, at 3 and 24 h in HG-cultured HREC were observed (Figure 9A,B). Further, increased expression of Lactosyl ceramide synthases 5 and 6 (B4GALT5 and B4GALT6), the LacCer biosynthetic genes, at 24 and 48 h in HG-treated HREC (Figure 9B,C) were observed. Increased expression of cell adhesion molecule Platelet and endothelial cell adhesion molecule 1 (PECAM1) (Figure 9A–C) and Vascular Endothelial Growth Factor A (VEGFA) (Figure 9C) were noted in HG-treated cells. In addition, elevated levels of inflammatory markers (Interleukin-6, IL-6; Tumor necrosis factor-α, TNF-α; Interleukin-18, IL-18) were observed in HG-treated HREC (Figure 9A–C). In contrast to HRECs, HG treatment in BV2 cells caused a significant increase in gene expression of Sptlc2, Sphingosine kinase 1 (Sphk1), and Glucosyl ceramide synthase (Gcs) (Figure 10A–C). In BV2, while there were no changes in vascular permeability genes, a significant increase in proinflammatory markers (Il6, Tnf-α, Il18, and inducible Nitric oxide synthase; iNos) was noted in HG-treated cells (Figure 10A–C).
Given the involvement of SPLs in a wide array of cellular functions, it follows logically that differences in the functional niches of different types of cells may be reflected by differences in their SPL composition and metabolism. SPLs are important not only as membrane components but also as mediators of numerous biological processes associated with specialized cell populations. In this study, we performed comprehensive SPL profiling of cultured immortalized retinal endothelial cells (HREC) and microglial cells (BV2), two distinct cell types which fill very different roles but whose concerted activity is tied to the pathophysiology of microvascular diseases such as diabetic retinopathy. We further analyzed how cellular stress from high glucose concentrations influences their SPL composition, SPL metabolic activity, and expression of genes tied to SPL regulation, vascular permeability and proliferation, and inflammation. We observed differences in SPL levels between HREC and BV2 cells in their respective cultured media conditions. In HREC, the levels of Cer and LacCer were higher than BV2 cells, while BV2 cells showed increased levels of HexCer compared to HREC (Figure 1A,B). HREC also showed approximately 2.5-fold elevated Cer relative composition compared to BV2 cells, while the BV2 HexCer relative composition was about 4-fold higher than HREC (Figure 1C,D). These differences in SPL composition could suggest that the roles of SPLs in cellular events may differ between cell types depending on their lineage or differences in their functional roles for that particular cell type. However, given that HREC are human-derived and BV2 cells are mouse-derived, we cannot rule out the possibility that these differences may reflect the differences in the species of origin. Therefore, the use of human microglial cell lines such as HMC3 (CRL-3304, ATCC) would have been a better choice to compare with HREC. However, considering the fact that lipidomes segregate more widely according to organs rather than species origin, we hypothesize that the lipid profiles of endothelial cells would match more closely to endothelial cells of other origins, and similarly, the lipidome of microglial cells with microglial cells of other origins [34]. A lipidomic study of 6 tissues from 32 mammalian species revealed that while genomic differences contribute to >80% of phylogenetic differences, lipidome’s contribution was only 1.9%, indicating that lipid concentrations evolve differently compared with genome sequences. Thus, the differences between HREC and BV2 are more likely due to their functional changes rather than their species of origin [34]. HREC cells showed elevated levels of LacCer and reduced SM in high-glucose conditions (Figure 3 and Figure 4), while levels of Cer remained unaltered. From the biochemical (lipidomic, SMase) and gene expression assays, it appears that these changes in the SPL profile might result from changes in SMase and LacCer synthase activity but not from the de novo or salvage pathway of Cer synthesis. Cer is a central intermediate of cellular SPL biosynthesis and generates potent signals for apoptosis and inflammation [2,3,9]. Therefore, the levels of Cer species in a cell are tightly regulated and strongly influence cell fate [5]. Although Cer levels were unchanged in glucose-stressed HREC cells, the presence of increased LacCer levels suggests conversion to forms of Cer that do not exude as strong a proapoptotic signal. We also observed increased Cer:SM and Cer:Sph ratios in high-glucose-cultured HREC (Figure 6A,C), while the S1P:Sph ratio remained unaltered (Figure 6E). These elevations of LacCer and changes in Cer:SM and Cer:Sph ratios suggest that high glucose induces changes in the SPL metabolic gradient in endothelial cells. Our earlier studies on human cadaver vitreous samples showed significantly elevated levels of Cer, LacCer, and SM from diabetic patients [35], indicating that SPL alterations could be a characteristic of DM pathology in the eye. In mouse models of STZ-induced Type 1 DM, there are decreased rates of cellular respiration and defects in the calcium retention capacity of mitochondria in cardiac tissue. These changes were associated with elevation of LacCer and could suggest an involvement of glycosphingolipids in diabetic cardiovascular pathology. In the same study, authors also reported increased expression of de novo Cer biosynthetic enzymes, Sptlc1 and CerS2, in cardiac tissue. Elevations of LacCer in knockout mouse models of neutral ceramidase deficiency further suggest a role of Cer and LacCer in the pathogenesis of glucoregulatory disorders [36]. We also demonstrated increased transcript levels of SPTLC1 and CERS2 in high-glucose-cultured cells (Figure 9), which further supports associations between glucose-induced cell stress and changes in Cer and LacCer metabolism. LacCer activation mediates signaling pathways that modulate cell proliferation, adhesion, migration, and angiogenesis [37]. Endothelial dysfunction in proliferative DR is associated with the generation of proangiogenic factors, namely, VEGF, which stimulates pathological neovascularization [38]. VEGF-mediated angiogenic activity with concomitant elevation of PECAM1 is a signature of neovascularization and may be associated with changes in LacCer in endothelial cells, suggesting that LacCer synthase could be a downstream effector of VEGF-induced angiogenesis [39,40]. In agreement with this, we observed increased transcript levels of VEGFA (Figure 9), LacCer synthase genes (B4GALT5 and B4GALT6), and PECAM1 (Figure 9) in high-glucose-cultured HREC. This correlation suggests the possibility that LacCer could be explored as a potential prognostic or diagnostic marker in hyperglycemia and DR, which may warrant further study. Along with the increase in LacCer, we noticed a 10% reduction in SM in HG-treated cells (Figure 3). SM is the most abundant SPL in any cell and the third most abundant SPL in the cell membrane and constitutes the majority of lipid rafts [5]. Changes in SM composition can significantly affect cell fate and signaling as this can change the biophysical properties of the lipid rafts and thus can influence the activity and functions of many proteins (enzymes and channels) associated with lipid rafts. Cellular stoichiometry of SM-Cer is maintained so that a 1% change in SM composition may cause a 30% change in Cer; Cer is a bioactive lipid and is well known for signaling inflammatory and apoptotic pathways [2,5]. Alternatively, Cer can also be converted to C1P or Sph, followed by S1P; all of them can have signaling roles. The reduction in SM in HG-treated cells could also be associated with increased aSMase activity, which acts either in lysosome or on the outer leaflet of the cell membrane after vesicular transport [41]. Finally, SMase activation and generation of Cer are known to be one of the major pathways by which TNFα induces cellular inflammatory signaling [2,5,42]. Thus, SMase activation in HG-treated cells and SM reduction may have important biological significance. Future studies are warranted to understand better the changes in LacCer and SM influencing cell fate and biophysical properties of the lipid rafts under high-glucose stress. Angiogenic factors play a major role in disrupting the blood–retinal barrier (BRB) in DR. High glucose stimulates the formation of microcapillaries by destabilizing tight junctions with the help of inflammatory factors [43]. The elevated activity of aSMase in endothelial cells in diabetic individuals increases Cer synthesis, which helps disrupt retinal vasculature and induce retinal inflammation [18,19]. We observed higher enzymatic activity of aSMase (Figure 8A) and elevated transcript levels of proinflammatory cytokines (IL-6, IL-18, and TNF-α) (Figure 9) in high-glucose-cultured HREC, which mimics the prevailing phenotypic characteristics observed in diabetic retinas. Likewise, in STZ-induced diabetic rat retinas, the breakdown of the BRB was accelerated by an aSMase-induced accumulation of Cer, which is associated with increases in proinflammatory cytokines IL-1β and IL-6 [44]. There is a delicate balance between short-chain Cer species, proinflammatory, and VLC Cer species, which help maintain tight junctions in the retina [45]. In the diabetic retina, this delicate balance is tilted towards higher levels of short-chain Cer species, contributing to compromised vascular integrity [33]. The decreased levels of VLC Cer might be related to the downregulated activity of ELOVL4 [33]. High-glucose treatment significantly downregulated the expression of ELOVL4 (Figure 7E) in HREC, resulting in the accumulation of C26:1 and C26:0 LacCer (Figure 7D) and SM (Figure 7B), and a reduction in VLC (C28-C30) Cer, LacCer, and SM species (Figure 7A,B,D). A similar downregulation of Elovl4 expression was observed in STZ-induced diabetic rat retinas [46]. Overexpression of ELOVL4 in cultured bovine endothelial cells was shown to restrict basal permeability and permeability induced by IL-1β and VEGF [33]. VLC SPL-mediated alteration of VEGF may affect tight junction protein ZO1 and claudin [33]. In the present study, we observed decreased VLC Cer, LacCer, and SM levels associated with increased transcript levels of VEGFA and PECAM1. Our observations of reduced levels of VLC LacCer in HG-treated HREC are supported by findings from the Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) Type 1 diabetes sub-cohort, which sought to identify glycosphingolipids as potential biomarkers in diabetic complications [47]. The study reported significantly decreased plasma VLC-LacCer levels associated with macroalbuminuria in Type 1 DM [47]. Thus, ELOVL4 and VLC SPL species might have specific roles in vascular abnormalities in diabetes, though further investigation is needed. In BV2 cells, we observed elevations of HexCer (Figure 5) and S1P (Figure S2D) and a significant reduction in SM (Figure 5). Unlike HREC, nSMase activity was increased in HG-treated BV2 cells but with no change in aSMase levels (Figure 8), which might have contributed to the decrease in SM. Additionally, the Cer generated might have been converted to HexCer to maintain cellular homeostasis by restricting Cer elevation [5]. The increase in expression of the Gcs gene may support that GCS activity might have been increased in HG-treated BV2 cells (Figure 10). Similarly, higher expression of Sphk1 (Figure 10) may support the higher activity of SPHK1 protein and increased levels of S1P in HG-treated BV2 cells. STZ-induced Type 1 diabetic rat retinas exhibited significantly increased levels of GlcCer and decreased levels of Cer, which could possibly induce endoplasmic reticulum stress, as evidenced by increased expression of Glucose-related protein 78 (GRP78) and chaperone protein CHOP [15]. Higher levels of short-chain HexCer (C14:0 and C18:0) in HG-induced BV2 cells (Figure S1B) and C16:0 LacCer in HREC (Figure S1A) follow a similar trend of pathological elevation of C16:1 and C18:2 GlcCer in the retinas of diabetic rat models [15]. Lipidomic studies of plasma in T2DM patients have also shown associations between short-chain C16:1 and C18:2 HexCer and characteristics of obesity vis-à-vis DM [48]. Earlier, we also reported higher levels of HexCer in diabetic vitreous samples [35], potentially supporting a pathological role of glycosphingolipids in DM. Glycosphingolipids, both GluCer and LacCer, are associated with plaque-related inflammation, and their significantly elevated presence in atherosclerotic plaques supports their pathological role in this regard [49,50]. GluCer and LacCer may also act as neuroinflammatory agents in Parkinson’s disease, supporting their role as inflammatory effectors within the CNS [51]. These findings highlight the potential role of SPLs in acting as stimuli or intermediate agents in inflammatory processes. Inflammation is one of the major causes of pathological changes in diabetic patients [52], and microglial cells play a critical role in inflammation within the CNS and the retina [20]. Other in vitro studies of high-glucose-cultured BV2 cells showed increased levels of inflammatory factors [53], while blocking the toll-like receptor-4/nuclear factor kappa-B pathway inhibits high-glucose-induced inflammation in BV2 cells [54], suggesting that elevation of glucose acts as a proinflammatory stimulus for microglia. Experimental models mimicking diabetic pathology using high-glucose/high-cholesterol conditions in zebrafish also showed the involvement of microglia and secretion of inflammatory cytokines [55]. Our observed elevations of inflammatory cytokine transcription in high-glucose-cultured HREC and BV2 cells also support a potential pathological effect of glycosphingolipids. Plasma samples of experimental Type 1 diabetic rats and Ins2Akita diabetic mice have shown significant elevations of bioactive S1P [56]. S1P acts as an agent for generating normal retinal vasculature in the retina. It also triggers the secretion of inflammatory cytokines and proliferating factors that can destabilize retinal vasculature [9]. The increased ratios of Cer:SM, Cer:Sph, and S1P:Sph (Figure 6B,D,F) in BV2 cells could suggest the conversion of S1P, while the increased transcript levels of Sphk1 (Figure 10A,C) indicate the elevation of S1P synthesis in high-glucose conditions. Omics studies of gene and protein expression provide helpful information for understanding tissue function. With the recent advances in shotgun lipidomics, the quantification of hundreds of lipids from a sample presents a very useful tool for correlative studies of gene and protein expression in understanding tissue function in normal and disease conditions [57]. The present study suggests that sphingolipid profiles may differ between cell populations based on their specialized functional roles, and that changes in sphingolipid metabolism induced by stressors may differ based on the identity or origin of the affected cell. In the present study, we noticed striking differences in SPL profiles between HREC and BV2 cells. In high-glucose-treated HREC, we observed alterations in levels of LacCer and VLC SPL, as well as changes in cell proliferative and proinflammatory factor gene transcription. Under similar conditions, BV2 cells showed elevations of HexCer and S1P levels and changes in proinflammatory factor transcription but no changes in proliferative factor gene transcription (Figure 9 and Figure 10). This in vitro pilot study suggests that alterations in SPL composition and metabolism are tied to the distinct responses of specialized cell types to high-glucose-induced stress, which could further imply that SPLs are involved in modulating critical processes specific to these cells. High-glucose-mediated lipid toxicity and altered metabolism are strongly implicated in the pathogenesis of DM and diabetes-related complications [11]. Hyperglycemic metabolic stress can trigger proinflammatory phenotypic changes in various cell types, including endothelial cells and immune cells, which are believed to contribute to macro- and microvascular pathology associated with DM. The interplay between these cell types is believed to contribute to the pathogenesis of DR, where chronic glucose-induced cellular stress leads to the development of retinal lesions, vascular abnormalities, and neuronal atrophy [58]. The endothelial cells lining the inner lumens of retinal blood vessels regulate vascular permeability, blood flow, and nutrient exchange, play a critical role in neovascularization (NV), and help modulate the activity and movement of immune cells. Long-term exposure to elevated glucose levels leads to chronic, low-level activation of inflammatory signaling in endothelial cells, causing changes to vascular permeability, endothelial proliferation and pathologic neovascularization, thickening of basement membranes, and loss of pericytes [59]. Vasculopathy in DR is also associated with microglial morphological changes such as process shortening and decreased ramification, which are consistent with inflammatory phenotype switching. The presence of microglial perivasculitis in DR highlights the interplay between endothelial and microglial cells in glucose-induced inflammation and vascular pathology [25]. Given the involvement of SPL signaling in inflammation and neovascularization, it is possible that conditions which induce these processes could result in changes in SPL metabolism in cells that take an active role in them. Taken together, the present study highlights the potential involvement of SPLs in high-glucose-mediated metabolic changes that could act as prognostic or/and therapeutic markers in different pathological complications, including diabetes. However, one limitation of our study is that our research does not present a causal connection between the changes in SPL profiles under high-glucose stress and SPL profile changes to observed cellular phenotypes. Further studies are needed to genetically (such as using siRNA) or pharmacologically inhibit one pathway or enzyme at a time and dissect the individual pathways or the enzyme’s causative role with the changes observed in the SPL profile or the phenotype of the cells. | true | true | true |
|
PMC9563780 | Rajamanikkam Kamaraj,Martin Drastik,Jana Maixnerova,Petr Pavek | Allosteric Antagonism of the Pregnane X Receptor (PXR): Current-State-of-the-Art and Prediction of Novel Allosteric Sites | 24-09-2022 | PXR,pregnane X receptor,allosteric site,AF-2 site,BF-3 site,PAM-antagonist,CAR | The pregnane X receptor (PXR, NR1I2) is a xenobiotic-activated transcription factor with high levels of expression in the liver. It not only plays a key role in drug metabolism and elimination, but also promotes tumor growth, drug resistance, and metabolic diseases. It has been proposed as a therapeutic target for type II diabetes, metabolic syndrome, and inflammatory bowel disease, and PXR antagonists have recently been considered as a therapy for colon cancer. There are currently no PXR antagonists that can be used in a clinical setting. Nevertheless, due to the large and complex ligand-binding pocket (LBP) of the PXR, it is challenging to discover PXR antagonists at the orthosteric site. Alternative ligand binding sites of the PXR have also been proposed and are currently being studied. Recently, the AF-2 allosteric binding site of the PXR has been identified, with several compounds modulating the site discovered. Herein, we aimed to summarize our current knowledge of allosteric modulation of the PXR as well as our attempt to unlock novel allosteric sites. We describe the novel binding function 3 (BF-3) site of PXR, which is also common for other nuclear receptors. In addition, we also mention a novel allosteric site III based on in silico prediction. The identified allosteric sites of the PXR provide new insights into the development of safe and efficient allosteric modulators of the PXR receptor. We therefore propose that novel PXR allosteric sites might be promising targets for treating chronic metabolic diseases and some cancers. | Allosteric Antagonism of the Pregnane X Receptor (PXR): Current-State-of-the-Art and Prediction of Novel Allosteric Sites
The pregnane X receptor (PXR, NR1I2) is a xenobiotic-activated transcription factor with high levels of expression in the liver. It not only plays a key role in drug metabolism and elimination, but also promotes tumor growth, drug resistance, and metabolic diseases. It has been proposed as a therapeutic target for type II diabetes, metabolic syndrome, and inflammatory bowel disease, and PXR antagonists have recently been considered as a therapy for colon cancer. There are currently no PXR antagonists that can be used in a clinical setting. Nevertheless, due to the large and complex ligand-binding pocket (LBP) of the PXR, it is challenging to discover PXR antagonists at the orthosteric site. Alternative ligand binding sites of the PXR have also been proposed and are currently being studied. Recently, the AF-2 allosteric binding site of the PXR has been identified, with several compounds modulating the site discovered. Herein, we aimed to summarize our current knowledge of allosteric modulation of the PXR as well as our attempt to unlock novel allosteric sites. We describe the novel binding function 3 (BF-3) site of PXR, which is also common for other nuclear receptors. In addition, we also mention a novel allosteric site III based on in silico prediction. The identified allosteric sites of the PXR provide new insights into the development of safe and efficient allosteric modulators of the PXR receptor. We therefore propose that novel PXR allosteric sites might be promising targets for treating chronic metabolic diseases and some cancers.
The pregnane X receptor (PXR, NR1I2) is a nuclear receptor superfamily member. Nuclear receptors are mostly activated by exogenous and endogenous compounds. Interestingly, the classified nuclear receptors conserve the DNA binding domain (DBD), but not its ligand-binding domain (LBD). PXR belongs to the NR1I subfamily together with the vitamin D receptor (VDR, NR1I1) and the constitutive androstane receptor (CAR, NR1I3) [1,2]. The PXR protein consists of an N-terminal conserved DBD and activation function domain 1 (AF-1), and the C-terminal contains an LBD with activation function 2 (AF-2). A flexible hinge region intervenes between the DBD and the LBD (Figure 1A). The artificial intelligence application AlphaFold has been used to predict the full length of the PXR structure; moreover, the error plot shows the expected position error for each residue in the sequence (Figure 1B) [3,4,5]. The PXR ligand-binding pocket (LBP) is formed by ~28 amino acids, which are typically hydrophobic residues. In the absence of an agonist, PXR can be found in the cytosol. A ligand-activated PXR forms a heterodimer with retinoid X receptor alpha (RXRα), and the complex binds coactivators (e.g., steroid receptor coactivator-1, SRC-1) (Figure 1C) [6,7,8]. The PXR is expressed in the liver, small intestine, and colon, and is traceable in the brain [9,10,11,12,13,14,15,16]. It functions as the master regulator of drug detoxification in the liver [17]. In a growing number of reports, ligand-activated PXRs alter the metabolic profile of a drug and increase the probability of drug–drug interactions (DDIs) [18,19,20]. Activation of PXRs causes significant clinically relevant DDIs with compounds whose clearance is critically dependent on hepatic biotransformation by inducible cytochrome P450s, such as CYP3A4 (Figure 1D). Under this scenario, the metabolism of the drug is significantly augmented by a PXR ligand (inducer, perpetrator) resulting in decrease in therapeutic efficacy, which may subsequently require shortening dosage intervals or increasing dosage. The most serious DDI interactions mediated by PXR activation have been reported for rifampicin [21,22,23,24]. Activated PXRs play a role in many metabolic pathways, including bile acid metabolism, lipids, and glucose homeostasis, as well as in inflammation [25]. Recently, their potential role in tumor growth, aggressiveness, relapse, and cancer drug resistance was also noted in mouse tumor xenografts for colon cancer [26,27,28,29,30]. These various physiological functions make PXRs a potentially valuable therapeutic target [25,31]. Numerous synthetic or natural ligands, mostly with agonist activities, have been discovered among currently used therapeutics and dietary or natural compounds [32,33]. Activation of PXRs using their ligands has been connected with many metabolic side effects (e.g., hypercholesterolemia or liver steatosis) as well as with intestinal cancer [10,25,26,34,35,36]. Therefore, PXR antagonists could have a significant therapeutic value, although designing a novel PXR antagonist is challenging due to the promiscuous nature of the PXR LBP [37,38,39,40,41,42]. This review summarizes and discusses our current knowledge about the allosteric modulation of PXRs. In addition, we describe our discovery of the binding function 3 (BF-3) of PXRs, which is the common allosteric binding site for other nuclear receptors. We also mention the novel allosteric site III based on in silico modeling. We propose that knowledge of allosteric modulation of PXRs as well as the characterization of the novel allosteric binding III and BF-3 sites will help us understand the biology of the PXR as well as discover novel efficient PXR antagonists.
The PXR shares common structural characteristics with other nuclear receptors [43]. The orthosteric binding site of the PXR receptor is large (>1600 Å3), dynamic, and flexible enough to bind bulky ligands [40]. The PXR LBD is characterized by an alpha-helical sandwich structure with a specialized five-stranded beta sheet [44]. The LBP is deeply embedded in the PXR LBD (Figure 2A) [45]. The orthosteric PXR LBP is formed by the helices α3/5/6/7/10/12 (Figure 2B) [46]. The PXR ligand-binding cavity was detected at the bottom of the LBD, with the ligand entrance located between the alpha 2 and 6 helices (Figure 2B) [40,44]. The PXR was observed as a homodimer in solution, and the crystal structure showed linking with β1′ strands and with support by six intermolecular hydrogen bonds in the monomer [43]. The PXR LBD shares structural similarities with the VDR and CAR, with a sequence homology of 49.4% and 48.6%, respectively. In humans, the main isoform of the PXR (NM_003889.4, variant 1) consists of 434 amino acids, and the main hydrophobic hot spots are the amino acid residues F288, W299, and Y306, which interact with all reported co-crystallized PXR ligands [40]. All available 49 PXR LBD crystal structures represent the PXR receptor in its active state and are available for coactivator interactions through helix 12 (H12) for the following transcriptional activation. Due to the destabilization of the PXR LBD, to date no crystal structure of a PXR antagonist has been reported. Therefore, as an alternative to crystallography, the biophysical technique of hydrogen-deuterium exchange mass spectrometry (HDX-MS) was used to analyze the PXR–antagonist interactions [41]. The structural characteristics of PXR ligands are determined by the nature of the PXR binding pocket: (i) the PXR LBP is large and highly hydrophobic; and (ii) the polar residues Ser247, Gln285, His407, and Arg410 define the ligand binding position in the PXR LBP via strong hydrogen bond (H-bond) interactions. Ser247 and Gln285 residues are involved in the orientation of the ligand binding and the His407 and Arg410 side chains are flexible to accommodate various sizes of ligands in the LBP [47]. Recently, a detailed examination of the PXR antagonist/inverse agonist SPA70 binding has been conducted. The distinction of the PXR antagonist/inverse agonist (SPA70) from agonist (SJB7) binding showed that the agonist should have the hydrogen bond to the polar residue of Ser247 in helix 10 as well as proximity to the helix 12 region. In contrast, the antagonist has interaction with Ser247 residue and/or has weak contacts with helix 12 [39]. Since the PXR agonist SJB7 and the antagonist SPA70 bind to the same ligand-binding pocket in the PXR in the combined agonist and antagonist conformation, it is difficult to discriminate agonist and antagonist binding sites in the PXR LBD [39,41]. Notably, a single PXR LBP mutation (W299A) converts the activity of SPA70 from inhibition to activation [48]. It has recently been reported that the designed PXR molecular glue SJPYT-195, which is composed of SPA70 linked with the CRBN ligand (thalidomide), degraded the GSPT1 translation termination factor instead of the PXR [49], but the loss of GSPT1 decreased the level of PXR protein in human colon cancer cells (SNU-C4) [49]. SJPYT-195 weakly bound the PXR LBD, suggesting that long-length linkers may be more favorable in the design of potent PROTACs for PXR [49,50]. We can also speculate that PXR allosteric sites could be another strategy for the successful targeting of PXR protein degradation (Figure 2) [51].
Allosteric modulators are usually structurally different from orthosteric ligands, and they bind to distinct sites that are spatially distant from orthosteric sites to modulate the activities of orthosteric ligands [53]. The essential features of a receptor allosteric modulation are: (1) orthosteric and allosteric sites not overlapping, i.e., there is no mutual biomolecular interaction in the binding; (2) the allosteric binding of one ligand to its site can affect the binding of the second ligand to the orthosteric site and vice versa; and (3) the effect of allosteric modulation can be positive or negative depending on the existing orthosteric ligand. This phenomenon is known as “probe dependence” [54,55]. Allosteric modulation may be inhibited due to protein–protein interactions (e.g., nuclear receptor–coactivator) or a co-bound receptor, such as in the case of the nuclear receptor homodimerization and heterodimerizations as was reported for G protein-coupled receptors [56]. Figure 3 represents the classification of the allosteric modulators and their allosteric properties: affinity modulation, efficacy modulation, reciprocity, and ceiling effect. Affinity modulation indicates the change in the structural conformation of an orthosteric LBP such that the binding affinity of an orthosteric ligand increases or decreases [57]. Efficacy modulation denotes the increase/decrease in intracellular responses (intrinsic efficacy), depending on the orthosteric ligands (agonist or antagonist) [57]. The ceiling effect or the saturability effect means that allosteric modulators are non-competitive and maintain a certain saturation level at a certain concentration [58]. Allosteric modulators may display the possibility of an absolute subtype selectivity for a target protein. For example, a 1700-fold selective allosteric inhibition for phospholipase D1 (PLD1) compared to its subtype PLD2 has been reported [59]. Allosteric modulators may improve the efficacy and potency of an agonist (e.g., from 2 to 100-fold) in a receptor or may have a synergistic effect [60,61]. Furthermore, reciprocity or allosteric activation may occur [54], which means that the receptor is activated directly, without the presence of an orthosteric ligand [62]. Currently, 202 allosteric modulators have been reported for nuclear receptors [65], with the following existing allosteric ligand binding surfaces: (i) the AF-2 site; (ii) the binding function 3 (BF-3 site); (iii) the ligand-binding pocket (synergistic); (iv) zinc fingers and response elements; and (v) the AF-1 site [66,67]. Allosteric modulators have been demonstrated for numerous nuclear receptors; for example, Gabler et al. reported imatinib as a first-in-class allosteric farnesoid X receptor (FXR) modulator that enhances agonist-induced FXR activation in a reporter gene expression assay. The imatinib analogues-16 (I-16) possess extraordinary efficacy (EC50 = 1.9 nm) and high selectivity over other nuclear receptors [68,69].
The PXR LBD is able to bind two synthetic ligands concomitantly to the orthosteric ligand binding site (Figure 4A) [70], a phenomenon which has been described previously for the PPARγ and ERβ nuclear receptors [71,72]. Two or more compounds accommodate the same binding sites in the receptors and occupy a space within the canonical LBP near the helix 3 region, leading to enhanced coactivator binding, transactivation, and target gene expression [70,71,72]. In the case of the PXR, it was found that binary cocktails of the pesticide trans-nonachlor (TNC) as well as 17α-ethinylestradiol (EE2), the active component of contraceptive pills, produce synergistic activation of the PXR and increase expression of its target CYP3A4 gene. Interestingly, the single chemicals (EE2 or TNC) were not observed to express the PXR target gene CYP3A4 in the human hepatocyte [70]. Synergistic activation of the PXR by EE2 and TNC reached full agonist activity compared to the potent agonist SR12813, but the compounds act as weak agonists when used separately (Figure 4B). Small molecules (<500 Da) bind and fill a limited portion of the PXR LBP, leaving empty volume available to accommodate a second compound [73]. This concept has been proposed as a “supramolecular ligand” assembled from two or more compounds that interact with each other within the LBP of a receptor. TNC binds to the LBP, and EE2 binds to the closely adjacent helix 12 in the PXR LBP (Figure 4C). This ectopic site near helix 12 may be considered an allosteric site, with bitopic ligands linking orthosteric and allosteric sites to achieve improved PXR affinity or selectivity [70,74]. The EE2 binding position is a resisted region of the PXR LBP that leaves a significant portion of the pocket unoccupied and available for additional interactions. Supramolecular ligands of the PXR and their properties have been reported to a very limited extent so far, and we can expect many more compounds with this activity [70].
It was found that ligands binding to the AF-2 function of the PXR alter the interactions between a coregulator peptide and the PXR. The ligand-dependent groove of the AF-2 is made up of helixes 3, 4, 5, and 12, which are mainly hydrophobic regions [37]. The orthosteric PXR LBP ends with a short helix 12 (H12), which is important for the structural organization of the AF-2 region to recruit transcriptional coregulators. Coregulators play an important functional role in the transduction of PXR signals. Both corepressors and coactivators bind to the AF-2 regions through a short amphipathic helical sequence containing the Leu-Xxx-Xxx-Leu-Leu (LXXLL) motif in coactivators or Ile/Leu-Xxx-Xxx-Ile/Val-Ile motifs in corepressors via an electrostatic interaction [75,76]. When the binding of a ligand is performed, helix H12 undergoes a significant conformational structural change that alters the overall shape of the AF-2 binding site. The PXR ligands alter the AF-2 site after binding into the PXR LBP, and thus modify the recruitment of coactivators or corepressors, resulting in different agonist or antagonist effects of the ligands [77]. All available PXR antagonist molecules act through similar agonist activation mechanisms, except for some AF-2 disruptors (Figure 5). The PXR’s ligand binding site is the same for both agonists and antagonists, but there are small residue differences with helix 12 [39,40,78]. In addition, none of the identified AF-2 site binding antagonists possess allosteric properties (selectivity, efficacy, or affinity) (Table 1) [54]. Targeting of the AF-2 site always involves issues of nuclear receptor selectivity because the AF-2 site is structurally conserved across subtypes of the nuclear receptor family [78,79]. For example, ketoconazole, the first identified PXR AF-2 site allosteric modulator and antagonist, is a common inhibitor of activated PXRs, CARs, LXRα/βs, and FXRs [80,81]. Ketoconazole binding to the surfaces of the AF-2 site suggests that ketoconazole directly blocks a coactivator (e.g., SRC-1) binding, a finding which was confirmed by a double mutant model (T248E/K277Q) in the AF-2 region of the PXR [81]. The phytoestrogen coumestrol is a natural PXR antagonist proposed to bind to a non-LBP. Biochemical binding assays and LBP-filled mutant (obliterating) studies confirm their surface binding is distinct from the orthosteric site of the LBP [82]. Additionally, computational pharmacophore and docking analyses showed that the known PXR antagonists coumestrol and sulforaphane also accommodate the AF-2 ligand-binding site [83]. In another study, the azole compound FLB-12, a derivative of the azole antifungal ketoconazole, antagonized activated PXRs in the hepatocyte cell line and in in vivo models. The triple mutant plasmid of the PXR LBP (S247W/S208W/C284W) was used to confirm its binding sites outside the PXR LBP. FLB-12 disrupts the interaction between the PXR and SRC-1, a finding which was verified by the protein pull-down assay, indicating the location of binding into the AF-2 sites. This antagonist was found to be selective and less toxic as compared to ketoconazole [84]. Ekins et al. reported residues of the PXR AF-2 ligand binding site. The identified AF-2 ligand binding site is predominantly hydrophobic, and it consists of 15 amino acids (Lys252, Ile255, Lys259, Phe264, Ile269, Glu270, Gln272, Ile273, Ser274, Leu276, Lys277, Pro423, Leu424, Glu427, and Leu428). Lys277 probably serves as a “charge clamp” for the interaction between the coactivator SRC-1 (His687) and the PXR, and it may play a significant role in the initial phase of accommodation of azole molecules into the binding groove of the PXR. Two more azole analogs, enilconazole and fluconazole, have also had interactions confirmed with the AF-2 ligand binding site, as well as antagonist activity shown towards the PXR [80,85]. Leflunomide, a drug used clinically for rheumatic arthritis therapy, acts as a PXR antagonist, but as an activator of the CAR [90]. It was shown that it inhibits the PXR/SRC-1 interaction as demonstrated by site-directed mutagenesis of the AF-2 sites [83,91]. The pentacyclic alkaloid camptothecin is known as a topoisomerase I inhibitor, and its analogs are approved for colon cancer therapy. Camptothecin was identified as attenuating CYP3A4 induction by blocking PXR activation via binding outside of the PXR LBP to prevent the recruitment of coactivators (such as SRC-1) [88]. Metformin is an antihyperglycemic agent used for the treatment of type 2 diabetes mellitus. In two-hybrid assays, we have reported that metformin interrupts the PXR’s interactions with the SRC1 coactivator and it antagonizes PXR-mediated regulation of the CYP3A4 gene in human hepatocytes. Metformin also inhibits gluconeogenesis by activating AMPK, which is necessary for the transactivation of the PXR. In addition, the compound exhibits a similar effect on the transactivation of other nuclear receptors such as the CAR and VDR [87]. Pazopanib, a tyrosine kinase inhibitor, has recently been identified as a novel potent selective antagonist of PXR activation. The compound was observed to act like camptothecin, a known coactivator disruptor. Pazopanib was claimed as an allosteric noncompetitive antagonist binding in the AF-2 site, which was confirmed with the limited proteolytic digestion technique as well as with the competitive ligand binding TR-FRET PXR coactivator assay. Still, biophysical studies are needed to authorize the positioning of the binding to AF-2 for this antagonist [86]. In addition, the T-cell lymphoma-targeting drug belinostat antagonizes drug-activated PXR-mediated gene expression. Binding assays confirmed belinostat binding to both the LBP and AF-2 binding sites, and molecular docking studies reveal that it is possible to bind to the helix 8 position to allosterically suppress PXR activation [92]. Mustonen et al. discovered a dual PXR and protein kinase inhibitor to prevent PXR-dependent chemoresistance in intestinal carcinoma cells. The two novel analogues of phenylaminobenzosuberone were identified as kinase inhibitors that concomitantly antagonize the PXR. Interestingly, these analogues 100 and 109 are structurally related but functionally different in the PXR. Compound 73 was identified as the mixed competitive and allosteric modulator of the PXR, which was confirmed with the LBP-filled triple mutant model [42].
The surface BF-3 site on nuclear receptors represents another attractive position for discovering antagonistic molecules to regulate the binding of coactivators [76,93,94,95]. The structural and functional data of the androgen receptor (AR) showed first the presence of another ligand binding site called BF-3, which is located in an area distinct from AF-2 but topographically adjacent. The novel BF-3 site allosterically influences the association of coregulators with AF-2. It was shown that the binding of 3,3‘, 5-triodothyroacetic acid (TRIAC) to BF-3 remodels the adjacent interaction site AF-2 to weaken coactivator binding, as was confirmed by X-ray crystallography. Subsequently, several allosteric inhibitors for the AR were developed, including flufenamic acid (FLUF), triiodothyronine (T3), and some novel compounds (ZINC ID: ZINC12342, ZINC2058890, ZINC3877300, and ZINC3445992) [93,95]. The AR-BF-3 is located on the N-terminal helix 1 (residues Gln670, Pro671, Ile672, and Phe673), helix 3 (Pro723, Gly724, Arg726, and Asn727), the loop between helices 3 and 4, and helix 9 (Phe826, Glu829, Leu830, Asn833, Glu837, and Arg840) in the AR (Table 2). The amino acid residues R726 and N727 interlink AF-2 and BF-3 sites, thus transforming allosteric signaling. These detailed structural features to locate BF-3 binding sites are useful for linking this concept to other nuclear receptors. Studies confirmed that the BF-3 position is also available in other nuclear receptors and the helix 3–4 loop (H3 and H4) is thought to be the signature sequence for the BF-3 position (Figure 6) [95,96]. Mutations in the AR-BF-3 site have shown a significant increase in AR activity, indicating that the AR-BF-3 site could be a co-repressor binding site, although this needs to be confirmed. Thus, its most remarkable feature is its interaction with the AF-2 surface conformation and its role in modulating the AF-2 capabilities to engage coactivator peptides [97,98,99]. Katja et al. identified novel GARRPR hexapeptide repeat sequence co-regulator motifs in the AR LBD that allow the binding of Bag-1L co-chaperon peptides. Biochemical assays and molecular modeling studies reveal that the allosteric BF-3 site is an essential domain for the interaction of Bag-1L peptides with the GARRPR motif. The disruption of Bag-1L/AR interactions via allosteric sites or residues in the BF-3 pocket represent targets for the treatment of prostate cancer [100]. The BF-3 ligand binding site, therefore, opens a new paradigm for the development of novel allosteric modulators, as opposed to the targeting of the complex orthosteric site in the PXR [37,38,51,100].
The crystal structures of nuclear receptors (PXRs, CARs, and VDRs) have enabled the identification of residues of the BF-3 site. Table 2 lists the BF-3 residues of the PXR that are located in the topologically equivalent position in the solved three-dimensional structure of AR-BF-3 residues (Figure 6). Additionally, we examined putative BF-3 sites of the CAR and VDR. Primary sequence similarity analysis reveals that the typical BF-3 residues are not conserved in other nuclear receptors, in contrast to the PXR (Figure 7A). The AR Q670 is conserved in all the nuclear receptor BF-3 sites. The PXR BF-3 site seems to be unique compared to other analyzed receptors (Table 2). The residue AR P723 is structurally conserved in the VDR (P249), FXR (P310), and CAR (P180), but not in the PXR (S262). Next, we compared structural data reported for the AR and FXR BF-3 residues by multiple sequence analysis using the Clustal Omega tool [46,96,97,100]. Based on our analysis, an alternative BF-3 ligand binding site has been proposed for the PXR (Figure 7B,C). However, more structural biophysical studies are needed to confirm the location and functionality of the allosteric site. High-throughput screening is required to test compounds in functional or biochemical assays. To accelerate the identification of selective or nonselective BF-3 modulators, new assays and technologies must be developed, since currently available methods only report the detection of a small molecule binding to the orthosteric ligand-binding site of the PXR.
Artificial intelligence (AI) is the fastest-growing technology in the life sciences and drug discovery. Computational docking, molecular dynamics simulations, and machine learning approaches are useful for designing and discovering new chemical entities with nuclear receptors [101,102]. Support vector machine algorithms (SVM) and pocket-based analysis have been used to predict allosteric sites in proteins. AI is receiving more attention in the drug discovery field and helping to advance the discovery of new allosteric modulators [103,104]. SVM is used to map and recognize similar data sets through the use of machine learning algorithms [105]. The AlloFinder web server represents a useful tool in new automated drug discovery strategies to identify allosteric modulators [104,106]. The tool was used to identify, e.g., the STAT3 inhibitor K116. Mutational and binding studies further confirmed the inhibition activity of K116 in a novel allosteric site [106]. Subsequently, AlloFinder was used to identify the allosteric site and modulators for the surface antigen CD38. The allosteric modulator LX-102 was found to target CD38 on the side opposite its enzymatic binding pocket, with this confirmed by surface plasmon resonance (SPR) and HDX-MS experiments [107]. Utilizing this web server, we identified a novel allosteric binding pocket for the PXR, which is distinct from the orthosteric LBP, AF-2, and BF-3 regions. The pocket has been termed allosteric site III. In the analysis, the PXR crystal structure was obtained from the protein data bank (PDB) (PDB-5 × 0 R) and the refined structure was submitted to the server to predict allosteric sites and for hotspot mapping (Figure 8). According to the computational calculation, the allosteric site III is located on helix 3 (α3) and the PXR unique beta sheets β1, and β1′. The identified allosteric binding residues are Thr165, Phe166, Ser167, Phe169, Asn171, Phe172, Leu174, Pro175, Val177, Val211, Leu213, Gln214, Leu215, Arg216, Trp223, Asn224, Tyr225, His242, Cys301, Arg303, Leu304, and Tyr306 (Table 3).
Antagonist molecules can correct inappropriate pathological signaling in two therapeutic settings: (i) prevention of pathological signaling before it is initiated; and (ii) reversal of such signaling once it is established. There are fundamentally two ways in which an antagonist molecule can interact with a receptor to block an agonist response: through an orthosteric or allosteric mechanism. An orthosteric blockade occurs when the antagonist physically binds to the agonist-binding site and prevents agonist binding [63]. The flexible and large LBP of the PXR enables the binding of a wide range of structurally unrelated endogenous and exogenous ligands [33,76]. Preventing pathological signaling before it is initiated by an agonist is difficult for competitive PXR antagonist molecules in the orthosteric PXR LBP. Inhibition is based on association and dissociation rates such as the offset (agonist dissociates from the receptor) and onset (antagonist binds to the receptor) of an agonist and antagonist [63]. It can be supposed that allosteric PXR modulators bind to their distinct sites on the receptor to cause a change in the conformation of the PXR protein that then alters the orthosteric agonist’s effect on the receptor (Figure 2). The new concept of the positive allosteric modulator antagonism for NRs (PAM antagonist) is based on existing GPCR allosteric modulators (e.g., palonosetron for a 5-HT3 receptor) [63,108]. PAM antagonists represent a unique class of negative allosteric modulators that antagonize the response of an agonist by increasing the affinity of the agonist to the receptor but decreasing its efficacy, making agonism overall less effective (Figure 9). These divergent reciprocities of the allosteric effect promote a “seek and destroy” mechanism of action [63,109]. PAM antagonism may offer a better way to block pathological receptor signaling than does orthosteric (antagonist) and allosteric (NAM) blockers in order to correct inappropriate pathological signaling in two therapeutic settings: (i) prevention of receptor activation; and (ii) reversal of preexisting agonist activation [63,110]. The development of PAM antagonists is not easy. Functional assays and other emerging technologies such as advanced biophysical and biochemical assays, high-throughput CRISPR engineering and mutant techniques, and molecular docking to X-ray crystal structures could be used for this unique challenge [63]. We can propose that PXR PAM antagonist molecules could be refined to target specific endogenous signaling pathways with special therapeutic effects [63].
Most ligands bind to the orthosteric PXR ligand binding pocket, which results in transcriptional upregulation (induction) or downregulation (transrepression) of its target genes. Allosteric modulation of the PXR opens up many questions with respect to clinical application and consequences for DDIs. At present, we do not have data for the clinical consequences of PXR inhibition on the putative downregulation of key target PXR genes. Moreover, PXR antagonism has been proposed to alleviate DDIs mediated by PXR inducers [87,112]. Therefore, we are at the beginning of the discovery of efficient PXR allosteric modulators that can help us titrate drug metabolism as well as eliminate PXR-mediated DDIs. A growing number of small molecules have been shown to bind to the PXR AF-2 coactivator binding site, although none of them have significant affinity (in nanomolar concentration) to the allosteric site (Table 1). The binding site of PXR has not yet been studied in detail because it is challenging to crystallize PXRs without a coactivator. Biochemical assays and technologies must be developed to accelerate the identification of more efficient and highly selective allosteric PXR modulators in the AF-2 site. In addition, the crystal structure of PXR and molecular modeling have enabled us to identify novel PXR allosteric sites. The novel allosteric ligand binding III and BF-3 sites have been proposed for the PXR in the report, although this awaits confirmation using structural and biophysical techniques. The identified allosteric sites (Figure 10) provide new information in the development of safe and efficient allosteric modulators of the PXR receptor, a promising target for treating chronic metabolic diseases and cancers. Moreover, other avenues for targeting the PXR are via the development of micro/siRNA strategies (to block or promote the degradation of PXR mRNA), or the design of peptidomimetic inhibitors of the PXR–coactivator interaction. Artificial intelligence combined with advanced chemical-biological and high-throughput screening technologies are expected to help us with the development of novel allosteric modulators of PXRs to fight against metabolic disorders, drug resistance, and colon and breast cancers. The overview presented in this review may stimulate interest and facilitate scientific effort towards the development of allosteric modulators for the PXR nuclear receptor. | true | true | true |
|
PMC9563794 | Soudeh Ghafouri-Fard,Tayyebeh Khoshbakht,Bashdar Mahmud Hussen,Mohammad Taheri,Mohammadreza Hajiesmaeili | A review on the role of LINC00467 in the carcinogenesis | 13-10-2022 | LINC00467,Cancer,lncRNA,Biomarker,Expression | LINC00467 is an example of long intergenic non-coding RNAs whose roles in human disorders are being identified. This gene coding LINC00467 is located on chromosome 1: 211,382,736 − 211,435,570 forward strand. This lncRNA has been firstly recognized through a microarray-based lncRNA profiling as an N-Myc target in neuroblastoma cells. Further studies have shown up-regulation of LINC00467 in different cancer including those originated from brain, gastrointestinal tract, lung and breast. It acts as a molecular sponge for miR-339, miR-138-5p, miR-107, miR-133b, miR-451a, miR-485-5p, miR-7-5p, miR-485-5p, miR-339-3p, miR-200a, miR-1285-3p, miR-299-5p, miR-509-3p, miR-18a-5p, miR-9-5p and miR-20b-5p. LINC00467 can regulate activity of NF-κB, STAT1, Wnt/b-catenin, Akt and ERK1/2 signaling pathways. Accumulating evidence indicates oncogenic role of LINC00467. The current review article aims at providing an overview of LINC00467 in the carcinogenesis. | A review on the role of LINC00467 in the carcinogenesis
LINC00467 is an example of long intergenic non-coding RNAs whose roles in human disorders are being identified. This gene coding LINC00467 is located on chromosome 1: 211,382,736 − 211,435,570 forward strand. This lncRNA has been firstly recognized through a microarray-based lncRNA profiling as an N-Myc target in neuroblastoma cells. Further studies have shown up-regulation of LINC00467 in different cancer including those originated from brain, gastrointestinal tract, lung and breast. It acts as a molecular sponge for miR-339, miR-138-5p, miR-107, miR-133b, miR-451a, miR-485-5p, miR-7-5p, miR-485-5p, miR-339-3p, miR-200a, miR-1285-3p, miR-299-5p, miR-509-3p, miR-18a-5p, miR-9-5p and miR-20b-5p. LINC00467 can regulate activity of NF-κB, STAT1, Wnt/b-catenin, Akt and ERK1/2 signaling pathways. Accumulating evidence indicates oncogenic role of LINC00467. The current review article aims at providing an overview of LINC00467 in the carcinogenesis.
Long non-coding RNAs (lncRNAs) are a group of transcripts having sizes larger than 200 nt. They are regarded as important epigenetic regulators that control epigenetic mechanisms principally in the nucleus, modulating transcription of genes through changing histone or DNA methylation and acetylation marks [1]. The majority of identified lncRNAs are transcribed by RNA polymerase II, thus having similar structures with mRNAs. While sharing many features with mRNAs, these widely expressed transcripts have distinct roles from mRNAs. Notably, function of lncRNAs is related with their particular subcellular localization [2]. In addition to modulation of chromatin function, lncRNAs can influence establishment and functions of nuclear bodies, change mRNAs stability and their translation and affect activity of signaling pathways [2]. GENECODE catalog of lncRNAs have classified these transcript into distinct categories of long intergenic non-coding (linc)-RNAs, antisense transcripts, intronic, and non-overlapping antisense transcripts [3]. LINC00467 is an example of the first group of lncRNAs whose roles in human disorders are being identified. This transcript is encoded by a gene located on chromosome 1: 211,382,736 − 211,435,570 forward strand. This gene has 28 transcripts with sizes ranging from 3536 bp (LINC00467-201) to 469 bp (LINC00467-204). This lncRNA has been firstly recognized as an N-Myc target in neuroblastoma cells through a microarray-based transcriptome profiling [4]. Further studies have indicated abnormal expression of LINC00467 in a wide variety of cancer cell lines and clinical samples. Moreover, several studies have assessed functional roles of LINC00467 in xenograft models of cancers. The current review article aims at providing an overview of LINC00467 in the carcinogenesis through summarization of three mentioned lines of evidence.
Expression of LINC00467 has been shown to be elevated in acute myeloid leukemia (AML) cell lines. LINC00467 silencing has inhibited the malignant features of these cells. Notably, expression of miR-339 has been up-regulated after LINC00467 silencing. Moreover, expression of miR-339 target gene SKI has been decreased following this intervention. Since miR-339 silencing can chiefly eliminate the impact of LINC00467 silencing in AML cell lines, miR-339/SKI axis has been proposed as the molecular axis mediating the effects of LINC00467 [5]. In breast cancer cells, LINC00467 silencing has impeded proliferation, migratory potential, invasive features and epithelial-to-mesenchymal transition (EMT), while its up-regulation has led to opposite impacts. LINC00467 could down-regulate miR-138-5p through functioning as a molecular sponge for this miRNA. Moreover, LINC00467 could enhance expression of LIN28B through directly interacting with it [6]. Another in silico study in breast cancer has shown possible role of LINC00467 in the regulation of peroxisomal lipid metabolism and immune response through targeting miRNAs [7]. In cervical cancer cells, expression assays have detected high expression of LINC00467 and KIF23, and down-regulation of miR-107. LINC00467 has been shown to be mainly localized in the cytoplasm, where it acts as a molecular sponge for miR-107. LINC00467 silencing or miR-107 overexpression has blocked proliferation and decreased migration, invasion, and EMT [8]. In squamous cell carcinoma cells, LINC00467 can also enhance EMT through influencing activity of miR-299‐5p/USP48 axis [9]. Moreover, LINC00467 can influence response of hepatocellular cancer cells to Axitinib via acting as a molecular sponge for miR-509-3p and enhancing expression of PDGFRA [10]. miR-18a‐5p/NEDD9 [11] and miR-9-5p/PPARA [12] molecular axes are other routes of participation of LINC00467 in the pathoetiology of hepatocellular carcinoma as revealed through in vitro assays. In osteosarcoma cells, LINC00467 has been shown to sponge miR‑217 and increase expression of KPNA4 [13] which facilitates progression of this type of cancer. Moreover, the sponging effect of LINC00467 on this miRNA leads to up-regulation of HMGA1 which enhances growth and metastatic abilities of these cells [14]. LINC00467 has also been shown to increase proliferation of lung adenocarcinoma cells through influencing miR-20b-5p/CCND1 activity [15]. Moreover, LINC00467 increases stemness of lung cancer cells through sequestering miR-4779 and miR‐7978 [16].
Experiments in bladder cancer cells have shown the role of LINC00467 in enhancement of proliferation and invasive properties of these cells. Mechanistically, LINC00467 directly binds to NF-kb-p65 transcript, enhances its stability and promotes its nuclear translocation for further activation of the NF-κB signaling [17]. SiRNA-mediated LINC00467 silencing has suppressed proliferation, invasiveness and metastatic potential of colorectal cancer cells. Mechanistically, LINC00467 could affect expression of Cyclins D1 and A1, CDK2, CDK4, Twist1 and E‑cadherin [18]. LINC00467 can also promote invasive properties and block apoptosis of squamous cell carcinoma cells through sponging miR-1285-3p and enhancing expression of TFAP2A [19]. In hepatocellular carcinoma cells, LINC00467 has been shown to bind with IGF2BP3 and stabilize TRAF5, thus promoting proliferation and metastatic abilities of these cells [20].
Expression of LINC00467 has been shown to be suppressed by N-Myc. In fact, N-Myc directly binds to the promoter of LINC00467 gene, decreasing its promoter activity. N-Myc has also inhibited expression of the down-stream gene of LINC00467, i.e. RD3 via directly binding to its promoter (Fig. 1). SiRNA-mediated silencing of LINC00467 has led to up-regulation of the tumor suppressor gene DKK1. This intervention has also decreased viability of neuroblastoma cells and increased their apoptosis. Notably, co-transfection of LINC00467 siRNA and DKK1 siRNA has blocked the effect of LINC00467 silencing [4]. Table 1 shows function of LINC00467 in cell lines derived from different types of cancers.
Up-regulation of LINC00467 has enhanced breast cancer growth, whereas its silencing has inhibited lung metastases in vivo [6]. Furthermore, LINC00467 knock down or miR-107 over-expression has suppressed tumorigenic ability of cervical cancer cell in xenograft models [8]. Similar studies in AML, bladder cancer, colorectal cancer, esophageal carcinoma, glioma, hepatocellular carcinoma, lung cancer and prostate cancer have consistently confirmed oncogenic effects of LINC00467 (Table 2).
Assessment of expression data from a GEO dataset and the TCGA database has revealed up-regulation of LINC00467 in bladder cancer samples and negative correlation between its expression and patients’ prognosis [17]. Expression assays in patients with breast cancer has also verified over-expression of LINC00467 in cancerous tissues compared with nearby normal samples. Moreover, up-regulation of LINC00467 has been associated with poor overall survival (OS) [6]. Another study has indicated association between LINC00467 over-expression and tumor metastases and poor prognosis. Genomic and epigenetic analyses have shown the impact of copy number amplification, chromatin configuration, and methylation status of DNA on expression of this lncRNA. Copy number amplification and up-regulation of LINC00467 has been associated with the lower levels CD8 + and CD4 + T cells infiltrations [7]. LINC00467 level has also been reported to be elevated in colorectal cancer tissues compared with normal colon mucosal counterparts. In silico analyses available datasets have confirmed correlation between over-expression of LINC00467 and poor OS and recurrent-free survival rate [18]. The association between over-expression of LINC00467 and poor clinical outcome has been verified in different cancers, including bladder cancer, breast cancer, colorectal cancer, glioma, lung cancer, osteosarcoma and testicular germ cell tumor (Table 3).
Numerous studies have indicated up-regulation of LINC00467 in different types of cancers. Mechanistically, this lncRNA can be up-regulated through DNA demethylation and copy number variations. The sponging effect of LINC00467 on miRNAs has been well assessed in different cancer cell lines. Through this mechanistical route, LINC00467 can affect activity of miR-339/SKI, miR-107/KIF23, miR-133b/FTL, miR-485-5p/DPAGT1, miR-7-5p/EGFR, miR-339-3p/IP6K2, miR-200a/E2F3, miR-1285-3p/TFAP2A, miR-299-5p/USP48, miR-509-3p/PDGFRA, miR-18a-5p/NEDD9, miR-9-5p/PPARA, miR-20b-5p/CCND1, miR-125a-3p/SIRT6, miR-217/KPNA4, miR-217/HMGA1 and miR-494-3p/STAT3 axes. Moreover, LINC00467 can influence activity of NF-κB, STAT1, Wnt/b-catenin, Akt and ERK1/2 signaling pathways. Most notably, LINC00467 has been shown to increase EMT in breast, cervical, colorectal, head and neck and prostate cancer as well as osteosarcoma. Thus, strategies to decrease expression of LINC00467 are expected to affect tumor invasion and metastasis. LINC00467 has a possible role in the tumor microenvironment and immune evasion. Copy number variations within LINC00467 have been associated expression levels of this lncRNA, immune infiltration in lung adenocarcinoma and poor clinical outcome [38]. Moreover, LINC00467 expression in breast cancer has been associated with immune infiltration [7]. Up-regulation of LINC00467 has been associated with poor prognosis of patients with bladder cancer, breast cancer, colorectal cancer, glioma, lung cancer, osteosarcoma and testicular germ cell tumor. Thus, LINC00467 is a putative prognostic marker in cancers. However, the potential of this lncRNA as a diagnostic marker has not well studied. Future studies should focus on this aspect. Expression assays of LINC00467 particularly in biofluids such as serum and urine would pave the way for establishment of non-invasive methods for cancer diagnosis. Identification of additional miRNA targets of LINC00467 is expected to clarify the molecular mechanisms and signaling pathways being affected by this lncRNA. This would help in design of novel and efficient targeted therapies for cancer. Based on the critical roles of LINC00467 in the regulation of cell apoptosis, it is expected that modification of its expression affects response of cancer cells to anti-cancer modalities. This function of LINC00467 has been verified in hepatocellular carcinoma cells where its silencing has enhanced sensitivity to Axitinib ([10]. Finally, the presence of single nucleotide polymorphisms within LINC00467 would affect expression or function of this lncRNA. Therefore, genotyping of these variants would help in recognition of risk factors for different types of cancer.
LINC00467 is regarded as an oncogenic lncRNA in humans. Thus, strategies to down-regulate its expression are theoretically effective in reduction of tumor burden. The most challenging issue in this regard is establishment of effective ways to convey LINC00467-targetted therapies in a specific way to cancer cells and avoid off-target effects. | true | true | true |
|
PMC9563928 | Maximilian N. Kinzler,Falko Schulze,Steffen Gretser,Nada Abedin,Jörg Trojan,Stefan Zeuzem,Andreas A. Schnitzbauer,Dirk Walter,Peter J. Wild,Katrin Bankov | Expression of MUC16/CA125 Is Associated with Impaired Survival in Patients with Surgically Resected Cholangiocarcinoma | 27-09-2022 | cholangiocarcinoma,mucins,CA-125 antigen,survival,surgical oncology | Simple Summary MUC16/CA125, a commonly used blood biomarker of ovarian cancer, is associated with cancer proliferation in several tumor entities. The data on MUC16 expression in cholangiocarcinoma (CCA) tissue are very limited. We observed a remarkable proportion of MUC16 (+) patients with surgically resected CCA as 57 of 168 (34%) patients with CCA had evidence of MUC16 expression in tumor tissue. We found a significantly impaired overall survival for MUC16 (+) patients (27.4 months) in comparison to MUC16 (−) patients (56.1 months). Our data demonstrate that MUC16 (+) is an independent risk factor for poor survival in CCA patients. Abstract MUC16/CA125 is associated with cancer proliferation in several tumor entities. The data on MUC16 expression in cholangiocarcinoma (CCA) tissue are very limited. The aim of this study was to assess the MUC16 status and its impact on survival in CCA patients. All the patients with surgically resected CCA that were diagnosed between August 2005 and December 2021 at the University Hospital Frankfurt were retrospectively analyzed. A 7-Mucin biomarker panel was assessed by immunohistochemistry. For overall survival (OS), Kaplan–Meier curves and Cox-regression analyses were performed. Randomly selected intrahepatic cholangiocarcinoma (iCCA) were further processed for differential expression profiling. A total of 168 patients with CCA were classified as MUC16 (−) (66%, n = 111) and MUC16 (+) (34%, n = 57). Subgroup analyses revealed a median OS of 56.1 months (95% CI = 42.4–69.9 months) and 27.4 months (95% CI = 15.8–39.1 months) for MUC16 (−) and MUC16 (+), respectively (p < 0.001). In multivariate analysis, MUC16 (+) (HR = 1.6, 95% CI = 1–2.6, p = 0.032) was an independent risk factor for poor prognosis. Prominently deregulated pathways have been identified following MUC16 expression, overrepresented in cell cycle and immune system exhaustion processes. These findings suggest including MUC16 in clinical routine diagnostics as well as studying its molecular pathways to identify further mechanistic key players. | Expression of MUC16/CA125 Is Associated with Impaired Survival in Patients with Surgically Resected Cholangiocarcinoma
MUC16/CA125, a commonly used blood biomarker of ovarian cancer, is associated with cancer proliferation in several tumor entities. The data on MUC16 expression in cholangiocarcinoma (CCA) tissue are very limited. We observed a remarkable proportion of MUC16 (+) patients with surgically resected CCA as 57 of 168 (34%) patients with CCA had evidence of MUC16 expression in tumor tissue. We found a significantly impaired overall survival for MUC16 (+) patients (27.4 months) in comparison to MUC16 (−) patients (56.1 months). Our data demonstrate that MUC16 (+) is an independent risk factor for poor survival in CCA patients.
MUC16/CA125 is associated with cancer proliferation in several tumor entities. The data on MUC16 expression in cholangiocarcinoma (CCA) tissue are very limited. The aim of this study was to assess the MUC16 status and its impact on survival in CCA patients. All the patients with surgically resected CCA that were diagnosed between August 2005 and December 2021 at the University Hospital Frankfurt were retrospectively analyzed. A 7-Mucin biomarker panel was assessed by immunohistochemistry. For overall survival (OS), Kaplan–Meier curves and Cox-regression analyses were performed. Randomly selected intrahepatic cholangiocarcinoma (iCCA) were further processed for differential expression profiling. A total of 168 patients with CCA were classified as MUC16 (−) (66%, n = 111) and MUC16 (+) (34%, n = 57). Subgroup analyses revealed a median OS of 56.1 months (95% CI = 42.4–69.9 months) and 27.4 months (95% CI = 15.8–39.1 months) for MUC16 (−) and MUC16 (+), respectively (p < 0.001). In multivariate analysis, MUC16 (+) (HR = 1.6, 95% CI = 1–2.6, p = 0.032) was an independent risk factor for poor prognosis. Prominently deregulated pathways have been identified following MUC16 expression, overrepresented in cell cycle and immune system exhaustion processes. These findings suggest including MUC16 in clinical routine diagnostics as well as studying its molecular pathways to identify further mechanistic key players.
Cholangiocarcinoma (CCA) represents a heterogenous group of highly malignant cancers with poor prognosis. Despite relatively rare occurrence, CCA is the second most common primary hepatic cancer after hepatocellular carcinoma (HCC) and its incidence is increasing globally [1,2]. Depending on its localization, CCA can be divided into intrahepatic (iCCA), perihilar (pCCA), or distal (dCCA) harboring distinct clinical and pathological features. The only curative therapy for all subclasses is surgical resection (R0 resection). However, even for CCA that is resected with curative intent, the prognosis is poor as most patients are not cured: the median survival ranges from about 30 months for iCCA to 38 months for pCCA [3]. Due to this devastating outcome of patients with CCA, the identification of further potential prognostically relevant markers is of utmost importance to improve the clinical management of these patients. Mucins are high molecular weight glycoproteins that are produced by epithelial cells that serve multiple functions including lubrication, cell signaling, and maintaining epithelial integrity by forming a chemical barrier [4]. However, the aberrant expression of mucins has been reported to promote cancer development in several entities including CCA [5,6,7,8,9] by mimicking epithelial composition and structure, thus evading potential immune surveillance by growth factor and cytokine capture motifs. MUC16 expresses the peptide epitope cancer antigen 125 (CA125), a routinely used blood biomarker of ovarian cancer, promoting cancer cell proliferation while inhibiting the anticancer immune response [10]. Furthermore, a crucial role of MUC16 is also suggested in pancreatic, colorectal, and non-small cell lung cancer [11,12,13] while its overexpression in breast cancer tissue regulates proliferation and anti-apoptosis via JAK-STAT signaling [6]. Apart from few studies evaluating the impact of preoperative CA125 serum levels on outcome [14,15], studies investigating the role of MUC16 expression in tumor tissue from CCA patients remain scarce so far. In more detail, only two studies indicated that the expression of MUC16, as well as the co-expression with mesothelin, is a risk factor for poor outcome in mass-forming iCCA and pCCA, respectively [16,17]. As the outcome of CCA patients remains poor even after surgical resection, novel therapeutic strategies for CCA patients are of high interest. Intriguingly, MUC16 has emerged as a potential target for novel cancer therapies using monoclonal antibodies or immunotherapy, particularly for ovarian and pancreatic cancer [18,19,20]. To test this hypothesis, we assessed the frequency of MUC16 in a large retrospective cohort of CCA patients that were treated in our tertiary hospital. In conclusion, this study evaluates the impact of MUC16/CA125 expression and potential pathway alterations in patients with CCA on survival after resection in curative intent for all known histopathological subtypes in a large cohort and its potential role as a prognostic marker.
All patients that were treated with surgically resected (R0, R1) cholangiocarcinoma at Frankfurt University Hospital between August 2005 and December 2021 were retrospectively analyzed. Histopathological confirmation was assessed independently by expert pathologists of the Dr. Senckenberg Institute of Pathology, University Hospital Frankfurt. Clinical data (date of birth, gender, tumor stage, tumor size, comorbidities, laboratory parameters, and follow-up) were collected from electronic medical records. The tissue samples that were used in this study were provided by the University Cancer Center Frankfurt (UCT). Written informed consent was obtained from all the patients and the study was approved by the Institutional Review Boards of the UCT and the Ethical Committee at the University Hospital Frankfurt (project-number: SGI-13-2018).
Formalin-fixed paraffin-embedded (FFPE) tissue samples were retrieved from the archive of the Dr. Senckenberg Institute of Pathology, University Hospital Frankfurt. For the construction of TMA, annotations of defined regions of interest (ROI) containing representative tumor area were set to a core diameter of 1 mm. Via slide overlay function, digital annotations were adapted to the donor tissue and transferred to a blank recipient paraffin matrix using the TMA Grandmaster (3DHISTECH, Budapest, Hungary). HE-stained slides were automatically processed on a Tissue-Tek Prisma Plus staining device (Sakura Finetek, Torrance, CA, USA). IHC was conducted using the DAKO FLEX-Envision Kit (Dako/Agilent, Santa Clara, CA, USA) and the fully automated DAKO Omnis staining system (Dako/Agilent, Santa Clara, CA, USA) according to the manufacturer’s instructions. We performed staining of MUC1 (Clone: E29; ready to use; incubation time: 15 min; Dako/Agilent, Santa Clara, CA, USA), MUC2 (Clone: CCP58; ready to use; incubation time: 30 min; Dako/Agilent, Santa Clara, CA, USA), MUC4 (Clone: 8G7; dilution: 1:200; incubation time: 30 min; Bio SB, Santa Barbara, CA, USA), MUC5AC (Clone: CLH2; ready to use; incubation time: 30 min; Dako/Agilent, Santa Clara, CA, USA), MUC6 (Clone: MRQ20; dilution: 1:500; incubation time: 30 min; Cell Marque, Rocklin, CA, USA), MUC16 (Clone: M11; ready to use; incubation time: 60 min; Dako/Agilent, Santa Clara, CA, USA), Ki-67 (Clone: MIB-1 ; ready to use; incubation time: 20 min; Dako/Agilent, Santa Clara, CA, USA), and PD-L1 (Clone: 22C3 ; ready to use; incubation time: 40 min; Dako/Agilent, Santa Clara, CA, USA). The stained slides were scanned with the Pannoramic slide scanner (3DHISTECH, Budapest, Hungary). For MUC antibodies, staining in more than 1% of the tumor cells was considered positive. IHC analysis was performed by two independent investigators. The assessment of Ki-67 and PD-L1 was performed by a pathologist with three years of experience (S.G.). TMA cores with either the absence of representative tumor tissue or the presence of staining artifacts were excluded from the analysis. Representative images of absence and presence of immunohistochemical expression of MUC16 in CCA tissue are shown in Figure 1. A representative control image of MUC16 staining in cancerous and paracancerous tissue is shown in Supplementary Figure S3.
Representative tumor material of the primary tumor was retrieved by a 1 mm core. The RNA was isolated using the truXTRAC FFPE total NA Kit (Covaris, Woburn, MA, USA) based on focused ultrasonification and column purification according to the manufacturer’s instructions. Nanostring nCounter® Platform and Immune Exhaustion Panel v1 were used to enrich a commercially available function-specific panel of 798 genes by hybrid capture technique (Nanostring, Seattle, WA, USA). Nanostring nSolver™ software v4 and implemented nCounter® Advanced Analysis module v2.0.115 were used for subsequent raw data processing and normalization by internal controls following differential supervised analysis between MUC1-positive (n = 7) and -negative CCA patients indicated by previous immunohistochemistry. ClueGO v.2.5.6 and CluePedia v.1.5.6 functional classification and network annotation was applied in order to identify enriched genes that were associated to overrepresented gene ontologies based on REACTOME_pathways (updated 17.08.2022) [21,22]. The Ingenuity Pathway Analysis (IPA v.76765844; QIAGEN, Hilden, Germany) tool was additionally used to predict network-associated biomarkers as well as functional terms (Fx) and therapeutical agents (Rx).
We compared baseline clinicopathological characteristics between patients with an absence and presence of MUC16. The categorial variables are presented as frequencies and percentages and the continuous variables are shown as means with standard deviations. The categorial and continuous variables were compared using the Student’s t-test and chi-square test, respectively. The overall survival (OS) was defined as the time of onset of disease until death or date of last follow-up. The date of last follow-up was treated as censored observation. Survival was compared using the log-rank test. The Kaplan–Meier curves for survival were derived to visualize the comparison between the absence and presence of MUC16 expression. Cox regression analysis was performed to assess the risk factors influencing patient survival. We preliminarily used univariate Cox regression analysis to screen our variables. We then included the variables with p < 0.05 into the multivariate Cox regression analysis. The adjusted common odds ratios are reported with 95% CIs to indicate statistical precision. The significance level was set to p < 0.05. For disease survival-rate, patients without recurrence and patients who were alive at the time of data collection were counted as patients with disease survival as we only enrolled patients with surgically resected CCA with curative intent. Furthermore, patients that were treated as censored observation due to lost to follow-up were separately classified as “lost to follow-up” in this analysis. Patients for whom the exact cause of death could not be determined were classified as “not available”. We used a strict definition for cancer-related death as only patients who died in a palliative setting in our hospital, patients with dedicated palliative care treatment, patients in best-supportive care situation as well as patients who died from CCA therapy (e.g., side effects from chemotherapy) were defined as cancer-related death. Furthermore, patients with postoperative death (during the same hospital stay as surgery) were classified as cancer-related death as well. All the data were analyzed with SPSS 27 (IBM; Armonk, NY, USA) statistical software.
In total, 260 patients with surgically resected CCA in our tertiary hospital were analyzed in this study. A total of 69.2% (n = 180) of the patients were suitable for TMA construction. Of the remaining 168, 66% (n = 111) were classified as MUC16 (−) while 34% (n = 57) were MUC16 (+) (Figure 2). The MUC16 (+) patients had more frequently positive serum CA-19/9 (p = 0.011), higher Pn status (p = 0.034), and increased Ki-67 proliferative index (p = 0.027). In contrast, MUC16 (−) was associated with larger tumor size (p < 0.001). Remarkably, MUC16 was more frequently expressed in iCCA compared to extrahepatic CCA (p < 0.001). MUC16 (−)/(+) patients with CCA did not differ in any other clinicopathological findings including common risk factors such as viral hepatitis, primary sclerosing cholangitis, liver cirrhosis, or cholelithiasis. As expected, the serum levels of CA125 were determined in only a very small minority during clinical routine diagnostics (n = 8). The baseline clinicopathological characteristics are summarized in Table 1. The details on distinct patterns of recurrence are depicted in Supplementary Table S4.
The median OS for all 180 patients with surgically resected CCA that were included in this study was 29.1 months. In line with the literature, the Kaplan–Meier curves revealed a significant impact on OS in CCA patients for the expression of MUC4 (p < 0.001) or MUC5AC (p = 0.006) while there was no difference in the survival for MUC1 (p = 0.44), MUC2 (p = 0.3) and MUC6 (p = 0.983) in our study cohort [23]. In addition, the Kaplan–Meier curves indicated a median OS of 56.1 months (95% CI = 42.4–69.9 months) for MUC16 (−) in comparison to 27.4 months (95% CI = 15.8–39.1 months) for patients with MUC16 (+), thus showing a significant difference between both groups (p < 0.001) (Figure 3A). To further investigate the impact of MUC16 expression on OS, the three CCA subtypes were analyzed separately. In iCCA, the OS rates were 53.4 months (95% CI = 37–69.9) for MUC16 (−) patients and 19.1 (95% CI = 10.6–27.6) for MUC16 (+) (p = 0.01) (Figure 3B). In line with these findings, for MUC16 (−)/(+) pCCA patients, the median OS was 57.8 months (95% CI = 32.2–83.4) and 20.8 (95% CI = 5.9–35.7), respectively (p = 0.028) (Figure 3C). Correspondingly, MUC16 (+) is associated with impaired OS rates in patients with dCCA (26.6 months (95% CI = 10–43.2)) in contrast to MUC16 (−) patients (52.7 (95% CI = 19.9–85.4)) although these results did not reach statistical significance (p = 0.317) (Figure 3D). Additionally, the disease survival rate was higher in MUC16 (−) compared to MUC16 (+) patients. However, these results did not reach statistical significance (p = 0.385) (Table 1).
As our results indicate a marked impact of MUC16 expression on survival rates in our study, we further performed univariate and multivariate Cox regression analysis to identify the correlating risk factors. Interestingly, the univariate analysis determined MUC16 (+) as a significant risk factor of OS (HR = 1.9, 95% CI = 1.3–2.8, p < 0.001). In addition, the presence of tumor marker CA-19/9 could be described as a significant risk factor as well (HR = 2.1, 95% CI = 1.4–3.2, p < 0.001). Furthermore, multiple tumors (HR = 1.8, 95% CI = 1.2–2.6, p = 0.002), pathological grade 3 (HR = 4.8, 95% CI = 1.2–20, p = 0.031), ECOG 1 (HR = 2.7, 95% CI = 1.8–3.9, p < 0.001), M1 status (HR = 2.7, 95% CI = 1.5–4.9, p = 0.001), and R1 status (HR = 1.6, 95% CI = 1.1–2.4, p = 0.018) also served as significant risk factors in univariate analysis. To further investigate independent risk factors, multivariate analysis was performed. Multivariate Cox regression analysis revealed that MUC16 (+) (HR = 1.6, 95% CI = 1–2.6, p = 0.032), ECOG 1 (HR = 2, 95% CI = 1.3–3.2, p = 0.003), multiple tumors (HR = 1.7, 95% CI = 1.1–2.7, p = 0.025), as well as the presence of tumor marker CA-19/9 (HR = 1.6, 95% CI = 1–2.6, p = 0.035) serve as independent risk factors for overall survival for CCA patients in our study (Table 2).
Randomly selected iCCA have been included for preliminary expression analysis in dependence of MUC16 status. In total, 324 genes have been identified as differentially regulated among groups with a log2 fold change that was larger than 1 (linear fold change ≥2) and a p-value ≤ 0.05. A p-value correction that was based on the Benjamini–Yekuteli method (BY p-value) allowed 197 genes to be selected as prominently deregulated between groups (BY p-value ≤ 0.2). The top 40 candidates (BY p-value ≤ 0.04, log2 fold change ranging from 2–4.71) have been selected for further functional classification and are displayed in a volcano plot (Supplementary Figure S1, Supplementary Table S1). The selection revealed distinct pathways such as cell cycle checkpoint processes (e.g., CCNB1 among others); cytokine and interleukin signaling in immune system (e.g., CCL3, CCL4 among others); and pathways of the adaptive immune system highlighting checkpoint key players such as CTLA4, CD274 (PD-L1), and intracellular signaling by second messengers (e.g., SNAI2, FGF10 among other) that may be altered, providing a first tentative insight into the molecular mechanisms in MUC16-expressing tumors (Figure 4, Supplementary Table S2). Next, we performed verification of PD-L1 as a clinically relevant target by IHC in the corresponding patient tissue. In line with our expression data, IHC assessment revealed 5.9% and 75% PD-L1-positive CCA tissue in MUC16 (−) or MUC16 (+) patients, respectively. Thus, positive PD-L1 status differed significantly between MUC16 (−)/(+) tissue (p < 0.001). Ingenuity pathway analysis (IPA) was further used to merge networks that were associated with cancer growth and survival to predict the activation and inhibition of additional biomarkers (Supplementary Figure S2). Besides the experimentally identified biomarkers Stat3, MYC, CXCL5, CCND1, and RELA among others are predicted to be activated. The given candidates are involved in cancer cell proliferation or resistance to chemotherapy. Additional significant expression changes are depicted in Supplementary Table S3.
The overexpression of MUC16 has been linked to worse prognosis in several malignancies [16,24,25]. Immunohistochemical expression of MUC16 in CCA has been studied to a limited extend, while novel prognostically relevant markers for this devastating tumor entity are urgently needed. Hence, the aim of the present study was to determine the impact of MUC16 expression on OS of patients with CCA after surgical resection. To our knowledge, this study is the first to investigate this clinically relevant issue in all CCA subtypes by immunostaining and differential expression analysis in order to preliminarily elucidate the pathomechanism behind its expression. In the present study, 66% of the patients were MUC16 (−), while 34% expressed MUC16. We thereby observed a lower proportion of MUC16 (+) patients compared to data from Higashi et al., although these data solely refer to mass-forming iCCA [16]. Besides the prevalence of MUC16 expression, we also analyzed its impact on clinical outcome. We found a significantly impaired overall survival for MUC (+) patients (27.4 months) in comparison to patients without MUC16 expression (56.1 months). Several aspects need to be discussed as potential cofactors for the negative impact of MUC16 (+) on survival, i.e., MUC16 (+) patients had higher Pn status, more frequently positive serum levels of tumor marker CA-19/9, and increased Ki-67 proliferative index. Especially, R-status, metastatic spread, recurrence, and common risk factors for CCA were comparable between both subgroups. In multivariate analysis, MUC16 (+) remained an independent risk factor for impaired survival as well as the known prognostic factor ECOG 1, multiple tumors, and the presence of tumor marker CA-19/9. In general, MUCs and their impact on survival are extensively studied in biliary tract cancer (BTC). Among these, MUC4 as well as MUC5AC are highly tumor-associated in BTC [26]. In the literature, the immunohistochemical detection of MUC1 and MUC4 serve as predictors of poor outcome in CCA [23,27] whereas we demonstrated a significant impact on survival only for MUC4 but not for MUC1 or MUC2. However, the impact of MUC1 on survival in the study from Park et al. was based solely on univariate analysis [23]. In addition, cellular and serum MUC5AC were thought to be related to advanced tumors and poor prognosis, respectively [23,28]. In contrast, a recently published study reported that the low expression of MUC5AC and MUC6 predicts poor prognosis solely for pCCA patients [29] whereas our study revealed a significant impact on OS for high MUC5AC expression. Additionally, the immunohistochemical expression of MUC6 is thought to be associated with well-differentiated CCA but not with poor survival, in line with our data [23]. However, data on the impact of MUC16 expression on the outcome of CCA patients are very limited as two studies only investigated the role of MUC16 in mass-forming iCCA or its co-expression with mesothelin in extrahepatic CCA, preventing comparability between both cohorts [16,17]. For the first time, we provide conclusive data on MUC16 expression in all three CCA subtypes from one cohort of our tertiary hospital, strengthening the role of MUC16 as a prognostically relevant marker. However, it still remains elusive whether MUC16 detection in tumor tissue and serum levels of CCA patients correlate or have similar prognostic value. In CCA patients with available CA125 serum levels, 100% (4/4) of MUC16 (−) patients were also negative for serum CA125 while 75% (3/4) of the MUC16 (+) patients were also positive for serum CA125 in our study, which may serve as the first indication of a possible correlation between MUC16 assessment by IHC and serum levels. However, these data have to be analyzed carefully as the patient cohort with available CA125 serum levels is very small in our study. As pre-operative serum levels of CA125 were recently found to predict poor prognosis in pCCA receiving surgery [14], we hypothesize that CA125 may play a crucial role in the future, especially in CCA patients that are lacking the well-known tumor marker CA-19/9. Therefore, investigations of CA125 serum levels in clinical routine diagnostics are warranted as an additional and readily accessible biological resource. In pancreatic cancer, combined chemo-immunotherapy with anti-MUC16 led to the development of specific T-cell immunity [19] while targeting MUC16 revealed prolonged progression-free and OS in a Phase II study in advanced ovarian cancer [30]. Next, the overcoming chemoresistance by targeting MUC16 may be beneficial in lung cancer, since MUC16 overexpression influences gemcitabine and cisplatin resistance in this entity [31]. Future studies should determine a specific therapeutic response for MUC16 (+) patients to enable the development of chemopreventive strategies in CCA. Interestingly, a correlation between therapeutical response and the quantity of MUC16 in ovarian cancer patients could be observed in two clinical trials [19,30]. In line with this, Wang et al. could show that patients with epithelial ovarian cancer and high MUC16 tissue expression had a significantly poorer prognosis compared to low MUC16-expressing patients [32]. Next, Gubbels et al. suggested that O-glycolisation on cell surface-bound MUC16 contributes to immune evasion in ovarian cancer cells. Interestingly, this observation depends on the MUC16 quantity since the expression of low levels of MUC16 correlated with an increased number of conjugates between malignant cells and natural killer cells [33]. Hence, the quantitative expression of MUC16 needs to be assessed and defined in further studies, as it could influence clinical outcome in CCA as well. By MUC16 discrimination of iCCA cases, cell cycle and immune response processes are overrepresented among deregulated candidate transcripts. Enhanced cancer cell proliferation highly intersects with chemotherapy resistance processes. MUC16 (+) patients had higher Pn status and increased Ki-67 proliferative index underlining the functional annotation of our expression data. By predicting the activation of STAT3, we corroborate the frequently discussed association of MUC16 and JAK-STAT signaling [6,13,34]. This link might be orchestrated by Type I interferons (IFN) and mediate further transcription factors such as IFN regulatory factors (IRF) that also emerged in our expression analysis [35]. In line with this, PD-L1 expression increased in several cancer entities by IFN via JAK-STAT signaling [36,37]. Interestingly, recently published data reported increased survival rates for palliative CCA receiving standard of care chemotherapy if a PD-L1 inhibitor is added [38] while we confirmed clinically relevant CD274 (PD-L1) deregulation in MUC16 (+) patients by companion IHC diagnostics of PD-L1. Since data about the interplay between MUC16 and JAK-STAT signaling in CCA are missing so far, our preliminary gene expression profiling should, therefore, encourage further research investigating the underlying molecular mechanisms that are potentially altered in MUC16-expressing CCA. We acknowledge the following limitations of our study. As a retrospective single center study, the sample size is modest and may lead to case selection bias, especially for pCCA and dCCA. Nevertheless, the case number of iCCA in our cohort is remarkably larger compared to Higashi et al. [16] while the number of cases for pCCA and dCCA is similar to Takihata et al. [17] and is limited due to the relatively rare occurrence of CCA in general. Since we analyzed data from a large TMA cohort, it should be considered that TMA cores only represent limited areas of tumor tissue. As we provide preliminary insights into genes that are particularly deregulated between MUC16-positive and -negative patients, we must acknowledge that this is a very small case series that is limited to iCCA only. Since the MUC16 antibody is well established in routine clinical practice for ovarian cancer, its applicability in CCA diagnosis is ensured without additional effort.
In summary, this study strengthens the limited published data on MUC16 expression in CCA tissue and its impact on survival by using a large cohort encompassing all CCA subtypes. MUC16 (+) is associated with impaired OS after surgical resection in curative intention, serving as an independent risk factor for poor prognosis. These findings strengthen MUC16 expression as a prognostically relevant marker for patients that are suffering from CCA and suggest including it in routine immunohistochemical staining. | true | true | true |
|
PMC9563957 | Changchuan Ye,Yuting Yang,Xi Chen,Lijie Yang,Xia Hua,Mengjie Yang,Xiangfang Zeng,Shiyan Qiao | Metabolic engineering of Escherichia coli BW25113 for the production of 5-Aminolevulinic Acid based on CRISPR/Cas9 mediated gene knockout and metabolic pathway modification | 13-10-2022 | E. coli,5-Aminolevulinic Acid,CRISPR/cas9,T7 Expression System,Metabolic engineering | Background 5-Aminolevulinic acid (ALA) recently received much attention due to its potential application in many fields. In this study, an ALA production strain of Escherichia coli was constructed by rational metabolic engineering and stepwise improvement based on known regulatory and metabolic information and CRISPR/Cas9 mediated gene knockout. Results A metabolic strategy to produce ALA directly from glucose in this recombinant E. coli via the C5 pathway was applied herein. The rational metabolic engineering by gene knockouts significantly improved ALA production from 662.3 to 1601.7 mg/L. In addition, we managed to synergistically produce ALA via the C4 pathway in recombinant strain. The expression of a modified hemA gene, encoding an ALA synthase from Rhodobacter sphaeroides, improved ALA production from 1601.7 to 2099.7 mg/L. After 24 h cultivation, a yield of 0.210 g ALA per g glucose was achieved by constructed E. coli D5:FYABD-RSA. Conclusion Our study revealed that an industrially competitive strain can be efficiently developed by metabolic engineering based on combined rational modification and optimization of gene expression. Supplementary Information The online version contains supplementary material available at 10.1186/s13036-022-00307-7. | Metabolic engineering of Escherichia coli BW25113 for the production of 5-Aminolevulinic Acid based on CRISPR/Cas9 mediated gene knockout and metabolic pathway modification
5-Aminolevulinic acid (ALA) recently received much attention due to its potential application in many fields. In this study, an ALA production strain of Escherichia coli was constructed by rational metabolic engineering and stepwise improvement based on known regulatory and metabolic information and CRISPR/Cas9 mediated gene knockout.
A metabolic strategy to produce ALA directly from glucose in this recombinant E. coli via the C5 pathway was applied herein. The rational metabolic engineering by gene knockouts significantly improved ALA production from 662.3 to 1601.7 mg/L. In addition, we managed to synergistically produce ALA via the C4 pathway in recombinant strain. The expression of a modified hemA gene, encoding an ALA synthase from Rhodobacter sphaeroides, improved ALA production from 1601.7 to 2099.7 mg/L. After 24 h cultivation, a yield of 0.210 g ALA per g glucose was achieved by constructed E. coli D5:FYABD-RSA.
Our study revealed that an industrially competitive strain can be efficiently developed by metabolic engineering based on combined rational modification and optimization of gene expression.
The online version contains supplementary material available at 10.1186/s13036-022-00307-7.
Synthetic biology plays a critical part in bio-based production of fuels, chemicals and materials from biomass. The sophistication of available genetic and biochemical tools has developed as synthetic biology applications have grown in complexity. In the past 20 years, a dramatic change in the scope and complexity of efforts within the space of synthetic biology has occurred [1]. Due to the great development of genetic engineering, accomplishments have progressed from simple reconstitution of biosynthetic pathways in heterologous hosts to complicated refactoring efforts [2–4] that can produce medicinally relevant compounds (strictosamide, taxadiene, etc.) [5, 6] or amino acids (L-valine, L-threonine, etc.) [7, 8] in a high titer. Among many systems available for heterologous protein production, the Gram-negative bacterium Escherichia coli remains one of the most attractive organisms because of its ability to grow rapidly and at high density on inexpensive substrates. Well-characterized genetics and availability of a large number of cloning vectors and mutant host strains [9]. As a commonly used organism for expression and optimization of heterologous biosynthetic pathways, E. coli has been the most widely used prokaryotic system that produces heterologous proteins for industrial production of bacterial metabolites by batch and fed-batch operations [10, 11]. E. coli is a facultative anaerobic, non-sporulating bacterium that is considered as the workhorse of modern biotechnology in the microbial production of bio-fuels and biochemicals [12]. Recent advances in the fundamental understanding of transcription, translation and protein folding in E. coli, together with developments of improved genetic tools have made E. coli more valuable than ever [9]. 5-Aminolevulinic acid (ALA) is a key intermediate involved in the biosynthesis of tetrapyrrole, which has attracted much attention for its potential applications in agriculture, cosmetics and cancer therapy. Due to its nontoxicity to crops, animals and humans, ALA is also used as selective biodegradable herbicide and insecticide in agriculture [13]. In living organisms, there are two major pathways described for ALA biosynthesis [13]. One is the C4 pathway, which occurs in mammals, birds, yeast and purple non-sulfur photosynthetic bacteria. In this pathway, ALA is formed through catalyzation of 5-aminolevulinate synthase, which condenses glycine and succinyl-CoA, an intermediate of the tricarboxylic acid (TCA) cycle [14]. The second pathway, C5, is present in higher plants, algae and many bacteria including E. coli [15]. In the C5 pathway, glutamate is the only substrate for biosynthesis of ALA. In E. coli, ALA biosynthesis is through the C5 pathway and is tightly regulated by feedback inhibition of heme, the end product of the C5 pathway [16]. Earlier study has developed an available metabolic strategy of ALA biosynthesis via C5 pathway in E. coli [17]. Through this strategy, ALA can be synthesized in E. coli by over-expressing the key genes, hemA and hemL, in a modified minimal medium using glucose as the sole carbon source. In the research reported herein, we used a similar process for ALA biosynthesis. We managed to integrate T7 RNA Polymerase into the genome of E. coli BW25113 in our earlier work. This mutant strain was named E. coli BW25113-T7 [18]. In this mutant strain, the key genes, hemA, hemL and RhtA, were expressed efficiently and controlled by the T7 Expression System. Our earlier work proved that this process for ALA biosynthesis in E. coli BW25113-T7 works. Through efficient and controllable protein expression using the T7-Lac promoter, E. coli BW25113-T7 accumulated 42.4% more ALA than E. coli BL21(DE3) did [18]. Here, we report the development of the genetically defined ALA overproducing E. coli strain by systems metabolic engineering. It was thus our goal to engineer E. coli to make a 5-Aminolevulinic acid overproducer with completely known genotype by systems-level metabolic engineering. In this study, rational metabolic engineering based on known regulatory and metabolic information was carried out to develop an E. coli strain capable of overproducing ALA.
We constructed an E. coli strain from BW25113 which achieved high efficiency of protein expression through T7 expression system in our earlier work [18]. This mutant E. coli BW25113-T7 strain was rationally engineered to produce ALA by over-expression of key genes. The biosynthetic pathways of ALA in E. coli and the strategies for constructing ALA production strain are shown in Fig. 1. The key genes of ALA synthesis via the C5 pathway in this strategy are hemA and hemL. The hemA gene encodes a NADPH-dependent glutamyl-tRNA reductase which catalyzes the reduction of glutamyl-tRNA to glutamate-1-semialdehyde (GSA) [17]. The hemL gene encodes glutamate-1-semialdehyde aminotransferase which quickly converts GSA to ALA. The eamA gene encodes an O-acetylserine or cysteine exporter which is capable of translocating dipeptides and amino acid analogs from the cytosol to the periplasm. Plasmid pET-ALA-LAA was constructed to over-express these three key genes. In this plasmid, these three genes can be tightly controlled and efficiently expressed by the T7-Lac promoter. Cells that harbored pET-ALA-LAA were cultured in M9YE medium. After 24 h of induced fermentation, an ALA titer (662.3 mg/L) was determined.
Glutamate and L-glutamyl-tRNA are both precursors of ALA synthesis via the C5 pathway. To increase glutamate and L-glutamyl-tRNA amounts for ALA synthesis, we overexpressed gltB, gltD and gltX genes alone or in combination. There are two pathways by which E. coli synthesizes glutamate from ammonia [19]. The pathway catalyzed by glutamate dehydrogenase (encoded by gdhA) is ATP-independent. The other is a cyclic and ATP-driven pathway (Shown in Fig. 2) which catalyzed by glutamate synthase (encoded by gltBD). Our modified minimal medium does not provide a high level of ammonia. In this case, the ATP-driven pathway functions (Fig. 1). Glutamate synthase is a tetramer of dimers, with each dimer having one large and one small subunit (gltB and gltD, respectively) [19]. Glutamate synthase catalyzes the single-step conversion of L-glutamine and alpha-ketoglutarate into two molecules of L-glutamate (Fig. 2). In doing so, glutamate synthase simultaneously operates as the major source of L-glutamate for the cell and as a key step in ammonia assimilation during nitrogen-limited growth [19–21]. The ammonia-dependent activity can be catalyzed very slowly by just the small subunit in the absence of the full complex [22]. Glutamate-tRNA ligase (GluRS) is a member of the family of aminoacyl-tRNA synthetases, which is encoded by gltX gene [23, 24]. GluRS charges GlutRNA for both protein and ALA synthesis [25]. In this study, we constructed four extra plasmids to overexpress these three genes (Table 1). These four plasmids were electroporated into cells that harbored pET-ALA-LAA. The resulting strains were cultivated in modified minimal medium (see ‘‘Section Materials and Methods’’) supplied with 10 g/L glucose for ALA accumulation analysis. After 24 h of fermentation, ALA concentration was determined (Table 2 and Fig. 3A). Strain BW25113-T7 which only contained pET-ALA-LAA was set as the control. Beyond our expectation, we found all strains that harbored extra plasmids exhibited decreased ALA accumulation with E. coli BDX having the greatest decrease in 5-ALA accumulation relative to the control (Table 2). We found that all strains harbored extra plasmids grew much more slowly than control strains (Table S1). These results indicate that overexpression of gltB, gltD and gltX genes may not have a positive effect on the ALA biosynthesis in E. coli. This finding suggests that the biosynthesis of L-glutamate and conversion of L-glutamate to GlutRNA are not rate-limiting steps of ALA biosynthesis.
To further improve the BW25113-T7 strain, following targeted genetic modifications were performed (Fig. 1). HemF is a strictly aerobic enzyme that requires molecular oxygen as the electron acceptor and produces hydrogen peroxide. The knock-out of hemF gene would mightily repress the biosynthesis of protoporphyrinogen IX (PPG) resulting in reduced endogenous loss of 5-ALA. After knocking out the hemF gene in BW25113-T7 (named D1:F), yield of 5-ALA increased 2.06-fold compared to the original strain (named as D0). This result indicated that knock-out of the hemF gene has a positive effect on ALA biosynthesis. L-glutamate is a precursor of ALA synthesis via the C5 pathway. Therefore, we attempted to increase ALA accumulation by reducing endogenous loss of L-glutamate. GdhA, which encodes glutamate dehydrogenase, catalyzes the ATP-independent amination of α-ketoglutarate to yield L-glutamate [26]. This reaction is reversible (Fig. 4A). YbdK catalyzes ATP-dependent ligation of glutamate with cysteine at a low catalytic rate [27] (Fig. 4B). GadB, a glutamate decarboxylase enzyme, catalyzes cleavage of L-glutamate into carbon dioxide and 4-aminobutanoate [28] (Fig. 4C). These three genes in the ALA biosynthetic pathway were knocked out (Fig. 1) individually or jointly using the E. coli D1:F strain. ALA accumulations of those mutant strains ranged from a reduction to a 2.30-fold increase in ALA accumulation (Table 3). Among the mutant strains, E. coli D3:FYB significantly increased ALA accumulation (2.30-fold) compared with E. coli D0. The periplasmic binding proteins—mppA, the L-alanyl-g-D-glutamyl-meso-diaminopimelate binding protein, or dppA, the dipeptide binding protein—actively import ALA through an interaction with the dipeptide inner membrane ATP-binding cassette transporter, DppBCDF, in E. coli [29]. We theorized that inactivation of dppA and/or mppA genes would reduce ALA assimilation and therefore increase ALA accumulation in the medium. As expected, inactivation of dppA improved ALA production but mppA did not (Table 3). To further improve ALA production, we used the E. coli D4:FYAB strain to perform dppA and mppA gene knockouts individually or jointly. We found that mutant strain, E. coli D5:FYABD, showed the highest production of ALA which was 2.42 fold greater than D0 (Table 3).
After knocking out these genes, growth of mutant strains in different media were examined to assess whether CRISPR/Cas9-mediated gene knock-out affected metabolic characteristics of the bacteria. These strains all harbored pET-ALA-LAA for producing ALA. These mutant strains were cultured in M9YE medium (see ‘‘Section Materials and Methods’’) or standard LB medium (Fig. 5). Growth rate among these strains in LB medium ranged from 0.25 to 0.3 (Table 4). When cultured in M9YE medium, growth rate of these strains varied from 0.3 to 0.5. Two mutant strains (D2:FA and D3:FYB) showed significant change in growth rate when culture in M9YE medium. These results reveal that knockouts of some genes may lead to growth retardation of bacteria. We easily selected D5:FYABD as the optimal mutant strain for ALA production based on its growth rate and ALA production (Fig. 6). The ALA production was increased 2.42-fold with similar growth rate when compared with the original-type (D0).
In E. coli, primary ALA biosynthesis is through the C5 pathway. To further increase accumulation of ALA, we tried to synergistically produce ALA via the C4 pathway. The photosynthetic bacterium, Rhodobacter sphaeroides, can accumulate ALA under certain conditions or after mutagenesis [30–32]. Through metabolic engineering, recombinant E. coli was also able to produce ALA from the C4 pathway through biotransformation [33]. In this aspect, the gene encoding for ALA synthase from R. sphaeroides was introduced into E. coli through genetic engineering [33]. Biosynthesis of glycine and succinyl-CoA are regulated in E. coli. Consequently, glycine and succinate (the precursor of succinyl-CoA) must be added to the culture medium artificially to provide sufficient substrates for enhanced ALA biosynthesis. In the current study, we overexpressed a modified heterologous hemA from R. sphaeroides using the T7 Expression System (Fig. 7). The amino acid sequence of hemARS was modified for better expression in E. coli. Glycine (2 g/L) was added to the M9YE medium as a substrate for the C4 pathway. These modifications increased ALA accumulation from 1602.72 mg/L with E. coli D5:FYABD to 2099.7 mg/L for the mutant strain containing the hemARS gene (Table 2 and Fig. 3B). After 24 h cultivation, a yield of 0.210 g ALA per g glucose was achieved (Fig. 8).
Metabolic engineering provides a powerful tool for regulating metabolic pathway towards accumulation of desired compounds [34]. Many bio-based products are now produced in E. coli through metabolic engineering or synthetic biology [17, 35–37]. In the present study, we developed a metabolic strategy to produce ALA directly from glucose and constructed a well-defined genetically-engineered E. coli strain based on known metabolic and regulatory information and CRISPR-Cas9 mediated gene knock-outs. After systematic metabolic engineering of E. coli, we were able to achieve a high yield of 0.210 g of ALA per gram of glucose. More importantly, the engineered strain developed in this study can be further improved because all of the modifications are clearly defined. We first constructed the ALA-producing E. coli base strain by amplifying three key genes. HemA encodes Glutamyl-tRNA reductase, which catalyzes the first step of porphyrin biosynthesis [38]. Glutamate-1-semialdehyde 2,1-aminomutase (hemL) catalyzes the pyridoxal 5'-phosphate-dependent transfer of the amino group from C2 of glutamate-1-semialdehyde (GSA) to C1, thereby forming ALA [39]. Synergistic functions of hemA and hemL efficiently converted glutamate to ALA. The eamA gene encoded an integral membrane protein which is implicated in exporting metabolites of the cysteine pathway [40]. Deletion of eamA in E. coli strain BL21 is associated with a decline in extracellular 5-aminolevulinate concentration [41] which suggests that eamA might be a transporter involved in export cysteine pathway metabolites. We overexpressed hemA, hemL and eamA genes using T7 system in E. coli BW25113-T7 and achieved a yield of 0.066 g of ALA per gram of glucose. Then several genes encoding those enzymes directly involved in Glutamate and L-glutamyl-tRNA biosynthesis were amplified. However, this modification did not increase accumulation of ALA. This result was similar to previous research by Kang [17]. In their study, amplifying gltX genes decreased ALA accumulation. Our result also suggested biosynthesis of L-glutamate and conversion of L-glutamate to GlutRNA are not rate-limiting steps in ALA biosynthesis. Overexpression of gltX, gltB and gltD genes may increase metabolic stress on cells which may lead to growth retardation and decreased ALA accumulation. Further improvement of the ALA-producing strain was achieved to modify metabolic pathway rationally by using the metabolic and regulatory information available in the literature (Fig. 1). In E. coli, biosynthesis of ALA via the C5 pathway is tightly regulated by the end product heme [16]. Synthesis of heme leads to an endogenous depletion of ALA. Additionally, heme also makes hemA unstable which inhibits accumulation of ALA [42]. Therefore, we expected that down-regulation of heme biosynthesis might lead to high ALA production. However, knock-outs of most genes encoding enzymes directly involved in heme biosynthesis (hemBCDEGH) lead to severe growth limitation or even death. Oxidative decarboxylation of coproporphyrinogen III to protoporphyrinogen IX is catalyzed by two enzymes in E. coli. Under aerobic conditions, the hemF gene product is active. Under anaerobic conditions, HemN catalyzes the coproporphyrinogen III dehydrogenase reaction. Hence, we believed that knock-out of hemF will ultimately restrict biosynthesis of heme. This belief was confirmed with the observation that the mutant strain E. coli D1:F increased ALA yield from 662.25 to 1361.22 mg/L. Our next step was to knock out the genes responsible for major competing pathways based on published metabolic and regulatory information (Fig. 1). The major competing pathways, including nitrogen metabolism, glutathione metabolism and butanoate metabolism, were suppressed in the present study. In E. coli, there are two pathways to synthesize glutamate from ammonia [19]. The pathway catalyzed by glutamate dehydrogenase (encoded by gdhA) directly from ammonia, α-ketoglutarate, and NADPH [26]. The other is a cyclic and ATP-driven pathway (Shown in Fig. 2) and is catalyzed by glutamate synthase (encoded by gltBD). Each turn of which utilizes one molecule each of ammonia, ATP, NADPH, and α-ketoglutarate to produce one molecule of glutamate [19]. When culture in our modified minimal medium (M9YE) which does not provide a high level of ammonia, the ATP-driven pathway functions (Fig. 1). As the reaction catalyzed by gdhA is reversible, we assumed that down-regulation of gdhA catalyzed reversible reaction is believed to reduce endogenous loss of L-glutamate and led to a raise of ALA production. The ybdK (encoding putative glutamate—cysteine ligase 2) and gadB (encoding glutamate decarboxylase B) genes were also knocked out to increase L-glutamate availability for ALA synthesis. In our study, all triple mutant strains (D3:FYA, D3:FYB and D3:FAB) yielded more ALA than double mutant strains (D2:FY, D2:FA and D2:FB). We propose that synergistical effects of knocking out these three genes (ybdK, gadB and gdhA) improves ALA yields because of the increased flux through L-glutamate. These observations led us to select mutant strain D4:FYAB for further modification even though this strain did not produce the highest yield of ALA. To avoid ALA accumulation inside cells and to produce more ALA in the medium, inactivation of mppA and/or dppA genes was carried out. Cultivation results showed that inactivation of dppA (D1:D) improved ALA production while inactivation of mppA (D1:M) did not. The D1:D strain harboring pET28b-ALA-LAA accumulated 27.9% more ALA than the D0 strain did (Table 3). Further knockout of dppA and/or mppA genes in strain D4:FYAB showed a similar result. The additional knockout of the mppA gene in D4:FYAB did not increase ALA production. Double mutant of dppA and mppA in D4:FYAB strain even showed a 29.2% decrease in ALA production. These results suggest that silencing the dppA gene is more beneficial for ALA biosynthesis than knocking out the mppA gene. According to metabolic and regulatory information available in the literature, we identified a mutant strain that allowed elevated ALA production with an acceptable growth rate. In the D5:FYABD strain, heme biosynthesis was limited and flux through glutamate was increased. Deletion of the dppA gene reduced ALA assimilation by E. coli which allowed increased ALA accumulation in the medium. The D5:FYABD strain harboring pET-ALA-LAA showed a 141.8% increase in ALA production compared with the D0 strain harboring the same plasmids. These results suggest that beneficial effects on ALA production of flux redistribution achieved in D5:FYABD strain could be further strengthened by engineering regulatory and export functions. There are two major pathways described for ALA biosynthesis in living organisms [13]. We successfully measured production of ALA via the C5 pathway in this study and tried to develop a strategy to improve ALA production in recombinant E. coli via the C4 pathway. The ALA synthase from R. sphaeroides was introduced into E. coli to produce ALA from the C4 pathway. This modified hemARS was able to work synergistically with native hemA and hemL from E. coli (Fig. 7). In R. sphaeroides, glycine and succinyl-CoA can be catalyzed to form ALA by hemA. Since biosynthesis of glycine and succinyl-CoA was also regulated in E. coli, we added glycine in the medium artificially to provide more substrates for ALA biosynthesis. Through this strategy, ALA production in D5:FYABD was increased from 1601.7 mg/L to 2099.7 mg/L, a 31.1% increase in ALA accumulation.
In summary, rational metabolic engineering based on known metabolic and regulatory information, and co-expression of native key genes and modified heterologous gene, allowed development of an E. coli strain capable of efficiently producing ALA. An impressively high yield of 0.210 g of ALA per gram of glucose could be achieved using a metabolically engineered strain in flask culture. Further optimization of the fermentation process could improve ALA production to even higher levels. The approaches described in this study can also be applied more broadly to develop strains for efficient production of other metabolites.
Molecular cloning and manipulation of plasmids were done with E. coli DH5α (TransGen, Beijing). BW25113-T7 strains were used for CRISPR/Cas9-induced Double Strain Break and recombination. All E. coli strains were cultured routinely in standard LB medium when not mentioned otherwise.
LB medium (10 g/L tryptone, 5 g/L yeast extract and 10 g/L NaCl, pH 7.2) was used in all DNA manipulations. During cultivation and fermentation, the modified M9 medium (M9YE) was used that contained 1 g/L NH4Cl, 0.5 g/L NaCl, 3 g/L KH2PO4, 17.1 g/L Na2HPO4·12H2O, 2 mM MgSO4, 0.1 mM CaCl2, 2 g/L yeast extract and 10 g/L glucose. Glycine (2 g/L) was added as indicated to serve as the substrate for the C4 pathway. Ampicillin (100 mg/mL), chloramphenicol (25 mg/mL) and kanamycin (50 mg/mL) were added to provide selective pressure for E. coli during cultivation when necessary. To induce expression of plasmid-borne genes, Isopropyl-β-D-thiogalactopyranoside (IPTG) was added to cultures which resulted in a final concentration of 0.1 mM. Considering cell growth and ALA stability, the pH was measured by a glass electrode and controlled at 6.5 ± 0.3 with 4 M NaOH.
The function and detailed message of the hemF gene (and other genes) was verified in NCBI and BioCyc Database. Sequence of the hemF gene (and other genes) in the BW25113-T7 genome was confirmed in NCBI. As the recognition site for sgRNA, N20 site directs the Cas9 protein to enable site-specific induction of a DSB. The N20 site was found by BROAD international design tool, which is available at: http://www.broadinstitute.org/rnai/public/analysis-tools/sgrna-design.
All primer pairs we designed for gene cloning and intermediate plasmid construction are listed in Table S2. Plasmid pCas (Fig. S1A) and pTarget were prepared in our laboratory. Plasmid pTarget-gene was constructed by Reverse-PCR and T4-Ligation (T4 DNA Ligase, NEB, England) to replace the N20 fragment (Fig. S1B) with specific primers (Table S2). Plasmid pACYCD-gene and pACYCD-Donor-Gene were both constructed for preparing Donor DNA (Fig. S2) by In-Fusion® HD Cloning Kit (Takara, Japan). Donor DNA was cloned from pACYCD-Donor-Gene by Hi-Fi PCR (Phusion® High-Fidelity PCR Master Mix, NEB, England). All plasmids used in this research are listed in Table 1.
For transformation, the plasmid or linear DNA were electroporated into competent cells in the pre-chilled cuvette (0.1 cm) using Bio-Rad MicroPulser (1.8 kV, time constant > 5.0 ms). For selection, 25 μg/mL chloramphenicol (Chl) or 50 μg/mL kanamycin (Kan) were used alone or in combination. For induction of λ-Red proteins and lac operator, 1 mM arabinose and 1 mM IPTG were used. To prepare cells harboring pCas, cells cultured at 37℃ (OD600 = 0.45–0.55) were made competent, mixed with pCas (100 ng) and subjected to electroporation, after which the cells were recovered in SOC medium (1 mL) for 1 h at 30℃, plated onto the Kan plate, and cultured at 30 ℃ for 18–24 h. For CRISPR/Cas9-mediated homologous recombination, cells harboring pCas were cultured at 30℃ in medium containing Kan and Arabinose and made competent. After co-electroporation of Donor DNA (400 ng) with pTarget-Gene (100 ng), cells were recovered in SOC (1 mL) medium for 1 h at 30℃, plated onto Chl/Kan plate, and cultured at 30℃ for 18–24 h. For elimination of pTarget-Gene, cells harboring both pCas and pTarget were cultured at 30℃ in medium containing Kan and IPTG for 2 h. Cells were plated onto Kan plates and cultured at 30℃ for 18–24 h. For elimination of pCas, cells harboring pCas were cultured at 37℃ in the medium without any antibiotic for 12–16 h. Then the cells were plated onto non-antibiotic plates and cultured at 37℃ for 12–16 h.
The targeted genetic modifications were rationally engineered by CRISPR/Cas9 in E. coli BW25113-T7. The six targeted genes (hemF, ybdK, gadB, gdhA, dppA and mppA) were cloned from E. coli BW25113-T7 to prepare for Donor DNA. The process of hemF gene knock-out is shown here as example. The map of hemF gene knock-out in BW25113-T7 in ideal condition is shown in Fig. S3. The homologous left arm (HRL) and the homologous right arm (HRR) were set near site of N20. Both homologous arms were approximately 400 bp, which could have a high efficiency of recombination [43]. After CRISPR/Cas 9 mediated gene knock-out, it would make a deletion of 100 bp DNA in targeted gene. To make CRISPR/Cas9-mediated homologous recombination in BW25113-T7, we electroporated pCas (encoding both Cas and λ-Red proteins) into E. coli BW25113-T7, followed by Arabinose (Ara) induction of pCas-encoded λ-Red proteins Gam, Bet and Exo. After preparing competent cells, pTarget-gene (such as pTarget-hemF) and Donor DNA were co-electroporated into cells (Fig. S4). To verify deletion in target locus, all mutated bacterial strains were selected for colony PCR (Fig. S5). The primer pairs for colony PCR are shown in Table S2. Characteristics of those mutant strains are shown in Table 1.
All E. coli strains were grown at 37℃ in conical flasks (250 mL) containing M9YE medium or LB medium with Kanamycin (50 μg/mL). Growth was measured by monitoring optical density at 600 nm (OD600) using a spectrophotometer. The growth rate was fitted with Sigmoidal-4PL curves. The OD of the stationary phase (ODsp) and the time required to reach it are calculated from this fitted curve. The ratio of the ODsp to the time required to reach it was set as growth rate (hr-1) of the strain in medium.
Flask cultivations were carried out in 100 mL conical flasks supplied with 30 mL modified minimal medium at 37℃ with agitation of 200 rpm. A 1% (v/v) inoculum from an overnight culture (12) was used. Cells were cultured in M9YE at 37℃ until OD600 reached 0.7. Then IPTG with a final concentration of 0.1 mM was added to the induced group. To analyze ALA production, culture (30 mL) after inducing for 24 h was centrifuged (12,000 × g for 2 min at 4℃). The supernatant was analyzed for extracellular ALA concentration. ALA concentration was analyzed using modified Ehrlich’s reagent [44]. Specifically, standard or sample (2 ml after diluted) was mixed with 1 ml 1.0 M sodium acetate (pH 4.6) in a cuvette, and 0.5 ml acetylacetone (2,4-pentanedione) was added to each cuvette. Then the mixtures were heated to 100 ℃ for 15 min. After cooling for 15 min, the reaction mixture (1 ml) and freshly prepared modified Ehrlich's reagent (1 ml) were mixed together. After 30 min, the absorbance at 554 nm was measured. Standard plot for ALA measurement is shown in Fig. S6. For analyzing glucose, 1 mL of culture was centrifuged (12,000 g for 2 min at 4 ℃) and the supernatant was then filtered through a 0.22 mm syringe filter for analysis. The HPLC system was equipped with a cation exchange column (HPX-87 H, BioRad Labs), and a differential refractive index (RI) detector (Shimadzu RID-10 A). A 0.5 mL/min mobile phase using 5 mM H2SO4 solution was applied to the column. The column was operated at 65 ℃.
Data for ALA production of were subjected to analysis of variance (ANOVA) by GraphPad Prism (version 7.00). Error bars indicate standard error of the mean (SEM). P values were calculated using Dunnett's multiple comparisons test (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). The mean of each column was compared with the mean of a control column. BW25113-T7 (D0) was set as control column.
Additional file 1: Fig S1. Map of plasmids which were constructed for CRISPR. A) Map of pCas, which harbored the temperature sensitive oriR101 with repA101ts, kanamycin resistance gene, the λ-Red operon encoding Gam, Bet, and Exo proteins under the control of arabinose-inducible promoter ParaB, S. pyogenes-derived cas9 driven by endogenous promoters and sgRNA guided to ori-p15a which is under the control of lac operator. B) Map of pTarget-gene, which harbored Chloramphenicol resistance, ori-p15a and sgRNA guided to E. coli BW25113-T7 targeted gene.Additional file 2: Fig S2. Construct of intermediate cloning vectors for preparing Donor DNA. (A) Fragment A cloned from BW25113-T7 was concatenated to pACYCD-Blank to assemble pACYCD-gene. (B) Reverse-PCR to constructed pACYCD-Donor. (C) The map of Donor DNA, which contains HRL and HRR.Additional file 3: Fig S3. Gene map for targeted Cas9-mediated gene Knock-In. (A) The knock-out site of hemF in BW25113-T7. (B) map of hemF knock-out in ideal condition and the location of PCR product (899 bp) for sequencing.Additional file 4: Fig S4. Schematic illustration of DSB induction and homologous recombination. After preparing competent cells, the pTarget-gene and Donor DNA which harbored homology arms (HRR and HRL) that targeted a chromosomal locus spanning the middle of targeted gene and the DSB site were electroporated into cells.Additional file 5: Fig S5. Detection of successful gene knock-outs of BW25113-T7.Additional file 6: Fig S6. Standard Plot for ALA measurement. x: the absorbance at 554 nm; y: ALA concentration of sample after diluted (mg/L).Additional file 7: Table S1. Growth rate of strain expressing various related genes in different medium.Additional file 8: Table S2. Primers for Plasmid constructions and Testing. | true | true | true |
|
PMC9564094 | 36197996 | Dhenugen Logeswaran,Yang Li,Khadiza Akhter,Joshua D. Podlevsky,Tamara L. Olson,Katherine Forsberg,Julian J.-L. Chen | Biogenesis of telomerase RNA from a protein-coding mRNA precursor | 05-10-2022 | basidiomycete,ncRNA,RNA processing,telomere | Significance Many noncoding RNA molecules are indispensable components of macromolecular complexes that govern critical cellular processes. Telomerase RNA is a long noncoding RNA essential for maintaining chromosome stability and cellular immortality in eukaryotes. In this study, we discover a basidiomycete fungal telomerase RNA that is processed from a messenger RNA (mRNA) transcript that encodes a conserved protein. Our finding demonstrates an unprecedented biogenesis pathway for generating a functional long noncoding RNA from a protein-coding mRNA precursor. | Biogenesis of telomerase RNA from a protein-coding mRNA precursor
Many noncoding RNA molecules are indispensable components of macromolecular complexes that govern critical cellular processes. Telomerase RNA is a long noncoding RNA essential for maintaining chromosome stability and cellular immortality in eukaryotes. In this study, we discover a basidiomycete fungal telomerase RNA that is processed from a messenger RNA (mRNA) transcript that encodes a conserved protein. Our finding demonstrates an unprecedented biogenesis pathway for generating a functional long noncoding RNA from a protein-coding mRNA precursor.
Many vital cellular processes that govern genetic information transfer from DNA to protein rely on a vast variety of RNA molecules. These RNAs include tens of thousands of messenger RNAs (mRNAs) that encode proteins as well as numerous noncoding RNAs (ncRNAs) that are not translated into proteins yet form crucial ribonucleoprotein complexes (1). The biogenesis of these diverse ncRNAs, such as ribosomal RNAs (rRNAs), transfer RNAs (tRNAs), small nuclear RNAs (snRNAs), and small nucleolar RNAs (snoRNAs), requires distinct RNA polymerases and diverse processing mechanisms. Among the diverse types of ncRNAs, long noncoding RNAs (lncRNAs) are generally defined as ncRNA molecules larger than 200 nucleotides and mainly regulate the expression of protein-coding genes at transcriptional, RNA processing, translational, and posttranslational levels (2). Telomerase RNA (TER) is a distinct class of lncRNA that functions as an integral component of the telomerase ribonucleoprotein enzyme. TER provides a short template that is reiteratively used by the telomerase reverse transcriptase (TERT) catalytic component to perform de novo synthesis of telomeric DNA repeats onto telomeres at chromosome ends. This telomere replenishment is crucial for preserving genomic stability and maintaining cellular reproductive capacity (3). While the TERT proteins show broad evolutionary conservation (4), TERs are extremely divergent in sequence, length, structure, and biosynthesis pathway (5). The computational search for TERs in distinct groups of eukaryotes is a daunting challenge due to the lack of sequence conservation among TERs. Although advanced bioinformatics strategies have made significant progress for specific lineages of eukaryotes (6–8), many TERs await to be uncovered in some eukaryotic clades that are evolutionarily distant. For these eukaryotes, biochemical purification of the telomerase holoenzyme from cell lysates remains most effective for TER identification (9, 10). Despite the poor conservation in primary sequence, TERs show secondary structure conservation in two domains that are required for telomerase activity (5). The 5′ proximal template-pseudoknot (T/PK) domain contains an RNA template that specifies the telomeric DNA repeat sequence to be synthesized, followed by a conserved pseudoknot structure that is indispensable for telomerase catalysis (11). A distal helical domain located downstream of the T/PK domain stimulates telomerase activity in the presence of T/PK and can independently bind the TERT protein (9, 12). This distal domain is known as conserved region 4/5 (CR4/5) in most animal taxa and filamentous fungi (6, 9, 13), eCR4/5 in other eukaryotic clades (14, 15), and helix IV in ciliates (16). The two-domain requirement for telomerase catalytic activity is universal and likely emerged during early evolution of eukaryotes (14). The biogenesis of TER has evolutionarily diversified by employing different RNA polymerases along distinct eukaryotic lineages. Flagellate (781-993 nucleotides [nt]), fungal (900 to 2,400 nt), and animal TERs (239 to 609 nt) are transcribed by RNA polymerase II (Pol II) (5) and predominantly contain a 5′ 2,2,7-trimethylguanosine (TMG) cap (17, 18). However, ciliate (140 to 210 nt) and plant (231 to 350 nt) TERs are smaller and transcribed by RNA polymerase III (Pol III) with a characteristic 5′ triphosphate and 3′ uridine-tail (19–22). The intriguing diversity of TER is further exemplified by the presence of lineage-specific accessory proteins that bind the 3′ region of TER to regulate 3′-end maturation. These protein repertoires mainly protect the TER 3′ end from exonucleolytic degradation. The 3′ proximal region of animal TERs contains a box H/ACA snoRNA-like domain bound by a heterotetrameric complex of dyskerin, NHP2, NOP10, and GAR1 proteins (6, 23, 24). Additionally in mammalian TERs, the terminal stem-loop of this domain contains a Cajal body (CAB) box bound by telomerase Cajal body protein 1 (TCAB1) for Cajal body localization (25, 26). In contrast, fungal TERs from budding and fission yeasts employ the snRNA biogenesis pathway for TER maturation by harboring conserved Sm binding sites at 3′ regions for binding the heptameric Sm ring complex (27, 28). The maturation of TER relies on a fine balance between RNA processing and decay. The 5′ end of fungal TER is typically protected by a TMG cap, and the 3′ end undergoes complex processing involving loading and unloading of protein complexes. In Saccharomyces cerevisiae TER (TLC1), the 3′-end maturation depends on the Nrd1-Nab3-Sen1 complex–mediated transcription termination and further stabilization by the Sm ring complex (29). Despite sharing the Sm binding site with S. cerevisiae, the 3′ end of Schizosaccharomces pombe TER (TER1) is generated by the first transesterification step of intron splicing without the subsequent religation of the exons (30). Filamentous fungal TERs undergo a similar 3′-end maturation by the spliceosomal cleavage of a terminal intron but use a distinct 5′-splice site sequence (31, 32). Here, we report a unique mRNA-derived lncRNA biogenesis pathway for the TER identified in the model fungus Ustilago maydis. The mature form of U. maydis TER (UmTER) possesses a 5′-monophosphate and is processed from an mRNA precursor that contains a 5′ 7-methylguanosine (m7G) cap, a long 3′ untranslated region (3′ UTR), and a 3′ poly(A) tail. Moreover, this mRNA precursor undergoes alternative splicing and contains an open reading frame (ORF) that encodes a conserved protein. The processing of mature UmTER from the 3′-UTR of a protein-coding mRNA precursor is an unprecedented mechanism for TER biogenesis.
Within the fungal kingdom, TER has been identified and extensively studied in Ascomycota but not in the sister phylum Basidiomycota (Fig. 1A). Past studies of telomerase using ascomycete fungal models such as S. cerevisiae (33), S. pombe (30), and Neurospora crassa (9) have led to many important findings in telomerase biogenesis mechanisms. To further explore TER biogenesis across fungal phyla, we set out to identify basidiomycete TER from the corn smut fungus U. maydis (Fig. 1A). To that end, we generated a recombinant U. maydis strain expressing a 3xFLAG-tagged U. maydis TERT (3xFLAG-UmTERT) protein for affinity purification of U. maydis telomerase (SI Appendix, Fig. S1 A and B). The recombinant U. maydis telomerase holoenzyme purified by anti-FLAG immunoprecipitation (IP) showed significant telomerase activity detected by the telomere repeat amplification protocol (TRAP) assay (SI Appendix, Fig. S1C). The RNA molecules copurified with the active telomerase holoenzyme were extracted and analyzed by Illumina next-generation sequencing, which generated over 109 million short reads. By employing a proven bioinformatics strategy (9, 34), we searched the U. maydis genome for TER candidates and identified 782 genomic loci that contain putative template sequences for synthesizing the telomeric DNA repeats (TTAGGG)n (Fig. 1B). Mapping the Illumina sequencing reads onto these loci identified TER candidates that were then ranked by read coverage (Fig. 1C). The top candidate sequence with the highest read coverage was the 18S rRNA, presumably due to its abundance in the cell (Fig. 1C). The second TER candidate was an unannotated RNA with a conserved homolog found in the closely related species Ustilago bromivora. Notably, the U. bromivora homolog also contained a conserved TER template sequence, 5′-UAACCCUAA-3′ (Fig. 1C). The remaining candidates lacked template-containing homologs in U. bromivora and were not pursued any further (Fig. 1C). Northern blot analysis verified the presence of the second TER candidate in U. maydis with a size of ∼1,300 nt (Fig. 1D), which is consistent with the length of the locus covered by the sequencing reads (Fig. 1E). We then mapped the 5′ end of this TER candidate by a cap-independent 5′ Rapid Amplification of complementary DNA (cDNA) Ends (5′-RACE) procedure (SI Appendix, Fig. S2A-C). For mapping the 3′ end by 3′-RACE, the RNA was first added with a guanosine/inosine (G/I) tail at the 3′ end, followed by an RT reaction using an oligo-dC reverse primer for cDNA synthesis and nested PCR for cDNA amplification (SI Appendix, Fig. S2 D–F). The 5′-RACE generated a major cDNA product, and the 5′ end of the TER candidate was determined by sequencing cloned cDNA products (SI Appendix, Fig. S2 B, lane 1 and SI Appendix, Fig. S2C). The 3′ ends of this TER candidate appeared to be slightly heterogeneous, while the majority of the 3′-RACE cDNA products revealed a 3′ end at position 1291 (SI Appendix, Fig. S2 E, lanes 1 and 2, and SI Appendix, Fig. S2F). More importantly, the RACE analyses performed on the RNA copurified with the 3xFLAG-UmTERT also indicated the same 5′ and 3′ ends (SI Appendix, Fig. S2 G–K). Thus, our 5′- and 3′-RACE results concluded that UmTER candidate #2 is a UmTERT-bound RNA with a size of 1,291 nt (Fig. 1E) which is consistent with the northern blot result (Fig. 1D). The homologs of the UmTER candidate #2 RNA were found by a bioinformatic search using the Basic Local Alignment Search Tool (BLAST) across three fungal taxonomy orders, namely, Ustilaginales, Urocystidales, and Violaceomycetales, and located between two protein-coding genes with a conserved gene synteny (SI Appendix, Fig. S3). To determine if UmTER candidate #2 is truly a telomerase component, we generated a UmTER gene knockout (ΔUmter) U. maydis strain with the 1,291-bp UmTER-encoding genomic region replaced by a hygromycin resistance gene cassette (SI Appendix, Fig. S4A) (35). Hygromycin-resistant ΔUmter clones were identified and verified by Northern blot analysis (SI Appendix, Fig. S4B). Two independent ΔUmter clones were analyzed by Terminal restriction fragment length (TRF) assay and showed progressive telomere shortening over 200 generations, while the wild-type (WT) strain was able to maintain telomere length during successive passages (SI Appendix, Fig. S4 C and D). Thus, UmTER candidate #2 is essential for telomere maintenance and indeed the authentic TER component of U. maydis telomerase.
The core telomerase enzyme consists of the catalytic TERT protein and the TER subunit that provides the template for DNA synthesis. As a prelude to performing detailed functional dissections of UmTER, we assembled in vitro transcribed UmTER with the in vitro synthesized recombinant UmTERT protein and successfully reconstituted telomerase activity (SI Appendix, Fig. S5A). In vitro reconstituted U. maydis telomerase was analyzed by direct primer extension assay and showed activity of template-directed DNA synthesis (SI Appendix, Fig. S5A, Bottom, lanes 1 and 6). The in vitro reconstituted U. maydis telomerase was not processive and capable of adding only a single repeat (SI Appendix, Fig. S5A, Bottom). This is in contrast to the ladder of bands observed using the TRAP activity assay with the telomerase holoenzyme from U. maydis cell lysate (SI Appendix, Fig. S1C, lane 1). We suspect that additional factors in the U. maydis telomerase holoenzyme may be responsible for the different telomerase processivity observed (36). TERs from all major eukaryotic groups contain two structural domains essential or crucial for telomerase catalysis (14). To identify structural domains within UmTER necessary for reconstituting telomerase activity, we carried out a series of truncation analyses on UmTER (SI Appendix, Fig. S5 B and C) and identified two minimal UmTER fragments, nt 171 to 861 and nt 1,042 to 1,141, which can assemble in trans with UmTERT to reconstitute activity (SI Appendix, Fig. S5C). It was noted that the longer 5F2 RNA fragment reconstituted a lower activity than the 5F3 RNA fragment (SI Appendix, Fig. S5C, Bottom Right, lanes 5 and 6), which could be due to a suboptimal folding of fragment 5F2 with an extended 5′ region compared to fragment 5F3 (SI Appendix, Fig. S5C, Left). The two minimal UmTER domains that bind UmTERT independently were consistent with the two regions covered by most sequencing reads from the Illumina RNA sequencing analysis of the RNA extracted from the purified U. maydis telomerase holoenzyme (Fig. 1E). More importantly, based on a phylogenetic comparative analysis of 18 Basidiomycota fungal TER homolog sequences (SI Appendix, Fig. S6 A and B), these two structural domains of UmTER fold into TER-specific structural domains, i.e., the T/PK and the template-distal CR4/5, which are commonly present in both animal (6) and Ascomycota fungal (9) TERs (Fig. 2A). Deletion analysis on these two structural domains further minimized the T/PK and CR4/5 core domains required for telomerase activity (Fig. 2 B–D), which include functionally crucial elements, such as the putative template boundary element (TBE) in the T/PK domain and the P6.1 stem-loop in the CR4/5 domain (Fig. 2A) (6, 9). An activity assay of telomerase reconstituted from the truncated T/PK and CR4/5 fragments identified the minimal fragments, namely, T/PK-Δ4 with three regions deleted and CR4/5(1046-1135)-D3 with most of the P6b stem deleted (Fig. 2B, lane 7, Fig. 2C, lane 5, and Fig. 2D, lane 3). Notably, telomerase reconstituted from one of the truncated T/PK fragments, namely, T/PK-Δ2 with only the region 723 to 784 deleted, showed significantly higher activity than T/PK-Δ4 (Fig. 2B, lanes 5 and 7), which was likely due to higher enzyme turnover. However, the underlying mechanism for such higher activity from a truncated T/PK domain remains to be explored. Taken together, our results identified the minimal regions required for enzymatic activity, which is consistent with the structural domains essential for telomerase function and conserved between animal and fungal TERs (6, 9).
Animal and Ascomycota yeast TERs are transcribed by RNA Pol II followed by hypermethylation of the 5' m7G cap by the TGS1 hypermethylase to generate a TMG cap (17, 18). The 5′ end of Trypanosoma TER is processed by transsplicing with a unique cap chemistry (37). To determine the 5′-end structure of UmTER, we first treated U. maydis total RNA with a decapping enzyme, namely, RNA 5′ pyrophosphohydrolase (RppH) (38), that removes the 5′-cap from the RNA Pol II transcripts leaving a 5′ monophosphate group that can then be ligated to an RNA adapter (Fig. 3A). The adapter-ligated RNA transcripts were then analyzed by a standard RNA ligase-mediated RACE (RLM-RACE) procedure using UmTER-specific reverse primers to amplify the 5′ cDNA of the UmTER transcripts (Fig. 3A). Intriguingly, the RppH decapping and RLM-RACE analysis detected multiple UmTER-related transcripts with various lengths. It appeared that the longer UmTER transcripts contain a 5′ cap and require RppH treatment to be detected by RLM-RACE (Fig. 3B, compare lanes 8 and 10 with R1 primer). In contrast, the short UmTER transcript contains a 5′ monophosphate and can thus be ligated to the RNA adapter for RACE detection with or without RppH treatment (Fig. 3B, compare lanes 8 and 10 with R1 primer). Moreover, the sequencing results of the PCR DNA product confirmed that the 5′ end of the short UmTER transcript is identical to the 5′ end of the mature UmTER determined previously by cap-independent template-switching RACE (SI Appendix, Fig. S2 A–C). We further confirmed the presence of a 5′ monophosphate in mature UmTER by treating the RNA with calf intestinal alkaline phosphatase (CIP) to remove 5′ phosphates. Upon CIP treatment, mature UmTER and 18S rRNA that also contains 5′-monophosphate cannot be detected by RLM-RACE (Fig. 3C, lane 5), which can then be rescued by T4 polynucleotide kinase (PNK) treatment to add back the 5′-monophosphate (Fig. 3C, lane 7). The 5S rRNA contained a 5′ triphosphate and was resistant to RLM-RACE (Fig. 3C, lanes 1 and 3), unless treated with CIP and T4 PNK sequentially to generate a 5′ monophosphate (Fig. 3C, lane 7). Furthermore, we confirmed the presence of a 5′ monophosphate in mature UmTER by treating the RNA with Terminator exonuclease that degrades specifically the RNAs with 5′ monophosphate such as rRNA but not the 5′-capped RNAs such as mRNA or snRNAs (Fig. 3D, Top). The RT-qPCR analysis of the Terminator exonuclease-treated RNA showed significant degradation of UmTER but not the U2 snRNA (Fig. 3D, Bottom). These results collectively confirm the presence of a 5′ monophosphate in the mature UmTER.
The presence of a 5′ monophosphate suggests that mature UmTER is posttranscriptionally processed at the 5′ end from larger precursors, likely from the 5′-capped UmTER transcripts detected by RLM-RACE in the RppH-treated sample (Fig. 3B, lanes 7 and 8). As expected, sequencing analysis of these longer cDNA products revealed a distinct 5′ end located further upstream of the 5′ end of mature TER (Fig. 3E). To directly detect the longer UmTER precursor transcripts, we performed Northern blot analyses with probes targeting specific regions in either the precursor or mature UmTER transcripts (Fig. 3E, Top). The radiolabeled riboprobe P1 detected the template region of UmTER, while the other three riboprobes P2, P3, and P4 targeted only the precursor transcripts at three different regions, i.e., near the transcription start site (TSS), upstream of the mature UmTER, and downstream of the mature UmTER, respectively (Fig. 3E, Top). The northern blot probed with the P1 riboprobe revealed a strong band of ∼1,300 nt RNA corresponding to the mature TER and a weak band of RNA with higher molecular weight (Fig. 3E, Bottom, lane 1). All three precursor-specific probes, namely, P2, P3 and P4, detected only the larger RNA, confirming the presence of the larger UmTER precursor transcripts (Fig. 3E, Bottom, lanes 2, 3, and 4). Interestingly, probe P4 targeting the region downstream of mature UmTER detected additional faint bands with smaller sizes (Fig. 3E, Bottom, lane 4), indicating that the 5′-end processing of the UmTER precursor likely precedes the 3′-end processing. To determine the identity of the 5′-cap structure of the larger UmTER transcripts, we performed 5′-cap specific RNA IP using anti-m7G, anti-TMG, or immunoglobulin G (IgG) antibodies, and we detected the affinity-purified RNA targets by RT-qPCR with primer sets specific to GAPDH mRNA, UmTER, or U2 snRNA. The results showed that the anti-m7G IP significantly enriched the GAPDH mRNA and UmTER precursor transcripts over the IgG antibody negative control, supporting the presence of 5′ m7G in the UmTER precursor transcripts, while the anti-TMG IP enriched only the 5′ TMG-capped U2 snRNA (Fig. 3F). A minor cross-reactivity of anti-m7G IP to the 5′ TMG of U2 snRNA was observed as previously reported (Fig. 3F) (39). Our results indicated that the UmTER-encompassing long transcripts are 5′-m7G capped. In addition to the presence of 5′-m7G cap, the promoter regions upstream of the UmTER precursor sequences were identified in 17 Ustilaginomycetes species and shared a conserved 5′-ACGCGAA-3′ sequence (SI Appendix, Fig. S7). This conserved promoter sequence was previously identified as a top-ranked binding site for the transcription factor Swi4 in a yeast chromatin IP study (40). The Swi4 transcription factor is part of a complex that activates mRNA transcription of multiple protein-coding genes involved in cell cycle regulation (41). Thus, the transcription of the UmTER precursor transcript may be regulated developmentally along the life cycle of U. maydis.
Sequencing analysis of the 5′-RACE DNA products indicated the presence of introns in the larger 5′ m7G-capped UmTER transcripts of various sizes (Fig. 3B, lanes 7 and 8). To analyze these UmTER isoforms quantitatively, we performed Nanopore long-read sequencing on a cDNA library enriched with the UmTER precursor transcripts (Fig. 4A). Briefly, we performed the RT reaction using a reverse primer targeting a region immediately downstream of the mature UmTER sequence to generate cDNA products of the precursor transcripts. Two rounds of nested PCRs were performed to enrich specifically the UmTER-containing sequences (Fig. 4A and SI Appendix, Materials and Methods). Nanopore sequencing of the UmTER-enriched library successfully generated 2.6 million valid long reads. Mapping and isoform analysis of these long sequencing reads revealed five isoforms, namely, A to E, of the UmTER precursor transcripts (Fig. 4B). All five isoforms encompass the mature UmTER sequence (Fig. 4B, Top) and could potentially serve as precursors for producing mature UmTER. Isoform A represented the unspliced UmTER precursor transcript and accounted for 57.5% of the reads (Fig. 4B, Bottom). Isoforms B, C, D, and E represented products of alternative intron splicing at two 5′- and two 3′-splice sites, each of which accounted for 14.6%, 9.7%, 9.7%, and 8.4% of the total reads, respectively (Fig. 4B, Bottom). The introns removed from each of the four spliced isoforms contain the canonical 5′ splice site, branch site, and 3′ splice site sequences (SI Appendix, Fig. S8A) that are universally conserved in fungal introns (42–44). The branch-site sequence, 5′-UGCUAAGA-3′ (branch point underlined), in the intron is conserved (42) and could base pair with the U. maydis U2 snRNA to form an adenosine bulge (SI Appendix, Fig. S8B). In addition to the presence of splicing isoforms, the UmTER precursor transcripts contain a 3′-poly(A) tail. The G/I tailing-mediated 3′-RACE located the 3′ end of the larger UmTER precursor transcripts at position 2295 from the TSS or 186 nt downstream of the 3′ end of the mature UmTER (SI Appendix, Fig. S9). Sequencing analysis of the cDNA product indicated the presence of a poly(A) tail with approximately 20 adenosine residues (SI Appendix, Fig. S9). In summary, the larger UmTER precursor transcripts showed hallmark characteristics of mRNA transcripts including a conserved mRNA promoter element, a 5′-m7G cap, alternative intron splicing, and a poly(A) tail.
To directly determine if the UmTER precursor is indeed processed to produce mature UmTER, we generated a U. maydis strain that expresses a recombinant UmTER precursor transcript from a Pol II mRNA promoter (Fig. 4C). This recombinant gene construct contains the hsp70 promoter to transcribe the first 390 nt of the U. maydis GAPDH protein-coding sequence, followed by the mature UmTER sequence with 200 bp upstream and 264 bp downstream flanking sequences, and the hsp70 terminator (SI Appendix, Fig. S10A). To differentiate the recombinant UmTER from WT UmTER, a 112-bp marker sequence was engineered to replace the region (257 to 368) (SI Appendix, Fig. S10A), which is located within a functionally dispensable region (208 to 440) (Fig. 2A). Following PCR validation of desired transformants (SI Appendix, Fig. S10 B and C), Northern blot analyses were performed using radioactive probes that detect either the engineered marker or WT sequence to confirm the expression of the recombinant UmTER transcript (Fig. 4C). The Northern blot showed the presence of the recombinant UmTER precursor transcript (Fig. 4 D, lane 4, red circle) and the processed mature UmTER (Fig. 4 D, lane 4, red triangle), suggesting that the expressed recombinant UmTER precursor can be processed to produce mature UmTER. The bands above the processed recombinant UmTER were likely transcripts partially processed at either the 5′ or 3′ end, but not both (Fig. 4D, lane 4). To quantitate the distribution of the recombinant precursor and mature UmTER transcripts, we performed Nanopore long-read sequencing using a similar strategy described in Fig. 4A but with a reverse primer targeting both the mature and precursor transcripts (Fig. 3C). More than one million (1,042,646) valid reads were obtained and 99% of the reads were mapped to the recombinant UmTER sequence with less than 0.5% of the reads mapped to the WT sequence. This dramatic mapping disparity is due to the overexpression of the recombinant UmTER compared to the WT UmTER gene (Fig. 4D, lane 4, black triangle). The mapping of the recombinant UmTER sequencing reads confirmed that the 5′-end position of the recombinant mature UmTER (Fig. 4E, red triangle) is identical to the 5′ end of the WT mature UmTER (SI Appendix, Fig. S2I). Similarly, we performed Nanopore sequencing to determine the 3′ ends of processed recombinant UmTER (Fig. 4F). Data analysis from over one million (1,175,036) valid reads confirmed the correct 3′-end processing of the recombinant mature UmTER (Fig. 4F, red triangle and SI Appendix, Fig. S2K). Furthermore, the 3′ end of the recombinant precursor transcript has the poly(A) tail at the same position as the WT precursor (Fig. 4F, Right and SI Appendix, Fig. S7). However, it was noticed that some RNA molecules smaller than the processed recombinant UmTER were detected in the Northern blot analysis (Fig. 4D, lane 4) but absent in the Nanopore sequencing analysis. We suspected these smaller RNAs were products of RNase degradation that possess a 3′ phosphate and thus were resistant to enzymatic G/I-tailing during Nanopore library preparation (Fig. 4D, lane 4). Overall, our result showed that a recombinant UmTER precursor expressed from a Pol II mRNA promoter can be correctly processed to generate mature UmTER.
Isoforms A, C, and E of the UmTER precursor transcripts contain an ORF that encodes a hypothetical protein, UMAG_03168 (NCBI RefSeq - XP_011389625.1). Protein BLAST searches using UMAG_03168 as a query identified homologs in both Basidiomycota (SI Appendix, Table S1) and Ascomycota (SI Appendix, Table S2) phyla. The homolog identified in S. cerevisiae with a BLAST E-value of 10−5 is a protein called Early meiotic induction protein 1 (Emi1). This Emi1 protein appears to induce the expression of a transcription factor called Inducer of meiosis (Ime1) during meiosis initiation in yeast (45, 46). To determine if UMAG_03168 is truly a homolog of yeast Emi1, we performed multiple sequence alignments of Basidiomycete homologs of UMAG_03168 (SI Appendix, Fig. S11A) and Ascomycete homologs of yeast Emi1 (SI Appendix, Fig. S11B) independently, which revealed two similar consensus sequences. A phylogenetic analysis of the aligned sequences for both UMAG_03168 and Emi1 homologs inferred a phylogenetic tree consistent with the established relationships between the two fungal phyla (SI Appendix, Fig. S11C), supporting the sequence homology between U. maydis UMAG_03168 and yeast Emi1 protein. In addition, we found the presence of highly conserved twin Cx9C motifs in both U. maydis UMAG_03168 and the yeast Emi1 protein sequence (Fig. 5A). Each Cx9C motif harbors a pair of cysteine residues spaced by 9 amino acids with positions 4 and 7 being predominantly hydrophobic (Fig. 5A) (47). In addition to the sequence analysis, we performed structure predictions of UMAG_03168 and yeast Emi1 protein sequences using Alphafold2 (48), which inferred two similar structures (SI Appendix, Fig. S11 D and E). These two structures clearly showed a high degree of structural homology with a similar positioning of the universally conserved residues in 5 alpha helices (α1 to α5) (SI Appendix, Fig. S11 D and E). In both predicted structures, the twin Cx9C motifs stabilize two antiparallel α-helices, namely, α2 and α4, through forming intramolecular disulfide bridges (Fig. 5 B and C). This type of hairpin-like arrangement has been observed in other proteins with the same twin Cx9C motifs (47). The conserved positions of the twin Cx9C motifs and the similar hairpin structure support structural homology between UMAG_03168 and the yeast Emi1 protein. Based on sequence and structural homology, we renamed UMAG_03168 as the U. maydis Emi1 protein (UmEmi1). However, UmEmi1 is likely to be encoded only in isoform A. Other alternatively spliced isoforms have much lower abundance (Fig. 4B) and would produce proteins with altered sequence and structure in helix α5 (SI Appendix, Fig. S11F). Although we suspect that UmEmi1 may play a role in meiosis induction in U. maydis (49), its exact function would require further functional studies. To experimentally confirm the expression of the 15.2-kDa UmEmi1 protein in U. maydis cells, liquid chromatography-tandem mass spectrometry analysis was performed. Briefly, U. maydis cell lysate was resolved on a sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel, and three gel slices covering three protein size ranges, namely, 14 to 16, 16 to 18, and 18 to 20 kDa, were excised and analyzed separately. Following in-gel trypsinization and mass spectrometry analysis, two tryptic peptides unique to UmEmi1, namely, ESAENVWELR and ESAENVWELRR, were detected with 100% and 92% probabilities, respectively (Fig. 5D and SI Appendix, Fig. S12A and Table S3). Protein BLAST analysis against the U. maydis proteome independently confirmed that both peptides are unique to UmEmi1. These two peptides differed by only one arginine at the C terminus, which was likely due to partial trypsin digestion. Importantly, both peptides were detected from only the SDS-PAGE gel slice covering protein sizes between 14 and 16 kDa and not from the other two gel slices, supporting that they were derived from the 15.2-kDa UmEmi1 protein. To further confirm the mass spectrometry results, we expressed a recombinant 3xFLAG-UmEmi1 protein in U. maydis cells (Fig. 5E). The expression of the recombinant protein was confirmed by anti-FLAG Western blot analysis (Fig. 5F). The recombinant 3xFLAG-UmEmi1 protein was then affinity purified by anti-FLAG IP and analyzed by the same mass spectrometry approach. From the purified 3xFLAG-UmEmi1 protein sample, three unique tryptic peptides, namely, YEQYVAEDVAYHK (×7), ESAENVWELRR (×3), and ESAENVWELR (×2), covering two separate regions of the UmEmi1 protein were detected multiple times (Fig. 5 G–I and SI Appendix, Fig. S12B and Table S3). Reproducible detection of the two peptides, ESAENVWELR and ESAENVWELRR, in both native UmEmi1 and the purified recombinant 3xFLAG-UmEmi1 protein samples validated the methodology employed and confirmed the expression of the UmEmi1 protein in U. maydis cells. Noticeably, spectra of the peptide fragment ESAENVWELRR from both native and recombinant UmEmi1 predominantly showed doubly protonated y ions (SI Appendix, Fig. S12), which was likely due to the protonation of both the arginine residues in the peptide (50). Collectively, the mass spectrometry-based detection of UmEmi1 in U. maydis cells supports the expression of the UmEmi1 protein and that the UmTER precursor is indeed a protein-coding mRNA.
The remarkable mRNA-derived biogenesis pathway employed by UmTER further demonstrates the incredible diversity of TER biogenesis across major eukaryotic kingdoms (Fig. 6). In most eukaryotic lineages including the basal branching trypanosome, TERs are predominantly transcribed by RNA Pol II (5), while ciliate and plant TERs are transcribed by RNA Pol III (7, 19–22). This dramatic Pol II-Pol III switch of TER transcription machinery presumably occurred early in eukaryotic evolution prior to the branching of two supergroups, namely, TSAR (Telonemia, Stramenopila, Alveolata, and Rhizaria) that includes ciliates and Archaeplastida (Chloroplastida) that include land plants (51, 52). Metazoa and fungi belong to the same supergroup Opisthokonta and use RNA Pol II for TER transcription. While it remains speculative which apparatus for TER transcription is employed in early eukaryotes, the Trypanosoma brucei TER being a Pol II transcript suggests that the RNA Pol II is the ancestral machinery for TER transcription (Fig. 6). The use of different transcription machinery inevitably led to distinct TER biogenesis mechanisms (Fig. 6). For example, the Pol III–transcribed TERs are relatively smaller in size and possess a 5′ triphosphate and a 3′ poly(U) tail that is typically bound and protected by an accessory protein with a La motif (53). In contrast, the Pol II–transcribed TERs from metazoa and fungi are significantly larger and possess a hypermethylated 5′-TMG cap and a processed 3′ end (Fig. 6). The 3′ ends of metazoan and ascomycete TERs are defined through snoRNA- and snRNA-like structural elements, respectively, and are processed by exosome-mediated exonucleases (Fig. 6). The UmTER precursor is transcribed by RNA Pol II as a typical mRNA transcript but then undergoes an unusual maturation process presumably through endonucleolytic cleavages in the 3′ UTR region to release the mature UmTER with a 5′ monophosphate (Fig. 6). In the absence of a protective 5′ cap, the mature UmTER would rely on stable RNA structural elements and/or protein binding (54–56) to protect the 5′ end from 5′-to-3′ degradation by exonucleases such as XRN1 (57). The 3′ end of ascomycete yeast TER contains an Sm binding site that is bound by the heptameric Sm ring protein complex to protect the 3′ end from 3′-to-5′ exonuclease degradation (27, 30). The 3′ region of mature UmTER also contains a putative Sm-binding site, 5′-AUUUUU-3′, located 9 nt upstream of the 3′ end (SI Appendix, Fig. S13), which suggests UmTER may undergo a 3′-end maturation process similar to the yeast TER. The S. pombe and N. crassa TERs employ a unique spliceosomal cleavage reaction to generate the 3′ end (30–32). We did not find any potential 5′ splice site at or near the 3′ end of the mature UmTER. Thus, the UmTER 3′-end maturation may not employ a terminal intron splicing but rather rely on an endonucleolytic mechanism that is yet to be identified. This mRNA-derived biogenesis of UmTER may have originated through either chromosomal translocations, inversions, or deletions that merged the Umter gene with the 3′ UTR of an mRNA gene, creating a unique polycistronic mRNA-lncRNA fusion gene (58). The conserved ORF in the UmTER precursor mRNA encodes a protein homologous to the yeast Emi1 protein (Fig. 5A-C and SI Appendix, Figs. S10 and S11). Yeast Emi1 is required for up-regulation of Ime1 that is the master regulator of yeast meiosis and is activated during early meiosis (45, 46). As U. maydis does not appear to harbor an Ime1 ortholog (49), the target genes of the UmEmi1 protein are yet to be identified. Since the UmEmi1 protein and UmTER are produced from the same polycistronic gene, the processing of UmTER from the mRNA precursor would presumably result in mRNA degradation and affect UmEmi1 protein expression. Thus, it remains to be explored if the UmTER biogenesis plays a role in U. maydis meiotic induction. Furthermore, disrupting the UmTERT coding gene (trt1) was shown to impair U. maydis teliospore production (59), a crucial step in the meiotic phase of U. maydis life cycle, which suggests telomerase may play a role in U. maydis meiosis. Elucidation of the detailed mechanism and regulation of this mRNA-derived UmTER biogenesis would bring more insights to RNA and telomere biology.
The identification of UmTER was carried out through affinity purification of recombinant U. maydis telomerase, TRAP telomerase activity assay, Illumina next-generation sequencing of TERT-bound RNA, and bioinformatics search of UmTER candidates. The validation of UmTER was performed via RNA secondary structure determination by phylogenetic comparative analysis, in vitro reconstitution of U. maydis telomerase activity, and TRF telomere length analysis of ΔUmTER strains. The analysis of UmTER biogenesis employed Northern blot analysis, specific enzymatic treatments, 5′-cap-specific IP, RT-qPCR analysis, and Nanopore long-read sequencing of both WT and recombinant UmTER precursor. The analysis of UmEmi1 protein expression employed in-gel trypsinization of native or recombinant UmEmi1 proteins and mass spectrometry analysis of tryptic peptides. Details of all materials and methods are available in SI Appendix, Materials and Methods. | true | true | false |
PMC9564188 | Xiao Ding,Lu Yu,Luo Chen,Yujie Li,Jinlun Zhang,Hanyan Sheng,Zhengwei Ren,Yunlong Li,Xiaohan Yu,Shuangxia Jin,Jinglin Cao | Recent Progress and Future Prospect of CRISPR/Cas-Derived Transcription Activation (CRISPRa) System in Plants | 28-09-2022 | CRISPRa,CRISPR/Cas,dCas9,genome editing,transcription activation | Genome editing technology has become one of the hottest research areas in recent years. Among diverse genome editing tools, the Clustered Regularly Interspaced Short Palindromic Repeats/CRISPR-associated proteins system (CRISPR/Cas system) has exhibited the obvious advantages of specificity, simplicity, and flexibility over any previous genome editing system. In addition, the emergence of Cas9 mutants, such as dCas9 (dead Cas9), which lost its endonuclease activity but maintains DNA recognition activity with the guide RNA, provides powerful genetic manipulation tools. In particular, combining the dCas9 protein and transcriptional activator to achieve specific regulation of gene expression has made important contributions to biotechnology in medical research as well as agriculture. CRISPR/dCas9 activation (CRISPRa) can increase the transcription of endogenous genes. Overexpression of foreign genes by traditional transgenic technology in plant cells is the routine method to verify gene function by elevating genes transcription. One of the main limitations of the overexpression is the vector capacity constraint that makes it difficult to express multiple genes using the typical Ti plasmid vectors from Agrobacterium. The CRISPRa system can overcome these limitations of the traditional gene overexpression method and achieve multiple gene activation by simply designating several guide RNAs in one vector. This review summarizes the latest progress based on the development of CRISPRa systems, including SunTag, dCas9-VPR, dCas9-TV, scRNA, SAM, and CRISPR-Act and their applications in plants. Furthermore, limitations, challenges of current CRISPRa systems and future prospective applications are also discussed. | Recent Progress and Future Prospect of CRISPR/Cas-Derived Transcription Activation (CRISPRa) System in Plants
Genome editing technology has become one of the hottest research areas in recent years. Among diverse genome editing tools, the Clustered Regularly Interspaced Short Palindromic Repeats/CRISPR-associated proteins system (CRISPR/Cas system) has exhibited the obvious advantages of specificity, simplicity, and flexibility over any previous genome editing system. In addition, the emergence of Cas9 mutants, such as dCas9 (dead Cas9), which lost its endonuclease activity but maintains DNA recognition activity with the guide RNA, provides powerful genetic manipulation tools. In particular, combining the dCas9 protein and transcriptional activator to achieve specific regulation of gene expression has made important contributions to biotechnology in medical research as well as agriculture. CRISPR/dCas9 activation (CRISPRa) can increase the transcription of endogenous genes. Overexpression of foreign genes by traditional transgenic technology in plant cells is the routine method to verify gene function by elevating genes transcription. One of the main limitations of the overexpression is the vector capacity constraint that makes it difficult to express multiple genes using the typical Ti plasmid vectors from Agrobacterium. The CRISPRa system can overcome these limitations of the traditional gene overexpression method and achieve multiple gene activation by simply designating several guide RNAs in one vector. This review summarizes the latest progress based on the development of CRISPRa systems, including SunTag, dCas9-VPR, dCas9-TV, scRNA, SAM, and CRISPR-Act and their applications in plants. Furthermore, limitations, challenges of current CRISPRa systems and future prospective applications are also discussed.
Gene expression involves multiple processes, including transcription of DNA into messenger RNA (mRNA), splicing of mRNA, translation, and post-translation modification. Accurate regulation of DNA transcription into mRNA is the first step to control the complex process of gene expression. Directional regulation of gene expression will contribute to our understanding of cell physiology, and it is essential for advances in biotechnology. For the regulation of endogenous gene expression, manipulating transcription factors (TFs) to target the specific target gene promoters to activate/inhibit gene transcription is the classical strategy and a successful one [1]. For example, in mammals, simultaneous up-regulation of four transcription factors reversed differential cells to pluripotent stem cells. However, it is difficult to regulate multiple genes’ transcription due to the specificity of TFs binding sites at DNA [2]. Researchers circumvented these limitations by designing a TFs binding site for promoters to regulate target genes transcription [3]. However, naturally occurring TFs have extensive DNA binding activity, which limits the specificity and efficiency of this method. Transcription factors can only bind to fixed sites and lack flexibility. Site-specific nucleases (SSNs) such as TALEN and CRISPR/Cas9 have emerged as multipurpose tools which can greatly enhance molecular biologists’ capability such as knock out, base edit, knock in, knock up and knock down the target gene [4,5,6,7]. These SSN systems emerged as genome editing tools, and introduce DNA double-strand breaks (DSBs) anywhere in a particular genome [8,9,10]. In known DNA binding modules, inactivation of Cas9 (dead Cas9, dCas9) fused to transcriptional activators or repressor [9,11] can effectively regulate multiple genes’ transcription under the guidance of different gRNAs [12]. Overexpression of foreign genes by traditional transgenic technology in plant cells is the routine method to verify gene function and shape gene regulation. However, there are still some limitations for the wide application of this strategy. One of the main limitations is the vector capacity constraint that makes it difficult to express multiple genes using the typical Ti plasmid vectors from Agrobacterium. The cumbersome of stacking gene cloning protocol is another obstacle that limits its application. The recent advancements in CRISPR-based gene activation have offered powerful and specific induction of gene expression that overcome the limitations of traditional gene overexpression methods. Multi genes activation can be effectively realized through the CRISPRa system which provides more possibilities for the application of plant genetic improvement in the future. In this paper, we summarized the emergence and development of transcriptional activation system based on CRISPR/dCas9 and its application in plant research. To provide a reference for plant researchers to compare the differences of different activation systems and to regulate the activation efficiency of endogenous genes in plants in the future. In addition, the potential applications, existing problems, and challenges for these new technologies were also discussed.
The CRISPR/Cas system was first investigated in 1987: scientists discovered a kind of unique DNA sequences from Escherichia coli genome, which are near the iap gene sequence and were called Clustered Regularly Interspaced Palindromic Repeats (CRISPR). However, its biological significance was unclear at that time [13]. Subsequent studies showed that CRISPR/Cas is a complex adaptive defense mechanism in prokaryotes against invading viruses or plasmid DNA [14]. Researchers found that about 40% bacteria and about 90% archaea are present in this powerful defense system [15]. Cas proteins involved in this defense mechanism have also been identified [16]. According to the repeat sequence identity of CRISPR and their Cas protein sequence homology, CRISPR/Cas system was classified into two classes and six types (as shown in Figure 1) [17,18,19]. Class I CRISPR/Cas systems require large effector protein complexes, which are classified into Type I, Type III, and Type IV. Class II CRISPR/Cas systems are classified into Type II, Type V, and Type VI, requiring only an RNA-directed endocytase to cut invading genetic components. The simplicity and efficiency of Class II system make it work as widely used genome editing tool. Type II CRISPR/Cas systems have been used in a variety of organisms, including microbes [20], fungi [21], animals [22], and plants [23].
Cas9 is a specific DNA endonuclease that existed in bacteria species such as Streptococcus sepsis, Staphylococcus aureus, and Streptococcus thermophilus. It is a multifunctional protein with two ribozyme domains HNH and RuvC. First, Cas9 forms a ribosome protein complex with two small non-coding RNAs, CRISPR RNAs (crRNAs) and trans activated crRNAs (tracrRNAs). Then the RNA complex will find and identify a suitable Protospacer Adjacent Motif (PAM), such as the 5‘-NGG-3’ sequence of the target sequence [24]. Finally, Cas9 with RNA complex search and match the crRNA target sequence, and then the HNH ribozyme domain will cut the target chain, while the RuvC domain will cut the reverse chain subsequently [25,26,27]. Qi et al. (2013) mutated two conserved endonuclease domains of Cas9 in CRISPR/Cas9 system. The aspartic acid at position 10 of RuvC domain was mutated to alanine (D10A) and histidine at position 840 of HNH domain was mutated to alanine (H840A), so that Cas9 protein lost endonuclease activity and became dead Cas9 (dCas9), which cannot cut DNA but still can bind to specific target DNA sequences with the guide RNA [28]. Therefore, dCas9 binds to the upstream region of the promoter transcription start site (TSS) region and disrupts RNA polymerase or transcription factor binding to the promoter, resulting in inhibition of gene expression without altering the genome. This gene transcription regulation strategy was defined as CRISPR interference (CRISPRi) [29]. The inhibition intensity of CRISPRi system can reach 1000-fold mostly in prokaryotes [30]. CRISPR interference is also widely used in plants such as Nicotiana tabacum, Zea mays, and Arabidopsis thaliana (as shown in Table 1) [31,32,33,34,35,36]. CRISPR interference can be used as an alternative tool for RNAi. Conversely, CRISPR/Cas-derived transcriptional activation (CRISPRa) also can be achieved by dCas9. Bikard et al. (2013) fused the dCas9 protein with ω subgroups (rpoZ) in E. coli, and this dCas9-ω complex increased the reporter gene transcriptional level up to 2.8-fold [37]. The models of CRISPR mediated transcriptional regulation are shown in Figure 2.
Guide RNA (gRNA) of CRISPR/Cas9 system is a specific RNA sequence composed of two elements: crRNA and tracrRNA. The gRNA recognizes the target DNA and directs Cas9 protein to produce double-strand breaks (DSB) in target DNA [10]. Any complementary gene or nucleotide of sgRNA sequences can be targeted by CRISPR systems. Furthermore, Cas9 and dCas9 can use multiple gRNAs effectively to expand the flexibility and multiplicity of CRISPR/Cas system by editing several different target genes simultaneously [38]. The selection of different regulatory elements and the change in regulatory efficiency can be achieved through the modification of scaffolds [39].
Transcription regulatory factors are essentially chimeric proteins, and their DNA binding domains are connected with the functional domains that control the transcription mechanism by promoting the recruitment of key cofactors to regulate transcription [35]. In the CRISPR/dCas9 meditated transcription regulation system, transcriptional regulation of target genes is achieved by fusing transcription regulators with dCas9. Transcriptional Repression Domains (TRD) include transcription repressors such as the Krüppel-Associated Box (KRAB) domain. Its function is to suppress transcription by collecting co-inhibitors KRAB-Associated Protein 1 (KAP-1), which leads to the formation of heterochromatin complexes that ultimately lead to gene silencing [40]. The transcriptional repressor SRDX originate from Ethylene-responsive element binding factor-associated Amphiphilic Repression (EAR) transcriptional repressor domain, which are effective plant transcriptional repressors [41]. On the other hand, transcriptional activators such as herpes simplex Virus Protein 16 (VP16) Transcriptional Activator Domain (TAD) or tetrameric repeat VP64 [42] could elevate the target genes’ transcription. Plant also has its own specific transcription regulators such as Ethylene Responsive Factor/Ethylene-Responsive Element Binding Proteins (ERF/EREBP). They can maintain high activity even in the presence of activating elements (such as VP16). This domain has been used as a tool for transcriptional inhibition in some studies [40]. Ethylene response factors in the ERF/EREBP family play a leading role in the response to biological and abiotic stress [43]. These transcriptional regulators bind to the APETALA2(AP2) DNA domain and various unnamed motifs [44,45,46]. These sequences are EDLL short sequences consisting of conserved glutamate (E), aspartic acid (D), and leucine (L). Several studies showed that the EDLL motif is a powerful tool for endogenous gene transcriptional activation [47,48]. Transcriptional activation domains and their detail description are summarized in Table 2.
As mentioned previously, transcriptional activators are essential for the regulation of transcriptional activation through dCas9. Regulation of target gene activation can be achieved by fusing transcriptional activators with dCas9, or by directing gRNA into scaffolds to recruit transcriptional activators. A comparison of different activation systems using these two strategies and evaluating their efficiency is summarized below:
Studies showed that the dCas9 protein fused with the transcription factor VP64 can recruit and stabilize the promoter complex to form dCas9-VP64 [49]. However, dCas9-VP64 is a low-intensity activator with less than 10-fold target gene transcription [32,34,36]. Although higher levels of expression may be desirable to observe more pronounced phenotypes associated with the function of certain genes, these results clearly show that Cas9-based transcription factors can activate endogenous gene expression [50]. Based on this, the researchers attempted to string other different activators together with dCas9 and this resulted in the following three systems:
Researchers found that the combination of multiple transcription factors with a single promoter significantly enhanced transcriptional activation of the downstream gene. This principle of signal amplification through protein polymers had been applied in biological system imaging and engineering design [51]. Based on this principle, Tanenbaum et al. (2014) developed a novel protein scaffold, a repeating peptide array called SunTag for transcription activation [52]. In the Super Nova Tag (SunTag) system, dCas9 is fused with tandem General Control Nonderepressible 4(GCN4) peptide repeats and each repeat connects a transcription regulator via an anti-GCN4 antibody called Single Chain Fragment Variable (scFv). The domain has highly affinity with relatively short nucleic acid sequences, allowing for protein polymerization on a single RNA template. Compared to the dCas9-VP64, SunTag enables multiple transcription factors such as Ten-Eleven Translocation (TET) and regulatory elements X to be recruited in one system. Therefore, multiple genes can be synergistically activated by different transcriptional regulators in tandem [53] as shown in Figure 3A.
In natural gene regulation system in cells, many transcription factor Activation Domains (ADs) can make changes in transcription through a coordinated collection of necessary activators. Chavez et al. (2015) hypothesized that transcriptional activation could be enhanced by combining multiple Ads [54]. A series of more than 20 known transcriptional effectors were fused to the C-terminus of dCas9 in an effort to enhance the transcriptional activation efficiency in human HEK 293T cells. The result show that dCas9-VP64, dCas9-p65 [55] and dCas9-Rta [56] were active. Then researchers used Cas9-VP64 as their starting scaffold and expanded up-regulation with p65 and Rta at the C-terminal, resulting in VP64-p65-Rta (VPR) activator complex as shown in Figure 3B. This novel activator displayed increased transcription levels of endogenous target gene, ranging from 20-fold to 320-fold over the original dCas9-VP64 activator [54].
Currently, the CRISPRa system is continually being optimized and improved in animal cells, whereas it is still at an early stage for plant cells. The dCas9-TV system is the first CRISPR/dCas9 activation system applied in plant cells [57]. Li et al. (2017) modified dCas9-VP64 with VP16’s octahedron VP128 instead of VP64 and this system activated LUC with a five-fold increase; an improved performance on the dCas9-VP64 (only a two-fold increase). TAD was then introduced as a second step to enhance dCas9-VP128 activity, including plant-specific EDLL and modified ERF2 (ERF2m) and TAD from the herpes simplex virus’ TALE element. Results showed that the combination of VP128 with a tandem ERF2m-EDLL motif (up to four copies) activated LUC transcription up to 12.6-fold compared to the base level, while combination of VP128 and TALE TAD (up to six copies) increased LUC transcription up to 55-fold. The results suggested that dCas9-6TAL-VP128 was a strong transcriptional activator and was called the dCas9-TV system [39,57] as shown in Figure 3C.
A common method for studying RNA localization is to insert multiple MS2-binding RNA scaffolds into the target RNA molecule and then tag the RNA molecule with GFP by recruiting MS2-GFP fusion proteins [51]. The inherently modular and programmable nature of RNA allows it to be used to coordinate biological assembly. First of all, RNA can recognize DNA targets via the complementary base pairing principle. Second, RNA contains a domain of RNA-protein interactions that are useful for recruiting specific proteins. Previous studies demonstrated that RNA scaffolds can coordinate the assembly of functional proteins [58]. Therefore, a second common strategy for CRISPR/dCas9 derived transcriptional activation system is to use gRNA as scaffold to recruit transcriptional activators.
Zalatan et al. (2015) constructed a scaffold RNA (scRNA) system and demonstrated that it can effectively activate gene transcription in yeast and human cells [29]. ScRNA systems are inspired by natural regulatory systems in which scaffold proteins physically assemble interacting components of cell signaling pathways. Similar scaffolding principles were applied in genomic modification, such as using Long-strand Non-Coding RNA (lncRNA) as assembly scaffolds to recruit key epigenetic modifiers at specific genomic sites [59,60]. In this system, the researchers fused the well-characterized viral RNA sequences including MS2, PP7, and Com into the 3′-end of gRNA, which are recognized by RNA binding proteins, respectively. Then, the transcriptional activation domain VP64 was fused to each corresponding RNA binding proteins, as shown in Figure 3D.
Synergistic Activation Mediator (SAM) continues to be constructed by modifying gRNA loops. Initially, the researchers tried to find the best anchoring position for the activation domain in the Cas9-sgRNA complex. Previously, dCas9-based transcription activators relied on trans activation domains fused to the N or C ends of dCas9 proteins. For an investigation regarding alternative anchorage sites, Konermann et al. (2015) investigated the crystal structure of dCas9 and revealed that the distal gRNA loops did not interact with dCas9 at all. Other research revealed that the gRNA loops could fuse with protein-interacting RNA domains to promote transcription factor recruitment to dCas9 [61]. They combined VP64 with the NF-κB activation p65 [62] and reintroduced the activation domain of Human Heat Shock Transcription Factor 1 (HSF1) for improving the efficiency of dCas9-mediated gene activation (Figure 3E). Finalization of MS2-p65-HSF1 fusion protein improved ASCL1 (12%) and MYOD1 (37%) transcription activation [61]. Some results demonstrated that the SAM gene activation platform can facilitate in vivo research and drug discovery [63].
The CRISPR-Act 2.0 system is an improved system for multiple transcriptional activation in plants, which is also dependent on MS2. It is capable of recruiting four VP64 proteins to gRNA and the fifth VP64 was carried by the dCas9-VP64 complex. As shown in Figure 3F, this system can carry a total of five activators to the target site. When compared to the original dCas9-VP64 system, CRISPR-Act 2.0 increased transcriptional activation efficiency by three to four folds [64]. Additionally, the CRISPR-Act 2.0 system enables simultaneous activation of multiple genes in vivo. Lowder et al. (2017) assessed the efficacy of the CRISPR-Act 2.0 system in rice. They simultaneously activated three independent endogenous genes Os11g35410, Os03g01240 and Os04g39780 in rice protoplasm [48,64]. The results indicated that CRISPR-Act 2.0 was more effective than the previous dCas9-VP64 systems.
Based on the CRISPR-Act 2.0 system, Pan et al. (2021) combined dCas9-VP64, gR 2.0 scaffolds, 10xGCN4 SunTag and a newly developed 2xTAD activator to build a novel CRISPR-activating system, CRISPR-Act3.0 [65], as shown in Figure 3G. Multigene activation was achieved by assembling gRNA with multiple activators. At the same time, the further combination with CRISPR-Cas12b and the SpCas9 variant SPRY may expand the target range of CRISPR-activation. The primary objective of functional genomics is to determine the causal relationship between gene expression and phenotypic characteristics. By regulating gene expression in plants, the CRISPRa system provides a novel, efficient method to simplify and accelerate these studies. The future holds a lot of potential for improving CRISPR-activation efficiency, flexibility, and scalability.
First, the researchers used A. thaliana and N. benthamiana for transcription activation test to evaluate the transcriptional activation activity of dcas9-VP64. The data suggested that dCas9-VP64 had weak activation when a single sgRNA was designed in the vector. Then, the adoption of multiple sgRNAs to target the same gene promoter resulted in higher level of gene activation [32,35,47,66]. Transcription activation domains from plant AP2 transcription factors as known as the EDLL motif and bacterial TALE protein have been used to construct dCas9-based transcriptional activators in plants [64,66]. Li et al. (2017) designed a transcriptional activation system dCas9-TV that works effectively in plants. High-level transcriptional activation of target genes in A. thaliana and O. sativa and rice cells was achieved with a 201-fold and 2745-fold increase, respectively [57]. Xiong et al. (2021) further optimized dCas9-TV system in rice to increase the transcription of OSER1 gene up to 4000-fold [67]. Researchers also tried to link the EDLL motif with VP64, in a similar fashion to the VPR strategy, in order to increase the efficiency of transcriptional activation in rice [68]. Selma et al. (2022) used multipliable CRISPR activator dCasEV2.1 activated 3 and 7 genes with gene activation levels ranging from 4- to 1500-fold in N. benthamiana [69]. Similarly, transcriptional activation of endogenous target genes was also efficiently applied in A. thaliana when SAM or SunTag were fused with the dCas9 [70,71]. When using the same gRNA, the CRISPR-Act2.0 system significantly outperformed the dCas9-VP64 system to activate target genes in A. thaliana cells with a 1500-fold increase [64]. CRISPR-Act3.0 recruits additional activators based on the SunTag system coupled to MS2-MCP which increased activation efficiency of endogenous genes in rice (250-fold), Arabidopsis (4000-fold) and tomato (240-fold), respectively [65]. It is important to note that even with effective activation systems, such as CRISPR-Act 2.0 and dCas9-TV, some endogenous genes are also difficult to activate effectively [64,66]. It is speculated that dCas9 and activators must compete with endogenous transcription activators to bind the specific region of promoters. The activation system and activation efficiency applied in plant species are summarized in Table 2. The current CRISPR/dCas9-based transcription system needs to be delivered in plant by Agrobacterium, Polyethylene Glycol (PEG) or particle bombardment techniques. The donor plants should harbor the T-DNA insertion with CRISPR/dCas9 fragment may be regulated under the biosafety framework for GM plants since they contain transgene element in the genome. This is maybe the major concern for wide application of the current CRISPR/dCas9 mediated transcription activation system in plant. In addition, with the miniaturization of Cas protein and the maturity of virus introduction technology, it will leave the tissue culture stage and accelerate the efficiency of related research. It is anticipated that transcriptional regulation using CRISPR will be further improved by overcoming the technical difficulties mentioned above in the near future.
This paper reviews the different strategies developed based on CRISPR/dcas9 activation system in recent years. These strategies have been widely employed to study gene transcription regulation in animal cells; however, their application in plants is still at its early stages. In plants, this system is mainly used in the model plant species such as Arabidopsis and rice, but not in other major crop plant species. Nevertheless, there is still a great deal of potential for research in this field. First, the multi gene activation system enables large-scale transcriptional regulation in plants in order to better understand gene regulatory networks. Second, the up-regulation of multiple key genes in the metabolic pathway can generally result in the production of valuable commercial products, and synthetic biology is likely to make a significant breakthrough in the field of agriculture [72]. Additionally, the improvement of the protein complexes function is mainly dependent upon the simultaneous activation of many different genes, and the multigene activation system is helpful for studying the functional region of complex proteins [73]. Finally, directed activation of multiple defense genes against pathogen attack is a potential strategy to improve plant immunity without affecting traits [74]. Plants are able to recognize insect molecules and respond accurately and defensively when insects are grazing. However, the effectors released by insects interfere with the host plant’s defensive response. We can use the CRISPR-activation system to avoid the destruction of plant defense mechanisms by insect effectors and create broad-spectrum insect-resistant plant materials. Several novel gene editing tools such as Cas12a T-containing PAM with short guide RNA (42–44 nucleotides of crRNA), appeared to be more effective in regulating multiple genes [75]. Ongoing efforts are being made to update the transcriptional activation system based on CRISPR/dCas9. For example, Chiarella et al. (2020) improved CRISPR/dCas9-based transcriptional activation using Chemogenetic Epigenetic Modifiers (CEMs). By employing endogenous chromatin activators, this system was able to activate target gene expression without requiring exogenous transcriptional activators, leading to dose-dependent activations of target genes [76]. Gamboa et al. (2020) integrates the heat shock switch with dCas9 complex to remotely control gene activation and inhibition with short-time heating pulses [77]. The experimental results demonstrated that the activation intensity of the dCas9-VP64 complex with heat as remote trigger depends on thermal pulse and can be substantially improved in only 15 min. With this heat-activated transcriptional activation system, CRISPR/dCas9 can activate transcription without invasive procedures. Optical regulation can be used as another new method for inducing and regulating endogenous genes in plants. Moreover, in an attempt to overcome the problems associated with the compatibility of optogenetic tools with plant growth requirements, Ochoa-Fernandez et al. (2020) developed Plant-Usable Light-Switch Elements (PULSE). A combination of blue light sensing inhibition regulation with red light sensing activation regulation resulted in gene expression being regulated only in the presence of red light. Combining PULSE with CRISPR/dCas9 mediated gene activation system (dCas9-TV) demonstrated light controlled activation of A. thaliana [78]. As opposed to ribozymes with cutting activity, which often cause uncertain genomic modifications and result in chromosome rearrangements or deletion during multiple point editing, dCas ribozymes offer a superior solution to these problems [79]. However, the challenge remains in creating long sequences of multiple gRNAs strung together and determining their editing efficiency for various targets [80]. Under a multilocus CRISPR editing system, the increased number of gRNAs leads to limited dcas9 competition between different gRNAs, which in turn leads to variations in target gene editing efficiency [81], as well as uncertainty in the regulation of target genes by gRNAs [82]. As mentioned previously, CRISPR/dCas9 derived transcription activation system exhibited obvious advantages over the traditional overexpression strategy to elevate the target genes’ transcription. Epigenetic regulation is an important way to regulate gene expression and has certain reversibility. The researchers used CRISPR/Cas9 to knock out epigenetic factors to determine the role of epigenetic factors in the regulation of endogenous genes in plants [83]. In addition, researchers can combine CRISPR/dCas with different epigenetic effector domains for specific epigenetic regulation of target sites [84]. In the future, with the improvement of episequencing technology and the in-depth study of epigenetic regulators, CRISPR-based epigenetic regulation research will have room for development. Finally, with the wide application of single-cell sequencing technology in plants [85], it will further improve the transcriptional regulation information of plants [86] and open up new space for the application of CRISPR-activation system in plants. | true | true | true |
|
PMC9564198 | Min Chen,Lei Wang,Min Li,Marietta M. Budai,Jin Wang | Mitochondrion-Mediated Cell Death through Erk1-Alox5 Independent of Caspase-9 Signaling | 29-09-2022 | caspase-9-independent cell death,Erk1,Alox5,ROS,lipid peroxidation | Mitochondrial disruption leads to the release of cytochrome c to activate caspase-9 and the downstream caspase cascade for the execution of apoptosis. However, cell death can proceed efficiently in the absence of caspase-9 following mitochondrial disruption, suggesting the existence of caspase-9-independent cell death mechanisms. Through a genome-wide siRNA library screening, we identified a network of genes that mediate caspase-9-independent cell death, through ROS production and Alox5-dependent membrane lipid peroxidation. Erk1-dependent phosphorylation of Alox5 is critical for targeting Alox5 to the nuclear membrane to mediate lipid peroxidation, resulting in nuclear translocation of cytolytic molecules to induce DNA damage and cell death. Consistently, double knockouts of caspase-9 and Alox5 in mice, but not deletion of either gene alone, led to significant T cell expansion with inhibited cell death, indicating that caspase-9- and Alox5-dependent pathways function in parallel to regulate T cell death in vivo. This unbiased whole-genome screening reveals an Erk1-Alox5-mediated pathway that promotes membrane lipid peroxidation and nuclear translocation of cytolytic molecules, leading to the execution of cell death in parallel to the caspase-9 signaling cascade. | Mitochondrion-Mediated Cell Death through Erk1-Alox5 Independent of Caspase-9 Signaling
Mitochondrial disruption leads to the release of cytochrome c to activate caspase-9 and the downstream caspase cascade for the execution of apoptosis. However, cell death can proceed efficiently in the absence of caspase-9 following mitochondrial disruption, suggesting the existence of caspase-9-independent cell death mechanisms. Through a genome-wide siRNA library screening, we identified a network of genes that mediate caspase-9-independent cell death, through ROS production and Alox5-dependent membrane lipid peroxidation. Erk1-dependent phosphorylation of Alox5 is critical for targeting Alox5 to the nuclear membrane to mediate lipid peroxidation, resulting in nuclear translocation of cytolytic molecules to induce DNA damage and cell death. Consistently, double knockouts of caspase-9 and Alox5 in mice, but not deletion of either gene alone, led to significant T cell expansion with inhibited cell death, indicating that caspase-9- and Alox5-dependent pathways function in parallel to regulate T cell death in vivo. This unbiased whole-genome screening reveals an Erk1-Alox5-mediated pathway that promotes membrane lipid peroxidation and nuclear translocation of cytolytic molecules, leading to the execution of cell death in parallel to the caspase-9 signaling cascade.
During intrinsic apoptosis, mitochondrial outer membrane permeabilization leads to the release of cytochrome c into the cytosol to form apoptosome with Apaf-1 and caspase-9, resulting in the activation of caspase-9 [1,2,3,4]. Caspase-9 is the initiator caspase that activates the downstream effector caspases, including caspase-3, caspase-6 and caspase-7 [5]. The cleavage of DFF45/ICAD by caspase-3 leads to the activation of DFF40/CAD that enters the nucleus to mediate DNA cleavage and apoptosis [6,7,8,9]. In addition to cytochrome c, Smac/DIABLO released from mitochondria can facilitate the activation of caspase-9 and caspase-3 by sequestering XIAP [10,11]. Moreover, AIF and endonuclease G (EndoG) have been shown to translocate from disrupted mitochondria to the nucleus to induce DNA damage [12,13]. In addition to caspase-dependent apoptosis, different non-apoptotic forms of cell death have been described [14]. In the absence of caspase activation, cell death signaling can induce Ripk1 and Ripk3 activation by autophosphorylation [15,16], followed by phosphorylation and activation of downstream pore-forming MLKL to trigger necroptosis [17,18]. Inflammasome-mediated activation of caspase-1 can induce the cleavage of gasdermin D (GSDMD), resulting in the generation of N-terminal pore-forming GSDMD-N to trigger pyroptosis [19]. Elevated intracellular iron and depletion of antioxidant glutathione can cause lipid peroxidation and ferroptosis, which is inhibited by glutathione peroxidase 4 (GPX4) [20,21,22]. Although the caspase-9-dependent pathway can rapidly trigger apoptosis after mitochondrial disruption, it appears that interruption of caspase-9-dependent apoptosis may shift cells to engage other cell death mechanisms. Caspase-9-knockout mice display perinatal lethality with prominent defects in brain development that may be associated with defective cell death in neuronal progenitor cells [23,24]. However, cell death in lymphocytes and other cell types in the absence of caspase-9 appears to be largely intact [25]. In addition, apoptosis is efficiently inhibited by the over-expression of Bcl-2, but not by the deletion of caspase-9 [26], suggesting that caspase-9-independent cell death mechanisms can efficiently carry out intrinsic apoptosis. Cells can often undergo caspase-independent cell death when caspase function is inhibited [27,28]. Erk-dependent induction of caspase-independent cell death has been shown in neuronal cells [29]. Induction of caspase-independent cancer cell death may also be a common feature for many chemotherapeutic drugs [30,31,32,33,34,35,36,37,38]. These studies indicate that caspase-9-independent mechanisms are involved in cell death under different physiological and pathological conditions. Caspase-8-independent cell death mechanisms have been intensively studied. Rip1 and Rip3 have been shown to mediate necroptosis when caspase-8 activation is inhibited [15,16,17,39,40,41]. While the release of cytochrome c into the cytosol triggers caspase-9 activation [1,2,3,4], ROS production by mitochondria has been suggested to induce caspase-independent cell death [42]. However, genes that are critical for ROS generation to induce cell death independent of caspase-9 have not been defined. Moreover, downstream signaling events that mediate caspase-9-independent cell death have not been systemically characterized. Mitochondrion-dependent cell death is important for the maintenance of T cell homeostasis and immune tolerance [43]. Whether caspase-9 signaling downstream of the mitochondrion can regulate T cell functions and homeostasis has not been extensively studied due to perinatal lethality of caspase-9−/− mice [23,24]. We therefore generated mice with conditional knockout of caspase-9 in T cells (T/casp9−/−). However, caspase-9−/− T cells efficiently underwent cell death induced by different stimuli. Moreover, we did not detect T cell accumulation in T/casp9−/− mice, suggesting that the existence of other cell death mechanisms independent of caspase-9 in T cells. To define the caspase-9-independent cell death pathway downstream of mitochondrion disruption, we performed a genome-wide siRNA library screening for genes involved in cell death using caspase-9-deficent Jurkat T cells. Our data suggest that mitochondrial disruption leads to Erk1-dependent phosphorylation and nuclear membrane translocation of Alox5, resulting in membrane lipid peroxidation, nuclear entry of nucleases and cytolytic molecules to induce cell death.
Lck-cre mice (The Jackson Laboratory) were bred with caspase-9flox mice [28] to obtain Lck-cre/caspase-9flox/flox mice with T cell-specific knockout of caspase-9 (T/casp9−/−). Alox5−/− mice (The Jackson Laboratory) were crossed with T/casp9−/− mice to obtain T/casp9−/−Alox5−/− mice. Experiments were performed according to federal and institutional guidelines and with the approval of the Institutional Animal Care and Use Committee of Baylor College of Medicine (AN-2099) and the Houston Methodist Research Institute (IS00006582).
JMR cells [44], a caspase-9-deficient human T cell leukemia Jurkat cell line, cultured in 384-well plates were robotically transfected with 40 nM of a pool of 4 siRNAs targeting each of a total of 21,121 genes in the human genome (Dharmacon, Lafayette, CO, USA) by reverse transfection using RNAiMax (Life Technologies, Carlsbad, CA, USA) at the MD Anderson Cancer Center siRNA Screening Facility. siRNAs for each gene were transfected into JMR cells in 6 replicates of 384-well plates. In addition to the targeting siRNAs, each plate also contained transfection with non-targeting siRNA as negative control, as well as siRNA targeting PLK (siPLK) to induce cell death [45]. Forty-eight hours after transfection, the cells were cultured with 6 μM ABT-263 or solvent control in triplicates for another 48 h. Live cells were quantitated by the Celltiter-Fluor cell viability assay (Promega, Madison, WI, USA). The percentage of cell death induced by ABT-263 treatment was calculated as follows: (control-treated)/control × 100%. We scored a gene as a positive hit if the killing rate after its silencing is more than 2× standard deviation (SD) above or below the average killing rate on its plate. According to this standard, we identified 776 such genes in the first round of siRNA screening. In the second round of screening, four siRNAs targeting these 776 gens were transfected into JMR cells separately. Killing with ABT-263 and quantitation of cell viability was performed in the same way as in the first round of screening. The candidate genes identified in the second round of screening were used for further analyses. The candidates for cell death and anti-cell death genes are listed in Tables S1 and S2. To characterize the function of some of these candidate genes identified in the second round of screening, two to three validated siRNAs targeting selected genes were obtained from Life Technologies and used for transfection of JMR cells at a final concentration of 100 nM and used for various analyses. The sequences for these siRNAs are listed in Table S3.
Cells were cultured in 96-well tissue culture plates in the presence or absence of indicated treatments for different time. The cells were stained with APC-annexin V (ThermoFisher, Waltham, MA, USA) and 5 μg/mL propidium iodide (PI) (ThermoFisher), followed by flow cytometry analyses. Percentage of cell death induced by treatments with different apoptosis stimuli was calculated from live cells of untreated and treated groups as follows: (untreated-treated)/treated × 100%. In some experiments, 10 μM necrostatin-1, necroX-5 (Enzo Life Sciences, Farmingdale, NY, USA) or carbobenzoxy-valyl-alanyl-aspartyl-O-methyl-fluoromethylketone (zVAD, Promega) was included in the culture. Cells were also stained with APC-conjugated Annexin V (BD Bioscience, San Jose, CA, USA) in some assays. Mouse T cells were stimulated cultured with 5 μg/mL Concanavalin A (con A) ConA and 100 U/mL IL-2 for 48 h and used for different assays. To determine the effects of phosphorylation of Alox5, substitution of Serine 271 to Alanine (S271A) in Alox5 was performed using the site-directed mutagenesis kit (Stratagene, San Diego, CA, USA). Plasmids expressing Alox5-GFP (Origene, Rockville, MD, USA), Alox5S271A-GFP or GFP alone were transfected into JMR cells by electroporation using the Neon transfection system (Life Technologies). After 24 h culture, the cells were treated with 6 μM ABT-263 or solvent control. The loss of live GFP+ cells after ABT-263 treatment was determined 24 h later.
To measure ROS, JMR cells were cultured in the presence or absence of ABT-263 for 12 h. The cells were then incubated with 5 μM Mito-SOX (Life Technologies) at 37 °C for 30 min. In some experiments, cells were also stained with APC-annexin V, followed by flow cytometry. To determine lipid peroxidation, JMR cells with or without ABT-263 treatment as above were incubated with 5 μM BODIPY 581/591 C11 (Life Technologies) at 37 °C for 1 h and analyzed by flow cytometry.
Cells treated with cell death stimuli for 12 h were added to glass slides by cytospin and incubated with monoclonal mouse anti-TIA-1 (Abcam, Boston, MA, USA), monoclonal mouse anti-AIF (Santa Cruz Biotechnologies, Dallas, TX, USA) or polyclonal rabbit anti-EndoG (ProSci, Fort Collins, CO, USA). The cells were then incubated with Alexa Fluor 488-conjugated goat anti-mouse or anti-rabbit IgG (Life Technologies). The nuclei were counterstained with DAPI (Thermo Scientific). Cells treated with cell death stimuli for 24 h were used for TUNEL staining using the Click-iT TUNEL assay kit (Life Technologies). The nuclei were counterstained with DAPI. Alox5-GFP or Alox5S271A-GFP expression plasmids were transfected into JMR cells by electroporation using the Neon transfection system (Life Technologies) and cultured for 24 h. The cells were then treated with ABT-263 or solvent control for 8 h, followed by cytospin onto glass slides and DAPI staining. The staining was examined using a Nikon Eclipse 80i fluorescence microscope.
pCMV6-Alox5-GFP (Origene) was transfected into 293T cells with vectors expressing FLAG-tagged MAPKs. The cells were lysed 24 h later and immunoprecipitated with anti-GFP (Origene), followed by Western blot with anti-phospho-Alox5S217 (Cell Signaling Technology, Danvers, MA, USA). The blots were also probed with anti-GFP. Total lysates were used for Western blot by probing with anti-FLAG (Sigma, St. Louis, MO, USA). JMR cells transfected with siRNA targeting specific genes or non-targeting siRNA control were lysed for Western blot by probing with specific antibodies. The following antibodies were used for Western blot: rabbit polyclonal or monoclonal antibodies to Alox5, ERK1, p38 MAPK, p44/42 MAPK (ERK1/2), phosphor-p44/p42 MAPK (p-ERK1/ERK2), RAD51, SAPK/JNK, SESN2 (Cell Signaling Technology), Alox5AP, DHODH, HADHA, MGST1, TIA-1 (Abcam), EndoG (ProSci), OXR1 (Bethyl Laboratory, Montgomery, TX, USA.) and PDP1 (Sigma); and mouse monoclonal antibodies to AIFM1, API5, PNKP (Santa Cruz Biotechnology), caspase-3, caspase-6 caspase-7, caspse-9, phosphor-p38 MAPK, phosphor-SPAK/JNK (Cell Signaling Technology), FLAG (Sigma) and GFP (Origene). The blots were also probed with mouse monoclonal anti-α tubulin (Santa Cruz Biotechnology) to ensure equal loading.
Data are presented as mean ± SD, and p values were determined by a two-tailed Student’s t-test using the GraphPad Prism software.
To define the roles for caspase-9-dependent intrinsic cell death in the regulation of T cell apoptosis and functions, we crossed Caspase-9flox mice [28] with lck-cre mice to generate T cell-specific knockout of caspase-9 (T/casp9−/−). As expected, caspase-9 deletion led to virtually complete suppression of activation of effector caspases, including caspases-3, -6 and -7, in T cells treated with etoposide (Figure 1a). Despite the lack of caspase activation, cell death could be induced efficiently in T cells by etoposide (Figure 1b). Similarly, cell death could be induced efficiently in caspase-9-deficient T cells by another apoptosis inducer, staurosporine (Figure 1b). In treatments with etoposide, staurosporine also induced DNA fragmentation by TUNEL staining in both caspase-9−/− T cells and wild-type controls (Figure 1c), indicating that cell death in the absence of caspase-9 also involves nuclear DNA damage. These results suggest that cell death can proceed efficiently in the absence of caspase-9 signaling cascade. We also treated T cells with ABT-263, a BH3-mimetic that induces mitochondrion-dependent apoptosis by specifically inhibiting Bcl-2 and Bcl-xL [46]. Compared to wild-type controls, caspase-9−/− T cells showed a small but reproducible decrease in killing by ABT-263 (Figure 1b). However, DNA fragmentation as shown by TUNEL staining was not significantly affected by caspase-9 deficiency (Figure 1c), suggesting that alternative mechanisms efficiently mediate DNA cleavage independent of caspase-9. This suggests that inhibition of Bcl-2 and Bcl-xL with ABT-263 triggers both caspase-9-dependent and -independent cell death. Why caspase-9−/− T cells showed decreased killing by ABT-263, but not by etoposide or staurosporine is not entirely clear. It is possible that ABT-263 induces cell death exclusively through mitochondria and is more severely influenced by caspase-9 deficiency. However, knockout of caspase-9 did not cause T cell expansion in T/casp9−/− mice (Figure S1), suggesting that caspase-9-independent cell death mechanisms are sufficient for the maintenance of T cell homeostasis in vivo. We found that a cell death inhibitor targeting oxidative stress, necroX-5 [47], significantly inhibited cell death induced by etoposide, staurosporine or ABT-263 in caspase-9−/− but not in wild-type T cells (Figure 1d). In contrast, a RIP1-specific necroptosis inhibitor, necrostatin-1 [48], did not inhibit cell death in caspase-9−/− T cells induced by these apoptosis stimuli (Figure 1d). In addition, ABT-263-induced cell death was partially inhibited by the pan-caspase inhibitor zVAD in wild-type but not in caspase-9−/− T cells (Figure S2a), indicating that cell death observed in caspase-9−/− T cells is caspase independent. Interestingly, wild-type and caspase-9−/− T cells that underwent ABT-263-induced cell death displayed positive staining by both Annexin V and Mito-SOX (Figure S2s). These data suggest that oxidative stress is required for cell death in caspase-9−/− T cells but dispensable for that in wild-type T cells.
We next characterized caspase-9-independent cell death mechanisms downstream of mitochondrial disruption using a caspase-9-deficient human T cell leukemia Jurkat cell line, JMR [44]. JMR cells can be readily transfected with siRNA to silence target genes, Bax and Bak (Figure S3). Cell death in JMR cells induced by an inhibitor for Bcl-2 and Bcl-xL, ABT-263 [46], was suppressed by silencing of BAX and BAK (Figure S3). To identify genes that either inhibit or enhance ABT-263-induced cell death in caspase-9-deficient JMR cells, we used a siRNA library targeting of 21,121 genes of the human genome (Dharmacon) to perform a genome-wide screening (Figure 2a). The pools of four siRNA oligos targeting each of the genes were transfected into JMR cells in six replicates of 67 sets of 384-well plates. Forty-eight hours later, the transfected cells were treated with either ABT-263 or solvent control in triplicates for another 48 h, followed by measurement of viability. The percentage of killing of JMR cells induced by ABT-263 was calculated for cells transfected with siRNAs targeting each gene. We scored a gene as a positive hit if the killing rate after its silencing is more than 2× the standard deviation (SD) above or below the average killing on its plate. We identified 776 such candidate genes in the first screening. siRNA oligos targeting each of these genes were transfected separately into JMR cells to perform the second round of screening. A gene was scored as positive if at least two out of four of its siRNAs significantly increased or decreased ABT-263-induced killing of JMR cells. We identified 70 candidate pro-cell death genes whose silencing decreased ABT-263-induced killing of JMR cells (Figure 2b, Table S1). We also found 53 candidate anti-cell death genes whose silencing increased ABT-263-induced cell death (Figure 2b, Table S2). Interestingly, positive hits for candidate pro-cell death genes include oxidative stress-related genes, an endonuclease, MAP kinases and previously identified pro-apoptotic genes (Figure 2b, Table S1). Besides previously identified anti-apoptotic genes, positive hits for candidate anti-cell death genes include anti-oxidative genes and those genes encoding DNA damage repair enzymes (Table S2). These results support the potential roles for oxidative stress and DNA damage in mediating caspase-9-independent cell death.
In addition to pro-apoptotic BAX, other genes encoding mitochondrial proteins whose silencing suppressed cell death were also identified in the initial screening (Table S1). Transfection of individual validated siRNAs (Table S3) confirmed that dihydroorotate dehydrogenase (DHODH), 3-hydroxyacyl-CoA dehydrogenase Trifunctional Multienzyme Complex Subunit Alpha (HADHA) and pyruvate dehydrogenase phosphatase catalytic subunit 1 (PDP1), which are associated with respiratory chain, fatty acid oxidation or glucose metabolism [49,50,51], were involved in ABT-263-induced cell death in JMR cells (Figure 3a,b). This indicates that the mitochondrial proteins encoded by these genes promote caspase-9-independent cell death signaling after mitochondrial disruption. Because caspase-9-independent cell death involves oxidative stress (Figure 1d), we determined whether these mitochondrial proteins are required for ROS generation during mitochondrion-dependent cell death. The mitochondrial pyrimidine biosynthesis enzyme DHODH, which is associated with mitochondrial electron transport, has been linked to ROS production and apoptosis [49,52], but the roles of other genes in cell death are unknown. Treatment with ABT-263 increased ROS production as measured by Mito-SOX staining in JMR cells transfected with a control siRNA (Figure 3c). Silencing DHODH, HADHA or PDP1 suppressed such increases in ROS (Figure 3c), suggesting that these genes are indeed involved in ROS production during caspase-9-independent cell death. Increased ROS production may cause membrane lipid peroxidation [53,54]. We therefore measured the levels of membrane lipid peroxidation in JMR cells by staining with BODIPY 581/591 C11 [55]. We found that treatment of JMR cells with ABT-263 indeed induced lipid peroxidation as measured by BODIPY staining, while silencing of DHODH, HADHA or PDP1 inhibited ABT-263-induced lipid peroxidation (Figure 3c). These data suggest that several mitochondrial proteins, including DHODH, HADHA and PDP1, are involved in promoting ROS production and membrane lipid peroxidation during caspase-9-independent cell death.
Because oxidative stress appears to be important for the induction of caspase-9-independent cell death, genes with pro-survival functions may function by protecting against oxidative stress. Indeed, some candidate anti-cell death genes identified in our screening have anti-oxidative stress functions (Table S2), including microsomal glutathione transferase 1 (MGST1), sestrin 2 (SESN2) and oxidative resistance 1 (OXR1). MGST1, SESN2 and OXR1 all have been implicated in the protection against oxidative stress and cell death by regulating glutathione metabolism or ROS scavenging [56,57,58,59,60]. We found that silencing of these genes increased the sensitivity of JMR cells to ABT-263-induced cell death (Figure 4a). Treatment with ABT-263 increased ROS levels in JMR cells by Mito-SOX staining, while silencing of these genes further promoted ROS production (Figure 4b, upper panels). Treatment with ABT-263 induced lipid peroxidation as measured by BODIPY staining, while silencing of MGST1, SESN2 or OXR1 further increased membrane lipid peroxidation (Figure 4b, lower panels). These results indicate that MGST1, SESN2 and OXR1 may inhibit caspase-9-independent cell death by inhibiting ROS production and lipid peroxidation.
Whether the induction of lipid peroxidation might be important for caspase-9-independent cell death is not known. Alox5, a lipoxygenase that promotes membrane lipid peroxidation and leukotriene biosynthesis [61,62], was identified as a potential caspase-9-independent cell death gene in our screening (Table S1). Interestingly, Alox5 has been implicated to play an important role in mediating cell death signaling in apoptosis, pyroptosis and ferroptosis [63]. We found that silencing of Alox5 significantly suppressed ABT-263-induced cell death (Figure 5a). Alox5-activating protein (Alox5AP), a nuclear membrane protein that promotes the enzymatic activity of Alox5 [64], was also a positive hit in our screening (Table S1). Silencing of Alox5AP also suppressed ABT-263-induced cell death (Figure 5b), while silencing of Alox5 and Alox5AP together did not further inhibit cell death (Figure 5c). This indicates that Alox5AP functions in the same pathway as Alox5 to mediate cell death. Silencing of Alox5 or Alox5AP blocked the induction of lipid peroxidation in JMR cells (Figure 5d), indicating that Alox5 and Alox5AP are required for mediating lipid peroxidation during caspase-9-independent cell death in JMR cells. These data suggest that Alox5-dependent lipid peroxidation contributes to caspase-9-independent cell death.
It has been shown that phosphorylation of Alox5 at Serine 271 affects nuclear membrane localization of Alox5 [65,66]. Whether phosphorylation regulates the cell death function of Alox5 is unknown. Interestingly, treatment of JMR cells with ABT-263 for 2 h led to increased phosphorylation of Serine 271 residue in Alox5 (Figure 6a). To investigate whether Alox5 phosphorylation might be important for Alox5 to induce cell death, we substituted Serine 271 of Alox5 with Alanine (Alox5S271A) in an Alox5-GFP fusion construct. Interestingly, transfection with Alox5S271A-GFP significantly suppressed ABT-263-induced cell death in JMR cells, whereas wild-type Alox5-GFP had no such effect (Figure 6b). This suggests that expressing more wild-type Alox5 is not sufficient to induce cell death, while modification by phosphorylation is required for its cell death activity. Overexpressed Alox5 mutant may compete with endogenous wild-type Alox5 for interaction with signaling molecules in this cell death pathway and inhibit the transmission of cell death signaling. MAP kinases play an important role in the regulation of cell survival and cell death [67,68]. Our siRNA library screening identified several MAPK network members as candidate genes involved in caspase-9-independent cell death, including Erk1/MAPK3, MEKK3/MAP3K3 and ERK3/MAPK6 (Table S1). We therefore investigated whether MAPKs might be involved in Serine phosphorylation of Alox5. We first determined which MAPKs might be activated after the induction of caspase-9-independent cell death by measuring their phosphorylation. Among three major mammalian MAPK subfamilies: extracellular signal-regulated kinase (Erk), c-Jun NH2 terminal kinase (JNK) and p38 kinase, we detected activation of Erk1/2, but not JNK1, JNK2 or p38, as early as one hour after treatment with ABT-263 (Figure 6c). The activation of Erk1/2 preceded Alox5-Ser271 phosphorylation (Figure 6a,c), we therefore examined whether Erk1/2 could phosphorylate Alox5. We co-transfected Alox5-GFP with Erk1, Erk2, JNK1, JNK2 and p38. We found that co-transfection with Erk1, but not Erk2 or the other MAPKs, significantly increased Alox5 phosphorylation at Ser271 (Figure 6d). Silencing of Erk1 inhibited Alox5-Ser271 phosphorylation (Figure 6e). Moreover, silencing of Erk1 inhibited ABT-263-induced cell death (Figure 6f) and lipid peroxidation in JMR cells (Figure 6g). Consistently, an Erk1/2 inhibitor, SCH772984 [69], also potently inhibited the induction of cell death (Figure 6h) and lipid peroxidation (Figure 6i) in JMR cells treated with ABT-263. These results suggest that Erk1 is activated after mitochondrial disruption, leading to phosphorylation of Alox5 at Serine 271 and the induction of cell death. Alox5 has been shown to migrate from the nucleus or cytoplasm to the nuclear membrane after activation [70]. However, whether such translocation to the nuclear membrane takes place during caspase-9-independent cell death is not known. We observed that Alox5-GFP was localized to the nucleus in JMR cells (Figure 7a). Interestingly, Alox5-GFP showed perinuclear localization after treatment with ABT-263 (Figure 7a), suggesting that Alox5 is translocated to the nuclear membrane during caspase-9-independent cell death. In contrast, Alox5S271A-GFP did not show nuclear membrane translocation after treatment with ABT-263 (Figure 7a), indicating Ser271 phosphorylation is important for such translocation during the induction of cell death. We also found that silencing of Erk1 inhibited perinuclear translocation of Alox5-GFP after ABT-263 treatment (Figure 7b). Together, these data suggest that Erk1 is required for phosphorylation of Alox5 at Ser271, leading to nuclear membrane translocation of Alox5 and the induction of membrane lipid peroxidation.
In order to determine whether Alox5 is also important for the induction of caspase-9-independent cell death in vivo, we crossed T/casp9−/− mice with Alox5−/− mice. We did not observe T cell accumulation in T/casp9−/− or Alox5−/− mice compared to wild-type controls (Figure 8a). In contrast, T/casp9−/−Alox5−/− mice displayed significant increases in the percentages and total numbers of CD4+ and CD8+ T cells (Figure 8a), while the numbers of B cells and NK cells were not affected (Figure S2). This indicates that caspase-9- and Alox5-dependent mechanisms can function in parallel to maintain T cell homeostasis in vivo. However, when both pathways are deficient, T cell homeostasis is significantly disrupted. Consistently, wild-type, caspase-9−/− and Alox5−/− T cells treated with etoposide or staurosporine showed similar levels of cell death (Figure 8b). However, caspase-9−/−Alox5−/− T cells displayed impaired cell death (Figure 8b). Similarly, ABT-263-induced cell death in caspase-9−/−Alox5−/− T cells was significantly reduced compared to that in wild-type, caspase-9−/− and Alox5−/− T cells (Figure 8b). TUNEL staining shows that ABT-263-mediated DNA damage was inhibited in caspase-9−/−Alox5−/−, but not caspase-9−/− or Alox5−/− T cells (Figure 8c). Treatment with ABT-263 induced increased BODIPY staining in wild-type and caspase-9−/− T cells, but not in Alox5−/− or caspase-9−/−Alox5−/− T cells (Figure 8d). While lipid peroxidation was suppressed in both Alox5−/− and caspase-9−/−Alox5−/− T cells, only caspase-9−/−Alox5−/− T cells showed significant reduction in nuclear DNA damage by TUNEL staining (Figure 8c). In caspase-dependent apoptosis, nuclear translocation of DDF40/CAD after cleavage of DDF45/ICAD leads to DNA fragmentation and apoptosis [7,8,9]. Our data suggest that Alox5 induces lipid peroxidation to promote nuclear damage in a caspase-9-indepenent manner.
Caspase-9-independent cell death involved nuclear DNA damages as indicated by TUNEL staining (Figure 1c). Several DNA repair genes, including RAD51 recombinase (RAD51) and Polynucleotide Kinase 3′-Phosphatase (PNKP) [71,72], were identified as candidate cell survival genes against caspase-9-independent cell death (Table S2). We found that silencing of RAD51 or PNKP increased cell death in JMR cells treated with ABT-263 (Figure S4a). Silencing of Apoptosis Inhibitor 5 (API5), an anti-apoptotic gene that may suppress DNA damage [73,74], also increased cell death in JMR cells treated with ABT-263 (Figure S4a). Moreover, silencing of these genes promoted ABT-263-induced DNA damage in JMR cells as shown by TUNEL staining (Figure S4b). These results indicate that the induction of DNA damage is important for the execution of caspase-9-independent cell death. The protective functions of DNA repair enzymes suggest that DNA damage is important for the execution of caspase-9-independent cell death. In caspase-dependent apoptosis, nuclear translocation of DDF40/CAD after cleavage of DDF45/ICAD leads to DNA fragmentation and apoptosis [7,8,9]. Lipid peroxidation has been shown to promote nuclear translocation of AIF to induce cell death [54]. Endonuclease G (EndoG) and Apoptosis Inducing Factor (AIF) have been shown to enter the nucleus to induce caspase-independent cell death [12,13]. EndoG was among the positive hits of cell death genes in our screening (Table S1). Silencing of EndoG showed an inhibitory effect on caspase-9-independent cell death (Figure S5). In contrast, silencing of AIF did not suppress caspase-9-independent cell death in JMR cells (Figure S5). This could be due to the redundancy of AIF with other effector molecules as well as its homologs [75]. T-cell intracellular antigen-1 (TIA-1) was also a positive hit for cell death promoting genes in our screening (Table S1). TIA-1 is originally identified as a granule-associated RNA-binding protein in cytotoxic T cells that can induce DNA fragmentation in target cells [76]. It is widely expressed in different cell types of lymphoid and non-lymphoid tissues and has been shown to regulate mitochondrial dynamics, apoptosis, autophagy and cell proliferation [77,78]. Consistently, TIA-1 is down-regulated in a variety of human tumors, and its knockdown promotes tumor growth and invasion in mice [79]. We found that silencing of TIA-1 inhibited the killing of JMR cells by ABT-263 (Figure S5), indicating that TIA-1 is an effector molecule to induce caspase-9-independent cell death. We next examined whether double knockouts of caspase-9 and Alox5 affect nuclear import of cytolytic molecules during cell death. We observed nuclear translocation of EndoG after ABT-263 treatment in T cells from wild-type, T/caspase-9−/− or Alox5−/− mice (Figure 8a). In contrast, T cells from T/caspase-9−/−Alox5−/− mice did not display EndoG nuclear translocation after ABT-263 treatment (Figure 9a). Endogenous TIA-1 in primary mouse T cells showed distinct perinuclear localization (Figure 9b). After treatment with ABT-263, nuclear translocation of TIA-1 was observed in wild-type, caspase-9−/− and Alox5−/− T cells, but not in caspase-9−/−Alox5−/− T cells (Figure 9b). Nuclear translocation of AIF was also decreased in caspase-9−/−Alox5−/− T cells (Figure 9c). Together, these results suggest that Alox5 promotes nuclear entry of cell death-inducing molecules, such as EndoG and TIA-1, to cause DNA damage when caspase-9 is absent.
Through a genome-wide siRNA screening using a caspase-9-deficient Jurkat T cell line, we identified an Erk1-Alox5-mediated cell death pathway in parallel to caspase-9 signaling cascade. Several mitochondrial proteins, including DHODH, HADHA and PDP1, were found to promote ROS production after treatment with a Bcl-2/Bcl-xL inhibitor, ABT-263. Mitochondrial disruption led to ERK1-dependent phosphorylation of Alox5 at Serine 271 and promoted translocation of Alox5 to nuclear membranes. An Alox5 mutant with Serine 271 substitution dominantly interfered with caspase-9-independent cell death. Activation of Alox5 was critical for inducing nuclear translocation of nucleases and cytolytic molecules and the induction of DNA damage. While knockout of caspase-9 in T cells did not cause T cell expansion in mice, double knockouts of caspase-9 and Alox5 in T cells led to severe defects in cell death with significant T cell accumulation. This study suggests that mitochondrial disruption induces the activation of Alox5-dependent membrane lipid peroxidation and translocation of nucleases to the nucleus, resulting in caspase-9-independent cell death. Although the features of caspase-9-independent cell death were grossly similar to the classic caspase-dependent apoptosis with nuclear translocation of DNases and DNA damages, it employed distinct cell death machinery involving oxidative stress, MAPK activation and lipid peroxidation to promote nuclear entry of cell death molecules. Our genome-wide screening suggests that several sets of genes with opposite functions in regulating oxidative stress and DNA fragmentation also counteract each other in mediating caspase-9-independent cell death. Notably, genes promoting ROS production and oxidative stress, including PDP1, DHODH, HADHA, Alox5 and Alox5AP, were involved in enhancing caspase-9-independent cell death. In contrast, genes with anti-oxidative functions, including MGST1, SESN2 and OXR1, showed significant protection of cell survival. In the downstream, genes involved in DNA fragmentation, such as EndoG and TIA-1, promoted cell death, while DNA repair enzymes, RAD51 and PNKP, inhibited cell death. These data support an important role for oxidative stress and nuclear DNA fragmentation in the induction and execution of caspase-9-independent cell death. Activation of Erk1 leads to its translocation to the nucleus where it phosphorylates transcription factors and cell cycle regulators to promote cell survival and proliferation [80]. A death-promoting role for Erk independent of caspases has been reported [29,81,82]. Erk1/2 may promote EndoG-mediated cell death independent of caspases [83]. Interestingly, Erk1/2 activation has been associated with oxidative stress injury to neuronal cells in humans and mice [84]. However, the precise molecular mechanisms orchestrated by Erk to promote caspase-independent cell death are not entirely clear. Our data identified a novel mechanism for Erk1 in promoting cell death through the phosphorylation of Alox5. Phosphorylation of Alox5 plays an important in the regulation of its biological activities [63]. We found that mutation of the Ser271 residue in Alox5 impaired nuclear membrane localization of Alox5. This Alox5 mutant also dominantly interfered with caspase-9-independent cell death in JMR cells, suggesting that phosphorylation of Alox5 at Ser271 is critical for the cell death functions of Alox5. After ABT-263 treatment, we detected Erk activation and Alox5 phosphorylation within the first two hours and nuclear membrane translocation within eight hours. This suggests that the Erk1-Alox5 signaling cascade takes place rapidly to mediate caspase-independent cell death. It has been shown that the Serine 271 of Alox5 resides within a nuclear localization signal [65]. The abolishment of nuclear membrane localization by mutating Serine 271 or Erk1 silencing suggests that phosphorylation of this residue is critical for Alox5 translocation during caspase-9-independent cell death. ROS production and lipid peroxidation have been implicated in apoptosis, autophagic cell death and ferroptosis [85]. Alox5-dependent lipid peroxidation may be important for the execution of cell death triggered by different pathways, through promoting nuclear membrane permeability to facilitate the nuclear translocation of cell death proteins. In intrinsic apoptosis, caspase-9 induces the activation of effector caspases, which in turn cleaves DDF45/ICAD and releases DDF/CAD into the nucleus to cleave DNA [7,9]. Our data suggest that Alox5-mediated lipid peroxidation is important for the nuclear entry of cell death-inducing molecules, EndoG and TIA-1. It has been shown that nuclear entrance of EndoG can lead to cleavage of chromatin DNA independent of caspases [13]. TIA-1 is an RNA-binding protein that regulates pre-mRNA splicing and selective translational silencing [76,86,87,88]. TIA-1 purified from cell lysates or as a recombinant protein can directly induce DNA fragmentation when added to digitonin-permeabilized thymocytes [76], suggesting that it can promote DNA cleavage. TIA-1 can indeed bind to AT-rich DNA in vivo [89]. It has also been shown to be associated with endonuclease PRM1 in stress granules [90]. Whether TIA-1 might promote other nucleases in causing DNA damage remains to be investigated. Knockdown of TIA-1 increased tumor growth and invasion in mice, while TIA-1 expression is down-regulated in a variety of human tumor tissues [79]. Whether the potential tumor suppressor function of TIA-1 is associated with its cell death functions remains to be determined. A lack of T cell accumulation in T/caspase-9−/− or Alox5−/− mice suggests that caspase-9- and Alox5-dependent cell death mechanisms function in parallel of each other. T cell accumulation in T/caspase-9−/−Alox5−/− mice indicates that Alox5-dependent cell death is indeed critical for carrying out intrinsic cell death when caspase-9-dependent pathway is absent. Caspase-9-independent intrinsic cell death is mediated by distinct signaling molecules without the involvement of caspase-9-dependent caspase cascade. Although the early signaling events are different from caspase-dependent apoptosis, the late stages, including nuclear translocation of DNase and DNA damages, are similar. This suggests that cells undergoing such caspase-9-independent intrinsic cell death are efficiently cleared by phagocytosis without causing inflammation. The Erk1-Alox5-dependent peroxidation of nuclear membrane is a potentially important signaling pathway for the execution of cell death shared by apoptosis and non-apoptotic forms of cell death. Its relationship to apoptosis, necrosis and ferroptosis remains to be determined. This caspase-9-independent cell death mechanism may play an important role in the regulation of diverse biological and pathological processes [85]. Erk-dependent induction of caspase-independent cell death has been observed in neuronal cells [29]. Inhibition of Alox5 translocation has been shown to protect against cerebral ischemia/reperfusion injury [91]. Whether caspase-9-independent mechanisms play a prominent role in neuronal cell death and neurodegeneration will be interesting to investigate. In addition, the induction of caspase-independent cell death is a common feature to many chemotherapeutic drugs [30,31,32,33,34,35,36,37]. Characterization of genes involved in the caspase 9-independent cell death pathways will likely provide more specific and effective targets for cancer therapy. | true | true | true |
|
PMC9564236 | Daniel Ortuño-Sahagún,Julia Enterría-Rosales,Vanesa Izquierdo,Christian Griñán-Ferré,Mercè Pallàs,Celia González-Castillo | The Role of the miR-17-92 Cluster in Autophagy and Atherosclerosis Supports Its Link to Lysosomal Storage Diseases | 26-09-2022 | autophagy,cholesterol,enzyme deficiency,lysosomal storage diseases,metabolism,vesicle trafficking | Establishing the role of non-coding RNA (ncRNA), especially microRNAs (miRNAs), in the regulation of cell function constitutes a current research challenge. Two to six miRNAs can act in clusters; particularly, the miR-17-92 family, composed of miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1, and miR-92a is well-characterized. This cluster functions during embryonic development in cell differentiation, growth, development, and morphogenesis and is an established oncogenic cluster. However, its role in the regulation of cellular metabolism, mainly in lipid metabolism and autophagy, has received less attention. Here, we argue that the miR-17-92 cluster is highly relevant for these two processes, and thus, could be involved in the study of pathologies derived from lysosomal deficiencies. Lysosomes are related to both processes, as they control cholesterol flux and regulate autophagy. Accordingly, we compiled, analyzed, and discussed current evidence that highlights the cluster’s fundamental role in regulating cellular energetic metabolism (mainly lipid and cholesterol flux) and atherosclerosis, as well as its critical participation in autophagy regulation. Because these processes are closely related to lysosomes, we also provide experimental data from the literature to support our proposal that the miR-17-92 cluster could be involved in the pathogenesis and effects of lysosomal storage diseases (LSD). | The Role of the miR-17-92 Cluster in Autophagy and Atherosclerosis Supports Its Link to Lysosomal Storage Diseases
Establishing the role of non-coding RNA (ncRNA), especially microRNAs (miRNAs), in the regulation of cell function constitutes a current research challenge. Two to six miRNAs can act in clusters; particularly, the miR-17-92 family, composed of miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1, and miR-92a is well-characterized. This cluster functions during embryonic development in cell differentiation, growth, development, and morphogenesis and is an established oncogenic cluster. However, its role in the regulation of cellular metabolism, mainly in lipid metabolism and autophagy, has received less attention. Here, we argue that the miR-17-92 cluster is highly relevant for these two processes, and thus, could be involved in the study of pathologies derived from lysosomal deficiencies. Lysosomes are related to both processes, as they control cholesterol flux and regulate autophagy. Accordingly, we compiled, analyzed, and discussed current evidence that highlights the cluster’s fundamental role in regulating cellular energetic metabolism (mainly lipid and cholesterol flux) and atherosclerosis, as well as its critical participation in autophagy regulation. Because these processes are closely related to lysosomes, we also provide experimental data from the literature to support our proposal that the miR-17-92 cluster could be involved in the pathogenesis and effects of lysosomal storage diseases (LSD).
Establishing the role of non-coding RNA (ncRNA) in the regulation of cell function has been one of the new research challenges of the 21st century, with microRNAs, commonly abbreviated as miRNAs, constituting a target of great interest. The miRNAs are small ncRNA fragments (about 22 nucleotides in length) that regulate gene expression at the post-transcriptional level. They are known to modify gene function by binding to specific fragments of messenger RNA (mRNA) and suppressing their translation or promoting their degradation to ultimately repress gene expression [1]. To date, about 38,589 miRNAs have been introduced into the miRBase [2]. Although it has been described that most miRNAs can act by themselves, some of them act in clusters, composed of at least two miRNAs and up to six, which share a proximal location in the genomic DNA [3]. This is the case for the miR-17-92 cluster. It is composed of six miRNAs; miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1, and miR-92a (or miR-92-1) and is located on chromosome 13 in humans and on chromosome 14 in mice. This cluster of six miRNAs regulates certain functions during embryonic development, including cell differentiation, growth and development, and morphogenesis [4], both as a cluster and individually. Members of miR-17 family, which share the seed sequence AAACUG, are located on three different clusters: miR-17-92 (six members), miR-106b-25 (three members: miR-106b, miR-93, and miR-25), and miR-106a-363 (six members: miR-106a, miR-18b, miR-20b, miR-19b2, miR-92a-2, and miR-363). These clusters contain members of another three miRNA families: miR-18 family (seed sequence AAGGUG), mir-19 family (seed sequence GUGCAA), and mir-92 family (seed sequence AUUCGA) [5].
Plentiful works have reported that the miR-17-92 cluster, also known as oncomiR-1, and its members, are overexpressed in several types of cancers and constitute, to date, one of the best-characterized oncogenic miRNAs clusters [6]. In addition, it has been identified as a crucial factor involved in the regulation of vascular integrity and angiogenesis [7,8], which mainly regulates endothelial cell function under pathophysiological conditions [9]. Nevertheless, the targetome of miR-17-92 is large and variated [10]. Interestingly, this cluster of miRNAs has been implicated as a key factor in the regulation of metabolic reprogramming of tumors. When its expression is absent, cell metabolism decreases, and when it is overexpressed, nutrient utilization by tumor cells increases [11], thus regulating glycolytic and mitochondrial metabolism. Therefore, the members of this cluster were also recently considered mitomiRs, or miRNAs present in mitochondria [12]. Additionally, the miR-17-92 cluster has been recognized as a specific biomarker in gestational diabetes mellitus (GDM) [13,14,15], which increases the interest in the possible participation of this cluster in regulating cellular energetic metabolism. The regulation of lipid metabolism constitutes a relevant part of metabolism and homeostasis, not only for energy production but also for lipid neogenesis and cholesterol metabolism. Interestingly, a few studies have proposed a relationship between the miR-17-92 family and lipid metabolism regulation. This family has been identified as actively participating in dyslipidemia pathophysiology in patients with coronary artery disease (CAD) [16], as well as in the process of steatosis in hepatic cells, specifically through miR-17, which regulates the expression of CYP7A1 (cytochrome P450 family 7 subfamily A member 1) and is a regulator of hepatic lipid metabolism [17]. However, the findings available so far only suggest a potential role of miR-17-92 cluster in the regulation of cellular lipid metabolism. Although individual members of the miR-17-92 family modulates lipid metabolism, whether the miR-17-92 cluster functions as a whole remains largely unexplored. Moreover, it has been described that dysregulation in the expression of lysosome-related genes may contribute to the incapacity of lysosomes to deal with high lipid consumption and the later development of atherosclerosis, which is caused by cholesterol accumulation [18]. In this case, miRNAs regulate atherosclerotic disease initiation and progression [19], which makes them suitable candidates as biomarkers for atherosclerosis as well as targets for pharmacological intervention as a hopeful therapeutic notion. Therefore, an alteration in the regulation of lipid metabolism can lead to dysfunction of the lysosomes and, subsequently, generation of atherosclerosis. Given that in addition to controlling cholesterol efflux, lysosomes also regulate autophagy, it is logical to suppose that a possible link between these two cellular processes exists. This link may be represented, at least partially, by the coordinating regulatory action of miR-17-92 family members on gene expression and cell function. Experimental evidence from the literature supporting this hypothesis is presented below, first for atherosclerosis and then for autophagy.
The plasma expression level of miR-17-5p is increased in patients with CAD, while, interestingly, VLDLR (very low-density lipoprotein receptor) mRNA expression is decreased in peripheral blood lymphocytes in these patients [20]. Another target repressed by miR-17-5p is ABCA1 (adenosine triphosphate (ATP)-binding cassette subfamily A member 1) [21], whose expression is the rate-limiting step in reverse cholesterol transport [22], which is considered the only mechanism by which the human body can clear excess cholesterol [23]. On the other hand, miR-17-5p can be downregulated by a hypoxia-induced long non-coding RNA (lncRNA) named MALAT1 (metastasis-associated lung adenocarcinoma transcript 1) to regulate the expression level of ABCA1 and then reduce cholesterol accumulation in oxidized-LDL (low-density lipoprotein)-induced macrophages. Therefore, the knockdown of MALAT1 may promote cholesterol accumulation through the miR-17-5p/ABCA1 axis [24]. Similarly, the knockdown of miR-17-5p can attenuate atherosclerosis [25] by suppressing inflammation and reducing lipid accumulation in atherosclerotic lesions [21]. Likewise, downregulation of miR-17-5p could be conducted by p53-dependent lincRNA-p21 expression, which consequently protects against atherosclerosis progression via SIRT7 (sirtuin 7) elevation, which is one of miR-17-5p’s targets [26]. Another member of the miR-17-92 cluster, miR-20a/b, together with other miRNAs, has been described as a regulator of ABCA1 expression: miR-20a/b decreases ABCA1 expression, thus increasing cholesterol accumulation in vitro. By inhibiting miR-20a/b, ABCA1 expression and cholesterol flux increase [27]; miR-20a also targets the PTEN (phosphatase and tensin homolog) gene, participating in the prevention of CAD by promoting the survival and proliferation of vein endothelial cells. Likewise, overexpression of miR-20a could down- or upregulate the expression levels of several atherosclerosis-related genes in CAD [28]. Furthermore, miR-20a was predicted to regulate myogenic factor 5 (MYF5) and proliferator-activated receptor-γ (PPARγ), two of the most important genes involved in adipogenesis in both mice and humans [29]. Another target of miR-20a is LDLR (low-density lipoprotein receptor) [30], involved in cholesterol homeostasis. Additionally, miR-20a-5p, which targets almost 500 genes [31], has been involved in non-alcoholic fatty liver disease (NAFLD) and in its advanced version, non-alcoholic steatohepatitis (NASH), which is characterized by steatosis, inflammation, ballooning, and fibrosis [32]. Downregulation of miR-20a-5p is accompanied by upregulation of one of its targets, CD36 (CD36 molecule), and increased lipid deposition, suggesting novel pathogenesis of non-alcoholic fatty liver disease and a potential therapeutic strategy for metabolic diseases [33]. A third member of the cluster, miR-92a, has been involved in lipid metabolism in hypoxic rats, which is relevant because hypoxia is known to induce lipolysis and inhibit fat synthesis [34]. Under these conditions, miR-92a expression levels decrease significantly in hypoxic rats compared with normoxic rats, suggesting that miR-92a deficiency could suppress lipolysis by regulating Fzd10/Wnt/β-catenin signaling [35]. Likewise, miR-19a-5p has recently been identified as a regulator of lipid metabolism in tilapia fish by participating in triglyceride synthesis, mainly because one of its target genes is the 3′UTR region of DGAT2 (diacylglycerol O-acyltransferase 2) [36]. In addition to the role of the miR-17-92 cluster in lipid metabolism, a second related miRNA cluster has been also involved in metabolic pathophysiology. Deletion of the miR-106b-25 microRNA cluster (which contains miR-106b, miR-93, and miR-25) attenuates atherosclerosis in Apoe (apolipoprotein E) knockout mice, presumably by regulating plasma cholesterol levels [37]. Moreover, two of its three members, miR-106b and miR-93, were shown to impair cholesterol efflux [38,39]. Similarly, miRNAs from the miR-17 and miR-19 families are involved in lipid and cholesterol metabolism, which is mostly performed in subcellular organelles such as mitochondria, peroxisomes, and lysosomes.
Besides enzyme-mediated lipid digestion and cholesterol metabolism, other metabolic lipid-related processes, such as vesicle trafficking and autophagy, are also essential for normal cell functioning. Vesicle trafficking is a means of intracellular transport; however, it is also implicated extracellularly, especially through exosomes. It is relevant to mention that miRNAs from the miR-17-92 cluster have been located within exosomes [40,41,42]. This suggests that this miRNA cluster is involved in cell-to-cell communication through exosomes, and therefore, in paracrine regulation from one cell to another. Autophagy is a basic and conserved cellular mechanism involved in the intracellular degradation of proteins and organelles, or pathogens, through the formation of autophagosomes [43], and it constitutes an extremely relevant metabolic process that needs to be highly regulated. As will be explained below, all members of the miR-17-92 cluster have been separately implicated in the regulation of autophagy in different situations. Paradoxically, there are virtually no studies involving all of them in the regulation of this central cellular phenomenon, probably because their role in autophagy has only recently been described. The most-studied member of this cluster concerning autophagy is miR-17-5p [44]. Initially, Comincini et al. [45] identified miR-17-5p as a modulator of different autophagy-related proteins (ATGs), demonstrating that anti-miR-17-5p administration results in an increase in MAP1LC3B (microtubule-associated protein 1 light chain 3 beta) and ATG7 (autophagy-related 7) protein expression, and subsequently, yields an activation of autophagy through autophagosome formation in glioblastoma T98G cells. ATG7 is one of the master regulators of the autophagy process, controlling autophagosome formation and vesicle progression [46], and is a direct target of miR-17-5p. Also, it has been recently demonstrated that miR-17 is downregulated in chemoresistant non-small-cell lung cancer (NSCLC) cells. In these same cells, a lncRNA, BLACAT1 (BLACAT1 overlapping LEMD1 locus), was shown to be significantly upregulated, together with ATG7, ABCC1 (ATP binding cassette subfamily C member 1), LC3-I/II, and BECN1 (beclin 1). Then, it was demonstrated that BLACAT1 targets miR-17 and negatively regulates it, and thus promotes ATG7 expression, which suggests that autophagy may be a novel target for overcoming drug resistance [47]. Additionally, two proteins relevant to autophagy and the autophagosome, ULK1 (Unc-51 like autophagy activating kinase 1) and LC3I/II, have been also identified as targets of miR-17-5p, being capable of down-regulating its expression in macrophage RAW264.7 cells in response to mycobacterial infection and therefore modulating phagosomal maturation [48]. Likewise, the levels of the anti-apoptotic myeloid cell leukemia 1 (MCL1), an apoptosis regulator member of the BCL2 family, and its transcriptional activator STAT3 (signal transducer and activator of transcription 3) are down-regulated by miR-17-5p (MCL1 suppresses autophagy through its ability to sequester BECN1), and overexpression of miR-17-5p also inhibits the phosphorylation of PRKCD (protein kinase C delta). Therefore, the miR-17-PRKCD-STAT3-MCL1 pathway emerges as a key regulating axis of autophagy during M. tuberculosis infection [49]. Further evidence extends these results to hepatic ischemia/reperfusion injury (IRI), in which high expression of miR-17-5p upregulates autophagy to promote hepatic IRI through the suppression of STAT3 expression [50]. Similarly, miR-17-5p overexpression decreases the transcription of STAT3 in vascular smooth muscle cells subjected to hypoxia-induced autophagy. Also, miR-17-5p could directly target BECN1, which mediates irradiation-induced autophagy activation in a glioma cell line [51]. Furthermore, miR-17 inhibition promotes cisplatin-induced autophagy of tongue squamous cell carcinoma CAL-27 cells through the STAT3 pathway [52]. Moreover, miR-17-5p is also able to suppress autophagy in an osteoarthritis (OA) mice model, in which decreased miR-17-5p expression induces autophagy mainly through suppressing the expression of another of its targets, SQSTM1/p62 (sequestosome 1), thereby contributing to osteoarthritis progression [53]. Interestingly, miR-17-5p has also been recently involved in osteosarcoma pathophysiology. It is upregulated in osteosarcoma cell lines and induces autophagy by targeting PTEN [54]. In addition, it has been recently demonstrated that miR-17-5p targets Mfn2 (Mitofusin 2), a mitochondrial fusion protein that plays a role in balancing autophagy and inhibits its expression, activating the PI3K/AKT/mTOR pathway and suppressing autophagy to promote cardiac hypertrophy [55]. Again, miR-17-5p downregulation inhibits autophagy and myocardial remodeling after myocardial infarction by targeting the STAT3 pathway [56]. It is relevant to mention that inhibition of MTOR (mechanistic target of rapamycin kinase) by miR-17-5p upregulates autophagy and slows down the aging process [57]. Consequently, miR-17-5p may be a novel central miRNA regulator in stress-induced cellular mechanisms, considering stressful situations like pathogen infections, hypoxia, irradiation, inflammation, hypertrophy, or cancer chemotherapy and chemoresistance, reacting by mediating autophagy responses that in turn appear to be regulated by miR-17-5p. A decade ago, miR-20a was also first appointed as an autophagy-related gene because of its ability to negatively regulate autophagy in C2C12 myoblasts [58]. Since then, cumulative evidence has been shown to support this idea. Under hypoxia, HIF1A (hypoxia-inducible factor 1 subunit alpha) suppresses miR-20a, which negatively regulates ATG16L1 (autophagy-related 16 like 1), an autophagy-related gene, solidifying the HIF1A-miRNA-20a-ATG16L1 regulatory axis as a critical mechanism for hypoxia-induced autophagy in osteoclast differentiation [59]. Additionally, in U-251 glioma cells, EMAPII (low-dose endothelial-monocyte-activating polypeptide-II) also downregulates miR-20a and induces autophagy by subsequently increasing the expression of ATG5 (autophagy-related 5) and ATG7, both of which are miR-20a targets [60]. miR-20a also directly targets and inhibits ATG7 and TIMP2 (TIMP metallopeptidase inhibitor 2) in SiHa cells [61]. Depletion of miR-20a suppresses proliferation and autophagy and promoted apoptosis by increasing the expression of one of its targets, THBS2 (thrombospondin 2), in cervical cancer cells [62]. Interestingly, resveratrol induces cell autophagy and decreases the inhibitory effect of miR-20a on another of its targets, PTEN, thus activating the PTEN/PI3K/AKT signaling pathway to attenuate liver fibrosis [63]. On the other hand, overexpression of miR-20a, induced by mycobacterial infection, inhibits the autophagy process in macrophages by targeting ATG7 and ATG16L1 and suppressing their expression [64]; miR-20a is also overexpressed in breast cancer, where it targets several genes related to autophagy, such as BECN1, ATG16L1, and SQSTM1/p62, downregulating them. Consequently, miR-20a inhibits autophagic flux and lysosomal proteolytic activity. If endogenous miR-20a is blocked, autophagic flux is increased [65]. miR-20a-5p gain-of-function also inhibits autophagy in ovary cancer via DNMT3B (DNA methyltransferase 3 beta)-mediated DNA methylation of RBP1 (retinol-binding protein 1) [66]. Therefore, lowering miR-20a expression activates autophagy flux by upregulating the expression of autophagy-related proteins, and contrarily, overexpression of miR-20a inhibits autophagy and lysosomal proteolytic activity by downregulating them. The first member of the miR-17-92 cluster that was involved in autophagy regulation was miR-18a. Its ectopic overexpression in HCT116 colon cancer cells promotes autophagy by upregulating ATM (ataxia telangiectasia mutated) gene expression, a Ser/Thr protein kinase, and a member of the PI3K (phosphoinositide 3-kinase)-related protein kinase (PIKK) family, and by inhibiting mTORC1 (mechanistic target of rapamycin complex 1) activity [67]. Additionally, miR-18a-5p downregulates apoptosis and upregulates resveratrol-induced autophagy in the kidney podocytes of db/db mice (diabetic mice), also through targeting of the Atm gene [68]. Similarly, miR-18a induces the degradation of the oncogenic protein hnRNPA1 (heterogeneous nuclear ribonucleoprotein A1) by forming a complex that is then degraded through the autolysosomal pathway, enhancing the autophagy pathway itself [69]. Furthermore, miR-18a-5p promotes autophagy in NSCLC, enhancing not only autophagosome formation but autophagy flux [70]. Similarly, downregulation of miR-18a promotes autophagy by upregulating BDNF (brain-derived neurotrophic factor) expression and by inactivating the AKT/mTOR axis in hypoxia-induced rat cardiomyocytes, modeled through the H9c2 cell line. Likewise, the upregulation of BDNF suppresses cell senescence through the downregulation of miR-18a [71]. A less-studied microRNA of the miR-17-92 cluster concerning autophagy is miR-19a. Its overexpression in cardiomyocytes ameliorates hypoxia-induced cell death by switching from apoptosis to autophagy through downregulation of its specific target BCL2L11 (BCL-2-like 11), an apoptotic activator [72]. In addition, reduced expression of miR-19a by propofol impedes autophagy and apoptosis caused by glutamate in PC12 cells, through an activation of AMPK (AMP-activated protein kinase) and mTOR signaling pathways [73]. Finally, and even less studied, miR-92a-3p has also been involved in endothelial cell autophagy and cardiomyocyte metabolism. In vitro, its inhibition promotes autophagy through the expression of ATG4A (autophagy-related 4 cysteine peptidase) [74]. Interestingly, the lncRNA MALAT1, acting as a sponge in cardiac cells induced to senescence, was identified as an exosomal transferred RNA that represses miR-92a-3p expression to unblock ATG4A [75]. Taken together, all this experimental evidence strongly suggests the direct involvement of the miR-17-92 cluster in the regulation of autophagy. To move forward, we must emphasize here that both autophagy and intracellular metabolism are functions mainly performed by lysosomes, and consequently, regulation of these functions must be connected, somehow, with the regulation of lysosomal functioning, and vice versa, as we will discuss below.
Many miRNAs are subject to various regulatory mechanisms, and one of them involves their inhibition through the binding of lncRNAs. These lncRNAs interact with miRNAs in a functional network that affects several processes. Only a few lncRNA are well-conserved among species, one of them being MALAT1, which blocks numerous miRNAs by providing non-functional binding sites and plays a crucial role in atherosclerosis [76,77,78] and autophagy [79,80]. Although to date there are no works that investigate the relationship between MALAT1 and the miR-17-92 cluster, it is noteworthy that there are six works that correlate this lncRNA with two of the members of the cluster: miR-17-5p and miR-92a. Moving on to miR-17-5p, it is interesting to note that diabetic patients who smoke present higher serum levels of MALAT1 and lower miR-17 levels in comparison to the serum of nonsmokers. This could be explained by the fact that cigarette smoke extract inhibits insulin production by upregulating TXNIP (thioredoxin interacting protein) via MALAT1-mediated downregulation of miR-17 [81]. Also, it has been reported that MALAT1 expression increases in HeLa and CaSki cells treated with Cas-II-gly, together with a suppression of Wnt (Wingless-related integration site) signaling. Hence, MALAT1 inhibits FZD2 (frizzled class receptor 2) expression by targeting miR-17-5p via inactivation of the Wnt signaling pathway [82]. Finally, and more interesting for our reviewed topic, as we mentioned before, MALAT1 regulates cholesterol accumulation via the microRNA-17-5p-ABCA1 axis [24]. On the other hand, miR-92a can also be regulated by MALAT1. In human coronary artery endothelial cell (HCAEC)-derived exosomes under hyperbaric oxygen conditions, MALAT1 is overexpressed, suppressing miR-92a expression, which in turn unblocks KLF2 (Kruppel like factor 2) to enhance angiogenesis [83]. These results have been replicated in a rat model of cardiac infarction [84]. Additionally, as previously commented, MALAT1 sponges miR-92a-3p expression to inhibit cardiac senescence by targeting ATG4A [75]. Considering that MALAT1 has been largely associated with both atherosclerosis and autophagy, that the miR-17-92 cluster is greatly involved in these two processes, and that its expression is directly regulated by MALAT1, analyzing this relationship in depth may yield interesting results.
Cellular metabolism is highly compartmentalized within each cell, made possible by the endomembrane system. Although the mitochondrion is essential to produce energy from metabolic sources (both the Krebs cycle and the beta-oxidation of fatty acids take place here), other organelles, such as peroxisomes and lysosomes, are also fundamental for maintaining cellular metabolism. Besides cellular detoxification, peroxisomes also perform beta-oxidation of fatty acids, and lysosomes constitute the main intracellular digestive system of the cell. Consequently, it is reasonable to deduce that general pathways of metabolic regulation may affect the function of more than one organelle. Therefore, if the miR-17-92 regulatory cluster can perturb genes related to mitochondrial metabolic function, it could be also related, in some way, to genes involved in lysosomal metabolic function. Lysosomes are intracellular organelles that, in form of small vesicles, participate in several cellular functions, mainly digestion, but also vesicle trafficking, autophagy, nutrient sensing, cellular growth, signaling [85], and even enzyme secretion. The membrane-bound lysosome has long been regarded as the waste management and recycling facility of the cell because of its many hydrolytic enzymes used to digest various biomolecules. For these digestive enzymes to work properly, the pH inside the organelle must be highly acidic in comparison to the neutral cytoplasm [86]. More recently, the lysosome has been proven to be a dynamic key signaling hub involved in various pathways related to cellular adaptation, immunity, metabolism, and intracellular communication. Cells have between 50 and 1000 lysosomes in their cytoplasm, and the importance of these structures in cell homeostasis is highlighted by the severity of phenotypes present in LSDs [87].
LSDs are a heterogeneous group of conditions brought about by congenital inborn errors of metabolism (IEM), characterized by lysosomal dysfunction that led to the pathogenic accumulation of diverse molecules. Although the underlying basic mechanism behind the over 70 LSDs known to date is mostly the same, the accumulated or stored substance and the clinical manifestations vary widely among particular conditions [88]. Most LSD are classified as rare or ultra-rare conditions, but when put together, they have a joint prevalence of 1 in 5000 to 5500 live births [88,89] and belong to the even larger group of inborn errors of metabolism with a strong genetic component. These disorders are all monogenic and mostly have an autosomal recessive pattern of inheritance except for a few X-linked diseases, such as Fabry disease and Hunter syndrome [88]. As was mentioned before, although lysosomal dysfunction has a central role in all these diseases, the mechanisms that cause this to happen, as well as the pathways that lead to cell death, differ widely among them [90]. The specific organ damage involved in each of these conditions is different, but most of them seem to have a strong effect on the central nervous system, with many of them deemed neurodegenerative, which indicates a special susceptibility to lysosomal dysfunction in neurons and other nervous system cells [88,90]. Being mainly a genetic disease, epigenetic factors must also play relevant roles in LSD. Some of these epigenetic regulatory elements are miRNAs that function as regulatory molecules, which have been linked to several physiological processes as well as numerous diseases, among which, their role in LSD pathology is starting to be acknowledged [91,92]. In addition, miRNAs are recognized as novel regulators of cholesterol biosynthesis and metabolism [93], which are tightly controlled by the expression and proteolytic activation of the sterol regulatory element-binding proteins (SREBPs). An intricate network of miRNAs regulates gene expression of certain key genes, such as APP (amyloid-beta precursor protein), which appears to be regulated by multiple miRNAs, some of them being members of the miR-17 family (such as miR-20a-5p, miR-106a/b-5p) [93]. However, there is a lack of research on the role of ncRNA in lysosome functioning and particularly in LSD [92]. We must also consider, regarding miRNAs, that not only those produced within the cell are relevant to define the regulation of which they are capable, as exosomal miRNAs have an important role in cell signaling. Consequently, it is necessary to keep the possibility of the role of exosomes in LSD in mind [94].
Although miRNAs are potentially vital in the pathogenic mechanisms underlying LSDs, much information remains to be discovered. Very little work has focused on differentially expressed miRNAs in LSDs, and even less on Niemann–Picks type C (NPC) disease. The first correlation between LSD and miRNAs was most probably a paper, published in 2010, by Ozsait and colleagues, which identified variations in the concentration of several miRNAs involved in lipid metabolism and molecular transport in NPC fibroblasts. Although only a small percentage of the upregulated miRNAs were specifically related to cholesterol and lipid metabolism, they found some members of the miR-17-92 cluster (miR-19a, miR-19b, and a related one, miR-106b) to be among the most downregulated miRNAs [95]. All of them are linked to the lipid and partly glycosphingolipid metabolic processes, but miR-19a and miR-19b, in particular, are related to cholesterol transport and cholesterol-related metabolic processes; this, together with cellular cholesterol accumulation, constitutes one of the most prominent phenotypes in these cells [95]. More recently, Niculescu et al. evaluated the potential of miR-92a in the reversal of hyperlipidemia in hamsters. They discovered that miR-92a targets ABCG4 (ATP-binding cassette G4) and NPC1 (NPC intracellular cholesterol transporter 1) proteins, and that anti-miR-92a restored ABCG4, NPC1, and SOAT2 (sterol O-acyltransferase 2) expression. Interestingly, they proposed that in vivo inhibition of miR-92a could be a potential approach to correct lipid metabolism dysregulation and even atherosclerosis [96]. Some other cumulative evidence independently links miRNAs from the miR-17-92 cluster with different LSDs. In the case of Gaucher disease (GD), caused by deficiency of GBA1 (glucocerebrosidase 1) enzyme [97], it has been reported that miR-19a-5p is one of the three miRNAs that strongly down-regulate SCARB2 (scavenger receptor class B member 2) expression, which is an important membrane receptor involved in GBA1 availability [98]. Also, protein Lrrk2 (leucine-rich repeat kinase 2), involved in the mitochondrial dysfunction pathway in GD is regulated by miR-19b-3p [99]. Another LSD is Mucopolysaccharidosis type I (MPS I), caused by deficiency of IDUA (alpha-L-iduronidase), which leads to the ubiquitous accumulation of two glycosaminoglycans (GAGs), dermatan, and heparan sulfates [100,101]. Pereira et al. examined gene expression of Neu1 (neuraminidase 1) and Ctsa (cathepsin A), two components of the lysosomal multienzyme complex (LMC) in the cerebellum of MPS I mice and controls. They found that miRNAs from the miR-17 family (miR-17, miR-20a, miR-20b, miR-93, miR-106a, and miR-106b) were predicted to bind them, thus suggesting that this family of miRNAs might play a role in the regulation of lysosomal multienzyme complex (LMC) gene expression [102]. Furthermore, studies focused on Fabry disease also found miRNAs from the miR-17-92 family to be involved in its pathology. This is the case for miR19a-3p, involved in TGF (transforming growth factor)-beta signaling pathways, which is significantly down-regulated in Fabry disease male patients with enzyme replacement therapy (ERT) [103]. However, in the case of Fabry disease, many other miRNAs have been involved [104]. Other LSDs also present alterations in miRNAs from the miR-17-92 family. This is the case for GM2-Gangliosidosis deficiencies, such as Tay–Sachs and Sandho diseases, in which a panel of nine miRNAs that included miR-19a have been identified as highly downregulated [105]. Regarding probable interventions and the use of miRNA therapies to alleviate the consequences of LSD, it is of interest to analyze the few results obtained in this approach. First, one miRNA that has been described as linked to LSD and neurodegenerative disorders and is not directly related to the miR-17-92 cluster is miR-155, which is considered an inflammatory master regulator. Nonetheless, ablation of the pro-miR-155 does not mitigate neuroinflammation or neurodegeneration in a vertebrate model of GD [106] and does not affect the neuroinflammatory trajectory in an infantile neuronal ceroid lipofuscinosis (INCL) mouse model, also known as CLN1-disease, a devastating neurodegenerative LSD [107]. Contrastingly, in 2010, Gentner et al. demonstrated that therapy with transplanted hematopoietic stem and progenitor cells (HSPCs) is useful in the treatment of LSD. Interestingly, they reported that members of the miR-17-92 cluster (miR-19, miR-93a, and miR-17-5p) were highly expressed in hematopoietic stem and progenitor cells [108] which could indicate that those precursors exert a modulatory effect to improve the metabolism in affected organisms, compensating the damage.
Niemann–Picks type C (NPC) disease is an LSD characterized by the pathogenic accumulation of unesterified cholesterol and other lipids due to mutations in the genes coding for the intracellular cholesterol transport proteins NPC1 (95% of cases) or NPC2 (NPC intracellular cholesterol transporter 2, 5% of cases), that leads to a progressive neurovisceral condition [109]. The disease prevalence is currently estimated to be 0.95 per 1 million people in the United States, but the diagnostic difficulty of atypical cases might make the real prevalence higher [110]. The genetic and hereditary mechanisms behind this disease and other LSDs are still to be completely elucidated, as most cases are compound heterozygotes and notable phenotypical differences have been found among siblings and twins with identical genetic variations. Because of this and the unusual genetics behind NPC, the role of miRNAs and other epigenetic mechanisms in its pathogenesis is particularly interesting [111]. Under normal conditions, the protein products of NPC1 and NPC2 are thought to work in unison to export cholesterol and other molecules from the lysosome to other organelles, such as the mitochondria. NPC2 collects cholesterol from the lysosomal lumen and transports it to the lysosomal-membrane-bound NPC1, which then translocates it to the exterior [112]. It is relevant to note that, during neuronal aging, the activation of the AKT-mTOR pathway triggers the degradation of NPC1 protein, which induces the accumulation of cholesterol in endosomal compartments [113], similarly to what occurs in NPC1 mutant and Niemann–Picks disease. Because the AKT/mTOR pathway can be upregulated, at least indirectly, by different members of the miR-17-92 cluster, such as miR17-5p, miR-20a, and miR18a (see previous mentions of this), it is logical to suppose that there is a strong relationship between this miRNA cluster and the metabolic consequences of Niemann–Picks disease.
Although the role of the miR-17-92 cluster during embryonic development in cell differentiation, growth, and morphogenesis, as well as in oncogenesis, has been well-established, its role in cell metabolism, mainly in lipid and cholesterol flux under pathological conditions such as atherosclerosis and in autophagy as a cellular response to different situations is not established. Here, we present comprehensive up-to-date experimental evidence that supports the fundamental role of the miR-17-92 cluster in regulating cellular energetic metabolism, mainly lipid and cholesterol flux and as a key regulator in atherosclerosis, as well as a critical participant in regulating autophagy. Because these cellular functions are closely related to lysosomes, we also propose that the miR-17-92 cluster would be somehow involved in LSD effects. A summary of the molecular network in which the miR.17-92 cluster is involved is presented in Figure 1. Alterations in the transport mechanism driven by NPC1 and NPC2 cause the abnormal storage of cholesterol and other macromolecules in lysosomes and late endosomes. This has many consequences, such as a general slowing down of the endocytic process, which prevents the binding of cholesterol vesicles to endosomes, disruption of autophagy, deregulation of proteins (due to lack of cholesterol in the endoplasmic reticulum and Golgi apparatus), and mitochondrial damage, which then leads to eventual cell death [112,114]. Accordingly, alterations that affect the lysosomal-mitochondria relationship and their metabolic equilibrium generate a defective metabolism, which contributes to disease progression [115]. Consequently, the identification of regulatory molecular links between these two organelles will most probably cause the rise of novel targets for the treatment of NPC. Therefore, we propose that members of the miRNA-17-92 cluster could be relevant actors in the clinical consequences of LSD, and therefore, could be considered pharmacological targets to, at least partially, alleviate this pathological condition. | true | true | true |
|
PMC9564368 | Islam M. Saadeldin,Bereket Molla Tanga,Seonggyu Bang,Chaerim Seo,Okjae Koo,Sung Ho Yun,Seung Il Kim,Sanghoon Lee,Jongki Cho | ROCK Inhibitor (Y-27632) Abolishes the Negative Impacts of miR-155 in the Endometrium-Derived Extracellular Vesicles and Supports Embryo Attachment | 10-10-2022 | miR-155,CRISPR/Cas9,extracellular vesicles,embryo,ROCK inhibitor | Extracellular vesicles (EVs) are nanosized vesicles that act as snapshots of cellular components and mediate cellular communications, but they may contain cargo contents with undesired effects. We developed a model to improve the effects of endometrium-derived EVs (Endo-EVs) on the porcine embryo attachment in feeder-free culture conditions. Endo-EVs cargo contents were analyzed using conventional and real-time PCR for micro-RNAs, messenger RNAs, and proteomics. Porcine embryos were generated by parthenogenetic electric activation in feeder-free culture conditions supplemented with or without Endo-EVs. The cellular uptake of Endo-EVs was confirmed using the lipophilic dye PKH26. Endo-EVs cargo contained miR-100, miR-132, and miR-155, together with the mRNAs of porcine endogenous retrovirus (PERV) and β-catenin. Targeting PERV with CRISPR/Cas9 resulted in reduced expression of PERV mRNA transcripts and increased miR-155 in the Endo-EVs, and supplementing these in embryos reduced embryo attachment. Supplementing the medium containing Endo-EVs with miR-155 inhibitor significantly improved the embryo attachment with a few outgrowths, while supplementing with Rho-kinase inhibitor (RI, Y-27632) dramatically improved both embryo attachment and outgrowths. Moreover, the expression of miR-100, miR-132, and the mRNA transcripts of BCL2, zinc finger E-box-binding homeobox 1, β-catenin, interferon-γ, protein tyrosine phosphatase non-receptor type 1, PERV, and cyclin-dependent kinase 2 were all increased in embryos supplemented with Endo-EVs + RI compared to those in the control group. Endo-EVs + RI reduced apoptosis and increased the expression of OCT4 and CDX2 and the cell number of embryonic outgrowths. We examined the individual and combined effects of RI compared to those of the miR-155 mimic and found that RI can alleviate the negative effects of the miR-155 mimic on embryo attachment and outgrowths. EVs can improve embryo attachment and the unwanted effects of the de trop cargo contents (miR-155) can be alleviated through anti-apoptotic molecules such as the ROCK inhibitor. | ROCK Inhibitor (Y-27632) Abolishes the Negative Impacts of miR-155 in the Endometrium-Derived Extracellular Vesicles and Supports Embryo Attachment
Extracellular vesicles (EVs) are nanosized vesicles that act as snapshots of cellular components and mediate cellular communications, but they may contain cargo contents with undesired effects. We developed a model to improve the effects of endometrium-derived EVs (Endo-EVs) on the porcine embryo attachment in feeder-free culture conditions. Endo-EVs cargo contents were analyzed using conventional and real-time PCR for micro-RNAs, messenger RNAs, and proteomics. Porcine embryos were generated by parthenogenetic electric activation in feeder-free culture conditions supplemented with or without Endo-EVs. The cellular uptake of Endo-EVs was confirmed using the lipophilic dye PKH26. Endo-EVs cargo contained miR-100, miR-132, and miR-155, together with the mRNAs of porcine endogenous retrovirus (PERV) and β-catenin. Targeting PERV with CRISPR/Cas9 resulted in reduced expression of PERV mRNA transcripts and increased miR-155 in the Endo-EVs, and supplementing these in embryos reduced embryo attachment. Supplementing the medium containing Endo-EVs with miR-155 inhibitor significantly improved the embryo attachment with a few outgrowths, while supplementing with Rho-kinase inhibitor (RI, Y-27632) dramatically improved both embryo attachment and outgrowths. Moreover, the expression of miR-100, miR-132, and the mRNA transcripts of BCL2, zinc finger E-box-binding homeobox 1, β-catenin, interferon-γ, protein tyrosine phosphatase non-receptor type 1, PERV, and cyclin-dependent kinase 2 were all increased in embryos supplemented with Endo-EVs + RI compared to those in the control group. Endo-EVs + RI reduced apoptosis and increased the expression of OCT4 and CDX2 and the cell number of embryonic outgrowths. We examined the individual and combined effects of RI compared to those of the miR-155 mimic and found that RI can alleviate the negative effects of the miR-155 mimic on embryo attachment and outgrowths. EVs can improve embryo attachment and the unwanted effects of the de trop cargo contents (miR-155) can be alleviated through anti-apoptotic molecules such as the ROCK inhibitor.
The pig is considered a crucial model for transgenic animals and xenotransplantation; however, the process of embryo production in vitro is quite challenging due to a drastic decrease in the embryonic cell number and blastocyst formation as compared toother farm animal species [1,2]. This raises several questions about whether the effects are endogenous or lack exogenous supportive signals. Preimplantation embryos are more competent when co-cultured with other embryos or maternal cells due to the production of paracrine or juxtracrine factors that interact to support the inefficient culture conditions associated with individually cultured embryos [3,4,5]. The cell number of in vivo–derived embryos is twice that of those that are generated in vitro, and the oviduct significantly affects the cell number and results in a 1.5-fold increase in cell number and hatching rates [6]. It shows that exogenous maternal factors are important for acquiring embryo developmental competence, and there are some other endogenous factors from the embryos themselves that may hamper the development of the embryo. Recent transcriptomics studies have shown vast differences between the porcine blastocysts that are produced in vivo and in vitro and demonstrate that upregulated gene expression of metabolism and arginine transporter contribute to the low developmental competence in in vitro–derived embryos [7]. To reveal the possible factors that regulate the blastocyst development and the rate of attachment in porcine embryos, we designed experiments to investigate the molecular impact of exogenous and endogenous signals responsible for embryonic–maternal crosstalk. For instance, the endogenous factors are formed or released by the embryos themselves, while the exogenous factors are released by the maternal cells and hamper the developmental competence such as miRNAs. Several cargos of protein, mRNA, miRNA, and metabolites are carried through the extracellular vesicles (EVs) and affect the growth of the embryo [4,8,9]. Moreover, the interplay between the EVs derived from the endometrium during embryo implantation in humans and animals has been investigated [10,11,12]. The porcine endogenous retrovirus (PERV) is secreted by all porcine cells and is considered a natural inhabitant of cells and biological fluids including the uterine cells. The endogenous retroviruses establish interplay between maternal and embryonic cells and are present in the exosomes released by the endometrium [13,14,15]. Studies also revealed a supportive role of Rho-associated coiled-coil-containing kinases (ROCK) in the development of cleaved embryos, while ROCK inhibition is critical during embryonic and pluripotent stem cell development [16,17,18], particularly trophoblast adhesion and differentiation [19,20]. The mechanism behind this interplay between the endogenous and exogenous factors that affect porcine embryo developmental competence remains unclear. Therefore, our study is an attempt to understand the interplay between the exogenous factors represented in endometrial EVs and the endogenous factors represented in PERV and ROCK pathways in the developmental competence of porcine embryos to enhance the production of more competent embryos that can meet the needs of cloning and xenotransplantation.
Unless otherwise specified, chemicals and reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA).
Porcine embryos were obtained through chemical parthenogenetic activation of in vitro matured oocytes as per our previous reports [5,21,22]. Porcine ovaries were collected from a slaughterhouse and transferred to the laboratory within 4 h in saline (NaCl 0.9%) at 30 °C. Cumulus–oocyte complexes (COCs) were retrieved through aspiration by an 18-gauge needle connected with a 10 mL syringe. Oocytes surrounded by compact layers of cumulus cells were selected using a stereomicroscope (SMZ 745T, Nikon, Tokyo, Japan) and washed three times in HEPES buffered Tyrode’s medium comprising 0.05% polyvinyl alcohol (TLH-PVA). COCs were cultured in 4-well dishes (Nunc, ThermoFisher Scientific, Roskilde, Denmark) containing 500 mL of a maturation medium comprising TCM-199 (Gibco, Waltham, MA, USA), 10% (v/v) porcine follicular fluid, cysteine (0.6 mM), sodium pyruvate (0.91 mM), epidermal growth factor (10 ng/mL), kanamycin (75 μg/mL), insulin (1 μg/mL), human chorionic gonadotrophin (10 IU/mL; Daesung Microbiological Labs; Uiwang, Korea), and equine chorionic gonadotrophin (10 IU/mL; Daesung Microbiological Labs) for 22 h. Then, the COCs were moved to the same culture conditions without the presence of the hormones for 22 h. Matured COCs were harvested, and cumulus cells were detached by gentle pipetting in hyaluronidase (0.6%) and then were washed in TLH-PVA and equilibrated in a pulsing medium consisting of mannitol (0.28 M), CaCl2 (0.1 mM), HEPES (0.5 mM), and MgSO4 (0.1 mM). Oocytes were then activated with a single direct current pulse of 1.5 kV/cm for 60 μs generated inside a glass chamber of two electrodes in an activation medium. The electric current was generated through a BTX Electro-Cell Manipulator 2001 (BTX Inc., San Diego, CA, USA). Activated oocytes were washed in TLH-PVA and cultured for 7 days in microdrops of porcine zygote medium-5 (PZM-5, Functional Peptides Research Institute Co. Ltd. (IFP), Yamagata, Japan) overlaid with mineral oil in a humidified atmosphere at 38.5 °C (5% O2, 5% CO2, and 90% N2). Blastocysts were obtained and washed in PBS and zona pellucida was removed by 0.1% pronase (w/v in PBS) to obtain zona-free embryos ready for further experiments.
Uterine tissues of diestrus multiparous sows were collected from the slaughterhouse and transported to the lab within 4 h. Endometrium was separated aseptically under a laminar flow hood [23]. Endometrium was chopped into 1 mm pieces and seeded on 100 mm tissue culture dishes with a minimal volume of culture medium that comprised DMEM, 10% fetal bovine serum, and penicillin/streptomycin (100 U/mL penicillin, 100 µg/mL streptomycin) at 38.5 °C in a humidified atmosphere of 5% CO2. Tissue attachment and primary cell outgrowths were observed on day-5 of culture and the culture medium was then changed to a fresh one. Primary culture monolayer was maintained until day-8, and the tissue remnants were mechanically discarded.
On day-8, endometrial cell monolayers were cultured in a serum-free culture medium for 24 h and the conditioned medium was aspirated and centrifuged at 300× g to discard cell debris pellets [24]. EVs were isolated through targeted protein binding and nanofiltration using PureExo Exosomes Isolation kits (101 Bio, Palo Alto, CA, USA) [25] to yield 100 µL of EVs in phosphate-buffered saline (PBS) solution. EVs were characterized through ZetaView PMX 110 (Particle Metrix, Meerbusch, Germany) nanoflow fluorescence cytometry and nanoparticle tracking analysis instrument associated with ZetaView 8.05.14 SP7 software and Microsoft Excel 365 (Microsoft Corp., Seattle, WA, USA) [26]. After calibration with 100 nm polystyrene particles (ThermoFisher Scientific), one mL of the sample (diluted 20X in 1× PBS) was loaded into the machine and eleven different positions and two reading cycles per position were automatically set to measure the mean, median, and mode sizes (indicated as diameter), concentrations, and outlier removal in each sample. EVs were examined through transmission electron microscopy (TEM) [4,27]. In brief, 4 μL of isolated EVs solution was stained with 2% uranyl acetate, mounted on the center of 200-mesh copper grids, dried, and visualized through an OMEGA-energy filtering TEM (ZEISS LEO 912, Carl Zeiss, Jena, Germany) at 120 kV. The EVs cargo contents of some selected mRNAs, miRNAs, and proteins were analyzed through reverse-transcription polymerase chain reaction and proteomics as discussed below.
Embryo attachment in feeder-free culture condition was established according to our previous method [28] with some modifications. Fifty μL microdrops of Matrigel basement membrane matrix (BD Biosciences, San Jose, CA, USA) were placed on 4-well dishes (Nunc) and incubated for 30 min at 38.5 °C. Matrigel was removed and replaced with 50 μL of culture medium that was composed of DMEM/F-12 supplemented with 10% fetal bovine serum, β-mercaptoethanol (0.1 mM), 1% nonessential amino acids (Invitrogen, Waltham, MA, USA), and 1% penicillin/streptomycin (100 U/mL penicillin, 100 µg/mL streptomycin). The microdrops were covered with mineral oil and incubated for 30 min before embryo placement. On day-7, embryos were collected and zona pellucidae were removed by pronase (0.1% in PBS) for 1 min at 38.5 °C. Embryos were washed with the culture medium before placing them into the Matrigel-coated microdrops. Embryos were then incubated in a humidified atmosphere of 5% CO2 at 38.5 °C and monitored for attachment and outgrowths on days 2–5 from culture.
First, embryos (n = 20 for 3 replicates) were divided into 4 groups: control, 10 µM ROCK (Rho-associated coiled-coil containing kinases) inhibitor (RI, Y-27632) [29], Endo-EVs (1.5 × 107 particles/mL) [25,30], or combined supplementation of RI and Endo-EVs for different durations (i.e., 36 h and extended to 5 days). The control group was cultured in a plain culture medium without supplementation. Embryos were monitored for attachment, cell number, apoptosis, outgrowths, immunofluorescence staining of pluripotency marker (Oct4) and trophoblast marker (Cdx2), and relative quantitation of some miRNA and mRNA transcripts’ expression that are related to apoptosis, cell attachment, cell cycle, and embryo development.
Based on the findings of EVs analysis, we designed experiments to explore the roles of miR-155 in maternal-embryonic communications. The mir-155 inhibitor was supplied to examine its effect on Endo-EVs supplementation on embryonic development and attachment. Moreover, the effects of miR-155 mimic on embryonic development and attachment were studied in combination with or without RI.
Based on the findings of EVs analysis, we targeted PERV with CRISPR/Cas9 to explore the role and impact of EVs derived from PERV-depleted endometrium on embryo development and attachment.
Before EVs isolation, a serum-free conditioned medium was mixed with the PKH26 lipophilic fluorescent stain (Invitrogen) according to the manufacturer’s instructions, and EVs were isolated to remove the excess free PKH67 dye following the manufacturer’s recommendations [31,32]. EVs were then supplemented (1.5 × 107 particles/mL) [25,30] with cultured embryos for 24 h to monitor their uptake through a fluorescent microscope (MshOt, Guangzhou Micro-shot Technology Co., Ltd., Guangzhou, China). For negative control staining, the plain conditioned medium was mixed with PKH26 and processed by the same EV labeling procedure.
Immunofluorescence staining of OCT4 and CDX2 was performed according to our previous protocol [33] with some modifications as follows: attached embryos on day-5 were fixed in 4% paraformaldehyde (w/v in PBS), pH 7.4 for 15 min at room temperature. Fixed embryos were washed in PBS, permeabilized with 0.1% Triton-X100 (v/v in PBS) for 10 min, and then were blocked by 1% goat serum (v/v; Invitrogen) for 30 min at room temperature. Primary antibodies specified against Oct4 (mouse monoclonal IgG2b, sc-5279, Santa Cruz Biotech. Inc., Santa Cruz, CA, USA) and Cdx2 (rabbit monoclonal IgG, ab76541, Abcam, Seoul, Korea) were diluted (1:100) and prepared in PBS. The attached embryos were incubated with the primary antibodies (1 h at 38.5 °C), washed in PBS three times, then incubated with the secondary antibodies (Alexa Fluor 488 goat anti-mouse IgG, A11001 and Alexa Fluor 568 goat anti-rabbit IgG, A11011; Invitrogen, Life Technologies Corp., Eugene, OR, USA), and the resulting solution was diluted (1:200) and kept in PBS for 1 h at 38.5 °C before washing in PBS three times. Embryonic cell nuclei were counterstained with Vectashield antifade mounting medium containing 40,60 -diamidino-2-phenylindole (DAPI; Vector Laboratories, Vector Laboratories, Burlingame, CA, USA) for 5 min, and the fluorescence signals were examined with a fluorescent microscope at 488 nm, and 568 nm for Oct4 and Cdx2, respectively. Images were captured and the fluorescence intensity pixel analysis was analyzed with ImageJ 1.53k software (National Institute of Health, Bethesda, MD, USA).
Labeling of DNA strand breaks and detection of apoptotic cells were examined through In Situ Cell Death Detection TUNEL assay Kit, Fluorescein (Roche Holding AG, Basel, Switzerland) according to the manufacturer’s protocol. Embryos were fixed in 4% paraformaldehyde and permeabilized in 0.1% TritonX and then incubated with the working solution of an enzyme (TdT) and a label solution (fluorescein-dUTP) for 1 h at 38.5 °C. Nuclei were counterstained with Vectashield antifade mounting medium as mentioned above. Green fluorescence positive cells (apoptotic cells) were captured and counted with ImageJ software.
We used Campylobacter jejuni CRISPR/Cas9 (cjCas9) vector to cleave PERV mRNA. We cloned cjCas9-based sgRNA targeting env gene of PERV in our cjCas9 vector with slight modification (D8A for inactivating RuvC domain) (Supplementary Figure S1). CRISPR/Cas9 vector (1 mg), miR-155 mimic (100 nM), and miR-155 inhibitor (100 nM) oligonucleotide sequences (Table 1) [34] were transfected to the embryos [35] with some modifications. The nucleic acids were incubated with Lipidofect-P transfection reagent (Cat # LDL-P001, Lipidomia, Gachon University IT Center, Gyeonggi-do, Korea) for 30 min at room temperature, and then the mixture was supplemented to the embryo culture medium and incubated for the attachment and further development.
Total RNA was extracted from the embryos (n = 5, 4 replicates) using RNeasy Micro Kit (Qiagen GmbH, Hilden, Germany, Cat #74004) that included DNase I for removing any of DNA residuals. NanoDrop 2000 (Thermo Scientific) was used to determine the quality of the extracted RNA. Values of > 1.8 of OD 260/280 and 260/230 ratios were used for the reverse transcription. Complementary DNA (cDNA) was synthesized using 2X RT Pre-Mix of QuantiNova Reverse Transcription Kit (Qiagen) with a total volume of 20 μL (1 μg of total RNA, 4 μL of 5× RT buffer, 1 μL of reverse transcriptase enzyme mix, 1 μL of oligo dT primer for mRNA or universal stem-loop primer for miRNAs (Table 1), and RNase-free distilled water to reach the volume of 20 μL). Pulsed reverse transcription was conducted to generate complementary DNA (cDNA). Twenty nanograms of total RNA in a 20 μL reaction volume was used as a template for cDNA synthesis in 60 cycles of 2 min at 16 °C, 1 min at 37 °C, and 0.1 s at 50 °C, and a final inactivation at 85 °C for 5 min [36,37]. For conventional PCR, cDNA was amplified by using 2X Taq PCR Pre-Mix (BioFACT, Seoul, Korea) in the following conditions: initial denaturation (2 min at 95 °C), 40 amplification cycles of denaturation (95 °C for 20 s), annealing (60 °C for 30 s), and extension (72 °C for 30 s), followed by a final extension step of 5 min at 72 °C. The PCR products were analyzed by a Gel Doc XR+ UV transilluminator with Image Lab Software (Bio-Rad, Berkeley, CA, USA) on a 2.5% agarose gel (Amresco, Cleveland, OH, USA) stained with ethidium bromide (BioFACT). Gel electrophoresis was performed using Mupid®-One (TAKARA, Tokyo, Japan) at 135 V for 25 min. Relative quantitative PCR was performed using the CFX Connect Real-Time PCR system (Bio-Rad) and SYBR 2X Real-Time PCR Pre-Mix (BioFACT). Details about the target genes, housekeeping mRNA and snRNA, primers, and product size are listed in Table 2. The reaction mixture (20 μL) comprised 10 μL of SYBR® Premix (2×), 2 μL of cDNA (100 ng), 2 μL of forward and reverse primers (10 μM), and 6 μL of distilled water. Cycling conditions were 95 °C (1 min) followed by 40 PCR cycles of 95 °C (5 s, DNA denaturation), 60 °C (30 s, primer annealing), and 72 °C (30 s, extension). Primer specificity was determined by melting curve protocol ranging from 65 to 95 °C and was confirmed with single peaks in the melt curves, gel electrophoresis, and cDNA-exempted samples. Transcripts of the target genes were compared to those of housekeeping genes (GAPDH-mRNA and U6-snRNA). The gene networks of these studied transcripts were analyzed through GeneMANIA webtool (https://genemania.org/, accessed on 30 September 2022) [21].
The protocol was performed according to our recent report [27]. In brief, EVs pellets were suspended and dialyzed against 10 volumes of 20 mM Tris-HCl (pH 8.0) (the molecular mass cutoff of was 10,000 Da). Protein concentration was estimated through the bicinchoninic acid method and then proteins were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). For Coomassie Brilliant Blue staining, the gels were destained by 50% acetonitrile and 10 mM ammonium bicarbonate solution [46] and then gels were rinsed twice with distilled water, followed by 100% acetonitrile, respectively and then dried with a speed vacuum concentrator. The gels were treated with mixture of 10 mM dithiothreitol and 100 mM ammonium bicarbonate at 56 °C, before treatment with 100 nM iodoacetamide to minimize alkylate S–S bridges. The gels were vortexed in three volumes of distilled water for washing and then dried with a speed vacuum concentrator. The gels were incubated in 50 mM ammonium bicarbonate and 10 ng/mL trypsin at 37 °C for 12–16 h for tryptic digestion. Tryptic peptides were retrieved after treatment with 50 mM ammonium bicarbonate and 50% acetonitrile containing 5% trifluoroacetic acid. Peptide extract was lyophilized and stored at 4 °C until further analysis. Tryptic peptide extract was suspended in 0.5% trifluoroacetic acid and 10 μL from each sample was loaded onto MGU30-C18 trapping columns (LC Packings) to concentrate peptides and clear extra chemicals. Concentrated tryptic peptides were eluted from the column and loaded onto a 10 cm × 75 μm I.D. C18 reverse-phase column (PROXEON, Odense, Denmark) at adjusted flow rate (300 nL/min). Peptides were retrieved by a gradient of 0–65% acetonitrile for 80 min. MS and MS/MS spectrum was obtained by using LTQ-Velos ESI ion trap mass spectrometer (Thermo Scientific, Waltham, MA, USA). MASCOT 2.4 was used to analyze MS/MS data with a false discovery rate of 1% as a cutoff value. Protein quantities were estimated through the exponentially modified protein abundance index (emPAI) and were expressed as mol %. Three technical replicates were performed. Functional analysis and gene ontology were performed through the Functional Annotation Tool, DAVID Bioinformatics Resources (NIAID/NIH; https://david.ncifcrf.gov/home.jsp, accessed on 16 August 2022) [36,47].
For each experiment, an average of 20 embryos and at least 3 replicates were used for the analysis. Lieven’s test and Kolmogorov–Smirnov test were used to confirm the homogeneity of variance and the normality of distribution, respectively. Data were expressed as mean ± standard error of means (SEM) or standard deviation (SD) and examined through using an unpaired Student t-test or with univariate analysis of variance (ANOVA) followed by Tukey’s multiple comparison test, respectively. Statistical significance was considered at p <0.05 or p <0.01. GraphPad Prism 5 (GraphPad Software Inc., San Diego, CA, USA) was used for the statistical analyses.
Endometrial cells were successfully cultured with the characteristic flattened epithelial properties and were maintained in culture until day-8 (Figure 1A,B). Cells were cultured in a serum-free culture medium for 24 h to retrieve the Endo-EVs. ZetaView analysis showed a presence of 1.1 × 108 particles/mL of average particle size 115.6 ± 28.4 nm in the isolated conditioned medium (Figure 1C). Furthermore, TEM images revealed the presence of lipid bilayer vesicles (Figure 1D) that characterize the appearance of EVs. The isolated Endo-EVs were analyzed with qPCR and showed the expression of certain miRNAs and mRNAs when compared to those of corresponding endometrial origin (Figure 2). Endo-EVs contained miR-100, miR-132, and miR-155, β-catenin, and PERV, but we could not observe GAPDH mRNA in the isolated EVs. Endo-EVs were further characterized through the profiling of protein contents by proteomics. We identified 82 proteins in the Endo-EVs (Supplementary Table S1), of which the top 20 proteins that constituted around 65% of the total proteins of the Endo-EVs included proteins associated with cytoskeleton structures (e.g., keratins), extracellular matrix (e.g., plasminogen), and calcium metabolism (e.g., vitamin-D-binding protein, and calcium-binding protein A2); however, they contain apoptosis-related proteins such as procathepsin (Table 3). Detailed information about the proteins and their functions are listed in Supplementary Tables S2 and S3.
The embryonic uptake of Endo-EVs was confirmed by the presence of intracytoplasmic fluorescence signals after labeling Endo-EVs with PKH26 stain and their incubation with the embryos for 30 h (Figure 3). We then investigated the embryonic attachment and the development after supplementing the culture medium with EVs or with RI or with their combination together. When compared with the control group, Endo-EVs supplementation for 36 h showed significant improvement in embryo attachment (65.55% vs. 34.43%, p < 0.01), increased embryonic cell number (26 vs. 21.8, p < 0.01), and a significant reduction in apoptosis (2.37% vs. 26.1%, p < 0.01) (Figure 4A–D). Based on these preliminary experiments, we found that RI can reduce apoptosis in embryonic stem cells (Figure 4D); however, it could not support the embryonic development (Figure 4C) compared to control and Endo-EVs groups. Therefore, we examined the beneficial effects of both Endo-EVs and RI to support the embryonic development, which showed 100% embryonic attachment with a significant increase in cell numbers (mean = 33.6) and a significant reduction in the ratio of apoptotic cells (1.57%) when compared to those of the experimental groups (Figure 4A–D). We followed up the development of embryos supplanted with Endo-EVs and RI for 5 subsequent days and compared that with the control group. The results showed a significant increase in the embryonic outgrowths on day-5 in the embryos supplemented with Endo-EVs and RI (72.9% vs. 32%) (Figure 5A,B). Furthermore, the expression of Oct4 and Cdx2 in the Endo-EVs and RI-treated embryonic cells was significantly increased by 2.45-fold and 3.48-fold, respectively, compared to those of the control group (Figure 6A–C). Additionally, the qPCR analysis showed that combined supplementation of Endo-EVs with RI significantly reduced the expression of Bax (0.6-fold) and miR-155 (0.17-fold) but increased the expression of Bcl2 (4.73-fold), Cdk2 (4.33-fold), PERV (2.55-fold), β-catenin (7.13), interferon-gamma (1.7-fold), Zeb1 (1.9-fold), PTN mRNAs (9.83-fold), miR-100 (3.79-fold), and miR-132 (16.1-fold) compared to those of the control group (Figure 7).
Based on the previous results, we speculated that the miR-155 contents of Endo-EVs could exert a negative impact on embryo attachment, and therefore we specifically targeted miR-155 with an inhibitor (miR-155 inhibitor). Treatment of embryos with Endo-EVs and miR-155 inhibitor significantly improved the attachment (90% vs. 50%, p < 0.01), increased the cell number (30 vs. 12, p < 0.01) but had no effects on the apoptosis ratio (2% vs. 3.16%, p = 0.27) as compared to those of Endo-EVs supplemented group (Figure 8A–D). Additionally, individual treatment with miR-155 mimic showed a significant reduction (p < 0.05) in the attachment (20.8%) and cell number (n = 6) and a significantly increased apoptosis ratio (48.3%). Moreover, these effects were significantly alleviated with an individual treatment of RI (45.6%, 14, and 14.66% for attachment ratio, cell number, and apoptosis ratio, respectively; p < 0.01) (Figure 9A–D). Hence, we speculated that RI could antagonize the negative impact of miR-155 on embryonic attachment and development.
Furthermore, based on the Endo-EVs cargo contents of PERV, we targeted PERV with CRISPR/Cas9 to examine the effects of PERV reduction on embryonic attachment and development. The designed CRISP/Cas9 was successfully transfected and expressed in the cells as indicated by the green fluorescence protein expression in Figure 10A,A’,A’’. PERV expression was significantly reduced (0.23-fold, p < 0.05) compared to that of control endometrium cells. Surprisingly, miR-155 expression showed a 6.16-fold increase (p < 0.05) in PERV-depleted endometrium compared with that of the control endometrium. Similarly, the derived EVs from PERV-depleted cells showed a significant reduction in PERV mRNA expression (0.27-fold, p < 0.05), and a significant increase in miR-155 (4.9-fold, p < 0.01). Moreover, supplementation of embryos with EVs from PERV-depleted cells significantly reduced the attachment and day-5 outgrowth ratios compared to those of control EVs (49% vs. 65.8% and 18% vs. 31%, respectively, p < 0.05).
In this study, as a continuation of our work [28], we generated a model of culturing porcine embryos in feeder-free culture conditions using Matrigel basement membrane matrix but on a microdrop level. This model achieved 100% of blastocyst attachment and embryonic outgrowths with the help of Endo-EVs supplementation and ROCK pathway inhibition. Recently, the interplay between EVs derived from the endometrium during the embryo implantation in humans has been investigated [10,11,12]. Our results showed that Endo-EVs enhanced embryonic attachment and adhesion to the Matrigel basement membrane matrix, which is in accordance with previous studies [11,48,49,50,51]. This improvement may be attributed to the transfer of proteins associated with cell attachment, cytoskeleton integrity, calcium metabolism as revealed by proteomics analysis. Additionally, embryonic development was improved because of β-catenin transfer through the Endo-EVs cargo that increased the expression of β-catenin in embryos. β-catenin plays a crucial role in endometrium functions and is considered an imperative signal in invasion and differentiation of trophoblasts and embryo implantation [39,52]. Moreover, Endo-EVs cargo contained miR-100, which is expressed in human endometrial cell-derived EVs and activates focal adhesion kinase (FAK) and c-Jun N-terminal kinase (JNK) and promotes the invasion and migration of human and goat trophoblasts [12,53,54]. Furthermore, Endo-EVs contained miR-132 that is expressed temporally in porcine endometrium at the time of embryo implantation [55] and is a potential factor for enhancing trophoblast invasion and embryo implantation by targeting death-associated protein kinase 1 (DAPK-1) [56]. In our study, several mRNA transcripts (antiapoptotic gene (BCL2), zinc finger E-box-binding homeobox 1 (Zeb1), β-catenin, interferon-γ (IFNG), protein tyrosine phosphatase non-receptor type 1 (PTPN1), and cyclin-dependent kinase 2 (CDK2)) were increased in the embryonic cells after EVs supplementation. Moreover, the pluripotency master Oct4 and the trophoblast associated gene CDX2 were also increased in the EVs supplementation which might be attributed to embryo developmental competence observed in this group [4,28,57,58,59]. These genes are of important roles in embryonic development, implantation, trophoblast attachment, and stem cell growth, as well as the cell cycle and survival as we revealed in our previous reports [5,21]. Furthermore, bioinformatics analysis indicated an existing physical interaction, shared protein domains, common pathways, and co-expression of the studies genes (Supplementary Figure S2). Detailed functions of the genes are listed in Table 2. On the other hand, Endo-EVs contained miR-155, which inhibits trophoblast migration and proliferation [60,61], increases preeclampsia in patients, and induces trophoblast apoptosis by targeting BCL2 (apoptosis regulator) [62]. Furthermore, some of the cargo proteins in the Endo-EVs included proapoptotic signals such as cathepsin and procathepsin. Paradoxically, miR-155 showed a 1.6-fold increase in the mouse uterus during the receptive phase of embryo attachment, which suggests a modulatory role of miR-155 during this critical stage in mice [63]. In this study, the ROCK inhibitor (Y-27632) abolished all negative impacts of miR-155 on embryo attachment and development. RI remarkably reduces the tumor necrosis factor (TNFα)-induced upregulation of miR-155 [64]. RI interferes with the cargo transfer from microparticles and impairs their ability to mediate extracellular signaling [65]. Therefore, we speculate the likelihood of other mechanisms that are associated with miRNA export upstream to RhoA/ROCK signaling. Another mechanism is the antiapoptotic action of RI can antagonize the apoptotic action of miR-155 on embryonic cells [18,66,67,68,69]. Additionally, ROCK pathway inhibition enhances trophoblast adhesion and viability in humans. Paradoxically, RI can reduce the trophoblast migration of human extravillous trophoblasts [70]. This is in accordance with our findings regarding the ameliorative effects of RI on the negative impacts of both individual and EVs-transmitted miR-155; however, individual RI improved the embryonic attachment and development. Therefore, RI synergize the actions of Endo-EVs through antagonizing the effects of the non-useful cargo contents of Endo-EVs, such as miR-155. Computational analysis of miR-155 targets (http://mirdb.org/, accessed on 16 August 2022) showed that they interfere with a cell-cycle-related gene (CDK2-associated cullin domain 1 (CACUL1)) (target score 82%) and an antiapoptotic gene (BCL2-associated athanogene 5 (BAG5)) (target score 70%), which had correlation with other proteins involved in cell apoptosis and growth, including BCL-2. Our qPCR data showed that RI ameliorates the negative effects on mRNA expression of CDK2 and BCL2, which could help the cell cycle and reduce apoptosis. Moreover, miR-155 targets catenin alpha 3 (CTNNA3) (target score 67%) which belongs to the catenin family and encodes a protein that plays a role in cell-to-cell adhesion. Similar findings in qPCR have been shown in β-catenin expression. Furthermore, miR-155 targets protein tyrosine phosphatase, non-receptor type 2 (PTPN2) (target score 84%), which regulates various cellular processes including cell growth, differentiation, and mitotic cycle. Moreover, we found positive effects of RI on the expression of PTPN1. Therefore, we inferred that miR-155 can interfere with several essential pathways related to cell growth and differentiation and cause apoptosis. The EVs cargo contained PERV mRNA, which coincides with some recent reports showing exosomes that contain mRNAs for ovine endogenous jaagsiekte retroviruses (enJSRV-ENV) [13] and human endogenous retroviruses [71]. There is a consensus about the essential roles of endogenous retroviruses in the early stages of embryo attachment physiological functions of trophoblasts and placentation [15,72,73,74,75]; however, the role of PERV in porcine embryo attachment remains unclear. A recent report showed that targeting PERV with CRISPR/Cas9 at the zygote stage impaired the blastocyst development and indicated the essential roles of PERV for the preimplantation embryonic development [14]. Our results indicate that PERV targeting in EVs could have reduced the embryonic attachment and development. This reduction might be due to the increased levels of miR-155 in the transferred cargo contents of EVs. The reason behind these increased levels is unclear and might be attributed to the essential roles of PERV in cellular viability, normal physiological functions, and the indel mutations caused by CRISPR/Cas9 [14]. Furthermore, avian endogenous retrovirus shows negative regulation with miR-155, which is suggestive of the interplay between ERVs and miR-155 [76]. The human and zebrafish microRNA-155 target the corresponding HERV and ZFERV env sequence, which indicates that miR-155 targeting ERVs env is mostly conserved in animals and may regulate ERVs activity [76]. We speculate that EVs can carry both useful and harmful cargo contents and ameliorating the de trop cargo contents (such as miR-155) can maximize the useful effects of EVs, especially during embryo implantation and maternal recognition of pregnancy.
To our knowledge, this is the first attempt to understand the roles of EVs cargo in determining embryo developmental competence and mediating molecular signaling between the embryo and the endometrium in the pig. Endometrial EVs improved embryo attachment, increased cell numbers and reduced apoptosis, and the unwanted effects of their de trop cargo contents of miR-155 can be alleviated through anti-apoptotic molecules such as the ROCK inhibitor. This model would help in establishing an extended culture system to understand early embryonic stem cell differentiation. The current model would provide a paradigm for studying the embryonic–maternal crosstalk and to develop pharmaceutical criteria for improving pregnancy outcomes in porcine species. | true | true | true |
|
PMC9565427 | Biao Feng,Jieting Liu,Eric Wang,Zhaoliang Su,Subrata Chakrabarti | Endothelial derived miRNA-9 mediated cardiac fibrosis in diabetes and its regulation by ZFAS1 | 14-10-2022 | Diabetic cardiomyopathy (DCM) is one of the most prevalent causes of morbidity and mortality in diabetic patients. Hyperglycemia induces increased expression/deposition of extracellular matrix (ECM) proteins including fibronectin (FN) and collagen (Col) and plays an important role in fibrosis in diabetic cardiomyopathy (DCM). The roles of RNAs including microRNA (miRNA) and long non-coding RNAs (lncRNA) have begun to be understood in many conditions. In this study, we investigated the role of a specific miRNA, miR-9, and its interactions with lncRNA ZFAS1 in mediating fibrosis in DCM. Treatment with 25 mM glucose (HG) decreased miR-9 expression and increased expressions of ZFAS1, ECM proteins and inflammatory markers, compared to 5 mM glucose (NG) in the HCMECs by using qRT-PCR. Glucose-induced upregulation of ECM proteins can be prevented by ZFAS1 siRNA or miR-9 mimic transfection. Luciferase assay was confirmed miR-9 binding to FN 3’-UTR. miR-9 expression can be regulated by ZFAS1 through polycomb repressive complex 2 (PRC2) components using RNA immunoprecipitation (RIP) and chromatin immunoprecipitation (ChIP) assays. In the in vivo experiment, hyperglycemia-induced the ECM production can be prevented by the miR-9 overexpression in the fibrosis in DCM. These studies showed a novel glucose-induced molecular mechanism in which ZFAS1 participates in the transcriptional regulation of ECM protein production in diabetes through miR-9. | Endothelial derived miRNA-9 mediated cardiac fibrosis in diabetes and its regulation by ZFAS1
Diabetic cardiomyopathy (DCM) is one of the most prevalent causes of morbidity and mortality in diabetic patients. Hyperglycemia induces increased expression/deposition of extracellular matrix (ECM) proteins including fibronectin (FN) and collagen (Col) and plays an important role in fibrosis in diabetic cardiomyopathy (DCM). The roles of RNAs including microRNA (miRNA) and long non-coding RNAs (lncRNA) have begun to be understood in many conditions. In this study, we investigated the role of a specific miRNA, miR-9, and its interactions with lncRNA ZFAS1 in mediating fibrosis in DCM. Treatment with 25 mM glucose (HG) decreased miR-9 expression and increased expressions of ZFAS1, ECM proteins and inflammatory markers, compared to 5 mM glucose (NG) in the HCMECs by using qRT-PCR. Glucose-induced upregulation of ECM proteins can be prevented by ZFAS1 siRNA or miR-9 mimic transfection. Luciferase assay was confirmed miR-9 binding to FN 3’-UTR. miR-9 expression can be regulated by ZFAS1 through polycomb repressive complex 2 (PRC2) components using RNA immunoprecipitation (RIP) and chromatin immunoprecipitation (ChIP) assays. In the in vivo experiment, hyperglycemia-induced the ECM production can be prevented by the miR-9 overexpression in the fibrosis in DCM. These studies showed a novel glucose-induced molecular mechanism in which ZFAS1 participates in the transcriptional regulation of ECM protein production in diabetes through miR-9.
Diabetes is a worldwide healthcare challenge. According to data from IDF, the number of adults affected with diabetes globally is estimated to reach 578 million by 2030. Patients with diabetes mellitus are at increased risk for developing to chronic complications, such as retinopathy, nephropathy and cardiomyopathy. Cardiomyopathy is one of the leading cause of morbidity and mortality among patients with diabetes [1, 2]. Diabetic cardiomyopathy (DCM) is characterized by left ventricular hypertrophy, myocardial fibrosis and diastolic dysfunction, which are not attributable to underlying coronary artery disease or hypertension [3]. One of the key manifestations of DCM, myocardial fibrosis, is caused by increased production and deposition of extracellular matrix (ECM) proteins, such as collagen and fibronectin (FN). Increased ECM protein deposition leads to structural changes in the heart in DCM include thickening of the capillary basement membrane, focal myocardial fibrosis etc. [4]. Endothelial cells (ECs) are initial targets of hyperglycemic damage and play a major role in the production of ECM proteins in all chronic diabetic complications [5]. Several regulatory mechanisms including specific microRNAs (miRNAs) and long non-coding RNAs (lncRNAs) may play significant roles increased ECM deposition in DCM. miRNAs are short (~22nt) RNA molecules with no protein coding function, so-called non-coding RNAs. miRNAs regulate biological processes through interactions with the 3’UTR of target mRNAs. miRNAs have been demonstrated to contribute to pathological processes and the development of cardiac fibrosis in DCM [6]. We have previously shown downregulation of miR-146a in association with upregulation of ECM protein transcripts in the heart, kidneys and retinas in diabetes [7–9]. Other investigators also demonstrated the involvement of other miRNAs in the regulation of ECM protein expression in diabetes [10, 11]. Several studies showed that alteration of miR-9 expression related to the cancers, such as cervical cancer, breast cancer, prostate cancer, etc., and Alzheimer’s disease [12]. miR-9 also mediates CALHM1-activated ATP-P2X7R signalling in diabetic neuropathy in diabetic rat model [13]. One study showed that miR-9 inhibits high glucose-induced proliferation, differentiation and collagen accumulation of cardiac fibroblasts by down-regulation of TGFBR2 [14]. Previous studies suggest that methylation of miR-9-1 and miR-9-3 favors diabetic retinopathy (DR) in patients with diabetes [15]. However, role of miR-9 in DCM has not reported. lncRNAs are RNA molecules longer than 200 nucleotides in length that are also not translated into functional proteins [16]. Apart from miRNAs, lncRNAs are another major class of non-coding RNAs that are widely expressed in mammals. lncRNAs regulate local and distal genes by various mechanisms and play key roles in diverse biological processes. They play important roles in physiological processes and cellular functions such as cell proliferation, resistance to apoptosis, and induction of angiogenesis [17, 18]. Recent evidence shows that dysregulation of target genes leads to abnormal lncRNA expression in diabetes [19]. A previous study from our lab showed that changes in expression of the lncRNA ANRIL plays an important role in DCM [20]. Nevertheless, the potential functions of lncRNAs and the related mechanisms in DCM are still largely unknown and needs further exploration. Recent evidences have highlighted the importance of epigenetic regulation in the pathogenesis of DCM [21–23]. The research data further indicate that certain epigenetic processes are fine-tuned by lncRNAs via direct interaction with cis-regulatory elements, or through interactions with key RNA binding proteins and chromatin modifiers [24]. Such epigenetic modifications include DNA methylation, acetylation and methylation of histones. These modifications can alter the production of specific miRNAs and lncRNAs, which play important roles through regulation of gene expression at both transcriptional and translational levels [25]. Furthermore, a regulatory interaction exists between specific miRNA and lncRNA ANRIL [26]. In this study, we examined the role of miR9 in diabetic conditions both in vitro and in vivo. To examine the role of miR-9 in DCM in vitro, we used endothelial-specific miR-9 overexpressing mice generated by us. We also investigated the role of ZFAS1 in regulating the production of ECM protein in human cardiac microvascular endothelial cells (HCMECs) and in DCM in mouse models. We further examined the interactions between ZFAS1 and miR-9 in mediating cardiac fibrosis in diabetes.
Human cardiac microvascular endothelial cells (HCMECs) were obtained from ScienceCell (Carlsbad, CA, USA) and were cultured in endothelial basal media-2 (EBM-2) supplemented with Endothelial Cell Growth Medium-2 SingleQuots (EGM-2) and 10% fetal bovine serum as previously described [7, 27]. Cells were serum-starved for 24 hours once they reached 80–90% confluency. HCMECs were then treated with serum-free media containing various concentrations of D-glucose (with 25 mM mannitol as osmotic control) for 48 hours. The cells were then collected for the further analysis. Mouse heart endothelial cells (MHECs), fibroblasts and myocytes were isolated as described previously [7, 27]. Briefly, on the day of the experiment, Dynabeads were washed in 0.1% BSA-PBS (B-PBS) for 3 times. The beads were then incubated with 20 μL of rat anti-mouse CD31 monoclonal antibody (Invitrogen, ON, Canada) at 4°C for 3 hours. Anti-CD31 antibody-conjugated Dynabeads (CD31-beads) were washed for 3 times using B-PBS and re-suspended in 200 μL B-PBS on ice for later use. The mice were euthanized, the heart tissues were collected, and tissue from the left ventricles were minced into 1 mm2 small pieces and digested with collagenase A for one hour at 37°C. The mixtures were flushed using 5 ml pipet and transferred to 50 mL tube containing cold DMEM with 2% FCS (DMEM-F) and pelleted by centrifugation. The pellets were re-suspended in the DMEM-F solution, passed through a 100 μm cell strainer, and pelleted again. The cells were further incubated with CD31-beads for one hour at 4°C and washed 5 times using B-PBS. Cells bound to CD31-beads (MHECs) were cultured in 2% gelatin coated 6 well plates (Falcon, BD, Canada) in EBM2 with supplemental kit same as above (Lonza, MD, USA). The unbound cells were incubated in 75cm2 flask at 37°C for one hour, after which the suspended cells (myocytes) were collected. Cells attached to the bottom of the flask (fibroblasts) were cultured with DMEM with 10% FCS and collected after purification using CD90 beads. Isolated cardiac ECs (estimated purity 90%), fibroblasts and myocytes were used for separate analyses. The cellular phenotypes were also confirmed by RT-PCR assay for cell type specific markers.
HCMECs were transfected with miRNA-9 mimic (20 nM) using Lipofectamine 3000 (Invitrogen Canada, ON, Canada). Scrambled mimics was used as a control. miRNA transfection efficiency was determined by TaqMan real-time RT-PCR technique [7, 28]. For siRNA experiment, ECs were transfected with ZFAS1 siRNA (20 nM) using Lipofectamine 3000. We initially used multiple siRNAs. They produced similar results and then we selected one with best results. Scrambled siRNA was used a control. For methylation inhibition experiments, 3-Deazaneplanocin A (DZNep, 5 μM, Cayman Chemical, Ann Arbor, MI, USA) was applied as described in previous studies to HCMECs for 1 hour prior to the addition of D-glucose [7, 27, 29]. Treated HCMECs and their controls were then subjected to various glucose treatments, then collected after 48 hours of glucose incubation. Cell viability was examined using MTT assay.
All animal experiments were performed in accordance with the Canadian Council on Animal Care. The protocols were approved by Western University Animal Care and Veterinary services. This investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996).
We generated miR-9 transgenic (miR9TG) mice with EC-specific promoter using previously described methodology [27]. A cDNA fragment containing miR-9 was inserted into the pg52pSPTg.T2FpAXK (pg52) plasmid which contains a Tie2 promoter, enhancer and an SV40 PolyA signal (kindly provided by Dr Sato, Nara Institute of Science and Technology Graduate School of Biological Sciences, Ikoma, Japan) [4]. The fragment containing Tie2-miR-9 was excised by using restriction enzyme SalI, purified, and then injected into mouse blastocysts from the C57BL/6 background. These blastocysts were then transferred into pseudopregnant female mice [28]. Transgenic mice founders were identified by PCR-based assays using the genomic DNA from tail-tip biopsy specimens according to the protocol described previously [27]. miR9TG mice for the experiments were also confirmed by the same method (S1 Fig). No behavioural or phenotypical alterations were observed in the transgenic mice. We used miR9TG mice and littermate controls (both of C57BL/6 background) mice (8 weeks) in this study. Transgenic and control mice were further randomly divided into diabetic and control groups. The animals were injected with freshly prepared streptozotocin (STZ) in sodium citrate buffer (I.P., pH 4.5, 50 mg/kg) or an equal volume of buffer only once a day for 5 consecutive days. Diabetes was confirmed by measuring blood glucose levels (>16.7 mmol/L) following the final STZ injection as described previously [28]. Following confirmation of diabetic for 2 months, the mice were sacrificed using isoflurane and cardiac tissues were collected. A small portion of ventricular tissue was fixed in 10% neutral-buffered formalin and were embedded in paraffin. After fixing 48 hours, the tissues were cut in 5 μm sections and stained with hematoxylin-eosin and trichrome for histologic analysis. The remainder of the cardiac tissues were stored at -80°C for later use.
miRNAs were extracted from the cells and tissues using the mirVana miRNA isolation kit (Thermo Fisher Scientific, MA, USA) following manufacturers’ instructions. Briefly, the cells were lysed and the tissues were homogenized using the Lysis/Binding solution. The miRNA additive from the kit was added to the lysed samples. Equal volume acid-phenol:chloroform was added to cell suspension solution. Following centrifugation, the aqueous phase was removed and 1.25-fold 100% ethanol was added to the mixture. The mixture was then passed through the filter column and miRNAs were eluted. Reverse transcription of RNA and real-time PCR (RT-PCR) were performed using a kit (Life Technologies, USA) as follows: 10 μL TaqMan 2X Universal PCR Master Mix, 8 μL Nuclease-free water, 1 μL TaqMan microRNA assay (Life Technologies, CA, USA) and 1 μL RT-product. U6 snRNA was used as an internal control [27].
Total RNA was extracted using TRIzol™ reagent (Invitrogen Canada Inc., ON, Canada) as previously described [27, 28]. Total RNA (1 μg) was used for cDNA synthesis by using high capacity cDNA reverse transcription kit (ThermoFisher, CA, USA). Real-time quantitative RT-PCR was performed using the LightCycler (Roche Diagnostics Canada, QC, Canada) and normalized to β-actin mRNA to control the amount of template in the reaction mixtures. The primer sequences are shown in Table 1.
The cells were lysed in RIPA buffer (MilliporeSigma, Canada). Total protein was collected and the concentration was measured by using BCA kit (Thermo Fisher Scientific Inc., IL, USA). ELISAs for human FN and IL6 were performed using a commercially available kits (R&D Systems and Millipore Corporation, USA.) according to the manufacturer’s instructions.
For luciferase reporter assay, a human FN 3’-UTR segment was amplified by PCR and inserted into the pMIR REPORT Luciferase vector with CMV promoter (Life Technologies, CA, USA) by using the Sac I and Hind III sites immediately downstream from the stop codon of luciferase. The following sets of primers were used to generate specific fragments for human FN 3’-UTR, forward primer, 5’-AGAGCTC TCATCTTTCCAATCCAGAGGAAC-3’; reverse primer, 5’-TCAAGCTT TAATCACCCACCATAATTATACC-3’ (underlined sequences indicate the endonuclease restriction site). Nucleotide substitutions were introduced by PCR to yield a mutated binding site. The primer sequences for human FN 3’UTR mutation cloning are listed in Table 1. The sequence of the cloned product was confirmed by sequencing analysis. The pMIR-FN 3’UTR, miR-9 mimic, and β-galactosidase control plasmids were then co-transfected into HEK293A cells for 24 hours. After transfection, luciferase activity was measured using the Dual-Light Chemiluminescent Reporter Gene Assay System (Life Technologies, CA, USA) and Chemiluminescent SpectraMax M5 (Molecular Devices, Sunnyvale, CA) following the manufacturer’s instructions [8, 30]. Luciferase activity was normalized for transfection efficiency by measuring β-galactosidase activity. The experiments were performed in triplicates.
Cell lysates from HCMECs cultured in NG or HG were used for immunoprecipitation using the Magna RIP RNA-binding protein immunoprecipitation kit (Millipore, ON, Canada) following the manufacturer’s instructions. Anti-IgG (control) and anti-EZH2 antibodies (Millipore, Canada) were used to co-precipitate the RNA-binding proteins of interest. The extracted RNAs were then reverse transcribed to cDNA, and analyzed by RT-PCR.
ChIP assays (Milipore, CA, USA) were carried out as previously described [29]. Briefly, HCMECs cultured in NG or HG were collected for immunoprecipitation. Cells were then fixed with 1% formaldehyde and then lysed. ChIP assays were performed using anti-IgG and anti-EZH2 (Millipore, Canada) antibodies. Anti-mouse IgG was used as a negative control. The immunoprecipitated DNA was detected by RT-qPCR using promoter-specific primers for miR-9 promoter region: forward: 5’-GAAATGGGACTGTGACTCCTAC-3’, reverse: 5’-AGAGGATACAAGAGGAGGAGAG-3’ [31].
The mouse heart tissues were collected, fixed and embedded in paraffin and cut it to 5 μm thickness section on positively charged slides. The sections were deparaffinized in xylene and stained with hematoxylin and eosin, or with Masson’s trichrome stain for evaluation as described previously [27, 32].
Data were expressed as mean ± standard error (SEM) and statistical significance of results were analyzed by ANOVA and Student’s t-test as appropriate. A p value of 0.05 or less (p < 0.05) was considered significant between 2 groups. The results were expressed as average of n = 6–8 animals per group.
As glucose induced alterations of endothelial cells is the key initiating factor for tissue damage in diabetes [33], we examined HCMECs for glucose induced changes. Initially we confirmed our previous finding that glucose-induced FN overexpression peaks at 48 hrs (not shown). Hence, we used this time point for all subsequent experiments. We first studied the expression of miR-9 in HCMECs after exposure to HG. miR-9 expression was decreased after treatment with HG for 48 hours (Fig 1A). In parallel, expressions of ECM proteins’ mRNA (FN, collagen 4α1, Collgen1α1,) were increased after exposure to HG at the same time point (Fig 1B–1D). Changes in mRNA expressions were mirrored by protein expression levels (Fig 1E). To establish a cause–effect relationship we examined the effect of miR-9 overexpression. miR-9 mimics transfection prevented glucose-induced upregulation of ECM protein production on both the mRNA and protein levels (Fig 1B–1E). Similarly, glucose induced augmented expression of inflammatory mediators (IL-1β, IL6) and transcription factor (NF-κB) were also prevented by miR-9 mimic transfection (Fig 1F–1H). To examine a direct regulatory relationship of miR-9 on its target gene, FN 3’UTR fragment was cloned to pMIR vector, after co-transfection with miR-9 mimics in to HEK293 cells for 24 hours, luciferase reporter activity was found to be significantly decreased in the wild type FN 3’UTR transfection group compared to the mutant FN 3’UTR transfection group (Fig 1I). These findings further supported a direct interaction between miR-9 and FN 3’UTR. No alteration in cell viability were seen in MTT assay (not shown).
As previously mentioned, the lncRNA ZFAS1 may regulate miRNA action. Hence, we examined whether nullifying the effects of ZFAS1 has any effect on miR-9 and the downstream targets. We first confirmed that high glucose upregulates ZFAS1 RNA expression (Fig 2A). We found that glucose-induced mRNA overexpression of ECM proteins including FN, collagen 4a1 was prevented by ZFAS1 silencing via siRNA transfection (Fig 2C and 2D), in association with correction of glucose induced reduced miR-9 levels (Fig 2B). We further measured the expressions of the pro-inflammatory factors (NF-κB, IL-6 and IL-1β). The expressions of these pro-inflammatory factors were also corrected following ZFAS1 silencing (Fig 2E–2G). Such changes were further reflected in IL-6 protein levels (Fig 2H). We further explored whether ZFAS1 acted through interactions with miR-9 by carrying out a rescue experiment. For this experiment, we transfected HCMECs using ZFAS1 siRNA, then transfected using miR-9 antagomir. We found that such approach rescued ZFAS1 siRNA-mediated suppression of the aforementioned factors at mRNA and protein levels (Fig 2C–2G). These data indicate that lncRNA ZFAS1 mediates its action through miR-9 in the current scenario.
To further understand the mechanistic actions of ZFAS1 and miR-9, we used in vitro experiments. It has been shown that ZFAS1 and polycomb repressive complex 2 (PRC2, a histone methyltransferase) have a regulatory relationship in other systems [34]. Hence, we examined such relationship in this context. We used a global histone methylation inhibitor, 3-deazaneplanocin A (DZNep). PRC2 has methyltransferase activity and is one of two classes of polycomb-group proteins. PRC2 has three subunits: EZH2, EED and SUZ12 [35, 36]. DZNep corrected glucose-mediated augmentation of two PRC2 components namely, EZH2, EED and SUZ12 mRNAs (Fig 3A–3C) and recovered glucose induced miR-9 downregulation (Fig 3D). We then tested the effects of ZFAS1 on polycomb repressive complex 2 (PRC2), by using ZFAS1 siRNA to silence ZFAS1. Silencing of ZFAS1 led to the repression of enhancer of zeste homolog 2 (EZH2) and EED, but not SUZ12 (Fig 3E–3G). Subsequently, we performed RIP and ChIP analyses using anti-EZH2 antibodies and found that ZFAS1 binding to EZH2 (PRC2) was enhanced under high glucose conditions (Fig 3H). Furthermore ChIP analysis showed enhance interaction of EZH2 (PRC2 complex)-ZFAS1 with miR-9 promoter in high glucose (Fig 3I). Collectively these analyses indicate that ZFAS1 regulates miR-9 through PRC2 complex.
We then explore whether such miR-9 mediated mechanisms play a role in a clinically relevant model of DCM. Fibrosis is one of the characteristics in DCM. To examine the potential role of miR-9 in this process, we generated miR-9 endothelial specific transgenic mice. We isolated various cells from mouse hearts, and confirmed miR-9 overexpression in the cardiac endothelial cells, but not in the fibroblasts and myocytes (Fig 4A). Diabetic animals showed hyperglycemia, glycosuria compared to control animals, these parameters were not affected by miR-9 overexpression. Clinical data have been depicted in Table 2. We then examined expressions of miR-9 and ZFAS1 in the mouse hearts and ECs from mouse heart (Fig 4B–4D). We found that the expression of miR-9 was downregulated and ZFAS1 was upregulated in the hearts of diabetic mice compared to control mice. ECM protein expressions (FN and Col1) were also examined in cardiac tissues. mRNA expressions of both FN and Col1 were increased in the hearts of wild-type mice with diabetes (Fig 4E and 4F). All such abnormalities were prevented in the miR-9 TG mice with diabetes (Fig 4E and 4F). In parallel, focal myocardial fibrosis was present in the hearts of wild-type diabetic animals (Fig 4H) and were prevented in the transgenic mice with diabetes (Fig 4J).
Our study demonstrated that endothelial derived miR-9 plays a key role in the pathogenesis of DCM. We have demonstrated that it regulates glucose induced increased ECM protein and inflammatory molecule production in the ECs. Using an EC specific miR-9 TG model, we further showed that such EC derived miR-9 regulate cardiac fibrosis in diabetes. In addition, we demonstrated a novel pathway in which lncRNA ZFAS1 regulates miR-9, to produce effects of glucose through PRC2 complex. Hyperglycemia-induced ECM overproduction could alter the structure and function of the heart in diabetes [37]. Studies have shown that several miRs play important roles in endothelial to mesenchymal transition (EndMT) in DCM [38]. Our research has previously demonstrated that miR-200b regulates diabetes-induced EndMT in the heart and retina [28]. We have also shown that miR-146a influences cardiac fibrosis in diabetes [7]. Others have shown a protective role of miR-30c in cardiac metabolism in diabetes [39]. We have also shown that other miRs, including miR-1 and miR-133 are protective of cardiac hypertrophy in diabetes [40, 41]. Our present and previous studies have demonstrated that regulation of miRs in endothelial cells, the primary target of glucose-induced vascular damage in diabetes, can prevent cardiac structural and functional changes through the anti-inflammatory mechanism [7, 27]. These findings indicate that endothelial cell dysfunction is probably a primary target of glucose induced cardiac damage in diabetes. MiRs possess a wide range of functions and are involved in most, if not all, chronic disease processes. MiR-9 is evolutionary conserved and play roles in various cellular activities in a context-dependent manner [42]. miR-9 is known to target important ECM proteins (fibronectin [FN], collagen [COL]) and their regulator (TGFβ1), multiple inflammatory mediators (IL-6, IL-1β, TNFα) and NF-κB—key molecules in DCM [43]. Hence, miR-9 may act as a common mediator that regulates cardiac inflammation and subsequent fibrosis in DC. Preliminary studies have demonstrated downregulation of miR-9, along with upregulation of inflammatory mediators, transcription factors and ECM proteins in ECs incubated with HG and in the hearts of diabetic animals (Fig 1). Another interesting part of the study is the identification of a novel regulatory mechanism. In this study, we have further confirmed that the lncRNA ZFAS1 regulates miR-9 through the PRC2. This is in keeping with similar regulation of other miRs. It is however interesting to note that silencing of ZFAS1 led to prevention of glucose induced upregulation of EZH2 and EED, but not SUZ12. Exact reasons for such findings are not clear. Further investigations may be needed to understand such phenomenon further. LncRNAs regulate gene expression at the epigenetic, transcriptional and translational levels in a variety of ways. Histone methylation and acetylation are implicated good epigenetic marks in diabetic complications [44, 45]. Histone methylation is a process by which methyl groups are transferred to amino acid residues of histone proteins. EZH2, the catalytic core subunit of PRC2, acts as an epigenetic silencer through the trimethylation of lysine 27 on histone H3. Moreover, EZH2 is implicated in promoting tumour angiogenesis [46, 47]. DZNep was reported to selectively inhibit trimethylation of lysine 27 on histone H3 (H3K27me3) and lysine 20 on histone H4 (H4K20me3) as well as reactivate silenced genes in cancer cells [48]. In our experiments, increased levels of ECM protein were accompanied by EZH2 upregulation in HCMECs. Histone methylation blockade resulted in reduction in ECM mRNA expression, showing histone methylation is associated with HG induced increased ECM. Although there are no previous reports on ZFAS1 alterations in diabetic cardiomyopathy, recent studies have shown that other lncRNAs play an important role in cardiac fibrosis. One study indicated that H19 negatively modulated the expression of DUSP5 gene in cardiac fibroblasts (CFs) and fibrotic tissues [49]. The study showed that the lncRNA myocardial infarction associated transcript (MIAT) is involved in myocardial infarction (MI). When MIAT is upregulated, it is accompanied by miR-24 down-regulation and TGF-β1 up-regulation [50]. The expression of MIAT was also significantly upregulated in mouse CFs treated with Ang II [50]. Our previous study showed that ANRIL knock out can prevented the elevated expressions of extracellular matrix (ECM) products in the tissues of heart and kidney in diabetic animals, and we also demonstrated the interaction of ANRIL with EZH2 (PRC2) complex in the regulation of VEGF [26]. MALAT1 also played an important role in the pathogenesis of chronic diabetic complications involving the heart and kidneys [51]. LncRNA ZFAS1 is highly expressed in the heart and is a regulator of organ development, cancer growth and metastasis, apoptosis and cell cycle regulation [52, 53]. ZFAS1 was originally discovered to play a vital role in hepatocellular carcinoma progression; there ZFAS1 may be a potential tumor suppressor [54]. ZFAS1 is located in the 20q13.13 region, which functions in breast cancer progression [55]. LncRNA ZFAS1 has also been reported to regulate cell proliferation, migration and invasion in bladder cancer by targeting miR-193a-3p/SDC1 [56]. However, the biological function and molecular mechanisms of ZFAS1 in diabetes remained unclear. Therefore, understanding the effects of ZFAS1 on diabetes can increase the essential knowledge and provide novel way to the diagnostic and treatment of diabetic patients. In our study, we found that ZFAS1 regulates miR-9 by binding to the PRC2 complex, subsequently regulating expressions of ECM and proinflammatory factors in DCM. However exact binding site of the PRC complex has not been characterised and need further investigation. In summary, we showed that glucose causes upregulation of ZFAS1 in the human cardiac microvascular endothelial cells and in the heart in diabetic animals. This upregulation is responsible for altered the expressions of ECM protein production and proinflammatory factors. ZFAS1 promoted ECM protein production by regulating PRC2 components and inhibiting miR-9 expression. Data from this study shed light on a potentially novel method to prevent DCM using an RNA based approach. A schematic of the regulatory process as observed in this study is outlined in Fig 5.
Click here for additional data file. Click here for additional data file. Click here for additional data file. | true | true | true |
||
PMC9565477 | Asmaa Mohammed,Olfat G. Shaker,Mahmoud A. F. Khalil,Mohammed Gomaa,Shaimaa A. Fathy,Abeer K. Abu-El-Azayem,Amira Samy,Mahmoud I. Aboelnor,Mohamed S. Gomaa,Othman M. Zaki,Randa Erfan | Long non-coding RNA NBAT1, TUG1, miRNA-335, and miRNA-21 as potential biomarkers for acute ischemic stroke and their possible correlation to thyroid hormones 10.3389/fmolb.2022.914506 | 30-09-2022 | long non coding RNA,NBAT1,TUG1,miRNA-335,miRNA-21,ischemic stroke | Objective: RNA-based mechanisms of epigenetic modification related to acute ischemic stroke (AIS) have been widely studied recently. The current work aimed to determine the potential roles of four ncRNAs (TUG1 and its target miR-21, NBAT1, and miR-335) as promising diagnostic biomarkers in AIS as well as their involvement in the disease pathogenesis. Methods: The levels of the studied lncRNAs and miRNAs were measured in the serum for two different groups, including patients with AIS (60) and healthy controls (60). All individuals were subjected to a full history investigation and clinical examination. Blood samples were tested for FBS, 2HPP, TAG, HDL, LDL, TSH, T3, and T4 levels. Results: The serum levels of TUG1 were significantly increased in AIS patients compared to control subjects. It is worthwhile to note that serum TUG1 levels were positively correlated with cholesterol, triglycerides, LDL, carotid IMT (Intima-media thickness), and miR-21, while they were negatively correlated with HDL levels. Our study showed that NBAT1 serum expression levels were elevated in AIS patients compared to controls. NBAT1 expression levels were observed to be positively correlated with triglycerides, TUG1, and miR-21. NBAT1 could distinguish between AIS patients and controls with a sensitivity of 100% and specificity of 100% at a cut-off point of 1.45. Regarding miR-335, we found that its expression levels were downregulated in AIS patients compared with healthy controls. It could distinguish between AIS patients and controls with a sensitivity of 73.3% and a specificity of 100% at a cut-off point of 0.796. Conclusion: Our results revealed that serum TUG1, miR-21, NBAT1, and miR-335 could be promising molecular diagnostic markers for AIS as these biomarkers could discriminate between AIS patients and healthy controls. | Long non-coding RNA NBAT1, TUG1, miRNA-335, and miRNA-21 as potential biomarkers for acute ischemic stroke and their possible correlation to thyroid hormones 10.3389/fmolb.2022.914506
Objective: RNA-based mechanisms of epigenetic modification related to acute ischemic stroke (AIS) have been widely studied recently. The current work aimed to determine the potential roles of four ncRNAs (TUG1 and its target miR-21, NBAT1, and miR-335) as promising diagnostic biomarkers in AIS as well as their involvement in the disease pathogenesis. Methods: The levels of the studied lncRNAs and miRNAs were measured in the serum for two different groups, including patients with AIS (60) and healthy controls (60). All individuals were subjected to a full history investigation and clinical examination. Blood samples were tested for FBS, 2HPP, TAG, HDL, LDL, TSH, T3, and T4 levels. Results: The serum levels of TUG1 were significantly increased in AIS patients compared to control subjects. It is worthwhile to note that serum TUG1 levels were positively correlated with cholesterol, triglycerides, LDL, carotid IMT (Intima-media thickness), and miR-21, while they were negatively correlated with HDL levels. Our study showed that NBAT1 serum expression levels were elevated in AIS patients compared to controls. NBAT1 expression levels were observed to be positively correlated with triglycerides, TUG1, and miR-21. NBAT1 could distinguish between AIS patients and controls with a sensitivity of 100% and specificity of 100% at a cut-off point of 1.45. Regarding miR-335, we found that its expression levels were downregulated in AIS patients compared with healthy controls. It could distinguish between AIS patients and controls with a sensitivity of 73.3% and a specificity of 100% at a cut-off point of 0.796. Conclusion: Our results revealed that serum TUG1, miR-21, NBAT1, and miR-335 could be promising molecular diagnostic markers for AIS as these biomarkers could discriminate between AIS patients and healthy controls.
Ischemic stroke (IS) is one of the leading causes of disability and the second leading cause of death globally, after ischemic heart disease (Benjamin et al., 2019). IS is caused by an embolus or thrombus occluding a cerebral artery, reducing oxygen and blood flow to the brain. When this happens repeatedly, it leads to irreparable brain damage because of neuronal cell death. Another typical hallmark of IS is increased oxidative stress and inflammation (Chehaibi et al., 2016). The most common modifiable risk factors for stroke include diabetes mellitus and hypertension. Both can cause dyslipidemia and atherosclerosis (Johansson, 1999; Tuomilehto and Rastenyte, 1999). Another mechanism for hyperglycemia-related stroke risk is the increased brain lactate production and hypoperfusion in the infarction area. Hyperglycemia (blood sugar levels above 140 mg/dl) also reduces early blood flow restoration (Cai et al., 2020). Similarly, thyroid hormones also vary in the serum of patients with acute stroke (Marouli et al., 2020). In addition to modifiable risk factors, increasing evidence suggests that epigenetics contributes to the etiology of cardiovascular diseases, specifically stroke susceptibility (Morgado-Pascual et al., 2018). MicroRNAs (miRNAs) and long non-coding RNAs (lncRNAs) are two RNA-based epigenetic regulatory mechanisms that are particularly noteworthy. Recently, lncRNAs (>200 nucleotides) have gained increased interest. LncRNAs can use both transcriptional and post-transcriptional pathways to control the expression of target genes (Chen et al., 2018). LncRNAs are involved in regulating the pathological process of ischemic stroke by altering atherosclerosis, inflammation, cell survival, and angiogenesis (Dykstra-Aiello et al., 2016). MiRNAs are small single-stranded non-coding RNAs of 18–22 nucleotides. MiRNAs have been linked to neuronal development, synaptic plasticity, differentiation, metabolism, proliferation, and neurodegenerative disorders (Bartel, 2009). Nevertheless, several clinical diseases that contribute to IS, such as atherosclerosis, dyslipidemia, and inflammation, have been linked to changes in miRNA levels (Rink and Khanna, 2011). Furthermore, changes in several phases of stroke had different miRNA levels, implying that they could be used for diagnostic, therapeutic, and prognostic purposes (Mirzaei et al., 2018). LncRNA taurine up-regulated gene 1 (TUG1) is a 7.1-kb lncRNA up-regulated by taurine (Khalil et al., 2009). TUG1 plays a role in developing non-small cell lung cancer, osteosarcoma, and bladder cancer (Zhang et al., 2013; Zhang et al., 2014; Liu et al., 2017). It is also linked to the development of atherosclerosis through modulating the miR-21/phosphatase and tensin homolog (PTEN) axis (Li et al., 2018). That is why both TUG1 and miR-21 can be implicated in AIS. LncRNA Neuroblastoma-associated transcript1 (NBAT1) is a lncRNA that plays a role in carcinogenesis, and earlier reports confirmed NBAT1 dysregulation in various types of cancer (Yan et al., 2017). Polycomb repressive complex 2 (PRC2) signaling is one of the routes through which NBAT1 promotes the genesis and growth of tumors (Hu et al., 2015). It is worth mentioning that NBAT1 functional interaction with EZH2 (Enhancer of Zeste Homolog 2), a member of the PRC2 complex, plays a critical role in neurogenesis (Pandey et al., 2014) and causes suppressed expressions of its target genes, which are implicated in cell proliferation and cell migration (Pandey and Kanduri, 2015). Besides, NBAT1 can regulate gene expression of DKK-1 (Dickkopf-1) through modulating the functions of PRC2 as overexpression of NBAT1 represses EZH2 functions, which results in increased DKK1 levels (Pandey and Kanduri, 2015). Increased Dkk-1 expression is involved in the atherosclerosis process. High serum Dkk-1 levels have been consistently reported in patients with ischemic stroke (He et al., 2016). Moreover, Zhu et al. (2019), found that Dkk-1 may be a potential prognostic biomarker for ischemic stroke. Recent research has related miR-335 to neural growth and development. Si and co-workers discovered a decrease in miR-335 expression in the adult rat brain after acute middle cerebral artery occlusion, leading to AIS-enhanced apoptosis (Si et al., 2019). On the other hand, the function of miR-335 in humans is uncertain. The current study target is to estimate levels of four ncRNAs (TUG1 and its target miR-21, NBAT1, and miR-335) in AIS patients. Additionally, this study investigated the possibility of utilizing these ncRNAs as diagnostic and prognostic biomarkers in AIS patients.
The current study protocol was authorized by Fayoum University Hospital’s local ethics committee. All procedures were carried out according to the Declaration of Helsinki. This report was performed following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) principles (Moher et al., 2012).
This is a case-control study of 60 Egyptian individuals diagnosed with AIS within the first 24 h. Patients were chosen from the Neurology Departments at Fayoum University Hospitals. Patients were diagnosed using World Health Organization (WHO) guidelines, and symptoms appeared within 24 h (Hatano, 1976). Patients with a history of ischemic stroke, epilepsy, intracerebral hemorrhage, chronic kidney diseases, tumors, neurological diseases, and liver diseases were excluded. Moreover, patients with congestive heart failure, renal insufficiency, malignant tumors, severe edema, systemic infections, febrile disorders, recent surgery or trauma within the last 2 months, and autoimmune diseases were excluded. As a control group, 60 participants were involved in this study. They were similar in age, sex, and traditional vascular risk factors to the patients’ group. They sought medical assistance for a headache or spondylosis at the neurology clinic.
The participants were examined regarding their medical history, a general and comprehensive neurological examination, and a score on the National Institute of Health Stroke Scale (NIHSS) ranging from 0 to 42, with higher values indicating more severe neurologic damage (Brott et al., 1989). To confirm the diagnosis, all patients had magnetic resonance imaging (MRI) with a diffusion scan (stroke protocol) of the brain at Fayoum University Hospital’s Radiology Department.
All the patients had their carotid arteries sonographically examined. They were positioned in a supine or semi-supine position, with their heads somewhat hyperextended and rotated 45° away from the side targeted by the exam. A lower frequency linear transducer (7 Megahertz) was used for the Doppler exam, while a higher frequency linear transducer (>7 Megahertz) was used to assess intima-media thickness and plaque morphology. A greyscale picture was used to evaluate intima-media thickness (carotid IMT) at the far wall of the common carotid artery, bulb, and internal carotid artery. The measurement included the media (echo-poor layer) and intima (echogenic layer). A thickness less than 1 mm in the intima-media was considered normal.
Erythrocyte sedimentation rate (ESR), serum C-reactive protein (CRP) level, Complete blood count (CBC), and liver and kidney function test to rule out the fasting blood glucose level, presence of metabolic or systemic disorders, and finally lipid profile were all performed as part of routine laboratory investigations. TUG1, miR-21, NBAT1, and miR-335 were examined for their relative expressions. Each subject’s blood sample was collected (10 ml) using a vacutainer apparatus. The collected blood samples were placed in tubes with separator gels stuck between the serum layer (top) and the packed cells and were left to clot for 15 min before centrifuging at 4,000 ×g for 10 min. The serum was extracted from clotted whole blood and stored at −80°C until it was employed in the RNA extraction process (Bush et al., 2001).
A total sample volume of 100 μl serum was used for RNA extraction, and the extraction was carried out utilizing a miRNeasy extraction kit (Qiagen, Valencia, CA, United States). First, 500 ml QIAzol lysis reagent was added to the reaction mixture and incubated for 5 min at room temperature. Then, chloroform (100 μl) was added; the mixture was vortexed for 15 s and incubated for 2–3 min at room temperature. Then, centrifugation was done at 12,000 ×g at 4°C for 15 min. After removing the top aqueous phase, 1.5 times its volume of 100% ethanol was added. Then, we centrifuged each 700 μl of this mixture at 8,000 ×g for 15 s at room temperature on an RNeasy Mini spin column in a 2 ml collection tube. RW1 buffer (700 μl) was added to each column after the mixture had entirely passed and centrifuged at 8,000 ×g for 15 s at room temperature. The column was then filled with 500 μl buffer RPE and centrifuged at 8,000 ×g for 15 s at room temperature. Then, another 500 μl of buffer RPE was added to the column and centrifuged at room temperature for 2 min at 8,000 ×g. the column was placed in a new 1.5 ml collection tube and centrifuged for 2 min at 8,000 ×g. Finally, we pipetted 50 μl RNase-free water straight onto the column, followed by centrifugation for 1 min at 8,000 ×g to elute the RNA. The sample was treated with DNase after extraction to remove any remaining DNA before being reverse transcribed into cDNA with the DNase Max Kit (Qiagen, Valencia, CA, United States). Then, a NanoDrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA, United States) measured the RNA at 260/280 nm.
A high-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA, United States) was used to reverse transcript 1 μg RNA in a 10 μl final reaction volume, according to the manufacturer’s instructions (incubated for 60 min at 37°C, for 5 min at 95°C, and then maintained at 4°C).
QPCR primers and the miScript SYBR Green PCR kit (Qiagen) were used for performing the Quantitative PCR (qPCR). The levels of miR-335 and miR-21 gene expression were examined using SNORD 68 as an internal control (Ayeldeen et al., 2018). GAPDH, which was frequently employed as an internal control for serum lncRNAs in various studies, was used as an endogenous control for analyzing TUG1 and NBAT1 per the manufacturer’s method (Shaker et al., 2017). Table 1 lists the primer sequences utilized for each of the genes investigated. The PCR cycling condition protocol was performed as follows: 95°C for 10 min, then 40 cycles of 15 s at 95°C and 60 s at 60°C. Using the Rotor gene Q System, the operation was carried out on a 20 μl reaction mixture (Qiagen). Target genes were quantified concerning their endogenous control using the cycle threshold (Ct) approach. By subtracting the Ct values of SNORD 68 from miR-335 and miR-21, the ΔCt of microRNAs was computed. ΔCt of lncRNAs was estimated by subtracting GAPDH Ct values from TUG1 and NBAT1 Ct values. The equation 2−ΔΔCt was utilized to calculate the expression levels of miR-335, miR-21, NBAT1, and TUG1 (Livak and Schmittgen, 2001). The fold change values for the control group were set to 1. A negative or down-regulation is indicated by a fold change value less than 1, while a positive or up-regulation is indicated by a fold change value greater than 1 (Santoro et al., 2016). The four biomarkers’ fold changes were studied, and their diagnostic utilities in AIS patients were the primary parameters. The relationship between fold changes in the examined biomarkers and clinical disease activity and clinical presentation was one of the secondary outcomes.
All data were collected and analyzed using SPSS version 22 (SPSS Inc., Chicago, IL, United States). The Chi-square test (for two or more groups) was used to analyze qualitative data, and the results were reported as numbers and percentages. An independent t-test (for two independent groups) and a one-way ANOVA (for two or more independent groups) were used to produce quantitative data in the form of standard deviation. Each research group’s quantitative data were tested for normality using the One-Sample Kolmogorov-Smirnov test, followed by inferential statistical analyses. The link between the variables was determined using a bivariate Pearson correlation test. The specificity and sensitivity of the studied variable were determined using a ROC curve analysis. P-Values of less than 0.05 were statistically significant.
The clinical and demographic characteristics of study groups are illustrated in Table 2. With a p-value of >0.05, the table revealed no significant difference in medical history, age, or sex between groups. There was a significantly higher triacylglycerides (TAG) mean and a lower HDL mean among cases with a p-value <0.001. The mean carotid IMT among cases was (0.879 ± 0.24), and the mean NIHSS score was (12.88 ± 4.6). For the thyroid profile, the mean TSH was 4.47 ± 5.7 uI U/ml, with a mean free T3 of 2.62 ± 1.31 pg/ml and a mean free T4 of 1.84 ± 0.36 ng/dl.
When compared to the control group, TUG1 and NBAT1 were strongly expressed in the serum samples of AIS patients. The fold change median of TUG1 was 1.42 in patients compared to controls with p < 0.001. The fold change median of NBAT1 was 12.5 in patients compared to controls with p < 0.001 (Table 3). Overexpressed miRNA-21 was demonstrated between the patients and control groups with a fold change median of 8.2. On the other hand, there was downregulation of the expression levels of miRNA-335 with a fold change median of 0.24 (Table 3) (Figures 1A–D).
Regarding the median of different biomarkers, no significant difference was observed between patients with comorbidities [diabetes mellitus (DM) and hypertension (HTN)], p > 0.05, as shown in Table 4.
There was a significant positive association between carotid IMT and NIHSS score with a p-value <0.05 (Figures 2, 3). Unlikely, we observed a negative correlation among cases with free T3 levels. Unlikely, no significant correlation was observed (p-value >0.05) in other thyroid profiles and biomarkers (Table 5).
Serum TUG1 levels were positively correlated with cholesterol, triglyceride, LDL, free T4, and carotid IMT, in addition to NBAT1 and miRNA-21, while negatively correlated with HDL levels. Additionally, serum NBAT1 levels were positively correlated with triglycerides, TUG1, and miRNA-21, while negatively correlated with HDL levels. Considering miRNA-21, there was a positive correlation between it and TAG, TUG1, and NBAT1, while a negative correlation was found with HDL levels. Finally, a positive correlation was found between miRNA-335, carotid IMT, and TSH levels, while a negative correlation was found with cholesterol levels, as shown in Table 6.
Biomarker tests in the diagnosis of AIS revealed good sensitivity and specificity in NBAT1, followed by miRNA-21 and miRNA-335 with a sensitivity of (100%, 93.3%, and 73.3%) and a specificity of (100% for all) at cut-off values of (1.45, 1.25, and 0.796), respectively. TUG1 showed high sensitivity (80%) and very low specificity (8.3%) (Table 7) (Figures 4A–D).
When cerebral blood flow is disrupted, oxygen and glucose availability to brain cells is reduced, resulting in AIS. Atherosclerosis in major intracranial arteries is one of the most common causes of stroke worldwide (Gorelick et al., 2008). The current study aimed to estimate the levels of four ncRNAs: TUG1 and its target miR-21, NBAT1, and miR-335 to determine their potential use as diagnostic and prognostic biomarkers in AIS patients and their possible relation to stroke-related risk factors and the patient’s thyroid profile. TUG1 levels in the blood were substantially higher in AIS patients than in controls. TUG1 was also able to distinguish between AIS patients and control subjects, with a sensitivity of 80% and an AUC of 73.3%. It is noteworthy that serum TUG1 levels were positively correlated with cholesterol, TAG, LDL, carotid IMT, and miR-21 and negatively correlated with HDL levels. However, TUG1 was not significantly correlated with the NIHSS score of the patients and was not associated with comorbidities (DM or HTN). This is consistent with previous work by Wang et al. (2019), who found that TUG1 microglia were involved in neuroinflammation following an ischemic stroke. TUG1 was increased in microglial cells and behaved as a sponge for miR-145a-5p. TUG1 silencing shifted the microglia phenotype (from M1 to M2) and reduced pro-inflammatory cytokines, promoting the production of anti-inflammatory cytokines, and improving cell survival. The deletion of TUG1 also inhibited the activation of the nuclear factor- κB (NF-κB) pathway produced by oxygen-glucose deprivation (OGD)/R. As a result, TUG1 is thought to synchronize microglia and inflammatory cytokine assembly immediately after an OGD insult. Besides, neuron apoptosis represents one of the primary causes of AIS in patients. Apoptosis in AIS can be influenced by both downregulation and upregulation of a specific ncRNA. TUG1 promotes apoptosis in neurons by sponging miR-9 and increasing Bcl2l11 expression (Chen et al., 2017). Moreover, the mentioned positive correlation between TUG1 and cholesterol, TAG, LDL, and carotid IMT confirmed the direct relation of this biomarker to atherosclerosis which is one of the most direct risk factors of AIS. Therefore, the current study agrees with the study of Zhang and co-workers, who detected that by regulating fibroblast growth factor1 (FGF1) via miR-133a, TUG1 knockdown suppressed hyperlipidemia (Zhang et al., 2018). We found that miR-21 was a target gene of TUG1 by Starbase prediction. Moreover, our findings are consistent with Li et al. (2018), who found that TUG1 was more expressed in patients with atherosclerosis than in healthy volunteers. Their study proved that TUG1 expression level was reversely correlated with PTEN expression in patients with atherosclerosis through competing with PTEN for miR-21 binding. So, they considered that TUG1 could be a potential target for treating atherosclerosis. They also discovered that miR-21 expression was positively linked with the expression of TUG1 and that downregulating TUG1 significantly decreased miR-21 expression. Our results agree with the previously mentioned study, as upregulation of miR-21 was demonstrated in the patients, and we found a positive correlation between miR-21 and TUG1. Furthermore, miR-21 and TAG positive association, in addition to miR-21 and HDL negative correlation, have been related to metabolic syndrome and implicated in the proliferation and development of human adipose tissue-derived mesenchymal stem cells (hASC) (Kim et al., 2012). A substantial correlation between plasma miR-21 and IMT was described by Fontanella et al. (2021) as a surrogate marker for early atherosclerosis. Chen et al. (2019) also discovered that angiotensin II-induced angiogenic sprouting in human microvascular endothelial cells (HMECs) was linked to the STAT3/miR-21 pathway. Their findings suggest that targeting the STAT3/miR-21 axis in conjunction with current atherosclerosis therapies could be effective. We conclude from the above that TUG1 and its target miR-21 have a possible relation to the pathogenesis of AIS and atherosclerosis which is one of the most important risk factors for AIS. To the best of our knowledge, this is the first report to investigate the levels of NBAT1 in AIS patients. The levels of NBAT1 in AIS patients were higher than in controls. Also, serum NBAT1 levels were positively correlated with TAG, TUG1, and miR-21 while negatively correlated with HDL levels. Moreover, it was found that NBAT1 could distinguish between AIS patients and control subjects with a sensitivity of 100% and a specificity of 100% at a cut-off point of 1.45. The specific role of NBAT1 in AIS pathogenesis is unclear but it may act through regulation of DKK-1 as overexpression of NBAT1 represses EZH2 functions, which results in increased DKK-1 levels (Pandey and Kanduri, 2015). Increased Dkk-1 expression is involved in the atherosclerosis process. High serum Dkk-1 levels have been consistently reported in patients with ischemic stroke (He et al., 2016). Moreover, Zhu et al. (2019), found that Dkk-1 may be a potential prognostic biomarker for ischemic stroke. The present study demonstrated that miR-335 expression was lower in AIS patients than in control subjects. It could distinguish between AIS patients and controls with a sensitivity of 73.3 percent and a specificity of 100 percent at a cut-off point of 0.796. These findings agree with that of Si et al. (2019), who discovered that miR-335 was downregulated during AIS in rat models, related to decreased stress granules (SG) formation, elevated Rho-associated protein kinase 2 (ROCK2) expression, and higher apoptotic levels. They also discovered that treating mice with miR-335 increased SG formation, reduced ROCK2 protein expression and apoptosis, and reduced ischemia-induced infarction. Their findings showed that miR-335 prevented apoptosis and enhanced SG formation in AIS via lowering ROCK2 expression, suggesting that it could be a therapeutic target for brain injury. Besides, we found for the first time a positive correlation between miR-335 and carotid IMT. Zhao et al. (2016) also found that plasma miR-335 levels were lower in AIS and negatively correlated with NIHSS scores, in contrast to plasma Calmodulin (CaM) levels. In patients who have never had intrinsic thyroid disease, changes in thyroid hormone concentrations are frequently related to critical illness. Non-thyroidal illness syndrome is the name for this condition (NTIS, or euthyroid sick syndrome). The most common hormone pattern in NTIS is T3 deficiency with normal thyroid-stimulating hormone (TSH) and thyroxine (T4) levels. The alterations in thyroid hormones can be due to NTIS or disturbances in the hypothalamic-pituitary-thyroid axis (Nagri et al., 2021). In our study, we found a negative correlation between T3 and the NIHSS. This is agrees with Zhang and co-workers, who state that a low FT3 value on admission was associated with stroke severity, subtype, and prognosis (Zhang et al., 2019). Various miRNAs and lncRNAs clearly affected by thyroid hormones and receptors have been identified (Zhao et al., 2019). One of these lncRNAs is TUG1. Lin and co-workers found that T3/TR treatment reduced TUG1 expression in vitro, resulting in the downregulation of alfa fetoprotein (AFP) mRNA in patients with non-hepatitis B/non-hepatitis C HCC (NBNC-HCC) (Lin et al., 2020). Our results showed that TUG1 was positively correlated with T4 in the presence of acute stroke. To our knowledge, no previous study has investigated this correlation in this clinical setting. Our work also found a positive correlation between miR-335 and TSH in AIS. Our study’s relatively small sample size is a limitation, necessitating additional investigations with bigger sample sizes. In conclusion, serum TUG1, miR-21, NBAT1, and miR-335 are promising molecular diagnostic markers for AIS. The correlation of TUG1 and miR-335 with T4 and TSH, respectively, in acute ischemic stroke opens a new channel to study how this correlation may play a role in severe acute illness. | true | true | true |
|
PMC9565569 | Marina Rodrigues Tavares,Kaplan Kirakci,Nikolay Kotov,Michal Pechar,Kamil Lang,Robert Pola,Tomáš Etrych | Octahedral Molybdenum Cluster-Based Nanomaterials for Potential Photodynamic Therapy | 26-09-2022 | polymer carrier,photodynamic therapy,octahedral molybdenum clusters | Photo/radiosensitizers, such as octahedral molybdenum clusters (Mo6), have been intensively studied for photodynamic applications to treat various diseases. However, their delivery to the desired target can be hampered by its limited solubility, low stability in physiological conditions, and inappropriate biodistribution, thus limiting the therapeutic effect and increasing the side effects of the therapy. To overcome such obstacles and to prepare photofunctional nanomaterials, we employed biocompatible and water-soluble copolymers based on N-(2-hydroxypropyl)methacrylamide (pHPMA) as carriers of Mo6 clusters. Several strategies based on electrostatic, hydrophobic, or covalent interactions were employed for the formation of polymer-cluster constructs. Importantly, the luminescent properties of the Mo6 clusters were preserved upon association with the polymers: all polymer-cluster constructs exhibited an effective quenching of their excited states, suggesting a production of singlet oxygen (O2(1Δg)) species which is a major factor for a successful photodynamic treatment. Even though the colloidal stability of all polymer-cluster constructs was satisfactory in deionized water, the complexes prepared by electrostatic and hydrophobic interactions underwent severe aggregation in phosphate buffer saline (PBS) accompanied by the disruption of the cohesive forces between the cluster and polymer molecules. On the contrary, the conjugates prepared by covalent interactions notably displayed colloidal stability in PBS in addition to high luminescence quantum yields, suggesting that pHPMA is a suitable nanocarrier for molybdenum cluster-based photosensitizers intended for photodynamic applications. | Octahedral Molybdenum Cluster-Based Nanomaterials for Potential Photodynamic Therapy
Photo/radiosensitizers, such as octahedral molybdenum clusters (Mo6), have been intensively studied for photodynamic applications to treat various diseases. However, their delivery to the desired target can be hampered by its limited solubility, low stability in physiological conditions, and inappropriate biodistribution, thus limiting the therapeutic effect and increasing the side effects of the therapy. To overcome such obstacles and to prepare photofunctional nanomaterials, we employed biocompatible and water-soluble copolymers based on N-(2-hydroxypropyl)methacrylamide (pHPMA) as carriers of Mo6 clusters. Several strategies based on electrostatic, hydrophobic, or covalent interactions were employed for the formation of polymer-cluster constructs. Importantly, the luminescent properties of the Mo6 clusters were preserved upon association with the polymers: all polymer-cluster constructs exhibited an effective quenching of their excited states, suggesting a production of singlet oxygen (O2(1Δg)) species which is a major factor for a successful photodynamic treatment. Even though the colloidal stability of all polymer-cluster constructs was satisfactory in deionized water, the complexes prepared by electrostatic and hydrophobic interactions underwent severe aggregation in phosphate buffer saline (PBS) accompanied by the disruption of the cohesive forces between the cluster and polymer molecules. On the contrary, the conjugates prepared by covalent interactions notably displayed colloidal stability in PBS in addition to high luminescence quantum yields, suggesting that pHPMA is a suitable nanocarrier for molybdenum cluster-based photosensitizers intended for photodynamic applications.
Photodynamic therapy (PDT) represents a very promising therapeutic modality which employs the light irradiation as external stimuli to activate compounds, so-called photosensitizers (PS), for the treatment of various malignant tumors. The crucial mechanism of this therapy is based on the interaction of a light-activated PS with molecular oxygen dissolved in the intracellular medium, thus producing reactive oxygen species, e.g., singlet oxygen, O2(1Δg), which can damage critical cellular components. Their cytotoxicity affects cell’s DNA, RNA, lipids, and proteins, resulting in tumor cell destruction. Selection of an appropriate PS is one of the most important factors influencing the efficacy of PDT. Various classes of compounds were described as PS for PDT, e.g., porphyrins, phtalocyanines, indocyanines, or bodipy dyes [1,2,3]. Nevertheless, the pharmacokinetics of PS plays a key role in the PDT treatment efficacy. Indeed, several research groups reported the encapsulation of numerous PS compounds to supramolecular systems, including polymer nanoparticles [4,5] and polymer micelles [4,6], and their controlled delivery. In particular, PS can be physically entrapped into nanoparticles by hydrophobic or electrostatic interactions of PS with biodegradable polymers, namely, poly(glycolic acid) [4,7], poly(lactic acid) [4,7], poly(lactic-co-glycolic acid) [8] and poly(ethylene glycol) [9]. Recently, octahedral molybdenum cluster compounds (Mo6) were reported as relevant photo/radiosensitizers for PDT [10,11,12,13,14], X-ray induced PDT [14,15,16,17,18], and photoinactivation of bacteria [11,19,20,21,22]. Mo6 clusters are nanometer-sized metallic aggregates where the distances between the atoms are similar to those found in corresponding bulk metals, evidencing electron delocalization on the whole cluster. For stabilization, the clusters are surrounded by eight strongly bonded inner ligands (Li), generally halogen (Cl, Br, I), and six apical ligands (La) that can be of either organic or inorganic nature to form a cluster denoted [Mo6Li8La6]n. Upon excitation with UV, blue light, or X-rays, these complexes form long-lived triplet states that relax via a broad red-NIR luminescence. This emission is efficiently quenched by oxygen leading to the formation of O2(1Δg) in high yields [23,24]. Even though a previous study showed the phototoxic activity of Mo6 complexes against the human cervical cancer cell line HeLa [25], their limited solubility and/or low stability in physiological conditions, as well as their lack of tumor selectivity, may result in a systemic toxicity, thus restricting their possible use in PDT [26]. To overcome these shortcomings, we hypothesized that the interaction of Mo6 with a suitable hydrophilic and biocompatible polymer carrier based on N-(2-hydroxypropyl)methacrylamide (pHPMA) may improve both their stability in physiological conditions and also pharmacokinetic properties. The polymer conjugation may prolong half-life in circulation, significantly decrease interaction with serum proteins, promote a superior tumor specific accumulation, and reduce possible adverse effects, thus opening possibilities for their real clinical application. Previous studies have shown that PS molecules bound to pHPMA copolymers accumulate in the tumor via the enhanced permeability and retention (EPR) effect [1], in which the principle relies on the tumor tissues defective blood vessels and leaky vasculature combined with the poor lymphatic drainage, ensuring the O2(1Δg) production in the tumor rather than in healthy tissues [2,27]. In this work, we describe the design, synthesis, and thorough evaluation of physico-chemical properties of HPMA-based polymer-Mo6 constructs. We aimed to explore various strategies, such as covalent or non-covalent interactions, to bind Mo6 clusters to the polymer carriers and to optimize their structures for possible future applications in the field of polymer therapeutics, more specifically on tumor-targeted photodynamic therapy. In this study, the term “complexes” is employed for the non-covalent constructs based on electrostatic and hydrophobic interactions, while the term “conjugates” is used for the constructs prepared by the covalent azide-alkyne “click chemistry” (Figure 1). The relationship between the selected coupling strategy and properties of polymer-Mo6 constructs is studied and described.
1,4-Dioxane, 2,2′-azobisisobutyronitrile (AIBN), 2-cyanopropan-2-yl dithiobenzoate (CTA-AIBN), 2-thiazoline-2-thiol, 2,4,6-trinitrobenzene-1-sulfonic acid (TNBSA), 3-azido-1-propylamine, 4,4′-azobis(4-cyanopentanoic acid) (ACVA), 4-cyano-4-(thiobenzoylthio)pentanoic acid (CTA-ACVA), dimethyl sulfoxide (DMSO), methacryloyl chloride, N,N-diisopropylethylamine (DIPEA), N,N-dimethylacetamide (DMA), phosphate buffered saline (phosphate buffer 0.01M and NaCl 0.154M, pH 7.4) (PBS), and t-butanol were purchased from Sigma-Aldrich (Prague, Czech Republic). 2,2′-Azobis(4-methoxy-2,4-dimethylvaleronitrile) (V-70) was from Fujifilm Wako Chemicals Europe (Neuss, Germany). 3-Amino-1-(11,12-didehydrodibenzo[b,f]azocin-5(6H)-yl)propan-1-one (DBCO-NH2) was from Click Chemistry Tools (Scottsdale, AZ, USA). N-(3-tert-butoxycarbonyl-aminopropyl)methacrylamide (APMA-Boc) was purchased from Polysciences, Inc. (Warrington, PA, USA), and 1-aminopropan-2-ol was from TCI Europe (Zwijndrecht, Belgium). All solvents and chemicals were of analytical grade.
The monomers HPMA [28], 3-methacrylamidopropanoylthiazolidine-2-thione (Ma-AP-TT) [29] and cholest-5en-3β-yl 6-methacrylamido hexanoate (MA-AH-cholesterol) [30] were prepared according to the literature. High Performance Liquid Chromatography (HPLC) analysis was employed to verify the purity of monomers and chain transfer agents. Analysis was performed using a Shimadzu HPLC system with a C18 reversed-phase Chromolith Performance RP-18e column and a diode array detector (Shimadzu SPD-M20A), using water/acetonitrile (gradient of 5–95% v/v acetonitrile) as eluent with 5 mL min−1 flow rate.
Statistical copolymers were prepared by reversible addition–fragmentation chain transfer (RAFT) polymerization of HPMA with respective monomers as follows: MA-AP-TT was employed for P0a, P0b, P1, and P5 copolymers; APMA-Boc for P2 [31]; MA-AH-cholesterol for P3; MA-AH-cholesterol and APMA-Boc for P4. The chain transfer agent CTA-AIBN was used for all precursors except for P2, in which CTA-ACVA was employed instead. Reaction conditions were adapted from our previous studies [32], using a mixture of t-butanol and DMA for all precursors except for P2, in which a mixture of water and dioxane was employed. Dithiobenzoate (DTB) groups originating from CTA were removed by reaction with an excess of AIBN as previously described [33]. For P2 and P4, t-butoxycarbonyl (Boc) groups were removed by heating in distilled water at 150 °C for 1 h [34]. Copolymer P1 with COOH groups along the polymer chain was prepared via hydrolysis of thiazolidine-2-thione (TT) groups of the polymer precursor poly(HPMA-co-MA-AP-TT) in phosphate buffer (pH 8.0). Detailed structures of copolymer precursors P1–P5 are shown in Figure 2, and their physico-chemical characterization is described in Table 1. Detailed synthetic procedures and the ratio between monomers, chain transfer agents, and initiators are described in Supplementary Materials. 1H NMR spectra of polymer precursors P3 and P4 are shown in Supplementary Materials.
Previously published procedures were employed to prepare Mo6 cluster compounds: [Mo6I8(OCOC4H8PPh3)6]Br4 (C1) [12], Na2[Mo6I8(OPOPh2)6] (C2) [11], Na2[Mo6I8(cholate)6] (C3) [35], and Na2[Mo6I8(N3)6] (C4) [36]. Schematic structures of clusters C1–C4 are shown in Figure 2. For characterization see Table 2.
Different conjugation strategies were exploited aiming to optimize the constructs’ structure for possible future applications in the field of polymer therapeutics. In this study, the term “complexes” is employed for the non-covalent constructs based on electrostatic (POL1–POL2) and hydrophobic (POL3–POL4) interactions, while the term “conjugates” is used for the constructs using covalent attachment of the Mo6 clusters to the polymer backbone (POL5–POL6). The physico-chemical characterization of polymer-cluster constructs POL1–POL6 is shown in Table 3, and their detailed synthetic procedures are described below.
Solutions of the polymer precursor and the cluster, each in 100 µL of methanol, were prepared separately. In case of cluster solution, addition of 10 µL of DMSO was necessary for complete dissolution. Aliquots of 50 µL from each solution were mixed together and vortexed for 10 min. The reaction mixture was added into 1 mL of distilled water and kept under stirring at room temperature for 1 h. Then, methanol was removed under vacuum, and water was added to adjust the volume to 1 mL. The resulting solutions were used for dynamic light scattering (DLS) and quantum yield (QY) measurements. POL1 was composed of P1 (10 mg, 1.3 µmol of -COOH) and cluster C1 (1.3 mg, 0.32 µmol) while POL2 was composed of P2 (10 mg, 2.35 µmol of -NH2) and cluster C2 (3.5 mg, 1.19 µmol).
The procedure for preparation of POL3 and POL4 was analogous to the one described in 2.5.1., except that the addition of DMSO was not necessary for dissolution of the cluster C3. POL3 was composed of P3 (10 mg) and cluster C3 (2.5 mg) while POL4 was composed of P4 (10 mg) and cluster C3 (2.5 mg).
A solution of P5 (300 mg, 145 µmol of DBCO groups) in dry methanol (2.4 mL) was added into a solution of cluster C4 (49.6 mg, 145 µmol of azide groups) in 0.75 mL of dry methanol under stirring. The reaction mixture was maintained under argon for 30 min; then, it was stirred at 25 °C overnight. Isolation and purification procedures were performed as follows: first, the polymer was precipitated into a mixture of acetone/diethyl ether (1/1) twice. The product was washed with pure diethyl ether and dried under vacuum, affording the polymer-cluster conjugate POL5 (296 mg, 93%). HPLC analysis was performed to control the course of reactions. POL6 was prepared analogously to POL5; however, part of DBCO groups was reacted with 3-azido-1-propylamine via a copper-free alkyne-azide cycloaddition as follows: 3-azido-1-propylamine (1.3 μL, 13.6 μmol) was added to a solution of P5 (150 mg, 72.5 μmol of DBCO groups) in dry DMA (2 mL) followed by cluster C4 (49.6 mg, 145 µmol of azide groups) in 0.75 mL of dry methanol. The reaction mixture was bubbled with argon; then, it was stirred for 3 h at 24 °C. The reaction mixture was diluted with methanol (2 mL) and purified using a Sephadex LH-20 column with methanol elution and UV detection. The conjugate-containing fraction was collected and concentrated under vacuum to 2 mL. The polymer was isolated by precipitation into the mixture of acetone and diethyl ether (2/1; 100 mL) and dried to yield POL6 (136 mg; yield 91%).
The number-average molecular weight (Mn), weight-average molecular weight (Mw), and dispersity (Ð) of polymer precursors P1–P5 and polymer-cluster conjugates POL5 and POL6 were determined by a Shimadzu HPLC system equipped with a size exclusion chromatography (SEC) column TSK 3000 SWXL column (Tosoh Bioscience, Tokyo, Japan). Evaluation was carried out using a multi-angle light scattering (MALS) DAWN HELEOS II (Wyatt Technology Co., Santa Barbara, CA, USA), photodiode array SPD-M20A (Shimadzu, Japan) and differential refractometer index Optilab®-rEX (Wyatt Technology Co., Santa Barbara, CA, USA) detectors. The analysis was performed using a mixture of methanol and 0.3 M sodium acetate buffer, pH 6.5 (4/1, v/v) as a mobile phase at a flow rate of 0.5 mL min−1. The ASTRA software (version 8.1, Wyatt Technology Co., Santa Barbara, CA, USA) was used for calculation of Mw and Ð values.
The hydrodynamic diameter (Dh) and surface zeta potential (ZP) of all samples was measured using a Nano-ZS instrument (ZEN3600, Malvern, UK) with a laser wavelength of 632.8 nm, and the intensity of the scattered light was detected at an angle θ = 173°. Precursors were dissolved in PBS at 1 mg mL−1. Clusters and polymer-cluster constructs were dissolved in methanol, and aliquots of these solutions were added to deionized water (pH ~ 6) or PBS to obtain the final concentration of 1 mg mL−1; their long-term stability was evaluated. The values were determined as a mean of at least five independent measurements.
UV–VIS spectrophotometry (Specord 205 ST, Analytic Jena AG, Jena, Germany) was used for determination of the content of TT, DBCO, and amine groups. The molar absorption coefficient of ε(TT) = 10,300 L mol−1 cm−1 (λmax = 305 nm) in methanol was used for determination of TT groups in the polymer precursors before hydrolysis (P1) or aminolysis (P5). In the case of DBCO groups, ε(DBCO) = 13,000 L mol−1 cm−1 (λmax = 292 nm) in methanol was employed. The content of amine groups in polymer precursors P2, P4, and polymer-cluster conjugate POL6 was determined using a modified TNBSA assay as published earlier [37]. For P2 and POL6, the solution was prepared in borate buffer (0.1 M Na2B4O7·10H2O, pH 9.3) at the concentration of 2 mg mL−1. An aliquot of 100 μL of this solution was mixed with borate buffer (875 μL) and 0.03 M solution of TNBSA in water (25 μL). The molar absorption coefficient ε (NH2) = 17,200 L mol−1 cm−1 (λmax = 500 nm) was used, and the absorbance was measured after 90 min of incubation. In the case of P4, a similar evaluation method was employed, but the sample was dissolved in a mixture of borate buffer and DMSO (9/1) due to the hydrophobic character of cholesterol moieties [30].
For determination of cholesterol content in polymer precursors P3 and P4, 1H NMR spectra were measured with a Bruker Avance III 600 spectrometer (Bruker, Karlsruhe, Germany) operating at 600.2 MHz using DMSO-d6 as solvent. Typical conditions for measurements of the spectra were as follows: π/2 pulse width 10 µs, relaxation delay 10 s, spectral width 10 kHz, acquisition time 3.21 s, 200 scans, and 5 mm NMR tubes were used. The content of cholesterol moieties statistically distributed along the polymer backbone was assessed using the integral intensities of signals at δ ≈ 4.71 ppm (1 H, OH) and δ ≈ 3.67 ppm (1 H, CH) from the HPMA monomer unit and the integral intensity of the signal at δ ≈ 5.34 ppm (1 H, CH) from C-6 of cholesterol moiety. NMR data for both polymer precursors are shown in the Supplementary Materials, Section S2.
ATR FTIR spectra were recorded using a Thermo Nicolet Nexus 870 FTIR spectrometer (Bruker, Karlsruhe, Germany) purged with dry air and equipped with a liquid-nitrogen-cooled MCT (mercury cadmium telluride) detector. All the spectra were acquired using a Golden Gate single reflection ATR accessory (Specac Ltd., Orpington, UK) equipped with a diamond internal reflection element. Measurements were performed at room temperature using the following parameters: resolution 4 cm−1 and 256 scans. All data were processed in the OMNIC software (ver. 8.3.103). The atmosphere spectrum was subtracted from the acquired spectra, and then, the resulting spectra were subjected to the baseline and ATR corrections. Measurements were performed for solutions of C4, POL5, and POL6 in distilled water at 0.2 wt. % of Mo6-cluster equivalent.
Absolute photoluminescence quantum yields and emission spectra in deionized water or PBS were measured using a Quantaurus QY C11347-1 spectrometer (Hamamatsu, Japan). The samples were prepared by adding small aliquots of concentrated methanol solutions of the polymer-cluster constructs to deionized water or PBS to reach the final concentration of 0.1 mg mL−1. All samples were excited at 400 nm. In order to perform measurements under various concentrations of dissolved oxygen, the aqueous solutions of the clusters and the corresponding polymer constructs were saturated with air or argon.
In order to prevent Mo6 clusters’ aggregation in aqueous solutions, we present novel synthetic strategies for the synthesis of biocompatible polymer-coated Mo6 cluster constructs by employing covalent or non-covalent interactions with various pHPMA copolymers differing in their structure. This study is focused on the physico-chemical and photophysical properties of the polymer-cluster constructs intended as nanomedicines for anti-tumor therapy.
Controlled RAFT polymerization technique was employed in order to prepare well-defined polymer precursors with an appropriate number of functional groups, such as TT, COOH, NH2, and DBCO moieties, which are used for further chemical modifications. First, poly(HPMA-co-MA-AP-TT) (P0a) containing reactive TT groups were prepared and subsequently hydrolyzed in phosphate buffer (pH 8.0) to yield poly(HPMA-co-MA-AP-COOH) (P1). Even though much higher amounts of negatively charged groups may be necessary for stronger electrostatic interactions between the polymer and molybdenum molecule, highly, negatively charged systems tend to be captured by the reticuloendothelial system (RES), mainly in the liver and spleen, thus impairing their use for in vivo applications [38]. Therefore, a lower ratio of MA-AP-TT related to HPMA was used for the polymerization of P0a precursor, resulting in copolymers with 1.9 mol. % of final reactive groups and Mw ≈ 18,700 g·mol−1. Another poly(HPMA-co-MA-AP-TT) (P0b) was prepared, but using a higher ratio of MA-AP-TT related to HPMA, as well as a higher amount of monomers related to CTA and initiator in the reaction mixture, resulting in a precursor containing 10.2 mol. % of reactive TT groups and Mw ≈ 39,200 g mol−1. Here, the respective content of functional groups was chosen aiming to afford a precursor with a higher number of moieties available for further covalent interactions. To prepare P5 containing 8 mol. % of DBCO groups, P0b was reacted with an amine-functionalized DBCO. The copper-free click chemistry approach was selected to avoid the use of copper in further reactions, thus reducing toxicity-related risks and also avoiding time-consuming and complicated purification steps [39]. Higher amounts of positively charged groups on polymer precursors may be necessary to achieve stronger electrostatic interactions with negatively charged molybdenum clusters. Unfortunately, a higher content of amine groups is known to generate toxicity in vitro and in vivo; therefore, the ratio of comonomers was optimized to afford poly(HPMA-co-APMA-Boc) (P2) containing approximately 5.5 mol. % of amine groups. The introduction of hydrophobic moieties, such as cholesterol or its derivatives, into the structure of the water-soluble polymer carrier switches the character of the polymers to their amphiphilic nature [30,40,41,42]. Such amphiphilic polymers can either self-assemble into the core-shell micellar structures or can form a coating of hydrophobic nanoparticles or liposomes via interaction of cholesterol moiety with hydrophobic compartments of those nanomaterials [43,44]. According to the described procedures [30,40], the amphiphilic polymer precursors P3 and P4 were prepared containing 2.3 and 2.5 mol. % of cholesterol moieties, respectively, and comparable molecular weights around 25,000 g mol−1 with narrow dispersity. For 1H-NMR spectra of P3 and P4, see Figures S1 and S2, respectively. The introduction of certain number of amine groups into the polymer precursor P4 was performed to adjust the negative charge of C3, thus improving the properties of the system for eventual in vivo applications. According to the published data, neutral nanomedicines exhibit more favorable pharmacokinetic properties such as plasma half-life, recognition by RES, and possible adhesion to vascular endothelium [38]. Linear precursors P1 and P2 exhibited hydrodynamic diameters in aqueous solution around 5 nm, typical for HPMA-based polymer random coils, whereas the diameter of P5 was significantly higher (DH ≈ 11 nm) indicating eventual association of the macromolecules due to the presence of the hydrophobic pendant DBCO group and a higher molecular weight of the polymers. As expected, P3 showed the highest hydrodynamic diameter (DH ≈ 40 nm) due to the hydrophobic characteristic of pendant cholesterol moieties, which enable P3 to form micelles in aqueous solution, as already observed in our previous studies of such HPMA copolymers with cholesterol derivatives [40,41]. Nevertheless, positively charged amine moieties distributed along the same polymer backbone (P4) may have contributed to the formation of smaller micelles (DH ≈ 27 nm), probably due to the increased hydrophilicity of the precursor, hence impairing the formation of larger micelles. Detailed structures and physico-chemical characterization of the polymer precursors P1–P5 are shown in Figure 2 and Table 1, respectively.
In this study, several conjugation strategies were exploited aiming to optimize the constructs’ structure for possible future applications in the field of polymer therapeutics. Our initial attempts relied on the use of non-covalent electrostatic and hydrophobic interactions for preparation of the non-covalent constructs (POL1–POL2) and (POL3–POL4), respectively, which are referred to as “complexes”. Afterwards, the covalent attachment of the Mo6 clusters was employed to prepare POL5–POL6, which are called by the term “conjugates”. The structures of precursors P1–P4 and clusters C1–C3, used for non-covalent complex formation, and precursor P5 and cluster C4, used for preparation of the covalent conjugates, are summarized in Figure 2. Polymer-cluster complexes POL1 and POL2 were prepared by employing electrostatic interactions between P1 and P2 with the positively and negatively charged Mo6 clusters C1 and C2, respectively. An equimolar ratio between the oppositely charged functional groups of the polymer precursors and clusters was used. Consequently, this resulted in various contents of the clusters within their respective complexes: POL1 (11.5 wt%, cluster charge 4+) and POL2 (25.9 wt%, cluster charge 2−). POL1 exhibited ZP value +9 mV which was slightly lower than that of cluster C1 (+13 mV) as a consequence of the electrostatic complex formation with the negatively charged polymer precursor P1. Although the change in ZP value was relatively low, the formation of the complex POL1 was accompanied with a dramatic change of the hydrodynamic diameter from 48 nm of the cluster C1 (that tends to aggregate in water) to 5 nm of the polymer-coated complex POL1. The cluster–polymer interaction in this particular case enabled to obtain the unimolecular complexes presented in single cluster molecules coated by the hydrophilic copolymer. Most probably, the cluster containing +4 charge is coated with one or two polymer chains affording a slightly positive complex and a small dimension in contrast to huge aggregates of C1 cluster molecules. When the polymer precursor P2 with amine groups was employed for the complex formation, a significant increase in the zeta potential of complex POL2 (+4 mV), reaching almost neutral complex, was observed in comparison to the highly negatively charged cluster C2 (−67 mV) used in complexation. Such behavior indicate that the cluster was successfully modified with the positively charged polymer. Importantly, the hydrodynamic size of complex POL2 was significantly higher (Dh ≈ 30 nm) compared with POL1 (Dh ≈ 5 nm). We hypothesize that this was caused by the presence of much higher content of amino groups on polymer than the carboxyl groups on the cluster. The polymer coated sufficiently the clusters, but one polymer chain was generally involved in the coating of more than one cluster molecule in solution, thus the crosslinking of the polymer-cluster C2 occurred and increased the size. Another non-covalent method of the polymer-cluster complex preparation explored in this study was based on hydrophobic interactions of polymer precursors P3 and P4, both bearing hydrophobic cholesterol moieties, with the cholate-based cluster C3, affording the polymer-cluster complexes POL3 and POL4, respectively. Loading of 20 wt. % of Mo6 cluster was chosen for both complexes POL3 and POL4. The introduction of amine groups to polymer P4 was performed to verify the influence of the positive charge upon the formation of the hydrophobic complexes (POL3 and POL4). After interaction of precursor P3, bearing cholesterol moieties and neutral charge, with C3 (ZP = −9 mV), the zeta potential of the resulting complex POL3 slightly dropped to −14 mV. Opposite to that, a neutral to slightly positive surface charge (+1 mV) was observed when amine groups were introduced into the structure of POL4, showing the benefit caused by the addition of amino groups to the complex. The hydrodynamic sizes of original P3 and P4 polymers, Dh ≈ 40 and 27 nm, respectively, showed formation of the micelles self-assembled from these amphiphilic copolymers. Importantly, after the complexation with C3, both complexes POL3 and POL4 showed smaller hydrodynamic sizes, Dh = 8.4 and 12 nm, respectively, in comparison to their polymer precursors. We summarize that the observed change in size upon the addition of the hydrophobic moieties containing cluster is most probably caused by disruption of the self-assembled micellar vesicles formed by the amphiphilic polymer precursors and subsequent rearrangement of the polymer chains around the hydrophobic cluster in aqueous solution. In contrast to POL2 and similarly to POL1, both POL3 and POL4 are rather formed by a single molecule of cluster coated by a small number of polymer chains. Physico-chemical characteristics of Mo6 clusters are shown in Table 2. With respect to the limited colloidal stability of the non-covalent polymer-cluster complexes, which will be further discussed within this manuscript, we evaluated the covalent binding of Mo6 clusters to the polymer carriers as an alternative approach. For this purpose, DBCO groups from polymer precursor P5 were used for the attachment of Mo6 cluster C4 via a copper-free alkyne-azide cycloaddition (“click reaction”), affording stable covalent polymer-cluster conjugates POL5 and POL6. For both conjugates, the content of Mo6 cluster was calculated considering the ratio 1/1 between DBCO groups on polymer precursor and azide groups on cluster moiety. It is generally known that the highly positively charged systems should be toxic for the body cells and can adhere to the negative vascular endothelium, leading to a lower concentration in the plasma along with an impaired EPR effect [38]. On the other hand, highly negatively charged systems are easily recognized taken up by the immune system, and then stacked in the kidneys. To avoid such effects and potential drawbacks, the ideal surface charge should be neutral or only slightly negative. Therefore, additional amine groups were introduced into the polymer-cluster conjugate POL6 via reaction of the conjugate POL5 with 3-azido-1-propylamine to neutralize the negative charge on the cluster and make the conjugate neutral in charge. Infrared spectroscopy was used to measure the conversion of the alkyne-azide cycloaddition. The efficiency of C4 conjugation to P5 was determined using the strong asymmetric stretching vibration mode of azide groups [45,46,47]. To assess the amount of unreacted azide groups, a reference was set up as the integral area of the asymmetric stretching vibrational band of azide groups centered at 2056 cm–1 of C4 (0.2 wt% Mo6-cluster in distilled water; Figure 3, dark grey line) [45,46]. This typical complex profile can be attributed to weak Fermi resonance of the vibration with combination tones of C–N stretching vibrations and partly to possible differences in local environments of the six azide groups [47,48]. The azide band’s shift to a lower wavenumber (2056 cm–1) compared to the typical position described in the literature was attributed to the presence of charged molybdenum and iodine atoms in the cluster’s structure [47]. Additionally, the cluster forms a stable colloid in water, and azide groups are involved in H-bonding with water molecules [48] which may contribute to the observed red-shifted position of the band. After conjugation of C4 with the precursor P5, affording POL5, and further introduction of amine groups along the polymer backbone affording POL6, the asymmetric stretching band intensity of the azide groups greatly decreased (Figure 3, blue and red lines) due to newly formed covalent bond. As the bands’ position remained in the same region, the peak area tool in the OMNIC software was used to compare their intensities in the spectra of POL5 and POL6 using C4 band as a reference. After the conjugation reaction, 9% ± 1% of azide groups remained unreacted for POL5, and similar value was found for POL6. This assessment indicates that the click reaction was successful as majority of the azide groups reacted. The zeta potential of POL5 conjugate (−17 mV) was similar to the respective cluster C4 (−16 mV). As mentioned above, negatively charged nanoparticles could exhibit unfavorable pharmacokinetic properties after the body injection. Therefore, to avoid such behavior, part of the DBCO groups of polymer precursor P5 was reacted with 3-azido-1-propylamine prior to addition of cluster C4 to compensate, at least partially, the negative zeta potential of the polymer-cluster conjugate. Such combinational approach provided the polymer precursor with 1.5 mol. % of amine groups along the polymer backbone and consequently polymer-cluster conjugate POL6 with zeta potential adjusted to −7 mV, which is much more suitable for future therapeutic application of the developed polymer-cluster conjugates. Both Mw and Dh values proved the formation of polymer-cluster conjugates. The Mw slightly increased for POL5 and POL6 (Mw ≈ 49,500 g mol−1 and 52,000 g mol−1, respectively) when compared to their precursor P5 (Mw ≈ 40,000 g mol−1) upon introduction of Mo6 clusters in the structure. The hydrodynamic size remains similar to that of the other complexes, thus most probably forming the unimolecular complexes presented with a single or low number of cluster molecules. For physico-chemical and photophysical characteristics of polymer-cluster constructs POL1–POL6, see Table 3. Importantly, we can summarize that we have successfully designed and synthesized several polymer-cluster systems differing in their inner structure and mode of the polymer-cluster interaction or bonding.
We further evaluated the colloidal stability and photophysical properties of the polymer-cluster constructs in PBS, a biologically relevant medium. In the case of POL1 and POL2, formed by electrostatic interactions, the complexes were not stable in any manner as the electrostatic interactions were not strong enough to keep the polymer-cluster complexes stable. POL1 and POL2 complexes in PBS immediately formed huge aggregates, and their precipitation was observed quite rapidly after their dissolution. In the case of POL3 and POL4, exploiting the hydrophobic interactions, a slightly better stability in PBS was observed. Nevertheless, aggregates were also formed, suggesting that the hydrophobic polymer-cluster interactions were disrupted in PBS. Consequently, it was not possible to properly perform DLS, SEC or luminescence spectroscopy studies in PBS for any of these non-covalent complexes. In terms of photophysical properties, the constructs were first studied in deionized water, where they displayed the typical broad emission band of the Mo6 clusters with maxima in the 685–697 nm range (Figure 4). The emissivity was high with luminescence quantum yield ranging from 0.16 for POL1 to 0.49 for POL4 in argon saturated water (see Table 3). The quenching of the emission by oxygen was efficient with a fraction of triplet states quenched by oxygen in an air saturated solution (FT = 1 − ΦL(air)/ΦL(Ar)) of approximately 0.8 for all constructs, indicating good accessibility of the clusters to dissolved oxygen. This feature suggests an effective production of O2(1Δg) which is a major factor for a successful photodynamic treatment. Overall, the luminescent properties of the clusters were preserved upon association with their respective polymers. In contrast to the non-covalent complexes which were not stable in PBS buffer, covalent polymer-cluster conjugates POL5 and POL6 displayed remarkable stability in PBS. Their hydrodynamic diameter of approximately 7–11 nm and ZP values did not change significantly even after 5 days (see Table 4). Additionally, no drastic changes were observed in the photophysical properties of the conjugates dissolved in PBS in comparison with deionized water solutions, except for slightly red shifted emission maxima and higher quantum yields (Table 4). The photophysical stability of the solutions was evaluated over a five-day period revealing no significant changes in the emission maxima, quantum yields, and oxygen quenching constant, which evidenced the high stability of the photosensitizing system (see Figure 5) in PBS. These features are attractive for photodynamic applications as a reasonable stability of the photosensitizing system is required for an effective PDT. Taken all together, even though suitable photophysical properties were found for the polymer-cluster complexes, their colloidal stability in PBS does not allow further in vitro and in vivo testing and application. For polymer-cluster conjugates POL5 and POL6 bearing the Mo6 cluster covalently bound to the polymer carrier, a very good stability and photophysical properties were observed. Both size and zeta potential were maintained after several days in PBS and even after months of storage as dried powders; hence, we believe that these conjugates are more suitable for further biomedical applications.
This study described the design and synthesis of hydrophilic HPMA-based polymer constructs with Mo6 clusters, potent singlet oxygen photosensitizers, and their structures’ optimization from the chemical and physico-chemical point of view. We investigated three methods of preparation of the constructs using electrostatic, hydrophobic, or covalent interactions between the polymer backbone and cluster moieties. The luminescent properties of the Mo6 clusters were preserved upon association with their respective polymers and all polymer-cluster constructs exhibited a production of O2(1Δg), which is a major factor for a successful photodynamic treatment. The conjugates prepared by covalent interactions, such as the azide-alkyne “click chemistry”, were the best in the series–they possessed a high colloidal stability in PBS and provided high luminescence quantum yields. Moreover, a significant advantage of the synthesis is the fact that copper is completely avoided during the procedure. Results from physico-chemical and photophysical evaluation indicate that the conjugates with Mo6 covalently attached to the polymer backbone are prospective candidates for biological evaluation including measurements of the cytotoxicity effect against selected cancer cells and in vivo experiments using suitable animal tumor models to assess the PDT efficacy of such system. | true | true | true |
|
PMC9568309 | Zhenzhen Wei,Jing Zhou,Hao Yu,Yunzhou Pu,Yuelei Cheng,Yi Zhang,Qing Ji,Huirong Zhu | Zuo Jin Wan Reverses the Resistance of Colorectal Cancer to Oxaliplatin by Regulating the MALAT1/miR-200s/JNK Signaling Pathway | 07-10-2022 | Background Oxaliplatin (L-OHP) is a common chemotherapy drug used in the treatment of colorectal cancer (CRC). Our previous work showed that Zuo Jin Wan (ZJW), a traditional Chinese medicine prescription, could improve sensitivity to L-OHP in the treatment of CRC, but the detailed mechanism is not clear. In previous mechanistic studies, we found that the miR-200s expression in CRC is associated with L-OHP sensitivity through regulation of MDR1/p-gp and the downstream c-JunN-terminal kinase (JNK) signaling pathway. Moreover, lncRNA-MALAT1 offers great potential in the regulation of drug resistance by interacting with miR-200s. Therefore, in this work, we explored whether ZJW could reverse L-OHP resistance in CRC by regulating MALAT1, miR-200s, and the downstream signaling pathway. Methods Cell Counting Kit-8 and flow cytometry were used to detect the effects of ZJW combined with L-OHP on chemotherapy tolerance and cell apoptosis of HCT116/L-OHP cells. Western blotting and quantitative real-time PCR (qRT-PCR) were used to detect the activation of the JNK signaling pathway and the protein and mRNA expression levels of the drug resistance-related MDR1/ABCB1 gene in HCT116/L-OHP cells treated with ZJW. The binding sites of MALAT1 and miR-200s were predicted by bioinformatics tools and confirmed by qRT-PCR. qRT-PCR was used to detect the expression of miR-200s and MALAT1 in HCT116/L-OHP cells treated with ZJW. A xenograft model of CRC in nude mice was established to observe the effect of ZJW combined with L-OHP on the growth of subcutaneously transplanted tumors. Apoptosis in tumor cells was detected by TUNEL staining. The activation of the JNK signaling pathway and the expression of drug resistance-related proteins were detected by immunohistochemistry and immunofluorescence. qRT-PCR was used to detect the expression of miR-200s and the MALAT1 gene in the tumors. Results Our study showed that ZJW could significantly decrease the proliferation and promote apoptosis of HCT116/L-OHP cells treated with L-OHP. We further proved that ZJW could reverse the drug resistance of HCT116/L-OHP cells by reducing MALAT1, indirectly upregulating miR-200s, alleviating the activation of the JNK signaling axis, and downregulating the expression of resistance proteins such as MDR1/ABCB1 and ABCG2. ZJW combined with L-OHP inhibited the growth of subcutaneously transplanted tumors and induced apoptosis in nude mice. ZJW reduced the expression of MALAT1 and upregulated the expression of miR-200s in transplanted tumors. In addition, ZJW also alleviated the activation of the JNK signaling pathway while reducing the expression of MDR1/ABCB1 and ABCG2. Conclusions Our study identified that MALAT1 promotes colorectal cancer resistance to oxaliplatin by reducing the miR-200s expression. ZJW may reverse chemoresistance by inhibiting the expression of MALAT1 and regulating the miR-200s/JNK pathway, providing an experimental basis for the clinical application of ZJW in relieving chemotherapy resistance. | Zuo Jin Wan Reverses the Resistance of Colorectal Cancer to Oxaliplatin by Regulating the MALAT1/miR-200s/JNK Signaling Pathway
Oxaliplatin (L-OHP) is a common chemotherapy drug used in the treatment of colorectal cancer (CRC). Our previous work showed that Zuo Jin Wan (ZJW), a traditional Chinese medicine prescription, could improve sensitivity to L-OHP in the treatment of CRC, but the detailed mechanism is not clear. In previous mechanistic studies, we found that the miR-200s expression in CRC is associated with L-OHP sensitivity through regulation of MDR1/p-gp and the downstream c-JunN-terminal kinase (JNK) signaling pathway. Moreover, lncRNA-MALAT1 offers great potential in the regulation of drug resistance by interacting with miR-200s. Therefore, in this work, we explored whether ZJW could reverse L-OHP resistance in CRC by regulating MALAT1, miR-200s, and the downstream signaling pathway.
Cell Counting Kit-8 and flow cytometry were used to detect the effects of ZJW combined with L-OHP on chemotherapy tolerance and cell apoptosis of HCT116/L-OHP cells. Western blotting and quantitative real-time PCR (qRT-PCR) were used to detect the activation of the JNK signaling pathway and the protein and mRNA expression levels of the drug resistance-related MDR1/ABCB1 gene in HCT116/L-OHP cells treated with ZJW. The binding sites of MALAT1 and miR-200s were predicted by bioinformatics tools and confirmed by qRT-PCR. qRT-PCR was used to detect the expression of miR-200s and MALAT1 in HCT116/L-OHP cells treated with ZJW. A xenograft model of CRC in nude mice was established to observe the effect of ZJW combined with L-OHP on the growth of subcutaneously transplanted tumors. Apoptosis in tumor cells was detected by TUNEL staining. The activation of the JNK signaling pathway and the expression of drug resistance-related proteins were detected by immunohistochemistry and immunofluorescence. qRT-PCR was used to detect the expression of miR-200s and the MALAT1 gene in the tumors.
Our study showed that ZJW could significantly decrease the proliferation and promote apoptosis of HCT116/L-OHP cells treated with L-OHP. We further proved that ZJW could reverse the drug resistance of HCT116/L-OHP cells by reducing MALAT1, indirectly upregulating miR-200s, alleviating the activation of the JNK signaling axis, and downregulating the expression of resistance proteins such as MDR1/ABCB1 and ABCG2. ZJW combined with L-OHP inhibited the growth of subcutaneously transplanted tumors and induced apoptosis in nude mice. ZJW reduced the expression of MALAT1 and upregulated the expression of miR-200s in transplanted tumors. In addition, ZJW also alleviated the activation of the JNK signaling pathway while reducing the expression of MDR1/ABCB1 and ABCG2.
Our study identified that MALAT1 promotes colorectal cancer resistance to oxaliplatin by reducing the miR-200s expression. ZJW may reverse chemoresistance by inhibiting the expression of MALAT1 and regulating the miR-200s/JNK pathway, providing an experimental basis for the clinical application of ZJW in relieving chemotherapy resistance.
Colorectal cancer (CRC) is one of the most common malignant tumors worldwide [1], with increasing incidence each year [2]. Oxaliplatin (L-OHP)-based chemotherapy improves the survival of CRC patients [3, 4]. However, the occurrence of drug resistance seriously damages the efficiency of chemotherapy [5, 6]. It has been reported that the response rate for L-OHP is 6%–31% and the median progression-free survival is 3.1–7 months [7]. Chemoresistance contributes greatly to treatment failure [8]. Therefore, it is urgent to develop effective strategies for enhancing the sensitivity of L-OHP to improve the survival and prognosis of patients with CRC. Traditional Chinese medicine (TCM) and its bioactive substances have been confirmed to improve treatment efficiency and prolong the survival of patients with tumors [9]. Zuo Jin Wan (ZJW) is a TCM prescription that has been widely used in gastrointestinal diseases. Our previous study demonstrated that ZJW could enhance the sensitivity to cisplatin and L-OHP in HCT116/L-OHP cells [10]. However, the mechanism by which ZJW reverses chemoresistance remains unknown. Considerable efforts have been made to elucidate the mechanism of platinum resistance through biochemical characterization and molecular aspects, including reduced platinum accumulation, enhanced DNA repair, decreased apoptosis, and inactivation by thiol-containing biomolecules, including glutathione [11]. P-gp, ATP-binding cassette (ABC) drug transporter noncoding RNA, and the microbiome were confirmed to regulate the molecular and biochemical processes of chemoresistance [12, 13]. P-gp acts as an ATP-dependent outflow pump which can pump a variety of drugs out of cells to prevent intracellular accumulation [14]. miRNAs regulate the expression of mRNA by inhibiting the translation or stability of mRNA [15]. Long noncoding RNAs (lncRNAs) are capable of interacting with miRNAs by acting as competing endogenous RNAs. The dysregulation of lncRNAs/miRNAs is associated with the development of cancer [16]. MALAT1 is a lncRNA distributed in the nucleus that is highly expressed in CRC, liver cancer, breast cancer, and other cancer tissues [17]. The c-JunN-terminal kinase (JNK) signaling pathway contributes to drug resistance by regulating transporters and blocking the cell cycle and cell apoptosis [18]. Our previous studies have found that the overexpression of miR-200c can inhibit the expression of MDR1/P-gp and deactivate the JNK pathway to increase the sensitivity of drug-resistant human CRC cells to L-OHP. Combined with previous research results, we speculate that ZJW reverses the drug resistance of CRC by regulating the MALAT1/miR-200s/JNK signaling axis. Thus, in this work, we investigated and confirmed the function of MALAT1 with respect to miR-200c through bioinformatics analysis and in vitro experiments. Using L-OHP-resistant cell lines and a xenograft model of CRC, we demonstrated the effect and the molecular mechanisms of ZJW in reversing chemoresistance in CRC. Overall, our study revealed the use of ZJW as a potential strategy for enhancing the antitumor efficacy of L-OHP in CRC treatment.
ZJW is composed of Rhizoma coptidis and Evodia in the proportion of 6 : 1. All Chinese herbal medicines were purchased from the traditional Chinese medicine store (Shang, China) of Shuguang Hospital affiliated with Shanghai University of Traditional Chinese medicine. The mixture (70 g) was refluxed and extracted twice with ethanol (1 : 8, v/v) for 1 hour each time. The filtrate was concentrated and dried under vacuum at −80°C, and the dry powder yield was 21.2%. The powder was stored at 4°C. The preparation was standardized and quality-controlled according to the standards set by the State Food and Drug Administration (SFDA). One gram of the ZJW powder was accurately weighed by an electronic balance and dissolved in 1 mL DMSO. The samples were vortexed and sonicated overnight and sterilized by ultraviolet irradiation to prepare 1 g/mL ZJW. In the in vitro experiment, ZJW was diluted with medium to three concentrations: low (5 μg/mL), medium (10 μg/mL), and high (20 μg/mL).
Human colorectal cancer HCT116 parental cells were purchased from Shanghai Fine Cell Bank (Shanghai, China). The HCT116/L-OHP multidrug-resistant cell line was established and maintained in our laboratory. The HCT116 cell line was cultured in a RPMI 1640 conditioned medium (HyClone, China) containing 10% (v/v) heat-inactivated fetal bovine serum, 100 units/mL penicillin, and 100 μg/mL streptomycin. HCT116/L-OHP cells were routinely cultured in a conditioned medium containing 5 μg/mL L-OHP (Sanofi, France). The culture conditions were 37°C and 5% CO2 saturation humidity.
A total of 36 male nude mice of SPF cleanliness class, aged 8 to 12 weeks, were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China, license no. SCXK2017-0005) and kept under specific pathogen-free conditions. The room temperature was 20°C, and the relative humidity was 60%. The 12 h cycle of light and shade was maintained, with freely available drinking water, and fasting was ensured for 12 hours before the experiment. The animal experiment was performed and approved by the Animal Ethics Committee of Shanghai University of Traditional Chinese Medicine. All experimental mice were tested in accordance with animal ethics standards and in line with the provisions and general recommendations of the regulations of the Administration of Experimental Animals of China. HCT116/L-OHP cells were grown in a medium, digested, separated by trypsin, washed, and resuspended in Hanks' solution (HBSS). There were 1 × 106 HCT116/L-OHP cells in the logarithmic phase per 0.2 mL of HBSS. When the average size of the tumor reached 100 mm3, the mice were randomly divided into 6 groups (n = 5 in each group). The mice in the first group were treated with distilled water every day as the control group. The mice in groups 2-5 were intraperitoneally injected with L-OHP once every two days, and the injection dose (5 mg/kg) was half of the MTD of L-OHP, as mentioned earlier. The mice in group 3, group 4, and group 5 were given ZJW by intragastric administration at doses of 1027.5 mg/kg, 2055 mg/kg, and 4110 mg/kg, respectively. The mice in the sixth group were only given intragastric administration of ZJW at a dose of 1027.5 mg/kg to eliminate the toxicity of evodiamine. The length (A) and width (B) of tumors were recorded every 2 days, and the tumor volume was estimated according to the following formula: V = π/6 × A ∗ B ∗ B. The tumor growth map was generated according to the tumor volume and time. On the 28th day after treatment, the mice were killed and the tumor tissue was removed and weighed.
The oxaliplatin-resistant human colorectal cancer cell line HCT116/L-OHP in the logarithmic growth phase was selected, and the cell concentration was adjusted to 1 × 104 mL. A total of 100 μL per well was inoculated in a 96-well cell culture plate at 37°C in a 5% CO2 saturated humidity incubator. After 3–5 hours, when the cells adhered to the wall, the culture medium was discarded. Compared with the blank group, L-OHP and L-OHP combined with different concentrations of ZJW were added. After 48 hours of culture, 100 μL of CCK-8 solution (Dojindo, Japan) was added to each well and then incubated in the incubator for 4 hours. The OD value (450 nm) was measured by an enzyme labeling instrument and the growth inhibition rate of the cells was calculated.
The cells were inoculated in 6-well plates (4 × 105/well). Twelve hours later, compared with the blank group, L-OHP and L-OHP combined with different concentrations of ZJW were added (the IC10 of ZJW in the CCK-8 assay was obtained). After 48 hours of incubation, the cells were digested by trypsin, washed twice with PBS, and counted. According to the instructions, 1× binding Buffer in the apoptosis kit (BD Pharmingen, USA) was used to produce a 1 × 106 mL cell suspension. A total of 300 μL of cell suspension was added to a new flow tube, and 5 μL Annexin V-FITC and 5 μL propidium iodide (PI) were added. After gently mixing, the reaction proceeded for 15 min at room temperature with light avoidance, 200 μL 1× binding buffer was added to each tube, and the results were detected on the computer within 1 hour. Flow cytometry was used to detect apoptosis by determining the relative number of Annexin V-FITC-positive and PI-negative cells. Nonstained cells, cells stained with Annexin V-FITC alone, and cells stained with PI alone were used as controls. Monochromatic cells were used to regulate electronic compensation in the FL1 and FL2 channels.
The excised tumor was washed with xylene twice, each time for 5 min. The cells were rinsed once with gradient ethanol (100%, 95%, 90%, 80%, 70%) and twice with PBS. Using Proteinase K working liquid, the tissue was treated for 15–30 min and the TUNEL reaction mixture was prepared according to the instructions. The specimens were then covered with glass slides or sealing film in a dark wet box at 37°C for 60 minutes, then rinsed 3 times with PBS. Then, 50 μL of converter-POD was added to the specimen after the slide was dried, and the slide or seal film was covered and reacted at 37°C in a dark wet box and rinsed 3 times within 30 minutes. Next, 50 μL DAB substrate was added to the tissue, which was reacted at 25°C and then rinsed with PBS for 10 minutes. After image acquisition, the tissue was stained with hematoxylin or methyl green and immediately rinsed with water, then subjected to gradient alcohol dehydration, xylene for transparency, and neutral gum sealing. A drop of PBS or glycerin was added to the field of vision, and cell counts and images were acquired with an optical microscope.
HCT116/L-OHP cells in the logarithmic growth period were used to establish the control group and groups of different doses of ZJW. L-OHP was combined with different concentrations of ZJW (0, 10, 15, and 20 μg/mL) for 48 h. The protein was extracted and quantified by a BCA protein assay kit (Beyotime, China), and the protein samples were treated. A 10% SDS-PAGE gel, sample protein, and marker (Thermo, China) were prepared. The proteins were then transferred to PVDF membranes after electrophoresis and sealed at room temperature for 2 hours. The closed membrane was immersed in the primary antibody and incubated overnight at 4°C. The next day, TBST was used to wash the film three times, each time for 10 min, and the membrane was soaked in the secondary antibody and incubated at room temperature for 2 hours. After washing the film, it was colored with ECL chemiluminescence solution (Cytiva, China), and the gray value was analyzed by ImageJ software. The primary antibodies used were MDR1/ABCB1 (CST, USA), ABCG2 (CST, USA), MRP4/ABCC4 (CST, USA), JNK2 (CST, USA), phospho-SAPK/JNK (CST, USA), and GAPDH (CST, USA). The secondary antibody used was HRP-labeled goat anti-rabbit/mouse IgG (H + L) (Beyotime, China). Each band was quantitatively analyzed using ImageJ software and normalized to the expression of GAPDH in the same lane.
RNA was extracted from cultured cells with TRIzol reagent (Takara, China). Total RNA was reverse-transcribed into cDNA according to the instructions of the reverse transcription kit (TIANGEN, China). An appropriate amount of cDNA was used as the template for PCR and amplified according to the qRT-PCR specifications (TIANGEN, China). Finally, the gene expression was analyzed with GAPDH and U6 as internal references, and relative expression levels were calculated by using the 2−△△Ct method. PCR was performed using an ABI 7500 instrument (Applied Biosystems, USA). The primers used for real-time PCR analysis are listed in Table 1.
The paraffin sections were incubated in the sealing solution (10% donkey serum + 5% skim milk + 4% BSA + 0.1% Triton X-100) for 10 minutes; then, hydrated sections were incubated with antibodies at 4°C overnight. After rinsing with phosphate-buffered saline (PBS), the slices were incubated with a diluted biotinylated secondary antibody for 30 minutes. Subsequently, the slides were washed with PBS again and incubated with the prefabricated avidin peroxidase macromolecular complex for 30 minutes. The peroxidase reaction was completed by incubation in PBS containing 0.01% hydrogen peroxide at room temperature for approximately 5 minutes. The slices were thoroughly washed in tap water, anti-stained in hematoxylin, dehydrated in anhydrous ethanol, deparaffinized in xylene, and examined under a microscope in synthetic resin.
Frozen tissue sections were fixed in 4% paraformaldehyde for 10 min at room temperature and then washed twice with PBS. Blocking buffer (DakoCytomation, Glostrup, Denmark) was added for 30 min, and samples were then stained with primary antibodies and FITC-conjugated goat anti-rabbit IgG (Millipore). Sections were imaged using a TCS SP2 spectral confocal system (Leica, Germany). All experiments were conducted according to instructions from the antibody manufacturer.
The starBase database is a widely used open-source platform for studying lncRNA interactions from CLIP-seq, degradome-seq, and RNA-RNA interactome data. Herein, the starBase database was introduced to analyze the expression correlation between miRNAs and genes or pseudogenes. Taking miR-200s as the target, the biological information in the lncRNA online database (https://starbase.sysu.edu.cn/browseNcRNA.php) should be used to predict the lncRNA molecules that may regulate miR-200s and select them according to the scoring results and related literature.
SPSS 26.0 statistical software was used for analysis. The measurement data are represented by ± s, and the t test was used for comparisons between the two groups. P < 0.05 indicates that the difference is statistically significant.
To avoid the effect of the cytotoxicity of ZJW on the inhibition of CRC cell proliferation, we detected the cytotoxic effect of ZJW on HCT116/L-OHP cells. The results showed that the IC10 dosage in HCT116/L-OHP cells was 20 μg/mL (Figure (a)1). Below this dose, there was no significant difference in cell survival between treated cells and untreated cells. Therefore, in all cell proliferation experiments, ZJW was used to treat cells in the concentration range of 20 μg/mL. The CCK-8 results showed that HCT116/L-OHP cells were more resistant to oxaliplatin than sensitive HCT116 cells (Figure (b)1). When HCT116/L-OHP cells were treated with different concentrations of oxaliplatin for 48 hours, the inhibitory effect of oxaliplatin on HCT116/L-OHP cells was not obvious, but the inhibitory effect of ZJW combined with oxaliplatin on HCT116/L-OHP cells was significantly stronger than that of oxaliplatin alone (Figure (c)1). These results suggest that ZJW can enhance the sensitivity of HCT116/L-OHP cells to oxaliplatin in vitro. To further explore the mechanism by which ZJW increases cell apoptosis, Annexin V and PI double staining was used to observe the changes in apoptosis induced by ZJW combined with oxaliplatin. The results showed that compared with the control group, the apoptosis rate of HCT116/L-OHP cells treated with oxaliplatin was 26% (Figures (d)1 and 1(e)). After intervention with different concentrations of ZJW combined with oxaliplatin, the apoptosis rate of HCT116/L-OHP cells increased to 36.2%, 52.8%, and 71.4%, respectively (Figures (d)1 and 1(e)), suggesting that ZJW can enhance the apoptosis of oxaliplatin-treated HCT116/L-OHP cells in a concentration-dependent manner.
Our previous studies showed that the overexpression of miR-200c inhibits the activation of the JNK signaling pathway and reverses tumor drug resistance. Therefore, we detected whether ZJW affects the expression level of the JNK signaling pathway indirectly by affecting the expression of miR-200s. The qRT-PCR results showed that the mRNA expression of miR-200s in HCT116/L-OHP cells was significantly lower than that in sensitive cells. The mRNA expression levels of miR-200s in HCT116/L-OHP cells treated with L-OHP and different concentrations of ZJW were significantly higher than that in the control group (Figure (a)2). At the same time, qRT-PCR and Western blot showed that compared with the control group, the relative protein expression of JNK in HCT116/L-OHP cells did not change, while the relative protein expression of p-JNK decreased after ZJW intervention, suggesting that the JNK signaling pathway was inactivated (Figures (b)2 and 2(c)). These results suggest that ZJW may regulate the drug resistance of CRC by increasing the miR-200s expression and inhibiting the JNK signaling pathway in vitro. MDR1/ABCB1, MRP4/ABCC4, and ABCG2 are all ABC transporter proteins associated with drug transport. We detected the expression levels of these proteins. The results showed that compared with sensitive cells, the relative expression of MRP4/ABCC4 protein in HCT116/L-OHP cells exhibited no significant change. The relative expression of MDR1/ABCB1 and ABCG2 proteins was increased, suggesting that the mechanism of drug resistance may be related to drug efflux on the cell membrane (Figure (a)3). After treatment with different concentrations of ZJW, the expression levels of MDR1/ABCB1 and ABCG2 resistance proteins in the low-, middle-, and high-dose ZJW intervention groups were significantly decreased compared with those in the control group (Figures (a)3 and 3(b)). In addition, compared with sensitive cells, the relative mRNA expression of MDR1/ABCB1 and ABCG2 in HCT116/L-OHP cells increased, while the relative mRNA expression of MDR1/ABCB1 and ABCG2 decreased in a dose-dependent manner after intervention with different concentrations of ZJW and the difference was statistically significant (Figure (c)3). These results suggest that the reversal effect of ZJW on drug resistance may be related to inhibiting the expression of MDR1/ABCB1 and ABCG2.
To study the existence of upstream regulatory genes in miR-200s, we used bioinformatics analyses to predict the existence of multiple complementary binding sites between MALAT1 and miR-200s (Figures (a)4 and 4(b)). Then, we verified the expression of MALAT1 and miR-200s in cells. We found that compared with sensitive cells, the relative mRNA expression of MALAT1 in HCT116/L-OHP cells was increased, while the relative mRNA expression of miR-200s was decreased, and the difference was statistically significant (Figures (a)2 and 4(c)). To further study the internal relationship between miR-200s and MALAT1, we constructed a siRNA against the MALAT1 gene. The application of siRNA-MALAT1 resulted in a significant decrease in MALAT1 mRNA expression in HCT116/L-OHP cells (Figure (e)4). Verified by qRT-PCR, the relative mRNA expression of miR-200s was increased in HCT116/L-OHP cells transfected with siRNA-MALAT1 (Figure (f)4). In addition, the use of ZJW reduced the MALAT1 expression in HCT116/L-OHP cells (Figure (d)4). Therefore, we proved the function of MALAT1 in indirectly regulating the miR-200s expression. Combined with the abovementioned results, we suggest that ZJW may reverse chemoresistance in CRC by regulating the MALAT1/miR-200s/JNK signaling pathway.
We used HCT116/L-OHP cells to inoculate nude mice to build a xenograft model. After successful modeling, the nude mice were divided into 6 groups in order to observe the tumor volume and body weight. With the passage of time, the tumor volumes of nude mice in the model group and experimental group gradually increased. After 28 days of the administration, there was no significant difference in tumor size between the L-OHP group and the model group, but the combined administration of L-OHP and ZJW could significantly inhibit tumor growth, and the shrinkage of tumor volume in mice was concentration-dependent (Figures (a)5–5(d)). These experiments provide evidence that ZJW enhances drug sensitivity during chemotherapy in vivo. To explore whether ZJW can also induce apoptosis in vivo, a TUNEL assay was used to observe the changes in subcutaneously transplanted tumors in nude mice. The results showed that compared with the normal saline control group, there was no significant change in apoptosis in subcutaneous tumor tissues of nude mice treated with L-OHP or ZJW alone. After intervention with ZJW combined with L-OHP, the expression of TUNEL in subcutaneous tumor tissue of nude mice increased significantly, and the difference was statistically significant (Figures (e)5 and 5(f)). These results suggest that ZJW can enhance the effect of L-OHP on the apoptosis of HCT-116/L-OHP cells in vivo, which is consistent with the in vitro experimental results.
The immunohistochemical results showed that the relative protein expression levels of p-JNK in the treatment groups with L-OHP combined with different concentrations of ZJW were lower than that in the control group (Figures (a)6 and 6(d)). The Western blot results were consistent with the immunohistochemistry results (Figures (b)6 and 6(c)). These results showed that ZJW could regulate the drug resistance of CRC by inhibiting the JNK signaling pathway in vivo. qRT-PCR detection showed that, in accordance with the in vitro results, the relative mRNA expression of miR-200s was significantly upregulated (Figure (e)6). The relative mRNA expression of MALAT1 was significantly decreased in the treatment groups with L-OHP combined with different concentrations of ZJW compared with the control group in vivo (Figure (f)6). The expression of ABCG2 was detected by immunohistochemistry, while the expression of MDR1/ABCB1 and MRP4/ABCC4 was detected by immunofluorescence to evaluate the effect of ZJW in vivo. In accordance with the results of the in vitro experiment, the expression levels of MDR1/ABCB1 and ABCG2 in the treatment groups with L-OHP combined with different concentrations of ZJW were lower than that in the normal saline control group (Figures (a)7 and 7(b)), but there was no significant change in the MRP4/ABCC4 expression (Figures (c)7 and 7(d)). Furthermore, Western blotting was used to verify the results. The relative protein expression of MDR1/ABCB1 and ABCG2 in the treatment groups with L-OHP combined with different concentrations of ZJW was significantly lower than that in the control group (Figures (e)7 and 7(f)). The qRT-PCR results coincided with the WB results (Figure (g)7). There was no significant change in the protein or mRNA expression of MRP4/ABCC4. These results suggest that ZJW may increase the sensitivity of CRC to L-OHP by decreasing the MDR1/ABCB1 and ABCG2 expressions. In summary, we proved that ZJW could increase the sensitivity to L-OHP, and the mechanism may be related to the downregulation of MALAT1/miR-200s/JNK signaling.
Chemotherapy is currently still one of the main nonsurgical approaches to impede tumor progression in CRC patients with advanced stages. The occurrence of drug resistance and side effects limits the response rate of CRC patients to chemotherapy. The mechanisms of chemoresistance are complex. Plant-based phytochemicals are large parts of TCM, which are now used to treat various physiological diseases. Recently, herbal medicine has gained preference owing to its purity, strength, and cost-effectiveness [19]. Natural TCM compounds slow the spread of cancer by promoting apoptosis and inhibiting metastasis [20]. ZJW is a common prescription for gastrointestinal diseases such as gastritis, ulcerative colitis, and colorectal cancer. Our study demonstrated that ZJW could reverse L-OHP resistance in CRC in vitro and in vivo. Meanwhile, bioactive components such as berberine and evodiamine have been demonstrated to inhibit the growth of tumors and reverse chemotherapy resistance in multiple tumors [21, 22]. The combination of berberine and evodiamine displays higher anticancer activity while reducing side effects in CRC cell lines [23]. In recent years, it has been found that miRNAs participate in chemoresistance by affecting drug efflux, cell apoptosis, and cell cycle arrest [24]. Liu found that exosomes derived from cisplatin-resistant oral squamous cell carcinoma cells can transfer miR-21 to sensitive cells and induce cisplatin resistance [25]. It is widely reported that miR-200c is involved in colon cancer progression. It has been reported that the overexpression of miR-200c inhibits the proliferation of colon cancer by targeting the FUT4/Wnt/β-catenin pathway [26]. Our study showed that miR-200c overexpression in CRC could inhibit the activation of the downstream JNK signaling pathway and reverse drug resistance. The addition of ZJW increased miR-200s and inhibited the activation of the JNK pathway. MALAT1 was one of the earliest discovered lncRNAs related to human diseases, and the high expression of MALAT1 predicts poor outcomes for CRC patients [27]. The expression level of MALAT1 is significantly increased in CRC patients and is related to advanced TNM stage, lymph node metastasis, and short survival [28]. MALAT1 has been proven to promote CRC cell proliferation and invasion by sponging miR-508-5p and enhancing RAB14 expression [29]. In addition, the expression of MALAT1 is prognostic for CRC patients treated with L-OHP. A high MALAT1 expression is associated with poor response to oxaliplatin-based chemotherapy and reduced survival in advanced CRC patients [30]. We found that MALAT1 was of great importance in the expression of miR-200s, and ZJW could regulate the expression of MALAT1 in drug-resistant CRC.
Our study identified that MALAT1 promotes colorectal cancer oxaliplatin resistance by reducing the miR-200s expression, and ZJW may reverse chemoresistance by inhibiting the expression of MALAT1 and regulating the miR-200s/JNK pathway. Thus, this study provides an experimental basis for the clinical application of ZJW to alleviate chemoresistance and improve the prognosis of CRC patients.
Zhenzhen Wei and Jing Zhou are co-first authors. | true | true | true |
||
PMC9568341 | Yun Jiang,Wei Ji,Jiaqi Zhu,Zihao Shen,Jianle Chen | Upregulation of miR-664a-3p Ameliorates Calcific Aortic Valve Disease by Inhibiting the BMP2 Signaling Pathway | 07-10-2022 | The development of calcific aortic valve disease (CAVD) is a complex process of ectopic calcification involving various factors that lead to aortic valve stenosis, hemodynamic changes, and, in severe cases, even sudden death. Currently, aortic valve replacement is the only effective method. The osteogenic differentiation of aortic valve interstitial cells (AVICs) is one of the key factors of valve calcification. Emerging evidence suggests that bone morphogenetic protein 2 (BMP2) can induce the proosteogenic activation of AVICs. However, the regulatory mechanism underlying this activation in AVICs is unclear. In the present study, we elucidated through high-throughput RNA sequencing and RT-qPCR that miR-664a-3p was evidently downregulated in the calcific aortic valve. We also proved that miR-664a-3p was involved in regulating osteogenic differentiation in AVICs. Target prediction analysis and dual-luciferase reporter gene assay confirmed that miR-664a-3p is preferentially bound to BMP2. Furthermore, the effect of the miR-664a-3p/BMP2 axis on osteogenic differentiation in AVICs was examined using the gain- and loss-of-function approach. Finally, we constructed a mouse CAVD model and verified the effect of the miR-664a-3p/BMP2 axis on the aortic valve calcification leaflets in vivo. In conclusion, miR-664a-3p regulates osteogenic differentiation in AVICs through negative regulation of BMP2, highlighting that miR-664a-3p may be a potential therapeutic target for CAVD. | Upregulation of miR-664a-3p Ameliorates Calcific Aortic Valve Disease by Inhibiting the BMP2 Signaling Pathway
The development of calcific aortic valve disease (CAVD) is a complex process of ectopic calcification involving various factors that lead to aortic valve stenosis, hemodynamic changes, and, in severe cases, even sudden death. Currently, aortic valve replacement is the only effective method. The osteogenic differentiation of aortic valve interstitial cells (AVICs) is one of the key factors of valve calcification. Emerging evidence suggests that bone morphogenetic protein 2 (BMP2) can induce the proosteogenic activation of AVICs. However, the regulatory mechanism underlying this activation in AVICs is unclear. In the present study, we elucidated through high-throughput RNA sequencing and RT-qPCR that miR-664a-3p was evidently downregulated in the calcific aortic valve. We also proved that miR-664a-3p was involved in regulating osteogenic differentiation in AVICs. Target prediction analysis and dual-luciferase reporter gene assay confirmed that miR-664a-3p is preferentially bound to BMP2. Furthermore, the effect of the miR-664a-3p/BMP2 axis on osteogenic differentiation in AVICs was examined using the gain- and loss-of-function approach. Finally, we constructed a mouse CAVD model and verified the effect of the miR-664a-3p/BMP2 axis on the aortic valve calcification leaflets in vivo. In conclusion, miR-664a-3p regulates osteogenic differentiation in AVICs through negative regulation of BMP2, highlighting that miR-664a-3p may be a potential therapeutic target for CAVD.
Calcified aortic valve disease (CAVD) is a high-risk disease in older people and is closely related to morbidity and mortality in patients with cardiovascular disease. It is also a major risk factor for cardiovascular disease complications, such as myocardial infarction [1–3]. So far, there are still no effective clinical interventions to reverse CAVD or halt its progression [4–6]. Valve replacement is the only effective clinical option [7]. Identification of the pathological mechanisms of CAVD can help its treatment. Aortic valve calcification is an active process involving complex changes, such as endothelial injury, lipid infiltration, chronic inflammation, matrix remodeling, cell differentiation, calcium salt deposition, and neovascularization [8]. Moreover, the active regulation of CAVD formation involves the synergistic action of multiple cells, including resident valvular endothelial cells, valvular interstitial cells, bone marrow-derived cells, and circulating inflammatory and immune cells [9]. Among them, valvular interstitial cells (VICs) play an important role in maintaining normal valve structure and function [10]. The activation of VICs in the normal dormant state is a major mechanism of the pathological process of aortic valve calcification. VICs can change to the osteoblast phenotype and accumulate calcium, phosphorus, and other inorganic salt ions, causing the cells to calcify [11]. The process of VIC transition may involve multiple signal transduction pathways [3, 12]. Therefore, the strategy of preventing VIC transformation by inhibiting osteoblast differentiation may lead to new therapeutic interventions to prevent and even reverse CAVD progression. Previous studies have shown BMP2 as an important proosteogenic factor involved in vascular and aortic valve calcification [13–15]. Inorganic phosphate osteogenic induction medium promotes VIC osteogenic differentiation via the BMP2 signaling pathway [16]. A previous study also claimed that concurrent upregulation of BMP2 and TGF-β1 is responsible for biglycan-induced proosteogenic reprogramming in human aortic VICs (AVICs) [17]. Therefore, the osteogenic effect of BMP2 on AVICs may play an important role in aortic valve calcification and CAVD progression. Studying the regulatory pathway of BMP2-induced osteogenic differentiation of AVICs will advance our understanding of the molecular mechanism of CAVD occurrence and development. In the search for effective therapeutics to treat CAVD, microRNA is an exciting candidate, as its expression can be manipulated using microRNA mimics or inhibitors [18]. MicroRNAs are a class of noncoding RNA with a size of 18–22 nt [19, 20]. They not mainly regulate the degradation of target mRNA but also regulate the translational inhibition depending on complementarity between the miRNA and mRNA [21]. MicroRNAs can regulate various physiological and pathological processes, such as cell proliferation, development, differentiation, and apoptosis; in fact, microRNA dysfunction often leads to impaired cell function [22–25]. Numerous studies have shown that the level of microRNAs changes dramatically during the osteogenic differentiation process in CAVD. For example, miR-138 suppresses the osteoblastic differentiation of VICs in degenerative CAVD [26]. miR-214 was found to inhibit aortic valve calcification in stretch-induced CAVD [27]. Moreover, miR-34a improved aortic valve calcification by regulating the Notch1-runt-related transcription factor 2 (Runx2) signaling pathway [28]. These studies indicate that microRNAs play a dual role in the process of aortic valve calcification. Previous studies have shown that microRNA regulates osteogenic differentiation by regulating BMP2 expression. For instance, BMP2 downregulated miR-30b and miR-30c to increase RUNX2 expression in human coronary artery smooth muscle cells and promote mineralization [19]. However, whether microRNA regulates the osteogenic differentiation of AVICs via the BMP2 pathway remains unclear. In the present study, we first compared the expression of microRNAs between the calcified aortic valve leaflets of CAVD patients and normal tissues through RNA sequencing. Subsequently, the results were verified by gain- and loss-of-function experiments to examine whether miR-664 participates in regulating AVIC osteogenic differentiation by influencing BMP2 expression. Finally, we constructed an animal model of CAVD and verified the effect of the miR-664a-3p/BMP2 axis on the calcification of aortic valve leaflets in vivo. Our findings show that miR-664a-3p participates in regulating AVIC osteogenic differentiation by negatively regulating BMP2 expression, highlighting that miR-664a-3p may be a potential therapeutic target for CAVD.
A total of 16 pairs of tissue samples were obtained from Jiangsu Province Hospital, China, including calcified aortic valve leaflets (CAVs), obtained via aortic valve replacement procedures and normal noncalcified aortic valves without thickening or nodules isolated via heart transplantation procedures. Samples with diseases such as rheumatic aortic valvulopathy, congenital valve disease, infective endocarditis, and autoimmune disease were excluded. All protocols using patient samples were approved by the Ethical Committee of Jiangsu Province Hospital. Written informed consent was obtained from the patients before surgery. All studies involving humans were conducted in accordance with the Helsinki Declaration as well as relevant guidelines and regulations.
MicroRNA was extracted and purified from human CAVs or normal aortic valve tissue by using the miRNeasy Mini Kit (217004; Qiagen, Germany). A small RNA Sample Prep Kit (RS200-0012; Illumina, Germany) was used to construct the microRNA library. MicroRNA expression was processed for 50 bp single-end reads using the miRDeep2 analysis pipeline. The cDNA was amplified by reverse transcription using primers that were complementary to the 3′ junction, and the microRNA library was obtained by 15 cycles of PCR amplification using the Illumina HiSeq 2000 high-throughput sequencing technology. The detailed operation was performed by Shanghai Kangcheng Biological Company. MicroRNA sequences were assigned names consistent with miRBase 20. R packages, DESeq2 and EdgeR, were utilized to normalize counts and calculate differential expression of microRNAs [29, 30].
Human AVICs were isolated from noncalcified heart valves using collagenase I, as described previously [28]. In brief, valve leaflets were digested in a medium containing 1.0 mg/mL collagenase type I for 30 min at 37°C. Valve endothelial cells were removed, and 1.0 mg/mL collagenase I fresh medium was added for 4-6 h at 37°C. After vortexing and repeated aspiration to break up the tissue mass, the cells were precipitated by centrifugation at 1000 rpm for 10 min. The temperature during centrifugation was 4°. Finally, the precipitated cells were resuspended and cultured in Dulbecco's Modified Eagle Medium (DMEM, Thermo Fisher, USA) containing 10% fetal bovine serum (Thermo Fisher) and 100 U/mL penicillin-streptomycin (Sigma, USA). Next, the cells were incubated in an osteogenic induction medium for 14 days to stimulate osteogenic differentiation, as previously described [31]. Complete DMEM was supplemented with 0.25 mmol/L L-ascorbic acid, 10 mmol/L β-glycerophosphate, and 10 nmol/L dexamethasone (Sigma).
BMP2 shRNA (F: 5′- AATTCAAAAAGTTCGAGTTGCGGCTGCTCAGCTCGAGCTGAGCAGCCGCAACTCGAAC-3′, R: 5′- CCGGGTTCGAGTTGCGGCTGCTCAGCTCGAGCTGAGCAGCCGCAACTCGAACTTTTTG-3′), scramble (F: 5′-AATTCAAAAAGCGCGATAGCGCTAATAATTTCTCGAGAAATTATTAGCGCTATCGCGC-3′, R: 5′-CCGGGCGCGATAGCGCTAATAATTTCTCGAGAAATTATTAGCGCTATCGCGCTTTTTG-3′), miR-664a-3p mimic/inhibitor (mimic-F: 5′- AUAAGUAAAUAGGGGUCGGAUGU-3′, mimic-R: 5′- AUCCGACCCCUAUUUACUUAUUU-3′, inhibitor: 5′- AUAAGUAAAUAGGGGUCGGAUGU-3′), and the corresponding control (mimics-NC-F: 5′-GAGAUGUUCAAUCGGGUAUUU-3′, mimics-NC-R: 5′- AUACCCGAUUGAACAUCUCUU-3′, inhibitor-NC: 5′- GAAUUACAUGCACCACUCAAU-3′) primers were all synthesized by Shanghai Gene Pharma. shRNA was inserted into the pLKO-TRC plasmid (Sigma-Aldrich). Human and mouse BMP2/Bmp2 coding sequences (human: F: 5′--GGGAGAAGGAGGAGGCAAAG --3′, R: 5′--GAAGCAGCAACGCTAGAAGACA--3′; mouse: F: 5′--GACATCCGCTCCACAAACGA--3′, R: 5′--CATCACTGAAGTCCACATACAAAGG--3′) were inserted into the overexpressed pcDNA3.1 plasmid vector (Invitrogen). All lentiviral solutions were obtained from Gene Pharma. VICs (1 × 105 cells/mL) were seeded in a 6-well plate and cultured in an osteogenic medium. After overnight incubation at 37°C for wall attachment, 200 μL of 1 × 108 transduction units (TU)/mL lentiviral solution was added. The medium was replaced after 20 h of infection. After 48 h of infection, 3 ng/μL puromycin (Solarbio) was added to screen for successful infection of cells.
A CAVD animal model was established as described previously [32]. Male ApoE−/− (C57BL/6 background) mice aged 6–8 weeks were purchased from Nanjing University Animal Center, China, and housed under specific pathogen-free conditions and 12 h dark-light cycle in individually ventilated cages at 22°C, with free access to food and water. The mice were fed a 0.2% high cholesterol diet for 24 weeks to develop aortic valve calcification [33]. Next, the mice were randomly divided into four groups of 12 mice per group, namely, control, miR-664a-3p mimic, Bmp2 overexpression (OE), and miR − 664a − 3p mimic + Bmp2 OE. The lentivirus of miR-664a-3p and Bmp2 OE vectors was obtained from GenePharma (Shanghai, China). The lentiviruses were injected into the mice via the tail vein twice a week for another 10 weeks. Animal care and euthanasia were carried out with the approval of the Institutional Animal Care and Use Committee of Nanjing Medical University.
Western blotting analysis was performed based on previous reports [34]. Protein was extracted from VICs and aortic valve tissues using RIPA lysis buffer with protease inhibitors (Thermo Fisher). Next, 20 μg of the protein solution was fractionated by 10% SDS-PAGE and then transferred to a polyvinylidene difluoride membrane (Millipore). After blocking with 5% nonfat milk, the membrane was incubated at 4°C overnight with primary antibodies, including anti-BMP2, anti-RUNX2, anti-MST2, anti-OSTERIX, and anti-GAPDH (as an internal reference). Subsequently, the membrane was washed with Tris-buffered saline-Tween 20 (TBST) solution and incubated with goat antirabbit HRP-conjugated secondary antibody for 2 h at room temperature. All antibodies were purchased from Abcam and used in accordance with the manufacturer's instructions. Protein bands were visualized using an enhanced chemiluminescence kit (Vazyme) and quantified using the ImageJ software.
Total RNA was isolated from VICs or aortic valve tissues using a miRNeasy Mini Kit (Qiagen) and then reverse-transcribed to cDNA using the PrimeScript RT Master Mix (Takara, Japan) with oligo (dT) or random primer. After that, RT-qPCR was performed using the PowerUp SYBR Green Master Mix (Thermo Scientific, USA) on an Applied Biosystems 7500 Fast Real-Time detection system (Applied Biosystems). U6 or GAPDH gene was used as the internal control for measuring relative gene expression. All experiments were independently repeated three times [35, 36]. The primer sequences for miR-664a-3p and U6 are listed in Table S1.
VICs were washed by PBS and fixed with 4% PFA for 30 min. The cells were subsequently stained with 2% Alizarin Red S solution (Sigma) at room temperature for 30 min. Red staining indicates the formation of calcified nodules. Next, the cells were washed by PBS and observed under a light microscope (Olympus). Quantitative analysis was performed using the ImageJ software.
VICs were washed twice with PBS. Proteins were then extracted with 1% Triton X-100 and centrifuged at 10,000 rpm for 10 min. ALP activity was measured using the 5-bromo-4-chloro-3-indolyl phosphate (BCIP)/nitro blue tetrazolium (NBT) alkaline phosphodiesterase chromogenic Kit (C3206; Beyotime) according to the manufacturer's protocol. The stained samples were observed under a light microscope (Olympus). Quantitative analysis was performed using the ImageJ software.
Mouse and human aortic valves were removed and fixed with 4% paraformaldehyde for 24 h. The tissues were then paraffin-embedded and cut into sections of 5 μm thickness. Subsequently, the sections were dewaxed and stained with HE (C0105S; Beyotime) according to the manufacturer's protocol. Finally, the stained sections were observed under a light microscope (Olympus).
A Cy3-labeled anti-digoxin miR-664a-3p probe (5′-ATAAGTATAGGGGTCGGATGT-3′) was obtained from GenePhama. VICs cultured in osteogenic differentiation medium were considered as “case,” and those cultured in the normal medium were considered as “control.” The VICs were inoculated at a density of 1 × 104 cells/well in 48-well plates (the wells were pretreated with appropriately sized coverslips). The cells were then mixed and incubated overnight at 37°C in a 5% CO2 incubator. The next day, the medium was removed, and the cells were washed twice with PBS. After that, the PBS was removed, and 100 μL of 4% paraformaldehyde was added to each well. The cells were then fixed for 15 min at room temperature. RNA-FISH kit (cell sliver; GenePhama) was used for the subsequent experiments, according to the manufacturer's instructions. Finally, the cell nuclei were stained with 4′-6-diamidino-2-phenylindole (DAPI), and images were captured using a Zeiss LSM 700 confocal microscope (Carl Zeiss).
BMP2 wild-type 3′-UTR containing putative miR-664a-3p binding sites was inserted into the pGL3 control luciferase reporter vector (Promega, USA). To assess binding specificity, the sequence interacting with miR-664a-3p was mutated by using the Q5® Site-Directed Mutagenesis Kit Protocol (New England Biolabs), and the mutated BMP2 (BMP2-MUT) 3′-UTR was also inserted into the pGL3 plasmid. VICs were cultured in 24-well plates and transfected with Lipofectamine 3000 (Thermo Scientific). Next, 1 μg of luciferase reporter plasmid was added to each well, followed by treatment with 0.2 μg of pRL-TK Renilla luciferase plasmid (internal control), 100 pmol/well of either miR-664a-3p mimic, miR-664a-3p inhibitor, or the corresponding control. Luciferase activity was measured at 48 h posttransfection using a dual-luciferase reporter system (Promega). The ratio of firefly to Renilla luciferase activity was determined to eliminate variations in transfection efficiency [32].
All experiments were repeated three times. The data were analyzed on SPSS 18.0, and the results are presented as the mean ± SEM. The student's t-test was used for the comparison of two value sets. One-way ANOVA or two-way ANOVA followed by Dunnett's T3 post hoc test was conducted to analyze differences between more than two groups. The association between two variables was evaluated by two-tailed Pearson's correlation analysis. GraphPad Prism 7.0 was used for data analysis. P < 0.05 indicates a statistically significant result. All measurements were performed in triplicate.
We first performed HE staining of human CAVs and normal AVs (Figure 1(a)), and the results showed that patients in the CAVD group had calcified tissue lesions, which caused the valve to lose its original uniform dense structure. Furthermore, we performed western blotting analysis to determine the protein expression of osteogenic genes, including BMP2, RUNX2, MST2, and OSTERIX, which are essential for bone formation [37–39]. As shown in Figure 1(b), BMP2, RUNX2, MST2, and OSTERIX were significantly upregulated in CAVD patients. To investigate the role of microRNAs in CAVD, RNA sequencing was performed to identify the difference in microRNA expression between 3 CAVD and normal samples each. The volcano plot showed differential expression of microRNAs, with miR-664a-3p (labeled in the plot; Figure 1(c)) being the most downregulated microRNA in the CAVD group. Clustering analysis of the top 20 differentially-expressed microRNAs was performed, and the results are shown as a heat map, with 4 downregulated and 16 upregulated microRNAs in the CAVD group. qPCR results confirmed that miR-664 was the most downregulated microRNA in CAVD tissues (Figure 1(e)).
We assessed changes in the expression of osteogenic differentiation-related indicators to examine whether miR-664a-3p influences the development of CAVD by modulating miR-664a-3p expression in VICs. RNA-FISH analysis results showed that miR-664a-3p was localized in the cytoplasm of VICs. Compared to the control, VICs cultured with osteogenic differentiation medium had a higher level of miR-664a-3p (Figure 2(a)). We used a mimic and an inhibitor to regulate miR-664a-3p expression in VICs and then measured gene expression by RT-qPCR (Figure 2(b)). Western blotting analysis showed a significant negative correlation between the expression of miR-664a-3p and osteogenic differentiation-related proteins, including RUNX2, MST2, and OSTERIX (Figure 2(c)). Alizarin Red staining and ALP activity assay results showed that the downregulation of miR-664a-3p promoted the formation of calcium nodules and increased ALP activity (Figures 2(d) and 2(e)). All these results suggested that miR-664a-3p inhibited the osteogenic differentiation and calcification of VICs.
The potential binding site between miR-664a-3p and the 3′-UTR of BMP2 was predicted (Figure 3(a)). A dual-luciferase reporter gene assay was conducted to examine the interaction between miR-664a-3p and BMP2 in VICs. The results showed that luciferase activity was significantly lower in the BMP2-WT group after the expression of miR-664a-3p was enhanced, whereas reducing the level of miR-664a-3p led to an opposite result. However, in the BMP2-MUT group, luciferase activity was not affected by the expression level of miR-664a-3p (Figure 3(b)). RT-qPCR and western blotting analyses revealed that miR-664a-3p negatively regulated the expression of BMP2 (Figures 3(c) and 3(d)). The cycle threshold (Ct) values of miR-664a-3p and BMP2 corresponding to each sample were obtained by RT-qPCR analysis of 5 CAV tissues and normal AV tissues each. Pearson's correlation analysis based on the Ct values revealed a significant negative correlation between the two: miR-664a-3p expression was lower, whereas BMP2 expression was higher in the disease group. As shown in Figure 3(d), the data in red boxes were from CAV tissues, whereas the data in blue boxes were from normal AV tissues. The above results fully confirmed that miR-664a-3p could target BMP2.
Given that BMP2 was overexpressed in CAVs and a target of miR-664a-3p, which was associated with the osteoblast differentiation of VICs, we performed overexpression and knockdown experiments in VICs to investigate whether BMP2 could reprogram VICs toward an osteogenic phenotype. First, western blotting results confirmed that BMP2 expression in VICs was successfully regulated (Figure 4(a)). After the expression of BMP2 was altered, we then examined the expression of other osteogenic differentiation-related proteins (RUNX2, MST2, and OSTERIX), which was positively correlated with BMP2 expression (Figure 4(b)). This result suggested that BMP2 regulated the osteogenic differentiation of VICs by affecting the expression of associated genes. Alizarin Red and ALP staining assays showed that the overexpression of BMP2 enhanced the calcification of VICs, whereas BMP2 knockdown inhibited VIC calcification (Figures 4(c) and 4(d)). Collectively, these results indicated that BMP2 played a positive role in the osteoblast differentiation and calcification of VICs.
To verify the above experimental results, we conducted animal experiments in CAVD -induced mice. We overexpressed miR-664a-3p and Bmp2 separately or simultaneously in CAVD mice by lentiviral transfection. Total RNA was extracted from mouse aortic valve tissues, and RT-qPCR analysis was performed. The results revealed a significant increase in miR-664a-3p and Bmp2 expression after infection with the corresponding virus (Figure 5(a)). In addition, Bmp2 expression was significantly downregulated after infection with the miR-664a-3p overexpression lentivirus (Figure 5(b)). The subsequent western blotting analysis yielded the same result. miR-664a-3p overexpression significantly downregulated the expression of Bmp2, Runx2, Mst2, and Osterix. Moreover, the overexpression of Bmp2 upregulated the expression of Runx2, Mst2, and Osterix; however, this phenomenon was reversed by the upregulation of miR-664a-3p (Figure 5(c)). This observation further suggested that miR-664a-3p also downregulated Bmp2 expression in an in vivo setting and that the latter may influence aortic valve leaflets osteogenic differentiation by positively regulating the expression of genes involved in osteogenic differentiation. HE staining of the aortic valve showed a uniform density of valve cells in the miR-664a-3p overexpression group, as well as inflammatory infiltration and new capillary formation in the valve after BMP2 upregulation, whereas the histology of the mice in both overexpression groups was similar to that in the control group (Figure 5(d)). The above experiments showed that miR-664a-3p improved the calcification of aortic valve leaflets by targeting Bmp2 in vivo. Finally, the biological process and mechanism of action of this paper are shown in Figure 6.
CAVD is a cardiovascular disease with high morbidity and mortality, especially in the elderly [40, 41]. Studies have shown that calcification plays an important role in the pathogenesis of CAVD, and the osteogenic differentiation of AVICs has been confirmed to be closely related to the pathological process of CAVD [42]. However, there are still no effective pharmacological treatments to prevent or treat this disease. Therefore, the study of the regulatory mechanism of AVIC osteogenic differentiation may contribute to a better understanding of the pathogenesis of CAVD and provide a new perspective for the treatment of the disease. In the present study, we detected the protein expression of BMP2 in CAVD patient samples and non-CAVD samples. The result showed that the BMP2 protein expression level in CAVD patient samples was significantly higher than that in normal AV samples. This result indicates that BMP2 is indeed involved in the regulation of CAVD and has a posttranscriptional regulation mechanism. RNA sequencing of CAVs and normal tissue revealed many differentially expressed microRNAs, and miR-664a-3p, which was the most downregulated microRNA in calcified valves, was selected for subsequent studies. However, several other microRNAs were also significantly downregulated and could be candidates for follow-up studies. In addition, many other microRNAs were significantly upregulated in CAVD tissues, showing great potential for further study. Although we have not explored these microRNAs, we will continue to study them in-depth in the hope of gaining a more comprehensive understanding of the important role of microRNAs in CAVD development. Subsequently, we examined the expression of osteogenic differentiation-related proteins after altering miR-664a-3p expression in VICs. Our findings showed that a low expression of miR-664a-3p exacerbated calcification in VICs. miR-664a-3p was found to target and bind to BMP2, thus downregulating BMP2 expression. The relationship between the two was confirmed by dual-luciferase reporter assay and Pearson's correlation analysis. Cardiac valve calcification is the active conversion of VICs to an osteoblast-like cell phenotype, and it involves the regulation of diverse osteogenic factors [43], including the promotion of BMP2 [16]. We again assessed the effect of BMP2 levels on calcification in VICs by altering the expression of BMP2. Not surprisingly, high BMP2 expression exacerbated calcification in VICs, as evidenced by ARS and ALP staining results. In addition, western blotting results showed that BMP2 levels were positively correlated with common osteogenic differentiation-related proteins such as RUNX2, MST2, and OSTERIX. A previous study reported that the formation of the RUNX2-SMAD regulatory complex was obligatory for activating a gene network that drives osteoblast differentiation. As a molecular endpoint, RUNX2 was required to execute and complete TGF-/BMP2 signaling in osteoblasts [44]. Finally, we used ApoE−/− mice to construct a CAVD model by controlling their diet. Next, we upregulated the levels of miR-664a-3p and Bmp2 in the CAVD mice by long-term lentiviral injection. We examined the morphological changes in the aortic valves after incubation for a period of time. Our results showed that a high expression of BMP2 exacerbated aortic valve calcification in mice, whereas the upregulation of miR-664a-3p had the opposite effect. This result confirmed, in an in vivo setting, the important role of miR-664a-3p/Bmp2 in the development of aortic valve calcification.
In conclusion, this study reveals, for the first time, that miR-664a-3p regulates aortic valve calcification by targeting BMP2. The findings of the present study advance mechanistic studies related to the development of CAVD and provide a potential therapeutic target for improving the outcome of CAVD. | true | true | true |
||
PMC9568347 | Runan Zhang,Genhua Zhang,Baohua Li,Juan Wang,Jvfang Wang,Jia Che,Xiaojun Wang,Zhen Zhang | Analysis of LINC01314 and miR-96 Expression in Colorectal Cancer Patients via Tissue Microarray-Based Fluorescence In Situ Hybridization | 07-10-2022 | Methods A tissue microarray (TMA) containing 76 individual colorectal tumor samples and 28 adjacent normal samples was constructed, and the expression levels of LINC01314 and miR-96 were detected by fluorescence in situ hybridization. Results The expression levels of both LINC01314 and miR-96 were upregulated in CRC tissues and were associated with vascular metastasis (p < 0.05). A significantly positive correlation was observed between LINC01314 and miR-96 expression in tumor tissues (p < 0.001, r = 0.870). Dominant expression of LINC01314 was a risk factor for both blood vessel invasion (p < 0.05) and poor 5-year survival (p = 0.001, hazard ratio = 4.144). The Kaplan–Meier analysis indicated that patients with LINC01314-dominant expression exhibited worse 5-year survival rates than those with miR-96-dominant expression (p < 0.05). Conclusion The expression patterns of both LINC01314 and miR-96 may be diagnostic of, and prognostic for, CRC. These findings will facilitate further exploration of the molecular mechanism of lncRNAs in CRC. | Analysis of LINC01314 and miR-96 Expression in Colorectal Cancer Patients via Tissue Microarray-Based Fluorescence In Situ Hybridization
A tissue microarray (TMA) containing 76 individual colorectal tumor samples and 28 adjacent normal samples was constructed, and the expression levels of LINC01314 and miR-96 were detected by fluorescence in situ hybridization.
The expression levels of both LINC01314 and miR-96 were upregulated in CRC tissues and were associated with vascular metastasis (p < 0.05). A significantly positive correlation was observed between LINC01314 and miR-96 expression in tumor tissues (p < 0.001, r = 0.870). Dominant expression of LINC01314 was a risk factor for both blood vessel invasion (p < 0.05) and poor 5-year survival (p = 0.001, hazard ratio = 4.144). The Kaplan–Meier analysis indicated that patients with LINC01314-dominant expression exhibited worse 5-year survival rates than those with miR-96-dominant expression (p < 0.05).
The expression patterns of both LINC01314 and miR-96 may be diagnostic of, and prognostic for, CRC. These findings will facilitate further exploration of the molecular mechanism of lncRNAs in CRC.
Colorectal cancer (CRC) is the second most common cause of cancer mortality worldwide, with more than 1.93 million diagnoses and 930,000 deaths annually (World Cancer Report 2020) [1]. In China, 550,000 cases and 280,000 deaths are reported each year [2]. Early CRC is curable, and the tumors can be removed surgically. Unfortunately, CRC is often advanced when diagnosed, associated with distant metastases [3]. Biomarkers that detect early cancer and/or predict prognosis have been extensively studied [4–6]. However, 5-year survival remains poor; CRC pathogenesis is not well-understood. Novel biomarkers facilitating early diagnosis, prognostic predictions, and treatment are urgently required. Long noncoding RNAs (lncRNAs) have recently attracted attention as biomarkers for CRC diagnosis and prognostication [7]. lncRNAs (>200 nucleotides in length) are widespread in many species [8]. Accumulating evidence suggests that lncRNAs engage in transcriptional regulation (thus gene-specific transcription) and posttranscriptional and epigenetic regulation [9, 10]. Several lncRNAs are aberrantly expressed in various cancers and function as tumor suppressors, promoters, or both under certain conditions by combining with proteins or nucleotide sequences to regulate downstream molecules [11–15]. For example, growth arrest specific 5 (GAS5) and LINC01559 exhibit antioncogenic effects in CRC development or progression. Low-level expression of lncRNA GAS5 and/or LINC01559 is associated with a poor prognosis in CRC patients [16, 17]. Meanwhile, HOX transcript antisense RNA (HOTAIR) and colon cancer-associated transcript 1 (CCAT1) were found to be upregulated in the early stages of colorectal carcinogenesis and associated with TNM stage and poor overall survival [18–20]. Similarly, LINC01314 was demonstrated to repress gastric cancer progression by modulating Wnt/β-catenin signaling [21]. Lv et al. reported that LINC01314 overexpression reduced hepatoblastoma cell proliferation and migration [22]. However, few reports have explored the expression levels, molecular effects, and clinical significance of LINC01314 in CRC patients. MicroRNAs (miRNAs) are single-stranded noncoding RNAs that regulate gene expression via base-pairing. Interactions between lncRNAs and miRNAs regulate several biological and pathological processes [23–25]. The molecular details of lncRNA–miRNA crosstalk in terms of CRC progression were summarized by Wang et al. [26]. TargetScan revealed that LINC01314 shared a binding site with miR-96; the molecules may interact. We thus explored LINC01314 and miR-96 expressions in CRC patients via tissue microarray- (TMA-) based fluorescence in situ hybridization (FISH). We speculated that LINC01314 and miR-96 expression patterns might be both diagnostic of, and prognostic for, CRC.
Tumor tissues and adjacent normal tissues were obtained from patients who underwent surgery to treat primary CRCs in the Tongde Hospital of Zhejiang Province (People's Republic of China) in 2015 and 2016. All specimens were independently diagnosed histologically by three experienced pathologists by reference to the NCCN Clinical Practice Guidelines in Oncology for Colon Cancer (ver. 3, 2013) [27]. Residual tissues were immediately frozen in liquid nitrogen. The study was approved by the Ethics Committee of Tongde Hospital, Zhejiang Province (reference number: 2021025). All patients provided written informed consent in accordance with the 1975 Declaration of Helsinki. Patients were followed-up every 6 months for 5 years after primary surgery; survival, the dates of any events, and the causes of death were recorded. The median overall survival was 48 months, and the patient age ranged from 34 to 95 years (median 70 years). Clinicopathological data (tumor size; pathological pattern; blood vessel invasion, lymph node metastasis, and nerve invasion statuses; and TNM stage) were retrieved from pathology reports lodged in the hospital information system. The clinical and pathological characteristics of 76 CRC patients are listed in Table 1.
A colorectal TMA was constructed as described previously [28–30]. Briefly, tumor and adjacent normal tissues were fixed in 4% (v/v) formalin and embedded in paraffin. Donor blocks were subjected to hematoxylin and eosin (H&E) staining to identify representative tumor regions. Tissue cylinders (6 mm diameter) were punched from target areas and inserted into recipient paraffin blocks using an automatic precision instrument. Each TMA block featured 76 individual colorectal tumor samples and 28 adjacent normal samples. Each TMA block was then cut into several 4 μm thick sections (HistoCore BIOCUT, Leica, Wetzlar, Germany), and the sections were mounted on glass slides for H&E staining and FISH.
LINC01314 and miR-96 in TMA samples were detected via FISH, as described previously [16, 31–33]. After deparaffinization and air-drying, TMA slides were immersed in DEPC-treated RNase-free water and incubated in 0.01 M citric acid buffer (pH 6.0) at 95°C for 10 min, followed by proteinase K digestion for 20 min. After prehybridization for 1 h, slides were incubated with a 1 μM solution of the Spectrum-CY3-labeled miR-96 probe (5′-CY3-GCAAAAATGTGCTAGTGCCAAA-CY3-3′) and the Spectrum-FAM-labeled LINC01314 probe (5′-FAM-GGTGGATGTGGGGATGGCGCTGTAAGGG-FAM-3′) in hybridization buffer overnight at 42°C in a humidified chamber. The slides were washed in graded SCC solutions (2×, 1×, and 0.5× SCC for 10 min each), and the nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Cell Signaling Technology, Danvers, MA, USA) for 8 min. Images were obtained using a fluromicroscope (Nikon ECLIPSE C1, Tokyo, Japan) at 100x magnification.
Statistical analysis was performed using SPSS ver. 26.0 (IBM Corp., Armonk, NY, USA) and GraphPad Prism ver. 9.0 (GraphPad Software, San Diego, CA, USA). LINC01314 and miR-96 levels in tumor tissues were subjected to the Pearson correlation analysis. Associations between LINC01314 and miR-96 expressions and pathological characteristics were evaluated using Student's t-test and ANOVA. Possible risk factors for tumor vascular invasion and 5-year mortality were explored using logistic regression and Cox's regression analyses. The survival rates of the LINC01314- and miR-96-dominant groups were compared using the Kaplan–Meier method. A p value < 0.05 was considered significant.
TMA blocks containing 76 and 28 tumor and normal tissue samples, respectively, were constructed. A complete H&E-stained block is shown in Figure 1(a). One H&E-stained TMA spot is shown in Figures 1(b) and 1(c). FISH was used to detect LINC01314 and miR-96 expressions (Figure 2). Representative TMA sections exhibiting LINC01314 and miR-96 expressions are shown in Figure 3. The extent of RNA expression was measured by recording the integrated optical density (IOD) using ImageJ software. As normal tissue sections barely fluoresced, data were compared based on the relative IODs (IOD of tumor tissues divided by the mean IOD of normal tissues). The mean relative IODs of LINC01314 and miR-96 were significantly higher in tumor tissues than normal tissues (77.3 ± 232.1 vs. 1 and 31.2 ± 62.5 vs. 1, respectively; data not shown). A significantly positive correlation was observed between LINC01314 and miR-96 expressions in tumor tissues (p < 0.001, r = 0.870, Figure 4).
Tissue samples were divided into several groups according to the clinicopathological characteristics, and LINC01314 and miR-96 expression levels were compared. In contrast to the vascular nonmetastasis group, tissues in the metastasis group expressed significantly lower levels of LINC01314 (21.89 ± 39.178 vs. 109.15 ± 284.897, p < 0.05) and miR-96 (10.30 ± 15.292 vs. 43.98 ± 75.010, p < 0.05, Table 1). ANOVA revealed that LINC01314 expression was associated with tumor invasion depth (p < 0.05, data not shown) and miR-96 expression was associated with tumor differentiation (p < 0.05). However, neither expression level differed by age, sex, tumor size, histological type, TNM stage, lymph node metastasis status, nerve invasion, or survival.
To further analyze the relationships of LINC01314 and miR-96 expressions with the clinical characteristics of CRC patients, we introduced the concept of dominant expression. If the relative IOD of LINC01314 was higher than that of miR-96 in a TMA section, the patient was considered to exhibit LINC01314-dominant expression and vice versa. Of the 76 CRC patients, 52 exhibited LINC01314-dominant expression and 24 miR-96-dominant expression. Potential risk factors for tumor vascular metastasis (age, sex, tumor size, tumor location, histological type, extent of differentiation, TNM grade, nerve invasion, and LINC01314-dominant expression) were evaluated via logistic regression analysis. As shown in Table 2, tumor vascular metastasis was significantly associated with both the TNM grade and LINC01314-dominant expression (p = 0.003 and p = 0.029, respectively). Cox's regression analysis indicated that age (p = 0.029, hazard ratio (HR) = 1.029), TNM grade (p = 0.015, HR = 0.470), and LINC01314-dominant expression (p = 0.001, HR = 4.144) were significantly associated with poor 5-year survival (Table 3). Thus, LINC01314-dominant expression is a risk factor for both tumor vascular metastasis and poor 5-year survival in CRC patients.
The mean survival time of the 76 CRC patients was 44.2 ± 17.1 months (range 12–60 months); 44.2% (23/52) of patients with LINC01314-dominant expression remained alive during follow-up. The 5-year survival rate of the patients with miR-96-dominant expression was 54.2% (13/24). The Kaplan–Meier analysis indicated a significant difference in survival curves between the two groups of patients (p = 0.048). As shown in Figure 5, patients with LINC01314-dominant expression exhibited worse 5-year survival rates.
CRC is one of the most frequently diagnosed cancers worldwide, including China, and over 80% cases are initially diagnosed at an advanced stage [34]. It is critical to develop indicators aiding diagnosis and/or predicting prognosis. We are the first to simultaneously evaluate LINC01314 and miR-96 expressions in CRC TMA blocks via FISH and to explore the correlations between expression patterns and clinicopathological characteristics. Both LINC01314 and miR-96 expression levels were significantly higher in tumor tissues and were associated with vascular metastasis. Cox's regression analysis showed that LINC01314-dominant expression was associated with an increased risk of death in CRC patients. LINC01314 and miR-96 expression patterns will aid the diagnosis and/or prognosis of CRC patients. miR-96 is involved in many critical cellular processes including proliferation, differentiation, and apoptosis [23, 35]. However, the role played by miR-96 in colorectal carcinogenesis remains unclear. Yue et al. reported that miR-96 triggers CRC development and progression via the AMPKα2-FTO-m6A/MYC axis [36]. Ress et al. suggested that lower miR-96 values were associated with metastases and shorter survival in CRC patients [37]. In vitro, overexpression of miR-96 reduced cellular growth as reflected by increased p27-CDKN1A and decreased cyclin D1 expression [37]. We found that miR-96 was expressed more highly in tumor than normal tissues and that lower expression was associated with vessel invasion, consistent with the Ress data. The Cancer Genome Atlas RNA-seq data show that LINC01314 is aberrantly expressed in various tumors, showing upregulation in thyroid carcinoma but downregulation in cholangiocarcinoma, esophageal carcinoma, kidney chromophobes, kidney renal papillary cell carcinoma, kidney carcinoma, lung adenocarcinoma, pheochromocytoma, and paraganglioma [12]. However, LINC01314 expression and function in CRC have not been investigated. We found that LINC01314 expression was higher in CRC tissues than normal tissues and that lower expression was associated with tumor invasion. These results improve our understanding of the role played by LINC01314 during colorectal carcinogenesis. The lncRNA–miRNA–mRNA axis is a novel regulatory mechanism featuring interactions among lncRNAs, miRNAs, and mRNAs, and it plays a crucial role in the pathophysiological steps of tumor carcinogenesis, progression, and metastasis [38–41]. Most CRC-related lncRNAs have been reported to be upregulated and appear to function as miRNA sponges [7]. lncRNAs are involved in a variety of tumor-related pathways, such as the estimated growth factor receptor (EGFR), Wnt, and p53 signaling pathways, by regulating miRNAs [41]. For example, nuclear-enriched abundant transcript 1 (NEAT1) promoted CRC tumorigenesis through various lncRNA/miRNA axes, such as the NEAT1/miR-495-3P/CDK6 [42], NEAT1/miR-34a/SIRT1/Wnt/-catenin [43], and NEAT1/miR-205-5p/VEGFA axes [44]. In the present study, LINC01314 and miR-96 expression levels were found to be positively correlated in CRC tumor tissues, and both were associated with vessel invasion. Bioinformatics analysis suggested that a binding site is shared by LINC01314 and miR-96. We suggest that LINC01314–miR96 is a novel epigenetic regulatory axis involved in CRC development. More importantly, LINC01314-dominant expression was associated with higher risks of vessel invasion and poorer survival compared with miR-96-dominant expression in CRC patients. We speculate that LINC01314 may promote the development of CRC by reducing the ability of miR-96 to slow tumor progression, thereby reducing the survival of CRC patients. These speculations and the underlying crosstalk mechanisms between LINC01314 and miR-96 in CRC development will be demonstrated in our future in vitro experiments. N6-methyladenosine (m6A) modification is among the most ubiquitous epigenetic modifications of mRNAs and noncoding RNAs (e.g., miRNAs and lncRNAs) [45]. Overwhelming evidence indicates that the dysregulation of m6A modification is significantly correlated with CRC tumorigenesis and progression [46, 47]. lncRNAs and miRNAs are not only important targets of m6A modification regulators; they also regulate m6A modification [48, 49]. Whether LINC01314 and miR-96 affect the development of CRC carcinogenesis through m6A methylation modification remains to be explored in a future study. LINC01314 and miR-96 expressions were detected via TMA-based FISH, which affords many advantages compared with traditional methods. Aggregation of many tissues and experimentation under identical conditions optimize standardization [50]. A single tumor block can be cut into many sections, and repeat evaluations are possible. However, there were two major limitations that we plan to address. First, any retrospective study is associated with a risk of selection bias. Second, the interactions between LINC01314 and miR-96 and their roles in CRC progression must be investigated in vitro. We will establish a prognostic CRC model featuring lncRNA and miRNA detection and perform a large, prospective cohort study. This research will improve our understanding of the molecular mechanism of lncRNAs and miRNAs in CRC development and provide a novel strategy for the clinical diagnosis and prognosis of CRC.
In conclusion, we found that LINC01314 and miR-96 expression levels were upregulated in CRC tissues and were associated with vascular metastasis. LINC01314-dominant expression was a risk factor for tumor vascular invasion and poor 5-year survival in CRC patients. LINC01314 and miR-96 may be used as novel biomarkers for the diagnosis, prognostic predictions, and treatment of CRC. Combined detection of the expression of LINC01314 and miR-96 in tumor tissues and expression pattern analyses will aid CRC diagnosis and prognosis. | true | true | true |
||
PMC9568348 | Bismark Opoku Mensah,Linda Ahenkorah Fondjo,W. K. B. A. Owiredu,Ben Adusei | Urinary PCA3 a Superior Diagnostic Biomarker for Prostate Cancer among Ghanaian Men | 07-10-2022 | Introduction Prostate cancer is one of the most commonly diagnosed cancers in men. Prostate-specific antigen (PSA) has been the biomarker of choice for screening and diagnosis of prostate cancer. However, inefficiencies exist with its diagnostic capabilities. This study thus evaluated the diagnostic and prognostic potential of urinary PCA3 as an alternative biomarker for prostate cancer in the Ghanaian population. Methods A hospital-based cross-sectional study was conducted at the Urology Department of the 37 Military Hospital, Accra, Ghana. A total of 237 participants aged 40 years and above with any form of suspected prostate disorder were recruited into the study after written informed consent was obtained. Total serum PSA levels was measured using the electrochemiluminescence method and transrectal ultrasound-guided systematic core needle biopsies were obtained from each study participant. Receiver operating characteristic curve (ROC) analysis was used to evaluate the diagnostic accuracies of serum PSA, DRE, and PCA3 as diagnostic tools for prostate cancer. These three diagnostic tools were also evaluated in various combinations to ascertain the combinations with the best diagnostic accuracy. Results Prostate cancer was diagnosed in 26.6% of the participants. Benign prostate hyperplasia and prostatitis were diagnosed in 48.5% and 24.9% participants, respectively. DRE had a sensitivity of 93.7% and a specificity of 12.1%. PSA had a sensitivity of 92.1% and a specificity of 16.1%. PCA3 had a sensitivity of 57.1% and a specificity of 85.6% and showed a better accuracy (AUC = 83.0) compared to PSA (AUC = 60.0) and DRE (AUC = 65.0) as individual diagnostic tools. The combination of DRE+PCA3 score had the best diagnostic accuracy (AUC = 0.80) with a sensitivity and specificity of 60.3% and 80.5%, respectively. Conclusion The urinary PCA3 assay showed a better diagnostic performance compared to serum PSA and DRE. PCA3 as a stand-alone and in combination with DRE could be a suitable complimentary marker in diagnosis and management of prostate cancer. | Urinary PCA3 a Superior Diagnostic Biomarker for Prostate Cancer among Ghanaian Men
Prostate cancer is one of the most commonly diagnosed cancers in men. Prostate-specific antigen (PSA) has been the biomarker of choice for screening and diagnosis of prostate cancer. However, inefficiencies exist with its diagnostic capabilities. This study thus evaluated the diagnostic and prognostic potential of urinary PCA3 as an alternative biomarker for prostate cancer in the Ghanaian population.
A hospital-based cross-sectional study was conducted at the Urology Department of the 37 Military Hospital, Accra, Ghana. A total of 237 participants aged 40 years and above with any form of suspected prostate disorder were recruited into the study after written informed consent was obtained. Total serum PSA levels was measured using the electrochemiluminescence method and transrectal ultrasound-guided systematic core needle biopsies were obtained from each study participant. Receiver operating characteristic curve (ROC) analysis was used to evaluate the diagnostic accuracies of serum PSA, DRE, and PCA3 as diagnostic tools for prostate cancer. These three diagnostic tools were also evaluated in various combinations to ascertain the combinations with the best diagnostic accuracy.
Prostate cancer was diagnosed in 26.6% of the participants. Benign prostate hyperplasia and prostatitis were diagnosed in 48.5% and 24.9% participants, respectively. DRE had a sensitivity of 93.7% and a specificity of 12.1%. PSA had a sensitivity of 92.1% and a specificity of 16.1%. PCA3 had a sensitivity of 57.1% and a specificity of 85.6% and showed a better accuracy (AUC = 83.0) compared to PSA (AUC = 60.0) and DRE (AUC = 65.0) as individual diagnostic tools. The combination of DRE+PCA3 score had the best diagnostic accuracy (AUC = 0.80) with a sensitivity and specificity of 60.3% and 80.5%, respectively.
The urinary PCA3 assay showed a better diagnostic performance compared to serum PSA and DRE. PCA3 as a stand-alone and in combination with DRE could be a suitable complimentary marker in diagnosis and management of prostate cancer.
Prostatic carcinoma is one of the cancers mostly diagnosed in men and a leading cause of cancer death in men. It is estimated that there are over seventy-five million prevalent cases, twenty-seven million incident cases, and seventeen million deaths expected globally by 2030 [1–3]. Across the African continent, countries such as Uganda, South Africa, Nigeria, Ghana, and Zimbabwe, the incidence of prostate cancer is reported to increase among men between the ages of 40 and 70 years [4, 5]. In Ghana, the clinical and laboratory screening for prostate cancer is mostly done with prostate-specific antigen (PSA) and digital rectal examination (DRE). Screening for prostate cancer with PSA has largely led to a decrease in prostate cancer mortality [6] and assisted clinicians in case management of men with prostate cancer worldwide. However, some inefficiencies have been reported with the use of PSA for prostate cancer screening and diagnosis. PSA is known to be elevated in prostatitis, trauma, benign prostatic hyperplasia, and other pathological and physiological conditions of the urinary system [7]. This makes the continuous reliance on PSA for clinical decision making in prostate cancer cases problematic particularly in the Ghanaian population where it is still the main diagnostic criteria. Considering the heterogenicity of prostate carcinoma and the factors that influence the release of PSA from the prostate and the limitations that exist with the use of PSA, it is imperative that the introduction of other biomarkers with higher sensitivity and specificity is explored to minimize overdiagnosis associated with PSA screening. One of such molecular biomarkers that has shown significant prospects in improving on some of the limitations of PSA is prostate cancer gene 3 (PCA3) [8, 9]. Prostate cancer gene 3 (PCA3) is specific to the prostate gland and expressed significantly in cancerous prostate tissues compared to benign prostate tissues [10]. This may give PCA3 a cancer specificity that may be lacking with PSA. PCA3 levels in the urine is associated with the extent of metastatic activity of cancerous cells in the prostate, which suggests that PCA3 could be valuable in the diagnosis of prostatic carcinomas [11]. The clinical use of PCA3 urine assay as a diagnostic and screening tool for prostate cancer in European and US men is well documented [12]. However, scanty or no clinical data is currently available on the potential of the prostate cancer gene 3 (PCA3) urine assay as a screening and/or diagnostic tool in other population especially in the African population where the incidence of prostate cancer is on the rise. This study evaluated the potential of PCA3 as a diagnostic biomarker and compared the performance characteristics of urine PCA3 and serum PSA as diagnostic tools for prostate carcinomas in Ghanaian men.
This study was a hospital-based cross-sectional study conducted at the Urology Department of the 37 Military Hospital, Accra, Ghana, from February 2019 to August 2020. The 37 Military Hospital is a teaching hospital located in Accra, the capital city of Ghana, and has several departments including Surgical, Medical, Paediatrics, Obstetrics, Gynecology, Dental, Pathology, Pharmacy, Physiotherapy, and Urology.
The study employed a nonprobability convenience sampling technique to recruit 237 men who visited the Urology Unit of the 37 Military Hospital. Men forty years and above reporting to the Urology Department for the first time with any form of suspected prostate disorder were eligible for the study. Men who had elevated total serum PSA (≥4.0 ng/ml) and or abnormal results on DRE were recruited as study participants after giving written informed consent. Excluded participants were men below 40 years of age and men who were taking drugs for the treatment or management of any urologic disorder.
Ethical approval for the study was obtained from the Committee on Human Research, Publication, and Ethics of the School of Medicine and Dentistry (SMD), Kwame Nkrumah University of Science and Technology (KNUST) (CHRPE/AP/537/19), and the ethical review board of the 37 Military Hospital (37MH-IRB IPN/306/2019), respectively. Participants enrolled onto the study willingly after written informed consent. All experiments were conducted in accordance with the Declaration of Helsinki (1964).
A well-structured and validated questionnaire was designed and administered to each study participant to obtain sociodemographic information including age, occupation, educational status, ethnicity, and behavioural activities (smoking and alcohol consumption). Smoking was defined as smoking at least one cigarette a day and alcohol consumption was defined as drinking at least a bottle of any alcoholic liquor weekly. The medical history of each participant was taken. These included a history of other chronic illnesses such as diabetes, hypertension, kidney disease, duration of such illness, medications, family history of chronic diseases, and a history of present and past medication. This information was verified from the medical records of each participant.
About five (5) ml of blood was drawn from the antecubital vein of participants observing all aseptic protocols prior to DRE and ultrasound scan examinations. The blood sample was dispensed into a plain-gel tube and then centrifuged, and the sera obtained were stored at -80°C and used for measurement of PSA levels.
Total PSA was measured using the electrochemiluminescence method (Cobas e411 Analyzer, Roche Diagnostics, Germany). The Cobas e411 automates the immunoassay reactions using a sandwich electrochemiluminescent immunoassay standardized using the Reference Standard/WHO 96/670.
Digital rectal examination was carried out on all participants by a certified urologist. The DRE was performed to evaluate prostate shape, size, consistency, presence or absence of nodules, symmetry, edge, tenderness, or the presence of any rectal mass.
Transrectal ultrasonography (TRUS) was performed on all study participants by a certified sonographer using the Mindray DP-50 digital ultrasound scan machine (Shenzhen Mindray Bio-Medical Electronics, China) with probes of frequency 5-12 MHz to determine the volume and the configuration of the prostate.
About 5 ml of first-catch urine samples was collected immediately after DRE had been performed on each study participant. The PCA3 assay kit (Gen-Probe Inc., San Diego, CA, USA) was used to measure the mRNA concentrations in the urine samples, and a PCA3 score based on the ratio of urine PCA3 to PSA mRNA was determined.
Transrectal ultrasound-guided systematic core needle prostate biopsies were taken from study participants who had PSA ≥ 4.0 ng/ml or abnormal and or suspicious results on DRE. The prostate biopsies were stained and examined by a certified pathologist who had no prior knowledge of the clinical conditions of the participants.
Statistical analyses were performed using SPSS ver. 22.0 (IBM, Armonk, NY, USA). Descriptive statistics were performed for demographic variables and were expressed as mean ± standard deviation (SD) for continuous variables with normal distribution. In cases of asymmetrical distribution, median and interquartile (IQR) values were used. Variables such as age, PSA, and prostate volume were compared using chi (χ2) tests, t-test, and Mann–Whitney u-test. Nonparametric values were compared using the Fisher exact test. ROC analysis was used to evaluate the accuracies in predicting positive outcomes and performance of combination of the tests. Multivariate logistic regression analyses were used to evaluate the relationships between PCA3 score, Gleason score, and percentage positive biopsy cores; a p value ≤ 0.05 was considered significant.
Prostate cancer was detected in 63 (26.6%) out of the 237 participants. Benign prostatic hyperplasia and prostatitis were the major nonmalignant conditions diagnosed in a greater number of the participants. Prostatitis and benign prostatic hyperplasia were diagnosed in 59 (24.9%) and 115 (48.5%) participants, respectively. Benign prostatic hyperplasia (37.4%) and prostatitis (44.1%) were diagnosed predominantly among participants within age group 50–59 years. Prostate cancer (34.9%) was most predominant among participants within the age group 70–79 years (Table 1). A higher proportion, 96 (55.2%) participants without prostate cancer and 27 (42.9%) cancer subjects were within the PSA range 4.0-10.0 ng/ml. Majority of the participants with prostate cancer had increased PSA levels (>20 ng/ml) compared to their counterparts without prostate cancer (Table 2). As expected, there was a significant difference in total serum PSA (p = 0.025) between participants with and without prostate cancer. A greater proportion (71.8%) of participants without prostate cancer had a PCA3 score < 15.0. Participants with prostate cancer had increased PCA3 score from 15.0 to 60.0 with majority (28.6%) of them having a PCA3 score greater than 60.0 (Table 2). Digital rectal examination detected 50 (21.1%) positive and 187 (78.9%) negative participants. There was a significant difference in PCA3 scores and DRE findings between subjects with and without prostate cancer (p = 0.0001) There was no significant difference in prostate volume (p = 0.210) between subjects with and without prostate cancer. Serum PSA had a sensitivity of 92.1% and a specificity of 16.1% at a PSA cut-off value of 4.0 ng/ml with positive and negative predictive values of 28.4% and 84.8%, respectively. Using a PCA3 score cut-off value of 30.0, PCA3 score had a sensitivity of 57.1% and a specificity of 85.6% with a negative predictive value of 84.7%. DRE had a sensitivity and specificity of 66.7% and 60.3%, respectively, and a negative predictive value of 83.6% (Table 3). ROC curve analysis performed on the performance of PSA, PCA3 score, and DRE as diagnostic methods with biopsy as the reference method (Figure 1) gave an area under the curve (AUC) of 0.60, 0.83, and 0.65, respectively. Among the three prostate cancer diagnostic tests, PCA3 score performed better yielding an AUC of 0.83 (95% CI: 0.74 to 0.90) (Figure 1). Multivariate logistic regression analysis showed that high PCA3 score (OR: 1.621, p = 0.001) and high levels of serum PSA (OR: 1.110, p = 0.031) had a significant correlation with high Gleason score (Table 4). An increase in the PCA3 score was found to be associated and positively correlated (r = 1.169) with an increase in prostate cancer incidence (p < 0.0001) (Figure 2). PCA3 cut-off score of 10 had the highest sensitivity (93.7%) with a specificity of 48.9%, and a cut-off of 50 had the lowest sensitivity (39.7%) and highest specificity (100.0%). The sensitivity and specificity were 69.8% and 72.4%, respectively, at a cut-off of 20. A cut-off of 30 combined the greatest cancer sensitivity (57.1%) and specificity (85.6%) (diagnostic accuracy = 0.78 and Youden index of 0.53). PCA3 cut-off score of 40 yielded a sensitivity of 44.4% and a specificity of 96.0%. Diagnostic sensitivity decreased with increasing PCA3 score as with negative predictive values while specificity and positive predictive values generally increased with increasing PCA3 scores (Table 5). The combination of PSA and DRE (PSA+DRE) had a sensitivity of 100% and a specificity of 2.9%. PSA+PCA3 score combination had a sensitivity and specificity of 71.4% and 52.9%, respectively. The combination of DRE and PCA3 score (DRE+PCA3 score) had a sensitivity of 60.3% and a specificity of 80.5%. The three combined parameters (PSA+DRE+PCA3 score) had a sensitivity and specificity of 60.3% and 78.2%, respectively (Table 6). The combination of DRE+PCA3 score had the best diagnostic accuracy (AUC = 0.80) (Figure 3) with a sensitivity and specificity of 60.3% and 80.5%, respectively (Table 6).
In the Ghanaian population, prostate-specific antigen (PSA) and digital rectal examination (DRE) are the key screening and diagnostic protocols for making clinical decisions when prostate cancer is suspected in men. The use of PSA and DRE as the basis for recommending patients for biopsy often leads to patients being subjected to unnecessary biopsies due to the lack of specificity of PSA and DRE. Thus, we evaluated the diagnostic and prognostic potential of urinary prostate cancer gene 3 (PCA3) as a biomarker for diagnosing prostate cancer and compared the performance characteristics of serum PSA and urinary prostate cancer gene 3 (PCA3). Prostate-specific antigen (PSA), digital rectal examination (DRE), and urinary prostate cancer gene 3 (PCA3) score were used as the diagnostic tools with biopsy as the reference diagnostic tool in the assessment of prostate disorders in the participants. In this study, 26.6% presented with prostate cancer among the study participants, a finding consistent with the study by Yeboah et al. in Kumasi, Ghana. A systematic review and a meta-analysis of forty (40) studies spreading across 16 African countries also reported a pooled prevalence of prostate cancer consistent with our findings [1]. However, reports from a population-based study from the Kumasi Cancer Registry (KsCR) [13] and a population-based study among West Africans [14] reported a much lower prostate cancer incidence of 13.2% and 7.0%, respectively. A small proportion, representing 6.4% of the participants between the ages 40 and 49 years were diagnosed with prostate cancer in this current study; this is in agreement with the assertion that only about 1 out of 350 men under the age of 50 years is likely to be diagnosed with prostate cancer [15]. Furthermore, this finding is also consistent with the SEER Cancer Statistics Review in 2013 which reported a lower incidence of prostate carcinoma among men between ages 40 and 49 years [16]. Majority, 42.9%, of the men diagnosed with prostate cancer were within the PSA range 4.0 ng/ml to 10.0 ng/ml. This is in conformity with prospective cohort studies in the United States and China by Jue et al. and Tang et al., respectively, who reported that about 44% of prostate cancer patients had PSA values ranged between 4.0 and 10.0 ng/ml [17, 18]. Moreover, a higher proportion of men diagnosed with nonmalignant conditions of the prostate (BPH and prostatitis) were also within the PSA range 4.0 ng/ml to 10.0 ng/ml (Table 2). This supports the fact that many nonmalignant pathologies other than prostate cancer lead to increase in total serum PSA [19–21]. In this current study, the sensitivity and specificity of PSA for prostate cancer detection were 92.1% and 16.1%, respectively, at a PSA cut-off of 4.0 ng/ml (Table 3). This agrees with findings of several studies across the globe which report very high sensitivity with low specificity for PSA in the detection of prostate cancers [22–24]. Our study found that the PCA3 scores for men diagnosed with prostate cancer were significantly higher than those for those diagnosed with nonmalignant conditions of the prostate. To predict the accuracy for predicting the accuracy for detecting prostate cancer, we report that PCA3 had an AUC of 83.0 compared to 59.5 for serum PSA (Figure 1), with urine PCA3 showing significant association with prostate cancer detection. Our findings are consistent with findings by Auprich et al. and Ploussard et al., who observed that PCA3 is valuable in detecting prostate cancer in men scheduled for initial biopsies [25, 26]. Chronic inflammation is a known risk factor for several forms of human cancer and now regarded as an “enabling characteristic” of human cancer [27, 28]. Considering the heterogeneous nature of bacterial and nonbacterial prostate inflammation and the recurrence rates [29], chronic inflammation of the prostate may have an effect on the diagnostic accuracy of urinary PCA3 and may impact the accuracy of urinary PCA3 in predicting prostate cancer. However, in this current study, the effect of chronic inflammation on the accuracy of urinary PCA3 was not explored. In this current study, urinary PCA3 had similar negative predictive value (NPV) of 84.7% compared with PSA (84.8%) and DRE (84.0%) but significantly higher positive predictive value (PPV) of 59.0% corroborating the findings of three community-based studies on men undergoing initial and repeated prostate biopsies [30–32]. PCA3 score also correlated significantly with the probability to detect a positive biopsy (Figure 2) supporting the hypothesis that the probability of detecting prostate cancer increases with increasing PCA3 scores [8, 33]. However, a multiparametric MRI was not used in the detection of prostate cancer in this study and thus acknowledged as a limitation of the study. Multiparametric magnetic resonance imaging (mpMRI) with or without targeted biopsy has a well-established role in the detection of clinically significant PCa (csPCa) and an appealing alternative to transrectal ultrasonography (TRUS) biopsy [34–36] that was employed in this current study. Compared with systematic transrectal ultrasonography-guided biopsy, mpMRI is associated with a 57.0% improvement in the detection of clinically significant PCa and a 77% reduction in the number of cores taken per procedure [35, 37]. Alkasab et al. in evaluating the performance of PCA3 and MpMRI also reported an NPV of 40% and 83%, respectively. However, adding mpMRI to high PCA3 scores augmented the NPV to 95% [38]. In comparing mpMRI, SelectMDx, and PSA as separate tools and in various combinations, the association of mpMRI and SelectMDx was reported to have the best performance in predicting PCa and csPCa after biopsy [39]. Thus, an MRI first pathway may have influenced the performance and accuracy of urinary PCA3 in the detection of prostate cancer in the Ghanaian population. Our findings suggest that PCA3 may not be a complete replacement for PSA as the appropriate choice of test for prostate cancer especially in the Ghanaian population but may however serve as a complimentary diagnostic biomarker which would be very beneficial in the management and treatment of malignant and nonmalignant prostate conditions. There are contrasting reports and differing views regarding the optimal cut-off of PCA3 values for discriminating men having prostate cancer and those with benign prostate conditions [40]. The PCA3 assay used in this study proposes a cut-off of 35 as the optimal threshold for improved sensitivity and specificity [41]. However, several experimental studies have suggested that the optimal threshold of PCA3 may be dependent on population characteristics [42–45]. In this current study, a PCA3 score of 30.0 gave the best combination of sensitivity and specificity among Ghanaian men (Table 5). Merola et al. and Luo et al. documented a much lower cut-off of 20.0 among Italian and Chinese men [41, 46], while multicentered hospital-based studies in Europe by Marks et al. and Roobol et al. reported specificities above 90.0% at a cut-off of 100.0 [30, 47]. In this study, high levels of serum PSA and increased PCA3 score significantly correlated with Gleason score; while age and prostate volume showed no significant correlation with Gleason score (Table 4). This is in consonance with reports on men with suspected prostatic carcinoma in the Zhejiang province of China [48]. This finding also supports the opinion expressed by Groskopf et al., that PCA3 as a tissue-based overexpressed biomarker can be a promising urinary marker that can support the diagnosis and management of prostate cancer in clinical practice [49]. Additionally, this study supports the report by De Luca and colleagues which suggested that PCA3 could be a main determinant for prostatitis, high-grade prostatic intraepithelial neoplasia (HG-PIN), and prostate cancer [50]. Several studies have reported that patients with small foci of prostate cancer mostly benefit from active surveillance than from immediate treatment. It has been suggested that patients with small foci of prostate cancers should be placed on active surveillance than on chemotherapy [51, 52]. However, the major concerns of most clinicians in placing patients under active surveillance are the uncertainty in the use of PSA and or DRE as the basis for such decisions [53]. From the findings of this study, we suggest that PCA3 holds the potential valuable in deciding which patients will benefit from active surveillance or immediate treatment. Often in clinical settings, a single biomarker is not sufficient to accurately assess the clinical significance of prostate cancer at the time of diagnosis. This is due to the heterogeneity in the pathogenesis of prostate cancer [54], a fact that has informed the opinion that the combination of multiple markers and diagnostic protocols could improve the rates of detection of prostatic carcinomas [54, 55]. In assessing the performance of combined diagnostic tools in detecting prostate cancer, results from this study show that combinations of diagnostic tests improve the rates of detection of prostate cancers compared to the use of single diagnostic tests (Table 6), corroborating the findings by Descotes and Dimakakos et al. [54, 55]. Another descriptive retrospective studies also reported similar findings among Algerian men [56]. Shimizu et al. also observed a higher detection rate for prostate cancer using a combination of PSA and DRE as compared to when DRE and PSA were used as single diagnostic tools [57]. Positive predictive value is a parameter that is very essential in the assessment of cancer detection as it gives vital clinical information on the frequency of redundant biopsies [58–60]. In this study, DRE+PCA3 combination had the highest positive predictive value followed by PSA+DRE+PCA3 combinations (Table 6). These findings suggest that the addition of urinary PCA3 to the current routine diagnostic protocols in Ghana can improve the outcome of prostate cancer screening and reduce the number of patients who may have to go through unnecessary invasive biopsy procedures.
The urinary PCA3 assay showed a better diagnostic performance compared to serum PSA and DRE. Urinary PCA3 assay can facilitate the selection of high-risk men who may benefit from prostate biopsy. PCA3 urine assay could therefore be a useful marker in detecting prostate cancer in our population. | true | true | true |
||
PMC9568360 | Haiyun Liu,Changpeng Zuo,Lijuan Cao,Naiquan Yang,Tingbo Jiang | Inhibition of miR-652-3p Regulates Lipid Metabolism and Inflammatory Cytokine Secretion of Macrophages to Alleviate Atherosclerosis by Improving TP53 Expression | 07-10-2022 | Purpose The aim was to elucidate the regulatory function of miR-652-3p on lipid metabolism and inflammatory cytokine secretion of macrophages in atherosclerosis. Methods miR-652-3p level in atherosclerosis patients, ox-LDL-treated macrophages, and their controls were monitored by Q-PCR. After ox-LDL treatment and miR-652-3p mimic, si-TP53 and their controls transfection, ELISA, and Q-PCR assays were used to detect IL-1ß, IL-6, and TNF-α levels. oil red O staining was processed to verify cholesterol accumulation. CE/TC and lipid metabolism were also detected. The protein levels of ABCA1, ABCG1, PPARα, CRT1, ADRP, and ALBP were detected by western blot assay. Based on the TargetScan database, the TP53 3′UTR region had complementary bases with miR-652-3p, which was also verified by dual-luciferase reporter gene assay. Finally, the regulation of miR-652-3p and TP53 was confirmed by rescue assay in atherosclerosis. Results miR-652-3p is highly expressed in atherosclerosis, miR-652-3p inhibitor decreased IL-1β, IL-6, and TNF-α expression after ox-LDL treatment. Knockdown of miR-652-3p reduces foam formation in ox-LDL-treated macrophages. miR-652-3p inhibitor ameliorates cholesterol accumulation and lipid metabolism disorder. miR-652-3p negatively regulated TP53 in atherosclerosis. Si-TP53 rescued the effect of miR-652 inhibitor in atherosclerosis. Conclusion miR-652-3p regulates the lipid metabolism of macrophages to alleviate atherosclerosis by inhibiting TP53 expression. It might be a potential target for atherosclerosis treatment. | Inhibition of miR-652-3p Regulates Lipid Metabolism and Inflammatory Cytokine Secretion of Macrophages to Alleviate Atherosclerosis by Improving TP53 Expression
The aim was to elucidate the regulatory function of miR-652-3p on lipid metabolism and inflammatory cytokine secretion of macrophages in atherosclerosis.
miR-652-3p level in atherosclerosis patients, ox-LDL-treated macrophages, and their controls were monitored by Q-PCR. After ox-LDL treatment and miR-652-3p mimic, si-TP53 and their controls transfection, ELISA, and Q-PCR assays were used to detect IL-1ß, IL-6, and TNF-α levels. oil red O staining was processed to verify cholesterol accumulation. CE/TC and lipid metabolism were also detected. The protein levels of ABCA1, ABCG1, PPARα, CRT1, ADRP, and ALBP were detected by western blot assay. Based on the TargetScan database, the TP53 3′UTR region had complementary bases with miR-652-3p, which was also verified by dual-luciferase reporter gene assay. Finally, the regulation of miR-652-3p and TP53 was confirmed by rescue assay in atherosclerosis.
miR-652-3p is highly expressed in atherosclerosis, miR-652-3p inhibitor decreased IL-1β, IL-6, and TNF-α expression after ox-LDL treatment. Knockdown of miR-652-3p reduces foam formation in ox-LDL-treated macrophages. miR-652-3p inhibitor ameliorates cholesterol accumulation and lipid metabolism disorder. miR-652-3p negatively regulated TP53 in atherosclerosis. Si-TP53 rescued the effect of miR-652 inhibitor in atherosclerosis.
miR-652-3p regulates the lipid metabolism of macrophages to alleviate atherosclerosis by inhibiting TP53 expression. It might be a potential target for atherosclerosis treatment.
Atherosclerosis (AS) is a complex pathophysiological process caused by the formation of plaques that accumulate cholesterol on the arterial walls [1]. Common diseases caused by AS include CAD, MI, stroke, and abdominal aortic aneurysm (AAA) [2, 3]. Cardiovascular disease even leads to death [4]. With the aging of the population, the incidence of AS shows a clear trend of increasing. The study of the pathogenesis of AS has been going on for nearly a century. Dysregulation of lipid metabolism activates the biological function of immune cells [5]. The formation of atherosclerotic inflammatory response is due to lipid activation of multiple signal transduction pathways related to inflammation and apoptosis [6]. In particular, multiple transcription factors such as NF-κB, NFAT, and STAT1/3 are activated, and each transcription factor regulates multiple downstream genes related to inflammation, oxidative stress, and cell cycle regulation [7, 8]. Macrophages are a major contributor to AS pathogenesis and development. Most foam cells are formed from macrophages, and lipid metabolism disorders in macrophages are prerequisites for forming foam cells. Under normal circumstances, a dynamic balance is maintained between macrophages' lipid intake, ester hydrolysis, and outer row [9]. However, extracellular lipid levels cause intracellular metabolism abnormalities, and the expression of receptors and enzymes related to macrophages and lipid metabolism will change [10]. Ox-LDL-induced macrophage abnormal lipid metabolism and increased inflammation are important factors leading to atherosclerotic plaque progression [11]. Monocyte-derived macrophages can secrete a series of inflammatory factors after excessive lipid uptake [12, 13]. It induces a local chronic inflammatory response in coronary arteries, which in turn triggers the occurrence and progression of atherosclerotic plaques [14]. However, AS is a complex inflammatory response formed after the long-term action of multiple factors [15]. Therefore, studying the gene expression in AS will be the key to a deeper understanding of the pathophysiological mechanism of AS formation. The emergence of miRNA provides a great possibility for the realization of this goal. MiRNAs are abnormally expressed in the intimal lesions of AS and vascular occlusion [16, 17]. Ji et al. [18] detected many aberrantly expressed miRNAs in the neovascularization of the intima of mouse arteries injured by balloon catheters by microarray analysis. In addition, Liang et al. [19] demonstrated aberrant expression of some miRNAs, such as let-7 miRNA in human and murine noninjured atherosclerotic vessels. Recently, miR-652-3p was referred to be a potential target for AS. For instance, Vegter et al. [20] verified that the expression of miR-652-3p was critical in AS and cardiovascular disease. Besides, Huang et al. [21] confirmed that miR-652-3p targeted cyclin D2, and further affect the endothelial repair and AS development. However, the function of miR-652-3p in AS is complicated. This study aimed to elucidate the role of miR-652-3p on AS by studying the effect of miR-652-3p on macrophage lipid metabolism in atherosclerotic plaques. It is hoped that it will play a role in promoting the research on the pathogenesis, diagnosis, prognosis, and efficacy judgment of AS.
A total of 60 cases were from patients treated in our hospital from June 2021 to June 2022. All cases are unrelated, and patients excluded by relevant examinations include acute and chronic infection, surgery, trauma, liver and kidney disease, malignant tumor, rheumatoid disease, secondary hypertension, heart failure, heart valve disease, and alcoholism. Based on coronary angiography, patients with less than 50% stenosis of major vessels were diagnosed with coronary atherosclerotic stenosis. All subjects fasted for more than 8 h, 4 mL of cubital venous blood was collected in the early morning, and the serum was separated for testing. The patients involved affix their signatures on informed consent.
The macrophage cells RAW264.7 were obtained from the ATCC (Manassas, VA). Cells were cultured in a carbon dioxide incubator with RPMI-1640 medium with 5% FBS. The cultured RAW264.7 was treated with ox-LDL (100 ms/L) for 3 h. The miR-652-3p inhibitor and NC were designed and synthesized by Guangzhou Ribo Biological Company. Macrophages were divided into four groups, including control, ox-LDL, ox-LDL + NC inhibitor, and ox-LDL + miR-652-3p inhibitor. The transfection steps were performed according to the instructions of the Lipofectamine™ 2000 reagent (Invitrogen, Carlsbad, CA). miR-652-3p mimic, si-TP53, and their controls were also transfected by Lipofectamine™ 2000 reagent.
Total RNA was collected from samples in each group by Trizol reagent, and RNA was transcribed into cDNA. Q-PCR detection was performed by the SYBR Primix Ex Taq detection kit using cDNA as a template. The reaction system contains 2 μL of cDNA, 10 μL of 2 × SYBR Primix Ex Taq, 1 μL of upstream and downstream primers, and 20 μL of ddH2O. The reaction program was set at 95°C for 5 min, followed by 40 cycles (95°C for 30 s, 60°C for 20 s, and 72°C for 20 s). The experiment was carried out 3 times, Ct was calculated, and U6 and GAPDH were chosen as the reference. The miR-652-3p level was analyzed by the relative quantitative 2−ΔΔCt method. The primer sequence information is shown in Table 1.
Cells from each group were collected in sterile tubes and centrifuged for 20 minutes, and the supernatant was carefully collected. PBS was used to dilute the cell suspension (106 cells/mL) when detecting intracellular components. Repeated freezing and thawing were processed for releasing intracellular components. ELISA kits were purchased from Neobioscience Co., Ltd. (China). Double antibody ELISA was used to show IL-1β, IL-6, and TNF-α levels in the supernatant according to the kit instructions. The OD450 was determined for each sample. A standard curve was drawn, and the protein expression is calculated according to the standard curve.
The 15 mm sterile slides were placed in a 12-well plate in advance and the cell count was 1 × 106 mL−1 After intervention for 72 h, the samples were washed, fixed with 4% paraformaldehyde for 10 min, soaked in 60% ethanol for 1 min, and treated with oil red O staining solution. The samples were rinsed and stained with hematoxylin for 5 min. After the slides were rinsed and dried, they were sealed with glycerin gelatin and observed under an oil microscope (Olympus).
The amplex red cholesterol detection kit was used for detection. To determine total cholesterol (TC) and free cholesterol (FC), cells were obtained by chloroform/methanol extraction (2 : 1 by volume). The chloroform phase layers were collected, dried, and then stayed in the reaction buffer. The content of cholesteryl ester (CE) was calculated by measuring TC and FC content in each sample. TC and FC were detected with an automatic biochemical analyzer (Beckman). CE/TC values were applied to assess lipid metabolism.
Cells in each group were collected, NP-40 lysate was added, and the total protein in cells was extracted on ice. SDS-PAGE gels were prepared, and equal amounts of protein were taken for electrophoresis. After the protein was separated, it was transferred to the PVDF membrane; the sample was soaked in 5% skimmed milk for blocking for 1 h, and the membrane was washed. After the corresponding primary antibody (ABCA1, ABCG1, PPARα, CRT1, ADRP, and ALBP, 1 : 1000, Abcam) was added, the samples were then incubated on a vertical shaker at 4°C for 10 h. After washing, a secondary antibody diluted 1 : 3000 was applied for incubation at room temperature for 2 h. ECL luminescent solution was added to the sample, followed by the development, exposure, and image acquisition. β-Actin was normalized as a reference, and the Quantity One software was used to analyze the data.
TargetScan online database predicted that the TP53 3′UTR region had complementary bases with miR-652-3p. According to the predicted results, TP53 wild-type (TP53-Wt) and TP53 mutant (TP53-Mut) luciferase recombinant vectors were constructed, respectively. Cells were seeded into 24-well plates at 5 × 104 cells/well and cultured in a 37°C incubator. The cells were divided into NC + TP53 − Wt, miR-652-3p mimic + TP53 − Wt, NC + TP53 − Mut, and miR-652-3p mimic + TP53 − Mut groups. The transfection procedure was performed according to the instruction manual of Lipofectamine 2000 transfection reagent and incubated for 2 d for reaction. A dual-luciferase reporter gene detection kit was used. The relative luciferase activity of cells in each group was calculated by normalizing the activity of Renilla luciferase.
Statistical analysis was undertaken by SPSS 21.0 software. The data were shown as mean ± SD. The t-test and the one-way analysis were chosen for two and multiple groups, respectively. Statistically significant was with P < 0.05.
Q-PCR assay was processed to detect the miR-652-3p expression in serum samples, ox-LDL-treated macrophage cell lines, and their controls. As shown in Figure 1(a), miR-652-3p was highly expressed in AS patients. Similar results were also obtained in the cell experiment. After treatment, miR-652-3p was higher than that in the control (Figure 1(b)). Thereby, miR-652-3p was critical in the pathogenesis of AS.
Macrophages could contribute to local inflammation by producing proinflammatory cytokines in AS [22]. ELISA and Q-PCR assays were undertaken to detect immune factors, including IL-1ß, IL-6, and TNF-α. In vivo, IL-1 is mainly responsible for the acute response. Cytokines of the IL-1 family were also a part of the host to resist infection [23]. Besides, TNF-α is a cytokine involving systemic inflammation, which is mainly secreted by macrophages [24]. Conversely, IL-6 systematically acts on the liver to produce acute proteins, such as CRP, fibrinogen, and osteoclast activation inhibitors [25]. The levels of IL-1β, IL-6, and TNF-α were higher in the ox-LDL group and ox-LDL + NC inhibitor group. The levels of IL-1β, IL-6, and TNF-α were lower in the DL + miR-652-3p inhibitor group (Figure 2(a)). The results of ELISA and Q-PCR assay were consistent. The ELISA assay showed that ox-LDL accelerated IL-1ß, IL-6, and TNF-α levels, while miR-652-3p knockdown reduced these proinflammatory cytokines (Figure 2(b)). According to the evidence, the knockdown of miR-652-3p reduced inflammation in ox-LDL-treated macrophages.
The formation of foam cells often occurs early in the early stage of atherosclerosis [26]. The blood cells of patients with atherosclerosis have differentiated under the endometrium, forming macrophages, and devouring a large amount of low-density cholesterol, forming foam cells [27]. Interestingly, ox-LDL significantly increased foam cell formation in macrophages, while the knockdown of miR-652-3p reduced it (Figure 3(a)). The intracellular CE/TC value and cholesterol level of the ox-LDL treatment group were higher. miR-652-3p inhibitor decreased CE/TC value and cholesterol level (Figures 3(b) and 3(c)). ABCA1 and ABCG1 are two important proteins for cholesterol transport. ABCA1 is mainly responsible for promoting cholesterol efflux from cells to lipid-poor apoA-I, and ABCG1 is mainly responsible for promoting cholesterol efflux to mature high-density lipoprotein particles. At the same time, the two also interact to jointly promote reverse cholesterol transport. According to the western blot assay, ABCA1 and ABCG1 levels significantly decreased in the ox-LDL group, while they were upregulated after the miR-652-3p inhibitor was transfected (Figure 3(d)). Based on these results, miR-652-3p inhibitor reduces cholesterol accumulation in ox-LDL-treated macrophages.
PPARα, CRT1, ADRP, and ALBP were biomarkers of lipid metabolism disorder. We performed Q-PCR and western blot to detect their expression. Interestingly, PPARα, ADRP, and ALBP expressed higher in ox-LDL groups, while they expressed lower in the ox − LDL + miR − 652 − 3p inhibitor group. However, CRT1 showed an opposite trend (Figures 4(a)–4(d)). Thereby, miR-652-3p inhibitor reduces lipid metabolism disorder in ox-LDL-treated macrophages.
Based on the miRWalk, TargetScan, and ENCORI online databases, a total of 10 crossover target genes were obtained, including TP53, CAPZB, HOXA9, and UBE2I, HSD3B7, NPTN, KPNA1, TNRC6A, RPL28, and ISL1 (Figure 5(a)). Q-PCR was undertaken to detect these genes. Interestingly, miR-652 mimic downregulated the expression of TP53, while no significant difference was found in the other 9 genes (Figure 5(b)). Therefore, TP53 would be downstream of miR-652. The binding site was predicted by the TargetScan database (Figure 5(c)). miR-652-3p mimic could reduce the expression of TP53 in the wild group but did not affect the mutant group (Figure 5(d)). Furthermore, TP53 in the serum of AS patients and controls was also detected by Q-PCR. TP53 was expressed lower in AS serum samples (Figure 5(e)). After ox-LDL treatment, TP53 was expressed lower in macrophages (Figure 5(f)). According to the evidence, TP53 was a target of miR-652-3p.
As the above results mentioned, miR-652-3p inhibitor decreased the levels of IL-1ß, IL-6, and TNF-α, while si-TP53 rescued their expression (Figure 6(a)). Besides, miR-652-3p inhibitor reduced foam formation and ameliorated cholesterol accumulation (Figures 6(b) and 6(c)). miR-652-3p inhibitor decreased PPARα, ADRP, and ALBP expression, while accelerated CRT1 level (Figures 6(d) and 6(e)). It was worth noting that the function of miR-632-3p inhibitor was rescued when si-TP53 was cotransferred.
AS is the main cause of various diseases, such as coronary heart disease [28]. Among the entire pathogenesis, lipid metabolism disorder is the pathological basis of AS [29]. In this study, we elucidated the regulatory mechanism of miR-652-3p on AS by studying the effect of miR-652-3p on macrophage lipid metabolism in vitro. Excitingly, we verified that miR-652-3p was highly expressed in AS. Moreover, it regulated the lipid metabolism of macrophages to participate in AS development via interacting with TP53. In the previous study, miR-652-3p was confirmed to be expressed higher in various diseases, such as non-small-cell lung cancer [30], lymphoblastic leukemia [31], and cerebral ischemia [32]. Grimaldi et al. [33] researched TICAGROR for patients with acute coronary syndrome and referred that miR-652-3p expressed differently in the process of treatment. Moreover, miR-652-3p was verified to be involved in endothelial repair, atherosclerotic disease, and rehospitalizations [20, 21]. Interestingly, miR-652-3p inhibitor attenuates inflammation in ox-LDL-treated macrophages in this study. Besides, we also referred that the levels of ABCA1 and ABCG1 significantly decreased in the ox-LDL group, while they were upregulated after the miR-652-3p inhibitor was transfected. As we all know, ABCA1 is the main decisive factor in the level of blood high-density lipoprotein [34]. Therefore, the protein is particularly important in drug research and development for the treatment of arterial cell cholesterol deposition. ABCG1 is an ABC transfer protein located on the surface of the cell surface. In the plasma, HDL plays a critical role in promoting the flow of macrophage cholesterol outflow and atherosclerosis [35]. Besides, previous research shows that mice models lacking ABCG1 may cause cholesterol to accumulate a lot in macrophages and liver cells [36]. ABCG1 is highly expressed in macrophages, which is critical in the cholesterol reversal of macrophages [37]. Importantly, an imbalance of lipid homeostasis in macrophages in the early stages of arteriosclerosis leads to the accumulation of intracellular cholesterol, which in turn leads to foam cell formation [38]. The formation of cholesterol-rich foam cells is a key event in the early stages of AS [39]. In this study, the knockdown of miR-652-3p reduces foam formation in ox-LDL-treated macrophages. Furthermore, miR-652-3p inhibitor ameliorates cholesterol accumulation. In addition to removing excess cholesterol in the body, ABCA1 and ABCG1 were also important in the suppression of inflammation, thus becoming critical targets for research on inhibiting AS progress [38]. Inflammation was an important immune defense mechanism of the body. Proinflammatory cytokines played a central role in infectious or noninfectious inflammatory diseases. Proinflammatory cytokines were mainly obtained by macrophages and involved in the upregulation of inflammatory responses. IL-1β, IL-6, and TNF-α were typical proinflammatory cytokines. These cytokines acted to contain and resolve inflammatory foci by activating local and systemic inflammatory responses. Based on the importance of these cytokines in the inflammatory response, we detected their expression and confirmed that miR-652-3p inhibitor attenuated inflammation via regulating inflammatory cytokines. It was worth noting that miR-652-3p targeted TP53 expression in AS based on biological information analysis and molecular experiments in this study. The relationship between miR-652-3p and TP53 has not been reported so far. TP53 expression products were found in macrophages, which participated in the adjustment of macrophages during the process of AS [39]. Moreover, Guevara et al. [40] researched the role of TP53 in the occurrence and development of AS in vivo, confirming that lacking TP53 was related to the size of the AS lesion, macrophages, and lipoprotein. In this study, TP53 was confirmed to be the target of miR-652-3p. Interestingly, si-TP53 rescued the effect of the miR-652-3p inhibitor on lipid metabolism and inflammation. Thereby, the rescue assay further confirmed the role of miR-652-3p inhibitor and si-TP53 in the inflammatory and lipid metabolism of macrophages in AS. In conclusion, miR-652-3p regulated lipid metabolism and inflammatory cytokine secretion of macrophages to alleviate AS by inhibiting TP53 expression. This study hopes to provide favorable clues to solve AS and related cardiovascular diseases through the miR-652-3p inhibitor's function in lipid metabolism and inflammatory cytokine secretion. | true | true | true |
||
PMC9568361 | Huiqing Zhang,Xiaohua Shen,Shuping Xiong,Lixiang Peng,Wenli Mai,Longxiang Xin | HEIH Promotes Malignant Progression of Gastric Cancer by Regulating STAT3-Mediated Autophagy and Glycolysis | 27-07-2022 | To study the clinical value of HEIH hyperexpression in gastric cancer and the molecular mechanism of promoting malignant proliferation of gastric cancer cells, qRT-PCR was used to detect the expression of HEIH in gastric cancer and nontumor gastric tissues. HEIH interference sequence was constructed to downregulate HEIH expression in MGC-803 and BGC-823 cell lines. CCK8, clonogenesis, and Transwell assay were used to detect the effects of HEIH on proliferation and invasion of tumor cells. The protein levels of STAT3, p-STAT3, P62, and LC3 were detected by Western blotting. The results showed that HEIH was highly expressed in gastric cancer (P < 0.01). Interference of HEIH expression in MGC-803 and BGC-823 cells reduced the proliferation and invasion of gastric cancer cells, and the results were statistically significant (P < 0.05). HEIH acts as a miRNA sponge for miR-4500. HEIH promotes gastric cancer development by inhibiting miR-4500. STAT3 is a downstream target of miR-4500. HEIH inhibits autophagy and promotes glycolysis. In conclusion, HEIH is highly expressed in gastric cancers. HEIH promotes malignant proliferation and development of gastric cancer cells. HEIH may be a new candidate site for pathological diagnosis and molecular drug therapy for future clinical treatment of gastric cancer. | HEIH Promotes Malignant Progression of Gastric Cancer by Regulating STAT3-Mediated Autophagy and Glycolysis
To study the clinical value of HEIH hyperexpression in gastric cancer and the molecular mechanism of promoting malignant proliferation of gastric cancer cells, qRT-PCR was used to detect the expression of HEIH in gastric cancer and nontumor gastric tissues. HEIH interference sequence was constructed to downregulate HEIH expression in MGC-803 and BGC-823 cell lines. CCK8, clonogenesis, and Transwell assay were used to detect the effects of HEIH on proliferation and invasion of tumor cells. The protein levels of STAT3, p-STAT3, P62, and LC3 were detected by Western blotting. The results showed that HEIH was highly expressed in gastric cancer (P < 0.01). Interference of HEIH expression in MGC-803 and BGC-823 cells reduced the proliferation and invasion of gastric cancer cells, and the results were statistically significant (P < 0.05). HEIH acts as a miRNA sponge for miR-4500. HEIH promotes gastric cancer development by inhibiting miR-4500. STAT3 is a downstream target of miR-4500. HEIH inhibits autophagy and promotes glycolysis. In conclusion, HEIH is highly expressed in gastric cancers. HEIH promotes malignant proliferation and development of gastric cancer cells. HEIH may be a new candidate site for pathological diagnosis and molecular drug therapy for future clinical treatment of gastric cancer.
Gastric cancer is one of the most common malignancies in the world and is the third leading cause of cancer-related deaths, with more than 1 million new cases per year. The 5-year overall survival rate of patients with localized, early gastric cancer is greater than 60%, and the 5-year overall survival rate of patients with gastric cancer with local and distant metastasis has substantially decreased to 30% and 5%, respectively [1–3]. The early clinical symptoms of gastric cancer are concealed and atypical, so more than 60% of patients are diagnosed with local or distant metastasis. Surgical resection is the best treatment option for patients with early gastric cancer; chemotherapy is the most important treatment for patients who cannot undergo surgical resection or have advanced metastasis [4]. However, patients with gastric cancer respond very poorly to chemotherapy due to intrinsic or acquired resistance, which becomes the most common cause of treatment failure. Low early diagnosis rate and chemoresistance are the main reasons for poor prognosis of gastric cancer [5]. Long noncoding RNA (lncRNA) is a class of noncoding RNA (ncRNA) with a length of more than 200 nucleotides that was discovered in recent years in the study of human genome [1]. lncRNAs do not have the ability to encode proteins but regulate gene expression through various pathways in the form of RNA, including tumor genes [2]. Relevant lncRNAs have been found to play a key regulatory role in the occurrence and development of malignant gastric cancer [4]. It has been reported that HEIH enhances paclitaxel tolerance in endometrial cancer cells by activating the MAPK signaling pathway [5]. In addition, lncRNA-HEIH in serum and exosomes may serve as a potential biomarker for HCV-associated hepatocellular carcinoma [6]. In colorectal cancer, HEIH promotes the development of colorectal cancer by regulating the miR-939/Bcl-XL axis [7]. However, the relationship between HEIH and EMT has not been reported. Studies have confirmed that miR-4500 plays a similar role as “tumor suppressor gene” in the progression of thyroid cancer, non-small-cell lung cancer, colorectal cancer, and other tumors [8–10]. However, the expression pattern of miR-4500 in gastric cancer and whether it plays the function of “tumor suppressor gene” in gastric cancer remain unclear. Signal transduction and transcription activation factor 3 (STAT3) is an important intracellular signal transduction factor, which is involved in various physiological and pathological processes of cells [11]. Autophagy is a mode of cell self-degradation, which can provide energy for cell metabolism and maintain the homeostasis of the intracellular environment [12, 13]. STAT3 and autophagy play important roles in tumor genesis and development [14]. Previous studies have shown that STAT3 and autophagy levels are altered and abnormal in a variety of tumors, and STAT3 and autophagy may play a synergistic or antagonistic role in different stages of tumor pathological changes, thus promoting or inhibiting the occurrence and development of tumors [15, 16]. The objective of this study was to determine the abnormal expression of HEIH in human gastric cancer and to investigate the relationship between HEIH differential expression and malignant degree of gastric cancer. To further study the function of malignant proliferation of HEIH tumor cells, we hope to provide a new theoretical basis for exploring the molecular mechanism of malignant proliferation of brain gastric cancer. At the same time, provide new research evidence for molecular therapy of clinical gastric cancer.
Thirty clinical gastric cancer samples and paracancer control samples were collected between September 2019 and December 2020. The specimens were obtained from patients who had been diagnosed with gastric cancer and underwent surgery. No local or systemic therapy was performed before surgery. All patients signed informed consent. Specimens were assessed by the pathology department. All specimens were frozen in liquid nitrogen immediately after excision. Afterward, they were stored in an -80°C refrigerator for later use. This project is approved by the Ethics Research Committee of Cancer Hospital of Nanchang University.
MGC-803 is a human gastric cancer cell. The morphological characteristics are epithelial-like and adherent growth. MGC-803 was established in a 53-year-old man with primary poorly differentiated myxoid adenocarcinoma of the stomach. BGC-823 is a human gastric adenocarcinoma cell (poorly differentiated). BGC-823 cells were derived from a 62-year-old patient with gastric cancer (undifferentiated adenocarcinoma). BGC-823 cells could express CEA. In this study, MGC-803 and BGC-823 cells were cultured in RPMI medium containing 10%FBS. Double antibody (100 U/mL penicillin and 100 mg/mL streptomycin) was added to the medium. Culture conditions: constant temperature incubator, 37°C, 5% CO2. According to the cell growth condition, the medium was changed from 1 to 3 days, and the culture was carried out when the culture dish was about 80% to 90% full of cell melt growth.
The interference sequences of HEIH (si-HEIH 1#, si-HEIH 2#) and out-of-order control (si-NC) were purchased from GenePharma. About 2 × 105 cells/wells of MGC-803 and BGC-823 cells were inoculated into 6-well culture plates. 100 pmol siRNA and si-NC or 4 μg pcDNA and pcDNA-STAT3 were diluted with 200 μL Opti-MEM and incubated at room temperature for 5 min. The incubated liposomes were mixed with plasmids or siRNA diluents, gently blown and mixed, and incubated at room temperature for 20 min. Then, drop evenly into the 6-well culture plate with 1.6 Opti-MEM in advance. Culture at 37°C for 6 h in the incubator with 5% CO2. Change the medium completely. Continue to cultivate. After transfection 48 h, cells were collected and total RNA was extracted for qRT-PCR analysis and other subsequent experiments.
Cells were collected and total cell RNA, cytoplasmic RNA, and nuclear RNA were extracted in strict accordance with the operation instructions of the kit. The expression of HEIH was detected by Takara reverse transcription reagent and qRT-PCR. β-Actin was used as cytoplasmic reference and U6 as nuclear reference.
The cells were cultured in 6-well plates, and their density was about 60%. The cells were collected 24 h after transfection, and the artificial cell count was 1000 cells/well, and the cells were inoculated into 96-well culture plates. Each sample is set with 6 multiple holes. After about 80% of the inoculated cells adhered to the wall, the cells continued to be cultured for 12 h. Add 20 μL CCK8 reaction solution directly. It was incubated at 37°C for 2 h away from light. The absorbance at 490 nm was measured by a microplate reader. The results were repeated three times.
The cells were inoculated on 6-well plates and cultured to an appropriate density of 60% before transfection. Cell counts were collected 24 h after transfection. The cells were seeded into 6-well plates at an appropriate density (about 500) and placed in a 5% CO2 incubator at 37°C for further culture for 14 days. Add medium 1 time if necessary. Culture was terminated when clones were visible in the petri dish. PBS was gently cleaned, and 1 mL methanol was added for 60 min. 1 mL 0.1% crystal violet was added and dyed for 60 min. No enzyme washing to remove residual dyeing solution, air drying. The number of clones with more than 10 cells was counted under a microscope (low magnification). Clone formation rate was calculated.
The density of transfected cells was adjusted to 3 × 106 cells/mL in serum-free medium, and 2 × 104 cells were injected into the upper compartment. 500 μL RPMI 1640 medium containing 10% FBS was injected into the lower chamber. It was placed in an incubator and cultured for 12 h. A cotton swab gently wipes away the superior ventricle cells that have not penetrated the submembrane, methanol fixation for 5 min, 0.1% crystal violet staining for 5 min, water cleaning. Photographs were taken under an inverted microscope. The experiment was repeated three times.
Collect each group of cells, and add TRIzol reagent to extract total RNA from cells. According to the reverse transcription kit, reverse transcribe to synthesize cDNA. The primers of the target gene and the internal reference β-actin gene were designed and synthesized by BGI. miR-4500 upstream primer: 5′-GGGGTGAGGTAGTAG-3′, downstream primer: 5′-CAGTGCGTGTCGTGGAGT-3′. U6 upstream primer: 5′-CTCGCTTCGGCAGCACA-3′, downstream primer: 5′-AACGCTTCACGAATTTGCGT-3′. STAT3 upstream primer: 5′-GGACATCAGCGGTAAGACCC-3′, downstream primer: 5′-CCTGGGTCAGCTTCAGGATG-3′. Reaction conditions: 95°C predenaturation for 30 s; 95°C denaturation for 5 s, 60°C annealing for 30 s, 72°C extension for 30 s, 40 cycles; 72°C final extension for 7 minutes. The 2-ΔΔCt method was used to calculate the mRNA expression level of each gene.
Bioinformatics software TargetScan predicted that the target gene of miR-4500 might be STAT3. Bioinformatics software Starbase predicted that the target gene of HEIH might be miR-4500.
The transfected cells were randomly divided into 4 groups: wild-type+miR-4500 group (transformed into wild-type plasmid+miR-4500), wild-type+miR-NC group (transformed into wild-type plasmid+miR-NC), mutant+miR-4500 group (transformed into mutant plasmid+miR-4500), and mutant+miR-NC group (transformed into mutant plasmid+miR-NC). Use dual luciferase reporter test kit for detection. The firefly luciferase activity and Renilla luciferase activity of the 4 groups of cells were analyzed. The experiment was repeated three times.
Cells in good growth state and in logarithmic growth phase were taken and placed into a 10 mm × 10 mm glass slide in a 24-well culture plate. The cells were inoculated into 24 empty plates at a density of 5 × 103 cells per well. After 24 h, the supernatant was removed and 4% paraformaldehyde was fixed. 0.1% Triton X-100 permeable, prehybridization solution prehybridized at 37°C. The probes were hybridized at 42°C for 16-20 h. Then, rinse with 2× SSC and drop DAPI into the section hybridization area for 10 min. PBS was cleaned and observed under fluorescence microscope.
Cells were taken from each group, and total protein was extracted by adding protein lysate. Protein concentration was quantified by BCA method. 20 μg protein was extracted and subjected to 8% SDS-PAGE. The isolated protein gel was transferred to the NC membrane. 5% skim milk powder was sealed for 120 min, then primary antibody (diluted at 1 : 1000) was added at 4°C to seal overnight, and TBST was washed for 5 min × 3 times. Diluted secondary antibody was added and incubated at room temperature for 1 h and then washed with TBST, exposure, development, photography. The gray value of strip was determined by ImageJ image analysis software. The ratio of gray value of each target protein to GAPDH gray value was used as the relative expression of target protein.
SPSS 20.0 was used for statistical analysis. Results were shown by mean ± standard deviation of 3 experiments. The difference between the two groups was tested by two-tailed Student's ST. One-way ANOVA was used for comparison between multiple groups. The statistical results were represented by P value, ∗P < 0.05, which was statistically significant, and ∗∗P < 0.01, which was statistically significant.
qRT-PCR was used to analyze the expression of HEIH in 30 cases of gastric cancer and nontumor brain tissues. The results showed that HEIH expression was upregulated in gastric cancer tissues compared with adjacent tissues (Figure 1(a)), and the difference was statistically significant. The expression of HEIH in the cytoplasm and nucleus was analyzed. The results showed that HEIH was mainly distributed in the cytoplasm (Figure 1(b)).
siRNA HEIH transfection in MGC-803 and BGC-823 cells was detected by qRT-PCR. The results showed that HEIH expression was decreased in siRNA transfected cells compared with the control group (Figure 2(a)). In MGC-803 and BGC-823 cell lines, the absorbance value of CCK8 at 490 nm was detected after transfection with si-HEIH. Compared with the control group (si-NC), the proliferation activity of MGC-803 and BGC-823 cells was significantly decreased after transfection with si-HEIH (Figure 2(b)). Cloning experiments suggested that knockdown HEIH significantly inhibited the clonogenesis of MGC-803 and BGC-823 cells (Figure 2(c)). Transwell experiment results showed that the number of invasive cells in the interference group was significantly reduced (Figure 2(d)), and the difference between si-HEIH and si-NC was statistically significant. The results of EMT-related markers showed that the expression of Twist1, Snail, Slug, and N-cadherin mRNAs in MGC-803 and BGC-823 cells was decreased by HEIH lowering (Figures 2(e)–2(h)), while the expression of E-cadherin mRNA was upregulated (Figure 2(i)).
Starbase, a bioinformatics software, predicted that miR-4500 might be the target gene of HEIH (Figure 3(a)). As can be seen from Figure 3(b), the relative luciferase activity of the wild-type+miR-4500 group was significantly lower than that of the wild-type+miR-NC group. These results indicate that miR-4500 can effectively inhibit the activity of wild-type plasmid luciferase. The relative luciferase activity of the mutant+miR-4500 group was significantly higher than that of the wild-type+miR-4500 group, indicating that miR-4500 could not inhibit the mutant plasmid luciferase activity. Fluorescence in situ hybridization was used to detect the colocalization of HEIH and miR-4500 in BGC-823 cell line. The results showed that HEIH and miR-4500 were colocated in the cytoplasm (Figure 3(c)). qRT-PCR was used to analyze the expression of miR-4500 in 30 gastric cancer tissues and nontumor adjacent tissues. The results showed that compared with the adjacent tissues, the expression of miR-4500 in gastric cancer tissues was decreased (Figure 3(d)), with statistically significant difference.
Real-time PCR results showed that the expression level of miR-4500 in the si-HEIH group was significantly higher than that in the si-NC group, and the difference between the two groups was statistically significant (Figure 4(a)). CCK8 results showed that compared with the si-HEIH+miR-NC group, the cell activity of the si-HEIH+miR-4500 inhibitor cotransfection group was enhanced (Figure 4(b)). Compared with the si-HEIH+miR-NC group, cell clonal formation and invasion ability were enhanced in the si-HEIH+miR-4500 inhibitor cotransfection group (Figures 4(c) and 4(d)). After HEIH silencing, the expression of Twist1 and N-cadherin decreased. After transfection with si-HEIH+miR-4500 inhibitor, the expression levels of Twist1 and N-cadherin were upregulated (Figures 4(e) and 4(f)). The expression of E-cadherin increased after HEIH silencing. After transfection with si-HEIH+miR-4500 inhibitor, the expression of E-cadherin was decreased (Figure 4(g)).
Bioinformatics software TargetScan predicted that the target gene of miR-4500 might be STAT3. The miR-4500 seed region has complementary pairing sequences with the 3′ untranslated region of STAT3 gene (Figure 5(a)). The results of survival analysis showed that patients with higher HEIH expression had worse prognosis. It was suggested that HEIH upregulation was positively correlated with worse prognosis in gastric cancer (Figure 5(b)). STAT3 expression was upregulated in gastric cancer tissues compared with paracancer tissues (Figure 5(c)). The correlation results of STAT3 and miR-4500 expression in gastric cancer tissues and paired normal tissues showed negative coexpression of STAT3 and miR-4500 (Figure 5(d)). As shown in Figure 5(e), the relative luciferase activity of the wild-type+miR-4500 group was significantly lower than that of the wild-type+miR-NC group, indicating that miR-4500 could effectively inhibit the wild-type plasmid luciferase activity. The relative luciferase activity of the mutant+miR-4500 group was significantly higher than that of the wild-type+miR-4500 group. These results indicated that miR-4500 could not inhibit the mutant plasmid luciferase activity. Therefore, only miR-4500 can effectively bind to the 3′ UTR region of the wild-type plasmid. The correlation results of STAT3 and HEIH expression in gastric cancer tissue showed that STAT3 and HEIH showed coexpression positive correlation (Figure 5(f)). Real-time PCR results showed that STAT3 expression level in the si-HEIH group was significantly lower than that in the si-NC group, and the difference between the two groups was statistically significant (Figure 5(g)).
Real-time PCR results showed that the expression level of STAT3 in the pcDNA3.1 STAT3 group was significantly higher than that in the NC group, with statistically significant difference between the two groups (Figure 6(a)). CCK8 results showed that compared with the si-HEIH+Vector-NC group, the cell activity of the si-HEIH+STAT3 cotransfected group was enhanced (Figure 6(b)). Compared with the si-HEIH+Vector-NC group, si-HEIH+STAT3 cotransfected cells had higher clonal formation and invasion ability (Figures 6(c) and 6(d)). After HEIH silencing, the expression of Twist1 and N-cadherin decreased. Meanwhile, after si-HEIH+STAT3 transfection, the expression levels of Twist1 and N-cadherin were upregulated (Figures 6(e) and 6(f)). The expression of E-cadherin increased after HEIH silencing. Meanwhile, after si-HEIH+STAT3 transfection, the expression of E-cadherin decreased (Figure 6(g)).
To further investigate the role of HEIH, we detected the protein levels of STAT3, p-STAT3, P62, and LC3 in the treated BGC-823 cells by Western blotting. The results showed that the protein levels of STAT3, p-STAT3, P62, and LC3 decreased after HEIH silencing. After si-HEIH+STAT3 transfection, the protein levels of STAT3, p-STAT3, P62, and LC3 were upregulated (Figure 7(a)). These results suggest that HIEH silencing can inhibit autophagy of gastric cancer cells. Subsequently, lactate production, glucose uptake, and ATP production in gastric cancer cells were measured after different treatments. The results showed that lactate production, glucose uptake, and ATP production decreased after HEIH addiction. These results showed that after HEIH silencing, glycolysis of gastric cancer cells was inhibited (Figures 7(b)–7(d)).
The occurrence and development of gastric cancer is a malignant tumor with multigene mutation, multisignal pathway, and multifactor and multistep participation. Under the background of in-depth research on molecular targeted therapy of tumors and increasing clinical application, a large number of studies have shown that lncRNA is related to the occurrence and development of gastric cancer. Gastric juice lncRNAs have high specificity and can be used as biomarkers for the diagnosis and prognosis of gastric cancer [18]. lncRNAs in a single gastric juice are insufficiently sensitive as biomarkers although they are highly specific. More combined studies can be carried out, such as the combination of multiple gastric juice lncRNAs, the combination of gastric juice lncRNAs with plasma lncRNAs, and the combination of gastric juice lncRNAs with serum tumor markers to improve sensitivity [19]. But the specific sources and molecular mechanisms of lncRNAs in gastric juice need to be further explored. lncRNA plays a dual role in malignant gastric cancer [17]. Li et al. [18] conducted cell transfection on the pathological tissues of 15 patients with malignant gastric cancer. Then, real-time quantitative reverse transcription polymerase chain reaction (RT-QPCR) analysis showed that NEAT1 expression was upregulated in malignant brain glial tissue. NEAT1 knockout can reduce the proliferation, invasion, and metastasis of tumor cells [19–21]. RNA-binding luciferase assay by immunoprecipitation confirmed that miRNA-449b-5p was bound to NEAT1 and acts as a “molecular sponge” to it. That is, the expression of miRNA-449b-5p can be negatively regulated. Shi et al. conducted statistical analysis on the expression of intergene-long noncoding gene H19 in 158 cases of gastric cancer combined with gastric cancer grade. The expression of H19 in high-grade gastric cancers was significantly higher than that in low-grade gastric cancers. Meanwhile, the expression level of H19 increased with the grade of gastric cancer. Inhibition of H19 expression by siRNA significantly reduced the invasion and migration ability of both malignant gastric cancer cells [22]. H19 may play a role in promoting the invasion and migration of gastric cancer cells as a precursor of miR-675. Ma et al. [23] analyzed the relationship between MALAT1 expression level and clinicopathological features of gastric cancer. The results showed that MALAT1 expression was increased in tumor tissues compared with adjacent normal brain tissues. The expression level was positively correlated with WHO grade and volume of gastric cancer. Overexpression of MALAT1 accelerates the growth of gastric cancer cells and enhances the ability of tumor cells to invade and metastasize. The mechanism may be that MALAT1 regulates the expression of genes related to cell metastasis and cell cycle [24, 25]. In order to investigate the clinical value of HEIH, qRT-PCR was used to confirm the high expression of HEIH in gastric cancer. In order to study the effect of HEIH on the biological function of gastric cancer cells, we downregulated and upregulated HEIH expression in gastric cancer cells. Subsequently, CCK8, clonogenesis, and Transwell assay were performed to determine the effects of HEIH on proliferation, clonogenesis, and invasion of gastric cancer cells. The results suggest that HEIH can regulate the malignant proliferation of gastric cancer cells. Mechanism studies have shown that upregulated HEIH binds miR-4500 to affect autophagy and glycolysis through the STAT3 signaling pathway and promote the proliferation and migration of gastric cancer cells. Autophagy and glucose metabolism reprogramming play an important role in tumor proliferation, drug resistance, invasion, and metastasis. Autophagy has been found to regulate glucose metabolism and affect the malignant progression of tumors. Bioinformatics software TargetScan predicted that the target gene of miR-4500 might be STAT3. The miR-4500 seed region has complementary pairing sequences with the 3′ untranslated region of STAT3 gene. Overexpression of STAT3 can be detected in a variety of tumor tissues and cells. Previous studies have found that autophagy level changes in a variety of tumors at different stages of tumor genesis and development. Autophagy can promote and inhibit it in different ways [26, 27]. The interaction between STAT3 and autophagy is a complex process [12]. Our results suggest that HEIH can inhibit autophagy by regulating STAT3 in gastric cancer cell lines. One of the characteristics of glucose metabolism in tumor cells is the use of glycolysis even when oxygen content is normal, namely, the Warburg effect. The internal mechanism of Warburg effect is very complex, which may be related to oncogene activation, tumor suppressor gene inactivation, abnormal expression of glucose metabolism enzymes, and changes in tumor microenvironment. The specific mechanism needs to be further studied. In this study, HEIH can also promote glycolysis of gastric cancer cells by regulating STAT3.
This study confirmed that HEIH is upregulated in gastric cancer tissues. HEIH can promote malignant proliferation of gastric cancer cells and promote tumor proliferation and invasion in vitro. It is suggested that abnormally high expression of HEIH may be one of the factors of poor prognosis in patients with gastric cancer. HEIH may serve as a risk factor for shorter survival and higher risk of metastasis in patients in the future. This study will enrich the molecular mechanism of lncRNAs regulating the occurrence and development of gastric cancer and provide a new experimental basis for clinical diagnosis and treatment of gastric cancer. This study may also improve new candidate sites for pathological diagnosis and molecular drug therapy for future clinical treatment of gastric cancer. | true | true | true |
||
PMC9568463 | 36241744 | Joan Sala-Gaston,Leonardo Pedrazza,Juanma Ramirez,Arturo Martinez-Martinez,Lettie E. Rawlins,Emma L. Baple,Andrew H. Crosby,Ugo Mayor,Francesc Ventura,Jose Luis Rosa | HERC2 deficiency activates C-RAF/MKK3/p38 signalling pathway altering the cellular response to oxidative stress | 14-10-2022 | Neurodevelopmental disorder,Angelman,Ubiquitin,MAPK,Cell stress | HERC2 gene encodes an E3 ubiquitin ligase involved in several cellular processes by regulating the ubiquitylation of different protein substrates. Biallelic pathogenic sequence variants in the HERC2 gene are associated with HERC2 Angelman-like syndrome. In pathogenic HERC2 variants, complete absence or marked reduction in HERC2 protein levels are observed. The most common pathological variant, c.1781C > T (p.Pro594Leu), encodes an unstable HERC2 protein. A better understanding of how pathologic HERC2 variants affect intracellular signalling may aid definition of potential new therapies for these disorders. For this purpose, we studied patient-derived cells with the HERC2 Pro594Leu variant. We observed alteration of mitogen-activated protein kinase signalling pathways, reflected by increased levels of C-RAF protein and p38 phosphorylation. HERC2 knockdown experiments reproduced the same effects in other human and mouse cells. Moreover, we demonstrated that HERC2 and RAF proteins form molecular complexes, pull-down and proteomic experiments showed that HERC2 regulates C-RAF ubiquitylation and we found out that the p38 activation due to HERC2 depletion occurs in a RAF/MKK3-dependent manner. The displayed cellular response was that patient-derived and other human cells with HERC2 deficiency showed higher resistance to oxidative stress with an increase in the master regulator of the antioxidant response NRF2 and its target genes. This resistance was independent of p53 and abolished by RAF or p38 inhibitors. Altogether, these findings identify the activation of C-RAF/MKK3/p38 signalling pathway in HERC2 Angelman-like syndrome and highlight the inhibition of RAF activity as a potential therapeutic option for individuals affected with these rare diseases. Supplementary Information The online version contains supplementary material available at 10.1007/s00018-022-04586-7. | HERC2 deficiency activates C-RAF/MKK3/p38 signalling pathway altering the cellular response to oxidative stress
HERC2 gene encodes an E3 ubiquitin ligase involved in several cellular processes by regulating the ubiquitylation of different protein substrates. Biallelic pathogenic sequence variants in the HERC2 gene are associated with HERC2 Angelman-like syndrome. In pathogenic HERC2 variants, complete absence or marked reduction in HERC2 protein levels are observed. The most common pathological variant, c.1781C > T (p.Pro594Leu), encodes an unstable HERC2 protein. A better understanding of how pathologic HERC2 variants affect intracellular signalling may aid definition of potential new therapies for these disorders. For this purpose, we studied patient-derived cells with the HERC2 Pro594Leu variant. We observed alteration of mitogen-activated protein kinase signalling pathways, reflected by increased levels of C-RAF protein and p38 phosphorylation. HERC2 knockdown experiments reproduced the same effects in other human and mouse cells. Moreover, we demonstrated that HERC2 and RAF proteins form molecular complexes, pull-down and proteomic experiments showed that HERC2 regulates C-RAF ubiquitylation and we found out that the p38 activation due to HERC2 depletion occurs in a RAF/MKK3-dependent manner. The displayed cellular response was that patient-derived and other human cells with HERC2 deficiency showed higher resistance to oxidative stress with an increase in the master regulator of the antioxidant response NRF2 and its target genes. This resistance was independent of p53 and abolished by RAF or p38 inhibitors. Altogether, these findings identify the activation of C-RAF/MKK3/p38 signalling pathway in HERC2 Angelman-like syndrome and highlight the inhibition of RAF activity as a potential therapeutic option for individuals affected with these rare diseases.
The online version contains supplementary material available at 10.1007/s00018-022-04586-7.
Hereditary neurodevelopmental disorders arise from alterations in central nervous system development and manifest perinatally or during infancy and childhood. Despite showing wide genetic and clinical heterogeneity, most share some common phenotypic features, such as developmental delay, impaired motor function and intellectual disability. The identification of genes responsible for these disorders has enabled genetic diagnosis, accurate genetic counselling, and better management [1]. The HECT and RCC1-like domain 2 (HERC2) gene encodes an unusually large protein with 4834 amino acid residues. The HERC2 protein is an E3 ubiquitin ligase that functions in ubiquitylation by accepting ubiquitin from ubiquitin-conjugating enzymes (E2) and transferring it to a target protein [2]. Ubiquitylation affects proteins in many ways, variously marking them for proteasome degradation or, affecting their activity, localisation or interactions with other proteins. Therefore, ubiquitin ligases are key regulators of many cellular processes, with their dysregulation being common in numerous cancers and neurodegenerative diseases [3]. For example, HERC2 mutations are associated with breast, skin (melanoma), gastric, colorectal, and haematological (leukaemia) cancers [4]. The underlying molecular mechanism could be that HERC2 regulates BRCA1, XPA, USP20 or RPA2 protein ubiquitylation, involved in regulating DNA repair and genomic stability [5–9]. HERC2 also regulates p53 transcriptional program by promoting p53 tetramerisation and subsequent activation, independent of its ubiquitin ligase activity [10–12]. Besides, HERC2 is essential during embryonic development and plays an important role in regulating motor coordination [13]. Moreover, it is highly expressed in the nervous system and has been linked with hereditary neurodegenerative disorders [14]. Biallelic HERC2 variants associated with HERC2 Angelman-like syndrome include missense and frameshift mutations with a premature stop codon that result in a loss of function. These cases are associated with a complete loss or markedly reduced levels of HERC2 protein [15–19]. The condition was first described in Amish/Mennonite communities, associated with homozygosity for a HERC2 (c.1781C > T, p.Pro594Leu) founder gene variant at increased frequency in the population (autosomal recessive mental retardation type 38; OMIM # 615516) [15, 16]. Proteomic studies of peripheral blood-derived lymphoblasts from individuals with this condition suggest derangements of multiple cellular pathways probably involving disparate pathogenic mechanisms [20]. Despite these efforts, the molecular mechanisms underlying HERC2-related disorders remain elusive, impeding efforts to find potential treatments for these rare diseases. Further investigation of their molecular basis could reveal not only the underlying pathology but also potential therapeutic targets. In this study, we analysed intracellular signalling pathways in skin fibroblasts from individuals with the pathological variant HERC2 Pro594Leu (HERC2 P594L). They displayed altered mitogen-activated protein kinase (MAPK) signalling that affected the oxidative stress response, with increases in C-RAF protein levels and MAPK p38 activation. These effects were reproduced in other human and mouse cells with HERC2 protein knockdown. Furthermore, we showed that HERC2 regulates C-RAF ubiquitylation and that HERC2 deficiency triggers MKK3/p38 pathway activation in a RAF-dependent manner. In line with this, cells with the HERC2 P594L variant had increased resistance to H2O2-induced oxidative stress, dependent on the activities of RAF and p38. Finally, we discuss both the implications of these findings for neurodevelopmental disorders caused by HERC2 variants and the potential therapeutic use of RAF inhibitors.
Samples of human skin fibroblasts were obtained with approved informed consent as previously described elsewhere [16]. U2OS, HEK 293T, H1299, RAW 264.7, mouse embryonic fibroblasts (MEFs) and human skin fibroblasts were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U/mL penicillin, and 0.1 mg/mL streptomycin sulphate. Mouse primary osteoblasts were cultured in Minimum Essential Medium α with 10% FBS, 2 mM l-glutamine, 1 mM pyruvate, 100 U/ml penicillin, and 0.1 mg/ml streptomycin with 50 μg/ml ascorbic acid and 4 mM β-glycerophosphate. All cells were maintained in a humidified incubator at 37 °C and 5% CO2 atmosphere.
Cells were treated with one of three inhibitors, as indicated: 1 µM LY3009120 (Selleckchem), 1 µM Sorafenib (Santa Cruz Biotechnology) or 10 µM SB203580 (Selleckchem). Different cellular stress types were induced using different stressors: oxidative stress by 500 µM or 50 µM hydrogen peroxide (H2O2) (Panreac), depending on the experiment; saline stress by 100 mM NaCl.
Plasmid transfection was performed using the Lipofectamine LTX method (15338; Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions. Myc-tagged fragments of HERC2 (F1, F2, F3, F4, F5 and F5CT) were kindly provided by Dr. Ohta [21]. Green fluorescent protein (GFP) and C-RAF fusion constructs (CR1, CR2, CR3 and full-length) were generated, sub-cloned and tested elsewhere [22]. Plasmids expressing HERC2 full-length protein pcDNA5 FRT/TO SF-HERC2 (ShB-R) (Addgene plasmid # 55613; http://n2t.net/addgene:55613; RRID:Addgene_55613) and pcDNA5 FRT/TO SF-HERC2 C4762S (ShB-R) (Addgene plasmid # 55614; http://n2t.net/addgene:55614; RRID:Addgene_55614) were a gift from David Chan [23]. His-Ubiquitin constructs were kindly provided by Dr. Erazo [24]. The plasmid expressing a biotinylatable version of ubiquitin had been previously described elsewhere [25]. For gene interference, siRNAs were transfected using the calcium phosphate method described elsewhere [10]. Custom double-stranded siRNA oligonucleotides were obtained from GeneCust (Boynes, France). The forward sequences were as follows: negative control (NC) = 5′-UUCUCCGAACGUGUCACGUTT; HERC2 (H2.2) = GACUGUAGCCAGAUUGAAATT; HERC2 (H2.4) = GGAAAGCACUGGAUUCGUUTT; HERC1 = CGGCAUGGAUGAACAAAUUTT; MKK3 = GGAAGAAGGAUCUACGGAUTT; C-RAF = UAGUUCAGCAGUUUGGCUATT; A-RAF = AACAACAUCUUCCUACAUGAGTT; B-RAF = AAAGAAUUGGAUCUGGAUCAUTT; p53 = GACUCCAGUGGUAAUCUACTT.
Lentiviral vectors were produced in HEK 293 T cells. Cells were transfected with 7 μg pMD2.G, 7 μg psPAX2 (VSV-G) and 7 μg of either empty pLKO.1-Puro or pLKO.1‐shHERC2 by the calcium phosphate method. Media containing lentiviral particles were collected, filtered using polyvinyl difluoride filters (Millex-HV filter 0.45 μm, Millipore SLHV033RB) and stored in aliquots at − 80 °C. Target cells were seeded at a confluence of 50–60% in a 6-well plate before adding 300 μL of the medium containing the lentiviral vectors to each well. Fresh medium, supplemented with 5 μg/mL polybrene, was added to make a total volume of 1 mL. Media with lentiviral vectors were removed the next day and after 24 h, 5 μg/mL puromycin was added for selection. MISSION shRNA clone of mouse HERC2 (TRCN0000039444) was purchased from Sigma-Aldrich. The plasmid vector pLKO.1—TRC control was a gift from David Root (Addgene plasmid #10879; http://n2t.net/addgene:10879; RRID:Addgene_10879) [26], and the VSV-G envelope expressing plasmid pMD2.G (Addgene plasmid #12259; http://n2t.net/addgene:12259; RRID:Addgene_12259) and the lentivirus packaging plasmid psPAX2 (Addgene plasmid #12260; http://n2t.net/addgene:12260; RRID:Addgene_12260) were a gift from Didier Trono.
For protein extraction, cells were washed twice in ice-cold phosphate-buffer saline after discarding the media. Cell lysis was performed by scrapping after adding of NP40 lysis buffer (50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 50 mM NaF, 0.5% NP40) containing protease and phosphatase inhibitors as previously described [27]. Lysates were maintained on ice under agitation for 20 min, and then centrifuged at 13,000×g at 4 °C for 10 min. Supernatants were collected before analysis using the Tris–Acetate PAGE system [28]. Band intensities were detected using a gel documentation system (LAS-3000, Fujifilm) and quantified with ImageJ software (Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA, https://imagej.nih.gov/ij/). We used the following antibodies: anti-HERC2 monoclonal (BD Biosciences #612366); anti-C-RAF (BD Biosciences #610151); anti-Clathrin Heavy Chain (TD.1) (Santa Cruz Biotechnology #sc12734); anti-P-ERK1/2 (Sigma-Aldrich #M 8159); anti-p44/42 MAPK (ERK1/2) (Cell signalling #9102); anti-phospho-p38 (Cell signalling #9211); anti-p38 (Santa Cruz Biotechnology #sc-535); anti-HERC1 (410) [10]; anti-P-MKK3 (Cell signalling #9231); anti-MKK3 (Proteintech #13898–1-AP); anti-A-RAF (A-5) (Santa Cruz #sc-166771); anti-B-RAF (F-7) (Santa Cruz Biotechnology #sc-5284); anti-HERC2 polyclonal (bvg3) [10]; anti-c-myc (clone 9E10) (Roche #1 667 149); anti-GFP (Abcam #ab13970); anti-Flag M2 (Sigma-Aldrich #F 3165); anti-p-HSP27 (Enzo Life Sciences #ADI-SPA-523); anti-HSP27 (Santa Cruz Biotechnology #sc-1049); anti-NRF2 (Cell signalling #12721); anti-ubiquitylated proteins (clone FK2; Biomol); and peroxidase-conjugated secondary antibodies (Invitrogen).
We seeded U2OS cells on glass coverslips and performed fixation by incubating cells at room temperature for 20 min in 4% paraformaldehyde. Then, cells were permeabilised for 20 min with 0.05% saponin in phosphate-buffered saline containing 0.5% bovine serum albumin. The primary antibody, anti-phospho-p38 (Cell signalling #9211) (1:200), was incubated at 37 °C for 1 h. After washing, Alexa-Fluor 488 secondary antibody (Invitrogen) (1:500) was incubated at 37 °C for 45 min. Actin filaments were stained by incubation with phalloidin-Alexa 647 (BioProbes) (100 ng/mL) for 20 min at room temperature. Nuclei were stained with DAPI (Sigma-Aldrich) (1 μg/mL). All images were acquired using a confocal laser scanning microscope (LSM 880 spectral, Carl Zeiss Microscopy GmbH, Jena, Germany).
For immunoprecipitation, cells were lysed with CHAPS buffer (10 mM Tris–HCl, pH 7.5, 100 mM NaCl, 0.3% CHAPS) containing protease and phosphatase inhibitors as described above. Cell lysates (input) were incubated with pre-immune serum or anti-HERC2 polyclonal antibody (bvg3) for 2 h at 4 °C with gentle rotation and immunoprecipitated with protein A-Sepharose (GE Healthcare) for 1 h at 4 °C. Beads were pelleted by centrifugation at 2500×g, washed five times with CHAPS buffer, and analysed by electrophoresis and immunoblot. For the GFP pull-downs, supernatants were incubated with 2 μL GFP-TrapA (ChromoTek) for 2 h at 4 °C. Pellets were washed five times with CHAPS buffer and analysed by electrophoresis and immunoblot. For ubiquitome proteomic experiments, biotin-pull-downs were performed in triplicates as previously described [25], in order to compare proteins more ubiquitinated in Flag-HERC2 WT-overexpressing cells, relative to Flag-HERC2 C4762S-overexpressing cells.
HEK 293 T cells were transfected with the indicated plasmids for 48 h. Before the ubiquitylation assay, the cells were treated for 4 h with 10 µM of the proteasome inhibitor MG132 (Sigma-Aldrich/Merck #C2211). Then, cells were lysed with denaturing buffer #1 (6 M guanidinium-HCl, 10 mM Tris, 100 mM Na2HPO4–NaH2PO4 buffer, pH 8) and the cells extracts were incubated with the nickel beads (Ni2+-NTA agarose; Qiagen) for 2 h at 4 °C under rotation. Beads were successively washed as follows: twice with 1 ml of denaturing buffer #1 plus 10 mM 2-mercaptoethanol; three times with 1 ml of buffer #2 (8 M urea, 10 mM Tris, 10 mM 2-mercaptoethanol, 100 mM Na2HPO4–NaH2PO4 buffer, pH 8); twice with 1 ml of buffer #3 (8 M urea, 10 mM Tris, 100 mM Na2HPO4–NaH2PO4 buffer, pH 6.3) containing 0.2% Triton X-100; once with 1 ml of buffer #3 containing 0.1% Triton X-100 and 0.5 M NaCl; and three times with 1 ml of buffer #3. Finally, proteins were eluted by incubating the beads with 200 mM imidazole in 5% SDS, 0.15 M Tris–HCl, pH 6.7, 30% (vol/vol) glycerol, 0.72 M 2-mercaptoethanol for 1 h at 37 °C with mixing. The samples were analysed by immunoblotting as indicated above.
Total RNA was isolated from U2OS cells using the TRIsure reagent according to the manufacturer’s protocol (Bioline). Total RNA (2 μg) was reverse-transcribed using the high-capacity cDNA Reverse Transcription kit (Applied Biosystems). PCR amplification reactions were performed with the ABI Prism 7900 HT Fast Real-Time PCR System. Applied Biosystems’ TaqMan Gene Expression Assays (ThermoFisher Scientific) were used to quantify the gene expressions of the following: GUSB (Hs00939627_m1), NFE2L2 (Hs00975960_m1), SOD1 (Hs00533490_m1), SOD2 (Hs00167309_m1), GPX1 (Hs00829989_Gh), and the housekeeping gene GAPDH (Hs99999905_m1), which was used to normalise.
Using 96-well plates, U2OS cells and human skin fibroblasts were seeded to final concentration of 10,000 cells/well or 15,000 cells/well, respectively. After incubation at 37 °C for 24 h in the cell incubator, we initiated treatments, as indicated and performed the MTT assay (M5655; Sigma/Merck) according to manufacturer’s instructions. Briefly, we added MTT at a final concentration of 0.5 mg/mL to each well, incubated the cells for 4 h in a humidified incubator, then discarded the media and solubilised the formazan crystals with isopropanol. Finally, absorbance at a wavelength of 570 nm was determined using a 96-well plate spectrophotometer.
To evaluate mitochondrial reactive oxygen species (ROS), human skin fibroblasts were seeded in a µ-Slide 8 well-chambered coverslip at a concentration of 15,000 cells/well. The next day, cells were stained with 1 µg/mL of Hoechst 33,342 (H3570, ThermoFisher, USA) for 30 min at 37 °C and with 2 µM MitoSOX Red (Invitrogen) for 15 min at 37 °C. Cells were examined in a Zeiss LSM 880 laser scanning confocal spectral microscope equipped with an incubation control system (37 °C, 5% CO2). Fluorescence intensity per cell was measured, quantified and expressed as arbitrary units (a.u). Images were analysed using ImageJ software (Rasband, W.S., ImageJ, U.S. National Institutes of Health, Bethesda, Maryland, USA, https://imagej.nih.gov/ij/).
For mitochondria staining, human skin fibroblasts were seeded in a µ-Slide 8 well-chambered coverslip at a concentration of 15,000 cells/well. The next day, cells were stained with 1 µg/mL of Hoechst 33,342 (H3570, ThermoFisher, USA) and 50 nM Mitotracker Red CMXRos (M7512, ThermoFisher, USA) for 30 min at 37 °C. Images were taken using a Zeiss LSM 880 laser scanning confocal spectral microscope equipped with an incubation control system (37 °C, 5% CO2). Fragmented mitochondrial percentage was calculated by counting spherical non-contiguous mitochondrial particles and dividing by the number of total structures comprised in the mitochondrial network. Images were analysed using ImageJ software (Rasband, W.S., ImageJ, U.S. National Institutes of Health, Bethesda, Maryland, USA, https://imagej.nih.gov/ij/).
The results indicate the means and standard error of the mean (± SEM) of, at least, three independent experiments. Individual data points are plotted as single dots. Significance was calculated by Student t-test and indicated as follows: *, **, or *** for p values of ≤ 0.05, ≤ 0.01, or ≤ 0.001, respectively. Figures were created, and statistical analysis was performed, using GraphPad Prism version 8.4.3 for Windows (GraphPad Software, San Diego, California USA), www.graphpad.com.
Several recessive mutations affecting the HERC2 gene cause developmental delay with Angelman-like features [14, 19]. Knowing how pathologic HERC2 variants affect intracellular signalling could reveal the underlying pathology and identify possible therapies. Therefore, we studied cells from an individual with the mutant HERC2 P594L variant described in most cases. Since HERC1 had previously been reported to regulate the ERK and p38 MAPK signalling pathways [22, 29], we wondered if HERC2 also had a modulatory role. As expected, cells with the HERC2 P594L mutation showed almost undetectable HERC2 protein levels (Fig. 1A–C). Interestingly, although they showed higher protein levels of C-RAF (Fig. 1A), this did not correlate with the canonical activation of the ERK signalling pathway, assessed by ERK phosphorylation (Fig. 1B). An increment in p38 phosphorylation was also detected while total p38 protein levels remained stable (Fig. 1C). In order to provide more evidence that these changes in MAPK signalling pathways are a general hallmark of disease in patients with biallelic HERC2 mutations, we analysed samples of two more individuals carrying the mutant HERC2 P594L variant. Consistently, patients with the HERC2 P594L mutation (P1, P2 and P3) showed lower HERC2 protein levels than the wild-type controls (C1, C2 and C3). In addition, C-RAF protein levels and p38 phosphorylation were upregulated in all three patients, but no changes were detected in ERK activation (Fig. 1D). These results showed that cells with the HERC2 P594L mutation exhibit altered MAPK signalling pathway activation, as reflected by higher C-RAF and phospho-p38 protein levels.
To delve deeper into the molecular mechanisms involved in the altered MAPK signalling pathway in HERC2 P594L cells, we considered human cells with low levels of HERC2 protein shared this alteration. In knockdown experiments performed in human U2OS cells, cells were transfected with either a negative control (NC) small-interfering RNA (siRNA), an siRNA against HERC2, or a positive control siRNA against HERC1. The positive control was chosen because previous work had shown that HERC1 controls ERK and p38 signalling pathways modulating C-RAF protein levels [22, 29]. HERC2 knockdown mimicked the effect observed in HERC2 P594L cells, with depletion of HERC2 correlating with increased C-RAF protein levels. As expected, this was also observed after HERC1 silencing (Fig. 2A). HERC2 depletion modified neither A-RAF nor B-RAF protein levels (Fig. 2B, C). These data indicated that RAF regulation by HERC2 is specific for the C-RAF isoform. Next, we analysed the RAF MAPK signalling pathway, in which canonical RAF activation triggers ERK phosphorylation [29]. We noted that C-RAF upregulation observed after HERC1 depletion correlated with increased phosphorylated ERK levels, while total ERK protein levels remained stable. However, we detected no changes in ERK phosphorylation in the HERC2-depleted cells (Fig. 2D). These results suggested that C-RAF upregulation caused by HERC2 depletion was not signalled through the canonical MEK/ERK pathway.
Given that HERC1 regulates the MKK3/p38 axis through a RAF-dependent mechanism [29], we decided to study if this mechanism was the same for HERC2. We analysed levels of p38 phosphorylation in U2OS cells transfected with a negative control siRNA, an siRNA against HERC2, and a positive control siRNA against HERC1. We observed the induction of p38 phosphorylation in HERC2-depleted cells, though with total p38 protein levels remaining stable and higher C-RAF protein levels (Fig. 3A). Analogous behaviour was detected in HERC1-depleted cells (Fig. 3A). The same results for p38 phosphorylation were obtained when silencing HERC2 with siRNAs containing different RNA sequences (HERC2 H2.2 and HERC2 H2.4) (Fig. 3B). The phosphorylation of p38 is associated with its activation and nuclear translocation. To check this, we analysed p38 subcellular localisation. Immunofluorescence experiments showed increased p38 nuclear localisation in HERC2-depleted cells (Fig. 3C). This was quantified assessing the nucleus:cytoplasm ratio, which was higher in HERC2-depleted cells compared with control cells (Fig. 3D). After HERC2 silencing, p38 activation, was replicated in other human cells, such as the p53-lacking human non-small lung carcinoma cell line (H1299) and the non-tumorigenic human kidney 293 T cell line (HEK 293 T) (Fig. 3E). In addition, the same results were obtained in mouse cells and when using a different HERC2 silencing method. RAW 264.7 macrophage cell line, primary mouse osteoblasts and MEFs were infected with lentiviral particles carrying either an empty vector as a control (plKO) or a short hairpin RNA (shRNA) against HERC2. All HERC2 knockdown cells presented higher phospho-p38 protein levels compared to controls, while total p38 protein levels remained constant (Fig. 3F). In conjunction, these results demonstrated that HERC2 participates in regulating p38 signalling.
MAPK kinase (MAPKK or MKK) mediates p38 activation through phosphorylation. MKK3 is the dominant isoform in human U2OS cell lines [29], and its activation has been analysed by measuring its phosphorylation at Ser189 [30]. Thus, we analysed MKK3 activation and its total protein expression in HERC2-depleted U2OS cells, revealing that neither MKK3 phosphorylation at Ser189 nor total MKK3 protein levels were altered compared with control cells (Fig. 4A). To confirm whether p38 phosphorylation triggered by HERC2 depletion depends on MKK3, we co-transfected U2OS with an MKK3 siRNA and either the negative control or the HERC2 siRNA. This revealed that MKK3 knockdown significantly abolished the increment in p38 phosphorylation after HERC2 depletion (Fig. 4B). These data suggested that MKK3 activation caused the increase in phospho-p38 independent of phosphorylation at Ser189. Given the finding that HERC2 regulates C-RAF and p38 activation, we used two specific RAF kinase inhibitors to identify a potential crosstalk mechanism between the two pathways. LY300912 was used to inhibit all RAF isoforms, and Sorafenib was used to inhibit only B-RAF and C-RAF. In absence of the inhibitors, cells showed an increase in p38 phosphorylation after HERC2 depletion; remarkably, however, this increase was clearly abrogated after incubation with LY3009120 or Sorafenib inhibitors for 1 h (Fig. 4C). Since RAF isoforms interact by forming different heterodimers [31], sometimes all isoforms must be depleted to rescue the regulatory effects mediated by RAF proteins. Therefore, we co-transfected U2OS cells with siRNAs against C-RAF or all three RAF isoforms (A-RAF, B-RAF and C-RAF) along with either the negative control siRNA or the siRNA against HERC2 to achieve knockdown (Fig. 4D). Although silencing C-RAF alone was insufficient to reduce p38 phosphorylation significantly, silencing all three isoforms led to a significant decrease in p38 activation in HERC2-depleted cells (Fig. 4D). Unlike pharmacological inhibition of RAF, triple knockdown failed to produce a complete abrogation of p38 phosphorylation after HERC2 depletion, which is probably due to the fact that siRNA silencing did not achieve sufficient RAF isoforms knockdown. Altogether these results confirm the existence of a crosstalk between the RAF and p38 signalling pathways regulated by HERC2.
To further investigate the mechanism behind C-RAF regulation by HERC2, we analysed whether these two proteins can interact. In immunoprecipitation experiments in U2OS cells with a specific anti-HERC2 antibody (bvg3), endogenous HERC2 and C-RAF immunoprecipitated, while HERC1 did not, indicating that the interaction of HERC2 and C-RAF was independent of HERC1 (Fig. 5A). RAF hetero-dimerisation between its isoforms is a well-reported process [31], and consistent with this, A-RAF and B-RAF were also detected in HERC2 immunoprecipitated complexes (Fig. 5B, C). The same results were obtained in the human 293 T cell line (Fig. 5D–F). To identify the region of HERC2 interacting with C-RAF, we co-expressed a GFP-C-RAF fusion protein with a series of Myc-HERC2 fusion proteins in HEK 293 T cells (Supplementary Fig. 1A), followed by pull-down assays with GFP-binding beads. Constructs F4, F5, and F5CT coimmunoprecipitated with GFP-C-RAF, indicating that the HERC2 and C-RAF protein interaction occurs mainly in the carboxyl-terminus of HERC2 polypeptide chain. F5CT construct, which contains the HECT domain holding the ubiquitin ligase activity, showed the highest affinity with C-RAF, suggesting that this is the most relevant interaction site (Supplementary Fig. 1A). HEK 293 T cells were then co-transfected with a Flag-HERC2 full-length fusion protein along with GFP (as a negative control) or the GFP-C-RAF fusion constructs (CR1, CR2, CR3 or full-length) to map the C-RAF region involved. In the GFP pull-down, Flag-HERC2 was coimmunoprecipitated with CR1, CR3, and the full-length constructs (Supplementary Fig. 1B). To characterise this interaction further, we co-expressed the F4 Myc-HERC2 construct with GFP-C-RAF fusion constructs and performed a GFP pull-down, which showed preferential co-immunoprecipitation of the F4 construct with CR3 (Supplementary Fig. 1C). In parallel, the same experiment was done but with the F5CT Myc-HERC2 construct instead of F4, and this revealed co-immunoprecipitation of F5CT with CR1 and CR3 (Supplementary Fig. 1D). In conjunction, pull-down experiments confirmed the interaction between HERC2 and C-RAF, and indicated the possible domains involved. The HERC2 HECT domain, contained in the F5CT construct, showed the highest affinity for C-RAF and its catalytic domain (CR3), suggesting that the HECT and CR3 domains could be the most relevant at the physiological level. Subsequent structural studies should confirm this relevance.
Having shown that the ubiquitin E3 ligase HERC2 interacts with C-RAF and regulates its protein levels, we wanted to dissect whether HERC2 regulates C-RAF ubiquitylation targeting it to proteasomal degradation. To determine this, we analysed C-RAF ubiquitylation both in control and HERC2-depleted cells in the absence and presence of the proteasome inhibitor MG132. Control and HERC2-depleted HEK 293 T cells were transfected with constructs expressing GFP-C-RAF or GFP as a negative control. Forty-eight hours later, cells were incubated for 6 h in the absence or presence of MG132 (10 µM). Lysates from these cells were pulled down using GFP resin. Inputs and pull-down proteins were analysed by PAGE/SDS and immunoblotted with the anti-ubiquitylated proteins antibody (FK2) or with specific antibodies against the indicated proteins. GFP-C-RAF polyubiquitylation slightly decreased in HERC2-depleted cells compared to control cells under basal conditions (Fig. 5G, lane 10 compared with lane 12). Treatment with MG132 efficiently caused accumulation of polyubiquitylated GFP-C-RAF due to proteasome degradation inhibition in control cells (Fig. 5G, lane 10 compared with lane 14). Remarkably, under MG132 treatment, GFP-C-RAF polyubiquitylation levels were much lower in HERC2-depleted cells (Fig. 5G, lane 14 compared with lane 16). Altogether, these results demonstrated C-RAF proteasome-dependent degradation and its regulation by HERC2. To confirm the role of HERC2 regulating C-RAF polyubiquitylation we performed an ubiquitylation assay. First, we checked expression of different HERC2 constructs in HEK 293 T cells. We transfected either a negative control plasmid (Flag-CTL), a plasmid encoding wild-type HERC2 protein tagged with Flag peptide (Flag-HERC2 WT), or a plasmid encoding a mutant variant lacking ubiquitin ligase activity (Flag-HERC2 C4762S). HERC2 overexpression occurred in both Flag-HERC2 WT and Flag-HERC2 C4762S transfected cells, being greater with the mutated form (Supplementary Fig. 2A). Next, HEK 293 T cells were also transfected with GFP-C-RAF and His-tagged ubiquitin constructs. Cells were lysed after incubation with the proteasome inhibitor MG132 to enrich ubiquitylated proteins degraded by the proteasome, and incubated with Ni–NTA agarose resin to pull-down His-tagged ubiquitin molecules and the proteins to which they were attached (Fig. 5H). The results showed a smear of high molecular weight GFP-C-RAF indicating much greater C-RAF ubiquitylation with wild-type HERC2 overexpression. The plasmid construct HERC2 C4762S, which is catalytically inactive, was used to determine if the effect on C-RAF ubiquitylation by HERC2 WT overexpression, was dependent on the ubiquitin ligase activity of HERC2. Despite greater overexpression of the mutated form (HERC2 C4762S), this did not correlate with increased C-RAF ubiquitylation (Fig. 5H). Ubiquitin proteomic analysis further supported these results. In short, HEK 293 T cells were transiently transfected with a biotinylatable version of ubiquitin, along with either Flag-HERC2 WT or Flag-HERC2 C4762S plasmids. Ubiquitylated proteome from each condition was then isolated by a biotin-based pull-down approach [25] and analysed by mass spectrometry. C-RAF protein appeared in Flag-HERC2 WT-overexpressing cells, but not in Flag-HERC2 C4762S-overexpressing cells (Supplementary Fig. 2B, C). These results confirmed that HERC2 regulates C-RAF ubiquitylation.
Given that a major function of p38 is to regulate cellular stress, we analysed the cellular response to oxidative stress. U2OS cells were transfected with a negative control siRNA or an siRNA against HERC2, and oxidative stress was induced by treating cells with 500 µM H2O2 for different durations. Protein levels of phosphorylated p38 were analysed by immunoblot. As expected, HERC2-depleted cells started from a more phosphorylated basal state (t = 0) (Fig. 6A). After 3 h of H2O2 stimulation, both control cells and HERC2-depleted cells clearly showed induced phosphorylation of p38 and reached a maximum peak intensity, which is higher in HERC2-depleted cells. Interestingly, while p38 phosphorylation levels in control cells were clearly reduced after 6 and 12 h of treatment, the HERC2-depleted cells maintained significantly higher levels at these times, resulting in a more pronounced and prolonged phosphorylation response curve (Fig. 6A). Consistent with HERC2 having a role in regulating the cellular antioxidant response, mRNA levels of the antioxidant genes NFE2L2, SOD1, SOD2, and GPX1 increased in the HERC2-depleted cells compared with control cells. By contrast, mRNA levels of GUSB, used as a negative control, did not change significantly (Fig. 6B). Protein levels of the nuclear factor erythroid 2-related factor 2 (NRF2), a master regulator of all these antioxidants genes, were also upregulated in HERC2-depleted cells (Fig. 6C). To determine whether the role of HERC2 regulating the cellular response to oxidative stress depends on the activation of the RAF/MKK3/p38 signalling pathway we used p38 (SB203580) and RAF inhibitors (LY3009120). As previously shown, in absence of the inhibitors, cells showed an increase in NRF2 protein levels after HERC2 depletion; however, this increase was abrogated after incubation with SB203580 or LY3009120 for 1 h (Fig. 6D). These results suggested that p38 acts upstream NRF2 activation and that the cellular response to oxidative stress regulated by HERC2 depends on the RAF/MKK3/p38 signalling pathway. We then evaluated if HERC2 also regulates other stress types modulated by p38. To assess osmotic stress, we treated control cells and HERC2-depleted cells with 100 mM NaCl for different durations; as with H2O2, the HERC2-depleted cells maintained higher levels of p38 phosphorylation after 6 and 12 h (Supplementary Fig. 3). Ultimately, these data showed a complex regulation of downstream p38 signalling dependent on HERC2, pointing out HERC2 as a modulator of the cellular response to oxidative and saline stresses.
Finally, to determine whether cells with the HERC2 P594L mutation showed an altered response to H2O2-induced oxidative stress, we treated them with 500 µM H2O2 for different durations. Both the controls (HERC2 WT) and the fibroblasts carrying the mutation (HERC2 P594L) responded with a strong induction of p38 phosphorylation by 1–3 h after H2O2 treatment. Notably, HERC2 P594L cells maintained higher p38 phosphorylation levels after 6 h, while levels in control cells had already decreased to baseline (Fig. 7A). These differences in p38 signalling correlated with differences in cell morphology spotted by optical microscopy. After 3 h of treatment with H2O2, the HERC2 WT cells had already begun to show a rounder morphology, probably due to the toxic effect of H2O2, and after 6 h, most cells showed this altered morphology. By contrast, the HERC2 P594L cells seemed to be more resistant to H2O2 exposure, appearing healthier and more attached to the plate culture surface than controls at both 3 and 6 h (Fig. 7B). To confirm the differences in cell viability and test their dependence on the C-RAF/MKK3/p38 signalling pathway, MTT assays were performed in the presence of a p38 inhibitor (SB203580) or the RAF inhibitors (LY3009120 or Sorafenib). After 6 h of treatment with 500 µM H2O2, cell viability fell to 13.7% and 44.8% in the control cells and the HERC2 P594L cells, respectively. The higher resistance of HERC2 P594L cells to H2O2-induced oxidative stress was abrogated by treatment with the inhibitors (Fig. 7C). We then evaluated this effect under prolonged but less aggressive exposure to H2O2 (50 µM for 24 h). Again, HERC2 P594L cell viability was higher compared to the controls after stress exposure and, this higher resistance was abrogated by the p38 or RAF inhibitors (Fig. 7D). Previous studies have reported that HERC2 depletion enhances cell proliferation due to impaired p53 transcriptional activity [10–12]. Given that HERC2 modulates the activity of p53, we wanted to determine whether the evaluated effects on cell viability also depend, in part, on this tumour suppressor protein. As expected, HERC2-depleted cells with functional p53 (WT p53), presented higher cell growth compared to the control cells (Fig. 7E). The differences in cell growth between the control and HERC2-depleted cells were abolished under p53 knockdown (siRNA p53) (Fig. 7E), so we used this model (p53-knockdown U20S cells) to repeat the cell viability assay after H2O2 exposure. In untreated conditions, no significant differences were observed between negative control cells (NC + p53) and HERC2-depleted cells (HERC2 + p53) (Fig. 7F), including after treatment with the inhibitors. However, after 24 h of treatment with 50 µM H2O2, cell viability reduced to 67.5% in control cells and only to 89.2% in HERC2-depleted cells. Again, the higher cell resistance of HERC2-depleted cells was abrogated by treatment with p38 (SB203580) and RAF (LY3009120) inhibitors (Fig. 7F). Taken together, these results demonstrated that cellular resistance to H2O2-induced oxidative stress acquired by HERC2 deficiency is independent of p53, instead being mediated through the C-RAF/MKK3/p38 signalling pathway. These above results suggested that cells with HERC2 deficiency are better equipped against oxidative stress, so we wondered how does increased protection against oxidative stress fits into pathology. Excessive reactive oxygen species (ROS) cause oxidative stress. However, ROS also play a physiological role in cell signalling. Thus, appropriate ROS production is essential to maintain redox balance. Overexpression of antioxidant enzymes, such as NRF2, may lead the cell to a more reduced state. This pathophysiological situation is known as reductive stress and can be as harmful as is oxidative stress [32–34]. To assess this, mitochondrial ROS levels were evaluated with MitoSOX staining. Cells with the HERC2 P594L mutation showed lesser production of mitochondrial ROS than control cells, suggesting a more reduced state in these cells (Supplementary Fig. 4A). In addition, mitochondria were stained using MitoTracker probes. HERC2 P594L cells presented a more fragmented mitochondrial network than control cells, indicating a possible mitochondrial disfunction (Supplementary Fig. 4B). Further experiments should confirm these preliminary observations and deepen how ROS levels and mitochondrial function participate in the neurological syndrome caused by the HERC2 P594L variant.
This study provides the first evidence that HERC2 controls the cellular response to oxidative stress through the p38 signalling pathway dependent on RAF. Our results demonstrate that HERC2 forms a complex with RAF proteins, consistent with the results of a previous proteomic analysis, in which C-RAF was identified to interact with the carboxyl-terminus domain of HERC2 [35]. Mechanistically, our data show that HERC2 regulates C-RAF ubiquitylation and protein degradation; thus, in individuals with the HERC2 P594L mutation, the resulting HERC2 deficiency, causes an increase in C-RAF protein levels. However, this increase is not signalled through the canonical MEK/ERK pathway, and instead, seems to affect the MKK3/p38 pathway specifically (Fig. 8). Activation of crosstalk between C-RAF and the MKK3/p38 pathway has also been described as a mechanism regulated by HERC1, the other member of the large HERC protein family [29]. This raises the question of whether this signalling mechanism is specific to large HERC proteins. In any case, our results demonstrated that the role of HERC2 in the C-RAF/MKK3/p38 signalling pathway is independent of HERC1. Several lines of evidences show this independent role: (1) HERC1 and HERC2 proteins do not interact [29]; (2) HERC1 is not present in the HERC2/C-RAF complex (Fig. 5A); and (3) while HERC1 depletion regulates ERK signalling, HERC2 does not (Fig. 2D). By contrast, activation dependent on HERC2 affects the cellular response to oxidative and saline stresses. Although the precise mechanism explaining the differences between HERC1 and HERC2 should be explored further, differences could be explained by the different complexes formed between RAF proteins and large HERC proteins or by the pleiotropy of the p38 pathway [36]. Many p38 MAPK substrates have been described, both in the cytosol and the nucleus, and each large HERC family member appears to direct p38 signalling towards different downstream targets, suggesting the participation of different HERC1 and HERC2 complexes. NRF2, a transcription factor encoded by the NFE2L2 gene, is considered the master regulator of the cellular antioxidant response [37]. A critical regulatory step leading to its activation is its dissociation from Cullin 3 (CUL3) and the ubiquitin ligase Kelch-like ECH-associated protein 1 (KEAP1). CUL3 ubiquitylates NRF2, targeting it to proteasomal degradation, and upon exposure to oxidative stress, the NRF2-KEAP1 complex is disrupted and NRF2 is stabilised for translocation to the nucleus. Nevertheless, the precise mechanism by which cellular stress signals end up reaching NRF2 and causing its dissociation of the complex remains unclear [38]. Indeed, several studies have pointed out that some MAPK pathways are responsible for regulating this signal transduction. The p38 MAPK can regulate NRF2 activity through its activation [39–41] and its repression [42] depending on the context [43]. We observed NRF2 protein levels increasing after HERC2 depletion (Fig. 6C), consistent with p38 activating NRF2. Moreover, the mRNA levels of NRF2-regulated antioxidant genes also increased (e.g. SOD1, SOD2 and GPX1) (Fig. 6B). The NFE2L2 gene contains an antioxidant response element within its promoter region, providing NRF2 the ability to activate its own transcription [44]. This could explain why we observed increased mRNA levels of NFE2L2 in addition to its protein levels. In addition, the fact that the inhibition of RAF or p38 activity, abolished upregulation of NRF2 in HERC2-depleted cells, confirmed that p38 acts upstream of NRF2 activation. Still, given the variety of p38 substrates we cannot discard that other transcription factors, apart from NRF2, could also be involved in the regulation of the studied antioxidant genes. The transcription factor ATF-2 is another important mediator of p38 in the induction of SOD2 expression upon H2O2-induced oxidative stress in MEFs [45]. This possible cooperation between NRF2 and ATF-2, or some other transcription factor targeted by p38, should be studied further. Overall, our findings may have both physiological and clinical repercussions. Physiologically, we revealed a pro-survival function of p38 that is regulated by HERC2. HERC2 potentially fine-tunes the cellular response to oxidative stress by controlling protein levels of C-RAF and, therefore, C-RAF/MKK3/p38 signalling to regulate antioxidant gene expression. Clinically, these findings may be relevant to cancer, as well as individuals with HERC2 Angelman-like syndrome due to biallelic HERC2 gene variants [16, 18]. Several HERC2 mutations have been associated with a wide number of tumours [4]. In renal cancer, higher HERC2 gene expression correlates with better patient prognosis [46], supporting the hypothesis that HERC2 may act as a tumour suppressor [4, 46]. We previously demonstrated that HERC2, and NEURL4, regulate the transcriptional activity of the tumour suppressor p53, facilitating its oligomerisation. HERC2 knockdown accordingly increases cell proliferation due to the impaired capability to arrest the cell cycle through p53 [10–12]. Equally, although the precise mechanism remains elusive, it is well established that the production of reactive oxygen species in tumour cells increases due to the higher metabolic rate, with the resulting excess being countered by an increased antioxidant cellular response [47]. Supporting this, in mice, oncogenic alleles of Kras, Braf and Myc, associated with increased Nfe2l2 expression. This appear to stably enhance NRF2 antioxidant program and lower intracellular reactive oxygen species [48]. Furthermore, in p53 mutated cancer cells, the NRF2-dependent antioxidant response was selectively modulated to enhance cancer cell survival [49]. Our data reveal a new mechanism by which HERC2 deficiency may contribute to tumour malignancy by impairing p53 transcriptional activity, and also by boosting the cellular antioxidant response making cancer cells more resistant to oxidative stress (Fig. 8). In this context, combination treatments with drugs causing non-genotoxic activation of p53 oligomerisation and FDA-approved RAF inhibitors, such as Sorafenib, represent potential therapeutic candidates for tumours associated with HERC2 deficiency. Finally, a previous proteomic analysis of human HERC2 mutants (including the p.Pro594Leu variant studied here) has already identified an enrichment of the NRF2-mediated oxidative stress response in HERC2 mutants compared to control group [20]. In addition, protein–protein interaction networks containing signal transduction proteins and MAPKs were found to be differentially expressed in HERC2 mutants [20]. Our results add to these observations and, importantly, provide a possible molecular mechanism explanation. It would be interesting for further research to study the implication of a chronic activation of the C-RAF/MKK3/p38 signalling pathway in neuronal cells with HERC2 deficiency. Alterations in p38 MAPK signalling in neurons have been linked to neurodegenerative diseases including Parkinson’s disease, Alzheimer’s disease and amyotrophic lateral sclerosis (ALS) [50]. Therefore, we cannot discount the possibility that alterations in this pathway could be associated with clinical outcomes in HERC2 Angelman-like syndrome. Consistent with this, previous studies have shown that SOD1 overexpression, in which gene variants are associated with ALS and whose mRNA levels we found to be increased following HERC2 depletion, is associated with defects in the cerebellar architecture [51, 52]. In addition, while excessive ROS elicit oxidative stress, their persistent depletion, as observed in HERC2-deficient cells (Supplementary Fig. 4A), leads to an opposite condition called reductive stress. Persistent activation of antioxidant signalling can cause reductive stress and lead to pathology. In HERC2 P594L cells, the overactivation of NRF2 signalling could be one of the causes. In fact, NRF2 sustained activation has already been linked to reductive stress [53]. Highlighting the importance of reductive stress on pathology, mutations in key components of the cellular reductive stress response can cause developmental diseases. For instance, FEM1B gain-of-function mutation, which cause a persistent activation of the reductive stress response, elicit developmental syndromes with some similarities to the HERC2 Angelman-like syndrome [33, 34]. An example of the damage that reductive stress can exert on cells is that it can induce mitochondrial dysfunction and impact on the correct cell function [54, 55]. Accordingly, we observed an increased number of fragmented mitochondria in HERC2 P594L cells, which is a common feature observed in neurodegeneration [56]. However, more experiments are needed to confirm these hypotheses and to associate these mechanisms with clinical outcomes in HERC2 Angelman-like syndrome. All things considered, the findings in this study identify p38 and RAF inhibitors as potential therapeutic options for individuals who present with such rare disease.
Below is the link to the electronic supplementary material.Supplementary file1 (PDF 633 KB) | true | true | true |
PMC9568475 | 35986859 | Qinfan Yao,Cuili Wang,Yucheng Wang,Xiuyuan Zhang,Hong Jiang,Dajin Chen | The integrated comprehension of lncRNA HOXA-AS3 implication on human diseases | 20-08-2022 | HOXA-AS3,Long non-coding RNAs,Tumor promoter,Mechanism,Clinical applications | Long non-coding RNA (lncRNA) is a non-protein-coding RNA with a length of more than 200 nucleotides. Studies have shown that lncRNAs have vital impacts on various pathological processes and participate in the development of human diseases, usually through acting as competing endogenous RNAs to modulate miRNA expression and biological functions. lncRNA HOXA Cluster Antisense RNA 3 (HOXA-AS3) was a newly discovered lncRNA and has been demonstrated to be abnormally expressed in many diseases. Moreover, HOXA-AS3 expression was closely correlated with the clinicopathologic characteristics in cancer patients. In addition, HOXA-AS3 exhibited significant properties in regulating several biological processes, including cell proliferation, invasion, and migration. Furthermore, HOXA-AS3 has provided promising values in the diagnosis, prognosis, and therapeutic strategies of several diseases such as liver cancer, glioma, lung cancer, oral cancer, gastric cancer, and even atherosclerosis. In this review, we discuss the abnormal expression of HOXA-AS3 in several human disorders and some pathobiological processes and its clinical characteristics, followed by a summary of HOXA-AS3 functions, regulatory mechanisms, and clinical application potential. | The integrated comprehension of lncRNA HOXA-AS3 implication on human diseases
Long non-coding RNA (lncRNA) is a non-protein-coding RNA with a length of more than 200 nucleotides. Studies have shown that lncRNAs have vital impacts on various pathological processes and participate in the development of human diseases, usually through acting as competing endogenous RNAs to modulate miRNA expression and biological functions. lncRNA HOXA Cluster Antisense RNA 3 (HOXA-AS3) was a newly discovered lncRNA and has been demonstrated to be abnormally expressed in many diseases. Moreover, HOXA-AS3 expression was closely correlated with the clinicopathologic characteristics in cancer patients. In addition, HOXA-AS3 exhibited significant properties in regulating several biological processes, including cell proliferation, invasion, and migration. Furthermore, HOXA-AS3 has provided promising values in the diagnosis, prognosis, and therapeutic strategies of several diseases such as liver cancer, glioma, lung cancer, oral cancer, gastric cancer, and even atherosclerosis. In this review, we discuss the abnormal expression of HOXA-AS3 in several human disorders and some pathobiological processes and its clinical characteristics, followed by a summary of HOXA-AS3 functions, regulatory mechanisms, and clinical application potential.
Long non-coding RNA (lncRNA) represents a non-coding functional RNA subtype with over 200 nucleotides in length [1–5]. Along with a growing number of long non-coding RNAs (lncRNAs) being identified in recent years, investigating the biological functions of lncRNAs has gained increasing attention [5–9]. Increasing evidence suggests that lncRNA dysregulation is linked to illness onset and progression, particularly malignancies [10–13]. In addition, functional investigations have also indicated that lncRNAs play an essential role in the pathogenesis of various diseases via several molecular processes, including cell proliferation, metabolism, migration, invasion, and apoptosis [12, 14]. Furthermore, lncRNAs have emerged as novel focuses of clinical applications due to the increasing in-depth studies on molecular mechanisms [14–18] for the functions of numerous lncRNAs [17, 19–21]. lncRNA HOXA Cluster Antisense RNA 3 (HOXA-AS3) is a newly discovered lncRNA with 25,952 bases and in the genomic location at human chromosome 7p15.2 started 900 nt downstream of the 3’ end of HOXB5 [22]. HOXA-AS3 expression has been implicated in a variety of human diseases and pathophysiological processes, including liver cancer [23–25], glioma [26, 27], lung cancer [28–30], oral cancer [31], colorectal cancer [32], gastric cancer [33], pancreatic cancer [34], endometriosis [35], atherosclerosis [36, 37], pulmonary arterial hypertension [38, 39], and even the lineage differentiation of mesenchymal stem cells (MSCs) [39], according to numerous studies. In these disorders, high HOXA-AS3 expression has been reported to closely relate to several clinicopathologic characteristics, such as pathological grade, TNM stage, tumor size, lymph node metastasis, invasion depth, and Helicobacter pylori infection status, overall survival, and disease-free survival. Research has further revealed that HOXA-AS3 exerted regulatory effects on the initiation and progression of various human disease types through the positive induction of many cellular processes such as cell proliferation, apoptosis, migration, invasion, chemotherapy resistance, endothelium inflammation, and MSCs differentiation. Furthermore, HOXA-AS3 has been shown to have a high potential for a variety of interesting therapeutic applications in diagnosis, prognosis, and therapy. In this review, we focus on the expression profiles, corresponding clinicopathologic features, biological roles, molecular mechanisms, and clinical applications of HOXA-AS3 in diverse disease types and pathophysiological processes.
Recent evidence showed that the overexpression of HOXA-AS3 was revealed in various types of human diseases and pathophysiological processes, including liver cancer [23–25], glioma [26, 27], lung cancer [28–30], oral cancer [31], colorectal cancer [32], gastric cancer [33], pancreatic cancer [34], endometriosis [35], atherosclerosis [36, 37], pulmonary arterial hypertension [38], and the lineage differentiation of mesenchymal stem cells (MSCs) [39]. In addition, high HOXA-AS3 expression has been identified to correlate with unfavorable clinicopathological features and poor prognoses, such as pathological grade, TNM stage, tumor size, lymph node metastasis, invasion depth, and Helicobacter pylori infection status, overall survival, and disease-free survival (Table 1). HOXA-AS3 was also involved in regulating biological functions and disease processes through various mechanisms, including cell proliferation, apoptosis, migration, invasion, drug resistance, endothelium inflammation, and MSCs lineage specification (Table 2). This section briefly introduces HOXA-AS3 expression changes, relevant clinicopathologic features, and the leading biological roles in diverse disease types and pathophysiological processes.
It was found that HOXA-AS3 expression was markedly upregulated in hepatocellular carcinoma cells (Hep3B, SNU-387, Li-7, SMMC-7721, HepG2, Huh7, and HCC-LM3) and tissues [23–25]. In addition, patients with high HOXA-AS3 levels were confirmed to possess shorter overall survival. Similarly, functional studies revealed that HOXA-AS3 increased cell proliferation, anti-apoptosis, migration, invasion, the epithelial–mesenchymal transition (EMT), and the MEK/ERK signaling pathway in Hep3B, HuH-7, SMMC-7721, and HepG2 cells, as well as tumor growth and lung metastasis in mouse xenograft models [23–25].
HOXA-AS3 was highly expressed in glioma tissues and cell lines (LN229, U251, SNB19, U87, U138, and H4) and was closely associated with poor prognoses such as worse overall survival and pathological grade [26, 27]. HOXA-AS3 was demonstrated, through functional experiments, to facilitate the processes of cell proliferation, anti-apoptosis, and migration in LN229, H4, and U251 cells and in vivo tumor xenograft models [26, 27].
In small cell lung cancer, HOXA-AS3 expression was shown to upregulate over fourfold in patients’ tissues acquired stable disease (SD)/progressive disease (PD) after first-line chemotherapy compared to partial response (PR) patients’ tissues [28]. Furthermore, in a dose- and time-dependent manner, HOXA-AS3 was dramatically overexpressed in non-small-cell lung cancer tissues and cell lines after cisplatin therapy [29, 30]. Moreover, HOXA-AS3 exerted chemoresistance functions by inducing anti-apoptosis and EMT in A549, PC-9, NCI-H358, and NCI-H1299 cells. The developed xenograft mice model confirmed that HOXA-AS3 knockdown increased cisplatin effectiveness in lung cancer [30].
In addition, HOXA-AS3 was also highly expressed in oral squamous cell carcinoma tissues and cell lines (TSCCA, CAL-27, SCC-9, and Tca8113) [31]. Furthermore, high levels of HOXA-AS3 indicated an undesirable pathological stage and overall survival. Moreover, HOXA-AS3 has also played proliferative roles on SCC-9 and CAL-27 cells [31].
Overexpression of HOXA-AS3 has been reported in colorectal cancer tissues and SW480, SW620, HCT116, COLO205, and LOVO cells. In addition, HOXA-AS3 showed strong abilities to stimulate cell proliferation, suppress cell apoptosis in COLO205 and LOVO cell lines, and accelerate tumor growth in vivo [32].
HOXA-AS3 was upregulated in gastric cancer cell lines (MGC-803, AGS, MKN45, SGC7901, and HGC-27) and tissues. In addition, high HOXA-AS3 levels were correlated with poor prognosis, including tumor size, lymph node status, invasion depth, Helicobacter pylori infection status, over survival, and disease-free survival. More crucially, HOXA-AS3 increased cell proliferation, migration, and invasion in MKN45 and SGC7901 cells and tumor development and lung metastasis in vivo [33].
HOXA-AS3 was overexpressed in pancreatic cancer tissues and Panc-1, Aspc-1, sw1990, and Bxpc-3 cells and was closely bound up with worse prognosis, aggressive TNM stage, and lymph node metastasis. HOXA-AS3 was also reported to exist in its pro-proliferative effects for Panc-1 and Bxpc-3 cells as well as subcutaneous xenograft tumors [34].
The elevation of HOXA-AS3 was observed in human umbilical vein endothelial cells (HUVECs) and associated with worse pathological conditions of the coronary wall, increased levels of TG, TC and LDL-C in the serum of mice model. By promoting HUVEC adherence to monocytes and monocyte movement across HUVEC monolayers [36], HOXA-AS3 was thought to be a crucial activator for endothelium inflammation. HOXA-AS3 was also involved in the anti-proliferative and apoptotic effects of ox-LDL-induced HUVECs and the advancement of angiogenesis [37].
HOXA-AS3 was also overexpressed hypoxia-treated human pulmonary artery smooth muscle cells (HPASMCs), which was regarded as in vitro model of pulmonary arterial hypertension (PAH). Besides, HOXA-AS3 has been proved to promote proliferation and migration but repress the apoptosis of HPASMCs [38].
The expression of HOXA-AS3 in mesenchymal stem cells (MSCs) was increased during adipogenic differentiation, whereas it was unchanged during osteogenic differentiation. Moreover, HOXA-AS3 has been proposed as a critical factor in epigenetic tuning that contributed to the lineage differentiation of MSCs. Downregulation of HOXA-AS3 in both human MSCs and mouse MSCs resulted in improved osteogenesis and impaired adipogenesis [39].
Numerous studies have found that lncRNAs primarily work by interacting with miRNAs and interfering with gene expression [40–44]. Functional analysis indicated that HOXA-AS3 participated in regulating various cell biological processes via numerous mechanisms, including cell proliferation, apoptosis, invasion, metastasis, chemotherapy sensitivity, endothelium inflammation, and MSCs-lineage differentiation. Following that, we will discuss the molecular mechanisms of fundamental biological tasks such as cell proliferation, migration, and invasion. Uncontrolled cell proliferation constitutes malignant transformation and eventually tumor occurrence [44–48]. Meanwhile, cell migration and invasion are indispensable properties for cancer metastasis, mainly attributed to cancer-related death [49–53]. It is clear that HOXA-AS3 exerts several functions in many cancer types. In hepatocellular carcinoma, HOXA-AS3 enhanced cell proliferation, migration, invasion, and the epithelial–mesenchymal transition (EMT) process by interacting with miR-29c to increase BMP1 expression in SMMC-7721 and HepG2 cells, as well as binding to miR-455-5p to upregulate PD-L1 expression in Hep3B and HuH-7 cells (Fig. 1) [23, 24]. Moreover, HOXA-AS3 was also shown to sponge miR-455-5p to increase USP3 levels in glioma LN229 and H4 cells, allowing them to proliferate, and migrate [27]. In addition, HOXA-AS3 combined with HOXA3 to induce EMT and inhibit cell apoptosis in non-small-cell lung carcinoma cells (A549, PC-9, NCI-H358, and NCI-H1299), therefore, weakening the effectiveness of cisplatin treatment [29]. Furthermore, in lung adenocarcinoma A549 cells, HOXA-AS3 was shown to be triggered by histone acylation and then bound to NF110 to raise HOXA6 levels, leading to cell proliferation, migration, and invasion [30]. In oral squamous cell carcinoma SCC-9 and CAL-27 cells, HOXA-AS3 performed the pro-proliferative functions through the interaction with miR-218-5p [31]. HOXA-AS3 also strengthened colorectal cancer cell proliferation and apoptosis in COLO205 and LOVO cells by repressing the expression of miR-4319 to activate SPNS2 expression and AKT signaling pathway [32]. In gastric cancer MKN45 and SGC7901 cells, HOXA-AS3 has been found to accelerate the processes of cell proliferation, migration, and invasion via restraining miR-29a-3p from raising the LTβR expression and then sensitizing the NF-κB signaling pathway [33]. Moreover, HOXA-AS3 exerts a pro-proliferative action in pancreatic cancer Panc-1 and Bxpc-3 cells through the activation of NF-κB signaling. Recently, HOXA-AS3 was also shown to play pivotal roles in non-cancer diseases. During the process of inflammatory atherosclerosis, HOXA-AS3 was revealed to enhance monocyte migration through HUVEC monolayers through the activation of NF-κB signaling and exhibit an anti-proliferative and apoptotic effect on ox-LDL induced HUVECs via miR-455-5p/p27 Kip1 axis. In osteogenic differentiation of MSCs, HOXA-AS3 regulated the lineage specification of MSC by combining with EZH2 and also influencing the H3 lysine-27 trimethylation of Runx2 [39]. Besides, HOXA-AS3 combined with miR-675-3p to elevate PDE5A levels, boosting the proliferation, migration, and anti-apoptosis of hypoxia-treated human pulmonary artery smooth muscle cells [38].
lncRNAs are considered potential biomarkers in illness diagnosis and prognosis based on the properties of tissue-specific lncRNA expression patterns and their extensive participation in numerous biological processes [54–57]. Recently, multiple published studies have shown the significance of HOXA-AS3 in the diagnostic and prognostic potentials for diverse diseases. The differences in the expression of HOXA-AS3 in a variety of diseases can be used to distinguish diseased from adjacent normal tissues and cell lines, contributing to good diagnostic value for disease screening. Moreover, the strong correlation between HOXA-AS3 expression with the clinical features in various disease types has also been confirmed to possess solid prognostic value for disease risk assessment. For example, Kaplan–Meier analysis demonstrated that patients with high HOXA-AS3 expression suffered lower overall survival rates and even disease-free survival rates. HOXA-AS3 expression was also verified as an independent prognostic predictor in various disorders, including hepatocellular carcinoma, glioma [26, 27], oral squamous cell carcinoma, and gastric cancer, using univariate and multivariate Cox regression analyses [23–25]. Moreover, the expression profile of HOXA-AS3 increased in a dose- and time-dependent manner after cisplatin treatment, suggesting the potential of HOXA-AS3 to predict cisplatin resistance [29]. Nevertheless, HOXA-AS3 expression was currently detected in only tissues and cell lines, while have not been extensively performed on blood and other accessible body fluids. In addition, restricted accessibility of tumor tissue, inconvenient tissue storage, traumatic nature, and high cost make tissue biopsy an unsatisfactory choice for clinical applications [58–64]. Based on the several deficiencies of biopsy tissues, non-invasive and minimally invasive body fluid assessments are more conducive to early disease identification, more convenient for monitoring disease prognosis repeatedly, and more helpful in guiding timely intervention and treatment [65–69]. Consequently, it is necessary to explore further the specificity and stability of HOXA-AS3 expression in biological samples with less invasive (such as blood and urine) in different diseases.
lncRNAs are involved in various malignant biological behaviors, and understanding the potential role of lncRNAs in diseases represents a new perspective to develop molecular targeted therapeutic strategies [70–74]. Previously, it has been reported that HOXA-AS3 was overexpressed in numerous disorders and exerted a critical role in driving the disease formation and progression. Besides, numerous studies have further shown that HOXA-AS3 modulated the biological processes of disease types through several mechanisms. HOXA-AS3 have been demonstrated to function as competing endogenous RNAs (ceRNA) by sponging miRNA and inhibiting miRNA function. Targeting HOXA-AS3 to repress its abnormal upregulation has been shown to suppress disease progression in many disease models, representing a potent approach to developing therapeutic agents for several types of disease. For example, suppression of HOXA-AS3 in glioma tumor models showed an apparent reduction in glioma tumor weight and size, suggesting the promising therapeutic potential of targeted-HOXA‐AS3 for glioma [27]. Besides, downregulation of HOXA-AS3 by CRISPR-dCas9 has also been indicated to impair the tumor growth of pancreatic cancer [34] Panc-1 cells in vivo. Moreover, the 5′-terminal region of HOXA-AS3 from nt 1 to 800 was verified to be associated with the modulation of NF-κB activity and, therefore, targeting the region of HOXA-AS3 from nt 1 to 800 may act as a promising therapeutic strategy for the treatment of multiple NF-κB-mediated inflammatory disorders [36]. Besides, HOXA-AS3 knockdown also dramatically alleviated the symptom of atherosclerosis and ameliorated the pathological change in the coronary wall, the levels of TG, TC, and LDL-C in serum of mice, exhibiting the powerful therapeutic effect on atherosclerosis [37]. In addition, chemoresistance [75–78] remains a primary obstacle in cancer therapy. Elucidating the underlying mechanism of chemoresistance could improve the curative effect of chemotherapy and guide the effective therapeutic method [79–83]. It has been reported that HOXA-AS3 expression was implicated in the development of small cell lung cancer chemotherapy [28] insensitivity and non-small-cell lung carcinoma cisplatin [29] resistance, knockdown of HOXA-AS3 enhanced the efficacy of chemotherapeutic drugs in lung cancer. These results suggest that suppressing HOXA‐AS3 expression may have therapeutic potential for numerous diseases. Moreover, it may have significant implications for studying and treating different clinical disorders. However, the clinical treatment strategies based on HOXA‐AS3-targeted agents have not yet been broadly applied. In future research, the safety, stability, and efficacy of HOXA‐AS3-targeted drugs should be adequately evaluated in large-scale randomized clinical trials. It is a promising field of HOXA‐AS3 in non-invasive disease detection, prognosis, and target treatment.
HOXA-AS3 expression was upregulated in several human disorders and pathophysiological processes, such as liver cancer, glioma, lung cancer, oral cancer, colorectal cancer, gastric cancer, pancreatic cancer, endometriosis, atherosclerosis, pulmonary arterial hypertension, and the osteogenic differentiation of MSCs. Furthermore, HOXA-AS3 overexpression was tightly associated with numerous clinicopathological features, such as tumor size, pathological grade, lymph node metastasis, infection status, overall survival, and disease-free survival. In addition, HOXA-AS3 also played a crucial role in regulating biological functions and participated in the pathogenesis of diseases by multiple mechanisms. It is also well established that HOXA-AS3 was considered a promising biomarker and had a beneficial impact on clinical applications, including diagnosis, prognosis, and treatment. Additional in vivo and clinical experiments of identifying HOXA-AS3 expression characteristics in non-invasive samples and testing the efficacy and safety of targeted-HOXA-AS3 drugs will provide helpful insights into future disease management. | true | true | true |
PMC9568505 | Olusola A. Ogunyewo,Omoaruemike E. Okereke,Sandeep Kumar,Syed Shams Yazdani | Characterization of a GH5 endoxylanase from Penicillium funiculosum and its synergism with GH16 endo-1,3(4)-glucanase in saccharification of sugarcane bagasse | 14-10-2022 | Biochemistry,Biotechnology,Microbiology,Molecular biology | The production of second-generation fuels from lignocellulosic residues such as sugarcane bagasse (SCB) requires the synergistic interaction of key cellulose-degrading enzymes and accessory proteins for their complete deconstruction to useful monomeric sugars. Here, we recombinantly expressed and characterized unknown GH5 xylanase from P. funiculosum (PfXyn5) in Pichia pastoris, which was earlier found in our study to be highly implicated in SCB saccharification. The PfXyn5 has a molecular mass of ~ 55 kDa and showed broad activity against a range of substrates like xylan, xyloglucan, laminarin and p-nitrophenyl-β- d -xylopyranoside, with the highest specific activity of 0.7 U/mg against xylan at pH 4.5 and 50 °C. Analysis of the degradation products of xylan and SCB by PfXyn5 showed significant production of xylooligosaccharides (XOS) with a degree of polymerization (DP) ranging from two (DP2) to six (DP6), thus, suggesting that the PfXyn5 is an endo-acting enzyme. The enzyme synergistically improved the saccharification of SCB when combined with the crude cellulase cocktail of P. funiculosum with a degree of synergism up to 1.32. The PfXyn5 was further expressed individually and simultaneously with a notable GH16 endoglucanase (PfEgl16) in a catabolite-derepressed strain of P. funiculosum, PfMig188, and the saccharification efficiency of the secretomes from the resulting transformants were investigated on SCB. The secretome of PfMig188 overexpressing Xyn5 or Egl16 increased the saccharification of SCB by 9% or 7%, respectively, over the secretome of PfMig188, while the secretome of dual transformant increased SCB saccharification by ~ 15% at the same minimal protein concentration. | Characterization of a GH5 endoxylanase from Penicillium funiculosum and its synergism with GH16 endo-1,3(4)-glucanase in saccharification of sugarcane bagasse
The production of second-generation fuels from lignocellulosic residues such as sugarcane bagasse (SCB) requires the synergistic interaction of key cellulose-degrading enzymes and accessory proteins for their complete deconstruction to useful monomeric sugars. Here, we recombinantly expressed and characterized unknown GH5 xylanase from P. funiculosum (PfXyn5) in Pichia pastoris, which was earlier found in our study to be highly implicated in SCB saccharification. The PfXyn5 has a molecular mass of ~ 55 kDa and showed broad activity against a range of substrates like xylan, xyloglucan, laminarin and p-nitrophenyl-β-d-xylopyranoside, with the highest specific activity of 0.7 U/mg against xylan at pH 4.5 and 50 °C. Analysis of the degradation products of xylan and SCB by PfXyn5 showed significant production of xylooligosaccharides (XOS) with a degree of polymerization (DP) ranging from two (DP2) to six (DP6), thus, suggesting that the PfXyn5 is an endo-acting enzyme. The enzyme synergistically improved the saccharification of SCB when combined with the crude cellulase cocktail of P. funiculosum with a degree of synergism up to 1.32. The PfXyn5 was further expressed individually and simultaneously with a notable GH16 endoglucanase (PfEgl16) in a catabolite-derepressed strain of P. funiculosum, PfMig188, and the saccharification efficiency of the secretomes from the resulting transformants were investigated on SCB. The secretome of PfMig188 overexpressing Xyn5 or Egl16 increased the saccharification of SCB by 9% or 7%, respectively, over the secretome of PfMig188, while the secretome of dual transformant increased SCB saccharification by ~ 15% at the same minimal protein concentration.
The rapid increase in energy demand along with the continuous increase in the price of fossil fuels constantly necessitates the need to diversify the channels of energy generation for every growing economy. The increasing shortage of non-renewable energy resources and huge environmental pressure from greenhouse gases released by the burning of fossil fuels continue to prompt growing studies in search of alternative production of cleaner fuels compared to petroleum-based fuels. Second-generation biofuels produced from lignocellulosic biomass continue to attract wide attention as bountiful sources as they neither compete with food requirements nor impact the food and feed markets. Sugarcane bagasse (SCB) is one of the most abundant and promising biomass sources in the world which is obtained from the processing of sugarcane with a worldwide annual production of approximately 54 million tons. On average, approximately 270 kg of bagasse (50% moisture) per metric ton of sugarcane is generated by the sugar factories as a major by-product. SCB predominantly consists of cellulose, hemicellulose and lignin that are strongly associated to form the plant cell wall. The bioconversion process of this biomass requires several arrays of lignocellulolytic enzymes which act synergistically to deconstruct the tightly packed polymeric carbohydrate components into monomeric sugars that can subsequently be converted to fuels and valuable chemicals. Previous studies have shown that several biomass conversion processes rely on biomass pretreatment before enzymatic hydrolysis, to remove lignin and reduce the recalcitrance nature of the biopolymer. However, this process often leads to a reduction in the hemicellulose content. Owing to this challenge, several mild pretreatment techniques are getting developed to reduce the loss in hemicellulose content of lignocellulosic biomass thereby giving attention to more characterization of hemicellulases and notable accessory enzymes beneficial to biomass deconstruction. Hemicellulases such as xylanases, lichenases and laminarases which are endoglucanases active on mixed-glucans have earlier been reported to improve the hydrolysis of xylan and cellulose thereby contributing to the reduction of enzyme dosage for complete biomass saccharification. Xylanases (EC. 3.2.1.8), are crucial enzymes employed for the hydrolysis of β-1,4-d-xylosidic linkages in the xylan biomass, releasing xylooligosaccharides (XOS) and xylose. This enzyme has been reported to demonstrate wide application in several industrial processes such as in biorefinery, food and biobleaching of paper and pulp. It has been shown that xylanase treatment reduces the content of hexenuronic acid (HexA) in the pulp. HexA is a component formed during chemical pulping from glucuronic acid, which is part of hemicellulose polymers. Based on the sequence and protein structure, xylanases are classified into many glycosyl hydrolase families (GH) such as 7, 8, 10, 11, 16, 30, 43, 51, 98, and 141, with most of the characterized fungal xylanases being GH 10 and 11. While there are extensive studies on GH5 xylanases from bacterial hosts, there is a dearth of information on fungal GH5 xylanase till date. Furthermore, due to the presence of β-(1,3)- and β-(1,4)-d-glucans in plant cell walls, the requirements for mixed-linked glucanases such as endo-1,4-β-glucanase (EC 3.2.1.4), endo-1,3(4)-β-glucanase (EC 3.2.1.6), endo-1,3-β-glucanase (EC 3.2.1.39), and endo-1,3–1,4-β-glucanase (EC 3.2.1.73) in cellulolytic enzyme cocktails could also be important for better saccharification performance. In an attempt to develop effective technology for utilization of SCB in biorefineries, especially for production of second-generation biofuels, we previously identified the fungus Penicillium funiculosum NCIM1228 as a suitable host for the production of potent lignocellulolytic enzymes for the deconstruction of diverse pretreated cellulosic feedstocks as its secretome exhibited outstanding saccharification potential when compared with secretomes from other fungal strains. Using a combination of both proteomic and biochemical approaches, we identified some relevant accessory enzymes such as xylanases (GH30), endo‑β‑1,3(4)‑glucanase (GH16), β‑1,3‑galactosidase (GH43), cutinase (CE5), and some glycoside hydrolases with unknown functions important for the conversion of SCB to monomeric sugars when the production media for cultivation of Penicillium funiculosum was modulated with SCB. It was seen that the remarkable induction of these enzymes in the secretome of P. funiculosum facilitated complete saccharification of SCB although at a high enzyme dosage. Therefore, this study was designed to further improve the quality of enzyme systems in the secretome of P. funiculosum for enhanced saccharification of SCB at low enzyme loading to reduce the overall enzyme requirement for SCB deconstruction. To achieve this, two accessory enzymes, i.e., a GH5 protein with an unknown function and a GH16 endoglucanase, both found to be highly upregulated in the secretome of P. funiculosum in the presence of SCB, were selected for overexpression in P. funiculosum. The unknown function enzyme from P. funiculosum belonging to glycosyl hydrolase family 5 (PfGH5) was first recombinantly expressed in Pichia pastoris and its characteristics were evaluated to decipher its role in biomass hydrolysis. Furthermore, its impact on SCB hydrolysis was investigated and then it was overexpressed along with a GH16 endoglucanase in the background of a catabolite-derepressed strain of P. funiculosum (PfMig188) to produce a more efficient cocktail specific for this biomass. The resulting enzyme cocktail from the engineered strain was then assessed for improved production of hemicellulolytic enzymes as well as for their saccharification efficiency toward SCB with minimal enzyme loading.
A unique GH5 family protein was one of the proteins that was found to be highly upregulated in our earlier study in the secretome of P. funiculosum in response to the pre-treated sugarcane bagasse and was implicated in the enhanced hydrolysis of the pretreated bagasse by the secretome. The sequence analysis of this GH5 family protein of P. funiculosum, PfGH5, showed that this protein has not yet been characterized since the functions of the most similar genes that have been deposited in NCBI as well as in the CAZY database are unknown. As a result, we attempted to characterize and assess the functionality of this gene. The sequence analysis indicated that the PfGH5 contains 1561 nucleotides which encode a chain of 454 amino acid residues with a theoretical molecular weight of 50.1 kDa. Further analysis showed that there were four potential N-glycosylation sites (Asn-Ala-Ser) found in the PfGH5 sequence using the NetNGlyc1.0 server (http://www.cbs.dtu.dk/services/NetNGlyc/), while there was no potential O-glycosylation predicted using the NetOGlyc 4.0 server (http://www.cbs.dtu.dk/services/NetOGlyc/). The PfGH5 amino acid sequence revealed 100% similarity to an uncharacterized glycoside hydrolase family 5 protein from P. occitanis (accession number PCH04120.1), while 99% similarity was found with Talaromyces cellulolyticus (GAM37412.1) and Talaromyces pinophilus (KAF3403216.1) as seen also from the phylogenetic tree (Fig. 1). Surprisingly, the protein also exhibited 82% similarity with a Talaromyces islandicus GH5 protein annotated to be a β-xylosidase suggesting its possibility as a xylan-degrading enzyme. Further analysis using the InterPro scan tool revealed the presence of a conserved domain of cellulase super-family glycoside hydrolase 5 between positions 162–171. The nucleotide sequence of PfGH5 was deposited in the NCBI database with accession number OM160816.
To understand the properties of this PfGH5 protein and its role in biomass deconstruction, the protein was first recombinantly expressed using Pichia pastoris as a host. P. pastoris holds a great advantage as a system for the production of recombinant proteins because of its ability to produce highly efficient enzymes in high abundance as well as for fast growth characteristics in inexpensive media. In addition, its capability for appropriate folding and transportation of proteins renders it an excellent tool for recombinant protein production. P. pastoris also secretes fewer native proteins thereby enabling easier purification and biochemical characterization of recombinant proteins in supernatant. To express this protein, the alpha secretion signal peptide of the pPICZαA vector was used to direct the secretion of the protein under the AOX1 promoter. For this, the PfGH5 was fused with the α-secretion signal of the pPICZα vector to obtain the vector pPICZαA-PfGH5, designated as pOAO6. The recombinant pOAO6 vector was linearized with SacI restriction enzyme to enable its integration into the P. pastoris X-33 genome by homologous recombination. Linearized vector was then transformed into competent P. pastoris X-33 for expression and the transformants were selected with Zeocin selection marker (100 μg/ml) for 72 h. Some positive transformants were randomly picked up and re-cultivated on YPDA plates at 30 °C with different zeocin concentrations between 200 and 1000 μg/ml for screening multi-copy positive transformants. A strain with multi-copy integration of the gene that resisted 1000 μg/ml zeocin was further selected, confirmed by PCR (Supplementary Fig. S1) and then inoculated into 50 ml inducing media at 30 °C and 220 rpm for producing the recombinant PfGH5. The induction was monitored every 24 h by loading some of the culture supernatants on SDS-PAGE electrophoresis (Fig. 2a; Supplementary Fig. S2a) and the recombinant enzyme was finally recovered from Pichia pastoris culture supernatant at approximately 1.8 g/L protein after 96 h of induction. PfGH5 was purified to apparent homogeneity by a single step of ion-exchange chromatography using Q-Sepharose Fast Flow (QSFF) and the purified enzyme was then used for all subsequent biochemical characterization (Supplementary Fig. S2b). The purified enzyme showed an intense single band on SDS-PAGE with a molecular mass close to 55 kDa (Fig. 2b), which was slightly higher than the theoretical mass of 50 kDa. The difference in sizes suggested that post-translational modifications occurred in PfGH5 during heterologous expression in P. pastoris and these post-translational modifications had effects on the migration rate of the PfGH5 on SDS-PAGE. To gain insights into the action of this glycoside hydrolase in biomass deconstruction, the enzyme was screened on various polysaccharides, for its specificity on glycosidic bonds (Table 1). The PfGH5 exhibited the highest activity toward beechwood xylan with a specific activity of 0.7 U/mg protein followed by xyloglucan and lichenan. The results also showed that the PfGH5 could also hydrolyse laminarin, carboxymethyl cellulose and para-nitrophenyl-β-d-xylopyranoside (pNPX) although the efficiencies were much lower than with xylan while there was no activity detected on Avicel, β-glucan and para-nitrophenyl-β-d-glucopyranoside (pNPG) (Table 1). Recognition of a broad range of substrates for catalysis indicated the promiscuous nature of the enzyme. However, the relatively higher specific activity of this enzyme towards both xylan and xyloglucan when compared with other substrates screened suggests the possibility of this PfGH5 be mainly a xylanase. Therefore, to further understand the mechanism of action of the enzyme and the nature of the products it produces, the hydrolysis reactions were set up with 5% xylan as well as with 5% pretreated SCB as the substrate and 25 mg/g protein dosage at 50 °C for 24 h and 72 h, respectively. Analysis of the hydrolysate showed significant production of xylooligosaccharides (XOS) with a degree of polymerization (DP) up to six (DP6) from the xylan in the test reaction containing the enzyme, while there was no trace of XOS in control reactions without enzyme in the reaction mixture after 24 h of incubation (Fig. 2c,d). We observed significant liberation of xylobiose and xylotriose with a very less amount of xylose in the hydrolysate. This mode of action was found to be similar to that of an endoxylanase from Trichoderma asperellum, thus, suggesting that the PfGH5 is an endo-acting enzyme that randomly cleaves the β-linkages in xylan. Furthermore, the chromatogram of the hydrolysate for the SCB test condition showed the presence of some glucose released in addition to xylose and XOS by PfGH5 (Supplementary Fig. S3a,b). The detected glucose in the SCB hydrolysates could be from the breakdown of xyloglucan or other mixed-linked glucans that are an integral component of the hemicellulose complex in SCB, as the enzyme showed some activities towards pure xyloglucan as well as mixed-linked glucan substrates (Table 1). To further characterize the enzyme, the effect of pH on the PfGH5 xylanase activity was examined from pH 3.0 to 6.0. The results showed that the enzyme exhibited maximum activity at pH 4.5, while it had a relative activity of 76% and 60% at pH 4.0 and 5.0, respectively (Fig. 2e). The optimum pH activity observed with this enzyme is in contrast to earlier characterised GH5 xylanases mostly of bacterial origin which exhibited optimum activity between pH 5 and 6. The enzyme exhibited good stability at room temperature between pH 3.5–5.5 after 24 h of incubation (Fig. 2e). It was most stable at pH 4.5 and retained about 96% of its original activity after 24 h of incubation at this pH. It was remarkable that the enzyme had a residual activity of 90% at both pH 4.0 and 5.0 after 24 h of incubation (Fig. 2e). The high activity of this enzyme at acidic pH range and stability across a wider pH range indicates that this enzyme could be potentially relevant as most saccharification assays with fungal enzymes are mostly performed at acidic conditions between pH 4 and 5. Evaluation of the temperature optima of this enzyme showed that PfGH5 displayed maximal activity at 50 °C while it exhibited 84% and 73% of its activity at 40 °C and 60 °C, respectively (Fig. 2f). The thermostability of the enzyme was monitored by incubating enzyme at various elevated temperatures for 5 h and then measuring its residual activity. The enzyme was found most stable between 40 and 50 °C where it retained 85% of its original activity at 50 °C after 5 h of incubation (Fig. 2g). At more elevated temperatures (70 and 80 °C), the activity declined drastically (Fig. 2g). The half-lives of PfGH5 at 60, 70 and 80 °C were 192, 45 and 10 min, respectively (Table 2; Supplementary Fig. S4). Significant enzyme activity at elevated temperature is advantageous for industrial biomass processing, as it not only improves substrate solubility and decreases viscosity but also minimizes microbial contaminations.
Since most of the characterized xylanases from the GH10 and GH11 families have been reported as notable accessory enzymes which work synergistically with cellulases in biomass deconstruction, we next evaluated the impact of this new enzyme in hydrolysis of complex heterogeneous biomass together with the secretome from P. funiculosum (PfMig188). For this experiment, pretreated SCB was used as cellulosic substrate. The hydrolysis reaction was set-up at 5% solid loading and a total protein dosage of 10 mg/g dry biomass weight (DBW) that was modulated in different ratios by replacing a portion of the P. funiculosum cellulase cocktail with PfGH5 xylanase for 72 h at 50 °C as shown in Table 3. For the enzyme replacement approach, varying amounts of the PfMig188 cellulases were replaced with PfGH5 xylanase, while the total amount of enzyme added was kept constant on a protein basis at 10 mg/g DBW. Interestingly, the supplementation of the cellulase cocktail with varying amounts of PfGH5 xylanase increased the saccharification potential of P. funiculosum cellulase cocktail synergistically. The results showed substantial boosting effect in hydrolysis with the 80:20 and 90:10 enzyme combination ratio, while the 95:5 ratio showed maximum improvement and the highest degree of synergism (Table 3). The synergistic effect between this accessory enzyme and the cellulase cocktail could be due to the creation of more binding sites as a result of the removal of non-cellulosic polysaccharides by the xylanase for cellulase to act upon. It was noticed that the total amount of monomeric sugars released with the 95:5 ratio increased by ~ 32% (10.78 g/L) as compared to the 100:0 ratio without the addition of PfGH5 (8.15 g/L), thus indicating that this enzyme holds great biotechnological prospect in improving the saccharification performance of P. funiculosum at reduced cellulase loading.
Upon validation of the function of this GH5 protein to be a xylanase, it was further designated as PfXyn5. Also, from the promising boosting characteristics it exhibits in SCB saccharification when combined with cellulase cocktail, we next decided to overexpress this gene in the background of the Mig1-repressor-deleted strain of P. funiculosum, i.e., PfMig188, as it has been shown to produce a larger amount of cellulolytic enzymes in our earlier study. To overexpress the xylanase gene in PfMig188, the expression vector containing the endogenous gene along with its promoter and terminator was constructed using the backbone of pBIF binary vector as mentioned in the method section. The 3.6-kb region containing the Xyn5 gene (Fig. 3a) was amplified from the genome of the parent NCIM1228 strain and cloned into the pBIF vector to generate the pOAO7 plasmid. The pOAO7 recombinant plasmid was confirmed by restriction digestion (Supplementary Fig. S5a) before transformation into the PfMig188 strain using the agrobacterium-mediated transformation method (AMTM). The hygromycin-resistant transformants obtained (Fig. 3b) were subsequently analysed by performing PCR to check for the integration of the new expression cassette into the genome of the fungus (Fig. 3c). The amplification of the expected 5 kb fragment in three of the selected transformants (PfOAO6) and found to be absent in the control (PfMig188) confirmed the integration of the Xyn5 cassette into the genome of the fungus. To understand the effect of increasing the copy number of this enzyme on its expression, the transcript abundance of the transformants was next examined at the mRNA level through quantitative PCR (qPCR). For this, cultures of NCIM1228, PfMig188, and the three positive transformants of PfOAO6 were grown in Mandel’s media containing 25 g/L Avicel and 15 g/L SCB for 60 h to obtain good mycelial growth. The transcript levels of all five strains under derepressing conditions were determined by real-time PCR (RT-PCR) with tubulin as a control, and the relative fold change was normalized to the level in NCIM1228 since it was the original strain from which all other mutant strains (PfMig188 and PfOAO6) were generated (Fig. 3d). The results under the overexpression showed about a 20-fold increase in Xyn5 transcripts for PfMig188, while there were about 35 to 40-fold increases in transcript levels for the PfOAO6 strains. To evaluate the expression level of the xylanase gene in the PfOAO6 transformants, the three transformants with increased transcripts were cultivated in a cellulase-inducing medium (CIM) for 5 days. The transformants were then screened based on the xylanase production capability of their culture supernatant and the results were compared with those for the parent strain, PfMig188 (Fig. 3e). The results showed a significant increase in xylanase activity in the range of 11 to 27%, in all the transformants compared to the parent strain, with the transformant T3 showing maximum xylanase production (Fig. 3e). Furthermore, we investigated if the overexpression of this gene will have an impact on the β-xylosidase content in the secretome since it is the last enzyme in the xylan degradation system that converts xylobiose to xylose. We found a close to ~ 9% increase in the β-xylosidase activity in two of the transformants (T3 and T5) while the β-xylosidase activity of transformants T1 was the same as the control strain (PfMig188). It was observed that these differences could be due to the differences in the integration loci of the expression cassette in the PfMig188 genome when performing random integration as seen in our earlier study. As expected, we saw that there were no changes in the FPase activity of the secretomes in comparison with that of the parent strain (Fig. 3e), as the GH5 xylanase was found not to be active on a pure cellulosic substrate which is the main substrate in the standard filter paper assay reaction mixture (Table 1). Since there was a significant improvement in the xylanase content of the transformants’ secretome as seen both from the transcriptional and translational studies, we next evaluated its corresponding effect on the digestion of pretreated SCB (Fig. 3f–h). For this, an SCB hydrolysis reaction was set up with a 15% solid load and minimal protein loading of 2.5 FPU/g DBW at 50 °C for 72 h. A low dose of 2.5 FPU/g was considered to avoid complete saccharification of the SCB, thereby enabling easy identification of the differences in the monomeric sugars released if any in the hydrolysate. The results obtained after hydrolysis showed no difference in the total glucose released from the secretomes of the transformants when compared with the native strain (Fig. 3f), while there was a noticeable improvement in the xylose content after 72 h as expected. The xylose yield with the secretomes of transformants T3 and T5 increased to 17 g/L which corresponded to 94% hemicellulose conversion when compared with the control that yielded 13.6 g/L xylose which corresponded to 76% hemicellulose conversion (Fig. 3g). Overall, the improvement in hemicellulose conversion in the transformants resulted in 44% total holocellulose conversion (Fig. 3h) which was ~ 9% improvement over that of the parent strain. The results indicate that the gene was successfully overexpressed and the integration of the additional copy of this gene in the fungus genome indeed helped in further improving the saccharification efficiency of its secretome in SCB deconstruction.
Since the overexpression of the PfXyn5 gene in PfMig188 from the earlier section improved only the saccharification of the hemicellulose component of SCB with almost no effect on the cellulose conversion, we next focused on overexpressing another gene, a GH16 endo‑β‑1,3(4)‑glucanase (Egl16) with accession number OM483862, which was also significantly upregulated in the presence of SCB from earlier proteomics study. This enzyme is one of the β-glucanases typically active on complex polysaccharides containing mixed linkages such as β-(1,3)- and β-(1,4)-d-glucans in plant cell wall. To overexpress the Egl16 gene of P. funiculosum in PfMig188 strain, the Egl16 nucleotide which consists of 1887 bp with one intron and 546 amino acids was constructed using the backbone of the pBIF binary vector along with its native promoter and terminator (Fig. 4a). The gene was amplified from the genome of the parent strain and first cloned into a PCR cloning vector pJET1.2 to generate the pOAO8 plasmid. The PstI/XbaI fragment was excised and cloned into the same sites in pBIF to create the pOAO9 vector. The pOAO9 plasmid was then digested with MunI and MluI restriction enzymes to confirm the presence of the Egl16 gene in the expression cassette and absence in the pBIF vector (Supplementary Fig. S5b). The confirmed pOAO9 plasmid was then transformed into PfMig188 strain using the ATMT procedure, and the hygromycin-resistant transformants (PfOAO7) obtained were analyzed for gene integration and protein expression (Fig. 4b). PCR was performed to confirm the integration of the Egl16 cassette into the genome as shown in Fig. 4c. PCR reactions performed with the primers PgpdA-F and TrpC-R, which were designed based on vector sequence, generated the anticipated amplification products (~ 5 kb fragment) in three out of the five transformants selected, but not in the control PfMig188, which confirmed the integration of the Egl16 cassette into the genome of PfMig188 strain (Fig. 4c). For assessment of the expression level of the overexpressed Egl16 gene, cultures of NCIM1228, PfMig188, and the three positive transformants of PfOAO7 were grown in Mandel’s media containing 25 g/L Avicel and 15 g/L SCB as carbon sources for 60 h for mRNA isolation and RT-PCR. The transcript levels of all five strains under derepressing conditions were evaluated with tubulin as a control, and the relative fold change was normalized to the level in NCIM1228 (Fig. 4d). The results showed that the transcript level of Egl16 remained almost unchanged for PfMig188 despite the deletion of the carbon catabolite repressor gene in its genome while there were about threefold increases in transcript levels for two of the PfOAO7 strains. To evaluate the expression level of the endoglucanase gene in the PfOAO7 transformants, the three transformants obtained were cultivated in a cellulase-inducing medium (CIM) for 5 days and the supernatants of the harvested cultures were used for enzyme assays on mixed-linked glucan substrates. The results showed an enhancement in lichenase and laminarase activities in two of the transformants when compared with the parent strain while one of the transformants (T3) showed almost the same activity as that of the parent strain (Fig. 4e). About 26% and 18% increase in laminarase and lichenase activities were found with two of the transformants (T1 and T4) over that of PfMig188. The transformant T4 gave the maximum laminarase and lichenase activities of 9.5 U/ml and 65 U/ml, respectively, while PfMig188 yielded 7.5 U/ml and 54.8 U/ml for laminarase and lichenase, respectively. Evaluation of the impact of overexpressing this gene on the total cellulase activity showed that there were no changes in the FPase activity of the secretomes in comparison with that of the parent strain (Fig. 4e). This result was not surprising though as the overexpressed enzyme is an accessory enzyme that may show little or no activity on a pure cellulosic substrate such as Avicel. To evaluate the corresponding impact of overexpressing this gene on the quality of PfMig188 secretome towards SCB hydrolysis, a hydrolysis reaction was set-up with a 15% solid load of SCB and 2.5 FPU/g DBW at 50 °C for 72 h. Upon completion of the hydrolysis, the resulting hydrolysates were analysed accordingly and the results obtained showed a significant increase in the glucose content after 72 h for the PfOAO7 transformants while there was no difference in the total xylose released from the secretomes of the transformants when compared with the native strain (Fig. 4f), The results showed that the concentration of glucose released from SCB increased to 29.6 g/L and 28.5 g/L for transformants T1 and T4, respectively when compared with the control that yielded 26.1 g/L glucose. The approximate 13% increase in glucose yield with the PfOAO7 transformants led to 42.8% cellulose conversion in contrast to PfMig188 which was 37.7% cellulose conversion at this minimum enzyme dosage. Overall, the enhancement in saccharification performance of the PfOAO7 transformants resulted in 46% total holocellulose conversion (Fig. 4g) which was ~ 7% improvement over that of the parent strain (43%). These results, therefore, indicate that the introduction of an extra copy of the Egl16 gene into the genome of the PfMig188 strain enabled increased production of its corresponding enzyme for better hydrolysis on complex heterogeneous substrates such as SCB.
In the above sections, we found that individual overexpression of Pfxyn5 and PfEgl16 provided remarkable hydrolysis efficiency against SCB at low enzyme dosage. The results provided an insight that simultaneously overexpressing the two genes may further facilitate enhanced deconstruction of SCB over what was achieved when individually overexpressed. To engineer the fungal strain for the dual overexpression of both Xyn5 and Egl16, a systematic approach was utilized for the construction of the desired pOAO10 vector containing the Xyn5/Egl16 expression cassette (Fig. 5a), as described in “Materials and methods”. For fungal transformation, the pOAO10 plasmid was confirmed by restriction digestion before transformation into the PfMig188 strain (Supplementary Fig. S5c). The hygromycin-resistant transformants (PfOAO8) obtained (Fig. 5b) were analysed by PCR with the primers PgpdA-F and TrpC-R which were designed based on vector sequence. The results showed that all the transformants generated the anticipated amplification products (~ 7 kb fragment) which was absent in the native strain thereby confirming the integration of the Xyn5/Egl16 cassette into the genome of the fungus (Fig. 5c). A further evaluation of the transcript level of the overexpressed genes (Xyn5 and Egl16) showed increased xylanase abundance with the PfOAO8 transformants over that of PfMig188 when grown in Mandel’s media containing 25 g/L Avicel and 15 g/L SCB as carbon sources (Fig. 5d). Transformants T1 showed the highest increased Xyn5 transcript which was twofold higher than that of PfMig188 while transformants T3 and T5 both showed a 1.8-fold increment in Xyn5 transcripts (Fig. 5d). For Egl16, the results obtained showed about three to ninefold increases in its transcript level after 60 h with transformants T5 having the highest Egl16 level (Fig. 5d). To evaluate the expression level of the two overexpressed genes at the translational level, the positive transformants of PfOAO8 were cultivated in a cellulase-inducing medium (CIM) for 5 days and the supernatants of the harvested cultures were used for enzyme assays as stated in “Materials and methods”. Analysis of the results showed significant improvement in enzyme activities across all the transformants when compared with the parent strain which was also in accordance with the mRNA levels (Fig. 5e). The PfOAO8 transformants showed ~ 17% and ~ 7% increase in xylanase and β-xylosidase activities, respectively, compared to PfMig188. Similarly, we found about 43% and 37% increases in lichenase and laminarase activities, respectively when compared to PfMig188 while the FPase activities in all the transformants remained unchanged just as it was observed for the PfOAO6 and PfOAO7 strains. For the evaluation of the corresponding impact of simultaneously overexpressing these two genes on the quality of PfMig188 secretome towards SCB hydrolysis, another set of hydrolysis reactions was set up with a 15% solid load of SCB and 2.5 FPU/g DBW at 50 °C for 72 h. Upon completion of the hydrolysis, the resulting hydrolysates were analysed accordingly and the results obtained showed a significant increase in both the glucose and xylose contents of the hydrolysates for the PfOAO8 transformants when compared with the native strain (Fig. 5f). The results showed between 11 and 17% increase in the concentration of glucose released for the PfOAO8 transformants when compared with the control. Furthermore, the concentration of xylose in the hydrolysates of the PfOAO8 transformants increased by 11%. The saccharification efficiency (holocellulose conversion) for PfOAO8 strains when analysed was found to be considerably higher than the individual overexpressing strains (PfOAO6 and PfOAO7) as they exhibited between 45 and 49% holocellulose conversion in contrast to the native strain (PfMig188) where ~ 40% holocellulose conversion was achieved (Fig. 5g). These results demonstrated that the Xyn5–Egl16 double overexpression provided a highly significant increase in SCB saccharification efficiency when compared to the individual ones and the engineered strains developed could potentially be used as promising bioresources needed for the production of more balanced cellulase and accessory enzyme cocktails required for the low-cost production of lignocellulose-based biofuels. In conclusion, an unknown glycosyl hydrolase 5 xylanase from the hypercellulolytic P. funiculosum was recombinantly expressed and characterized and its mechanism of action in the deconstruction of lignocellulose especially SCB was further deciphered in this study. The notable biochemical properties of PfXyn5 suggest it is a suitable candidate for various industrial applications most especially in bioenergy. Using a combined application of various fungal genetic tools, a new fungal strain was also created for simultaneous overexpression of notable accessory enzymes whose secretome demonstrated enhanced saccharification performance on pretreated SCB with minimal protein loading.
All the strains and plasmids used in this study are listed in Table 4, while all the primer sets used in the study are itemized in Table 5. The nucleotide and protein sequences for Xyn5 and Egl16 for the study have been deposited in the NCBI database with accession numbers OM160816 and OM483862, respectively. Escherichia coli DH5α was used for plasmid propagation throughout the experiments. The Agrobacterium tumefaciens LBA4404 strain used for fungal transformation was maintained on a low-sodium LB medium (10 g/L tryptone, 5 g/L yeast extract, 5 g/L sodium chloride) containing 100 µg/ml kanamycin and 30 µg/ml rifampicin. The pBIF vector which was used as the backbone vector for fungal transformation, contains hygromycin and kanamycin resistance genes as selective markers for the selection of transformants (35). The pJET vector containing the ampicillin resistance gene was used as a shuttle vector for the sub-cloning of the Egl16 gene before re-cloning into pBIF for overexpression. P. funiculosum NCIM1228 and its derivative P. funiculosum Mig188 (PfMig188), the fungal strains used for this study, were routinely cultivated on Petri dishes containing low-malt extract–peptone (LMP) agar for approximately 14 days until full sporulation. For the selection and maintenance of fungal transformants, an LMP medium supplemented with hygromycin B at 100 µg/ml was used.
NaOH-treated Sugarcane bagasse was used as the cellulosic substrate in this study. The SCB was kindly provided by Natems Sugars Private Limited, Hyderabad, Telangana, India. SCB was washed, dried in a convection oven at 60 °C to constant weight, and then knife-milled using a 1 mm sieve in the mill. The alkaline pretreatment of SCB was done at 90 °C using 0.4 M NaOH at 10% solid loading for 6 h. After pretreatment, the solids were washed three times was then repeatedly washed until the pH became neutral. Compositional analysis conducted according to National Renewable Energy Laboratory (NREL) procedure TP510-42618 on the pretreated sugarcane bagasse yielded a cellulose content of 46.1%, a hemicellulose content of 24.5% and a lignin content of 17%.
The Xyn5 sequence was retrieved from the draft genome sequence of P. funiculosum NCIM1228 available in our laboratory. The amino acid sequences of some closely similar glycoside hydrolase 5 proteins from other species were retrieved from the NCBI database, and molecular phylogenetic analysis by the maximum likelihood method and the Tamura-Nei model was conducted using MEGA X software. Multiple-sequence alignment of PfXyn5 with other GH5 proteins obtained from the phylogenetic analysis was performed using the Clustal Omega multiple-sequence alignment program. The signal peptide and conserved domains were analyzed at Signal P 5.0 server (http://www.cbs. dtu.dk/services/SignalP/) and ScanProsite (http://www.expasy.ch/tools/ScanProsite), respectively. N-Glycosylation and O-glycosylation sites were predicted using NetNGlyc 1.0 (http://www.cbs.dtu.dk/services/NetNGlyc/) and NetOGlyc 4.0 server (http://www.cbs.dtu.dk/services/NetOGlyc/), respectively.
The xylanase sequence used in the present study from P. funiculosum was synthesized by GenScript USA, Inc. (New Jersey, USA) with codon optimization for expression in P. pastoris. The synthesized sequence was ligated into the vector pPICZαA, named pPICZαA-PfGH5 (or pOAO6). The pOAO6 vector was transformed into Escherichia coli (DH5α), and subsequently, the positive clone was screened, and the plasmid was extracted using Qiagen mini-prep kit. The pOAO6 plasmid was linearized using SacI restriction endonuclease and transformed into P. pastoris X-33 competent cells using the lithium chloride method described by Kumar et al.. The transformation liquid was plated on a YPD plate containing 100 µg/ml zeocin, and the positive transformants were further reselected on a higher concentration of zeocin (1000 µg/ml) for 48 h before being screened using polymerase chain reaction (PCR) with the AOXI-F and AOX1-R primer pairs (Table 5). The selected positive transformants were subsequently inoculated into buffered glycerol-complex medium (BMGY) for 24 h before transferring into methanol-complex medium (BMMY) for xylanase expression with the addition of 1% methanol every 12 h till the end of the fermentation process. The induction of the expressed protein was monitored by collecting samples from the cultures every 24 h and loaded on sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). To purify the protein, the culture supernatants were recovered by pelleting the cells by centrifugation at 4000 rpm for 5 min, 4 °C and filtered on 0.45-µm filters (Millipore, Molsheim, France) to remove any remaining cells. The recovered supernatant was dialyzed against 50 mM sodium phosphate buffer, pH 7.5 overnight with two equal changes of buffer at 12 h intervals until the pH becomes 7.5. After adjusting the pH to 7.5, the supernatants were filtered once more on 0.2-µm filters and loaded onto a FPLC HiTrap Q Sepharose XL 5 ml column (GE Healthcare) that was equilibrated with 50 mM sodium phosphate buffer (pH 7.5). Proteins were eluted using a linear gradient of NaCl (25–1000 mM) in the same buffer mentioned above at flow rate of 2.0 ml/min. Fractions containing the enzyme activity were pooled and subjected to SDS-PAGE to check protein purity. Protein concentrations were determined using the bicinchoninic acid (BCA) method using bovine serum albumin (BSA) as standard.
To characterize the recombinant PfXyn5, the protein was screened on various polysaccharides which include Avicel, beechwood xylan, xyloglucan, laminarin, sodium carboxymethyl cellulose, lichenan, β-glucan, p-nitrophenyl-β-d-glucopyranoside (pNPG) and p-nitrophenyl-β-d-xylopyranoside (pNPX). The substrates were incubated with the appropriate amount of the enzyme at 50 °C for 30 min and the amount of reducing sugar liberated was quantitated using the DNS method where appropriate. For determination of the mechanism of action of the enzyme, hydrolysis reactions were set up using 5% xylan and 5% pretreated SCB as substrates and protein dosage of 25 mg/g DBW at 50 °C for 24 h and 72 h, respectively. The resulting hydrolysates obtained were analysed using an Agilent 1260 series HPLC instrument coupled with a 1290 Infinity evaporative light scattering detector (ELSD). The degradation products were estimated using a Rezex RSO-oligosaccharide Ag+ (4%) column (200 by 10 mm) with a Rezex RSO-oligosaccharide Ag+ (4%) guard column (60 by 10 mm) (Phenomenex), with a mobile phase of 100% water. Throughout the analysis, the column was kept at 80 °C, and the mobile phase flow rate was maintained at 0.3 ml/min under isocratic conditions for 60 min. The chromatographic run was initiated by first equilibrating the column for 5 min followed by injecting 20 µl of the sample solution into the column. The ELSD was maintained at 45 °C throughout. The nebulizer (nitrogen) gas pressure was set at 2.8 × 105 Pa, and the detector gain was set at 9 × 105 Pa. The concentration of different XOS generated was calculated from calibration curves of external standards purchased from Megazyme. The pH dependence of PfXyn5 was determined by using 50 mM citrate–phosphate buffer (pH 3.0–6.0) containing purified enzyme and xylan at 50 °C as described above. To determine the effect of pH on the stability of PfXyn5, the enzyme was incubated in relevant buffers of varying pH (3.0–6.0) without substrate for 24 h at room temperature. The residual xylanase activity was determined according to the standard assay procedure earlier described. The temperature dependence was determined by incubating the reaction with the optimal pH at 40–80 °C. To determine the dependence on pH and temperature, the maximum activity obtained was defined as 100%. The thermal stability of PfXyn5 was analysed by preincubating it at 40–80 °C for 5 h and its residual activity was then determined.
The effect of supplementing the cellulase cocktail present in the secretome of P. funiculosum with PfXyn5 in SCB saccharification was determined by setting up hydrolysis reactions with pretreated 5% solid loading and a total protein dosage of 10 mg/g DBW at 50 °C for 72 h with constant shaking. The 10 mg/g protein concentration was maintained at different combination ratios of 60:40, 70:30, 80:20, 90:10, 95:05 and 100:0 for P. funiculosum to PfXyn5. After hydrolysis the samples were centrifuged to separate the solid residue from the digested biomass. Supernatants that were recovered after enzymatic hydrolysis of the pretreated SCB were analysed by high-performance liquid chromatography with an Aminex HPX-87H anion-exchange column (Bio-Rad, USA) and a refractive index (RI) detector to analyse the released monosaccharides (glucose and xylose) by anion-exchange chromatography. The concentration of each monosaccharide was calculated from calibration curves of external standards (xylose and glucose) purchased from Absolute Standards Inc. The degree of synergism between the PfXyn5 and cellulase was calculated following the method described by Goncalves et al. and Li et al. with the formula:where α and β correspond to the mass ratio of the enzymes in each reaction. Y1+2, indicates the yield of reducing sugar achieved by the two enzymes working simultaneously, and Y1 and Y2 indicate the yields of reducing sugar, achieved by each enzyme working separately.
All vectors for fungal expression described in this study were constructed on the backbone of the pBIF vector that was previously constructed using the binary vector pCAMBIA1300 as a backbone. Binary vectors for the overexpression of the Xyn5 and Egl16 genes from P. funiculosum were constructed as follows. The endogenous gene encoding Xyn5 was amplified from the genome of P. funiculosum NCIM1228 using the primers PfXyn5-F and PfXyn5-R (Table 5). The primers were designed according to the Xyn5 sequence containing both its native promoter and terminator, which was obtained from the draft genome sequence of the strain. The PCR product obtained was digested with the restriction enzymes PstI and BamHI before being ligated into the pBIF vector previously digested with the same enzyme sets to generate the pBIF-Xyn5 (pOAO7) vector. For the overexpression of Egl16, the Egl16 gene was amplified from the genome of P. funiculosum NCIM1228 using the primers PfEgl16-F1 and PfEgl16-R1 spanning its native promoter and terminator (Table 5). The PCR product was first subcloned into pJET1.2 to obtain the pOAO8 vector. The PstI/XbaI fragment containing the Egl16 fragment was excised and cloned into the same sites in pBIF to create the pBIF-Egl16 vector (pOAO9) vector. The ligated products of pOAO7 and pOAO9 were then transformed into E. coli DH5α cells and selected on 50 µg/ml kanamycin. The resulting colonies were screened for positive transformants by colony PCR, followed by restriction digestion of the corresponding plasmids. To create a strain for the combined overexpression of both the Xyn5 and Egl16 genes, the pOAO7 vector was used as the base vector upon which the Egl16 gene was cloned. The Egl16 gene with its native promoter and terminator was re-amplified from the genome of P. funiculosum NCIM1228 using the primers PfEgl16-F2 and PfEgl16-R2 containing the ApaI and BamHI restriction sites, respectively. The PCR product obtained was digested with the ApaI and BamHI restriction enzymes before being ligated into the corresponding sites of the pOAO7 vector to obtain the pBIF-Xyn5/Egl16 (pOAO10) vector. The verified pOAO7, pOAO9, and pOAO10 plasmids were transformed into the PfMig188 fungal strain according to the agrobacterium-mediated transformation method (AMTM) as previously described. After transformation, the resulting hygromycin-resistant transformants were screened according to the method described previously by Fang and Xia. The transformants were verified by PCR for the integration of the Xyn5 and Egl16 gene expression cassettes. For PCR analysis, the primers PgpdA-F and TrpC-R were used to screen for Xyn5 and Egl16 integration, while transformants of Xyn5/Egl16 were screened using the primer set PgpdA IR-F and TrpC-R (Table 5).
For real-time PCR experiments, cultures of all strains were grown in minimal Mandel’s medium containing 25 g/L Avicel and 15 g/L SCB for 60 h. Mycelia were harvested by filtration and frozen in liquid nitrogen. RNA was extracted using an RNeasy kit (Qiagen) according to the manufacturer’s instructions. RNA was treated with DNase (Invitrogen) before cDNA synthesis. One microgram of RNA was used as the template for each quantitative real-time PCR (qRT-PCR). A cDNA synthesis control was performed to ensure the absence of DNA contamination. qRT-PCR was carried out using iTaq universal SYBR green supermix (Bio-Rad) and a Bio-Rad CFX96 qPCR detection system. Primers for transcripts to be tested were designed using the boundary sequence of two exons to avoid any amplification from contaminant genomic DNA. qRT-PCR was performed in biological triplicates with tubulin as the endogenous control. Relative expression levels were normalized to the level of tubulin, and the fold changes in RNA levels were calculated as the ratios of the relative expression levels in PfMig188 and the corresponding transformants of Xyn5 and Egl16 to that in NCIM1228 under cellulase-inducing conditions.
Penicillium funiculosum NCIM1228, PfMig188, and the resulting PfOAO6, PfOAO7, and PfOAO8 transformants were cultivated on Petri dishes containing low-malt extract agar until full sporulation. After 14 days of incubation, spores were recovered with sterile water, filtered through sterile Mira cloth, and quantified using a hemocytometer. The primary culture of each strain was prepared by culturing 107 conidiophores in potato dextrose broth (PDB) for 36 h. Primary cultures of the strains were added to a cellulase-inducing medium in Erlenmeyer flasks at a final concentration of 10%. Cellulase-inducing medium (CIM) contained soya peptone (24 g/L), wheat bran (14 g/L), microcrystalline cellulose (MCC) (16 g/L), pretreated SCB (15 g/L), KH2PO4 (12.4 g/L), K2HPO4 (2.68 g/L), (NH4)2SO4 (0.28 g/L), CaCO3 (2.5 g/L), corn steep liquor (1%), urea (0.52 g/L), and yeast extract (0.05 g/L), with the final pH adjusted to 5.0. The flasks were kept at 28 °C for 5 days with orbital shaking at 150 rpm (Innova 44; Eppendorf AG, Germany). Induced cultures were centrifuged at 9000 rpm for 10 min at 4 °C, and the cellulolytic supernatants were collected and stored at 4 °C until use.
All enzymatic activities performed in this study were routinely determined following standard assay procedures. Endoglucanase, xylanase, lichenase, laminarase, β-glucosidase and β-xylosidase activities were determined by incubating appropriate dilution of the enzyme with 2% CMC (Sigma), 2% beechwood xylan (HiMedia), 1% lichenan (Megazyme), 1% laminarin (Sigma) and p-nitrophenyl-β-d-xylopyranoside (Megazyme), respectively, for 30 min, after which the amount of reducing sugars released was measured as previously reported. One unit of CMCase, xylanase, lichenase and laminarase activity is defined as the amount of enzyme releasing 1 µmol of reducing sugar per min while one unit of β-xylosidase activities was defined as the amount of protein that released 1 μmol of p-nitrophenol (pNP) per min. Total cellulase activity in the secretome was measured in terms of filter paper units (FPU) per milliliter of original (undiluted) enzyme solution. The assay requires a fixed degree of conversion of the substrate, from 50 mg of filter paper within 60 min at 50 °C. One FPU is defined as the amount of enzyme required to produce 2 mg of glucose from 50 mg of filter paper within 60 min of incubation. To quantitate the amount of secreted proteins in the various PfMig188 secretomes as well as its derivatives, the crude secretomes were first buffer exchanged with the help of a 10-kDa cut-off membrane using citrate– phosphate buffer (pH 4.0) and then the total protein of each secretome was estimated by the bicinchoninic acid (BCA) method using bovine serum albumin (BSA) as standard.
The saccharification efficiency of the secretomes of all the strains used in the study with pretreated SCB was determined according to the method described previously by Ogunyewo et al., with some modifications. The performance of the secretomes toward NaOH-treated SCB was evaluated at 15% dry weight of biomass using an enzyme concentration of 2.5 FPU/g DBW. Saccharification was performed in 50 ml screw-cap Falcon tubes in an incubator shaker at 50 °C for 72 h. The reaction mixture included NaOH-treated SCB under 15% dry weight loading in a 5 ml final reaction volume. The total cellulase activities of each secretome of all the fungal strains tested were measured, and the appropriate volume of the desired protein concentration (2.5 FPU/g DBW) was added to the reaction mixture. The reactions were set up in 50 mM citrate phosphate buffer (pH 4.0), and the mixtures were incubated at 50 °C with constant shaking at 220 rpm for 72 h. Samples were collected at the end of the 72 h incubation period and analysed for the production of fermentable sugars. Control experiments were carried out under the same conditions, using substrates without enzymes (enzyme blank) and enzymes without substrates (substrate blank); a substrate-free negative control was set up by filling the Falcon tubes with 50 mM citrate phosphate buffer (pH 4.0), and the background of soluble sugars present in the respective biomass was determined by incubating each biomass in the absence of the enzyme. Following the completion of hydrolysis at each time point, the Falcon tubes were centrifuged at 3500 rpm for 10 min in a swinging-bucket centrifuge (Eppendorf, Germany) to separate the solid residue from the digested biomass. Supernatants that were recovered after enzymatic hydrolysis of the pretreated SCB were analysed by high-performance liquid chromatography with an Aminex HPX-87H anion-exchange column (Bio-Rad, USA) and a refractive index (RI) detector to analysed the released monosaccharides (glucose and xylose) by anion-exchange chromatography. The filtered mobile phase (4 mM H2SO4) was used at a constant rate of 0.3 ml/min with the column, and the RI detector temperature was maintained at 35 °C. The concentration of each monosaccharide was calculated from calibration curves of external standards (xylose and glucose) purchased from Absolute Standards Inc. The theoretical conversions of cellulose and hemicellulose (in percentage) into monomeric sugars were calculated using the equations provided in NREL’s LAP TP-510-43630 as earlier reported.
All experiments were performed in triplicate, and the results are presented as the means and standard deviations. The data were compiled in a Microsoft Excel spreadsheet, where the averages and standard errors of the means were determined. All graphs were created using GraphPad Prism 8.0 software. The data were further evaluated by one-way analysis of variance (ANOVA) and multiple t-tests using GraphPad Prism 8.0 software where appropriate.
Supplementary Figures. | true | true | true |
|
PMC9568537 | Wei Zhou,Wen-hui Wu,Zi-lin Si,Hui-ling Liu,Hanyu Wang,Hong Jiang,Ya-fang Liu,Raphael N. Alolga,Cheng Chen,Shi-jia Liu,Xue-yan Bian,Jin-jun Shan,Jing Li,Ning-hua Tan,Zhi-hao Zhang | The gut microbe Bacteroides fragilis ameliorates renal fibrosis in mice | 14-10-2022 | Microbiome,Renal fibrosis | Renal fibrosis is an inevitable outcome of various manifestations of progressive chronic kidney diseases (CKD). The need for efficacious treatment regimen against renal fibrosis can therefore not be overemphasized. Here we show a novel protective role of Bacteroides fragilis (B. fragilis) in renal fibrosis in mice. We demonstrate decreased abundance of B. fragilis in the feces of CKD patients and unilateral ureteral obstruction (UUO) mice. Oral administration of live B. fragilis attenuates renal fibrosis in UUO and adenine mice models. Increased lipopolysaccharide (LPS) levels are decreased after B. fragilis administration. Results of metabolomics and proteomics studies show decreased level of 1,5-anhydroglucitol (1,5-AG), a substrate of SGLT2, which increases after B. fragilis administration via enhancement of renal SGLT2 expression. 1,5-AG is an agonist of TGR5 that attenuates renal fibrosis by inhibiting oxidative stress and inflammation. Madecassoside, a natural product found via in vitro screening promotes B. fragilis growth and remarkably ameliorates renal fibrosis. Our findings reveal the ameliorative role of B. fragilis in renal fibrosis via decreasing LPS and increasing 1,5-AG levels. | The gut microbe Bacteroides fragilis ameliorates renal fibrosis in mice
Renal fibrosis is an inevitable outcome of various manifestations of progressive chronic kidney diseases (CKD). The need for efficacious treatment regimen against renal fibrosis can therefore not be overemphasized. Here we show a novel protective role of Bacteroides fragilis (B. fragilis) in renal fibrosis in mice. We demonstrate decreased abundance of B. fragilis in the feces of CKD patients and unilateral ureteral obstruction (UUO) mice. Oral administration of live B. fragilis attenuates renal fibrosis in UUO and adenine mice models. Increased lipopolysaccharide (LPS) levels are decreased after B. fragilis administration. Results of metabolomics and proteomics studies show decreased level of 1,5-anhydroglucitol (1,5-AG), a substrate of SGLT2, which increases after B. fragilis administration via enhancement of renal SGLT2 expression. 1,5-AG is an agonist of TGR5 that attenuates renal fibrosis by inhibiting oxidative stress and inflammation. Madecassoside, a natural product found via in vitro screening promotes B. fragilis growth and remarkably ameliorates renal fibrosis. Our findings reveal the ameliorative role of B. fragilis in renal fibrosis via decreasing LPS and increasing 1,5-AG levels.
It is estimated that approximately 10% of the general population suffers from chronic kidney disease (CKD). Renal fibrosis that inevitably results in CKD progression is characterized by the proliferation of fibroblasts and myofibroblasts. Myofibroblasts in turn are characterized by the production of alpha-smooth muscle actin (α-SMA) fibers, collagen and extracellular matrix (ECM) proteins. Incessant ECM production leads to decreased glomerular filtration rate and renal injury. A total halt in disease progression or perhaps induction of renal fibrosis regression can alleviate CKD. Currently, the management of CKD involves the use of antihypertensive drugs, such as angiotensin-converting–enzyme inhibitors and angiotensin-receptor blockers, and the adoption of other blood pressure (BP)-controlling measures, including reductions in protein and salt intake. Keeping the BP and even blood glucose levels in check could avoid the incidence of acute kidney injury. Aside from dialysis and surgery (i.e., kidney transplantation), there seems to be no available effective treatment for renal fibrosis and end-stage kidney disease. This situation therefore calls for better and effective alternatives in the management and treatment of CKD. The association of the gut microbiota with renal diseases has recently gained traction. The gut microbiota composition comprises nearly 1 trillion microbes with diverse genetic make-up. A healthy gut microbiota regulates physiological homeostasis; dysbiosis therefore could be an avenue for the promotion of chronic diseases such as kidney disease. There is a crosstalk between the kidneys and the gut in terms of their relationship with the gastrointestinal (GI) environment, gut epithelial barrier permeability and diseased state (such as CKD). There is a unique but reciprocal relationship between CKD and the gut microbiota. CKD incidence alters gut microbiota makeup and functions, eventually leading to dysbiosis. On the other hand, the gut microbiota regulates various processes that result in CKD onset and progression. These include the accumulation of uremic toxins, such as indoxyl sulfate (IS), p-cresyl sulfate (p-CS) and trimethylamine n-oxidase (TMAO). The uremic toxins are produced by the increased abundance of Enterobacteriaceae, Clostridiaceae, Pseudomonadaceae, and Bacteroidiaceae, and the decreased levels of Lactobacillaceae, Bifidobacteriaceae, and Prevotellaceae. Chronic inflammation and interruption of intestinal barrier function via increased ammonium production also enhanced the production of uremic toxins. Besides, the reductions in CKD-protective metabolites derived from microbiota-like indolepropionic acid (IPA), short-chain fatty acids (SCFAs) produced by Clostridium and Eubacterium and the neurotransmitters (γ-aminobutyric acid and acetylcholine) produced by Lactobacillaceae, Prevotellaceae and Bifidobacteriaceae could accelerate the progression of CKD. Thus, prevention of dysbiosis and restoration of homeostasis could be a potential strategy for the prevention and management of CKD. In this study, we found that the abundance of Bacteroides fragilis (B. fragilis) in the fecal samples of CKD patients and UUO mice was significantly decreased. B. fragilis is an anaerobic obligate gram-negative commensal that resides in the lower gut of mammals and is known to influence their susceptibility to inflammatory diseases. B. fragilis strains have been shown to inhibit inflammation in different organs, including the peritoneum (such as intra-abdominal abscess), intestinal tract (intestinal bowel diseases), brain (autism spectrum disorder) and lungs (asthma), by producing polysaccharide A (PSA) and SCFA to stimulate interleukin-10-producing CD4+Foxp3+ T-regulatory cells. However, there have been no reports on the role of B. fragilis in CKD. Therefore, both unilateral ureteral obstruction (UUO) and adenine mice models were first used to evaluate the effects of oral administration of B. fragilis on CKD. Then, untargeted and targeted metabolomics, together with label-free quantitative proteomics, were used to elucidate the mechanisms that underlie its renoprotective function. We have found that B. fragilis alleviates renal fibrosis by decreasing LPS levels. It also inhibits the Nrf2/Keap1 and TGF-β/Smad signaling pathways owing to increased levels of 1,5-anhydroglucitol (1,5-AG) in the blood. 1,5-AG was found to be a substrate of sodium-glucose cotransporter 2 (SGLT2). B. fragilis restored the reduced expression of SGLT2 in the kidneys of UUO and adenine models. We found 1,5-AG to be a TGR5 agonist, and knockdown of TGR5 abolished the anti-fibrotic effect of 1,5-AG and Nrf2/Keap1 activation in vitro. Finally, a natural product, madecassoside (Mad), was found to significantly promote B. fragilis growth and remarkably relieve renal fibrosis. Thus, regulating B. fragilis abundance in the gut might be a strategy to treat CKD.
We recruited 10 CKD patients and 10 age- and sex-matched healthy controls from Renmin Hospital of Wuhan University as microbiota discovery set. We performed a detailed and comparative gut microbial profiling using 16 S rDNA bacteria gene sequencing of their fecal samples. We found that the relative abundance of the phyla (Bacteroidetes, Firmicutes and Bacteroidetes/Firmicutes ratio) and genera (Bacteroides, Streptococcus, Akkermansia, Bifidobacterium, Fecalibacterium etc.) did not differ significantly (by false discovery rate adjustment) between the 2 groups (Supplementary Fig. 1A–1F). However, we found significantly decreased relative abundance of B. fragilis in patients with CKD by qPCR (Fig. 1A). We recruited another set of CKD patients (15 in total) and 15 age- and sex-matched healthy controls from the Putuo People’s Hospital as microbiota validation set. The abundance of B. fragilis in patients with CKD was also significantly decreased (Fig. 1B). Moreover, the abundance of B. fragilis had a significant negative correlation with BUN and Scr (Fig. 1C). Therefore, we focused on the B. fragilis to investigate the relationship between this specie and CKD and to clarify any underlying mechanisms.
To determine the effect of B. fragilis on the progression of CKD, UUO mice were treated with live B. fragilis by daily oral gavage for 2 weeks. We examined the abundance of B. fragilis in the sham, UUO-, HF-, and BF-treated mice by qPCR (Supplementary Fig. 2). The results indicated that treatment with B. fragilis could restore the decreased B. fragilis level in the fecal sample from the UUO mice. Oral administration of live, but not heat-killed, B. fragilis significantly improved renal morphology and reduced renal index of the UUO mice (Fig. 1D, E). Treatment with B. fragilis significantly attenuated the increase in BUN, Scr, triglyceride and total cholesterol levels, suggesting an improved renal function in the UUO mice (Fig. 1F). Immunofluorescence showed a visibly decreased expression of vimentin protein and a visibly increased expression of E-Cadherin in the UUO mice treated with B. fragilis (Fig. 1G, H). The administration of B. fragilis greatly reduced the levels of several pro‑fibrotic markers, including pro-fibrotic proteins collagen I, fibronectin and alpha-smooth muscle actin (α-SMA) (Fig. 1I, J). Haematoxylin and eosin (H&E) and Masson staining showed that tubular dilatation, tubular atrophy, and widening of the interstitial space with severe inflammatory cell infiltration were attenuated by administration of B. fragilis (Fig. 1K, L). Used together, these findings indicate that the replenishment of live B. fragilis by oral gavage exerts strong anti-fibrotic effects in UUO mice.
The fecal and serum levels of LPS were assessed in UUO mice and CKD patients. The results indicated that fecal and serum LPS levels were higher in CKD patients and UUO-mice compared to the controls (Supplementary Fig. 3A–D). Moreover, we found that oral administration of live B. fragilis dramatically decreased the fecal and serum LPS concentrations in UUO mice (Supplementary Fig. 3C, D). B. fragilis treated mice had significantly lower fecal and serum levels of pro-inflammatory cytokines including interleukin (IL)−1β, IL-6 and tumor necrosis factor-α (TNF-α) (Supplementary Fig. 3E, F). The mRNA expressions of these pro-inflammatory cytokines were also decreased in mice treated with live B. fragilis (Supplementary Fig. 3G). These results showed that B. fragilis attenuated inflammation in UUO mice.
The TGF-β/Smad pathway plays a key role in renal fibrosis. The progression of CKD has been shown to result in inflammation and oxidative stress. Thus, we evaluated the effect of B. fragilis on oxidative stress and TGF-β/Smad signaling pathway in UUO mice treated with B. fragilis. The expressions of TGF-β, Smad2, Smad3 in the UUO mice were all significantly increased compared to controls. Treatment with B. fragilis mediated significant attenuation of UUO-induced increase in TGF-β, Smad2 and Smad3 expressions (Fig. 2A, B). The mRNA levels of TGF-β, Smad2 and Smad3 exhibited similar patterns (Supplementary Fig. 4). Furthermore, we found that UUO caused significant increases in Keap1 and ROS-generating molecules including 12-Lox and Rac1. However, the expression of antioxidant protein (Nrf2) was significantly downregulated in the UUO mice (Fig. 2C, D). After treatment with B. fragilis, the upregulation of Keap1, 12-Lox and Rac1 was significantly reduced, while the downregulation of Nrf2 was significantly enhanced (Fig. 2C, D). In addition, the mRNA levels of Keap1, Rac1, p67 (Ncf2) and p47 (Nsfl1c) exhibited similar patterns (Supplementary Fig. 5A). These results indicate that B. fragilis can inhibit oxidative stress and the TGF-β/Smad signaling pathway in the UUO mice.
The profound influence of gut microbiota on the host is strongly associated with complex interactions comprising a series of host-microbe metabolic axes. To assess metabolic alterations in response to the treatment with B. fragilis, metabolome profiles were generated from serum samples by untargeted metabolomics using gas chromatography–mass spectrometry (GC-MS). A supervised OPLS-DA was performed using data from sham, UUO and UUO + BF group. OPLS-DA score plots could readily be divided into several clusters (Fig. 2E), indicating that the metabolic states of all UUO mice were significantly changed in relation to the sham mice. Additionally, a clear distinction in the metabolomes between UUO group and UUO + BF group was observed (Fig. 2E), suggesting the metabolic states of UUO + BF group were significantly different from the UUO group. An unsupervised PCA was also performed using the aforementioned data and the PCA plots exhibited patterns similar to the OPLS-DA plots (Supplementary Fig. 6A). Fourteen differentially expressed metabolites between the sham and UUO groups were identified by variable importance in the projection (VIP) ≥ 1.0 and adjusted p values ≤ 0.05 (Supplementary Table 1). A systematic pathway analysis based on these 14 differential metabolites is shown in Fig. 2F. Then, a heatmap was created to visualize the relative levels of the 14 differential metabolites in each group (Fig. 2G). Notably, of these 14 identified metabolites, 1,5-AG has been reported as a reliable glycaemic marker in type 2 diabetics with CKD. Moreover, low 1,5-AG levels are associated with higher risk of incident ESRD independent of baseline kidney function. In this study, the serum levels of 1,5-AG were significantly decreased in the UUO mice, while treatment with B. fragilis significantly upregulated the decreased level of 1,5-AG using GC-MS based targeted metabolomics (Fig. 2H). The results indicate that treatment with B. fragilis can significantly upregulate the serum level of 1,5-AG in UUO mice.
We conducted GC-MS based untargeted metabolomics to evaluate the inherent differential serum metabolic profiles between healthy controls and CKD patients (115 patients and 113 healthy subjects from the Affiliated Hospital of Nanjing University of Chinese Medicine). The OPLS-DA results indicated significant alteration in metabolic profiles between the healthy controls and CKD patients (Fig. 2I). An unsupervised PCA was also performed on the same data and the PCA plots exhibited patterns similar to the OPLS-DA plots (Supplementary Fig. 6B). Differential metabolites were identified and summarized in Supplementary Table 2. A systematic pathway analysis based on 23 differential metabolites is shown in Supplementary Fig. 7A. Then, a heatmap was created to visualize the relative levels of the 23 differential metabolites in each group, and the level of 1,5-AG was significantly reduced in the CKD patients relative to the healthy controls (Supplementary Fig. 7B). We also confirmed the significantly decreased serum level of 1,5-AG in CKD patients by external validation 1 using GC-MS based targeted metabolomics (110 patients and 110 healthy subjects from Ningbo Hospital of Zhejiang University) (Fig. 2J). The accurate quantification of 1,5-AG showed that the concentration of 1,5-AG in serum was 15.15 ± 5.21 μg/mL in the healthy control group, and 5.56 ± 3.11 μg/mL in the CKD group. In addition, we further confirmed the significantly decreased serum level of 1,5-AG in CKD patients by external validation 2 using LC-MS based targeted metabolomics (100 patients and 100 healthy subjects from the Putuo People’s Hospital) (Fig. 2K). The concentration of 1,5-AG in serum was 18.24 ± 6.64 μg/mL in the control group, and 4.24 ± 4.43 μg/mL in the CKD group.
To assess whether B. fragilis attenuates renal fibrosis by upregulation of the level of 1,5-AG, we examined the renoprotective effect of 1,5-AG in UUO-mice. Administration of 1,5-AG significantly improved renal morphology of the UUO-mice (Fig. 3A), reduced renal index of the UUO-mice (Fig. 3B) and reduced the levels of several pro‑fibrotic markers, including pro-fibrotic proteins collagen I, fibronectin and alpha-smooth muscle actin (α-SMA) (Fig. 3C, D). Treatment with 1,5-AG significantly attenuated the increase in BUN, Scr, triglyceride and total cholesterol levels, suggesting that 1,5-AG improved renal function (Fig. 3E). H&E and Masson staining showed that tubular dilatation, tubular atrophy, and widening of the interstitial space with severe inflammatory cell infiltration were attenuated by 1,5-AG (Fig. 3F, G). In addition, 1,5-AG decreased LPS level and inhibited inflammation in UUO mice (Supplementary Fig. 8). Moreover, 1,5-AG activated the Nrf2/Keap1 and inhibited TGF-β/Smad signaling pathways in UUO mice (Fig. 3H–K, Supplementary Fig. 5B). Taken together these results indicate that 1,5-AG significantly attenuated renal fibrosis in vivo.
The transmembrane G protein-coupled bile acid receptor (TGR5) is a cell membrane bile acid receptor, which is closely associated with fibrosis and inflammation. In this study, immunohistochemical staining showed that the TGR5 levels of the IgA nephropathic patients were markedly decreased (Fig. 4A, B). We then examined the expression of TGR5 in UUO and adenine models. Consistent with the results in human subjects, the results indicated that the levels of TGR5 were significantly decreased in the UUO and adenine mice (Fig. 4C). Treatment with 1,5-AG significantly upregulated the levels of TGR5 (Fig. 4C). We further examined the protective effects of 1,5-AG against fibrosis and inflammation in vitro. Treatment with 1,5-AG significantly inhibited the expression of the pro-fibrotic mediators (collagen I, vimentin, fibronectin) in HK-2 cells and inflammatory markers (TNF-α, IL-1β, IL-6) in HMC cells (Supplementary Fig. 9A–D). The mRNA expressions of these pro-inflammatory cytokines (IL-1β, IL-6, P67) were also decreased in HMC cells treated with 1,5-AG (Supplementary Fig. 9E). In TGF-β1 + high glucose-induced primary mouse renal tubular cells (PRTC), 1,5-AG significantly inhibited the expression of the pro-fibrotic mediators (collagen I, vimentin, fibronectin) (Fig. 4D, E). In addition, the levels of TGR5 were significantly restored after treatment with 1,5-AG in PRTC (Fig. 4D, E). TGR5 activation increases cAMP production. Therefore, we examined the effects of 1,5-AG on cAMP levels. The results indicated that 1,5-AG increased cAMP levels in adenine mice (Fig. 4F). To explore the binding of 1,5-AG to TGR5, in silico molecular docking and then molecular dynamics (MD) simulation were performed. The root mean square deviation (RMSD) of protein backbone atoms and 1,5-AG during the MD simulations is shown in Fig. 4G. The RMSD of TGR5 increased to 6 Å in 100 ns and then equilibrated at approximately 5.5 Å. 1,5-AG was stably bound at the pocket during all of the MD simulation. The binding mode of 1,5-AG to TGR5 was determined through the MD simulation. As shown in Fig. 4H, 1,5-AG formed five hydrogen bonds with TGR5 by interacting with ASN76, GLY73 and PRO72. The binding energy calculated through the Molecular Mechanics-Poisson Bolzmann Surface Area (MM-PBSA) method was −4.48 kcal mol−1. These results indicated that 1,5-AG might bind to TGR5. Taken together, these results suggested that 1,5-AG is an agonist for the TGR5 receptor. It has been reported that TGR5 attenuated liver ischemia–reperfusion injury by activating the Nrf2/Keap1 signaling pathway in mice. In this study, we have demonstrated that 1,5-AG inhibited oxidative stress by activating the Nrf2/Keap1 signaling pathway in PRTC (Fig. 4I, J).
To investigate whether the anti-fibrotic effect of 1,5-AG relies on the TGR5 receptor, we knocked down TGR5 expression in PRTC. TGR5 expression in the knockdown group was visibly reduced compared to the control group (Supplementary Fig. 10A, B). Treatment with 1,5-AG could significantly decrease the pro-fibrotic proteins including collagen I, vimentin, and fibronectin in PRTC (Fig. 4D, E). However, the knockdown of TGR5 abolished the protective effect of 1,5-AG in PRTC (Fig. 4K, L). Likewise, the attenuation of the pro-fibrotic proteins observed after 1,5-AG treatment was reversed by the administration of SBI-115, a potent selective antagonist of TGR5 (Fig. 4M, N).
The label-free quantitative proteomics was performed to profile the proteome differences between the sham and the UUO mice. Compared to the sham mice, we found 216 upregulated proteins and 215 downregulated proteins in the UUO mice based on the criterion of fold change ≥ 5, and adjusted-p value ≤ 0.01 (Supplementary data file, Supplementary Fig. 11B). Among them, SGLT2 expression was significantly downregulated in the UUO mice (Fig. 5A). GO and KEGG pathway analyses are presented in Supplementary Fig. 11A, C. Western blot analysis also verified this result using a specific anti-SGLT2 antibody (Fig. 5B, C). We also analyzed the mRNA level of SGLT2 in the kidneys of patients with lupus nephritis, hypertensive nephropathy and IgA nephropathy using the GEO database. The results indicated that SGLT2 mRNA level was significantly decreased in these nephropathic patients compared to healthy subjects (Fig. 5D). The SGLT2 is responsible for the tubular reabsorption of filtered glucose from the kidney into the bloodstream. The chemical structure of 1,5-AG is similar to that of glucose. Importantly, administration of B. fragilis significantly upregulated the UUO-induced decrease in mRNA and protein levels of the SGLT2 (Fig. 5B, C, E). B. fragilis transports more 1,5-AG from the kidney into the bloodstream possibly owing to upregulation of SGLT2 expression.
To explore the binding of 1,5-AG to SGLT2, in silico molecular docking and molecular dynamics (MD) simulation were performed. The root mean square deviation (RMSD) of protein backbone atoms and 1,5-AG during the MD simulation is shown in Fig. 5F. The RMSD of SGLT2 increased to 8 Å in 100 ns and then equilibrated at approximately 8 Å. 1,5-AG was stably bound at the pocket during all of the MD simulations. The binding mode of 1,5-AG to SGLT2 was determined through MD simulation. As shown in Fig. 5G, H, 1,5-AG formed five hydrogen bonds with SGLT2 by interacting with Lys321, Glu99, Phe98, Gln457 and Trp291. The binding energy calculated through the MM-PBSA method was −19.73 kcal mol−1. These results indicated that 1,5-AG might bind to SGLT2. Empagliflozin is an inhibitor of SGLT2 and works by increasing sugar loss in urine. We found that administration of empagliflozin significantly decreased the serum concentration of 1,5-AG compared to the control mice (Fig. 5I). Next, cellular uptake experiments of 1,5-AG were performed in stably SGLT2-transfected HEK293 versus wild-type (WT) cells. We found that the uptake of 1,5-AG in HEK293 cells transfected with SGLT2 was 1.7-fold higher than that in WT cells (Fig. 5J), suggesting that 1,5-AG was the substrate of SGLT2. Taken together, these results demonstrate that SGLT2 is responsible for renal reabsorption of 1,5-AG.
Previous studies have shown that herbal medicines are able of maintaining intestinal flora homeostasis. We have demonstrated that B. fragilis ameliorates renal fibrosis. Thus, we hypothesized that, active components from herbs could attenuate renal fibrosis via the action of B. fragilis. To this end, we assessed the growth-modulating effect of 14 active components associated with CKD on B. fragilis in vitro (Supplementary Fig. 12). The results indicated that only madecassoside (Mad) promoted the growth of B. fragilis. As shown in Fig. 6A, B, oral administration of Mad improved renal morphology and the renal index. Oral administration of Mad improved renal histology (Fig. 6C, D). Moreover, oral administration of Mad improved renal function in the UUO mice (Fig. 6G). However, Mad had little renoprotective effect in the UUO mice when administered by intraperitoneally (i.p.), which suggests that p.o. Mad does not exert its anti-renal fibrotic effect by direct absorption into the blood stream, but the intestinal tract might be its primary site of action (Fig. 6A–D). When Mad is administered p.o., it is eventually hydrolyzed to madecassic acid (MA) by intestinal flora. We found that MA did not improve renal fibrosis, renal histology and renal function in the UUO-mice (Fig. 6E–I). To investigate whether the renoprotective effects of Mad are dependent on the presence of gut microbiota, we treated the UUO mice with a cocktail of antibiotics including ampicillin, gentamicin, neomycin, metronidazole and vancomycin. We found that when the gut microbiota was suppressed by the antibiotics cocktail, the renoprotective effects of Mad were abolished (Fig. 6E–I). Moreover, Mad inhibited oxidative stress and activation of the TGF-β/Smad signaling pathway in the UUO mice (Fig. 6J–M, Supplementary Fig. 5C). In addition, oral administration of Mad (80 mg/kg) did not affect the morphology of colon tissue (Supplementary Fig. 13A, B). Taken together, these results indicate that Mad protects mice against renal fibrosis in a gut microbiota–dependent manner. To confirm the growth-promoting effect of Mad on B. fragilis in vivo, we examined the abundance of B. fragilis in the sham, UUO- and Mad-treated mice by qPCR (Fig. 6N). The results indicated that treatment with Mad could restore the decreased B. fragilis level in the UUO mice.
Here, we confirmed the renoprotective effects of B. fragilis, 1,5-AG and Mad using an adenine mouse model. H&E and Masson staining showed that tubular dilatation, tubular atrophy, and widening of the interstitial space with severe inflammatory cell infiltration were attenuated by administration of B. fragilis, 1,5-AG or Mad in adenine-treated mice (Fig. 7A, B). Treatment with B. fragilis, 1,5-AG or Mad significantly attenuated the increase in BUN and Scr levels, suggesting that B. fragilis, 1,5-AG or Mad improved renal function in the adenine-induced CKD mice (Fig. 7C). Adenine-treated mice that received B. fragilis, 1,5-AG or Mad showed significant reductions in the levels of several pro-fibrotic markers, including the pro-fibrotic proteins collagen I, fibronectin and α-SMA (Fig. 7D–I). However, MA (80 mg/kg) did not improve renal fibrosis, histology or function in the adenine-induced CKD mice (Supplementary Fig. 14). Moreover, the adenine-induced CKD mice showed significantly decreased SGLT2, while B. fragilis significantly upregulated the decreased SGLT2 level (Fig. 7J, K). The fecal and serum levels of LPS were assessed in adenine-treated mice. The results indicated that fecal and serum LPS levels were higher in the adenine-treated group in relation to the controls (Supplementary Fig. 15A, B). Oral administration of live B. fragilis or 1,5-AG dramatically decreased the fecal and serum LPS concentrations in the adenine-treated mice (Supplementary Fig. 15A, B). We then examined the intestinal tight junction expression of occludin and ZO-1 in the adenine-induced CKD mice. Adenine stimulation obviously decreased the expressions of occludin and ZO-1 in comparison with the control group (Supplementary Fig. 15C, D). B. fragilis treatment improved the protein level of occludin and ZO-1 compared to the adenine group (Supplementary Fig. 15C, D).
Increasing evidence suggests that the gut microbiota plays a key role in the development of CKD. Quantitative and qualitative alterations in the gut microbiota have been noticed in patients with CKD and ESRD. Previous studies suggest that prebiotics and probiotics play pivotal roles in maintaining a metabolically balanced gut microbiota and reducing progression of CKD and uremia-associated complications. Here, we found that the abundance of B. fragilis is lower in the gut microbiome of patients with CKD and UUO-mice. To our knowledge, this is the first study to report that oral administration of live B. fragilis can protect mice against renal fibrosis. The anti-fibrotic effect of B. fragilis may be associated with the downregulation of LPS and upregulation of 1,5-AG, which inhibits inflammation, oxidative stress, and the TGF-β/Smad signaling pathway. Additionally, we report a new function of SGLT2 as a 1,5-AG transporter in the kidney. 1,5-AG reabsorption in the kidney is improved by upregulation of SGLT2 expression. In addition, we demonstrate that Mad functions as B. fragilis growth modulator in vitro and in vivo, and show that the oral administration of Mad is effective in the upregulation of 1,5-AG to prevent development of renal fibrosis. Bacterial species belonging to the phyla Bacteroidetes and Firmicutes dominate the gut microbiota. Among the Bacteroides species, B. fragilis is an important obligate anaerobe that colonizes the mammalian lower gastrointestinal tract. B. fragilis has been extensively studied and shown to be equally effective in preventing colitis and experimental allergic encephalomyelitis (EAE) in murine models. It has been reported that PSA produced by B. fragilis induces an anti-inflammatory milieu involving the stimulation of interleukin-10-producing CD4 + Foxp3 + T-regulatory cells in the intestine, thereby reducing pathological gastrointestinal symptoms in a mouse model of colitis. Kasper et al. demonstrated that B. fragilis can protect against neuroinflammation in mouse models of multiple sclerosis. In general, these studies raised the possibility that B. fragilis might be important for the establishment of beneficial intestinal microbiota and could be developed into a probiotic therapy. However, whether or not B. fragilis could protect against renal fibrosis has not been studied. In this work, we found that the B. fragilis treatment effectively alleviated the disruption of serum biochemistry, renal histopathology and attenuated renal fibrosis in UUO/Adenine mice. In addition, treatment with B. fragilis could protect the kidneys by inhibiting inflammation, oxidative stress and the TGF-β/Smad signaling pathway. Lipopolysaccharide (LPS), a component of the outer membrane of gram-negative bacteria, is a potent ligand for toll-like receptor 4 (TLR4), leading to the canonical activation of NF-κB and the associated expression of pro-inflammatory mediators, such as tumor necrosis factor α (TNFα) and interleukin-6 (IL-6). It is important to note that oral administration of B. vulgatus and B. dorei dramatically decreased colon LPS concentrations and offered protection against atherosclerosis. However, the effect of B. fragilis on LPS concentrations in UUO and adenine mice was not known until now. In the present study, we found that fecal and serum LPS levels were higher in CKD patients with a lower abundance of B. fragilis. Oral administration of live B. fragilis dramatically decreased the fecal and serum LPS concentrations and protected mice against inflammation. We examined the intestinal tight junction expression of occludin and ZO-1 in adenine-induced CKD mice. Adenine stimulation obviously decreased the production of occludin and ZO-1 in comparison to the control group. B. fragilis treatment improved the protein levels of occludin and ZO-1 compared to the levels in adenine-induced CKD mice. These results indicated that adenine-induced CKD mice showed breakdown of the gut epithelial barrier, which led to high levels of LPS in the serum. Moreover, the decreased LPS level in feces indicated that B. fragilis dramatically decreased the production of LPS 1,5-AG, a naturally occurring 1-deoxy form of glucose, is derived primarily from dietary sources. Normally, in the kidneys, 1,5-AG is filtered and completely reabsorbed. However, with elevated serum glucose concentrations, glucose is not completely reabsorbed by the kidney, and serum 1,5-AG falls due to competitive inhibition of renal tubular reabsorption. As such, low 1,5-AG is a marker of hyperglycemia over a period of approximately 1–2 weeks. A global, untargeted metabolomics study discovered that the level of 1,5-AG, out of 204 metabolites examined, was an independent risk factor for chronic kidney disease in the Atherosclerosis Risk in Communities Study. Rebholz et al. found that low 1,5-AG levels are associated with higher risk of incident ESRD independent of baseline kidney function. In this study, we also confirmed the significantly decreased serum level of 1,5-AG in patients with CKD compared to controls. Since 1,5-AG is 1-deoxy form of glucose, we speculated that SGLT2 acts as a 1,5-AG transporter in the kidneys. It was reported that the expression of glucose transporters was decreased in nephrectomized rats. As renal impairment or tubular damage progresses, it is possible that reabsorption of 1,5-AG would be reduced as a result of decreases in SGLT2 number and aggravating damage of glucose cotransporters, which might be responsible for the reduction of the serum level of 1,5-AG in patients with CKD. Thus, we examined the mRNA level of SGLT2 in kidneys from sham, and UUO-mice. The results indicated that SGLT2 mRNA level was significantly decreased in UUO mice, showing a trend similar to 1,5-AG. Furthermore, we demonstrated that administration of B. fragilis significantly upregulated the UUO-induced decrease in SGLT2 expression at the mRNA and protein levels. Molecular docking and MD simulation indicated the binding of 1,5-AG to SGLT2. Cellular uptake experiments suggested 1,5-AG as the substrate for SGLT2. Administration of empagliflozin significantly decreased the serum 1,5-AG levels. Our results demonstrated that B. fragilis treatment can upregulate the level of SGLT2 which contributes to renal reabsorption of 1,5-AG in UUO mice while replenishment of 1,5-AG improved renal dysfunction and attenuated inflammation and renal fibrosis. TGR5, a cell membrane bile acid receptor, expression varied in multiple tissues. Increasing evidence has reported the crucial role of TGR5 in many biological functions, including energy homeostasis and glucose metabolism. Activation of TGR5 was found to prevent kidney disease in obese and diabetic mice by inhibiting oxidative stress I In addition, recent studies have proved that TGR5 contributes significantly to ameliorating inflammation. We have demonstrated here that, 1,5-AG is a TGR5 agonist and knockdown of TGR5 abolishes the anti-fibrotic effect of 1,5-AG in vitro, suggesting the anti-fibrotic effect of 1,5-AG might be dependent on TGR5. In this study, we demonstrated that treatment with B. fragilis can ameliorate renal fibrosis. Thus, we speculated that certain small molecular compounds could attenuate renal fibrosis by remodeling the microbiota composition, especially B. fragilis. We assessed the growth-modulating effect of 14 active components from herbal medicines associated with CKD on B. fragilis in vitro. The results indicated that Mad significantly promotes the growth of B. fragilis in vitro. Mad is a pentacyclic triterpene isolated from Centella asitica (L.). A number of studies have suggested that this compound may exhibit anti‑inflammatory, antioxidant, anticancer and anti-pulmonary fibrotic effects. As a triterpenoid saponin, orally administered Mad is extremely difficult to be absorbed. When Mad is administered p.o. it is eventually hydrolyzed to MA which is the main product of intestinal metabolism. Thus, the level of Mad in plasma or tissue is much lower than the minimal effective concentration required for inhibition of lung fibroblasts. We also examined whether Mad could promote the growth of B. fragilis and exert any anti-renal fibrotic effect in vivo. Our results suggest that oral administration of Mad, not MA, protected the mice against renal fibrosis in a gut microbiota–dependent manner. Moreover, treatment with Mad could restore the decreased B. fragilis abundance in the UUO mice, implying the in vivo growth-promoting effect of Mad on B. fragilis. In summary, the results of our study show that oral administration of live B. fragilis could attenuate renal fibrosis and the anti-fibrotic effect of B. fragilis may be associated with the downregulation of LPS and upregulation of SGLT2 which contribute to renal reabsorption of 1,5-AG. 1,5-AG as an agonist of TGR5, attenuated renal fibrosis via inhibition of oxidative stress and inflammation. As a B. fragilis growth modulator, the small molecular compound Mad was effective in upregulating 1,5-AG to prevent the development of renal fibrosis. Our findings should be of value in modulating the microbiome as a useful therapeutic strategy for progressive renal fibrosis (Fig. 8).
All procedures were approved by the medical ethics committee of the Affiliated Hospital of Nanjing University of Chinese Medicine and followed the tenets of the Declaration of Helsinki (2019NL-109-02). All subjects were informed of the use of their feces and blood, and written informed consent was obtained. The collection, preservation and processing of the fecal samples were conducted according to a previous protocol. Feces from 10 CKD patients and 10 age- and sex-matched healthy subjects as microbiota discovery set were collected from the Department of Nephrology, Renmin Hospital of Wuhan University (Supplementary Table 3). Feces from 15 CKD patients and 15 age- and sex-matched healthy controls as microbiota validation set were enrolled from the Putuo People’s Hospital (Supplementary Table 4). Sera from 115 CKD patients and 113 age- and sex-matched healthy subjects for GC-MS based untargeted metabolomics were collected from the Affiliated Hospital of Nanjing University of Chinese Medicine (Supplementary Table 5). Sera from 110 CKD patients and 110 age- and sex-matched healthy subjects for external validation 1 using GC-MS based targeted metabolomics were collected from the Ningbo Hospital of Zhejiang University (Supplementary Table 6). Sera from 100 CKD patients and 100 age- and sex-matched healthy subjects for external validation 2 using LC-MS based targeted metabolomics were collected from the Putuo People’s Hospital (Supplementary Table 7). Slices of renal tissues from patients with category III immunoglobulin A nephropathy were provided by the Ningbo Hospital of Zhejiang University. Patients with acute kidney injury, liver disease, active vasculitis, gastrointestinal pathology or cancer were excluded from the study. As defined in international guidelines, CKD refers to a glomerular filtration rate (GFR) of less than 60 ml/min per 1.73 m2, or a marker of kidney damage, or both, for a duration of at least 3 months.
Animal experiments were conducted in accordance with the Guidelines for Animal Experimentation of China Pharmaceutical University (Nanjing, China), and the protocols were approved by the Animal Ethics Committee of this institution (No: 202002001). Mice were housed in pathogen-free and ventilated cages in a 12 h light/dark cycle, with room temperature at 25 ± 2 °C and humidity between 40 and 60%. All mice used in this work were male. Eight-week-old ICR male mice (18 to 22 g) were allowed free access to water and regular chow and their body weights taken every week. Mice models of renal fibrosis were established by known procedures for UUO. After general anesthesia, complete UUO was carried out by double-ligation of the left ureter by 4–0 silk following a dorsal incision. The ureters of sham operated mice were exposed, but not ligated. For the adenine-induced mouse model, control group animals received saline daily (0.2 ml/100 g) by oral gavage for 6 weeks. The experimental groups received daily adenine injectable suspension of 80 mg/kg by oral gavage for 3 weeks. Starting on 4th week, the treatment groups received intervention for 3 weeks.
For UUO study, the experimental mice were randomly grouped as follows: (1) sham control; (2) UUO; (3) UUO + 1,5-AG (J&K Scientific, Bailingwei Technology, China; 100 mg/kg in saline administered i.p. for 7 days); (4) UUO + Mad (FeiYu Biotech Co Ltd, Nantong, China; 80 mg/kg in saline administered p.o. or i.p. daily for 14 days); (5) UUO + MA group (FeiYu Biotech Co Ltd, Nantong, China; 40 mg/kg given p.o. daily for 14 days); (6) UUO + Mad + antibiotics [YuanYe Bio-Technology Co Ltd (Shanghai, China); Ampicillin (5 mg/mL); Gentamicin (5 mg/mL); Neomycin (5 mg/mL); Metronidazole (5 mg/mL); Vancomycin (2.5 mg/mL); gavage mice with 200 μL of antibiotic mix]. For the adenine model, the experimental mice were randomly grouped as follows: (1) Control; (2) Adenine (J&K Scientific, Bailingwei Technology, China; 80 mg/kg by oral gavage); (3) Adenine + 1,5-AG (100 mg/kg in saline administered i.p. for 21 days); (4) Adenine + Mad (80 mg/kg in saline administered p.o. for 21 days); (5) Adenine + MA (80 mg/kg in saline administered p.o. for 21 days). (6) Control+ empagliflozin (GlpBio Technology, CA, USA; 10 mg/kg administered p.o. for 14 days). After the mice were sacrificed, their kidneys were harvested by surgical procedure and stored at −80 °C for further analysis. At least six animals were included in each group, and at least three independent experiments were performed. No animal was excluded from experiments unless for technical reasons.
DNA from human or mouse fecal samples was extracted using the E.Z.N.A. ®Stool DNA Kit (D4015, Omega, Inc., USA) according to manufacturer’s instructions. The V3-V4 region of the prokaryotic small-subunit (16 S) rRNA gene was amplified with primers 341 F (5′-CCTACGGGNGGCWGCAG-3′) and 805 R (5′-GACTACHVGGGTATCTAATCC-3′). The 5′ ends of the primers were tagged with specific barcodes per sample and sequencing universal primers. PCR amplification was performed in a total volume of 25 μL reaction mixture containing 25 ng of template DNA, 12.5 μL PCR Premix, 2.5 μL of each primer, and DEPC water to adjust the volume. The PCR products were confirmed with 2% agarose gel electrophoresis. The PCR products were purified by AMPure XT beads (Beckman Coulter Genomics, Danvers, MA, USA) and quantified by Qubit (Invitrogen, USA). The amplicon pools were prepared for sequencing and the size and quantity of the amplicon library were assessed on Agilent 2100 Bio-analyzer (Agilent, USA) and with the Library Quantification Kit for Illumina (Kapa Biosciences, Woburn, MA, USA), respectively. The samples were sequenced on an Illumina NovaSeq platform by LC-Bio Technology Co., Ltd (Hang Zhou, Zhejiang Province, China) according to the manufacturer’s recommendations. Paired-end reads were merged using FLASH. Quality filtering on the raw reads were performed under specific filtering conditions to obtain the high-quality clean tags according to the Fqtrim (v0.94). Chimeric sequences were filtered using Vsearch software (v2.3.4). After de-replication using DADA2, we obtained feature table and feature sequence. Relative abundance was used in bacteria taxonomy, alpha diversity and beta diversity were analyzed by QIIME2 process, and figures were drawn by R (v4.1.0). The sequence alignment of species annotation was performed by QIIME2 plugin feature-classifier, and the alignment database was SILVA and NT-16S. For 16 S rDNA bacteria gene sequencing of fecal samples in CKD patients and healthy subjects, the rarefaction number was 25134. Adjusted p-values less than 0.05 for phylum and genus were considered statistically significant in comparisons between CKD patients and healthy subjects.
Fecal DNA was extracted using the TIANamp Stool DNA Kit [TIANGEN Biotech (Beijing) Co. Ltd., DP328] according to the manufacturer’s protocol, and the concentration was measured by 24-well plate reader (BioTek, Winooski, VT, USA). qPCR assays were performed using the AceQ qPCR SYBR Green Master Mix (Vazyme Biotech Co. Ltd.) with primers that amplify the genes encoding 16S rRNA from B. fragilis and all bacteria (Supplementary Table 8) by the StepOne Real-Time PCR System (A&B, Waltham, MA, USA).
B. fragilis (NCTC 9343) was purchased from the National Collection of Type Cultures (NCTC) and cultivated in sterilized thioglycolate medium. An anaerobic chamber containing 10% CO2, 10% H2, and 80% N2 was used for all anaerobic microbiological works. Cultures were collected in log phase and diluted with sterile phosphate-buffered saline (PBS) to 2 × 108 colony-forming units/ml for gavage. For the sham-controlled trials, B. fragilis were heat-killed at 121 °C (treatment duration, 15 min). The UUO-mice were gavaged daily with either live bacteria or sham (0.2 mL/10 g).
The cultivated B. fragilis was seeded in 96-well plates under the initial OD600 value of approximately 0.1, and treated with 14 active natural products associated with CKD, including madecassoside (Mad), asiatic acid (AA), asiaticoside (Aad), madecassic acid (MA), ginsenoside Re, Rc, Rg1, Rh1, Rb1, artemisinin, emodin, astragaloside IV, dihydroartemisinin (DHA), and rhein at the final concentration of 100 µM. Then, the OD600 value was recorded after co-cultivation at different time-course of 0 h, 12 h, 24 h, 36 h, 48 h, 60 h, and 72 h using microplate reader. Modulators were screened through observation of the effects on the growth curve of B. fragilis in vitro.
HK-2 cells (the Cell Resource Center, Peking Union Medical College) and HMC cells (FuHeng Biology, Shanghai) were cultured in DMEM/F-12 (C11330500BT, Gibco) and DMEM (01-052-1ACS, BI), respectively, supplemented with 10% FBS (04-001-1ACS, BI) and 1% penicillin/streptomycin (B540732, Sangon Biotech). The primary mouse renal tubular epithelial cells (PRTC cells) were isolated and cultured according to our previous report. HK-2 cells were treated with 10 ng/mL recombinant human TGF-β1 protein (R&D system, USA), HMC cells were treated with 15 μg/ml LPS (L2630, Sigma, USA), while PRTC cells were treated with 10 ng/mL recombinant human TGF-β1 protein and 30 mM high glucose (A100188, Sangon Biotech). The concentration of 1,5-AG (BYOC-ALD-070-1g, Omicron Biochrmicals) for 24 h treatments of the HK-2, HMC and PRTC cells was 50 μM. The concentration of SBI-115 (TGR5 antagonist, HY-111534, MedChemExpress, Monmouth Junction, NJ, USA) for 48 h treatment of the PRTC cells was 10 μM.
Blood analyses were performed on an automatic biochemistry analyzer (Chemray 240, Redu Life Technology). Kidney tissue was fixed in 10% formaldehyde and then embedded in paraffin. Five µm-thick paraffin sections were cut. Sections were dewaxed and hydrated through graded alcohols and dipped in water, and then were stained with conventional H&E. Western blot protocol is as follows: Protein concentration was measured by Enhanced BCA Protein Assay Kit (P0009, Beyotime, China). The 10–20 µl of total protein was fractionated by Polyacrylamide gel and transferred to a 0.45 μm hydrophobic PVDF transfer membrane (IPVH00010, Merk Millipore, Germany). After incubated for 2 h in 5% non-fat milk blocking buffer, the membranes were incubated overnight at 4 °C with primary antibody. The secondary antibodies of goat anti-rabbit (1:2000, #7074, CST, USA), horse anti-mouse (1:2000, #7076, CST, USA) were incubated with 2 h at room temperature. The following antibodies were used: anti-collagen I (ab260043, Abcam, rabbit, dilution 1:1000 for WB), anti-fibronectin (ab2413, Abcam, rabbit, dilution 1:2000 for WB), anti-α-SMA (ab124964, Abcam, rabbit, dilution 1:5000 for WB), anti-GAPDH (HRP-60004, Proteintech, dilution 1:4000 for WB), anti-TGF-β1 (21898-1-AP, Proteintech, rabbit, dilution 1:1000 for WB), anti-Smad2 (#5339, CST, rabbit, dilution 1:1000 for WB), anti-Smad3 (#9523, CST, rabbit, dilution 1:1000 for WB), anti-12-LO (C-5) (sc-365194, Santa Cruz Biotechnology, mouse, dilution 1:100 for WB), anti-Rac1 (66122-1-Ig, Proteintech, mouse, dilution 1:1000 for WB), anti-Nrf2 (16396-1-AP, Proteintech, rabbit, dilution 1:1000 for WB), anti-Keap1 (10503-2-AP, Proteintech, rabbit, dilution 1:1000 for WB), anti-HO-1 (66743-1-Ig, Proteintech, mouse, dilution 1:1000 for WB), anti-SGLT2 (ab37296, Abcam, rabbit, dilution 1:1000 for WB), ZO-1(21773-1-AP, Proteintech, rabbit, dilution 1:2000 for WB), Occludin (66378-1-Ig, Proteintech, mouse, dilution 1:5000 for WB), anti-TGR5 (ab72608, Abcam, rabbit, dilution 1:1000 for WB and 1:100 for IHC), anti-IL-1β (ab254360, Abcam, rabbit, dilution 1:1000 for WB), anti-TNF-α (ab215188, Abcam, rabbit, diltion 1:1000 for WB), anti-IL-6 (ab259341, Abcam, rabbit, dilution 1:1000 for WB), anti-Vimetin (#5741, CST, rabbit, dilution 1:1000 for WB and 1:50 for IF), E-cadherin (#14472, CST, mouse, dilution 1:50 for IF), HRP-linked anti-rabbit IgG (#7074, CST, rabbit, dilution 1:2000 for WB), HRP-linked anti-mouse IgG (#7076, CST, mouse, dilution 1:2000 for WB).Semi-quantitative analysis of each protein was performed using ImageJ software (version 1.5), and the band densities were normalized to the band density of GAPDH.
PRTC cells were plated in 6-well plates for 24 h before transfection. siRNA targeting TGR5 was acquired from Sangon Biotech. The sequences of TGR5-siRNA were as follows: sense: 5′-CUCUGUUAUCGCUCAUCUCAUTT-3′ and antisense: 5′-AUGAGAUGAGCGAUAACAGAGTT-3′. The siRNA was transfected using INTERFER transfection reagent according to the manufacturer’s protocol.
Plasmids encoding SLC5A2 vector (HBLV-h-SLC5A2-3xflag-PURO) and control vector (HBLV-PURO) were designed and lentivirus was provided by HANBIO (Shanghai, China). HEK293 cells (Stem Cell Bank, Chinese Academy of Sciences, GNHu 43) were transfected with SLC5A2 or control virus at a multiplicity of infection (MOI) of 2. The cells were then treated with 1 μg/ml puromycin for 14 days. The stably expressed SCL5A2 colonies were verified by Western blot. We found overexpression of SGLT2 in HEK293 cells was established (Supplementary Fig. 16).
Total mRNA was extracted using a High Pure RNA Isolation Kit (RNAiso Plus, Takara Bio, Japan) according to the manufacturer’s instructions. Total RNA was reverse transcribed by a HiScript II QRT SuperMix for qPCR according to the manufacturer’s instructions (+gDNA wiper, R233-01, Vazyme, Nanjing, China). Quantitative real-time PCR (qRT-PCR) was carried out by the Step One System (A&B, Waltham, MA, USA) using AceQ qPCR SYBR Green Master Mix (High ROX Premixed, Q141-02, Vazyme Biotech Co. Ltd.). The mRNA levels of the genes were calculated by normalization to the levels of Gapdh. Primer sets for genes are presented in Supplementary Table 8.
For immunofluorescence of vimentin and E-Cadherin in the kidney tissues, the slides of the tissues were incubated with vimentin (#5741, CST, concentration, 1:50) or E-Cadherin (#14472, CST, concentration, 1:50) overnight at 4 °C in a humidified dark chamber, and then incubated with Alexa Fluor 555-labeled goat anti-rabbit IgG (H + L) (ab150078, Abcam, concentration, 1:200) or FITC-labeled goat anti-mouse IgG (H + L) (SA00003-1, Proteintech, concentration, 1:200) for 1 h at room temperature. Subsequently, they were washed with PBS and stained with DAPI for 10 min. The image was acquired by confocal laser scanning microscope (LSM700, Zeiss, Jena, Germany).
Sections (5 μm) of kidney tissues were obtained and deparaffinized in xylene, hydrated in graded ethanol solutions, and rinsed with tap water and distilled water. Then, endogenous peroxide activity was blocked by incubation in 0.3% hydrogen peroxide in methanol for 30 min. For antigen retrieval, the kidney tissue sections were incubated with 10 μmol/L citrate buffer solution (pH: 6.0) and boiled for 10–15 min. Subsequently, the sections were blocked with 10% normal goat serum for 1 h at room temperature and then incubated overnight at 4 °C with the TGR5 antibody (ab72608, Abcam, concentration, 1:75). After washing with PBS, the HRP-linked goat anti-rabbit IgG (AFIHC003, AiFang biology, no dilution) was added, and the sections were incubated at 37 °C for 1 h. Finally, the kidney tissue sections were exposed to diaminobenzidine peroxidase substrate for 5 min and counterstained with haematoxylin and eosin. Images of the sections were obtained using a Leica DMi8 fluorescence microscope (Leica, Germany).
Sera were collected from the blood of human and mice by centrifuging at 2000 × g for 10 min. For fecal samples preparation, feces were accurately weighed, and homogenized for 5 min in 9-fold volume of PBS. The supernatant was collected by centrifugation at 5000 × g for 10 min. For cell sample preparation, cells were collected using a sterile container, and diluted for 1 million/mL using PBS. Intracellular components were released by repeated freeze-thaw cycles. The supernatant was collected by centrifugation at 2000 × g for 10 min. For tissue sample preparation, the tissue samples were accurately weighted, and rapidly frozen with liquid nitrogen, then homogenized by grinders at 4 °C after melting. The supernatant was collected by centrifugation at 5000 × g for 10 min. After treatment, the levels of LPS, pro-inflammatory cytokines and cAMP were assayed. Human LPS (KT98561, MSKBIO, maximum concentration of standard: 960 ng/L) and mouse LPS (KT37561, MSKBIO, maximum concentration of standard: 640 pg/mL), IL-1β (KT21178, MSKBIO, maximum concentration of standard: 160 ng/L), IL-6 (KT99854, MSKBIO, maximum concentration of standard: 240 pg/mL), TNF-ɑ (KT99985, MSKBIO, maximum concentration of standard: 1600 ng/L) in the supernatants of serum and feces were quantified by ELISA. cAMP (E-EL-0056c, Elabscience, maximum concentration of standard:100 ng/L) in the supernatants of cells and renal samples were quantified by ELISA according to the manufacturer’s protocol.
Mice were anesthetized with 10% urethane, and blood samples were obtained by carotid artery cannula on the 14th day. Blood was centrifuged at 2000 × g for 10 min and the sera were collected and stored at −80 °C.
All the blood samples were immediately centrifuged at 2000 × g for 10 min, and sera were transferred into clean Eppendorf tubes. The serum samples were stored at −80 °C.
Internal standard solutions (10 μL of myristic acid-1,2-13C2 in methanol, 1 mg/mL) were added to 200 μL of serum. The mixed solution was extracted with 600 μL of methanol and chloroform (3:1, v/v) and vortexed for 30 s. The mixture was stored at room temperature for 10 min and centrifuged at 15,000 × g for 10 min at 4 °C. The resulting supernatant (600 μL) was transferred to a sample vial for vacuum drying at room temperature. The residue was redissolved in 40 μL of a methoxyamine solution (15 mg/mL in pyridine) and vortexed for 1 min. An oximation reaction was performed at 37 °C for 1.5 h. Then, 80 μL of BSTFA (containing 1% TMCS) was added to the solution, and the solution was vortexed for 30 s. The sample was kept at 70 °C for 1 h and vortexed for 10 s. The supernatant was then transferred to a sample vial for GC-MS analysis. Representative total ion chromatograms are presented in Supplementary Fig. 17.
1, 5-AG standard solutions was diluted to different concentrations. The 600 μL solution was transferred to a sample vial for vacuum drying at room temperature. The residue was redissolved in 50 μL of a methoxyamine solution (15 mg/mL in pyridine) and vortexed for 1 min. An oximation reaction was performed at 37 °C for 1.5 h. Then, 50 μL of BSTFA (containing 1% TMCS) was added to the solution, and the solution was vortexed for 30 s. The sample was kept at 70 °C for 1 h and vortexed for 10 s. The supernatant was transferred to a sample vial for GC-MS analysis. The calculation curve is shown in Supplementary Fig. 18.
The samples were analyzed using an Agilent 7890 chromatograph coupled with a 5977B MS system (Agilent Technologies, Santa Clara, CA, USA). Separation was achieved on a DB-5 ms capillary column coated with 95% dimethyl 5% diphenyl polysiloxane (30 m× 0.25 mm i.d., 0.25-μm film). The initial GC oven temperature was set at 60 °C for 1 min, followed by a 10 °C/min oven temperature ramp to 325 °C, which was maintained for 10 min. The temperature of the inlet, transfer line, and ion source was set to 250, 290, and 250 °C, respectively. The injection volume for untargeted metabolomics was 1 μL with splitless. The injection volume for targeted metabolomics was 1 μL with splitless. Helium was used as the carrier gas with a constant flow rate of 0.87 mL/min. Measurements were made with electron impact ionization (70 eV) in full scan mode (m/z 50 − 650). Quantification analysis of 1,5-AG was performed with SIM mode with m/z 259.0.
For analytical method assessment, 50 µl each of all the samples were pooled to get a quality control sample (QC) that would be tested during the analysis. QC samples were analyzed five times at the beginning of the run and injected once after every 10 injections of the random sequenced samples. The precision and repeatability were validated by the duplicate analysis of six injections of the same QC sample and six parallel QC samples prepared using the same preparation and method, respectively. From this QC sample, extracted ion chromatographic peaks of five ions (7.72_72.1, 11.91_234.1, 17.07_259.1, 19.39_267.1, 27.68_369.1) were selected for method validation. The RSD of peak area and retention time were below 3.1% and 0.04% respectively and the reproducibility and precision were satisfactory for metabolomic analysis. The stability of samples was tested by analyzing two QC samples kept in autosampler at room temperature for 12, 24 and 72 h. The RSD of peak areas of the serum sample for metabolomic analyses was 2.0% to 5.1%.
The raw mass spectrometry data were exported to data format (mzdata) files by Mass Hunter Workstation Software (version B.06.00, Agilent Technologies). Data pre-treatment procedures, such as nonlinear retention time alignment, peak discrimination, filtering, alignment, matching, and identification, were performed in XCMS package (Scripps Center for Metabolomics and Mass Spectrometry, La Jolla, California). The matrix result was reduced by replacing the missing values, and data with more than 20% missing values were removed. The resulting data set, including retention time, sample names and peak areas were introduced into the SIMCA-P 14.0 Software package (Umetrics, Umea, Sweden) for multivariate statistical analysis. The significance of each metabolite was analyzed by the Mann-Whitney-Wilcoxon test with false discovery rate (FDR) correction via Benjamini–Hochberg method. The discrimination of variables was identified by Orthogonal partial least-squared discriminant analysis (OPLS-DA). Adjusted p-values less than 0.05 were considered statistically significant. Differential metabolites were screened by those with variable importance in the projection (VIP) ≥ 1.0 obtained from OPLS-DA and adjusted p-values less than 0.05, where VIP indicates the contribution of each variable to group differences. Differential metabolites were identified by a library search (NIST and Fiehn) and confirmed by available references.
UltiMate® 3000 ultra-performance liquid chromatography system (DIONEX, Sunnyvale, CA, USA) equipped with an ACQUITY UPLC® BEH Amide (2.1 × 100 mm, 1.7 μm, Waters Co., Milford, MA, USA) was used to perform the chromatographic separation of 1, 5-AG. Mobile phase A was water containing10 mM ammonium acetate, while mobile phase B was acetonitrile and water (95: 5), containing 10 mM ammonium acetate. The mobile phase gradient was as follows: 0–1 min, 85 % B, 1–6 min, 85–10% B, 6–8 min, 10–85% B, 8–12 min, 85% B. The flow rate was 0.3 mL/min. The column oven temperature was set to 50 °C. TSQ VantageTM triple quadrupole mass spectrometer (Thermo Fisher Scientific Inc.) was used for the determination of 1, 5-AG. The ESI source was set as negative ion mode with the following parameters: spray voltage, 2.8 kV; capillary temperature, 300 °C; sheath gas flow rate, 45 arb; aux gas flow rate, 15 arb; vaporizer temperature, 300 °C. The MRM parameters of 1, 5-AG were as follows: Parent ion, 163; Product ion, 101; CE, 15; S-lens, 55. The MRM parameters of 1,5-AG-13C6 were as follows: Parent ion, 169; Product ion, 105; CE, 15; S-lens, 55. For sample preparation, an aliquot of 50 μL serum was mixed with 200 μL MeOH containing 5 μg/mL of 1,5-AG-13C6 in 1.5 mL polypropylene test tube. The mixture was then vortexed for 3 min and centrifuged at 13500 × g for 10 min to remove the precipitated protein. The 200 μL supernatant was transferred to another tube and evaporated to dryness by a vacuum drier. The dried residue was redissolved in 80 μL 50% MeOH and centrifuged at 13500 × g again for 10 min, and an aliquot of 2 μL supernatant was subjected to LC-MS system. The quantitation of 1,5-AG in the validation of center 2 was performed with reference to the corresponding isotope-labeled internal standard (20 μg/mL of 1,5-AG-13C6) according to previous report.
Each mouse kidney tissue was ground into powder in nitrogen., then lysis buffer (0.1 M Tris-HCL, pH 7.5, 4% SDS, 0.1 M DTT) was added, and the mixture was sonicated with 20 cycles of pulses (30 s each on/off, 80% power; CosmoSonic II Ultra Sonicator). After heating for 10 min at 95 °C, the lysate was centrifuged at 16,000 × g for 20 min at room temperature. The protein concentrations were determined by measuring tryptophan fluorescence. 100 μg protein was digested by the FASP method. Each sample peptides were loaded onto a 20-cm column packed in-house with C18 3 μM ReproSil particles (Dr. Maisch GmbH), with an EASY-nLC 1200 system (Thermo Fisher Scientific) coupled to the mass spectrometer (Q Exactive Plus, Thermo Fisher Scientific). Column temperature was maintained at 50 °C. Peptides were separated with a 120 min gradient at a flow rate of 300 nL/min. Each sample was detected twice. The raw data files were processed using software MaxQuant (http://www.maxquant.org.) version 1.6.2.10 with an FDR < 0.01 at the levels of proteins and peptides. The MS/MS spectra were searched against the Homo sapiens protein database in UniProt (January 2021). Bioinformatics analyses were carried out with R version 4.1.0 (https://www.r-project.org/) statistical computing software.
The crystal structure of human SGLT2 was predicted by AlphaFold2 (PDB: AF-P31639-F1-model_v1). The crystal structure of human TGR5 was downloaded from the Protein Data Bank (PDB: 7BW0). Then, the binding of 1,5-AG to SGLT2 and TGR5, was performed using the AutoDock program. The genetic algorithm was applied for conformational analysis. To assess the conformational space of 1,5-AG as completely as possible, we performed 100 individual genetic algorithm runs to generate 100 docked conformations. The size of the docking box was properly set to enclose the possible binding pocket. The protein structure was fixed during molecular docking. The docked protein-ligand complex structure was further relaxed through molecular dynamics (MD) simulation by using the Amber program. The most populated structure during the 500 ns MD simulation was obtained via cluster analysis.
Stably transfected HEK293 cells and control vector-transfected HEK293 cells were cultured in high glucose DMEM with 10% FBS, 1% penicillin-streptomycin, and seeded onto 10 cm dish and incubated at 37 °C, 5% CO2, and 95% humidity. Daily changes of medium were performed. At 48 h after plating, uptake experiments were conducted. Cells were first washed 3 times with prewarmed PBS (pH 7.4) and then were incubated with 4 mL of PBS containing 1,5-AG-13C6 at concentrations of 100 μM. Cells were incubated for 10 min at 37 °C. After incubation, uptake was stopped by aspirating the incubation solution and washing each well 3 times with ice PBS. The cells were lysed in 1 mL of 0.1 M NaOH for 30 min, and then neutralized by adding equal volumes of 0.1 M HCl. Some of the cell lysates were used as protein concentration determination by BCA method and the others were used for quantification by GC-MS above. The following Eqs. (1 and 2) were used to calculate the uptake rates (U) and the SGLT2 uptake ratio (UR). A(analyte)lysate is the intensity of 1,5-AG-13C6 in cell lysates. A(IS)lysate is the intensity of internal standard (myristic acid-1,2-13C2) in cell lysates. P is the protein concentration in cell lysates. T is the incubation time. UHEK293-SGLT2 is the uptake rate obtained in HEK293 stably transfected SGLT2. UHEK293-MOCK is the uptake rate obtained in control vector-transfected HEK293. The uptake experiment was conducted according to our previous study.
Data are shown as means ± SD. Multigroup comparisons were performed using one-way ANOVA. Student’s t-test or Mann-Whitney-Wilcoxon test was used for comparisons between two groups. At least three independent experiments were performed. Analyses were performed with Prism Software (GraphPad Software 9.0). All results were considered statistically significant at p-value < 0.05.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Supplementary information Reporting Summary | true | true | true |
|
PMC9569040 | Dandan Wu,Jia Liu,Liji Yu,Shaofang Wu,Xiaomei Qiu | Circular RNA hsa_circ_0000144 aggravates ovarian Cancer progression by regulating ELK3 via sponging miR-610 | 15-10-2022 | circ_0000144,miR-610,ELK3,Ovarian cancer,Proliferation | Background Ovarian cancer is a common cause of death among women and a health problem worldwide. Circ_0000144 has been confirmed to be an oncogene involved in cancer progression, such as gastric cancer. However, the role of circ_0000144 in ovarian cancer remains unclear and needs to be elucidated. This retrospective study aimed to investigate the underlying mechanism of circ_0000144 in ovarian cancer. Methods Differentially expressed circ_0000144 expression in ovarian cancer and normal tissues was identified by quantitative reverse transcriptase polymerase chain reaction (qRT-PCR). In vitro assays were performed to explore the biological functions of circ_0000144 in ovarian cancer cells. An in vivo xenograft model was used to investigate the efficacy of circ_0000144 in the progression of ovarian cancer. Results Circ_0000144 was significantly upregulated in ovarian cancer cells and tissues. Circ_0000144 overexpression significantly promoted ovarian cancer cell proliferation, migration, and invasion. This study further demonstrated that circ_0000144 downregulated ELK3 levels by sponging miR-610 in ovarian cancer cells. Moreover, circ_0000144 significantly promotes ovarian cancer tumorigenesis in vivo. Conclusion Our data indicate that circ_0000144 could enhance the carcinogenesis of ovarian cancer by specifically targeting miR-610, which may serve as a novel target for the diagnosis and prognosis of ovarian cancer. | Circular RNA hsa_circ_0000144 aggravates ovarian Cancer progression by regulating ELK3 via sponging miR-610
Ovarian cancer is a common cause of death among women and a health problem worldwide. Circ_0000144 has been confirmed to be an oncogene involved in cancer progression, such as gastric cancer. However, the role of circ_0000144 in ovarian cancer remains unclear and needs to be elucidated. This retrospective study aimed to investigate the underlying mechanism of circ_0000144 in ovarian cancer.
Differentially expressed circ_0000144 expression in ovarian cancer and normal tissues was identified by quantitative reverse transcriptase polymerase chain reaction (qRT-PCR). In vitro assays were performed to explore the biological functions of circ_0000144 in ovarian cancer cells. An in vivo xenograft model was used to investigate the efficacy of circ_0000144 in the progression of ovarian cancer.
Circ_0000144 was significantly upregulated in ovarian cancer cells and tissues. Circ_0000144 overexpression significantly promoted ovarian cancer cell proliferation, migration, and invasion. This study further demonstrated that circ_0000144 downregulated ELK3 levels by sponging miR-610 in ovarian cancer cells. Moreover, circ_0000144 significantly promotes ovarian cancer tumorigenesis in vivo.
Our data indicate that circ_0000144 could enhance the carcinogenesis of ovarian cancer by specifically targeting miR-610, which may serve as a novel target for the diagnosis and prognosis of ovarian cancer.
Despite great progress in the diagnosis and treatment of cancer over the past decade, ovarian cancer is one of the three most common gynecological malignancies and has the highest mortality rate among gynecological tumors worldwide [1]. Despite continuous progress in surgical, chemotherapy, and radiotherapy treatments, the prognosis of ovarian cancer remains unsatisfactory, and its morbidity and mortality are increasing annually [2, 3]. Therefore, there is an urgent need to identify potential biomarkers for early diagnosis to obtain optimal clinical antitumor therapies. In eukaryotes, circular RNAs (circRNAs) are produced by reverse splicing of mRNA precursors of exons of thousands of genes [4]. CircRNAs are highly abundant and specifically expressed in tissues, and thousands of circRNAs are expressed differently in tumor and normal tissues [5]. Previous studies have shown that circRNAs regulate gene expression via multiple mechanisms, such as sponging of microRNAs (miRNAs) [6]. CircRNAs are involved in many biological functions, especially in cell cycle regulation and extracellular interactions, and play gene regulatory roles in multicellular organisms through interactions with nucleic acids, proteins, and especially microRNAs [7]. The regulation of circRNAs on gene expression controls a variety of biological functions, such as cell growth and apoptosis, development, embryogenesis and the pathogenesis of human diseases, especially cancer. Dysregulated circRNAs play a role in biological processes including cell proliferation, migration, invasion, apoptosis and angiogenesis, thereby affecting tumorigenesis [8, 9]. Some circRNAs have been shown to play an important role in ovarian cancer progression and have been used as biomarkers for the diagnosis and prognosis of ovarian cancer. For instance, circ_0078607 contributes to ovarian cancer carcinogenesis by sponging oncogenic miR-518a-5p to induce FAS expression [10]. However, the development of cancer is complex and may involve multiple signaling pathways. Therefore, the expression profiles and functions of circRNAs in human ovarian cancer need to be investigated. Recently, circ_0000144 was discovered to be an important factor in various tumors [11, 12]. Specifically, circ_0000144 overexpression has been found in gastric cancer (GC) [13] and can also promote the development of bladder cancer by regulating miRNAs [14]. However, its biological function in ovarian cancer remains unknown. ETS transcription factor ELK3 (ELK3) is an ETS domain protein that forms a ternary complex with DNA and serum response factor (SRF) [15]. ELK3 is a transcription suppressor that is converted into a transcription activator by phosphorylation of extracellular signal-regulated kinase 1/2 (ERK1/2) in response to Ras signaling [16]. ELK3 is highly expressed in a variety of cancers, including basal-like malignant breast cancer, and coordinates metastasis during tumor progression [17]. Previous studies have shown that ELK3 plays a vital role in the development of breast and bladder cancer [18]. Overexpression of ELK3 also occurs in ovarian cancer cell lines and human tumors [19]. However, the potential gene regulatory mechanism of ELK3 in human ovarian cancer remains unclear. Therefore, we investigated the association between circ_0000144 and ELK3 while detecting the expression of circ_0000144 in ovarian cancer. In the present study, we first identified that circ_0000144 is significantly upregulated in ovarian cancer tissues and cell lines. Functional experiments revealed that circ_0000144 enhanced ovarian cancer cell proliferation, invasion, and migration and promoted tumor growth in mice. Furthermore, mechanistic investigations indicated that circ_0000144 exhibited a tumor promoter role by sponging miR-610 and increasing ELK3 expression in ovarian cancer cells. Moreover, circ_0000144 knockdown significantly inhibit the expression of ELK3 and suppress ovarian cancer progression. This study revealed a probable pathway mechanism of circ_0000144 in ovarian cancer progression.
Primary ovarian cancer samples were obtained from 60 patients diagnosed with ovarian cancer before treatment between August 1, 2013, and June 30, 2016. Tumor and paired non-carcinoma tissue samples were acquired from patients with ovarian cancer at Department of Obstetrics and Gynecology in the First Hospital of Quanzhou Affiliated to Fujian Medical University. Peripheral blood samples were collected from the patients and healthy control subjects when they attended the clinic at the start of the study. Serum was prepared and stored at − 80 °C until further processing. The study protocol was approved by the Medical Ethics Committee of First Hospital of Quanzhou Affiliated to Fujian Medical University and the code of ethical number was 2020–185. The procedures involving human participants in this study were conducted in accordance with the guidelines of the Declaration of Helsinki. Each participant provided informed consent. Human ovarian cancer cell lines (SKOV3, ES-2, and OVCAR3) and normal human ovarian cells (IOSE80) were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). All cells were cultured in DMEM (Gibco, USA) containing 10% fetal bovine serum (FBS, Gibco, USA) and 0.1% penicillin-streptomycin and incubated in an incubator at 37 °C and 5% CO2.
The synthetic circ_0000144 sequence was subcloned into the pcDNA3.1 vector as pcDNA3.1-circ_0000144 vector. Small interfering RNAs (siRNAs) against circ_0000144 (si-circ_0000144, 5′-AGGGAGAGAGAGGTAGAACTA-3′) were used to reduce circ_0000144 expression [13]. MiR-610 mimics were used to overexpress miR-610 in the cells, and NC mimics served as negative controls. MiR-610 inhibitors were used to knock down miR-610 expression in cells. Full-length ELK3 was constructed into pcDNA3.1 vector, and blank pcDNA3.1 was used as a control. Cells were transfected using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions.
Total RNA from tissues and indicated treated cells was extracted with TRIzol reagent. cDNA was synthesized from RNA by reverse transcription using a PrimeScript RT reagent kit (RR036A, Takara Biotechnology, Shiga, Japan). Real-time qPCR was performed in accordance with the SYBR® Premix Ex TaqTM II kit instructions (RR820A, Takara Biotechnology, Shiga, Japan) using the ABI7500 Real-Time Fluorescence Quantitative PCR System (7500, ABI, USA). The relative expression of circ_0000144, miR-610p, and ELK3 was calculated using the 2 −ΔΔCt method, with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) serving as the endogenous control. Primers were performed the amplification according to previous reports [13, 20], and the sequences were as follows: GAPDH, forward: 5′-TATGATGATATCAAGAGGGTAGT-3′ and reverse: 5′-TGTATCCAAACTCATTGTCATAC-3′; Circ_0000144, forward: 5′-GAGTGTTGGCCTGTCCTCAA-3 and reverse: 5′-TTGTGCCCAGTTGCCTGTAT-3′; SLAMF6, forward, 5′-GAGTGTTGGCCTGTCCTCAA-3′ and reverse, 5′-TTGTGCCCAGTTGCCTGTAT-3′; MiR-610, forward: 5′-GAGCTAAATGTGTGCTGG-3′ and reverse: 5′-GAACATGTCTGCGTATCTC-3′.
The total genomic DNA (gDNA) was extracted using the Blood/Cell/Tissue Genomic DNA Extraction Kit (Tiangen, China) with the manufacturer’s instructions. Based on NCBI reference sequences, convergent and divergent primers were designed using Primer 5.0 software to validate the existence of circRNA. All primers used in this study were synthesized by Sangon (Sangon Biotech, Shanghai, China). For each PCR amplification, cDNA or gDNA was used with 2 × Taq Master Mix (Vazyme, China), and 40 cycles of PCR cycling condition were performed. PCR products were examined by 1% agarose gel electrophoresis. PCR products from that amplified with divergent primers only from cDNA template were sent for Sanger sequencing by Sangon Biotech Co., Ltd.
SKOV3 and ES-23 cells at a density of 5 × 104 were seeded into a 96-well plate for 24 h, followed by incubation with 50 μmol/l EdU for 2 h at 37 °C. After fixation and permeabilization, the cells were incubated in DAPI solution and observed by fluorescence microscopy (Olympus, Tokyo, Japan).
To examine the stability of hsa_circ_0000144 in SKOV3 and ES-23 cells, 10 μg of total RNA was incubated with RNase R (40 U; Epicenter Biotechnologies, Madison, WI, USA). At 1 h post-digestion, the enrichment of hsa_circ_0000144 and GAPDH was analyzed using RT-qPCR.
The SKOV3 and ES-23 cells were transfected with the indicated plasmids. The medium in the lower chamber contained 10% FBS, and the cells (5 × 105) suspended in Matrigel were added to the upper chambers at the same time. The cells were incubated at 37 °C for 72 h. Cells that passed through the membrane were stained with methanol and 0.1% crystal violet and photographed. Cell invasion was quantified using direct microscopic visualization and counting.
Plasmid constructs carrying wild-type or mutant circ_0000144/ELK3 3UTR in the psiCHECK vector were co-transfected with miR-610 mimic or miR-NC into SKOV3 and ES-23 cells, respectively, using Lipofectamine 2000 (Thermo Fisher Scientific, USA). The cells were lysed 48 h after transfection, and luciferase activity was determined using the dual-luciferase reporter assay method. The PierceTM Renilla-Firefly Luciferase Dual Assay Kit (Thermo Fischer Scientific) was used to determine luciferase activity. Each sample was normalized by dividing the activity of the test firefly luciferase by the expression of the control Renilla luciferase. Both tests were performed in triplicate.
The migration of SKOV3 and ES-23 cells after transfection with plasmids, mimics, or inhibitors was assessed using a wound assay. The cells were inoculated in 6-well plates (1 × 106 cells), and the single cell layer was scratched with a 200 μL pipette tip. The cells were washed three times with PBS to remove cell fragments. RPMI-1640 without fetal bovine serum was added to each well and incubated at 37 °C and 5% CO2 for 48 h. Wound images were observed using a microscope at the same scratch location at 0, 24, and 48 h.
For the colony formation assay, SKOV3 and ES-23 cells (5 × 102) after transfection with plasmids, mimics, or inhibitors were seeded into 6-well plates. After 15-d routine culture, the generated colonies were fastened using 4% paraformaldehyde (Beyotime), stained with crystal violet (Beyotime), counted using ImageJ software (NIH, Bethesda, MD, USA), and photographed.
Based on the manufacturer’s protocol (Thermo Fisher Scientific), the Pierce Magnetic RNA-Protein Pull-Down Kit was used for the RNA pull-down assay. Cell protein extracts were mixed with the specifically biotinylated RNA probes to circ_0000144, and magnetic beads with streptavidin were added for 1 h. The pull-down of the mixture was monitored by RT-qPCR.
Radioimmunoprecipitation assay buffer was used to lyse the cells. An 8% SDS-PAGE gel was used to separate equal amounts of proteins, followed by transfer onto nitrocellulose membranes. The membranes were then blocked with 5% milk, followed by incubation with anti-ELK3 and anti-GAPDH antibodies (Santa, 1:1000). The membranes were then incubated with an HRP-conjugated secondary antibody. ECL reagents (Pierce, USA) were used to visualize proteins. The gray value of each band was analyzed using ImageJ (NIH, USA), with GAPDH as the endogenous reference.
Circular RNA interactome (https://circinteractome.nia.nih.gov/index.html) and TargetScan (http://www.targetscan.org/mamm_31/) were used to identify the target genes of circ_0000144 and miR-610, respectively.
Four-week-old female BALB/c nude mice were randomly divided into two groups (n = 6 per group). SKOV3 cells (5 × 106), which were transduced with circ_0000144 or NC vectors, were mixed in 150 μL of Matrigel. Then, the flanks of nude mice from different groups were handled by subcutaneous injection with a mixture. The equation for calculating the tumor volume was as follows: tumor volume = length × width × width/2. All animal experiments were performed in accordance with the relevant guidelines and regulations and were approved by First Hospital of Quanzhou Affiliated to Fujian Medical University.
SPSS17.0 software was adopted for the experimental data analysis. Data are expressed as mean ± SD, and Student’s t-test was used for between-group comparisons. Differences between groups were analyzed using one-way analysis of variance. Survival analysis was conducted using Kaplan-Meier survival analysis. Differences were considered statistically significant at p < 0.05. All statistical analyses were performed using GraphPad Prism 5 software (GraphPad Software, La Jolla, CA, USA).
To determine the expression profile, we measured the expression of circ_0000144 in tissues, serum, and cells. The results showed that circ_0000144 expression was significantly higher in tumor tissues than in normal samples (Fig. 1A). We also detected expression of linear SLAMF6 mRNA, which is the linear isomer of circ_0000144. Expression of SLAMF6 was significantly higher in tumor tissues (Fig. 1B). In addition, circ_0000144 and SLAMF6 expression was also notably elevated in the serum of patients with ovarian cancer compared to that in normal samples (Fig. 1C and D). Kaplan-Meier analysis showed that the overall survival (OS) rate and disease-free survival (DFS) rates of patients with high circ_0000144 expression were significantly shorter than those with low circ_0000144 expression (Fig. 1E and F). In addition, we found circ_0000144 and SLAMF6 upregulation in ovarian cancer cell lines (SKOV3, ES-2, and OVCAR3) compared to that in normal human ovarian cells (IOSE80) (Fig. 1G and H). In addition, distinct PCR products with the expected size were amplified using convergent and divergent primers. The circ_0000144 were validated by PCR amplification using divergent primers from cDNA, but not from gDNA, of SKOV3 and ES-2 cell lines (Fig. 1I). The back-splicing sites were verified using Sanger sequencing (Fig. 1J). Moreover, the expression level of linear mRNA, rather than circ_0000144, was decreased in SKOV3 and ES-2 cells after RNase R digestion and actinomycin D treatment, suggesting a stable structure of hsa_circ_0000144 (Fig. 1K and L). Collectively, circ_0000144 was significantly upregulated in ovarian cancer tissues, serum, and cells, hinting that circ_0000144 affected ovarian cancer development.
Circ_0000144 expression was the most upregulated in SKOV3 and ES-2 cells, as indicated in Fig. 1G; therefore, these two cell lines were selected for further study. To assess the biological function, circ_0000144 was overexpressed in the cells. qRT-PCR showed that circ_0000144 was upregulated in SKOV3 and ES-2 cells after transfection with circ_0000144 recombinant expression vector, but SLAMF6 showed no change in expression (Fig. 2A and B). To better investigate the function of circ_0000144, we performed an EdU assay to determine the function of circ_0000144 in SKOV3 and ES-2 cell proliferation. The results showed that circ_0000144 overexpression accelerated ovarian cancer cell proliferation (Fig. 2C). Migration and invasion capacities were also detected, and we found that circ_0000144 upregulation effectively increased the number of migrated and invaded SKOV3 and ES-2 cells (Fig. 2D and E). To test the biological role of circ_0000144 in vivo, nude mice were subcutaneously injected with SKOV3 cells transfected with pcDNA3.1-circ_0000144 or pcDNA3.1-NC. The results showed that the tumor volume in the circ_0000144 group was remarkably increased compared to that in the control group, indicating that circ_0000144 may promote ovarian cancer progression in vivo (Fig. 2F). Collectively, these findings indicate that circ_0000144 affects the occurrence and development of ovarian cancer, both in vitro and in vivo.
To explore the mechanisms by which circ_0000144 regulates ovarian cancer cell aggressive behavior, analysis of the bioinformatics tool circRNA interactome was carried out within SKOV3 and ES-2. Analysis of the circBank database revealed that circ_0000144 contains miR-610 binding sites (Fig. 3A). The results of the biotin-labeled RNA pull-down assay indicated that circ_0000144 could be pulled down by the biotin-miR-610 mimic rather than biotin-miRNA NC (Fig. 3B). In addition, qRT-PCR results showed that miR-610 was expressed at low levels in ovarian cancer tissues (Fig. 3C). Moreover, the expression level of circ_0000144 was significantly negatively correlated with miR-610 expression (Fig. 3D). A luciferase reporter assay was conducted using SKOV3 and ES-2 cells. Compared to miR-NC, the overexpression of miR-610 inhibited luciferase activity of wild-type circ_0000144 luciferase activity. After mutating the predicted binding site of circ_0000144, its inhibitory effect disappeared (Fig. 3E). In addition, transfection of circ_0000144 mutant plasmid did not affect the expression of circ_0000144, indicating binding sites mutation was not affect the circularization of circ_0000144 insert (Fig. 3). Subsequently, miR-610 expression increased after circ_0000144 knockdown in SKOV3 and ES-2 cells compared with si-NC (Fig. 3G). Taken together, these results indicated that miR-610 is a target of circ_0000144.
Rescue experiments were performed in SKOV3 and ES-2 cells by transfection with si-circ_0000144 + NC inhibitor or si-circ_0000144 + miR-610 inhibitor, si-NC, or si-circ_0000144 serving as the respective controls. The EdU assay showed that knockdown of circ_0000144 slowed the growth of SKOV3 and ES-2 cells, while the miR-610 inhibitor partially increased cell growth (Fig. 4A). Consistently, circ_0000144 knockdown decreased the colony formation capacity of cells, while the miR-610 inhibitor partially increased cell colony formation (Fig. 4B). Furthermore, migration and invasion of SKOV3 and ES-2 cells were suppressed after circ_0000144 knockdown, and co-transfection with miR-610 inhibitor partially increased these abilities (Fig. 4C and D). Taken together, these results suggest that circ_0000144 regulates ovarian cancer progression by sponging miR-610.
The bioinformatics tool TargetScan predicted that there is an miR-610 binding site within the ELK3 non-coding region (Fig. 5A). Luciferase reporter assays were first performed in SKOV3 and ES-2 cells to determine the miR-610 function on ELK3 and verify the targeting relationship. miR-610 overexpression enhanced ELK3-WT-luciferase activity relative to miR-NC. Nonetheless, after the predicted binding site of ELK3 was mutated, its promotional effect disappeared (Fig. 5B). Subsequently, we found that miR-610 overexpression could inhibit ELK3 protein expression levels in SKOV3 and ES-2 cells, and ELK3 protein expression increased after incubation with the miR-610 inhibitor (Fig. 5C), indicating that miR-610 could act as an miRNA sponge that competes with miRNA for binding to ELK3. Functionally, we tested the ability of ELK3 to induce the proliferation of SKOV3 and ES-2 cells. The results showed that ELK3 knockdown reduced cell proliferation compared to that in the NC group (Fig. 5D and E). In line with this, we found that the knockdown of ELK3 suppressed the migration and invasion of cells (Fig. 5F and G). Taken together, these findings indicate that miR-610 impairs ELK3 expression to block ovarian cancer development in vitro.
Expression analysis was performed in SKOV3 and ES-2 cells. As shown in Fig. 6, the expression of ELK3 impaired by transfection with si-circ_0000144 was partially strengthened by transfection with the si-circ_0000144 + miR-610 inhibitor, indicating that circ_0000144 downregulation could weaken the level of ELK3 by sponging miR-610.
Ovarian cancer, one of the most common cancers worldwide, remains an important human health problem that can lead to cancer-related threats. Thus far, the underlying regulatory mechanisms of ovarian cancer development remain unknown. Increasing evidence indicates that circRNAs contribute to elucidating the molecular mechanisms of cancer progression [21]. Circ_0000144 has been shown to participate in some types of cancer initiation and progression in previous studies [11, 13]. In the present study, we determined whether circ_0000144 enhances ovarian cancer tumorigenesis by regulating the downstream pathways. Since circRNAs are involved in regulating ovarian cancer carcinogenesis, studies on their roles have received increased attention in recent years, including circCSPP1 [22], circular RNA Cdr1as [23], circEXOC6B, and circN4BP2L2 [24]. In a previous study, circ_0000144 was reported to be upregulated in gastric cancer cells and tissues [13]. In our study, we found that circ_0000144 expression was also increased in ovarian cancer cells and tissues. In addition, circ_0000144 also showed increased expression in peripheral blood samples of patients in our study and was correlated with the survival rate of patients, suggesting that circ_0000144 can be used as a diagnostic biomarker for ovarian cancer. The following functional experiments presented that circ_0000144 upregulation accelerated ovarian cancer cells viability, colony formation capacity, migration and invasion, and promoted tumorigenesis. Based on these results, we found that circ_0000144 plays a carcinogenic role in ovarian cancer, suggesting that circ_0000144 may be involved in the occurrence and development of ovarian cancer and is expected to become a therapeutic target of ovarian cancer, but the mechanisms still need to be studied. It has been reported that miRNA sponge effects achieved by circRNA formation are now regarded as a general phenomenon in human malignancies [25]. Next, we analyzed the miRNAs known to be bound by circ_0000144 and identified miR-610 as a circ_0000144 associated miRNA. The sponge adsorption effect of circ_0000144 on miR-610 was further verified by dual-luciferase reporter gene and RNA pull-down assays. Growing evidence has shown that miR-610 dysregulation has been identified in different cancer types [20, 26, 27]. Nonetheless, the mechanisms by which miR-610 regulates the progression of ovarian cancer remain unknown. In this study, we found that miR-610 was downregulated in ovarian cancer tissues and negatively correlated with circ_0000144 in ovarian cancer tissues. In addition, low expression of miR-610 can promote proliferation, migration, and invasion of ovarian cancer cells. Further functional studies showed that miR-610 inhibitors functionally restored the ability of circ_0000144 knockdown in ovarian cancer cells. These results suggest that circ_0000144 regulates ovarian cancer progression by acting as an miR-610 sponge. Together with miRNAs and their targets, the circRNA-miRNA-mRNA axis may function as an extensive regulatory network in gene expression, and their dysregulation may cause disease progression, including cancer development [28]. In the present study, we found that ELK3 was co-overexpressed with circ_0000144 in ovarian cancer cells. ELK3 has been reported to participate in cancer genesis and development, including gastric [29], breast [17], and liver cancers [30]. Although previous studies have also demonstrated that ELK3 expression is altered in ovarian cancer cell lines and tumors through overexpression of miR-378 [19], the regulatory mechanism by which ELK3 regulates ovarian cancer remains unclear. According to our findings, ELK3 was predicted to be a direct target of miR-610 by bioinformatics analysis. Moreover, knockdown of ELK3 inhibited proliferation, migration, and invasion of ovarian cancer cells. This phenomenon is similar to that observed for circ_0000144 in ovarian cancer cells. We further demonstrated that miR-610 suppressed ELK3 expression and that circ_0000144 could promote ELK3 expression by sponging miR-610 in ovarian cancer cells, which enhanced ovarian cancer cell proliferation, migration, and invasion. Therefore, we speculate that circ_0000144 acts as a sponge for miR-610 to enhance ELK3 expression and inhibit ovarian cancer cells. Therefore, the circ_0000144/miR-610/ELK3 network may promote a new treatment strategy for patients with ovarian cancer. Although our study demonstrated the relationship between circ_0000144 and clinicopathological features of ovarian cancer, the present work still has some limitations, including the lack of a deeper understanding of the underlying molecular mechanism between miR-610 and circ_0000144, and functional experiments were only performed in the cell line. Therefore, further studies in nude mice are required to confirm our conclusions.
This study provides the first evidence that circ_0000144 is significantly upregulated in ovarian cancer cells, tissues, and serum samples. Meanwhile, circ_0000144 overexpression promotes ELK3 protein expression through the sponging of miR-610, causing ovarian cancer cell proliferation, migration, and invasion. In addition, circ_0000144 overexpression markedly accelerated ovarian cancer tumorigenesis in a mouse xenograft model. These data indicate that circ_0000144 is a novel candidate therapeutic biomarker for ovarian cancer and a specific clinical diagnostic and prognostic biomarker and therapeutic target for ovarian cancer. | true | true | true |
|
PMC9569155 | Gonzalo Vidal,Carlos Vidal-Céspedes,Macarena Muñoz Silva,Carlos Castillo-Passi,Guillermo Yáñez Feliú,Fernán Federici,Timothy J Rudge | Accurate characterization of dynamic microbial gene expression and growth rate profiles | 15-10-2022 | Inverse problem,characterization,dynamical systems,web application,gene expression | Abstract Genetic circuits are subject to variability due to cellular and compositional contexts. Cells face changing internal states and environments, the cellular context, to which they sense and respond by changing their gene expression and growth rates. Furthermore, each gene in a genetic circuit operates in a compositional context of genes which may interact with each other and the host cell in complex ways. The context of genetic circuits can, therefore, change gene expression and growth rates, and measuring their dynamics is essential to understanding natural and synthetic regulatory networks that give rise to functional phenotypes. However, reconstruction of microbial gene expression and growth rate profiles from typical noisy measurements of cell populations is difficult due to the effects of noise at low cell densities among other factors. We present here a method for the estimation of dynamic microbial gene expression rates and growth rates from noisy measurement data. Compared to the current state-of-the-art, our method significantly reduced the mean squared error of reconstructions from simulated data of growth and gene expression rates, improving the estimation of timing and magnitude of relevant shapes of profiles. We applied our method to characterize a triple-reporter plasmid library combining multiple transcription units in different compositional and cellular contexts in Escherichia coli. Our analysis reveals cellular and compositional context effects on microbial growth and gene expression rate dynamics and suggests a method for the dynamic ratiometric characterization of constitutive promoters relative to an in vivo reference. | Accurate characterization of dynamic microbial gene expression and growth rate profiles
Genetic circuits are subject to variability due to cellular and compositional contexts. Cells face changing internal states and environments, the cellular context, to which they sense and respond by changing their gene expression and growth rates. Furthermore, each gene in a genetic circuit operates in a compositional context of genes which may interact with each other and the host cell in complex ways. The context of genetic circuits can, therefore, change gene expression and growth rates, and measuring their dynamics is essential to understanding natural and synthetic regulatory networks that give rise to functional phenotypes. However, reconstruction of microbial gene expression and growth rate profiles from typical noisy measurements of cell populations is difficult due to the effects of noise at low cell densities among other factors. We present here a method for the estimation of dynamic microbial gene expression rates and growth rates from noisy measurement data. Compared to the current state-of-the-art, our method significantly reduced the mean squared error of reconstructions from simulated data of growth and gene expression rates, improving the estimation of timing and magnitude of relevant shapes of profiles. We applied our method to characterize a triple-reporter plasmid library combining multiple transcription units in different compositional and cellular contexts in Escherichia coli. Our analysis reveals cellular and compositional context effects on microbial growth and gene expression rate dynamics and suggests a method for the dynamic ratiometric characterization of constitutive promoters relative to an in vivo reference.
Gene expression and growth are subject to variation due to changes in environmental and internal conditions, which may be divided into cellular and compositional contexts. The cellular context is related to the cell or chassis itself such as the strain and to operational conditions such as media and carbon source. Microbial populations themselves create intrinsically dynamic conditions which can be separated into distinct phases of the growth cycle. An initial lag phase of adaptation is followed by exponential growth and finally a minimal growth stationary phase (1). These transitions in growth phases are driven by cell internal changes caused by depletion of nutrients, accumulation of waste products or due to biochemical and physical signals (1). These signals can be part of the environment that the cell senses and responds to and can be experimentally determined. Cellular context changes the metabolic activity of the cell and leads to distinct phenotypes (2). The effects of changes in external factors on growth and gene expression levels have been extensively studied (3–5) and correlations between them established. For example, ribosome and RNA polymerase are positively correlated with peak growth rate in different media (6). However, changes in the dynamics of gene expression due to cellular context are less clear (7), since different genes are known to be expressed differentially in each growth phase. This is at least partly due to the regulation of distinct sigma factor RNA polymerase sub-units, such as σ70 which peaks in exponential growth phase (8), as well as potentially gene-specific transcription factors, ribosome numbers and other translation factors (9). The compositional context is related to the composition or sequence of the inserted DNA and can be seen at the transcription unit (TU), plasmid or genomic level. Synthetic biologists separate DNA into standard parts with determined sequences and function (10). A composition of parts that enables transcription is a TU, and the transcribed RNA can be then translated into proteins. A TU capable of gene expression can be assembled with the parts promoter (Pro), ribosome-binding site (RBS), coding sequence (CDS) and terminator (Ter) in that exact composition order. The behavior of these parts varies depending on the surrounding sequences. While the magnitude of gene expression from a TU is known to be influenced by its surrounding sequences, orientation or its position relative to other parts in a plasmid or in the genome, their effect on dynamics is not well understood (11–14). Measuring and analyzing the dynamics of gene expression are thus fundamental to understanding cellular regulation by both natural and synthetic gene networks in the face of different cellular and compositional contexts. Typical experiments to measure gene expression rates in bacteria and other microorganisms utilize fluorescent reporters to track the expression levels of lineages of cells (7,15–17). The total biomass of these lineages is also tracked, typically using optical density measurements or colony size (18, 19). The genes measured are often fusions of promoters of interest with a downstream fluorescent reporter, and their expression rate profiles are taken to be indicative of the transcription rate of the promoter (20, 7). Promoters have been characterized relative to a standard promoter (20–22) that is measured under the same conditions as the promoter of interest. This approach has been used under steady-state conditions but has not been applied to dynamics. Thus, there is a need for methods to characterize the dynamics of gene expression and growth as phenotypic parameters. Reconstructing dynamic microbial gene expression and growth rate profiles from data is difficult because, particularly at low biomass, measurements suffer from significant noise. Typical methods involve data smoothing and differentiation of the resulting signal—herein referred to as the indirect method (23–26). We consider two data smoothing filters; an anti-causal zero-phase digital filter (27) and the Savitzky–Golay polynomial filter (28). The indirect methods are sensitive to noise leading to the development of a more robust method based on the linear inversion of differential Equation models (29)—herein referred to as the direct method. Inverse problems present an approach to infer the values of parameters or functions on which measurements depend (30). The basic requirements to solve them are appropriate measurements and a mathematical model of the process that generates them. Two difficulties in inverse problems are that it cannot determine correlated parameters and that most problems of interest are ill-posed. An ill-posed problem is one that does not satisfy one or more of the well-posed properties: a solution exists, it is unique and its behavior changes continuously with the initial conditions. However, regularization methods, basis transformations and constraints can often be used to transform ill-posed into similar well-posed problems. We present a method for the reconstruction of gene expression and growth rate profiles using inverse problems that achieves several times lower error than the direct and indirect methods. Our approach more accurately reproduced features of dynamic gene expression and growth rate profiles, including the lag phase and peak rates. Using this method, we characterized the dynamics of a collection of synthetic TUs in different cellular and compositional contexts relative to an in vivo reference, revealing uncertainty due to the genetic composition and external environmental factors. Our results suggest an approach to ratiometric characterization of the dynamics of gene expression.
The rate at which a protein is synthesized by a genetic network or circuit varies over time, giving rise to dynamic gene expression rate profiles, which may include rich behaviors such as bistability (31) and oscillations (32). To see the importance of measuring these gene expression rate profiles, consider a genetic circuit composed of regulatory proteins. In the typical case of short half-life mRNAs, we may assume quasi-steady state and, for a genetic circuit with N proteins at concentrations pi, with degradation rates γi, in cells growing at rate . The functions ϕi define the interactions in the circuit by mapping the protein concentrations to protein synthesis or gene expression rates. The dependence on time is due to variation in the cellular context such as growth phase transitions, which cause observable changes in the growth rate profile . It is, therefore, essential for the analysis of genetic circuit operation to estimate the gene expression rate profiles ϕi and growth rate profile µ, allowing the parameterization of models such as those in Equation 1. The simplest genetic circuits consist of constitutive TUs which do not regulate each other, in which case is only a function of time. To model the growth and reporter gene expression measurements from such a circuit in a population of cells, we use the following equations: where B is a measure of sample biomass, is the instantaneous relative growth rate, yi is the intensity of reporter, is the instantaneous expression rate and γi is the reporter degradation rate for TU i. Here we have assumed that reporter intensity , with pi the protein concentration per biomass. While this is a reasonable assumption in constant conditions (33), this relationship may not always hold leading to inaccuracy in the reconstructed profiles, which is a general problem with gene expression reporters. More complex measurement models might be constructed, given sufficient information about the transformation of gene expression level into reporter intensity. We leave this for future work. From these equations we wish to accurately and robustly estimate the growth rate and the expression rates , given an estimate of the reporter degradation rates γi. Typically, reporter proteins are stable and so in the following we will use (34) (see Supporting Information, Figure S1 for the effect of underestimating γi). Note that the expression rate is different from the rate of change of fluorescence and from the rate of change of fluorescence concentration , which both depend on protein degradation and dilution due to growth (and may be negative). What is being estimated in this work is the underlying fluorescent protein synthesis rate (strictly positive), which is independent of dilution and degradation processes and which we assume is proportional to the underlying cellular gene expression rate (see Discussion).
Reconstructing the functions and represents an inverse problem, which is underdetermined and ill posed (30). In order to reduce the dimensionality of the inverse problem, we exploit prior knowledge of the functions and to construct a simple basis as follows. Expression rates and growth rates may be reasonably assumed to be strictly positive and smooth on typical timescales of transcription and translation. We propose the following approximation, given a function f(t) that meets our assumptions, with which represents a sum of n Gaussian curves Gk with weight , variance Δ and regularly spaced over time t at intervals Δ. Here Δ determines the timescale of variation or smoothness of the representation of the function f(t). Since here we consider bulk culture experiments, gene expression bursting and noise in growth cannot be observed, and the relevant timescales are the gene expression reporter half-life and the culture doubling time, typically <1 h. Choosing Δ greater than the sampling interval of the data makes the system over-determined and regularized in the sense that it is constrained to be smooth. In the following we choose h, the typical timescale of protein synthesis, which is larger than the usual sampling interval of 10–15 min (see Methods). In Figure 1A and B we can see examples of reconstructed growth rate and gene expression rate profiles in turquoise and the function to be fitted in dashed black. The effect of this approximate Gaussian basis can be seen from the dependence on Δ of the maximum slope of the basis functions , which scales as . This means that the sharpest change in expression rate and growth rate profiles that can be reconstructed by our method is determined by Δ (see Supporting Information, Figure S2 for effect of Δ on reconstruction error). The model given in Equations 2 and 3 combined with the approximation of Equations 4 and 5 represents the forward models of the inverse problems for the reconstruction of and . In practice, the measurements used to estimate B and y are discrete, will contain background signal and are subject to noise. The background signals Bʹ and yʹ are typically estimated by measuring appropriate control samples containing no cells (Bʹ) and cells with no reporter expression (yʹ) (see Methods). After subtracting these background measurements, we are left with the noisy estimates and . We then wish to parameterize the forward models given by Equations 2 and 3 such that and are minimized. Note that while this approach uses a smooth basis to represent the reconstructed profiles, it does not filter or smooth the input data as in the indirect method. Here we approximate the growth and gene expression rate profiles using a superposition of Gaussian functions to represent a continuous function as a discrete vector of parameters. The measurements and the estimated parameters are inputs of the forward model. The forward model then generates simulated measurements using different estimated parameters, and we compute the ones that minimize the difference between the measurements and the model.
The inverse problem for the reconstruction of the growth rate can be stated as, with, We minimize the difference between noisy estimates of Biomass and the result of the forward model . Since this problem is ill-posed we regularize using a Tikhonov penalty term λ. Θ is the vector of parameters to optimize, containing the initial biomass B0 and the Gaussian basis weights to represent . The hyperparameter λ was chosen to minimize error within a reasonable range of values (see Supporting Information, Figure S2). This problem is a nonlinear least squares optimization, which we solve using the trust region reflective algorithm (35). We generated simulated data from Equations 2 and 3 using 100 randomly parameterized Gompertz growth models (1) and three different levels of measurement noise and characterized the growth rate using the inverse, direct and indirect methods (Supporting information Figures S3–S5). Hyperparameters for each method were optimized to minimize error within a reasonable range of values (see Supporting Information, Figure S2). We compare our results to the direct linear inversion method and show that our approach reduces mean squared error by more than 29-fold (, Welch’s t-test) (Figure 2A). Then we compared to the indirect method and show that the inverse method reduced the mean squared error by >2-fold (, Welch’s t-test) (Figure 2A). Finally, we compared to a different indirect method that uses an anti-causal zero-phase digital filter and show that the inverse method reduced the mean squared error by more than a 1000-fold (Supporting information Figure S17). The inverse method maintained the best performance at high noise levels. Furthermore, our method is more robust to noise in early biomass measurements, where the direct method overestimates the initial values and the indirect produces noisy reconstructions that usually goes below zero at the beginning (Figure 2B). The inverse method correctly reconstructed the lag phase, which is missing from the linear inversion solution. Our method also reconstructed better the growth rate peak, effectively distinguishing between lag, exponential and stationary growth phases (Figure 2C and D).
In a similar way to growth rate, we wish to find the optimal parameters, where Θ is the vector of parameters to optimize, containing the initial reporter intensity y0 and the Gaussian basis with weights and the forward model . The problem is again a nonlinear least squares optimization, which we solve using the same numerical procedure as for growth rate. To test this approach, we generated 100 random gene expression rate profiles from smoothed lognormal random walks, with random Gompertz models for the biomass, and three measurement noise levels (Supporting information Figures S6–S8) (see Methods). Again, we compared the inverse method to the direct linear inversion method and show that the mean squared error is more than 4-fold lower (, Welch’s t-test) and close to 3-fold lower than the indirect method ( Welch’s t-test) (Figure 3A). Finally, we compared to a different indirect method that uses an anti-causal zero-phase digital filter and show that the inverse method reduced the mean squared error by more than a 100-fold (Supporting information Figure S17). The gene expression simulations are not constrained to start with low values or to have an initial peak which make the three methods have similar initial errors (Figure 3B). We find that the direct method does not correctly reconstruct early peaks in gene expression rate profiles and that the indirect method produces extremely noisy solutions (Figure 3C and D). The inverse method is the best performer capturing the shape of the growth rate profile, although this method suffers with sharp changes due to its dependence with Δ (Figure 3C and D). The inverse method requires knowledge of the timescale of the process that you want to measure to set the hyperparameter Δ (see Supporting information).
Using the inverse, direct and indirect methods we reconstructed growth and gene expression dynamics from experimental data of Escherichia coli carrying a synthetic triple TU plasmid pAAA (Figure 4A and B). This plasmid contains three TUs with the same synthetic σ70 constitutive promoter J23101 (36) in different plasmid compositional contexts determined by its position in the plasmid and in different TU compositional contexts determined by different sets of promoter downstream elements RBS-CDS-Terminator (Figure 4A). Each CDS encoded a different fluorescent protein as reporter. Assays were performed using a 96-well microplate reader to measure fluorescence in each reporter channel as a proxy for protein concentration and optical density as a proxy for biomass (see Methods). To assess the effects of cellular context on growth and gene expression dynamics we measured two strains of E. coli carrying the pAAA plasmid, growing on two different carbon sources (26) (Figure 4B–E) (see Supporting information, Figure S9 for an example of the raw fluorescence and OD 600 data). The inverse method captured the shape of the growth and gene expression rate profiles in a smooth way, consistent with population averages on the timescale of protein synthesis, while the direct and indirect methods produced noisy solutions. Our method reconstructed a peak in growth rate consistent with the transition between lag, exponential and stationary phases, as in the Gompertz growth model (37). Furthermore, it captured a peak in the expression rate coincident with the growth rate peak, which is consistent with promoter dependence on σ70 and the abundance of this factor, as well as ribosomes, during peak growth (Figure 4B–E) (6, 9). The gene expression rate profiles were different for TUs in the same plasmid under the control of the same promoter due to both different cellular and compositional contexts (Figure 4B–E). While in all cases peak gene expression rate coincided with peak growth rate, in some contexts multiple peaks were observed. The cellular context (strain and carbon source) as well as the compositional context (promoter downstream elements and position and orientation in the plasmid) clearly change gene expression and growth dynamics, leading to different peak timing and overall shape. All growth rates characterized using the inverse method exhibited clear lag, exponential and stationary phases, which are not apparent with the direct method (Figure 4B–E). The timing of growth phase transitions was different in each cellular context (Figure 4B–E).
Gene expression magnitude has been characterized relative to a standard in vivo reference containing promoter J23101, in order to normalize for cellular context (22). Plasmid pAAA provides such a reference, with three TUs containing the same J23101 promoter in different compositional contexts. We hypothesized that this reference plasmid could be used to characterize the dynamics of gene expression in a standardized fashion. Each TU in the pAAA plasmid presents a standard reference for a particular compositional context—the promoter downstream elements, position and orientation in the plasmid. We then wish to characterize TUs with arbitrary promoters relative to these reference TUs, allowing us to describe their dynamics in a concise way. In order to compare gene expression rate dynamics from different experiments we synchronized profiles and normalized each one by subtracting its mean and dividing by its standard deviation. The timing of growth phase transitions is variable due to differences in initial conditions and experimental variability, which leads to differences in the timing of gene expression rate profiles. In order to correct for these differences, we used the reconstructed growth rate peak time t0 to synchronize the expression rate profiles, shifting time to , such that τ = 0 is the time of peak growth rate. We tested this approach on a collection of 14 combinatorial three-reporter plasmids, combining 10 different TUs which were each driven by one of seven promoters (38). Each plasmid contained three TUs producing red fluorescent protein (RFP), yellow fluorescent protein (YFP) and cyan fluorescent protein (CFP), containing the same promoter downstream elements as the corresponding reference TU. The promoter downstream elements RBS, CDS and terminator for each reporter were maintained constant and we refer to them using the reporter name (Supporting information Tables S1 and S2). The CFP TU was maintained the same in all plasmids, to serve as a control (22, 20). This collection of plasmids presented a variety of promoters in different TU compositional contexts, that is, with different promoter downstream elements. Each TU was assembled into multiple plasmid compositional contexts, that is, in the presence of different upstream or downstream TUs. Three of the TUs in the collection contained a promoter from a family of constitutive promoters created by mutating a consensus sequence (36), which includes the reference TU promoter, J23101. The RFP reference TU matched the gene expression rate profile shape of the RFP TU with the constitutive promoter J23106 over different downstream TU contexts (Figure 5A). The YFP reference TU matched the gene expression profile shape of the YFP TU with the constitutive promoters J23107 (Figure 5B) and J23101 (Supporting information Figure S10B) over different upstream TU contexts. The CFP TU was consistent across all compositional contexts (Supporting information Figure S10A). These results suggest that the dynamics are not affected by the tested plasmid compositional contexts. Two of the TUs contained promoters that are repressible by a transcription regulating protein that binds to the promoter. In the case of the TetR-responsive TU (containing promoter R0040), the RFP reference TU matched the gene expression rate profile over different downstream TU contexts and over different cellular contexts (Figure 5C). This is in spite of the fact that the genome of the strain MG1655z1 contains a constitutive TetR gene, while Top10 does not produce the protein. This result suggests that the dynamics of the TetR-repressible TU are not affected by the action of the repressor. The other repressible TU contained promoter R0011, regulated by LacI. LacI is produced by both strains and partly regulated by cAMP in Top10 (39). The YFP reference TU matched the LacI-repressible YFP TU gene expression rate profiles in M9-glycerol but did not match them in M9-glucose (Figure 5D). In MG1655z1 growing on glucose gene expression rates became negatively correlated with growth rate, and in Top10 on glucose they exhibited a second peak. These results suggest that transcription regulation can significantly affect the dynamics of gene expression compared to constitutive expression profiles. The remaining two TUs contained promoters activated by different one-component signaling systems, measured in the absence of signal and regulator to study their basal expression. The RFP TU responsive to C6 homoserine lactone, containing promoter pLux76, was consistent with the constitutive reference TU in all contexts (Supporting information Figure S10C). However, for the YFP TU responsive to C12 homoserine lactone, containing promoter pLas81, the gene expression rate profile inverted its correlation with growth rate when downstream of the TU containing promoter R0040 and growing on glucose in strain MG1655z1. All the magnitudes of these measurements can be found in Supporting information Figure S13. These results show the uncertainty that can be introduced in gene expression rate dynamics due to changing compositional and cellular context (Supporting information Figure S10D).
The accurate characterization of dynamic gene expression and growth rate profiles is essential for the characterization of genetic circuits and the inference of gene regulatory interactions in natural networks (40–42). We have demonstrated an inverse problem approach to reconstructing dynamic gene expression rate and growth rate profiles from noisy kinetic measurement data. We compared our method to the current state-of-the-art algorithms, direct linear inversion (29) and indirect smoothing and differentiation (25). Our approach reduced the mean squared error of reconstructions from simulated data of growth rate by almost 30-fold and gene expression rate by more than 4-fold with respect to the direct method. The comparison showed that the direct method often fails to capture peaks at the beginning of the profiles. Indirect methods, even after filtering the noise, fail to reconstruct early stages of growth. This is likely because the biomass can have very low and even negative values after background correction and dividing by these values can result in the amplification of noise. This highlights that particular attention should be paid to reducing noise from experimental procedures and measurement techniques since all methods attain lower reconstruction errors with less noise. Surprisingly, our indirect method often performed better than the direct method, but our inverse problems approach improved on the indirect method by almost 3-fold for gene expression rates, and more than 2-fold for growth rate. Furthermore, we were able to reconstruct features of both growth rate and gene expression rate profiles, such as exponential phase and peak growth, that were not apparent from the direct linear inversion method nor the indirect method. The growth rate peak is an important feature captured better with our method, allowing the synchronization of gene expression rate profiles. While in terms of computation time our method is relatively slow, it is the most intuitive to adjust by estimating the timescale of dynamics to obtain Δ. The indirect methods require knowledge of the signal processing filters used, and the direct method requires tuning the ‘insignificant value’ which is rather obscure. Using our method we showed that the dynamic form of gene expression rates, not only their magnitude, is determined both by cellular and compositional contexts (Figure 5, Supporting information Figures S10 and S13–S15). We examined two types of compositional context. Firstly, the composition of parts within a TU, not only the sequence of the promoter, determined the dynamics of gene expression rates. Secondly, gene expression rate profiles were largely independent of the context in which the TU was placed, that is the upstream and downstream TUs. In most cases, gene expression rates peaked in exponential growth phase, with some promoters exhibiting a second peak in stationary phase. This may be an effect of the abundance of different sigma factors in each growth phase and the sensitivity of the promoter to them. It is known that the binding sites for σ70, most abundant in exponential growth phase, and σS, most abundant in stationary growth phase, are very similar (43). Therefore, a promoter with peaks in both exponential and stationary phases may be activated by both σ70 and σS to different extents and the peaks caused by variations in the sigma factor abundance in each growth phase. In some cases gene expression rate profiles inverted their correlation with growth rate, highlighting the uncertainty introduced by changing circuit composition. This uncertainty may be due to various mechanisms that modulate gene expression at sequence level such as DNA supercoiling (12). Furthermore, our results also showed that cellular context, that is media and strain, changed dynamic gene expression and growth rate profiles. Our results suggest that while the dynamic characterization for TUs under constitutive or leaky expression relative to an in vivo reference could be useful, uncertainty due to compositional and cellular context must be taken into account. This highlights the need for strategies to mitigate compositional context effects, such as gene expression load (44), five prime untranslated region (45) as well as techniques to predict interactions of genetic elements at the sequence level (11,21,46). Typically, measurements of promoter–reporter fusions are used as a proxy for transcription rates under various levels of transcription factors, external signals and other determinants of gene expression (20,21,47,48). However, applying our method to multiple reporter TUs with the same promoter, we showed that gene expression rate profiles should not be taken as indicative of intrinsic characteristics of promoters, since they are affected by both the promoter and downstream genetic elements, gene position and orientation and external factors such as carbon source and host strain (see Supporting information Figure S16). Synthetic biology aims to design novel genetic networks or circuits from compositions of transcription units. It relies heavily on the characterization of the functions of these TUs. Fundamentally, gene expression rates must be reconstructed from noisy measurement data in a range of conditions, including concentrations of inducer chemicals. The function of the circuit may then be mathematically modeled as in Equation 1. Methods such as ours will enable such approaches for dynamical systems (32, 31) where the dynamic profile of gene expression rates is essential to the operation of the circuit. In order to model circuit operation in this way, calibration of the fluorescence and biomass signals with respect to standard references (15, 18) or the use of relative (ratiometric) quantification (21, 20) are necessary and can be easily incorporated into our workflow. Our results show that profiles for constitutive gene expression were consistent across a range of promoters in a range of contexts suggesting that an in vivo reference gene may be used to infer the expression profiles of other genes in a circuit without directly measuring them. The inverse problems approach provides a framework that could be extended easily to fit more complex models of gene expression and also for regulatory parameters (e.g. Hill functions) as well as dynamic profiles, providing an accurate and flexible characterization method for synthetic biology.
Kinetic gene expression and growth assays were made culturing triple reporter plasmid pAAA containing bacteria (Figure 4A) using two different strains as well as two different carbon sources, as described in Flapjack (26). Monoclonal colonies of E. coli strain TOP10 or MG1655z1 transformed with plasmid DNA were picked and cultured for 14–15 h overnight in M9 media with 50 g/ml kanamycin, 0.2% w/v casaminoacids and 0.4% w/v glucose or 0.4% w/v glycerol. Overnight cultures were diluted 1000 times in 2-ml tubes. All the tubes were filled with 1996µ l of fresh M9 media, 2µ l of kanamycin and 2µ l of the bacteria liquid culture obtaining a final volume of 2 ml. In each well of a 96-well plate were added 200 µ l, 4 wells with M9 media with the proper carbon source and kanamycin, 4 wells with non-transformed bacteria of the same strain and 10 wells of bacteria transformed with the appropriate plasmid to analyze from the previously prepared 2-ml tubes. Optical density and fluorescence in three channels (RFP, YFP and CFP) were measured approximately every 15 min for 24 h in a Synergy HTX plate reader with Gen5 software. Each assay was repeated on three different days, with 10 replicates on each day, and each 96 well plate contained experiments with the same carbon source and strain following the methods in Flapjack (26).
Computational methods were performed using Flapjack (26), Python (49), Numpy (50), Scipy (51), Pandas (52), Matplotlib (53), Plotly (54), Jupyter (55), Google Colaboratory (56) and Matlab. The direct method was computed using the WellFARE package (29). The indirect method with anti-causal zero-phase digital filter was computed using the Matlab function ‘butter’ (27).
Flapjack (26) (http://flapjack.rudge-lab.org) was extended to compute gene expression rate and growth rate profiles using the methods described above, using the WellFARE package (29) for the direct method. The indirect method is computed by Flapjack by filtering the measured signals (biomass and reporter levels) using a Savitzky–Golay filter (28) and then differentiating the resulting smooth polynomial interpolation. Flapjack is a systems and synthetic biology data storage and analysis tool, built as a web app that provides a user-friendly web interface and a representational state transfer (REST) and web socket application programming interface (API). The system allows upload of kinetic gene expression data from a variety of sources and links it to metadata about experimental conditions. These data may then be queried and filtered and used to reconstruct gene expression and growth rates. Flapjack automatically subtracts background signal from both reporter and biomass measurements, taking the average of control samples (untransformed cells or media with no cells) at each time point.
In order to generate a range of gene expression rate profiles, with minimal assumptions about their form, we generate lognormal random walks as, with and . The profiles ϕt were then smoothed using a second-order Savitzky–Golay filter (28) with window size 21 and normalized to . To generate random growth rate profiles, we used the Gompertz equation, with the maximal growth rate uniformly distributed on per hour, λ the lag phase length uniformly distributed in hours, , where the maximal biomass and minimum biomass . The growth rate profile implied by this equation is given by,
Equations 2 and 3 were solved using the forward Euler integration scheme with time step h for a period of 24 h. Noise and background were added according to the following equations with and , where ϵt and ζt are uncorrelated white noise with variance σ2, due to the measurement process. Simulated measurements were generated using LOICA (57), then uploaded to Flapjack (26) and analyzed using the API via Python (see Supporting information Figures S11 and S12 for an example of reporter and biomass raw data, respectively).
Each of the methods tested in the main text is dependent on one or more hyperparameters. In the case of the direct method this is the so-called insignificant value ϵL (29), for the indirect method it is the Savitzky–Golay filter window size and for the inverse method they are Δ and λ. For reconstructions of simulated data these parameters were optimized by scanning a range of reasonable values and choosing the parameter that minimized the mean squared error (see Supporting information). For experimental data, the value of λ was chosen using the L-curve method (58), ϵL was taken from the original paper (29), the value of Δ was fixed at 1 h and the Savitzky–Golay window size fixed at 11. For the anti-causal zero-phase digital filter, a second-order Butterworth filter (27) with cut-off frequency 4/33 was used.
Click here for additional data file. | true | true | true |
|
PMC9569213 | Jinghua Li,Junyi Liu,Yuanping Zhang,Xu Zha,Hong Zhang,Yongying Tang,Xueying Zhao | miR-181d-5p Protects against Retinal Ganglion Cell Death after Blunt Ocular Injury by Regulating NFIA-Medicated Astrocyte Development | 08-10-2022 | Background Traumatic optic neuropathy (TON) refers to damage to the optic nerve resulting from direct and indirect trauma to the head and face. One of the important pathological processes in TON is the death of retinal ganglion cells (RGCs), but the cause of RGCs death remains unclear. We aimed to explore the mechanisms of RGCs death in an experimental TON model. Methods Optic nerve crush injury was induced in ten New Zealand white rabbits. On the 1st, 3rd, 7th, 14th, and 28th days after the operation, the retinal tissues of the rabbits were observed pathologically by hematoxylin-eosin staining. The expression of POU-homeodomain transcription factor Brn3a and glial fibrillary acidic protein (GFAP) was measured by immunofluorescence to evaluate the number of RGCs and astrocytes, respectively. miRNA expression and protein levels were assessed by RT-qPCR and western blot methods, respectively. Finally, the malondialdehyde content, superoxide dismutase activity, and proinflammatory factor levels were measured by ELISA. Western blot and dual-luciferase reporter assays were used to elucidate the relationship between miR-181d-5p and nuclear factor I-A (NFIA). Results Blunt ocular trauma increased oxidative stress and apoptosis and reduced ganglion cell layer (GCL) density. The expression of miR-181d-5p was decreased in retinal tissues, and its overexpression relieved RGCs death, astrocyte development, oxidative stress, and inflammation of the retina, which were reversed by NFIA overexpression. Conclusion miR-181d-5p can protect against the deterioration of TON by inhibiting RGCs death, astrocyte development, oxidative stress, and inflammation by targeting NFIA. This study provides new insight into early medical intervention in patients with TON. | miR-181d-5p Protects against Retinal Ganglion Cell Death after Blunt Ocular Injury by Regulating NFIA-Medicated Astrocyte Development
Traumatic optic neuropathy (TON) refers to damage to the optic nerve resulting from direct and indirect trauma to the head and face. One of the important pathological processes in TON is the death of retinal ganglion cells (RGCs), but the cause of RGCs death remains unclear. We aimed to explore the mechanisms of RGCs death in an experimental TON model.
Optic nerve crush injury was induced in ten New Zealand white rabbits. On the 1st, 3rd, 7th, 14th, and 28th days after the operation, the retinal tissues of the rabbits were observed pathologically by hematoxylin-eosin staining. The expression of POU-homeodomain transcription factor Brn3a and glial fibrillary acidic protein (GFAP) was measured by immunofluorescence to evaluate the number of RGCs and astrocytes, respectively. miRNA expression and protein levels were assessed by RT-qPCR and western blot methods, respectively. Finally, the malondialdehyde content, superoxide dismutase activity, and proinflammatory factor levels were measured by ELISA. Western blot and dual-luciferase reporter assays were used to elucidate the relationship between miR-181d-5p and nuclear factor I-A (NFIA).
Blunt ocular trauma increased oxidative stress and apoptosis and reduced ganglion cell layer (GCL) density. The expression of miR-181d-5p was decreased in retinal tissues, and its overexpression relieved RGCs death, astrocyte development, oxidative stress, and inflammation of the retina, which were reversed by NFIA overexpression.
miR-181d-5p can protect against the deterioration of TON by inhibiting RGCs death, astrocyte development, oxidative stress, and inflammation by targeting NFIA. This study provides new insight into early medical intervention in patients with TON.
Traumatic optic neuropathy (TON) refers to optic nerve damage that results from direct or indirect trauma to the head and face; this condition is rare but can cause severe and irreversible vision loss [1]. TON treatment is mainly divided into three methods: high-dose steroid therapy, surgical decompression, and combination steroid and optic nerve decompression therapy [1, 2]. Unfortunately, the reliance on these methods has been controversial [3, 4]. On the one hand, although complications of steroid therapy are rare, there is no obvious evidence that steroid therapy has benefit in improving vision in patients with TON [2]. On the other hand, surgical treatment of the optic nerve carries a clear risk of complications, such as postoperative cerebrospinal fluid leakage and meningitis, and there is no evidence that surgery produces any additional benefit [5]. Therefore, it is crucial to discover a new and effective therapeutic strategy for TON. Retinal ganglion cells (RGCs) axons and supporting cells make up the optic nerve (ON) cells [6]. Because they are central neurons, RGCs lack endogenous regenerative capacity. Therefore, RGCs undergoing apoptosis cannot be replaced, and damaged ON cells cannot be regenerated, leading to irreversible blindness [7]. Trauma leads to a series of pathological events, including inflammation [8] and oxidative stress [9]. These risk factors are closely related to the apoptosis of RGCs. Research has found that RGCs death is associated with TON, and resveratrol treatment can delay the loss of RGCs and the loss of pupillary light response after optic nerve compression [8], suggesting that a protective strategy for RGCs may be a promising next-generation therapy. Most forms of neurological disease are associated with reactive astrocytes, ranging from acute injury to degeneration [10]. Impairment of axonal regeneration and functional recovery results from damage to the central nervous system, enabling the transformation of naive astrocytes into reactive astrocytes and ultimately into scar-forming astrocytes [11]. Interestingly, a recent study found that the injured ON promotes astrocyte accumulation, glial scar formation, and RGCs death in the retinal layer [12]. In addition, astrocytes are emerging as central regulators of retinal ON inflammatory responses because they also have strong proinflammatory potential [13]. However, how trauma promotes astrocyte accumulation remains largely unknown. It is important to investigate the influence of astrocyte aggregation and the inflammatory response on TON and its underlying mechanisms. As a transcription factor, NFIA not only regulates astrocyte development but also affects inflammation [14] and oxidative stress [15]. These risk factors are strongly associated with the pathogenesis of TON [8, 9]. Considering the critical role of inflammation and oxidative stress in causing RGCs death, NFIA may regulate these processes and play a role in TON. In addition, it has been reported that the NFIA protein is highly expressed in reactive astrocytes during human neurological injury, such as multiple sclerosis, hypoxic-ischemic encephalopathy, and spinal cord [16–18], but the expression of NFIA in TON has not been reported. In particular, NFIA localizes the nucleus to the inner nuclear layer and the nerve fiber layer, thereby regulating retinal development [19], suggesting that NFIA is important in the development of the retina. However, whether NFIA regulates astrocyte accumulation in TON remains undefined. In this study, we found that NFIA is upregulated in the retina of TON. miR-181d-5p is an upstream target of NFIA. miR-181d-5p inhibited RGCs loss and astrocyte development by downregulating NFIA. This study revealed a new molecular mechanism of RGCs death during the TON process and provided a potential therapeutic target for the development of new treatment methods for TON.
New Zealand rabbits (clean grade) were obtained from the Experimental Animal Center of Southern Medical University (Guangzhou, China), and all rabbits were housed in a light-dark (12 : 12) cycle temperature-controlled barrier facility with enough food and water at a temperature of 22 °C and a humidity of 50%. All experiments and their protocols involved in this study were approved by the Institutional Committee for Animal Care and Utilization of Kunming Medical University (protocol reference No. kmmu20211586). As described previously [20], rabbits were anesthetized with 3% phenobarbital, and the neurobone canal (or ring) surrounding the optic nerve was isolated through an operating microscope. After inserting the Yasargil aneurysm clip (65742) into the bony canal (or bony ring) to clamp the nerve for 30 seconds, the clip was removed. In this study, rabbits were divided into the model group, sham group, or miR-181 (miR-181d-5p mimic) group and were treated for four weeks as follows: model rabbits receiving agomiR-181d-5p treatment as described previously (5 nmol agomiR-181d-5p applied 4 times daily) [21] and (4) miR-181+oe-NFIA (NFIA overexpression lentiviral vector) group: model rabbits receiving agomiR-181d-5p and oe-NFIA (Guangzhou RiboBio Biotechnology Co., Ltd., Guangzhou, China). All experimental rabbits were euthanized on days 1, 3, 7, 14, or 28 of treatment and their tissues were used for further analysis.
The rabbit eyeballs were removed and fixed with 4% paraformaldehyde for 20 min, and 5 μm thick tissue pieces were stained with H&E at 25 °C for 10 min. A Zeiss fluorescence microscope (Axio imager microscope) with identical acquisition settings was used to observe H&E staining (magnification 50x). Three sections were taken from each eye, and the data were analyzed by GCL cell density in each visual field.
The rabbit eyeballs were fixed with FAS solution at room temperature for 24 h and sectioned into paraffin-embedded sections with a thickness of 5 μm. The sections were dewaxed, heated at high pressure for antigen repair, sealed with goat serum at room temperature for 1 h, and incubated with primary antibodies against glial fibrillary acidic protein (GFAP) (dilution ratio, 1 : 1000; No. ab7260; Abcam, UK) and Brn3a (dilution ratio, 1 : 1000; No. ab245230; Abcam) at 4 °C overnight. After the addition of Alexa Fluor-conjugated secondary antibody, the cells were incubated at 25 °C for 1 h. Subsequently, cell nuclei were stained with 0.1% DAPI for 5 min at 25 °C. Immunofluorescence staining was visualized at 50x magnification using a Zeiss fluorescence microscope. Three sections were taken from each eye, and the data were analyzed by ganglion cell layer (GCL) cell density in each visual field.
Retinal tissue from rabbits was harvested, and the MDA content and SOD activity in retinal tissue were determined using the corresponding ELISA kits (Sangon Biotech, Shanghai) following the manufacturer's instructions.
The total proteins from retinal tissue were extracted utilizing RIPA lysis buffer (Sangon Biotech, Shanghai) and using a BCA assay (Sangon Biotech, Shanghai) to determine the total protein content. A 10% SDS-PAGE gel was used to separate the extracted total proteins, which were then transferred to polyvinylidene fluoride membranes by a constant current flow at 200 mA. Subsequently, PVDF membranes were incubated with Bcl-2 (1 : 500; No. ab196495; Abcam), Bax (1 : 1000; ab32503; Abcam), NFIA (1 : 1000; ab228897; Abcam), and GAPDH (1 : 5000; ab8245; Abcam) antibodies for 12 h at 4 °C after blocking with 5% skim milk. The PVDF membranes were washed with TBS buffer and incubated with corresponding secondary antibodies (Abcam) at 25 °C for 1 h. Immunoblots were visualized using chemiluminescent reagents (Yeasen, Shanghai, China), and grayscale analysis was performed by the ImageJ software.
RNA was isolated from retinal tissue and RGCs using a Total RNA Extractor (Sangon Biotech) and a first-strand cDNA synthesis kit (Vazyme, Nanjing, China) to reverse transcribe it into cDNA. Subsequently, RT-qPCR was performed using a universal high-specificity, dye-based, quantitative PCR detection kit (Vazyme, Nanjing, China) in an ABI 7300 sequence detection system (Applied Biosystems) with thermal cycling conditions of 94 °C for 5 min, 40 cycles of denaturation at 94 °C for 15 s, and annealing at TM value (60 °C) for 30 s. The U6 gene was selected as the reference gene, and the relative expression of the target gene was calculated by the 2-∆∆Ct method [22].
In this study, dual-luciferase reporter vectors containing wild-type (WT) and mutant-type (MUT) binding sites for NFIA sequences were constructed by a rapid cloning kit (Vazyme, Nanjing, China) and named WT-NFIA and MUT-NFIA, respectively. Subsequently, WT-NFIA and MUT-NFIA vectors were transfected into 293T cells (Chinese Academy of Sciences Culture Collection) with miR-181d-5p mimic and negative controls. After transfection for 48 h, the dual-luciferase reporter assay (Promega) was used to detect luciferase activity.
In this study, following the instructions of the ELISA kit (Abcam), the retinal tissues of rabbits were collected after treatment, and the content of inflammatory cytokines, including TNF-α, IL-1, IL-1β, and IL-6, was measured at the corresponding wavelength (450 nm) optical density using a microplate spectrophotometer (BioTek, USA).
The GraphPad Prism 8 software was used to analyze and prepare graphs of the experimental data. In this study, all data are shown as the mean ± standard deviation (SD). Data from two groups and multiple groups were analyzed by unpaired Student's t-test and one-way analysis of variance followed by Tukey's post hoc test, respectively. The P value for statistical significance was 0.05.
To understand and investigate the potential impact of trauma on retinal laminar structure, retinal sections were histologically assessed by H&E staining in this study. As shown in Figure 1(a), GCL density gradually decreased with time after trauma, and the lowest point was measured at 14 d, which showed significant changes at 7 d, 14 d, and 28 d. The activity of SOD, an antioxidative enzyme, gradually decreased with time after trauma and reached the lowest point at 7 d (Figure 1(b)), while the content of MDA, an oxidative stress marker, gradually increased with time and reached the highest point at 7 d (Figure 1(c)). We further investigated the apoptotic effect of traumatic conditions in the retina. Western blotting was used to measure the expression of Bcl-2 and Bax. As represented in Figure 1(d), Bcl-2 gradually decreased with time after trauma and reached the lowest point at 14 d, while Bcl-2 gradually increased with time and reached the highest point at 14 d. As shown in Figures 1(a), 1(b), and 1(d), compared with the sham group, these data showed a large difference at 14 d. Thus, we chose 14 d after trauma for the following experiments. These results demonstrate that trauma increases retinal oxidative stress and RGCs death.
MicroRNAs (miRNAs) can be involved in the pathophysiologic process of many diseases by binding to target mRNAs [23]. A recent study identified that ocular hypertension and TON induce significant changes in RGCs miRNAs [24]. To explore a functional miRNA in TON, we queried the RNA-seq data between normal RGCs and injured RGCs (seven days postoptic nerve crush) and selected the top 12 differentially expressed miRNAs; the expression of these 12 miRNAs was analyzed by RT-qPCR. Among these abnormally expressed miRNAs, we were concerned with miR-181d-5p, which is one of the lowest expressed miRNAs in injured RGCs (Figure 2(a)) and injured retinal tissues (Figure 2(b)). These data indicate that miR-181d-5p was highly expressed in traumatic retinal tissues and RGCs.
We next overexpressed miR-181d-5p in rabbit eyeballs (Figure 3(a)). miR-181d-5p overexpression dramatically increased the GCL density of the retina compared with that of the model group (Figure 3(b)). RGCs death and astrocyte development are important indicators of TON [25, 26]. Immunofluorescence assays showed that the expression of POU-homeodomain transcription factor Brn3a, an RGCs marker, was reduced in the injured retinal tissues but was increased by miR-181d-5p overexpression (Figure 3(c)). In contrast, as an astrocyte marker, GFAP was elevated in the traumatic retinal tissues, and the increase was rescued by miR-181d-5p overexpression (Figure 3(d)). In addition, miR-181d-5p overexpression remarkably enhanced SOD activity and reduced MDA content (Figures 3(e) and 3(f), respectively). The expression of Bcl-2 was increased, while the expression of Bax was decreased under the overexpression of miR-181d-5p (Figure 3(g)). We further tested the content of TNF-α, IL-1β, and IL-6 in the retinal tissues by ELISA. We noticed that their levels were elevated after trauma but this elevation was reduced by upregulating miR-181d-5p (Figure 3(h)). It was discovered that miR-181d-5p rescues RGCs death, oxidative stress, astrocyte development, and inflammation in retinal cells.
NFIA was upregulated in the injured retinal tissues (Figure 4(a)). Identifying the targets of miR-181d-5p is a very important part of this mechanism. In this study, starBase v2.0 software (http://starbase.sysu.edu.cn/targetSite.php) was used to predict miR-181d-5p Targeted binding site with NFIA (Figure 4(b)) [27]. Meanwhile, to confirm the interaction function between miR-181d-5p and NFIA, WT or MUT-3′UTR of NFIA was cloned into the luciferase reporter vector. We next overexpressed miR-181d-5p in 293T cells (Figure 4(c)), other than the finding that the luciferase activity of the WT-NFIA reporter vector was significantly restrained by the overexpression of miR-181d-5p (Figure 4(d)). In addition, the expression of NFIA was significantly reduced in 293T cells transfected with miR-181d-5p mimic (Figure 4(e)). These data indicate that miR-181d-5p directly targets the NFIA.
Next, to determine the effects of the miR-181d-5p/NFIA axis on RGCs death, astrocyte development, oxidative stress, and inflammation of the retina, we used miR-181d-5p mimics and an NFIA overexpression vector to change miR-181d-5p and NFIA expression. As shown in Figure 5(a), trauma lowered GCL density, which was reversed with overexpression of miR-181d-5p, but GCL density was ultimately repressed by overexpression of NFIA. Immunofluorescence analysis showed that Brn3a expression was significantly inhibited by induced trauma but enhanced after miR-181d-5p overexpression, which was terminally repressed by upregulating NFIA (Figure 5(b)). In contrast, GFAP protein levels were elevated after induced trauma but reduced with miR-181d-5p overexpression, which was terminally boosted by NFIA overexpression (Figure 6(a)). Overexpression of miR-181d-5p alleviated the trauma-induced SOD activity reduction and MDA content increase, and NFIA overexpression reversed the effect of miR-181d-5p overexpression (Figures 6(b) and 6(c)). The expression of NFIA was elevated after induced trauma but reduced with miR-181d-5p overexpression, which was terminally boosted by NFIA overexpression (Figure 6(d)). The expression of Bax showed a similar trend as NFIA expression, but the expression of Bcl-2 showed the opposite trend (Figure 6(d)). Finally, the levels of TNF-α, IL-1β, and IL-6 were elevated after induced trauma but reduced with miR-181d-5p overexpression, which was terminally boosted by NFIA overexpression (Figure 6(e)). Taken together, our data show that miR-181d-5p relieves RGCs death, astrocyte development, oxidative stress, and inflammation by downregulating NFIA.
Trauma to the head and face can directly or indirectly cause damage to TON cells [1]. The major cellular component of the retina is RGCs, the loss of which can lead to retinopathy, including photoreceptor degeneration, diabetic retinopathy, glaucoma, and TON [28–30]. In our study, we observed RGCs death, astrocyte development, oxidative stress, and inflammation in the retinas of the TON animal model. A recent study identified that ocular hypertension and TON induce significant changes in RGCs miRNAs [24]. Through RT-qPCR analysis, we identified the top 12 differentially expressed miRNAs in retinas and RGCs. Among these abnormally expressed miRNAs, we identified miR-181d-5p to play a critical role in RGCs death in TON. According to previous reports, miRNAs are involved in gene regulation and other cellular processes and play a wide range of roles [23], including RGCs death [31]. Here, we found that miR-181d-5p targets the NFIA gene. Notably, accumulating evidence shows that NFIA levels regulate astrocyte development, oxidative stress, and inflammation [14, 15, 17]. Overexpression of miR-181d-5p relieves RGCs death, which may be related to astrocyte development, oxidative stress, and inflammation, suggesting that miR-181d-5p and NFIA are critical for preventing RGCs death in the retina after induced trauma. In the process of retinal degeneration, cellular oxidative stress plays an indispensable role, including TBI and TON [9]. Our results indicate that the overexpression of miR-181d-5p alleviated the trauma-induced SOD activity reduction and MDA content increase in this study. Previous studies have shown that with the increase in oxidative stress in the mouse retina, RGCs death was decreased by TON and inhibition of oxidative stress [8, 32]. We utilized the expression of Brn3a measured by immunofluorescence to evaluate RGCs death, and Brn3a was significantly decreased in the model group, while miR-181d-5p increased this value, indicating that miR-181d-5p has the ability to prevent RGCs loss. Notably, as a rheostat, Bcl-2/Bax can regulate cellular antioxidant pathways and death [33]. We observed that miR-181d-5p promoted the expression of Bcl-2 while decreasing the expression of Bax. Our study supports this view and suggests that miR-181d-5p may inhibit oxidative stress-linked RGCs death, which is beneficial for TON prevention. Astrocytes have extensive proinflammatory capabilities and are regulators of inflammatory responses in the CNS [13]. Reactive astrocyte accumulation has been hypothesized to underlie RGCs apoptotic processes after TON [12]. In addition, studies have shown that excessive activation of astrocytes is detrimental to the repair of retinal ganglion cells after optic nerve injury [34]. Inhibition of neuroinflammatory reactive astrocyte formation also significantly reduced the death of RGCs in mouse models of glaucoma [35]. In our study, significant astrocyte development in the retina of TON animal models was observed and miR-181d-5p effectively reduced the number of astrocytes. Moreover, the levels of TNF-α, IL-1β, and IL-6 were elevated after induced trauma but were reduced with miR-181d-5p overexpression. Furthermore, oxidative stress in astrocytes and neurons is triggered by astrocyte dysfunction which may lead to neurodegeneration [36]. Here, we found that the number of astrocytes was elevated in the retina of TON animal models, and that the oxidative stress response was also increased. Interestingly, NFIA is a transcription factor that not only regulates astrocyte development but also affects inflammation [14] and oxidative stress [15]. NFIA in astrocytes has an endogenous prodifferentiation function [37], and NFIA is highly expressed in astrocytes responsive to human neural injury [17]. In addition, miRNA can play a role through targeted regulation of NFIA, for example, miR-424 can prevent astrocyte proliferation after cerebral ischemia/reperfusion in elderly mice by regulating NFIA [38]. In the present study, we determined that the expression of BFIA was upregulated in the retinal tissues of TON animal models. However, miR-181d-5p rescued the upregulated expression of NFIA. Furthermore, overexpression of NFIA reversed the effects of miR-181d-5p against RGCs death, astrocyte development, oxidative stress, and inflammation. These results indicated that miR-181d-5p protects against TON by downregulating NFIA. In this study, RGCs death and associated astrocyte activation were observed in traumatic optic neuropathy, which was consistent with previous studies, but this study proposed a new regulatory mechanism and for the first time explored the effect of mir-181d-5p/NFIA molecular axis on RGCs death. In conclusion, miR-181d-5p can protect against the deterioration of TON by inhibiting RGCs death, astrocyte development, oxidative stress, and inflammation through downregulation of NFIA. These results provide new insights for early medical interventions in patients with TON. | true | true | true |
||
PMC9569233 | Xiaokang Wang,Shulong Wu,Yi Yang,Jingjing Zhao | LncRNA CARMN Affects Hepatocellular Carcinoma Prognosis by Regulating the miR-192-5p/LOXL2 Axis | 08-10-2022 | Background Hepatocellular carcinoma (HCC) is aggressive cancer with a poor prognosis. It has been suggested that the aberrant expression of LOXL2 is associated with the development of HCC, but the exact mechanism remains unclear. This research is aimed at examining the expression level and prognostic value of LOXL2 in hepatocellular carcinoma and its relationship with immune infiltration and at predicting its upstream noncoding RNAs (ncRNAs). Method The transcriptome data of HCC was first downloaded from The Cancer Genome Atlas (TCGA) database to investigate the expression and prognosis of LOXL2. Then, the starBase database was used to find the upstream ncRNAs of LOXL2, and correlation analysis and expression analysis were performed. Finally, the Tumor Immune Estimation Resource (TIMER) was used to explore the association between LOXL2 and immune cell infiltration. Result CARMN was considered to be the potential upstream lncRNA for the hsa-miR-192-5p/LOXL2 axis in HCC. Furthermore, the level LOXL2 was markedly positively associated with tumor immune cell infiltration and immune checkpoint expression in HCC. Conclusion Higher expression of LOXL2 mediated by microRNA (miRNA) and long noncoding RNAs (lncRNA) is associated with poor overall survival (OS), immune infiltration, and immune checkpoint expression in HCC. | LncRNA CARMN Affects Hepatocellular Carcinoma Prognosis by Regulating the miR-192-5p/LOXL2 Axis
Hepatocellular carcinoma (HCC) is aggressive cancer with a poor prognosis. It has been suggested that the aberrant expression of LOXL2 is associated with the development of HCC, but the exact mechanism remains unclear. This research is aimed at examining the expression level and prognostic value of LOXL2 in hepatocellular carcinoma and its relationship with immune infiltration and at predicting its upstream noncoding RNAs (ncRNAs).
The transcriptome data of HCC was first downloaded from The Cancer Genome Atlas (TCGA) database to investigate the expression and prognosis of LOXL2. Then, the starBase database was used to find the upstream ncRNAs of LOXL2, and correlation analysis and expression analysis were performed. Finally, the Tumor Immune Estimation Resource (TIMER) was used to explore the association between LOXL2 and immune cell infiltration.
CARMN was considered to be the potential upstream lncRNA for the hsa-miR-192-5p/LOXL2 axis in HCC. Furthermore, the level LOXL2 was markedly positively associated with tumor immune cell infiltration and immune checkpoint expression in HCC.
Higher expression of LOXL2 mediated by microRNA (miRNA) and long noncoding RNAs (lncRNA) is associated with poor overall survival (OS), immune infiltration, and immune checkpoint expression in HCC.
Hepatocellular carcinoma (HCC) is the main type of liver cancer (LC). As the sixth most common cancer in the world, it is the second largest cause of cancer-related death and the most common primary liver cancer with poor prognosis [1, 2]. The morbidity and mortality rates of HCC are expected to significantly increase in the next few years [2]. Epidemiological data show that hepatitis virus infection [3], aflatoxin [4], type 2 diabetes [5], alcohol consumption [6], and smoking [7] are all predisposing factors for liver cancer. Despite remarkable improvements in the diagnosis and treatment of HCC, such as surgical resection [8] and sorafenib-regorafenib sequential therapy [9], patients with HCC often exhibit local invasion and metastasis resulting in a poor overall survival (OS) rate [10, 11]. Therefore, early screening and diagnosis of HCC are particularly crucial, and there is an urgent need to find specific and sensitive biomarkers. Lysyl oxidase (LOX), an extracellular enzyme, plays a key role in the covalent cross-linking of collagen fibers by oxidizing deamino-specific lysine and hydroxylysine in the telopeptide structural domain of the collagen molecule to form allantoin [12]. In addition to LOX, there are four other members in the lox protein family, namely, LOX-like proteins (LOXL1, LOXL2, LOXL3, and LOXL4) [13, 14]. Among these proteins, LOXL2 is considered to be an important regulator of tumor progression, and previous studies reported that LOXL2 is significantly overexpressed in human HCC tissues compared to nontumor tissues [15]. Studies demonstrated that elevated levels of LOXL2 might contribute to tumor progression and metastasis by promoting tumor cell invasion and remodeling of the tumor microenvironment [16, 17]. Considering the major role of LOXL2, we further investigated the role of LOXL2 in the development of HCC progression based on previous studies. In this study, the expression level of the LOXL2 and its relationship with prognosis were first analyzed in various common cancers. Next, we found some noncoding RNAs (microRNAs (miRNAs)) and long noncoding RNAs (lncRNAs) as regulatory molecules of LOXL2 by bioinformatics analysis, so as to establish the LncRNA-miRNA-mRNA regulatory network and explore the mechanism of HCC at a deeper level. Moreover, the correlation of LOXL2 expression with immune cell infiltration, biomarkers of immune cells and immune checkpoints was finally discussed.
The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer. gov/) is a collaboration between the National Cancer Institute (NCI) and the National Human Genome Research Institute (NHGRI) for cancer research. It provides a large and free reference for cancer research by collecting and organizing various cancer-related histological data. A total of 33 cancer types are currently included. Data in this research was obtained from the Liver Hepatocellular Carcinoma (LIHC) cohort, and then Log2 transformed. The expression levels of LOXL2 in tumor tissues were compared with normal tissues using Wilcoxon rank sum test in the eighteen cancers (BLCA, BRCA, CHOL, COAD, ESCA, GBM, HNSC, KICH, KIRC, KIRP, LIHC, LUAD, LUSC, PRAD, READ, STAD, THCA, and UCEC), and then visualized by box plots. The analysis was performed using the R software “limma” and “ggplot2” packages [18].
The GEPIA database (http://gepia.cancer-pku.cn/) integrates TCGA cancer data with GTEx normal tissue data, exploiting bioinformatics techniques to drill down into novel cancer targets and markers. In this research, we analyzed the expression LOXL2 in tumor and normal samples and prognosis in pan-cancer by the GEPIA database [19]. The boxplots and Kaplan–Meier plots were downloaded for visualizing the results of differential expression analysis and survival analysis.
The starBase database (https://starbase.sysu.edu.cn/) is used to analyze data related to multiple cancers integrated from the TCGA project. It provides a platform for predicting miRNA targets by searching for miRNA targets through high-throughput CLIP-Seq experimental data and degradome experimental data which include lncRNAs, miRNAs, snoRNAs, and mRNAs [20]. We use this platform to detect the upstream miRNA of LOXL2 in this research. After searching with the keyword “LOXL2” in “Target Gene” module and selecting the “miRNA-mRNA” option, the upstream miRNAs that LOXL2 may bind to would be presented. The filtering condition for screening is that the miRNAs would be predicted in two or more programs. The regulatory network between these predicted miRNAs and LOXL2 was demonstrated with Cytoscape software [21]. Among them, miRNAs with correlation coefficients greater than 0.2 were included in the subsequent analysis. In addition, expression analysis for the selected miRNA was also performed.
The upstream lncRNA of the miRNAs selected in the prior step was also identified in starBase database [20]. It has been known that miRNAs can bind to target mRNAs and inhibit their translation or cause mRNA degradation to achieve the function of posttranscriptional regulation of gene expression. The ceRNA theory represents a new model of gene expression regulation. ceRNA molecules (lncRNA, circRNA, etc.) can compete to bind the same miRNA through miRNA Response Element (MRE) to regulate each other's expression levels [22–24]. Therefore, the eligible lncRNA should be negatively correlated with miRNA and positively correlated with mRNA.
The TIMER database (https://cistrome.shinyapps.io/timer/) is a comprehensive resource for the systematic analysis of immune infiltrates in diverse cancer types [25]. TIMER database is used to detect six types of immune cell (including dendritic cells, macrophages, neutrophils, CD4+ T cells, CD8+ T cells, and B cells) infiltration in tumor tissues with RNA-Seq expression profiling data. In this study, the TIMER database was used to estimate the correlation between LOXL2 and the extent of infiltration of specific immune cell subpopulations. Additionally, considering the potential oncogenic role of LOXL2 in HCC, the relationship of LOXL2 with immune checkpoints (involving CTLA4/PDCD1/CD274) was assessed as well.
The statistical analysis was automatically calculated by the online database or statistical software. p value less than 0.05 was considered statistically significant.
To explore the role of LOXL2 in the development of HCC progression, we first analyzed the expression level of LOXL2 in 18 types of cancer based on the TCGA database, which found that LOXL2 was markedly upregulated in 16 cancer types, including BLCA, BRCA, CHOL, COAD, ESCA, GBM, HNSC, KIRC, KIRP, LIHC, LUAD, LUSC, READ, STAD, THCA, and UCEC and was significantly downregulated in PRAD. However, there was no significant difference between KICH and normal tissues (Figure 1(a)). Next, to further confirm this result, the GEPIA database was also used to evaluate the LOXL2 expression levels in these cancers. As presented in Figure 1(b), LOXL2 was significantly increased in CHOL, ESCA, GBM, HNSC, KIRC, LIHC, and STAD, whereas it was decreased in PRAD, compared with normal tissues. Also, no statistically significant difference was observed in BLCA, BRCA, COAD, KICH, KIRP, LUAD, LUSC, READ, THCA, and UCEC (Figures 1(c)–1(s)). Taken together, LOXL2 was upregulated in CHOL, ESCA, GBM, HNSC, KIRC, LIHC, and STAD and downregulated in PRAD, which indicated that LOXL2 was in connection with the development of the above cancer types.
GEPIA platform was used to analyze the prognostic value of LOXL2 in pan-cancers, and OS is selected to be the outcome indicator. As shown in Figure 2, higher expression of LOXL2 was associated with worse OS in LIHC, LUAD, and LUSC. Consequently, combining the expression of LOXL2 between tumor and normal tissues and its prognostic value, LOXL2 may be utilized as an unfavorable prognostic biomarker in patients with LIHC.
The expression analyses of LOXL2 in unpaired samples, paired samples, and different clinical subgroups are shown in Figure 3. With the increased expression level of LOXL2, the pathologic stage of HCC increases accordingly. Also, we found that in female, the expression level of LOXL2 was higher than that in male. Table 1 listed the LOXL2 expression data and clinical data for 374 HCC patients. We observed a significant association between LOXL2 expression and clinicopathological features, such as gender, T stage, and histologic grade (p < 0.05). As for survival, we further performed subgroup analyses to assess the impact of LOXL2 expression on OS of patients with HCC according to age, gender, pathological stage, and differentiation grade. We found that high-expression of LOXL2 resulted in poor survival in patients older than 60 years old and in male patients (Figure 4).
MiRNAs represent a class of noncoding single-stranded RNA molecules of approximately 22 nucleotides in length encoded by endogenous genes, which are involved in the posttranscriptional regulation of gene expression. Twenty-two possible miRNAs were found in the starBase database according to the rules, and the regulatory network was presented in Figure 5(a). Notably, the miRNA that regulates the mRNA must be negatively correlated with that mRNA. Therefore, the screening criteria were set: the value of the correlation coefficient was less than -2 and the p value was less than 0.05. Finally, only hsa−miR−192−5p met the conditions (Figure 5(b)). Furthermore, hsa−miR−192−5p was found to be lowly expressed in HCC tissues and patients with lower hsa−miR−192−5p expression had a better prognosis. (Figures 5(c) and 5(d)).
Also, the upstream lncRNAs of miRNAs can be searched in the starBase database. After downloading the twelve relevant lncRNAs from the database, the regulatory network was visualized using Cytoscape software (Figure 6(a)). However, among these lncRNAs, only CARMN was negatively correlated with hsa−miR−192−5p (R < −2, P < 0.05) and positively correlated with LOXL2 (R > 2, P < 0.05) (Figures 6(b) and 6(c)). Besides, expression analysis showed that CARMN was highly expressed in the tumor samples (Figure 6(d)). Thus, CARMN was selected as a promising upstream lncRNA for the miR-192-5p/LOXL2 axis in HCC.
The relationship between LOXL2 and immune cell infiltration was investigated using the timer database because the level of immune cells is associated with the proliferation and development of tumor cells. The expression of LOXL2 was positively correlated with the infiltration of dendritic cells, neutrophils, macrophages, CD8+ T cells, CD4+ T cells, and B cells (Figure 7). Among them, CD4+ T Cells, macrophages, and neutrophils showed the strongest positive correlation.
CD274, PDCD1, and CTLA-4 were known as critical immune checkpoints that are associated with immune escape in cancers. So, the relationship between LOXL2 and these checkpoints was analyzed via online tools. Both TIMER data analysis and GEPIA data analysis found a significant positive correlation of LOXL2 with CD274, PDCD1, and CTLA-4 (Figures 8(a)–8(f)).
HCC is one of the most common malignant tumors in the world, and with the continuous development of medical technology and research on the molecular biology of tumors, targeted therapy for hepatocellular carcinoma is developing rapidly [1, 2]. TERT, MLL4, CCNE1, TP53, and CTNNB1 were identified as commonly mutated genes in HCC [26–30]. Despite the availability of many potential therapeutic targets, the incidence and mortality rates of patients with HCC are still increasing currently [8–11]. The overall survival rate remains suboptimal, so it is crucial to explore the underlying molecular mechanisms and oncogenes to provide new ideas for the diagnosis and treatment of HCC. Previous studies had uncovered the value of LOXL2 in different tumors, but the exact mechanism remains unclear [15, 31–33]. In this study, we discussed the role of LOXL2 in HCC through bioinformatics analysis to further understand the potential value of LOXL2. In this present study, the expression of LOXL2 in various cancers was analyzed using different databases (TCGA database and GEPIA database), and it was concluded that LOXL2 was significantly differentially expressed in HCC. The next survival analysis using GEPIA online tool revealed that LOXL2 overexpression leads to a poorer prognosis in patients with HCC. Wu et al. constructed LOXL2-small interfering RNA using a lentiviral vector and investigated the effect of LOXL2 on the proliferation of HCC cell lines by reverse transcription-quantitative polymerase chain reaction and other experimental methods [34]. The results showed that LOXL2 was highly expressed in HCC tissues and that LOXL2 silencing reduced cell number, proliferation, colony formation, and cell growth, induced cell cycle arrest, and increased apoptosis [34]. This study and our results both demonstrated the oncogenic role of LOXL2 in HCC. The starBase database contains seven programs that can be used to predict miRNAs, including TargetScan, miRmap, miRanda, PicTar, RNA22, PITA, and microT. Through these programs, twenty-two potential miRNAs were found for LOXL2. Among all these 22 possible miRNAs, only hsa-miR-192-5p showed a significant negative correlation with LOXL2. Subsequently, the data of miRNAs were downloaded from the TCGA database, and the differential analysis showed that hsa-miR-192-5p was lowly expressed in the tumor samples, and its low-expression was related to the poor prognosis of HCC. During the past decade, more and more studies are focusing on the role of miR-192-5p in cancers. For example, in Wang et al.'s experiment, they found that miR-192-5p-modified tumor-associated macrophages-derived exosome suppressed endometrial cancer progression by targeting IRAK1/NF-κB signaling [35]. The role of miR-192-5p in HCC is also noteworthy. Previous studies revealed that miR-192-5p loss enhanced glycolysis and over produced lactate might further increase HCC malignant features via interacting with environmental nontumor cells [36]. Previous findings are consistent with our predictions. To summarize, hsa-mir-192-5p was found to be a vital regulatory molecule of LOXL2 in HCC. According to the ceRNA theory, it is known that ceRNA inhibits the inhibitory effect of miRNA on mRNA by binding to it. Therefore, it is essential to find the upstream lncRNA of hsa-mir-192-5p because it plays a role in cancer development. StarBase bioinformatics software predictions showed that CARMN may be the upstream lncRNA of miR-192-5p. The expression analysis also showed that CARMN was highly expressed in the tumor samples. There is a paucity of research on CARMN in cancers. Sheng et al. found that overexpression of CARMN can promote the prognosis and chemosensitivity in breast cancer [37]. Other reports on the role of lncRNA CARMN in cancers are currently scarce. However, the results of this paper suggested that CARMN might influence the progression of HCC by regulating the miR-192-5p/LOXL2 axis. The tumor microenvironment is a hot topic of research in recent years. The immune microenvironment, consisting of tumor-infiltrating lymphocytes (B cells, T cells) and other immune cells (such as dendritic cells, macrophages, and neutrophils), is an important part of the tumor microenvironment and is considered as the “seventh hallmark feature” of tumors, so more research is urgently needed to focus on the link between immune cell infiltration and tumors [38–40]. In our research, a positive correlation between LOXL2 expression and dendritic cells, neutrophils, macrophages, CD8+ T cells, CD4+ T cells, and B cell in HCC was observed. In addition, the correlation between LOXL2 expression and immune checkpoint markers (CD274, PDCD1, and CTLA-4) suggested a role for LOXL2 in immune regulation of tumor immunity. Consequently, it is hypothesized that tumor immune infiltration exerts an influential role in LOXL2-mediated HCC development. The limitation of this study should be mentioned. First, this is a bioinformatics study based on an online database that lacks experimental validation. Second, in identifying CARMN as a potential upstream LncRNA for hsa-miR-192-5p in HCC, serval extreme values may leverage the true correlation and may lead to overestimation of the correlation of CARMN with hsa-miR-192-5p. Third, the results need to be interpreted carefully because many of the data do not have high-correlation values. In short, CARMN was identified as a possible upstream lncRNA for miR-192-5p, which affects LOXL2 expression and promotes the progression of HCC. Our study suggests that LOXL2 may exert its tumorigenic effects by potentiating tumor immune cell infiltration and immune checkpoint expression. However, in the future, more relevant studies are still needed to verify these predictions by bioinformatics. | true | true | true |
||
PMC9569602 | 36232960 | You Lan,Bo Qian,Hai-Yan Huang,Pan Wang,Ting Li,Qi Yuan,Han-Yu Zhang,Yu-Chun Lin,Zhong-Ning Lin | Hepatocyte-Derived Prostaglandin E2-Modulated Macrophage M1-Type Polarization via mTOR-NPC1 Axis-Regulated Cholesterol Transport from Lysosomes to the Endoplasmic Reticulum in Hepatitis B Virus x Protein-Related Nonalcoholic Steatohepatitis | 01-10-2022 | hepatitis B virus x protein,prostaglandin E2,lysosomal mTOR-NPC1 signal axis,cholesterol transport,endoplasmic reticulum stress,macrophage M1-type polarization,nonalcoholic steatohepatitis | Lipid metabolic dysregulation and liver inflammation have been reported to be associated with nonalcoholic steatohepatitis (NASH), but the underlying mechanisms remain unclear. Hepatitis B virus x protein (HBx) is a risk factor for NASH. Based on metabolomic and transcriptomic screens and public database analysis, we found that HBx-expressing hepatocyte-derived prostaglandin E2 (PGE2) induced macrophage polarization imbalance via prostaglandin E2 receptor 4 (EP4) through in vitro, ex vivo, and in vivo models. Here, we revealed that the M1-type polarization of macrophages induced by endoplasmic reticulum oxidoreductase-1-like protein α (ERO1α)-dependent endoplasmic reticulum stress was associated with the HBx-related hepatic NASH phenotype. Mechanistically, HBx promoted Niemann–Pick type C1 (NPC1)/oxysterol-binding protein-related protein 5 (ORP5)-mediated cholesterol transport from the lysosome to the endoplasmic reticulum via mammalian target of rapamycin (mTOR) activation. This study provides a novel basis for screening potential biomarkers in the macrophage mTOR–cholesterol homeostasis–polarization regulatory signaling pathway and evaluating targeted interventions for HBx-associated NASH. | Hepatocyte-Derived Prostaglandin E2-Modulated Macrophage M1-Type Polarization via mTOR-NPC1 Axis-Regulated Cholesterol Transport from Lysosomes to the Endoplasmic Reticulum in Hepatitis B Virus x Protein-Related Nonalcoholic Steatohepatitis
Lipid metabolic dysregulation and liver inflammation have been reported to be associated with nonalcoholic steatohepatitis (NASH), but the underlying mechanisms remain unclear. Hepatitis B virus x protein (HBx) is a risk factor for NASH. Based on metabolomic and transcriptomic screens and public database analysis, we found that HBx-expressing hepatocyte-derived prostaglandin E2 (PGE2) induced macrophage polarization imbalance via prostaglandin E2 receptor 4 (EP4) through in vitro, ex vivo, and in vivo models. Here, we revealed that the M1-type polarization of macrophages induced by endoplasmic reticulum oxidoreductase-1-like protein α (ERO1α)-dependent endoplasmic reticulum stress was associated with the HBx-related hepatic NASH phenotype. Mechanistically, HBx promoted Niemann–Pick type C1 (NPC1)/oxysterol-binding protein-related protein 5 (ORP5)-mediated cholesterol transport from the lysosome to the endoplasmic reticulum via mammalian target of rapamycin (mTOR) activation. This study provides a novel basis for screening potential biomarkers in the macrophage mTOR–cholesterol homeostasis–polarization regulatory signaling pathway and evaluating targeted interventions for HBx-associated NASH.
Nonalcoholic fatty liver disease (NAFLD) has become an emerging public health problem, with a global prevalence of 25%, including a range of liver diseases, from fatty liver to nonalcoholic steatohepatitis (NASH) [1,2]. NASH is characterized by lipid accumulation and inflammatory damage following lipid metabolism disorders [3,4]. Regional immune regulation and immuno-metabolic mechanisms in the liver are the current frontiers and hot spots in the study of NASH. Metabolic perturbations affect the immune cell phenotype and function, disrupting hepatic homeostasis and promoting the development of NASH [5,6]. The increasing number of cases of hepatitis B and hepatic steatosis indicates an association between the hepatitis B virus (HBV) and NAFLD [7]. Additionally, studies have suggested that hepatitis B virus x protein (HBx) is a possible risk factor for a fatty liver. Wu et al. found that HBx induces hepatic steatosis by enhancing the expression of liver fatty acid binding protein [8]. We also found that HBx triggers hepatic steatosis via the induction of cyclooxygenase 2 (COX-2) in hepatocytes [9]. COX-2 catalyzes the synthesis of prostaglandin E2 (PGE2), an inflammatory lipid mediator, from arachidonic acid [10]. Su et al. found that senescent preosteoclasts promote metabolic syndrome associated with osteoarthritis through the COX-2/PGE2 axis [11]. PGE2 stimulates different immune cells, such as macrophages, dendritic cells, neutrophils, and T cells, to exert various biological effects through PGE2 receptors (EP1-EP4) [12]. The specific role and related mechanism of PGE2 in HBV/HBx-related NASH need further exploration. Lipotoxicity induced by cholesterol accumulation within macrophages induces the formation of foam cells with an M1-type polarization and accelerates the transition of NAFLD from simple steatosis to NASH [13,14]. Macrophages obtain cholesterol primarily through the uptake of low-density lipoprotein (LDL) from the surrounding environment [15]. LDL then enters the lysosomes (Lyso) and is degraded to free cholesterol (FC) by lysosomal acid lipase (LAL), and through feedback inhibition, the increase in intracellular FC levels can prevent LDL receptors from reducing cholesterol uptake [16]. Unlike LDL, oxidized low-density lipoprotein (ox-LDL) accumulates in the lysosome, and intracellular FC levels do not negatively affect scavenger receptors, leading to cholesterol accumulation in macrophages [17]. As a result, excess cholesterol is transferred from the lysosome to other organelles and causes stress and dysfunction [18]. In atherosclerotic disease, cholesterol accumulation in the endoplasmic reticulum (ER) can trigger ER stress through PERK, IRE1α, and ATF6 pathways, which can launch proinflammatory responses, reduce cholesterol efflux, or/and increase cholesterol uptake, which can lead to foam cell formation via lipid droplet accumulation [19]. Our previous studies found that ER stress mediated the inflammatory response triggered by ultra-small superparamagnetic iron oxide nanoparticles [20]; HBx induced cell-cycle arrest and apoptosis via ER stress [21]; and aflatoxin B1 induced COX-2 upregulation and inflammasome activation via the modulation of endoplasmic reticulum oxidoreductase-1-like protein α (ERO1α)-dependent ER redox homeostasis and ER stress [22]. Therefore, we hypothesized that cholesterol accumulation-induced ER stress is also involved in the proinflammatory M1-type polarization of macrophages, contributing to the development of HBx-associated NASH. Cholesterol transport between organelles is mainly divided into vesicular and non-vesicular forms [23]. Recent studies have proposed that inter-organelle membrane contact sites (MCS), as non-vesicular lipid transfer hubs, are central to the metabolic integration and signal transduction between organelles by facilitating the exchange of small molecules, such as cholesterol, in local regions through tightly arranged bundle complexes [24]. For example, Höglinger et al. found that Niemann–Pick type C1 (NPC1) regulates endocytic organelle–ER contact to mediate lipoprotein-derived cholesterol transport from the lysosome to the ER [25]. In addition, cholesterol transporters bridging MCS between the lysosome and ER include oxysterol-binding protein (OSBP)-related proteins (ORPs) [26], steroidogenic acute regulatory protein-related lipid transfer domain (START) [27], GramD [25], and SLC38A9 [28]. Notably, the mammalian target of rapamycin (mTOR) response to various environmental cues, from cytokine stimulation to nutrients to stress, may be associated with cholesterol transport [29]. In a previous study, we found that COX-2 modulated Ca2+ transport across the mitochondria-associated ER membrane (MAM) to mediate superparamagnetic iron oxide nanoparticle-induced hepatotoxicity [30]. Therefore, we hypothesized that Lyso-ER cholesterol transport induces ER stress and is regulated by mTOR signaling. We constructed in vitro, ex vivo, and in vivo experimental models of HBx-expressing hepatocytes to activate macrophages in the current study. We found that PGE2-EP4-mTOR signaling promotes Lyso-ER cholesterol transport via NPC1 and aggravates ER stress-induced macrophage M1-type polarization through the ERO1α-mediated disturbance of ER redox in HBx-related NASH. EP4 may be a potential targeted intervention for the occurrence and progression of NASH induced by HBx.
Our previous research found that HBx combined with aflatoxin B1 triggers hepatic steatosis via COX-2 [9]. To further investigate the effect of HBx expression on hepatic metabolism, we performed metabolomic assays in the livers of HBx-Tg mice. Using untargeted metabolomics, we found that most metabolite annotations were in carbohydrate metabolism, lipid metabolism, and the digestive system (Figure S1A). Bile secretion was enriched in the KEGG pathway (Figure S1B). A total of 104 upregulated and 35 downregulated differential metabolites were found in the HBx group (Figure 1A). The levels of PGE2 and hydroxy-PGE2 in specific metabolites of prostaglandins were upregulated, and the level of the PGE2 precursor PGH2 was downregulated (Figure 1B). Since HBx-Tg mice omics suggested the activation of lipid and prostaglandin metabolism, we further tested the levels of ox-LDL and PGE2 in the serum and liver tissue. The levels of ox-LDL and PGE2 in the sera and livers of the HBx group were significantly increased (Figure 1C,D). In addition, IF experiments revealed increased levels of the macrophage marker F4/80 and lipid droplets in liver sections of the HBx group (Figure 1E). The expression of COX-2 and mPGES-1, two key enzymes for PGE2 synthesis, were increased in the NASH liver and correlated with the NASH activity score [31]. Meanwhile, we screened the GEO database (GSE83148) and found that the mRNA level of the COX-2-encoding gene (PTGS2) was upregulated in the liver tissues of HBV-infected patients (Figure S1C). To explore whether PGE2 signaling is involved in macrophages in NASH, we queried the GEO database (GSE104901) to analyze the PGE2 receptor family in the liver macrophages of NASH mice induced by a high-carbohydrate, high-fat, and high-cholesterol diet. As shown in the figure, the level of Ptger4 in the liver macrophages of NASH mice was significantly increased; in addition, biomarkers of the mTOR complex (Mtor), ER stress (Atf4, Chop), and polarization (Nos2, Mrc1) were also significantly different and correlated (Figure S1D,E). Based on the clues provided by GEO, we further verified the co-culture model of HepG2 cells and THP-1 cells. A schematic diagram of the co-culture model is shown in Figure 1F. The HBx-induced upregulation of COX-2 protein levels in HepG2 cells increased the PGE2 release level and upregulation of the EP4 protein level in THP-1 cells (Figure 1G–I). In the co-cultured THP-1 cells in the HBx-expressing group, the M1 polarization marker iNOS increased, the M2 polarization marker CD206 level decreased, and the level of IL-6 in the supernatant was upregulated (Figure 1I,J). The protein levels of p-AKTSer473 and p-mTORSer2448 in THP-1 cells in the HBx-expressing group increased (Figure 1K). The protein levels of ER stress markers PERK, ATF4, and ATF6 in THP-1 cells in the HBx-expressing group increased, and the level of ATF4 in the nucleus increased (Figure 1L,M). Compared with the ox-LDL+PGE2 group, the protein levels of p-AKTSer473 and p-mTORSer2448 decreased after antagonizing the EP4 receptor with ONO-AE3-208 (Figure S1F); the protein levels of PERK, p-PERKT982, ATF4, and CHOP decreased after inhibiting mTOR with rapamycin (Rapa) (Figure S1G). After inhibiting ATF4, compared with the ox-LDL+PGE2 group, ATF4 expression and the nuclear translocalization level significantly decreased, the iNOS protein level decreased, the CD206 protein level increased (Figure S2A,B), and the level of IL-6 in the supernatant decreased (Figure S2C). The above results suggest that HBx induces COX-2-expressing hepatocyte-derived PGE2 release, activates AKT-mTOR signaling in macrophages via EP4, mediates ER stress-dependent polarization imbalance (M1), and promotes inflammatory responses.
The increased free cholesterol in the ER may induce ER redox homeostasis disturbances and mediate ER stress enhancement. To prove this point, we performed transcriptomic assays on the livers of HBx-Tg mice. There were 859 upregulated and 1066 downregulated differentially expressed genes (DEGs) in the HBx group (Figure 2A). The differentially enriched GO pathways in the HBx group included arachidonic acid metabolism, prostaglandin metabolism, cholesterol homeostasis, cholesterol metabolism, cholesterol biosynthesis, and other biological processes (Figure S3A). Hepatitis B, nonalcoholic fatty liver disease, arachidonic acid metabolism, cholesterol metabolism, and other processes were enriched in the KEGG pathway (Figure S3B). We further performed heatmap cluster analysis on cholesterol metabolism-related genes and found that the transcription level of the cholesterol transport-related gene Npc1 was increased, and the transcription levels of the lipoprotein uptake-related receptor Ldlr and cholesterol hydroxylase Cyp7a1 were downregulated (Figure 2B). In the same GEO database, the correlation analysis showed positive correlations between Mtor, Npc1, Ero1l, and Atf4, and the transcription levels of Npc1 and Ero1l in the liver macrophages of NASH mice were significantly increased (Figure S3C,D). THP-1 cells treated with ox-LDL were morphologically enlarged, and the phagosome/lysosome was significantly enlarged and increased. On this basis, the number of lipid droplets in the ox-LDL+PGE2 group was also increased, and the distance between the lysosome and ER was close (Figure S3E). Compared with the ox-LDL group and ox-LDL+PGE2 group, after Rapa inhibited mTOR, we found that NPC1 and ORP5 protein levels were decreased in the ox-LDL+PGE2 group, but not STARD3 or SLC38A9 (Figure S3F, Left); the colocalization of free cholesterol with the ER (blue-green) was decreased, and its colocalization with the lysosome (magenta) was enhanced (Figure S3G, Upper). Compared with the ox-LDL+PGE2 group, NPC1 and ORP5 protein levels decreased after siNPC1 intervention, and fluorescence experiments revealed the diminished colocalization of free cholesterol with the ER (blue-green) and its enhanced colocalization with the lysosome (magenta) (Figure S3F, Right; Figure S3G, Lower). ERO1α, PDI, and PERK protein levels decreased after siNPC1 intervention (Figure S3H). Protein folding in the ER is an oxidative process that relies on protein disulfide isomerase (PDI) and ERO1α. ERO1α plays an important role in oxidative protein folding by recycling reduced PDI, so ERO1α dysfunction critically affects ER homeostasis or disease states [32]. After inhibiting ER redox with siERO1A, compared with the ox-LDL+PGE2 group, the levels of ERO1α, PDI, PERK, ROS, and GSSG were decreased, and the GSH level was increased (Figure S4A–E); CD206 was increased, and iNOS and IL-6 levels in the supernatant were decreased (Figure S4A,F). Compared with THP-1-pCDH cells, ERO1α, PDI, PERK, iNOS, and IL-6 were upregulated, and CD206 was downregulated in ERO1α-overexpressing THP-1-pCDH-ERO1A cells (Figure S4G,H). Compared with the control group, enhanced NPC1-ORP5 interaction was detected in the IP assay, the colocalization of NPC1 and ORP5 was observed in the IF assay, and increased cytoplasmic levels of free cholesterol and enhanced ER colocalization (blue-green) were observed in the IF assay in THP-1 cells of the HBx-expressing group (Figure 2C–E); the levels of ERO1α, PDI, PERK, ROS, and GSSG were increased, and the GSH level was decreased in THP-1 cells of the HBx-expressing group (Figure 2F–H). The above results suggest that PGE2 induces mTOR signaling activation in ox-LDL-loaded macrophages, regulates NPC1-ORP5-mediated Lyso-ER cholesterol transport, and mediates M1 polarization in macrophages via ER redox and ER stress homeostasis regulation.
To verify the roles of mTOR, NPC1, and ERO1α in macrophage (MC) polarization imbalance induced by HBx-expressing hepatocyte-derived PGE2, primary hepatocytes (PHCs) and primary macrophages (PMCs) were co-cultured by HBx-Tg mouse liver perfusion isolation and divided into the WT, HBx, HBx+Rapa (inhibition of mTOR in PMCs), HBx+U18666A (inhibition of NPC1 in PMCs), and HBx+EN460 (inhibition of ERO1α in PMCs) groups. In the HBx group, COX-2 was increased in PHCs; the levels of PGE2 and ox-LDL were increased in the supernatant (Figure 3A–C); iNOS was increased and CD206 was decreased in PMCs (Figure 3A); the IL-6 level was increased in the supernatant (Figure 3D); the proportion of M1 polarization (CD86+) was increased; and the proportion of M2 polarization (CD206+) was decreased (Figure 3E,F). In the groups with mTOR, NPC1, and ERO1α interventions, the above indicators were rescued to some extent, suggesting that PGE2/ox-LDL in HBx-expressing mouse livers mediates intercellular communication between PHCs and PMCs, regulates mTOR-NPC1 in PMCs to regulate Lyso-ER cholesterol transport, and induces ER stress-dependent PMC polarization imbalance (M1).
The experimental model and treatment pattern of mice for in vivo tests are shown in Figure 4A. In the model construction, there was a tendency toward inhibited body-weight gain in the HBx group, with significant weight gain after 49 days of feeding a high-fat/cholesterol diet (HFCD), suggesting an abnormal metabolic pattern in HBx-Tg mice (Figure 4B). Although there was no significant change in the overall liver/body-weight ratio, there was an increasing trend in the HBx + HFCD group and a significant decrease after intervention with ONO-AE3-208, suggesting that EP4 intervention could rescue HBx + HFCD-induced liver injury (Figure 4C). Compared with the WT group and the corresponding control groups, PGE2, ox-LDL, and ALT in the serum were elevated in the HBx and HBx + HFCD groups and decreased in the HBx + ONO and HBx + HFCD + ONO groups (Figure 4D–G). A further test of relevant indicators in liver homogenates showed that COX-2 protein expression levels were upregulated (Figure S5A), while PGE2, ox-LDL, IL-6, TG, TC, and FC levels in the liver were elevated in the HBx and HBx + HFCD groups and decreased in the HBx + ONO and HBx + HFCD + ONO groups (Figure S5B–E). The HBx and HBx + HFCD groups showed the ballooning of hepatocytes, vacuolization, and inflammatory foci, increased Oil Red O–stained lipid droplets, and increased NASH scores; the ONO-AE3-208 intervention group showed the reduced liver pathological characteristics and the decreased NASH scores (Figure 4H, Table 1). These results suggest that EP4 intervention can reduce HBx-dependent lipid accumulation and inflammation-related liver injury in NASH phenotypes in mice.
PMCs were further isolated by liver perfusion from different HBx-Tg mouse models to investigate the regional immunoregulatory function of liver macrophages. In the HBx-Tg and HBx-Tg + HFCD groups, the levels of p-AKTser473, p-mTORser2448, NPC1, ORP5, ERO1α, PDI, PERK, and ATF4 proteins increased in PMCs (Figure S6A,B,E); NPC1 and ORP5 protein levels increased in the liver (Figure S6C); cytoplasmic levels of free cholesterol and enhanced ER colocalization (blue-green) increased in PMCs (Figure S6D); the proportion of F4/80+/CD11b+ mononuclear-derived KC increased (Figure S6F); the proportion of M1 polarization of CD86+/CD206− increased, and the proportion of M2 polarization of CD206+/CD86− decreased (Figure 5A,B); EP4 and iNOS protein expression levels increased, and CD206 protein expression levels decreased in PMCs (Figure 5C); F4/80 and iNOS increased, and CD206 decreased in the liver (Figure 5D); and IL-6 was elevated in the serum and liver (Figure 5E,F). Moreover, after the ONO-AE3-208-induced inhibition of EP4, there was some improvement in the above indicators. Altogether, HBx-Tg mice subjected to HFCD exposure expressed HBx-associated NASH phenotypes; in the in vivo assay, it was demonstrated that HBx-expressing hepatocyte-derived PGE2 mediated hepatic MC polarization dysregulation via the mTOR-NPC1-ER stress signaling pathway; and EP4 receptor-targeted intervention rescued HBx-associated NASH mediated by hepatic MC polarization imbalance (Figure 6).
In the present study, we found an improved MC polarization imbalance after the inhibition of ERO1α and ATF4 in cholesterol-loaded macrophages, suggesting that ER redox and ER stress may mediate hepatic immune inflammatory responses by regulating MC polarization. Transient ER stress, a defense mechanism, protects hepatocytes from overexposure to nutrients such as lipids and/or exogenous agents such as viruses. Moreover, the chronic activation of ER stress is a central process in the transition from the simple steatosis stage to the NASH stage in NAFLD, including triggering cell death, inducing inflammatory responses, and accelerating metabolic disorders [33]. The initiation-sensing proteins of ER stress, PERK, IRE1α, and ATF6 inhibitors can alleviate the activation of cell death and inflammatory pathways as well as metabolic disorders. So, ER stress may be a new idea and a promising target for NASH intervention and treatment [34]. The latest study indicated that ER stress in innate immune cells contributes to NASH. For example, ATF6, by mediating a proinflammatory synergy between ER stress and TLR activation, is involved in the development of liver injury [35]. XBP1-mediated ER stress promoted steatohepatitis via NLRP3 inflammatory vesicle-mediated M1-type polarization [6]. Therefore, the interaction between liver inflammation and ER stress and the elucidation of ER stress-related mechanisms in macrophages can contribute to the prevention and treatment of NASH. In this study, we found that the inhibition of cholesterol transport from lysosomes to ER alleviated ER cholesterol accumulation-induced ER redox and ER stress by interfering with NPC1-ORP5, suggesting that NPC1-ORP5-mediated cholesterol transport between Lyso and ER may regulate hepatic immunometabolism. The downregulation of NPC1-ORP5 and the inhibition of cholesterol transport from Lyso to the ER were found by interfering with mTOR, suggesting that NPC1-ORP5-mediated cholesterol transport between Lyso and the ER may be regulated by mTOR. The role of Lyso-ER cholesterol transport-related molecules in macrophage polarization has been poorly reported and is an area that remains to be elucidated. Xu et al. found that simvastatin upregulates NPC1 via CYP7A1/LXRα signaling in ox-LDL-loaded macrophages to promote FC efflux from the lysosome, reduce the secretion of proinflammatory cytokines, and inhibit the M1 polarization phenotype, which is important for atherosclerosis intervention [36]. Borthwick et al. found that the overexpression of STARD3 caused significant upregulation of the receptor ABCA1 in macrophages, alleviated cholesterol ester accumulation by promoting cholesterol efflux, and suggested that enhanced lysosomal cholesterol transport induces an anti-atherogenic macrophage lipid phenotype [37]. The overexpression of ORP1L in macrophages causes decreased ABCG1 expression, and impaired cholesterol efflux promotes inflammatory signaling, leading to an M1-type polarization shift and exacerbating atherosclerotic lesions in Ldlr−/− mice [38]. In addition, it has been suggested that ox-LDL-induced lysosomal cholesterol accumulation plays a specific role in triggering inflammation and is a driver of atherosclerosis and NASH, and further attention should be paid to how to stimulate lysosomal cholesterol transport and block ox-LDL particle uptake interventions to prevent cholesterol accumulation in the lysosome [39]. Head et al. found that NPC1 mediates cholesterol transport via mTORC1 to promote angiogenesis [40]. These studies suggest that Lyso-ER cholesterol transport regulates the proinflammatory phenotype of foam macrophages, which is also associated with mTORC1. In this study, we found that the HBx expression of hepatocyte-derived PGE2 induced the upregulation of M1-type polarization and the downregulation of M2-type polarization in macrophages via EP4 receptors, suggesting that PGE2/EP4 may mediate communication between hepatic HC and MC cells to regulate the local immune microenvironment in the liver. PGE2 has been shown to influence NAFLD progression in insulin resistance, hyperglycemia, hepatic lipid accumulation, and inflammation [41]. The regulatory effect of PGE2 on macrophage polarization is not invariable and may be related to the specific microenvironment [42,43]. The EP4 receptor has been well described as an effective target for possible interventions in prostate, liver, colon, breast, skin, and vulvar cancers; NSAID-induced enteropathy; bone resorption, metabolism, and formation; atherosclerotic inflammation; rheumatoid arthritis; and other diseases [44]. The EP4 antagonist ONO-AE3-208 has been shown to inhibit malignant tumor invasion, migration, and metastasis [45]. However, as of today, no studies related to EP4 as an intervention target for NAFLD or NASH have been reported. Our previous study confirmed that COX-2 expression was upregulated in aflatoxin B1-treated hepatocytes and induced inflammatory liver injury. The inhibition of COX-2 with celecoxib improved the local immune microenvironment of the liver and blocked inflammatory injury [22]. On this basis, we proceeded to find that ONO-AE3-208, as an EP4 receptor antagonist of PGE2 inhibition, could improve the disturbance of hepatic lipid accumulation and KC polarization balance in mice to some extent, attenuate inflammatory injury, and ultimately rescue the HBx-associated NASH phenotype, suggesting for the first time that the EP4 receptor of PGE2 may become a potential target for intervention. In this study, we found that by improving the disturbance of hepatic cholesterol accumulation and KC polarization balance, we were able to reduce inflammatory injury and ultimately rescue the HBx-associated NASH phenotype, suggesting that the dysregulation of cholesterol homeostasis and MC polarization can be used as biomarkers of HBx-associated NASH immune injury. Reduced high-density lipoprotein cholesterol (HDL-C) and an elevated monocyte-to-HDL-C ratio (MHR) are significantly associated with mortality in HBV patients and can be used as prognostic biomarkers [46,47]. There is a complex association between the polarized phenotype of macrophages and HBV infection, with M1 activation indicating a strong immune response to HBV infection and M2 indicating persistent HBV infection, which is associated with disease progression [48]. In summary, we identified that PGE2-EP4-mTOR signaling regulates NPC1-ORP5 interaction mediating Lyso-ER cholesterol transport. ER-redox-dependent ER stress regulation induced by ER cholesterol is involved in mediating the M1 polarization of macrophages in HBx-induced ox-LDL-dependent NASH-associated inflammatory responses.
HBx-Tg mice were constructed as mentioned in our previous study [49]. For the transcriptome and untargeted metabolomic analyses of mouse liver, the complete liver lobes of mice in the WT group and the HBx group were used, with 2 mice in each group, and each liver lobe was tested twice, with a total of 8 samples. For processing and testing, the samples were entrusted to NovoGen Medical (Beijing, China). After the samples for transcriptomics were pretreated, mRNA extraction and detection, library construction and quality control, and up-sequencing were performed to obtain mRNA sequence information. Statistical methods such as DESeq, the negative binomial distribution model, and the padj BH value were used to compare gene mRNA-level differences and screen-related genes. The analysis process included differential significance analysis, GO function enrichment analysis, and KEGG pathway enrichment analysis. Using non-targeted metabolomics based on high-resolution mass spectrometry detection technology, we detected the characteristic peaks of the sample molecules and matched the metabolite identified in biological systems with the mzCloud database constructed by standards and with MassList and mzVault databases. The specific process included simple screening by the retention time, mass-to-charge ratio, and other parameters, adjusting the peak alignment according to the retention time and mass deviation, and then extracting the peak based on the information of the given mass deviation, signal-to-noise ratio, and additive ions. Then, the identified metabolites were compared with the database. Metabolites with a coefficient of variation of less than 30% in the quality control samples were retained as the final identification results. Finally, the identified metabolites were annotated with functions and classifications using databases such as KEGG. Multivariate statistical methods, such as principal component analysis and partial least-squares discriminant analysis, were used to filter the differential metabolites by downscaling and regression analysis. The raw histological data were normalized and, using the online bioinformatics website (https://hiplot.com.cn/, accessed on 2 April 2022), visualized as volcano maps, heat maps, and bubble maps. Comprehensive gene expression dataset analysis included the analysis of the transcript levels of the COX-2 gene (PTGS2) involved in the catalytic synthesis of PGE2 from liver tissues of HBV-infected patients in the GEO database (GSE83148) and the analysis of the transcription of PGE2 receptors, mTOR components, lysosomal cholesterol transporters, ER redox and stress, and M1/2 polarization markers in NASH mouse liver macrophages in the GEO database (GSE104901). Differential genes were screened out to provide clues for subsequent experiments.
PGE2 was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA); ox-LDL was purchased from Guangzhou Yiyuan Bio (Guangzhou, China); U18666A and EN460 were purchased from MedChemExpress (Monmouth Junction, NJ, USA); ONO-AE3-208 was purchased from CSNpharm (Arlington Heights, IL, USA); siNPC1, siERO1A, siATF4, and siNC were purchased from RiboBio (Guangzhou, China); Rapa was purchased from Sigma (St. Louis, MO, USA); and doxycycline hydrochloride was purchased from Macklin (Shanghai, China). Primary antibodies: Anti-ERO1α, -EP4, -iNOS, -CD206, and -XBP1 were purchased from Abcam (Waltham, MA, USA); anti-PERK, -mTOR, -p-mTORSer2448, and -IRE1α were purchased from Cell Signalling Technology (Danvers, MA, USA); anti-NPC1, -STARD3, -HBx, and -β-actin were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA); anti-ORP5 was purchased from Bioss (Beijing, China); anti-SLC38A9 was purchased from NOVOS Biologicals (Littleton, CO, USA); anti-COX2 was purchased from HUABIO (Hangzhou, China); anti-Akt and -p-AKTSer473 were purchased from Affinity Biosciences (Liyang, China); anti-ATF4 was purchased from Proteintech (Chicago, IL, USA); anti-CHOP was purchased from Ruiying Biological (Suzhou, China); and anti-ATF6 and -p-PERKT982 were purchased from ABclonal Technology (Wuhan, China). Secondary antibodies: Anti-mouse IgG (H+L) was purchased from Invitrogen (Carlsbad, CA, USA), anti-rabbit IgG (H+L) was purchased from Thermo Scientific (Waltham, MA, USA), and fluorescent secondary antibodies 488-labeled and 647-labeled IgG were purchased from Beyotime (Shanghai, China).
HBx-expressing hepatocytes: With pcDNA3.1 plasmid as a control, HepG2 cells were transfected with a pcDNA3.1-HBx plasmid (1 μg) for 8 h and then recovered for 1 day; with dimethyl sulfoxide as a control, the expression of HBx was induced in HepG2-Teton-HBx cells by treatment with doxycycline hydrochloride (1 μg/mL, 2d). Co-culture of hepatocytes and macrophages: HepG2 cells recovered from transfection were plated on the upper layer of the chamber, and THP-1 cells pretreated with ox-LDL (1 μg/mL, 1d) were plated on the lower layer of the chamber. Culturing and processing of mononuclear-derived macrophage THP-1: THP-1 cells were treated with phorbol ester (100 ng/mL, 1 d) to induce their differentiation into macrophages. Using dimethyl sulfoxide as a control, cells were treated with ox-LDL (50 μg/mL) for 1 day to build a cholesterol-load model, ox-LDL+PGE2 (0.25 μM) for 1 day to establish a liver injury exposure model, ox-LDL+PGE2+ONO-AE3-208 (1 μM) for 1 day to construct the EP4 intervention model, and ox-LDL+PGE2+Rapa (50 nM) for 1 h to construct the mTOR intervention model. With siNC as a control, cells were transfected with ox-LDL+PGE2+siNPC1 (50 nM) for 8 h and recovered for 1 day to construct the NPC1 intervention model, ox-LDL+PGE2+siERO1A was used to construct the ERO1α intervention model, and ox-LDL+PGE2+siATF4 was used to construct the ATF4 intervention model. The pcDH-ERO1A cell line with high expression of ERO1α was established with the pcDH vector as the control.
Western blotting analysis was performed as previously described [50]. According to the standard procedures commonly used in the laboratory, detection operations were performed: gel preparation, electrophoresis, transfer, blocking, primary antibody incubation, secondary antibody incubation, and luminescence development.
According to the corresponding experimental design, when approaching the end point of the treatment, referring to the instructions of the kit, PGE2 (MEIMIAN, Yancheng, China), IL-6 (SINOBESTBIO, Shanghai, China), and ox-LDL (RUIDA HENG HUI, Beijing, China) were measured by a multi-function microplate reader at 450 nm.
Immunofluorescence analysis was performed as described previously [51]. Cells were incubated with the lipid droplet probe (BODIPY™ 493/503, green) (Thermo Scientific, MA, USA), free cholesterol probes (Filipin complex, blue) (Sigma, MO, USA), lysosome probes (Lyso-Tracker, red) (Molecular Probes, Waltham, MA, USA), endoplasmic reticulum probes (ER-Tracker, green) (Molecular Probes, MA, USA), or target protein antibody or nuclear dye DAPI near the end of the treatment according to the appropriate experimental design. A high-sensitivity laser confocal microscope (LSM 880, Zeiss, Jena, Germany) was used to observe and analyze the colocalization of free cholesterol and the lysosome or ER, the expression and colocalization of NPC1 and ORP5, and the level of ATF4 in the nucleus.
According to the corresponding experimental design, the DCFH-DA probe (Beyotime, Shanghai, China) was incubated for 30 min according to the kit instructions at the end of the treatment, and the green fluorescence intensity was recorded to detect the ROS level by the FITC channel parameter using a flow cytometer.
Transmission electron microscopy analysis was performed as described previously [51]. Referring to standard procedures commonly used in the laboratory, electron microscopy samples were prepared near the end of the treatment process and sent to the electron microscopy room to observe the morphology of lysosomes and the ER.
According to the corresponding experimental design, the operation was carried out according to the kit (NJJCBIO, Nanjing, China) instructions at the end of the treatment. Reduced glutathione (GSH) = Total glutathione − 2 × oxidized glutathione (GSSG).
Basic operations were performed as described previously [22]. In brief, the mice were anesthetized by intraperitoneal injection of 10% urethane with 1% body weight. The calcium-free perfusate was slowly injected until the liver was swollen. The collagenase perfusate was slowly injected. The perfusion was stopped and followed by digestion for 10 min at 37°C. After filtration with a cell strainer and centrifugation at 70× g for 3 min, the precipitate was primary hepatocytes (PHCs), and the supernatant (non-parenchymal cells) was taken to further isolate primary macrophages (PMCs). The supernatant was centrifuged at 500× g for 3 min, the supernatant was removed, and the pellet was resuspended; this process was repeated twice. The pellet was resuspended in 5 mL of the medium, the suspension was centrifuged at 50× g for 3 min, and the supernatant was transferred; this process was repeated twice. Supernatants were collected and centrifuged at 500× g for 3 min for a total of 3 times, and the precipitate was PMCs.
The PHCs of the WT group and the HBx group were placed on the upper layer of the chamber, and PMCs were placed on the lower layer of the chamber. After adhering to the wall, Rapa or U18666A and EN460 were added as a treatment for 1 day.
Eight–twelve-week-old C57BL/6 and C57BL/6-HBx-Tg mice were randomly divided into 5 groups (n = 5): the control group (wild type), HBx group, HBx + HFCD group (high-fat and high-cholesterol diet for 10 weeks, positive NAFLD/NASH model group), HBx + ONO-AE3-208 group (5 mg/kg, intraperitoneal injection, once every other day for 2 weeks, EP4 intervention group), and HBx-Tg + HFCD + ONO-AE3-208 group. The animals were kept in an independent ventilation cage system to ensure sufficient water and food and a clean environment. The whole process complied with the normative requirements of the Xiamen University Laboratory Animal Management Ethics Committee.
Referring to the kit instructions, the levels of ALT (NJJCBIO, Nanjing, China), AST (NJJCBIO, Nanjing, China), ox-LDL, PGE2, and IL-6 were detected.
Referring to the kit instructions, the levels of ox-LDL, TG, TC, FC (APPLYGEN, Beijing, China), PGE2, and IL-6 were detected and corrected with the protein concentration.
Block Fc receptors: A total of 0.5–1 μg of CD16/32 monoclonal antibody was added and incubated at room temperature for 15 min. Cell staining: A total of 5 μL each of flow-through antibodies PerCP/Cyanine5.5-F4/80, FITC-CD11b, PE-CD206, and APC-CD86 (Elabscience, Wuhan, China) was added, mixed, and incubated for 45 min in the dark. The cells were resuspended in PBS with 1% BSA, the cell suspension was centrifuged at 300× g for 5 min, and the supernatant was discarded. Cells were resuspended in 1% BSA in PBS and filtered into flow tubes for detection and analysis by flow cytometry.
Liver samples of suitable size were fixed in 4% paraformaldehyde at 4°C overnight; then, the paraformaldehyde was replaced, and the samples were used to prepare paraffin sections. Liver samples of suitable size were embedded in OCT gel, placed at −80°C overnight, and then used to prepare frozen sections. Immunofluorescence (IF) assay: Frozen sections were taken, incubated with a lipid droplet probe (BODIPY™ 493/503, Thermo Scientific, MA, USA) and antibody to the macrophage marker F4/80, and the expression and distribution of lipid droplet levels were observed and analyzed with F4/80 using a high-sensitivity laser confocal microscope (LSM 880). Immunohistochemistry (IHC) assay: IHC experiments were performed using IHC kits (Maxim, Fuzhou, China), and the expression of related proteins was observed by incubating cells with anti-F4/80, -iNOS, -CD206, -NPC1, and -ORP5 according to standard laboratory procedures. Hematoxylin–eosin staining: Paraffin sections were taken following the standard laboratory procedures for dewaxing and hydration, hematoxylin staining, eosin staining, dehydration, and mounting to observe the level of liver tissue damage by microscopic examination. Oil Red O staining: Frozen sections were taken, and the level of lipid accumulation in the liver was observed according to the instructions of the kit (NJJCBIO, Nanjing, China). NASH scoring [52]: According to the analysis of histological features in the slices, the degree of liver steatosis was divided into 4 grades (0–3), the number of inflammatory reaction foci was divided into 4 grades (0–3), and the degree of hepatocyte damage (balloon degeneration) was divided into 3 grades (0–2). A total score of 0–3 was judged as non-NASH (N), while a score of 4–8 was considered NASH.
GraphPad Prism 7.0 and CorelDRAW-X4-SP2 software were used to analyze and draw charts, in which the data are presented in the form of mean ± standard deviation (SD). The specific statistical analysis methods were as follows: An independent sample t-test was used for the comparison of two groups of data. One-way analysis of variance was used for data comparison between multiple groups. Further analysis of the differences between pairs was performed by Dunnett’s t-test. The correlation was analyzed by Pearson’s correlation coefficient, and p < 0.05 means that the difference is statistically significant. | true | true | true |
PMC9569612 | 36232714 | Ekaterina I. Romanova,Anatoliy V. Zubritskiy,Anna V. Lioznova,Adewale J. Ogunleye,Vasily A. Golotin,Anna A. Guts,Andreas Lennartsson,Oleg N. Demidov,Yulia A. Medvedeva | RUNX1/CEBPA Mutation in Acute Myeloid Leukemia Promotes Hypermethylation and Indicates for Demethylation Therapy | 27-09-2022 | AML,DNA methylation,RUNX1,CEBPA,TET2,epigenetics,BIK | Acute myeloid leukemia (AML) is a rapidly progressing heterogeneous disease with a high mortality rate, which is characterized by hyperproliferation of atypical immature myeloid cells. The number of AML patients is expected to increase in the near future, due to the old-age-associated nature of AML and increased longevity in the human population. RUNX1 and CEBPA, key transcription factors (TFs) of hematopoiesis, are frequently and independently mutated in AML. RUNX1 and CEBPA can bind TET2 demethylase and attract it to their binding sites (TFBS) in cell lines, leading to DNA demethylation of the regions nearby. Since TET2 does not have a DNA-binding domain, TFs are crucial for its guidance to target genomic locations. In this paper, we show that RUNX1 and CEBPA mutations in AML patients affect the methylation of important regulatory sites that resulted in the silencing of several RUNX1 and CEBPA target genes, most likely in a TET2-dependent manner. We demonstrated that hypermethylation of TFBS in AML cells with RUNX1 mutations was associated with resistance to anticancer chemotherapy. Demethylation therapy restored expression of the RUNX1 target gene, BIK, and increased sensitivity of AML cells to chemotherapy. If our results are confirmed, mutations in RUNX1 could be an indication for prescribing the combination of cytotoxic and demethylation therapies. | RUNX1/CEBPA Mutation in Acute Myeloid Leukemia Promotes Hypermethylation and Indicates for Demethylation Therapy
Acute myeloid leukemia (AML) is a rapidly progressing heterogeneous disease with a high mortality rate, which is characterized by hyperproliferation of atypical immature myeloid cells. The number of AML patients is expected to increase in the near future, due to the old-age-associated nature of AML and increased longevity in the human population. RUNX1 and CEBPA, key transcription factors (TFs) of hematopoiesis, are frequently and independently mutated in AML. RUNX1 and CEBPA can bind TET2 demethylase and attract it to their binding sites (TFBS) in cell lines, leading to DNA demethylation of the regions nearby. Since TET2 does not have a DNA-binding domain, TFs are crucial for its guidance to target genomic locations. In this paper, we show that RUNX1 and CEBPA mutations in AML patients affect the methylation of important regulatory sites that resulted in the silencing of several RUNX1 and CEBPA target genes, most likely in a TET2-dependent manner. We demonstrated that hypermethylation of TFBS in AML cells with RUNX1 mutations was associated with resistance to anticancer chemotherapy. Demethylation therapy restored expression of the RUNX1 target gene, BIK, and increased sensitivity of AML cells to chemotherapy. If our results are confirmed, mutations in RUNX1 could be an indication for prescribing the combination of cytotoxic and demethylation therapies.
Acute myeloid leukemia (AML) is a hematologic malignancy characterized by the proliferation of immature myeloid progenitors and substantial genetic, cytogenetic, and epigenetic abnormalities. Recent cohort studies have shown that about 45% of AML patients are characterized by a normal karyotype [1] and the presence of a few core driver mutations [2] in transcription factors and epigenetic regulators (DNMTs, TET2, ASXL1 and IDH1/2 [3]). Cytotoxic drug Cytarabine (Ara-C), in combination with etoposide or alone, remains the first line of treatment for AML patients. DNA-Ademethylating drugs Decitabine and Azacitidine supplement cytotoxic therapy of AML with mutations in epigenetic regulators [4]. Recently, new drugs have been approved for use, including Enasidenib and Ivosidenib—inhibitors targeting recurrent mutations in IDH1/2 genes, respectively, thus extending the therapeutic landscape for AML [5,6]. RUNX1 and CEBPA, transcriptional factors essential for normal hematopoiesis, are also frequently mutated in AML [7]. Point mutations in RUNX1 are detected in 6–33% of cytogenetically heterogeneous AML patients [8,9,10,11]. RUNX1 mutations are almost mutually exclusive to AML, with recurrent genetic abnormalities, and they co-occur with a complex pattern of gene mutations, frequently involving mutations in epigenetic modifiers (ASXL1, IDH2, KMT2A, EZH2). RUNX1-mutated AML is associated with distinct clinicopathologic features and inferior prognosis, depending on a spectrum of co-occurring mutations [12]. CEBPA mutations occur in about 7–15% of AML cases, and most of them are double (bi-allelic) mutations [13,14]. Biallelic CEBPA mutation (biCEBPA) in AML with a normal karyotype is distinguished as a distinct AML subtype and is associated with a favorable clinical outcome [15,16,17]. Monoallelic CEBPA-bZip mutations are also associated with favorable prognosis in children [18]. TET enzymes play a major role in active DNA demethylation [19]. TET2, which is commonly mutated in AML [20], does not have a DNA binding domain, and requires molecular partners—usually transcriptional factors (TFs)—to guide it to specific genomic locations [21,22,23,24,25]. In normal conditions, both RUNX1 and CEBPA could contribute to attracting TET2 to their binding sites (TFBS), causing demethylation of regulatory regions and keeping the corresponding genes active [22,23,25,26]. We hypothesized that this mechanism might be disturbed in AML in patients with RUNX1 or CEBPA mutation. We presume that the lack of RUNX1/CEBPA may reduce TET2-induced demethylation near RUNX1/CEBPA binding sites, indirectly affecting the DNA methylation of regulatory regions and keeping RUNX1 and CEBPA-regulated genes inactive. In line with this hypothesis, we show that chemical demethylation improves sensitivity of such cells to state-of-the-art AML chemotherapy, at least partially through activation of the pro-apoptotic gene BIK.
For this study, we used data on DNA methylation and gene expression from The Cancer Genome Atlas (TCGA) AML cohort of 186 patients. Of 186 patients, 17 (9%) and 13 (7%) patients had mutations in RUNX1 and CEBPA, respectively. The types of the mutations are described in Supplementary Table S1. In patients with RUNX1/CEBPA mutation, 5/1 patients had nonsense mutations and 4/8 patients had frameshifts, respectively. These mutations most likely affect the protein structure or function. The rest of the mutations were missense or in-frame insertions or deletion, the effects of which, on the protein function, could be subtle. We split patients in two groups and compared DNA methylation levels in patients with and without a corresponding mutation. CpGs positions within close proximity to RUNX1/CEBPA binding sites (TFBS) showed an increase in DNA methylation in AML patients with corresponding mutations in comparison to patients without mutations (Figure 1A,B,D,E,G,H). For the majority of CpGs, only a mild gain in methylation in RUNX1-mutated AML patients was observed (Figure 1A, Supplementary Figure S1A, average Δmeth = 0.06); conversely, for the differentially methylated CpG in close proximity to RUNX1 TFBS, the gain in DNA methylation was significantly higher (Figure 1A, Supplementary Figure S1A, average Δmeth = 0.18). A similar but less significant tendency was observed for the patients with mutations in CEBPA: the gain in DNA methylation genome-wide was significantly lower than that near CEBPA TFBS (Figure 1G, Supplementary Figure S1C, average Δmeth = 0.09 vs. average Δmeth = 0.13). This tendency did not occur for CEBPA TFBS near hypermethylated CpG TL, but the difference was insignificant, due to a low number of such CpGs (Figure 1J,K). CEBPA mutation leads to a higher level of genome-wide hypermethylation (Figure 1G, gray dots, average Δmeth = 0.09) when compared to RUNX1 mutation (Figure 1A, gray dots, average Δmeth = 0.06). In the case of RUNX1 mutation, genome-wide hypermethylation was almost lost for CpG TL (Supplementary Figure S1B, gray violines)—functional CpG sites associated with changes in gene expression (see Methods). However, in the case of CEBPA mutation, CpG TL tended to demonstrate hypermethylation independent of the presence of CEBPA TFBS (Supplementary Figure S1D, gray violines). The mechanism behind genome-wide hypermethylation of functional CpGs in patients with CEBPA mutation is unclear, since the majority of genomic CpGs do not have CEBPA TFBS nearby. This suggests that CEBPA regulation of the associated genes might be indirect or specific to a subpopulation of patients, as shown in the work of Figueroa [27]. On average, genes associated with CpGs that dramatically change DNA methylation (strong CpG TL) located near RUNX1 were significantly downregulated (Supplementary Figure S2). On the contrary, CEBPA mutation leads to downregulation of only a few genes. We focused on genes with the most significantly changed methylation and expression levels in the case of TF mutation (FC > 2, FDR < 0.05, absolute expression value > 0.5). We found 12 and 11 genes that meet these criteria for AML patients with RUNX1 and CEBPA mutation, respectively (Supplementary Tables S3 and S4). For CEBPA mutation, these genes are HOXA10, HOXA9, PTRF, TNS3, FSTL1, GPR109B, PI4K2A, ECE1, LOC283663, SCHIP1, and MFSD2A. Only two of these genes (HOXA9, HOXA10) show general prognostic significance. Hypermethylation of HOXA9 and HOXA10, independent of the patients’ mutation profile is linked to a significantly improved survival rate (Supplementary Figure S3). For RUNX1 mutation, these genes are BIK, OSBPL5, LGALS3BP, KRT18, CACNA2D4, C20orf197, HOXB3, TNFRSF10C, C10orf91, VSTM1, ZBTB16, and C16orf93. Downregulation of OSBPL5 has been previously reported in a subtype of AML [28], while downregulation of BIK was observed in multiple cancers (reviewed in [29]). Hypermethylation of BIK is associated with a worse survival rate in the long run (t > 500 days, p-value = 0.002); however, for the overall survival rate, the difference is insignificant, due to a very similar probability of short-term survival (Supplementary Figure S4). Methylation levels of strong CpG TL in AML patients with RUNX1/CEBPA mutation resemble those in immature cells observed in normal hematopoiesis (Figure 1F,L). Cell type deconvolution analysis [30] showed a significant decrease in peritoneal macrophages in AML patients with both mutations and in monocytes in the case of CEBPA mutation (Supplementary Figure S5A,B). This observation suggests that demethylation of RUNX1 and CEBPA binding sites is required for normal myelopoiesis, and that this program is disrupted in AML with RUNX1 or CEBPA mutations.
In a recent study, Suzuku et al. showed, through co-immunoprecipitation, that RUNX1 and CEBPA directly interacts with TET2 and, in this way, could recruit it to their TFBS in human cell lines [23]. We hypothesized that this mechanism might be implicated in AML patients with RUNX1 and CEBPA mutation. We presume that the lack of fully functional RUNX1/CEBPA may reduce TET2-induced demethylation of nearby RUNX1/CEBPA binding sites, indirectly affecting the DNA methylation profile of the patients and keeping RUNX1 and CEBPA-regulated genes inactive. To verify this hypothesis, we used the TET2 profile determined by ChIP-seq (see Methods for the details). Indeed, the amount of TET2 is significantly increased in close proximity of RUNX1 and CEBPA TFBS in cells with intact RUNX1 and CEBPA (Figure 1C,I), supporting the suggestion that TET2 could be involved in demethylation of these CpGs in normal conditions. Several genes regulated by RUNX1 or CEBPA demonstrated a significant increase in DNA methylation and a decrease in expression in AML patients with RUNX1 or CEBPA mutation, respectively (Supplementary Tables S3 and S4). Next, we focused on genes with the most significantly changed methylation and expression levels in the case of TF mutation (FC > 2, FDR < 0.05, absolute expression value > 0.5). BIK/OSBPL5 and HOXA9 were among those genes in the case of RUNX1 and CEBPA mutation, respectively. ChIP-qPCR confirmed reduced TET2 presence in OSBPL5 and BIK genes in the OCI-AML5 cell line (a line with a reported RUNX1 mutation (Figure 2A,B)), and in HOXA9 in the Kasumi-6 cell line (a line with a reported CEBPA mutation (Supplementary Figure S6)), supporting the role of TET2 in TFBS demethylation.
Cytarabine (Ara-C), in combination with etoposide or alone, remains the first line of treatment for AML patients. AML cell lines with RUNX1 mutations OCI-AML5 and Mono-Mac1 were more resistant to Ara-C than wild-type (wt) RUNX1 AML cell line OCI-AML2 (Figure 2C). To restore the sensitivity of RUNX1-mutated AML cells to chemotherapy, we pretreated AML cells with the demethylating agent azacytidine (AZA). AZA did not potentiate Ara-C-dependent cytotoxicity in RUNX1 wild-type OCI-AML2 cells, but significantly reduced the viability of AML cells bearing mutant RUNX1, OCI-AML5, and MonoMac1 (Figure 2D). Demethylation significantly potentiated the cytotoxic effect of chemotherapy in AML with RUNX1 mutations. Moreover, treatment with a demethylating agent increased the expression of pro-apoptotic gene BIK, both at mRNA and protein levels, which can explain the sensitization of RUNX1-mutated leukemic cells to chemotherapy by Ara-C (Figure 2E,F).
One of the distinctive features of AML is a significant disturbance in epigenetic regulation. Mutations in IDH1/2, TET2, and DNMT3A are known to be early events in the development of AML and can be detected years before diagnosis [31,32]. We presented evidence that in AML patients, TET2 is specifically involved in the DNA demethylation near TFBS of RUNX1 and CEBPA, particularly in demethylation of CpG TL—functional CpG positions in which methylation is a marker of the gene expression nearby. Thus, a lack of fully functional RUNX1 or CEBPA proteins in AML patients with a corresponding mutation could prevent the interaction of TET2 with the regions around their TFBS, leading to methylation of regulatory regions and suppression of several genes’ expressions. Regulatory regions of RUNX1/CEBPA target genes were affected by lack of TET2 and, as a result, demonstrated extensive hypermethylation. Changes in DNA methylation followed different scenarios in patients with mutations of both types: in the case of RUNX1 mutation, the most pronounced changes happened in CpG TL near RUNX1 TFBS, while, in case of CEBPA mutation, we observed a DNA methylation change in genome-wide CpG TL, independent of the presence of CEBPA TFBS, supporting the idea that CEBPA may not only affect DNA methylation through attraction of TET2. Recently, it has been shown that CEBPA interacts with DNMT3A N-terminus and, in this way, blocks DNMT3A from accessing DNA substrate, thereby inhibiting its activity [33]. Thus, this suggests that regulation of DNA methylation by CEBPA has a complex nature. However, we were able to detect only a few genes that respond to hypermethylation caused by CEBPA mutation. Among those genes whose expression is strongly affected by CEBPA mutation, we identified HOXA9 and HOXA10. We showed that hypermethylation of HOXA9 and HOXA10 genes was linked to significantly improved rates of long-term survival, most likely indirectly contributing to the long-term survival rates detected in patients with double CEBPA mutations [17]. Thus, methylation levels of HOXA9/HOXA10 could be considered prognostic markers in AML. It has previously been shown [8] that patients with RUNX1 mutation demonstrate lower survival rates. Changes in TET2-dependent levels of DNA methylation close to RUNX1 TFBS in AML patients with a corresponding mutation lead to the repression of multiple regulated genes, including pro-apoptotic gene BIK. Although patients with both RUNX1 and CEBPA mutations demonstrated the most significant increase in DNA methylation near their TFBS in CpG TL—functional CpG positions whose methylation correlates to the expression of a nearby gene—the downstream effect on gene expression of RUNX1 is more pronounced. This observation is in line with the more severe consequences of RUNX1 mutation on patients’ survival. The chromosomal translocation t(8;21)(q22;q22), generating the RUNX1/RUNX1T1 fusion gene, is the most prevalent chromosomal rearrangement in AML, with an incidence rate of about 15% in children and young adults [34]. The translocation produces a fusion protein composed of the RUNX1 DNA-binding Runt domain and the almost complete open reading frame of RUNX1T1. Since the resulting fusion protein lacks a TET2 binding domain, we believe that such a rearrangement should prevent the fusion protein from attracting TET2 to its TFBS and demethylating the regions [35,36]. This observation is supported by the epigenetic heterogeneity in patients with RUNX1/RUNX1T1 fusion [37]. We used the OCI-AML5 cell line as a proxy for AML cells with a RUNX1 mutation. However, this cell line also has 41 more mutated genes, including TET2 mutation (S825*) [38]. A more accurate study of the mutations in OCI-AML5 cell line shows bi-allelic mutations S825*/Y1148C in the TET2 gene [39]. It is not entirely clear how all these mutations may affect TET2-dependant regulation. Since one allele of TET2 in OCI-AML5 contains only a substitution to relatively similar amino acids, we hypothesize that levels of functional TET2 protein could be partially reduced, but not totally eradicated. However, global DNA methylation levels are not changed in this cell line [40], suggesting that even partially functional TET2 can perform demethylation. Although the evidence is indirect, we believe that the changes we observed in RUNX1 target genes in OIC-AML at least partially occur due to RUNX1 mutation. Similarly, we used the Kasumi-6 cell line as a cell line with a CEBPA mutation. However, this cell line also has RUNX1 overexpressed. It has been previously shown [23] that overexpression of RUNX1 leads to the increased levels of TET2 near RUNX1 binding sites. This is despite RUNX1 and CEBPA biding sites being quite different (Supplemental Figure S7), so it is highly unlikely that increased binding of RUNX1 will affect binding of CEBPA in any way, or that it will increase the TET2 signal near CEBPA TFBS. Even if, in some cases, there is a cooperative RUNX1-CEBPA binding, the increased binding of RUNX1 should have an increased TET2 signal; however, we observed a decreased TET2 signal in the KASUMI-6 cell line, supporting the hypothesis that overexpression of RUNX1 should not interfere with CEBPA mutation-related effects. Due to a lack of TET2 ChIP-seq data in AML-related cells, our results on genome-wide TET2 distribution should be considered with caution. We managed to support these results for several genes with ChIP-qPCR (Figure 2A,B, Supplementary Figure S4), suggesting that genome-wide distribution of TET2 in MCF cells could be considered a decent proxy for the TET2 distribution in other cells. Our results indicate that methylation patterns are significantly changed not only in AML patients with mutations in epigenetic regulators, but also in AML patients carrying mutations in RUNX1 or CEBPA. Currently, mutations in epigenetic regulators, such as IDH1/2 and DNMT3A, in AML patients serve as molecular predictors of a good response to therapy with hypomethylating agents (HMA) such as Azacitidine or Decitabine. Recently, a therapeutic strategy for the treatment of CEBPA-mutated leukemia with DNA-hypomethylating agents has been suggested. Moreover, we demonstrated the reactivation of the expression of pro-apoptotic protein BIK (Bcl-2-interacting killer) by HMA in AML cells. We also showed that hypermethylation of BIK might have an effect on long-term patient survival. This is in line with the previously proposed idea that therapeutic approaches to activate the pro-apoptotic BH3-only genes, including BIK, might improve the clinical outcome of chemotherapy treatments in drug-resistant AML [29]. BIK-associated mechanisms may be partially responsible for the observed complementary effect of the recently FDA-approved protocol for the treatment of AML in elderly patients, which implements a combination of Bcl-2 inhibitor Venetoclax and HMA [41]. Our data advocates for the rationale of prescribing epigenetic therapy with hypomethylating agents for the subtypes of AML with mutations in transcriptional factor RUNX1, although this hypothesis would need to be confirmed in a prospective AML cohort.
Mutations, DNA methylation (Illumina 450k Array beta-values) and gene expression (RNA-seq, RPKM) profiles for the bone marrow of 186 AML patients were obtained from The Cancer Genome Atlas (TCGA) [42]. Methylation beta-values (Illumina 450k array) were obtained by the TCGA consortium. DNA methylation data were generated using the ‘EGC.tools’ R package (version 1.3.0) after processing raw IDAT files for each sample with the ‘methylumi’ R package (version 2.3.22) by the TCGA consortium.
DNA methylation profiles in different stages of normal granulopoiesis (Illumina 450k beta-values) were obtained from FACS-sorted bone marrow cells of voluntary healthy donors [43]. The data were preprocessed with the Genome Studio module 1.8. by [43]. Genome-wide TET2 location profiles (ChIP-Seq) in the MCF7 cell line were obtained from the work of Wang and colleagues [44].
RUNX1 and CEBPA TFBS were annotated in all CpG-centered 200bp regions (+/−100 bp) using positional weight matrices (PWM) from HOCOMOCO v11 (p-value < 0.001) [45]. We predicted 53,443 sites for RUNX1 and 47,157 for CEBPA. To reduce the number of false positives, we filtered out those predicted TFBS that did not co-locate with the ChIP-seq peak for a corresponding TF from Cistrome (A, B and C categories only) [46]. As a result, we obtained 10570 sites for RUNX1 and 5263 for CEBPA, respectively.
To determine CpG sites critical for gene regulation, we used a methodology similar to the one we recently published [47]. Briefly, for each 336478 CpG associated with a gene (Illumina 450k Array annotation), we calculated Spearman correlation coefficient (SCC) between the CpG methylation and the expression profile across all patients. A CpG dinucleotide with a significant SCC (FDR < 0.005, Benjamini–Hochberg method (BH)) was considered critical for gene regulation and referred to as a CpG Traffic Light (CpG TL). In this way, we detected 25928 CpG TLs. We defined strong CpG TLs as those CpG TLs that not only match the criteria of CpG TL, but also demonstrated a significant change in gene expression and DNA methylation (see below) in AML patients with vs. without a mutation in RUNX1/CEBPA gene, respectively. We detected differentially methylated CpGs (DM CpG) between patients with a mutation in RUNX1/CEBPA and those without a mutation (Student’s t-test, FDR < 0.05, BH). We considered a DM CpG as differentially hypermethylated CpG (DHM CpG) if the mean difference in AML patients with RUNX1/CEBPA mutation and those without it was greater than 0.15. Differentially expressed genes (DEG) were selected in a similar way (Student’s t-test, FDR < 0.05, BH). TET2 ChIP-Seq analysis was performed with Deeptools [48]. Detection of a cell type composition was performed with an Epidish package [30]. To cover AML cell types more accurately, the default set of DNA methylation profiles was supplemented with the DNA methylation profiles of mononuclear cells from AML patients [49] and immature blood cells [43].
Chromatin immunoprecipitation (ChIP) was performed as described previously [50] with the antibody to TET2 in OCI-AML2 cells (wt RUNX1) and OCI-AML5 (mutated RUNX1). Eighteen million AML-2 or AML-5 cells were washed twice with ice-cold PBS and resuspended in 6 mL of ice-cold PBS. The cells were first fixed with 2 mM DSG for 30 min with rotation at room temperature (RT), after which 37% formaldehyde was added to achieve a final concentration of 1%, and samples were rotated for another 10 min. Glycine was added to a final concentration of 125 mM, and the fixation was quenched for 5 min at RT. Fixed cells were washed twice with ice-cold PBS and resuspended in the Sodium dodecyl sulfate (SDS) buffer. After 10 min rotation at RT, cells were spun down for 2 min at 12,000 RCF, resuspended in the IP buffer and sonicated with Bioruptor (Diagenode). Upon sonication, chromatin was diluted with an SDS-free IP buffer to reach the final SDS concentration of 0.1%. Cellular debris was cleared by centrifugation at 16,900 RCF (4 °C) for 20 min, leaving a chromatin-rich suspension. After centrifugation, the supernatant was divided into input (1% vol), IP (67% vol) and mock (32% vol). The IP sample was incubated with 1ug anti-TET2 (N-term) Rb polyclonal antibody (ab230358) for 18 h with rotation at 4 °C. Mock samples were incubated with no antibody under the same conditions. Next, the IP and mock samples were incubated for 3 h with Protein G SureBeads (BioRad, Hercules, CA, USA), pre-blocked with BSA and salmon sperm DNA, with rotation at 4 °C. Beads were washed 3 times with a low-salt buffer, 2 times with a high-salt buffer, 1 time with an IP buffer containing 0.1% SDS. Protein G beads were resuspended in the decrosslinking buffer and incubated overnight at 65 °C with Proteinase K treatment. After decrosslinking, the DNA was extracted with the Qiagen MinElute kit. For the primers sequences used for qPCR, see Supplementary Table S2.
Ara-C (20 μm; Cytosar, Astellas Pharma Inc., Tokyo, Japan) were added to experimental wells after 48 h of cultivation. To test the combination therapy with AZA (3 μM, decitabine, Sigma-Aldrich, MO, Saint-Louis, USA), cells at a concentration of 1 million per ml were precultured for 48 h with 3 μM AZA. Thereafter, Ara-C was added at a concentration of 12.8 μM, and culture was continued for an additional 72 h.
The CyQUANT™ XTT Cell Viability Assay kit was used to analyze cell viability. To test the residual viability after exposure to cytarabine (Ara-C), 1 million cells per ml were seeded on a 96-well plate with various concentrations of Ara-C (up to 10 μM) and cultured in humid conditions at 37 °C and 5% CO2 for 72 h. The viability test was carried out using the XTT reagent spectrophotometrically at a wavelength of 490 nm.
Western blots were performed using anti-BIK antibodies (Cell Signaling, Danvers, MA, USA). | true | true | true |
PMC9569777 | 36232782 | Irshad Ahmad | CRISPR/Cas9—A Promising Therapeutic Tool to Cure Blindness: Current Scenario and Future Prospects | 29-09-2022 | CRISPR,retinal degeneration,eye diseases,blindness,eye therapeutics,viral vectors,non-viral vectors | CRISPR-based targeted genome editing is bringing revolutionary changes in the research arena of biological sciences. CRISPR/Cas9 has been explored as an efficient therapeutic tool for the treatment of genetic diseases. It has been widely used in ophthalmology research by using mouse models to correct pathogenic mutations in the eye stem cells. In recent studies, CRISPR/Cas9 has been used to correct a large number of mutations related to inherited retinal disorders. In vivo therapeutic advantages for retinal diseases have been successfully achieved in some rodents. Current advances in the CRISPR-based gene-editing domain, such as modified Cas variants and delivery approaches have optimized its application to treat blindness. In this review, recent progress and challenges of the CRISPR-Cas system have been discussed to cure blindness and its prospects. | CRISPR/Cas9—A Promising Therapeutic Tool to Cure Blindness: Current Scenario and Future Prospects
CRISPR-based targeted genome editing is bringing revolutionary changes in the research arena of biological sciences. CRISPR/Cas9 has been explored as an efficient therapeutic tool for the treatment of genetic diseases. It has been widely used in ophthalmology research by using mouse models to correct pathogenic mutations in the eye stem cells. In recent studies, CRISPR/Cas9 has been used to correct a large number of mutations related to inherited retinal disorders. In vivo therapeutic advantages for retinal diseases have been successfully achieved in some rodents. Current advances in the CRISPR-based gene-editing domain, such as modified Cas variants and delivery approaches have optimized its application to treat blindness. In this review, recent progress and challenges of the CRISPR-Cas system have been discussed to cure blindness and its prospects.
The eye is an important mammalian organ with a complex anatomical structure that is produced and regulated by different types of multiple ocular tissues. The disruption of any ocular tissue can cause visual impairment and vision loss, which can significantly influence the quality of life in mammals, including human beings. Damaging of the retina and optic nerve are the major causes of vision loss due to aging. Recent data from WHO shows that 1.3 billion people are affected with vision impairment among which the majority of people are above the age of 50 years. Age-related muscular degeneration (AMD), cataract, dry eye, and glaucoma are the common types of ocular diseases in the world [1]. During the past two decades, a lot of research work has been done to discover different molecular therapeutic tools to treat ocular diseases. These strategies include gene therapy through viral delivery methods, implants, hydrogels, liposomes, and nanoparticles-based delivery to repair ocular diseases. However, the mentioned strategies have major drawbacks, such as immune rejections, and biocompatibility issues of hydrogels, implants, and nanoparticles. The recently developed gene-editing technology, clustered regularly interspaced short palindromic repeats (CRISPR/Cas9) is revolutionizing every field of biological science research as well as ophthalmology. The CRISPR/Cas9 is a precise gene-editing technology that has emerged as a novel therapeutic tool to treat ocular diseases by restoring the vision in human beings. CRISPR/Cas9 was discovered in 1987 and was reported for the first time in 2013 for editing the mammalian genome. Consequently, it has been used as a therapeutic agent to treat genetic diseases such as retinal dystrophies, neurodegenerative diseases, etc. There are three classes of CRISPR-Cas system, I to III. Each class of Cas system has been divided into the CRISPR subtypes [2]. The mechanism is very simple, designing the guide RNA (gRNA) according to any target gene and then combining it with Cas9 to form the ribonucleoprotein (RNP) (Figure 1). The RNP can be directed by gRNA to cut the specific nucleotide sequence of DNA that triggers the DNA repair system, non-homologous end joining (NHEJ). The NHEJ DNA repair system is not efficient and could lead to small insertion or deletions (INDELS) that ultimately knock-out (KO) the gene of interest. It may cause the frameshift mutation in the particular gene aimed to be knocked out in the specific genome. Similarly, if the donor template has been provided with an insert of any foreign gene or to correct the single-nucleotide polymorphism (SNP), that leads to a DNA repair system e.g., homologs direct repair system. In this case, the provided donor single-stranded DNA can be used as a template DNA to repair the target DNA and ultimately knock-in (KI) the gene of interest. These strategies make the CRISPR/Cas9 system less costly as compared to conventional gene-editing tools [3]. Rapid advancement in the genome editing field has led to the development of precise base and prime editing technologies. Due to favorable anatomical eye structure and privileged immunogenic characteristics, inherited retinal diseases are at the forefront of CRISPR/Cas-based therapies. Genome-wide screening and CRISPR/Ca-based gene knockout techniques can be used to investigate eye disease mechanisms for the development of novel therapies. Leber congenital amaurosis type 10 is a genetic eye disease developing in children due to a mutation in the CE9290 gene that causes the loss of CEP290 protein and gradual loss of eyesight. Recently it has been fully cured by fixing the mutant gene through the CRISPR gene knock-in (KI) strategy. An ongoing CRISPR/Cas9-based clinical trial (NCT03872479) to cure LCA is in phase III and has been sponsored by Editas Medicine, Inc. In this research CRISPR/Cas9 has been used to correct the CEP290 gene and in near future, it will be approved to treat LCA [4]. Herpes Simplex virus (HSV) keratitis is a viral infection of the cornea that causes blindness. Worldwide, an estimated 491 million people aged 15–49 are affected by HSV. A CRISPR/Cas9-based clinical trial (NCT04560790) to treat HSV keratitis diseases is in phase II and has been sponsored by Shanghai BDgene Co., Ltd., Shanghai, China. In this study, CRISPR/Cas9 has been employed to disturb the UL8/UL29 genes to inhibit the growth of HSV [5,6]. This article reviews and highlights the CRISPR/Cas9 as an emerging therapeutic tool for the treatment of genetic eye diseases. Its potential applications have been elaborated to cure blindness in human beings and other mammalian species.
The current development in gene therapy research has played an important role in understanding the genetic origin of structural and functional defects in common human eye diseases for instance age-related macular degeneration (AMD), glaucoma, cataract, myopia, Stargardt’s disease (SD), retinitis pigmentosa (RP), Marfan syndrome (MFS), polypoidal choroidal vasculopathy (PCV), uveal melanoma (UM), and optic atrophy. Table 1 shows the gene variants associated with the mentioned eye diseases. AMD commonly occurs in old people that causes the loss of central vision as a result of macular degeneration which is the vital area of the retina [7,8,9]. Multiple SNPs occur in many genes such as NOS2A which increases the likelihood of AMD, especially SNPrs8072199 [10]. Recently 3 SNPs in the MMP-9 locus (rs4810482, rs17576, and rs17577) have been reported that are highly associated with an increased risk of AMD [11]. Moreover, TIMP-3 risk alleles and mutations are directly linked with retinopathies such as AMD and SFD [12,13]. Another study has reported the association of HTRA1 and ERCC6 with AMD [14]. Researchers have reported that CALM2 might be the cause of glaucoma since it takes part in the death of retinal ganglion cells (RGC) that disturbs the cell communication system. Moreover, the MPP7 was found to be involved in primary open angle glaucoma (PAOG) through its dysfunction [15]. Additionally, multiple defective genetic regions have been identified that are related to glaucoma such as CYP1B1, and optineurin. LOX1 was also recognized to be one of the causes of exfoliation type glaucoma. Early-onset glaucoma can be due to mutations in PAX6, FOXC1, MYOC, LTBP2, and PITX2. Finally, gene testing can help in recognizing people who can participate in clinical trials for the treatment of glaucoma [16,17]. It was discovered that inherited cataract is affiliated to more than 50 loci however the genetics of many cataracts are still unidentified. Multi-gene panel have been used to study these mutations such as GEMIN4 mutation in cataract individuals. A parent mutation has been discovered in RIC1 that exists in patients with cataract and by sequencing of the cataract gene it was exposed as a truncating mutation in TAPT1 [18]. Recently, it was reported that more than 10 mutations can occur in MIP which is associated with cataracts [19,20,21]. Myopia is a short-sightedness where light is concentrated in front of the retina rather than on the retina. It was identified that HGF and C-MET are associated with myopia in Asians but mutations in the C-MET receptor in Asians were not reproduced in Caucasians [22]. Further studies showed that mutation in UMODL1 may have a role that contributes to the susceptibility of myopia. Moreover, it was discovered that a functional SNP at 3′UTR of PAX6 can contribute to the increased risk of myopia [23]. Stargardt’s disease (SD), is shortsighted due to the fatty material accumulation on the macula of the eye. There are more than 800 mutations identified in ABCA4 that are associated with SD [24]. In one report two distinct phenotypes (LCA and SD) have been identified to cause severe retinal degeneration. The cases with LCA and SD were due to mutation in CRB1 and ABCA4, respectively [25]. Retinitis pigmentosa (RP) is, a disorder that affects the light-sensitive layer of retina, and the patients exhibit gradual loss of sight. There are more than 250 mutations identified for 100 genes but only 50% of the defected have been identified [26,27,28,29]. A higher proportion of RP2-mediated XLRP has been reported in Danish people [30]. After extensive studies, it was reported that two rare and deleterious variants p. Arg281Cys and p. Arg487* were discovered in AGBL5 [31]. Marfan syndrome (MFS), caused by genetic mutations in the connective tissues such as mutations in FBN1 on chromosome 15 have been discovered in patients which can lead to many abnormalities that includes increased TGF-β1 causing irregularity in signaling [32,33,34,35]. SNPs that can change the TGFBR2 are connected to MFS [36]. Polypoidal choroidal vasculopathies (PVC) can be genetically separated into polypoidal CNV and typical PCV. The gene variants C2 and CFB have been revealed to be associated with polypoidal CNV [37]. Other studies have reported the association of C3, SERPING1 and PEDF with PCV [38,39]. Uveal melanoma (UM), caused due to BAP1 suppressor mutations that are linked to the development of different tumors in pigment cells [40]. Hence, the roles of CDKN2A/P16INK4A, CDK4, and P14ARF germline mutations were tested. In addition, the BRCA1/2 germline mutations and the patient history were also evaluated in the monitoring of UM [41,42]. Inherited optic neuropathies is the most typical inherited form of vision loss due to an injury in the mitochondria affecting young males. The greater number of cases with this disorder is because of a three-point mutation in mtDNA affecting complex I or ND genes (G11778A, G3460A and T14484C), while dominant optic neuropathy (DOA) is a non-sex chromosomal hereditary optic nerve neuropathy that is the most widely known inherited neuropathy. Despite the fact that it is heterogenous, a main locus has been mapped to chromosome 3q28 with a mutation in OPA1 [43,44]. Gene-editing methods such as RNAi and CRISPR/Cas9 technology can be used to correct the mentioned gene variants associated with human eye diseases.
There are many ophthalmological processes such as replacing the lens after a cataract and fluid leakage control from retinal vessels which are already considered the principle guideline in clinical care. Lately, gene therapy was shown to be effective in enhancing visual performance for individuals that suffer from LCA and other hereditary disorders [45]. Although there is huge progress in this field, from the patient’s perspective, this progress is very slow. This is due to adverse challenges that exist in the retina such as how complex it is and that there is no regeneration potential. Therefore, therapeutic procedures that are being used until now are very limited to and targeted only to impede the degeneration of vision. There are many reports each year about new therapeutic options and different procedures to treat IRDs that comprise stalling cell degeneration such as neuroprotection, spurring the remaining retina-like retinal prostheses, and supplying the cells with a duplicate that has the same function as the gene of interest, e.g., gene therapy. Retinal prostheses are a type of bionic eye or implant electronic device that stimulate the sensation of vision. These implants have been used to treat retinitis pigmentosa. The stimulation of visual sensation can be achieved in two ways, either directly to ganglion cells or by using electric impulses that excite the internal retina, which is directed to the ganglion cells, requiring a thorough analysis of the retinal reorganization as it can vary from one patient to another. Normally, many factors constitute the retinal prostheses such as a small camera, an image processing unit, a stimulator chip, and an intraocular electrode array. The camera takes a photo of the field of sight and transmits it to the image processing unit, where the image is translated into the right stimulating pattern that is sent to the stimulator chip which delivers the stimulation to the intraocular electrode array [46]. Recently, FDA approved the second generation of external hardware (glasses and processing unit; Argus2s) to practice with Argus II implants though Second Sight has suspended Argus II implants since 2019 in favor of visual cortical prosthesis systems with the electrode neural interface moved from the retina to the visual cortex [47,48]. Prosthetic devices can be used to target other parts of the visual pathway; for instance, the optic nerve, lateral geniculate nucleus, and visual cortex have been developed that may be of specific assistance to patients with severe retinal damage. Examples of the ongoing studies with cortical prostheses are Orion Cortical Visual Prosthesis System (Second Sight; NCT03344848) and Intracortical Visual Prosthesis (NCT0463438) [49,50]. Choosing a therapeutic procedure for implementation relies on the stage of disease, however none of the therapeutic strategies that are being used have ever reported complete restoration of normal vision that was already lost [51]. For example, when retinal degeneration reaches the late stage, the retina will contain only a few photoreceptor cells which prevent the use of treatment options that depends on the existence of healthy cells, so when the degeneration reaches an advanced stage, the only proper therapeutic options are cell transplantation, retinal prostheses, and optogenetics [52]. Neuroprotective therapeutic strategies have been used to enhance the survivability of neuron cells by inhibiting apoptosis directly or by strengthening the means that enhance cell survival [53]. The basic issues were associated with the low ability to copy the results of the clinical trials which are usually done on animal models that have distinct compositions than humans [54]. Optogenetics is a method used to insert light-sensitive proteins into cells that facilitate neural activity control and can be monitored by using light. These light-sensitive proteins encoded by opsin genes, especially microbial opsins can serve as light-responsive ion pumps or sensory receptors [49,55]. This method can be used in the recovery of sight which is due to the degeneration of photoreceptor cells, however the leftover cells from the bipolar and ganglion retinal cells that existed in the retina can be targeted by optogenetics [51]. Moreover, optogenetics obligates the use of goggles during treatment since it is used for projecting the visual field onto the retina in intensity and spectrum that coincides with the opsins used; however, opsins are limited to a narrow range of light intensities [56]. The stimulation of retinal cells in invertebrates can occur by injuries and it is possible to regenerate the photoreceptor cells. Typically, adult mammals have stem cells that are present in the retina but they cannot spontaneously re-enter the cell cycle [57]. In addition, injuries can cause cell degeneration instead of cell stimulation. So, the transplantation of retinal progenitor cells or stem cells to the damaged retina has a large prospect for the treatment of retinal diseases [58]. Finally, stem cell transplantation is an interesting prospect for treating late-stage retinal degeneration but the level of organization of the retina cells must be analyzed carefully before the commencement of treatment since a high level of organization can hinder the development of new cells that can form networks inside the retina [59]. However, there are still many unknown factors in this strategy such as the way to increase cell survival, integration, and differentiation in a patient’s retina [60]. The classical therapeutic approaches to restore vision are time consuming, laborious, less efficient, and costly. These limitations can be overcome by an efficient and cost-effective RNA guided gene-editing technology (CRISPR/Cas9) that is currently in progress to overcome the limitations of conventional recombination methods [3].
In 1987, scientists have observed a rare structure comprising 29 vastly conserved nucleotides located in the 3′-end adjoining region of the iap gene in E. coli. Afterward analogous repeats have been found in many strains of bacteria and archaea by random sequencing of their complete genomes [61]. These sequences of a clustered repeat were termed short regular spaced repeats (SRSRs) that are frequently spaced via unique intervening sequences of constant length and have been classified as an exclusive gene family existing in the immune systems of prokaryotes providing them an acquired resistance against viruses [62]. The existence of these inward short, inverted repeats in the repetitive units shows the same physiognomies of known sites intended for specific DNA-binding proteins. Many queries have been raised from these findings, such as whether their existence may be due to some of the earliest sequences or have been diverged through the process of evolution. At the beginning of the 20th century, this innovative repetitive family ‘CRISPR’ was discovered via in silico analysis by revealing many genes that existed in line with the cluster repeats that were denoted as CRISPR-linked genes or Cas genes. Later on, these gene families of CRISPR were recognized as stably maintained secondary structures of RNA demonstrating efficient preservation through the process of evolution [63]. These structures are considered to perform significant functions in the prokaryotic gene repair and regulation to instigate their defense against different viruses. Subsequently, tremendous applications of these structures have been studied to edit animal and human genes using stem cells in order to generate restrictive mutations to find out the inside working environment of the cell. Certainly, the applications of the CRISPR-Cas9 system seem to be beyond our imagination. The CRISPR-Cas9 system was first investigated in prokaryotes which are using it as an adaptive immune system to counter phage viruses. The system comprises two important constituents, e.g., an endonuclease (Cas9) and a sole guide RNA (sgRNA) used to detect a specific sequence in the genome. The Cas9 act as an endonuclease by cleaving a specific sequence in the genome containing two catalytical positions (RuvC and HNH), nuclease domains that produce DSBs by cleaving the reverse strands of DNA [51]. SgRNA contains two components to enhance its action that comprises CRISPR RNA (crRNA) and an intergenic trans-activating crRNA (tracrRNA) used as a guided tool required for the activity of Cas9 activity. The crRNA contains a sequence ~20 bp that binds to its target sequence following the complementary base pairing rules of Watson–Crick model. The crRNA recognizes the target based on a DNA sequence that is short and conserved called protospacer-adjacent motif (PAM) which is positioned adjacent to the target DNA. SgRNA can be simply designed and cast off rapidly. It unites the Cas9 as a ribonucleoprotein (RNP) complex in the CRISPR structure and SgRNA directs the Cas9 to the targeted position in the anticipated gene and at that point, Cas9 exactly cut the DNA at 2–3 nucleotides earlier than the PAM sequence to generate DSBs. Generally, HDR or NHEJ are the two key mechanisms in DNA repair therefore DSBs can be repaired by any of these mechanisms. Similarly, the sgRNA may be restructured by modest variations in the DNA sequence to retarget novel DNA sequences. Furthermore, the CRISPR-Cas9 system can be used to change numerous genes through a solitary nuclease employing diverse sgRNAs simultaneously (Figure 1). An alternative form of Cas 9 is Cas9 nickase (nCas9) which varies in a single point mutation at position D10A or H840A as of Cas9 and cuts single-stranded DNA sequences targeted by sgRNA. Likewise, the double mutation in H840A and D10A of the HNH and RuvC domains of Cas9 may disable the endonuclease action of Cas9 and can reinforce dead Cas9 (dCas9) [42] which are unable to cut the DNA positions in order to generate DSBs; Therefore, it may be attached to transcriptional activators to make a CRISPR activation (CRISPRa) system capable of altering the gene expression by rewriting the epigenetic marks on the target gene; moreover, It may be attached with transcriptional inhibitors to develop a CRISPR inhibition (CRISPRi) system used to overwhelm the process of transcription [64].
Different types of injection routes have been used to deliver to ocular cells such as topical administration, intracameral injection, and subconjunctival, that are commonly used for gene transfer, and that can be used for the transfer of CRISPR/Cas9 components. The plasmids DNA gene has been delivered for transgene expression through topical administration. The intracameral injection is usually used to transfer the genes to the interior part of ocular cells such as the Conroy endothelial, and ciliary body. The intravitreal rout injections are used to transfer the genes to the inner retina cells, especially into the ganglion cells. The subretinal injections have been used successfully to transfer transgenes to the outer parts of the retina. Multi variants types of vectors have been designed and used to transfer the Cas9 and RNA components to targeted ocular cells. The Cas9 as of Streptococcus pyogenes (SpCas9) is considered as the most studied ortholog however, due to its greater size (4.2 kb), confining its capacity for in vivo use to be packed with gRNAs into a single adeno-associated virus (AAV) vector. To overcome this constraint, packaging into lentiviral vectors can be employed for twin AAV vectors to express the whole SpCas9 distinctly as of gRNAs, or to use the split-Cas9 by separating the SpCas9 expression on its messy element V713–D718 and remaking the entire protein via trans-splicing of split-intein protein [65]. Furthermore, SaCas9 of Staphylococcus aureus and CjCas9 of Campylobacter jejuni contain minor Cas9 orthologs that can be combined and accompanied by gRNAs simultaneously in the AAV vector [66]. The first time the CRISPR technology was used in the human eye was in the year 2019. The clinical test has assessed the subretinal AAV-facilitated transport of SaCas9 combined with a couple of gRNAs to mark a profound intronic mutation in the CEP290 gene in order to treat type 10 Leber congenital amaurosis (NCT03872479) [67]. Remarkably, even with the theoretical gain to use solitary viral vector for clinical translation, different researchers have observed that genome editing efficacy of the smaller Cas9 alternatives is mediocre with regards to the dual-vector delivery of SpCas9 and gRNAs [66]. Even though viral vectors can be efficiently used as genome editing tools in the retinal cells, the constant viral expression of Cas9 endonuclease may cause off-target effects. In contrast to conservative gene expression or biofactory approaches, the CRISPR systems no longer entail enduring transgene expression, wherever the continued existence of Cas9 may activate indifferent mutations. A different approach intended for clinical use is the straight delivery of recombinant Cas9 proteins along with gRNAs as ribonucleoprotein complexes (RNPs) in the eye that can instantly influence the cleavage of DNA to be quickly degraded in cells, thereby minimizing the off-target effects as well as cytotoxicity [68]. The straight use of RNPs in human cells proved to be an effective gene cleavage method compared to the transfection of plasmid by means of ~79% on-target mutation with less off-target effects and if carried out in the subretinal region for targeting VEGF in the mouse eyes, it resulted in ~40% decline in a CNV laser-induced model. Regardless of the RNP advantages, the Cas9 protein delivery into the nucleus is still a difficult task, mostly attributable to endosomal entrapment in the cytoplasmic matrix. In this regard, the cell-penetrating peptides (CPPs) can enhance the transport efficacy by ~80% [69]. In summary, due to size limitation capacity, mostly the adenosine-associated viral vector (AAV) has been widely used to transfer Cas9 and gRNA cargo into ocular cells. The AVV has a better balance of efficiency and stability. The non-viral vectors can be used in some circumstances, such as off-target effects and immunogenic reactions. With the rapid development of ocular gene delivery vectors, the advanced tailer-designed vectors will enable more efficient and effective CRISPR/Cas9 delivery.
AMD is a multi-genetic ailment and the key source of blindness worldwide. A neovascular form of AMD is the wet AMD caused via irregular growth of choroidal vessels in the macula area of the retina ensuing damage to central vision. The macula is controlling the color vision and bright light actions due to their abundant cone photoreceptors [45]. The excess proliferation of vascular endothelial growth factor (VEGF) causes neovascularization in AMD, therefore anti-VEGF mediators turn out to be an excellent therapy [46]. Recently such types of mediators (aflibercept, bevacizumab, and ranibizumab) can be given as an intravitreal injection to treat wet AMD patients [47]. AMD patients can be treated with the AAV-CRISPR tool designed based on CjCas9 (Campylobacter jejuni) [38,70,71] as well as type-V CRISPR-Cas systems, LbCpcf1 nucleases. The CjCas9 gene with its matching sgRNA sequence and marker gene has been packaged into an AAV vector. CjCas9 delivered through AAV can specifically cut a restricted number of sites in the human or mouse genome that can cause desired mutations in the RPE cells. In this regard, the Vegfa or Hif1a gene in RPE cells can be targeted by CjCas9 to reduce the size of laser-induced choroidal neovascularization and this approach can be developed into in vivo genome editing therapy for AMD. After six weeks an AAV-CjCas9 injection intravitreally has resulted from an Indel efficiency of 22 ± 3% in Vegfa and 31 ± 2% in Hifla genes, respectively. Additionally, the outcomes of the Indels have been observed at the protein level e.g., substantial decline in VEGF-A protein was detected in RPE cells as compared to the control set. Another research group used lipofectamine 2000 for the delivery of sgRNA/Cas9 expressing plasmid with Cas9 RNPs which showed 82 ± 5% and 57 ± 3% of indel in NIH3T3 and ARPE-19 cells, respectively. In this study, Cas9 RNPs have been observed to be highly active w.r.t plasmid on the second day of transfection. After treatment with Cas9 RNPs, a reduction in the VEGF A mRNA and protein levels of 40 ± 8% and 52 ± 9% has been observed in the adult retinal pigment epithelial cells (ARPE). To evaluate the in vivo efficacy, Cy3-characterized RNPs have been delivered through intravitreal injection. After three days of injection, Cy3 dye was accumulated into RPE cells with detection of 25 ± 3% of indel at the delivery site in RPE cells. Furthermore, to mimic wAMD a CNV mouse model has been developed using laser tracked by an injection of Cas9 RNPs in the subretinal region. Three days later 22 ± 5% indel was detected in RPE cells for the VEGFA gene. The treatment with Cas9 RNPs has tremendously condensed the CNV area by 8 ± 4% along with a reduced level of VEGF A protein (Figure 2) [72]. Recently a lentiviral system has been reported to deliver mRNA encoding an extended Cas9 protein (SpCas9) and gRNA instantaneously named mLP-CRISPR. The mLP specifies mRNA-carrying lentiviral particles that are used to prevent the progress of wAMD in a laser-persuaded CNV mouse model. Subretinal injection of mLP-CRISPR displays an enhanced tissue specificity in the retinal pigment epithelium (RPE) cells that are considered to be the main cause of VEGFA in the outer portion of adult eyes. Additionally, mLP-CRISPR did not induce anti-Cas9 IgG in blood or T-cell permeation in the eyes. An injection of mLP-CRISPR has disrupted 44% of Vegfa genes in RPE while decreasing 63% of the laser-induced CNV area in the wAMD mouse model. This has been done by using a gene-editing tool without any off-target effects. This type of mLP technique can be used to carry mRNA encoding numerous Cas9 nucleases, base, prime, and epigenome editors [73].
Glaucoma is an assemblage of eye diseases caused due to advanced and permanent disintegration of the ganglion cells in the retina which axonal projections establish the optic nerve [74]. At present, it is considered the foremost source of irreversible blindness globally [71] that could affect >76 million people by the year 2020. Presently, glaucoma is mainly treated in clinics by reducing the intraocular pressure (IOP) in affected individuals by medicine, laser action, or surgical treatment. However, surgical and laser therapy brings risks that frequently need additional interference or combinational tactics supplemented with more topical therapies during a patient’s lifetime [47]. The patients frequently require consistent treatment through numerous types of eye drops that can be used many times daily. It means that patient treatment becomes a challenge with the passage of time and even numerous of them have experienced continuous visual loss instead of their reduced IOP [48,49]. Consequently, novel therapeutic methods can be developed that can be obtainable by an injection into the eye directly with enduring or perpetual beneficial consequences. A considerable number of adult individuals with glaucoma have an indistinct, varied cause of disease linking numerous genetic, environmental, and individual risk factors [38,75]. Due to these reasons, the commencement of glaucoma in adult individuals can be controlled by gene therapies that are mainly focused on neuroprotection, which encompasses the reduced loss of RGCs by changing their physiological status in order to reduce the severity of the disease. This can be achieved by either enhancing the action of innate survival pathways in RGC or inhibiting the development of programmed cell death. With the recent developments in specificity and efficacy of CRISPR-mediated genome editing technology, there is a hope to target these mutations. Currently, researchers have targeted the dominant MYOC mutations in a mouse model of myocilin-linked POAG through adenoviruses that are expressing the CRISPR/Cas9 components (Ad5-cas9 and Ad5-crMYOC). In this study, Cas9 has successfully knock-out the mutant MYOC gene, reduced IOP in the eyes of treated mice, and stopped further glaucomatous damage in mouse eyes. Similarly, the same constructs have been used for the treatment of trabecular meshwork tissue in the human eyes cultured ex vivo. A decrease in the myocilin mRNA suggests the possibility of using this technology to overcome the problems of patients with MYOC mutations [76]. Primary open-angle glaucoma (POAG) has been considered the main global source of blindness, comprising an imperative risk factor of elevated intraocular pressure (IOP). The pathological variations in the trabecular meshwork (TM) increase POAG IOP linked with changes caused by an enhanced level of TGFβ2. Recently, CRISPR interference has been used to explicitly deacetylate histones in order to decline TGFβ2 in the TM. The CRISPR interference system has been observed to constrain TGFβ2 expression in human TM cells through accurately designing sgRNA that have targeted the TGFβ2 gene promoter. The sgRNA has targeted the CMV promoter of the Ad5-CMV-TGFβ2 viral vector and it has been observed that lentivirus-mediated KRAB-dCAS9 and sgRNA expression were capable to constrain Ad5-CMV-TGFβ2-induced OHT in the eyes of C57BL/6J mice (male and female). The reduction in OHT was linked with a diminished level of TGFβ2 and extracellular matrix proteins in the mouse eye. These findings suggest that CRISPR interference can be used as a tool for gene inhibition containing the therapeutic potential for the treatment of TGFβ2-induced OHT [77]. Currently, a practical gene therapy tool has been used to reduce IOP via specific disruption of the aqueous humor synthesis in the ciliary body ensuing an intravitreal injection. In this regard, explicit gene editing was done by means of the smaller S. aureus-derived CRISPR-Cas9 that is proficient to be carried in a single recombinant AAV vector which is considered a benchmark by US-FDA and has been approved for visual gene therapy. In order to target a gene vital for a well-maintained functional development relatively to correct a single explicit mutation, a collective approach deprived of excessive personalized tactic is manageable. IOP decrease can be accomplished through knock outing Aquaporin 1 (Aqp1) in the ciliary body. Aquaporins are a group of water-transferring transmembrane proteins that are broadly expressed all over the human body and transgenic mice lacking Aqp1 have shown lesser IOP due to abridged influx to form aqueous humor deprived of contrary consequences [78].
RP is considered an important source of progressive blindness affecting 1 in 4000 individuals [50]. The rod-cone dystrophy is a usual RP that is recognized via shaft vision ultimately leading to the loss of peripheral vision. Its early symptoms contain nyctalopia which leads to night blindness and patients have complications in adjusting to the dark that happened to the injury of rod function in early childhood [51]. Due to the damaged photoreceptors, RPE begins to lose its pigment eventually leas to the buildup of intraretinal melanin deposits, which appeared as a bone spicule conformation. Though, the central vision is still integral till the last stages. Different genetic modes are responsible for its transmission, such as autosomal dominant and recessive or X-linked and are heterogeneously linked with mutations as a minimum of 79 genes [79]. There are primarily two types of RP which are MERTK connected and RPGR X-linked. The apical membrane of RPE comprises photoreceptors that are sensitive to light and mediated by MERTK (Mer tyrosine kinase) receptors that are engaged in the rods and cones phagocytosis. To play an effective role, the continuous recycling of these photoreceptors is very important that is interrupted by mutations in MERTK, which proceed to degradation and ultimately decline of the photoreceptors [41]. It has been observed in meta-analyses that ~3% of MERTK type RP are a result of autosomal recessive transmission which are causing macular atrophy and childhood photoreceptor abnormality [43,44]. Alterations in the RP GTPase regulator (RPGR) that is an X-linked RP (XLRP) has been observed in 1 out of 3500 individuals. RPGR, as well as the δ subunit of rod cGMP phosphodiesterase, controls the proteins and its disorder is causing continuous damage to the central vision and progressing to night blindness [51]. CRISPR/Cas9 technology has been used in some in vitro and in vivo studies to treat RP e.g., applied in a rat model of adRP to remove mutation in the rhodopsin gene (RhoS334). In this experiment the sgRNA/Cas9 plasmid has been used to target exon 1 closely upstream of a PAM exclusive toward the RhoS334 locus was directed intravitreally in S334ter-3 rats. In two different rats, a cleavage efficacy of 33 and 36% was confirmed in transfected retinal cells through genome analysis. Subsequent injection of sgRNA/Cas9 plasmid has confirmed an enhanced visual insight and widespread protection of the retina through immunohistological examination [80]. Moreover, the same strategy was used to edit the RHO gene mutations. In this research, an intended plasmid containing an insert for 2 sgRNAs targeting the RHO gene was used to generate DSB and subsequently NHEJ. As a result of this research, the RHO gene was successfully edited with additional downregulation of the expressing RHO protein. In another study, CRISPR/Cas9 system was used in iPSCs attained from a patient with photoreceptor degeneration to treat XLRP to correct RPGR (a pathogenic point mutation). In this research, 21 different sgRNAs have been screened for editing, and among them, g58 was observed as utmost activity. Thus, the plasmid containing g58/Cas9 was considered to transfect the iPSCs together with RPGR single-stranded oligo deoxy ribonucleotide (ssODN) that plays a supportive role in the HDR pathway. After deep sequencing analysis, the data showed an effective modification of mutation in 13% of the transfected cells [81]. Additionally, the premature stop codon TAG was successfully substituted by the wild-type codon GAG encoding glutamate at position 1024. In contrast, the untransfected iPSCs did not show any variations in the mentioned mutation. It was determined that the 13% correction rate was meaningfully productive that can be further enhanced by reducing error-prone NHEJ through DNA ligase IV inhibition on the DNA cleavage site. Moreover, a CRISPR/Cas-based approach was established to edit RHO gene mutations in the ADRP mouse model using a plasmid comprising an insert designed for two sgRNAs to target the RHO gene (exon 1) containing P23H dominant mutation. Primarily, HeLa cells have been used for gene editing in vitro showing an indel frequency of 70%, 76%, and 82% by means of sgRNA1, sgRNA3, and 2sgRNA, respectively. Further, the Real-time Taqman PCR has been used to observe RHO expression where 35%, 25%, and 20% of reduced expression level has been detected in the cells treated with sgRNA1, sgRNA3, and 2sgRNA, respectively. Later on, the CRISPR/Cas plasmid has been used by electroporation in P23 RHO transgenic mice comprising 2sgRNA together with green fluorescence protein (GFP) to achieve subretinal expression. To evaluate Cas9 expression, the GFP expressing section of the retina has been isolated where the expression was inadequate in the cells articulating GFP together with 84 edited sequences [81]. The scope of treatment is the inadequate beginning embryonic day (E) E16 to P2 due to the rate of retinal deterioration in adRP patients that is heterogeneous. Therefore, an ablation therapy for long duration could be useful to maintain vision in animal models. Thereby, a slower degenerating adRP model having a common adRP mutation (rhodopsin P23H) has been used for clinical application. In adRP patients, the P23H (RHOP23H) is a very common mutation that can be observed in the rod cell-specific gene rhodopsin (RHO) which is an effective mark to study mutation-specific ablation approaches via CRISPR [10,13]. Usually, an excessive level of rhodopsin misfolding (class II mutation) and mistrafficking (class I mutation) as a result of P23H mutation is accumulated in the endoplasmic reticulum (ER) [14,15,16]. The P23H line-3 rats experience photoreceptor degeneration from P15 which shows the level of vision retained with the passage of time equal to decades in patients accepted for treatment at various disease stages. In this research, the subretinal delivery of AAV-CRISPR/Cas9 has been shown as a benign approach that can cure photoreceptors and vision intended for 15 months in Rhodopsin P23H-3 rats (Figure 3) [82].
LCA is considered a high overwhelming retinal dystrophy due to its potential of causing congenital blindness in <1 year of age. Until now 14 mutant genes have been known based on linkage examination, homozygosity mapping, and genome scrutiny in LCA patients among which 70% are children with retinal degeneration [40]. LCA patients are generally connected with plain imperfections comprising rambling eye movements termed nystagmus. In children, the symptoms are usually sluggish pupil reactions and deficiency of electroretinographic reactions [41,42]. Genes intricate LCA encode the proteins that are accountable for retinal roles comprising vitamin A cycling (LRAT, RPE65, RDH12), photoreceptor morphogenesis (CRB1, CRX), guanine synthesis (IMPDH1), phototransduction (AIPL1, GUCY2D), outer segment phagocytosis (MERTK), and intra-photoreceptor ciliary transport progressions (CEP290, RPGRIP1, LCA5, TULP1) [40]. Until now, the investigated genes for LCA stand mutations in the RPE (RPE65) gene that is encoding retinoid isomerase [30], although, the maximum going on mutations are connected with the CEP290 (15%), GUCY2D (12%), and CRB1 (10%), respectively. About 20% of north-western Europe patients have an intronic CEP290 mutation (p.Cys998X). In the mice, an AVV-CRISPR system has been studied to investigate the in vivo action of autosomal dominant retinitis pigmentosa (adRP) and LCA10. In this research, the AAV-SpCas9 vector was carried through subretinal injections targeting the rhodopsin (RHO) or CEP290 and Nrl (neural retina leucine zipper transcription factor) gene using a mouse model for adRP. The results are very promising which show an enhanced level of spCas9 protein expression in the retinal cells of mice for a period of nine-and-a-half months. Additionally, the researchers have arrayed diverse AAV serotypes besides changed vector doses and have found a real repair of RP or LCA10 phenotype lacking off-target properties or adverse toxic responses [43,44]. The same approach has been used successfully in human cells to resolve the RHO gene mutation. This study confirms CRISPR/Cas9 as an effective system to target gene/alleles in a well-organized way indicating that it may be cast off for the treatment of RP and further dominant human genetic disorders [83].
Usher syndrome is considered a frequent form of syndromic IRD with a prevalence of 1 out of 20,000 individuals with its exclusive characteristics comprising RP and hearing damage [70]. Based on the development and intensity of the hearing injury along with the RP outset, this heterogenous syndrome is segregated into three subtypes which are usher syndrome type 1 (USH1) utmost critical, USH2 frequently observed with modest to severe symptoms, and USH3 considered as a moderate phenotype that varies case by case regarding disease outset and its progression [71]. USH1 is considered to be the main source of deaf-blindness that is inherited in an autosomal recessive manner in humans causing vestibular ailment, deep congenital heart loss, and RP. USH1 is triggered as a result of mutations in myosin VIIA encoding an organelle transport protein in the RPE [46]. CRISPR/Cas9 gene editing was proposed as a gene-editing tool to reinstate the c.2299delG mutation in the USH2A gene. The gene-editing experiment was carried out in human dermal fibroblasts (HDFs) cells taken from a USH2 patient having c.2299delG mutation. By means of nucleofection, a Cas9 RNPs containing 15 µg Cas9 and 20 µg sgRNA has been transfected into the normal individual HDFs producing 18% indel rates. Afterward, delivery of RNPs combined with ssODN-2299 has produced 5% HDR efficacy. Correspondingly, patient HDFs with c.2299delG mutation have been transfected with ssODN having a WT sequence with a deleted PAM sequence resulting in 6% indel efficacy with 2.5% deletion of HDR [72]. Recently a pig model of USH imitates the mix of deafness, vestibular dysfunction, and vision loss found in USH individuals. Visual impairment has been observed in USH1C pigs through ophthalmologic checkups, studying their behavioral examinations go together with morphologic changes in the photoreceptor cells. Additionally, the photoreceptor and primary cilia of fibroblasts have been constantly extended in USH1C pigs which could help to conduct gene-editing experiments as potential therapeutics. Investigative gene therapeutic tools can verify the improvement of visual function in USH1C pigs (Figure 4) [84].
In the last decade, a lot of safety concerns have been reported in gene-editing studies regarding off-target effects due to double-strand break (DSB) in CRISPR/Cas9 experiments. The DSB generates DNA lesions which leads to deleterious chromosomal rearrangements, duplication, inversion, deletions, and translocation that generates instability in the edited genomes [85]. However, the Cas9-nickases as Cas9 variant that generates the single strand break was the first preference of researchers for the integration of foreign DNA pieces through HDR. Before the development of CRISPR/Cas9-based gene-editing genetic therapies, there would be deep knowledge of DNA repairs system such as NHEJ. The NHEJ mostly leads to off-target effects, which is why further development of highly sensitive unbiased gene-editing techniques is a dire need of the current research. As an alternative, CRISPR/Cas9, which should edit the genes without the DSB and NHEJ, was investigated. To reduce the off-target effects, recently researchers have developed new precise gene-editing technologies, such as base and prime editing. The working principle of the base editor consists of fusing the Cas9 nickase with the cytosine or adenine deaminase which converts the base pair such as C to A or A to G without generating the double-strand break in the DNA strands. The first base editor was developed by combining the Cas9-nickase with the cytosine deaminase enzyme, APOBEC1 which converts the C to T. Through a direct evolution process the RNA adenosine deaminase was generated and the second base editor was developed by combining the Cas9 nickase with RNA adenosine that converts the A to G. Recently a dual base editor has been generated which can covert adenine and cytosine basis simultaneously on DNA strands. As compared to CRISPR/Cas9 tool, which only work in dividing cells, conversely the base editor also works in non-dividing cells. The base editors only cause nucleotide conversion at a specific point rather than DSB by Cas9, and there is a negligible chance of insertions, deletions, and transversion chromosomal rearrangement in base-editing experiments. Recently a new technology, prime editing has been developed by combining the Cas9-nickase and engineered reverse transcriptase. The Cas-nickase is guided through prime editing guide RNA (pegRNA) consisting of 3′ end-extended gRNA. The 3′end extended portion of pegRNA act as a primer for the functioning of the reverse transcriptase (RT) initiation. After finding the complementary sequences the RT activated that synthesized the complementary DNA (cDNA), which is incorporated at target sites after a series of flap removal events. Prime editing is a versatile technique that can introduce small deletion and insertion efficiently. The highest efficiencies with no off-target effects were observed in editing the HK293T cells. In the last two years, base and prime editing have been employed for eye diseases [86,87,88]. Lebar congenital amaurosis (LCA) disease characterized by the loss of function mutation in the Rpe65 gene, which is the crucial factor in LCA emergence. This non-sense mutation was corrected through the base editors efficiently with zero off-target effects in mice that helped in the regeneration of visual chromophores which profoundly enhanced and restored the visual function in mice. The adenine and cytosine base editors were delivered and transduced efficiently into the retinal cells. However, small insertion and deletions were observed in the mice [89]. The prime editing has successfully edited the Dnmt1 gene in mouse retina with fewer off-target effects. Recently, six autosomal inherited genes associated with inherited retinal diseases (IRD) have been reported [90]. Primer editing has rectified 89% of genetic mutations that could be employed to modulate six pathogenic variants of IRD. As the new enzymatic modifications are coming day by day to reduce off-target effects, the base and prime editors would be at the forefront of the development of regenerative therapies.
Currently CRISPR/Cas9 has shown better potential to cure blindness. However, it poses inherent issues of controllability and stability that needs to be resolved before its clinical application [91]. The technology specifically targets a genome with high editing efficiency, facilitating quick and easy gene targeting and screening of recombinant viruses. The DNA repair induced by CRISPR-Cas9 seems to be quick, effective, and controllable, with recombination stability and fidelity in the case of multisite editing [92].
The delivery of CRISPR/Cas9 machinery to the target cells is influenced by safety and therapeutics efficacy. The viral and non-viral strategies have been used to deliver the CRISPR/Cas9 components to targeted cells. In viral delivery, mostly three types of vectors, (1) adenoviral, (2) lentiviral, and (3) adenosine associated viral (AAV), have been used to deliver the genes of Cas9 and gRNA to a specific target locus. The adenoviral method has a harmful potential to trigger an immune response in host cells [93]. The lentiviral and AAV methods causes immunotoxicity, oncogenes insertion, and random mutations. Multiple types of mutations have been generated using viral vector deliveries such as deletion of Rep protein that has increases the transducer efficiency, but still the immunogenic problem sustained [94]. The AAV delivery is limited due to packaging capacity that makes AAV vector difficult to use for transferring Cas9/gRNA with selectable markers and regulatory elements such as single strand homologues DNA for gene knock-in in the retinal cells. Newer approaches have been explored to reduce the immunotoxicity, oncogenes insertion and module the packaging capacity of AAV [95]. The limitations associated with viral delivery methods such as immune system trigger could be compensated through the use of non-viral delivery methods such as nanoparticles. Numerous nanoparticles-based delivery approaches have been employed and approved by FDA for treatment of eye diseases which are under clinical trials. These include liposomes (NCT00121407) to treat AMD, aptamer-polymer nanoparticles (NCT00549055) to treat the wAMD, and lipid-based nanoparticles (NCT03093701) to treat the molecular edema. The nanoparticles-based delivery methods can be modulated for targeted delivery, sustained release of cargo, and increased exposure time, but the application of nanoparticles delivery is limited by biocompatibility and low in vivo delivery [96]. In recent years, researchers have modulated the nanoparticles for stimulus-based delivery to targeted cells for CRISPR/Cas9 based gene-editing, that could be used in future to treat retinal diseases [97].
The specific or precision editing is also a limitation of CRISPR/Cas9. The HDR pathway which is highly crucial for gene insertion or knock-in has very low efficiency as compared to NHEJ for gene knock-out. Currently the efficiency of HDR has been increased through inhibition of NHEJ pathway by employing the modulators of NHEJ suppression enzymes such as Ku [98] and DNA ligases IV [99]. Other strategies to increase the HDR pathway efficiency by adding the single strand DNA homologous to targeted regions has increased the 60% chances of correct and precise/specific gene editing in zebrafish and mouse models [100]. It has been proved that homologous arm of 48 bp upstream and downstream can significantly improve the precise gene insertion in zebrafish genome through CRISPR/Cas9 gene editing [101]. Moreover, cell cycle stages can play a key role in determining the DNA repair system HDR or NHEJ. The HDR DNA repair pathway is restricted to S and G2 phases. Currently the S phase has been arrested for longer time than usual to increase the HDR efficiency through the addition of aphidicolin [102]. Advances in genome editing has allowed the use of base editors with specific gene editing capability to modulate the DNA repair system. In this new gene editing system, inactive dead Cas9 (dCas9) fused with an enzyme deaminase can catalyze and convert the nucleotides without the requirement of DNA repair system (HDR, NHEJ). Consequently, highly specific adenine base editors have been developed that convert adenosine to guanine and cytosine base editors which convert cytosine to thymine with precise gene editing [87]. Recently prime editing technology has been developed which can edit the genome in a precise way without the double strand break. Researchers have combined the Ca9 with reverse transcriptase enzyme and prime editing gRNA (pegRNA) which contain the RNA template to specifically insert the desired sequence [88]. In future studies, base and prime editors can be optimized for precise/specific gene editing in the eye cells to restore the vision.
One of the major issues with application of CRISPR/Cas9 for gene therapy is unintended mutations or off-target effects. Sometimes these off-target effects are observed above 50% [103]. Many strategies have been adopted to mitigate off-target effects at multiple levels such as delivery, Cas9 variants, and modulation of DNA repair pathways. One of the more effective strategies is the use of the Cas9 variant, Cas9 nickase (Cas9n), which generates the single strand break at one side of DNA with one guide RNA and a second cut at an opposite site of DNA through second gRNA. The application of Cas9 nickase has generated less off-target effects compared to wild type Cas9 [104]. In many gene editing studies, spCas9-HF1 has been used with any off-targets detected in mouse and zebrafish models, as compared with wild type spCas9 nuclease [105]. Other types of Cas variants have been developed includes evoCas9 and HifiCas9 by altering the amino acids in Rec3 domain, which is involved in the recognition of homologous DNA nucleotides. This mutation has desensitized the Rec3 domain of protein and ultimately reduced unintended mutations with high editing efficacy [106]. In another strategy, gRNA has been used by reduction in size at 5′, GC content, and optimizing the sequence. Many algorithmic based computational tools have been developed, such as CasOFFinder, and E-Crisp allowing researchers to generate the sequences with most optimized gRNA nucleotides. These tools have significantly helped to reduce the off-target effects [107,108].
In the past 10 years, CRISPR/Cas system has emerged as a revolutionizing genome editing tool that has been used to treat many ophthalmic diseases affecting the lives of thousands of patients. CRISPR/Cas9-based clinical trials are in progress to treat LCA10 to mutate the CEP290 [109]. Highly efficient gene disruption through CRISPR/Cas9 makes it a promising technique to treat the dominant mutations in rhodopsin and retinitis pigmentosa. A large number of point mutations can be generated through the HDR pathway. Numerous mammalian and non-mammalian eye disease model organisms have been generated to make the CRISPR/Cas9 a promising therapeutics tool to cure blindness. Current ongoing research in ophthalmology has shed a light on the pathogenic mechanisms of non-genetic eye diseases. In the near future, the development of highly targeted editing CRISPR tools in ophthalmology will be at the forefront to treat genetic eye diseases to cure blindness. However, the CRISPR tool application is limited due to off-target effects, heterogeneity of diseases, and ethical concerns. The early CRISPR/Cas9 tool relied on double-strand DNA repair, NHEJ, or HDR. NHEJ and HDR always lead to INDELS formation and off-target effects. The off-target effects can mutate the cancer tumor-suppressing genes. Therefore, it is crucial to predict the off-target effects before starting the experimental application of CRISPR/cas9. The off-target effects are determined by the PAM site and gRNA homologous regions. Many bioinformatics tolls have been generated that can be employed to predict the off-target effects before their application. The off-target effects can be significantly reduced with improved Cas9 variants, SpCas9-HF1, limited expression of Cas9, and improved genome editing methods, such as base editing and prime editing. In near future, researchers are trying to engineer Cas9 variants by direct evolution that will be able to decrease the off-target effects significantly [85]. The heterogeneity of eye diseases remains a challenging factor for their treatment. Many diseases have hundreds of causative mutations [110]. Using the iPSC stem cells and editing with CRISPR/Cas9 and then differentiating it into photoreceptors could be a better strategy to avoid heterogeneity. This strategy is still in its initial trials, but successful findings of these trails would support the use of CRISPR/Cas9 technology from laboratory to clinics. Researchers have focused on making the CRISPR/Cas9 as an effective and safe therapeutic tool that could be used in clinical trials. Advances in CRISPR/Cas9 technology such as base and prime editing could boost the application of CRISPR/Cas9 to treat retinal diseases in an efficient way with less cost. One of the main limitations of CRISPR/Ca9 technology would be the manufacturing and scale up for in vivo editing. With less cost and precision gene editing abilities the CRISPR/Ca9 will be a promising technology to cure blindness in the near future. | true | true | true |
PMC9569794 | 36232922 | Ryutaro Yasudome,Naohiko Seki,Shunichi Asai,Yusuke Goto,Yoshiaki Kita,Yuto Hozaka,Masumi Wada,Kan Tanabe,Tetsuya Idichi,Shinichiro Mori,Takao Ohtsuka | Molecular Pathogenesis of Colorectal Cancer: Impact of Oncogenic Targets Regulated by Tumor Suppressive miR-139-3p | 01-10-2022 | microRNA,expression signature,miR-139-3p,tumor suppressor,colorectal cancer,KRT80,HK2,AKT | We recently determined the RNA sequencing-based microRNA (miRNA) expression signature of colorectal cancer (CRC). Analysis of the signature showed that the expression of both strands of pre-miR-139 (miR-139-5p, the guide strand, and miR-139-3p, the passenger strand) was significantly reduced in CRC tissues. Transient transfection assays revealed that expression of miR-139-3p blocked cancer cell malignant transformation (e.g., cell proliferation, migration, and invasion). Notably, expression of miR-139-3p markedly blocked RAC-alpha serine/threonine-protein kinase (AKT) phosphorylation in CRC cells. A combination of in silico database and gene expression analyses of miR-139-3p-transfected cells revealed 29 putative targets regulated by miR-139-3p in CRC cells. RNA immunoprecipitation analysis using an Argonaute2 (AGO2) antibody revealed that KRT80 was efficiently incorporated into the RNA-induced silencing complex. Aberrant expression of Keratin 80 (KRT80) was detected in CRC clinical specimens by immunostaining. A knockdown assay using small interfering RNA (siRNA) targeting KRT80 showed that reducing KRT80 expression suppressed the malignant transformation (cancer cell migration and invasion) of CRC cells. Importantly, inhibiting KRT80 expression reduced AKT phosphorylation in CRC cells. Moreover, hexokinase-2 (HK2) expression was reduced in cells transfected with the KRT80 siRNAs or miR-139-3p. The involvement of miRNA passenger strands (e.g., miR-139-3p) in CRC cells is a new concept in miRNA studies. Our tumor-suppressive miRNA-based approach helps elucidate the molecular pathogenesis of CRC. | Molecular Pathogenesis of Colorectal Cancer: Impact of Oncogenic Targets Regulated by Tumor Suppressive miR-139-3p
We recently determined the RNA sequencing-based microRNA (miRNA) expression signature of colorectal cancer (CRC). Analysis of the signature showed that the expression of both strands of pre-miR-139 (miR-139-5p, the guide strand, and miR-139-3p, the passenger strand) was significantly reduced in CRC tissues. Transient transfection assays revealed that expression of miR-139-3p blocked cancer cell malignant transformation (e.g., cell proliferation, migration, and invasion). Notably, expression of miR-139-3p markedly blocked RAC-alpha serine/threonine-protein kinase (AKT) phosphorylation in CRC cells. A combination of in silico database and gene expression analyses of miR-139-3p-transfected cells revealed 29 putative targets regulated by miR-139-3p in CRC cells. RNA immunoprecipitation analysis using an Argonaute2 (AGO2) antibody revealed that KRT80 was efficiently incorporated into the RNA-induced silencing complex. Aberrant expression of Keratin 80 (KRT80) was detected in CRC clinical specimens by immunostaining. A knockdown assay using small interfering RNA (siRNA) targeting KRT80 showed that reducing KRT80 expression suppressed the malignant transformation (cancer cell migration and invasion) of CRC cells. Importantly, inhibiting KRT80 expression reduced AKT phosphorylation in CRC cells. Moreover, hexokinase-2 (HK2) expression was reduced in cells transfected with the KRT80 siRNAs or miR-139-3p. The involvement of miRNA passenger strands (e.g., miR-139-3p) in CRC cells is a new concept in miRNA studies. Our tumor-suppressive miRNA-based approach helps elucidate the molecular pathogenesis of CRC.
According to the World Health Organization (Globocan 2020), colorectal cancer (CRC) is the third most common cancer (over 1,800,000 cases) worldwide and the second leading cause of cancer-related deaths (over 880,000 deaths) [1]. In clinical practice, the prognosis of CRC is relatively good if diagnosed early. However, the prognosis is consistently poor in advanced cases, with a 5-year survival rate of approximately 14% (stage III or stage IV metastatic disease) [2]. At the time of the initial diagnosis, approximately 14–18% of patients with CRC have metastases, and the treatment strategies for unresectable cases are limited [3]. The oncogenesis of CRC is illustrated by a well-known multistep model of cancer cells [4,5,6]. From previous studies, mutations in various genes involved in the oncogenesis of CRC (e.g., APC, TP53, SMAD4, KRAS, and PIK3CA) and activation of cancer signaling pathways (e.g., WNT, RAS/MAPK, PI3K, TGF-β, P53, and DNA mismatch-repair) caused by these gene mutations have been reported [4]. CRC cells have highly heterogeneous properties, requiring new therapeutic parameters for CRC from genetic and genomic points of view. As a result of this molecular heterogeneity, recent genome-wide transcriptome analyses have revealed that CRC cells can be molecularly classified into four consensus molecular subtypes (CMS1 to CMS4) [7]. The future treatment strategies for CRC patients will be based on these subtypes. As a result of the Human Genome Project, it has become clear that a vast number of functional non-coding RNA molecules (ncRNAs) are present in the human genome [8]. Current studies have shown that numerous ncRNAs play important roles in various biological activities such as the stabilization of RNA molecules and regulation of gene expression and the cell cycle [9,10]. Extensive research to date has revealed that ncRNA dysregulation is deeply involved in the initiation and development of human diseases, including cancer [11]. Among ncRNAs, microRNAs (miRNAs), consisting of only 19–22 nucleotides, have been well studied in cancer research fields. They function as fine-tuners of gene expression control in a sequence-dependent manner [12]. A single miRNA controls numerous genes, and in turn, a single gene is controlled by numerous miRNAs [13]. As a result, miRNAs and their target genes form a very complex network within cells, and it is easy to imagine that aberrant expression of miRNAs disrupts this RNA network. Many studies have shown aberrant expression of miRNAs in CRC cells, and these miRNAs act as oncogenes and/or tumor suppressors by targeting cancer-related genes in CRC cells [14,15,16]. More recently, to identify aberrantly expressed miRNAs in CRC cells, we determined the miRNA expression signature of CRC by RNA sequencing [17]. Our signature revealed that both the guide and passenger strands of 16 miRNAs (e.g., miR-9, miR-28, miR-29c, miR-30a, miR-99a, miR-100, miR-125b, miR-129, miR-133a, miR-139, miR-143, miR-145, miR-218, miR-195, miR-490, and miR-497) derived from pre-miRNAs were downregulated [17]. Our recent studies showed that some passenger strands of miRNAs (e.g., miR-30a, miR-99a, miR-143, miR-145, and miR-490) act as tumor-suppressive miRNAs in a wide range of cancers [17,18,19,20,21]. Interestingly, the genes regulated by a specific miRNA differ depending on the type of cancer. In this study, we focused on miR-139-3p (the passenger strand of pre-miR-139) and investigated its functional significance and target oncogenes in CRC cells. Notably, ectopic expression of miR-139-3p markedly blocked the phosphorylation of RAC-alpha serine/threonine-protein kinase (AKT) in CRC cells. Our search strategy for miRNA targets revealed a total of 29 genes as putative candidate targets of miR-139-3p in CRC cells. Of these, keratin 80 (KRT80) was found to be a direct target of miR-139-3p, and its aberrant expression enhanced the malignant transformation of CRC cells. Involvement of the passenger strand of miRNA and its gene targets in CRC pathogenesis is a new concept and provides novel insights into the molecular pathogenesis of CRC.
Recently, we determined the miRNA expression signature of CRC by miRNA sequencing using CRC clinical specimens (GEO accession number: GSE183437). Analysis of the signature showed that 84 miRNAs were upregulated, and 70 were downregulated in CRC tissues (Figure 1A). Among downregulated miRNAs in CRC tissues, we focused on miR-139-5p (the guide strand) and miR-139-3p (the passenger strand), because both strands of miRNAs derived from pre-miR-139 were significantly downregulated in CRC tissues. Our interest is to clarify how the passenger strand of miRNA is involved in the malignant transformation of CRC cells. The mature sequences of the two microRNAs are shown in Figure 1B. CRC tissues and noncancerous tumor-adjacent tissues (27 paired) were used to verify the expression status of miR-139-5p, miR-139-3p, and their target genes. Clinical information of these specimens is shown in Table S1. The expression levels of miR-139-5p (p < 0.001) and miR-139-3p (p < 0.001) were significantly lower in CRC tissues than normal tissues (Figure 1C). Next, we examined the expression levels of miR-139-5p and miR-139-3p in two CRC cell lines, HCT116 and DLD-1. In these cell lines, the expression levels of miR-139-5p and miR-139-3p were lower than those in normal epithelial tissues (Figure 1C). Furthermore, a positive correlation was detected between miR-139-5p and miR-139-3p expression levels by Spearman’s rank analysis (r = 0.559, p < 0.001; Figure 1D).
To investigate the tumor-suppressive functions of miR-139-5p and miR-139-3p, we ectopically expressed mature miR-139-5p and miR-139-3p in two CRC cell lines, HCT116 and DLD-1, and performed functional assays, e.g., cancer cell proliferation, migration, and invasion. After miR-139-5p transfection, cancer cell migration in both cell lines was significantly inhibited (Figure 2A–C). In contrast, the malignant phenotypes of cancer cells, e.g., proliferation, migration, and invasion, were significantly reduced by miR-139-3p transfection in both cell lines (Figure 2A–C). Representative images from the migration and invasion assays are shown in Figure S1.
Based on these expressions and functional analysis, aberrant expression of miR-139-3p and disruption of its gene regulation mechanisms were considered to be more deeply involved in the malignant pathogenesis of CRC. We focused on miR-139-3p (passenger strand) for further validation. We investigated whether epigenetic modifications affect the downregulation of miR-139-3p in CRC cells. After treatment of Trichostatin A (TSA) in CRC cells, the expression level of miR-139 was increased compared to TSA untreated cells (Figure S2A). In addition, miR-139-3p expression level was elevated by 5-aza-2-deoxycytidine (5-aza-dC) treatment in CRC cells (Figure S2B). These results suggest that histone deacetylation and DNA methylation are closely involved in the downregulation of miR-139-3p in CRC cells. The following hypotheses regarding miR-139-3p target genes in CRC cells were made: the target genes of miR-139-3p have one or more binding site(s), are downregulated after miR-139-3p transfection in CRC cells, and are upregulated in CRC tissues. We combined the gene expression data from two databases (TargetScan and GEPIA2) with gene expression data from miR-139-3p-transfected CRC cells (GSE155659) to search for genes that meet these three criteria. A flowchart of the search strategy is shown in Figure 3. A total of 95 putative targets of miR-139-3p in CRC cells were identified. We assessed the expression levels of putative miR-139-3p target genes in CRC clinical tissues using The Cancer Genome Atlas database via the GEPIA2 platform. A total of 29 genes were significantly upregulated in CRC clinical specimens (colon adenocarcinoma or rectal adenocarcinoma) in this database (p < 0.01: Table 1, Figure S3). GEPIA2 analysis revealed that the expression level of KRT80 was fairly low in normal tissues (Figure S3). Genes expressed exclusively in cancer cells are appropriate therapeutic targets for CRC. We focused on KRT80 in the subsequent functional analyses in CRC cells.
In CRC cells transfected with miR-139-3p, both the mRNA and protein levels of KRT80 were significantly downregulated (Figure 4A). Next, RNA immunoprecipitation (RIP) analysis was performed to confirm that KRT80 mRNA was incorporated into the RNA-induced silencing complex (RISC) after miR-139-3p transfection. The RIP assay concept is illustrated in a schematic in Figure 4B. In samples subjected to immunoprecipitation using an Argonaute2 (AGO2) antibody, quantitative real-time reverse-transcription PCR (qRT-PCR) showed that the KRT80 mRNA level was significantly higher than that in mock and miRNA control-transfected cells (p < 0.001; Figure 4B). Ago2-bound miR-139-3p and KRT80 mRNA were isolated by immunoprecipitation using the AGO2 antibody, suggesting that the RISC plays a central role in miRNA biogenesis (Figure 4B). Finally, a dual-luciferase reporter assay was performed to confirm that miR-139-3p binds directly to the 3′ untranslated regions (UTR) of KRT80. Luciferase activity was significantly reduced following co-transfection with miR-139-3p and a vector containing the miR-139-3p-binding site within the 3’-UTR of KRT80 (Figure 4C). In contrast, co-transfection with a vector containing the KRT80 3’-UTR in which the miR-139-3p-binding site was deleted resulted in no change in luciferase activity (Figure 4C).
To assess the functional significance of KRT80 in CRC cells, we performed knockdown assays using siRNAs corresponding to KRT80 mRNA. First, the inhibitory effects of two different siRNAs (siKRT80-1 and siKRT80-2) targeting KRT80 in two cell lines were examined. Both KRT80 mRNA and protein levels were effectively suppressed after transfection of each siRNA into HCT116 and DLD-1 cells (Figure S4). Knockdown of KRT80 slightly inhibited cell proliferation (Figure 5A) and markedly inhibited migration and invasion in both HCT116 and DLD-1 cells (Figure 5B,C). Representative photographs from the migration and invasion assays are shown in Figure S5. Based on the previous report that overexpression of KRT80 induced epithelial-mesenchymal transition (EMT)-related genes and activated AKT signaling via phosphorylation of AKT (Ser 473) [22], Western blotting for phosphorylation of AKT was performed on KRT80 and miR-139-3p. Notably, transfection of the KRT80 siRNAs suppressed the phosphorylation of AKT (Figure 5D). In addition, expression of miR-139-3p markedly inhibited the phosphorylation of AKT in CRC cells, according to Western blot analysis.
Protein expression of KRT80 was assessed by immunohistochemistry in CRC clinical specimens. Overexpression of KRT80 protein was detected in cancer lesions (Figure 6).
To explore KRT80-regulated RNA networks in CRC, we performed comprehensive gene expression analyses in KRT80-knockdown CRC cells. A total of 52 genes were identified as downregulated in both KRT80-knockdown CRC cell lines (log2 fold change < −1.0: Table 2). Our expression data were deposited in the GEO database (GEO accession number: GSE208785). In this study, we focused on hexokinase 2 (HK2) because it was identified as a miR-139-3p target in CRC cells (Table 1). HK2 was commonly regulated by miR-139-3p and KRT80 in CRC cells (Figure 7A). Moreover, HK2 was directly regulated by miR-139-3p in CRC cells, by RIP assay and dual luciferase reporter assay (Figure S6). In addition, HK2 expression was upregulated in CRC tissues (Figure S3), and a vast number of studies showed that aberrant expression of HK2 enhances cancer cell malignant transformation in various types of cancers. Our results showed that HK2 expression was reduced in cells transfected with siKRT80 (Figure 7B) or miR-139-3p (Figure 7C).
In the analysis using surgical specimens (27 paired normal and cancerous tissues), we observed marked suppression of miR-139-3p and marked upregulation of KRT80 in cancer tissues (Figure 1C and Figure S7A). In addition, a negative correlation was observed between the expression of miR-139-3p and KRT80 in CRC specimens (Figure S7B). Contrary to the TCGA data analysis, we did not find any significant upregulation of HK2 in our cancerous samples.
Because CRC is a heterogeneous disease, as indicated by our genome-wide transcriptome analysis, it is necessary to search for diagnostic markers and therapeutic target molecules in an individualized manner. Recently, we determined the miRNA expression signature of CRC using RNA sequencing [17]. In that study, we found that miR-490-3p acted as a tumor-suppressive miRNA in CRC cells, and expression of its gene targets (IRAK1, FUT1, and GPRIN2) was significantly predictive of 5-year overall survival in CRC patients [17]. This new miRNA expression signature of CRC will be a useful tool for elucidating the molecular pathogenesis of this disease. Aberrant expression of miRNAs is frequently observed in several types of cancers [14,15,16]. A vast number of studies showed that epigenetic modification (histone modifications and promoter DNA methylation) is closely involved in the silencing of miRNAs expression in cancer cells [23,24,25,26]. A recent study showed that miR-139 was epigenetically silenced by histone H3 lysine 27 trimethylation (H3K27me3) in lung cancer cells [25]. Our present data (TSA and 5-aza-dC treatment) showed that both events of histone deacetylation and DNA methylation were closely involved in the silencing of miR-139-3p on CRC cells. It has been shown that miR-139-3p silencing plays a pivotal role in human oncogenesis. Our recent studies revealed that some passenger strands of miRNAs are closely involved in the molecular pathogenesis of a wide range of human cancers, e.g., miR-30c-2-3p, miR-101-5p, miR-143-5p, and miR-145-3p [19,21,27,28]. Based on our CRC signature, we focused on miR-139-3p (the passenger strand derived from pre-miR-139) in this study. We have analyzed the passenger stand miR-139-3p in several types of cancers and found that it acts as a tumor-suppressive miRNA in bladder cancer, renal cell carcinoma, and head and neck squamous cell carcinoma by targeting several genes closely linked to cancer pathogenesis [29,30,31]. Here, the function of miR-139-3p in CRC cells was clarified and found to be consistent with previous reports. As we have discussed, our in vitro assays showed that miR-139-3p acted as a tumor suppressive miRNA in CRC cells. However, the endogenous expression levels of passenger strands of miRNAs are little, and the full picture of the functions of passenger strands of miRNAs in vivo remains unknown. In order to investigate the in vivo functions of miRNAs, it is essential to generate and analyze cells that constitutively express miRNAs or cells in which miRNA expression is completely knocked out. Several oncogenic signaling pathways are activated in CRC cells, of which PI3K/AKT/mTOR signaling is frequently activated [32,33,34,35]. Therefore, inhibiting activation of this signaling pathway is an attractive strategy for CRC treatment [32,36,37,38]. The AKT serine/threonine kinase is activated by phosphatidylinositol-3 kinase (PI3K) or phosphoinositide-dependent kinases via phosphorylation of Thr308 or Ser473 in AKT and activated AKT phosphorylates various downstream protein substrates (e.g., mTOR, glycogen synthase kinase 3 beta, and forkhead box protein O1) [39]. Aberrant expression and activation of AKT have been observed in many types of cancers, including CRC [40]. Notably, ectopic expression of miR-139-3p inhibited the phosphorylation of AKT in CRC cells in this study. Next, we searched for target genes regulated by miR-139-3p in CRC cells, particularly those involved in AKT phosphorylation. A unique feature of miRNAs is that they regulate different sets of genes depending on the cancer cell type. We identified 29 genes as tumor-suppressive targets of miR-139-3p in CRC cells. Of these, we focused on KRT80 because its expression was significantly different between cancer and normal tissues. Ideally, a therapeutic target molecule for cancer is not expressed in normal cells. Expression levels of KRT80 in normal tissues were assessed using previous large-scale transcriptional analysis data [41]. Expression of KRT80 was detected in skin, esophagus, and salivary glands. In contrast, KRT80 was hardly expressed in other tissues (Figure S8). We showed that aberrant expression of KRT80 enhanced the malignant phenotypes of cancer cells (i.e., proliferation, migration, and invasion). Interestingly, overexpression of KRT80 induced EMT-related genes and activated the AKT signaling through phosphorylation of AKT (Ser 473) [22]. Considering our present data and previous reports, it was strongly suggested that the miR-139-3p/KRT80/p-AKT axis influences the migration and invasive abilities of CRC cells. In ovarian cancer, overexpression of KRT80 induced the expression of genes related to epithelial–mesenchymal transition and activated both MEK and ERK [42]. In gastric cancer, overexpression of the circular RNA CircPIP5K1A induced expression of KRT80 and activated the PI3K/AKT pathway via miR-671-5p adsorption [43]. Moreover, KRT80 expression was significantly correlated with clinical parameters, such as lymph node metastasis and pathological stage, in CRC and ovarian cancer [22,42]. Together, these data suggest that KRT80 is a potential therapeutic target for CRC. We also investigated genes affected by KRT80 in CRC cells. In CRC cells, the expression of several genes was suppressed after the knockdown of KRT80 expression. Among these genes, we focused on HK2. The four members of the HK family (HK1-4) in mammals catalyze the conversion of glucose to glucose-6-phosphate, and they are involved in the first and rate-limiting step of glycolysis [44,45,46]. Previous studies reported that Akt and HK2 are overexpressed in cancer cells and that there is a positive correlation between activation of the PI3K/Akt/mTORC1 pathway and HK2 expression [47,48,49]. These findings indicate that simultaneous inhibition of glycolysis and the AKT/mTOR signaling pathway is effective in suppressing the growth of cancer cells [50].
Fifty-four clinical specimens (27 CRC tissues and 27 normal colon tissues) were used to evaluate the expression status of miR-139-5p/3p. All specimens used in this study were obtained by surgical resection at Kagoshima University Hospital between 2014 and 2017. Normal colon tissue was collected from adjacent sites to the specimen from which each CRC tissue sample was taken. All patients provided written informed consent for the use of their specimens. This study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the Ethics Committee of Kagoshima University (approval number 160038 (28–65); date of 19 March 2021). The clinical information was described in our previous study [17].
Two CRC cell lines, HCT116 and DLD-1, were used in this study. HCT116 cells were obtained from the RIKEN Cell Bank (Tsukuba, Ibaraki, Japan), and DLD-1 cells were obtained from the Cell Resource Center for Biomedical Research Bank (Sendai, Miyagi, Japan). HCT116 was cultured in DMEM medium supplemented with 10% concentration of fetal bovine serum (FBS), and DLD-1 was cultured in RPMI-1640 medium, also supplemented with 10% concentration of fetal bovine serum (FBS).
The protocols used for RNA extraction and qRT-PCR were described in our previous studies [51,52]. In brief, Total RNA was isolated from cell lines using TRIzol reagent according to the manufacturer’s protocol. RNA samples were reverse transcribed using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, MA, USA). qPCR was performed using PCR Master Mix (Applied Biosystems, Waltham, MA, USA) and Fast SYBR Gren Master Mix (Applied Biosystems, Waltham, MA, USA), StepOnePlus real-time PCR system (Applied Biosystems, Waltham, MA, USA). Gene expressions were quantified relatively by the delta-delta Ct method (used GUSB as internal control). TaqMan assays used in this study are summarized in Table S2. The sequences of primers for SYBR green assays are summarized in Table S3.
Cells were treated with 5-Aza-2′-deoxycytidine (5-aza-dC; Wako, Osaka, Japan) at concentrations of 0.5, 1,2,5, and 10 μmol/L for 96 h. Cells were first cultured in growth medium; after 24 h of incubation, the medium was replaced with fresh medium containing 5-aza-dC or Dimethyl sulfoxide (DMSO, negative control), and cells were incubated for another 48 h; after 48 h of treatment, the medium was again replaced with fresh medium containing 5-aza-dC or DMSO and cells were cultured for additional 48 h. After 120 h treatment, total RNA was isolated. The expression levels of miR-139-3p were measured by qRT-PCR.
Cells were treated with Tricostatin A (TSA; Wako, Osaka, Japan) at 0.1 or 0.5 μmol/L concentration for 24 h. Cells were first grown in growth medium. After 24 h of incubation, the medium was replaced with fresh medium containing TSA or DMSO and the cells were incubated for an additional 24 h. After 48 h treatment, total RNA was isolated. Expression levels of miR-139-3p were measured by qRT-PCR.
The protocols used for transient transfection of miRNAs and siRNAs were described in our previous studies [51,52]. The miRNA precursors and siRNAs used in this report were detailed in Table S2. Opti-MEM (Gibco, Carlsbad, CA, USA) and LipofectamineTM RNAiMax Transfection Reagent (Invitrogen, Waltham, MA, USA) were used for miRNA and siRNA transfection of miRNAs and siRNAs into CRC cell lines. All miRNA precursors and siRNAs were transfected into the CRC cell line at 10 nM. Mock transfection consisted of cells without precursors or siRNAs. Control groups were transfected with the negative control precursor.
The tumor-suppressive functions of miRNAs were evaluated by transient transfection assays using mature miR-139-5p and miR-139-3p. The tumor-promoting functions of KRT80 (loss-of-function assays) were assessed by siRNA transfection assays using siRNAs targeting KRT80. Functional assays (proliferation, migration, and invasion assays) were performed according to procedures of previous studies [51,52]. Briefly, for proliferation assays, HCT116 or DLD-1 cells were transferred into 96-well plates at 3.0 × 103 cells/well. Cell proliferation was assessed using XTT assay kit II (Sigma-Aldrich, St. Louis, MO, USA) 72 h after the transfection procedure. For the migration and invasion assay, HCT116 and DLD-1 cells were transfected in 6-well plates at 3.0 × 105 cells/well; 48 h later, transfected HCT116 and DLD-1 cells were added to each chamber at 1.0 × 105 cells/well. Corning BioCoatTM cell culture chambers (Corning, Corning, NY, USA) were used for the migration assay and Corning BioCoat Matrigel Invasion Chambers were used for the invasion assay. cells on the underside of the chamber membrane were stained and counted for analysis. All experiments were performed in triplicate. The details of the reagents used in these analyses are listed in Table S2.
To identify oncogenic targets controlled by miR-139-3p in CRC cells, data were merged from the following sources to narrow down the targets: (1) Target Scan Human 8.0 database (http://www.targetscan.org/vert_80, accessed on 6 August 2021) [53], (2) gene expression data from miR-139-3p transfected CRC cells (GEO accession number, GSE155659), and (3) gene expression database from CRC clinical tissues using the GEPIA2 platform (http://gepia2.cancer-pku.cn/#index; accessed on 10 April 2022) [54].
The assay for RIP was performed according to previous studies [55]. Briefly, CRC cells were cultured in 6-well dish at 3.0 × 105/well concentration. Negative control miRNA precursors and miR-139-3p precursors were transfected. After 12 h, immunoprecipitation was performed using the MagCaptureTM microRNA Isolation Kit, Human Ago2, obtained from FUJIFILM Wako Pure Chemical Corporation (Wako, Osaka, Japan) according to the manufacturer’s protocol. Expression levels of KRT80 and HK2 bound to Ago2 were measured by qRT-PCR. TaqMan assays used in this study are summarized in Table S2. The sequences of primers for SYBR green assays are summarized in Table S3.
The dual-luciferase reporter assay was performed to determine whether miR-139-3p binds directly to the 3′-UTR of KRT80. A partial wild-type sequence, including the seed sequence, of the KRT80 3′-UTR, was inserted into the psiCHECK-2 vector (C8021; Promega, Madison, WI, USA). Alternatively, the same KRT80 3′-UTR sequence but with the miR-139-3p binding site deleted was also inserted into the same vector to create the deletion-type construct. The design of each vector cloning sequence into wild-type and deletion-type were shown in Figures S9 and S10 mRNA sequences of KRT80 and HK2 were cited from National Center for Biotechnology Information database [56]. The dual-luciferase reporter assay was performed according to previous studies [17,52]. The reagents used in the assay are listed in Table S2.
The procedures for Western blot and immunohistochemical analyses were performed according to our previous studies [51,52]. In brief, 72 h after transfection, cells were collected, and lysates were prepared. Next, 18 μg/lane of protein lysate was separated on e-PAGEL (ATTO, Tokyo, Japan), transferred to PVDF membranes, and incubated with primary antibody overnight at 4 °C and with secondary antibody for 1 h at room temperature. GAPDH was used as an internal control. The antibodies used are listed in Table S2, and the clinical specimens evaluated by immunohistochemistry are shown in Table S4.
JMP Pro 15 (SAS Institute Inc., Cary, NC, USA) was used for the statistical analyses. Differences between two groups were assessed using Welch’s t-test and those among multiple groups using Dunnett’s test. Spearman’s test was used for the correlation analyses. A p-value less than 0.05 was considered statistically significant.
Based on the miRNA expression signature of CRC obtained by RNA sequencing, the expression of miR-139-3p (the passenger strand) was significantly reduced in CRC tissues. Functional assays revealed that expression of miR-139-3p attenuated cancer cell malignant phenotypes, indicating that miR-139-3p acts as a tumor suppressor in CRC cells. KRT80 was identified as a direct target of miR-139-3p, and aberrant expression of KRT80 was confirmed in CRC clinical specimens. Moreover, HK2 expression was regulated by both miR-139-3p and KRT80 in CRC cells. Exploration of miRNA-regulated molecular networks provides important information for identifying therapeutic targets for CRC. | true | true | true |
PMC9569848 | 36232699 | Li Liu,De-Sheng Pei | Insights Gained from RNA Editing Targeted by the CRISPR-Cas13 Family | 27-09-2022 | CRISPR/Cas13,Cas13d,Cas13X,RNA cleavage activity,CRISPR-Cas VI system | Clustered regularly interspaced short palindromic repeat (CRISPR)/CRISPR-associated protein (Cas) systems, especially type II (Cas9) systems, have been widely developed for DNA targeting and formed a set of mature precision gene-editing systems. However, the basic research and application of the CRISPR-Cas system in RNA is still in its early stages. Recently, the discovery of the CRISPR-Cas13 type VI system has provided the possibility for the expansion of RNA targeting technology, which has broad application prospects. Most type VI Cas13 effectors have dinuclease activity that catalyzes pre-crRNA into mature crRNA and produces strong RNA cleavage activity. Cas13 can specifically recognize targeted RNA fragments to activate the Cas13/crRNA complex for collateral cleavage activity. To date, the Cas13X protein is the smallest effector of the Cas13 family, with 775 amino acids, which is a promising platform for RNA targeting due to its lack of protospacer flanking sequence (PFS) restrictions, ease of packaging, and absence of permanent damage. This study highlighted the latest progress in RNA editing targeted by the CRISPR-Cas13 family, and discussed the application of Cas13 in basic research, nucleic acid diagnosis, nucleic acid tracking, and genetic disease treatment. Furthermore, we clarified the structure of the Cas13 protein family and their molecular mechanism, and proposed a future vision of RNA editing targeted by the CRISPR-Cas13 family. | Insights Gained from RNA Editing Targeted by the CRISPR-Cas13 Family
Clustered regularly interspaced short palindromic repeat (CRISPR)/CRISPR-associated protein (Cas) systems, especially type II (Cas9) systems, have been widely developed for DNA targeting and formed a set of mature precision gene-editing systems. However, the basic research and application of the CRISPR-Cas system in RNA is still in its early stages. Recently, the discovery of the CRISPR-Cas13 type VI system has provided the possibility for the expansion of RNA targeting technology, which has broad application prospects. Most type VI Cas13 effectors have dinuclease activity that catalyzes pre-crRNA into mature crRNA and produces strong RNA cleavage activity. Cas13 can specifically recognize targeted RNA fragments to activate the Cas13/crRNA complex for collateral cleavage activity. To date, the Cas13X protein is the smallest effector of the Cas13 family, with 775 amino acids, which is a promising platform for RNA targeting due to its lack of protospacer flanking sequence (PFS) restrictions, ease of packaging, and absence of permanent damage. This study highlighted the latest progress in RNA editing targeted by the CRISPR-Cas13 family, and discussed the application of Cas13 in basic research, nucleic acid diagnosis, nucleic acid tracking, and genetic disease treatment. Furthermore, we clarified the structure of the Cas13 protein family and their molecular mechanism, and proposed a future vision of RNA editing targeted by the CRISPR-Cas13 family.
The structure of the CRISPR-Cas system has gone through more than 20 years from discovery to function revelation, and its basic structure and molecular mechanism have been gradually elucidated in recent years. In 1987, a Japanese scientist stumbled upon a 29 bp repeating sequence while studying the isozyme conversion of alkaline phosphatase in Escherichia coli. Unfortunately, there is not enough DNA sequence data available, and Ishino et al. could not explain the function of these unique sequences [1]. Mojica et al. first identified a 30 bp DNA fragment that repeated at a regular distance in the genome of Haloferax mediterranei, calling this element “short regularly spaced repeats” (SRSRs) in 1993 [2]. In 2002, these SRSRs were named “CRISPR” by Mojica [3]. Three years later, several bioinformatics teams almost simultaneously found that CRISPR spacer sequences were consistent with foreign DNA sequences of invading bacteria, suggesting that the CRISPR system might play a major role in microbial immunity [4,5,6]. In 2011, the CRISPR-Cas system was first classified according to the DNA structure and the dynamic evolution of the CRISPR-Cas system [7]. In 2012, Jinek et al. elucidated the mechanism of CRISPR/Cas9 and highlighted the potential use of this system for programmable genome editing [8]. The next year, Mali et al. successfully applied the technique to the human genome and other organisms [9]. Two years later, two studies reported type II CRISPR systems from Streptococcus thermophilus and Streptococcus pyogenes and successfully edited the genomes of mammalian cells [10]. In 2015, the discovery of Cpf1 (Cas12a) provided a more flexible way of selecting DNA targets compared to Cas9. Cpf1 (Cas12a) recognizes AT-rich PAMs, and it can target AT-rich regions. However, Cpf1 (Cas12a) has difficulty finding targets in GC-rich regions [11]. Therefore, Cpf1 (Cas12a), together with Cas9, expands the target sites for selection. Cas13, a type VI CRISPR system-related protein discovered in 2016, is an RNA-guided and RNA-targeted ribonuclease protein family, which avoids permanent damage to the DNA of organisms [12,13,14]. Since then, due to the emergence of RNA editing targeted by the CRISPR-Cas13 family, CRISPR technology has broadened its research scope and ushered in a bright prospect for the future. The CRISPR-Cas systems are an adaptive immune system derived from prokaryotes, which widely exist in 40% of bacteria and 90% of archaea to protect bacteria and archaea from exogenous genetic invasion through the interaction of RNA and CRISPR-related proteins [15,16,17]. The CRISPR-Cas system consists of three parts: the leader sequence (LS), an operon containing a set of cas genes, and a CRISPR DNA array. The leading sequence is about 200–500 bp upstream of the first CRISPR repeat sequence rich in A/T and provides recognition sites [18,19]. The CRISPR DNA array is composed of 21–48 bp length repeats and spacer sequences, and a hairpin structure can be formed and repeated up to 250 times, which is the binding region of Cas protein [20]. Spacers are separated by repeated regulatory sequences and form the CRISPR array with leader sequences and repeated sequences [5,6]. The CRISPR-associated gene is highly conserved near the CRISPR locus. The Cas protein encoded by the CRISPR-associated gene contains endonucleases, helicases, and a binding domain with ribonucleic acid, which can recognize foreign DNA and cut the invading DNA through site-specific cleavage [21]. Bacteria and archaea use the adaptive immune system to protect exogenous genetic components from phages and nucleic acids. CRISPR-mediated adaptive immunity consists of three phases: adaptation, maturation, and interference (Figure 1) [21]. In the adaptation phase, the CRISPR-Cas system acquires new spacer sequences. When the small gene fragments from the invader enter bacteria and archaea containing the CRISPR-Cas system, the host’s CRISPR-associated protein complex binds to the protospacer adjacent motifs (PAM) of the gene fragments. A new spacer sequence was formed under the action of related proteins and established structural motifs for the acquired immunity of bacteria [22]. Mature CRISPR RNA (crRNA) was generated by the CRISPR-Cas expression system during the maturation phases. When foreign nucleic acids invaded again, the CRISPR sequence of the CRISPR-Cas system was transcribed into a sequence-specific CRISPR RNA precursor (pre-crRNA), which was further processed and spliced by Cas proteins to form mature CRISPR RNA (crRNA). Interference phases: ribonucleoprotein complex with crRNA and Cas protein can detect and bind the target invading genome (DNA/RNA) by crRNA, and eventually leads to the breakage or degradation of the invader genome (DNA/RNA) [23,24]. This study mainly reviews the systematic classification, defense mechanism, and the application prospect of RNA editing targeted by the CRISPR-Cas13 family, and especially highlights the latest research progress of the CRISPR-Cas13 family, including Cas13d and Cas13X. The existing problems of RNA editing systems based on the Cas13 family are also analyzed.
The CRISPR system is classified mainly according to its evolutionary development of the CRISPR system, Cas protein sequence, and genomic locus architecture [7]. The adaptation module in several CRISPR-Cas systems possesses endonuclease Cas1 and structural subunit Cas2, which are critical for acquiring spacers [25]. According to the structural composition of the Cas effector protein complex, the currently known CRISPR-Cas system can be divided into two classes (class 1 and class 2), and each can be further divided into three types. Among them, the class 1 system (including type I, III, and IV) exist in bacteria and archaea, requiring complex multiple subunits and crRNA to form the CRISPR-Cas complex. Type I and type III effectors have a common origin [26,27]. In contrast, the class 2 system (including type II, V, and VI) exists almost exclusively in bacteria, requiring a single Cas protein and a crRNA to form the CRISPR-Cas complex [13,26,28,29]. In the class 2 system, Cas9 is a type II effector complex containing two unrelated nuclease domains (HNH and RuvC), while the type V effector Cpf1 (Cas12a) contains only one nuclease domain (RuvC-like). Of note, type VI CRISPR systems including the CRISPR-Cas13 family are unique because their Cas proteins contain two conserved “higher eukaryotic and prokaryotic nucleotide binding (HEPN)” domains with RNase activity [30,31,32,33]. The property difference between Cas9 and Cas13 can be seen in Table 1. Due to the complexity of multi-polymerization and the large size of Cas proteins in the class 1 system, Cas proteins in the class 2 system, especially CRISPR-Cas13 of type VI CRISPR systems, provide promising applications for gene editing and RNA editing.
Recently, researchers identified a CRISPR-Cas13 system in type VI CRISPR systems that can target RNA with high specificity when screening Cas proteins in microbial genome and metagenomics data. To date, the CRISPR-Cas13 system is the only known RNA editing in prokaryotes [12,13,34,35,36]. The Cas13 protein effectors belong to type VI CRISPR systems in the class 2 CRISPR system, and the lack of a RuvC domain of Cas13 effectors is the main difference from other class 2 CRISPR-Cas effectors (type II and V). However, the Cas13 proteins family has two HEPN domains, which can process precursor RNA into mature crRNA and cleave target RNA mediated by crRNA (except for Cas13X). CRISPR-Cas13 members can be divided into four subtypes (A, B1, B2, C, D, X, and Y) according to their phylogeny. A homology analysis indicated that those four subtypes shared low homology except for the HEPN domain (Figure 2). Interestingly, the most well-studied Cas protein of type VI CRISPR systems is Cas13a, previously called the C2c2 effector [12]. The Cas13a effector (~1000 aa) is the earliest discovered Cas endonuclease that cleaves target ssRNA. Cas13a is a bilobed effector protein, consisting of a crRNA recognition (REC) lobe and a nuclease (NUC) lobe. The REC lobe contains an N-terminal domain (NTD) and a Helical-1 domain, which can bind and recognize crRNA, respectively. The NUC lobe contains two HEPN domains, of which Helical-2 and Helical-3 are separated between two HEPN domains [12]. In 2016, Cas13b (C2c6) was identified via a computational pipeline that searches for CRISPR-Cas locus conserved sequences from the microbial genome and metagenomic data. The size of the Cas13b protein is about 1100 to 1200 aa, and its structure contains a crRNA recognition lobe (REC) and two NUC lobes. Only two HEPN domains of N-terminal and C-terminal from different species are conserved, while other domains show no obvious similarity in amino acid sequence, implying that there are different structures among Cas13b family members [32]. As a member of type VI CRISPR systems, Cas13c (C2c7) likely possesses RNA cleavage activity, but further studies are needed due to there not being enough structural and functional data available. Yan et al. developed a computational pipeline to construct an expanded database of class 2 CRISPR-Cas systems from microbial genomic and metagenomic data. Then, a new CRISPR-associated ribonuclease was discovered, named Cas13d as a type VI-D CRISPR system [36]. The type VI-D CRISPR-Cas locus was originally identified from gram-positive intestinal bacteria Eubacterium and Ruminococcus. Cryo-electron microscopy analysis of Cas13d indicated that it was a ~930 aa endoribonuclease and shared less similarity except for two conserved R-X4–6 H HEPN motifs (HEPN1 and HEPN2), compared to previous type VI CRISPR systems [14]. Cas13d protein from different species generally owns a bilobed architecture of class 2 effectors. This protein contains five distinct functional domains around the central guide crRNA, which has a REC lobe and a NUC lobe. In the tertiary crystal structure of the complex, the REC lobe possesses an N-terminal domain (NTD) and a Helical 1 domain, while the NUC lobe includes HEPN-1, Helical-1, Helical-2, and HEPN-2 domains. The HEPN-1 domain functions as a structural scaffold to link the REC and NUC lobes [33]. In the main sequence, the structure of Cas13d almost has the same functional domain as Cas13a except for the Helical-1 domain. Besides, Cas13d has a compact structure, and five functional domains of Cas13d are essential for RNase activity, compared to other Cas13 effectors. Moreover, the Cas13d locus as a type VI-D CRISPR system is significantly different from that of other Cas13 effectors (type VI-A and type VI-B) [33]. The crRNA repeat region of Cas13d was relatively conserved in the spacer and its secondary structure. The total length of crRNA is 36 nt with an 8–10 nt stem, 4–6 nt A/U rich ring, a 5–10 nt 5’ terminal, and an AAAAC motif 3’ terminal [14]. Besides, most type VI-D orthologs contain a WYL domain, which enhances both the targeted and the collateral ssRNA cleavage activity in a dose-dependent manner [33,36]. In 2021, Xu et al. developed a computational pipeline to search for previously uncharacterized CRISPR-Cas13 systems from metagenomic datasets. Using CRISPR arrays as search anchors, they identified 340425 putative CRISPR repeat arrays and sequentially analyzed seven Cas13 variants based on conserved stem-loop structure and BLAST alignment similarity, and finally divided them into two groups, including Cas13X (Cas13X.1 and Cas13X.2) and Cas13Y (Cas13Y.1 to Cas13Y.5) [13,37]. The newly identified Cas13X shares some similarity with the previously identified Cas13 family effectors, but possesses a shorter and more compact structure [31]. The crystal structure analysis revealed that the Cas13X protein adopts a bilobed structure consisting of REC and NUC lobes, and the crRNA direct repeat could be anchored in REC leaves. The NUC lobe contains HEPN1 and HEPN2 domains, while the REC lobe is composed of Helical 1, Lid, and Helical 2 domains interleaved with each other. HEPN1 and HEPN2 regions are connected to the REC lobe by inter-domain linkers 1 (IDL1) and 2 (IDL2) [31]. However, currently, no detail structure information of Cas13Y is reported. Cas13bt has been identified as the most ultrasmall family of small Cas nucleases a [38]. The phylogenetic analysis suggested that Cas13bt proteins evolved from larger ancestral Cas13b proteins through multiple deletions. Its crRNA consists of the 5-nucleotide spacer (guide) segment and the 36-nucleotide direct repeat region. The direct repeat region includes stem 1, an internal loop, stem 2, and a hairpin loop. The electron density was less distinct for the spacer region, suggesting its flexibility in the Cas13bt3-crRNA binary complex structure [31].
The adaptive immune mechanism of type VI CRISPR bacteria is significantly different from that of other types of bacteria due to the differences in the types of foreign nucleic acids. It has been found that, similar to some type III CRISPR systems, Cas1 naturally fuses with reverse transcriptase to form a reverse transcriptase (RT)-Cas1 fusion complex in some type VI CRISPR systems [39,40]. RT associated with the RNA-targeting CRISPR-Cas13 system forms an integrase complex together with Cas1 and Cas2, facilitating the acquisition of RNA molecules, which are integrated into the CRISPR array as a new spacer (Figure 3) [41]. Cas1 proteins play a catalytic role in spacer acquisition from DNA [40]. Recently, it was found that the type VI-A system may include two types of RT-Cas. Type VI-A/RT1 and RT2 systems encompass either one or two CRISPR arrays, each containing two to six spacers of 30–49 nt in length. In type VI-A/RT1 systems, the direct repeats (37-nt) are 6/8-nt longer than those of type VI-A/RT2 systems (29/31-nt). The Cas13a protein sequences of the type VI-A/RT1 and VI-A/RT2 systems clustered separately into two distinct clades with other Cas13a sequences lacking RT-Cas1. Thus, the CRISPR arrays of type VI-A/RT1 and RT2 systems are characteristic of type VI-A systems. Therefore, their association with Cas13a preceded the acquisition of the RT-Cas1/Cas2 adaptation modules [39]. Notably, the RNA immune mechanism of other effectors of the Cas13 family including Cas13b, Cas13c, Cas13X, and Cas13Y are still unclear and need to be studied further.
Different subtypes of Cas13 proteins share a common mechanism when binding pre-crRNA and recognizing target RNA molecules (except for the type VI-X and VI-Y). The Cas13 protein recognizes and binds pre-crRNA to form an intermediate transition state. Functionally, most of the Cas13 effectors are crRNA-guided RNases with two distinct and independent catalytic centers, one of which directly processes pre-crRNAs and the other which cleaves ssRNA. Nakagawa et al. found that pre-crRNA treatment experiments showed that Cas13X did not process its pre-crRNA in vitro [31]. The second catalytic center that can cleave ssRNA possesses two R-X4-H motifs, which are typical nucleotide-binding domains in higher eukaryotes and prokaryotes. Generally, the conversion of pre-crRNA to crRNA is processed independently by Cas13 in a metal-independent manner (except for the types VI-D, VI-X, and VI-Y). In Cas13a, the conserved sites of the HEPN-2 domain and Helical-1 domain undergo conformational changes, which enable Cas13a to process pre-crRNA with RNase activity and form a mature Cas13a-crRNA complex. After crRNA maturation, the 5’ end of crRNAs of subtypes VI-A, VI-C, and VI-D contain palindromic repetition regions; however, the crRNA of the VI-B, VI-X, and VI-Y subtypes own the palindromic repetition region at the 3’ end of crRNA. In the second step, in Cas13a, the spacer sequences of target ssRNA and crRNA complement each other and result in a synergistic change to form Cas13a and crRNA complex, which makes the HEPN1 domain closer to the HEPN2 domain. These conformational changes extend the positively charged channel of the NUC lobe to combine crRNA, then unfold the R-X4-H motif of the HEPN1 domain for spatial closer access to the second R-X4-H motif, resulting in activating the HEPN nuclease site to cleave ssRNA in a nonspecific manner. Abudayyeh et al. found that the first nucleotide in the protospacer of 3’ flanks preferred A, U, or C (H) rather than G for LshCas13a, which was confirmed by in vitro RNA degradation experiments. G significantly reduced the activity of HEPN nuclease, but A, U, and C enhanced this activity [12,42]. Cas13b-mediated ssRNA cleaves are restricted by a double-side protospacer flanking sequence (PFS), such as a 5′ PFS of D (A, U, or G) and 3′ PFS of NAN or NNA [43]. However, the cleavage of the Cas13d targeted RNA appears to be independent of the original PFS [14]. Cas13d may bind pre-crRNA and cleave it into a 30 nt 5′ palindromic repeat region and a 20 nt 3′ spacer region [36]. Cas13d then recognizes the 5’ direct repeat region of crRNA with the aid of NTD and HEPN2 domains and clamps 2 nt in the head region of the direct repeat region, while the 3’ spacer region is trapped between the Helical-1 and Helical-2 domains. The HEPN1 domain acts as a hinge and provides a structural scaffold to connect the two functional lobes of Cas13d, which may form a positively charged solvent to unwind the RNA spacer. In this compact structure, Cas13d becomes a “surveillance complex” to search and identify complementary RNA target sites [33]. The combination of Cas13d and crRNA mainly occur at the 3’ end of the crRNA repeat region, which is crucial for the correct localization and binding of crRNA and Cas13d [14,33]. Furthermore, the sequence and structure of the crRNA repeat region play an important role in the cleavage activity of Cas13d, and the presence of Mg2+ and other auxiliary components, such as the WYL domain, increase its nuclease activity. Recent studies indicated that Mg2+ is a key factor that enhances the affinity between Cas13d and crRNA, which may stabilize the conformation of the Cas13d-crRNA repeat region. However, Mg2+ does not affect the mature processing of Cas13d on pre-crRNA [33]. EsCas13d and RspCas13d activate RNA cleavage by the WYL1 protein structural domain, indicating a common regulatory mechanism of Cas13d orthologs [36]. Cas13d has a relatively flexible structure, and its WYL1 domain can further stabilize Cas13d by connecting it with crRNA, while the binary Cas13d and crRNA complex recognizes ssRNA and activates Cas13d’s RNA cleavage activity [44]. Due to the activation by the target RNA, all domains of Cas13d further form new conformations and fully integrate with the sugar-phosphate backbone of the crRNA spacer region to form a Cas13d-crRNA-targeted ssRNA ternary complex [45]. The cleavage of the Cas13X targeted RNA is also independent of the original PFS [13]. First, the crRNA DR is recognized by the Helical-1, Lid, and Helical-2 domains. The crRNA is kinked at the DR-spacer junction with the spacer region surrounded by the HEPN1, Helical-1, Lid, and Helical-2 domains. Three nucleotides in the spacer interact with the HEPN1, Lid, and Helical-1 domains within the protein molecule. This structure implied that the DR-proximal region in the spacer cannot serve as a seed sequence that initiates base pairing with a target RNA [31]. Then, in the ternary complex, the crRNA DR is anchored within the REC lobe in the binary complex. The crRNA spacer base matches the target RNA, while the terminal five base pairs are disordered. The crRNA-target RNA duplex is bound to the groove formed by the Helical-1, Lid, HEPN1, and HEPN2 domains [31]. The crRNA-target RNA duplex is recognized by the Helical-1, Lid, HEPN1, and HEPN2 domains through interactions with its sugar-phosphate backbone [31].
Although different subtypes of Cas13 effectors have similar structure and function domains and share common molecular mechanisms, subtle changes in their structures can result in differences in binding pre-crRNA and recognizing target RNA. The structural and functional differences of different Cas13 effectors are listed as follows (Table 2). (1) Due to the difference in primary structure, the protein sizes of Cas13 effectors are significantly different. Among them, Cas13a is about 1250 aa, Cas13b is about 1150 aa, Cas13d is about 930 aa, Cas13Y is about 790 aa, and Cas13X is the smallest, with only about 775 aa. (2) During the process of pre-crRNA, different domains of Cas13a, Cas13b, Cas13d, Cas13X, and Cas13Y participate in crRNA maturation. In Cas13a, the Helical-1 is responsible for the process of pre-crRNA; in Cas13b, the RRI-2 domain is critical for the pre-crRNA process and produces a 66 nt mature crRNA with a 30 nt 5’ spacer and a 36 nt 3’ direct repeat; the compact Cas13d lacks those domains, and the HEPN-2 domain takes on that role. However, Cas13X is not involved in pre-crRNA processing, which is processed by an unknown enzyme [31]. (3) There are different PFS requirements for ssRNA cleavage. For Cas13a, the first nucleotide in the 5’ flanking of ssRNA showed a preference for A, U, or C instead of G. For Cas13b, the selection of ssRNA is diverse based on Cas13b effectors derived from different species. Cas13b effectors from Bergeyella Zoohelcum and Prevotella buccae prefer 5′ D (A, U, or G) and 3′ NAN or NNA of ssRNA; To date, no PFS sequence has been detected in any Cas13d and Cas13X orthologs, and Cas13Y has not been studied yet. (4) The crRNA structures are different for Cas13a, Cas13b, Cas13d, and Cas13X. For Cas13a, Cas13b, and Cas13d, their crRNA contains a 35–39, 36, or 36 nt direct repeat sequence and forms a conserved 5–6, 3–6, or 8–10 nt stem and a 7–9, 9–14, or 4–6 nt loop, respectively. For Cas13a, Cas13b, and Cas13d, their crRNA contains a 35–39, 36, or 36 nt direct repeat sequence and forms a conserved 5–6, 3–6, or 8–10 nt stem and 7–9, 9–14, or 4–6 nt loop, respectively. However, Cas13X differs slightly from them because the crRNA of Cas13X contains a direct 36 nt repeat sequence and forms stem structures consisting of two separate stems. Stem 1 includes four canonical Watson-Crick base pairs (G-C–G-C), and stem 2 contains a non-canonical G-U wobble base pair and five canonical base pairs (C-G–C-G and C-G), with U flipped out from the stem. The crRNA of Cas13X contains a 5 nt loop [31].
The VI CRISPR-Cas system has attracted extensive attention due to its advantages of high efficiency, high specificity, programmable RNA targeting, and autonomous pre-crRNA processing. Based on the characteristics of the Cas13 family, researchers developed websites for high-throughput crRNA design: https://gggenome.dbcls.jp/ (22 September 2022) and https://cas13design.nygenome.org/ (22 September 2022). Meanwhile, the RNA editing of Cas13 effectors and their trans-ssRNA cleavage activity accelerate their promising application, such as the clinic diagnosis of pathogens and the treatment of the disease (Table 3). Notably, the plasmid information of the Cas13 system used in the previous studies was summarized in Table 4.
Recently, with an in-depth understanding of the structure and function of the CRISPR-Cas13 family, the CRISPR-Cas13 system has been widely used in biological basic research fields, such as gene function, RNA interference, basic medicine, genetic development, etc. Mendez-mancilla et al. found that the chemical modification of crRNA in CRISPR-Cas13d can increase RNA knockdown efficiency in human cells, and demonstrated that the complex of chemically modified crRNA and Cas13d effectors can edit transcript RNA in human primary T cells [49]. Mahas et al. confirmed that CasRx can target a single plant virus or two plant RNA viruses simultaneously with high interference efficiency in Nicotiana Benthamiana [50]. Due to RNA-targeting characteristics of the CRISPR-Cas13 system may avoid permanent damage to organisms’ genome, Cas13 can be used as a reliable RNA interference tool for RNA manipulation, including the establishment of RNA editing platform, cell apoptosis and cancer therapy studies, and the manipulation of gene knockdown in animal disease models [45,94,95]. Kushawah et al. developed an efficient, specific, economical, and direct application platform based on CRISPR-RfxCas13d for gene function study during embryogenesis in an animal model with a 76% knockout efficiency [91]. Similarly, Buchman et al. knocked down endogenous gene expression and confirmed its feasibility using a RfxCas13d (CasRx) system in Drosophila melanogaster [96].
Nucleic acid detection and diagnosis technology is a measurement method integrating molecular biology and application techniques. The sensitivity, specificity, quick detection, and low cost are major considered issues for point-of-care testing (POCT). The adaptive immune system of CRISPR-Cas13 members in microorganisms can cleave RNA or trans-cleavage nucleic acids, which may be used for nucleic acid detection and diagnosis (Figure 4a) [47]. Gootenberg et al. designed a fast, cheap, and sensitive portable nucleic acid diagnosis platform, called SHERLOCK, based on the RNA trans-cleavage activity of Cas13 type VI CRISPR systems and strict crRNA complementary base pairing, and successfully applied it for nucleic acid detection and diagnosis [97]. Later, it was gradually applied to the studies of COVID-19, cancer mutations, and plant genetic traits improvement [48,51,52,53]. Currently, researchers are conducting a large-scale clinical study to evaluate SHERLOCK’s availability as a cheap and reliable diagnostic tool for COVID-19 [98,99]. Moreover, other nucleic acid diagnostic tools based on CRISPR-Cas13 have also been developed, including quantitative detection of microRNA and virus surveillance via microfluidic chip. Of note, combinatorial arrayed reactions for the multiplexed evaluation of nucleic acids (CARMEN) can simultaneously perform the high-throughput detection of different species, which may monitor the spread and evolution of infectious diseases [100]. Zhang et al. also developed a hepatitis virus detection technique based on RCA amplification combined with Cas13a, and successfully tested liver tissue samples of hepatitis-infected patients for virus infection with a minimum detection limit of 1 copy/microliter [55]. Due to the COVID-19 pandemic, our society urgently needs fast, labor-saving, and highly efficient diagnostic tools; SHERLOCK has been further simplified as a “streamlined highlighting of infections to navigate epidemics” (SHINE) diagnostic tool for detecting SARS-CoV-2 RNA from unextracted samples [54]. Furthermore, Cunningham et al. modified SHERLOCK and successfully detected plasmodium parasites from different clinical samples in the Democratic Republic of the Congo, Uganda, and Thailand, and achieved 73% sensitivity and 100% specificity [101].
The visualization of RNA localization and dynamics in living cells is an important technique for the in-depth study of RNA subcellular localization and function. Currently, single-molecule fluorescence in situ hybridization (smFISH) of RNA is the most commonly used method to observe the dynamics of intracellular RNA. However, CRISPR-Cas13 tagging fluorescent markers accelerate the development of RNA tracking and imaging in living cells (Figure 4b) [56,102]. In 2017, for the first time, the Cas13 family effector Cas13a was applied for intracellular nucleic acid imaging, in which dLwaCas13a-NF can image stress granule formation in living cells and realize the dynamic observation of intracellular transcripts [46]. The dPspCas13b and dPguCas13b fusing different fluorescent proteins successfully achieved quick and efficient tracking of target RNA [58,89]. Besides, dRfxCas13d combined with fluorescence-labeled crRNA may observe RNA transcription in living cells using CRISPR Live-cell fluorescent in situ hybridization (LiveFISH) [56]. Moreover, a new system based on dCas13a-SunTag-BiFC was developed by fusing dLwaCas13a and SunTag systems. In this system, dLwacas13a was used as a tracker targeting specific RNA, while SunTag recruits split Venus fluorescent proteins to label the targeted RNA [60].
Immunotherapy and precision medicine based on RNA editing become effective alternatives for antiviral therapies in animals and plants due to their specific target delivery and non-permanent genetic effect [103]. Previous studies indicated that ssRNA viruses infecting humans, vertebrates, and plants occupied approximately 51%, 44%, and 70% of all viruses, respectively, implying that the application of Cas13 targeting ssRNA is promising and has a broad market potential (Figure 4c) [50,103]. For antiviral therapy in mammals and plants, Freije et al. showed that the Cas13 effector can be programmed to target and destroy the genomes of multiple mammalian single-stranded RNA viruses, including lymphocytic choriomeningitis virus (LCMV), influenza A virus (IAV), and vesicular stomatitis virus (VSV) [104]. Moreover, the strategy of Cas13-mediated RNA virus interference is expected to be an effective tool for plant immunity against plant RNA viruses. Using the CRISPR/Cas13a system, Zhang et al. constructed an antiviral protection tool for monocotyledons, which can target and cleave viral RNA genomes to resist RNA viruses [63]. Abbott et al. developed prophylactic antiviral CRISPR in human cells (PAC-MAN) for simultaneously targeting multiple coronaviruses and influenza viruses via highly conserved regions of the virus genome [61]. Cui et al. developed an all-in-one expression system expressing Cas13b and double crRNA to meet the needs for synchronal delivery of Cas13b proteins and crRNA, which can eliminate PRRSV infection [62]. Because of the rapid spread of novel coronavirus, some researchers have proposed RNA virus treatment strategies based on CRISPR/Cas13 family effectors for coronavirus infection [82,105]. Ashraf et al. have preliminarily achieved the treatment of hepatitis C in mammalian cells [64]. Zhang et al. developed a powerful tool based on the CRISPR/Cas13d system to treat Seneca Valley virus (SVV). In this system, CasRx-de-NSL can be expressed in human cells, and SVV RNA can be cleaved and treated. The results showed that the experimental group reduced the replication ability of SVV by 57% [106]. As the smallest type VI nucleic acid effector, Cas13X is an excellent candidate for treating genotypic diseases. Yang et al. utilized the AAV-PHP.eB capsid and GFAP promoter to drive the specific expression of Cas13X-NLS-HA-sgPtbp1 in astrocytes and showed that the PTBP1 signal, a key signal, was decreased after two weeks, one month, and two months of intravenous treatment, respectively. The PTBP1 S signal decreases sequentially from 74.79% to 18.42%, 14.96%, and 11.98%, indicating that astrocytes can be successfully converted into neurons (AtN) [107].
CRISPR/Cas13-based gene therapy strategies can treat many diseases, such as dominant and recessive genetic diseases, cancer, cardiovascular diseases, and neurodegenerative diseases [108]. The CRISPR-Cas13 system may knock down intracellular RNA to block the occurrence of mutagenic and genetic diseases. Cas13d, the smallest Cas13 protein, has developed a clinic platform for RNA therapy (Figure 4d). To better apply the CRISPR/Cas13 system for the treatment of human and animal diseases, researchers have developed multiple plasmids and RNA delivery techniques, including the intravenous injection of Cas protein and crRNA, the stable delivery of AAV lentiviral vector, and PB transposon systems fused with CRISPR/Cas13 and crRNA [69]. Zhao et al. found that the subtly designed crRNA combined with Cas13a could inhibit the transcriptional expression level of the KRAS mutant in pancreatic cancer, and ultimately suppressed oncogene expression in mice, suggesting that the CRISPR/Cas13 system may be useful for tumor therapy by perturbing oncogenes at the transcriptional level [74]. He et al. proposed that the delivery of AAV (adenovirus) mediated CasRx and Pcsk9 sgRNA into the mice liver could successfully reduce serum Pcsk9 and cholesterol levels, which provided a permanent genetic therapy using Cas13 proteins [109]. Due to the continuous outbreak of SARS-CoV-2, researchers have also applied CRISPR/Cas13 for COVID-19 treatment [109,110,111]. Xiao et al. evaluated the therapeutic potential of a Cas13-derived RNA base editor to correct mutations in myosin VI (Myo6) transcripts that cause hearing loss in a mouse model [66]. In previous studies, oncogenes and toxic RNAs were knocked down by Cas13a under the control of a minimal promoter regulated by NF-κB transcription factor binding [112]. In addition, in the gene therapy study of Cas13a, the researchers induced EGFR overexpression by regulating the key oncogenic genes in glioma cells and caused the apoptosis of glioma cells [113]. Tian et al. used an alternative polyadenylation (APA) reporter to screen a set of dCas13 proteins, including Cas13a, Cas13b, and Cas13d, and showed that the CRISPR- dPguCas13b system had the most significant efficiency of APA manipulation [114].
Cas13 family proteins are also widely used for the mRNA process because they are currently found to be CRISPR effectors that specifically cleave RNA [115]. The first other application is to participate in alternative splicing (SA) and APA of mRNA maturation by the CRISPR-Cas13 system. Konermann et al. reported that CRISPR-dCas13d targeted splicing elements and engineered fusion into the Gly-rich C-terminal domain of hnRNPa1, one of the most abundant hnRNP families, which can successfully interfere with exon exclusion or inclusion in endogenous gene reporter systems. This demonstrated that the CRISPR-Cas13 system is a useful tool for the gene functional study of AS events [14]. Second, the CRISPR-Cas13 system can be applied for RNA modification, including programmable regulation of alternative splicing, A-to-I and C-to-U editing, and m6A modifications [89,116,117]. Third, the CRISPR-Cas13 system can be used for the design of cell fate. Previous studies utilized the single-base editing technology of CRISPR/dCas13 to convert C to U for nucleic acid mutations. Furthermore, STAT3 and β-catenin pathways were activated via the CRISPR-Cas13 system to accelerate the growth of HEK293FT cells and HUVECs [118]. In addition, it was reported that the demethylation of m6A in PTH1R mRNA by the photoactivatable RNA m6A editing system using CRISPR-dCas13 (PAMEC) in bone marrow mesenchymal stem cells could hinder osteogenic differentiation and reduce translation efficiency without changing RNA stability [119].
As we know, because RNA is one of the most important messengers in cellular processes, characterization of its function and dynamic metabolism is crucial for understanding life processes. The use of the RNA-targeted CRISPR-Cas13 system will contribute to the understanding of the RNA’s function. Cas13 protein is an effector with dual ribonuclease activities (including pre-crRNA processing and crRNA-mediated RNA cleavage), and has unique advantages in RNA-specific recognition, RNA cleavage, and trans-activating activity). It has a wide application prospect in biomedical basic research, animal disease models, genetic disease diagnosis, nucleic acid subcellular localization, and disease treatment. Presently, although the CRISPR-Cas13 family can avoid permanent damage to the genome of organisms because of RNA targeting, it should be considered for the toxic effects of RNA-targeted cleavage. Cas13X in type VI CRISPR systems possesses a small size (~775 aa) without a PFS limit, enabling it to be conveniently fused with multiple RNA arrays and improving its high-throughput application. Taken together, RNA editing targeted by the CRISPR-Cas13 family further expands the CRISPR toolkit for RNA manipulation and greatly enhances diagnostic capabilities. | true | true | true |
PMC9569903 | 36233204 | Alena A. Kozlova,Anastasia N. Vaganova,Roman N. Rodionov,Raul R. Gainetdinov,Nadine Bernhardt | Assessment of DDAH1 and DDAH2 Contributions to Psychiatric Disorders via In Silico Methods | 07-10-2022 | DDAH1,DDAH2,bipolar disorder,schizophrenia | The contribution of nitric oxide synthases (NOSs) to the pathophysiology of several neuropsychiatric disorders is recognized, but the role of their regulators, dimethylarginine dimethylaminohydrolases (DDAHs), is less understood. This study’s objective was to estimate DDAH1 and DDAH2 associations with biological processes implicated in major psychiatric disorders using publicly accessible expression databases. Since co-expressed genes are more likely to be involved in the same biologic processes, we investigated co-expression patterns with DDAH1 and DDAH2 in the dorsolateral prefrontal cortex in psychiatric patients and control subjects. There were no significant differences in DDAH1 and DDAH2 expression levels in schizophrenia or bipolar disorder patients compared to controls. Meanwhile, the data suggest that in patients, DDAH1 and DDHA2 undergo a functional shift mirrored in changes in co-expressed gene patterns. This disarrangement appears in the loss of expression level correlations between DDAH1 or DDAH2 and genes associated with psychiatric disorders and reduced functional similarity of DDAH1 or DDAH2 co-expressed genes in the patient groups. Our findings evidence the possible involvement of DDAH1 and DDAH2 in neuropsychiatric disorder development, but the underlying mechanisms need experimental validation. | Assessment of DDAH1 and DDAH2 Contributions to Psychiatric Disorders via In Silico Methods
The contribution of nitric oxide synthases (NOSs) to the pathophysiology of several neuropsychiatric disorders is recognized, but the role of their regulators, dimethylarginine dimethylaminohydrolases (DDAHs), is less understood. This study’s objective was to estimate DDAH1 and DDAH2 associations with biological processes implicated in major psychiatric disorders using publicly accessible expression databases. Since co-expressed genes are more likely to be involved in the same biologic processes, we investigated co-expression patterns with DDAH1 and DDAH2 in the dorsolateral prefrontal cortex in psychiatric patients and control subjects. There were no significant differences in DDAH1 and DDAH2 expression levels in schizophrenia or bipolar disorder patients compared to controls. Meanwhile, the data suggest that in patients, DDAH1 and DDHA2 undergo a functional shift mirrored in changes in co-expressed gene patterns. This disarrangement appears in the loss of expression level correlations between DDAH1 or DDAH2 and genes associated with psychiatric disorders and reduced functional similarity of DDAH1 or DDAH2 co-expressed genes in the patient groups. Our findings evidence the possible involvement of DDAH1 and DDAH2 in neuropsychiatric disorder development, but the underlying mechanisms need experimental validation.
Mental disorders remain among the top ten leading causes of burden worldwide, which makes research to establish causal pathways imperative for effective prevention and treatment [1]. In recent years, nitric oxide (NO) signaling has been implicated in the pathophysiology of several mental illnesses, such as schizophrenia and affective disorders, comprising bipolar disorder and major depressive disorder [2]. For example, in schizophrenia patients, NO metabolism is impaired in various organs, including the brain [3,4,5], and high NO levels are found in post-mortem samples of the prefrontal cortex and hippocampus [6]. NO is a gaseous molecule acting as the second messenger of the NMDA receptor, thereby regulating glutamatergic transmission [7]. NO further interacts with the dopaminergic and serotonergic systems [8] and is involved in the storage, uptake, and release of transmitters, such as acetylcholine, noradrenaline, GABA, taurine, and glycine [9]. In the brain, NO regulates synaptic plasticity, neurodevelopment, and cerebral blood flow [10]. Meanwhile, excessive amounts of free radical NO lead to neurotoxicity and neurodegeneration [10,11]. There are three isoforms of enzymes generating NO: the neuronal (nNOS or NOS1), the inducible (iNOS or NOS2), and the endothelial NO synthases (eNOS or NOS3) [9]. Genetic variations in NOS genes have been associated with several psychiatric conditions: (i) NOS1 polymorphisms are associated with an increased risk of schizophrenia development [12]; (ii) NOS1 and NOS3 alleles are involved in modifying an individual’s susceptibility to bipolar disorder, depression, or risk of suicide attempts, and impact glutamatergic neurotransmission [13,14,15,16,17,18,19]; (iii) some NOS3 variants demonstrate a protective role in bipolar disorder [18]; (iv) NOS2 involvement in psychiatric diseases was demonstrated in animal models [1] and population studies [20], but the knowledge of this association is limited. In addition, the NOS1 gene is methylated differently in schizophrenic patients and healthy individuals [21]. NOS1 coupling to the NMDA receptor is regulated by the NOS1 adapter protein (NOS1AP) [13], whose polymorphisms are associated with schizophrenia [22] and the severity of posttraumatic stress disorder [23]. Thus, there is increasing evidence that NOS1 and NOS3 are promising drug targets for treating schizophrenia and affective disorders [24,25,26]. NOS activity and NO levels are regulated by dimethylarginine dimethylaminohydrolases (DDAHs). There are two known isoforms whose amino acid sequences are 50% identical. DDAH1 is responsible for the degradation of N (omega), N (omega) dimethyl-L-arginine (ADMA), the major competitive inhibitor of NOS [27]. DDAH2 also contributes to the regulation of NO levels, although it is still being debated if through ADMA degradation or ADMA-independent mechanisms. Both isoforms of DDAHs are expressed in the brain in a regional and cell-type complementary fashion [28]. DDAH1 alleles are associated with the risk of developing autistic spectrum disorder or obsessive-compulsive disorder [29], and animal models of autism and schizophrenia endophenotypes present with increased DDAH1 levels [30,31]. At the same time, experiments in mice fed an Mg-restricted diet resulted in depression-like behavior and decreased DDAH1 expression [32]. However, the data on DDAH1 expression levels in different brain structures and psychiatric conditions are sparse. Thus far, reduced DDAH1 expression has been found in the anterior cingulate cortex in schizophrenic patients [33], whereas in the prefrontal cortex, a downregulation was transient and detectable only in the first years after the onset of the disease [34]. In addition, a strong downregulation of the hsa-miR-219-5p that is suggested to control DDAH1 expression was observed in schizophrenia patients [35]. DDAH2 gene variants are associated with schizophrenia and bipolar disorder susceptibility [36]. In schizophrenia patients, the DDAH2 gene is aberrantly methylated in both the prefrontal cortex and blood, and DDAH2 brain mRNA levels are significantly increased [37]. Further, loss of methylation was shown in schizophrenia patients with suicide attempts [38]. Loss or overexpression of NOS or other genes directly involved in this pathway may define altered NO-signaling in psychiatric disorders. The impact of the spatial expression pattern and the level of co-expressed genes may also be relevant. For example, malformation of NOS1 positive GABAergic interneurons was described in schizophrenia [39]. In addition, the mRNA for NPY, expressed by many NOS1/NADPH-d GABA-ergic neurons, is selectively decreased in neurons within the superficial white matter of subjects with psychosis. While NO facilitates blood flow through the cortical microvasculature, NPY mediates microvessel constriction; therefore, its deregulation leads to microcirculatory dysfunction [40]. Thus, associations of NOS gene mutations and expression deregulation with psychiatric disorders are well studied. However, relatively less is known about the contribution of other modulators of NO-mediated processes in the pathogenesis of these diseases. Suggesting that the context-dependent consequences of NO-signaling may differ in the cortex of psychiatric and non-psychiatric subjects, we attempted to estimate the differences of DDAHs co-expressed gene patterns in publicly available datasets. We performed a careful functional analysis of these co-expressed gene sets. Specifically, we evaluated evidence of DDAH1 and DDAH2 involvement in regulating processes associated with major psychotic disorders, schizophrenia, and bipolar disorder.
DDAH1 and DDAH2 mRNA were identified in all dorsolateral prefrontal cortex samples in the selected datasets (refer to Table 1) in both patients and non-psychiatric controls. The DDAH1 expression levels were greater than DDAH2. However, no significant differences in the DDAH1 and DDAH2 expression levels were identified when their expression was compared between the control group and patients with either bipolar affective disorder or schizophrenia (refer to Figure 1a). Slight upregulation of DDAH2 was identified in the tissue samples from patients with bipolar disorder compared to the control group in the GSE112523 dataset. However, this finding did not remain statistically significant after the adjustment (Padj > 0.05). The DDAH1 and DDAH2 expression values are congruent in different datasets. DDAH2 expression is lower than DDAH1; however, its estimation is more prone to fluctuations and bias, particularly in the dataset GSE87194, where expression is lower than the other datasets. The dataset GSE112523 combines data from subjects with schizophrenia, bipolar affective disorder, and non-psychiatric controls; thus, it was further analyzed to compare DDAH1 and DDAH2 co-expressed gene sets in these psychiatric disorders. The dataset contains dorsolateral prefrontal cortex samples (mainly BA46 area) of seven patients with schizophrenia, ten patients with bipolar disorders, and seventeen non-psychiatric control subjects. Study group data are summarized in Table 2.
The distribution of the Pearson correlation coefficient demonstrates that the median values for all three study groups fluctuate near the zero level. In addition, a predominance of positive correlation coefficients was observed for both DDAH1 and DAAH2 in the schizophrenia group (refer to Figure 1b).
Suggesting the semantic similarity score between two genes mirrors the functional linkage of these genes, the most DDAH1 and DAAH2 co-expressed genes (r > 0.8, p > 0.05) were selected for the comparative semantic similarity analysis of GO biological process terms, with which these genes were annotated. Genes selected based on the correlation levels were used as clusters for further analysis. Both DDAH1 and DDAH2 co-expressed genes in control subjects have higher (p < 0.001) functional relationships compared to bipolar disorder or schizophrenic patients (refer to Figure 1c). As a number of direct and indirect protein–protein interactions of DDAH1 and DDAH2 are identified and represented in public databases such as STRING [42], BioGRID [43], MINT [44], and HPRT [45,46], we compared our co-expression pattern with the data deposited in these resources. This approach also allows us to more precisely identify the genes, which may be functionally linked with DDAHs in our co-expressed gene sets and to select them for further analysis. We analyzed the protein–protein interaction databases to select all genes for which interactions with DDAH1 or DDAH2 were previously identified. From now on, we refer to these genes collectively as the “DDAH1 cluster” and “DDAH2 cluster”, respectively (see Material and Methods for the sources and cluster formation and Supplementary S1, S2 for the lists of genes included in these clusters). To designate the genes of the “DDAH1 cluster” and “DDAH2 cluster” that are co-expressed with DDAH1 or DDAH2, respectively, in the control group, bipolar disorder patients and schizophrenia patients, we compared these clusters with the gene sets derived from our co-expression analysis (cut-off r > 0.3, p < 0.05). All study groups had low overlap between the “DDAH1 cluster” or “DDAH2 cluster” and sets of genes co-expressed with DDAH1 or DDAH2, respectively. However, Venn diagrams show that DDAH1 or DDAH2 co-expressed gene sets in each group include some genes involved in DDAHs-related biological processes. The greatest overlap between the co-expressed gene set and the “DDAH1 cluster” was identified in patients with bipolar affective disorder. Forty-five common genes were identified between these two gene sets (refer to Figure 2a). In contrast, the DDAH2 co-expressed gene set in controls includes ten genes from the “DDAH2 cluster”, whereas in the samples from patients with bipolar affective disorder and schizophrenia, the overlap was even lower (refer to Figure 2b).
Despite the small number of genes included in the “DDAH1 cluster” or “DDAH2 cluster” and co-expressed with DDAH1 or DDAH2, respectively, in the tissue samples studied in the analyzed dataset, we performed GO term enrichment analysis in these narrow gene subsets to explain their specific biological function. We found that the large group of genes for which involvement in DDAH1-related functions was previously identified (n = 45) is co-expressed with DDAH1 in the bipolar affective disorder group stochastically, and no significant GO term enrichment results were revealed in this gene set. In contrast, several GO groups were enriched in the constricted clusters of DDAH functionally associated genes, co-expressing with DDAH1 in schizophrenic patients and controls or with DDAH2 in all study groups. In the group of schizophrenia patients, DDAH1 co-expressed genes of the “DDAH1-cluster” (n = 5) are found to associate with protein localization and amino-acid metabolism and transport (refer to Figure 2a”, Supplementary data S3, Figure S1A). In contrast, in the control group (n = 4), the terms describing exocytosis and cell response were predominant (refer to Figure 2a’, Supplementary data S3, Figure S1B). DDAH2 co-expressed and functionally linked gene cluster enrichment results were congruent in different groups, with specific features in all cases. The most enriched GO terms in the control group (n = 10) and patients (n = 9 in patients with bipolar disorder and n = 3 in schizophrenic patients) describe the response to hypoxic conditions (refer to Figure 2b’–b’”, Supplementary data S3, Figure S2A–C).
To uncover whether the DDAH1 and DDAH2 co-expressed genes are regulated by common transcription factors in patients and non-psychiatric controls, we compared the enriched transcription factors-binding motives in genes co-expressed with DDAH1 and DDAH2 (i.e., r > 0.3, p < 0.05), respectively, in the different study groups. In healthy subjects, genes whose promoters contain short motif CG, which is recognized by zinc finger-CxxC proteins, are enriched in the DDAH1 co-expressed gene set (refer to Table 3, Supplementary data S4, Table S4.1). This association is completely lost in both the schizophrenia and bipolar disorder group. In the control groups’ transcriptomic data, genes whose promoters contain the AGGGGGA motif, which is recognized by several C2H2 zinc finger transcription factors, are enriched in the DDAH2 co-expressed gene set (refer to Table 3, Supplementary Sdata 4, Table S4.2). In contrast, in patients with bipolar affective disorder, several types of promoters, including the promoters C2H2 zinc finger transcription factors binding sites, are over-represented in the DDAH2 co-expressed gene set (refer to Table 3, Supplementary data S4, Table S4.3). In the meantime, we did not observe any over-represented motives in the promoters of DDAH2 co-expressed genes in patients with schizophrenia.
For disease ontology terms, 74 terms that predominantly characterize gene involvement in neoplastic disease (benign tumors and cancer) were enriched in DDAH1 co-expressed genes in the control group (refer to Supplementary data S5, Table S5.1). In genes co-expressed with DDAH1 in samples from schizophrenia patients, five terms corresponding to non-cancerous disease were enriched (refer to Supplementary data S5, Table S5.2). No DO terms were enriched in genes co-expressed with DDAH1 in the patients with bipolar affective disorder. In the context of the mental health terms, only one term, “DOID:0060037: a developmental disorder of mental health” (refer to Figure 3a), is significantly over-represented in the set of genes co-expressed with DDAH1 in the control group. Genes that regulate membrane potential, axonogenesis, and cell junction assembly contribute most to the enrichment result for this term (i.e., enrichment core; refer to Figure 3a’). No other associations with mental disorders were identified in DDAH1 co-expressed genes in any study group. Over one hundred DO terms were enriched in DDAH2 co-expressed gene groups in samples from patients with bipolar disorder and non-psychiatric subjects (refer to Supplementary data S5, Tables S5.3 and S5.4). Conversely, in genes co-expressed with DDAH2 in samples from schizophrenic patients, only eighteen terms are enriched (refer to Supplementary data S5, Table S5.5). As in the DDAH1 co-expressed gene set, genes of the term “DOID:0060037: a developmental disorder of mental health” are enriched in the control group. In addition, the genes corresponding to the term “DOID:0060041: autism spectrum disorder”/“DOID:12849:autistic disorder” are significantly over-represented in this set (refer to Figure 3b,b”). The functional characteristics of the enrichment core of “DOID:0060037: a developmental disorder of mental health” in DDAH2 co-expressed genes in non-psychiatric subjects differ slightly from DDAH1 co-expressed genes. The enrichment core genes are associated with synapse organization and cognitive functions such as learning, memory, and cognition (refer to Figure 3b’). Curiously, the top three over-represented functions in “DOID0060041:autism spectrum disorder”/“DOID:12849 autistic disorder” in the enrichment core are the same (refer to Figure 3b’”).
This study found DDAH1 and DDAH2 expression in all dorsolateral prefrontal cortex samples from patients and non-psychiatric control subjects. The expression level of DDAH1 was considerably higher than the DDAH2 expression levels in all subjects. We did not observe changes in DDAH1 expression levels here. In line, normal DDAH1 expression levels have previously been shown in chronic schizophrenic patients, while DDAH1 upregulation was found in patients with short-term schizophrenia [33]. For DDAH2 expression, upregulation is reported in association with schizophrenia patient-specific methylations and promoter region SNPs [36]. Other studies have also shown that DDAH2 mRNA levels were significantly elevated in brain tissue in schizophrenia, although the brain region was not specified in this study [37]. DDAH2 expression upregulation has also been described in the prefrontal cortex of patients with bipolar disorder [47]. While a similar trend was found in the present study, it does not reach statistical significance. The discrepancy between our results and published data may be attributed to the differences in study groups and applied methods. Affective disorders such as major depression or bipolar disorder are associated with an aberrant expression pattern of NOS in the dorsolateral prefrontal cortex. Deregulation does, however, not influence the expression level of NOS1-, NOS2-, or NOS3-mRNA in whole cortex samples, but it was accompanied by changes in protein localization in cortical layers [48]. Thus, NO deregulation in the brain of psychiatric patients does not solely depend on gene expression levels. It also may be associated with the disturbance of expression patterns in the complex multicellular cortex structure and disease-associated deregulation of biological processes in the cortex. Considering that the gene’s co-expression mirrors the similar biologic function of these genes [49], we attempted to compare DDAH1 and DDAH2 co-expression patterns in control subjects and patients with psychiatric disorders. The Pearson correlation coefficient is useful for estimating gene co-expression [50], revealing that subsets of genes for which co-expression with DDAH1 or DDAH2 is predicted in different groups hardly overlap. As the GO semantic similarity score between genes is related to the involvement of their products in the protein–protein interaction network [51], the functional relationship between DDAH1 or DDAH2 co-expressed genes appears to be higher in non-psychiatric controls. This difference may be related to the deregulation of DDAH-associated processes. Further analysis of DDAH co-expressed genes may uphold this assumption. To increase the stringency of the selection of DDAHs interacting partners in each study group, we selected genes with the correlation coefficient (r) > 0.3, p < 0.05, which are suggested to interact with DDAHs. The overlap between sets of genes that are co-expressed with DDAHs in different conditions and “DDAH1 cluster” or “DDAH2 cluster” was low. Meanwhile, the occurrence of low overlap between an experimentally identified gene co-expression pattern and protein–protein interaction data described in the literature or databases was observed and discussed in previous studies [52,53]. In the non-psychiatric control subjects, these genes are involved in vesicular transport in both directions. In contrast, in schizophrenic patients, this association is lost, and genes associated with protein localization and amino acid metabolism predominate. Enrichment of any biological process in DDAH1-interacting genes co-expressed with DDAH1 is completely lost in patients with bipolar disorder patients. The significance of NO-signaling for normal transcytosis functioning, i.e., vesicular traffic across the interior of cells in the blood–brain barrier, has been demonstrated [54]. In addition, NO diffusion stimulates the release of vesicles in the synaptic cleft [55]. Thus, the association of DDAH1 co-regulated genes with vesicle transport seems reasonable. Psychiatric disorders such as schizophrenia and bipolar disorder may be associated with brain-blood barrier dysfunction [56]. Disruption of eNOS is one of the suggested reasons for increased blood–brain barrier permeability [57]. Losing the association of DDAH1 co-expressed genes with the vesicular traffic (exocytosis and endosome trafficking) may mirror this or any other aspect of NO-dependent pathway disruption in the prefrontal cortex in psychiatric patients. The association of DDAH2 co-expressed genes with the response to the hypoxic stress identified in this study is expected in light of considerable evidence of DDAH2 upregulation in hypoxic conditions. The growth of DDAH2 expression levels in response to hypoxia was described in monocytes [58,59], endothelium [60], and myotubes [61]. In the studied group, DDAH2 co-expressed genes in the prefrontal cortex are stably associated with the response to hypoxia, despite the psychiatric diagnosis. Deregulation of DDAH1 and DDAH2-associated processes in psychiatric patients was also confirmed in the analysis of protein-binding motif enrichment in their promoter regions. In schizophrenic patients, the over-representation of any specific promoter motive is completely lost, both in DDAH1 and DDAH2 co-expressed genes. In patients with bipolar disorder, the enrichment is also lost in the DDAH1 co-expressed gene set. While, in genes co-expressed with DDAH2 in the prefrontal cortex of patients with bipolar disorder, several protein-binding patterns are significantly over-represented. Notably, AP-2 transcription factors play essential roles in sleep regulation in the nematode Caenorhabditis elegans and the fruit fly Drosophila melanogaster [62]. The AP-2 paralogous transcription factors Tfap2a and Tfap2b control sleep behavior in mice, allowing for bidirectional control of sleep quality [63]. Over-representation may thus link prefrontal DDAH2 functionality with sleep disturbance, a core symptom of bipolar disorder [64]. Genes co-expressed with DDAH1 in non-psychiatric subjects frequently harbor the CG motif in their promoters. Human CxxC-binding domains display different structures and selectivity [65]. The CG pattern is the sole DNA-binding domain of CGBP, which is implicated in the expression of genes associated with CpG islands and the regulation of cytosine methylation [66]. ShinyGO software also demonstrated the enrichment of TET1 CxxC-binding protein, which binds predominantly on CGCGAT motifs [65], whose role in the expression regulation is also dualistic. However, TET1 binds and represses CpG-rich promoters by interacting with the polycomb repressive complex 2 [67]. In the brain structure, it is involved in regulating synapse development and functioning, memory, neuronal death and repair, and neuro-glial communication. The lost association of DDAH1 expression with CG promoter motif genes may mirror the overexpression of TET1 in cortical structures in patients with schizophrenia or bipolar disorder [68]. As the co-expressed genes are suggested to share their function, the demonstrated loss of correlation between expression of DDAH1 and genes harboring CG patterns in their promoters may distort the DDAH1 role in memory, learning, and neuron functioning. Still, further experimental work is needed on this matter. The C2H2 zinc finger MZF-1 binding pattern is over-represented in DDAH2 co-expressed genes in non-psychiatric subjects and patients with bipolar disorder, but in schizophrenic patients, this association is lost. The transcription factor MZF-1 is a known tumor suppressor [69]. A significant portion of genes specifically expressed in the cortex, hindbrain, and midbrain harbor MZF-1 binding sites in their promoters [70], but the significance of MZF-1 expression in the central nervous system remains not well understood. In oxygen–glucose deprivation conditions, MZF-1 mediates the protective effect of human umbilical cord blood cells on both neurons and oligodendrocytes in mixed cultures [71,72]. This transcription factor also seems involved in gene regulation after peripheral nerve injury [73]. However, in contrast to non-psychiatric controls and schizophrenia, many other enriched promoter motifs in genes co-expressed with DDAH2 are found in patients with bipolar disorder samples. Of note, neither TET1 binds the DDAH1 promoter, nor does MZF1 regulate DDAH2 expression directly, as summarized in the Chip Seq Atlas [74] or Signaling Pathways Project [75]. Thus, the DDAHs co-expression with TET1- or MZF1-regulated genes needs an explanation, which is more complex than the co-regulation of the same transcription factors motivating further research in this direction. Disease Ontology was designed for researchers to study gene–disease relationships [76]. It has a hierarchical structure [77]; thus, the “DOID:0060037: developmental disorder of mental health” term covers the term “DOID0060041: autism spectrum disorder” and other terms corresponding to a learning disability, intellectual disability, attention deficit hyperactivity disorder, communication disorder, eating disorder and some other specific developmental disorders. The co-expression of DDAH1 and DDAH2 with the genes annotated with these terms was identified in dorsolateral prefrontal cortex samples in non-psychiatric controls but was lost both in schizophrenic and bipolar disorder patients. Developmental mental disorders, including autism spectrum disorders, demonstrate behavior and cognitive disabilities, which are also intrinsic to major psychoses [78]. At the same time, schizophrenia and bipolar disorder patients exhibit a developmental lag [79], and the deregulation of genes involved in neurogenesis and neurodifferentiation was identified in schizophrenia patients at disease onset [34]. The genetic and molecular backgrounds of these diseases share numerous similarities. Genes with a documented association with neurodevelopmental and neuropsychiatric disorders are predominantly involved in transcription, synaptic transmission, cell–cell communication, ion transmembrane transport, intracellular signaling pathways, cell cycle, metabolic processes, nervous system development, and neuron death [78,80,81,82,83]. The common genetic etiology mirrors the high comorbidity of these psychiatric diseases and developmental mental disorders [84]. Hence, in the latest version of the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), schizophrenia has been listed in proximity to neurodevelopmental disorders [85]. Currently, schizophrenia and bipolar disorder may be considered neurodevelopmental disorders with a widening of the neurodevelopmental spectrum [86] the loss of co-expression between genes characterized by the term “DOID:0060037: a developmental disorder of mental health”, including genes involved in synaptogenesis, axonogenesis and cognitive functions, and DDAHs is of particular interest. Thus far, most of the evidence on DDAH/ADMA axis relevance for neuropsychiatric disorders comes from case-control clinical studies and measurements of peripheral ADMA and NO levels without offering in-depth mechanistic insight. Nevertheless, connections between the DDAH/ADMA axis and oxidative stress markers, as well as molecules important for cognitive processes, have been shown and support the findings of this study. For example, plasma levels of the oxidative stress-induced lipid peroxidation product 4-HNE were increased and correlated positively with plasma ADMA levels in depression [87] and schizophrenia patients [88]. In terms of cognition, a correlation between plasma ADMA levels and cognitive deficits has been established, with decreases in ADMA levels leading to improvements in working memory and attention [89,90]. ADMA infusion decreases BDNF, a factor highly associated with cognitive functions [91]. Lastly, the G allele of DDAH2 (−449 G/C) was positively associated with leukoaraiosis and high ADMA levels [92]. Progression of leukoaraiosis, a condition frequently met in neuropsychiatric disorders, relates to cognitive decline and thus could explain the link of DDAH2-cluster with cognition (learning and memory, cognition, synapse organization). In addition, ADMA levels in patients with leukoaraiosis were significantly higher than those in healthy controls [93], and these high concentrations of ADMA were associated with cognitive dysfunction in leukoaraiosis patients [94]. Our findings must be seen with some limitations. (i) The transcriptomic datasets are generated with different sequencing depths. Although all measures were taken to normalize the data, full uniformity and overcoming batch effects are unattainable. (ii) Only a few datasets in the GEO [95] repository were relevant for the study. Further study groups are relatively small and heterogeneous, and patient information is restricted. (iii) The mRNA abundance has a limited capacity as the indicator of downstream expression. The gene expression level and activity of gene products depend on multiple factors, including RNA stability, modifications, the translation rate and protein turnover, localization in the cell, and availability of ligands or co-interacting proteins. (iv) In addition, the size of gene clusters selected for the GO enrichment analysis was small and sometimes appeared less than the clusters used for the enrichment analysis. However, considering that the enrichment results are more informative than raw data about the GO groups in which the identified genes are included, we include GO enrichment results in this paper. Despite this limitation, we received statistically significant results, with gene ratios of 0.5–1 for several GO groups. (v) The used protein–protein interactions databases STRING [42], BioGRID [43], and HPRT [45,46] may be misleading for DDAH2 co-expressed genes because they are based on text mining and the assumption that DDAH2 is purely metabolizing ADMA, which is still debated. (vi) The analyzed data were restricted to the dorsolateral prefrontal cortical area and therefore do not allow for generalization to other brain regions. (v) Finally, our study lacks experimental validation. Despite all these limitations, the present work provides preliminary evidence that DDAH1 and DDAH2 are co-regulated with genes involved in mental disorder development and derangement. Experimental studies are now needed for confirmation and to gain mechanistic insight.
The expression data were derived from the public database of Gene Expression Omnibus (GEO) [95]. We used the terms “schizophrenia” and “bipolar disorder”; the filter “Expression profiling by high throughput sequencing” for the series type; and the filter “Homo sapiens” for the organism. Datasets that comprise the prefrontal cortex samples were selected (refer to Table 1). Unfortunately, the data for other brain structures were unrepresented or represented by a single dataset, making it impossible to compare them, and we did not include them. The dataset GSE112523 combining data for the subjects with schizophrenia, bipolar disorder, and non-psychiatric controls was selected for further analysis of DDAH1 and DDAH2 co-expressed genes. This dataset was generated by the 75 bp paired-end sequencing, which was performed on an Illumina NextSeq 500 sequencer [41]. The genes of DDAH1- and DDAH2-interacting proteins (“DDAH1 cluster” and “DDAH2 cluster”, refer to Supplementary data S1, S2) for the comparative analysis were selected from the public databases STRING [42], MINT [44], BioGRID [43], and HPRD [45,46]. STRING database was searched for the human data in the full STRING network (i.e., for both functional and physical protein associations), and filtered for the data received by the text mining, experiments, databases, co-expression, or co-occurrence by the basic settings options. BioGRID data were filtered for the interactions identified in human studies. All other databases were searched with default settings. Then, the genes of all identified proteins were included in the “DDAH1 cluster” or “DDAH2 cluster”.
Raw counts were count per million (CPM)-normalized by edgeR package [96]. CPM values above the threshold level 1 were considered positive. The distribution of CPM-normalized expression levels in the analyzed samples was visualized by the beeswarm R package. To estimate the differential gene expression, raw counts were normalized using the Trimmed Mean of M-values (TMM) method by the edgeR package [96] to avoid batch effects. Differentially expressed genes were identified by the glmQLFTest test using the edgeR package [96]. p values were adjusted for multiple testing corrections using the Benjamini–Hochberg method. Genes were considered differentially expressed if adjusted p values (Padj) < 0.05.
Before co-expression measurement, CPM-normalized data were filtered to exclude genes that are expressed below the threshold (i.e., CPM = 1 for all samples in the study groups). Data for different study groups were filtered independently. DDAH1 and DDAH2 co-expressed genes were selected by Pearson’s correlation coefficient (r > 0.3, p < 0.05). Genes co-expressed with DDAH1 or DDAH2 in the different study groups were included in separate gene clusters (i.e., DAAH1-co-expressed genes in the control group, DAAH1-co-expressed genes in schizophrenic patients, DAAH1-co-expressed genes in patients with bipolar disorder, DAAH2-co-expressed genes in the control group, DAAH2-co-expressed genes in schizophrenic patients, and DAAH2-co-expressed genes in patients with bipolar disorder). The comparative analysis of the selected clusters was performed as described below.
Gene Ontology (GO) semantic similarity was calculated by Wang’s method in the GOSemSim package [97] employing the “Biological process” GO terms. The difference in semantic similarity scores in different gene clusters was estimated by Brown–Forsythe and Games–Howell post hoc tests.
The clusters of DDAHs co-expressed genes, identified in different study groups, were compared to each other, and to the “DDAH1 cluster” or “DAAH2 cluster”, which were identified by searching public databases as described above. DDAH1 co-expressed gene clusters were compared in the control group, schizophrenic patients, and patients with bipolar disorder, and then matched with the “DDAH1 cluster”. Similarly, the DDAH2 co-expressed gene clusters were evaluated. The overlap between identified gene clusters was visualized by the VennDiagram R package. To increase the stringency of DDAHs interacting patterns, the genes that were common for DDAHs co-expressed groups and the “DDAH1 cluster” or “DAAH2 cluster” were selected for further analysis, as it was described elsewhere [52,98]. Enriched motifs in promoters of the studied gene sets were identified using the ShinyGo 0.76 [99] web tool (available at http://bioinformatics.sdstate.edu/go/, accessed 10 June 2022). The upstream 300 bp region was specified as the promoter. GO enrichment analysis (identification of GO terms that are significantly enriched by the genes of the selected set) was performed in the identified co-expressed gene clusters, and visualization of results was performed by the clusterProfiler Bioconductor package [100]. Disease Ontology (DO) enrichment analysis was performed in gene lists ranging from the highest value of the Pearson correlation of the gene expression level with the levels of DDAH1 or DDAH2 expression to the lowest. The clusterProfiler [100] and DOSE [101] R packages were used. We considered significant enrichment results only for GO biological process terms, transcription factors, or DO terms with a false discovery rate value of <0.05.
Our results suggest a possible involvement of DDAH1 and DDAH2 in the pathophysiology of psychiatric disorders. While mRNA levels in the dorsolateral prefrontal cortex of psychiatric patients remain unchanged, a functional shift occurs that is reflected in dramatic changes in the expression of genes whose products interact with DDAH1/2. We found that correlations between expression levels of DDAH1 or DDAH2 and genes associated with mental illness are lost in cortical samples from psychiatric patients. DDAH1 and DDAH2 co-expressed genes were generally less integrated into shared functions in psychiatric patients than in non-psychiatric controls. Furthermore, the overlap between genes co-expressed with DDAHs in control subjects and psychiatric patients is low, possibly due to the complex deregulation of transcription factor activity. These data suggest that DDAHs are associated with the processes that form the molecular basis of mental and cognitive functions and thus may be potential therapeutic targets in psychiatric disorders. | true | true | true |
PMC9569910 | 36232391 | Andrej Minich,Veronika Lišková,Ľubica Kormanová,Ján Krahulec,Júlia Šarkanová,Mária Mikulášová,Zdenko Levarski,Stanislav Stuchlík | Role of RNAIII in Resistance to Antibiotics and Antimicrobial Agents in Staphylococcus epidermidis Biofilms | 21-09-2022 | biofilm,oxacillin,phenol soluble modulins,vanillin | Staphylococcus epidermidis is a known opportunistic pathogen and is one of the leading causes of chronic biofilm-associated infections. Biofilm formation is considered as a main strategy to resist antibiotic treatment and help bacteria escape from the human immune system. Understanding the complex mechanisms in biofilm formation can help find new ways to treat resistant strains and lower the prevalence of nosocomial infections. In order to examine the role of RNAIII regulated by the agr quorum sensing system and to what extent it influences biofilm resistance to antimicrobial agents, deletion mutant S. epidermidis RP62a-ΔRNAIII deficient in repressor domains with a re-maining functional hld gene was created. A deletion strain was used to examine the influence of oxacillin in combination with vanillin on biofilm resistance and cell survival was determined. Utilizing real-time qPCR, confocal laser scanning microscopy (CLSM), and crystal violet staining analyses, we found that the RNAIII-independent controlled phenol soluble modulins (PSMs) and RNAIII effector molecule have a significant role in biofilm resistance to antibiotics and phenolic compounds, and it protects the integrity of biofilms. Moreover, a combination of antibiotic and antimicrobial agents can induce methicillin-resistant S. epidermidis biofilm formation and can lead to exceedingly difficult medical treatment. | Role of RNAIII in Resistance to Antibiotics and Antimicrobial Agents in Staphylococcus epidermidis Biofilms
Staphylococcus epidermidis is a known opportunistic pathogen and is one of the leading causes of chronic biofilm-associated infections. Biofilm formation is considered as a main strategy to resist antibiotic treatment and help bacteria escape from the human immune system. Understanding the complex mechanisms in biofilm formation can help find new ways to treat resistant strains and lower the prevalence of nosocomial infections. In order to examine the role of RNAIII regulated by the agr quorum sensing system and to what extent it influences biofilm resistance to antimicrobial agents, deletion mutant S. epidermidis RP62a-ΔRNAIII deficient in repressor domains with a re-maining functional hld gene was created. A deletion strain was used to examine the influence of oxacillin in combination with vanillin on biofilm resistance and cell survival was determined. Utilizing real-time qPCR, confocal laser scanning microscopy (CLSM), and crystal violet staining analyses, we found that the RNAIII-independent controlled phenol soluble modulins (PSMs) and RNAIII effector molecule have a significant role in biofilm resistance to antibiotics and phenolic compounds, and it protects the integrity of biofilms. Moreover, a combination of antibiotic and antimicrobial agents can induce methicillin-resistant S. epidermidis biofilm formation and can lead to exceedingly difficult medical treatment.
The formation of biofilms is the key survival strategy of Staphylococcus epidermidis in the host environment [1]. Biofilm is a multicellular bacterial community defined as a consortium of bacteria coated with extracellular polymeric substances (EPSs) composed of polysaccharides, proteins, and eDNA (extracellular DNA) [2]. EPSs and the heterogeneity of the biofilm are the main factors in the development of antimicrobial-resistant strains such as methicillin-resistant strains of S. epidermidis (MRSE) [3]. In biofilms, poor antibiotic penetration, slow growth, nutrient limitation, and the formation of persister cells are hypothesized to be responsible for drug resistance. Several mechanisms have been described in S. epidermidis that can recognize and overcome the physical and chemical protection of the host. The most important mechanisms include communication systems such as quorum sensing (QS) regulating multiple virulence factors and biofilm formation [4]. The S. epidermidis agr QS contributes to virulence in model infections associated with the biofilm, including endocarditis [5] and osteomyelitis [6], although the exact role of the agr QS system differs from the infection development and site of infection. The model agr locus represented by S. aureus consists of two distinct operons driven by the P2 and P3 promoters [7]. P3 directs the transcription of RNAIII, the effector molecule of the agr locus from which the δ-hemolysin encoded by the hld gene is translated [8]. In S. epidermidis, RNAIII is a 510-nt-long intracellular effector molecule responsible for regulating the expression of many virulence genes [9]. The main principle of RNAIII-dependent regulation of targeted genes expression is the formation of RNAIII duplexes with 5′ un-translated regions (5′UTRs) of targeted genes [10]. By this mechanism, RNAIII mainly inhibits the production of a series of predominantly surface protein A, coagulase (Coa), responsible for maintaining a compact mature form of EPS, adhesins, and in S. aureus, for producing transcription factor repressor (rot) of virulent factors [11]. In addition, RNAIII-dependent regulation leads to an increased production of lipases (Geh) and proteases (SspA, SspB) [12]. In an RNAIII-independent manner, AgrA directly upregulates the transcription of psm genes at the psmα and psmβ operons by binding to the appropriate promoter sequences [13]. Phenol-soluble modulins (PSMs) belong to the family of staphylococcal peptide toxins. In S. aureus, PSMα show cytolytic activity against human neutrophils and erythrocytes. Overall modulins affect cell motility and act as antimicrobials against specific bacterial species such as Streptococcus [14]. S. epidermidis has lower cytotoxic activity than S. aureus and preferably produces β-type PSMs, suggesting that this species begins its defense earlier in the onset of an infection [15]. Moreover, biofilm-forming strains can be more resistant to antibiotics than planktonic cells, since the matrix of biofilm structures acts as a diffusion barrier [16] due to biofilm heterogeneity and different phenotypes such as the expression of efflux pumps and persister cells inside the biofilm [17]. Since RNAIII is the main effector molecule of the agr QS system, inhibition of RNAIII may, therefore, be an effective method for a reduction in the production of toxins and other virulence factors [18]. In recent years, natural antimicrobial compounds have seemed to be suitable anti-QS and antibiofilm agents and many experiments determined the antimicrobial potential of many natural compounds against biofilm formation. Phenolic compounds have been shown to be effective in inhibiting the formation of biofilms and cell growth in pathogenic bacteria [19]. Moreover, these agents have shown modulatory effects on antibiotic resistance in S. aureus and S. epidermidis by inhibiting the function of efflux pumps, which are one of the important factors in biofilm maturation [20]. In this study, the effect of the previously studied vanillin [21] in combination with oxacillin was examined to investigate the possibility of a synergistic effect in treatment. Next, analyzing the role of RNAIII in biofilm resistance to antibiotics and phenolic compounds, an S. epidermidis RP62a-ΔRNAIII deletion mutant with deleted repressor domains was prepared. The deletion strain was used to determine any change in the susceptibility to oxacillin and vanillin and their combination. Moreover, the change in the phenotype was analyzed by implementing CLSM and Congo red agar.
In our previous study [21], vanillin and syringic acid showed a significant antibiofilm and anti-QS effect against S. epidermidis biofilms. The next step was to examine whether the selected phenolic compounds can potentiate the antibiotic susceptibility of methicillin-resistant S. epidermidis. The addition of oxacillin in a 1/2MIC (MIC 4 mg/mL) concentration determined for the S. epidermidis clinical strain 108 was used in combination with vanillin and syringic acid and showed significant induction of biofilm formation (Figure 1). Vanillin and syringic acid in combination with oxacillin in the determined sub-inhibitory concentrations enhanced biofilm formation. All three concentrations showed that the dose-dependent biofilm induction increased by up to threefold for both phenolic compounds compared to the control. The addition of oxacillin had no effect on biofilm formation, as expected. The same effect of biofilm induction was determined for two more methicillin-resistant clinical strains 745 and 817 with different biofilm genotypes (see the Supplementary Materials S2), which were also studied previously in [21]. To identify to what extent the agr QS system and the RNAIII effector molecule potentiate biofilm formation induction and resistance to antimicrobial agents, a deletion mutant S. epidermidis RP62a-ΔRNAIII was prepared. This specific genetic manipulation resulted in the deletion of repressor domains with a remaining functional hld gene.
To achieve an allelic exchange and deletion of the part of the RNAIII effector molecule without hld gene deletion, a plasmid pIMAY-delRNAIII was constructed (Figure 2A) containing a deletion cassette created from two fragments downstream (green) and upstream (blue) from RNAIII (all the details are listed in Supplementary Material S1) (Figure 2B). The plasmid was transformed into S. epidermidis RP62a. Transformation and allelic exchange was accomplished using the protocol for staphylococci by Monk et al. [22]. The deletion of the RNAIII fragment was verified by PCR using the primers DEL1-F and DEL1-R, resulting in a length change of 375 bp compared to wild-type RP62a (Figure 2C). Since deletion of the agr locus is frequent when deletion around the agr locus is performed, retained agr genes were verified by PCR (see Supplementary Material S1) and possible spontaneous mutations in the deletion mutant S. epidermidis RP62a-ΔRNAIII were ruled out by sequencing.
Vanillin was selected to analyze the effect of the phenolic compound in combination with oxacillin on biofilm formation and resistance. First, the 24 h growth curve was determined for both strains (Figure 3). This growth curve for RP62a shows that an addition of 1/2MIC oxacillin had no effect on the growth during 24 h compared to the control. All three sub inhibitory concentrations of vanillin used showed a dose-dependent reduction in bacterial growth. The combination of the highest concentrations of vanillin and oxacillin led to a decrease in the optical density of 67%. In the case of RP62a-ΔRNAIII, the addition of 1/2MIC oxacillin had a higher inhibition effect and lowered bacterial growth by 46% after 24 h. Moreover, the presence of vanillin and oxacillin together resulted in inhibition of bacterial growth. These results were similar to the methicillin-susceptible strain S. epidermidis 108 (see Supplementary Material S2). Next, the biofilm treatment of both strains using only vanillin was analyzed since biofilm as a structure is more resistant to treatment than planktonic cells (Figure 4). The highest concentration of 1/20MIC vanillin used lowered the biofilm formation by 27% compared to the control. The concentrations of 1/40MIC and 1/60MIC also showed similar effects and inhibited biofilm formation by 50% or up to 70%. However, different results were obtained for the analyzed strains when oxacillin in combination with vanillin was used to treat biofilm (Figure 5). As mentioned, vanillin and oxacillin at the 1/2MIC concentration induced biofilm formation in RP62a of up to 300% compared to the control. On the contrary, when RP62a-ΔRNAIII was treated with oxacillin in combination with vanillin, inhibition of biofilm formation was observed. Moreover, the use of 1/2MIC only oxacillin lowered the biofilm mass by 37%. The combination of both agents decreased the biofilm formation and the highest concentration of vanillin used allowed the formation of only 27% of the biofilm mass compared to the control.
To obtain a better insight into changes in the gene expression levels, biofilm formation tests using crystal violet staining were supported by relative mRNA expression analysis of biofilm genetic determinants (icaA and aap), two representative genes of agr QS system (agrA and agrD), and RNAIII-independent genes (psmβ1 and psmβ2). Calculation of the relative mRNA expression of the biofilm determinants aap and icaA genes from RP62A and RP62a-ΔRNAIII in different conditions showed almost identical results (Figure 6). When the 1/2MIC concentration of oxacillin was added to a RP62a biofilm culture, the change in relative mRNA expression was insignificant. Different results were obtained when treated only with vanillin. Aap expression increased by 30% and icaA gene expression decreased by 66%. In the presence of oxacillin and vanillin, the same trend in expression was observed and an increase of 40% for the aap gene and 55% for the icaA gene were detected. For RP62a-ΔRNAIII, similar changes in relative mRNA expression were obtained. Compared to the wild type, the expression of aap and icaA decreased by 4- to 5-fold when only oxacillin was added. Approximately, a 70% decrease was also observed for the deletion mutant without the addition of any agent. Equal changes in the relative mRNA expression were also calculated for the other analyzed conditions, leading to a reduction of up to 84% depending on the agent. When the agr QS genes were analyzed (agrA and agrD), similar results for both genes were obtained (Figure 7). Comparing both strains, for the deletion mutant, no expression was detected for agrA. For agrD, in all conditions, expression was inhibited by up to 99% or no expression was observed. Relative mRNA expression changes for RP62a showed the same trend as for icaA. Oxacillin did not affect the expression, the addition of vanillin led to a decrease in the expression from 30% to 60% and the combination of both agents showed an even more pronounced decrease in the expression of up to 70%. To complete the analysis, the expression of RNAIII-independent genes psmβ1 and psmβ2 was determined (Figure 8). For both psmβ1 and psmβ2, in the case of RP62a-ΔRNAIII, relative mRNA expression was decreased by up to 99% or no expression was determined in all analyzed conditions, with no difference related to the agents used. These results were opposite to the data obtained for RP62a. As for agrA and agrD or the biofilm determinant genes aap and icaA, in the presence of oxacillin, no change was observed. Vanillin increased expression by 30%. The highest change was determined in the presence of oxacillin and vanillin combined, resulting in a 35% increase in psmβ1 and 80% in psmβ2.
In order to complete the analyses of differences in genotypes, a comparison of strains by CLSM and Congo Red agar was carried out to determine any phenotypic differences (Figure 9, Figure 10 and Figure 11). Both methods agree with the results described in Section 2.3 and Section 2.4. CLSM confirmed the enhancement of the production of polysaccharides in strain RP62a compared to the control (Figure 9) when the strains were treated with oxacillin and vanillin. The same trend was observed for proteins and eDNA. RP62a-ΔRNAIII showed the same results and higher signal, suggesting overexpression of the EPS components. Even the real-time qPCR and microtiter-plate analyses showed an inhibition effect. CLSM revealed a change in the structure of the biofilm of RP62a-ΔRNAIII (Figure 10). The strain deficient in the RNAIII gene showed a different composition of EPS (Figure 10, marked by red square) and contained perforated structures throughout the biofilm compared to the wild-type strain (this microscopic picture was selected for representation of perforations). To support the findings of the CLSM, Congo Red agar was used (Figure 11) to determine differences in biofilm formation. A drop of an overnight culture of both strains and the bacterial streak showed differences in their phenotypes. RP62a produced more polysaccharides and appeared black on Congo Red agar. Oppositely, the deletion strain RP62a-ΔRNAIII appeared red after 24 h, showing lower polysaccharide production and supporting the real-time qPCR and microtiter plate analyses.
Due to the unique properties of biofilm formation, according to the WHO, S. epidermidis causes up to 80% of nosocomial infections [23,24]. Based on their capability to colonize various living and non-living surfaces, up to 31% of diseases are associated with bloodstream infections, especially after the introduction of central venous catheters, urea catheters, artificial joint replacements, and heart valves [25]. Phenolic compounds have potential to be suitable antimicrobial agents since an antimicrobial effect on the planktonic stages of bacterial cells and inhibition of mature cells was observed in biofilms without an antibacterial effect, especially in some pathogens such as Listeria monocytogenes [26], Pseudomonas aeruginosa [27], Streptococcus mutans [28], S. aureus, and S. epidermidis [29]. In our previous study, vanillin and syringic acid were selected as antimicrobial agents and their antibiofilm and anti-QS properties were determined [21]. In the present study, vanillin was used to study the potential enhancement of susceptibility to antibiotics, in this case, oxacillin, since similar results of enhancement were obtained for biofilms of S. aureus [30], Klebsiella pneumoniae [31], and E. coli [32]. As an example, previously analyzed flavonoids (−)-epicatechin gallate and (−)-epigallocatechin gallate diminished β-lactam MICs to the antibiotic breakpoint in S. aureus [33,34]. The same effect of the combination of an antimicrobial agent with an antibiotic was determined for patulin combined with tobramycin, resulting in effective killing of bacterial cells [35]. To determine if vanillin or syringic acid can potentiate the antibiotic susceptibility of methicillin-resistant S. epidermidis, oxacillin at a concentration of 1/2MIC (see the material and methods) determined for the susceptible strain S. epidermidis 108 was used with the same concentrations of vanillin and syringic acid (concentrations in the materials and methods) from a previous study [21] (Figure 1). The addition of 1/2MIC of oxacillin and vanillin or syringic acid significantly enhanced the biofilm formation at all used sub-inhibition concentrations of phenolic compounds by up to 300% compared to the control. Both combinations did not influence the culture treated only with oxacillin. The same effect of induction was observed for P. aeruginosa and Agrobacter tumefaciens, where vanillin and the other examined phenolic compounds alone increased the capacity of biofilm formation by 2- to 7-fold depending on the agent used [36]. However, induction was observed only for the use of antimicrobial agent, not in combination with antibiotic. To assess the influence of RNAIII repressor domains on the induction of biofilm formation and resistance to antimicrobial agents, RNAIII-deficient mutant S. epidermidis RP62-ΔRNAIII was prepared with the functional hld gene (Figure 2). Similar mutants of S. epidermidis were prepared, in which the whole agr QS locus was deleted [13]. In this study, only part of the RNAIII coding repressor domains was deleted. An allelic exchange cassette was constructed by amplification of downstream and upstream fragments from the targeted sequence and fused to the functional cassette. (Figure 2A). Successful allelic replacement was then verified by PCR using selected primers (Figure 2B), resulting in a length change of 385 bp (Figure 2C) compared to the wild-type RP62a. Since agr locus deletion can occur frequently, confirmation PCR analysis of the agrA, agrD, and hld genes was performed for wild-type and deletion mutant strains (Figure S1 in Supplementary Material S1). Moreover, parts of the agr locus (agrB, agrD, and agrC, part of agrA) were sequenced and no difference was determined. Next, only the effect of vanillin was examined since the results from our previous study [21] were comparable to those of syringic acid. The growth curves of cultures that were untreated, treated with oxacillin, and treated with oxacillin and vanillin were determined (Figure 3). To determine the growth curves of the analyzed strains, BHI medium was used instead of TSB supplemented with 1% glucose since media with a high content of NaCl or glucose induces biofilm formation [37,38]. The curves of the 24 h untreated cultures showed relatively minor changes in the optical density (OD600). For RP62a-ΔRNAIII (OD600 2.34), OD600 decreased by 13% compared to wild-type RP62a (OD600 2.71). A more significant difference was observed for the treatment with the 1/2MIC oxacillin concentration, where OD600 of the RNAIII-deficient mutant decreased by 45% compared to the control and wild-type RP62a. The same effect of oxacillin was observed for the methicillin-susceptible strain S. epidermidis 108 (see Supplementary Material S2). A combination of 1/2MIC of oxacillin and the analyzed vanillin inhibited the bacterial growth of RP62a-ΔRNAIII completely. The same effect of inhibition was determined only for high doses of antibiotics for infected cardiovascular-implantable electronic device (CIEDs) pathogens P. aeruginosa, S. aureus, and E. coli, which were resistant to antibiotics used for treatment [39]. Since biofilms are 10 to 1000 times more resistant to common antibiotics than their planktonic state [2], the effect of vanillin alone or in combination with 1/2MIC oxacillin on biofilm formation was studied (Figure 4 and Figure 5). As data show (Figure 4), vanillin alone at all sub-inhibitory concentrations showed the same trend of inhibition of biofilm formation for both strains. The highest used sub-inhibitory concentration of vanillin (1/20MIC) lowered biofilm formation by up to 27% compared to the control (Figure 4). Different results were obtained when a combination of antibiotic and phenolic compound was used. As mentioned before, biofilm formation of RP62a increased by up to 300%. On the contrary, when RP62a-ΔRNAIII was treated with antibiotic in combination with vanillin, an inhibitory effect of biofilm formation was observed, allowing the formation of only 27% of the biofilm mass (Figure 5). Moreover, the use of 1/2MIC oxacillin only decreased the biofilm mass by 37%. Some studies have demonstrated that low doses of antibiotics can induce biofilm formation, indicating that biofilm regulation includes the presence of antibiotics. However, this phenomenon is currently unclear and remains under investigation [40,41]. These findings also suggest that repressor domains of RNAIII can regulate other genes as described in Gupta et al. [42]. This study showed RNAIII regulation of mgrA through binding to two regulation sites of mgrA, one of which was deleted in this study. Moreover, the study of Gupta et al. showed no effect on resistance to oxacillin. The combination of vanillin and oxacillin, which affects the survival of viable cells, could lead to stress enhancement and the involvement of unknown factors, enhancing biofilm formation In this study, the results also showed significant enhancement of the susceptibility to oxacillin. These findings suggest a significant impact of the RNAIII repressor domains on the regulation of other virulence factors through transcriptional regulators such as mgrA; however, more research in this direction is needed. These results also show that biofilm EPS provides high defense against antimicrobial agents and antibiotics [13]. To support the microtiter plate method findings, real-time qPCR analysis was performed. Calculation of the relative mRNA expression of the biofilm determinants aap and icaA genes from RP62A and RP62a-ΔRNAIII in different conditions showed almost identical results (Figure 6). Oxacillin at the 1/2MIC concentration changed the relative mRNA expression. Regarding the biofilm cultures treated with only vanillin, aap expression increased by 30% and icaA gene expression decreased by more than 60%. In the presence of oxacillin and vanillin, aap expression decreased by 40% and 55% for the icaA gene. For RP62a-ΔRNAIII, compared to the wild type, the expression of aap and icaA decreased only when oxacillin was added, resulting in a 4- to 5-fold decrease. The absence of both agents had a similar effect and equal change in the relative mRNA expression, which was calculated from cultures in the presence of both agents, of up to 84% (Figure 6). When agr QS genes were analyzed (agrA and agrD), comparable results for both genes were obtained for the RP62a strain (Figure 7). When the expression of agr genes influenced by the analyzed antimicrobial agent or oxacillin decreased, biofilm formation increased. These results agree with a previous finding, where agr deficiency or lower expression enhanced biofilm formation [43]. Comparing both strains, for the deletion mutant, no expression was detected for agrA. For agrD, in all conditions, expression was inhibited by up to 99% or no expression was observed. Interestingly, even the absence of agrA and agrD expression in all conditions for the RNAIII-deficient mutant had no influence on biofilm formation and this strain formed only 10% less biofilm mass compared to the wild-type strain RP62a. The same effect was observed for the deletion of the agrD gene in S. aureus, where no change, or even enhancement, in biofilm formation was observed [44]. A partial explanation could be that under certain conditions, quorum sensing may play a role in biofilm development, inducing genes responsible for the attachment to the surface but may not play an obvious role in biofilm development under all conditions [43]. The agr QS system is thought to exercise regulatory control over a wide range of staphylococcal genes such as the analyzed psmβ1 and psmβ2, many lipases, exoproteases, or virulence factors [45]. However, conditions that result in an increase biofilm formation were not determined and remain unknown. Greater insight into the crucial part of RNAIII repressor domains in biofilm resistance was offered by the analysis of psmβ1 and psmβ2 gene expression (Figure 8). As the data shows, relative mRNA expression increased by up to 37% for psmβ1 and 80% for the psmβ2 gene when wild-type RP62a was treated with the combination of oxacillin and vanillin compared to the control. These findings are contradictory to the finding published by Queck et al. [13]. This study showed enhanced PSM production even when the expression of agrA-regulating psm genes was decreased. On the contrary, no agrA expression led to no expression of PSMs in RP62a-ΔRNAIII. Increased PSM expression in the wild-type strain can be influenced by other non-identified regulator molecules; however, more analyses are needed to further explain the observed effect. CLSM confirmed the overexpression of PSMs for RP62a (Figure 9), where the almost 2-fold increase detected by qPCR correlated with the enhanced signal of protein in the sample under microscopy. These findings agree with microtiter plate analyses (Figure 3, Figure 4, Figure 5 and Figure 6). Moreover, as CLSM shows, other components were enhanced, such as polysaccharides and eDNA. Interestingly, CLSM showed the same results for RP62a-ΔRNAIII. The signals of all three components were 3 to 4 times higher compared to the control. Even staining with crystal violet showed an inhibitory effect (Figure 6). One of the hypotheses is the role of PSMs in the biofilm structure. All PSMs of S. aureus and S. epidermidis PSMβ peptide structure biofilms are so far the only bacterial factors for which roles in the in vivo dissemination of biofilm-associated infection could be directly demonstrated [15,39]. In isogenic S. aureus, psm gene deletion mutants led to highly disrupted formation of biofilm channels, abolishment of the characteristic waves of biofilm detachment and regrowth, and loss of control of biofilm expansion [46]. The described characteristics of disrupted biofilm channels were determined for the RNAIII-deficient strain RP62a-ΔRNAIII (Figure 10) compared to the wild type. This may explain why the biofilm of the deficient strain showed an increased signal even when the inhibition effect was determined. The hypothesis is that with no PSM expression, biofilm formed but was less rigid and flatter, without the typical “mushroom form” 3D structure and other layers of cells were captured by CLSM (Figure 9). The Congo Red method was used to supplement the other method and detect any possible changes in the phenotype. The biofilm 24 h culture of RP62a produced more polysaccharides and appeared black. Oppositely, the deletion strain RP62a-ΔRNAIII appeared red after 24 h, displaying lower polysaccharide production (Figure 11). This study confirms that vanillin alone can act as an antibiofilm and anti-QS agent. When the selected phenolic compounds are combined with antibiotic, in this case, oxacillin, possible stress conditions lead to enhanced expression of PSMs by unknown mechanism, and polysaccharides and eDNA form rough EPS and protect cells inside the biofilm of methicillin-resistant S. epidermidis. The RNAIII-deficient strain produced very low or no PSMs became susceptible to oxacillin and with vanillin treatment had synergistic effects and no growth was detected. These results provide a hypothesis that for S. epidermidis-producing biofilms determined as a cause of infection, it is important to determine whether the strain is resistant or susceptible to antibiotics used for treatment since when treated with antimicrobial agents or their combination, it can lead to an increase in biofilm formation, and lead to more difficult treatment of infection.
Briefly, the fragments DEL1 and DEL2 were amplified by specific primers (the primers are listed in Supplementary Material S1) from genomic DNA of RP62a. Next, both fragments were joined together by fusion PCR using only the outer primers of each fragment and a cassette for allelic exchange was prepared. For PCR, Dream Taq DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA, cat. no. EP0701) was used. The cassette delRNAIII was cloned into the plasmid pIMAY using the restriction sites NotI and EcoRI and the plasmid pIMAY-delRNAIII was constructed. The protocol by Monk et al. [20] was used for electroporation and allelic exchange in S. epidermidis.
The biofilm-producing strain S. epidermidis RP62A (ATCC 35984) and the prepared deletion mutant S. epidermidis RP62a-ΔRNAIII was examined. Strains were maintained in BHI medium for determination of the growth curve. For the biofilm assays, a TSB medium supplemented with 1% glucose was used to induce biofilm formation. For the selection of positive clones, BHI agar was used as described in the protocol by Monk et al. [22].
Vanillin was isolated from oak bark lignin waste and was provided by the Slovak Technical University in Bratislava. Vanillin was stored at room temperature and diluted at subinhibitory concentrations in 99.9% dimethylsulfoxide (DMSO). The range of used subinhibitory concentrations of vanillin (1/60MIC, 1/40MIC, 1/20MIC) and syringic acid in this work was prepared from stock solution at a concentration of 250 mM/mL. Stocks were stored at 4 °C when used for the assays. The sub-inhibitory 1/2MIC concentration of oxacillin was prepared from stock solution at a concentration of 4 mg/mL.
The biofilm formation assay was analyzed by crystal violet staining for staphylococci [47]. Overnight cultures were cultivated at 37 °C and were diluted with TSB supplemented with 1% glucose at a 1:100 ratio, and 200 µL of bacterial culture was poured into the well. For the inhibition effect analysis, 20 µL of antimicrobial agent was poured into 180 µL of bacterial culture in the well and biofilm was grown at 37 °C for 24 h. For staining, a solution of 0.1% crystal violet was used. The optical density (OD) was measured at 570 nm using a microtiter-plate reader (Varioskan Flash, Thermo Fisher Scientific, Waltham, MA, USA).
During determination of the bacterial growth rate for 24 h, strains were grown in BHI medium at 37 °C, with shaking at 160 rpm. These conditions were maintained by a microtiter plate reader (Varioskan Flash, Thermo Fisher Scientific, Waltham, MA, USA). The optical density (OD) was measured every hour at 570 nm.
WGA (Wheat Germ Agglutinin, Alexa Fluor™ 488 Conjugate, Thermo Fisher Scientific, Waltham, MA, USA, cat.no. W11261) at a final concentration of 5 µg/mL was used to stain PIA/PNAG in EPS, PI at a final concentration of 3 µg/mL was used to stain eDNA, and FilmTracer™ SYPRO™ Ruby Biofilm Matrix Stain (Thermo Fisher Scientific, Waltham, MA, USA, cat.no. F10318) at the final concentration in the protocol was used to stain proteins. Welco wells B.V. (Netherlands) glass-bottom Petri dishes were used to form biofilm. After 24 h, biofilms were washed with PBS, and stained with appropriate dye. Microscopy was performed on a FluoroView FV1200 (Olympus LifeSciences), and images were adjusted and correlated by FluoroView FV1200 Software. An oil objective with 60× zoom was used for imaging.
To analyze the differences in the phenotypic production of biofilm, the modified Congo Red agar method of Kaiser et al. [48] was used. BHI agar with sucrose (5%), Congo Red (0.08%), NaCl (1.5%), glucose (2%), and vancomycin (0.5 mg/mL) was prepared. A volume of 20 µL of overnight culture was dropped on the agar plate, air-dried, and incubated for 24 h at 37 °C.
The biofilm was grown in a polystyrene Petri dish for 24 h in static conditions at 37 °C and washed with PBS (pH 7.2) to remove planktonic cells. The biofilm was scraped off, nuclease-free water (1 mL) was added to the dish, and cells were transferred to a microcentrifuge tube and pelleted by centrifugation at room temperature at 13,400 rpm (MiniSpin® Eppendorf—rotor F45-12-11) for 5 min. The pellet was resuspended with 100 μL of RNase-free water, 100 μL of a phenol: chloroform mixture at a 1:1 ratio was added, and the mixture was incubated at 70 °C for 30 min with shaking followed by centrifugation. The water fraction was then precipitated with isopropanol. Pellet was washed with 70% ethanol two times, air-dried, and resuspended in nuclease-free water. The concentration of RNA was measured by a NanoDrop (Thermo Fisher Scientific, Waltham, MA, USA).
A One-Step PCR kit (Qiagen), following the standard manufacturer’s protocol, was used. For all genes, the primer annealing temperature was set to 60 °C, and for amplification, Roche LightCycler® 480 96-well half-skirted plates were used. Real-time analysis was performed in triplicates on the Light Cycler® 480 System by Roche. The primers are listed in the Supplementary Materials (Table S1).
All the measurements were performed in biological triplicates. Values as mean ± SD were obtained from Microsoft Excel, and statistical significance was evaluated by the t-test (p < 0.05) or one-way ANOVA (p < 0.01).
The presented study demonstrates the effect of oxacillin in combination with vanillin as an antimicrobial agent against methicillin-resistant S. epidermidis. For this research, vanillin was selected based on our previous study and potential enhancement of antibiotic susceptibility was determined. To analyze the role of RNAIII in biofilm resistance to antimicrobial agents, the S. epidermidis RP62a-ΔRNAIII strain was prepared. Our findings show that treatment of biofilms with oxacillin in combination with vanillin can enhance the biofilm formation of wild-type RP62a by up to threefold compared to the control. The opposite effect of these agents was observed for the RP62a-ΔRNAIII strain, where biofilm formed only up to 27% compared to the control. These results were supported by real-time qPCR. Data showed lowered relative mRNA expression or no expression compared to the control of all analyzed genes. The CLSM analysis and the Congo Red method supplemented the findings of importance of the RNAIII repressor domains and its important role in the biofilm structure and resistance to antimicrobial agents and antibiotics. When PSMs were not expressed, the biofilm was flatter with a perforated structure in EPS and lowered polysaccharide and protein production led to the formation of more susceptible biofilm to antimicrobial agents and antibiotics. Oppositely, stress conditions triggered by vanillin and oxacillin led to increased biofilm production when treated with oxacillin and vanillin, suggesting the importance of knowing whether the strain is antibiotic resistant or susceptible to select the optimal treatment. | true | true | true |
PMC9569911 | 36233029 | Taylor S. Campbell,Katelyn M. Donoghue,Urmi Ghosh,Christina M. Nelson,Tania L. Roth | Early Life Stress Affects Bdnf Regulation: A Role for Exercise Interventions | 03-10-2022 | early life stress,neurotrophins,Bdnf,epigenetics,aerobic exercise | Early life stress (ELS) encompasses exposure to aversive experiences during early development, such as neglect or maltreatment. Animal and human studies indicate that ELS has maladaptive effects on brain development, leaving individuals more vulnerable to developing behavioral and neuropsychiatric disorders later in life. This result occurs in part to disruptions in Brain derived neurotrophic factor (Bdnf) gene regulation, which plays a vital role in early neural programming and brain health in adulthood. A potential treatment mechanism to reverse the effects of ELS on Bdnf expression is aerobic exercise due to its neuroprotective properties and positive impact on Bdnf expression. Aerobic exercise opens the door to exciting and novel potential treatment strategies because it is a behavioral intervention readily and freely available to the public. In this review, we discuss the current literature investigating the use of exercise interventions in animal models of ELS to reverse or mitigate ELS-induced changes in Bdnf expression. We also encourage future studies to investigate sensitive periods of exercise exposure, as well as sufficient duration of exposure, on epigenetic and behavioral outcomes to help lead to standardized practices in the exercise intervention field. | Early Life Stress Affects Bdnf Regulation: A Role for Exercise Interventions
Early life stress (ELS) encompasses exposure to aversive experiences during early development, such as neglect or maltreatment. Animal and human studies indicate that ELS has maladaptive effects on brain development, leaving individuals more vulnerable to developing behavioral and neuropsychiatric disorders later in life. This result occurs in part to disruptions in Brain derived neurotrophic factor (Bdnf) gene regulation, which plays a vital role in early neural programming and brain health in adulthood. A potential treatment mechanism to reverse the effects of ELS on Bdnf expression is aerobic exercise due to its neuroprotective properties and positive impact on Bdnf expression. Aerobic exercise opens the door to exciting and novel potential treatment strategies because it is a behavioral intervention readily and freely available to the public. In this review, we discuss the current literature investigating the use of exercise interventions in animal models of ELS to reverse or mitigate ELS-induced changes in Bdnf expression. We also encourage future studies to investigate sensitive periods of exercise exposure, as well as sufficient duration of exposure, on epigenetic and behavioral outcomes to help lead to standardized practices in the exercise intervention field.
Over 30 years ago, Dr. David Barker presented the Barker Hypothesis, which stated that the perinatal environment sets off a chain reaction of neural programming that determines cognitive function and emotional health in adulthood [1]. In line with this idea, the perinatal period is marked by a degree of neural plasticity that is highly sensitive to environmental influences and not seen during any later period of life [2,3,4]. During this developmental period, a cascade of neural processes work in tandem to program the brain [4]. Negative experiences during early development, such as stress, can alter the epigenetic regulation of neurotrophins [5,6,7,8] and thereby increase an individual’s susceptibility to later development of neuropsychiatric and neurodegenerative disorders [6,7,8,9]. Epigenetics refers to the addition of molecules, such as methyl groups, to DNA strands that alter the way that DNA can be read and used in the body. Hence, the consequences of these early disruptions can become embedded in the DNA and stay with a person through their entire lifespan. Neurotrophins regulate brain development during infancy and adolescence, and in adulthood, they facilitate synaptic plasticity and neuronal survival [6,10,11,12]. Perturbations during development often dysregulate neurotrophin expression, leading to brain-region-specific maladaptive changes in expression [13,14,15,16]. In contrast, aerobic exercise typically upregulates neurotrophin expression, and is thought to be neuroprotective [17,18,19,20]. Most notably, exercise has a neuroprotective effect on the brain by preventing neuronal cell death [21] and facilitating adaptive cellular processes, including synaptogenesis (formation of synapses) [21] and neurogenesis [22]. For example, rodent studies show that aerobic exercise promotes neurogenesis in the hippocampus, a brain region where neurogenesis is abundant throughout adulthood [22,23]. This is important when considering that neurogenesis is reduced in the hippocampus of rodents exposed to developmental stress, making exercise a potential treatment mechanism for the maladaptive biological effects of stress [24,25,26,27]. Consistent with findings in animal subjects, imaging studies in humans indicate that child maltreatment leads to a decrease in hippocampal volume [28,29,30], as well as decreased hippocampal activation during threat detection [31] and memory tasks [32]. In this review we discuss the literature on rodent models of early life stress (ELS) and exercise interventions. We pay special attention to the effect of these experiences on neurotrophin regulation, with the capacity of exercise to correct biological processes and aberrant behavior associated with early stress.
In humans, ELS encompasses many different experiences. These range from physical and emotional abuse and neglect to living in a war-torn country, experiencing extreme poverty, or the death of a caregiver. In the laboratory, researchers recapitulate similar experiences in animal models to investigate the biological consequences of stress during early development. Two common models used are the limited bedding and nesting (LBN) model and the maternal separation (MS) model. In the LBN model, a rodent mother is not provided with sufficient nesting material to properly care for the pups. This causes stress in the dam and elicits more aversive behaviors from her towards the pups, such as actively avoiding and rough handling them, and less frequent or fragmented maternal behaviors such as arched back nursing and hovering over the pups [33,34]. This model is used to study the consequences of disrupted infant–caregiver interactions during early life. In the MS models, pups are separated from the dam for varying periods of time during early development [33,35] to approximate the experience of caregiver neglect. Rodent models of aerobic exercise either use involuntary treadmill running or voluntary wheel running paradigms. Currently, the timing and duration of exercise exposure as a treatment intervention has not been standardized (see Figure 1). In voluntary models, rodents have continuous free access to running wheels, usually via a window cut out in the cage that allows the subjects to freely migrate to an attached wheel. Involuntary exercise models employ a treadmill apparatus that keeps the subject’s feet on the treadmill, forcing the subject to locomote for a set duration of time and intensity. Additional studies have employed both voluntary and involuntary aerobic exercise to compare the effects of the two running models. One study found that voluntary running wheel exercise decreased immobility time during a forced swim test of adolescent rats exposed to maternal separation; however, this antidepressant-like effect was lost in the involuntary exercised subjects [36]. Studies of this nature question the efficacy of forced exercise models as they likely upregulate the subject’s stress response, leading to a diminished neuroprotective effect. For example, Ke and colleagues [37] employed an animal model of stroke to measure the ability of aerobic exercise to recover motor behavior function and increase brain-derived neurotrophic factor (BDNF) protein levels in the hippocampus. Their results showed that voluntary, but not involuntary, exercise improved motor behavior overtime and increased BDNF in the hippocampus. In contrast, rats in the involuntary exercise group showed increased corticosterone (CORT) levels as well as a decrease in BDNF levels in the hippocampus compared to the control group [37]. These results are further supported by a 2016 study that found involuntary exercise to be maladaptive for stroke recovery in animal models. Svensson and colleagues [38] showed that, in a model of ischemic stroke, involuntary exercise increased anxiety-like behaviors on the open-field test (OFT), increased neuron loss in the right hippocampus, and increased fecal CORT levels following the OFT. They also noted a positive correlation between CORT levels and neuron loss [38]. In a 2014 study by Uysal and colleagues [39], voluntary exercise led to a decrease in basal CORT levels compared to both sedentary and involuntary exercised rats that was also accompanied by an increase in locomotion on the OFT in voluntarily exercised rats. Results from this study also showed that female rats exposed to voluntary exercise had increased BDNF protein levels in the prefrontal cortex (PFC) compared to sedentary rats. In male rats, BDNF levels increased in both exercise groups compared to the sedentary rats; however, there was a significantly greater increase in BDNF levels in the PFC of voluntarily exercised rats compared to the involuntary exercise group [39]. Taken together, these studies indicate that voluntary exercise models may be more advantageous when investigating the anxiolytic and neuroprotective effects of exercise, as involuntary exercise may exacerbate neural insults, including those caused by ELS, by upregulating CORT reactivity in the brain. In addition to the type of exercise, the duration of exercise and age of exposure may play a role in its effectiveness. For example, Greenwood and colleagues [40] showed that six, but not three, weeks of voluntary wheel running was sufficient to prevent learned helplessness behaviors when subjects were exposed to uncontrolled tail shocks later. This same research group also showed that chronic voluntary exercise exposure is more rewarding in rats compared to short-term exposure. Six weeks, but not two weeks, of voluntary exercise led to exercise-induced changes in gene expression and receptor activity in the mesolimbic dopamine pathway that were accompanied by preference for a chamber that was previously paired with wheel running exposure on a conditioned place preference task [41]. Six weeks of voluntary exercise is also sufficient to decrease habituation time to future stress (loud noise exposure) as measured by significantly reduced plasma CORT levels compared to sedentary rats [42]. A breakdown of the ELS models and exercise intervention methods used in the current literature is provided in Table 1.
In laboratory models, the effects of ELS combined with later exercise experience vary based on the specific experimental parameters (see Table 2). Though nuances are present in the literature, a frequent finding suggests that rats with a history of ELS show increased anxiety- [52,53,58,60,62] and depressive-like [6,53,54,56,57,58,62] phenotypes, and these phenotypes are ameliorated by exercise exposure [36,51,52,53,54,56,57,58]. These outcomes are mostly illustrated in male rats exposed to MS, with information on female rodent outcomes and sex differences, as well as rats exposed to other models of early stress, severely lacking. However, there is some evidence that stress and exercise differentially affect sexes. For example, James and colleagues [52] showed that exercise ameliorates anxiety-like behaviors in male rats exposed to MS but worsens these behaviors in females exposed to MS. An outcome such as this underscores the importance of studying sex differences, including the effect of hormones on behavior and susceptibility to stress. Indeed recent studies have shown that estrogen can significantly impact the effect of trauma on the brain and susceptibility to psychiatric disorders [63,64]. The effects of exercise are not always consistent on ameliorating ELS-phenotypes, with increases in anxiety behavior [49,52] or no effects [36] sometimes observed. Rats may use the wheels to facilitate an escape behavior [44], which could have consequences for anxiety behavior. Another factor contributing to the inconsistency in exercise effects reflects the use of a voluntary WR treatment verses an involuntary treadmill (TM) exercise treatment. For example, Sadeghi and colleagues [36] reported that WR exposure decreased depressive-like behavior in rats but that TM exposure had no effect. Outside of affecting any ELS outcomes, exercise did bolster behavioral performance on cognitive and memory tasks in several studies [49,61], as is a common finding in the exercise literature [65].
One way our experiences can get under the skin to affect genes, including the Bdnf gene, is through epigenetic mechanisms. Epigenetics refers to modifications to DNA that affect gene expression without making changes to the genetic sequence. One form of epigenetic regulation is called DNA methylation, wherein a methyl group is added to the cytosine at a CG site (cytosine-guanine dinucleotide) on the DNA [66,67]. CG sites are highly potent surrounding the promotor regions of most genes, making them a prime target for gene regulation. Increased methylation at promoter regions typically leads to decreases in gene expression because methyl groups recruit repressor proteins, interact with chromatic structure, and inhibit transcription factors from binding [66,68]. As researchers look to understand how early life stress can have long-term behavioral consequences and how exercise can reprogram the brain to have neurotherapeutic effects, focus often turns to neurotrophins, especially BDNF (see Figure 2 for our theoretical framework). Neurotrophins are a family of proteins which induce the development, survival, and function of neurons [6,69,70]. BDNF’s neurotrophic actions are vital for brain development and plasticity, and BDNF exhibits activity-regulated release in the central nervous system [71,72,73]. BDNF is a neurotrophin important for neural development, neural plasticity, learning, memory, and synaptic plasticity later in life, especially within the hippocampus [74,75]. Typically, increased methylation of the Bdnf gene is associated with decreased expression of its genetic material [68]. Methylation at any of Bdnf’s nine promotor regions can lead to decreased transcription of total Bdnf mRNA [76]. Stress during neonatal development has the capacity to alter Bdnf methylation for the long haul [5,77,78], which is important as decreased BDNF protein expression is found in patients with neurogenerative diseases and neuropsychiatric disorders [79,80,81,82]. These data highlight Bdnf as an important genetic locus for studies investigating epigenetic-behavioral interactions.
Many studies have shown that ELS reduces levels of both Bdnf mRNA and BDNF protein in multiple brain regions, including the prefrontal cortex and hippocampus [5,70,77,78,83,84,85]. Early life exposure to stress, especially within a caregiving environment, can result in a decrease in Bdnf gene expression through increased methylation of the Bdnf gene [5,86,87,88]. This impact of developmental stress extends to humans. For example, Bdnf DNA methylation correlates with the number of aversive childhood experiences in patients with bipolar disorder [89]. Further research in humans sheds light on the transgenerational effects of ELS on BDNF, in that babies born to mothers who experienced ELS show changes in Bdnf methylation and expression in blood cells obtained from the umbilical cord based on infant sex and degree of maternal fear [90]. While ELS generally decreases neurotrophin levels, aerobic exercise increases neurotrophin expression and is thought to be neuroprotective [17,91,92,93]. Exercise has positive impacts on neurotrophin expression, which directly impact neuronal survival and neurogenesis. Previous studies have identified exercise as a behavioral mechanism that specifically increases Bdnf expression [55,56,93] and decreases Bdnf methylation [94,95]. Several studies have also reported significant associations between exercise-induce BDNF upregulation and improved cognition [96] and depression symptoms [97]. The effect of exercise on neurotrophin expression in rodents exposed to ELS is understudied, with only 6 studies beginning to elucidate this relationship to date. Within these studies, exercise affected Bdnf/BDNF expression in a nonuniform and nuanced manner. Three studies showed that voluntary WR increased Bdnf mRNA in the hippocampus following 3hrs/daily MS during the first 2-3 weeks of life [48,55,59], and this increase was associated with rescued hippocampal neurogenesis in the dentate gyrus [48]. At the protein level, one study found that exercise increased BDNF expression in the striatum but not the ventral hippocampus compared to sedentary MS-exposed rats [56]. Given the complicated nature of Bdnf expression and gene regulation, Wearick-Silva and colleagues [61] investigated the exon-specific effects of exercise in the hippocampus. They reported that MS decreased Bdnf exon IV expression and increased exon IX expression, while exercise had an opposite effect on exon IX and increased Bdnf exon I expression. To shed light on the mechanism behind exercise-induced increases in Bdnf/BDNF expression, future studies should measure Bdnf exon-specific methylation in conjunction with expression to determine if specific genetic loci act in tandem to alter de novo expression. Little is known regarding the effect of exercise on other neurotrophins in this ELS context. Marais and colleagues [56] produced the only current study investigating NT-3 and nerve growth factor (NGF) in this model, where they reported no significant effects on these neurotrophins in the ventral hippocampus and striatum. However, further investigation is warranted given that areas known to be highly impacted by stress and exercise, including the cerebellum, PFC, and dorsal hippocampus [98,99], have been overlooked.
ELS alters long-term neurotrophin expression in the brain. These epigenetic changes contribute to an individual’s risk of numerous neurological, immune, and psychiatric disorders. We propose use of aerobic exercise as a treatment mechanism in future studies to understand the capacity of exercise to bolster the brain and body against ELS-induced disruption of biological processes. Support for our line of thinking comes from studies showing that exercise improves spatial memory, autoimmune and neurodegenerative disease symptomatology, muscle function, and gut health with concomitant changes in various neurotrophin expression levels. Exercise has become a popular research area in neuroprotective research fields for its promising effects on brain health. A 2017 study following the natural aging of older adults (for ~10 years) showed that exercise is positively related to total cerebral and hippocampal volumes, and negatively related to developing Alzheimer’s Disease and dementia [100]. This suggests exercise is neural protective against neuronal and glial cell loss through the lifespan. It’s widely accepted that exercise has a positive effect on BDNF expression and epigenetic regulation [94,101], however, the data on other neurotrophins is much more nuanced and understudied [102]. Currently, only two studies have investigated the effect of exercise on NT-3 expression in humans [103,104]. A 2021 cutting-edge study reported that 12 weeks of high-intensity interval training increased serum levels of BDNF, NGF, NT-3, and NT-4 in elderly, obese, Chinese subjects [104]. Importantly, future research must critically examine exercise and participant parameters as these variables seem to be important when comparing participant outcomes. For example, in a 2018 clinical trial of adult obese males, high-intensity interval training (HIIT) had no effect on blood BDNF, NT-3, or NT-4 [103]. However, resistance training increased NT-3 and NT-4 levels, and combined exercise (resistance training plus HIIT) increased NT-3 and BDNF levels. In the combined exercise group, BDNF and NT-3 levels were positively correlated [103]. Taken together, these studies underscore the importance of continued research at the intersection of exercise science and neuroscience. Further research indicates that aerobic exercise may be a valuable treatment mechanism for neurodevelopmental, autoimmune and psychological disorders. Aerobic exercise is a popular intervention strategy in studies investigating the molecular and cognitive effects of fetal alcohol spectrum disorders, with exercise exposure ameliorating Bdnf dysregulation [94], corpus callosum volume deficits [105], and executive functioning [106]. Recent work also demonstrates that aerobic exercise decreases the pathogenesis of multiple sclerosis [107], while increasing peripherally circulating BDNF and NGF [108]. Moderate physical exercise for 6 weeks is sufficient to reduce self-reported depression levels and increase peripheral BDNF and NGF levels in postmenopausal woman [109]. The therapeutic benefits of exercise on disease states underline a promising future for exercise intervention models. | true | true | true |
PMC9569918 | 36233020 | Qi Zhang,Zhonghao Li,Xianyan Liu,Ming Zhao | Recombinant Humanized IgG1 Antibody Protects against oxLDL-Induced Oxidative Stress and Apoptosis in Human Monocyte/Macrophage THP-1 Cells by Upregulation of MSRA via Sirt1-FOXO1 Axis | 03-10-2022 | oxidative stress,apoptosis,MSRA,SIRT1-FOXO1 axis | Oxidized low-density lipoprotein (oxLDL)-induced oxidative stress and apoptosis are considered as critical contributors to cardiovascular diseases. Methionine sulfoxide reductase A (MSRA) is a potent intracellular oxidoreductase and serves as an essential factor that protects cells against oxidative damage. Here, we firstly provide evidence that recombinant humanized IgG1 antibody treatment upregulated the expression of MSRA in THP-1 cells to defend against oxLDL-induced oxidative stress and apoptosis. It was also observed that the upregulation of MSRA is regulated by the forkhead box O transcription factor (FOXO1), and the acetylation of FOXO1 increased when exposed to oxLDL but declined when treated with recombinant humanized IgG1 antibody. In addition, we identified that silent information regulator 1 (SIRT1) suppresses FOXO1 acetylation. Importantly, SIRT1 or FOXO1 deficiency impaired the anti-oxidative stress and anti-apoptotic effect of recombinant humanized IgG1 antibody. Together, our results suggest that recombinant humanized IgG1 antibody exerts its anti-oxidative stress and anti-apoptotic function by upregulation of MSRA via the Sirt1-FOXO1 axis. | Recombinant Humanized IgG1 Antibody Protects against oxLDL-Induced Oxidative Stress and Apoptosis in Human Monocyte/Macrophage THP-1 Cells by Upregulation of MSRA via Sirt1-FOXO1 Axis
Oxidized low-density lipoprotein (oxLDL)-induced oxidative stress and apoptosis are considered as critical contributors to cardiovascular diseases. Methionine sulfoxide reductase A (MSRA) is a potent intracellular oxidoreductase and serves as an essential factor that protects cells against oxidative damage. Here, we firstly provide evidence that recombinant humanized IgG1 antibody treatment upregulated the expression of MSRA in THP-1 cells to defend against oxLDL-induced oxidative stress and apoptosis. It was also observed that the upregulation of MSRA is regulated by the forkhead box O transcription factor (FOXO1), and the acetylation of FOXO1 increased when exposed to oxLDL but declined when treated with recombinant humanized IgG1 antibody. In addition, we identified that silent information regulator 1 (SIRT1) suppresses FOXO1 acetylation. Importantly, SIRT1 or FOXO1 deficiency impaired the anti-oxidative stress and anti-apoptotic effect of recombinant humanized IgG1 antibody. Together, our results suggest that recombinant humanized IgG1 antibody exerts its anti-oxidative stress and anti-apoptotic function by upregulation of MSRA via the Sirt1-FOXO1 axis.
Atherosclerosis has been the leading cause of cardio-cerebrovascular disease with increasing morbidity and mortality [1,2]. Atherosclerosis is a complex disease that involves chronic inflammation in combined with metabolic risk factors and autoimmune response [3]. Atherosclerosis is characterized by activated monocytes/macrophages ingesting oxidized lipids, such as oxidized low-density lipoprotein (oxLDL), and becoming lipid-laden “foam” cells; this results in the production of proinflammatory cytokines and chemokines, further leading to progressive necrotic lipid core formation and eventually developing into fibrous plaque. Moreover, macrophages decrease phagocytic activity and undergo polarization to pro-inflammation macrophages upon oxLDL recognition and internalization [4]. Apoptosis of lesional monocytes/macrophages and their defective function on cleaning up dead cells leads to plaque instability and thrombosis [5]. Since monocytes/macrophages play a central role in the stability of the atherosclerotic plaque, it is necessary to clarify the underlying apoptosis mechanism of the lesional monocytes/macrophages. Moreover, it is known that mitochondrial DNA mutations also contribute to atherosclerosis initiation and progression [6]. Proteins with sulfur-containing amino acids cysteine and methionine residues are particularly sensitive to oxidation by reactive oxygen species (ROS) and this oxidation process can subsequently be reduced by the action of certain reductases [7]. Particularly, oxidation of methionine results in two enantiomers [8], namely S-form or R-form of oxidized methionine sulfoxide. Methionine sulfoxide reductase A (MSRA) is one of methionine sulfoxide reductases that are responsible for reducing methionine sulfoxide to methionine in proteins [9]. MSRA specifically reduces the S-form of both free and peptide-bound methionine sulfoxide. Interestingly, the essential role of MSRA in defending against oxidative stress has been studied extensively. It has been reported that MSRA could decrease susceptibility to oxidative stress and increase longevity in Drosophila [10,11], yeast [12], and mice [13,14]. However, the detailed mechanism remains unknown. The forkhead box O (FOXO) transcription factors in mammals comprise four members, namely FOXO1, FOXO3, FOXO4 and FOXO6, which are involved in regulating many cells’ physiological processes, such as oxidative stress, apoptosis, cell cycle, and cell survival and differentiation [15]. FOXO1 is mainly expressed in insulin-responsive tissues, such as pancreas, liver, skeletal muscle and adipose tissue, and also acts as a master regulator of energy metabolism [16]. As a transcription factor, the activity of FOXO1 is primarily dependent on its protein post-translational modifications, such as phosphorylation and acetylation. It has been reported that silent information regulator 1 (SIRT1), a nicotinamide adenine dinucleotide (NAD+)-dependent histone deacetylase, deacetylates FOXO1 and regulates its activity [17,18]. Therefore, it may be a promising strategy to attenuate oxidative stress and apoptosis through improving the SIRT1-FOXO1 axis. We previously had reported that the recombinant humanized IgG1 antibody reduces atherosclerosis in ApoE−/− mice fed with a high-fat-diet after immune therapies, and activates macrophage polarization from M1 toward M2 [19]. The purpose of the present work was to determine the modulation of expression of MSRA upon exposure of THP-1 cells to oxLDL. We found that recombinant humanized IgG1 antibody alleviated oxLDL-mediated THP-1 cells’ apoptosis by upregulation of MSRA. We also demonstrated that recombinant humanized IgG1 antibody decreased FOXO1 acetylation by targeting SIRT1 expression. Correspondingly, ectopic SIRT1 expression blocked oxLDL-induced FOXO1 acetylation and THP-1 cells’ apoptosis. Our data highlight a functional role of recombinant humanized IgG1 antibody in protection against oxidative stress and apoptosis and uncover its possible mechanisms.
Our previous work has revealed that recombinant humanized IgG1 antibody inhibits oxLDL-induced apoptosis in CD14+ monocytes derived from human peripheral blood (data not shown). To determine whether the antibody has the same effect on human monocyte/macrophage cell line THP-1, we treated the THP-1 cells with oxLDL along with or without recombinant humanized IgG1 antibody (14 Ab) to detect apoptosis by flow cytometry for Annexin V staining. As shown in Figure 1A, exposure to oxLDL led to a significant increase in apoptosis induction (16.15%) compared to normal control group (11.03%); 14 Ab intervention moderately but statistically significantly alleviated oxLDL-induced apoptosis (13.44%). Treatment of THP-1 cells with just 14 Ab did not have the anti-apoptotic effect in comparison with the control experiment (Figure 1A). In addition, oxLDL treatment enhanced caspase-3 activation, which was reversed by 14 Ab. A similar result was observed regarding the activation of caspase-8 and caspase-9, the two major upstream initiators of caspase-3 in the extrinsic and intrinsic pathway of apoptosis, respectively (Figure 1B). Moreover, 14 Ab treatment restored the mitochondrial transmembrane potential, inhibited ROS production (characterized by the fluorescence intensity of the DCFH-DA), and enhanced ATP content (Figure 1C-E). These results suggest that recombinant humanized IgG1 antibody not only exerts an anti-apoptosis effect on oxLDL-induced apoptosis, but also helps cells to resist oxidative stress in presence of oxLDL.
To investigate the underlying mechanism of recombinant humanized IgG1 antibody in anti-apoptosis and anti-oxidative stress, TMT-labeled quantitative proteomics were applied to compare the protein expression profiles between any two groups (control, oxLDL, and oxLDL + 14 Ab) in isolated and purified human peripheral blood CD14+ monocytes. A total of 5116 proteins were identified, of which 4428 proteins were quantitated. The property and functions of qualified proteins were classified with Gene Ontology (GO) and Cluster analysis of GO, KEGG pathway and protein domain (Supplementary Materials). Only when the variation of those proteins’ abundance was more than 1.2 times and t-test p-value < 0.05, were they accepted as differentially expressed proteins. The number of upregulated and downregulated proteins in pairwise comparison was shown in Figure 2A. Finally, we found 16 proteins that were downregulated in oxLDL treatment compared to normal control group, but upregulated in oxLDL plus 14 Ab group compared to oxLDL only group (Figure 2B). Among the 16 candidates, methionine sulfoxide reductase A (MSRA), which functions as a repair enzyme for proteins that have been inactivated by oxidation, aroused our great interest (Table 1). We further verified that 14 Ab treatment along with oxLDL in THP-1 cells significantly reversed the reduction of MSRA protein expression induced by oxLDL only, which is consistent with the proteomics sequencing results (Figure 3A).
We then designed siRNA to block the endogenous MSRA expression to determine whether MSRA is involved in apoptosis regulation. As demonstrated in Figure 3B and C, the endogenous MSRA protein expression and mRNA transcription was knocked down when applied to MSRA siRNA-296, so we chose MSRA siRNA-296 for subsequent working conditions. As shown in Figure 3D–E, transfection with MSRA siRNA-296 increased cell apoptosis and reduced mitochondrial membrane potential. Meanwhile, the enhancement of ATP content induced by 14 Ab was potently downregulated due to the reduced expression of MSRA (Figure 3E). These results suggested that the anti-apoptosis and anti-oxidant stress properties of 14 Ab are mediated by the upregulation of MSRA when cells are incubated with oxLDL.
Evidence has shown that the transcription factor FOXO1 regulates a number of cellular processes, such as cell cycle progression, apoptosis and oxidative stress [15]. We therefore investigated whether FOXO1 participates in the effects of 14 Ab on MSRA expression and MSRA-mediated anti-apoptosis and anti-oxidant stress. The results showed that 14 Ab significantly enhanced the protein levels of FOXO1 in THP-1 cells (Figure 4A). Thereafter, THP-1 cells were transfected with FOXO1 siRNA for 48 h, followed by adding 14 Ab along with oxLDL. As indicated in Figure 4B and C, transfection of THP-1 cells with FOXO1 siRNA reduced the protein level of FOXO1, and almost abolished the amplification of 14 Ab on MSRA expression. Moreover, we cloned 4 different lengths of the MSRA promoter region into pGL3-basic vector to obtain the human MSRA reporter and co-transfected into the HepG2 cells together with renilla luciferase control, as THP-1 cells were difficult to transfect transiently. The results indicated that the 1683bp MSRA reporter obtained the highest luciferase activity, followed by 1378bp MSRA reporter compared with 2166bp MSRA reporter (Figure 4D). Next, we transfected the 1683bp MSRA reporter into HepG2 cells for 24 h, followed by adding oxLDL with or without 14 Ab for another 24 h, and found that 14 Ab treatment significantly restored the reduction of MSRA promoter activity induced by oxLDL (Figure 4E). However, the mutation of the FOXO1 binding site on 1683bp MSRA reporter predicted by JASPAR website almost eliminated the luciferase activity, even though in the presence of 14 Ab (Figure 4E). These results suggested that 14 Ab enhanced MSRA expression in the presence of oxLDL is regulated by the transcription factor FOXO1.
Since the acetylation modification attenuates the ability of FOXO1 to bind DNA and suppress transcription [20,21], we examined the FOXO1 acetylation in oxLDL-induced apoptosis. As shown in Figure 5A and B, THP-1 cells exposed to oxLDL manifested remarkably increased FOXO1 acetylation, while treatment with recombinant humanized IgG1 antibody could reverse the acetylation modification of FOXO1 induced by oxLDL. Moreover, compared with oxLDL treatment alone, recombinant humanized IgG1 antibody treatment clearly elevated the nuclear level of FOXO1 but reduced the nuclear level of FOXO1 acetylation oxLDL (Figure 5C). Interestingly, SIRT1, a deacetylase responsible for FOXO1 deacetylation [22], was upregulated by recombinant humanized IgG1 antibody (Figure 5D). Importantly, we also found that EX-527, which blocks SIRT1 activity, decreased the expression of FOXO1 and MSRA (Figure 5D, H–J), as well as increased the level of acetylated FOXO1 (Figure 5E and F). In addition, knockdown of SIRT1 exacerbated FOXO1 acetylation (Figure 5G). These findings strongly suggest that the increased FOXO1 acetylation result from inhibition of SIRT1 is involved in oxLDL-induced decline of MSRA expression. Consistently, when there is knockdown of SIRT1 or FOXO1, oxLDL-induced cell apoptosis was further increased, even in the presence of recombinant humanized IgG1 antibody (Figure 6). Similarly, the recombinant humanized IgG1 antibody could not inhibit the decreased mitochondrial membrane potential induced by oxLDL upon knockdown of SIRT1 or FOXO1 (Figure 6). Consequently, we concluded that the upregulation of MSRA by recombinant humanized IgG1 antibody could inhibit oxLDL-induced apoptosis via a SIRT1-FOXO1 axis.
Intracellular redox status is tightly regulated by oxidant and antioxidant systems. An imbalance between these systems causes ROS accumulation which leads to oxidative stress and inflammation. Oxidized low-density lipoprotein (oxLDL)-induced oxidative stress and apoptosis are considered as critical contributors to cardiovascular diseases such as atherosclerosis. Various studies have shown that oxLDL promotes ROS generation in endothelial cells, vascular smooth muscle cells, and macrophages. oxLDL induces pro-inflammatory responses, pro-oxidative conditions and endothelial cell apoptosis. Macrophages are considered immune sentinels and play a role in maintaining tissue homeostasis. Histological examination of human atherosclerotic plaques has confirmed that macrophage subsets are associated with plaque progression [23]. Several drugs that inhibit atherosclerosis by targeting macrophages have been reported [24]. Regulation of macrophages may thus be a therapeutic strategy for atherosclerosis [25]. Our previous study had demonstrated that an antibody may reduce atherosclerosis by activating monocyte/macrophage polarization in ApoE−/− mice [19]. In this paper we have further explored the anti-apoptotic effect of the antibody on monocytes/macrophages. MSRA, one of the antioxidant defenses in cells, is important in the maintenance of redox homeostasis and in the prevention of oxidative stress-related disease. It has been confirmed that MSRA plays a protective role against hypoxia/reoxygenation-induced cell death in cardiac myocytes [26] and neuronal cells [9]. MSRA mutations or deficiency in E. coli and yeast were particularly sensitive to oxidative damage [27,28]. MSRA knockout mice are highly sensitive to oxidative stress and show nerve damage and shortened lifespans [14], whereas MSRA transgenic Drosophila show significantly enhanced anti-oxidation and anti-aging characteristics [11]. Our study demonstrates that recombinant humanized IgG1 antibody inhibits ROS production in THP-1 cells caused by oxLDL as well as protects against apoptosis. The protective functions of recombinant humanized IgG1 antibody are attributed to an increased MSRA expression, indicating that the antioxidant and anti-apoptotic effects of recombinant humanized IgG1 antibody are closely associated with MSRA. In previous studies, it was shown that the expression of MSRA is mainly regulated by transcription factor Forkhead box group O 3a (FOXO3a) [29,30,31]; this is a member of the family of Forkhead transcription factors, which controls the expression of many endogenous antioxidant-encoding genes, including manganese superoxide dismutase and catalase [32,33]. However, our results show that FOXO1, another member of the family of Forkhead transcription factors, also participates in the regulation of MSRA expression in THP-1 cells treated with recombinant humanized IgG1 antibody. As FOXO proteins are tightly regulated to determine cell survival and cell death responsive to specific environmental conditions, we also confirm that knockdown of endogenous FOXO1 exacerbates THP-1 cells apoptosis. The activity of FOXO1 is strictly regulated by modifications on its protein, which ensures that transcription of its downstream target genes is tightly responsive to environmental signals. Phosphoinositide 3-kinase/protein kinase B (PI3K/PKB) phosphorylates FOXO1 protein results in disrupted interactions between the FOXO1 protein and its target DNA and lead to the translocation of the FOXO1 protein from the nucleus to the cytoplasm, thus suppressing FOXO1-dependent transcription. C-Jun N-terminal kinase (JNK) can also phosphorylate FOXO1, which results in the import of the FOXO1 protein from the cytoplasm to the nucleus, thereby antagonizing the action of PI3K/PKB. FOXO1 protein activity is also regulated by reversible acetylation modification. Acetylation of FOXO1 attenuates the ability of FOXO1 to bind DNA and suppress transcription. Interestingly, it was also found that acetylation regulated the function of FOXO1 by influencing its sensitivity for phosphorylation. Our data show that recombinant humanized IgG1 antibody decreased acetylation of FOXO1 induced by oxLDL, which in turn increased FOXO1 transcriptional activity. FOXO1 acetylation is well known to be regulated by SIRT1, an NAD+-dependent class III deacetylase and widely found in human tissues and cells [17,20,34]. It has been reported that SIRT1 deacetylated FOXO1 and decreased H2O2-induced granulosa cell apoptosis [35]. Curcumin alleviates oxidative stress and inhibits apoptosis in diabetic cardiomyopathy via SIRT1-FOXO1 pathways [36]. In the present study, we also found that SIRT1 deacetylated FOXO1 in THP-1 cells upon recombinant humanized IgG1 antibody treatment, and this was suppressed by SIRT1 inhibitor EX-527. Moreover, SIRT1 deficiency impaired the protective effect of recombinant humanized IgG1 antibody against apoptosis, suggesting the SIRT1-FOXO1 axis as a potential target of recombinant humanized IgG1 antibody against oxidative stress and anti-apoptosis. We know that the antibody regulates immune responses through interacting with Fc receptors [37]. There are 4 Fc receptors of IgG, namely Fcgamma-RI, -RIIa, -RIIb, and -RIII. FcgammaRIIb is the only one which has negative signal transduction with immunoreceptor tyrosine-based in-activation motif (ITIM), while the others have a ITAM (immunoreceptor tyrosine-based activation motif) with a positive signaling transduction. We previously reported that the antibody inhibits oxLDL-induced macrophage MCP-1 release through the FcgammaRIIb signal transduction pathway [38]. Thus, whether FcgammaRIIb and its downstream signal transduction participates in the regulation of MSRA expression as well as reduction of oxLDL-induced apoptosis upon 14 Ab treatment, requires further exploration. In conclusion, our findings for the first time imply that recombinant humanized IgG1 antibody upregulates MSRA and inhibits oxLDL-induced THP-1 cells’ apoptosis as a result of increased SIRT1 expression and decreased FOXO1 acetylation. This reveals the critical role of the SIRT1-FOXO1 axis in recombinant humanized IgG1 antibody mediated upregulation of MSRA expression and anti-apoptosis.
Dulbecco’s modified Eagle’s medium (DMEM), RPMI 1640 medium, fetal bovine serum (FBS) and phosphate-buffered saline (PBS) were purchased from Invitrogen (Carlsbad, CA, USA). The TRIzol Reagent was from Takara Bio Inc. (Kusatsu, Japan). Antibody against MSRA, FOXO1 and SIRT1 was purchased from Proteintech (Wuhan, China), Cell Signaling Technology (Danvers, MA, USA) and Abcam (Cambridge, UK), respectively. Anti-acetyl Lysine antibody was from Abcam (Cambridge, UK). Inhibitor of SIRT1 (EX-527) was from MCE (Shanghai, China). Annexin V-FITC Apoptosis Detected Kit was purchased from BD Biosciences (Franklin Lakes, NJ, USA). Mitochondrial membrane potential assay kit and ATP Assay Kit was from Beyotime (Beijing, China). DCFH-DA were purchased from Invitrogen (CA, USA). The specific siRNA for MSRA, FOXO1 and SIRT1 were designed and synthesized by GenePharma (Suzhou, China). OxLDL was purchased from Yiyuan (Guangzhou, China). The recombinant humanized IgG1 antibody (clone: NO. 14, namely 14 Ab) was produced and maintained by our laboratory, and the detailed properties of the antibody were determined in our previous study [19]. THP-1 cells were maintained in our lab.
Human peripheral blood mononuclear cells (PBMCs) were heparinized and layered over Ficoll-Isopaque (Pharmacia, Freiburg, Germany) density gradient reagent according to the manufacturer’s instructions. Mononuclear cells were separated by centrifugation at 400 g for 30 min at room temperature. Mononuclear cells were collected and washed two times with PBS without Ca2+ and Mg2+ by centrifugation at 250 g for 20 min at 4 °C. Cells were then diluted with complete RPMI 1640 medium. Human monocytes were purified with MACS CD14 microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s instructions.
CD14+ monocytes and THP-1 cells (human monocytic leukemia cell line) were grown in RPMI 1640 medium supplemented with 10% FBS under standard culture conditions (5% CO2, 37 °C), and exposed to oxLDL (50 µg/mL) [39] with or without recombinant humanized IgG1 antibody (100 µg/mL) [19] for 48 h. The THP-1 cells were then collected for extraction of proteins to analyze MSRA and FOXO1 expression, while CD14+ monocytes were collected for proteomics analysis.
Cells were sonicated three times on ice using a high intensity ultrasonic processor (Scientz, Ningbo, China) in lysis buffer (8 M urea, 1% Protease Inhibitor Cocktail). The remaining debris was removed by centrifugation at 12,000 g at 4 °C for 10 min. Finally, the supernatant was collected and the protein concentration was determined with BCA kit according to the manufacturer’s instructions. For digestion, the protein solution was reduced with 5 mM dithiothreitol for 30 min at 56 °C and alkylated with 11 mM iodoacetamide for 15 min at room temperature in darkness. The protein sample was then diluted by adding 100 mM TEAB to urea concentration less than 2 M. Finally, trypsin was added at 1:50 trypsin-to-protein mass ratio for the first digestion overnight and 1:100 trypsin-to-protein mass ratio for a second 4 h-digestion. After trypsin digestion, peptide was desalted by Strata X C18 SPE column (Phenomenex, Torrance, CA, USA) and vacuum-dried. Peptide was reconstituted in 0.5 M TEAB and processed according to the manufacturer’s protocol for TMT kit. Briefly, one unit of TMT reagent was thawed and reconstituted in acetonitrile. The peptide mixtures were then incubated for 2 h at room temperature and pooled, desalted and dried by vacuum centrifugation. The tryptic peptides were fractionated into fractions by high pH reverse-phase HPLC using Thermo Betasil C18 column (5 μm particles, 10 mm ID, 250 mm length). Briefly, peptides were first separated with a gradient of 8% to 32% acetonitrile (pH 9.0) over 60 min into 60 fractions. Then, the peptides were combined into 18 fractions and dried by vacuum centrifuging. The tryptic peptides were dissolved in 0.1% formic acid (solvent A), directly loaded onto a home-made reversed-phase analytical column (15-cm length, 75 μm i.d.). The gradient was increased from 6% to 23% solvent B (0.1% formic acid in 98% acetonitrile) over 26 min, 23% to 35% in 8 min, climbing to 80% in 3 min, then holding at 80% for the last 3 min, all at a constant flow rate of 400 nL/min on an EASY-nLC 1000 UPLC system. The peptides were subjected to NSI source followed by tandem mass spectrometry (MS/MS) in Q ExactiveTM Plus (Thermo, Waltham, MA, USA) coupled online to the UPLC.
The Gene Ontology (GO) annotation proteome, containing cellular component, molecular function and biological processes, was derived from the UniProt-GOA database (www.http://www.ebi.ac.uk/GOA/, accessed on 10 January 2022). Identified proteins domain functional descriptions were annotated by InterProScan (a sequence analysis application) based on protein sequence alignment method, and the InterPro domain. Proteins were classified by GO annotation into three categories: biological process, cellular compartment and molecular function. For each category, a two-tailed Fisher’s exact test was employed to test the enrichment of the differentially expressed protein against all identified proteins. The GO with a corrected p-value < 0.05 is considered significant. The Encyclopedia of Genes and Genomes (KEGG) database was used to identify enriched pathways by a two-tailed Fisher’s exact test, to test the enrichment of the differentially expressed proteins against all identified proteins.
Western blot analysis was conducted as described previously [40]. Briefly, the total proteins were extracted, after which their concentrations were determined using a BCA kit (Beyotime; Beijing, China). The proteins (20 μg per lane) were subsequently separated by 10% SDS-PAGE and transferred to a PVDF membrane, which was immuno-blotted using antibodies against GAPDH, MSRA, FOXO1 and SIRT1, respectively. Following a series of rinses in TBS-T, the membrane was incubated with a peroxidase-conjugated secondary antibody. Lastly, the proteins were visualized by enhanced chemiluminescence (ECL; Merck Millipore, Darmstadt, Germany), and the relative expression levels were assessed by Image J via densitometry.
THP-1 cells that had reached approximately 80% confluence were transfected with small-interfering RNA (siRNA) specific for SIRT1, FOXO1 and MSRA, respectively, using LipofectamineTM 3000 (Invitrogen), according to the instructions of manufacturers. After transfection for 48 h, cells were harvested for RT-qPCR and Western blot analysis to evaluate the silencing efficiency or after treatment with recombinant humanized IgG1 antibody (100 µg/mL) with or without oxLDL (50 µg/mL) for another 24 h to detect apoptosis and mitochondrial membrane potential by FACS.
A 2166 bp fragment of human MSRA promoter was amplified by PCR from the genome of the THP-1 cells and then cloned into the reporter vector pGL3-basic (Promega, Madison, WI, USA); truncated MSRA-Luc vector and mutated MSRA-Luc vector were constructed by the same strategy. Transfection experiments were carried out in 24-well plates using LipofectamineTM 3000 (Invitrogen). 24 h after transfection, HepG2 cells were treated with or without recombinant humanized IgG1 antibody for another 24 h. A Dual-Luciferase Reporter Assay System (Promega) was used to evaluate luciferase activity, in accordance with the manufacturer’s instructions. The data was expressed as relative luciferase activity (firefly luciferase activity/renilla luciferase activity).
All data were obtained from at least three indepent experiments and were analyzed via one-way ANOVA and the Student–Newman–Keuls (SNK) post hoc multiple comparison test by GraphPad Prism 6 software (GraphpadPrism6, GraphPad Software, San Diego, CA, USA). The results are expressed as the mean ± standard deviation (SD). Results with p < 0.05 were considered statistically significant. *, p < 0.05; **, p < 0.01; ***, p < 0.001. | true | true | true |
PMC9569926 | 36233075 | Fuxiu Shi,Xinyue Chen,Yi Wang,Yujie Xie,Junpei Zhong,Kangtai Su,Miao Li,Yuqiu Li,Qing Lin,Youjia Zhou,Jie Wang,Lixia Xiong | HOTAIR/miR-203/CAV1 Crosstalk Influences Proliferation, Migration, and Invasion in the Breast Cancer Cell | 04-10-2022 | breast cancer,HOTAIR,miR-203,CAV1,ceRNA,invasion,migration | In recent years, malignant breast cancer metastasis has caused a great increase in mortality. Research on the genetic and molecular mechanisms of malignant breast cancer has continued to deepen, and targeted therapy has become the general trend. Among them, competing endogenous RNA (ceRNA)-related molecules have received much attention. Homeobox transcript antisense RNA (HOTAIR) has been reported to function extensively as a ceRNA in breast cancer. Notably, miR-203 and Caveolin 1 (CAV1) have also been found to play a role in breast cancer. However, the relationship between the three remains unclear. In this study, we present a new mechanic through bioinformatics tool and basic experiments: the HOTAIR/miR-203/CAV1 axis, which complemented the role network of HOTAIR as a ceRNA, thus, it will provide a novel potential idea for breast cancer research and therapy. | HOTAIR/miR-203/CAV1 Crosstalk Influences Proliferation, Migration, and Invasion in the Breast Cancer Cell
In recent years, malignant breast cancer metastasis has caused a great increase in mortality. Research on the genetic and molecular mechanisms of malignant breast cancer has continued to deepen, and targeted therapy has become the general trend. Among them, competing endogenous RNA (ceRNA)-related molecules have received much attention. Homeobox transcript antisense RNA (HOTAIR) has been reported to function extensively as a ceRNA in breast cancer. Notably, miR-203 and Caveolin 1 (CAV1) have also been found to play a role in breast cancer. However, the relationship between the three remains unclear. In this study, we present a new mechanic through bioinformatics tool and basic experiments: the HOTAIR/miR-203/CAV1 axis, which complemented the role network of HOTAIR as a ceRNA, thus, it will provide a novel potential idea for breast cancer research and therapy.
According to the latest statistics released by the International Agency for Research on Cancer (IARC) in December 2020, breast cancer accounted for 11.7% of new tumor cases in 2020, surpassing lung cancer, which accounted for 11.4% [1]. Breast cancer has become the most common malignant tumor type in the world. Therefore, the prevention, diagnosis, and treatment of breast cancer are urgent. The significance of analyzing breast cancer’s metastasis mechanism and studying the mechanism of malignant proliferation, invasion, and migration of breast cancer cells has reached its peak. Homeobox (HOX) transcript antisense RNA (HOTAIR) was first discovered accidentally in 2007 and is located at 12q13.13, between HOXC11 and HOXC12 on chromosome 12 [2]. Therefore, it is also called intergenic long non-coding RNA (lncRNA), which contains 2158 nucleotides and 6 exons. HOTAIR can regulate gene expression epigenetically in a manner similar to scaffolding and can also interfere with the expression of signaling molecules associated with breast cancer development in breast cancer. Qian L. et al. found that up-regulated HOTAIR in breast tissue can enhance the expression of the DNA damage repair factors and regulate cell cycle and apoptosis [3]. The overexpression of HOTAIR can activate estrogen receptor transcription, inducing breast cancer drug-resistant and promoting cancer cell proliferation [4]. A recent study analyzed the in situ expression of HOTAIR in breast cancer patients and showed that a high expression of HOTAIR in tumor tissue is closely related to lymph node metastasis [5]. A population-based case-control study found that HOTAIR is associated with breast cancer risk and prognosis [6]. Based on the important role of HOTAIR in breast cancer progression, a recent review summarized the potentially promising therapeutic function of HOTAIR in breast cancer [7]. However, the underlying mechanism of HOTAIR in breast cancer requires more research. Caveolin 1 (CAV1) is the gene encoding for caveolin-1. The expression of the CAV1 gene is associated with many human diseases, including pulmonary hypertension, hypertriglyceridemia, and cancer. It has been reported that CAV1 plays a key role in breast cancer cell proliferation, apoptosis, autophagy, invasion, migration, and breast cancer metastasis [8,9,10,11,12]. Another study confirmed the absence of CAV1 gene mutations in human breast cancer [13]. It is only that the expression level of CAV1 is regulated by other factors, and it plays a dual role in inhibition or promotion in the process of breast cancer [14]. In the study of mechanisms in breast cancer, CAV1 has been our target, and we strive to identify the key molecules that act on it. It has been reported that lncRNAs regulate mRNA expression by binding specific microRNAs (miRNAs) as competing endogenous RNAs (ceRNAs) [15]. Therefore, the physical interaction of ceRNA-miRNAs leads to a complex regulatory network of ceRNA/miRNA/mRNA that controls gene expression at the transcriptional, post-transcriptional, and post-translational levels [16,17]. Some studies have found that these interacting ceRNA/miRNA/mRNA networks are involved in many cellular processes and human diseases, including tumorigenesis [18,19,20]. The ceRNA/mRNA network has also been reported in the study of breast cancer mechanisms. Using a weighted gene co-expression network analysis (WGCNA) algorithm, they found a ceRNA- lncRNA TRPM2 that promotes the proliferation of BRCA cells and inhibits apoptosis through the TRPM2-AS/miR-140-3p/PYCR1 axis [21]. This further inspires us to explore the miRNA that can connect lncRNAHOTAIR and CAV1 and improve the role of the ceRNA/mRNA network in the progression of breast cancer. In this study, we first analyzed the differential expression of HOTAIR in breast cancer and normal tissues by using database and online messaging tools, screened the possible HOTAIR-binding miRNA, enriched some of the miRNAs, verified the role of miR-203 in cancer, and chose the miR-203 which we are interested in. Then, we analyzed the key protein molecules and their corresponding genes by protein–protein interaction (PPI) and found that CAV1 was included in the PPI network. The dual-luciferase reporter gene, cell proliferation, migration, and invasion in vitro were used to verify the sponge effect of HOTAIR on miR-203 and the regulation of miR-203 on CAV1 expression. Finally, the role of the HOTAIR/miR-203/CAV1 axis in the proliferation, invasion, and migration of the breast cancer cell line MDA-MB-231 was clarified.
To reveal the function of HOTAIR in cancer, we first used GEPIA (http://gepia.cancer-pku.cn/detail.php, accessed on 12 March 2022) to analyze the expression of HOTAIR in different cancers. The results showed that, compared with normal tissues, HOTAIR was highly expressed in breast cancer (Figure 1A). Through the UALCAN (http://ualcan.path.uab.edu/index.html, accessed on 12 March 2022) online tool to analyze the TCGA database, the results showed that HOTAIR was significantly overexpressed in luminal breast cancer, HER2-positive breast cancer, and triple-negative breast cancer (TNBC) (Figure 1B). Based on the subclasses, the UALCAN results showed that HOTAIR was overexpressed in breast invasive carcinoma compared with normal tissues (Figure 1C). The Kaplan–Meier curves and log-rank analysis results suggested that patients with a high expression of HOTAIR may have a worse prognosis (Figure 1D).
The expression of HOTAIR was higher in MDA-MB-231 than in MCF-7 (Figure 2A); thus, we selected a highly invasive breast cancer cell line, MDA-MB-231 (a type of TNBCcells), for cell transfection in the subsequent experiments. The results of CCK-8 showed that the proliferation rate of down-regulated HOTAIR cells decreased, while the proliferation rate of overexpressed HOTAIR cells increased significantly (Figure 2B). The transwell assay results showed that the ability of migration and invasion in overexpressed HOTAIR cells was significantly increased, while in down-regulated HOTAIR cells, it was decreased (Figure 2C). The wound healing assay results revealed that HOTAIR could strikingly enhance the migration of MDA-MB-231 (Figure 2D). These results all proved that the up-regulation of HOTAIR could promote the proliferation, migration, and invasion of MDA-MB-231 cells.
To study the relation between HOTAIR (as ceRNA) and miR-203 in breast cancer, we predicted the potential miRNAs targets of HOTAIR by using ENCORI (StarBase, https://starbase.sysu.edu.cn/index.php, accessed on 12 March 2022) and DAVID (https://david.ncifcrf.gov/tools.jsp, accessed on 12 March 2022). We exported the data (Table S1) from the two databases. Software Cytoscape 3.0 was used to construct a network map of miRNAs that may bind to HOTAIR (Figure 3A). Of note, miR-203 was one of the targets (Figure 3A). Then, we performed GO functional annotation and pathway enrichment analysis using miEAA 2.0 (https://ccb-compute2.cs.uni-saarland.de/mieaa, accessed on 12 March 2022) to examine the functions of these miRNAs. The GO analysis was conducted at three sublevels: biological process (BP), molecular function (MF) and molecular composition (CC). The results of MF and CC are the same. The results showed that the overexpression of miR-203 was significantly enriched in the regulation of angiogenesis, cell proliferation, and apoptosis, such as angiogenesis, blood vessel morphogenesis, and the positive regulation of the cell population in the BP category (Figure S1); negative regulation of anoikis, wound healing, B-cell lymphoma-2 family protein complex, and the extrinsic apoptotic signaling pathway in the MF and CC categories (Figures S2 and S3). KEGG’s cell signaling pathway was also conducted. The miR-203-enriched KEGG pathways were MicroRNAs in cancer, pathways in cancer and Adherens junction (Figure S4). Compared with normal tissues, miR-203 was overexpressed in breast invasive carcinoma through the UALCAN analysis of the TCGA database (Figure 3B). Moreover, miR-203 was significantly overexpressed in luminal breast cancer, HER2-positive breast cancer, and TNBC (Figure 3C). The Kaplan–Meier analysis suggested that patients with a high expression of miR-203 may have a worse prognosis (Figure 3D). From the above analysis, it can be concluded that miR-203 plays an important role in the proliferation, survival, migration, and invasion of cells, especially cancer cells. Notably, these roles of miR-203 are similar to HOTAIR. Therefore, we have reason to believe that HOTAIR is one of the ceRNAs of miR-203. To further verify the binding of HOTAIR-miR-203 in cancer cells, we constructed the wild-type plasmid HOTAIR-wt of HOTAIR that contains the complementary sequence of miR-203, and the mutant plasmid HOTAIR-mu with the mutation of the complementary binding sequence (Figure 3E). Compared with the NC group, the luciferase activity of the miR-203 mimics + HOTAIR-wt group was decreased, while the miR-203 mimics + HOTAIR-mu group luciferase activity changes were less pronounced (Figure 3F), showing the existence of binding between HOTAIR and miR-203.
To further verify the role of HOTAIR/miR-203 in breast cancer progression, we transfected the MDA-MB-231 cells with the HOTAIR plasmid. The RT-qPCR results showed that the expression level of miR-203 was significantly increased when HOTAIR was down-regulated, and the expression level of miR-203 was significantly decreased when HOTAIR was up-regulated (Figure 4A). The results of the CCK assay showed that inhibiting the expression of miR-203 would promote the proliferation of MDA-MB-231 cells, and down-regulating HOTAIR reversed the effect of the miR-203 inhibitor on the proliferation of MDA-MB-231 cells (Figure 4B). The transwell assays showed that the inhibition of miR-203 expression promoted the migration and invasion of MDA-MB-231 cells, and the down-regulation of HOTAIR reversed the effect of the miR-203 inhibitor on the promotion of MDA-MB-231 cell migration and invasion (Figure 4C). The wound healing assay results showed that inhibiting the expression of miR-203 would promote the migration of MDA-MB-231 cells, and down-regulating HOTAIR reversed the effect of the miR-203 inhibitor on promoting the migration of MDA-MB-231 cells (Figure 4D). The above results all proved that HOTAIR/miR-203 could regulate the proliferation, migration, and invasion of MDA-MB-231 cells.
The last link is missing in our constructed ceRNA/miRNA/mRNA axis. Therefore, GO (GO-Biological Process) in GeneCards (https://www.genecards.org/, accessed on 12 March 2022) for the miR-203 involved in gene silencing by miRNA was analyzed (GO:0035195, Figure 5A). In addition, we used the online tool, miEAA 2.0, to enrich the target genes corresponding to the miRNAs (including miR-203) that may be bound by HOTAIR, and screened all the target genes containing miR-203 in the enrichment (Table S2). Then, STRING (https://cn.string-db.org/, accessed on 12 March 2022) was used to make a network map of the interaction (PPI) between the protein encoded by the target gene and the protein encoded by CAV1 (Figure 5B). Notably, CAV1 connected five edges as one of the nodes (Figure 5B), indicating that CAV1 may be in the regulatory network of miR-203. Therefore, we transfected the MDA-MB-231 cells with miR-203 mimics and miR-203 inhibitors to verify the effect of miR-203 on CAV1 expression. The Western blot results showed that when the expression of miR-203 was down-regulated, the expression level of caveolin-1 was significantly increased. When the expression of miR-203 was up-regulated, the expression level of caveolin-1 was decreased (Figure 5C). To further verify whether miR-203 binds to CAV1, we constructed the wild-type plasmid CAV1-wt that contains the complementary sequence of miR-203, and mutant plasmid CAV1-mu, with a mutation of the complementary binding sequence (Figure 5D). The experimental results obtained by the dual-luciferase reporter gene assay are shown in Figure 5E. Compared with the NC group, transfection miR-203 could significantly down-regulate the luciferase activity of the CAV1-wt group. However, transfection of miR-203 could not significantly change the luciferase activity of the CAV1-mu group, which proved that miR-203 could directly bind to CAV1. Therefore, miR-203 could interact with CAV1, and the overexpression of miR-203 could inhibit the expression of CAV1.
We further investigated, in vitro, that miR-203/CAV1 regulated the proliferation, invasion, and migration of MDA-MB-231 cells. The results of CCK-8 showed that the overexpression of CAV1 could promote the proliferation of MDA-MB-231 cells, and up-regulation of the expression of miR-203 could reverse the promoting effect of the overexpression of CAV1 on the proliferation rate of MDA-MB-231 cells (Figure 6A). The transwell assay showed that the overexpression of CAV1 could promote the migration and invasion of MDA-MB-231 cells, and up-regulation of miR-203 reversed the promoting effect of CAV1 overexpression on the migration and invasion ability of MDA-MB-231 cells (Figure 6B). The wound healing assay showed that the overexpression of CAV1 could promote the migration of MDA-MB-231 cells, and up-regulation of miR-203 reversed the promotion of CAV1 overexpression on the migration ability of MDA-MB-231 cells (Figure 6C). The above in vitro experiments all proved that the overexpression of CAV1 could promote the progression of breast cancer cells, and the interaction of miR-203/CAV1 could regulate the proliferation, invasion, and migration of MDA-MB-231 cells.
LncRNA is a type of non-coding RNA, exerting as ceRNA. In recent years, the role of the ceRNA networks in cancer has been extensively studied, linking the functions of protein-coding mRNAs to those of non-coding RNAs [17]. miRNAs are also non-coding RNAs with a length of about 22 nt and play a crucial role in regulating gene expression [22]. They exert regulatory effects by binding to the 3′-untranslated region of target mRNAs, leading to mRNA degradation or silencing in a sequence-specific manner [22]. Given that any transcript containing miRNA-response elements could theoretically function as a ceRNA, they likely represent a broad form of post-transcriptional regulation of gene expression in physiology and pathology. HOTAIR is the first lncRNA discovered to have trans-regulatory roles [2]. Since the discovery of HOTAIR, multiple studies have elucidated its critical role in tumor growth, apoptosis, invasion, metastasis, tumor stem cell differentiation, and drug resistance. Notably, HOTAIR has been reported to function extensively as a ceRNA in breast cancer. Zhao W. et al. found that HOTAIR affected MDA-MB-231 cell growth and apoptosis through the miR-20a-5p/HMGA2 pathway [23]. Wu D. et al. found that the HOTAIR/miR-129-5p/FZD7 axis could accelerate breast cancer progression [24]. In addition, HOTAIR could increase radioresistance in breast cancer by promoting HSPA1A expression through binding to miR-449b-5p in MDA-MB-231 cells [25]. Breast cancer studies are limited to a few cell lines, with MCF-7, T47D, and MDA-MB-231 accounting for the majority of studies [26]. Notably, MDA-MB-231 is a highly aggressive and poorly differentiated triple-negative breast cancer cell line [27]. In recent years, articles on tumor metastasis, drug resistance, and treatment based on MDA-MB-231 have been emerging. Research on HOTAIR has been studied in MDA-MB-231. It is of interest to continue to study MDA-MB-231 to construct a more complete network of actions of HOTAIR. Additionally, in our study, we found that the level of HOTAIR was higher in MDA-MB-231 than in MCF-7 (Figure 2A). In addition, the results (Figure 1C) showed that HOTAIR was higher in TNBC, as known to all, that MDA-MB-231 is a type of TNBC cell. Therefore, we chose the MDA-MB-231 cell line. In this study, we demonstrated that the overexpression of HOTAIR significantly promoted the proliferation, invasion, and migration in MDA-MB-231 cells, which was consistent with the predictions using bioinformatics analysis andprevious research. Further, miR-203, a member of the miRNA family, has been confirmed to be involved in regulating the proliferation, differentiation, metastasis, invasion, and apoptosis of tumor cells [28,29]. Currently, only three studies have addressed the role of miR-203′s link with HOTAIR in cancer [30,31,32]. The connection between the two deserves further investigation. In this study, we predicted the role of miR-203 in biological and cancer processes through database information. Notably, HOTAIR can compete for binding to miR-203, which is a molecular sponge of miR-203. The dual-luciferase reporter gene assay showed the combination of the two. Furthermore, we demonstrated that the expression of miR-203 was significantly decreased after the overexpression of HOTAIR. The inhibition of HOTAIR reversed the promotion of breast cancer cell progression induced by miR-203 down-expression. That is, HOTAIR could reverse the inhibitory effect of miR-203 on the proliferation, invasion, and migration of MDA-MB-231 cells. Notably, through the TCGA database, we analyzed the expression of miR-203 in breast cancer and found that the expression of miR-203 was different in different breast subtypes. However, its expression was higher than that in normal tissues. In addition to our results, other existing experimental data showed that miRNA was used as a tumor suppressor, such as miR-203, identified as a transcript of matrix stiffness inhibition, which was negatively correlated with the malignant proliferation of breast epithelium [33]. Further, miR-203 also inhibited the migration ability of MDA-MB-231 cells by targeting PRKCQ [34]. There are also studies that directly conclude that LncRNA DLG1-AS1 may promote cancer cell proliferation in TNBC by down-regulating the tumor suppressor, miR-203 [35]. This suggests that the role of miR-203 in breast cancer is intriguing, and the mechanism behind its expression, regulation by other molecules, and tumor inhibition is worthy of further exploration. Recently, CAV1 has been found to be actively involved in human tumor progression [36]. On the one hand, high expression of CAV1 has been reported to drive tumorigenesis by inhibiting apoptosis and promoting anchorage-independent growth, drug resistance, and metastasis [37,38,39]. For example, CAV1 expression in pancreatic cancer cells was found to be positively correlated with cachectic states [40]. On the other hand, CAV1 acts as a tumor suppressor in some cases, as its low expression favors tumor progression. For instance, Geletu et al. found that CAV1 down-regulation accelerated the proliferation of lung cancer cells via the Cadherin-11/Stat3 axis [41]. Notably, in recent studies, the dual role of CAV1 is evident in breast cancer progression. Loss of stromal CAV1 in human breast cancer was associated with a poor clinical outcome [42]. CAV1 could suppress breast cancer development [43]. However, Dong et al. found that a high expression of CAV1 promoted the migration and invasion of breast cancer cells, supporting its oncogenic role [9]. The dual role of CAV1 in breast cancer deserves further exploration. Moreover, the regulation of CAV1 expression also remains poorly understood. Therefore, our team has been devoted to the related research of CAV1, trying to elucidate the mechanism of CAV1 in breast cancer progression. Note, however, that the role of CAV1 in association with miR-203 in cancer has only been elucidated in renal cell carcinoma [44]. In this study, the down-regulation of CAV1 expression attenuated the proliferation, invasion, and migration of the breast cancer cell MDA-MB-231. We used bioinformatics technology to find the miR-203 molecule that can affect the expression of CAV1. The binding of CAV1 and miR-203 was further verified by basic experiments. Moreover, the overexpression of miR-203 significantly down-regulated CAV1’s expression and reversed the promotion effect of the MDA-MB-231 cells’ progression induced by CAV1.
Gene Expression Profiling Interactive Analysis (GEPIA; http://gepia.cancer-pku.cn/detail.php, accessed on 12 March 2022) [45] was used to analyze the expression of HOTAIR in invasive breast cancer. UALCAN (http://ualcan.path.uab.edu/index.html, accessed on 12 March 2022) was used to analyze the differential expression of HOTAIR and miR-203 in breast cancer subtypes and survival analysis in the TCGA database [46]. The ENCORI (StarBase, https://starbase.sysu.edu.cn/index.php, accessed on 12 March 2022) database [47] and DAVID (https://david.ncifcrf.gov/tools.jsp, accessed on 12 March 2022) database [48] were used to analyze the miRNAs interacting with HOTAIR. Cytoscape (3.9.0) was used to draw the HOTAIR-miRNA network map and select the key molecules on the network map. Gene Ontology/Kyoto Encyclopedia of Genes and Genomes (GO/KEGG) analysis on the miRNA set was performed by using the miRNA enrichment analysis and annotation tool, miEAA 2.0, online (https://ccb-compute2.cs.uni-saarland.de/mieaa, accessed on 12 March 2022) [49], and GeneCards (https://www.genecards.org/, accessed on 12 March 2022). Then, enrich the related target genes of miR-203, and use the STRING (https://cn.string-db.org/, accessed on 12 March 2022) database [50] for the proteins encoded by the target genes to export the related protein–protein interaction (PPI) network. Finally, combined with the CAV1, we focused on and analyzed whether it has a role with the above proteins.
The human breast cancer cell line (MDA-MB-231and MCF-7) was purchased from Shanghai Cell Bank, the Chinese Academy of Sciences. The human embryonic kidney cell line (HEK-293T) was provided by Shanghai Hanheng Biotechnology Co., Ltd., Shanghai, China. The cells were cultured in DMEM (high glucose) (Hyclone, Logan, UT, USA), supplemented with 10% fetal bovine serum (FBS) (BI, KibbutzBeit Haemek, Israel) and 1% penicillin-streptomycin (Solarbio, Beijing, China) in a cultured environment of 37 °C constant temperature, 5% CO2 incubator.
Before transfection, 3.5 × 105 cells were seeded in each well in a six-well plate. When the cells grew to about 70% of the bottom of the culture dish, Lipofectamine 3000 (Thermo Fisher, 81 Wyman Street, Waltham, MA, USA) and Opti-MEM serum-free medium (Gibco, Grand Island, NY, USA) were used for HOTAIR overexpression plasmid (HOTAIR homo pcDNA3.1, oeHOTAIR). Further, siRNA (siRNA 536, siHOTAIR), miRNA mimics (miR-203 mimics), miRNA inhibitor (miR-203 inhibitor), CAV1 overexpression plasmid (CAV1 homo pcDNA3.1, oeCAV1), siRNA (siRNA710, siCAV1), negative control (NC, down-regulated NC is siCON, overexpressed NC is oeCON) were prepared as a transfection complex and added to six wells. The cells were transfected in the plate. The above pcDNA and siRNA were constructed by Suzhou Gene, and the cells were collected after 24 h to extract total RNA to determine the transfection efficiency by qPCR. The sequences are shown in Table S3.
Total RNA was isolated from the cells 24 h post-transfection after RNA quantification according to the kit instructions (Omega, Madison, WI, USA), and the reaction system was prepared according to the instructions. RNA (2–5 ug per system) was reverse transcribed into cDNA by using the PerfectStartTM Green qPCR SuperMIX (TransGen Biotech, Beijing, China) and StepOneTM qRT PCR machine (ABI, Waltham, MA, USA) real-time device under the following conditions: 42 °C for 30 min, 85 °C for 5 s. The qPCR system was prepared according to the instructions of the TransStart® Tip Green qPCR SuperMix kit (TransGen Biotech, Beijing, China). The qPCR was performed under the following conditions: 94 °C for 30 s, 40 cycles of 94 °C for 5 s, and 62 °C for 30 s. All results were normalized to U6 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The relative expression level of miRNA was detected by the −2∆∆Ct method. All primer pairs were synthesized by Sangon (Shanghai, China), and the sequences are shown in Table S4
The Cell Counting Kit-8 (CCK-8) (TransGen Biotech, Beijing, China) was used to detect the proliferation of MDA-MB-231, and the cells in the logarithmic growth phase that had been transfected in the early stage were used to inoculate 1000 cells per well in a 96-well plate. More than 4 duplicate wells were set in the group, and the zero-adjustment group only contained medium and CCK. Three 96-well plates were set for 24 h, 48 h, and 72 h. The cells were treated for the first 2 h, washed three times with sterile PBS, and then 90 μL of culture medium and 10 μL of CCK reagent were added to each well. Incubate in the incubator for 1 h, use a microplate reader to measure the OD value of each well at 450 nm and record, and the CCK value is the value obtained by subtracting the blank well from the measured value.
The wound healing assay, as well as the transwell assay, were used to assess cell migration. For the wound healing assay: cells were seeded into triplicate wells of a 6-well plate and cultured to 30–50% confluence, and then a 20 μL pipette tip was used to create artificial scratches. Cell images were collected under a microscope at 0 h, 24 h, and 48 h, and the area of the scratched area was measured by software to calculate the migration rate. For the transwell assay: 24 h after transfection, 1 × 105 cells were inoculated in a 200 μL serum-free medium in the upper Transwell chamber. A total of 500 μL complete medium containing 10% FBS was added into the lower chamber as a chemoattractant. After 48 h of incubation, we used a cotton swab to remove non-invasive cells manually. Subsequently, the cells were fixed in 4% paraformaldehyde for 30 min, stained with crystal violet for 20 min, and then counted under a microscope.
A Transwell chamber (Corning, Corning, NY, USA) with a Matrigel matrix (Corning, Corning, NY, USA) was used to measure the invasive ability of cells. Before cell plating, the Matrigel matrix was prepared with serum-free DMEM at 200 µg/mL, and 100 μL was evenly spread in each chamber. After clotting, 1 × 105 serum-free cultured cells were seeded. An amount of 500 μL complete medium containing 10% FBS was added into the lower chamber as a chemoattractant. After 48 h of incubation, we used a cotton swab to remove non-invasive cells manually. Subsequently, the cells were fixed in 4% paraformaldehyde for 30 min, stained with crystal violet for 20 min, and then counted under a microscope.
The HEK-293T cells were seeded in 96 wells of human embryonic kidney cells, and the density reached 50–70%. An amount of 10 µL DMEM, 0.16 µg HOTAIR/CAV1 target plasmid (HanBIO, Shanghai, China), and 5 pmol miR-203 mimics/negative control (NC) were mixed. After thorough mixing (Solution A), let this stand for 5 min at room temperature. Mix 10 µL DMEM with 0.3 µL transfection reagent (HanBIO, Shanghai, China, 0.8 mg/mL) thoroughly (Solution B), and let this stand for 5 min at room temperature. Mix solution A and solution B well and leave them at room temperature for 20 min. Add the transfection mixture to the cell culture medium and mix at 37 °C, 5% CO2. After 6 h of transfection, the fresh medium was changed, and 48 h after transfection, the cells were collected with a kit (Promega Dual-Luciferase, Madison, WI, USA) to measure the luminescence value of the reporter gene.
Forty-eight hours after the transfection of MDA-MB-231 cells, cell lysates were collected using 100 μL of RIPA lysate (Applygen, Beijing, China) and 10 μL of PMSF per well of a six-well plate. The protein concentration of cell lysates was measured with a BCA protein assay kit (Applygen, Beijing, China). Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and electrotransferred to nitrocellulose (PVDF) (Millipore, Boston, MA, USA) membranes for Western blotting. The membrane was blocked with 5% nonfat dry milk (BD, Franklin, LA, USA) for 2 h, and then diluted with primary antibody caveolin-1 (Affinity, Shanghai, China, 1:1000). β-actin (ZSGB-Bio, Beijing, China, 1:1000) was used as an internal reference. After overnight incubation, the membranes were washed 3 times with TBST (supplemented with 0.1% Tween 20) and incubated with HPR-labeled goat anti-rabbit IgG (ZSGB-Bio, Beijing, China, 1:1000) dilution for 2 h at room temperature. After washing with TBST, it was reacted with a luminescent solution (TransGen Biotech, Beijing, China) and developed under a gel imaging system.
We statistically analyzed and calculated the mean ± standard deviation ( ± SD). The charts and experimental data analyzed were created by using GraphPad Primer 8. A student’s t-test was used to analyze the differences between two groups, and one-way analysis of variance (ANOVA) was used to analyze the comparison among multiple groups. A p < 0.05 was defined as the differences that were statistically significant. Further, ns represented no statistically significant difference between data. Image Pro Plus v.6.0 was used to analyze the wound healing and transwell assays. ImageJ software (1.8.0.112) was used to analyze the results from the Western blots.
We successfully constructed the HOTAIR/miR-203/CAV1 network and explored the regulatory mechanism of ceRNA-miRNA-mRNA in this network for MDA-MB-231 cells. If only the overexpression of HOTAIR or overexpression of CAV1, or the inhibition of the expression of miR-203 is considered, the proliferation, invasion, and migration of breast cancer MDA-MB-231 cells can be significantly improved. Notably, in our HOTAIR/miR-203/CAV1 network, HOTAIR competes for binding to miR-203, which, in turn, silences CAV1 expression. Therefore, we can boldly propose that, once the expression of HOTAIR is extremely low, the resistance to miR-203 expression is successfully reduced. Then, a large amount of miR-203 is expressed, which, in turn, blocks the production of a large amount of CAV1, thereby successfully inhibiting the progression of breast cancer cells (Figure 7). As our initial validation demonstrates, knockdown of HOTAIR inhibited the proliferation, invasion, and migration of breast cancer cells. Although the HOTAIR/miR-203/CAV1 network is of great significance to the progression of breast cancer in MDA-MB-231, what other regulators are involved? The anti-tumor properties of miR-203 and the mechanism of differential expression of related molecules in invasive and non-invasive breast cancer cells deserve further study. However, can these relevant molecules be used as targets for inhibiting breast cancer metastasis? It deserves more exploration. In the future, we will fully decipher this ceRNA/mRNA axis and provide new ideas, directions, and an experimental basis for breast cancer diagnosis, targeted therapy, and prognosis assessment. | true | true | true |
PMC9569934 | 36232914 | Sifaddin M. R. Konis,Jonathan R. Hughes,Jason L. Parsons | TRIM26 Maintains Cell Survival in Response to Oxidative Stress through Regulating DNA Glycosylase Stability | 01-10-2022 | DNA damage,DNA repair,OGG1,NEIL1,NTH1,TRIM26,ubiquitin | Oxidative DNA base lesions in DNA are repaired through the base excision repair (BER) pathway, which consequently plays a vital role in the maintenance of genome integrity and in suppressing mutagenesis. 8-oxoguanine DNA glycosylase (OGG1), endonuclease III-like protein 1 (NTH1), and the endonuclease VIII-like proteins 1–3 (NEIL1–3) are the key enzymes that initiate repair through the excision of the oxidized base. We have previously identified that the E3 ubiquitin ligase tripartite motif 26 (TRIM26) controls the cellular response to oxidative stress through regulating both NEIL1 and NTH1, although its potential, broader role in BER is unclear. We now show that TRIM26 is a central player in determining the response to different forms of oxidative stress. Using siRNA-mediated knockdowns, we demonstrate that the resistance of cells to X-ray radiation and hydrogen peroxide generated as a consequence of trim26 depletion can be reversed through suppression of selective DNA glycosylases. In particular, a knockdown of neil1 or ogg1 can enhance sensitivity and DNA repair rates in response to X-rays, whereas a knockdown of neil1 or neil3 can produce the same effect in response to hydrogen peroxide. Our study, therefore, highlights the importance of TRIM26 in balancing cellular DNA glycosylase levels required for an efficient BER response. | TRIM26 Maintains Cell Survival in Response to Oxidative Stress through Regulating DNA Glycosylase Stability
Oxidative DNA base lesions in DNA are repaired through the base excision repair (BER) pathway, which consequently plays a vital role in the maintenance of genome integrity and in suppressing mutagenesis. 8-oxoguanine DNA glycosylase (OGG1), endonuclease III-like protein 1 (NTH1), and the endonuclease VIII-like proteins 1–3 (NEIL1–3) are the key enzymes that initiate repair through the excision of the oxidized base. We have previously identified that the E3 ubiquitin ligase tripartite motif 26 (TRIM26) controls the cellular response to oxidative stress through regulating both NEIL1 and NTH1, although its potential, broader role in BER is unclear. We now show that TRIM26 is a central player in determining the response to different forms of oxidative stress. Using siRNA-mediated knockdowns, we demonstrate that the resistance of cells to X-ray radiation and hydrogen peroxide generated as a consequence of trim26 depletion can be reversed through suppression of selective DNA glycosylases. In particular, a knockdown of neil1 or ogg1 can enhance sensitivity and DNA repair rates in response to X-rays, whereas a knockdown of neil1 or neil3 can produce the same effect in response to hydrogen peroxide. Our study, therefore, highlights the importance of TRIM26 in balancing cellular DNA glycosylase levels required for an efficient BER response.
The base excision repair (BER) pathway plays a vital role in the repair of oxidative DNA base damage and DNA single-strand breaks, and consequently in the maintenance of genome stability. This is exemplified by the fact that cellular DNA is continually subject to reactive oxygen species (ROS) generated by endogenous and exogenous sources, such as through oxidative metabolism and by ionizing radiation, respectively. Estimates of the levels of DNA lesions are ~10,000 per cell per day [1], and if these are not efficiently and effectively repaired, this can lead to mutagenesis and ultimately to the development of human diseases, such as premature ageing, neurodegenerative diseases, and cancer. BER is initiated by one of eleven damage-specific DNA glycosylases [2,3], which act to excise the DNA lesion, and which then stimulates downstream repair activities including AP site incision by AP endonuclease-1, single-nucleotide incorporation by DNA polymerase β, and DNA ligation by the DNA ligase IIIα-X-ray cross-complementing protein 1 complex [4]. The most critical DNA glycosylases that are involved in the recognition and repair of the majority of oxidative DNA base damages are 8-oxoguanine DNA glycosylase (OGG1), endonuclease III-like protein 1 (NTH1; also known as NTHL1), and the endonuclease VIII-like proteins 1–3 (NEIL1–3). However, other DNA glycosylases such as MutY DNA glycosylase (MYH) and thymine DNA glycosylase can also excise oxidized DNA bases. Nevertheless, OGG1 and NTH1 are considered the major enzymes responsible for the repair of 8-oxoguanine and oxidized pyrimidines, respectively [5,6]. In contrast, NEIL1 and NEIL2 appear to have roles in the repair of oxidative lesions in single-stranded DNA [7,8] and, therefore, during transcription and DNA replication. NEIL3 has also been observed to be more active on single-stranded DNA-containing lesions [9], with indications of roles specifically in the repair of interstrand crosslinks and quadruplex DNA [10,11,12,13,14]. It is known that the BER pathway is tightly controlled by individual protein post-translational modifications that act to control the levels, activities, and interactions of the proteins required for an efficient cellular DNA damage response and, therefore, to suppress the accumulation of DNA lesions [15,16]. Protein ubiquitylation, catalyzed ultimately by E3 ubiquitin ligases, has in particular been shown to be a key mechanism through which the cellular protein levels of BER are regulated, both at a steady-state level but also coordinated in response to oxidative stress [17,18]. Approximately 600–700 E3 ubiquitin ligases are present in the human genome, but each have their own target substrate specificity, where they act to transfer ubiquitin moieties predominantly onto specific lysine residues present within the protein. The addition of multiple ubiquitin units through internal lysine residues leading to polyubiquitylation can target the protein for degradation by the 26S proteasome. However, ubiquitylation is a reversible protein post-translational modification that is catalyzed by deubiquitylation enzymes (DUBs) [19]. Our previous evidence has demonstrated that the cellular levels of DNA polymerase β are coordinated by the E3 ubiquitin ligases Mcl-1 ubiquitin ligase E3 (Mule) and C-terminal of Hsc70-interacting protein (CHIP), which is counterbalanced by the DUB-ubiquitin-specific protease 47 (USP47) [20,21,22]. However, we also have more recently demonstrated that Mule can target NEIL1 for ubiquitylation-dependent degradation [23], and so is involved in regulating the two proteins at different steps in the BER process. This demonstrates that the BER response to DNA damage can be efficiently controlled through ubiquitylation driven by a specific set of E3 ubiquitin ligases. We also recently demonstrated that another E3 ubiquitin ligase, tripartite motif 26 (TRIM26), plays a major role in the regulation of the protein levels of both NEIL1 and NTH1 required for cell survival in response to DNA damage stress [23,24], further demonstrating that a single E3 enzyme can target multiple BER proteins. These studies were performed using different sources of oxidative stress, ionizing radiation and hydrogen peroxide, and so at this stage, it was difficult to understand the contribution of TRIM26-dependent regulation of the two DNA glycosylases under both of these conditions. TRIM26 is one member of the tripartite motif proteins, many of which contain a N-terminal RING finger domain that catalyzes ubiquitylation and has broad cellular roles, including in autophagy and innate immunity, and whose dysregulation is implicated in several cancer types [25,26]. TRIM26 specifically has been shown to be downregulated in hepatocellular carcinoma and papillary thyroid carcinoma [27,28], but upregulated in bladder cancer [29]. Alteration of the expression of trim26 in cell lines from these tumors was found to impact cell proliferation and migration. In terms of molecular mechanisms and separate from its role in BER, TRIM26 has been demonstrated to target the transcription factor IRF3 for ubiquitylation-dependent degradation, leading to reduced interferon β production and an antiviral response [30]. TRIM26 is also thought to control the inflammatory innate immune response through polyubiquitylation of TAB1 and enhancing NF-κB and MAPK signaling [31]. Moreover, TRIM26 has been indicated in the regulation of ZEB1 protein degradation via ubiquitylation, which alongside USP39, controls the epithelial-to-mesenchymal transition and ultimately the growth of hepatocellular carcinoma cells [32]. It is therefore clear that TRIM26 has multiple cellular targets and roles through its action as an E3 ubiquitin ligase. Here, we utilized an siRNA knockdown of trim26 to further define its role in the regulation of the DNA glycosylases required for the response to oxidative stress. Interestingly, we discovered that the increased resistance and DNA repair activity of trim26-knockdown cells to ionizing radiation can be suppressed in combination with a knockdown of neil1 or ogg1, whereas this phenotype in response to hydrogen peroxide appears dependent on neil1 and neil3. We also demonstrate that purified TRIM26 can ubiquitylate NEIL1, OGG1, NTH1, and NEIL3 proteins in vitro, suggesting that the enzyme has a wider role in the regulation of DNA glycosylases than previously thought.
Our previous evidence has demonstrated that TRIM26 can ubiquitylate both NEIL1 and NTH1 in vitro, which is important in promoting cell survival in response to ionizing radiation and hydrogen peroxide, respectively [23,24]. We, therefore, initially further explored the substrate specificity of purified TRIM26 protein against other DNA glycosylases, using in vitro ubiquitylation assays. We demonstrate that not only can TRIM26 ubiquitylate NEIL1 and NTH1 (Figure 1A,B) in keeping with our previous data, but it can also promote ubiquitylation of OGG1 and NEIL3 (Figure 1C,D). The degree of ubiquitylation efficiency of the DNA glycosylases by TRIM26 varies, and, in some cases (such as NTH1), appears to achieve a level of saturation. Nevertheless, this suggests that at least in vitro, TRIM26 has a broad substrate specificity and can target multiple DNA glycosylases for ubiquitylation.
We previously showed that trim26 knockdown U2OS cells are more resistant to the cell-killing effects of ionizing radiation, as a consequence of the accumulation of steady-state levels of NEIL1 protein [23]. To explore this phenotype further, we performed a double knockdown of neil1, nth1, ogg1, or neil3 along with trim26 in order to identify specific combinations that led to restoration of cellular radiosensitivity. We first show that we are able to suppress the levels of TRIM26 protein in U2OS cells using a targeted siRNA knockdown compared to a non-targeting (NT) control siRNA (Figure 2A, compare lanes 1 and 2), and that the knockdown efficiency is retained in the various combinations targeting both trim26 and the DNA glycosylases (Figure 2A, compare lanes 1 and 3–5). Using clonogenic assays, and as expected, we observe a significantly (p < 0.03) acquired resistance of cells to X-ray irradiation in the absence of trim26 compared to the NT control siRNA-treated cells (Figure 2B,C). However, we demonstrate that this radiosensitivity can be restored using a double knockdown of both trim26 and neil1, which is significant from trim26 knockdown alone (p < 0.0001), which highlights the association of the radioresistance of trim26-deficient cells with an accumulation of NEIL1 protein, which we previously observed [23]. In contrast, an siRNA knockdown of either nth1 (Figure 2D,E) or neil3 (Figure 2F,G) was unable to enhance the radiosensitivity of trim26-deficient cells. Surprisingly, we discovered that the combination of ogg1 and trim26 siRNA led to radiosensitivity that was significantly different from trim26 knockdown alone (p < 0.002), and similar to that observed in the NT control siRNA-treated cells (Figure 2H,I). This suggests that cellular resistance to ionizing radiation in the absence of trim26 is dependent on neil1 and ogg1.
To correlate the effects of the DNA glycosylase knockdowns in combination with trim26 on X-ray-induced cell survival relative to DNA repair, we analyzed the rates of repair of DNA damage in U2OS cells using alkaline comet assays. In the absence of trim26, we observed an accelerated rate of repair of alkali-labile sites and DNA single-strand breaks compared to the NT control siRNA-treated cells (Figure 3A), suggesting that accumulating DNA glycosylase levels are responsible for enhanced repair activity. In the absence of both trim26 and neil1, the rates of DNA damage repair were restored similar to that of the NT control siRNA (Figure 3A). However, no changes in the rates of repair of alkali-labile sites and DNA single-strand breaks were observed with the combination of a siRNA knockdown of both trim26 and nth1 compared to a trim26 knockdown alone (Figure 3B). Interestingly, and similar to experiments involving neil1 siRNA, targeting ogg1 for an siRNA-mediated knockdown also led to a slower rate of repair of radiation-induced DNA damage compared to the trim26-deficient cells (Figure 3C). These effects are consistent with the changes in radiosensitivity of the cells observed above (Figure 2A–I) and show that the increased resistance of trim26 siRNA knockdown cells is driven through repair coordinated by either NEIL1 or OGG1. To provide supporting evidence for this, we overexpressed neil1 and ogg1 individually (Figure 3D,G) and demonstrate that this leads to significantly (p < 0.04) enhanced resistance of U2OS cells to X-ray irradiation compared to control-transfected cells (Figure 3E,F). We furthermore show that these cells also harbor increased rates of repair of radiation-induced alkali-labile sites and DNA single-strand breaks compared to control cells (Figure 3H,I), correlating with the increased cellular radioresistance due to the higher expressed levels of NEIL1 or OGG1 protein. We have previously shown that TRIM26 controls the steady-state levels of NEIL1 protein, and that a trim26 knockdown generates cellular resistance to X-ray irradiation due to an accumulation of NEIL1 [23]. Given our new data that OGG1 also appears to play a role in radiation-induced DNA damage repair and cellular resistance in a TRIM26-dependent manner, we analyzed DNA glycosylase protein levels by quantitative immunoblotting. In whole cell extracts, we observe that the steady-state levels of both NEIL1 and OGG1 increase by ~2.2 and ~1.8-fold, respectively, in the absence of TRIM26 (Figure 4A,B), whereas protein levels of NTH1 remain unchanged. We also analyzed protein levels in response to X-ray irradiation following biochemical fractionation. In keeping with our previously published data [23], we observe that NEIL1 is present in U2OS cells in a soluble fraction (S) and not strongly bound to chromatin (CB; Figure 4C, compare lanes 1 and 2). We also find that NEIL1 accumulates in response to X-ray irradiation (Figure 4C,E), but that the protein levels are moderately ~1.2-fold higher in trim26-siRNA-treated cells, particularly at 0.5–1 h post-irradiation, compared to the NT control siRNA (Figure 4D,E). However, it should be noted that protein levels are normalized relative to their respective control, and that the steady-state levels of NEIL1 are already ~2.2-fold higher in trim26 knockdown cells. Analysis of NTH1 protein reveals that this is majorly chromatin-bound (Figure 4C, compare lanes 1 and 2) as we previously observed [24], but that there are no substantial differences in protein levels in the presence or absence of trim26 following irradiation (Figure 4C,D,F). This suggests that NTH1 under these conditions is not regulated in a TRIM26-dependent manner. OGG1 protein is found to be present in both a soluble and chromatin-bound form (Figure 4C, compare lanes 1 and 2), and similar to NEIL1, the levels of the protein are higher in trim26 knockdown cells, particularly in the soluble fraction at 1–6 h post-irradiation, compared to the NT siRNA control irradiated cells (Figure 4C,D,G,H). Nevertheless again, protein levels are normalized relative to their respective control, and the steady-state levels of OGG1 are already ~1.8-fold higher in trim26 knockdown cells. Cumulatively, this demonstrates that both NEIL1 and OGG1 protein levels are controlled by TRIM26, which mediates the response to X-ray irradiation.
We performed a double knockdown of neil1, nth1, ogg1, or neil3 along with trim26 in U2OS cells and analyzed the sensitivity in response to oxidative stress induced by hydrogen peroxide compared to a trim26 knockdown alone. Similar to our previous evidence acquired in HCT116 cells [24], we observed that U2OS cells with a trim26 siRNA-mediated depletion display a significantly (p < 0.004) increased resistance to hydrogen peroxide compared to NT siRNA treated cells (Figure 5A). With a double knockdown of both trim26 and neil1, cellular sensitivity is restored to NT siRNA-treated levels, which was significantly (p < 0.0001) different compared to trim26-deficient cells (Figure 5A). In contrast, an siRNA knockdown of either nth1 (Figure 5B) or ogg1 (Figure 5C) has no significant impact on the resistance of trim26-deficient cells to hydrogen peroxide. Interestingly, we discovered that the combination of neil3 and trim26 siRNA knockdown led to cellular sensitivity to hydrogen peroxide that was similar to that observed in the NT control siRNA, which was again significantly (p < 0.0001) different compared to trim26-deficient cells (Figure 5D). To correlate these effects on cell survival following hydrogen peroxide with DNA damage repair, we analyzed the rates of repair of alkali-labile sites and DNA single-strand breaks in U2OS cells with the various siRNA knockdown combinations. As expected in the absence of trim26, there was an accelerated rate of repair of the DNA damage compared to the NT-control-siRNA-treated cells, but which could be suppressed in combination with a knockdown of neil1 (Figure 5E). In the absence of both trim26 and nth1 (Figure 5F) or of trim26 and ogg1 (Figure 5G), the kinetics of DNA damage repair were similar to that of the trim26-siRNA-treated-only cells. Targeting neil3 for an siRNA-mediated knockdown in trim26-depleted cells, similar to neil1, led to a slower rate of repair of DNA damage induced by hydrogen peroxide (Figure 5H). Effects on DNA repair rates are consistent with the changes observed in cellular sensitivity (Figure 5A–D), and reflect that neil1 and neil3 are the major drivers of increased resistance in trim26 siRNA knockdown cells. Additional support for this, at least focused on neil1, is provided by our observations that NEIL1 overexpression leads to significantly (p < 0.02) enhanced resistance of U2OS cells to hydrogen peroxide compared to control-transfected cells (Figure 5I), and that there are also associated increases in the kinetics of repair of alkali-labile sites and DNA single-strand breaks under these conditions (Figure 5J). We analyzed the endogenous protein levels of NEIL1, NTH1, and NEIL3 in trim26 knockdown compared to NT control siRNA-treated cells following hydrogen peroxide treatment. The levels of NEIL1 protein within the soluble fraction were ~1.4–1.5-fold higher in trim26 siRNA-treated cells at 1–6 h post-treatment compared to the NT control siRNA cells (Figure 6A–C). No significant differences in chromatin bound NTH1 protein levels in the presence or absence of trim26 following treatment were found (Figure 6A,B,D). Similar to NEIL1, NEIL3 protein was observed to be largely present within the soluble fraction and not chromatin bound but also the levels of the protein were ~1.5–1.7-fold higher in trim26 knockdown cells at 2–6 h post-treatment with hydrogen peroxide compared to the NT siRNA control cells (Figure 6A,B,E). These data indicate that both NEIL1 and NEIL3 protein levels are tightly controlled by TRIM26, which mediates the response to oxidative stress induced by hydrogen peroxide.
BER is an essential DNA repair pathway that responds to cellular oxidative stress and is critical in maintaining genome stability and in preventing mutagenesis. Within this pathway, OGG1, NTH1, and NEIL1–3 are the principal DNA glycosylases that recognize and excise oxidative DNA base lesions, which then promotes subsequent repair coordinated by AP endonuclease-1, DNA polymerase β, and DNA ligase IIIα-X-ray cross-complementing protein 1 complex. A number of studies have demonstrated that the efficiency of the BER pathway is subject to tight control by post-translational modifications, of which ubiquitylation as a mechanism for regulating individual repair protein levels has been increasingly found [15,16]. DNA glycosylases specifically are a target for regulation at both the transcriptional and post-translational level [33], which avoids the build-up of potentially more toxic BER intermediates. The importance of controlling DNA glycosylase levels is displayed by the altered protein expression observed in several diseases, including neurodegenerative diseases and cancer. For example, an altered expression and activity of OGG1 has been observed in head and neck cancers [34,35], of NTH1 in gastric cancer [36], and of NEIL3 in various human cancers [37,38], whereas a loss of NEIL1 causes memory and brain defects indicative of early-onset neurodegenerative disease, similar to those observed in Alzheimer’s and Parkinson’s diseases [39,40]. We previously identified that the E3 ubiquitin ligase TRIM26 can regulate the cellular protein levels of NEIL1 and NTH1 in response to X-ray radiation and hydrogen peroxide-induced stress, respectively [23,24]. In this study, we now provide evidence that TRIM26 can also ubiquitylate OGG1 and NEIL3 in vitro, but that it controls the different DNA glycosylases based on the form of DNA damage stress. Specifically, NEIL1 and OGG1 are responsive to X-ray radiation in a TRIM26-dependent manner, whereas NEIL1 and NEIL3 respond following treatment with hydrogen peroxide. OGG1 and NTH1 are well established as being the major DNA glycosylases involved in the repair of 8-oxoguanine and oxidized pyrimidines, respectively. In contrast, the NEIL glycosylases appear to have more defined cellular roles, such as in transcription and replication as a consequence of their activity on DNA lesions within single-stranded DNA [3], and in DNA crosslink repair [41]. Therefore, it is understandable that their regulation may be responsive to different types and sources of DNA damage. Interestingly, we observed that the control of NEIL1 is important for the response to both X-rays and hydrogen peroxide in terms of promoting survival and efficient DNA damage repair, indicating that a common DNA lesion is being generated by these sources of oxidative stress that is highly dependent on NEIL1. However, maintenance of OGG1 by TRIM26 only occurs following X-ray irradiation, whereas NEIL3 is tightly controlled by TRIM26 in response to hydrogen peroxide. The reasoning for this is currently unclear but suggests a different DNA lesion dependence that requires either OGG1 or NEIL3 for repair. Interestingly, recent evidence has suggested that 8-oxoguanine may act as a transcriptional regulator and can negatively affect gene transcription when in non-transcribed DNA, but can alternatively promote gene expression, such as when present within a G-quadruplex sequence [42,43,44]. It is therefore tempting to speculate that the levels of OGG1, and possibly NEIL1 and NEIL3, are differentially regulated not only for maintaining 8-oxoguanine throughout the genome, but specifically for its roles in DNA transcription and epigenetic regulation [45,46]. However, this requires more detailed investigation. In addition to the selective control of the DNA glycosylase levels by TRIM26 relative to the DNA damage stress, an unanswered question is how the mechanism is coordinated. Predictably, this could also be achieved at the post-translational level, either through a competing DUB that is able to reverse the effects of TRIM26-dependent ubiquitylation and degradation of the DNA glycosylase, or through an alternative post-translational modification that either stimulates or inhibits TRIM26 activity. In support of the former, we have previously identified that the DUB USP47 can control the protein levels of DNA polymerase β and provides competition for ubiquitylation catalyzed by CHIP and Mule [20], although no DUBs for the DNA glycosylases OGG1, NTH1, NEIL1, and NEIL3 have yet been identified. In terms of alternative post-translational modifications, OGG1 has been previously demonstrated to be subject to acetylation [47] and phosphorylation [48], which could interfere with ubiquitylation. Similarly, NEIL1 is reportedly phosphorylated [49] and acetylated [50]. However, to our knowledge, no post-translational modifications of NTH1 and NEIL3 have yet been identified. Furthermore, poly(ADP-ribosyl)ation catalyzed predominantly poly(ADP-ribose) polymerase-1 (PARP-1) plays a critical role in coordinating BER, and, thus, in controlling cell survival in response to genotoxic stress [4]. Therefore it could be speculated that this post-translational modification may also have an underlying role in regulating DNA glycosylase stability. Nevertheless, further research needs to be established in order to examine any potential crosstalk between TRIM26-dependent ubiquitylation and other post-translational modifications of the DNA glycosylases. Despite this, our research highlights a central role for TRIM26 in controlling and coordinating the cellular response to DNA damage through DNA glycosylase modulation.
NEIL1 antibodies were kindly provided by Dr. T. Rosenquist. Antibodies against TRIM26 (ab89290), NTH1 (ab70726), OGG1 (ab124741), and fibrillarin (ab4566) were from Abcam (Cambridge, UK). NEIL3 (sc-393703) and lamin a/c antibodies (sc-7292) were from Santa Cruz Biotechnology (Dallas, TX, USA), and tubulin antibodies (T6199) were from Merck (Gillingham, UK). Bacterial expression plasmids and protein purification of TRIM26, OGG1, NTH1, NEIL1, and NEIL3 proteins was performed as previously described [23,24,51].
U2OS cells were cultured at 37 °C in 5% CO2 in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum, 2 mM L-glutamine, 1 × penicillin-streptomycin, and 1 × non-essential amino acids. Cells were authenticated using short-tandem-repeat (STR) profiling and were routinely tested to ensure the absence of mycoplasma infection. To perform siRNA knockdowns, cells were cultured in 35 mm dishes for 24 h to 30–50% confluence and then treated with 2 µL of Lipofectamine RNAiMAX transfection reagent (Life Technologies, Paisley, UK) in the presence of either 80 nM (NT, TRIM26, NEIL1, and NEIL3) or 160 nM (OGG1) siRNA for an additional 72 h. The following siRNA sequences were used: Qiagen AllStars Negative Control siRNA (Qiagen, Southampton, UK), TRIM26 siRNA (5′-CCGGAGAAUUCUCAGAUAA-3′), or the appropriate ON-TARGETplus siRNA pools against OGG1, NTH1, NEIL1, or NEIL3 (Horizon Discovery, Cambridge, UK). For overexpression of NEIL and OGG1, 0.2 µg of pCMV-Tag3a mammalian expression plasmids (as previously described [24,51]) was similarly transfected into cells using Lipofectamine 2000 transfection reagent (Life Technologies, Paisley, UK) for 24 h prior to subsequent analysis. Control samples for overexpression were treated with transfection reagent only.
Cells cultured in 35 mm dishes were treated with 1–4 Gy X-rays using the 130 MeV CellRad X-ray irradiator (Faxitron Bioptics, Tucson, AZ, USA), or with 250–1000 µM hydrogen peroxide for 15 min. Cells were washed with PBS, trypsinized, counted, and a defined number seeded in triplicate into 6-well plates. Cells were incubated at 37 °C in 5% CO2 for 9 days to promote colony growth, and these were then fixed and stained with 6% glutaraldehyde and 0.5% crystal violet for 30 min. Plates were washed, left to air-dry overnight, and colonies counted using the GelCount colony analyzer (Oxford Optronics, Oxford, UK). The surviving fraction was determined using the number of colonies per treatment level versus the number of colonies achieved in the untreated control. Statistical analysis of the differences across the treatment doses comparing the various gene knockdowns/overexpressions was performed using the CFAssay for R package [52].
Cells were washed, harvested in ice-cold PBS, and whole-cell extracts prepared as previously described [53]. Alternatively, biochemical fractionation was performed immediately to generate soluble and chromatin-bound protein fractions as previously described [23]. In brief, cell pellets were resuspended in two packed cell volumes (PCVs) of buffer containing 20 mM Tris–HCl (pH 7.8), 2.5 mM MgCl2, 0.5% (v/v) IGEPAL CA-630, 100 µM PMSF, 1 mM N-ethylmaleimide (NEM), and 1 µg/mL of the protease inhibitors (leupeptin, aprotinin, chymostatin, and pepstatin), and incubated for 10 min on ice. Extracts were centrifuged at 10,000 rpm for 2 min at 4 °C and the supernatant containing soluble proteins (S) was collected. The nuclear pellet was similarly extracted with two PCVs of buffer containing 20 mM NaPO4 (pH 8.0), 0.5 M NaCl, 1 mM EDTA, 0.75% (v/v) Triton X-100, 10% (v/v) glycerol, 100 µM PMSF, 1 mM NEM, and 1 µg/mL of each protease inhibitor and incubated on ice for 10 min. Following centrifugation, the supernatant containing chromatin-bound proteins (CB) was collected. For immunoblotting analysis, 40–70 µg of protein from the S fraction and the same corresponding volume of the CB fraction were used, and proteins were visualized and quantified using the Odyssey image analysis system (Li-cor Biosciences, Cambridge, UK).
Ubiquitylation reactions were performed as previously described [23,24,51]. Briefly, reactions containing either histagged-OGG1 (5.2 pmol), NTH1 (5.8 pmol), NEIL1 (4.6 pmol), NEIL3 (3 pmol), and/or TRIM26 (11–22 pmol) were incubated with 0.7 pmol GST-E1 activating enzyme, and 2.5 pmol H5a, H5b, and H5c E2-conjugating enzymes; 0.6 nmol ubiquitin in buffer containing 25 mM Tris-HCl (pH 8.0), 4 mM ATP, 5 mM MgCl2, 200 µM CaCl2, and 1 mM DTT were prepared and incubated in LoBind protein tubes (Eppendorf, Stevenage, UK) for 1 h at 30 °C with agitation. After terminating the reactions through the addition of SDS-PAGE sample buffer (25 mM Tris-HCl (pH 6.8), 2.5% β-mercaptoethanol, 1% SDS, 10% glycerol, 1 mM EDTA, and 0.05 mg/mL of bromophenol blue), these were heated for 5 min at 95 °C and analyzed by SDS-PAGE and immunoblotting.
The alkaline comet assay was performed as previously described, utilizing in-gel DNA repair activities [54]. In brief, cells were trypsinized, diluted to ~1 × 105 cell/mL, and 250 µL aliquots placed in the wells of a 24-well plate on ice. Following X-ray irradiation (1.5 Gy) or treatment with hydrogen peroxide (10 µM) for 5 min, cells were embedded in 1% low-melting-point agarose (Bio-Rad, Hemel Hempstead, UK), which was added to a microscope slide precoated and dried with 1% normal-melting-point agarose. The agarose was then allowed to set for 2–3 min on ice, and then placed in a humidified chamber for up to 120 min to stimulate DNA repair. Following this, cell lysis was performed by placing slides in 2.5 M NaCl, 100 mM EDTA, 10 mM Tris-HCl, pH 10.5, 1% (v/v dimethyl sulfoxide (DMSO), and 1% (v/v) Triton X-100 for at least 1 h at 4 °C. Slides were transferred to a darkened comet assay tank (Appleton Woods, Birmingham, UK), incubated for 30 min in fresh cold electrophoresis buffer (300 mM NaOH, 1 mM EDTA, and 1% (v/v) DMSO, pH 13) to allow the DNA to unwind, and electrophoresis was performed at 25 V, 300 mA for 25 min. Slides were carefully removed from the tank and neutralized three times with 5 min washes of 0.5 M Tris-HCl (pH 8.0), prior to air-drying overnight. Following rehydration of the slides for 30 min in water (pH 8.0), the DNA was stained using SYBR Gold (Life Technologies, Paisley, UK) diluted 1:20,000 in water (pH 8.0) for 30 min, and then again allowed to air-dry. For imaging, cells (50 per slide, 2 slides per time point) were analyzed using the Komet 6.0 image analysis software (Andor Technology, Belfast, Northern Ireland) and average % tail DNA values were determined from three independent, biological experiments. | true | true | true |
PMC9569948 | 36232946 | Veronika Banicka,Marie Christine Martens,Rüdiger Panzer,David Schrama,Steffen Emmert,Lars Boeckmann,Alexander Thiem | Homozygous CRISPR/Cas9 Knockout Generated a Novel Functionally Active Exon 1 Skipping XPA Variant in Melanoma Cells | 01-10-2022 | DNA repair,nucleotide excision repair,XPA,CRISPR,knockout,protein variant,melanoma,A375 | Defects in DNA repair pathways have been associated with an improved response to immune checkpoint inhibition (ICI). In particular, patients with the nucleotide excision repair (NER) defect disease Xeroderma pigmentosum (XP) responded impressively well to ICI treatment. Recently, in melanoma patients, pretherapeutic XP gene expression was predictive for anti-programmed cell death-1 (PD-1) ICI response. The underlying mechanisms of this finding are still to be revealed. Therefore, we used CRISPR/Cas9 to disrupt XPA in A375 melanoma cells. The resulting subclonal cell lines were investigated by Sanger sequencing. Based on their genetic sequence, candidates from XPA exon 1 and 2 were selected and further analyzed by immunoblotting, immunofluorescence, HCR and MTT assays. In XPA exon 1, we established a homozygous (c.19delG; p.A7Lfs*8) and a compound heterozygous (c.19delG/c.19_20insG; p.A7Lfs*8/p.A7Gfs*55) cell line. In XPA exon 2, we generated a compound heterozygous mutated cell line (c.206_208delTTG/c.208_209delGA; p.I69_D70delinsN/p.D70Hfs*31). The better performance of the homozygous than the heterozygous mutated exon 1 cells in DNA damage repair (HCR) and post-UV-C cell survival (MTT), was associated with the expression of a novel XPA protein variant. The results of our study serve as the fundamental basis for the investigation of the immunological consequences of XPA disruption in melanoma. | Homozygous CRISPR/Cas9 Knockout Generated a Novel Functionally Active Exon 1 Skipping XPA Variant in Melanoma Cells
Defects in DNA repair pathways have been associated with an improved response to immune checkpoint inhibition (ICI). In particular, patients with the nucleotide excision repair (NER) defect disease Xeroderma pigmentosum (XP) responded impressively well to ICI treatment. Recently, in melanoma patients, pretherapeutic XP gene expression was predictive for anti-programmed cell death-1 (PD-1) ICI response. The underlying mechanisms of this finding are still to be revealed. Therefore, we used CRISPR/Cas9 to disrupt XPA in A375 melanoma cells. The resulting subclonal cell lines were investigated by Sanger sequencing. Based on their genetic sequence, candidates from XPA exon 1 and 2 were selected and further analyzed by immunoblotting, immunofluorescence, HCR and MTT assays. In XPA exon 1, we established a homozygous (c.19delG; p.A7Lfs*8) and a compound heterozygous (c.19delG/c.19_20insG; p.A7Lfs*8/p.A7Gfs*55) cell line. In XPA exon 2, we generated a compound heterozygous mutated cell line (c.206_208delTTG/c.208_209delGA; p.I69_D70delinsN/p.D70Hfs*31). The better performance of the homozygous than the heterozygous mutated exon 1 cells in DNA damage repair (HCR) and post-UV-C cell survival (MTT), was associated with the expression of a novel XPA protein variant. The results of our study serve as the fundamental basis for the investigation of the immunological consequences of XPA disruption in melanoma.
In the last years, the defects of different DNA repair pathways have been associated with an increased response to immune checkpoint inhibition (ICI) in cancer patients [1,2,3]. Certainly, the most prominent clinical example is the first site/tissue agnostic approval of pembrolizumab by the U.S. food and drug administration for mismatch repair (MMR) deficient (or microsatellite instability-high) cancer [4]. Besides MMR, there are two other pathways that repair DNA single-strand breaks (SSB): base excision repair (BER); and nucleotide excision repair (NER). Remarkably, the immunological consequences of NER defects are far less explored than those of MMR and BER, although people suffering from the NER defect disease, xeroderma pigmentosum (XP), had an impressive response of their skin tumors to anti-programmed cell death-1 (PD-1) ICI in case reports [2,5,6,7,8,9,10,11,12]. Of note, we just recently revealed that expression of XP genes was predictive of response in two cohorts of anti-PD-1 treated melanoma patients [13]. However, the fundamental basis of our observation is not yet clear and needs further investigation by, e.g., cell experiments. We therefore used Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) and CRISPR-associated (Cas) protein 9 to target XPA, the central coordinator of NER and scaffold provider for other NER proteins, in the well-established human melanoma cell line A375 [14,15,16,17]. CRISPR/Cas9 is an innovate DNA-editing system, which was originally discovered as a prokaryotic adaptive immunity mechanism for the cleavage of invading nucleic acids [18,19,20,21]. When utilizing in genome engineering, specific single guide (sg)RNAs are designed to bind to a DNA target sequence in the gene of interest. Complementary pairing of a sgRNA to its target results in the generation of a blunt-end double strand break (DSB) by Cas9 [18,21]. In eukaryotic cells, those are predominately repaired by error-prone non-homologous end joining (NHEJ) and to a lesser extent by more precise homology driven repair (HDR) [18,21]. NHEJ simply re-ligates the two strands without a homologous template and thereby often generates insertions and/or deletions (InDels) at the cleavage site [22]. Hence, it usually fails in restoring the original genetic sequence, leading to nonsense sequences, rendering the resulting protein inactive or at least less active. NHEJ is also fundamental for CRISPR/Cas9-mediated gene knockouts [21,23,24]. Upon targeting one site in a coding exon, knockouts arise from InDel-mediated frameshift mutations. They characteristically introduce premature stop codons, which disrupt gene expression or can lead to proteins missing the N-terminal region due to alternative translation initiation (ATI) [21,24]. By applying CRISPR/Cas9, we generated three different XPA-mutated A375 subclonal cell lines. Targeting exon 1 resulted in two knockout cell lines for wild type XPA—a homozygous (c.19delG; p.A7Lfs*8) and a compound heterozygous (c.19delG/c.19_20insG; p.A7Lfs*8/p.A7Gfs*55)—both leading to short N-terminal protein fragments. By targeting exon 2, we established a compound heterozygous cell line bearing an in-frame 3-base pair deletion in one of the two XPA alleles, while having a frameshifting mutation on the other (c.206_208delTTG/c.208_209delGA; p.I69_D70delinsN/p.D70Hfs*31). Notably, homozygous XPA exon 1 knockout resulted in expression of a novel XPA protein variant, which was associated with an improved survival following UVC irradiation compared to the other two cell lines.
For the CRISPR/Cas9-mediated knockout of XPA, we used PX459 plasmids encoding five sgRNAs (T1.1, T1.8, T1.9, T2.1 and T2.2), targeting the first two exons of the XPA gene. After transfection into A375 melanoma cells and puromycin selection [25] (Figure 1), PCR sequencing of the targeted exons confirmed the generation of polyclonal XPA-mutated cell lines (Figure 2). In this context, those polyclonal cell lines were considered particularly promising, that included the most additional superimposed peaks in the sequencing chromatogram (Figure 2, auspicious XPA-T1.8 compared to barely altered XPA-T1.1). Two polyclonal cell lines (targeted with T1.8 in exon 1 and with T2.1 in exon 2) were finally chosen to generate stable subclones bearing mutations on both XPA alleles through single clone expansion.
Subclonal single cell lines, named after their well of origin, were analyzed by Sanger sequencing of the targeted XPA exons (Figure 3A,B). Manifest genetic outcomes included subclones containing the wildtype (G1), or mutations on one or both alleles in the targeted region. Biallellic mutations were either homozygous (exon 1: B1) or compound-heterozygous (exon 1: A9, D10 and G7; exon 2: A3 and G3). Based on the Sanger sequencing results, two exon 1 knockout cell lines bearing frameshifting mutations on both alleles leading to early stop codons—the homozygous B1 (c.19delG) and the compound-heterozygous A9 (c.19delG/c.19_20insG)—were selected for further investigations (Figure 3B). From the exon 2 targeted subclones we chose the compound heterozygous A3 subclone (c.206_208delTTG/c.208_209delGA).
The genetic code of both alleles was used to predict the order of amino acids (AA) of the most likely resulting proteins presuming that transcription and translation start sites were unaltered (Figure 4A). DNA alterations in exon 1 targeted cell lines, B1 (c.19delG) and A9 (c.19delG/c.19_20insG), caused frameshift mutations leading to premature stop codons (p.A7Lfs*8; p.A7Lfs*8/p.A7Gfs*55) and thereby short N-terminal protein fragments. Only the first six AA were predicted to correspond to the XPA wildtype. Accordingly, in both exon 1 knockout subclones, immunoblotting revealed a loss of the ~37 kDa protein band when incubated with the monoclonal 12F5 XPA antibody (unpublished epitope; Figure 4B, upper panel). Our exon 2-mutated cells (A3) were anticipated to bear an in-frame mutation on one allele and a frameshift mutation on the other (p.I69_D70delinsN/p.D70Hfs*31). Consistently, the 12F5 antibody recognized XPA, though the signal was much weaker (Figure 4B, upper panel). Of note, when the homozygous exon 1 knockout cells (p.A7Lfs*8) were incubated with the monoclonal D9U5U antibody (targets residues surrounding R158) or the polyclonal STJ96279 antibody (binds C-terminally at AA 180–260), a new band of approximately ~31–33 kDa was detected (Figure 4B, middle and bottom panel). We hypothesized, that this constituted a de novo XPA protein variant missing the N-terminal start of the wildtype protein, and explored its functional consequences following UVC irradiation.
6-4 pyrimidine pyrimidone photoproducts (6-4PPs) and cyclobutane pyrimidine dimers (CPDs) are the main DNA lesions repaired by NER and they were analyzed by immunofluorescence before and after UV-C irradiation [16] (Figure 5A). XPA wildtype (WT) showed DNA defects only after irradiation with UVC, while 6-4PPs and CPDs were surprisingly detected even in non-irradiated B1 and A9 knockout cells. In exon 2-mutated A3, only a weak basal DNA damage signal could be detected. 100 J/m2 UVC irradiation produced clearly visible DNA damage in all tested cell lines, however, it was most detrimental to A9, characterized by lower cell numbers observed after irradiation. In the host-cell reactivation (HCR) assay, the repair of the irradiated (1000 J/m2 UVC) firefly plasmid was measured and set in relation to the baseline luminescence activity of non-irradiated firefly plasmids. All subclones had a diminished repair compared to the WT (Figure 5B). In general, repair of both exon 1 mutated cell lines was reduced to a greater extent. Among each other, repair in homozygous B1 cells was less impaired than in its heterozygous counterpart A9, possibly due to the expression of a compensatory new XPA protein variant.
Next, we systematically assessed the metabolic activity of subclonal cell lines compared to the XPA wildtype in the (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) tetrazolium (MTT) assay after irradiation with increasing doses of UVC (0–250 J/m2), that have been previously determined by our dosis-finding experiments. Again, the CRISPR/Cas9-generated subclones were more sensitive to UVC than the parental A375 XPA WT cells. Since the MTT assay is commonly used for determination of cell survival, we calculated that 50% of the WT cells were still living after 101 J/m2 UVC (defined as their lethal dose [LD]50), while LD50 doses were much lower in the CRISPR/Cas9-altered cell lines (Figure 6). Importantly, from the exon 1 knockout cell lines, homozygous B1 cells were again less affected by UVC irradiation than heterozygous A9 cells. Remarkably, metabolic activity was most restricted in A3 cells with a significant reduction compared to the WT at doses of 100 J/m2 and 200 J/m2.
Using CRISPR/Cas9, we generated several single cell-derived A375 subclonal cell lines, in which we successfully disrupted both XPA alleles within its first two exons. From the cells targeted in exon 1, we chose two similarly mutated knockout cell lines. While B1 cells were homozygously altered lacking one guanine (G) on either XPA allele (c.19delG), A9 cells had one allele identically mutated as B1 and an inserted G on the other (c.19delG/c.19_20insG) (Figure 3B). These alterations, provided translation start at the regular site, caused frameshift mutations resulting in premature termination codons (PTCs) and thereby leading to short N-terminal XPA protein fragments. However, functional consequences following UVC irradiation differed between B1 and A9. In this regard, better performance of B1 cells in HCR (Figure 5B) and MTT (Figure 6) assays was consistent with the expression of a novel smaller XPA protein variant, that we detected in our immunoblot analysis (Figure 4B). CRISPR/Cas may have diverse on-target impacts on protein level rather than exclusive knockout [23,24]. After introduction of frameshift mutations, the fate of subsequent mRNAs differs. First, mRNAs can be recognized by the cell as aberrant and degraded by nonsense mediated decay (NMD) [24,26,27]. Second, if a transcript bearing frameshifting mutations resists NMD, it can be translated in a truncated protein. Third, during RNA processing, alternative splicing (AS) events may occur, including particularly skipping of exons affected by mutations [24]. These splice variants can still provide partial functionality of the original protein or produce proteins with alternate functions. Fourth, ATI beginning at other start codon can give rise to various new protein variants [23,26,27]. For our B1 cells, all these possible mechanisms leading to different protein variants were taken into consideration. mRNA expression was assessed by qPCR with primers targeting exon 1 and exons 2–3 (Supplementary Figure S1A). NMD was deemed implausible since WT and B1 had overlying amplification plots of exon 1 specific primers (Supplementary Figure S1B). With the other primer pair, we analyzed a second section of the mRNA transcript sequence downstream of the mutation-bearing exon 1 and compared its expression with expression of the first target. Provided similar PCR efficiency, WT cells revealed a similar expression of both primer targets, while B1 cells expressed significantly more of the second target (Supplementary Figure S1C,D). This is consistent with the expression of a second transcript lacking exon 1, possibly due to AS. In addition, B1 DNA sequence was scanned for possible downstream ATI sites, revealing in-frame start codons in exon 1 (M37) and exon 2 (M59) (Figure 4A). Notably, according to predictions with the Peptide Mass Calculator (available at www.peptidesynthetics.co.uk/tools, accessed on 30 September 2022) a novel B1 protein variant, possibly originating from AS or ATI or even both, and lacking the mutated region of exon 1, would match with the approx. −4 to −6 kDa smaller protein detected by XPA antibody clones D9U5U and STJ96279 (Figure 4B). A3 cells (p.I69_D70delinsN/p.D70Hfs*31) possessed a frameshift mutation only in one allele, whereas the other was predicted to encode for a slightly altered XPA-variant with two AA lacking and a new one inserted (Figure 4A). Accordingly, immunoblot with all three tested XPA antibodies detected a protein band, however, it was much weaker than in the WT (Figure 4B). Correspondingly, the load of basal DNA damage was the lowest in immunofluorescence (Figure 5A), and the DNA repair capacity was least reduced in the HCR assay (Figure 5B). Surprisingly, this cell line was massively impaired in the MTT assay (Figure 6), which is commonly used to investigate cellular survival. The MTT assay reflects the cellular metabolic activity, which usually correlates with cell viability [28,29]. Nevertheless, several conflicting cases are documented in the literature showing that MTT is not always accurate in mapping cell numbers. Moreover, in some cases metabolism-independent cell viability assays contradicted those results generated by MTT [28,30,31,32]. Additionally, in XP patients clinical sun sensitivity and post-UV survival of patient’s cell lines, as assessed by MTT, did not always match with each other [33]. That is why we questioned whether the XPA mutations in A3 might influence metabolic activity and thereby could interfere with the MTT assay. The targeted mutation in A3, particularly the deletion of D70, occurred in a region of XPA exon 2, indispensable for recruitment and function of the endonuclease ERCC1 [14,15,17,34,35,36,37,38]. Indeed, ERCC1 depletion in cancer cells resulted in lower nicotinamide adenine dinucleotide phosphate hydrogen (NADPH), NADP+, NADH and NAD+ levels [39]. Because MTT assay is NAD(P)H-dependent, lower ERCC1 recruitment in A3 might explain their poor MTT performance [28,31]. The immunofluorescence was used to visualize 6-4PPs and CPDs as the main DNA lesions repaired by NER. Of note, in our CRISPR/Cas 9 altered cells a weak signal indicating DNA damage could be detected even without direct UVC irradiation. We assume that this DNA damage was acquired during the standardized handling of all cells including washing steps and mock irradiation. Importantly, this incidental UV exposure was not sufficient to cause detectable DNA lesions in wildtype cells, further supporting increased sensitivity of XPA mutated cells. In our study, a transfection based CRISPR/Cas 9 approach was performed due to its lower probability of off-target effects [40]. Nevertheless, and as one limitation of our study, off-target effects may not be fully excluded. To this end, it would be necessary to sequence the whole genome of the generated subclonal cell lines. However, the most probable predicted exonic off-targets—to the best of our knowledge—do not influence the cellular functions analyzed. Furthermore, only one parental melanoma cell line (A375) was originally utilized in our study, though it was applied to generate multiple subclonal cell lines. In conclusion, our study provides the first generation and detailed characterization of different CRISPR/Cas 9 XPA knockouts in A375 cells. In the context of our ICI-focused research, we will use these cell lines to investigate the immunological consequences of differently disrupted XPA proteins, including a novel XPA protein variant. Those experiments will involve the analysis of the expression of molecules, e.g., programmed-death-ligand 1 (PD-L1), that influences effective antitumor immune response and are thereby indispensable for the activity of anti-PD-1 ICI [41,42].
A375 melanoma cells (CRL-1619) were obtained from the American Type Culture Collection (ATTC, Manassas, VA, USA) and cultured in Dulbecco’s Modified Eagle Medium (DMEM) (Thermo Fisher Scientific, Waltham, MA, USA) with 10% Fetal Bovine Serum (PAN Biotech, Aidenbach, Germany), 100 U/mL Penicillin and 100 μg/mL Streptomycin (Merck, Darmstadt, Germany) at 37 °C with 5% CO2. The six XPA exons of parental A375 were sequenced (Sanger Sequencing at Eurofins Genomic, Ebersberg, Germany; primer sequences are presented in Supplementary Table S1) to validate the wildtype. All cells were regularly tested for mycoplasma contamination.
For the design of single-guide (sg)RNAs targeting the first two exons of XPA, CRISPR/Cas9 target online predictor (CCTop, University of Heidelberg, available at https://cctop.cos.uni-heidelberg.de:8043/index.html, accessed on 30 September 2022) was used. Oligonucleotides were ordered from Sigma-Aldrich (Merck, Darmstadt, Germany). pSpCas9(BB)-2A-Puro (PX459) V2.0 was a gift from Feng Zhang (Addgene plasmid #62988; http://n2t.net/addgene:62988, accessed on 30 September 2022; RRID:Addgene_62988). PX459 was digested with BbsI and phosphorylated using Calf Intestinal Alkaline Phosphatase. Multiple single-stranded sgRNA oligonucleotide pairs were annealed at 95 °C, dephosphorylated using T4 Polynucleotide kinase and ligated into the PX459 plasmid (all enzymes from New England Biolabs, Ipswich, MA, USA). Ligation products were transformed by heat shock into chemocompetent DH5α Escherichia coli bacteria (New England Biolabs, Ipswich, MA, USA). Successful integration of the oligonucleotides in PX459 was verified by Sanger sequencing.
The different sgRNA-PX459 plasmids were transfected into A375 cells (ViaFect Transfection Reagent, Promega, Madison, WI, USA), followed by selection with puromycin (InvivoGen, San Diego, CA, USA). Single-cell expansion of polyclonal lines was achieved by seeding of serially diluted cells in 96 well-plates. After DNA extraction (DNeasy, Qiagen, Hilden, Germany), Sanger sequencing was performed. For sequence analyses and sgRNA annotation SnapGene Viewer (GSL Biotech LLC, San Diego, CA, USA) was used. In case of heterozygous subclonal cell lines, superimposed peaks in the chromatogram were annotated as upper and lower peaks. For manual assignment, we analyzed those peaks searching for the wildtype sequence, that would follow an induced mutation after a possible insertion or deletion. Once identified, the peaks at each location were assigned to the alleles. The allele determinations were verified using the online tools DECODR, CRISPR-ID and Synthego (available at https://decodr.org/, http://crispid.gbiomed.kuleuven.be and https://ice.synthego.com, respectively. All accessed on 30 September 2022).
A total of 75,000 cells/well were seeded on glass cover slips in 24 well-plates and incubated for 48 h. Cells were either irradiated with 100 J/m2 UVC (UVC 500 Crosslinker, Amersham Biosciences, Buckinghamshire, UK) or left non-irradiated. Prior to all irradiation experiments cells were washed with phosphate-buffered saline (PBS, PAN Biotech, Aidenbach, Germany) and all liquid was carefully removed before irradiation with UVC. DMEM was replaced and the cells were incubated 90 min before fixation with 4% PFA (Merck, Darmstadt, Germany) in PBS. They were probed with antibodies against 6-4PP (clone 64M-2, Abcam, Cambridge, UK) and CPD (clone TDM-2, Absolute Antibody, Cleveland, UK). The secondary anti-mouse antibody was AlexaFluor488-coupled (Thermo Fisher Scientific, Waltham, MA, USA). Cells were mounted with Fluoroshield Mounting Medium With DAPI (Abcam, Cambridge, UK) and analyzed with ZEISS Axio Imager.M2 (Carl Zeiss, Wetzlar, Germany).
Total cellular proteins were extracted at 4 °C using radioimmunoprecipitation assay (RIPA) buffer (Thermo Fisher Scientific, Waltham, MA, USA) containing protease inhibitors (Roche, Basel, Switzerland). Protein concentrations were determined using Pierce BCA assay kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. Proteins (10–20 μg) were resolved on 10% SDS–polyacrylamide gels and transferred to Amersham Protra 0.45 μm NC nitrocellulose membranes (Thermo Fisher Scientific, Waltham, MA, USA). They were probed with antibodies against XPA with three different antibody clones: 12F5 (Santa Cruz Biotechnology, Dallas, TX, USA), D9U5U (Cell Signaling Technology, Danvers, MA, USA) and STJ96279 (St John’s Laboratory, London, UK). β-Actin (detected by antibody clone AC-15, Merck, Darmstadt, Germany) served as loading control. Detection using horseradish peroxidase (HRP)-conjugated secondary antibodies (Agilent Dako, Santa Clara, CA, USA) was performed in the ECL Chemocam Imager (Intas Science Imaging Instruments, Göttingen, Germany).
A total of 30,000 cells/well were seeded in 24 well-plates. After overnight incubation, XPA wildtype and CRISPR-rendered cells were transfected with non-irradiated renilla plasmid 50 ng/well (as transfection control), and either with non-irradiated or UVC-irradiated (1000 J/m2) firefly plasmid 250 ng/well using FuGene HD, according to the manufacturer’s protocol. After 48 h incubation, cell lysis, transfer to 96 well-plates and luminescence measurement using GloMax were performed according to the Dual-Luciferase Reporter Assay System protocol (Luciferase plasmids, Luciferase assay agents, transfection reagent and GloMax all from Promega, Madison, WI, USA). DNA repair capacities were determined by first calculating firefly to renilla luminescence ratios and then dividing irradiated to non-irradiated measurements. Finally, all calculated repair capacities were set in relation to those of A375 XPA wildtype (set as value 1).
In total, 10,000 cells/well were seeded in 96 well-plates 24 h prior to UVC. Initial dosis-finding with increasing UVC irradiation was performed with A375 wildtype cells in between 0 and 500 J/m2. Non-UVC-treated cells and wells containing only DMEM served as control and blank, respectively. After 48 h incubation in 100 µL DMEM, 15 µL of MTT dye solution per well were added. After 4 h incubation 100 µL stop solution (MTT dye and stop solution from Promega, Madison, WI, USA) were added, and plates were incubated overnight. Absorptions were captured with Tecan Sunrise (Tecan Trading AG, Männedorf, Switzerland) and the differences between the readings of two wavelengths (550 nm and 650 nm) were indicative of measured cellular metabolic activity. All results were set in relation to the respective non-irradiated cells (set as value 1).
For every single experiment, 250,000 cells/well were seeded each in three wells of a 6-well plate as biological replicates. After 48 h incubation, total cellular RNA was extracted (RNeasy Mini Kit, Qiagen, Hilden, Germany) and the concentration was measured (NanoVue Plus, Biochrom, Thermo Fisher Scientific, Waltham, MA, USA).
RNA was reverse transcribed at a concentration of 100 ng/µL (High Capacity cDNA Reverse Transcription Kit, Thermo Fisher Scientific, Waltham, MA), followed by 1:20 dilution. qPCR was performed in 384-well plates using QuantStudio™ 5 Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA) with 1 µL cDNA in a 10 µL reaction volume (PowerUp SYBR Green Master Mix, Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer´s protocol. Primer sequences are deposited in Supplementary Table S1. Actin beta was used as housekeeping gene for normalization. ΔΔCt method was applied to analyze obtained data.
Values are depicted as means ± SEM of data obtained from three independent experiments. Results were statistically analyzed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). Calculated relative DNA repair capacities, metabolic activity indicating cell survival and fold changes in mRNA expression were analyzed with one-way analysis of variance (ANOVA). *, **, *** indicate significance levels of p < 0.05, p < 0.01, p < 0.001, respectively. | true | true | true |
PMC9569992 | 36232436 | Tingzhen Wang,Mingjia Liu,Yang Wu,Yufeng Tian,Yingyan Han,Chaojie Liu,Jinhong Hao,Shuangxi Fan | Genome-Wide Identification and Expression Analysis of MAPK Gene Family in Lettuce (Lactuca sativa L.) and Functional Analysis of LsMAPK4 in High- Temperature-Induced Bolting | 22-09-2022 | MAPK,lettuce,genome-wide analysis,bioinformatics analysis,high temperature,expression pattern,bolting,virus-induced gene silencing,LsMAPK4 | The mitogen-activated protein kinase (MAPK) pathway is a widely distributed signaling cascade in eukaryotes and is involved in regulating plant growth, development, and stress responses. High temperature, a frequently occurring environmental stressor, causes premature bolting in lettuce with quality decline and yield loss. However, whether MAPKs play roles in thermally induced bolting remains poorly understood. In this study, 17 LsMAPK family members were identified from the lettuce genome. The physical and chemical properties, subcellular localization, chromosome localization, phylogeny, gene structure, family evolution, cis-acting elements, and phosphorylation sites of the LsMAPK gene family were evaluated via in silico analysis. According to phylogenetic relationships, LsMAPKs can be divided into four groups, A, B, C, and D, which is supported by analyses of gene structure and conserved domains. The collinearity analysis showed that there were 5 collinearity pairs among LsMAPKs, 8 with AtMAPKs, and 13 with SlMAPKs. The predicted cis-acting elements and potential phosphorylation sites were closely associated with hormones, stress resistance, growth, and development. Expression analysis showed that most LsMAPKs respond to high temperatures, among which LsMAPK4 is significantly and continuously upregulated upon heat treatments. Under heat stress, the stem length of the LsMAPK4-knockdown lines was significantly shorter than that of the control plants, and the microscope observations demonstrated that the differentiation time of flower buds at the stem apex was delayed accordingly. Therefore, silencing of LsMAPK4 significantly inhibited the high- temperature-accelerated bolting in lettuce, indicating that LsMPAK4 might be a potential regulator of lettuce bolting. This study provides a theoretical basis for a better understanding of the molecular mechanisms underlying the MAPK genes in high-temperature-induced bolting. | Genome-Wide Identification and Expression Analysis of MAPK Gene Family in Lettuce (Lactuca sativa L.) and Functional Analysis of LsMAPK4 in High- Temperature-Induced Bolting
The mitogen-activated protein kinase (MAPK) pathway is a widely distributed signaling cascade in eukaryotes and is involved in regulating plant growth, development, and stress responses. High temperature, a frequently occurring environmental stressor, causes premature bolting in lettuce with quality decline and yield loss. However, whether MAPKs play roles in thermally induced bolting remains poorly understood. In this study, 17 LsMAPK family members were identified from the lettuce genome. The physical and chemical properties, subcellular localization, chromosome localization, phylogeny, gene structure, family evolution, cis-acting elements, and phosphorylation sites of the LsMAPK gene family were evaluated via in silico analysis. According to phylogenetic relationships, LsMAPKs can be divided into four groups, A, B, C, and D, which is supported by analyses of gene structure and conserved domains. The collinearity analysis showed that there were 5 collinearity pairs among LsMAPKs, 8 with AtMAPKs, and 13 with SlMAPKs. The predicted cis-acting elements and potential phosphorylation sites were closely associated with hormones, stress resistance, growth, and development. Expression analysis showed that most LsMAPKs respond to high temperatures, among which LsMAPK4 is significantly and continuously upregulated upon heat treatments. Under heat stress, the stem length of the LsMAPK4-knockdown lines was significantly shorter than that of the control plants, and the microscope observations demonstrated that the differentiation time of flower buds at the stem apex was delayed accordingly. Therefore, silencing of LsMAPK4 significantly inhibited the high- temperature-accelerated bolting in lettuce, indicating that LsMPAK4 might be a potential regulator of lettuce bolting. This study provides a theoretical basis for a better understanding of the molecular mechanisms underlying the MAPK genes in high-temperature-induced bolting.
Mitogen-activated protein kinase (MAPK) is a serine/threonine protein kinase [1]. Through stepwise phosphorylation, MAPK, MAPKK (mitogen-activated protein kinase kinase), and MAPKKK (mitogen-activated protein kinase kinase kinase) collectively form a highly conserved MAPKKK → MAPKK → MAPK cascade, signal-transduction pathway [2]. When MAPKKK is activated by a sensor/receptor it can phosphorylate serine/threonine residues of S/TXXXXXS/T, in which X represents any amino acid, to activate downstream MAPKK, after which MAPKK continues to phosphorylate the conserved threonine and tyrosine residues in the TXY motif downward to activate MAPK [3,4,5,6,7,8]. MAPK cascades respond to almost all biological or abiotic stresses [9,10,11], such as mechanical damage [12], high salt [13], low temperature [14], drought [15] and pathogen invasion [16] and widely participate in the process of plant growth and development [17,18]. MAPK is located downstream of the entire MAPK signaling cascade. After phosphorylation, MAPK has three targets: cytoplasmic protein kinases or transcription factors (TFs), cytoplasmic cytoskeleton, and nuclear transcription factors [19]. Many plant MAPK genes have been studied, and sequence analysis of the whole genome showed that Arabidopsis thaliana has 20 MAPK genes, which can be separated into four groups, A, B, C, and D, according to their phosphorylation motif. The phosphorylation motif of groups A, B, and C was TEY, and that of group D was TDY [9,20]. The C-terminus of MAPK in groups A, B, and C contains the CD domain, which is speculated to be the binding region of upstream MAPKK, phospholipase, and various downstream substrates. The CD domain of group C is modified. Group D does not contain this region [21,22]. In addition to Arabidopsis, 15 MAPK genes were identified in rice [23], 11 in chrysanthemum [24], 43 in cultivated strawberry [25], and 12 in grape [26]. The functions of many MAPK genes have been identified and are known to participate in almost all activities in the plant life cycle, including plant growth and development, response to stress signals, activation of stress resistance gene expression, and improvement in adaptability to adversity [5]. MAPKs play an important role in biotic and abiotic stress responses. Arabidopsis AtMAPK4 has been reported to negatively regulate plant disease resistance and hypertonic resistance [27,28]. In addition, MAPKs are widely involved in hormone synthesis and signal transduction. AtMAPK6 is involved in regulating the polar transport of auxin and promotes ethylene synthesis together with AtMAPK3 [29,30]. Ethylene/jasmonic acid (JA) and MAPK3/MAPK6 cooperate to regulate plant defense [31,32]. ABA induces the expression of the MKK3–MAPK6–MYC2 cascade and actively regulates ABA biosynthesis and signal gene transcription [33]. Soybean GmMAPK4s negatively regulates salicylic acid accumulation and defense [34]. More importantly, MAPKs are largely involved in plant growth and development. Rice OsMAPK2 is involved in flower development, and AtMAPK4 is required for cell plate formation and cell division of male-specific meiosis in Arabidopsis [35]. The model TES/STUD/AtNACK2-AtANPs-AtMKK6-AtMPK4 was proposed to regulate cell meiosis [36,37]. In addition, MAPK also plays an important role in anther development [38,39], ovule development [40], pollen development [17], and seed development [41]. Lettuce is rich in nutrients and is an important component in modern diet and nutrition [42]. Lettuce is sensitive to high temperatures, which can induce premature bolting, resulting in poor growth and reductions in yield and quality. The symbol of bolting is flower bud differentiation, which is the process of rapid stem elongation after flower bud differentiation at the stem tip, accompanied by the transformation from vegetative growth to reproductive growth [43,44]. Bolting is restricted by its development and environment, such as hormones, temperature, and light, and its genetic characteristics, such as cell division, elongation, differentiation, and aging [45,46]. Moreover, the specific regulatory mechanism of high-temperature bolting in lettuce remains unclear, and MAPK has not been reported to play a role in the high temperature-induced bolting of lettuce. Early in this research, we screened differentially expressed MAPK cascade members involved in lettuce bolting at high temperatures, including LsMAPK3, LsMAPK4, LsMAPK9, and LsMAPKK5 protein, using comparative proteomics [47]. Recognizing the crucial role of MAPK in plant growth, development, and stress response, we conducted a genome-wide screening, identification, and bioinformatics analysis of the MAPK gene family in lettuce, determined the differential expression of the MAPK gene family at high temperature, and conducted a virus-induced gene silencing (VIGS) evaluation on LsMAPK4 with significant differential expression.
By homology comparison and domain analysis, we identified a total of 17 dependable LsMAPK gene family members in the lettuce genome. According to the unified nomenclature of Arabidopsis AtMAPKs, we comprehensively analyzed the evolutionary relationship between LsMAPKs and AtMAPKs (Figure 1). Based on the analysis, we systematically renamed them according to the phylogenetic relationship (LsMAPKs in Table 1). Among them, LsMAPK4 and LsMAPK4* are identical genes (Figure S1). The corresponding gene name, gene registration number, chromosome position, isomer numbers, protein size, molecular weight, isoelectric point (pI), subcellular localization, and transmembrane domain are summarized in Table 1. The lengths of the LsMAPK proteins ranged from 284 (LsMAPK4-3) to 782 (LsMAPK16-4) amino acids, with a molecular weight ranging from 33.00 to 89.70 kDa. The isomer numbers varied from 1 to 4, and the theoretical pI value scope was from 4.74 (LsMAPK4-3) to 9.08 (LsMAPK16). The subcellular localization of all MAPKs was located in the nucleus without a transmembrane domain. Chromosome mapping results showed that the 17 LsMAPKs were unevenly distributed on 8 lettuce chromosomes, ranging in number from 1 to 3, with none on chromosome 6 (Table 1, Figure 1).
To analyze the phylogenetic relationship and classification of LsMAPKs, a phylogenetic tree was constructed based on AtMAPKs of Arabidopsis thaliana and LsMAPKs of lettuce (Figure 2). LsMAPK members were clustered into four groups: A, B, C, and D, which included 3 (LsMAPK3/3-2/6), 4 (LsMAPK4/4*/4-2/4-3), 4 (LsMAPK1/1-2/1-3/7), and 6 (LsMAPK9/9-2/16/16-2/16-3/16-4) LsMAPKs, respectively. Multisequence alignment analysis showed that all LsMAPKs contained a highly conserved tripeptide motif of TXY (Figure S2). Groups A, B, and C contain TEY, and group D contains TDY. Group A and B members also contain the conserved CD domain in (LH)DXXDE(P)XC, which is considered to be a binding site for MAPKK. In addition, more plant species were used to analyze the differentiation of the MAPK gene family in plants (Figure 3), including Physcomitrella patens, Chlamydomonas reinhardtii, Selaginella moellendorffii, Arabidopsis thaliana, Oryza sativa, Zea mays, Solanum lycopersicum, and Populus trichocarpa. All MAPK genes fell into five groups, and the new group of MAPKs (group O) was mainly composed of lower eukaryotic and fern members, which were CreinMAPK4-1, CreinMAPK7, and SmMAPK10. Compared with higher eukaryotic angiosperms, lower plants encode very few MAPK genes. Among mosses, algae, and ferns, C. reinhardtii lacks the ‘group B’ type of MAPKs. All three species lacked the ‘group A’ type. It can also be seen from the phylogenetic tree that group A had the fewest members and group D had the most members.
We conducted analyses to characterize the sequences of the 17 LsMAPKs by motif distribution, conserved domain, and exon–intron structure (Figure 4). Fifteen distinct motifs were predicted (Figure 4B), among which motifs 1-8 were highly conserved and appeared in almost the same order and position in all LsMAPKs proteins, although LsMAPK16 lacked motif 8 and LsMAPK9-2 lacked motifs 5 and 8. Motifs 9 and 10 were unique to groups A and B, and motif 15 was specific to group D. Compared with other group D members, LsMAPK16, LsMAPK16-2, and LsMAPK16-4 specifically had motifs 11-14. Conservative domain analysis showed that all LsMAPKs proteins had the same PKc-like superfamily domain (Figure 4C), in which groups A, B, and C generally had the STKc_TEY_MAPK domain, whereas group D generally had the STKc_TDY_MAPK domain. LsMAPK16, 16-2, and 16-4 had one additional domain. The results of exon–intron structure analysis showed that the number of exons and introns in groups A, B, and C was relatively conservative (Figure 4D). In contrast, the number of introns and exons in group D was distinctly different from that in groups A, B, and C, and the numbers in Group D were generally greater than those in groups A, B, and C.
Collinearity analysis can further reveal the potential evolutionary mechanism of species’ gene families. Therefore, the evolutionary relationship and homology between LsMAPK members, lettuce and Arabidopsis, and lettuce and tomato were predicted. Among the 4291 collinear pairs in lettuce, five collinear pairs belong to LsMAPKs, among which LsMAPK1-2 was collinear with LsMAPK1-3, LsMAPK1, and LsMAPK7 at the same time (Figure 5). Furthermore, the comparison of lettuce with Arabidopsis and tomato at the genomic level showed that 12,492 collinear pairs were obtained between lettuce and Arabidopsis, and 15,558 collinear pairs were obtained between lettuce and tomato (Figure 6), among which the numbers belonging to MAPK members were 8 and 13, respectively. The collinearity pairs of LsMAPK1, 4, and 7 accounted for a total of 12 pairs, indicating the substantial role of these genes in evolution.
After cis-acting element analysis of the LsMAPK promoter region, a total of 64 types of elements were identified and divided into four subgroups: phytohormones, plant development, stress responsiveness, and light responsiveness (Figure 7). The members of each subgroup were evenly distributed on all MAPKs, with the most elements related to stress response and the fewest elements related to development (Figure 7A–D). Among a total of 706 cis-acting elements, MYB elements occurred most frequently at 101 times, followed by MYC elements, which appeared 77 times (Figure 7E). However, MYC elements were absent from LsMAPK16-4. Members with close relatives may not show similar types of cis-acting elements. Here, for example, LsMAPK16 in group D, LsMAPK3 in group A, and LsMAPK4 and 4* in group B were classified into the same group according to the type and number of cis-acting elements, suggesting that they may be closely related in function.
Activation of protein kinases is usually achieved by phosphorylation, so the analysis of phosphorylation sites is vital to understanding the mechanisms of protein action. The predicted phosphorylation sites of LsMAPK gene family members showed that a total of 1131 reliable phosphorylation sites (the same site could be phosphorylated by multiple kinases) were widely distributed in the sequence (Figure 8A). The number of nonspecific sites (phos-unsp) was the largest, with a total of 436, followed by phosphokinase C-specific sites (phos-PKC), with a total of 154, and casein kinase II sites, with a total of 105 (Figure 8B). Only one specific phosphorylation site of protein kinase B (phos-PKB) was predicted in LsMAPK16 and 16-2. LsMAPK16, 16-2, and 16-4 had the most nonspecific and total phosphate sites, which may be attributed to their sequence length. However, the number of phosphorylation sites was not necessarily related to the length of the amino acid sequence. LsMAPK4-3 had the shortest amino acid sequence length, but its phosphorylation sites were still more abundant than those of LsMAPK1, 3, and 7.
To investigate the potential regulatory mechanism of the MAPK gene in high-temperature bolting, the expression patterns of LsMAPKs were quantitatively assessed in response to high temperature in the stems of lettuce (Figure 9). Our previous research demonstrated that the experimental material GB-30 generally began bolting on the 8th day after high-temperature treatment. The results showed that the 16 LsMAPKs demonstrated a specific expression pattern during high-temperature bolting in lettuce. The expression levels of LsMAPK1, LsMAPK1-3, LsMAPK4-2, and LsMAPK7 on the 8th day were significantly lower than those of the control. Among them, no difference was found on the 16th day in LsMAPK1, LsMAPK1-3, and LsMAPK7. On the 24th day, the expression of LsMAPK1 and LsMAPK7 was lower, but that of LsMAPK1-3 was higher. The expression of LsMAPK4-2 remained lower than that of the control after HT treatment. The transcriptional levels of LsMAPK1-2, LsMAPK4-3, and LsMAPK16 were higher than those of the control only on the 16th day, showing no difference on the 8th and 24th days after HT treatment. LsMAPK3-2, LsMAPK4, and LsMAPK6 were abundantly expressed after HT treatment. LsMAPK3-2 peaked on the 8th day, and LsMAPK4 and LsMAPK6 peaked on the 24th day. The expression level of LsMAPK9-2 was higher than that of the control on the 8th and 24th days and showed no difference on the 16th day. LsMAPK16-2, LsMAPK16-3, and LsMAPK16-4 presented similar expression patterns. Their expression levels were substantially higher than those of the control on the 8th and 16th days, and except that LsMAPK16-4 was lower, there was no difference on the 24th day. In addition, the expression level of LsMAPK3 was higher than that of the control on the 8th day, then decreased, and was lower than that on the 16th and 24th day. Overall, qRT–PCR analysis suggested that LsMAPKs had a high likelihood of involvement in the bolting of lettuce induced by high temperature.
In the previous comparative proteomics, we found that among LsMAPKs proteins, LsMAPK3/4/9 were differentially expressed in high-temperature bolting of leaf lettuce [47]. The subsequent qRT–PCR indicated that the transcription level of LsMAPK4 among the above three LsMAPKs was markedly and continuously upregulated during high-temperature bolting. To further explore gene function, VIGS-mediated LsMAPK4 gene silencing in lettuce was conducted. A 355 bp fragment of LsMAPK4 was successfully cloned into the pTRV2 vector (Figure 10A). Three weeks after Agrobacterium infection, RT–PCR amplification results showed that the pTRV2 empty vector and pTRV2-LsMAPK4 vector had been transferred into lettuce and effectively expressed (Figure 10B,C). Then, to finally determine whether LsMAPK4 was silenced, the expression was measured by qRT–PCR (Figure 10D). The results showed that the expression of LsMAPK4 in the silenced group was significantly lower than that in the blank control group and the empty vector group. In the third week after the VIGS infection, the three groups of lettuce were treated with HT. There was no significant difference in stem length among the three groups at 0 and 2 days after HT treatment (Figure 10E). On the 4th day, the stem length of the silent group began and continued to be significantly shorter than that of the blank control group and the empty vector group. On the 8th day, the stems of the blank control and the empty vector group elongated rapidly, and both of them began bolting (Figure 10F). At this time, the flower bud differentiation status of the stem tips of the three groups was detected by paraffin sectioning to judge whether the lettuce was bolting (Figure 10G). The results showed that the growth cone of the stem tips of the blank control and the empty vector group was round, blunt, and hypertrophic, without obvious protuberance, and had entered the early stage of flower bud differentiation. However, the growth cone of the silent group was semicircular with obvious protuberances, indicating that it was still in the vegetative growth stage and had not yet entered the early stage of flower bud differentiation to start bolting. Overall, silencing LsMAPK4 significantly delayed the bolting of lettuce under high temperatures, indicating that LsMAPK4 may play a promoting role in high-temperature bolting in lettuce.
The MAPK cascade is a functional module that widely exists in eukaryotes [48] and has been well studied in Arabidopsis, rice, and other plants. However, there are few reports regarding its occurrence in lettuce. Our study identified a total of 17 MAPK family members in the lettuce genome, of which LsMAPK4 and LsMAPK4* were found to be identical. The 17 MAPK family members found in lettuce is fewer than the number in Arabidopsis (20 genes) [49], corn (19 genes) [50], and poplar (21 genes) [51] but more than the number in tomato (16 genes) [52], C. reinhardtii (6 genes), P. patens (10 genes), and S. moellendorffii (6 genes) [53] and the same as the number found in rice [51]. This indicates that there is no correlation between the number of MAPKs in plants and a plant species’ genome size. The chromosome distribution pattern of the MAPK family indicates that they are located on chromosomes individually or in clusters, but they do not occur on all chromosomes. None of the LsMAPK family members are located on chromosome 6, just as no MAPK members are distributed on chromosome 5 of Brachypodium distachyon [54] or chromosome 3 of cucumber [55]. Through phylogenetic analysis and multiple sequence alignment, LsMAPK family members can be divided into four groups according to the TXY motif. To study the molecular evolution and phylogenetic relationship of MAPKs in lettuce, we conducted phylogenetic analysis on 132 MAPKs from different plant groups. In lettuce, dicotyledons, and monocotyledons, MAPKs in group A have orthologs and paralogs of MAPK3 and MAPK6 but no orthologs of MAPK10. MAPK3 and MAPK6 are either present together or absent, and the decisive functions of these two MAPKs in growth and development and response to biotic and abiotic stresses have been confirmed in various plants [41,56,57,58], implying that MAPK3 and MAPK6 orthologs are indispensable. Another member of group A, the ortholog of MAPK10, was identified only in mustard plants [53], indicating that the ortholog of MAPK10 is only conserved in the Brassicaceae family of dicotyledons and has been lost from other families during evolution. In group B, MAPK5 and MAPK11 are similar to MAPK10, and MAPK11 has been lost even in the mustard family. In lettuce, all group B LsMAPKs are orthologs or paralogs of MAPK4, and there are two identical LsMAPK4s, indicating that a wide range of repetitive events may occur in its genome. LsMAPKs in group C exist in the form of MAPK1 and MAPK7, similar to the tomato genome. Group D is a type of MAPK unique to plants. None of the studied species lacks group D MAPKs. With the increase in biological complexity, the group D MAPKs expand through gene replication [22], which is indispensable in the green plant lineage. Collinearity analysis further revealed the potential evolutionary mechanism of the species gene family. There were five collinear pairs among LsMAPK members of lettuce and eight pairs among AtMAPKs of Arabidopsis thaliana. There were 13 pairs of SlMAPKs in tomato. This indicated that the genetic relationship between tomato and lettuce was closer than that of Arabidopsis. Gene structure is closely related to gene expression and function. All LsMAPKs proteins contain motifs 1-7 (except LsMAPK4-3, which lacked motif 3). The C-terminal motif is more conserved than the N-terminal motif. There are unique characteristic motifs among distinct groups, indicating their evolutionary relationship and different functional divisions [59]. Transcript information indicates that LsMAPKs have different exon arrays in their genes. The distribution of introns and exons in the gene structure of LsMAPKs showed regularity among diverse groups. Both group A and group B contain 6 exons, group C contains 2–3 exons, and group D contains more exons, including 8–12 introns. Similar structural patterns were also observed in other plants, with high conservation within groups and an elevated level of variation between groups [49,53]. SmMAPK10 of S. moellendorfii contains up to 15 exons, and some intron-free MAPKs also exist in higher eukaryotic plants, such as PaMAPK2, PaMAPK3, PaMAPK7-1, and PaMAPK20 of Picea abies [53]. The plant MAPK cascade pathway mediates the response to salt, drought, cold, high temperature, and pathogenic bacteria [5] and is cross-linked with the hormone signaling pathway to regulate plant growth and development [60,61]. The promoter of the LsMAPK gene is rich in various cis-acting elements, such as ABA response elements (ABREs), MYB elements, MYC elements, methyl jasmonate reaction motifs (TGACG motifs and CGTCA motifs), root-specific expression elements (as-1 s), and various light response elements, which is strong evidence that LsMAPKs may participate in various life activities. Analysis of phosphorylation sites is very important for understanding the mechanism of action of the LsMAPK protein kinase. In addition to nonspecific sites, phosphokinase C (phos-PKC), casein kinase II (phos-CKII), and cell division cyclin 2 (phos-cdc2) have the largest number of specific phosphorylation sites, which have been reported to participate in plant life activities such as stress resistance, cell apoptosis, flowering regulation, and cell division [62,63,64,65]. It is speculated that LsMAPK is related to growth and development and the regulation of stress resistance. Plants suffer from various adversity stresses during their growth and development. Among them, high-temperature stress affects almost all aspects of plant growth, development, reproduction, and yield, resulting in shedding, flowering, fruit abortion, and even whole plant death [15,66,67]. Activation of MAPK plays a key role in the perception and transduction of temperature signals [68]. MAPKs in B. distachyon are temperature-sensitive, 60% of which are induced by high temperature [69]. High temperature can induce the activity of AtMAPK6 in Arabidopsis, but it has also been reported that AtMAPK6 negatively regulates the high-temperature response [70,71]. Most cucumber CsMAPKs (except CsMAPK3 and CsMAPK7) were upregulated after high-temperature treatment [55]. In chrysanthemum, except for the increased expression levels of CmMAPK4.1, CmMAPK6, and CmMAPK13 after 1 h of heat shock treatment, the other MAPKs decreased or remained unchanged [24]. The high-temperature treatment times used in the above literature were mostly between 1 and 8 h. In this paper, the changes in LsMAPK expression in lettuce after high-temperature treatment were continuously monitored and detected at 0, 8, 16, and 24 days after high-temperature treatment with a higher sampling frequency and longer observation time. Among them, the most obvious were LsMAPK3-2, LsMAPK4, and LsMAPK6, which were significantly and continuously higher than the control after high temperature. LsMAPK4-2 was significantly and continuously lower than the control, and the rest of the LsMAPKs had a tendency to decrease or to rise first and then decrease compared with the control, but not every stage had a significant difference. At present, studies on MAPK4 gene function mainly focus on cell division and stress response. Tomato MAPK4 [72], rice MAPK4 [73], and Arabidopsis MAPK4 [74] have been reported to play a role in low-temperature stress. In addition, MAPK4 was more reported to be related to cell division, flower organ formation, and photosynthesis [18,37,75], and little has been reported about the high-temperature stress response. In our study, we screened LsMAPK4 and LsMAPK4* on chromosome 3, which are located in different positions but are very close to each other, with a distance of about 180,000 bp. Their gene sequences and open reading frame sequences are identical. No same gene was found in the front and back of the two chromosome positions, indicating that this is a small and segmental duplication event. Compositae species have experienced the triplication of the whole genome in the evolutionary process, and leaf lettuce also shows the triplication of the genome, which leads to a large number of replication events [76]. The segmental duplication of LsMAPK4 is probably the result of the retention of duplicate copies. We have noticed that the promoter sequences of LsMAPK4 and LsMAPK4* were completely different, and there were differences in cis-acting elements (Figure 7), which may lead to functional differences between the two genes. But it is also very likely that the functions of the two genes will overlap under a specific condition. The gene replication of LsMAPK4 is likely to make its expression more extensive and efficient. Since the expression of LsMAPK4 was significantly upregulated by high temperature, we further conducted VIGS experiments to explore the function of LsMAPK4 in the high-temperature bolting of lettuce. According to the cytological and morphological experiments of flower buds and stems, after 8 days of high-temperature treatment, bolting occurred in the blank control group and the empty vector group but not in the silenced group. Easy-bolting lettuce GB-30 was used as the experimental material. After high-temperature treatment, GB-30 began bolting on the 8th day [46]. This is very consistent with the VIGS phenotype. High temperature significantly promoted the expression of LsMAPK4, which played a positive regulatory role in the bolting process of lettuce at high temperatures.
Whole-genome data for lettuce (https://phytozome.jgi.doe.gov/pz/portal.html; accessed on 6 August 2021) were downloaded and constructed into a local protein database. The Arabidopsis thaliana MAPK protein sequences from the Arabidopsis thaliana genome database (http://www.arabidopsis.org/; accessed on 6 August 2021) were used as the search seed for a BLASTP homology search of lettuce whole-genome data with an e-value of 1 × 10−5 and a minimum identity of 50% as a threshold. The HMMER3.0 program (http://hmmer.org/; accessed on 10 September 2021) was used to remove redundant sequences by applying the serine/threonine protein kinase-like domain (PF00069) as a query for hidden Markov model (HMM) searches with a threshold of 1 × 10−5.
By querying the annotation file of the lettuce genome, the chromosome location, direction, and isomer number of MAPK proteins were obtained, and the chromosome location was visualized by TBtools. The amino acid number, isoelectric point, and molecular weight of the MAPK protein were predicted by using the ProParam tool (https://web.expasy.org/protparam/; accessed on 5 July 2022). Cell-PLoc 2.0 was used to predict subcellular localization (http://www.csbio.sjtu.edu.cn/bioinf/Cell-PLoc-2/; accessed on 5 July 2022). TMHMM was used to predict transmembrane domains (https://dtu.biolib.com/DeepTMHMM; accessed on 5 July 2022). The MAPK members of each species were obtained from the corresponding plant database, and the species database was downloaded from Phytozome V13 (https://phytozome-next.jgi.doe.gov/; accessed on 11 July 2022). The construction of the phylogenetic tree was completed by MEGA 7.0 software and enhanced using Evolview (http://www.evolgenius.info/evolview/#/login; accessed on 13 July 2022). The conserved domains of proteins were analyzed by NCBI-CD search (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi; accessed on 12 July 2022). MEME Suite analyzed conserved motifs of proteins (http://meme-suite.org/tools/meme; accessed on 12 July 2022). The cis-acting elements were analyzed by Plant CARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/; accessed on 6 July 2022). Analysis of phosphorylation sites was completed by using netphos (https://services.healthtech.dtu.dk/service.php?NetPhos-3.1; accessed on 24 July 2022). Intron–exon structure, whole genome collinearity analysis of lettuce, and collinearity analysis of lettuce with Arabidopsis and tomato were analyzed by TBtools. The visualization of all the above analyses was completed by TBtools and further enhanced using Adobe Illustrator.
Easily bolting lettuce variety ‘GB30’ was used as the experimental material, which was from seeding plant protection Co., LTD in Inner Mongolia Bameng Fidelity, conserved in our laboratory. It was planted in a matrix of sand, soil, and peat (1:1:1 by volume) and cultured in the computer greenhouse of Beijing Agricultural College under the following conditions: temperature of 20 ± 2 °C (day)/13 ± 2 °C (night), photoperiod of 14 h (day)/10 h (night), relative humidity of 60 ± 5%, and light intensity of 12,000 lux. The plants were transplanted at the trefoil stage, and high-temperature treatment was carried out when lettuce had six true leaves [46]. The control group maintained the original culture conditions, whereas the high-temperature group increased the temperature to 33 ± 2 °C (day)/25 ± 2 °C (night), with the other conditions unchanged. Samples were taken at 0, 8, 16, and 24 days of treatment, and the sampling site was the stem. Every five plants were used as a replicate, and the sampling was repeated three times independently.
Total RNA was extracted using a Quick RNA Isolation kit (Huayueyang Biotech, Beijing, China). The first-stand cDNA of lettuce was synthesized by TransScript® Uni All-in-One First-Strand cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China). Primer3 Plus software was used to design qRT–PCR primers (Table S1), and 18S rRNA (HM047292.1) was used as an internal reference gene. qRT–PCR was performed using TB Green® Premix Ex Taq™ II (Takara Bio, Beijing, China) and a CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA). The amplification program was as follows: pre-denaturation at 95 °C for 3 min, denaturation at 95 °C for 20 s, and annealing at 55 °C for 30 s for 40 cycles. The 2−ΔΔCT method was used to analyze the data.
A 355 bp sequence in the CDS region of LsMAPK4 was amplified by VIGS-LsMAPK4-F/R primers (Table S1). The pTRV2 vector was linearized by EcoRI and XhoI. A ClonExpress Ⅱ One Step Cloning Kit (Vazyme Biotech, Beijing, China) was used to complicate the homologous recombination of the amplified band and linearized vector. After the recombinant plasmid was transferred into Trans1-T1 Phage Resistant Chemically Competent Cells and sequenced correctly, it was transferred into GV3101 Chemically Competent Cells (Weidi Biotechnology, Shanghai, China). Before VIGS injection, the same amount of pTRV11 and pTRV2 bacterial solution was mixed. There were 30 lettuces in each group, and there were three treatment groups in total: the blank control group (no injection), empty vector group (injection with equal pTRV2 and pTRV1), and silent group (injection with equal pTRV2-LsMAPK4 and pTRV1). Three weeks after injection, a high-temperature treatment was carried out. Before treatment, RNA from new leaves was taken to detect whether the vector was transferred and expressed, and the expression level of LsMAPK4 was measured. The vector detection primers were pTRV2-F/R for the empty vector group and VIGS-LsMAPK4-F/R for the silencing group (Table S1). The stem length of each group was measured at 0, 2, 4, 6, 8, and 10 days after high temperature, and the stem tips of the plants were paraffin-sectioned on the 8th day.
All tests were performed in triplicate. One-way ANOVA was performed on the data using statistical analysis software SPSS 12.5 (International Business Machine, Chicago, IL, USA), and graphs were drawn by Origin 9 (Origin Lab, Northampton, MA, USA). The standard error is indicated. The * stands for p < 0.05 and the ** stands for p < 0.01, followed by Student’s t-test. Different letters represent significant differences as determined using one-way ANOVA followed by Duncan’s test. p < 0.05.
In this study, we identified a total of 17 MAPK gene family members from the whole genome of lettuce. The physicochemical properties, chromosomal localization, phylogeny, gene structure, family evolution, cis-acting elements, and phosphorylation sites were comprehensively analyzed. To explore the role of LsMAPKs in high-temperature bolting of lettuce, the expression pattern of LsMAPKs was evaluated within 24 days after high-temperature treatment, and the positive regulatory role of LsMAPK4 in high-temperature bolting was verified by VIGS. This provides a theoretical basis for understanding high-temperature bolting mechanisms in lettuce and identifies potential paths for improving the annual production of lettuce. | true | true | true |
PMC9570013 | 36233293 | Hui-Min Zhang,Zi-Yi Li,Zhou-Tong Dai,Jun Wang,Le-Wei Li,Qi-Bei Zong,Jia-Peng Li,Tong-Cun Zhang,Xing-Hua Liao | Interaction of MRPL9 and GGCT Promotes Cell Proliferation and Migration by Activating the MAPK/ERK Pathway in Papillary Thyroid Cancer | 09-10-2022 | MRPL9,GGCT,Papillary Thyroid Cancer (PTC),proliferation,migration | Thyroid cancer remains the most common endocrine malignancy worldwide, and its incidence has steadily increased over the past four years. Papillary Thyroid Cancer (PTC) is the most common differentiated thyroid cancer, accounting for 80–85% of all thyroid cancers. Mitochondrial proteins (MRPs) are an important part of the structural and functional integrity of the mitochondrial ribosomal complex. It has been reported that MRPL9 is highly expressed in liver cancer and promotes cell proliferation and migration, but it has not been reported in PTC. In the present study we found that MRPL9 was highly expressed in PTC tissues and cell lines, and lentivirus-mediated overexpression of MRPL9 promoted the proliferation and migration ability of PTC cells, whereas knockdown of MRPL9 had the opposite effect. The interaction between MRPL9 and GGCT (γ-glutamylcyclotransferase) was found by immunofluorescence and co-immunoprecipitation experiments (Co-IP). In addition, GGCT is highly expressed in PTC tissues and cell lines, and knockdown of GGCT/MRPL9 in vivo inhibited the growth of subcutaneous xenografts in nude mice and inhibited the formation of lung metastases. Mechanistically, we found that knockdown of GGCT/MRPL9 inhibited the MAPK/ERK signaling pathway. In conclusion, our study found that the interaction of GGCT and MRPL9 modulates the MAPK/ERK pathway, affecting the proliferation and migration of PTC cells. Therefore, GGCT/MRPL9 may serve as a potential biomarker for PTC monitoring and PTC treatment. | Interaction of MRPL9 and GGCT Promotes Cell Proliferation and Migration by Activating the MAPK/ERK Pathway in Papillary Thyroid Cancer
Thyroid cancer remains the most common endocrine malignancy worldwide, and its incidence has steadily increased over the past four years. Papillary Thyroid Cancer (PTC) is the most common differentiated thyroid cancer, accounting for 80–85% of all thyroid cancers. Mitochondrial proteins (MRPs) are an important part of the structural and functional integrity of the mitochondrial ribosomal complex. It has been reported that MRPL9 is highly expressed in liver cancer and promotes cell proliferation and migration, but it has not been reported in PTC. In the present study we found that MRPL9 was highly expressed in PTC tissues and cell lines, and lentivirus-mediated overexpression of MRPL9 promoted the proliferation and migration ability of PTC cells, whereas knockdown of MRPL9 had the opposite effect. The interaction between MRPL9 and GGCT (γ-glutamylcyclotransferase) was found by immunofluorescence and co-immunoprecipitation experiments (Co-IP). In addition, GGCT is highly expressed in PTC tissues and cell lines, and knockdown of GGCT/MRPL9 in vivo inhibited the growth of subcutaneous xenografts in nude mice and inhibited the formation of lung metastases. Mechanistically, we found that knockdown of GGCT/MRPL9 inhibited the MAPK/ERK signaling pathway. In conclusion, our study found that the interaction of GGCT and MRPL9 modulates the MAPK/ERK pathway, affecting the proliferation and migration of PTC cells. Therefore, GGCT/MRPL9 may serve as a potential biomarker for PTC monitoring and PTC treatment.
Thyroid cancer (TC) is currently the most common endocrine system malignancy [1]. The incidence of thyroid cancer has been on the rise worldwide over the past few decades [2,3]. The most common histological type of TC is papillary thyroid carcinoma (PTC), which accounts for approximately 80–90% of all TC cases [4,5]. Although the vast majority of PTC patients have a good prognosis with conventional treatment, recurrence and distant metastasis occur in approximately 30% of patients, thus significantly reducing survival [6,7]. In addition, the occurrence and progression of PTC are influenced by multiple factors, such as genetic mutations, environmental exposures, and epigenetic alterations [8]. Therefore, new molecular biomarkers should be discovered and used as useful tools for diagnosis and treatment management to better characterize the malignancy and aggressiveness of lesions, while providing theoretical support for the search for new PTC diagnostic and therapeutic targets. Mitochondrial proteins (MRPs) are essential components of the structural and functional integrity of the mitochondrial ribosomal complex [9]. In mammals, more than 80 Mrp genes have been identified, and this group of genes is divided into two broad categories: Mrpl, a component of the large subunit, and Mrps, a component of the small subunit [10]. MRPL9 genes are a component of the mitochondrial ribosomal large subunit, which mediates translation in mitochondria [11]. It has been reported that MRPL9 is highly expressed in liver cancer and is associated with poor prognosis [12,13], and knockdown of MRPL9 inhibits the proliferation and migration of liver cancer cells [13]. Additionally, loss-of-function of MRPL9 inhibited the colony-forming unit (CFU) potential of MDA-MB-231 and BT-549 TNBC models and increased their sensitivity to paclitaxel [14]. It can be seen that MRPL9 may be related to the proliferation and migration ability of cancer cells. However, the function of MRPL9 in PTC is unknown. GGCT (γ-glutamylcyclotransferase) is one of the major enzymes in glutathione metabolism [15,16], catalyzing the reaction to generate 5-oxoproline and free amino acids from γ-glutamyl peptide. Oakley et al., cloned a cDNA encoding human GGCT and found that GGCT is identical to the putative protein chromosome 7 open reading frame 24 (C7orf24), which was previously registered as a putative open reading on the chromosome 7 (7p15-14) frame [17]. Before Oakley’s report, C7orf24 was known as a cancer-associated protein. Xu et al., identified 46 common cancer signature genes from a pooled DNA array database of previously reported human cancers and reported that one of the highly expressed genes was C7orf24 [18]. In addition, Kageyama et al. also characterized C7orf24 as an up-regulated protein in urothelial carcinoma specimens by proteomic analysis [19,20]. Studies have shown that GGCT is highly expressed in tumorous breast tissue [21], and that patients with high GGCT expression have a poor prognosis. In addition, GGCT is highly expressed in colorectal cancer [22], gastric cancer [23], prostate cancer [24], and human glioma [25]. Our previous study showed that GGCT is highly expressed in PTC, and knockdown of GGCT inhibited the migration ability of PTC cells [26]. In this paper, we further explore the molecular mechanism of GGCT regulating the proliferation and migration of PTC cells on the basis of previous research, and provide theoretical support for the treatment of PTC. In this study, we evaluated MRPL9 expression levels at the clinical and cellular levels and the effects of MRPL9 deletion or enhancement on the biological behavior of PTC cells. In addition, the interaction between MRPL9 and GGCT was demonstrated, and the regulatory effect of MRPL9/GGCT on the MAPK/ERK signaling pathway was further explored. Finally, the effects of MRPL9 and GGCT on tumor growth and metastasis were further investigated in vivo.
TCGA database analysis found that MRPL9 was highly expressed in PTC (Figure 1A). In addition, we found that the expression level of MRPL9 was associated with the prognosis of thyroid cancer, and patients with high MRPL9 expression showed poorer overall survival (http://gepia2.cancer-pku.cn/#index, accessed on 21 March 2022) (Figure 1B). To further explore the expression level of MRPL9 in PTC, 26 pairs of cancer tissues and adjacent tissues from PTC patients were collected. Specimens were collected for immunohistochemical staining, and the results were scored and counted. The results showed that MRPL9 was highly expressed in PTC tissues (Figure 1C,D). We further divided patients into a low expression group (n = 13) and a high expression group (n = 13) according to the median expression level of MRPL9. The correlation between the expression level of MRPL9 and the clinicopathological characteristics of PTC patients, such as age, gender, TNM stage, multifocality, distant metastasis, and lymph node metastasis, is shown in Table 1. The high MRPL9 expression group was associated with a higher degree of TNM progression, extrathyroidal extension and lymph node metastasis (p ≤ 0.05). However, there was no significant correlation between MRPL9 level and patient’s gender and age (p > 0.05). The protein expression level of MRPL9 in patients’ tissues was further analyzed by western blot experiment, and western blot quantitative analysis was performed. The results showed that MRPL9 was highly expressed in PTC cancer tissues (Figure 1E, F). The above results indicated that MRPL9 was highly expressed in PTC tissues. In addition, we further explored the expression level of MRPL9 in PTC cell lines, qRT-PCR and Western blot results showed that, compared with Nthy-ori 3.1 cell lines, the mRNA and protein levels of MRPL9 were higher in TPC-1, K1 and BCPAP cell lines (Figure 1G–I).
In view of the high expression of MRPL9 in PTC cell lines and to explore the role of MRPL9 in PTC cell lines, K1 and BCPAP cell lines were infected with lentivirus to obtain stable cell lines that stably overexpressed or stably knocked down MRPL9. Western blot and quantitative results showed that the protein level of MRPL9 in the LV- MRPL9 group was significantly increased in K1 and BCPAP cell lines compared with the LV-Vector group (Figure 2A,B). Compared with the LV-shNC group, the protein expression level of MRPL9 in the LV-sh MRPL9 group was significantly down-regulated in K1 and BCPAP cell lines (Figure 2C,D). This indicated that we had successfully constructed cell lines stably overexpressing or stably knocking down MRPL9. The cell clone formation assay showed that K1 and BCPAP cells grew faster in the overexpression MRPL9 group. Conversely, knockdown of MRPL9 inhibited cell growth (Figure 2E,F). The results of wound healing and transwell assays showed that overexpression of MRPL9 promoted the migration ability of K1 and BCPAP cells, while knockdown of MRPL9 had the opposite function (Figure 2G–J). The above results indicated that MRPL9 promoted the proliferation and migration ability of PTC cells.
In order to further explore the molecular mechanism of MRPL9-mediated PTC cell proliferation and migration, starbase database (http://starbase.sysu.edu.cn/index.php, accessed on 22 March 2022) predicted that the expression levels of MRPL9 and GGCT were positively correlated (Figure 3A). In addition, the results of immunofluorescence detection showed that the localization of MRPL9 and GGCT in cells partially overlapped (Figure 3B), and the immunofluorescence homologous double-labeling experiment of PTC tissue further demonstrated the partial overlap of the intracellular localization of MRPL9 and GGCT (Figure 3C). These results suggested the possibility of an interaction between MRPL9 and GGCT. Western blot results indicated that overexpression of GGCT could promote the protein expression of MRPL9, while knockdown of GGCT inhibited the expression of MRPL9 (Figure 3D–F). To further validate the interaction between GGCT and MRPL9, Co-IP experiments were performed on K1 and BCPAP cells. Compared with the IgG group, Anti-GGCT enriched a stronger signal of MRPL9 (Figure 3G), which further proved the interaction between GGCT and MRPL9.
To explore the expression level of GGCT in PTC, 26 pairs of cancer tissues and para-cancerous tissues of PTC patients were collected. The collected samples were stained with IHC, and the IHC results were scored and counted. The results showed that GGCT was highly expressed in PTC tissues (Figure 4A,B). In addition, statistical analysis of the IHC score of GGCT found that the expressions of GGCT and MRPL9 were positively correlated (Figure 4C), which further confirmed the interaction between GGCT and MRPL9. In order to further prove the high expression of GGCT in PTC, the protein expression level of GGCT in the patients’ tissues was analyzed by western blot experiment and the quantitative analysis of western blot was performed. The results showed that GGCT was highly expressed in PTC cancer tissue (Figure 4D,E). The above results indicated that GGCT was highly expressed in PTC tissues. In addition, we further explored the expression level of GGCT in PTC cell lines, qPCR and Western blot results showed that compared with Nthy-ori 3.1 cell lines, GGCT was highly expressed in mRNA and protein levels in TPC-1, K1, and BCPAP cell lines (Figure 4F–H).
Next, we explored the molecular mechanism by which GGCT and MRPL9 promotes the proliferation and migration of PTC cells. Thyroid cancers have been reported in the literature to be predominantly MAPK-driven cancers, and approximately 70% of thyroid cancers are caused by mutations that activate this pathway [27]. Further studies showed that knockdown of GGCT/MRPL9 could significantly inhibit the expression of p-P38/p-ERK, and the inhibition was more pronounced in the sh-GGCT+sh-MRPL9 group (Figure 5A–C). These results suggested that MRPL9 regulated the MAPK/ERK pathway through interaction with GGCT.
In order to explore the role of GGCT/MRPL9 in vivo, nude mice were subcutaneously xenografted or injected with GGCT/MRPL9 in the tail vein to explore the effect of GGCT/MRPL9 on the proliferation and migration of PTC cells in vivo. Luciferase-labeled K1 cells were infected with lentivirus containing shGGCT/shMRPL9/shGGCT + shMRPL9, and then equal amounts of cells were injected subcutaneously into nude mice. The bioluminescence intensity of the subcutaneous tumors was detected on the Xenogen IVIS Lumina System 28 days after the subcutaneous injection and the xenografts were harvested. The results showed that the fluorescence intensity of the subcutaneous xenografts in the shGGCT/shMRPL9 group was weakened (Figure 6A,B), and the mass and volume of the subcutaneous tumors decreased (Figure 6C–E), while the fluorescence intensity, mass and volume of the xenografts in the shGGCT + shMRPL9 group were the smallest (Figure 6A–E). For the nude mouse lung metastasis model, luciferase-expressing sh-NC-K1/shGGCT-K1/shMRPL9-K1/shGGCT + shMRPL9-K1 cells were injected into the tail vein, and the bioluminescence signal in the lungs of nude mice was recorded every 7 days. With the prolongation of observation time, the fluorescent signal in the lungs of nude mice in the shGGCT/shMRPL9 group weakened compared with the control group, and the fluorescent signal in the shGGCT + shMRPL9 group was the weakest (Figure 6F,G). The nude mice were sacrificed on the 28th day after injection. From the bioluminescence signals in the living lungs of nude mice on day 28, we preliminarily concluded that knockdown of GGCT/MRPL9 attenuated the ability of K1 cells in the lung metastases of nude mice, while knockdown of GGCT and MRPL9 had the best inhibitory effect (Figure 6H). The results of further detection of the fluorescence signal of the nude mouse lung tissue in vitro were consistent with the change trend of the fluorescence intensity of the nude mice in vivo (Figure 6I). Further H&E staining of the lung tissue of nude mice showed that the number of tumor nodules in the shGGCT + shMRPL9 group was reduced compared with the control group, and the number of tumor nodules in the shGGCT + shMRPL9 group was the least (Figure 6J and Supplementary Figure S1). The above results indicated that knockdown of GGCT/MRPL9 inhibited the proliferation and migration of tumor cells in vivo, and simultaneous knockdown of GGCT and MRPL9 had the strongest inhibitory effect on the proliferation and migration of PTC cells.
In this study, GGCT and MRPL9 were shown to be involved in regulating the biological behavior of PTC tumor cells and promoting the further development of cancer, suggesting that GGCT and MRPL9 may serve as potential biomarkers. Further research found that GGCT/MRPL9 promotes the MAPK/ERK signaling pathway to promote the proliferation and migration of PTC cells. The function of GGCT/MRPL9 in PTC cells was also confirmed in in vivo experiments. Thyroid cancer originates from follicular epithelial cells or parafollicular C cells. Follicular cell-derived thyroid cancers are classified into 4 histological types: papillary thyroid cancer (PTC 80–85%), follicular thyroid cancer (FTC 10–15%), poorly differentiated thyroid cancer (PDTC, <2%) and anaplastic thyroid cancer (ATC, <2%) [28]. PTC is primarily associated with mutations that activate the MAPK (mitogen-activated protein kinase) signaling pathway, such as RET, NTRK (neurotrophic receptor tyrosine kinase) and ALK (anaplastic lymphoma kinase) gene rearrangements or RAS (rat sarcoma) and BRAF (rapidly accelerating fibrosarcoma type B) activating point mutations, almost always found in an exclusive manner, suggesting that oncogenic activation of one member of this pathway was sufficient to drive transformation [29,30,31]. Furthermore, Xing M demonstrated that activation of the MAPK signaling pathway resulted in the up-regulation of tumor-promoting genes (such as VEGFA, MET, HIF1A, UPA, UPAR, TGFB1 and TSP1) and the down-regulation of tumor suppressor and thyroid genes (such as TIMP3, SLC5A8, DAPK1, NIS, TSHR and TPO) [32]. Our study found that functional inhibition of MRPL9/GGCT would inhibit the activity of the MAPK signaling pathway, and, simultaneously, inhibiting the protein expression of MRPL9 and GGCT would aggravate the inhibition of the MAPK signaling pathway. This indicated that MRPL9/GGCT promoted the occurrence and deterioration of PTC by activating the activity of MAPK signaling pathway. Dysregulation of mitochondrial bioenergetics has been recognized as a key metabolic hallmark of malignancy [33]. In recent years, some MRPs have been found to be involved in multiple cellular processes, such as cell proliferation, ribosome cycle regulation in vitro, and apoptosis, and the expression level of MRPL20 was significantly down-regulated in androgen-independent prostate cancer [34,35,36]. The levels of MRPL37 and mRNA were also significantly elevated in different lymphoma tissues. MRPL33 is required for mitochondrial function and has been implicated in tumor progression [34,36,37]. Recently, PTCD3 (also known as MRPS39) was shown to be critical for the maintenance of Myc-driven lymphomas [38]. MRPL44 expression was identified as a predictor of lymph node metastasis in papillary thyroid carcinoma [39], and MRPL13 inhibition was shown to be a key upstream regulator of OXPHOS dysfunction and hepatoma cell invasiveness [40]. These results suggest that dysregulation of MRPs may be responsible for bioenergetic dysregulation and play a key role in tumor progression. MRPL9 has been reported to be highly expressed in HCC and associated with poor prognosis [12,13], and knockdown of MRPL9 was determined to significantly inhibit cell proliferation and migration in HCC [13]. Our study found that MRPL9 was highly expressed in PTC, and MRPL9 promoted the proliferation and migration ability of PTC cells, which was further proved in the nude mouse subcutaneous xenograft model and lung metastasis model. Further studies revealed the interaction between MRPL9 and GGCT, which activated the MAPK/ERK signaling pathway and further promoted the malignancy of PTC. Gamma-glutamyl cyclotransferase (GGCT, 188 amino acids, 21 kDa) is an enzyme involved in the metabolism of glutathione, which catalyzes the conversion of γ-Glu-AA to pyroglutamate [15]. GGCT has been reported to be upregulated in several cancers, and depletion of GGCT can exert anticancer effects in these cancers, including prostate, esophageal squamous cell carcinoma, breast, gastric, and ovarian cancers [23,41,42,43,44]. GGCT was shown to promote colorectal cancer migration and invasion through epithelial-mesenchymal transition [22]. GGCT depletion in gastric cancer significantly inhibited cell proliferation and colony-forming capacity in MGC80-3 and AGS cells, induced apoptosis in early and late gastric cancer cells and induced gastric cancer cell cycle arrest in G2/M phase [23]. A study identified inhibition of cancer cell proliferation by disrupting GGCT, highlighting the potential to treat types of malignancies by inhibiting GGCT [45]. Our study found that GGCT was highly expressed in PTC, and the interaction between GGCT and MRPL9 jointly affected the MAPK/ERK signaling pathway. Our previous study showed that GGCT knockdown reduced the proliferation, migration and invasion abilities of PTC cells in vitro and blocked the EMT process. Mechanistically, it was found that GGCT can bind to CD44 and that GGCT plays a role in stabilizing CD44 to prevent its degradation [26]. Whether the interaction between GGCT and MRPL9 in this study is based on the modification of MRPL9 by GGCT remains to be further investigated. In conclusion, our study shows that MRPL9/GGCT acts as an oncogene in PTC to promote the malignant progression of PTC by activating the MAPK/ERK signaling pathway. The findings provide new insights into the diagnosis and treatment of PTC.
The cancer tissues and adjacent normal tissues (n = 26) of PTC patients involved in this study were provided by Tongji Hospital of Huazhong University of Science and Technology. All participants gave informed consent and did not receive radiotherapy or chemotherapy before surgery. The research protocol was approved by the Ethics Committee of Tongji Hospital, Huazhong University of Science and Technology.
Human normal thyroid cell line Nthy-ori 3-1; PTC cell line TPC-1, BCPAP and K1; human embryonic kidney 293T (HEK293T) cell line was purchased from Shanghai Cell Bank, Chinese Academy of Sciences. Nthy-ori 3-1, TPC-1, BCPAP cell lines were reared with RPMI-1640 complete medium (Meilunbio, Dalian, China), while 293T and K1 were reared in DMEM-H complete medium (Meilunbio, Dalian, China). All media contained 10% fetal bovine serum (FBS) (ExCell Bio, Shanghai, China) and 1% penicillin-streptomycin (Beyotime, Shanghai, China), and the cell culture conditions were 37 °C, 5% carbon dioxide.
For the construction of the expression plasmid for GGCT/MRPL9, the GGCT/MRPL9DE CDS sequence was amplified from human cDNA and ligated into the pLVX-puro (Addgene, Watertown, MA, USA) vector. The procedure followed the instructions for the cloning kit and EcoRI restriction endonuclease sites were used (10911ES20, Yeasen, Shanghai, China). The plasmid of shGGCT/shMRPL9 was synthesized in Tsingke Biotechnology Co., Ltd. (Tsingke Biotechnology, Beijing, China). The PCR amplification primers and sh-RNA sequences mentioned above are provided in Supplementary Table S1. For lentivirus packaging, PEI (40820ES04, Yeasen, Shanghai, China). was used to transfect the target plasmid, pCMV-VSV-G (Beyotime, Shanghai, China) and pCAG-dR8.9 (Beyotime, Shanghai, China) in 293T cells (according to the ratio of target plasmid: pCMV-VSV-G: pCAG-dR8.9 = 4:3:1). The medium was replaced with fresh medium 6 h after transfection, and the virus stock solution was collected 48 h after transfection. For stably transfected cell lines, PTC cells were infected with LV-GGCT, LV-MRPL9, LV-sh-GGCT, LV-sh-MRPL9, LV-pLVX, LV-pLKO.1, then cells were treated with 1 ug/mL puromycin (Meilunbio, Dalian, China).
Total RNA was extracted with Trizol (R0016, Beyotime, Shanghai, China), 1 µg RNA was reverse transcribed into cDNA by reverse transcription kit (R212, Vzayme, Nanjing, China), and qPCR reaction was performed with SYBR Green (11201ES03, Yeasen, Shanghai, China). The qRT-PCR program was set as follows: 95 °C for 5 min, followed by 40 cycles of 95 °C for 10 s, 60 °C for 20 s, and 72 °C for 20 s (CFX96, Bio-Rad, Hercules, CA, USA). Actin was used as normalization and the relative levels of gene expression were determined by the 2−ΔΔCT method. Primer information is provided in Supplementary Table S1.
Total protein from cells or patient tissue was extracted using RIPA lysis buffer (Beyotime, Shanghai, China), and an equal amount of protein was electrophoresed on a polyacrylamide gel and then transferred to a polyvinylidene fluoride (PVDF) membrane (Millipore, Burlington, MA, USA). The obtained PVDF membrane was incubated with 5% BSA at room temperature for 1 h, and then the PVDF membrane was incubated with the primary antibody and the secondary antibody successively, and developed on the imaging system (Bio-Rad, Hercules, CA, USA) after adding the highly sensitive ECL reagent. The experimental results were quantitatively analyzed with Image J. The antibody information used in the study is as follows: Anti-GGCT (1:1000, Proteintech, 16257-1-AP, Wuhan, China); Anti-MRPL9 (1:1000, Proteintech, 15342-1-AP, Wuhan, China), Anti-Actin (1:5000, ABclonal, AC026, Wuhan, China); anti-P-p38 (1:1000, CST, 29216, Danvers, MA, USA), anti-p38 (1:1000, ABclonal, A5049, Wuhan, China), anti-ERK1/2 (1:1000, CST, 4695, Danvers, MA, USA), anti-p-ERK1/2 (1:1000, CST, 4370, Danvers, MA, USA).
The transwell chambers (Corning, 3422, Corning, NY, USA) were coated with Matrigel and placed in 24-well plates to incubate overnight with culture medium. The treated cells were resuspended in FBS-free medium, and 2 × 104 cells/well were seeded into the upper chamber. A medium containing 10% FBS was then added to the lower chamber and incubated for 24 h. After removal of non-migrating cells, the chambers were fixed in 4% paraformaldehyde (MA0192, Meilunbio, Dalian, China) for 30 min and stained with 1% crystal violet (MA0148, Meilunbio, Dalian, China) for 30 min, and images were acquired using a microscope.
A total of 1 × 105 treated cells were seeded into each well of a 6-well plate and cultured to 90% confluence. Then, a 200 µL pipette tip was used to create a wound in the middle of each well. After washing with PBS, cells were incubated with medium containing 1% FBS. Wound images were acquired at 0 and 36 h. Each experiment was performed in triplicate.
For cell colony formation experiments, treated cells were seeded into 6-well plates (5 × 103). After 12 days of incubation at 37 °C in RPMI-1640/DMEM-H medium containing 10% FBS, the plates were washed with PBS, and stained with 0.1% crystal violet for 30 min at room temperature. Each experiment was repeated ≥3 times.
For immunofluorescence experiments, briefly, K1/BCPAP cells were seeded into 24-well plates (1 × 104/well) containing cell slides, where K1/BCPAP cells stably overexpressed flag-tagged GGCT. After cells adhered, cells were fixed with 4% paraformaldehyde (MA0192, Meilunbio, Dalian, China), permeabilized with 0.1% Triton X-100 (P0096, Beyotime, Shanghai, China), blocked with 5% BSA, and incubated with primary antibody and fluorescent secondary antibody, respectively. Finally, the nuclei were stained with DAPI and observed under a confocal microscope. Information on the primary and secondary antibodies used is as follows: MRPL9 (1:1000, Proteintech, 15342-1-AP, China), anti-flag-tag (1:50, ABclonal, AE005, China), FITC-labeled goat anti-mouse IgG (1:100, Servicebio, GB22301, Wuhan, China), Cy3-labeled goat anti-rabbit IgG (1:100, Servicebio, GB21303, China).
For Co-IP assay, whole cell lysates were obtained by Western and IP cell lysate (Beyotime, Shanghai, China) lysis buffer and incubated overnight with anti-GGCT (1:100, Proteintech, 16257-1-AP, Wuhan, China) or anti-IgG (1:100, ABclonal, AC005, Wuhan, China) at 4 °C. It was then incubated with rProtein A/G Plus MagPoly Beads (RM09008, ABclonal, Wuhan, China) for 2 h at 4 °C. Finally, the protein bound to the beads was eluted and detected by western blotting using antibodies against GGCT (1:1000, Proteintech, 16257-1-AP, Wuhan, China) or MRPL9 (1:1000, Proteintech, 15342-1-AP, Wuhan, China).
For the mouse xenograft model, 16 BALB/c nude mice (four weeks old, female) were randomly divided into four groups: K1-shNC, K1-shGGCT, K1-shMRPL9, shGGCT + shMRPL9. Luciferase-labeled K1 cells transduced with K1-shNC, K1-shGGCT, K1-shMRPL9, shGGCT + shMRPL9 viruses were harvested, resuspended in PBS, and then inoculated subcutaneously into nude mice (1 × 107). Twenty-eight days later, the subcutaneous tumors were harvested for subsequent analysis after imaging with the Xenogen IVIS Lumina System (Caliper Life Sciences, Waltham, MA, USA). For the lung metastasis mouse model, the same grouping was used as the mouse xenograft model. Luciferase-labeled K1 cells transduced with K1-shNC, K1-shGGCT, K1-shMRPL9, shGGCT + shMRPL9 were harvested and resuspended in PBS, and injected into nude mice via tail vein (1 × 106). Bioluminescence imaging (BLI) was used to monitor the growth of lung K1-fLuc metastases every 7 days. After 28 days, the lung tissues of nude mice were harvested for subsequent analysis. All animal protocols were approved by the Laboratory Animal Ethics Committee of Wuhan University of Science and Technology. This research was conducted according to the principles of the Declaration of Helsinki.
Tissues were sent to Sevier Bio (Wuhan, China) for embedding and sectioning. For H&E staining, sections were deparaffinized and then stained according to the instructions of the hematoxylin-Eosin (H&E) staining kit (E607318, Sangon Biotech, Shanghai, China). For immunohistochemistry, after the paraffin sections were dewaxed and rehydrated, the tissue sections were placed in a repair box filled with citric acid antigen retrieval buffer (PH 6.0), heated in a microwave oven for antigen retrieval, and then blocked endogenous peroxidase activity with 3% hydrogen peroxide solution. After incubation with specific primary and secondary antibodies, color is developed by DAB color development kit (G1212, Servicebio, Wuhan, China). The required antibody information is as follows: GGCT (1:100, Proteintech, 16257-1-AP, Wuhan, China), MRPL9 (1:100, Proteintech, 15342-1-AP, Wuhan, China). The immunohistochemical results were evaluated as follows: staining intensity was divided into 3 grades (0 = no staining, 1 = weak staining, 2 = moderate staining, 3 = strong staining), and the percentage of positive areas was divided into 4 grades (0% (0), <10% (1), 10–30% (2), 31–70% (3), 71–100% (4)). The final score for immunohistochemistry was determined by multiplying the staining intensity score by the percent positive area score, up to a maximum of 12 points. Proteins with high expression (≥6) and low expression (<6) were based on the final score. The protein expression of MRPL9 and GGCT (score) was assessed by two independent pathologists.
Briefly, paraffin sections were dewaxed and the sections were placed in a retrieval box containing EDTA antigen retrieval buffer (pH 8.0) (G1206, Servicebio, Wuhan, China) for antigen retrieval in a microwave oven. Then, the sections were placed in a 3% hydrogen peroxide solution to block endogenous peroxidase. After blocking with 3% BSA, add Anti-GGCT and incubate overnight at 4 degrees Celsius. After incubation with secondary antibody, add FITC-TSA (G1222, Servicebio, Wuhan, China) and incubate in the dark for 10 min. Next, the tissue sections were placed in a repair box filled with EDTA antigen retrieval buffer (PH8.0) and heated in a microwave oven to remove the primary and secondary antibodies that had been bound to the tissue, and then incubated with Anti-MRPL9 and fluorescent secondary antibodies (1:100, GB21303, Servicebio, Wuhan, China) respectively. After counterstaining the nuclei with DAPI, an autofluorescence quencher (G1221, Servicebio, Wuhan, China) was added, and finally the slides were mounted.
The TCGA database (https://www.cancer.gov/aboutnci/organization/ccg/research/structural-genomics/tcga, accessed on 14 March 2022) analyzed the expression level of MRPL9 in thyroid cancer, and the GEPIA 2 (http://gepia2.cancer-pku.cn/#index, accessed on 14 March 2022) website predicted the survival and prognosis of MRPL9 in thyroid cancer. In addition, the Starbase database website (http://starbase.sysu.edu.cn/index.php, accessed on 5 April 2022) predicted the expression correlation between MRPL9 and GGCT.
Statistical analysis was performed using SPSS 19.0 software (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 8. The measurement data (mean ± standard deviation) between the two groups were compared by paired t-test, and multiple groups were compared by one-way analysis of variance (ANOVA). p < 0.05 was considered statistically significant. | true | true | true |
PMC9570021 | 36233185 | Andrew S. Mason,Claire L. Varley,Olivia M. Foody,Xiang Li,Katie Skinner,Dawn Walker,Tony R. Larson,Daisuke Wakamatsu,Simon C. Baker,Jennifer Southgate | LPCAT4 Knockdown Alters Barrier Integrity and Cellular Bioenergetics in Human Urothelium | 06-10-2022 | urothelium,transcriptomics,lipidomics,transepithelial electrical resistance,glycerophospholipids,phosphatidylcholine,wound healing,AGPAT7,LPEAT2,LPLAT10 | Urothelium is a transitional, stratified epithelium that lines the lower urinary tract, providing a tight barrier to urine whilst retaining the capacity to stretch and rapidly resolve damage. The role of glycerophospholipids in urothelial barrier function is largely unknown, despite their importance in membrane structural integrity, protein complex assembly, and the master regulatory role of PPARγ in urothelial differentiation. We performed lipidomic and transcriptomic characterisation of urothelial differentiation, revealing a metabolic switch signature from fatty acid synthesis to lipid remodelling, including 5-fold upregulation of LPCAT4. LPCAT4 knockdown urothelial cultures exhibited an impaired proliferation rate but developed elevated trans-epithelial electrical resistances upon differentiation, associated with a reduced and delayed capacity to restitute barrier function after wounding. Specific reduction in 18:1 PC fatty acyl chains upon knockdown was consistent with LPCAT4 specificity, but was unlikely to elicit broad barrier function changes. However, transcriptomic analysis of LPCAT4 knockdown supported an LPC-induced reduction in DAG availability, predicted to limit PKC activity, and TSPO abundance, predicted to limit endogenous ATP. These phenotypes were confirmed by PKC and TSPO inhibition. Together, these data suggest an integral role for lipid mediators in urothelial barrier function and highlight the strength of combined lipidomic and transcriptomic analyses for characterising tissue homeostasis. | LPCAT4 Knockdown Alters Barrier Integrity and Cellular Bioenergetics in Human Urothelium
Urothelium is a transitional, stratified epithelium that lines the lower urinary tract, providing a tight barrier to urine whilst retaining the capacity to stretch and rapidly resolve damage. The role of glycerophospholipids in urothelial barrier function is largely unknown, despite their importance in membrane structural integrity, protein complex assembly, and the master regulatory role of PPARγ in urothelial differentiation. We performed lipidomic and transcriptomic characterisation of urothelial differentiation, revealing a metabolic switch signature from fatty acid synthesis to lipid remodelling, including 5-fold upregulation of LPCAT4. LPCAT4 knockdown urothelial cultures exhibited an impaired proliferation rate but developed elevated trans-epithelial electrical resistances upon differentiation, associated with a reduced and delayed capacity to restitute barrier function after wounding. Specific reduction in 18:1 PC fatty acyl chains upon knockdown was consistent with LPCAT4 specificity, but was unlikely to elicit broad barrier function changes. However, transcriptomic analysis of LPCAT4 knockdown supported an LPC-induced reduction in DAG availability, predicted to limit PKC activity, and TSPO abundance, predicted to limit endogenous ATP. These phenotypes were confirmed by PKC and TSPO inhibition. Together, these data suggest an integral role for lipid mediators in urothelial barrier function and highlight the strength of combined lipidomic and transcriptomic analyses for characterising tissue homeostasis.
Glycerophospholipids are the major constituent of biological membranes [1,2]. These amphipathic molecules consist of glycerol, two hydrophobic fatty acids, likely of variable length and saturation, and a hydrophilic phosphate ester. All glycerophospholipids are synthesised from glycerol-3-phosphate through the Kennedy Pathway [3,4,5], with the addition of different phosphate esters defining the major subgroups: phosphatidic acid (PA), phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylinositol (PI), phosphatidylglycerol (PG), and phosphatidylserine (PS). PC is typically the most abundant glycerophospholipid in mammalian cell membranes, with abundance highly dependent on species, cell type and cellular compartment [1,6]. Glycerophospholipids are remodelled and diversified in membranes through the Lands’ Cycle, where deacylation by phospholipase A2 family members is followed by reacylation, incorporating new fatty acyls by acyltransferases with substrate specificity [7]. Such diversity is highly important for membrane thickness and fluidity, intrinsic curvature and lateral pressure profile. In addition to their structural role within membranes, glycerophospholipid derivatives, such as lysophospholipids, eicosanoids and diacylglycerol (DAG), are critical components of cell signalling [8]. Despite their importance to cellular homeostasis and physical properties, the glycerophospholipid environment remains poorly understood in many tissues. The human urinary tract is lined by urothelium, a transitional epithelium that uses intercellular junctions anchored in cell membranes to form a highly specialised tight epithelial barrier to urine retained despite undergoing continuous physical changes from urine flow and storage. Urothelial differentiation is regulated by the nuclear receptor peroxisome proliferator-activated receptor gamma (PPARγ), well-recognised in lipid research as the master regulator of adipogenesis [9,10,11,12]. Despite this, little is known about human urothelial glycerophospholipid biology, even though both benign and malignant urothelial disease undergo disruption of barrier integrity and restitution [13,14,15], and urine has the potential to provide non-invasive diagnosis and monitoring [15,16,17], as long as the healthy state is fully understood. Here, we exploited our ability to derive biomimetic, functional urothelium from in vitro-propagated normal human urothelial (NHU) cells derived from surgical samples [18,19] to perform the first joint transcriptomic and lipidomic characterisation of urothelial differentiation. Whilst the overall glycerophospholipid profile did not change significantly upon differentiation, Lysophosphatidylcholine acyltransferase 4 (LPCAT4) was significantly upregulated. Stable LPCAT4 knockdown in NHU cells resulted in reduced proliferation and, after differentiation, a tighter epithelial barrier but impaired wound restitution ability. Specific transcriptomic and glycerophospholipid changes in the LPCAT4 knockdowns supported a role for excess LPC to inhibit protein kinase C activity and alter barrier mechanics, and also, independently, limit the availability of intracellular ATP. Together, these results suggest that lipid mediators play a significant functional role in human urothelial homeostasis.
Twin cultures of undifferentiated and in vitro-redifferentiated human urothelial cells from three donor backgrounds were profiled by LC-MS (Liquid Chromatography-Mass Spectrometry)/MS. Lipid species with a relative abundance greater than 1% in at least one sample were manually annotated from both positive (Figure 1A) and negative ion modes (Figure 1B). A total of 12 distinct PC, plasmanyl-PC (O-PC) and sphingomyelin (SM) lipids were found in both modes, 7 triglycerides (TG) were identified in the positive mode alone, and 2 PE and 5 plasmenyl-PE species were identified only in the negative mode. There were no significant spectra consistent with PG, PI or PS glycerophospholipids identified in either mode, however these subgroups were detectable from the lipid standards. Pooled TG species abundance increased by 279% upon differentiation (p = 0.048). Conversely, glycerophospholipid intensities were not, overall, differentiation-associated. Glycerophospholipid intensities in the negative mode were highly similar across samples (Figure 1B), and sample lipid profiles were strongly correlated within treatment groups (ρ = 0.945; p = 9.48 × 10−12). PC, O-PC and SM abundances also correlated well between modes (n = 6; average ρ = 0.913), reducing confidence in statistically significant PC abundance changes observed only in the positive mode analysis. Whilst glycerophospholipid abundance was not greatly affected by differentiation, there was evidence of remodelling, with a significant increase in utilised chain length (p = 0.014; Figure 1C) and slight enrichment of the unsaturated fatty acyl chain proportion (p = 0.070), both driven by increased abundance of polyunsaturated C20 fatty acyls upon differentiation. PC was the most abundantly detected lipid class, with a stable PC:PE:SM ratio of ~24:8:1 across samples and differentiation states. PC molecules predominantly contained monounsaturated C18 fatty acyls (49.7%; sd 2.3%; Figure 1D).
Urothelial differentiation elicits broad transcriptomic changes, reflecting the major shift from an actively proliferating cell monolayer to a transitional, stratified and quiescent epithelium with a tight electrophysiological barrier, as previously reported [19]. Strikingly, genes involved in lipid regulation and metabolism were highly enriched within the significantly changing genes (30.89% of expressed lipid regulatory genes; 23.26% of all expressed genes; p = 1.39 × 10−7) (Figure 2A). The undifferentiated state shows differential expression of genes involved in lipid biogenesis, including regulation of mitochondrial transport and fatty acid oxidation (Figure 2A,B). Differentiated urothelium switches from lipid biosynthesis to catabolic processes and lipid remodelling, including a striking example of the “sphingolipid rheostat” model [20], where both ceramide and the proliferation- and migration-inducing sphingosine-1-phosphate are remodelled in favour of the quiescence- and differentiation-associated sphingosine (Figure 2C). Differentiation also induced a sphingosine kinase switch from the pro-survival isoform 1, to the growth-inhibiting isoform 2; the latter at lower resting expression (Figure 2A,C). This is fully consistent with the well-described transcriptomic pattern of the sphingolipid rheostat associated with reduced sphingosine-1-phosphate abundance [21]. Given the dominance of PC in urothelial membranes (Figure 1A,B), the greater than 5-fold expression increase in the LPC-dominant acyl-CoA transferase Lysophospholipid acyltransferase 4 (LPCAT4; previously known as AGPAT7 [22] and LPEAT2 [23], and recently proposed for a change of nomenclature to LPLAT10 [24]) upon differentiation was striking (Figure 2D). Interrogation of the differentiation time course by RT-qPCR showed that LPCAT4 was elevated immediately upon differentiation and then sustained (Figure 2E). Upregulation of the LPI-specific Membrane Bound O-Acyltransferase Domain Containing 7 (MBOAT7) did not contribute to any significant PI species by LC-MS/MS, suggesting an alternative role to membrane composition, likely in secondary signalling. Consistent with the LC-MS/MS profiling (Figure 1B), there were few changes across the glycerophospholipid de novo synthesis “Kennedy” pathway upon differentiation (Figure 2D).
We further investigated the role of LPCAT4 in NHU cytodifferentiation, formation of an epithelial barrier and wound repair by establishing stable shRNA knockdowns in 6 independent NHU donor cell lines. Control and LPCAT4 knockdown shRNA cultures were differentiated before assessing LPCAT4 expression by RT-qPCR. LPCAT4 expression was donor-dependent (Figure 3A), but was knocked down to an average of 50.88% across six independent cell lines (SEM 4.71; range 33.67–66.15). LPCAT4 protein knockdown was evaluated by Western blot and densitometry analysis in three donors, with a congruent average knockdown of 57.16% (Figure 3B). Initial observations of the shRNA LPCAT4 and shRNA control cultures suggested that growth rate was affected in the LPCAT4 knockdown cells (Figure 3C,D). Quantitative analysis of undifferentiated shRNA-transduced NHU cell cultures using the Phasefocus LiveCyte™ showed significant inhibition of population size at each measured time point following seeding at identical densities (Figure 3E). Control shRNA cells exhibited a linear growth phase up to 60 h, when they became confluent. Comparison of the growth rates across this first 60 h period revealed a doubling time of 22.3 h in the control shRNA cells compared to 60.3 h in the LPCAT4 knockdown cells. Similarly, knockdown cells exhibited significantly reduced motility compared to time-matched and seeding-density-normalised control cultures (Figure 3F). In both conditions, culture motility reduced as cells approached confluence (but at a non-significantly different rate; 0–21 h; p = 0.62). Resting normalised culture optical flow velocity at confluence appeared slightly reduced in the control, but non-significantly. Furthermore, there was no evidence of elevated cell death or sloughing in the knockdown cultures, supporting truly reduced growth and motility rates following LPCAT4 knockdown. Once induced to differentiate, formation of a tight epithelial barrier was determined by measuring transepithelial electrical resistance (TEER). Both control and knockdown shRNA cultures formed tight barriers (≥500 Ω.cm2), but LPCAT4 knockdown generated a tight barrier more quickly, and sustained a significantly elevated barrier throughout the time course (maximal in both conditions at 192 h; Figure 3G). Absolute TEER values were donor-dependent, but were elevated in the LPCAT4 knockdown lines (Figure 3H). Control and LPCAT4 shRNA cultures showed no significant differences in early (CK13) or late (CLDN4) differentiation markers at the protein level (Figure 3I). We investigated whether knockdown influenced urothelial repair after wounding, using a scratch-wound model of stratified cultures on glass in 2 independent donor backgrounds. Rate of wound restitution was donor-dependent, but typically wounds began to repair from 3 h and were resolved by approximately 10 h post-wound (representative images in Figure S1). LPCAT4 knockdown significantly impaired scratch wound healing (p = 3.45 × 10−6 at 3 h and p = 5.00 × 10−7 at 6 h post-wound), with a highly donor-dependent closure rate (Figure 3J). As all wounds were eventually resolved, LPCAT4 knockdown elicited a delayed response, rather than compromising the tissue entirely.
We assessed the impact of LPCAT4 knockdown on the glycerophospholipid environment by performing LC-MS/MS on differentiated shRNA control and knockdown cell lines from three donor backgrounds, utilising the same donors reported in Figure 1. Analysis focused on the negative ion channel where PC, PE and SM were represented (Figure 4A). Across all detected lipids, the only significant difference was a decrease in PC 18:1_18:1 molecules in LPCAT4 knockdown cells (p = 0.013), which was also observed in the positive channel (p = 0.04; Supplementary Table S1). Overall PC proportion did not change, with averages of 58.1% in controls, and 58.0% in the knockdown. LPCAT4 has a known preference for 18:1 CoA [25], and concordantly the impact on 18:1 CoA acylation of LPC was only reliably observed in overall reduction of 18:1_18:1 PC. LPCAT4 preference for PC over PE was also observed, with no significant difference in PE 18:1_18:1 (p = 0.348) or P-PE 18:1_18:1 (p = 0.253). Slight, non-significant changes were observed in overall fatty acyl chain length usage across glycerophospholipids upon knockdown (Figure 4B), with the only significant difference a reduction in 18:1 usage in PC molecules (Figure 4C). Intriguingly, similar PC composition differences were not observed upon differentiation (Figure 1B), when LPCAT4 gene expression (Figure 2C and Figure 3A) and protein abundance (Figure 3B) were significantly lower in the undifferentiated state.
LPCAT4 is not a transcription factor, and as an ER-bound protein it has limited potential for direct influence upon the cellular transcriptome. However, we investigated the overall impact of the altered lipid environment induced by LPCAT4 knockdown on the transcriptome by performing mRNAseq at an early and late differentiation time point in four donor backgrounds (Table S2). Generally, changes to the global transcriptome were limited and typically exhibited low fold change, but specific genes relevant to the knockdown phenotypes described above were identified. Early in the differentiation protocol (Figure 5A) the changes between control and knockdown states were less numerous than at the later time point (Figure 5B). Whilst some of the identified genes were not modulated at both time points (Figure 5C), the same key processes were represented: cell-cell and cell-matrix interactions, matrix metallopeptidase activity, and regulation of lipid transport related to maintenance of the acyl-CoA pool. Markers of mature urothelial differentiation were not affected (Figure 5D). Tight junction component genes CLDN1 and CLDN8 (a specific recruiter for CLDN4 [26]), as well as the adherens junction component CDH6, were upregulated in the knockdown cultures, while desmoglein DSG1 was initially reduced. LPCAT4 knockdown cells appeared less primed for matrix degradation with increased expression of the matrix metallopeptidase inhibitor TIMP3 as well as the SERPIN family genes SERPINA1, SERPINB7 and SERPINB2 (latter switches fold change between 48 and 144 h), which actively inhibit wound healing driven by plasmin and TGF-β signalling in urothelium [27]. Upregulation of the metal ion regulating MT2A and MT1E also supported the modified activation state of metallopeptidases (e.g., CPA4, CPZ and ADAM28) upon knockdown. Furthermore, cell-matrix cross-linking genes were typically upregulated in the knockdown, including PLOD2, RHOB, ICAM5, LCN2, COL16A1, TFF3 and TGM2 (transglutaminase 2; TG2). Previous work found an increased expression of matrix-degrading inhibitors by NHU cells grown on Matrigel matrix, which was also associated with reduced migratory phenotype [28]. Interestingly, the changing dynamics of the acyl-CoA pool also impacted its broader regulation, with decreased TSPO transcript at both time points suggesting reduced efflux into the mitochondria [29] and differential stabilisation of the acetyl-CoA to acyl-CoA ratio to increase acyl-CoA (reduced acetyl > acyl activity via MTHFD2 reduction, and increased acetyl > acyl via AKR1B10). Despite these broad CoA modulators, there were no specific changes in Lands’ Cycle or Kennedy Pathway genes upon LPCAT4 knockdown (Figure 5E).
Based on the LPCAT4 knockdown transcriptomics, we identified two candidate, potentially impacted signal transduction pathways which might direct the observed changes in urothelial biology. We went on to investigate whether specific modulation of either of these pathways would phenocopy the effects of LPCAT4 knockdown in NHU cells. TGM2 transcript was significantly increased in LPCAT4 knockdown cells (Figure 5C). As TG2 protein directly inhibits Phospholipase C delta 1 (PLCδ1) hydrolysis of the PI derivative PIP2 to IP3 and DAG [30], we hypothesised that limited availability of DAG would reduce protein kinase C (PKC) activity. PKC-δ (PRKCD) was the most abundant family member in NHU cells, is differentiation-associated (Figure 2A and Figure 6A), and is DAG-dependent whilst independent of calcium [31]. We examined the effect of reduced PKC activity on urothelial physiology using Go6983, a well-characterised pan-PKC inhibitor [32]. Consistent with LPCAT4 knockdown, PKC inhibition resulted in both a tighter urothelial barrier resistance (Figure 6B) and a significant delay to wound healing (Figure 6D). TSPO was the most significantly downregulated transcript after LPCAT4 knockdown (Figure 5C), and we confirmed TSPO protein was also reduced by Western blotting (Figure 6C). Furthermore, consistent with the LPCAT4 knockdown, TSPO inhibition with PK11195 [34] led to a significant elevation of urothelial barrier resistance (Figure 6B). Impact on wound healing was non-significant (Figure 6D), but there was high variability between the cultures. TSPO has a broad role in cellular bioenergetics, including directly influencing ATP availability [34,35]. An ATP assay demonstrated significantly reduced ATP in cells treated with increased concentrations of PK11195 (Figure 6E), and further established that differentiated LPCAT4 knockdown cells also contained significantly lower concentrations of endogenous ATP (Figure 6E).
The master regulator of the urothelial differentiation programme is PPARγ [10], yet despite its parallel role in fatty acid regulation in adipogenesis [11], the regulation and modification of lipids in normal human urothelial differentiation is not fully understood. Here, LC-MS/MS profiling revealed large increases in TG abundance upon differentiation. TG biosynthesis genes were not upregulated transcriptomically, supporting a change of metabolism between proliferative and quiescent states. The corollary of this phenomenon is well known in epithelial cancers, including urothelial carcinoma [15,36]. Genes involved in lipid biosynthesis, regulation and signalling were enriched in genes differentially expressed between urothelial differentiation states, with an overall switch from biosynthesis to remodelling upon differentiation. Consistently, LC-MS/MS revealed no significant differentiation-associated changes in glycerophospholipid or sphingolipid abundance, however the transcriptomics supported the “sphingolipid rheostat” model [20] switch towards sphingosine from both ceramide and sphingosine-1-phosphate, as well as PC and PI remodelling through the Lands’ Cycle acyl transferases LPCAT4 and MBOAT7. Whilst PI abundance across diverse mammalian membranes is lower than both PC and PE [2], its absence from the LC-MS/MS data supports a key role for phosphoinositide signalling in urothelial differentiation. PC is the most abundant glycerophospholipid in urothelial membranes, so the 5-fold increase in LPCAT4 expression upon differentiation warranted further study. Whilst differentiation-associated, our c.50% knockdown of LPCAT4 transcript and protein in six independent biological backgrounds did not impact early and late markers of urothelial differentiation at either transcript or protein levels. However, LPCAT4 knockdown did result in a significantly tighter epithelial barrier. Whilst barrier function is typically a marker of healthy urothelium [19], LPCAT4 knockdown also resulted in delayed wound restitution, potentially driven by cell–cell and cell-matrix changes, particularly the enhanced expression of metalloproteinases known to inhibit classical plasmin-driven epithelial wound repair [37,38]. LC-MS/MS of the differentiated knockdown lines identified a specific reduction in 18:1 fatty acyls in PC; a previously reported LPCAT4 preference [25] not detected in a study of LPCAT4 activity in chondrocyte differentiation [39]. Pertinently, 18:1 abundance did not increase upon differentiation, even though LPCAT4 abundance is much lower in the undifferentiated state, suggesting expanded glycerophospholipid diversity is more important in the fully differentiated quiescent tissue. Whilst membrane PC changes in LPCAT4 knockdown lines were specific, the magnitude of change was considered unlikely to have caused the broader impacts on urothelial tissue biology, specifically the increased barrier tightness and inhibited wound repair. We therefore examined a role for altered lipid signalling dynamics when LPC reacylation is reduced, although the specific mechanism remains unknown. Transcriptomic analysis of the LPCAT4 knockdown suggested PLCδ1-driven hydrolysis of PIP2 may be impaired, reducing DAG availability for PKC activation. Previous epithelial studies have highlighted the role of PKC in regulating the activity of metallopeptidases and other cell-matrix regulators [40,41], as well as regulating tight junction assembly and dissolution [32]. Consistent with the LPCAT4 knockdown phenotype, and these previous studies, direct inhibition of PKC activity in NHU cells elevated barrier tightness and reduced the rate of wound repair. The importance of DAG-reliant signalling in urothelial barrier integrity is further supported in the present study by identifying the upregulation of LPI-reacylator MBOAT7 upon differentiation, the absence of high-abundance/low-diversity PI species in urothelial membranes, and a Kennedy Pathway phosphatidate phosphatase (PA→DAG) switch from LPIN1 to LPIN2/3 upon differentiation (known activators of PPARG expression in adipogenesis [42]). Our data support a role for elevated LPC concentrations in disrupting this key urothelial signalling pathway. LPCAT4 knockdown also resulted in downregulation of TSPO at the transcript and protein level. Again, the direct regulatory mechanism is unknown, although, pertinently, PKC-ε has been implicated in steroidogenic cells [43]. Whilst the impact of pharmacological TSPO inhibition on barrier tightness and wound healing was less striking (yet still consistent with LPCAT4 knockdown), it highlighted reductions in endogenous ATP availability observed strongly in the LPCAT4 knockdown environment, consistent with reduced proliferation rate and delays to urothelial wound healing, as described previously [44]. Taken together, these data support a key role for elevated LPC in modifying lipid signalling pathway outcomes in human urothelium. In summary, we have performed the first paired lipidomic and transcriptomic characterisation of normal human urothelium, highlighting the importance of lipid mediators in the regulation of both urothelial differentiation and tissue homeostasis. The purpose of glycerophospholipid diversity in urothelium is yet to be fully elucidated. Knockdown of the differentiation-associated LPCAT4 resulted in a tighter epithelial barrier but limited wound healing, consistent with LPC-driven changes to lipid signalling affecting both PKC activity and endogenous ATP availability. Barrier integrity and wound restitution are commonly compromised in both benign and malignant bladder disease, emphasising the need for a more complete characterisation of barrier regulation beyond protein localisation.
NHS Research Ethics Committee approval was granted for the collection of human ureters from discarded tissue following renal transplant surgery from patients with no history of urothelial malignancy. The study was approved by The University of York Department of Biology Research Ethics Committee.
Finite NHU cell lines were maintained in keratinocyte serum-free medium (KSFM; Invitrogen Europe Ltd., Inchinnan, UK) supplemented with recombinant epidermal growth factor, bovine pituitary extract and 30 ng/mL cholera toxin (Merck UK, Poole, UK), to derive KSFM “complete” (KSFMc). Under these conditions, NHU cells grow as an undifferentiated monolayer and maintain a proliferative phenotype. As previously described [19], NHU cells were induced to differentiate and form a biomimetic tissue with a physiological “tight” epithelial barrier [45] by culturing the cells in KSFMc supplemented with 5% adult bovine serum (ABS) for 3–4 days, then seeding 3–6 replicate 1.13 cm2 ThinCert® membranes (Greiner Bio-One Ltd., Stonehouse, UK) at 5 × 105 cells/membrane, cultured for a further 7 days in KSFMc with 5% ABS and [Ca2+] elevated to 2 mM. TEER was measured with the EVOM™ Voltohmmeter (World Precision Instruments, Hitchin, UK), or assessed continuously after membrane seeding using the cellZscope2 (nanoAnalytics, distributed through Labtech, Heathfield, UK).
In anticipation of generating targeted gene knockdowns, a non-specific, control short hairpin RNA (shRNA) construct provided with the Clontech RNAi-Ready pSIREn-RetroQ vector kit (Takara Bio UK Ltd., London, UK) was used to provide an experimental model control to future knockdowns. This control shRNA construct, complete with restriction overhangs for directional cloning and a MluI restriction site, was ligated into the RNAi-Ready pSIREN-RetroQ vector. Following bacterial transformation using XL-1 Blue supercompetent cells (Agilent Technologies LDA UK Ltd., Stockport), ligation was verified by MluI restriction digest. PT67 retroviral packaging cells were transfected with the verified pSIREN-RetroQ vector, and retroviral particles were collected from the medium before being passed through low-binding 0.45 μm Tuffryn® filters (VWR International Ltd., Lutterworth, UK). As previously described [46], the filtered medium containing retrovirus was applied to undifferentiated, actively proliferating NHU cells, and transduced cells were selected using 1 μg/mL puromycin. To evaluate the impact of transduction on NHU differentiation state, triplicate cultures from two independent NHU cell lines were either transduced with the control shRNA, treated with only a puromycin-selection mock control, or not manipulated. Cells from each culture condition were differentiated as described above and RT-qPCR for early and late differentiation markers, as well as TEER values, were highly congruent, suggesting limited non-specific effects.
mRNA sequencing data from donor-matched, undifferentiated (PRJNA847878) and in vitro differentiated (PRJNA610264) NHU cells, previously generated by our group, were used to identify urothelial differentiation-associated genes involved in lipid biosynthesis and regulation. Reads were quality-checked using FastQC v0.11.7 [47] and trimmed to remove adapters and low quality read ends using trimmomatic v0.36 [48]. Reads were pseudoaligned to the Gencode v35 human transcriptome using kallisto v0.46.0 [49] and gene-level transcripts per million (TPM) expression values derived using tximport v1.14.0 [50]. Donor-aware differential expression analysis at each timepoint was conducted using the likelihood ratio test in Sleuth v0.30.0 [51]. Gene set enrichment analysis was performed using the prerank module of GSEApy v0.10.2 [52], a Python3 wrapper for the Broad Institute’s GSEA tool [53], and the MSigDB v7.2 hallmark and gene ontology collections, run with 1000 permutations.
LPCAT4 protein functional and transmembrane domain locations were identified using InterPro v85.0 [54] and used to educate knockdown strategy. shRNA oligonucleotides were designed against the shared exonic sequences of the LPCAT4 (ENSG00000176454) protein-coding transcripts in Gencode v28, using DSIR (http://biodev.cea.fr/DSIR/DSIR.html, accessed 8 January 2018) [55] and InvivoGen siRNA Wizard™ v3.1 (https://www.invivogen.com/sirnawizard/index.php, accessed 8 January 2018), followed by addition of a hairpin loop, restriction overhangs for directional cloning, and a MluI restriction site. Three unique sequences were designed: LPCAT4_shRNA1: gatccGCACCTGTTCCAACAAGAATTCAAGAGATTCTTGTTGGAACAGGTGCTTTTTTACGCGT (within ENSE00001253149); LPCAT4_shRNA2: gatccGTAGGGAGCTTACCTGTGATTTTCAAGAGAAATCACAGGTAAGCTCCCTACTTTTTTACGCGTg (within ENSE00003524206); LPCAT4_shRNA3: gatccGAATGATCAGCCAGGAAGAGTTTCAAGAGAACTCTTCCTGGCTGATCATTCTTTTTTACGCGTg (within ENSE00003630257). Each construct was transfected into retroviral packaging cells as described above. Two independent sets of donor-matched NHU cell lines were generated for the three LPCAT4 shRNA constructs, and the control shRNA construct, and differentiated as described above. The impact of transduction on LPCAT4 expression was assessed, as well as the knockdown efficiency of each LPCAT4 shRNA construct by RT-qPCR and western blotting, and the culture doubling time was measured. These initial studies showed LPCAT4 expression was unaffected by control transduction, and that the LPCAT4_shRNA2 construct gave the most effective LPCAT4 transcript knockdown. 6 pairs of donor-matched NHU cell lines transduced with either control shRNA or LPCAT4_shRNA2 constructs were then generated and used for all further analyses.
Whilst actively proliferating, transduced NHU cells were regularly imaged to assess culture doubling time. Doubling rate was formally quantified using the Phasefocus LiveCyte™ instrument (Phase Focus Ltd., Sheffield, UK) housed in the University of York (York, UK) Bioscience Technology Facility. Transduced cells from two independent backgrounds were seeded at 104 cells/cm2 in triplicate onto 24-well Corning™ Primaria™ plasticware (suppled by Thermo Fisher Scientific UK, Loughborough, UK), allowed to adhere and acclimatise overnight in KSFMc, with cells counted every 12 h for 96 h.
Videos of cell cultures obtained from the Phasefocus LiveCyte™ instrument were further analysed to assess cell migration rate, measured by normalised optical flow. Cells were counted from image frame sequences at 30 min intervals using ImageJ v1.53c (https://www.afterdawn.com/software/desktop/image_editing/imagej.cfm, last accessed 3 October 2022). First, non-uniform background was removed using the “Sliding Parabvoloid” function with a rolling ball radius of 50 pixels, then images were converted to 8-bit with a threshold applied in the range 0–210. Small visual artefacts in cell cytoplasm were removed using the “Fill Holes” function and watershedding applied to separate partially overlapped cells. Finally, “Analyse Particles” was applied to count all particles with radius greater than 50 pixels. Following this preprocessing, images were loaded into Matlab 2021b (Computer Vision Toolbox, Mathworks, Natick, MA, USA) and converted to greyscale. The Matlab Image Processing toolbox function “estimateFlow” was then used to apply the opticalFlowHS “Horn-Schunck” method of estimating flow velocity characteristics on a per pixel basis between subsequent frames in each series. Velocity matrices were processed to obtain pixel velocity magnitude and orientation, with frame-wide averages normalised by observed cell number.
RNA was extracted from cell cultures using Trizol™ (Invitrogen Europe Ltd.) and cDNA synthesis performed as previously described [9]. Quantification assays were performed on an ABI Prism Real-Time PCR System (Thermo Fisher Scientific UK) using SYBR-green™ PCR master mix and PCR primers for LPCAT4, UPK2, KRT13 and GAPDH. LPCAT4 PCR primers (5′-AGCAGGATACCAAGGGTTTGG; 5′-GCCAGACGAGTTAGCTCTTCCA) were designed on the Applied Biosystems Primer Express™ Software, and existing primers were used for the other gene targets [9]. All values were quantified using the ΔΔCt method, and normalised to endogenous GAPDH expression.
Whole cell lysates (20 µg) were resolved on 4–12% Bis-Tris gels (Thermo Fisher Scientific UK) and transferred to PVDF membranes using standard immunoblotting techniques. Membranes were incubated with each primary antibody (Table 1) for 16 h at 4 °C. Secondary antibodies used were goat anti-mouse LI-COR IgG IRDyeTM 680 (Molecular Probes supplied by Thermo Fisher Scientific UK) or goat anti-rabbit IgG DyLight 800 (Generon, Slough, UK). Antibody binding was detected on the OdysseyTM Infra-red Imaging System (Li-Cor Biosciences, Cambridge, UK) and the densitometry normalized to β-actin (Table 1).
Transduced urothelial cells were seeded onto 12-well Multispot slides (Hendley-Essex, Loughton, UK) at 3.5 × 104 cells/well, allowed to attach, then maintained for 7 days in differentiated culture conditions as described above. After the cells were scratch-wounded with a pipette tip, replicate slides were fixed in methanol:acetone (v/v) before wounding, and then at 0 h, 3 h, 6 h, 10 h and 72 h post-wound. Hoechst 33258 (0.1 µg/mL; Merck UK) was used to visualise nuclei. Wound size at each timepoint was measured using ImageJ and normalised to the 0 h condition-average scratch size.
Twin cultures from three ureteric NHU cell lines (transduced donor backgrounds 1, 3 and 4) transduced with the scrambled shRNA sequence were expanded in proliferative conditions. From each twin culture, an undifferentiated, proliferative state was sampled by harvesting cells at ~70–80% visual confluence. The remaining culture was grown to visual confluence and then differentiated using medium supplemented with 5% ABS and 2 mM [Ca2+] for a further six days on plastic before harvest. Differentiated donor-matched LPCAT4 shRNA knockdown cultures were also generated in the same manner. Harvested cells were pelleted and submitted to the University of York Bioscience Technology Facility for analysis. Cell pellets were lyophilised using a Genevac EZ-2.3 Elite Centrifugal Evaporator (Biopharma Group, Winchester, UK) and prepared for analysis by LC-MS/MS by the Bioscience Technology Facility Metabolomics and Proteomics staff, as described previously [56]. Downstream data analysis was performed in R, using the Bioconductor package XCMS [57] and processed as previously described [56]. MS1 spectra were matched against the LipidMaps Structural Database [58], and matching MS2 spectra were searched against the Lipid Match [59] and Lipid Blast databases [60]. Features were quantified relative to spiked-in deuterated standards (Avanti Polar Lipids Inc. SPLASH® Lipidomix®, Stratech 330707-AVL). Lipid spectra were corrected using null and culture media controls, and species with abundance ≥ 1% were manually annotated by evaluating database results against likely MS1 adducts and retention times in positive and negative ionisation modes, and inspection of MS2 daughter ions and neutral losses. PC, PE, SM and TG species abundances were quantified using the class-specific SPLASH® Lipidomix® standards. Lipid profiles were compared by 2-way ANOVA, significance values corrected using Benjamini-Hochberg, and pairwise comparisons assessed by Tukey’s test.
A single ureteric NHU cell line was used to generate a differentiation time course of LPCAT4 expression by RT-qPCR. LPCAT4 expression was upregulated 24 h after media supplementation with ABS, and remained elevated throughout, validating previous data from microarray [61]. This suggested LPCAT4 knockdown could affect initial differentiation and progression, as well as the final establishment of the differentiated phenotype. Twin cultures of four donor-matched (transduced donor backgrounds 1–4) scrambled shRNA control and LPCAT4 shRNA knockdown cell lines were expanded in proliferative culture conditions as described above. To assess early effect on barrier formation, cells were seeded onto triplicate 12-well 1.13 cm2 ThinCert™ culture membranes in KSFMc, allowed to adhere overnight, then the media supplemented with 5% ABS and 2 mM [Ca2+]. Half of each twin culture was harvested after two days in supplemented media, and the other after six days. TEER was assessed by EVOM™ Voltohmmeter before harvest. Total RNA was extracted using Trizol™ and submitted to the Oxford Genomics Centre (Oxford, UK) for polyA library synthesis and mRNA sequencing using the Illumina NovaSeq6000 (Illumina, Cambridge, UK) instrument, generating 150 bp paired-end reads (deposited in PRJNA848077). Sequencing data was checked for quality, processed and analysed as described above (independently at both time points), though no read trimming was required.
NHU cells from 2 donors were cultured in triplicate and differentiated as described above. Cells from each culture were seeded onto 1.13 cm2 ThinCert® membranes (6 technical replicates per culture) as described, supplementing the differentiating medium with either 100 nM [32] pan-Protein Kinase C (PKC) inhibitor Go6983 (Tocris Bioscience, Avon, UK), or 100 nM [34] Translocator Protein (TSPO) inhibitor PK11195 (Tocris). Cultures were monitored for TEER, restitution of barrier after scratch-wounding, and rate of wound closure as mentioned above.
Actively proliferating NHU cells were seeded on plastic at 104 cells/well, allowed to attach, and treated for 48 h with either 0.1% DMSO (control), the TSPO inhibitor PK11195 at 100 nM or 500 nM, or with 3 mM ketamine, previously shown to exhibit a 23.2% reduction in urothelial cellular ATP [33]. Cellular ATP was quantified using the CellTiter-Glo® 2.0 assay (Promega UK Ltd., Southampton, UK) with 5 technical replicates per condition, and normalised using a bicinchoninic acid protein assay kit (Thermo Fisher Scientific UK). Cells were lysed in CellTiter-Glo® 2.0 reagent and assayed for luciferase activity in a Clariostar luminescence plate reader (BMG Labtech Ltd., Aylesbury, UK). Donor-matched control shRNA and LPCAT4 shRNA knockdown differentiated cultures were also assayed for cellular ATP concentration. | true | true | true |
PMC9570089 | 36232738 | Léa Toury,Diane Frankel,Coraline Airault,Frédérique Magdinier,Patrice Roll,Elise Kaspi | miR-140-5p and miR-140-3p: Key Actors in Aging-Related Diseases? | 28-09-2022 | microRNA,miR-140,aging-related diseases | microRNAs (miRNAs) are small single strand non-coding RNAs and powerful gene expression regulators. They mainly bind to the 3′UTR sequence of targeted mRNA, leading to their degradation or translation inhibition. miR-140 gene encodes the pre-miR-140 that generates the two mature miRNAs miR-140-5p and miR-140-3p. miR-140-5p/-3p have been associated with the development and progression of cancers, but also non-neoplastic diseases. In aging-related diseases, miR-140-5p and miR-140-3p expressions are modulated. The seric levels of these two miRNAs are used as circulating biomarkers and may represent predictive tools. They are also considered key actors in the pathophysiology of aging-related diseases. miR-140-5p/-3p repress targets regulating cell proliferation, apoptosis, senescence, and inflammation. This work focuses on the roles of miR-140-3p and miR-140-5p in aging-related diseases, details their regulation (i.e., by long non-coding RNA), and reviews the molecular targets of theses miRNAs involved in aging pathophysiology. | miR-140-5p and miR-140-3p: Key Actors in Aging-Related Diseases?
microRNAs (miRNAs) are small single strand non-coding RNAs and powerful gene expression regulators. They mainly bind to the 3′UTR sequence of targeted mRNA, leading to their degradation or translation inhibition. miR-140 gene encodes the pre-miR-140 that generates the two mature miRNAs miR-140-5p and miR-140-3p. miR-140-5p/-3p have been associated with the development and progression of cancers, but also non-neoplastic diseases. In aging-related diseases, miR-140-5p and miR-140-3p expressions are modulated. The seric levels of these two miRNAs are used as circulating biomarkers and may represent predictive tools. They are also considered key actors in the pathophysiology of aging-related diseases. miR-140-5p/-3p repress targets regulating cell proliferation, apoptosis, senescence, and inflammation. This work focuses on the roles of miR-140-3p and miR-140-5p in aging-related diseases, details their regulation (i.e., by long non-coding RNA), and reviews the molecular targets of theses miRNAs involved in aging pathophysiology.
The first microRNA (miRNA), lin-4, was described in 1993 in the nematode Caenorhabditis elegans, by the Ambros and Ruvkun groups [1,2]. Since then, more than 48,800 mature miRNAs have been discovered from 271 organisms; among them, 2654 mature miRNAs sequences have been reported in humans (miRBase v.22) [3,4]. miRNAs are small non-coding RNAs involved in gene expression regulation. The average length of miRNAs is 22 nucleotides (nt), of which seven correspond to the seed region (nucleotides 2 to 8). Genes encoding microRNAs are intergenic or intragenic. In the first case, a specific promoter regulates transcription; in the second case, miRNA genes are mainly located in introns, whereas few of them are located in exons [5,6,7]. Conventionally, miRNAs genes are transcribed to primary miRNAs (pri-miRNAs; >1 kb) using RNA polymerase II. Then, pri-miRNA is processed by a complex including DiGeorge Syndrome Critical Region 8 protein (DGCR8) and a RNase III enzyme Drosha, which cleaves the pri-miRNA into a 70 nt stem-loop precursor miRNA (pre-miRNA) [8,9,10,11,12]. Pre-miRNA exportation to the cytoplasm is controlled by RanGTP/exportin 5 complex [13,14,15]. Lastly, the cytoplasmic RNase III endonuclease Dicer removes the terminal loop, leading to a ≈ 25 nt miRNA duplex [16,17]. This duplex is integrated in “RNA-induced silencing complex”, called “RISC-complex”, containing the Argonaute protein (AGO). The passenger strand is eliminated from the duplex, whereas the guide strand, corresponding to the mature (≈22 nt) miRNA, is conserved [18]. Both the 5p and 3p strands, arising from the 5′ or 3′ end of the duplex, respectively, can be the guide strand, depending to different factors; e.g., the cell type, developmental stage, enrichment of purines at the 5′end, and miRNA thermodynamic stability [19]. Several non-classical miRNA biogenesis pathways have also been described, independent of Drosha, Dicer, or AGO proteins [5]. Mirtrons are a class of intronic but non-canonical miRNAs that are processed by a DGCR8/Drosha-independent pathway and are generated by spliceosome from direct intron splicing. “Debranched” mirtrons are similar to canonical pre-miRNAs and are exported by Exportin-5 in the cytoplasm. Their further maturation is identical to the canonical manner [20,21,22]. Intronic miRNAs genes are embedded into a host gene and are mostly processed according to the protein-encoding gene in which they are hosted. In a few cases, intronic miRNA expression differs from the expression of the host gene: miRNA transcription can be independent from the host gene promoter, as the miRNA genes can be regulated by their own independent promoter [23,24,25,26]. MiRNAs mainly regulate gene expression due to messenger RNA (mRNA) interactions, inside the RISC complex: the seed region located to the 5′ end of microRNAs usually targets the 3′ untranslated transcript region (3′UTR) of mRNA, leading to mRNA degradation or translation inhibition, according to the base complementarity between these two sequences. In a few cases, a miRNA targets the 5′UTR of mRNA or coding sequence, and this interaction has been reported to induce translation under specific conditions (e.g., in quiescent cells) [27,28,29,30]. It has also been well established that some miRNAs can be reimported into the nucleus, where they directly interact with gene promoters, thus inducing gene transcription [31,32]. Considering all these mechanisms, miRNAs are considered major actors in gene expression regulation. In addition, one microRNA is rarely specific to one mRNA, and one mRNA can be targeted by different microRNAs sharing the same seed region. Many studies and reviews are available reporting miRNA involvement in physiological pathways and developmental processes, as well as their deregulation under pathological conditions leading to diseases, such as cancers or genetic diseases. They also have been proposed as potential biomarkers or therapeutic targets in various pathological contexts [33,34,35,36,37].
Gene encoding microRNA-140-5p/-3p is located on chromosome 16q22.1, and is hosted in the 15th intron of the WW domain containing E3 Ubiquitin protein ligase 2 (WWP2) human gene. In mice, miR-140 gene is localized in the 16th intron of WWP2 gene [7,38]. Its transcription generates two transcripts, miR-140-3p and miR-140-5p, processed from the same pre-miR. miR-140 gene transcription is independent of the transcription of its host gene, and is regulated by transcription factors, such as Nuclear factor of activated T-cells 3 (NFAT3) and Mothers against decapentaplegic homolog 3 (SMAD3) [38]. Both miR-140-3p and miR-140-5p can be the guide strand and silence mRNAs. Interestingly, as these two miRNAs do not share the same seed region (Figure 1), the predicted targeted mRNAs are thus different [38]. miR-140 (i.e., miR-140-5p in the former studies) was first described as cartilage specific [39]. A large number of studies have demonstrated its key role in osteoarthritis (OA) development. Both miR-140-5p and miR140-3p have also been proposed as plasmatic biomarkers for several diseases and their involvement in cancers has largely been described [40]. In cancer, as well as in an aging context, miR-140-5p and miR-140-3p target molecules involved in cell senescence: anti-senescent molecule forkhead box Q1 (FOXQ1) [41] is a target of miR-140-3p in bladder carcinoma [42], and miR-140-5p enhances cell senescence due to Peptidyl-prolyl cis-trans isomerase NIMA-interacting 1 (PIN1) repression, in a model of progenitor cell aging [43]. This review focuses on the roles of miR-140-3p and miR-140-5p in neurodegenerative diseases, atherosclerosis, visual acuity loss, osteoporosis, and osteoarthritis, which are all age-related diseases, and details the molecular targets of these miRNAs involved in aging pathophysiology.
Neurodegenerative diseases, such as Alzheimer’s disease (AD), Parkinson’s disease, or post-stroke infarction, lead to cognitive impairment with aging. Several miRNAs have recently been considered as neuroregulators, as well as neuroprotectors, leading to their consideration as biomarkers or therapeutic targets for neuronal cell death [44]. To our knowledge, no study has evidenced miR-140-5p/-3p involvement in Parkinson’s disease. In contrast, two recent publications have highlighted the neurotoxic role of miR-140-5p in AD, based on the study of different AD models. AD is caused by accumulation of β-amyloid (Aβ) in the brain, forming neurotoxic amyloid plaques. miR-140-5p overexpression was observed in hippocampal tissues from a rat model of AD, and associated with neurological function impairment. In hippocampal mouse neurons from an in vitro model of AD, miR-140-5p overexpression (using mimics transfection) led to reactive oxygen species (ROS) production and mitochondrial dysfunction, by directly targeting phosphatase and tensin homolog-induced putative kinase 1 (PINK1), involved in mitochondrial function preservation [45]. miR-140-5p is also upregulated in post mortem brains from AD patients, compared to cognitively normal controls, causing a deleterious circle as it directly targets A Disintegrin and Metalloproteinase 10 (ADAM10) and SRY-box transcription factor 2 (SOX2), a transcription factor that positively regulates ADAM10 transcription. ADAM10 expression is thus reduced in AD, leading to Aβ plaque accumulation [46]. Plasmatic miR-140-5p/-3p has also been proposed as a predictive tool for cognitive function evaluation during aging: miR-140-5p levels are positively associated with cognitive performance measures [47]. miR-140-3p, in combination with other defined miRNAs, could represent a plasmatic biomarker, predictive for mild cognitive impairment (MCI—an intermediate state between normal aging and dementia) in elderly populations [48]. Similarly, the miR-140-5p plasmatic level is enhanced during acute ischemic stroke in humans [49]. Interestingly, in a mouse model of ischemic stroke (middle cerebral artery occlusion), miR-140-5p expression was drastically decreased in response to ischemic stroke, in comparison to control mice; while miR-140-5p ectopic overexpression in brain prevented neuronal apoptosis and restrains the development of ischemic injury in this model, via toll-like receptor 4/nuclear factor-kappa B (TLR4/NF-κB) axis regulation [50]. Similar results were obtained in a rat model of ischemia stroke [51,52]. miR-140-3p was also described as neuroprotective from cerebral ischemia-reperfusion injury in an in vitro model (PC12 cells) [53]. Moreover, miR-140-5p directly targets the 3’UTR of Vascular Endothelial Growth Factor A (VEGFA), a pro-angiogenic factor and inhibits endothelial cell properties in vitro (proliferation, migration, and pseudo-capillar formation) under normoxic and hypoxic conditions [52]. These studies suggest that miR-140-5p is downregulated after ischemic stroke. Preventing the downregulation of miR-140-5p and, to a lesser extent, miR-140-3p could be neuroprotective and represent a new therapeutic approach, by inhibiting neuro-inflammation and abrogating angiogenesis following ischemic brain injury.
Atherosclerosis (AS) is one of the main factors leading to cardiovascular diseases, which are the leading cause of death worldwide [54]. AS is characterized by formation of a plaque in the arterial intima, essentially composed of lipids, such as triglycerides and oxidized Low-Density Lipoprotein (ox-LDL), generating immune and inflammatory responses. Ultimately, plaque growth or plaque rupture causes damage as it induces vessel obstruction [55,56]. In atherosclerotic plaques, phagocytosis is activated by ox-LDL; macrophages accumulate lipids and derive into foam cells. Ox-LDL are the origin of a vicious circle, as they stimulate ROS production, thus accentuating ox-LDL generation. They cause cellular damage in macrophages and they stimulate macrophages phagocytosis and activation into foam cells [57]. During AS, vascular smooth muscle cell (VSMC) dysfunction was also evidenced. Phenotypic alterations, such as gain of proliferation and migration, contribute to plaque progression and vascular injury [58]. Opposite results were published concerning the role of miR-140-5p in lipid accumulation into macrophages-derived foam cells: in a cellular model of AS macrophages (ox-LDL treated THP1 cells), miR-140-5p expression was inversely correlated with ox-LDL macrophages content, suggesting that in AS, miR-140-5p is downregulated. miR-140-5p could participate as a brake in AS, by targeting TLR4. Indeed, miR-140-5p mimic transfection in AS macrophages prevented lipids accumulation, ROS production, and cell apoptosis [57]. In contrast, Zhao and colleagues demonstrated, in similar in vitro model, that miR-140-5p expression was enhanced in ox-LDL-treated macrophages. They showed that miR-140-5p silences Regulatory Factor X7 (RFX7), a transcription factor that binds to the promoter of ABCA1 gene to induce its expression [59]. This gene encodes ATP-binding cassette transporter A1 (ABCA1), an ABC transporter involved in cholesterol efflux outside macrophages [60]. miR-140-5p could thus amplify the lipid accumulation in atherosclerotic plaque macrophages, participating to AS aggravation. Other arguments support these results. In ApoE−/− mice, an in vivo model of AS, overexpression of miR-140-5p accelerated plaque formation in aorta, in association with reduced levels of RFX7 and ABCA1 in this tissue [59]. miR-140-5p is also involved in VSMC alterations during AS. In AS aorta from ApoE−/− mice, miR-140-5p was enhanced compared to wild type mice, thus increasing oxidative stress, resulting from the increase of ROS production and the decrease of antioxidant molecules (superoxide dismutase (SOD) and gluthathione (GSH)) [61]. Sirtuin 2 (SIRT2) and nuclear factor erythroid 2-related factor 2 (NRF2), two direct targets of miR-140-5p, may strongly participate in oxidative stress in AS. The mechanistic role of SIRT2 in oxidative stress remains unclear as SIRT2 regulates hypoxia inducible factor 1 (HIF1) expression, but its positive regulation (stabilization) or negative regulation (degradation) is still debated [61,62]. In response to oxidative conditions, NRF2 is translocated into the nucleus, where it acts as a transcription factor activating anti-oxidative gene transcription [61,63]. miR-140-5p expression was also significantly higher in artery tissues from AS patients versus healthy controls, as well as in ox-LDL-treated human VSMCs compared to control conditions. In addition, miR-140-5p inhibition represses AS VSMC migration, proliferation, and stimulates apoptosis, via rescue of roundabout guidance receptor 4 (ROBO4) expression, a vascular-specific receptor involved in angiogenesis [64]. LDL-cholesterol clearance from blood circulation also contributes to reducing atherosclerotic plaque formation. Plasmatic LDL-cholesterol endocytosis, via LDL-receptors (LDLR) on hepatocytes, is a major event controlling plasmatic LDL-cholesterol levels. In humans, the seed region of miR-140-5p was predicted to hybridize with 3′UTR of LDLR. This prediction was experimentally confirmed in human hepatic cells, in which miR-140-5p inhibits the cell surface LDLR expression and reduces LDL-cholesterol absorption, while simvastatine reverts this phenomenon, as this drug is able to reduce miR-140-5p expression [65]. miR-140-5p expression seems to be endogenously regulated during AS. Prostate cancer antigen 3 (PCA3), a long non-coding RNA (lncRNA), physiologically sponges miR-140-5p in macrophage-derived foam cells, thus repressing its deleterious role in AS [59]. Moreover, the plasmatic level of miR-140-5p is inversely correlated with the lncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), which may inhibit miR-140-5p [66]. Concerning miR-140-3p, this miRNA was significantly downregulated in AS aorta from ApoE−/− mice, leading to MKK6 and TP53RK proteins overexpression, compared to control aorta. These two kinases have been reported to be involved in signaling cascades playing a pivotal role in endothelial injury [67,68]. Similarly, in a commercial cell line of human coronary endothelial cells (HCAECs), miR-140-3p directly targets mitogen-activated protein kinase 1 (MAPK1), encoding a serine/threonine protein kinase that belongs to the MAPK family. Activation of MAP kinase signaling results in the production of inflammatory cytokines and chemokines via NF-kB nuclear importation [69,70]. In a vascular injury model (i.e., in-stent restenosis), miR-140-3p was downregulated in arterial smooth muscle cells (ASMC) isolated from restenosis artery wall compared to normal arteries, resulting in abnormal ASMC proliferation enhancement and apoptosis inhibition, which contributed to arterial damage. In this model, C-Myb and B-cell lymphoma 2 (BCL-2) are no longer targeted by miR-140-3p, causing their upregulation. C-Myb is a transcription factor known to induce ASMC proliferation, while BCL-2 is an outer membrane mitochondrial anti-apoptotic protein [71]. As described for miR-140-5p (see above), several non-coding RNAs are predicted or proven to sponge miR-140-3p in AS. In plasma from coronary atherosclerotic heart disease patients, miR-140-3p was inversely correlated with the lncRNA nuclear enriched abundant transcript 1 (NEAT1) [69]; and in AS aorta from ApoE−/− mice, miR-140-3p was negatively regulated by circRNA_36781, also called circRNA_ABCA1 [67], leading to miR-140-3p downregulation and facilitating endothelial injury in these models.
During aging, visual acuity loss occurs. Age-related macular degeneration (AMD) is the leading cause of visual impairment in Europe [72]. AMD is divided into two clinical entities: wet and dry AMD [73]. In a cohort of 33 wet AMD patients versus 31 controls, the expression of a panel of 384 microRNAs was investigated in plasma. In this study, miR-140-3p, among 10 others, was found to be downregulated in the AMD group, suggesting that plasmatic miRNAs levels could be predictive for wet AMD. Nevertheless, this observation needs to be confirmed in a larger cohort, and no molecular mechanisms were explored, to highlight the potential role of miR-140-3p in wet age-related macular degeneration [74]. In age-related cataract, another cause of visual acuity loss occurring during aging, no studies have shown a role for miR-140-5p/-3p, unlike other miRNAs [75].
Osteoporosis is a systemic bone metabolic disorder, characterized by dysregulation of bone turnover, associated with low bone density and mass and accelerated bone loss, leading to bone fragility and fractures [76]. Osteoporosis affects life quality during aging, notably in postmenopausal women. To prevent this worldwide health problem, regular physical exercise is an effective method for stimulating bone osteogenesis in osteoporotic patients, and which improves bone density and reduces osteoporosis-induced risk of fracture [77]. Osteogenic differentiation was induced in a model of bone marrow mesenchymal stem cells (BMSCs) submitted to tensile strain, a mechanical stimulation mimicking physical exercise at the cellular level, together with lncRNA-MEG3 expression enhancement; the latter represses miR-140-5p expression through a “sponging” mechanism [78]. Nevertheless, in a cohort of idiopathic osteoporosis patients with prevalent low-traumatic fractures (pre- and postmenopausal women, and men), seric miR-140-5p, among other miRNAs, was downregulated in comparison to healthy subjects and was inversely correlated to body mass index (BMI) [79]. Conversely, an elevated seric level of miR-140-3p could be a biomarker of osteoporosis, osteopenia, and fractures in post-menopausal women, as miR-140-3p was overexpressed in the serum of osteoporotic post-menopausal women, compared to healthy post-menopausal women [80]. These results were reinforced by Yin et al., who reported higher levels of miR-140-3p in PBMC and serum from osteoporotic post-menopausal women compared to healthy controls. Moreover, in an osteoblastic progenitor cell line (C2C12 cells), miR-140-3p silencing induced cell proliferation and differentiation, and inhibited apoptosis, by targeting Phosphatase and TENsin homolog (PTEN) and thus, inactivating the Phosphatase and TENsin homolog/ Phosphoinositide 3-kinase/ serine/threonine kinase 1 (PTEN/PI3K/AKT) signaling pathway [81]. These results corroborate an older publication, which highlighted the overexpression of miR-140-3p in BMSCs from a model of osteoporotic rats, compared to control animals. The inhibition of miR-140-3p in this cell model promoted osteogenesis [82]. The role of miR-140-3p in osteoblastic differentiation has still not been totally elucidated, as opposite results have been reported [83]: Fushimi et al. demonstrated that miR-140-3p overexpression activates osteoblasts differentiation by targeting Transforming Growth Factor-β3 (TGF-β3) [84]. In contrast, Mao et al. showed in MC3T3-E1, an osteoblastic cell line obtained from mouse calvarium, that upregulation of miR-140-3p suppresses viability and differentiation, by targeting MCF.2 cell line derived transforming sequence-like (MCF2L) [85].
Osteoarthritis (OA) is a chronic degenerative joint disease that affects quality of life in elderly populations. OA is associated with articular pain and stiffness, leading to functional disabilities. Several predisposing factors have been identified such as age, gender, obesity, and traumatic injury. Extracellular matrix (ECM) remodeling is impaired, in response to chondrocytes senescence and lesser proteoglycan secretion by chondrocytes. High levels of proinflammatory cytokines (e.g., Interleukin-6 (IL-6)) and matrix-degrading enzymes (e.g., Matrix metallopeptidase 13 (MMP13) and a disintegrin and metalloproteinase metallopeptidase with thrombospondin type 1 motif 5 (ADAMTS5)) are secreted in OA cartilage and synovial fluid. The elevated levels of reactive oxygen species (ROS) associated with oxidative stress cause oxidative damage, leading to OA [86]. During aging, various epigenetic changes occur, such as DNA methylation, histone modifications, and miRNAs regulation. Theses mechanisms are also present in aging-related OA. miR-140-3p and miR-140-5p are physiologically expressed in cartilage and chondrocytes, and have been highlighted as key actors in OA [87,88]. In normal conditions, miR-140-5p expression increases during chondrogenesis, in parallel to chondrogenic differentiation markers (SRY-related high-mobility group box 9 (SOX9) and Collagen type-II (COL2A1)), as the levels are higher in mature chondrocytes than in Mesenchymal stem cells (MSCs) [89]. Mouse models were generated to investigate the miR-140-5p/3p involvement in cartilage development and regeneration. Specific deletion of the miR-140 gene sequence in the intronic region of Wwp2 gene led to the abolishment of both miR140-3p and miR-140-5p expression in these miR-140−/− mice, whereas transgenic mice overexpressing both miR140-3p and miR-140-5p in cartilage were generated, using pri-miR-140 insertion near the cartilage-specific Col2a1 promoter [90]. Mouse developmental observations showed that miR-140−/− mice exhibited a mild skeletal phenotype with a short stature, and during aging, OA damage was associated with ECM degradation and cartilage fibrosis. Moreover, transgenic mice were protected from OA development in a model of antigen-induced arthritis [90]. miR-140-5p was reduced in joint tissues from old mice compared to healthy young mice [91] and many teams have described a downregulation of miR-140-5p expression in human OA-cartilage, in comparison to healthy cartilage [38,89,92]. This decrease of miR-140-5p in OA-cartilage was corroborated by in vitro models of OA-chondrocytes, which have also objectified miR-140-5p expression decreases [89,93,94]. Regulatory mechanisms for miR-140-5p are opposed, independently of those regulating WWP2: NFAT3 increases miR-140-5p levels in OA-chondrocytes, and SOX9 physiologically stimulates miR-140-5p expression during chondrogenesis [38,95]. Conversely, several mechanisms negatively regulate miR-140-5p expression in OA-chondrocytes: TGF-β/SMAD3, hypermethylation of CpG sites of the regulatory region of miR-140 gene, and Interleukin-1β (IL-1β) stimulation [38,88,96]. Moreover, in chondrocytes extracted from OA cartilage tissue, LINC01534 was overexpressed in comparison to healthy cartilage tissue, and this lncRNA directly targets and sponges miR-140-5p. LINC01534 overexpression is mediated by IL-1β [97]. In an in vitro model of OA chondrocytes (IL-1β treatment), miR-140-5p was decreased compared to non-stimulated chondrocytes. In this model, miR-140-5p transfection prevented chondrocytes senescence, and ECM degradation, as the expression of two major ECM components (COL2 (collagen type II) and ACAN (Aggrecan)), was restored, in association with a reduction of ECM degradation enzymes [89]. Comparatively, in a rat OA model, intra-articular injection (IAJ) of miR-140-5p attenuates OA progression and prevents chondrocyte senescence [93,94,98]. During OA, miR-140-5p downregulation contributes to cartilage remodeling impairment. The absence of inhibition of miR-140-5p targets in chondrocytes results in chondrocytes senescence (via NUMB-like endocytic adaptator protein (NUMBL) and Jagged1 (JAG1) in the Notch pathway), chondrocytes pyroptosis (apoptosis occurring in an inflammatory state) (via cathepsin B/Nod-like receptor protein 3), inflammation (via IL-1β, Interleukin-6 (IL-6), SMAD3, C-X-C motif chemokine receptor 4 (CXCR4)), ECM catabolism (MMP13, ADAMTS5), and inhibition of chondrocytes proliferation (fucosyltransferase 1 (FUT1)) (Figure 2) [88,90,99,100,101,102,103,104,105]. miR-140-3p is also expressed in cartilage to a greater extent than miR-140-5p [87]. Similarly to miR-140-5p, miR-140-3p is downregulated in OA-chondrocytes [38]. This downregulation may be the result of direct targeting of miR-140-3p by the lncRNA LINC01385 in OA-tissues [106]. A few targets of miR-140-3p have been identified as being involved in OA progression: miR-140-3p attenuates OA progression and regulates chondrogenesis due to CXCR4 and Ras-like proto-oncogene (RALA) inhibition, respectively [99,100]. miR-140-3p level has been monitored in the serum from OA patients before and after high tibial osteotomy (HTO), which represents one option for OA-treatment, as cartilage regeneration is expected after this surgery. miR-140-3p expression is significantly up-regulated 6 and 18 months post-surgery, indicating that the miR-140-3p seric level may represent a prognostic biomarker for the cartilage repair process [107]. Considering all these data, miR-140-5p/-3p represent serious targets for OA treatment. As described previously, intra-articular injection of miR-140-5p in animal models of OA slows down OA progression [93,94,98]. To protect miR-140-5p from degradation, one team generated exosomes overexpressing miR-140-5p (miR-140-5p transfected cells) or not (untransfected cells) and derived from human synovial mesenchymal stem cells. These overexpressing miR-140-5p exosomes stimulated chondrocytes proliferation in vitro and prevented OA progression in a rat model [108]. In humans, gene editing has been performed using the CRISPR/Cas9 method, in order to silence miR-140-5p/-3p expression in primary OA chondrocytes, without WWP2 expression modulation. This recent development has only been used at this stage to identify miR-140-5p/-3p targets, but opens up great possibilities as a tool to further evaluate the role of miRNAs in AO and to assess potential new therapies [109].
miR-140-5p/-3p have been widely studied in the context of cancer, but also in diseases related to aging, particularly in osteoarthritis. Both miR-140-5p and miR-140-3p are derived from the same gene, but have different targets. Despite this, these two miRNAs are often involved in the same pathologies with similar or opposite actions. They are key actors in the pathophysiology of many diseases and/or can be used as predictive circulating biomarkers. A list of miRNAs regulating aging-related pathways, such as senescence, DNA damage response, has been established; the miRNAs involved in physiological aging are thus, grouped under the term “geromiRs” [110]. Even if miR-140-5p/-3p are not listed as “GeromiRs”, this review highlights that miR-140-5p/3p, and particularly miR-140-5p, are involved in the most common aging-related pathologies. Considering the major involvement of miR-140-5p/-3p in physiological aging, it would be interesting to analyze the role of these miRNAs in the pathophysiology of diseases with accelerated and premature aging. | true | true | true |
PMC9570155 | 36233123 | Xiaodan Zhang,Xiaohu Zhou,Midi Wan,Jinxiang Xuan,Xiu Jin,Shaowen Li | PINC: A Tool for Non-Coding RNA Identification in Plants Based on an Automated Machine Learning Framework | 05-10-2022 | plant,ncRNA identification,AutoGluon,tool | There is evidence that non-coding RNAs play significant roles in the regulation of nutrient homeostasis, development, and stress responses in plants. Accurate identification of ncRNAs is the first step in determining their function. While a number of machine learning tools have been developed for ncRNA identification, no dedicated tool has been developed for ncRNA identification in plants. Here, an automated machine learning tool, PINC is presented to identify ncRNAs in plants using RNA sequences. First, we extracted 91 features from the sequence. Second, we combined the F-test and variance threshold for feature selection to find 10 features. The AutoGluon framework was used to train models for robust identification of non-coding RNAs from datasets constructed for four plant species. Last, these processes were combined into a tool, called PINC, for the identification of plant ncRNAs, which was validated on nine independent test sets, and the accuracy of PINC ranged from 92.74% to 96.42%. As compared with CPC2, CPAT, CPPred, and CNIT, PINC outperformed the other tools in at least five of the eight evaluation indicators. PINC is expected to contribute to identifying and annotating novel ncRNAs in plants. | PINC: A Tool for Non-Coding RNA Identification in Plants Based on an Automated Machine Learning Framework
There is evidence that non-coding RNAs play significant roles in the regulation of nutrient homeostasis, development, and stress responses in plants. Accurate identification of ncRNAs is the first step in determining their function. While a number of machine learning tools have been developed for ncRNA identification, no dedicated tool has been developed for ncRNA identification in plants. Here, an automated machine learning tool, PINC is presented to identify ncRNAs in plants using RNA sequences. First, we extracted 91 features from the sequence. Second, we combined the F-test and variance threshold for feature selection to find 10 features. The AutoGluon framework was used to train models for robust identification of non-coding RNAs from datasets constructed for four plant species. Last, these processes were combined into a tool, called PINC, for the identification of plant ncRNAs, which was validated on nine independent test sets, and the accuracy of PINC ranged from 92.74% to 96.42%. As compared with CPC2, CPAT, CPPred, and CNIT, PINC outperformed the other tools in at least five of the eight evaluation indicators. PINC is expected to contribute to identifying and annotating novel ncRNAs in plants.
RNA is the template that codes for the proteins required to create cellular functions. RNA is structurally similar to DNA, but its function and chemical composition are fundamentally different. At a higher level, RNA is divided into two main groups: coding RNA that accounts for approximately 2% of all RNAs, and non-coding RNA (ncRNA) that accounts for the majority (>90%) of RNAs [1]. Non-coding RNA refers to all RNAs that are transcribed from DNA but do not code for proteins. Additionally, ncRNA can be categorized into two groups according to the size of the sequence: long non-coding RNAs (lncRNAs) with sequences >200 nucleotides and small non-coding RNAs (sncRNAs) with sequences shorter than 200 nucleotides [2]. In previous research, ncRNAs have frequently been referred to as “useless genes” or transcriptional “noise” [3,4]. In contrast, a growing number of experiments have demonstrated that ncRNAs play important biological roles in a variety of biological processes, including gene regulation/expression, gene silencing, RNA modification and processing, as well as multiple important roles in life activities [5,6,7]. Numerous plant-specific biological processes, including the regulation of plant nutrient homeostasis, development, and stress responses, have been linked to ncRNAs [8,9,10]. MiRNAs and trans-acting siRNAs, for instance, contribute to leaf senescence in Arabidopsis; miR164 and its target ORE1 control leaf senescence in Arabidopsis, and as miR164 expression declines, ORE1 expression eventually increases [11]. In addition, overexpression of miR398b has been shown to decrease the transcript levels of genes encoding superoxide dismutase (CSD1, CSD2, SODX, and CCSD), which resulted in the production of reactive oxygen species (ROS) and increased rice resistance to Magnaporthe oryzae [12,13]. In recent years, to facilitate subsequent analyses and research of transcripts, ncRNA identification has been one of the tasks that needs to be addressed. Numerous bioinformatics methods and experiments have been developed for ncRNA identification and to evaluate their functions [14,15]. Genomic SELEX, microarray analysis, and chemical RNA-Seq are the most commonly used experimental techniques [16]; however, they are costly and time-consuming. Therefore, bioinformatics may be a more effective means of addressing the biological aspects of the problem. Kong et al. developed the Coding Potential Calculator (CPC) in 2007 [17]. The CPC selected a number of biologically significant features, including ORF quality, coverage, and integrity. These features were incorporated into a support vector machine for coding potential identity, but its performance was dependent on sequence comparisons. CPC was revised in 2017 with the release of CPC2 [18]. CPC2 is faster and more accurate than CPC, and, as an input to the SVM model, it uses ORF size and integrity, a Fickett score, and the isoelectric point extracted from the original RNA sequence. CPC2 is a relatively neutral tool, which makes it somewhat more applicable to transcriptomes of non-model organisms. CPAT, developed by Wang et al. in 2013, is a logistic-regression-model-based ncRNA identification tool that classifies ncRNAs and cRNAs based on features such as ORF size and coverage, Fickett score, and hexamer score [19]. CNCI was proposed by Liang et al. in 2013, and while it is also based on the same SVM classifier as CPC2, it uses different features, categorizing ncRNA and cRNA based on ANT features [20]. CNIT is an updated version of CNCI that was released in 2019. CNIT employs the more robust integrated machine model XGBoost for classification [21]. Tong et al. introduced CPPred in 2019 [22] as an SVM-based tool. This tool distinguishes between ncRNAs and coding RNAs using the same ORF features as CPC2, as well as the isoelectric point, stability index, gravity three peptide, hexamer score, CTD, and Fickett score features. A number of tools have been published that can distinguish between ncRNA and coding RNA; however, the tools have some limitations, for example, their application is mainly limited to vertebrates and mammals. In addition, these tools rarely consider using plants for model training. Most tools only use the model plant Arabidopsis, and rarely involve other non-model plants. Moreover, since ncRNAs of animals are mainly transcribed by polymerase II, while ncRNAs of plants are mainly transcribed by RNA polymerase II, IV, and V [23], and ncRNAs are characterized by low-level expression and cross-species conservation [24], these tools for ncRNA identification in animals cannot guarantee the reliability in plants. Therefore, it is necessary to construct a powerful tool for ncRNA identification in plants. Automatic machine learning (AutoML) is the process of applying machine learning to real-world problems in an automated manner. Since 2013, frameworks have been developed that have been based on the AutoML concept. AutoWEKA was the first AutoML framework to emerge [25]; it automatically selected models and hyperparameters. Additionally, H2O [26] and TPOT [27] were created. H2O is a JAVA-based framework that supports multiple types of grid searches to identify the optimal parameters following the generation of an integrated model. At its core, TPOT is a tree-based process optimization tool based on a genetic algorithm. Today, more and more frameworks, such as AutoGluon [28] and AutoKeras [29], have been developed based on the concept of AutoML. These frameworks have also been applied to Alzheimer’s disease diagnosis [30], biomedical big data [31], and additional bioinformatics fields [32]. In this experiment, we developed PINC, an AutoML-based instrument for the identification of ncRNAs and cRNAs in plants. The AutoML framework does not require a great deal of effort and time to optimize the model; it simply accepts the processed data as an input, tunes and sets the framework’s parameters, and then outputs the model automatically. Our experimental results include a number of significant contributions: (1) By combining the F-test and variance threshold, 10 out of 91 features were identified as being able to strongly distinguish between ncRNAs and coding RNA in plants. (2) Using the AutoML framework, a neutral model for non-coding RNA identification was obtained. (3) We combined the two previous points and developed a tool called PINC for ncRNA identification. After comparing PINC with the CPC2, CPAT, CNIT, and CPPred identification tools on nine independent test sets to validate the performance of PINC, we discovered that PINC performed exceptionally well on these independent test sets. This suggests that PINC is a reliable method for ncRNA identification in plants. In addition, users can upload their data for identification, which facilitates the study of plants that have received less attention.
Once the features were selected, the models were tuned to find the best parameters, and the results were validated using a five-fold cross-validation procedure. A benchmark dataset of 4000 randomly selected data from each class was constructed for training and validation. Meanwhile, to ensure the validity of the experiments, we repeated the above experiments 100 times. As shown in Figure 1A, the highest accuracy of the 100 experiments was 95.32% and the lowest was 94.52%, mostly distributed between 94.6% and 94.9%, with very small fluctuations. For further proof, we averaged the accuracy of every fifth experiment, as shown in Figure 1B, and the curve fluctuates even less. This result shows that the randomly selected data is representative of the entire data set. Therefore, we took 4000 randomly selected data from each class as our baseline dataset.
In this research, 91 features were filtered using four feature selection methods: F-test, variance threshold, RF, and variance threshold combined with F-test (VT-F). These feature selection methods were compared in order to assess their usefulness. These feature selection methods use learning curves to continuously reduce the number of available features and to select the most appropriate features. The maximum validation set accuracy was 94.77 percent when the first 31 features were chosen using F-test filtering, and it was 94.29 percent when the first 25 features were chosen using variance threshold filtering. For VT-F, features below the mean were first filtered out using a variance threshold, and then the remaining features were filtered using the F-test, with a maximum accuracy of 95.25 percent when the first 10 features were selected. The evaluation of the three previously described feature selection methods was based on the AutoGluon model. For RF, the range of features was narrowed down based on the importance of the features, with the highest accuracy of 93.27 percent when the first 21 features were selected, and the 21 features were then fed into AutoGluon with an accuracy of 94.72 percent. In addition to accuracy, we compared SE, F1, MCC, and SPC performance metrics. Table 1 demonstrates that the method combining the F-test and variance threshold for feature selection outperformed the other methods, and the 10 features it selected were GC content, score, cdsStop, cdsSize, and T, C, GT, GC, ACG, and TAT frequencies. The experiments analyzed the distribution of ncRNAs and coding RNAs on the dataset for these 10 features, and based on Figure 2, it can be seen that they play a significant role in the identification of discriminatory power. In addition, we conducted a correlation analysis between the ten features selected for the classification task. Figure 3 showed that, GC content had a weak correlation with the other features. Score, cdsStop, and cdsSize showed a stronger correlation with the other features. T, C, GT, GC, ACG, and TAT frequencies had the strongest correlation with the other features.
Regarding the validation set, in this study, we compared the five-fold cross-validation results of four AutoML frameworks, AutoGluon, TPOT, H2O, and AutoKeras, to those of three conventional machine learning models, i.e., random forest, SVM, and Naive Bayes (Table 2). It is evident that, in general, conventional machine learning models are less effective than AutoML; three of the four automated machine learning frameworks produced more effective models than the random forest, the best performing conventional machine learning model. AutoGluon achieved the best results for five of the eight evaluation metrics within the AutoML framework: ACC, F1, MCC, NPV, and SE. H2O achieved the best results for AUC, while Autokeras achieved the best results for PPV and SPC. It is evident that the AutoGluon framework is more effective than the other frameworks, possibly because AutoGluon employs per-variable embedding, which improves quality via gradient flow, whereas the other frameworks merely apply the standard feed-forward architecture to hot-coded data. The Autokeras effect, which is based on NAS that combines multiple search strategies such as random search, grid search, etc., is only marginally weaker than the AutoGluon effect. The goal of NAS is to reduce human intervention and to allow the algorithm to design the neural network automatically, which consists of three key components: the search space, the search strategy, and the evaluation strategy. However, this process is typically very time-consuming. H2O had the highest AUC score, but its overall performance was comparable to that of conventional machine learning models and TPOT. H2O is a distributed machine learning platform based on the Java programming language, unlike other AutoML frameworks. TPOT was the least effective AutoML and the only framework with overall lower results than conventional machine learning models. This is likely due to the genetic algorithm employed by TPOT, which tends to converge on a locally optimal solution prematurely. Consequently, the comparison demonstrates that the models created by the AutoGluon framework are superior to those created by the other four automatic machine learning frameworks and the three conventional machine learning models.
To evaluate the accuracy of PINC in ncRNA and coding RNA identification, we compared it to CPC2, CPAT, CNIT, and CPPred. We compared the identification accuracy for nine plant species from four databases, GreeNC, CANTATA, RNAcentral, and Phytozome, using five different tools. It is evident from the results shown in Figure 4 that our tool has the highest degree of precision for all nine plants. The large fluctuation of CPPred indicates that it has poor generalization performance, whereas the other three tools have some stability. However, it can be seen that the identification accuracy of PINC is greater than that of the other three tools, indicating that our tool performs the best among the different plant species. To compare the performances of these five tools further, we used eight metrics: sensitivity (SE), specificity (SPC), accuracy (ACC), F1-score, PPV, NPV, MCC, and AUC to evaluate and compare the five tools for these nine independent test sets (Table 3). We plotted the ROC curve (Figure 5); it can be seen that the ROC curve for PINC differs from the other tools. A true positive rate is rapidly achieved (1.0) at the cost of a relatively high false positive rate. Therefore, we have also plotted PR curves (Figure 6) to further illustrate the performance of PINC. The results showed that the PR curve of PINC did not fluctuate markedly and had a decreasing trend when the threshold was greater than 0.8. Meanwhile, PR curves illustrated that Precision and Recall values of five plants (Cicer arietinum, Manihot esculenta, Nymphaea colorata, Sorghum bicolor, and Zea mays) were higher than the other tools at the same threshold. All those results showed that PINC had the superior performance for distinguishing ncRNAs from coding RNAs. Solanum tuberosum outperformed the other tools in seven of the eight evaluation metrics and at least five of the remaining eight test sets, namely, SE ACC, F1, NPV, and MCC. The high Se score indicates that the probability of missing is small; therefore, PINC is the best choice for ncRNA identification. For the specificity SPC score, only one dataset was higher than the other tools, with four datasets performing best on CNIT and two datasets performing best on CPC2 and CPAT, respectively. However, the difference between the SPC of PINC and the SPC of the other tools was not large, and all tools had high performances above 86.99%. Among the five tools, PINC was the most effective for ncRNA identification in the nine plants. This indicates that our tool has a strong generalization to plants, which is crucial for non-model plants.
In the field of bioinformatics, automated machine learning methods are now beginning to be implemented. In our experiments, we compared four automatic machine learning frameworks that are good matches for the more recently introduced frameworks and the older frameworks. For all the automatic machine learning frameworks, we used the same preprocessing methods to process the data as a raw input, then, we adjust the parameters of each framework in order to find the most suitable parameters, and finally we output the model. In general, we consider automatic machine learning frameworks to be black boxes and do not examine frame-specific methods for automatically optimizing parameters and integrating the model for direct output. Automated machine learning frameworks automatically optimize models, thereby reducing the time and effort devoted by researchers and, to a certain extent, allowing non-experts in machine learning to solve bioinformatics problems. Utilizing high-quality features is one way to improve performance in machine learning. It is necessary to find features that are suitable for ncRNA identification in the study because providing or discovering good features is one of the most important tasks in machine learning. We extracted k-mer frequency features, coding sequence features, and other features during our experiments. Despite the fact that traditional k-mer features have been used in a variety of studies, such as gene identification [33], subcellular localization [34], and sequence analysis, it has been demonstrated that the k-mer frequency is highly effective at detecting ncRNAs [35]. Many tools have also used features related to coding sequences and some other features [36]. Ninety-one extracted features were filtered using our feature selection method; the filtered features successfully identified ncRNAs and it was the most precise tool, to date, for ncRNA identification in all plant species. For ncRNA identification, there are additional factors to consider, such as the trade-off between sensitivity and specificity. At present, the number of ncRNAs is small as compared with the number of coding RNAs identified. To prevent ncRNAs from being missed, high sensitivity is important. Currently, CPAT, CNCI, CPPred, and CPC2 are less sensitive and focus more on identifying coding RNAs, but this requires an additional step to screen for non-coding RNAs. In contrast, the high sensitivity of PINC reduces the necessity for additional filtering processes. Moreover, PINC demonstrated a higher rate of accuracy than any other tool among the nine plants evaluated. Although some tools for non-coding RNA identification have reached over 85 percent accuracy, increased accuracy is not meaningless, as large amounts of data have become available due to advances in sequencing technology, and it is possible that for every one percent increase, hundreds of additional correct RNAs can be identified. Here, PINC achieves a high degree of ncRNA identification precision. This may be because the model in PINC adopts the stacking strategy, while other tools use single models such as SVM, logistic regression, and xgboost. For a long time, the performance of combining the predicted results of multiple models has been better than that of a single model, and the variance has been significantly reduced [37]. In the experiment, we selected the default basic model in the AutoGluon framework. Here, the basic model is trained separately, and then the prediction of the basic model is used as a feature to train the stacked model. Stacked models can improve the shortcomings of a single-model prediction and can take advantage of their interactions to improve the prediction ability [38]. In addition, it can be seen from the feature level distribution map described earlier that these features also have strong discrimination ability. In addition, we plan to continue research in two areas: first, deep learning, which can automatically extract features, reduce the time required to extract features, and can improve the accuracy of cross-species recognition. In contrast, we should consider machine learning techniques to gain a deeper understanding of these RNA types and to investigate their biological significance. In addition, for plants, only a handful of ncRNA functions have been identified; once these functions are identified, new mechanisms can be explored and new features can be added to PINC to improve our tool further.
Figure 7 depicts the tool’s overall workflow, which consists of three steps: (1) dataset construction, (2) feature extraction and selection, and (3) model construction. To create the dataset, RNA sequences were obtained from the GreeNC, CANTATA, RNAcentral, and Phytozome databases. Secondly, feature selection methods were used to extract and filter features. Finally, machine learning models were compared to determine the most effective model for ncRNA identification.
To construct the experimental dataset, we considered two factors. On the one hand, the diversity of plants and the abundance of annotation data were taken into consideration. On the other hand, considering the balance of the data, we chose four plants as our training and validation datasets (Table 4), which included two model plants, i.e., Arabidopsis thaliana and Oryza sativa, in addition to two non-model plants, i.e., Glycine max and Vitis vinifera. We used ncRNAs as the positive sample data and coding RNAs as the negative sample data in the dataset. Negative samples were obtained from Phytozome.v13 [39]. Positive samples were obtained from three public databases, including GreeNC [40], CANTATA [41], and RNAcentral [42]. For all data, first, we used cd-hit-est-2D in the CD-hit tool [43] to eliminate redundant sequences between the test and training sets at a threshold of 80% [22,35,44,45]. Second, in order to balance the datasets, random selections of 4000 data were made for each plant, of which 2000 were positive samples and 2000 were negative samples. The positive sample data consisted of 1800 lncRNAs and 200 sncRNAs, and the negative sample data consisted of 2000 mRNAs (Table 5) [18,46]. Thus, the baseline dataset consisted of a total of 16,000 protein sequences from four plants. Meanwhile, we analyzed the length distribution of the positive and negative datasets, as shown in Figure 8. The median length of the coding RNAs data was 1029 and the data were mostly concentrated in the range of 0–2000. The ncRNA data had a median length of 321 and the data were mostly concentrated in the range of 0–1000. Finally, we proportionally divided the dataset into 70% training data and 30% validation data. Additionally, nine independent test sets were created for nine plants. (Table 6): Cicer arietinum, Gossypium darwinii, Lactuca sativa, Manihot esculenta, Musa acuminata, Nymphaea colorata, Solanum tuberosum, Sorghum bicolor, and Zea mays. To eliminate redundant sequences, the data for these nine independent test sets were taken from the four databases mentioned above and filtered at a threshold of 80%.
This experiment initially extracted 91 features (Table 7). The 86 features of k-mer frequency, sequence length, and GC content were obtained using the Python script program (https://github.com/midisec/PINC, accessed on 22 August 2022); the five features of Score and CDS were obtained using the UCSC Genome txCdsPredict program in the browser (http://hgdown-load.soe.ucsc.edu/admin/jksrc.zip, accessed on 11 November 2014) [47]. These features can be classified into three categories: k-mer frequency features, CDS-related features, and other features. The k-mer frequency describes all possible frequencies for the presence of k nucleotides in a sequence, based on methods that have initially been implemented in whole genome shotgun assemblers. When k = 1, each nucleotide can contain a maximum of four A, C, G, or T. When k equals 2, the calculation involves the dinucleotide frequency (i.e., AA, AT, AG, AC, …, TT) and a total of = 16 species. When k = 3, the calculated three-nucleotide frequencies (i.e., AAA, AAT, AAG, AAC, …, TTT) are computed for a total of = 64 species. By combining 1–3-mer frequencies for a total of 84 features, k-mer frequencies can capture rich statistical information about negative profiles in plant genomes, according to some research [48]. CDS is the result of encoded proteins that are interchangeable with ORF in some ways, but differ slightly [49]. The features Score, cdsStarts, cdsStop, cdsSize, and cdsPercent comprise the second major category of features. Score is the predicted protein score; if it is >800, there is a 90% chance that it is a protein, and if it is >1000, it is virtually certain that it is a protein. cdsStop is the end of the coding region in the transcript, cdsSize is cdsStop minus cdsStart, and cdsPercent is the ratio of cdsSize to the total sequence length. Other features include sequence length and GC content, which are widely used for ncRNA identification. Sequence length indicates the total length of the sequence. GC content is the ratio of guanine and cytosine to the other four DNA bases. There may be redundant features among the 91 features listed above; therefore, we employed feature selection to filter them. For the feature selection method, redundant features were filtered out using a combination of variance threshold filtering and the F-test. Variance threshold filtering is used to filter features based on their own variance. The smaller a feature’s variance, the less significant its variation, and these insignificant features are eliminated. F-test is a method to determine the relationship between each feature and label. The GC content, Score, cdsStop, cdsSize, and T, C, GT, GC, ACG, and TAT frequencies were among the 91 features identified by this combined feature selection method. Finally, these 10 features were used as the model input.
Machine learning (ML) is currently utilized in a variety of fields to solve numerous difficult problems. Nevertheless, model construction for machine learning requires human intervention. Manual intervention is required in the feature extraction, model selection, and parameter adjustment processes, which require professionals to optimize and can waste a significant amount of time and resources if errors occur. To reduce these repetitive development costs, the concept of automating the entire machine learning process, automatic machine learning, has been conceived (AutoML). The definition of AutoML is that it is a combination of automation and ML [50]. From an automation standpoint, AutoML can be viewed as the design of a framework to automate the entire machine learning process, allowing models to automatically learn the correct parameters and configurations without manual intervention. From the standpoint of machine learning, AutoML is a system that is highly capable of learning and generalizing given data and tasks. Recent research on AutoML has focused on the neural network architecture search (NAS) method, which employs a search strategy to test and evaluate a large number of architectures in a search space, and then selects the one that best meets the objectives of a given problem by maximizing the adaptation function. However, the NAS faces two obstacles to the method: first, the amount of computation is excessive, resulting in increased resource consumption. Second, instability may vary each time and the search structure is altered, resulting in varying precision. In our experiments, we compared four automatic machine learning frameworks, AutoGluon, H2O, TPOT, and Autokeras, with three conventional machine learning models, SVM, RF, and Naive Bayes. We determined that AutoGluon was the superior framework, and therefore it was used as the classifier. AutoGluon contains 26 base models including random forest, XGBoost, and a neural network, and in our experiments, we used all the base models for training the model [51]. AutoGluon is an open-source machine learning training framework for tabular data. It is a framework that attempts to avoid a hyperparametric search as much as possible, training multiple models concurrently and weighting them using a multi-layer stacking strategy to obtain the final output.
Several widely used performance metrics were evaluated in the experiments, including sensitivity (SE), specificity (SPC), accuracy (ACC), F1-score, positive predictive value (PPV), negative predictive value (NPV), and the Matthews correlation coefficient (MCC). To evaluate the performance of the classifier numerically and visually, the area under the curve (AUC) and ROC curves were also used. These definitions are as follows: TP represents true positives, the number of correctly identified positive samples, while FN, TN, and FP represent false negatives, true negatives, and false positives, the number of incorrectly identified positive samples, correctly identified negative samples, and incorrectly identified negative samples, respectively.
Various tools have been developed to distinguish between ncRNAs and coding RNAs, the majority of which have used scientific computational methods to differentiate sequences and to accelerate the annotation of various human genes. In addition to nucleotides with high discriminatory power in 1–3-mer, we also extracted other features such as the sequence’s definition, composition, and function. Moreover, we combined F-test and variance threshold filtering and found that the combined method was superior to the individual methods of F-test and variance threshold filtering. A number of automated machine learning and traditional machine learning frameworks were also used for modeling, in which the validation set was carefully evaluated and analyzed, including the use of cross-validation on the validation set available, with AutoGluon performing the best. Then, we compiled these into a tool called PINC and compared it to nine other tools on nine test sets, demonstrating that PINC performed better than other tools on all of these species. For user convenience, a user-friendly web (http://www.pncrna.com/, accessed on 22 August 2022) has been developed, where the output can be obtained simply by entering a FASTA sequence or file. Overall, PINC has excellent predictive properties, permits cross-species plant identification, and is a practical and user-friendly tool. | true | true | true |
PMC9570156 | 36232579 | Naoufal Lakhssassi,Dounya Knizia,Abdelhalim El Baze,Aicha Lakhssassi,Jonas Meksem,Khalid Meksem | Proteomic, Transcriptomic, Mutational, and Functional Assays Reveal the Involvement of Both THF and PLP Sites at the GmSHMT08 in Resistance to Soybean Cyst Nematode | 24-09-2022 | PLP,THF,SCN resistance,SHMT,soybean,mutational analysis,composite hairy root transformation | The serine hydroxymethyltransferase (SHMT; E.C. 2.1.2.1) is involved in the interconversion of serine/glycine and tetrahydrofolate (THF)/5,10-methylene THF, playing a key role in one-carbon metabolism, the de novo purine pathway, cellular methylation reactions, redox homeostasis maintenance, and methionine and thymidylate synthesis. GmSHMT08 is the soybean gene underlying soybean cyst nematode (SCN) resistance at the Rhg4 locus. GmSHMT08 protein contains four tetrahydrofolate (THF) cofactor binding sites (L129, L135, F284, N374) and six pyridoxal phosphate (PLP) cofactor binding/catalysis sites (Y59, G106, G107, H134, S190A, H218). In the current study, proteomic analysis of a data set of protein complex immunoprecipitated using GmSHMT08 antibodies under SCN infected soybean roots reveals the presence of enriched pathways that mainly use glycine/serine as a substrate (glyoxylate cycle, redox homeostasis, glycolysis, and heme biosynthesis). Root and leaf transcriptomic analysis of differentially expressed genes under SCN infection supported the proteomic data, pointing directly to the involvement of the interconversion reaction carried out by the serine hydroxymethyltransferase enzyme. Direct site mutagenesis revealed that all mutated THF and PLP sites at the GmSHMT08 resulted in increased SCN resistance. We have shown the involvement of PLP sites in SCN resistance. Specially, the effect of the two Y59 and S190 PLP sites was more drastic than the tested THF sites. This unprecedented finding will help us to identify the biological outcomes of THF and PLP residues at the GmSHMT08 and to understand SCN resistance mechanisms. | Proteomic, Transcriptomic, Mutational, and Functional Assays Reveal the Involvement of Both THF and PLP Sites at the GmSHMT08 in Resistance to Soybean Cyst Nematode
The serine hydroxymethyltransferase (SHMT; E.C. 2.1.2.1) is involved in the interconversion of serine/glycine and tetrahydrofolate (THF)/5,10-methylene THF, playing a key role in one-carbon metabolism, the de novo purine pathway, cellular methylation reactions, redox homeostasis maintenance, and methionine and thymidylate synthesis. GmSHMT08 is the soybean gene underlying soybean cyst nematode (SCN) resistance at the Rhg4 locus. GmSHMT08 protein contains four tetrahydrofolate (THF) cofactor binding sites (L129, L135, F284, N374) and six pyridoxal phosphate (PLP) cofactor binding/catalysis sites (Y59, G106, G107, H134, S190A, H218). In the current study, proteomic analysis of a data set of protein complex immunoprecipitated using GmSHMT08 antibodies under SCN infected soybean roots reveals the presence of enriched pathways that mainly use glycine/serine as a substrate (glyoxylate cycle, redox homeostasis, glycolysis, and heme biosynthesis). Root and leaf transcriptomic analysis of differentially expressed genes under SCN infection supported the proteomic data, pointing directly to the involvement of the interconversion reaction carried out by the serine hydroxymethyltransferase enzyme. Direct site mutagenesis revealed that all mutated THF and PLP sites at the GmSHMT08 resulted in increased SCN resistance. We have shown the involvement of PLP sites in SCN resistance. Specially, the effect of the two Y59 and S190 PLP sites was more drastic than the tested THF sites. This unprecedented finding will help us to identify the biological outcomes of THF and PLP residues at the GmSHMT08 and to understand SCN resistance mechanisms.
Soybeans are the largest source of proteins and the second-largest source of oil worldwide. The production value of soybeans in the United States amounted to USD 46.06 billion in 2020 [1]. Soybean production, however, is affected by the presence of a microscopic parasitic roundworm, soybean cyst mematode (SCN), which contributes dramatically to increased yield loss in soybean crops nationwide, causing an estimated USD 1.5 billion in damage [2]. Emerging SCN populations have adapted to the resistance found in certain varieties of soybean, rendering the plant susceptible to infection. In fact, more than 95% of cultivated soybeans in the U.S. use SCN-resistant varieties based on the PI 88788 source of resistance, and 3% of varieties carry resistance from Peking. Due to SCN adaptation, a reduction in the effectiveness of resistant cultivars is taking place [3]. The shift in virulence of the pathogen resulted in 80% of fields in Midwest having SCN that can reproduce on PI 88788 [4]. Peking-type of resistance presents a sustainable alternative to breed for soybean lines with broad resistance to SCN. Cloning novel genes and understanding Peking-type resistance is essential and key to creating soybean varieties. Peking-type reaction has been reported to be bigenic, requiring both the rhg1 and the Rhg4 loci [5]. The gene underlying resistance to SCN at the Rhg4 locus, the GmSHMT08, has been identified and functionally characterized [2,6,7]. Additionally, copy numbers of the Rhg4 were shown to play an essential role in broad resistance to SCN against five nematode races [8]. Although the soybean genome encodes at least 13 GmSHMT members, only the cytosolic GmSHMT08c was shown to play a role in SCN resistance, with the absence of functional redundancy by the other GmSHMT members, including the other cytosol-targeted GmSHMT05, the four nucleic-targeted GmSHMTs, the two plastidial-targeted GmSHMTs, and the five mitochondrial-targeted GmSHMTs [7]. The serine hydroxymethyltransferase (SHMT) is commonly present in plant and animal species. SHMT plays essential role in methionine synthesis, one-carbon metabolism, and the maintenance of redox homeostasis during photorespiration [9,10,11,12]. The SHMT is involved in the interconversion of serine/glycine and tetrahydrofolate (THF)/5,10-methyleneTHF through a transaldimination reaction [9]. The enzymatic co-factor THF is involved in the biosynthesis of various biologically important molecules including purine and pyrimidine nucleotides [13]. On the other hand, PLP acts as a coenzyme in all transamination reactions and in certain decarboxylation, deamination, and racemization reactions of amino acids. PLP, the active form of vitamin B6, is required for hundreds of different reactions in human metabolism, primarily for the synthesis of amino acids and amino acid metabolites and for the synthesis and/or catabolism of certain neurotransmitters and degradation pathways [14]. The SHMT enzyme is therefore essential to directing one-carbon units to the folate-mediated one-carbon metabolism that is required for nucleotide biosynthesis, methyl group biogenesis, and vitamin and amino acid metabolism during glycine biosynthesis [15]. During serine biosynthesis, SHMT plays a major role in the photorespiration metabolic reaction and is therefore essential for C3 plants. Through the glyoxylate cycle, SHMT plays a role in the maintenance of redox homeostasis, involving the gluthatione synthase, peroxidases, and other related genes. It is known that mutations at the mitochondrial AtSHMT1 cause a photorespiratory deficiency in the plant model Arabidopsis thaliana [16]. Mutations in the human SHMT protein were shown to cause cancers and cardiovascular diseases [17,18,19]. SCN resistance in Forrest is derived from Peking (PI 548402) and is considered to be a promising cultivar that confers resistance to SCN that differs from SCN resistance in PI 88788. Two naturally occurring mutations, P130R and N358Y, distinguish the Forrest GmSHMT08 allele from the susceptible soybean alleles contained in Essex and Williams 82 [2]. The GmSHMT enzyme contains several PLP and THF binding and catalysis sites that are essential to carrying out the transaldimination reaction [7]. Recently, the two Forrest-specific polymorphic substitutions (P130R and N358Y) that differ from the susceptible Essex have been reported to impact the mobility of a loop near the entrance of the (6S)-tetrahydrofolate binding site [20]. Both ligand binding and kinetic studies indicate a severe reduced affinity for folate, which dramatically impaired enzyme activity in Forrest GmSHMT08 [20]. In the current study, we performed proteomic analysis of a set of protein complexes that was immunoprecipitated using GmSHMT08 antibodies under SCN infected soybean roots. Although soybean cyst nematodes infect soybean roots, leaves play an important role by supplementing nematodes with most of the nutrients that they use to grow and complete their life cycle. In the current study, root and leaf transcriptomic analysis of differentially expressed genes under SCN infection supported the data from LC-MS. In fact, integration of proteomic and transcriptomic data pointed to the involvement of several proteins that belong mainly to pathways that use glycine/serine as a substrate/precursor. Therefore, the obtained data pointed to the involvement of the interconversion reaction carried out by the serine hydroxymethyltransferase protein. Most importantly, site-directed mutagenesis combined with composite hairy root transformations in addition to mutational analysis of the previously identified 18 EMS Gmshmt08 Tilling mutants derived from Forrest [2,6,7] uncovered the impact of the four THF cofactor binding sites, the four PLP cofactor binding sites, and the two PLP cofactor catalysis sites at the GmSHMT08 protein on SCN resistance. This study reveals for the first time the large effect of both PLP cofactor binding and PLP cofactor catalysis sites on SCN resistance when compared to THF cofactor binding sites.
To identify components related to the SCN resistance mechanism, we analyzed the mass spectrometry data from the immunoprecipitated protein complex that was conducted using anti-GmSHMT08 antibodies immobilized to beads in a chromatography column. Several peptides related to SCN infection were present when comparing non-infected and SCN-infected root-eluted fractions. Under non-SCN-infected conditions, only the GmSHMT08 was present in the analyzed proteomic root fractions of Forrest and Essex soybean (Supplemental Table S1). For SCN-infected soybean roots, mass spectrometry analysis showed that the obtained fragmented peptides belong to 37 proteins in the resistant genotype “Forrest” only, while three proteins were identified in the susceptible genotype “Essex” only (Figure 1). Additionally, 23 proteins were common between Forrest and Essex (Figure 1). In addition to translation, growth, cell differentiation, response to stress, flower development, and carbohydrate metabolic processes that were found in both cultivars, many other additional categories (biological processes), such as cellular metabolic process, transport, biosynthetic process, and signal transduction, were contained in the resistant cultivar only (Figure 1). When comparing the biological processes that differentiate the SCN resistance reaction from susceptibility, two categories were mainly dominant in the susceptible cultivar Essex. A total of 75% of the genes that correlate with the presence of the nematode were linked to translation versus 26% in the resistance line. Surprisingly, the growth process occupied 25% in Essex versus 5% in Forrest. The presence of these two main processes in Essex is coherent with the development and growth of plant root cells in the susceptible lines to form syncytial feeding structures. In silico analysis of the fragmented peptides obtained from the LC-MS analysis (Supplemental Table S1) identified 62 genes that belong to the 37 proteins that were identified in Forrest only, 3 genes that belong to the 3 proteins that were identified in Essex only, and 74 candidate genes that belong to the 23 members that were common between Essex and Forrest (Figure 2A). As expected, mass spectrometry analysis revealed the presence of the GmSHMT08, GmSNAP18, and GmPR08-BetVI proteins, which is coherent with the previously reported physically interacting GmSHMT08/GmSNAP18/GmPR08-BetVI protein complex in resistance to SCN [21]. Interestingly, we were able to identify several proteins (i.e., glycine decarboxylase (GLDC), glycine dehydrogenase (GlyDH), serine/glycine hydroxymethyltransferase (SHMT/GHMT), etc.) that belong mainly to cycles that use glycine/serine as substrate/precursor and therefore that are directly related to the interconversion reaction carried out by the serine hydroxymethyltransferase protein (Supplemental Table S2).
Although a nematode interacts mostly during its life cycle with soybean roots, most of the nutrients that the nematode uses to grow and complete its life cycle are transported from the leaves. Therefore, to understand and identify the biological pathways that are linked to the presence of nematodes, the current study explored the expression of genes in both leaves and roots in response to SCN infections. Using an integrated approach combining the mass spectrometry data of 136 genes (identified in Essex MS and Forrest MS infected roots) and RNAseq data of 1538 DEG (SCN-infected Forrest root) and 8282 DEG (SCN-infected Forrest leaves), we identified several genes that were differentially expressed under SCN infection in soybean roots (Figure 2C). Transcripts of these identified genes were induced up to 3.9 Log2FoldChange in the SCN-infected soybean roots and up to 10.42 Log2FoldChange in the SCN-infected soybean leaves. Transcripts were downregulated up to 5.92 Log2FoldChange in the SCN-infected soybean roots and up to 11.41 Log2FoldChange in the SCN-infected soybean leaves (Figure 2D). These data reveal the importance of soybean leaves during root SCN infections in the resistant reaction. Of 136 genes identified in Forrest MS, 78 were differentially expressed under SCN infection in the resistant Forrest, including root and leave tissues. Out of the 78 genes, 56 were differentially expressed in Forrest leaves, 8 were differentially expressed in Forrest roots, and 14 were differentially expressed in both Forrest roots and leaves. Most of the identified genes belong mainly to 30 gene families, of which most were found to be related to redox hemeostasis, serine/glycine conversion, glyoxylate cycle, glycolysis, succinyl-CoA, heme biosynthesis related enzymes, cytoskeleton-related enzymes, and ATP mitochondrial related genes (Supplemental Table S2, Figure 3).
Most of the identified genes were mapped to QTLs for resistance to SCN using different mapping populations. In fact, 34 genes were located within reported SCN QTLs; 21, 4, and 2 genes were located ~3Mbp, ~4–6 Mbp, and ~11Mbp away from reported SCN QTLs, respectively (Supplemental Table S2). The most reported gene that mapped to QTLs for resistance to SCN is the glutamine synthase (Glyma.18G041100) gene after the GmSHMT08 at the Rhg4 locus. The sucrose synthase 1 (Glyma.15G182600) was reported frequently in SCN QTL mapping analysis, followed by the 6-phosphogluconate dehydrogenase (Glyma.08G254500) and the mitochondrial GmSHMT08m (Glyma.08G274400), showing the contribution of the glycolysis cycle to SCN resistance (Supplemental Table S2). Within the SCN QTLs, we were able to identify high frequency genes that belong to glycolysis (30 SCN QTLs), followed by cytoskeleton-related genes (18 SCN QTLs), glyoxylate cycle (13 SCN QTLs), redox hemeostasis (9 SCN QTLs), and ATP-mitochondrial-related genes (8 SCN QTLs) (Supplemental Table S2).
The current study revealed many candidate genes related to redox hemostasis (Supplemental Table S2, Figure 3). The annotation of the set containing these genes showed the presence of four genes that encode glutathione S-transferase (GST), three genes encoding glutathione peroxidases, two genes encoding NAD(P)H dehydrogenase, and two genes encoding glutamate dehydrogenases (Supplemental Table S2). The obtained mass spectrometry data are coherent with previous studies showing that modulation of the SHMT serine/glycine interconversion impact important maintenance of redox homeostasis occurs via both glutathione synthase and glutathione peroxidases [23]. RNAseq analysis showed that transcripts from the previous 11 genes were significantly induced under SCN infection in both root and leaves (Figure 2E, Supplemental Figure S1A).
Glycolysis cycle provides several products that support nematodes growth. This study showed several candidate genes related to the glycolysis cycle including four sucrose synthase 1-related genes, three glyceraldehyde 3-phosphate dehydrogenase, three enolases, two glutamine synthetases, two 6-phosphogluconate dehydrogenases, 1 fructose-bisphosphate aldolase, and one glutamyl-tRNAGlu reductase (Supplemental Table S2). Transcripts from all these 16 genes were induced under SCN infection in both root and leaves (Figure 2E, Supplemental Figure S1B).
Glyoxylate products support early nematode development. Mass spectrometry analysis revealed the presence of many candidate genes related to the glyoxylate cycle. The annotation of the set containing these genes showed the presence of five malate dehydrogenases (MDH1), five NADP-dependent malic enzymes, four glycine dehydrogenases (GLDC, gcvP), and two NADPH-specific isocitrate dehydrogenases (Supplemental Table S2). Transcripts from all these 16 genes were induced under SCN infection in both root and leaves (Figure 2E, Supplemental Figure S1C).
Several genes related to the serine and glycine synthesis were identified by mass spectrometry including six serine hydroxymethyltransferases (GmSHMT08c, GmSHMT02m, GmSHMT08m, GmSHMT09m, GmSHMT14m, and GmSHMT18m), three glycine hydroxymethyltransferases, and one glycine decarboxylase. Additionally, mass spectrometry showed the presence of two dihydrolipoyllysine-residue succinyltransferases and two methylmalonate-semialdehyde dehydrogenases (Supplemental Table S2). Dihydrolipoyllysine-residue succinyltransferase and Methylmalonate-semialdehyde dehydrogenase are key enzyme to synthesize Succinyl-CoA that—together with glycine, the SHMT product—produce the ALA, an important component of and precursor to the production of heme. It is well known that SCN requires heme source for its survival. RNAseq data showed that transcripts from all previous 12 enzymes were induced under SCN infections (Figure 2E, Supplemental Figure S1D).
Several components of cytoskeleton-related genes were found, including two actin, two actin-7, one tubulin-A, and nine tubulin beta-4 in addition to several ATP- and ADP-mitochondrial-related genes, such as two ADP/ATP carrier protein 1, an ATP synthase subunit beta-1, an ATP synthase subunit alpha, and two ADP-ribosylation factor 2-B-like genes that regulate the interaction of tubulin-folding cofactor D with native tubulin (Supplemental Table S2). Interestingly, transcripts from all the 21 genes were induced under SCN infection in both root and leaves (Figure 2E, Supplemental Figure S1E,F).
Mass spectrometry analysis pointed to the importance of the GmSHMT08 in the interconversion of serine and glycine, two substrates that are essential for redox hemeostasis, serine/glycine conversion, glyoxylate cycle, glycolysis, succinyl-CoA, and heme biosynthesis (Figure 3). This interconversion relies on two essential sites at the GmSHMT08 enzyme, the PLP and the THF cofactor sites. To carry out the GmSHMT08 protein homology modeling, an available SHMT crystal structure from of serine hydroxymethyltransferase from glycine max cultivar Essex already complexed with PLP-glycine and 5-formyltetrahydrofolate residues was used as the template (Figure 4). Next, all four THF cofactor and six PLP cofactor sites at the GmSHMT08 were mapped against the model (Figure 4). To visualize the effect of the site directed mutagenesis on the THF and PLP residues, rotamers tools have been used to mutate the four THF and 6 PLP residues on the GmSHMT08 protein model (Figure 5). The PLP molecule binds to different residues in the GmSHMT08 PLP binding pocket, as shown in Figure 4. Lys-244 forms a covalent Schiff base linkage (internal aldimine) with PLP (Figure 4). Nearby residues from both chains of the obligate dimer (Tyr-59′, Glu-61′, Ser-107, Asp-215, Thr-241, Arg-250 (prime indicates chain B) (Figure 4)) assure conserved interactions with the phosphate, N1, and O3 hydroxyl of PLP, whereas the pyridine ring of PLP stacks against His-134. The SHMT8-PLP-Gly complex represents an intermediate step of the THF-dependent catalytic mechanism, in which L-Ser attacks the Schiff base linkage between Lys-244 (Figure 4) and PLP to form a PLP-Ser external aldimine. Formaldehyde is next liberated when the active site general base deprotonates the hydroxyl side chain of L-Ser. Once synthesized, formaldehyde is next attacked by THF N5, transferring the side chain of L-Ser to THF, resulting in an external aldimine/quinonoid product called PLP-Gly. In silico analysis revealed that the mutated GmSHMT08ΔL129A, GmSHMT08ΔL135A, GmSHMT08ΔF284A, GmSHMT08ΔN374A, GmSHMT08ΔY59A, GmSHMT08ΔG106A, GmSHMT08ΔG107A, GmSHMT08ΔH134A, GmSHMT08ΔS190A, and GmSHMT08ΔH218A alleles and the two polymorphisms between the SCN resistant cultivar Forrest and the SCN susceptible Essex are predicted to impact negatively their conserved interactions with the phosphate, N1, and O3 hydroxyl of PLP near the PLP binding pocket (Figure 5). These mutations are expected to affect the GmSHMT08’s ability to bind PLP substrate and the interconversion of serine/glycine and tetrahydrofolate (THF)/5,10-methylene THF, which may impact resistance to SCN.
To gain more insight into the impact of the eighteen EMS Gmshmt08 mutants identified earlier [2,6,7] on the PLP/THF cofactor binding and catalysis, all previous mutants were mapped and mutated using the rotamers tool. E61K and G71D are found close to the Tyr59 residue, that is required for PLP cofactor catalysis. E61K, G62S, and P285S are located very close to the Phe284 that is required for THF binding. Disruption of this site profoundly altered substrate binding and catalytic activity in E. coli [24]. Therefore, Gmshmt08E61K, Gmshmt08G62S, and Gmshmt08p285S mutations are likely to have the same conformational deficiency by impacting the THF cofactor binding and PLP cofactor catalysis at the GmSHMT08. The Forrest polymorphism R130P and Gmshmt08G132D mutant were located close to the Leu129 residue that is required for THF binding. Gmshmt08G132D is also close to the two essential and conserved histidine residues: His134 and His137. Since proline has a conformational rigidity due to its direct incorporation of the α-carbon into its side chain, this may cause drastic conformational changes, interfering with this catalysis (Figure 6). Both polymorphic substitutions between Essex and Forrest (P130R and N358Y) were shown to impact the mobility of a loop near the entrance of the THF binding site at the GmSHMT08 protein, resulting in reduced affinity for folate substrate, subsequently impairing the enzymatic activity of GmSHMT08 [20]. Gmshmt08G357R mutation is located one residue away from the Forrest polymorphic substitution N358Y and therefore is predicted to impact the THF site’s binding to folate. Another mutation, Gmshmt08Q226*, which resulted in a loss of SCN resistance in Forrest, was mapped close to His218 residues involved in PLP cofactor binding. Gmshmt08G106S is located at the PLP cofactor binding site Gly106, in addition to being located very close to the other PLP cofactor binding site Ser107. Gmshmt08G106S, Gmshmt08A302V, and Gmshmt08L299F, were closely located to the Thr-241 and Arg-250 where the pyridine ring of PLP stacks against His-134 (PLP catalysis). In addition, the previous mutations were mapped close to the Lys-244 that forms a covalent Schiff base linkage (internal aldimine) with PLP.
The four THF cofactor binding sites (L129, L135, F284, N374) at the GmSHMT08 protein contribute to the interconversion of tetrahydrofolate (THF) and 5,10-methylene THF [7]. To test the effect of each THF binding sites on SCN resistance, we introduced independent mutations in each one of the four THF-related residues at the GmSHMT08 coding sequence from the resistant Forrest allele, then overexpressed it in the ExF12 RIL, carrying the SCN-resistant GmSNAP18+ from Forrest and the SCN-susceptible GmSHMT08− allele from Essex. To conduct the GmSHMT08ΔL129A, GmSHMT08ΔL135A, GmSHMT08ΔF284A, GmSHMT08ΔN374A, and GmSHMT08 overexpression analysis, the 1416-bp nucleotide coding sequence of the different GmSHMT08 alleles were overexpressed under the control of a soybean ubiquitin promoter using a transgenic hairy root system. Interestingly, unlike the GmSHMT08 wild-type allele from Forrest that reduced the cyst numbers of the ExF12 RILs by 91% in the susceptible ExF12 background, reductions of the cyst numbers at the induced mutations, including GmSHMT08ΔL129A, GmSHMT08ΔL135A, GmSHMT08ΔN374A, and GmSHMT08ΔF284A, were limited to 50%, 58%, 66%, and 78%, respectively (Figure 7). Thus, induced mutations at GmSHMT08ΔL129A, GmSHMT08ΔL135A, GmSHMT08ΔN374A, and GmSHMT08ΔF284A affected the GmSHMT08’s ability to reduce the number of cysts by 40%, 32%, 24%, and 12%, respectively. Statistical analysis showed that of the four THF sites, site directed mutagenesis of the GmSHMT08ΔL129A, GmSHMT08ΔL135A, and GmSHMT08ΔF284A were significantly different (p < 0.0001) from the GmSHMT08 wild-type allele. The mutation at the GmSHMT08ΔN374A THF residue presented the lowest reduction of cyst numbers.
Four PLP cofactor binding sites (G106, G107, S190A, H218) and two PLP cofactor catalysis sites (Y59 and H134) at the GmSHMT08 protein are involved in the interconversion of serine and glycine [7]. To test the real effect of each PLP cofactor binding and catalysis sites on SCN resistance, we introduced independent mutations in each one of the six PLP related residues at the GmSHMT08 coding sequence from the resistant Forrest allele, then overexpressed it in the ExF12 RIL. To conduct the GmSHMT08ΔY59A, GmSHMT08ΔG106A,107A, GmSHMT08ΔH134A, GmSHMT08ΔS190A, GmSHMT08ΔH218A and GmSHMT08 overexpression analysis, the 1416-bp nucleotide coding sequences of the different GmSHMT08 alleles were overexpressed under the control of a soybean ubiquitin promoter using a transgenic hairy root system. Surprisingly, unlike the GmSHMT08 wild-type allele from Forrest that reduced the cyst numbers of the ExF12 RILs by 91% in the SCN-susceptible ExF12 background, reductions of the cyst number at the induced mutations, including GmSHMT08ΔS190A, GmSHMT08ΔY59A, GmSHMT08ΔG106A,107A, GmSHMT08ΔH218A, and GmSHMT08ΔH134A were limited to 6%, 42%, 61%, 61%, and 63%, respectively (Figure 7). Thus, induced mutations at GmSHMT08ΔS190A, GmSHMT08ΔY59A, GmSHMT08ΔG106A,107A, GmSHMT08ΔH218A, and GmSHMT08ΔH134A affected the GmSHMT08’s ability to reduce the number of cysts by more than 84%, 48%, 29%, 29%, and 27%, respectively. Most importantly, the PLP cofactor binding site (S190A) and the PLP cofactor catalysis site (Y59) presented higher impact on SCN resistance when compared to the four THF binding sites that were tested previously.
Soybean cyst nematode is the most destructive pathogen to soybeans [3]. Most of the efforts to understand the SCN resistance mechanism were focused on deciphering the genes for resistance to SCN within two types of SCN resistance: the PI88788, which uses the rhg-1b; and Peking-type resistance, which uses a combination of rhg1-a and Rhg4 loci [2,6]. Although the gene that confers resistance to SCN at the Rhg4 locus has been cloned and identified a decade ago [2]; the involvement of the four THF cofactor binding sites, four PLP cofactor binding sites, two PLP cofactor catalysis sites at the GmSHMT08 in resistance to SCN has not been revealed yet. The current study revealed the presence of enriched cycles that mainly use glycine as a substrate, including the glyoxylate cycle, redox homeostasis, and heme biosynthesis. Several key enzymes involved in the glycolysis cycle were also identified, which is coherent with QTL analysis studies that were reported earlier (Supplemental Table S2). Although glucose and glutamine are the main sources that are used to maintain the glycolysis pathway, serine plays an essential role in the glycolysis pathway through de novo serine biosynthesis. Serine derived from a branch of glycolysis can be reintegrated into the glycolysis pathway to synthesize pyruvate but can also be converted to glycine, which provides carbon units for one carbon metabolism. One carbon metabolism is essential for the synthesis of proteins, lipids, nucleic acids, and other precursors through a complex metabolite network based on the chemical reactions of folate compounds. Thus, the main products of SHMT, serine, and glycine provide precursors for the biosynthesis of proteins, nucleic acids, and lipids, which are essential for both host and pathogen growth. The war between nematodes and soybean for metabolites and how plants can fight underground attacks is complex and still requires investigation [25]. Proteomic and transcriptomic assays of SCN-infected soybeans identified seven enzymes that play essential roles in the glycolysis and four enzymes in the glyoxylate cycle. Previous studies reported the ability of nematodes to steal nutrients from host plants [26]. Soluble sugars, such as fructose, glucose, and sucrose, were previously found to increase significantly in tomato leaves and roots during early infection by root-knot nematodes (RKNs) [27], which is coherent with the 7 genes identified at the glycolysis cycle in the current study. Another plant-parasitic nematode, Heterodera schachtii, has been shown to stimulate plant root cells to form syncytial feeding structures which synthesize all nutrients required for successful nematode development [27,28]. During glycolysis, a series of enzymatic reactions will convert sugars, typically sucrose to glucose, fructose, and then to pyruvate [29]. Products derived from glycolysis and glyoxylate cycle support early nematode development [30]. Nematodes will metabolize energy through the standard metabolic pathways, which is reflected by high metabolic activity, elevated sucrose levels, and the formation of starch [31,32]. The root cells affected by nematode attack show altered metabolisms—especially increased allocation of soluble sugars. Sugar importation into syncytia follows the symplasmic path during later stages of development [33,34,35]. On the other hand, it is known that glycine powers the biosynthesis of heme. Since SHMT catalyzes the conversion of serine to glycine, any disruption of the PLP/THF cofactor binding/catalysis sites, as shown in the EMS GmSHMT08 mutants and by site-directed mutagenesis, may negatively impact the production of glycine and therefore the biosynthesis of heme. Heme is considered a major nutrient for nematodes from the plant host. Nematodes such as Rhabditis maupasi, Caenorhabditis elegans, and Heterodera glycines require heme source or any related iron porphyrin for feeding and survival [27,36,37]. This may explain the presence of four heme-related genes that were obtained via LC-MS and their differential expression during SCN infection. Two out of the four identified genes (Glyma.08G066600 and Glyma.07G183600, belonging to the malonate-semialdehyde dehydrogenase gene family), were mapped at two SCN QTLs in previous studies (Supplemental Table S2). Several genes belonging to the redox hemostasis pathway that were identified by LC-MS in the current study, such as glutathione peroxidases, NADP(H) dehydrogenase, glutathione S-transferases (GSTs), and glycine dehydrogenases, were differentially expressed under SCN infection. GSTs catalyze the conjugation of glutathione (GSH) to xenobiotic substrates for detoxification [38,39,40]. GST activity is dependent upon GSH supply from the glutathione synthetase enzyme and the activity of some transporters to remove GSH conjugates from the cell [41,42]. Most of the identified ROS proteins from LC-MS were differentially expressed under SCN infection. This is coherent with previous transcriptomic analysis, in which both glutathione peroxidase and glutathione transferase transcripts, among other ROS-scavenging enzymes, were shown to be significantly modulated under SCN infection (in syncytia) [43]. Recently, mitochondrial OsSHMT and NbSHMT have been demonstrated to play a role in broad-spectrum resistance via the ROS pathway [44]. Cytoskeletons (i.e., microtubules) play an important role during the intracellular transport of mitochondria [45,46]. The interaction of some mitochondrial components with certain cytoskeletal proteins was found to be involved in the coordination of mitochondrial function [47,48]. In fact, interaction between the microtubule-associated C4HC3-type E3 Ligase (MEL) and the mitochondrial SHMT1 leads to SHMT1-dependent mitochondrial ROS generation, activation of MAPK cascades, and reprogramming of defense-related transcripts, ultimately leading to attenuated pathogen invasion. The interacting MEL-SHMT1 complex mediates regulation of plant immunity involving microtubules and mitochondria. Infections by multiple pathogens induce MEL transcription. This is followed by the formation of MEL homodimers, which activate MEL E3 ligase activity, subsequently triggering SHMT1 degradation by the 26S [44]. Cytoskeleton-including actin filaments are dynamic structures that can grow and shrink rapidly via the addition or removal of tubulin proteins. During cellular homeostasis responses, mitochondria organelles are considered the major source for the generation of intracellular ROS by supplying ATP and biosynthetic intermediates for redox, cell death, and energy metabolism [49,50,51,52]. The current study found several potential substrates/components of cytoskeleton including Actins, Tubulin A, Tubulin beta-4, and several ATP and ADP mitochondrial related genes including ADP/ATP carrier protein 1, ATP synthase subunit beta-1, ATP synthase subunit alpha, and ADP-ribosylation factor 2-B-like that regulates the interaction of tubulin-folding cofactor D with native tubulin. Tubulin and actin cytoskeletons have been continuously reported to be implicated in plant defense against pathogenic fungi, oomycetes, and bacteria [52,53,54,55]. We also found the presence of cyclophilin, which are known to be modulated by microtubules [44]. The role of cyclophilin in plant pathogenesis has been reported earlier [56]. The cyclophilin GmCYP1 (Glyma.11G098700) has been suggested to play a role in soybean defense via its interaction with the isoflavonoid regulator GmMYB176 [57], which is known to play major roles in resistance to cyst nematodes in Arabidopsis [58] and in SCN [59]. Microtubule disruption of hematopoietic cells cause a dramatic subcellular redistribution of cyclophilin-A and pin1 from the nucleus to the cytosol and plasma membrane [60]. Another microtubule, MAP65-3 microtubule-associated protein, has been shown to be essential for cytokinesis in somatic cells and also play an important role during nematode-induced giant cell ontogenesis in Arabidopsis [61]. In fact, MAP65-3 is associated with mini cell plates that are required for the formation of a functional nematode feeding cell. In giant cell map65-3 mutants, a defect in mini cell plate formation prevents the development of functional feeding cells, which resulted in the death of the nematode [61]. The identification of several cytoskeleton components from the current study reinforces their involvement in resistance to SCN, which is coherent with QTL SCN analysis where the identified 22 cytoskeleton-related and ATP mitochondrial-related genes were mapped to 26 reported SCN QTLs.
Forrest and Essex soybean cultivars were infected using SCN (HG0), as described earlier [62]. Root and leaf samples from three biological replicates containing five SCN (HG0)-infected and five non-SCN-infected soybeans were washed and frozen in liquid nitrogen three days after infection. Total root proteins from SCN-infected and non-infected soybean “Forrest” and “Essex” cultivars were extracted in a lysis buffer containing 5mM DTT, 1% (v/v) NP40, 1mM sodium molybdate, 1 mM NaF, 1 mM PMSF, 1.5 mM Na3VO4, 100 mM NaCl, 2 mM EDTA, 50 mM Tris–HCl at pH 7.5, 10% (v/v) glycerol, and one tablet from the plant protease and phosphatase inhibitors at 1:100 mL (Thermo Scientific), as previously shown [21]. Coomassie Bradford Protein Assay Kit was used to quantify protein concentrations (Thermo Fisher Scientific, Waltham, MA, USA). For in planta immunoprecipitation analysis, anti-GmSHMT08 polyclonal antibodies [21] were immobilized in a column (Pierce Co-Immunoprecipitation Kit). Then, immunoblot analysis of root protein fraction samples from soybean Forrest and Essex was incubated overnight with the immobilized antibodies. After three washes, the associated proteins were eluted as described by the Pierce Co-Immunoprecipitation Kit. The eluted fraction was then used for mass spectrometry analysis.
Peptide digestion, microsequencing analyses, and protein characterization of the SHMT-associated proteins from non-infected and SCN-infected Forrest and Essex roots 3 DAI were carried out in the Charles W Gehrke Proteomics Center at the University of Missouri-Columbia, as previously shown [21]. The eluted fractions obtained from the immunoprecipitation experiment using anti-GmSHMT08 polyclonal antibodies were briefly subjected to lyophilization. Then, all proteins were subsequently digested with trypsin, resulting in one main fraction representing the three biological replicates. Furthermore, samples were acidified, lyophilized, and re-suspended in 21 µL of a 5% acetonitrile, 0.1% formic acid solution, and peptides were analyzed via LC-MS (18 µL injection), as previously described [63]. Liquid chromatography gradient conditions were carried out as previously shown [63]. The Proxeon Easy nLC HPLC system was attached to an LTQ Orbitrap XL mass spectrometer. BSA was used for quality control on the column. Searches of Swiss-Prot-all species and NCBI-Gmax were conducted using Sorcerer-Sequest.
Four plant soybean tissues were used for RNA-seq, including SCN-infected (3 DAI) soybean root, non-SCN-infected soybean root, SCN-infected (3 DAI) soybean leaves, and non-SCN-infected soybean leaves. Three biological replicates that correspond to three independent experiments where each experiment contained five SCN (HG0) infected and five non-SCN-infected soybean plants were washed and frozen in liquid nitrogen three days after infection. Total RNA for each sample was extracted from 100 mg of frozen grounded samples using RNeasy QIAGEN KIT (Cat. No./ID: 74004, Germantown, Maryland). Total RNA was treated with DNase I (Invitrogen, Carlsbad, CA, USA). RNA-seq libraries preparation and sequencing were performed at Novogene INC. (Cambridge, UK) using Illumina NovaSeq 6000. The four libraries were multiplexed and sequenced in two different lanes generating 20 million raw pair end reads per sample (150 bp). Quality assessment of sequenced reads was performed using fastqc version 0.11.9 [64]. After removing the low-quality reads and adapters with trimmomatic version V0.39 [64], the remaining high-quality reads were mapped to the soybean reference genome Wm82.a2.v1 using STAR, version v2.7.9 [65,66]. Uniquely mapped reads were counted using Python package HTseq v0.13.5 [67]. Read count normalization and differential gene expression analysis were conducted using the Deseq2 package v1.30.1 [68] integrated in the OmicsBox platform from BioBam (Valencia, Spain). DEGs were considered significant if p value < 0.05, Log2FoldChange no less than ±0.6. Expression profiling was visualized through a heatmap using Heatmapper [22].
The GmSHMT08 coding sequence from the Forrest WT (Rhg4) was amplified from soybean Forrest root cDNA via RT-PCR using the GmSHMT08c-AscI-Fw primers (ggcgcgccATGGATCCAGTAAGCGTGTGGGGTA) and the GmSHMT08c-AvrII-Rv primers (ggatccCTAATCCTTGTACTTCATTTCAGATACC) and cloned into the pG2RNAi2 vector under the control of the soybean ubiquitin (GmUbi) promoter [21,23]. Cloning was carried out between AscI and AvrII cloning sites at the pG2RNAi2 vector to generate pG2RNAi2::GmSHMT08, which was used as a positive control (Supplemental Figure S2). All mutations described in this study were introduced using site directed mutagenesis. The following mutated residues, GmSHMT08ΔL129A, GmSHMT08ΔL135A, GmSHMT08ΔF284A, GmSHMT08ΔN374A, GmSHMT08ΔY59A, GmSHMT08ΔG106A, GmSHMT08ΔG107A, GmSHMT08ΔH134A, GmSHMT08ΔS190A, and GmSHMT08ΔH218A, were cloned into the pG2RNAi2 vector to generate pG2RNAi2::GmSHMT08ΔL129A, pG2RNAi2::GmSHMT08ΔL135A, pG2RNAi2::GmSHMT08ΔF284A, pG2RNAi2::GmSHMT08ΔN374A, pG2RNAi2::GmSHMT08ΔY59A, pG2RNAi2::GmSHMT08ΔG106A, pG2RNAi2::GmSHMT08ΔG107A, pG2RNAi2::GmSHMT08ΔH134A, pG2RNAi2::GmSHMT08ΔS190A, and pG2RNAi2::GmSHMT08ΔH218A constructs, respectively (Supplemental Figures S3–S11). All clones were target-sequenced to confirm that the genes and associated mutations were inserted correctly (Supplemental Figures S2–S11).
The ExF12 RIL used for composite hairy root soybean transformation carrying the resistant GmSNAP18+ allele from Forrest but the susceptible GmSHMT08− allele from Essex was developed and genotyped as described by [62].
The functional characterization of the four THF binding sites (L129, L135, F284, N374), four PLP binding sites (G106, G107, S190A, H218), and the two PLP catalysis sites (Y59 and H134) at the GmSHMT08 protein has been validated using the transgenic hairy root system experiment. Williams 82 composite hairy roots transformed with pG2RNAi2:: empty vectors were used as a negative control. The pG2RNAi2 vector has a sGFP-selectable marker in planta [21,23]. Transgenic Williams 82 composite hairy roots transformed with pG2RNAi2::GmSHMT08 and the ten different mutated GmSHMT08 PLP and THF cofactor binding/catalysis sites (pG2RNAi2::GmSHMTΔPLP and pG2RNAi2::GmSHMTΔTHF) were produced by injecting agrobacterium bacterial suspensions three times into the hypocotyl directly below soybean cotyledons using a 3 mL needle (BD#309578) as shown earlier [21]. After injection, composite hairy roots from at least 50 independent soybean transgenic plants per construct were grown and propagated in medium vermiculite. Transgenic soybeans were covered with plastic humidity domes sprayed consistently with water, maintained in a growth chamber for 1–2 weeks, and fertilized once per week with NPK 20-20-20 fertilizer. GFP-positive composite hairy roots at ~2–3 inches long were transferred into a steam-pasteurized sandy soil and packed into plastic containers as mentioned earlier [62]. Each container held 25 tubes and was suspended over water baths maintained at 27 °C. At least 15 plants from the control lines (WI82 and ExF12) were arranged in a randomized complete block design. Two days after transplanting, each plant was inoculated with ~2000 SCN (HG0) eggs. After 30 days, cysts were counted under a stereomicroscope. The experiment was independently conducted three times to obtain a minimum of 15 to 20 independent composite hairy root lines per construct per experiment. The results were plotted and analyzed for statistical significance by using analysis of variance (ANOVA) using the JMP Pro V12 software as described earlier.
The availability of new crystal structure of the GmSHMT08 in soybeans [20] enhanced our knowledge of how the previously identified 18 EMS Gmshmt08 TILLING mutants can affect the PLP/THF cofactor binding and catalysis sites. Thus, we performed in silico mutational analysis of the thirteen Gmshmt08 EMS mutants that were identified using forward genetic screening [6], the three Gmshmt08 EMS mutants identified using forward genetics [7], and the two Gmshmt08 EMS mutants that were identified using Gel-TILLING [2].
Homology modeling of a putative GmSHMT08 protein structure was conducted using Deepview and Swiss-Model Workspace software, as previously shown [21,23]. Briefly, protein sequences from Forrest and the available SHMT crystal structure from soybean Glycine max cultivar Essex (PDB accession 6uxj.1) were used as templates. Residues 2–470 were modelled against their corresponding template with a sequence identity of 99.57% (according to the Protein Data Bank database). The structure of serine hydroxymethyltransferase from Glycine max cultivar Essex was complexed with PLP-glycine and 5-formyltetrahydrofolate residues [20]. Visualization of the THF cofactor binding sites (L129, L135, F284, N374), PLP cofactor binding and catalysis residues (Y59, G106, G107, H134, S190A, H218)—in addition to the two polymorphisms (P130R and N358Y)—and the 18 EMS-induced GmSHMT08 mutations was performed using the UCSF Chimera package [69]. To study the impact on the THF/PLP cofactor binding/catalysis and EMS mutations that were located close to the PLP/THF cofactor sites, the mapped induced mutations at the PLP cofactor sites, THF cofactor sites, and EMS-induced mutations were mutated using the structural editing tool from the UCSF Chimera package. Then, the rotamers tool that is incorporated within the Chimera package software was used to mutate the corresponding residues [24]. The rotamers tool allows amino acid sidechain rotamers to be viewed, evaluated, and incorporated into structures in which a given residue can be changed into different amino acids to predict the impact and effect of the mutations on the adjacent residues surrounding the mutated residue.
Our data are coherent with previous studies showing that glutathione peroxidase transcription, among other ROS scavenging enzymes, was significantly modulated under SCN infection in syncytia [51]. The Arabidopsis thaliana Atshmt1-1 mutant showed a greater accumulation of H2O2, which is known to induce salicylic acid biosynthesis [70,71]. The implications of phytohormones, such as SA and CK, have been previously shown to be involved in a crosstalk between SCN-resistant genes (GmSHMT08 and GmSNAP18) and SCN defense genes (GmPR08-Bet VI) [21]. Maintenance of a certain level of ROS homeostasis at low levels is required for parasitic nematodes to cause and maintain pathogenic disease [72,73]. However, disruption of this homeostasis (overaccumulation of ROS) can cause termination of syncytial formation or syncytial apoptosis [72,73,74]. Taken together, modulation of the SHMT serine/glycine interconversion may impact important maintenance of redox homeostasis that occurs via ROS. Maintenance of balanced SHMT expression appears to be highly important in plants. The current study uncovered for the first time the involvement of the interconversion reaction carried out by the serine hydroxymethyltransferase protein involving the two cofactors at the GmSHMT08c protein, the four THF cofactor biding sites, and the six PLP cofactor binding/catalysis sites in resistance to SCN. | true | true | true |
PMC9570250 | 36233212 | Jéssika Cristina Chagas Lesbon,Taismara Kustro Garnica,Pedro Luiz Porfírio Xavier,Arina Lázaro Rochetti,Rui Manuel Reis,Susanne Müller,Heidge Fukumasu | A Screening of Epigenetic Therapeutic Targets for Non-Small Cell Lung Cancer Reveals PADI4 and KDM6B as Promising Candidates | 07-10-2022 | NSCLC,epigenetic targets,metastasis,KDM6B,PADI4 | Despite advances in diagnostic and therapeutic approaches for lung cancer, new therapies targeting metastasis by the specific regulation of cancer genes are needed. In this study, we screened a small library of epigenetic inhibitors in non-small-cell lung cancer (NSCLC) cell lines and evaluated 38 epigenetic targets for their potential role in metastatic NSCLC. The potential candidates were ranked by a streamlined approach using in silico and in vitro experiments based on publicly available databases and evaluated by real-time qPCR target gene expression, cell viability and invasion assays, and transcriptomic analysis. The survival rate of patients with lung adenocarcinoma is inversely correlated with the gene expression of eight epigenetic targets, and a systematic review of the literature confirmed that four of them have already been identified as targets for the treatment of NSCLC. Using nontoxic doses of the remaining inhibitors, KDM6B and PADI4 were identified as potential targets affecting the invasion and migration of metastatic lung cancer cell lines. Transcriptomic analysis of KDM6B and PADI4 treated cells showed altered expression of important genes related to the metastatic process. In conclusion, we showed that KDM6B and PADI4 are promising targets for inhibiting the metastasis of lung adenocarcinoma cancer cells. | A Screening of Epigenetic Therapeutic Targets for Non-Small Cell Lung Cancer Reveals PADI4 and KDM6B as Promising Candidates
Despite advances in diagnostic and therapeutic approaches for lung cancer, new therapies targeting metastasis by the specific regulation of cancer genes are needed. In this study, we screened a small library of epigenetic inhibitors in non-small-cell lung cancer (NSCLC) cell lines and evaluated 38 epigenetic targets for their potential role in metastatic NSCLC. The potential candidates were ranked by a streamlined approach using in silico and in vitro experiments based on publicly available databases and evaluated by real-time qPCR target gene expression, cell viability and invasion assays, and transcriptomic analysis. The survival rate of patients with lung adenocarcinoma is inversely correlated with the gene expression of eight epigenetic targets, and a systematic review of the literature confirmed that four of them have already been identified as targets for the treatment of NSCLC. Using nontoxic doses of the remaining inhibitors, KDM6B and PADI4 were identified as potential targets affecting the invasion and migration of metastatic lung cancer cell lines. Transcriptomic analysis of KDM6B and PADI4 treated cells showed altered expression of important genes related to the metastatic process. In conclusion, we showed that KDM6B and PADI4 are promising targets for inhibiting the metastasis of lung adenocarcinoma cancer cells.
Lung cancer is the second most diagnosed cancer worldwide and was responsible for 1,796,144 deaths in 2020, according to GLOBOCAN [1]. Non-small cell lung cancer (NSCLC) represents 85% of the cases, of which 80% are adenocarcinomas (AdCs), adenosquamous carcinomas, or squamous-cell carcinomas (SqCCs) [2,3]. Unfortunately, almost 50% of lung cancer cases are metastatic resulting in a poor prognosis and limited therapeutic options, with a critical five-year overall survival (OS) of only 10% and 1% in patients with stage IVA and IVB respectively [3,4,5]. The low survival rates are mainly attributed to chemoresistance, low detection rate of mutations in target genes, compromised choice of targeted therapy, and late diagnosis of lung cancer patients [3,6,7]. Therefore, the identification of novel therapeutic targets and their inhibitors is urgent [8]. Proteins that modify the epigenetic code are promising targets for the development of new anti-metastasis and anti-invasion drugs for NSCLC [8]. Histone posttranslational modifications (PTMs) represent epigenetic modifications that are frequently altered in cancer and contribute to tumor migration, metastasis, and aberrant cellular growth [9]. Many histone deacetylase (HDAC) inhibitors (HDIs), such as vorinostat and panobinostat, have shown promising results in preclinical and clinical investigations of NSCLC [9] and new molecules for epigenetic targets are being developed and explored for their use in the treatment of diverse cancers [10]. However, there remains a need to validate these targets in large-scale clinical trials [8]. The Structural Genomics Consortium (SGC) is an international public–private partnership with the goal of supporting research for a better understanding of human disease biology and to enable the discovery of new medicines (https://www.thesgc.org, accessed on 25 January 2019). To this end, SGC develops and makes available highly specific inhibitors (chemical probes) to the scientific community [11,12,13]. Most available epigenetic probes are inhibitors of bromodomains (BRDs) and protein methyltransferases (PMTs). These molecules have been shown to be effective in several tumor models by inhibiting or attenuating several characteristics relevant to tumor development, such as metastatic capacity and resistance to conventional treatment [14,15,16]. However, epigenetic inhibitors used at adequate doses can inhibit tumorigenic phenotypes without being overtly cytotoxic to cells from healthy tissues [17]. Here, we performed a streamlined set of in silico and in vitro experiments to rank and validate epigenetic targets that regulate the metastatic process in lung cancer cells, using SGC chemical probes as specific inhibitors.
All 38 epigenetic targets were analyzed and ranked based on the significance of inverse association between survival of patients with NSCLC and gene expression. According to the selection criteria, from 1082 patients, 590 and 492 patients were selected for adenocarcinoma and squamous cell carcinoma, respectively. Of the 38 epigenetic targets, eight were inversely associated with low survival of lung adenocarcinoma patients (Hazard Ratio (HR) > 1, p < 0.05, Table 1), and none were associated with patients with squamous cell carcinoma (n = 492).
The systematic review initially resulted in 98 publications related to the epigenetic target Enhancer Of Zeste 2 Polycomb Repressive Complex 2 Subunit (EZH2), 6 related to Bromodomain Containing 4 (BRD4), 5 related to Protein Arginine Methyltransferase 1 (PRMT1), 4 related to each Lysine Demethylase 6B (KDM6B) and Bromodomain Containing 9 (BRD9) targets; one related to Coactivator Associated Arginine Methyltransferase 1 (CARM1), and no work related to Bromodomain Adjacent To Zinc Finger Domain 2 (BAZ2A) or Peptidyl Arginine Deiminase 4 (PADI4) as targets in lung cancer. Specific targets, such as PRMT1, KDM6B, CARM1, BRD4, and EZH2, have been shown to be associated with malignant phenotypes of lung cancer, influencing cell proliferation and metastatic processes (regulation of epithelial–mesenchymal transition and cell invasion). Lysine Demethylase 6A (KDM6A) expression was not correlated with poor survival in lung cancer patients. Thus, we selected the targets that presented the highest risk rate (HZ) and the lowest number of publications in the literature (<5), with the aim of studying potential new epigenetic targets for lung cancer. Therefore, the epigenetic targets selected for further analysis were KDM6B, CARM1, BAZ2A and PADI4.
Through the analysis of gene expression in silico, three cell lines showed elevated expression of the chosen potential targets (A549, H2126 and H1568). The cell lines (H2126 and H1568) were collected from metastatic sites, pleural effusion, and lymph nodes, whereas A549 cells were collected from the primary tumors (Figure S1). However, the H2126 cell line does not have the potential for in vitro invasion [18]. The expression of the epigenetic targets in healthy lung tissue showed the following results: KDM6B (Z-score = 2.8), CARM1 (Z-score = −0.5), BAZ2A (Z-score = −1.0) and PADI4 (Z-score = 0.58), Z-score < 5, suggesting non-expression in healthy lung tissue.
The expression of CARM1, BAZ2A, KDM6B and PADI4 was evaluated in the cell lines H2126, H1568 and A549. All three cell lines, A549, H2126, and H1568, showed a higher level of target expression in general (Figure S2).
Cytotoxic potential of the four epigenetic inhibitors, TP-064 (CARM1), GSK2801 (BAZ2A/B), GSK-J4 (KDM6A/B), and GSK484 (PADI4) in A549 cells was assessed GSK-J4 (KDM6A/B) had an IC50 value of 8.21 µM, whereas the other probes were not cytotoxic, even at 10 µM. Therefore, the inhibitors showed low cytotoxic potential, even when using very high doses not recommended for use, demonstrating high safety. However, we selected a concentration of 1000nM for further experiments (Figure 1).
Inhibition of the epigenetic targets KDM6A/B and PADI4 reduced cell invasiveness compared to the control group (p < 0.05) in the A549 and H1568 cell lines (Figure 2).
Treatment with the PADI4 inhibitor GSK484 led to 152 differentially expressed genes, of which 62 genes were downregulated and 90 were upregulated in A549 cancer cells, whereas the KDM6A/B inhibitor, GSK-J4, altered the expression of 190 genes, of which 56 genes were downregulated and 134 were upregulated in A549 cancer cells (FDR < 0.05 and LogFC > 1; <−1) (Table S1). Functional enrichment analysis showed that treatment with PADI4 and KDM6A/B inhibitors was associated with processes linked to the collagen-containing extracellular matrix, extracellular matrix, extracellular space, cell periphery-related genes, and processes related to metastasis. In cells treated with the PADI4 inhibitor, we found nine genes differentially regulated and six genes in cells treated with the KDM6A/B inhibitor, of which the following five genes were common among the treatments: the genes for Fibrinogen Alpha Chain (FGA), Nidogen 2 (NID2), Inter-Alpha-Trypsin Inhibitor Heavy Chain 2 (ITIH2), Peroxidasin (PXDN) and Heparin Binding EGF Like Growth Factor (HBEGF). All of the five genes common among the treatments are related to adhesion proteins, cell ligands, and protein stabilizing proteins of the extracellular matrix, suggesting that these genes participate in the regulation of metastasis (Figure 3).
Metastasis, one of the biggest problems of solid epithelial cancers, begins with the migration of tumor cells from the confined primary tumor to adjacent tissue, where tumor cells cross the basement membrane and lamina propria to invade the underlying connective tissue. Unlike normal epithelial cells, which undergo apoptosis when they lose contact with their native extracellular matrix, tumor cells develop mechanisms to detach from the primary tumor associated with epithelial organization, closely followed by the expression of mesenchymal markers [18]. These changes are the result of altered gene expression, which can be driven by epigenetic processes, thereby opening the possibility of affecting these changes by epigenetic regulation. Here, we performed a streamlined approach with in silico and in vitro analyses starting from 38 epigenetic targets to select the most relevant for lung cancer cell treatment and showed that the inhibition of PADI4 and KDM6B proteins controls the metastatic process, inhibiting cancer cell migration and invasion by altering their transcriptomes. Protein-arginine deiminase Type-4 (PADI4) is a calcium-dependent enzyme that is known for its role in converting arginine to citrulline residues. Its downstream signaling has been studied in the progression of a variety of human cancers, but there is a lack of studies showing the efficacy of PADI4 in lung cancer [19,20]. Recently, Liu et al. (2019) demonstrated that PADI4 is overexpressed in lung cancer and contributes to cell growth and metastasis. Knockdown of PADI4 in A549 lung cancer cells resulted in a striking reduction in the EMT-associated Snail Family Transcriptional Repressor 1 (Snail1/mothers) against the decapentaplegic homolog ¾ transcriptional complex, which was consistent with alterations in migratory and invasive phenotypes of A549 lung cancer cells. On the other hand, the lysine demethylase 6B (KDM6B) is a histone demethylase that removes methyl groups from lysine and arginine residues on histone tails. It is a member of the Fe(II)- and α-ketoglutarate-dependent demethylases that activates gene expression by removing H3K27me3 marks on gene promoters [21]. KDM6B has been shown to be involved in tumor progression via the regulation of cell proliferation, migration, and senescence [22]. High levels of KDM6B induce the expression of mesenchymal genes, such as Snail and Slug (Snail Family Transcriptional Repressor 2), which promote TGF-β-induced (Transforming Growth Factor Beta 1) EMT and tumor metastasis [23]. Knockdown of KDM6B inhibited EMT induced by TGF-β, inhibiting breast cancer cell invasion [21]. Another study provided evidence of pulmonary metastasis of osteosarcoma in an in vivo model in which osteosarcoma cells were injected into the medullary cavity of nude mice. Intraperitoneal administration of GSK-J4 at concentrations above 5 mg/kg significantly inhibited the pulmonary metastasis of osteosarcoma cells in vivo. These results strongly suggest the potential of KDM6B as a target for highly metastatic osteosarcoma [24]. Thus, KDM6B may present a target for cancer metastasis. One point to consider is that GSK-J4 could also inhibit lysine demethylase 5B (KDM5B) histone demethylase and not only KDM6A/B. KDM5B has been implicated in several cancers, including NSCLC, and was recently described as a therapeutic target for cancer therapy [25]. However, GSK-J4 is more selectively potent for KDM6B than for KDM5B. Interestingly, the treatment of cancer cells with non-cytotoxic doses of PADI4 and KDM6B inhibitors induced similar transcriptomic profiles, regulating genes related to cell adhesion and the extracellular matrix, which was associated with decreased capacity of cancer cells to invade and migrate in the in vitro model. For both inhibitors, upregulation of FGA, NID2 and ITIH2 genes, and downregulation of PXDN and HBEGF, was observed. Fibrinogen is an extracellular matrix protein composed of three polypeptide chains, fibrinogen alpha (FGA), beta (FGB), and gamma (FGG), and is involved in tumor angiogenesis and metastasis. FGA may play a suppressive role by inhibiting tumor growth and metastasis. FGA administration is considered a novel therapeutic approach to inhibit the growth and metastasis of lung adenocarcinoma [26]. Nidogen-2 (NID2) is ubiquitously present in the basement membrane and maintains its integrity and stability of the basement membrane by connecting laminin and collagen IV networks in the extracellular matrix (ECM). The restoration of NID2 expression in cancer cells was shown to have a negative regulatory role in Epidermal Growth Factor Receptor (EGFR) and integrin signaling pathways, suggesting that NID2 elicits in vitro migration/invasion suppression and in vivo metastasis inhibition effects through negative modulation of these two oncogenic pathways [27]. The other gene upregulated by both inhibitors was ITIH2, the inter-alpha-trypsin inhibitor 2, belonging to a family of plasma protease inhibitors, contributing to the stability of the extracellular matrix by covalently binding to hyaluronan. Loss or downregulation of ITIH2 expression was observed in 70%, 71%, and 70% of breast, lung, and kidney tumors, respectively. In addition, careful densitometric evaluation of hybridization signals revealed downregulation in 56% of gastric cancers, 61% of rectal carcinomas, and 50% of prostate cancers [28]. Epigenetic inhibitors downregulated two genes in common: PXDN and HB-EGF. Peroxidasin (PXDN) is an extracellular matrix protein with peroxidase activity and has been reported to participate in epithelial mesenchymal transition processes, playing a promoting role in the proliferation, invasion, and migration of ovarian cancer cells through the regulation of PI3K (Phosphatidylinositol-4,5-Bisphosphate 3-Kinase Catalytic) pathway activation Pl3k/Akt (AKT Serine/Threonine Kinase), and is considered a potential target for therapy [29]. Heparin-bound epidermal growth factor-like growth factor (HB-EGF) is a member of the heparin-bound EGF family (Epidermal Growth Factor) and is more widely expressed in tumors than in normal tissues. HB-EGF can be produced in a membrane-anchored form (pro-HB-EGF) and further processed into a soluble form (s-HB-EGF), although a significant amount of pro-HB-EGF remains cleaved on the surface of the cell. In addition, wild-type s-HB-EGF or HB-EGF induced the expression and activity of the metalloproteases MMP-9 (Matrix Metallopeptidase 9) and MMP-3 (Matrix Metallopeptidase 3), leading to increased cell migration [30]. PADI4 inhibitor treatment in cancer cells downregulated four genes related to metastatic cancer phenotypes: Laminin Subunit Gamma 2 (LAMC2), C-X-C Motif Chemokine Ligand 8 (CXCL8), Niban Apoptosis Regulator 1 (FAM129A), and Pleckstrin 2 (PLEK2). Laminin Subunit Gamma 2 (LAMC2) is a subunit of the heterotrimeric glycoprotein laminin-332 (LAM-332, formerly laminin-5) consisting of α3, β3, and γ2 chains. Although LAMC2 is an important structural component of the epithelial basement membrane (BM) in various normal tissues, there is emerging evidence of a pathological role for the LAMC2 monomer in cancer [31]. LAMC2 promotes migration and invasion via EMT, which is dependent on TGF-β1 and ZEB1 (Zinc Finger E-Box Binding Homeobox 1) integrin [31]. CXCL8, also known as interleukin-8 (IL-8), is a prototypic chemokine belonging to the CXC family and is responsible for the recruitment and activation of neutrophils and granulocytes to the site of inflammation [32]. Recent studies have shown that CXCL8 is essential for tumor cells to acquire and maintain this aggressive phenotype. A member of the family with sequence similarity 129, member A (FAM129A), inhibited apoptosis and promoted migration and proliferation in human cancers. One study revealed that FAM129A promoted tumor invasion and proliferation by upregulating the expression of MMP2 (Matrix Metallopeptidase 2) and cyclin D1, which was due to increased FAK (Protein Tyrosine Kinase 2) phosphorylation at Tyr 397 and Tyr 576. Overexpression of FAM129A was associated with tumor progression and predicted low survival of NSCLC patients [33]. Pleckstrin2 (PLEK2) is a 353 amino acid protein that is widely expressed in a variety of tissues and is highly expressed in NSCLC. PLEK2 promotes NSCLC proliferation and metastasis via a BRD4-dependent PI3K/AKT signaling pathway that functions as an epigenetic reader and binds to acetylated lysine residues (KAc) that regulate chromatin structure and gene expression [34]. Treatment of cancer cells with a KDM6B inhibitor downregulated FUT1 (Fucosyltransferase 1) gene, which is also related to metastasis. Fucosylation is a posttranslational modification that links fucose residues with protein- or lipid-linked oligosaccharides. Certain genes in the fucosylation pathway are aberrantly expressed in several types of cancer, including non-small cell lung cancer, and this aberrant expression is associated with poor prognosis in cancer patients. Fucosylation pathway genes, including fucosyltransferase 1/2/3/6/8 (FUT1, FUT2, FUT3, FUT6, FUT8) and GDP-L-fucose synthase (TSTA3), were correlated with poor patient survival in these patients. In this study, the inhibition of FUTs by 2F-peracetyl-fucose (2F-PAF) suppressed transforming growth factor β (TGFβ)-mediated Smad3 (SMAD Family Member 3) phosphorylation and nuclear translocation in NSCLC cells. Furthermore, transwell wound healing and migration assays demonstrated that 2F-PAF inhibited the TGFβ-induced migration and invasion of NSCLC cells [35]. Our work emphasized the inhibition of important epigenetic targets related to the process of migration and invasion of tumor cells that favor cancer metastasis. Thus, these inhibitors have great potential to add to antitumor therapy, and can be added to other drugs already in clinical use, such as chemotherapy, immunotherapy and targeted therapy, contributing to the increase of antitumor effects, overcoming resistance to drugs already used and activation of the host’s immune response. Indeed, chemotherapy is still a traditional method in advanced cases, in which surgical excision is not possible, so the emergence of chemoresistance remains a major problem in cancer therapy. Thus, the combination of epigenetic drugs with other chemotherapeutics can not only promote a potent suppression of tumorigenesis, but also resensitize tumor cells to radiotherapy and chemotherapy [36]. Immunotherapy has been used as a promising candidate for both first- and second-line treatment in metastatic NSCLC. However, about 50% of NSCLC expressed PD-L1. There is no consensus predictive biomarker and resistance to immunotherapy can occur [37,38,39]. This fact limits the use of immunotherapy and overcoming immunotherapy resistance can be challenging due to the complex and dynamic interplay between malignant cells and the defense system. In the case of resistance, the epigenetic inhibitors could act as reactivating tumor suppressor genes and repress cancer cell growth. Some studies have shown that epigenetic inhibitors, such as BET, LSD1 and EZH2 inhibitors, are already used in combination with anti-PD1 therapy activating the antitumor immune response by increasing the persistence of T cells in the tumor microenvironment [3]. A study by Rohrbach and collaborators elucidates the relation between PAD4 activation and immune cells. PAD4 is expressed in granulocytes, which are essential for innate immunity and the formation of neutrophil extracellular traps (NETs). Anti-PAD4 therapies have been proposed for inflammatory and cancer conditions, but we need a better understanding regarding the role of neutrophils in cancer. The tumor microenvironment is composed of adaptive immune cells, which play important roles in tumor growth and metastasis [40]. Shi et al., 2020, transplanted Padi4 wild-type and Padi4-knocknout breast cancer cells into inguinal mammary fat pad areas of immunodeficient mice, which lacked functional T cells, B cells and NK, and found that tumor derived PADI4 facilitated metastasis, at least partially independent of the adaptative immune cells. Those findings together suggested that PADI4 inhibition can negatively affect the immune cells; however, the effects on metastatic cancer cells remained [41]. Lysine demethylase 6b (KDM6B) is essential for the generation and proper functioning of CD8+ effector T cells during acute infection and tumor eradication, being indispensable for proper effector functions and tumor protection, and KDM6B inhibition exhibits a memory-defective T cell response. Therefore, KDM6B may act as an epigenetic modulator of CD8+ T cell fate determination by regulating effector-associated gene expression and chromatin accessibility [42]. As members of the KDM6 family have been therapeutic targets for several cancers, it is necessary to properly understand their intrinsic role in T cell function. More studies are necessary to better understand the interaction between epigenetic protein inhibition and immunotherapy.
A streamlined set of in silico analyses coupled with in vitro analyses (Figure 4) was performed to evaluate and rank potential epigenetic targets based on epigenetic probes from the SGC (https://www.thesgc.org, accessed on 25 January 2019). The screening was based on a list of available epigenetic inhibitors from SGC, followed by an analysis of the association of the epigenetic target with NSCLC survival, subsequent selection by a systematic review of potential new cancer targets, in silico analysis of protein expression in NSCLC cell lines, and real-time qPCR expression to evaluate target expression in the cell lines.
The Kaplan-Meier Plotter software [43] (http://kmplot.com/analysis, accessed on 4 February 2019).) was powered with public data from 14 repositories with information on gene expression and clinical samples, totaling 2438 cases of NSCLC. Thirty-eight epigenetic targets (Table S2) were individually analyzed for their association with the survival rates of 590 patients diagnosed with adenocarcinoma and 492 patients with squamous cell carcinoma (Table 2). The patient selection criteria considered the histological types, grouping them into adenocarcinoma and squamous cell carcinoma, patients in stages I, II, III, and IV of the disease, of both sexes, smokers and non-smokers, and if patients had started any type of treatment, such as surgery, chemotherapy, and radiotherapy. The follow-up time for each patient was evaluated from the time of diagnosis to the time of death. For all analyses, the results (p < 0.05) were considered, according to the log-rank test (chi-square), to compare whether there was a statistical difference between the curves of high and low gene expression and the use of the hazard ratio (HR) with a 95% confidence interval. A hazard ratio equal to one means no association between treatments, a rate greater than one suggests an increase in risk and below one suggests a decrease in risk. For the selection of the cases, the following inclusion criteria were applied: use of cases that presented patient survival; quality control of array chips excluding chips with outliers (>95% of total arrays) from analysis for any of the following parameters: percentage of calls present, background, rawQ, bioB-/C-/D-spikes, GAPDH (Glyceraldehyde-3-Phosphate Dehydrogenase) and ACTB (Actin Beta) 3 ratio for 5. As recommended by the authors, the Jetset Best probe set was always used to analyze the expression of genes of interest, and a high and low expression group based on quartiles (25%, Q1 × Q4) was created for survival analysis. Analyses were performed independently for adenocarcinoma and squamous cell carcinoma cases. Statistical analysis was performed using univariate Cox regression, generating p-values and hazard ratios.
An independent systematic literature review was performed for each epigenetic target in the PubMed database (https://www.ncbi.nlm.nih.gov/pubmed, accessed on 7 February 2019). For literature selection, a specific set of keywords was used, presenting the abbreviation of the name of the epigenetic target and the term “lung cancer,” as being mandatory in the titles in order to find studies that specifically evaluated epigenetic targets and lung cancer. The analysis was performed on publications published from 2000 to 2019.
The CellExpress software [44] (http://cellexpress.cgm.ntu.edu.tw, accessed on 11 March 2019) was used to perform gene expression analysis on more than 4000 tumor cell lines and clinical samples obtained from public datasets. For expression analysis, the databases of gene expression studies of cell lines NCI-60 Human Tumor Cell Lines Screen (GSE32474), Cancer Cell Line Encyclopedia-CCLE (GSE36133), and Sanger Cell Line Project (GSE68950) were used. Microarray data obtained on the same platform were normalized using a quartile normalization algorithm to remove systematic biases. Expression data from the GSE36133 study from the CCLE database were used, which presented a more complete list of cell lines of interest, using the selection of the Jetset Best probe and the expression of the four potential epigenetic targets in lung adenocarcinoma cell lines. The probes selected to assess gene expression levels in cells were from the Jetset Best probe set, being (41386_at) for KDM6B (212512_at) for CARM1, (201353_at) for BAZ2A and (220001_at) for PADI4, which were also used in the Kaplan-Meier survival analysis. The endogenous gene, GAPDH, was used to normalize the expression levels of the genes of interest. The results were generated by calculating the relative expression, which showed similar gene expression values between the analyzed cells. Evaluation of the expression of targets in healthy lung tissue was performed using the CellNavigator software (https://medicalgenomics.org/rna_seq_atlas, accessed on 15 March 2019) through microarray analysis coupled to the RNA-seq Atlas platform through the Human Genome Set U133 (HG-U133). Background correction, normalization and summarization was performed by applying the frma function from the fRMA package to AffyBatch with default options. The Z-Score transformation was calculated using the barcode function of the fRMA package to standardize gene values from the Microarray data. The barcode options were set for the corresponding platform and the output method was set to ‘z-score’. Then, the Z-Scores were averaged for each tissue and each pathological state (healthy, cancer), Z-Score > 5 suggests that the gene is expressed in that tissue. Finally, the Z-Score was averaged for each tissue and state (healthy, cancer) and stored in the PostgreSQL database.
The lung cancer cell lines A549, H1568, and H2126 were donated by Dr. Lucy M. Anderson from the Laboratory of Comparative Carcinogenesis at the Frederick National Laboratory for Cancer Research, Frederick, MD, USA, and maintained as previously described [45]. Briefly, cell lines were maintained in 75 cm2 flasks at 37 °C and 5% CO2 in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic (Pen-Strep). Cell passaging was performed when the cells were 85% confluent using TrypLETM Express trypsin. Culture evolution was evaluated daily using optical microscopy (Axio Vert A1, Zeiss, Jena, Germany). All the reagents used for cell culture were purchased from Thermo Fisher Scientific (USA). All cell lines were authenticated at the Laboratory of Molecular Diagnosis of the Cancer Hospital of Barretos (Hospital de Amor HA) as previously reported [46] before the experiments and were free for Mycoplasma spp. by real-time PCR (Myco-Sniff-Valid™ Mycoplasma PCR Detection Kit).
Total RNA was extracted from A549, H1568, and H2126 cell lines using TRIzol (Invitrogen; Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. cDNA was synthesized from total RNA (1000 ng) using a High-Capacity cDNA Reverse Transcription kit (Applied Biosystems; Thermo Fisher Scientific, Inc.) using the following parameters: 25 °C for 10 min, 37 °C for 120 min, and 85 °C for 5 min. qPCR was performed using the SYBR Master Mix (Roche Diagnostics, Basel, Switzerland). Gene expression analyses were performed by real-time PCR using the StepOne System (Thermo Fisher Scientific). Specific primers were designed using Primer-BLAST [47] and dimers and hairpins were verified using AutoDimer software [48]. Primers were also analyzed using in silico polymerase chain reaction (PCR) (https://genome.ucsc.edu/cgi-bin/hgPcr, accessed on 7 March 2019) to confirm specificity. The primer sequences are listed in (Table S3). PCR was carried out using Fast SYBR Green Master Mix in a final volume of 10 µL. The conditions for quantitative PCR were as follows: 95 °C for 20 s; 40 cycles at 95 °C for 3 s for denaturation and 60 °C for 30 s for anneal/extension; melt curve analysis was performed at 95 °C for 15 s and 60 °C for 60 s. The housekeeping gene used was 18 s ribosomal RNA, and the analysis of relative gene expression data was performed according to the ΔΔCt method [49]. The experiments were performed twice and in triplicate. All reagents were purchased from Thermo Fisher Scientific. The experiment was performed twice and in triplicate.
The epigenetic probes (Cayman Chemical, Ann Arbor, MI, USA) were dissolved in dimethylsulfoxide (DMSO) to a concentration of 20 mM. A549 cells were seeded at 5000/well in 96-well plates (Corning, NY, USA) containing 100 µL of supplemented media, as described previously. After 24 h, the medium was replaced with fresh culture medium containing different concentrations of epigenetic probes, ranging from 10 µM to 13.72 nM. Epigenetic probes were added in six replicates per concentration and the experiments were performed in triplicate. After 72 h, 10 µL of 3-(4.5-dimethylthiazol-2-yl)-2.5-diphenyl tetrazolium bromide (MTT, 5 mg/mL) was added to each well and formazan crystals were produced over a 2 h incubation period. One hundred microliters of DMSO were added to dissolve the crystals. The optical density at 540 nm was measured using Fluorstar Optima (BMG Labtech, Ortenberg, Germany). The concentration of the compound corresponding to the IC50 was calculated using a nonlinear regression test performed in GraphPad Prism (version 6.00 for Windows, GraphPad Software, USA).
The cells were treated with DMSO (control) or 100 nM of TP-064, GSK2801, GSK-J4, and GSK484 for 72 h. Cells were cultured for 24 h in serum-free medium. Transwell inserts were placed in 24 well plates and filled with 100 μL of ECM gel (Sigma Aldrich Saint Louis, MO, USA) in RPMI-40 medium (1:5). After, 2 × 104; A549 and 2.5 × 104 H1568 cells were resuspended in 100 μL serum-free medium and plated on inserts. The bottom well was filled with 600 μL of RPMI-40 medium supplemented with 20% fetal bovine serum (FBS), used as a chemoattractant, and after 48 h, a cotton swab was used to remove non-invasive cells from the top of the inserts. As a fixative, 5% glutaraldehyde was used for 10 min at room temperature and inserts were stained with 1% crystal violet in 2% ethanol for 20 min. The invasive cells were observed and photographed under an optical microscope in five random fields at 100× magnification using the ZEISS ZEN 2 Microscope Software (ZEISS, Germany). Finally, the invasive cells were counted using ImageJ software version 1.8.0_112 [46]. The experiment was performed thrice in duplicate.
To assess the genes affected by treatment with GSK 484 (PADI4) and GSK-J4 (KDM6A/B), A549 cells were treated with 100 nM of these inhibitors. Duplicates of each treatment and control group were prepared. A549 cells were treated with 100 nM GSK-J4, GSK484, and DMSO (control) for 72 h and then RNA was extracted using TRIzol (Invitrogen; Thermo Fisher Scientific, Inc.). RNA quality and quantity were assessed using automated capillary gel electrophoresis on a Bioanalyzer 2100 with RNA 6000 Nano Labchips, according to the manufacturer’s instructions (Agilent Technologies, Cork, Ireland). Only samples that presented an RNA integrity number (RIN) higher than 8.0 were considered for sequencing. RNA libraries were constructed using the TruSeq™ Stranded mRNA LT Sample Prep Protocol and sequenced on an Illumina HiSeq platform. 2500 equipment in HiSeq Flow Cell v4 using a HiSeq SBS Kit v4 (2 × 100 bp).
Sequencing quality was evaluated using FastQC software (http://www.bioinformatics.babraham.ac.uk/projects/fastqc, accessed on 3 March 2019), and no additional filtering was performed. Sequence alignment against the human reference genome (GRHC38) was performed using STAR [50], according to standard parameters and including the annotation file (Ensembl release 89). Secondary alignments, duplicated reads, and reads failing vendor quality checks were removed using Samtools [51]. The alignment quality was confirmed using Qualimap [52]. Gene expression was estimated by read counts using HTseq [53] and normalized to counts per million reads (CPM). Only genes presenting at least one CPM in at least four samples were retained for differential expression (DE) analysis. DE was performed using the EdgeR package [54] in the R environment based on a negative binomial distribution. The Benjamini-Hochberg procedure was used to control the false discovery rate (FDR), and transcripts with FDR ≤ 0.05, and log-fold change (LogFC) > 1; <−1 were considered differentially expressed (DE). Functional enrichment analysis of the DE genes was performed using STRING [55,56].
The IC50 was calculated using a nonlinear regression test. Gene expression was analyzed by one-way ANOVA with Tukey’s post-hoc test. One-way ANOVA followed by Student’s t-test was used for invasion assays. For functional enrichment analyses, p-values were adjusted for multiple tests, and the Benjamin and Hochberg method was used to test multiple categories in a group of functional gene sets. Differences were considered statistically significant at p < 0.05.
In summary, a streamlined approach of in silico and in vitro experiments allowed us to select, from 38 different epigenetic targets, the two most promising candidates for NSCLC drug development: PADI4 (GSK 484) and KDM6B (GSK-J4). The inhibition of these epigenetic proteins regulates molecular pathways in NSCLC, affecting the ability of cancer cells to migrate and invade, thereby controlling the metastatic cascade. Treatment with the identified inhibitors regulates common genes linked to tumor metastasis. | true | true | true |
PMC9570315 | 36232428 | Florencia Haase,Rachna Singh,Brian Gloss,Patrick Tam,Wendy Gold | Meta-Analysis Identifies BDNF and Novel Common Genes Differently Altered in Cross-Species Models of Rett Syndrome | 22-09-2022 | Rett syndrome,WGCNA,MECP2 | Rett syndrome (RTT) is a rare disorder and one of the most abundant causes of intellectual disabilities in females. Single mutations in the gene coding for methyl-CpG-binding protein 2 (MeCP2) are responsible for the disorder. MeCP2 regulates gene expression as a transcriptional regulator as well as through epigenetic imprinting and chromatin condensation. Consequently, numerous biological pathways on multiple levels are influenced. However, the exact molecular pathways from genotype to phenotype are currently not fully elucidated. Treatment of RTT is purely symptomatic as no curative options for RTT have yet to reach the clinic. The paucity of this is mainly due to an incomplete understanding of the underlying pathophysiology of the disorder with no clinically useful common disease drivers, biomarkers, or therapeutic targets being identified. With the premise of identifying universal and robust disease drivers and therapeutic targets, here, we interrogated a range of RTT transcriptomic studies spanning different species, models, and MECP2 mutations. A meta-analysis using RNA sequencing data from brains of RTT mouse models, human post-mortem brain tissue, and patient-derived induced pluripotent stem cell (iPSC) neurons was performed using weighted gene correlation network analysis (WGCNA). This study identified a module of genes common to all datasets with the following ten hub genes driving the expression: ATRX, ADCY7, ADCY9, SOD1, CACNA1A, PLCG1, CCT5, RPS9, BDNF, and MECP2. Here, we discuss the potential benefits of these genes as therapeutic targets. | Meta-Analysis Identifies BDNF and Novel Common Genes Differently Altered in Cross-Species Models of Rett Syndrome
Rett syndrome (RTT) is a rare disorder and one of the most abundant causes of intellectual disabilities in females. Single mutations in the gene coding for methyl-CpG-binding protein 2 (MeCP2) are responsible for the disorder. MeCP2 regulates gene expression as a transcriptional regulator as well as through epigenetic imprinting and chromatin condensation. Consequently, numerous biological pathways on multiple levels are influenced. However, the exact molecular pathways from genotype to phenotype are currently not fully elucidated. Treatment of RTT is purely symptomatic as no curative options for RTT have yet to reach the clinic. The paucity of this is mainly due to an incomplete understanding of the underlying pathophysiology of the disorder with no clinically useful common disease drivers, biomarkers, or therapeutic targets being identified. With the premise of identifying universal and robust disease drivers and therapeutic targets, here, we interrogated a range of RTT transcriptomic studies spanning different species, models, and MECP2 mutations. A meta-analysis using RNA sequencing data from brains of RTT mouse models, human post-mortem brain tissue, and patient-derived induced pluripotent stem cell (iPSC) neurons was performed using weighted gene correlation network analysis (WGCNA). This study identified a module of genes common to all datasets with the following ten hub genes driving the expression: ATRX, ADCY7, ADCY9, SOD1, CACNA1A, PLCG1, CCT5, RPS9, BDNF, and MECP2. Here, we discuss the potential benefits of these genes as therapeutic targets.
Rett syndrome (RTT) is one of the most common genetic causes of intellectual disabilities in females and affects one in 10,000 births [1]. RTT is an X-linked dominant disorder caused by mutations in the MECP2 gene, which encodes the Methyl-CpG Binding Protein 2 (MeCP2) protein. The molecular pathogenesis of RTT remains poorly understood, with patients presenting with numerous complex disabilities, which are likely due to the pleiotropic molecular functions of MeCP2 and its ubiquitous expression. At a cellular level, MeCP2 expression is critical for neuronal maturation and neuronal function. In hemizygous male patients, loss-of-function mutations in MECP2 cause neonatal encephalopathy, which is usually fatal before the age of two, and in heterozygous female patients it results in a severe neurological phenotype [2,3]. MeCP2 is a pleiotropic protein mediating early events of neurodevelopment, including neurogenesis, migration, and patterning, where it predominantly acts as a methyl-DNA binding protein and controls transcriptional regulation [4,5]. The protein is highly conserved among mammals showing 95% protein sequence identity between human and mouse, and to a lesser extent, between humans and zebrafish (48%). However, the genomic structure and expression patterns of MeCP2 in zebrafish and mammals are similar, suggesting probable conserved functions [6]. MeCP2 contains two highly evolutionarily conserved domains, the methyl-CpG-binding domain (MBD) and the transcriptional repression domain (TRD), and acts as a transcriptional repressor and activator. Upon binding to CpG islands, MeCP2 forms an inhibitory transcription complex through interactions of the transcriptional repression domain with cofactors, including Sin3A and histone deacetylase 1. MeCP2 also acts as a transcription activator to regulate gene expression by either long-range chromatin remodeling or by regulating RNA splicing [7,8,9]. MeCP2-deficiency leads to an excitation/inhibition (E/I) imbalance in the brain and is recognized as the leading cellular and synaptic hallmark of the disorder resulting in stereotypic hand movements, impaired motor coordination, breathing irregularities, seizures, and learning/memory dysfunctions [10]. Mice harbouring mutations in the Mecp2 gene represent one of the most clinically relevant models for RTT as they recapitulate many of the features observed in RTT patients, such as seizures and motor and cognitive dysfunction, which has assisted in our understanding of the underlying pathophysiology [11]. However, despite the vast majority of RTT patients being female, most gene therapy and other preclinical studies in animal models of RTT have used male mice, which is not truly representative of the patient population. Even though the phenotype of the RTT mouse models is very robust, there are many differences in brain development and structure between humans and mice that may confound findings in translational preclinical studies [12]. For example, the origin of cortical neurons in brain development differs in humans and mice with the subventricular zone, where human neurogenesis mostly occurs, that is significantly reduced in mice [13]. Thus, regardless of the significant insight gained from these models, inconsistencies between mouse models and human disease may affect the validity of preclinical findings. Immortalised cell lines and post-mortem brain tissue have also been used extensively to study the pathophysiology of RTT. However, the use of post-mortem brain tissue is limited, as the tissue only reflects end-stage disease and cannot be used in live-cell testing studies such as electrophysiology and immortalised cell lines do not represent the complex organisation of the brain [8]. More recently, stem cells, including human embryonic stem cells (hECSs) and induced pluripotent stem cells (iPSCs), have come to play an important role in in vitro disease modelling. hECSs are generated from early-stage human embryos and have the potential to differentiate into various cell types, whereas iPSCs are derived from patients and can be differentiated into any cell type [14]. Reprogramming of somatic cells to iPSCs through the overexpression of transcription factors was demonstrated over a decade ago [15] and this technology has now strengthened the utility of stem-cell-based disease models [16]. Over the past few years, several studies have successfully generated iPSC lines from RTT fibroblasts and have differentiated these lines into neural progenitor cells (NPCs), neurons, and glial cells [17]. Stem-cell-based modelling has been demonstrated to be effective for RTT research, because iPSC lines can harbour pathogenic MECP2 mutations and thus can demonstrate neuronal morphological defects, such as reduced dendritic branching, spine density, and smaller soma size [18,19]. Several studies have reported differentiated neuronal cells from RTT-iPSCs in two-dimensional (2D) cultures, with a smaller soma size compared to that of controls [14,20,21,22,23]. Additionally, the dysregulation in cellular maturation and morphological complexities in RTT-iPSC neurons have recapitulated the findings of mouse studies and in human post-mortem brain tissues [24]. The complexities of RTT at a clinical level and MeCP2 function have resulted in significant challenges for developing safe and effective therapies [25]. It is unclear whether novel therapies that have shown promising preclinical efficacy would effectively mitigate systemic manifestations of the disease when administered in the clinic. This is partly due to the lack of models that cover all aspects of the disease. Thus, well-characterised, disease-relevant models are critical to uncovering underlying molecular, cellular, and physiological intermediate phenotypes in the pathophysiology of RTT that may provide insights into potential therapies. Therefore, we hypothesise that by taking advantage of all existing models, both old and new (Figure 1), useful insights into the pathophysiology of RTT may be gleaned, which will drive the discovery of novel therapeutic targets. To do this, we conducted a meta-analysis of the transcriptomic data from three different RTT models: mouse brain, post-mortem human brain tissue, and iPSC-derived neurons. Weighted gene correlation network analysis (WGCNA) offers a powerful method to untangle novel disease pathways compared to approaches, such as differential gene expression. Thus, this study has utilised WGCNA to examine three previously published transcriptomic datasets of human post-mortem brain tissue, iPSC-derived neurons, and mouse brain samples. After identifying a consensus module between the three datasets, we analysed the genes in that module against another two datasets which could not be included in the WGCNA analysis, using differential gene expression.
Publicly available genome-wide transcriptomic datasets of iPSC-derived neurons, post-mortem human brain tissue, and mouse brains were retrieved from the NCBI Gene Expression Omnibus database. These included: GSE75303 (post-mortem), GSE123753 (iPSC-derived neurons) [26], and GSE96684 (mouse brain) [27] (Table 1). The post-mortem and mouse datasets included RTT and wild-type samples, whereas the iPSC-derived neurons included RTT and isogenic controls. The post-mortem dataset included sequencing results from both the temporal and frontal cortex. The age of the patients ranged from 17 to 20 years, and all subjects were female, harbouring three different mutations: c.378-2A > G, c.763C > T and c.451G > T. The mouse samples were from the brain cortex, and all were mouse nomenclature Mecp2 knockout males [18]. The iPSC-derived neurons were females harbouring a deletion between exons 3 and 4 of MECP2. All samples were included in this study. All datasets were normalised and filtered prior to WGCNA analysis (Figure 2). Briefly, abnormal samples were first filtered through hierarchical clustering, where any missing data count was eliminated. The genes in the mouse dataset were homologated to the human genome, with only the common genes being included in the study. Overall, there were a total of 9864 genes included in this analysis (Figure 2). Specific details on the normalisation used for the three different datasets can be found in Section 4.2.
Weighted gene co-expression networks were constructed based on the identified genes following the soft threshold analysis using all three datasets combined. An optimal soft-thresholding power is needed to calculate co-expression similarity. Hence, to assess the similarity between genes at the expression and network topology levels, we created a topological overlap matrix (TOM) which was achieved by calculating the adjacency and correlation matrices of the gene expression profile. As shown in Figure 3A in the scale free topology plot, power 8 was the lowest power where all three datasets reached a topology fit index of 0.9. Hence, it was chosen to produce the hierarchical clustering tree (dendrogram). Using the hierarchical average linkage clustering method in combination with the TOM, we proceeded to identify gene modules of each gene network. The dynamic tree cut algorithm highlighted all gene modules and each was identified by a colour (Figure 3B). Each tree branch constitutes a module, and each leaf in the branch is one gene.
The module–trait associations were analysed by correlating module–sample eigengenes with clinical traits to identify significant associations. The colours of all the modules were selected at random to distinguish between modules. Correlation coefficients were assigned to each module and the disease status trait (RTT vs. wild-type (WT)). Subsequently, only modules with a significant correlation to the disease trait (p < 0.05) were identified for all three datasets (Figure 4).
Through the identification of modules with a significant correlation coefficient to the disease trait across all models and tissues, four modules were found to be significantly dysregulated across all datasets (p < 0.5 eigenscore associated with disease trait): brown4, blue, magenta, and skyblue (Figure 4).
To better understand the biological functions of the genes in the four modules, each module was subjected to KEGG pathway enrichment analysis (Figure 5). The level of significance of each pathway enrichment was calculated and expressed in adjusted p-values using the Bonferroni correction method. We then focussed on those pathways that had higher adjusted p-values (depicted in yellow in Figure 5). The brown4 module was highly enriched in cytokine–cytokine receptor interaction, the TGF-beta signalling pathway, fluid shear stress, and atherosclerosis, and signalling pathways regulating pluripotency of stem cells. The blue module was highly enriched in pathways including ribosomes, COVID-19 disease, thermogenesis, and basal transcription factors. The magenta module was enriched in cysteines and methionine metabolism, biosynthesis of amino acids, biosynthesis of cofactors, glycosaminoglycan degradation, mucin-type 0-glycan biosynthesis, pantothenate and CoA biosynthesis, and 2-oxocarboxylic acid metabolism. Finally, the skyblue module was enriched in GABAergic synapses, morphine addiction, vibrio cholerae infection, thyroid hormone synthesis, purine metabolism, cortisol synthesis and secretion, alcoholism, ovarian steroidogenesis, glutamatergic synapsis, cholinergic synapse, the longevity regulating pathway, growth hormone synthesis, secretion, and action, inflammatory mediator regulation of transient receptor potential (TRP) channels and relaxin signalling.
Given the relevance to the known pathophysiology of RTT of the pathways identified in the skyblue module, we investigated this module further. Furthermore, focussing on the disease trait (WT vs. RTT), the skyblue module exhibited the highest correlation (Figure 4) and more disease-relevant enrichment (Figure 5). Therefore, this module was identified as a key module in RTT and was subjected to further analysis. Interestingly, we found that the skyblue module was driven by hub genes: MECP2, BDNF, SOD1, PLCG1, CCT5, RPS9, ADCY9, ADCY7, ATRX, and CACNA1A (Figure 6 and Table 2). Hub genes are defined as genes with connectivity (degree) greater than 10 in the genetic interaction network. All genes were shown to be downregulated in RTT except for CCT5 and ADCY9.
To broaden the study and determine whether the skyblue module genes were also dysregulated in other RTT studies, the expression profiles obtained from two other datasets were analysed. The MT dataset (GSE6955 [28]) consisted of post-mortem human brain tissue, and OH (GSE107399 [29]) of iPSC-derived neurons (Table 3). Due to the constrains of WGCNA analysis, these datasets could not be included in the original analysis as the OH dataset had less than six samples (considering each experimental replicate as one sample) and the MT dataset was analysed using single cell RNA sequencing. The MT dataset consisted of six post-mortem superior frontal gyri samples, two belonging to female RTT patients, and four non-RTT controls (Table 3). The OH dataset consisted of seven iPSCs-derived neuron samples, of which two were RTT patients analysed in replicate, and three were the corresponding isogenic controls. After performing differential gene expression analysis, 12,625 genes were identified in the MT dataset, of which 156 were significantly upregulated and 61 significantly downregulated (p < 0.05). Conversely, analysis of the OH data demonstrated 20,055 differentially expressed genes. Of these, 655 were significantly upregulated and 992 were significantly downregulated (p < 0.05). Next, the gene expression profiles of MT and OH were cross-referenced with those genes in the skyblue module to identify common dysregulated genes across all datasets. Overall, there were 71 genes shared between the skyblue module and the MT dataset. TOX3, FABP7, ATRX, and SGMS1 had the largest positive log-fold changes (Figure 7A). BDNF, GNG11, FAM168B, HMOX1, and VCL were identified as having the largest negative log-fold changes (Figure 7A). A comparable number of genes (107) were common between the skyblue module and the OH dataset. The top five upregulated genes with the greatest positive log-fold change were NPAS4, FABP7, HECW2, TOX3, and CACNA1A (Figure 7B). Conversely, FREM2, BDNF, HMOX1, KDELR2, and CXADR had the highest negative log-fold changes (Figure 7B). The expression of the hub genes identified in the skyblue module was also cross-examined in the MT and OH datasets (Figure 8). From the meta-analysis, CCT5 and ADCY7 were upregulated in the skyblue module, whilst the remaining eight (MECP2, BDNF, CACNA1A, ADCY9, ATRX, RPS9, SOD1, and PLCG1) were downregulated. When compared with the other datasets, ATRX was significantly upregulated (p < 0.05) in the MT dataset and CACN1A was upregulated in the OH dataset. Interestingly, BDNF was significantly downregulated in all datasets (Figure 8A,B).
The overarching aim of this study was to identify common pathways and genes that intersect RTT transcriptomic studies spanning different species and models with the premise of identifying universal and robust disease drivers and therapeutic targets. To do so, a meta-analysis and bioinformatics approach consisting of the identification of gene modules rather than differential gene expression was employed to interrogate the transcriptomic landscape of RTT using human post-mortem brain tissue, mouse models, and patient-derived neurons. After identifying the statistically significant dysregulated modules between all the RTT samples and controls, the module with the highest correlation to disease status and genes with the highest connectivity within the module were interrogated to identify key genetic drivers across all tissue samples and models. Reassuringly, the identified hub genes included MECP2 and BDNF, where the correlation between the two genes in RTT is well recognized, with BNDF being a well-established target gene of MeCP2 [30,31].
Through this meta-analysis four modules of genes that were significantly dysregulated in the RTT transcriptome relative to the controls was identified. The pathways that were enriched in each of the four modules were investigated and the brown4 module was identified to be mostly enriched in pathways related to immunological aberrations, which is consistent with previously published studies in RTT, including one of our own [16,25,26]. Next, the magenta module was enriched for pathways primarily involving the metabolic system, which also aligns with previously reported literature [32,33]. On the other hand, enrichment of the blue module did not produce any pathways known to be of relevance in RTT. The fourth module, skyblue, consisted of enriched modules, including glutamatergic, GABAergic, and cholinergic synaptic pathways, as well as protein export, and was identified to have the most enriched pathways relevant to the neuropathology of RTT, hence supporting our further focus on this module.
The ten hub genes that were identified as the main drivers of the skyblue module were ATRX, ADCY7, ADCY9, SOD1, CACNA1A, PLCG1, CCT5, RPS9, BDNF, and MECP2. They are therefore surmised to play key roles in the pathology of RTT and may assist in understanding the underlying disease pathophysiology, as well as identifying disease drivers and drug targets. ATRX (ATRX Chromatin Remodeler) has recently been implicated in RTT as a binding partner of MeCP2 where together they modulate pericentric heterochromatin (PCH) organization in neurons [34]. Mutations in ATRX cause ATR-X syndrome, implicated in abnormal brain development and associated with severe intellectual disability [34,35]. The downregulation of ATRX in this meta-analysis supports previous reports of an interaction with MeCP2, where MeCP2 recruits the helicase domain of ATRX to heterochromatic foci in a DNA methylation dependent manner, as shown in living mouse cells [36]. Furthermore, it has been shown that the heterochromatin location of ATRX is disrupted in Mecp2-null mice neurons. These data together suggest that a MeCP2–ATRX interaction leads to pathological changes that contribute to the mental retardation phenotype. Interestingly, as an epigenetic modifier, ATRX has been implicated in cancer and has received a level of attention in the identification of expression modifying drugs [37]. ATRX loss leads to increased DNA damage and general genomic instability [38], and thus drugs or small molecules aimed at increasing the stability of the genome may be potential therapeutic options for RTT. Adenylate Cyclases 7 and 9 (ADCY7 and ADCY9) are membrane-bound enzymes that catalyze the formation of cyclic AMP from ATP and are highly expressed in the brain. De novo mutations in ADCY7 have been reported in autism spectrum disorders (ASD) where the gene has been proposed to be a risk factor [39]. ASD and RTT share some commonalities with RTT individuals showing some ASD-like behaviors [40,41]. ADCY7 mRNA is highly expressed in microglia and plays an important role in presynaptic GABA release, and evidence suggests that ADCY7 is involved in mood regulation and plays an essential role in the immune response [42]. Conversely, despite ADCY9 being highly expressed in the brain, its function in the CNS remains largely unknown. However, some findings have suggested that ADCY9 may regulate cognitive function and learning and memory [42]. Interestingly, ADCY9 has been shown to be downregulated in Mecp2 null embryonic cortexes, suggesting ADCY9 as a target of MeCP2 [43]. This effect is lost postnatally, suggesting the crucial role of ADCY9 in embryogenesis [36,43]. Interestingly, both genes are involved in two common pathways: the GPER1 signaling and integrin pathway, which provides potential therapeutic targets to explore in RTT. SOD1 plays a crucial role in the oxidative stress response and systemic redox alterations, and the related oxidative stress is well reported in RTT [44]. It is therefore not surprising to find the free radical scavenger SOD1 enzyme downregulated in this meta-analysis. Loss of SOD1 has been hypothesized to result in an accumulation of mitochondrial reactive oxygen species, leading to oxidative damage and mitochondrial dysfunction [45]. Animal studies have suggested a possible direct correlation between Mecp2 mutations and increased ROS levels, and the debate continues regarding whether oxidative stress is a cause or consequence of RTT. The voltage-dependent P/Q-type calcium channel subunit alpha-1A (CACNA1A) gene has been implicated in epileptic encephalopathy, familial hemiplegic migraine, episodic ataxia, and spinocerebellar ataxia [46] and has recently been reported in a small number of atypical Rett patients previously lacking known genetic mutations [47]. Voltage-sensitive calcium channels mediate the entry of calcium ions into excitatory neurons and are also involved in a variety of calcium-dependent processes and neurotransmitter release. Our findings suggest that the downregulation of CACNA1A in this meta-analysis may be contributing to the epileptic encephalopathy of RTT. Among its related pathways are the CREB and integrin signaling pathways. PLCG1 (Phospholipase C, gamma 1) is a protein involved in cell growth, migration, apoptosis, and proliferation. Among its related pathways is theBDNF-TrkB signaling pathway. Even though no direct link to MECP2 has been reported in the literature, it is known that activation of the neurotrophin receptor TRKB by BDNF triggers downstream PLCG1 signaling [48]. No direct relation has been reported between CCT5 and RPS9 and MECP2. However, CCT5 is implicated in the cellular pathways related to trafficking to the periciliary membrane and cell cycle and has also been linked to intellectual disabilities and early onset motor neuropathies [49,50,51]. On the other hand, RPS9 is linked to RNA binding and is a structural constituent of the ribosome, and as ribosomal dysfunction has been previously reported in RTT iNeurons by Rodrigues et al., 2020 [26], the dysregulation of RPS9 in this study supports these findings and provides further evidence of ribosomal dysfunction in RTT. In addition, common cellular pathways, such as the CREB and integrin signaling pathways, are common amongst the hub genes. The CREB pathway has previously been reported to be implicated in RTT, where overexpression of CREB signaling in RTT forebrain neurons rescued the phenotype of neurite growth, dendritic complexity, and mitochondrial function [52]. Furthermore, pharmacological activation of CREB in female RTT mice rescued several behavioral phenotypes [52]. These findings support the motion to investigate the CREB pathway as a potential therapeutic target [52]. In addition, while the integrin pathway has not been reported in RTT, it has been previously implicated in dendritic development, autism spectrum disorder, and intellectual disabilities [53] suggesting that this pathway too could also be a potential target for future RTT therapeutics.
Through this study, three synaptic pathways enriched in the skyblue module were identified, namely the cholinergic, glutamatergic, and GABAergic pathways. A loss of excitation/inhibition (E/I) balance in the neural circuit is a major hallmark of RTT pathology, causing many neurological symptoms, such as loss of purposeful hand movements, impaired motor coordination, breathing irregularities, and seizures, amongst others [10]. This loss of E/I balance is caused by MeCP2 deficiency, leading to a dysregulation of the glutamatergic and GABAergic pathways. Furthermore, downstream genes affected in RTT such as BDNF play an important role influencing neurotransmission activity. Many drugs have been tested to improve the E/I balance in RTT, including glutamatergic modulators such as AMPAkines to increase excitatory synapsis and enhance BDNF expression, ketamine, and NMDAR antagonist to enhance neuronal activity [54,55]. GABAergic modulators have also shown potential in aiding with behavioral dysfunction in RTT patients and mice. However, while respiratory alterations were ameliorated by treatment using benzodiazepines and Midazolan in mice, the phenotype was not fully rescued [56].
A comparison of differentially expressed genes in the MT and the OH datasets with the skyBlue module identified 71 and 107 commonly expressed genes, respectively. Interestingly, TOX3 was upregulated in both datasets and the skyblue module. TOX3 plays a role in shaping DNA and altering chromatin structure and while the protein has been shown to be a neuron survival factor [57], it is yet to be linked with neurodevelopmental disorders and specifically to RTT. BDNF and HMOX1 were also commonly dysregulated in all the datasets where they were observed to be significantly downregulated. HMOX1 is a heme oxygenase responsible for the degradation of heme to biliverdin/bilirubin and free iron and heavily implicated in aging and disease. The expression of HMOX1 is confined to small populations of neurons and glia and is upregulated by a wide range of pro-oxidant and other stressors [58]. While there have been no reports linking HMOX1 to RTT pathology, its downregulation confirms the role of oxidative stress in the pathology of RTT. ATRX and CACNA1A were identified to be dysregulated in the skyblue module as well in either the MT or OH datasets where the expression of ATRX was upregulated in the MT dataset and CACNA1A was upregulated in OH. Additionally, BDNF, an identified hub gene for skyblue, was also downregulated in both the MT and OH datasets. This is the first time BDNF has been demonstrated to be consistently downregulated in a bioinformatic meta-analysis examining dysregulated genes across species and models.
Of the ten identified hub genes in the meta-analysis, eight were downregulated, suggesting that wild type MeCP2 transcriptionally activates these genes, and two (ADCY9 and CCT5) were upregulated, suggesting that MeCP2 transcriptionally represses these genes. From the differential gene expression analysis performed on the OH and MT datasets, we showed that BDNF was downregulated in both studies, and ARTX and CACN1A were upregulated in the MT and OH datasets, respectively. We showed an overall trend of upregulation in the five tested genes ATRX, ADCY7, ADCY9, CACNA1A, and SOD1. These results were different to that found in the meta-analysis as only ADCY9 was upregulated. This disparity in expression between the meta-analysis and differential gene expression points to the complexity of RTT and the context dependent expression of MECP2. Here, we have used two different analytical tools (WGCNA and DGE), two species, and three models to identify dysregulated genes that drive the disease pathology of RTT. The identification of BDNF as the only consistent gene to be downregulated relative to controls across all models comes as no surprise given the known association with RTT. BDNF has been explored as a therapeutic target for RTT. However, as BDNF has a low blood–brain barrier permeability, this limits its bioavailability for peripheral administration as a therapy [59]. Three clinical trials aimed at augmenting BDNF expression, trialing Copaxone (glatiramer acetate) [60] and Fingolimod, have been conducted [61]. However, to date, no therapies have entered the clinic, with the glatiramer acetate trial being withdrawn due to reported potential life-threatening reactions [62]. Additional compounds have been described to increase BDNF levels and improve RTT-like symptoms in mice, however, none have reached human clinical trials, alluding to the complexity of the disorder and difficulty of this approach [59].
The three datasets included in the WGCNA analysis were obtained from the NCBI Gene Expression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/ (last accessed on 20 June 2022)): GSE75303 (Post-mortem), GSE123753 (iPSC-derived neurons), and GSE96684 (Mouse Brain). The GSE75303 dataset contained 12 samples in total, including three female RTT patient frontal and temporal cortexes harboring mutations at c.378-2A > G, c.763C > T and c.451G > T and three female age-matched controls. The GSE123753 dataset consisted of six female samples: three patients involving rearrangements that removed exons 3 and 4, creating a functionally null mutation, and their three corresponding isogenic controls. The GSE96684 dataset consisted of eight male mouse samples: four MECP2 knockout and four wild type mice. The characteristics of the samples in each dataset are summarized in Table 1.
Since the three datasets were from different sequencing platforms, we performed pre-processing according to a previously published WGCNA pipeline. Specifically, GSE75303 (human post-mortem brain) array data were quantile normalized, GSE123753 (iPSCs derived neurons) provided the already quantile normalized data as a supplementary file through GEO (Gene Expression Omnibus), and the raw sequencing data were obtained for GSE96684 from SRA (PRJNA3779366) and mapped to mm10 using STAR (Version 2.7). Mouse gene symbols were mapped to human gene symbols using biomart as an R package. It is important to also note that one requirement of WGCNA with multiple models is that the same list of genes for each tissue sample should be used in the analysis, which may lead to missing the information on genes that are not represented in every sample. Briefly, raw counts and probe intensity data were pre-processed using the Limma package [63] in the R environment. Count data were transformed by mean-variance modelling at the observational level (voom) [64] before all studies were subjected to quantile normalization and data quality control as recommended for WGCNA. Finally, for differential gene expression analysis, raw count data were voom transformed and array data were log-transformed in the Limma package [63] in the R environment, followed by quantile normalization and data quality control.
Unsigned co-expression networks were built using the WGCNA 1.63 package in R software [65]. Clusters of genes that behaved similarly were grouped together into different color modules. These modules were related to specific traits. In heatmaps, red represents genes upregulated within that dataset and green represents genes downregulated within that dataset. The top 1000 connections within a gene network were determined by WGCNA. For the multiple array consensus analysis, WGCNA was performed on the individual datasets first, as suggested by Langfelder and Horvath’s tutorial [65], using “1 step function for network construction and detection of consensus modules”. The default WGCNA soft thresholding power β in which co-expression was raised was chosen to calculate the adjacency of each data set. The soft thresholding power β was used to allow us to compare each data set by approximate scale-free topology, thus compensating for scale differences between data sets.
The correlation between module eigengenes and clinical traits was analyzed to identify modules of interest that were significantly associated with clinical traits. For the purpose of this study, we identified the modules that were significantly correlated with disease status in all three datasets. The correlation values were then displayed within a heatmap. Gene significance (GS) was defined as the correlation between gene expression and each trait. In addition, module membership (MM) was defined as the association between gene expression and each module eigengene. Subsequently, the correlations between GS and MM were examined to verify certain module–trait associations. The correlation analyses in this study were performed using Pearson correlation as described in the WGCNA package [65].
The genes in each module of interest were extracted from the network and enrichment analysis was performed to further explore the functions of the respective modules. The R package ‘clusterProfiler’ was used to perform Kyoto Encyclopedia of Genes and Genomes (KEGG) [66,67] pathway enrichment analysis. A statistical p value of <0.05 was set as the significance threshold, and the enrichment results of KEGG pathways in each module of interest module were obtained.
The intramodular connectivity of genes in the corresponding modules of interest was measured using module eigengene-based connectivity (kME). The top 30 genes of each module of interest, which represent the central status in the module gene network, were selected to visualize the subordinate module using String software [68]. Subsequently, one key module was chosen that exhibited the highest levels of positive or negative correlation with RTT to search for hub genes for RTT in the modules. The top ten genes with the highest kME were selected as the hub genes in the corresponding module [65] and their gene significance (GS) for RTT (disease status) and intramodular connectivity kME were determined to confirm the reliability of these hub genes.
Differential gene expression analysis was performed on six patient-derived datasets from the MT study (GSE6955) and seven iPSC-derived neuronal samples from the OH study (GSE107399). Samples in the MT study were taken from the superior frontal gyrus of patients with RTT and age-matched controls. From the OH study, seven samples were utilized for analysis. Of these, four were RTT mutants (including two experimental controls) and three isogenic controls. The datasets were analyzed in R using the EdgeR package (R Bioconductor). Firstly, genes with low expression and a CPM value ≤ 1 were filtered. Then, the remaining counts were used to generate linear models and statistical analysis was conducted. To identify overlapping differentially expressed genes from the OH and MT datasets corresponding to the skyblue module, the log fold change was noted for genes that overlapped were identified.
Through this meta-analysis and sub-analysis of datasets belonging to a mouse model, using post-mortem brain and iPSC-derived neurons to identify dysregulated genes that underpin the RTT pathology, a set of genes common to all models were identified. Some genes, such as BDNF, ADCY9, ATRX, and CACNA1A, have previously been linked to RTT, while others, including CCT5, RPSP, and PLCG1, are potential disease-modifying genes. Interestingly, a previous transcriptomic study using DGE only in RTT human samples did not cidentify some of the molecular network hub genes identified in the current study [69]. Validating BDNF, a known target of MeCP2, demonstrates the utility of this bioinformatic approach in identifying therapeutic genes targets. Further exploration of these known and novel disease-modifying genes may provide a better understanding of the molecular mechanisms of RTT and pave the way for the investigation of novel therapeutic candidates. | true | true | true |
PMC9570364 | 36233277 | Xue Yong,Tangchun Zheng,Yu Han,Tianci Cong,Ping Li,Weichao Liu,Aiqing Ding,Tangren Cheng,Jia Wang,Qixiang Zhang | The miR156-Targeted SQUAMOSA PROMOTER BINDING PROTEIN (PmSBP) Transcription Factor Regulates the Flowering Time by Binding to the Promoter of SUPPRESSOR OF OVEREXPRESSION OF CO1 (PmSOC1) in Prunus mume | 09-10-2022 | Prunus mume,SQUAMOSA PROMOTER BINDING PROTEIN,miR156,SUPPRESSOR OF OVEREXPRESSION OF CO1,flowering time | Prunus mume, a famous perennial ornamental plant and fruit tree in Asia, blooms in winter or early spring in the Yangtze River area. The flowering time directly determines its ornamental and economic value, so it is of great significance to study the molecular mechanism of flowering time. SQUAMOSA PROMOTER BINDING PROTEIN (SBP), often regulated by miR156, is an important flowering regulator, although its function is unknown in P. mume. Here, 11 miR156 precursors were analyzed and located in five chromosomes of the P. mume genome. The expression pattern showed that PmSBP1/6 was negatively correlated with miR156. The promoters of PmSBP1/6 were specifically expressed in the apical meristem. Overexpression of PmSBP1/6 in tobacco promoted flowering and changed the length ratio of pistil and stamen. Moreover, PmSBP1 also affected the number and vitality of pollen and reduced the fertility of transgenic tobacco. Furthermore, ectopic expression of PmSBP1/6 caused up-regulated expression of endogenous SUPPRESSOR OF OVEREXPRESSION OF CO1 (NtSOC1). The yeast-one hybrid assay showed that PmSBP1 was bonded to the promoters of PmSOC1s. In conclusion, a miR156-PmSBP1-PmSOC1s pathway was formed to participate in the regulation of flowering time in P. mume, which provided references for the molecular mechanism of flowering time regulation and molecular breeding of P. mume. | The miR156-Targeted SQUAMOSA PROMOTER BINDING PROTEIN (PmSBP) Transcription Factor Regulates the Flowering Time by Binding to the Promoter of SUPPRESSOR OF OVEREXPRESSION OF CO1 (PmSOC1) in Prunus mume
Prunus mume, a famous perennial ornamental plant and fruit tree in Asia, blooms in winter or early spring in the Yangtze River area. The flowering time directly determines its ornamental and economic value, so it is of great significance to study the molecular mechanism of flowering time. SQUAMOSA PROMOTER BINDING PROTEIN (SBP), often regulated by miR156, is an important flowering regulator, although its function is unknown in P. mume. Here, 11 miR156 precursors were analyzed and located in five chromosomes of the P. mume genome. The expression pattern showed that PmSBP1/6 was negatively correlated with miR156. The promoters of PmSBP1/6 were specifically expressed in the apical meristem. Overexpression of PmSBP1/6 in tobacco promoted flowering and changed the length ratio of pistil and stamen. Moreover, PmSBP1 also affected the number and vitality of pollen and reduced the fertility of transgenic tobacco. Furthermore, ectopic expression of PmSBP1/6 caused up-regulated expression of endogenous SUPPRESSOR OF OVEREXPRESSION OF CO1 (NtSOC1). The yeast-one hybrid assay showed that PmSBP1 was bonded to the promoters of PmSOC1s. In conclusion, a miR156-PmSBP1-PmSOC1s pathway was formed to participate in the regulation of flowering time in P. mume, which provided references for the molecular mechanism of flowering time regulation and molecular breeding of P. mume.
Flowering time is a very important agronomic trait, which directly determines the ornamental value and economic benefits of ornamental plants. Plant flowering marks the end of vegetative growth and the beginning of reproductive growth, which is a very complex and important link in the process of plant growth and development. It is not only affected by external environmental factors (light, temperature, etc.) but also regulated by internal physiological factors. So far, six pathways regulating the flowering time of plants have been identified. These pathways include environmental factor pathways: photoperiod pathway, vernalization pathway, and temperature pathway; and the physiological factor pathways: autonomous pathway, gibberellin pathway, and aging pathway. Flowering time is jointly regulated by these pathways, which can affect each other and integrated by a series of genes, such as SUPPRESSOR OF OVEREXPRESSION OF CO1 (SOC1), LEAFY (LFY), FLOWERING LOCUS T (FT), CONSTANS (CO), FLOWERING LOCUS C (FLC), etc. [1]. The aging pathway is based on the phenomenon that plants must grow to a certain age before the commencement of reproductive growth. The aging pathway is mainly regulated by microRNA156 (miR156) and its target gene SQUAMOSA PROMOTER BINDING PROTEIN-LIKE (SPL), which was also named SQUAMOSA PROMOTER-BINDING PROTEIN (SBP). The expression of miR156 and SPL genes is negatively correlated. For example, from the juvenile to adult phase of Arabidopsis thaliana, the expression level of miR156 was gradually decreased, while the expression level of its target gene AtSPL, increased with the increase of age, and reached a high level in adulthood [2,3,4]. In addition, their expression patterns were conservative in many species [5,6,7,8]. The miR156-SPL mode plays an important role in the flowering process of plants. In Arabidopsis, constitutive expression of miR156 could prolong the vegetative growth, and overexpression of its target gene SPL resulted in early flowering [3,4,9,10,11]. Furthermore, the aging pathway can mediate photoperiod, gibberellic acid, and endogenous flowering pathways, and co-regulate plant flowering by affecting flowering integrators, such as SOC1 [12,13], CO [14], FT [2,15], AP1, FUL, and LFY [11,16]. However, AtSPL8, a member of SBP-box genes in Arabidopsis, does not affect the flowering time, but affects the development of the flower organs and reduces the fertility of plants [17]. Mei (Prunus mume), which belongs to Prunus in Rosaceae, is an important perennial woody ornamental plant and fruit tree. P. mume is widely spread in temperate regions of Asia and has been cultivated for more than 3000 years in China. The early flowering determines the ornamental, economic, and cultural value of Mei. So far, some flowering-related genes in P. mume have been studied, such as PmSOC1 [18], PmSVP [19], and PmLFY [20]. The SBP-box genes have been identified in P. mume [21]. However, the molecular mechanism of PmSBPs in flowering regulation is still not clear. Here, two members of the SBP-box gene family (PmSBP1 and PmSBP6) were cloned, and their expression patterns, gene function, and regulation mechanism were studied. The outcomes, thus, lay a foundation for clarifying the molecular mechanism of flowering time regulation of P. mume and other Prunus species.
In P. mume, the precursors of pmu-miR156 were encoded by 11 genomic sequences (pmu-MIR156a-k), which were located in five chromosomes (Supplementary Figure S1A) and cut into four types of mature sequences (pmu-miR156a–d, e–f, g–j, k). An ML phylogenetic tree of 36 precursor sequences (11 from P. mume, 10 from Arabidopsis, and 15 from tobacco) was built and the miR156 family was divided into three subgroups (Supplementary Figure S1B). Among them, the pmu-MIR156i/j/g/h were clustered with ath-MIR156a–f and nta-MIR156c–i, and the pmu-MIR156a/b/c/d/f/k were clustered with ath-MIR157a–d and nta-MIR156a/b/j–o, but the pmu-MIR156e was a subgroup alone. Sequence alignment showed that only two differentiated bases were found between the miR156 mature sequences from P. mume, Arabidopsis, and tobacco (Supplementary Figure S1C). The pmu-miR156f can fully bind to the miR156 site on the mRNA of PmSBP1, and they are located on the same chromosome.
The flower development period of P. mume consists of two phases with overlapping parts: flower bud morphological differentiation and dormancy, which were divided into nine stages (S1–S9): flower primordium forming stage (S1), sepal forming stage (S2), petal forming stage (S3), stamen forming stage (S4), pistil forming stage (S5), anther and ovule forming stage (S6), morphological differentiation finished stage (S7), alabastrum intumescence stage (S8), and flower upcoming to bloom stage (S9) (Figure 1A,B). Among them, S1–S7 belong to the flower bud morphological differentiation phase, and S6–S9 belong to the flower bud dormancy phase. To verify the function of PmSBP1/6 and the regulation relationship with miR156f, their expression patterns in flower development were analyzed. The expression level of PmSBP1 gradually increased (S1–S4) and remained stable (S4–S6), but decreased sharply when flower bud morphological differentiation was completed (S7), then recovered slightly at S8 and then maintained a very low level before blooming (S9) (Figure 1C). The expression level of PmSBP6 increased first (S1–S2) and decreased gradually from S2 to S4, then suddenly raised at pistil forming stage (S5) and then dropped when flower bud morphological differentiation was completed (S7), and maintained at a low level (S8–S9) (Figure 1D). In general, PmSBP1/6 shows relatively high expression in the flower bud morphological differentiation phase, but low expression after differentiation. On the other hand, the expression level of pmu-miR156f decreased from S1 to S4, but suddenly increased at S5, then fell back and remained at a comparatively stable level (S6–S9). In conclusion, the expression patterns of PmSBP1/6 and pmu-miR156f were negatively correlated.
To verify the activity of the PmSBP1/6 promoters, the 2000 bp upstream sequences of PmSBP1/6 were cloned and fused with the β-glucuronidase (GUS) gene and transiently transformed into the N. benthamiana leaves by A. tumefaciens. Finally, clear blue spots were observed in the tobacco leaves with PmSBP1/6 promoters (Figure 2A,B) compared with the control (Figure 2C,D). For further study of the tissue spatial expression profile of PmSBP1/6 promoters, the promoters with GUS were stably overexpressed in tobacco. In the PmSBP1-promoter transgenic tobacco, obvious blue tissues were detected in the apical meristem (Figure 2E). While in the PmSBP6-promoter transgenic tobacco, the blue tissues were detected in both the apical meristem and lateral bud (Figure 2F).
The promoter cis-elements can affect the gene expression. The cloned promoter sequences of PmSBP1/6 (about 2000 bp) were predicted on the PlantCARE website (Supplementary Table S2 and Supplementary Data S1). Statistical results of cis-elements showed that gibberellin response elements, anaerobic induction elements, and light response elements were contained in the promoters of both PmSBP1 and PmSBP6 (Figure 2G,H). In addition, the promoter of PmSBP6 also contained meristem expression elements (Figure 2H). To analyze whether the expression of PmSBP1/6 is regulated by gibberellin, the qRT-PCR was carried out with the flower buds treated with 100 mg/L GA3 solutions. The expression patterns of PmSBP1 and PmSBP6 are shown in Figure 2I,J. The expression level of PmSBP1 and PmSBP6 significantly decreased after gibberellin treatment for 48 h and 2 h, respectively (Figure 2I,J).
The subcellular localization of PmSBP1/6 was detected by the transient transformation of the PmSBP1/6-GFP fusion proteins into tobacco leaves. In the control 35S::GFP, the GFP fluorescence signals were examined in the cytoplasm and nucleus. In the 35S::PmSBP1/6-GFP, although the GFP fluorescence signals were detected mainly in the nucleus and cytoplasm, the fluorescence intensity is weaker than the control (Figure 3).
The PmSBP1, PmSBP1tb, and PmSBP6 were overexpressed in tobacco under CaMV35S. To confirm the function of PmSBP1 and miR156, the miR156 site of PmSBP1 was synonymously mutated and named PmSBP1tb. The schematic map of PmSBP1/1tb recombinant vectors is shown in Figure 4A. More than 10 transgenic lines of each gene were obtained and confirmed by a reverse transcription polymerase chain reaction (RT-PCR) assay, and three lines were selected for subsequent analyses. In the juvenile period of transgenic T3 generation plants, the vegetative growth of the 35S::PmSBP1tb transgenic seedlings was significantly less than that of the wild type (WT) (Figure 4C), but the 35S::PmSBP1 transgenic lines were slightly larger than that of the WT (Figure 4B). Besides, the vegetative growth of the three 35S::PmSBP6 transgenic lines was also a little larger than that of the WT (Figure 5A). As plants grew older, the difference in vegetative growth between all the transgenic seedlings and wild type gradually disappeared (Figure 4D,E and Figure 5B). The flowering time of transgenic plants was different when they under different photoperiod conditions. In short-day conditions, both PmSBP1 (Figure 4F), PmSBP1tb (Figure 4G), and PmSBP6 (Figure 5C,D) transgenic plants bloomed earlier than WT. In addition, the flowering time of PmSBP1 transgenic plants was earlier than that of PmSBP1tb. While under the long-day conditions, the flowering time of PmSBP1 (Figure 4H) and PmSBP6 (Figure 5E, F) transgenic lines was similar to that of WT (Figure 4I). Interestingly, the flowering time of the PmSBP1tb transgenic lines was significantly earlier than that of WT (Figure 4J). In the reproductive growth stage, the 35S::PmSBP1tb transgenic tobacco was normal, but the 35S::PmSBP1 and 35S::PmSBP6 transgenic tobacco were different from the wild-type tobacco. In the wild-type tobacco, the length of style and filaments were the same, allowing them to be self-bred and bear fruit (Figure 5G). While in the 35S::PmSBP1 (Figure 6A) and 35S::PmSBP6 (Figure 5H) transgenic tobacco, the length of style and filaments were inconsistent. In the 35S::PmSBP1 transgenic tobacco, the length ratio of style to filament showed two phenotypes: one was the high-style (the style was higher than the filament) and the other was the short-style (the style was shorter than the filament). Among them, the high-style flower was the main phenotype (96%), with smaller size, withered anthers, less pollen, and fast-drying filaments (Figure 6A,B), which could not be self-pollinated. The pods became smaller and the seed number in one pod was less after artificial self-pollination (Figure 6C). The pollen viability of the high-style flower was tested by TTC staining and pollen germination in vitro. As shown in Figure 6D and Supplementary Figure S2, in the wild-type tobacco and three 35S::PmSBP1 transgenic lines S1-3, S1-5, and S1-6, the pollen staining rates were 85.8%, 41%, 32.5%, and 13.8%, respectively, and the pollen germination rates were 78.8%, 18.2%, 13.7%, and 12.5%, respectively. In short, the results of the two methods showed that the pollen viability of 35S::PmSBP1 transgenic plants was lower than that of WT. The flower phenotype in the 35S::PmSBP6 transgenic tobacco was partly different from the 35S::PmSBP1 transgenic tobacco. In the 35S::PmSBP6 transgenic tobacco, all the flowers showed a high-style phenotype with normal size, but the flowers could bear fruit normally after artificial self-pollination (Figure 5H).
To investigate the early flowering transgenic tobacco, the expression level of eight endogenous flowering-related genes (NtSOC1, NtCO, NtAP1, NtMADS3, NtMADS4, NtMADS11, NtNFL1, and NtNFL2) in the shoot tip of tobacco (90 d in short-day conditions, 100 d in long-day conditions) was detected by qRT-PCR. The expression level of these eight endogenous flowering-related genes in the three PmSBP1 transgenic tobacco lines is shown in Figure 7A. In short-day conditions, their expression was higher than that of the wild-type tobacco; while in long-day conditions, the expression level was lower than that of wild-type tobacco, except for NtSOC1 showing similar expression to wild-type tobacco. As shown in Figure 7B, the expression level of the endogenous NtSOC1 in the PmSBP1tb transgenic tobacco was increased under both short-day and long-day conditions. In short-day conditions, the expression of the other seven flowering-related genes was similar to WT. In long-day conditions, the expression of NtCO only increased sharply in the PmSBP1tb transgenic tobacco line S1tb-7; however, it changed slightly in the other two lines. The expression level of NtAP1, NtMADS3, NtMADS4, and NtMADS11 in all three lines decreased; however, the expression of NtNFL1 and NtNFL2 fluctuated marginally. As shown in Figure 8, the expression level of these eight flowering-related genes in the three PmSBP6 transgenic lines under short-day conditions was higher than that of WT. Under long-day conditions, their expression levels in the three lines fluctuated up and down at the relative expression value of WT. The expression level of NtSOC1 was directly proportional to the phenotype of early flowering in these transgenic plants.
To verify the regulatory relationship between PmSBP1 and PmSOC1s, the promoter sequences of PmSOC1s (about 850 bp) were cloned and analyzed (Supplementary Data S1). Among them, except for PmSOC1-1, the promoters of PmSOC1-2 and PmSOC1-3 contained four SBP binding sites ‘GTAC’ respectively (Figure 9A). The 100 bp fragments of PmSOC1s promoters containing SBP binding sites were used as baits. The B1 fragment of PmSOC1-2 (C2B1) and the B2 fragment of PmSOC1-3 (C3B2) contained two and three SBP binding sites, respectively. The interactions between PmSBP1 and the fragments of PmSOC1s promoters were detected by yeast one-hybrid (Figure 9B). All transformed yeast grew normally on the SD/-Leu solid medium. As shown in Figure 9B, PmSBP1 could bind to the B2 fragments of PmSOC1-2 (C2B2) and PmSOC1-3 promoters (C3B2) but could not associate with the other fragments (C2B1, C2B3, and C3B1). This result suggested that PmSBP1 could activate PmSOC1-2 and PmSOC1-3 by binding to the B2 sites of their promoters. Furthermore, we compared the eight SBP-binding fragments containing the core ‘GTAC’ (11 bp) with the SBP-binding sequence of Arabidopsis AtSPL9 (ID: MA1322.1, homologous with PmSBP1), and found that the three bases adjacent to the core ‘GTAC’ in C2B2 and C3B2 were consistent with the high-frequency bases in the SBP-binding sequence of AtSPL9 (Figure 9C), which indicated that the three bases adjacent to ‘GTAC’ were critical for the binding with PmSBP1.
Plant microRNAs (miRNAs), 19–25 nt long, are highly conserved small non-coding RNAs, and play important role in juvenile-to-adult phase transition and flowering time in Arabidopsis [22,23]. The miR156 is quite conservative among plant species [24]. In this study, the sequence alignment results showed that the miR156 mature sequences in P. mume, Arabidopsis and tobacco were indeed highly conserved, with only two SNPs (Supplementary Figure S1A). The miR156-SPL is the famous age pathway; SPL is the target gene of miR156, and is regulated by miR156 mainly through complementation base-pairing at the post-transcriptional level [23]. In our previous study, the miR156-mediated PmSBPs mRNA cleavage was detected by 5′-RACE [25]. In Arabidopsis, there are 16 AtSPLs, which are clustered into two clades: clade I contains two subclades AtSPL7 and AtSPL1/12/14/16; clade II contains four subclades AtSPL3/4/5, AtSPL2/10/11, AtSPL9/15/6/13, and AtSPL8, and all members in clade II have miR156/157 sites except for AtSPL8 [26]. In this study, PmSBP1 were clustered with the AtSPL9/15, and PmSBP6 were clustered with AtSPL3/4/5 together with PmSBP7/8 [21]. Both PmSBP1 and PmSBP6 have miR156 binding sites, but the difference is that the miR156 binding site of PmSBP1 is in the coding region, while the miR156 binding site of PmSBP6 is in the 3′ UTR region [25]. Here, the expression patterns of pmu-miR156f and PmSBP1/6 during the flower development period were negatively correlated. The spatiotemporal expression pattern of genes is closely related to gene function. So far, the expression pattern of SBP in the flower development period has been studied in some perennial plants. The expression profile of AtSPL9/15 homologous genes in some species is as follows. In loquat, the expression of EjSPL9 decreased with the development of flower bud [27]. In chestnut, the expression level of CmSPL9 was increased with the development of flower bud [28], and so does the JrSBP23 in walnut [29]. In Betula luminifera, BlSPL8 was highly expressed in the early and middle stages of male inflorescence [30]. Unlike them, the expression of PmSBP1 was relatively stable and high in the flower bud morphological differentiation phase of P. mume, which indicated that PmSBP1 may play a vital role in flower bud differentiation. The expression pattern of AtSPL3/4/5 homologous genes in other species is as follows. In loquat, the expression of EjSPL3 and EjSPL 4 reached the peak at flower bud initiation, the expression level of EjSPL4 and EjSPL5 was suddenly raised in the middle stage of flower bud development [27]. The BlSPL15 of Betula luminifera was significant highly expressed in the early stage of female inflorescence [30]. But partly like EjSPL4, EjSPL5 and BlSPL15, in this study, the expression level of PmSBP6 was significantly high at the stamen differentiation stage (S5), which implied that it may participate in the stamen differentiation or regulate fertility. Different expression patterns in different species indicated that they may have different functions during flower bud development. The promoter is the switch of the gene, which can regulate the gene expression, or bind to the transcription factors to start or close the gene expression. In this study, the 2000 bp promoter sequences of PmSBP1/6 were cloned, both had typical promoter core structure regions, and the GUS staining results showed they had driven activity. Some fragment deletions were found in the cloned promoter sequences of PmSBP1/6 when compared with the genome sequence of wild P. mume. This may be the evolutionary result of the changes in environmental factors during the long-term cultivation process. In addition, the two promoter sequences contained multiple cis-acting elements, such as the light response elements, the anaerobic inducing elements, hormone response elements, and the endosperm expression element or meristem expression element. This suggested that PmSBP1 and PmSBP6 can be regulated by external environmental stimuli (light, water, and hormones), and may be expressed in some specific plant tissues. Besides, the exogenous GA3 downregulated the expression of PmSBP1/6, which was similar to the result of CmSPL9 in chestnut [28]. In addition, the tissue spatial expression profile showed that the PmSBP1 promoter was initiated in the apical meristem, and the PmSBP6 promoter was initiated in both the apical meristem and lateral bud. This displayed that they may play an important role in the initiation of flower buds. Recently, the function of AtSPLs in clade II of Arabidopsis was extensively studied. AtSPL8 functions in both the male and female fertility of Arabidopsis by affecting the anther development [17,31,32,33] and gynoecium patterning [33]. Based on the function research results, the other ten AtSPLs with miR156/157 sites are divided into three groups. Group 1 contains AtSPL2/9/10/11/13/15, which play role in both the transition from childhood to adulthood and from nutrition to reproduction, and AtSPL9/13/15 [34] play more important roles than AtSPL2/10/11 [10]. Group 2 contains AtSPL3/4/5, which can promote the transformation of flower meristem, but play a minor role in the change of vegetative stage or flower induction [16,35]. Group 3 contains AtSPL6, which does not play a major role in shoot morphogenesis, but may play an important role in some physiological processes [11]. As is known, the miR156-SPL is an important regulatory model in plant growth. Overexpression of miR156 in Arabidopsis prolongs childhood, while silencing miR156 makes plants mature early [3]. In Arabidopsis [2], tobacco [36], and several other plants [6,30,37], the content of miR156 was high and the content of SPL was low in childhood. With the increase of plant age, the content of miR156 decreased and the content of SPL increased. In addition, their expression can be regulated by each other [3,9]. In our study, PmSBP1 and its synonymous mutation PmSBP1tb and PmSBP6 were constitutively overexpressed in tobacco and showed their role in the regulation of plant growth, flowering time, and reproductive organ development. As our results showed, in the childhood stage, the 35S::PmSBP1tb transgenic tobacco was significantly smaller than that of the wild-type tobacco (Figure 4C), while the 35S::PmSBP1 transgenic seedlings were only slightly larger than the wild-type tobacco (Figure 4B). This may be related to the regulatory balance between the content of excessive SBP transcript and miR156. Unexpectedly, the 35S::PmSBP6 transgenic seedlings were also slightly larger than WT in the childhood stage (Figure 5A). Like PmSBP1tb, The CDS region of PmSBP6 does not contain a miR156 binding site, which cannot be regulated by nta-miR156 in tobacco, but they do have opposite phenotypes (Figure 4B and Figure 5A). We speculated that there was still a regulatory relationship between PmSBP6 and miR156 after removing the 3′ UTR sequence (containing miR156 binding site); however, further research should be conducted to confirm the new findings. The development of floral organs is very important for the sexual reproduction of plants, especially the stamen and pistil. In Arabidopsis, the AtSPL8 (without miR156 binding site) and other AtSPLs (with miR156 binding site) are necessary for the production of fully fertile flowers [23,26]. The knocking out of AtSPL8 resulted in abnormal anther development [17], but the overexpressing of AtSPL8 caused anther non-dehiscence [32]. When overexpressed miR156 in the spl8 mutant, the plant showed complete male sterility, but the overexpression of other AtSPL with miR156 binding sites in the spl8 mutant can alleviate the semi-sterile phenotype to a certain extent [26]. Another miR156 targeted gene AtSPL2 can affect plant fertility by affecting pollen production and fertilization rate. The fertility of both the gain-of-function mutant 35S::SPL2SRDX and the loss-of-function mutant spl2 decreased, and the fertility of the loss-of-function mutant was lower [38]. In this research, the overexpression of PmSBP1 (homologous to AtSPL9/15) and PmSBP6 (homologous to AtSPL3/4/5) in tobacco caused changes in the length of style and filament, and led to changes in the pollination mode, especially PmSBP1, which caused pollen abortion by reducing the number and vitality of pollen. However, overexpression of the PmSBP1tb gene neither caused changes in flower organs nor affected the pollen number and activity, and the fertility of PmSBP1tb transgenic plants was not affected. This indicated that both PmSBP1 and PmSBP6 can change the length ratio of style to the filament and affect the pollination mode, and the miR156 locus of PmSBP1 is indispensable in affecting plant fertility. Similar to Arabidopsis AtSPL8 [23,26] and AtSPL2 [38], all these genes affect plant fertility, but in different ways. Among these SBPs from different species, those not clustered together by phylogenetic analysis also have similar functions, this may be due to the species evolution. Ectopic expression of PmSBP1 and PmSBP6 in tobacco caused early flowering, which was similar to their homologous genes in Arabidopsis [15,34,35]. However, in our study, the flowering time of each transgenic plant was not consistent under different photoperiod conditions. The flowering time of PmSBP1 and PmSBP6 transgenic tobacco was earlier than that of WT only in short-day conditions, but the flowering time of PmSBP1tb transgenic tobacco was earlier in both long-day and short-day conditions. This indicated that PmSBP1 and PmSBP6 were affected by miR156 in promoting flowering in tobacco through mediating the photoperiod pathway. Furthermore, we examined the expression pattern of eight endogenous flowering-related genes in transgenic plants under different photoperiod conditions. Combined with the phenotype and expression level, we found that the expression level of endogenous NtSOC1 in early flowering transgenic plants was significantly higher than that of wild-typeunder corresponding photoperiod conditions. This suggested that PmSBP1 and PmSBP6 can regulate flowering time by regulating NtSOC1. Furthermore, the results of the yeast-one hybrid showed that PmSBP1 can regulate PmSOC1-2 and PmSOC1-3 by directly binding to their promoters. However, the relationship between PmSBP6 and the promoter of PmSOC1s cannot be verified by yeast-one hybrid because the introduction of PmSBP6 makes yeast grow abnormally. In previous reports, AtSPL3/4/5 mediated flowering time regulation by cooperating with FT-FD complexes in the photoperiod pathway [15]. Moreover, the AtSPL3/4/5 can be regulated by AtSOC1 by directly binding to their promoters [13]. However, the study on how SBP protein regulates flowering time by directly binding to the promoter of SOC1 has not been reported. Our findings opened a new way for SBP to regulate flowering time and laid a foundation for molecular breeding of P. mume and its research in other species.
Prunus mume ‘Sanlun Yudie’, grown in the campus of Beijing Forestry University was used in this study. The flower buds and young leaves were collected for gene and promoter cloning, respectively. The developmental periods of flower buds were identified by hand sectioning and flower buds were collected from July (initiation of flower bud differentiation) to March (flowering) in the next year for expression pattern analysis. To detect the response of PmSBP1 on gibberellin, the 100 mg/L GA3 was sprayed on the flower buds of cut-off one-year-old branches after pistil formation and during bud dormancy, the buds were sampled after 0 h, 2 h, 6 h, 12 h, 24 h, 48 h, and 72 h. The Nicotiana tabacum and Nicotiana benthamiana were grown at 25 °C, 16 h light/8 h dark (long-day conditions) or 8 h light/16 h dark (short-day conditions) in the greenhouse. The stem tips of transgenic and wild-type tobacco were sampled for endogenous gene detection at 100 d in short-day conditions and 110 d in long-day conditions. All samples were collected in liquid nitrogen and stored at −80 °C for RNA isolation.
The mature sequences and precursor sequences of pmu-miR156 in P. mume, ath-miR156 in Arabidopsis, and nta-miR156 in Nicotiana tabacum were obtained from our previous study [25], the miRbase (www.mirbase.org/ftp.shtmlmL, accessed on 2 June 2021), and published article [39], respectively. The precursor sequences of pmu-miR156 were aligned to the genome data of P. mume by blastn, and the chromosome position was drawn through the MG2Cv2 (http://mg2c.iask.in/mg2c_v2.0, accessed on 2 June 2021). The DNAMAN software was used to analyze the differences of miR156 mature sequence in P. mume, Arabidopsis and tobacco. The phylogenetic tree of the precursor sequences of three species was obtained through the maximum likelihood method in MEGA software after Clustalx alignment.
The total RNA of flower buds was isolated by EASYspin Plus Plant RNA Kit (Aidlab, Beijing, China), and the first strand of cDNA was synthesized by TransScript RT Kit (TIANGEN, Beijing, China). The genome DNA of young leaves was extracted by DNAsecure Plant Kit (TIANGEN). The cDNA was used as a template to amplify the coding sequences (CDS) of PmSBP1 and PmSBP6. The promoter sequences of PmSBP1/6 (about 2000 bp) and PmSOC1s (about 850 bp) were cloned from the genome DNA by specific primers. The primers were designed by Oligo7 and are listed in Supplementary Table S1. The total volume of the 50 μL polymerase chain reaction (PCR) system includes 25 μL PrimeSTAR HS (Premix) (TaKaRa, Beijing, China), 1 μL forward primer, and 1 μL reverse primer, 2 μL cDNA or 1 μL genome DNA. The PCR procedure was 94 °C pre-denaturation for 2 min; 35 cycles of 98 °C for 10 s, 55 °C for 5 s, 72 °C for 90 s; and 72 °C extensions for 10 min. The PCR products were separated from 0.8% agarose gel using TIANgel Midi Purification Kit (TIANGEN) and cloned into the pCloneEZ-NRS-Omni-Amp/HC vector (Clone Smarter, Houston, TX, USA) and transformed into Escherichia coli DH5α for sequencing (TIANGEN). The cis-acting element composition of the PmSBP1/6 (about 2000 bp) promoter sequences was predicted on the PlantCARE website [40].
The total RNA of flower buds in different flower development stages and treated with GA3, and the stem tip of tobacco was isolated by the EASYspin Plus Plant RNA Kit (Aidlab). The FastQuant RT Kit (with gDNase) (TIANGEN) was used to synthesize the first strand of cDNA, and the SYBR Premix ExTaq II (TaKaRa) was used for qRT-PCR following the instructions. The miRNA RT/qPCR Detection kit (Aidlab) was used to synthesize cDNA and complete Poly(A) tailed qRT-PCR of miR156f. The qRT-PCR was performed on the PikoReal real-time PCR system (Thermo Fisher Scientific, Waltham, MA, USA). The primers of the housekeeping gene (protein phosphatase 2A, PmPP2A) [41], miR156f [25], and PmSBP1/6 [21] of P. mume were referred to in the previous study. The reaction with 10 µL volume (1 µL of cDNA, 5 µL of SYBR Premix ExTaq II (Takara), and 0.2 µL of each primer) were conducted as follow: 30 s in 95 °C, 40 cycles of 5 s in 95 °C and 30 s in 60 °C, and finally end in 20 °C, and each reaction was repeated in triplicate. The qRT-PCR primers of the housekeeping gene (NtActin) and eight endogenous flowering-related genes in tobacco used were also from the previous study [42,43,44]. The relative expression level was calculated by the 2−∆∆Ct method [45]. The error line was drawn according to the standard deviation (calculated by EXCEL) of the three technical repetitions. The significant difference analysis of the gene expression level under GA3 treatment was performed by F-test in EXCEL.
To analyze the subcellular localization of PmSBP1/6, the CDSs of PmSBP1/6 were cloned into the vector pSuper1300-GFP between Sal I and Spe I restriction sites to generate the PmSBP1/6-GFP fusion gene driven by CaMV35S. The recombinant plasmid was transformed into Agrobacterium tumefaciens GV3101, cultured in 15 mL liquid LB medium (with 50 mg/L kanamycin and 50 mg/L rifampicin) until the OD600 = 0.8, then diluted to OD600 = 0.5, and injected into the leaves of 4–6 weeks old Nicotiana benthamiana. After injection for 24–72 h, the leaves were sectioned and stained in 4′,6-diamino-2-phenylindole (DAPI) solution for 10 min, and observed under TCS SP8 (Leica, Wetzlar, Germany) confocal laser scanning microscope. The fluorescence of GFP and DAPI were detected at 488 nm and 405 nm excitation wavelengths, respectively.
To analyze the promoter activity, the 2000 bp promoters of PmSBP1/6 were inserted into the vector pCAMBIA1305.4-GUS through Sal I/Bam HI restriction sites. After instantaneous transformation in N. benthamiana leaves, the leaves were cut and stained in histochemical GUS staining reagent for 24 h, and then rinsed in 75% alcohol to remove chlorophyll.
The miR156-sensitive gene PmSBP1, miR156-insensitive genes PmSBP1tb, and PmSBP6 were cloned into the plant expression vector pSuper1300 using Spe I/Kpn I restriction sites under the CaMV35S. PmSBP1tb is a synonymous mutation in the miR156 binding site of PmSBP1, and this mutation was completed by Sangon Biotech Company (Shanghai, China). The recombinants 35S::PmSBP1/6 and 35S::PmSBP1tb were transformed into A. tumefaciens GV3101, respectively. The A. tumefaciens contain 35S::PmSBP1/6, 35S::PmSBP1tb, and PmSBP1/6-promoter::GUS vectors were separately cultured in 30 mL liquid LB medium (with 50 mg/L kanamycin and 50 mg/L rifampicin) until the OD600 = 0.6–0.8, then diluted to OD600 = 0.2–0.5 and infected tobacco by leaf disc method [46]. The phenotypes of T3 generation of PmSBP1/6 and PmSBP1tb transgenic seedlings were observed and photographed. To determine the tissue expression specificity of the promoter of PmSBP1/6, different tissues of T1 generation of PmSBP1/6-promoter transgenic seedlings were stained in GUS solution for GUS activity detection.
The phenotype changes of these transgenic and wild-type tobacco were observed, recorded, and photographed. The flower morphology was observed throughout the flowering period, and the anthers were collected when the flower was just opened in each transgenic line. The 2,3,5-triphenyl tetrazolium chloride (TTC) staining and pollen germination in vitro were used to test the pollen viability. Three plants were used in each transgenic line. The TTC staining and pollen germination were observed under the microscope. The number of pollen grains in each field was more than 100, and six fields of each plant were counted. The average value and standard deviation of the three technical repetitions were calculated by EXCEL, and the error line was drawn according to the standard deviation. For transgenic plants of PmSBP1 and PmSBP6, artificial self-pollination was performed at the beginning of flower blooming. The pods were harvested when their color turned brown and cracked. The seeds and pods were photographed, and the seeds were stored in bags with desiccant.
In the cloned promoter sequences of PmSOC1s, there are several ‘GTAC’ SBP binding sites. Fragments of about 100 bp including the SBP binding sites during the 850 bp promoters were inserted into the Sac I/Sal I-cleaved pAbAi vector as baits. The prey vector pGADT7-PmSBP1/6 was constructed by Bam H I/Eco R I sites. The primers were designed by Oligo7 and shown in Supplementary Table S1. The bait recombinants plasmid was transformed into the yeast Y1HGold strains after being digested by Bst BI, and their tolerance to Aureobasidin A (AbA) was detected in the SD/-Ura/AbA medium. Subsequently, the prey vector was transferred into the yeast-containing bait vectors by the Quick & Easy Yeast Transformation Mix (TaKaRa) following its procedure, which was selected on the SD/-Leu/AbA medium.
Overall, PmSBP1/6 were important regulatory genes in flowering time and fertility, negatively regulated by miR156 and exogenous gibberellin. Meanwhile, PmSBP1 can directly bind to the promoters of PmSOC1-2 and PmSOC1-3 to regulate flowering. In conclusion, a miR156-PmSBP1-PmSOC1s pathway was formed to participate in the regulation of flowering time in P. mume. This study lays a foundation for revealing the molecular mechanism of flowering time regulation in P. mume. | true | true | true |
PMC9570445 | 36232448 | Indre Valiulyte,Aiste Pranckeviciene,Adomas Bunevicius,Arimantas Tamasauskas,Hanna Svitina,Inessa Skrypkina,Paulina Vaitkiene | Associations of miR-181a with Health-Related Quality of Life, Cognitive Functioning, and Clinical Data of Patients with Different Grade Glioma Tumors | 22-09-2022 | miR-181a,glioma,GBM,IDH1,survival,health-related quality of life | Gliomas are central nervous system tumors with a lethal prognosis. Small micro-RNA molecules participate in various biological processes, are tissue-specific, and, therefore, could be promising targets for cancer treatment. Thus, this study aims to examine miR-181a as a potent biomarker for the diagnosis and prognosis of glioma patients and, for the first time, to find associations between the expression level of miR-181a and patient quality of life (QoL) and cognitive functioning. The expression level of miR-181a was analyzed in 78 post-operative II-IV grade gliomas by quantitative real-time polymerase chain reaction. The expression profile was compared with patient clinical data (age, survival time after the operation, tumor grade and location, mutation status of isocitrate dehydrogenase 1 (IDH1), and promoter methylation of O-6-methylguanine methyltransferase). Furthermore, the health-related QoL was assessed using the Karnofsky performance scale and the quality of life questionnaires; while cognitive assessment was assessed by the Hopkins verbal learning test-revised, trail-making test, and phonemic fluency tasks. The expression of miR-181a was significantly lower in tumors of grade III and IV and was associated with IDH1 wild-type gliomas and a worse prognosis of patient overall survival. Additionally, a positive correlation was observed between miR-181a levels and functional status and QoL of glioma patients. Therefore, miR-181a is a unique molecule that plays an important role in gliomagenesis, and is also associated with changes in patients’ quality of life. | Associations of miR-181a with Health-Related Quality of Life, Cognitive Functioning, and Clinical Data of Patients with Different Grade Glioma Tumors
Gliomas are central nervous system tumors with a lethal prognosis. Small micro-RNA molecules participate in various biological processes, are tissue-specific, and, therefore, could be promising targets for cancer treatment. Thus, this study aims to examine miR-181a as a potent biomarker for the diagnosis and prognosis of glioma patients and, for the first time, to find associations between the expression level of miR-181a and patient quality of life (QoL) and cognitive functioning. The expression level of miR-181a was analyzed in 78 post-operative II-IV grade gliomas by quantitative real-time polymerase chain reaction. The expression profile was compared with patient clinical data (age, survival time after the operation, tumor grade and location, mutation status of isocitrate dehydrogenase 1 (IDH1), and promoter methylation of O-6-methylguanine methyltransferase). Furthermore, the health-related QoL was assessed using the Karnofsky performance scale and the quality of life questionnaires; while cognitive assessment was assessed by the Hopkins verbal learning test-revised, trail-making test, and phonemic fluency tasks. The expression of miR-181a was significantly lower in tumors of grade III and IV and was associated with IDH1 wild-type gliomas and a worse prognosis of patient overall survival. Additionally, a positive correlation was observed between miR-181a levels and functional status and QoL of glioma patients. Therefore, miR-181a is a unique molecule that plays an important role in gliomagenesis, and is also associated with changes in patients’ quality of life.
Astrocytic origin gliomas (astrocytomas) are central nervous system (CNS) tumors. According to the World Health Organization’s (WHO) 2016 classification of CNS tumors, these tumors are graded as grade II—diffuse astrocytoma, grade III—anaplastic astrocytoma, and grade IV—glioblastoma (GBM), of which GBM is the most common and malignant primary brain tumor [1]. The treatment of GBM remains a challenge. Despite available treatment options (surgical resection, adjuvant radiotherapy, and chemotherapy), the average life expectancy of patients diagnosed with GBM is only slightly above one year [2,3]. The dismal patient outcome is due to a tumor’s ability to regrow (~90%) and molecular heterogeneity [4]. Genomic and transcriptome studies revealed that histologically identical GBM forms may belong to different molecular subtypes, leading to different responses to treatment and patient life expectancy [5,6,7,8]. Therefore, more precise molecular identification of gliomas is necessary to prescribe more effective individualized therapies that would prolong patients’ survival times. This strategy is already widely used in the diagnosis of various other oncological diseases [9,10]. Nevertheless, the increased life expectancy of patients should not be the only aim of improved GBM therapies; quality of life with the oncological disease is of equal importance. Research indicates that the quality of life of GBM patients remains extremely poor [11]. Therefore, it is important to discover novel molecules that could more accurately predict the behavior of the tumor and the patient’s quality of functioning after the surgery. In recent years, it was demonstrated that micro RNAs (miRNAs) are associated with tumor progression and drug resistance by targeting genes associated with drug resistance or by affecting genes involved in cancer cell proliferation, cell cycle, and apoptosis [12,13,14]. Mature miRNAs are short, non-coding, regulatory RNAs of 21–25 nucleotides involved in the post-transcriptional regulation of gene expression by binding to the 3′ UTR of an mRNA. According to numerous studies, miRNAs regulate about a third of human genes and are involved in many biological processes, such as nervous system regulation, angiogenesis, cell cycle control, cell differentiation, proliferation, apoptosis, and even the immune response [15,16]. Importantly, a single miRNA targets many genes, has a high specificity for tissue, and is sensitive to tumor progression. As a result, depending on the organ or tissue, miRNA molecules may act as inhibitors and/or oncogenes and could be used as a non-invasive way to diagnose and predict disease [17]. One of the most studied and promising biomarkers with predictive value for the prognosis of cancer progression is miRNA-181a, which belongs to the miR-181 family. The family of miR-181 is composed of four different mature forms, namely miR-181a, miR-181b, miR-181c, and miR-181d, localized to three separate chromosomes (1, 9, and 19) [18]. The research studies have reported the involvement of miR-181a in diverse cellular functions such as cell growth, proliferation, death, survival, and maintenance, as well as gliomagenesis [18,19,20]. Therefore, the idea of this study is to examine miR-181a as a potent biomarker for the diagnosis and prognosis of glioma patients and to find associations between the expression level of miR-181a and patients’ health-related symptoms.
To reveal the importance of miR-181a in the pathogenesis of astrocytomas, the mRNA expression of miR-181a was analyzed in 78 different malignancy grade tumors. In Figure 1, it was demonstrated that the gene expression was diversified. The majority of GBM patients with lower than average (<−1.26) mRNA expression died within 2 years, while patients with lower-grade tumors had a higher expression of miR-181a and survived 2 to 6 years. As follows, to determine whether expression changes of miR-181a were significantly associated with patient clinicopathological characteristics, miR-181a expression was divided into “low” (<mean of miR-181a mRNA expression) and “high” (≥mean) gene expression groups. The analysis reveals that a higher expression of miR-181a is significantly associated with younger age of patients (<54-year, p = 0.006), lower tumor malignancy grade (p = 0.036), and gliomas with a mutant-type of IDH1 (p = 0.002) (see Table 1; Figure 2a–c). According to the Kaplan–Meier analysis, patients with a higher expression of miR-181a have a significantly higher chance of longer survival, compared with patients with low gene expression values (Log-rank test, χ2 = 4.465, df = 1, p = 0.035) (see Figure 2d). The median survival time was 8.9 months longer in the patient group with higher expression of miR-181a. The univariate cox regression analysis reveals that patients’ clinical characteristics such as age, tumor malignancy grade, and IDH1 status, as well as the expression of miR-181a, are significantly associated with patient overall survival (OS). However, according to the multivariate cox regression analysis, the tumor stage and IDH1 status are the only covariates significantly associated with the OS of glioma patients (see Table 2).
In addition, it was noted that all IDH1 mutant GBM tumors were detected in the higher miR-181a mRNA expression group, and the majority of IDH1 wild-type GBM tumors (62%, 32/52) were detected in the lower gene expression group (see Figure 1 and Figure 2b). According to Student’s t-test, the noted difference was statistically significant (p = 0.005). To find out whether miR-181a expression was associated with patient survival, only IDH1 wild-type GBM tumors were selected and divided into two groups, according to the median of miR-181a expression. Kaplan–Meier analysis shows the tendency that patients with a higher expression of miR-181a have a significantly higher chance of longer survival, compared with patients with low gene expression values (see Figure 3). According to the log-rank test, the difference was not statistically significant (χ2 = 2.64, df = 1, p = 0.104); however, the Gehan–Breslow–Wilcoxon method, which gives more weight to deaths at early time points, shows a statistically significant difference (χ2 = 5.83, df = 1, p = 0.016).
The correlation analysis was performed to reveal miR-181a expression associations with variables related to the quality of functioning (see Table 3). The results reveal that the expression of miR-181a positively correlates with general quality of life (EORTC QLQ-C30; r = 0.310, p = 0.010) and functional status, evaluated by a clinician (KPS; r = 0.237, p = 0.049) in all glioma patients. Interestingly, a statistically significant positive correlation between miR-181a and a better quality of life was reflected in the group of men (p = 0.009) rather than women (p = 0.469). In addition, as patients with GBM experience more severe symptoms, it was decided to analyze this subgroup separately. The expression levels of miR-181a show a significantly positive correlation with patient quality of life (EORTC QLQ-C30; r = 0.290, p = 0.041), but an inverse correlation with patient memory (HVLT-R; r = −0.291, p = 0.040). However, no statistically significant correlations of miR-181a with tumor-related symptoms (EORTC QLQ-BN20), depression (PHQ-9), cumulative learning (HVLT-R), psychomotor speed (TMT-A), executive functioning (TMT-B), and/or verbal fluency were determined.
Numerous studies have shown that miRNAs, which regulate biological processes such as cell proliferation, apoptosis, metabolism, and/or differentiation, are thought to have clinical potential in cancer prognosis and treatment [21]. Among the so-far-characterized miRNAs, miR-181a is involved in several types of cancer [18]. A significant upregulation of miR-181a level has been found in breast cancer [22], ovarian cancer [23], pancreatic cancer [24], hepatocellular carcinoma [25], and oral squamous cell carcinoma [26], whereas evident downregulation of miR-181a has been detected in non-small cell lung cancer [27] and prostate cancer [28], as compared with healthy controls. Our study also reveals that the expression levels of miR-181a decrease during astrocytoma progression. Comparable results were observed by other researcher groups, which demonstrates the downregulation of miR-181a in all grade glioma tumors (WHO II–IV), GBM cell lines, and glioma stem cells (GSCs), as compared with normal brain tissue, astrocytes, and differentiated GBM cells, respectively [29,30,31,32]. Importantly, the overexpression of miR-181a inhibited proliferation, migration, invasion, epithelial-mesenchymal transition (EMT), and induced apoptosis of GBM cells [30,33,34]. The process of apoptosis was modulated by targeting the apoptosis-related genes (p53, Bax, Bcl-2, Bim, etc.) [19], while proliferation was modulated by the downregulation of the MAPK pathway [35]. In addition, the upregulation of miR-181a sensitized the GBM cells to temozolomide (TMZ) and radiation treatment [34,36] also suppressed the formation of GSCs and inhibited GBM tumorigenesis [31]. More importantly, the study by Wu et al. showed the clinical significance of circulating miR-181a in patients with glioma tumors. Before the operation, circulating miR-181a was found to be downregulated in the plasma of GBM patients as compared with lower-grade tumors [37]. After 10 days, the levels of miR-181a increased more than 10-fold. In addition, a lower expression of circulating miR-181a was significantly associated with poor OS [37], as was demonstrated in our study with astrocytoma tumors. Therefore, the aberrantly downregulated miR-181a could be a critical factor that contributes to the malignant appearance of astrocytoma. Our study also demonstrates a significant association between the expression levels of miR-181a and patient age, as well as IDH1 status. However, according to the multivariate Cox regression analysis, only the tumor grade and IDH1 mutation were the best predictors of patient OS in our study cohort. Importantly, we noted that the mutant form of IDH1, which is the factor of a good astrocytoma patient survival prognosis, was found in GBM tumors with an increased expression of miR-181a. All these GBM patients with IDH1 mutation survived more than 14 months, and one of them survived even more than 57 months. We hypothesize that the expression of mirR-181a may influence the activity of the IDH1 protein which affects several major metabolic processes of the cells [38] and, therefore, has an impact on the OS of GBM patients. This relation between IDH1 and miR-181a in GBM patients was also observed by Sippl et al. [39]. They determined the inverse correlation between the expression of miRNA-181a2 and mRNA expression of IDH1 (p = 0.06, r = −0.55). Nevertheless, studies on a larger sample would be needed to confirm our observations. Next, we wanted to find out whether the survival of patients with IDH1 wild-type GBM tumors depends on miR-181a expression level. Since the survival of GBM patients after surgery is generally short, we additionally used the Gehan–Breslow–Wilcoxon method, which gives more weight to deaths at early time points. The analysis demonstrates that patients with a lower expression of miR-181a have significantly worse OS as compared with those with a higher gene expression level. However, the opposite effect was noted by another research group. They demonstrated that in patients with IDH1 wild-type GBMs, low miR-181a2 expression correlated with a prolonged OS (p = 0.019), [39]. The discrepancy may have occurred because different miR-181a isoforms were used in the analyses, which might have distinct biological functions. In our study, we analyzed the miR-181a1 isoform, which is located on chromosome 1 (37.p5), while miR-181a2 is situated on chromosome 9 (37.p5) [18]. Both isoforms produce almost identical mature miR-181a, but could be regulated by distinct molecules and, therefore, be differently expressed. For example, in human blood natural-killer cells, pri-miR-181ab-2 levels were higher than pri-miR-181ab-1. The immunosuppressive cytokine, TGF-β, suppressed pri-miR-181ab-1 expression while elevating pri-miR-181ab-2 expression. On the contrary, interleukins −2, −15, and −12/−18 increased the expression of pri-miR-181ab-1, but inhibited pri-miR-181ab-2 [40]. To the best of our knowledge, this was the first study comparing expression levels of miR-181a with patient quality of life. The analysis reveals that miR-181a positively correlates with the general quality of life subjective reported by patients themselves, as well as functioning status, evaluated by the treating clinician. However, no statistically significant correlations were found between miR-181a expression and more specific tumor-related symptoms, levels of depression, or cognitive functioning. A relatively small sample size, especially of tumor grades II and III, could affect the statistical power of our analysis. Furthermore, cognitive impairment and the profile of tumor-related symptoms are highly dependent on the localization of the tumor and other clinical factors, such as edema, tumor-induced compression to nearby tissues, or the frequency of seizures. Thus, it might be that relationships were lost due to the heterogeneity of the sample regarding tumor locations and other clinical characteristics. However, previous findings by our team group showed that miR-34a or miR-181b/d expression levels were related to patients’ functioning and tumor-related symptoms [41,42]. Therefore, we believe that current findings support the idea that levels of miR-181a expression might be valuable not only in predicting longer survival but also in better general functioning. Still, further studies are needed to clarify the mixed findings in male and female subgroups.
Seventy-eight samples from patients with the diagnosis of II-IV grade glioma tumors were analyzed in the study. All patients underwent neurosurgery at the Department of Neurosurgery, Hospital of Lithuanian University of Health Sciences, from 2015 to 2018. After surgical resection, tumor samples were frozen in liquid nitrogen. The diagnosis was confirmed by the pathologists. The study was approved by the Kaunas Regional Biomedical Research Ethics Committee, and written patient consent was taken from each patient before inclusion in the study. The clinical data, such as gender, age at the time of surgery, tumor grade, isocitrate dehydrogenase 1 (IDH1) status (the R132H mutation in the IDH1 gene), and methylation of O-6-methylguanine methyltransferase (MGMT) were collected from medical records. According to the WHO classification 2016 [1], there were 14 tumor samples of grade II, 6 of grade III, and 58 of grade IV (GBM). There were 35 women and 43 men, with a mean age of 54 years (range: 24–80 years). The overall survival of the patient was calculated from the date of tumor resection to the date of patient death or database closure (5 October 2021).
Patient functional status was assessed by a treating neurosurgeon during the hospital stay using the Karnofsky performance scale (KPS) [43]. The KPS measures a patient’s ability to carry on his/her normal daily activities and dependence on help and nursing care using an 11-point rating scale. The total KPS score ranges from 100 (normal functioning) to 0 (death), with higher scores indicating better daily functioning and higher functional independence. Data on functional status was available for 70 (88.6%) of patients.
Quality of life (QoL) assessment was performed 2–3 days before the neurosurgery. Patients were asked to fill out self-report questionnaires addressing their symptoms and quality of life. If assistance was needed due to reading, motor, or visual problems, questionnaires were filed with the help of a medical psychologist. Data on QoL was available for 68 (87.2%) of the patients. The European Organization for Research and Treatment of Cancer quality of life questionnaires QLQ-30 [44] and QLQ-BN20 [45] for brain tumor-related symptoms were used in this study. The QLQ-C30 contains 30 items that were designed to assess global health status, subjectively reported functional status, role functioning, emotional functioning, cognitive functioning, social functioning, and various cancer-related symptoms. Raw scores were linearly transformed to 0–100 scales, with higher scores indicating better quality of life. The QLQ-BN20 is a 20-item self-rating scale specifically developed for the assessment of health complaints in brain tumor patients. The questionnaire contains many common BT-related symptoms, including future uncertainty, visual disorder, cognitive impairment, etc. The QLQ-BN20 scores were linearly transformed to a 0–100 scale, with a higher score indicating greater BT-related symptom severity. In addition to QoL instruments, we also included a measure of depression, as depression is significantly related to decreased QoL in many patient populations. The patient health questionnaire-9 (PHQ-9) [46] was chosen for the assessment of current depressive symptoms. The PHQ-9 is based on the diagnostic statistical manual-IV depression diagnostic criteria, and it is recognized as a valid and reliable tool for depression screening in glioma patients [47,48].
A cognitive assessment was performed by a medical psychologist 2–3 days before the neurosurgery at the Department of Neurosurgery of the Hospital of LUHS. A set of neuropsychological tests, recommended for the assessment of treatment outcomes in glioma studies, was used [49]. The Hopkins verbal learning test-revised (HVLT-R) was used for verbal memory assessment [50]. The test consists of 12 words that are read aloud for three trials, each trial followed by a patient’s free recall. After an approximately 20 min. delay, during which other tests are administered, the patient is asked to recall the list of words. Two scores: cumulative learning (total number of words recalled in trials 1, 2, and 3) and delayed recall (number of words recalled after a delay) were analyzed in this study. The Trail-making test (TMT, parts A and B) was used for the assessment of psychomotor speed and executive functioning [51]. During the task, a patient is asked to connect a sequence of 25 targets (numbers 1, 2, 3, etc. in Part A, and numbers and letters in Part B) on a sheet of paper. The time of completion (in seconds) is considered as an indicator of psychomotor speed and executive functions. Verbal fluency was measured using phonemic fluency tasks [52]. Patients were asked to produce as many words as possible beginning with a specific letter within the one-minute interval. Three trials using the letters K, A, and S were performed. The total number of words produced during three trials is used as a verbal fluency indicator. As cognitive performance is sensitive to aging processes, all primary scores of neuropsychological tests were transformed to age-adjusted T scores (Mean 50, SD 10) using available norms for the Lithuanian population, with higher scores indicating better cognitive function. Memory and verbal fluency data were available for 68 (87%) of the patients, and data on psychomotor speed and executive function was available for 62 (79.5%) of the patients, as some patients were not able to perform these tasks due to visual or motor impairment.
Small RNAs were extracted from frozen tumor tissue using a mirVana™ miRNA Isolation Kit (Life Technologies, Carlsbad, CA, USA, cat. no. AM1560), according to the manufacturer’s instructions. The quality and concentration were determined by NanoDrop 2000 (Thermo Fisher Scientific, Wilmington, DE, USA). Following this, 10 ng of purified micro RNAs were synthesized to cDNA using the “TaqMan Advanced miRNA cDNA Synthesis Kit” (Thermo Fisher Scientific, Pleasanton, CA, USA, cat. no. A25576). The RT-PCR with TaqMan™ Fast Advanced Master Mix (Life Technologies, Carlsbad, CA, USA, cat. no. 4444965) was performed to analyze miR-181a expression changes in II-IV grade glioma tumors and healthy human brain RNA sample “FirstChoice Human Brain Reference RNA” (RHB; Ambion, Austin, TX, USA, cat. no. AM7962). The reaction consisted of 6 µL of TaqMan® Fast Advanced Master Mix, 0.6 µL of hsa-miR-181a-3p probe (Applied Biosystems, Foster City, CA, USA, Assay ID: 479405_mir), 3 µL of cDNA sample, and nuclease-free water to a total volume of 12 µL. The reactions were performed in the RT-PCR System “Applied Biosystems 7500 Fast” (Applied Biosystems, Foster City, CA, USA) using a fast-cycling program. In addition, the housekeeping genes has-miR191-5p (Assay ID: 477952_mir), has-miR361-5p (Assay ID: 478056_mir), has-miR345-5p (Assay ID: 478366_mir), and has-miR103a-3p (Assay ID: 478253_mir) were measured to normalize the data. The average of the housekeeping genes was used for the comparative 2−∆∆Ct method, in which gene expression in tumor samples was compared to healthy brain tissue.
Statistical programs SPSS (version 25.0, IBM SPSS, Armonk, NY, USA) and GraphPad Prism (version 7.0, Graph-Pad Software, San Diego, CA, USA) were used for data analysis. The normality was confirmed by the Kolmogorov–Smirnov test. The Student’s independent t-test was applied to evaluate the difference in miR-181a expression between the two groups. Meanwhile, one-way ANOVA with Tukey’s correction was used for the three groups. The Chi-square test was used for the comparison of categorical variables and the Pearson correlation was used for the quantitative variables. Patient survival was evaluated according to the Kaplan–Meier curves with log-rank or Breslow tests. Univariate and multivariate (with the backward conditional method) Cox regression analyses were used to evaluate the relationships between clinical, psychological, and molecular variables and patient overall survival. The significance level was defined as p < 0.05.
The study suggests that miR-181a could be a promising biomarker for glioma patients, since the expression of miR-181a decreases during tumor progression and the downregulation of the gene is significantly associated with the worst patient survival prognosis. More importantly, lower expression levels of miR-181a are also related to wild-type IDH1, older age, as well as worse patient quality of life and functioning (see Figure 4). However, the findings of this study should be seen in light of some limitations. One of them is a small number of tumor samples and incomplete sample information. Unfortunately, not all of the patients had information about IDH1, MGMT status, and functional data. Furthermore, patients’ genetic and medical histories were not considered during this study, which might affect the evaluation of patients’ health-related quality of life. This may have affected the reliability of the results. Therefore, follow-up studies are needed to confirm the prognostic value of miR-181a in glioma patients. | true | true | true |
PMC9570462 | 36232884 | Tomasz Szaflik,Hanna Romanowicz,Krzysztof Szyłło,Radosław Kołaciński,Magdalena M. Michalska,Dariusz Samulak,Beata Smolarz | Analysis of Long Non-Coding RNA (lncRNA) UCA1, MALAT1, TC0101441, and H19 Expression in Endometriosis | 30-09-2022 | lncRNA,UCA1,MALAT1,TC0101441,H19,endometriosis | Endometriosis is a disease of complex etiology. Hormonal, immunological, and environmental factors are involved in its formation. In recent years, special attention has been paid to genetic mechanisms that can have a significant impact on the increased incidence of endometriosis. The study aimed to analyze the expression of four long non-coding RNA (lncRNA) genes, UCA1, MALAT1, TC0101441, and H19, in the context of the risk of developing endometriosis. The material for genetic testing for the expression of lncRNA genes were tissue slices embedded in paraffin blocks from patients with endometriosis (n = 100) and the control group (n = 100). Gene expression was determined by the RT-PCR technique. The expression of the H19 gene in endometriosis patients was statistically significantly lower than in the control group. A statistically significant association was found between H19 gene expression in relation to The Revised American Society for Reproductive Medicine classification of endometriosis (rASRM) in the group of patients with endometriosis. Research suggests that H19 expression plays an important role in the pathogenesis of endometriosis. | Analysis of Long Non-Coding RNA (lncRNA) UCA1, MALAT1, TC0101441, and H19 Expression in Endometriosis
Endometriosis is a disease of complex etiology. Hormonal, immunological, and environmental factors are involved in its formation. In recent years, special attention has been paid to genetic mechanisms that can have a significant impact on the increased incidence of endometriosis. The study aimed to analyze the expression of four long non-coding RNA (lncRNA) genes, UCA1, MALAT1, TC0101441, and H19, in the context of the risk of developing endometriosis. The material for genetic testing for the expression of lncRNA genes were tissue slices embedded in paraffin blocks from patients with endometriosis (n = 100) and the control group (n = 100). Gene expression was determined by the RT-PCR technique. The expression of the H19 gene in endometriosis patients was statistically significantly lower than in the control group. A statistically significant association was found between H19 gene expression in relation to The Revised American Society for Reproductive Medicine classification of endometriosis (rASRM) in the group of patients with endometriosis. Research suggests that H19 expression plays an important role in the pathogenesis of endometriosis.
Endometriosis is an estrogen-dependent, gynecological disease that, due to the accompanying ailments and chronic nature, is a very important medical, social and economic problem [1]. For endometriosis, subtypes of endometriosis, treatment, and outcomes, different definitions are used. This has important implications for research and clinical practice. A paper prepared by the International Working Group of AAGL, ESGE, ESHRE, and WES presented a list of 49 terms used in endometriosis; including the definition of endometriosis and its subtypes, different locations, interventions, symptoms, and outcomes [2]. It will also provide an appropriate definition of endometriosis. Endometriosis is defined as an inflammatory process characterized during surgery by the presence of an epithelium and/or endometrial-like stroma outside the endometrium and myometrium, usually accompanied by an inflammatory process. Recently, in the context of endometriosis, long non-coding RNAs (lncRNAs) have become of interest, which covers more than 200 bp and are a subtype of non-coding RNAs (ncRNAs) [3,4,5,6,7]. Unlike a group of short, non-coding RNAs (sncRNAs) such as microRNAs (miRNAs), long non-coding RNAs tend to show greater sequence matching and, thus, specificity of action relative to target genes [8]. Non-coding RNA molecules participate in the processes of regulation at virtually all stages of the transmission of genetic information: from DNA to protein. Particularly spectacular is the involvement of certain non-coding RNA molecules in the mechanisms leading to the switching on or off of the expression of individual genes. In the research of Zhou et al., 388 of the lncRNA transcripts tested showed overexpression, and 188 showed reduced levels of expression in the ectopic endometrium compared to the eutopic endometrium [3]. It is known that the expression of many lncRNA is subject to changes in (a) the serum of women with endometriosis compared to healthy women, (b) the eutopic endometrium of women with endometriosis compared to healthy women, (c) in the ectopic endometrium of the ovaries compared to the eutopic endometrium in women with endometriosis [4]. Recent literature data suggest that the following lncRNAs are associated with endometriosis: UCA1, MALAT1, TC0101441, and H19 [5,6]. In the work of Mejer et al., it was presented that, specifically, these four lncRNAs play an important role in pathogenesis and the development of endometriosis [6]. Although the above lncRNA sequences are somehow correlated with endometriosis, the exact meaning and context of this interaction remain unclear and need to be explained. This work is designed to expand the still vague and incomplete knowledge of the role of selected lncRNA in endometriosis. There is limited literature data on these lncRNAs, and until now, none of them has been studied in endometriosis in the Polish population. In our study, we presented an analysis of the expression of UCA1, MALAT1, TC0101441, and H19lncRNAs in endometriosis patients compared to controls, to possibly correlate these lncRNAs with the risk of endometriosis.
The patient’s characteristics are shown in Table 1. Patients with endometriosis were between the ages of 23 and 58 (mean age 36.4 ± 6.0). Control patients were aged 29 to 61 years (mean age 38.3 ± 6.2). The clinical stage of patients with endometriosis was determined according to the rASRM (The Revised American Society for Reproductive Medicine) classification of endometriosis 1996) [9]. H19, UCA1, MALAT1, and TC0101441 expression results in endometriosis patients and controls, measured by relative quantification, are presented in Figure 1, Figure 2, Figure 3 and Figure 4, respectively. There were no statistically significant differences in the expression of UCA1, MALAT1, and TC0101441 between patients and the control group. Statistically, there was a significantly lower expression of H19 in cases of endometriosis compared to controls. Table 2 presents the comparisons of expressions of the studied lncRNA in patients and controls. The expression of UCA1, MALAT1, TC0101441, and H19 lncRNA sequences were also statistically analyzed for correlation with clinical-pathological data on age, parity, spontaneous abortions, hormone replacement therapy, and BMI. All the correlations listed above turned out to be statistically insignificant. lncRNA expression was also analyzed statistically for correlation with the clinical stage of endometriosis. An association was observed between H19 gene expression and rASRM classification in the endometriosis group (Table 3).
Despite the fact that research on biomarkers of endometriosis is still ongoing, there are still no satisfactory results, which makes it impossible to carry out effective laboratory diagnostics used in the diagnosis and monitoring of the treatment of the disease [10]. Recently, in the context of endometriosis, long non-coding RNAs have become the subject of interest. However, the clinical significance and biological mechanism of lncRNA in the development of endometriosis remain largely unknown. There are reports in the scientific press suggesting a link between lncRNA expression and the development of endometriosis [5,6,7,11,12,13,14,15]. This study aimed to analyze the level of expression of four long non-coding RNA (UCA1, MALAT1, TC0101441, and H19) in endometriosis. In the context of lncRNA expression levels, the following were evaluated: clinical stage of endometriosis, age, BMI, parity, and spontaneous abortions. An interesting aspect of such an assessment is whether lncRNA can serve as a biomarker in a group of patients affected by endometriosis. On this basis, it is possible to identify patients belonging to the risk group whose prognosis would be worse and the therapeutic process should be more aggressive—even in the case of endometriosis at an early clinical stage. The only lncRNA in which our statistical analysis showed a significant correlation with endometriosis was H19. In cases of the disease, a reduced level of H19 expression was shown. In addition, H19 levels correlated with the clinical degree (staging) of endometriosis. There are reports similar to our results. Ghazal et al. showed that H19 expression was reduced in eutopic endometrial patients with endometriosis compared to the control group [16]. Reduced H19 expression leads to increased let-7 bioavailability, which in turn inhibits IGF1R expression at the post-transcriptional level, thus, contributing to reduced stromal cell proliferation. Disruption of the H19/Let-7/IGF1R regulatory pathway may contribute to impaired endometrial preparation and receptivity in women with endometriosis. Liu et al. showed that H19 lncRNA was downregulated in mononuclear cells from peritoneal fluid (PFMC) in endometriosis patients [17]. Korucuoglu et al. presented that H19 expression is reduced in endometrial tissues of infertile women for an unexplained cause [18]. There are studies showing that the expression of H19 lncRNA in the ectopic and eutopic endometrial tissues of endometriosis is much higher than in normal endometrial tissues [19,20,21]. H19 overexpression in endometriosis lesions is associated with infertility, relapse, bilateral ovarian lesions, elevated CA125 levels, and progression in the altered stage of the American Fertility Society disease (rAFS). Multivariate logistic regression analysis showed that H19 overexpression in endometriosis lesions is a prognostic factor for endometriosis recurrence [21]. Further studies of H19 expression in endometriosis showed that expression could be differentiated in endometriosis. In the case of endometriosis tissues and/or cells, both up- and downregulated H19 were observed. For body fluids or lesion microenvironment H19 was downregulated [5]. The discrepancy in the researchers’ H19 expression results is likely the result of differences in race and perhaps the stage of the disease. Further genetic testing is, therefore, needed. Endometriosis is known to be an estrogen-dependent disease. Therefore, lncRNAs involved in the estrogen pathway or those directed by estrogen signaling may play a role in this disease. H19 is positively regulated by estrogen, and its expression in the endometrium increases in the proliferative phase of the menstrual cycle [18]. H19 has been shown to regulate several pathways that are important in endometriosis, IGF1R, ITGB3, IER3, and ACTA2 [16,17,20,22]. H19 can act as a miRNA sponge in endometriosis. Reduced H19 levels have been shown to be associated with an increase in let-7 miRNA activity. The consequence is the inhibition of IGF1R expression, which results in reduced proliferation of endometrial stromal cells [20]. The H19/let7/IGF1R pathway may contribute to impaired endometrial receptivity in women suffering from this disease. H19 regulates cell proliferation and invasion of endometrial ectopic cells by increasing ITGB3 expression through the miR-124-3p sponge [17]. H19 is also involved in the impaired immune response of women with the disease, acting as a sponge for miR-342-3p, which regulates the IER3 pathway. This pathway is involved in the differentiation of Th-17 cells and the proliferation of endometrial stromal cells in ectopic sites in women with this disease. In our study, three of the four sequences we studied (UCA1, MALAT1, TC0101441) were found to be statistically insignificant from the point of view of endometriosis risk. Our results are in contrast to the literature data that support the role of these lncRNAs in endometriosis. Long non-coding RNA urothelial carcinoma-associated 1 (UCA1) has been proven to be involved in the pathogenesis of various diseases in humans, including various types of ovarian diseases, such as ovarian cancer [23]. Recent research indicates that UCA1 participates in the pathogenesis of endometriosis [24]. Research by the team of Huang et al. showed that in most patients, the level of expression of lncRNA UCA1 was significantly higher in the ectopic endometrial tissues than in paired eutopic endometrial tissues [24]. Compared with healthy controls, serum lncRNA-UCA1 levels were lowered in patients with ovarian endometriosis, and serum UCA1 levels decreased further as the disease progressed. Serum levels of lncRNA-UCA1 l showed no significant correlation with either the age of the patients or their life habits. After treatment, serum UCA1 levels increased, and serum UCA1 levels on the day of discharge were significantly lower in relapsed patients than in non-relapse patients [24]. Based on these data, it can be concluded that the downregulation of UCA1 is involved in the pathogenesis of ovarian endometriosis and may be a promising diagnostic and prognostic biomarker of this disease. Since endometriosis is known as a precursor to some types of ovarian cancer, similar biological functions of lncRNAs detected in cancer may also determine the development of endometriosis [25]. UCA1 can interact with various miRNAs, which then change gene expression profiles and control tumor progression. The participation of genetic variants of UCA1 in the development of endometriosis and related infertility is indicated [26]. It is known that UCA1 may be involved in the occurrence and development of endometriosis by promoting cell proliferation and inhibiting apoptosis [27]. MALAT1 expression is enhanced in many cancer tissues, and MALAT1 is involved in the proliferation, apoptosis, migration, invasion, and spread of cancer cell metastasis [28]. Endometriosis can be considered a mild metastatic disease and, in addition, due to the ability of endometrial tissue to infiltrate, form metastases and relapses, like tumors, it is very similar to cancer [29]. Epidemiological data suggest that endometriosis has malignant potential [30]. MALAT1 lncRNA expression has been shown to be significantly elevated in endometrial ectopic tissues compared to eutopic endometrial tissues [31]. Yu et al. found that MALAT1 lncRNA can promote endometrial cell apoptosis and regulate MMP-9 expression through the NF-κB/iNOS pathway, thus, mediating the pathogenesis of endometriosis [32]. MALAT1 levels were significantly higher in ectopic endometrial tissues than in paired eutopic endometrial tissues [33]. lncRNA-MALAT1 can also work by regulating autophagy under hypoxia conditions in endometriosis. MALAT1 silencing has been shown to inhibit hypoxia-induced autophagy in human endometrial stromal cells [34]. TC0101441 is a newly identified lncRNA associated with cancer metastasis [35,36]. Since endometriosis has a prometastatic effect similar to those observed in cancer, it was decided to determine the role of TC0101441 in this disease. Analyses TC0101441 concentration in extracellular vesicles contributes to migration/invasion of endometrial cyst stromal cells (ECSC) [37]. It was shown that lncRNA-TC0101441 is strongly expressed in ectopic lesions and that serum extracellular follicular TC0101441 levels were significantly increased in patients with stage III/IV endometriosis compared to patients with stage I/II endometriosis and control. This indicates the potential of extracellular vesicular TC0101441 as a biomarker of endometriosis [38]. lncRNA-TC0101441 in epithelial ovarian cancer (EOC) promotes EOC cell migration/invasion [39]. Extracellular transport of TC0101441 via exosomal vesicles is known to enhance the migration and invasion of endometriosis [39]. All the literature reports cited above encourage the conclusion that the study of selected lncRNA sequences is supported by convincing premises, and research in this field must be continued. The results presented in this paper on the analysis of lncRNA genes reveal the existence of certain relationships of their expression with endometriosis. It should be noted, however, that the presented research covered a small population (100 patients and 100 control) and requires further work carried out on much larger groups of respondents. The study groups we use may simply be quantitatively unsatisfactory to draw the right conclusions. Second, the cases and controls are not exactly homogeneous—to some extent, they differ in age, spontaneous abortion, and the use of hormone replacement therapy, which can falsify the results. What is more, the cases are not a group of disease-free women, but they were all treated surgically for a mild gynecological condition—symptomatic uterine fibroids. Some reports suggest that some lncRNA sequences (i.e., lncRNA H19) may exhibit a correlation with uterine fibroids’ tissue as such [40], yet in our study, the genetic assays were performed strictly on the selected endometrial tissues and not on uterine fibroids. With all of the above findings in mind and realizing the limitations of our study, we dare say that this research has shed new light on lncRNAs in endometriosis and contributes to a growing—but still unclear—knowledge of these non-coding sequences in endometriosis.
The study group consisted of 100 patients diagnosed with endometriosis in the laparoscopic and pathological examination. On the other hand, the control group consisted of 100 patients who had no endometriosis during the surgical procedure, and the eutopic endometrium was normal in the histopathological examination. These patients were operated on for uterine fibroids. The exclusion criterion was concomitant cancer in patients pre-classified to the study group and existing cancers in patients from the control group. Patients were selected for studies in the period between 2015 and 2016. The material for analysis was RNA isolated from paraffin blocks obtained from material collected during surgery or curettage of the uterine cavity in the Department of Gynecology, Gynecology Oncology and Endometriosis Treatment of the Polish Mother’s Memorial Hospital Research Institute, Lodz. Paraffin blocks came from the archives of the Department of Clinical Pathology of the Polish Mother’s Memorial Hospital Research Institute, Lodz. All preparations have been characterized histologically. A formal consent (No. 98/2015) was obtained from the Bioethical Committee of the Institute-Polish Mother’s Memorial Hospital in Lodz, Poland.
Total RNA was isolated from formalin-fixed paraffin-embedded (FFPE) tissues using the High Pure RNA isolation kit (Roche Diagnostics GmbH, Mannheim, Germany), according to the manufacturer’s instructions. The FFPE samples were placed in 2 mL Eppendorf tubes, dewaxed with 100% xylene, washed in 100% ethanol, and dried at 55 °C for 10 min. The dried tissue was suspended in 100 μL of paraffin tissue lysis buffer (included in the kit) and digested with proteinase K at 55 °C overnight. The resulting total RNA was used for cDNA synthesis or stored at −80 °C until use.
The purity of the obtained RNA preparations was determined by the spectrophotometric method by twice measuring the absorbance of each sample at wavelengths of 260 nm and 280 nm. The adopted criterion for RNA purity was A260/A280 within 1.8–2.0. The concentration of RNA was determined by the spectrophotometric method based on the absorbance value measured at a wavelength of 260 nm. This value corresponded to the following relationship: The mean values of purity and RNA concentration obtained for the test material and the correct one met the necessary criteria.
The reverse transcription was carried out using the Maxima First Strand cDNA synthesis kit (ThermoFisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer’s protocols. 500 ng of total RNA was used as the starting material. Reverse transcription was performed under conditions optimized for use with this kit (25 °C for 10 min, 50 °C for 30 min, 85 °C for 5 min). The cDNA samples were stored frozen at −20 °C. Quantification of lncRNA was carried out using TaqMan™ (Applied Biosystems, Lincoln Centre Dr, Foster City, CA, USA). In addition, GADPH Assay has been used as an endogenous control. qPCR reactions were performed in a volume/reaction of 10 μL, including 10 ngcDNA, 5 μL TaqMan Fast Advanced PCR Master Mix, and 0.5 μL of a suitable primer (20×) (OriGene Technologies GmbH, Schillerstr.532052, Herford, Germany). The expression of GAPDH served as an endogenous control. The following primer sequences were used: TC0101441: forward, 5′ AAGGCAGGTGAGAACGAGT3′; reverse, 5′CTCGACTTAGGGAGCTGCAC3′, UCA1: forward 5′-TTTATGCTTGAGCCTTGA-3′; reverse 5′-CTTGCCTGAAATACTTGC-3′. MALAT1: forward 5′ GACTTCAGGTCTGTCTGTTCT3′; reverse 5′ CAACAATCACTACTCCAAGC3′, H19: forward 5′ ATCGGTGCCTCAGCGTTCGG3′; reverse 5′ CTGTCCTCGCCGTCACACCG3′ and GAPDH: forward, 5′ AGCCACATCGCTCAGACAC3′; reverse 5′ GCCCAATACGACCAAATCC3′. Each assay was performed in triplicate. The samples were incubated in a 96-well plate at 95 °C for 3 min, followed by 40 cycles of 95 °C for 1 s and 60 °C for 20 s. The relative level of expression was determined by the 2-ΔΔCq method.
Statistical analysis of the obtained results was carried out using the STATISTICA 11 program (StatSoft, Krakow, Poland). The analysis of the significance of differences in the level of gene expression at the mRNA level was carried out using non-parametric tests (Mann–Whitney U test, Kruskal–Wallis test) due to the lack of normality of the distribution of the obtained results, which was confirmed by the Shapiro–Wilk test. The correlation analysis between the variables was performed using the R-Spearman test. Statistical significance was observed at p < 0.05.
In view of the relevance of the results obtained, these studies should be continued. The results obtained in the work contribute to the broadening of knowledge on the subject of molecular mechanisms conducive to the development of endometriosis. The study showed differences in lncRNA expression between the studied groups of patients that could potentially promote the development of the disease. Due to the small number of studies on lncRNA expression analyses in patients with endometriosis in the Polish population, the effect of the research makes a significant contribution to expanding knowledge about the impact of genetic factors on the development of endometriosis. Understanding the relationship between gene expression and endometriosis may contribute to the development of new therapeutic strategies in the context of this disease. Gene expression testing could be a future target for personalized therapy. The current state of knowledge about lncRNA in endometriosis is still limited. Further research is warranted to further explore this topic. | true | true | true |
PMC9570855 | 36233480 | Katarzyna Zdanowicz,Anna Bobrus-Chcociej,Karolina Pogodzinska,Agnieszka Blachnio-Zabielska,Beata Zelazowska-Rutkowska,Dariusz Marek Lebensztejn,Urszula Daniluk | Analysis of Sphingolipids in Pediatric Patients with Cholelithiasis—A Preliminary Study | 23-09-2022 | cholelithiasis,gallstones,sphingolipids,children | (1) Background: Disturbances in the sphingolipid profile are observed in many diseases. There are currently no data available on the evaluation of sphingolipids and ceramides in cholelithiasis in children. The aim of this study was to evaluate the concentrations of sphingolipids in the sera of pediatric patients with gallstones. We determined their relationship with anthropometric and biochemical parameters. (2) Methods: The concentrations of sphingolipids in serum samples were evaluated using a quantitative method, ultra-high-performance liquid chromatography–tandem mass spectrometry. (3) Results: The prospective study included 48 children and adolescents diagnosed with gallstones and 38 controls. Serum concentrations of total cholesterol (TC); sphinganine (SPA); ceramides—C14:0-Cer, C16:0-Cer, C18:1-Cer, C18:0-Cer, C20:0-Cer and C24:1-Cer; and lactosylceramides—C16:0-LacCer, C18:0-LacCer, C18:1-LacCer, C24:0-LacCer and C24:1-LacCer differed significantly between patients with cholelithiasis and without cholelithiasis. After adjusting for age, gender, obesity and TC and TG levels, we found the best differentiating sphingolipids for cholelithiasis in the form of decreased SPA, C14:0-Cer, C16:0-Cer, C24:1-LacCer and C24:0-LacCer concentration and increased C20:0-Cer, C24:1-Cer, C16:0-LacCer and C18:1-LacCer. The highest area under the curve (AUC), specificity and sensitivity were determined for C16:0-Cer with cholelithiasis diagnosis. (4) Conclusions: Our results suggest that serum sphingolipids may be potential biomarkers in pediatric patients with cholelithiasis. | Analysis of Sphingolipids in Pediatric Patients with Cholelithiasis—A Preliminary Study
(1) Background: Disturbances in the sphingolipid profile are observed in many diseases. There are currently no data available on the evaluation of sphingolipids and ceramides in cholelithiasis in children. The aim of this study was to evaluate the concentrations of sphingolipids in the sera of pediatric patients with gallstones. We determined their relationship with anthropometric and biochemical parameters. (2) Methods: The concentrations of sphingolipids in serum samples were evaluated using a quantitative method, ultra-high-performance liquid chromatography–tandem mass spectrometry. (3) Results: The prospective study included 48 children and adolescents diagnosed with gallstones and 38 controls. Serum concentrations of total cholesterol (TC); sphinganine (SPA); ceramides—C14:0-Cer, C16:0-Cer, C18:1-Cer, C18:0-Cer, C20:0-Cer and C24:1-Cer; and lactosylceramides—C16:0-LacCer, C18:0-LacCer, C18:1-LacCer, C24:0-LacCer and C24:1-LacCer differed significantly between patients with cholelithiasis and without cholelithiasis. After adjusting for age, gender, obesity and TC and TG levels, we found the best differentiating sphingolipids for cholelithiasis in the form of decreased SPA, C14:0-Cer, C16:0-Cer, C24:1-LacCer and C24:0-LacCer concentration and increased C20:0-Cer, C24:1-Cer, C16:0-LacCer and C18:1-LacCer. The highest area under the curve (AUC), specificity and sensitivity were determined for C16:0-Cer with cholelithiasis diagnosis. (4) Conclusions: Our results suggest that serum sphingolipids may be potential biomarkers in pediatric patients with cholelithiasis.
The incidence of cholelithiasis is constantly increasing in the pediatric population, ranging from 1.9% to 4% [1,2]. Factors that predispose to the development of cholelithiasis include hemolytic disorders, ceftriaxone therapy, total parenteral nutrition (TPN), cystic fibrosis, obesity, genetic predisposition and hipercholesterolemia [3]. However, obesity is not only a growing problem in the pediatric population worldwide, but is also the most common cause of cholelithiasis in children today [1,4]. Additionally, similarly to adults, a significantly greater incidence of gallstone disease is observed in girls than in boys [5]. The formation of cholesterol gallstones is connected with hypersecretion of cholesterol and supersaturated bile in the gallbladder. Despite the known predisposing factors, such as genes and lifestyle, the mechanism of this disease has not been clearly defined [6,7]. Novel omics techniques such as metabolomics and lipidomics are promising tools that allow the measurement of carbohydrates, lipids, amino acids, amines or steroids. Sphingolipids are a class of bioactive lipids that include, i.a., ceramides (CER), lactosylceramides (LacCer), sphingosine (Sph) and sphinganine (SPA). The roles of these lipids are complex and involve the inflammatory response and cellular metabolism, migration and signaling. CER may also be involved in growth and apoptosis [8]. Several studies have found that sphingolipids may be associated with obesity, insulin resistance, inflammation-related illnesses and cancer [9,10]. To the best of our knowledge, there are currently no data available on the evaluation of sphingolipids and ceramides in cholelithiasis in both children and adults. Changes in the concentrations of adipokines may be indirect indicators of lipid disturbances. In the studies published so far in children, the effect of chemerin on the presence of cholelithiasis has been observed [11]. Chemerin, as an adipokine involved in adipogenesis, glucose homeostasis and energy metabolism [12], may be associated with lipid disturbances. The aim of our study was to analyze the concentrations of sphingolipids, including C16:0-LacCer, C18:0-LacCer, C18:1-LacCer, C24:0-LacCer, C24:1-LacCer, C14:0-Cer, C14:0-Cer, C16:0-Cer, C18:0-Cer, C18:1-Cer, C20:0-Cer, C22:0-Cer, C24:0-Cer, C24:1-Cer, SPA and Sph, in the sera of pediatric patients with gallstones to broaden the knowledge of the pathogenesis of this disease. Moreover, we analyzed the relationships between the concentrations of the sphingolipids and anthropometric measurements and biochemical lipid profile parameters.
This prospective study involved children with an initial diagnosis of cholelithiasis who were admitted to our Department from January 2017 to December 2018. They were diagnosed by abdominal ultrasound. None of the patients were treated with ursodeoxycholic acid at study entry. Patients included in the study group had no stone formation in other organs confirmed by abdominal ultrasound or magnetic resonance cholangiopancreatography (MRCP). In their family histories there were no diseases predisposing to cholelithiasis (e.g., cystic fibrosis). The family history was irrelevant, since only a few patients reported a positive family history of cholelithiasis. All patients had body mass index (BMI) calculated based on the World Health Organization; children were overweight or obese if their BMIs were ≥85th percentile. The control group included 38 children without any somatic organ pathology. Patients were excluded from the analysis if they were diagnosed with complications of cholelithiasis (for example, gallstone pancreatitis), bile duct defects, hemolytic or infectious diseases. None of the patients were taking medications that affect lipid or carbohydrate metabolism. Written informed consent was obtained from the parents of all the study participants. The protocol was approved by the local Bioethics Committee prior to patient recruitment, and the study was in accordance with the Helsinki Accords (approval number: R-I-002/393/2016, APK.002.464.2020).
All participants underwent a physical examination. Blood samples were taken from all participants after a 10 h overnight fast and were immediately centrifuged and frozen at −80 °C for further analysis. Total cholesterol (TC) and triglycerides (TG) levels were measured by a homogenous enzymatic colorimetric method, and values < 5.17 mmol/L for TC and 0.4–1.69 mmol/L for TG were considered normal.
The content of sphingolipids was measured using ultra-high performance liquid chromatography–tandem mass spectrometry (UHPLC/MS/MS) according to Bielawski et al. [13]. An internal standard mixture (Sph-d7, SPA-d7, C15:0-d7-Cer, C16:0-d7-Cer, C18:1-d7-Cer, C18:0-d7-Cer, 17C20:0-Cer, C24:1-d7-Cer, C24-d7-Cer, C16-LacCer, C17-LacCe, C18-LacCer) (Avanti Polar Lipids, Alabaster, Al, USA) and an extraction mixture (isopropanol:ethyl acetate, 15:85; v:v) were added to each sample (100 uL of serum). The mixture was vortexed, sonicated and then centrifuged for 5 min at 3000× g, 4 °C. The supernatant was transferred to a new tube, and the pellet was re-extracted. After centrifugation, the supernatants were combined and evaporated under nitrogen. The dried sample was reconstituted in 100 μL of LC Solvent B B (2 mM ammonium formate, 0.15% formic acid in methanol) or LC/MS/MS analysis. Sphingolipids were analyzed using a Sciex Qtrap 6500 + triple quadrupole mass spectrometer (SCIEX, Framingham, MA, USA) using a positive ion electrospray ionization (ESI) source (except S1P, which was analyzed in negative mode) with multiple reaction monitoring (MRM), against standard curves constructed for each analyzed compound. The chromatographic separation was performed on a reverse-phase Zorbax SB-C8 column, 2.1 × 150 mm, 1.8 um (Agilent Technologies, Santa Clara, CA, USA), in a binary gradient using 1 mM ammonium formate and 0.1% formic acid in water as solvent A, and 2 mM ammonium formate and 0.1% formic acid in methanol as solvent B at the flow rate of 0.4 mL/min.
Comprehensive data were processed using IBM SPSS Statistics 25.0 (Chicago, IL, USA) and shown as the median, minimum and maximum values. Statistical analysis was performed using the Mann–Whitney test for quantities and the chi-square test for categorical variables. Spearman’s correlation test was used to analyze the correlations between variables. A generalized multivariable linear model was created to determine the association between ceramides levels and the presence of cholelithiasis after adjusting for presence of age, obesity, sex and levels of TG and TC. Receiver operating characteristic (ROC) curves were generated by using the presence of cholelithiasis as a classification variable and concentrations of sphingolipids as prognostic variables (data were analyzed using Statistica 13.3 package, TIBCO Software Inc., Cracow, Poland). Statistical significance was considered to be met when the p value was less than 0.05.
The prospective study included 48 children and adolescents diagnosed with gallstones and 38 controls. Among patients with cholelithiasis, 21 (43.75%) were overweight/obese, and 27 (56.25%) had normal BMIs, whereas 15 (39.47%) were overweight/obese and 23 (60.53%) had normal BMIs in the control group. The demographic data and laboratory results of each group are presented in Table 1. All subjects included in the study were Caucasian.
Serum concentrations of TC, SPA, C14:0-Cer, C16:0-Cer, C18:1-Cer, C18:0-Cer, C20:0-Cer, C24:1-Cer, C16:0-LacCer, C18:0-LacCer, C18:1-LacCer, C24:0-LacCer and C24:1-LacCer differed significantly between patients with cholelithiasis and without cholelithiasis (Table 1).
The correlations between selected sphingolipids (with the greatest prognostic significance for cholelithiasis) and BMI, TC and TG are summarized in Table 2. The following significant correlations of ceramides with BMI were noted: C16:0-Cer, C12:0-Cer and C24:1-Cer; significant correlations with TG were found for C14:0-Cer, C24:1-Cer and C24:0-LacCer; and significant correlations for TG were determined for C14:0-Cer, C16:0-Cer, C18:1-LacCer, C24:1-LacCer and C24:0-LacCer. The ROC analysis presented in Table 3 and Figure S1 Supplementary Materials was performed to determine which sphingolipids had the best predictive value for distinguishing children with cholelithiasis from those without gallstones. The best result was determined for C16:0-Cer. A cut-off value of 59.692 ng/mL discriminated children with cholelithiasis with 97.9% sensitivity and 100% specificity (AUC = 1.0). Additionally, a high AUC value (0.99) was obtained for C14:0-Cer at the cut-off value 1.363 ng/mL with 95.8% sensitivity and 97.4% specificity.
After adjusting for age, gender, obesity and TC and TG levels, we found the best differentiating sphingolipids for cholelithiasis in the form of decreased SPA, C14:0-Cer, C16:0-Cer, C24:1-LacCer and C24:0-LacCer concentrations; and increased C20:0-Cer, C24:1-Cer, C16:0-LacCer and C18:1-LacCer (Table 4).
To our knowledge, this is the first report on sphingolipid analysis in pediatric patients with cholelithiasis. The available data on the role of ceramides in gallstone disease were obtained in animal models. In two animal studies, inhibition of ceramide biosynthesis by myriocin suppressed gallstone formation [14,15]. Myriocin is a natural inhibitor of serine palmitoyltransferase, an enzyme involved in the initial synthesis of sphingolipids [16]. This observation may suggest that sphingolipids are involved in gallstone pathogenesis. Unfortunately, the ceramide profile was not analyzed in this study. In another study in genetically engineered mice on a lithogenic diet, cholesterol gallstone formation was observed in 70% of cases, as opposed to 40% of mice on the same diet with supplementation of myrocin. At the same time, significant increases in the serum and bile ceramide concentrations were observed in mice on a lithogenic diet. In the group additionally treated with myrocin, the levels of ceramides were lower [15]. According to the authors, it is not known whether increases in serum and bile ceramides were directly related to the formation of gallstones; unfortunately, the ceramide profile was not analyzed in this study. In addition, in the group without myrocin and with increased serum ceramides also showed increased liver and ileum expression of ABCG5 and ABCG8 mRNA—i.e., the ATP binding cassette (ABC) transporter was noted. The ABC polymorphism has been reported to increase the risk of gallstone formation through the efflux of cholesterol in the liver [17]. Another factor that may link sphingolipids with gallstone formation is the presence of alkaline sphingomyelinase in human bile and liver, an enzyme that hydrolyzes sphingomyelin to ceramide in a bile-salts-dependent manner [18]. In our study, in patients with cholelithiasis, regardless of age, body weight, gender and levels of TC and TG, significant differences were observed in concentrations of SPA, C14:0-Cer, C16:0-Cer, C20:0-Cer, C24:1-Cer, C16:0-LacCer, C18:1-LacCer, C24:1-LacCer and C24:0-LacCer compared with children without gallstones. Our analysis showed that the most sensitive and specific markers of cholelithiasis among ceramides were decreased C14:0-Cer and C16:0-Cer. Regarding results for other diseases in children, elevated levels of ceramides C14:0-Cer and C16:0-Cer have been observed in non-alcoholic fatty liver disease (NAFLD) [19]. Moreover, Chang et al. described a bidirectional association between occurrence of NAFLD and cholelithiasis [20]. However, the mechanism of the coexistence of these two diseases has not been clearly identified. It is likely that insulin resistance, which increases de novo lipogenesis and activation of bile acid signaling pathways, may be involved in this process [21]. Due to the possible influence of circulating sphingolipids, patients with NAFLD were not included in our study. Maldonado-Hernández et al. also found higher levels of C14:0-Cer in adolescents with hepatic steatosis (HS). C14:0-Cer significantly correlated with total cholesterol in obese patients with HS [22]. Impaired cholesterol metabolism plays an important role in gallstone formation. In children with cholesterol gallstones, low intestinal cholesterol absorption was measured by increased serum plant sterols [23]. The accumulation of cholesterol and plant sterols in gallstones mirrors increased liver secretion of cholesterol [24]. We observed a significantly lower concentration of TC in children with cholelithiasis, regardless of body weight. However, we did not determine cholesterol metabolites and phytosterols. In our study, the composition of gallstones was not specified. We have also demonstrated that a decrease in C24:0-LacCer or C24:1-LacCer, or an increase in C18:1-LacCer, may predispose one to cholelithiasis. In pediatric patients with type 1 diabetes, lower levels of C18-LacCer and C24-LacCer were associated with an increased risk of progression in chronic kidney disease [25]. In other study conducted in adults, lactosylceramides were not associated with diabetes [26]. On the other hand, a higher amount of liver C24:0-LacCer was observed in patients with non-alcoholic steatohepatitis (NASH) than in the controls [27]. The above data may suggest that, depending on the disease, the concentrations of individual lactosylceramides may have protective or aggravating effects on the course of the disease. Metabolic syndrome is a potential risk factor for the increased occurrence of gallstones [28]. However, it is not known whether gallstones are a determinant of the development of metabolic syndrome and insulin resistance. Type 2 diabetes is one of the important risk factors for the development of gallstone disease in adult patients [29]. In our study, we found elevated C20:0-Cer levels in patients with cholelithiasis, but none of our patients had been diagnosed with metabolic syndrome at the time of evaluation. In adult studies, C20:0-Cer levels were higher in obese patients with type 2 diabetes mellitus; unfortunately, in this study there were no data on the coexistence of gallstone disease [30]. The plasma level of C20:0-Cer was higher not only in patients with type 2 diabetes mellitus, but also 3 years before diagnosis [31]. It would be interesting to see if patients already have abnormal level of C20:0-Cer or other ceramides prior to development of cholelithiasis. In an adult study it was reported that the risk of type 2 diabetes could be predicted using concentrations of C16:0-Cer, C18:0-Cer and C18:1-Cer. In addition, the loss of body weight is accompanied by significant reductions in the ceramide index and the risk of diabetes [32]. However, more research is needed to assess the effects of selected sphingolipid levels on the incidence of diabetes and gallstone development. Dietary fat consumption is also a risk factor for the development of cholesterol gallstones. In animal models, a high-fat diet was associated with increases in body weight, fasting glucose, and insulin levels. Moreover, the imbalance in glucose metabolism resulted in increased levels of C18:1-Cer, C18:0-Cer, C24:1-Cer and C24:0-Cer in submandibular gland cells [33]. In other studies, a relationship has also been observed between the increased dietary intake of saturated fats and the serum level of ceramides [34,35]. Ceramides are cholesterol-independent biomarkers of an overlapping but distinct spectrum of diseases [35]. Zabielski et al. also found higher concentrations of C20: 0-Cer in serum and liver in rats on a high-fat diet [36]. Other studies also attempted to assess the usefulness of ceramides in the courses of various diseases. Interestingly, in our study we found increased serum C16:0-LacCer in children with cholelithiasis independently of the other risk factors, such as overweight/obesity and gender. The same lactosylceramide was found to be very specific to children with Crohn’s disease, a type of inflammatory bowel disease [9]. It would be interesting to investigate whether the ileal expression of ABCG5 and ABCG8 is associated with this inflammatory bowel disease, which has as-yet-unknown pathogenesis, and to evaluate the effect of myrocin treatment on the course of intestinal inflammation in an animal model of inflammatory bowel disease. This is the first study in pediatric patients with cholelithiasis to investigate serum sphingolipids. The novelty of these findings is the main strength of our study and may be an introduction to further analyses in this age group. However, our work has several potential limitations. First, the small number of participants did not allow us to generalize our results. Our study was designed as an explanatory study to generate only pathophysiological theories of disease. The number of patients enrolled in the study was low due to the time span and the monocentric nature of the study. We understand that our results may be subject to errors of omission (type II error), and we did not interpret non-significant statistical results as underlying a true lack of differences. We also did not analyze the eating habits, physical activity and insulin resistance, which may also impact the results.
Our results suggest that serum sphingolipids are potential biomarkers for patients with cholelithiasis. However, it is important to conduct further research and answer the question of whether sphingolipids are factors regulating the formation of gallstones, or changes in their concentration are only secondary effects of cholelithiasis. It would be interesting to see if patients already have abnormal level of ceramides prior to the development of cholelithiasis. | true | true | true |
PMC9571230 | Neda K. Dezfuli,Ian M. Adcock,Shamila D. Alipoor,Babak Salimi,Sharareh Seifi,Mohammad Chehrazi,Mohammad Varahram,Esmaeil Mortaz | The miR-196a SNP Rs11614913 but not the miR-499 rs37464444 SNP is a Risk Factor for Non-small Cell Lung Cancer in an Iranian Population | 01-01-2022 | rs11614913, miR196a,rs3746444,miR-499,Lung Cancer,NSCLC | Background: Globally, lung cancer represents a major cause of cancer-related deaths. The regulation of gene expression is modulated by small noncoding RNAs called miRNAs that can act as both tumor suppressors and oncogenes. The maturation, expression and binding to target mRNAs is affected by single nucleotide polymorphisms (SNPs) in miRNA genomic regions thereby contributing to cancer susceptibility. SNPs Rs11614913 in miR196a and Rs3746444 in miR-499 are implicated in the development of cancers such as non-small cell lung cancer (NSCLC) in non-Arabic subjects. Materials and Methods: A small cohort of 204 participants including 104 lung cancer patients and 100 non-cancer controls subjects were enrolled into the study. The allele frequencies were determined by Polymerase Chain Reaction– Restriction Fragment Length Polymorphism (PCR-RFLP) and their correlation with lung cancer risk was determined. Results: The miR-196a rs11614913 polymorphism increased the risk of NSCLC (CC vs. TT+TC: OR= 2.26, 95%CI= 1.28 – 3.98, P= 0.0046) in a dominant genetic model. No statistically significant association was found between the miR-499 rs37464444 polymorphism and NSCLC. Conclusion: The rs11614913 polymorphism in miR-196a, but not the miR-499 rs37464444 polymorphism, increased the risk of NSCLC. Further studies with larger sample sizes in correlation with functional outcomes at the cellular level should be undertaken. | The miR-196a SNP Rs11614913 but not the miR-499 rs37464444 SNP is a Risk Factor for Non-small Cell Lung Cancer in an Iranian Population
Globally, lung cancer represents a major cause of cancer-related deaths. The regulation of gene expression is modulated by small noncoding RNAs called miRNAs that can act as both tumor suppressors and oncogenes. The maturation, expression and binding to target mRNAs is affected by single nucleotide polymorphisms (SNPs) in miRNA genomic regions thereby contributing to cancer susceptibility. SNPs Rs11614913 in miR196a and Rs3746444 in miR-499 are implicated in the development of cancers such as non-small cell lung cancer (NSCLC) in non-Arabic subjects.
A small cohort of 204 participants including 104 lung cancer patients and 100 non-cancer controls subjects were enrolled into the study. The allele frequencies were determined by Polymerase Chain Reaction– Restriction Fragment Length Polymorphism (PCR-RFLP) and their correlation with lung cancer risk was determined.
The miR-196a rs11614913 polymorphism increased the risk of NSCLC (CC vs. TT+TC: OR= 2.26, 95%CI= 1.28 – 3.98, P= 0.0046) in a dominant genetic model. No statistically significant association was found between the miR-499 rs37464444 polymorphism and NSCLC.
The rs11614913 polymorphism in miR-196a, but not the miR-499 rs37464444 polymorphism, increased the risk of NSCLC. Further studies with larger sample sizes in correlation with functional outcomes at the cellular level should be undertaken.
Lung cancer is currently the commonest cancer worldwide and has a poor prognosis being associated with very high mortality rates. Lung cancer is divided broadly into two main subtypes: small-cell (SCLC) and non-small-cell lung carcinoma (NSCLC). NSCLC accounts for 80–85% of all lung cancer cases whilst SCLC includes 12% of lung cancer cases particularly those with a high mortality (1, 2). Despite recent advances in the diagnosis and therapies available, lung cancer is still a major cause of death worldwide (3). Understanding the molecular pathology of lung cancer is critical for obtaining early diagnosis and thereby enabling initiation of timely and effective therapies. Although smoking is recognized as the major risk factor for the development of lung cancer, the disease also occurs in nonsmokers (4). Thus, genetics and lifestyle characteristics including diet, smoking and exposure to other environmental pollutants are important factors in the susceptibility and development of lung cancer (5). Inherited familiar gene changes in P53, Myc and breast cancer gene (BRCA)1 have been described in lung cancer (1, 6). Noncoding small RNAs (ncRNA) are key factors in the development and progression of lung cancer (7). MicroRNAs are members of the small ncRNA family and act post-transcriptionally to modulate gene expression (8). MicroRNAs regulate various biological functions and may act to control the expression of oncogenes and/or tumor suppressors (8). Dysregulation of miRNA expression provokes cancer invasion, metastasis and angiogenesis (8). MicroRNA networks coordinately modulate numerous genes in the body (2, 7). A SNP occurs in just under every 300bp of the genome including the coding and non-coding regions (9). However, most (93%) of SNPs that affect miRNA function are distributed within non-coding regions (10) and it is well known that SNPs associated with cancer susceptibility (2). The induction of the single nucleotide polymorphisms (SNPs) at a specific site, especially in non-coding regions, affects the induction and maturation of miRNAs in cancer (8). For example, miRNA 196a rs11614913 T/C and the miR-499 rs3746444 A/G polymorphisms are associated with the development of breast (11,12), lung (13–15), gastric (16), esophageal (17), hepatocellular (18), head and neck (19) and colorectal (20, 21) cancers. We hypothesize that polymorphisms within these miRNAs varies according to the ethnicity and geographical area of the patient. Thus, in this study we aimed to assess the possible association between miR-499 rs3746444, miR-196a rs11614913SNP in Iranian NSCLC patients.
One hundred and four patients with newly diagnosed based on pathology and clinical manifestation of NSCLC at age 58.1 ± 8.0 years old (mean ± SD) were recruited at the Masih Daneshvari hospital (Tehran, Iran) between April 2015 and September 2019. One hundred age- and gender-matched healthy controls subjects with a negative history of cancer and other inflammatory diseases were also enrolled in the study. The Ethics Committee of the Dr. Masih Daneshvari Hospital approved the study and all subjects gave written informed consent (Ethics committee approval number: IR.SBMU.MSP.REC.1397.525).
2 ml whole blood was collected into EDTA-containing tubes from all participants and genomic DNA isolated using a DNA High Pure PCR Template Preparation Kit, (Mannheim, Roche, Germany, Version 20, Cat.No.11796828001) as described by the manufacturer. The DNA concentration was measured by Nanodrop 2000 (Thermo Fisher, MA, USA). Specific SNPs were genotyped using polymerase chain reaction–restriction fragment length polymorphism (PCR-RFLP) with the PCR reaction performed using Taq DNA polymerase master mix (Invitrogen, Massachusetts USA), in a thermal cycler (Bio-Rad, California, USA). The primer sequences for each PCR reaction are shown in Table 1. PCR cycles were as follows: initial denaturation at 95°C for 5 min, 35 cycles of denaturation at 94°C for 30 sec, annealing at 58°C for 1min and extension at 72°C for 1 min and a final extension at 72°C for 10 min. To identify the mir-196 C/T polymorphism, the PCR product was digested with the restriction enzyme Taal (Thermo Fisher, USA) by incubating the samples at 65°C for 4h. The mir-499 T/C polymorphism PCR product was incubated at 37°C for 4h with the restriction enzyme TSP45I (Thermo Fisher, USA) and the digestion products were detected by 3% agarose gel electrophoresis.
The differences in genotype distribution for the two analyzed SNPs between patients and healthy subjects were analyzed using Chi-square test. All statistical analyses were carried out using SPSS-25 software (SPSS, Inc.). P values <0.05 were considered statistically significant.
The demographic information of participants including histological subtype, stage and smoking status are demonstrated in Table 2. 104 NSCLC patients and 100 healthy controls were enrolled in this study with mean age of 58.1 and 51.7 years, respectively (Table 2). For rs11614913 of miR196 the uncut PCR product size was 431bp and digested products shows bands at 281 and 150bp (Table 1, Figure 1A). The PCR product size for rs3746444 of miR499 was 302bp and the digested products showed bands of 111 and 191bp (Table 1, Figure 1B). The allele frequencies for rs11614913 and rs3746444 in patient and control groups are indicated in Table 3. The CC genotype of mir-196a rs11614913 was associated with an increased risk of lung cancer using a dominant genotype model (CC vs. TT+TC: OR= 2.26, 95%CI= 1.28 – 3.98, P= 0.0046). In contrast, the mir-499 rs3746444 variant was not associated with NSCLC in any inheritance model tested (Table 3). Rs3746444 variants were not associated with smoking status (Table 4) whereas the CC genotype frequency in smoking patients was higher than in smoking control subjects with the rs11614913 variant (OR=2.82, 95%CI=1.01–7.83, P=0.045) (Table 4). In addition, the C allele frequency of the rs11614913 variant correlated with stage II and III NSCLC (OR=0.09, 95%CI=0.012–0.67, P=0.019 and OR=0.14, 95%CI=0.02–0.92, P=0.041, respectively). There was no association between the rs3746444 SNP and disease stage (Table 5). Moreover, the rs11614913 SNP in miR196 was not associated with types of NSCLC disease, however, the C allele frequency in rs3746444 of miR499 variant was higher in the large cell carcinoma (LCC) subtype (OR=0.06, 95%CI=0.007–0.59, p=0.015) (Table 5).
The current study reports that the miR-196a2 rs11614913 polymorphism, but not the miR-499 rs3746444 polymorphism, was significantly associated with Iranian NSCLC patients. In addition, there was an association of the rs11614913 polymorphism with the CC genotype in smoking NSCLC patients and of the rs11614913 C allele frequency with stage II and III disease. There was also a higher frequency of the rs3746444 C allele in LCC patients. Previous studies reported that the rs11614913 polymorphism is associated with increased risk of lung cancer (13, 15). Meta-analysis of published studies reported that the rs11614913 polymorphism was associated with an increased risk of lung cancer particularly in Asian populations (15, 22). In our study, the rs11614913 polymorphism did not show any association with the types of NSCLC disease. Little information is available in Iranian NSCLC patients regarding miRNA polymorphisms although one study from North-East Iran also failed to find any significant association between the miR-196a2 rs11614913 polymorphism and lung cancer (23). This discrepancy between our study and the one from North-East Iran of similar size may be due to practical reasons such as usage of different enzymes, different participants from the two regions and type of NSCLC. Thus, larger, multi-centred studies across Iran using the same standardised methodology are required to confirm the role of the rs11614913 polymorphism in NSCLC patients in Iran. The CC genotype of rs11614913, which significantly increased the expression of the mature miR-196a, was associated with decreased survival of NSCLC patients (13). Furthermore, individuals carrying the TC or CC genotype of rs11614913 had an increased risk of lung cancer compared to those possessing the TT genotype among Chinese non-smoking females (14). In addition, miR-196a2 rs11614913 variant homozygote CC was associated with approximately 25% significantly increased risk of lung cancer in the Chinese population (24). Yoon and et al. reported that the rs11614913 genetic variant positively correlated with a better recurrence-free survival (RFS) in stage II and stage III of lung cancer. Overall, these findings indicated that the rs11614913 polymorphism is strongly associated with prognosis in NSCLC patients who undergo lung resection (25). Moreover, the rs11614913 genotypes were significantly associated with overall survival (OS) and disease-free survival (DFS) in women and in patients with stage II+IIIA disease, but not in men and patients with stage I disease (26). However, there was no difference in genotype-related adjusted hazard ratios (aHR) between the different subgroups of NSCLC (26). In contrast, the CC genotype in rs11614913 was associated with lower survival compared with TT/CT genotypes in NSCLC patients (13). In addition, the C allele frequency in rs11614913 was higher in stage II and III NSCLC in the current study. Importantly, a previous study has shown that the TC vs. TT genotype of rs11614913 is protective for NSCLC and may reduce the risk of NSCLC in the non-SCC subgroup (27). Furthermore, an earlier study reported a significant association between the miR-196a2 rs11614913 (CT/TT) genotype with NSCLC patients who are active smokers in a Korean population (28). Our data also showed that the rs11614913 polymorphism was associated with the CC genotype in smoker patients. The rs11614913 polymorphism is also associated with other cancers such as head and neck (19), hepatocellular (22) and breast cancers (29). The presence of any variant allele was associated with a significantly reduced risk of HNSCC but homozygous variant allele carriers with pharyngeal tumors had significantly reduced survival compared to wild-type and heterozygous forms (19). Moreover, the CC polymorphic genotype demonstrated associated with a decreased risk of breast cancer and the presence of the T allele was significantly associated with an increased risk of breast cancer (29). The functional SNP rs3746444 T/C within the miR499 gene causes an A/G transition in the mature miR-499 (13). The ethnic background of patients with the miR-499 rs3746444 polymorphisms may affect its impact on the susceptibility to lung cancer. The rs3746444 polymorphism is associated with decreased expression of miR-499 and poor survival in Chinese lung cancer patients (30). In contrast, Hau Qiu and co-workers found that the miR-499a rs3746444 genotypic distribution was not different in NSCLC cases and controls but this polymorphism elevated the susceptibility of NSCLC in a never smoking subgroup (adjusted P=0.035 for GG vs. AA genetic model and adjusted P=0.049 for GG vs. AA/AG genetic model) (27). Furthermore, Serena Vinci and colleagues did not find any association between miR-499 genotype and risk of NSCLC in 101 Italian NSCLC patients (31). A meta-analysis showed no association of the rs3746444 polymorphism and lung cancer in East Asian populations (32). However, another meta-analysis has shown an association between the microRNA-499 rs3746444 A/G polymorphism and cancer susceptibility in Asians, but not in Caucasians (33). Subgroup analysis of other cancer types, demonstrated no risk of breast, liver, or lung cancers with the microRNA-499 rs3746444 A/G polymorphism (33). An association between miR-499 rs3746444 and the susceptibility to cervical squamous cell carcinoma, prostate cancer, hepatocellular carcinoma (34), chronic obstructive pulmonary disease (35) and colorectal cancer has been reported (36). The G Allele of rs3746444 was also shown to be associated with the decreased risk of prostate cancer progression in a Serbian population (37). In addition, miR-499 rs3746444 increased the risk of cancer (38) but not for breast cancer (20, 39, 40). In another study, the rs3746444 G allele was associated with an increased cancer risk factor in Chinese subjects especially for breast cancer (41). This discrepancy may be due to difference in ethnic background, since, there was a significant association of rs3746444 with the susceptibility to cancers in Asians (12, 42, 43) but not in Caucasians (44). Hashemi and colleagues have shown that the miR-499 rs3746444 polymorphism increased the risk of prostate cancer in an Iranian population (45) whilst the rs3746444 T > C polymorphism was associated with high prevalence of cancer in Iranian and Chinese populations but low prevalence with esophageal cancer (44). There are some limitations in our study. First, this is a single center retrospective study. Due to the untimely COVID-19 epidemic, the sample size was small and lacked longitudinal samples. Future studies should include multi-centre studies across Iran and other middle eastern countries with different stages of NSCLC stages to verify these results. In addition, it will be important examine the functional impacts on the survival of NSCLC patients.
In conclusion, the current study data suggests an ethnic difference in the impact of rs3746444 T/C polymorphism in NSCLC lung cancer incidence. Our findings proposed that miR-196a2 rs11614913 polymorphism increased the risk of NSCLC lung cancer. In addition, the results do not support an association between the genetic variant of miR-499 rs37464444 and the risk of developing lung cancer. Additional larger clinical studies together with an analysis of the related cell functional outcomes are required. | true | true | true |
|
PMC9571420 | Rigbat Rozi,Yubo Zhou,Kai Rong,Pingbo Chen | miR-124-3p sabotages lncRNA MALAT1 stability to repress chondrocyte pyroptosis and relieve cartilage injury in osteoarthritis | 15-10-2022 | Osteoarthritis,microRNA-124-3p,LncRNA MALAT1,Pyroptosis,LncRNA stability,Chondrocytes | Background Osteoarthritis (OA) is a prevalent inflammatory joint disorder. microRNAs (miRNAs) are increasingly involved in OA. Aim Our study is proposed to clarify the role of miR-124-3p in chondrocyte pyroptosis and cartilage injury in OA. Methods OA mouse model was established via the treatment of destabilization of the medial meniscus (DMM), and the in vitro cell model was also established as mouse chondrocytes were induced by lipopolysaccharide (LPS). Mouse cartilage injury was assessed using safranin-O-fast green staining, hematoxylin–eosin staining, and OARSI grading method. Expressions of miR-124-3p, MALAT1, KLF5, and CXCL11 were determined. Cartilage injury (MMP-13, osteocalcin), inflammation (IL-6, IL-2, TNF-, IL-1β, and IL-18)- and pyroptosis-related factors (Cleaved Caspase-1 and GSDMD-N) levels were detected. Mechanically, MALAT1 subcellular localization was confirmed. The binding relationships of miR-124-3p and MALAT1 and MALAT1 and KLF5 were verified. MALAT1 half-life period was detected. Then, miR-124-3p was overexpressed using agomiR-124-3p to perform the rescue experiments with oe-MALAT1 or oe-CXCL11. Results miR-124-3p was downregulated in DMM mice and LPS-induced chondrocytes where cartilage injury, and increased levels of inflammation- and pyroptosis-related factors were found. miR-124-3p overexpression relieved cartilage injury and repressed chondrocyte pyroptosis. miR-124-3p bounds to MALAT1 to downregulate its stability and expression, and MALAT1 bounds to KLF5 to enhance CXCL11 transcription. Overexpression of MALAT1 or CXCL11 annulled the repressive function of miR-124-3p in chondrocyte pyroptosis. Conclusion miR-124-3p reduced MALAT1 stability and inhibited the binding of MALAT1 and KLF5 to downregulate CXCL11, thereby suppressing chondrocyte pyroptosis and cartilage injury in OA. | miR-124-3p sabotages lncRNA MALAT1 stability to repress chondrocyte pyroptosis and relieve cartilage injury in osteoarthritis
Osteoarthritis (OA) is a prevalent inflammatory joint disorder. microRNAs (miRNAs) are increasingly involved in OA.
Our study is proposed to clarify the role of miR-124-3p in chondrocyte pyroptosis and cartilage injury in OA.
OA mouse model was established via the treatment of destabilization of the medial meniscus (DMM), and the in vitro cell model was also established as mouse chondrocytes were induced by lipopolysaccharide (LPS). Mouse cartilage injury was assessed using safranin-O-fast green staining, hematoxylin–eosin staining, and OARSI grading method. Expressions of miR-124-3p, MALAT1, KLF5, and CXCL11 were determined. Cartilage injury (MMP-13, osteocalcin), inflammation (IL-6, IL-2, TNF-, IL-1β, and IL-18)- and pyroptosis-related factors (Cleaved Caspase-1 and GSDMD-N) levels were detected. Mechanically, MALAT1 subcellular localization was confirmed. The binding relationships of miR-124-3p and MALAT1 and MALAT1 and KLF5 were verified. MALAT1 half-life period was detected. Then, miR-124-3p was overexpressed using agomiR-124-3p to perform the rescue experiments with oe-MALAT1 or oe-CXCL11.
miR-124-3p was downregulated in DMM mice and LPS-induced chondrocytes where cartilage injury, and increased levels of inflammation- and pyroptosis-related factors were found. miR-124-3p overexpression relieved cartilage injury and repressed chondrocyte pyroptosis. miR-124-3p bounds to MALAT1 to downregulate its stability and expression, and MALAT1 bounds to KLF5 to enhance CXCL11 transcription. Overexpression of MALAT1 or CXCL11 annulled the repressive function of miR-124-3p in chondrocyte pyroptosis.
miR-124-3p reduced MALAT1 stability and inhibited the binding of MALAT1 and KLF5 to downregulate CXCL11, thereby suppressing chondrocyte pyroptosis and cartilage injury in OA.
Osteoarthritis (OA) is defined as a kind of degenerative joint disorder with the traits, such as long-term pain and organic dysfunction, and is becoming epidemic worldwide along with the growing rates of obesity and aging [1]. OA may occur in all joints and is prevalent due to overweight, excessive smoking, poor nutrition, occupational injury, physical activities, aging, joint disorganization, hereditary factor, and unhealthy lifestyle [2]. To date, joint replacement is the only effective strategy for OA, but its side effect might induce unsatisfactory clinic consequences and even discounted living quality [3]. Primarily, OA brings about severe cartilage injury and a high disability rate for subjects with OA [4]. Meanwhile, inflammatory factor release is observed in different sites of OA joints [5]. Furthermore, intensive and persistent pyroptosis in soft tissues might induce cartilage injury, augmented anguish, or hypersensitivity when subjected to injurious stimulation and eventually secrete cytokines to strengthen pain from OA [6]. Against this background, the management to alleviate chondrocyte pyroptosis and repair cartilage injury is warranted for OA therapy. microRNAs (miRNAs) are modulators of gene translation and metabolic activities of OA chondrocytes [7]. Besides, given their role in regulating gene expression and physiopathology of OA, they can be also used as makers of the disease and therapeutic targets [8]. miRNAs mediate the cellular biological behaviors including pyroptosis, to affect the pathogenesis and development of human disorders [9]. Significantly, miR-124-3p expression is decreased in OA [10]. miR-124-3p strengthens the anti-inflammatory efficacy of the candidate medicine in suppressing cartilage injury and protecting joint structure [11]. Moreover, miR-124-3p reverses the increasing levels of pyroptosis indicators [12]. Primarily, miRNAs influence the molecular modification and gene phenotypes of long non-coding RNAs (lncRNAs) to control the stability and expression of lncRNAs in different biological behaviors [13]. LncRNAs, a family of intensively researched RNAs, are also candidate biomarkers for OA by mediating OA development and providing diagnostic and prognostic insights for OA [14]. LncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) is accumulated in individuals assaulted by OA and exacerbates inflammatory reactions and cartilage injury [15]. On the other hand, MALAT1 can stimulate the release of inflammatory cytokines and exert pro-pyroptotic influences on pathological changes [16]. LncRNAs are involved in human diseases by interacting with transcription factors to modulate molecular activities [17]. Consistently, MALAT1 can positively regulate the level of transcription factor Krüppel-like factor 5 (KLF5) to enhance cellular damage in diabetic nephropathy [18]. KLF5 intensifies inflammatory symptoms, chondrocyte apoptosis, and cartilage degradation in OA [19]. Taking the above-listed evidence into consideration, we hypothesize that miR-124-3p might mediate chondrocyte pyroptosis and cartilage injury in OA via the modulation of MALAT1 and its downstream pathway.
This study was approved and supervised by the ethics committee of Traditional Chinese Medicine Hospital Affiliated to Xinjiang Medical University. The protocol was also approved by the Institutional Animal Care and Use Committee of Traditional Chinese Medicine Hospital Affiliated to Xinjiang Medical University and the Guidelines for the Care and Use of Laboratory Animals provisions of administration and usage of laboratory animals [20]. Significant efforts were made to minimize both the number of animals and their respective suffering.
Forty-eight male C57BL/6 mice [12 weeks old, Laboratory Animals Monitoring Institute, Guangzhou, Guangdong, China, SCXK (Guangdong) 2018-0044] were kept in 12-h light–dark cycles with constant temperature (22 ± 2 °C) and relative humidity (60%) for 1 week before the OA mouse model was established.
After a week of adjustable feeding, mice were subject to destabilization of the medial meniscus (DMM) to establish the OA mouse model. The mice were divided into the DMM group and the sham group. Mice in the DMM group were anaesthetized through an intraperitoneal injection of 2% pentobarbital sodium (40 mg/kg) and then subject to the right knee articular capsule incision at the medial patellar tendon, with the medial meniscotibial ligament (MMTL) transected using microsurgical scissors to establish the DMM model. Mice were given fluid resuscitation (1 mL/mouse sterile saline via subcutaneous injection) immediately after surgery. Mice in the sham group were treated in the same manner as the DMM group except for the transfection of MMTL. All mice had access to adequate food and water after surgery. After 48 h of the DMM surgery, agomiR-124-3p was injected into the joint of the OA mice, with agomiR-NC as the negative control (NC).
After 8 weeks of surgery, 1 mL of whole blood was collected by removing the eyeball. Next, the blood was rested for 40 min and centrifuged at 1000 × g for 10 min to separate the serum, which was preserved in a − 80 °C refrigerator for further experiments. After blood collection, mice were euthanized via an injection of 200 mg/kg sodium pentobarbital, and knee joint tissues were collected from each group. Six pairs of knee tissues from each group were prepared into tissue sections. Briefly, knee tissues were fixed in 4% paraformaldehyde for 24 h, decalcified in 10% ethylenediaminetetraacetic acid for 8 weeks, dehydrated in gradient ethanol, paraffin-embedded, and made into sections (5 μm) for subsequent experiments. Six pairs of knee tissues from each group were prepared into tissue homogenates for the following experiments.
The prepared 5-μm sections were stained with H&E or S–O for the observation of knee joint structure using an optical microscope (Olympus Optical Co., Ltd, Tokyo, Japan). OA cartilage degeneration was assessed by 3 independent researchers according to the Osteoarthritis Research Society International (OARSI) grading method [21]. The grading ranges from Grade 0 (normal) to Grade 6 as intact cartilage and surface, Grade 0; intact surface, Grade 1; surface incontinuity, Grade 2; vertical fracture, Grade 3; erosion, Grade 4; denudation, Grade 5; and deformation, Grade 6.
Primary mouse chondrocytes were isolated from neonatal C57BL/6 male mice [5 days old, Laboratory Animals Monitoring Institute, SCXK (Guangdong) 2018-0044] via the collagenase digestion method according to a previous study [22]. Briefly, mice were euthanized via the intraperitoneal injection of 200 mg/kg sodium pentobarbital to remove articular cartilage, which was detached using 3 mg/mL collagenase D at 37 °C with 5% CO2 for 90 min, detached using 0.5 mg/mL collagenase D 37 °C overnight, and then centrifuged at 400 × g for 10 min to discard the supernatant. The cell precipitate was resuspended in Dulbecco’s modified Eagle medium containing 100 IU/mL penicillin and 0.1 mg/mL streptomycin (Gibco Company, Grand Island, NY, USA). Next, chondrocytes (8 × 103 cells/cm2) were seeded into culture dishes, and the medium was refreshed after 2 d. The isolated chondrocytes were confluent after 6–7 d. The cells of passage 2 were employed for subsequent analysis. agomiR-124-3p, agomiR-NC, overexpression (oe)-MALAT1, oe-CXCL11, oe-NC, small interfering (si) RNAs of KLF5 (si-KLF5-1 and si-KLF5-2) and si-NC (all from GenePharma Co, Ltd, Shanghai, China) were transfected into cells following the instructions of Lipofectamine 2000 (Invitrogen Inc., Carlsbad, CA, USA). Subsequently, cells were preserved in the medium for 48 h for subsequent analysis.
The CCK-8 kit (Keygen Biotech Co., Ltd, Nanjing, Jiangsu, China) was employed to assess chondrocyte activity. Chondrocytes (5 × 103 cells/cm2) were seeded in 96-well plates, and 10 μL CCK-8 reagent was supplemented into each well, followed by the culture at 37 °C for 2 h. Optical density value at the wavelength of 450 nm was determined. Each procedure was repeated 3 times.
The total RNA was extracted using the TRIzol reagent (Invitrogen) and RNA was reverse-transcribed into cDNA via the RT kits (R&D Systems Inc., Minneapolis, MN, USA), and the purity and concentration of RNA were determined using Nano-Drop ND-1000 (NanoDrop Technologies Inc., Wilmington, DE, USA). U6 served as the internal reference of miR-124-3p, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as the internal reference of other genes. cDNA was screened using the QuantiTect SYBR Green PCR kits (Thermo Fisher, Shanghai, China) for RT-qPCR. All samples were employed for RT-qPCR via the ABI7500 qPCR system (Applied Biosystems, Inc., Carlsbad, CA, USA). The RT-qPCR primers are seen in Table 1. The relative expressions of genes were calculated using the 2−ΔΔCt method. Each procedure was repeated 3 times.
The total proteins were extracted from tissues or cells using PRO-PREPTM protein extraction buffer (iNtRON Biotechnology, Seongnam, South Korea), and the protein concentration was measured using the Bio-Rad protein assay kits (R&D). Protein samples were separated through sodium dodecyl sulfate–polyacrylamide gel electrophoresis and then transferred onto polyvinylidene fluoride membranes, followed by membrane culture with Tris-buffered saline with Tween 20 (TBST) containing 5% skim milk (BD Biosciences, Sparks, MD, USA) and incubation with rabbit monoclonal antibody Cleaved Caspase-1 (89332, 1:1000, Cell Signaling Technology, Beverly, MA, USA), rabbit monoclonal antibody gasdermin D N-terminal (GSDMD-N, ab8245, 1: 500, Abcam Inc., Cambridge, MA, USA) and rabbit polyclonal antibody GAPDH (ab9485, 1: 2500, Abcam) overnight. Next, the membranes were washed, incubated with goat anti-rabbit immunoglobulin G (IgG, ab6721, 1: 2000, Abcam), and then subjected to 3 washes with TBST. Bands were developed using enhanced chemiluminescence (Thermo Scientific, Pierce, Rockford, IL, USA), and the results were analyzed by the Image J software (National Institutes of Health, Bethesda, MD, USA), with GAPDH as the internal reference. Each experiment was conducted 3 times.
Osteocalcin (OC) and matrix metalloproteinase (MMP)-13 levels in the mouse serum were detected to evaluate mouse cartilage injury. The levels of interleukin (IL)-6, IL-2, and tumor necrosis factor-α (TNF-α) were assessed to analyze mouse inflammation, and variances of IL‐1β and IL‐18 in mouse serum and cell suspension were detected to evaluate cell pyroptosis under the instructions of the kits (R&D).
The binding relation between miR-124-3p and MALAT1 was predicted via StarBase (http://starbase.sysu.edu.cn/) [23]. The subcellular localization of MALAT1 was predicted through the LncATLAS website (http://lncatlas.crg.eu/) [24]. The binding relations between MALAT1 and KLF5 and between KLF5 and CXCL11 were predicted via the RNAInter database (http://www.rna-society.org/rnainter/) [25].
The PARIS™ kit (R&D) was employed for the fractionation of cytoplasm and nucleus in compliance with the manufacturers’ instructions. Cells (1 × 107 cells/well) were collected, resuspended in cell fractionation buffer for the separation of cytoplasm and nucleus, and then placed on ice for 10 min, and centrifuged at 4 °C and 5000 × g for 30 s. Next, following the instructions of the cell fractionation buffer, the supernatant and nuclear precipitation were separated and preserved for RNA extraction.
Wild type (WT) and mutant type (MUT) of MALAT1 fragments containing the binding sites of miR-124-3p were constructed into the pmirGLO-reporter vector (Beijing Huayueyang Biotechnology, Beijing, China). The constructed luciferase reporter plasmids were co-transfected with agomiR-NC or agomiR-124-3p into mouse chondrocytes. After 48 h, cells were collected, lysed, and the luciferase activity was evaluated according to the instructions of the luciferase assay kit (K801-200; Biovision, Mountain View, CA, USA). All steps were repeated 3 times.
RIP assay was conducted using the RIP kit (R&D) to verify the binding relations between miR-124-3p and MALAT1 and between MALAT1 and KLF5. Chondrocytes (1 × 107) were suspended in RIP lysis buffer (Thermo Fisher), upon which lysates were incubated with magnetic bead-coupled Argonaute-2 (ab186733, 1: 50, Abcam) or IgG (ab6709, 1: 1000, Abcam) at 4 °C overnight. The immunoprecipitation was collected, followed by the extraction of protein using protease K (Thermo Fisher) to extract RNA. The enrichment of miR-124-3p, MALAT1, and KLF5 was analyzed by RT-qPCR.
The Pierce™ Magnetic RNA–Protein Pull-Down Kit (R&D) was applied for RNA pull-down assay to assess the binding relation between MALAT1 and KLF5. The biotin was used to label the wild-type probe (Bio-MALAT1 probe-WT), the mutant-type probe (Bio-MALAT1 probe-MUT), and the negative control (Bio-control). Then, mouse chondrocytes were lysed using lysis buffer (Thermo Fisher), followed by incubation with biotin-labeled RNA prober (GenePharma, Shanghai, China). Subsequently, cells were treated with RNase-free DNase I (Thermo Fisher) and subject to RNeasy Mini Kit (R&D) to extract the binding RNA, which was applied for RT-qPCR analysis to detect the abundance of KLF5.
To evaluate the RNA half-life period of MALAT1, cells in the agomiR-124-3p group and the agomiR-NC group were treated with actinomycin D (2 μg/mL). Total RNA in cells was extracted at 0, 2, 4, 6, and 9 h after actinomycin D treatment. MALAT1 expression at the different time points after actinomycin D treatment was detected by RT-qPCR, with GAPDH serving as the internal reference. Each experiment was conducted 3 times.
SPSS 21.0 software (IBM Corp. Armonk, NY, USA) was appointed for data analysis, and GraphPad Prism 8.0 software (GraphPad Software Inc., San Diego, CA, USA) was used for graphing. The results were presented as mean ± standard deviation. All data were inspected with normality distribution and homogeneity test of variance. The Student’s t-test was appointed for comparison analysis between two groups, and one-way or two-way analysis of variance (ANOVA) was appointed for comparison analysis among multiple groups, and Tukey's multiple comparisons test was appointed for the post-test of data. The p value was attained using a two-tailed test, a value of p < 0.05 indicated a statistical significance, and a value of P < 0.01 indicated a highly statistical significance.
To investigate the regulatory role of miR-124-3p in OA, we established a DMM model to mimic OA mice. The results showed that DMM mice showed obvious OA signs, such as cartilage fiber and cartilage erosion in the knee joint, increased hyaline cartilage (HC) thickness, disorganized articular chondrocytes with a tendency to thicken (Fig. 1A), and elevated OARSI score (p < 0.01, Fig. 1B). The OC content in the serum of DMM mice was decreased (p < 0.01, Fig. 1C), while the contents of MMP-13, IL-6, IL-2, and TNF-α levels were increased (p < 0.01, Fig. 1D, E). In addition, miR-124-3p expression was downregulated in DMM mouse tissue (p < 0.01, Fig. 1F). Therefore, miR-124-3p expression was upregulated in vivo via injection of agomiR-124-3p (p < 0.05, Fig. 1F). miR-124-3p overexpression alleviated OA symptoms in DMM mice (Fig. 1A), lowered OARSI score (p < 0.05, Fig. 1B), increased the OC content in the serum (p < 0.01, Fig. 1C), and reduced the levels of, MMP-13, IL-6, IL-2, and TNF-α in the serum (p < 0.01, Fig. 1D, E).
Then, we tried to find out the molecular mechanism of pyroptosis in OA cartilage damage. miR-124-3p overexpression declined the levels of IL‐1β, IL‐18, Cleaved Caspase-1, and GSDMD-N in DMM mice (p < 0.01, Fig. 2A, B). Then, lipopolysaccharide (LPS) was employed to induce inflammatory injury in mouse chondrocytes and downregulated miR-124-3p expression in cells (p < 0.01, Fig. 2C). Next, miR-124-3p was overexpressed in chondrocytes using agomiR-124-3p (p < 0.01, Fig. 2C, D). Compared to the LPS group, the LPS + agomiR-124-3p group had enhanced cell activity (p < 0.01, Fig. 2E) and diminished the expressions of IL‐1β, IL‐18, Cleaved Caspase-1, and GSDMD-N (p < 0.01, Fig. 2F, G).
The binding relation between miR-124-3p and MALAT1 was revealed through the StarBase website (Fig. 3A), and then, we verified the binding of miR-124-3p to LncRNA MALAT in chondrocytes (p < 0.01, Fig. 3B, C), and miR-124-3p quenched MALAT1 stability (p < 0.05, Fig. 3D). In addition, MALAT1 expression was robustly upregulated after modeling, and miR-124-3p overexpression reduced MALAT1 expression (p < 0.05, Fig. 3E, F).
To elucidate the effects of MALAT1 on OA chondrocyte pyroptosis, oe-MALAT1 was transfected into cells and successfully overexpressed MALAT1 expression (p < 0.01, Fig. 4A), followed by rescue experiments with agomiR-124-3p. Compared with the LPS + agomiR-124-3p group, the LPS + agomiR-124-3p + oe-MALAT1 group showed decreased cell activity (p < 0.01, Fig. 4B), elevated IL‐1β and IL‐18 contents (p < 0.01, Fig. 4C), and increased expressions of Cleaved Caspase-1 and GSDMD-N (p < 0.01, Fig. 4D).
Next, the subcellular localization of MALAT1 was predicted via the LncATLAS website, and it was found that MALAT1 was principally localized in the nucleus (Fig. 5A). In addition, MALAT1 was mainly localized in the nucleus of chondrocytes (Fig. 5B). Via the RNAInter database, the binding relations between MALAT1 and KLF5 and between KLF5 and CXCL11 were predicted (Fig. 5C). Besides, the binding relation between KLF5 and MALAT1 was verified in chondrocytes (p < 0.01, Fig. 5D, E). si-KLF5-1 or si-KLF5-2 was transfected into cells and successfully downregulated KLF5 expression (p < 0.01, Fig. 5F), and it was found that KLF5 knockdown reduced CXCL11 expression (p < 0.05, Fig. 5G). In the DMM and LPS groups, KLF5 and CXCL11 expressions were highly expressed, miR-124-3p overexpression effectively decreased the expression levels of KLF5 and CXCL11, and the decrease was reversed by MALAT1overexpression (p < 0.05, Fig. 5H–K).
Eventually, we investigated the effects of CXCL11 on OA chondrocyte pyroptosis. oe-CXCL11 was transfected into cells to upregulate CXCL11 (p < 0.01, Fig. 6A), followed by rescue experiments with agomiR-124-3p. Compared with the LPS + agomiR-124-3p group, the LPS + agomiR-124-3p + oe-CXCL11 group showed decreased cell activity (p < 0.01, Fig. 6B), elevated IL‐1β and IL‐18 contents (p < 0.01, Fig. 6C), and increased the expressions of Cleaved Caspase-1 and GSDMD-N (p < 0.01, Fig. 6D).
OA refers to serious cartilaginous disease with high morbidity and leads to mounting health and welfare burdens and loss of functions [26]. Pyroptosis-induced chondrocyte absence and degradation is the principal culprit of inflammatory infiltration and cartilage injury during OA development [27]. Non-coding RNAs, particularly miRNAs, are closely correlated to musculoskeletal injuries in the diseases [8, 28]. miR-124-3p expression is repressed in OA, which is accompanied by joint degradation and dysfunction [11]. In addition, miR-124 affects the expressions of pyroptotic factors to manipulate cell viability and migration [29]. Based on the information, we attempt to discuss the exact mechanism of miR-124-3p in cartilage injury and chondrocyte pyroptosis in OA (Fig. 7). miRNAs are necessary players and significant indicators of cartilage homeostasis and chondrocyte reproduction in OA as their deficiency exasperates osteal ill-growth and degradation and their involvement regulates molecular progression, cell death, and self-renewal of chondrocytes [30]. miR-124 was weakly expressed in OA, while its expression was reversed with the treatment of a possible medicine [31]. To investigate the role of miR-124-3p in OA, the OA mouse model was established through DMM treatment, and the results unveiled that OA mice showed cartilage fibrosis and erosion in the knee joint, increased HC thickness, disorganized articular chondrocytes with a tendency to thicken, elevated OARSI score, lowered the OC content, increased MMP-13, IL-6, IL-2, and TNF-α levels in the serum, and decreased miR-124-3p expression level in tissue, while these symptoms were all reversed upon agomiR-124-3p treatment. Expectedly, miR-124 overexpression mitigated cartilage tissue damage and encouraged chondrocyte self-renewal [32]. In patients with osteoporosis, OC content was quenched and miR-124-3p was downregulated [33]. Besides, miR-124-3p attenuated inflammatory damages in DMM rats by reducing the contents of IL-6, IL-2, and TNF-α [10]. In summary, miR-124-3p was poorly expressed in OA mice, while miR-124-3p overexpression alleviated cartilage damage in OA mice. Mechanically, as a process resulting from inflammation and related to cellular programmed death, pyroptosis takes an active part in OA pathological progression [34]. miRNAs mediate pyroptotic changes in many human disorders [35]. Although miR-124-3p can render its benign effects on pyroptosis elimination in pulmonary hypertension [12], the mechanism of miR-124-3p in pyroptosis in OA is poorly understood. To find out the molecular mechanism of chondrocyte pyroptosis in OA, LPS was employed to induce inflammatory injury in OA chondrocytes with agomiR-124-3p treatment, after which IL‐1β, IL‐18, Cleaved Caspase-1, and GSDMD-N levels were all diminished. miR-124 can abrogate the inflammatory reactions elicited by LPS [36]. Interestingly, miR-124 upregulation plays a protective role in cerebral ischemia–reperfusion injury by restricting pyroptosis as evidenced by the limited expressions of IL‐1β, IL‐18, Caspase-1, and GSDMD [37]. Collectively, miR-124-3p suppressed chondrocyte pyroptosis in OA. The regulatory effects of miRNAs on lncRNA transcription and stability in chronic inflammatory conditions have attracted much attention [38]. LncRNAs influence a variety of gene biological behaviors, and they are closely associated with the incidence and severity of OA [39]. The binding of miR-124-3p and MALAT1 was verified in this experiment. To confirm the effects of miR-124-3p on MALAT1 stabilization, cells with miR-124-3p overexpression were treated with actinomycin D, upon which the half-life period of MALAT1 was inhibited. Importantly, a recent finding demonstrated that miR-124-3p and MALAT1 were negatively correlated, and miR-124-3p inhibited MALAT1 expression [40]. The above data demonstrated that miR-124-3p bounds to MALAT1 to reduce its stability and expression. MALAT1 expression was aggregated in OA, which quenched cell viability and enhanced cartilage injury [41]. The pyroptosis property of lncRNAs is a crucial mechanism in the research and treatment of several diseases [42]. To further elucidate the effects of MALAT1 on OA chondrocyte pyroptosis, oe-MALAT1 was transfected into LPS-induced cells to overexpress MALAT1 expression, and to conduct rescue experiments with agomiR-124-3p. MALAT1 overexpression decreased cell activity, elevated IL‐1β and IL‐18 contents, and increased expressions of Caspase-1 and GSDMD-N. MALAT1 overexpression catalyzed pyroptosis in diabetic nephropathy to accelerate kidney impairment with the involvement of increased IL‐1β, IL‐18, Caspase-1, and GSDMD [43, 44]. Essentially, MALAT1 upregulation could augment the inflammatory injury induced by LPS treatment, to debilitate hindlimb mobility of rodents with acute spinal cord injury [45]. It is plausible that MALAT1 overexpression reversed the protective role of miR-124-3p overexpression in chondrocyte pyroptosis. Subsequently, we found that MALAT1 could bind to KLF5. It was previously discovered that the binding of miR-124-3p, MALAT1, and KLF5 could modulate pulmonary artery hypertension development, and MALAT1 was positively correlated to KLF5 [46]. Furthermore, KLF5 overexpression was responsible for chondrocyte hypertrophy and inflammatory damage in DMM-triggered OA [47]. Afterwards, the binding of KLF5 and CXCL11 was confirmed. CXCL11 was upregulated in rheumatoid arthritis, with the manifestation of exacerbated inflammatory injury and limited cell activity [48]. To further elucidate the effects of CXCL11 on OA chondrocyte pyroptosis, oe-CXCL11 was transfected into cells with LPS treatment to overexpress CXCL11 and to conduct rescue experiments with agomiR-124-3p. CXCL11 overexpression aggravated pyroptosis of OA chondrocytes. CXCL11 stimulated tissue lesions under an inflammatory microenvironment together with elevated IL‐1β and IL‐18 levels [49]. Additionally, CXCL11 ablation reduced cartilage degradation and inflammatory injury in LPS-induced rheumatoid arthritis [50]. Collectively, CXCL11 overexpression reversed the protective role of miR-124-3p overexpression in chondrocyte pyroptosis. In conclusion, our findings supported that miR-124-3p bounds to MALAT1 to reduce its stability and expression, repressed the binding of MALAT1 and the transcription factor KLF5 and suppressed CXCL11 transcription, thereby inhibiting chondrocyte pyroptosis and cartilage damage in OA. These findings hinted at a therapeutic strategy for OA alleviation. However, our research just explored the mechanism at the cellular level. In the future, we will further work out to probe the effects of miR-124-3p/MALAT1/KLF5/CXCL11 axis on chondrocyte pyroptosis and cartilage damage in OA through animal experiments. | true | true | true |
|
PMC9571429 | Masahiro Yura,Kazumasa Fukuda,Satoru Matsuda,Tomoyuki Irino,Rieko Nakamura,Hirofumi Kawakubo,Hiroya Takeuchi,Yuko Kitagawa | Effects of let-7a microRNA and C–C chemokine receptor type 7 expression on cellular function and prognosis in esophageal squamous cell carcinoma | 15-10-2022 | Esophageal squamous cell carcinoma,C–C chemokine receptor type 7,Let-7a microRNA,Invasive ability,Metastasis | Background C–C chemokine receptor type 7 (CCR7) participates in chemotactic and metastatic responses in various cancers, including in esophageal squamous cell carcinoma (ESCC). The microRNA (miRNA) let-7a suppresses migration and invasion of various types of cancer cells by downregulating CCR7 expression. Methods The expression levels of CCR7 and let-7a were measured in the cell lines, tumor, and peritumoral tissues of ESCC patients. KYSE cell lines were transfected with synthetic let-7a miRNA and a let-7a miRNA inhibitor, and their CCR7 expression levels as well as invasive ability were evaluated. A highly invasive cell line was established via an invasion assay, and CCR7 expression level along with let-7a level was subsequently evaluated. Cancer cells overexpressing CCR7 were injected subcutaneously into mice, and the animals were monitored for tumor growth along with lymph node metastasis. Results A negative correlation between CCR7 and let-7a expression was observed in the ESCC cell lines as well as in tissue samples from patients. Synthetic let-7a decreased CCR7 expression level, while the let-7a inhibitor increased it. In vitro, the established highly invasive cancer cells with high and low levels of CCR7 and let-7a expression, respectively, exhibited a greater invasive ability than the wild-type cell line. The cells were associated with tumor growth and lymph node metastasis in mice. Patients in the high-CCR7/low-let-7a group had the worst prognosis, with a five-year recurrence free survival (5-RFS) rate of 37.5%, followed by the high-CCR7/high-let-7a (5-RFS: 60.0%) and low-CCR7 (5-RFS: 85.7%; p = 0.038) groups. Conclusions The expression of CCR7 was downregulated by let-7a miRNA in esophageal cancer cells. The decrease in let-7a expression level led to the increased expression level of CCR7 in ESCC cells, consequently increasing their invasive ability and malignancy and resulting in a worse prognosis for ESCC patients. Trial registration. Retrospectively registered. Supplementary Information The online version contains supplementary material available at 10.1186/s12885-022-10178-2. | Effects of let-7a microRNA and C–C chemokine receptor type 7 expression on cellular function and prognosis in esophageal squamous cell carcinoma
C–C chemokine receptor type 7 (CCR7) participates in chemotactic and metastatic responses in various cancers, including in esophageal squamous cell carcinoma (ESCC). The microRNA (miRNA) let-7a suppresses migration and invasion of various types of cancer cells by downregulating CCR7 expression.
The expression levels of CCR7 and let-7a were measured in the cell lines, tumor, and peritumoral tissues of ESCC patients. KYSE cell lines were transfected with synthetic let-7a miRNA and a let-7a miRNA inhibitor, and their CCR7 expression levels as well as invasive ability were evaluated. A highly invasive cell line was established via an invasion assay, and CCR7 expression level along with let-7a level was subsequently evaluated. Cancer cells overexpressing CCR7 were injected subcutaneously into mice, and the animals were monitored for tumor growth along with lymph node metastasis.
A negative correlation between CCR7 and let-7a expression was observed in the ESCC cell lines as well as in tissue samples from patients. Synthetic let-7a decreased CCR7 expression level, while the let-7a inhibitor increased it. In vitro, the established highly invasive cancer cells with high and low levels of CCR7 and let-7a expression, respectively, exhibited a greater invasive ability than the wild-type cell line. The cells were associated with tumor growth and lymph node metastasis in mice. Patients in the high-CCR7/low-let-7a group had the worst prognosis, with a five-year recurrence free survival (5-RFS) rate of 37.5%, followed by the high-CCR7/high-let-7a (5-RFS: 60.0%) and low-CCR7 (5-RFS: 85.7%; p = 0.038) groups.
The expression of CCR7 was downregulated by let-7a miRNA in esophageal cancer cells. The decrease in let-7a expression level led to the increased expression level of CCR7 in ESCC cells, consequently increasing their invasive ability and malignancy and resulting in a worse prognosis for ESCC patients. Trial registration. Retrospectively registered.
The online version contains supplementary material available at 10.1186/s12885-022-10178-2.
Cancer incidence and mortality are rapidly increasing worldwide. In 2018, esophageal cancer ranked seventh and sixth in terms of incidence and overall mortality, respectively [1]. Although the survival rate of esophageal cancer has improved due to the development of multidisciplinary treatment, the five-year post-esophagectomy survival rate is approximately 50% [2, 3]. Therefore, there is an urgent need to investigate the mechanisms of cancer progression and metastasis to improve the prognosis of patients with esophageal cancer. Chemokines regulate tumor cell proliferation, infiltration, and metastasis [4, 5]. In a previous study, we demonstrated the significant clinicopathological relationship and functional causality between C–C chemokine receptor type 7 (CCR7) expression and lymph node metastasis in patients with esophageal squamous cell carcinoma (ESCC) [6]. MicroRNAs (miRNAs) are non-coding small RNA molecules that can control the translation of mRNAs and regulate several cancer-related genes [7–9]. The miRNA let-7a suppresses various types of cancers [10–13]. For example, in breast cancer, let-7a suppresses the expression of CCR7 and reduces the ability of cancer cells to migrate and invade. Furthermore, a recent prospective study has reported that high let-7a expression level can be a predictive factor for favorable response to chemotherapy [14]. For gastric cancer, a lack of let-7a expression increases CCR7 expression level and is associated with metastasis, contributing to a poor prognosis [12]. Similarly, in esophageal cancer, plasma levels of let-7a miRNA are significantly lower in cancer patients than in healthy participants [15]. However, the effect of let-7a on molecular expression is still unclear, and it has not been confirmed whether the downregulation of let-7a miRNA expression is responsible for increased CCR7 expression levels in ESCC tissues. Here, we investigated the relationship between CCR7 and let-7a miRNA expression as well as the underlying regulatory mechanism, in esophageal cancer cell lines, tumor tissues, and peritumor tissues of patients with ESCC.
Tissue samples were obtained during a biopsy from 17 ESCC patients at the Keio University School of Medicine in Japan. Immediately after the procedure, the samples were frozen in liquid nitrogen and stored at -70 °C until use.
For this study, we used six established ESCC cell lines from the KYSE series (KYSE-350, 510, 590, 1260, 1440, and 2400), purchased from the Japanese Collection of Research Bioresources Cell Bank of the National Institutes of Biomedical Innovation, Health, and Nutrition.
Cells were maintained and cultured in a Roswell Park Memorial Institute 1640 medium supplemented with 10% fetal bovine serum, 100 units/mL penicillin, and 100 µg/mL streptomycin at 37 °C and 5% CO2 atmospheric content. For the experiments, cells were collected and prepared as single-cell suspensions in phosphate-buffered saline.
The sequence of double-stranded RNA (dsRNA) used in transfection experiments as scrambled small interfering RNA was 5′-UCACAACCUCCUAGAAAGAGUAGA-3′, that of synthetic let-7a-5p miRNA was 5′-UGAGGUAGUAGGUUGUAUAGUU-3′, while inhibitor let-7a-5p miRNA was 5′-UGAGGUAGUAGGUUGUAUAGUU-3′. These RNAs were synthesized by Applied Biosystems (Tokyo, Japan). The cell lines were transfected with dsRNA using the reagents Lipofectamine RNAiMAX and Lipofectamine 2000 (Invitrogen, Tokyo, Japan), according to the reagent manufacturer’s instructions. The cell lines were harvested 2 d after transfection and subjected to various analyses.
Isolation of let-7a miRNA from the cell lines was conducted using the mirVana™ miRNA Isolation Kit. The RNA concentration was quantified via the NanoDrop ND-1000 Spectrophotometer (Thermo Fisher Scientific, Tokyo, Japan). The complementary DNA (cDNA) was synthesized by the reverse transcription of let-7a miRNA. A quantitative reverse transcription polymerase chain reaction (RT-PCR) was conducted using ViiA™ 7 (Applied Biosystems); the thermal cycling consisted of an initial cycle of 2 min at 50 °C and 10 min at 95 °C, followed by 40 cycles for 15 s at 95 °C and 60 s at 60 °C. Relative let-7a expression level was quantified using the 2−ΔΔCt method.
Isolation of total RNA from cell lines was conducted using an RNeasy® Micro Kit (QIAGEN, Tokyo, Japan). The RNA concentration was quantified using a NanoDrop ND-1000 Spectrophotometer (Thermo Fisher Scientific). Then, synthesis of cDNA from total RNA was performed using an RNA-to-cDNA kit (Applied Biosystems). The quality and quantity of the cDNA samples were evaluated via standard electrophoresis and a NanoDrop ND-1000 Spectrophotometer (Thermo Fisher Scientific).
The quantitative RT-PCR analysis was performed via ViiA™ 7 (Applied Biosystems) and the Fast SYBR® Green Master Mix (Applied Biosystems). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and CCR7 expression levels were evaluated using PCR primers. The thermal cycling consisted of an initial 20 s at 95 °C, followed by 40 cycles for 1 s at 95 °C and 20 s at 60 °C, using GAPDH as an internal control. The relative amount of CCR7 expression in KYSE cell lines was calculated using the 2−ΔΔCt method, wherein the KYSE-350 expression level was defined as 1.
A synthetic let-7a and a let-7a inhibitor were used for transfection. We diluted 15 μL of the Lipofectamine® RNAiMAX reagent (Life Technologies) in 2 mL of Opti-MEM® (Life Technologies), then added 2 mL of diluted let-7a miRNA (300 pmol) to diluted Lipofectamine® RNAiMAX and incubated the mixture for 5 min at room temperature. Next, we plated the transfection reagent onto a 10 cm plate, which was incubated for 10–20 min in a CO2 incubator at 37 °C. A suspension of 5 × 105 cells in an antibiotic-free medium was added to the plate and incubated for 3 days in a CO2 incubator at 37 °C.
To create highly invasive cell lines, we used Corning BioCoat Matrigel Invasion Chambers containing a polyester membrane with 8 µm pores and a thin layer of a Matrigel Basement Membrane Matrix. We prepared a suspension of 2 × 104 cells/mL in culture medium for 24-well chambers and incubated it in an invasion chamber for 72–96 h in a humidified tissue culture incubator at 37 °C in 5% CO2 atmosphere. After incubation, the non-invading cells were removed from the upper surface of the membrane while the cells in the lower surface were collected and reseeded to the chamber. After repeating this process six times, a highly invasive cell population was obtained and isolated using a cell dissociation solution, solubilized, and stained using a staining solution. The number of infiltrated cells were quantified by measuring the absorbance of the solution.
A lymph node metastasis model was developed through the following steps: 1) a xenograft tumor was created subcutaneously by injecting 1 × 107 wild-type and invasive KYSE-510 cells into five-week-old mice, visualized with Green Fluorescent Protein (GFP) transfection; 2) after growth (4–6 weeks post injection), the resulting tumors were resected and cut into small pieces (approximately 2 to 8 mm3), then transplanted into the elbows of different five-week-old mice; 3) the accessory axillary lymph nodes were removed and examined 8 weeks after transplantation.
Statistical analysis was done using the IBM SPSS statistics 26.0 software. The continuous variables were expressed as mean ± SD and compared with the Student’s t-test. Recurrence-free survival (RFS) was measured from the time of surgery until the recurrence of tumor or death, whichever came first. The RFS curves were estimated with the Kaplan–Meier method and compared by the log-rank test. Univariate analysis was then performed using Cox’s proportional hazard model, and a p-value of 0.05 was considered as statistically significant.
Using quantitative RT-PCR analysis, CCR7 mRNA expression was confirmed in all esophageal cancer cell lines (Fig. 1A), and the level of let-7a miRNA expression was also determined (Fig. 1B). KYSE-510 and 1440 exhibited significantly high CCR7 and low let-7a expression levels, whereas KYSE-590 had significantly low CCR7 and high let-7a expression levels. The expression levels of the CCR7 and let-7a were normalized based on the expression levels in KYSE-350 cell, which were defined as 1 (Table S1).
As determined via quantitative RT-PCR analysis, the expression of CCR7 was downregulated by the transfection of let-7a in five of the six KYSE cell lines (Fig. 2A). The decrease in CCR7 expression level was most prominent in KYSE-1440; CCR7 expression level was the highest in the wild-type.
The inhibitory effect of let-7a was determined by using synthetic anti-let-7a oligonucleotides in KYSE-590 esophageal cancer cells, which expressed a low level of CCR7 and a high level of let-7a in the wild-type. After transfection with synthetic anti-let-7a oligonucleotides, the level of CCR7 expression in KYSE-590 cells increased (p < 0.001; Fig. 2B), along with their invasive ability (p = 0.061; Fig. 2C).
After six courses of the invasion assay, the invasive ability of the highly invasive lines increased (KYSE-510: p = 0.004, KYSE-590: p = 0.001, Fig. 3A–B). Compared to that in the wild-type, CCR7 expression level increased (KYSE-510: p = 0.04, KYSE-590: p = 0.203, Fig. 3C) in the invasive type, but let-7a expression level decreased (KYSE-510: p < 0.001, KYSE-590: p = 0.001, Fig. 3D). Since the KYSE-510 cell line alone could be injected into mice for the in vivo experiment, we used it for the mouse model. The implanted primary tumor of invasive KYSE-510 was both larger (6.5 ± 3.51 vs. 13.0 ± 3.31 mm, p = 0.025) and heavier (0.158 ± 0.161 vs. 0.481 ± 0.263 g, p = 0.070) than that of wild-type cells (Fig. 4A–D). Similarly, the percentage of positive lymph node metastasis of the invasive-type cells, labeled with GFP, was higher than that of the wild-type — 25% (1/4) and 0% (0/5), respectively (Fig. 4E–F).
Tissue samples were obtained from 17 ESCC patients during a biopsy, and their characteristics are listed in Table 1. Figure 5A shows the expression levels of CCR7 and let-7a in cancer tissue relative to those in normal tissue, defined as 1 based on quantitative RT-PCR. The graph uses a logarithmic notation, thus presenting negative values when the CCR7 and let-7a expression level ratio is less than 1. In 11 ESCC patients, CCR7 expression levels were higher in malignant tissues than in normal tissues (positive value). In contrast, in a different group of 11 ESCC patients, let-7a expression was downregulated more in malignant tissues than in normal tissues (negative value). Around nine ESCC patients exhibited a negative correlation between CCR7 and let-7a expression.
Recurrence of tumor was observed in six patients, and the most predominant sites were the lymph nodes. Using the median value, patients were split into two respective groups with low and high CCR7 expression levels; patients with high CCR7 expression levels were further divided into two sub-groups with low and high let-7a expression levels. The high-CCR7/low-let-7a group (5-RFS: 37.5%) had the worst prognosis, followed by the high-CCR7/high-let-7a (5-RFS: 60.0%), and then the low-CCR7 group (5-RFS: 85.7%; p = 0.038, Fig. 5B). Cox proportional hazards analysis identified the high-CCR7/low-let-7a group as a significant prognostic factor (Table 2).
This study showed that synthetic let-7a and the let-7a inhibitor respectively decreased and increased CCR7 expression levels in KYSE cells, demonstrating a negative correlation between CCR7 and let-7a expression. Therefore, the downregulation of let-7a miRNA is a factor contributing to increased CCR7 expression levels in esophageal cancer cells and the development of malignant tumors. High CCR7 expression level is related to lymphatic tumor metastasis and prognosis in gastric as well as colorectal cancer [16, 17]. Crucially, the elevated CCR7 expression level caused by a deficiency in let-7a expression level is intricately linked to the metastasis and progression of cancer [12]. Furthermore, CCR7 is correlated with lymphatic metastasis and poor prognosis in different types of breast cancer, while let-7a has been found to suppress breast cancer cell migration and invasion through the downregulation of CCR7 expression [10, 18]. Thus, our study confirmed the negative correlation between CCR7 and let-7a expression in ESCC, which strongly suggests that let-7a, suppresses the expression of CCR7. Regarding the direct interaction between CCR7 and let-7a, Kim et al. used the luciferase assay to show that let-7a directly regulates CCR7 expression by binding with its 3′-UTR [10]. In addition, we can check what type of gene sequence can be regulated by specific miRNA at Target Scan Human 8.0 [19]; it showed that CCR7 is regulated by let-7a binding its 3′-UTR. In normal human cells, CCR7 is mainly expressed in differentiated lymphocytes as well as the surface of dendritic cells, and it is thought to mediate lymphocyte migration [20]. The invasive ability of cancer cells is affected by the expression of CCR7 [10, 18]. In this study, the invasive ability of ESCC increased after transfection with synthetic anti-let-7a because of the increased and decreased CCR7 and let-7a expression levels, respectively. After the invasion assay, the expression level of CCR7 in the invasive-type cells increased while that of let-7a decreased, in comparison with those in the wild-type. Therefore, the respective increase and decrease in CCR7 and let-7a expression levels are important factors affecting the invasive ability of ESCC. We also used the highly invasive cell line for in vivo experiments on mice. The implanted primary tumor of the invasive-type was larger in size and mass than that of the wild-type, and the percentage of positive lymph node metastasis was higher as well. Therefore, the highly invasive cell line was more likely to cause and accelerate lymph node metastasis in vivo. We also investigated the expression of CCR7 and let-7a miRNA in ESCC patient tissues and demonstrated that they tend to have a higher CCR7 and a lower let-7a miRNA expression levels than normal tissues. Furthermore, we found that the expression levels of CCR7 and let-7a miRNA can stratify the prognosis of the ESCC patients. Patients in the two high-CCR7 groups had a worse prognosis than those in the low-CCR7 group, with the low-let-7a group exhibiting a worse prognosis than the high-let-7a group. These results suggest that during carcinogenesis, the suppression of CCR7 expression diminishes owing to the decreased expression level of let-7a, and that these changes may affect the prognosis of ESCC patients. Supporting this finding, He et al. [15] found that the plasma levels of let-7a miRNA were significantly lower in ESCC patients than in healthy participants. However, our study has three main limitations. First, while we confirmed the negative correlation between CCR7 and let-7a, we did not account for the potential influence of other oncogenes controlled by let-7a. Second, in the mouse experiment, we demonstrated the in vivo malignancy of the synthesized highly invasive cell line but did not measure CCR7 or let-7a expression levels in metastatic lymph nodes. Nevertheless, the relationship between CCR7 expression level and lymph node metastasis in ESCC cells has been examined in a previous study [6]. Third, the number of cell lines and patient tissue samples used in this study was relatively small; therefore, additional studies are required to validate these results.
The expression of CCR7 was found to be downregulated by let-7a miRNA in ESCC cells. Thus, a decrease in let-7a expression level led to an increase in the expression level of CCR7 in ESCC cells, which consequently acquired increased invasive ability and malignancy, resulting in a worse prognosis for patients with ESCC.
Additional file 1: Table S1. Quantitative RT-PCR analysis. | true | true | true |
|
PMC9571456 | Yifan Tang,Yanqing Sun,Junkai Zeng,Bo Yuan,Yin Zhao,Xiangwu Geng,Lianshun Jia,Shengyuan Zhou,Xiongsheng Chen | Exosomal miR-140-5p inhibits osteogenesis by targeting IGF1R and regulating the mTOR pathway in ossification of the posterior longitudinal ligament | 15-10-2022 | OPLL,miR-140-5p,MSC,Exosome | Background Ossification of the posterior longitudinal ligament (OPLL) is a disabling disease whose pathogenesis is still unclear, and there are no effective cures or prevention methods. Exosomal miRNA plays an important role in the osteogenesis of ectopic bone. Therefore, we focused on the downregulation of miR-140-5p in OPLL cell-derived exosomes to explore the mechanism by which exosomal miR-140-5p inhibits osteogenesis in OPLL. Results Exosomes were isolated by differential centrifugation and identified by transmission electron microscopy, nanoparticle tracking analysis, and exosomal markers. Exosomal RNA was extracted to perform miRNA sequencing and disclose the differentially expressed miRNAs, among which miR-140-5p was significantly downregulated. Confocal microscopy was used to trace the exosomal miR-140-5p delivered from OPLL cells to human mesenchymal stem cells (hMSCs). In vitro, we verified that exosomal miR-140-5p inhibited the osteoblast differentiation of hMSCs by targeting IGF1R and suppressing the phosphorylation of the IRS1/PI3K/Akt/mTOR pathway. In vivo, we verified that exosomal miR-140-5p inhibited ectopic bone formation in mice as assessed by micro-CT and immunohistochemistry. Conclusions We found that exosomal miR-140-5p could inhibit the osteogenic differentiation of hMSCs by targeting IGF1R and regulating the mTOR pathway, prompting a further potential means of drug treatment and a possible target for molecular therapy of OPLL. Supplementary Information The online version contains supplementary material available at 10.1186/s12951-022-01655-8. | Exosomal miR-140-5p inhibits osteogenesis by targeting IGF1R and regulating the mTOR pathway in ossification of the posterior longitudinal ligament
Ossification of the posterior longitudinal ligament (OPLL) is a disabling disease whose pathogenesis is still unclear, and there are no effective cures or prevention methods. Exosomal miRNA plays an important role in the osteogenesis of ectopic bone. Therefore, we focused on the downregulation of miR-140-5p in OPLL cell-derived exosomes to explore the mechanism by which exosomal miR-140-5p inhibits osteogenesis in OPLL.
Exosomes were isolated by differential centrifugation and identified by transmission electron microscopy, nanoparticle tracking analysis, and exosomal markers. Exosomal RNA was extracted to perform miRNA sequencing and disclose the differentially expressed miRNAs, among which miR-140-5p was significantly downregulated. Confocal microscopy was used to trace the exosomal miR-140-5p delivered from OPLL cells to human mesenchymal stem cells (hMSCs). In vitro, we verified that exosomal miR-140-5p inhibited the osteoblast differentiation of hMSCs by targeting IGF1R and suppressing the phosphorylation of the IRS1/PI3K/Akt/mTOR pathway. In vivo, we verified that exosomal miR-140-5p inhibited ectopic bone formation in mice as assessed by micro-CT and immunohistochemistry.
We found that exosomal miR-140-5p could inhibit the osteogenic differentiation of hMSCs by targeting IGF1R and regulating the mTOR pathway, prompting a further potential means of drug treatment and a possible target for molecular therapy of OPLL.
The online version contains supplementary material available at 10.1186/s12951-022-01655-8.
Ossification of the posterior longitudinal ligament (OPLL) is a kind of degenerative disease in which heterotopic ossification in the posterior longitudinal ligament compresses the spinal cord and nerve roots, produces a series of clinical manifestations, and leads to dysfunction of the spinal cord. OPLL occurs mostly in the cervical spine (less often in the thoracic and lumbar spine) and can be accompanied by diffuse idiopathic skeletal hyperostosis [1–3]. OPLL has a high incidence in Japan and other East Asian countries; therefore, OPLL has also been called "Japanese disease" [4]. A study in Japan showed that the prevalence of OPLL in people over 30 years old is 1.9% to 4.3% [5]. The prevalence in other East Asian countries is similar. The average prevalence in Chinese individuals is 3.08% [6], but the prevalence in Koreans is lower, approximately 0.6% [7]. The prevalence of OPLL in the United States and Europe is only 0.01% to 1.7% [4]. The main symptoms of OPLL include limb movement and sensory disorders and sphincter function damage. The continued progression of the disease may lead to paraplegia, severely affecting patients’ quality of life and increasing family and socioeconomic burdens. However, the cause of OPLL is still unclear. To investigate the cause, mesenchymal stem cells were isolated from human spinal ligaments, including the posterior longitudinal ligament, and studied [8, 9]. OPLL has the characteristic of family inheritance, so most scholars believe that the combination of genetics, environment, and other factors together lead to osteogenic differentiation of human mesenchymal stem cells (hMSCs), resulting in the occurrence of OPLL, but it is still at the stage of speculation and theory [10–12]. The pathological mechanism of OPLL is probably related to changes in a variety of extracellular microenvironmental regulatory substances, which induce the osteogenic differentiation of hMSCs. Extracellular microvesicles (EVs) are effective microenvironmental regulatory components and potential biomarkers. The various vesicles secreted by cells can be distinguished by their diameters. Exosomes, with a diameter of 50 nm-150 nm, are considered to be active substances with strong biological effects [13, 14]. Exosomes are generally released through multivesicular exocytosis and participate in intercellular signal transduction. They can deliver a series of signal transduction proteins, mRNAs and microRNAs (miRNAs, miR) to exert biological effects, participating in the occurrence and development of many diseases, including atherosclerosis, coronary artery disease, hematological diseases, diabetes, and cancer [15–17]. Many miRNAs are abnormally expressed in (Additional file 4: Table S1) diseased tissues, and some miRNAs, as important biomarkers, are important indicators for the diagnosis and prognosis of diseases. Studies have found that many miRNAs are delivered from cell to cell in the form of vesicles. Due to the protection of vesicles, these miRNAs can avoid degradation and exert biological effects on target cells [18–20]. We found that miR-140-5p was significantly downregulated in exosomes derived from OPLL cells, suggesting that exosomal miR-140-5p may play an important role in OPLL. Here, we explored the mechanism by which exosomal miR-140-5p inhibits osteogenesis in OPLL.
To clarify the role of exosomes in the pathogenesis of OPLL, we hypothesized that the differentially expressed miRNAs in OPLL cell-derived exosomes played an important role in this process. Therefore, we collected tissue samples from patients with OPLL during surgical operations, cultured cells in vitro using the ligaments close to the ossified mass, and collected exosomes in the culture supernatant of OPLL cells. In addition, we collected the posterior longitudinal ligament (PLL) of patients with cervical spine trauma (non-OPLL) and obtained exosomes from the culture supernatant of PLL cells as a control group. There were no statistic differences between OPLL and non-OPLL patients’ characteristics (Additional file 1: Table S1). The typical double-layer membrane structure of exosomes was observed under transmission electron microscopy (Fig. 1a). Nanoparticle tracking analysis (NTA) showed that the collected extracellular vesicles in each group were approximately 108 nm to 120 nm in size (Fig. 1b). The exosomal markers CD63 and TSG101 (Fig. 1c) were detected. Subsequently, next-generation sequencing (NGS) technology was used to analyze the differentially expressed miRNAs in exosomes (Fig. 1d). We chose the top three downregulated miRNAs for follow-up studies. Next, we used real-time polymerase chain reaction (PCR) to verify NGS data, and the results showed that miR-140-5p was significantly downregulated in OPLL cell-derived exosomes (Fig. 1e).
To clarify whether miR-140-5p was secreted by OPLL cells and delivered to hMSCs, OPLL cells were transfected with lentivirus (LV-miR-140-5p) or its negative control to overexpress miR-140-5p (Fig. 2a), and then the cell culture supernatant was collected to isolate exosomes. Quantitative real-time PCR (qPCR) showed that miR-140-5p was enriched in exosomes (Fig. 2b). We labeled miR-140-5p-overexpressing exosomes with PKH67 and then added them to hMSC cultures. PKH67 fluorescence was positive in the cytoplasm of hMSCs (Fig. 2c), indicating that miR-140-5p was delivered by exosomes into hMSCs. Moreover, the expression level of miR-140, the precursor of miR-140-5p, was analyzed in hMSCs cultured with miR-140-5p-overexpressing exosomes or control exosomes (Fig. 2d), showing no significant difference. These results indicated that the upregulation of miR-140-5p in hMSCs was induced by exosomal delivery rather than endogenous miR-140 transcription.
The prerequisite for the formation of heterotopic ossification is the initiation of osteogenic differentiation of hMSCs. Therefore, to clarify the role of exosomal miR-140-5p in the pathogenesis of OPLL, we studied the effect of exosomal miR-140-5p on the osteogenic differentiation of hMSCs. We transfected OPLL cells (Fig. 3a) with lentivirus LV-miR-140-5p or LV-sponge to overexpress or downregulate miR-140-5p, respectively; the negative control lentivirus LV-miR-NC or LV-sponge-NC was also transfected. We collected cell-conditioned medium to isolate the following exosomes: miR-140-5p-exo, miR-NC-exo, sponge-exo, and sponge-NC-exo. qPCR verified that miR-140-5p was successfully overexpressed or downregulated, as expected, in exosomes (Fig. 3b). After culturing hMSC with the above exosomes for 24 h, qPCR showed that the expression of miR-140-5p in hMSCs was upregulated after miR-140-5p-exo treatment (Fig. 3c), indicating that miR-140-5p was delivered by exosomes into hMSCs. Subsequently, osteogenic induction medium was used to induce osteogenic differentiation of hMSCs. After 7 days, hMSCs were stained with alkaline phosphatase (ALP). After 14 days, hMSCs were stained with Alizarin red. The results showed that the positive rate of hMSC staining after miR-140-5p-exo treatment was significantly lower than that of the other groups (Fig. 3d, e). qPCR and Western blotting showed that the expression of osteogenesis-related genes, including osteocalcin (OCN), collagen type I alpha 1 (COLIA1), runt-related transcription factor 2 (RUNX2), and ALP, in hMSCs treated with miR-140-5p-exo was significantly suppressed (Fig. 3f, g, and Additional file 2: Figure S1a). Taken together, these results indicated that exosomal miR-140-5p inhibited the osteogenic differentiation of hMSCs. The lack of miR-140-5p in exosomes could promote the osteogenic differentiation of hMSCs and the expression of osteogenesis-related genes.
To clarify the mechanism by which miR-140-5p inhibits the osteogenic differentiation of hMSCs, we used the TargetScan database to predict target genes of miR-140-5p, and 434 transcripts were found. We also used the miRDB database to predict the targets, and there were 413 predicted targets for miR-140-5p in miRDB. To narrow down the possible candidate genes, we determined the overlap of the above two databases, which showed 191 genes in total (Fig. 4a). One of them was IGF1R, which belongs to the receptor tyrosine kinase family. The insulin receptor subfamily is located on the cell membrane and can be activated by insulin-like growth factors (IGF1 and IGF2) to cause phosphorylation of its own tyrosine kinase domain, initiates intracellular signal transduction, and regulates cell growth, differentiation, and various life activities, such as the growth, development, and aging of organisms [21, 22]. We tested the mRNA level of IGF1R in hMSCs after miR-140-5p-exo treatment, and the results showed that the mRNA level of IGF1R was significantly reduced (Fig. 4b). We used a luciferase reporter assay to explore whether IGF1R was the direct target of miR-140-5p. The target sequences of IGF1R (wt 3’UTR and mt 3’UTR) were cloned into a vector (Fig. 4c). Subsequently, hMSCs were transfected with the vector and cocultured with miR-140-5p-exo. The results showed that the fluorescence intensity in the wt 3'UTR carrier was significantly reduced, while the fluorescence intensity in the mt 3'UTR carrier was not significantly different from that in the control group (Fig. 4d). In addition, we also directly transfected miR-140-5p mimic or inhibitor into hMSCs, and the results showed that IGF1R mRNA and IGF1R protein were downregulated by miR-140-5p mimic (Fig. 4e, f, and Additional file 2: Figure S1b). Taken together, these results indicated that IGF1R was a direct target of miR-140-5p. Exosomal miR-140-5p interacted with the IGF1R 3'UTR and exerted a transcriptional inhibitory effect in hMSCs.
To clarify that the inhibitory effect of exosomal miR-140-5p on the osteogenic differentiation of hMSCs was mediated by the inhibition of IGF1R, we used IGF1R siRNA to directly knock down IGF1R in hMSCs (Fig. 5a). After osteogenic induction of the hMSCs, the positive rate of ALP and Alizarin Red staining of hMSCs in the knockdown group was significantly lower than that of the control group (Fig. 5b, c). The expression of OCN, COLIA1, RUNX2, and ALP was also suppressed (Fig. 5d, e, and Additional file 2: Figure S1c). We also performed an IGF1R rescue assay. LV-IGF1R was transfected into hMSCs to overexpress IGF1R (see Fig. 5f for transfection efficiency), and osteoinduction medium was used for differentiation. The results showed that after IGF1R was upregulated, ALP and Alizarin Red staining were strongly positive (Fig. 5g, h). The expression levels of OCN, COLIA1, RUNX2, and ALP also increased (Fig. 5i, j, and Additional file 2: Figure S1d). Moreover, after hMSCs were transfected with LV-IGF1R, we treated them with miR-140-5p-exo, and the effect of IGF1R on promoting osteogenesis was reversed (Fig. 5k, l, and Additional file 2: Figure S1e). To explore the molecular mechanism of exosomal miR-140-5p inhibiting osteogenesis through IGF1R, we mainly examined the mTOR pathway related to the osteogenic differentiation of hMSCs because it has been reported in the literature that the combination of IGF1 and IGF1R can promote the osteogenesis of MSCs by activating the mTOR pathway, resulting in improvement of osteoporosis in mice [23]. After we treated hMSCs with miR-140-5p-exo, IGF1 did not stimulate the phosphorylation of IGF1R, IRS1, PI3K, Akt, or mTOR, while after miR-NC-exo, sponge-exo, or sponge-NC-exo treatment, IGF1 stimulated the phosphorylation of IGF1R, IRS1, PI3K, Akt, and mTOR (Fig. 6a and Additional file 3: Figure S2a). To clarify the upstream and downstream relationships of IRS1, PI3K, Akt, and mTOR, we treated hMSCs with the IRS1 inhibitor NT157. The results showed that IGF1 could not cause the phosphorylation of PI3K, Akt, or mTOR. LY294002 (PI3K inhibitor) treatment reduced the phosphorylation of Akt and mTOR caused by IGF1 but did not affect the phosphorylation of IGF1R and IRS1. In addition, MHY1485 (mTOR inhibitor) treatment did not inhibit the phosphorylation of IRS1, PI3K, or Akt caused by IGF1 (Fig. 6b and Additional file 3: Figure S2b). Therefore, the upstream and downstream relationship of the above key molecules was IGF1R/IRS1/PI3K/Akt/mTOR. To further clarify the role of the IGF1R/IRS1/PI3K/Akt/mTOR axis, we respectively used NT157 (IRS1 inhibitor), LY294002 (PI3K inhibitor), MK2206 (Akt inhibitor), or MHY1485 (mTOR inhibitor) to interfere with the osteogenic induction of hMSCs. The results showed that NT157, LY294002, MK2206, or MHY1485 inhibited the osteogenic differentiation of hMSCs, respectively (Fig. 6c, d). Taken together, these data indicated that exosomal miR-140-5p inhibited the osteogenic differentiation of hMSCs by targeting IGF1R and regulating the mTOR pathway.
To further explore the function of miR-140-5p in vivo, we conducted ectopic bone formation experiments in nude mice. We cultured hMSCs with miR-140-5p-exo, miR-NC-exo, sponge-exo, or sponge-NC-exo for 48 h. After osteogenic induction, the above hMSCs and Bio-Oss Collagen were thoroughly mixed and cultured for 48 h. Finally, the mixture of scaffold and cells was implanted under the skin of the backs of nude mice (Fig. 7a). The animals were sacrificed 8 weeks later, and the bone volume/tissue volume (BV/TV) and bone mineral density (BMD) were calculated by micro computed tomography (micro-CT) scans (Fig. 7b). The results showed that after miR-140-5p-exo treatment, the BV/TV and BMD of the ectopic bone significantly decreased (Fig. 7c and d). Subsequently, we performed immunohistochemical detection of ectopic bone, and the results showed that the expression of OCN, COLIA1, RUNX2, and ALP in ectopic bone was inhibited after miR-140-5p-exo treatment, while the control groups were positive (Fig. 7e). In addition, the ectopic bone sections were also immunostained by IGF1R, and the results showed that the positive rate of IGF1R in the ectopic bone treated with miR-140-5p-exo was significantly lower than those of the control groups (Fig. 7e). The above results indicated that miR-140-5p had an inhibitory effect on bone formation in vivo. In summary, our study found that the expression of exosomal miR-140-5p in OPLL cells was significantly lower than that in PLL cells. MiR-140-5p was transferred into hMSCs by exosomes, targeted IGF1R, regulated IRS1/PI3K/Akt, and ultimately inhibited osteogenic differentiation through the mTOR pathway. Overexpression of miR-140-5p in OPLL cells inhibited the osteogenic differentiation of hMSCs, which is a potential new strategy for OPLL treatment. A schematic diagram of this mechanism is shown in Fig. 7f.
OPLL is a common disease requiring spinal surgery and is one of the causes of cervical spinal cord compression and paralysis. It is highly prevalent in East Asian populations [2, 3]. Due to insufficient understanding of the mechanism of spinal ligament ossification, there are currently no effective nonsurgical methods to prevent or cure OPLL. Patients often experience poor surgical outcomes, including ossified mass growth and recurrence or even aggravation of symptoms after surgery [24]. Therefore, the development of nonsurgical treatments for early intervention of OPLL is particularly important, especially exploring the key molecules and signaling pathways in the pathogenesis of OPLL for possible therapeutic targets. It was reported in the literature that the occurrence of OPLL may be related to multiple factors, such as heredity, environment, and lifestyle. However, how these factors induce ectopic osteogenesis through specific molecular mechanisms has not yet been elucidated [25, 26]. At present, it is generally accepted that multiple factors promote the osteogenic differentiation of MSCs in the posterior longitudinal ligaments, thereby initiating heterotopic ossification. Although studies have shown that MSCs exist in spinal ligaments (e.g., ligamentum flavum, posterior longitudinal ligament, and interspinous ligament) [8, 27], the specific osteogenic differentiation mechanism has not been elucidated in detail. The extracellular microenvironment plays an important role in cell proliferation, differentiation, metabolism, and other biological activities, and changes in microenvironmental components may lead to abnormal cell differentiation [28, 29]. Almost all somatic cells can secrete microvesicles into the extracellular microenvironment; extracellular microvesicles with a diameter of 50–150 nm are called exosomes [14]. Exosomes contain DNA, mRNA, noncoding RNA (including miRNA, lncRNA, circRNA, etc.), lipids and proteins. These biologically active substances participate in the regulation of a variety of physiological activities of recipient cells [14]. Exosomal miRNA participates extensively in the regulation of cell physiological activities after entering recipient cells and may play a role in promoting the osteogenic differentiation of MSCs. According to the literature, miR-193b-3p can regulate chondrogenesis and the metabolic activity of chondrocytes; exosomal miR-320c can promote the differentiation of bone marrow MSCs to chondrocytes; and osteoclast exosomal miR-214-3p can inhibit the bone formation process of osteoblasts [30–32]. These studies all suggested that exosomal miRNA plays an important regulatory role in the process of bone formation after entering the recipient cells. We isolated the exosomes secreted by OPLL cells, analyzed the differentially expressed miRNAs, and found that the expression of miR-140-5p decreased significantly. We constructed miR-140-5p-overexpressing exosomes and labeled the exosomes with PKH67. We proved that exosomes could deliver miR-140-5p into hMSCs. In addition, we demonstrated through a series of functional experiments that miR-140-5p could inhibit osteogenic differentiation after being delivered by exosomes into hMSCs. Therefore, we speculate that the lack of miR-140-5p relatively weakened the inhibition of osteogenic differentiation and finally initiated the osteogenic process. IGF-1 binds to IGF1R to activate it, resulting in the regulation of cell growth and differentiation and promoting the mineralization process coupled with osteoblasts, thereby promoting bone formation [33, 34]. The activation of IGF1R can activate autophagy, which is a joint process that stimulates the differentiation of early osteoblasts [35]. Goto et al. [36] found that IGF-1 was highly expressed in OPLL tissue sections and induced osteogenic differentiation. Therefore, IGF-1 might be involved in the ossification process of OPLL. Similar to these studies, we found that IGF1R was one of the candidate targets when predicting miR-140-5p target genes. Subsequently, we used a luciferase reporter assay to demonstrate the interaction between miR-140-5p and IGF1R. In addition, we used LV-IGF1R or siRNA to treat hMSCs to overexpress or downregulate IGF1R in hMSCs, respectively. It was found that the osteogenic differentiation of hMSCs was promoted or inhibited as expected, proving the role of IGF1R in regulating osteogenesis. We further clarified that IGF1R promoted the phosphorylation of IRS1/PI3K/Akt and activated the mTOR pathway to promote osteogenesis. In one study, Xian et al. [23] found that IGF-1 improved osteoporosis through the mTOR pathway, which also supported our results. In our study, overexpression of miR-140-5p exosomes reversed the promotion in osteogenesis caused by the upregulation of IGF1R and inhibited the osteogenic differentiation of hMSCs, suggesting that drugs targeting IGF1R on the surface of MSCs in the posterior longitudinal ligament (such as exosomes delivering miR-140-5p) may serve as a therapy for OPLL in the future. There are some limitations in this study. OPLL is considered to be a polygenic genetic disease, and multiple factors costimulate the expression of susceptible genes, which leads to the initiation of heterotopic ossification. However, the content of the extracellular microenvironment is extremely rich, and the role of extracellular vesicles represented by exosomes in the pathogenesis of OPLL may have limitations. In addition, exosomes carry a large number of biologically active molecules, including miRNA, mRNA, and proteins. Therefore, the role of exosomal miR-140-5p in the extracellular microenvironment needs to be further investigated. The role of exosomal miR-140-5p in vivo was examined in nude mice with a Bio-Oss Collagen scaffold. The reason is that there is no reliable animal model of OPLL. In addition to the ossification of the spinal ligament, the tip-toe walking mice commonly used by scholars show ossification and mineralization of multiple tissues and organs throughout the body, including the thoracic aorta, carotid artery, heart, spleen, lung, eye, kidney, and liver [37, 38]. Thus, this mouse model cannot completely simulate the pathological process of OPLL in the human body because OPLL is an isolated disease, usually not accompanied by ossification of other tissues and organs. Taken together, this study demonstrated that exosomal miR-140-5p inhibited the osteogenic differentiation of hMSCs by targeting IGF1R and regulating the mTOR pathway and played an important role in the pathogenesis of OPLL. This research focused on biologically active molecules in the extracellular microenvironment of OPLL for the first time and clarified the role and mechanism of hMSCs in the formation of OPLL, which could provide a new theoretical basis for the future development of targeted drugs for the treatment of OPLL.
Our data indicated that exosomal miR-140-5p could inhibit the osteogenic differentiation of hMSCs by targeting IGF1R and regulating the mTOR pathway. The findings in this study indicated that exosomes, as a drug delivery system, might serve as a potential means of drug treatment for OPLL treatment in the future. IGF1B might also be a target for future molecular therapy of OPLL.
A total of 20 OPLL patient samples and 18 cervical spine trauma patients' PLL samples were collected. The characteristics of patients were displayed in Additional file 1: Table S1. All experiments in this study were approved by the ethics committee of Shanghai Changzheng Hospital. The patients signed an informed consent form before the operation. The OPLL samples came from patients who had received anterior cervical corpectomy and fusion (ACCF). The inclusion criteria were patients [1] who were diagnosed with cervical OPLL or cervical spine trauma through X-ray, MRI and CT; [2] who received ACCF. The exclusion criteria were patients with [1] a history of infection, tumor, osteoporosis or other serious neurological diseases; [3] a history of substance abuse. After the vertebral body was removed during the operation, the ossified posterior longitudinal ligament of the posterior wall of the vertebral body was carefully separated, washed with normal saline, and stored aseptically in complete cell culture medium, which was prepared for primary OPLL cell culture. The normal posterior longitudinal ligament was derived from patients with cervical spine trauma who received ACCF. The posterior longitudinal ligament behind the posterior wall of the vertebral body was also separated for primary PLL cell culture. Specifically, tissues from patients with OPLL or PLL were cut into pieces with a diameter of approximately 1 mm and placed in a petri dish with complete cell culture medium (90% high-glucose Dulbecco’s modified Eagle’s medium (DMEM, Gibco, USA), 10% fetal bovine serum (FBS, Gibco, USA), and 1% penicillin/streptomycin (Gibco, USA)). The petri dishes were placed into a 37 °C constant temperature incubator in a humidified atmosphere containing 5% CO2. Half of the medium was changed every 3 days, and media was completely changed every 7 days. After 10–14 days, a mass of fibroblast-like cells migrated out of the tissue and grew adherently, which can be observed microscopically. When the cell confluence reached 80%, the cells were passaged at a ratio of 1:3. According to literature, cells derived from OPLL mainly consist of fibroblasts, osteoblasts, and chondrocytes [39, 40]. In this article, we define these cells as OPLL cells. The hMSCs used in this experiment were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The complete cell culture medium consisted of 90% low-glucose DMEM (Gibco, USA), 10% FBS (Gibco, USA), and 1% penicillin/streptomycin (Gibco, USA). The culture environment was the same as that of OPLL cells.
The exosomes were separated from the cell culture medium with gradient centrifugation. The OPLL cell or PLL cell supernatant was centrifuged at 300 × g for 15 min to remove floating cells and then centrifuged again at 820 × g for 15 min and 10,000 × g for 5 min, and cell debris was removed through a 0.8 μm filter. Finally, the samples were centrifuged at 100,000 × g for 2 h (Beckman L-90 K, USA) to obtain an exosome pellet. The exosomal pellet was resuspended in PBS and ultracentrifuged again for further experiments. All centrifugation was performed at 4 °C. The size distribution and concentration of exosomes were examined by nanoparticle tracking analysis with ZetaView PMX 110 (Particle Metrix, Germany) and the corresponding software ZetaView 8.04.02. Exosomes were diluted in PBS to measure the particle size and concentration. NTA measurements were recorded and analyzed at 11 positions. The ZetaView system was calibrated using 110 nm polystyrene particles. All procedures were performed at room temperature. The exosome samples were prepared and observed with a transmission electron microscope (H-7650, Hitachi, Japan) as described previously [41]. The voltage was 80 kV.
Total exosomal RNA of OPLL or PLL cells was extracted using a Total Exosome RNA Isolation Kit (Invitrogen, USA) according to the manufacturer’s instructions. MiRNA library construction and miRNA sequencing were performed as described previously [42] to identify the differentially expressed miRNAs derived from OPLL or PLL cell exosomes. Next, target genes of miR-140-5p were predicted by TargetScan (http://www.targetscan.org) and miRDB (http://mirdb.org). The intersection of these two databases was created by Venny 2.1.0 (https://bioinfogp.cnb.csic.es/tools/venny).
Exosomes were collected and labeled with the green lipophilic fluorescent dye PKH67 (Sigma, USA) and then cocultured with hMSCs for 1 h. After that, hMSCs were washed with PBS and fixed with 4% paraformaldehyde. The cell nuclei were stained with 4′6-diamidinophenylidole (DAPI, Beyotime, China). The samples were observed with confocal microscopy (FV10i, Olympus, Japan).
With the help of restriction endonuclease site primers, we synthesized a human wild-type IGF1R 3'UTR fragment containing the miR-140-5p conserved binding site by PCR and cloned it into the pMIR reporter vector. This is called the wild-type (wt) IGF1R 3'UTR. By mutating the miR-140-5p binding site of the wt IGF1R 3'UTR, a mutant (mt) IGF1R 3'UTR was synthesized and cloned into the vector. Human 293 T cells were transfected with the above two vectors and cultured with miR-140-5p-exo or control conditions. A dual luciferase reporter gene detection system (Promega, USA) was used to detect luciferase activity.
Agomirs, antagomirs, and siRNAs were synthesized by Sangon Biotech (Shanghai, China). The IGF1R vector for cDNA delivery was synthesized by Sangon Biotech (Shanghai, China). Lentivirus production, purification, and titration were performed as described previously [43]. Transfection of miR-140-5p mimic or inhibitor into cells was performed by using Lipofectamine 2000 (Invitrogen, USA) according to the manufacturer’s instructions. Transfection of miR-140-5p mimic or inhibitor into exosomes was performed by using Exo-Fect Exosome Transfection Kit (SBI, CA) according to the manufacturer’s instructions.
After miR-140-5p-exo, miR-NC-exo, sponge-exo, and sponge-NC-exo were collected, they were added to hMSC culture for 48 h and then replaced by osteogenic induction culture for 14 days. The cells were resuspended, thoroughly mixed with Bio-Oss collagen (Geistlich, GEWO GmbH, Germany) and cultured for 48 h. Finally, the mixture of scaffold and cells was implanted into the backs of nude mice (4-week-old BALB/c homozygous nude) subcutaneously. After antiseptic preparation of the operative site, a 2 mm-long incision was made. The incision was located on the back of the mouse to avoid possible bites from itself. The incision was normally 0.5 mm deep, which was enough to reach the surface of the muscle. The space superior to the muscle was bluntly dissected. Then the prepared scaffold was implanted into the space and the skin was sutured. After 8 weeks, the specimen was removed and fixed in 4% paraformaldehyde for further immunohistochemical analysis. Images of animal models after 8 weeks were shown in Additional file 4: Figure S3.
The scaffold samples were decalcified in 10% ethylene diamine tetraacetic acid (EDTA) for 1 month. Then, they were dehydrated and embedded in paraffin. The paraffin tissue block was cut into 5 μm slices for hematoxylin and eosin (H&E) staining. The sections were blocked with 3% BSA for 30 min and then incubated with primary antibodies (Abcam, USA), including anti-OCN, anti-COLIA1, anti-RUNX2 anti-ALP, and anti-TGF1R. Then, samples were placed at 95 °C in citrate buffer to retrieve antigens for 10 min. Subsequent sections were incubated again with primary antibodies at 4 °C overnight. Sections were observed with an Olympus microscope equipped with an Olympus DP70 camera (Olympus, Japan).
The scaffold samples were harvested after intervention and fixed in 4% paraformaldehyde. Subsequently, the samples were scanned by micro-CT (Quantum FX microCT, PerkinElmer, USA); RigakuTM software (Rigaku, Japan) and RadiAnt DICOM ViewerTM (Medixant, Poland) were used for 3D reconstruction and image processing.
Osteogenesis induction medium (Cyagen, China) contained 10% FBS, 1% penicillin–streptomycin, 10 mM L-glutamine, 50 μM L-ascorbic acid, 10 mM β-glycerophosphate, and 100 nM dexamethasone. When the degree of hMSC fusion reached 60%-70%, the culture medium was replaced with osteogenic medium for osteoinduction. The medium was changed every 3 days. After 14 days, the induction medium was removed. Then, the cells were washed with PBS and fixed with 4% paraformaldehyde. Cells were stained by using an ALP staining kit (Beyotime, China) according to the manufacturer’s instructions. Similarly, cells were stained with Alizarin red S (Cyagen, China) 14 days after induction.
Briefly, total RNA from cultured cells was isolated by TRIzol (Invitrogen, USA) according to the manufacturer's instructions. Then, the extracted RNA was reverse transcribed into cDNA by a ReverTra Ace® qPCR RT Kit (Toyobo, Japan). qPCR was performed with real-time PCR (ABI 7500, Applied Biosystems, USA), and the expression levels of several osteogenesis-related genes, including OCN, COLIA1, RUNX2 and ALP, were calculated by the 2−ΔΔCt method. The sequences of qPCR primers were presented in Additional file 5: Table S2.
First, cells were lysed with high-efficiency RIPA lysis buffer (Solarbio, China), and the protein content was determined by a BCA protein assay kit (Thermo Fisher Scientific, USA). Second, protein samples were separated by Bis–Tris gel (Invitrogen, USA), transferred to a nitrocellulose filter membrane (Bio-Rad, USA), blocked with 5% w/v skimmed milk in TBST buffer (1X Tris-buffered saline and 0.1% Tween 20) for 1 h at room temperature, and incubated with primary antibodies overnight at 4 °C. Third, the membranes were washed and incubated with secondary antibodies for 2 h at room temperature. Finally, the Odyssey imaging system (Li-Cor, Lincoln, USA) was used to determine the fluorescent signals. Anti-GAPDH or anti-β-actin was used as an endogenous control.
All data analysis was performed using SPSS 21.0 software (Chicago, IL, USA). All data are expressed as the mean ± SD. Student’s t test was selected for the comparison of two independent variables. One-way ANOVA was selected for the comparison of multiple independent variables. A P value < 0.05 was considered to be statistically significant.
Additional file 1: Table S1. Characteristics of patients.Additional file 2. Figure S1. Quantitative analysis of the osteogenesis-related proteins.Additional file 3. Figure S2. Quantitative analysis of the phosphorylation of the proteins.Additional file 4. Figure S3. Representative photos of the animal model.Additional file 5: Table S2. The sequences of qPCR primers. | true | true | true |
|
PMC9571494 | Gerrit A. Martens,Cornelia Geßner,Carina Osterhof,Thomas Hankeln,Thorsten Burmester | Transcriptomes of Clusterin- and S100B-transfected neuronal cells elucidate protective mechanisms against hypoxia and oxidative stress in the hooded seal (Cystophora cristata) brain | 15-10-2022 | Clusterin,S100B,Hypoxia,Oxidative stress,Transcriptome,Neurons,Brain,Hooded seal,Marine mammals | Background The hooded seal (Cystophora cristata) exhibits impressive diving skills and can tolerate extended durations of asphyxia, hypoxia and oxidative stress, without suffering from irreversible neuronal damage. Thus, when exposed to hypoxia in vitro, neurons of fresh cortical and hippocampal tissue from hooded seals maintained their membrane potential 4–5 times longer than neurons of mice. We aimed to identify the molecular mechanisms underlying the intrinsic neuronal hypoxia tolerance. Previous comparative transcriptomics of the visual cortex have revealed that S100B and clusterin (apolipoprotein J), two stress proteins that are involved in neurological disorders characterized by hypoxic conditions, have a remarkably high expression in hooded seals compared to ferrets. When overexpressed in murine neuronal cells (HN33), S100B and clusterin had neuroprotective effects when cells were exposed to hypoxia. However, their specific roles in hypoxia have remained largely unknown. Methods In order to shed light on potential molecular pathways or interaction partners, we exposed HN33 cells transfected with either S100B, soluble clusterin (sCLU) or nuclear clusterin (nCLU) to normoxia, hypoxia and oxidative stress for 24 h. We then determined cell viability and compared the transcriptomes of transfected cells to control cells. Potential pathways and upstream regulators were identified via Gene Ontology (GO) and Ingenuity Pathway Analysis (IPA). Results HN33 cells transfected with sCLU and S100B demonstrated improved glycolytic capacity and reduced aerobic respiration at normoxic conditions. Additionally, sCLU appeared to enhance pathways for cellular homeostasis to counteract stress-induced aggregation of proteins. S100B-transfected cells sustained lowered energy-intensive synaptic signaling. In response to hypoxia, hypoxia-inducible factor (HIF) pathways were considerably elevated in nCLU- and sCLU-transfected cells. In a previous study, S100B and sCLU decreased the amount of reactive oxygen species and lipid peroxidation in HN33 cells in response to oxidative stress, but in the present study, these functional effects were not mirrored in gene expression changes. Conclusions sCLU and S100B overexpression increased neuronal survival by decreasing aerobic metabolism and synaptic signaling in advance to hypoxia and oxidative stress conditions, possibly to reduce energy expenditure and the build-up of deleterious reactive oxygen species (ROS). Thus, a high expression of CLU isoforms and S100B is likely beneficial during hypoxic conditions. Supplementary Information The online version contains supplementary material available at 10.1186/s12868-022-00744-6. | Transcriptomes of Clusterin- and S100B-transfected neuronal cells elucidate protective mechanisms against hypoxia and oxidative stress in the hooded seal (Cystophora cristata) brain
The hooded seal (Cystophora cristata) exhibits impressive diving skills and can tolerate extended durations of asphyxia, hypoxia and oxidative stress, without suffering from irreversible neuronal damage. Thus, when exposed to hypoxia in vitro, neurons of fresh cortical and hippocampal tissue from hooded seals maintained their membrane potential 4–5 times longer than neurons of mice. We aimed to identify the molecular mechanisms underlying the intrinsic neuronal hypoxia tolerance. Previous comparative transcriptomics of the visual cortex have revealed that S100B and clusterin (apolipoprotein J), two stress proteins that are involved in neurological disorders characterized by hypoxic conditions, have a remarkably high expression in hooded seals compared to ferrets. When overexpressed in murine neuronal cells (HN33), S100B and clusterin had neuroprotective effects when cells were exposed to hypoxia. However, their specific roles in hypoxia have remained largely unknown.
In order to shed light on potential molecular pathways or interaction partners, we exposed HN33 cells transfected with either S100B, soluble clusterin (sCLU) or nuclear clusterin (nCLU) to normoxia, hypoxia and oxidative stress for 24 h. We then determined cell viability and compared the transcriptomes of transfected cells to control cells. Potential pathways and upstream regulators were identified via Gene Ontology (GO) and Ingenuity Pathway Analysis (IPA).
HN33 cells transfected with sCLU and S100B demonstrated improved glycolytic capacity and reduced aerobic respiration at normoxic conditions. Additionally, sCLU appeared to enhance pathways for cellular homeostasis to counteract stress-induced aggregation of proteins. S100B-transfected cells sustained lowered energy-intensive synaptic signaling. In response to hypoxia, hypoxia-inducible factor (HIF) pathways were considerably elevated in nCLU- and sCLU-transfected cells. In a previous study, S100B and sCLU decreased the amount of reactive oxygen species and lipid peroxidation in HN33 cells in response to oxidative stress, but in the present study, these functional effects were not mirrored in gene expression changes.
sCLU and S100B overexpression increased neuronal survival by decreasing aerobic metabolism and synaptic signaling in advance to hypoxia and oxidative stress conditions, possibly to reduce energy expenditure and the build-up of deleterious reactive oxygen species (ROS). Thus, a high expression of CLU isoforms and S100B is likely beneficial during hypoxic conditions.
The online version contains supplementary material available at 10.1186/s12868-022-00744-6.
The hooded seal (Cystophora cristata) is an excellent breath-hold diver, performing dives for up to 1 h, while diving over 1 km deep [24, 98]. Physiological adaptations, such as increased oxygen stores (hemoglobin, myoglobin), a decreased heart rate (bradycardia) and a redirection of blood flow to vital organs (selective peripheral vasoconstriction) have evolved to facilitate this diving lifestyle [6, 7, 23, 81, 82, 94]. However, during repetitive diving bouts oxygen partial pressure may drop dramatically, as shown in other deep-diving seals [69, 84], and would lead to neuronal damage in humans [66]. Neurons from the hooded seal have an intrinsic hypoxia tolerance that cannot be explained by these physiological adaptations. Isolated hooded seal brain slices maintained their membrane potential during hypoxic treatment, while those of mice lost their functional integrity [25]. The molecular basis of this intrinsic hypoxia tolerance is not well understood. In a comparative transcriptomics analysis, the calcium-binding protein S100B and the molecular chaperone clusterin (CLU) emerged as highly expressed in the hooded seal visual cortex, when compared to the ferret (Mustela putorius furo). More precisely, S100B expression was 38-fold higher in the hooded seal than in the ferret [20]. CLU exhibited the highest mRNA levels in the hooded seal cortex, with a fourfold increase compared to the ferret [20]. The remarkably enhanced transcription of S100B was confirmed in laser-excised hooded seal neurons, in which S100B was 82-fold more highly expressed than in neurons of mice (Mus musculus) [30]. The observed overexpression suggests that both genes might contribute to the hypoxia tolerance of the hooded seal brain [20, 30]. Both, S100B and CLU, are associated with many neurological disorders such as Alzheimer’s disease and Parkinson’s disease in humans that involve hypoxia and oxidative stress [27, 71]. However, their role and molecular mechanisms in these conditions are controversial and ambiguous. The S100 proteins are appreciably conserved among different species [21, 31], which may indicate crucially conserved biological roles. S100B is a calcium-binding protein that is involved in a broad range of Ca2+-dependent pathways [72]. Human and rodent studies demonstrated S100B’s dual role, acting as an intracellular regulator on the one hand and as an extracellular signal substance on the other hand [17]. Intracellular S100B is involved in many processes such as proliferation, differentiation and survival [72]. For instance, in melanoma cells S100B leads to improved tumor survival, preventing p53-mediated apoptosis [64, 65]. Although mainly present in astrocytes [71], S100B was also found to be located in neurons [88], in which it reduced apoptosis and nerve growth factor (NGF)-induced differentiation [2]. S100B may be secreted from astrocytes and neurons in conditions of metabolic stress and other external stimuli [18, 29], where its role depends on its concentration [72]. When released into the extracellular microenvironment, S100B acts as a damage-associated molecular pattern (DAMP) protein through its interaction with the receptor for advanced glycation end products (RAGE) [13, 96]. RAGE is a multi-ligand receptor of the immunoglobulin superfamily which is mainly expressed by neurons and microglia and which mediates inflammatory responses by activating multiple signaling pathways [1, 44]. At nanomolar concentrations, S100B demonstrates neurotrophic effects, promoting neurite extension and neuron survival, modulating long-term potentiation and counteracting neurotoxicants like reactive oxygen species (ROS) [71]. Neuron survival may be facilitated by different RAGE-dependent pathways [17, 42, 59]. By contrast, persistent activation of RAGE by micromolar concentrations of S100B produces increased amounts of ROS, leading to lipid peroxidation and consequently induction of apoptosis [99]. However, studies report varying results at what concentration S100B exerts neurotrophic or neurotoxic effects [17]. In neuronal disorders such as acute brain injury and neurodegenerative diseases, S100B has been found at high levels serving as a biomarker of disease progression [71]. Nevertheless, in proteinopathies like Alzheimer’s and Parkinson’s disease, S100B might also be involved in clearance of detrimental protein aggregates [72]. CLU, a multifunctional glycoprotein, is a constitutively secreted chaperone in its predominant form (soluble CLU, sCLU), but truncated forms localized to the nucleus (nuclear CLU, nCLU) have also been found [38, 45]. The different isoforms of CLU target distinct cellular or subcellular localizations in the rat and human brain, where they demonstrate different functions [38]. The sCLU isoform is translated as a pre-protein of 36–39 kDa, which contains an N-terminal endoplasmic reticulum (ER)-signaling peptide and two nuclear localization sequences. After removal of the signaling peptide, the pre-protein is phosphorylated and glycosylated in the ER and Golgi body. Cleavage of the intermediate glycoprotein results in two subunits linked by disulfide-bonds. The resulting mature antiparallel, heterodimeric glycoprotein (70–75 kDa), commonly referred to as sCLU, is then secreted [8]. In contrast, nCLU is a truncated isoform of 45–50 kDa, lacking the ER signaling peptide and is primarily detected in the cytosol and nucleus [62, 83]. Other CLU forms targeted to the mitochondria have also been described, where they may function as antiapoptotic proteins during stress conditions or facilitate mitochondrial respiration [38, 89, 103]. Although some studies question the relevance and existence of CLU isoforms [93], it has been demonstrated that sCLU and nCLU exhibit distinct functions and that they regulate certain cellular processes in opposite manners [38]. CLU function and expression is regulated by a wide variety of signals including growth and transcription factors, as well as several stress conditions [27]. For instance, during oxidative stress conditions, intracellular CLU promotes cardiomyocyte survival [47]. Due to its stress-increased expression and extracellular chaperone activity CLU has been compared to heat shock proteins [27]. Indeed, CLU might mediate neuroprotection by preventing stress-induced precipitation and aggregation of proteins, by mediating clearance of extracellular misfolded proteins and aggregates, and by promoting their cellular uptake [100]. Additionally, cytosolic CLU might have an important role in intracellular protein homeostasis (proteostasis), by transporting misfolded proteins to the proteasome and/or autophagy for degradation [34, 68, 102]. Similar to S100B, in proteinopathies like Alzheimer’s and Parkinson’s disease, CLU has been found to associate with protein aggregates and might be involved in their clearance [27, 61]. However, in advanced disease stages CLU has been found to promote neurotoxicity [101]. Its concentration in peripheral blood was identified as a potential biomarker for neurodegenerative diseases, like Alzheimer’s disease [4]. Still, CLU’s downstream pathways and molecular mechanisms are not well characterized. Different cell-types, varying levels of model complexity, and conditions that represent distinct physiological situations might lead to differing conclusions [27]. Furthermore, lack of discrimination between different CLU isoforms as well as structural differences between these proteins (e.g., glycosylation levels) may be key in explaining the variability in CLU effects in apoptosis and cell death pathways. Remarkably, the protective effects of S100B and CLU could be demonstrated in cell culture experiments [31]. Murine neuronal cell cultures (HN33) that were transfected with S100B, sCLU and nCLU demonstrated elevated viability when exposed to hypoxia [31]. Furthermore, overexpression of S100B, sCLU and, to a lesser degree, nCLU, led to a reduction of ROS and lipid peroxidation [31]. While these findings confirmed neuroprotective roles of S100B and CLU in hypoxia and oxidative stress, their molecular mechanisms remained obscure. In order to improve our understanding of the roles of these proteins in cell metabolism, we analysed the transcriptomes of neuronal cell cultures transfected with S100B, nCLU and sCLU and then exposed to normoxia, hypoxia and oxidative stress, with the aim to identify possible molecular targets and pathways of S100B and CLU at different oxygen regimes.
Transfected HN33 cell lines demonstrated stable mRNA overexpression of the S100B, sCLU and nCLU transgenes compared to the endogenous expression levels measured in mock-transfected cells. Overexpression in qPCR experiments ranged from 712-fold for S100B to 9230–9723-fold for nCLU and sCLU, respectively (Additional file 1: Fig. S1). Expression of transfected genes was also observed in the RNA-Seq transcriptome data, when mapped to the mouse reference genome with added transgenic sequences (Additional file 1: Fig. S2).
Cell lines exhibited significant differences in ATP amounts, as shown by the CTG cell viability assay after exposure to normoxia, hypoxia and oxidative stress (275 µM H2O2) for 24 h (Fig. 1). Cells transfected with sCLU and S100B were more viable at normoxic conditions than the mock and nCLU cell lines (sCLU: false discovery rate (pFDR) < 0.001; S100B: pFDR < 0.001). At hypoxic conditions, only sCLU demonstrated significantly elevated ATP levels compared to mock cells (pFDR < 0.01). Viability of nCLU and S100B cell lines was with 85% insignificantly higher than that of mock cells with 84%. When exposed to oxidative stress, ATP levels of the sCLU and S100B cell lines were significantly higher than in the mock cell line (sCLU: pFDR < 0.01; S100B: pFDR < 0.01), while the nCLU cell line exhibited a similar ATP concentration as did the mock cell line.
Three replicates per cell line and condition were sequenced, with the exception of the mock cell line at oxidative stress conditions with two replicates. An average of 51 million RNA-Seq reads per sample were generated, with a minimum of 27 million and a maximum of 82 million reads per sample. Around 75% of all reads mapped to the GRCm39 mouse reference genome (Additional file 1: Table S1) across all cell lines. Cells were compared to the mock cell line at the respective stress condition, to identify differentially expressed genes (DEGs). Principal component analysis of DEGs revealed that cells clustered according to stress conditions (Fig. 2). Hypoxia and oxidative stress elicited very distinct responses that were well-distinguishable. Still, within stress conditions, differences in gene expression for transfected cell lines could be determined (Fig. 3). At normoxia, the nCLU cell line demonstrated the least variable DEGs (p < 0.05, TPM > 1), with only 69 DEGs, while the sCLU and S100B cell lines exhibited 676 and 1484 DEGs, respectively (Fig. 3). At hypoxia, numbers were more similar across cell lines, with 1018, 2316 and 2003 DEGs for nCLU, sCLU and S100B cell lines, respectively. Proportions shifted, when at oxidative stress the nCLU cell line displayed 5556 DEGs whereas sCLU and S100B cells only had 143 and 156 DEGs. The sCLU and S100B cell lines shared a substantial amount of DEGs at normoxia (142 up- and 201 downregulated genes), while the nCLU cell line only contributed few genes to the shared DEG pool of all cell lines (9 up- and 8 downregulated genes). The overlap of DEGs between cell lines was larger at hypoxia (192 up- and 135 downregulated genes), but was almost non-existent at oxidative stress (21 upregulated genes).
In normoxia, the nCLU cell line demonstrated the least differentially expressed genes (DEGs) in comparison to the mock cell line. The only enriched pathway was the inhibited neuropathic pain signaling in dorsal horn neurons [z-score (z) = − 1, − log(p) = 2.9] (Fig. 4) with an upregulated potassium channel (KCNN3) within this pathway [binary logarithm of fold change (log2FC) = 1.19]. Some developmental genes also showed high expression, demonstrated by activation of upstream regulator eomesodermin (EOMES, z = 2.0, p < 0.0001). Homozygous silencing of EOMES leads to neurodevelopmental disorders such as microcephaly in the human brain [3] and full knockout leads to embryonic lethality in mice [90]. One of EOMES’ targets, semaphorin receptor plexin A4 (PLXNA4, (log2FC = 1.01), as well as another gene involved in axon guidance, ephrin type-A receptor 8 (EPHA8) exhibited high expression levels (log2FC = 2.16). Axon guidance may be an important process for development of the nervous system [46]. In contrast, the expression of opioid receptor delta 1 (OPRD1) was decreased in nCLU cells after normoxia (log2FC = − 1.49). In mouse astrocyte cell culture, activation of the delta opioid receptor increased expression of excitatory amino acid transporters, suggesting a role in glutamate uptake and prevention of glutamate-induced neuroexcitotoxicity [63]. Similar neuroprotection by delta opioid receptor activation has been found in mouse neuronal cell culture, which attenuated neuronal injury in normoxic and hypoxic conditions [37, 104]. Downregulation of OPRD1 may therefore indicate reduced capacity of nCLU-transfected cells to respond to stress conditions.
When exposed to hypoxia, the nCLU-transfected neuronal cells demonstrated elevated hypoxia response pathways in comparison to the mock cells (Fig. 5). Nevertheless, the mock cells displayed a hypoxia response as well. When comparing its gene expression in hypoxic conditions to normoxic conditions, cellular response to hypoxia was among top 10 enriched pathways (FE = 15.89, pFDR < 0.05). Even when only looking at highly differentially expressed genes (log2fc > 1 or log2fc < − 1) cellular response to hypoxia was among top 10 enriched pathways (FE > 100, pFDR < 0.05). However, hypoxia response of the nCLU-transfected cells appeared to be even stronger in comparison to mock cells. In hypoxia, most eukaryotic cells can shift their primary metabolic strategy from predominantly mitochondrial respiration towards increased glycolysis [51]. However, it is not clear if neurons are able to undergo such a shift in metabolism [15]. GO term glycolytic process [fold enrichment (FE) = 9.8, false discovery rate (pFDR) < 0.01], as well as IPA pathways glycolysis I (z = 2.6, − log(p) = 4.1) and gluconeogenesis (z = 1.6, − log(p) = 3.1) were enhanced in nCLU-transfected cells. Additionally, GO term glycogen biosynthetic process (FE = 14.0, pFDR < 0.05) and IPA pathway glycogen biosynthesis (− log(p) = 3.0) were increased in the nCLU-transfected cells, which is a common response to hypoxia [80]. Thus, enzymes for glycogen metabolism which mainly promote glycogen accumulation (e.g., glucan phosphatase EPM2A, log2FC = 1.03) [86] were upregulated in the nCLU-transfected cell lines. Glycogen has been demonstrated to protect cerebellar and cortical mouse neurons from hypoxic stress-induced cell death in cell culture [91] and glycogen storage is increased in the seal brain [14, 50], which illustrates its significance in dealing with hypoxia. In a transcriptome analysis, glycogenolysis-associated genes were, thus, found to be upregulated in hooded seal neurons compared to neurons of mice [30]. In general, energy metabolism through glycogen biosynthesis and glycolysis were enhanced in the nCLU-transfected cells in hypoxic conditions. Response to hypoxia is substantially mediated by the hypoxia-inducible factor (HIF1), which is a master regulator that activates the transcription of many genes involved in energy metabolism, apoptosis and oxygen delivery. In the present study, nCLU transfected cell lines increased HIF1A signaling (z = 2.7, − log(p) = 3.0), while HIF1A was also found to be an upstream regulator (z = 1.4, p < 0.01). Additionally, in GO analyses, cellular response to hypoxia was enhanced in the nCLU cell line (FE = 21.0, pFDR < 0.01). However, as in normoxic conditions, OPRD1 demonstrated low expression in the nCLU cell line (log2FC = − 1.32), which might be disadvantageous in dealing with hypoxic conditions.
Oxidative stress is an inevitable by-product of oxidative metabolism and reflects a state of imbalance between reactive oxygen species (ROS) and substances that are involved in their detoxification, which may cause damage to proteins, lipids and DNA. Diving mammals may counteract oxidative stress through elevated antioxidant levels in its brain [20, 30]. At oxidative stress conditions, antioxidants, such as the glutathione S-transferase alpha 4 (GSTA4, log2FC = 1.80), thioredoxin 1 (TXN1, log2FC = 1.40), peroxiredoxin 3 (PRDX3, log2FC = 1.40), selenoprotein F (SELENOF, log2FC = 1.27) and the putative glutathione peroxidase 8 (GPX8, log2FC = 1.00) were increased in the nCLU cell line, which may partly be ascribed to activation of nuclear factor, erythroid 2-like 2 (NFE2l2, also abbreviated as NRF2, z = 2.2, p < 0.05) as upstream regulator. The transcription factor NRF2 plays a pivotal role controlling the expression of antioxidant genes that exert neuroprotective functions [92]. Nevertheless, the nCLU cell line exhibited a profound response to oxidative stress (FE = 2.8, pFDR < 0.05), characterized by mitochondrial dysfunction (− log(p) = 8.1), enhanced oxidative phosphorylation (z = 6.6, − log(p) = 4.3) and TCA cycle (z = 4.0, − log(p) = 5.9) (Fig. 6). Mitochondrial complexes such as cytochrome c oxidase subunits COX7A2 (log2FC = 1.91), COX4I1 (log2FC = 1.28), COX5B (log2FC = 1.10), COX7C (log2FC = 1.09), NADH:ubiquinone oxidoreductase subunits NDUFB1 (log2FC = 1.37), NDUFC2 (log2FC = 1.36), NDUFB9 (log2FC = 1.34), NDUFV2 (log2FC = 1.20), NDUFA7 (log2FC = 1.12), NDUFS4 (log2FC = 1.06), NDUFA9 (log2FC = 1.03), isocitrate dehydrogenases IDH3A (log2FC = 1.08), IDH1 (log2FC = 1.05) and succinate dehydrogenase subunit SDHB (log2FC = 1.09) exhibited high expression levels. Dysregulated aerobic respiration may have led to apoptotic mitochondrial changes (FE = 4.3, pFDR < 0.05), increased proteasome assembly (FE = 6.0, pFDR < 0.05) and autophagy (z = 3.3, − log(p) = 1.7), which might serve to clear cells from dysfunctional mitochondria. Important genes involved in autophagy processes such as BCL2 interacting protein 3 (BNIP3, log2FC = 1.11), VPS35 retromer complex component (VPS35, log2FC = 1.01), as well as FUN14 domain containing 1 (FUNDC1, log2FC = 1.76) were upregulated in the oxidative-stress-challenged nCLU cell line. Ultimately, autophagy processes and redox imbalance may have induced ferroptosis (z = 1.2, − log(p) = 1.9), which is an iron-dependent form of non-apoptotic cell death [97] and necroptosis signaling pathway (z = 4.2, − log(p) = 1.5), another form of non-apoptotic cell death [74]. In accordance, low expression of glial cell line derived neurotrophic factor (GDNF, log2FC = − 1.45) may reduce survival pathways [12]. In summary, the nCLU-transfected cells attempted to counter the deleterious effects of dysregulated mitochondrial aerobic respiration by increasing antioxidant expression, but ultimately experienced elevated cell death at oxidative stress conditions.
In contrast to the nCLU cell line, the sCLU-transfected cells already demonstrated substantial differences in gene expression at normoxic conditions, which included energy metabolism and autophagy processes. Adenosine triphosphate (ATP) provides the energy necessary to drive all energy-demanding processes in living cells. Elevated ATP levels of sCLU-transfected cells have already been demonstrated in cell viability assays [31] and were confirmed in this study. At aerobic conditions, the main ATP production usually takes place at the mitochondrial electron transport chain. However, this process is also associated with the generation of ROS, which can be detrimental in high concentrations. The high expression of pyruvate dehydrogenase kinase 1 (PDK1, log2FC = 1.20) in sCLU cells might inactivate pyruvate dehydrogenase and consequently inhibit the first step of the citric acid cycle. In carcinoma and fibroblast cell cultures, overexpression of PDK1 shifted ATP production from mitochondrial respiration to glycolysis, thereby attenuating hypoxic ROS generation and rescuing cells from hypoxia-induced apoptosis [52, 79]. Additionally, low expression of myeloid translocation gene 16 product (MTG16, also abbreviated as CBFA2T3, log2FC = − 1.14) was observed in the sCLU cell line. MTG16 reduced the expression of PDK1 and genes involved in glycolysis in lymphoma cells [57], and reduced levels of MTG16 might therefore indicate enhanced glycolytic capacity. Consistent with this, sCLU expression correlated with high expression of the monocarboxylate transporter 4 (MCT4, also abbreviated as SLC16A3, log2FC = 1.20). MCT4 has an important role in tissues reliant on glycolysis [36], by facilitating lactate efflux and preventing pyruvate efflux, thereby enabling conversion of pyruvate to lactate and regeneration of NADH for glycolysis [36]. MCT4 was also upregulated in hooded seal brain slices that were exposed to hypoxia and reoxygenation in vitro [39] and MCT4 was more highly expressed in hooded seal than in mouse neurons [30]. Exported lactate by MCT4 might be further metabolized by neighboring astrocytes in the hooded seal brain that exhibit high levels of lactate dehydrogenase b (LDHB) [40], as suggested by the ‘reverse lactate shuttle’ hypothesis, first presented by Mitz et al. [73]. Additionally, aerobic metabolism was found to be decreased in the visual cortices of hooded seals compared to ferrets [20]. In contrast, Geßner et al. [30] concluded that mitochondrial function and numbers may have been enhanced, while glycolytic capacity was slightly lower, in neurons of the hooded seal compared to mice. Arguably, these differences might be related to the choice of the non-diving model organism, i.e., ferrets vs mice, which are known to maintain quite different basal metabolic rates [30]. Further, we here consider the effect of particular genes, whereas Geßner et al. [30] analysed the neuronal transcriptome as a whole. In summary, alterations of pathways by sCLU during normoxia may indicate a preparation or pre-adaptation of hooded seals to upcoming diving-associated stress conditions, such as low oxygen levels or ROS production. To prepare for these conditions, capacity for aerobic respiration might be decreased and capacity for anaerobic glycolysis might be increased. Cellular stress responses comprise mechanisms that minimize acute damage and promote cell survival. Oxidative stress might denature proteins, thereby disrupting protein homeostasis (proteostasis) necessary for biological function and cell metabolism [10]. Chaperones can help defend the cell against damage by facilitating protein folding, ensuring that proteins assume their necessary shape. Misfolded proteins may be degraded by proteasomes or autophagy, to remove potentially toxic aggregates. Clusterin (CLU) may play an important role as stress-induced secreted chaperone protein, mediating proteasomal degradation of misfolded proteins [45, 93], and CLU is known to protect neuronal cells against intracellular protein aggregation and cytotoxicity [34]. Interestingly, in IPA analyses the endoplasmic reticulum stress pathway (z = 0.4, − log(p) = 5.23) and unfolded protein response (z = 0.3, − log(p) = 2.84) were slightly activated in the normoxic sCLU-transfected cells, suggesting the contribution of CLU to protein homeostasis in the hooded seal brain. In addition, the chaperone peptidylprolyl isomerase C (PPIC, log2FC = 1.15) was upregulated in sCLU cells, which may also be important for coping with oxidative stress [58, 67]. Additionally, autophagy (z = 0.2, − log(p) = 2.51), which promotes degradation of damaged proteins was slightly activated in the sCLU-transfected cell line. In particular, BNIP3, which is necessary for clearing dysfunctional mitochondria with low membrane potential and reducing the buildup of ROS to promote cell survival [78], was observed to be more highly expressed in the sCLU cell line (log2FC = 1.33). Furthermore, NRF2-mediated oxidative stress response (z = − 2.7, − log(p) = 1.5) (Fig. 7), which coordinates the basal and stress-inducible activation of a vast array of cytoprotective genes, like antioxidants [92], was downregulated in sCLU cells. A reduction of this pathway may imply decreased stress and reduced necessity to detoxify ROS. In accordance with our results, Geßner et al. [31] found no increased caspase activity and ROS amount for sCLU-transfected HN33 cells at normoxia, indicating no increased stress. In conclusion, sCLU might contribute to the stress tolerance of the hooded seal brain by improving autophagy and protein folding pathways as well as glycolytic capacity.
When exposed to hypoxic stress, the sCLU cell line exhibited a similar transcriptome response to the nCLU-transfected cells. IPA pathways glycolysis I (z = 3.0, − log(p) = 3.4), gluconeogenesis (z = 2.1, − log(p) = 2.7) and glycogen biosynthesis (− log(p) = 2.0) and GO terms glycolytic process (FE = 5.7, pFDR < 0.05) and glycogen biosynthetic process (FE = 8.1, pFDR < 0.05) were enhanced (Fig. 8). The related glycogen storing enzyme EPM2A, important for accurate accumulation of glycogen [86], also demonstrated increased transcription (log2FC = 2.07). On the other hand, aerobic respiration (FE = 3.3, pFDR < 0.05) was decreased in sCLU-transfected cells. Aerobic respiration may have been reduced by high expression of PDK1 (log2FC = 1.01) and low expression of MTG16 (log2FC = − 0.46) in sCLU-transfected cells, which inhibit the first step of the TCA cycle and improve glycolytic capacity [52, 57, 79]. However, regulation of PDK1 and MTG16 genes was not as strong as at normoxia. This was probably related to that mock cells to some extent also downregulated aerobic respiration at hypoxia and the difference between these cell lines may have become less distinct. GO term cellular response to hypoxia (FE = 21.0, pFDR < 0.05) in the sCLU cell line may have also been mediated by HIF1A as upstream regulator (z = 2.0, p < 0.01) and HIF1A signaling (z = 3.1, − log(p) = 1.5). Furthermore, egl-9 family hypoxia-inducible factor 1 (EGLN1) was highly expressed (log2FC = 1.06), which hydroxylates HIF proteins and thereby targets them for degradation [26]. However, the hydroxylation reaction might have been attenuated by limited oxygen availability. In addition to high expression of BNIP3 (log2FC = 1.36), HIF1-upregulated mitochondria-localized glutamic acid rich protein (MGARP, log2FC = 1.33) was highly expressed in sCLU cells and might further support mitophagy [43], which may reduce deleterious ROS by clearance of dysfunctional mitochondria. Additionally, upregulation of PPIC (log2FC = 2.06) may have further improved protein folding and reduced ER stress in the sCLU-transfected cell line [58, 67]. According to described neurotrophic functions, intrinsic apoptotic signaling (FE = 3.6, pFDR < 0.05) as well as MYC mediated apoptosis (z = − 1.7, − log(p) = 1.6) were found to be downregulated in the sCLU cell line. Decreased apoptosis is in line with previous findings, e.g. as reflected by decreased caspase activity and phosphatidylserine exposure in transfected cells [31]. Overall, sCLU cells might enhance stress resistance and reduce apoptosis in response to hypoxia.
When applying oxidative stress, the sCLU cell line exhibited a limited DEG response. This seems counterintuitive, since sCLU cells had a significantly higher viability in oxidative stress than nCLU and mock cells [31]. Normoxic (i.e., constitutional) differences in gene expression might therefore already have prepared sCLU-transfected cells for oxidative stress, indicating a pre-adaptive response to upcoming stress conditions. Nevertheless, the upregulation of some genes might assist sCLU in protein folding such as cAMP responsive element binding protein 3-like 2 (CREB3L2, log2FC = 1.26) and PPIC (log2FC = 2.17). CREB3L2 protected cells from ER stress-induced death in a neuroblastoma cell line [55], while the chaperone PPIC might have further promoted protein folding and reduce oxidative stress [67]. Additionally, pathways involved in synaptic signaling were elevated in the sCLU cell line at oxidative stress such as calcium signaling (z = 2.5, − log(p) = 1.8), semaphorin neuronal repulsive signaling (z = 1.3, − log(p) = 1.9) and synaptic long term depression (z = 1.9, − log(p) = 2.8) (Fig. 9). High expression of voltage-dependent calcium channels CACNA1G (log2FC = 1.48) and CACNA1I (log2FC = 1.32) may have facilitated calcium flux and subsequent binding of calcium to synaptotagmin I (SYT1, log2FC = 1.33), which may have triggered neurotransmitter release at the synapse [22]. Especially serotonin may have functioned as neurotransmitter in the oxidative stress-exposed sCLU cell line. Tryptophan hydroxylase 2 (TPH2, log2FC = 1.74), which catalyzes the first rate-limiting step in serotonin biosynthesis [41], as well as the serotonin receptor HTR3A (log2FC = 1.43) demonstrated increased expression. Binding of serotonin to HTR3A causes fast, depolarizing responses in neurons [5], but may also regulate the development of the mammalian central nervous system [19]. Described genes might support observed decrease in lipid peroxidation and caspase activity in a previous study [31] and protect sCLU-transfected cells from oxidative stress induced cell death. The majority of preventive measures in sCLU cells though may have already been taken at normoxic conditions.
Similar to the sCLU cell line, the S100B-transfected cells exhibited elevated ATP levels at normoxic conditions, which is in accordance to a previous study [31]. Likewise, glycolytic process (FE = 6.6, pFDR < 0.05) and MCT4 expression (log2FC = 1.87) were elevated in the S100B cells at normoxia, while the TCA cycle may have been inhibited by high expression of PDK1 (log2FC = 1.14). Consequently, capacity for glycolytic metabolism may have been increased and aerobic respiration decreased in the S100B cell line at normoxic conditions. Furthermore, S100B may play a role in neurodegeneration or neuroprotection [27]. Although genes associated with the GO term neuron development (FE = 2.3, pFDR < 0.05) demonstrated reduced expression in S100B-transfected cells, increased expression of neurotrophic factors such as growth-associated protein 43 (GAP43, log2FC = 1.52) and brain-derived neurotrophic factor (BDNF, log2FC = 0.83) might enhance neuron growth and survival. GAP43 may regulate synaptic plasticity and neurite outgrowth [95] and is also an important mediator of the neuroprotective effects of BDNF in connection with excitotoxicity [35]. The growth factor BDNF is one of the most widely distributed and extensively studied neurotrophins in the mammalian brain [56], and is associated with neuronal maintenance, survival, plasticity, and neurotransmitter regulation [32]. Furthermore, the neurotrophic growth factor pleiotrophin (PTN, log2FC = 0.37) and its receptor anaplastic lymphoma kinase (ALK, log2FC = 1.20) were upregulated in S100B-transfected cells [48]. Loss of PTN in pericyte-ablated mice has been linked to a rapid neurodegeneration cascade [75], which illustrates its role in neuroprotection. Additionally, tissue-type plasminogen activator (PLAT, log2FC = 1.03) and a PLAT inhibitor (SERPINI1, log2FC = 1.07), which are involved in synaptic plasticity [9] and have been reported to facilitate neuron survival depending on concentration and isoform [11], were upregulated in S100B cells. In summary, although most genes involved in neuron development were slightly downregulated in normoxic S100B cells, highly upregulated important neurotrophic factors may have promoted neuronal plasticity and survival. During synaptic signaling, the presynaptic neuron membrane at the synapse depolarizes, causing influx of calcium ions and release of neurotransmitters. Reestablishing the resting membrane potential, calcium gradient and neurotransmitter levels, are all highly energy-intensive processes in the brain. S100B is involved in calcium homeostasis and thereby presumably also in regulating synaptic plasticity [16]. It has been suggested that its high expression in marine mammals may help prevent excitotoxicity by reducing free intracellular Ca2+ and thereby attenuate continued release of glutamate and other neurotransmitters [28]. In the S100B-transfected cells neurotransmitter release might be reduced by activation of opioid signaling (z = 1.2, − log(p) = 2.1) (Fig. 10), which can prevent calcium ion influx and facilitate potassium ion efflux, thereby causing membrane hyperpolarization and a reduced neurotransmitter release [33]. The upregulated OPRD1 (log2FC = 0.80) may have facilitated clearance of neurotransmitters from the synaptic cleft and prevent neuroexcitotoxicity [37]. Furthermore, levels of purkinje cell protein 4 (PCP4) were decreased in S100B cells (log2FC = − 1.96). PCP4 is a small calmodulin-binding protein that promotes calcium exchange and neurotransmitter release [87]. Its downregulation might further support a reduction of neurotransmitter levels in S100B cells. A previous study that compared the transcriptomes of hooded seal and mouse neurons also indicated reduced glutamatergic transmission in the seal, by reduced expression of glutamate receptors, while glutamate uptake was increased [30]. Noh et al. [77] stated that glutaminergic synapse function might have been commonly positively selected in pinnipeds, indicating its importance for the adaptation to the marine environment. In summary, S100B may contribute to the neuronal hypoxia tolerance by reducing neurotransmission and thus, our findings support the observations of Geiseler et al. [28] and Geßner et al. [30]. A reduction in neurotransmission might ultimately serve to reduce energy consumption and thereby the oxygen needs of neuronal cells.
In analogy to the other transfected cell lines, hypoxia response of the S100B cell line might have been mediated by activated HIF1A signaling (z = 3.1, − log(p) = 1.4) (Fig. 11). GO terms tricarboxylic acid cycle (FE = 6.37, pFDR < 0.05), oxidative phosphorylation (FE = 6.11, pFDR < 0.01) and aerobic respiration (FE = 5.4, pFDR < 0.01) as well as IPA pathway TCA cycle (z = − 2.6, − log(p) = 2.1) were decreased in S100B-transfected cells, but PDK1 upregulation (log2FC = 0.65) was not as prominent in inhibiting the first step of TCA cycle as in normoxic conditions. In accordance with the putative function of S100B GO term calcium-ion regulated exocytosis (FE = 6.1, pFDR < 0.05) was enriched in upregulated genes of the S100B-transfected cells. Furthermore, opioid signaling (z = 1.9, − log(p) = 1.9) and synaptogenesis signaling pathway (z = 1.6, − log(p) = 1.4) were enhanced in the S100B-transfected cells when subjected to hypoxia. As mentioned before, high expression of OPRD1 (log2FC = 0.99) and low expression of PCP4 (log2FC = − 1.89) may reduce neurotransmitter levels and protect cells from hypoxia-induced excitotoxicity [37, 87]. Additionally, GAP43 (log2FC = 1.45) and BDNF (log2FC = 1.17) may facilitate neuron survival [32, 35]. According to described neurotrophic functions, intrinsic apoptotic signaling (FE = 4.8, pFDR < 0.01) was enriched in downregulated genes of S100B-transfected cells at hypoxic conditions. Therefore, in addition to decreasing aerobic respiration, S100B may facilitate neuroprotection of neuronal cells by downregulating synaptic signaling and upregulating neurotrophic factors at hypoxic conditions.
Similar to the sCLU-transfected cells, the S100B cell line did not exhibit a diverse DEG response at oxidative stress. However, elevated viability [31] may point to pre-adaptive mechanisms already carried out at normoxia. When exposed to oxidative stress, the only two activated pathways in IPA analysis were semaphorin neuronal repulsive signaling pathway (z = 1.9, − log(p) = 3.07) and IL-15 production (z = 1, − log(p) = 1.51) (Fig. 12). The former already demonstrated activation at normoxia (z = 0.8, − log(p) = 2.98). Semaphorin such as SEMA6D (log2FC = 1.35) and the semaphorin co-receptor neuropilin 1 (NRP1, log2FC = 1.02) may have played an essential role in axonal guidance signaling (− log(p) = 2.4) and thereby overall nervous system development at oxidative stress conditions [46]. Additionally, high expression of OPRD1 (log2FC = 1.06) and low expression of PCP4 (log2FC = − 1.86) may have facilitated neuroprotection by reduction of neurotransmitter levels as mentioned before [37, 87]. These mechanisms may aid in protecting S100B-transfected cells from oxidative stress induced cell death. However, metabolic alterations at normoxia may have already prepared cells for imminent stress conditions.
Clusterin (CLU) and S100B are highly expressed in the hooded seal brain and probably represent two of numerous factors that contribute to its intrinsic hypoxia tolerance. In order to investigate their potential roles, we transfected HN33 cells with soluble clusterin (sCLU), nucleus clusterin (nCLU) and S100B, subjected these cell lines to three challenges; normoxia, hypoxia and oxidative stress, and studied viability and differential gene expression (DEG) responses (Tables 1, 2, 3). We found that aerobic metabolism was reduced in the sCLU and S100B cell lines, and that synaptic signaling pathways were reduced in S100B-transfected cells, at normoxic conditions. These transcriptomic responses might reduce production of reactive oxygen species (ROS), while also reducing the energy consumption of neuronal cells. Additionally, autophagy processes appeared to be important for cellular homeostasis in sCLU-transfected cells, which might ultimately protect cells from apoptosis. When oxidatively stressed, sCLU- and S100B-transfected cells did not mount similar gene regulatory responses, but nevertheless demonstrated improved viability compared to mock-transfected cells, presumably due to a pre-adaptive (constitutional) response, seen already under normoxic conditions, in preparation for upcoming stress conditions. In contrast to this effect, the nCLU cell line exhibited elevated stress and apoptosis pathways in response to oxidative stress, which suggests a reduced basal protection against oxidative damage in this cell line. Furthermore, known hypoxia response genes and pathways, such as HIF1A signaling and glycogen metabolism, were enhanced in transfected cells when exposed to hypoxic conditions. While the roles of CLU and S100B in neurodegenerative diseases are being debated, we found evidence for the upregulation of neuroprotective effects in cell lines overexpressing these genes, in response to hypoxia and oxidative stress. The findings of the present study have been demonstrated in a cell culture model and effects would still need to be confirmed in vivo. Unfortunately, it is not feasible to obtain samples from naturally diving hooded seals that experience hypoxia. Alternatively, fresh brain slices exposed to hypoxia in vitro could mimic more closely natural conditions than cell culture. However, capturing hooded seals and performing experiments on fresh tissue requires great effort and have only been done on rare occasions [14, 39]. Cell culture experiments therefore represent a great possibility to mimic hypoxic conditions. In this study, we highlighted pathways and targets of hypoxia tolerance that may provide clues to tackle neurodegenerative diseases such as Alzheimer’s disease and Parkinson’s disease. While the cell culture experiments indicated neuroprotective effects of CLU and S100B at hypoxia and oxidative stress, these results yet require confirmation in in vivo studies.
HN33 cells (murine hippocampal neurons × neuroblastoma) [60] (American Type Culture Collection, Rockville, MO) had been stably transfected with each candidate gene (nuclear clusterin (nCLU), soluble clusterin (sCLU) and S100B) and an empty vector (mock), respectively [31]. The four cell lines were cultivated in Dulbecco’s Modified Eagle Medium (DMEM) (Biowest, Darmstadt) containing 10% fetal calf serum (FCS) (Biowest, Darmstadt, Germany) and 1% of a mixture of penicillin and streptomycin (PAA, Pasching, Austria) at 37 °C in a humidified atmosphere and 5% CO2. The medium of the transfected cells was supplemented with 700 µg/ml geniticin (PAA, Pasching, Austria).
The successful overexpression of target genes was verified before and after experiments by qRT-PCR as described in Geßner et al. [31]. For that purpose, RNA was extracted from cells using the Crystal RNA Mini Kit (BiolabProducts, Gödenstorf, Germany) including an on-column DNA digestion with RNase-free DNase (Qiagen, Germany). First-strand cDNA was synthesized from 1 µg of isolated RNA with Oligo(dT)18 primer using the Fermentas RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Scientific, Schwerte, Germany). The qPCR was performed with a 7500 Fast Real-Time PCR System and the Power SYBR Green master mix (Applied Biosystems, Darmstadt, Germany) using a standard PCR protocol (step 1–2: 50 °C 2 min, 95 °C 10 min; 40 cycles step 3–5: 95 °C for 15 s, 60 °C for 30 s, 72 °C for 30 s) including melting curve analysis. Absolute mRNA copies were calculated with the 7500 System Sequence Detection Software 2.0.6 (Applied Biosystems) using recombinant plasmid dilutions of 102–108 as standard curve, and then normalized to 1 µg of total RNA.
Experiments were conducted in 96-well plates containing 3.75 × 104 cells per well diluted in 50 µl DMEM medium (10% FCS, 1% Penicillin/Streptomycin) of each transfected cell line at passage 38, including a cell line transfected with an empty vector (mock cell line). Cells were exposed to normoxia (21% O2, 5% CO2, 37 °C), hypoxia (1.2% O2, 5% CO2, 93.8% N2, 37 °C) and oxidative stress (275 µM H2O2 in 50 µl DMEM per well, 21% O2, 5% CO2, 37 °C) for 24 h, respectively. After trypsinization every 6 wells were pooled and used as one replicate, generating four replicates per cell line and treatment. Samples were centrifuged at 180×g for 5 min, supernatant removed and pellets stored at − 20 °C until further processing. We note that we considered 21% as normoxic condition. The HN33-cells used in this study were cultured at 21% O2 since their dissociation and somatic cell fusion with neuroblastoma cells [60]. In these conditions, the cells displayed expression of neurofilaments and electrophysiological behavior typical of hippocampal neurons [60]. While other hippocampal neurons may experience 21% O2 as hyperoxic, the HN33-cells have been exposed to 21% O2 over many generations and hence arguably perceive this condition as normoxic. Related studies (e.g. [31, 53, 54]) also considered 21% as normoxia.
Cell viability was assessed by CellTiter-Glo® (CTG) Luminescent Cell Viability Assay Kit (Promega, Mannheim, Germany) according to the manufacturer’s instructions. The assay determines the ATP content of the cells and serves as reliable indicator of the number of healthy, metabolically active cells [76]. After incubation at normoxia, hypoxia and oxidative stress, as mentioned above, CTG reagent was added and luminescence measured by a DTX 880 Multimode Detector (Beckmann Coulter, Krefeld, Germany). Statistical analysis was conducted in R version 4.1.2 [85]. Robust triplicates were determined and intensities normalized to the mock cell line at normoxic conditions. Pairwise t-test and false discovery rate (FDR) multiple correction testing was performed using the compare_means function of ggpubr package, with the mock cell line as reference group at each respective condition [49].
Total RNA of frozen cell pellets was extracted with the Crystal RNA Mini Kit (BiolabProducts, Gödenstorf, Germany) after the manufacturer’s instructions, including an on-column DNA digestion with RNase-free DNase (Qiagen, Germany). RNA integrity and quantity were assessed with the Agilent 4200 TapeStation system (Agilent Technology, Sanat Clara, USA) and triplicates determined for sequencing. The cDNA libraries were generated with 500 ng of total RNA after rRNA depletion, and sequenced on a NovaSeq platform with a setting of 150 bp paired-end reads and an estimated output of 50 million reads (GeneWiz, Leipzig, Germany). The raw transcriptome files are available from the NCBI Sequence Read Archive (SRA) from cell lines transfected with mock vector, nCLU, sCLU and S100B at normoxic, hypoxic and oxidative stress conditions (Additional file 1: Table S1).
Sequencing files were uploaded to a Galaxy platform in fastq.gz format for further analysis. A sequencing quality report was generated using FastQC (Galaxy Version 0.72) and MultiQC (Galaxy Version 1.7). Since read quality was good (average Phred score > 35, Additional file 1: Table S1), no further read trimming was performed. Reads were mapped against the mouse reference genome GRCm39 (http://www.ncbi.nlm.nih.gov/assembly/GCF_000001635.27/, genomic FASTA and GTF) with Bowtie2 (Galaxy Version 2.3.4.2) with the very sensitive end-to-end preset setting (–very-sensitive). Mapped reads from generated genome BAM files were filtered by a minimum mapping quality of 10 and determined with featureCounts (Galaxy Version 1.6.3 + galaxy2), counting aligned fragments (even when only one paired read mapped) and excluding chimeric fragments. Reads were allowed to contribute to one feature only. Differentially expressed genes (DEGs) were determined from count tables using DESeq2 (Galaxy Version 2.11.40.6), performing pairwise comparisons with the mock cell line at the respective condition as reference. Only genes with a corrected FDR p-value < 0.05 were considered significant. Principal component analysis (PCA) was performed on count tables from all cell lines and treatments with DESeq2 (Version 1.32.0) in R (Version 4.1.0). Gene Ontology (GO) analysis was performed using PANTHER Overrepresentation Test (Protein Analysis Through Evolutionary Relationships, http://go.pantherdb.org/, GO Ontology database 10.5281/Zenodo.5228828 Released 2021-08-18) [70]. The annotated mouse genes in the PANTHER DB were used as a reference list, and overrepresentation was tested in GO and GO Slim terms and Reactome pathways with Fisher's Exact Test with FDR multiple testing correction. Only categories with corrected p-values < 0.05 were considered significant. Enrichment in Canonical Pathways and Upstream Regulator Analysis were performed with Qiagen’s Ingenuity Pathway Analysis (IPA, Qiagen, Hilden, Germany) Core analysis tool.
Additional file 1: Figure S1. Expression of endogenic and transgenic CLU and S100B sequences in transfected HN33 cell lines at normoxia, determined by qPCR experiments. Differences in Ct-values between endogenic and transgenic nCLU, sCLU and S100B were 13.17 (with Ct of 40 for endogenic nCLU), 13.25 and 9.48, respectively. The fold-expression difference for nCLU, sCLU and S100B therefore were 213.17, 213.25 and 29.48, respectively. Figure S2. TPM values of endogenic (CLU, S100B) and transgenic [CLU (Ccr), S100B (Ccr)] sequences in transfected cell lines (mock, nCLU, sCLU, S100B) at normoxia, hypoxia and oxidative stress. Table S1. Sequencing and mapping overview. Triplicates were sequenced per cell line and oxygen treatment. For replicate mock-H2O2-2 sequencing failed and was discarded. Around 51 million reads per sample were generated of which around 75% mapped to the GRCm39 mouse reference genome. | true | true | true |
|
PMC9571541 | 36235643 | Yunyi Xie,Han Qi,Wenjuan Peng,Bingxiao Li,Fuyuan Wen,Fengxu Zhang,Ling Zhang | SNPs in lncRNA KCNQ1OT1 Modulate Its Expression and Confer Susceptibility to Salt Sensitivity of Blood Pressure in a Chinese Han Population | 26-09-2022 | salt sensitivity,acute salt loading,blood pressure,lncRNA,single-nucleotide polymorphism | Long noncoding RNA (lncRNA) plays an important role in cardiovascular diseases, but the involvement of lncRNA in salt sensitivity of blood pressure (SSBP) is not well-known. We aimed to explore the association of sixteen single-nucleotide polymorphisms (SNPs) in five lncRNA genes (KCNQOT1, lnc-AGAP1-8:1, lnc-IGSF3-1:1, etc.) with their expression and susceptibility to SSBP. A two-stage association study was conducted among 2057 individuals. Quantified expression of the lncRNA was detected using real-time PCR. Genotyping was accomplished using the MassARRAY System. The expression quantitative tra2it loci test and the generalized linear model were utilized to explore the function of SNPs. One-sample Mendelian randomization was used to study the causal relationship between KCNQOT1 and SSBP. Significant effects were observed in KCNQ1OT1 expressions on the SSBP phenotype (p < 0.05). Rs10832417 and rs3782064 in KCNQ1OT1 may influence the secondary structure, miRNA binding, and expression of KCNQ1OT1. Rs10832417 and rs3782064 in KCNQ1OT1 were identified to be associated with one SSBP phenotype after multiple testing corrections and may be mediated by KCNQ1OT1. One-sample Mendelian randomization analyses showed a causal association between KCNQ1OT1 and SSBP. Our findings suggest that rs10832417 and rs3782064 might be associated with a lower risk of SSBP through influencing the KCNQ1OT1 secondary structure and miRNA binding, resulting in changes in KCNQ1OT1 expression. | SNPs in lncRNA KCNQ1OT1 Modulate Its Expression and Confer Susceptibility to Salt Sensitivity of Blood Pressure in a Chinese Han Population
Long noncoding RNA (lncRNA) plays an important role in cardiovascular diseases, but the involvement of lncRNA in salt sensitivity of blood pressure (SSBP) is not well-known. We aimed to explore the association of sixteen single-nucleotide polymorphisms (SNPs) in five lncRNA genes (KCNQOT1, lnc-AGAP1-8:1, lnc-IGSF3-1:1, etc.) with their expression and susceptibility to SSBP. A two-stage association study was conducted among 2057 individuals. Quantified expression of the lncRNA was detected using real-time PCR. Genotyping was accomplished using the MassARRAY System. The expression quantitative tra2it loci test and the generalized linear model were utilized to explore the function of SNPs. One-sample Mendelian randomization was used to study the causal relationship between KCNQOT1 and SSBP. Significant effects were observed in KCNQ1OT1 expressions on the SSBP phenotype (p < 0.05). Rs10832417 and rs3782064 in KCNQ1OT1 may influence the secondary structure, miRNA binding, and expression of KCNQ1OT1. Rs10832417 and rs3782064 in KCNQ1OT1 were identified to be associated with one SSBP phenotype after multiple testing corrections and may be mediated by KCNQ1OT1. One-sample Mendelian randomization analyses showed a causal association between KCNQ1OT1 and SSBP. Our findings suggest that rs10832417 and rs3782064 might be associated with a lower risk of SSBP through influencing the KCNQ1OT1 secondary structure and miRNA binding, resulting in changes in KCNQ1OT1 expression.
Salt sensitivity of blood pressure (SSBP) is determined as a quantitative trait in which the blood pressure (BP) for parts of the population displays variants that parallel changes in sodium loading [1]. Studies have revealed that individuals in the population have various BP responses to salt load and display various SSBP phenotypes [2]. Individuals who exhibit elevations in BP paralleled with high salt intake are viewed as salt-sensitive (SS), whereas others are viewed as salt-resistant (SR) [2]. The Genetic Epidemiology Network of Salt Sensitivity (GenSalt) study declared that the prevalence of SS, which is viewed as a qualitative trait of SSBP, is generally 30% in North Chinese adults [2,3]. SSBP is not only the intermediate phenotype for developing hypertension but also is detrimental to the development of cardiovascular diseases (CVDs) that increase mortality [3,4]. Long-term follow-up research for normotensive and hypertensive individuals showed evidence for SSBP increasing mortality [5]. Thus, the early detection of SSBP, subsequent salt-limiting interventions, and personalized medicine may benefit from reducing the burden of CVDs. Genes can participate in SSBP pathology via different pathways such as the renin–angiotensin–aldosterone system and ion–water channel [1,6]. Single-nucleotide polymorphisms (SNPs) are the most frequent genetic variants and are associated with gene expression, function, and diseases [7]. SNPs in some coding genes or noncoding genes have been found to be associated with SSBP. Some genome-wide association studies (GWAS) and candidate gene research have been conducted for SSBP or SS [8,9]. The GenSalt study identified eight novel loci for BP responses to dietary sodium and potassium intervention and the cold pressor test in the Han Chinese population [8,10]. Citterio et al. [11] demonstrated that SNPs located in the PRKG1 gene are associated with SSBP in Caucasian individuals with mild hypertension under the acute salt-loading test in a GWAS study. Our previous study suggested that SNPs are located in protein-coding genes such as PRKG1 and SLC8A1 [10,12]. While about 2 percent of the human genome encodes proteins, most of them are detectably transcribed in certain circumstances [11,13]. Long noncoding RNAs (lncRNAs) are defined as a type of non-protein-coding transcripts of more than 200 nucleotides [12,14]. LncRNAs have been suggested to perform several functions such as transcriptional regulation [13,15]. Our previous studies have reported the transcriptome profiles of SSH and constructed a ceRNA network to help elucidate the mechanism of SSH [14,16], and reported that lncRNAs could participate in the biological pathways in SS. We further detected that lncRNA lnc-IGSF3-1:1, SCOC-AS1, and SLC8A1-AS1 could perform as circulating biomarkers of SS [15,17]. SNPs in lncRNAs may be involved in the disease by linking to modification of the lncRNA sequence or altering their gene expression levels and influencing their regulatory capacity [18]. Additionally, SNP-caused structural disturbance within the lncRNAs could disturb lncRNAs’ molecular functions and, thus, is likely to be involved in the physiological pathways of disease [13,15]. In the present research, based on our current discovered SS-related lncRNAs, we integrated epidemiological analysis and a bioinformatics prediction method to explore the associations between five lncRNAs (KCNQOT1, lnc-AGAP1-8:1, lnc-IGSF3-1:1, SCOC-AS1, and SLC8A1-AS1) and SSBP in order to ascertain the involvement of lncRNA-SNPs in SSBP susceptibility and its potential mechanism. The flow chart for this study is shown in Figure 1.
A total of 1684 unrelated individuals from the Systems Epidemiology Study on Salt Sensitivity (EpiSS) between July 2014 and July 2016 were enrolled in this two-stage association study, the detailed information of which has been published previously [17,19]. In North China, individuals recruited from Tieling were used as discovery sets, and those recruited from Beijing were analyzed as replication sets. Written informed consent was obtained from enrolled individuals before any research-specific tests. The modified Sullivan’s acute oral saline load and diuresis shrinkage test (MSAOSL-DST) was used to assess the SSBP [10,12]. DNA samples were available for 1684 participants. Fasting venous blood samples were assembled before the assessment and utilized for serum biochemical and RNA examination for 251 participants out of the 1684 participants. The study was approved by the Capital Medical University ethics committee (no. 2013SY22) and was registered in the WHO International Clinical Trials Registry Platform (No: ChiCTR-EOC-16009980).
The determination of SSBP was conducted using MSAOSL-DST, the details of which have been previously reported [16,19,20]. All the participants were asked to pause taking antihypertensive drugs for a full day before the assessment. Each fasting participant received an oral administration of 1000 mL 0.9% saline solution within half an hour and orally took 40 mg furosemide two hours after the saline loading. After a 5 min break in the sitting position, automatic sphygmomanometers (Omron HEM-7118, Kyoto, Japan) were used to test BP two times at 1 min intervals. The mean value was computed as the final BP value. The BP was tested 3 times in the following order: before the test, 2 h after the participant finished drinking the given saline solution (acute salt-loading process), and 2 h after taking oral furosemide (diuresis shrinkage process). Systolic BP (SBP) and diastolic BP (DBP) were recorded. Mean arterial pressure (MAP) was calculated using the equation MAP = (SBP + 2 × DBP)/3 [21]. Participants with a rise in MAP of at least 5 mmHg after salt loading and (or) a reduction of more than 10 mmHg after diuresis shrinkage were viewed as SS; otherwise, they were viewed as salt-resistant (SR). SSBP phenotypes are defined as several continuous variables including MAP change 1 (MAP after the acute salt-loading process minus the baseline MAP) and MAP change 2 (MAP after the diuresis shrinkage process minus MAP after the acute salt-loading process).
Five lncRNAs (KCNQOT1, lnc-AGAP1-8:1, lnc-IGSF3-1:1, SCOC-AS1, and SLC8A1-AS1) were selected based on our previous study [15,17]. RNA was extracted using a PAXgene Blood RNA Kit (cat. No. 762174, QIAGEN GmbH, Hilden, Germany) following the instructions stipulated by the manufacturer. Quantitative RT-PCR (qRT-PCR) assays were conducted to detect the levels of lncRNAs by utilizing the SYBR Green qPCR Master Mix reagent kit (MedChemExpress) on an ABI 7900HT Real-Time PCR System (Applied Biosystem, Foster City, CA, USA), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was utilized as the internal control. The relative changes in gene expression were calculated by using the 2−ΔΔct method [18,22]. Each sample was measured in triplicate using the average value. The primer sequences for qRT-PCR have been summarized in our previous study [15,17].
Candidate SNPs within the selected 5 lncRNAs genes and their ±5 kb flanking regions were searched with the following criteria: ① Minor allele frequencies (MAF) ≥ 5%, P for the Hardy–Weinberg equilibrium (HWE) ≥ 0.05, and calling rate ≥ 95%. ② Linkage disequilibrium (LD) r2 ≥ 0.8 was selected from the 1000 Genomes CHB population (http://www.internationalgenome.org/ accessed on 15 August 2022). ③ The Genotype-Tissue Expression (GTEx) database was used to find the expression quantitative trait loci (eQTLs), RegulomeDB (https://regulomedb.org/ accessed on 15 August 2022) [19,23], HaploReg (https://pubs.broadinstitute.org/ accessed on 15 August 2022) [20,24], and 3DSNP (https://omic.tech/3dsnpv2/ accessed on 15 August 2022) of the selected functional SNPs [21,25]. Genomic DNA was isolated from 200 μL of a suspension of EDTA-anticoagulated peripheral blood leukocytes utilizing the Magnetic Beads Whole Blood Genomic DNA Extraction Kit by using automatic nucleic acid extraction apparatus (BioTeke, Beijing, China). A NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) was used to measure the concentration and purity of the extracted DNA. All SNPs were genotyped utilizing the high-throughput sequencing method on the Sequenom Mass ARRAY Platform (Sequenom, San Diego, CA, USA).
As lncRNA-SNP may be involved in disease pathology by changing the expression or functioning of downstream target genes through several mechanisms, bioinformatics analyses were conducted to further explore their functions. The minimum free energy (MFE) structure algorithm was used to predict the RNA secondary structure [26]. The MFE estimation is used to determine the predicted structure with the lowest free energy because it is presumed that the lower the value, the more reliable and possible the structure. MFE structure predictions were calculated using the Vienna RNA package RNAfold [23,27]. The binding sites of lncRNA and miRNA influenced by SNPs were predicted using lncRNASNP2 (http://bioinfo.life.hust.edu.cn/lncRNASNP#!/ accessed on 15 August 2022) [24,28].
The SSBP phenotypes were defined continuously as MAP change 1 during the acute salt-loading process and as MAP change 2 during the diuresis shrinkage process in the MSAOSL-DST. We characterized the distributions of continuous variables according to the mean and standard deviation (SD) of the normal distribution variables or the median and interquartile range (IQR) for skewed distribution variables. For continuous phenotypes with a normal distribution, Student’s t-test was conducted to measure the differences between the two groups. The Wilcoxon rank–sum nonparametric test was utilized to test the continuous variables with non-normal distribution and rank variables. Chi-square (χ2) analysis was utilized to analyze Hardy–Weinberg equilibria (HWE). The effect of each SNP was calculated using additive models and allelic models. Generalized linear models were conducted to measure the associations of SNPs with lncRNAs and SSBP phenotypes which were estimated by β and 95% CI. A cumulative genetic risk score (c-GRS) was utilized to test the integrated effect of multiple lncRNA-SNPs on SSBP. The c-GRS was divided into quartiles. Multiple linear regression was conducted to analyze the association between c-GRS quartile groups and SSBP adjusted for potential confounders. To resolve multiple comparisons and control the false positives and false negatives, the false discovery rate (FDR) was used [25,29]. For lncRNA data, log2-transformation was performed. The lncRNA data used in this paper were all log2-transformed. The statistical analysis was carried out in R software (version 3.4.4). In mediation analysis, we considered each SNP which can influence lncRNA expression level to be the independent variable, SSBP phenotypes to be the outcome, and the corresponding lncRNA to be the mediator that may explain a portion of the SSBP risk. We utilized a two-step method using the R package “mediate” [26,30]. The model-based causal mediation test was measured in two steps. In step one, a mediator model and an outcome model were fitted. The mediator model was a linear regression of log2 (KCNQ1OT1) with the SNP, age, gender, and hypertension as the predictors. The outcome model was a linear regression model for SSBP phenotypes with the following covariates: SNP, log2 (KCNQ1OT1), SNP ∗ log2 (KCNQ1OT1) interaction term, age, gender, and hypertension. After the two models were fitted, the average causal mediation effect and average direct effect were computed through a general algorithm [30]. We used 1000 iterations and p ≤ 0.05 was viewed as nominally significant. With one-sample Mendelian randomization (MR) analyses, the causal association from genetically determined KCNQ1OT1 (relative expression unit) to SSBP (MAP change 1 or/and MAP change 2) was estimated by utilizing the instrumental variable analysis with two-stage least-squares regression (2SLS). The statistical analysis was conducted using the R package “AER” [31]. Quanto version 1.2.4 was used to calculate statistical power. Among our second-stage participants, the minimal MAF of the KCNQ1OT1 SNPs was 0.05. By calculating, we found that with our sample size, the power to find an OR of 1.5 was greater than 0.8.
This study included 1684 participants selected from the EpiSS study, and detailed information has been published in the previous paper [19]. Supplementary Table S1 shows the baseline characteristics for all individuals. In all subjects, the participants who were male, drinking alcohol, smokers, and those with hypertension had a positive correlation with SSBP risk (p < 0.05) and a higher LDL-C level, and had a negative correlation with SSBP risk (p < 0.05).
Our previous paper suggested the potential role of lnc-IGSF3-1:1, SCOC-AS1, and SLC8A1-AS1 as susceptible biomarkers for SS [15,17]. We further explored the effects of these lncRNAs on SSBP. In the first stage, we detected the relative expression of KCNQOT1, lnc-AGAP1-8:1, lnc-IGSF3-1:1, SCOC-AS1, and SLC8A1-AS1 by conducting qRT-PCR among 251 individuals, and analyzed their relationship with SSBP (MAP change 1 and MAP change 2). Our study discovered that three lncRNAs (KCNQ1OT1, lnc-AGAP1-8:1, and lnc-IGSF3-1:1) were significantly associated with SSBP (p < 0.05): the KCNQ1OT1 expression level was positively associated with MAP change 1 and was negatively associated with MAP change 2; the lnc-AGAP1-8:1 and the lnc-IGSF3-1:1 expression level was positively associated with MAP change 1. The results are shown in Figure 2.
Several studies have demonstrated that SNPs in lncRNA may affect the expression of lncRNA [28,32]. We assumed that the expression of SSBP-related lncRNAs was affected by genotypes of corresponding SNPs. A total of 13 candidate SNPs in three SSBP-related lncRNA (KCNQ1OT1, lnc-AGAP1-8:1, and lnc-IGSF3-1:1) genes were selected in this study. All SNPs were compatible with HWE (p > 0.05) and the MAFs of these SNPs ranged from 5 to 38%. Detailed information on these SNPs is shown in online Supplementary Table S2. Multilinear regression analysis showed that participants with the nine SNP minor alleles (rs10832417-T, rs3782064-A, rs7925578-G, rs11023840-T, rs71034996-T, rs58956504-C, rs11023582-A, rs2411884-C, and rs12577654-T) had a lower expression of KCNQ1OT1 before and after adjusting for age and gender, respectively (p < 0.05). There was no significant difference between lnc-AGAP1-8:1 and lnc-IGSF3-1:1 expression with different corresponding SNP alleles (rs71402704 g and rs995060-G). The results are shown in Figure 3 and Supplementary Table S3.
Previous research declared that lncRNA SNPs could affect the binding site efficiency for specific miRNAs and, therefore, impact the expression levels of lncRNA [29,30,33,34]. Thus, we assumed that the above nine positive SNPs could influence KCNQ1OT1 expression by changing the lncRNA secondary structure or microRNA-binding sites. MFE change, as well as local change in the structure located around the altered nucleotide, was shown in rs10832417 and rs33782064 (Figure 4). Bioinformatics analysis reported that rs10832417-T could decrease the binding efficiency of the hsa-miR-8068; rs33782064-A could increase the binding efficiency of the hsa-miR-6834, etc., and could decrease the binding efficiency of the hsa-miR-423-5p, etc. (Table 1).
The first-stage study showed that rs58956504-C significantly increased MAP change 2 (β = 2.766, p = 0.012) playing a protective role against SSBP using the multilinear regression model (Supplementary Table S4). However, there were no statistically significant results between other SNPs and SSBP (MAP change 1 and MAP change 2) (p > 0.05). In the second stage, we conducted the multilinear regression analysis on 1443 individuals. Four SNPs were found to be significantly associated with MAP change 2 (rs3782064, β = 0.762, FDR < 0.05; rs7925578, β = 0.610, FDR < 0.05; rs11023840, β = 0.779, FDR < 0.05; rs12577654, β = 0.653, FDR < 0.05). Two SNPs reached borderline significance (rs10832417-T, β = 0.547, FDR < 0.1; rs7103496, β = 0.676, FDR < 0.1). There was no statistically significant result in rs58956504 and rs11023582 after FDR correction (FDR > 0.1). The individuals in the third quartile of the c-GRS decreased the MAP change 1 by 0.710 mmHg compared with those in the lowest c-GRS quartile (p = 0.039), and the individual in the highest quartile increased the MAP change 2 by 0.420 mmHg compared with those in the lowest c-GRS quartile (p = 0.009). The results are shown in Table 2.
We further explored the mediation effect of KCNQ1OT1 expression on the association between the above six risk SNPs and SSBP in 251 participants. Mediation models were set up with the KCNQ1OT1 for SSBP as a mediator to detect the direct and indirect effects of the SNPs on SSBP, and the results are shown in Table 3. We found significant mediating effects in KCNQ1OT1 as a mediator from four SNPs (rs10832417, rs3782064, rs7103496, and rs12577654) to both MAP change 1 and MAP change 2, respectively (indirect effect p < 0.05; total effect p > 0.05). There were no significant mediating effects for rs7925578 and rs11023840 (indirect effect p > 0.05; total effect p > 0.05). When the indirect effect is significant but the total effect is not (and no suppression is present), one likely lacks the power to identify the total effect [31,35]. Increasing the sample size will clear the issue [32,36]. As expected, four SNPs (rs10832417, rs3782064, rs7103496, and rs12577654) were found to be significantly associated with SSBP in the second stage; the results are described above.
In observational analyses, a one-unit KCNQ1OT1 increase was associated with 4.147 mmHg in MAP change 1 (p < 0.001) and was associated with −2.829 mmHg in MAP change 2 (p = 0.023). Corresponding estimates in one-sample Mendelian randomization analyses were 5.581 mmHg in MAP change 1 (p = 0.020) and −3.464 mmHg in MAP change 2 (p = 0.014), respectively, by using seven SNPs (rs10832417, rs3782064, rs7103496, rs58956504, rs11023582, rs2411884, and rs12577654) as instruments. The results are shown in Figure 5.
Our study showed that three lncRNAs (KCNQ1OT1, lnc-AGAP1-8:1, and lnc-IGSF3-1:1) were significantly associated with SSBP and that participants with the nine SNP minor alleles (rs10832417-T, rs3782064-A, rs7925578-G, rs11023840-T, rs71034996-T, rs58956504-C, rs11023582-A, rs2411884-C, and rs12577654-T) had a lower expression of KCNQ1OT1. Four SNPs (rs10832417, rs3782064, rs7103496, and rs12577654) affected SSBP by modulating the KCNQ1OT1 expression. Among them, SNPs rs10832417 and rs3782064 in the KCNQ1OT1 gene might be associated with a low risk of SSBP through changing the KCNQ1OT1 secondary structure and miRNA binding, resulting in changes in KCNQ1OT1 expression. SSBP, as the intermediate phenotype for developing hypertension, plays a critical role in the occurrence of CVDs. The previous studies had studied SSBP from different levels, including genomics [2,3], metabolomics [33,37], and transcriptomics [15,17], to enlighten the pathogenic mechanism of SSBP. LncRNA deregulation plays an essential role in complex diseases such as CVDs [34,38]. Nearly 90% of the phenotype-associated SNPs discovered by GWAS are located beyond the protein-coding regions and map to the noncoding regions such as lncRNA. In this study, we demonstrated that three lncRNAs (KCNQ1OT1, lnc-AGAP1-8:1, and lnc-IGSF3-1:1) were significantly associated with SSBP. The results were consistent with our previous study which illuminated that lncRNAs, such as KCNQ1OT1, lnc-IGSF3-1:1, lnc-GNG10-3:1, SCOC-AS1, and SLC8A1-AS1, had an up-regulated expression in SS compared with SR [15,17]. LncRNA-SNP may change the expression or functioning of downstream target genes through several mechanisms. Zhang et al. showed that rs7130280 in the lncRNA NONHSAT159216.1 was associated with a low risk of Behcet’s disease and uveitis and influenced the interaction between lncRNA and its target genes [30,34]. Feng et al. suggested that rs140618127 in the lncRNA LOC146880 decreased ENO1 phosphorylation by increasing the binding efficiency for miR-539-5p [35,39]. Chaoqin Shen et al. revealed that rs1317082 at lncRNA CCSlnc362 decreased the susceptibility to CRC by creating a binding site for miR-4658 [36,40]. As lncRNAs do not code for a protein, their structure is considered to be important for their function. Some RNAs can yield strong structural variation upon SNP change [37,41]. So, we performed an eQTL analysis to figure out whether lncRNA-SNPs affect the expression of the genes in which they are located and LncRNA-SNP functioning mechanisms. The binding sites of lncRNA and miRNA influenced by SNPs were predicted using lncRNASNP2. RNA secondary structure prediction methods are established in thermodynamics and, usually, the MFE structure is determined. In our study, MFE structure predictions were performed using RNAfold [33,37]. We found that participants with the nine SNP minor alleles (rs10832417-T, rs7925578-G, rs3782064-A, rs11023840-T, rs71034996-T, rs58956504-C, rs11023582-A, rs2411884-C, and rs12577654-T) had a lower expression of KCNQ1OT1. Among them, rs10832417 and rs3782064 in KCNQ1OT1 could influence the secondary structure, miRNA binding, and relative expression of KCNQ1OT1 through association study and bioinformatic methods. Later, we hypothesized that SNP was related to SSBP through affecting KCNQ1OT1 expression. In the two-stage association study, we found that four SNPs were found to be significantly associated with MAP change 2 (rs7925578, rs3782064, rs11023840, and rs12577654). Two SNPs reached borderline significance (rs10832417 and rs7103496). Mediation analyses were conducted in the first stage. We found significant mediating effects in KCNQ1OT1 as a mediator from four SNPs (rs10932417, rs3782064, rs7103496, and rs12577654) to both MAP change 1 and MAP change 2. However, no significant associations were found in the total effects, which seems odd to explain, because either they had a suppression effect (indirect effect) which was not in our research, or they had other explanations. Kenny et al. demonstrated that the test of the indirect effect is more powerful than the test of the total effect [31,35]. As such, when total effects are not large effects, it is more likely to find indirect effects as significant than finding total effects. We conducted multilinear regression models on 1443 individuals in the second stage to detect the association (total effect) between the above SNPs and SSBP. The results have been discussed above. Generally, we found that 4 SNPs (rs10832417, rs3782064, rs7103496, and rs12577654) may affect SSBP by modulating the KCNQ1OT1 expression. It is worth noting that SNPs located in KCNQ1OT1 were related to MAP change 2, but not with MAP change 1, which indicated that SNP-induced changes in KCNQ1OT1 expression only affect SSBP during the diuresis shrinkage process, not during the acute salt-loading process. The GenSalt study identified rs10832417 in lncRNA KCNQ1OT1 had a protective effect on mean arterial pressure response to a high-sodium diet using the GWAS method [38]. We found that the rs10832417-T/G variant in an exon of KCNQ1OT1 was significantly associated with MAP change during the diuresis shrinkage process in the second stage, which was consistent with the GenSalt result [38,42]. We proceeded to explore the potential function of GWAS’s significant SNP rs10832417 through the eQTL analysis and bioinformatic prediction, which were previously described. Our results showed that rs10832417 and rs3782064 might be associated with a low risk of SSBP through influencing the KCNQ1OT1 secondary structure and miRNA binding, and resulting in changes in KCNQ1OT1 expression. Furthermore, one-sample Mendelian randomization analysis showed that KCNQ1OT1 may have a causal relationship with SSBP using the above eQTL SNPs as instruments. KCNQ1OT1 functioned at the epigenetic level, creating a positive effect on the formation of a repressive chromatin structure, and participates in the CVD process [39,43]. KCNQ1OT1 could play an essential mediator role in endothelial cell physiologic development, and endothelial dysfunction could affect the pathogenesis of SSBP [40,44]. The regulation between KCNQ1OT1, adiponectin receptors, and the p38 MAPK/NF-kB pathway was declared [41,45], which was involved in inflammation, leading to CVDs. KCNQ1OT1 acts with miR-183-3p to up-regulate CTNNB1 in vascular smooth muscle cells (VSMCs) and subsequently influences the proliferation and apoptosis of VSMCs [42,46]. Although our results reveal that rs10832417 and rs3782064 in lncRNA KCNQ1OT1 played important roles in SSBP susceptibility through changing KCNQ1OT1 expression, there are some limitations. First, the findings were concluded by association analysis and bioinformatic prediction. Our research group plans to conduct more experiments such as a luciferase assay to validate the results of our study in the future. Next, the findings of the current research were mainly reached using human peripheral blood specimens; therefore, further investigations in animal models are required. Third, because of the number of samples in lncRNA qRT-PCR, the statistical power of the lncRNA analysis was limited. Fourth, although there is no gold-standard method to determine SSBP, the chronic dietary salt-loading protocol is relatively more accurate and stable than the MSAOSL-DST. Finally, our research was limited to Han Chinese, and our findings need to be validated in different populations. Generally, our present study elucidated that three lncRNAs (KCNQ1OT1, lnc-AGAP1-8:1, and lnc-IGSF3-1:1) were significantly associated with SSBP. Nine SNPs’ minor alleles (rs10832417-T, rs7925578-G, rs3782064-A, rs11023840-T, rs71034996-T, rs58956504-C, rs11023582-A, rs2411884-C, and rs12577654-T) had a lower expression of KCNQ1OT1 compared with their major alleles. SNPs rs10832417 and rs3782064 in KCNQ1OT1 were negatively associated with the susceptibility of SSBP, which might influence the KCNQ1OT1 secondary structure and miRNA binding, and result in changes in KCNQ1OT1 expression. These data highlighted a potential relationship between gene variation and lncRNAs, potentially contributing to a pathological outcome, which would be a promising pathogenic mechanism and therapeutic target for SSBP. Further functional molecular experiments of the genetic variant would be of great interest.
The SNPs rs10832417 and rs3782064 in KCNQ1OT1 were negatively associated with the susceptibility of SSBP, which might function through influencing the KCNQ1OT1 secondary structure and miRNA binding using bioinformatic predictions, resulting in changes in KCNQ1OT1 expression. | true | true | true |
PMC9571561 | 36233452 | Yaling Zhang,Dejian Chen,Daojuan Wang,Lei Wang,Yajing Weng,Hongwei Wang,Xiaoke Wu,Yong Wang | Moderate Aerobic Exercise Regulates Follicular Dysfunction by Initiating Brain-Derived Neurotrophic Factor (BDNF)-Mediated Anti-Apoptotic Signaling Pathways in Polycystic Ovary Syndrome | 23-09-2022 | aerobic exercise,brain-derived neurotrophic factor,apoptosis,neuroendocrine,polycystic ovary syndrome | Polycystic ovary syndrome (PCOS) is a common endocrine disorder among women. Moderate aerobic exercise intervention is considered an initial treatment strategy for managing PCOS. Brain-derived neurotrophic factor (BDNF) is an important molecular mediator and a beneficial response to exercise. We aimed to investigate the expression pattern and underlying molecular mechanisms of this neurotrophic factor during follicle development in ovarian tissues. The PCOS model was established by subcutaneous injection of 60 mg/kg dehydroepiandrosterone (DHEA) into the neck of Sprague Dawley rats for 35 consecutive days. PCOS rats then received aerobic exercise for 8 weeks. Body/ovarian weight and peripheral serum hormone levels were observed. Immunohistochemistry combined with Western blot analysis and fluorescence quantitative polymerase chain reaction were used to detect the changes in BDNF-TrkB/p75NTR pathway, apoptosis, and inflammatory factors. We show that moderate aerobic exercise not only reverses the PCOS phenotype but also activates the BDNF-TrkB pathway and initiates downstream targets. p-TrkB upregulates and phosphorylates phosphatidylinositol 3-kinase (PI3K) and protein kinase B (Akt) to inhibit apoptosis. In addition, aerobic exercise therapy reduces the high expression of p75NTR in the ovarian tissue of PCOS rats and initiates the anti-apoptotic effect from the downstream pathway of NF-κB/JNK. Our in vitro results state that treatment with BDNF ameliorated dihydrotestosterone (DHT)-induced granulosa cells (GCs) apoptosis by provoking p-TrkB activation and upregulating the PI3K/AKT pathway. The present study suggests that moderate aerobic exercise regulates follicular dysfunction in PCOS-like rats. One possible mechanism is to initiate the BDNF-mediated anti-apoptotic signaling pathway. | Moderate Aerobic Exercise Regulates Follicular Dysfunction by Initiating Brain-Derived Neurotrophic Factor (BDNF)-Mediated Anti-Apoptotic Signaling Pathways in Polycystic Ovary Syndrome
Polycystic ovary syndrome (PCOS) is a common endocrine disorder among women. Moderate aerobic exercise intervention is considered an initial treatment strategy for managing PCOS. Brain-derived neurotrophic factor (BDNF) is an important molecular mediator and a beneficial response to exercise. We aimed to investigate the expression pattern and underlying molecular mechanisms of this neurotrophic factor during follicle development in ovarian tissues. The PCOS model was established by subcutaneous injection of 60 mg/kg dehydroepiandrosterone (DHEA) into the neck of Sprague Dawley rats for 35 consecutive days. PCOS rats then received aerobic exercise for 8 weeks. Body/ovarian weight and peripheral serum hormone levels were observed. Immunohistochemistry combined with Western blot analysis and fluorescence quantitative polymerase chain reaction were used to detect the changes in BDNF-TrkB/p75NTR pathway, apoptosis, and inflammatory factors. We show that moderate aerobic exercise not only reverses the PCOS phenotype but also activates the BDNF-TrkB pathway and initiates downstream targets. p-TrkB upregulates and phosphorylates phosphatidylinositol 3-kinase (PI3K) and protein kinase B (Akt) to inhibit apoptosis. In addition, aerobic exercise therapy reduces the high expression of p75NTR in the ovarian tissue of PCOS rats and initiates the anti-apoptotic effect from the downstream pathway of NF-κB/JNK. Our in vitro results state that treatment with BDNF ameliorated dihydrotestosterone (DHT)-induced granulosa cells (GCs) apoptosis by provoking p-TrkB activation and upregulating the PI3K/AKT pathway. The present study suggests that moderate aerobic exercise regulates follicular dysfunction in PCOS-like rats. One possible mechanism is to initiate the BDNF-mediated anti-apoptotic signaling pathway.
Polycystic ovary syndrome (PCOS) is a common endocrine disorder among women, mainly characterized by hyperandrogenemia, hyperinsulinemia, and chronic anovulation [1,2]. Because of the complex pathogenesis of PCOS, no mechanism-based treatments have been presented against it. However, previous studies support that elevated androgen levels contribute to PCOS [3]. Previous studies show that hyperandrogenism has induced oxidative stress (OS), ovarian fibrosis, chronic low-grade inflammation, mitochondrial dysfunction, and excessive endoplasmic reticulum (ER) stress in the ovary that ultimately affected follicle development [4,5,6,7,8]. Recent evidence has revealed that the pathophysiological mechanisms of PCOS are associated with neuroendocrine impairments [9,10,11]. Thus, increased androgen signaling in the brain and ovary may be a potential mechanism in the pathophysiology of PCOS. Regular physical activity can prevent or reduce the risk of many diseases [12]. Moderate aerobic exercise is the preferred treatment for PCOS. Exercise can reduce androgen and insulin levels in PCOS, improve the endocrine environment of the ovary, and play an important role in promoting ovulation [13,14]. Our previous research has shown that aerobic exercise can alleviate hyperandrogenism-induced ER stress and reverse ovarian granulosa cells (GCs) apoptosis in PCOS-like rats [15]. In addition, exercise stimulates the release of neurotransmitters and neurotrophins such as nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3, and neurotrophin-4/5 in an activity-dependent manner [16,17]. They play an important role in regulating neuronal survival and differentiation and in non-neuronal tissues such as ovarian follicles [18]. Some studies have shown that endometriosis is associated with low follicular-fluid BDNF levels, and diminished ovarian reserve is associated with increased follicular-fluid NGF levels [19]. In vitro treatment with an appropriate concentration of BDNF can promote oocyte maturation and embryonic development [20]. In addition, BDNF affects the development of the materno-fetal-placental unit in terms of differentiation, proliferation, and placental nutrient transport [21]. However, it has also been reported in some studies that the BDNF levels of PCOS patients in plasma and in follicular fluid were higher than values obtained in healthy controls [22]. Subsequent findings suggest that chronic low-dose inflammation in PCOS may interact with BDNF to contribute to the development of depression [23]. In conclusion, BDNF signaling events have substantial roles in the ovary, and BDNF expression and levels have been linked with follicle organization during ovarian development, follicle recruitment, and growth and oocyte maturation [24]. However, BDNF expression patterns in non-neuronal tissues, such as the ovary, and the underlying mechanism of oocyte maturation remain unclear. BDNF binds to the high-affinity tropomyosin-related kinase receptor type B (TrkB) and exerts its pro-survival effects by activating the downstream signaling pathways, including the phosphatidylinositol 3-kinase (PI3K)-Akt pathway [25]. BDNF can also activate the pan-neurotrophin low-affinity co-receptor p75 (p75NTR) [26,27,28]. In many species, BDNF can affect oocyte maturation and early embryo development [29]. However, the role of BDNF and its receptor in PCOS remains to be determined. In addition, the role of BDNF-TrkB/p75NTR signaling in follicular development is limited and incompletely understood. In particular, exercise has been shown to promote the expression of BDNF in hippocampus and cerebral cortex [30]. Whether exercise can directly affect the expression of BDNF and its receptors in ovarian tissues remains to be further explored. In this study, we explored the potential mechanism by which moderate aerobic exercise may modulate follicular dysfunction by increasing the expression of BDNF in PCOS. As our understanding of the expression and underlying molecular mechanisms of these neurotrophic factors in the human ovary grows, new diagnostic and therapeutic applications for the management of patients with infertility and ovarian pathology, as well as improvement in oocyte quality, will be developed.
The Shanghai Xipuer-Bikai Laboratory Animal Co., Ltd. (Shanghai, China) provided 60 specific pathogen-free (SPF) Sprague Dawley® (SD) female rats (21 days old, 50–60 g). All animal experiments were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) and approved by the Institutional Research Animal Committee of Nanjing University. The rats were housed in closed cages under controlled temperature (22 ± 2 °C) and light (12:12 h, light:dark). All SD rats were initially divided into three groups: the control group (no treatment, n = 10), model group (DHEA treatment, n = 10), and exercise group (DHEA treatment + treadmill training treatment, n = 10). The PCOS rats were induced with dehydroepiandrosterone (DHEA) (6 mg/100 g BW) for 35 consecutive days.
After the PCOS model was established by DHEA subcutaneous injection, treadmill aerobic training was conducted at 10:00 a.m. the next day. In the first week, adaptive treadmill training was performed. Rats were trained at 5 m/min–10 m/min for 10 min on the first day, and training length increased by 30 min per day, ending at up to 15 m/min for 60 min on the sixth day. Then, seven-week regular exercise training was performed at 15 m/min for 60 min per day (5 m/min for the first 5 min, 10 m/min for 10 min, and 15 m/min for the remaining 45 min), six days per week for eight weeks. Furthermore, at the end of the 8-weeks treatment period, blood samples and tissues were collected immediately.
In order to enhance multiple follicular development, 23-day-old immature female rats were injected with pregnant mare serum gonadotropin (PMSG) (20 IU) 48 h in advance. The rats were killed by cervical dislocation and were disinfected with 75% alcohol for 20 min. The ovaries were quickly removed on a super-clean bench, and the follicles were punctured with microscopic tweezers to release GCs. Cell debris was removed using 100 μm cell strainers. The primary GCs was cultured in DMEM-F12 containing 10% fetal bovine serum (FBS, Gibco, New York, NY, USA) and 1% penicillin-streptomycin solution (Gibco, New York, NY, USA), in a cell incubator at 37 °C (95% relative humidity, 5% CO2). Cell growth was monitored intermittently.
GCs were seeded in 96-well plates (1 × 105 cells/well) for 48 h and then treated with various concentrations of dihydrotestosterone (DHT) and BDNF (Cayman Chemical, Ann Arbor, MI, USA) for the indicated times. Cell counting kit-8 (CCK-8) solution (10 μL; A311-02-AA, Vazyme Biotech, Nanjing, China) was added to each well, followed by incubation for 4 h at 37 °C. The absorbance at 450 nm was measured using a microplate reader.
Cell apoptosis was assessed using an Alexa Fluor 488-conjugated annexin V and propidium iodide (PI) (C1062L, Beyotime, Shanghai, China) detection kit. Primary GCs were treated with or without DHT, BDNF, and ANA-12, and then, 5 μL of PI and 10 μL of annexin V–FITC was added to the cells. FITC–Annexin V-positive cells were considered apoptotic cells.
Paraffin slides were stained with hematoxylin and eosin to examine the pathological structural alterations of the rat ovary and hippocampus under an optical microscope (Leica Microsystems, Weztlar, Germany).
The rats were anesthetized with 1% pentobarbital sodium (40 mg/kg, ip), and blood was drawn from the superior vena cava. The serum was separated immediately and stored at −80 °C for further determination of testosterone (T), luteinizing hormone (LH), and follicle-stimulating hormone (FSH) levels through enzyme-linked immunosorbent assay (ELISA) (Elabscience Biotechnology, Wuhan, China).
Nissl staining was performed to assess neuronal survival. The sections were rinsed in deionized water, dipped in a warm (50 °C) solution of 1% thionine for 45 min, and differentiated with 70% alcohol for approximately 5 min.
Serum was collected for detecting IL-1β using ELISA kits (Elabscience Biotechnology). The ELISA provided a detection range ranging from 31.25 to 2000 pg/mL, and the sensitivity of the ELISA kit was 18.75 pg/mL. According to the manufacturer’s instructions, 450 nm was regarded as the most suitable wavelength for measuring absorbance.
After antigen retrieval, dewaxed and rehydrated sections were treated with 3% hydrogen peroxide and then 5% BSA and incubated overnight at 4 °C with primary antibody BDNF (1:100, Wanleibio, Shenyang, China). Slides were then incubated with a secondary goat anti-rabbit IgG horseradish peroxidase (HRP) at 37 °C for 30 min. Sections were consequently stained with diaminobenzidine for 10 min, counterstained with hematoxylin (Beyotime, Shanghai, China), covered with coverslips, and observed under an optical microscope. The histology was quantified with Image Pro Plus 6.0 based on optical density. The tissue steps and cell climbing pieces before the sections were incubated with the primary antibody were the same for immunohistochemical staining. Sections were incubated overnight at 4 °C with antibodies against BDNF (1:100, Wanleibio), TrkB (1:100, Wanleibio), p75NTR (1:100, 55014-1-AP, Proteintech, Chicago, IL, USA), and Cleaved-Caspase-3 (1:100, 66470-2-Ig, Proteintech) at a 1:100 dilution. Fluorescently labeled secondary antibodies were diluted (1:1000) and incubated in the dark at 25 °C for 2 h. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (C1002, Beyotime) at a dilution of 1:2000 for 30 min. Images were photographed using an Olympus laser scanning confocal microscope (FV3000, Tokyo, Japan). Fluorescence intensity was quantified using Image-Pro Plus 6.0 (Media Cybernetics, Rockville, MD, USA).
TUNEL assays were performed to detect Bcl-2 and caspase 3 levels in ovary sections. A Fluorescein (FITC) TUNEL Cell Apoptosis Detection Kit (G1501-100T, Servicebio, Wuhan, China) was used in accordance with the manufacturer’s instructions. Slides treated with DNase I for 30 min served as positive controls. DAPI was used to stain the nuclei.
The total RNA was extracted from the GCs in the ovaries with TRIzol reagent (Beyotime), and the cDNA was synthesized with a reverse transcription kit (Vazyme, China). Quantitative RT-PCR was performed with the ABI Viia7 real-time PCR system (ABI, Los Angeles, CA, USA) by using the SYBR™ Green PCR Master Mix (Vazyme). Quantitative RT-PCR was performed as follows: Stage 1, pre-denaturation (Rep: 1, 95 °C, 30 s); Stage 2, circular reaction (Reps: 40, 95 °C, 10 s; 60 °C, 30 s); Stage 3, melting curve (Rep: 1, 95 °C, 10 s; 60 °C, 60 s; 95 °C, 15 s). The primers used in this study are shown in Table 1. The critical threshold cycle (Ct) value was determined for each reaction for relative quantification using the 2−∆∆Ct method. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an internal control.
The tissue samples were homogenized through mechanical disruption in RIPA lysis buffer (P0013B, Beyotime) containing 1 mM PierceTM phosphatase inhibitor (B15001, Selleck, TX, USA) and 0.1% HaltTM protease inhibitor cocktail (B14001, Selleck). The gel electrophoresis system of Bio-Rad (12% SDS polyacrylamide) was used to separate proteins from samples that contained the same protein quantity (30 µg); the proteins were then transferred onto the polyvinylidene difluoride membranes (IPVH00010, Merck Millipore, Burlington, MA, USA). Target bands were incubated at 4 °C overnight with corresponding primary antibodies against BDNF (1:500, WL0168, Wanleibio, Wuhan, China), TrkB (1:500, WL00839, Wanleibio), p-TrkB (1:500, WL02988, Wanleibio), p75NTR (1:500, 55014-1-AP, Proteintech, Chicago, IL, USA), PI3K (1:1000, 4292S, CST, Boston, MA, USA), p-PI3K (1:1000, AF3242, Affinity, Boston, MA, USA), AKT (1:1000, 4691T, CST), p-AKT (1:1000, 4060T, CST), NF-κB (1:1000, 8242S, CST), p-NF-κB (1:1000, 3033S, CST), JNK (1:1000, 9252T, CST), p-JNK (1:1000, 4668S, CST), IL-1β (1:1000, ab9722, Abcam, Cambridge, UK), IL-6 (1:1000, 21865-1-AP, Proteintech), Bcl-2 (1:500, Wanleibio), Bax (1:1000, 50599-2-Ig, Proteintech), cleaved caspase-3 (1:1000, 66470-2-Ig), AR (1:1000, ab52615, Abcam), FSHR (1:1000, BS5724, Bioworld, Nanjing, China), Cyp11α1 (1:1000, BS6578, Bioworld), Cyp19α1 (1:1000, BS6580, Bioworld), and GAPDH (1:5000, Bioworld), followed by the addition of HRP-labeled secondary antibodies. The blots were visualized using chemiluminescent detection (Merck Millipore, NJ, USA). Densitometric analysis was performed with Image J.
Values were expressed as the mean ± SEM. All statistical analyses were performed with GraphPad (Prism 7.0). Multiple comparisons were implemented through one- and two-way ANOVA followed by Tukey’s post-hoc test. Binary variables were compared using a t-test, and a p-value < 0.05 was regarded as statistically significant.
In this study, the bodyweight of rats induced with DHEA was higher than that in the control group, but after 8 weeks of aerobic (treadmill) exercise, their weight was significantly reduced (Figure 1A). The ovarian weight of PCOS rats also markedly increased after the exercise treatment (Figure 1B). In addition, compared to the control group, the ovaries in PCOS rats had more atypical follicles but almost no corpus luteum. However, multiple immature follicles were substantially reduced after exercise treatment, whereas the number of corpus luteum evidently increased (Figure 1C). Furthermore, the elevated T and LH/FSH levels in PCOS rats declined to normal levels after the exercise treatment (Figure 1D,E). The estrus cycle of all rats was continuously observed for two cycles (10 days). The results show that the rats in the PCOS group lost their regular estrus cycle. After exercise intervention, the estrus cycle returned to normal (Figure 1F). Notable changes were also observed in genes related to follicular development after 8 weeks. AR, Cyp11α1, and Cyp19α1 levels increased, while FSHR levels decreased in the ovaries of PCOS rats (Figure 2A,I). In addition, fasting blood sugar (FBG) was measured using a glucometer, and compared with the control group, the FBG levels in the PCOS group were increased. Relative to the PCOS group, the PCOS + exercise group had decreased FBG levels. However, there was no statistical difference (Table 2).
The effects of exercise in the brain are most apparent in the hippocampus [31], and thus, H&E and Nissl staining were conducted to explore the effects of aerobic treadmill exercise in the hippocampus of DHEA-induced PCOS rats. Compared with the control group, Nissl stains in the hippocampal CA1, CA3, and DG regions of PCOS rats were lighter. After treadmill running, neurons were orderly and densely arranged in the hippocampal regions, and the number of neurons in hippocampal DG and CA3 areas was notably enhanced, similar to the hippocampal regions of the control group (Figure S1A,B). Analysis of immunofluorescence data showed that the BDNF level in the PCOS group was significantly lower than that in the other groups. A significant rise in BDNF in the PCOS group was observed after 8 weeks of exercising (Figure S1C,D). Western blot analysis showed that the expression of BDNF, TrkB, and p-TrkB proteins decreased and p75NTR proteins increased in the hippocampus of the PCOS group compared to the control group. After the treadmill exercise, the above phenomenon was corrected (Figure S1E,F).
BDNF and its receptors are thought to be key regulatory proteins in the development of the ovary and ovarian regulation in a follicle-stage-dependent manner [32,33]. To detect the location and expression of BDNF, TrkB, and p75NTR in the ovary, immunohistochemistry and double-labeling immunofluorescence were performed. BDNF could be detected in granulosa and membranous cell layers in the ovarian follicle. However, the expression levels of BDNF and TrkB in the PCOS group were lower than those in the control group, but the expression levels of p75NTR were higher. After 8 weeks of aerobic exercise, the PCOS group had substantially enhanced BDNF and TrkB expression, but with decreased levels of p75NTR (Figure 3A–E).
The expression levels of p75NTR, TrkB, and BDNF can be modulated with response to stimuli through lowering p75NTR or ameliorating BDNF-TrkB signaling [28]. To further confirm that aerobic exercise could activate BDNF signaling in the ovary of DHEA-induced PCOS rats, we analyzed the mRNA and protein expression levels of BDNF, TrkB, and p75NTR in the ovarian tissue from different groups. Based on RT-qPCR and Western blot analysis, the expression level of BDNF in the PCOS group was lower than that in the other groups. Moreover, the expression pattern of TrkB was similar to BDNF, but that of p75NTR was higher in the PCOS group after 8 weeks of aerobic exercise. BDNF expression was substantially enhanced, accompanied by the upregulated phosphorylation of TrkB (p-TrkB) (Figure 4A–C). In addition, the level of p75NTR significantly decreased in rats after exercising (Figure 4F,H). TrkB and p75NTR are known to engage in disparate signaling pathways downstream of ligand activation. TrkB promotes PI3K/AKT, while p75NTR stimulates NF-κB and JNK pathways. Consistently, Western blot analysis confirmed that PI3K and p-AKT markedly decreased in the PCOS group after 8 weeks of aerobic exercise, and the expression levels of PI3K and p-AKT in the ovarian tissue of rats were significantly upregulated (Figure 4D,E). Furthermore, both NF-κB and JNK pathways were activated in DHEA-induced rats. Treatment with aerobic exercise significantly reduced the expression of phosphorylated protein levels in NF-κB (p-NF-κB) and JNK (p-JNK) pathways (Figure 4G,I).
A TUNEL assay was performed to examine the apoptosis of ovarian tissues in each group. The number of TUNEL-positive cells was significantly higher in the PCOS group than in the control group, but the PCOS + exercise group had fewer TUNEL-positive cells than the PCOS group. In addition, double immunofluorescence staining showed that aerobic exercise increased the expression of anti-apoptotic protein Bcl-2 and decreased the level of cleaved caspase-3 in ovarian tissue of PCOS rats (Figure 5A,D). Consistently, Western blot results demonstrate that hyperandrogenism significantly increased levels of cleaved caspase-3 and Bax and decreased the anti-apoptotic protein Bcl-2 in ovarian tissues compared to the control group. However, levels of cleaved caspase-3 and Bax in ovarian tissues of rats treated with aerobic exercise were remarkably reduced, while the level of Bcl-2 increased (Figure 5E,F). We observed greater production of IL-1β in serum samples of each group. The concentration of serum IL-1β (pg/mL) in the PCOS group was greater than that in the control group. After exercise therapy, the serum IL-1β level was decreased significantly in the PCOS + exercise group (Figure 5G). Moreover, the protein levels of inflammatory cytokines (IL-10, IL-6, IL-1β) in ovarian tissue significantly decreased after aerobic exercise therapy (Figure 5H,I). Our findings indicate that aerobic exercise therapy is capable of suppressing DHEA-induced apoptosis and inflammation in the PCOS model.
Our previous research showed that hyperandrogenism can induce ovarian GCs pyroptosis [7]. Our data showed that BDNF in PCOS rats increased significantly after exercise. Thus, GCs were treated with 5 μM DHT and various BDNF concentrations (0, 10, 100, 1000 ng/mL) for 48 h, and GCs viability after BDNF treatment was analyzed using CCK-8. As expected, the cell viability was rescued after BDNF treatment (Figure S2A,B). To test this hypothesis of BDNF/TrkB signaling, we used the inhibitor for TrkB receptor ANA-12. Results from immunofluorescence and Western blotting assays show that the provoking action of BDNF was abolished in the presence of ANA-12 with diminished expression of p-TrkB compared with DHT + BDNF group (Figure S2C–F). To further confirm the anti-apoptotic effect of BDNF is dependent on the activation of the PI3K/AKT pathway, we used PI3K inhibitor LY294002 in in vitro studies. The apoptosis rate of GCs was also analyzed by PI and FITC-Annexin V staining. As expected, BDNF treatment reversed DHT-induced apoptosis of GCs, and PI3K inhibitor LY294002 eliminated the anti-apoptotic effect of BDNF (Figure 6A). Moreover, BDNF treatment exerted an anti-apoptotic effect on DHT-induced GCs, and cleaved caspase-3 level was decreased using immunofluorescence assay (Figure 6B,C). In addition, Western blotting analysis showed that BDNF treatment caused increased expression of p-PI3K and p-AKT, but LY294002 inhibits their expression in GCs (Figure 6D,E).
Follicular dysplasia in PCOS patients may be caused by abnormal endocrine and paracrine factors and changes in follicular microenvironment. Recently, a growing number of studies have shown that neurotrophins and their receptors are also expressed throughout the reproductive system [34]. BDNF is a neurotrophic protein first discovered in the pig brain by scientists in 1982. It is mainly expressed in the central nervous system, with the highest content in the hippocampus and cortex. Studies on the role of BDNF and its receptor in the ovaries of humans and other mammals are increasing. It is inferred that BDNF may be a physiological regulator promoting follicle development, granulosa cell proliferation, and oocyte maturation [24,35]. BDNF is typically synthesized as a large precursor protein (pro-BDNF), which exhibits exclusive binding to p75NTR. Pro-BDNF can be cleaved to form mature BDNF (m-BDNF), and both are biologically active. Mature BDNF signals through its high-affinity receptor TrkB [36,37]. However, the interactions between BDNF-TrkB and pro-BDNF-p75NTR are complex and can be modified at various levels (alternative forms, alternative receptors/signaling pathways). Therefore, the underlying mechanism is not fully understood and should be further explored. Exercise can enhance the expression of BDNF. Exercise plays an important role in the prevention and treatment of PCOS as a non-drug intervention method. In addition, exercise requires high intensity, duration, and muscle group allocation. Moderate exercises are beneficial for most women and could improve fertility for those with anovulatory disorders such as PCOS [38,39]. There was also a study showing that acute high-intensity intermittent exercise increased serum BDNF concentration and attenuated the emotional states of tension, depression, and anger [40]. The findings suggested exercise as a strategy to attenuate the deleterious sensations occasioned by ovarian hormonal fluctuations and regulate ovarian function. In this study, we first established that DHEA is a major androgen precursor. Hyperandrogenism-induced PCOS rats showed weight gain, ovarian weight loss and polycystic changes, and hormone disturbance. When the rats were subjected to treadmill exercise (6 d/wk, 1 h/d at a pace of 15/min for 8 week), PCOS rats lost body weight, ovarian morphology returned to normal, and hormone disturbance was corrected. The intensity of the exercise was based on References [41,42]. This is consistent with previous studies that showed moderate exercise may help to recover optimal hormonal balance and restore ovulation in overweight and obese women with PCOS [43]. Moderate aerobic exercise can improve follicular dysfunction in PCOS rats, which may be associated with the increased expression of BDNF. Interestingly, we found that BDNF and its receptors TrkB and p75NTR were not only expressed throughout ovarian follicle development but also in the hippocampus after exercise. BDNF is inferred to cross the blood–brain barrier in a bidirectional manner [44,45]. Moreover, the level of circulating BDNF further increased, but in addition to the brain, BDNF was also released in the skeletal muscles, PBMCs, vascular endothelial cells, and platelets [46,47]. Nevertheless, we considered that our study mainly focused on the changes in the PCOS model rats after they received exercise training, rather than explaining the effect of exercise. Although the experimental purpose has been reached, due to the lack of the control + exercise group, the comparison between the changes in the control group after exercise and the changes in the model group after exercise is still inconclusive, which is also the limitation of this study. In addition, the understanding of BDNF is still in its infancy, and further exploration of the candidate sources and release mechanisms of BDNF in exercise is needed. In addition to its well-established neurotrophic action, BDNF also possesses anti-apoptosis, anti-oxidation, and autophagy-suppressing qualities [48,49]. Previous studies have shown that neurotrophic factors directly regulate the function of somatic and granulosa cells in the ovary by binding to corresponding receptors rather than regulating the development of primary to secondary follicles through the ovarian nerves [50,51]. Moreover, studies have shown that BDNF in the ovary is secreted by cumulus and granulosa cells; therefore, the concentration of BDNF must be related to the proliferation of follicular granulosa cells. The BDNF level decreased when granulosa cells were apoptotic [52], and in this study, DHEA-induced PCOS rats showed increased expression of apoptotic proteins (cleaved caspase-3/Bax) and decreased expression of anti-apoptotic B-cell lymphoma 2 (Bcl-2) in the ovary. The results are consistent with previous findings that the beneficial effects of BDNF involve the induction of anti-oxidative thioredoxin, with the resultant expression of Bcl-2. In addition, the number of TUNEL-positive cells was significantly higher in the ovarian tissue of PCOS rats. After 8 weeks of aerobic exercise therapy, the PCOS phenotype was reversed, the expression of apoptotic proteins was downregulated, and the anti-apoptotic protein was upregulated in the ovary. We infer that aerobic exercise increases BDNF, plays an anti-apoptotic role, and corrects ovarian dysfunction. We analyzed the mRNA and protein expression level of BDNF, TrkB, and p75NTR in ovarian tissues and found that the activation of the BDNF-TrkB pathway initiated downstream targets. p-TrkB upregulates and phosphorylates phosphatidylinositol 3-kinase (PI3K) and Akt to inhibit apoptosis. Moderate aerobic exercise therapy may also reduce the high expression of p75NTR in the ovarian tissue of PCOS rats, initiate the anti-apoptotic effect mediated by the downstream pathway of NF-κB/JNK, and thus reduce apoptosis of ovarian oocytes and granulosa cells. In addition, the protein levels of inflammatory cytokines (IL-6 and IL-1β) in ovarian tissue of PCOS rats significantly decreased after aerobic exercise therapy. However, it remains completely unclear whether aerobic exercise benefits BDNF in the process of exerting anti-inflammatory effects. These findings suggest that BDNF could be a candidate neurotrophic factor capable of improving follicular dysfunction. In conclusion, we show that low levels of BDNF in the ovarian follicle of PCOS rats may be the potential cause of follicular development disorders. Moderate aerobic exercise can enhance the expression of BDNF and initiate the BDNF-mediated anti-apoptotic signaling pathway, thus reversing the ovarian phenotype of PCOS. Exercise plays an important role in the prevention and treatment of PCOS as a non-drug intervention method. However, the limitations of this study included that the exercise protocol was relatively simple, as the rats were only trained on a treadmill and not engaged in other types of aerobic exercise, such as swimming or voluntary wheel running. There is also a lack of studies evaluating exercise in this study. In addition, the temporal and spatial differences in the expression of BDNF and its receptor during ovarian and follicular development suggest that the physiological functions of the BDNF signaling pathway vary in follicles at different developmental stages. Presently, the mechanism for signal transduction of BDNF and its receptors in oocytes, follicular granulosa, and stromal cells is not completely clear. Thus, the expression of BDNF in follicles at various stages, whether as pathogenesis of PCOS or just a biochemical indicator, in the pathophysiological process of PCOS, need to be further studied. From a clinical point of view, due to the high prevalence of PCOS in reproductive-aged women, it is important that future studies about the effectiveness of lifestyle interventions in this PCOS patient population are robustly designed to better provide new ideas for future clinical practice guidelines/recommendations. | true | true | true |
PMC9572001 | 36235459 | Ling Lei,Dan Wu,Chao Cui,Xiang Gao,Yanjie Yao,Jian Dong,Liangsheng Xu,Mingming Yang | Transcriptome Analysis of Early Senescence in the Post-Anthesis Flag Leaf of Wheat (Triticum aestivum L.) | 01-10-2022 | wheat,transcriptome,early senescence,flag leaf,post-anthesis | Flag leaf senescence is an important determinant of wheat yield, as leaf senescence occurs in a coordinated manner during grain filling. However, the biological process of early senescence of flag leaves post-anthesis is not clear. In this study, early senescence in wheat was investigated using a high-throughput RNA sequencing technique. A total of 4887 differentially expressed genes (DEGs) were identified, and any showing drastic expression changes were then linked to particular biological processes. A hierarchical cluster analysis implied potential relationships between NAC genes and post-anthesis senescence in the flag leaf. In addition, a large set of genes associated with the synthesis; transport; and signaling of multiple phytohormones (JA, ABA, IAA, ET, SA, BR, and CTK) were expressed differentially, and many DEGs related to ABA and IAA were identified. Our results provide insight into the molecular processes taking place during the early senescence of flag leaves, which may provide useful information in improving wheat yield in the future. | Transcriptome Analysis of Early Senescence in the Post-Anthesis Flag Leaf of Wheat (Triticum aestivum L.)
Flag leaf senescence is an important determinant of wheat yield, as leaf senescence occurs in a coordinated manner during grain filling. However, the biological process of early senescence of flag leaves post-anthesis is not clear. In this study, early senescence in wheat was investigated using a high-throughput RNA sequencing technique. A total of 4887 differentially expressed genes (DEGs) were identified, and any showing drastic expression changes were then linked to particular biological processes. A hierarchical cluster analysis implied potential relationships between NAC genes and post-anthesis senescence in the flag leaf. In addition, a large set of genes associated with the synthesis; transport; and signaling of multiple phytohormones (JA, ABA, IAA, ET, SA, BR, and CTK) were expressed differentially, and many DEGs related to ABA and IAA were identified. Our results provide insight into the molecular processes taking place during the early senescence of flag leaves, which may provide useful information in improving wheat yield in the future.
Wheat (Triticum aestivum L.) is a worldwide staple food crop. It is grown on about 200 million hectares across diverse global environments [1], and it is consumed by 2.5 billion people in 89 countries, accounting for 20% of the global caloric intake (Centro Internacional de Mejoramientode Maizy Trigo (CIMMYT)). CIMMYT has predicted that 70% more wheat will be required by 2050 to continue supplying a similar proportion of the global population. To achieve this goal, many challenges from physiological, agronomic, socioeconomic, environmental, and managerial aspects need to be overcome. The grain yield of wheat is determined by the successful formation, partitioning, translocation, and accumulation of assimilates during grain filling. Assimilates mainly come from two sources [2]. First is the accumulation of assimilates formed by photosynthesis; approximately 70–90% of the overall grain yield is supplied by photoassimilates [3]. Leaves, being the major site of photosynthetic activity, play an important role in determining the grain yield of wheat. Flag leaves, compared with other leaves, contribute the most photosynthates (41–43%) [4]. It has been reported that about 60% of grain saccharides are derived from the photosynthates of the flag leaf [5]. Second, the remobilization of pre-anthesis assimilates stored in vegetative organs. With the senescence of vegetative organs, the structure, metabolism, and gene expression of cells have undergone a large number of orderly changes, accompanied by the transport of the degradation products of proteins, lipids, and nucleic acids to the reservoir tissue. The disintegration of the photosynthetic apparatus is a major event during senescence, mainly as ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) [6]. With the hydrolysis of Rubisco, a large amount of free amino acids will be produced and transported to the grain [7]. Photosynthetic organelle chloroplasts are the main organelles for iron storage (80%), and a large number of Fe released after the disintegration of the organelle will be reused by the grains, so that the Fe content in the leaves decreases by nearly 10-fold when the grains mature [8]. In addition, up to 75% of the reduction in nitrogen in leaves is from the chloroplasts [9], which provide most of the nitrogen that is redistributed to wheat grain [10]. The redistribution of N is also found in other crops, such as 50–90% of the N in rice ears [11] and 60–80% of the N in corn kernels [12] from N transport in leaves. Therefore, leaf senescence occurs in a coordinated manner during grain filling. Premature aging will affect the formation of assimilates, and delayed aging will affect the optimal time for nutrients to be transferred to reproductive organs. The senescence time of a flag leaf is one of the important factors affecting the crop yield and grain mineral element content. Aging is affected by many internal and external factors. The external factors include cultivation conditions, such as temperature, sunshine, water and nutrition supply, diseases and pests, etc. The internal factors include the developmental stage and hormone level. Although various factors are interrelated and influence each other, the internal factors play a leading role. Different techniques have been used to identify many senescence-associated genes (SAGs), which are mainly involved in degradation, biosynthetic and regulatory processes, stress responses, and transport [13,14,15,16]. In SAGs, NAC (no apical meristem (NAM), ATAF1/2, and cup-shaped cotyledon (CUC)) and WRKY transcription factors (TFs) play important roles in regulating gene expression changes during leaf senescence. For example, delaying senescence by declining the expression of NAM-B1 reduced the contents of protein, zinc, and iron in wheat grain by more than 30% [13,17]. The NAM-1 allele was also related to delaying senescence and prolonging the grain-filling period of wheat [18]. ZmNAC126 could be induced by ethylene, the ectopic overexpression of ZmNAC126 in Arabidopsis, and maize enhanced chlorophyll degradation and promoted leaf senescence [19]. The ABA signaling pathway and antioxidant enzyme systems are involved in TaNAC29-mediated stress tolerance mechanisms and plays an important role in the senescence process [20]. Expression of the TaSNAC11-4B gene was induced by all the senescence-promoting phytohormones, namely abscisic acid (ABA), jasmonic acid (JA), ethylene, and salicylic acid (SA) [21]. EIN3 (a key transcription factor in the ethylene signaling pathway) and ORE1/NAC2 accelerate chlorosis during ethylene-mediated leaf senescence by directly activating chlorophyll catabolic genes in Arabidopsis [22]. TaWRKY13-A, TaWRKY40-D, and TaWRKY42-B have also been reported to be involved in the regulation of leaf senescence and related to the JA or ABA pathway [23,24,25]. In conclusion, NAC and WRKY transcription factors seem to be inextricably linked with plant hormones, such as ABA, JA, and ethylene, and play crucial roles in senescence. In addition, it has been reported that the interaction between different hormones is also related to leaf senescence [26]. Although there are many studies on senescence in wheat, Arabidopsis, rice, maize, etc., most of them are based on the regulation of a single gene. Using 9K probe sets that were designed against unigenes from 35 wheat tissue cDNA libraries, Gregersen et al. (2006) [27] conducted a microarray experiment on the transcriptome of a wheat flag leaf during its senescence to obtain information about the senescence process. However, Poole et al. (2007) [28] pointed out the insufficiency of commercial and homemade microarrays used in wheat studies and found no comparability existed among these results. Moreover, the huge size, polyploid complexity, and estimated 75% repetitive DNA sequence content of the wheat genome [29] also largely limited their use. The relationship between flag leaf photosynthesis, the timing of senescence, and grain filling are not clear [30]. In recent years, with the development of high-throughput sequencing technology, transcriptome sequencing (RNA-seq) has become a powerful tool for studying complex biological processes and identifying candidate genes related to specific biological functions at the molecular level. It not only contains genomic information but also reveals the flow of biological information to cells in the genome and can find unknown transcripts, fusion genes, and genetic polymorphisms, which is impossible for chips. Its high sensitivity will also have a better detection of low abundance transcripts. The availability of a large number of transcriptome datasets provides an opportunity to comprehensively analyze the senescence process of wheat. Therefore, in this study, we used RNA-seq technology and the model wheat variety “Chinese Spring” that was sequenced by the whole genome to characterize possible senescence-related genes at different development stages of flag leaves of wheat, an important food crop. The present research extends the wheat cDNA array analysis of Gregersen and Holm (2006) [27] by examining the expression of extensively annotated wheat genes and further separately analyzing the substance translocation, transcriptional regulation, and phytohormone signaling during leaf senescence. It contributes to a better regulation understanding of light assimilate formation and the redistribution of senescence materials and identifies the genes involved in early leaf senescence, further providing the possibility for the simultaneous improvement of wheat yield and quality.
The bread wheat cultivar Chinese Spring was field grown at the Wheat Breeding Center of Northwest A&F University in Yangling, Shaanxi Province (34°26′ N, 108°14′ E) at an altitude of 527 m. It belongs to a warm temperate monsoon climate, with an annual average temperature of 12.9 °C, average annual precipitation of 500–700 mm, and a frost-free period of 211 d. During the wheat flowering until harvest (15 April–2 June), the daily maximum temperature is 33 °C, the daily average temperature is 10–26 °C, the average relative humidity is 66.87%, and the daily average relative humidity is 40.67–97.31% (Figure S1). Meteorological data were from https://rp5.ru/ (accessed on 15 September 2022). The amount of fertilizer applied before sowing was 120 kg/ha urea and 300 kg/ha diammonium phosphate. The sowing amount of wheat seeds was 150 kg/ha, and row spacing was 0.20 m, the plot area 35 m2 (5 m × 7 m). Irrigation was carried out at the jointing stage of wheat, and 225 kg/ha of urea was applied. The main culm spikes were tagged upon anthesis, and the flag leaves were sampled at 0 (20 April), 15 (5 May), 25 (15 May), and 30 (20 May) days after anthesis (DAA). The weather conditions on the day of sampling are shown in Figure S1. At 6 p.m. on each date, using the five-point sampling method, 10 flag leaves were collected from different plants. Samples were snap frozen in liquid nitrogen and stored at −80 °C for later use.
Total RNA was extracted using an RNA extraction kit, following the manufacturer’s protocol (Biotech, Beijing, China). The integrity, quantity, and purity of each RNA sample were examined by 1% agarose gel electrophoresis using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and a NanoDropTM 1000 (Thermo Fisher Scientific, Waltham, MA, USA). Ten equimolar high-quality RNA samples for each time point were mixed and sent to Novogene Co., Ltd. (Beijing, China) for cDNA library preparation. Finally, four libraries were sequenced on the Illumina HiSeq 2500 platform for paired-end sequencing with a read length of 125 bp (Novogene, Beijing, China).
Raw data were first preprocessed by filtering adaptor contaminants and reads with ≥10% poly (N), as well as ≥50% low-quality bases (SQ ≤ 5). The remaining data were utilized for subsequent analyses. The reference genome of the bread wheat cultivar Chinese Spring used for mapping was produced by the International Wheat Genome Sequencing Consortium (ftp://ftp.ensemblgenomes.org/pub/current/plants/fasta/triticum_aestivum/dna/Triticum_aestvum, accessed on 9 June 2015; IWGSP2.25.dna.toplevel.fa.gz/, accessed on 9 June 2015). TopHat v2.0.12 [31] was applied, where all clean reads were aligned against the reference genome, and no more than two nucleotide mismatches were allowed. Subsequently, the reads that simultaneously mapped to multiple transcripts were eliminated, and only the unambiguously mapping reads were retained. Meanwhile, novel transcripts were constructed and identified by the reference annotation-based transcript (RABT) assembly method that was built on Cufflinks v2.0.2 with default parameters [32].
Uniquely mapped reads were counted in union mode with HTSeq v0.6.1 [33]. The RPKM (reads per kilobase of exon model per million mapped reads) value was calculated as the expression amount of the transcript. Prior to normalization, only the transcripts that were found in at least two libraries and more than three mapped reads were kept. The differential expression analysis was conducted in a timepoint-wise manner by employing DEGSeq R package v1.12.0 [34] after all read counts were normalized by the TMM method [35]. Whether the gene has significant differential expression was evaluated by the difference multiple of expression and the p-value corrected by the Benjamini and Hochberg method [36]. When the expression difference of a gene in the two libraries was more than 2-fold (|log2[foldchange]| > 1) and the corrected p-value was less than 0.005, the gene was considered a significantly differentially expressed gene (DEGs).
Heat map clustering of DEGs was performed using the heatmap.2 function of Bioconductor package v 1.0.1 ggplots (Clustering mode: hclust; Distance metric: Pearson correlation; linkage method: average). At the same time, the relative expression levels of the DEGs were calculated using Multi Experiment Viewer v4.7.4, DBF_L1 samples as controls, and the expression trend clustering model of these genes was established using the same hierarchical clustering parameters.
Using a Hidden Markov model (HMM) of protein families from the TIGR (http://blast.jcvi.org/web-hmm, accessed on 15 June 2015) and PFAM databases (http://pfam.xfam.org, accessed on 15 June 2015), differentially expressed novel transcripts were annotated by HMMscan. The parameters that controlled the results were based on a previous study [37]. The annotations for other DEGs and the enrichment analysis were subsequently conducted with GOseq R package v1.20.0 based on Wallenius’ noncentral hypergeometric distribution. KOBAS v2.0 was utilized to identify biological pathways in which DEGs were involved and calculate the statistical significance of each pathway using the default criteria [38]. The GO terms and KEGG pathways with corrected p-values of less than 0.05 were considered significantly enriched.
All DEGs were blasted against protein sequences from the Leaf Senescence Database (http://www.eplantsenescence.org, accessed on 16 June 2015) and the Arabidopsis Phytohormone Database (http://ahd.cbi.pku.edu.cn, accessed on 16 June 2015) by employing BLASTx (E value ≤ 1 × 10−10). In addition, the differentially expressed new genes were compared with the NCBI nonredundant protein database (NR) (BLASTx, E value ≤ 1 × 10−10), and the protein sequences with the highest degree of similarity obtained and the protein sequences of other DEGs extracted from Enesembl Plant were analyzed to identify the transcription factors in them; at the same time, the hidden Markov model of all transcription factors in Pln TFDB was used to analyze the possible transcription factors and their types in all DEGs with HMMER3.0 (http://hmmer.janelia.org/, accessed on 16 June 2015). After that, all the above identification results were confirmed with the NCBI tool BLASTn. Finally, expression clustering was performed according to the clustering analysis of DEGs.
Total RNA extraction and quality detection were performed as described in Section 2.2, and EasyScript® First-strand cDNA synthesis Supermix (TRAN, Beijing, China) was used for cDNA synthesis. Using a QuantStudioTM 7 FleX Real-Time PCR Detection System (Thermo Fisher Scientific, Waltham, MA, USA), qRT-PCR was done using the PerfectStart® Green qPCR Supermix (TRAN, Beijing, China). The qRT-PCR was completed in the three biological replications where the same samples were used for RNA-seq analysis. To standardize the relative expression, the actin gene was used as a housekeeping gene. The relative expression values for each of the 12 selected genes were obtained by the delta–delta threshold cycle (2−ΔΔCt) method. The primers used in the qRT-PCR analysis are listed in Table S1. The relative expression data were subjected to ANOVA in SPSS (version 22, SPSS Inc., Chicago, IL, USA). The correlation between RNA-seq and qRT-PCR was calculated by a Pearson correlation coefficient.
The chlorophyll content and the absorbance of 0.1 g flag leaf at 645 and 663 nm in four periods were measured by a spectrophotometer. Refer to the Plant Chlorophyll Content Assay Kit (Boxbio AKPL003M, Beijing, China) for detailed experimental procedures and Chla and Chlb content calculations. The soluble protein content was determined using Coomassie brilliant blue [39].
With Q30 percentages (sequencing error rate 0.1%) over 92.15%, a total of 269,360,020 paired-end raw reads were generated for four libraries. More than 97% of the raw reads were determined as clean data after quality filtering from individual libraries, and 62.37–65.57% were mapped to a single transcript. Most of these reads (over 88.80%) were aligned to exons (Table 1). In sum, 76,872 transcripts found in at least two libraries and more than three aligned reads were obtained, of which 11,771 were novel. The raw data were submitted to the Sequence Read Archive (SRA) at the National Center for Biotechnology Information (accession number: SRP067916).
Based on the cutoff values (|log2[foldchange]| > 1, p < 0.005), 4887 DEGs were identified among any two time points. At 0–15 DAA, 15–25 DAA, and 25–30 DAA, most of the DEGs (90.96%, 75.30%, and 77.04%) did not show significant changes in their expressions, respectively. The biggest shift of upregulated DEGs took place in the interval of 15–25 DAA (820) compared to other intervals, while downregulated DEGs continuously increased until the interval of 25–30 DAA (666). Only 29 DEGs were shared by all three intervals, and common genes between 0–15 DAA and either 15–25 DAA (92) or 25–30 DAA (67) were less than that between the latter two intervals (446). Around 55.08% of DEGs only showed differential expressions during the periods across one or two-time points (e.g., 0–25 DAA, 25–30 DAA, and 0–30 DAA). The DEGs in each interval are shown in Figure S2.
As shown in Figure 1A, biological processes, including cell morphogenesis (GO:0000902), transport (GO:0006810), abiotic stimulus-response (GO:0009628), and phosphorus metabolism (GO:0006793), were enriched in the first 15 DAA. Expressions of most DEGs that related to the first three terms were upregulated, whereas DEGs involved in phosphorus metabolism, especially protein phosphorylation (GO:0006468), were downregulated. Carbohydrate (GO:0005975), lipid (GG:0006629), and amino acid metabolisms (GO:0006520) were active processes after anthesis. Upregulated DEGs (58) associated with lipid metabolism became remarkably enriched compared to the downregulated DEGs (19) at 15–25 DAA (Figure 1B). Compared with results during 15–25 DAA, downregulated DEGs for both carbohydrate and amino acid metabolisms were more significantly enriched than the upregulated DEGs at 25–30 DAA (Figure 1C). In contrast to the increase at 15–25 DAA, transcripts of all DEGs involved in the tricarboxylic acid (TCA) cycle (GO:0006099) and the coding of phosphofructokinase (PFK) (GO:0005945) were exclusively dropped at 25–30 DAA. Additionally, significant changes in chloroplast-related DEGs were common traits at both 15–25 DAA and 25–30 DAA (Figure 1). The DEGs with reduced expression levels progressively increased until almost no upregulated DEGs were noted at 25–30 DAA, especially the genes that encoded functional units of photosystem I (PSI) (GO:0009523) and photosystem II (PSII) (GO:0009522) (Table S2). The upregulated DEGs, including a transmembrane transporter (GO:0055085) and an amino acid transporter (GO:0006865), were enriched at 15–25 DAA and 25–30 DAA, respectively. In addition, cysteine-type peptidase DEGs and antioxidant-encoded DEGs were specifically enriched at 25–30 DAA.
More KEGG pathways were significantly enriched for both up- and downregulated DEGs at 15–25 DAA (13/7) and 25–30 DAA (15/11) than at 0–15 DAA (5/4). Two pathways, glyoxylate and dicarboxylate metabolism and carbon fixation in photosynthetic organisms, were vigorous in all three sequential periods. Additionally, more pathways were shared by the 15–25 DAA and 25–30 DAA (3/5) intervals (Figure S3). All DEGs were sorted into 26 groups based on their expression dynamics, and 12 of these are shown in Figure 2. Although an overall downward trend was observed for major photosynthesis-related genes, minor distinctions in expressions were still detected among DEGs that individually related to carbon fixation in photosynthetic organisms (Figure 2B), carotenoid metabolism (Figure 2C,D), photosystem (Figure 2D), and chlorophyll metabolism (Figure 2C,J). Notably, only the transcripts of Rubisco-coded genes kept dramatically reducing after flowering (Figure 2A). DEGs that participated in carbohydrate, lipid, and protein metabolisms showed various expression patterns. For example, the average transcript amounts of major DEGs involved in glyoxylate and dicarboxylate metabolism and the pentose phosphate pathway dropped from 15 DAA and 25 DAA, respectively (Figure 2B,D). Transcript amounts of DEGs for lipid metabolism reached their maximum at 25 DAA (Figure 2F,G) or 30 DAA (Figure 2K). Moreover, the genes associated with the metabolism of multiple amino acids were generally upregulated after 25 DAA (Figure 2F,H). As a crucial cofactor for metal element transport, DEGs for nicotinamide metabolism showed globally significant transcript accumulations at 25–30 DAA and 15–30 DAA (Figure 2H,K). The same trends were also observed for DEGs involved in the regulation of autophagy (Figure 2H,K).
Over one-third of the DEGs were identified as putative SAGs (1780/4887) in this study, which was close homologs of SAGs previously found in 14 species. As expected, a large number of genes played roles in multiple substance metabolisms (30.24%), including the lipid, carbohydrate, amino acid, protein, nucleic acid, chlorophyll, and secondary metabolites. In addition, other SAGs were also involved in transcription regulation (5.22%), redox regulation (6.35%), transport (7.42%), and phytohormone and signal transduction (8.66%), among others. All information relating to putative SAGs is shown in Table S3.
A total of 273 putative transporter-encoded DEGs were confirmed, and they were involved in the inter- or intracellular trafficking of a range of substances, including phytohormones, carbohydrates, amino acids, and minerals. Followed by carbohydrate transporter genes (16.48%), metal translocator genes were the most abundant group comprising over one-third of the transporter-coded DEGs (23.08%) (Table S4). According to the expression patterns, the transcripts of amino acid permease 3 (AAP3), bidirectional amino acid transporter 1 (BAT1), cationic amino acid transporter 2 (CAT2), and part of proline transporter 1 (ProT1) kept rising after anthesis, whereas that growth for most lysine histidine transporters (LHTs) started at 15 DAA (Figure 3A–D). The expressions were discrepant among carbohydrate translocator DEGs to some extent. Apart from an over five-fold (5.17) consistent transcript increase for one glucose-6-phosphate/phosphate translocator (GPT) at 15–30 DAA, others were only upregulated at 15–25 DAA (Figure 3E). Unlike various trends for other polyol transporters (PLTs), the upregulation of most PLT5s was observed at 15–30 DAA (Figure 3F). Followed by a slight increase (1.01–1.14-fold), the expression of a tonoplast sugar-efflux transporter, early responsive to dehydration 6-like 4 (ERD6-like4), was remarkably promoted before 25 DAA (2.71–3.44-fold) (Figure 3G). DEGs that code aquaporins showed significant transcription enhancements (2.35–4.28-fold) at 0–15 DAA and complex change trends after that (Figure 3H). Multiple metal transporter genes were differentially regulated. Two copper transporter genes, copper chaperone (CCH) and partial antioxidant 1 (ATX1), were upregulated at 15–30 DAA (Figure 3I), whereas the magnesium transport genes, mitochondrial RNA splicing 2-A (MRS2-A) and nonimprinted in Prader-Willi/Angelman syndrome region protein 2 (NIPA2), were downregulated during this period (Figure 3J). About 34.92% of the metal transporter-encoded DEGs were confirmed as potassium-related, and they showed diversity in their expression trends (Figure 3K–N). Iron and zinc transporter genes were also differentially expressed. With a flat expression at 0–15 DAA (<1.5-fold), ferric reductase oxidase 7 (FRO7) transcripts showed dramatic drops afterward (Figure 3O). DEGs that were homologous with two heavy metal transporters (HMAs) and two vacuolar iron transporters (VITs) were identified. The expression pattern of HAM1 reversed that of FRO7, and VIT1 showed opposite changes during the first two periods, namely a steep rise at 0–15 DAA, followed by a sharp drop at 15–25 DAA (Figure 3Q). Four different yellow stripe-likes (YSLs) were DEGs, and most of them were upregulated after flowering (except one YSL12 and one YSL14) (Figure 3S).
There were 164 differentially expressed TFs that belonged to 27 different families, and NAC (19.52%) constituted the largest TF family (Table S5). All TFs were divided into 14 groups using a hierarchical cluster analysis (Figure 4). Most TFs (54.88%) were concentrated in three clusters, namely clusters 11 (32), 12 (26), and 14 (32) (Figure 4K–N). Among these genes, TFs of clusters 11 and 12 showed an average of 11.44-fold upregulated expressions and 5.34-fold downregulated expressions after a flat change at 0–15 DAA (1.16/1.02-fold), respectively. Over half of the NAC genes (18) belonged to cluster 11, including the homologs of AeCUC2, AeNAC29, TuNAC8, and TaNAC6. Additionally, the closest ortholog of auxin response factor 18 (OsARF18) was in this group. Major TFs with putative roles in basic biological activities were in cluster 12, such as the homolog genes of golden-like 2 (TuGLK2), agamous-like 10 (TaAGL10), CO-like 4 (OsCOL4), and rice RNA polymerase sigma factor genes. Compared with the nearly 4-fold rise in cluster 11 (3.95-fold) at 25–30 DAA, the overall expression change of the DEGs in cluster 14 was insignificant (1.30-fold). The homologs of eight NACs (AeCUC2 and AeNAC29), IAA-leucine resistant 3 (OsILR3), OsMYB1R1, and RBA-related 1 (TaRBR1) were in this class. Notably, only three previously characterized NAC genes (AeNAM-D1, TtNAM-A1, and TtNAM-B2), which control the reserve substance redistribution and flag leaf senescence in wheat, were found in cluster 13. After a rapid rise in the transcript levels (7.0–15.99-fold), their increases became slower after 15 DAA (1.08–2.55-fold) (Figure 4M). Only 6.10% of DEGs were defined as WRKY genes, and these genes mainly showed two general types of expression trends (Figure 4B–D,F,H). At 0–15 DAA, the transcripts of all WRKY genes were dramatically reduced (4.81–8.99-fold), except for a slight decline (1.42-fold) for one TuWRKY6 (OsWRKY1) homolog. After this point, four WRKY genes (including TaWRKY74) in cluster 8 continued to be modestly downregulated until 30 DAA (<4.50-fold). However, the expressions of TaWRKY53, TaWRKY14, and TaWRKY2 in cluster 2 (>8.54-fold) were promoted more sharply than those of TaYWRK27 in cluster 4 (4.71-fold) and TuWRKY41 in cluster 3 (1.97-fold) (Figure 4).
A local blast facility against an Arabidopsis phytohormone database was applied to identify 124 putative phytohormone-associated DEGs, some of which were simultaneously implicated in multiple phytohormones. These genes were likely involved in the synthesis, transport, and signal transduction of seven phytohormones, including abscisic acid (ABA), jasmonic acid (JA), ethylene (ET), cytokinin (CTK), brassinosteroid (BR), salicylic acid (SA), and auxin (IAA). IAA (44), JA (33), and ABA (28) were the three phytohormones with the most abundant DEGs (Table S6). The expression trends of some core components for these processes are shown in Figure 5. Unlike the transcript increase of pyracbactin resistance-like 8 (PYL8) after 15 DAA, other ABA receptors, namely genomes uncoupled 5 (GUN5) and PYL4, declined after 15 DAA or 0 DAA, respectively (Figure 5A). Moreover, type 2C protein phosphatase (PP2C) family genes, the negative signaling regulator of ABA, were all upregulated during 0–25 DAA, while the positive regulator respiratory burst oxidase (RBOH) was only upregulated at 15–25 DAA. Multiple IAA transporter genes were also differentially expressed. Genes encoding influx transporter like-AUX1 3 (LAX3) and efflux transporter aminopeptidase (M1APM1) were down- and upregulated during 0–30 DAA, respectively. Efflux transporter P-glycoprotein 11 (PGP11) was upregulated only until 25 DAA (Figure 5B). Evident expression correlations among DEGs for JA biosynthesis were not observed. For instance, the expression patterns of omega-3 fatty acid desaturase 7 (FAD7) and phospholipase D (PLD) at 15–30 DAA reversed that of another FAD7, 12-oxophytodienoate reductase 1 (OPR1), as well as 3-ketoacyl-CoA thiolase 2 (KAT2) (Figure 5F). A difference between two core components of JA signal transduction, namely coronatine insensitive1 (COI1) and TIFY10A/B, was detected only at 25–30 DAA (Figure 5F). Information about the DEGs associated with other phytohormones was sparse and complex. As with the signaling repressors of BR, the expression dynamics for growth regulating factor 3 (GRF3) and shaggy-related protein kinase (ASK41) were diverse (Figure 5C). Despite disparate functions of histidine kinase 3 (AHK3) and the two-component response regulator ARR12 (ARR12) in the CTK signal pathway, their encoded genes were similarly upregulated (Figure 5D). The expression of ETHYLENE INSENSITIVE3 (EIN3), a positive regulator in the ET-signaling system, was partially coordinated with that of the ET synthesis gene s-adenosylmethionine synthase (SAM2) (Figure 5E).
To validate the results of RNA-seq, we selected 12 genes from different pathways of SAGs to perform a quantitative real-time PCR (qRT-PCR) analysis. These results showed a similar trend to our RNA-seq results, giving further credence to our sequencing findings (Figure S4). For instance, among the transporter-related genes, Traes_2AS_9DE16F020 and Traes_2AS_9DE16F020 had a trend of decreasing first and then increasing, and both had the highest expression at 30 DAA. The expression of four TF genes, Traes_7DS_F5A240B02, Traes_2AL_8A23618BA, Traes_1BS_EF67E5A24, and Traes_1AS_F3EAEC435 increased from post-anthesis to 30 DAA. Traes_5BL_62D9B877B, phytohormone-related genes, had the same trend. Chlorophyll synthesis-related genes Traes_6DL_0FF72D765 and Traes_2DL_30F23E577 gradually decreased from 0 DAA or 15 DAA to 30 DAA. The expression of Traes_2BL_98439EA10 and Traes_2DS_AE6E354A9 in 0–15 DAA gradually increased but then gradually decreased. The correlation heatmap showed a good correlation between RNA-seq and qRT-PCR, which indicated the accuracy of our sequencing data (Figure S5).
The flag leaf is the last leaf that grows before the emergence of the wheat spike, and it is an important sign of wheat entering the booting stage. Under suitable growth conditions, 80% of the dry matter of wheat grains is accumulated after anthesis, and the contribution rate of “functional leaves”, i.e., flag leaves, to wheat grain yield can be as high as one-third. In addition to photosynthetic compounds, the senescence of flag leaves also provides some basic materials for the vigorous growth of grains. Previous studies have shown that the advance in senescence of vegetative tissues will be conducive to the accumulation of more protein and essential trace elements in grains [13]. The wheat flag leaf serves as the major N source that provides a large amount of assimilates for grains at the post-anthesis stage and serve as a model tissue to study leaf senescence and N remobilization [16]. Therefore, the senescence of wheat flag leaf is an important factor that affects the main agronomic traits, including yield and nutritional quality. The study on the senescence mechanism of flag leaf will help to improve these agronomic characteristics. The cluster analysis in the present study showed that expression trends of major DEGs were different before and after 15 DAA (Figure S6). Compared with 0–15 DAA, more metabolic pathways were enriched at 15–25 DAA and 25–30 DAA. Among them, more genes involved in the metabolism of fat, amino acids, and carbohydrates are more active in 15–25 DAA. Additionally, there was more similarity between these two stages than at 0–15 DAA. This suggested that 15 DAA was a crucial turning point for the metabolism in the flag leaf, and it reflected a response of metabolic changes in the flag leaf to the demand of the grain as well. Grain development consists of two phases after anthesis: cell division and grain filling. In the first stage, the kernel length and width are established until approximately 14 DAA [40], but not much dry matter is accumulated. This corresponds with the slow-rate dry matter accumulation in the grain before 12 DAA [41]. The downregulated genes increased continuously after anthesis and reached the maximum at 25–30 DAA. This indicates that, at the molecular level, 15–25 DAA may be a transitional period for the metabolic degree of Chinese Spring flag leaf cells. At this time, there may be catabolism and anabolism in the cell, and anabolism is stronger than catabolism. In addition, multiple metabolisms at 25–30 DAA became less active than those at 15–25 DAA, likely due to the progressive function loss of chloroplasts after 15 DAA. In the mature grains, 66.7% of the total N was remobilized from the pre-anthesis accumulation in the biomass, while the remaining 33.3% was derived from the N taken up during post-anthesis. From anthesis to 2 weeks after the anthesis stage, the flag leaf remobilized 3.67 mg of N outwards, and the ear remobilized 3.87 mg of N inwards from the pre-anthesis accumulation in each plant [42]. The DEGs related to photosynthesis also changed significantly from 15 DAA. Whether it is the functional components of the photosynthetic system or other related genes involved in photosynthesis, the number and types of 25–30 DAA downregulated genes were significantly increased compared with 15–25 DAA. Differently, only enriched downregulated DEGs encoding components of the photosynthetic system were observed at 25–30 DAA. Although leaf yellowing is a major visible symptom of senescence, the main factor that causes photosynthesis inactivation during this process remains under discussion. In wheat, parallel losses of photosynthetic activity and chlorophyll content have been reported [43], whereas others found no obvious correlation existed [44]. With a flat expression at 0–15 DAA, DEGs that encode enzymes such as pheophorbide an oxygenase (PaO) and chlorophyll (ide) b reductases (CBR) for chlorophyll breakdown were significantly upregulated after 15 DAA, while DEGs involved in chlorophyll synthesis were downregulated. Therefore, the rapid reduction in the chlorophyll content may start at 15 DAA or later (Table S7 and Figure S7). Additionally, Grover et al. (1993) [45] (pp. 225–255) found that chloroplast function loss was associated with the decline in photochemical activities of PSI and PSII due to the chloroplast disintegration in wheat. In the present study, dramatic changes in the DEG coding functional photosystem units were observed after 15 DAA. Furthermore, more DEGs related to PSII (11) showed reduced expressions than those related to PSI (6) at 15–25 DAA, whereas, after, a larger increase in the number of PSI-encoded genes (18) was detected compared to PSII (18) genes. Several studies have indicated diverse physiological changes between PSI and PSII during leaf senescence. For example, PSII was more susceptible to senescence than PSI [46,47]. It has also been reported that the decline of PSII activity preceded that of PSI activity in wheat leaves during heat-promoted senescence [10]. Despite the rate-limiting enzyme for photosynthesis in the C3 plant [48], the role of Rubisco in photosynthetic activity decline has been debated [45,49,50,51,52]. Interestingly, Rubisco-encoded genes were the only DEGs that showed consistently and significantly decreased expressions after anthesis in the present study. Considering a normally rapid drop of net photosynthesis after 15 DAA [53], the overall results suggested that a chlorophyll decrease or photosystem breakdown had major effects on photosynthetic inactivation, rather than the reduced amounts of Rubisco (Figure S7). Zhou et al. (2018) [42] also reported that Rubisco might play a critical role in N deposition. Regardless of the photoassimilate or storage resource, transporters are undoubtedly indispensable for their distributions of grain. As one of the major long-distance transport forms of organic nitrogen, the translocation of amino acids depends on various membrane-integral transporters in plants [54]. The expression trend of each transporter was complex, and the expression of genes involved in the same function was also different. AAPs were involved in the phloem trafficking of amino acids supplied to the sink [55,56]. The rapid transcript rise of AAPs in the flag leaf 15 DAA may also reflect their response to the growing demand for grain. Furthermore, magnesium (Mg) ion was not only the central atom of the chlorophyll molecule but also specifically affected the half-lives of chloroplast RNA [57]. Three Mg transporter DEGs were remarkably downregulated during 15–30 DAA, which likely resulted from chloroplast degradation. TaFRO1 and TaFRO2 (ferric reduction oxidation) encode oxidoreductases localized to the chloroplast membrane. This enzyme was responsible for the reduction of Fe3+ to Fe2+ before the transport of Fe3+ to the chloroplast cavity. Consistent with previous studies, the expression of TaFRO1 and some TaFRO2 genes decreased at 15–30 DAA [27]. Both HMA1 and FRO7 reside in chloroplast membranes; however, they are functionally different. HMA1 is a member of the Zn/Co/Cd/Pb-transporting HMA family, and its close homolog in rice is shown to be able to export metals across the chloroplast [58]. In the present study, expression changes implied that the efflux of Zn/Fe from chloroplasts was more active than their influx at 15–30 DAA. The vacuole was a crucial organelle to buffer the concentration of metal in the cell. A few tonoplast-integrated transporter genes were differentially expressed. A previous study suggested a potential contribution of OsVIT1 to the recycling of Zn/Fe during senescence [59]. At 0–30 DAA, the expression level of TaVIT1 decreased by about 2.4–4.5-fold, while Ta VIT2 decreased from 15 DAA. Consistent with previous research, only NRAMP2, a member of the NRAMP family, was identified as the DEG during grain filling [60]. Moreover, the continuous upregulation at 15–30 DAA confirmed that this gene facilitated the iron efflux from the vacuole to satisfy the material demand during grain development. Many identified NAC and WRKY TFs have been associated with leaf senescence in Arabidopsis, barley, and wheat [15,21,61,62]. Over half the NAC genes identified in this study showed a correlation with senescence based on their expression patterns, including NAC21/22, NAC29, NAC8, NAC7, and three other characterized NAM genes. Although less direct evidence has been established, several studies ultimately proved the involvement of their homologs in aging or stress responses. For example, it was shown that ectopic expression of TaSNAC11-4B in Arabidopsis promotes ROS accumulation and significantly accelerated age-dependent, as well as drought- and ABA-induced, leaf senescence. TaSNAC11-4B was shown to positively regulate the expressions of AtrbohD and AtrbohF, which encode catalytic subunits of ROS-producing NADPH oxidase [21]. TaNAC21/22 negatively regulates the resistance of wheat to stripe rust [63], while TaNAC8, TaNAC29, and TaNAC7 are also related to the defense responses [64,65,66]. NAM-A1, NAM-B2, and NAM-D1 were expressed in a generally similar trend to that established in a previous study in tetraploid wheat [13]. The present study verified the dramatic increase in their transcript amounts of those regulatory genes exclusively at 0–15 DAA in early senescence. Research conducted by Lu et al. (2022) [67] showed that the yield of a premature senescence mutant (GSm) was considerably lower than that of the wild type (WT). Many physiological indexes are lower than those of the WT, except malondialdehyde. The transcriptomic analysis indicated that blockades of chlorophyll and carotenoid biosynthesis accelerate the degradation of chlorophyll and diminish the photosynthetic capacity in mutant leaves and that brassinolide may facilitate chlorophyll breakdown and, consequently, accelerate leaf senescence. This research also found that NAC genes positively regulate the senescence process. Compared with NAC genes, expressions of the WRKY and MYB genes are induced earlier in the mutant, possibly due to increased levels of reactive oxygen species and plant hormones (e.g., brassinolide, salicylic acid, and jasmonic acid), thereby accelerating leaf senescence. Furthermore, this research ascertained that the antioxidant system plays a role in minimizing oxidative damage in the mutant [67]. Mostly, the genes detected in the carotenoid biosynthetic pathway initiated by ABA (PYR/PYL, PP2C, SnRK2, and ABF) expressed higher in either no or low N than high N. ABA is known to promote senescence in the plant by the stomatal closure, which restricts cellular growth and induces SAGs, non-yellow coloring 1(NYCI), stay green (SGR), PHEOPHYTINASE (PPH), and pheophorbide an oxygenase (PAO) gene expressions. The higher expression of carotenoid-related genes may be attributed to the chlorophyll degradation observed in the N-limited conditions (no and low N) compared to the high N [16]. In the past 20 years, WRKY TFs have been broadly investigated in plants, and they were extensively proven to be involved in numerous physiological processes, including leaf senescence. For example, WRKY14, WRKY2, WRKY53, WRKY13, WRKY40, and WRKY42 are putative SAGs, as their close homologs are positive regulators of JA- and ET-mediated signaling [23,24,25,68]. Likewise, in the present study, these genes were upregulated as the flag leaf senesced. Inverse expression patterns between phytochrome-interacting factor 3 (PIF3), PIF5, and GLK2 confirmed the nature of their interactive relationships. PIFs are central mediators in phytochrome signaling [69], and a study on Arabidopsis proved that certain ones (PIF3, PIF4, and PIF5) are required to restrain photomorphogenic development by repressing chloroplast activity maintainer gene GLK2 [70]. Accumulated evidence implies a more complicated network is involved in the regulation of leaf senescence. Although it has been suggested that CTK and IAA have positive roles in plant development, many studies have found these phytohormones also have functions in leaf senescence [71]. CTK-induced changes in nitrogen remobilization and chloroplast ultrastructure in wheat retain the sink activity of the older leaves by inhibiting amino acid and sugar export to the phloem and stimulating assimilate accumulation in the chloroplasts of the older leaves [72]. The JA, ABA, SA, and BR-related genes promoted leaf senescence in low N, whereas the IAA, GA, and CTK genes inhibited leaf senescence in hight N [16]. In line with a similar study in cotton, IAA polar transporter genes showed active changes as well [73]. Furthermore, the efflux of IAA was more vigorous than its influx in the present study. This result again verified the changes in IAA transport, rather than the endogenous IAA level itself, which could confer growth plasticity. Additionally, many DEGs were simultaneously modulated by multiple phytohormones, which demonstrated their extensive and complicated crosstalk during our sampling period. To date, few core genes have been identified for signaling pathways. This restricts a clear picture of the role of each phytohormone, and their relationships in the regulation of senescence were explored based on the expressions of related genes.
By analyzing the changes of the flag leaf transcriptome of Chinese Spring wheat at 0DAA, 15DAA, 25DAA, and 30DAA, the differential expression of senescence-related genes was investigated, which is helpful to understanding the relevant pathways and regulation of wheat senescence. We confirmed that different periods lead to the rearrangement of glyoxylic acid, dicarboxylic acid metabolism, carotenoid metabolism, photosystem, chlorophyll metabolism, carbohydrate, lipids, and protein metabolism. Some transporters, such as amino acid permease 3, bidirectional amino acid transporter 1 (BAT1), cationic amino acid transporter 2 (CAT2), part of proline transporter 1 (ProT1), lysine histidine transporters (LHTs), metal transporter genes, etc., were identified to be associated with senescence. Some NAC and WRKY family transcription factors such as TaWRKY53, TaWRKY14, and TaWRKY2, and senescence genes of ABA, JA, ET, CTK, BR, SA, and IAA related to plant hormones were identified. The results provided in this paper enrich the senescence gene resources and further provide a basis for improving the wheat yield and quality at the same time. Future research may clone some of these genes and further analyze their functions and regulatory mechanisms. | true | true | true |
PMC9572210 | 36233442 | Joao Carlos Batista Liz,Fernanda Genre,Verónica Pulito-Cueto,Sara Remuzgo-Martínez,Diana Prieto-Peña,Ana Márquez,Norberto Ortego-Centeno,María Teresa Leonardo,Ana Peñalba,Javier Narváez,Luis Martín-Penagos,Lara Belmar-Vega,Cristina Gómez-Fernández,José A. Miranda-Filloy,Luis Caminal-Montero,Paz Collado,Diego De Árgila,Patricia Quiroga-Colina,Esther F. Vicente-Rabaneda,Ana Triguero-Martínez,Esteban Rubio,Manuel León Luque,Juan María Blanco-Madrigal,Eva Galíndez-Agirregoikoa,Javier Martín,Oreste Gualillo,Ricardo Blanco,Santos Castañeda,Miguel A. González-Gay,Raquel López-Mejías | IgA Vasculitis: Influence of CD40, BLK and BANK1 Gene Polymorphisms | 22-09-2022 | IgA vasculitis,Henoch–Schönlein purpura,CD40,BLK,BANK1,polymorphisms | CD40, BLK and BANK1 genes involved in the development and signaling of B-cells are identified as susceptibility loci for numerous inflammatory diseases. Accordingly, we assessed the potential influence of CD40, BLK and BANK1 on the pathogenesis of immunoglobulin-A vasculitis (IgAV), predominantly a B-lymphocyte inflammatory condition. Three genetic variants within CD40 (rs1883832, rs1535045, rs4813003) and BLK (rs2254546, rs2736340, rs2618476) as well as two BANK1 polymorphisms (rs10516487, rs3733197), previously associated with inflammatory diseases, were genotyped in 382 Caucasian patients with IgAV and 955 sex- and ethnically matched healthy controls. No statistically significant differences were observed in the genotype and allele frequencies of CD40, BLK and BANK1 when IgAV patients and healthy controls were compared. Similar results were found when CD40, BLK and BANK1 genotypes or alleles frequencies were compared between patients with IgAV stratified according to the age at disease onset or to the presence/absence of gastrointestinal or renal manifestations. Moreover, no CD40, BLK and BANK1 haplotype differences were disclosed between patients with IgAV and healthy controls and between patients with IgAV stratified according to the clinical characteristics mentioned above. Our findings indicate that CD40, BLK and BANK1 do not contribute to the genetic background of IgAV. | IgA Vasculitis: Influence of CD40, BLK and BANK1 Gene Polymorphisms
CD40, BLK and BANK1 genes involved in the development and signaling of B-cells are identified as susceptibility loci for numerous inflammatory diseases. Accordingly, we assessed the potential influence of CD40, BLK and BANK1 on the pathogenesis of immunoglobulin-A vasculitis (IgAV), predominantly a B-lymphocyte inflammatory condition. Three genetic variants within CD40 (rs1883832, rs1535045, rs4813003) and BLK (rs2254546, rs2736340, rs2618476) as well as two BANK1 polymorphisms (rs10516487, rs3733197), previously associated with inflammatory diseases, were genotyped in 382 Caucasian patients with IgAV and 955 sex- and ethnically matched healthy controls. No statistically significant differences were observed in the genotype and allele frequencies of CD40, BLK and BANK1 when IgAV patients and healthy controls were compared. Similar results were found when CD40, BLK and BANK1 genotypes or alleles frequencies were compared between patients with IgAV stratified according to the age at disease onset or to the presence/absence of gastrointestinal or renal manifestations. Moreover, no CD40, BLK and BANK1 haplotype differences were disclosed between patients with IgAV and healthy controls and between patients with IgAV stratified according to the clinical characteristics mentioned above. Our findings indicate that CD40, BLK and BANK1 do not contribute to the genetic background of IgAV.
B-lymphocytes are key cells for an effective immune response, mainly because they produce immunoglobulins (Igs) [1]. Cluster of differentiation 40 (CD40), a glycoprotein expressed on the surface of B-cells, participates in the activation [2], survival, proliferation and differentiation of these lymphocytes and in the isotype switching of Igs [3]. B-lymphoid kinase (BLK) and B-cell scaffold protein with ankyrin repeats 1 (BANK1) are components of the B-cells’ signalosome [4]. In this regard, BLK is a src family nonreceptor tyrosine kinase [5] with a relevant role in the development and receptor signaling of B-cells [6], whereas BANK1 is an adaptor/scaffold that participates in B-cell activation and signalization [7,8]. Interestingly, CD40 [9,10,11,12,13], BLK [7,14,15,16,17,18] and BANK1 [7,8,18,19] genes are identified as susceptibility loci for several inflammatory diseases. Likewise, CD40 and BLK variants are known as susceptibility factors for different forms of vasculitis, specifically for the development of ischemic manifestations in patients with giant cell arteritis [20,21] and for Kawasaki disease [13,22], supporting the relevance of B-cell activation in the pathophysiology of both vasculitides. Immunoglobulin-A vasculitis (IgAV), or Henoch–Schönlein purpura (HSP), is a small-sized blood vasculitis, more common in children and rarer but more severe in adults [23,24,25]. Besides the classic clinical triad of palpable purpura, arthralgias/arthritis and gastrointestinal (GI) tract involvement [26,27], renal damage is also presented in patients with IgAV, constituting the most serious complication of this vasculitis [28,29]. Abnormal IgA deposits in the vessel walls are the characteristic pathophysiologic feature of IgAV [23], supporting the theory that this vasculitis is predominantly a B-cell disease. The etiology of IgAV has not been completely elucidated. Nevertheless, numerous pieces of evidence support the claim that genetics is crucial in the pathogenesis of this condition [30,31,32]. Accordingly, we aimed to determine, for the first time, the potential influence of CD40, BLK and BANK1 on the pathogenesis of IgAV. For this purpose, eight polymorphisms (three within CD40, three within BLK and two within BANK1) were genotyped in the largest series of Caucasian IgAV patients ever assessed for genetic studies.
The study group of the present work encompassed a total of 382 unrelated patients who fulfilled Michel et al.’s criteria [33] and/or the American College of Rheumatology’s classification criteria [34] for IgAV-HSP and/or the 2012 revised International Chapel Hill Consensus Conference Nomenclature [35] definition for IgAV. All these patients were Spaniards of European ancestry and were recruited in the following healthcare centers: Hospital Universitario Marqués de Valdecilla (Santander), Hospital Universitario Clínico San Cecilio (Granada), Hospital Universitario de Bellvitge (Barcelona), Hospital Universitario Lucus Augusti (Lugo), Hospital Universitario Central de Asturias (Oviedo), Hospital Universitario Severo Ochoa and Hospital Universitario de La Princesa (Madrid), Hospital Universitario Virgen del Rocío (Sevilla) and Hospital Universitario de Basurto (Bilbao). A description of the main clinical characteristics of the patients with IgAV included in this study is presented in Table 1. GI manifestation was considered present if bowel angina and GI bleeding were observed as previously described [31]. Renal manifestations were defined to be present if hematuria, proteinuria or nephrotic syndrome at any time over the clinical course of the disease and/or renal sequelae (persistent renal involvement) at last follow up was disclosed [31]. In addition, 955 sex- and ethnically matched, unrelated individuals without a history of cutaneous vasculitis or any other autoimmune disease from Hospital Universitario Marqués de Valdecilla (Santander) and National DNA Bank Repository (Salamanca) were also included in our work as healthy controls. All subjects gave their informed consent to be included in the study. The procedures followed were in accordance with the ethical standards of the approved guidelines and regulations, in accordance with the Declaration of Helsinki. All experimental protocols were approved by the local ethics committees of each participant hospital (approval code 15/2012 and date of approval 11 May 2012).
Three polymorphisms within CD40 (rs1883832, rs1535045 and rs4813003) and BLK (rs2254546, rs2736340, rs2618476) genes as well as two genetic variants within BANK1 (rs10516487 and rs3733197) were selected in this study. These eight specific variants were selected considering that they were related to numerous inflammatory disorders [7,8,9,10,11,12,13,14,15,16,17,18,19]. In addition, potential functional consequences were previously proposed for some of these polymorphisms [8,18,19,36]. DNA from all the IgAV patients and healthy controls included in the study was extracted from peripheral blood samples using the REALPURE “SSS” kit (RBME04, REAL, Durviz S.L., Spain). All individuals were genotyped for the eight genetic variants mentioned above using predesigned TaqMan genotyping probes (C__11655919_20 for rs1883832, C___1260189_10 for rs1535045, C___1260313_20 for rs4813003, C__16036468_10 for rs2254546, C___1886931_30 for rs2736340, C__16036467_10 for rs2618476, C____313748_30 for rs10516487 and C___1793403_1_ for rs3733197) in a QuantStudioTM 7 Flex real-time polymerase chain reaction system, according to the conditions recommended by the manufacturer (Applied Biosystems, Foster City, CA, USA). Negative controls and duplicate samples were incorporated in our study to check the genotyping accuracy.
CD40 rs1883832, CD40 rs1535045, CD40 rs4813003, BLK rs2254546, BLK rs2736340, BLK rs2618476, BANK1 rs10516487 and BANK1 rs3733197 genotypes were examined for deviation from the Hardy–Weinberg equilibrium (HWE). To test for association, we compared CD40, BLK and BANK1 frequencies between patients with IgAV and healthy controls as well as between patients with IgAV stratified according to specific clinical characteristics of the disease (age at the disease onset or presence/absence of GI or renal manifestations). CD40, BLK and BANK1 variants were evaluated independently. Genotype and allele frequencies were calculated and compared between the groups mentioned above using chi-square test or Fisher test. Strength of association was estimated using odds ratio (OR) and 95% confidence intervals (CIs). Then, we carried out the allelic combination (haplotype) analysis for the three CD40 genetic variants studied as well as for the three BLK polymorphisms assessed and the two BANK1 variants evaluated. Haplotype frequencies were calculated by the Haploview v4.2 software (http://broad.mit.edu/mpg/haploview) (accessed on 20 September 2022) and compared between the groups mentioned above using chi-square test. The strength of association was estimated by OR and 95% CI. We considered results with p-values <0.05 as statistically significant. All statistical analyses were conducted with the STATA statistical software 12/SE (Stata Corp., College Station, TX, USA).
No deviation from HWE was detected for CD40 rs1883832, CD40 rs1535045, CD40 rs4813003, BLK rs2254546, BLK rs2736340, BLK rs2618476, BANK1 rs10516487 and BANK1 rs3733197 in healthy controls. The genotyping success rate was greater than 98% for the eight polymorphisms evaluated in this study. Both genotype and allele frequencies of CD40, BLK and BANK1 variants assessed were in accordance with those reported in the 1000 Genomes Project (http://www.internationalgenome.org/) (accessed on 20 September 2022) for European populations.
Genotype and allele frequencies of CD40, BLK and BANK1 polymorphisms assessed independently were compared between patients with IgAV and healthy controls. In this respect, similar genotype and allele CD40, BLK and BANK1 frequencies were observed in patients with IgAV when compared to healthy controls (Table 2).
Subsequently, we evaluated whether differences in the genotype and allele frequencies of each CD40, BLK and BANK1 polymorphism could exist between IgAV patients stratified according to specific clinical characteristics of the disease. Since IgAV is often a benign and self-limited pathology in children and a more severe condition in adults, we analyzed potential differences in CD40, BLK and BANK1 genotype and allele frequencies between patients with IgAV stratified according to the age at disease onset. As shown in Table 3, no statistically significant CD40, BLK and BANK1 differences were found in children when compared to adults. Moreover, we analyzed CD40, BLK and BANK1 genotype and allele frequencies between patients with IgAV stratified according to the presence/absence of GI or renal manifestations. In this regard, similar CD40, BLK and BANK1 frequencies were observed when IgAV patients were stratified according to the presence/absence of GI manifestations (Table 3). This was also the case when patients with IgAV who developed renal manifestations were compared to those without these complications (Table 3).
Moreover, we investigated whether CD40, BLK and BANK1 haplotype frequencies differed between IgAV patients and controls as well as between IgAV patients stratified according to the specific clinical characteristics of the disease above mentioned. In this regard, no statistically significant CD40, BLK and BANK1 haplotypes differences were disclosed in patients with IgAV when compared to healthy controls (Table 4). Likewise, CD40, BLK and BANK1 haplotype frequencies were similar between IgAV patients stratified according to the age at disease onset or to the presence/absence of GI or renal manifestations (Table 5).
B-lymphocytes play essential functions in regulating immune responses [37], being rigorously regulated with respect to both development and activation [1]. Abnormalities in these processes contribute to the pathogenesis of autoimmune diseases [38]. Cumulative knowledge clearly demonstrated that CD40, BLK and BANK1 are key proteins involved in the development and signaling of B-cells [2,3,4,5,6,7,8]. Additionally, CD40, BLK and BANK1 genes are identified as shared susceptibility loci for several inflammatory diseases [7,8,9,10,11,12,13,14,15,16,17,18,19]. Taking these considerations into account, we evaluated whether CD40, BLK and BANK1 are also implicated in the pathogenesis of IgAV, predominantly a B-cell inflammatory condition, involving small blood vessels. For that purpose, three polymorphisms within CD40 and BLK as well as two genetic BANK1 variants, previously associated with several inflammatory diseases [7,8,9,10,11,12,13,14,15,16,17,18,19], were evaluated in the largest series of Caucasian patients with IgAV ever assessed for genetic studies. Some of these variants also exhibit different functional consequences [8,18,19,36]. In particular, the CD40 rs1883832C allele influences the translational efficiency of nascent CD40 mRNA transcripts, resulting in elevated CD40 levels [8]. BLK rs2736340 is in tight linkage disequilibrium with rs13277113 (D’ = 1, r2 = 0.99 in Europeans), wherein the A allele is associated with lower levels of mRNA BLK [36]. Finally, BANK1 rs10516487 leads to a substitution of Arg to His at amino-acid position 61 of BANK1 protein, whereas BANK1 rs3733197 causes an Ala to Thr substitution at amino-acid position 383 (creating a site for threonine kinases that affects the B-cell signaling) [18,19]. Interestingly, our findings revealed no influence of CD40, BLK and BANK1 on IgAV susceptibility when we studied each of the polymorphisms separately or together, conforming haplotypes. Several studies described the influence of different polymorphisms on the increased risk of nephritis or GI disease in IgAV [39,40,41,42]. Accordingly, we also evaluated the potential association of CD40, BLK and BANK1 with the increased risk of nephritis or GI complications in our study. Nevertheless, our results do not support a role of CD40, BLK and BANK1 variants with clinical features of IgAV, suggesting that these genes do not contribute to IgAV severity. Notwithstanding, our results do not exclude the potential implication of other polymorphisms related to B-cells in the pathogenesis of IgAV. Consequently, further studies are needed to clarify this issue. Shared molecular mechanisms among different vasculitides have been described [43,44]. However, as observed in our series of IgAV, no association of BANK1 rs10516487 and BANK1 rs3733197 variants with the susceptibility and clinical expression of patients with giant cell arteritis [45], a primary systemic vasculitis that involves large- and middle-sized blood vessels in people older than 50 years, was disclosed. It was also the case for the potential implication of BLK rs2736340 and BANK1 rs10516487 in Takayasu arteritis [46], another primary large-vessel vasculitis that involves mainly young individuals. In contrast, a potential influence of CD40 rs1883832 [20] and BLK rs2736340 [21] polymorphisms was previously reported on the development of ischemic manifestations in patients with giant cell arteritis. Similarly, CD40 rs4813003 [13], BLK rs2254546 [13], BLK rs2736340 [22] and BLK rs2618476 [22] were identified as susceptibility markers for Kawasaki disease, a vasculitis affecting small- and medium-sized arteries. Our findings suggest that IgAV may not be a state of increased B-cell activation, pointing to IgAV as a different entity from other types of vasculitis. In keeping with our results, genome-wide association studies in IgA nephritis, which is pathophysiologically similar to IgAV [47,48], have not identified CD40, BLK and BANK1 genes as significant players in the pathogenesis of the disease [49,50,51]. With respect to this, genes affecting the mucosal immune defense, having an impact on IgA production by plasma cells in mucosa and previously reported as susceptibility loci in IgA nephropathy [51], may also be implicated in the pathogenesis of IgAV, and further studies assessing this issue would be of potential interest.
In summary, although complex genetic interactions appear to be involved in the pathogenesis of IgAV, based on a large series of patients, we could not observe a contribution of the CD40, BLK and BANK1 genes to the genetic network underlying this small-vessel vasculitis. Further studies should be performed to fully explore the role of B-cells in the pathogenesis of IgAV. | true | true | true |
PMC9572342 | 36235125 | Sara Julietta Lozano-Herrera,Gabriel Luna-Bárcenas,Ramón Gerardo Guevara-González,Rocio Campos-Vega,Juan Carlos Solís-Sáinz,Ana Gabriela Hernández-Puga,Haydé Azeneth Vergara-Castañeda | Fermentation Extract of Naringenin Increases the Expression of Estrogenic Receptor β and Modulates Genes Related to the p53 Signalling Pathway, miR-200c and miR-141 in Human Colon Cancer Cells Exposed to BPA | 05-10-2022 | colon cancer,ERβ,BPA,fermented extract of naringenin,miR-200c,miR-141 | The estrogenic receptor beta (ERβ) protects against carcinogenesis by stimulating apoptosis. Bisphenol A (BPA) is related to promoting cancer, and naringenin has chemoprotective activities both can bind to ERβ. Naringenin in the colon is metabolized by the microbiota. Cancer involves genetic and epigenetic mechanisms, including miRNAs. The objective of the present study was to evaluate the co-exposure effect of colonic in vitro fermented extract of naringenin (FEN) and BPA, to elucidate molecular effects in HT-29 colon cancer cell line. For this, we quantified genes related to the p53 signaling pathway as well as ERβ, miR-200c, and miR-141. As an important result, naringenin (IC50 250 µM) and FEN (IC50 37%) promoted intrinsic pathways of apoptosis through phosphatase and tensin homolog (PTEN) (+2.70, +1.72-fold, respectively) and CASP9 (+3.99, +2.03-fold, respectively) expression. BPA decreased the expression of PTEN (−3.46-fold) gene regulated by miR-200. We suggest that once co-exposed, cells undergo a greater stress forcing them to mediate other extrinsic apoptosis mechanisms associated with death domain FASL. In turn, these findings are related to the increase of ERβ (5.3-fold with naringenin and 13.67-fold with FEN) gene expression, important in the inhibition of carcinogenic development. | Fermentation Extract of Naringenin Increases the Expression of Estrogenic Receptor β and Modulates Genes Related to the p53 Signalling Pathway, miR-200c and miR-141 in Human Colon Cancer Cells Exposed to BPA
The estrogenic receptor beta (ERβ) protects against carcinogenesis by stimulating apoptosis. Bisphenol A (BPA) is related to promoting cancer, and naringenin has chemoprotective activities both can bind to ERβ. Naringenin in the colon is metabolized by the microbiota. Cancer involves genetic and epigenetic mechanisms, including miRNAs. The objective of the present study was to evaluate the co-exposure effect of colonic in vitro fermented extract of naringenin (FEN) and BPA, to elucidate molecular effects in HT-29 colon cancer cell line. For this, we quantified genes related to the p53 signaling pathway as well as ERβ, miR-200c, and miR-141. As an important result, naringenin (IC50 250 µM) and FEN (IC50 37%) promoted intrinsic pathways of apoptosis through phosphatase and tensin homolog (PTEN) (+2.70, +1.72-fold, respectively) and CASP9 (+3.99, +2.03-fold, respectively) expression. BPA decreased the expression of PTEN (−3.46-fold) gene regulated by miR-200. We suggest that once co-exposed, cells undergo a greater stress forcing them to mediate other extrinsic apoptosis mechanisms associated with death domain FASL. In turn, these findings are related to the increase of ERβ (5.3-fold with naringenin and 13.67-fold with FEN) gene expression, important in the inhibition of carcinogenic development.
Colon cancer is one of the most common cancers worldwide, ranking third after breast cancer in women, prostate cancer in men, and lung cancer in both sexes [1,2]. This type of cancer is inversely related to dietary habits, especially the consumption of fruits and vegetables rich in fiber and antioxidant compounds [3]. Among antioxidants, flavonoids, especially flavonones such as naringenin, have become known for their antioxidant and anti-inflammatory effects against colon cancer through different mechanisms such as transactivation of estrogen receptors (ER) [4,5,6,7,8]. There are two different isotypes of ER, α (ERα) and β (ERβ), encoded by the ESR1 and ESR2 genes, respectively. The function of ERβ is related to apoptotic processes that reduce carcinogenic progress. ERα is associated with anti-apoptotic processes. Therefore, modulation of ERs activity and expression by estrogenic disruptors could have a preventive or promoting effect on colon cancer [9,10]. Some natural estrogenic disruptors, such as naringenin, can bind to ERβ, promoting its own expression and apoptotic processes, which has a protective effect on the colon. However, other disruptors of synthetic origin, such as BPA, a compound used to polymerize plastics used in various products such as food containers, can migrate into the food matrix and be ingested by the consumer, promoting colon cancer because they have an antagonistic effect on ERβ in the presence of estradiol [11,12]. Interestingly, simultaneous exposure to both types of disruptors, naringenin and BPA, has been studied in the development of breast cancer, and naringenin had an antagonistic effect on ERα associated with proliferation processes even in the presence of simultaneous exposure to BPA [13,14,15]. The antiproliferative or apoptotic molecular mechanisms of ER are diverse, including signalling associated with the p53 pathway; there are also epigenetic mechanisms such as miRNAs that can interfere with the expression of genes. The miRNAs are small RNA fragments with a length of 18–25 nucleotides that primarily inhibit the translation process by binding to the 3′UTR region of mRNA. Their activity is associated with the promotion of carcinogenesis once they inhibit the translation of tumour suppressor genes. In this regard, miR-200c has been reported to play a key role in suppressing tumour progression by inhibiting epithelial–mesenchymal transition and metastasis in various cancer [16]. Naringenin is a flavonoid that belongs to the subclass of flavanones. It is found in citrus fruits such as oranges, grapefruits, and tomatoes. After digestion, a large amount of naringenin enters the colon, where it is metabolised by the resident microbiota through a fermentation process. This process produces by-products such as hippuric acid, 3(3-hydroxyphenyl) propionic acid, and 3(3,4-dihydroxyphenyl) propionic acid, or phloroglucinol, which have been reported previously [17,18,19]. These compounds have antioxidant and anti-inflammatory activity [20,21]. It is therefore hypothesized that consumption of foods rich in flavonoids such as naringenin could attenuate the negative effects of BPA, but information on their effects as estrogenic disruptors is lacking. Therefore, the aim of this study was to investigate the anticarcinogenic effect of naringenin and its fermented extract from colon upon simultaneous exposure to the disruptor BPA in human adenocarcinoma cells HT-29, and to elucidate the molecular mechanisms involved.
FEN was analyzed by the method UPLC-MS, identifying naringenin and other compounds such as 3(3-hydroxyphenyl) propionic acid (3-HPPA), apigenin, phenylacetic acid, and secoisolariciresinol (Table 1). Nevertheless, only apigenin and 3-HPPA can be considered as by-products of naringenin, since the other two compounds were also detected in the blank, the culture medium used for fermentation. Several authors have found dehydroxylations, hydroxylations, and deglycosylations of phenolic compounds generated by microbial metabolism to produce various byproducts, as we show (Figure 1) [22,23].
The antioxidant capacity of both naringenin and FEN is shown in Table 1. The results show a decrease in the antioxidant capacity of FEN by DPPH, a synthetic radical that is not similar to the conditions prevailing in the organism; therefore, the ORAC method is the most accurate to obtain information at the physiological level. Regarding the ORAC method, there is no statistical difference between the antioxidant capacity of FEN and naringenin, so they have similar biological capacity to transfer hydrogen atoms.
Cell viability experiments were performed to determine the effects of naringenin and FEN on HT-29 cells. Cells treated with naringenin showed a decrease in cell viability with a dose–response effect. Naringenin successfully suppressed cell growth, with an IC50 value of 250 µM at 24 h of exposure time (Figure 2a). To determine the IC50 of FEN (37%), cell viability was measured at concentrations of 25–43% of FEN (Figure 2b) after 24 h and a linear decrease in cell viability was also observed. Cells treated with BPA showed no significant effect on cell viability under our study conditions, but the responses at the molecular level were different, but cell viability decreased as much with simultaneous exposure to BPA (4.4 µM) + naringenin or +FEN, as it did without BPA (Figure 3).
Flow cytometry assays with Annexin V were performed to evaluate cell death. Naringenin (250 µM) induces death in 52 ± 0.7% of cells, of which 41 ± 0.7% correspond to apoptosis, with or without co-exposition to BPA (4.4 µM), and 7 ± 1.3% of cells die by necrosis (Figure 4). Cells treated with FEN maintain the trend of viability without (54 ± 0.3%) or with BPA (59 ± 0.7%), dying mainly by apoptosis (39 ± 1.5 and 37 ± 1.5%; respectively) (Figure 4). In addition, when the cells were treated with BPA alone, 92% of cells died by apoptosis and the other 7% died by necrosis, a type of cell death that causes cell damage and inflammatory processes. To confirm cell death by necrosis, LDH assay, an indirect biomarker for necrosis, was performed (Figure 5). Necrosis was detected in 7–12% of cells under the different treatments, which is consistent with our results from the flow cytometer.
To determine the modulation of the endogenous antioxidant system by the experimental treatments, the activity of SOD enzyme and the amount of GSH were measured in the previously described treatments. Cells treated with FEN showed the highest percentage of O2 inhibition, consistent with the high activity of SOD (13.18 ± 2.0%) (Table 2). On the other hand, GSH, the most potent antioxidant in the body, was significantly decreased by naringenin and naringenin + BPA (110.75 ± 1.5 and 129.19 ± 0.9 Nmol/mL, respectively) (Table 2), suggesting that GSH, as an electron donor, undergoes oxidation–reduction reactions to neutralize free radicals generated by oxidative stress in the cell.
The relative mRNA expression of ERβ and GPR30 in HT-29 cells is shown in Figure 6. Our results show that naringenin and FEN treatments increase the ERβ expression by 5.3 ± 0.77-fold and 13.67 ± 2-fold, respectively, while BPA-treated cells increase by 2.53 ± 0.2 fold compared with the negative control. On the other hand, GPR30 is an estrogen-sensitive G protein-coupled receptor that can trigger various signalling pathways such as proliferation, apoptosis, and cell migration. In the current study, negative transcriptional regulation of GPR30 by naringenin and BPA was detected (Figure 6b). The decrease in GPR30 receptor expression needs further investigation to identify the triggering mechanism. The expression of genes related to the p53 pathway was determined using a real-time PCR array, and the results are presented in the form of fold-change of expression levels (Table 3). As shown, some important genes involved in the apoptosis process are up- or down-regulated. This is the case with the overexpression of the Caspase 9 gene induced by the treatments with naringenin (3.99-fold), FEN (2.03-fold), naringenin + BPA (2.36-fold), and FEN + BPA (12.27-fold). Bcl-2 is an antiapoptotic gene whose expression was decreased by naringenin (-2.17-fold) but not by FEN (1.70-fold). It was increased by naringenin + BPA (40.54-fold) and FEN + BPA (146.89-fold) treatments, suggesting activation of BPA in co-exposure responsive mechanisms to prevent apoptosis, which are triggered only in the presence of naringenin or its metabolites. Activation of both extrinsic and intrinsic apoptosis pathways is suggested by the co-exposure treatments in the present study, as an increase in FasLG gene expression with naringenin + BPA (35.45-fold), FEN (4.63-fold), and FEN + BPA (52.37-fold), the Fas gene in cells treated with naringenin + BPA (12.09-fold), FEN + BPA (15.3-fold), and FADD (1.19-fold) and (2.05-fold), respectively. Another gene involved in apoptosis is caspase-2 and CRADD [24]. Naringenin increased the expression of caspase-2 (2.14-fold) and CRADD (2.96-fold). In addition, TNFRSF10D was also overexpressed by naringenin (4.98-fold), naringenin + BPA (14.79-fold) and FEN + BPA (9.7-fold). PTEN is involved in the modulation of several cancer processes, including apoptosis [25,26]. In the current study, a reduction in PTEN was observed with BPA treatment (-3.46-fold), but co-exposure with naringenin (1.20-fold) or FEN (6.98-fold). Co-exposure treatments naringenin + BPA (21.69-fold), FEN + BPA (36.22-fold), and FEN (2.85-fold), increased MDM2 gene expression. In addition, overexpression of the TP73 gene was observed among treatments with naringenin (3.99-fold), naringenin + BPA (75.89-fold), FEN (2.65-fold), and FEN + BPA (101.66-fold). Once activated, TP73 can bind to PTEN, which was overexpressed under naringenin treatment (2.70-fold), FEN (1.72-fold) and down-regulated under BPA treatment (-3.46-fold); this binding activates BBC3 (for Bcl-2 binding component 3 or PUMA), which is negatively regulated by BPA (−4.96-fold). Our results show an up-regulation of MLH-1 by naringenin treatment (4.48-fold) and a down-regulation by BPA (−7.94-fold). In this study, we also found a decrease in ATR expression in HT-29 cells treated with BPA (-2.16-fold). Therefore, this might not have a beneficial effect on cell cycle arrest; whereas co-exposed to naringenin (3.28-fold) or FEN (5.63-fold) resulted in an increase in expression, suggesting a contribution to cell cycle arrest, as previously informed for apigenin, which is present in FEN. In addition, RPRM is overexpressed in BPA-treated cells in the current study (3.63-fold), and even more strongly in naringenin + BPA (63.17-fold), FEN (5.11-fold), and FEN + BPA (33.01-fold) treatments. MYOD is a gene involved in this process. Our results showed that cells treated with naringenin (3.99-fold), naringenin + BPA (75.89-fold), FEN (5.20-fold), and FEN + BPA (26.82-fold), had an increase in MYOD expression, while conversely BPA treatment (-1.10-fold) decreased MYOD expression (−1.10-fold). The results support intrinsic and extrinsic apoptosis in cells treated with naringenin and FEN and suggest the relationship with ER. On the other hand, they support the malignant mechanisms of BPA treatment.
Regarding to miRNAs expression, both miR-200c and miR-141, which belong to the miR-200 cluster were analyzed (Figure 7). An increase in the expression of miR-200c was observed under BPA treatment (2.94 ± 0.44-fold) and a lower expression of miR-141 was observed under FEN treatment (0.01 ± 0.01 fold). The RT2 Profiler PCR Data Analysis software from GeneGlobe, Qiagen, suggests PTEN as a target gene, and we ratify with miRmap software. This software presents the results in an algorithm that is thermodynamic, evolutionary, probabilistic, and sequence-based, finally giving a score from 0–100. Our miRNAs have 63.52 (miR-200c) and 90 (miR-141) with PTEN gen0065 (Figure 8).
The results obtained in this study show the generation of 3-HPPA by c-ring cleavage and of apigenin by dehydrogenation of naringenin during the fermentation process. The formation of these compounds and others such as apiferol, eriodyetol, or some hydroxycinnamic acids by hydroxylations, hydrogenations, hydrolysis, or methylations depends on the composition of the microbiome in the colon. The above findings are relevant because these compounds have antioxidant capacities and activities associated with disease prevention [22,23,27]. FEN exhibited a lower antioxidant capacity than naringenin as measured by DPPH. This is probably because the by-products found in FEN differ from naringenin in the amount and position of the hydroxyl groups. Apigenin, a byproduct of FEN, has a higher antioxidant potential than naringenin because of the 2,3-dehydrogenation reaction leading to conjugative C=C binding. However, the interaction of different metabolites leads to antagonism that reduces the ability to donate electrons to the DPPH radical, which is the fundamental basis of the assay. Second, in the evaluation of FEN and naringenin by ORAC method, which is based on the ability of the molecule to transfer hydrogen atoms to oxygen radicals, there is no difference between FEN and naringenin [28,29]. Thus, it can be seen that naringenin has a great ability to transfer electrons and has the same ability to transfer hydrogen atoms to oxygen radicals as the group of by-products formed after fermentation. The antioxidant capacity of naringenin shows how it can interact with some proteins, either by reduction ROS, chelation of metals, or in a pro-oxidant manner, which in a cell metabolism altered by cancer would promote apoptosis and eventually inhibition of cell proliferation. In addition to these mechanisms, naringenin has shown several others by which it inhibits proliferation. These include some signaling pathways such as EGFR/MAPK, Akt, and the Wnt/β-catenin pathway, among others [30,31]. Naringenin (IC50 between 180–360 µM) has been shown to have an inhibitory effect on the proliferation of HT-29 colon cancer cells [32]. Currently, there is no evidence that FEN has the same effect. This is the first study to investigate the antioxidant capacity of the fermentation extract of naringenin in colon cancer cells. After gastrointestinal digestion, about 84% of naringenin enters the colon bound to an indigestible carbohydrate or in free form. Moreover, naringenin is one of the by-products of quercetin, another widely used flavonoid, so its transformation during the fermentative process is very important to determine an effect closer to the biological [33,34]. Some authors have reported the formation of diverse by-products such as 3-HPPA, which has been shown to have antioxidant and anti-inflammatory effects and also to promote apoptotic processes [34]. Within the antioxidant capacity, the activity of enzyme systems is very important, one of them is SOD, whose biological function is to dismute the peroxide anion into hydrogen peroxide. There are three SOD isoenzymes, Cu/Zn SOD (SOD1), which is found in greater quantity in the cytoplasm, Mn SOD (SOD2), located in the mitochondria, and EcSOD (SOD3), which is found mainly in the extracellular space. SOD3 is associated with inhibition of tumour growth and metastasis, while SOD1 and SOD2 are elevated in different stages of several cancers, including colon cancer, and can trigger proliferative and apoptotic signals depending on the amount of H2O2 produced [35,36]. In our assay, the total activity of SOD was measured, so we hypothesize that the increase in activity leads to H2O2 formation that promotes apoptosis under the FEN treatment. Another antioxidant system is GSH. Although it is one of the most potent antioxidant systems, large amounts of GSH have been associated with metastasis and chemoresistance in cancer, as an adaptive response of the cancer cell to large amounts of free radicals. We hypothesize that GSH depletion after naringenin treatment, as reported by other authors, is mediated by multidrug resistance protein 1 (MRP1), which promotes increased sensitivity of cancer cells [6,30,37,38]. In the cells treated with BPA, it was observed that the viability of the cells did not decrease significantly and the cells that diet were mostly due to necrosis. Previously, it was reported that testicular cancer cells treated with BPA (0.01–10 µM µM) and exposed for 24 h maintained the same trend of decreasing cell proliferation (>25%); however, at concentrations of 10 µM an increase in cell migration (wound assay) was detected, in addition to increased expression of 50 migration-related proteins such as Gal-1, supporting the results of the current study, in which molecular responses of migration were observed under BPA treatment [39]. On the other hand, naringenin and FEN promoted apoptosis related to signalling pathway. Other reports suggest that naringenin (10 µM) induces apoptosis via the p38/MAPK pathway, which is modulated by ERβ agonist activation in DLD-1 colon cancer cells [40,41]. Moreover, simultaneous exposure of naringenin (100 µM) and BPA (10 µM) in breast cancer cells has shown an apoptotic effect due to a decrease in anti-apoptotic Bcl-2 proteins, mechanisms that may also play a role in the present study [42]. Moreover, naringenin and the by-products of FEN could promote the formation of ROS, which is related to the induction of the apoptotic process [43,44]. The differential expression of ERβ under naringenin and FEN treatment is the result of the position at which the ligands bind to the receptors and determine their positive or negative modulation. Kuiper et al. described in terms of relative affinity units (RBA) with respect to 17β-estradiol, that BPA had similar affinity for both estrogen receptor α and β, naringenin had higher affinity for ERβ and similarly apigenin, which is a component of FEN, so it can be assumed that the affinity of flavonoids for this receptor positively modulates its expression [45]. The reduction of ERβ expression has been associated with poor prognosis in cancer, since its main function is to promote the apoptotic process. [46,47] The lack of regulation of ERβ mRNA by BPA treatment is consistent with the reports of Hess–Wilson et al. on 1nM BPA-treated LNCaP prostate cancer cells [48]. In our study, treatment had no effect on GPR30. Dong et al. showed that BPA (10 µM) in breast cancer cells activates the Erk1/2 signalling pathway and transcriptional regulation of c-fos through GPR30 [49]. In addition, the flavonoid compounds genistein and baicalein were reported to activate GPR30, suggesting an antiproliferative effect. [50,51] Under the conditions studied, this receptor is not related to the results obtained. We found high expression of caspase 9 in naringenin and FEN, suggesting that mitochondrial apoptosis may be activated by the stimulation of ROS, which may be related to the increase seen in SOD activity by FEN and the low amount of GSH by naringenin. In addition, naringenin decreases BCL-2 expression. Kang et al. demonstrated the fisetin, a flavone with antiangiogenic and antioxidant activity, increases caspase-9 activity and decreased Bcl-2 protein expression in NCI-H460 lung cancer cells at 75 μM for 24 h [33]. Our results show extrinsic apoptosis when BPA is co-exposed to naringenin or FEN. Lee et al. observed that kaempferol (40 and 60 μM, 48 h, HT-29 colon cancer cell line) promotes apoptosis through both extrinsic and intrinsic pathways. This process occurs through an increase in FAS-L protein expression, which activates caspase-8 and eventually leads to cleavage of Bid and activation of caspase-3, and through the increase in released cytochrome C, which activates caspase-9 and caspase-3 [52]. FASLG is a gene that is not constitutively expressed in the HT-29 cell line, and tumour necrosis factor alpha (TNF-α) has been shown to induce the expression of functional FasL on the cell surface of HT29 cells, so it may be able to induce apoptosis through this pathway. [53] This is related to the results shown, in which an increase in TNF expression is observed under treatments with naringenin + BPA, FEN, and FEN + BPA. Other studies reported that polyphenols such as crocetin (100 μM) promote the FAS/FADD interaction in HT29 cells in a p73-dependent manner. This interaction activates BID, which subsequently activates apoptosis via mitochondrial [54]. Another gene involved in apoptosis is caspase-2, which encodes a protein involved in apoptotic and inflammatory mechanisms. CRADD is a gene associated with the pathway triggered by caspase-2. It encodes the protein containing the caspase recruitment domain and death domain (DD) (also known as RAIDD), which together with PIDD1 are necessary for the activation of caspase-2 as a pro-apoptotic protein. [24,54,55,56,57] In addition, naringenin increases the expression of caspase-2 and CRADD. Moreover, TNFRSF10D was also overexpressed by naringenin, naringenin + BPA, and FEN + BPA. TNFRSF10D is a gene encoding TRAILR1 receptor, which is associated with the promotion of proliferation processes as apoptosis and could regulate the extrinsic pathway of apoptosis [24,53]. Another important gene for promoting apoptosis is p53. This gen is post-translational activated by cellular stress regulated by MDM2 in the signalling cascade of mitochondrial apoptosis, which increases transcription of MDM2, resulting in inhibition of p53 activity. However, the p53 gene is mutated in the HT-29 cell line, causing the cell to activate other signalling pathways to reduce its viability [58]. Zeng et al. showed that MDM2 can modulate the activation of TP73 but not its expression in cells with mutated p53 gene [59]. The co-exposure treatments naringenin + BPA, FEN + BPA, and FEN increased MDM2 gene expression. In addition, overexpression of TP73 gene was observed among naringenin, naringenin + BPA, FEN, and FEN + BPA. Dabiri et al. investigated the role of TP73 in HT-29 cells and found the presence of TP73 in both the nucleus and cytosol of cells treated with bortezomil, a chemotherapeutic agent used as a proteasome inhibitor and promoter of cellular cycling and apoptotic death. The authors suggested that the role of TP73 in cellular apoptosis was due to the presence of a mutation in p53 upon treatment with naringenin and FEN [60]. Once TP73 is activated, it can bind to PTEN, which was overexpressed under naringenin treatment, FEN and down-regulated under BPA treatment; this binding activates BBC3 (for Bcl-2 binding component 3 or PUMA), which is negatively regulated by BPA. BBC3 is a proapoptotic gene encoding the BBC3 protein located in mitochondria and released into the cytosol by mitochondrial permeabilization along with other proapopototic molecules and cytochrome C, dependent or independent of p53 [61]. In HT-29 cells harbouring a p53 mutation, its activation by Sp1 and p73 was detected [62]. Tili et al. found overexpression of PTEN protein in SW480 cells treated with resveratrol (50 μM), which triggered apoptosis processes, similar to those observed when treated with naringenin and FEN [63]. On the other hand, Li et al. reported a decrease in PTEN protein expression in breast cancer cells treated with 10 μM BPA, related to an increase in cell proliferation and a higher percentage of cells in the S phase of the cell cycle, while 1 μM curcumin counteracted this effect [8]. In the current study, a reduction in PTEN was observed with BPA treatment, but the co-exposure with naringenin, or FEN, offsets this effect. PTEN is also involved in the modulation of cell adhesion, migration, and invasion through inhibition of the adaptor protein Shc and the focal adhesion protein kinase Fak. [64,65,66,67,68] Overexpression of the PTEN gene by treatments with naringenin and FEN suggests a decrease in the metastatic phenotype of tumour cells through inhibition of cell migration. We hypothesise that the low expression of PTEN in BPA treatments is due to epigenetic modulation, i.e., high expression of miR-200c or miR-141. Chen et al. demonstrated the binding of miR-200 homologs to the 3’UTR region of PTEN in endometrial cancer cells using a luciferase assay, and the importance of E2 in regulating miR-200c expression was also demonstrated via ERα overexpression, which is important for the development of this type of cancer [65]. Under our experimental conditions, BPA induced miR-200c expression and PTEN gene repression, so it could increase cellular proliferation through a mechanism modulated at the molecular level by increasing ERα expression (ESR1). Mei et al. demonstrated the immunosuppressive role of miR-200c through the inhibition of PTEN and the increase of MDSCs (ROS in myeloid-derived suppressor cells), which suppress the antitumor immune response, suggesting various negative effects of BPA through the suppression of PTEN mediated by miR-200c, such as the inhibition of apoptosis, the increase of cell cycle and immune responses [66]. PTEN is also a target gene of miR-141, and studies performed on cells from patients with nasopharyngeal carcinoma have shown that miR-141 expression decreases PTEN expression of after treatment with cisplatin, related to chemoresistance [67]. The MLH-1 gene is one of the genes belonging to MMR, which encodes proteins that recognize and repair errors in microsatellite sequences of newly synthesized DNA. Therefore, its downregulation by methylation or mutation is common in colon cancer [69]. Our results show up-regulation of MLH-1 up-regulation by naringenin treatment and down-regulation by BPA. Lu et al. observed an increase in MLH-1 gene in HT-29 and SW480 cells treated with 17β-estradiol, suggesting a role of ERβ as an upstream mechanism [70]. We suggest positive regulation of MLH-1 by ERβ under naringenin treatment. DNA damage is also detected by the ATM-Rad3-Related gene (ATR), whose activation is arrested at G1, S, or G2 stages of the cell cycle, a process mediated by the action of the checkpoint kinases CHK1 and CHK2 [71]. In this study, we also found a decrease in ATR expression in HT-29 cells treated with BPA. Therefore, this might not have a positive effect on cell cycle arrest, while the co-exposure to naringenin or FEN resulted in an increase in expression, suggesting a contribution to cell cycle arrest as previously informed for apigenin, which is present in FEN. On the other hand, Reprimo (RPRM) is a gene encoding a protein involved in cell cycle arrest in G2/M phase by regulating the activity of Cdc2 and cyclin B1 in a p53 and p73- dependent manner, and its loss of expression has been associated with more invasive stages of gastric cancer [72,73]. In the present study, all treatments except BPA, increased the expression of RPRM, which mediates cell cycle arrest. In addition, hypermethylation of RPRM related to ERα (ESR1) has been observed in several cancers such as breast cancer [74,75,76,77], which is overexpressed in cells treated with BPA in the present study, and even more so in treatments with naringenin + BPA, FEN, and FEN + BPA. ERα expression, implicated in the development and progression of colon cancer, has been associated with lower survival in epidemiological studies [78]. Huang et al. have evaluated different concentrations of BPA (0.1–1000 nM) in prostate epithelial cells and reported 98% viability but with an increase in the expression of both α and β estrogenic receptors [79]. In the present study, the increase in ERα gene was found to be under BPA treatment, which is explained by the affinity that naringenin has for the ERα as the apinenin contained in FEN. Ye et al. support in their in silico analysis that the number of hydroxyl groups and their position affect the affinity for the receptor, probably the other naringenin byproducts present in FEN have synergy with the receptor [80]. It is important to highlight that ERβ expression is predominant in differentiated normal colon epithelium, but its expression decreases due to the hypoxic microenvironment in the colon when malignancy resulting from cancer progresses [81]. Therefore, naringenin and FEN treatments induced ERβ expression (Figure 6), which could promote apoptotic processes (Figure 4), even if there is an increase in ERα expression. SIRT1 is a nicotinamide adenosine dinucleotide (NAD/NADH) dependent histone deacetylase (HDAC) that acts by deacetylation of histones (H1, H3, and H4) and is associated with gene expression of ER [80,81]. Transcription of SIRT1 is regulated by FOXO3A, p53, E2F1 and other transcription factors [80]. This explains the increase in E2F1 and the increase in SIRT1 in co-exposure treatments. In ovarian cancer cells, loss of SIRT1 mRNA expression is associated with higher expression of ERβ [82]. Therefore, SIRT1 could possibly act as a coactivator or repressor of the different ERs depending on the cell line and its conditions, which could support the low expression of SIRT1 gene with the low expression of ERβ and high expression of ERα genes under the treatment with BPA. However, this mechanism does not apply to the other treatments; therefore, the role of SIRT1 is not fully elucidated [82,83,84]. In addition to the processes of apoptosis and cell cycle inhibition, the inhibition of metastasis is an important mechanism mediated by flavonoids. MYOD is a gene involved in this process by decreasing mRNA expression of E-cadherin and vimentin, both promoters of epithelial-mesenchymal transition (EMT) [85]. Our results showed that cells treated with naringenin, naringenin + BPA, FEN and FEN + BPA exhibited an increase in MYOD expression, and conversely, BPA treatment decreased MYOD expression. The promotion of metastasis mediated by BPA and its inhibition by naringenin and FEN in HT-29 cells is thus demonstrated in addition to the PTEN gene results.
Naringenin (#N5893), BPA (#239658), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (#M2128), casein peptone, HCl, NaCl, KCl, MgSO4, KH2PO4, NaHCO3, CaCl2, NaOH, pepsin, pancreatin, Tween salts, hematin, glucose, crystal violet, formaldehyde, ethanol, LDH assay kits (#MAK066), SOD assay kits (#19–160), were purchased from Sigma-Aldrich Co. (St. Louis, MO, USA). HT-29 cells were purchased from the American Type Culture Collection (ATCC). Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum (FBS), antibiotic-antimycotic, trypsin, and ethylenediaminetetraacetic acid (EDTA) were purchased from Gibco or Thermo Fisher Scientific, Inc. (Waltham, MA, USA). Anexin-V kit (#602128), were purchased from Millipore, Merk. Silica-gel columns kit (#PP-210S) and SCRIPT cDNA kit (#77176) were purchased from Jena Bioscience GmbH (Jena, Germany). Radiant TM Green Hi-ROX qPCR kit (#QS2005) were purchased from Radiant molecules, Alkali Scientific Inc. Finally, the Human RT2 Profiler Real Time PCR Array System (PAHS-027A) was purchased from Qiagen, SABiosciences, USA.
In vitro simulation of the digestion process of naringenin in the mouth, stomach, and small intestine was performed according to the procedure of Campos-Vega et al [86]. Briefly, 5 mL of saliva was collected from four volunteers, two males and two females, healthy, with no chronic degenerative diseases or contact infections in the past 30 days, no use of antibiotics, medications in general, or alcohol in the past 20 days. Subjects had to be of legal age and not eating a vegetarian or keto diet. The sample was diluted with 10 mL of distilled water. The samples were incubated with 1 g naringenin for 5 min. Samples were then adjusted to a pH of 2.0 with a 150 mM HCl solution to simulate gastrointestinal tract conditions. Pepsin (0.055 g) was dissolved in 0.94 mL of 20 mM HCl solution, and the samples were incubated at 37 °C for 2 h. Later, intestinal conditions were simulated with 3 mg of bovine bile and 2.6 mg pancreatin mixed with 5 mL of Krebs-Ringer buffer (118 mM NaCl, 4.7 mM KCl, 1.2 mM, MgSO4, 1.2 mM KH2PO4, 25 mM NaHCO3, 11 mM glucose and 2.5 mM CaCl2; pH 6.8). This solution was gasified with a gas mixture (10:10:80, H2:CO2:N2). The fermentation process in the colon was simulated according to the method of Campos-Vega et al [87]. Briefly, human stool inoculum from healthy donors (two males and two females) was used. Three grams of the stool sample was homogenised with 27 mL of phosphate buffer pH 7.0. One millilitre of the stool solution was collected and added to sterile basal culture medium pH 7.0 containing: 2 g/L peptone water, 2 g/L yeast extract, 0.1 g/L NaCl, 0.04 g/L K2HPO4, 0.04 g/L KH2PO4, 0.01 g/L MgSO4-7H2O, 0.01 g/L CaCl2-2H2O, 0.01 g/L NaHC32, 0.5 g/L L-cysteine, 0.5 g/L bile salts, 2 mL/L Tween-80, and 5 mL hematin solution (0.2 g dissolved in 5 mL NaOH). Finally, 1 mL of the sample obtained from in vitro gastrointestinal digestion was added, and the samples were placed in a fermenter containing a gas mixture (10:10:80, H2:CO2:N2) and allowed to ferment for 24 h at 37 °C, using raffinose as a fermentation control (100 mg, R0514, Sigma-Aldrich, St. Louis, MO, USA) [87]. The obtained extract was designated as fermentation extract of naringenin (FEN) and the basal nutrient medium as blank.
Fermentation-derived metabolites were identified according to the modified methodology proposed by Abu-Reidah et al [88]., using ultrahigh performance liquid chromatography (UPLC) (Waters UPLC Acquity I-class with PDA detector) coupled to a time-of-flight mass spectrometer VION QtoF. Leucine enkephalin was injected at 200 pg/μL at a flow rate of 10 μL/min. Injections were performed in MSe mode. Data were analysed using the UNIFI 1.9 program with Small Molecule Option.
The trolox equivalent antioxidant capacity (TEAC) of FEN, blank, of FEN and naringenin was evaluated by ORAC and DPPH assay. For the DPPH assay, the sample was incubated in the dark for 30 min and the absorbance was measured at 525 nm [89]. The ORAC method is based on the difference in the decay of fluorescein between the blank and the sample. Samples were read at 493 nm as excitation λex and 515 nm as emission λem (Varioskan Lux, Thermo Fisher Scientific, Waltham, MA, USA). In total, 1 × 10−2 M of fluorescein solutions in PBS (75 mM) 0.6 M AAPH in PBS (75 mM) were prepared. The sample contained 21 μL fluorescein, 2,899 μL PBS, 30 μL of the tested extract, and 50 μL AAPH [90].
The human colon adenocarcinoma HT-29 cell line was obtained from initial plate cultures (ATCC, Manassas, VI, USA). DMEM culture medium containing glucose (4.5 g/L) and L-glutamine was used for maintenance. The medium was supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. Cells were incubated at 37 °C in an atmosphere containing 5% CO2 and 95% humidity. Subcultures were performed under sterile conditions in a laminar flow hood. Cells were cultured with 90% confluence by trypsinization, for 5 min at 37 °C. Then, the trypsin was inactivated with culture medium supplemented with 10% FBS, and the cell solution was centrifuged at 1500× g by 3 min (Hermle Z323 K, Hermle Labortechnik GmbH, Wehingen, Germany). The cell pellet was resuspended in 1 mL of culture medium count the cells using a Neubauer chamber.
For the different experiments, the treatments included naringenin, BPA, naringenin with BPA, extract from the fermentation of naringenin (FEN), FEN with BPA, and the blank of the fermentation (10%), which corresponds to the culture medium required for the bacterial fermentation process. Naringenin was dissolved by stirring (SP131635Q ceramics hotirrert plate, Cimarec, Thermo Fisher Scientific, Waltham, MA, USA) for 24 h in DMEM medium supplemented with 2% FBS and 1% penicillin–streptomycin. BPA was diluted in 0.01% dimethyl sulfoxide (DMSO) and mixed with DMEM medium supplemented with 2% FBS and 1% penicillin-streptomycin. FEN was omitted using PVDF membrane filters with a pore size of 0.45 um pore sizes (MILFSLHV033RB syringe filters, Millex, Merck KGaA, Darmstadt, Germany), and mixed with DMEM medium supplemented with 2% FBS and 1% penicillin-streptomycin. For the co-exposure, both compounds were administered simultaneously. For all treatments, cells were exposed for 24 h.
Cell viability was determined by the MTT assay. Briefly, the cells were seeded in 96-well plates (1 × 104 cells/well) and grown for 24 h in DMEM supplemented with 10% FBS. Then, the culture medium was discarded and replaced by the different treatments, consisting of naringenin (50–500 µM), FEN (25–43%), 4.4 µM BPA diluted in 0.01% dimethyl sulfoxide (DMSO) and all dissolved in DMEM supplemented with 2% FBS. The BPA concentration used in this study was selected based on the U.S. Environmental Protection Reference Dose for Chronic Oral BPA Exposure (50 µg/ kg body weight/day), considering an average body weight of 70 kg and a total water intake of 3 l [91]. For co-exposure, the IC50 of naringenin plus 4.4 µM BPA and the IC50 of FEN plus 4.4 µM BPA were used. All treatments were incubated for 24 h. After incubation, the medium was removed and a MTT solution containing 0.5 mg/mL 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide dissolved in DMEM was added and incubated at 37 °C for 60 min. The solution was removed, and the formazan crystals were dissolved in DMSO at room temperature for 5 min. The absorbance of each sample was measured using a Multiskan Ascent (Thermo Fisher Scientific, Waltham, MA, USA, 51118307) flash spectral scanner at 540 nm. Measurements from cells without treatment were used for normalization. All experiments were performed three times each in triplicate.
Apoptosis was detected by flow cytometric analysis using Annexin-V, which identifies the externalization of phosphatidylserine. For analysis, cells were seeded in 6-well plates (4 × 106 cells per 2 mL per well) in DMEM medium containing 10% FBS. After 24 h of incubation, cells were treated with naringenin (250 µM, equivalent to IC50), BPA (4.4 µM), FEN (37%, equivalent to IC50) or in co-exposure. After 24 h treatment, floating cells in the culture medium were separated, while adherent cells were collected by trypsinization. The two cell populations were then pooled and centrifuged (778× g, for 5 min, at 4 °C). The supernatant was discarded and 100 µL of Anexin-V was taken, which was prepared under the supplier’s conditions along with 100 µL of resuspended cells in DMEM medium supplemented with 1% SFB. They were incubated for 20 min at room temperature and analyzed in a Merck Muse Cell Analyzer (Merck KGaA, Darmstadt, Germany) [92].
To determine cellular necrosis, LDH activity was measured. For this purpose, cells were seeded in 96-well microplates (1 × 104 cells/well) in DMEM culture medium with added FBS (10%). After 24 h of incubation, the medium was replaced by DMEM supplemented with 2% FBS containing the different concentrations of the evaluated treatments and incubated for 24 h. Cells without treatments were used as negative control and cells treated with Triton X-100 and maintained with DMEM medium containing 2% FBS were used as positive control. After 24 h of incubation, the supernatant was collected and transferred to a 96-well microplate, where 100 µL of the reaction mixture was then added according to the method recommended by the supplier (Sigma-Aldrich, St. Louis, MO, USA, MAK066), and incubated at room temperature for 20 min protected from light. Finally, absorbance was measured at 492 nm (Multiskan Ascent 51118307, Thermo Fisher Scientific, Waltham, MA, USA).
HT-29 cells with the different treatments were rinsed with PBS, and the cell sediments were sonicated. SOD activity was measured using the superoxide dismutase assay kit (19–160, Sigma-Aldrich, St. Louis, MO, USA), according to the manufacturer’s protocol. Absorbance was measured at 440 nm in the microplate reader. The SOD assay uses tetrazolium salt to detect superoxide radicals generated by xanthine oxidase (XO). The activity of SOD is expressed as the percentage inhibition of OX.
Reduced glutathione was quantified using the kit #CS0260 (Sigma-Aldrich, St. Louis, MO, USA), which provides the 5-sulfosalicylic acid needed to deproteinize the biological sample. For this technique, 1 × 108 cells were seeded, washed with PBS after the treatments, later recovered and centrifuged (600× g, for 10 min). 5-sulfosalicylic solution equivalent to 3 times the volume of sediment was added, then frozen in liquid nitrogen and thawed in a water bath 37 °C. After centrifugation of the cells (10,000× g, 10 min) (Z323 K, Hermle Labortechnik GmbH, Wehingen, Germany), the supernatant was transferred to 90-well plates, where the working mixture was added (95 mM potassium phosphate buffer pH = 7, 0.95 mM EDTA, 48mM NADPH, 0.115 units/mL reduced glutathione and 0.24% 5-sulfosalicylic). Finally, absorbance was measured at 412 nm. Calculations were performed based on the previously established GSH standard curve.
Expression of ERβ and GPR30 genes was determined by qPCR using β-actin as housekeeping gene for normalization. Extracted RNA from treated and untreated HT-29 cells was performed using silica-gel columns (Jena, Germany). Subsequently, cDNA was synthesized using oligo dT primers and the SCRIPT cDNA kit under the following conditions: 50 °C for 40 min, 70 °C for 10 min. For qPCR, the RadiantTM Green Hi-ROX qPCR kit (Thermo Fisher Scientific, Waltham, MA, USA) was used under the following conditions: 95 °C for 2 min, 95 °C for 5 s, alignment time (Supplementary Table S1) for 20 s, and 65 °C for 10 s. Subsequently, the expression of 84 genes was assessed using the Human RT2 Profiler Real Time PCR Array (PAHS-027A, Qiagen, Germantown, MD, USA) according to the manufacturer’s user manual. The array includes genes related to p53. Data were analyzed using RT2 Profiler PCR Data Analysis software from GeneGlobe (Qiagen, Germantown, MD, USA), based on the ΔΔC t method with normalization of raw data to housekeeping genes (β-actin and GAPDH). We considered sequences as potential target genes if the change between the control group and treatments was more than 2-fold (up- or down-regulated genes).
For miR-200c and miR-141 expression, 2 µL of RNA (100 ng/µL) was collected and cDNA was synthesized using Stem Loop for miR-200c and miR-141 as primer and SCRIPT cDNA Synthesis Kit under the following conditions: 50 °C for 40 min, 70 °C for 10 min; then, qPCR was performed using universal antisense and specific sense primer for miR-200c and miR-141, U6 was taken as the constitutive gene synthesized using the oligo dT primer and the corresponding FWD and REV primers for qPCR (StepOne™ 48-well, Thermo Fisher Scientific, Waltham, MA, USA) (Supplementary Table S2). The qPCR conditions were as follows: 95 °C for 2 min, 95 °C for 5 s, 60 °C for 20 s, and 65 °C for 10 s using the RadiantTM Green Hi-ROX qPCR kit (QS2100, Alkali Scientific Inc., FL, USA) The predictive power of suppression of protein-coding gene expression was evaluated using miRmap software, Swiss Institute of Bioinformatcs, Université De Genève.
The results obtained were analyzed using a two-way and one-way variance test (ANOVA), followed by a post hoc Tukey test to compare the treatment and control groups. Analyzes were performed using the Prism 8 program (GraphPad Software, San Diego, CA, USA).
The results of the present study showed that microbial metabolism during fermentation of naringenin in the colon produced by-products of biological interest such as apigenin and 3HPP, and that both naringenin and FEN dose-dependently reduced the viability of HT-29 colon cancer cells, both alone and with BPA co-exposure. Naringenin promoted extrinsic apoptosis through overexpression of TNFRST10D/CRADD/CASP-2 and intrinsic apoptosis through PTEN/BBC3/APAF-1/CASP-9 on HT-29 cells (Figure 9). FEN promoted apoptotic mechanisms through intrinsic pathways likely induced by H2O2 generated by SOD activity (Figure 10). In contrast, BPA decreased the expression of BBC3, an important gene that promotes apoptosis, decreases tumour suppressor genes such as PTEN and also decreases MLH-1, which plays a role in DNA repair related to cell proliferation. In addition, BPA-mediated overexpression of miR-200c is thought to be related to the down-regulation of PTEN, which regulates the cell cycle and promotes apoptosis on HT-29 cells. Furthermore, simultaneous exposure of naringenin and FEN to BPA triggers overexpression of the TNF gene, which causes overexpression of the FASL gene, mediating activation of apoptosis via the extrinsic pathway (Figure 11). The decrease in cell viability as well as the recovery of ERβ gene expression on HT-29 cells under treatment with naringenin and its extract FEN suggest that the expression of this receptor is associated with apoptotic mechanisms and DNA damage repair, although the increase in ERα gene expression related to proliferation processes, increased ERβ expression plays a very important role in inhibiting colon cancer development. Finally, ERβ could modulate the low expression of miR-141 targeting PTEN in HT-29 cells treated with FEN. Overall, our results demonstrated that naringenin and its fermentation extract activate apoptotic signalling pathways even in the presence of BPA, suggesting the effect of naringenin in inhibiting colon cancer development. Finally, this provides an important perspective for further exploration of naringenin as an agent with enhanced efficacy against the identified molecular targets given the high BPA burden worldwide. | true | true | true |
PMC9573026 | 36234731 | Lin Wang,Tingting Pan,Yan Wang,Jiewen Yu,Peiyi Qu,Yue Chen,Hua Xin,Sicen Wang,Junxing Liu,Yan Wu | Effect of Nanoparticles of DOX and miR-125b on DNA Damage Repair in Glioma U251 Cells and Underlying Mechanisms | 21-09-2022 | glioma,nanoparticles (NPs),DNA damage,miR-125b | Glioma is the most common primary craniocerebral malignant tumor, arising from the canceration of glial cells in the brain and spinal cord. The quality of life and prognosis of patients with this disease are still poor. Doxorubicin (DOX) is one of the most traditional and economical chemotherapeutic drugs for the treatment of glioma, but its toxic effect on normal cells and the resistance of tumor cells to DOX make the application of DOX in the treatment of glioma gradually less effective. To solve this problem, we co-encapsulated DOX and endogenous tumor suppressor miR-125b into nanoparticles (NPs) by nanoprecipitation methods, and passively targeted them into glioma cells. In vitro experiments show that miR-125b and DOX can be effectively encapsulated into nanoparticles with different ratios, and by targeting YES proto-oncogene 1 (YES1), they can affect the adenosine 5′-monophosphate (AMP)-activated protein kinase (AMPK)/p53 pathway and induce brain glioma cell apoptosis. They can also affect the DNA damage repair process and inhibit cell proliferation. The obtained data suggest that co-delivery of DOX and miR-125b could achieve synergistic effects on tumor suppression. Nanosystem-based co-delivery of tumor suppressive miRNAs and chemotherapeutic agents may be a promising combined therapeutic strategy for enhanced anti-tumor therapy. | Effect of Nanoparticles of DOX and miR-125b on DNA Damage Repair in Glioma U251 Cells and Underlying Mechanisms
Glioma is the most common primary craniocerebral malignant tumor, arising from the canceration of glial cells in the brain and spinal cord. The quality of life and prognosis of patients with this disease are still poor. Doxorubicin (DOX) is one of the most traditional and economical chemotherapeutic drugs for the treatment of glioma, but its toxic effect on normal cells and the resistance of tumor cells to DOX make the application of DOX in the treatment of glioma gradually less effective. To solve this problem, we co-encapsulated DOX and endogenous tumor suppressor miR-125b into nanoparticles (NPs) by nanoprecipitation methods, and passively targeted them into glioma cells. In vitro experiments show that miR-125b and DOX can be effectively encapsulated into nanoparticles with different ratios, and by targeting YES proto-oncogene 1 (YES1), they can affect the adenosine 5′-monophosphate (AMP)-activated protein kinase (AMPK)/p53 pathway and induce brain glioma cell apoptosis. They can also affect the DNA damage repair process and inhibit cell proliferation. The obtained data suggest that co-delivery of DOX and miR-125b could achieve synergistic effects on tumor suppression. Nanosystem-based co-delivery of tumor suppressive miRNAs and chemotherapeutic agents may be a promising combined therapeutic strategy for enhanced anti-tumor therapy.
Glioma is the most common primary malignant intracranial tumor caused by the canceration of brain and spinal cord glial cells. It mainly originates from the glial cells constituting brain parenchyma [1]. The incidence of glioma accounts for about 60% of the total incidence of intracranial tumors [2]. Currently, the primary treatment for glioma is surgery, which is supplemented by radiotherapy, chemotherapy, and immunotherapy [3]. Surgical removal of glioma still faces the problem of an extremely high recurrence rate, possibly because of the intense stress and inflammatory reactions in surgery which induce immunosuppression and promote the local occurrence of cancer [4]. As for the poor effects of radiotherapy and chemotherapy, the most probable causes are the intrinsic or acquired resistance to radiotherapy and chemotherapy drugs and the invasion of other sites in post-operative recurrence [5]. This suggests that traditional treatment strategies should be further improved to maximize therapeutic effects. DNA damage repair is realized via the interactions between a series of signaling pathways and enzymes. Ataxia-telangiectasia mutated (ATM) protein kinase, as a member of the kinase family, has sequence homology with phosphoinositide-3-kinase (PI3K), and constitutes the core of DNA damage repair. Its main functions include regulating the activation of checkpoints under DNA double-strand break or oxidative stress and coordinating DNA repair, cell cycle progression, and cell metabolism [6]. The pathogenic germline mutations of ATM play a role in DNA damage reactions and cell cycle checkpoints, so ATM has now become an important marker of cancers and a new target for cancer treatment [7]. Histones play a pivotal role in the composition and gene function regulation of chromatin structures in eukaryotes. The nucleosome constituting chromatin is mainly composed of four core histones (H2A, H2B, H3, and H4) and one monomer histone H1, while another protein subtype H2AX plays a critical part in DNA damage repair. DOX is one of the most traditional and economical chemotherapy drugs for treating glioma. However, the toxic effect of DOX on normal cells and the resistance of tumor cells to DOX have reduced the use of DOX in glioma treatment. Resistance to DOX has also become a research emphasis in clinical treatment with DOX. So far, many molecular pathways and mechanisms of resistance to DOX have been identified [8]. Non-coded RNA also plays a vital role in resistance to DOX. MicroRNA-125b (miR-125b) is one of the most important miRNAs that regulate all kinds of physiological and pathological processes. Currently, the role of miR-125b in many cancers has been well confirmed. By binding with the 3′ untranslated region (3′-UTR) of target mRNA, it causes the degradation or translational inhibition of target mRNA [9]. It has been proven that miR-125b experiences expression dysregulation in many cancers [9], and affects the development of cancer by affecting cell proliferation and apoptosis-related pathways such as PI3K [10] and NF-κB [11]. Studies have shown that the expression level of miR-125b drops significantly in breast cancer patients, and that the overexpression of miR-125b enhances the sensitivity of DOX-resistant breast cancer cell MCF-7/R to DOX [12]. Therefore, we speculate that miR-125b also increases the sensitivity of tumors to DOX in other tumors, and this effect may be achieved through its downstream target genes. Through bioinformatics analysis, a possible target of miR-125b is the yes1 gene. YES1 is a tyrosine kinase coded by proto-oncogene yes1. As a member of the src protein family, YES1 plays a critical part in cell proliferation, adherence, and differentiation [13]. NPs show great application potential in the medical field [14]. Compared with free drugs, NPs can directly transport drugs to specific cells or tissues, thereby greatly reducing the side effects of drugs and increasing the site-specific delivery of drugs [15]. At the same time, packaging such as nanoparticles can also reduce drug toxicity and improve efficacy. Preliminary experiments [16] have demonstrated that nanoparticles have great application potential in the delivery of antitumor drug DOX. In our present studies, nanocarriers can deliver dual drugs in addition to single drugs. Nano-complexes were prepared using nanoprecipitation which simultaneously encapsulate DOX and miR-125b mimics. The particle size, surface zeta potential, morphology, DOX and miR-125b encapsulation efficiency were characterized. The synergetic effects and mechanisms of DOX and miR-125b in glioma therapy were further investigated in vitro.
The morphology of NPs was observed by TEM. The NPs appeared to be typical spheres in shape with good dispersion (Figure 1A). The average size, size distribution, and surface potentials of NPs, NPs+miR-125b, DOX+NPs, and DOX+NPs+miR-125b were measured via DLS (Figure 1B,C). The average size of NPs, NPs+miR-125b, DOX+ NPs, and DOX+NPs+miR-125b was 256.08 nm, 237.2664 nm, 220.7669 nm, 235.6625 nm, respectively. All the NPs showed a narrow size distribution (PDI < 0.2). The zeta potential for the NPs, NPs+miR-125b, DOX+NPs, and DOX+NPs+miR-125b was −17.8 mV, −17.5 mV, −19.9 mV and −19.1 mV, respectively (Table 1). Then we examined the encapsulation efficiency for DOX and miR-125b at different material mass ratios (Table 2). An optimal proportion of 8:1 (DOX+NPs: miR-125b, mass ratio) was chosen for the following studies.
First, we measured the mRNA level of YES1 in miR-125b mimics-transfected cells (Figure 2B). As expected, the mimics of miR-125b significantly downregulated the mRNA level of YES1 when compared with control group. In order to make sure whether miR-125b can directly target YES1 3′UTR, we predicted the binding site of miR-125b to the yes1 3′UTR region (Figure 2A) and mutated this binding site (Figure 2C). Luciferase report vectors were constructed with wild-type YES1 3′UTR (YES1 3′UTR WT) and mutated YES1 3′UTR (YES1 3′UTR MT). The YES1 luciferase activity was measured for describing miR-125b function on luciferase translation. The results showed that luciferase activity of wild-type YES1 3′UTR was significantly inhibited by miR-125b overexpression, yet mutated YES1 3′UTR terminated this effect (Figure 2D). Taken together, we confirmed that YES1 is the dir000ect target of miR-125b. Thus, we reached the conclusion that miR-125b targets YES1 and regulates YES1 expression.
miR-125b was transfected into cells, and the cell viability of U251 cells transfected with different concentrations of miR-125b at 24 h and 48 h was determined by MTS assay (Figure 3A). It can be seen that with the increase of miR-125b concentration and duration, the viability of U251 cells gradually decreased. This result suggests that miR-125b inhibits U251 in a time- and dose-dependent manner. U251 cells were treated with DOX, NPs, DOX+NPs, and DOX+NPs+miR-125b for 24 h and 48 h, respectively (the drug concentration is based on DOX concentration). Cell viability was determined by MTS assay (Figure 3B,C). According to the results, with the increase of DOX drug concentration and the increase of action time, the cell viability of each group decreased gradually. However, there was no difference in cell viability between NPs, DOX-NPs and the same concentration of DOX (p > 0.05), which indicated that nanoparticles had no effect on the viability of U251 cells. We chose the conditions (DOX concentration of 1 μg/mL, DOX+NPs and miR-125b ratio of 8:1, and the effect time of 24 h) with cell viability around 50% and statistically significant (p < 0.05) for our next experiments. Cellular uptake of DOX+NPs+miR-125b was assessed by CLSM measurements. DOX+NPs+miR-125b was incubated with U251 cells at 37 °C for 0.5 and 1 h. We used the red autofluorescence of DOX and the green fluorescence of FITC to study the cellular uptake of nanoparticles and the intracellular localization of DOX. As shown in Figure 4A, a small amount of red fluorescence was observed in U251 cells incubated with DOX+NPs+miR-125b for 0.5 h, and when the cells were incubated for 1 h, strong DOX red fluorescence appeared around the cytoplasm. This suggests that DOX+NPs+miR-125b may be taken up by cells through a nonspecific endocytic mechanism, and DOX molecules are released in the endocytic compartment. To evaluate the endosomal escape behavior of nanocarriers, DOX+NPs+miR-125b was incubated with U251 cells for different durations. At predetermined time points (0, 2, 4 h), co-localization analysis was performed to examine the extent of nanoparticle and lysosome overlap (Figure 4B). Green fluorescence indicates the location of the micellar nanocarriers. Red fluorescence from a commercial probe (LysoTracker® Red DND-99, Invitrogen, Waltham, MA, USA) indicates the site of the lysosome. The overlap of green and red produces yellow due to co-localization of nanocarriers and lysosomes. The green fluorescence of NPs can be seen within the organelle by overlaying the fluorescence images of LysoTracker Red (BOC Sciences, New York, NY, USA), which is concentrated in the lysosome, with only a small fraction randomly distributed in the cytoplasm. These results suggest that delivery of DOX-loaded NPs via lysosomes is likely to enhance DOX release, which we demonstrate occurs efficiently at lysosomal pH. Regarding DOX+NPs+miR-125b, the degree of colocalization remained high over the course of the experiment (i.e., 4 h), suggesting a lower degree of endosomal escape.
After treatment with different DOX concentrations of DOX, DOX+NPs and DOX+NPs+miR-125b, the cell cycle was detected by flow cytometry (Figure 5A), and all concentrations of DOX, DOX+NPs and DOX+NPs+miR-125b could induce G0/G1 arrest in U251 cells. The G0/G1 phases of U251 cells treated with DOX at concentrations of 0.25, 0.5 and 1 μg/mL were 55.13%, 61.04% and 65.18%, respectively. After treatment with DOX+NPs at DOX concentrations of 0.25, 0.5 and 1 μg/mL, the G0/G1 phase distribution of U251 cells was 54.67%, 60.43% and 64.54%, respectively. After treatment with DOX+ NPs+miR-125b at DOX concentrations of 0.25, 0.5 and 1 μg/mL, the G0/G1 phase distribution of U251 cells was 60.10%, 65.44% and 71.08%, respectively. While the untreated cells showed only 51.12% of the G0/G1 phase distribution of the treated U251 cells, the G0/G1 phase arrest was significantly higher than that of the untreated U251 cells (p < 0.001). It is worth noting that compared with the same concentration of DOX group, DOX+NPs+miR-125b very significantly increased the G0/G1 phase arrest of U251 cells (p < 0.001), while DOX+NPs and the same concentration of DOX group did not have this effect (Figure 5B). This suggests that miR-125b, but not nanoparticles, can significantly enhance the cell cycle arrest effect of DOX. Cell cycle checkpoints were detected by Western Blot and immunofluorescence (Figure 6). Cell cycle checkpoints were detected by Western Blot in U251 cells treated with different drugs, and it was found that compared with untreated U251 cells, the expression levels of ATM, H2AX, p-ATM and γ-H2AX were increased in drug-treated cells, and the difference was statistically significant (p < 0.05). Compared with the DOX group, the cells treated with DOX+NPs+miR-125b also had significantly higher expression of the above proteins, and the difference was statistically significant (p < 0.01). However, there was no significant difference in protein expression between the DOX+NPs group and the DOX group, indicating the effect of miR-125b on aggravating DOX-induced cell cycle arrest (Figure 6A). The results of immunofluorescence assay for γ-H2AX and p-ATM were the same as those of Western Blot (Figure 6B). It can be seen from the figure that the fluorescence intensity of the control group is very weak, and the fluorescence intensity of the treated cells is higher than that of the control group. Moreover, the fluorescence intensity of cells treated with the same drug also increased with the increase of DOX concentration in the drug. After treatment with the same DOX concentration, the fluorescence intensity of DOX+NPs+miR-125b group was higher than that of DOX group, indicating that DOX+NPs+miR-125b induced more severe cell cycle arrest compared with DOX.
Studies have shown that YES1 promotes the expression of YES1-associated protein (YAP1), which can also be regulated by AMPK, and its expression is increased in the presence of AMPK inhibition [16]. However, the relationship between YES1 and AMPK is still uncertain. In view of the important role of AMPK/p53 in apoptosis, we inhibited the expression of YES1 by transfecting miR-125b, detected the changes of AMPK expression, and analyzed the relationship between YES1 and AMPK. After transfecting U251 cells with different concentrations of miR-125b for 24 h (Figure 7), the expression levels of YES1 and p-YES1 gradually decreased with the increase of miR-125b concentration, and the difference was statistically significant (p < 0.05). Especially after the action of medium and high concentrations of miR-125b, the expression of YES1 and p-YES1 decreased more significantly (p < 0.01). With the increase of miR-125b concentration, the expressions of AMPK, p-AMPK, p53 and p-p53 gradually increased (p < 0.05), and the change of phosphorylated protein expression was more significant than that of total protein. (p < 0.01). From the results, we can speculate that YES1 protein may have an inhibitory effect on AMPK, and that miR-125b indirectly activates the AMPK/p53 signaling pathway and promotes apoptosis by inhibiting the expression of YES1. U251 cells were treated with the same DOX concentration of DOX, DOX+NPs, DOX+NPs+miR125b (Figure 8), and compared with the control group. The expressions of YES1 and p-YES1 in the drug-treated group were decreased, and the difference was statistically significant (p < 0.05). However, AMPK, p-AMPK, p53, and p-p53 showed the opposite trend, that is, compared with the control group, the expression of the above four proteins increased in the drug treatment group (p < 0.05), and the changes of p-AMPK and p-p53 were more significant (p < 0.01). It is worth noting that compared with the DOX group, the expression levels of YES1 and p-YES1 in the DOX+NPs+miR-125b group were significantly decreased (p < 0.01), while the expression levels of AMPK, p-AMPK, p53 and p-p53 were significantly increased (p < 0.01). However, compared with the DOX group, DOX+NPs had no significant effect on protein expression (p > 0.05).
Glioma is characterized by high cellular heterogeneity, fast proliferation, and strong invasiveness. Currently glioma is mainly treated by surgery, together with radiotherapy and chemotherapy, but post-operative recurrence is still unavoidable [17]. DNA damage repair is realized via the interactions between a series of signaling pathways and enzymes. Ataxia-telangiectasia mutated (ATM) protein kinase, as a member of the kinase family, has sequence homology with PI3K, and constitutes the core of DNA damage repair. Its main functions include regulating the activation of checkpoints under DNA double-strand break or oxidative stress and coordinating DNA repair, cell cycle progression, and cell metabolism [18]. The pathogenic germline mutations of ATM play a role in DNA damage reactions and cell cycle checkpoints, so ATM has now become an important marker of cancers and a new target for cancer treatment [19]. Histones play a pivotal role in the composition and gene function regulation of chromatin structures in eukaryotes. The nucleosome constituting chromatin is mainly composed of four core histones (H2A, H2B, H3, and H4) and one monomer histone H1, while another protein subtype H2AX plays a critical part in DNA damage repair. The DOX+NPs+miR-125b NPs constructed in this study can inhibit the proliferation of glioma cells. One of the possible mechanisms lies in that NPs can induce DNA damage and inhibit DNA repair in glioma. Compared to DOX used alone, DOX+NPs+miR-125b double-loaded NPs significantly up-regulate the expression levels of γ-H2AX and p-ATM. Histone H2AX is an important cell cycle checkpoint. In case of DNA double-strand break (DSB) damage, H2AX experiences rapid phosphorylation, resulting inγ-H2AX. γ-H2AX can recruit cell cycle-related proteins and repair proteins to the damage site. These proteins and γ-H2AX form γ-H2AX foci. As one of the protein complexes recruited in the early stage after occurrence of DNA damage, γ-H2AX foci provide binding sites for other repair proteins, such as BRCA1 (breast cancer gene 1) and 53BP1 (p53 binding protein), and contribute to DSB repair. Thus, γ-H2AX is a biomarker that can characterize DNA DSB damage and repair [20]. A higher expression level of γ-H2AX suggests that DOX+NPs+miR-125b induces more serious DNA DSB damage than DOX and has a stronger lethality against tumors. Meanwhile, DNA damage initiates DNA damage response (DDR). ATM, as the core protein of DNA damage repair, automatically experiences phosphorylation and further phosphorylates downstream target cell cycle checkpoint kinase 2 (CHK2). CHK2 is a protein kinase, and one important substrate of CHK2 is cell division cycle 25 homolog A (CDC25A). CDC25A activates cyclin-dependent kinase 2 (CDK2), and promotes cells to develop from G1 phase to S phase. However, when CDC25A is phosphorylated by p-CHK2, its activity declines, and its functions are inhibited, making it impossible for cells to enter S phase and resulting in G0/G1 phase arrest. As a result, DNA damage cannot be normally repaired, and the proliferation of tumor cells is inhibited. YES1 is a tyrosine kinase coded by proto-oncogene yes1. As a member of the src protein family, YES1 plays a critical part in cell proliferation, adherence, and differentiation [21]. Dual-luciferase detection experiments have confirmed the role of yes1 as a target gene of miR-125b, and shown that miR-125b negatively regulates the expression of YES1. Adenosine monophosphate-activated protein kinase (AMPK) is a critical molecule in bioenergy metabolism regulation, and a centrin linking up anabolism and catabolism. Its role in diabetes and other metabolic diseases is widely known. However, recent studies have revealed that AMPK also plays a significant role in the regulation of apoptosis, autophagy, and cell cycle [22]. Activation of AMPK/p53 pathway is important for apoptosis. In this study, in control U251 cells, the expression levels of YES1 and p-YES1 were relatively high, while the expression levels of p-AMPK and p-p53 were relatively low. After miR-125b transfected cells, the expression of YES1 decreased under the action of miR-125b because of its negative regulation of the expression of YES1, and the expression level gradually decreased with the increase of miR-125b concentration. At the same time, the expression levels of p-AMPK and p-p53 increased with the increase of miR-125b concentration (Figure 7). According to the experimental results, we have reason to believe that YES1 has an inhibitory effect on AMPK, and miR-125b alleviates its inhibitory effect on AMPK by inhibiting the expression of YES1, resulting in an increase in the active form of AMPK (p-AMPK), thereby activating AMPK/p53 pathway and induction of apoptosis. We used our constructed double-loaded nanoparticles to act on U251 cells, and the results showed (Figure 8) that although DOX can also activate the AMPK/p53 pathway by inhibiting the expression of YES1, the effect is far less than that of DOX and miR-125b. Compared with DOX or mi-125b alone, dual-loaded nanoparticles inhibited YES1 and activated the AMPK/p53 pathway more significantly, and induced tumor cell apoptosis more strongly. In conclusion, miR-125b can aggravate the cell cycle arrest caused by DNA damage by DOX, and may also attenuate its inhibitory effect on AMPK by inhibiting YES1, thereby activating the AMPK/p53 pathway. AMPK/p53 pathway plays an important role in apoptosis. Therefore, we speculate that miR-125b can promote the apoptosis of U251 cells by inhibiting YES1 to activate the AMPK/p53 pathway, but further experiments are needed to prove this. Compared with DOX or miR-125b alone, the double-loaded nanoparticles combined with DOX and miR-125b could not only accurately enter tumor cells, but also had stronger activation effects on DNA damage and apoptosis pathways.
DOX was provided by Beijing Huafeng United Technology (Beijing, China). Antibodies against GAPDH, ATM, p-ATM, H2AX, p-H2AX, YES1, p-YES1, p53, p-p53, AMPKα, p-AMPKα were purchased from Abcam (Cambridge, UK). Dulbecco’s modified Eagle’s high glucose medium and Dulbecco’s modified Eagle’s medium were provided by GIBCO (Grand Island, New York, NY, USA). Dual-Luciferase Reporter Assay System Kit (Thermo Fisher Scientific, Waltham, MA, USA) and pSI-CHECK2 vector were provided by Hanbio Biotechnology (Wuhan, China). EntiLink™ 1st Strand cDNA Synthesis Kit and EnTurbo™ SYBR Green PCR SuperMix Kit were purchased from ELK Biotechnology (Wuhan, China).
DOX and miR-125b were encapsulated into copolymer NPs by the nano-precipitation method, as detailed in previous research [16]. The sizes, distribution, and surface potentials of prepared NPs were measured by dynamic light scattering (DLS).
Human glioma cells U251 and human embryonic kidney cell line 293T were from Shanghai Cell Bank (Shanghai, China), Chinese Academy of Science. U251 was cultured in Dulbecco’s modified Eagle’s high glucose medium with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. To investigate the role of DOX-miR-125b-NPs in glioma, human glioma U251 cells were exposed to a certain concentration of DOX, DOX NPs, miR-125b, DOX-miR-25b-NPs for a period of time. The 293T was cultured in DMEM used for the dual-luciferase assays. The medium was supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Cells were cultivated in a humidified incubator containing 5% CO2 at 37 °C.
The proliferation effects of drugs on U251 cells were determined by 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-soufophenyl)-2H-tetrazolium, inner salt(MTS) assay. In brief, cells were plated in a 96-well plate overnight to adhere, then different drugs were administrated and incubation continued for 24 or 48 h. MTS solution (20 µL per well) was added and incubated for another 4 h at 37 °C, discarding the super natant and using dimethyl sulfoxide (DMSO) to dissolve products for 10 min at 37 °C. Microplate reader (BioTek, Winooski, VT, USA) was used to measure the 96-well plate at 490 nm, and the cell viability (%) was calculated by OD values.
For the luciferase reporter assay, the 3′UTR of YES1 containing the wild or mutant miR-125b target sites was cloned using primers with NotI and XhoI cleavage sites. The wild or mutant type 3′UTR fragment was inserted into the corresponding site of the pSI-CHECK2 vector and then co-transfected into 293T cells with miR-125b mimics/mock. After 48 h transfection, the cells were harvested and the Dual-Luciferase Reporter Assay System Kit was used for detecting dual-luciferase activity, according to the manufacturer’s instructions.
Total RNA was isolated using the TRIzol reagent (ELK Biotechnology, Wuhan, China) according to the manufacturer’s protocol. cDNA was synthesized using the EntiLink™ 1st Strand cDNA Synthesis Kit. All real-time polymerase chain reactions (PCRs) were performed using EnTurbo™ SYBR Green PCR SuperMix Kit on a StepOne™ Real-Time PCR system (Life Technologies, New York, NY, USA). GAPDH was used as an internal control for normalization of the relative expression levels. Gene expression levels were calculated using the 2–ΔΔCT method relative to that of the internal control. Primers are listed in Table 3.
U251 were treated with different concentrations of drugs. Cells were centrifuged, then washed twice with PBS and lysed with loading buffer for 45 min in 4 °C. Then, the cell lysate was boiled for 10 min and stored at −80 °C. Protein samples were resolved by 8–10% SDS-PAGE, transferred to PVDF membranes (Millipore, Billerica, MA, USA). The PVDF membranes were blocked with 5% skimmed milk and then incubated with the primary antibodies overnight at 4 °C. Subsequently, the membranes were incubated with HRP-secondary antibody at 37 °C for 1 h. Finally, the image was detected by Tanon 5200 (Tanon, Beijing, China).
Cells were harvested and fixed for 20 min at 4 °C in 90% ethanol. Thereafter, the cells were washed twice with phosphate buffered saline (PBS) and then stained with the PI/RNase staining buffer (Sungene Biotech, Shanghai, China). Cell-cycle distribution was determined using flow cytometry. Each experiment was repeated three times.
Cells were inoculated in a six-well culture plate at the density of 1 × 105 cells/well for 24 h, and each well was placed with a sterile coverslip. After cell adherence, culture medium was added with fluorescein isothiocyanate (FITC)-labeled free drugs for culture at 37 °C for certain time, respectively. For the purpose of detecting intracellular localization, cells were further incubated with lysosomal red fluorescent probe 75 nM LysoTracker Red DND-99 for 30 min for lysosome labeling. The supernatant was carefully removed, and cells were washed three times with precooled PBS. Samples were detected by CLSM using Olympus FV1000 (Olympus, Shibuya, Japan). kex 488 nm and kem 510 nm were used for LysoTracker (BOC Sciences, New York, NY, USA) and FITC-labeled NPs (lysosomal red fluorescent probe and FITC; λex 488 nm and λem 510 nm).
Cells were inoculated in a six-well culture plate at the density of 1 × 105 cells/well for 24 h, and each well was placed with a sterile coverslip. After cell adherence, cells were washed once with PBS and fixed with 4% formaldehyde for 15 min, followed by three times of washing with PBS (5 min each time). After 10 min of permeation at room temperature (0.2%), cells were washed three times with PBS (5 min each time), blocked with 5% BSA+PBS at room temperature for 45 min, and incubated in 4 °C primary antibody overnight. After that, they were washed three times with PBS (5 min each time), and incubated in fluorescent secondary antibody for 30–40 min in dark place, followed by three times of washing with PBS (15 min each time). After addition of 30–50 µL DAPI dye liquor to each well, staining continued for 3–5 min. Finally, cells were washed three times with PBS (5 min each time), sealed, and observed under a laser confocal scanning microscope.
All data were presented as mean ± standard deviation. Experiments in each group were repeated at least three times. The statistical significance between two groups was analyzed using t-test. Intergroup significance was determined by one-way analysis of variance (ANOVA). The difference was considered statistically significant when * p < 0.05, ** p < 0.01, *** p < 0.001.
In this study, we report the successful application of DOX+NPs+miR-125b prepared by nanoprecipitation method. Our data showed that DOX+NPs+miR-125b inhibited U251 more significantly than DOX alone. The reason is that miR-125b can aggravate the DNA damage caused by DOX, leading to more severe cell cycle arrest; it can also attenuate the inhibitory effect of YES1 on the AMPK/p53 pathway and promote cell apoptosis. Drugs and miRNAs are linked together by nanoparticles, which can not only amplify the effect of the drug itself, but also bring the corresponding role of miRNAs into play. Since the co-delivery method has the advantage of simultaneously inhibiting tumor growth and migration, the co-delivery of miRNAs and chemotherapeutic drugs through nanosystems has shown great potential as a combination therapy strategy in anticancer therapy. Owing to the advantages of the co-delivery approach for the simultaneous inhibition of tumor growth, co-delivery of miRNAs and chemotherapeutic drugs by nano-systems demonstrates a great potential as combined therapeutic strategy in anti-cancer treatment. It provides new ideas for research and development into tumor drugs, especially to improve the sensitivity of anti-tumor drugs. | true | true | true |
PMC9573083 | 36235327 | Xiangnan Li,Longming Zhu,Zhenxing Wu,Jianjian Chen,Tingzhen Wang,Xiaoli Zhang,Gaofu Mei,Jian Wang,Guihua Lv | Classification and Expression Profile of the U-Box E3 Ubiquitin Ligase Enzyme Gene Family in Maize (Zea mays L.) | 21-09-2022 | maize (Zea mays L.),U-box E3 (PUB) gene family,classification,expression profile | The U-box E3 (PUB) family genes encode the E3 ubiquitin ligase enzyme, which determines substrate specific recognition during protein ubiquitination. They are widespread in plants and are critical for plant growth, development, and response to external stresses. However, there are few studies on the functional characteristic of PUB gene family in the important staple crop, maize (Zea mays L.). In this study, the PUB gene in maize was aimed to identify and classify through whole-genome screening. Phylogenetic tree, gene structure, conserved motif, chromosome location, gene duplication (GD), synteny, and cis-acting regulatory element of PUB member were analyzed. The expression profiles of ZmPUB gene family in maize during development and under abiotic stress and hormones treatment were analyzed by the RNA-seq data. A total of 79 PUB genes were identified in maize genome, and they were stratified into seven categories. There were 25 pairs of segmental duplications (SD) and 1 pair of tandem duplication (TD) identified in the maize PUB gene family. A close relationship was observed between the monocot plant maize and rice in PUB gene family. There were 94 kinds of cis-acting elements identified in the maize PUB gene family, which included 46 biotic- and abiotic-responsive elements, 19 hormone-responsive elements, 13 metabolic and growth-related elements. The expression profiles of maize PUB gene family showed characteristics of tissue specificity and response to abiotic stress and hormones treatment. These results provided an extensive overview of the maize PUB gene family. | Classification and Expression Profile of the U-Box E3 Ubiquitin Ligase Enzyme Gene Family in Maize (Zea mays L.)
The U-box E3 (PUB) family genes encode the E3 ubiquitin ligase enzyme, which determines substrate specific recognition during protein ubiquitination. They are widespread in plants and are critical for plant growth, development, and response to external stresses. However, there are few studies on the functional characteristic of PUB gene family in the important staple crop, maize (Zea mays L.). In this study, the PUB gene in maize was aimed to identify and classify through whole-genome screening. Phylogenetic tree, gene structure, conserved motif, chromosome location, gene duplication (GD), synteny, and cis-acting regulatory element of PUB member were analyzed. The expression profiles of ZmPUB gene family in maize during development and under abiotic stress and hormones treatment were analyzed by the RNA-seq data. A total of 79 PUB genes were identified in maize genome, and they were stratified into seven categories. There were 25 pairs of segmental duplications (SD) and 1 pair of tandem duplication (TD) identified in the maize PUB gene family. A close relationship was observed between the monocot plant maize and rice in PUB gene family. There were 94 kinds of cis-acting elements identified in the maize PUB gene family, which included 46 biotic- and abiotic-responsive elements, 19 hormone-responsive elements, 13 metabolic and growth-related elements. The expression profiles of maize PUB gene family showed characteristics of tissue specificity and response to abiotic stress and hormones treatment. These results provided an extensive overview of the maize PUB gene family.
The ubiquitin proteasome system (UPS) is an energy dependent system that modulates protein activity and degeneration. It is involved in cellular growth, differentiation and apoptosis, secretion and endocytosis, gene transcription and expression, as well as signal transduction and immune response [1,2]. UPS is composed of the ubiquitin-activating (E1), ubiquitin-conjugating (E2), and ubiquitin ligase (E3) enzymes. Ubiquitination is a common post-translational modification involving proteins. The process is as follows: E1 binds ubiquitin using a high-energy, thiol-ester bond in an ATP-based system. The E1-ubiquitin complex then interacts with E2 and transfers the active ubiquitin to a cysteine residue in E2. E3 next recognizes the substrate proteins and catalyzes the formation of an isopeptide bond between the C-terminal carboxyl group of ubiquitin and free lysl-amino groups within the protein substrate [3,4,5,6,7]. Finally, the ubiquitinated protein is degraded by 26S protease [1]. The ubiquitin ligase enzymes (E3) determine ubiquitin-specificity via recognition of substrate proteins, and they constitute the largest family among the three enzymes. According to the functional mechanism and subunit composition, E3 can be classified into four distinct categories: homologous to E6-AP COOH-terminus (HECT), U-box, really interesting new gene (RING), and cullin-RING ligase (CRLs) [8]. The U-box E3 (PUB) ligase enzymes contain the U-box domain, which are made up of approximately 70 amino acids. Moreover, they are highly conserved in eukaryotes, such as plants, animals, and yeast [9]. The prototype U-box protein was initially reported in yeast [10]. Subsequently, the U-box protein was also recognized in mammals, wherein it influences ubiquitination in combination with E1 and E2 and in the absence of additional E3 components [11]. To date, the PUB gene family was detected in numerous species, namely Arabidopsis (Arabidopsis thaliana, 64), rice (Oryza sativa, 77), Chinese cabbage (Brassica rapa ssp. Pekinesis, 101), soybean (Glycine max, 125), barley (Hordeum vulgare, 67), tomato (Lycopersicon esculentum, 62), citrus (Citrus clementina, 56), grape (Vitis vinifera, 56), cotton (Gossypium raimondii, 93), and Medicago (Medicago truncatula, 41) [12,13,14,15,16,17,18,19,20,21]. The PUB gene family modulates protein ubiquitination and degradation, and it plays essential roles in growth, development, reproduction, biotic and abiotic stress, as well as hormones [1,11,22,23]. In rice, OsPUB75 encodes an active PUB ligase enzyme. A mutation in this gene decreases cellular proliferation, disorganizes cellular files in aerial organs, and eventually causes the dwarf phenotype [24]. In barley, the brh2 and ari-l mutant exhibits PUB ligase enzyme deficiency and produces a relatively strong semi-dwarf phenotype [25]. In Arabidopsis, AtPUB4 serves as a global modulator of cellular proliferation and division, which are critical elements of the root architecture [26]. In Brassica, ARC1 promotes the ubiquitination and proteasomal destruction of compatibility factors within the pistil, thereby resulting in pollen rejection [27]. The PUB gene family is also known to modulate abiotic and biotic stress responses as well. The PUB ligase enzyme CMPG1 is critical for plant defense and disease resistance in tobacco (Nicotiana tabacum) and tomato, and it is homologous to the encoding proteins of AtPUB20 and AtPUB21 in Arabidopsis [28]. GmPUB1 is up-regulated after Phytophthora sojae infection, and GmPUB1 gene silencing in soybean produces a loss of race-specific Phytophthora resistance [29]. Arabidopsis AtPUB30 participates in salt-stress tolerance as a negative modulator of the germination stage in root tissues [30]. OsPUB67 strongly regulates drought tolerance in rice, and over-expression of OsPUB67 enhances drought-stress tolerance by augmenting reactive oxygen scavenging capability and closing the stomata [31]. VaPUB encodes a novel PUB ligase enzyme, which is upregulated by cold stress. Overexpression of VaPUB augments cold- and salt-stress tolerance [19]. Arabidopsis AtPUB19 is highly up-regulated by salt, drought, cold, and heat stress. Knock out of AtPUBI9 substantially decreased the resistance to high temperature but enhanced the resistance to drought [32]. In addition, there are multiple PUB genes that play distinct roles in hormonal response. AtPUB10 negatively regulates the abscisic acid (ABA) in Arabidopsis [33]. In rice, a DSG1-encoded PUB ligase enzyme is negatively modulated by brassinosteroid (BR), ethylene (ETH), auxin (AUX), and salicylic acid (SA), and its mutant is less responsive to brassinosteroid compared to the wild-type rice [34]. The PUB ligase enzyme PHOR1 serves an essential function in the gibberellin (GA) response, and its function is conserved in potato (Solanum tuberosum) and Arabidopsis [35,36]. The PUB gene family serves essential functions in regulating plant development as well as response to biotic, abiotic, and hormonal stressors. However, the current published studies were primarily focused on model plants such as Arabidopsis, cotton, tomato, rice, and Medicago. Hence, the examination of the PUB gene family in maize is rather insufficient, particularly the assessment of its gene expression profile. Herein, we aimed to screen and classify the PUB gene family in maize and performed systematic and comprehensive analyses. This study would provide a basis for the in-depth study of the physiological activities associated with these genes.
Overall, 79 putative PUB genes were screened with HMMER, using default parameters and significant e−3 value against the maize genome. We named these genes as ZmPUB1 to ZmPUB79. According to Pfam and SMART, all 79 ZmPUBs contained the U-box domain (PF04564). We further analyzed the gene locus, chromosomal location, sequence length, exon amount, molecular weight (MW), isoelectric point (PI), and subcellular location of all ZmPUB genes (Table S1). The protein length of ZmPUBs was between 94–1353 amino acids, and the MW was between 10.75–144.72 kDa. The PI range was 4.57 to 9.80. The subcellular localization was estimated using the WoLF PSORT website. Among the 79 ZmPUB proteins, 19 were predicted to be cytoplasmic proteins; 14 were located in the nucleus; 10 were in the plasma membrane; and the rest were localized in various organelles, such as the chloroplast (29), endoplasmic reticulum (3), golgi apparatus (2), mitochondrion (1), and peroxisome (1). To explore the relationship among different PUB genes, we generated a neighbor-joining phylogenetic tree with the U-box domain sequences of ZmPUBs, AtPUBs, and OsPUBs. Based on our analysis, the 79 ZmPUB proteins were stratified into seven categories, namely group I to group VII (Figure 1). Group IV consisted of the most members, namely 20 ZmPUB proteins, followed by group VII (19), and group II (17). Groups I and V possessed the least amounts of members, including only four and five ZmPUB proteins, respectively. The protein structure domain analysis was carried out using SMART and Pfam databases (Table S1). All ZmPUB proteins harbored the U-box domain. From group I to group VII, the proportion of ZmPUB proteins containing only the U-box domain was 50% (2/4), 11.7% (2/17), 42.8% (3/7), 35% (7/20), 100% (5/5), 0 (0/7), and 89.4% (17/19). We also identified an additional 10 structural domains in the ZmPUB proteins. In group I, the RS4NT or TPR domain was detected in two ZmPUB proteins. In groups II and IV, a total of eight domain types were observed, including the ARM, pkinase-tyr, coil coil region, UFD2P_CORE, USP, TPR, WD40, and KAP domains. In groups V, VI, and VII, there were three major domain types, which were all in the coil coil region, as well as the ZnF_TTF and ARM domains, and they were located in 19 ZmPUB proteins. In maize, most ZmPUB proteins carried only the U-box domain, accounting for 45.6% (36/79) of total proteins. There were similar proportions surveyed in Arabidopsis, rice, barley and citrus, and they were 25.0% (16/64), 25.9% (20/77), 31.3% (21/67), and 30.3% (17/56), respectively [12,13,18,20]. A previous study revealed that the ARM domain was related to substrate recognition during the ubiquitination process, and it is the second populated domain in rice and Arabidopsis [21]. In maize, 25 ZmPUB proteins harbored the ARM domain, and it was similar to rice and Arabidopsis. Additionally, approximately 70–74% of ZmPUB proteins were clustered into three groups (groups II, IV, and VII) in rice, Arabidopsis, and maize (Figure 1). This suggested that the maize PUB gene family was evolutionarily conserved.
To better elucidate the association between gene function and evolution, we explored the structural organization and conserved motifs of PUB genes (Figure 2, Table S2). We observed between 1 to 16 exons, with a mean of 3.61 exons per gene. There were about 37.9% ZmPUB genes with one exon but no intron. The proportions of ZmPUB genes containing two, three, and four exons were 9.8%, 8.8%, and 13.5%, respectively. The largest number of exons was 16, and they were detected in ZmPUB4 and ZmPUB69. In group I, the exon number in ZmPUB genes was between 5–8, with a mean of 6.75. In groups II and IV, we detected the maximum variation of exons, which was 1–16 and 2–9, respectively. In group III, there were 2–4 exons per gene. Around 84–100% of ZmPUB genes from groups V, VI, and VII harbored only one exon. The ZmPUB genes from the same group shared comparable gene structure, thereby suggesting functional conservation among the maize PUB gene family. Additionally, the exon variations in ZmPUB genes may indicate the function diversity of the maize PUB gene family. We screened ten conserved motifs among the 79 ZmPUB genes (Figure 2, Tables S2 and S3). About 83% of ZmPUB genes contained motifs 1, 3 and 4 simultaneously. Motifs 6, 7, and 9 mainly existed in the ZmPUB genes of group II. The tandem repeats of motifs 2 and 5 were characteristic of ZmPUB genes in group IV, and it likely served a distinct biological function. Motif 8 was prevalent exclusively among groups IV and VI. Motif 10 existed in all groups except for groups I and II. The conserved motif analysis suggested that the PUB genes reserved the U-box domain and accumulated additional motifs over the course of evolution. Lastly, similar motifs among ZmPUB genes might suggest the conserved evolutionary relationship and similar biological function.
To examine the chromosomal distribution of ZmPUB genes, we screened the genomic database, based on the DNA sequence of individual ZmPUB genes, and drew the chromosomal location map via the MapChart software (Figure 3). In total, 77 ZmPUB genes (97.4%) were mapped to 10 chromosomes, and they were unevenly distributed within each chromosome. We identified 10, 12, and 13 ZmPUB genes at chromosomes 1, 4, and 5, respectively. On chromosomes 2, 3, 9, and 10 were identified 9, 8, 8, and 7 ZmPUB genes, respectively. Lastly, only two ZmPUB genes were mapped to the chromosome 8. Gene duplication (GD) involves segmental duplication (SD) and tandem duplication (TD), and it promotes the expansion of the gene family. We conducted GD analysis to reveal the expansion process of the maize PUB genes (Figure 4). We identified 25 pairs (43 ZmPUB genes) of SD and one pair of TD (ZmPUB43/ZmPUB44) (Figure 3 and Figure 4). GD occurred on one or two loci. Using synteny analysis, we identified eight ZmPUB genes (ZmPUB27, 32, 43, 45, 48, 55, 58 and 62) with duplications on two loci, whereas the rest of the genes exhibited duplication on one locus (Figure 4). To elucidate the evolutionary pressures acting on ZmPUB genes, we computed the Ka and Ks of the duplicated gene pair, which represented nonsynonymous and synonymous substitution rates, respectively (Table 1). The Ka values were between 0.02–1.00. The Ks values were between 0.17–2.81. The Ka/Ks values of 26 gene pairs were between 0.05 and 0.99, which suggested that these ZmPUB genes evolved under strong purifying selection (Ka/Ks < 1). In addition, we computed the probable dates of the duplication events, and they occurred between 13.06 Mya (Ks = 0.17, Million years ago) and 215.77 Mya (Ks = 2.81, Million years ago), with an average of 93.55 Mya (Ks = 1.21, Million years ago). To further explore the evolutionary associations of the PUB genes in Arabidopsis, maize, and rice, we assessed syntenic interactions among the three species using the software MCScanX (Figure 5). A total of 64, 79, and 77 PUB genes were used from Arabidopsis, maize, and rice, respectively [12,13]. There were 53 syntenic PUB gene pairs screened between maize and rice, while only 5 syntenic PUB gene pairs were found in Arabidopsis and maize.
To better elucidate the transcriptional modulatory pathways associated with ZmPUB genes, we performed cis-acting regulatory element analysis with the 1500 bp upstream region from the transcription start site in PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/, (accessed on 1 March 2022)). In total, we screened 94 cis-acting regulatory elements in the promoter sequences of 79 ZmPUB genes (Tables S4 and S5, Figure 6). This included 3 core promoter elements, 19 hormone-responsive elements, 46 biotic- and abiotic-stress-response elements, 13 metabolism- and development-related elements, and 13 unknown function elements. The core promoter elements (CAAT-box, TATA-box) were present in all ZmPUB genes. The hormone-response elements consisted of MeJA and JA (4), ABA (4), ETH (3), GA (3), IAA (3), and SA (2). There were around 2–12 hormone response elements identified in each ZmPUB gene. Among these hormone-response elements, the MeJA and JA response element (as-1 and CGTCA-motif), ETH response element (MYC), and ABA response element (ABRE) were present in more than 80 percent of ZmPUB genes. Moreover, 46 distinct elements contributed to the biotic and abiotic stress responses, which included pathogen defense (3), light (28), drought (5), wound (3), cold (2), hypoxia (2), heat (1), salt (1), low osmotic pressure (1), and other abiotic stress (1). Among these, the light-response element (G-box) was present in the promoter regions of 84.4% ZmPUB genes. Additionally, the metabolism- and development-related elements harbored five seed or endosperm development elements (O2-site, GCN4_motif, OCT, RY-element, and MBSI), three meristem expression-associated elements (CAT-box, CCGTCC-box, and HD-Zip 1), two gene expression-related elements (CARE and AT-rich element), one circadian regulatory element (circadian), one root-specific regulatory element (motif I), and one cell cycle regulatory element (MSA-like).
To explore a possible role of ZmPUB genes during root development, we employed the RNA-seq data from the MaizeGDB website, which is a submission by prior study [37]. The RNA-seq data revealed that ZmPUB genes were differentially expressed during root development (Figure 7, Table S6). Three days after seed sowing, 21 ZmPUBs were upregulated in the differentiation zone of the primary root, whereas, only eight and four ZmPUBs were upregulated in the meristematic zone and stele of the primary root, respectively. Further, 6–7 days after seed sowing, 19 and 18 ZmPUBs exhibited elevated expression in the primary and seminal roots. In the V7 stage, 31, 21, and 2 ZmPUBs were highly expressed in the crown root nodes 1 through 3, crown root node 4, and crown root node 5, respectively. In the V13 stage, a total of 26 ZmPUBs revealed elevated expression in the crown root node 5. Interestingly, we noticed that ZmPUB4, ZmPUB8, ZmPUB16, ZmPUB23, ZmPUB27, ZmPUB35, ZmPUB43, ZmPUB49, ZmPUB66, and ZmPUB78 showed a relatively high expression in varying root locations (FPKM > 10) (Table S6). Among these genes, four ZmPUB genes contained the GA response element, five ZmPUB genes contained the IAA response element, and nine ZmPUB genes contained the meristem expression-related element (Tables S4 and S5). Lastly, seven ZmPUB genes were not at all expressed in the root.
To explore the function of ZmPUBs, we analyzed the expression profiles of ZmPUBs during leaf development, and the gene expression data were provided by a previous study [38]. In the previous study, the fully expanded third leaf of the seeding maize B73 plants was sliced into 15 sections (M1–M15) from the base to the tip, and a gradient gene expression analysis was performed from immature to mature leaf tissue. There were 34 highly expressed ZmPUB genes, which were mainly concentrated within the basal part of the leaf (Figure 8, Table S7). These evidences suggested that these genes may contribute to cell division and elongation. Eight ZmPUB genes were identified, which were highly expressed only in the middle region of the leaf. Three ZmPUB genes, namely ZmPUB1, ZmPUB48, and ZmPUB62, were only upregulated in the tip of the leaf. In addition, we identified ZmPUB43, ZmPUB44, ZmPUB53, ZmPUB54, ZmPUB62, and ZmPUB78 with very high expression (FPKM > 20) in the different parts of the leaf, indicating that these may significantly modulate leaf development. The promoter regions of the highly expressed genes were further analyzed; six genes contained the MeJA, JA, ETH and light-response elements, five genes contained the ABA response elements, five genes contained the meristem expression-related element, and four genes contained the GA response elements (Tables S4 and S5). Therefore, it is our belief that these highly expressed genes in various locations of leaves may be crucial for photosynthesis, differentiation, and hormonal responses. Lastly, we also found that seven ZmPUB genes were not at all expressed in the leaves.
To explore the function of ZmPUB genes during seed development, we gathered the seed development RNA-seq data from prior investigations (Table S8) [39,40]. Overall, 50 ZmPUB genes were expressed during the whole-seed development. Among them, 31 ZmPUB genes were present within the embryo, 9 ZmPUB genes were present in the endosperm, and 12 ZmPUB genes were present in both the endosperm and during embryo development (Figure 9). We also observed multiple highly expressed PUB genes, including ZmPUB3, ZmPUB4, ZmPUB8, ZmPUB23, ZmPUB24, ZmPUB43, and ZmPUB52 (FPKM > 10), which suggested that they may associate with seed development (Table S8). Among these genes, seven genes harbored the ETH and ABA response elements, six genes contained the meristem expression-related element, four genes contained the gliadin metabolic regulatory element, three genes contained the MeJA and JA response elements, two genes contained the GA response elements, one gene contained the IAA response elements, and one gene contained the endosperm formation regulatory element (Tables S4 and S5). In previous studies, hormones were reported as information transfer substances, which served essential functions in crop development, particularly, during grain filling [41,42,43]. This indicates that these highly expressed genes may participate in seed development. Additionally, we also identified 10 ZmPUB genes that were not expressed during seed development.
To elucidate changes in the transcriptome response to abiotic stress, RNA-seq was carried out in maize under control and stress conditions [44]. Under cold stress (7 °C for 16 h), 31 ZmPUB genes were markedly elevated in maize seedlings relative to control, 12 of which showed high expression (fold change >1). ZmPUB45 was the most up-regulated gene (fold change >36). Twenty-four ZmPUB genes were down-regulated in maize seedlings under cold stress. Among them, ZmPUB29, ZmPUB61, and ZmPUB73 displayed the most down-regulated expression, which were 31%, 24%, and 35% of control, respectively (Figure 10, Table S9). Under heat stress (50 °C for 4 h), sixteen ZmPUB genes were highly expressed in maize seedlings relative to control, seven of which showed extremely high expression (fold change >1). ZmPUB30 showed the highest up-regulated expression, which was up-regulated by 12.5 times. Forty ZmPUB genes were down-regulated, nineteen of which were significantly down-regulated by more than 50% (Figure 10, Table S9). Under salt stress, twenty-six ZmPUB genes were up-regulated in maize seedlings, and among them, eight genes were up-regulated by more than 1-fold, and ZmPUB45 and ZmPUB57 were the most up-regulated genes, with increases by 25.5 and 84.8 times, respectively. Thirty-two genes were down-regulated in maize seedlings as compared to controls, and four genes (including ZmPUB1, ZmPUB27, ZmPUB40, and ZmPUB78) showed the most down-regulated expression, with reductions of 67%, 69%, 82%, and 63%, respectively (Figure 10, Table S9). Under UV stress, thirty-nine ZmPUB genes were up-regulated, and seven genes were up-regulated more than 1-fold; among them, ZmPUB27 and ZmPUB45 had the most up-regulated expression, which were 5.18 and 19.8 times, respectively. Twenty-five ZmPUB genes were down-regulated, four of which were down-regulated by more than 50% (Figure 10, Table S9). In addition, nine ZmPUB genes, including ZmPUB7, ZmPUB9, ZmPUB12, ZmPUB28, ZmPUB41, ZmPUB56, ZmPUB74, ZmPUB77, and ZmPUB79, were not influenced by various abiotic stressors. Drought stress affected plants particularly during reproduction. To investigate the potential role of ZmPUB genes in drought stress, we collected the RNA-seq data of drought-exposed and appropriately watered fertilized ovary and basal leaf meristem tissue (Figure 10, Table S9) [45]. Under drought stress, 36 up-regulated and 30 down-regulated genes were screened in the ovary, among which 41 genes were altered more than 50%. ZmPUB49 was the gene with the maximum change in expression, which was up-regulated by 6.5 times. There were 32 up-regulated and 30 down-regulated genes in the basal leaf meristem following drought stress, among which 20 genes were altered more than 50%. Four genes (including ZmPUB2, ZmPUB31, ZmPUB57, and ZmPUB78) exhibited the maximum alterations in expression, which increased by 14.0, 5.3, 5.6, and 8.0 times, respectively. Lastly, nine genes were not expressed under drought stress in the ovary or basal leaf meristem tissue.
To further elucidate whether hormone affect ZmPUBs expression, we conducted RNA-seq under the hormone treatments (ABA, IAA, and GA). Based on our observations, 42 (53.16%) ZmPUB genes were detected in the leaf tissue (FPKM > 1) (Figure 11, Table S10). Thirty-nine ZmPUB genes responded to the hormones ABA, IAA, and GA but with variable degrees of sensitivity. Under ABA treatment, 35 ZmPUB genes were significantly altered in expression pattern, and among them, 13 genes were down-regulated, 20 genes were up-regulated, and 2 genes were down-regulated at the early stage (3 or 6 h) and up-regulated at the late stage (12 h). Under IAA treatment, 15 ZmPUB genes were altered, among which 5 were reduced, and 10 were augmented. Under GA treatment, 17 ZmPUB genes were altered, 6 of which were significantly diminished at one stage, and 11 were significantly elevated at one or two stage. As a result, we identified eight ZmPUB genes that showed the highest degree of hormone-induced expression, and they were ZmPUB1, ZmPUB4, ZmPUB43, ZmPUB45, ZmPUB57, ZmPUB64, ZmPUB68, and ZmPUB78, respectively. These evidences suggested that these ZmPUB genes contributed to the hormone-responsive network.
The PUB genes are common among mammals, plants, and microorganisms, and they encode the PUB ligase enzyme, a key enzyme in ubiquitin proteasome-mediated degradation [8]. The genome-wide analysis of PUB genes was performed in numerous species, and they were found to play essential roles in modulating plant development and response to biotic, abiotic, and hormonal stressors [1,11,22,23]. Herein, we screened 79 ZmPUB genes in the maize genome and designated ZmPUB1 through ZmPUB79 based on their location on the chromosome. Phylogenetic analysis revealed that these ZmPUB genes can be clustered into seven categories. Similar results were found in rice, soybean, barley, citrus, and banana (Musa paradisiaca), wherein the genes were classified into 6–8 groups using phylogenetic analysis [15,16,18,46]. Further analysis revealed that the 36 ZmPUB genes encoded proteins that harbor only the U-box domain, and 43 ZmPUB genes encoded proteins that contain the U-box domain, along with additional 10 domains, including the ARM repeat (25), coil coil region(10), pkinase-tyr (9), TPR (2), UFD2P_CORE (2), ZnF_TTF (1), WD40 repeat (1), KAP (1), RS4NT (1), and USP (1) domains (Table S1). In Arabidopsis, 16 AtPUB proteins harbor only the U-box domain, whereas 48 AtPUB proteins contain both the U-box domain and several other domains, such as ARM repeat (27), kinase (14), WD40 repeat (2), MIF4G (2), cyclophilin (1), UFD2 (1), and TPR domains (1) [13]. In rice, the domain composition of PUB proteins are similar to Arabidopsis; all OsPUB proteins contain the U-box domain, and 57 OsPUB proteins contain six types of domains except MIF4G [12]. In higher eukaryotes, genomic evolution can lead to the gene expansion and functional differentiation of gene families. Being the core domain of the PUB gene family, the U-box domain is evolutionarily conserved. Meanwhile, there are plenty of other domains observed in numerous species, which might introduce functional diversity to the PUB genes. In eukaryotes, genes are usually composed of exons, introns and untranslated regions (UTRs) at the 5′ and 3′ ends. The structural organization is closely related to gene function and family evolution [47]. In the maize PUB gene family, there was a wide variation in the number of exons; the mean number of exons per gene was 3.61, and it ranged from 1 to 16. Around 37.97% of the ZmPUB genes contained only one exon and no intron (Figure 2, Table S2). In prior investigations, about 1 to 17 exons were identified in the PUB genes of tomato and Medicago, with an average of four exons per gene, and the PUB genes with only one exon and no intron accounted for 34.14% and 40.32% of all PUB genes [17,21]. The structural organization of PUB genes are similar in different species. It is suggested that the PUB gene family is evolutionarily conserved. Exons possess the core information needed by the cell to synthesize protein, whereas introns protect the coding proteins from randomly generating deleterious mutations [48]. In the PUB gene family, numerous intron-less genes were reported in multiple species, such as tomato, grape, citrus, Medicago, and Chinese cabbage [14,17,18,21,49], suggesting a strong structural integrity of PUB genes. Furthermore, the conserved motif distributions suggested a strong structural and functional similarity among maize PUB genes (Figure 2, Table S2). For instance, motifs 1, 3, and 4 are highly conserved and exist in almost all PUB genes. We also observed some unique features in some groups of the maize PUB gene family. Motifs 5 and 2 tandem sequences were mainly located in group IV. Motifs 6, 7, and 9 were the exclusive feature of group II. These analyses are highly beneficial to scientists who are interested in the evolution, structure, and function of the PUB gene family. Gene duplication (GD), mutation, and natural selection are the major sources of new genes and functions, and they provided a basis for biodiversity [50]. Gene duplication and syntenic analysis identified 25 pairs (43 ZmPUB genes) of SD and one pair of TD (ZmPUB43/ZmPUB44) in the maize PUB gene family (Table 1). GD is speculated to be the major contributor of expansion and diversity in the maize PUB gene family. We next evaluated the domain composition of the maize duplicate genes. New domains were introduced to the newly duplicated genes; five genes were added to the ARM domain (ZmPUB10, ZmPUB19, ZmPUB39, ZmPUB43, and ZmPUB63), two genes to the coil coil region (ZmPUB62 and ZmPUB8), and one gene to the KAP domain (ZmPUB49) (Table 1 and Table S1). The addition of new domains to the duplicated genes may have introduced diversification of function to PUB genes. The maize PUB gene family was predicted to harbor GD events around 13–84 Mya and 129–215 Mya (Table 1). Based on our analysis, there were approximately 0.05–0.99 synonymous substitutions per site, which indicated that the maize PUB gene family evolved under purifying selection, whereby deleterious mutations were eradicated while the core functional domains were conserved [51]. In addition, the maize PUB genes displayed a stronger association with rice than Arabidopsis (Figure 5). This suggested that rice is a more suitable reference for the functional study of the maize PUB gene family. Cis-acting regulatory element analysis identified 94 kinds of elements in the promoter sequences of the maize PUB gene family. These were 19 hormone-response elements, 46 biotic- and abiotic-response elements, 13 metabolism- and development-related elements, and 13 unknown function elements (Tables S4 and S5). Similar element patterns were reported in tomato, citrus, and Medicago, which suggested a common relationship between PUB genes and stresses as well as hormonal and developmental modulatory mechanisms in plants [17,18,52]. Emerging evidences revealed that the PUB genes played essential functions in modulating plant development and in developing tolerance to abiotic stresses such as cold, heat, salt, and drought [53,54,55,56]. Using available transcriptome data and RNA-seq analysis under various treatments, we examined the expression profile of the maize PUB genes in this research. Our work could provide new clues to the biological function of PUB genes in maize. The PUB genes were known to regulate development of roots, stems, leaves, flowers, and fruits in plants [17,57]. Herein, we identified different expression profiles of maize PUB genes during the root development (Figure 7, Table S6). In total, 10 genes (ZmPUB4, ZmPUB8, ZmPUB16, ZmPUB23, ZmPUB27, ZmPUB35, ZmPUB43, ZmPUB49, ZmPUB66, and ZmPUB78) were highly expressed at certain stages of root development (FPKM > 10). Further analysis revealed that GA, IAA, and meristem expression-related elements were present in the promoter of these genes (Tables S4 and S5). The root apical meristem-specific genes played essential roles in root development during early somatic embryogenesis [58]. IAA strongly promoted the development of adventitious roots, lateral roots, root hairs, and primary roots. GA determined cell growth polarity in the root cortex of maize [59,60,61]. It is suggested that these genes participate in root morphogenesis. Six highly expressed genes were observed during leaf development, and they were ZmPUB43, ZmPUB44, ZmPUB53, ZmPUB54, ZmPUB62, and ZmPUB78, respectively (Figure 8, Table S7). We screened the MeJA, JA, ETH, and light-response elements in the promoters of these genes (Tables S4 and S5). MeJA, JA, and ETH regulated plant growth. The enhanced endogenous MeJA levels induced morphological alteration within leaves in the transgenic soybean plants [62]. JA and ETH may positively regulate cotton leaf senescence [63]. Leaf is an important photosynthetic organ in plants, and light is known to modulate plant morphogenesis and the leaf ultrastructure [64]. The presence of MeJA, JA, ETH, and light-response elements provided evidence that these genes may be involved in leaf development. Similarly, seven highly expressed genes (ZmPUB3, ZmPUB4, ZmPUB8, ZmPUB23, ZmPUB24, ZmPUB43, and ZmPUB52) were spotted during seed development, four of which harbored the gliadin metabolic regulatory element (O2-site) in their promoter regions (Figure 9, Table S8). Gliadin is the most abundant storage protein in maize seed, and it is critical for the nutritional profile of maize seeds [65]. Based on our gene expression analysis, these genes were ubiquitously expressed at 2–14 days of whole-seed development. This suggested that they may participate in seed development by modulating protein accumulation within the endosperm. The PUB genes modulate response to abiotic and hormonal stresses [52,58]. Based on the RNA-Seq data, there were several stress- and hormone-responsive genes, such as ZmPUB1, ZmPUB27, ZmPUB30, ZmPUB31, ZmPUB43, ZmPUB45, ZmPUB57, and ZmPUB78 (Tables S9 and S10). ZmPUB1 and OsPUB75 were homologous genes. OsPUB75 encoded a cytosolic RING-type E3 ubiquitin ligase, which was a crucial negative modulator of abiotic stress. Under salinity and mannitol stress, OsPUB75 was transcriptionally repressed in Arabidopsis [66]. In maize, ZmPUB1 displayed an expression profile similar to OsPUB75, and it was down-regulated by 26–83% under heat, salt, UV, drought, and ABA treatments (Tables S9 and S10). This suggested that ZmPUB1 may participate in the negative regulation of abiotic stresses and hormonal responses. ZmPUB27 was homologous with AtPUB16 and OsPUB3. In Arabidopsis, AtPUB16 participated in the GA pathway, which favored self-pollination [67]. In maize seedling, a similar ZmPUB27 up-regulation was observed at 3 and 12 h after GA treatment. This indicated that ZmPUB27 may contribute to the GA response. In rice, OsPUB3 was a positive regulator of cold stress, and its over-expression in transgenic plants showed enhanced tolerance to cold stress compared to wild-type plants [53]. ZmPUB27 expression was elevated under conditions of cold, drought, and UV, and it was reduced after heat and salt treatment. This suggested that ZmPUB27 was involved in abiotic stresses. ZmPUB30 and ZmPUB57 were a pair of duplicated genes, and they were homologous to OsPUB5, AtPUB18, and AtPUB19. In rice, OsPUB5 encoded a copper (Cu) transporter 6 protein, which was highly expressed in the node, and it blocked the cadmium (Cd) upward transport [68]. AtPUB18 and AtPUB19 were positively regulated by ABA and salt, and their double mutants could not respond well to ABA and the salt-based suppression of seed germination relative to wild-type plants [69]. Under cold, heat, salt, and ABA treatments, ZmPUB30 and ZmPUB57 were up-regulated by 0.5–12.5 or 3.67–84.82 times. The conserved motif analysis revealed that ZmPUB57 contained more of motifs 2, 5, and 8 than ZmPUB30. This may greatly enhance the stress response ability of ZmPUB57. ZmPUB31 and ZmPUB45 were a pair of duplicated genes. Under cold, heat, salt, drought, UV, and ABA stresses, ZmPUB31 and ZmPUB45 were up-regulated by 0.19–3.8 and 3.1–36.6, respectively (Tables S9 and S10). MeJA and JA were naturally occurring physiologically active materials that respond to exoteric stimulations by transmitting stress signals and activating the stress-resistant genes in plants [70,71,72]. Compared to ZmPUB31, there were more JA and MeJA response elements in the promoter sequence of ZmPUB45 (Tables S4 and S5). This may enhance the responsive capacity of ZmPUB45 under environmental stress. OsPUB45 was homologous with ZmPUB31 and ZmPUB45. In rice, OsPUB45 was up-regulated many-folds under salt, drought, and cold stresses [73]. We observed a similar expression pattern between ZmPUB45 and OsPUB45, which indicated that these two genes may possess comparable functions. ZmPUB43 was the homologous gene of OsPUB8. In rice, OsPUB8 was markedly up-regulated during drought stress [74]. ZmPUB43 was up-regulated by 29–36% in the maize ovary and meristem under drought stress (Table S9). Further analysis revealed that there were significant alterations in expression of ZmPUB43 after salt, UV, and ABA treatments. ABA was critical for multiple biological processes, such as seed dormancy, germination, and adaptive responses to abiotic stresses [75,76]. This suggested that ZmPUB43 may contribute to abiotic stress responses via the ABA pathway. ZmPUB78 was homologous to AtPUB11. AtPUB11 was up-regulated under drought and ABA treatments, and its silent mutant was highly tolerant to drought stress compared to the wild type [77]. In maize, the expression levels of ZmPUB78 were up-regulated by 2.8–8.0 fold after ABA and drought treatments, which suggested that ZmPUB78 may negatively regulate the ABA-mediated drought response. In contrast, ZmPUB78 was down-regulated by the cold, heat, and salt stresses, thus indicating that ZmPUB78 may also be involved in muti-abiotic stresses. Above all, these results provided insights into the possible biological functions of the maize PUB gene family. Our extensive analyses aided in the selection of specific PUB genes for additional functional research so as to improve the genetic agronomic characteristics and strengthen environmental resistance within maize.
The maize genome sequences (B73 RefGen_v4) were obtained from the Ensembl Plants database (https://plants.ensembl.org/index.html (accessed on 1 March 2022)). The Hidden Markov Model (HMM) profile of the U-box domain (PF04564) was retrieved from the Pfam database (http://pfam.xfam.org/ (accessed on 5 March 2022)). The HMMER program was searched in the maize genome with default parameters and a significant e−3 value. Next, we confirmed the putative ZmPUB genes using the Pfam (http://pfam.xfam.org/ (accessed on 5 March 2022)), SMART (http://smart.embl-heidelberg.de/ (accessed on 5 March 2022)), and NCBI CDD (https://www.ncbi.nlm.nih.gov/cdd (accessed on 5 March 2022)) databases. The putative ZmPUB protein MW (molecular weight) and PI (isoelectric point) were analyzed using ExPASy (http://www.expasy.org (accessed on 5 March 2022)). WoLF PSORT (https://www.genscript.com/wolf-psort.html (accessed on 5 March 2022)) was employed to estimate the subcellular localization of ZmPUB proteins.
Multiple sequence alignments were performed using the U-box domain sequences of the 219 PUB proteins (79 ZmPUBs, 77 OsPUBs, and 64 AtPUBs) using MEGA 7 [78]. A phylogenetic tree was generated using the neighbor-joining (NJ) method and the following parameters: Poisson correction, complete deletion, and 1000 bootstrap replicates, and visualization was done in the EvolView v2 software [79,80]. The Oryza sativa and Arabidopsis thaliana PUB protein sequences were obtained from the NCBI website (https://www.ncbi.nlm.nih.gov/ (accessed on 6 May 2022)), as reported in previous studies [12,13].
The genetic structure analysis was conducted via comparison of the CDS and DNA sequences of the ZmPUB genes in GSDS 2.0 (http://gsds.cbi.pku.edu.cn/ (accessed on 6 May 2022)), SMART (http://smart.embl-heidelberg.de/ (accessed on 6 May 2022)) and Pfam (http://pfam.xfam.org/ (accessed on 6 May 2022)) softwares [81]. The conserved motifs were analyzed in MEME using parameters as follows: maximum motif number (10) and motif length (6–100 amino acid residues) (http://meme-suite.org (accessed on 6 May 2022)) [82], and visualization was performed in Tbtools (https://github.com/CJ-Chen/Tbtools (accessed on 6 May 2022)).
The gene lengths and position data were acquired from the maize genome (B73 RefGen_v4). Next, MapChart was employed for the construction of the chromosomal localization map [83]. To assess GD events, the coding sequences of PUB genes from Arabidopsis, maize, and rice were aligned in BLASTp using an E-value < 1 × e−10 cut-off. MCScanX was employed to assess GD events and examine syntenic association between different species, and visualization was performed in Circos [84,85]. The Kaks_Calculator computed the nonsynonymous (Ka) and synonymous (Ks) substitution rates of duplicated gene pairs, and the approximate time of GD was estimated as follows: T = (Ks/2λ) × 10−6, where λ = 6.5 × 10−9 [86].
We downloaded the 1500 bp upstream flanking sequences from the transcription start site of each ZmPUB gene. Then the cis-acting regulatory element analysis was carried out in PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/ (accessed on 10 May 2022)) [87].
The RNA-seq information for ZmPUB genes in numerous tissues, developmental stages, and stress conditions were acquired from NCBI (https://www.ncbi.nlm.nih.gov/ (accessed on 6 June 2022)) and MaizeGDB (https://www.maizegdb.org/ (accessed on 6 June 2022)) and were reported in previous studies [38,39,40,88]. The expression heat maps of ZmPUB genes were drawn in the omicshare website (https://www.omicshare.com/ (accessed on 10 June 2022)). Zea mays cv. B73 was employed for the examination of gene profile in response to multiple hormonal treatments. All seeds were cultivated in commercial soil at 28 °C in a photoperiod of 16 h light/8 h dark. Three-leaf stage seedlings were treated with 100 μmol/L ABA, 100 μmol/L IAA, and 100 μmol/L GA in 0.1‰ Tween-20, respectively. The controls were treated with water in 0.1‰ Tween-20. The leaves were harvested at 0, 3, 6, 12, and 24 h after hormonal treatments. Three replicates of three plants were used in each treatment. All samples were flash-frozen in liquid nitrogen prior to storage in −80 °C until subsequent analysis. Total RNA extraction employed Trizol (Invitrogen, Carlsbad, CA, USA). The quality of RNA was assessed via the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Transcripts were enrich-fragmented into small pieces, which were then converted to cDNA. The cDNA library was generated with the NEBNext Ultra RNA Library Prep Kit for Illumina (NEB #7530, New England Biolabs, Ipswich, MA, USA), and sequencing was completed with Illumina Novaseq 6000 by Gene Denovo Biotechnology Co., (Guangzhou, China). To quantify gene abundance, we filtered reads and mapped the clean reads to the reference genome with the fastp, HISAT2. 2.4, and StringTie v1.3.1 softwares [89,90,91]. Finally, the FPKM (fragment per kilobase of transcript per million mapped reads) gene values were computed in the RSEM software [92].
To compare the variation of gene expression, SPSS 12.0 (SPSS Inc., Chicago, IL, USA) was employed to carry out the least significant difference (LSD) test. Data processing and visualization were done with GraphPad Prism 5. Individual data were presented as mean ± standard error (SE) of three experimental replicates.
Herein, we conducted an extensive analysis of PUB genes in maize. We identified 79 ZmPUB genes, which were stratified into 7 categories. Each group exhibited similar exon-intron structures and motif compositions. Using gene duplication and synteny analysis of PUB genes, we obtained important clues regarding the evolutionary profiles of maize PUB genes. PUB genes strongly regulated plant development as well as response to abiotic stresses and hormones. Our phylogenetic and gene expression analyses provided useful information for enhancing our comprehension of the biological roles of the PUB genes in maize. | true | true | true |
PMC9573493 | 36235659 | Hong-Ren Yu,Jiunn-Ming Sheen,Chih-Yao Hou,I-Chun Lin,Li-Tung Huang,You-Lin Tain,Hsin-Hsin Cheng,Yun-Ju Lai,Yu-Ju Lin,Mao-Meng Tiao,Ching-Chou Tsai | Effects of Maternal Gut Microbiota-Targeted Therapy on the Programming of Nonalcoholic Fatty Liver Disease in Dams and Fetuses, Related to a Prenatal High-Fat Diet | 27-09-2022 | prenatal,high-fat diet,DOHaD,nonalcoholic fatty liver disease,Lactobacillus reuteri,butyrate | Metabolic disorders can start in utero. Maternal transmission of metabolic phenotypes may increase the risks of adverse metabolic outcomes, such as nonalcoholic fatty liver disease (NAFLD); effective intervention is essential to prevent this. The gut microbiome plays a crucial role in fat storage, energy metabolism, and NAFLD. We investigated the therapeutic use of probiotic Lactobacillus reuteri and postbiotic butyrate gestation in the prevention of perinatal high-fat diet-induced programmed hepatic steatosis in the offspring of pregnant Sprague–Dawley rats who received regular chow or a high-fat (HF) diet 8 weeks before mating. L. reuteri or sodium butyrate was administered via oral gavage to the gestated rats until their sacrifice on day 21 of gestation. Both treatments improved liver steatosis in pregnant dams; L. reuteri had a superior effect. L. reuteri ameliorated obesity and altered the metabolic profiles of obese gravid dams. Maternal L. reuteri therapy prevented maternal HF diet-induced fetal liver steatosis, and reformed placental remodeling and oxidative injury. Probiotic therapy can restore lipid dysmetabolism in the fetal liver, modulate nutrient-sensing molecules in the placenta, and mediate the short-chain fatty acid signaling cascade. The therapeutic effects of maternal L. reuteri on maternal NAFLD and NAFLD reprogramming in offspring should be validated for further clinical translation. | Effects of Maternal Gut Microbiota-Targeted Therapy on the Programming of Nonalcoholic Fatty Liver Disease in Dams and Fetuses, Related to a Prenatal High-Fat Diet
Metabolic disorders can start in utero. Maternal transmission of metabolic phenotypes may increase the risks of adverse metabolic outcomes, such as nonalcoholic fatty liver disease (NAFLD); effective intervention is essential to prevent this. The gut microbiome plays a crucial role in fat storage, energy metabolism, and NAFLD. We investigated the therapeutic use of probiotic Lactobacillus reuteri and postbiotic butyrate gestation in the prevention of perinatal high-fat diet-induced programmed hepatic steatosis in the offspring of pregnant Sprague–Dawley rats who received regular chow or a high-fat (HF) diet 8 weeks before mating. L. reuteri or sodium butyrate was administered via oral gavage to the gestated rats until their sacrifice on day 21 of gestation. Both treatments improved liver steatosis in pregnant dams; L. reuteri had a superior effect. L. reuteri ameliorated obesity and altered the metabolic profiles of obese gravid dams. Maternal L. reuteri therapy prevented maternal HF diet-induced fetal liver steatosis, and reformed placental remodeling and oxidative injury. Probiotic therapy can restore lipid dysmetabolism in the fetal liver, modulate nutrient-sensing molecules in the placenta, and mediate the short-chain fatty acid signaling cascade. The therapeutic effects of maternal L. reuteri on maternal NAFLD and NAFLD reprogramming in offspring should be validated for further clinical translation.
Evidence indicates that maternal obesity will be a major determinant of the “developmental origins of health and disease” in the next generation. Obesity before pregnancy and gestational weight gain, especially in early pregnancy, may increase the risk of obesity in the offspring [1,2,3]. Adverse cardiovascular risks in the offspring, such as coronary heart disease, type 2 diabetes, and stroke, also increase from childhood to adulthood [1,4]. Non-alcoholic fatty liver disease (NAFLD), an excessive accumulation of fat in the liver, is the most common form of chronic liver disease and is closely related to obesity. A meta-analysis conducted in Asia between 1999 and 2019 showed that the incidence of NAFLD was 50.9 cases per 1000 person-years [5]. Yunosi et al. estimated the global prevalence of NAFLD to be as high as 25% [6]. In an autopsy study conducted in San Diego, the authors enrolled 742 children and found that the prevalence of fatty liver was 9.6%, whereas the prevalence in obese children was as high as 38% [7]. NAFLD is strongly associated with metabolic syndrome and insulin resistance. NAFLD can eventually progress to hepatic cirrhosis, cardiovascular problems, malignancy, and chronic kidney disease [8,9]. Hepatic steatosis in children can be observed in early childhood developmental stages and even in utero [10]. The prenatal predisposing factors for NAFLD include maternal obesity, maternal metabolic syndrome, and gestational diabetes [11,12]. Fetuses nurtured by dams with high-fat (HF) diets during pregnancy can develop fatty livers [13]. The disease process of NAFLD is currently considered to be related to abnormal lipid metabolism, inflammation, and oxidative stress [14,15]. In a previous animal study of obese dams fed HF diets, we showed that the alteration of the placenta through oxidative stress and metabolism-related transcriptomes and a change in the short-chain free fatty acid cascade leads to lipid dysmetabolism and steatotic changes in the fetal liver. Pregnant women are more likely to be on HF diets and tend to be obese; therefore, providing a safe and effective strategy to prevent fatty liver in the offspring during pregnancy is an important issue. Many studies have shown an association between gut dysbiosis and metabolic syndrome, obesity, type 2 diabetes, and NAFLD [16,17,18,19]. Ge et al. showed that adult patients with NAFLD demonstrated lower diversity and a Firmicutes/Bacteroides (F/B) ratio than the controls [20]. Chierico et al. found an increased abundance of Bradyrhizobium, Anaerococcus, Peptoniphilus, Propionibacterium acnes, Dorea, and Ruminococcus, and a decreased abundance of Rikenellaceae and Oscillospira in the gut microbiome of children with NAFLD [16]. In animal studies, the gut microbiota of pregnant dams with hepatic steatosis correlating with a HF diet showed a low level of alpha diversity and Lachnospiraceae genus, but more Romboutsia and Akkermansia genera than the control group [13]. Current strategies to alleviate NAFLD primarily include lifestyle changes, such as exercise, weight loss, and dietary control. Since the gut microbiome plays a crucial role in body fat storage, energy metabolism, and inflammatory response, these mechanisms also correspond to the fatty liver. Manipulation of the gut microbiota has the potential to be a deprogramming strategy for NAFLD in pregnant mothers and even offspring. The interaction between the gut microbiota and the host is mediated via its metabolites, which are produced by the fermentation of food substances by the gut microbiota [21]. Postbiotics are substances released or produced through the metabolic activities of microorganisms that exert beneficial effects on the host [22]. Short-chain fatty acids (SCFAs) are polysaccharide fermentation products produced by the gut microbiota. SCFAs have been shown to modulate glucose homeostasis, lipid metabolism, and immunity [23]. SCFAs have also been shown to improve the gut barrier and mucosal immune function, and ameliorate a HF diet-induced fatty liver [24,25,26,27]. The most well-known products of SCFAs are acetic acid, propionic acid, and butyric acid. Among these, butyrate acids have received the most attention. Butyrate is the primary energy source for colonocytes. However, it is uncertain whether treatment with butyrate can ameliorate hepatic abnormalities generated by diet-induced maternal obesity. Observations on the long-term effects of maternal obesity on offspring have significant public health implications. The prevalence of obesity in women of reproductive age is increasing worldwide, thus contributing to an increased risk of exposure of the offspring to an “obese intrauterine environment.” The latter environment perpetuates the vicious cycle of maternal–offspring obesity and increases the burden of chronic diseases. There is an urgent need to develop safe and effective interventions to stop this vicious cycle. In this study, we examined the use of Lactobacillus reuteri (L. reuteri) and sodium butyrate to attenuate the effects of maternal obesity on the offspring of pregnant dams.
Seven-week-old virgin female Sprague–Dawley (SD) rats purchased from BioLASCO (Taipei, Taiwan) were housed in a humidity-, temperature-, and light-controlled environment [13]. Food and sterile tap water were provided ad libitum. At one week of adaptation to the experimental environment, the rats were weight-matched and assigned to receive either a HF diet or a regular control diet. The Institutional Animal Care and Use Committee of the Chang Gung Memorial Hospital approved the experimental protocol (no. 2019053001). The HF diet (D12331) and the control diet feeds were purchased from the Research Diets (Research Diets, New Brunswick, NJ, USA) and Fwusow Industry (Fwusow Industry, Taichung, Taiwan), respectively. The ingredients of the dietary feeds are listed in Supplementary Table S1. The rats were fed the specified diet for eight weeks and then mated for two days. Dams and their fetuses were divided into four groups randomly: control chow (CC), high-fat (HF), high-fat and L. reuteri (H + L), and high-fat and sodium butyrate H + B (n = 6 for each group). Two pregnant dams were assigned to each group, and each group produced approximately 20 offspring. “Group I CC”: maternal rats commenced a control chow diet prior to mating, and were continued on this diet until their sacrifice on gestational day 21 (GD21). “Group II HF”: maternal rats commenced a HF diet 8 weeks prior to mating and were fed this diet until their sacrifice on GD21. “Group III H + L”: maternal rats commenced a HF diet 8 weeks prior to mating and the pregnant rats were administered L. reuteri (GMNL-89, GenMont Biotech, Inc., Taipei, Taiwan) (1 × 109 colony-forming units (CFU)/day) by gavage from GD0 until their sacrifice on GD21. “Group IV H + B”: maternal rats commenced a HF diet 8 weeks prior to mating and the pregnant rats were administered 1% (w/v) sodium butyrate (B5887, SIGMA, St. Louis, MO, USA) in drinking water (150 mg/kg/day) by gavage from GD0 until their sacrifice. For further experiments on L. reuteri, other groups were included. “Group I CC”: maternal rats were foddered a control chow diet before mating and during gestation until sacrifice on GD21. “Group II HF”: maternal rats were foddered a HF diet for 8 weeks before mating and during gestation until sacrifice on GD21. “Group III L”: maternal rats were foddered a control chow diet for 8 weeks before mating and during gestation, and the pregnant rats were administered L. reuteri (1 × 109 CFU/day) from GD0 until GD21. “Group IV H + L”: maternal rats were foddered a HF diet for 8 weeks before mating and during gestation, and the pregnant dams were administered L. reuteri (1 × 109 CFU/day) from GD0 until sacrifice on GD21.
The dams were sacrificed by anesthetization with Zoletil (25 mg/kg) (tiletamine-zolazepam, Virbac; Carros Cedex, France) and Rompun (23.32 mg xylazine hydrochloride, Bayer, Korea) in a 1:1 mixture by intramuscular injection, followed by cardiac puncture and perfusion. Heparinized blood samples were obtained via a cardiac puncture [13,28]. After the dams were sacrificed, the placenta, liver tissue, and fetal liver were collected by cesarean section as previously described [13]. A portion of the tissue was fixed in 10% formalin for histological analysis, and the remainder was stored at −80 °C for quantitative reverse transcription PCR (RT-qPCR) study.
The dams were subjected to weekly BW checks. The indirect tail-cuff method (BP-2000, Visitech Systems, Apex, NC, USA) was used for BP measurements five days before sacrifice.
For the intraperitoneal glucose tolerance test (IPGTT), the rats were fasted for 8 h, and 50% glucose (4 mL/kg of BW) was injected intraperitoneally. The tail vein was used to determine serum sugar levels via a blood glucose meter (Accu-Chek, Roche, Germany) before glucose injection and at 15, 30, 60, and 120 min after glucose injection [13,29]. The integrated area under the curve (AUC) of IPGTT was computed using the trapezoidal method.
Glutamic-oxaloacetic transaminase (GOT), glutamic-pyruvic transaminase (GPT), and total cholesterol (T-chol) levels were measured using an automatic biochemical analyzer (Hitachi model 7450; Hitachi, Tokyo, Japan), according to the manufacturer’s manuals [30]. An enzyme-linked immunosorbent assay (ELISA) kit (Abcam, Cambridge, MA, USA) was used to determine the serum leptin levels (n = 6 per group). The ELISA assay principle relies on the formation of an antibody/antigen complex that attaches the target to the surface of the detection plate and allows the target to be detected and quantified.
Placenta and liver tissues were fixed with 4% paraformaldehyde at 4 °C overnight, dehydrated in a gradient of ethanol, hyalinized in xylene, and embedded in paraffin wax. Formalin-fixed tissues were cut and stained with hematoxylin and eosin (H&E). The slides were then scanned using a 3DHISTECH PANNORAMIC.SCAN slide scanner. Lipid accumulation in the liver was quantified using ImageJ software (Fiji version 1.8.0). Oxidative stress of the tissue was measured using 8-hydroxy-2-deoxyguanosine (8-OHdG), a product of DNA oxidation. Briefly, after transferring to polylysine-coated slides, the tissue sections were stained with an anti-8-OHdG antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA for 60 min at room temperature and a secondary antibody for 30 min after rinsing. Avidin and biotinylated horseradish peroxidase H were used to visualize the staining.
Total RNA was extracted from the tissues using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Messenger RNA (mRNA) expression in the fetal liver and placenta tissue was determined by RT-qPCR, as previously reported [13,31]. mRNA primers are listed in Supplementary Table S2. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and 18S ribosomal RNA (rRNA) were used as the housekeeping genes for the liver and placenta, respectively [13]. The comparative threshold cycle method was used to determine relative mRNA quantification [31].
Fecal samples were collected and stored at −80 °C for 1 week prior to sacrifice. The EZNA Soil DNA Kit (Omega. Bio-tek, Norcross, GA) was used to extract microbial DNA from stool samples. The bacterial 16S rRNA gene was amplified by RT-qPCR as previously described [32]. Sequencing amplicons were quantified using QuantiFluor-ST (Promega, San Luis Obispo, CA, USA) and analyzed on an Illumina MiSeq platform. Amplicon sequencing was performed using 300 bp paired-end raw reads, and the entire paired-end reads were assembled using FLASH v1.2.11. We carried out de-multiplexing based on barcode identification. For quality control, low-quality reads (Q < 20) were excluded from the QIIME v1.9.1 pipeline. The reads were truncated for three consecutive bases (<Q20). The dataset was retained for more than 75% of the original length using split_libraries_fastq.py script in QIIME. UCHIME was used as a chimera check to obtain effective tags, and it was filtered from the data set before operational taxonomic unit (OTU) clustering at 97% sequence identity using the UPARSE function in the USEARCH v7.0.1090 pipeline. The RDP classifier (v2.2) algorithm was employed to annotate the taxonomic classification for each representative sequence based on information retrieved from the Silva Database v132, and was performed with an 80% minimum confidence threshold to record an assignment. We filtered out sequences with one-time occurrences (or if they were present in only one sample). Multiple sequence alignments were conducted using PyNAST software (v1.2) against the core-set dataset in the Silva Database v132 to investigate the sequence similarities among different OTUs. A phylogenetic tree was constructed with a set of sequences representative of the OTUs using FastTree. For sequence similarities among different OTUs, multiple sequence alignments were conducted by using the PyNAST software (v1.2) against the core-set dataset in the Silva Database v132. A phylogenetic tree was constructed with a set of sequences representative of the OTUs using FastTree. To normalize the variations in sequence depths across samples, OTU abundance information was rarefied to the minimum sequence depth using the QIIME script (single_rarefaction.py). Subsequent analyses of alpha and beta diversities were performed using normalized data.
Gas chromatography was used to measure the plasma concentrations of acetic acid, propionic acid, and butyric acid [32]. Briefly, 5 μL of 100 μM internal standard was mixed with 100 μL of the sample and 100 μL of propyl formate. After vortexing and centrifugation, the supernatant was subjected to gas chromatography (GC) analysis using a Shimadzu QPlus 2010 gas chromatograph.
Continuous data were analyzed using the Kruskal–Wallis test or the one-way analysis of variance (ANOVA) with Tukey post hoc tests, as indicated; p value < 0.05 was considered statistically significant. Values are expressed as the mean ± standard error of the mean. The BW difference and the IPGTT test data between the groups were determined by repeated measures analysis. The interaction between the group and time (G × T) was calculated for each variable. Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) software version 19. We calculated the sample size as four rats per group with the glucose level at a 15 min mean difference between the two groups of 152.87 (=386.20 − 233.33) mg/dL (standard deviation) 50.97), this provided 80% power to detect such a difference using a two-sample t-test with a two-sided type I error of 0.05. We recruited 24 rats (six per group) to achieve roughly four rats per treatment group. The power test was performed using the G*Power 3.1.9.4.
Changes in the weekly BW of the animals are illustrated in Figure 1 and listed in Supplementary Table S3. After one week of diet manipulation, the dams receiving the HF diet had heavier BW than those receiving the control diet (234.30 ± 6.45 g vs. 210.28 ± 3.06 g; p = 0.032). In this study, L. reuteri and butyrate were administered after breeding. We found that L. reuteri treatment decreased the obesity of dams receiving the HF diet after two weeks of treatment (HF group vs. HF plus L. reuteri group; 358.26 ± 12.86 g vs. 325.41 ± 11.61 g; p = 0.021). Before delivery, the HF group dams maintained a higher BW than the other groups. L. reuteri treatment, rather than butyrate treatment, improved HF-diet-related BW increases. The biochemical markers are in Supplementary Table S4. For IPGTT, the HF group showed a higher glucose level at 15 min than the control group (HF group vs. control group; 233.33 ± 42.89 mg/dL vs. 386.20 ± 57.93 mg/dL; p = 0.021). Only L. reuteri treatment improved the IPGTT glucose level at 15 min (HF group vs. HF plus L. reuteri group; 386.20 ± 57.93 mg/dL vs. 231.44 ± 25.63 mg/dL; p = 0.022). (Supplementary Figure S1). Neither L. reuteri nor butyrate treatment affected the area under the curve (AUC) level of IPGTT. The HF diet increased caloric intake in the dams. L. reuteri treatment led to a decrease in both the total intake (Figure 2A) and intake per unit of BW (Figure 2B) in dams. This suggests that L. reuteri therapy may reduce the appetite of dams. A HF diet intake led to BP elevation, increased plasma leptin levels, and increased retroperitoneal fat deposits (Figure 2). Both L. reuteri and butyrate therapy improved the elevated BP related to the HF diet (Figure 2D). L. reuteri and butyrate treatments did not decrease plasma leptin levels or retroperitoneal fat deposits (Figure 2C,E). Thus, L. reuteri treatment during pregnancy ameliorates HF diet-related obesity through a leptin-independent pathway.
HF diets may lead to liver steatosis and oxidative stress in pregnant dams [13]. We determined the preventive effects of L. reuteri and butyrate intake on HF diet-related fatty liver disease in pregnant dams. HF intake resulted in macrovesicular steatosis (a single lipid droplet filling the entire cell with a displaced nucleus) and microvesicular steatosis (several small cytoplasmic vacuoles with a central nucleus) (Figure 3A). L. reuteri treatment during pregnancy greatly improved macrovesicular and microvesicular steatosis in dams induced by a HF diet. Butyrate treatment during pregnancy primarily improved macrovesicular steatosis compared to microvesicular steatosis. Exposure to a HF diet also resulted in liver oxidative stress in dams, as illustrated by 8-OHdG staining, one of the major products of DNA oxidation. Both L. reuteri and butyrate therapies improved liver oxidative stress in the dams (Figure 3B).
Maternal HF diet intake also caused steatosis and oxidative stress damage in the fetal liver (Figure 4A,B). We evaluated the preventive effects of maternal L. reuteri and butyric acid intake during pregnancy on fetal liver steatosis. We found that, although both maternal L. reuteri and butyric acid could ameliorate oxidative stress injury in the fetal liver, only L. reuteri was effective in improving fetal fatty liver related to the maternal HF diet. Since only maternal L. reuteri treatment could prevent hepatic steatosis in fetuses with prenatal HF exposure, we studied the mechanism by which prenatal L. reuteri intake could prevent fetal fatty liver. We found that acetyl-CoA carboxylase (ACC1) and lipoprotein lipase (LPL), coding for key enzymes implicated in lipid metabolism, were significantly increased in the maternal HF diet [13]. Prenatal L. reuteri treatment significantly improved the altered expressions of ACC1 and LPL (Figure 5).
The mechanisms by which HF diet intake during pregnancy leads to obesity in the offspring include remodeling of the placenta and altered maternal gut microbiota composition. These mechanisms result in perturbed lipid metabolism in the fetal liver [13]. Placental remodeling and maternal gut microbiome shaping after L. reuteri treatment during maternal pregnancy were investigated. Placental histology showed that L. reuteri intake during pregnancy reformed placental remodeling related to the maternal HF diet. L. reuteri treatment during pregnancy increased the labyrinth zone and decreased the transitional zone compared to those in the HF group (Figure 6A). L. reuteri treatment during pregnancy diminished placental oxidative injury caused by maternal HF diet exposure (Figure 6B). The placenta bridges maternal and fetal circulation. The placenta plays a key role in fetal development, inclusive of nutrient transportation, hormone production, and acting as an immune barrier [33]. In response to changes in maternal nutrition, the placenta undergoes functional and structural changes that affect the supply of nutrients, hormones, and other molecules to the fetus. Maintaining a balance between caloric intake and energy storage is important for ensuring good health. Cells rely on nutrient sensing to recognize and respond to fuels. Thus, nutrient sensing plays a key role in regulating tissue growth. Gene expressions of several important nutrient-sensing molecules in placental tissue were investigated. L. reuteri treatment during pregnancy broadly increased the mRNA expression of nutrient-sensing molecules including Sirtuin 1 (SIRT-1), peroxisome proliferator-activated receptor α (PPAR-γ), mammalian target of rapamycin (mTOR), and peroxisome proliferator-activated receptor gamma coactivator 1α (PGC1α ) (Figure 7).
Dietary manipulation of the host can lead to significant changes in the gut microbial composition and subsequent alteration of microbial metabolites. Therefore, we investigated the ability of L. reuteri treatment during pregnancy to alter the maternal gut microbiome or metabolites related to a HF diet. The HF diet for dams led to lower α-diversity and higher β-diversity of the gut microbiome in comparison to the CC diet (Figure 8A,B). This result suggested that taxa abundance was influenced by the HF diet. L. reuteri treatment did not alter the abundance of lower taxa related to the HF diet. Neither HF diet exposure nor L. reuteri treatment changed the F/B ratio (Figure 8C). In a previous study, we showed that an HF diet caused gut dysbiosis in dams [13]. In this study, we aimed to determine whether L. reuteri treatment during pregnancy can ameliorate dysbiosis associated with an HF diet. Among the taxa with the greatest difference between the CC and HF groups, the top ten most abundant and least abundant taxa in the CC group compared to the HF group were presented in the heatmap (Figure 8D). The corresponding taxa of the L group and H + L groups were also presented. In agreement with our previous findings, there was a higher presence of the Akkermansia genus in the gut microbiota of HF dams and less Lachospiraceae genus compared to CC dams [13]. In this study, HF dams also demonstrated more Tannerellaceae and Parabacteroides and lower Turicibacter genera than the CC group. The gut microbiota of the H + L group was similar to that of the HF group (Figure 8D). Since L. reuteri treatment could improve liver steatosis in dams and offspring related to the maternal HF diet, further work was carried out to identify the characteristic bacteria between the maternal HF diet with and without L. reuteri treatment. The tax abundance of HF dams with and without L. reuteri treatment was investigated using linear discriminant analysis (LDA) and the effect size (LEfSe) analysis (Figure 8D,E). The length of the histogram (i.e., LDA score), which was based on the size of the different species, was set to be significant, with a difference of more than 3 log10. At the genus level, the gut microbiomes of the HF and H + L groups showed similar performances. LEfSe also showed an increase only in L. reuteri in the H + L group compared with that in the HF group. Considering the significant effect of probiotic treatment on metabolic modulation in offspring, probiotic treatment is likely to achieve therapeutic effects through relevant metabolites. SCFAs are important metabolites fermented from dietary fiber by the gut microbiota. G-protein-coupled receptors (GPR)41 and GPR43 are important SCFA receptors that respond to acetate, butyrate, and propionate. In the placental analysis, we found that probiotic treatment improved the corresponding mRNA expression of GPR43 relative to maternal HF diet exposure (Figure 9B), whereas the mRNA expression of GPR41 and GPR120 was unaffected (Figure 9A,C). Our results also revealed that the plasma acetate and butyrate levels were similar across the four groups (Figure 9D,F). Probiotic treatment improved low plasma propionate outcomes following exposure to a maternal HF diet (Figure 9E).
In this study, we tested the therapeutic effects of L. reuteri (a probiotic) and butyrate (a postbiotic) on metabolic problems during pregnancy in high-fat-fed obese dams. We found that L. reuteri intervention during pregnancy ameliorated obesity and altered the metabolic profiles of obese dams. Both L. reuteri and butyrate treatments improved liver steatosis related to the HF diet in pregnant dams; however, L. reuteri was more effective than butyrate. It was found that L. reuteri intake rather than butyrate intake was an effective preventative strategy for maternal HF diet-induced fetal liver steatosis. Maternal L. reuteri treatment was found to have the ability to reform placental remodeling and decrease placental oxidative injury induced by maternal HF diet. Maternal L. reuteri therapy also modifies the SFCAs signaling cascade, transcription of nutrient-sensing molecules in the placenta, and lipid metabolism in the fetal liver (Supplementary Figure S3. Graphic abstract). The diagnostic criteria for NAFLD are equal to or more than 5% liver fat accumulation and exclusion of other secondary hepatic fat accumulation causes, such as autoimmune liver disease, viral hepatitis, and significant alcohol intake [34]. NAFLD can be subdivided into steatosis (increased liver fat without inflammation) and steatohepatitis (increased liver fat with inflammation and hepatocellular injury). Hepatic steatosis and steatohepatitis may mirror different disease entities [35]. Under certain conditions, inflammation may precede steatosis in steatohepatitis. Whilst under other conditions, some patients with simple steatosis may develop steatohepatitis. Tilg et al. proposed the “multiple parallel hits” hypothesis to explain these diverse courses [35]. They include endoplasmic reticulum stress, endotoxins, energy metabolism changes by gut microbiota, aryl hydrocarbon receptor activation by trans-fatty acids and fructose, and an imbalance of adipocytokines. In our study, HF diet exposure in pregnant dams led to hepatic steatosis in dams and fetuses, rather than steatohepatitis. This may be due to the difference in hits. In addition, the development of steatohepatitis would likely have taken a longer time and was therefore not visible in the fetal period. The gut microbiome plays an important role in the development of NAFLD. Studies have shown (in comparison to control animals) that received fecal transplants from insulin-sensitive mice, germ-free mice transplanted with gut microbiota from insulin-resistant mice that received a HF diet showed similar body weight and food consumption, but increased liver fat deposition [36]. The expression of key transcription factors that regulate lipogenesis was also increased in the liver. Gut bacteria also influence the host metabolism through their metabolites. One example of this is SCFAs. SCFAs are one of the end products of polysaccharides after fermentation by intestinal bacteria and they play a role in nutrient sensing and energy balance [37]. In a previous study, we found that pregnant dams fed a HF diet showed gut dysbiosis and a decrease in plasma propionate levels [13]. In this study, the intestinal flora of L. reuteri-treated pregnant dams exposed to the HF diet did not differ from those fed the HF diet alone. The gut microbiome of the L. reuteri-treated group showed a high abundance of L. reuteri; this can be explained by the probiotic treatment. However, L. reuteri treatment increased the serum butyrate levels of pregnant dams and the gene expression of GPR43 in placental tissue. A previous report showed that New Zealand black × New Zealand white (NZB/W) F1 mice supplemented with L. reuteri GMNL-89 had improved hepatic apoptosis and inflammatory markers [38]. In this study, L. reuteri treatment of pregnant dams showed dual effects of treating NAFLD in pregnant women with an HF diet and re-programming the fatty liver change in the fetus. These therapeutic effects may be partly mediated by the modulation of postbiotics. To carry out their health effects, probiotics must overcome inappropriate lethal conditions during processing, and maintain their survival under inappropriate storage and digestive system conditions. Postbiotics may successfully overcome these adverse effects and become ideal substitutes for probiotics. In this study, we assigned butyrate as a postbiotic for the treatment of pregnant dams fed a HF diet. Previous studies showed that butyrate treatment relieves HF diet-induced hepatic steatosis in mice by upregulating miR-150 expression, downregulating C-X-C chemokine receptor type 4, enhancing insulin-induced gene activity, and suppressing lipogenic genes [25,39]. Modulation of intestinal tight junctions is also involved in the regulation of lipid metabolism in db/db mice [40]. In our study, we observed the beneficial effect of butyrate on hepatic steatosis in dams, even though L. reuteri treatment was more effective. Butyrate treatment did not have a therapeutic effect on fetal steatosis in mothers with obesity. Postbiotics are bioactive metabolites produced by gut bacteria during fermentation or may be structural fragments of bacteria. In addition to SCFAs, teichoic acids, polysaccharides, vitamins, peptides, plasmalogens, and organic acids are also important postbiotics with different functional properties [41]. Butyrate treatment alone cannot completely replace probiotics, and it is important to recognize the characteristics and functions of each postbiotic. Pregnant obese women may pass on their metabolic phenotypes to their offspring, leading to a cycle of obesity across generations. Offspring of obese women are at an increased risk of developing NAFLD. A nationwide study in Sweden that included 165 young adults aged <25 years with liver biopsy-proven fatty liver disease and 717 age- and sex-matched non-fatty liver controls found that the offspring of obese mothers had a higher risk of developing fatty liver disease [42]. The increase in liver fat in the offspring of these obese mothers may be partially due to the delivery of more nutrients to the fetus. Fetal subcutaneous fat storage capacity is limited, so, the fetus is forced to use the liver to store fat. This results in a fatty liver [43]. The persistence of hepatic fat storage increase in the offspring of obese mothers may be due to changes in hepatic lipogenesis, fatty acid oxidation, or lipoprotein output process [13,43,44]. It has been shown that the expression of nutrient-sensing mitochondrial proteins SIRT1 and SIRT3 is reduced in the livers of offspring born to dams exposed to a HF diet and obese dams, respectively [45,46]. SIRT1 can modulate the transcription of peroxisome proliferator-activated receptor α (PPARα) and PPARγ, which in turn control fatty acid oxidation and lipogenesis, respectively. This indicates that fatty liver in the offspring of obese mothers is associated with nutrient-sensing defects and mitochondrial dysfunction. From our study, maternal L. reuteri treatment (rather than maternal HF diet alone) enhanced the mRNA expressions of several nutrient sensing molecules. The modulatory effects of nutrient-sensing molecules could be another therapeutic mechanism for prenatal L. reuteri treatment. Our study has several limitations. Ideally, the expression of target genes should be evaluated at both the transcriptional and translational levels. Limited sample sizes inhibited the ability to analyze both transcriptional and translational levels in this study. As such, only mRNA expression, rather than protein abundance, was provided in this study. Further studies are needed to validate our observations at both the protein and functional levels. In our study, the α-diversity and β-diversity (Figure 8A,B) of the gut microbiome in the HF and H + L groups showed large deviations. In addition, the gut microbiomes of the HF and H + L groups were found to be significantly discrete (Figure 8C). Since food and sterile tap water were provided ad libitum in this study, it is possible that the gut microbiome was affected by the inconsistent eating status of each mouse. Therefore, the gut microbiome exhibits intragroup variation. In this study, we found that pregnant dams exposed to a HF diet with L. reuteri had greater reductions in hepatic fat compared to dams fed the HF diet alone. It is possible that this effect was partially mediated by weight loss since the dams with L. reuteri intake lost body weight compared to those consuming only the HF diet. Because the caloric intake per body weight of dams fed L. reuteri was lower than that of dams not fed L. reuteri, L. reuteri consumption appears to reduce the appetite of dams fed an HF diet. Pharmacological treatments for NASH include aldafermin, pioglitazone, vitamin E, obeticholic acid, resveratrol, liraglutide, and lanifibranor [47,48]. L. reuteri has the potential to be used as adjuvant therapy for lifestyle interventions in the treatment of NAFLD in pregnant mothers and as a de-programming strategy for NAFLD in the next generation. The therapeutic effects of L. reuteri should be validated in future clinical trials. | true | true | true |
PMC9573645 | 36235052 | Jaeseok Lee,Youngjun Lee,Young Mee Jung,Ju Hyun Park,Hyuk Sang Yoo,Jongmin Park | Discovery of E3 Ligase Ligands for Target Protein Degradation | 02-10-2022 | target protein degradation,PROTAC,E3 ligase ligand | Target protein degradation has emerged as a promising strategy for the discovery of novel therapeutics during the last decade. Proteolysis-targeting chimera (PROTAC) harnesses a cellular ubiquitin-dependent proteolysis system for the efficient degradation of a protein of interest. PROTAC consists of a target protein ligand and an E3 ligase ligand so that it enables the target protein degradation owing to the induced proximity with ubiquitin ligases. Although a great number of PROTACs has been developed so far using previously reported ligands of proteins for their degradation, E3 ligase ligands have been mostly limited to either CRBN or VHL ligands. Those PROTACs showed their limitation due to the cell type specific expression of E3 ligases and recently reported resistance toward PROTACs with CRBN ligands or VHL ligands. To overcome these hurdles, the discovery of various E3 ligase ligands has been spotlighted to improve the current PROTAC technology. This review focuses on currently reported E3 ligase ligands and their application in the development of PROTACs. | Discovery of E3 Ligase Ligands for Target Protein Degradation
Target protein degradation has emerged as a promising strategy for the discovery of novel therapeutics during the last decade. Proteolysis-targeting chimera (PROTAC) harnesses a cellular ubiquitin-dependent proteolysis system for the efficient degradation of a protein of interest. PROTAC consists of a target protein ligand and an E3 ligase ligand so that it enables the target protein degradation owing to the induced proximity with ubiquitin ligases. Although a great number of PROTACs has been developed so far using previously reported ligands of proteins for their degradation, E3 ligase ligands have been mostly limited to either CRBN or VHL ligands. Those PROTACs showed their limitation due to the cell type specific expression of E3 ligases and recently reported resistance toward PROTACs with CRBN ligands or VHL ligands. To overcome these hurdles, the discovery of various E3 ligase ligands has been spotlighted to improve the current PROTAC technology. This review focuses on currently reported E3 ligase ligands and their application in the development of PROTACs.
Proteins are the basic machinery of the cellular system and execute genetically programmed behaviors for cellular survival. Around 20,000 proteins have been identified from human cells and their balance in the protein network is extremely important to maintain the healthy status of cells. The dysfunction or breakdown of a single protein could lead to the disease status of cells. Therefore, discovering the small molecules that are modulating the dysregulated protein is the key strategy for the drug discovery process. However, only a limited number of proteins has been reported as druggable proteins and traditional drug discovery has been focused on those druggable proteins. Considering the fact that a numerous number of proteins in cells are essential for cellular homeostasis, targeting undruggable proteins could be a solution for the treatment of incurable diseases. Targeted Protein Degradation (TPD) is an emerging therapeutic strategy which is considered a solution to overcome the limitations of conventional drug discovery. TPD is a powerful chemical biology approach for inducing the degradation of proteins known as undruggable, and it has had a tremendous impact on the field of recent drug discovery. TPD manipulates the E3 ligase to selectively degrade the protein of interest (POI) via the Ubiquitin Proteasome System (UPS), an intracellular proteolysis mechanism. A targeted protein degrader consists of a ligand that binds to E3 ligases, a ligand that targets the POI, and a chemical linker between the ligands. These degraders are also called proteolysis-targeting chimeras (PROTACs), and the induced proximity between the target protein and the E3 ligase promotes the ubiquitination and proteasomal degradation of the target protein. Along with PROTAC, a protein degradation strategy using the protein tag has been developed to understand the inherent function of proteins. The additional aminoacidic signal sequences such as dTAG, AiD, and SMASh Tag induce proximity-based POI ubiquitination for its degradation via the ubiquitin–proteasome system. The development of tag-based strategies is well documented in the following references [1,2,3]. Even though there are over 600 types of E3 ligases in human cells, only a very limited number of E3 ligases have been used for PROTAC technology (CRBN, VHL, IAP, and MDM2). The majority of PROTACs developed so far have been restricted to CRBN or VHL, which are ubiquitously expressed in the human body. However, the recent emergence of drug resistance for CRBN- or VHL-based PROTACs strongly suggested that the discovery of various E3 ligase ligands is highly demanding to fully exploit the PROTAC strategy [4,5,6]. In addition, considering that a number of E3 ligases have been shown to overexpress in specific types of cells or tissues—for example, brain (FBXL16, KCTD8), pancreas (ASB9), skeletal muscle (KLHL40, KLHL41), testis (DCAF4L1), fallopian tube (DCAF8L1)—harnessing new E3 ligases may offer better opportunities for PROTACs having a higher selectivity and specificity for the efficient disease treatment. In this context, an overview of the PROTAC design strategies is highly necessary to develop efficient protein degraders. This review focuses on the currently reported E3 ligase ligands and their application in the development of PROTACs.
In the 1950s, thalidomide (1) was first developed by Grünenthal as a sedative for morning sickness in pregnant women. However, thalidomide was withdrawn from the market due to its severely teratogenic effects in the early 1960s [7]. During the following decades, thalidomide was further studied extensively, redeveloped as a promising immunomodulatory imide drug (IMiD), and approved for the treatment of erythema nodosum leprosum (ENL) and multiple myeloma. The thalidomide derivatives, such as pomalidomide and lenalidomide, have shown an excellent efficacy in the treatment of multiple myeloma with their immunomodulatory activities (Figure 1) [8,9,10]. However, the mechanism of action of thalidomide had not been elucidated until Ito et al. revealed that the target protein of thalidomide is cereblon, a subunit of the E3 ubiquitin ligase CUL4-RBX1-DDB1-CRBN(CRL4CRBN), in 2010 [11]. They demonstrated a thalidomide–CRBN complex-induced teratogenic effect in in vivo models. In addition to thalidomide, the target protein of pomalidomide and lenalidomide was identified as CRBN by Zhe et al. in 2011 [12]. The antimultiple myeloma activity of pomalidomide (2) and lenalidomide (3) was CRBN expression dependent. In 2014, Fischer et al. presented the cocrystal structure of the DDB1–CRBN complex bound to thalidomide, lenalidomide, and pomalidomide. This research showed that CRBN is the substrate receptor of CRL4CRBN and binds to thalidomide derivatives in an enantioselective manner. They further revealed that the thalidomide–CRBN complex recruits IKZF1 or IKZF3 and induces their degradations [13]. Thalidomide derivatives are composed of the phthalimide and glutarimide group (Figure 2A). The cocrystal structure (PDB code: 4CI1) of the thalidomide and CRBN complex showed that the glutarimide group of thalidomide derivatives plays an important role in CRBN binding via two major interactions: (1) the H-bond between carbonyl and amide groups of the glutarimide group and His380 and Trp382 of CRBN; and (2) van der Waals interactions between the glutarimide group and the hydrophobic pocket of CRBN composed of Phe404, Trp388, and Trp402 (Figure 2B). A carbonyl group of the phthalimide group also contributes a H2O-mediated hydrogen bonding with the His359 of CRBN. This cocrystal structure revealed that the solvent-exposed site of thalidomide is the benzene ring of the phthalimide group, which can be further conjugated for the PROTAC design without a loss of the binding affinity [14]. Based on this structural information, the development of thalidomide derivatives has been able to be accelerated. In 2015, the Crews and Bradner group reported that bromodomain-containing protein 4 (BRD4) targets PROTACs. ARV-825 (DC50 < 1 nM) [15] and dBET1 (EC50 = 430 nM) were synthesized by conjugating a BRD4 ligand, JQ1, with a thalidomide derivative for CRBN engagement (Figure 3). Both PROTACs effectively triggered the degradation of BRD4 in the cells and inhibited the cell proliferation. dBET1 showed an excellent antiacute myeloid leukemia (AML) efficacy in vitro and in vivo [16]. After dBET1 was reported as the first PROTAC in vivo, various thalidomide derivatives have been explored to improve the efficiency of the PROTAC technique. In 2018, the Crew group reported 22 thalidomide analogs with a rapid one-pot synthesis without purification. They measured the binding affinities of newly synthesized thalidomide analogs to CRBN with a surface plasmon resonance analysis. In addition, the ability to induce the degradation of Aiolos and CK1α was investigated. Among them, three thalidomide analogs (4, 5, 6) showed improved pharmacological properties and good CRBN binding affinities (KD for 4, 5, and 6 were 55, 549, and 111 nM, respectively). They found that the chemical modification of the phthalimide part in thalidomide did not induce a significant deterioration in the CRBN binding affinities [17]. In 2019, the Yang group synthesized PROTAC, DGY-08-097, which induces hepatitis C virus (HCV) NS3/4A protease degradation by conjugating 6 with Telaprevir, an FDA-approved drug for the treatment of HCV [18] (IC50 = 247 nM, DC50 = 50 nM). DGY-08-097 showed antiHCV activity in the cellular infection model. C4 Therapeutics also filed a patent for the synthesis of CRBN-targeting moieties (Degrons) that can be conjugated to target protein ligands. They reported various piperidine-2,6-dione derivatives by conjugating O- and N-linked heterocycles. A fluorescence polarization (FP) assay demonstrated that 7 was one of the most promising CRBN ligands. The degrons are capable of functioning as molecular glues and downregulating the levels of the Aiolos or Ikaros protein, which can lead to the treatment of leukemia, acute myeloid leukemia, chronic lymphoblastic leukemia, and multiple myeloma just like other IMiDs. They conjugated the developed E3 ligase ligand to JQ1 and named them Degronimers. The representative Degronimers 1 and 2 showed an excellent binding affinity to CRBN [19] (KD < 10 μM). They also reported various SMARCA2 degraders using the degrons. The N-linked degrons, 8, was developed as PROTAC (compound 156) for SMARCA2 degradation with a nanomolar efficacy through a HiBiT degradation assay (DC50 = 3 nM) [20]. The Hwang group designed aminobenzotriazino glutarimides as novel CRBN ligands and discovered TD-106 (9). The TD-106 (9) exhibited a better degradation efficiency of IKZF1/3 than that of pomalidomide in NCI-H929 cells. After the confirmation of TD-106 (9) as a direct CRBN binder through a thermal stability shift, an in vivo xenograft model study demonstrated that intraperitoneally injected TD-106 (9) showed antitumor activity after 14 days of administration [21]. They further synthesized a BRD4-targeting PROTAC, TD-428 (DC50 = 0.32 nM), by conjugating JQ1 and TD-106 (9). The successful BRD4 degradation in 22Rv1 cells induced by TD-428 was confirmed. They also reported an androgen receptor (AR) degrader, TD-802 (DC50 = 12.5 nM), for the treatment of metastatic castration-resistant prostate cancer using the TD-106 (9) ligand [22]. In 2019, Arvinas unveiled that the ARV-110 (DC50 ~ 1 nM) and ARV-471 (DC50 ~ 1 nM) are targeting AR and estrogen receptor (ER), respectively. Both degraders have been spotlighted since they are currently in phase two of clinical trials [23,24]. ARV-110 was synthesized via the conjugation of a thalidomide derivative (4) and AR ligand. ARV-471 was developed via the conjugation of a thalidomide derivative (5) and ER ligand. Both shared the same linker structure for the conjugation of CRBN and target protein ligands. The most interesting feature of these PROTACs is that both can be administered orally, which was not easily achievable in other PROTACs. Kymera Therapeutics reported various CRBN ligands having two rings conjugated with diverse linkers in 2019. They measured their affinities toward CRBN through a time-resolved fluorescence resonance energy transfer (TR-FRET) assay [25]. They synthesized a PROTAC, I-265 (DC50 < 0.1 μM), by conjugating 10 and an interleukin-1 receptor activated kinase (IRAK) inhibitor. They showed that I-265 degraded the IRAK4 protein in human peripheral blood mononuclear cells (PBMCs) [26]. One of their ligands, SB572027 (11), was used for the synthesis of BTK-targeting PROTACs by the Chinese biotechnology company Beigene. Beigene synthesized a series of PROTACs that conjugated a BTK inhibitor with various thalidomide derivatives capable of recruiting the target protein to the E3 ubiquitin ligase. They performed an ELISA assay for BTK detection to evaluate the activity of a series of the PROTACs. Among them, compound 155 using SB57027 (11) as a CRBN ligand showed nanomolar degradation activity (DC50 = 7.2 nM) [27]. In 2021, the Rankovic group reported highly stable CRBN binders and their application for PROTACs. They replaced the hydrolysis-labile phthalimide moieties of thalidomide with phenyl groups to synthesize phenyl glutarimide (PG, 12). PG (12, IC50 = 2.191 μM) showed outstanding stability (t1/2 > 24 h) compared with thalidomide (t1/2 = 3.3 h) in cell media without affecting the binding affinity with CRBN (IC50 for thalidomide = 1.282 μM, lenamide = 0.699 μM, and pomalidomide = 0.4 μM). PG (12) also showed superior metabolic stability in mouse and human liver microsomes. A PG-based PROTAC (SJ995973, DC50 = 0.87 nM) showed potent antiproliferative efficacy in both MV4-11 (IC50 = 3 pM) and HD-MB03 cells (IC50 = 1.83 nM) [28]. In a follow-up study, the Rankovic group used PG (12) to synthesize Janus kinase (JAK)-targeting PROTACs. Through the structure activity relationship (SAR) study, they found a PG (12)-based PROTAC, SJ10542, with highly selective JAK degradation and reduced GSPT1 degradation compared with thalidomide-based PROTACs (JAK2 DC50 = 14 nM, JAK3 DC50 = 11 nM) [29]. In 2021, Wang et al. filed a patent about the synthesis of various CRBN ligands for PROTAC design [30]. Among their compounds, 13 and 14 were conjugated to spirocyclic AR ligands to synthesize AR-degrading PROTACs for the treatment of prostate cancer. They measured AR degradation in VCaP cells and compound 311 showed the efficient degradation of AR with a nanomolar concentration (DC50 < 10 nM). Moreover, they demonstrated the superior oral bioavailability of compound 311 [31]. In 2021, Novartis developed a BRD9 degrader for cancer treatment. BRD9 is a subunit of the SWI/SNF complex, which has been reported as a drug target for the treatment of synovial sarcoma and acute myeloid leukemia. A series of BRD9 PROTACs were synthesized via the conjugation of BRD9-targeting ligands and CRBN-targeting E3 ligase ligands with various types of short and rigid piperidine linkers. Among the synthesized PROTACs, compound B6 and compound E32, based on the CRBN binders 15 and 16, respectively, showed BRD9 DC50 at the nanomolar concentration [32] (DC50 = 1 nM for both compound B6 and E32).
The von Hippel–Lindau (VHL) protein is the substrate receptor protein of the Cullin 2 E3 ligase. Hypoxia-inducible factor 1α (HIF-1α) is one of the substrate proteins of VHL. Considering the HIF-1α-mediated upregulation of proangiogenic factors, an inhibitor of VHL would increase erythropoietin and can be used for the treatment of chronic anemia and cancer chemotherapy. In 2012, the Ciulli and Crews group reported a series of small molecule inhibitors targeting VHL for the first time. They rationally designed inhibitors of the VHL/HIF-1α interaction in silico. After the discovery of their initial hit, they synthesized a focused library of hydroxy proline derivatives with solid phase synthesis. They found VHL ligand 17 with single-digit micromolar activity through an FP assay (Figure 4). The cocrystal structure of the ligand bound to VHL confirmed that the ligand mimics the binding mode of the HIF-1α to VHL [33]. They demonstrated that the major binding affinity of VHL ligands originates from the H-bond between the hydroxyl group of pyrrolidine and His115 and Ser111, as well as the interaction between the amide NH group and a carbonyl group of His110 (Figure 5A). In addition, the phenyl group, adjacent to the amide group, contributes to the high binding affinity of the VHL ligand via a π-π interaction with Tyr98. In a follow-up study, the Ciulli group designed and optimized the initial VHL ligands based on X-ray crystal structures and reported new VHL ligands, VH032 (18, KD = 185 nM) and 19 (KD = 291 nM), with nanomolar binding affinities, in 2014 [34]. Using VH032 (18) and 19, the Ciulli group reported bromodomain and extra-terminal domain (BET) protein-targeting PROTACs MZ1 (KD = 149 nM) and MZ3 (KD = 311 nM) (Figure 6) [35]. They observed the MZ1-mediated degradation of BRD2, BRD3, and BRD4 at the single digit micromolar concentration (BRD4 DC50 < 100 nM). In 2015, the Crews group proposed HaloPROTAC using their VHL ligands. Based on their previously reported crystallographic evidence, they identified the possible linker positions by confirming the solvent-exposed sites of the VHL ligands. Not only the linker position but also a proper linker length were studied for VHL-mediated protein degradation. Finally, they successfully designed HaloPROTAC3 (DC50 = 19 nM) by conjugating a chloroalkane linker to the VHL ligand (20, IC50 = 0.34 μM) for the degradation of the HaloTag fusion protein [36]. The Crews group revealed that the linker position of the VHL ligand largely affects the substrate specificity of PROTAC. They synthesized PROTACs with the linkers on the left-handed amide side and right-handed phenyl side of the VHL ligand, respectively (Figure 5B). The PROTAC (SJF-6683) conjugated a p38 MAPK ligand, foretinib, to the right-hand phenyl side of the VHL ligand (20) selectively and strongly degraded p38δ, a specific p38 isoform (p38δ DC50 = 46.17 nM) [37]. In 2018, Ciulli reported the structure-guided rational optimization of VH032 (18). Increasing the lipophilicity of the VHL ligands led to a higher cell permeability and higher binding affinity to the VHL protein. They discovered VH298 (21, KD = 52 nM) and VH101 (22, KD = 16 nM), which showed the effective protein–protein inhibition between the VHL and HIF-1α protein. [38]. Their SAR study provided novel VHL ligands, which can be used for VHL-based PROTACs. Note that this discovery expanded the landscape of PROTAC research to other E3 ligases, which was only restricted to CRBN. VH101 (22) was used in the SAR study to develop BRD7 and BRD9 degraders by the Ciulli group. A PROTAC (VZ185) was developed by coupling the BRD7/9 inhibitor, BI-7273, and VH101 (22). VZ185 efficiently degraded BRD7 and BRD9 simultaneously (BRD7 DC50 = 4.5 nM, BRD9 DC50 = 1.8 nM) [39]. The Hodgkinson group reported histone deacetylase (HDAC)-targeting PROTACs. They synthesized HDAC-degrading PROTACs with various linkers and VHL ligands and monitored their degradation activities. JPS036, a PROTAC composed of VH101 (22), was developed as a selective degrader of HDAC3 with a submicromolar activity (DC50 = 0.44 μM) [40]. In 2018, the Ciulli group reported VHL ligands with a different stereochemistry of fluoro-hydroxyproline (F-Hyp). They synthesized four diastereoisomers of 3-fluoro-4-hydroxyproline containing VHL ligands and found that VHL can stereoselectively recognize the (3R,4S) epimer of F-Hyp (23). A JQ1-based PROTAC (compound 15a) using the (3R,4S) epimer of F-Hyp selectively degraded BRD4 at nanomolar concentrations, despite its weak affinity for VHL (KD = 3.08 μM, BRD4 DC50 = 1~3 nM, BRD2 and BRD3 DC50 = 10 nM). This discovery was an important advance in expanding the chemical space of TPD toward low affinity molecules [41]. In 2019, the Wang group discovered new VHL ligands via the introduction of an (S)-methyl group on VH101 based on previous work [42,43]. Through a SAR study, they found that appending an amide group to the (S)-methyl group increased the potency of the VHL ligands. The FP-based binding assay showed that VHL-e (24) binds to VHL with a high affinity (IC50 = 190 nM). In addition, they confirmed that the introduced stereochemistry was crucial for their binding affinity to VHL. With VHL-e (24), an effective AR degrader, ARD-69, was discovered after the optimization of the linker length and linking site on the ligands. ARD-69 showed effective AR degradation activity in LNCaP (DC50 = 0.86 nM), VCap (DC50 = 0.76 nM), and 22Rv1 prostate cancer cell lines and in a VCaP xenograft mouse model [43]. The Wang group further optimized the AR-targeting PROTAC with a shorter linker length and a low-affinity VHL ligand. ARD-266 using a weak binding VHL ligand, VHL-g (25), showed a much higher AR degradation activity than other PROTACs with higher-affinity VHL ligands (DC50 = 0.5 nM in LNCaP cell line, DC50 = 1.0 nM in VCaP cell line). Along with compound 15a, this study demonstrated that a low-affinity E3 ligase ligand-based PROTAC could induce the successful formation of a ternary complex with the POI for efficient degradation [44]. Subsequently, the Wang group extensively studied the SAR of an ER degrader based on a VHL ligand (26) and FDA-approved ER modulator, Raloxifene. As a result, a very potent ER degrader (ERD-308) was developed with a subnanomolar activity (DC50 = 0.17 nM). ERD-308 degraded the ER and inhibited cell growth more than those of the FDA-approved selective ER degrader molecule Fulvestrant [45]. In this report, changing the VHL ligand to the CRBN ligand completely abolished the ER degradation activity of ERD-308.
Inhibitors of apoptosis proteins (IAPs) are the regulators of cell death and they control apoptotic events triggered by diverse stimuli. In 2007, the Vucic group developed a cellular inhibitor of the apoptosis 1 and 2 (c-IAP1 and c-IAP2) antagonist (MV1, 27) that binds to the baculovirus IAP repeat (BIR) domains of IAP proteins, leading to the autoubiquitination and proteasomal degradation of c-IAPs (KD = 5.8 nM) (Figure 7). The degradation of c-IAPs by MV1 induced TNF signaling-pathway-dependent cell death [46]. Sekine et al. reported a different cIAP1 ligand, bestatin-methyl ester (ME-BS, 28), which binds to the BIR3 domain of cIAP1 and induces autoubiquitination followed by the proteasomal degradation of cIAP1 [47]. Itoh et al. developed bifunctional small molecules using the two ligands described above. One of the molecules (Compound 4b) was designed by the conjugation of ME-BS (28) and all-trans retinoic acid (ATRA) with a polyethylene glycol (PEG) linker (Figure 8). The other was developed via the conjugation of MV1 (27) and ATRA (Compound 6) [48,49]. These compounds were found to induce the degradation of both cIAP1 and cellular retinoic acid binding protein-II (CRABPII). They named this degrader the specific and nongenetic IAP-dependent protein eraser (SNIPER). It was later utilized on other targets, such as ER [50], BRD4 [51], and BCR-ABL [52]. In 2012, Genentech discovered a potent antagonist of cIAP1/2, ML-IAP, and XIAP. The SAR study using the crystal structure led to the development of a broad spectrum IAP inhibitor, GDC-0152 (29) (Ki values for XIAP-BIR3 = 28 nM, MLXBIR3SG = 14 nM, cIAP1-BIR3 = 17 nM, and cIAP2-BIR3 = 43 nM) [53]. In cocrystal structures of GDC-0152 with ML-IAP or cIAP1, the critical interaction was the H-bond between the Asp (Asp138 for ML-IAL, Asp320 for cIAP1) and the amide group of the ligand (Figure 9). It allows for the proper positioning of the α-methyl group of the ligand to reside in the P1 cavity. In addition, another hydrophobic interaction was observed between the phenyl group of the ligand and P4 hydrophobic pocket in both crystal structures. Since most of the ligand structure is exposed to the solvent, the target protein ligands can be conjugated to diverse positions of the IAP ligands. The Hennessy group discovered a series of aminopiperidine-based inhibitors of the IAP by mimicking the IAP binding residues of the second mitochondrial activator of caspases (Smac). They found that a bicyclic piperidine, 30 (KD for XIAP-BIR3 = 0.9 μM), fixed in a boat form was a potent inhibitor of cIAP1 and effectively induced cIAP1 degradation in MDA-MB231 cells (EC50 = 5 nM) [54]. In 2013, the Cosford group reported the synthesis of a potent IAP antagonist via a highly efficient application of the Ugi four-component reaction. Their optimized IAP antagonist (31) showed the best binding affinity to IAPs, especially against ML-IAP (Ki = 2 nM). 31 showed a powerful anticancer activity in breast, ovarian, and prostate cell lines and had no general toxicity to noncancerous human foreskin fibroblast (HFF) cells. They performed molecular modeling to reveal key interacting residues of IAP proteins with 31 [55]. The Zheng group used 31 to synthesize IAP-recruiting BCL-XL PROTACs, compound 8A (IC50 = 62 nM, DC50 < 500 nM). Compound 8A showed efficient BCL-XL degradation in the T-cell lymphoma cell line, while it had reduced the human platelet toxicity [56]. In 2014, Bristol-Myers Squibb reported that 32 embedded bivalent heterodimeric IAP antagonists showed a high affinity for the BIR2 domain and an excellent IAP inhibitory activity (IC50 up to 3.6 nM) [57]. Pfizer synthesized BC5P, a PROTAC that degrades BTK using 32 (DC50 = 182 nM). They confirmed that BC5P bound to only the BIR3 domain of IAP1 but not BIR1 or BIR2 by using biolayer interferometry (BLI). They utilized molecular modeling, solution NMR, and X-ray crystallography to elucidate the structural insights of the IAP-BC5P-BTK ternary complex [58]. In 2017, the Naito group developed an ER-targeting SNIPER, SNIPER(ER)-87, via the conjugation of an ER ligand, 4-hydroxytamoxifen, and an IAP ligand, LCL-161 (33) [50]. SNIPER(ER)-87 effectively reduced ERα protein levels at nanomolar concentrations in vitro (DC50 = 3 nM, IC50 = 15.6 nM in MCF-7 cell line, IC50 = 9.6 nM in T47D cell line). SNIPER(ER)-87 showed good metabolic stability in the serum. The intraperitoneal administration of SNIPER(ER)-87 reduced the growth of ER-positive human breast tumors in vivo. They also conjugated LCL-161 derivatives to JQ1, a PDE4 inhibitor, and dasatinib for SNIPERs targeting BRD4, PDE4, and BCR-ABL proteins to demonstrate the usefulness of LCL-161 derivatives for the development of various targeted protein degraders. In 2018, the Naito group optimized their previously reported SNIPER(ER)-87 by incorporating various LCL-161-derivative IAP ligands (29, 30, 35, 36). With the improved IAP binding affinities of the LCL-161 derivatives, the optimized SNIPER(ER)s exhibited better binding affinities toward cIAP1, cIAP2, and XIAP. However, the E3 ligase binding affinities of SNIPER(ER)s were not exactly correlated to their target protein degradation efficiencies. The GDC-0152 (29)-based SNIPER(ER)-131 did not efficiently degrade ERα, despite its higher IAPs affinity than that of SNIPER(ER)-87 (SNIPER(ER)-131: ERα DC50 > 33.8 nM, IC50 = 80 nM, SNIPER(ER)-87: ERα DC50 < 3 nM, IC50 = 110 nM). 30-based SNIPER(ER)-118 had a low ERα degradation efficiency compared to the original compound SNIPER(ER)-87 (SNIPER(ER)-118: ERα DC50 > 100 nM, IC50 = 230 nM). SNIPER(ER)-110 and SNIPER(ER)-126 were the most potent ERα degraders with the lowest DC50 (< 3 nM) among the SNIPER(ER)s (SNIPER(ER)-110: ERα DC50 < 3 nM, IC50 = 120 nM, SNIPER(ER)-126: ERα DC50 < 3 nM, IC50 = 83 nM). SNIPER(ER)-110 showed the best ERα degradation efficiency and excellent antitumor activity in the in vivo tumor xenograft model [59]. Astex Pharmaceuticals successfully discovered a nonpeptidomimetic cIAP1 and XIAP inhibitor, AT-IAP (34), through a fragment-based drug discovery using structure information from X-ray crystallography, computational studies, and NMR solution structure analysis. AT-IAP (34) showed a strong dual antagonistic efficacy toward XIAP and cIAP1 (XIAP EC50 = 5.1 nM, cIAP1 EC50 = 0.32 nM). An oral administration of AT-IAP (34) in a mouse xenograft model effectively inhibited the tumor growth without affecting the body weight of the mouse [60]. In 2020, GlaxoSmithKline reported a palbociclib-based PROTAC with a CRBN ligand, a VHL ligand, and an IAP-binder (37) for the degradation of CDK4 and CDK6. They conjugated three different E3 ligase ligands with palbociclib, an FDA-approved anti-breast-cancer agent, using various linkers. With previously reported CRBN-based CDK4/6-targeting PROTACs [61,62], VHL- and IAP-based PROTACs showed an ability to effectively degrade CDK4/6, (DC50 < 10 nM) which could not be achieved with the previously reported VHL- and IAP-based PROTACs [63]. This work again emphasized the importance of linker structures in the development of PROTACs [64]. Moreover, all the Palbociclib-based PROTACs with three different E3 ligases showed the preferential degradation of CDK6 over CDK4 with a marginal degradation efficiency difference. 37 and AT-IAP (34) were also utilized in the development of a series of receptor-interacting serine/threonine protein kinase 2 (RIPK2) PROTACs including compound 20 and compound 22 by GlaxoSmithKline (compound 20: RIPK2 pDC50 = 9.1, compound 22: RIPK2 pDC50 = 9.8). RIPK2 plays a crucial role in the innate immune system. Therefore, the dysregulation of RIPK2 signaling pathways is highly related to various inflammatory diseases such as inflammatory bowel disease [65], severe pulmonary sarcoidosis [66], and multiple sclerosis [67]. Among the synthesized RIPK2 PROTACs, compound 20 showed the best profile with excellent solubility, a strong RIPK2 degradation ability, and TNFα inhibition [68].
The mouse double minute 2 homolog (MDM2) protein is an E3 ubiquitin ligase that regulates the ubiquitination of p53 and the subsequent proteasomal degradation of p53. In 2004, Roche reported a potent and selective small molecule inhibitor of the MDM2-p53 interaction. They screened a diverse library of synthetic compounds and identified Nutlin-3 (38) as a hit compound (Figure 10) [69]. Notably, two enantiomers of cis-imidazoline Nutlin-3 (38) possessed a highly different binding affinity to MDM2 (IC50 of enantiomer a = 13.6 μM, enantiomer b = 0.09 μM). In 2008, the Crews group first reported MDM2-based PROTAC 14 by conjugating Nutlin-3 (38) and a nonsteroidal AR ligand with a PEG linker (Figure 11) [70]. The Sheng group reported a Nutlin-3 (38)-based homo-PROTAC 11a that induces the self-degradation of MDM2 to inhibit the MDM2-p53 interaction (DC50 = 1.01 μM). Homo-PROTAC 11a induced effective MDM2 dimerization and triggered the proteasomal degradation of MDM2 in A549 non-small cell lung cancer cells. In addition, homo-PROTAC 11a-1, one of the stereoisomer of homo-PROTAC 11a, showed the highest antitumor activity in the A549 xenograft model (IC50 = 1.0 μM) [71]. In 2013, Roche optimized the original Nutlin-3 (38) compound based on the crystal structure of the p53-MDM2 complex and synthesized a new MDM2 inhibitor, RG7112 (39), with dimethyl substitution on the imidazoline ring and replacement of the methoxy group with a tert-butyl group (IC50 = 18 nM) [72]. RG7112 was the first orally available p53-MDM2 inhibitor under clinical trials. In a follow-up study, they discovered a new ligand with an improved affinity based on the crystal structure of MDM2. They replaced the imidazoline structure of RG7112 with a pyrrolidine moiety and introduced stereochemical configurations for the higher affinity. They reported a second-generation clinical MDM2 inhibitor, RG7388 (40), with an excellent efficacy and selectivity through a SAR study [73] (IC50 = 6 nM). In 2020, the Calabretta group reported that RG7112 (39)-based PROTAC YX-2-233 showed a strong degradation of CDK4 and CDK6 in Ph+ ALL cells and suppressed S-phase [74]. The Crews group reported that an RG7388-based PROTAC, A1874, showed a 98% degradation of the BRD4 protein in HCT116 cells with a submicromolar concentration. This was a substantial improvement in target potency compared with their first nutlin-based PROTAC 14 (PROTAC 14: DC50 = 10 μM vs. A1874: DC50 = 32 nM). In addition, A1874 increased p53 stability due to the RG7388 moiety. The dual mode of action, BRD4 degradation and p53 stabilization, by A1874 strongly suppressed cancer cell viability compared with VHL ligand-based PROTACs [75].
Sulfonamide derivatives have drawn attention due to their antibacterial, antifungal, antiviral, and anticancer activities. Recent studies reported that the sulfonamide derivatives indisulam (41), E7820 (42), and chloroquinoxaline sulfonamide (CQS, 43) function as a molecular glue that induces the protein–protein interaction between the E3 ligase and the target protein (Figure 12) [76,77]. In 2017, the Nijhawan group revealed that sulfonamide induces the proteasomal degradation of RNA-binding motif protein 39 (RBM39) by interacting with the DCAF15-DDB1-CUL4 complex [76]. In a follow-up study, they investigated how E7820 (42) recruits RBM39 to DCAF15 with kinetic analysis and crystal structures (KD = 22 μM) [77]. In an independent study, the Owa group reported that sulfonamide induces ubiquitination and the proteasomal degradation of CAPERα through the formation of CAPERα-sulfonamide-DCAF15 [78]. The Chen group reported an E7820 (42)-based PROTAC (DP1)-targeting BRD4 by employing JQ1 as a target protein ligand (Figure 13). DP1 showed an excellent degradation of BRD4 in SU-DHL-4 cells (DC50 = 10.84 ± 0.92 μM, Dmax = 98%) and inhibited tumor growth in a mouse model [79]. In 2019, the Cravatt group used a chemical proteomic approach that utilized a FKBP12 ligand, SLF, conjugated with the cysteine-directed electrophilic fragments for the discovery of a nuclear localized E3 ligase. After identifying the electrophilic fragments that induce FKBP12 degradation in the nucleus, they used pull-down-based proteomics and identified DCAF16 E3 ligase as their target protein. One of the electrophilic fragments, KB02 (44), were extensively studied after conjugation with the FKBP ligand (KB02-SLF) to monitor its FKBP12 degradation ability (DC50 < 2 μM). In addition, they also designed a BRD4-targeting PROTAC, KB02-JQ1 (DC50 ~ 20 μM), by coupling KB02 (44) to JQ1, and observed DCAF16 as a valuable nuclear localized E3 ligase [80]. With a similar screening strategy, the Cravatt group also developed electrophilic ligands (45) of E3 ligase DCAF11. Using the discovered ligand 45, they synthesized 21-ARL for AR degradation with a recruited DCAF11 [81].
In 2019, the Nomura group reported a set of binders to the E3 ligase RNF4 using the ABPP-based covalent ligand screening approach. The optimized covalent ligand CCW16 (46) (Figure 14) reacted with one of two zinc-coordinated cysteines in the RING domain without affecting the zinc binding ability of RNF4 (IC50 1.8 μM). They demonstrated a covalent PROTAC (CCW 28-3) for BRD4 degradation (Figure 15) [82]. The same group also reported that Nimbolide (47), a natural product exhibiting anticancer activities, was identified as a covalent binder for the E3 ligase RNF114. They used activity-based protein profiling (ABPP) chemoproteomic platforms to discover that Nimbolide (47) binds to a cysteine residue of RNF114. Based on the structure of Nimbolide (47), they developed a covalent PROTAC, XH2, targeting BRD4 (IC50 = 240 nM) [83]. They also demonstrated that Nimbolide (47) can be used as a BCR-ABL-targeting PROTAC by recruiting RNF114. They synthesized the degrader BT1 by coupling the BCR-ABL inhibitor, dasatinib, with Nimbolide (47). They demonstrated that BT1 selectively degraded BCR-ABL rather than c-ABL, which was also observed in previously reported CRBN or VHL-based BCR-ABL PROTACs [84]. In 2021, they used the same ABPP-based approach to discover the fully synthetic covalent ligand EN219 (48), which targets RNF114 (IC50 = 470 nM). The mode of action study showed that EN219 (48) mimics the function of a complex natural product, Nimbolide (47). They developed a covalent PROTAC (ML 2-14) by conjugating EN219 (48) to the BET inhibitor JQ1 for BRD4 degradation [85] (BRD4 [short] DC50 = 14 nM, BRD4 [long] DC50 = 36 nM).
In 2019, the Naito group developed a novel PROTAC recruiting the aryl hydrocarbon receptor (AhR) E3 ligase complex. They conjugated an AhR ligand (β-NF, 49) (Figure 16) with ATRA resulting in a chimeric molecule β-NF-ATRA (Figure 17). β-NF-ATRA, a PROTAC recruiting CRABPs, induced the degradation of CRABPI and CRABPII in an AhR-dependent manner via the ubiquitin–proteasome pathway [86].
The CUL2 E3 ligase FEM1B was recently discovered as an important regulator of the cellular response to reductive stress. In 2022, the Nomura group discovered a chloroacetamide-based covalent ligand, EN106 (50) (Figure 16), as a FEM1B ligand (IC50 = 2.2 μM). They found the formation of a direct covalent bond between EN106 (50) and a cysteine residue on FEM1B by ABPP. They demonstrated that an EN106 (50)-based PROTAC can be used for the FEM1B recruitment toward target protein degradations. The conjugation of EN106 (50) to JQ1 and dasatinib generated NJH-1-106 (DC50 = 250 nM) and NJH-2-142 and showed the successful degradation of BRD4 and BCR-ABL (Figure 17) [87].
Kelch-like ECH-associated protein-1 (KEAP1) has been known to interact with nuclear factor erythroid 2-related factor-2 (Nrf2) to regulate cellular protective proteins. Therefore, the discovery of the protein–protein interaction inhibitor of KEAP1-Nrf2 has attracted attentions for the treatment of stress-related diseases [88]. In 2020, the Nomura group reported a reversible covalent binding PROTAC using a known KEAP1 ligand bardoxolone methyl (CDDO-Me, 51) (Figure 16). They synthesized a bardoxolone-based PROTAC, CDDO-JQ1 (Figure 17) by conjugating bardoxolone to the BET inhibitor JQ1 (DC50 < 100–200 nM). They found that the deletion of a Michael acceptor in CDDO-JQ1 reduced the BRD4 degradation activity. This indicates that the formation of a reversible covalent bond between the cysteine of KEAP1 and CDDO-Me (51) is essential for the recruitment of BRD4 to the KEAP1 E3 ligase [89]. In 2021, the Jin group reported that the E3 ligase KEAP1 can be utilized to develop PROTACs using a highly selective and noncovalent KEAP1 ligand. They developed a KEAP1-recruiting PROTAC, MS83, using the previously reported KEAP1 small molecule ligand (52, KD = 1.3 nM) [90]. MS83 showed a more durable degradation of BRD3 and BRD4 than dBET1 in MDA-MB-468 cells (DC50 < 500 nM) [91]. In 2022, the Lv group discovered a nature product, Piperlongumine (PL, 53), as an E3 ligase ligand. They first confirmed that PL (53) was bound to multiple E3 ligases with a competitive ABPP assay. They synthesized PROTAC, 955, by coupling PL (53) with a CDK9 selective inhibitor (SNS-032) and observed the proteasomal degradation of CDK9 by 955 (DC50 = 9 nM). Interestingly, they identified that KEAP1 was the only E3 ligase protein recruited by 955 via a covalent attachment of PL using the TurboID-bait assay [92]. This observation suggested that E3 ligase selectivity could be changed after the conjugation of the E3 ligase ligand to a target protein ligand.
PROTAC has been developed as a new strategy for disease treatment in the chemical biology community over the past 20 years. PROTACs harness an intracellular proteolytic system for the degradation of a POI. This strategy has emerged as an alternative to overcome the limitations of conventional drug discovery by targeting undruggable proteins. For effective TPD, the selection of E3 ligase ligands and target protein ligands is critical for PROTAC design. Although a number of ligands for various proteins have been reported during the traditional drug discovery campaigns so far, only a few E3 ligase ligands are currently available for TPD. For those reasons, most of the PROTAC research has focused on the demonstration of the TPD concept for the various drug target proteins using CRBN or VHL ligands during the last decades. However, the desired target protein degradations often are compromised due to the cell type or tissue type dependent expression profiles of CRBN or VHL. Moreover, resistance to CRBN- or VHL-based PROTACs has been recently observed. Therefore, the discovery of ligands for new E3 ligases has drawn attention as they can be used for effective PROTAC development. In addition, there is still a lack of understanding of the previously developed PROTACs. For example, a given drug molecule connected to different E3 ligase ligands has shown to exhibit different PROTAC efficiencies [45,74], target selectivities [64], and drug resistance profiles in various cancer cells [4]. Although there is a report on the systematic approach taken to select the E3 ligase ligand in PROTAC designs, it is restricted to the kinase degraders and only three E3 ligase ligands were studied [93]. Their kinase-targeting PROTACs mostly prefer CRBN ligands over VHL or IAP ligands for the efficient protein degradation. Additional studies on other protein families are needed for a comprehensive understanding of E3 ligase ligand selection for efficient protein degradation. This suggests that the ternary complexes that are recruited by bifunctional PROTACs could be susceptible to ligand orientation, the ligand’s hydrophobicity, and the physicochemical properties of the chemical linkers. Further studies related to various ternary complexes should provide clearer pictures on future PROTAC designs. Herein, we reviewed the discovery of various E3 ligase ligands and their applications for the design of PROTACs. Considering more than 600 types of E3 ligases, the development of ligands for various E3 ligases would involve the expansion of the toolbox in TPD to overcome the current limitation of PROTACs. Additionally, the cell or tissue type specific unexplored E3 ligases would be the basis for a new type of PROTAC, which could control a certain protein in a spatially specific way. The spatial protein degradation in lesions would treat diseases very effectively without toxicity or side effects in normal tissues. Therefore, the discovery of novel E3 ligase ligands would be an important goal to expand the landscape of PROTACs toward promising therapeutics in the future. | true | true | true |
PMC9573787 | Zachary Brodie,Erin McCartney,Sergio Toledo | An Adrenal SMARCB1/INI1 Deficient Proximal Epithelioid Sarcoma in a Middle-Aged Female: A Case Report | 16-09-2022 | ini1,proximal epitheloid sarcoma,left adrenalectomy,adrenal mass,smarcb1 | Proximal epithelioid sarcomas are rare soft tissue sarcomas that have been documented in a diverse range of presentations. However, there have been few cases describing adrenal presentations. These neoplasms are thought to be driven by a loss of SWItch/sucrose non-fermentable (SWI/SNF)-related matrix-associated actin-dependent regulator of chromatin subfamily B member 1 (SMARCB1), also known as integrase interactor 1 (INI1). SMARCB1/INI1 is a tumor suppressor gene thought to play a role in multiple malignancies with varying degrees of gene expression. Complete loss of SMARCB1/INI1 has most commonly been described in the English scientific literature as malignant rhabdoid tumors of renal origin within pediatric populations and proximal epithelioid sarcomas in adult populations. We describe a case of a primary adrenal proximal epithelioid sarcoma demonstrating complete loss of SMARCB1/INI1 in a middle-aged adult female. | An Adrenal SMARCB1/INI1 Deficient Proximal Epithelioid Sarcoma in a Middle-Aged Female: A Case Report
Proximal epithelioid sarcomas are rare soft tissue sarcomas that have been documented in a diverse range of presentations. However, there have been few cases describing adrenal presentations. These neoplasms are thought to be driven by a loss of SWItch/sucrose non-fermentable (SWI/SNF)-related matrix-associated actin-dependent regulator of chromatin subfamily B member 1 (SMARCB1), also known as integrase interactor 1 (INI1). SMARCB1/INI1 is a tumor suppressor gene thought to play a role in multiple malignancies with varying degrees of gene expression. Complete loss of SMARCB1/INI1 has most commonly been described in the English scientific literature as malignant rhabdoid tumors of renal origin within pediatric populations and proximal epithelioid sarcomas in adult populations. We describe a case of a primary adrenal proximal epithelioid sarcoma demonstrating complete loss of SMARCB1/INI1 in a middle-aged adult female.
SWItch/sucrose non-fermentable (SWI/SNF)-related matrix-associated actin-dependent regulator of chromatin subfamily B member 1 (SMARCB1), also known as integrase interactor 1 (INI1) is a tumor suppressor gene first discovered in yeast in the 1980s and thought to play a role in chromatin remodeling [1,2]. It has been proposed to be a tumor suppressor gene involved in several signaling pathways [2]. SMARCB1/INI1 deficient tumors are rare and typically classified by the degree of nuclear loss observed on immunohistochemical staining, ranging from partial to complete. Complete loss is most commonly described as malignant rhabdoid tumors (MRT) of renal origin within pediatric populations and as proximal epithelioid sarcomas (ES) in adult populations [1,2]. In children, these neoplasms are thought to be due to germline mutations, typically presenting before the age of three years, and follow a lethal and aggressive course [3]. In adult populations, their origin is less understood. SMARCB1/INI1 deficient proximal ES in adults have been documented in various gastrointestinal, genitourinary, neurological, and cutaneous locations [1,3]. However, very few cases in the literature have described these malignancies arising from the adrenal gland [4-6]. While a loss of SMARCBI/INI1 links these tumors, little is understood regarding other relationships [1]. We describe a case of a malignant primary adrenal proximal epithelioid sarcoma demonstrating complete loss of SMARCB1/INI1 in a middle-aged adult female.
A 44-year-old female patient with a past medical history significant for gastroesophageal reflux disease and hiatal hernia presented to the emergency department (ED) in May of 2021 with a complaint of left-sided abdominal pain with associated bloating for four months. Physical examination findings were remarkable for tenderness to palpation over the mid and left lateral abdomen with no peritoneal signs. Laboratory values were within normal limits. A non-contrast computed tomography (CT) scan of the abdomen and pelvis was performed as part of the workup, and the findings demonstrated a 4x2.4 cm left-sided adrenal mass (Figure 1). She was discharged from the ED with recommendations to follow-up with a primary care provider for further imaging and workup. Upon further evaluation by her primary care provider, a biochemical workup was initiated for pheochromocytoma and she was referred to general surgery. The results of the biochemical workup were within normal limits. However, during evaluation by surgery, she reported anxiety, occasional panic attacks, palpitations, and tremors. A follow-up triple-phase, contrasted CT of the abdomen and pelvis with adrenal protocol was done which demonstrated a 4.3 cm left-sided adrenal mass which was indeterminate based on washout (Figure 2). However, surgical removal was recommended due to the size (>4 cm). Due to the patient's reported symptoms, there was a concern for a subclinical pheochromocytoma, and alpha blockade was started prior to surgical intervention. Laparoscopic left adrenalectomy was performed. A 7.2x4.9x2.3 cm adrenal mass with calcifications was removed and sent for further analysis. The surgical course was uncomplicated, and she was discharged on postoperative day one. There were no postoperative complications. Pathologic examination identified a nodular lesion replacing the entire medulla with a central cavitary area of hemorrhage with clear surgical margins. Microscopic evaluation revealed rhabdoid appearing, sheets and nests of malignant epithelial cells with small nucleoli, high mitotic activity, areas of tumor necrosis, and lymphovascular and perineural invasion. Immunohistochemistry was completed by two laboratories. The pathology results are listed in Table 1. Analysis also demonstrated no rearrangement of Ewing sarcoma RNA-binding protein 1 (ESWR1) but did reveal a complete loss of INI1, consistent with a SMARCB1/INI1-deficient malignant neoplasm. The case was reviewed at tumor board, and a positron emission tomography (PET) scan was recommended to assess for metastasis or recurrent disease. PET CT was chosen due to its high sensitivity (98.5%) and specificity (92%) for assessing malignant adrenal tumors [7]. A PET scan one month postoperatively did not identify any areas of hypermetabolic activity suggesting metastasis or recurrent tumor in the area of the left adrenal gland. Six-month follow-up PET scan was recommended and the patient has had an uneventful follow-up with oncology.
It can be difficult to differentiate between proximal-type ES and MRT as previous literature has reported that both can demonstrate a complete loss of SMARCB1/INI1 expression. MRTs are typically described in pediatric populations, whereas ES are more common in adults [8]. Proximal-type ES have been described most commonly within the pelvic girdle and perineal region [9]. However, some differentiation can be made using immunohistochemical staining. CD34 is positive in half of ES presentations and typically negative in MRT [8,9]. Given the positive staining for CD34 in this case, the tumor is more likely a proximal-type ES rather than MRT. Additionally, immunohistochemistry was negative for S100 and desmin, consistent with proximal ES [9,10]. Proximal ES typically have an aggressive course with poor five-year survival rates. In adult populations, they demonstrate a male predominance [1,10]. Proximal ES have high recurrence rates, reported as high as 77%, and approximately half will have metastatic disease [10]. These malignancies are thought to metastasize via lymphatic spread [10]. If there is metastasis, it will most commonly be to the lungs followed by local lymph nodes [10]. Proximal ES arising from the adrenal gland has been rarely reported. We believe that this is the fourth case described in the literature [4-6]. Compared to previously reported cases, this presentation is consistent with vague abdominal pain and a mass discovered on imaging [4-6]. Non-functional biochemical workup is consistent across published cases. The age of presentation has ranged from 31 to 81 years of age. Of proximal ES arising from the adrenal gland, three out of four cases have been reported in female patients and occurred in the left adrenal gland [4-6]. Half of the adrenal cases had suspected lymph node involvement at resection [4-6]. In both instances, those with lymph node involvement had recurrence or progression of disease after surgical resection [4,5]. Unfortunately, due to such a small number of cases available in the literature, no specific conclusions can be drawn regarding the epidemiology of proximal ES arising in the adrenal gland. This is most likely a primary tumor rather than metastatic disease. Initial imaging was only apparent for an adrenal mass, and follow-up PET CT postsurgical resection did not demonstrate any areas of increased metabolism. The tumor was positive for CDX-2 which can be a marker of colorectal metastasis [11,12]. However, the rest of the clinical history and microscopic investigation were not consistent with adenocarcinoma, making a metastatic colorectal source less likely. Immunohistochemical staining did not identify any other likely tissue source of potential metastasis including lung, thyroid, breast, or myoepithelial tissue [12]. Additionally, with the typically aggressive nature of SMARCB1/INI1 tumors, a metastatic origin would have likely been apparent on PET CT as increased tumor burden within the lungs.
Although rare, the aggressive nature of these malignancies necessitates early removal. While rarely associated with adrenal masses, this case highlights the importance of removing non-functional adrenal masses that are larger than 4 cm. Because of the high rate of metastatic disease at presentation, follow-up after diagnosis and surgical removal should include a PET CT to assess tumor burden/metastasis. Follow-up should be consistent with national soft tissue sarcoma guidelines and based on initial tumor location. | true | true | true |
|
PMC9573893 | Nicholas J. Queen,Xunchang Zou,Jacqueline M. Anderson,Wei Huang,Bhavya Appana,Suraj Komatineni,Rachel Wevrick,Lei Cao | Hypothalamic AAV-BDNF gene therapy improves metabolic function and behavior in the Magel2-null mouse model of Prader-Willi syndrome | 27-09-2022 | Prader-Willi syndrome,BDNF,hypothalamus,hypothalamic,metabolic,metabolism,adeno-associated virus,AAV,gene therapy,molecular therapy | Individuals with Prader-Willi syndrome (PWS) display developmental delays, cognitive impairment, excessive hunger, obesity, and various behavioral abnormalities. Current PWS treatments are limited to strict supervision of food intake and growth hormone therapy, highlighting the need for new therapeutic strategies. Brain-derived neurotrophic factor (BDNF) functions downstream of hypothalamic feeding circuitry and has roles in energy homeostasis and behavior. In this preclinical study, we assessed the translational potential of hypothalamic adeno-associated virus (AAV)-BDNF gene therapy as a therapeutic for metabolic dysfunction in the Magel2-null mouse model of PWS. To facilitate clinical translation, our BDNF vector included an autoregulatory element allowing for transgene titration in response to the host’s physiological needs. Hypothalamic BDNF gene transfer prevented weight gain, decreased fat mass, increased lean mass, and increased relative energy expenditure in female Magel2-null mice. Moreover, BDNF gene therapy improved glucose metabolism, insulin sensitivity, and circulating adipokine levels. Metabolic improvements were maintained through 23 weeks with no adverse behavioral effects, indicating high levels of efficacy and safety. Male Magel2-null mice also responded positively to BDNF gene therapy, displaying improved body composition, insulin sensitivity, and glucose metabolism. Together, these data suggest that regulating hypothalamic BDNF could be effective in the treatment of PWS-related metabolic abnormalities. | Hypothalamic AAV-BDNF gene therapy improves metabolic function and behavior in the Magel2-null mouse model of Prader-Willi syndrome
Individuals with Prader-Willi syndrome (PWS) display developmental delays, cognitive impairment, excessive hunger, obesity, and various behavioral abnormalities. Current PWS treatments are limited to strict supervision of food intake and growth hormone therapy, highlighting the need for new therapeutic strategies. Brain-derived neurotrophic factor (BDNF) functions downstream of hypothalamic feeding circuitry and has roles in energy homeostasis and behavior. In this preclinical study, we assessed the translational potential of hypothalamic adeno-associated virus (AAV)-BDNF gene therapy as a therapeutic for metabolic dysfunction in the Magel2-null mouse model of PWS. To facilitate clinical translation, our BDNF vector included an autoregulatory element allowing for transgene titration in response to the host’s physiological needs. Hypothalamic BDNF gene transfer prevented weight gain, decreased fat mass, increased lean mass, and increased relative energy expenditure in female Magel2-null mice. Moreover, BDNF gene therapy improved glucose metabolism, insulin sensitivity, and circulating adipokine levels. Metabolic improvements were maintained through 23 weeks with no adverse behavioral effects, indicating high levels of efficacy and safety. Male Magel2-null mice also responded positively to BDNF gene therapy, displaying improved body composition, insulin sensitivity, and glucose metabolism. Together, these data suggest that regulating hypothalamic BDNF could be effective in the treatment of PWS-related metabolic abnormalities.
Prader-Willi syndrome (PWS) is a contiguous gene syndrome that occurs in approximately 1 in 15,000 individuals. Individuals with PWS display developmental delays, cognitive impairment, excessive appetite, obesity, hypothalamic hypogonadism, obsessive compulsive behavior, anxiety, and temper tantrums.2, 3, 4 Current treatments to address metabolic dysfunction and behavioral abnormalities are limited to strict supervision of daily food intake and growth hormone (GH) therapy., While GH therapy remains the standard of care for PWS, it has several shortcomings, as (1) it requires daily administration and thus patient compliance, (2) it does not target the hypothalamic origins of PWS, (3) data about its efficacy in older adults are scant, and (4) exclusion criteria include common comorbidities such as severe obesity, uncontrolled diabetes, and active psychosis. Management of individuals with PWS is otherwise largely supportive and results in high levels of caregiver burden.9, 10, 11, 12 Loss of function of MAGEL2 (MAGE Family Member L2) is thought to contribute to several aspects of PWS pathophysiology, including alterations in the hypothalamic leptin-proopiomelanocortin (POMC) pathway. This pathway interprets peripheral signals of energy needs to drive feeding or fasting through opposing neuronal populations. In mice, Magel2 is required for leptin-mediated depolarization of hypothalamic anorexigenic POMC neurons. Furthermore, loss of Magel2 reduced anorexigenic α-melanocyte-stimulating hormone (α–MSH) axons,, while orexigenic agouti-related peptide (AGRP) fibers remained unchanged., Thus, in the PWS-driven absence of Magel2, the response to peripheral hormones is blunted, resulting in an inability to maintain energy homeostasis. Brain-derived neurotrophic factor (BDNF) serves as a potential therapeutic target, as it functions downstream in the leptin-POMC pathway and regulates energy homeostasis and behavior.18, 19, 20, 21, 22, 23, 24, 25 Furthermore, individuals with PWS display reductions in peripheral BDNF and PWS-related transcriptomic alterations in the leptin-POMC pathway may result from reduced BDNF expression. Administration of native or recombinant BDNF to the brain for therapeutic means proves difficult due to (1) the need for fine delivery localization, because, as a growth factor, BDNF has pleiotropic roles across various brain regions; (2) unfavorable pharmacokinetics; and (3) the need for repeated dosing.27, 28, 29 Previous work by our laboratory has shown that hypothalamic injection of a recombinant adeno-associated virus (rAAV)-BDNF vector in mice efficiently alleviates deficits in the leptin-POMC signaling pathway, thus mitigating obesity, hyperphagia, hyperglycemia, and hepatic steatosis. Moreover, adeno-associated virus (AAV)-BDNF gene therapy has been shown to reduce anxiety- and depression-like behavior and improve cognitive function.31, 32, 33 To facilitate clinical translation, we developed an autoregulatory AAV system to control therapeutic BDNF expression and mimic the body’s natural feedback systems. This autoregulatory approach leads to a sustainable plateau of body weight after substantial weight loss is achieved—minimizing the risk of cachexia—and has been validated in various physiologically related animal models.,,34, 35, 36 Given these data, we hypothesized that applying hypothalamic BDNF gene therapy to the Magel2-null murine model of PWS would ameliorate the metabolic dysregulation that results from deficiencies in the hypothalamic leptin-POMC signaling pathway. In this preclinical study, we assess the translational potential of hypothalamic AAV-BDNF gene therapy as a first-in-class therapeutic for PWS.
Prior to vector administration, an EchoMRI was performed to assess body composition and allow for experimental group randomization. Consistent with genotype, female Magel2-null mice exhibited increased fat mass and decreased lean mass over wild-type controls at baseline (Figures S1A–S1D). Adult (16–20 weeks old) female wild-type and Magel2-null mice were injected with either AAV-yellow fluorescent protein (YFP) or AAV-BDNF vectors (Figure 1A) to the arcuate nucleus of the hypothalamus (ARC)/ventromedial hypothalamus (VMH), known sites of BDNF and POMC/AGRP action., In vivo monitoring continued over 23 weeks, beginning with metabolic measures (Figure 1B). As early as 3 weeks post injection, AAV-BDNF-treated female Magel2-null mice exhibited a significant decrease in body weight and prevention of excessive body weight gain over genotype-matched counterparts injected with AAV-YFP (Figures 1C and 1D). This reduction in body weight was maintained over the 23-week experiment. Magel2-null mice do not exhibit hyperphagia, despite manifestation of various other metabolic abnormalities., Consistent with previous reports, we observed reduced food intake in Magel2-null mice compared with wild-type counterparts (Figure 1E). AAV-BDNF-treated Magel2-null mice displayed a further reduction in food intake, consistent with BDNF’s anorexigenic actions.,, At 23 weeks post AAV injection, female Magel2-null mice treated with AAV-YFP displayed a significant increase in relative fat mass and decrease in relative lean mass over wild-type controls, whereas AAV-BDNF gene therapy rescued wild-type-like fat and lean mass composition (Figures 1F and 1G). AAV-BDNF-driven reductions in total body weight could be accounted for by losses in both absolute fat mass and lean mass (Figures S1E and S1F). Additional body composition measurement time points can be found in Figure S2. At 6 weeks post AAV injection, an insulin tolerance test was performed. AAV-YFP-injected female Magel2-null mice exhibited no changes in sensitivity to an exogenous insulin bolus over wild-type counterparts, whereas AAV-BDNF-treated Magel2-null mice displayed improved insulin sensitivity (Figures 1H and 1I). At 8 weeks post AAV injection, a glucose tolerance test was performed to assess glycemic processing ability. AAV-YFP-treated Magel2-null mice exhibited worsened processing of an exogenous glucose bolus over wild-type controls; this deficit was rescued with AAV-BDNF administration (Figures 1J and 1K). Notably, compared with wild-type counterparts, AAV-YFP-injected Magel2-null mice exhibited elevated fasting blood glucose at baseline (t = 0) for both the insulin (Figure 1H) and glucose tolerance tests (Figure 1J), which was reversed by AAV-BDNF treatment. Between 10 and 12 weeks post AAV injection, mice were subjected to indirect calorimetry. Body composition has known effects on energy expenditure (EE); larger animals typically have higher absolute rates of EE, due in part to increases in total metabolically active mass, whereas smaller animals typically have higher per-kilogram rates of EE. At the time of indirect calorimetry, AAV-YFP-treated female Magel2-null mice exhibited increased body weight over wild-type counterparts (Figure 2A), while AAV-BDNF-treated mice had reduced body weight. Accordingly, we observed lower hourly EE in the lighter AAV-BDNF-treated mice (Figures 2B and 2C), likely a byproduct of reductions in total metabolically active mass following gene therapy. Next, we compared the relationship between EE and body weight for each experimental group. Female AAV-BDNF-treated Magel2-null mice displayed increased EE per gram of body weight (Figure 2D), denoting increased energy efficiency over AAV-YFP genotype-matched counterparts (p = 0.049) and wild-type counterparts (p = 0.061). The respiratory exchange ratio (RER; calculated as VCO2/VO2) provides an indirect measurement of the predominant metabolic source (carbohydrate versus fat) for whole-body EE. Magel2-null AAV-YFP mice displayed reduced RER (Figures 2E and 2F) compared with wild-type counterparts, indicating fat as the main fuel source. This genotype-driven deficit was ameliorated by AAV-BDNF gene therapy; the RER indicated a shift toward carbohydrate use over fat use. Similar to previous reports,, Magel2-null were hypoactive compared with wild-type counterparts (Figures 2G and 2H). Importantly, AAV-BDNF gene therapy did not ameliorate genotype-driven lethargy; increases in various metrics of metabolic function could not be explained by increased locomotor activity. Additional weight-normalized and absolute measures of VO2 and VCO2 can be found in Figure S3.
A similar experiment was initiated in male mice (between 16 and 22 weeks old at time of AAV injection). As seen in the female mice, metabolic measures were normalized in AAV-BDNF-treated male mice through 10 weeks post AAV injection. AAV-BDNF-treated male Magel2-null mice exhibited a significant decrease in body weight and prevention of excessive body weight gain over genotype-matched counterparts injected with AAV-YFP (Figures 3A and 3B). We observed a lack of hyperphagia in male AAV-YFP-injected Magel2-null mice compared with wild-type counterparts (Figure 3C). AAV-BDNF-treated Magel2-null mice displayed reduced food intake compared with genotype-matched controls (Figure 3C). At 4 weeks post AAV injection, male Magel2-null mice treated with AAV-YFP displayed a significant increase in relative fat mass and decrease in relative lean mass over wild-type controls, whereas AAV-BDNF gene therapy rescued wild-type-like fat and lean mass composition (Figures 3D and 3E). AAV-BDNF-driven reductions in total body weight could be accounted for by losses in both absolute fat mass and lean mass (Figures 3F and 3G). At 6 weeks post AAV injection, AAV-YFP-injected male Magel2-null mice exhibited no alterations in sensitivity to an exogenous insulin bolus over wild-type counterparts, whereas AAV-BDNF-treated Magel2-null mice displayed improved insulin sensitivity (Figures 3H and 3I). At 8 weeks post AAV injection, AAV-YFP-treated male Magel2-null mice exhibited a reduction in glycemic processing ability over wild-type controls; this deficit was rescued with AAV-BDNF administration (Figures 3J and 3K).
To assess safety and determine whether hypothalamic BDNF gene therapy alters behavior in female Magel2-null mice, comprehensive behavioral profiling was performed following metabolic assessments (Figure 4A). At 13 weeks post AAV injection, an open field test was performed, as Magel2-null mice exhibit alterations in exploratory activity and anxiety-like behavior., Magel2-null AAV-YFP-treated mice exhibited a trending—but not significant (p = 0.081)—decrease in exploratory activity over wild-type counterparts (Figure 4B). Magel2-null AAV-BDNF mice exhibited a significant increase in exploratory activity over AAV-YFP-treated Magel2-null counterparts Figure 4B). No changes were observed in percentage time spent in the periphery (Figure 4C) or center (Figure 4D) of the open field arena, indicating no changes in anxiety-like behavior. At 14 weeks post injection, the novel object recognition test was performed. Female Magel2-null mice have been shown to be averse to novel objects and environments. Consistent with the literature, we observed a reduced discrimination index in Magel2-null YFP-treated mice over wild-type controls (Figure 4E). This deficit was ameliorated in the Magel2-null mice receiving AAV-BDNF gene therapy treatment (Figure 4E). One interpretation of these results indicates that AAV-BDNF treatment resolves non-discrimination of novel objects previously reported in Magel2-null mice. The alternative interpretation of these data suggests that AAV-BDNF gene therapy improves short-term memory, in Magel2-null mice. At 16 weeks post injection, mice were subjected to the marble burying test to assess repetitive behaviors., Consistent with previous work, Magel2-null female mice buried fewer marbles than wild-type counterparts (Figure 4F). Wild-type-like burying behavior was observed in Magel2-null mice receiving AAV-BDNF gene therapy. Previous reports suggest Magel2-deficient mice have altered social phenotypes., At 19 weeks post injection, mice were subjected to the three-chamber sociability test to assess social affiliation and social novelty preference. Mice were given the opportunity to wander three chambers while investigating social stimuli. In the first test phase—assessing social affiliation—mice were exposed to a novel confined peer and an opposing empty chamber. We observed no genotype- or AAV-BDNF-induced alterations in social preference (Figure 4G) during this first test phase. During the second test phase—assessing social novelty engagement—mice were exposed to both familiar and novel confined peers in opposing chambers. Similarly, we observed no genotype- or AAV-BDNF-induced alterations in social novelty-seeking behavior (Figure 4H). At 21 week post injection, mice were subjected to the tail suspension test to assess depression-like behavior. AAV-YFP-treated Magel2-null mice exhibited increased levels of immobility compared with wild-type counterparts (Figure 4I). AAV-BDNF gene therapy yielded a trending, but not yet significant (p = 0.060), reduction in immobility of Magel2-null mice during this test (Figure 4I).
Female mice were euthanized at 23 weeks post AAV injection and tissues were collected. No changes were observed in relative brown adipose tissue (BAT) weight (Figure 5A). Consistent with EchoMRI observations, AAV-YFP-treated Magel2-null mice displayed increased relative mass of three white adipose tissue (WAT) depots—the inguinal white adipose tissue (iWAT), gonadal white adipose tissue (gWAT), and retroperitoneal adipose tissue (rWAT)—compared with wild-type counterparts (Figure 5A). Increases in Magel2-null mouse WAT depot size were ameliorated with AAV-BDNF gene therapy (Figure 5A) and relative liver weight was normalized. Serum was profiled to assess changes in systemic metabolism following AAV-BDNF gene therapy. Leptin serves as a central-peripheral messenger to maintain energy homeostasis and its production is positively correlated with adipose tissue mass. Consistent with gross adipose observations, AAV-YFP-treated Magel2-null mice exhibited increased serum leptin over wild-type counterparts (Figure 5B). Heightened serum leptin in Magel2-null mice was ameliorated with AAV-BDNF administration (Figure 5B). Adiponectin plays roles in insulin sensitivity, glucose homeostasis, and systemic metabolic function; circulating adiponectin levels are negatively correlated with adipose tissue mass and impaired glucose tolerance. Magel2-null mice exhibited no significant increase in total (Figure 5C) or high-molecular-weight (HMW) adiponectin (Figure 5D) compared with wild-type counterparts. AAV-BDNF gene therapy increased both total and HMW adiponectin in Magel2-null mice over wild-type AAV-YFP-treated mice (Figures 5C and 5D). We observed no genotype- or AAV-BDNF-induced alterations changes in the total/HMW adiponectin ratio (Figure 5E). Increased adiponectin/leptin ratios are correlated with reduced body mass index and cardiometabolic risk., AAV-YFP-treated Magel2-null mice exhibited reduced adiponectin/leptin ratios compared with wild-type controls; this deficit was ameliorated with AAV-BDNF gene therapy (Figure 5F). No genotype- or gene-therapy-induced changes were observed in circulating glucose levels (Figure 5G). Magel2-null AAV-BDNF-treated mice exhibited a significant decrease in fasting serum insulin over AAV-YFP-treated counterparts (Figure 5H). Similar results were observed in the HOMA-IR (homeostatic model assessment for insulin resistance) index; together, both are indicative of improved insulin sensitivity following AAV-BDNF gene therapy (Figure 5I). No changes were observed in two systemic inflammatory markers, chemokine (C-C motif) ligand 2 (CCL2, also known as MCP-1) and plasminogen activator inhibitor-1 (PAI-1) (Figures 5J and 5K). We observed a gene-therapy-induced increased in serum corticosterone (Figure 5L), consistent with previous reports that suggest BDNF activates the hypothalamic-pituitary-adrenal (HPA) axis.59, 60, 61 The GH/IGF-1 (insulin-like growth factor 1) axis is controlled in part by the hypothalamus and is dysfunctional in PWS patients and Magel2-null mice. Importantly, GH administration remains the current standard of care for PWS patients., AAV-YFP-treated wild-type and Magel2-null mice displayed no difference in GH levels, whereas BDNF gene therapy treatment resulted in a strong induction of circulating GH (Figure 5M). Serum IGF-1 was reduced in AAV-YFP-treated Magel2-null mice compared with wild-type counterparts; this genotype-driven decrease in IGF-1 was not ameliorated by AAV-BDNF gene therapy (Figure 5N).
Previous work by our laboratory described a brain-adipose axis driven by hypothalamic BDNF expression. Environmental- or genetic-driven increases in hypothalamic BDNF confer a white-to-brown adipose phenotypic shift via increased sympathetic tone. Accordingly, we profiled expression of various metabolism-relevant genes in the largest visceral WAT depot, the gWAT (Figure 6A). Magel2-null mice showed a trend of Lep upregulation in gWAT (not reaching significance) compared with wild type, which appeared to be corrected by BDNF gene therapy. It is possible that the genotype-induced increase in circulating levels of leptin is not caused solely by transcription changes, but instead reflects increases in total fat mass. We observed no significant differences in Adipoq (encoding adiponectin), Hsl (encoding hormone-sensitive lipase), and Pten (encoding phosphatase and tensin homolog) expression. We observed a genotype-induced reduction in Adrb3 (encoding adrenoceptor β3). Magel2-null mice exhibited increased Ppargc1a (encoding peroxisome proliferator-activated receptor gamma coactivator 1-alpha) expression following gene therapy. Consistent with gross tissue observations, Magel2-null mice exhibited increased adipocyte size over wild-type counterparts; genotype-driven increases in adipocyte size were ameliorated with AAV-BDNF gene therapy (Figures 6B–6D).
To verify the expression of our vector, transgene expression was probed using various methods. Hypothalamic gene expression profiling revealed a 4.5-fold upregulation of Bdnf in AAV-BDNF-treated mice over AAV-YFP controls (Figure 7A). No changes were observed in BDNF receptor TrkB-FL (encoding tropomyosin receptor kinase B full length), Agrp, or Pomc gene expression. Consistent with obese states, we observed a baseline reduction in Mc4r (encoding melanocortin-4 receptor) expression and increase in Insr (encoding insulin receptor) expression in Magel2-null mice. Obrb (encoding leptin receptor long form) expression was increased in Magel2-null mice following BDNF gene therapy. We observed reduced expression of Crh (encoding corticotrophin-releasing hormone) in AAV-YFP-injected Magel2-null mice compared with wild-type counterparts; this deficit was ameliorated with AAV-BDNF gene therapy. Consistent with increased Bdnf expression, we observed a strong upregulation of Vgf (non-acronymic) in AAV-BDNF-treated mice over AAV-YFP-treated mice. Immunoblotting of a hemagglutinin (HA) tag was performed as a proxy for BDNF transgene expression. As expected, we observed no HA signal in AAV-YFP-treated controls and positive HA protein in AAV-BDNF-treated mice (Figure 7B). For spatial targeting validation, representative brain samples were sectioned and YFP fluorescence was observed in the ARC and VMH of the hypothalamus (Figure 7C).
For the past two decades, GH therapy and strict supervision of food intake have remained the standard of care for PWS. Unfortunately, caregiver burden remains high, and, in some estimates, exceeds that of caregivers for persons with dementia, Alzheimer’s, and traumatic brain injury., Recent work has focused on therapeutics that can regulate food intake and EE through mechanisms targeting the hypothalamic leptin-POMC pathway. Melanotan II (MT-II) and setmelanotide are MC4R agonists that have known roles in food intake reduction. Although Magel2-null mice are sensitive to both compounds,, setmelanotide translation into the clinic proved less than fruitful (NCT02311673). More recently, intranasal carbetocin was proposed for treatment of hyperphagia, anxiousness, and distress associated with PWS (CARE-PWS; NCT03649477). Study trial difficulties stemming from early 2020 COVID-19 pandemic lockdowns limited efficacy data and ultimately led the US Food and Drug Administration (FDA) to deem the phase 3 study’s data to be insufficient for approval, although the drug was regarded as generally safe and well tolerated. To date, no FDA-approved therapies exist to target the day-to-day challenges PWS patients and families face surrounding excessive hyperphagia and emotional reactivity. Notably, our work represents the first attempt at an AAV-based, PWS-targeted gene therapy. The gene-therapy field has grown dramatically over the last decade as researchers and clinicians have focused efforts on treating genetic diseases such as PWS. With all gene therapies, the risk/benefit ratio must be considered. While general obesity can be better managed with interventions targeting exercise and nutrition, certain severe genetic forms of obesity may indicate the use of CNS-targeted gene therapies. Importantly, typical obesity interventions—such as self-monitoring of food intake—remain difficult in the PWS population due to incessant food-seeking behavior and associated cognitive rigidity. Patients are highly dependent upon caregivers for maintenance of “food security; restricting of food access and control of food intake occurs through drastic measures such as locks on refrigerators and pantries. To some, surgical methods may seem to be a better alternative to the lock-and-key method. While sleeve gastrectomy was well tolerated in Magel2-null mice, bariatric procedures in PWS patients result in far greater levels of complications than those observed in non-PWS obese patients and furthermore require patient or caregiver post-surgical compliance to guarantee adequate weight loss. Exercise and behavioral interventions have proved successful70, 71, 72 to treat symptoms but remain limited in their ability to address the root causes of satiety imbalance. In sum, therapies targeting the hypothalamic and genetic roots of PWS are sorely needed. Energy homeostasis requires dynamic feedback systems; peripheral tissues are in constant communication with the brain to convey the need for fasting or feeding. Our autoregulatory AAV-BDNF vector provides significant advantages over other pharmacological interventions, as transgene expression is inherently tied to central-peripheral feedback systems and thus reflects the host’s physiological needs. In contrast to repeated dosing of MC4R agonists, our therapeutic vector would need to be administered only once for long-term efficacy and would require limited patient/caregiver dosing compliance. Notably, AAV vectors have shown reasonable safety and tolerability when administered to the human CNS, although administration methods vary.73, 74, 75, 76 Development and optimization of minimally invasive administration methods continues to be an area of focus in the field. Importantly, intracranial injection of hypothalamic tissue in humans—as in this proof-of-concept experiment—remains highly challenging due to the region’s deep location within the brain. Alternative administration methods (e.g., delivery of AAV to the hypothalamus via endoscopic endonasal procedures) must be considered on the path to clinical development. As with any preclinical AAV work, future experiments will need to validate the vector in larger animal models, determine the optimal serotype and administration method, and assess toxicity and immune response. Furthermore, determination of the proper dose will be necessary to observe a functional response, but not one that is clinically dangerous. Special care must be taken to avoid hypoglycemic states, as individuals with PWS remain insulin sensitive. While cachexia is unlikely due to the autoregulatory nature of the vector, it remains unclear whether body weight and food intake reduction below that of healthy individuals would be clinically indicated, as was observed in our AAV-BDNF-treated Magel2-null mice. Our data demonstrate that hypothalamic BDNF gene therapy results in sustained metabolic improvement in Magel2-null mice. Several caveats and limitations of the current study remain. MAGEL2 is just one of many genes located within the PWS deletion region of chromosome 15q11-q13. Mouse models with large deletions at the PWS locus were found to display high levels of lethality, leading researchers to favor single-gene knockout models targeting various PWS genes including Magel2, Snord116, and Ndn., As such, current preclinical models remain limited, as they recapitulate some, but not all, PWS phenotypes., Furthermore, animal models cannot fully mirror nuanced human emotionality and behavioral aspects of food-seeking behavior. Regarding specific limitations of the model used herein, Magel2-null mice do not mirror human hyperphagia,, despite broad systemic metabolic dysfunction and impaired sensitivity to peripheral hormones. Notably, Schaaf-Yang syndrome is driven by loss of function of MAGEL2; individuals with Schaaf-Yang syndrome display many PWS-like phenotypes, albeit with a notable lack of hyperphagia and obesity.80, 81, 82 Despite this modeling issue, we observed an anorexigenic effect of AAV-BDNF administration that may prove useful in PWS-relevant translation efforts. Moreover, this BDNF gene therapy reversed hyperphagia in Mc4r-deficient mice, a model for the most common monogenic form of human obesity. On a different note, while AAV-BDNF gene therapy ameliorated metabolic deficits in both female and male mice, sex differences may exist for other outcomes. Notably, BDNF has known interactions with estrogen that may influence metabolic outcomes,83, 84, 85 and previous work suggests female and male Magel2-null mice display some variations in endocrine responses. Further investigation is warranted prior to translational efforts. The hypothalamic leptin-POMC pathway remains a prime target for PWS therapeutics. In Magel2-null mice, leptin administration fails to (1) reduce food intake, (2) depolarize anorexigenic POMC+ neurons in the ARC, and (3) induce STAT3 phosphorylation. Additional work suggests that MAGEL2 regulates leptin receptor cell surface abundance, subcellular localization, and lysosomal degradation. In tandem with unchanged levels of orexigenic AGRP fibers, alterations in leptin sensitivity,, and POMC activity underlie metabolic dysregulation in the Magel2-null mouse model. Here, we did not observe changes in POMC expression as measured by qPCR. It is unclear whether normalization occurred over time or an alternative explanation exists. Future work by our laboratory will provide an unbiased screening of transcriptomic-wide changes in the hypothalamus of Magel2-null mice following short-term AAV-BDNF gene therapy. On a different note, BDNF works downstream of aberrant leptin-POMC signaling to regulate food intake and energy homeostasis. Previous work by our group describes how environmental or genetic-driven increases in hypothalamic BDNF yield elevated β-adrenergic signaling, thermogenesis, and browning in adipose tissue alongside reductions in fat depot size and circulating leptin levels. We observed reduced circulating leptin levels and increased hypothalamic Obrb expression in Magel2-null mice following BDNF gene therapy. Notably, reduced leptin levels restore leptin sensitivity in POMC neurons, providing one feasible mechanism for observations presented herein. Additional work may delineate whether BDNF gene therapy alters central leptin sensitivity through direct or indirect means, thus normalizing central-peripheral feedback systems in the Magel2-null mouse model. Such work may provide additional mechanistic insights relevant to PWS etiology and potential druggable targets. In this preclinical study, we conducted several behavior tests as safety assessments and observed no adverse behavioral changes in female Magel2-null mice following AAV-BDNF treatment. Unexpectedly, AAV-BDNF gene therapy rescued genotype-associated behavioral alterations in female Magel2-null mice, improving novel object recognition and increasing exploratory activity. Improved cognition in the novel object recognition test is of particular interest, as it is thought to be hippocampus dependent. It remains to be seen whether systemic metabolic improvements induced by hypothalamic BDNF gene transfer reflect back to extrahypothalamic regions of brain, thereby ameliorating cognitive and affective deficits of PWS models. Regarding the open field and marble burying tests, we observed genotype-driven behavioral deficits were ameliorated by AAV-BDNF gene therapy. While baseline reductions in total distance traveled and marbles buried (Figures 4B and 4F) can feasibly be explained by the hypoactivity observed in the Magel2-null model, the rescue of these behavioral outcomes following BDNF gene therapy cannot be explained by a mere amelioration of hypoactivity; AAV-BDNF-treated Magel2-null mice displayed no change in locomotion over AAV-YFP-injected Magel2-null counterparts (Figures 2G and 2H). Another interesting finding is that Magel2-null mice displayed increased immobility in tail suspension test—a commonly used test for depression-like behavior—while BDNF gene therapy rescued this deficit. Notably, scarce data regarding depression-like behavior have been reported. It is possible that our observations reflect differences in lean muscle mass and overall weight, known confounds of the tail suspension test. Nevertheless, these behavioral data encourage further investigation of potential therapeutic benefits beyond metabolic improvement. In summary, we present preclinical data that suggest hypothalamic AAV-BDNF gene therapy is efficacious and safe for treatment of metabolic dysfunction in female and male Magel2-null mice. Furthermore, we observed BDNF gene therapy has no adverse effects on behavior in female Magel2-null mice. These proof-of-concept data indicate BDNF as a promising molecular target for PWS and other genetic forms of obesity.
Magel2-null mice harbor a maternally inherited imprinted/silenced wild-type allele and a paternally inherited Magel2-lacZ knockin allele that abolishes endogenous Magel2 gene function. Male mice containing the Magel2-lacZ allele (Jackson Labs #009062) were bred with female C57BL/6 mice to produce both wild-type mice and Magel2-null littermates. Mice were genotyped from ear notch biopsies. Identification of mutant offspring was performed by polymerase chain reaction genotyping with Magel2 and LacZ oligonucleotide primers (common forward, 5′-ATGGCTCCATCAGGAGAAC; Magel2 reverse, 5′-GATGGAAAGACCCTTGAGGT; and LacZ reverse, RW4237, 5′-GGGATAGGTCACGTTGGTGT). Magel2-null mice develop metabolic deficiencies over time,; systemic manifestation occurs by 16 weeks of age. Therefore, all mice were between 16 and 22 weeks of age at experiment start date. All mice had ad libitum access to food (normal chow diet, 11% fat, caloric density 3.4 kcal/g, Teklad) and water. Mice were housed in standard laboratory cages (19.4 × 18.1 × 39.8 cm) within temperature (22°C–23°C) and humidity (30%–70%) controlled rooms under a 12:12 light:dark cycle. All animal experiments were in accordance with the regulations of The Ohio State University’s Institutional Animal Care and Use Committee (IACUC).
The rAAV vector (Figure 1A) was designed to be self-regulating in nature to allow for sustainable, safe weight loss. HA-tagged human BDNF (HA-BDNF, referred to as BDNF) or destabilized YFP control (dsYFP, referred to as YFP) was inserted in a multiple cloning site following a cytomegalovirus enhancer and a chicken β-actin promoter. Woodchuck posttranscriptional regulatory element (WPRE) and a bovine GH polyadenosine (BGH poly(A)) tail followed BDNF or YFP transgenes. In the BDNF vector, a second regulatory cassette included an AGRP promoter driving a microRNA targeting BDNF (miR-Bdnf) and WPRE/polyA elements. As body weight decreases and AGRP is physiologically induced, microRNA expression is activated to inhibit BDNF transgene expression; this leads to a sustainable plateau of body weight after a substantial weight loss is achieved, thus limiting the risk of cachexia. All vectors were packaged into serotype 1 capsids and purified by iodixanol gradient centrifugation as previously described.
Mice received either AAV-BDNF-miR-Bdnf (denoted as AAV-BDNF throughout the text) or AAV-YFP to the hypothalamus, 1 × 1010 viral genomes per site, bilaterally. Prior to surgery, mice were treated with Ethiqua XR (3.25 mg/kg body weight), an extended-release buprenorphine solution, for pain management. Mice were anesthetized with a single intraperitoneal dose of ketamine/xylazine (80 mg/kg and 5 mg/kg) and secured via ear and incisor bars on a stereotaxic frame (Kopf, Tujunga, CA). Anesthesia was further maintained with 2.5% isoflurane at 1 L/min during the injection procedure. A single midline incision was made through the scalp to expose the skull, and two small holes were made with a dental drill above the injection sites. rAAV vectors were administered bilaterally (0.5 μL per site) by a 10-μL Hamilton syringe (Reno, NV) and a Micro4 Micro Syringe Pump Controller (World Precision Instruments, Sarasota, FL) at 200 nL/min to the arcuate/ventromedial hypothalamus (AP, −1.1 mm; ML, ±0.50 mm; DV, −6.20). When the infusion was finished, the syringe was slowly retracted from the brain and the scalp was sutured. Animals were returned to clean cages resting atop a heating pad. Mice were provided with supplemental care—HydroGel (ClearH2O, Westbrook, ME) and mash—and were carefully monitored post surgery until fully recovered.
Food was measured at the cage level on a weekly basis throughout the experiment. Data were excluded if they were collected in weeks during which invasive profiling occurred that might disrupt normal feeding patterns (e.g., removal from home cages for indirect calorimetry monitoring and fasting periods for glucose tolerance tests).
An EchoMRI was utilized to measure fat and lean mass in live mice without anesthesia. At baseline, 4 weeks, between 14 and 15 weeks, and 23 weeks post AAV injection, body composition analysis was performed with a 3-in-1 Analyzer (EchoMRI LLC, Houston, TX) according to manufacturer instructions. Mice were subjected to a 5-Gauss magnetic field and whole-body masses of fat, lean, free water, and total water were determined during separate cycles by manufacturer software comparison with a canola oil standard.
At 6 weeks post AAV injection, mice were injected intraperitoneally with an insulin solution (0.5 U of insulin per kilogram body weight) following a 2-h fast. Tails were cut (<0.5 mm) with surgical scissors to elicit blood flow. Blood was obtained from the tail at baseline, 15, 30, 60, 90, and 120 min after insulin injection. Blood glucose concentrations were measured with a portable glucose meter using default manufacturer settings (Bayer Contour Next). Following blood collection, styptic powder was used to stop bleeding.
At 8 weeks post AAV injection, mice were injected intraperitoneally with glucose solution (2.0 g glucose per kilogram body weight) after a 16-h overnight fast. Tails were cut (<0.5 mm) with surgical scissors to elicit blood flow. Blood was obtained from the tail at 0, 15, 30, 60, 90, and 120 min after glucose injection. Blood glucose concentrations were measured with a portable glucometer using default manufacturer settings (Bayer Contour Next). Following blood collection, styptic powder was used to stop bleeding.
Between 10 and 12 weeks post AAV injection, mice underwent indirect calorimetry using the Comprehensive Laboratory Animal Monitoring System (CLAMS; Columbus Instruments, Columbus, OH). In short, the indirect calorimetry system pushes a flow of fresh air through gas-tight animal housing cages. The system collects and mixes expired air, measures the flow rate, and determines the incoming/outgoing O2 and CO2 concentrations. Mice were placed in calorimetry chambers and allowed to habituate to the testing environment for 16-18 h. Various physiological and behavioral parameters (VO2, VCO2, RER, EE, and ambulation) were recorded for 24 h. Mice were returned to their home cages after indirect calorimetry was performed.
At 13 weeks post AAV injection, mice were individually placed into the center of an open square arena (60 × 60 cm, enclosed by walls of 48 cm). Each mouse was allowed to explore the arena for 10 min, during which time and locomotion—in the center and the periphery of the open field—was recorded and analyzed via TopScan (Clever Sys) software. Between each trial, the arena was cleaned with 70% ethanol to remove odor cues.
At 14 weeks post AAV injection, mice were subjected to the novel object recognition test. For the test familiarization period, mice were placed in an open arena (60 × 60 cm, enclosed by walls of 48 cm) with two identical objects (either two Falcon tubes filled with water or two stacks of large Legos; habituation objects were varied between mice). Mice were allowed to explore the identical objects until 20 s of cumulative exploration time was achieved. Mice were returned to their home cage while the arena/objects were cleaned with 70% ethanol. During the novel object recognition test session, the two training objects were replaced with one matched item from the training session and a novel item. Mice were placed in the arena and allowed to explore both the novel and learned object until 20 s of cumulative exploration time was achieved, with a maximal time of 10 min to reach the criteria; those that did not meet the criteria were excluded from statistical analyses. Time spent exploring each respective object was recorded. Mice were returned to their home cages and the arena/objects were cleaned with 70% ethanol. As previously described, exploration activity was defined as “directing the nose toward the object at a distance less than or equal to 2 cm” and time spent climbing on or chewing on objects was not deemed exploration activity. The discrimination index was calculated by dividing the novel object exploration time by the total amount of object exploration during the test.
At 16 weeks post AAV injection, mice were individually placed into cages (19.4 × 18.1 × 39.8 cm) with evenly spaced glass marbles arranged in a three-by-four grid on the surface of clean aspen bedding (5 cm in depth). An experimenter who was unaware of genotype/treatment scored the number of marbles buried after 30 min. Marbles were washed with mild detergent and water between each trial.
At 19 weeks post AAV injection, mice were placed in an apparatus consisting of three connected plexiglass chambers (18 × 41 × 20 cm each) with removable dividers between each chamber. Each test subject was individually placed in the center plexiglass chamber for 5 min of habituation. In the first phase—testing social affiliation—another, unfamiliar mouse was placed in either the right or left chamber in a small wire cage, while another wire cage remained empty in the opposite chamber. The wire cage restricted social or aggressive interactions between the two mice beyond nose contact. Chamber dividers were lifted after the habituation period to allow the test subject to move freely about all three chambers for a 10-min observation period. A second 10-min test—assessing novel social engagement—was performed immediately afterward, using the conspecific from the first test (now denoted as a familiar mouse) and a novel unfamiliar mouse in the opposite chamber. Between each trial, the arena was cleaned with Opticide to remove odor cues. Trials were video recorded. An observer unaware of genotype/treatment used open-source event-logging software, BORIS, to create a user-scored ethogram with timestamps and length calculations for behavioral activities (i.e., chamber entry, mouse investigation). Once the behavioral coding process was completed, observation data were exported and analyzed. The social preference index was calculated by dividing the amount of time spent with the novel mouse by the amount of time spent in the empty chamber during the first test phase. The social novelty index was calculated by dividing the amount of time spent with the novel unfamiliar mouse by the amount of time spent with the familiar mouse during the second test phase.
At 21 week post AAV injection, a small plastic cylinder generated from a cut syringe (Becton, Dickinson and Company, Franklin Lakes, NJ) was slipped over each mouse’s tail to prevent climbing motion and escape from the test; mice that circumvented this anti-climb apparatus were excluded. Mice were suspended in air individually by tape attached to a shelf (64-cm height with bubble wrap laid beneath for safety) for 6 min. Trials were video recorded An observer unaware of genotype/treatment used the open-source event-logging software, BORIS, to create a user-scored ethogram with timestamps and length calculations for behavioral activities (i.e., immobility, mobility). Once the behavioral coding process was completed, observation data were exported and analyzed.
At 23 weeks post AAV injection, mice were euthanized following a 4-h fast. Mice were anesthetized with 2.5% isoflurane (1.0 L/min) and then decapitated to collect trunk blood. Tissues to be used for mRNA and protein analyses were flash frozen on dry ice and stored at −80°C until further analysis. Hypothalamus was collected under a dissection microscope at sacrifice, with the left and right sides being collected separately to allow for RNA and protein isolation from each mouse.
Trunk blood was collected at euthanasia, clotted on ice, and centrifuged at 10,000 rpm for 10 min at 4°C. The serum component was collected and stored at −20°C until further analysis. R&D Systems ELISA kits were used to assay serum leptin (#DY498), insulin-like growth factor (#DY791), PAI-1 (#DY3828), and CCL2 (#DY479). Additional ELISAs were performed for insulin (ALPCO #80-INSMSU-E01), corticosterone (Enzo #ADI-900-097), GH (EMD Millipore #EZRMGH-45K), and adiponectin (ALPCO #47-ADPMS-E01). Caymen Chemical colorimetric assay kits were used to assay triglycerides (#10010303) and glucose (#10009582). Homeostatic model assessment for insulin resistance (HOMA-IR) index was calculated as [fasting serum glucose (mmol/L) × fasting serum insulin (pmol/L)/22.5] as described elsewhere.
Following tissue sonication, RNA was isolated using the RNeasy Mini kit (QIAGEN #74804) with RNase-free DNase treatment. cDNA was reverse transcribed using Taqman Reverse Transcription Reagents (Applied Biosystems #N8080234). qRT-PCR was completed on the StepOnePlus Real-Time PCR System using Power SYBR Green (Applied Biosystems #A25742) PCR Master Mix. Primer sequences are available in Table S1. We calibrated data to endogenous controls—Hprt1 for hypothalamus, Actinb for gWAT—and quantified the relative gene expression using the 2−ΔΔCT method.
Tissue samples were homogenized in RIPA buffer (Pierce #89901) containing 1× PhosSTOP (Roche #4906845001) and protease inhibitor cocktail III (Calbiochem #539134). Tissue lysates were separated by gradient gel (4%–20%, Mini-PROTEAN TGX, Bio-Rad #4561096) and then transferred to a nitrocellulose membrane (Bio-Rad #1620115). Blots were incubated overnight at 4°C with the following Cell Signaling Technology primary antibodies: beta actin #3700, 1:500; HA tag #3724, 1:500. Blots were rinsed and incubated with HRP-conjugated secondary antibodies (Bio-Rad, 1:3,000). Chemiluminescence signal was detected and visualized by Odyssey Fc imaging (LI-COR Biotechnology, Lincoln, NE).
Portions of fat depots were fixed in 10% neutral buffered formalin for 72 h and dehydrated in 70% ethanol prior to processing. Paraffin sections (6 μm) were created and an H&E stain was performed by the Comparative Pathology & Digital Imaging Shared Resource of The Ohio State University’s College of Veterinary Medicine. Adipose tissue sections were imaged at 20× magnification using an Olympus BX43 microscope with an Olympus SC30 color camera attachment and Olympus cellSens software. Downstream analysis was performed as previously described. For each field, adipocyte area was measured in FIJI using a semi-automated custom method based on the Adiposoft algorithm.93, 94, 95 Briefly, each image was split into H&E color channels by Color Deconvolution, then made binary by the Otsu thresholding algorithm. The images were then filtered with opening and median operators with the Morphological Filters plugin to sharpen cell boundaries. Automated particle subtraction and manual removal of extraneous particles were performed to reduce noise. A cutoff of 100 μm2 was used as the minimum size for an adipocyte. Six images were analyzed per animal. Adipocyte distribution curves show all sampled adipocytes within a given group. Average adipocyte sizes were calculated on a per-animal basis and then compared.
A subset of mice was anesthetized and transcardially perfused with phosphate-buffered saline (PBS), followed by 4% paraformaldehyde (PFA) (Sigma, St. Louis, MO) in PBS. Fixed brains were extracted and incubated overnight in 4% PFA on a rocker at 4°C. Brains were rinsed three times with PBS for 30 min before being submerged overnight in 30% sucrose with 0.03% sodium azide on a rocker at 4°C. The sucrose solution was replaced and left at 4°C for at least 3 days. Brains were embedded in optimal cutting temperature (O.C.T.) compound (Sakura Finetek, Torrance, CA), sectioned into 30-μm slices on a Thermo Scientific HM525NX cryostat (Waltham, MA), and placed on glass microscope slides. Fluorescence microscopy was performed on a Zeiss microscope (Thornwood, NY), and images were captured with Zeiss Zen software.
Data are expressed as means ± SEM. Microsoft Excel, IBM SPSS v.25, GraphPad Prism 9, and R v.4.1.3. software were used to analyze data. One-way ANOVAs with Tukey’s post hoc test were utilized for comparisons between three groups. Time course data (BWs, GTT, ITT, indirect calorimetry) were analyzed using a mixed ANOVA and area-under-the-curve calculations were performed where applicable. Normality was tested using the Shapiro-Wilk method. Outliers were determined and removed using the robust regression and outlier removal (ROUT) method.
The authors confirm that data supporting the findings and conclusion of this study are presented within the article and supplemental information. | true | true | true |
|
PMC9574740 | 36102610 | Keerthiraju E Ravichandran,Lars Kaduhr,Bozena Skupien‐Rabian,Ekaterina Shvetsova,Mikołaj Sokołowski,Ros´cisław Krutyhołowa,Dominika Kwasna,Cindy Brachmann,Sean Lin,Sebastian Guzman Perez,Piotr Wilk,Manuel Kösters,Przemysław Grudnik,Urszula Jankowska,Sebastian A Leidel,Raffael Schaffrath,Sebastian Glatt | E2 / E3 ‐independent ubiquitin‐like protein conjugation by Urm1 is directly coupled to cysteine persulfidation | 14-09-2022 | oxidative stress,persulfidation,sulfur transfer,ubiquitin‐like,Urm1,Post-translational Modifications & Proteolysis,Structural Biology | Abstract Post‐translational modifications by ubiquitin‐like proteins (UBLs) are essential for nearly all cellular processes. Ubiquitin‐related modifier 1 (Urm1) is a unique UBL, which plays a key role in tRNA anticodon thiolation as a sulfur carrier protein (SCP) and is linked to the noncanonical E1 enzyme Uba4 (ubiquitin‐like protein activator 4). While Urm1 has also been observed to conjugate to target proteins like other UBLs, the molecular mechanism of its attachment remains unknown. Here, we reconstitute the covalent attachment of thiocarboxylated Urm1 to various cellular target proteins in vitro, revealing that, unlike other known UBLs, this process is E2/E3‐independent and requires oxidative stress. Furthermore, we present the crystal structures of the peroxiredoxin Ahp1 before and after the covalent attachment of Urm1. Surprisingly, we show that urmylation is accompanied by the transfer of sulfur to cysteine residues in the target proteins, also known as cysteine persulfidation. Our results illustrate the role of the Uba4‐Urm1 system as a key evolutionary link between prokaryotic SCPs and the UBL modifications observed in modern eukaryotes. | E2 / E3 ‐independent ubiquitin‐like protein conjugation by Urm1 is directly coupled to cysteine persulfidation
Post‐translational modifications by ubiquitin‐like proteins (UBLs) are essential for nearly all cellular processes. Ubiquitin‐related modifier 1 (Urm1) is a unique UBL, which plays a key role in tRNA anticodon thiolation as a sulfur carrier protein (SCP) and is linked to the noncanonical E1 enzyme Uba4 (ubiquitin‐like protein activator 4). While Urm1 has also been observed to conjugate to target proteins like other UBLs, the molecular mechanism of its attachment remains unknown. Here, we reconstitute the covalent attachment of thiocarboxylated Urm1 to various cellular target proteins in vitro, revealing that, unlike other known UBLs, this process is E2/E3‐independent and requires oxidative stress. Furthermore, we present the crystal structures of the peroxiredoxin Ahp1 before and after the covalent attachment of Urm1. Surprisingly, we show that urmylation is accompanied by the transfer of sulfur to cysteine residues in the target proteins, also known as cysteine persulfidation. Our results illustrate the role of the Uba4‐Urm1 system as a key evolutionary link between prokaryotic SCPs and the UBL modifications observed in modern eukaryotes.
Ubiquitin‐related modifier 1 (Urm1) is an evolutionarily conserved member of the ubiquitin family that adopts the β‐grasp fold characteristic of all eukaryotic ubiquitin‐like proteins (UBLs; e.g., ubiquitin, SUMO, NEDD8, ATG12, UFM1; Bedford et al, 2011). Urm1 acts as a sulfur carrier protein (SCP) and is an essential member of the modification cascade that thiolates the C2 position of wobble uridine (s2U34) in eukaryotic tRNA anticodons via the Ncs2‐Ncs6 sulfur‐transferase complex (Nakai et al, 2008; Schlieker et al, 2008; Leidel et al, 2009; Noma et al, 2009). The attachment of a sulfur atom to the wobble base in the tRNA anticodon optimizes ribosomal dynamics during translational elongation (Nedialkova & Leidel, 2015; Ranjan & Rodnina, 2016) and promotes cellular responses to nutrient starvation (Laxman et al, 2013; Gupta et al, 2019; Bruch et al, 2020). The disruption of tRNA anticodon modifications can result in dramatic proteome perturbations and the onset of severe human diseases, including cancer and neurodegenerative pathologies (Rezgui et al, 2013; Nedialkova & Leidel, 2015; Schaffrath & Leidel, 2017; Close et al, 2018; Hawer et al, 2018). In addition to its role as SCP in tRNA thiolation, reports showed that Urm1 covalently attaches to target proteins in conditions of oxidative stress in different organisms, reminiscent of other UBLs (Furukawa et al, 2000; Goehring et al, 2003; Schlieker et al, 2008; Van der Veen et al, 2011; Jüdes et al, 2015; Khoshnood et al, 2017; Wang et al, 2019; Tan et al, 2022). The conjugation reaction of a UBL to its target protein starts with adenylation of its C‐terminus by ATP‐dependent E1 ubiquitin‐activating enzymes. Next, the activated C‐terminus of the UBL is relayed via active site cysteines of an E1‐, E2‐, and E3‐enzyme cascade (Passmore & Barford, 2004). Finally, the UBL is attached to a specific target protein via a covalent isopeptide bond formed between a highly conserved diglycine motif at its C‐terminus and a lysine side chain of the target protein (Hochstrasser, 2000; McDowell & Philpott, 2013; Stewart et al, 2016; Cappadocia & Lima, 2018). Similarly, the initial activation steps for SCPs start with adenylation of the C‐terminus by specialized E1‐like proteins. Although SCPs typically do not form thioesters, our recent work revealed the necessity of a thioester intermediate for the activation of Urm1 by its dedicated E1 enzyme, Uba4 (Termathe & Leidel, 2018; Pabis et al, 2020). Following the formation of a thioester between Uba4 and Urm1, the rhodanese domain of Uba4 transfers a persulfide group to the activated C‐terminus of Urm1 (Kaduhr et al, 2021). The resulting thiocarboxylated C‐terminus (COSH) is present in all prokaryotic and archaeal SCPs (Kessler, 2006) and is required for the thiolation of nucleic acids as well as in the biosynthesis of Molybdenum cofactor (Moco) and thiamine (Leimkühler, 2017; Shigi, 2018). However, amongst eukaryotic UBLs only Urm1 is known to carry this unique terminal modification (Urm1‐SH). Hence, Urm1 and Uba4 are located at a crucial evolutionary branchpoint between prokaryotic SCPs and eukaryotic UBLs since they combine molecular features otherwise exclusively present in either SCPs or UBLs (Xu et al, 2006; Pedrioli et al, 2008; Jüdes et al, 2015). As the role of Urm1‐SH for tRNA thiolation is well established and no E2 or E3 enzymes for Urm1 have been identified so far, the potential conjugation of Urm1 to proteins (also referred to as protein “urmylation”), analogous to other UBLs, has been questioned. Nonetheless, several mass spectrometry studies have identified proteins conjugated with Urm1 under specific environmental conditions (Schlieker et al, 2008; Van der Veen et al, 2011; Khoshnood et al, 2017). In vivo, the most robust evidence for conjugation of Urm1 was observed for yeast Ahp1 (Goehring et al, 2003; Van der Veen et al, 2011; Jüdes et al, 2015; Brachmann et al, 2020), a peroxiredoxin that acts as an antioxidant enzyme in scavenging reactive oxygen species (Lee et al, 1999). However, the biochemical details including which E2 and E3 enzymes are required for the conjugation reaction to Ahp1, or other targets remain undefined. Furthermore, as the unique COSH moiety at the C‐terminus of Urm1 is absent from all other eukaryotic UBLs, neither its functional role nor its fate after the conjugation reaction is clear. Finally, the necessity of oxidative stress and a redox‐active (also called peroxidatic) cysteine for the conjugation of Ahp1 (Brachmann et al, 2020) raised the question of whether a preferred site of conjugation exists in other proteins and whether the reaction requirements would be similar. We sought to close these gaps by determining the mechanism of Urm1 conjugation to target proteins. We show that purified Urm1‐SH can be efficiently attached to Ahp1 and other target proteins in vitro using mild oxidative stress conditions in the absence of E2 enzymes or E3 ligases. We demonstrate that the conjugation depends on the thiocarboxylated C‐terminus of Urm1 and a redox‐active cysteine in the target protein, and that these covalent linkages can occur on a variety of lysine, serine, and threonine residues. Our high‐resolution crystal structures of Ahp1 before and after the conjugation reaction confirm our biochemical findings and provide in‐depth insights into the molecular mechanism of Urm1 attachment. Strikingly, we observed that Urm1‐SH transfers its sulfur moiety to the urmylated target protein, resulting in directed persulfidation of specific redox‐active cysteines. Persulfidation of cysteines (also called protein S‐sulfhydration) has been directly linked to aging and may be regulated by intracellular H2S levels (Zivanovic et al, 2019). However, the pathway responsible for targeted cysteine persulfidation has remained elusive. Oxidation of specific cysteines triggers the process of Urm1 conjugation, providing a molecular mechanism that enables cells to protect vulnerable cysteine residues from reactive oxygen species (ROS) by a highly selective process. Finally, we used our mechanistic insights to engineer artificially thiocarboxylated proteins and recapitulate the conjugation reaction on a variety of model proteins, including other UBLs and GFP, providing a platform to covalently modify proteins of interest with proteinaceous tags through this selective reaction. Our findings demonstrate how Urm1 protects proteins during oxidative stress and reveal a critical evolutionary link between sulfur transfer and covalent modification by ubiquitin family proteins.
To study the function of Urm1 and understand the details of its protein conjugation reaction, we produced large quantities of purified Urm1‐SH using two methods. In vivo, the cysteine desulfurase Nfs1 and the sulfur‐transferase Tum1 relay sulfur to the rhodanese domain of Uba4, which transfers an activated sulfur group to the C‐terminus of Urm1 (Fig 1A; Noma et al, 2009). The use of thiosulfate as a sulfur source for Uba4 makes it possible to bypass Nfs1 and Tum1, thereby catalyzing the direct thiocarboxylation of Urm1 in vitro (Termathe & Leidel, 2018). We used this approach to produce Urm1‐SH from purified Saccharomyces cerevisiae Urm1 and Uba4 proteins (ScUrm1 and ScUba4). As an alternative strategy, we generated Urm1‐SH or Urm1‐OH from Homo sapiens and Chaetomium thermophilum, a thermophilic fungus harboring highly stable proteins (Bock et al, 2014), using a chitin‐based inducible cleavage (intein) system (Kinsland et al, 1998; Fig EV1A). We verified all purified proteins by mass spectrometry and independently confirmed the presence of a sulfur atom in Urm1‐SH by its increased gel retardation in PAGE supplemented with [(N‐acryloylamino)phenyl]mercuric chloride (APM) (Fig 1B; Igloi, 1988). Of note, the sequences of most eukaryotic Urm1 proteins do not contain cysteine residues, which would affect their behavior in the used APM gels. We removed the nonconserved surface cysteine (Cys55) in CtUrm1 to circumvent super‐shifting in the APM gels and facilitate the analyses. Positioning a cysteine at the C‐terminus of Urm1 (CtUrm1G111C) also led to a retardation of the purified protein in the APM gel (Fig EV1A). The well‐studied in vivo target Ahp1 is a 2‐Cys peroxiredoxin that consists of a resolving cysteine (CR) and a peroxidatic cysteine (CP; Lian et al, 2012). CP reacts with reactive oxygen species to form sulfenic acid and the CR from the other monomer reacts with the sulfenylated CP forming a disulfide bridge. In vivo, Urm1 conjugation requires the presence of oxidative stressors and peroxidatic cysteine (Cys62) in ScAhp1 (Van der Veen et al, 2011; Brachmann et al, 2020). Surprisingly, we observed that mutating the resolving cysteine in Ahp1 (C31S) not only permitted the attachment of Urm1 in vivo but even promoted the conjugation reaction in absence of N‐Ethylmaleimide (NEM), a compound that irreversibly alkylates cysteines and is generally required to detect ubiquitin‐like conjugates in cell lysates (Fig 1C). Hence, ScAhp1C31S represented a very promising tool to recapitulate the conjugation reaction in vitro. First, we tested purified ScUrm1‐OH and ScUrm1‐SH in combination with various purified variants of ScAhp1 (ScAhp1C31S, ScAhp1C31R, ScAhp1C62S, and ScAhp1C31S C62S; Fig 1C). We were not able to detect Urm1 conjugation to the Ahp1 substrate in vitro by incubating wild‐type ScAhp1 with ScUrm1‐OH in the presence of the oxidizing agent tert‐butyl hydroperoxide (TBH) previously shown to stimulate the reaction in vivo (Van der Veen et al, 2011; Brachmann et al, 2020). Using wild‐type ScAhp1 with ScUrm1‐SH led to the formation of Urm1 conjugates, but their levels remained close to the detection limit. However, ScAhp1C31S and ScAhp1C31R very efficiently formed specific conjugates with Urm1‐SH in the presence of TBH, resulting in an increase in the molecular weight of Ahp1 roughly by the mass of one Urm1 molecule (Figs 1D and EV1B). Of note, the conjugates formed in presence of Urm1‐SH prepared by either thiocarboxylation protocol but not with Urm1‐OH or in the absence of TBH (Fig 1D). To further characterize the conjugates, we treated the samples after the reaction with either Tris(2‐carboxyethyl)phosphine (TCEP) to reduce disulfide bonds, hydroxylamine (HA) to reduce thioester bonds, or DTT to reduce both (Termathe & Leidel, 2018). None of these reagents disrupted the conjugates, suggesting that a covalent isopeptide bond is formed between Urm1‐SH and ScAhp1C31S (Fig 1D). In agreement with previous in vivo data (Brachmann et al, 2020), the mutation of the peroxidatic cysteine (Cys62) in Ahp1 (ScAhp1C31S C62S) completely abolishes Urm1 attachment in vitro (Fig 1D). In addition, the Urm1‐conjugation efficiency of mutated monomeric Ahp1 is strongly reduced in vitro (Appendix Fig S1) and almost completely abolished in vivo (Brachmann et al, 2020). As different oxidative stressors can induce Urm1 conjugation in vivo (Van der Veen et al, 2011; Jüdes et al, 2015), we tested whether “urmylation” can similarly be triggered by different oxidative compounds in vitro. Indeed, conjugate formation likewise occurred in the presence of hydrogen peroxide (HP), peroxynitrite, and diamide. On the contrary, methylglyoxal, di‐tert‐butyl disulfide, and NEM did not promote the conjugation reaction but also did not influence the stability of Urm1 conjugation if added after the reaction (Fig EV1C). None of the reagents compromised ScAhp1 integrity (Fig EV1D), and only the alkylating agents NEM and peroxynitrite negatively affected the stability of the thiocarboxyl group at the C‐terminus of Urm1 (Fig EV1E). This observation suggests that NEM does not promote the urmylation of Ahp1 in vivo but rather facilitates its detection by blocking potential inhibitory factor(s) or deconjugating enzyme(s). In addition, we used CtUrm1 and variants of CtAhp1 to independently confirm the necessity of the C‐terminal thiocarboxyl group, oxidative stress conditions, and redox‐active cysteines for the reaction to occur (Fig EV1F). In summary, our results demonstrate that Urm1, like all other canonical eukaryotic UBLs, can form a covalent bond with a target protein. However, Urm1 conjugation does not depend on the canonical cascade of E2 ubiquitin‐conjugating enzymes and/or E3 ligases. Instead, the conjugation reaction requires a thiocarboxylated C‐terminus of Urm1 and a peroxidatic cysteine in the target protein, which in the case of Ahp1 appears to be promoted by site‐specific substitution of the resolving cysteine.
To gain detailed molecular insights into the mechanisms of Urm1 conjugation, we purified and determined the structures of urmylated and unmodified CtAhp1 by macromolecular crystallography. The purified CtAhp1C30S‐Urm1C55S complex showed a clear shift in size in comparison to the individually purified Ahp1 and Urm1 proteins (Fig 2A). Of note, CtAhp1C30S also formed conjugates with wild‐type CtUrm1, showing that Cys55 is not required for the conjugation reaction in Chaetomium thermophilum (Fig EV1G). We independently crystallized CtAhp1 and CtAhp1C30S, which is equivalent to ScAhp1C31S, and collected complete datasets at 1.75 and 1.85 Å resolution, respectively (Table 1). We solved both structures by molecular replacement using ScAhp1 structure (PDB ID 4DSR; Lian et al, 2012) and refined the corresponding atomic models to R/Rfree values of 18.8%/20.2% and 15.7%/18.3%, respectively (Table 1, PDB ID 7Q68 and 7Q69). The structures confirm that CtAhp1, like ScAhp1, forms a homodimer, which is maintained by conserved hydrophobic residues (e.g., Phe56) at the center of the dimer interface (Fig 2B; Trivelli et al, 2003). Wild‐type CtAhp1 exists as a homodimer and forms a disulfide bond between Cys30 and Cys60 of opposite monomers, which indicates that in the absence of thioredoxins the dimer becomes trapped in the (post)oxidized state after the release of H2O. CtAhp1C30S cannot form the disulfide bond between the two redox‐active cysteines, resulting in a structural rearrangement of the entire loop region harboring the peroxidatic cysteine (Cys60). As the mutated residue (Ser30) remains in an almost identical position as in the wild‐type protein, CtAhp1C30S structurally mimics the reduced form that can accumulate oxidized/sulfenylated Cys60 (Fig 2B and Appendix Fig S2). Due to the presence of a reducing agent in the crystallization condition, there were no indications for the oxidation of the thiol group of Cys60. Next, we obtained crystals of the CtAhp1C30S‐Urm1C55S complex, which after several rounds of optimization, diffracted to an overall resolution of 2.5 Å (Table 1). We solved the structure by molecular replacement using the high‐resolution model of CtAhp1C30S in combination with the model of CtUrm1 taken from the Uba4C202K–Urm1 complex (PDB ID 6YUC; Pabis et al, 2020). The final model was refined to R/Rfree values of 23.4%/27.2% obeying perfect stereochemistry (Table 1, PDB ID 7Q5N). The structure shows an almost unchanged homodimer of CtAhp1C30S, where the Lys63 residue from each Ahp1 molecule forms a covalent isopeptide bond between the lysine side chain and the C‐terminus of Urm1 (Fig 2C). The overall dimeric structure of Ahp1 and the position of its peroxidatic cysteines remain almost identical in the presence or absence of Urm1 (Fig 2C). The asymmetric unit contains three individual CtAhp1C30S‐Urm1C55S dimers, which are all linked via the same peptide bond between Ahp1Lys63 and Urm1Gly111 (Appendix Fig S3A). The conformations of the three urmylated Ahp1 homodimers are superimposable, but the relative position of the respective Urm1 molecules, the quality of the density map, and the associated B‐factors differ between the individual copies (Fig 2C and Appendix Fig S3B). These variations arise most likely due to different packing environments in the crystal lattice and might result from the fact that Ahp1 and the attached Urm1 do not form an extensive interaction surface. However, there are no indications that a hexameric form of CtAhp1C30S‐Urm1C55S (Appendix Fig S3C) or dimerization of Urm1 between neighboring complexes (Appendix Fig S3D) occurs in solution or has any functional significance in vivo. Of note, the Urm1‐Urm1 packing interfaces are almost identical, and changes in relative positioning are not caused by structural changes of the individual Ahp1 and Urm1 proteins but are strictly related to variations in the flexible C‐terminal region of Urm1 (Appendix Fig S3E). Unexpectedly, we found three distinct types of Urm1 positioning relative to the attachment site around Ahp1Lys63 (Fig 2D). These differences are not only visible at the domain level but also the specific conformation of interacting residues in Urm1 and Ahp1 varies between the observed attachment sites. The structure of ScAhp1 bound to its recycling factor thioredoxin 2 (Trx2) has been determined in different reaction states (Lian et al, 2020, 2012), showing that Trx2 directly interacts with Cys30 of Ahp1 to resolve the disulfide bond formed between the redox‐active cysteines (Fig EV2A). By contrast, our urmylated Ahp1 structure places Urm1 close to the peroxidatic Cys62, indicating that certain conformations of Urm1 attachment are compatible with the binding of Trx2 while others are sterically impossible (Fig EV2B). Our crystallographic data confirm the formation of a covalent isopeptide bond between Ahp1 and Urm1. Furthermore, we identify Lys63 as one of the main attachment sites in CtAhp1 and reveal the structural variability of the conjugated complex, which is linked closely to the peroxidatic cysteine.
Although the crystal structure of CtAhp1C30S‐Urm1C55S displays an exclusive linkage with unambiguous density between CtUrm1Gly111 and CtAhp1Lys63 (Fig 3A), we previously found that Urm1 can be conjugated to different lysine residues of ScAhp1 in vivo (Brachmann et al, 2020). Hence, we analyzed the products of the in vitro conjugation reactions by mass spectrometry to independently confirm the presence of HGG‐K linked peptides after digestion with chymotrypsin (Fig 3B). Since ubiquitin is known to form poly‐ubiquitin chains on target proteins, we asked whether we could detect this kind of polymerization for Urm1 conjugates. In our dataset, the discrete molecular weights of the conjugates indicated that internal conjugation events did not occur between Urm1 molecules under these conditions. In contrast to the exclusive Lys63‐mediated linkage in our crystals, we identified additional lysine residues of CtAhp1 (Lys44, Lys99, Lys141, Lys156, and Lys171) conjugated to the C‐terminus of CtUrm1. Furthermore, we used a CtAhp1C30S K63R mutant, lacking the primarily conjugated residue Lys63, to promote the occurrence of rare conjugation events and detected urmylation of Lys71, which is in close proximity to Cys60. Similarly, in ScAhp1, which harbors more lysine residues than CtAhp1 (Fig 3B), we detected various additional attachment sites by mass spectrometry (Lys32, Lys47/Lys48, Lys79, Lys81, Lys107, Lys124, and Lys156). Next, we mutated various lysine residues in CtAhp1 and ScAhp1 individually and performed in vitro conjugation reactions using these purified Ahp1 variants. Our results clearly demonstrated that even after simultaneously mutating the major attachment sites in CtAhp1 (CtAhp1K44R K63R K99R), residual conjugates were still detectable (Fig 3C). In ScAhp1, only the simultaneous mutation of several lysine residues (ScAhp1K32R,K47R,K48R,K124R,K156R) slightly reduced urmylation levels (Fig 3D). Last, we tested the effects of these lysine mutants in S. cerevisiae and did not detect any preferred modification site in vivo. After mutating different lysine residues, we neither observed any phenotypic consequences in the presence of TBH (Fig 3E) nor the complete inhibition of conjugation (Fig 3F). Of note, the linked lysine residues in CtAhp1 and ScAhp1 are not evolutionarily conserved, and their spatial distribution does not reveal a specific distribution pattern (Fig 3G). However, our complementary mass spectrometry and mutational analyses show that the attachment of Urm1 to lysine residues in Ahp1 exclusively depends on the presence of a peroxidatic cysteine in the target protein. Our results in ScAhp1 emphasize that in the absence of a particular lysine residue a different residue will be used as a target with similar efficiency, indicating low specificity for the particular residue, motif, or structural context at which urmylation can occur.
As noncanonical linkages to serine, threonine or cysteine residues have also been described for other UBLs (McDowell & Philpott, 2013), we analyzed our mass spectrometry data for any such possible Urm1 linkage on Ahp1 (ScAhp1C31S and CtAhp1C30S). Strikingly, we found Urm1 being conjugated to several serine and threonine residues in ScAhp1C31S and few in CtAhp1C30S (Fig 3B). To test whether lysine residues are essential for catalyzing the conjugation reaction, we simultaneously replaced all lysine residues in ScAhp1C31S and CtAhp1C30S with arginine residues (ScAhp1C31S‐KtoR and CtAhp1C30S‐KtoR). Nonetheless, we still observed the formation of Ahp1‐Urm1 conjugates with these lysine‐less Ahp1 variants in the presence of Urm1‐SH and mild oxidative stress in vitro (Fig 4A). We mutated semi‐conserved serine and threonine residues in the spatial proximity of the peroxidatic cysteine and other nearby residues in CtAhp1 (Arg151) or CtUrm1 (Leu108, His109; Fig 4B). Whereas none of the mutations affected the conjugation efficiency in the presence of lysine residues, the simultaneous depletion of all possible serine and threonine attachment sites that are in proximity to the peroxidatic cysteines in ScAhp1C31S‐KtoR and CtAhp1C30S‐KtoR strongly reduced the formation of Ahp1‐Urm1 conjugates in vitro (Fig 4C and Appendix Fig S4A). Therefore, these serine and threonine residues can act as alternative attachment sites but play no significant role during the canonical reaction when lysines are present. Since the lysine‐less variants still showed traces of conjugates, we used mass spectrometry to unambiguously test for other low‐frequency attachment sites. This approach allowed us to confirm noncanonical ester linkages between ScUrm1 and Ser59 and Thr61 sites on ScAhp1 and between CtUrm1 and Thr57 and Thr59 on CtAhp1 (Fig 4D). In summary, our detailed mutational analyses indicate that Urm1‐SH preferably conjugates to lysine residues but is also able to attach to other amino acid side chains as has been shown for other UBLs.
To the best of our knowledge, Urm1 is the only eukaryotic UBL that carries a C‐terminal thiocarboxyl group, which we showed is essential for the conjugation reaction. Intriguingly, this unique sulfur atom was undetectable at the site of linkage in our mass spectrometry analyses above. Hence, we were curious about the fate of the attached sulfur moiety during and after isopeptide bond formation. To address this question, we generated Urm1‐SH carrying radioactive sulfur (35S) at its C‐terminus (see Materials and Methods and Appendix Fig S4B) for in vitro “urmylation” reactions (Appendix Fig S4C). We incubated 35S‐labeled thiocarboxylated ScUrm1 and CtUrm1C55S with various variants of ScAhp1 and CtAhp1 in the presence or absence of TBH (Fig 4E and Appendix Fig S4D–G). Strikingly, we detected a very efficient transfer of 35S from Urm1‐35SH to ScAhp1 and CtAhp1 variants (ScAhp1C31S, ScAhp1C31S‐KtoR, CtAhp1C30S and CtAhp1C30S‐KtoR). The mutation of the peroxidatic cysteine not only diminished the Urm1 conjugation reaction but also blocked the sulfur transfer between Urm1‐SH and Ahp1 (ScAhp1C62S, CtAhp1C60S, and CtAhp1C30S C60S). In the case of ScAhp1C31S C62S and ScAhp1C31S‐KtoR, we still observed a background of low‐level sulfur transfer, which appears to be conjugation‐independent and could be mediated by an additional cysteine residue in ScAhp1 (Cys120). Neither the unspecific reaction with Ahp1 nor the additional cysteine residue is present in CtAhp1. Interestingly, the 35S transfer to Ahp1 cannot only occur on the Urm1‐conjugated molecule of the Ahp1 homodimer but also on the other unconjugated chain of the homodimer. Next, we searched our mass spectrometry datasets of the complexes for possible forms of peptide S‐sulfhydration. Strikingly, we identified persulfidation of the catalytically active peroxidatic cysteine in both ScAhp1 and CtAhp1 (Fig 4F), indicating that Urm1‐SH transfers its sulfur atom to the sulfenylated cysteine on its target protein. The archaeal TfuA protein, which is also C‐terminally thiocarboxylated, was recently shown to replace oxygen with sulfur in the peptide backbone of its target protein (Liu et al, 2021). We reanalyzed our crystallographic data to identify the precise spatial position of the post‐translational modification. A prominent density next to the sulfur atom of the cysteine side chain was identifiable in the crystal structure of the CtAhp1C30S‐Urm1C55S complex, which was absent from individual Ahp1 structures alone (Fig 4G and Appendix Fig S5). To confirm the position and identity of the additional sulfur moiety, we collected diffraction data for the CtAhp1C30S‐Urm1C55S crystals at suitable wavelengths and calculated anomalous difference Fourier maps (Fig 4H). Foremost, we did not observe an anomalous signal at the position of the backbone oxygen atom, excluding the possibility of a sulfur substitution reaction, like TfuA. Furthermore, we obtained a structure at 1.1 Å resolution of the nonreacted CtAhp1C30S, which we repurified after the conjugation reaction with Urm1‐SH (Table 1, PDB ID 7Q6A). Although this Ahp1 sample underwent an identical treatment, we did not observe any extra densities around Cys60 without Urm1 conjugation (unreacted Ahp1; Fig 4G). Our complementary biochemical, mass spectrometry, and crystallographic analyses show that the detected persulfide group is absent in the starting material and that the post‐translational thio‐modification of the cysteine side chain in Ahp1 occurred during the conjugation reaction with Urm1‐SH. We propose a very simple reaction mechanism that directly couples urmylation with the observed sulfur transfer to the target protein (Fig 4I). In detail, the exposure of a redox‐active cysteine in Ahp1 to oxidative stress leads to the sulfenylation of its side chain (‐SOH). Urm1‐SH recognizes the sulfenylated cysteine and after condensation forms a short‐lived acyl disulfide intermediate (Kang et al, 2018), which we were not able to directly observe in our experiments due to its transient character. The reduction in the transient acyl disulfide intermediate can lead to the recycling of Urm1‐SH. Otherwise, a lysine, serine, or threonine residue at a proximal distance from the linked Cys‐Cys pair triggers a nucleophilic attack on the acyl disulfide. This reaction scheme simultaneously produces a persulfidated cysteine and a covalent conjugation site between the C‐terminus of Urm1 and respective lysine, serine, or threonine residue in the target protein (Fig 4I). Our results show that the Urm1‐conjugation reaction is mechanistically and directly coupled to the process of cysteine persulfidation. Foremost, our results demonstrate that Urm1 can specifically transfer the sulfur atom of its thiocarboxyl group to redox‐active cysteine residues in Ahp1 if the target is exposed to oxidative conditions.
It is important to emphasize that Ahp1, an antioxidant enzyme, represents an unusual target of Urm1 under oxidative stress conditions. For instance, even under optimized experimental conditions, we detected only weak conjugation activity of wild‐type Ahp1 in vitro and we had to use ScAhp1C31S or CtAhp1C30S to detect substantial conjugation activity. To identify and characterize additional Urm1 targets in vitro, we used a list of 547 candidate proteins that had been identified in yeast by LC–MS/MS after Urm1 pull‐down. Twenty one of the hits were significantly enriched upon NEM treatment, recapitulating the behavior of Ahp1 (Fig 5A). We decided to focus on a shortlist of potential Urm1 targets from yeast and humans based on available purification protocols. Hence, we used Tdh3, Ses1, and Pyk1 (Weygand‐Durasevic et al, 1987; Jurica et al, 1998; Liu et al, 2012) as well as their human homologs (PRDX5 (human homolog of Ahp1), GAPDH (human homolog of Tdh3), SARS1 (human homolog of Ses1) and PKM2 (human homolog of Pyk1)). Strikingly, we were able to detect Urm1 conjugates for the entire set of yeast and human targets in vitro using either ScUrm1‐SH or HsUrm1‐SH, respectively (Fig 5B). Formation of these conjugates required the addition of an oxidizing agent (Fig 5B) and only occurred in the presence of Urm1‐SH but not Urm1‐OH (Appendix Fig S6A). Unlike Ahp1, these targets showed robust Urm1 conjugation without introducing additional mutations or removing potential resolving cysteines (Appendix Fig S6B). To confirm the formation of covalent peptide bonds, we performed exhaustive mass spectrometry analyses to identify numerous HGG‐K, HGG‐T and HGG‐S linked peptides and map the individually conjugated residues (Fig 5C, and Appendix Fig S6C and D). As the conjugation of Urm1 to Ahp1 requires at least one peroxidatic cysteine, we tested whether known catalytic cysteines are necessary for the conjugation reaction to the additional target proteins. We observed a complete loss of Urm1 conjugation to HsPRDX5, ScTdh3, and HsGAPDH when a single peroxidatic cysteine was mutated (Fig 5D). The mutation of individual cysteine residues in the other targets only partially reduced the Urm1 conjugation efficiency, indicating that several cysteines can support the Urm1 conjugation reaction in these targets (Appendix Fig S6B). Furthermore, we used Urm1‐35SH for in vitro conjugation of yeast targets and observed the direct transfer of sulfur during Urm1 conjugation onto ScTdh3, ScSes1, and ScPyk1 in the presence of diamide (Fig 5E). In addition, we used mass spectrometry to verify that the sulfur atom from Urm1‐SH was transferred, leading to persulfidation of individual cysteine residues in HsGAPDH, ScTdh3 and HsPRDX5 (Fig 5C). Consistent with the observation that mutation of individual cysteines did not abolish Urm1 conjugation, we found that indeed more than one peroxidatic cysteine was persulfidated in ScSes1, HsSARS1, ScPyk1 and HsPKM2 (Fig 5C). Our data thus uncover the previously unknown source of the sulfur donor for targeted persulfidation of eukaryotic proteins.
We established that Urm1 can be covalently conjugated to various target proteins in vitro without the need for specific E2 enzymes and E3 ligases. The intein system that we used to generate Urm1‐SH above enables us to produce virtually any protein with a thiocarboxylated C‐terminus. Therefore, we tested whether other UBLs that harbor the C‐terminal GlyGly‐motif can also be conjugated to the identified Urm1 target proteins in response to oxidative stress. Using purified ubiquitin‐SH or SUMO‐SH (Fig 6A), we were able to attach these synthetically thiocarboxylated UBLs to ScAhp1 and ScTdh3 under the same experimental conditions that worked for Urm1‐SH (Fig 6A). As expected, ubiquitin‐OH and SUMO‐OH that lack the thiocarboxyl group did not form conjugates with the target proteins. We used mass spectrometry to map all conjugation sites. Interestingly, the pattern of conjugation sites differs between the different thiocarboxylated UBLs (Fig 6B). Nonetheless, we confirmed the identity of the peptides linked to the C‐termini and found that the sulfur is similarly deposited as a persulfide on the same oxidizable cysteines in the target proteins (Fig 6B). Next, we tested whether the highly conserved C‐terminal GG motif of UBLs is required for the conjugation reaction. We replaced the terminal Gly99 of ScUrm1 with all other canonical amino acids and produced the thiocarboxylated versions via the intein system. Strikingly, the presence of nearly any amino acid at the C‐terminus permitted the conjugation reaction, as long as the protein remained thiocarboxylated (Fig 6C). ScUrm1G99C was not thiocarboxylated by the intein system, since it undergoes autocleavage before the final cleavage and elution step could be induced with ammonium sulfide. We also observed reduced conjugation efficiencies for ScUrm1G99E and ScUrm1G99H for reasons that remain unclear. When removing both C‐terminal glycine residues (ScUrm1HΔGG), histidine (His97) is located at the C‐terminus, which also led to strongly reduced conjugation efficiency. However, when histidine is replaced by alanine (ScUrm1H97AΔGG), or when both glycine were mutated to alanine or serine (ScUrm1G98A,G99A, ScUrm1G98S,G99S), Urm1 did conjugate even without the characteristic GG‐motif. In summary, our data show that Urm1 variants carrying any thiocarboxylated amino acid at their C‐termini can undergo a covalent attachment to defined target proteins in vitro. Next, we tested whether the attachment requires the typical β‐grasp fold of UBLs or whether an unrelated thiocarboxylated model protein could be attached to Urm1 targets in vitro. We generated thiocarboxylated versions of GFP (GFP‐SH) and a fusion protein between GFP and ScUrm1 (GFP‐Urm1‐SH). Both artificially thiocarboxylated proteins can be conjugated to ScAhp1C31S or ScPyk1 (Fig EV3A). As GFP‐SH contains a C‐terminal lysine residue, we verified the formation of a conjugated XK‐K peptide (Fig EV3B) by mass spectrometry. The results of our in vitro studies suggest that the highly conserved C‐terminus of Urm1 is functionally not required for conjugation. However, it is important to highlight that by using the intein system to produce thiocarboxylated proteins, we can bypass the naturally occurring reaction catalyzed by the ATP‐dependent E1 activating enzyme Uba4. Our recent work showed that the C‐terminus of Urm1 is positioned deep inside the active site of Uba4 and that it most likely remains buried throughout the entire process of thiocarboxylation (Fig EV3D; Pabis et al, 2020). We performed an ATP hydrolysis assay to test whether different Urm1 variants and model proteins are able to activate Uba4 in vitro (Fig EV3C). Strikingly, only Urm1 and the GFP‐Urm1 fusion protein induced the ATPase activity of Uba4. Any variation in the C‐terminus of Urm1 (ScUrm1HGΔG, ScUrm1HΔGG, ScUrm1HG99W, ScUrm1HAA, ScUrm1HSS) failed to activate Uba4 (Fig EV3C). Furthermore, the use of ubiquitin, SUMO, GFP, or a GFP variant carrying the six most C‐terminal residues of Urm1 (GFP‐U6) did not induce a detectable Uba4 response (Fig EV3C). Finally, we performed Uba4‐mediated thiocarboxylation of GFP‐Urm1 in vitro and subsequently confirmed its direct conjugation to ScAhp1, ScTdh3, and ScSes1 (Fig EV3E). In summary, our findings highlight the broad reactivity of various thiocarboxylated proteins with Urm1 target proteins in vitro under mild oxidative stress conditions. Even if all other UBLs are highly similar and carry the same C‐terminal GlyGly‐motif, the high specificity, and selectivity between Urm1 and Uba4 strongly suggest that Uba4 exclusively thiocarboxylates Urm1 in vivo, thereby ensuring that Urm1 is unique in its ability to covalently attach to target proteins independent of E2 or E3 enzymes and simultaneously catalyze persulfidation of its target.
Hitherto, the gaseous H2S molecule was thought to constitute the major cellular response signal to initiate the protection of cysteines by persulfidation upon oxidative stress (Kimura & Kimura, 2004; Mustafa et al, 2009; Zivanovic et al, 2019). Cys3, Cys4, and Tum1 are key components of the trans‐sulfuration pathway in yeast and were shown to regulate intracellular H2S levels. After obtaining experimental evidence for the existence of Urm1‐SH mediated persulfidation of cysteine residues in target proteins, we asked whether the previously known pathway is directly or indirectly linked to Urm1 and its role in tRNA and protein thiolation. We generated a panel of yeast strains carrying individual deletions of trans‐sulfuration or urmylation pathway components (ahp1Δ, urm1Δ, tum1Δ, ncs6Δ, cys3Δ, and cys4Δ) or double deletions of a trans‐sulfuration pathway component and URM1 (cys3Δurm1Δ and cys4Δurm1Δ). We found that only the deletion of AHP1 leads to a measurable growth defect in response to the presence of TBH (Fig 7A). The deletion of URM1 did not affect this response, confirming that Urm1 conjugation is not essential for the function of Ahp1 (Van der Veen et al, 2011; Brachmann et al, 2020). We also observed a modest growth defect for cys3Δ and cys4Δ mutants, in contrast to deletions of URM1 or known components of the tRNA thiolation pathway, which grew normally. To test whether the trans‐sulfuration pathway is linked to Urm1 activity and whether the Uba4/Urm1 pathway regulates H2S levels independently, we analyzed the generation of H2S in the different deletion strains. As expected, the cys3Δ and cys4Δ strains displayed dramatically altered H2S levels. However, the deletion of URM1, TUM1, or NCS6 had no detectable impact on intracellular H2S levels (Fig 7B). The cys3Δurm1Δ and cys4Δurm1Δ double mutants display phenotypes similar to cys3Δ and cys4Δ strains alone. Our data thus indicate that Urm1 does not regulate H2S directly or via Cys4/Cys3. Next, we analyzed in vivo conjugation rates of Urm1 to Ahp1 and tRNA thiolation levels in the various deletion strains. Importantly, loss of Tum1, but not Ncs6, strongly reduces Urm1 conjugation to Ahp1 in vivo (Fig 7C). Therefore, Urm1 conjugation and protein persulfidation do not require tRNA thiolation. Consistent with previous studies, the elimination of components of the tRNA thiolation cascade, like Urm1 or Ncs6, completely abolishes the formation of mcm5s2U34 in tRNAs from these strains (Fig 7D). By contrast, the inactivation of Cys3 or Cys4 leads to a reduction of tRNA modification levels, while the AHP1 knockout has no effect. The reduction in tRNA thiolation levels could be affected by the availability of intracellular sulfur sources that are regulated by Cys4 and Cys3 in the cysteine biosynthesis pathway. Nonetheless, the inactivation of Cys3 or Cys4 does not affect the Urm1 conjugation levels of Ahp1, indicating that Urm1‐SH is formed at similar levels. Foremost, we show that persulfidation of cysteines by Urm1 is independent of the previously established trans‐sulfuration pathway. Finally, our results indicate that protein conjugation and cysteine persulfidation by Urm1 are also independent of its role in tRNA thiolation (Fig 7E). Altogether, we reveal an unexpected function of Urm1, which appears to act as a universal SCP for the regulation of amino acid persulfidation and tRNA base thiolation.
The key importance of post‐translational modification by UBLs for molecular processes and cellular functions is undisputed. Numerous novel UBLs and noncanonical mechanisms of ubiquitination have recently been discovered (Bhogaraju et al, 2016, 2019; Qiu et al, 2016; Kalayil et al, 2018). While some bacterial and archaeal SCPs were shown to be conjugated to substrate proteins in vivo (Shigi, 2012; Hepowit et al, 2016; Xu et al, 2019), the mechanisms for conjugation remained unknown. Our results reveal a novel and surprising link between ubiquitin‐like conjugation of Urm1, a highly conserved eukaryotic SCP, to target proteins and post‐translational persulfidation of specific cysteine residues in these targets. Several molecular characteristics indicate that the Urm1‐Uba4 system may represent an ancient precursor of ubiquitin‐like conjugation reactions in eukaryotes (Iyer et al, 2006; Burroughs et al, 2009; Schulman & Harper, 2009; Eme et al, 2017; Imachi et al, 2020). Urm1, with its thiocarboxylated C‐terminus and its bifunctional role as SCP and UBL, takes a unique position amongst all members of the eukaryotic UBL protein family. Our work establishes that Urm1‐SH is essential to catalyze the persulfidation of cysteines, while the lysine/serine/threonine‐directed attachment of UBLs in eukaryotes may have evolved as a side reaction. For instance, we found that the covalent attachment of Urm1 depends on the sulfur group that is simultaneously deposited at a redox‐active cysteine residue during the conjugation, highlighting the central role of the sulfur transfer during a highly coordinated reaction. Furthermore, the target motif only requires the modifiable cysteine residue and its preceding oxidation without the need for E2 enzymes or E3 ligases to direct Urm1 towards attachment sites and convey specificity. The observed high reactivity and broad specificity of the thiocarboxyl group might thus have been selected against the evolution of Urm1‐specific E2 and E3 enzymes and possibly, promoted the loss of the thiocarboxyl group in other UBLs. The identified Urm1 conjugation sites appear weakly conserved between homologs from different organisms (e.g., surface lysines in ScAhp1 versus CtAhp1) and removing all lysine residues still permits the sulfur transfer. Intriguingly, we did not detect functional consequences of deleting various lysine residues in Ahp1 in vivo, suggesting that the conjugation of Urm1 after the transfer of sulfur appears functionally dispensable. If this holds true, it would provide an optimal framework for the evolution of more complex E2/E3‐based conjugation systems around the formed GG‐K/GG‐S/GG‐T bond, potentially leading to functional specialization of the array of UBLs we see across eukaryotes today. Furthermore, it is very likely that most of the Urm1 conjugates are removed by one or more unidentified Urm1‐specific deconjugating/deubiquitinating enzyme(s) (DUBs) that can be blocked by NEM. Therefore, the attachment of Urm1 might be transient and the cysteine persulfidation of the target protein would outlast the actual modification by attachment of a Urm1 molecule. The existence of a potential “deurmylase” might have led to a strong underestimation of the number of cellular targets and caused the inconsistencies between Urm1 target lists from various studies and organisms (Schlieker et al, 2008; Van der Veen et al, 2011; Khoshnood et al, 2016; Wang et al, 2019). As Urm1 can also attach to serines and threonines, the recently described class of ubiquitin esterases (de Cesare et al, 2021) could be highly relevant for the efficient removal of Urm1 from its target proteins in vivo. We suggest that the overall number of Urm1 targets and the comparison with the known list of persulfidated proteins need to be reassessed after experimentally eliminating the possibility of continuous removal of Urm1 from its targets by known DUBs, esterases, or a specific deurmylase in cells. Urm1 conjugation is induced by oxidative stress in human cell lines, flies, plants, and yeast (Furukawa et al, 2000; Goehring et al, 2003; Van der Veen et al, 2011; Jüdes et al, 2015, 2016; Khoshnood et al, 2016; Wang et al, 2019). A recent study showed that the homolog of Urm1 in Toxoplasma gondii, an opportunistic parasite, also functions as a protein modifier in oxidative stress response. URM1 deletion strains of T. gondii display reduced proliferation, replication, invasion, and virulence in mice. Hence, TgURM1 plays a pivotal role in T. gondii survival and Urm1 could represent a novel target for the treatment of toxoplasmosis (Tan et al, 2022). Our data show that the transfer of the unique thiocarboxylated C‐terminal glycine of Urm1 requires the oxidation of catalytic or peroxidatic cysteines. Urm1‐SH condenses on the oxidized cysteine and may form a rather short‐lived acyl disulfide intermediate. Of note, we assume that diamide, which does not represent an oxidative stressor in vitro, is able to catalyze Urm1 conjugation by promoting the formation of the acyl disulfide intermediate. A nearby residue attacks the acyl disulfide, resulting in a persulfidated cysteine and the simultaneous formation of a covalent Urm1 conjugate. Hence, our in vitro observations provide a direct rationale for the dependency on cysteine oxidation. We further show that Urm1 conjugation and persulfidation are mechanistically coupled in all targets that we tested from yeast and human systems, providing substantial evidence that Urm1 is the elusive pathway that enables cysteine persulfidation throughout evolution. Recent studies revealed that motifs surrounding persulfidated cysteines are highly enriched with lysine residues (Longen et al, 2016; Fu et al, 2020). These observations provide further logical support for our conclusions on a proteome‐wide scale and support the concept that Urm1 may be directly responsible for the modification of a significant fraction of cysteines known to be persulfidated in cells. It was shown that cysteinyl‐tRNA synthetase can charge tRNACys with already persulfidated cysteine, which can be incorporated into proteins during ribosomal translation (Akaike et al, 2017). However, Urm1‐mediated persulfidation occurs post‐translationally, allowing for a direct molecular response to conditions of oxidative stress and damage. Persulfides exhibit strong redox‐scavenging activities and contribute to redox signaling regulation, due to their nucleophilicity. However, the key role of persulfide modifications in proteins is to protect thiols in cysteine residues against irreversible oxidation (Nishida et al, 2012; Kasamatsu et al, 2016; Kimura, 2017). An overoxidized persulfidated cysteine can be reversed to a fully reduced cysteine, whereas an overoxidized cysteine can hardly be reduced under physiological conditions (Kasamatsu et al, 2016; Dóka et al, 2020). Selective persulfidation of cysteine residues constitutes an evolutionarily conserved defense mechanism against oxidative stress that is critical during aging (Zivanovic et al, 2019). The levels of persulfidation decrease throughout the lifetime of an organism, and likely correlate with a decline in the efficiency of the inducible protection mechanism. Furthermore, persulfidation can regulate the activity of proteins, and the modification of the catalytic cysteine (Cys152) in HsGAPDH represents one of the best‐studied examples (Mustafa et al, 2009)—and we confirmed in this study that Urm1 is responsible for Cys152 persulfidation in GAPDH. Our work shows that cysteine persulfidation can be installed post‐translationally by Urm1 in direct response to oxidative stress, providing potential new avenues for the development of Urm1‐based therapeutic strategies to protect the proteome against oxidative stress and cellular aging. The essential role of Urm1 in tRNA thio‐modification (Termathe & Leidel, 2021) has long outshone its role in protein conjugation. Despite our unexpected findings, it remains undisputed that Urm1‐SH interacts with the Ncs2/Ncs6 thiolase and mediates the final transfer of sulfur to U34 of several tRNAs (Leidel et al, 2009; Yoshida et al, 2015). On the other hand, it remains to be shown whether other SCPs in prokaryotes that carry a thiocarboxylated C‐terminus are also able to catalyze cysteine persulfidation of specific key targets. Taken together, our work illustrates that Urm1‐SH can transfer its sulfur group to cysteine side chains of proteins. Foremost, this process is tightly linked to its E2/E3‐independent protein conjugation unknown among other eukaryotic UBLs.
Wild‐type (Appendix Table S1), mutants, and truncated proteins (Appendix Table S2) were expressed in BL21(DE3) pRARE in TB media at 37°C for 6 h and overnight induction with 0.5 M IPTG at 20°C. Bacterial pellets were resuspended in lysis buffer (20 mM Tris–HCl pH 7.5, 200 mM NaCl, 10 mM imidazole, 0.15% TX‐100, 10 mM MgCl2, 1 mM β‐mercaptoethanol, 10 mg/ml DNase, 1 mg/ml lysozyme, 10% glycerol and a cocktail of protease inhibitors) and lysed to homogeneity using a high pressure homogenizer Emulsiflex C3 (Avestin). The proteins were purified with Ni‐NTA agarose (Qiagen) under standard conditions. Tags were cleaved with TEV protease and removed with a second Ni‐NTA purification step. Subsequently, the proteins were purified by size‐exclusion chromatography (SEC) on HiLoad 26/600 Superdex 200 and HiLoad 16/600 Superdex 75 prep grade columns (GE Healthcare) using ÄKTA™ start. Purified proteins were stored at −80°C in a storage buffer (20 mM Tris–HCl pH 7.5, 200 mM NaCl, and 1 mM DTT). Of note, ordering details of used chemicals and reagents are listed in Appendix Table S3.
To obtain thiocarboxylated HsUrm1, ScUrm1 CtUrm1, ScUbiquitin, ScSumo, and GFP, the sequences of the respective proteins were N‐terminally Intein‐CBD‐His6 fused and overexpressed in Escherichia coli and purified according to (Kinsland et al, 1998; Termathe & Leidel, 2018) with modifications. In brief, the bacterial pellet was resuspended in lysis buffer without a reducing agent (30 mM Tris–HCl pH 7.5, 300 mM NaCl, 30 mM imidazole, 0.15% TX‐100, 10 mM MgSO4, 10 mg/ml DNase, 1 mg/ml lysozyme, 10% glycerol and a cocktail of protease inhibitors) and lysed to homogeneity. The lysate was passed through a Ni‐NTA column, and, following washes, the fusion protein was eluted with elution buffer (30 mM Tris–HCl pH 7.5, 300 mM NaCl, 500 mM imidazole, and 10% glycerol). The eluates were dialyzed overnight to chitin‐column buffer (30 mM Tris–HCl pH 7.5 and 500 mM NaCl) and applied on a chitin column. The column was washed with chitin‐column buffer and the cleavage of the tag was induced through incubation with cleavage buffer (30 mM Tris–HCl pH 8.5, 500 mM NaCl, and 35 mM ammonium sulfide or 50 mM DTT) for 16 h at 4°C. This approach enabled us to purify proteins with C‐termini that were carboxylated (‐OH) by using DTT or thiocarboxylated (‐SH) by using ammonium sulfide and without additional residues at the N‐terminus (Kinsland et al, 1998). The eluted proteins were further purified by size‐exclusion chromatography on a HiLoad 16/600 Superdex 75 column on ™KTA™ start system and stored at −80°C in storage buffer (20 mM Tris pH 7.5 and 200 mM NaCl) for further use. The presence of thiocarboxylated C‐terminus was confirmed by running the protein on a polyacrylamide gel containing [(N‐acryloylamino)phenyl]mercuric chloride (APM). The Laemmli sample buffer without any reducing agent was used. For protein visualization, the gels were stained with Coomassie Brilliant Blue.
Uba4‐mediated thiocarboxylation of Urm1 was performed as described (Termathe & Leidel, 2018). Briefly, 20 μM of Urm1 was mixed with 10 μM of Uba4 in the thiocarboxylation buffer (20 mM HEPES pH 8.5, 150 mM NaCl, and 2 mM MgCl2), supplemented with 5 mM ATP, 5 mM TCEP and 180 μM of sodium thiosulfate (Na2S2O3) was added to the reaction mix and incubated for 1 h at 30°C. For Urm1‐SH, samples were desalted using PD SpinTrapTM G‐25 (Cytiva) columns and either loaded on SDS–PAGE gels supplemented with 20 μM of APM to visualize the shift of the thiol group or buffer exchanged into 40 mM ammonium acetate and analyzed by ESI‐MS. Thiocarboxylation of Urm1 was scaled up and Urm1‐SH was purified by using Superdex 200 Increase 10/300 GL (Cytiva) on ÄKTA™ start system. The purified Urm1‐SH was snap frozen and stored at −80°C in storage buffer (20 mM Tris pH 7.5 and 200 mM NaCl) for further use.
Carboxylated (‐OH) and thiocarboxylated (‐SH) protein samples were analyzed using micrOTOF‐Q II mass spectrometer (Bruker Daltonics, Germany) equipped with an electrospray ionization source. The instrument was calibrated prior to measurements with ESI‐L Low Concentration Tuning Mix (Agilent Technologies). Samples were desalted on Amicon Ultra‐0.5 3 K (Millipore) using 0.05–0.1% formic acid (FA) as a washing solution. Samples at a protein concentration of about 0.2 mg/ml in formic acid (from 0.0125% up to 1%) and acetonitrile (from 25 to 50%) were directly infused into mass spectrometer with a syringe pump at a flow rate of 6 μl/min. Mass spectrometer was operated in positive mode with a spray voltage of 4,500 V and dry gas temperature of 180°C. MS scans were acquired over a mass range of m/z 500–3,000. The MS spectra were processed with the Maximum Entropy Deconvolution algorithm in Data Analysis 4.1 software (Bruker Daltonics, Germany).
Twenty micromolar of Ahp1 and 10 μM of carboxylated or thiocarboxylated Urm1 were mixed in reaction buffer (20 mM Tris pH 7.5 and 200 mM NaCl). 0.5 mM tert‐Butyl hydroperoxide (TBH) was included and excluded as indicated. The reaction mix was incubated for 30 min at 37°C for Chaetomium thermophilum, and 30°C for Saccharomyces cerevisiae and Homo sapiens proteins. For the confirmation of the isopeptide bond TCEP, DTT and hydroxylamine were added at a final concentration of 5 mM after the initial 30‐min reaction and further incubated for 5 min at 37 or 30°C, respectively. The reactions were stopped by adding Laemmli sample buffer containing DTT and incubated for 5 min at 95°C. Subsequently, the samples were loaded on Bolt™ 4–12% Bis‐Tris Plus Gels (Thermo Fisher Scientific). For protein visualization, the gels were stained with Coomassie Brilliant Blue.
Twenty micromolar of Ahp1 and 10 μM of thiocarboxylated Urm1 were mixed in reaction buffer (20 mM Tris pH 7.5 and 200 mM NaCl). Oxidative agents such as tert‐Butyl hydroperoxide (TBH), hydrogen peroxide (HP), diamide, peroxynitrite, methylglyoxal, and Di‐tert‐butyl disulfide (DTB‐disulfide) at a final concentration of 0.5 mM were included and excluded as indicated. The reaction mix was incubated for 30 min at 37°C for C. thermophilum, and 30°C for S. cerevisiae. The reactions were stopped by adding Laemmli sample buffer containing DTT and incubated for 5 min at 95°C. Subsequently, the samples were loaded on Bolt™ 4–12% Bis‐Tris Plus Gels (Thermo Fisher Scientific). For protein visualization, the gels were stained with Coomassie Brilliant Blue.
Ten micromolar of Ahp1 or 10 μM of thiocarboxylated Urm1 were added to the reaction buffer (20 mM Tris pH 7.5 and 200 mM NaCl) and oxidative agents such as TBH, HP, diamide, peroxynitrite, methylglyoxal, and DTB‐disulfide at a final concentration of 0.5 mM were included or excluded as indicated. The reaction mix was incubated for 30 min at 30°C. The reactions were stopped by adding Laemmli sample buffer containing DTT for Ahp1 and Laemmli buffer without reducing agents for Urm1‐SH and incubated for 5 min at 95°C. Subsequently, the samples of Ahp1 were loaded on Bolt™ 4–12% Bis‐Tris Plus Gels (Thermo Fisher Scientific). Samples of Urm1‐SH were loaded on SDS–PAGE gels supplemented with 20 μM of APM. For protein visualization, the gels were stained with Coomassie Brilliant Blue.
Large‐scale urmylation of CtAhp1C30S and ScAhp1C31S was carried out for crystallization purposes. Two hundred and fifty micromolar of CtAhp1C30S and 125 μM of CtUrm1C55S‐SH were mixed in 4 ml reaction buffer (20 mM Tris pH 7.5 and 200 mM NaCl) and TBH was added for the final concentration of 0.5 mM. The reaction was carried out for 30 min at 37°C. The reaction mixtures were applied to an ion exchange column HiTrap Q FF GL column. The fractions were collected, and samples were run on SDS–PAGE to locate the fractions containing the CtAhp1C30S‐CtUrm1C55S complex. The fractions containing the complex were pooled and concentrated for further purification. The concentrated sample was loaded into HiLoad 16/600 Superdex 75 pg GL column on ÄKTA™ pure system that was equilibrated with 20 mM Tris pH 7.5, 200 mM NaCl, and 1 mM TCEP. The integrated peak areas corresponding to the complex CtAhp1C30S‐CtUrm1C55S and individual proteins CtAhp1C30S and CtUrm1C55S‐SH were calculated using the UNICORN 7.0 software. Subsequently, the samples from the fractions were loaded on Bolt™ 4–12% Bis‐Tris Plus Gels (Thermo Fisher Scientific). For protein visualization, the gels were stained with Coomassie Brilliant Blue. The fractions containing the complex 30 kDa in size were collected and up concentrated to 25 mg/ml and proceeded with crystallization or snap frozen and stored at −80°C for further use. For urmylation of ScAhp1C31S and sample preparation of crystallization for the ScAhp1C31S‐ScUrm1 complex, similar procedures as CtAhp1C30S were used.
For crystallization, CtAhp1 and CtAhp1C30S were concentrated to 25 mg/ml in 20 mM Tris pH 7.5, 200 mM NaCl and 1 mM DTT. Crystals of the protein were grown at 21°C by vapor diffusion in sitting drops composed of equal volumes (1 μl each) of protein solution and crystallization buffer (2 M ammonium sulfate and 0.1 M sodium cacodylate pH 6.5). Crystals collected from reservoirs containing 2 M ammonium sulfate and 0.1 M sodium cacodylate‐pH 6.5 were cryoprotected by serial transfer into the cryoprotecting buffer (2 M ammonium sulfate and 0.1 M sodium cacodylate pH 6.5 and 50% glycerol). X‐ray diffraction data for CtAhp1 at 1.75 Å resolution were recorded at the ESRF beamline in Grenoble, France. X‐ray diffraction data for CtAhp1C30S at 1.85 Å resolution were recorded at BESSY in Berlin, Germany. Crystallization of unreacted CtAhp1C30S was carried out by pooling the unconjugated protein from the urmylation reaction and further loading into the HiLoad 16/600 Superdex 75 pg GL column on ÄKTA™ Pure system. The fractions of CtAhp1C30S were pooled and concentrated to 25 mg/ml 20 mM Tris pH 7.5, 200 mM NaCl and 1 mM TCEP. Crystals of the protein were grown at 21°C by vapor diffusion in sitting drops composed of equal volumes (1 μl each) of protein solution and crystallization buffer (2 M ammonium sulfate and 0.1 M sodium cacodylate pH 6.0). Crystals collected from reservoirs containing 2 M ammonium sulfate and 0.1 M sodium cacodylate pH 6.0 were cryoprotected by serial transfer into the cryoprotecting buffer (2 M ammonium sulfate and 0.1 M sodium cacodylate pH 6.0 and 30% glycerol). X‐ray diffraction data for CtAhp1C30S at 1.10 Å resolution were recorded at BESSY in Berlin, Germany. Crystals of CtAhp1C30S‐CtUrm1C55S complex were grown at 21°C using the hanging drop vapor diffusion technique. In each drop, 2 μl of protein sample was mixed with 2 μl of reservoir solution (0.05 M zinc acetate dihydrate, 0.1 M sodium cacodylate pH 6.5, 4% w/v polyethylene glycol 8000, and 30% w/v ethylene glycol). The crystals appeared after 48 h and grew to maximal size in 2 weeks. The crystals were cryoprotected by serial transfer into the cryoprotecting buffer (0.05 M zinc acetate dihydrate, 0.1 M sodium cacodylate pH 6.5, 4% w/v polyethylene glycol 8000, and 50% w/v ethylene glycol) and snap frozen in liquid nitrogen. X‐ray diffraction data at 2.40 Å resolution were collected at BESSY in Berlin, Germany. For data collection details see Table 1. The structures of CtAhp1, CtAhp1C30S and CtAhp1C30S‐CtUrm1C55S complex were determined by molecular replacement using Phaser (McCoy et al, 2007) with S. cerevisiae Ahp1 (PDB ID 4DSR; Lian et al, 2012) and Urm1 (PDB ID 2QJL; Yu & Zhou, 2008), respectively. Structures were refined using Phenix (Adams et al, 2010) and Refmac5 in CCP4 (Winn et al, 2011) programs. The comprehensive validation was done by MolProbity (Davis et al, 2007). Structural visualization was done with PyMOL (https://pymol.org/2/). For structure refinement statistics see Table 1. Anomalous density collection for Zn peak and below Zn peak data was collected for the CtAhp1C30S‐CtUrm1C55S complex at BESSY in Berlin, Germany. A complete list of all used software is listed in (Appendix Table S4).
Twenty micromolar of target proteins such as HsPRDX5, ScTdh3, HsGAPDH, ScSes1, HsSARS1, ScPyk1, HsPKM2 and 20 μM of carboxylated or thiocarboxylated Urm1 were mixed in reaction buffer (20 mM Tris–HCl pH 7.5 and 200 mM NaCl). Diamide at a final concentration of 0.5 mM was included or excluded as indicated. The reaction mix was incubated for 30 min at 30°C. The reactions were stopped by adding Laemmli sample buffer containing DTT and incubated for 5 min at 95°C. Subsequently, the samples were loaded on Bolt™ 4–12% Bis‐Tris Plus Gels (Thermo Fisher Scientific). For protein visualization, the gels were stained with Coomassie Brilliant Blue.
Samples were prepared and measured as described (Pabis et al, 2020) with minor changes. The flow rate during peptide separation on the analytical column was 250 nl/min. Moreover, different proteolytic enzymes were used depending on what modification was to be detected. Chymotrypsin was used to find urmylation and GFP conjugation, while trypsin to identify ubiquitination. To detect sumoylation, trypsin, chymotrypsin, and V8 protease were employed.
Protein samples were prepared in 20 mM Tris and 200 mM NaCl. A sample amount corresponding to about 10 μg of the protein of interest was used for digestion. Final volume of the digestion solution was 60 μl. Urea was added to the concentration of 0.5 M. Digestion solution was filled up to 60 μl with 50 mM ammonium bicarbonate. Chymotrypsin was used in the enzyme‐to‐protein ratio of 1:20. Digestion solution was supplemented with CaCl2 to the concentration of about 10 mM. Samples were incubated at 25°C overnight. The next day, digestion was stopped by adding trifluoroacetic acid to a concentration of 0.5%. Approximately 1% of digested samples were injected for LC–MS/MS analysis, which was performed the same as for samples from gel bands.
The LC–MS/MS data were processed using the Proteome Discoverer platform (v.1.4; Thermo Scientific) and searched using an in‐house MASCOT server (v.2.5.1; Matrix Science, London, UK) against cRAP database (https://www.thegpm.org/crap/) supplemented with the sequences of the proteins of interest. The following modifications were included in search parameters depending on the sample: urmylation (HisGlyGly tag after chymotrypsin, ∆ mass = 251.101839); ubiquitination (GlyGly tag after trypsin, ∆ mass = 114.042927); sumoylation (GlyGlyIleGlnGlu tag after trypsin, ∆ mass = 503.246552; GlyGlyIleGln tag after chymotrypsin or V8 protease, ∆ mass = 374.203959); GFP conjugation (Lys tag after chymotrypsin; ∆ mass = 128.094963); cysteine carbamidomethylation (∆ mass = 57.021464); cysteine persulfidation (∆ mass = 31.972071); carbamidomethylated cysteine persulfidation (∆ mass = 88.993534); methionine oxidation (∆ mass = 15.994915).
To produce free sulfide from L‐Cys, 10 μM of E. coli desulfurase IscS was incubated at 25°C overnight with 1 mM L‐Cys and 20 μM PLP cofactor in a desulfurase buffer (20 mM Hepes pH 8.5, 150 mM NaCl, 2 mM MgCl2). To examine the fate of sulfur in the yeast system, 10 μM of ScUba4 and 40 μM of ScUrm1 were added, with 2.5 mM ATP and 1 mM TCEP, and incubated for 1 h at 30°C. Analogously, for Ct., 10 μM of CtUba4 and 40 μM of CtUrm1C55S were added with 2.5 mM ATP and 1 mM TCEP, and incubated for 1 h at 37°C. Following incubation, the samples were desalted using PD SpinTrapTM G‐25 (Cytiva) columns, buffer exchanged into 40 mM ammonium acetate, and analyzed by ESI‐MS as previously described.
To generate thiocarboxylated Urm1 labeled with 35S, we purchased 35S‐containing L‐cysteine (Perkin Elmer, NEG022T001MC) and used it as a sulfur source that simulates one of the plausible sulfur sources in vivo. Approximately 0.1 mCi of 35S L‐cysteine was desulfurated using a 20 μM recombinant IscS desulfurase from E. coli for 1 h at room temperature in a desulfuration buffer containing 20 mM HEPES pH 8.5, 150 mM NaCl, 2 mM MgCl2, and 80 μM pyridoxal phosphate (PLP) necessary for IscS activity. Next, 10 μM Uba4 and 70 μM Urm1 proteins together with 2 mM ATP and 1 mM TCEP were added to the tube containing desulfurated cysteine. The Urm1 thiocarboxylation reaction was executed for 1 h at 30°C. To ensure that most of the 35S atoms were transferred to Urm1, 180 μM of sodium thiosulfate was added to the reaction mix and incubated for 15 min. Samples were desalted using Amersham WB MiniTrap kit (Cytiva) following the manufacturer's protocol to get rid of free cysteine, 35S, and reducing agent. Fresh 35S‐thiocarboxylated Urm1 was used for urmylation reactions in combination with 20 μM corresponding target proteins in presence of 0.5 mM oxidizing agents (TBH or diamide). The urmylation reaction was performed at 30°C for 1 h after which samples were denatured at 95°C in sample buffer without any reducing agents. Proteins were separated using SDS–PAGE, and gels were stained with Coomassie and dried for 2 h in the Gel Dryer Model 583 (BioRad). Dry gels were exposed overnight to storage phosphor screens (Cytiva) and the accumulated signal was visualized using the Personal Molecular Imager system (BioRad). All assays shown herein were repeated at least three times.
To examine the activatory potential of different UBLs on the ATPase activity of Uba4, we used a commercially available Malachite Green Kit (Sigma‐Aldrich). Assembled reactions contained 40 μM Uba4 and 60 μM of the UBL of interest, resuspended in 100 mM MES pH 6.0, 100 mM NaCl, 2 mM MgCl2 and 160 μM ATP. The addition of 5 mU/reaction of inorganic pyrophosphatase (Thermo Fischer Scientific) allowed us to convert all inorganic pyrophosphate crated by Uba4 to phosphate molecules that can be quantified using Malachite Green. The reaction took place for 90 min at 37°C. The reaction product was diluted 20× in water and developed according to the manufacturer's instructions. Absorbance was measured at 620 nm using SpectraMax 190 Microplate Reader (Molecular Devices). Data were acquired in three independent experiments with two technical replicates each.
ScUba4‐mediated thiocarboxylation of GFP‐ScUrm1 was done as described for ScUrm1. Briefly, 20 μM of GFP‐ScUrm1 was mixed with 10 μM of ScUba4 in the thiocarboxylation buffer (20 mM Hepes pH 8.5, 150 mM NaCl, 2 mM MgCl2), supplemented with 5 mM ATP and 5 mM TCEP and incubated for 1 h at 30°C. The GFP‐ScUrm1‐SH sample was loaded on Superdex 200 Increase 10/300 GL (Cytiva) to isolate GFP‐ScUrm1‐SH.
Conjugation reactions between 20 μM of GFP‐ScUrm1‐SH and 20 μM of target proteins ScAhp1, ScPyk1, and ScSes1 were done in the presence of 0.5 mM diamide for 30 min at 30°C. Carboxylated GFP‐ScUrm1‐OH was used as a negative control. The protein samples were denatured at 95°C in the presence of Laemmli sample buffer and analyzed on Bolt™ 4–12% Bis‐Tris Plus Gels (Thermo Fisher Scientific). For protein visualization, the gels were stained with Coomassie Brilliant Blue.
Saccharomyces cerevisiae strains used and generated in this study (Appendix Table S5) were grown on complete (YPD) or synthetic (SC) media (Sherman, 1991) at 30°C unless otherwise indicated. 80 μg/ml L‐cysteine hydrochloride was supplemented for the growth of auxotrophic strains when required. Yeast gene deletions were generated by PCR using the pUG plasmid system (Gueldener et al, 2002) and gene‐specific oligonucleotides (Appendix Table S6). Correct gene replacements were confirmed by PCR using primer pairs located outside of the target loci (Appendix Table S6). For the construction of yeast expression vectors (Gietz & Akio, 1988) carrying Ahp1 lysine to arginine mutants, we used the FastCloning technique (Li et al, 2011). In brief, the mutated AHP1 gene sequence was amplified from the corresponding pETM‐30 vector (Appendix Table S2) using primer pairs (Appendix Table S6) adding overlapping ends homologous to the linearized pAJ31 target plasmid. All mutations were verified by Sanger‐based DNA sequencing. The yeast expression plasmid pHA‐URM1 (Appendix Table S7) was used as a template to N‐terminally insert a polyhistidine affinity tag (8xHis) by PCR‐based site‐directed mutagenesis (Wang & Malcolm, 1999) with the appropriate oligonucleotide primers (Appendix Table S7) resulting in pLK20 (Appendix Table S7) that was used for in vivo urmylation assays. Transformation of yeast cells with PCR products or plasmids (Appendix Table S7) was performed as previously published (Gietz & Woods, 2002).
In vivo urmylation studies were performed as previously described (Jüdes et al, 2015). In brief, yeast cells grown to an OD600nm of 1.0 were harvested and lysed mechanically with glass beads in a buffer (10 mM K‐HEPES pH 7.0, 10 mM KCl, 1.5 mM MgCl2, 0.5 mM PMSF, and 2 mM benzamidine) containing complete protease inhibitors (Roche) and 2.5 mM N‐ethylmaleimide (NEM). Protein concentrations were determined according to Bradford assay (Bradford, 1976) and lysates were mixed with SDS sample buffer (62.5 mM Tris–HCl pH 6.8, 2% SDS, 10% glycerol, 0.002% bromophenol blue, and 5% β‐mercaptoethanol) according to Laemmli (Laemmli, 1970). For Western blot analyses, proteins were transferred to PVDF membranes and incubated with primary anti‐HA antibodies (F7, Santa Cruz Biotechnology or 2–2.214, Invitrogen). Unconjugated Ahp1 was detected using anti‐Ahp1 serum (Iwai et al, 2010) kindly provided by Dr Kuge (Tohoku Pharmaceutical University, Japan). Equal protein loading was verified with anti‐Cdc19 antibodies donated by Dr Thorner (University of California‐Berkeley, USA). Detection of target proteins involved horseradish peroxidase‐conjugated secondary goat anti‐mouse or anti‐rabbit IgGs (Jackson ImmunoResearch) and the WesternBright ECL Spray (Advansta Inc.) according to manufacturers' instructions. For visualization of target proteins, the Odyssey® Fc Imaging System (LI‐COR, Inc.) was used.
Overnight yeast cultures were diluted to an OD600nm of 1.0. Ten‐fold serial dilutions were prepared ranging from 10−1 to 10−3 and transferred to YPD medium using a pin tool. For toxicity assays, the medium was supplemented with the organic peroxide TBH as indicated. Plates were incubated at 30°C unless otherwise stated and monitored after 36–48 h.
Total tRNA was extracted from yeast as previously described (Krutyhołowa et al, 2019), using NucleoZOL reagent according to the manufacturers' instructions. The pelleted tRNA was washed once with 75% ethanol and stored in 100% ethanol for subsequent northern blot analysis. 0.4 μg of total tRNA was separated by electrophoresis on 12% denaturing polyacrylamide gels (7 M Urea, 0.5× TBE buffer), visualized by SYBR Gold (Invitrogen), and transferred to a nylon membrane (Immobilon‐Ny+) at 400 mA for 45 min using a Trans‐Blot® SD Semi‐Dry Transfer Cell (BioRad). To analyze the tRNA thiolation levels, the gels were supplemented with [(N‐acryloylamino)phenyl]mercuric chloride (APM) at a final concentration of 60 μg/ml (APM stock solution: 3 μg/μl in formamide). For APM+ gels the transfer was performed for 1 h and the transfer buffer was supplemented with 10 mM DTT to improve the transfer of thiolated tRNA. Membranes were hybridized at 42°C to a 32P‐5′‐end‐labeled DNA probe 5′‐tggctccgatacggggagtcgaac‐3′, which is complementary to a 3′ part of tRNAUUCGlu. Labeling of the probe with [γ‐32P]‐ATP, hybridization, and subsequent steps of the Northern blotting procedure were performed as described in (Leidel et al, 2009). The quantitative analysis was performed using a GelAnalyzer system. Thiolation levels (%) were calculated for three biological replicates. Aliquots from the spotting experiment were thawed on ice and total RNA was extracted by using hot phenol/chloroform extraction. One microgram of total RNA was resolved on an 8% PAGE containing 0.5 × TBE, 7 M Urea, and 50 μg/ml APM (Igloi, 1988). Northern blot analysis was performed as described previously by using the probe against tRNAUUCGlu (5′‐tggctccgatacggggagtcgaac‐3′; Leidel et al, 2009).
Aliquots (3 OD600 units) from spotting experiments were thawed on ice and total proteins were extracted as previously described (von der Haar, 2007). Total protein extracts were resolved by SDS–PAGE and transferred by semi‐dry blotting onto a PVDF membrane. Membranes were probed using a monoclonal anti‐HA antibody (Covance MMS‐101R).
The yeast strain used in this experiment was BY4742. The wild‐type strain together with the single knockout strains ahp1Δ, urm1Δ, tum1Δ, ncs6Δ, cys3Δ, and cys4Δ and the double knockout deletion strains cys3Δurm1Δ and cys4Δurm1Δ were maintained and grown on yeast extract‐peptone‐dextrose medium (YPD). For the Bismuth Glucose Glycine Yeast agar (BiGGY) screening, strains were grown overnight in a YPD medium and diluted to an OD600nm = 0.8 and OD600nm = 0.4. Four microliters of each strain were spotted on BiGGY agar plates and incubated for 72 h at 27°C. The production of H2S was evaluated using a color scale dependent on the production of sulfide, the more precipitation of bismuth sulfide, the darker the colony. Taking this into consideration, the scale comprises the following colors: white, cream, light brown, and dark brown (Mezzetti et al, 2014; Cirigliano et al, 2016).
Keerthiraju E Ravichandran: Conceptualization; data curation; formal analysis; validation; investigation; visualization; methodology; writing – original draft; writing – review and editing. Lars Kaduhr: Formal analysis; investigation; visualization; methodology; writing – review and editing. Bozena Skupien‐Rabian: Data curation; investigation; visualization; methodology. Ekaterina Shvetsova: Investigation; methodology. Mikołaj Sokołowski: Data curation; formal analysis; investigation; visualization; methodology; writing – review and editing. Ros´cisław Krutyhołowa: Data curation; formal analysis; investigation; visualization; methodology; writing – review and editing. Dominika Kwasna: Investigation; writing – review and editing. cindy Brachmann: Data curation; investigation; methodology. Sean Lin: Data curation; investigation; methodology. Sebastian Guzman Perez: Data curation; formal analysis; validation; investigation; visualization; methodology. Piotr Wilk: Data curation; formal analysis; validation; investigation; methodology. Manuel Kösters: Data curation; investigation. Przemyslaw Grudnik: Data curation; formal analysis; supervision; validation; investigation; visualization; methodology. Urszula Jankowska: Data curation; formal analysis; supervision; validation; investigation; visualization; methodology; writing – review and editing. Sebastian A Leidel: Conceptualization; data curation; formal analysis; supervision; funding acquisition; validation; writing – original draft; project administration; writing – review and editing. Raffael Schaffrath: Conceptualization; data curation; formal analysis; supervision; funding acquisition; validation; writing – original draft; project administration; writing – review and editing. Sebastian Glatt: Conceptualization; data curation; formal analysis; supervision; funding acquisition; validation; investigation; visualization; methodology; writing – original draft; project administration; writing – review and editing.
KER and SG are inventors on a filed patent application (EP 22461521.1) relating to certain aspects of the presented work in this article.
Click here for additional data file. Click here for additional data file. Click here for additional data file. | true | true | true |
PMC9575007 | 36228259 | Esra Tekcan,Nurten Kara,Hasan Murat Aydın,Ümmet Abur,Mohsen Abbaszadeh | Evaluation of the promoter methylation status of hypoxia factor 3A and interleukin-6 genes and expression levels of mir-130b and mir-146b in childhood obesity | 07-10-2022 | DNA methylation,Epigenomics,MicroRNA,Obesity | SUMMARY OBJECTIVE: Obesity, which causes many serious diseases, is increasing exponentially in childhood across the world. Epigenetic changes, as well as genetics, play an important role in the process of adipogenesis. Therefore, we aimed to examine the expression levels of obesity-related MicroRNA-130b and MicroRNA-146b and the methylation status of hypoxia factor 3A and interleukin-6 genes associated with obesity in children. METHODS: This study was performed with 98 individuals (49 obese children and 49 controls) whose DNA was isolated from peripheral blood. Gene promoter methylations were analyzed by methylation-specific Polymerase chain reaction. In addition, expression levels of MicroRNAs were determined by quantitative real-time Polymerase chain reaction in 30 children (15 obese children and 15 controls). RESULTS: Methylation status of interleukin-6 gene was 93.9% in obese children (n=46/49) and 100% (n=49/49) in control group (p>0.05). There was no methylation for hypoxia factor 3A gene (p>0.05). As a result of the study, there was no statistically significant difference in terms of methylation status for hypoxia factor 3A and interleukin-6 genes in the obese group compared to the control group. However, we found that expression levels of MicroRNA-130b (p<0.01) and MicroRNA-146b (p<0.001) were higher in the obese group. CONCLUSIONS: Results support that MicroRNA-130b and MicroRNA-146b are potential biomarkers for the prevention and early diagnosis of obesity. This is the first study on childhood obesity in the Middle Black Sea region of Turkey. We believe that the results obtained by expanding the studies in our country and neighboring countries will be more decisive. | Evaluation of the promoter methylation status of hypoxia factor 3A and interleukin-6 genes and expression levels of mir-130b and mir-146b in childhood obesity
Obesity, which causes many serious diseases, is increasing exponentially in childhood across the world. Epigenetic changes, as well as genetics, play an important role in the process of adipogenesis. Therefore, we aimed to examine the expression levels of obesity-related MicroRNA-130b and MicroRNA-146b and the methylation status of hypoxia factor 3A and interleukin-6 genes associated with obesity in children.
This study was performed with 98 individuals (49 obese children and 49 controls) whose DNA was isolated from peripheral blood. Gene promoter methylations were analyzed by methylation-specific Polymerase chain reaction. In addition, expression levels of MicroRNAs were determined by quantitative real-time Polymerase chain reaction in 30 children (15 obese children and 15 controls).
Methylation status of interleukin-6 gene was 93.9% in obese children (n=46/49) and 100% (n=49/49) in control group (p>0.05). There was no methylation for hypoxia factor 3A gene (p>0.05). As a result of the study, there was no statistically significant difference in terms of methylation status for hypoxia factor 3A and interleukin-6 genes in the obese group compared to the control group. However, we found that expression levels of MicroRNA-130b (p<0.01) and MicroRNA-146b (p<0.001) were higher in the obese group.
Results support that MicroRNA-130b and MicroRNA-146b are potential biomarkers for the prevention and early diagnosis of obesity. This is the first study on childhood obesity in the Middle Black Sea region of Turkey. We believe that the results obtained by expanding the studies in our country and neighboring countries will be more decisive.
Obesity is defined as abnormal or excessive fat accumulation that may impair health and cause morbidity and finally mortality by affecting many organs . It is also an increasingly growing problem in childhood worldwide and it is a serious nutritional disorder that may lead to serious health problems in future if no measures are taken in early stage . In 2016, approximately 40 million children under the age of 5, as well as 340 million children between the ages of 5 and 19 years, were affected by overweight or obesity . According to COSI-TUR (Childhood Obesity Surveillance Initiative), the rate of overweight was 14.6% and that of obesity was 9.9% in primary school second-grade children in Turkey in 2016 . This rapid increase in obesity is not only due to genetic factors but also due to environmental factors that cause epigenetic effects . Hypoxia occurs due to the decrease in the amount of blood entering adipocytes with the growth of adipose tissue in obese subjects. Methylation of hypoxia factor 3A (HIF3A) gene is related to body mass index (BMI) and adiposity. It was also reported that obesity is associated with a low degree of chronic inflammation and increased interleukin-6 (IL-6) levels . It is known that miRNAs play an important role in adipose tissue differentiation, proliferation, and lipid and glucose metabolism . Identifying obesity-associated miRNAs that cause inadequate or overexpression of proteins in cells may be a therapeutic target for future treatment by using anti-miRNA oligonucleotides targeting these miRNAs. There are several other miRNAs associated with obesity; however, due to limited cost, we compared the expression levels of only two miRNAs in this study. The tissue-specific epigenetic is subject to stunning changes during childhood development. A study has reported a correlation between methylation in adipose tissue and peripheral blood in adults, but it has not proved whether this holds true for children . Therefore, we aimed to observe the methylation status of HIF3A and IL-6 genes and expression levels of miR-130b and miR-146b, which are thought to be associated with childhood obesity from peripheral blood. As methylation is more tissue specific, it was necessary to use adipose tissue for most genes. In this study, blood sample was used in order to look for the methylation of obesity-related genes whose expression can be observed in the blood. Our study individuals involved adolescent in the Central Black Sea region and present previously unexplored data. It aims to support the data presented in previous studies in different populations. This is the first study to epigenetically examined HIF3A and IL-6 genes among Turkish obese children.
This study was carried out among children aged 6–12 years who were diagnosed with obesity (BMI>95th) before puberty period and who visited Ondokuz Mayıs University Faculty of Medicine Pediatric Endocrinology from the Middle Black Sea region of Turkey. A total of 98 children (49 obese children and 49 healthy controls) were selected to examine methylation status. The sample size was determined by statistical power analysis. To evaluate obese and control DNA methylations, 20 obese patients and 20 controls were selected, with 97% power to detect d=0.122 difference (with a maximum deviation of 0.11) at the 95% confidence interval. For miRNA expression analysis, a total RNA of 30 (15 obese subjects and 15 controls) samples were isolated from the whole peripheral blood samples of 98 volunteers. Again, 12 obese patients and 12 controls were calculated for 96% power to detect d=0.138 difference (with a maximum deviation of 0.15) at the 94% confidence interval to evaluate obese and control miRNA expression. This study was approved by Ethics Committee of the OMÜ KAEK (registration number 2016/301). In this study, chronic disease, chronic drug use, and endocrine, metabolic, or genetic short status were determined as exclusion criteria.
DNA was isolated from peripheral blood samples using Pure Link® Genomic DNA Mini Kit (Invitrogen, USA). EpiJET Bisulfite Conversion Kit (Thermo Fisher Scientific, Lithuania) was used for bisulfite modification of DNA. Specific primers for methylation-specific PCR analysis were designed with the “MethPrimer” database (Table 1). DreamTaq™ Hot Start DNA polymerase was used (Thermo Fisher Scientific). A 25-μl total volume of PCR reaction conditions was as follows: For IL-6: initial denaturation was at 97°C for 5 min, followed by 40 cycles of 96°C for 45 s, 56°C for 45 s, and 72°C for 60 s. The final extension was at 72°C for 7 min. For HIF3A: initial denaturation was at 98°C for 10 min, followed by 40 cycles of 97°C for 45 s, 60°C for 45 s, and 72°C for 70 s. The final extension was at 72°C for 7 min.
A 2% of micropore (Prona, Nu micropor) agarose gel was prepared to display the MSP product. A 50-bp DNA marker (Thermo Fisher Scientific, Lithuania) was used as the marker and gel and viewed using UV transilluminator.
Total RNA was isolated from whole peripheral blood samples using Quick-RNA™ Whole Blood kit (Zymo Research, USA). Total RNA was reverse transcribed to cDNA using Ipsogen® RT Kit 33, V1 kit (QIAGEN, Germany). The reaction mixture was 15 μl. Incubation conditions were as follows: 16°C for 30 min, 42°C for 30 min, and 85°C for 5 min (Thermal Cycler GeneAmp PCR System 9700, Applied Biosystems, USA). For miR-146b and miR-130b qRT-PCR analysis, Taqman™ MicroRNA Assay (Applied Biosystems) and Premix Ex Taq™ master mix (Perfect Real Time, Takara, Japan) were used. RNA was normalized using the ABL housekeeping gene (İpsogen, BCR-ABL1 Mbcr IS-MMR) in each sample . Using standards with a known number of molecules, one can establish a standard curve and determine the amount of target present in the test sample. To ensure accurate standard curves, we use four standard dilutions for ABL. The kit also includes an IS-MMR calibrator allowing conversion of results to the international scale. For ABL gene, raw CT values obtained from plasmid standard dilutions are plotted according to the log copy number. We compare the ABL gene with known copy number in the kit with our miRNAs whose copy number is unknown, and we obtain information about the copy number. We compare our ABL-normalized obese and control miRNAs with each other. As a result, with CT values, we obtained information about copy numbers increase or decrease of obese individuals compared to controls. Rotor-gene Q 5 PLEX HRM (Germany, USA) device used for qRT-PCR. The miR-130b microRNA sequence is ACUCUUUCCCUGUUGCACUAC, while miR-146b microRNA sequence is UGAGAACUGAAUUCCAUAGGCUG (http://www.mirbase.org). The relative abundance of each miRNA transcript in each sample was determined using the comparative 2−ΔΔCt method and an endogenous housekeeping control gene. Cycle threshold (CT) values of participants (obese and control) and increase or decrease ratio of expression levels in miR-146b-5p/ABL and miR-130b/ABL were calculated according to the Livak’s formula :
The relationship between parameters was evaluated using Fisher’s exact test using OpenEpi version 3.01 (last updated: May 6, 2013). SPSS software was used for data analyzing (version 22.0, SPSS Inc., Chicago, IL, USA). The statistical significance of miRNA expression was analyzed by the Student’s t-test, considering statistically significant at p<0.05.
The mean age of 49 obese groups (22 females, 27 males) whose methylation profile was analyzed was 10.0±1.9 years, and the mean age of 49 healthy controls (29 females, 20 males) was 9.2±2.0 years. HIF3A gene methylation was not detected in either group. IL-6 gene methylation was detected in 93.9% (46/49, p>0.05) of the obese group and 100% (49/49, p>0.05) of the control group. According to results of the tests, there was no statistical significance between the patient and control groups in terms of the promoter methylation profile of HIF-3A and IL-6 genes (p>0.05, Table 2).
The expression of obese miR-146b according to the control group was increased in 14 of 15 children, with an average 25.5±17.3-fold (p<0.001). Similarly, the expression of miR-130b increased in 14 of 15 obese children, with an average 8.5±5.3-fold (p<0.01). The qRT-PCR analysis results of the expression levels of miR-146b and miR-130b in obese and control groups were given in Table 3 and Figure 1.
One of the most studied mechanisms as an epigenetic determinant of childhood obesity is DNA methylation . In many studies, there was evidence showing that the HIF3A gene is associated with anthropometric parameters in childhood . Previous studies have reported that methylation of the HIF3A gene is associated with BMI and adiposity . In our study, we observed that there was no significant difference in methylation status of HIF3A genes in the obese group compared to the control group (p<0.01). The temporal sequence between epigenetic changes and the onset of childhood obesity is uncertain because epigenetics may be altered by a wide range of stimuli, including metabolic changes associated with obesity itself. Most of the studies included in the review use a cross-sectional design that makes it impossible to decipher the temporal order of events. Some DNA methylation studies have used longitudinal designs or statistical techniques such as Mendelian randomization to resolve this relationship . In the ALSPAC (avon longitudinal study of parents and children) cohort, childhood BMI was associated with methylation at HIF3A gene in adolescence, but childhood methylation was not robustly associated with BMI in adolescence, and based on Mendelian randomization, HIF3A methylation did not play a causal role on BMI . The children participating in our study were in pre-adolescent, so we may not have observed HIF3A gene methylation in our study. Cytokines production causes an increase in white adipose tissue in obesity subjects . In our study, there was no significant difference in methylation status of IL-6 gene in the obese group compared to the control group (p<0.01), similar to results obtained by Zhang et al. . Na et al. reported that IL-6 methylation can be used as a molecular biomarker for the assessment of obesity risk, finding which is in contrast to our results. The mechanism that can explain the increased IL-6 levels in obese can be attributed to the ability of fat tissue to produce and secrete IL-6 . In our study, DNA methylation may not reflect the actual methylation profile since DNA methylation was analyzed from peripheral blood, not directly from the affected adipose tissue. miRNAs, which are considered to be epigenetically important in regulating the expression of genes, play an active role in insulin secretion, cholesterol biosynthesis, fat metabolism, and adipogenesis. In our study, miRNA profiles were examined in 30 children (15 obese and 15 control) whose proficiency was supported by power analysis due to project budget limitations. However, circulating mir-130b and mir-146b expression levels were found to be high in obese children, although the number of samples studied was limited. Mansego et al. in 24 children (12 obese children and 12 controls) and Quyang et al. in a total of 12 children (6 obese children and 6 controls) reported that the miRNA profiles were associated with early childhood obesity from peripheral blood. Previous study showed that miR-130b expression in abdominal subcutaneous adipose tissue and plasma was low in obese. Lacomino and Siani concluded that the level of miR-130b in the blood was found to be higher in obese children, as in our study (p<0.01). Wang et al. determined miR-130b as a potential biomarker for overweight, hypertriglyceridemia, and metabolic syndrome. TGF-β signaling pathway is significant in the arrangement of energy homeostasis and studies suggest that miR-130b secreted from adipose tissues mediate the metabolic regulatory effect of TGF-β and TGF-β can stimulate miR-130b secretion from adipocytes . This explains the 8.5±5.3-fold increase in miR-130b expression in the obese group in our study. Prtas-Puig et al. proved that circulating miRNAs were unregulated in prepubescent obese children, which is similar to our results. Therefore, early detection of abnormal miRNA profile will be useful in the early diagnosis of obesity. In addition, miR-146b expressed during adipogenic differentiation and its level increased in overweight, obese, and high fat diet mice . The fact that the miR-146b expression level of obese individuals was 25.5±17.3-fold increase higher than the control group in our study supports this situation (p<0.001). Overexpression of Kruppel-like factor 7 (KLF7) inhibits expression of some adipogenic transcription factor genes and suppresses adipogenesis . The inhibitory effects of miR-146b on KLF7 expression and experimentally demonstrated that KLF7 is the direct target of miR-146b . Therefore, increased mir-146b expression in obesity supports adipogenesis indirectly. Sharma et al. confirmed that the expression of miR-146a and miR-146b significantly correlated with BMI in adipose tissues. Studies indicate that miR-146b can make significant contributions to obesity formation in children . Also, obese mice who were administered anti-miR-146b for 3 days to destroy miR-146b have shown losing weight . These results indicate that miR-146b may be a potential target for the treatment of obesity.
The role of epigenetic pathways in the prevention and treatment of increasingly common obesity is one of the current research areas. More studies are needed in both larger and different populations to support the results. | true | true | true |
PMC9575064 | 36128646 | Cong Ma,Jianling Du,Xiang Ma | tRNA ‐derived fragment tRF ‐1020 ameliorates diabetes‐induced retinal microvascular complications | 20-09-2022 | angiogenic function,retinal microvascular complication,tRNA‐derived RNA fragment,Wnt signalling | Abstract Transfer RNA (tRNA)‐derived fragments are the non‐coding single‐stranded RNAs involved in several physiological and pathological processes. Herein, we investigated the role of tRF‐1020, a tRNA fragment, in diabetes‐induced retinal microvascular complications. The results showed that the levels of tRF‐1020 expression were down‐regulated in diabetic retinal vessels and retinal endothelial cells following high glucose or H2O2 stress. Overexpressing tRF‐1020 led to decreased endothelial cell viability, proliferation, migration, and tube formation and alleviated retinal vascular dysfunction as shown by decreased retinal acellular capillaries, vascular leakage, and inflammation. By contrast, tRF‐1020 silencing displayed the opposite effects. tRF‐1020 regulated endothelial angiogenic functions and retinal vascular dysfunction by targeting Wnt signalling. Moreover, the levels of tRF‐1020 expression were reduced in aqueous humour and vitreous samples of the patients with diabetic retinopathy. Collectively, tRF‐1020 is a potential target for the diagnosis and treatment of diabetic retinopathy. | tRNA ‐derived fragment tRF ‐1020 ameliorates diabetes‐induced retinal microvascular complications
Transfer RNA (tRNA)‐derived fragments are the non‐coding single‐stranded RNAs involved in several physiological and pathological processes. Herein, we investigated the role of tRF‐1020, a tRNA fragment, in diabetes‐induced retinal microvascular complications. The results showed that the levels of tRF‐1020 expression were down‐regulated in diabetic retinal vessels and retinal endothelial cells following high glucose or H2O2 stress. Overexpressing tRF‐1020 led to decreased endothelial cell viability, proliferation, migration, and tube formation and alleviated retinal vascular dysfunction as shown by decreased retinal acellular capillaries, vascular leakage, and inflammation. By contrast, tRF‐1020 silencing displayed the opposite effects. tRF‐1020 regulated endothelial angiogenic functions and retinal vascular dysfunction by targeting Wnt signalling. Moreover, the levels of tRF‐1020 expression were reduced in aqueous humour and vitreous samples of the patients with diabetic retinopathy. Collectively, tRF‐1020 is a potential target for the diagnosis and treatment of diabetic retinopathy.
Diabetic retinopathy (DR) is a major cause of visual loss among the working‐age people. It is also one of the most common microvascular complications of diabetes. The pathogenesis of DR is tightly associated with increased vascular permeability, vascular occlusion, and neovascularization. Diabetes‐induced retinal microvascular complications can lead to vitreous haemorrhage, tractional retinal detachment, neovascular glaucoma, and eventually blindness. Current treatment methods include intravitreal pharmacologic agents, panretinal laser photocoagulation, and vitreous surgery, which can retard vision loss to a certain extent. However, these methods can also cause several side effects, such as rare infectious, retinal detachment, and increased intraocular pressure. , Further studies are still required to search for the potential targets or new alternative therapies to prevent or delay the progression of DR. Transfer RNAs (tRNAs) are a class of noncoding RNA transcripts corresponding to the delivery of specific amino acids to ribosome for translation. tRNAs can be cleaved by specific nucleases (e.g., Dicer and angiogenin) to produce tRNA‐derived small RNAs (tsRNAs), including tRNA‐derived fragments (tRFs) and tRNA‐derived stress‐induced RNAs (tiRNAs). tRFs are derived from the mature or primary tRNAs, which can be classified tRFs into five types by their derivations: tRF‐1, tRF‐2, tRF‐3, tRF‐5, and i‐tRF. tiRNAs are cleaved from the anti‐codon loop of tRNAs under stress such as oxidative stress, hypoxia, and heat shock. Previous studies have revealed that tRNA fragments participate in gene regulation in an RNAi manner or RNA silencing manner. tRFs and tiRNAs have been reported to be enriched in the body fluids. The composition and quantity of these RNAs are tightly dependent on cell type and disease condition. Stress‐induced tRFs and tiRNAs have been discovered to play a key role in cancers, metabolic diseases, and nervous disorders. , Nevertheless, the expression and clinical significance of tRFs and tiRNAs in DR remains unclear. In this study, we mainly investigated the role of tRF‐1020, a tRNA fragment, in diabetes‐induced retinal vascular complications. The result showed that the levels of tRF‐1020 expression were significantly down‐regulated in diabetic retinal vessels, vitreous samples, and aqueous humour samples of DR patients. Increased tRF‐1020 expression could retard the progression of retinal vascular dysfunction via the inactivation of Wnt signalling. tRF‐1020 may serve as a potential target for the diagnosis and treatment of DR.
Animal experiments were conducted according to the guidelines of the Statement for the Use of Animals in Ophthalmic and Vision Research (ARVO) and approved by the Animal Care and Use Committee of the author's institute (No. 20190302–42). The clinical samples were collected according to the Declaration of Helsinki. All included patients were given the informed consent.
Eight‐week‐old male C57BL/6 mice were housed under a controlled temperature with free access to water and food. The diabetes was induced with intraperitoneal (i.p.) injection of 50 mg/kg streptozotocin (STZ, 572201, Sigma‐Aldrich) or vehicle (0.1 M citrate buffer, pH 4.5) for 5 consecutive days. The levels of fasting blood glucose (FBG) were detected by an Accu‐Chek Performa glucometer system (Roche Diagnostic). The mice with FBG levels greater than 16.7 mmol/ L were considered to be hyperglycaemic.
Human retinal vascular endothelial cells (HRVECs) were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco) containing 10% fetal bovine serum, 1% penicillin–streptomycin solution (PS, Gibco), and endothelial cell growth supplement (Sigma Aldrich Corp) at 37 °C with 5% CO2 in a humidified atmosphere. One day prior to transfection, they were seeded into the 24‐well plates. When the confluency reached at 85%, HRVECs were transfected with negative control (NC) mimic, tRF‐1020 mimic, tRF‐1020 inhibitor, negative control (NC) inhibitor, or left untreated (Ctrl) using lipofectamine 6000 (Beyontime) following the manufacturer's protocol. After 12 h, the medium was replaced and cells were cultured at 37°C for additional 36 h for the subsequent studies. The sequences were shown below: tRF‐1020 mimic, 5’‐UCGGAGGCUUUGUUUU‐3′; NC mimic, 5’‐UCGCUUGGUGCACGUCGGG‐3′; tRF‐1020 inhibitor, 5’‐AAAACAAAGCCUCCGA‐3′; NC inhibitor, 5’‐UCUCCGAACGUGUCACGUU‐3′. Inhibitors and miRNA mimics were synthesized by GenePharma. During transfection, the concentrations of mimics and inhibitors were 30 nM.
Total RNAs were extracted using the TRIzol reagent (Ambion). Quantitative reverse transcription polymerase chain reactions (qRT‐PCRs) were conducted on an Applied Biosystems StepOne Real‐time system (Applied Biosystems). These RNAs were pre‐treated by rtStar™ tRF&tiRNA Pretreatment Kit (Arraystar) and then reversely transcribed into cDNAs using the rtStar™ First‐Strand cDNA Synthesis Kit (Arraystar). The reactions were carried out using the SYBR Premix Ex Taq II kit (Takara). The 2−ΔΔCt method was used for quantifying tRF‐1020 expression normalized with U6.
Cell viability was determined by Cell Counting Kit‐8 (CCK8) assay (Dojindo, Kumamoto, Japan) according to the manufacturer's protocols. Briefly, HRVECs were seeded into 96‐well plates at a density of 3 × 103 cells/well with 100 μl of 10% FBS DMEM medium. After the required treatment, 10 μl of CCK‐8 solution was added to each well of the plate and then incubated for 3 h. The absorbance at 450 nm wavelength was detected by a microplate reader (Molecular Devices).
Cell proliferation was determined by ki67 immunofluorescence staining. Briefly, HRVECs were seeded in a 24‐well plate and processed with the required treatments. They were fixed in 4% formaldehyde (Biosharp) for 15 min and then blocked with 5% bovine serum albumin (BSA) (143,066, Biofroxx) for 1 h. They were incubated with ki67 antibody (Abcam) overnight at 4°C and then incubated with Cy3‐conjugated secondary antibody (Life Technologies) for 3 h at room temperature. Cell nuclei were labelled with 4′,6‐diamidino‐2‐phenylindole (DAPI, C1002, Beyotime). The images of stained cells were observed under a fluorescence microscope.
Cell migration ability was determined using a Transwell chamber (8.0 μm pores, Corning). Briefly, HRVECs were resuspended in a serum‐free DMEM medium. 100 μl cell suspension (1╳105 cells) were added to the upper chamber and 400 μl normal complete medium was added to the lower chamber. The migrated cells were fixed in methanol and stained with 0.5% crystal violet solution (C805211, Macklin). These non‐migrated cells in the upper chamber were removed. The cells that passed the filter membrane were counted under a bright‐field microscope.
A precooled 24‐well plate was coated with the Growth Factor Reduced Matrigel (356234, BD Biosciences). After the required treatment, HRVECs were seeded on Matrigel at a cell density of 1 × 105/well cells and incubated at 37°C in 5% CO2 and 95% humidity. After 8 h culture, tube formation ability was observed under a light microscope and quantified by Image J software.
The 3′‐end biotinylated tRF‐1020 or control miRNA (RiboBio, 50 nM) were transfected into HRVECs for 12 h. The biotin‐coupled RNA complex was obtained by incubating cell lysates with the streptavidin‐coated magnetic beads (Life Technologies). The amount of DVL mRNA or GAPDH mRNA in the bound fraction was determined by qRT‐PCR assay.
Retinal vascular permeability was determined by Evans blue assay. Briefly, the mice were given intraperitoneal anaesthesia with ketamine (80 mg/kg) and xylazine (10 mg/kg). Evans blue dye (100 mg/mL) was injected into the femoral vein at the dosage of 45 mg/kg and 0.2 ml blood samples were collected. After the dye circulated for 1 h, the chest cavity was opened and perfused with citrate buffer (0.05 M, pH 3.5). After the perfusion, the eyes were enucleated and fixed in 4% paraformaldehyde for 30 min at room temperature. The retina was carefully dissected after removing the cornea, lens, and vitreous humour. For the quantitative assessment, the retina was incubated with formamide (Sigma‐Aldrich) overnight at 78°C and centrifuged at 4°C for 30 min. Evans blue dye in the supernatant was detected at the absorbance of 620 nm and 740 nm and compared with the blood samples treated similarly except for solubilization. The concentration of dye was calculated based on the standard curve of Evans blue in formamide and normalized to dry retina weight.
The eyes were enucleated and fixed in 4% paraformaldehyde for 24 h. The retinas were removed, gently shaken in water at room temperature overnight, and then incubated with 3% trypsin (BD, Difco) at 37 °C for 1 h. After trypsin digestion, the tissue was shaken gently to free vessel network, washed, and mounted on the glass slides. Dried retinal vessels were stained with periodic acid‐Schiff haematoxylin (PAS‐haematoxylin). Ten fields were randomly selected in each retina, and the number of acellular capillaries was counted in each field.
ELISA assays were conducted to detect the levels of VEGF, IL‐6, and TNF‐α in retinal lysates. The levels of VEGF, IL‐6, and TNF‐α in the supernatant were detected by the enzyme‐linked immunosorbent assay (ELISA) kit (R&D Systems) according to the manufacturer's instructions. Finally, optical density was determined at a 450 nm wavelength and plotted according to the numerical values.
Aqueous humour samples were harvested from the patients with diabetes retinopathy (DR, n = 20 eyes) and cataract (non‐DR, n = 20 eyes). Vitreous samples were harvested from the subjects with idiopathic macular hole (non‐DR, n = 20 eyes) and PDR (n = 20 eyes) during pars plana vitrectomy. These clinical samples were collected in tubes, placed immediately on ice, centrifuged for 20 min, and stored at −80 °C until use. The patients with cancers, serious infection, and active chronic inflammatory diseases were excluded.
Data were first determined for the normality using the D'Agostino Pearson omnibus normality test. For normally‐distributed data with equal variance, significant difference was evaluated by Student t‐test (2‐group comparison) or 1‐way anova followed by the post hoc Bonferroni test (multi‐group comparison). For non‐normally‐distributed data or data with unequal variances, a significant difference was evaluated by a nonparametric Mann–Whitney U test (2‐group comparison) or Kruskal–Wallis test followed by the post hoc Bonferroni test (multi‐group comparison). p < 0.05 was considered significant. All statistical analysis was conducted using GraphPad prism version 6.0.
Diabetes was induced by intraperitoneal injection of streptozotocin (STZ) for five consecutive days using the 8‐week‐old C57BL/6J mice. Retinal vessels were extracted from the mouse retinas at 1 month, 3 months, and 5 months after diabetes induction. qRT‐PCR assays revealed that the levels of tRF‐1020 expression were significantly down‐regulated in diabetic retinal vessels compared with that in the nondiabetic controls (Figure 1A). Retinal vascular endothelial cells (HRVECs) were exposed to high glucose or oxidative stress to mimic diabetic stress in vitro. Compared with the control group, high glucose (25 mM) or oxidative stress (H2O2, 50 μm) led to decreased levels of tRF‐1020 expression in HRVECs after 24‐hour or 48‐hour treatment (Figure 1B,C).
We then determined whether tRF‐1020 regulated endothelial biology in vitro. tRF‐1020 mimics or inhibitors were transfected into HRVECs to determine the potential function of tRF‐1020 in endothelial cells. The transfection of tRF‐1020 mimics led to enhanced levels of tRF‐1020. By contrast, the transfection of tRF‐1020 inhibitors led to reduced levels of tRF‐1020 (Figure 2A). CCK‐8 assays showed that the transfection of tRF‐1020 mimics significantly decreased the viability of HRVECs (Figure 2B). Ki67 staining assays showed that the transfection of tRF‐1020 mimics reduced the proliferation ability of HRVECs (Figure 2C,D). Transwell migration assays showed that the transfection of tRF‐1020 mimics decreased the migratory ability of HRVECs (Figure 2E,F). Tube formation assays showed that the relative tube length was significantly decreased in tRF‐1020 mimic‐transfected group compared with the control group (Figure 2G,H). By contrast, the transfection of tRF‐1020 inhibitors led to increased cell viability, proliferative ability, migration ability, and tube formation ability (Figure 2B–H). Taken together, the above‐mentioned evidence suggests that tRF‐1020 is an important regulator of endothelial angiogenic functions.
We further determined the role of tRF‐1020 in diabetes‐induced retinal vascular dysfunction in vivo. Diabetic C57BL/6 mice received an intravitreous injection of tRF‐1020 mimics or inhibitors to regulate the levels of tRF‐1020. Injection of tRF‐1020 mimics led to increased levels of tRF‐1020, while injection of tRF‐1020 inhibitors led to reduced levels of tRF‐1020 (Figure 3A). Evans blue assays demonstrated that compared with diabetic retinas, the injection of tRF‐1020 mimics alleviated diabetes‐induced retinal vascular leakage (Figure 3B,C). The acellular vessel is an important pathological feature of diabetic retinas. Trypsin digestion assays revealed that injection of tRF‐1020 mimics decreased the number of acellular vessels in diabetic retinas (Figure 3D,E). ELISA assays showed that the injection of tRF‐1020 mimics alleviated diabetes‐induced retinal inflammation as shown by decreased expression of VEGF, interleukin (IL)‐2, and TNF‐α (Figure 3F–H). By contrast, injection of tRF‐1020 inhibitors aggravated diabetes‐induced retinal vascular leakage, increased the number of acellular vessels in diabetic retinas, and aggravated diabetes‐induced retinal inflammation (Figure 3B–H).
We next explored the molecular mechanism of tRF‐1020 in endothelial cells. The potential target genes of tRF‐1020 were predicted by TargetScan and miRanda databases according to the presence of binding sites in the 3′‐UTR. To reduce the prediction scope, we evaluated the overlapping gene results from TargetScan and miRanda databases. We selected DVL‐2, GRAP, and Col23a1 for the subsequent study due to their roles in regulating angiogenic effects. qRT‐PCR assays showed that the expression level of DVL‐2 was significantly decreased after the transfection of tRF‐1020 mimics. By contrast, the transfection of tRF‐1020 inhibitors led to an increased expression level of DVL‐2. However, the transfection of tRF‐1020 inhibitors or mimics did not alter the expression levels of GRAP and Col23a1 (Figure 4A). We further conducted the luciferase assays to verify the direct binding between tRF‐1020 and DVL‐2. The binding sites for tRF‐1020 on the wild‐type DVL‐2 3′‐UTR are shown in Figure 4B. The luciferase activity of wild‐type DVL‐2 3′‐UTR was significantly reduced after transfection of tRF‐1020 mimic but markedly increased after transfection of tRF‐1020 inhibitor. However, altered tRF‐1020 expression did not affect the luciferase activity of mutant DVL‐2 3’‐UTR (Figure 4C). RNA pull down assays were conducted to detect the interaction between tRF‐1020 and DVL‐2. We observed a greater enrichment of DVL‐2 in tRF‐1020‐captured fraction in comparison with the negative control, biotinylated miR‐335 (Figure 4D). VEGFA is shown as a key regulator during retinal vascular dysfunction. The results show that tRF‐1020 intervention did not alter the expression levels of VEGFA, suggesting that tRF‐1020 did not have off‐sites or non‐specific targets (Figure 4E). DVL‐2 is an important regulator of Wnt/β‐catenin signalling. We next investigated whether tRF‐1020 intervention affected the expression of the downstream genes of Wnt signalling, such as c‐Myc, cyclin D1, and peroxisome proliferator‐activated receptor (PPAR) δ. The results showed that tRF‐1020 mimics reduced the expression levels of c‐Myc, cyclinD1, and PPARδ, whereas tRF‐1020 inhibitors decreased the expression levels of c‐Myc, cyclinD1, and PPARδ (Figure 4F). Collectively, these results indicate that tRF‐1020 directly regulates DVL‐2 expression in endothelial cells.
We further conducted the rescue experiments to determine whether tRF‐1020 regulated endothelial function by targeting DVL‐2 in HRVECs. Transfection of tRF‐1020 mimic led to decreased cell viability, proliferation ability, migration ability, and tube formation ability of HRVECs. DVL‐2 knockdown could achieve the similar effects as tRF‐1020 mimic on endothelial angiogenic effects (Figure 5A–D). However, DVL‐2 overexpression could partially reversed the inhibitory effects of tRF‐1020 mimic on endothelial angiogenic effects, which could lead to increased cell viability, proliferation ability, migration ability, and tube formation ability compared with tRF‐1020 mimic group (Figure 5A–D).
We next evaluated the diagnostic value of tRF‐1020 for DR. Aqueous humour is an important body fluid in the eye, which is known to be related to various ocular diseases. The baseline demographics of the patients for aqueous humour collection is shown in Table 1. The levels of tRF‐1020 were significantly down‐regulated in the aqueous humour samples of the patients with DR, but not in other patients with cataract (Figure 6A). We then determined the expression levels of tRF‐1020 in the vitreous samples of the patients with DR and non‐diabetic controls. The baseline demographics of the patients for vitreous sample collection is shown in Table 2. qRT‐PCR assays showed that the levels of tRF‐1020 were significantly down‐regulated in the vitreous samples of the patients with DR (Figure 6B). To evaluate the discriminative power of tRF‐1020 between non‐DR controls and the patients with DR, the area under the ROC‐AUC was calculated. The ROC‐AUC for tRF‐1020 for differentiating the patients with DR from non‐DR controls was 0.9086 (Figure 6C; 95% CI: 0.8537–0.9712; sensitivity 0.8963, specificity 0.8053), suggesting that tRF‐1020 is a promising marker for the diagnosis of DR. We also determined the expression levels of tRF‐1040 in the aqueous humour and vitreous samples of the patients with DR and non‐diabetic controls. The results showed that the levels of tRF‐1040 expression did not show significant difference between the patients with DR and non‐DR patients in the aqueous humour and vitreous samples (Figure 6D,E), suggesting that tRF‐1020 is a specific biomarker for the diagnosis of DR.
DR remains a leading cause of vision loss in the working‐age population. Its pathogenesis is highly complicated, in which vascular, inflammatory, and neuronal mechanisms are involved. Vascular complication mediates structural and molecular changes in DR. , However, the molecular mechanisms underlying retinal microvascular complications are not completely characterized. In this study, we demonstrate that tRF‐1020 is significantly down‐regulated in diabetic retinal vessels and retinal endothelial cells following high glucose or H2O2 stress. tRF‐1020 mimic suppresses endothelial angiogenic functions and alleviates diabetes‐induced retinal vascular complications by targeting DVL‐2‐mediated Wnt signalling. The levels of tRF‐1020 are significantly down‐regulated in the aqueous humour and vitreous samples of DR patients. Collectively, tRF‐1020 plays an important role in DR and is a potential target for the prevention/treatment of DR. Diabetes‐induced retinal vascular complications are the major causes of blindness. Endothelial cells are the constitutive part of the vasculature and are involved in the regulation of angiogenesis, haemostasis, and vascular tone. Hyperglycaemia leads to decreased levels of tRF‐1020 in endothelial cells and retinal vessels. Reduced levels of tRF‐1020 can induce abnormal proliferation, migration, and activation of endothelial cells. In the diseased condition, the unceasing or excessive proliferation and migration of endothelial cells occur in retinal microvascular complications. Reduced levels of tRF‐1020 aggravate diabetes‐induced retinal vascular complications as shown by increased vascular leakage, acellular capillary number, and inflammation response. We thus speculate that reduced tRF‐1020 level is a predisposing factor of diabetes‐induced retinal vascular complication. tRNA‐derived small RNAs (tsRNAs), including tRNA‐derived fragments (tRFs) and tRNA‐derived stress‐induced RNAs (tiRNAs), can provide additional regulatory layers of gene expression. Accumulating evidence has shown that tsRNAs play important roles in RNA silencing through the complementation between tsRNAs and target mRNAs. In addition, some tsRNAs are associated with Argonaute proteins and regulate gene expression in an RNA interference (RNAi) manner. Previous studies have shown that tRF‐3019a directly regulates FBXO47 expression via binding a site in the 3′‐UTR and Ago2. 5′‐tiRNAVal inhibits the FZD3‐mediated Wnt/β‐catenin signalling pathway in breast cancer cells via binding to the 3′‐UTR of FZD3. Herein, RNA pull down assay, luciferase activity assay, and qRT‐PCR assay indicate that tRF‐1020 exerts its biological role via a similar mechanism. It can directly bind to the 3′‐UTR of DVL‐2 and lead to the inhibition of DVL‐2 expression. Molecular mechanism study has revealed that tRF‐1020 directly regulates DVL‐2 expression in endothelial cells. DVL‐2 belongs to the dishevelled family and is a key regulator of Wnt/β‐catenin signalling. A high level of DVL‐2 can also efficiently activate Wnt/β‐catenin signalling. Increased DVL‐2 expression can aid in β‐catenin release from cytosolic Axin/GSK‐3/APC complex. , Moreover, DVL‐2 can mediate the formation of the DVL/c‐Jun/β‐catenin/TCF functional complex, leading to the stabilization of β‐catenin‐TCF interaction. All events can ultimately trigger the activation of Wnt/β‐catenin target genes. Dysregulation of Wnt signalling can mediate pathological vascular growth in proliferative retinopathy. In this study, the transfection of tRF‐1020 mimic leads to decreased levels of DVL‐2. By contrast, transfection of tRF‐1020 inhibitor leads to increased levels of DVL‐2. Altered expression of tRF‐1020 could affect the activation of Wnt signalling. This regulatory mechanism provides a novel insight into retinal microvascular complications.
In conclusion, this study demonstrates that tRF‐1020 is an important player in the progression of DR. tRF‐1020 shows great potentials for regulating endothelial angiogenic functions and treating diabetes‐induced retinal microvascular complications. Clinical data show that the levels of tRF‐1020 expression are down‐regulated in the aqueous humour and vitreous samples of DR, which can differentiate DR patients from non‐DR controls. As a result, tRF‐1020 is an appealing therapeutic target for the diagnosis and treatment of DR.
jianling du: Conceptualization (equal); data curation (equal); formal analysis (equal); funding acquisition (equal); project administration (equal); resources (equal). cong ma: Data curation (equal); formal analysis (equal); software (equal); supervision (equal); validation (equal); visualization (equal). xiang Ma: fuding aquisition (equal); project administration (equal).
The authors declare that there are no competing interests associated with the manuscript. | true | true | true |
PMC9575130 | 36073321 | Wei Guan,Songlin Li,Zhimin Zhang,He Xiao,Juan He,Jian Li,Xuan He,Jia Luo,Yun Liu,Lin Lei,Jungang Ma,Lizhao Chen,Chuan Chen | Promotor methylation status of MAPK4 is a novel epigenetic biomarker for prognosis of recurrence in patients with thymic epithelial tumors | 08-09-2022 | DNA methylation,prognosis,pyrosequencing,recurrence,thymic epithelial tumors | Abstract Background The prognosis of thymic epithelial tumors (TETs) currently relies on the commonly adopted WHO classification and Masaoka staging system, which cannot reflect the undefined biological behaviors limiting them as prognostic factors. Methods In this study, we first identified 40 genes and 179 genes, respectively that were epigenetically upregulated and silenced, corresponding to a total of 509 functionally methylated CpG sites between thymomas and thymic carcinomas by using the TCGA dataset. Results The methylation β‐values of cg20068620 in MAPK4 and cg18770944 in USP51 were significantly associated with recurrence‐free survival (RFS). In the independent validation cohort, only WHO classification and methylation β‐values of cg20068620 in MAPK4 were independent prognostic factors for RFS in Chinese patients with TETs. A linear weighted model including these two factors was used to calculate the recurrence risk score (RRS). Time‐dependent ROC curve analysis revealed that RRS was overwhelmingly superior to WHO classification for predicting 3‐, 5‐, and 10‐year RFS and Masaoka stage for 3‐ and 5‐year RFS. Conclusions These results suggested that the methylation site cg20068620 in MAPK4 can improve the accuracy of the WHO classification alone regarding the prognostic value of TETs recurrence. | Promotor methylation status of MAPK4 is a novel epigenetic biomarker for prognosis of recurrence in patients with thymic epithelial tumors
The prognosis of thymic epithelial tumors (TETs) currently relies on the commonly adopted WHO classification and Masaoka staging system, which cannot reflect the undefined biological behaviors limiting them as prognostic factors.
In this study, we first identified 40 genes and 179 genes, respectively that were epigenetically upregulated and silenced, corresponding to a total of 509 functionally methylated CpG sites between thymomas and thymic carcinomas by using the TCGA dataset.
The methylation β‐values of cg20068620 in MAPK4 and cg18770944 in USP51 were significantly associated with recurrence‐free survival (RFS). In the independent validation cohort, only WHO classification and methylation β‐values of cg20068620 in MAPK4 were independent prognostic factors for RFS in Chinese patients with TETs. A linear weighted model including these two factors was used to calculate the recurrence risk score (RRS). Time‐dependent ROC curve analysis revealed that RRS was overwhelmingly superior to WHO classification for predicting 3‐, 5‐, and 10‐year RFS and Masaoka stage for 3‐ and 5‐year RFS.
These results suggested that the methylation site cg20068620 in MAPK4 can improve the accuracy of the WHO classification alone regarding the prognostic value of TETs recurrence.
Thymic epithelial tumors (TETs) are the most common epithelial neoplasms of the anterior mediastinum. The World Health Organization (WHO) classification and Masaoka staging are the most commonly used prognostic factors for TETs because they reflect their histological types, clinical findings, and prognosis. , , The WHO classification system divides TETs into thymomas (type A, AB, B1, B2, and B3) and thymic carcinoma based on the morphology of epithelial tumor cells, degree of atypia, and relative proportion of the nontumoral lymphocytic component. In contrast, the Masaoka staging system divides TETs into four stages based on invasiveness. However, many biological behaviors of TETs remain unclear. Thus, the current WHO classification and Masaoka staging system cannot reflect these undefined biological behaviors of TETs, limiting them as prognostic factors for TETs. For example, even with complete resection, some TET patients still experience metastatic or local recurrence, which renders significant obstacles to the long‐term survival of TET patients. Unfortunately, the recurrence mechanism remains unclear, and no biomarkers can accurately predict recurrence in TETs. Tumor recurrence is a complex process that involves many factors, including aberrant DNA methylation. DNA methylation is an essential epigenetic mechanism that regulates gene expression. Thus, abnormal DNA methylation leads to abnormal gene expression, which leads to intracellular signal pathway disorders and ultimately tumor development and progression, including recurrence. , Importantly, dysregulated DNA methylation is also detected in TETs, and several gene methylation statuses are closely correlated with the prognosis of TETs, , suggesting that abnormal DNA methylation plays a crucial role in TET progression and may be a prognostic biomarker. This study aimed to identify methylation markers that can predict TET recurrence. Here, we demonstrated different DNA methylation patterns between thymomas and thymic carcinomas, and different methylation levels led to differences in gene expression and signaling pathways between thymomas and thymic carcinomas. In addition, we demonstrated that the WHO classification and methylation site cg20068620 in MAPK4 are independent predictors of recurrence in TET patients. Significantly, the combination of the WHO classification and methylation site cg20068620 in MAPK4 can more accurately predict the recurrence of TET patients. Additionally, we identified that the Masaoka stage could predict the recurrence of thymomas.
The methylation dataset TCGA.THYM.sampleMap/Human Methylation450 (version 2017‐09‐08) of TETs was downloaded from UCSC (https://xenabrowser.net/datapages) and used to identify differentially methylated sites between the thymoma and thymic carcinoma groups. This dataset included 113 cases of thymoma and 11 cases of thymic carcinoma, and the clinicopathological characteristics of the patients are shown in Table 1. A raw count matrix of gene‐level RSEM values from 120 TET tumor samples was downloaded from http://gdac.broadinstitute.org/runs/stddata_2015_11_01/ and used to identify differentially expressed genes (DEGs) between thymoma and thymic carcinoma. The clinicopathological characteristics of the 120 patients are shown in Table 2.
In total, 392 653 DNA methylation probes were included for methylation site analysis after removing probes with missing values. For each probe, the Wilcoxon rank‐sum test was used to evaluate differences in methylation β‐values between thymoma and thymic carcinoma. The Bonferroni procedure adjusted per test p‐values for multiple comparisons. The annotation for probes was performed by using the package “IMA” (Illumina methylation analyzer, version 3.1.2) and the annotation file “fullannotInd.rda”. DEGs in the comparison of thymic carcinomas versus thymomas were determined with the R/Bioconductor limma package after filtering genes with reading counts <10 in at least 80% of cases, leaving 13,581 genes included in the analysis. DNA hypermethylation or hypomethylation events that could functionally regulate mRNA expression were respectively identified through the following criteria: (1) probe at the promoter or the first exon region of one gene (TSS1500, TSS200, 1stExon), (2) the mean methylation in thymic carcinomas increased >50% compared with that in thymomas with adjusted p < 0.05 as well as a mean methylation in thymic carcinoma >30%, and (3) log2Ratio < −1 and FDR <0.05 for the corresponding gene; or (1) probe at the promoter or the first exon region of one gene (TSS1500, TSS200, 1stExon), (2) the mean methylation in thymic carcinomas decreased >50% compared with that in thymomas with adjusted p < 0.05 as well as a mean methylation in thymomas >30%, and (3) log2Ratio >1 and FDR <0.05 for mRNA expression of the corresponding gene. MSigDB c2 gene set was used to infer enrichment of genes that were potentially impacted by hypermethylation or hypomethylation with R package “clusterProfiler”. Because none of the clinicopathological characteristics, including sex (male vs. female), age, WHO classification (type C vs. type A‐B3), Masaoka stage (III–IV vs. I‐II) and adjuvant radiotherapy (yes vs. no), were significantly associated with RFS in the TCGA thymoma datasets revealed by univariate Cox regression, only univariate Cox regression was used to evaluate the prognostic value for each of 509 identified functional CpG sites. Only probes that showed significance with crude p‐values < 0.05 were identified as candidate probes for further analysis. Hierarchical clustering and visualization of methylation β‐values of 509 functional CpG sites across 124 TET cases was carried out through the “aheatmap” function in the R package “NMF” (version 0.1.3).
Overall, 95 patients with histologically confirmed thymoma or thymic carcinoma were enrolled and hospitalized between October 2013 and October 2016 for thoracic surgery at the Daping Hospital of the Army Medical University. This study was approved by the Medical Ethics Committee of Daping Hospital. Written informed consent was obtained from all patients prior to their enrollment. The clinicopathological characteristics of the patients are shown in Tables 2 and 3. Genomic DNA was extracted from formalin‐fixed, paraffin‐embedded (FFPE) TET tissues using the QIAamp DNA FFPE Tissue kit (Qiagen). The DNA concentration and purity were determined with a spectrophotometer (NanoDrop2000, Thermo Scientific). Bisulfite conversion of 500 ng purified DNA in each sample was performed with an EZ DNA Methylation‐Gold kit according to the manufacturer's instructions (cat. no. D5006, Zymo Research Corporation, Orange). The bisulfite‐converted DNA was amplified with TaKaRa EpiTaq HS (cat. no. R110A, Takara Biomedical Technology [Beijing] Co., Ltd) with the following reaction: 10 ng bisulfite‐treated DNA, 0.4 μM forward primer and reverse primer, 2.5 μl 10 × EpiTap PCR Buffer, 2.5 mM MgCl2, dNTP mixture (0.264 mM each), and EpiTap HS (0.025 U/μl) in 25 μl per reaction. The following thermal cycle conditions were used: denaturation at 98°C for 10 s, annealing at 55°C for 30 s, extension at 72°C for 30 s executed for 35 cycles followed by extension at 72°C for 1 min, and hold at 4°C. The amplicons were then subjected to pyrosequencing with PyroMark Q96 (Qiagen). All primers used are shown in Table S1.
A weighted model was constructed for the prognostic model. First, the prognostic values of sex (male vs. female), age, history of myasthenia gravis (yes vs. no), WHO classification (C vs. A‐B3), Masaoka stage (III–IV vs. I–II), adjuvant radiotherapy (yes vs. no), adjuvant chemotherapy (yes vs. no) and methylation β‐values in two candidate CpG sites, cg20068620 and cg18770944, were determined using univariate Cox regression. Only factors significantly associated with RFS were included in the multivariable Cox regression with a stepwise forward selection procedure for the identification of independent prognostic factors in which one covariate was included with criteria p < 0.05 and excluded with criteria p > 0.10 based on the likelihood ratio test. The proportional hazards assumption for the Cox proportional hazards regression model was assessed through the Schoenfeld residuals test. The recurrence risk score (RRS) was constructed using the linear predictor of the finalized model. The whole cohort was dichotomized into low‐ and high‐risk subgroups by median RRS. All β values and other continuous variables are represented by the median values and interquartile ranges and visualized with box plots. The differences in methylation β‐values in two candidate CpG sites between thymoma and thymic carcinoma were evaluated by the Kruskal–Wallis test. The cutoff value of methylation level that was used to define low‐ and high‐methylation subgroups with maximum log‐rank statistics in terms of RFS was determined by function “surv_cutpoint” in R package “survminer”. The Kaplan–Meier method and the log‐rank test were used to compare the RFS between low‐ and high‐methylation subgroups or low‐ and high‐risk subgroups. The predictive efficiency of RRS, Masaoka's stage and WHO classification for 3‐, 5‐ and 10‐year RFS was determined with time‐dependent ROC curve analysis using the “time ROC” function. Comparisons between two time‐dependent AUCs were performed with the “compare” function embedded in the R package “timeROC” (version 0.3 published in 2015‐03‐25). All other statistical analyses were performed using SPSS 17.0 (IBM SPSS). All tests were bilateral, and p < 0.05 was considered statistically significant.
Because a previous report showed that recurrences occurred more frequently in thymic carcinomas than thymomas, we first identified differentially methylated sites between thymomas and thymic carcinomas using the TCGA dataset. In total, 17 384 probes were identified as differentially methylated sites between thymic carcinomas and thymomas, of which 9021 probes were annotated (Figure 1a, Table S2). Among them, 1530 CpG sites were hypomethylated and 7491 CpG sites hypermethylated in thymic carcinomas, and these CpG sites covered 3460 genes (Table S2). We used Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analysis to demonstrate that these genes are closely associated with several signaling pathways, including the neuroactive ligand‐receptor interaction (hsa04080, adjusted p = 1.86 × 10−15), MAPK signaling pathway (hsa04024, adjusted p = 1.60 × 10−5), and cAMP signaling pathway (hsa04024, adjusted p = 1.20 × 10−10) (Figure 1b, Table S3). Next, we determined the differentially expressed genes between thymomas and thymic carcinomas from the TGCA dataset. The results showed that 2267 genes were differentially expressed between thymomas and thymic carcinomas (Table S4). Among these genes, 1229 were downregulated, and 1038 were upregulated in thymic carcinomas compared to thymomas. After overlapping with genes that were found to have differentially methylated CpG sites in their promotor or the first exon regions, 40 genes and 179 genes were considered epigenetically upregulated and silenced, respectively, corresponding to a total of 509 functionally methylated CpG sites (Tables S5 and S6). The hierarchical clustering of these methylation β‐values across all 124 patients is shown in Figure 1c. Enrichment analysis revealed that epigenetically silenced by hypermethylation genes were mainly involved in tumorigenesis such as mammary stem cell, gastric cancer and prostate cancer (Table S7, Figure S1a). The genes impacted by hypomethylation were enriched in more diverse pathways, including thyroid cancer, TNF pathway, and glioblastoma mesenchymal (Table S7, Figure S1b). This analysis revealed distinct methylation profiles that distinguish thymic carcinomas from thymomas. These findings also suggest that distinct methylation profiles may contribute to the malignant behavior of thymic carcinomas by dysregulating gene expression and signaling pathways.
In order to identify candidate methylation sites that had a potential impact on prognosis, we used univariate Cox regression to identify methylation sites that were closely related to RFS in TETs from 509 functional methylation sites. The results showed that 52 CpG sites were significantly associated with RFS in TETs (Table S8). Based on their possible involvement in the progression of carcinomas, as revealed by previously published articles, , cg20068620 in MAPK4 and cg18770944 in USP51 were ultimately selected as candidate methylation sites for further analysis (Table S9). The methylation β‐values of 0.2030 in the cg20068620 CpG site and 0.3636 in the cg18770944 CpG site were chosen as cutoff values to categorize patients into low and high methylation subgroups by using curv_cutpoint in terms of RFS. Patients in the low methylation subgroup exhibited extremely superior RFS to those with high methylation in the cg20068620 CpG site (HR = 10.64, 95% CI: 1.309–86.42, p = 0.027, Log‐rank χ2 = 7.421, p = 6.448 × 10−3). This result was more pronounced in the cg18770944 CpG site (HR = 16.94, 95% CI: 3.411–84.17, p = 0.00054, log‐rank χ2 = 20.878, p = 4.894 × 10−6) (Figure 2a–b). Moreover, the whole population was further divided into three groups: patients with low methylation levels in both two candidate CpG sites were categorized into the low‐risk group, patients with both high methylation levels into the high‐risk group and the rest into the intermediate‐risk group. This stratification had a strong prognostic capacity (log rank χ2 = 24.839, p = 4.037 × 10−6) (Figure 2c). Moreover, the prognostic efficacy of these two candidate CpG sites was further evaluated as a linear combination of coefficients of Cox regression of two CpG sties and corresponding β‐values with the formula: 7.861 × methylation β‐value of cg20068620 + 6.778 × methylation β‐value of cg18770944. Univariate Cox regression showed that the linear combination value was significantly associated with RFS (HR = 2.454, 95% CI: 1.436–4.193, p = 0.00103). After adjusting for age, sex, WHO classification, Masaoka stage, and adjuvant radiotherapy, the linear combination value as a continuous covariate was the only independent prognostic factor for RFS (HR = 2.728, 95% CI: 1.278–5.823, p = 0.0095) (Table S10). Thus, these two candidate methylation sites were finally chosen for further validation.
The results observed from the TCGA dataset were further confirmed in our patient cohort. Our data show that the recurrence rate of thymic carcinomas was 75.0% (9/12), which was significantly higher than the 22.9% recurrence rate of thymomas (19/83). The 3‐, 5‐, and 10‐year RFS rates were 98.6, 92.8, and 50.4%, respectively, in thymoma patients and 83.3, 37.4, and 16.7%, respectively, in thymic carcinoma patients. Consistently, the RFS time was significantly shorter in thymic carcinoma patients than in thymoma patients (log‐rank p < 0.001), suggesting that thymic carcinoma recurs more frequently than thymoma. Next, our patient cohort further confirmed the correlation of cg20068620 in MAPK4 and cg18770944 in USP51 with RFS of TETs. Consistent with TCGA dataset analysis results, our data also showed that the β values of both cg20068620 in MAPK4 and cg18770944 in USP51 were significantly increased in patients with thymic carcinoma compared to patients with thymomas (median (IQR), cg20068620: 65 (42–86) versus 20 (16–45), p < 0.001; cg18770944: 71 (62–99) versus 40 (25–66), p < 0.001). We also indicated significantly increased β‐values of cg20068620 in MAPK4 and cg18770944 in USP51 in advanced stage (Masaoka stage III–IV) compared to early stage (Masaoka stage I–II) (median (IQR), cg20068620: 59 (20–64) versus 20 (14–29), p < 0.001; cg18770944: 66 (40–71) versus 32 (21–62), p < 0.001). To illustrate the prognostic efficiency, the entire population was categorized into two groups with low or high methylation levels according to optimal cut‐off β‐values 0.51 of cg20068620 and 0.67 of cg18770944, respectively. As shown in Figure 2d–f, in line with the results obtained from TCGA THYM dataset, the patient with low methylation levels in cg20068620 or cg18770944 had superior RFS than those with high methylation levels (cg20068620: Log‐rank χ2 = 83.260, p = 7.195 × 10−20, HR = 94.57, 95% CI: 12.8–701.3, p = 8.34 × 10−6, cg18770944: Log‐rank χ2 = 49.465, p = 2.019 × 10−12, HR = 9.867, 95% CI: 4.524–21.52, p = 8.73 × 10−9). The patients with high methylation in the both CpG sites had the worse RFS (χ2 = 85.531, p = 2.674 × 10−19). Taken together, these results suggested that hypermethylation of cg20068620 in MAPK4 and cg18770944 in USP51 is closely associated with the aggressiveness of TETs. In addition, univariate Cox regression analysis revealed that age, WHO classification, Masaoka stage, adjuvant radiotherapy, adjuvant chemotherapy and two candidate methylation sites (cg20068620 and cg18770944) were significantly associated with RFS (Table 4). However, multivariable Cox regression with forwarding selection for covariates showed that only WHO classification and methylation β‐values of cg20068620 in MAPK4 were independent prognostic factors for RFS in TETs (Table 4). Therefore, the recurrence risk score (RRS) was calculated using these two factors as a weighted linear combination: 1.468 × WHO classification (C vs. A‐B3) + 0.097 × methylation β‐value in cg20068620. The median value and IQR of RRS in the whole validation set was 2.62 (1.55–5.72). The whole population of the validation set was categorized into low‐ and high‐risk subgroups according to the median value of RRS. This grouping had robust discriminative efficiency for the prognosis of RFS. The RFS in the high‐risk group was significantly shorter than that in the low‐risk group (median RFS: 84.3 months vs. not reached, log‐rank p = 1.04 × 10−9) (Figure 3a). In particular, all 28 recurrence events occurred in the high‐risk group. Multivariable Cox regression with forward selection for covariates verified that RRS was the only prognostic factor for RFS among all clinicopathological characteristics (HR = 2.718, 95% CI: 2.005–3.685, p = 1.173 × 10−10). Furthermore, time‐dependent ROC curve analysis revealed that RRS was overwhelmingly superior to WHO classification for predicting 3‐, 5‐, and 10‐year RFS and Masaoka stage for 3‐ and 5‐year RFS (Table 5, Figure S2a–c). These results suggested that integrating the β values of cg20068620 in MAPK4 into the commonly adopted clinical prognostic factors can yield a more precise prognosis in patients with TETs.
Most patients with TETs are diagnosed with thymomas, some patients with thymomas still experience recurrence, and no biomarkers can predict the recurrence of thymomas. Thus, we used our cohort (n = 83 thymoma patients) to investigate the possibility of cg20068620 in MAPK4 and cg18770944 in USP51 as predictors of thymoma recurrence. Multivariable Cox regression with forwarding selection demonstrated that Masaoka stage (III–IV vs. I–II) was the only prognostic factor for RFS in thymomas (HR = 26.289, 95% CI: 5.986–115.46, p = 1.491 × 10−5) among the clinicopathological features that showed a significant association with RFS in univariate Cox regression analysis. The combination of the Masaoka stage and cg18770944 provided more prognostic information than Masaoka stage alone based on the likelihood ratio test (Table 6). However, Cox regression with the Masaoka stage and cg20068620 as covariates showed that the Masaoka stage was no longer significant and that cg20068620 was strongly associated with RFS, suggesting that cg20068620 alone could more precisely predict RFS in this clinical set (Table 6). The patients with low‐ and high‐risk in this subgroup, according to median values of RRS calculated from model 1 and model 2, showed significant differences in RFS (Figure 3b–c). However, the AUCs for predicting 5‐ and 10‐year RFS of model 1 were not significantly higher than those of Masaoka stage alone (mean 95% CI: 0.943 (0.862–1.000) versus 0.847 (0.789–0.904), p = 0.055; 1.000 (1.000–1.000) versus 0.951 (0.882–1.000), p = 0.303). The same conclusion held true for model 2 (mean 95% CI: 0.909 (0.815–1.000) versus 0.847 (0.789–0.904), p = 0.356; 1.000 (1.000–1.000) versus 0.951 (0.882–1.000), p = 0.296).
The WHO classification divides thymic epithelial tumors into thymomas and thymic carcinoma histologically, while clinical data show that thymic carcinoma is more aggressive than thymomas. For example, the 5‐year survival rate of patients with thymomas is approximately 78%; however, in patients with thymic carcinomas, it is only approximately 40%. According to Khandelwal et al., the recurrence rate (including metastasis) of thymoma patients is 20%, while the recurrence rate of thymic carcinoma patients is as high as 67%. Consistent results were also observed in the present study. Notably, such clinical outcome differences between thymic carcinoma and thymomas may be caused by their different molecular biology. Comprehensive genomic analysis suggests that thymic carcinoma is molecularly distinct from thymomas. Enkner et al. also observed genetic differences between thymic carcinomas and thymomas. In the present study, we demonstrated that thymic carcinoma and thymomas have distinct DNA methylation profiles and that distinct DNA methylation may cause different gene expression patterns and signaling pathway activation. Consistently, Hirose et al. reported that aberrant DNA methylation was more frequent in thymic carcinomas than in thymomas. Altogether, these findings suggested that thymic carcinomas are highly aggressive tumors and that aberrant DNA methylation may be closely associated with the malignant behavior of thymic carcinoma. Studies have shown that metastatic or local recurrence occurs in 5%–31% of TET patients, suggesting that recurrence is one challenge of TET treatment. , Thus, reliable and accurate predictive markers to identify which subsets of TET patients are vulnerable to recurrence are urgently needed. Previously, Marx et al. reported that the WHO classification is a prognostic factor for recurrence and survival in TET patients, but the findings are controversial. Other studies have shown that WHO classification does not significantly predict TET recurrence. In this study, we used the TCGA dataset and the cohort analysis demonstrated that the WHO classification and DNA methylation site cg20068620 in MAPK4 are independent predictors of RFS in TET patients, and combining the WHO classification and cg20068620 in MAPK4 can more accurately predict the recurrence of TET patients than the WHO classification alone. Notably, our independent cohort analysis showed that all recurrent cases occurred in the high recurrence risk group set according to the combination of the WHO classification and DNA methylation site cg20068620 in MAPK4. These findings suggest that combining the WHO classification and cg20068620 in MAPK4 may be a valuable method for predicting recurrence among TET patients. However, the sample size of this study was limited. Thus, further confirmation in larger sample size is needed before clinical application. In particular, more thymic carcinoma patients need to be included. It has been reported that several CpG sites in the MAPK4 promoter region, including cg20068620, have been associated with overall survival in low‐grade glioma in a previously published article. These results suggested that epigenetic regulation of kinases, including MAPK4, may play an essential role in the biology of carcinomas. Although the survival time of thymoma patients is long, 20% of thymoma patients experience postoperative recurrence. In the present study, we identified that Masaoka stage was closely correlated with thymoma recurrence. We also indicated that the methylation sites cg20068620 in MAPK4 and cg18770944 in USP51 were associated with thymoma recurrence, suggesting that Masaoka stage, cg20068620 in MAPK4 and cg18770944 in USP51 are useful candidates as predictors of thymoma recurrence. USP51 is identified as a ZEB1 deubiquitinase in serval reports. USP51 can stabilize ZEB1 protein to enhance epithelial‐mesenchymal transition and metastasis. , However, the impact of epigenetic regulation of USP51 on cancer progression has not been reported. Recently, several genes with hypermethylation in thymic carcinoma compared with thymoma have been identified, including GAD1, GNG4, GHST, HOXD9, and SALL3. , Soejima et al. reported that patients with TETs with high GAD1 DNA hypermethylation and high mRNA and protein expression levels had significantly shorter relapse‐free survival rates than those with low levels. In our analysis, GAD1 was also identified as a gene impacted by hypermethylation in thymic carcinoma (Table S2). Further investigation into the mechanism by which methylation of these genes regulates malignant progression was needed. In conclusion, our findings indicate that different DNA methylation patterns are closely associated with the malignancy of TETs and that the combination of the WHO classification and methylation site cg20068620 in MAPK4 is better than the WHO classification alone in terms of the prognostic value of TET recurrence. In addition, we identified that the Masaoka stage, cg20068620 in MAPK4 and cg18770944 in USP51, are valuable candidates as prognostic factors for the recurrence of thymomas.
The authors have no conflict of interest.
Click here for additional data file. Click here for additional data file. Click here for additional data file. | true | true | true |
PMC9575132 | 36056690 | Shuo Chen,Qian‐hui Li,Xi Chen,Hai‐Juan Bao,Wu Wu,Fan Shen,Bing‐Feng Lu,Ru‐Qi Jiang,Zhi‐hong Zong,Yang Zhao | SNORA70E promotes the occurrence and development of ovarian cancer through pseudouridylation modification of RAP1B and alternative splicing of PARPBP | 03-09-2022 | alternative splicing,ovarian cancer,pseudouracil modification,RAP1B,SNORA70E | Abstract The present study demonstrated for the first time that SNORA70E, which belongs to box H/ACA small nucleolar noncoding RNAs (snoRNAs) who could bind and induce pseudouridylation of RNAs, was significantly elevated in ovarian cancer tissues and was an unfavourable prognostic factor of ovarian cancer. The over‐expression of SNORA70E showed increased cell proliferation, invasion and migration in vitro and induced tumour growth in vivo. Further research found that SNORA70E regulates RAS‐Related Protein 1B (RAP1B) mRNA through pseudouracil modification by combing with the pyrimidine synthase Dyskerin Pseudouridine Synthase 1 (DKC1) and increase RAP1B protein level. What's more, the silencing of DKC1/RAP1B in SNORA70E overexpression cells both inhibited cell proliferation, migration and invasion through reducing β‐catenin, PI3K, AKT1, mTOR, and MMP9 protein levels. Besides, RNA‐Seq results revealed that SNORA70E regulates the alternative splicing of PARP‐1 binding protein (PARPBP), leading to the 4th exon‐skipping in PARPBP‐88, forming a new transcript PARPBP‐15, which promoted cell invasion, migration and proliferation. Finally, ASO‐mediated silencing of SNORA70E could inhibit ovarian cancer cell proliferation, invasion, migration ability in vitro and inhibit tumorigenicity in vivo. In conclusion, SNORA70E promotes the occurrence and development of ovarian cancer through pseudouridylation modification of RAP1B and alternative splicing of PARPBP. Our results demonstrated that SNORA70E may be a new diagnostic and therapeutic target for ovarian cancer. | SNORA70E promotes the occurrence and development of ovarian cancer through pseudouridylation modification of RAP1B and alternative splicing of PARPBP
The present study demonstrated for the first time that SNORA70E, which belongs to box H/ACA small nucleolar noncoding RNAs (snoRNAs) who could bind and induce pseudouridylation of RNAs, was significantly elevated in ovarian cancer tissues and was an unfavourable prognostic factor of ovarian cancer. The over‐expression of SNORA70E showed increased cell proliferation, invasion and migration in vitro and induced tumour growth in vivo. Further research found that SNORA70E regulates RAS‐Related Protein 1B (RAP1B) mRNA through pseudouracil modification by combing with the pyrimidine synthase Dyskerin Pseudouridine Synthase 1 (DKC1) and increase RAP1B protein level. What's more, the silencing of DKC1/RAP1B in SNORA70E overexpression cells both inhibited cell proliferation, migration and invasion through reducing β‐catenin, PI3K, AKT1, mTOR, and MMP9 protein levels. Besides, RNA‐Seq results revealed that SNORA70E regulates the alternative splicing of PARP‐1 binding protein (PARPBP), leading to the 4th exon‐skipping in PARPBP‐88, forming a new transcript PARPBP‐15, which promoted cell invasion, migration and proliferation. Finally, ASO‐mediated silencing of SNORA70E could inhibit ovarian cancer cell proliferation, invasion, migration ability in vitro and inhibit tumorigenicity in vivo. In conclusion, SNORA70E promotes the occurrence and development of ovarian cancer through pseudouridylation modification of RAP1B and alternative splicing of PARPBP. Our results demonstrated that SNORA70E may be a new diagnostic and therapeutic target for ovarian cancer.
Ovarian cancer is one of the three major malignant tumours of the female reproductive system. Its deep location and lack of specific screening strategies for early detection make the occurrence of ovarian cancer insidious, and metastasis may occur rapidly. Most patients present with distant metastasis at diagnosis, and their prognosis is poor, with a five‐year survival rate of approximately 29.2%, which seriously threatens the life and health of women. , Therefore, ovarian caner's molecular mechanism, occurrence, and development should be analysed, and suitable markers for early diagnosis and therapy should be identified. Recent research has found that although noncoding RNAs (ncRNAs) lack the ability to encode proteins, they can regulate the expression or modification of proteins from multiple angles and participate in various biological functions, such as cell proliferation and migration, thus becoming a hot spot of cancer research. NcRNAs mainly include microRNAs (miRNAs), piRNAs, long noncoding RNAs (lncRNAs), circular RNAs (circRNAs), and small nucleolar RNAs (snoRNAs). , SnoRNAs are non‐coding RNAs of about 60–300 nt in size, most of which are located in nucleoli, hence their name. According to their molecular structure characteristics, snoRNAs can follow the principle of base complementary pairing to bind to their targets and induce pseudouridylation or 2'‐O‐ribose methylation to modify rRNA, tRNA, and other RNAs. , Recent studies have shown that snoRNAs are abnormally expressed in a variety of tumours and play an important role in cancer tumorigenesis and development. , , For example, SNORA23 is highly expressed in pancreatic cancer tissues. Overexpression of SNORA23 in pancreatic ductal adenocarcinoma (PDAC) cells can mediate sequence‐specific pseudouridine acidification of ribosomal RNA and by increasing the expression level of SYNE2 (encoding spectrin repeat containing nuclear envelope protein 2), promotes pancreas carcinogenesis and development, which provides new avenues for the study of molecular mechanisms of tumorigenesis and development. However, although numerous studies have shown that ncRNAs have important functions in ovarian cancer development, we know little about snoRNAs' function and their specific regulatory mechanisms in ovarian cancer. Our team screened snoRNAs that might function in ovarian cancer development through The Cancer Genome Atlas (TCGA) database and found that SNORA70E is an unfavourable prognostic factor for ovarian cancer. There is no gene expression data of normal ovarian tissue in the TCGA database; therefore, we further used quantitative real‐time reverse transcription PCR (qRT‐PCR) to detect the expression of SNORA70E in normal ovarian tissues and ovarian cancer tissues, and found that its expression in ovarian cancer was significantly increased. Hence, we suggested that SNORA70E might promote ovarian cancer occurrence and progression; however, the potential role and molecular mechanism of SNORA70E in tumours has not been reported. SNORA70E belongs to the group of box H/ACA snoRNAs, which have two hairpin structures, the hinge region is box H (ANANNA), and the last 4–6 nucleotides at the 3′ end of the molecule are the highly conserved box ACA. In the human body, snoRNAs, dyskerin pseudouridine synthase 1 (DKC1), and the other three core proteins NHP2 ribonucleoprotein (NHP2), GAR1 ribonucleoprotein (GAR1), nucleolar protein 10 (NOP10) form the snoRNP complex, bind to target RNA (e.g., rRNA and tRNA) through base pairing, and participate in various pathophysiological processes through DKC1's function of catalysing the pseudouridylation of specific sites of the target RNA. , Recently, Karijolich et al. showed that mRNA can also be modified by pseudouracil. They further identified a number of pseudouracil synthetases that can act on mRNA, including DKC1. This suggested that box H/ACA snoRNAs might be combined with the principle of base complementation under the action of pseudouracil synthase DKC1 and to modify mRNA by pseudouridylation, thus changing the expression of target mRNAs or proteins, which might affect the development of tumours. In addition, studies have shown that some snoRNAs can also be processed into miRNA precursors, and form snoRNA‐like microRNAs (sno‐miRNAs) with a size of about 18–30 bp through the co‐function of DICER1, and argonaute (AGO), the key enzyme for miRNA processing and maturation. These sno‐miRNAs function as miRNAs, which mighty also lead to the silencing of downstream target genes. , , Therefore, the present study aimed to use in vitro cell experiments to detect changes in the cell cycle, apoptosis, proliferation, migration or invasion ability, and in vivo nude mouse experiments to analyse SNORA70E's molecular mechanism in ovarian cancer.
Patients who underwent gynaecological surgery at the Third Affiliated Hospital of Guangzhou Medical University (Guangzhou, China) provided specimens of ovarian cancer tissues (n = 70), borderline tumours (n = 12), benign ovarian tumours (n = 7), and normal ovarian tissues (n = 14) with patient consents. Samples were immediately frozen in liquid nitrogen. Two pathologists confirmed the tissue specimens independently. The Ethics Committee of the Guangzhou Medical University approved the study (No. 2020‐053).
Jennio Biotech (Guangzhou, China) or ATCC (Manassas, VA, USA) provided the human ovarian cancer cell lines (SKOV3, OVCAR3, CAOV3, and A2780). Dulbecco's modified Eagle's medium (DMEM; HyClone, Logan, UT, USA) was used to grow A2780 cells. McCoys' 5A medium was used to grow SKOV3 cells. Roswell Park Memorial Institute (RPMI)‐1640 medium (HyClone) was used to grow OVCAR3, and CAOV3 cells. Penicillin/streptomycin (100 U/ml) and 10% fetal bovine serum (FBS) were used to supplement all media. Cells were grown at 37°C in a 5% CO2 atmosphere. Transfection of plasmids, ASOs (amido‐bridged nucleic acid‐flanked antisense oligonucleotides), and small interfering RNA (siRNA) used Lipofectamine 3000 following to the manufacturer's protocol (Invitrogen, Carlsbad, CA, USA). A SNORA70E overexpressing plasmid was used to upregulate SNORA70E expression (TCTCTGGCTAACTAGAGAACCCACTGCTTACTGGCTTATCGAAATTAATACGACTCACTATAGGGAGACCCAAGCTGGCTAGCGTTTAAACGGGCCCTCTAGACTGCAACCAATTAAGCCGACCTAGTTCCTTTCCTCTTTGGGGCCTGGTGTTCAATAGCTGCAAACAGCAGCTTCCTTGGTAGTGTATGCAGCCTGTTTCTTGTATGGGTTGCTCTAAAGGACCTTGGAGACAGCCTCTAGATAGTTAAACCGCTGATCAGCCTCGACTGTGCCTTCTAGTTGCCAGCCATCTGTTGTTTGCCCCTCCCCCGTGCCTTCCTTGACCCTGGAAGGTGCCACTCCCACTGTCCTTTCCTAATAAAATGAGGAAATTGCATCGCATTGTCTGAGTA), and a SNORA70E ASO (GTGTTCAATAGCTGCAAACA, Ruibo, China) was used to knockdown SNORA70E expression. The sequence of si‐DKC1 was AGCCTGGATTCGACGGATA, for si‐RAP1B (targeting RAP1B (encoding Ras‐related protein Rap‐1b)) was ACCTAGTGCGGCAAATTAA.
Cell viability was determined using a Cell Counting Kit‐8 (CCK‐8) assay. Cells (2000 per well) were seeded in 96‐well plates and then CCK‐8 solution (10 μl per well) (Bintech Co., Ltd.) was added at 0, 24, 48, and 72 h. The plates were incubated for 2 h at 37°C in 5% CO2. A microplate spectrophotometer (BioTek Instruments, Winooski, VT, USA) was then used to measure the absorbance at 450 nm.
Cells were seeded in 96‐well plates, after which cell proliferation was measured using a Click‐iT™ Plus EdU Alexa Fluor™ 555 Imaging Kit (Invitrogen) under the guidance of manufacturer's protocol.
Collected cells were resuspended in 100 μl of 1 × buffer with fluorescein isothiocyanate (FITC)‐labelled annexin V and PE or 7AAD and PE (5 μl each, BD) and incubated for 15 min in the dark. Thereafter, cells were added with buffer (400 μl) and flow cytometry was carried out within 1 h to determine the rate of cell apoptosis.
A wound‐healing assay was used to assess cell migration. Cells at 6 × 105 cells per well were seeded in 6‐well plates. After the cells grew to 80% confluence, a vertical wound was made in the confluent cell layer using a 200‐μl pipette tip. Three washes with phosphate‐buffered saline (PBS) were used to remove excess cells. To each well, 2 ml of serum‐free medium was added and the cells were incubated at 37°C in 5% CO2. Images of the wound were acquired under an optical microscope at 0, 24, and 48 h. Image J software (National Institutes of Health, Bethesda, MD, USA) was then used to calculate the size of the wound area. The following equation was used to calculate the wound‐healing rate: (original wound area—wound area at each time point)/original wound area × 100%.
Transwell chambers were used to assess cell invasion. Serum‐free medium was used to dilute Matrigel matrix (BD Biosciences, San Jose, CA, USA) 1:10, which was added to the upper Transwell chamber. The whole Transwell apparatus was placed in an incubator for 3–4 h for coagulation. Thereafter, 50,000 cells in serum‐free medium (200 μl) were added to the upper chamber, and as a chemoattractant, complete cell culture medium (600 μl) was placed the lower chamber. After incubation for 48 h, three washes of the chambers with PBS were performed, and the cells in the lower chamber were fixed for 15 min using 5% paraformaldehyde, before being stained for 15 min with crystal violet solution. Finally, cell invasion was evaluated by counting the cells under a microscope (Olympus, Tokyo, Japan).
The TRIzol reagent (Takara, Shiga, Japan) was used to extract total RNA from cells and tissues. The TRIzol mixture (1 ml) was added with chloroform (200 μl), mixed, and centrifuged. The upper aqueous phase was removed into a fresh tube, and the RNA was precipitated by the addition of an equal volume of isopropanol. The sample was centrifuged, the supernatant was discarded, and the pellet was washed using 75% ethanol. After drying, diethyl pyrocarbonate (DEPC) water was used to dissolve the precipitate. An ultraviolet spectrophotometer (Unico, Shanghai, China) was used to determine the OD value at 260 nm to calculate the RNA concentration. Then, cDNA was synthesized from the RNA using reverse transcription following the supplier's protocol (Takara). The cDNA was then used as the template in a real‐time quantitative PCR reaction, performed using a SYBR PrimeX EX‐TAQ Patent II Kit (Takara). Finally, to assess the target gene's relative expression, the cycle threshold (Ct) values of the control GAPDH (encoding glyceraldehyde‐3‐phosphate dehydrogenase) gene, and the target gene were compared according to the 2−ΔΔCt method (GenePharma, Shanghai, China).
Radioimmunoprecipitation assay (RIPA) buffer with protease inhibitors was used to lyse the cells. The total proteins were quantified, and 40 μg of denatured protein was subjected to 10% sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE), and then the separated proteins were electro‐transferred onto a methanol pre‐activated polyvinylidene difluoride (PVDF) membrane. After washing the membrane for 2 min with Tris‐buffered saline and Tween 20 (TBST), 3% bovine serum albumin (BSA) was used to block non‐specific binding by incubation for 2 h at room temperature. Thereafter, primary antibodies against RAP1B, β‐catenin, phosphatidylinositol‐4,5‐bisphosphate 3‐kinase (PI3K), matrix metalloproteinase 9 (MMP9), protein kinase B (AKT), mammalian target of rapamycin (mTOR) (1:1000, Proteintech Group, Chicago, IL, USA), and GAPDH, α‐Tublin, β‐actin (1:10000, Proteintech Group) were incubated with the membrane at 4°C overnight. Next day, TBST was used to wash the membrane three times and then anti‐rabbit/mouse secondary antibodies were incubated with the membrane for 1.5 h at room temperature. After another three washes with TBST, the ECL system (Santa Cruz Biotechnology, Santa Cruz, CA, USA) was used to visualize the immunoreactive proteins on the membrane.
The RIP assay used a Magna RIP RNA‐Binding Protein Immunoprecipitation Kit (Millipore, Bedford, MA, USA). Briefly, RIP lysis buffer was used to lyse cells, and the resultant cell extract was incubated with RIP buffer comprising magnetic beads conjugated with human anti‐AGO2 antibodies, anti‐DKC1 antibodies, anti‐RAP1B antibodies, or normal rabbit IgG (negative control). Thereafter, the samples were incubated with proteinase K to digest the proteins. Then, we isolated the immunoprecipitated RNA and subjected it to qRT‐PCR analysis.
High performance liquid chromatographic instruments manufactured by Waters (Milford, MA, USA) were used for detection. The chromatographic column was ECOSIL C185 μm 4.6 mm × 250 mm, and the mobile phase was 2.5 mM ammonium acetate (PH = 4.0) buffer containing 5% acetonitrile. The detection wavelength was 263 nm, the flow rate was 0.8 ml/min, the column temperature was 25°C, and the sample injection was 10 μl. A pseudouridine reference standard was accurately weighed, and double distilled water was added to prepare a solution containing 2.7 mg per 1 ml, which was used as the reference standard stock solution. An appropriate amount of pseudouridine stock solution was accurately aspirated and successively diluted with re‐distilled water, to prepare a range of standard solutions containing pseudo‐uridine, and 10 μl was precisely aspirated, injected, and the chromatogram was recorded. The chromatographic conditions were observed, and the chromatograms and peak areas were recorded. The standard curve was plotted with peak area (A) as the vertical coordinate and injection concentration (μg/ml) as the horizontal coordinate. A 500 μl sample was centrifuged at 12,000 × g for 20 min at 4°C, and the supernatant was collected. The sample was measured according to the above chromatographic conditions, the chromatographic diagram and the peak area were determined, and the content of the pseudouridine in the sample was calculated using the results for the standards and a regression equation.
CAOV3 cells were transfected with the SNORA70E overexpression plasmid or empty vector. Total RNAs were isolated using Trizol (Takara, Japan). For the RNA sample preparations, 1 μg RNA per sample was used as input material. An Neb Next Ultra RNA Library Prep Kit for Illumina (Neb, Ipswich, MA, USA) was used to generate the sequencing libraries following the supplier's protocol. In addition, an index code was added to determine the sequence attributes of each sample. Briefly, mRNA was purified, cleaved at elevated temperature, and cDNA was synthesized. An AMPure XP system (Beckman Coulter, Beverly, CA, USA) was used to purify the library fragments, and those with a length of 150–200 bases were screened. PCR was carried out using universal primers, index primers, and polymerase. Finally, the AMPure XP system was used to purify the PCR products and an Agilent Bioanalyzer 2100 system (Agilent, Santa Clara, CA, USA) was used to evaluate the quality of the library. The cBot Cluster Generation System was used to cluster the index‐coded samples employing a TruSeq PE Cluster Kit v3‐cBot‐HS (Illumina, San Diego, CA, USA) following the supplier's protocols. Thereafter, the Illumina NovaSeq platform was used to sequence the prepared libraries, generating 150 bp paired end reads.
FISH assay was performed following the manufacturer's instructions (GenePharma, Shanghai, China), as we had introduced before. The probe used for SNORA70E was 5′–GGGTAAAACTCCCTACCTGGTGTCTCCGT–3′ (Hanbio Biotechnology, China).
Vital River Laboratories (Beijing, China) supplied the BALB/c nude mice, which were housed in a specific pathogen‐free (SPF) environment. SNORA70E overexpressing CAOV3 cells or control CAOV3 cells (1 × 107 in 150 μl of FBS‐free media) were injected subcutaneously into 5‐week‐old female mice to establish a subcutaneous dissemination model (n = 6). Mice were euthanized at various times post‐injection, the tumour nodes were resected, and their volumes were measured. Besides, OVCAR3 cells (5 × 106 in 150 μl of FBS‐free media) were injected subcutaneously into 5‐week‐old female mice (n = 12), when most of the tumour volume reached 5 mm in diameter, the mice were randomly divided into two treatment groups, followed by injection with 250 mg/kg/week of ASO‐SNORA70E or ASO‐NC administered subcutaneously for 4 weeks. The Guangzhou Medical University Animal Care and Use Committee approved the animal experiments. The Guide for the Care and Use of Laboratory Animals (published by the National Institute of Health) was followed when carrying out the animal experiments.
SPSS 22.0 software (IBM Corp., Armonk, NY, USA) was used to perform the statistical analyses, the Mann–Whitney U test and paired samples t‐test were used to compare the means of different groups. Three repeats were performed for each experiment, and the overall parameters for each group of data were represented by the mean ± SD. Statistical significance was accepted at a p‐value <0.05.
According to the TCGA database, SNORA70E is an unfavourable prognostic factor for ovarian cancer (Figure 1A). Thus qRT‐PCR was used to assess the expression of SNORA70E in ovarian carcinoma tissues, borderline tumour tissues, benign tissues, and normal ovarian tissues. SNORA70E expression was significantly higher in epithelial ovarian cancer tissues than in borderline tumour tissues, benign tissues, and normal ovaries (Figure 1B, p < 0.05, Table S1). Compared with that in stage I disease, SNORA70E expression was higher in stages II–IV (Figure 1C, p < 0.05). In addition, compared with that in pathologically well classified disease, the SNORA70E expression was higher in the moderately and poorly pathological classified disease (Figure 1D, *p < 0.05), besides, SNORA70E expression was higher in ovarian serous adenocarcinoma than in the other types (Figure 1E, *p < 0.05, Table S2). These results indicated that SNORA70E participates in ovarian cancer tumorigenesis and progression, and might be related to poor prognosis.
In order to choose suitable cell lines for in vitro experiments, qRT‐PCR was used to assess the expression of SNORA70E in four ovarian cancer cell lines. SNORA70E expression was lowest in CAOV3 cells and highest in OVCAR3 cells (Figure 1F). Therefore, we used CAOV3 cells for transfection with SNORA70E overexpression plasmid and OVCAR3 cells for transfection with SNORA70E ASO. The transfection efficiency was confirmed using qRT‐PCR (Figure 1G,H, *p < 0.05).
To assess invasion, migration, and proliferation of ovarian cancer cells, Transwell, wound healing, cell apoptosis, and CCK‐8 assays were used. The overexpression of SNORA70E in CAOV3 cells induced cell proliferation (Figure 2A,B); decreased cell apoptosis (Figure 2C, p < 0.05); and induced cell migration and invasion (Figure 2D,E, p < 0.05). When mice were injected with CAOV3 cells overexpressing SNORA70E, compared with that in the control group (Figure 2F,G), tumorigenicity was induced (Figure 2H, p < 0.05), with larger tumour volumes (Figure 2I, p < 0.05). These results indicated that SNORA70E promotes ovarian cancer tumorigenesis and progression.
We further confirmed the function of SNORA70E through designing a specific ASO, which could downregulate SNORA70E expression. Our results showed that when SNORA70E was downregulated in OVCAR3 cells, compared with those in the negative control, cell proliferation was inhibited significantly (Figure 3A, p < 0.05). There was an increases in cell apoptosis (Figure 3B, p < 0.05), but a decrease in migration and invasion (Figure 3C,D, p < 0.05). In vivo results showed that compared with that in the ASO‐NC group, tumorigenicity was reduced after ASO‐SNORD70E injection subcutaneously (Figure 3E–G, p < 0.05), with smaller tumour volumes (Figure 3H, p < 0.05). Suggesting that SNORA70E may be a new diagnostic and therapeutic target in ovarian cancer.
In order to konw the molecular mechanism of SNORA70E, we first checked the cellular lacation through FISH, we found that SNORA70E not only locates in nuclear but also in cytoplasm (Figure 4A). We further overexpressed SNORA70E in CAOV3 cells and performed western blotting. The results showed that SNORA70E overexpression increased β‐catenin, PI3K, AKT1, mTOR, and MMP9 protein levels (Figure 4B). These results indicated that SNORA70E could regulate β‐catenin, PI3K, AKT1, mTOR, and MMP9 protein expression.
Studies had reported that box H/ACA snoRNAs might combine with pseudouracil synthase DKC1 and to modify mRNA by pseudouridylation, besides, it may also process into miRNA precursors and form sno‐miRNA through the co‐function of DICER and AGO. Through RIP experiments, we found that SNORA70E combine with DICER but not with AGO2, which means that SNORA70E might have a “miRNA like” function but need further research. Besides, we found that SNORA70E bound to DKC1, the pyrimidine synthase (Figure 5A,B, p < 0.05), and we further found that after silencing SNORA70E expression, the level of pseudouridine‐modified metabolite pseudouridine was decreased significantly, while overexpression of SNORA70E in CAOV3 cells could increase the level of pseudouridine‐modified metabolite pseudouridine (Figure 5C,D, p < 0.05). These results indicated that SNORA70E may combine with DKC1 to modify mRNA by pseudouridylation, thus influence the expression of target mRNAs or proteins.
In SNORA70E‐overexpressing CAOV3 cells, silencing DKC1 (Figure 6A, p < 0.05) reduced cell viability (Figure 6B, p < 0.05), and cell proliferation (Figure 6C) promoted apoptosis (Figure 6D, p < 0.05) and inhibited cell migration and invasion (Figure 6E,F, p < 0.05), and siliencing DKC1 in OVCAR3 cells whose basic SNORA70E expression is high had the same function (Figure S1). Besides, silencing DKC1 in SNORA70E‐overexpressing CAOV3 cells also decreased β‐catenin, PI3K, AKT1, mTOR, and MMP9 protein levels (Figure 6G).
Box H/ACA snoRNAs can form complementary pairs with modified sites on the target RNA by forming a SNORNP complex with DKC1. Through BLAST sequence alignment, we found that there is a binding site between SNORA70E and RAP1B (Figure 7A). The overexpression of SNORA70E increased the RAP1B protein level significantly (Figure 7B), and RIP detection found that DKC1 can bind to RAP1B mRNA (Figure 7C, p < 0.05). We also found that both silencing RAP1B in SNORA70E overexpression CAOV3 cells and in OVCAR3 cells whose basic SNORA70E expression is high inhibited cell proliferation, migration, and invasion, and induced apoptosis (Figure 7D–H, Figure S1), and reduced the levels of RAP1B, β‐catenin, PI3K, AKT1, mTOR, and MMP9 proteins (Figure 7I). Besides, siliencing β‐catenin, PI3K, and MMP9 in SNORA70E overexpression CAOV3 cells also inhibited cell proliferation, migration, and invasion, and induced apoptosis (Figure S2). These results indicated that SNORA70E could promote the occurrence and development of ovarian cancer through pseudouridylation modification of RAP1B.
The RNA‐Seq results revealed that SNORA70E regulated the alternative splicing (AS) of PARPBP (encoding PARP1 binding protein): The 4th Exon was lost in NM_001319988.1 (PARPBP‐88), forming a new transcript NM_017915.4 (PARPBP‐15) (Figure 8A). PCR results confirmed this Exon skipping event after SNORA70E overexpression (Figure 8B–C). Further studies showed that the overexpression of PARPBP‐15 promoted ovarian cancer cell proliferation (Figure 8D, p < 0.05), migration (Figure 8F, p < 0.05), and invasion (Figure 8G, p < 0.05), inhibited cell apoptosis (Figure 8E, p < 0.05), while the overexpression of PARPBP‐88 did not. Besides, PARPBP‐15 overexpression in SNORA70E‐ASO transfected OVCAR3 cell lines could significantly reversed the inhibition ability by SNORA70E downregulation (Figure S3). Above, these results suggested that SNORA70E‐regulated alternative splicing of PARPBP induced ovarian tumorigenesis and progression.
To identify markers for the diagnosis and treatment of epithelial ovarian cancer, we screened small noncoding RNAs in the TCGA database, and found that the H/ACA box‐type snoRNA, SNORA70E, which pseudouridylates ribosomal RNAs in a sequence‐specific manner, is an unfavourable prognostic factor for ovarian cancer. We further checked SNORA70E in our own clinical samples, and found that compared with those in normal ovaries, the SNORA70E expression levels were markedly increased in epithelial ovarian cancer tissues, and was associated with FIGO stage and differentiation, which suggested that SNORA70E participates in ovarian cancer tumorigenesis and progression, and might be related to poor prognosis. We assessed SNORA70E's function in ovarian cancer cells using in vivo and in vitro assays. CAOV3 cells transfected with the SNORA70E overexpression plasmid showed reduced cell apoptosis and increased cell invasion, migration, and proliferation, besides, SNORA70E overexpression induced tumour growth in vivo. However, in ovarian cancer, what is the mechanism by which SNORA70E affects tumorigenesis and progression? Recent studies showed that gene expression can be regulated by snoRNAs via gene‐related ribosome modulation or by snoRNA‐derived miRNA‐like molecules in the cytoplasm. , Target genes are regulated by an RNA‐induced silencing complex comprising argonaute 2 (Ago2) and the miRNA. However, RIP experiments showed that SNORA70E does not combine with AGO2, but does combine with DKC1, the pyrimidine synthase. Moreover, after silencing SNORA70E expression, the level of pseudouridine‐modified metabolite pseudouridine decreased significantly. Silencing DKC1 expression in SNORA70E stably overexpressing ovarian cancer cells inhibited the increased cell proliferation and migration ability. Thus, we suggest that SNORA70E might promote ovarian cancer tumorigenesis and progression through pseudouridine modification of downstream genes. Recent studies have shown that mRNA can also be modified by pseudouracil, and a number of pseudouracil synthetases that can act on mRNA have been identified, including DKC1. In a BLAST search for downstream genes complementary to the SNORA70E sequence, we found that there may be binding sites between SNORA70E and RAP1B, which functions as a tumour promoter. RAP1B is mainly located in the nucleus, and studies have shown that RAP1B expression is associated with tumorigenesis and metastasis in, for example, ovarian cancer, oesophageal squamous cell carcinoma, and gastric cancer, and can act as a tumour promoter by regulating multiple signalling pathways such as Wnt/β‐catenin, PI3K/AKT/mTOR pathway. , , , , We confirmed that the overexpression of SNORA70E induced the expression of the RAP1B protein significantly. Through RIP detection, we found that DKC1 can bind to RAP1B mRNA, thus we suggest that SNORA70E regulates RAP1B mRNA through pseudouracil modification, affecting its protein expression. Silencing of RAP1B in SNORA70E overexpressing CAOV3 cells or high expression OVCAR3 cells inhibited cell proliferation, migration and invasion, increased apoptosis, and decreased the levels of β‐catenin, PI3K, AKT1, mTOR, and MMP9. Studies have reported that Wnt/β‐catenin signalling pathways play an active role in cancer stem cells (CSCs) and carcinogenesis in many tumours, including ovarian cancer subtypes. Besides, the cell survival, growth, and proliferation of ovarian cancer id regulated by the PI3K/AKT/mTOR signalling pathway. Thus, we suggest that SNORA70E might regulate RAP1B and further regulates the expression of β‐catenin, PI3K, AKT1, mTOR, and MMP9 to exert cancer‐promoting effects. RNA variable shearing is an important component of eukaryotic gene expression regulation. That is, the exon of RNA generated by the main gene or mRNA precursor transcription is reinable by RNA shearing, thereby producing different mRNAs, which might play different effects. The RNA‐Seq results showed that in SNORA70E‐regulated alternative splicing (AS) of PARPBP, the 4th Exon was lost in PARPBP‐88, forming transcript PARPBP‐15, which was confirmed using PCR. Alternative splicing refers to the process by which the exons of the RNA transcribed from the main gene or mRNA precursor are variously rearranged via RNA splicing. The resulting different mRNAs may be translated into different protein constructs. Therefore, a gene might encode multiple proteins. There are seven types of AS, including mutually exclusive exons (ME), alternate terminator (AT), alternate promoter (AP), alternate acceptor site (AA), alternate donor site (AD), retained intron (RI), and exon skipping (ES). Alternative splicing occurs when different exons or introns are retained or excluded to generate alternative mRNA transcripts, and this process significantly increases the proteome diversity and cell complexity. , AS profoundly alters the function of proteins by changing their stability, adding or deleting structural domains, and modifying their protein–protein interactions. AS has been increasingly implicated in human diseases, especially cancer. The alternative splicing of genes modifies proteins involved in many malignant activities, including proliferation, invasion, metastasis, apoptosis, hypoxia, metabolic changes, and immune escape. Aberrant alternative splicing is a potential biomarker of tumorigenesis and prognosis and is also a therapeutic target in malignancy. , Further studies showed that the overexpression of PARPBP‐15 could promote the invasion, migration, and proliferation of ovarian cancer cells, while PARPBP‐88 could not. Besides, PARPBP‐15 overexpression in SNORA70E‐ASO transfected OVCAR3 cell lines could significantly reversed the inhibition ability by SNORA70E‐ASO. Above, these results suggested that SNORA70E‐regulated AS of PARPBP‐induced ovarian tumorigenesis and progression. Furthermore, ASO‐mediated silencing of SNORA70E could inhibit cell proliferation, invasion, migration ability and induce apoptosis in vitro, and inhibit tumorigenicity in vivo, suggesting that SNORA70E may be a new diagnostic and therapeutic target in ovarian cancer. In conclusion, SNORA70E, which is highly expressed in ovarian cancer, regulates RAP1B mRNA through pseudouracil modification, affecting its protein expression, and further regulates β‐catenin, PI3K, AKT1, and mTOR pathways to promote the development of ovarian cancer and epithelial ovarian cancer tumorigenesis and progression. Besides, SNORA70E also regulates the alternative splice of PARPBP to promote the development of ovarian cancer.
Shuo Chen: Conceptualization (equal); data curation (equal); formal analysis (equal); funding acquisition (equal); methodology (equal); writing – original draft (equal); writing – review and editing (equal). Qian‐hui Li: Investigation (equal); methodology (equal). Xi Chen: Investigation (equal); methodology (equal). Hai‐juan Bao: Investigation (equal). Wu Wu: Investigation (equal). Fan Shen: Investigation (equal). Bing‐feng Lu: Formal analysis (equal); investigation (equal). Ru‐qi Jiang: Investigation (equal). Zhi‐Hong Zong: Funding acquisition (equal); methodology (equal). Yang Zhao: Conceptualization (lead); funding acquisition (equal); writing – review and editing (lead).
The authors declare no conflict of interests.
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 |
PMC9575554 | Jielin Wang,Xuan You,Yanmin He,Xiaozhen Hong,Ji He,Sudan Tao,Faming Zhu | Simultaneous genotyping for human platelet antigen systems and HLA-A and HLA-B loci by targeted next-generation sequencing | 29-09-2022 | human leukocyte antigen,human platelet antigens,platelet transfusion refractoriness,next-generation sequencing,platelet blood donors | In order to treat the alloimmunization platelet transfusion refractoriness (PTR), human leukocyte antigen (HLA)-type and/or human platelet antigen (HPA)-type matched platelets between donors and patients are usually used. Therefore, genotyping of HLA-A and HLA-B loci, as well as HPA systems, for donors and patients, is of great significance. However, there is a rare report of genotyping for HLA-A and HLA-B loci as well as HPA systems at the same time. In this study, a high-throughput method for simultaneous genotyping of HLA-A and HLA-B loci, as well as HPA genotyping, was developed. A RNA capture probe panel was designed covering all exon sequences of the GP1BA, GP1BB, ITGA2, CD109, ITGB3, and ITGA2B genes and HLA-A and HLA-B loci. The HLA-A, HLA-B, and 34 HPA systems were genotyped using a targeted next-generation sequencing (NGS) method. The genotypes of the HLA-A and HLA-B loci, as well as the HPA, were assigned based on the nucleotides in the polymorphism sites. Using the NGS method, 204 unrelated blood specimens were successfully genotyped for all 34 HPA systems as well as HLA-A and HLA-B loci. The accuracy of the NGS method was 100%. Only HPA-2, HPA-3, HPA-5, HPA-6w, HPA-15, and HPA-21w showed polymorphism with frequencies of 0.9412, 0.6863, 0.9853, 0.9779, 0.4314, and 0.9951 for a allele, respectively. Thirty-two single nucleotide variants (SNVs) were detected. Of them, 12 SNVs can lead to amino acid change. HLA-A*11:01 and HLA-B*46:01 are the most common alleles for HLA-A and HLA-B loci. A targeted next-generation sequencing method for simultaneously genotyping HPA systems and HLA-A and HLA-B loci was first established, which could be used to create a database of HLA-typed and/or HPA-typed unrelated donors. | Simultaneous genotyping for human platelet antigen systems and HLA-A and HLA-B loci by targeted next-generation sequencing
In order to treat the alloimmunization platelet transfusion refractoriness (PTR), human leukocyte antigen (HLA)-type and/or human platelet antigen (HPA)-type matched platelets between donors and patients are usually used. Therefore, genotyping of HLA-A and HLA-B loci, as well as HPA systems, for donors and patients, is of great significance. However, there is a rare report of genotyping for HLA-A and HLA-B loci as well as HPA systems at the same time. In this study, a high-throughput method for simultaneous genotyping of HLA-A and HLA-B loci, as well as HPA genotyping, was developed. A RNA capture probe panel was designed covering all exon sequences of the GP1BA, GP1BB, ITGA2, CD109, ITGB3, and ITGA2B genes and HLA-A and HLA-B loci. The HLA-A, HLA-B, and 34 HPA systems were genotyped using a targeted next-generation sequencing (NGS) method. The genotypes of the HLA-A and HLA-B loci, as well as the HPA, were assigned based on the nucleotides in the polymorphism sites. Using the NGS method, 204 unrelated blood specimens were successfully genotyped for all 34 HPA systems as well as HLA-A and HLA-B loci. The accuracy of the NGS method was 100%. Only HPA-2, HPA-3, HPA-5, HPA-6w, HPA-15, and HPA-21w showed polymorphism with frequencies of 0.9412, 0.6863, 0.9853, 0.9779, 0.4314, and 0.9951 for a allele, respectively. Thirty-two single nucleotide variants (SNVs) were detected. Of them, 12 SNVs can lead to amino acid change. HLA-A*11:01 and HLA-B*46:01 are the most common alleles for HLA-A and HLA-B loci. A targeted next-generation sequencing method for simultaneously genotyping HPA systems and HLA-A and HLA-B loci was first established, which could be used to create a database of HLA-typed and/or HPA-typed unrelated donors.
There are a variety of alloantigen expression on the human platelet membrane surface, including some erythrocyte blood group antigens, human leukocyte antigen (HLA) class I proteins, human platelet antigens (HPA), and CD36 antigen (1–4). Nonetheless, the blood group system antigens, HLA class I antigen, and CD36 antigen in the human platelet membrane are shared with other cells (5–7). Classical HLA class I genes consist of HLA-A, HLA-B, and HLA-C loci. It has been reported that the HLA-A, HLA-B, and HLA-C loci are highly polymorphic, and their distribution in various populations is different (8). To date, 35 HPA have been identified and officially nominated by the International Platelet Immunology Nomenclature Committee of the International Society of Blood Transfusion (ISBT) (9). All HPA are located on the GPIIb/IIIa, GPIb/V/IX, GPIa/IIa, and CD109 platelet membrane glycoprotein complexes (1, 10). The antigens of the HPA systems showed a certain degree of polymorphism, but HPA a and b antigens were only found in HPA-1 to HPA-6w, HPA-15, and HPA-21w in the Chinese Han population (11). Due to the difference in platelets’ alloantigen expression in the individuals, it can trigger a corresponding immune response to form antibodies through transfusion, pregnancy pathways, and allotransplantation. Antibodies against HLA-A and HLA-B, HPA, and CD36 antigens are responsible for some clinical syndromes and transfusion-related conditions, such as platelet transfusion refractoriness (PTR), posttransfusion purpura (PTP), and neonatal alloimmune thrombocytopenia (NAIT) (1, 12, 13). Currently, the transfusion of platelets for both prophylaxis and treatment of bleeding is relevant to all areas of clinical medicine. PTR is a common complication of patients receiving multiple transfusions, which is defined as an unsatisfactory posttransfusion platelet count increment. PTR can be separated into immune and nonimmune causes (14). In the alloimmunization PTR, most cases can be found with anti-HLA, anti-HPA, and/or anti-CD36 (15–17). In order to treat and prevent the alloimmunization of PTR, a common method was used using HLA-typed and/or HPA-typed matched platelets (18, 19). However, this needs to establish a database for HLA-typed and/or HPA-typed unrelated donors. It shows that 5,000, 18,000, and 25,000 donor candidates would be necessary to find at least five completely compatible donors in the Japanese, European Caucasoid, and North American Caucasoid populations, respectively (19). Although platelets’ membrane surfaces express the HLA-A, HLA-B, and HLA-C molecules, platelet donors are not routinely matched for HLA-C antigen during the HLA match platelet procedure (20). In order to genotype for HLA-typed and/or HPA-typed unrelated donors, it needs to provide the corresponding methods for HLA-A and HLA-B loci, and HPA systems. Currently, many methods have been developed to detect HLA-A and HLA-B loci and HPA systems, including PCR sequence-specific primers (PCR-SSP), PCR sequence-based typing (PCR-SBT), and real-time PCR (11, 21–23). However, these methods always detect single gene separately and need too many wells and multiple amplification for the test. Now, next-generation sequencing (NGS) has been widely used for genotyping HLA loci. When compared with PCR-SBT, it has some advantages in terms of throughput and cost (24–26). Besides short-read sequencing, long-read sequencing technologies have been reported, such as single-molecule real-time sequencing (SMRT; Pacific Biosciences (PacBio)) and nanopore sequencing (Oxford Nanopore Technologies (ONT)) (27–29). Compared with short-read sequencing, these two long-read technologies provide lower per-read accuracy and require high data processing equipment. Furthermore, some studies have reported that the NGS method was used for HPA genotyping (30–34). Vorholt et al. reported an amplicon-based approach to genotyping for HPA-1 to HPA-5, HPA-9w, HPA-10w, HPA-16w, HPA-19w, HPA-27w, and HPA-34w, which required being amplified into 12 different fragments in each specimen (33). Davey et al. developed a targeted enrichment, high-sensitivity HaloPlex assay for 29 HPA systems, and 47 samples were sequenced (34). However, there is a rare report of simultaneously genotyping HLA-A and HLA-B loci as well as HPA systems. Here, in order to establish a database for HLA-typed and/or HPA-typed unrelated donors, a simultaneous genotyping method for HLA-A and HLA-B loci and HPA-1–HPA-30 and HPA-32–HPA-35 systems was established based on the target enrichment technology. The frequencies of genotypes and alleles of 34 HPA systems and HLA-A and HLA-B loci were also analyzed in the Chinese platelet blood donors.
In total, 204 unrelated blood specimens were collected from healthy platelet blood donors who have at least donated the platelet three times at the Blood Center of Zhejiang Province, China. The ethnic background of all individuals is Zhejiang Han. Informed consent was obtained from all participants. This study was approved by the Ethical Scientific Committee of the Blood Center of Zhejiang Province, China (2020-005). From each platelet donor, 5 ml peripheral blood with EDTA anticoagulant was collected. To validate the accuracy of the method, nine HPA reference specimens from the 14th, 16th, and 20th Platelet Immunology Workshop of ISBT have been chosen for detection, which contained HPA-1ab, HPA-1bb, and HPA-4ab genotypes, respectively.
The genomic DNA was extracted using a commercial DNA extraction reagent kit (RBC Bioscience, Taiwan) and an automatic Magcore nucleic acid extraction instrument (RBC Bioscience, Taiwan) according to the manufacturer’s instructions. The OD260/280 ratio of the DNA was 1.6 to 1.8. The final DNA concentration was adjusted to 30 ng/μl.
A RNA capture probe panel covering all exon sequences of the GP1BA, GP1BB, ITGA2, CD109, ITGB3, and ITGA2B genes and HLA-A and HLA-B loci was designed based on the GRCh37 reference sequence and synthesized by a commercial company (Lianchuan Biotechnology Co. Ltd., Hangzhou, China). This panel included 498 nonoverlapping probes, covering 97.88% of target regions. The length of each probe was 120 bp, and all probes were designed by end-to-end tiling to match the reference sequence. The 5′ end of all capture probes is coupled with biotin and then all probes were mixed. The information on the covered region in the gene for HPA and HLA-A and HLA-B genotyping are listed in Table 1 . The sequences of all probes are listed in Supplementary Table S1 .
DNA fragmentation and amplification were prepared from genomic DNA using the VariantBaits™ Target Enrichment System (Lianchuan Biotechnology Co., Ltd., Hangzhou, China) according to the manufacturer’s instructions. Briefly, a total of 200 ng genomic DNA for each specimen was sheared to 200–250 bp by a focused ultrasonicator (M220-Covaris, Auburn, MA, USA) at 4°C, followed by end-repairing, A-tailing, index ligation, and purification. The purification DNA was then amplified in a 50 µl reaction system, including 20 µl DNA fragments, 5 µl primers, and 25 µl DNA polymerase master mixture, by initial denaturation at 98°C for 45 s, followed by 7 cycles at 98°C for 15 s, 60°C for 30 s, 72°C for 30 s, and final extension at 72°C for 1 min. The PCR amplicons with 200–500 bp length were obtained by size selection using purification magnetic beads (ThermoFisher, San Jose, CA, USA). The PCR amplicon concentration and quality were determined by the Qubit instrument (ThermoFisher, San Jose, CA, USA) and the Agilent 4200 Bioanalyzer (Agilent Technologies Inc., Santa Clara, CA, USA).
Every four indexed PCR amplicons (500 ng) were pooled into one tube for hybridization and subsequently added 7.5 µl of different blocking solutions. The tube was evaporated in a vacuum freeze drier (Heto-Holten, ThermoFisher Scientific, San Jose, CA, USA) at 40°C. In total, 7.5 μl 2× hybridization buffer and 3 μl blocking solution B were added to the dry tube and incubated at 95°C for 5 min and 65°C for 5 min. After, 4.5 µl VariantBaits™ biotinylated probes (Lianchuan Biotechnology Co. Ltd., Hangzhou, China) were added, and the hybridization was performed at 65°C for 16 to 24 h. In this step, PCR amplicons that contained the target sequences of the HPA systems and HLA-A and HLA-B loci would be specifically hybridized into biotinylated capture probes. In the next reaction, 40 µl of Dynabeads™ MyOne™ Streptavidin T1 magnetic beads (ThermoFisher, San Jose, CA, USA) was used to separate and purify the captured product from the above hybridization system, and the product was resuspended using 20 µl nuclease-free water. The 20 µl separated DNA was amplified with a 50 µl reaction mix containing 5 µl primer mixture and 25 µl DNA polymerase master mixture, performed as follows: 98°C for 45 s; 16 cycles of 98°C for 15 s, 60°C for 30 s, and 72°C for 30 s; and a final extension at 72°C for 1 min. The PCR products were purified with magnetic beads, and the quality was detected by the Agilent 4200 Bioanalyzer.
The aforementioned PCR products from different tubes were pooled into one new tube equally, which formed the DNA libraries. The quality and size of the pooled library were detected by the Agilent Bioanalyzer 4200. The concentration of the library was determined by the Qubit instrument, prepared to a final concentration of 12 pmol/L, and then sequenced using a MiSeq instrument (Illumina, San Diego, CA, USA) with a standard v2 Reagent Kit (2*318 cycle; Illumina, San Diego, CA, USA).
The FASTQ files generated by the MiSeq instrument were analyzed for all exon sequences, including the nucleotides in the polymorphism sites of the HPA systems using CLC benchwork 23.11 (Qiagen Company, Stockach, Germany) according to the manufacturer’s instructions. The sequences of the GP1BA (NG_008767), GP1BB (NG_007974), ITGA2 (NG_008330), CD109 (NG_0033971), ITGB3 (NG_008332), and ITGA2B (NG_008331) were set as the reference sequences, respectively. The FASTQ files were aligned with the different reference sequences using CLC benchwork 23.11. Sequence data with quality over Q30 and depth of coverage over 30× were accepted. The genotypes for HPA-1 to HPA-35 (except for HPA-31) were assigned manually according to the nucleotides in the polymorphism sites from the versiti-HPA database (www.versiti.org). The genotypes of HLA-A and HLA-B loci were assigned using the TypeStream Visual Software version 2.0 (One Lambda Inc., Canoga Park, CA, USA).
In order to validate the results of the established NGS method, the specimens were also genotyped for HPA-1–HPA-28 systems using a PCR-SBT method according to our previous report (11). Three new primer pairs were added for genotyping HPA-29 to HPA-35 systems, and the procedure is the same as in our previous report except for the primers. In brief, 21 of the specific primers were divided into seven groups; therefore, one group has three primer pairs and mixed into one well. The nucleotide sequence of each HPA system was amplified using the primer mixture, and the amplicons were then Sanger sequenced using a Big Dye Terminator v3.1 cycle sequencing kit (ThermoFisher Scientific, Shanghai, China). The genotypes for HPA systems were assigned according to the nucleotides in the polymorphism sites of the HPA systems.
The HLA-A and HLA-B loci were also genotyped using an AllType NGS 9-Loci Amplification Kit (One Lambda Inc.) according to the manufacturer’s instructions. The HLA-A and HLA-B genotypes were assigned using the TypeStream Visual Software version 2.0 with the IMGT/HLA Database version 3.46.0.0 (One Lambda Inc.) as previously reported (35).
A Hardy–Weinberg equilibrium (HWE) was determined for each HPA system using the Chi-square test. p-values of less than 0.05 were considered statistically significant.
A total of 204 specimens from the blood donors were detected by the NGS method. The results of the HLA-A and HLA-B loci and HPAs in all specimens consisted of those of the commercial NGS or PCR-SBT in-house. The accuracy of the NGS method for HLA and HPA genotyping was 100%. The HPA genotypes in the nine reference specimens by the NGS method were in concordance with the reference results and demonstrated in Supplementary Table S2 . Some quality parameters of the NGS procedure were as below. All the observed sequence lengths were 150 bps, and the distribution of GC-content fits normal distribution, and the relative GC-content of a sequence of R1 and R2 in 20%–80% were 99.81% and 99.54%, respectively. The relative N-content of a sequence less than 1% of R1 and R2 were 99.95% and 99.99%. In addition, the distribution of average sequence quality scores over 30 of R1 and R2 was 98.78% and 98.08%, respectively. The mean value of the reading depth in the exon regions for each gene is shown in Supplementary Table S3 . Meanwhile, the mean reading depth frequency value of minor alleles of single nucleotide variants (SNVs) in the heterozygous positions is shown in Supplementary Table S4 .
The genotype distributions of the 34 HPA systems (HPA-31w was not analyzed by the NGS method) were fitted with Hardy–Weinberg equilibrium (p > 0.05). Among these HPA systems, only HPA-2, HPA-3, HPA-5, HPA-6bw, HPA-15bw, and HPA-21bw systems observed polymorphism in this study. However, a/a homozygote individuals were found in the other 28 HPA systems. The genotypes and allele frequencies of the systems with polymorphism are listed in Table 2 . Among these 34 HPA systems, HPA-3 and HPA-15 systems showed high polymorphisms, with frequencies of 0.6863 and 0.3137 for HPA-3a and HPA-3b alleles and 0.4314 and 0.5686 for HPA-15a and HPA-15b alleles, respectively.
Thirty-two SNVs were detected in the exon regions of the ITGB3, ITGA2B, ITGA2, CD109, and GP1BA genes. Only one SNV has not existed in the dbSNP database. No SNV was found in the GP1BB gene. The position of these SNVs and allele frequencies are shown in Table 3 . Of these 32 SNVs identified, 12 SNVs can result in an amino acid change, while the other SNVs are synonymous changes. A new SNV site, c.2878G>A located on the exon 23 of CD109, can result in an amino acid change, p.Gly960Ser, the novel SNV has been submitted to Genbank, the Accession number is OP434394. The amino acid was located in the A-macroglobulin receptor-binding domain according to the Simple Modular Architecture Research Tool (SMART) analysis based on the CD109 antigen (Q6YHK3-4). The probably damaging effect was predicted for this SNV using the PolyPhen-2 software (v2.2.3r406) in silico, and the score was 1.000 (score 0.000–0.452 = benign, 0.453–0.956 = possibly damaging, 0.957–1.000 = probably damaging).
The numbers of HLA-A and HLA-B alleles were 23 and 49 in the 204 specimens, respectively. The HLA-A*11:01 allele and HLA-B*46:01 allele are the most common alleles for each locus, with a frequency of 22.55% and 14.22%, respectively. The allele distributions of the HLA-A and HLA-B loci are listed in Supplementary Table S5 .
Thrombocytopenic patients with immune-based PTR would have significantly increased the risk of a major spontaneous or life-threatening bleed. In order to treat these patients, one of the common methods is to provide cross-match-compatible or HLA-matched/compatible platelet units (15, 18). In order to provide matching platelets, it needs a large number of donors with known HLA and/or HPA genotypes (18, 19). At present, many DNA-based methods have been used for HLA and/or HPA genotyping in platelet donors (20–28, 30, 31). It has been reported that PCR-SBT and NGS methods have been routinely used for HLA genotyping in the laboratory (24–26, 35, 36). The NGS methods for HLA genotyping are divided into targeted and amplicon-based methods (34, 35). Classical HLA-A, HLA-B, and HLA-C molecules were all expressed in the platelet, but the HLA-C molecule was expressed at a low level. Therefore, only HLA-A and HLA-B loci were routinely analyzed to search for HLA-matched/compatible platelet units (20). The HPA genotyping methods included PCR-SSP, TaqMan assay, PCR-SBT, and NGS (22, 23, 33, 34). However, there is a rare study for HPA and HLA genotyping simultaneously. Now genotyping for all HPA systems and HLA-A and HLA-B loci needs multiple PCR amplifications and/or sequencing reactions using the PCR-SSP and PCR-SBT methods, which do not simultaneously detected the all HPA systems and HLA-A and HLA-B loci. Here, we report an NGS method for simultaneously genotyping HPA-1~35 systems (except HPA-31w) and HLA-A and HLA-B loci in platelet donors. In this study, target-enrichment panel probes for the HLA-A, HLA-B, ITGB3, ITGA2B, ITGA2, CD109, GP1BA, and GP1BB genes were designed based on GRCh37 as a reference. Due to the HPA-31w polymorphism, the site is located in the GP9 gene and it was not nominated when we initiated this study (10). Therefore, the target probes were not for the GP9 gene, and the protocol cannot assign the HPA-31w system. The design probes overlapped one by one to cover all exon regions. Therefore, all known human platelet antigens except for HPA-31w and novel variations in the encoding regions can be detected. In the panel, the probes are RNA-based, which has better hybridization efficiency and stability to DNA complexes compared to DNA probes (37, 38). However, the RNA probes are expensive and require a lower temperature to store. In the hybridization reaction, four tubes were pooled into one to assure the volume was less than 31 μl after pooling; otherwise, it will affect hybridization efficiency. In order to validate the results of the established NGS method, the specimens were genotyped for all HPAs by a PCR-SBT method in-house and for HLA-A and HLA-B loci using a commercial NGS method. It showed that the accuracy of the established NGS method for HLA and HPA was 100%. The accuracy of the NGS method depends on some influence factors. Davey et al. reported the results of the NGS method for HPA were in concordance 100% with historical HPA genotypes (34). However, Vorholt et al. reported that the concordance with TaqMan assay and NGS results for HPAs was 99.8%, and the discrepancies between them were due to allele dropout in real-time PCR (33). It is generally considered that a depth of coverage of over 30× is necessary for some SNV detection (39). Due to the depth of coverage being different among the genes in most NGS methods, it needs to assure a depth of coverage of over 30× for each gene. The accuracy can be improved with increased coverage (40). However, the coverage can also be influenced by the GC content, which may affect the efficiency of PCR. There was also evidence that the same volume had more coverage than the same mole (33), but the specimens that were mixed in the same mole were better than the same volume in our experiment. In addition, we found a repetitive sequence in exon 2 of the GP1BA gene, which can affect the analysis accuracy when using the CLC benchwork 23.11. Therefore, the SNVs in the GP1BA gene by the NGS method were also confirmed by PCR-SBT in-house. Besides the depth of coverage, the breadth of coverage also helps to ensure that sensitivity and specificity are sufficient for supporting variant detection. However, the coverage profiles may not be uniform and a lack of coverage in key regions, such as exons, may affect the accuracy of the typing. Although software analysis programs will usually have built-in filters to define the minimum coverage required for accurate typing, certain circumstances may permit going below this threshold, such as when the polymorphisms of two alleles of a locus are phased or when the region with low coverage does not affect the genotyping (i.e., introns, untranslated regions). Furthermore, besides the depth of coverage, for HLA typing data analysis, the assessment of adequate allele balance is important for detecting issues due to allele dropout. The distribution of HLA-A, HLA-B, and HPA systems was similar to those of our previous reports (11, 35). Furthermore, some SNVs were detected in the coding region, which is located in the structural domain of the platelet membrane glycoproteins ( Table 3 ). c.882T>C and c.1143A>C in the ITGB3 were both located in the VWA domain (41), which are synonymous changes. c.3063C>T in the ITGA2B is located in the transmembrane region (42), a hydrophobic region of the protein. c.327G>A in the ITAG2 is located in the integrin alpha region of the GPIa (43); c.759C>T, c.789T>C, c.825G>A, and c.993A>G in the ITAG2 are located in the VWA domain of the GPIa; and c.3252C>T and c.3324T>C are located in the transmembrane region of the GPIa. c.4173G>T in the CD109 is located in the A2M_receptor-binding domain, which is the receptor-binding domain (RBD) of alpha-2-macroglobulin proteins (44, 45). Therefore, some SNVs may affect the domain function, but this needs further study. In summary, a targeted NGS method can be successfully established to genotype simultaneously 34 HPA systems as well as HLA-A and HLA-B loci. Its advantages include high throughput and simultaneous testing for HLA-A, HLA-B, and HPA genotypes. The method can be used to establish a database bank of HLA-typed and/or HPA-typed unrelated donors, which can help provide matching platelets for PTR, PTP, and NAIT patients.
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 in the article/ Supplementary Material .
Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.
JW, XY and YH performed the experiments, collected the data, and wrote the rough draft. XH and JH analyzed the data. ST and FZ interpreted the results and drafted the paper. All authors contributed to the article and approved the submitted version.
This work was supported by the Science Research Foundation of Zhejiang Healthy Bureau (2021KY137 and WKJ-ZJ-2021).
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 |
|
PMC9575788 | Yu-Fang Huang,Ming-Wei Liu,Han-Biao Xia,Rong He | Expression of miR-92a is associated with the prognosis in non-small cell lung cancer: An observation study | 14-10-2022 | lymph node metastasis,miR-92a,non-small cell lung cancer,prognosis | With the development of molecular biology technology, the discovery of microRNAs (miRNAs) has provided new ideas for the diagnosis, treatment, and prognosis of lung cancer and laid a foundation for the study of this malignancy. To assess the potential prognostic value of miR-92a as a new biomarker in non-small cell lung cancer (NSCLC) via clinical evaluation, a total of 100 patients with NSCLC admitted to the Respiratory and Intensive Care Department of Suining Central Hospital in Sichuan Province between August 2007 and April 2020 were retrospectively analyzed. The correlation between miR-92a expression and prognosis of patients with NSCLC was also evaluated in the present study. The expression level of miR-92a was measured by PT-PCR and in situ hybridization. Chi-square test was adopted to explore the relationship of miR-92a expression and clinical features. Kaplan–Meier survival curves were plotted to delineate the overall survival rate of patients with NSCLC. Cox regression analysis was performed to evaluate the prognostic significance of miR-92a expression in NSCLC. The miR-92a expression in NSCLC tissue samples was significantly higher than that in normal lung tissues (P < .001) and significantly correlated with the Eastern Cooperative Oncology Group score, histological type, and distant metastasis (P < .05). Survival curve revealed that patients with NSCLC and high miR-92a expression had relatively higher mortality than those with low PAK4 expression (P = .001). Cox regression analysis explained that miR-92a expression was associated with the prognosis of patients with NSCLC (HR = 1.8, 95% CI: 1.0–3.2, P = .036). In summary, miR-92a was highly expressed in NSCLC tissues and could act as a prognostic factor for patients with NSCLC. These results illustrate that miR-92a expression plays an important role in the invasion and metastasis of NSCLC, and miR-92a can be used as a new biomarker to determine the prognosis of this cancer. | Expression of miR-92a is associated with the prognosis in non-small cell lung cancer: An observation study
With the development of molecular biology technology, the discovery of microRNAs (miRNAs) has provided new ideas for the diagnosis, treatment, and prognosis of lung cancer and laid a foundation for the study of this malignancy. To assess the potential prognostic value of miR-92a as a new biomarker in non-small cell lung cancer (NSCLC) via clinical evaluation, a total of 100 patients with NSCLC admitted to the Respiratory and Intensive Care Department of Suining Central Hospital in Sichuan Province between August 2007 and April 2020 were retrospectively analyzed. The correlation between miR-92a expression and prognosis of patients with NSCLC was also evaluated in the present study. The expression level of miR-92a was measured by PT-PCR and in situ hybridization. Chi-square test was adopted to explore the relationship of miR-92a expression and clinical features. Kaplan–Meier survival curves were plotted to delineate the overall survival rate of patients with NSCLC. Cox regression analysis was performed to evaluate the prognostic significance of miR-92a expression in NSCLC. The miR-92a expression in NSCLC tissue samples was significantly higher than that in normal lung tissues (P < .001) and significantly correlated with the Eastern Cooperative Oncology Group score, histological type, and distant metastasis (P < .05). Survival curve revealed that patients with NSCLC and high miR-92a expression had relatively higher mortality than those with low PAK4 expression (P = .001). Cox regression analysis explained that miR-92a expression was associated with the prognosis of patients with NSCLC (HR = 1.8, 95% CI: 1.0–3.2, P = .036). In summary, miR-92a was highly expressed in NSCLC tissues and could act as a prognostic factor for patients with NSCLC. These results illustrate that miR-92a expression plays an important role in the invasion and metastasis of NSCLC, and miR-92a can be used as a new biomarker to determine the prognosis of this cancer.
The diagnosis and treatment of lung cancer have been updated and continuously improved. However, the mortality rate of this disease remains high because of the lack of specific clinical symptoms and signs during the early stages of lung cancer. Thus, patients are more likely to ignore it, resulting in a missed optimal treatment period.[ Some studies have shown that the 5-year survival rate of patients with early stage lung cancer after surgery is more than 70%, whereas that of patients with middle-advanced stage lung cancer is only approximately 20%. Therefore, the 5-year survival rate of patients with lung cancer is closely associated with early diagnosis and treatment.[ With the development of molecular biology technology, the discovery of microRNAs (miRNAs) has provided new ideas for the diagnosis, treatment, and prognosis of lung cancer and laid a foundation for the study of this malignancy. Clinical studies have found that compared with normal lung tissue, the expression levels of miR-21 in lung squamous cell carcinoma tissue, atypical hyperplasia tissue, and metastatic cancer tissue are increased by 9.1, 4.5, and 11.8 times, respectively,[ suggesting that miR-21 is closely related to lung squamous cell carcinoma. The occurrence and progression of cancer are closely related.[ Introducing exogenous miR-21 analogs into lung cancer H2170 cells to upregulate the expression of miR-21 in cells can lead to significantly higher cell proliferation ability in those with high miR-21 expression than those with low expression. miR-21 can promote the proliferation of cancer cells by inhibiting apoptosis and exert the function of oncogenes.[ miR-150 is highly expressed in non-small cell lung cancer (NSCLC) and is closely related to clinicopathological features, such as tumor node metastasis (TNM) stage, lymph node metastasis, and distant metastasis.[ Kaplan–Meier analysis showed that the 5-year overall survival rate of patients in the high expression group of miR-150 was 40.8%, whereas that in the low expression group was 69.2%, suggesting that high expression of miR-150 is associated with poor prognosis of patients.[ Recent studies have found that miR-92a promotes the proliferation, invasion, and migration of cervical cancer cells by directly targeting phosphoinositide 3 kinase regulatory subunit 1.[ Downregulated miR-18a and miR-92a inhibit the growth of NSCLC by targeting Sprouty 4.[ Current studies have found that miR-92a can be used as a marker for the diagnosis of colon cancer and is an important indicator for prognosis.[ However, whether miR-92a can be used as an important indicator for assessing the prognosis of NSCLC remains unknown. Therefore, the purpose of this study was to evaluate whether miR-92a can be used as a prognostic indicator for NSCLC.
This study was approved by the Ethics Committee of Suining Central Hospital in Sichuan Province. All NSCLC tissue samples were collected on the premise that the patients signed informed consent forms. Wax blocks of 100 surgically resected specimens were collected from patients with confirmed NSCLC in the Department of Respiratory and Critical Care Medicine, Suining Central Hospital between August 2007 and April 2018. The patient screening steps are shown in Figure 1. The patients included 78 men and 22 women aged 39–77 years (median, 61 years). Based on the 2002 version of the Union for International Cancer Control (UICC) standard, there were 38 stage I cases, 37 stage II cases, and 25 stage III cases. The samples included 57 cases of lymph node metastasis, 54 of squamous cell carcinoma, 46 of adenocarcinoma, 6 of highly differentiated cancer, 57 of moderately differentiated cancer, and 37 of poorly differentiated cancer. Thirty paracancerous normal lung tissue wax blocks were used as the control group. The inclusion criteria were as follows: patients who had not received any form of chemotherapy, radiotherapy, or targeted drug therapy; tissue specimens were obtained by puncture biopsy in all cases; paraffin-embedded sections of the specimens were excised during surgery; the pathological diagnosis was NSCLC; the patient had complete clinical data; and the acquisition and processing of specimens complied with the ethical norms and operating procedures of clinical trials. Patients with a clinical pathological staging of 0 or precancerous lesions were excluded from the study.
Quantitative real-time PCR (qPCR) kits were purchased from Thermo Fisher Scientific (USA). TRIzol reagents and RNA extraction kits were purchased from Beijing Solarbio Science and Technology Co., Ltd. (Beijing, China). miR-92a and U6 primers were synthesized by Shanghai Shenggong Biological Engineering Co., Ltd. (Shanghai, China).
Total RNA from the cell culture was extracted using the mirVana miRNA isolation kit, according to the manufacturer’s instructions. Revert Ace kits (TOYOBO, Osaka, Japan) were used to reverse transcribe the RNA samples into cDNA using different reverse transcribed primers. The primer sequences were as follows: U6, upstream 5-GCTTCGGCAGCACATATACTAAAAT-3, downstream 5-CGCTTCACGAATTTGCGTGTCAT-3; miR-92a, upstream 5-AGCTCTACGACTGTCACTCG-3, downstream 5-GTATGCATTCTATCGTAG-3′. The reaction conditions were as follows: 95°C for 10 minutes, 45 cycles of denaturation at 95°C for 10 seconds, and annealing and extension at 60°C for 60 seconds. The melting curve was determined. After the test was completed, the Ct value of each sample was automatically analyzed using a computer system, and 2−ΔΔCt was used to calculate the relative expression of miRNA. The experiment was repeated three times.
The tissues were stored overnight at 65°C. The tissue was dewaxed by adding xylene twice for 10 minutes each. The tissue was rehydrated and treated with 30% acidic sodium sulfite at 50°C for 20 to 30 minutes. The nucleic acid hybridization rinse solution was used to rinse the tissue twice for 5 minutes each time. The samples were enzymatically digested and dehydrated. The sample was then immersed in acetone for 2 minutes, dried, placed on a glass slide, and covered at 56°C for 3 minutes, after which it was immersed in a denaturing solution at 73°C for 5 minutes. The following operations were performed under protection from light. The sample was subjected to gradient alcohol dehydration at −20°C, preheated at 45°C to 50°C for 2 to 5 minutes, and hybridized overnight with a probe at 42°C. The glass slide was washed with formamide and dried, and DAPI was added to the slide for double staining. The samples were observed under a fluorescence microscope after 20 to 30 minutes. The positive signal for miR-92a was determined by blue staining of the cytoplasm and red staining of the nucleus.
The 100 enrolled patients were numbered based on their admission orders and hospital numbers to protect their privacy to the greatest extent. The patient’s personal phone number, email address, and other contact information were recorded by the HIS system based on the data registered when the patient was admitted to the hospital. When the patient was discharged from the hospital, the patient or the patient’s family was instructed to follow up at the outpatient clinic of our hospital at 1, 2, 3, and 6 months after discharge. After 1 to 3 years, the patient visited the hospital for follow-up every 6 months. After 3 years, the patient visited our hospital for follow-up every year. During follow-up, the patient’s current survival status was evaluated, including complications and secondary conditions if and when recurrence occurred. Detailed information was also recorded. The patient’s last follow-up was in January 2021. If the enrolled patient died or was lost to follow-up before October 2020, the overall survival time or disease-free survival (DFS) time was recorded as the time of death or loss to follow-up. The lost follow-up data were deleted during data analysis to control for confounding factors. In this study, none of the 100 patients with NSCLC was lost to follow-up.
SPSS19.0 was used to test the normality of the measurement data in the experiment. When the data conformed to a normal distribution, an independent-sample t-test was used. The Mann–Whitney test was performed with α = 0.05 as the test level. The results are expressed as median (interquartile range). To determine the diagnostic efficacy of lung cancer tissue miRNAs for NSCLC, the receiver operating characteristic (ROC) curve was used to analyze the samples. The Kaplan–Meier method was used to evaluate the survival of patients with lung cancer. Univariate and multivariate Cox proportional hazards models were used to analyze the importance of survival variables. The independent variables included in multivariate Cox analysis were statistically significant in univariate Cox analysis.
The results of in situ hybridization staining showed that the expression of miR-92a in the patient’s NSCLC tissues was significantly higher than that in the adjacent tissues (Fig. 1).
The expression of miR-92a in NSCLC tissues was higher than that in adjacent lung tissues. The expression was also related to clinical parameters, such as vascular invasion, TNM staging, and lymph node metastasis (Table 1).
The area under the ROC curve (AUC = 0.669, P < .001) suggested that high miR-92a expression was correlated with NSCLC (Fig. 2A). Kaplan–Meier survival analysis was performed to explore the prognostic value of miR-92a in patients with NSCLC. The results showed that the overall survival (OS) of patients with high miR-92a expression was significantly higher than that of patients with low miR-92a expression (χ2 = 8.364, P = .003; Fig. 2B). However, there was no statistically significant difference in DFS between the two groups (χ2 = 0.831, P = .367; Fig. 2C). The abnormally high expression of miR-92a may have had an effect on the prognosis of patients with NSCLC (Fig. 2B and C).
Univariate Cox analysis showed that lymph node metastasis (P < .0001), distant metastasis (P = .0435), tumor differentiation (P = .0472), TNM staging (P = .009), and miR-92a expression (P = .003) were significantly associated with the OS rate of patients with NSCLC (Table 2). Further study suggested an inherent connection between miR-92a and poor prognosis of patients with NSCLC. Statistically significant variables in univariate Cox analysis were included in multivariate Cox regression analysis, and the results showed that lymph node metastasis (HR = 4.5, 95%CI: 2.1–7.7, P < .0001) and miR-92a (HR = 1.8, 95% CI: 1.0–3.2, P = .036) were independent predictors of OS in patients with NSCLC (Table 3).
miRNAs are involved in biological processes such as cell division, proliferation, differentiation, apoptosis, development, metastasis, angiogenesis, and immune responses.[ In the process of tumor occurrence and development, miRNA not only plays an important role in promoting the tumor gene but also the tumor suppressor gene.[ Relevant studies have found a close relationship between the expression profile of miRNA and the embryonic origin of tumors, and the expression of miRNA is usually expressed in a tissue-specific manner.[ miRNAs are highly conserved, tissue-specific, time-sequential, and stable; their changes in organisms occur earlier than the appearance of protein markers.[ Studies have shown that the abnormal expression of miR-92a is closely related to the occurrence and development of tumors, and it plays an important role in the proliferation, apoptosis, invasion, and migration of various types of tumor cells.[ Recent studies have shown that the expression of miR-92a is upregulated in colon cancer, gastric cancer, and other tumors, which is related to tumor progression and angiogenesis.[ Zhou et al[ reported that the expression of miR-92a is upregulated in cervical cancer, and it can promote cancer cell proliferation and invasion by targeting F-box with 7 tandem WD40. Zhang et al[ reported that miR-92a induces colorectal cancer epithelial cell–mesenchymal transition to regulate cell proliferation, migration, and invasion by inhibiting phosphatase and tensin homolog expression. Ren et al[ showed that the upregulation of the expression of miR-92a in NSCLC can promote cancer cell proliferation, migration, and invasion; reduce cell apoptosis; and enhance chemotherapy resistance. The expression of miR-92a is inhibited, which showed the opposite effect. They confirmed that phosphatase and tensin homolog is a direct target of miR-92a. Studies have shown that the upregulated expression of miR-92a in colorectal cancer can promote tumor angiogenesis.[ Studies have shown that the abnormal expression of miRNAs is related to the occurrence and development of tumors, and its expression profile has obvious prognostic significance. In addition, miR-92a can inhibit the occurrence and development of hepatocellular carcinoma, gastric cancer, colorectal cancer, and other cancers.[ The characteristics of miRNAs as promoting genes for lung cancer tumors have been primarily studied in miR-17-92 clusters. These clusters also play an important role in the promotion of other tumors.[ miR-92a, a member of the miR-17-92 family, affects the migration of lung cancer cells via the TAT3→miR-92a→RECK axis.[ Some studies have also shown that high expression of miR-92a in serum is closely related to lymph node metastasis.[28] Studies on NSCLC have shown that high miR-92a expression is related to NSCLC chemotherapy sensitivity and poor prognosis.[ Among the known miRNAs, miR-92a is one of the most attractive, and different institutions have shown that miR-92a has potential prognostic significance in patients with cancer.[28] However, the results of different studies have been inconsistent. In this work, the overall risk of miR-92a in tumor prognosis was evaluated through clinical studies, thereby laying the foundation for subsequent studies and prognostic assessments of NSCLC. In our clinical studies, a comparison of the relative expression level of miR-92a in lung cancer tissues and normal tissues adjacent to the cancer tissues showed that the relative expression level of miR-92a in the lung cancer tissues was 1.84 ± 0.76, whereas the relative expression level in the adjacent normal tissues was 0.45 ± 0.18. The relative expression level of miR-92a in the lung cancer tissues was significantly higher than that in the adjacent normal tissues, and the difference was statistically significant (P < .05). The relative expression of miR-92a in the lung cancer tissues of patients with NSCLC was not related to their clinicopathological type (P > .05). In this study, the expression of miR-92a in patients with positive lymph nodes, poor tumor differentiation, and late clinical staging significantly increased. This result indicated that the invasion and metastatic potential of lung cancer were closely related to the high expression of miR-92a. Analysis of the ROC curve showed that the AUC of miR-92a was 0.613, sensitivity was 50%, and specificity was 68% (P = .083). Kaplan–Meier analysis showed that the median survival time of patients with NSCLC in the high miR-92a expression group was 15 months, whereas that in the low miR-92a expression group was 24 months. The difference between the two groups was statistically significant (P < .05). Studies have shown that miR-92a and lymph node metastasis are independent risk factors for OS in patients with NSCLC. In this study, univariate Cox analysis of patients with NSCLC showed that lymph node metastasis (P < .0001), distant metastasis (P = .0435), tumor differentiation (P = .0472), TNM staging (P = .009), and miR-92a expression (P = .003) were significantly associated with the OS rate of patients with NSCLC. Statistically significant variables in univariate Cox regression analysis were included in multivariate Cox regression analysis. The results showed that lymph node metastasis (HR = 4.5, 95% CI: 2.1–7.7, P < .0001) and miR-92a expression (HR = 1.8, 95% CI: 1.0–3.2, P = .036) were independent predictors of OS in patients with NSCLC. In summary, high expression of miR-92a is closely related to the poor prognosis of patients with NSCLC. Therefore, miR-92a could be used as a biomarker to assess the prognosis of these patients. In future studies, large sample sizes and high-quality studies are needed for further confirmation.
Y-FH, ML, H-BX, and RH contributed to data acquisition and analysis. Y-FH and M-WL contributed to data interpretation. H-BX and RH made contributions to the conception and design of the study and drafted the manuscript. H-BX and RH made contributions to the design of the study and drafted the manuscript. Y-FH, M-WL, and H-BX prepared Figures 1–3. Y-FH, M-WL, H-BX, and RH confirmed the authenticity of all raw data. All the authors have read and approved the final manuscript. Conceptualization: Yu-Fang Huang, Ming-Wei Liu, Rong He. Data curation: Ming-Wei Liu, Han-Biao Xia. Formal analysis: Ming-Wei Liu. Funding acquisition: Han-Biao Xia, Rong He. Investigation: Ming-Wei Liu, Han-Biao Xia. Methodology: Ming-Wei Liu. Project administration: Yu-Fang Huang. Resources: Han-Biao Xia. Software: Yu-Fang Huang, Han-Biao Xia, Rong He. Validation: Yu-Fang Huang, Han-Biao Xia, Rong He. Visualization: Han-Biao Xia, Rong He. Writing – original draft: Ming-Wei Liu. Writing – review & editing: Rong He. | true | true | true |
|
PMC9575882 | Lingling Zhou | Comment on “Diagnosis of long noncoding RNA LINC00173 in patients with melanoma is controversial” | 24-06-2022 | Comment on “Diagnosis of long noncoding RNA LINC00173 in patients with melanoma is controversial”
Dear Editor, We were very pleased to read the article entitled “Diagnostic and prognostic significance of long noncoding RNA LINC00173 in patients with melanoma” by Wang et al. in which they revealed that LINC00173 expression could serve as an unfavorable prognostic biomarker for melanoma patients. However, some views should be raised in my opinion. The main problem of the study was that the reliability of conclusions has been questioned in a study published recently. A study found that the LINC00173 was a potential target for the diagnosis, prognosis, and/or treatment of melanoma . However, this article was recently retracted because the authors were unable to provide satisfactory original data for their study . As can be seen in patients and tissue samples section, 163 melanoma tissues and their pair-matched nontumor specimens in this study were obtained from patients who underwent radical resections at The First People’s Hospital of Jinan City from May 2012 to July 2015. Nevertheless, LINC00173 was first reported in 2017 . It is obviously unreasonable. In conclusion, due to the above reason, diagnosis of long noncoding RNA LINC00173 in patients with melanoma is controversial. | true | true | true |
|||
PMC9575897 | Hong Zhu,Qian Ma,Xianguo Wang | Comment on “Diagnostic and prognostic significance of long noncoding RNA LINC00173 in patients with melanoma” | 24-06-2022 | Comment on “Diagnostic and prognostic significance of long noncoding RNA LINC00173 in patients with melanoma”
Dear Editor, We were very pleased to read the article entitled “Diagnostic and prognostic significance of long noncoding RNA LINC00173 in patients with melanoma” by Wang et al. In this study, the authors revealed that LINC00173 expression was abnormally elevated in melanoma and may serve as a novel biomarker for predicting diagnosis and clinical progression of melanoma patients. However, some concerns need to be raised from our opinion. The main problem of the study was that it lacks general demographic information, inclusion criteria, and exclusion criteria. There are some factors that affect the prognosis of the melanoma, including size of tumor, status of lymph node, distant metastasis, and complication. Inclusion and exclusion criteria should also be provided. Some chronic diseases such as hypertension and diabetes that affect prognosis should be excluded. Another concern is that the definition of high expression for LINC00173 was not provided. In this study, 163 melanoma tissues and their pair-matched nontumor specimens were obtained from patients who underwent radical resections at The First People’s Hospital of Jinan City from May 2012 to July 2015. LINC00173 was first reported in 2017 . Therefore, we can assume that the researchers used frozen samples for the experiment. It is not clear whether freezing storage of general samples would lead to RNA degradation, and whether the researchers considered the effect of freezing time on RNA levels. | true | true | true |
|||
PMC9576185 | Si-Yu Liu,Chao Li,Li-Yang Sun,Ming-Cheng Guan,Li-Hui Gu,Dong-Xu Yin,Lan-Qing Yao,Lei Liang,Ming-Da Wang,Hao Xing,Hong Zhu,Timothy M. Pawlik,Wan Yee Lau,Feng Shen,Xiang-Min Tong,Tian Yang | ASAP Score versus GALAD Score for detection of hepatitis C-related hepatocellular carcinoma: A multicenter case-control analysis | 30-09-2022 | hepatocellular carcinoma,hepatitis C virus,alpha-fetoprotein,lens culinaris agglutinin-reactive fraction of alpha-fetoprotein,protein induced by vitamin K absence or antagonist-II,diagnosis,biomarker | Background The GALAD and ASAP scores are two well-recognized algorithms to estimate the risk of hepatocellular carcinoma (HCC) based on gender, age, alpha-fetoprotein (AFP), protein induced by vitamin K absence or Antagonist-II (PIVKA-II) and AFP-L3 (included in the GALAD score but not in the ASAP score). The current study sought to compare the diagnostic performance of each score to detect HCC among patients infected with hepatitis C virus (HCV). Methods A multicenter case-control study was undertaken in which blood samples were collected from HCVinfected patients with and without HCC. Using the area under the receiver operating characteristic curve (AUROC), ASAP and GALAD scores were compared relative to diagnostic performance to detect any stage HCV-HCC and early-stage HCV-HCC. Results The analytic cohort included 168 HCV-HCC patients and a control group of 193 HCV-infected patients. The ASAP score had a higher AUROC to detect any stage HCV-HCC versus the GALAD score, both in the overall group (0.917 vs. 0.894, P=0.057) and in the cirrhosis subgroup (0.909 vs. 0.889, P=0.132). Similar results were noted relative to the detection of early-stage HCV-HCC, whether defined by BCLC staging (stage 0-A: 0.898 vs. 0.860, P=0.026) or 8th TNM staging (stage I: 0.899 vs. 0.870, P=0.070). In subgroup analysis to detect AFP-negative HCV-HCC, the ASAP score also demonstrated a higher AUROC than the GALAD score to detect any stage HCV-HCC in the AFP-negative subgroup (0.815 vs. 0.764, P=0.063). Conclusions The ASAP score had better diagnostic performance for early detection of HCV-HCC compared with the GALAD score. The ASAP score may be preferrable to the GALAD score for HCC screening and surveillance among HCV-infected patients. | ASAP Score versus GALAD Score for detection of hepatitis C-related hepatocellular carcinoma: A multicenter case-control analysis
The GALAD and ASAP scores are two well-recognized algorithms to estimate the risk of hepatocellular carcinoma (HCC) based on gender, age, alpha-fetoprotein (AFP), protein induced by vitamin K absence or Antagonist-II (PIVKA-II) and AFP-L3 (included in the GALAD score but not in the ASAP score). The current study sought to compare the diagnostic performance of each score to detect HCC among patients infected with hepatitis C virus (HCV).
A multicenter case-control study was undertaken in which blood samples were collected from HCVinfected patients with and without HCC. Using the area under the receiver operating characteristic curve (AUROC), ASAP and GALAD scores were compared relative to diagnostic performance to detect any stage HCV-HCC and early-stage HCV-HCC.
The analytic cohort included 168 HCV-HCC patients and a control group of 193 HCV-infected patients. The ASAP score had a higher AUROC to detect any stage HCV-HCC versus the GALAD score, both in the overall group (0.917 vs. 0.894, P=0.057) and in the cirrhosis subgroup (0.909 vs. 0.889, P=0.132). Similar results were noted relative to the detection of early-stage HCV-HCC, whether defined by BCLC staging (stage 0-A: 0.898 vs. 0.860, P=0.026) or 8th TNM staging (stage I: 0.899 vs. 0.870, P=0.070). In subgroup analysis to detect AFP-negative HCV-HCC, the ASAP score also demonstrated a higher AUROC than the GALAD score to detect any stage HCV-HCC in the AFP-negative subgroup (0.815 vs. 0.764, P=0.063).
The ASAP score had better diagnostic performance for early detection of HCV-HCC compared with the GALAD score. The ASAP score may be preferrable to the GALAD score for HCC screening and surveillance among HCV-infected patients.
Hepatocellular carcinoma (HCC) is the most common primary liver malignancy with most patients developing HCC due to chronic liver diseases. Unfortunately, HCC has a morality-to-incidence ratio that approaches 1 (1). Of note, HCC cases in China accounted for roughly 50% of the new liver cancer cases and deaths that occurred worldwide during 2012 (2). Among the etiological factors associated with HCC, hepatitis C virus (HCV) infection is a common cause. According to the World Health Organization statistics, the global HCV infection rate is about 3%, and there are about 180 million people infected with HCV globally with about 40~60 million cases in China (3–5). Patients with HCV infection have a 2% annual risk and a 7% to 14% five-year risk of developing HCC (6). The benefits of HCC surveillance on survival of HCC at any stage, particularly at early stages, have been codified in the many guidelines recommending HCC surveillance for patients with chronic hepatitis B virus (HBV) and HCV infection (7–9). The early diagnosis of HCC is essential to initiate curative treatments to improve short-term and long-term prognosis. Therefore, highly-effective methods are needed to detect HCC at an earlier stage (10, 11). Although imaging techniques such as ultrasound and magnetic resonance imaging (MRI) have markedly improved the accuracy of HCC diagnosis, their applications have been limited due to disadvantages such as high cost, invasiveness, and insensitivity to small tumors (12). Serum biomarkers play an essential role in diagnosing HCC, as biomarkers are often more convenient, inexpensive, non-invasive, and reproducible (13–15). Alphafetoprotein (AFP) is a widely used biomarker for HCC diagnosis. The diagnostic accuracy of AFP is limited, however, due to its high false-negative rate to detect small or early-stage tumors. As previous studies have demonstrated, the sensitivity of AFP among patients with HCC was 52% for tumors > 3cm and dropped to only 25% for tumors < 3cm. In addition, AFP may also be elevated in some benign liver diseases, such as chronic hepatitis and cirrhosis even in the absence of HCC (16). Therefore, the application of AFP in the early screening of HCC has been controversial (17). Over the years, novel biomarkers for HCC have been suggested, such as prothrombin induced by vitamin K absence-II (PIVKA-II, also known as des-gamma-carboxy prothrombin) and lens culinaris agglutinin-reactive fraction of alpha-fetoprotein (AFP-L3%) (13). These two biomarkers have similar limitations in clinical application due to their insufficient diagnostic performance if used alone (18–20). Therefore, employing a combined multiple biomarkers approach is essential to improve the accuracy of early HCC diagnosis and reduce the missed diagnosis rate. To this end, Johnson et al. developed a serum-based tool (GALAD) to detect HCC based on objective measures including gender, age, and three serologic biomarkers (i.e., AFP, AFP-L3%, and PIVKA-II) (21). More recently, our team developed the ASAP score comprised of age, sex, AFP, and PIVKA-II based on a Chinese population of HBV-infected patients (22). The reason why AFP-L3 was not enrolled in the ASAP score was that AFP-L3 was not an independent predictor in the regression analysis used to construct the prediction algorithm of the ASAP score (22). Despite having one less predictive variable, the ASAP score performed better than the GALAD score to diagnosis HBV-HCC (AUC: 0.935 vs. 0.921, P < 0.006).22] However, the ASAP score has not been verified in any other etiological factors of HCC. In particular, the diagnostic performance of the ASAP score to detect HCC among patients with other types of chronic liver diseases, such as HCV infection, remains unknown. In addition, whether the ASAP score or the GALAD score is more suitable to detect HCC, particularly early-stage HCC among Chinese HCV-infected patients, remains unclear. Therefore, the current study sought to assess the diagnostic performances of the ASAP score, the GALAD score, and AFP, AFP-L3, and PIVKA-II to detect early HCC in a large cohort of Chinese patients with HCV infection. In addition, we determined and compared the diagnostic performances of the ASAP score versus GALAD score to detect HCC, irrespective of the presence of liver cirrhosis. Specific subgroup analyses were performed to assess detection of early-stage HCC relative to suitable biomarkers or panels of biomarkers that were candidates for HCC screening and surveillance among high-risk populations.
Patients from March 2018 to May 2021 with chronic HCV infection irrespective of the presence of HCC were retrospectively identified from the databases of four Chinese hospitals (Zhejiang University Lishui Hospital, Eastern Hepatobiliary Surgery Hospital of Shanghai, the First Affiliated Hospital of Soochow University, and Zhejiang Provincial People’s Hospital). Patients who met any of the following exclusion criteria were excluded from the study: 1) less than 18 years old, 2) classified as Barcelona Clinic Liver Cancer (BCLC) stage D, 3) unknown etiology of HCC, 4) HBV and hepatitis C virus (HCV) co-infection, 5) receipt of any anti-HCC treatment before blood collection, 6) incomplete medical record or missing data on variables and outcomes including HCC biomarkers test. HCV infection was defined as HCV-RNA positivity within six months prior to surgery. HCC was diagnosed with imaging including ultrasound, computerized tomography, and magnetic resonance imaging and confirmed by biopsy or postoperative histopathological examination (9). Cirrhosis was diagnosed based on clinical evidence of portal hypertension or hepatic decompensation according to the established guidelines (9). HCV-HCC was defined as HCC among patients with HCV infection but without evidence of any other underlying liver diseases. Two tumor staging systems were used to determine the stage of disease: BCLC staging and 8th edition American Joint Committee on Cancer tumor-node-metastasis (TNM) staging. Early-stage HCC was defined as BCLC stage 0+A and/or 8th edition TNM stage I. Written informed consent was obtained from all participants of this study. The ethics committee approved the study of each study site, and the study conduction conformed to the ethical guidelines of the 1975 Declaration of Helsinki.
Peripheral blood samples were collected from each participant before any HCC treatment. Serum samples were separated from the blood samples by centrifugation at 700g for 10 min. Serum samples were subsequently aliquoted and frozen at -80°C until testing. The sample storage facilities and conditions were standardized at each study site. Serum samples were sent to the Abbott testing centers on a regular basis, and liquid nitrogen was used during transportation. AFP and PIVKA-II serum concentrations were measured with the commercially available ARCHITECT immunoassay per the defined protocol (Abbott Diagnostics). Serum measurements of AFP-L3% were determined utilizing the Fujirebio assay (Fujirebio Diagnostics). The lower limits of detection for AFP, PIVKA-II, and AFP-L3% assays were 0.6 ng/mL, 5.0 mAU/mL, and 0.5%, respectively. The technicians performing the laboratory tests were blinded to the diagnosis of the participants. No adverse events related to serum sample collection were observed.
The ASAP score was calculated using the following equation: ASAP score = -7.58 + 0.05 × age - 0.58 × gender + 0.42 × Ln (AFP [ng/ml]) + 1.11 × Ln (PIVIKA-II [mAU/ml]), where gender = 0 for males and 1 for females. The GALAD score was calculated using the following equation: GALAD score = -10.08 + 0.09 × age + 1.67 × gender + 2.34 × Lg (AFP [ng/ml]) + 0.04 × AFP-L3%% + 1.33 × Lg (PIVKA-II [mAU/ml]), where gender = 0 for females and 1 for males. Data count distributions were compared between groups using the χ2 test, with Fisher’s exact test utilized for small sample sizes. To ensure that the normality assumption was met, measurement data were compared between groups using the student’s t-test and the analysis of the variance model on the log scale. Categorical variables were expressed as percentages and continuous variables as means (standard deviations) or medians (interquartile ranges). The receiver operating characteristic (ROC) curves were used to determine the area under the curve (AUC) for AFP, PIVKA-II, or AFP-L3% alone and for combinations of two or three biomarkers to predict HCC. Comparisons among AUC values were performed using DeLong’s test (23). To evaluate the diagnostic performances of biomarker combinations, binary logistic regression was used to predict the probability of developing HCC. The AUC, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were used to report diagnostic performances. All statistical analysis were performed using Medcalc version 19.6.3 (MedCalc Software, Ostend, Belgium) for Windows and SPSS software version 25.0 (SPSS Inc., an IBM Company, Chicago, IL, USA). A two-tailed value of P < 0.05 was considered statistically significant.
Among 585 HCV-infected patients, 361 patients met eligibility criteria and were enrolled into the study ( Figure 1 ). The clinical characteristics of the HCV-HCC and HCV-control groups are summarized in Table 1 . Several clinical characteristics differed among patients in the HCV-HCC versus HCV groups including age, sex, Child-Pugh grading, cirrhosis, individual biomarkers of AFP, PIVKA-II, and AFP-L3%, and diagnostic scores of ASAP and GALAD, with all P < 0.05. Of note, median levels of the tumor biomarkers AFP, PIVKA-II, and AFPL3% were higher among HCV-HCC patients versus HCV-control patients (37.1 versus 4.3 ng/mL for AFP, 288.0 versus 22.0 mAU/mL for PIVKA-II, and 8.6% versus 0.5% for AFP-L3%; all P < 0.05) ( Table 1 , and Figure 2 ).
ROC curve analysis was used to evaluate the performance of individual biomarkers, as well as combination of the two prediction models. As shown in Table 2 and Figure 3 , both ASAP and GALAD demonstrated better diagnostic performance to diagnose HCC than the three individual biomarkers alone. The sensitivity, specificity, PPV, and NPV were presented in Table 2 . Using 20 ng/mL for AFP, 40 mAU/mL for PIVKA-II, and 10% for AFP-L3% as the clinical cut-off values, the sensitivity of AFP, PIVKA-II, and AFP-L3% to detect HCC among patients with HCV-HCC were 59.5%, 76.8%, and 48.2%, respectively; the specificity was 89.6%, 87.1%, and 80.3%, respectively. Of note, the AUC of PIVKA-II alone to -HCC was 0.859 (0.819-0.893), which was better than the performance of AFP alone (0.804 [0.759-0.843]; P = 0.084) or AFP-L3% alone (0.727 [0.678-0.773], P <0.001. The ASAP score had a similar diagnostic ability compared with the GALAD score (AUC of ASAP = 0.917 [0.884-0.943]; AUC of GALAD = 0.894 [0.858-0.924]).
HCV patients without HCC had a markedly lower prevalence of cirrhosis compared with HCV-HCC patients (72.0% vs. 87.5%, P < 0.05). Further analyses to assess the diagnostic performance of the ASAP score, the GALAD score, AFP, PIVKA-II, and AFP-L3% to detect HCV-HCC among patients with HCV-cirrhosis (HCC, n = 147; control group, n = 139) were then performed. ROC curve analysis was then used to assess the performance of the ASAP score, the GALAD score, and AFP, PIVKA-II, or AFP-L3% alone to diagnose HCV-HCC among patients with HCV-cirrhosis ( Figure 3 ). Similar to the overall cohort, subgroup analysis of patients with HCVcirrhosis demonstrated that the ASAP score and the GALAD score had higher sensitivity (80.3% and 81.0%, respectively) than any individual biomarker (60.5% for AFP, 76.9% for PIVKA-II, and 50.3% for AFP-L3%, respectively). The ASAP score had a better diagnostic ability versus the GALAD score with a higher AUC of 0.909 (0.870-0.940) (P < 0.001) among patients with HCC-cirrhosis ( Table 2 ).
Clinical characteristics of early-stage HCV-HCC according to the BCLC and TMN staging systems are shown in Supplement Table 1 . The diagnostic performance of the ASAP score, the GALAD score, AFP, PIVKA-II, and AFP-L3% to detect early-stage HCV-HCC was evaluated and compared. Among patients with early-stage HCC (BCLC stage 0+A, n = 91), PIVKA-II demonstrated better diagnostic performance (AUC 0.828, 0.779-0.870) compared with AFP (AUC 0.755, 0.701-0.804) alone or AFP-L3% (AUC 0.684, 0.627-0.738) alone. In addition, the ASAP score (AUC 0.898, 0.856-0.930) and the GALAD score (AUC 0.860, 0.815-0.899) performed better than any individual biomarker ( Figure 4 ). Similar findings were obtained using the 8th TNM staging system to define early-stage HCC. Specifically, the ASAP score and the GALAD score achieved higher AUC of 0.899 (0.857-0.931) and 0.870 (0.825-0.907), respectively, compared with individual biomarkers (AUC ranging from 0.704 to 0.839). Similar results were noted in the subgroups of liver cirrhosis patients. In particular, individual biomarkers and the two models achieved similar results in early-stage subgroups with AUC values ranging from 0.697 to 0.834 when defined by BCLC stage 0+A, and 0.714 to 0.847 when defined by the 8th TNM stage I ( Table 3 , and Figure 4 ). Results of sensitivity, specificity, PPV, NPV, and percent correctly classified when using different cutoff points are presented in Table 3 .
The diagnostic performances of ASAP and GALAD were further examined among HCV-HCC patients who had the HCC diagnosis missed using AFP alone. In particular, the diagnostic performances of the ASAP and GALAD scores in subgroups of HCV-HCC and HCC-cirrhosis among AFP-negative patients was evaluated. As noted in Table 4 and Figure 5 , the ASAP score demonstrated a better ability to distinguish AFP-negative HCC from HCV controls than the GALAD score (AUC of ASAP = 0.815 vs. AUC of GALAD = 0.764, P < 0.01) in the subgroups of patients with HCV-HCC; in fact, the sensitivity was 59.4% versus 73.9% and the specificity was 89.6% versus 70.0%. Moreover, the ASAP score demonstrated a better performance to discriminate AFP-negative HCC from HCC arising in the setting of cirrhosis (AUC of ASAP = 0.796 vs. AUC of GALAD = 0.752, P < 0.01) compared with the GALAD score with a sensitivity of 61.2% versus 76.3% and specificity of 87.1% versus 66.9%, respectively.
HCC surveillance is recommended especially among high-risk individuals as the mortality-to-incidence ratio of this disease is approaching one and curative-intent therapeutic options are limited among patients who have already progressed to intermediate or advanced stage disease when HCC is detected (9, 24). In the current multicenter study, the ASAP score demonstrated comparable diagnostic ability to the GALAD score among HCVinfected patients for HCC detection; in addition, the ASAP score had better ability than any individual biomarkers including AFP, PIVKA-II and AFP‐L3%. Similar results were noted in subgroup analyses among patients with HCV-cirrhosis, as well as patients with early-stage HCV-HCC, irrespective of which staging system (BCLC or TNM) was adopted. From the public or global health standpoint, both the ASAP score and the GALAD score can be used in parts of the world where medical resources are limited, and liver ultrasound is not widely available or easily affordable. One main advantage of these scores is that each is easy to calculate and can serve as excellent screening tests. As such, utilization of these diagnostic scores may increase the uptake of and compliance with surveillance and consequently improve the effectiveness of surveillance programs among at-risk populations of HCC. To achieve early detection of HCC, serological AFP and liver ultrasound are conventionally recommended (25, 26). However, ultrasound interpretation is operator-dependent and can be problematic in patients with central obesity or underlying cirrhosis (27). A recent single-center study analyzing 941 patients with cirrhosis reported that ultrasound alone was inadequate to exclude HCC in up to 20% of patients (28). As such, serologic biomarkers are needed to complement ultrasound to detect HCC (29). Use of serologic biomarkers scores may decrease the risk of surveillance-related variation associated with false positive ultrasound results, while also maximizing the potential benefits of early HCC detection by identifying patients with false negative ultrasound results (28). Serum biomarkers are promising tools for surveillance and early diagnosis of HCC owing to their noninvasive, objective, and reproducible characteristics. Among many proposed biomarkers, AFP, PIVKA-II, and AFP‐L3% are the HCC-specific biomarkers commonly used in current clinical practice. Although AFP has been widely used as a serum biomarker of tumor response for HCC, one of its significant limitations is that approximately 30-50% of patients with HCC are AFP “non-secretors” (30, 31). In fact, we noted that 43.7% of patients in our cohort were AFP non-secretors, and in this subgroup, PIVKA-II was a useful alternative serum biomarker as changes in its levels tracked with treatment response in 65% of AFP non-secretors. Data from the current study suggested that as an individual biomarker, PIVKA-II demonstrated the ability to diagnose HCV-HCC accurately; moreover, the diagnostic ability of PIVKA-II was better than that of AFP or AFP‐L3% used alone, with higher AUC (0.859 [0.819-0.893]) values and greater sensitivity (76.8% [69.7-82.9%]) and specificity (87.1% [81.5-91.4%]) than the other two biomarkers. PIVKA-II had comparable diagnostic efficacy for HCC detection independent of AFP-positive or negative status, making PIVKA-II a valuable supplement to AFP assessments. These conclusions were consistent with several other studies (32–34). Given that individual biomarkers are likely insufficient to detect HCC, utilizing combinations of these complementary markers may be helpful to diagnose HCC. Johnson P. J. et al. developed a serum-based tool (i.e., the GALAD model and associated GALAD score) for surveillance of HCC based on a UK cohort (21, 35). In 2019, the ASAP score was first developed based on participants recruited from 11 Chinese hospitals using a statistical model that could determine the risk of developing HCC in individual HBV patients using objective measures, which were mainly serological tumor markers (22). The ASAP score included age, sex, AFP, and PIVKA-II, while the GALAD score utilized age, sex, AFP, PIVKA-II, and AFP-L3%. AFP-L3 is one of the three glycosylated forms of AFP, and the level of AFP-L3% largely depends on the level of AFP. However, at the most frequently used cut-off value (10%), AFP-L3% has a specificity of 99.4% with a low sensitivity of 18.8%, indicating a poor ability to diagnose HCC as sensitivity takes priority over specificity in surveillance (36, 37). For the diagnosis of early-stage HCC, AFP-L3% is not recommended because of the need for an elevated AFP level, which limits its effectiveness; while AFP-L3% may serve as a supplementary for AFP, highly sensitive AFP-L3% measurements are region-restricted and costly (38, 39). As such, we constructed the ASAP score excluding AFP-L3% because the measurement of this marker is complex, time-consuming, expensive, and requires up to a 400-μL volume of serum sample; in addition, the contribution of AFP-L3% to risk prediction of HCC was low. As demonstrated in the current study’s results, the ASAP score performed comparably to the GALAD score. Furthermore, the ASAP score had a better discriminatory ability than the GALAD score in subgroup analyses of patients with HCV-cirrhosis and patients with early-stage HCV-HCC. From the public or global health standpoint, both the ASAP score and the GALAD score can be used in parts of the world where medical resources are limited, and liver ultrasound is not widely available or easily affordable. One main advantage of these scores is that each is easy to calculate and can serve as excellent screening tests. As such, utilization of these diagnostic scores may increase the uptake of and compliance with surveillance and consequently improve the effectiveness of surveillance programs among at-risk populations of HCC. The current study had several limitations. The retrospective study design had the inherent defect of selection bias. The present study population’s viral status (in particular, the serum HCV-RNA levels) was unknown. The study failed to account for anti-HCV therapy treatment history/status could have impacted the results. Furthermore, the diagnostic performances between liver ultrasound and serum biomarkers could not be compared in the present study. Considering that the GALAD score was constructed based on an UK cohort and the ASAP score was built on the basis of a Chinese cohort, the AUC of the ASAP score may have had a slight advantage with a database of Chinese patients used for the model validation. These two prediction models have not been compared in the international setting or relative to different stages of disease. In addition, the diagnostic performance of these two scores among HCC patients with other etiologies of chronic liver diseases, such as HBV infection, alcoholic liver disease, or nonalcoholic fatty liver disease, deserves to further evaluate in the future.
The ASAP score demonstrated excellent diagnostic performance among a large cohort of Chinese patients with HCV-HCC. The ASAP score had comparable or even better diagnostic ability compared with the GALAD score to detect any stage and early-stage HCV-HCC. In the future, comparison of the ASAP score with the GALAD score will be performed in a more extensive prospective multicenter cohort study and the costeffectiveness of the ASAP score versus the GALAD score with/without AFP-L3% will be further investigated before the comprehensive implementation of the prediction model in clinical practice.
The raw data supporting the conclusions of this article will be made available by the authors upon request, without undue reservation.
The studies involving human participants were reviewed and approved by The ethic committee of the Lishui Municipal Central Hospital. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. Written informed consent was not obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.
S-YL, CL, L-YS, and M-CG contributed equally to this work. Study concept and design: S-YL, CL, L-YS, M-CG, TP, X-MT, and TY. Drafting the manuscript: S-YL, CL, L-YS, and M-CG. Acquisition of data: S-YL, CL, L-YS, M-CG, LL, HX, M-DW, L-QY, FS, X-MT, and TY. Statistical analysis: S-YL, L-YS, TP, HX, and TY. Approval of the final version manuscript: All authors. Accountable for all aspects of the work: All authors. X-MT and TY had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors contributed to the article and approved the submitted version.
This work was supported in part by the Health Science and Technology Plan of Zhejiang Province (No. 2022KY453 for S-YL), the National Natural Science Foundation of China (No. 81972726 for TY), and Zhejiang Province Administration Foundation of Traditional Chinese Medicine (No. 2020ZB305 for S-YL).
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 |
|
PMC9576197 | D. Rajaraman,L. Athishu Anthony,P. Nethaji,Ravali Vallangi | One-pot synthesis, NMR, quantum chemical approach, molecular docking studies, drug-likeness and in-silico ADMET prediction of novel 1-(2,3-dihydrobenzo[b][1,4]dioxin-6-yl)-2-(furan-2-yl)-4,5-diphenyl-1H-imidazole derivatives | 15-10-2022 | Imidazole,DFT,Molecular docking,ADMET,Drug likeness | A novel drug to treat SARS-CoV-2 infections and hydroxyl chloroquine analogue, 1-(2,3-dihydrobenzo[b][1,4]dioxin-6-yl)-2-(furan-2-yl)-4,5-diphenyl-1H-imidazole (DDFDI) compound has been synthesized in one pot reaction. The novel compound DDFDI had been characterized by FT-IR, 1H-NMR and 13C-NMR spectral techniques. The geometrical structure was optimized by density functional theory (DFT) method at B3LYP/6-31G (d, p) as the basis set. The smaller energy value provides the higher reactivity of DDFDI compound than hydroxyl chloroquine and was corrected by high electrophilic and low nucleophilic reactions. The stability and charge delocalization of the molecule were also considered by natural bond orbital (NBO) analysis. The HOMO-LUMO energies describe the charge transfer which takes place within the molecule. Molecular electrostatic potential has also been analysed. Drug likeness and oral activity have been carried out based on Lipinski's rule of five. Molecular docking studies are implemented to analyse the binding energy of the DDFDI compound against Covid-19/6W41, COVID-19/6WCF, COVID-19/6Y84 and COVID-19/6W4B receptors and found to be considered as a better antiviral agents. | One-pot synthesis, NMR, quantum chemical approach, molecular docking studies, drug-likeness and in-silico ADMET prediction of novel 1-(2,3-dihydrobenzo[b][1,4]dioxin-6-yl)-2-(furan-2-yl)-4,5-diphenyl-1H-imidazole derivatives
A novel drug to treat SARS-CoV-2 infections and hydroxyl chloroquine analogue, 1-(2,3-dihydrobenzo[b][1,4]dioxin-6-yl)-2-(furan-2-yl)-4,5-diphenyl-1H-imidazole (DDFDI) compound has been synthesized in one pot reaction. The novel compound DDFDI had been characterized by FT-IR, 1H-NMR and 13C-NMR spectral techniques. The geometrical structure was optimized by density functional theory (DFT) method at B3LYP/6-31G (d, p) as the basis set. The smaller energy value provides the higher reactivity of DDFDI compound than hydroxyl chloroquine and was corrected by high electrophilic and low nucleophilic reactions. The stability and charge delocalization of the molecule were also considered by natural bond orbital (NBO) analysis. The HOMO-LUMO energies describe the charge transfer which takes place within the molecule. Molecular electrostatic potential has also been analysed. Drug likeness and oral activity have been carried out based on Lipinski's rule of five. Molecular docking studies are implemented to analyse the binding energy of the DDFDI compound against Covid-19/6W41, COVID-19/6WCF, COVID-19/6Y84 and COVID-19/6W4B receptors and found to be considered as a better antiviral agents.
In modern chemical synthesis of organic chemistry, multicomponent reactions have emerged as a powerful weapon to give a single synthetic product from the complex organic molecules. These multicomponent reactions have emerged mostly in organic synthesis and medicinal chemistry [1], [2], [3]. Multicomponent reactions are also known as tetrasubstituted compound due to their short reaction time, better efficiency, atom economy and operational simplicity [4,5]. In the current arena, interest for the synthesis of complex molecules in heterocyclic compounds, polymers, natural products, pharmaceutical and drug discovery are the new valuable approaches in MCR's [6], [7]. Development of a one pot MCR's especially polysubstituted compounds have been one of the current interest method of study amongst the researchers [8]. Amongst the entire heterocyclic compound, imidazole possessed the central core due to their numerous applications as nitro groups are present in the structure [9]. Imidazole shows a wide spectrum of biological activities like antibacterial [10], antagonist [11], anti-cancer [12], anti-viral and anti-inflammatory [13], anti-oxidant [14], antifungal [15], cytotoxicity [16]. This imidazole shows great applications in natural compounds such as histamine [17], algeacidal [18], pilocarpine alkaloids [19]. In the imidazole skeleton which contain drugs are losartan (antihypertension), etomidate (hypnotic agent) and flumazenil [20]. These drugs are broadly supplied and employed. With these abundant applications of imidazole, researchers are putting their efforts to construct more imidazole moiety. From the above observation, development of more practical and versatile method from the available resources is one of the most vital features. Thus in this study, we aim to synthesize a novel derivative of 1-(2,3-dihydrobenzo[b][1,4]dioxin-6-yl)-2-(furan-2-yl)-4,5-diphenyl-1H-imidazole (DDFDI)compound in one-pot reaction of 2,6-bis(4-chlorophenyl)-3-methylpiperidine-4-one, acetic acid and 2,4,6-trichlorophenyl)hydrazine in the presence of ethanol as a catalyst. The prediction of molecular geometry parameters, HOMO-LUMO orbitals, intramolecular charge transfer, NBO activity and molecular electrostatic potential (MEP) of the target molecule were carried out in B3LYP/6-31G (d,p) level of theory to study all the reactivity and biological importance of the target compound. The DDFDI compound has been experimented in the molecular docking studies and found better binding results in the pharmaceutical arena than the target drugs.
All the chemicals and solvents were purchased from Sigma Aldrich and chemical suppliers, used as received without further purification. The DDFDI compound was recorded by the Fourier transform-infrared (FTIR) spectra in the range of 4000-400cm−1 using AVATAR 300FTIR. The 1H-NMR and 13C-NMR spectra of the title compound were recorded at 400MHz on Bruker AMX 400MHz spectrometer and 100MHz on BRUKER AMX400MHz spectrometer using Dimethyl sulfoxide (DMSO4) as a solvent. The melting points were verified by open capillaries and uncorrected.
In the present work, quantum mechanical calculations were performed with Gaussian-03 program. At the very first step of DDFDI calculations, the geometry occupied from the preliminary structure was fully optimized by density functional theory (DFT) using the Becke3-Lee-Yang-Parr (B3LYP) with standard 6-31 G (d, p) as the basis set [21,22].
Molecular docking recreation was performed with the Argus Lab 4.0. The readied 3D structures of different protein was downloaded from the protein information bank (see http//www.rcsb.org/pdb) and tying webpage was made by picking "Making tying website for this protein" alternative. The ligand was then presented and docking figuring was permitted to run utilizing shape-based pursuit calculation and A Score scoring capacity. The scoring capacity is in charge of assessing the vitality between the ligand and protein target. Adaptable constructing so as to dock was permitted lattices over the coupling locales of the protein and vitality based turn is set for that ligand gathering of molecules that do not have rotatable bonds. For every pivot, torsions and made and postures (adaptation) are created amidst the docking process. For every perplexing 10 free runs were led and one posture was returned for every run. The best docking model was chosen by least tying vitality computed by Argus lab and the most suitable tying adaptation was chosen on the premise of hydrogen security connection between the ligand and protein close to the substrate tying site. The most minimal vitality postures show the most elevated tying proclivity as high vitality creates the temperamental adaptations. The subsequent receptor model was spared to Brookhaven PDB document from the record the 2D and 3D connections are seen in revelation studio 4.5 renditions [23].
Drug likeness and ADMET are the current vital method to predict the potential drug candidates. At first, it emerged for the preliminary assessment of the pharmacokinetics, physicochemical and drug likeness parameters in the drug discovery process. ADMET outlooks for Absorption, Distribution, Metabolism, Excretion and toxicity. Prediction of the drug likeness of the title compound was evaluated by rule based filters from Lipinski, Ghose, veber, Egan, Muegge and synthetic accessibility difficulty level was from 1-10 [24].
A mixture benzil (6.0mmol), ammonium acetate (24.0mmol), 2,3-dihydrobenzo[b][1,4]dioxin-6-amine (27.0mmol) and furan-2-carbaldehyde (9.0mmol) in absolute ethanol (20ml) in the presence of C4H10BF3O (2/3drops) as a catalyst. The reaction mixture was refluxed for about 24hrs at the boiling point of ethanol (78°C) and upon completion of the reaction, the reaction mixture was cooled and Thin Layer Chromatography is a technique used to isolate non-volatile mixtures. The experiment is conducted on a glass which is coated with a thin layer of adsorbent material. The material usually used is aluminium oxide.. Thin layer chromatography (TLC) technique was monitored by using ethyl acetate: benzene (2:8) as the eluent. The reaction mixture was extracted with dichloromethane and the product was purified by column chromatography. The final product was recrystallized from ethanol by slow evaporation and harvested the pure compound of 1-(2,3-dihydrobenzo[b][1,4]dioxin-6-yl)-2-(furan-2-yl)-4,5-diphenyl-1H-imidazole (DDFDI).
White solid: m.p. = 178-190°C and yield 92%: IR (Neat, cm−1);1599 (C=N), 1245 (C-O), 1441, 1403, 1294 (C=C), 2319-3052 (C-H),1H-NMR (CDCl3, ppm);4.39 (m, CH2benzodioxane), 6.34(d, H-20, furfural moiety), 6.71 (d, H-22, furfural moiety),7.14-8.53ppm (aromatic proton).13C-NMR (CDCl3, ppm);53.29, 64.33 (2CH2, benzodioxane), 109.25, 111.33, 123.02 (carbon signal for furfural moiety), 156 (C=N), 124.04-141.31(aromatic carbon). Chemical Formula: C27H20N2O3, Exact Mass: 420.15 Molecular Weight: 420.46 m/z: 420.15 (100.0%), 421.15 (29.5%), 422.15 (5.0%) Elemental Analysis: C, 77.13; H, 4.79; N, 6.66; O, 11.42.
A schematic diagram of compound DDFDI is shown in Scheme 1 . The FT-IR spectral analysis of the DDFDI compound is discussed below. Generally, imines indicate a strong C=N stretching vibration in the region 1500-1600cm−1. The DDFDI compound demonstrate strong absorption band and were detected in the region 1541-1618cm−1 due to the functional group of imine. The C=N stretching band appears at 1599cm−1 in DDFDI compound showed high absorption band. The existence of C=N stretching band supports the skeleton of the imidazole ring. A concrete absorption band appeared at 1245cm−1 due to furan and dioxin C-O stretching. The aromatic C-H stretching band arises in the broad region 2319-3052cm−1. The observed imine, aliphatic and aromatic C-H stretching frequencies were confirmed the compound DDFDI.IRspectral values of compound DDFDI is displayed in Fig. 1 . 1H-NMR spectra of DDFDI have been recorded in CDCl3 solvent. The signals were obtained in the 1H-NMR spectra were consigned and established accordingly to their position, multiplicities and integral values. In general, the aromatic proton signals emerged in the higher frequency region at 7.00ppm due to the magnetic anisotropic effect. In the 1H-NMR spectrum of DDFDI compound, the signals appeared in the region7.14-8.53ppm resemble to sixteen protons integral due to aromatic protons. A multiplet peak detected at 4.39ppm is assigned to two CH2 protons (2CH2 forbenzodioxane moiety). A douplet peak detected at 6.34ppm is assigned to H-20 proton signal of furfuryl moiety. A doublet peak was appeared at 6.71ppm was assigned to H-22 proton signal for furfural moiety.1H-NMRspectral values of compound DDFDI is displayed in Fig. 2 . The 13C-NMR spectra of 1-(furan-2-yl)methyl)-4,5-diphenyl-2-(p-tolyl)-1H-imidazole have been recorded. The 13C-NMR chemical shifts are quoted after rounding off to one decimal point. The 13C-NMR spectrum of compound DDFDIwas recorded at 400MHZ instrument.In 13C-NMR spectrum of compound DDFDI, the two CH2 carbon signals of benzodioxane moiety appeared at 53.29ppm and 64.33ppm. The C-8, C-9 and C-10 carbon signals were appeared at 109.25, 111.33 and 123.02ppm due to furfural moiety. The aromatic and ipso-carbon signals are appeared in the region of 124.04-141.31ppm. The imidazole compound containing C=N carbon signal appeared at 156.02ppm.13C-NMR spectral values of compound DDFDI is displayed in Fig. 3 .
The DDFDI compound was computed in Gaussian 03 software package using functional B3LYP at 6-31G (d,p) as the basic level theory. The substitution of functional groups in this compound, electron donor and electron acceptor are observable in the wavelength and oscillator strengths. In DDFDI compound, the bond lengths are observed in C30-O29, C24-N27, C24-N26, C38-O50, C23-N27, C45-O49 and C25-C23 are 1.45Å, 1.33Å, 1,52Å, 1.42Å, 1.44Å, 1.43Å and 1.49Å respectively. The bond angles in C31-C30-O29, C30-O29-C28, C28-C24-N27, C24-N27-C23, C24-N26-C25, C46-O50-C38, C45-O49-C39 and C40-C39-O49 are 111.49°, 102.52°, 123.43°, 102.25°, 98.79°, 112.64°, 112.61° and 120.11° respectively. While the dihedral angles at H33-C30-O29-C28, N27-C24-C25O29, C2-C25-C23-N27, N26-C24-C28-O29, C36-C37-C38-O50, H47-C45-O49-C39, C41-C40-C39-O49 and H48-C46-O50-C38 are identified at 167.79°, 0.79°, -129.85°, 178.79°, 176.15°, 170.90°, 176.21° and 170.82° respectively. When we compared our theoretical (DFT) values with experimental values of bond lengths O1-C23, N1-C7 and N2-C22 are 1.35Å, 1.32Å and 1.44Å, the bond angles are C23-O1-C26, C7-N1-C8 and C7-N2-C22 are 105.90°, 105.84° and 126.23° and the dihedral angles of N2-C15-C16C17 and O1-C23-C24-C25 are 96.00° and -2.05° respectively were found to be alike [23]. From this observed bond length, bond angle and dihedral angle, the compound (FMDI) have planar geometry. The optimized structure of DDFDI compound is shown in Fig. 4 . The optimized parameters namely bond lengths, bond angles and dihedral angles are of an isolated molecule in gaseous phase (Table 1 ).
Natural bond orbital (NBO) is a special technique to study intramolecular and intermolecular bonding, hybridization, interaction bonding and charge transfer in the molecular structure [23]. The NBO calculation was carried out in the Gaussian 03 package at the DFT/B3LYP level using 6-31G (d,p) as the basis set to recognize numerous second order interactions between occupied orbitals and unoccupied orbitals is shown in Table 2 . The second order Fock matrix was performed to calculate the donor-acceptor interactions in the NBO analysis. For each donor (i) and acceptor (j), the stabilization energy E2 is related with the delocalization i-j is estimated as Where qi = donor orbital occupancy εi and εj = diagonal elements F (i, j) = off diagonal NBO Fock matrix element. The stabilization energy (E2) values verified the hyperconjugative interaction and charge transfers by the orbital overlap were determined between πC28-C32 to σ*C37-H42 for anti-bonding orbital with stabilization energy 81.59kJ/mol. The larger energy value of E2 gives more chemical stability in the molecular interaction between electron donors and electron acceptors [25].
The highest occupied molecular orbitals (HOMO's) and lowest unoccupied molecular orbitals (LUMO's) are the types of frontier molecular orbitals. The HOMO has capacity to donate an electron whereas LUMO represents the ability to accept electron. These orbitals are important to decide the stability of the chemical compound. The HOMO and LUMO energy values was calculated through DFT method using B3LYP/6-31G (d,p) level of theory are shown below. Their main energy gap = 0.205eV, 0.198eV, 0.198eV. The transfer of electron from ground state to the excited state is labelled by one electron from highest occupied molecular orbitals to the lowest unoccupied molecular orbitals. As we can see from the given figure, HOMO shows the charge density contained over the entire phenyl ring. The LUMO part is situated on imidazole, furfuraldehyde and 1,4-benzodioxan-6-amine ring.The displayed diagram of HOMO-LUMO energy value is shown in Fig. 5 . The energy gap between HOMO and LUMO gives the detail interaction between the molecules and chemical stability [26].
The map of molecular electrostatic potential delivers the isosurface values with the position of positive and negative electrostatic potentials and broadly used for electrophilic reactions, nucleophilic reactions, biological process, hydrogen bonding and intermolecular interactions. MEP map of the DDFDI compound was calculated by DFT method at B3LYP/6-31G (d,p) as the level of theory. The various values of the electrostatic potential at the surface are symbolized by different colours i, e, red colour denotes negative charge, partial blue colour indicate positive charge, yellow colour shows electron rich region and white colour represents region of zero potential [27]. In the present work, molecular geometry and anharmonic vibrational spectra of o-, m-, p-iodonitrobenzene have been studied. The anharmonic frequencies were calculated using second order perturbative (PT2) approach with basis set 3-21G⁄ on iodine and 6-311G(d,p) on other atoms at DFT(B3LYP) level of theory and were compared to experimental values. The assignments of vibrational modes of isomeric iodonitrobenzenes were done by using potential energy distribution (PED) and vibrational assignments of benzene, nitrobenzene and iodobenzene. The combination and overtone bands are also assigned. The electronic spectra were recorded as well as simulated using polarizable continuum model (PCM) at TD-B3LYP/6-311G(d,p)/3-21G⁄ level of theory. The vibrational and electronic spectra are interpreted. Moreover, atomic charges, MEP mapping, HOMO–LUMO, NBO analysis and various thermodynamics and molecular properties are reported. The MEP map shows the chemical active location and chemical reactivity of atoms as shown in Fig. 6 .
Mulliken atomic charge plays a significant role in the application of quantum chemical calculation to molecular system. In the mulliken atomic charge calculations, atomic charges affect the dipole moment, polarizability, electronic structure and the properties of molecular systems. The compounds DDFDI are calculated mulliken atomic charge are recorded in Table 3 . Thus, from the distribution charge calculation, we can see all the heteroatoms displayed major electron density. Therefore, from the given table …the carbon atoms C1, C2, C3, C4, C5, C6, C12, C13, C14, C16, C17, C31, C32, C37, C40 and C41 possess small negative charges whereas carbon atoms C15, C23, C24, C25, C28, C36, C38, C39, C45 and C46 possess positive charges. The DDFDI compound illustrated that the high positive charges in a molecule are C24=0.36 and C39=0.33. The positive regions are associated to nucleophilic reactivity. These data clearly demonstrate that compound DDFDI are the most reactive parts taking place towards substitution reactions [28].
The hyperpolarizability (β0), dipole moment (µ) and polarizability (α) of the title compound were calculated using DFT method with 6-31G (d,p) as the level of theory. The entire equations for calculating the total dipole moment (µ), polarizability (α) and first hyperpolarizability(β0) using x, y, z components from Gaussian 03W output is shown below: herefore, the higher values of dipole moment, polarizability and hyperpolarizability (β0) are the efficient potential in NLO materials are shown in Table 4 . Our current work reveals that the π-π interactions give more intra-molecular interaction and hence the polarizability of the molecule increases. The β0 value of the DDFDI (2.40 × 10−30esu) compound is ∼7 times greater than that of urea (0.3728 × 10−30esu) [28]. Thus, we can say that the DDFDI compound has more potential in nonlinear optical properties.
The molecular docking analysis of DDFDI ligand with COVID-19/6W41, COVID-19/6WCF, COVID-19/6Y84 and COVID-19/6W4B receptors was performed. For structure-based drug design, the molecular docking plays an important role in the biological studies and to understand the ligand receptor interactions. The particular treatment for COVID-19 is vacant till date, so from researchers many antiretroviral drugs against COVID-19 were reported and existing such as Atazanavir, Darunavir ritonavir, lopinavir, oseltamivir, remdesivir, chloroquine and hydroxyl chloroquine. Some of the imidazole drugs also reported against COVID-19 [29]. We decided to do molecular docking analysis of DDFDI compound that we can recommend against COVID-19. The molecular docking mechanism between DDFDIligand and the COVID-19/6W41, COVID-19/6WCF, COVID-19/6Y84 and COVID-19/6W4B receptor was examined and evaluated. Firstly, 6W41 is Crystal structure of SARS-CoV-2 receptor binding domain in complex with human antibody CR3022. In the binding mode, the DDFDI compound was attractively bound to 6W41 via Vander Waals, carbon hydrogen bond, π-sigma, π-lone pair, π-π T-shaped, π-alkyl bond interactions. The residue ALA621, CYS490 is binding with phenyl group in the imidazole ring with distance 6.33Å, 4.13Å by π-alkyl bond. The residue LYS246 attached in phenyl ring, imidazole and furan ring having a bond distance 4.65Å, 4.66Å, 4.33Å by Vander Waals bond. The residues LEU489, ILE244, LEU616 is binding with phenyl, imidazole and furan moiety with various bond distance 5.01Å, 5.09Å, 5.12Å,4.77Å, 4.17Å, 5.65Å by π-sigma bond. The residue HIS618 is binding with dioxane moiety having a bond distance of 4.84Å, 3.13Å, 6.55Å by π-lone pair bond. But, the standard drug hydroxylchloroquine is enclosed with carbon-hydrogen bond, mixed alkyl and Pi-alkyl bond interactions of the amino acids GLU438 (3.45, 3.08), PHE441, VAL466, LEU467, PHE473, TRP535 residues having a bond distance 4.07Å, 5.09Å, 4.38Å, 4.89Å, 3.39Å and 6.12Å via π-alkyl and alkyl bond. The binding energy of DDFDI compound is -11.08kcal/mol while the standard drug (hydroxyl chloroquine) is - 9.21kcal/mol. The Docking of 2D and 3D images of compound DDFDI and standard drug (hydroxylchloroquine) with 6W41 receptor were shown in Fig. 7. Secondly, the interactions between DDFDI ligand and 6WCF receptor. 6WCF is a Crystal Structure of ADP ribose phosphatase of NSP3 from SARS-CoV-2 in complex with MES. SARS-CoV-2 main protease has a vital role in the processing of polyprotein that is translated from viral RNA, and the protease is considered as the main role for viral survival and growth. The compound was well bound to 6WCF via π-π stacked, amide-π stacked, π-sigma and Vander Waals. The π-π stacked, amide-π stacked interactions of the amino acids GLY47, PHE132, ALA38 have bond distance 3.96Å, 3.18Å and 4.67Å is binding with phenyl moiety. The amino acids ILL131 have bond distance 5.19Å, 4.01Å, 3.05Å, 5.40Å, 2.50Å is binding with furan, imidazole, phenyl and dioxane moiety via π-sigma. The Vander Waals has bond distance 2.08Å of GLY130 residue is attached to dioxane moiety. But the standard hydroxylchloroquine is enclosed with mixed alkyl and π-alkyl interactions, conventional hydrogen bond and π-lone pair between ligand and receptor. The residues ALA129, LEU126, VAL49, ILE131 are attached to methyl ring and chorophenyl moiety have bond distance 3.03Å, 4.53Å, 4.93Å, 4.68Å via alky and π-alkyl interactions. The amino acids of ASP157, PHE156 with a bond distance 2.98Å, 2.67Å is attached with oxane moieties. The residue of PHE132 have bond distance 3.62Å, 2.84Å is attached to pyridine and chlorophenyl moiety via π-lone pair bond. The binding energy of compound DDFDI is -9.75kcal/mol while the standard drug (hydroxyl chloroquine) is - 8.31kcal/mol. The docked 2D and 3D images of compound DDFDI and standard drug (hydroxyl chloroquine) with 6WCF receptor were shown in Fig. 8. Thirdly, the interactions between DDFDI ligand and 6Y84 receptor. 6Y84 is the COVID-19 main protease with un-liganded active site. SARS-CoV-2 main protease has a vital role in the processing of polyprotein that is translated from viral RNA, and the protease is considered as the main role for viral survival and growth. The compound was well bound to 6Y84 via π-donor hydrogen bond and π-π T-shaped. The pi-donor hydrogen bond and π-π T-shaped interactions of the amino acids TYR101 (2.81Å) and PHE103 (3.61Å) is binding with phenyl moiety. But the standard hydroxylchloroquine is enclosed with Vander Waals, carbon hydrogen bond, mixed alkyl and π-alkyl interactions between ligand and receptor. The residues PHE294 is attached to pyridine and chlorophenyl moiety have bond distance 4.16Å, 3.18Å and 3.27Å via carbon hydrogen bond. The amino acids of PRO293, ILE249, CYS300, LEU253, VAL297, PRO252 with a bond distance 4.64Å, 3.55Å, 4.01Å, 5.12Å, 4.73Å, 3.50Å, 6.38Å, 5.06Å, 5.28Å is attached with methyl, chlorophenyl and phenyl moieties. The binding energy of compound DDFDI is - 10.62kcal/mol while the standard drug (hydroxyl chloroquine) is -8.68kcal/mol. The docked 2D and 3D images of compound DDFDI and standard drug (hydroxyl chloroquine) with 6Y84 receptor were shown in Fig. 9. Lastly, the interactions between DDFDI ligand and the 6W4B receptor. 6W4b is the crystal structure of Nsp9 RNA binding protein of SARS CoV-2. SARS-CoV-2 main protease has a vital role in the processing of polyprotein that is translated from viral RNA, and the protease is considered as the main role for viral survival and growth. The compound was well bound to 6W4b via π-sigma, π-sulfur, π-π T-shaped and π-alkyl. The π-π T-shaped interaction of the amino acids was attached by the residue PHE91 (4.3Å) while the residue LEU89, VAL0 is attached to the phenyl group has bond distance 4.42Å, 4.20Å via π-alkyl bond. The residue CYS74, MET0 is attached to phenyl ring, imidazole and furan moiety have bond distance 3.01Å, 5.10Å, 4.83Å via π-sulfur bond whereas the amino acids LEU107, LEU104, LEW5 are attached to phenyl, furan and imidazole moiety with bond distance 3.09Å, 3.76Å, 3.41Å, 4.58Å, 4.82Å, 5.09Å, 5.30Å via π-sigma bond. But the standard hydroxylchloroquine is enclosed with alkyl and π-alkyl interactions between ligand and receptor. The residues PHE76, LEU46, ALA79, LEU5, LEU113, LEU104, VAL111, LEU107 with bond distance 4.13Å, 3.68Å, 4.64Å, 4.26Å, 4.64Å, 4.54Å, 3.13Å, 3.68Å, 5.15Å, 3.77Å, 5.01Å, 4.58Å were attached to phenyl, chlorophenyl and methyl moiety. The binding energy of compound DDFDI is -12.25kcal/mol while the standard drug (hydroxyl chloroquine) is -10.18kcal/mol. The docked 2D and 3D images of compound DDFDI and standard drug (hydroxyl chloroquine) with 6Y84 receptor were shown in Fig. 10. The binding energy values for synthesized compound DDFDI and standard drug for different proteins are shown in Table 5 . From the above molecular docking results, we concluded that DDFDI compound can be considered as potential agent against COVID-19/6W41/6WCF/6Y84/6W4b receptors (Table 4 ).
Computational techniques and prediction of the various physicochemical and pharmacokinetics of novel drug candidate is one of the angle for the drug development process to save time, effort and cost. The word drug-likeness indicated the stability between the structure characteristic and various molecular properties which determine the drug discovery and production. The five rules of Lipinski plays the main role in the discovery of drugs and in this designed compound the five rules were employed to determine the bioavailability of bulk materials to examine the drug-likeness properties and the results shows no violation of the rule. The physicochemical properties with their ranges as described by the Lipinski's rule of five which includes; Molecular weight is found to be 420.45g/mol (<500g/mol), hydrogen bond donor is 0 (<5), hydrogen bond acceptor is 4 (<5), High lipophilicity LogP is 4.93 (<5), Van der Walls topological polar surface area (TPSA) value is 49.42Å^2(120Å^2). Hence, the compound DDFDI obeys the five rules of Lipinski's. The ADMET of the designed compound were used admet SAR database (http://lmmd.ecust.edu.cn/admetsar1/predict). The features of ADMET properties of the title compound to act as drug leads such as blood-brain barrier (BBB) penetration and gastrointestinal absorption (GI), water soluble capability, lipophilicity and CYP1A2 inhibitor. The ADMET result shows that the studied compound has well absorbed by the gastrointestinal tract and easily flow into the brain to reach the target enzyme. Similarly, the effect of cytochrome P450 enzymes (CYP1A2, CYP2C19 and CYP3A4) for drug metabolism in humans and the toxicity properties were non-inhibitors of CYP2C9 and CYP2D6 cytochrome P450 enzymes. Skin permeability is considered in transdermal drug delivery and evaluates product efficacy, the studied compound show low skin permeability at -4.98cm/s. Here, skin permeability shows that the compound has the ability to reduce skin allergy when directed and the results proved that the compound is harmless and cure for skin allergies [30], [31], [32]. The ADMET prediction results of compound are given in Table 6.
1-(2,3-dihydrobenzo[b][1,4]dioxin-6-yl)-2-(furan-2-yl)-4,5-diphenyl-1H-imidazole derivatives (DDFDI) have been synthesized in the presence of C4H10BF3O. The compound DDFDI have been characterized by IR, 1H-NMR and 13C-NMR spectral techniques. Theoretical calculation was carried out for compound DDFDI using DFT/B3LYP/6-31G (d,p) basis set and their optimized bond parameters were calculated.The stabilization energy (E2) values verified the hyperconjugative interaction and charge transfer by the orbital overlap were determined between π(C28-C32) to σ*(C37-H42) for anti-bonding orbital with stabilization energy 81.59kJ/mol. The larger energy value of E2 gives more chemical stability in the molecular interaction between electron donors and electron acceptors.HOMO shows the charge density contained over the entire phenyl ring. The LUMO part is situated on imidazole, furfuraldehyde and 1,4-benzodioxan-6-amine ring. The molecular energies of HOMO, HOMO-1, HOMO-2 levels are -0.222, -0.224 and -0.226eV respectively while LUMO, LUMO+1, LUMO+2 levels are 0.017, 0.026, 0.028eV, respectively. The band gap energy between HOMO and LUMO is 0.205eV.Nonlinear optical studies revealed that the π-σ* interactions tend more intra-molecular interaction and hence the polarizability of the molecule increases. The β0 value of the DDFDI compound is ∼7 times greater than that of urea. Drug likeness predicts the oral activity and ADMET property analysis gives an idea about the pharmacokinetic properties of the title molecule. Molecular docking studies reveals that the compound DDFDI exhibit more binding energy are -11.08, -9.75, -10.62 and -12.25kcal/mol while the standard drug (hydroxyl chloroquine) is -9.21, -8.31, -8.68 and -10.18kcal/mol with different Covid-19/ 6W41/ 6WCF/ 6Y84/ 6W4B receptors. Finally, molecular docking results have shows that the compound DDFDI can be considered as a potential antiviral agent.
D. Rajaraman: Conceptualization, Supervision, Investigation, Methodology, Resources, Formal analysis, Data curation, Writing – original draft. L. Athishu Anthony: Conceptualization, Investigation, Methodology, Resources, Formal analysis, Data curation, Writing – original draft. P. Nethaji: Software, Resources. Ravali Vallangi: Software, Resources.
I declare no conflict of interest. | true | true | true |
|
PMC9576395 | Yuhua Cai,Yunpeng Li | LncRNA Gm43843 Promotes Cardiac Hypertrophy via miR-153-3p/Cacna1c Axis | 10-10-2022 | Long noncoding RNAs (lncRNAs) have been reported to engage in many human diseases, including cardiac hypertrophy. Cardiac hypertrophy was mainly caused by excessive pressure load, which can eventually lead to a decline in myocardial contractility. Gm43843, a novel lncRNA, has not been well explored in cardiac hypertrophy so far. Herein, we are going to search the function and the underlying molecular mechanism of Gm43843 in cardiac hypertrophy. Gm43843 levels were measured via qRT-PCR in mouse myocardial cells when they are treated with angiogenin II (Ang II) or transfected with different plasmids. Western blot assay was implemented to detect the cardiac hypertrophy-related protein markers, while the cell was analyzed via immunofluorescence (IF) assay to evaluate the hypertrophy. Meanwhile, the binding of Gm43843 and the putative targets was examined based on mechanistic assay results. We found that Gm43843 expression was increased with the elevated concentration of Ang II. Inhibited Gm43843 was detected to reduce the hypertrophy of mouse myocardial cells. Meanwhile, Gm43843/miR-153-3p/Cacna1c axis was found to modulate cardiac hypertrophy. In short, Gm43843 promotes cardiac hypertrophy via miR-153-3p/Cacna1c axis. | LncRNA Gm43843 Promotes Cardiac Hypertrophy via miR-153-3p/Cacna1c Axis
Long noncoding RNAs (lncRNAs) have been reported to engage in many human diseases, including cardiac hypertrophy. Cardiac hypertrophy was mainly caused by excessive pressure load, which can eventually lead to a decline in myocardial contractility. Gm43843, a novel lncRNA, has not been well explored in cardiac hypertrophy so far. Herein, we are going to search the function and the underlying molecular mechanism of Gm43843 in cardiac hypertrophy. Gm43843 levels were measured via qRT-PCR in mouse myocardial cells when they are treated with angiogenin II (Ang II) or transfected with different plasmids. Western blot assay was implemented to detect the cardiac hypertrophy-related protein markers, while the cell was analyzed via immunofluorescence (IF) assay to evaluate the hypertrophy. Meanwhile, the binding of Gm43843 and the putative targets was examined based on mechanistic assay results. We found that Gm43843 expression was increased with the elevated concentration of Ang II. Inhibited Gm43843 was detected to reduce the hypertrophy of mouse myocardial cells. Meanwhile, Gm43843/miR-153-3p/Cacna1c axis was found to modulate cardiac hypertrophy. In short, Gm43843 promotes cardiac hypertrophy via miR-153-3p/Cacna1c axis.
Cardiac hypertrophy is a slow but effective compensatory function, which happened under long-term pressure overload [1]. Pathological cardiac hypertrophy poses a high risk of myocardial ischemia, which will lead to a deficiency in myocardial contractility, and eventually causes heart failure [2]. The major treatment for cardiac hypertrophy is surgical therapy [3]. Nevertheless, a lack of knowledge in the pathogenesis of cardiac hypertrophy makes the prevention of this disease difficult [4]. Thence, it has become vitally important for us to explore the underlying molecular mechanism in cardiac hypertrophy. Noncoding RNAs (ncRNAs) are a group of genes, lacking protein-coding ability, but play important roles in modulating the biological behavior of cells [5–8]. Meanwhile, increasing long ncRNAs (lncRNAs) have been found to have a connection with the progression of cardiac hypertrophy by acting as competing endogenous RNAs (ceRNAs), that is, to sponge microRNAs (miRNAs) and modulate downstream messenger RNA (mRNA) expression [9]. For instance, lncRNA MIAT has been revealed to sequester miR-93 and regulate the expression of TLR4 in cardiac hypertrophy, functionally promoting the progression of cardiac hypertrophy [10]. Additionally, lncRNA MIAT contributes to cardiac hypertrophy by modulating the miR-93/Akt3 axis [11]. LncRNA-ROR can also modulate the progression of cardiac hypertrophy via miR-133 [12]. Gm43843 is a novel lncRNA that has not been well investigated so far. Limited evidence has suggested that lncRNA Gm15834 is allowed to facilitate myocardial hypertrophy by serving as a miR-30b-3p sponge and elevating ULK1 expression [13]. In our study, we are going to search for the function of Gm43843 in cardiac hypertrophy. MiR-153-3p is a crucial regulator identified in various diseases. Specifically, in cardiac diseases, it has been found that miR-153-3p contributes to mitochondrial fragmentation in cardiac hypertrophy [9]. The regulatory influence of miR-153-3p on cardiomyocyte apoptosis by directly targeting βII spectrin has also been uncovered [14]. Recent evidence has also pointed out that miR-153-3p is able to affect cardiomyocyte apoptosis induced by formaldehyde [14]. Herein, we aim to figure out whether miR-153-3p is a participant of the Gm15834-centered ceRNA regulatory axis. mRNA calcium voltage-gated channel subunit alpha1C (Cacna1c) has been recognized as the effector of a wide range of neuropsychiatric syndromes [15]. Importantly, it has once been revealed that Cacna1c targeted by miR-221/222 is related to the change in cardiac ion channel expression and current density [16]. Cacna1c is also essential for cardiac electrophysiological development and maturation [17]. Therefore, it is one of the major targets of our study to uncover the function of Cacna1c in cardiac hypertrophy. Atrial natriuretic factor (ANF), brain natriuretic peptide (BNP), and β-myosin heavy chain (β-MHC) have been known as the biomarker of cardiac hypertrophy progression [11, 18]. In our study, these markers were detected to reflect hypertrophy. Meanwhile, the surface area of mouse myocardial cells was also assessed for investigating the hypertrophy variation. In this study, the specific molecular mechanism of lncRNA Gm43843 in cardiac hypertrophy will be scrutinized, with the ceRNA network taken into consideration.
The mouse myocardial cells (H9C2 and MCM) were available from the ATCC (Manassas, VA). The cell culture environment was kept with 5% CO2 at 37°C. Dulbecco's modified Eagle's medium (DMEM) supplemented with 1% antibiotics and 10% fetal bovine serum (FBS) was procured from Gibco (Grand Island, NY). To induce cardiac hypertrophy, H9C2 and MCM cells were severally processed with angiotensin II (Ang II; Sigma–Aldrich, St. Louis, MI) at 0.5 or 1 mmol/L concentration for 48 h.
Total RNA was isolated from cardiomyocytes utilizing TRIzol reagent (Thermo Fisher Scientific, Waltham, MA), followed by reverse transcription with the application of a reverse transcription system (Thermo Fisher Scientific). qRT-PCR was conducted on ABI 7900 Detection System (Applied Biosystems, Foster City, CA) by use of the SYBR-Green PCR Master Mix kit (Takara, Shiga, Japan). Relative expression of genes, normalized to GAPDH or U6, was calculated via the 2−ΔΔCt approach.
The synthesized short hairpin RNAs (shRNAs) and control-shRNAs (GenePharma, Shanghai, China) were available to silence Gm43843 and Cacna1c using Lipofectamine3000 (Thermo Fisher Scientific). In addition, the miR-153-3p mimics and NC mimics, miR-153-3p inhibitor and NC inhibitor, as well as the pcDNA3.1-Cacna1c and pcDNA3.1-NC were all available from GenePharma for 48 h of plasmid transfection.
Total protein was extracted from cells by use of RIPA lysis buffer (Beyotime, Shanghai, China). Thereafter, the separation of proteins was achieved by using 10% SDS-PAGE (Bio-Rad, Hercules, CA), and the samples were then moved to PVDF membranes (Millipore, Bedford, MA). Following sealing with 5% fat-free milk, the membranes were cultivated at 4°C overnight with primary antibodies for ANP (1 : 2000; Abcam, Cambridge, MA), BNP (1 : 2000; Abcam), β-MHC (1 : 2000; Abcam) and control GAPDH (1 : 2000; Abcam). Later, secondary antibodies (1 : 5000; Abcam) were added to the culture for 1 h at 37°C. Finally, proteins were evaluated by the ECL detection system (Pierce, Rockford, IL, USA).
Processed H9C2 and MCM cells were fixed utilizing cold methanol (Sigma–Aldrich), followed by incubation with primary antibodies against α-actin (Abcam) and secondary antibodies conjugated to Alexa Fluor 488. Images were at last taken via fluorescence microscope (Olympus, Tokyo, Japan). After randomly examining 50 cells in 3 independent experiments, we obtained the average value for analyses. Image-Pro Plus 6.0 software was used to evaluate the surface area.
The subcellular localization of Gm43843 was examined through the FISH kit (Roche, Mannheim, Germany). Cells were cultured with a hybridization solution containing a specific Gm43843 probe (Sigma–Aldrich). The nuclei were dyed in DAPI (Sigma–Aldrich) for 10 min. At last, cells were captured via a fluorescence microscope to record images of fluorescence.
Nuclear/cytoplasmic fractionation PARIS Kit (Thermo Fisher Scientific) was used for collecting nuclear and cytoplasmic fractions of cells. The qRT-PCR was performed to determine the relative expression of Gm43843, GAPDH (cytoplasmic reference), and U6 (nuclear reference).
The wild-type and mutated Gm43843 or Cacna1c fragments covering miR-153-3p binding sites were subcloned into pmirGLO dual-luciferase vector (Promega, Madison, WI). The acquired Gm43843-WT/Mut and Cacna1c-WT/Mut reporter vectors were cotransfected into cells with miR-153-3p mimics or NC mimics for 48 h. Luciferase activity was measured with a dual-luciferase reporter assay system (Promega).
The wild-type and mutated miR-153-3p fragment covering Gm43843 or Cacna1c binding sites were labeled separately with biotin into Bio-miR-153-3p-WT/Mut probes. Biotinylated RNA was incubated with cell lysates and magnetic beads, and the RNAs in the complexes pulled down were purified and detected by qRT-PCR.
StarBase website (https://starbase.sysu.edu.cn/) was employed for projecting candidate miRNAs targeted by Gm43843 with no specific condition. This database was also applied for screening potential mRNA likely binding with miR-153-3p in the subset of microT. The potential binding sequences of Gm43843 and Cacna1c covering miR-153-3p binding sites were obtained from starBase as well.
All assays were run thrice. Values were shown as mean ± SD. GraphPad Prism 7 software (GraphPad Software, La Jolla, CA, USA) was utilized for statistical analysis with Student's t-test for two groups or one-way/two-way ANOVA for three or more groups with one or two variables. Tukey and Dunnett's approaches were applied as post hoc tests. The level of significance was specified as p < 0.05.
Reportedly, lncRNAs play vital parts in the progression of cardiac hypertrophy [10, 19, 20]. Herein, we discussed the role of lncRNA Gm43843 in cardiac hypertrophy. Above all, it was noticed that Gm43843 expression increased upon the elevation of Ang II concentration in mouse myocardial cells (H9C2 and MCM) (Figure 1(a)). Thence, we predicted that Gm43843 played its regulatory function in H9C2 and MCM cells. Furthermore, inhibition efficiency of Gm43843 was detected via qRT-PCR assay (Figure 1(b)). As presented in Figures 1(c)-1(d), Ang II (1 mmol/L) significantly increased the mRNA and protein levels of the biomarkers of cardiac hypertrophy (ANF, BNP, and β-MHC) in H9C2 and MCM cells. When Gm43843 was inhibited, their expressions were decreased. These findings indicated that Ang II (1 mmol/L) induced H9C2 and MCM cell hypertrophy. However, Gm43843 inhibition decreased the hypertrophy symptom. Meanwhile, the IF assay assessed the cell surface area (Figure 1(e)). Results found that cell surface area expanded after Ang II (1 mmol/L) treatment, but reduced again by silenced Gm43843. In conclusion, inhibited Gm43843 relieved cardiac hypertrophy.
It has been scrutinized that cytoplasmic lncRNAs can function as ceRNAs to regulate downstream RNA expression [21]. In our study, we firstly investigated the location of Gm43843 in H9C2 and MCM cells (Figures 2(a)-2(b)). We could see from the result that Gm43843 was mainly located in the cytoplasm in H9C2 and MCM cells. Thence, we further searched the miRNAs that were predicted to bind to Gm43843 via the starBase website and their expressions were assessed in H9C2 and MCM cells treated with Ang II (1 mmol/L) (Figure 2(c)). We could see that mmu-miR-153-3p was down-regulated in H9C2 and MCM cells upon Ang II (1 mmol/L) treatment. Meanwhile, the binding site of Gm43843 and mmu-miR-153-3p was manifested based on starBase prediction (Figure 2(d)). After the high miR-153-3p overexpression efficiency was verified (Figure 2(e)), we confirmed the binding relationship between Gm43843 and miR-153-3p as the luciferase activity of Gm43843-WT was weakened due to miR-153-3p augmentation (Figure 2(f)). RNA pull-down assay further supported the above finding, since Bio-miR-153-3p-WT probes largely pulled down Gm43843 but Bio-miR-153-3p could not (Figure 2(g)). To sum up, miR-153-3p is directly targeted by Gm43843 in H9C2 and MCM cells.
In this part, we further explored the target gene of miR-153-3p to complete the ceRNA network. According to the starBase website, we found 700 mRNAs in the subset of microT. All of these mRNAs were implemented into qRT-PCR assay to detect the most suitable mRNA whose expression could be affected by inhibited Gm43843 and overexpressed miR-153-3p in H9C2 cells with or without Ang II treatment (Figure 3(a)). Xkr4 and Cacna1c were found. Meanwhile, the expression of Xkr4 and Cacna1c was investigated with different concentrations (0.5 and 1 mmol/L) of Ang II in H9C2 and MCM cells via qRT-PCR assay (Figure 3(b)). We detected that only Cacna1c expression increased with the elevation of Ang II concentration. Thence, Cacna1c was selected as the target. We presented the binding site of Cacna1c and mmu-miR-153-3p (Figure 3(c)). The binding affinity between Cacna1c and mmu-miR-153-3p was corroborated, as the wild type of Cacna1c luciferase activity was observed to be reduced on account of miR-153-3p up-regulation (Figure 3(d)). Cacna1c was also substantially pulled down by the wild type of Bio-miR-153-3p in H9C2 and MCM cells (Figure 3(e)). In a word, Cacna1c could bind to miR-153-3p in H9C2 and MCM cells.
Previous studies have claimed the function of Cacna1c in the progression of cardiomyocyte hypertrophy [22–24]. In our study, we further investigated the function of Cacna1c in H9C2 and MCM cells. We first knocked down Cacna1c in H9C2 and MCM cells (Figure 4(a)). Then, we found that ANF, BNP, and β-MHC expressions were both inhibited by silenced Cacna1c (Figures 4(b)-4(c)). Meanwhile, the IF assay delineated that cell surface area was reduced when Cacna1c was inhibited in H9C2 and MCM cells (Figure 4(d)). To conclude, Cacna1c inhibition could relieve the hypertrophy of H9C2 and MCM cells.
We further studied the function of the Gm43843/miR-153-3p/Cacna1c axis in cardiac hypertrophy. We found that Cacna1c expression was significantly decreased with the silencing of Gm43843, but recovered with the inhibition of miR-153-3p (Figure 5(a)). After that, we overexpressed Cacna1c and the overexpression efficiency was proved to be high (Figure 5(b)). It was detected that RNA expressions and protein levels of ANF, BNP, and β-MHC were inhibited by Gm43843 depletion, but then, they were rescued by miR-153-3p inhibition or Cacna1c overexpression (Figures 5(c)-5(d)). Meanwhile, the IF assay found that lessened cell surface area due to knockdown of Gm43843 was increased again after miR-153-3p inhibition or Cacna1c augment (Figure 5(e)). In conclusion, Gm43843 could modulate the hypertrophy of H9C2 and MCM cells through miR-153-3p/Cacna1c.
Accumulating evidence has proved that lncRNA can play regulating roles in cardiac hypertrophy. For example, Plscr4 increment can reduce Ang II-induced cardiomyocyte hypertrophy by regulating the expression of miR-214 and Mfn2 [25]. LncRNA CASC15 upregulated in cardiomyocytes treated with Ang II can increase the cell surface area of cardiomyocytes in cardiac hypertrophy by modulating the miR-432-5p/TLR4 pathway [26]. Meanwhile, H19 inhibition can activate cardiomyocyte hypertrophy, and H19 can regulate miR-675 targeting CaMKIIδ in cardiac hypertrophy [27]. In our study, we searched the role of Gm43843 in mouse myocardial cells treated with Ang II for inducing hypertrophy. Gm43843 expression was found up-regulated with the Ang II concentration increasing. Furthermore, Gm43843 inhibition was observed to relieve the cardiac hypertrophy of mouse myocardial cells. Meanwhile, the IF assay reassured that silenced Gm43843 played an inhibitory role in cardiac hypertrophy. Previous studies have reported the ceRNA character of lncRNAs, which indicated that lncRNAs can function as miRNA sponges to form a miRNA/mRNA pathway to modulate the progression of human disease [28–30]. For example, lncRNA HOXD-AS1 can sponge to miR-130a-3p activating the expression of SOX4 to enhance the progression of liver cancer [31]. LncRNA TDRG1 can modulate cervical cancer cell growth, migration, and invasion via the miR-326/MAPK1 axis [32]. In our study, we firstly located Gm43843 in the cytoplasm in mouse myocardial cells. Mmu-miR-153-3p was corroborated to bind to Gm43843. Furthermore, Cacna1c was validated to bind to mmu-miR-153-3p. Increasing molecular genetic testing has suggested that Cacna1c-linked disorders account for pathogenic variants and clinical findings. Cacna1c is related to calcium channel function and individuals with a pathogenic variant of this gene have a risk for cardiovascular disease [33]. Specifically, previous studies have unveiled that Cacna1c was involved in the progression of cardiomyocyte hypertrophy [22–24]. CACNA1C expression could be inhibited by miR-135b in cardiomyocytes to relieve the symptom of pathological cardiac hypertrophy [23]. In our study, we found the expression of Cacna1c was up-regulated with Ang II inducement. It was worth noting that Cacna1c inhibition decreased the expression and protein level of ANF, BNP, and β-MHC, as well as reduced the cell surface area of mouse cardiomyocytes. Last but not least, it was verified that miR-153-3p inhibition or Cacna1c augmentation was able to abrogate the suppressive impact of Gm43843 deficiency on the hypertrophy of mouse cardiomyocytes. To conclude, Gm43843 promotes cardiac hypertrophy via miR-153-3p/Cacna1c axis. Although human cells and clinical samples need to be involved in the future study for further confirmation of the validity of the axis, our study can still provide a novel perspective for a more in-depth understanding of the molecular mechanism in cardiac hypertrophy. | true | true | true |
||
PMC9576420 | Yuping Zhu,Xuanying Wang,Linfeng Zheng,Dechuan Li,Zhuo Liu,Lisong Teng | The lncRNA NEAT1 Inhibits miRNA-216b and Promotes Colorectal Cancer Progression by Indirectly Activating YY1 | 10-10-2022 | Background Nuclear Paraspeckle Assembly Transcript 1 (NEAT1) is commonly considered an oncogene in various cancers. The long noncoding RNA NEAT1 has been reported to be overexpressed in colorectal cancer (CRC). However, the exact role of NEAT1 in CRC remains unknown. Our research aimed to explore the function of NEAT1 in the tumorigenesis and the development of CRC. Methods Real-time quantitative PCR (qRT-PCR) was used to detect the NEAT1, miR-216b, and YIN-YANG-1 (YY1) mRNA levels in CRC tissues and cells, then immunohistochemistry (IHC) was used to detect the expression of YY1 in CRC tissues. Luciferase reporter, qPCR, western blot, and DNA pulldown assays were conducted to study the relationships between NEAT1, miR-216b, and YY1. Flow cytometry analysis was performed for cell cycle and apoptosis analyses, and a colony formation assay was performed to test cell proliferation. Transwell assays were performed to detect cell invasion and migration. Results The NEAT1 expression was significantly upregulated in CRC tissues compared with its expression in normal tissues, and downregulation of NEAT1 suppressed the proliferation, migration, and invasion of CRC cells. Moreover, we found NEAT1 decreased the miR-216b level directly, and the suppression of miR-216b could inhibit the function of downstream YY1. However, overexpression of YY1 accelerated CRC cell proliferation, migration, and invasion. Conclusion Our results indicated that NEAT1 acted as an oncogene in CRC and promoted the progression of CRC by directly sponging miR-216 b expression to activate the expression of YY1. The NEAT1/miR-216b/YY1 axis may be a novel therapeutic target for CRC. | The lncRNA NEAT1 Inhibits miRNA-216b and Promotes Colorectal Cancer Progression by Indirectly Activating YY1
Nuclear Paraspeckle Assembly Transcript 1 (NEAT1) is commonly considered an oncogene in various cancers. The long noncoding RNA NEAT1 has been reported to be overexpressed in colorectal cancer (CRC). However, the exact role of NEAT1 in CRC remains unknown. Our research aimed to explore the function of NEAT1 in the tumorigenesis and the development of CRC.
Real-time quantitative PCR (qRT-PCR) was used to detect the NEAT1, miR-216b, and YIN-YANG-1 (YY1) mRNA levels in CRC tissues and cells, then immunohistochemistry (IHC) was used to detect the expression of YY1 in CRC tissues. Luciferase reporter, qPCR, western blot, and DNA pulldown assays were conducted to study the relationships between NEAT1, miR-216b, and YY1. Flow cytometry analysis was performed for cell cycle and apoptosis analyses, and a colony formation assay was performed to test cell proliferation. Transwell assays were performed to detect cell invasion and migration.
The NEAT1 expression was significantly upregulated in CRC tissues compared with its expression in normal tissues, and downregulation of NEAT1 suppressed the proliferation, migration, and invasion of CRC cells. Moreover, we found NEAT1 decreased the miR-216b level directly, and the suppression of miR-216b could inhibit the function of downstream YY1. However, overexpression of YY1 accelerated CRC cell proliferation, migration, and invasion.
Our results indicated that NEAT1 acted as an oncogene in CRC and promoted the progression of CRC by directly sponging miR-216 b expression to activate the expression of YY1. The NEAT1/miR-216b/YY1 axis may be a novel therapeutic target for CRC.
Colorectal cancer (CRC), including colon cancer and rectal cancer, is one of the most common malignancies in the world, which has threatened the global health [1]. Although the existing methods of diagnosis and treatment continue to improve, it also exhibits the low rate of 5-year overall survival and poor prognosis [2]. Due to its complex process of pathogenesis involving many factors, the clear mechanism of CRC still retain unknown, which motivate the clinical treatment to explore the CRC progression. Increasing attention has focused on the effect of long noncoding RNAs (LncRNAs), a subfamily of noncoding RNAs, in regulating the CRC progress [3]. Emerging evidence has shown that lncRNAs are important in the occurrence and progression of cancers by regulating cell cycle progression, apoptosis, and metastasis [4]. PANDAR, HOTAIR, HULC, H19, MALAT1, XIST, and NEAT1 have been considered as the oncogenes in various cancers, which can act as a “sponge” and compete for the binding of miRNAs of other genes [5–10]. LncRNA NEAT1 (NEAT1), encoding two variants of Neat1_v1 (3.7kb) and Neat1_v2 (23kb), is transcribed from a gene locus called multiple endocrine tumor type 1 in familial tumor syndrome on chromosome 11, which abnormally expressed in many malignant tumors, including CRC [11]. Nevertheless, the underlying molecular mechanism of NEAT1 needs further elucidation. Moreover, NEAT1 could sponge microRNAs such as miR-150-5p, miR-193a, miR-205-5p, and miR-34a to regulate CRC progression, leading to downstream signal transduction to induce chemotherapy resistance or metastasis [12–15]. In our research, we detected the expression of NEAT1 in human CRC tissues and cells. Subsequently, functional experiments were prepared to verify the carcinogenic effect of NEAT1. In mechanism, the regulatory relationship between NEAT1 and micRNA-216b was further studied which has not reported previously. Notably, we provided evidence that NEAT1 directly sponged the expression of miR-216b, which consequently removed functional regulation of the downstream YY1. This study provides new insights into clinical treatment and further intervention targets for CRC.
All 57 paired CRC tissues and normal tissues stored in biobank were obtained from those diagnosed CRC patients who underwent no preoperative therapy prior to surgical resection at Zhejiang Cancer Hospital (Hangzhou, China) from 2013 to 2017. Tissue collection and manipulation were authorized by the Ethics Committee of Zhejiang Cancer Hospital (ethics number: IRB-2020-26). Informed consents were collected from all subjects. The specimens were frozen in liquid nitrogen after surgery and kept at −80°C. The detailed clinical and pathological data from these 57 CRC specimens are shown in Table 1. The overall survival time and time from diagnosis to death were recorded to detect the prognostic status. Conventional receiver operating characteristic (ROC) curve analysis was used to confirm the cut-off value of NEAT1 expression. Furthermore, Youden index, which consisted of specificity and sensitivity, was prepared to help calculate the cut-off of the expression level (specificity%+sensitivity%−100%), the largest one of which was the cut-off value.
Total RNA was isolated from the paired tissues using the Trizol reagent (number: 15596–018, Invitrogen, Carlsbad, USA) according to the standard protocol in the kit. After reverse transcription of total RNA by using a cDNA reverse transcriptase kit (Takara, code number: 6215A, purchased from ElifeBio, Hangzhou, China), we obtained the cDNA. The expression levels of LncRNA NEAT1 and miR-216b were measured using Applied Biosystems 7500 Real Time PCR System with a Green One-Step qRT-PCR SuperMix Kit (Transgen, code number: AQ211-01, Beijing, China). The reaction conditions were as follows: predenaturation at 95°C for 5 min, denaturation at 95°C for 10 s, and annealing at 60°C for 30 s for 35 cycles. The results were normalized to the expression levels of GAPDH or U6 snRNA using the 2-ΔΔCT method for quantification. Primers were devised and produced by Sangon Biotech (Shanghai, China), and the sequences of PCR primers are shown in Table 2.
The details of western blot assays were performed as described in our previous published research [10]. Fresh tumor tissues and normal tissues treated with liquid nitrogen were crushed and then lysed in RIPA lysis buffer supplemented with proteinase inhibitor cocktail (Roche, number: 5892970001). In brief, lysed proteins were isolated by SDS-PAGE, transferred to PVDF membranes, and incubated with the following primary antibodies: anti-GAPDH antibody (1 : 3000 dilution, rabbit, Abcam, ab9485) andanti-YY1 antibody (1 : 500 dilution, rabbit, Abcam, ab109228). Following incubation with the appropriate HRP-conjugated secondary antibodies, the bands were visualized using Pierce™ ECL western blotting substrate (Thermo Scientific, number: 32106, USA). The signal intensity of the protein was further determined by ImageJ software (Madison, WI, USA).
The detail of the assay referred to the previous study [16]. After cut into 4-µm sections all the tissues were deparaffinized and treated with EDTA (pH 9.0) to antigen retrieval in a microwave for 20 min. Then, Autostainer Link 48 machine (Dako, Denmark A/S, Denmark) was performed for staining. Subsequently, primary anti-YY1 antibody (1 : 100, Abcam, ab109228) was added to the sections as well as their coordinate secondary antibody, while PBS buffer was used as a blank control instead of the antibody. EnVision Flex Kit (K802321–2CN) was used as the second antibody (Dako, Denmark A/S, Denmark). Percentage of positively stained cells and staining scores were used to assess the IHC results and the detail is shown in Table S1.
All cell lines (LoVo, Caco-2, SW620, HT29, HCT-116, and the normal colon cell line NCM460) were purchased from ATCC and cultured in RPMI 1640 medium (RPMI 1640; Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (FBS; Gibco, Grand Island, NY, USA) and 1% penicillin/streptomycin (ElifeBio, Hangzhou, China) with 5% CO2 at 37°C. HCT-116 and SW620 cells (1 × 105 cells) were seeded in 6-well plates in RPMI 1640 medium for 12 h. Then, the cells were treated with NEAT1-or YY1-specific shRNA as well as the negative control with 0.2% Lipofectamine 3000 in RPMI 1640 medium. The plasmids YY1 and NEAT1 were purchased from ElifeBio (Hangzhou, China). The miR-216 b inhibitor, miR-216 b mimics, and mimic normal control were purchased from ElifeBio (Hangzhou, China). The detailed sequences are listed in Table 2. After 5 h, the transfection reagent was replaced with RPMI 1640 medium. Forty-eight hours later, the cells were collected for further experiments.
Cell proliferation or viability was determined by a CCK-8 assay (HY–K0301, MCE). HCT-116 and SW620 cells (1 × 105 cells/well) transfected with or without the plasmid were seeded in 96-well plates. In brief, on the day of measuring the growth rate of the treated cells, 100 µl of spent medium was replaced with an equal volume of fresh medium containing 10% CCK-8 reagent. Cells were incubated at 37°C for 1 h, and the absorbance was subsequently detected at 450 nm with a microplate reader.
HCT116 and SW6200 cells, transfected with plasmid or not, were seeded in 6-well plates in RPMI1640 medium at 37°C, half of the medium was replenished on day 5 and the medium was discarded on day 14. Cells were washed with phosphate buffer saline (PBS), fixed with methanol for 10 min, and stained with crystal violet for 5 minutes, and then observed under a microscope. Cells in five fields were selected for counting.
Cell invasion and migration were tested by a transwell chamber with Matrigel (8-μm pore size; BD Biosciences, USA) according to the manufacturer's protocol. The detail referred to the previous study [12].
Flow cytometry assay for cell cycle and apoptosis was prepared referring to a previous article [17].
The DNA fragment containing the full-length NEAT1 sequence or negative control sequence was PCR amplified by a T7-containing primer and then bound to GV394 (Invitrogen, Shanghai, China). The obtained plasmid DNAs were digested to a linear form by the restriction enzyme XhoI. Biotin-labeled RNAs were conversely transcribed, and then qRT-PCR was used to analyze target the RNA expression according to the method described previously [18].
All experimental data were analyzed by GraphPad Prism 8.0, and the results are shown as the mean ± SD (standard deviation). Chi-square tests were performed to detect the correlation between the NEAT1 expression and clinicopathological factors. Student's t-test was performed to compare the differences between the two groups. Comparisons among multiple groups were made by one-way ANOVA. Survival analysis was performed via the Kaplan–Meier method accompanied by Cox regression analysis for univariate and multivariate analyses. P < 0.05 was considered significant.
According to the public, database “UALCAN” (http://ualcan.path.uab.edu) from the Cancer Genome Atlas (TCGA) in cancers, we found NEAT1 overexpressed in tumor tissues (Figure 1(a), p < 0.01) as well as upregulated higher in stage IV and stage III than that in stage I and stage II (Figure 1(b), p < 0.01), which was also identified in “Gepia” database (http://gepia.cancer-pku.cn/) (Figure 1(c), p < 0.01). NEAT1 showed much higher expression in mucinous adenocarcinoma than adenocarcinoma(Figure 1(d), p < 0.01). Higher level expression of NEAT1 also indicated the severe lymph node metastasis stage (Figure 1(e), p < 0.01) as well as poor disease-specific survival (Figure 1(f), p=0.036) and overall survival (Figure 1(g), p=0.018) in cohort GSE17536. Referred to the results from the public database, we have tested the NEAT1 expression in CRC tissues and cell lines that its actually overexpressed in CRC tissues (Figure 2(a), p < 0.01) and cells (Figure 2(b), p < 0.01) compared with normal. It was similar with the database result that NEAT1 showed the high expression in poorly differentiated tissues not the well ones (Figures 2(c) and 2(k), p < 0.01). Moreover, we found NEAT1 overexpressed in CRC patients with the stage III/IV compared with the stage I/II (Figure 2(d), P < 0.05). Furthermore, CRC patients with metastasis especially nodal metastasis and liver metastasis exhibited a higher expression level of NEAT1 than nonmetastasis ones (Figures 2(e)–2(g), p < 0.01). In order to identify the association of NEAT1 in prognosis, we explored the cut-off value of NEAT1 in our CRC specimen, the result of which showed 2.888 was the threshold for NEAT1 (compared with GAPDH). Exactly mRNA levels of NEAT1 >2.888 was considered as “high” and those ≤2.888 as “low” (Figure 2(j), AUC = 0.9070). In our result, higher expression level of NEAT1 displayed the lower overall survival time (Figure 2(h), p=0.0012) and progression-free survival time (Figure 2(i), p=0.0064). We also found high expression of NEAT1 mRNA significantly correlated with distant metastasis (p < 0.01), lymph node metastasis (P=0.0001), and tumor grade (P=0.011) (Table 1).
After transfection, the expression of NEAT1 significantly decreased in the siRNA NEAT1 group compared with the siRNA negative control group (Figure 3(a), P < 0.05). The CCK-8 assay showed that at 48, 72, and 96 hours, siRNA NEAT1 significantly repressed cell viability compared with the negative control group in SW620 (Figure 3(b), P < 0.05). We also detected NEAT1 knockdown inhibited cell proliferation by the colony formation assay (Figure 3(f)). Suppression of NEAT1 restrain much more cells than the negative control group in G0/G1 which inhibited cell proliferation (Figure 3(c), P < 0.05). Loss function of NEAT1 elevated the apoptotic rate in SW620 and HCT-116 by the flow cytometry apoptosis assay (Figure 3(d), P < 0.05). Afterwards, the transwell assay further indicated the silenced of NEAT1 decreased invasive and migratory cell numbers, which revealed NEAT1 led to CRC invasion and migration (Figure 3(e), P < 0.05).
We identified miR-216 b as a most potential candidate by bioinformatics database “Diana Tools” (Diana, http:// diana.imis.athena-innovation.gr/DianaTools) (Figure 4(a)). Since then, we constructed plasmids of NEAT1 to detect whether the decreased expression of NEAT1 could lead to the increased expression of miR-216 b as well as designed the luciferase reporter containing exact or mutant miR-216 b binding sites to determine the binding effect between NEAT1 and miR-216 b. We found NEAT1 knockdown resulted in the increased expression of miR-216 b by qPCR (Figure 4(b)). Furthermore, we found the decreased miR-216 b expression in the CRC tissue (Figure 4(c), P < 0.05). The luciferase reporter showed miR-216 b mimic significantly reduced luciferase activity of the wild-type NEAT1 plasmid, however, miR-216b inhibitor significantly increased it in HCT-116 ell line (Figure 4(d), P < 0.05). Moreover, miR-216b inhibitor significantly upregulated the expression of NEAT1(Figure 4(e), P < 0.05). In contrast, knockdown of NEAT1 also induced the upregulation of miR-216b (Figure 4(f), P < 0.05). Subsequently, the biotin-labeled pulldown system was applied to further confirm whether NEAT1 could pulldown miR-216 b. We observed miR-216 b in CRC cells pulled down by biotinylated NEAT1 which indicated miR-216b could directly bind to NEAT1 at the microRNA recognition site (Figure 4(g), P < 0.05). The transwell assay was prepared to explore the function of NEAT1 and miR-216b in migration and invasion, the result from which identified SW620 and HCT-116 transfected with NEAT1-silenced and cotransfected with miR-216b mimic significantly repressed the cell invasion and migration but failed when cotransfected with the miR-206 inhibitor (Figure 4(h), p < 0.01). The SW620 and HCT-116 cell lines with NEAT1 silence showed poor ability in proliferation and it achieved better when cotransfected with miR-216b mimic. However, when cell lines with NEAT1-silenced and cotransfected with miR-216b inhibitor displayed better ability in proliferation than that with NEAT1-silenced and cotransfected with miR-216b mimic(Figure 4(i), p < 0.01).
To explore candidate target of miR-216b, bioinformatics online tools (miRDB, http://mirdb.org/) were prepared to predict the suitable one that finding miR-216b possessed the matched binding site with YY1 (Figure 5(a)), which overexpressed in the CRC tissues not normal tissues (Figure 5(b), p < 0.01). In addition, the IHC staining indicated the strong staining of YY1 in the CRC tissues and liver metastatic tissues compared with the almost no staining in the CRC situ tissues especially in the liver metastasis tissues (Figure 5(h)), which suggested the oncogenic role of YY1 in CRC. It also showed the decreased mRNA and protein expression of YY1 in SW620 and HCT-116 with NEAT1-silence or miR-216b mimic (Figures5(c), 5(d), and 5(e), p < 0.01), nevertheless, it decreased much better when SW620 and HCT-116 with NEAT1-silence cotransfected with miR-216b mimic(Figure 5(f), p < 0.01). The cell migration and invasion ability in SW620 or HCT-116 was inhibited significantly when transfected with YY1 plasmid (Figure 5(g), p < 0.01). Taken together, NEAT1 sponged miR-216b to activate YY1 to accelerate the ability of cell viability, apoptosis, and invasion on colorectal cancer cells.
NEAT1 participates in many critical biological processes and acts as a potential predictor or target for prognosis in CRC, lung cancer, gastric cancer, and breast cancer [9, 12, 19, 20]. However, further research studies are required to elucidate how NEAT1 function. Herein, we discovered high expression of NEAT1 in the CRC tissue, and associated with histological differentiation, overall survival, distant metastasis, and nodal metastasis, acting as an independent prognostic factor for overall survival in patients with CRC, indicating the poor prognosis. In the present study, we also observed that NEAT1 knockdown or miR-216b overexpression remarkably attenuated the ability of cell viability, apoptosis, and invasion on colorectal cancer cells. Mechanismly, our results disclosed that NEAT1 sponged the miR-216b to facilitate the function of the downstream regulators YY1, thus to promote CRC proliferation, invasion, and migration. Accumulating evidence has indicated that LncRNA could regulate the expression of microRNA as ceRNA [3, 4, 21–23]. NEAT1 displayed multiple function in regulating microRNA that it not only sponged miR-133b to promote migration and invasion of breast cancer cells [24], but also inactivated miR-101 to play potential oncogene in breast cancer [25]. Moreover, in lung cancer, NEAT1 competed against let-7a to contribute to nonsmall-cell lung cancer proliferation and metastasis [26], and another research in lung cancer also revealed downregulation of NEAT1 led to cell invasion in NSCLC via sponging miR-153-3p [27]. In CRC, it was illustrated that the NEAT1 expression level was an independent prognostic factor for disease-free survival and overall survival which mechanismly promoted colorectal cancer progression by competitively binding miR-34a [12] resulting in activation of Wnt/beta-catenin signaling pathway. What's more, NEAT1 also could facilitate the sensitivity of 5-FU in CRC cells via miR-150-5p [14]. Although a little study displayed the anti-cancer role of NEAT1 in CRC, a flood of investigations have still been verified its oncogenic role especially acting as ceRNA to regulate microRNAs [20, 27]. However, Xiong's team revealed that the NEAT1 expression in the CRC tissues were not significantly different compared with the normal tissues [28]. Therefore, the function of NEAT1 in CRC needs to be further explored. Like previous studies, our research also identified NEAT1 acting as oncogene in CRC and considered as an independent prognostic factor for disease-free survival and overall survival. With the help of online bioinformatics tools “DIANA” and “MIRDB”, we have explored microRNA-216b that NEAT1 may sponge to discover the deep mechanism. MicroRNA-216 b from our result was first reported molecular which can be regulated by NEAT1. MicroRNA-216 b was considered as the tumor suppressor in lung cancer, gastric cancer, and pancreatic cancer which could regulate cell proliferation, apoptosis, and invasion [29, 30]. Here, we found that miR-216b, acting as a connecting carrier, targets both NEAT1 and the 3′-UTR of YY1 by luciferase reporter assays. YY1 showed the oncogenic role in CRC, especially playing the epigenetic regulative role on cancer stem cell transcription factors to accelerate tumor metastasis [31, 32]. This constitutes the identification of a competitive endogenous RNA network in colorectal cancer. Although we verified NEAT1 sponged the miR-216b to facilitate the function of the downstream regulators YY1, our limitations should not be ignored in this study: our results were from experiments in vitro not in vivo, which made us to finish the further studies as well as the epigenetic regulation of YY1 in this axis.
Our results illustrate that NEAT1 functions as an oncogenic lncRNA to facilitate the carcinogenesis and progression of CRC by competitively sponging miR-216b to activate YY1. The present results elucidate a potential mechanism underlying the tumor-oncogenic role of NEAT1 in colorectal cancer and indicate that NEAT1 could serve as a useful marker and potential therapeutic target in CRC. | true | true | true |
||
PMC9576421 | Wanshun Liu,Qi Zhang,Kai Shen,Keran Li,Jie Chang,He Li,Ao Duan,Sheng Zhang,Yumin Huang | Long Noncoding RNA LINC00909 Induces Epithelial-Mesenchymal Transition and Contributes to Osteosarcoma Tumorigenesis and Metastasis | 10-10-2022 | Background Osteosarcoma (OS) is a malignant tumor that is highly metastatic with a high mortality rate. Although mounting evidence suggests that LINC00909 is strongly associated with the malignant progression of various tumors, the exact role of LINC00909 in OS remains unknown. Therefore, the current study was designed to investigate the relationship between LINC00909 and the malignant progression of OS. Methods LINC00909 expression was measured in OS cell lines and clinical specimens using RT-qPCR assays. The effects of LINC00909 on OS proliferation, invasion, and migration were calculated both in vitro and in vivo. Apart from that, bioinformatics analyses, FISH, RIP, and luciferase reporter assays were carried out to investigate the downstream target of LINC00909. Rescue experiments were also conducted to investigate the potential mechanisms of action of competitive endogenous RNAs (ceRNAs). Results In this study, we found that LINC00909 was highly expressed in OS cell lines and clinical specimens. In vivo and in vitro experiments demonstrated that LINC00909 induces epithelial-to-mesenchymal transition (EMT) and contributes to OS tumorigenesis and metastasis. FISH, RIP, and luciferase assays indicated that miR-875-5p is a direct target of LINC00909. Moreover, HOXD9 was validated as the downstream target of miR-875-5p in a luciferase reporter assay and western blotting experiments. Rescue experiments revealed that HOXD9 reversed the effect of LV-sh-LINC00909 on OS cells by positively regulating the PI3K/AKT/mTOR signaling pathway. Conclusion Collectively, LINC00909 induces EMT and contributes to OS tumorigenesis and metastasis through the PI3K/AKT/mTOR pathway by binding to miR-875-5p to upregulate HOXD9 expression. Targeting the LINC00909/miR-875-5p/HOXD9 axis may have potential in treating OS. | Long Noncoding RNA LINC00909 Induces Epithelial-Mesenchymal Transition and Contributes to Osteosarcoma Tumorigenesis and Metastasis
Osteosarcoma (OS) is a malignant tumor that is highly metastatic with a high mortality rate. Although mounting evidence suggests that LINC00909 is strongly associated with the malignant progression of various tumors, the exact role of LINC00909 in OS remains unknown. Therefore, the current study was designed to investigate the relationship between LINC00909 and the malignant progression of OS.
LINC00909 expression was measured in OS cell lines and clinical specimens using RT-qPCR assays. The effects of LINC00909 on OS proliferation, invasion, and migration were calculated both in vitro and in vivo. Apart from that, bioinformatics analyses, FISH, RIP, and luciferase reporter assays were carried out to investigate the downstream target of LINC00909. Rescue experiments were also conducted to investigate the potential mechanisms of action of competitive endogenous RNAs (ceRNAs).
In this study, we found that LINC00909 was highly expressed in OS cell lines and clinical specimens. In vivo and in vitro experiments demonstrated that LINC00909 induces epithelial-to-mesenchymal transition (EMT) and contributes to OS tumorigenesis and metastasis. FISH, RIP, and luciferase assays indicated that miR-875-5p is a direct target of LINC00909. Moreover, HOXD9 was validated as the downstream target of miR-875-5p in a luciferase reporter assay and western blotting experiments. Rescue experiments revealed that HOXD9 reversed the effect of LV-sh-LINC00909 on OS cells by positively regulating the PI3K/AKT/mTOR signaling pathway.
Collectively, LINC00909 induces EMT and contributes to OS tumorigenesis and metastasis through the PI3K/AKT/mTOR pathway by binding to miR-875-5p to upregulate HOXD9 expression. Targeting the LINC00909/miR-875-5p/HOXD9 axis may have potential in treating OS.
Osteosarcoma (OS) is a common primary bone tumor mostly seen in children and adolescents [1]. The treatment of primary tumors in patients with OS is mainly chemotherapy and surgery [2]. Despite great improvements in neoadjuvant chemotherapy and surgery, the possibility of long-term survival for patients with metastases is still low [3]. Therefore, it is critical to elucidate the mechanism underlying OS metastasis. Long noncoding RNAs (lncRNA) are a class of RNA molecules greater than 200 nucleotides that are involved in various biological processes such as metastasis or tumorigenesis. lncRNAs affect these processes by modulating epigenetic, transcriptional, and posttranscriptional gene expression [4, 5]. Notably, lncRNAs may play a role as competing endogenous RNAs (ceRNAs), which bind miRNAs that target mRNA expression [6]. Moreover, lncRNAs modulate numerous OS biological processes, like epithelial-mesenchymal transition (EMT), cell growth, and apoptosis [7–9]. LINC00909 is upregulated in many tumor types and regulates tumor malignant progression through the ceRNA mechanism [10, 11]. Nevertheless, LINC00909's role in OS is still unclear. In EMT, epithelial cells lose their polarity and tight and adhesion junctions between cells and gain infiltration and migration capabilities to become cells with mesenchymal characteristics [12, 13]. During the malignant evolution of tumors, EMT enables tumor cells to invade and metastasize [14, 15]. Therefore, inhibiting the progression of EMT may be a potentially effective treatment for OS. MicroRNAs (miRNAs) are noncoding RNAs that contain 22–28 nt, which promote target mRNA decomposition and suppress its translation through binding to its 3′-untranslated regions (UTR) [16]. It has been discovered that miRNAs are linked to tumor occurrence and progression [17]. Downregulated miR-875-5p expression can be detected in numerous types of tumors and associated with a favorable prognosis [18, 19]. However, the function of miR-875-5p in OS progression is still unknown. HOXD9, a transcription factor (TF), has an essential function in the HOX family, which encodes for transcription factors with crucial roles in development [20]. Studies have shown that HOXD9 is highly expressed in diverse cancers, which predicts a dismal prognostic outcome [21, 22]. In addition, HOXD9 expression is linked to tumor proliferation, invasion, and distant metastasis [23, 24]. However, the relationship between HOXD9 and osteosarcoma progression has not been reported. The present study analyze LINC00909's function in OS progression by experiments in vitro and in vivo and explore the related mechanisms. LINC00909 functions as a ceRNA to bind miR-875-5p, thereby upregulating HOXD9 expression, and contributes to OS tumorigenesis and metastasis through the PI3K/AKT/mTOR pathway. This study illustrates a new possible mechanism of OS development and provides a rationale for a novel anti-OS therapeutic strategy.
We collected 60 OS patients who received tumor biopsies prior to radiotherapy and chemotherapy in the Department of Orthopedics of Jiangsu Provincial People's Hospital between 2014 and 2020. Three pathologists confirmed the histological diagnosis of intraoperatively resected OS samples. Both tumor and matched noncarcinoma samples were frozen in liquid nitrogen. All patients provided informed consent for participation. This study gained approval from the Evaluation Committee and Ethics Committee of the First Affiliated Hospital of Nanjing Medical University. Clinicopathological features of patients are presented in Table 1.
OS cells were cultivated within DMEM (Gibco, CA, USA) that contained 1% penicillin/streptomycin (P/S) (Gibco) as well as 10% fetal bovine serum (FBS, Gibco, NY) at 37°C, while hFOB1.19 osteoblasts were cultivated in DMEM that contained 1% P/S and 10% FBS at 33.5°C. The above cell lines were subject to incubation in 5% CO2.
The LINC00909 overexpression lentiviral plasmid LV-LINC00909, lentiviral plasmid containing short hairpin RNA (shRNA) targeting LINC00909, plasmid containing shRNA targeting HOXD9, HOXD9 overexpression plasmid, and the corresponding negative control plasmids were purchased from Tsingke (Nanjing, China). Additional file 1: Table S1 provides the shRNA sequences. RT-qPCR was conducted to verify transfection efficiency. LV-sh-LINC00909#3 and sh-HOXD9#1 showed the greatest knockdown efficiency, which was therefore chosen as the optimal shRNA for later studies. The miR-875-5p inhibitor and miR-875-5p mimic, together with corresponding negative controls, were provided by RioBio (Guangzhou, China). Lipofectamine 3000 (Invitrogen, CA, USA) was used to transfect OS cells according to the manufacturer's instructions.
The lncRNA-sequencing data of corresponding clinical information OS were downloaded from the TARGET database; R package limma was used to analyze lncRNA-sequencing data. The downstream miRNA targets of LINC00909 were predicted using the DIANA tool and the lncRNASNP2 tool. Gene microarray data (GSE12865 and GSE14359) for OS were downloaded from the Gene Expression Omnibus database. The starBase tool was used to find the targets of miR-875-5p. The lncRNASNP2 tool was used to predict the targeted relationship between miR-875-5p and LINC00909. TargetScanHuman 7.2 was used to predict the targeted relationship between miR-875-5p and HOXD9. Gene Set Enrichment Analysis (GSEA) was performed to detect functions of LINC00909 and the downstream signaling pathways of HOXD9 in OS based on the TARGET database.
TRIzol (Invitrogen, United States) was used to isolate total tissue and cell RNA. A NanoDrop spectrophotometer was utilized to measure RNA content and purity. In addition, RNeasy/miRNeasy Mini Kits (Qiagen) were employed to extract miRNAs. We conducted RT-qPCR according to previous publications [25]. U6 and β-actin were the reference for normalizing miRNA and LINC00909/HOXD9 levels, respectively. Every experiment was carried out in triplicate. The 2−ΔΔCT method was used to determine relative gene levels. Sequences of all primers utilized are displayed in Additional file 1: Table S1.
The total proteins of tissues and OS cells were extracted by RIPA lysis buffer (YEASEN, China); then, the BCA protein detection kit (Thermo Fisher Scientific, USA) was adopted for measuring protein concentration. Samples were boiled, denatured, and separated using electrophoresis. Proteins were then transferred onto PVDF membranes. Rapid blocking solution was used to block membranes for a 30 min period. Primary antibodies were then used to incubate membranes at 4°C overnight. Afterwards, membranes were further incubated with secondary antibody for an additional 1 h at room temperature (RT). The membranes were then exposed to ECL reagent (Millipore, USA) with the Tanon 4200 automatic chemiluminescence imaging analysis system. Antibody information is shown in Additional file 2: Table S2.
To measure OS cell proliferation, 5-Ethynyl-2-deoxyuridine (EdU), colony formation, and Cell Counting Kit-8 (CCK-8) assays were performed as described previously [25].
Scratch and transwell assays were carried out to determine the impacts on OS cell invasion and migration according to previous experiments [25].
LINC00909 levels in OS, matched noncarcinoma tissues, and OS cells were detected using a fluorescence in situ hybridization (FISH) assay. Probes against LINC00909 and miR-875-5p were synthesized by Servicebio (Wuhan, China). FISH was performed according to the manufacturer's protocol. An upright fluorescence microscope (Nikon, Japan) was used for imaging.
The cytoplasmic and nuclear fractions were extracted using the PARIS™ Kit (Thermo Fisher, MA, USA). RT-qPCR was conducted to analyze the expression levels of LINC00909, 18S (cytoplasmic control transcript), and U6 (nuclear control transcript).
LINC00909's binding sites for miR-875-5p were predicted using the lncRNASNP2 tool (http://bioinfo.life.hust.edu.cn/lncRNASNP). HOXD9 binding sites for miR-875-5p were predicted using the TargetScan database (http://www.targetscan.org/vert_72/). Mutant (MUT) and wild-type (WT) LINC00909, named MUT-LINC00909-3′ UTR and WT-LINC00909-3′ UTR, respectively, together with MUT and WT HOXD9, named MUT-HOXD9-3′ UTR and WT-HOXD9-3′ UTR, respectively, were prepared using GenScript (Nanjing, China). MG63 and 143B cells were first transfected with miR-NC or miR-875-5p mimic, followed by cotransfection with MUT-LINC00909-3′ UTR, WT-LINC00909-3′ UTR, MUT-HOXD9-3′ UTR, and WT-HOXD9-3′ UTR for a 48 h period. We then utilized the double Luciferase Assay System (Promega, USA) to measure luciferase activity. Normalization was based on Renilla luciferase activity.
Magna RIP RNA-binding Protein IP Kit (Millipore, Billerica, MA) was used to perform the RIP assay. 143B and MG63 cells were lysed using RIP lysis wash buffer. After being centrifuged for half an hour, the supernatant was subjected to immunoprecipitation with anti-IgG or anti-Ago2-coated magnetic beads. RT-qPCR was used to detect RNA levels in the precipitates.
A 4% paraformaldehyde solution was used to fix nude mouse and human cancer samples; samples were then embedded in paraffin and sliced into 4 μm sections for immunohistochemistry (IHC). Antigen retrieval and blocking were performed, followed by incubation with anti-vimentin and anti-Ki-67 primary antibodies at 4°C overnight. Slides were then incubated for 1 h with secondary antibodies at ambient temperature. Sections were then treated with freshly prepared 3,3-diaminobenzidine solution. The staining intensity and positive tumor percentage were measured from five randomly selected fields of view.
Twenty 6-week-old female athymic BALB/c nude mice were divided into four groups for xenograft experiments: LV-LINC00909, LV-sh-LINC00909, and the corresponding NC groups, N = 5 for each. An OS cell suspension (200 μl) of 2 × 107 luciferase-expressing cells/ml was injected into the anterior right armpit of each animal. We measured tumor dimensions every four days and calculated volumes according to the following formula: volume = (width)2 × length/2. At 28 days postcell implantation, the fluorescence intensity of tumors was detected using the IVIS Imaging System (Caliper Life Sciences, USA) and tumor tissues were resected, weighed, and fixed for IHC assays. Additionally, the OS lung metastasis model was constructed in 6-week-old female athymic BALB/c nude mice. Mice were grouped similarly to above, and 100 μl of cell suspension containing 2 × 107/ml of OS cells transfected with a luciferase-expressing vector was injected into each mouse through the tail vein. Twenty-eight days postinjection, we measured the fluorescence intensity of lung metastases using the IVIS imaging system and lung tissue was removed for formalin fixation and subsequent hematoxylin and eosin.
Results were presented in a form of mean ± SD. All assays were carried out in triplicate. SPSS22.0 (SPSS Inc., Chicago, Illinois, USA) was employed for statistical analysis. The significance between two groups was compared using a Student t-test, whereas a one-way ANOVA was adopted for comparison across several groups. p < 0.05 was considered statistically significant.
To identify potential lncRNAs which participate in OS metastasis, we first explored the TARGET database and found that lncRNA LINC00909 was highly expressed in OS metastatic specimens compared with nonmetastatic samples (Figures 1(a) and 1(b)). Moreover, the results of GSEA indicated that LINC00909 may facilitate the EMT process in OS (Figure 1(c)). We therefore chose LINC00909 for further study. LINC00909 expression in OS cells and HFOB1.19 was detected through RT-qPCR. The expression level of LINC00909 was higher in OS cells compared with healthy hFOB 1.19 cells (Figure 1(d)). Additionally, LINC00909 levels were measured in 60 OS tissues and matched noncarcinoma tissues. LINC00909 expression levels were higher in OS samples compared with matched noncarcinoma samples (Figure 1(e)). At last, in comparison with the nonmetastatic group, LINC00909 levels increased within the metastatic group (Figure 1(f)). Moreover, we further verified that LINC00909 was highly expressed in OS samples compared with matched noncarcinoma samples by FISH (Figure 1(g)). We also performed nuclear mass separation and FISH assays to determine the subcellular localization of LINC00909 in OS cells; LINC00909 was found mainly in the cytoplasm (Figures 1(h) and 1(i)). IHC analysis of clinical samples revealed that Ki-67 and vimentin expression in the high-LINC00909 group was significantly higher relative to that in the low-LINC00909 group (Figure 1(j)). In addition, we examined the association of LINC00909 levels with clinicopathological characteristics in the 60 OS cases (Table 1), LINC00909 expression was in direct proportion to tumor size, metastasis, and TNM stage. Taken together, LINC00909 is highly expressed in OS tissues and cells and may be associated with patient prognosis.
We investigated the effect of overexpressing LINC00909 on proliferation, migration, invasion, and EMT in OS cells. LV-LINC00909 was transfected into MG63 and 143B cells. RT-qPCR was conducted to assess gene expression posttransfection (Figure 2(a)). The CCK-8 assay data suggested that the LV-LINC00909 group had significantly higher cell proliferation (Figure 2(b)). In addition, OS cells showed enhanced colony formation abilities in the LV-LINC00909 group (Figure 2(c)). Furthermore, based on the EdU assay, LINC00909 overexpression resulted in an increase in the mitotic cell proportion in the LV-LINC00909 group (Figure 2(d)). Scratch and transwell assays were performed to investigate the function of LINC00909 in OS cell migration and invasion. LV-LINC00909 overexpression increased the proportion of migrating cells and promoted cell invasion (Figures 2(e) and 2(f)). The LV-LINC00909 group showed an elevated migration rate in the scratch assay (Figure 2(g)). Subsequently, we examined the association between LINC00909 and EMT-associated proteins through WB. Vimentin and N-cadherin levels were upregulated after LINC00909 overexpression, while E-cadherin levels were attenuated (Figure 2(h)), suggesting that LINC00909 activates EMT to enhance tumor metastasis. Taken together, in vitro experiments demonstrate that LINC00909 stimulates OS cell growth, migration, invasion, and EMT.
We next investigated whether LINC00909 knockdown affected the proliferation, migration, invasion, and EMT of OS cells. 143B and MG63 cell lines were transfected with LV-sh-LINC00909 (Figure 3(a)). We chose LV-sh-LINC00909#3 for shRNA experiments, since it had the greatest knockdown activity. To investigate whether LINC00909 exerts a critical role in OS cell proliferation, we performed CCK-8, colon formation, and EdU assays after LINC00909 downregulation. Cell proliferation, colon formation abilities, and the proportion of mitotic cells were all decreased after LINC00909 downregulation (Figures 3(b)–3(d)). We also investigated the effect of LINC00909 knockdown on the invasion and migration of OS cells using scratch and transwell assays. Migrating cell numbers decreased with LINC00909 knockdown (Figure 3(e)). The low expression level of LINC00909 also resulted in a reduced invasive capacity of OS cells (Figure 3(f)). Furthermore, the LV-sh-LINC00909 group had a decreased migration rate (Figure 3(g)). Vimentin and N-cadherin expression was also decreased after LINC00909 downregulation, whereas E-cadherin expression increased (Figure 3(h)). In summary, the in vitro data suggest that LINC00909 inhibition suppresses proliferation, migration, invasion, and EMT of OS cells.
We conducted a mouse model to further investigate the role of LINC00909 in OS tumorigenesis in vivo. We constructed a tumor model in nude mice using subcutaneous injection of fluorescein-expressing stably transfected OS cells. After four weeks, the LV-LINC00909 group had elevated tumor weight and volume relative to the LV-NC group, whereas the LV-sh-LINC00909 group had reduced tumor weight and volume relative to the LV-sh-NC group (Figures 4(a)–4(c)). The LV-LINC00909 group had a significantly increased tumor volume compared with the LV-NC group, whereas the LV-sh-LINC00909 group had a decreased tumor volume compared with the LV-sh-NC group, as shown in the in vivo imaging experiments (Figure 4(d)). As revealed by tissue IHC, vimentin and Ki-67 levels were elevated in the LV-LINC00909 group, while vimentin and Ki-67 levels were decreased in the LV-sh-LINC00909 group (Figures 4(e) and 4(f)). We also constructed an in vivo nude mouse model of OS lung metastasis for elucidating the effects of LINC00909 on OS metastasis. We assigned animals into four groups: LV-LINC00909, LV-sh-LINC00909, and the corresponding NC groups. The LV-LINC00909 group had remarkably enhanced lung metastases, whereas the LV-sh-LINC00909 group had decreased lung metastases. These observations were confirmed through in vivo imaging (Figure 4(g)). Lung metastatic lesions were verified through hematoxylin and eosin staining (Figure 4(h)). Altogether, our results show that LINC00909 promotes OS tumorigenesis and metastasis in vivo.
lincRNAs can bind various miRNAs to suppress miRNA expression and reduce their regulation of target mRNAs. By using the online databases, DIANA and lncRNASNP2, we predicted three potential target miRNAs with high binding scores (Figure 5(a)). Among the three candidate target miRNAs, we selected miR-875-5p for subsequent experiments because it participates in different cancers and has an important function in suppressing cancer development. miR-875-5p levels were measured in hFOB1.19 osteoblasts and five types of OS cells using RT-qPCR. And miR-875-5p levels within OS cells remarkably decreased compared with hFOB1.19 cells (Figure 5(b)). RT-qPCR analysis also demonstrated downregulation of miR-875-5p in OS tissues compared with matched noncarcinoma tissues (Figure 5(c)). In addition, miR-875-5p showed a negative correlation with LINC00909 in OS clinical samples (Figure 5(d)). Based on the endogenous miR-875-5p expression levels in OS cells, 143B and MG63 cell lines were transfected with miR-875-5p mimics or miR-875-5p inhibitor, respectively. RT-qPCR was later conducted to assess transfection efficiency (Additional file 3: Figure S1(a)). To investigate the function of miR-875-5p in the proliferation of OS cells, we performed CCK-8 and EdU assays after miR-875-5p upregulation. The results demonstrate that cell proliferation was inhibited after miR-875-5p upregulation (Additional file 3: Figures S1(b) and S1(c)). We also explored the effect of miR-875-5p on the migration and invasion of OS cells using transwell assays. The number of migrating cells decreased (Additional file 3: Figure S1(d)), and the invasive capacity of OS cells was reduced (Additional file 3: Figure S1(e)) when miR-875-5p was upregulated. Furthermore, we observed that LINC00909 overexpression downregulated miR-875-5p expression levels, and LV-sh-LINC00909 upregulated miR-875-5p expression levels in 143B and MG63 cells (Figure 5(e)). Moreover, the dual-luciferase reporter assay showed that miR-875-5p overexpression dramatically decreased WT LINC00909 luciferase activity but had little effect on MUT LINC00909 (Figure 5(f)). FISH analysis indicated that LINC00909 was bound to miR-875-5p (Figure 5(g)). RIP assay showed that miR-875-5p and LINC00909 were enriched in Ago2 immunoprecipitants compared with the control IgG immunoprecipitant (Figure 5(h)). These data implied that LINC00909 could function as a molecular sponge for miR-875-5p.
To find miR-875-5p's downstream targets, the starBase online tool was used and then overlapped with DEGs from GSE14359 and GSE12865, which resulted in nine shared genes (Figure 6(a)). We then adopted a Kaplan–Meier approach for exploring the association of overall survival with these nine genes in patients based on the TARGET database (Additional file 4: Figure S2). As suggested by the Kaplan–Meier survival method, HOXD9 upregulation predicted poor overall survival for OS cases (p < 0.05) (Figure 6(b)). Therefore, we predicted HOXD9 as a putative target for miR-875-5p. As a follow-up, we conducted WB and RT-qPCR on 60 clinical sample pairs and six cell samples. As a result, HOXD9 was dramatically upregulated in OS cells and tissues (Figures 6(c)–6(f)). Moreover, HOXD9 showed a negative correlation with miR-875-5p but a positive correlation with LINC00909 in OS clinical samples (Figures 6(g) and 6(h)). miR-875-5p overexpression reduced WT HOXD9 luciferase activity but made no significant difference to MUT HOXD9, as evidenced by the dual-luciferase reporter assay (Figure 6(i)). Additionally, miR-875-5p overexpression downregulated HOXD9, while the miR-875-5p inhibitor upregulated HOXD9 expression levels in OS cells, as assessed by RT-qPCR and WB assays (Figure 6(j)). Moreover, LINC00909 overexpression upregulated HOXD9 expression levels while LINC00909 downregulation reduced HOXD9 expression levels (Figure 6(k)). In the rescue experiment, LINC00909 downregulation eliminated miR-875-5p inhibitor-induced HOXD9 upregulation (Figure 6(l)). These data implied that HOXD9 is upregulated and is a target gene of miR-875-5p in OS.
Based on endogenous HOXD9 expression level in OS cells, we transfected sh-HOXD9 plasmid into MG63 and 143B cells. The expression level after transfection was assessed by RT-qPCR and WB (Figures 7(a) and 7(b)). We chose sh-HOXD9#1 for shRNA experiments, since it had the greatest suppression activity. For exploring the function of HOXD9 during OS cell proliferation, we performed CCK-8, colony formation, and EdU assays after HXOD9 knockdown. The colony forming ability of OS cells was decreased after HOXD9 downregulation (Figure 7(c)). CCK-8 assay revealed remarkably inhibited cell proliferation following HOXD9 downregulation (Figure 7(d)). In addition, the EdU proliferation assay showed that the mitotic cell proportion declined following decreased HOXD9 expression (Figure 7(e)). We assessed the role of HOXD9 in the migration and invasion of OS cells using scratch and transwell assays. The migrating cell number decreased when HOXD9 was knocked down (Figure 7(f)). The ability of cell migration and invasion decreased in the sh-HOXD9 group (Figures 7(g) and 7(h)). Furthermore, N-cadherin and vimentin levels were attenuated after HOXD9 downregulation, while E-cadherin was upregulated (Figure 7(i)), suggesting that HOXD9 activates EMT to enhance tumor metastasis. Taken together, the in vitro experiments showed that downregulation of HOXD9 inhibited OS cell proliferation, migration, invasion, and EMT.
We next sought to confirm the role of LINC00909 as a ceRNA that binds miR-875-5p and regulates HOXD9 expression to promote the proliferation, migration, invasion, and EMT of OS cells by a series of rescue experiments. CCK-8, colony formation, and EdU assays showed that cell proliferation was significantly inhibited with LINC00909 downregulation, while overexpression of HOXD9 or downregulation of miR-875-5p could reverse this effect (Figures 8(a)–8(c)). WB showed that LINC00909 downregulation attenuated expression levels of N-cadherin and vimentin, while E-cadherin levels were upregulated, and HOXD9 overexpression or downregulation of miR-875-5p could reverse this effect (Figure 8(d)). Scratch and transwell assays demonstrated that overexpression of HOXD9 or downregulation of miR-875-5p reversed the suppressive impact of LV-sh-LINC00909 on invasion and migration of OS cells (Figures 8(e)–8(g)). In conclusion, LINC00909 promotes OS cell proliferation, migration, invasion, and EMT by acting as a ceRNA to bind miR-875-5p while modulating HOXD9.
To explore the downstream mechanism of the LINC00909/miR-875-5p/HOXD9 axis, we performed GSEA to measure the downstream signaling pathway of HOXD9 in OS cells. Enrichment of the PI3K/AKT/mTOR pathway was observed when HOXD9 was highly expressed (Figures 9(a) and 9(b)). As reported previously, the PI3K/AKT/mTOR pathway is involved in tumor progression. We next investigated whether the LINC00909/miR-875-5p/HOXD9 axis promotes OS progression through the PI3K/AKT/mTOR pathway. Upregulating HOXD9 increased protein levels of HOXD9, P-PI3K, P-AKT, and P-mTOR, but downregulating LINC00909 reversed these effects (Figure 9(c)). Moreover, treating OS cells with AKT agonist (SC79) reversed the inhibitory effect of sh-HOXD9 on the expression of the G1/S checkpoint proteins and EMT-related proteins (Figure 9(d)). Overall, the above findings verify the involvement of the LINC00909/miR-875-5p/HOXD9 axis in regulating OS malignant progression via the PI3K/AKT/mTOR pathway (Additional file 5: Figure S3).
The literature has shown that lncRNAs play an extremely significant role in the malignant progression of OS. For example, the lncRNA lncARSR promotes OS progression by activating AKT [26]. lncRNA SNHG4 facilitates OS progression by binding miR-224-3p [27]. LINC00909 is highly expressed in a variety of tumors and is closely related to a poor prognosis. For instance, LINC00909 is highly expressed in ovarian cancer and significantly enhances the proliferation and metastasis of ovarian cancer cells [10]. However, the role of LINC00909 in OS has yet to be explored. In our study, we found that LINC00909 was highly expressed in OS metastatic cases in the TARGET database and positively correlated with EMT-associated gene signatures through bioinformatics analysis and RT-qPCR shows that LINC00909 levels are upregulated in OS cells and tissues. Moreover, LINC00909 expression showed a positive correlation with tumor size, ALP, metastasis, and TNM stage. According to experiments in vivo and in vitro, LINC00909 enhances the proliferation, migration, invasion, and EMT of OS cells. Therefore, we conclude that LINC00909 can serve as an oncogenic lncRNA in OS. Previous studies have found that miRNAs play an important role in the pathogenesis of tumors [28–30]. Furthermore, miRNAs play a vital role in the occurrence and development of OS [31, 32]. Although miR-875-5p is involved in inhibiting tumor progression in a variety of tumors [18, 19], the mechanism of action of miR-875-5p in OS has not previously been studied. We found that miR-875-5p expression was markedly decreased in OS cells and tissues. In vitro experiments also showed that the proliferation, migration, and invasion ability of OS was significantly reduced after miR-875-5p overexpression. Moreover, FISH, RIP, and luciferase assays demonstrated that miR-875-5p was a downstream target of LINC00909. These data suggest that LINC00909 exerts oncogenic effects by binding to miR-875-5p in OS. Recent studies have demonstrated that lncRNAs can play the role of ceRNAs that bind to miRNAs, which are involved in regulating target genes, thus affecting the development of tumors, including OS [33, 34]. For example, in vivo and in vitro experimental results show that lncRNA NEAT1 significantly promotes OS progression through the ceRNA mechanism [33]. Therefore, whether LINC00909 regulates the malignant progression of OS through the ceRNA mechanism should be explored. HOXD9 is one of the homeobox family members, which has a critical function in the morphogenesis of multicellular organisms [35]. Moreover, the expression of HOXD9 increases in diverse cancers and is correlated with patient prognosis [36]. Through a bioinformatics analysis, we suggest that HOXD9 is a potential target of miR-875-5p. HOXD9 is significantly upregulated in OS cells and tissues and is negatively correlated with miR-875-5p expression. Dual-luciferase reporter assay confirmed that HOXD9 is a downstream target of miR-875-5p. Rescue experiments also confirmed that LINC00909 promotes OS cell proliferation, migration, invasion, and EMT by playing the role of the ceRNA to bind to miR-875-5p and regulate HOXD9. The PI3K/AKT/mTOR pathway has been suggested to be involved in tumorigenesis [37, 38]. Moreover, the occurrence and development of OS are closely related to the signaling pathway [39]. Through a bioinformatics analysis of HOXD9 expression and the verification of proteins associated with the pathway by WB, we found that the LINC00909/miR-875-5p/HOXD9 axis regulates OS cell proliferation and EMT through the PI3K/AKT/mTOR pathway.
This work was conducted first to identify whether LINC00909 induces EMT and contributes to OS tumorigenesis and metastasis via the PI3K/AKT/mTOR pathway by binding to miR-875-5p to elevate HOXD9 expression. The present work illustrates the possible OS development mechanism and provides a novel anti-OS therapeutic strategy. Nonetheless, the OS pathogenic mechanism remains to be further elucidated. | true | true | true |
||
PMC9576429 | Zifeng Yang,Menglong Zhang,Jian Zhang,Cunkun Chu,Bijuan Hu,Liyin Huang | miR-302a-3p Promotes Radiotherapy Sensitivity of Hepatocellular Carcinoma by Regulating Cell Cycle via MCL1 | 10-10-2022 | Background The relationship between tumor suppressor gene miR-302a-3p and radiotherapy for hepatocellular carcinoma (HCC) remains unclear. This study intended to illustrate the molecular mechanism how miR-302a-3p regulated radiotherapy sensitivity of HCC. Methods miR-302a-3p expression in HCC tissues and cells was examined by qRT-PCR. The effect of miR-302a-3p on HCC radiotherapy sensitivity were detected by CCK-8, colony formation, and flow cytometry assays. The expression levels of cell cycle-related proteins were detected by Western blot. The influence of miR-302a-3p on radiotherapy sensitivity of HCC was further investigated via cell cycle inhibitor (Caudatin) treatment. The target gene (MCL1) of miR-302a-3p was obtained by bioinformatics analysis, and their binding relationship was confirmed by RNA-binding protein immunoprecipitation assay. The mechanisms of miR-302a-3p regulating cell cycle and affecting radiotherapy sensitivity of HCC cells through MCL1 were further explored through the rescue experiments. Results miR-302a-3p expression was remarkably reduced in radiotherapy-resistant tissues and cells of HCC. miR-302a-3p overexpression restored sensitivity of radiotherapy-resistant HCC cells to radiotherapy. Treatment with cell cycle inhibitor Caudatin could reverse suppressive effect of miR-302a-3p downregulation on sensitivity of HCC to radiotherapy. Additionally, miR-302a-3p could restrain MCL1 expression. In vitro cell assays further revealed that miR-302a-3p/MCL1 axis could enhance radiotherapy sensitivity of HCC cells by inducing G0/G1 arrest. Conclusions miR-302a-3p facilitated radiotherapy sensitivity of HCC cells by regulating cell cycle via MCL1, which provided a new underlying target for radiotherapy resistance of HCC patients. | miR-302a-3p Promotes Radiotherapy Sensitivity of Hepatocellular Carcinoma by Regulating Cell Cycle via MCL1
The relationship between tumor suppressor gene miR-302a-3p and radiotherapy for hepatocellular carcinoma (HCC) remains unclear. This study intended to illustrate the molecular mechanism how miR-302a-3p regulated radiotherapy sensitivity of HCC.
miR-302a-3p expression in HCC tissues and cells was examined by qRT-PCR. The effect of miR-302a-3p on HCC radiotherapy sensitivity were detected by CCK-8, colony formation, and flow cytometry assays. The expression levels of cell cycle-related proteins were detected by Western blot. The influence of miR-302a-3p on radiotherapy sensitivity of HCC was further investigated via cell cycle inhibitor (Caudatin) treatment. The target gene (MCL1) of miR-302a-3p was obtained by bioinformatics analysis, and their binding relationship was confirmed by RNA-binding protein immunoprecipitation assay. The mechanisms of miR-302a-3p regulating cell cycle and affecting radiotherapy sensitivity of HCC cells through MCL1 were further explored through the rescue experiments.
miR-302a-3p expression was remarkably reduced in radiotherapy-resistant tissues and cells of HCC. miR-302a-3p overexpression restored sensitivity of radiotherapy-resistant HCC cells to radiotherapy. Treatment with cell cycle inhibitor Caudatin could reverse suppressive effect of miR-302a-3p downregulation on sensitivity of HCC to radiotherapy. Additionally, miR-302a-3p could restrain MCL1 expression. In vitro cell assays further revealed that miR-302a-3p/MCL1 axis could enhance radiotherapy sensitivity of HCC cells by inducing G0/G1 arrest.
miR-302a-3p facilitated radiotherapy sensitivity of HCC cells by regulating cell cycle via MCL1, which provided a new underlying target for radiotherapy resistance of HCC patients.
Hepatocellular carcinoma (HCC) ranks fifth in prevalence of all cancers and second in cancer-related deaths worldwide [1]. The prognosis of HCC was poor with a 5-year survival rate of less than 30% [2]. However, due to unapparent symptoms at the early stage, most HCC patients are detected at an advanced stage, at which time patients cannot tolerate a surgical resection [3]. Radiation therapy is especially important for patients who are not eligible for a resection. The mainstream radiotherapy currently is the radiation therapy with a particle accelerator [4]. In addition, recent studies have shown that radioactive iodine-125 particle implantation therapy is an effective nonsurgical treatment for HCC patients who are unable to undergo resection. Iodine-125 particles are synthetic radionuclides that emit X- and γ-rays, which can damage tumor DNA and thus cause free radical production in tissues, which in turn kill tumor cells [5]. Currently, the feasible and well-tolerated radiotherapy has gradually become a noninvasive treatment for local ablation of HCC patients. However, radiotherapy-induced radiation resistance has seriously affected the control effect of radiotherapy on tumor [6]. The generation of radiotherapy resistance is related to many biological factors; the specific mechanism remains unclear. Therefore, elucidating the molecular mechanisms involved in radiotherapy resistance may help to explore therapeutic targets to improve the effectiveness of radiotherapy, thus achieving better therapeutic outcomes in HCC patients. MicroRNAs (miRNAs) play imperative regulatory roles in multiple biological and pathological processes of cancer [7]. Accumulating evidence has found that miRNAs affect radiotherapy resistance of various types of malignant tumors. For example, miR-612 was bound by TRPM2-AS in gastric cancer to increase FOXM1 expression and enhance radiotherapy resistance of gastric cancer [8]. MiR-208a could promote proliferation and radiotherapy resistance of lung cancer cells through targeting p21 [9]. In addition, miR-302a-3p can be a tumor suppressor in endometrial cancer, non-small-cell lung cancer, and melanoma [10–12], suggesting a tight relationship between miR-302a-3p and antitumor radiotherapy resistance. However, no studies have been reported on regulation of radiotherapy resistance of HCC via miR-302a-3p. Therefore, the present study preliminarily explored the potential impact and mechanism of miR-302a-3p on HCC radiotherapy resistance. DNA damage response induced by ionizing radiation is a highly complex and coordinated system. Ionizing radiation can affect cell cycle progression by activating DNA damage checkpoints, which are specific points that prevent or slow down the cell from entering the next stage in the cell cycle [13]. The G2/M checkpoint is modulated by many proteins in eukaryotic cells, like cell division cycle 2 protein (Cdc2) and cyclin B protein (cyclin B), and reduced expression of Cdc2 and cyclin B can trigger G2/M arrest [14]. It is noteworthy that the radiosensitivity of cells shows varying characteristics in different cell cycle phases [15]. Specifically, G2/M phase is the most sensitive stage to radiation, while cells in S phase are the most resistant to radiation. Therefore, drugs that alter the course of the cell cycle are usually effective radiotherapy modulators [16, 17]. In addition, cyclin-dependent kinases (CDKs) have been shown to drive normal cells from G1 phase into the cell cycle process. Storch and Cordes [18], for example, found that CDK9 loss delayed cell transition from G1 to S phase, thereby enhancing radiotherapy sensitivity of head and neck squamous cell carcinoma cells. It seems that cell cycle condition correlates much to the efficacy of radiotherapy. Therefore, we assumed that miR-302a-3p enhances radiotherapy sensitivity by modulating cell cycle in HCC. Herein, we revealed that miR-302a-3p expression is related to tumor radiotherapy resistance, and further cell experiments found that miR-302a-3p could induce G0/G1 arrest and enhance the radiotherapy sensitivity of HCC. Our study experimentally proved that miR-302a-3p regulated cell cycle progression and facilitated radiotherapy sensitivity of HCC by downregulating MCL1 expression. These findings shed new insights into the mechanism of radiotherapy resistance of HCC and provided theoretical basis for miR-302a-3p as a promising radiotherapy sensitization target.
The mRNA expression profile of HCC was obtained from TCGA database, and then, the TargetScan (http://www.targetscan.org/vert_72/), miRDB (http://mirdb.org/), and starBase (http://starbase.sysu.edu.cn/) databases were used to predict downstream target genes of miR-302a-3p to obtain the differentially expressed mRNAs with binding site to miR-302a-3p. The predicted results were intersected with differentially downregulated mRNAs. Finally, the target gene was determined through correlation analysis.
With the approval of the Ethics Committee of Ganzhou People's Hospital, we collected 30 HCC tissue samples from patients who were sensitive or resistant to radiotherapy. The mean age of the patients included in this study was 60.26 years, including 13 females and 17 males. All the patients were in intermediate and advanced clinical stages. Cancer tissue sample collection was performed after the patients received the first radiotherapy. All samples were from patients receiving radical radiotherapy without a distant metastasis in Ganzhou People's Hospital. Samples were taken during surgery and then quickly frozen and stored at -80°C. Radiotherapy was performed according to standard treatment regimens, and efficacy was assessed at the end of radiotherapy to determine whether patients (and corresponding specimens) are resistant or sensitive to radiotherapy. All patients had signed the informed consents.
Human HCC cell line HepG2 (BNCC338070) was provided by BeNa Culture Collection (BNCC, China). HepG2 cells were cultured in DMEM +10% fetal bovine serum (FBS) medium and placed in an incubator at 37°C with 5% CO2. The establishment of radiation-resistant HCC cells was performed as per the method described by Chen et al. [19]. Specifically, the HepG2 cells were treated with 6-MV X-rays generated by a linear accelerator (Varian 2300EX, dose rate of 2 Gy/min; Varian, USA). The cells were cultured after the initial radiation of 2 Gy and subcultured twice. The surviving cells were then exposed to a series of gradually increasing radiation doses (4, 6, 8, and 10 Gy) twice. The total radiation dose was 50 Gy, and the whole selection process was finished within 6 months. The final surviving cells were defined as HepG2/RR (radiation-resistant). We used parental HepG2 cells as controls and were defined as HepG2/RS (radiation-sensitive). HepG2/RR and HepG2/RS were cultured at 37°C for 24 h before further analyses.
miR-302a-3p mimic (miR-mimic), miR-302a-3p inhibitor (miR-inhibitor), and their negative controls (mimic-NC or NC-inhibitor) were all provided by Ribobio (China). The plasmids of miR-mimic/miR-NC, miR-inhibitor/NC-inhibitor, and oe-NC/oe-MCL1 were transfected into HCC cell line HepG2/RR by using Lipofectamine 2000 kit (Invitrogen, USA). The cells were then treated with X-rays and cultured for 24 h for following experiments.
Caudatin (Medherb Biotechnology, China) was dissolved in dimethyl sulfoxide, and HepG2/RR cells were treated with 100 μM Caudatin for 24 h for subsequent experiments.
Total RNA was separated with RNAiso Plus (Takara, Japan), and Prime Script TM RT Master Mix (Takara, Japan) was recommended for reverse transcription. qPCR was completed on SYBR Premix Ex Taq II (Takara, Japan). The operation for RNA extraction and PCR analysis were carried out as per the methods described by predecessors [20]. U6 snRNA and GAPDH were served as internal references to quantify miRNA and mRNA in cells. qPCR was performed with StepOnePlus Real-Time PCR system (AB, USA). The PCR results were quantified by 2−ΔΔCt, and the experiment was repeated 3 times. Table 1 exhibits the primer sequences.
The required proteins were extracted from cells, and the protein concentration was determined by BCA kit (Beyotime, China). The same amount of proteins (50 μg) was added to each well of SDS-PAGE gel. After electrophoresis for 1.5 h, the protein was transferred to PVDF membrane. The PVDF membrane was placed in 0.5% skim milk powder and blocked at room temperature for 1.5 h. The PVDF membrane was then probed with the primary antibodies and incubated overnight (4°C). After washing three times with TBST, the membrane was placed in the second antibody and incubated for 2 h (room temperature). ECL kit (Thermo Scientific, USA) was used for color developing. Detailed steps of Western blot were performed according to previously described methods [21], and the experiment was repeated 3 times. The primary antibodies included rabbit anti-cyclin D1 (ab134175), rabbit anti-cyclin E1 (ab224819), rabbit anti-P27 (ab137736), and rabbit anti-GAPDH (ab181602), and the secondary antibody was goat anti-rabbit IgG (ab205718), all purchased from Abcam (Abcam, UK).
To investigate sensitivity of HepG2/RS and HepG2/RR cells to radiotherapy, cells that had received different ionizing radiation doses (0, 2, 4, 6, 8, and 10 Gy) were inoculated into a 96-well plate (2 × 104 cells/well). After 24 h of culture, CCK-8 reagent was added as per the instructions of CCK-8 kit (Beyotime, China), and absorbance of cells was determined at 450 nm. To detect sensitivity of miR-302a-3p-transfected HepG2/RR cells to radiotherapy, cell viability was detected on d1, d2, d3, d4, d5, d6, d7, and d8 according to kit instructions after 8 Gy radiation treatment, and the experiment was repeated 3 times.
Transfected or untransfected HepG2/RR cells were plated into a 6-well plate (1 × 103 cells per well). The cells received different ionizing radiation doses (0, 2, 4, 6, 8, and 10 Gy). Culture medium was replaced after 24 h, and then, the cells were maintained for 10-14 days. When visible cell colonies appeared, the medium was discarded. The cells were rinsed 3 times with cold PBS, followed by fixing with paraformaldehyde for 15 min and staining with 0.5% crystal violet solution for 15 min. After washing with PBS, cells were dried, and colonies were photographed and counted using camera.
For cell cycle detection, transfected cells in each group were cultured to logarithmic growth phase. After receiving 8 Gy ionizing radiation and being cultured for 24 h, the cells were digested with trypsin, washed with PBS, and fixed in 70% ethanol overnight. Then, the cells were dyed with propidium iodide. The distribution of cell cycle was analyzed by flow cytometry.
RIP assay was carried out according to the steps of Zhang et al. [22]. In brief, HepG2 cells were treated with miR-NC or miR-mimic. 48 h later, the transfected cells were subjected to RIP assay with the Magna RIP™ RNA Binding Protein Immunoprecipitation Kit (Millipore, USA). Subsequently, the cells were probed with anti-Ago2 antibody or negative control IgG. The relative enrichment degree of MCL1 was determined by qRT-PCR.
Data graphs and data analysis in this paper were plotted and analyzed using GraphPad Prism 8 (GraphPad Software, USA), and intergroup comparison was performed using ANOVA or t-test. The data were presented as the mean ± standard deviation (SD) from at least three independent experiments. p < 0.05 meant that the difference was significant.
To study the role of miR-302a-3p in regulating HCC resistance to radiotherapy, we first measured the expression of miR-302a-3p expression in tumor tissues of HCC patients with radiotherapy sensitivity or radiotherapy resistance. The result displayed that miR-302a-3p was dramatically underexpressed in tumor tissues with radiotherapy resistance (Figure 1(a)). Then, HepG2 cells were used to establish radiotherapy-sensitive HepG2/RS and radiotherapy-resistant HepG2/RR. qRT-PCR was utilized to detect miR-302a-3p level in HepG2/RS and HepG2/RR cells. The result displayed a notable reduction in miR-302a-3p expression in radiotherapy-resistant HepG2/RR cells compared to its parental radiotherapy-sensitive HepG2/RS cells (Figure 1(b)). Subsequent CCK-8 result displayed that radiotherapy resistance of HepG2/RR cells was significantly enhanced compared with HepG2/RS, indicating that HepG2/RR cells obtained higher radiotherapy resistance, with a median lethal radiation dose of 8 Gy (Figure 1(c)). In conclusion, we successfully constructed radiotherapy-sensitive HepG2/RS and radiotherapy-resistant HepG2/RR cells and found that miR-302a-3p was dramatically underexpressed in HCC tissues and cells with radiotherapy resistance.
To preliminarily explore effects of miR-302a-3p on cell viability, survival rate, and cycle of radiotherapy-resistant HCC cells, we constructed HepG2/RR cells with overexpression miR-302a-3p (miR-mimic) and its negative control (mimic-NC). First, the transfection efficiency of miR-302a-3p was evaluated by qRT-PCR, and the result indicated a prominent increase in miR-302a-3p expression in miR-mimic group (Figure 2(a)). Then, the cell viability was detected by CCK-8. It was demonstrated that miR-302a-3p overexpression remarkably reduced viability of HepG2/RR cells compared with control group (Figure 2(b)). Besides, the result of colony formation assay demonstrated that survival rate of HepG2/RR cells with miR-302a-3p mimic was notably reduced (Figure 2(c)) after cells were exposed to radiation, indicating that radiotherapy sensitivity of HepG2/RR cells was improved. Then, flow cytometry was adopted to detect cell cycle distribution in each treatment group. It was found that cells in G0/G1 phase were prominently increased and those in S phase were remarkably decreased in miR-mimic group (Figure 2(d)). Western blot analysis showed that expression levels of cyclin D1 and cyclin E1 were significantly downregulated, while expression of cell cycle regulator P27 was remarkably upregulated after miR-302a-3p overexpression (Figure 2(e)). In conclusion, these results indicated that overexpression of miR-302a-3p enhanced radiotherapy sensitivity, inhibited cell viability, and led to cell cycle arrest at G0/G1 phase.
To further analyze the impact of miR-302a-3p on sensitivity of HCC cells by affecting cell cycle, we set the following cell groups: negative control group (NC-inhibitor+DMSO), miR-302a-3p low expression group (miR-inhibitor+DMSO), and miR-302a-3p low expression with 100 μM cell cycle inhibitor Caudatin group (miR-inhibitor+Caudatin). First, CCK-8 was utilized to detect effects of different treatments on cell viability. Compared to control group, the HepG2/RR cells with downregulated miR-302a-3p expression showed enhanced viability, which was reversed by the simultaneous addition of cell cycle inhibitor (Figure 3(a)). Then, colony formation assay was introduced to measure survival rate of HepG2/RR cells in each treatment group. The experimental result demonstrated that survival rate of HepG2/RR cells with downregulated miR-302a-3p was significantly increased after radiation exposure, while simultaneous addition of cell cycle inhibitor reversed effect of miR-302a-3p on survival rate of HepG2/RR cells (Figure 3(b)). Subsequently, flow cytometry was applied to analyze cell cycle distribution of each treatment group. According to the results, the number of G0/G1 phase cells was notably reduced in miR-inhibitor+DMSO group and prominently increased in miR-inhibitor+Caudatin group compared to NC-inhibitor+DMSO group (Figure 3(c)). Additionally, Western blot was applied to evaluate expression of cell cycle-related proteins. According to the experimental results, cyclin D1 and cyclin E1 expression was upregulated, and P27 expression was downregulated in miR-inhibitor+DMSO group compared to NC-inhibitor+DMSO group, while cyclin D1 and cyclin E1 expression was downregulated, and P27 expression was upregulated in miR-inhibitor+Caudatin group compared to miR-inhibitor+DMSO group (Figure 3(d)). These results implied that miR-302a-3p could make HCC cells sensitive to radiotherapy by regulating cell cycle.
Studies have shown that miRNA mainly participates in the molecular regulatory pathway by regulating expression of target genes. We further explored downstream target mRNAs of miR-302a-3p through bioinformatics methods. First, differential analysis on the mRNAs in TCGA-HCC database were performed. Then, the obtained differentially downregulated mRNAs were intersected with the target genes of miR-302a-3p predicted by TargetScan, starBase, and miRDB databases. By this way, differential mRNAs with binding site to miR-302a-3p were obtained (Figure 4(a)). After reviewing the literature, we found that MCL1 was a key protein related to cell cycle [23] and that MCL1 was significantly overexpressed in tumor tissues with radiotherapy resistance (Figure 4(b)). Therefore, we selected MCL1 as the research object to explore its influence on the radiotherapy sensitivity of HCC cells. Subsequently, the binding site between miR-302a-3p and MCL1 was predicted by bioinformatics analysis (Figure 4(c)). The binding relationship between miR-302a-3p and MCL1 was further verified by RIP experiment, and the result showed that MCL1 was significantly enriched in miR-mimic cells (Figure 4(d)). To investigate whether miR-302a-3p can affect cell cycle progression by targeting MCL1, we first constructed miR-302a-3p overexpression cells (miR-mimic+oe-NC), MCL1 overexpression cells (mimic-NC+oe-MCL1), and miR-302a-3p and MCL1 simultaneous overexpression cells (miR-mimic+oe-MCL1) using HepG2/RR cells. According to qRT-PCR result, the mRNA expression of MCL1 was significantly reduced in miR-mimic+oe-NC group in comparison with control group, but it was recovered in miR-mimic+oe-MCL1 cotransfection group (Figure 4(e)). With regard to CCK-8 result, the viability of HepG2/RR cells in mimic-NC+oe-MCL1 group was enhanced compared to the mimic-NC+oe-NC, but it was restored in miR-mimic+oe-MCL1 group (Figure 4(f)). Then, flow cytometry result indicated the number of G0/G1 cells in mimic-NC+oe-MCL1 was significantly reduced compared to mimic-NC+oe-NC group, but the number in miR-mimic+oe-MCL1 group was recovered (Figure 4(g)). Western bolt was used to evaluate the expression of cell cycle-related proteins. According to the result, cyclin D1 and cyclin E1 expression was upregulated, and P27 expression was downregulated in mimic-NC+oe-MCL1 group compared to mimic-NC+oe-NC group, while expression levels of these proteins were recovered in miR-mimic+oe-MCL1 group (Figure 4(h)). These results indicated that miR-302a-3p could promote radiotherapy sensitivity of HCC cells by regulating HepG2/RR cell cycle progression via downregulation of MCL1.
Radiotherapy resistance is an important factor leading to clinical radiotherapy failure in HCC patients. Radiotherapy resistance can be attributed to the inherent radioresistance of tumor cells in hypoxic microenvironment or to the resistance acquired during hyperfraction radiotherapy [24]. Studies have shown that radiation exposure can increase the levels of intracellular free radical species, cause DNA strand breaks, and lead to subsequent dysfunction of some organelles such as mitochondria and endoplasmic reticulum [25]. These radiation-induced cellular events facilitate proapoptotic signal activation and ultimately cause tumor cell killing [26]. Nevertheless, the key molecules involved in radiation-induced radiotherapy resistance are still poorly understood. Hence, it is an urgent task to study the molecular mechanism of radiotherapy resistance and find new therapeutic targets, which are of great importance to overcome cancer radiotherapy resistance. The regulatory mechanisms of miRNAs in the progression of many complex diseases have been extensively studied. miR-302a-3p plays a potential cancer-promoting role. For example, Zhang et al. [27] found that miR-302a-3p directly targets SOCS5 to promote STAT3 phosphorylation and induce transcription of STAT3 target genes, thereby promoting metastasis of pancreatic cancer cells. However, miR-302a-3p acts as an important tumor suppressor in most cases, where it inhibits the biological processes like proliferation, invasion, and migration of various human cancers, including HCC, colon cancer, and gastric cancer [28–30]. It is worth noting that miR-302a, as the precursor of miR-302a-3p, has been proved to participate in the inhibition of tumor chemotherapy resistance development [31, 32]. In addition, Liang et al. [33] discovered that miR-302a sensitizes radiation-resistant breast cancer cells to radiotherapy both in vivo and in vitro. Yu et al. [34] obtained similar results and found that miR-302a overexpression hampers proliferation of non-small-cell lung cancer, promotes cell apoptosis, and reduces cell radioresistance. Therefore, it is reasonable to speculate that miR-302a-3p is associated with HCC radiotherapy resistance. In this study, miR-302a-3p was remarkably downregulated in radiotherapy-resistant cells (HepG2/RR) compared with radiotherapy-sensitive cells (HepG2/RS). However, miR-302a-3p overexpression could enhance the radiotherapy sensitivity of HCC cells, confirming that miR-302a-3p mediated radiotherapy resistance in HCC. Radiation resistance is a complex cellular response involving many signaling pathways and genes. Ionizing radiation can induce DNA damage, including DNA single-strand break, DNA base damage, and DNA double-strand break [35]. In addition, DNA damage induced by ionizing radiation can activate a series of cell cycle checkpoints [36]. Hence, cell cycle checkpoint block is involved in the regulation of tumor radiotherapy resistance. For example, Wang et al. [37] showed that cyclin D1 and cyclin E1 are specific targets of miR-16-5p and that miR-16-5p overexpression can downregulate expression of cyclin D1 and cyclin E1 and induce cell cycle arrest in G0/G1 phase, which enhances radiosensitivity of prostate cancer cells. Although numerous studies have revealed the role of periodic checkpoint block in tumor radiotherapy resistance, there is little research on the regulatory role of miR-302a-3p. In the present study, in vitro cell experiments confirmed that miR-302a-3p induced cell arrest in G0/G1 phase by regulating expression of key cell cycle proteins cyclin D1, cyclin E1, and P27 and promoted the sensitivity of radiation-resistant HCC cells to radiotherapy, which was consistent with previous studies. In addition, this study predicted a new downstream target of miR-302a-3p, MCL1, through bioinformatics analysis, and verified the targeted relationship between the two. MCL1 has been proved to be overexpressed in many human cancers and contributes to cancer occurrence and inhibits apoptosis [38]. Specific targeting of MCL1 may overcome the antiapoptotic ability of malignant tumor cells. For example, BAG3 can upregulate MCL1 by downregulating miR-29b, thereby inducing chemotherapy resistance to paclitaxel in ovarian cancer [39]. In addition, Yu et al. found that MCL1 overexpression significantly inhibited mulanin-induced autophagy and cell cycle arrest in colorectal cancer cells [40]. In the present study, rescue experiments demonstrated that miR-302a-3p induced cell cycle arrest and promoted sensitivity of HCC cells to radiotherapy by targeting MCL1 expression, which was in accordance with the results of previous reports. Our results confirmed that miR-302a-3p could enhance radiotherapy sensitivity of radiation-resistant HCC cells by regulating the cell cycle. In addition, our study revealed influences of miR-302a-3p/MCL1 axis on the radiotherapy sensitivity of HCC cells for the first time. However, the mechanism of miR-302a-3p enhancing the radiotherapy sensitivity of HCC was analyzed at the cellular level, which has not been verified in vivo at the animal level. Meanwhile, it also lacked the exploration of relevant signal pathways, which was the deficiency of this study. In summary, this study for the first time clarified the modulatory mechanism of miR-302a-3p in HCC radiotherapy resistance, suggesting that miR-302a-3p may be a potential sensitizer of radiotherapy. Our results provided new evidence for the radiotherapy resistance mechanism of HCC cells and can help overcome the difficulties in cancer radiotherapy in the future. | true | true | true |
||
PMC9576446 | Lixue Liu,Ru Bai,Debang Li,Bai Dai,Ya Tuo | LncRNA BANCR Promotes Endometrial Stromal Cell Proliferation and Invasion in Endometriosis via the miR-15a-5p/TRIM59 Axis | 10-10-2022 | Long non-coding RNA (LncRNA) emerges as a regulator in various diseases, including endometriosis (EM). This study aims to uncover the role of long non-coding RNA BRAF-activated non-protein coding RNA (lncRNA BANCR)-mediated competing endogenous RNA mechanism in endometrial stromal cell (ESC) proliferation and invasion in EM by regulating miR-15a-5p/TRIM59. ESCs were isolated from eutopic and ectopic endometrial tissues, followed by the determination of Cytokeratin 19 and Vimentin expressions in cells. Then, expressions of lncRNA BANCR, microRNA (miR)-15a-5p, and tripartite motif-containing 59 (TRIM59) in tissues and cells were determined by real-time quantitative polymerase chain reaction or Western blot assay, and cell proliferation and invasion were evaluated by cell counting kit-8 and transwell assays. After that, the subcellular localization of lncRNA BANCR and binding of miR-15a-5p to lncRNA BANCR or TRIM59 were analyzed. LncRNA BANCR was upregulated in ectopic endometrial tissues and ectopic ESCs (Ect-ESCs). Silencing lncRNA BANCR suppressed Ect-ESC proliferation and invasion. LncRNA BANCR inhibited miR-15a-5p to promote TRIM59 expression. miR-15a-5p downregulation or TRIM59 overexpression both reversed the effects of silencing lncRNA BANCR on Ect-ESC proliferation and invasion. In summary, our findings suggested that lncRNA BANCR facilitated Ect-ESC proliferation and invasion by inhibiting miR-15a-5p and promoting TRIM59. | LncRNA BANCR Promotes Endometrial Stromal Cell Proliferation and Invasion in Endometriosis via the miR-15a-5p/TRIM59 Axis
Long non-coding RNA (LncRNA) emerges as a regulator in various diseases, including endometriosis (EM). This study aims to uncover the role of long non-coding RNA BRAF-activated non-protein coding RNA (lncRNA BANCR)-mediated competing endogenous RNA mechanism in endometrial stromal cell (ESC) proliferation and invasion in EM by regulating miR-15a-5p/TRIM59. ESCs were isolated from eutopic and ectopic endometrial tissues, followed by the determination of Cytokeratin 19 and Vimentin expressions in cells. Then, expressions of lncRNA BANCR, microRNA (miR)-15a-5p, and tripartite motif-containing 59 (TRIM59) in tissues and cells were determined by real-time quantitative polymerase chain reaction or Western blot assay, and cell proliferation and invasion were evaluated by cell counting kit-8 and transwell assays. After that, the subcellular localization of lncRNA BANCR and binding of miR-15a-5p to lncRNA BANCR or TRIM59 were analyzed. LncRNA BANCR was upregulated in ectopic endometrial tissues and ectopic ESCs (Ect-ESCs). Silencing lncRNA BANCR suppressed Ect-ESC proliferation and invasion. LncRNA BANCR inhibited miR-15a-5p to promote TRIM59 expression. miR-15a-5p downregulation or TRIM59 overexpression both reversed the effects of silencing lncRNA BANCR on Ect-ESC proliferation and invasion. In summary, our findings suggested that lncRNA BANCR facilitated Ect-ESC proliferation and invasion by inhibiting miR-15a-5p and promoting TRIM59.
Endometriosis (EM) is termed a complex gynecological disorder, diagnosed by the presence of endometrial-like tissue, glands, and stroma outside the uterine cavity [1]. EM will affect about 10% of the female population and not only can it result in heavy social and economic burdens, including, pain, infertility, depression and anxiety, health care costs, and indirect productivity loss, but data are accumulating that malignant transformation is an important consideration [2, 3]. EM is a benign disease similar to malignancy in some perspectives, including estrogen-stimulated proliferation, recurrence, and tendency to metastasis [4]. Pathologically, endometrial stromal cells (ESCs), the major resident cells in human endometrium, are switched to invasive and proliferative phenotypes under the stimulation of cytokines and chemokines, thus exacerbating EM progression [5, 6]. Most current therapies focus on alleviation of pain symptoms and prevention of recurrence but lack the curable intent [7]. Therefore, it is prudent to explore novel molecules sensitive to EM proliferation and invasion, in an effort to improve the prognosis of EM patients. Long non-coding RNA (LncRNA), the well-studied RNA molecules that regulate multiple cellular and biological processes, are documented to participate in hormone response, behaviors of ESCs, autophagy, and immune disorder in the context of EM, thus affecting EM progression [8]. LncRNA BRAF-activated non-protein coding RNA (BANCR), a well-established oncogenic lncRNA in human cancers, is involved in cell proliferation, migration, invasion, apoptosis, and epithelial to mesenchymal transition [9, 10]. More importantly, lncRNA BANCR inhibitor is potent to reduce the volume of eutopic endometrium in animal models by suppressing the production of angiogenic factors [11]. However, the expression pattern and role of lncRNA BANCR in ectopic ESCs (Ect-ESCs) remain indistinct. Some lncRNAs can act as competing endogenous RNAs (ceRNAs) to compete for the microRNA (miRNA) binding sites by means of partial complementarity and further modulate endogenous mRNAs [12]. Especially, lncRNA BANCR could serve as a sponge for various miRNAs in cancers, such as miR-590-5P, miR-195-5p, and miR-338-3p [13–15]. In this study, our bioinformatics and experimental data suggested a correlation between lncRNA BANCR and miR-15a-5p/tripartite motif-containing 59 (TRIM59). miR-15a-5p is a non-coding RNA that binds to its complementary messenger RNA (mRNA) and inhibits mRNA translation by regulating mRNA degradation [16, 17]. Of note, miR-15a-5p has been demonstrated to show downregulated expression patterns in endometriotic tissues and Ect-ESCs-derived extracellular vesicles [18, 19], and inhibition of miR-15a-5p promotes ESC migration and invasion [20]. Moreover, TRIM59, a member of the TRIM protein family that is associated with autophagy, immunity, anti-virus, and carcinogenesis and plays roles in the pathogenesis of osteoarthritis, sepsis, and myocardial ischemia-reperfusion injury [21–25], and TRIM59 overexpression can facilitate the invasion of ectopic ESCs (Ect-ESCs) [26]. Nevertheless, the ceRNA network of lncRNA BANCR/miR-590-5p/TRIM59 in EM has not been studied before and warrants further investigation. Taking the aforementioned associations into consideration, we raised a hypothesis that lncRNA BANCR mediates a ceRNA network with miR-15a-5p/TRIM59 to affect Ect-ESC proliferation and invasion. In this manner, the present study sought to investigate the molecular mechanism of lncRNA BANCR in EM, hoping to provide a novel theoretical reference for EM treatment.
A total number of 20 EM patients (28–40 years old) who received tissue resection in The Affiliated Hospital of Inner Mongolia Medical University were included in this study for the collection of eutopic and ectopic endometrial tissues. The procedures of sample collection were ratified by the medical ethics committee of The Affiliated Hospital of Inner Mongolia Medical University and conformed to the Declaration of Helsinki. The written informed content was signed by each patient. The collected samples were stored in liquid nitrogen and kept at –80°C before the subsequent uses. Inclusion criteria of EM patients were as follows [27]: (1) at reproductive age (19–45 years old); (2) at the proliferative phase of the menstrual cycle; (3) EM was confirmed by laparoscopic surgery and postoperative histological examination. Exclusion criteria of EM patients were as follows: (1) with any history of malignancy; (2) with autoimmune or metabolic disorders; (3) taking any hormonal medications or dietary supplements within the last three months before the surgery; (4) pregnancy or lactation.
hESCs were obtained from eutopic and ectopic endometrial tissues of EM patients in The Affiliated Hospital of Inner Mongolia Medical University. The procedures of cell isolation were as follows: firstly, endometrial tissues were shredded, detached with 4% collagenase (60 min), and centrifuged (500 × g, 5 min); secondly, the cell suspensions were centrifuged (3000 × g, 10 min); thirdly, cell precipitates were resuspended in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS, Gibco, Carlsbad, CA, USA), 100 U/mL penicillin, and 100 mg/L streptomycin (Solarbio, Beijing, China). Then, isolated ESCs were placed in an incubator and kept at 37°C, and the medium was renewed after the first 24 h of cell culture and then replaced every 2–3 days.
Cells were fixed with 4% paraformaldehyde, washed with phosphate buffer saline (PBS), and incubated overnight with rabbit monoclonal antibody Cytokeratin 19 (CK19, ab52625, 1 : 400, Abcam, Cambridge, MA, USA) or rabbit monoclonal antibody Vimentin (ab92547, 1 : 200, Abcam) at 4°C. Next, cells were incubated with the secondary antibody IgG (ab6721, 1 : 1000, Abcam) at 37°C for 30 min, followed by color-developing with diaminobenzidine (Sigma Co., St. Louis, MO, USA) and counterstaining with hematoxylin for 1 min. Afterwards, cells were observed and photographed using an optical microscope (Eclipse E200, Nikon Co., Tokyo, Japan). The protein levels were analyzed with the help of Image-pro Plus software (Media Cybernetics, San Diego, CA, USA).
Two strands of small interfering (si) RNA targeting lncRNA BANCR (si-BANCR#1 and si-BANCR#2), miR-15a-5p inhibitor (miR-inhibitor), pcDNA3.1 vector-packaged overexpressed (oe)-TRIM59, and their negative controls were supplied by GenePharma (Shanghai, China). After Ect-ESCs were cultured in 6-well plates (3 × 105cells/well) for 24 h, the above plasmids were transiently transfected into Ect-ESCs applying Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA) according to manufacturer's requirements.
Ect-ESCs were placed in 96-well plates (3 × 103cells/well) 24 h after cell transfection. At certain time points of cell culture (0, 24, 48, and 72 h), Ect-ESCs were incubated with 10 μL of CCK-8 solution at 37°C for 1 h. After that, the absorbance at a wavelength of 450 nm was measured on a microplate reader (Bio-Rad, Hercules, CA, USA).
Ect-ESCs were placed into 6-well plates (500 cells/well) for 10 days of cell culture. After that, Ect-ESCs were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet solution (Sigma, St. Louis, MO, USA) for 20 min. Next, the plate was washed with PBS and dried, and the number of colonies in each well was counted and analyzed.
The invasion potential of Ect-ESCs was analyzed using transwell assay. Simply put, the apical transwell chamber was pre-coated with Matrigel (BD Biosciences, San Jose, CA, USA), followed by cell seeding (4 × 104cells/0.1 mL). Then, the basolateral transwell chamber was supplemented with DMEM supplemented with 10% FBS. After 24 h of cell culture, cells were stained with crystal violet. Non-invasive cells were wiped out, and the number of invasive cells was counted under a microscope.
After extraction of the total RNA content in cells and tissues using the TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA), the extracted RNA was reverse-transcribed into the complementary DNA using RT-PCR assay kits (K1621, Thermo Fisher Scientific, Waltham, MA, USA). The quantification of gene expression was achieved by real-time PCR. The relative amount of gene expression was quantified according to the 2-ΔΔCt method, with GAPDH and U6 as endogenous references [28]. The primers used in RT-qPCR are shown in Table 1.
The total protein samples were extracted from tissues or cells using radioimmunoprecipitation assay buffer. According to the instructions of the producer, the protein concentration was determined using the bicinchoninic acid protein assay kit (ab102536, Abcam, Cambridge, UK). An equal amount of protein (30 μg) was separated by 10% sodium dodecyl sulfate-polyacrylamide gel and transferred onto polyvinylidene fluoride membranes. After blockade with Tris-HCl buffered saline containing 0.1% Tween-20 and 5% skim milk, membranes were incubated with primary antibodies against TRIM59 (ab166793, 1 : 500, Abcam) and GAPDH (ab181602, 1 : 10000, Abcam) overnight and then with secondary antibody against IgG (ab6721, 1 : 2000, Abcam) for 1 h. Membranes were visualized using the enhanced chemiluminescence-PLUS (Amersham Biosciences, Sweden). Eventually, the grayscale was quantified using Image-Pro Plus (Media Cybernetics, San Diego, CA, USA).
The subcellular localization of lncRNA BANCR was predicted on the lncLocator database (http://www.csbio.sjtu.edu.cn/bioinf/lncLocator/?tdsourcetag=s_pcqq_aiomsg) [29]. The downstream miRNAs of lncRNA BANCR (Table S1) and the binding sites of lncRNA BANCR to miR-15a-5p were predicted on the RNA22 database (https://cm.jefferson.edu/rna22/Interactive/) [30]. The downstream target genes of miR-15a-5p were predicted on Starbase (http://starbase.sysu.edu.cn/) [31] and miRWalk (http://mirwalk.umm.uni-heidelberg.de/) [32] databases (Table S2-S3). The binding sites of miR-15a-5p to TRIM59b were predicted on the Starbase database.
Following the manufacturer's protocol, the nucleus and cytoplasm were separated using PARIS kits (AM1921, Thermo Fisher Scientific). In brief, Ect-ESCs were collected, lysed with the preparation buffer, and then centrifuged. The fractionation of the supernatant and nuclear pellet was achieved using the cell disruption buffer, and the nuclear pellet was preserved for RNA extraction and analysis. Lastly, lncRNA BANCR was analyzed by RT-qPCR, with U6 (the control of nucleus) and GAPDH (the control of cytoplasm) as internal references.
The synthesized BANCR fragments containing the binding sites to miR-15a-5p and BANCR fragments containing the mutant binding sites to miR-15a-5p, as well as TRIM59 fragments containing the binding sites to miR-15a-5p and TRIM59 fragments containing the mutant binding sites to miR-15a-5p were inserted into pmirGLO-reporter plasmids. The above-constructed luciferase reporter plasmids (BANCR-WT or BANCR-MUT, TRIM59-WT or TRIM59-MUT) were co-transfected into Ect-ESCs with mimics-NC or miR-15a-5p mimics (miR-mimics). After 48 h, cells were harvested and lysed, and the luciferase activity was quantified using the luciferase assay kits (16186, Thermo Fisher Scientific). All plasmids were supplied by GenePharma.
Data analysis and graphing were conducted with the help of SPSS21.0 software (IBM Corp, Armonk, NY, USA) and GraphPad Prism 8.0 software (GraphPad Software Inc., San Diego, CA, USA). Measurement data were formalized as mean ± standard deviation (SD). First, data were subjected to normality and homogeneity of variance tests, which revealed that data conformed to normal distribution and homogeneity of variance. Then, pairwise comparisons were analyzed using the t test, and multi-group comparisons were analyzed using one-way or two-way analysis of variance (ANOVA), followed by Tukey's multiple comparison test. The P values were attained by two-sided tests, and a value of P < 0.01 was indicative of extremely statistical significance.
A previous study has documented the upregulation of lncRNA BANCR in EM [11]. First, we determined the expression levels of lncRNA BANCR in eutopic and ectopic endometrial tissues and found that lncRNA BANCR was upregulated in ectopic endometrial tissues (P < 0.05, Figure 1(a)). Then, we isolated ESCs from eutopic and ectopic endometrial tissues. Through immunocytochemistry, it was found that cytoskeleton protein CK19 was negatively expressed and Vimentin was positively expressed (Figure 1(b)), indicating the successful extraction of ESCs. Accordingly, the subsequent experiment revealed higher expression levels of lncRNA BANCR in Ect-ESCs relative to eutopic ESCs (P < 0.05, Figure 1(c)). Several studies have documented that suppression of abnormal proliferation or invasion of ESCs mitigates EM progression [33–35]. To further explore the role of lncRNA BANCR in ESC proliferation and invasion, Ect-ESCs were transfected with si-BANCR (si-BANCR#1 and si-BANCR#2) to downregulate lncRNA BANCR expression (P < 0.05, Figure 1(d)). It was found that the proliferation potential of Ect-ESCs was significantly reduced upon silencing lncRNA BANCR (P < 0.05, Figure 1(e)-1(f)), and so was the invasion potential of Ect-ESCs (P < 0.05, Figure 1(g)). Altogether, the preceding data suggested that silencing lncRNA BANCR suppressed Ect-ESC proliferation and invasion.
To explore the downstream mechanism of lncRNA BANCR, we predicted the subcellular localization of lncRNA BANCR on the lncLocator database and found that lncRNA BANCR was predominantly expressed in the cytoplasm (Figure 2(a)). The subcellular fractionation assay further validated that lncRNA BANCR was mainly expressed in the cytoplasm of Ect-ESCs (P < 0.05, Figure 2(b)). LncRNA BANCR as a ceRNA can act as a sponge of miRNA [14]. Next, the downstream miRNAs of lncRNA BANCR were predicted on the RNA22 database. Among these, miR-15a-5p is documented to be downregulated in EM [18]. According to the predicted binding sites of lncRNA BANCR to miR-15a-5p (Figure 2(c)), the dual-luciferase assay was performed to testify the binding relationship between lncRNA BANCR and miR-15a-5p (P < 0.05, Figure 2(d)). miR-15a-5p expression levels in ectopic endometrial tissue and Ect-ESCs were lower than those in eutopic endometrial tissue and eutopic endometrial stromal cells (P < 0.05, Figure 2(e)-2(f)). miR-15a-5p was significantly upregulated by silencing lncRNA BANCR (P < 0.05, Figure 2(g)). In addition, lncRNA BANCR was negatively correlated with miR-15a-5p in 20 cases of EM patients (P < 0.05, Figure 2(h)). Altogether, the preceding data suggested that lncRNA BANCR targeted and inhibited miR-15a-5p.
To probe the role of miR-15a-5p in Ect-ESC proliferation and invasion, Ect-ESCs were transfected with miR-15a-5p inhibitor (miR-inhibitor) to downregulate miR-15a-5p expression (P < 0.05, Figure 3(a)), followed by combined treatment with si-BANCR#1 which had better knockdown effectiveness. Through CCK-8 and colony formation assays, it was found that the proliferation potential of Ect-ESCs was enhanced upon miR-15a-5p downregulation (P < 0.05, Figure 3(b)-3(c)). Through transwell assay, it was observed that the invasion potential of Ect-ESCs was also upregulated upon miR-15a-5p downregulation (P < 0.05, Figure 3(d)). Altogether, the preceding data suggested that miR-15a-5p downregulation reversed the effects of silencing lncRNA BANCR on inhibiting Ect-ESC proliferation and invasion.
To further explore the downstream mechanism of miR-15a-5p, the downstream target genes of miR-15a-5p were predicted on the Starbase and miRWalk databases, and our attention was paid to TRIM59 (Figure 4(a)). In a previous study, TRIM59 was elevated in EM [26]. Then, according to the binding sites of miR-15a-5p to TRIM59 predicted on the Starbase database (Figure 4(b)), the dual-luciferase assay was performed to testify the binding relationship between miR-15a-5p and TRIM59 (P < 0.05, Figure 4(c)). Next, we determined the expression levels of TRIM59 in tissues and cells and found that TRIM59 expression levels were augmented in ectopic endometrial tissues and Ect-ESCs (P < 0.05, Figure 4(d)-4(f)), declined by silencing lncRNA BANCR, and upregulated in response to miR-15a-5p downregulation (P < 0.05, Figure 4(g)-4(h)). Besides, TRIM59 mRNA level was positively correlated with lncRNA BANCR (P < 0.05, Figure 4(i)) and negatively correlated with miR-15a-5p (P < 0.05, Figure 4(j)) in 20 cases of EM patients. Altogether, the preceding data suggested that miR-15a-5p targeted and inhibited TRIM59 expression.
To evaluate the role of TRIM59 in Ect-ESC proliferation and invasion, Ect-ESCs were transfected with oe-TRIM59 to upregulate TRIM59 expression (P < 0.05, Figure 5(a)-5(b)), followed by combined treatment with si-BANCR#1 which had better knockdown effectiveness. Our subsequent experiments revealed that TRIM59 overexpression enhanced the proliferation potential of Ect-ESCs (P < 0.05, Figure 5(c)-5(d)), as well as the invasion potential of Ect-ESCs (P < 0.05, Figure 5(e)). Altogether, the preceding data suggested that TRIM59 overexpression reversed the effects of silencing lncRNA BANCR on inhibiting Ect-ESC proliferation and invasion.
EM is a gynecological disorder with malignant potential and remains a challenge in the clinic due to the lack of therapies with curable efficacy [7]. ECSs are identified as the major resident cells in the endometrium and alterations in their behaviors, such as proliferation, invasion, migration, and epithelial-mesenchymal transition, are associated with the pathogenesis of EM [33, 36, 37]. In addition, lncRNAs are involved in gene regulation in EM and the alteration of lncRNA expression level affects EM progression [38]. Apart from lncRNAs, miRNAs and circRNAs also play crucial roles in the regulation of EM progression [39–41]. In the present study, our findings supported that lncRNA BANCR promotes Ect-ESC proliferation and invasion via the miR-15a-5p/TRIM59 axis. LncRNA BANCR is a putative biomarker of cell malignancy [9, 10]. A pioneering study by Zhu and colleagues has uncovered that knockdown of lncRNA BANCR restrains angiogenesis in ectopic endometrial tissues by targeting the extracellular signal-regulated kinase/mitogen-activated protein kinase signaling pathway [11]. In this study, augmented lncRNA BANCR levels were evident in ectopic endometrial tissues of EM patients. Then, ESCs were isolated from eutopic and ectopic endometrial tissues and it was found that lncRNA BANCR was prominently expressed in Ect-ESCs. To further evaluate the functionality of lncRNA BANCR in Ect-ESC proliferation and invasion, lncRNA BANCR was silenced in Ect-ESCs using si-BANCR, upon which the proliferation and invasion potentials of Ect-ESCs were declined. It might be correlated with the effects of lncRNA BANCR on promoting the expression levels of ESCs mobility-related proteins, such as matrix metalloproteinase 1 and 2 [42–44]. Therefore, it is plausible to suggest that silencing lncRNA BANCR suppresses Ect-ESC proliferation and invasion. LncRNAs further function as the sponges of miRNAs to play performance in human diseases [45, 46]. miRNAs can serve as desirable biomarkers for EM diagnosis and treatment due to their stability in body fluids, specific and distinct signatures relative to clinical controls, and biological relevance to EM pathogenesis [47]. To further analyze the downstream mechanism of lncRNA BANCR, the lncLocator database and subcellular fractionation assay validated the cytoplasmic localization of lncRNA BANCR in Ect-ESCs, suggesting the possibility of lncRNA BANCR as a ceRNA in Ect-ESCs. A prior microarray analysis disclosed the downregulation of miR-15a-5p in endometriotic tissues and its association with angiogenesis-related proteins [48]. In addition, miR-15a-5p was previously demonstrated to repress proliferation, migration, and invasion of Ect-ESCs and mobility and angiogenesis of endometrial mesenchymal stem cells [18, 20], further highlighting the participation of miR-15a-5p in EM pathogenesis. Accordingly, the dual-luciferase assay verified the binding relationship between lncRNA BANCR and miR-15a-5p, and decreased miR-15a-5p expression levels were found in ectopic endometrial tissues and Ect-ESCs and elevated in response to silencing lncRNA BANCR, suggesting that lncRNA BANCR negatively regulated miR-15a-5p in EM. Subsequently, miR-15a-5p was downregulated in Ect-ESCs in the si-BANCR group using the miR-15a-5p inhibitor, upon which the proliferation and invasion potentials of Ect-ESCs were enhanced. In accordance, miR-15a-5p was capable of improving pro-apoptosis B-cell lymphoma 2-associated X protein (Bcl2) and decreasing anti-apoptosis B-cell lymphoma 2 (Bax) wherein the ratio of Bax/Bcl2 is associated with apoptosis of ESCs [49, 50]. Collectively, the above findings and shreds of evidence initially demonstrated that lncRNA BANCR promotes Ect-ESC proliferation and invasion by inhibiting miR-15a-5p. Furthermore, the subsequent bioinformatic data directed our attention to TRIM59. As indicated by a previous study, TRIM59 functioned as a positive regulator of Ect-ESC invasion [26]. The binding relationship between miR-15a-5p and TRIM59 was testified by the dual-luciferase assay. Meanwhile, increased TRIM59 expression levels were found in ectopic endometrial tissues and Ect-ESCs, downregulated by silencing lncRNA BANCR and upregulated by inhibition of miR-15a-5p, indicative of a positive correlation between lncRNA BANCR and TRIM59 and a negative correlation between miR-15a-5p and TRIM59. To confirm the effects of TRIM59 on Ect-ESC proliferation and invasion, TRIM59 was overexpressed in Ect-ESCs in the si-BANCR group using oe-TRIM59, upon which Ect-ESC proliferation and invasion were accelerated. Likewise, TRIM59 acts as a booster of proliferation and invasion of endometriosis-associated cancer, such as ovarian, breast, and cervical cancers [51–53]. Altogether, the above findings initially supported that lncRNA BANCR promotes Ect-ESC proliferation and invasion by inducing TRIM59 upregulation.
To conclude, our study was the first of its kind that unveiled the promotive role of the lncRNA BANCR-mediated ceRNA network in Ect-ESC proliferation and invasion and provided a novel theoretical reference for the clinical study of lncRNA BANCR in EM. However, whether other downstream miRNAs of lncRNA BANCR and downstream target genes of miR-15a-5p are involved in functions of Ect-ESCs remains unknown. In the next step, experimentation will be designed to investigate other downstream mechanisms of lncRNA BANCR and miR-15a-5p in Ect-ESCs and the upstream mechanism of lncRNA BANCR in Ect-ESCs. | true | true | true |
||
PMC9576506 | Wenbo Chen,Wuyang Zheng,Shixiao Liu,Qiang Su,Kangxi Ding,Ziguan Zhang,Ping Luo,Yong Zhang,Jianming Xu,Chundong Yu,Weihua Li,Zhengrong Huang | SRC-3 deficiency prevents atherosclerosis development by decreasing endothelial ICAM-1 expression to attenuate macrophage recruitment | 03-10-2022 | SRC-3,endothelial cell,ICAM-1,bufalin,NF-κB signaling | Steroid receptor coactivator 3 (SRC-3) is a member of the p160 SRC family. This factor can interact with multiple nuclear hormone receptors and transcription factors to regulate the expression of their target genes. Although many physiological roles of SRC-3 have been revealed, its role in atherosclerosis is not clear. In this study, we found that SRC-3-/-ApoE-/- mice have reduced atherosclerotic lesions and necrotic areas in their aortas and aortic roots compared with SRC-3+/+ApoE-/- mice after Western diet (WD) feeding for 12 weeks. RNA-Seq and Western blot analyses of the aorta revealed that SRC-3 was required for maintaining the expression of ICAM-1, which was required for macrophage recruitment and atherosclerosis development. siRNA-mediated knockdown of SRC-3 in endothelial cells significantly reduced WD-induced atherosclerotic plaque formation. Additionally, treatment of ApoE-/- mice with SRC-3 inhibitor bufalin prevented atherosclerotic plaque development. SRC-3 deficiency reduced aortic macrophage recruitment. Accordingly, ICAM-1 expression was markedly decreased in the aortas of SRC-3-/-ApoE-/- mice and ApoE-/- mice with endothelial SRC-3 knockdown mediated by AAV9-shSRC-3 virus. Mechanistically, SRC-3 coactivated NF-κB p65 to increase ICAM-1 transcription in endothelial cells. Collectively, these findings demonstrate that inhibiting SRC-3 ameliorates atherosclerosis development, at least in part through suppressing endothelial activation by decreasing endothelial ICAM-1 expression via reducing NF-κB signaling. | SRC-3 deficiency prevents atherosclerosis development by decreasing endothelial ICAM-1 expression to attenuate macrophage recruitment
Steroid receptor coactivator 3 (SRC-3) is a member of the p160 SRC family. This factor can interact with multiple nuclear hormone receptors and transcription factors to regulate the expression of their target genes. Although many physiological roles of SRC-3 have been revealed, its role in atherosclerosis is not clear. In this study, we found that SRC-3-/-ApoE-/- mice have reduced atherosclerotic lesions and necrotic areas in their aortas and aortic roots compared with SRC-3+/+ApoE-/- mice after Western diet (WD) feeding for 12 weeks. RNA-Seq and Western blot analyses of the aorta revealed that SRC-3 was required for maintaining the expression of ICAM-1, which was required for macrophage recruitment and atherosclerosis development. siRNA-mediated knockdown of SRC-3 in endothelial cells significantly reduced WD-induced atherosclerotic plaque formation. Additionally, treatment of ApoE-/- mice with SRC-3 inhibitor bufalin prevented atherosclerotic plaque development. SRC-3 deficiency reduced aortic macrophage recruitment. Accordingly, ICAM-1 expression was markedly decreased in the aortas of SRC-3-/-ApoE-/- mice and ApoE-/- mice with endothelial SRC-3 knockdown mediated by AAV9-shSRC-3 virus. Mechanistically, SRC-3 coactivated NF-κB p65 to increase ICAM-1 transcription in endothelial cells. Collectively, these findings demonstrate that inhibiting SRC-3 ameliorates atherosclerosis development, at least in part through suppressing endothelial activation by decreasing endothelial ICAM-1 expression via reducing NF-κB signaling.
Atherosclerosis is a chronic inflammatory disease of the large arterial walls and is one of the main causes of death worldwide 1. The pathology of atherosclerosis is characterized by the subendothelial accumulation of inflammation, vascular wall cells, extracellular matrix, and apolipoprotein B-containing lipoproteins 2. The entry and retention of apolipoprotein B containing lipoproteins into the intima is a key initiating event in atherosclerosis 3. The activation of endothelial cells can increase the expression of leukocyte adhesion molecules (VCAM-1 and ICAM-1) and chemoattractants (MCP-1), promoting the entry of bone marrow-derived monocytes into the intima, where these cells differentiate into macrophages 4, 5. Macrophage internalizes oxLDL through CD36 and SR-1 to form foam cell, which leads the formation of atheromas and plays a critical role in plaque initiation and development 6, 7. The typical pathology of vulnerable plaques includes large areas of necrotic cores and thinning of the fibrous cap. Plaque rupture of the fibrous cap leads to thrombus formation, which is the primary contributor to cardiovascular events, such as stroke, myocardial infarction, and sudden cardiac death 8. Therefore, understanding the underlying molecular mechanisms can provide information for the development of effective preventive and therapeutic strategies. Steroid receptor coactivator 3 (SRC-3) belongs to the p160 coactivator family, which interacts with multiple nuclear receptors and other transcription factors to enhance their effects on target gene transcription 9. Our previous studies showed that SRC-3 expression in macrophages could protect against LPS-induced endotoxic shock by regulating the proinflammatory cytokine balance 10 and protect against Escherichia coli-induced septic peritonitis by enhancing the phagocytosis of bacteria and reducing macrophage apoptosis 11. Our recent study demonstrated that SRC-3-/- mice displayed delayed clearance of Citobacter rodentium and more severe tissue pathology after oral infection with C. rodentium due to decreased production of the chemokines CXCL2 and CXCL5 and decreased recruitment of neutrophils, indicating that SRC-3 plays a critical protective role in bacteria-induced colitis 12. It has been reported that SRC-3 has an ER-dependent vasoprotective role in vascular trauma by suppressing vascular cell proliferation 13, and deletion of SRC-1/3 causes noncompaction cardiomyopathy in the hearts of newborn and adult mice by inhibiting cardiomyocyte proliferation and differentiation 14, suggesting that SRC-3 has a protective role in cardiovascular; however, the biological role of SRC-3 in atherosclerosis remains poorly defined. In this study, we used SRC-3-/-ApoE-/- mice to assess the role of SRC-3 in atherosclerosis. We found that SRC-3-/-ApoE-/- mice exhibited fewer atherosclerotic lesions and necrotic areas in the aortas and aortic roots than SRC-3+/+ApoE-/- mice after being fed a high-fat diet for 12 weeks. Furthermore, knockdown of SRC-3 in endothelial cells ameliorated atherosclerosis development. SRC-3 accelerated atherosclerosis development, at least in part through increasing endothelial adhesion and the transmigration of leukocytes by promoting ICAM-1 expression in endothelial cells via enhancing NF-κB function.
SRC-3-/- mice on a C57BL/6×129Sv background 15 were bred with the ApoE-/- mice on a C57BL/6J background (a gift from Jiahuai Han, Xiamen University) to generate SRC-3+/-ApoE-/- mice. Male SRC-3+/+ApoE-/- and SRC-3-/-ApoE-/- mice (8 weeks old) were fed a WD for 12 weeks. The WD was produced by Vital River Laboratories (Beijing, China; fed to SRC-3-/-ApoE-/- mice and mice that were injected with AAV-shSRC-3 or AAV-hICAM-1) or Ready Biotechnolgy (Shenzhen, China; fed to mice that were injected with AAV-shICAM-1) according to the protocol from Harlan (TD.88137). Animal experiments were approved by the Laboratory Animal Center of Xiamen University, Xiamen, Fujian, China.
Human lower limb aortic tissue specimens were obtained from 16 patients with arteriosclerosis obliterans who had undergone surgery at Gannan Medical University. Written informed consent was obtained from all individuals, and the study protocol was approved by the Institute Research Ethics Committee at the First Hospital of Xiamen University and Gannan Medical University.
HUVECs (purchased from National Stem Cell Translational Resource Center, mycoplasma negative, DFSC-EC-01) were maintained in EGM2 media (Lonza) supplemented with 10% fetal bovine serum (Gibco) and the EGM-2 bullet kit at 37 °C and 5% CO2. After reaching 60% confluence, the HUVECs were transfected with SRC-3 siRNA (Genepharma, Shanghai) or scrambled siRNA by using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. The cells were cultured for 48 h or 72 h before further experiments.
For en face analysis of atherosclerotic lesions, the whole aorta was removed and stained with Oil Red O, and the lesion areas were quantified with Image-Pro Plus 6.0 (Image Metrology, Copenhagen, Denmark). Hearts were fixed in 4% paraformaldehyde and embedded in paraffin or O.C.T. for histological analysis. Five-micrometer sections of the atrioventricular valve region of the heart were cut and stained with hematoxylin and eosin (H&E) to analyze atherosclerosis lesions or Masson reagents to analyze collagen contents. Atherosclerotic lesions were analyzed in six consecutive sections from 4-6 different littermates in each group using Image-Pro Plus 6.0. The plaque stability score = (α-SMC area + collagen area) / (macrophage area + lipid area) 16.
Five-micrometer consecutive sections of aortic roots were cut, deparaffinized and rehydrated. Antigens were retrieved by soaking in citrate buffer (pH 6.0) with microwave heating for 20 min. The sections were blocked in 10% goat serum for 30 min and incubated with the following primary antibodies overnight: anti-F4/80 (Cell Signaling Technology, D4C8V, #30325) and anti-α-smooth muscle actin (Abcam, E184, ab32575). The sections were incubated with 3% H2O2 for 10 min at the room temperature to eliminate endogenous peroxidase activity and then incubated with Elivision Plus kits (Maixin) at room temperature. DAB reagent was used to visualize the stained proteins. The positive areas were analyzed with Image Pro-Plus 6.0.
Total RNA samples were isolated from the aortas of SRC-3-/-ApoE-/- and SRC-3+/+ApoE-/- mice using TRIZOL reagent (Invitrogen) and treated with RNase-free DNase I. The extracted RNA samples were sent to GENEWIZ for RNA-sequencing analysis. Each sample represented a pooled sample of three representative mice. P<0.05 and fold change>1.5 were defined as the thresholds for significantly differential gene expression. Pathway enrichment analysis and GO analysis were based on the KEGG pathway database and Gene Ontology Database, respectively.
The shuttle plasmids were cotransfected into HEK-293T cells with endothelial cell-enhanced RGDLRVS-AAV9-cap plasmids (a gift from O. J. Müller, Universitat Heidelberg, Germany) and pHelper 17. The AAV viral particles were isolated and purified according to a previously reported protocol 18. For AAV-mediated ICAM-1 overexpression and SRC-3 and ICAM-1 knockdown, 1 × 1010 viral genomes were injected into male 8-week-old ApoE-/- mice via the tail vein before the mice were fed a WD.
For the preservation model, 8-week-old male ApoE-/- mice were intraperitoneally injected with 1.0 mg/kg bufalin 6 times per week (once daily) 19 for 13 weeks and were fed a WD. For the regression model, 8-week-old male ApoE-/- mice were fed a WD for 10 weeks followed by 1.0 mg/kg bufalin administration 6 times per week (once daily) for another 13 weeks while being fed a WD. DMSO containing 0.9% NaCl was used as a vehicle. The mice were killed and dissected. Oil Red O staining was used to determine atherosclerotic plaque formation.
Total RNA was extracted from tissues and cells using TRIzol reagent (Invitrogen). RNA integrity was assessed by absorbance spectroscopy, and samples with an OD260/OD280 of approximately 1.9 were used for experiments. One microgram of total RNA was reverse transcribed using a ReverTra Ace Qpcr RT Master Mix kit (TOYOBO). Real-time PCR was performed using FastStart Universal SYBR Green Master Mix (Rox) (Roche). The relative mRNA level was calculated by normalization to GAPDH. The primer sequences are available upon request.
Total cholesterol and high-density lipoprotein levels were measured by kits (Nanjing Jiancheng) according to the manufacturer's instructions. Blood glucose levels were measured by a glucometer (Roche).
HUVECs and the whole aorta were lysed in RAPA buffer (150 mM NaCl, 50 mM Tris, 0.1% SDS, 1 mM EDTA, 1% Triton X-100, 1 mM PMSF and protease inhibitors). Proteins were quantified with a BCA assay. Equal amounts of proteins were loaded onto 8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels and transferred onto polyvinylidene difluoride membranes (Millipore), followed by immunoblotting with anti-SRC-3 (Cell Signaling Technology, 5E11, #2116), anti-human ICAM-1 (Abcam, EPR4776, ab109361), anti-mouse ICAM-1 (Abcam, EPR16608, ab179707), anti-p65 (Cell Signaling Technology, D14E12, #8284), anti-p-p65 (Cell Signaling Technology, Ser536, #3031), anti-GAPDH (Cell Signaling Technology, D16H11, #5174) and anti-β-actin (Sigma, AC-15, #A5441). Western blots were analyzed using a Tonen Image System.
Full-length human ICAM-1 cDNA was amplified by polymerase chain reaction (PCR) from HUVECs with the following primer sets: forward, ICAM-1 5'-CCGGAATTCATGGCTCCCAGCAGCCC-3'; reverse, ICAM-1 5'-TGCTCTAGATCAGGGAGGCGTGGCTTGTG-3' and then inserted into the pCR3.1-HA vector at the EcoRI and XbaI sites to yield the human ICAM-1 expression plasmid. The human ICAM-1 promoter 1007-bp fragment was amplified by PCR using mouse genomic DNA as a template with the following primer sets: forward, ICAM-1 (-1092) 5'-CGGGGTACCCTTAAGAGTACCCAGCCTCGAC-3'; reverse, ICAM-1 (-88) 5'-CGACGCGTCCCCTCCGGAACAAATGCT-3'. The fragment was ligated into the KpnI and MluI sites of the luciferase reporter vector PGL3-basic to produce the ICAM-1 promoter reporter plasmid. To generate site mutations in NF-κB, the transcription factor recognition site was abrogated by a one-step site-directed and site-saturation mutagenesis method 20. The primers for site-directed mutation were yielded as follows: mNF-κB, forward, 5'-TaagAATTCCGGAGCTGAAGCGGCCAGCG-3', reverse, 5'-TCAGCTCCGGAATTcttAAGCTAAAGCAAT-3'. Lowercase letters indicate mutated sites. DNA sequencing was used to verify the nucleotide sequences of these constructs. Luciferase activity was examined by a Dual Luciferase Reporter Assay System (Promega, Madison, WI). Renilla luciferase activity was used to normalize the transfection efficiency.
Wild-type HUVECs or transient SRC-3-knockdown HUVECs were used for chromatin immunoprecipitation (ChIP) assays, which were performed according to the method described by Abcam (Cambrige, MA). The following primers were used: ICAM-1 promoter NF-κB binding site, forward, 5'-GTCATCGCCCTGCCACC-3' and reverse, 5'-ATTTCCGGACTGACAGGGTG-3'. Anti-SRC-3 antibodies (C-20, sc-7216) and anti-p65 (D14E12, #8284) antibodies were purchased from Santa Cruz Biotechnology and Cell Signaling Technology, respectively.
For the monocyte-endothelial adhesion assay, confluent HUVECs were transfected with scramble or SRC-3 siRNA for 48 h and then the scramble or SRC-3 siRNA-transfected HUVECs were stimulated with 10 ng/ml TNF-α or 2 ng/ml IL-1β for 3 h or 6 h. After 3 h or 6 h, Calcein (Yeasen)-labeled THP-1 monocytes were cocultured with the stimulated HUVECs for 30 min followed by two rounds of gentle washing to remove nonadherent THP-1 cells, and the cells were fixed and imaged with fluorescence microscopy at 488 nm. At least five random fields were imaged to quantify adherent THP-1 cells per condition. For the transendothelial migration assay, confluent HUVECs were transfected with scramble or SRC-3 siRNA for 24 h, and then the scramble or SRC-3 siRNA-transfected HUVECs were seeded onto 6.5 mm Transwell filters coated with 0.1% gelatin (Sigma-Aldrich) for another 24 h. The HUVECs were stimulated with 10 ng/ml TNF-α or 2 ng/ml IL-1β for 6 h followed by the addition of calcein-labeled THP-1. After being incubated for 24 h, the transmigrated THP-1 cells were imaged and quantified.
Data were collected from at least two independent experiments. All data are expressed as the mean ± SEM. Statistical significance was determined by unpaired two-tailed Student's t test. P<0.05 was considered statistically significant.
To determine whether SRC-3 is involved in the development of atherosclerosis, we measured SRC-3 expression in human atherosclerotic plaques. We collected atherosclerotic plaques and plaque-adjacent vasculature in the lower limb aorta and assayed SRC-3 by Western blotting. As shown in Figure 1A, SRC-3 protein expression in the atherosclerotic plaques in lower limb aortas was significantly higher than that in the plaque-adjacent vasculature. To further determine whether SRC-3 was increased in mouse models of atherosclerosis, whole aortas were isolated from 12-week chow-fed and WD-fed male ApoE-/- mice for Western blotting. The results showed that SRC-3 protein expression was significantly increased in the aortas of WD-fed ApoE-/- mice compared with chow-fed ApoE-/- mice (Figure 1B). These results suggest that SRC-3 may be involved in atherosclerosis development. To assess whether SRC-3 deficiency affects the development of atherosclerosis, SRC-3-/-ApoE-/- mice were generated by crossing SRC-3+/- mice with ApoE-/- mice. Since a LacZ reporter was inserted into the 10th amino acid of SRC-3, which is controlled by the endogenous SRC-3 promoter in SRC-3-/-ApoE-/- mice 15, we performed an X-gal staining assay to determine the pattern of SRC-3 expression in the aortic roots of SRC-3-/-ApoE-/- mice. As shown in Figure 1C, SRC-3 expression (blue) was observed in endothelial cells (ECs) and vascular smooth muscle cells (VSMCs) in the aortic roots of chow-fed SRC-3-/-ApoE-/- mice, and the signals in the aortic roots of WD-fed SRC-3-/-ApoE-/- mice were stronger. After eight-week-old male SRC-3-/-ApoE-/- and SRC-3+/+ApoE-/- mice were fed a WD for 12 weeks, the Oil Red O-stained areas in the aortas of SRC-3-/-ApoE-/- mice were reduced by 60% compared with those in SRC-3+/+ApoE-/- mice (6.68% versus 2.43%; Figure 1D). Next, we analyzed the advanced atherosclerotic lesions in the aortic roots by H&E staining. Quantitative analysis of representative images showed that the lesion areas were much smaller in SRC-3-/-ApoE-/- mice than in SRC-3+/+ApoE-/- mice (Figure 1E). We also analyzed the necrotic cores of plaques, which are critical determinants of plaque vulnerability. Quantification of this index revealed a pronounced decrease in the total necrotic area in the lesions of SRC-3-/-ApoE-/- mice (Figure 1F). Since body weight, plasma lipid levels, and glucose levels are associated with atherosclerosis development, we measured body weight, plasma lipid levels, and glucose levels in SRC-3-/-ApoE-/- and SRC-3+/+ApoE-/- mice after the mice were fed a WD for 12 weeks. Decreased plasma lipid levels and glucose levels were observed in SRC-3-/-ApoE-/- mice after being fed a WD for 12 weeks (Supplementary Figure 1A). Collectively, these results suggest that SRC-3 deficiency attenuates the development of atherosclerosis. To further examine the effect of SRC-3 deficiency on the stability of atherosclerotic plaques, we investigated the properties that affect plaque stability, including α-SMA content, collagen content, fibrous cap thickness, lipid accumulation, and macrophage numbers. The α-SMA content of aortic root lesions in SRC-3-/-ApoE-/- mice was decreased compared with that in SRC-3+/+ApoE-/- mice (Figure 1G). However, the collagen content of aortic root lesions was significantly increased in SRC-3-/-ApoE-/- mice compared with SRC-3+/+ApoE-/- mice (Figure 1H). Fibrous caps in the aortic root lesions of SRC-3-/-ApoE-/- mice were thicker than those of SRC-3+/+ApoE-/- mice (Figure 1I). Lipid staining with Oil Red O revealed a robust 36% decrease in lipid accumulation in the aortic root lesions of SRC-3-/-ApoE-/- mice compared with SRC-3+/+ApoE-/- mice (Figure 1J). Immunohistological staining of macrophages (F4/80-positive), which can form foam cells that are critical for atherogenesis, was performed, and there was greatly reduced staining of F4/80-positive cells in the aortic root lesions of SRC-3-/-ApoE-/- mice compared with SRC-3+/+ApoE-/- mice (Figure 1K). In summary, aortic plaque stability was significantly increased in SRC-3-/-ApoE-/- mice compared with SRC-3+/+ApoE-/- mice (Figure 1L). These results suggest that SRC-3 reduces the stability of atherosclerotic plaques in the aortic root.
Our results demonstrated that SRC-3 was highly expressed in endothelial cells (Figure 1B). Therefore, we hypothesized that endothelial SRC-3 may play an important role in promoting the development of atherosclerosis. It has been reported that the transduction efficiency of the RGDLRVS-AAV9-cap plasmid in endothelial cells is significantly higher than both wild-type AAV2 and AAV9 in vitro 17, and the RGDLRVS-AAV9-cap plasmid has been used for endothelial-specific gene expression or knockdown in mouse models 18, 21. To downregulate SRC-3 expression in endothelial cells in vivo, we infected ApoE-/- mice with endothelial-specific RGDKRVS-AAV9-mediated SRC-3 short hairpin RNA (shSRC-3) and negative control short hairpin RNA (shCtrl) by tail vein injection before feeding the mice a WD. The aortas of ApoE-/- mice injected with AAV9-mediated shSRC-3 exhibited significantly reduced SRC-3 protein expression after the mice were fed a WD (Figure 2A). There were comparable plasma cholesterol levels and glucose levels between the AAV-shSRC-3 group and the AAV-shCtrl group after WD feeding (Supplementary Figure 1B). En face analysis of Oil Red O-stained atherosclerotic lesion area showed an approximate 2-fold decrease in the AAV-shSRC-3 group compared with the AAV-shCtrl group after WD feeding (Figure 2B). Staining of smooth muscle actin, collagen, lipids, and the macrophage marker F4/80 in the aortic roots in the AAV-shSRC-3 group and AAV-shCtrl group was performed to assess plaque composition. Consistent with the en face results, enhanced staining for collagen but reduced staining for lipids and the macrophage marker F4/80 were observed in AAV-shSRC-3 aortic roots (Figure 2C-F). These results demonstrate that SRC-3 in endothelial cells plays an important role in promoting the development of atherosclerosis.
To examine the mechanism of SRC-3-mediated promotion of atherosclerosis, we performed messenger RNA (mRNA) sequencing in the aortas of WD-fed SRC-3-/-ApoE-/- and SRC-3+/+ApoE-/- mice. KEGG enrichment pathway analysis revealed eight enriched pathways (Figure 3A and E), including the NF-kappa B signaling pathway, cell adhesion molecules (CAMs), and ECM-receptor interaction. Gene Ontology (GO) analysis of enriched biological processes indicated that SRC-3 was associated with cell adhesion and the inflammatory response (Figure 3B). Because the aortas of WD-fed SRC-3-/-ApoE-/- mice had reduced macrophage recruitment (Figure 1K), we focused on the expression of cell adhesion molecules, CC and CXC chemokines and their receptors, which are responsible for leukocyte recruitment (Figure 3C). The RNA-Seq results revealed decreased expression of CCL3, CD68, ICAM-1, and VCAM-1 in the aortas of WD-fed SRC-3-/-ApoE-/- mice (Figure 3C); however, real-time quantitative PCR showed that only ICAM-1 expression was significantly reduced in the aortas of WD-fed SRC-3-/-ApoE-/- mice (Figure 3D). ICAM-1 is involved in the progression of atherosclerosis in ApoE-/- mice 22, 23, 24, 25, 26. Therefore, ICAM-1 was selected for further investigation. Western blot analysis showed that the protein levels of ICAM-1 were significantly decreased in the aortas of WD-fed SRC-3-/-ApoE-/- mice compared with SRC-3+/+ApoE-/- mice (Figure 3E), suggesting that SRC-3 upregulates ICAM-1 expression. Consistently, the protein levels of SRC-3 were positively correlated with the protein levels of ICAM-1 in human atherosclerotic lesions (Figure 3E and F). Taken together, these results suggest that SRC-3 promotes ICAM-1 expression during atherosclerosis development.
Since both the mRNA and protein levels of ICAM-1 were significantly decreased in the aortas of WD-fed SRC-3-/-ApoE-/- mice compared with SRC-3+/+ApoE-/- mice, SRC-3 may regulate ICAM-1 expression at the transcriptional level. To determine whether SRC-3 can regulate ICAM-1 expression in endothelial cells, we transfected SRC-3-specific small interfering RNA (siSRC-3) into human umbilical vein endothelial cells (HUVECs) to knock down SRC-3 and examined ICAM-1 expression. As shown in Figure 4A and B, the protein and mRNA levels of ICAM-1 in siSRC-3-transfected HUVECs were significantly reduced compared with scrambled siRNA (siCtrl)-transfected HUVECs before and after TNFα or IL-1β treatment. Restoration of SRC-3 expression in SRC-3-knockdown HUVECs could rescue ICAM-1 expression after TNFα or IL-1β treatment (Supplementary Figure 2A and B). Furthermore, we examined whether SRC-3 could promote ICAM-1 transcription by transfecting the ICAM-1 promoter-reporter plasmid into wild-type and SRC-3-knockdown HUVECs. As shown in Figure 4C, TNFα or IL-1β treatment markedly induced ICAM-1 promoter activity in scramble HUVECs, whereas SRC-3 knockdown significantly reduced the induction of ICAM-1 promoter activity. These results suggest that SRC-3 promotes ICAM-1 expression at the transcriptional level. It has been reported that ICAM-1 expression can be induced by NF-κB 27. Because TNFα or IL-1β treatment can induce NF-κB activation (Figure 4A) and SRC-3 can interact with p65 to enhance p65-mediated NF-κB reporter activation (Figure 4D, upper panel), we hypothesized that SRC-3 may enhance NF-κB-mediated ICAM-1 expression. We tested this hypothesis by transfecting HUVECs with an ICAM-1 promoter-reporter plasmid with SRC-3, p65, or both SRC-3 plus p65 expression plasmids. Transfection of p65 or SRC-3 alone induced a 10-fold or 1.3-fold increase in ICAM-1 promoter activity (Figure 4D, lower panel), respectively, whereas cotransfection of p65 and SRC-3 induced a 13-fold increase in ICAM-1 promoter activity (Figure 4D, lower panel). These results suggest that SRC-3 can enhance NF-κB-mediated ICAM-1 expression. To further confirm the role of p65 in NF-κB-mediated activation of the ICAM-1 promoter, we mutated the NF-κB binding site on the ICAM-1 promoter. As shown in Figure 4E, cotransfection of p65 and SRC-3 induced the activity of the wild-type ICAM-1 promoter by approximately 15-fold, while mutating the NF-κB binding site dramatically abolished NF-κB-mediated ICAM-1 promoter activity, suggesting that the NF-κB binding site is essential for NF-κB-mediated activation of the ICAM-1 promoter. We next determined whether SRC-3 and p65 could be recruited to the NF-κB binding site on the promoter after TNFα or IL-1β treatment. As shown in Figure 4F, SRC-3 and p65 bound to the NF-κB binding site on the ICAM-1 promoter after TNFα or IL-1β treatment, whereas SRC-3 knockdown in HUVECs markedly reduced the recruitment of p65 and SRC-3 to the ICAM-1 promoter, suggesting that SRC-3 and p65 are recruited to the NF-κB binding site to enhance NF-κB-mediated activation of the ICAM-1 promoter after TNFα or IL-1β treatment. In response to endothelial cell activation, various leukocytes are recruited, which is a basic event in vascular inflammation. The processes that participate in inflammatory cell recruitment include leukocyte-endothelial adhesion and transmigration. To determine the role of endothelial SRC-3 in vascular inflammation, we performed proinflammatory cytokine-mediated monocyte-endothelial cell adhesion and transmigration assays. As shown in Figure 4G, the siSRC-3-transfected HUVEC monolayer exhibited a marked decrease in the number of adhered and transmigrated THP-1 cells compared with the siCtrl-transfected HUVEC monolayer after TNFα or IL-1β treatment. We next determined whether the restoration of ICAM-1 expression in SRC-3-knockdown HUVECs could rescue the number of adhered and transmigrated THP-1 cells after TNFα or IL-1β treatment. As shown in Figure 4H and I, ectopic expression of ICAM-1 in SRC-3-knockdown HUVECs rescued the number of adhered and transmigrated THP-1 cells after TNFα or IL-1β treatment. These results suggest that SRC-3 promotes the adhesion and transmigration of THP-1 cells at least in part by increasing ICAM-1 expression.
To verify that ICAM-1 in endothelial cells promotes atherosclerotic plaque formation in mice, we infected ApoE-/- mice with endothelial-specific RGDKRVS-AAV9-ICAM-1 or scramble virus by tail vein injection before feeding the mice a WD. Mice that were injected with endothelial-specific AAV9-ICAM-1 virus showed significantly higher ICAM-1 expression (Figure 5A) and markedly increased plaque formation (Figure 5B) and lesion area (Figure 5C) than mice that were injected with the scramble control virus. Furthermore, staining for lipids and the macrophage marker F4/80 in the aortic roots in the AAV9-scramble control group and AAV9-ICAM-1 expression group was performed to assess plaque composition. Consistent with the en face results, increased staining for lipids and the macrophage marker F4/80 was observed in AAV9-hICAM-1-infected aortic roots (Figure 5C and E). There was no difference in plasma lipid content levels and plasma glucose levels between the AAV9-scramble and AAV9-hICAM-1 groups (Supplementary Figure 3A). Conversely, we knocked down mouse ICAM-1 by using an endothelial-specific AAV9-shICAM-1 virus. Plasma lipid levels and plasma glucose levels were comparable between the AAV-scramble control and AAV-shICAM-1 virus-infected groups (Supplementary Figure 3B). Mice that were injected with the endothelial-specific AAV9-shICAM-1 virus exhibited significantly decreased atherosclerotic plaque formation and lesion areas (Supplementary Figure 4A and B). Consistent with the en face results, reduced staining for lipids and the macrophage marker F4/80 was observed in AAV9-shICAM-1 virus-infected aortic roots (Supplementary Figure 4C-E). These results demonstrate that ICAM-1 in endothelial cells is required for the development of atherosclerosis.
Since SRC-3 deficiency prevents atherosclerosis development, we assessed the protective effect of bufalin, a small-molecule inhibitor of SRC-3 that has been used in tumor therapy 28, 29, in the development of atherosclerosis. We first performed proinflammatory cytokine-mediated monocyte-endothelial cell adhesion and transmigration assays to determine whether bufalin alleviates macrophage recruitment in vitro. As shown in Figure 6A, the expression of SRC-3 and ICAM-1 in HUVECs was significantly reduced by bufalin treatment. Additionally, bufalin treatment significantly decreased the expression of p-p65 (Figure 6A). Consequently, a significant reduction in adhered and transmigrated THP cells was observed in the presence of bufalin-treated HUVECs in response to TNFα or IL-1β treatment (Figure 6B). We then determined the protective effect of bufalin on atherosclerosis by using ApoE-/- prevention model. Bufalin was administered to ApoE-/- mice (1.0 mg/kg, six times a week) by intraperitoneal injection for 13 weeks during WD feeding (Figure 6C). Bufalin-treated mice exhibited decreased atherosclerotic plaque formation compared with those treated with vehicle only (Figure 6D). Furthermore, the protein levels of SRC-3 and ICAM-1 in the aortas from ApoE-/- mice treated with bufalin were also significantly reduced after the mice were fed a WD for 13 weeks (Figure 6E). Reduced p-p65 was observed in the aortas of mice treated with bufalin after WD feeding for 13 weeks (Figure 6F). Bufalin treatment did not significantly affect plasma lipid or glucose levels (Supplementary Figure 5A). It is well known that atherosclerosis is an advanced disease, and so we performed a regression model by treating ApoE-/- mice that were fed WD alone for 10 weeks with bufalin for 13 weeks during WD feeding (Figure 6F). Bufalin treatment further diminished plaque development (Figure 6G). The levels of SRC-3, ICAM-1, and p-p65 in the aortas of bufalin-treated mice were also reduced compared with those of vehicle-treated mice (Figure 6H). Bufalin treatment significantly increased TC and HDL-C, but decreased glucose levels (Supplementary Figure 5B). These results demonstrate that the pharmacological SRC-3 inhibitor bufalin can ameliorate the development of atherosclerosis.
Lipid-mediated inflammation of the vessel wall is a key initiating event in atherosclerosis that activates endothelial cells to recruit macrophages in a leukocyte adhesion molecule-dependent manner 30. In this study, we found that global SRC-3 deficiency or endothelial SRC-3 knockdown on an ApoE-/- background ameliorated the development of atherosclerosis, suggesting that SRC-3 could promote atherosclerosis development. Mechanistically, SRC-3 promoted atherosclerosis development by increasing ICAM-1 expression at the transcriptional level by enhancement of NF-κB function in endothelial cells to promote macrophage recruitment (Figure 7). SRC-3 depletion or pharmacological inhibition of SRC-3 by bufalin ameliorated atherosclerosis development, at least in part by decreasing endothelial ICAM-1 expression via reduction of NF-κB function (Figure 7). In this study, we showed that the genetic ablation of SRC-3 prevented the development of atherosclerosis in ApoE-/- mice, indicating that SRC-3 injures the vasculature during atherosclerosis. A previous study showed that SRC-3 was required for estrogen-induced inhibition of neointima formation during vascular trauma, which may be caused by suppressing of vascular cell proliferation 13. Recently, it has been reported that SRC-1/3 deficiency can cause noncompaction cardiomyopathy-like abnormalities in the hearts of newborn and adult mice, which are caused by the inhibition of cardiomyocyte proliferation and differentiation 14, suggesting that SRC-3 has a vasoprotective role. These studies demonstrate that SRC-3 has injures or protects the vasculature in different cardiovascular diseases through diverse mechanisms. Advanced atherosclerotic lesions, which are essentially a nonresolving inflammatory condition, can cause the formation of the vulnerable plaques, increasing the risk of plaque rupture. Vulnerable plaques are characterized by the formation of necrotic cores and thinning of the fibrous cap. The necrotic core is the result of increased macrophage death and impaired efferocytosis 8. H&E staining of the aortic roots showed that the total necrotic areas in the aortic root lesions of SRC-/-ApoE-/- mice were smaller than those in SRC-3+/+ApoE-/- mice, and fibrous caps in the aortic root lesions of SRC-3-/-ApoE-/- mice were thicker than those of SRC-3+/+ApoE-/- mice. In human atherosclerotic-associated cardiovascular events, plaque morphology is a more critical predictor of plaque rupture than plaque size, and the size of the necrotic core plays a primary role in plaque vulnerability 16. Other features, such as high levels of inflammatory cytokines and significant accumulation of lipids also contribute to the vulnerability of atherosclerotic plaques 31. In this study, analysis of atherosclerotic plaques revealed that SRC-3-/-ApoE-/- mice exhibited notably decreased infiltration of macrophages, decreases in α-SMA levels and lipid accumulation and an increase in collagen content. Advanced necrotic cores and other deteriorated features of vulnerable plaques are responsible for the majority of clinical events. Our results showed that global SRC-3 deficiency on the ApoE-/- background ameliorated the development of atherosclerosis. Previous studies have shown that RGDLRVS-AAV9-cap plasmid-based virus preferentially mediates genes expression in endothelial cells in vitro and in vivo 17, 18, 21. Therefore, ApoE-/- mice were infected with endothelial-specific RGDKRVS-AAV9-shSRC-3 or control shRNA virus by tail vein injection before being fed a WD, and a model of atherosclerosis was constructed. We found that mice on the ApoE-/- background that were infected with endothelial-specific AAV9-shSRC-3 also exhibited significant reductions in atherosclerotic plaques, similar to the effects of global SRC-3 deficiency on the ApoE-/- background, suggesting that endothelial SRC-3 contributes to atherosclerosis development. Additionally, SRC-3-/-ApoE-/- mice exhibited markedly decreased levels of plasma lipids and glucose compared with SRC-3+/+ApoE-/- mice after being fed a WD for 12 weeks. However, mice on the ApoE-/- background that were infected with endothelial-specific AAV9-shSRC-3 virus did not exhibit significantly altered levels of plasma lipids or glucose. Therefore, the decrease in plasma lipid and glucose levels due to global SRC-3 deficiency on the ApoE-/- background could contribute to the reduction in atherosclerotic plaque formation in a manner that is independent of the endothelial cells. ICAM-1 belongs to the Ig superfamily and is composed of a transmembrane domain and a short cytoplasmic domain 32. This factor is expressed by endothelial cells, epithelial cells, all leukocyte subsets 33, platelets 34, and vascular smooth muscle cells 35. ICAM-1 can mediate leukocyte rolling and firm adhesion to the vessel wall and leukocyte transmigration across the endothelial layer 36, promoting vascular inflammation. It has been reported that ICAM-1 deficiency or inhibition protects against the development of atherosclerosis in ApoE-/- mice 22, 23, 24, 25, 26. In this study, we found that ICAM-1 overexpression in the vascular endothelial cells of mice infected with AAV9-hICAM-1 accelerated atherosclerotic plaque formation after the mice were fed a WD for 12 weeks, whereas ICAM-1 knockdown in the vascular endothelial cells of mice infected with AAV9-shICAM-1 virus alleviated atherosclerotic plaque formation after being fed WD for 12 weeks. Furthermore, ectopic ICAM-1 expression in SRC-3-knockdown HUVECs increased the number of adhered and transmigrated THP-1 cells compared with that in SRC-3-knockdown HUVECs. These results suggest that endothelial ICAM-1 plays an import role in promoting the development of atherosclerosis, at least in part through increasing leukocyte adhesion and transmigration and is sufficient to promote atherosclerosis development. Bufalin is a major bioactive monomer of the traditional Chinese medicine Chan Su, which is acquired from the skin and parotid venom glands of the Chinese toad 37, 38. It has been reported that bufalin is a small-molecule inhibitor of SRC-3 that can directly bind to the receptor interacting domain of the SRC-3 protein to promote its degradation 28. In this study, bufalin treatment of ApoE-/- mice ameliorated atherosclerotic plaque formation in an ApoE-/- prevention model and regression model, suggesting that bufalin plays an important protective role against atherosclerosis. Furthermore, the protein levels of SRC-3 and ICAM-1 were reduced in the aortas of bufalin-treated ApoE-/- mice after the mice were fed a WD for 13 weeks. We found that bufalin suppressed the protein expression of p-p65 and p65 in the aortas of bufalin-treated ApoE-/- mice after the mice were fed a WD for 13 weeks. Additionally, bufalin-treated HUVECs exhibited significant reductions in adhered and transmigrated monocytes in response to TNFα or IL-1β, and bufalin treatment reduced the protein levels of SRC-3, ICAM-1, and p-p65 in HUVECs, similar to the effect on bufalin-treated ApoE-/- mice, suggesting that endothelial cells are targets of bufalin. These results demonstrate that SRC-3 pathway inhibited by bufalin was involved in atherosclerosis development, at least in part through reducing ICAM-1 expression via reducing NF-κB signaling. Recent studies have shown that bufalin exerts anticancer activity on various cancers 29. Furthermore, Chan Su has been widely used in the clinical treatment of malignancies in China and exhibits good therapeutic efficacy 39, 40. Thus, the SRC-3 inhibitor bufalin may not only represent a new drug for the prevention and treatment of atherosclerosis, but may also be safe for SRC-3-targeted cancer therapies at the experimental and clinical levels. NF-κB can regulate endothelial ICAM-1 expression at the transcriptional level 41, 42. Furthermore, endothelial NF-κB deficiency decreased the expression of leukocyte adhesion molecules, attenuated macrophage recruitment to atherosclerotic plaques and decreased the expression of proinflammatory cytokines in the aorta 43, 44, suggesting that endothelial NF-κB is required for promoting atherosclerotic plaque formation. The present study showed that NF-κB signaling was inhibited in aortas of SRC-3-/-ApoE-/- mice after WD feeding for 12 weeks or SRC-3-knockdown HUVECs after TNFα and IL-1β treatment, suggesting that SRC-3 can activate NF-κB. Additionally, it has been reported that SRC-3 can serve as a coactivator for p65 to enhance its transcriptional activity 12, 45, 46. Our results showed that SRC-3 could be recruited to the p65 binding site on the ICAM-1 reporter/promoter and that SRC-3 could cooperate with p65 to enhance ICAM-1 transcription, suggesting that SRC-3 coactivates NF-κB to upregulate ICAM-1 transcription. Although we found a consistent transcription factor AP-1 binding site (TGACTCGCA) at the ICAM-1 promoter, our results showed that SRC-3 could not cooperate with AP-1 to induce ICAM-1 promoter activity (Supplementary Figure 6). Certainly, we could not exclude that SRC-3 could coactivate other key transcription factors in the endothelial cells to affect ICAM-1 expression. Given that macrophage foam cell formation is critical for the atherosclerosis development, we investigated the effect of SRC-3 on macrophage foam cell formation demonstrated by affecting oxLDL uptake. As shown in Supplementary Figure 7B and C, oxLDL uptake was significantly reduced in two SRC-3-knockdown stable RAW264.7 cell lines compared with control RAW264.7 cell line (shCtrl), suggesting that SRC-3 knockdown reduces macrophage foam cell formation. To determine whether bufalin treatment affects SRC-3 function in macrophages, we first assessed the effect of bufalin treatment on SRC-3 expression in macrophages. As shown in Supplementary Figure 7D, bufalin treatment didn't decreased SRC-3 expression in macrophages, whereas bufalin treatment could significantly reduced SRC-3 expression in HUVECs. It suggests that the effect of bufalin on SRC-3 expression may be cell context-dependent. In summary, our study shows that endothelial SRC-3 deficiency or pharmacological inhibition prevents atherosclerosis development by decreasing endothelial ICAM-1 expression, resulting in the inhibition of macrophage recruitment. Thus, SRC-3 is a potential target for atherosclerosis prevention and therapy.
Click here for additional data file. | true | true | false |
|
PMC9576509 | Yu-Ming Wang,Qi-Wu Zhao,Zhi-Yong Sun,Hai-Ping Lin,Xin Xu,Min Cao,Yu-Jie Fu,Xiao-Jing Zhao,Xiu-Mei Ma,Qing Ye | Circular RNA hsa_circ_0003823 promotes the Tumor Progression, Metastasis and Apatinib Resistance of Esophageal Squamous Cell Carcinoma by miR-607/CRISP3 Axis | 21-09-2022 | hsa_circ_0003823,miR-607,crisp3,apatinib,esophageal squamous cell carcinoma | Background: Circular RNAs (CircRNAs) have attracted a growing interest of research in cancer. The regulatory roles and mechanisms of circRNAs in progression, metastasis and drug resistance of esophageal squamous cell carcinoma (ESCC) needed to be clarified. Our previous study revealed the crucial role of Apatinib in ESCC therapy. However, the correlation between circRNAs and Apatinib resistance remained unclear. Methods: 3 pairs of tumor and paracancerous tissues of ESCC patients were used for RNA sequencing. Western blot analysis, RNA immunoprecipitation (RIP), dual-luciferase reporter assays, apoptosis and animal assays were conducted to confirm the roles and specific mechanisms of hsa_circ_0003823 as well as the effects of it on Apatinib sensitivity in ESCC. Results: Our results revealed that hsa_circ_0003823 was highly expressed in ESCC and associated with poor prognosis. Further results indicated that hsa_circ_0003823 promoted proliferation and metastasis ability of ESCC. In the section of mechanism experiments, hsa_circ_0003823 regulated CRISP3 by targeting microRNA-607 (miR-607) to promote progression of ESCC. Besides, we found that silencing hsa_circ_0003823 improved Apatinib sensitivity. hsa_circ_0003823 resulted in Apatinib resistance by miR-607/CRISP3 axis. Conclusions: In this study, we elucidated the function of hsa_circ_0003823 and its role in promoting tumor progression, metastasis and Apatinib resistance of ESCC through miR-607/CRISP3 axis. | Circular RNA hsa_circ_0003823 promotes the Tumor Progression, Metastasis and Apatinib Resistance of Esophageal Squamous Cell Carcinoma by miR-607/CRISP3 Axis
Background: Circular RNAs (CircRNAs) have attracted a growing interest of research in cancer. The regulatory roles and mechanisms of circRNAs in progression, metastasis and drug resistance of esophageal squamous cell carcinoma (ESCC) needed to be clarified. Our previous study revealed the crucial role of Apatinib in ESCC therapy. However, the correlation between circRNAs and Apatinib resistance remained unclear. Methods: 3 pairs of tumor and paracancerous tissues of ESCC patients were used for RNA sequencing. Western blot analysis, RNA immunoprecipitation (RIP), dual-luciferase reporter assays, apoptosis and animal assays were conducted to confirm the roles and specific mechanisms of hsa_circ_0003823 as well as the effects of it on Apatinib sensitivity in ESCC. Results: Our results revealed that hsa_circ_0003823 was highly expressed in ESCC and associated with poor prognosis. Further results indicated that hsa_circ_0003823 promoted proliferation and metastasis ability of ESCC. In the section of mechanism experiments, hsa_circ_0003823 regulated CRISP3 by targeting microRNA-607 (miR-607) to promote progression of ESCC. Besides, we found that silencing hsa_circ_0003823 improved Apatinib sensitivity. hsa_circ_0003823 resulted in Apatinib resistance by miR-607/CRISP3 axis. Conclusions: In this study, we elucidated the function of hsa_circ_0003823 and its role in promoting tumor progression, metastasis and Apatinib resistance of ESCC through miR-607/CRISP3 axis.
Esophageal cancer (EC) is one of the most prevalent malignant tumor worldwide, especially in China 1, 2. Esophageal squamous cell cancer is the major histopathological subtype of esophageal cancer and accounts for 90% of EC patients 3. Although significant advances have been achieved in the diagnosis and treatment strategies, the 5-year survival rates of ESCC patients with advanced stage remain poor due to early recurrence and metastasis 4, 5. For advanced ESCC patients, targeted therapy might provide a good choice for comprehensive treatment of ESCC and be an important supplement to chemotherapy 6-8. Our previous study focused on Apatinib, a novel vascular endothelial growth factor receptor-2 (VEGFR-2) tyrosine kinase inhibitor, and elucidated its crucial role in inhibiting progression and metastasis as well as sensitizing paclitaxel in ESCC. However, drug resistance was still the problem we needed to confront and tackle 9. Therefore, it has become extremely important to clarify the specific molecular mechanisms of progression, metastasis and Apatinib resistance for ESCC and discover effective therapeutic targets. Circular RNAs (CircRNAs) are a novel and large class of non-coding RNAs with a covalently closed loop structure that are produced by backsplicing. Most circRNAs belong to non-coding RNAs (ncRNAs) without 5′ to 3′ polarity or 3′ polyA tail and are formed by a single exon or multiple exons of protein-coding genes 10. CircRNAs are mainly located in the cytoplasm, and not easily degraded by the exonuclease RNase R 11. Recently, many studies have revealed that circRNAs were involved in the occurrence, development and drug resistance in cancer 12-17. CircRNAs have many important non-coding functions, mainly including acting as miRNA sponges, interacting with miRNAs, regulating gene transcription, and interacting with RNA-binding protein to regulate other RNAs 18-20. In addition, a subset of circRNAs performs independent translation functions under certain conditions, although the vast majority of circRNAs are considered non-coding 21, 22. MicroRNAs are short non-coding RNAs with 19-25 nucleotides that are able to inhibit translation of mRNAs and participate in multiple cellular processes, including cell cycle, proliferation, apoptosis, invasion and migration 23. It has been reported that miRNAs are involved in various cellular signaling pathways and their dysfunction could lead to occurrence and progression of cancer 24. miR-607 has been reported as a tumor suppressor. In pancreatic ductal adenocarcinoma, low serum miR-607 levels were regarded as a prognostic biomarker 25. In non-small cell lung cancer, inhibition of miR-607 promoted tumorigenesis and invasion of cancer cells 26. And in osteosarcoma, miR-607 was associated with tumor proliferation 27. However, the role of miR-607 in ESCC has not been reported. Cysteine-rich secretory protein 3 (CRISP3) is a member of cysteine-rich secretory proteins and preferentially expressed in pancreas, prostate and salivary of human 28, 29. CRISP3 has been reported to be associated with inflammation and innate immunity 30. Recent studies revealed that CRISP3 was inextricably linked to cancer. In prostate cancer, CRISP3 could drive invasion and migration of cancer cells 31. In non-small cell lung cancer, decreasing CRISP3 expression levels inhibited progression and development of cancer 32. And in mammary carcinoma, patients with low expression levels of CRISP3 had a favorable prognosis 33. In ovarian cancer, CRISP3 was reported to be associated with drug resistance 34. So far, there were not relevant studies of CRISP3 in ESCC. In this study, we first used RNA-seq to screen the differential expression of circRNAs between tumor and paracancerous tissues, and identified a new circular RNA—circCEP70 from CEP70 that was significantly up-regulated in ESCC, the circBase ID of which was hsa_circ_0003823. Subsequently, we discussed the clinical significance of hsa_circ_0003823, deeply illustrated the role of hsa_circ_0003823 in the occurrence and development of ESCC, and explored the underlying molecular mechanisms. The results showed that hsa_circ_0003823 was significantly up-regulated in ESCC tissues, which was related to the pathological stage and prognosis of ESCC patients, and positively related to the expression of CRISP3. Further functional and mechanistic studies have shown that hsa_circ_0003823 could act as a sponge for miR-607, alleviating its inhibition on the target gene CRISP3, thereby promoting tumor progression, metastasis and Apatinib resistance. This study explored the expression, function, regulatory mechanism, and drug resistance of hsa_circ_0003823 in ESCC for the first time, which might provide new ideas and directions for diagnosis and prognosis of ESCC.
We collected a total of 38 pairs of tumor and paracancerous tissues of ESCC patients from the Department of Thoracic Surgery, Renji Hospital affiliated to Shanghai Jiaotong University School of Medicine, of which 3 pairs were used for RNA sequencing. All tissue samples were snap-frozen in liquid nitrogen and stored at -80°C until use. All enrolled patients signed informed consent before surgery, had no history of other malignancies, and had not received chemoradiotherapy. We analyzed basic clinical data of the patients, including age, gender, TNM stage, and tumor size. This study was approved by the Ethics Committee of Shanghai Jiao Tong University School of Medicine.
Human ESCC cell lines (TE-1, TE-13, ECA-109, EC9706, KYSE-150) and human esophageal epithelial cell line Het-1A as well as 293T cell line were purchased from American Type Culture Collection (ATCC) (Manassas, VA, USA). TE-1, TE-13, ECA-109, EC9706, KYSE-150 and 293T cell lines were cultured in DMEM or RPMI-1640 medium (Sigma, St. Louis, MO, USA) with 10% FBS (Gibco, USA), 100 U/ml penicillin (Millipore, TMS-AB2-C) and 100 U/ml streptomycin at 37°C with 5% CO2 in a humidified incubator. Het-1A cells were cultured in Bronchial epithelial cell basal medium (BEGM) with all the additives (Lonza, MD, USA). We used an intermittent stepwise selection protocol to establish Apatinib-resistant ESCC cells (ECA-109/AR and KYSE-150/AR cells) from ECA-109 and KYSE-150 cells over 6 months and confirmed the half maximal inhibitory concentration (IC50) according to the dose-response curves. Circ siRNAs, CRISP3 siRNAs and miRNA mimics/inhibitors were synthesized by Bioegene (Shanghai, China) and transfected into the ESCC cell lines using lipofectamine 3000 (Invitrogen, USA). These experiments were conducted in accordance with the manufacturer's protocols. The full-length cDNA of hsa_circ_0003823 was synthesized and inserted into the expression vector pcDNA3.1 (Bioegene, Shanghai, China), while there was no hsa_circ_0003823 sequences in the mock vector which was considered as the negative control. Then cells were treated with puromycin (Sigma, USA) until hsa_circ_0003823 overexpressed cells were stably constructed. Sequences of circ siRNA, CRISP3 siRNA and miRNA mimics and inhibitors used in this study were listed in Table S1.
Total RNA was extracted from ESCC tissues and cell lines applying TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Reverse transcription was carried out by the HiScript III RT SuperMix (Vazyme, Nanjing, China), and AceQ universal SYBR qPCR Master Mix (Vazyme, Nanjing, China) was used to detect total RNA under recommended conditions. GAPDH was used as the internal reference for mRNA and circRNA, and U6 was used for miRNA. Sequences of primers were listed in Table S2.
We performed FISH to evaluate the subcellular location of hsa_circ_0003823 using ESCC cell lines (ECA-109 and KYSE-150). After pre-hybridization was conducted at 55 °C for 2 h, cell slides were incubated with specific Cy3-labeled hsa_circ_0003823 probe (Bioegene, Shanghai, China) at 37 °C overnight and stained with DAPI. The slides were photographed with the fluorescence microscope (Leica, Germany).
We conducted in situ hybridization with a specific digoxin-labeled circRNA probe to detect the relative expression of hsa_circ_0003823 (Servicebio, Wuhan, China) on TMAs (Superbiotec, Shanghai, China), containing 60 paraffin-embedded ESCC samples. TMAs were digested with proteinase K, hybridized with the specific hsa_circ_0003823 probe overnight at 4 ℃, and then treated with anti-Digoxin-AP at 4 ℃ (Roche, Basel, Switzerland). Tissues were stained with NBT/BCIP and qualified (Roche, Basel, Switzerland).
For cell counting kit-8 (Dojindo, Japan) assays, 2x10^3 ESCC cells (ECA-109 or KYSE-150 cells) and Apatinib-resistant ESCC cells (ECA-109/AR or KYSE-150/AR cells) were plated in the 96-well plates for 5 days after cells were transfected for 48 h and the absorbance at 450 nm was measured by microplate reader. For colony formation assays, 1x10^3 ECA-109 or KYSE-150 cells were seeded in the 6-well plates and incubated for approximately 2 weeks, then stained and fixed by 1% crystal violet for 15 minutes before being photographed and counted. For apoptosis assays, ESCC cells (ECA-109 and KYSE-150) and Apatinib resistant cells (ECA-109/AR and KYSE-150/AR) were collected after transfection and stained using 3 μl FITC-Annexin V and 5 μl propidium iodide (PI, 50 μg/ml) for 15 minutes. FACS Caliber system (BD Biosciences, USA) was used to analyze apoptosis data. We set three independent events for each group.
Transfected ESCC cells (5x10^4 cells) were suspended in serum-free medium and seeded in the top chambers. The lower chambers were added into 700 μL medium with 10% fetal bovine serum. After 24 h, ESCC cells migrated from the top chambers were fixed with 1% crystal violet for 15 minutes and then photographed and counted. We set three independent events for each group.
Protein samples from cells or tissues which were lysed by RIPA buffer were subjected to 10% SDS-PAGE and then transferred onto polyvinylidene fluoride membranes. Primary antibodies including anti-CRISP3, anti-N-Cadherin, anti-E-Cadherin, anti-β-catenin, anti-Vimentin, anti-Snail were used (Table S3). GAPDH was used as the internal control.
RIP was conducted using Magna RIP kit (Millipore, MA, USA) in accordance with the manufacturer's instructions. miR-607 mimics or miR-NC were transfected into ECA-109 cells and cells were collected after 48 h and then lysed in 100 % RIP lysis buffer. RIP lysates were incubated with magnetic beads conjugated with anti-Argonaute2 (AGO2) (Millipore, MA, USA) or IgG antibody (Millipore, MA, USA) as the negative control. qRT-PCR and Western blot were used to detect the enrichment of the immunoprecipitated RNA and protein.
Male BALB/c nude mice (4-6 weeks) were purchased from the institute of zoology, Chinese Academy of Sciences of Shanghai and all animal experiments were performed strictly in accordance with the Guide for the Care and Use of Laboratory Animals and approved by the Committee of Animals Use and Care of Shanghai Jiaotong University School of Medicine (Approval ID: A-2018-024). ECA-109 cells were transfected using overexpression and mock vector and selected by puromycin to construct stably over-expressed cell lines. Lentiviruses (Bioegene, Shanghai, China) carrying sh-NC and sh-circ were transfected into ECA-109 cell lines, which were then selected with puromycin to obtain sh-NC and sh-circ cell lines which were named LV NC and LV circ0003823, respectively. 2 x 10^6 transfected ECA-109 cells were subcutaneously inoculated into the right flanks of male BALB/c nude mice (n=4 for each group). Mice in the Apatinib group received the drug (60mg/kg) daily by oral gavage. Tumor volume was measured once a week and calculated by the formula: V=0.5×length×width2. All mice were sacrificed after 4 weeks and xenograft tumors were removed to be weighed and fixed for immunohistochemistry (IHC) staining. IHC analysis was performed according to the manufacturer's instructions (Immunostain SP kit, Dako Cytomation, USA). Primary antibodies against Ki67 (Cell Signaling Technology), β-catenin (Signalway Antibody), CRISP3 (Proteintech), E-cadherin (Cell Signaling Technology) were used. IHC results were judged by staining intensity and number of positive cells and assessed by at least three pathologists in a single-blind method.
The sequences of hsa_circ_0003823, CRISP-3'UTR, and the matched mutant sequences without miR-607 binding sites were synthesized and cloned into the pmirGLO luciferase reporter vector (Promega, Madison, WI, USA). All plasmids were transfected into HEK-293T cells. Dual Luciferase Assay Kit (Promega, Madison, WI, USA) were used to measure the relative luciferase activities according to the manufacturer's protocols.
SPSS 20.0 (IBM, SPSS, Chicago, IL, USA) and GraphPad Prism 7.0 (GraphPad Software Inc., CA, USA) were used to analyze data which were expressed as mean ± standard deviation (SD). Student's t test, one-way ANOVA and χ2 test were used to assess differences between groups. Kaplan-Meier method was used to evaluate survival rates. Pearson correlation was used to analyze the correlation between groups. A receiver operating characteristic (ROC) curve was used to assess the diagnostic value. P value < 0.05 was considered statistically significant.
First, we elucidated the expression of mRNA and circular RNA by RNA-sequencing of tumor tissues and paracancerous tissues in 3 ESCC patients. By setting the criteria of fold change > 2.0 and p value < 0.05, we found that a total of 333 circRNAs were differentially expressed in tumor and paracancerous tissues, of which 212 circRNAs were significantly up-regulated, and 121 circRNAs were down-regulated (Figure 1A). We annotated the 20 circRNAs that were most up- or down-regulated in tumor tissues in Figure 1B as a heatmap. Among them, hsa_circ_0003823 was the most up-regulated circular RNA. In addition, we also analyzed mRNA sequencing results. According to the same screening criteria, a total of 1424 mRNAs were found and differentially expressed in tumor tissues and paracancerous tissues, of which 677 mRNAs were up-regulated in tumor tissues, while 747 mRNAs were down-regulated (Figure 1C and 1D). 20 mRNAs with the most pronounced up- or down-regulation were shown in a heatmap format (Figure 1E). Through Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of sequencing results, Wnt signaling pathway and cell molecules adhesion (CAMs) were the most enriched signaling pathways (Figure 1F and 1G), and CRISP3 was the most significantly up-regulated gene in tumor tissues and closely related to tumor progression and metastasis. Tumor metastasis played an important role in the occurrence and development of ESCC. Most deaths of ESCC patients were due to tumor cells spreading to other organs, proliferating and resisting conventional treatment, eventually leading to failure of important organs 35. Therefore, this study mainly focused on the roles and mechanisms of hsa_circ_0003823 and CRISP3 in ESCC progression and metastasis.
hsa_circ_0003823 consisted of 522 nucleotides and was derived from the splicing of exons 3, 4, 5, and 6 of the CEP70 gene, located at Chr3:138570318 to 138572984, also known as circCEP70 (Figure 2A). To evaluate the existence of hsa_circ_0003823, we first amplified it in 293T cells, and then performed Sanger sequencing on the PCR products, which confirmed that the head-to-tail splicing of the PCR products was in conformity to the expected size and site (Figure 2B). Considering that head-to-tail splicing was not only the result of reverse splicing of cDNA, but also the result of gene rearrangement, we designed polymer primers and divergent primers for hsa_circ_0003823 respectively, and amplified them with cDNA or gDNA of 293T cells. Results showed that hsa_circ_0003823 could only be amplified from cDNA, not gDNA (Figure 2C). In addition, the stability was also an important feature of circular RNAs 11, 36, 37. To confirm the stability of hsa_circ_0003823, RNase R was used to treat ECA-109 and KYSE-150 cell lines, and the results indicated that the expression level of CEP70 in the RNase R group was significantly reduced, while hsa_circ_0003823 expression was not obviously affected (Figure 2D and 2E). Since the function of circular RNA was usually related to its localization in cells, we further conducted FISH experiments to explore intracellular localization of hsa_circ_0003823 in ESCC cell lines, and the results showed that hsa_circ_0003823 was mostly localized in the cytoplasm (Figure 2F). Next, qRT-PCR was used to detect the expression level of hsa_circ_0003823 in 38 pairs of ESCC tumor and paracancerous tissues, and we found that the expression level of hsa_circ_0003823 was significantly higher in tumor tissues (Figure 2G). Then we divided 38 pairs of ESCC patients into high-expression group and low-expression group according to levels of hsa_circ_0003823, and further analyzed their clinical characteristics. The results showed that patients in high-expression group had higher T stage, N stage and TNM stage. However, there was no difference in age, gender and tumor size (Table 1). Besides, we also detected the expression level of hsa_circ_0003823 in ESCC cell lines (TE-1, TE-13, ECA-109, EC9706, KYSE-150) and the human esophageal epithelial cell line Het-1A, and the results indicated that the expression level was higher in the ESCC cell lines, especially ECA-109 and KYSE-150, so we chose the above two cell lines for subsequent experiments (Figure 2H). To further verify our conclusions, we performed in situ hybridization on tissue chips containing 60 ESCC tumor tissues and paracancerous tissues to detect the expression of hsa_circ_0003823, and the results indicated that the expression level was significantly higher in tumor tissues (Figure 2I). We divided 60 ESCC samples into high hsa_circ_0003823 expression group (n=30) and low expression group (n=30). We found that the high expression group had a higher proportion of lymph node metastasis and III-IV of TNM stage (Figure 2J and 2K). The Kaplan-Meier survival curve showed that the high expression group had significantly shorter overall survival and progression-free survival time than low expression group (Figure 2L and 2M). The ROC curve revealed that hsa_circ_0003823 had high sensitivity in distinguishing tumor tissues from paracancerous tissues (Figure 2N). In addition, we also examined the expression level of CRISP3 in 38 pairs of ESCC tumor and paracancerous tissues, and the results showed that the expression level of CRISP3 was significantly higher in tumor tissues (Figure 2O). Pearson correlation analysis revealed a positive correlation between the expression of hsa_circ_0003823 and CRISP3 (Figure 2P). The above results confirmed the RNA-seq data and pointed out that hsa_circ_0003823 and CRISP3 might play a synergistic role in ESCC progression.
To explore the role of hsa_circ_0003823 in ESCC cells, we first constructed hsa_circ_0003823 knockdown and overexpression cell lines, and qRT-PCR was used to validate the knockdown and overexpression effects of hsa_circ_0003823 in ECA-109 and KYSE-150 cell lines (Figure 3A-3C). For the knockdown experiments, we designed 3 shRNA sequences named sh-circ-1, sh-circ-2 and sh-circ-3. qRT-PCR results showed that the knockdown effects of sh-circ-2 and sh-circ-3 were better than sh-circ-1, so we chose the former for subsequent experiments. Next, we evaluated the effects of hsa_circ_0003823 on the proliferation of ESCC cell lines using CCK8 assays. The results showed that the up-regulation of hsa_circ_0003823 significantly enhanced the proliferation of cells, while its down-regulation inhibited the proliferation of cells (Figure 3D-3K). Plate colony formation experiments were used to further validate our conclusions, showing that knockdown of hsa_circ_0003823 reduced the number of plate clones, whereas overexpression of it significantly increased the number of plate clones (Figure 3L-3Q). Figure 3R-3U indicated knockdown of hsa_circ_0003823 significantly reduced the ability of ESCC cells for invasion and migration, while overexpression of hsa_circ_0003823 increased the number of cells undergoing invasion and migration. The effects of knockdown of hsa_circ_0003823 on the expression levels of tumor metastasis-related proteins were investigated by Western blot experiments. The results revealed that knockdown of hsa_circ_0003823 increased the expression level of E-cadherin and inhibited expression levels of N-cadherin, β-catenin, Vimentin and Snail (Figure 3V-3Z). The above results exhibited that hsa_circ_0003823 could enhance the proliferation, invasion and migration ability of ESCC cells. In addition, we also found that the expression level of CRISP3 was reduced after knockdown of hsa_circ_0003823, which further confirmed the correlation between the two.
To verify the role of hsa_circ_0003823 in proliferation, invasion and migration of ESCC in vivo, we subcutaneously injected ECA-109 cells with knockdown or overexpression of hsa_circ_0003823 into adult male nude mice. The results showed that, compared with the control group, the subcutaneous tumorigenicity in hsa_circ_0003823 overexpression group was higher (Figure 4A). The volume and weight of tumors in the overexpression group were also higher than those in the control group (Figure 4B and 4C). However, the subcutaneous tumorigenicity, tumor volume and weight were all lower in hsa_circ_0003823 knockdown group than that of the control group (Figure 4D-4F). IHC results indicated that the expression levels of Ki67 and β-catenin in the tumor tissues of the hsa_circ_0003823 overexpression group were significantly higher than the control group, while the E-cadherin level was lower in the hsa_circ_0003823 overexpression group (Figure 4G-4I). The expression levels of Ki67, β-catenin and E-cadherin in the hsa_circ_0003823 knockdown group were contrary to the above results (Figure 4J-4L). In addition, we also detected the expression level of CRISP3 and found that overexpression of hsa_circ_0003823 increased the expression level of CRISP3, while knockdown of hsa_circ_0003823 resulted in downregulation of CRISP3 (Figure 4G-4L). The above results revealed the importance of hsa_circ_0003823 in the occurrence and development of ESCC in vivo, which might be related to CRISP3 and metastasis-related signaling pathways.
In order to elucidate the downstream molecular mechanism of hsa_circ_0003823 in ESCC, we first predicted the potential miRNA targets of hsa_circ_0003823 through circbank and circ interactome databases. The results showed that hsa_circ_0003823 had a targeted binding site on the sequences of miR-607 (Figure 5A). Next, we used qRT-PCR to detect the expression level of miR-607 in 38 pairs of ESCC tumor and paracancerous tissues, and found that the expression level of miR-607 in tumor tissues was lower (Figure 5B). To verify the targeting effect of hsa_circ_0003823 on miR-607, we performed a dual-luciferase reporter experiment in 293T cells. The full-length sequences of hsa_circ_0003823 of wild type or mutant type without miR-607 binding sites were subcloned into the luciferase reporter vector pmirGLO, which was transfected into ESCC cells, and then transfected with miR-607 NC and mimics, respectively. The results showed that compared with the control group, miR-607 mimics could significantly reduce the luciferase activity in the wild type group, while there was no significant effect on the mutant group, indicating that there was a direct interaction between hsa_circ_0003823 and miR-607 (Figure 5C). In addition, we verified our conclusions by anti-AGO2 RNA immunoprecipitation (RIP) assays in ECA-109 cells. The results showed that compared with the IgG group, AGO2 antibody could obviously pull down hsa_circ_0003823. And compared with the miR-607 NC group, cells transfected with miR-607 mimics could enrich hsa_circ_0003823 more efficiently (Figure 5D and 5E). hsa_circ_0003823 was knocked down in ECA-109 and KYSE-150 cell lines and the expression level of miR-607 was detected by qRT-PCR. We found that compared with the control group, the expression levels of miR-607 in the hsa_circ_0003823 knockdown groups increased (Figure 5F and 5G). Conversely, overexpression of hsa_circ_0003823 in ESCC cell lines suppressed the expression level of miR-607 (Figure 5H). Pearson correlation analysis revealed a negative correlation between hsa_circ_0003823 and miR-607 expression in 38 pairs of ESCC tumor tissues (Figure 5I). Our previous results have shown that hsa_circ_0003823 was positively correlated with the expression of CRISP3, and hsa_circ_0003823 could act as a miR-607 sponge to inhibit its function. To explore the interaction among hsa_circ_0003823, miR-607 and CRISP3, we used the TargetScan database for prediction and found that miR-607 was the target miRNA of hsa_circ_0003823, and CRISP3 was the target gene of miR-607 (http://www.targetscan.org/vert_72/). miR-607 inhibitor was used to inhibit its expression (Figure 5J), and miR-607 mimics for increasing its expression (Figure 5K). qRT-PCR results showed that compared with the control group, miR-607 inhibitor significantly increased the level of CRISP3, while miR-607 mimics reduced the level of CRISP3 (Figure 5L and 5M). The dual luciferase reporter assay showed that compared with the CRISP3 3'UTR-Mut group, miR-607 mimics could obviously reduce the fluorescence intensity of the luciferase reporter vector carrying the CRISP3 3'UTR-WT sequences (Figure 5N). Western blot results showed that compared with the control group, miR-607 inhibitor could significantly increase the expression level of CRISP3, while miR-607 mimics could inhibit the expression level of CRISP3 (Figure 5O-5Q). qRT-PCR was used to detect the level of CRISP3 in the previously constructed hsa_circ_0003823 knockdown or overexpression ESCC cell lines. Results indicated that compared with the control group, knockdown of hsa_circ_0003823 inhibited the level of CRISP3, while overexpression of hsa_circ_0003823 increased the level of CRISP3 (Figure 5R-5T). The above results suggested that hsa_circ_0003823 acted as a miR-607 sponge to inhibit it and activate the expression of CRISP3 to promote tumor progression of ESCC.
In order to explore the specific mechanism of hsa_circ_0003823 exerting its biological function, miR-607 mimics were transfected into ESCC cell lines under the premise of overexpression of hsa_circ_0003823. Western blot results showed that overexpression of hsa_circ_0003823 up-regulated the expression levels of N-cadherin, β-catenin, Vimentin, Snail, and down-regulated the expression level of E-cadherin, while the transfection of miR-607 mimics reversed the above phenomena (Figure 6A). In addition, we found that the upregulation of CRISP3 caused by overexpression of hsa_circ_0003823 was also reversed by the transfection of miR-607 mimics (Figure 6A). The results of migration and invasion experiments showed that overexpression of hsa_circ_0003823 could promote cell migration and invasion, while the miR-607 mimics group not only inhibited the migration and invasion ability of ESCC cells, but also reversed the increase in the number of migrating and invading cells due to hsa_circ_0003823 (Figure 6B-6E). Moreover, the results of cell migration and invasion experiments indicated that knockdown of hsa_circ_0003823 could reduce the migration and invasion ability of cells, while inhibiting the expression of miR-607 could increase the number of migrating and invading cells, and reverse the reduction in the number of migrating and invading cells caused by knockdown of hsa_circ_0003823 (Figure 6F-6I). CRISP3 siRNAs were transfected into ESCC cell lines to reveal the function of CRISP3 in this process. Western blot results indicated that knockdown of CRISP3 down-regulated the expression levels of N-cadherin, β-catenin, Vimentin, Snail, and up-regulated the expression level of E-cadherin (Figure 6J). The results of cell migration and invasion experiments showed that knockdown of CRISP3 reduced the migration and invasion ability of cells (Figure 6K-6N). CCK8 assays were performed to evaluated the effects of CRISP3 on the proliferation of ESCC cell lines. The results showed that knockdown of CRISP3 significantly inhibited the proliferation of cells (Figure 6O-6R). Plate colony formation experiments were conducted to further validate our conclusions, indicating that knockdown of CRISP3 reduced the number of plate clones (Figure 6S-6T). The above results suggested that hsa_circ_0003823 promoted progression, invasion and migration of ESCC through miR-607/CRISP3 axis.
To test the effects of hsa_circ_0003823 on Apatinib sensitivity, we set 7 different Apatinib concentrations to treat previously constructed hsa_circ_0003823 knockdown and overexpression cell lines (ECA-109 and KYSE-150). The results showed that inhibiting hsa_circ_0003823 significantly weakened the viability of Apatinib-treated cells, while overexpression of hsa_circ_0003823 obviously enhanced Apatinib resistance (Figure 7A-7D). IC50 values were calculated and presented in Figure 7E. To further explore the regulation of hsa_circ_0003823 on Apatinib sensitivity, we constructed Apatinib-resistant ECA-109/AR and KYSE-150/AR cells. Compared with normal ESCC cell lines, the expression levels of hsa_circ_0003823 were significantly increased in Apatinib-resistant cells (Figure 7F). ECA-109/AR and KYSE-150/AR cells were transfected with hsa_circ_0003823 siRNA or miR-607 inhibitor and then treated with different concentrations of Apatinib. Results indicated that knockdown of hsa_circ_0003823 weakened viability of Apatinib-treated ESCC/AR cells, while inhibiting miR-607 was able to reverse the above phenomenon (Figure 7G-7I). Flow cytometry with double staining of Annexin V and PI was used to analyze the effects of hsa_circ_0003823 knockdown or miR-607 inhibitor on the apoptosis of ECA-109 and KYSE-150 cell lines treated with Apatinib, and the results showed that inhibiting miR-607 decreased the apoptosis ratio of ESCC cell lines, while knockdown of hsa_circ_0003823 could reverse the effects of miR-607 inhibitor (Figure 7J-7K). CRISP3 siRNA was transfected into normal and Apatinib-resistant ESCC cells. Results indicated that knockdown of CRISP3 decreased viability of both normal and Apatinib-resistant ESCC cells and increased apoptosis rates (Figure 7L-7P). We also validated the above conclusions by in vivo experiments. The normal ECA-109 cell line or cells transfected with mock or hsa_circ_0003823 was subcutaneously injected into adult male nude mice. After two weeks, all mice were assigned into 4 groups: control group, Apatinib group, Apatinib and mock group, Apatinib and circ0003823 group. All the mice were sacrificed four weeks after the drugs administration, and the tumors were removed to measure the volume and weight. The results showed that the tumorigenicity of ECA-109 cells in Apatinib group was significantly reduced, and the tumor volume and weight were significantly lower than those of the control group (Figure 8A-8C). Overexpression of hsa_circ_0003823 could reverse the inhibitory effects of Apatinib on tumor (Figure 8A-8C). IHC results showed that the expression levels of Ki67, β-catenin and CRISP3 in the Apatinib group were significantly decreased, while the expression level of E-cadherin was increased. Similarly, the above phenomenon was reversed by overexpression of hsa_circ_0003823 (Figure 8D-8F). Western blot was performed to further verify our conclusions and results showed that Apatinib could inhibit expression levels of N-cadherin, β-catenin, Vimentin, Snail and CRISP3, increase the level of E-cadherin, while overexpression of hsa_circ_0003823 reversed the above phenomenon (Figure 8G-8H). The above in vitro and in vivo experiments confirmed that hsa_circ_0003823 regulated the sensitivity of ESCC cells to Apatinib through miR-607/CRISP3 axis in vitro and in vivo.
ESCC is one of the most malignant tumor with high incidence and lethality, which is prone to early distant metastasis and drug resistance, and the regulatory mechanism of ESCC is still unclear 3, 38-40. In recent years, there have been a large amount of studies on circular RNAs, and now we have better understanding of the roles of circular RNAs in biogenesis and biology. However, the regulatory functions and corresponding mechanisms in many diseases, especially tumors, are still not thoroughly studied 41-46. At present, there are relatively few reports on the development, metastasis and drug resistance of circRNAs in ESCC, and the specific mechanism needs to be clarified. Wang J revealed that knockdown of circ_0087378 could repress the tumorigenesis and progression of ESCC by modulating the miR-140-3p/E2F3 axis 47. Liu Z focused on the role of circDOPEY2 and concluded that circDOPEY2 inhibited CPEB4-mediated Mcl-1 translation process and enhanced chemosensitivity of ESCC 48. Liang Y reported that CircIMMP2L promoted ESCC progression via CtBP1 nuclear retention dependent epigenetic modification 49. We discovered a novel circRNA named hsa_circ_0003823 by identifying the circRNA and mRNA expression profiles of 3 pairs of ESCC tumor and paracancerous tissues, which was also the most up-regulated circRNA in tumor tissues. Next, we detected the expression level of hsa_circ_0003823 in 38 pairs of ESCC tumor and paracancerous tissues, and determined that it was highly expressed in tumor tissues, and was closely related to the TNM stage, especially N stage of ESCC patients, Kaplan-Meier survival curve and other prognostic indicators. Further functional experiments showed that knockdown of hsa_circ_0003823 could inhibit the proliferation of ESCC cells, and weaken the ability of cells to invade and migrate, while overexpression of hsa_circ_0003823 had the opposite effects. And we also found that hsa_circ_0003823 affected expression levels of metastasis-related proteins. These experimental results indicated that hsa_circ_0003823 played an important role in the occurrence, development and metastasis of ESCC. The downstream mechanism of circRNAs is related to the localization. CircRNAs are generally acting as ceRNAs by sponging miRNAs when localized in the cytoplasm 50-52. It was reported that circLPAR3 could act as a miR-198 sponge to promote the invasion and migration of ESCC 53. CircRNA_2646 performed the function as the ceRNA to promote progression of esophageal squamous cells by inhibiting miR-124/PLP2 signaling pathway 54. Another study found that ciRS-7 accelerated ESCC progression through acting as a miR-876-5p sponge to enhance MAGE-A family expression 55. We confirmed that hsa_circ_0003823 was mainly located in the cytoplasm of ESCC cells using FISH experiments. By predicting the potential miRNA targets of hsa_circ_0003823 using relevant databases, we found that there were targeted binding sites between hsa_circ_0003823 and miR-607. It has been reported that miR-607 plays an important role as the tumor suppressor in various tumors including pancreatic cancer, non-small cell lung cancer, and osteosarcoma 25-27, however, its role in ESCC has not yet been elucidated. We found that the expression level of miR-607 in ESCC tumor tissues was significantly lower than that in paracancerous tissues. Dual luciferase reporter and anti-AGO2 RNA immunoprecipitation (RIP) assays confirmed the interaction between hsa_circ_0003823 and miR-607. Knockdown of hsa_circ_0003823 in ESCC cell lines upregulated miR-607, while overexpression of hsa_circ_0003823 suppressed miR-607 expression. Pearson correlation analysis showed that there was a negative correlation between the two. Therefore, we speculated that hsa_circ_0003823 might act as a role in accelerating tumor progression in ESCC through sponge of miR-607. CRISP3 is a member of the cysteine-rich secretory proteins, and it has been reported that CRISP3 was involved in the occurrence, development and drug resistance in a variety of tumors including prostate cancer, non-small cell lung cancer, breast cancer and so on 31-34, however, its role in ESCC remained unknown. Through RNA-seq in 3 pairs of ESCC tumor and paracancerous tissues and qRT-PCR experiments on 38 pairs of tumor and paracancerous tissues, we found that the expression level of CRISP3 in tumor tissues was significantly higher, and it was closely related to tumor development and metastasis. Pearson correlation analysis showed that there was a strong positive correlation between CRISP3 and hsa_circ_0003823. Knockdown or overexpression of hsa_circ_0003823 in ESCC cell lines significantly decreased or increased the expression level of CRISP3, further confirming the close relationship between the two. hsa_circ_0003823 could adsorb miR-607 through sponges and inhibit its function, thereby promoting ESCC progression. TargetScan database prediction showed that CRISP3 was one of the potential target genes of miR-607. Dual-luciferase reporter assays confirmed that miR-607 could directly target the 3-untranslated regions of CRISP3. Inhibition of miR-607 significantly increased the mRNA and protein levels of CRISP3, whereas overexpression of miR-607 had the opposite effects. Therefore, we confirmed that CRISP3 could be positively regulated by hsa_circ_0003823, which acted as the sponge of miR-607. KEGG enrichment analysis of RNA-seq results showed that metastasis-related signaling pathways were the most enriched pathways. Therefore, in terms of mechanism exploration, we focused on whether hsa_circ_0003823 affected ESCC progression and metastasis through the miR-607/CRISP3 signaling axis. Our study showed that overexpression of hsa_circ_0003823 significantly enhanced the invasion and migration ability of tumor cells, and increased the expression levels of CRISP3 and metastasis-related proteins, while knockdown of hsa_circ_0003823 had the opposite effects. At the same time, we found that miR-607 partially reversed the changes in invasion and migration ability of ESCC and the expression levels of CRISP3 and metastasis-related proteins caused by hsa_circ_0003823. Functional experiments showed that knockdown of CRISP3 inhibited the proliferation, migration and invasion of ESCC cells, and affected expression levels of metastasis-related proteins. This study demonstrated that hsa_circ_0003823, as a ceRNA, promoted CRISP3-mediated tumor progression and metastasis in ESCC by inhibiting miR-607. Apatinib is a novel VEGFR-2 tyrosine kinase inhibitor and has been reported not only to inhibit tumor progression but also to increase the sensitivity of tumor cells to chemotherapy drugs. Our previous study revealed the important role of Apatinib in ESCC and found that Apatinib could inhibit proliferation, migration and invasion, induce ER stress, autophagy and apoptosis, and potentiate cell sensitivity to paclitaxel in ESCC 56. However, drug resistance was prone to occurrence for advanced ESCC patients. Several studies have reported the correlation between circRNA and drug resistance. Circ0008399 was reported to promote cisplatin resistance through interacting with WTAP in bladder cancer 15. CircRNA-SORE could mediate sorafenib resistance in hepatocellular carcinoma by stabilizing YBX1 16. Another study focused on the function of circSNX6 and revealed that it could promote sunitinib resistance in renal cell carcinoma through miR-1184/GPCPD1/lysophosphatidic acid axis 14. However, the role of circRNA in drug resistance of ESCC has been rarely reported. Our study explored the effects of hsa_circ_0003823 on Apatinib sensitivity and found that inhibiting hsa_circ_0003823 significantly weakened the viability of Apatinib-treated cells through miR-607/CRISP3 axis. For in vivo experiments, overexpression of hsa_circ_0003823 could reverse the inhibitory effects of Apatinib on tumorigenicity. Our study confirmed that hsa_circ_0003823 regulated the sensitivity of ESCC cells to Apatinib through miR-607/CRISP3 axis in vitro and in vivo. Of course, this study had some limitations. The entire study was based on commercially purchased ESCC cell lines, which couldn't provide the most reliable in vivo and in vitro experimental results. Therefore, if necessary, further validation of experimental results using ESCC cells derived from tumor tissues of patients or PDX models was required.
In summary, in current studies, for the first time, we elucidated the function of hsa_circ_0003823 in ESCC and the underlying mechanism by regulating the expression of CRISP3 through sponge adsorption of miR-607, thereby promoting the progression, metastasis and Apatinib resistance of ESCC. Our findings suggested that hsa_circ_0003823 might be a potential biomarker and a novel target in the diagnosis and treatment of ESCC. The study of hsa_circ_0003823/miR-607/CRISP3 axis would expand our knowledge in understanding the underlying pathogenesis of ESCC.
Click here for additional data file. | true | true | true |
|
PMC9576539 | Ankit Srivastava,Tomas Bencomo,Ishani Das,Carolyn S. Lee | Unravelling the landscape of skin cancer through single-cell transcriptomics | 17-10-2022 | Skin,Single-cell transcriptomics,Single-cell RNA-sequencing,Spatial transcriptomics,Melanoma,Basal cell carcinoma,Cutaneous squamous cell carcinoma,Merkel cell carcinoma | Highlights • Human skin is a multifunctional and heterogeneous organ with a vast cellular pool. • Single-cell RNA-sequencing provides an unprecedented opportunity to investigate and understand the heterogeneous landscape of skin cancers. • Analysis of single-cell skin cancer transcriptomes opens new avenues of investigation, including the potential for precision-targeted therapy. | Unravelling the landscape of skin cancer through single-cell transcriptomics
• Human skin is a multifunctional and heterogeneous organ with a vast cellular pool. • Single-cell RNA-sequencing provides an unprecedented opportunity to investigate and understand the heterogeneous landscape of skin cancers. • Analysis of single-cell skin cancer transcriptomes opens new avenues of investigation, including the potential for precision-targeted therapy.
Cellular diversity and heterogeneity underlie the skin's many functions, which include sensation, stabilization of body temperature, maintaining fluid and electrolyte balance, vitamin and hormone synthesis, protection from external insults, and immunologic defense [1], [2], [3], [4].The skin is divided into three broad compartments- the epidermis, the dermis, and hypodermis/subcutaneous layer. Each compartment contains specific anatomical structures and diverse cell populations that contribute to skin physiology and homeostasis [1], [2], [3], [4], [5], [6]. The epidermis is arranged in stratified epithelial layers that establish the skin permeability barrier. While keratinocytes comprise >90% of the epidermal cell population, the epidermis also contains melanocytes, Langerhans cells, and Merkel cells [1], [2], [3], [4], [5], [6]. By comparison, the dermis is normally paucicellular and characterized by an extracellular matrix of collagen and elastin fibers that provides structural support as well as a niche for sparsely distributed fibroblasts and a variety of immune cells [[1], [2], [3], [4],7]. Skin adnexa, blood and lymphatic vessels, and nerve bundles are also housed in the dermis [1], [2], [3], [4]. The hypodermis is primarily composed of adipocytes organized into fat lobules and loose connective tissue; however, it also contains fibroblasts and macrophages (Fig. 1). Crosstalk between different cell populations, such as structural epithelial cells and immune cells, is increasingly appreciated to play an important role in both epidermal homeostasis as well as disease states [1,3,4]. Our aging population, rising global temperatures, and challenges to changing sun-related attitudes and behaviors all contribute to the rising incidence of skin cancer. Keratinocyte carcinomas, namely, basal cell carcinomas (BCC) and cutaneous squamous cell carcinomas (cSCC) [8], account for the majority of these cases and are the most common cancers in the United States [9,10]. While less common, melanoma and Merkel cell carcinoma (MCC) also demonstrate growing incidence rates. Histopathologically, these tumors are characterized by nests or sheets of neoplastic cells within the epidermis and/or involving the dermis as well immune cell infiltrates that reflect the host immune response. Local treatment of small and/or early skin cancers by surgical excision or radiation therapy is often curative; however, effective treatments for advanced or metastatic tumors have met with more measured success. The heterogeneity of tumor cells and cells comprising the tumor microenvironment (TME) often underlies therapeutic resistance, suggesting that a deeper understanding of these cell populations and their interactions might facilitate the development of more effective targeted treatments. Single-cell RNA-sequencing (scRNA-seq) now provides an unprecedented opportunity to study malignant cell heterogeneity and TME complexity at single-cell resolution. A number of recent review articles details how scRNA-seq has recently been applied to study skin homeostasis as well as cutaneous inflammatory diseases and cancers [8,11,12]. Here we discuss how our understanding of skin cancer biology has advanced using bulk transcriptomics as well as single-cell approaches and highlight different methods of scRNA-seq analysis.
When compared to hybridization-based microarrays, next-generation RNA sequencing (RNA-seq) possesses several strengths, including greater dynamic range and the capacity for de novo transcriptome assembly. In the skin cancer setting, whole transcriptome profiling by RNA-seq has been used to identify novel genes, such as non-coding RNAs and lowly expressed mRNA isoforms/transcripts. More recently, RNA-seq-based approaches have also been used to predict treatment response, nominate new therapeutic targets, and advance our understanding of disease progression as well as pathogenesis. For example, a recent analysis of melanocytic nevi and primary melanomas by Kunz et al. using RNA-seq revealed separate transcriptome signatures, with enrichment of genes driving BRAF/MEK inhibitor resistance in NRAS-mutant nevi and primary melanomas, while genes associated with resistance to PD-1 inhibition were expressed in lesions with wild type NRAS [13]. This work defined two different trajectories for melanoma development, suggesting that treatment resistance is determined early in melanomagenesis [13]. Additional work by Svedman et al. using targeted RNA sequencing (Ion AmpliSeq) to characterize melanomas following immune checkpoint inhibition similarly demonstrated enrichment of key pathways such as DNA replication, chromatin remodeling, and cell cycle, which may predict long-term response to therapy [14]. RNA-seq has also been used to highlight novel targets and signaling pathways in BCC. Wan et al. analyzed the transcriptomes of BCC with high tumor purity compared to patient-matched normal skin and identified enrichment of C2H2-type Zinc Finger genes, including Gli transcription factors as well as many others not functionally characterized in this malignancy [15]. This study, along with a larger pooled analysis of publicly available RNA-seq data by Litvinov et al. confirmed known dysregulated pathways in BCC and also suggested new therapeutic targets in the Wnt/β-catenin as well as IL-17 signaling pathways [15,16]. Other groups have used RNA-seq to elucidate the molecular underpinnings of skin malignancies. Using human specimens that represent increasingly dysplastic states on the cSCC spectrum as well as the equivalent tissues in SKH-1E cSCC-prone mice, Chitsazzadeh et al. distilled four TFs – ETS2, SP1, FOXF2 and AP1 – that appear to regulate the continuum of cSCC development [17]. Similar results were obtained by Das Mahapatra et al., who compared cSCC to unmatched normal skin controls and also reported enrichment of genes regulating immunological pathways, highlighting the importance of immunotherapy in this malignancy [18]. In MCC, which is divided into virus-positive and virus-negative categories based on the presence or absence of Merkel cell polyomavirus (McPyV) transcripts, Starrett et al. used RNA-seq to compare virus-positive to virus-negative MCC transcriptomes and demonstrated that genes regulating the cell cycle were increased in the former while DNA repair genes were downregulated in the latter [19]. These findings demonstrated a role for McPyV in controlling the MCC transcriptome and provide an explanation for the high mutation burden observed in virus-negative tumors [19]. As demonstrated by the aforementioned studies, bulk RNA-seq-based analyses are a valuable mainstay of investigative skin cancer biology. There are many advantages to using this approach, such as detection of mRNA isoforms and RNA species, high read capture capacity which enables detection of low abundance genes, low signal-to-noise ratio, and cost effectiveness. However, when applied to cancer, a few important limitations should be considered. First, details regarding the cellular context of the observed mRNA changes are not readily provided by bulk RNA-seq, although methods now exist to deconvolute this data into known cell populations using reference signatures. Second, bulk sequencing measures the average expression of a pooled population of cells, raising the possibility that transcriptomic changes of rare populations may be masked. This is true of many tissues including skin, where heterogeneity is present not only at the level of diverse cell types and states (Fig. 1), but also evolves from disease progression and treatment. The emergence of therapeutic resistance following cancer chemotherapy, for example, is a setting in which bulk sequencing can fall short due to its inability to distinguish between drug-sensitive and resistant clones. Additionally, the ability to explore RNA trajectory, cell-specific multi-omics or cell-cell communication is limited with bulk RNA-seq. Thus, these and similar questions are increasingly being addressed by single-cell approaches that provide a more granular view of cell types and states, potentially revealing new targetable transcriptomic changes. The two widely used methods for single cell sequencing are 10X genomics and Smart-seq2. Smart-seq2 has been shown to detect low abundance transcripts and isoforms, while 10X genomics is better at detecting rare cell types [20]. Using these methods and scRNA-seq data analysis approaches, researchers have highlighted cellular/spatial context of skin cancer transcriptome and deciphered novel cell-to-cell communications.
A key goal of scRNA-seq is to characterize the cell types present within a sample, as this allows researchers to understand cell-type specific gene expression. An important first step when analyzing scRNA-seq data is to consider batch effects, or gene expression patterns caused by non-biological factors that can cause erroneous interpretation of cell types. Removing this technical noise for proper biological interpretation is referred to as integration. Methods like Seurat CCA, Harmony, scVI, scanorama, and scMC have been developed to integrate cells from multiple technical conditions [21], [22], [23], [24], [25]. Although integration methods help remove technical noise, when applied incorrectly they can overcorrect and remove biological variations. Many approaches have been developed to identify cell types. Manual annotation is a popular technique combining dimensionality reduction and unsupervised clustering followed by manual inspection of marker genes. This approach works best for smaller studies or when most of the cells being analyzed are expected to have few known marker genes. For larger studies of samples with previously profiled cell types, automated classification methods can greatly speed up annotation. Some of these approaches, such as SingleR, CellAssign, and Garnett rely on transcriptomic reference profiles or gene signatures to label cell types [26], [27], [28], [29], [30]. By contrast other methods, including Azimuth, ProjecTILs, and scArches projects new cells to a reference atlas of previously annotated cell types [27,31,32]. Large consortium efforts such as Human Cell Atlas are currently cataloging major cell types and generating reference atlases that can be used to map and classify data from new studies [33]. After identifying broad cell types, further clustering is often performed on cells within a specific lineage (e.g., tumor cells, fibroblasts, T-cells) to identify subpopulations of cells with transcriptionally distinct states. This analysis is usually exploratory in nature and thus performed manually unless prior studies have already determined marker genes for distinct cell states. In the cancer setting, identification of cell types and subclusters is often used to characterize intra- and intertumoral heterogeneity and can also be applied to immune and stromal cells to study the transcriptional heterogeneity of the TME. For example, in the first scRNA-seq study of skin cancer, Tirosh et al. analyzed 4645 cells isolated from 19 patients with melanoma and showed that across all tumors, malignant cells from the same melanoma contained two distinct transcriptional states not distinguishable by bulk RNA-seq that were characterized by either high expression of MITF or AXL. Cells with high AXL expression exhibited drug resistance to RAF/MEK inhibition, suggesting they drive tumor recurrence [34]. This work also highlighted differential T-cell activation, expansion, and clonal exhaustion programs across different patients [34]. By sequencing single-cells from short-term culture of three melanomas with differing BRAF/NRAS genotypes, Gerber et al. derived gene expression modules from different cell subpopulations that included those with high MITF or AXL and also revealed targetable upregulation of the CDK4 and CDK2 cell cycle-dependent kinases responsive to palbociclib [35]. Yao and colleagues, identified similar patterns of intratumoral heterogeneity in BCC by utilizing scRNA-seq to characterize tumor cells based on expression of nuclear myocardin-related transcription factor (nMRTF), which correlates with Smoothened (SMO) inhibitor resistance and is targetable using AP-1 inhibitors [36]. In cSCC, Ji et al. also identified four different keratinocyte subpopulations, including tumor-specific keratinocytes (TSKs). TSKs are observed at the leading edge of the tumor and are thus positioned to act as a hub in cSCC tumor-stromal interactions [37]. Tumor subpopulation clustering was also used by Paulson et al. to elucidate mechanisms of treatment relapse in McPyV+ MCC by demonstrating tumor cells adapt to the immunological pressure created by autologous McPyV-specific CD8+ T cells and immune checkpoint inhibitors by suppressing the HLA specific to the targeted McPyV epitope [38]. When applied to cancer-associated immune and stromal cells, the identification of cell types and subclusters has broadened our understanding of the transcriptional heterogeneity of the TME. Several studies have examined the landscape of T-cells in skin cancers, finding a diverse range of cells including cytotoxic, effector, exhausted, regulatory, and helper T- cells. Yost et al. coupled scRNA and T-cell receptor sequencing (scTCR-seq) to study advanced BCC following PD-1 blockade and showed that novel T-cell clones infiltrate the TME of advanced BCC upon treatment, replacing their pre-existing exhausted counterparts rather than reinvigorating them [39]. Deng et al. performed a meta-analysis of 59 melanomas and catalogued the transcriptional states of CD8+ T-cells. They defined seven cytotoxic, exhausted, and naive/memory subpopulations, including a subset of exhausted T-cells associated with poor prognosis and characterized by high expression of PMEL, TYRP1, and EDNRB [40]. Frazzette et al. recently used scTCR-seq and gene expression data to compare the T-cell landscape in cSCC from skin cancer-prone immunocompromised organ transplant recipients (OTR) to that of immunocompetent individuals [41]. This effort highlighted several differences between the tumor-infiltrating lymphocytes (TILs) of these two groups, with a reduction in cytotoxic T-cell number, TCR clonotypes, and clonal expansion observed in OTR [41]. Other studies, such as that by Davidson and colleagues, have focused on characterizing cancer-associated fibroblasts (CAFs) [42]. Using the murine B16 melanoma model, these authors defined three functionally distinct CAF subpopulations that are positioned to play roles in immune crosstalk, extracellular matrix remodeling, and cytoskeletal reorganization [42].
The availability of expression profiles from multiple cell types within a single sample allows detection of intercellular signaling via ligand-receptor expression analysis [43,44]. Several computational methods have been developed, including CellPhoneDB and CellChat [45,46]. These tools utilize databases with prior knowledge of ligand-receptor pairs or complexes to infer cell-cell communication based on expression of ligands and receptors from annotated cell groups (Fig. 2). This approach has proved to be especially useful for investigating interactions between tumor cells and their microenvironment as well as immune-stromal crosstalk in skin cancers (Fig. 2). Within the TME, CAFs are well known to modulate tumor growth and progression. In their work identifying the TSK keratinocyte subcluster in cSCC, Ji and colleagues used NicheNet [37] to highlight TSK signaling to CAFs through ligand-receptor pairing of MMP9-LRP1 and TNC-SDC137. Similarly, work by Davidson et al. used CellPhoneDB to identify a subset of CAFs in the murine melanoma TME that express the immunomodulatory C3 protein and are in close contact with C3aR-positive myeloid cells; disrupting this interaction reduced tumor growth and resulted in fewer myeloid cells that suppress CD8+ T-cells [42]. Deng et al. also applied CellPhoneDB to their re-analysis of publicly available single-cell melanoma transcriptomes to show that ligand-receptor interactions involving exhausted T-cells, such as CCL5-CCR5 and CD74-MIF, are central to communication between different CD8+ subpopulations [40]. Guerrero-Juarez et al. used CellChat to investigate tumor-stromal interactions in BCC and identified that the nearby fibroblasts act as an inflammatory hub regulating BCC growth through induction of heat shock proteins [47]. Knowledge of ligand-receptor pairing was also essential to work performed by Miao and colleagues, who performed adoptive cytotoxic T-cell transfer (ACT)-based immunotherapy in a mouse model of cSCC to investigate mechanisms of immune evasion [48]. Using scRNA-seq and lineage tracing, these authors identified a population of CD80+ tumor-initiating stem cells that are refractory to ACT and engage with CTLA4-expressing cytotoxic T-cells, reducing their activity and promoting tumor relapse [48].
Genetic variation is traditionally assayed using exome or genome sequencing, which provides data on mutations, structural variants, and genomic copy number. When applied to cancer, these methodologies allow researchers to query the clonal structure of tumors, estimate tumor purity and ploidy, and distinguish neoplastic from healthy samples. Single cell DNA-sequencing (scDNA-seq) enables elucidation of genetic heterogeneity at higher resolution than bulk exome/genome approaches; however, it is not used as widely as scRNA-seq due to its high cost, lack of commercial kits, and data quality issues [49]. Method developers have attempted to circumnavigate these issues by estimating single-cell copy number profiles from scRNA-seq data with tools such as inferCNV, honeyBadger, copyKat, and CaSpER [50], [51], [52]. These approaches have been used to distinguish neoplastic from normal cells and understand clonal lineages within tumors. For example, in their scRNA-seq analysis of melanoma, Tirosh et al. classified melanocyte-like cells as malignant, non-malignant, or intermediate based on a copy number alteration score [34]. Only malignant cells were included in subsequent analyses of tumor gene expression programs, thus avoiding contamination from normal melanocytes that can occur when analyzing bulk RNA-seq data [34]. Yost and colleagues demonstrated concordance between copy number variants inferred from single-cell transcriptomes and those called using whole exome sequencing data in BCC [39].
Many biological processes such as differentiation create cells that exist along a continuum rather than distinct transcriptional states. Trajectory inference methods have been developed to infer the relationship between cells across a continuous process [53,54]. RNA velocity analysis is a related technique that uses the proportion of unspliced and spliced mRNA molecules within a cell to provide temporal ordering along the trajectory [55] (Fig. 3). When applied to skin cancer, trajectory analysis has also been used to study tumor evolution. Wouters et al. employed Monocle-2 to characterize phenotype switching in melanoma [56]. Their findings demonstrate that melanoma cells transition from a melanocytic transcriptional state to a mesenchymal-like one through a stable intermediate state characterized by unique chromatin features and transcriptionally regulated by SOX6, NFATC2, EGR3, ELF1 and ETV456. Another trajectory study by Su and colleagues found that BRAF V600E melanomas exhibit a bifurcated developmental trajectory in response to BRAF inhibition that is dependent on MITF expression [57]. Trajectory analyses have also provided new insight into the TME response to treatment. Using scRNA and T-cell receptor sequencing, Yost et al. showed that novel T-cell clones infiltrate the TME of advanced BCC following PD-1 blockade, replacing their pre-existing exhausted counterparts. The authors used trajectory analysis to analyze the evolution of T-cells in response to anti-PD-1 therapy and discovered they follow a bifurcated trajectory, becoming either terminally activated or exhausted, with the latter state marked by increased expression of PDCD1 and HAVCR239. Deng and colleagues observed similar findings in melanoma, with their trajectory analysis suggesting that CD8+ T-cells convert to either exhausted or cytotoxic terminal states [40].
While scRNA-seq offers new opportunities to study skin cancer, bulk RNA-seq remains more common due to its lower cost and ease of use. Transferring findings from single-cell experiments to bulk datasets to take full advantage of available data resources is thus an important part of single-cell analysis. The existence of several publicly available skin cancer datasets with bulk microarray or RNA-seq data allows researchers to validate findings from single-cell experiments in larger cohorts of bulk data such as The Cancer Genome Atlas (TCGA) [58]. One common approach is to derive cell-type or tumor subpopulation-specific gene signatures from single-cell data and then score bulk data using these signatures, allowing signature activity to be correlated with clinical outcomes such as overall survival or metastasis [34]. More advanced deconvolution algorithms such as CIBERSORTx, MuSiC, and others use single-cell or bulk reference profiles to estimate cell-type abundance and gene expression from bulk tumor samples [59], [60], [61]. In the skin cancer setting, deconvolution has successfully been applied to study tumor-stromal interactions. Tirosh and colleagues used gene signatures derived from single-cell melanoma transcriptomes to deconvolute 471 TCGA melanoma samples and showed that CAF abundance correlates with AXL expression [34]. They also identified a set of CAF-expressed genes that correlates with T-cell infiltration [34]. Similarly, after deriving signatures for T-cell exhaustion and cytotoxicity from single-cell data and scoring TCGA melanoma samples with these signatures, Deng et al. identified cytotoxic and exhausted subpopulations that predict prognosis [40]. Other studies have used deconvolution algorithms to link levels of infiltrating immune cells with survival and nominate genes that may be involved in this process [62], [63], [64]. Deconvolution was also been applied to cSCC by Ji et al., who linked TSK-like expression with CAF activity in several TCGA cancer types [37].
Single-cell approaches are increasingly being combined with spatial transcriptomics (ST) to gain further insight into the complex nuances of different cell types and states. By providing tissue context, ST enables researchers to understand gene expression as it relates to cell position within a tissue. The two widely used methods for ST are imaging‑based spatially resolved transcriptomics and in situ barcoding‑based spatially resolved transcriptomics. Imaging‑based spatially resolved transcriptomics methods such as MERFISH, seqFISH and ABER-FISH relies on multiplexed fluorescence in situ hybridization (FISH) and expansion microscopy to detect several RNA molecules [65], [66], [67], [68]. In situ barcoding‑based spatially resolved transcriptomics methods such as 10X Visium, Slide-seq and high-definition spatial transcriptomics (HDST) utilizes DNA barcoding to map out gene expression [68], [69], [70], [71]. Although these methods can interrogate the entire transcriptome, they are limited by their tissue resolution, which varies from 55 μm to 2 μm. ST methods have been applied to skin cancer to localize tumor subpopulations and may shed light on their role in cancer progression (Fig. 4). Using a highly multiplexed and sensitive platform called CEL-Seq2 Baron and colleagues used scRNA-seq to characterize zebrafish melanomas and identified a tumor subpopulation that expresses a stress-like transcriptional program [72]. ST was used to demonstrate that these cells are enriched in tumor regions, but not in the normal surrounding tissue [72]. After confirming the presence of stress-like cells in human melanomas and other cancer types, the authors showed that the stressed state efficiently seeds new tumors in zebrafish and is associated with resistance to MEK and BRAF inhibitors. In another study, Thrane et al. used the 10X Visium ST platform to characterize melanoma lymph node metastases, highlighting differential gene expression profiles as well as multiple melanoma signatures across the same region in a single tumor biopsy [73]. Poor patient survival was associated with greater expression heterogeneity in the area of transition between melanoma and lymphoid tissue [73]. Ji et al. similarly used the 10X Visium ST platform to demonstrate that cSCC TSKs and basal cells localize to the tumor leading edge and confirm the presence of vascular and CAF-enriched transcripts in the surrounding fibrovascular niche [37]. By incorporating a novel multi-parameter tissue imaging workflow named Pick-Seq, Nirmal et al. showed that invasive cutaneous melanoma possesses a unique cellular microenvironment at the tumor-stromal boundary that harbors different cell types supported by cytokine gradients [74]. Some of the limitations of ST are that it lacks single-cell resolution and the read quality as well as the number of transcriptomes captured is often tissue-dependent. HDST notably enables users to reach a resolution of 2µM [68,70], providing the possibility of subcellular resolution. This method currently requires specialized analytics as well as bioinformatic expertise, and is limited by low sensitivity of mRNA capture [75]. Moreover, both scRNA-seq and ST do not consistently detect genes with lower expression. To overcome these barriers, approaches such as RNA in situ hybridization, which offers a targeted but more sensitive alternative, can be used to generate spatial information at single-cell resolution. A recent study by Tran et al. developed a multimodal strategy, Spatial Transcriptomic and RNA in situ Hybridization (STRISH), to study cancer-immune cell crosstalk at the genome-wide level in keratinocyte cancers [76]. This approach starts by inferring ligand-receptor interactions in the cancer-immune cell context using scRNA-seq and ST. Cell-cell interactions are then visualized by RNA in situ hybridization and quantitated using digital droplet PCR [76]. Using this analysis pipeline, the authors demonstrated co-expression of the IL34-CSF1R and THY1-ITGAM ligand-receptor pairs in cancer cell nests and areas of immune cell infiltrate [76]. We predict that similar strategies combining different technologies to overcome the limitations of each approach in isolation will be increasingly adopted to discover and confirm critical cell-cell interactions in the cancer setting.
Bulk RNA-seq remains a valuable method for defining the coding and non-coding transcriptome; however, when applied to the skin and its attendant malignancies, scRNA-seq has revealed robust expression differences between cell types and states, providing new insight into the cellular diversity and heterogeneity that occurs in these settings (Table 1). While this approach is accompanied by its own limitations, most notably dropout events and the scalability of existing data analysis methodologies, the aforementioned studies demonstrate how scRNA-seq has enabled an improved understanding of molecular events regulating skin cancer progression. We predict that future studies will utilize single-cell transcriptomics to interrogate larger cohorts of treatment-resistant and relapsing skin cancers, with an eye towards predicting therapy response. As sequencing costs fall, scRNA-seq may also increasingly be used in the clinical setting to devise targeted personalized treatment strategies for advanced skin cancers that cannot be managed surgically.
Ankit Srivastava: Conceptualization, Writing – review & editing. Tomas Bencomo: Conceptualization, Writing – review & editing. Ishani Das: Conceptualization, Writing – review & editing. Carolyn S. Lee: Conceptualization, Writing – review & editing.
None. | true | true | true |
|
PMC9576543 | Isha Rakheja,Asgar Hussain Ansari,Arjun Ray,Dheeraj Chandra Joshi,Souvik Maiti | Small molecule quercetin binds MALAT1 triplex and modulates its cellular function | 23-09-2022 | MT: Oligonucleotides: Therapies and Applications,MALAT1,triple helix,quercetin,small-molecule binding,isothermal titration calorimetry,in silico docking,RNA FISH | The triple-helix structure at the 3′ end of metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), a long non-coding RNA, has been considered to be a target for modulating the oncogenic functions of MALAT1. This study examines the binding of quercetin—a known triplex binding molecule—to the MALAT1 triplex. By employing UV-visible spectroscopy, circular dichroism spectroscopy, and isothermal titration calorimetry, we observed that quercetin binds to the MALAT1 triplex with a stoichiometry of 1:1 and Kd of 495 ± 61 nM, along with a negative change in free energy, indicating a spontaneous interaction. Employing real-time PCR measurements, we observed around 50% downregulation of MALAT1 transcript levels in MCF7 cells, and fluorescence in situ hybridization (FISH) experiments showed concomitantly reduced levels of MALAT1 in nuclear speckles. This interaction is likely a result of a direct interaction between the molecule and the RNA, as indicated by a transcription-stop experiment. Further, transcriptome-wide analysis of alternative splicing changes induced by quercetin revealed modulation of MALAT1 downstream genes. Collectively, our study shows that quercetin strongly binds to the MALAT1 triplex and modulates its functions. It can thus be used as a scaffold for further development of therapeutics or as a chemical tool to understand MALAT1 functions. | Small molecule quercetin binds MALAT1 triplex and modulates its cellular function
The triple-helix structure at the 3′ end of metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), a long non-coding RNA, has been considered to be a target for modulating the oncogenic functions of MALAT1. This study examines the binding of quercetin—a known triplex binding molecule—to the MALAT1 triplex. By employing UV-visible spectroscopy, circular dichroism spectroscopy, and isothermal titration calorimetry, we observed that quercetin binds to the MALAT1 triplex with a stoichiometry of 1:1 and Kd of 495 ± 61 nM, along with a negative change in free energy, indicating a spontaneous interaction. Employing real-time PCR measurements, we observed around 50% downregulation of MALAT1 transcript levels in MCF7 cells, and fluorescence in situ hybridization (FISH) experiments showed concomitantly reduced levels of MALAT1 in nuclear speckles. This interaction is likely a result of a direct interaction between the molecule and the RNA, as indicated by a transcription-stop experiment. Further, transcriptome-wide analysis of alternative splicing changes induced by quercetin revealed modulation of MALAT1 downstream genes. Collectively, our study shows that quercetin strongly binds to the MALAT1 triplex and modulates its functions. It can thus be used as a scaffold for further development of therapeutics or as a chemical tool to understand MALAT1 functions.
Less than 2% of the human genome codes for proteins. However, we now know that transcription is pervasive across the genome and leads to a wide array of functional coding and non-coding transcripts regulated by cellular states. In the class of non-protein-coding RNA transcripts, long non-coding RNAs (lncRNAs) comprise the largest subset and are characterized by having a length of 200 nucleotides or more. Although there are close to 100,000 lncRNA transcripts encoded by the human genome, only a few have been functionally characterized. lncRNAs are generally poorly conserved among species at the sequence level, are expressed at low levels in cells, and are heterogeneous in function and expression pattern. They may be encoded in gene bodies or gene deserts and be transcribed from intergenic locations such as introns, exons, or enhancer regions.5, 6, 7 This variability and lack of sequence conservation seen among lncRNAs suggests a role for lncRNAs that may be sequence independent. From the ones already annotated, many lncRNAs are reported to undergo mRNA-like processing, in that they are transcribed by RNA polymerase II (Pol II), undergo 5′ end capping, are polyadenylated, and are even spliced. Several lncRNAs are known to play important roles in functions like regulation of RNA and proteins, transcriptional modulation, modification of chromatin structure, epigenetic regulation, organization of nuclear domains, gene imprinting, dosage compensation, and splicing regulation.8, 9, 10, 11, 12, 13, 14 However, despite the large number of lncRNAs that have been reported in the literature, very little is known so far about their structure-function relationships and their effects in modulating cellular states. lncRNAs interact with proteins not just through their primary sequence, but also through secondary and tertiary structural motifs. An increasing number of studies point toward the importance of RNA structure as a critical regulator of function, and such structures might be formed both locally (within the RNA) and with other intracellular molecules.15, 16, 17, 18 One such lncRNA, MALAT1 (metastasis-associated lung adenocarcinoma transcript 1), has an expression that is dysregulated in many cancers.19, 20, 21 High MALAT1 levels in patients also confer resistance to chemotherapy and radiotherapy., Among certain cancers, such as human breast cancer, it is upregulated and its expression leads to poor relapse-free survival. A role for MALAT1 has also been shown in pancreatic cancer, where high MALAT1 levels are found in metastasized cells compared with localized tumors. Similarly, high MALAT1 levels in human pancreatic ductal adenocarcinoma tissues are correlated with low patient survival. In gall bladder cancer, high levels of MALAT1 are shown to be correlated with larger tumor size, lymphatic metastasis, and poor patient survival rate. Similar associations have been reported for MALAT1 in renal cell carcinomas, bladder cancer, ovarian cancer, and a number of other cancer types.27, 28, 29 Furthermore, it has also been shown that targeting MALAT1 using antisense oligonucleotides (ASOs) leads to reduced metastasis in mouse lung cancer and the development of fewer tumor nodules after primary tumor implantation. This indicates that the targeting of MALAT1 can have potential therapeutic aspects. Among its other functions, MALAT1 in nuclear speckles alters the phosphorylation of SR (serine/ arginine-rich) proteins and modulates the cellular distribution and activity of SR splicing factors, thereby influencing alternative splicing of pre-mRNAs. MALAT1 is not polyadenylated, yet is transcribed at very high levels inside the nucleus by RNA Pol II. MALAT1 owes its nuclear stability to the triple-helix motif at its 3′ end, which sequesters the end from recognition and degradation by 3′–5′ exonucleases, which in turn leads to a higher half-life of the lncRNA., A U-rich hairpin motif termed element for nuclear expression (ENE) (which was later deemed a misnomer), was reported to modulate the stability of the lncRNA. The ENE, in reality, sequesters a genomically encoded A-rich tract (from the 3′ end of the lncRNA) to form the triple helix, which in turn protects the lncRNA from degradation., This triple helix is flanked on both ends by stem-loop structures named the P1 helix and the P2 apical helix. Targeting lncRNAs and modulating their function with small molecules has not been well studied, primarily because of limited understanding of lncRNA structures. Interestingly, for MALAT1, the blunt-ended triplex structure formed at its 3′ end is well characterized. Brown et al. deciphered the structure of the MALAT1 triplex region of the MALAT1 3′ end at a 3.1 Å resolution and implicated it in counteracting the rapid phase of RNA decay. The structure was described as an ENE (U-rich hairpin) structure, with an A-rich tract sequestered within (ENE + A). The MALAT1 transcript is not polyadenylated; rather, its A-rich tract is genomically encoded. A bipartite triple helix (that is, two shorter U⋅A-U triple helices) is formed with U⋅A-U triplets separated by an intervening C+⋅G-C triplet and a C-G doublet. The C-G doublet maintains the alignment between the Hoogsteen and the Watson-Crick strand. Similar triple-helical regions have also been observed in some viral RNAs, but with a lower number of triples and an absence of the intervening doublet in their triple-helical region. Despite these differences, there also exist structural similarities in the triplex between viral and vertebrate RNA. Up to six consecutive base triplets have been observed in nature, in contrast to 11 base triplets when chemically synthesized. The MALAT1 triplex has recently been characterized using several biophysical techniques., It is proposed that, since the triplex structure formed in MALAT1 provides stability to the RNA, selective inhibition of this triple helix may lead to a reduction in MALAT1 levels for better prognosis in cancer. A small molecule binding stably to a triplex may also preclude binding of other factors that stabilize the triple helix, leading to its opening up and affecting the degradation of the lncRNA. The triple helix is a structural motif that is susceptible to targeting with small molecules. A recent report published by Abulwerdi et al. shows the use of a small-molecule microarray of more than 26,000 molecules to identify two molecules that cause a decrease in MALAT1 RNA levels. The identified molecules were shown to physically bind with the MALAT1 triplex with unique modes of interaction. Two reports by Donlic et al. also demonstrated the screening of small molecules based on the diphenylfuran scaffold by using molecular docking to determine selectivity toward the triple helix., These studies revealed that rod-shaped ligands have the maximum specificity for the triplex structure. Using RNaseR exonuclease assays, they also showed that these molecules protect MALAT1 from exonucleolytic degradation. However, detailed understanding of the interaction of small molecules with the MALAT1 triplex has been lacking, particularly with respect to the complete biophysical characterization of such interactions. In this study, we have characterized a known triplex-binding small molecule, quercetin, for its ability to bind to the MALAT1 triple-helix structure. Quercetin is a flavonoid (subgroup flavonol) that occurs naturally in plants, many of which form a part of our diet. It is reported to be an antioxidant, with anti-inflammatory and anticarcinogenic properties. It has a generally recognized as safe (GRAS) status and is safe for animal and human consumption in low doses when taken in purified form. Further, quercetin has been reported to bind the RNA U-A-U triplex. However, there is no report that describes the binding of quercetin to a naturally existing triplex that is biologically relevant, like that of MALAT1. We have studied the interaction between quercetin and the MALAT1 triplex using different biophysical experiments. We have also checked the levels of MALAT1 upon treatment with quercetin in cells and in nuclear speckles by quantitative real-time PCR and small-molecule fluorescence in situ hybridization (FISH) techniques, respectively. Further, we have studied the effect on the splicing of genes upon MALAT1 perturbation by quercetin. Taken together, our study reveals that quercetin binds the MALAT1 triplex selectively and modulates its biological functions.
We employed circular dichroism (CD) to assess the structural aspect of triplex formation using the full-length 94-nt wild-type sequence. We also used a truncated 73-nt sequence as a control in our study, where 21 nt of the A-rich strand at the 3′ end of the sequence have been omitted so that the canonical triplex structure cannot form (Figures 1A and 1B). CD spectra of both oligonucleotides were measured at 25°C at pH 7. The 94-nt as well as the control 73-nt structures exhibit the typical spectra of a nucleic acid duplex in the “A” conformation; that is, the positive band at around ∼260 nm has a magnitude similar to that of the negative band at ∼210 nm. In addition to these peaks, a weak positive peak at around 219 nm and a negative peak at ∼243 nm were seen for the 94-nt sequence, whereas a weak negative peak at ∼243 nm was observed for the 73-nt sequence. The 94-nt sequence showed a larger magnitude band at 210 nm compared with the 73-nt, which may be characteristic of triple-helix formation (Figure 2A). UV melting experiments were conducted at pH 7.0 and were used to characterize the helix-coil transition of the structured RNAs under these conditions. Figure 2B shows typical melting curves for both oligonucleotides. Melting curves for both follow the characteristic sigmoidal behavior for the unfolding of a nucleic acid helix. Reversible melting and refolding profiles were obtained for both oligonucleotides without any hysteresis. UV melting profiles for both oligonucleotides show two transitions, a broad transition at lower temperature and a sharp transition at higher temperature. Plots of the first derivative of the absorbance at 260 nm with respect to temperature are presented in Figure 2C. The derivative curve for the 94-nt oligonucleotide indicated that there were two resolved transitions with peaks at 58°C and a high-temperature transition at 78°C (Figure 2C). The derivative curve for the 73-nt sequence showed similar behavior, with two resolved transitions with peaks at 45°C and a high-temperature transition at 76°C. The high melting temperatures for the 94-nt as well as the 73-nt sequence (76°C and 78°C, respectively) correspond to the denaturation of the Watson-Crick base-paired duplex from the P2 region into single strands. This is consistent with the observation made earlier under similar conditions. The transition at 58°C for the 94-nt oligonucleotide structure is due to the displacement of the Hoogsteen base-paired strand, as was reported earlier. Transition at 45°C for the 73-nt oligonucleotide is due to the 5-bp P1 helix. Ageeli et al. have reported similar melting behavior in that melting of the P1 helix occurs together with the triplex. Taken together, from these observations it may be concluded that the 94-nt sequence forms a stable triplex (hereafter termed the MALAT1 triplex) flanked by two duplexes, P1 and P2, and the 73-nt oligonucleotide, which lacks the A-rich tail, forms a stable duplex (hereafter termed the control duplex) under the conditions used in this study.
UV-visible (UV-vis) absorption spectroscopy is routinely used to gain insight into the interaction between a ligand and a nucleic acid structure. To understand the binding behavior between quercetin (Figure 1C) and the MALAT1 triplex, UV spectra of this ligand in the absence and presence of the MALAT1 triplex and the control duplex were obtained at 25°C. Figure 3 shows the changes in the absorption spectrum of 2.5 M quercetin when titrated against increasing amounts of triplex up to 1 M. Quercetin shows an absorbance spectrum with a positive peak at 369 nm. Upon addition of the MALAT1 triple-helix structure, a 23% hypochromic shift (Figure 3C) was observed along with a red shift of 9 nm from 369 to 378 nm (Figure 3A). In contrast, only a 10% hypochromic shift (Figure 3C) along with a red shift of 3 nm from 369 to 372 nm (Figure 3B) was observed when the control duplex was added. A higher change in hypochromicity and red shift in the absorption spectrum of quercetin in the presence of the MALAT1 triplex, in comparison with the changes in the presence of the control duplex, indicate that quercetin binds to the MALAT1 triplex stronger than to the control duplex. Hence, quercetin shows binding to the MALAT1 triplex, demonstrated by the appearance of isosbestic points in the UV spectra. Similar observations in the binding of flavonoids to an RNA triplex were also made by others.,
Advances in the sensitivity of isothermal titration calorimetry allow for more accurate determination of thermodynamic parameters such as the binding enthalpy change (ΔH), Gibbs free energy change (ΔG), entropy change (ΔS), and number of binding sites (n). Therefore, this technique was used to characterize the thermodynamic profiles of the MALAT1 triplex and quercetin complex formation. Preformed triple helix and control duplex were titrated with quercetin at 25°C (Figure 4), and an isothermal titration calorimetry (ITC) binding isotherm was generated. ITC was also performed to determine heat of dilution of quercetin, where identical amounts of quercetin were injected into the buffer alone (Figure 4A). The representative raw ITC profiles resulting from the titrations of triplex and control duplex with quercetin are shown in Figures 4B and 4C, respectively. Each of the injections of quercetin into MALAT1 triplex solution produced a large exothermic heat signal (Figure 4B), which continued to decrease until the 12th injection and then became similar to the titration of quercetin into buffer (Figure 4A). In contrast, when quercetin was injected into the control duplex solution, exothermic heat signals for each and every injection were more or less the same and also were similar to the heat of dilution (Figure 4C). The resulting corrected injection heats plotted against the respective molar ratios are shown in Figure 4D. In Figure 4D, the data points reflect the experimental injection heats, while the solid line reflects the calculated fit of data using a single site binding model. We observed that quercetin binds to MALAT1 triplex with a stoichiometry of 1:1 with the Kd in the range of 495 ± 61.6 nM, accompanied by negative changes in free energy (−8.61 kcal mol−1), indicating that the interaction was spontaneous. Further investigation showed that the negative free energy results from a net balance of favorable entropy (+7.08 cal mol−1 K−1) and favorable enthalpy contributions (−6.5 ± 0.184 kcal mol−1). A favorable binding enthalpy is associated with the presence of strong hydrogen bonds and van der Waals interactions between quercetin and triplex. A favorable binding entropy is usually associated with the release of bound water molecules upon complex formation.
The interaction of the quercetin molecule with the core ENE hairpin and A-rich tract from MALAT1 (PDB: 4PLX) was studied using in silico docking. The search-space encompassing the complete structure revealed that the ligand predominantly binds to major grooves (Figure 5A). The most efficient binding pose of quercetin was found to interact with the binding energy of −5.98 kcal mol-1 (Figure 5B). Atomistic detail of the docked conformation revealed six non-covalent electrostatic interactions between the ligand and the nucleotides of the MALAT1 triplex helix crystal structure (Figure 5C), where the O5 and O7 atoms of quercetin participated in two electrostatic bonds each, while atoms O1 and O6 participated in one bond each. The six non-covalent bonds can be further categorized as five of them being “moderate, mostly electrostatic,” while one is “weak, electrostatic.” The nucleotides involved are U58, U59, G61, C62, and A86 (Figure 5B). The nucleotides U59 and A84 and A85 participated also in the hydrophobic interaction, as depicted in Figure 5C (with red sun-flared semicircle). There were four oxygen-oxygen interacting pairs and two oxygen-nitrogen non-covalent interaction pairs found in the quercetin-MALAT1 triplex complex. As shown in Figure S1, there was a mismatch between the 76-nt sequence of the crystal structure (4PLX, used for docking) and the 94-nt MALAT1 triplex sequence. The alignment shown highlights the missing regions in the PDB structure. All the nucleotide positions mentioned above for docking reflect the numbering of the 94-nt MALAT1 triplex sequence. To assess the effect of quercetin binding on the MALAT1 triplex, we used temperature-scanning UV experiments to measure the melting temperatures of the MALAT1 triplex bound to quercetin. As is shown in Figure S2, quercetin binding does not appreciably influence the melting temperature of the MALAT1 triplex.
Recently, utilizing a novel small-molecule microarray strategy, Abulwerdi et al. identified two small molecules that bind very selectively to MALAT1 triplex. They observed reduced MALAT1 expression in HEK293 T cells on treatment with these compounds. As observed from our study, quercetin also binds strongly to the MALAT1 triplex, so we investigated the MALAT1 expression level in MCF7 cells upon treatment with quercetin. After treatment of cells with quercetin at concentrations ranging from 10 nM to 16 μM for 24 h, the total RNA was isolated (TRIzol, Invitrogen) and qRT-PCR was carried out for MALAT1 (Figure 6A). The U6 snRNA level was used to normalize the expression levels of MALAT1. Indeed, a significant and sustained decrease in MALAT1 levels was observed in cells treated with quercetin compared with the cells treated with 1% DMSO that served as controls. A 50% decrease in MALAT1 RNA levels was observed when cells were treated with 1 μM quercetin. As a comparison, we tested for the levels of NEAT1, which is another lncRNA and is encoded by the same chromosome as MALAT1 and also forms a triple helix similar to the one found at the 3′ end of MALAT1. Using the same protocol of qRT-PCR (and normalization), we observed no change in the levels of NEAT1 lncRNA, indicating that the effect of quercetin on MALAT1 is specific (Figure 6B). Further, to query whether the difference in MALAT1 was the result of a physical interaction between the small molecule and the RNA or a result of transcription level alteration, a transcription-stop assay was performed using α-amanitin. After 12 h of treatment with the transcription inhibitor, when MCF7 cells were exposed to quercetin, a chase experiment for MALAT1 levels using RT-PCR revealed that at 20 min post-quercetin treatment, MALAT1 RNA levels were brought down to half of the initial amount (Figure 6C). This indicated that the reduction in MALAT1 level was not a result of an effect on transcription; rather, it was due to a physical interaction between the molecule and the RNA.
Subcellular localization determination via qRT-PCR and FISH demonstrated that MALAT1 localizes to both the nucleus and the cytoplasm. However, it is more abundant in the nucleus and forms a characteristic speckled pattern. From the qRT-PCR studies, we observed that MALAT1 levels in cells are suppressed by quercetin treatment. However, it was not indicated whether MALAT1 suppression by quercetin affects MALAT1 located in both the nucleus and the cytoplasm or whether suppression occurs only in the nucleus or in the cytoplasm. As MALAT1 is mostly expressed in the nucleus, quercetin requires entry into the nucleus to suppress MALAT1 expression effectively. However, in drug-resistant cancer cells, only a small percentage of drugs finally reach the nucleus, as cancer cells develop intracellular resistance mechanisms to limit the access of cytosolic drugs to the nucleus. To determine if quercetin is able to reach the nucleus and suppress MALAT1 effectively, we examined the abundance of MALAT1 using single-molecule RNA FISH, along with co-immunostaining using an antibody specific for the SR splicing factor SC35. In line with the previous reports, our FISH experiments along with co-immunostaining showed that MALAT1 is located in the nuclear speckles (Figure 7D). However, upon treatment with quercetin, the FISH signals decrease, indicating a decrease in MALAT1 concentration in the nuclear speckles (Figures 7A and 7B). These experiments clearly demonstrate that quercetin is able to localize in the nucleus and target MALAT1, thus reducing the MALAT1 concentration in speckles. The number of MALAT1 puncta reduces by ∼50% on treatment of cells with 1 μM quercetin (Figure 7C), which is in line with what was observed by qRT-PCR.
One of the major roles of MALAT1 in regulating cell fate has been through its modulation of RNA splicing through SR protein phosphorylation. The control of alternative splicing has been studied extensively, and large ongoing efforts have been made to develop synthetic modulators of alternative splicing that can control cellular function, resulting in potential novel therapeutics. Since the treatment with quercetin reduced MALAT1 expression levels, we asked whether the altered expression of MALAT1 in the presence of quercetin may affect the alternative splicing scenario that is under the control of MALAT1 coordination. In previous studies, MALAT1 perturbation has been performed using sequence-specific knockdown approaches leading to discovery of diverse molecular pathways regulated by MALAT1. To explore the effect of small-molecule-induced MALAT1 changes on the transcriptome, we treated MCF7 cells with 1 μM quercetin and performed whole-RNA sequencing on two biological replicates. We were able to obtain only 168 transcripts (Table S1) that were significantly differentially regulated between 1% DMSO (control)- and quercetin-treated samples (Figure 8A). Surprisingly, we found that MALAT1 perturbation through quercetin, too, recapitulated major RNA regulatory pathways (ribosome biogenesis, RNA transport, etc.) at the level of transcript expression (Figure S3). This indicates that the pattern of gene expression in quercetin-treated MALAT1 knockdown cells is similar to that in cells where MALAT1 is specifically downregulated using siRNA. Alternative splicing being a significant outcome of MALAT1-mediated SR protein phosphorylation, we investigated if quercetin treatment recapitulated a similar regulation at the transcript level. We observed more than 100,000 alternative splicing local events each in the control and the quercetin-treated samples (Figure 8B), which were classified as events on mapping to Gencode data. This high number of events come up due to the specific annotation used for exon, intron boundaries, etc., in Gencode. However, 669 events were observed when we looked at only the significantly differential alternative splicing events between control and quercetin-treated cells. These may be classified into six different categories based on the type of event (Figure 8B) (Table S2). Pathway analysis for this set of alternatively spliced genes showed a prominent association with RNA-regulatory pathways (spliceosome, spliceosomal complex, mRNA splicing, etc.) and mitochondrial pathways (Figures S4A and S4B). This indicates that quercetin treatment-led MALAT1 downregulation affects the splicing of genes involved in RNA regulation as well as mitochondrial processes. Sashimi plots of a few obtained transcripts showing their differently spliced events are shown in Figure S5. There were some splicing events obtained that appeared exclusively in the control condition, and not in the treatment, and vice versa. We understood these events to appear as a result of quercetin-induced MALAT1 downregulation. Closer inspection of these mutually exclusive splicing events between the two conditions revealed an enrichment of exon-skipping events, alternative first exon usage, and alternative 5′ exon events in the quercetin-treated samples (Figure S6). These lead to a decrease in alternative 3′ exon usage, intron retention, and alternative last exon usage. Overall, we can conclude that quercetin treatment in MCF7 led to gross changes in alternative splicing, possibly acting through the MALAT1 splicing axis.
We asked whether quercetin treatment causes any global change in the lncRNA transcriptome. Our RNA-sequencing data analysis identified 168 transcripts (Table S1) that were significantly differentially regulated (Figure 8A), of which only 6 were lncRNAs (RP5-837J1.3, MIR181A1HG, FAM225B, AC138035.2, RP11-497H16.9, RP11-15L13.5). We checked the sequences of these lncRNAs for putative U-A-U triplex-forming motifs and found that no such motifs were present in any of these six lncRNA transcripts, suggesting that these lncRNAs (unlike MALAT1) are probably not targets of quercetin-triplex interaction and may be an outcome of other non-specific effects.
lncRNAs have recently been well studied and have emerged as important regulators of gene expression. They can fold into complex structures and interact with proteins, thus modulating biological activity inside a cell. Generally, lncRNAs are structurally conserved, and these structural motifs are crucial for maintaining their function. MALAT1 is an extensively characterized multifunctional lncRNA that is involved in transcriptional regulation, alternative splicing, microRNA sponging, and many more functions. Recent studies indicate that elevated MALAT1 expression levels promote cell proliferation and are correlated with poor overall survival in various cancer types. It has also been shown that an elevated MALAT1 level correlates with larger tumor size, advanced tumor stage, and overall poor prognosis. Thus, selective inhibition of MALAT1 can be a strategy to halt its oncogenic activity from a therapeutic perspective. The 3.1-Å-resolution crystal structural study has recently identified a triple helix at the 3′ end of MALAT1 containing stacks of five and four U⋅A-U triples separated by a C+⋅G-C triplet and C-G doublet, extended by two A-minor interactions and a 3′-terminal A-rich tract. Similar triple-helix-forming structural elements were also observed in other non-coding and genomic RNAs of diverse viruses. Targeting these intramolecular triple-helical structures with small molecules could thus provide valuable insights into the roles of these structures in lncRNA functions and will consequently help to establish these structures as small-molecule-druggable targets. It also offers opportunities to therapeutically modulate numerous cellular processes. UV melting studies showed that the MALAT1 triplex melts in two sequential well-resolved steps. The high-temperature transition is due to the melting of the apical P2 stem and the low-temperature transition to the thermally induced release of the third strand from the triplex to form the core duplex and the free single strand. These observations are consistent with the recent data reported by Ageeli et al. The U-rich third strand separation from the triplex occurs at 59°C ± 1.0°C, while the duplex strand denaturation occurs at 78°C ± 1.0°C in 10 mM sodium cacodylate (pH 7.0), 150 mM Na+, and 0.5 mM MgCl2. We also observed two melting transitions for the control duplex. The high-temperature transition for the control duplex occurred within the same temperature range as observed in the case of the MALAT1 triplex, confirming that the transition corresponds to apical P2 stem melting. However, low-temperature transition for the control duplex happened at 45°C ± 1.0°C, which is 14°C lower than the triplex melting. The amplitude of this transition is around 40% lower than the triplex transition, confirming that the transition is due to melting of the short P1 helix. Melting of the P1 helix at lower temperature was also noticed by Ageeli et al. There are a considerable number of small molecules known in the literature that specifically bind to triplex structures. We chose quercetin to determine biophysical insights into the interaction between quercetin and MALAT1 triplex because quercetin is known to bind specifically to RNA triplexes and is known for its potential pharmacological relevance., The results presented above clearly indicate that quercetin binds to the MALAT1 triplex. The binding of quercetin to the MALAT1 triplex resulted in a modest red shift (9 nm) as well as a modest hypochromicity (23%) at λ366nm. In general, small molecules binding to nucleic acid structures through an intercalation mode are generally accompanied by large red shifts and high hypochromicities. On the other hand, in the case of minor groove binding, a hypochromic effect can be observed, but the λmax absorption band virtually does not change.55, 56, 57 In the present study, since we observed modest shifts in the UV spectra of quercetin upon binding with the MALAT1 triplex, it can be argued that the binding mode is different from intercalation or minor groove binding. Currently, there are no high-resolution structures of a triple helix in complex with the aforementioned compounds, although various binding modes are predicted, such as intercalation and minor groove binding. The perturbation in the UV spectrum with hypochromic effect and red shifts upon increase in triplex concentration can be interpreted as an indication that the bound ligand is in a less polar environment. Our computational study supports the observation that the ligand predominantly binds to major grooves. This was further confirmed by the UV melting study, where we observed that the Tm of the triplex motif does not change in the presence of quercetin. Although not many small molecules are known that bind to RNA tertiary structures through a major groove, there are many natural products, such as pluramycins, aflatoxins, azinomycins, leinamycins, amino sugars, and neocarzinostatins, that are known to bind to DNA tertiary structures through the major groove. It has been shown that major groove binders do not stabilize nucleic acid tertiary structures, corroborating our observation. ITC experiments showed that, while quercetin binds to the MALAT1 triplex with a Kd of 495 ± 61.6 nM, it does not bind to the control duplex, hence confirming the selectivity of quercetin toward the MALAT1 triplex. Quercetin binding to MALAT1 triplex at 25°C yielded a negative enthalpy change (ΔH) of −6.5 ± 0.184 kcal mol−1 and an entropy change (ΔS) of +7.08 cal mol−1 with an overall favorable free energy (ΔGobs) of −8.61 kcal mol−1. The favorable binding of quercetin comes from a combination of enthalpy and entropy terms. The negative values of ΔH and positive values of ΔS are consistent with the characteristics of a combination of van der Waals, hydrophobic, and electrostatic interactions in the binding process. However, a large amount of entropy gain (TΔS = 2.11 kcal mol−1) was observed. Upon analysis of large scale data from existing experimental observations on thermodynamic features of drug-DNA interactions, Chaires et al. in previous reports have found that intercalation is primarily enthalpy driven, whereas groove binding is entropy driven. Correlating with this observation, the large entropy gain in the binding process indicates that quercetin interacts with MALAT1 triplex through the major groove binding mode. We observed an overall favorable free energy (ΔGobs) of −8.61 kcal mol−1 upon interaction, which is also more favorable than that obtained in the case of intercalation. For example, binding free energies (ΔGobs) of −6.7 and −7.49 kcal mol−1 were determined for ethidium and propidium, respectively, at 25°C in a buffer containing 0.2 M NaCl., Therefore, an additional contribution to the favorable binding free energy could indicate the formation of H bonds, in addition to the hydrophobic and van der Waals interactions, which generally happen in cases of interaction binding for ethidium and propidium. All these findings are also largely consistent with our computational observation. We found that a large network of hydrophobic and electrostatic interactions stabilize the quercetin-MALAT1 complex. Our real-time qPCR data clearly indicate that quercetin is capable of targeting the triplex and bringing pharmacological perturbation to MALAT1 expression levels. We observed almost 50% downregulation in the MALAT1 levels in MCF7 cells upon treatment of 1 μM or higher quercetin without affecting the expression of NEAT1, which contains a structurally similar ENE triplex compared with MALAT1, thereby confirming the structural selectivity of quercetin toward the MALAT1 triplex. Further, by transcription-stop assay, a physical interaction between the small molecule and the RNA is indicated (over a perturbation in transcription) as a probable cause of MALAT1 RNA level decrease. Decreased levels of MALAT1 were further supported by the observation of depletion of MALAT1 in nuclear speckles in MCF7 cells upon treatment with quercetin. As MALAT1’s role in alternative splicing has been discussed, any perturbation in the levels of MALAT1 by such targeting of the triplex by small molecules could be expected to manifest into visible effects in alternative splicing. Indeed, we observed successful splicing perturbation events in more than 600 cases due to quercetin treatment (leading to MALAT1 level downregulation). Moreover, quercetin exhibits minimal effects on the lncRNA transcriptome. The role of small-molecule-mediated targeting of triple helixes is gaining momentum with the recent focus on identification of highly specific lead molecules., A few comprehensive studies have reported the development of novel tools for targeting the triple helix in the lncRNA MALAT1. These studies have largely relied upon unbiased screening of large libraries of compounds using the MALAT1 triple helix as a target. Although these lead molecules are expected to bind to the MALAT1 triple helix with high specificity, detailed studies under complex intracellular conditions will further ascertain the utility of these compounds, but these are going to be restricted due to their lack of commercial availability. Quercetin binds to the MALAT1 triplex with dissociation constants in the nanomolar range and is as strong as the newly discovered MALAT1 triplex binders. Commercial availability of quercetin may help to overcome such restrictions. To sum up, this study entails an integrative approach wherein we combine both biophysical methods and molecular assays to demonstrate quercetin as a scaffold to target a triplex that is present at the 3′ end of the MALAT1 lncRNA. The spectroscopic and calorimetric studies have unanimously shown efficient and selective binding of quercetin to MALAT1 triplex. Upon binding to the triplex of MALAT1, quercetin decreases MALAT1 levels, which in turn modulates alternative splicing for certain genes. Overall, our study helps to establish the triplex that is present in the 3′ end of MALAT1 as a druggable target.
Quercetin (3,3′,4′,5,7-pentahydroxyflavone, 2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxy-4H-1-benzopyran-4-one, C15H10O7) was obtained from Sigma (Q4951, purity ≥95%, solid) and used without any further purification. All other reagents were of analytical grade. MilliQ water was used throughout all the experiments. The concentration of quercetin was determined by measuring the absorbance at 370 nm using the molar extinction coefficient of 1.4 × 104 M−1 cm−1. The lyophilized and HPLC-purified 94-nt (sequence: 5′-GGAAGGUUUUUCUUUUCCUGAGAAAACAACACGUAUUGUUUUCUCAGGUUUUGCUUUUUGGCCUUUUUCUAGCUUAAAAAAAAAAAAAGCAAAA-3′) and 73-nt control ENE (sequence: 5′-GGAAGGUUUUUCUUUUCCUGAGAAAACAACACGUAUUGUUUUCUCAGGUUUUGCUUUUUGGCCUUUUUCUAGC-3′) unmodified RNA oligonucleotides were obtained from Genscript (Biotech Desk, India) and dissolved in nuclease-free water. The solution concentrations of each of the oligonucleotides were determined optically at 260 nm and 25°C using the molar extinction coefficients (M−1 cm−1 of strands) of 955,700 for the 94-nt oligonucleotide and 713,000 for the 73-nt oligonucleotide. The folded triplex was obtained by heating the solutions to 100°C for 5 min and then keeping on ice for 5 min, followed by a slow return to room temperature. All biophysical experiments were performed in a 10 mM sodium cacodylate buffer (pH 7.0) containing 150 mM NaCl and 0.5 mM MgCl2 at 25°C, unless otherwise specified. All experiments were replicated three times.
Absorption spectra were measured on a Cary 100 UV-vis double-beam spectrophotometer with a 1-cm path length quartz cell. UV-vis absorption titrations were carried out by the stepwise addition of triplex and control duplex solutions to a quartz cuvette of 1-cm path length containing 700 μL of 0.5 μM quercetin. The triplex was titrated into the cuvette from an initial concentration of 0.025 μM up to a concentration of 1.0 μM. The absorbance spectra were recorded from 300 to 500 nm at 25°C. Experiments were replicated thrice.
Absorbance vs. temperature profiles (melting curves) for triplex (1 μM) and control duplex (1 μM) were measured at 260 nm with a thermoelectrically controlled Cary 100 (Varian) spectrophotometer. A temperature range of 25°C–95°C was used to monitor the absorbance at 260 nm at a heating/cooling rate of 0.25°C min−1. First derivatives of the melting profiles were then computed using Origin 7.5 software (OriginLab). The melting temperatures, Tms, were determined by fitting the derivative plots using multiple peak Gaussian fitting algorithms within the software. All reported Tms are the average of at least three independent measurements and within ±1°C. Experiments were replicated thrice.
CD spectra were recorded in a JASCO 815 spectropolarimeter equipped with a thermoelectrically controlled cell holder and a cuvette with a path length of 1 cm. CD spectra for the triplex and control duplex (both 2 μM) were recorded between 220 and 350 nm at 25°C, and the spectrum obtained was the average of three scans.
ITC measurements were performed using a Microcal PEAQ-ITC (Malvern Panalytical) instrument. The triplex (6 μM) was kept in the sample cell, and quercetin (100 μM) in the same buffer (with 2% final DMSO concentration) was placed in a syringe of volume 40 μL. Quercetin was added sequentially in 2-μL aliquots (for a total of 16 injections, 4 s duration) into the cell with 150 s of spacing time. All experiments were performed at 25°C. The injections were done from a stirring syringe rotating at 750 rpm, with detection on high feedback. The heat of dilution was determined in parallel experiments by titrating a quercetin solution of the same concentration into the same buffer in the cell. The generated heat burst curve (microwatts per second) was integrated with respect to time to give total heat per injection. The data were fit according to one set of site parameters using MicroCal PEAQ ITC Analysis software.
Virtual docking of quercetin was performed against the core ENE hairpin and A-rich tract from MALAT1 (PDB: 4PLX) using Autodock 4. The ligand was drawn using Marvin Sketch (https://www.chemaxon.com), a computational tool for drawing three- and two-dimensional chemical structures. The ligand’s structure was randomized and minimized prior to docking. A blind docking study was performed for the ligand wherein the complete receptor lncRNA was considered for search-space. Two hundred GA (genetic algorithm) runs were performed and the conformations with the 2.0 cluster tolerance from the 200 independent runs were selected for the cluster analysis. The best-docked position was further minimized using GROMACS 5.14 with 50,000 minimization steps and 0.01 step size. The analysis was performed using Autodock ADT and UCSF Chimera and LigPlot+.65, 66, 67
In cellulo experiments were performed in the MCF7 breast cancer cell line (procured from the European Collection of Authenticated Cell Cultures [ECACC]). MCF7 was maintained in DMEM (GIBCO) with 10% FBS (GIBCO) at 37°C with 5% CO2. Quercetin in solid (powder) form was procured from SIGMA (Q4951; HPLC purified with >95% purity) and reconstituted in 100% DMSO. All experiments had volume-equivalent vehicle (DMSO)-treated cells as controls. The final concentration of DMSO was kept at 1% or below in all cell culture experiments (referred to as 1% DMSO). Fresh quercetin solution was prepared for each experiment.
Total RNA was isolated from approximately 500,000 cells using Ambion TRIzol (cat. no. 15596018) reagent. RNA was quantified using the Nanodrop spectrophotometer. Only RNA with ideal ratios (A260nm/A280nm > 1.8 and A260nm/A230nm > 2.0) were processed for qPCR. DNase digestion and cDNA preparation were performed using the Qiagen Quantitect reverse transcription kit for 1 μg of RNA in a 20-μL reaction volume. Gene-specific reverse primers were used instead of the random reverse primers provided in the kit. RNA was diluted two times in nuclease-free water (AMBION) before use (to get 25 ng per reaction) in qRT-PCR. cDNA was amplified using the SYBR Green master mix (TaKaRa Tli RNase H PLus) in a Roche LightCycler 480 instrument with compatible reaction plates. Cycling conditions for qRT-PCR were as follows: 95°C for 3 min followed by 40 cycles of 95°C for 10 s, 60°C for 30 s, and 72°C for 30 s. Three technical replicates were set on the qPCR plate. Amplification of MALAT1 lncRNA (NR_002819.4) was done to produce a 217-bp amplicon from the 3′ end of MALAT1, and U6 snRNA (NR_004394.1) was amplified to give a 93-bp amplicon under the same conditions, as a control. A no-RT control was used to check for any residual DNA contamination, which may contribute toward amplification. A high Ct value (different by >10 Ct) was obtained each time to indicate no residual DNA in the sample. A PCR negative control (with no cDNA) was also used to check for any contamination of reagents or during setting up the reaction. A single melting curve was observed at the end of 40 cycles, which indicated the generation of a single species of product. Fold change for the analysis was performed using the 2−ΔΔCt method. Three biological replicates were used to calculate average fold change.
Sixty percent confluent MCF7 cells (maintained as described above) were treated with 10 μg/mL α-amanitin (A2263, Sigma Aldrich) for 12 h in a 12-well plate format. Following this, the cells were treated with 1 μM quercetin or volume-equivalent DMSO (as vehicle control). At fixed time points post-treatment, quercetin-treated as well as DMSO-treated cells were washed with PBS and harvested in Ambion TRIzol for RNA isolation, followed by qPCR using protocols as described above. Results presented are an average of three biological replicates.
Total RNA was isolated from MCF7 cells using Ambion TRIzol reagent. RNA was quantified using the Qubit RNA BR Assay. RNA purity was checked using QIAxpert, and RNA integrity was assessed on TapeStation using RNA ScreenTapes. The NEB Ultra II Directional RNA-Seq Library Prep kit protocol was used to prepare the libraries for total RNA sequencing (NEB cat. no. E7760L). Ribosomal RNA (which constitutes approximately 95% of the total RNA population) was first removed from 200 ng of total RNA by using biotinylated, target-specific oligos combined with Ribo-Cop rRNA removal beads. Following purification, the ribo-depleted RNA was fragmented using divalent cations under elevated temperatures. The cleaved RNA fragments were copied into first-strand cDNA using reverse transcriptase. Second-strand cDNA synthesis was performed using DNA Pol I and RNase H enzyme. The cDNA fragments were then subjected to a series of enzymatic steps that repair the ends and tail the 3′ end with a single A base, followed by ligation of the adapters. The adapter-ligated products were then purified and enriched using the following thermal conditions: initial denaturation at 98°C for 30 s; 13 cycles of denaturation at 98°C for 10 s, annealing at 65°C for 75 s; final extension at 65°C for 5 min. PCR products were then purified and checked for fragment size distribution on TapeStation using the High Sensitivity D1000 ScreenTape assay (Agilent, cat. no. 5067-5584). Prepared libraries were quantified using the Qubit High Sensitivity Assay (Invitrogen, cat. no. Q32854) and were pooled and diluted to the final optimal loading concentration before cluster amplification. The clustered flow cell was loaded on the Illumina NovaSeq 6000 instrument to generate 150-bp paired-end reads.
MCF7 cells were seeded onto coverslips introduced into a six-well plate. One day post-seeding, the cells were treated with quercetin. Twenty-four hours after treatment, the cells were washed with PBS and then fixed and permeabilized for immunofluorescence staining with SC35 antibody (Abcam Ab11826) in a humid chamber overnight. The next day, the cells were washed and incubated with a secondary antibody labeled with Alexa Fluor having an excitation wavelength of 488 nm (Thermo Fisher A11011). To probe for MALAT1, these cells were further fixed and permeabilized according to the Stellaris FISH protocol (LGC Bioresearch Technologies) and then hybridized with the stellaris MALAT1 probe, which was conjugated with Quasar 570 dye (cat. no. VSMF-2210-5), in a humidified chamber overnight. The next day, the coverslip was washed and then briefly stained with DAPI before fixing onto a slide for visualization. The images were captured at 60× in three channels using the z-stacks option in the DeltaVision microscope. Images were projected for maximum intensity and then analyzed in ImageJ. The far-red channel (for MALAT1 intensity) was processed by converting to 8 bits and then subjected to auto-thresholding. Particles in the size range of 0–1,000 pixels were counted and compared between treated and untreated samples. The results presented are averaged over 100 cells per sample.
Libraries were generated using the TruSeq Stranded Total RNA Sample Prep Kit (Illumina) followed by sequencing on the Illumina NovaSeq 6000 platform (>45 million reads per condition). Raw reads were checked for quality using FastQC (v.0.11.5). Trimmomatic (v.0.39) was used to trim adaptor sequences and filter low-quality reads. High-quality reads were mapped to the Human Gencode (GRCh38.p13) reference genome and transcriptome using STAR (v.2.7.9a) aligner with RSEM (v.1.3.1) to estimate gene expression., Gene read counts from RSEM were normalized using TMM (weighted trimmed mean of M values). This was followed by differential expression analysis using quasi-likelihood methods in edgeR (v.3.30.3)., Genes showing differential expression compared with the control condition were plotted using EnhancedVolcano (v.1.6.0). Gene enrichment and pathway analysis were done using the R interface of Enrichr.75, 76, 77
Reads were aligned against the Human Gencode (GRCh38.p13) transcriptome by Salmon (v.1.4.0) in quasi-mapping-based mode. Following this, SUPPA2 was used to generate, calculate, and compare the alternative splicing events (SE, A3, A5, MX, RI, AF, AL) between the quercetin-treated and the DMSO control samples., Splicing events with delta percentage or proportion spliced-in (ΔPSI) ≥ |0.10| and p < 0.05 were considered significant differential events. Pie charts were generated using the ggpubr and ggplot2 R package.
Data were expressed as the mean ± SEM for control and experiment cases for three biological replicates. Statistical analysis was performed by two-tailed Student’s t test using MS Excel. Experimental results leading to p < 0.05 were considered statistically significant. Asterisks denote p values in the figures as ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001, respectively. Biological replicates were similar to those employed in the field. Primer sequences used in this study were as follows: MALAT1 forward primer, CTTCCTGTGGCAGGAGAGAC; MALAT1 reverse primer, CGCTTGAGATTTGGGCTTTA; NEAT1 forward primer, TCTCCTGGCTATTCCAGGCT; NEAT1 reverse primer, TAGCCACACAGTGGCAGAAG; U6 forward primer, CTCGCTTCGGCAGCACATATACT; U6 reverse primer, ACGCTTCACGAATTTGCGTGTC. | true | true | true |
|
PMC9576564 | Eman Selem,Asmaa F. Mekky,Wesam A. Hassanein,Fifi M. Reda,Yasser A. Selim | Antibacterial and antibiofilm effects of silver nanoparticles against the uropathogen Escherichia coli U12 | 23-09-2022 | Silver nanoparticles,Biofilm,Adhesion | The drug-resistant bacterial strains' emergence increases day by day. This may be a result of biofilm presence, which protects bacteria from antimicrobial agents. Thus, new approaches must be used to control biofilm-related infections in healthcare settings. In such a study, biological silver nanoparticles were introduced in such a study as an anti-biofilm agent against multidrug-resistant E. coli U12 on urinary catheters. Seven different silver nanoparticles concentrations were tested for their antimicrobial activities. Also, anti-biofilm activities against E. coli U12 were tested. Using the dilution method, the silver nanoparticles concentration of 85 μg/ml was the MIC (Minimum Inhibitory Concentration) that had excellent biocompatibility and showed significant antibacterial activity against E. coli U12. Scanning electron microscopy (SEM) confirmed that the highest efficient dose of silver nanoparticles was 340 μg/ml at 144 h that reduced adhesion of E. coli U12 to the urinary catheter. E. coli U12 cells ruptured cell walls and cell membranes after being examined using transmission electron microscopy (TEM). Thus, biologically prepared silver nanoparticles could be used to coat medical devices since it is effective and promising to inhibit biofilm formation by impregnating urinary catheters with silver nanoparticles. | Antibacterial and antibiofilm effects of silver nanoparticles against the uropathogen Escherichia coli U12
The drug-resistant bacterial strains' emergence increases day by day. This may be a result of biofilm presence, which protects bacteria from antimicrobial agents. Thus, new approaches must be used to control biofilm-related infections in healthcare settings. In such a study, biological silver nanoparticles were introduced in such a study as an anti-biofilm agent against multidrug-resistant E. coli U12 on urinary catheters. Seven different silver nanoparticles concentrations were tested for their antimicrobial activities. Also, anti-biofilm activities against E. coli U12 were tested. Using the dilution method, the silver nanoparticles concentration of 85 μg/ml was the MIC (Minimum Inhibitory Concentration) that had excellent biocompatibility and showed significant antibacterial activity against E. coli U12. Scanning electron microscopy (SEM) confirmed that the highest efficient dose of silver nanoparticles was 340 μg/ml at 144 h that reduced adhesion of E. coli U12 to the urinary catheter. E. coli U12 cells ruptured cell walls and cell membranes after being examined using transmission electron microscopy (TEM). Thus, biologically prepared silver nanoparticles could be used to coat medical devices since it is effective and promising to inhibit biofilm formation by impregnating urinary catheters with silver nanoparticles.
Nanomedicine is a relatively new trend in medicine. Metal nanoparticles have been shown to have antibacterial, antifungal, and antiviral properties. Silver nanoparticles represent the common antimicrobial agent. Due to recent technological advancements (Patra et al., 2018), silver nanoparticles have resurfaced in the medical field. Because of their low toxicity to mammalian cells and stronger antimicrobial activity, Silver nanoparticles have been used in a variety of disciplines. Silver nanoparticles are utilized for the treatment of biofilms associated with medical devices that threaten life (Bruna et al., 2021). Biofilms are formed when bacteria attach to surfaces and form structures. These biofilm formations are the bacterial natural survival strategy for invading the host (Camele et al., 2019). They are more resistant to routinely used antimicrobial treatments, making control more challenging. As a result, the infection becomes more severe (Roy et al., 2018, Shaikh et al., 2019, Muhammad et al., 2020). Bacterial resistance to antibiotics and the potential to colonize abiotic surfaces through the formation of biofilm are major causes of medical implant-associated infections, leading to prolonged hospital stays and patient mortality. Various strategies have been adopted in medical settings to prevent biofilm-associated infections (Li et al., 2021). Nosocomial infections also referred to as healthcare-associated infections (HAI), are infections acquired during the process of receiving health care that was not present during the time of admission (Rosenthal et al., 2012). The most prevalent nosocomial infection was catheter-associated urinary tract infection, which was caused by pathogens that developed biofilms on urinary catheters. Catheterization for a long time with urinary catheters causes biofilm development and pathogen adhesion on the catheters (Almalki and Varghese, 2020). New bacteria may not colonize the biofilm if they are exposed to silver nanoparticles. As a result, it's critical to look for anti-biofilm molecules that can efficiently reduce and eliminate biofilms that are associated with infections (Skóra et al., 2021). We have examined whether Silver nanoparticles could prevent harmful bacteria from forming biofilms in some hospital isolates. This study's objective was to thoroughly assess the antibacterial and anti-biofilm potential of the biologically produced silver nanoparticles against Escherichia coli. This study provides a good strategy to control biofilm formation associated with serious diseases such as urinary tract infections.
i) In this study, silver nanoparticles were synthesized using Aloe vera leaf extract and evaluated for their antibacterial and antibiofilm activities against biofilms of multidrug resistant (MDR) uropathogen. Our study showed that, silver nanoparticles exhibited significant antimicrobial and anti-biofilm activities against E. coli U12. ii) Different concentrations of silver nanoparticles efficiently reduced biofilm formation of E. coli U12on urinary catheter. Therefore, pre-coating urinary catheters with silver nanoparticles can be largely utilized as anti-biofilm agent in medical fields to control uropathogens biofilms formation. This work was conducted in vitro so that it may be further examined for practical application.
A previous collection of 50 bacterial isolates from urinary catheters was done. Morphological characters, biochemical tests, and 16S rRNA gene sequencing were used to identify these isolates. These isolates were tested towards different antibiotics and were assayed for their potential to produce biofilm (Mekky et al., 2022). As a result, E. coli U12 with the accession number MT498270 was chosen as the most MDR uropathogen used in this investigation (Fig. S1).
The technique reported by Tippayawat et al. (2016) was used to biosynthesize silver nanoparticles with certain modifications.
Aloe vera seeds were obtained from the Agricultural Research Center in Cairo, Egypt, and young leaves and branches were harvested from young adult trees.
Aloe vera leaves were washed several times with double distilled water to remove debris and particles and dried in the shade at room temperature. The leaves were washed with double-distilled water and shade dried, ground into fine powder, and stored in an air-tight container in dark at room temperature for extraction. 50 g of Aloe vera leaf powder was mixed with 100 ml of double-distilled water in 250-ml beaker. The above mixture was heated on magnetic heating stirrer at 40 °C for 10 min. The aqueous Aloe vera leaf extract was then filtered using Whatman filter paper No. 1 and stored at 4 °C and used for the preparation of silver nanoparticles (Fig. S2).
In the preparation of silver nanoparticles samples, AgNO3 (0.17 g) was first dissolved in 100 ml of deionized water and mix 95 ml with 5 ml of Aloe vera extract solution, incubate for 30 min at 60 °C. Silver nanoparticles was observed after about 30 min which was indicated by change in color of the solution from colorless to brownish yellow and red, respectively. The formation of silver nanoparticles was confirmed by UV– visible spectroscopy. The silver nanoparticles was purified by frequent centrifugation at 18,000 rpm for 25 min, washed with double-distilled water, and were redispersed in deionized water for further characterizations.
The silver nanoparticles were first examined utilizing the Rigol ultra-3660 UV–vis spectroscopy in the 400–450 nm region. The functional groups and other phytochemical components responsible for the reduction and stability of the produced nanoparticles were then identified using FTIR. To verify the existence of Ag and evaluate the crystallite structure and size, the powdered specimen was exposed to CuKα1-X ray diffractometer radiation with λ = 1.5406 A° and operated at 30 mA and 40 kV with 2θ (30°-140°). Sonication was done on silver nanopowder that was mixed with ethanol and placed on a copper grid, which was then dried before being inspected using TEM (JEOL-2100 HR).
Different concentrations of silver nanoparticles were examined for their antibacterial activity towards E. coli U12 through the observation of the growth turbidity in liquid medium (Brain heart infusion) after incubation for 24 h at 37 °C. Then, the MIC was determined as the minimum inhibitory concentration that prevented the visual growth of E. coli U12 in liquid medium. Also, the minimum bactericidal concentration was detected as the lowest concentration at which there is no growth on solid media (Parvekar et al., 2020).
Different silver nanoparticles concentrations were tested for their antibacterial effect towards E. coli U12 using the CFU technique according to the modified Chapman et al. (2013) technique with incubation of silver nanoparticles and E. coli U12 for seven days.
Brain heart infusion (37 g/L), sucrose (80 g/L), agar no.1 (10 g/L), and Congo red stain (0.8 g/L) were used to make Congo red agar medium. Congo red was prepared as a concentrated aqueous solution and autoclaved at 121 °C for 15 min, separately from the other medium constituents, and was then added when the agar had cooled to 55 °C. Different concentrations of silver nanoparticles were mixed with E. coli U12 (105 CFU/mL) and kept at 37 °C for 48 h. Each mixture (0.1 ml) was streaked on a Congo red agar plate and kept at 37 °C for 24 h. Control consists of E. coli U12 suspension without silver nanoparticles. The appearance of black, dry colonies with a crystalline surface indicates biofilm formation. Weak biofilm producers usually remained pink with occasional darkening at the centres of colonies, and the appearance of pink colonies was indicated as negative biofilm bacteria. The experiment was performed in triplicate and repeated three times (Bose et al., 2009).
The pre-coating method outlined by Gudiña et al. (2010) was used in this test for assaying the anti-biofilm effect of different silver nanoparticles concentrations towards E. coli U12 using ELISA (Central Laboratory, Faculty of Pharmacy, Zagazig University, Egypt). Microliter plate wells were coated with different concentrations of silver nanoparticles and an E. coli U12 suspension (105 CFU/mL) was added after the wells dried. The microliter plate was maintained at 37 °C for 48 h and, by using CV, the fixed biofilm was dyed for 20 min. Control consists of E. coli U12 suspension without silver nanoparticles. After solubilization of the stained bounded biofilm, the Fracchia et al. (2010) equation was used to calculate the percentage of microbial adhesion: Where AC is the optical density of the well with silver nanoparticles, and Ao is the optical density of the control well.
This method was performed in vitro to inhibit biofilm development of E. coli U12 on urinary catheters using different concentrations of silver nanoparticles. The pre-coating method outlined by Ezeonu and Kanu (2016) with modifications was used in this test. In test tubes, the pre-coated catheter segments (3 segments for each) were incubated with E. coli U12 (105 CFU/ml) at 37 °C for seven days. After that, CFU/cm2 and the degree of CV coloring on the urinary catheter were used to determine the biofilm development. Untreated catheter segments were used as a positive control.
SEM confirmed E. coli U12 biofilm formation on the urinary catheter after treatment with silver nanoparticles according to Li et al. (2015). Urinary catheter segments were taken from the above method (CFU) for up to seven days and were processed using standard techniques. The SEM technique was performed in the Mansoura laboratory of electron microscopy, Mansoura University, Egypt. Each sample was gold coated using a gold sputter and viewed by SEM (JEOL JSM 6510 IV).
The TEM technique was performed in the Mansoura laboratory of electron microscopy, Mansoura University, Egypt, and was carried out according to Mirzajani et al. (2011). An examination was carried out using a JEOL JEM-2100 TEM.
One-way ANOVA was used to analysis the data. As described by Dytham (1999), the programme SPSS, version 14 (SPSS, Richmond, USA), was used.The significance of the results was determined at p < 0.05 by using Duncan's multiple range test. Each result is the average of 3 replicates ± standard error.
The shift in solution color from greenish-yellow to brown during silver nanoparticles synthesis indicated the reduction of silver oxide salt (Fig. S3).
The characteristics of silver nanoparticles are usually apparent at wavelengths between 400 and 600 nm. The UV–vis spectra of the synthesized silver nanoparticles utilizing the aqueous extract of Aloe vera indicate a blue shift of the absorption band as the concentration of AgNO3 increases. The absorption peak for 1 mM, 2 mM, 3 mM, 4 mM, and 5 mM specimens is located between 400 and 450 nm (Fig. S4). According to the results, silver nanoparticles were produced in the extract where the Ag+ was reduced to Ago.
The FTIR spectra of the synthesized silver nanoparticles using the green technique illustrated a peak at 442 cm−1 that corresponds to hexagonal Ag symmetric bending vibrations. In addition, a peak at 878 cm−1 owing to the weak vibration of Ag (Fig. 1).
As shown in Fig. 2, the X-ray diffraction pattern of the produced silver nanoparticles was taken. All of the sample's diffraction peaks demonstrated that the silver nanocrystalline production was pure and the same as the hexagonal phase with Wurtzite structures.
By TEM examination, the crystalline properties and size of the produced nanoparticles can be estimated. Fig. 3 demonstrates photos at various magnifications (50 and 100 nm) taken with the JEOL-2100. TEM photographs exhibit particle sizes ranging from 9.26 to 31.18 nm. Otherwise, it confirmed the hexagonal structure of the synthesized silver nanoparticles.
In a liquid dilution experiment, the lowest silver nanoparticles concentration that inhibits the visible growth of E. coli U12 in liquid medium was 85 μg/ml (MIC). Also, the lowest concentration of silver nanoparticles at which no growth of E. coli U12 on solid medium was 127.5 μg/ml (MBC) (Fig. S5). Using the CFU method, the results revealed that there was a parallel decrease in E. coli U12 growth as a function of silver nanoparticles concentration (from 85 to 340 μg/ml) during time, increasing up to 7 days compared with the control (Fig. S6).
In the Congo red agar method, the results showed that all tested silver nanoparticles concentrations completely inhibited E. coli U12 biofilm formation. E. coli U12 grew as dry, crystalline black colonies in the absence of silver nanoparticles, demonstrating biofilm formation (Fig. S7). In the crystal violet assay method, there was a gradual reduction in the production of biofilm with increasing the concentrations of silver nanoparticles compared with the positive control (Figs. S8 and S9).
In CFU, there was a parallel reduction in E. coli U12 adherence to the urinary catheter as a function of silver nanoparticles concentration (from 85 to 340 μg/ml) over time, increasing up to seven days compared with the positive control (Fig. S10). Staining the bounded bacteria on urinary catheter segments with CV indicated that, with increasing the concentration of silver nanoparticles, there was a gradual decrease in the degree of crystal violet stain color that was bound with biofilms on urinary catheters (Fig. S11).
In SEM analysis, the results demonstrated that E. coli U12 adherence to urinary catheter segments was inhibited after treatment with silver nanoparticles concentrations of 85, 255, and 340 μg/ml at 24, 72, and 144 h, respectively, compared with the control (Fig. S12). In TEM analysis, Fig. (S13) indicated that there was a rupture in the cell wall of E. coli U12 cells after exposure to Sub-MIC (42.5 μg/ml) of silver nanoparticles for 24 h. Also, some cells undergo partial lysing of their cytoplasmic contents compared with control.
All of the extract's secondary metabolites are important in the reducing as well as capping mechanisms for nanoparticle synthesis (Marslin et al., 2018). In the present study, silver nanoparticles were produced using Aloe vera aqueous extract and had an absorption peak of between 400 and 450 nm. FTIR is performed as a method of confirmation for nanoparticle production and provides an overview of the rotational and vibrational modes of the compounds that already exist, which aids in the identification of functional and phytochemical compounds that are associated with silver nanoparticles reduction and stability. The broad peaks at 3434 cm−1 and 1117 cm−1, respectively, point to the occurrence of OH and C-OH stretching vibrations. According to the FTIR findings, it can be concluded that the presence of proteins, enzymes, and metabolites like carboxylic acid, polyphenols, flavonoids, and alkaloids “that persisted attached to silver nanoparticles following multiple washings” are involved in zinc ion reduction to silver nanoparticles. The existence of free amino as well as carboxylic groups that have bonded to the zinc surface could likely account for the stability of the produced silver nanoparticles. Moreover, the proteins in the medium help to stabilise silver nanoparticles by producing a coat that covers the metal nanoparticles and prevents them from clumping together (Sri Sindhura et al., 2014). In addition, the (20) peaks angles at 31.77°, 34.42°, 36.26°, 47.54°, 56.60°, 62.86°, 66.38, 67.95°, 69.09, 72.57 and 76.97° correlate to the reflection from (1 0 0), (0 0 2), (1 0 1), (1 0 2), (1 1 0), (1 0 3), (2 0 0), (1 1 2), (2 0 1), (0 0 4) to (2 0 2) crystal planes, respectively according to JCPDS 36–1451 card (John and Rajakumari, 2012). The mean crystallite size of silver was measured to be 15.22 nm utilizing Scherrer's equation, that was determined from the FWHM of the peak that is more intense corresponding to the 101 plane at 36.26°. This lies inside the size range of 9.26 to 31.18 nm, determined via TEM. It has been proven that reducing particle size improves their effectiveness as antibacterial and anticancer agents owing to the high surface-to-volume proportion (Masum et al., 2019). By decreasing the size of the silver particles to the nanoscale, the antibacterial and anti-biofilm activities of the silver were increased as the surface area of the particles increased. As a result, the level of Ag+ release is greater than that of silver particles in their elemental form. Consequently, silver nanoparticles have a better ability to adhere, penetrate, and aggregate inside the cell membrane of bacteria, resulting in a large amount of silver ions being released within the cell. The presence of water channels all over the biofilm could explain silver nanoparticles biofilm inhibitory action. These pores were important in nutrient transport, and silver nanoparticles could pass right through these pores and reveal their antibacterial action (Ansari et al., 2014, Muzammil et al., 2018). In the present study, different concentrations of silver nanoparticles (85, 127.5, 170, 212.5, 255, 297.5, and 340 μg/ml) were examined for their potential to exhibit antimicrobial effect towards E. coli U12 via observing the growth of E. coli U12 visually in liquid media and CFU/ml. The results revealed that the MIC and MBC of silver nanoparticles against E. coli U12 were 85 μg/ml and 127.5 μg/ml, respectively. The findings revealed that the CFU of E. coli U12 proliferation was proportional to silver nanoparticles concentration (from 85 to340 μg/ml) over time (up to seven days). Our results are in agreement with Rodríguez-Serrano et al. (2020). They discovered that when silver nanoparticles concentration and time increased, the proliferation of uropathogenic E. coli was reduced. Skóra et al. (2021) found that silver nanoparticles had a greater antimicrobial activity towards E. coli, S. aureus, and P. aeruginosa. Regarding the anti-biofilm efficiency, silver nanoparticles were investigated against E. coli U12 grown on CRA enriched with and without silver nanoparticles, as well as the CV assay method. When E. coli U12 was cultured without silver nanoparticles (control), it developed black, dry crystalline colonies, indicating that exopolysacharide generation (EPS) is required for biofilm creation. When the uropathogenic E. coli U12 was treated with silver nanoparticles, the bacterial growth and production of biofilms were both suppressed at all concentrations. Our findings are consistent with those of Ansari et al. (2014). They found that E. coli as well as K. pneumoniae treated with silver nanoparticles did not establish biofilm over CRA medium because the generationof glycocalyx matrix and exopolysaccharide synthesis were inhibited. In addition, the current study revealed that when silver nanoparticles concentrations increased, E. coli U12 biofilm development in microtitre plate wells decreased gradually. Our findings are similar to those of Ramachandran and Sangeetha (2017). They discovered that rising silver nanoparticles concentrations from 12.5 to100 μg/ml prevented biofilm development in Klebsiella pneumoniae, Escherichia coli, as well as Pseudomonas aeruginosa within 24 h. Also, Rodríguez-Serrano et al. (2020) showed that silver nanoparticles concentrations ranging from 7.5 to 35 mg/L reduced and impaired the development of biofilms generated by uropathogenic E. coli by 97 %. Also, the present investigation was extended in vitro to control E. coli U12 biofilm production on urinary catheter segments. Silver nanoparticles concentrations were investigated for their potential to reduce E. coli U12 adherence to urinary catheter segments using CFU/cm2 as well as crystal violet staining of adherent bacteria over urinary catheter segments. The results indicated that the adhesion of E. coli U12 to the catheter decreased linearly with a rise in the concentration of silver nanoparticles (from 85 to 340 μg/ml) over time (up to seven days). As the silver nanoparticles concentration increased, the color intensity of the CV stain that was associated with adherent bacteria on the catheter decreased gradually. Our results are compatible with previous published studies that indicated silver nanoparticles biofilm elimination potential is directly correlated with their concentrations. As the concentration of silver nanoparticles increases, the efficiency of biofilm destruction increases (Masák et al., 2014, Thuptimdang et al., 2017). Our findings showed a similar pattern, demonstrating that silver nanoparticles are dose-dependently and can be utilized to treat multidrug-resistant E. coli U12. Skóra et al. (2021) found that the efficiency of silver nanoparticles was dose-dependent, with stronger biofilm elimination at concentrations of 1 to 2 μg/mL than at 0.5– 0.125 μg/mL doses of silver nanoparticles. Kostenko et al. (2010) showed that silver nanoparticles significantly reduced CFU counts and Pseudomonas aeruginosa adhesion. The development of S. aureus and E. coli biofilms was reduced following long-term exposure to silver nanoparticles. Kumar et al. (2012) found that silver nanoparticles suppressed biofilm development in E. coli, S. aureus, Salmonella typhii, and Vibrio cholerae. Ebrahimi et al. (2018) found that, silver nanoparticles prevented the biofilm development of P. aeruginosa, A. baumannii, E. faecalis, and S. aureus with a 90 % inhibitory activity. It is very important during the current investigation to confirm the reduction of E. coli U12 adhesion following pre-coating of the urinary catheter utilizing silver nanoparticles after SEM inspection. The results revealed that pre-coating with silver nanoparticles lowered the adhesion of E. coli U12 as the concentration of silver nanoparticles increased. Transmission electron microscopy study revealed that silver nanoparticles concentration of85 μg/ml had a dramatic effect on E. coli U12 cells for 24 h. This results in the rupturing of the cell wall and cell membrane, hence the release of the cell contents. Ansari et al. (2014) showed that, after examining biofilms produced by E. coli on glass slides for 24 h with SEM, their shape had changed and silver nanoparticles at a concentration of 20 g/ml had an impact on the roughness of the cell surface. Ramachandran and Sangeetha (2017). SEM investigation explained that silver nanoparticles suppressed bacterial growth and exopolysaccharide development by K. pneumonia on glass slides for 24 h. Singh et al. (2021) reported that, after inspection with SEM, the bacterial cells of P. aeruginosa and E. coli showed changes in morphology with apparent membrane pores, intracellular content leaking, as well as cell lysis after treatment with silver nanoparticles.
From these findings, silver nanoparticles have substantial anti-microbial and anti-biofilm activities towards E. coli U12. As a result, pre-coating urinary catheters with silver nanoparticles is a good idea and can be widely used in medical applications as anti-biofilm compounds to inhibit and limit the development of bacterial biofilm communities linked to major disorders, including urinary tract infections.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. | true | true | true |
|
PMC9576582 | Xiankui Gao,Xiujuan Li,Chengan Chen,Can Wang,Yuqi Fu,ZiZhen Zheng,Min Shi,Xiaolong Hao,Limei Zhao,Minghua Qiu,Guoyin Kai,Wei Zhou | Mining of the CULLIN E3 ubiquitin ligase genes in the whole genome of Salvia miltiorrhiza | 08-10-2022 | Salvia miltiorrhiza,Gene family,Expression pattern,Phenolic acid,Tanshinone,CULLIN E3 ubiquitin ligase | CULLIN (CUL) proteins are E3 ubiquitin ligases that are involved in a wide variety of biological processes as well as in response to stress in plants. In Salvia miltiorrhiza, CUL genes have not been characterized and its role in plant development, stress response and secondary metabolite synthesis have not been studied. In this study, genome-wide analyses were performed to identify and to predict the structure and function of CUL of S. miltiorrhiza. Eight CUL genes were identified from the genome of S. miltiorrhiza. The CUL genes were clustered into four subgroups according to phylogenetic relationships. The CUL domain was highly conserved across the family of CUL genes. Analysis of cis-acting elements suggested that CUL genes might play important roles in a variety of biological processes, including abscission reaction acid (ABA) processing. To investigate this hypothesis, we treated hairy roots of S. miltiorrhiza with ABA. The expression of CUL genes varied obviously after ABA treatment. Co-expression network results indicated that three CUL genes might be involved in the biosynthesis of phenolic acid or tanshinone. In summary, the mining of the CUL genes in the whole genome of S. miltiorrhiza contribute novel information to the understanding of the CUL genes and its functional roles in plant secondary metabolites, growth and development. | Mining of the CULLIN E3 ubiquitin ligase genes in the whole genome of Salvia miltiorrhiza
CULLIN (CUL) proteins are E3 ubiquitin ligases that are involved in a wide variety of biological processes as well as in response to stress in plants. In Salvia miltiorrhiza, CUL genes have not been characterized and its role in plant development, stress response and secondary metabolite synthesis have not been studied. In this study, genome-wide analyses were performed to identify and to predict the structure and function of CUL of S. miltiorrhiza. Eight CUL genes were identified from the genome of S. miltiorrhiza. The CUL genes were clustered into four subgroups according to phylogenetic relationships. The CUL domain was highly conserved across the family of CUL genes. Analysis of cis-acting elements suggested that CUL genes might play important roles in a variety of biological processes, including abscission reaction acid (ABA) processing. To investigate this hypothesis, we treated hairy roots of S. miltiorrhiza with ABA. The expression of CUL genes varied obviously after ABA treatment. Co-expression network results indicated that three CUL genes might be involved in the biosynthesis of phenolic acid or tanshinone. In summary, the mining of the CUL genes in the whole genome of S. miltiorrhiza contribute novel information to the understanding of the CUL genes and its functional roles in plant secondary metabolites, growth and development.
Salvia miltiorrhiza is a famous Chinese medicinal plant used in medicine and health food for thousands of years (Qian et al., 2022). It has been used to treat cardiovascular and cerebrovascular diseases in many countries. So far, the S. miltiorrhiza has become a model of Chinese herbal medicine due to its characteristics of being widely and deeply studied (Shi et al., 2016; Huang et al., 2021; Zhou et al., 2021a, b; Sun et al., 2022). The active ingredients of S. miltiorrhiza include two groups: one group is diterpenoid tanshinone, including tanshinone I, tanshinone IIA, tanshinone IIB, dihydrotanshnone I and cryptotanshinone, exhibits various pharmacological activities including antioxidant, antitumor and anti-inflammatory properties; the other group is water-soluble phenolic acids, such as rosmarinic acids, salvianolic acids and lithospermic acid, functions as antibacterial, anti-oxidative and antiviral reagents. (Sun et al., 2022; Zhao et al., 2022). These components have been shown to exhibit various biological activities, including anti-tumor, anti-inflammatory, and antibacterial effects (Liu et al., 2022; Sun et al., 2022; Zhao et al., 2022). During their life courses, plants are repeatedly exposed to various abiotic stresses such as drought, salt, and low temperatures, resulting in oxidative damage and adverse effects (Gupta et al., 2020; Smokvarska et al., 2020; Wang et al., 2021). Plants have evolved complex, efficient mechanisms to cope with unfavorable environment. The response of transcriptional regulation, post-transcriptional modification, epigenetic regulation, and secondary metabolism to abiotic stress has been studied in previous studies in S. miltiorrhiza (Marino et al., 2013; Dou et al., 2021; Karre et al., 2021; Tong et al., 2021; Wang et al., 2021). But, ubiquitination modification and degradation of functional proteins regulating the synthesis of medicinal active substance in S. miltiorrhiza is still unclear. Ubiquitination is a crucial post-translational modification (Chen et al., 2021; Wang et al., 2021). The ubiquitin/26S proteasome system (UPS) is a pervasive and effective route for protein removal in eukaryotes. UPS include ubiquitin (Ub), ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzyme (E2), ubiquitin-ligating enzyme (E3), and the 26S proteasome (Trujillo and Shirasu 2010; Chen et al., 2021). Ub is bound to specific proteins and functions in target proteins’ degradation by the E1–E2–E3 multi-enzyme cascade, while E3 are thought to be the key factor to define substrate specificity during the process of ubiquitination and degradation (Richburg et al., 2014; Serrano et al., 2018). E3 were classified into four main types as U-box, HECT (Homology to E6-Associated Carboxy-Terminus), RING (Really Interesting New Gene) and Cullin–RING ligases (CRLs) through their reaction mechanism and subunit compositions (Vierstra 2009). CUL proteins are molecular scaffolds and play a crucial role in ubiquitin-mediated post-translational modification of cellular proteins. CUL proteins are also present in model organisms like Drosophila melanogaster, Caenorhabditis elegans, Arabidopsis thaliana and yeast (Chen et al., 2009; Sarikas et al., 2011; Ban and Estelle 2021). CUL proteins possess a substrate-targeting function, often through an adaptor protein and a RING finger component (Sarikas et al., 2011; Liu et al., 2017). All the complexes known so far have been grouped into four main CRLs. The classes are consist of: 1) the CUL/RING/Skp/F-box CRLs proteins acting as substrate receptors while Skp1 or related proteins serving as adaptors; 2) the CUL/RING/BTB CRLs (BTB CRLs) protein being characterized by the lack of additional adaptors and containing proteins with BTB domains as substrate receptors directly attach to CULs (Christians et al. 2009, 2012; Marin 2009); 3) the CUL/RING/DDB/DCAFs CRLs (DDB CRLs) protein related to mammalian DAMAGE-SPECIFIC DNA-BINDING PROTEIN 1 protein (DDB1) serving as adaptors and with WD40 domains acting as substrate receptors (Marin 2009); 4) the complex receptors (BC-box CRLs) consisting of CUL/RING/Elongin/SOCSB boxes as substrate receptors and containing elongin proteins as adaptors (Marin 2009; Chahtane et al., 2018; Julian et al., 2019; Chico et al., 2020). The CUL-organized CRLs recruits the substrate and the E2 ubiquitin-conjugating enzymes, which transfer ubiquitin from the E2-conjugating enzymes to the substrate. In addition, conjugation of CULs with the ubiquitin-like molecule Nedd8 modulates activation of the corresponding CRL complex through conformational regulation of the interactions between CUL's carboxyterminal tail and CRL's RING subunit. In plants, CRLs are probably the best-characterized E3s to date, participating in plant growth and development (Roberts et al., 2011; Chen et al., 2013; Genschik et al., 2013; Chahtane et al., 2018). The dissection of the whole genome of S. miltiorrhiza provides an excellent molecular biology platform for its gene family analysis, functional gene mining, genome evolution, and so on (Schwechheimer 2018). So far, the CUL gene family of S. miltiorrhiza is rarely studied. CUL genes were thought to play vital roles in regulating the growth and development of S. miltiorrhiza, therefore, it is essential to investigate the CUL gene family in S. miltiorrhiza. The present study systematically studied the CUL genes number, gene structures, conserved domains and subgroup classification in the whole genome of S. miltiorrhiza. Moreover, we investigated gene expression profiles in different tissues along with the ABA treatment, providing a valuable reference for the functional identification of CUL genes.
To identify the potential CUL E3 in S. miltiorrhiza, the genome sequence was downloaded from the S.miltiorrhiza database (Xu et al., 2016) (ftp://danshen.ndctcm.org:10402/). Then, the seed file of the CULLIN domain (PF00888) was retrieved from the Pfam database. The HMMER program was used to identify the potential CUL genes in S. miltiorrhiza (Eddy 2011; Finn et al., 2011). All candidate CUL genes obtained from the result of HMMsearch were further submitted to SMART website (http://smart.embl-heidelberg.de/) to determine completeness of CUL conserved domain (Letunic and Bork 2018; Wang et al., 2021). In addition, ExPASy-Compute pI/Mw tool was used to calculate the amino acid number, molecular weight, theoretical pI, instability index, and aliphatic index as well as GRAVY (Grand Average of Hydropathicity) (Chen et al., 2022). A Cell PLoc 2.0 prediction was introduced to determine the subcellular localization of CUL gene candidates (http://www.csbio.sjtu.edu.cn/bioinf/plant-multi/) (Emanuelsson et al., 2000; Chen et al., 2022).
CUL protein sequences from A. thaliana and S. miltiorrhiza were collected. In final, the total of 14 CUL protein sequences were downloaded. Sequence alignment was performed using MEGA 6.0 software with the ClustalW function (Tamura et al., 2013). The phylogenetic tree was constructed by MEGA 6.0 with the Neighbor-joining method. The ITOL tool was used to create the tree visualization (https://itol.embl.de/) (Letunic and Bork 2019; Sharma and Taganna 2020; Sharma et al., 2021).
To identify and visualize the structural organization (introns, exons and untranslated regions) of the S. miltiorrhiza CUL genes, the GSDS (Gene Structure display System) tool was utilized (http://gsds.cbi.pku.edu.cn/) (Hu et al., 2015). The novel reserved motifs of S. miltiorrhiza CUL genes were identified using a motif-based sequence analysis tool by MEME suite (http://meme-suite.org/) (Bailey et al., 2009; Sharma and Taganna 2020; Sharma et al., 2021). For a total of three motifs and a width limit of 50 amino acids were used in this study. Meanwhile, the CUL genes structures and conserved domains are visualized using the Tbtools software (Chen et al., 2020a; Du et al., 2022).
A length of 3000-bp in the upstream of initiation codon of the CUL genes was specified as putative promoter sequences. Eight promoter sequences of CUL genes were retrieved using Tbtools (Chen et al., 2020a). The cis-acting regulatory elements of the promoter sequences were predicted by PlantCare online (Rombauts et al., 1999). According to the functional annotations of cis-acting elements, the candidate elements were gathered for further research and the cis-acting elements with the same functional annotations were incorporated into the same group. Additionally, the word art image of cis-acting elements in the promoters was generated with the WordArt tool (https://wordart.com) (Mi et al., 2005; Sharma and Taganna 2020).
Sterile S. miltiorrhiza plants were cultivated on Murashige and Skoog (MS) media at 25 °C under a 16 h light/8 h dark photoperiod (Deng et al., 2020b; Zhou et al., 2021b). The Agrobacterium rhizogenes strain C58C1 (pRi A4) cultures were used to infect the sterile stems and/or leaves of S. miltiorrhiza to generate hairy roots (Cao et al., 2018; Huang et al., 2019). The well-grown S.miltiorrhiza hairy roots were used to perform the different treatment with ABA, and the hairy roots were collected after 0h, 0.5h, 1h, 2h, 4h and 8h of treatment for RNA isolation and cDNA synthesis (Du et al., 2018; Cao et al., 2021; Zheng et al., 2021). Reverse transcription was performed with the cDNA Synthesis Kit (Clontech, USA) according to protocols. Double stranded cDNAs were separated on agarose gel, and were recovered for the RNA-seq (Chen et al., 2022). Construction of the cDNA library was performed by the Majorbio Bio-pharm Technology (Shanghai, China) and was sequenced by Illumina HiSeq TM 2500 with PE100. All reads have been uploaded to the National Center for Biotechnology Information (NCBI) public database with the SRA access number SRP307198. De novo assembly of the Illumina sequenced short length reads was conducted as reported previously (Zhou et al., 2017). Using the RNA-seq data, the expression levels of these genes were quantified by RPKM values, and TBtools visualized the expression results (Chen et al., 2020a). Based on the FPKM values of genes in transcriptome database, the co-expression regulatory network of tanshinone and salvianolic acid biosynthetic genes with candidate CUL genes was constructed by Pearson's correlation test (r > 0.8 and P-value < 0.05). Then, Cytoscape software is used to visualize the above results (Shannon et al., 2003).
Different tissues (roots, stems, leaves, and flowers) and hairy roots grown for 40 days were treated with 50 μM ABA for RNA isolation, and was converted into cDNA using a plant RNA prep pure kit (Tiangen Biotech Co., Ltd., Beijing, China) (Zhou et al., 2021a). cDNA of each sample was synthesized and qRT-PCR assay was carried out using a Super Real PreMix Plus (SYBR Green) kit (Tiangen, China) on ABI Step One Plus Real-TimePCR Systems (Applied Biosystems, USA) as described previously. SmActin gene was used as an internal control. The primer pairs for qRT-PCR are summarized in Table S1. The 2−ΔΔCt method was introduced to perform the quantification of gene expression (Liu et al., 2022). Each generated data point represented the average of three independent experiments.
In this study, we used a strict pipeline to identify CUL genes in the S. miltiorrhiza genome. The HMM (Hidden Markov Model) profile of the CUL domain was obtained from the Pfam database. The HMMER tools were employed to convert the multiple sequence alignment into the position-specific scoring system, then to conduct large-scale sequence analysis. In final, we obtained eight putative sequences using HMMER with default parameters and a significant P-value of 0.01. We have analyzed the molecular weight, number of amino acids, gene length, pI, instability index, grand average of hydropathy, introns, class, and subcellular localization of all the CUL genes (Table S2). The molecular weight of the CUL genes ranges from 32.55 to 103 kDa, and the number of amino acids varies from 280 to 920. The pI value is from 4.98 to 8.40. Most of the proteins were predicted to be unstable and hydrophilic. From the protein subcellular localization, it was predicted that most of the CUL proteins might locate in the cytoplasm, while the remaining members were predicted to exist in the chloroplast (Table S2).
To investigate the evolutionary history of CUL genes in S. miltiorrhiza, we constructed a phylogenetic tree using the MEGA 6.0 tool based on the CUL proteins from S. miltiorrhiza (8 members), Arabidopsis (6 members) (Fig. 1). According to the domains related to the function of CUL proteins, these specific proteins linked to CUL were classified as CUL-SCF (S-phase kinase-associated PROTEIN 1 (SKP1) -CUL-F-box), CUL-BTB (Bric a BRAC, Tramtrack and Broad Complex), CUL-DDB1 (UV-damaged DNA Binding Protein 1) and CUL-APC/C (Anaphase Promoting Complex), respectively, which were divided into four groups: CUL-SCF (Group I), CUL-BTB (Group II), CUL-DDB1 (Group III) and CUL-APC/C (Group IV). Interestingly, the total number of CUL genes in S. miltiorrhiza and Arabidopsis is comparatively secure, and it indicates the conservative features of this gene family. In order to evaluate the degree of gene expansion or loss during evolution, the CUL genes in each group were counted. In S. miltiorrhiza, it was found that Groups I–IV contained 4, 2, 1, 1 CUL genes, respectively. In Arabidopsis, Groups I to IV contained 2, 2, 1, 1 CUL genes, respectively. Comparison of S. miltiorrhiza and Arabidopsis, the increased gene number in group I implies the presence of gene expansion in S. miltiorrhiza.
To study the structure of CUL genes, we compared the full-length cDNA sequences of all genes with the corresponding genomic DNA. By comparing the number and location of exons and introns (Fig. 2a), we found that 8 CUL genes identified from S. miltiorrhiza had different numbers of exons, ranging from 1 to 19. SmCUL1 and SmCul1-likeA had the largest number of exons, and all the 8 CUL genes contained CUL domains and about 12.5% of them had no introns. The difference in the number of exons may indicate that the CUL gene families have different functions involved in the secondary metabolites biosynthesis, growth and development in S. miltiorrhiza. All of the identified CUL genes were analyzed for the presence of the novel and uncapped motifs using MEME suite utilizing a two-component finite mixture model (Fig. 2b). It was found that there are 1–3 motifs distributed in CUL members (Fig. S1). This discovery provides a novel evidence for predicting gene biological functions. The common motifs among the gene sequences are indicative of conserved evolutionary relatedness and similar cellular functions.
Usually, transcription factors regulate the expression level of target genes by binding to its cis-acting element in the promoter of target genes in specific biological processes. Thus, cis-acting elements were important clues for predicting the gene function. In order to further explore the function of the SmCUL gene, the PlantCare database was introduced to predict the cis-acting elements in the putative promoter region of the SmCUL genes. According to the predicted data, 18 cis-acting elements related to stress, hormones, plant growth and development in the promoters of the eight CUL genes were annotated and selected to further study the distribution pattern in the promoter. As shown in Fig. 3a, different distribution patterns were observed in the promoter region of the SmCUL genes, indicating that they have different biological functions. Especially, the cis-acting elements related to hormone regulation, such as abscisic acid (ABA), salicylic acid (SA), gibberellin (GA), auxin and methyl jasmonate (MeJA) are very important in most of the CUL genes (Fig. 3a and b). ABRE, as a key cis-acting element in response to ABA treatment, has been identified in 6 SmCUL genes (including SmCUL1, SmCUL1-likeA, SmCUL1-likeB, SmCUL3A, SmCUL3B and SmCUL4), which indicates that most of the CUL genes in S. miltiorrhiza may be particularly sensitive to ABA treatment. In addition, 7 SmCUL genes (including SmCUL1, SmCUL2, SmCUL1-likeA, SmCUL1-likeB, SmCUL3A, SmCUL3B and SmCUL4) are confirmed to have MeJA responsive elements, and 6 SmCUL genes (including SmCUL1, SmCUL1-likeA, SmCUL1-likeB, SmCUL3A, SmCUL3B and SmCUL4) have cis-acting elements related to drought, indicating that these genes may have special resistance under drought stress. It is worth noting that the MYB transcription factor (TF) binding elements exists in the six SmCUL genes (including SmCUL1, SmCUL1-likeA, SmCUL1-likeB, SmCUL3A, SmCUL3B and SmCUL4), indicating that the six SmCUL genes may be regulated by MYB genes in response to drought stress. The promoter elements are clustered and represented by a word cloud image. As shown in Fig. S2, these cis-acting elements including light responsive element (ATTAAT), abscisic acid (ABA) responsive element (ACGTG), MeJA responsive element (TGACG and CGTCA), low temperature responsive element (CCGAAA), MYB drought-induced binding site (CAACTG), auxin responsive element (AACGAC), salicylic acid responsive element (CCATCTTTTT), gibberellin responsive element (CCTTTTG and TCTGTTG) and stress responsive element (ATTCTCTAAC), are abundant in the promoters of CUL ubiquitin ligase genes (Fig. 3a and b), among of them, light, ABA and MeJA responsive element got the highest abundance, implying that the CUL genes are might be closely related to plant growth and development.
ABA has been validated to act as an abiotic inducer promoting tanshinone and phenolic acid biosynthesis in S. miltiorrhiza (Li et al., 2018; Deng et al., 2020b). Coincidentally, the promoters of CUL genes also have a large number of ABA cis-acting elements. So, we collected six ABA-treated RNA-Seq samples to study the expression pattern of the CUL genes through RNA sequencing analysis. The result showed that the identified eight CUL genes were all responded to ABA stress (Fig. 4a and Table S3). Meanwhile, according to our real-time PCR results, SmCUL1, SmCUL1-likeA, SmCUL3A and SmCUL3B were significantly induced by ABA, and their expression levels peaked at 4 h (Fig. 4b). These results suggested that CUL genes might play an important role in ABA regulation activity.
The expression profiles of the eight CUL genes in vegetative and reproductive tissues (leaf, stem, root and flower) were examined to explore the function of the gene participating in plant growth, development and secondary metabolism in S. miltiorrhiza (Fig. 5). The results showed that the expression levels of six CUL genes (including SmCUL1, SmCUL1-likeA, SmCUL1-likeB, SmCUL2, SmCUL4 and SmAPC2) exhibited the highest expression level in the vegetative tissue of stem. Whereas, SmCUL3A got the highest expression level in root, which is used as the medicinal harvesting tissue in traditional chinese medicine. The differential expression of CUL genes in various tissues indicated the diverse function in plant secondary metabolite synthesis, growth and development in S. miltiorrhiza.
Previous studies showed that ABA could promote the phenolic acids and tanshinone accumulation in hairy roots by activating the expression of phenylalanine ammonia-lyase (PAL), tyrosineamino transferase (TAT) (Zhang et al., 2013; Ding et al., 2017; Deng et al., 2020b; Zhou et al., 2021b).The co-expression network of eight CUL genes with phenolic acids and tanshinone biosynthetic genes was constructed, and the result indicated that 3 out of 8 CUL genes (including SmAPC2, SmCUL2 and SmCUL4) showed a negative correlation with phenolic acids biosynthetic genes (Fig. 6 and Table S4). Co-expression analysis revealed that 1 CUL gene (SmCUL2) showed a positive correlation with CPS gene in tanshinone biosynthetic pathway with Pearson correlation coefficient(r) > 0.8 and P-value <0.05 as a cutoff (Fig. 6 and Table S4). Overall, these results suggested that 3 out of 8 CUL genes might participate in phenolic acids and tanshinone biosynthesis.
CUL ubiquitin ligase genes widely exist in plants and have been validated to participate in diverse biological functions (Adams and Spoel 2018; Ma et al., 2021; Yu et al., 2021). Due to CUL genes acting as an essential role during plant development, they have been identified in many plant species. S. miltiorrhiza, one species from Lamiaceae, is a valuable traditional Chinese herbal plant being historically used to treat cardiovascular and cerebrovascular diseases (Deng et al., 2020a; Zhou et al., 2021a). Genome-wide identification of the CUL genes is an essential process towards further functional characterization of these genes in S. miltiorrhiza, but this work is poorly studied till now. In this study, 8 CUL genes were identified in S. miltiorrhiza genome by HMMER analysis using the Pfam and InterPro databases, and the total number of CUL genes in S. miltiorrhiza is comparable to that of A. thalania (11) (Thomann et al., 2005; Liu et al., 2017). In Arabidopsis, 6 out of 11 CUL genes have complete C-terminal or N-terminal sequence, and the other five CUL genes cannot be translated normally because they do not have complete N-terminal and C-terminal sequence. Based on the results of evolutionary tree clustering, eight CUL genes in S. miltiorrhiza were named according to the names of CUL genes in A. thaliana (Fig. 1). Phylogenetic analysis showed that a total of 14 CUL protein members in these two species (8 in S. miltiorrhiza while 6 in A. thalania), were grouped into four groups (Fig. 1) (Ban and Estelle 2021). Except to the CUL domain, these members of SmCUL1, SmCUL1-likeA, SmCUL1-likeB, SmCUL3A, SmCUL3B, SmCUL4 and SmAPC2 also contain Cullin_Nedd8, APC2 or ANAPC2 domains (Zhuang et al., 2009). In this study, the CUL proteins can be divided into four groups, including CUL-SCF (Group I), Cul-BTB (Group II), Cul-DDB1 (Group III) and Cul-APC/C (Group IV). CUL ubiquitin ligases can form multisubunit enzymes with complex structures (Thomann et al., 2005). The interaction of multisubunit enzymes with substrates requires specific connexin to form specific complexes in order to act as E3 ubiquitin ligases properly, of which it indicates the functional complexity and diversity of the CUL gene family (Chen et al., 2015; Liu et al., 2017; Chico et al., 2020).The diverse structure and organization of the CUL genes, is associated with the evolution and functional differentiation of this gene family in different species (Wu and Krainer 1998; Thomann et al., 2005). In the present study, some CUL genes in S. miltiorrhiza either have no introns or more than three introns (Fig. 2). It was thought that a large number of introns in CUL genes might act as a mutational buffer to protect the coding sequence and keep away from functionally deficient mutations (Wu and Krainer 1998; Thomann et al., 2005). The results of gene structure and motif analysis of CUL genes in S. miltiorrhiza will be valuable for predicting the gene evolution and identifying the function of candidate genes. Through analyzing cis-acting elements within promoters, it indicates that the CUL gene family is involved in stress-related mechanisms, hormone regulation, growth and development (Fig. 3A) (Belda-Palazon et al., 2019; Chen et al., 2020b; Dou et al., 2021). In our study, most of CUL genes contained ABA responsive elements in putative promoter regions (Fig. 3B). In particular, six of the eight CUL gene (SmCUL4, SmCUL3A, SmCUL3B, SmCUL1-likeA, SmCUL1-likeB and SmCUL1) promoters all contained ABA responsive elements, among of which SmCUL1 contains the largest number of ABA responsive elements reaching to six. The result indicates that the CUL gene may play an essential role in the ABA signal transduction process in S. miltiorrhiza. In A. thaliana, AtCUL3 was validated to interact with AtHB6 to respond to ABA induction (Lechner et al., 2011). Meanwhile, AtCUL3 promoted the degradation of AtMYB56 and AtWRI1 to regulate fatty acid accumulation in seeds and to affect flowering (Chen et al. 2013, 2015; Škiljaica et al., 2019). Herein, many MYB transcription factors binding sites referred to drought induction within the promoter region of CUL genes (including SmCUL1, SmCUL1-likeA, SmCUL1-likeB, SmCUL3A, SmCUL3B and SmCUL4) were identified in S. miltiorrhiza, suggesting that CUL genes might be regulated by related MYB genes mediating the drought stress signaling (Park et al., 2008; Baldoni et al., 2015; Chen et al., 2015), and this hypothesis needs to verify by further experiments. We also found light responsiveness elements, low-temperature responsive elements, and gibberellin-responsive elements in the promoter regions of CUL genes in S. miltiorrhiza. These results indicated that CUL genes might participate in diverse biological processes during growth and development in S. miltiorrhiza (Roberts et al., 2011; Morimoto et al., 2017). The CULgenes were thought to participate in various abiotic stress and hormone induction (Zhang et al., 2014; Orosa et al., 2017). Due to the highest occurrence frequency of the ABA responsive elements in the promoters of CUL genes (Fig. 3), it pushed us to investigate the gene expression pattern of the CUL genes responding to ABA treatment. Based on RNA-sequencing databases, it revealed that except to SmAPC2, the other seven CUL genes could be induced by ABA treatment (Fig. 4). Our quantitative detection of the expression level of all CUL genes exposed to ABA treatment were consistent with the transcriptome database (Fig. 4b) In fact, we found that many cis-elements in the CUL gene family are associated with hormone regulation not only contain ABA, but also include SA, GA and other auxins (Fig. 3). In conclusion, we have characterized the CUL gene family in S. miltiorrhiza based on the whole genome, transcriptome dataset and qRT-PCR expression analysis. Our research is the first systematic and comprehensive analysis of the CUL genes family in S. miltiorrhiza, and provides a valuable information for further elucidating the molecular mechanism of CUL genes responding to ABA induction. It may also help us to recognize the diverse biological functions of CUL genes in other species.
Xiankui Gao: Writing – original draft, Drafting the manuscript, Validation, Methodology, Formal analysis, Funding acquisition, of, Data curation. Xiujuan Li: Writing – original draft, Drafting the manuscript, Formal analysis, Funding acquisition, of, Data curation. Chengan Chen: Formal analysis. Can Wang: Formal analysis. Yuqi Fu: Resources, Investigation. ZiZhen Zheng: Resources, Investigation. Min Shi: Writing – review & editing, Supervision. Xiaolong Hao: Writing – review & editing. Limei Zhao: Methodology. Minghua Qiu: Writing – review & editing. Guoyin Kai: Conceptualization, and design of study, Acquisition of data, Revising the manuscript. Wei Zhou: Writing – original draft, Drafting the manuscript, Formal analysis, Conceptualization, and design of study, Acquisition of data, Revising the manuscript, Approval of the version of the manuscript to be published.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. | true | true | true |
|
PMC9576592 | 35338348 | Yarui Li,Junbi Hu,Dan Guo,Wenhui Ma,Xu Zhang,Zhiyong Zhang,Guifang Lu,Shuixiang He | LncRNA SNHG5 promotes the proliferation and cancer stem cell-like properties of HCC by regulating UPF1 and Wnt-signaling pathway | 25-03-2022 | Cell biology,Cancer stem cells,RNAi,Cancer genetics | The role of long noncoding RNA (lncRNAs) had been demonstrated in different types of cancer, including hepatocellular carcinoma. This study was intended to investigate the role of lncRNA small nucleolar RNA host gene 5 (SNHG5) in HCC proliferation and the liver CSC-like properties. Through functional experiments, we determined that knockdown of SNHG5 repressed HCC cell proliferation and CSC-like properties, while over-expression of SNHG5 promoted cell growth. At the same time, CSC markers (CD44, CD133, and ALDH1) and related transcription factors (OCT4, SOX2, and NANOG) were downregulated when SNHG5 was knocked down. Mechanically, RNA immunoprecipitation (RIP) and RNA pulldown assay showed that SNHG5 regulated the proliferation and CSC-like properties of HCC by binding UPF1. Further investigations showed that expression of critical components of Wnt/β-catenin pathway (β-catenin, TCF4, c-myc, cyclinD1, and c-Jun) were upregulated with depletion of UPF1 in liver CSCs, which were downregulated with depletion of SNHG5. After use of the inhibitor of Wnt/β-catenin pathway, the formation of liver CSCs sphere decreased. Taken together, SNHG5 plays a critical role to promote HCC cell proliferation and cancer stem cell-like properties via UPF1 and Wnt/β-catenin pathway. | LncRNA SNHG5 promotes the proliferation and cancer stem cell-like properties of HCC by regulating UPF1 and Wnt-signaling pathway
The role of long noncoding RNA (lncRNAs) had been demonstrated in different types of cancer, including hepatocellular carcinoma. This study was intended to investigate the role of lncRNA small nucleolar RNA host gene 5 (SNHG5) in HCC proliferation and the liver CSC-like properties. Through functional experiments, we determined that knockdown of SNHG5 repressed HCC cell proliferation and CSC-like properties, while over-expression of SNHG5 promoted cell growth. At the same time, CSC markers (CD44, CD133, and ALDH1) and related transcription factors (OCT4, SOX2, and NANOG) were downregulated when SNHG5 was knocked down. Mechanically, RNA immunoprecipitation (RIP) and RNA pulldown assay showed that SNHG5 regulated the proliferation and CSC-like properties of HCC by binding UPF1. Further investigations showed that expression of critical components of Wnt/β-catenin pathway (β-catenin, TCF4, c-myc, cyclinD1, and c-Jun) were upregulated with depletion of UPF1 in liver CSCs, which were downregulated with depletion of SNHG5. After use of the inhibitor of Wnt/β-catenin pathway, the formation of liver CSCs sphere decreased. Taken together, SNHG5 plays a critical role to promote HCC cell proliferation and cancer stem cell-like properties via UPF1 and Wnt/β-catenin pathway.
Hepatocellular carcinoma (HCC) is the sixth common human malignancy in the world [1, 2]. Although the diagnosis and treatment techniques have improved significantly in recent years, the long-term survival rate among the HCC patients is still very low, this is mostly because of the fast metastatic property and higher rate of recurrence [3, 4]. Therefore, finding an alternate therapeutics and the identification of the underlying mechanism of progression is crucial in this situation. Tumor cells characteristically present in a heterogeneous manner and thus influencing growth, metastasis, and recurrence. Heterogenicity might result from cells showing stem cell like character, otherwise known as cancer stem-cells (CSCs) [5]. CSCs have the abilities to self-renew, differentiate, and uncontrollable growth, that results in the formation of new growth in the local or a distant organ, which integrates with the non-CSCs [6–8], thus, contributing to the progression, metastasis, and recurrence of cancers. The existence of liver cancer stem cells (CSCs) is a known fact, and this group of cells has been characterized by several makers, such as CD133, CD13, CD90, and EpCAM [9–12]. Long noncoding RNA (lncRNA) length larger than 200nt does not possess the ability to encode proteins, thus remarked as “dark matter” in human diseases. LncRNAs have been demonstrated to regulate important biological functions, such as stem cell properties and tumor progression [13]. However, the functions and mechanisms of lncRNAs in CSCs are controversial. The small nucleolar RNA host gene 5 (SNHG5), one of the well-defined cytoplasmic lncRNAs, also known as U50HG, is 524 bp in length. Our previous research confirmed that SNHG5 was highly expressed in HCC tissues and was related to the prognosis of HCC patients, and further investigations showed that lncRNA SNHG5 plays a role in HCC metastasis [14]. Recently, many researches have revealed that lncRNAs have the ability to bind to DNA or RNA by a complementary sequence, which plays critical roles in mRNA splicing, RNA decay and translation, additionally, the post-translational modification of proteins can be modulated by lncRNAs [15, 16]. UPF1 is a key player in RNA-degradation pathways, and also essential for accomplishing DNA replication. Additionally, UPF1 interacts with many RNA substrates and promotes mRNA stability [17]. The important thing is that we find that UPF1 contains a potential binding site with SNHG5, which prompted our interest in investigating the biological roles and relationships of SNHG5 and UPF1 in HCC CSCs. The present study aims to analyze the effect of SNHG5 on the proliferation and establish the cancer stem cell-like properties of HCC, explore the function of the SNHG5 in regulating the properties of HCC CSCs through UPF1 and Wnt-signaling pathway in vitro, and in promoting tumorigenesis in vivo.
The human HCC cell lines HepG2 and Huh7 were obtained from the Chinese Academy of Sciences Cell Bank (Shanghai, China). The cell lines have been tested and authenticated by short tandem repeat (STR).The HCC cells were cultured in the DMEM/High Glucose (Hyclone, USA) medium in a humidified incubator at 37 °C temperature and 5% CO2 concentration. About 10% FBS (fetal bovine serum, Gibco USA) and penicillin–streptomycin (100 U/mL and 100 μg/mL, respectively) were added in the DMEM/High Glucose medium prior to culture.
SNHG5 and UPF1 overexpression plasmids, SNHG5-knockdown plasmids (SNHG5 shRNA with a corresponding negative control shRNA-NC), SNHG5-Mut/WT plasmids (pCMV-SNHG5- Mut vector containing mutations at the putative UPF1-binding site was generated by site-directed mutagenesis) were designed by Genechem (Shanghai, China). The GV248 vector was used, and the stable clones were selected by 5 μg/ml puromycin-containing medium. The puromycin-resistant cell clones were established after 4 weeks. Gene-expression level was evaluated by quantitative real-time PCR. The siRNA (small-interfering RNA) against UPF1 was designed by Genepharma (Shanghai, China). According to the manufacturer’s protocol, HCC cells were transfected with plasmid by using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA).
MTT (0.5 mg/ml) was added into transfected cells and kept in the dark for 4 h. Then, the supernatant was removed, 150 μl of DMSO was added, and the optical density (OD) was measured at 490 nm.
Transfected cells were seeded in to a 96-well plate (2 × 103) embedded with complete growth medium. Then according to the manufacturer’s protocol, the process was carried out with EdU-detection kits (Keygen, Nanjing, China). The experiment was done in triplets. The cells were imaged with an inverted fluorescent microscope (Nikon Eclipse Ti-S) (20X).
After routine incubation, transfected cells were trypsinized, centrifuged, counted, and replanted at a density of 500 cells/6 cm plate. After 12 days, the cell colonies (one colony containing at least 50 cells) were fixed with 37% methanol, stained with 0.1% crystal violet, and counted under a microscope.
Ultra-low-attachment culture dishes (Corning, USA) were used to culture HepG2 and Huh7 cells with DMEM/F12 (Gibco, USA) added with 1% FBS, 20 ng/mL epithelial growth factor, and 20 ng/mL fibroblast growth factor for two weeks. The formation and the number of spheroids were detected by a stereomicroscope (Olympus, Japan).
The total RNA was extracted from the cultured cells and the collected HCC tissues by using Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. The Prime Script TM RT Master Mix Kit (Takara, Japan) and Mir-X miRNA qRT-PCR SYBR Kit (Takara, Japan) were used to obtain the cDNA. Quantitative real-time PCR (qRT-PCR) was performed with SYBR Premix Ex Taq™ II (Takara)on Thermal Cycler CFX6 System (Bio- Rad). β-actin as the endogenous control of qRT-PCR. The 2−ΔΔCt method was used to calculate relative gene expression. Primer sequences for PCR were presented in Supplementary Table 1.
The total protein from the cultured HCC cells and the tissue samples was isolated by RIPA (Beyotime, Haimen, China) supplemented with proteinase and phosphatase inhibitors. BCA detection kit (Keygen, Nanjing, China) was used for the quantification according to the manufacturer’s protocol. During electrophoresis, 5% gel for concentration and 10% for separation were used. The proteins were then transferred on a PVDF membrane (Merck Millipore) and were blocked by 5% nonfat milk for 1 h. Then, the PVDF membrane was incubated overnight at 40C with the primary antibodies (Supplementary Table 2). On the next day the secondary antibody (Zhuangzhi Biology, China) was diluted in TBST in a 1:5000 ratio, and the membranes were re-incubated for 1 h. The protein bands were evaluated by ECL immunoblotting kit following the manufacturer’s protocol (Millipore, USA).
Cells were cultured on glass coverslips for 24 h, then fixed in 4% paraformaldehyde at room temperature for 15 min, and washed by PBS. The adherent cells were permeabilized using 0.5% Triton X-100, and blocked with 10% goat serum for 1 h. After that they were incubated with primary antibody at 4 °C overnight and secondary antibodies with an appropriate dilution. The cells were washed gently with PBS for 3 times, then the coverslips were stained with DAPI and imaged with an inverted fluorescent microscope (Nikon Eclipse Ti-S) (100X).
Four-week aged BALB/c nude mice were purchased from the Central Laboratory of Animal Science, Xi’an Jiaotong University, China. They were randomly divided into two groups, 5 mice in each group. The mice were kept under sterile specific pathogen-free (SPF) environment. A subcutaneous injection of 5 × 106/200 μl HepG2 cells stably transfected with SNHG5–shRNA or NC-shRNA were given to each mice. The tumor formation was carefully observed every 4-day interval. Eight days following the injection, the palpable tumors were observed (blinded to the group allocation). Four weeks following the injection, the subcutaneous formatted tumor nodes were executed for further detection(cervical-dislocation method executed experimental animals). This study was done according to the Guide line for the “Care and Use of Laboratory Animals of the National Institutes of Health” and was approved by the Medical Ethics Committee of the Experimental Animal Center of Xi’an Jiaotong University.
The cytoplasmic and nuclear fractions of HepG2 or Huh7 CSCs were isolated using the Nuclear/Cytosol Fractionation Kit (Cell Biolabs). The detailed steps of the experiment were executed according to the manufacturer’s protocol.Then, RIP–PCR was performed to detect the SNHG5 expression of cytoplasmic and nuclear.
Millipore EZ‐Magna RIP RNA Binding Protein Immunoprecipitation kit (Millipore) was applied to performed RIP assays according to the manufacturer’s protocol. Rabbit polyclonal IgG (Millipore) and antibodies to UPF1 (Abcam) were used in RIP assays. Then, RIP‐PCR was performed, and total RNA was used as input controls.
The length of synthetic biotinylated SNHG5 was synthesized by Genechem (Shanghai, China). Biotin RNA Labeling Mix (Roche) was used to perform RNA pulldown experiment. First, biotinylated SNHG5 or SNHG5-Mut was incubated with cell-protein extractions (1 mg), which were then targeted with streptavidin beads (Invitrogen) and washed. The associated proteins were resolved by gel electrophoresis. Specific bands were excised and identified by western bolt.
The cell- and molecular-biology experiments have been implemented three times. Data were expressed as mean ± SD. Statistical analysis was performed using SPSS 23.0 (IBM, SPSS, Chicago, IL, USA) and GraphPad Prism V7.0 and Student’s t-test and one-way ANOVA were done to analyze the results. A two-tailed P < 0.05 was considered as statistically significant and P < 0.01 was considered highly significant.
The effect of SNHG5 on HCC cells in vitro was investigated by upregulation and knockdown of the expression of SNHG5. The results showed that the expression of SNHG5 dramatically increased after transfection with the pCMV–SNHG5 vector in HepG2 and Huh7 cells while compared with the empty vector (Fig. 1A). Cellular functional-validation experiments were performed in HepG2 and Huh7 cells, these included MTT-proliferation assay, colony-formation assay, and Edu assay. The MTT assay showed that overexpression of SNHG5 induced HepG2 and Huh7 cell proliferation in comparison with the control group (Fig. 1B). Upregulation of SNHG5 results in an elevated growth tendency of the HCC cells in the Edu and colony-formation assays (Fig. 1C, D). To further evaluate the role of SNHG5 on HCC cell proliferation, specific lentivirus-mediated short-hairpin RNA (shRNA) targeting SNHG5 were transfected into HepG2 and Huh7 cells, resulting in a significant decrease in SNHG5 expression (Fig. 1E). MTT, Edu assay and plate colony-formation assays showed that knockdown of SNHG5 inhibited the HepG2 and Huh7 proliferation compared with the control group (Fig. 1F–H). Moreover, we check for the recovery of the SNHG5 phenotype after shSNHG5 knockdown, the MTT assay showed that SNHG5 vector can restore the inhibition of sh-SNHG5 on cell proliferation (Fig. S1). All these results suggest that SNHG5 has a significant role in HCC cell proliferation.
Sphere formation is a selection method that enriches CSCs. To confirm the expression pattern of SNHG5 in liver CSCs, we enriched liver CSCs through inducing hepatoma spheroid formation and examined SNHG5 expression in the self-renewing spheroids and the attached cells. As shown in Fig. 2A, the SNHG5 expression was increased dramatically in liver CSCs. To analyze the role of SNHG5 in self-renewal characteristics of liver CSC, SNHG5 expression in HepG2 and Huh7 CSCs was reduced by SNHG5–shRNA (Fig. 2B). Then, the sphere-formation assays were performed to observe the number of primary spheres per 1000 single-liver CSCs and secondary spheres (P2) per 100 single-liver CSCs after SNHG5 knockdown. The adherent cells/spheroids were observed every 4-day interval. The result showed that the sphere-formation rate was dramatically impaired in SNHG5–shRNA cells compared with the sh-NC cells (Fig. 2C–E). We also examined the effects of SNHG5 on the expression of stem factors (Oct4, Sox2, and Nanog) and markers (CD133, CD44 and ALDH1) in two groups. The result of qRT-PCR showed that decreased SNHG5 expression inhibited the enrichment of these CSC markers in HepG2 and Huh7 CSCs, while overexpression of SNHG5 showed the opposite effect (Fig. 2F). IF (Fig. 2G) and Western blot analysis (Fig. 2H) reconfirmed that knockdown of SNHG5 inhibits the stemness of HCC CSCs. These results indicated the contribution of SNHG5 in liver CSC-like properties.
Recently, several studies have found the role of RNA‐binding proteins (RBPs) in lncRNA-related pathway. In order to identify the proteins that are associated with SNHG5, the starBase and RBPsuite database were used for biological information prediction. The result showed that SNHG5 contains potential binding sites for UPF1. The expression of UPF1 in HCC tissues and healthy tissues from TCGA using GEPIA2 shows that UPF1 expression is reduced in HCC tissues (Figure S2). In order to confirm the regulatory relationship between SNHG5 and UPF1, we first tested the expression of UPF1 in HCC cells after SNHG5 knockdown or overexpression. The result showed that the expression of UPF1 increased in HCC cells and HCC CSCs with depletion of SNHG5 (Fig. 3A, C), while overexpression of SNHG5 significantly inhibited the expression of UPF1 (Fig. 3B, D). Consistently, Western blot and IF assay were performed to detect the expression of UPF1 in liver CSCs, revealing that knockdown of SNHG5 promotes the expression of UPF1 (Fig. 3E, F). To further investigate the interaction between SNHG5 and UPF1, first, we evaluated the cellular orientation of SNHG5 in HCC CSCs. Nuclear and cytoplasmic segments were gained from HCC CSCs. Then, RNA was extracted independently. SNHG5 was discovered mainly in the cytoplasmic fraction (Fig. 3G). Since UPF1 directly binds to SNHG5, we performed RNA immunoprecipitation (RIP) with UPF1 antibody, and observed an enrichment of SNHG5 with UPF1 antibody as compared with the nonspecific antibody (IgG control) (Fig. 3H, I). To further validate the interaction between SNHG5 and UPF1, we obtained the binding sites of SNHG5 and UPF1 through the RBPsuite database. RBPsuite is a database for predicting the binding sites of RNA and RBP. RBPsuite first divides the RNA sequence into 101-nucleotide fragments and scores the interaction between the fragments and RBP. According to the prediction results of the database, the binding site of SNHG5 fragment 3 and UPF1 scored the highest (Fig. S3), and we showed the schematic diagram in Fig. 3J. Therefore, we constructed SNHG5-wild-type(WT) and SNHG5-mutant (Mut) RNA probes (Fig. 3K) and performed RNA-pulldown experiments, the pullsown result showed that UPF1 binds to SNHG5 in HCC CSCs (Fig. 3L).These observations suggested that SNHG5 binds to UPF1 and inhibits UPF1 expression in HCC CSCs.
We report that SNHG5 suppressed the expression of UPF1, however, it was unclear whether UPF1 regulates SNHG5 expression. First, we depleted the expression of UPF1 by a specific siRNA against UPF1 gene transcript, and upregulated UPF1 by pCMV–UPF1 vector in HepG2 and Huh7 CSCs. The results of qRT-PCR showed that the expression of UPF1 dramatically decreased after transfection with siRNA–UPF1 (Fig. 4A), while pCMV–UPF1 vector increased the expression of UPF1 in HepG2 and Huh7 CSCs compared with the empty vector (Fig. 4C). However, it was interesting that neither downregulation or upregulation of UPF1 did not affect the expression of SNHG5(Fig. 4B, D). This indicated that UPF1 is a downstream gene of SNHG5, and SNHG5 regulates UPF1 in one direction. To explore the role of UPF1 in liver CSC properties, we performed sphere-formation assays, and the result showed that knockdown of UPF1 promoted the liver CSC-like properties of HCC cells (Fig. 4E), the number of spheres per 1000 single-liver CSCs increased after UPF1 downregulation (Fig. 4F). Additionally, qRT-PCR showed that the expression of CSC markers (ALDH1, CD44, and CD133) and stem factors (NANOG, OCT4, and SOX2) was remarkably upregulated with depletion of UPF1, while overexpression of UPF1 showed the opposite effect (Fig. 4G). Meanwhile, this result was verified in Western blot analysis (Fig. 4H, I). These data ultimately suggested that knockdown of UPF1 enhances liver CSC properties.
Our previous studies have demonstrated a relationship between SNHG5 and the Wnt/β-catenin signaling pathway [14], which is an important pathway for some stem cells. Hence, we speculated that Wnt/β-catenin pathway may be a key factor to liver CSC-like properties. To examine the speculation, we detected the Wnt-family members in HepG2 CSCs and Huh7 CSCs. The result showed that the expressions of Wnt1, Wnt3a, and Wnt10a were downregulated with depletion of SNHG5, and were upregulated with knockdown of UPF1 (Fig. 5A). The result of qRT-PCR also showed the same trend (Fig. 5B). Additionally, the critical components of Wnt/β-catenin pathway (TCF4, c-myc, cyclinD1, and c-Jun) were detected in HepG2 CSCs and Huh7 CSCs by Western blot. We observed that these key components were downregulated following knockdown of SNHG5, while the expression level was increased with depletion of UPF1 (Fig. 5C). It is known that the β-catenin is a key factor of Wnt/β-catenin pathway. Therefore, we detected the expression of β-catenin in HepG2 CSCs and Huh7 CSCs with IF analysis. The result illustrated that the nucleic β-catenin protein expression amplified in liver CSCs with depletion of UPF1, while SNHG5 knockdown attenuated the level of nucleic β-catenin (Fig. 5D).To better prove the rationality of the SNHG5/UPF1 signal axis, we performed rescue (adding-back) experiment, that is, to detect whether si-UPF1 can restore the sh-SNHG5 cell phenotype. The result of qRT-PCR showed that knockdown of SNHG5 significantly inhibited the expression of β-catenin, but downregulating UPF1 can partially restore the inhibitory effect of sh-SNHG5 on β-catenin expression (Fig. 5E). At the same time, MTT experiment revealed that downregulating UPF1 can partially restore the inhibitory effect of sh-SNHG5 on HCC cell proliferation, while UPF1 vector intensified the inhibitory effect of sh-SNHG5 (Fig. 5F, G). To further validate the function, XAV-939, a Wnt/β-catenin pathway inhibitor was adopted. With XAV-939, the sphere formation of liver CSCs was impaired and the number of spheres per 1000 single liver CSCs decreased (Fig. 6H–J). These results suggested that SNHG5 regulates the activation of the Wnt pathway through UPF1.
HepG2 and Huh 7 cells transfected with SNHG5–shRNA and negative control(NC) were subcutaneously injected into male nude mice for 5 weeks. Tumor growth curve revealed that HCC cells transfected with SNHG5–shRNA greatly inhibited tumor growth compared with NC group (Fig. 6A). We also observed the tumor volume and tumor weight among the two groups. The results showed that downregulation of SNHG5 suppressed tumor volume and weight, therefore inhibiting the tumor growth effectively (Fig. 6B, C). Additionally, the expression of SNHG5 and UPF1 was detected in xenograft tumors. The result of qRT-PCR indicated that the expression of SNHG5 greatly decreased, while the expression of UPF1 greatly increased in SNHG5–shRNA xenograft tumors (Fig. 6D). To test all their results at the same baseline level, we verified the expression levels of β-catenin, CSC markers, and stem factors in animal tumor tissues after knockdown of SNHG5 by qRT-PCR. The results showed that the expression levels of β-catenin, CSC markers and stem factors were decreased compared with the control group (Fig. 6D, E), which is also consistent with the results obtained in in vitro experiments. In summary, this study basically confirmed that SNHG5 promotes HCC proliferation and cancer stem cell-like properties by regulating UPF1 to activate the Wnt-signaling pathway (Fig. 7).
CSCs are a subgroup of cells with self-renew and differentiation properties, which have been isolated in many solid cancers [18–20]. Liver CSCs were proved to contain various subtypes,these were characterized by different surface makers, including CD133+CD13+, EpCAM+CD24+OV6+, CD133+CD44+CD24+ EpCAM+, and so on [21]. CSCs contribute to the epithelial–mesenchymal transition (EMT), metastasis, drug resistance, and radio resistance through varieties of mechanisms [22–24]. Numerous efforts have been made to dissect the molecular mechanisms involved in the regulation of liver CSCs with the intent to identify novel therapeutic strategies to improve the poor prognosis of HCC. Recently, it has been revealed that lncRNAs played critical roles in CSCs. For example, in glioma, the downregulation of lncRNA–ROR promoted the proliferation of cancer cells and the formation of a sphere of stem cells with the down expression of stem cell factor KLF4 [21]. In the case of HCC, it also has been verified that many lncRNAs are responsible to drive CSC self-renewal and tumor progression through various mechanisms [25, 26]. Based on this, lncRNA is expected to become an important therapeutic agent for HCC. The diverse functional repertoire of lncRNAs reveals various opportunities for their therapeutic targeting, the means of which need to be adjusted to the mode of action of the lncRNA [27]. Battistelli C et al. [28] designed a HOTAIR deletion mutant form, named HOTAIR-sbid, which was proven to reduce cellular motility, invasiveness, anchorage-independent growth, and responsiveness to TGFβ-induced EMT. These data provide evidence on a lncRNA-based strategy to effectively impair tumor metastases. Although studies have confirmed that lncRNAs function as critical regulators of gene expression in embryonic and induced pluripotent stem cells, the previous understanding of their role in CSCs has been limited. SNHG5 has been widely proven to play an important role in a variety of tumors, such as colorectal cancer, gastric cancer and osteosarcoma, and even myeloid leukemia [29–32]. In the present study, we unraveled a critical role for SNHG5 in liver CSC properties,SNHG5 alterations (knockdown or upregulation) significantly influenced the proliferation and self-renewal capacity of liver CSCs. Similarly, the expression of stem cell markers and stem factors decreased after downregulation of SNHG5. Our findings provided important insights into the relationship between SNHG5 and liver CSCs. In order to further investigate the mechanism of SNHG5 regulating the properties of liver CSCs, bioinformatic methods were used to find the potential target genes of SNHG5. We found that the key factor of nonsense-mediated mRNA decay (NMD) UPF1 contains a potential binding site with SNHG5.UPF1 plays a critical role in RNA degradation pathways, and also promotes the decay of mRNAs encoding many other proteins that oppose the proliferative, undifferentiated cell state [33]. UPF1 acts, in part, by destabilizing the NMD substrate encoding the TGFβ inhibitor, Smad7, and stimulating TGF signaling [34], and several studies have shown that UPF1 exerts suppressive roles in tumor progression [35]. As expected, UPF1 inhibited the stemness of liver CSCs. Mechanistically, we found that SNHG5 combined with UPF1, and the overexpression of SNHG5 following the downregulation of UPF1 with the downexpression of surface makers and stem factors. These results revealed that SNHG5 promotes the proliferation and cancer stem cell-like properties of HCC by regulating UPF1. However, the detailed binding sites within SNHG5 with UPF1 are still unclear, this needs to be further studied. Recently, several researches demonstrated that Wnt/β-catenin signaling played a critical role in cancer stem cells. For example, in glioma, the self-renewal and tumorigenicity of CSCs were regulated by dysregulated Wnt–FoxM1/β-catenin signaling pathway [36]. However, its role in liver CSCs was not completely known. Our data indicate that inhibition of the Wnt pathway results in an obviously impaired sphere formation capacity. We have also demonstrated that UPF1 is responsible for liver CSC characteristics, UPF1 mediates the activation of the Wnt pathway in liver CSCs by regulating the expression of Wnt, and SNHG5 mediated the activation of the Wnt pathway in liver CSCs by regulating UPF1 expression. In summary, we conclude that SNHG5 promoted HCC cell proliferation in vitro and in vivo, and was responsible for the sphere formation of liver CSCs and the CSC properties. The underlying mechanism of SNHG5 promoting the proliferation and CSC-like properties of HCC was by regulating UPF1 and activation of the Wnt-signaling pathway. Our current data imply that SNHG5, along with its downstream mechanism and pathways, could shed light on new potential therapeutic targets against liver CSCs.
Supplementary figures Supplementary tables | true | true | true |
PMC9576593 | 35352023 | Rihan El Bezawy,Stefano Percio,Chiara Maura Ciniselli,Michelandrea De Cesare,Gennaro Colella,Matteo Dugo,Silvia Veneroni,Valentina Doldi,Silvia Martini,Dario Baratti,Shigeki Kusamura,Paolo Verderio,Marcello Deraco,Paolo Gandellini,Nadia Zaffaroni,Valentina Zuco | miR-550a-3p is a prognostic biomarker and exerts tumor-suppressive functions by targeting HSP90AA1 in diffuse malignant peritoneal mesothelioma | 29-03-2022 | Cancer genetics,Mesothelioma | Diffuse malignant peritoneal mesothelioma (DMPM) is a rare and rapidly lethal tumor, poorly responsive to conventional treatments. In this regards, the identification of molecular alterations underlying DMPM onset and progression might be exploited to develop novel therapeutic strategies. Here, we focused on miR-550a-3p, which we found downregulated in 45 DMPM clinical samples compared to normal tissues and whose expression levels were associated with patient outcome. Through a gain-of-function approach using miRNA mimics in 3 DMPM cell lines, we demonstrated the tumor-suppressive role of miR-550a-3p. Specifically, miRNA ectopic expression impaired cell proliferation and invasiveness, enhanced the apoptotic response, and reduced the growth of DMPM xenografts in mice. Antiproliferative and proapoptotic effects were also observed in prostate and ovarian cancer cell lines following miR-550a-3p ectopic expression. miR-550a-3p effects were mediated, at least in part, by the direct inhibition of HSP90AA1 and the consequent downregulation of its target proteins, the levels of which were rescued upon disruption of miRNA-HSP90AA1 mRNA pairing, partially abrogating miR-550a-3p-induced cellular effects. Our results show that miR-550a-3p reconstitution affects several tumor traits, thus suggesting this approach as a potential novel therapeutic strategy for DMPM. | miR-550a-3p is a prognostic biomarker and exerts tumor-suppressive functions by targeting HSP90AA1 in diffuse malignant peritoneal mesothelioma
Diffuse malignant peritoneal mesothelioma (DMPM) is a rare and rapidly lethal tumor, poorly responsive to conventional treatments. In this regards, the identification of molecular alterations underlying DMPM onset and progression might be exploited to develop novel therapeutic strategies. Here, we focused on miR-550a-3p, which we found downregulated in 45 DMPM clinical samples compared to normal tissues and whose expression levels were associated with patient outcome. Through a gain-of-function approach using miRNA mimics in 3 DMPM cell lines, we demonstrated the tumor-suppressive role of miR-550a-3p. Specifically, miRNA ectopic expression impaired cell proliferation and invasiveness, enhanced the apoptotic response, and reduced the growth of DMPM xenografts in mice. Antiproliferative and proapoptotic effects were also observed in prostate and ovarian cancer cell lines following miR-550a-3p ectopic expression. miR-550a-3p effects were mediated, at least in part, by the direct inhibition of HSP90AA1 and the consequent downregulation of its target proteins, the levels of which were rescued upon disruption of miRNA-HSP90AA1 mRNA pairing, partially abrogating miR-550a-3p-induced cellular effects. Our results show that miR-550a-3p reconstitution affects several tumor traits, thus suggesting this approach as a potential novel therapeutic strategy for DMPM.
Diffuse malignant peritoneal mesothelioma (DMPM) is a rare tumor that develops from the mesothelial cells lining the peritoneal cavity. DMPM includes three histological subtypes, epithelioid, sarcomatoid, and biphasic (epithelioid and sarcomatoid), with the epitheliod subtype being the most frequent and less aggressive subtype [1]. Although locally aggressive, DMPM is characterized by poor prognosis, and patient survival does not exceed 1 year following treatment with palliative surgery and systemic or intraperitoneal chemotherapy [2]. The only treatment that meaningfully impacts the natural history of DMPM is aggressive cytoreductive surgery (CRS) combined with hyperthermic intraperitoneal chemotherapy (HIPEC), which is currently regarded as the gold-standard initial treatment for selected DMPM patients as it was found to significantly extend median survival time to 34–92 months [3–6]. However, for recurrent patients and for those who are not eligible for CRS + HIPEC there is an urgent need for alternative effective treatments. DMPM is an understudied disease and, although there might be an association with asbestos exposure, its pathogenesis is mostly unknown. A better understanding of the disease biology, leading to the identification of molecular alterations underlying disease onset and progression, could provide a source of novel therapeutic targets. A few studies carried out thus far indicated that a fraction of DMPM is characterized by the presence of mutations in BRCA1 associated protein 1,Neurofibromin 2, DEAD-Box Helicase 3 X-Linked, SET Domain Containing 2, and Histone Lysine Methyltransferase genes, as well as by the loss of 3p21 locus, which includes chromatin modifiers and epigenetic regulatory genes [7, 8]. In addition, ALK rearrangements have been described in a small subset of younger women affected by DMPM [9]. However, it is still unclear whether such molecular alterations are causative of the disease or impact disease progression. MicroRNAs (miRNAs) are small non-coding evolutionarily conserved RNA molecules, involved in post-transcriptional gene silencing by binding to 3’untranslated region (3’UTR) of target mRNAs, thus controlling a variety of important biological processes [10, 11]. Dysregulated miRNAs have been causatively associated with the pathogenesis of several diseases, including cancer. Depending on their expression levels, cellular context, and target genes, miRNAs can act as oncogenes or tumor suppressors [12]. In recent years, miRNA functional involvement in human cancer has raised an increasing interest toward their exploitation as therapeutic targets and tools [13]. Thus far, very little information is available on the expression and functional role of miRNAs in DMPM. Indeed, the current knowledge is limited to two miRNAs, miR-34a and miR-380-5p, that are negligibly expressed in DMPM and the ectopic expression of which in DMPM cell models exerts tumor-suppressive effects dealing with the inhibition of MET and AXL expression [14] and the perturbation of telomerase activity [15], respectively. In the current study, we investigated the expression levels of miR-550a-3p and its association with clinical outcome in a cohort of DMPM patients who underwent CRS + HIPEC. Furthermore, we assessed the biological effects induced by miR-550a-3p ectopic expression in DMPM patient-derived cell lines to provide the preclinical basis for the design of a novel miRNA-based therapeutic approach. In addition, to broaden the relevance of our findings, we extended the analysis of miR-550a-3p-induced effects to cell lines of other human tumor types, such as ovarian and prostate cancer. Results are reported herein.
Frozen DMPM lesions from 45 adult patients treated with CRS + HIPEC from 1997 to 2013 at the Fondazione IRCCS Istituto Nazionale dei Tumori (INT) of Milan were available for miR-550a-3p expression analysis. The H&E stained slides of all cases were reviewed, and the tumors were classified as epithelial, sarcomatoid, or biphasic according to the WHO classification. Eleven normal peritoneum specimens were also obtained from patients who underwent surgery for non-oncologic diseases. The study was approved by the Institutional Review Board. Written informed consent was obtained from all patients to donate the leftover tissue to INT after diagnostic and clinical procedures.
The DMPM cell lines STO, MP8, and MP115 were established in our laboratory from clinical samples of epithelioid (STO and MP8) and biphasic (MP115) DMPM [14–18]. Human prostate carcinoma cell lines, DU145 and PC3, were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). A human ovarian carcinoma cell line, IGROV-1, was established from a patient with ovarian adenocarcinoma as described [19], and its resistant subline IGROV-1/Pt1 was generated by continuous exposure of parental cells to platinum drugs and was characterized by mutations in the TP53 gene [19, 20]. DMPM cells were cultured in DMEM F-12 medium (Lonza, Basel, Switzerland). Ovarian and prostate cells were maintained in RPMI 1640 medium (Lonza). Both media were supplemented with 10% fetal bovine serum in a 37 °C humidified 5% CO2 incubator. All cells were human mycoplasma-free. All cell lines were authenticated by single tandem repeat analysis by the AmpFISTR Identifiler PCR amplification kit (Applied Biosystems, Waltham, Massachusetts, USA). Synthetic miR-550a-3p mimic (hereafter miR-550a-3p) and mimic negative control (Neg) were purchased as Pre-miR™ miRNA precursor molecules (Thermo Fisher Scientific Inc, Waltham, Massachusetts, USA). Cells were transfected for 24 h with 20 nM of miR-550a-3p or miR-Neg, using Lipofectamine® RNAiMAX Transfection Reagent (Thermo Fisher Scientific Inc) with Opti-MEM I (Gibco, NY, USA) according to the manufacturer’s instructions. In miR-Mask experiments, 20 nM of HSP90AA1-miScript Target Protector (Qiagen, Hilden, Germany) was transfected alone or in combination with miR-550a-3p mimic, under the same transfection conditions described above.
Total RNA from tissue specimens and cell lines was isolated using the miRNeasy Mini Kit (Qiagen) according to the manufacturer’s guidelines, and 1 μg of total RNA was reverse transcribed to cDNA by miScript II RT Kit (Qiagen). miR-550a-3p and HSP90AA1 mRNA expression levels were quantified by quantitative RT-PCR (qRT-PCR) using miScript SYBR Green PCR Kit (Qiagen) and TaqMan®gene expression assays (Thermo Fisher Scientific Inc), respectively (detailed in Material and Methods Supplementary Information). The primers for qRT-PCR were as follows: miScript Primer Assays specific for miR-550a-3p (MS00023807) and normalized on SNORD48 (MS00007511) (Qiagen); HSP90AA1 (Hs00743767, Thermo Fisher Scientific Inc) and normalized on GAPDH (Hs.PT.39a.22214836, Integrated DNA Technologies, Inc. Coralville, Iowa, USA). Amplifications were run on the 7900HT Fast Real Time PCR System (Applied Biosystem). Data were analyzed by SDS 2.2.2 software (Thermo Fisher Scientific Inc) and reported as 2−ΔCt or as relative quantity (RQ = 2−ΔΔCt): being ΔCt the difference between the threshold cycle (Ct) of the target gene and the Ct of the housekeeping gene and ΔΔCt the difference between ΔCt of the sample and ΔCt of the calibrator. The calibrator corresponded to the sample transfected with miR-Neg.
To assess the effect of miR-550a-3p on cell proliferation, cells were trypsinized at different intervals from transfection with miR-Neg or miR-550a-3p, and counted in a particle counter (Beckman Coulter Inc., Brea, California, USA). Results were expressed as percent variation in the number of miR-550a-3p-transfected cells compared with Neg-transfected cells. The anchorage-independent growth assay was performed as described by Cuccuru et al. [21]. Briefly, 24 h after miR-Neg or miR-550a-3p transfection, cells were trypsinized and suspended in a medium containing 0.33% of agarose (Sigma–Aldrich, St. Louis, Missouri, USA) and was layered onto semisolid agarose (0.5% of agarose in medium) in duplicate on 9.6 cm2 dishes. After 10 days of incubation at 37 °C, cell colonies were stained with p-iodonitrotetrazolium violet (Sigma–Aldrich) and counted by Image J software.
Whole-cell lysates were resolved by SDS-PAGE, transferred to nitrocellulose membranes, and probed with specific antibodies, as previously described [22]. The following primary antibodies were used: anti-Caspase-3 (#9662, 1:1000) and Caspase-9 (#9502, 1:1000) (Cell Signaling Inc, Beverly, MA, USA), anti-HSP90 alpha (TA332385, 1:2000; Origene, Rockville, MD, USA), anti-Cdc37 (sc-5617, 1:000), and anti-Raf-1 (sc-133, 1:1000) (Santa Cruz Biotechnology Inc, Santa Cruz, CA, USA), anti-Akt (610861, 1:1000; BD Biosciences, San Jose, CA, USA); anti-p53 (DO-7 M700101-2, 1:1000, Dako, Agilent, Santa Clara, CA, USA). Anti-Vinculin (V9131, 1:5000; Sigma–Aldrich) was used as an equal protein loading control. The filters were then incubated with the secondary peroxidase-linked whole antibodies and detailed in Material and Methods Supplementary Information. Bound antibody was detected using the Novex ECL, HRP Chemiluminescent substrate Reagent Kit (Thermo Fisher Scientific Inc). Membranes were cropped to allow simultaneous incubation of different primary antibodies on the same samples. Membranes were stripped and successfully reincubated with a second antibody, where appropriate. For the preparation of figures, we cropped the original western blot to generate the appropriate figure panels with the relevant lanes. The cropped image was then subjected to uniform image enhancement of contrast and brightness. Molecular weights were determined using the Precision Plus Protein™ Standard (Bio-Rad Laboratories, Hercules, California, USA), which yields a colorimetric image only and has been removed from the chemoluminescence blot image.
At different time points after transfection with miR-Neg or miR-550a-3p, floating and adherent cells were harvested and processed for apoptosis evaluation by TUNEL assay according to manufacturer’s instructions (Roche, Basel, Switzerland). The cells were subjected to FACS analysis (BD Accuri™ C6 Cytometer, Becton Dickinson, Basel, Switzerland).
Invasion assay was performed 72 h after transfection with miR-Neg or miR-550a-3p by using Transwell membranes previously coated with 3.47 μg Matrigel/well (Boyden chamber with 8 mm pore size filter in the inset chambers (Costar, Corning Inc., NY, USA)) according to the protocol of our previous studies [14]. Cells were suspended in 300 µL serum-free medium and seeded into the insert chambers. After 24 h of incubation at 37 °C in 5% CO2, cells that migrated into the bottom chamber containing 1 ml of complete medium were fixed in 95 % ethanol, stained with a solution of 0.4% sulforhodamine B in 0.1% acetic acid, photographed, and counted under an inverted microscope.
Animal experiments were approved by the Ethics Committee for Animal Experimentation of INT, authorized by the Italian Ministry of Health according to the national law (Project approval code: 1120/2015-PR), and performed in compliance with international policies and guidelines. SCID mice (8-week-old female) were purchased from Charles River Laboratories (Charles River Laboratories, Wilmington, Massachusetts, USA). Cells were transfected with miR-550a-3p or miR-Neg for 24 h, and then subcutaneously injected into the mouse right flank (5 × 106cells/mouse). Each experimental group was composed of five mice. Inoculated animals were inspected daily to establish the time of tumor onset. Tumor growth was measured every 2–3 days using a Vernier caliper. The subcutaneous tumor volume was calculated as follows: TV (mm3) = d2 × D / 2, where d and D are the shortest and the longest diameter, respectively. Volume inhibition percentage (TVI %) in tumors derived from miR-550a-3p- over Neg-transfected cells was calculated as follows: TVI% = 100 − (mean miR-550a-3p TV / mean Neg TV × 100).
Putative targets of miR-550a-3p were selected using miRWalk2.0 (http://www.ma.uniheidelberg.de/apps/zmf/mirwalk) algorithm. Predicted targets of miRWalk2.0 are obtained by integration of predicted miRNA targets produced by 12 established miRNA-target prediction programs (miRWalk microt4, miRanda, mirbridge, miRDB, miRMap, miRNAMap, PicTar2, PITA, RNA22, RNAhybrid, and TargetScan). We only selected the targets predicted by at least six of these programs which employ different algorithm predictions.
Twenty-four hours after transfection with miRNA mimics (miR-Neg and miR-550a-3p) STO cells were lysed as described above. RNA was extracted from three independent transfections of STO cells with miRNA mimics and analyzed on Illumina BeadStudio v4 gene expression platform. Scanned images were collected using Illumina BeadStudio v3.3.8 and processed using the lumi package [23] from Bioconductor v3.0 [24]. Raw data were log2-transformed, normalized with Robust Spline Normalization, and filtered, keeping only probes with a detection p-value < 0.01 in at least one sample; probes not associated with the official gene symbol were removed. Expression data were deposited in the Gene Expression Omnibus repository (GEO) with accession number GSE165341. Differentially expressed genes between the two conditions were identified using the limma package [25], and significance was assessed by Benjamini–Hochberg false discovery rate (FDR) method in order to take into account the multiple-testing correction.
The effect of miR-550a-3p ectopic expression in cell-based assays was assessed by using the nonparametric Wilcoxon or Kruskal–Wallis tests according to the number of considered groups [26] and corresponding p-values were estimated according to exact test or via Monte Carlo approaches. The nonparametric Sign Test was used to compare the sample distribution to a given value. For the animal experiment, a mixed model (with a compound symmetry covariance matrix) was fitted to assess the tumor growth (on a logarithmic scale) as a function of time and experimental groups (fixed factors) with mice considered as a random factor. For clinical data, relapse-free survival (RFS) was calculated as the time from surgery to the first relapse and overall survival (OS) as a time to death due to any cause. The prognostic role of miR-550a-3p expression (considered on its original continuous scale) on RFS and OS was investigated using a Cox regression model in a univariate fashion [27]. The relationship between miR-550a-3p expression and outcome was investigated using a regression model based on restricted cubic splines. Subsequently, miR-550a-3p was dichotomized according to the median value and the patterns of RFS and OS were estimated using the Kaplan–Meier method [28], and the survival curves were compared using log-rank tests. All statistical analyses were carried out with SAS (Statistical Analysis System, RRID:SCR_008567, version 9.4.; SAS Institute, Inc., Cary, NC) by adopting an alpha level of 5%.
By comparatively assessing miRNA profiles in 51 DMPM and 5 normal mesothelium samples on a microarray platform (GSE99362) [15], we initially identified miR-550a-3p as a downregulated miRNA in tumors (p = 0.019) (Fig. 1A). The significantly reduced abundance of miR-550a-3p was then confirmed by qRT-PCR in an independent DMPM cohort, including 45 DMPM and 11 normal mesothelium samples (p = 0.003) (Fig. 1B). Consistently, miR-550a-3p showed a trend of down-modulation also in DMPM cell lines (STO, MP8, and MP115) developed in our lab from DMPM clinical samples (Fig. 1C). Thanks to the availability of clinico-pathogical and follow-up information of DMPM patients, we then investigated whether miR-550a-3p expression level is associated with patient clinical outcomes. Clinico-pathological characteristics of the cohort are reported in Table 1. Median follow-up was 31 months (interquartile, 11–136 months). Univariate Cox analysis revealed a statistically significant association between miR-550a-3p expression, considered on its continuous scale, and RFS (HR: 0.802, 95% CI: 0.671; 0.959, p = 0.015) (Supplementary Fig. 1). The significance was retained also when we pursued the analysis by dichotomizing the miR-550a-3p expression (p = 0.007) (Fig. 1D). No statistically significant association was observed between miR-550a-3p expression and patient OS (HR: 0.876, 95% CI: 0.711; 1.080), possibly because of the low number of events.
To assess the biological role of miR-550a-3p in DMPM, we transiently transfected the three cell lines with synthetic miR-550a-3p mimic and miRNA negative control. A marked increase in miRNA abundance was observed in all cell lines at 24 h from transfection with the miRNA mimic, which was still appreciable, to a comparable extent, at 144 h (Fig. 2A, p < 0.05). miR-550a-3p ectopic expression inhibited the proliferation of all DMPM cell lines in a time-dependent manner, although a more rapid cell growth decline was observed in MP8 and MP115 cells than in STO cells (Fig. 2B). Moreover, results obtained in an anchorage-independent growth assay showed the ability of miR-550a-3p to significantly reduce the clonogenic potential of STO and MP8 cells (p < 0.001) (Fig. 2C). The antiproliferative effects of miR-550a-3p were paralleled by the induction of a marked apoptotic response in both cell lines, as indicated by the enhanced percentage of TUNEL-positive cells compared to cells transfected with the miRNA negative control (p < 0.05) (Fig. 2D), as well as by the presence of the cleaved forms of caspase-3 and caspase-9 (Fig. 2E). Interestingly, miR-550a-3p was also able to significantly reduce the invasive capabilities of DMPM cells in a matrigel-based assay (p < 0.001) (Fig. 2F). The proapoptotic effects of miR-550a-3p reconstitution were observed also in MP115 cells (Supplementary Fig. S2). To assess whether the tumor-suppressive functions of miR-550a-3p could be extended to models of tumor types other than DMPM, we ectopically expressed miR-550a-3p in two ovarian cancer (IGROV-1 and its platinum-resistant derivative IGROV-1/Pt1) and two prostate cancer (DU145 and PC3) cell lines (Fig. 3A, p < 0.05). miR-550a-3p overexpression was found to inhibit cell growth in all cell lines, although to a variable extent and with different kinetics (Fig. 3B), and to induce an apoptotic response, as detected by the increase in the percentage of TUNEL-positive cells (p < 0.05) (Fig. 3C and Supplementary Fig. S2) and/or the presence of the cleaved form of caspase-3 (Fig. 3D).
The in vitro tumor growth inhibitory effect of miR-550a-3p was challenged in the in vivo setting by subcutaneously transplanting STO cells transiently transfected with the miRNA mimic and the negative control into nude mice to generate xenografts. No differences were appreciable in the tumor take rate, which was 100% in both experimental groups. However, the growth of tumors originating from miR-550a-3p overexpressing cells was significantly delayed compared to those arising from control cells throughout the experiment (p = 0.02) (Fig. 4A, B), with a maximum tumor volume inhibition of 59% recorded at 19 days after cell inoculum. In light of the in vitro results, such a growth delay might be ascribed to reduced cell proliferation (Fig. 2B, C) and limited local invasive capabilities (Fig. 2F) of miR-550a-3p overexpressing cells.
In the search for molecular determinants through which miR-550a-3p affects proliferation, apoptosis, and invasion, we conducted an in silico target prediction analysis by using the miRWalk2.0 tool [29]. The 220 predicted targets were then crossed with the 123 genes found to be downregulated in STO cells following miR-550a-3p ectopic expression, which we identified by comparatively evaluating gene expression profiles of miR-550a-3p overexpressing and negative control cells (GSE165341) (Fig. 5A and Supplementary Table 1). The resulting intersection included six genes, which are listed in Fig. 5B. Among them, we focused on the heat shock protein 90 alpha family class A member 1 (HSP90AA1) gene coding for the heat shock protein 90 alpha (Hsp90α), the stress-inducible isoform of the molecular chaperone Hsp90, which interacts and supports numerous proteins that promote oncogenesis and is associated with each hallmark of cancer [30]. One 7mer-m8 site (intended as having an exact match to positions 2–8 of the mature miRNA, including the seed and position 8) complementary to miR-550a-3p is actually evident in position from 247 to 253 of HSP90AA1 3’UTR (Fig. 5C). To functionally address this point, a target protection approach was pursued. Specifically, IGROV-1/Pt1 cells overexpressing miR-550a-3p were transfected with a miR-Mask, a custom oligonucleotide designed to be fully complementary to miR-550a-3p binding site within HSP90AA1 3’UTR, to assess whether the disruption of miRNA-target interaction could abolish the miR-550a-3p-mediated repression of HSP90AA1 mRNA. Notably, the miR-Mask was able to completely restore HSP90AA1 transcript levels, thus confirming HSP90AA1 as a direct target of the miRNA (Fig. 5D). Ectopic expression of miR550a-3p in DMPM, ovarian cancer, and prostate cancer cell lines consistently decreased HSP90AA1 expression at both mRNA and protein levels (Fig. 6A,B). In addition, miR-550a-3p overexpression caused a decline in the abundance of HSP90 client proteins (Raf-1, Akt) and co-chaperons (Cdc37) (Fig. 6B). Interestingly, consistent with what was observed with HSP90 inhibitors, such as 17-(allylamino)-17-demethoxygeldanamycin [31, 32], miR-550a-3p induced opposing effects on wild-type and mutant p53. Specifically, p53 was upregulated in wild-type p53-expressing cell lines (STO and IGROV-1) and downregulated in mutant p53 tumor cells (MP8, IGROV-1/Pt1, and DU145) (Fig. 6C). Finally, to prove that the oncosuppressive effects induced by miR-550a-3p are mediated by direct targeting of HSP90AA1, STO, and IGROV-1/Pt1 cells were co-transfected with the miR-550a-3p mimic and the miR-Mask. The presence of miR-Mask was able to almost completely restore HSP90 alpha protein expression levels (Fig. 6D), and also partially prevented Raf-1 downregulation by miR-550a-3p in both cell lines (Fig. 6D). Moreover, the rescue of HSP90 alpha protein expression by the miRNA-Mask reduced the antiproliferative (p < 0.05) (Fig. 6E) and proapoptotic effects (p < 0.05) (Fig. 6F) induced by miR-550a-3p. Collectively, these results suggest that miR-550a-3p directly targets HSP90AA1 and that the observed tumor-suppressive effects caused by miRNA ectopic expression are mediated, at least in part, by the interference with the HSP90 alpha-client protein axis.
Compelling evidence about the functional involvement of miRNAs in cancer onset and progression has emerged from a huge amount of studies carried out on experimental models and clinical samples of a variety of human tumor types. However, concerning malignant mesothelioma, almost all available information has been generated on the pleural variant, and only a couple of studies dealt with DMPM [33]. Specifically, El Bezawy et al. [14] showed that miR-34a is downregulated in DMPM clinical specimens and cell lines. Moreover, miR-34a reconstitution in DMPM cells was found to inhibit proliferation and tumorigenicity, to induce an apoptotic response, and to decline invasion ability, mainly through the downregulation of its targets c-MET and AXL and the interference with the activation of downstream signaling. Cimino-Reale et al. [15] showed that the ectopic expression of miR-380-5p, a miRNA negligibly expressed in telomerase-positive DMPM clinical specimens, negatively interferes with telomerase activity and growth of DMPM cell lines by targeting the telomerase associated protein 1. In this study, we demonstrated that miR-550a-3p is downregulated in DMPM clinical samples and cell lines, and that its expression levels are inversely associated with patients’ outcomes, thus suggesting a possible oncosuppressive role of the miRNA in the disease. Consistently, functional experiments revealed that miR-550a-3p ectopic expression in DMPM cells impairs proliferation and invasiveness, enhances apoptosis, and reduces the growth of xenografts in mice. Interestingly, we also observed antiproliferative and proapoptotic effects following miR-550a-3p ectopic expression in prostate cancer and ovarian cancer cell lines, thus suggesting that the miRNA is endowed with tumor-suppressive functions also in these tumor types. By crossing in silico predicted targets and genes found to be downregulated in DMPM cells following miR-550a-3p ectopic expression, we identified and focused on HSP90AA1 as a direct target through which the miRNA exerts, at least in part, its oncosuppressive functions. Indeed, HSP90AA1 is the gene coding for HSP90 alpha, the stress-inducible isoform of HSP90 belonging to the family of molecular chaperones that have a key role in the stabilization of oncogenic proteins, such as Raf-1, ErbB2, Akt, and mutant p53, thus promoting survival of cancer cells [34]. The involvement of HSP90 in all of the hallmarks of cancer supports the functional role of HSP90 alpha in the pleiotropic effects induced by miR-550a-3p reconstitution in tumor cells. Chaperone signaling pathways are dysregulated in a wide range of tumors [35]. Recent studies reported that HSP90AA1 expression levels are upregulated in tissue and plasma of several cancers and correlated with poor prognosis [36–38]. As a consequence of HSP90AA1 inhibition, the reconstitution of miR-550a-3p resulted in the degradation of HSP90 client proteins, including Raf-1 and Akt. The involvement of HSP90 was also supported by a different p53 modulation upon miR-550a-3p reconstitution in our cellular models. Specifically, p53 was upregulated in wild-type p53-expressing cell lines (STO and IGROV-1 cells) and downregulated in mutant p53 tumor cells (MP8, IGROV-1/Pt1, and DU145 cells). These p53 status-dependent modulations have been previously described after treatment of cancer cells with HSP90 inhibitors, i.e., geldanamycin [31, 32], and are consistent with the cytotoxic and proapoptotic effects we observed in all our cell models. Akt and Raf-1 kinases play an important role in the control of pathways that regulate proliferation and apoptosis. The inhibition of functional HSP90 could coordinately block the transduction of growth factor signaling via the Akt and Raf-1 pathways. Indeed, Jones et al. previously reported that the depletion of Akt and Raf-1 tyrosine kinases after HSP90 inhibition resulted in apoptotic cell death as a consequence of the loss of prosurvival signals [39]. Moreover, the depletion of cdc37, an intracellular cofactor of HSP90, following miR-550a-3p reconstitution may further contribute to the disruption of the HSP90 chaperone machinery by impairing the association of client proteins and preventing protein maturation. Such blockage has been previously described to suppress multiple pathways and to induce growth inhibition in human cancer cells [40, 41]. Consistent with an oncosuppressive role of miR-550a-3p, Ho et al. [38] previously reported that the miRNA was downregulated in breast cancer clinical samples and cell lines, and that its ectopic expression in breast cancer cells impaired proliferation, invasion, and migration as well as tumorigenesis in a xenograft mouse model. Conversely, Yang et al. [42] showed that the miR-550a-3p was upregulated in non-small cell lung cancer (NSCLC) tissues compared to surrounding normal tissues, and that forced overexpression of the miRNA in NSCLC cells promoted proliferation, invasion, and migration through the inhibition of TIMP2. Xiong et al. [43] reported that miR-550a-3p was one of the most upregulated miRNAs in plasma extracellular vesicles from melanoma patients compared to healthy individuals, and that a signature consisting of high expression levels of miR-550a-3p, CDK2, and POLR2A and low expression levels of miR-150-5p was associated with reduced overall survival of melanoma patients. Again, by analyzing TCGA miRNA expression profiles of hepatocellular carcinoma, Qin et al. [44, 45] showed that high miR-550-3p expression levels were associated with reduced progression-free survival. The finding that miR-550a-3p displays opposite roles in different human tumor types is not surprising since miRNAs have been acknowledged to exert either tumor-suppressive or oncogenic functions depending on their expression levels, cell/tissue context, and availability of target genes [12]. This notion reinforces the necessity of a detailed understanding of the functions exerted by a specific miRNA and the precise identification of its key targets relevant to the disease of interest in view of its possible exploitation as a novel target/tool for anticancer therapy. In conclusion, due to the inherent resistance of DMPM to chemotherapy and the lack of alternative effective treatments for patients who are not suitable for or fail after CRS + HIPEC, novel miRNA-based therapeutic approaches are highly desirable. Preclinical data generated in the present study form a solid foundation for promoting the clinical use of a miR-550a-3p-based approach as a novel therapeutic strategy to be pursued for DMPM and potentially extended to other more frequent tumor types, such as prostate and ovarian cancers. miRNA mimics and inhibitors already entered the clinical therapeutic armamentarium in oncology. However, the number of clinical studies with miRNA modulators carried out so far is still limited. The first miRNA-based compound entering a phase I study in patients with advanced solid tumors was MRX34, a miR-34a mimic encapsulated in lipid nanoparticles, showing evidence of activity in a subset of patients. However, the study was closed early due to serious immune-mediated adverse events [46]. The activity of a miR-16 mimic delivered by bacterial minicells targeted to EGFR (TargomiRs) was then evaluated in a phase 1 study of patients with recurrent malignant pleural mesothelioma, showing a partial response in one out of 22 evaluable patients [47, 48]. Moreover, a locked nucleic acid-modified oligonucleotide inhibitor of miR-155 (Cobomarsen, MRG-106) is currently under clinical investigation in patients with cutaneous T-cell lymphoma [49]. A major constraint towards a concrete application of miRNA-based strategies to human cancer therapy is related to the need for safe and efficient delivery. Although several approaches have been tested, including lipid carriers, oligonucleotides with different chemical modifications, viral vectors, and, more recently, extracellular vesicles, strong in vivo evidence for improved efficiency and limited toxicity is still lacking. Dosage concerns and off-target effects also remain major challenges to be overcome for the clinical development of miRNA-based therapies as well as the need for an improvement of currently available information concerning pharmacokinetics of miRNA mimics and inhibitors [50]. Moreover, it must be considered that, as anticancer drugs, also miRNA-based therapeutics might activate resistance mechanisms. In this context, El Bezawy et al. [14] provided evidence of a cytoprotective mechanism against miR-34a reconstitution in DMPM cells through the persistent activation of ERK1/2 and Akt signaling.
Material and Methods Supplementary Information Legends supplementary information Supplementary figureS1 Supplementary figureS2 Table S1 Supplementary Information | true | true | true |
PMC9576757 | Tiantian Xu,Qisheng Ma,Yanan Li,Qing Yu,Peichen Pan,Yawen Zheng,Zhijian Li,Xiufang Xiong,Tingjun Hou,Bin Yu,Hongmin Liu,Yi Sun | A small molecule inhibitor of the UBE2F-CRL5 axis induces apoptosis and radiosensitization in lung cancer | 17-10-2022 | Drug development,Lung cancer,Drug discovery | Protein neddylation is catalyzed by a neddylation activating enzyme (NAE, E1), an E2 conjugating enzyme, and an E3 ligase. In various types of human cancers, the neddylation pathway is abnormally activated. Our previous study validated that the neddylation E2 UBE2F is a promising therapeutic target in lung cancer. Although the NAE inhibitor MLN4924/pevonedistat is currently under clinical investigation as an anti-cancer agent, there are no small molecules available that selectively target UBE2F. Here, we report, for the first time, the discovery, via structure-based virtual screen and chemical optimization, of such a small molecule, designated as HA-9104. HA-9104 binds to UBE2F, reduces its protein levels, and consequently inhibits cullin-5 neddylation. Blockage of cullin-5 neddylation inactivates cullin-RING ligase-5 (CRL5) activity, leading to accumulation of the CRL5 substrate, NOXA, to induce apoptosis. Moreover, HA-9104 appears to form the DNA adduct via its 7-azaindole group to induce DNA damage and G2/M arrest. Biologically, HA-9104 effectively suppresses the growth and survival of lung cancer cells and confers radiosensitization in both in vitro cell culture and in vivo xenograft tumor models. In summary, we discovered a small molecule, designated HA-9104, that targets the UBE2F-CRL5 axis with anti-cancer activity alone or in combination with radiation. | A small molecule inhibitor of the UBE2F-CRL5 axis induces apoptosis and radiosensitization in lung cancer
Protein neddylation is catalyzed by a neddylation activating enzyme (NAE, E1), an E2 conjugating enzyme, and an E3 ligase. In various types of human cancers, the neddylation pathway is abnormally activated. Our previous study validated that the neddylation E2 UBE2F is a promising therapeutic target in lung cancer. Although the NAE inhibitor MLN4924/pevonedistat is currently under clinical investigation as an anti-cancer agent, there are no small molecules available that selectively target UBE2F. Here, we report, for the first time, the discovery, via structure-based virtual screen and chemical optimization, of such a small molecule, designated as HA-9104. HA-9104 binds to UBE2F, reduces its protein levels, and consequently inhibits cullin-5 neddylation. Blockage of cullin-5 neddylation inactivates cullin-RING ligase-5 (CRL5) activity, leading to accumulation of the CRL5 substrate, NOXA, to induce apoptosis. Moreover, HA-9104 appears to form the DNA adduct via its 7-azaindole group to induce DNA damage and G2/M arrest. Biologically, HA-9104 effectively suppresses the growth and survival of lung cancer cells and confers radiosensitization in both in vitro cell culture and in vivo xenograft tumor models. In summary, we discovered a small molecule, designated HA-9104, that targets the UBE2F-CRL5 axis with anti-cancer activity alone or in combination with radiation.
Protein modification via neddylation regulates the stability, activity, or function of substrate proteins by covalently attaching a ubiquitin-like peptide NEDD8 (neural precursor cell expressed developmentally downregulated protein 8) to a substrate protein. The physiological substrates of neddylation modification are cullins, a family of proteins, serving as the molecular scaffolds responsible for assembling cullin-RING ligases (CRLs). Typical neddylation modification is catalyzed sequentially by three enzymes. The NEDD8 is activated by an E1 NEDD8-activating enzyme in the presence of ATP, then transferred to an E2 NEDD8-conguating enzyme via a thioester bond. Finally, an E3 ligase binds both NEDD8-loaded E2 and a substrate (e.g. cullin) to promote the covalent attachment of NEDD8 to the lysine residue on the cullin, leading to activation of CRLs. In mammalian cells, there is a single E1 (NAE), consisting of a catalytic subunit, UBA3/NAEβ and a regulatory subunit, APPBP1/NAE1; two E2s UBE2F and UBE2M (also known as UBC12), and several E3s. UBE2F couples with SAG/RBX2 to promote neddylation of cullin-5, whereas UBE2M couples with RBX1 to promote neddylation of cullins 1–4. Cullin neddylation triggers a conformational change of cullins to activate CRLs. CRLs are multi-component-containing and the largest family of E3 ubiquitin ligases, which degrade ~20% of cellular proteins doomed for proteasome degradation through UPS, thus regulating many key biological processes, including cell cycle progression, DNA replication and repair, signal transduction, and tumorigenesis. A wealth of data from many studies have accumulated in the past decade, showing that in many human cancers, the neddylation pathway is over-activated via overexpression of NEDD8, UBE2F, UBE2M, SAG/RBX2, which is often associated with poor patient survival. Thus, the neddylation pathway can serve as an attractive anti-cancer target. Indeed, MLN4924, also known as Pevonedistat, is a first-in-class NAE inhibitor that has been in many phase I-III clinical trials for patients with leukemia, lymphoma, melanoma, and several advanced solid tumors, including a phase II trial of pevonedistat plus docetaxel in patients with advanced non-small cell lung cancer (NCT03228186), based upon our previous study. Given the fact that MLN4924, as a NAE inhibitor, inhibits the entire neddylation pathway, which is essential for many physiological processes, its cytotoxic side-effect appears unavoidable. Up to now, 13 years since the first report, pevonedistat is still an investigational drug of which the efficacy and safety have not been fully demonstrated. In July 2020, FDA granted pevonedistat as breakthrough therapy designation for the treatment of patients with higher-risk myelodysplastic syndromes (HR-MDS) (https://www.takeda.com/newsroom/newsreleases/2020/takeda-announces-u.s.-fda-breakthrough-therapy-designation-granted-for-pevonedistat-for-the-treatment-of-patients-with-higher-risk-myelodysplastic-syndromes-hr-mds/). However, the Phase III PANTHER (Pevonedistat-3001) study was recently ended for not achieving the primary endpoint of event-free survival at the statistically significant level (https://www.takeda.com/newsroom/newsreleases/2021/takeda-provides-update-on-phase-3-panther-pevonedistat-3001-trial/). Therefore, the discovery of small molecule inhibitors that target down-stream enzyme in neddylation pathway should, in theory, have improved specificity and selectivity with reduced side-effects. Our recent study validated neddylation E2 UBE2F as an attractive target for lung cancer. However, no small molecule inhibitor of UBE2F, to the best of our knowledge, has been reported. In this study, we report the discovery of such a small molecule, designated HA-9104, via a structure-based virtual screen and multiple rounds of chemistry-based optimization. HA-9104 has potent growth suppression and radiosensitizing activities via targeting the UBE2F-CRL5 axis and causing DNA damage, leading to induction of apoptosis and G2/M arrest in lung and pancreatic cancer cells. Thus, HA-9104 may serve as a chemical prototype for future development into a new class of anti-cancer agents or radiosensitizers.
In an effort to identify small molecules with the potential to disrupt the UBE2F-UBA3 interaction via targeting two binding pockets (F56 and V30) on the UBE2F surface, we conducted a structure-based virtual screen of Specs chemical library (http://www.specs.net/) with a total of 240,000 compounds, and have identified HA-1141, a small molecule NAE E1 inhibitor targeting the F56 pocket. Here, we report the discovery, structural optimization, and characterization of HA-9104, a novel small molecule inhibitor of cullin-5 neddylation via virtually targeting the V30 pocket of UBE2F (Fig. 1a). A total of 90 top-ranked compounds identified from the virtual screen and 72 of their homologs were screened via the Western blotting for their potential inhibition of cullin-5 neddylation, a consequence of UBE2F inhibition in H358 lung cancer cells after a 24-h treatment at 20 μM. Among them, the compound iV26 showed the best activity (shown in Supplementary Fig. 1a is 12 such compounds with MLN4924 as positive control) and thus was chosen as a hit compound for further optimizations. For the first round of structural modifications, we took two approaches: First, we modified the 3-chlorobenzoyl group with diverse acyl groups (the R1 group), while retaining the 4-methylbenzoyl group intact, leading to 14 derivatives (iV26-1 to 9, 16, 18, 19, 22, 25). Second, we modified the 4-methylbenzoyl group attached to the piperazine ring with three different acyl groups to generate 11 derivatives (iV26-10 to 15, 17, 20, 21, 23, 24) (Supplementary Fig. 1b). The follow-up Western blotting revealed that compound iV26-9 bearing the cinnamon acyl moiety (Supplementary Fig. 1c) was the best one on both inhibiting cullin-5 neddylation and causing accumulation of NOXA, a known substrate of CRL5 (Supplementary Fig. 1d). The further structural optimizations were focused on the introducing different cinnamon acyl moieties, leading to the discovery of iV26-9-10 bearing an indole ring (Supplementary Fig. 1e), which showed the best activity in inducing NOXA accumulation in both H358 and H2170 lung cancer cells with partial inactivation of cullin-5 neddylation (Supplementary Fig. 1f). The final optimization was focused on the indole ring with different substitutions on the indole ring or nitrogen-hybridized indole ring, and identified iV26-9-10-4 (Fig. 1a) as the best derivative in both selectively inhibiting neddylation of cullin-5 among other cullin family members, and causing remarkable NOXA accumulation (Fig. 1b). The iV26-9-10-4 was designated as HA-9104 and was used in the rest of the studies. To determine the potential binding between HA-9104 and UBE2F in vitro, we used a mesylate form of HA-9104 (Supplementary Fig. 5b) to increase compound solubility. By an in-vitro thermal stability assay, we found that purified UBE2F is rather heat-liable and HA-9104 stabilizes it in all temperatures tested (Fig. 1c). We next performed an in vitro fluorescent thermal shift assay (also called differential scanning fluorimetry, DSF), with continuous temperature increases, and found that HA-9104 caused an increase in melting temperature of UBE2F up to 4 °C in a dose-dependent manner (Fig. 1d). To determine the exact step of NEDD8 transfer that is inhibited by HA-9104, we set up three in-vitro neddylation assays, using purified proteins in the test tubes, to measure the formation of E1~N8 thioester, E2~N8 thioester, and cullin-5 neddylation, mimicking each step of the enzymatic cascade of substrate neddylation. We first validated our assay using MLN4924, a well-characterized inhibitor of neddylation E1, as a positive control. Expectedly, MLN4924 completely inhibited the formation of E1~N8 thioester, but HA-9104 had no such effect at the concentration up to 200 μM (Fig. 1e), indicating that HA-9104 did not target neddylation E1. We next established an E2~N8 thioester assay using purified E1, E2, and NEDD8 in a reaction mixture containing DMSO or increasing concentrations of HA-9104, and found that HA-9104 indeed caused a dose-dependent inhibition of E2/UBE2F~N8 thioester (Fig. 1f) with an IC50 at ~49 μM (Fig. 1g). Finally, we measured in vitro cullin-5 neddylation in a reaction mixture, containing purified neddylation E1, E2, E3, NEDD8, and various concentrations of HA-9104 and found that HA-9104 inhibited cullin-5 neddylation in a dose-dependent manner (Fig. 1h) with an IC50 of ~30 μM (Fig. 1i). Taken together, these in vitro assays clearly demonstrate that HA-9104 inhibits cullin-5 neddylation via blocking the formation of UBE2F~NEDD8 thioester. To determine possible UBE2F pocket(s) onto which HA-9104 binds, we used Glide docking simulations to predict the binding geometries and identified three top binding pockets, ranking from the highest to the lowest: N152 (docking score of −4.7), V98 (docking score of −3.862) and V30 (docking score of −3.325). The N152 pocket involved HA-9104 binding residues of Leu151, Asp154, Phe153, Asn152, His165, Ala162, and Ile159, with residues Asp154 and His165 forming two hydrogen bonds, and Phe153 forming cation-π interaction with HA-9104 (Supplementary Fig. 1g, left panel). The V98 pocket involved residues Arg121, Leu120, Lys97, Val98, Thr112, Cys116 (catalytic cysteine residue), Ile115, Glu114, and Lys99 for HA-9104 binding with three hydrogen bonds (two side chains of Lys97 and Lys99, and one backbone on Ile115), and one cation-π interaction on Arg121 (Supplementary Fig. 1g, middle panel). The V30 pocket involved residues Asn60, Lys61, His63, Asp34, Lys35, and Lys39 with the formation of two hydrogen bonds involving Asp34 side chain and Asn60 backbone (Supplementary Fig. 1g, right panel). We then made three UBE2F mutants targeting these sites, respectively, by replacing the surrounding amino acid residues on these sites with residue alanine to disrupt the hydrogen bonds and cation-π interaction between HA-9104 and UBE2F. These mutants were N152 (F153A/D154A/H165A), V98 (K97A/K99A/R121A), and V30 (D34A). Note that it is predicted to be impossible to disrupt the hydrogen bond between HA-9104 and backbone residues Ile115 (V98 pocket) and Asn60 (V30 pocket). These residues were, therefore, excluded from these mutants. We then determined the thermo-stability of these mutants in comparison with wt (wild-type) UBE2F at a very high final concentration of 40 µg/mL, using Coomassie brilliant blue staining, and found that the V30 mutant had a significantly reduced thermo-stability (Supplementary Fig. 1h), which was excluded for further analysis to keep a fair comparison with UBE2F-wt. We next performed the in vitro thermal shift assay using Western blotting on UBE2F-wt and UBE2F-V98 and UBE2F-N152 mutants in the absence and presence of HA-9104. The results showed that compared to UBE2F-wt, N152 mutant was very thermo-stable and largely resistant to HA-9104-induced stabilization, whereas V98 mutant was a little more thermo-stable and less stabilized by HA-9104 (Supplementary Fig. 1i). We further compared the enzymatic activity between UBE2F-wt and two mutants in catalyzing cullin-5 neddylation, and found that both mutants, particularly V98 (near catalytic cysteine residue/Cys116), had significantly reduced catalytic activity (Supplementary Fig. 1j, k), an effect similar to HA-9104 inhibition, suggesting V98 and N152 pockets play the critical role to facilitate cullin-5 neddylation. Taken together, these results suggest that HA-9104 likely binds to UBE2F through the N152 and/or V98 pockets.
Next, we determined the effect of HA-9104 on the cellular levels of UBE2F in several lines of lung cancer cells. HA-9104 was found to reduce the protein levels of UBE2F in a dose-dependent (Fig. 2a, Supplementary Fig. 2a), as well as a time-dependent manner (Fig. 2b, Supplementary Fig. 2b), while having no or moderate effect on UBE2F mRNA (Fig. 2c, Supplementary Fig. 2c). Interestingly, HA-9104 did not affect the protein levels of UBE2M, a family member of UBE2F (Fig. 2a, b, Supplementary Fig. 2a, b, note that UBE2M was not detectable in H1650 cells), suggesting its selectivity toward UBE2F. Given UBE2F couples with RBX2/SAG to promote cullin-5 neddylation, whereas UBE2M couples with RBX1 to promote neddylation of cullins 1-4, the selectivity of HA-9104 toward UBE2F would suggest its selectivity to inhibit cullin-5 neddylation. Indeed, in all three lung cancer lines and one pancreatic cancer line tested, HA-9104 preferentially inhibited cullin-5 neddylation and caused accumulation of CRL5 substrate NOXA, in both dose- and time-dependent manners, with minor, if any, inhibition of cullins 1-4, the other cullin family members (Fig. 2d, Supplementary Fig. 2d, e). We then investigated the possible mechanism by which HA-9104 reduced the levels of UBE2F protein. We first determined whether HA-9104 enhanced UBE2F protein degradation, and found that it was not the case, since either proteasome inhibitor MG132 or lysosome inhibitor CQ (chloroquine) or even in combination failed to rescue UBE2F reduction by HA-9104 (Fig. 2e, Supplementary Fig. 2f). Given HA-9104 did not affect UBE2F mRNA, nor UBE2F degradation, we next determined whether HA-9104 affected the translation of UBE2F protein, using the ribosome profiling analysis. In comparison to DMSO control, HA-9104 treatment caused a minor-to-moderate increase of the 80S monosome peak, but a moderate decrease of the polysome peaks (Fig. 2f, Supplementary Fig. 2g), suggesting that HA-9104 may have a general impairment effect on mRNA translation. However, UBE2F translation appeared not to be inhibited, since the qRT-PCR analysis of each fraction showed an equal amount of UBE2F mRNA regardless of HA-9104 treatment (Fig. 2g, Supplementary Fig. 2h). Taken together, it appears that HA-9104 reduces UBE2F protein via a mechanism of neither enhanced degradation, nor reduced translation. The ribosomal profile experiments suggested that HA-9104 may have a general impairment effect on mRNA translation. We next carried out a quantitative proteomic analysis to evaluate the inhibition selectivity of HA-9104. H2170 cells were treated with HA-9104 at 10 µM for 24 h, followed by proteomic profiling analysis. Only 46 proteins were identified with a twofold reduction (Supplementary Fig. 2i) (for more details, see http://www.proteomexchange.org with the dataset identifier no. PXD036191). Thus, it appears that HA-9104 is not a general inhibitor of protein synthesis.
We next determined whether HA-9104-induced NOXA accumulation was due to reduced degradation as a result of inhibition of cullin-5 neddylation. Indeed, HA-9104 treatment significantly inhibited NOXA polyubiquitination (Fig. 3a, Supplementary Fig. 3a) and prolonged NOXA half-life (Fig. 3b, c, Supplementary Fig. 3b, c) with no or moderate effect on NOXA mRNA in all three lung cancer cell lines tested (Supplementary Fig. 3d). Since NOXA is a pro-apoptotic protein, we determined the biological consequence of NOXA accumulation. In all three lung cancer cell lines tested, HA-9104 induced significant apoptosis in the dose- and time-dependent manners, as demonstrated by cleavage of PARP and caspase-3 (Fig. 3d, Supplementary Fig. 3e), Annexin V-FITC/PI FACS analysis (Fig. 3e, Supplementary Fig. 3f), and DNA fragmentation (Fig. 3f, Supplementary Fig. 3g). HA-9104 also induced apoptosis in MIAPaCa-2 cells, as evidence by obvious PARP cleavage (Supplementary Fig. 2e). Finally, we investigated if accumulated NOXA was causally associated with apoptosis induction via siNOXA-based rescue experiments and found that NOXA knockdown indeed significantly attenuated the degree of apoptosis (Fig. 3g, h, Supplementary Fig. 3h, i), indicating HA-9104-induced apoptosis is mainly mediated by accumulated NOXA.
We further found, via FACS profiling, that HA-9104 also significantly induced a dose-dependent G2/M arrest in lung cancer cells (Fig. 4a, Supplementary Fig. 4a), which cannot be rescued by NOXA knockdown (Supplementary Fig. 4b), indicating a NOXA independent event. As G2/M arrest is usually induced by DNA damage, we measured the levels of γH2AX upon HA-9104 exposure and found a dose- and time-dependent induction, as revealed by both Western blotting and fluorescent foci assays (Fig. 4b–d, Supplementary Fig. 4c). We then measured what type of DNA damage responses were triggered by HA-9104 by comparing the levels of pATR/pRPA32/pCHK1 and pATM/pCHK2 in non-chromatin and chromatin fractions. HA-9104 treatment caused dose-dependent induction of pATRT1989 and pRPA32S33, with minimal effect, if any, on pCHK1/pATM/pCHK2 (Fig. 4e, Supplementary Fig. 4d). ATR/RPA32 activation is usually triggered by replication stress, which phosphorylates H2AX as a way to monitor proper DNA replication. To explore the potential underlying mechanism by which HA-9104 triggered the replication stress, we carefully examined HA-9104 chemical structure and found a 7-azaindole group, which is likely to form the N-H type of hydrogen bonds with thymine base (T), in analogs to adenine base, leading to the possible formation of HA-9104-DNA adduct, thus triggering replication stress (Supplementary Fig. 4e). To test this hypothesis, we examined four HA-9104 derivatives with a series of substitution of trans-3-indoleacrylcamide, for position 4, 5, and 6 of azaindole moiety, designated as compound HA-91011, HA-91012, HA-91013, and with double methyl-substituted indole as compound HA-9101 (Supplementary Fig. 4f). By chemical structures, neither of these derivatives could form two hydrogen bonds with thymine as HA-9104 did, so their potential binding to DNA was expected to be much weaker. Indeed, in comparison to HA-9104, none of these derivatives reduced UBE2F levels, inhibited cullin-5 neddylation, or induced NOXA accumulation as effective as HA-9104. Although to some extent, HA-91011 and HA-91012 induced RPA32S33 and H2AXS139 phosphorylation in H1650 cells, and HA-91011 also induced G2/M arrest in H358 cells, they were far less effective than HA-9104 (Fig. 4f, g, Supplementary Fig. 4g, h). Thus, it appears that the 7-azaindole group in the HA-9104 structure is responsible for triggering replication stress, γH2AX activation, and G2/M arrest, which is related to UBE2F reduction. Since replication stress could also trigger apoptosis-like cell death, accompanied by ROS production, which in turn enhances the occurrence of replication stress, we next measured the ROS levels upon HA-9104 exposure to lung cancer cells, and found a significant ROS induction in a dose- and time-dependent manner (Fig. 4h). Consistently, all four HA-9104 derivatives had much less effect on ROS generation (Fig. 4i, Supplementary Fig. 4i). Finally, we found that HA-9104 induced DNA damage irreversibly since γH2AX level remained high upon its withdrawal for up to 24 h (Fig. 4j, Supplementary Fig. 4j). Taken together, it appears that HA-9104 mimics adenine to pair with thymine, thus likely forming the HA-9104-DNA adduct to trigger replication stress and ROS generation, which in turn arrests cells at the G2/M phase as a cellular defensive response.
We next determined the anti-cancer activity of HA-9104 in three lines of lung cancer cells and one pancreatic cancer cell. Using ATPlite-based 72-h cell proliferation assay, HA-9104 showed growth suppression activity with the IC50 values ranging from 1 to 5 μM among all cancer cell lines (Fig. 5a, Supplementary Fig. 5a). To increase the solubility of HA-9104, we made three salty forms of HA-9104 (HA-9104-mesylate/#7; HA-9104-triflate/#8; and HA-9104-hydrochloride/#9) (Supplementary Fig. 5b). All salty forms inhibited cullin-5 neddylation, caused NOXA accumulation (Supplementary Fig. 5c) and induced apoptosis (Supplementary Fig. 5d). All had similar or little improved IC50 values, as compared to parental HA-9104 (Supplementary Fig. 5e). We further used a 6-day growth assay to determine the potency of HA-9104 in growth inhibition of lung cancer cells and found an obvious dose-dependent suppression (Fig. 5b). Clonogenic survival assay also demonstrated a dose-dependent suppression of colony formation in H1650 lung cancer cells (Fig. 5c). Note that both H358 and H2170 cells were unable to form colonies under normal culture conditions. Since HA-9104 induced the G2/M arrest, and cells in the G2/M phases were usually more susceptible to radiation exposure, and UBE2F had been reported to confer radiation resistance in cancer cells. We next investigated the radiosensitizing activity of HA-9104. Indeed, HA-9104 significantly sensitized both H1650 and MIAPaCa-2 cells to radiation, with an SER (sensitizing enhancement rate) of 1.41 (Fig. 5d, e) and 1.38, respectively (Supplementary Fig. 5f, g). Thus, HA-9104 is a potent anti-cancer agent with radiosensitizing activity.
Finally, we evaluated the in vivo anti-cancer activity of HA-9104 using the nude mice xenograft models. The H1650 cells (5 × 106) were inoculated into both flanks of nude mice. When tumors reached the size of 80–100 mm3, tumor-bearing mice were then randomly allocated into solvent control and HA-9104-mesylate (#7) treatment groups. The #7 (30 mg/kg) was administrated via intraperitoneal injection once a day for continuous 19 days. The tumor size was measured 2–3 times every week and a growth curve was plotted. The results showed that #7 inhibited in vivo tumor growth, as evidenced by reduced tumor size and significantly delayed growth rate (Fig. 6a) and reduced tumor weight (Fig. 6b) without much toxicity as measured by the body weight (Fig. 6c) and H&E staining of important organs (Supplementary Fig. 6a). At the end of the experiment, we selected five tumors from each group randomly and performed both Western blotting and immune-histochemical staining (IHC) to evaluate the in vivo effect of HA-9104. Indeed, the tumor tissues derived from HA-9104-treated mice had increased NOXA levels, enhanced cleavage of PARP (apoptosis) (Fig. 6d), reduced staining of UBE2F (target), Ki67 (proliferation), and increased staining of γH2AX (DNA damage) (Fig. 6e, f). We also performed in vivo HA-9104 radiosensitization assay with #7 (30 mg/kg) in combination with radiation (1 Gy per day for continuous 15 days, 2 h post compound dosing), as compared with radiation alone. The combination of #7 with radiation achieved a greater suppression of tumor growth than radiation alone (Fig. 6g, h, Supplementary Fig. 6b). Specifically, 4 out of 15 tumors in the radiation alone group, whereas 9 out of 16 tumors in the combination group had complete tumor regression. Thus, HA-9104 showed radiosensitization activity in in vivo xenograft lung tumor model. It is worth noting that severe toxicity, as evidenced by greater than 25% loss of body weight, was observed in one mouse in each group. The fluctuation of body weight was shown in Fig. 6i. In general, the combinational group caused slightly more weight loss than the radiation alone group.
It has been more than a decade since the discovery of the first neddylation inhibitor, MLN4924 (Pevonedistat) targeting NAE with 41 phases I–III clinical trials conducted so far alone or with the combination of several chemotherapeutic drugs (https://clinicaltrials.gov/ct2/results?cond=&term=MLN4924). Many NAE inhibitors have been reported thereafter (for review, see refs. ), including one from our group recently. The most significant one, TAS4464, has completed the phase I clinical trial recently in patients with advanced solid tumors but discontinued due to liver toxicity. Quite a few small molecular inhibitors targeting DCN1-UBE2M/UBC12 interaction have also been reported (for review, see ref. ), including a recent one that protects mice from liver toxicity induced by acetaminophen. Finally, our group has recently reported that gossypol has some inhibitory activity against cullin neddylation by targeting the complex of SAG-CUL5 and RBX1-CUL1. However, there is no report on small molecule inhibitors targeting UBE2F so far. In this study, we report the discovery, through a computer-based virtual screen and structure-based SAR optimization, of HA-9104, as a novel small molecule inhibitor of UBE2F. We provide the following lines of supporting evidence: (1) thermo-stability and thermo-shift assays, as well structure-based docking assay showed that HA-9104 binds to UBE2F likely on the N152 and/or V98 pockets; (2) enzyme-based in vitro assays demonstrated that HA-9104 abrogates UBE2F~NEDD8 thioester formation and inhibits cullin-5 neddylation; and (3) cell-based in vitro assays showed that HA-9104 reduces UBE2F protein levels, and selectively inhibits cullin-5 neddylation to inactivate CRL5 activity and cause NOXA accumulation. To the best of our knowledge, HA-9104 is the first small molecule inhibitor reported to target UBE2F. It is worth noting that our virtual screening was in an attempt to identify small molecules that disrupt the UBE2F-UBA3 binding on the V30 pocket. However, the structure-based docking simulation, and thermo-stability and thermo-shift assays of UBE2F mutants suggest that HA-9104 is likely targeting N152 and/or V98 pockets, leading to altered protein stability and loss of enzymatic activity. One unsolved puzzling issue in our study is how HA-9104 causes a reduction in UBE2F protein levels. HA-9104 did not change the levels of UBE2F mRNA (Fig. 2c, Supplementary Fig. 2c), excluding possible regulation at the transcriptional levels; failed to alter the rate of UBE2F translation (Fig. 2g, Supplementary Fig. 2h), excluding possible regulation at the translational levels. Furthermore, the failure in blockage of HA-9104-induced UBE2F reduction by proteasome inhibitor MG132 or lysosome inhibitor CQ (Fig. 2e, Supplementary Fig. 2f) excluded the possible regulation at the post-translational levels through the degradation by UPS or lysosome system. Nevertheless, we found that the effect of HA-9104 on UBE2F is rather specific since its several structural derivatives failed to reduce UBE2F levels (Fig. 4f, Supplementary Fig. 4g). Future investigation will be geared to elucidate the underlying mechanism of HA-9104 action, such as promoting UBE2F proteolysis. It is worth noting that although the ribosomal profiling assay did not reveal inhibition of UBE2F translation (Fig. 2g, Supplementary Fig. 2h), HA-9104, however, did moderately increase the 80S monosome fraction and reduced polysome fractions (Fig. 2f, Supplementary Fig. 2g), indicative of stalled translation initiation and reduced ribosomal translation process in general. However, HA-9104 is not a general inhibitor of protein translation since the mass spectrometry-based proteomic analysis detected only 46 proteins with a twofold reduction upon HA-9104 treatment (Supplementary Fig. 2i) (also in PXD036191). It is reasonable to speculate that replication stress and ROS generation induced by HA-9104, could impair the mRNA translation to some extent through oxidation of the cysteine on translational regulatory proteins. The detailed involving mechanism is an interesting project for future investigation. It is known that ATR activation converges a variety of the replication stress responses through its major downstream effector kinase CHK1 for replication restart and DNA repair. Interestingly, while ATR activation did cause RPA32 activation to trigger the replication stress response, it failed to activate CHK1 via Ser345 phosphorylation, implying a failure in replication restart and DNA repair. This is indeed consistent with the observation that HA-9104 caused irreversible DNA damage, as evidenced by the failure in the recovery of elevated γH2AX levels back to the basal levels even 24 h after compound removal (Fig. 4j, Supplementary Fig. 4j). We acknowledge that we did not provide experimental evidence to show that HA-9104 directly interacts with DNA, thereby inducing DNA damage. We have attempted to biotin-labeling HA-9104 and its precursor iV26-9 (Fig. S1C). Unfortunately, attachment of biotin to the piperazine group of iV26-9 or 7-azaindole group of HA-9104 abrogated their activities in (1) reducing UBE2F protein levels, (2) blocking cullin-5 neddylation and inducing NOXA accumulation, and (3) inducing DNA damage. Thus, we were unable to label HA-9104 with a fluorescent dye or similar molecule for a direct DNA binding assay. HA-9104 displayed sound activity in the suppression of growth and survival of lung cancer cells in in vitro cell-based assays, with the IC50 values ranging around 1–5 μM. However, the in vivo anti-tumor activity in xenograft tumor models is not very substantial, largely due to two reasons: (1) HA-9104 has a poor solubility even in salty form, and (2) HA-9104 has a very short half-life of 6 min, as measured by an in vitro liver microsomal metabolic stability assay (Supplementary Fig. 5h). The emerging biomaterials, like liposomes, albumin nanoparticles and polymeric micelles may overcome these shortages to some extent, but ultimate improvement relies on thorough SAR optimization to make it oral available with much better pharmacokinetics. Finally, our study, using both in vitro and in vivo tumor models, demonstrated that HA-9104 has sound radiosensitizing activity in both lung and pancreatic cancer cells. This effect is likely attributable to its activity in inducing apoptosis via NOXA accumulation and G2/M arrest via replication stress and ROS generation. In summary, we proposed the following working model in which HA-9104 acts as a novel class of anti-cancer small molecule. HA-9104, on one hand, reduces UBE2F levels (via a yet-to-defined mechanism) to inhibit cullin-5 neddylation, resulting in CRL5 inhibition and NOXA accumulation to trigger apoptosis. HA-9104, on the other hand, causes DNA base adduct (most likely) to activate ATR/RPA32 and replication stress, ROS generation, and G2/M arrest for radiosensitization (Fig. 6j).
Animal experiments were approved by the Animal Ethics Committee of Zhejiang University; animal care was provided under the principles and procedures of the regulatory standards at Zhejiang University Laboratory Animal Center.
Human lung cancer cell lines H2170, H1650, H358, and human pancreatic cancer cell line MIAPaCa-2 were obtained from American Type Culture Collection (ATCC). The culture medium RPMI 1640 was used for H2170, H1650, and H358 cells, whereas DMEM was used for MIAPaCa-2, both containing 10% fetal bovine serum and 1% penicillin/streptomycin. All cells were incubated at 5% CO2, 37 °C, with 95% humidity.
MLN4924 was from ApexBio (#B1036). Chlorhexidine (CHX) was from Sigma-Aldrich (#C7698), whereas hydroxyurea (HU) was from MedChem Express (HY-B0313). SYPRO orange was from Sigma-Aldrich (S5692). The iFluor 680 was from ATT Bioquest (1240).
The plasmids encoding human SAG-CUL5, UBE2F, UBA3/APPBP1, and NEDD8 were constructed and purified as described previously. UBE2F-wt and three UBE2F mutants (V30, V98, and N152) were cloned into the pET-28a vector with a SUMO-tag by Tsingke Biotech (Beijing, China), expressed in Rosetta 2(DE3) pLysS (Novagen). The proteins were purified by Ni-NTA agarose beads (QIAGEN) with His-SUMO tags cleaved by the His-tagged Ulp1 enzyme. The protein mixture was incubated with Ni-NTA agarose beads which bind His-Ulp1 and the cleaved His-SUMO tags, and un-tagged recombinant protein was collected as flow-through, and stored in wash buffer (25 mM HEPES pH 7.5–7.8 and 50 mM NaCl) at −80 °C.
Cells were treated with various compounds, and the protein levels were measured by standard Western blotting analysis. The antibodies used were obtained from the following vendors: CUL-5 (sc-373822, Santa Cruz/SC), CUL-1 (sc-11384, SC), CUL-2 (ab166917, Abcam), CUL-3 (2759S, Cell Signaling Technology/CST), CUL-4A (2699S, CST), CUL-4B (12916-1-AP, Proteintech), NOXA (OP180, EMD Millipore), UBE2F (sc-398668, SC), UBE2M (sc-390064, SC), PARP (9542S, CST), Caspase 3 (9662S, CST), CUL-5 CTD (AV35127, Sigma-Aldrich), CUL-1 CTD (12895-1-AP, Proteintech), NEDD8 (ab81264, Abcam), p-ATR T1989 (2853S, CST), ATR (2790S CST) p-RPA32 S33 (ab211877, Abcam), RPA32 (ab2175, Abcam), p-CHK1 S345 (2348, CST), CHK1 (sc-8408, SC), p-ATM S1981 (5883S, CST), ATM (2873S, CST), p-CHK2 T68 (2197P, CST), CHK2 (6334S, CST), γH2AX (05-636, EMD Millipore), DAPI (D1306, Invitrogen), β-Actin (M1210-2, HuaBio) and α-Tubulin (T8203, Sigma-Aldrich).
All docking simulations were carried out by the Glide module in Schrödinger 9.0, with the crystal structure of human UBE2F (PDB entry: 3FN1), as described with the maximum root-mean-square deviation value setting to 0.3 Å, and the scaling factors for van der Waals radii and partial atomic charge cutoff value setting to 0.8 and 0.15, respectively. The structure of HA-9104 was prepared by the LigPrep module with protonated states generated at pH = 7.0 ± 2.0 with all other parameters set to the default values. HA-9104 was docked into the structure of UBE2F, and the standard precision was used to score and rank the binding affinities.
The assays were performed as described recently. Briefly, purified UBE2F (at a final concentration of 0.4 μg/mL) was incubated with 20 μM HA-9104-mesylate, along with DMSO control, in 50 μL reaction buffer at 25 °C for 10 min, and then heated at various temperatures for 5 min. UBE2F protein levels were measured by Western blotting.
The melting temperature (Tm) values were measured using an in vitro fluorescent thermal shift assay (or differential scanning fluorimetry, DSF) as previously described. Briefly, 2 μM pure UBE2F was incubated with indicated concentrations of HA-9104-mesylate mixed with SYPRO orange, and heated to a temperature gradient of 1 °C/min from 25 to 95 °C. The fluorescence was monitored in an RT-PCR machine (Applied Biosystems StepOneTM) with the ROX filters.
The assays were performed as described recently. Briefly, purified E1 (UBA3/APPBP1, 50 nM) and NEDD8 (200 nM) were incubated with indicated compounds (final DMSO 1%) in a reaction buffer at 4 °C for 10 min. The reaction was initiated by the addition of 200 μM ATP, incubated at 16 °C for 10 min, followed by quenching with SDS loading buffer and Western blotting.
The assays were performed as previously described. Briefly, purified NEDD8 protein was labeled with iFluor 680 dye and designated as iNEDD8. The reaction in a buffer, containing 200 nM iNEDD8, 50 nM UBA3/APPBP1, and 50 nM UBE2F, was initiated by the addition of 200 μM ATP after a 10-min pre-incubation with HA-9104 at various concentrations. The reaction was quenched after 10 min of incubation at 25 °C by adding SDS loading buffer, subjected to SDS–PAGE gel, and detected by an Odyssey two-color infrared laser imaging system (LI-COR, USA).
The assays were performed as previously described. Briefly, the reaction in a buffer, containing 300 nM NEDD8, 25 nM UBA3/APPBP1, 200 nM UBE2F, and 200 nM SAG-Cullin 5 E3 complex (SAG-CUL5CTD) was initiated by the addition of 200 μM ATP after a 20-min pre-incubation with HA-9104 at various concentrations. The reaction was quenched after 15 min of incubation at 25 °C by adding SDS loading buffer, followed by Western blotting.
The assays were performed as previously described. Briefly, the reaction in a buffer, containing 3 μM NEDD8, 50 nM UBA3/APPBP1, 1 μM UBE2F (wt or its V98 and N152 mutants), and 1 μM SAG-CUL5CTD was initiated by the addition of 200 μM ATP. The reaction was quenched after indicated incubation at 25 °C by adding SDS loading buffer, followed by SDS-PAGE gel separating and Coomassie brilliant blue staining.
The assays were performed as described recently. The primers used for qRT-PCR were as follows:5′-GAC TGT TCG TGT TCA GCT CG-3′ and 5′-CAC TCG ACT TCC AGC TCT GCT-3′ for NOXA; 5′-GAC CGG GCA TGG TGT TGG-3′ and 5′-ACC ATC GTC ACG CTT CAG TT-3′ for UBE2F; 5′-GGA GTC AAC GGA TTT GGT-3′ and 5′-GTG ATG GGA TTT CCA TTG AT-3′ for GAPDH as an internal control.
The assays were performed as recently described, using an Auto Gradient Fractionator (Biocomp, Canada). The total RNA was isolated from collected fractions (600 μL/per fraction), followed by the synthesis of cDNAs and qRT-PCR.
H2170 cells were treated with DMSO or HA-9104 (10 μM) for 24 h in duplicates, then washed with ice-cold PBS 3 times, and harvested for quantitative proteomic analyses by Jingjie PTM Biolab Co, Ltd. (Hangzhou, China). The resulting MS/MS data were processed and analyzed using the MaxQuant search engine (v.1.6.15.0). A twofold change threshold, and CV (coefficient of variation) < 0.1 was set as the cut-off in which the difference after HA-9104 treatment was considered significant.
Cells were treated with HA-9104 in the presence of CHX (50 μg/mL) for various time points, followed by Western blotting and densitometry quantification via ImageJ software (NIH).
The assays were performed as previously described. Briefly, cells were transfected with His-tagged ubiquitin for 24 h, then treated with HA-9104 (10 μM), along with DMSO control for another 24 h. MG132 (20 μM) was added to the medium 6 h before harvest. Cells were lysed using a 6 M guanidinium denaturing buffer. An equal amount of whole cell lysate was incubated with Ni-NTA beads at room temperature for 4 h. Beads were then washed with a series of buffers, as previously described. Proteins were then eluted from beads, and subjected to Western blotting for NOXA polyubiquitination with the anti-NOXA antibody.
Cells were treated with HA-9104, stained with Annexin V-FITC staining kit (BD Pharmingen, Germany), and then analyzed by a CytoFLEX S flow cytometer (Beckman Coulter, USA).
Cells after HA-9104 treatment were fixed in ice-cold 70% ethanol at 4 °C overnight and suspended in 500 μL PI/RNase staining buffer (BD Pharmingen, USA) for 15 min in the dark after 3 × PBS washing. The samples were analyzed by a CytoFLEX S flow cytometer for cell cycle profiling.
ROS was measured using dichlorofluorescin diacetate (DCFH-DA) reagent (Beyotime, China) according to the manufacturer’s instructions. Briefly, cells after HA-9104 treatment were washed with serum-free medium, incubated with DCFH-DA at 37 °C for 20 min, followed by 3 × PBS washing. The levels of ROS were detected by a CytoFLEX S flow cytometer.
The assays were performed as previously described. Briefly, cells after HA-9104 treatment were lysed in a lysis buffer. Genomic DNA was isolated and subjected to a 1.8% agarose gel (Invitrogen), and photographed by Gel Doc XR + System (Bio-Rad, USA).
The siRNA transfection was conducted in lung cancer cells with the GenMute siRNA Transfection Reagent (SignaGen Laboratories, USA). The siRNA sequences were as follows: siNOXA: UGC ACG UUU CAU CAA UUU GTT; and siCont: UUC UCC GAA CGU GUC ACG UTT.
Cells after indicated treatment were fixed with ice-cold methanol (30 min at −20 °C), followed by 3× PBS washing, and then stained with anti-γH2AX Ab and DAPI. Cells were then photographed with a Nikon A1 Ti confocal microscope (Nikon, Japan).
Cells after indicated treatment were lysed in lysis I buffer (50 mM HEPES, pH = 7.5, 1 mM EDTA, 150 mM NaCl, 0.1% Triton X-100 (v/v), 1 mM PMSF and 1% protease inhibitor cocktail) on ice. After centrifugation, the supernatant was collected as non-chromatin protein. The pellet was dissolved in lysis II (10 mM Tris-HCl, pH = 7.5, 5% SDS (w/v), and 1% protease inhibitor) with sonication and used as chromatin fraction.
Cells were seeded in 96-well plates in triplicate and treated with HA-9104 at various concentrations for 72 h. Cell growth was assayed by Cell Counting Kit-8 (CCK-8) (MedChem Express) at OD450 in a microplate reader (SpectraMax iD3, Molecular Devices), or by the ATPlite 1 step Luminescence Assay System (PerkinElmer) in a microplate reader. The inhibition curve was drawn by GraphPad Prism software.
Cells were seeded in 60 mm dishes in duplicate and treated with HA-9104 for 24 h after adherence. Cells were then exposed to radiation (X-RAD 160; Precision X-Ray, Inc., Kentwood), cultured at 37 °C for another 7 days, and stained with Coomassie brilliant blue solution. Colonies with greater than 50 cells were counted under an inverted microscope. Survival curves were generated as described. The radiosensitization assay with a calculation of sensitizing enhancement rate (SER) was performed as described. For the clonogenic assay, H1650 cells were seeded 600 per well in six-well plates in triplicate, followed by incubation at 37 °C for another 8 days with HA-9104 (1, 2, 3 μM). The colonies were then stained and counted.
H1650 cells (5 × 106 cells suspended in 100 μL PBS per tumor) were inoculated subcutaneously in both franks into female BALB/C nude mice (5–6 weeks old) (Shanghai Slac laboratory animal Co., Ltd., China). The mice were randomized into control or experimental groups when the tumor size reached approximately 80–100 mm3. #7 in DMSO (30 mg/kg) or DMSO vehicle was administrated to mice by intraperitoneal injection, once a day for 19 consecutive days. For combination treatment, radiation (1 Gy) was given once a day for 15 consecutive days, two hours post #7 injection. Tumor growth/tumor size and body weight were monitored 2–3 times a week, and average tumor volumes were calculated by the formula (L × W2)/2. At the end of the experiment, tumors were harvested, weighed, and photographed. We used a humane protocol in our mouse xenograft tumor growth assay with the endpoints of tumor volume <1500 mm3.
The statistical analysis was assessed using GraphPad Prism software v 7.0 (San Diego, CA. USA). The student’s t-test was used for the comparison of parameters between groups. Three levels of significance (*p < 0.05, **p < 0.01, ***p < 0.001) were presented.
Supplementary Figure | true | true | true |
|
PMC9576931 | Yi Cheng,Nan Huang,Qingqing Yin,Chao Cheng,Dong Chen,Chen Gong,Huihua Xiong,Jing Zhao,Jianhua Wang,Xiaoyu Li,Jing Zhang,Shuangshuang Mao,Kai Qin | LncRNA TP53TG1 plays an anti-oncogenic role in cervical cancer by synthetically regulating transcriptome profile in HeLa cells 10.3389/fgene.2022.981030 | 04-10-2022 | TP53TG1,apoptosis,cervical cancer,RNA-binding proteins,alternative splicing | Long non-coding RNAs (lncRNAs) have been extensively studied as important regulators of tumor development in various cancers. Tumor protein 53 target gene 1 (TP53TG1) is a newly identified lncRNA in recent years, and several studies have shown that TP53TG1 may play oncogenic or anti-oncogenic roles in different cancers. Nevertheless, the role of TP53TG1 in the development of cervical cancer is unclear. In our study, pan-cancer analysis showed that high expression of TP53TG1 was significantly associated with a better prognosis. We then constructed a TP53TG1 overexpression model in HeLa cell line to explore its functions and molecular targets. We found that TP53TG1 overexpression significantly inhibited cell proliferation and induced apoptosis, demonstrating that TP53TG1 may be a novel anti-oncogenic factor in cervical cancer. Furthermore, overexpression of TP53TG1 could activate type I interferon signaling pathways and inhibit the expression of genes involved in DNA damage responses. Meanwhile, TP53TG1 could affect alternative splicing of genes involved in cell proliferation or apoptosis by regulating the expression of many RNA-binding protein genes. Competing endogenous RNA (ceRNA) network analysis demonstrated that TP53TG1 could act as the sponge of several miRNAs to regulate the expression level of target genes. In conclusion, our study highlights the essential role of lncRNA TP53TG1 in the development of cervical cancer and suggests the potential regulatory mechanisms. | LncRNA TP53TG1 plays an anti-oncogenic role in cervical cancer by synthetically regulating transcriptome profile in HeLa cells 10.3389/fgene.2022.981030
Long non-coding RNAs (lncRNAs) have been extensively studied as important regulators of tumor development in various cancers. Tumor protein 53 target gene 1 (TP53TG1) is a newly identified lncRNA in recent years, and several studies have shown that TP53TG1 may play oncogenic or anti-oncogenic roles in different cancers. Nevertheless, the role of TP53TG1 in the development of cervical cancer is unclear. In our study, pan-cancer analysis showed that high expression of TP53TG1 was significantly associated with a better prognosis. We then constructed a TP53TG1 overexpression model in HeLa cell line to explore its functions and molecular targets. We found that TP53TG1 overexpression significantly inhibited cell proliferation and induced apoptosis, demonstrating that TP53TG1 may be a novel anti-oncogenic factor in cervical cancer. Furthermore, overexpression of TP53TG1 could activate type I interferon signaling pathways and inhibit the expression of genes involved in DNA damage responses. Meanwhile, TP53TG1 could affect alternative splicing of genes involved in cell proliferation or apoptosis by regulating the expression of many RNA-binding protein genes. Competing endogenous RNA (ceRNA) network analysis demonstrated that TP53TG1 could act as the sponge of several miRNAs to regulate the expression level of target genes. In conclusion, our study highlights the essential role of lncRNA TP53TG1 in the development of cervical cancer and suggests the potential regulatory mechanisms.
Cervical cancer is the fourth most common malignancy diagnosed in women worldwide (Tsikouras et al., 2016). Nearly all cases of cervical cancer result from the infection of human papillomavirus (HPV) (Vu et al., 2018). In recent years, developed countries have adopted vaccination against HPV and conducted cervical screening with primary HPV testing followed by treatment of precancerous lesions, which is very effective in preventing and controlling the development of cervical cancer. However, cervical cancer incidence and mortality rates remain particularly high in developing countries, mainly in terms of chemotherapy resistance and metastasis (Tsikouras et al., 2016; Kori et al., 2019). It was already known that two main tumor suppressor proteins, p53 and retinoblastoma protein (pRb), are known to be inactivated by the HPV proteins, which disrupts both the DNA repair mechanisms and apoptosis, leading to rapid cell proliferation. Cell proliferation genes and Multiple genes involved in DNA repair become highly expressed in cervical cancer (Balasubramaniam et al., 2019). However, the molecular mechanisms of cervical cancer tumorigenesis and metastasis are far more complex. There is an urgent need to discover new accurate biomarkers and therapeutic targets. Long non-coding RNAs (lncRNAs) have been reported to be important regulators of gene expression and innovative molecular biomarkers in cervical cancer (Silva et al., 2015). Previous studies reported that lncRNAs could act as competing endogenous RNAs (ceRNAs) or molecular sponges to interact with miRNAs and participate in the physiological and pathological process. For example, lncRNA TPT1-AS1 promotes cell growth and metastasis in cervical cancer via acting as a sponge for miR-324-5p which mediates the regulation of SP1 expression (Ji et al., 2019). SNHG7 regulated cervical cancer progression by sponging miR-485-5p, thereby up-regulating JUND expression (Zhao et al., 2020a). Furthermore, lncRNAs could interact with DNA, RNA, protein molecules and/or their combinations, acting as an essential regulator in chromatin organization, and transcriptional or post-transcriptional regulation (Yang et al., 2014). For example, lncAB bound KH-type splicing regulatory protein (KHSRP) and also decreased the expression of KHSRP in papillary thyroid carcinoma, thus increasing CDKN1a (p21) expression and decreasing CDK2 expression to repress cell proliferation (Gou et al., 2018). In addition, lncRNAs may function in alternative splicing by modulating splicing factors (Tripathi et al., 2010). The lncRNA 91H has been verified to regulate hnRNPK-mediated alternative splicing (AS), affecting cancer processes such as metastasis in CRC (Gao et al., 2018). To date, few studies in cervical cancer investigate the role of lncRNAs in post-transcriptional regulation in cervical cancer. The lncRNA TP53TG1 (TP53 target gene 1) is a newly discovered p53-responsive lncRNA induced by DNA damage. It which exerts tumor-suppressive activities by inhibiting cell cycle progression, inducing apoptosis, or suppressing oncogene-induced transformation. Diaz-Lagares et al. (2016) found that the cancer growth suppressor features of TP53TG1 are linked to its ability to bind to RNA-binding protein YBX1 to prevent its nuclear localization and thus the YBX1-mediated activation of oncogenes (Diaz-Lagares et al., 2016). Several studies have suggested that TP53TG1 may act as an oncogene in different tumors. For instance, TP53TG1 promotes the growth and progression of pancreatic ductal adenocarcinoma by acting as a sponge to competitively bind to miR-96 and regulate KRAS expression (Aricò et al., 2019). TP53TG1 under glucose deprivation may accelerate cell proliferation and migration by influencing the expression of glucose metabolism-related genes in glioma (Gao et al., 2021). In spite of its bidirectional functions both in pro-tumor and anti-tumor effect, the role of TP53TG1 in cervical cancer remains unclear is not very clear. We also looked into whether TP53TG1 could be involved in transcriptional regulation and post-transcriptional control during the progression of cervical cancer. To address the above-mentioned issues, we constructed a TP53TG1 overexpression model in HeLa cell line, coupled with RNA-seq sequences, to comprehensively reveal the biological functions of TP53TG1 in HeLa cell and analyze the alterations of transcriptional profiles upon TP53TG1 overexpression. In particular, we identified TP53TG1-regulated alternative splicing events (TP53TG1-RAS) and constructed a co-disturbed network between TP53TG1-RAS and RNA-binding protein genes, the expression of which is regulated by TP53TG1. An integrated “lncRNA/miRNA/mRNA” competing endogenous RNA (ceRNA) network was constructed to reveal potential regulatory relationships mediated by TP53TG1. Moreover we used LncSEA to analysis the biological function of TP53TG1 such as: epithelial mesenchymal transformation, immune, Autophagy/apoptosis, cell growth, coding ability, which may reveal that its biological function is very broad. Our study highlights the essential role of lncRNA TP53TG1 in the development of cervical cancer and discusses new regulatory mechanisms.
The data of 24 cancer types from TCGA project, including gene expression profile and clinical information, were downloaded from the UCSC XENA database (https://xenabrowser.net/datapages/) to determine the expression levels of TP53TG1. Prognosis analysis of TCGA data was performed with GEPIA2 (Tang et al., 2017).
Unique molecular identifier (UMI) count matrix of single-cell RNA-seq data from one cervical cancer tissue sample and one normal adjacent tissue sample was downloaded from GSE168652 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE168652). The 10x UMI count matrix was converted into a Seurat object using the R package Seurat (Butler et al., 2018) (version 4.0.4). Cells with UMI numbers <500 or with detected genes <200 or with over 15% mitochondrial-derived UMI counts were considered low-quality, and thus were removed. Genes detected in less than three cells were removed for downstream analysis. After quality control, the UMI count matrix was log-normalized. Then the top 2,000 variable genes were used to create potential anchors with Find Integration Anchors function of Seurat. Subsequently, Integrate Data function was used to integrate data. To reduce the dimensionality of the scRNA-Seq dataset, principal component analysis (PCA) was performed on an integrated data matrix. With Elbowplot function of Seurat, the top 50 PCs were used to perform the downstream analysis. The main cell clusters were identified by the Find Clusters function offered by Seurat, with resolution set as default (res = 0.6). Finally, cells were clustered into 14 major cell types. Then they were visualized using tSNE or UMAP plots. To identify the cell type for each cluster, gene markers for each cell cluster were detected using the “Find Markers” function in Seurat package (v4.0.4). Then cell types were annotated using the cell markers provided in Gene Expression Omnibus (GEO) dataset (Ho et al., 2019).
pcDNA3.1-TP53TG1 (TP53TG1-overexpressing plasmid) was purchased from Youbio Biotech (Changsha, PRC). Full-length sequence of the plasmid: ccctgtctccagtgggcgtcttgggccccggctctattctgggctgcgggcctgggaagggctcgccgggtgccaaatgagctgtcctaactctgcggggctgcagcttcctgcatgatgctggggagcttggcgcctgacccaggatctagaaggcactctgggcaggccgcgctccgcccacgaaggtacccaaccctctgggatagatgcaggaagcgatggttaagacccattttcacccaacttctcgccgcaggtctggcttaccacacgctcctccccattcccagtgagccgctttttgcagcaccaggcgaacacttacaccagtgctttgtaaaggaatcttattgtccaccccgtgtcttggcaaaagaacagtgatcacacagattcctacttgggctctttcctttaatcttcggaggctgagtttgcccaactcaggtttaaccaccaaggactctgagagctggcaggtctgagtaaccctggtaacaattctcttcaccttatcaaaacctgagctaaaaccaatgcatcagctgatgatgacagcagagagtggcagggctgaggacccaaagtcatttcccaggctggcggagaataaactgccagggagaagaatgagaagacaggagacaaactgtttggaaagctaaatcttccctcttaatgaataaaggtttttgccttgtcttaaaaaaaaa.
GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was used as a control gene for assessing the effects of TP53TG1 overexpression. cDNA was synthesized according to standard procedures and RT-qPCR was performed on the Bio-Rad S1000 with Hieff qPCR SYBR® Green Master Mix (Low Rox Plus; YEASEN, China). The information of primers is presented in Additional File 1. The concentration of each transcript was then normalized to GAPDH mRNA level using 2−ΔΔCT method (Livak and Schmittgen 2001).
The cell proliferation assay was conducted using a Cell Counting kit-8 (CCK-8; Dojindo Molecular Technologies, Inc., Shanghai, China). Briefly, TP53TG1-overexpressing Hela cells and control Hela cells were seeded with 6,000 cells/well in 96-well culture plates. Cells treated with an equal volume of phosphate-buffered saline (PBS) served as controls and vials without cells were used as blank controls. After incubated for 24, 48, and 72 h, 20 μl CCK-8 solution was added to the culture medium and incubated for an additional 3 h. Subsequently, the optical density (OD) of the cells was measured with a PerkinElmer/envision at an absorbance of 450 nm. The cell proliferation rate was calculated using the formula: proliferation rate = (experimental OD value − blank OD value)/(control OD value − blank OD value) × 100%.
For the flow cytometric analysis of cell apoptosis, the transfected cells were incubated at 37°C for 48 h and the living cells were then harvested and washed twice with ice-cold PBS. Viable cells were double-stained with 7-amino actinomycin D and FITC-conjugated Annexin V (Beijing 4A Biotech Co., Ltd.). The percentage of apoptotic cells was calculated as the sum of the right lower and upper quadrants. The number of stained cells was quantified using a flow cytometer (CytoFLEX; Beckman Coulter, Inc.). Cell cycle distribution was quantified using multi-cycle software (FlowJo 10.5.3; FlowJo, LLC).
Total RNA was extracted using TRIzol reagent and was purified twice in phenol-chloroform. To remove DNA, the purified RNA was then treated with RNase-free RQ1 DNase (Promega Corp.) and its quality and quantity were determined by measuring the absorbance at 260 nm/280 nm (A260/A280) using a Smartspec Plus (Bio-Rad Laboratories, Inc.). The integrity of RNA was then verified by 1.5% agarose gel electrophoresis. A total of 10 μg total RNA from each sample was used to prepare a directional RNA-seq library. First, the polyadenylated mRNAs were concentrated with oligo (dT)-conjugated magnetic beads (Invitrogen; Thermo Fisher Scientific, Inc.). The concentrated mRNAs were then iron-fragmented at 95°C, end-repaired and ligated to a 5′ adaptor. Reverse transcription (RT) was performed with RT primer harboring a 3′ adaptor sequence and randomized hexamer. The purified cDNAs were amplified and stored at −80°C for subsequent sequencing. Following the manufacturer’s instructions, the libraries were prepared for high-throughput sequencing. The Illumina Novaseq 6000 (Illumina, Inc.) was used to collect data from 150-bp paired-end sequencing (Illumina, Inc.).
Raw sequencing reads containing more than 2-N bases were first discarded. Subsequently, adaptors and low-quality bases were trimmed off the raw reads using a FASTX-Toolkit (v.0.0.13; http://hannonlab.cshl.edu/fastx toolkit/). Short reads of less than 16 nt were dropped. Then clean reads were subsequently aligned to the GRch38 genome by HISAT2 (Kim et al., 2015). Uniquely mapped reads were ultimately used to calculate read number and paired-end fragments per kilobase of exon per million fragments mapped (FPKM) for each gene.
The gene expression levels were measured using FPKM. The DEGs were screened using the software DEseq2 (Love et al., 2014), which analyzes DEGs. Using the fold change (FC ≥ 2 or ≤0.5) and false discovery rate (FDR<0.05) as the cutoff, the results were analyzed to determine whether a gene was differentially expressed.
The alternative splicing events (TP53TG1-AS) and regulated alternative splicing events (TP53TG1-RAS) between TP53TG1 overexpression (TP53TG1-OE) samples and control samples were defined and quantified using the SUVA pipeline described previously (Cheng et al., 2021a). The frequency and proportion of reads of SUVA AS events (pSAR) were calculated.
TP53TG1-mediated ceRNA network. Miranda (score ≥ 150) and Rnahybrid (p-value ≤ 0.05) were used to predict the target relationship between miRNA and TP53TG1. Finally, the results of the two methods were intersected. miRNA-mRNA target pairs from miRDB (http://mirdb.org) and targetscan (http://www.targetscan.org/vert_80/) databases were used to predict the target relationship between miRNA and TP53TG1-regulated DEGs. The regulatory network was established using Cytoscape (https://cytoscape.org/).
Using the KOBAS 2.0 server (Xie et al., 2011), GO analyses and enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were adopted to predict the functions of genes and calculate the distribution frequency in each functional category. The enrichment of each pathway (corrected p < 0.05) was defined using hypergeometric tests and the Benjamini-Hochberg FDR controlling procedure.
One-way analysis of variance and GEPIA2 were used for comparison of expression levels of TP53TG1 among 24 TCGA cancer types, with disease state (Tumor or Normal) as the variable for calculating differential expression. Student’s t-test was used for all other comparisons between the TP53TG1-OE and control groups. For each assay, the results are presented as the mean ± standard error of the mean values of three experiments. The data were analyzed using R software (v3.5.3, https://www.r-project.org/). p < 0.01 or FDR<0.05 was set as the threshold to indicate a statistically significant difference.
Inspired by previous discoveries that TP53TG1 may play oncogenic or anti-oncogenic roles in different cancers (Zhang et al., 2019b; Islam et al., 2021), TP53TG1 expression was compared between tumor and normal tissues in 24 cancer types from TCGA. TP53TG1 expression in tumors was significantly up-regulated (p-value < 0.05) in 15 of the 24 cancer types and was significantly down-regulated only in two cancer types, including colorectal and gastric cancer (Figure 1A). This is consistent with that p53 transcriptionally activates TP53TG1 when the DNA-damaging agent is used except that tumor-specific promoter CpG island hypermethylation-associated silencing of the lncRNA TP53TG1 occurs in colorectal and gastric cancer cells (Diaz-Lagares et al., 2016). Furthermore, GEPIA2 was used to comprehensively analyze the association of TP53TG1 expression with survival rates in 2,376 patients. Pan-cancer analysis shows that higher TP53TG1 expression was positively correlated with a higher overall survival rate (Figure 1B), which confirmed that TP53TG1 exhibits tumor suppressor-like feature in cancer patients (Diaz-Lagares et al., 2016). Furthermore, the association of TP53TG1 expression with survival rates in various types of cancer was also depicted using the survival map (Figure 1C). Surprisingly, we found that TP53TG1 showed positive or negative correlations with prognosis in different cancers, and that this correlation was significant in only five of them. Taken together, these results suggest that TP53TG1 may have a dual role in cancer development, with a predominant oncogenic function. Its function is associated with regulatory networks in cancerous tissue. For cervical cancer (CESC), the expression of TP53TG1 in tumor tissue was higher than that in adjacent normal tissue (Figure 1A; Supplementary Figure S1A). Prognostic analysis showed that high expression of TP53TG1 was associated with a good prognosis (Figure 1B; Supplementary Figure S1B). Since bulk RNA-seq obtains the average gene expression of various cells, to know whether TP53TG1 is highly expressed in malignant cells or in other cell types, we downloaded and analyzed the single cell data of cervical cancer tissue and adjacent tissue, respectively. (GSE168652) (Li et al., 2021) (Figure 1D). We found that TP53TG1 was highly expressed in cell types C0 (cancer cells), C2 (cancer cells), C6 (endothelial cells), C9 (cancer cells) and C13 (cancer cells) (Figure 1E). These results reveal that TP53TG1 is activated in many malignant tumors and has a potential inhibitory effect on cancer development, resulting in a better prognosis.
To further investigate the function of TP53TG1 in cervical cancer, a TP53TG1-overexpression (TP53TG1-OE) cell model was constructed by transfecting a TP53TG1-overexpression plasmid into HeLa cells, with empty vector as negative control (NC). The overexpression of TP53TG1 was verified by RT-qPCR. The result showed that TP53TG1 gene expression significantly increased in TP53TG1-OE group (p < 0.001; Figure 2A). To characterize the role of TP53TG1 in regulating cell apoptosis and proliferation in HeLa cells, flow cytometric analyses and CCK8 assay were performed, respectively. Cell proliferation significantly decreased in the TP53TG1-OE group after 24 h (p < 0.01; Figure 2B; Supplementary Figure S2A). Meanwhile, TP53TG1-OE significantly promoted cell apoptosis (p < 0.001; Figure 2C). Together, these results indicated that TP53TG1 plays a tumor suppressor role in the progression of Hela cells.
LncRNAs participate in the regulation of gene expression and play vital roles in various biological and pathological processes (Liao et al., 2016). To investigate TP53TG1-mediated transcriptional or post-transcriptional regulation in HeLa cells, cDNA libraries for TP53TG1-OE cells and control cells were constructed for RNA-seq using the Illumina HiSeq Novaseq 6000 platform. Three biological replicates for each group were used and a total of 113–135 Million (M) 150-nucleotide paired-end raw reads per sample were obtained. After removing adaptors and low-quality reads, 103.5–125.4 M clean reads were aligned to the human GRCH38 genome using HISAT2, of which 86.27%–90.24% were uniquely aligned for further analysis (Supplementary Table S1). The level of gene expression was calculated in the units of FPKM (Supplementary Table S2). A total of 26,951 expressed genes were assessed by RNA-seq. Effective overexpression of TP53TG1 was further confirmed by the use of RNA-seq analysis (Figure 3A). The differentially expressed genes (DEGs) between the TP53TG1-OE and control cells were determined using an absolute FC ≥ 2 and a 5% FDR as criteria coupled with the DESeq2 package (Love et al., 2014). A total of 704 up-regulated and 470 downregulated DEGs were identified (Supplementary Table S3). The DEGs associated with TP53TG1-OE are displayed in a volcano plot (Figure 3B). To further explore the potential biological roles of these DEGs, GO (Supplementary Table S4) and KEGG (Supplementary Table S5) enrichment analyses were performed. The top 10 GO terms in the category biological process were shown in Figure 3C and Figure 3D. And the top 10 KEGG terms in the category biological process were shown in Supplementary Figures S3A,B. The upregulated genes in the TP53TG1-OE cells are mainly enriched in type I interferon signaling pathways, response to virus, and positive regulation of cell migration terms (Figure 3C). Previously, type I interferon was reported to inhibit the growth of tumor stem cells and promote the apoptosis of cancer cells (Bekisz et al., 2010; Aricò et al., 2019). The downregulated genes were mostly associated with DNA damage responses, such as cellular response to DNA damage stimulus and DNA repair. Genes involved in RNA splicing and regulation of cell migration were also significantly enriched in downregulated genes (Figure 3D). DNA damage is closely related to apoptosis in cancer (Zhang et al., 2007; Figueroa-González and Pérez-Plasencia, 2017). These results indicated that TP53TG1 could active the type I interferon signaling pathways and suppress DNA damage responses in HeLa cells, which ultimately promotes the apoptosis of cancer cells. To confirm the regulatory function of TP53TG1 in gene expression in the HeLa cell line, several DEGs were involved in type I interferon signaling pathways, cellular response to DNA damage stimulus and DNA repair, which were validated by RT-qPCR (Figure 3E). And these genes might be related to prognosis (Supplementary Figure S3C).
LncRNAs were shown to interact with transcriptional factors and thus affect transcription. They are also able to regulate the alternative splicing (AS) by interacting with splicing factors (SFs) or other RNA-binding proteins (RBPs) (Zhang et al., 2019a). Inspired by these previous discoveries, we compared the DEGs with reported RBP gene lists (Castello et al., 2012; Gerstberger et al., 2014; Castello et al., 2016; Hentze et al., 2018). We found that 72 RBP genes were upregulated and 40 RBP genes were downregulated in TP53TG1-OE cells (Figure 4A). Herein we propose that TP53TG1 could regulate alternative splicing by modulating the expression of RBPs. We used the recently published AS analysis software SUVA to identify AS events significantly altered between TP53TG1-OE and NC group (TP53TG1-RAS). As shown in Figure 4B, by applying a stringent cut-off of p ≤ 0.05 and an AS ratio ≥0.2, 470 high-confidence RASEs were identified (Supplementary Table S8) (Cheng et al., 2021a). There are five RAS types: 249 alt3p, 156 alt5p, 13 contain, 14IR and 38 Olp. The splicing events were corresponding to classical splicing events, in which A3SS events accounted for a large proportion (Figure 4C), which may be one of the characteristics of TP53TG1-RAS. As shown in Supplementary Figure S4B, Alt3p and alt5P events, A3SS and A5SS events dominate the alternative splicing events. Further, about 60% of TP53TG1-RAS events were complex splicing events (Supplementary Figures S4C,D), indicating the complexity of AS regulation by TP53TG1. A splicing event involves two transcripts, and these two transcripts may only account for a very small part of the expression of the whole gene. We hope to find a more dominant transcript undergo AS which was quantified as “pSAR” value by SUVA. Here principal component analysis (PCA) was conducted by using the splicing ratio of different variable splicing events, and it could be seen that the principal components of the two groups could be well separated. And the samples’ correlation is strong (Supplementary Figure S4E). The number of splicing events accounting for different pSAR was shown (Figure 4C). Finally, 470 main splicing events with pSAR ≥ 50% were selected for further analysis (Figure 4D). As shown in Supplementary Figures S4C,D, the TP53TG1-OE group and the NC group can be clearly distinguished by the splicing ratio of TP53TG1-RAS (pSAR ≥ 50%). Overall, genes linking with TP53TG1-RAS were primarily enriched in the proliferation, apoptosis (Supplementary Table S7) and signal transduction by P53 mediator pathways (Figure 4E). As shown in Figure 4F, the expression of differentially expressed RBPs and the splicing ratio of TP53TG1-RAS events associated with proliferation and apoptosis were used for correlation analysis (Pearson’s correlation coefficient ≥0.8, p ≤ 0.01). These RBPs might potentially regulate TP53TG1-RAS (Supplementary Figure S4F). The size of nodes in the figure represents the number of gene/splicing events associated with them, including two intron retention (IR), one A5SS, two ES and 1 CE which were located on BCLAF (Supplementary Figure S4G).
Here represents the number of gene/splicing events associated with them, including two intron retention (IR), one A5SS, two ES and 1 CE which were located on MVD, CNBP (Figures 5A,B).
CeRNAs have emerged as an important mechanism for lncRNA and miRNA regulatory network (Su et al., 2021). To investigate TP53TG1-mediated ceRNA network, miranda (Enright et al., 2003) and rnahybrid (Rehmsmeier et al., 2004) were used to predict the target relationship between miRNA and TP53TG1, respectively. Then, seven overlapping miRNAs, containing has-miR-6799-5p, has-miR-1273h-5p, has-miR-6779-5p, has-miR-6807-3p, has-miR-6510-3p, has-miR-6732-5p, and has-miR-1972, were screened. The seven miRNAs were mapped into miRDB (http://mirdb.org) and targetscan (http://www.targetscan.org) databases to explore their target genes, accompanied with TP53TG1-regulated DEGs. Finally, integrating the TP53TG1/miRNA interactions with the miRNA/DEGs interactions, a ceRNA network was constructed (Figure 6). The result reveals that lncRNA-TP53TG1 could adsorb miRNAs as a sponge RNA. The predicted ceRNA network comprised 426 nodes (13 miRNAs and 413 mRNAs) and 743 interactions. In this network (Figure 6), several nodes and interactions involving the feature genes play essential roles in CESC, such as hsa-miR-6779-5p (target genes: OAS3, HDAC2 and IFI6), hsa-miR-1972 (target genes: PINX2, OAS2, ZC3H12C), and hsa-miR-6807-3p (target genes: PDCD6IP, KIN, and EZH2), and all these target genes were upregulated. MiR-6779-5p was also identified as a prognosis-related miRNA (Yang et al., 2022).
In this study, we investigated the role of TP53TG1 in cervical cancer and explored whether TP53TG1 could function in transcriptional or post-transcriptional regulations. First, by analyzing Pan-cancer data from TCGA, a total of 17 out of 24 cancer types exhibited upregulation of TP53TG1 and only two cancer types exhibited downregulation. High expression of TP53TG1is associated with a good prognosis. Single-cell data and bulk RNA-seq analysis revealed that TP53TG1 is mainly activated in malignant cells and has a potential inhibitory effect on CESC cells, suggesting that TP53TG1 may play an important role in CESC development. On the basis of these findings, the HeLa cell line was used as a model to analyze the consequences of TP53TG1 overexpression. The overexpression of TP53TG1 promoted cell apoptosis and inhibited cell proliferation. Transcriptome sequencing demonstrated that the overexpression of TP53TG1 had a broad effect on gene expression. Functional clusters of DEGs highly enriched in cancer-associated terms were obtained. The upregulated DEGs were enriched in the type I interferon pathway and the downregulated DEGs were enriched in DDR pathway. Furthermore, overexpression of TP53TG1 significantly regulated the AS of genes involved in the mRNA splicing, apoptotic process and DNA repair pathways mediated by the expression alternation of dozens of TP53TG-regulated RBPs. Finally, a predicted ceRNA network was constructed. TP53TG1, known as a p53-induced lncRNA promotes apoptosis of cancer cells. In colorectal and gastric cancer, TP53TG1 is downregulated due to DNA methylation (Diaz-Lagares et al., 2016). Consistently, we found that the expression of TP53TG1 is up-regulated in most cancers except in colorectal and gastric cancers, perhaps due to chemotherapy and DNA methylation. Pan-cancer analysis also shows that higher expression of TP53TG1 was associated with a better prognosis, indicating that TP53TG1 inhibits the progression of cancer. However, TP53TG1 may act as a tumor suppressor gene or an oncogene in different tumors (Zhang et al., 2019b; Lu et al., 2021). It functions as an oncogene in pancreatic ductal adenocarcinoma (PDAC), retinoblastoma, and nasopharyngeal carcinoma (Yuan et al., 2017; Zhang et al., 2019b; Shao et al., 2020; Tang et al., 2020; Cheng et al., 2021a; Gao et al., 2021; Lu et al., 2021; Wang et al., 2021). In some other cancers, such as non-small cell lung cancer, hepatocellular carcinoma, gastrointestinal cancer, and cutaneous melanoma, TP53TG1 serves as a cancer suppressor (Benfodda et al., 2018; Xiao et al., 2018; Chen et al., 2021; Masoumi et al., 2021). Through TCGA analysis, expression of TP53TG1 is positively or negatively correlated with overall survival time in different cancers (Zhang et al., 2019b). These findings suggest that TP53TG1 plays a dual regulatory function in cancer and that its function varies in different cancer types and environment. TP53TG1 has been reported to be cancer-promoting in cervical cancer (Liao et al., 2022), but prognosis analysis showed its cancer-inhibiting function. Moreover, we also performed single cell RNA-seq analysis of tumor and adjacent normal tissue and found for the first time that TP53TG1 is predominantly expressed in cancer cells. Previous studies reported that the high expression of TP53TG1 inhibited cancer cell apoptosis in PDAC and breast cancer (Zhang et al., 2019b; Shao et al., 2020) and promoted cancer cell apoptosis in non-small cell lung cancer, hepatocellular carcinoma, gastrointestinal cancer, and cutaneous melanoma (Benfodda et al., 2018; Xiao et al., 2018; Chen et al., 2021; Masoumi et al., 2021). In this study, we constructed a TP53TG1-OE model in HeLa cell line to explore its role in cervical cancer. Conclusively, TP53TG1 significantly promotes apoptosis in Hela cells, inconsistent with existing reports that TP53TG1 inhibits apoptosis in cervical cancer (Liao et al., 2022). Whereas, the results we obtained were consistent with the results of the prognostic analysis. Previous studies have reported that TP53TG1 has a dual role in cancer (Lu et al., 2021). We hypothesized that TP53TG1 may have both tumor-suppressing and tumor-promoting effects in different cells under different conditions. In the future, we will explore the functions of TP53TG1 in more cell lines. Over the last decade, type I interferons (IFNs) have been studied extensively for inducing apoptosis in tumor cells (Pokrovskaja et al., 2005). IFN treatment could induce the tumor suppressor gene p53 (Bekisz et al., 2010). Furthermore, IFNs can up-regulate caspase-4 and caspase-8, which can activate the initiator caspases-8 and 9, as well as the effector caspase-3, USP18 and STAT3 in increased sensitivity of cells to pro-apoptotic (Matsui et al., 2003, Schreiber and Piehler 2015; Thyrell et al., 2002). Currently, upregulated gene of TP53TG1-OE were mostly enriched in type I interferon signaling pathway, which may be the direct reason for its pro-apoptotic effect. The activation of IFNs may induce p53 and further induce the expression of TP53TG1 to form a positive cycle. Next, we compared the RNA-seq data in cancer cells with those in clinical samples. We found five up-regulated expression IFNI-associated genes such as IFI6, OAS1, OAS3, IFIT1 and OASL. These five upregulated genes are generally associated with tumorigenesis and progression, including impressed cell proliferation and induced apoptosis (Matsui et al., 2003; Mullan et al., 2005; Hovanessian, 2007; Niu et al., 2016; Boehmer et al., 2021). Interestingly, overexpression of TP53TG1 also upregulates the expression of genes which promote cell migration, indicating the beneficial effect of TP53TG1 on cancer cell metastasis, this finding being consistent with the previous report that TP53TG1 may act as a tumor suppressor gene or an oncogene in different tumors (Zhang et al., 2019b). DNA damage response (DDR) is crucial for maintaining genome stability in cancer cells (Yoshida and Miki 2004). TP53TG1 is a lncRNA that is critical for the correct response of p53 to DNA damage (Diaz-Lagares et al., 2016). This prompted us to investigate how TP53TG1 regulates the DDR pathway in cancer cells, which could affect the sensitivity of chemotherapy drugs. Surprisingly, we found that genes involved in DDR pathways were significantly downregulated in/by the overexpression of TP53TG1. The overexpression of TP53TG1 could significantly reduce the DNA damage repair efficiency in cancer cells, so as to promote apoptosis of cancer cells. These gene, such as FOXP1, YY1, USP28, NUCKS1, and PCLAF were related to DDR pathway. PCLAF is highly expressed in neuroblastoma, which can accelerate neuroblastoma cell proliferation and cell cycle progression and restrain cell apoptosis (Liu et al., 2022). These genes are generally associated with tumorigenesis and progression, including induced cell proliferation and inhibited apoptosis (Wu et al., 2018; Zhao et al., 2019; Zhao et al., 2020b; Liu et al., 2020). Their downstream targets are regulated by TP53TG1 and may influence cancer progression in cells or clinical samples. The mechanism of lncRNAs regulating AS involved interactions with RNA-binding proteins especially with splicing factors, which control AS through the interaction with pre-RNA sequences (Gordon et al., 2019). TP53TG1 binds to RNA-binding protein YBX1 to block its function and acts as an anti-cancer agent (Diaz-Lagares et al., 2016). In this study, we found that TP53TG1 regulates the expression of a large number of RBPs and we speculated that TP53TG1 may play a prominent role in post-transcriptional regulation, such as in alternative splicing, by regulating the expression of RBPs. We are particularly concerned that the alternative splicing of the p53 pathway could prevent and repair damaged DNA and create feedback loops that enhance or attenuate p53 activity and communicate with other signal transduction pathways. Some proliferation- and apoptosis-related genes are significantly regulated by TP53TG1. These genes could provide a new mechanism for TP53TG1 in post transcriptional regulation. Meanwhile, we also established a ceRNA network based on predicted interactions among TP53TG1, miRNAs, and DEGs. TP53TG1 is predicted to serve as a sponge of miRNAs in gene expression. In the future research, we aspire to discover how TP53TG1 regulates the expression of important genes and RBPs, thereby affecting the splicing. | true | true | true |
|
PMC9576935 | Yunwei Wang,Ming Li,Jiaoxia Zeng,Yunshu Yang,Zengshan Li,Sijun Hu,Fangfang Yang,Na Wang,Wenlan Wang,Jun Tie | MiR-585-5p impedes gastric cancer proliferation and metastasis by orchestrating the interactions among CREB1, MAPK1 and MITF | 04-10-2022 | MITF,CREB1,MAPK1,MiR-585-5p,Gastric Cancer | Background Gastric cancer (GC) is one of the most malignant and lethal cancers worldwide. Multiple microRNAs (miRNAs) have been identified as key regulators in the progression of GC. However, the underlying pathogenesis that miRNAs govern GC malignancy remains uncertain. Here, we identified a novel miR-585-5p as a key regulator in GC development. Methods The expression of miR-585-5p in the context of GC tissue was detected by in situ hybridization for GC tissue microarray and assessed by H-scoring. The gain- and loss-of-function analyses comprised of Cell Counting Kit-8 assay and Transwell invasion and migration assay. The expression of downstream microphthalmia-associated transcription factor (MITF), cyclic AMP-responsive element-binding protein 1 (CREB1) and mitogen-activated protein kinase 1 (MAPK1) were examined by Immunohistochemistry, quantitative real-time PCR and western blot. The direct regulation between miR-585-5p and MITF/CREB1/MAPK1 were predicted by bioinformatic analysis and screened by luciferase reporter assay. The direct transcriptional activation of CREB1 on MITF was verified by luciferase reporter assay, chromatin immunoprecipitation (ChIP) and electrophoretic mobility shift assays (EMSAs). The interaction between MAPK1 and MITF was confirmed by co-immunoprecipitation (Co-IP) and immunofluorescent double-labelled staining. Results MiR-585-5p is progressively downregulated in GC tissues and low miR-585-5p levels were strongly associated with poor clinical outcomes. Further gain- and loss-of-function analyses showed that miR-585-5p possesses strong anti-proliferative and anti-metastatic capacities in GC. Follow-up studies indicated that miR-585-5p targets the downstream molecules CREB1 and MAPK1 to regulate the transcriptional and post-translational regulation of MITF, respectively, thus controlling its expression and cancer-promoting activity. MiR-585-5p directly and negatively regulates MITF together with CREB1 and MAPK1. According to bioinformatic analysis, promotor reporter gene assays, ChIP and EMSAs, CREB1 binds to the promotor region to enhance transcriptional expression of MITF. Co-IP and immunofluorescent double-labelled staining confirmed interaction between MAPK1 and MITF. Protein immunoprecipitation revealed that MAPK1 enhances MITF activity via phosphorylation (Ser73). MiR-585-5p can not only inhibit MITF expression directly, but also hinder MITF expression and pro-cancerous activity in a CREB1-/MAPK1-dependent manner indirectly. Conclusions In conclusion, this study uncovered miR-585-5p impedes gastric cancer proliferation and metastasis by orchestrating the interactions among CREB1, MAPK1 and MITF. | MiR-585-5p impedes gastric cancer proliferation and metastasis by orchestrating the interactions among CREB1, MAPK1 and MITF
Gastric cancer (GC) is one of the most malignant and lethal cancers worldwide. Multiple microRNAs (miRNAs) have been identified as key regulators in the progression of GC. However, the underlying pathogenesis that miRNAs govern GC malignancy remains uncertain. Here, we identified a novel miR-585-5p as a key regulator in GC development.
The expression of miR-585-5p in the context of GC tissue was detected by in situ hybridization for GC tissue microarray and assessed by H-scoring. The gain- and loss-of-function analyses comprised of Cell Counting Kit-8 assay and Transwell invasion and migration assay. The expression of downstream microphthalmia-associated transcription factor (MITF), cyclic AMP-responsive element-binding protein 1 (CREB1) and mitogen-activated protein kinase 1 (MAPK1) were examined by Immunohistochemistry, quantitative real-time PCR and western blot. The direct regulation between miR-585-5p and MITF/CREB1/MAPK1 were predicted by bioinformatic analysis and screened by luciferase reporter assay. The direct transcriptional activation of CREB1 on MITF was verified by luciferase reporter assay, chromatin immunoprecipitation (ChIP) and electrophoretic mobility shift assays (EMSAs). The interaction between MAPK1 and MITF was confirmed by co-immunoprecipitation (Co-IP) and immunofluorescent double-labelled staining.
MiR-585-5p is progressively downregulated in GC tissues and low miR-585-5p levels were strongly associated with poor clinical outcomes. Further gain- and loss-of-function analyses showed that miR-585-5p possesses strong anti-proliferative and anti-metastatic capacities in GC. Follow-up studies indicated that miR-585-5p targets the downstream molecules CREB1 and MAPK1 to regulate the transcriptional and post-translational regulation of MITF, respectively, thus controlling its expression and cancer-promoting activity. MiR-585-5p directly and negatively regulates MITF together with CREB1 and MAPK1. According to bioinformatic analysis, promotor reporter gene assays, ChIP and EMSAs, CREB1 binds to the promotor region to enhance transcriptional expression of MITF. Co-IP and immunofluorescent double-labelled staining confirmed interaction between MAPK1 and MITF. Protein immunoprecipitation revealed that MAPK1 enhances MITF activity via phosphorylation (Ser73). MiR-585-5p can not only inhibit MITF expression directly, but also hinder MITF expression and pro-cancerous activity in a CREB1-/MAPK1-dependent manner indirectly.
In conclusion, this study uncovered miR-585-5p impedes gastric cancer proliferation and metastasis by orchestrating the interactions among CREB1, MAPK1 and MITF.
GC (GC) is a highly malignant and lethal malignancy worldwide, having over 1 million estimated novel cases annually (1). Although advances were achieved for early diagnosis and therapy in GC, cases of unresectable GC are limited to life-prolonging palliative care options (2). Moreover, the underlying mechanism that governs GC malignancy remains uncertain. Hence, it’s urgent to explore the intrinsic molecular mechanism to find new effective therapeutic targets. MicroRNAs (miRNAs), a family of critical small non-coding RNAs, directly bind onto 3’-untranslated regions (3’UTRs) of designated transcripts, leading to translational inhibition or obliteration, thus endowing miRNAs with an crucial role in regulating multiple biological processes (3). Multiple miRNAs were found to be essential regulators within GC progression. In particular, we previously discovered that miR-218-5p is highly downregulated within GC tissue-types and displays pivotal inhibition of oncogenesis and the development of GC (4, 5). The gene encoding miR-218-5p is situated within intron of SLIT2 and SLIT3 (6), and the miR-585 gene was delineated within same region based on gene cluster analysis. As different miRNAs affiliated with the same gene cluster mostly display tightly synchronized expression and similar functions (7), we speculate that miR-585-5p might be important for regulating malignancy of GC. Nonetheless, the role of miR-585-5p in tumour pathogenesis is rarely defined. Recently, reports revealed that miR-585 is downregulated and acts as a tumour-suppressive miRNA within colon cancer, non-small-cell lung cancer, glioma and GC (8–11). Hu et al. (11) confirmed that miR-585 downregulation is linked with invasion/TNM stage/lymph node invasion levels and poor prognosis levels. However, the potential role and intrinsic mechanisms governing miR-585-5p function are poorly understood. Microphthalmia-associated transcription factor (MITF) was parsed out as a core-target based on target prediction of miR-585-5p. MITF, a melanocytic lineage-specific transcription factor, has been certified as a master regulator of melanocyte homeostasis and melanoma progression (12). However, the biology of MITF in GC requires further research. This investigation highlighted that miR-585-5p is downregulated within GC and is associated with poor prognosis, revealing miR-585-5p influence upon suppressing GC proliferative/metastatic properties. Furthermore, this investigation demonstrated that MITF had a straight positive regulation of GC proliferative/metastatic properties. CREB1 activates MITF transcription, and MAPK1 enhances the activity of MITF via phosphorylation at serine 73, boosting the cancer-promoting effect of MITF in GC. MiR-585-5p directly simultaneously inhibits expression of MITF, CREB1 and MAPK1 in a post-transcriptional manner. Consequently, miR-585-5p also indirectly restrains MITF transcription and activity through directly inhibiting the expression of CREB1 and MAPK1. Overall, we report for the first time that miR-585-5p suppresses GC proliferative/metastatic properties by orchestrating the interactions among CREB1, MAPK1 and MITF.
The GC tissue microarray for in situ hybridization (ISH) of miR-585-5p was purchased from Shanghai Outdo Biotech: HStmA180Su15 contains 16 cases of unpaired GC tissue-types and 82 cases of paired gastric adenocarcinoma and paraneoplastic tissue-types with one point for each tissue, all with long-term clinical follow-up records. The GC tissue microarray for immunohistochemistry of MITF was purchased from Avilabio: DC-Sto11020 contains 10 unpaired normal tissue-types and 45 paired GC and paraneoplastic tissue-types.
IHC and ISH staining were performed to quantify expression of the MITF protein and miR-585-5p, respectively. GC tissue microarrays were immuno-stained for MITF (Abcam, # ab270262). The digoxin-labelled nucleic acid probe for miR-585-5p was developed through GenePharma using the reverse complement of the following sequence: 25-CUAGCACACAGAUACGCCCAGA-46. The microarrays were observed/imaged through Pannoramic 250FLASH Scanner (3DHISTECH). IHC and ISH staining were concomitantly assessed through two blinded observers for individual clinical case clinico-pathological profiles. H-scoring was adopted based upon intensity and extent of staining by an experienced pathologist, and graded as: 0, negative staining; 1+, weak staining; 2+, moderate staining; and 3+, strong staining. The H-score was computed by multiplying the different intensities in 4 gradations with each percentage of positive tumour cells: H-score = 1× (% cells 1+) + 2× (% cells 2+) + 3× (% cells 3+). Finally, a score from 0 to 300 points was obtained (13). Median H-score within cohort was applied as a cut-off for distinguishing high- and low-expression.
Bioengineered novel recombinant miR-585-5p mimics, inhibitors and complementing negative controls were procured through Rongqingchang Biotech (China). Cultures including AGS, BGC823 and HGC27 lines at a confluence of 50% were exposed to transfection with miR-585-5p mimics or inhibitors employing Lipofectamine 3000® reagent (Invitrogen™, # L3000015) in line with kit protocols, with mimics-NC or inhibitors-NC, respectively, as controls. Ectopic expression efficiency tests and follow-up experiments were conducted at 48 h following transfection. Lentivirus expression plasmids for MITF-overexpression, MITF-mutant (S73A), shMITF (Target sequence of shMITF-3: GGTGAATCGGATCATCAAG), CREB1-overexpression, shCREB1 (Target sequence of shCREB1-3: acATTAGCCCAGGTATCTATG), MAPK1-overexpression and shMAPK1 (Target sequence: caAAGTTCGAGTAGCTATCAA) were constructed by GeneChem™ (Shanghai, China). Target cultures were exposed to 1×107 lentivirus transducing units within presence of HitransG P reagents (GeneChem, #REVG005, 1:25). Homologous empty lentiviral vectors acted as negative controls. The cultures were employed following infection and antibiotic selection for 4 weeks.
All animals in this investigation were purchased from the Experimental Animal Center of Fourth Military Medical University. All procedures were conducted in line with ARRIVE guidelines and accepted by the Ethics Committee of Fourth Military Medical University. Parental BGC823 cultures (5×105 cultures in 200 μl of PBS) were subcutaneously inoculated within ventral flank of 6-week-old male Balb/c nude murines (ten murines/cohort). Tumour diameter was quantified every two days, and following successful establishment of orthotopic xenograft tumorigenicity on the 13th day (the tumour diameters reached approximately 3-5 mm), 10 μg of miR-585-5p mimics or normal saline or the negative control was introduced within tumours every 48h for 2 weeks. Ten micrograms of mimics and 1.2 μl of in vivo jetPEI reagents (Polyplus Transfection, #PT-201-50G) were dissolved in 12.5 μl of 10% glucose, and sterile water was added to 25 μl. Ultimately, the two were incubated to form the internal delivery system. Twenty-nine days following subcutaneous tumour injection, all murines were sacrificed, and all tumours were removed, weighed, harvested and paraffin-embedded. Tumour volume (mm3) was determined depending upon longest/shortest diameters as:
A SimpleChIP® Enzymatic Chromatin IP Kit® (Cell Signaling Technology™, #9003) was employed for conducting ChIP assay. About 1×107 HGC27 cultures were cross-linked within 1% formaldehyde (Fuyu Fine, Tianjin), at room temperature for 15 minutes. The nuclear protein was extracted in 1× ChIP buffer within 1 M DDT (dithiothreitol) in line with kit instructions. Micrococcal Nuclease (Cell Signaling Technology, #10011) was used for digesting DNA into fragments of 150-900 bp. Following ultrasonication, the supernatants containing cross-linked chromatin were collected by centrifugation at 9,400×g and consequently placed into incubation with anti-CREB1 antibodies at 4°C with gentle rotation overnight. Equal amounts of Normal Rabbit IgG (Cell Signaling Technology™, #2729) and Histone H3 (D2B12) XP® Rabbit mAb (Cell Signaling Technology™, #4620) acted as the negative control and positive control, accordingly. ChIP-Grade Protein G Magnetic Beads® (Cell Signaling Technology™, #9006) were added the next day for 2-hour incubation with gentle rotation. Following the bead-antibody complexes were washed employing 1× ChIP buffer for 3 times on the DynaMag™-2 Magnetic Separation Rack (Invitrogen™, #12321D), the complexes were eluted and de-crosslinked with 5 mg/mL Proteinase K and 5 M NaCl and at 65°C for 2 h. The DNA was subsequently purified and subjected to PCR with the MITF promoter primers MITF ChIP Forward, AGAACTCCAGCCCTAACATC, and Reverse, TCTCATTTTGGTGTTTGGCC.
EMSA was employed for determining direct binding for CREB1 protein to the MITF DNA promotor in vitro employing LightShift™ Chemiluminescent EMSA Kit (Thermo Fisher, #20148). A prokaryotic vector for CREB1 expression was transformed into BL21 Escherichia coli and driven through isopropyl-beta;-d-thiogalactoside (IPTG). Bacterial pellets were lysed, and the recombinant CREB1 protein was purified through Ni-chelating affinity chromatography. Double-stranded DNA probes were synthesized employing the following sequences: MITF-wt-FF (5`6-FAM(FITC)-fluorescently labelled primers forward), TGGATGTCTTTTCTGATGTGAAATTAAA; MITF-wt-R (unlabelled primers reverse), TTTAATTTCACATCAGAAAAGACATCCA; MITF-mut-FF, TGGATGTCTTTCTCAGCATGAAATTAAA; MITF-mut-R, TTTAATTTCATGCTGAGAAAGACATCCA. The binding reactions, including 1 μl of ddH2O, 2 μl of binding buffer (5×), 6 μl of recombinant CREB1 protein and 1 μl of labelled-MITF-wt probes or labelled-MITF-mut probes, proceeded at 25°C for 20 min. The reaction products were added to 1 μl of EMSA/gel-shift loading buffer and segregated through SDS-PAGE. The gel was exposed and photographed employing an FLA-9000 apparatus (FujiFilm, Japan).
HGC27 cell pellets were lysed within IP lysis buffer (Beyotime, #P0013) harboring protease inhibitors (Boster, #AR1182-1) for 1 h with gentle rotation at 4°C. Supernatant lysates were collected by centrifuging at 12,000 rpm/min. Normal IgG antibody and protein A/G magnetic beads (Thermo Scientific™, #88802) acted for preclearing to lessen nonspecific binding, and then the lysates were incubated with the indicated antibodies. Equivalent protein lysates were premixed with an anti-flag antibody, an anti-HA antibody, or a normal rabbit IgG antibody as the negative control with gentle rotation at 4°C for 120 minutes and then placed into incubation with protein A/G magnetic beads (Thermo Scientific™, #88802) overnight. Resultant complexes were twice-washed with IP lysis buffer and three times with PBS, and the beads were resuspended in an equal volume of 2 × loading buffer and denatured at 100°C for 10 minutes on heat blocks, subsequently subjected to western blot analysis.
GraphPad Prism 8.0® software was employed in this case. All quantitative data-points reflected mean ± SEM. Statistical significance among multiple cohort analyses were identified through Analysis of variance (ANOVA) with Tukey’s post hoc test. Student’s t-test was employed for comparing mean variables of two cohorts. Kaplan-Meier method and log-rank t-test for significance were employed for survival analysis. Detailed information is listed within Appendix S1.
In order to quantify miR-585-5p expression-profile within GC, it was examined in a set of tissue microarray samples employing in situ hybridization. Compared with para-cancerous tissue-types, miR-585-5p expression in primary GC tissue-types was severely downregulated, and staining showed that miR-585-5p was mostly located within cytoplasm of adenocytes in gastric glands ( Figure 1A ). ISH quantitative analysis of 80 pairs of GC tissue-types and neighboring healthy tissue-types revealed a clear reduction in miR-585-5p levels in GC ( Figure 1B ). Based on quantitative analysis of miR-585-5p levels by H-scoring, 93 cases of GC (HStmA180Su15) were segregated within high miR-585-5p expression (n=47) and low miR-585-5p expression (n=46) cohorts, and correlation across miR-585-5p level and overall survival was analysed. Kaplan-Meier analyses showed that cases of positive miR-585-5p expression had prolonged overall survival ( Figure 1C ). Such dataset outcomes suggest that miR-585-5p might participate within GC carcinogenesis/development.
We assessed miR-585-5p expression-profile within multiple GC cell lines, with dataset outcomes demonstrating comparatively low levels of miR-585-5p in AGS and BGC823 cultures but obviously high expression in HGC27 cultures ( Supplementary Figure 1 ). Consequently, loss-of-function experiments to validate miR-585-5p were carried out employing HGC27 cultures and gain-of-function experiments in AGS and BGC823 cultures. MiR-585-5p mimics and inhibitors were utilized to overexpress and knockdown miR-585-5p in GC cultures. Following transient transfection for 48 h, the level of miR-585-5p was increased in AGS and BGC823 cultures and downregulated in HGC27 cultures ( Figures 2A, B ). Along with the enhancement of miR-585-5p, cell viability declined highly. Transwell assays showed that miR-585-5p upregulation remarkably decreased the migration and invasion of AGS and BGC823 cultures in vitro ( Figure 2A ). Conversely, inhibition of miR-585-5p dramatically increased cell proliferative, migrative and invasive effects in HGC27 cultures compared with negative controls ( Figure 2B ). Furthermore, in nude murines with subcutaneous inoculation of BGC823 cultures, tumour growth as well as both the volume and weight of orthotopic xenograft tumours were highly reduced with miR-585-5p intra-tumour therapy, in contrast to normal saline or negative controls ( Figures 2C, D ). In summary, miR-585-5p plays a pivotal part in regulating the proliferative/metastatic properties of GC at multiple levels.
The intrinsic mechanism by which miR-585-5p acts in GC development remains to be elucidated. One of the most important modes of miRNA functioning is miRNA-mediated post-transcriptional mRNA transcript repression via binding to 3’UTRs (14). To uncover the exact antitumour mechanism of miR-585-5p in GC, miRWalk 3.0 (http://mirwalk.umm.uni-heidelberg.de/) was employed for predicting target-genes. MITF-3’UTR was found to contain miR-585-5p-binding sites, indicating MITF could act as a target-gene for miR-585-5p. Interestingly, we carefully examined all predicted target-genes and found that among them, two have been reported to be associated with MITF: CREB1 and MAPK1. CREB1 has been widely confirmed to bind to a site of the MITF promotor, upregulating its expression (15). Phosphorylation of MITF by MAPK1 at serine 73 enhances its transcriptional activity (16). To further ascertain promising binding sites, we employed alignments for miR-585-5p seed sequence and 3’ UTR sequences of MITF, CREB1 and MAPK1, combined with another methodology of free energy calculation (Detailed predicted information is provided in Appendix S2). The above analysis implies that MITF might be the core target by which miR-585-5p works in GC progression. Considering that the prediction of CREB1 and MAPK1 suggest the probability of MITF regulation, we sought to determine the role of MITF. Based on IHC staining, we detected MITF upregulation in GC tissue-types compared to para-cancerous tissue-types ( Figure 3A ). Validating MITF function within GC, further gain-of-function experiments indicated that MITF overexpression increases the proliferative, migrative and invasive effects of GC cultures but that downregulation has opposing effects ( Figures 3B, C ). To assess whether miR-585-5p directly inhibits MITF translation, we constructed 3’UTR reporter genes for MITF-containing miR-585-5p binding sites as reporter-genes for binding-site mutations ( Figure 4A ) and co-exposed to transfection miR-585-5p mimics into 293T and BGC823 cultures. Luciferase activity outcomes demonstrated ectopic overexpression of miR-585-5p impaired activity of the MITF-3’UTR-wt reporter, though there was no notable activity shifts within mutated reporter ( Figure 4B ). Furthermore, RT-qPCR and western blot analyses were carried out employing GC cultures to explore the regulatory MITF function by miR-585-5p. Upregulation of miR-585-5p led to transcriptomic/proteomic MITF downregulation ( Figure 4C ). Decreased levels of miR-585-5p led to dramatic proteomic MITF upregulation; however, no significant change in MITF mRNA was found within miR-585-suppressed HGC27 cultures ( Figure 4D ). Given the crucial role of MITF in GC development, we wondered whether MITF contributes to the miR-585-mediated GC phenotype. To clarify whether miR-585-5p regulates GC cell growth and metastasis inhibition through MITF, we adopted an MITF antagonism-of-function strategy employing the MITF expression vector without miR-585-binding sites in miR-585-5p-co-expressing AGS and BGC823 cultures and found that MITF overexpression reverses the in vitro inhibitory effect of miR-585-5p on proliferative/metastatic properties in GC ( Figure 4E ).
In our previous study, bioinformatic analysis demonstrated CREB1 to act as an underlying target for miR-585-5p. Multiple studies have confirmed that CREB1 directly activates transcription of the MITF gene (17). Consequently, we evaluated whether CREB1 is an imperative intermediate bridging miR-585-5p and MITF. To explore presumed miRNA-mRNA interactions between miR-585-5p and CREB1, the entire CREB1 3’UTR harboring potential miR-585 binding sites and the corresponding mutant CREB1 3’UTR were fused to a reporter vector downstream of the firefly luciferase gene ( Figure 5A ). The resulting plasmid was exposed to transfection into 293T and BGC823 cultures along with miR-585-5p mimics and a transfection negative control. As expected, miR-585-5p overexpression highly interfered with the luciferase activity of the CREB1-wt reporter, and the inhibitory effect was antagonized by transfection of the CREB1-mut reporter ( Figure 5B ). Moreover, cellular CREB1 levels were robustly reduced by introduction of miR-585-5p into AGS and BGC823 cultures ( Figure 5C ). Conversely, the use of miR-585-5p inhibitors in HGC27 cultures promoted proteomic CREB1 augmentation though not at transcriptomic level ( Figure 5D ). Taken together, such results show that CREB1 is directly targeted by miR-585-5p. To determine whether the suppressive effects on GC phenotypes exerted by miR-585-5p are related to CREB1, we first specified the role of CREB1 in GC. CREB1 overexpression lentivirus was used to infect HGC27 cultures, and exogenous CREB1 overexpression potently stimulated cell proliferative, migrative and invasive effects in vitro ( Supplementary Figure 2A ). In contrast, CREB1 knockdown in AGS and BGC823 cultures employing shRNA lentiviruses exhibited opposite effects on GC phenotypes ( Supplementary Figure 2B ). In order to increase clarity on the functional links across miR-585-5p/CREB1, we applied a CREB1 gain-of-function approach in miR-585-expressing AGS and BGC823 cultures, and found that CREB1 (Δ3’UTR) upregulation partly rescued inhibitory effects of miR-585-5p on GC cell proliferative/metastatic properties ( Figure 5E ). Collectively, Such dataset outcomes suggest that miR-585-5p might regulate proliferative/metastatic properties in a CREB1-dependent manner. As an important transcription factor, CREB1 binds to the cAMP response element (CRE) consensus motif situated across -140 and -147 bp from the transcription site of the MITF promoter to enhance its expression (15). Multiple studies have affirmed the positive regulation of MITF transcription by CREB1 in malignant melanoma (18–20). However, whether CREB1 directly activates MITF transcription in GC is still unclear. Therefore, we predicted the interactive mode of the CREB1 protein and MITF promotor region employing the JASPAR database (http://jaspar.genereg.net/), and the highest scoring CREB1 binding site, i.e., TCTGATG (-1357 to -1351), was selected via bioinformatics analysis ( Figure 6A ). To confirm the hypothesis that MITF is a key target-gene of CREB1 in GC, ChIP-PCR was performed, and nucleic acid electrophoresis analysis showed that the CREB1 protein bound directly to the MITF promotor at the indicated sites in HGC27 cultures ( Figure 6B ). Subsequently, a probe for this binding site was designed and synthesized, and EMSA was carried out. Such dataset outcomes demonstrated proteomic CREB1 strongly binds to the MITF-wt probe; however, the interaction was highly but not completely hindered, which further confirms that CREB1 might bind to this site but that it is not the only binding site ( Figure 6C ). Additionally, the CREB1 expression vector was co-exposed to transfection with the MITF full-length promoter reporter gene or the binding-site mutant reporter gene into 293T cultures. CREB1 highly upregulated the transcriptional activity of the MITF wild-type promoter; the regulatory effect was inhibited but still existed within presence of the -1357 to -1351 binding site mutation, suggesting that CREB1 regulates MITF transcriptional expression through this binding site but not the only binding site ( Figure 6D ). A CREB1 overexpression lentivirus was used to infect HGC27 cultures, and shCREB1 lentivirus was applied to knockdown CREB1 in AGS and BGC823 cultures. Upregulated CREB1 obviously upregulated transcriptomic/proteomic MITF expression ( Figure 6E ), whereas downregulating CREB1 resulted in dramatically decreased MITF expression ( Figure 6F ). Based on the finding that CREB1 accelerates proliferative/metastatic properties in GC cultures, this investigation probed the possibility whether CREB1 functions in an MITF-dependent manner. Hence, we infected CREB1-overexpressing HGC27 cultures with shMITF lentivirus or shNC lentivirus and carried out CCK-8 and Transwell assays. Overall, knockdown of MITF diminished the CREB1-mediated promotion of cell growth and metastasis in GC ( Figure 6G ). In addition, upregulation of MITF rescued the dampened tumour proliferative/metastatic properties driven by shRNA-mediated silencing of CREB1 ( Supplementary Figure 3 ). These findings suggest that CREB1-induced MITF overexpression promotes GC proliferative/metastatic properties. In summary, such dataset outcomes verify that miR-585-5p inhibits MITF transcription through direct targeting of CREB1.
Based on the previous prediction of MAPK1 as a target-gene of miR-585-5p, this investigation examined if miR-585-5p is able to regulate MAPK1 in GC. The luciferase reporter system validated that miR-585-5p could be directly-bound onto MAPK1-3’UTR at the indicated sites ( Figures 7A, B ). Moreover, the transcriptomic/proteomic MAPK1 downregulation occurred by ectopic expression of miR-585-5p in AGS and BGC823 cultures ( Figure 7C ). Consistently, miR-585-5p inhibitors led to elevated expression of MAPK1 ( Figure 7D ). It is widely recognized that MAPK1 plays has pivotal parts within development and progression of various cancers, including GC (21–23). However, whether MAPK1 is implicated within miR-585-mediated tumour proliferative/metastatic properties remains uncertain. Resistance-of-function experiments were conducted via overexpression of MAPK1 with or without miR-585-5p-binding sites, showing that MAPK1 reverses the anti-proliferative, anti-migrative and anti-invasive capacities of miR-585-5p in AGS and BGC823 GC cultures ( Figure 7E ). Overall, dataset outcomes suggest MAPK1 is a functional target of miR-585-5p within GC malignant phenotype. The transcriptional and MITF functional activity is founded upon post-translational modifications and the availability of cooperating partners. MAPK1 was confirmed to phosphorylate the MITF protein at serine 73, enhancing its activity (16, 24). However, whether this effect exists in GC is indeterminate. To assess the interplay between MAPK1 and MITF, we overexpressed HA-tagged MAPK1 and flag-tagged MITF mutants (S73A) or wild-type MITF at similar levels in HGC27 cultures and used immunoprecipitation to evaluate the interaction. The MAPK1-MITF interaction was confirmed by blotting complexes containing endogenous MITF precipitated with flag-specific antibodies and probing for MAPK1. Conversely, an anti-HA antibody was used for precipitation, and the isolated complexes were blotted with an anti-MITF antibody. The interaction was partly abolished when the S73A mutant protein was used ( Figure 8A ), and an endogenous interaction assay studying GC cultures demonstrated that MAPK1 interacts with MITF and that serine 73 is a crucial site. Furthermore, immunofluorescence was employed to determine the cellular distributions of the two proteins, and the superimposition of green fluorescence indicating MAPK1 over red fluorescence indicating MITF validated the association between MAPK1 and MITF ( Figure 8B ). Interestingly, we found that MAPK1 overexpression or silencing had no significant influence on MITF levels at either the mRNA or protein level ( Supplementary Figures 4A, B ), suggesting that MAPK1 might only be implicated within phosphorylation of MITF but not in its degradation. To assess MAPK1 phosphorylation of the MITF protein at serine 73, we conducted combinatorial expression of NC(MAPK1) + MITF, MAPK1 + MITF and MAPK1 + MITF(S73A) in HGC27 cultures. Phosphorylated MITF was analysed by employing an anti-phosphorylation (serine) antibody to detect the immunoprecipitated MITF protein, showing that overexpressing MAPK1 strongly increased global MITF phosphorylation in GC cultures but that overexpressing the S73A mutant did not. In contrast, MAPK1 silencing attenuated the phosphorylation level of MITF, and S73A mutants sustained low-level phosphorylation, suggesting that serine 73 is a principle site for phosphorylation ( Figure 8C ). Collectively, the above results reveal that MAPK1 binds to MITF and phosphorylates serine 73 in GC cultures. In order to increase clarity regarding MAPK1 function in MITF serine phosphorylation, HGC27 cultures were engineered to express elevated levels of MAPK1 and MITF or mutant MITF (S73A) alone or in combination. We also overexpressed wild-type MITF or the S73A mutant in HGC27 cultures with downregulated MAPK1 to comprehensively assess the functional effect of MAPK1-mediated phosphorylation on MITF serine73. CCK-8 and Transwell assays showed that compared with NC + NC cultures, MAPK1 or MITF alone enhanced proliferative/metastatic properties, and the MAPK1 + MITF co-expressing cultures exhibiting the most strongly enhanced malignant phenotypes. However, MAPK1+MITF (S73A)-overexpressing cultures presented highly dampened proliferative and metastatic capacities relative to MAPK1 + MITF co-expressing cultures, which indicated that MAPK1 overexpression notably enhances the cancer-promoting characteristics of MITF but not of the MITF (S73A) mutant ( Figure 8D ). Taken together, our findings indicate that MAPK1 enhances the cancer-promoting characteristics of the MITF protein via serine 73 phosphorylation. In summary, results indicated miR-585-5p suppresses MITF activity by directly targeting MAPK1.
This investigation revealed miR-585-5p is markedly downregulated in GC tissue-types and that cases of positive miR-585-5p expression have better clinical outcomes than cases of negative miR-585-5p expression. Furthermore, we identified for the first time the melanocyte master regulator MITF as promoting carcinogenesis in GC and acting as a direct and essential mediator of miR-585-5p-impeded malignant phenotypes. Such dataset outcomes demonstrated that overexpression of miR-585-5p highly inhibits proliferative/metastatic properties of GC by directly targeting MITF, CREB1 and MAPK1. Moreover, CREB1 directly regulates MITF transcription, and MAPK1 directly phosphorylates MITF at serine73. This interaction network shows that miR-585-5p directly or indirectly regulates MITF expression and activity by simultaneously repressing target molecules, accounting for the anti-tumour role of miR-585-5p in inhibiting GC growth and metastasis. In summary, our study supports that miR-585-5p suppressed GC proliferative/metastatic properties through CREB1/MAPK1/MITF pathway ( Figure 9 ). MiRNAs, an abundant class of endogenous noncoding RNAs, have been confirmed to be ectopically expressed in malignant gastric tissue-types. Dysregulated miRNAs are important in regulating proliferative/metastatic properties of GC (25, 26). The gene encoding miR-585-5p is located at 5q35.1 within intron of SLIT3, together with miR-218-5p (6). Notably, miR-218-5p was previously identified as a tumour suppressor in GC by our research cohort (4, 5), and we tentatively proposed that miR-585-5p has a similar expression pattern and phenotypic effects as miR-218-5p in GC, considering that the two miRNAs belong to the same gene cluster. MiR-585 has been previously characterized to be a tumour suppressor in non-small-cell lung cancer (9), triple-negative breast cancer (27), cervical cancer (28), glioma (10), tongue squamous cell carcinoma (29), and colorectal cancer (30), among others. However, few studies have determined the specific role of miR-585 in GC. Some reports have identified that miR-585 exhibits relatively low expression in GC tissue-types, predicting poor prognosis (11, 31), but the molecular mechanisms implicated are far from clear. Several studies report that miR-218 directly targets MITF to exert its biological functions (32–34), and we accordingly screened for potential targets of miR-218-5p and miR-585-5p; MITF was co-pinpointed by bioinformatic prediction algorithms. However, inconsistent with existing studies, our findings showed that miR-218-5p overexpression had no impact on MITF expression ( Supplementary Figure 5 ) but that miR-585-5p was affirmed to directly target and regulate MITF in a post-transcriptional manner. Herein, we focus on miR-585-5p implications and its downstream molecular mechanisms in GC. We not only reconfirmed decreased expression of miR-585 in GC but, more importantly, highlighted that miR-585-5p hinders GC proliferative/metastatic properties by directly or indirectly adjusting MITF expression and biological activities at different levels of gene expression regulation. In addition, we identified that miR-585-5p simultaneously targets CREB1 and MAPK1 mRNAs, downregulating transcriptomic/proteomic expression. Functionally, miR-585-5p overexpression inhibits GC cell proliferative, invasive, and migrative properties by targeting CREB1 and MAPK1 in resistance-of-function analyses. In agreement with our results, Hu et al. (11) reported that miR-585 is downregulated in both GC tissue-types and cultures and that ectopic overexpression highly suppresses the malignant phenotype of GC by directly targeting MAPK1, with no evidence that miR-585-5p directly regulates CREB1. Additionally, miR-585 binds to the 3’UTR of F-box protein 11 (FBOX11), and overexpression of miR-585 inhibits the GC cell proliferation and migration driven through FBXO11 (31). Taken together, our findings indicate that miR-585-5p expression is often decreased in GC, which is of great value in GC treatment via suppression of tumour growth and metastasis. Such dataset outcomes illustrate for the first time that the anti-tumour effect of miR-585-5p in GC is due to direct simultaneous inhibition of MITF, CREB1 and MAPK1 expression. MITF, widely identified as one of the most classic and pivotal regulators in malignant melanoma, has been associated with phenotypic switching between predominantly invasive and proliferative behaviours of melanoma. Notably, MITF has been reported as a tumour suppressor in inhibiting tumour growth and metastasis in GC (35, 36). Several bodies of proof partly support that MITF plays an oncogenic function within GC development. For instance, miR-876-5p suppresses GC cell viability/migrative properties though induces apoptosis by targeting MITF (36). Li et al. (35) inferred that CSE1L silencing promotes apoptosis and inhibits tumour growth and metastasis by decreasing MITF expression. Nevertheless, whether MITF is aberrantly expressed and has a pathogenic function within GC, remains largely elusive. This investigation detected MITF expression in GC tissue-types for the first time, revealing that MITF was upregulated in GC tissue-types in comparison with peri-carcinomatous tissue-types. Within subsequent gain- and loss-of-function analyses, we observed that MITF overexpression conspicuously promotes HGC27 cell proliferation, invasion and migration but that MITF downregulation highly suppresses the malignant phenotypes of AGS and BGC823 cultures. Hence, in contrast to the controversial context-dependent regulation of MITF within tumorigenesis and progression of melanoma, our findings show that MITF acts straightforwardly as an oncogene in GC. In terms of the underlying mechanisms driving de novo MITF expression in GC, MITF was proven to be a direct target of miR-585-5p. Such dataset outcomes suggest that the high ectopic expression of the MITF protein in GC tissue-types might be due, at least in part, to inhibitory regulation by miR-585-5p. Collectively, we report that MITF is not only a potent marker predicting prognosis but as a direct downstream target of miR-585-5p, also a pro-proliferative and pro-metastatic gene in GC. Although the direct post-transcriptional miRNA-mRNA regulatory interaction matters greatly in regulating MITF protein expression, many transcription factors govern MITF transcriptional expression (17). As previously illustrated, we have shown that CREB1 and MITF are both the downstream targets of miR-585-5p, and it has been widely demonstrated that CREB1, a canonical bZIP transcription factor, identifies the CRE motif ‘-TGACGTCA-’ within MITF promotor (15), thus enabling varied cAMP levels to influence MITF expression (37, 38). This idea is further perpetuated by various studies showing that CREB1 directly upregulates MITF expression in human melanocytes or melanoma cultures (39–41). Nevertheless, there is a paucity of evidence regarding whether CREB1 directly activates MITF transcription in GC. This investigation is the first to confirm in GC cultures that CREB1 directly binds to the newly identified MITF promotor sites -1357 to -1351 bp, positively regulating MITF transcription. As expected, MITF mRNA and protein levels increase in response to CREB1 overexpression, whereas CREB1 knockdown results in highly decreased MITF expression. Further resistance of function showed that the deficiency of MITF in CREB1-overexpressing HGC27 cultures counteracted the pro-proliferation and pro-metastasis effects due to CREB1, underscoring that CREB1-dependent MITF upregulation is crucial to stimulate GC development. Interestingly, our results showed that the -TCTGATG- (-1357 to -1351) site might not be the only binding site for CREB1 within MITF promotor, and there is a possibility that other sites also contribute to CREB1 binding. Considering the widely confirmed CREB1 binding site TGACGTCA (-140 to -147 bp) within MITF promotor, we speculate that the two candidate sequences might both act immensely within transcriptional regulation of MITF. However, this should be further explored in GC cultures. Together with our data from ChIP, EMSA and luciferase reporter assays, we determined that CREB1 positively regulates MITF transcription in GC. In addition to transcriptional and post-transcriptional regulation of MITF expression, regulation of MITF activity contributes to its function in tumours. Phosphorylation is a prerequisite modification for modulating MITF activity, and the MITF protein has been reported to be phosphorylated by MAPK1 at Ser73, by P90RSK at Ser409, by GSK3 at Ser69, Ser298, Ser397, Ser401, and Ser405, and by P38MAPK at Ser307. Hemesath et al. (42) first reported that MAPK1-mediated MITF phosphorylation at Ser73 enhances MITF-dependent transactivation in melanocytes, and Ser73 phosphorylation by MAPK1 is proposed to be required for recruitment of the P300/CBP transcriptional coactivator within transactivation domain of MITF (16). Phosphorylation at MITF Ser73 is predominantly responsible for MITF activation to promote malignant phenotypes in melanoma (24, 43); however, this interaction has not been reported in GC. Hence, we conducted IP and double-labelling immunofluorescence of HGC27 cultures, showing endogenous interaction between MAPK1 and MITF protein in GC, and the effect of MAPK1 phosphorylation on MITF was determined employing anti-phosphoserine antibodies. Consistent with previous studies, the S73A mutant counteracted the above effects, indicating that the serine 73 site is required for MAPK1-mediated phosphorylation of the MITF protein. The results of subsequent functional assays demonstrated that phosphorylation of serine 73 is important to facilitate MITF-enhanced GC proliferative/metastatic properties. Nonetheless, our results show that MAPK1 does not affect MITF protein levels, which is incompatible with the proposal that S73-phosphorylated MITF is a prerequisite for MITF degradation. As an E2 SUMO conjugating enzyme, UBC9 was found to be responsible for the degradation of MITF in response to S73 phosphorylation (44). Moreover, Azam et al. (45) reported that sargaquinoic acid increases phosphorylation of MAPK1 and MITF (Serine73), ultimately inducing proteasomal degradation of MITF, in melanoma cultures. Controversially, both Hemesath et al. (42) and Wellbrock et al. (46) found that the S73A mutation has no effect on MITF protein stability, in line with our results. Considering the tumour heterogeneity between melanoma and GC, together with our results, it is reasonable to speculate that MAPK1 promotes GC proliferative/metastatic properties via phosphorylation of MITF (Serine73), enhancing its activity instead of stability. Additionally, accumulating lines of evidence have shown that kinase-regulated CREB1 phosphorylation activates CREB1-dependent transcription and acts as an essential cascade in oncobiology (47). Leduc et al. reported that ERK1 instead of ERK2 was necessarily required for CREB1 phosphorylation and activation in mouse pancreatic beta cells (48). Chen et al. illustrated that MAPK1 could activate CREB1 to bind to the promotor of miR-212-3p. But no specific evidence have certificated the direct interaction between CREB1 and MAPK1. Our study demonstrated that CREB1 and MAPK1 act through different approaches to promote MITF-mediated GC progression, with one enhancing MITF transcriptional expression and the other promoting phosphorylation of MITF proteins. However, our study did not address the possible regulation of CREB1 by MAPK1 in GC. Therefore, detailed analysis regarding the relationship between MAPK1 and CREB1 is needed for further investigation in GC. Besides, although the direct regulation of miR-585-5p targeting on MITF, CREB1 and MAPK1 were confirmed, miR-585-5p inhibitor did not result in significant change within mRNA levels. The results reflected that the inhibition of miR-585-5p might only make a difference to translational process instead of the stability or degradation of mRNAs of these 3 targets. And further exploration is desired for the seemingly contradictory results. Overall, this investigation not only revealed the tumour-inhibiting role of miR-585-5p in GC proliferative/metastatic properties but also defined the mechanism of miR-585-5p downstream signalling through MITF regulation in multiple aspects of gene expression. We identified another two targets of miR-585-5p that regulate MITF: CREB1 activates MITF transcription upregulating MITF expression, and MAPK1 phosphorylates MITF (Ser73) activating MITF pro-cancerous activity. Ultimately, miR-585-5p suppresses GC development by simultaneously targeting MITF in both direct post-transcriptional and indirect transcriptional (CREB1/MITF) or post-translational (MAPK1/MITF) manners. Accordingly, miR-585-5p/MITF-based targeted therapy might be a promising strategy for cases of GC. In conclusion, this study uncovered miR-585-5p suppressed GC proliferative/metastatic properties by orchestrating the interactions among CREB1, MAPK1 and MITF. And miR-585-5p/MITF-based targeted therapy might represent a potential therapeutic strategy.
The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors.
All procedures involving animals were carried out in accordance with the principles of ARRIVE and approved by the Ethics Committee of Air Force Medical University.
YW, WW and JT designed research. YW, ML, JZ and YY performed the experiments. YY, ML, and JZ analyzed statistics. YW, JZ and ZL purchased reagents. YW and JT wrote the article. All authors read and approved the final manuscript. All authors contributed to the article and approved the submitted version.
This work was supported by the National Natural Science Foundation of China [81773071, 81972226] and Key R&D Program of Shaanxi Province [2020ZDLSF01-05].
The authors would like to thank Hanbio Tech (shanghai, China) for assisting us in bioinformatic analysis. The diagram of “MITF”, “CREB1” and “MAPK1” in Figure 9 were modified from Servier Medical Art (https://smart.servier.com) under CC BY 3.0 license (https://creativecommons.org/licenses/by/3.0/).
The authors declare that the research was conducted within 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 |
|
PMC9577252 | Shaochuan Shi,Shiya Zhang,Jie Wu,Xintong Liu,Zhao Zhang | Identification of long non-coding RNAs involved in floral scent of Rosa hybrida | 04-10-2022 | Rosa,floral scent,lncRNA,terpenoids,phenylpropanoids,benzenoids,fatty acid derivatives | Long non-coding RNAs (lncRNAs) were found to play important roles in transcriptional, post-transcriptional, and epigenetic gene regulation in various biological processes. However, lncRNAs and their regulatory roles remain poorly studied in horticultural plants. Rose is economically important not only for their wide use as garden and cut flowers but also as important sources of natural fragrance for perfume and cosmetics industry, but presently little was known about the regulatory mechanism of the floral scent production. In this paper, a RNA-Seq analysis with strand-specific libraries, was performed to rose flowers in different flowering stages. The scented variety ‘Tianmidemeng’ (Rosa hybrida) was used as plant material. A total of 13,957 lncRNAs were identified by mining the RNA-Seq data, including 10,887 annotated lncRNAs and 3070 novel lncRNAs. Among them, 10,075 lncRNAs were predicted to possess a total of 29,622 target genes, including 54 synthase genes and 24 transcription factors related to floral scent synthesis. 425 lncRNAs were differentially expressed during the flowering process, among which 19 were differentially expressed among all the three flowering stages. Using weighted correlation network analysis (WGCNA), we correlate the differentially-expressed lncRNAs to synthesis of individual floral scent compounds. Furthermore, regulatory function of one of candidate lncRNAs for floral scent synthesis was verified using VIGS method in the rose. In this study, we were able to show that lncRNAs may play important roles in floral scent production in the rose. This study also improves our understanding of how plants regulate their secondary metabolism by lncRNAs. | Identification of long non-coding RNAs involved in floral scent of Rosa hybrida
Long non-coding RNAs (lncRNAs) were found to play important roles in transcriptional, post-transcriptional, and epigenetic gene regulation in various biological processes. However, lncRNAs and their regulatory roles remain poorly studied in horticultural plants. Rose is economically important not only for their wide use as garden and cut flowers but also as important sources of natural fragrance for perfume and cosmetics industry, but presently little was known about the regulatory mechanism of the floral scent production. In this paper, a RNA-Seq analysis with strand-specific libraries, was performed to rose flowers in different flowering stages. The scented variety ‘Tianmidemeng’ (Rosa hybrida) was used as plant material. A total of 13,957 lncRNAs were identified by mining the RNA-Seq data, including 10,887 annotated lncRNAs and 3070 novel lncRNAs. Among them, 10,075 lncRNAs were predicted to possess a total of 29,622 target genes, including 54 synthase genes and 24 transcription factors related to floral scent synthesis. 425 lncRNAs were differentially expressed during the flowering process, among which 19 were differentially expressed among all the three flowering stages. Using weighted correlation network analysis (WGCNA), we correlate the differentially-expressed lncRNAs to synthesis of individual floral scent compounds. Furthermore, regulatory function of one of candidate lncRNAs for floral scent synthesis was verified using VIGS method in the rose. In this study, we were able to show that lncRNAs may play important roles in floral scent production in the rose. This study also improves our understanding of how plants regulate their secondary metabolism by lncRNAs.
Approximately 90% of the eukaryote genome is transcribed (Wilhelm et al., 2008), but only 1-2% of the genome has a protein-coding capacity (Consortium, 2007), and the majority of the genome is transcribed as non-coding RNAs (ncRNAs). Small ncRNAs with length of less than 200 bp, such as microRNAs (miRNAs), small interfering RNAs (siRNAs), piwi-interacting RNAs (piRNAs), transacting siRNAs (ta-siRNAs), and natural antisense transcript siRNAs (NAT-siRNAs), have received considerable attention in the last decade for their essential roles in post-transcriptional and transcriptional regulation in eukaryotes (Simon and Meyers, 2011; Samad et al., 2017; Sun et al., 2019). Among them, the most well-known miRNAs are a class of RNAs with lengths of 20–24 bp that are highly conserved throughout evolution and regulate the growth and development of organisms by cleaving and degrading target gene transcripts or inhibiting translation through complementary pairing with the bases of target sites. In contrast, lncRNAs are typically larger than 200 bp but poorly conserved; they interact with large molecules, such as DNA, RNA, and proteins, and regulate protein modification, chromatin remodeling, protein functional activity, and RNA metabolism in vivo through cis- or trans-activation at the transcriptional, post-transcriptional, and epigenetic levels (Chekanova, 2015). In the past decade, thousands of lncRNAs have been identified in plants, including Arabidopsis thaliana (Di et al., 2014; Wang et al., 2014; Zhu et al., 2014; Moison et al., 2021; Liu et al., 2022), Medicago truncatula (Wen et al., 2007), Triticum aestivum (Xin et al., 2011; Zhang et al., 2016; Lu et al., 2020), Oryza sativa (Shin et al., 2018; Yu et al., 2020; Zhang et al., 2020; Chen et al., 2021), Zea mays (Boerner and McGinnis, 2012; Li et al., 2014), Manihot esculenta Crantz (Li et al., 2022), Solanum lycopersicum (Zhu et al., 2015; Jiang et al., 2019), Cuscuta spp. (Wu et al., 2022), Populus trichocarpa (Shuai et al., 2014), P. tomentosa (Chen et al., 2015; Chen et al., 2022), and P.×euramericana (Wang et al., 2017). Although the regulatory mechanisms of lncRNAs have been elucidated widely, they are mostly derived from animals, and only a few lncRNA mechanisms in plants have been revealed, resulting in a lack of systematic and consensus lncRNA regulatory mechanisms in the plant (Wu et al., 2020). Two novel intergenic lncRNAs in tomato, lncRNA1459 and lncRNA1840, play a regulatory role in tomato fruit ripening (Zhu et al., 2015), while a lncRNA in rice, referred to as long-day-specific male-fertility-associated RNA (LDMAR), regulates photoperiod-sensitive male sterility (Ding et al., 2012). In rice and maize, there is an association of some lncRNAs and their polymorphisms with agricultural traits (Wang et al., 2015). Two lncRNAs—COOLAIR (cool-assisted intronic non-coding RNA) and COLDAIR (cold-assisted intronic non-coding RNA)—are found to regulate vernalization by negatively regulating a MADS-box transcription factor FLC that represses flowering in Arabidopsis (Heo and Sung, 2011; Sun et al., 2013). Some lncRNAs were found to be endogenous target mimics (eTMs) of miRNAs, indicating a new mechanism for regulating miRNA activity (Wu et al., 2013; Jiang et al., 2019). In Arabidopsis thaliana, lncRNA SABC1 recruited the polycomb repressive complex 2 to its neighboring gene NAC3 to decrease its transcription via H3K27me3 (Liu et al., 2022). This evidence, highlighting the essential and varied functions of lncRNAs, demonstrates the importance of discovering and identifying lncRNAs in different biological processes and the need to elucidate their functional mechanisms. Floral scent primarily attracts pollinators to angiosperms to facilitate in fertilization (Dudareva et al., 2004), but also functions in plant defense (Caruso and Parachnowitsch, 2016), brings mental pleasure to humans, and provides essential oils and flavors for the food and perfume industries (Grammer et al., 2003). Increasing numbers of flower volatile biosynthesis genes have been cloned but the complete regulatory mechanism(s) has yet to be elucidated (Muhlemann et al., 2014), and lncRNAs in floral scent synthesis remain predominantly unknown. Consequently, the identification and characterization of novel lncRNAs is crucial to understand the function of lncRNAs in floral scent. Rose is one of the most commonly cultivated ornamental plants in the world, popular in gardens and as cut flowers, but are also important sources of essential oils for perfumes and cosmetics due to their floral scent (Magnard et al., 2015). However, in the process of rose breeding over hundreds of years, the focus on cut flowers and visual attributes has disadvantaged scent traits (Vainstein et al., 2001). Rose probably manufactures the most diverse scent compounds based on the emission of hundreds of volatile molecules. Any variation in the composition of the volatile molecules, in both quality and quantity, could lead to different rose scent profiles (Joichi et al., 2005; Bendahmane et al., 2013). Three major scent molecule classes were involved in roses: the terpenes, including rose oxide, geraniol, linalool, citronellol, nerol and so on; the benzenoids/phenylpropanoids, such as 2-phenylethanol (2-PE), 2-phenylethyl acetate, 3,5-dimethoxytoluene (DMT), 1,3,5-trimethoxybenzene (TMB) methyleugenol, methylisoeugenol and so on; the fatty acid derivatives, including cis-3-hexenyl-1-alcohol, 2-hexenyl acetate, cis-3-hexenyl acetate (Schnepp and Dudareva, 2008). Regulatory factors were revealed to promote the production of floral scent compounds more efficiently compared with synthetase genes, indicating a powerful tool to modify the floral scent trait (Zvi et al., 2012). However, although dozens of genes in synthesis pathway of rose floral scent have been identified and functionally validated in the past decade, there is limited information available on the regulatory mechanism in the rose (Shi and Zhang, 2022). Only one transcription factor (TF), RhMYB1 was found to probably play a role in rose floral scent production, but its function has not been validated (Yan et al., 2011). In rose petals, the miR156-SPL9 regulatory hub is proposed to orchestrate the production of both colored anthocyanins and certain terpenes, by permitting the complexation of preexisting MYB-bHLH-WD40 proteins (Raymond et al., 2018). This study used the rose cultivar ‘Tianmidemeng’ with a heavy floral scent to identify and analyze lncRNAs through strand-specific RNA-seq of petal samples from three flowering stages of the rose. Based on genome location and differential expressions of the lncRNAs, together with weighted gene co-expression network analysis (WGCNA), lncRNAs related to floral scent were identified. A total of 13,957 putative lncRNAs were discovered. Rose lncRNAs are shorter and harbor fewer exons and less coding potential compared with the protein-coding genes. Hundreds of lncRNAs showed significantly differential expression among the three flowering stages of the rose, and target prediction for lncRNAs coupled with WGCNA supported the role of these lncRNAs in floral scent production. Moreover, WGCNA further correlated the differentially expressed lncRNAs to individual floral scent compounds. Findings from the study suggest that lncRNAs are instrumental in the regulation of floral scent production and provide new insights into the study of floral scent.
The plant material ‘Tianmidemeng’ (R. hybrida) was planted in the natural environment of the campus of China Agricultural University in Haidian district, Beijing. Based on the open state of the flower, we divided the flower development into three stages: 1) early-flowering (EF), of which the sepals are slightly unfolded while the petals are still closed, the petals are becoming red and have little fragrance; 2) semi-flowering (SF), of which the outer 2-3 layers of petals are unfolded while the inner part are still closed, the petal color is rose-red, and the fragrance is rich; 3) late-flowering (LF), of which the petals are all unfolded but begin to wilt, the petal color begins to fade, and some fragrance still remains ( Figure 1A–C ). All fresh petal samples from development stages were collected at 9:00 a.m. The flower materials collected for each sample were divided in half: one part was used for gas chromatography-mass spectrometry (GC-MS) analysis, and the other part was for RNA-seq analysis after immediately freezing in liquid nitrogen. For every sample, three replicates were prepared.
For each sample, 3 g petals were quickly placed into a 100-mL sample vial, and 10 μL ethyl caprate (0.865 μg·μL-1; Sigma Ltd. Co., New York, USA) was subsequently added as the internal standard. The vial was then sealed rapidly with a rubber septa. For extracting and concentrating the floral volatiles in the vial, a solid-phase microextraction (SPME) manual headspace sampler was used with a 100-μm polydimethylsiloxane (PDMS) fiber embedded in it (Supelco, Bellefonte, PA, USA). The extraction and concentration were lasted for 40 min at 30°C. GC-MS was carried out using a Trace DSQ-GC-MS (Thermo Corporation, Waltham, MA, USA). The flow rate of the helium carrier gas in the DB-5MS fused-silica capillary column (30 m × 0.25 mm × 0.25 mm film) was 1.00 mL·min-1. Then, the sample was injected into the injector port at the temperature of 200°C. The column temperature was programmed as follows: the initial temperature was set at 50°C for 1 min, and then increased to 200°C at a rate of 5°C·min-1, finally increased to 230°C at 8 °C·min-1 and maintained for 8 min. The volatile compounds were identified by matching the resulting mass spectra with the NIST 11 library (National Institute of Standards and Technology, Gaithersburg, MD, USA), retention index and relative reports from the literature. Quantitative analysis was carried out by comparing peak areas of volatile compounds with that of the internal standard (Feng et al., 2014). The mass fraction was calculated as compound emission rate (μg·g-1·h-1) = {peak area of compound/peak area of internal standard × concentration of internal standard (μg·μL-1) × volume of internal standard}/sample mass (g)/extraction time (h).
The total RNA of each sample was extracted using a universal RNA extraction kit (Tiangen Biotech Co., Ltd., Beijing, China) according to the manufacturer’s instructions. RNA concentration and quality were determined with a Qubit 2.0 fluorometer (Life Technologies, Carlsbad, CA, USA), and a spectrophotometer (NanoPhotometer; Implen, Calabasas, CA, USA), respectively. RNA integrity was measured using a Bioanalyzer 2100 system with the RNA 6000 Nano Assay kit (Agilent, Carlsbad, CA, USA). Nine strand-specific RNA libraries were prepared with an insert size of ~250–500 nucleotides using a UTP method (Parkhomchuk et al., 2009), and then were sequenced by Biomarker Technologies Corporation (BMK, Beijing, China) on the Illumina HiSeq 2000 platform with the 150-bp paired-end method and a sequencing depth of ~53 million reads per library ( Table 1 ).
Barcode and adaptor sequences were removed from the sequencing reads by the quality checking and trimming processes. Any rRNA sequences were eliminated by aligning all reads to plant rRNA sequences using the Short Oligonucleotide Analysis Package (SOAP2; http://soap.genomics.org.cn/soapaligner.html). The clean reads from each library were then aligned with the reference genome of the rose ‘Old blush’ (ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/002/994/745/GCF_002994745.1_RchiOBHm-V2/GCF_002994745.1_RchiOBHm-V2_genomic.fna.gz) using Hisat2 (version 2.1.0; https://ccb.jhu.edu/software/hisat2/index.shtml). The alignments were used to assemble transcripts using StringTie (version v1.3.6, http://ccb.jhu.edu/software/stringtie/).
The assembled transcripts from each library were merged by Cuffmerge to remove those with uncertain directions or those shorter than 200 nt. Cuffcompare was then used to compare transcripts with the rose genome annotated protein sequences (ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/002/994/745/GCF_002994745.1_RchiOBHm-V2/GCF_002994745.1_RchiOBHm-V2_protein.faa.gz). The non-redundant transcripts exhibiting significant alignment (P<1.0E-10, identity >90%, coverage >80%) with rose proteins were excluded. According to rose genome annotation, all resulting transcripts that aligned to housekeeping ncRNAs (including rRNAs, tRNAs, snRNAs, and snoRNAs) were also removed. Transcripts with short ORFs (<100 amino acids) were detected for the open reading frame (ORF) filter. The longest consecutive codon chain was defined as the putative ORF of the lncRNA candidate. In addition, transcripts were aligned to the protein family (Pfam) database using the HMMER 3.0 program (profile hidden Markov model software) (Finn et al., 2011) with an E-value threshold of 10−5 to filter transcripts containing a known protein domain. The resulting transcripts were tested for protein-coding potential using the Coding Potential Calculator (CPC) software, and only transcripts with a CPC score of <0 were retained and considered as lncRNAs (Kong et al., 2007). The intersection of transcripts with no coding potential in the results of the two software analyses were considered as rose lncRNAs.
Based on location relative to the nearest protein-coding genes, the annotated lncRNAs were subdivided into four categories: (i) antisense lncRNAs, which overlap with exons of a protein-coding transcript on the opposite strand; (ii) lncRNAs without any overlap with other protein-coding genes are classified as intergenic lncRNAs (lincRNAs); (iii) lncRNAs with some overlap with genes on the same strand are classified as sense overlapping lncRNAs; and (iv) lncRNAs in some protein-coding loci but without any overlap with exons of protein-coding genes are classified as sense intronic lncRNAs (Harrow et al., 2012).
LncRNAs and protein-coding genes were analyzed for transcript length and exon number as followings (Zhu et al., 2015). Transcript length categories were <300, 300–400, 400–500, 500–600, 600–700, 700–800, 800–900, 900–1000, and >1000 nucleotides. Exon number categories were: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and >10. The proportions of different kinds of lncRNAs and protein-coding transcripts were then calculated.
There are two predominant mechanisms by which lncRNAs regulate target genes (Schmitt and Chang, 2016). Co-location means that a lncRNA may regulate the adjacent protein-coding genes, while co-expression means that a lncRNA regulates downstream genes through correlated expression. The threshold for the co-location mechanism was set to 100 kb upstream or downstream of the lncRNA location in the chromosome. For co-expression prediction, the pearsonr function was called by the python statistics module of scipy.stats to calculate pearson correlation coefficients of expression levels between lncRNAs and mRNAs in the trans-loci. It was conducted only when the sample number was bigger than five and the threshold for the pearson correlation coefficient was set to greater than 0.95 (Kopp and Mendell, 2018; Bao et al., 2019).
Using the cuffdiff program, both differentially expressed lncRNAs and mRNAs among flower developmental stages were identified (Trapnell et al., 2012). LncRNAs and mRNAs exhibiting |log2 (fold change)| ≥1 and adjusted P-values <0.05 were selected as differentially expressed.
Key lncRNAs correlated to flower volatiles were identified based on dynamic lncRNA expression changes in tissues of different flowers using the R package WGCNA (Langfelder and Horvath, 2008). For the sample number—including the biological replicates—needed by WGCNA was at least 15, the GC-MS and RNA-seq data of our another three rose cultivars’ flowers were recruited. The three cultivars—’Elle’, ‘First blush’ and ‘Qingge’—were parents and sister of ‘Tianmidemeng’ and possessed distinctive floral scent profiles, respectively. The GC-MS data was obtained from their flowers in the same condition and method as ‘Tianmidemeng’, while the RNA-seq data was obtained from their flowers in the same condition but with non-strand-specific RNA sequencing method. LncRNAs were isolated from RNA-seq data of the three cultivars and their expression levels were calculated with the same methods as ‘Tianmidemeng’. Parameters were set up as power = 6, minModuleSize = 6, deepSplit = 4, mergeCutHeight = 0.1, and MEDissThres = 0.15. The TO value (topological overlap, unsigned) was calculated for each pair of lncRNAs (Ravasz et al., 2002; Li and Horvath, 2006; Yip and Horvath, 2007) and a lncRNA cluster tree was subsequently constructed by hierarchical clustering method and further split into modules by the method of dynamic treecut (Langfelder and Horvath, 2008). Eigengenes (ME) of each module were evaluated by principal component analysis (PCA). To correlate flower volatiles with modules, the contents of flower volatiles in every tissue in the developmental process were listed and assembled into a matrix. The coefficient factors between the matrix and MEs were calculated. For each flower volatile, modules with top three coefficient factors were selected.
Validation of the RNA-seq results was conducted by qRT-PCR analysis to ten potential floral-scent-related lncRNAs. RNA samples for the three flowering stages were isolated from the same flower tissues as RNA-seq libraries, respectively. The cDNA for each sample was then synthesized using ReverTra Ace qPCR RT Master Mix (Toyobo, Japan). Primers for qRT-PCR were designed using Primer Premier software (version 5.0), listed in Supplementary Table 1 . A StepOnePlusTM Real-Time PCR System (Life Technologies, Carlsbad, CA, USA) was used to detect relative lncRNA expression levels with the SYBR® Green Real-Time PCR Master Mix (Toyobo, Japan). Three biological replicates were performed, and the reactions were performed in triplicate for each run. The quantification of the relative expression of the genes at different times was performed using the delta-delta Ct method as described by Livak and Schmittgen (Livak and Schmittgen, 2001). All data were expressed as means ± standard deviation (SD) after normalization. GAPDH was used as an internal control. Linear regression analysis was conducted using the fold-change values of qRT-PCR and RNA-Seq.
Virus-induced gene silencing (VIGS) was used to examine the functions of the candidate lncRNAs for floral scent production in the rose. LncRNA fragments of 300-500bp were amplified from the cDNA of ‘Tianmidemeng’, and then inserted into the vector of pTRV2 with a homologous recombination method The pTRV1, pTRV2 and pTRV2-lncRNA constructs were transformed into the competent cells of Agrobacterium strain GV3101, respectively. Monoclonal colonies of GV3101 with pTRV1, pTRV2 or pTRV2-lncRNA vectors were cultured in LB medium (pH 5.6) containing 10 mM MES and 20 μM acetosyringone with kanamycin, gentamycin, and rifampicin antibiotics at 28 °C for 24 h, The cultures were collected by centrifugation and then resuspended in the infiltration buffer (10 mM MgCl2, 200 μM acetosyringone, 5% sucrose) until the OD600 of the resuspension buffer arrived to 1.2-1.5. The infection buffer was prepared by mixing the resuspension buffers of pTRV1 and pTRV2, or pTRV2-lncRNA at a ratio of 1:1, and then placed in the dark for 2-3 h. Flowers of rose cultivar ‘Tineke’ in the EF stage were pricked by a needle in four directions, and then submerged in the infection buffer and subjected to a vacuum at 0.8-1 bar twice, each for 60 s. Infiltrated flowers were washed with distilled water and then grown in clean water at 8 °C in dark for 3 d and then in the greenhouse for another 3 d. Flowers infiltrated by infection buffer with pTRV1 and pTRV2 were set as controls. Each infiltration was carried out with 10 biological replicates. The collection and measurement of samples for GC-MS and qRT-PCR were conducted as mentioned above. Primers for VIGS and qRT-PCR were designed using Primer Premier software (version 5.0), listed in Supplementary Table 1 .
The compositions and contents of floral volatiles were detected by GC-MS for three flowering stages of the rose ‘Tianmidemeng’, respectively ( Figure 1 ). The non-floral-scent components, such as aliphatic hydrocarbons, alkenes, alcohols, aldehydes, acids, and esters, were excluded according to the criteria for the three classes of floral scent compounds and relevant reports about rose floral scent (Joichi et al., 2005; Schnepp and Dudareva, 2008; Bendahmane et al., 2013). Then the floral scent components in three flowering stages were obtained, of which the numbers were 16, 36, and 44, respectively, including terpenoids, phenylpropanoids/benzenoids, and fatty acid derivatives ( Supplementary Table 2 ). Based on the internal standard, the release amount (release rate) of each floral scent component in the three flowering periods was calculated ( Figures 2A–D ). Total release in the initial stage was 4.478 ug·g-1·h-1, and this increased significantly to 51.84 ug·g-1·h-1 in the semi-flowering stage but was 20.93 ug·g-1·h-1 in the final stage, which was a significant decrease compared with that in the semi-flowering stage ( Figure 2A ). This showed that floral scent synthesis of ‘Tianmidemeng’ changed significantly following the development of flower and reached the highest value in the semi-flowering stage. When floral scent components were classified into the three classes, it was found that the terpenoids accounted for the majority of the floral scent in the three flowering periods: 98.9% at the bud stage, 82.3% at the semi-flowering stage, and 83.8% at the late-flowering stage, and the changes of release amount were consistent with those of the overall floral scent ( Figure 2B ). The release changes of phenylpropanoids/benzenoids were the same as for the terpenoids ( Figure 2C ), while the release changes of fatty acid derivatives were similar to those of terpenoids and phenylpropanoids/benzenoids but the decrease was not significant from semi-flowering to late-flowering stages ( Figure 2D ). These results demonstrated that the synthesis of fatty acid derivatives was less affected by petal senescence compared with the other two classes of compounds (terpenoids and phenylpropanoids/benzenoids). In summary, floral scent synthesis of the rose ‘Tianmidemeng’ was regulated by the flowering process, and the semi-flowering stage was the period with the highest synthesis of various floral scent components. However, the responses of three kinds of floral scent compounds to the flowering process were slightly different; terpenoids and phenylpropanoids/benzenoids were sensitive to flower opening and aging, while the fatty acid derivatives were less sensitive to flower aging.
To identify lncRNAs in rose flowers, paired-end ssRNA-Seq for early-flowering, semi-flowering, and late-flowering stages of ‘Tianmidemeng’ was performed in three biological replicates. A total of ~476 million clean reads were obtained ( Table 1 ; Figure 3 ), and 102,426 unique transcripts were assembled ( Figure 3 ). To distinguish lncRNAs, five sequential stringent filters were applied to the 102,426 transcripts ( Figure 3 ). First, the transcripts were filtered with rose ncRNAs. A total of 263 rRNAs, 486 tRNAs, 472 snoRNAs, and 131 sRNAs were excluded according to the ‘Old blush’ genome annotation, leaving 101,074 transcripts. Among these, transcripts with a single exon were filtered for low reliability, and 95,188 transcripts with an exon number ≥2 were selected ( Figure 3 ). The resulting transcripts were then filtered with rose coding gene sequences. Almost 42% (39,571) of transcripts were coding genes, and the remaining 58% (55,617) might potentially be non-coding transcripts, consistent with other studies and showing that ncRNAs were widely transcribed ( Figure 3 ) (Heo et al., 2013). Among the 55,617 potential non-coding transcripts, 14,479 were annotated by the ‘Old blush’ genome annotation, with 10,887 annotated as lncRNAs ( Figure 3 ). The unannotated 41,138 transcripts were further analyzed for novel lncRNAs. Two criteria—longer than 200 nucleotides and unable to encode polypeptides longer than 100 amino acids—were applied to the 41,138 transcripts, and 41,115 transcripts were recovered (Li et al., 2014; Shuai et al., 2014) ( Figure 3 ). Coding potential is the key condition to judge whether a transcript is lncRNA. Transcripts with potential protein-coding domains were therefore further filtered by comparison with the Pfam database. Finally, after the assessment by CPC software, transcripts without protein-coding potential were obtained as the novel lncRNAs. After employing three stringent criteria, 3070 transcripts were considered as novel lncRNAs. Thus, a total set of 13,957 transcripts were obtained and defined as rose lncRNAs, including 10,887 annotated lncRNAs and 3070 novel lncRNAs ( Supplementary Table 3 ).
LncRNAs were further classified into four types according to the location relative to the nearest protein-coding genes. These types are: intergenic (lincRNA), sense intronic, sense overlapping, and antisense lncRNAs ( Figure 4A ) (Harrow et al., 2012). Most of the lncRNAs—5999 lncRNAs (43.0%)—were located in intergenic regions, whereas 2453 (17.6%) and 24 (0.2%) of the lncRNAs were either antisense of or overlapped with protein-coding genes ( Figure 4A ). This observation was consistent with previous studies (Li et al., 2014). In addition, 5481 lncRNAs (39.3%) were transcribed from inside genes (most from introns), which was similar to the result obtained for Arabidopsis but widely divergent to the result in tomato (Wang et al., 2014; Zhu et al., 2015). The numbers of three types of lncRNAs—lincRNA, sense intronic, and sense overlapping—from plus and minus strands (Watson and Crick strands) were similar ( Figure 4B ), while the number of sense overlapping lncRNAs was different from these ( Figure 4C ). Plant lncRNAs are reported to be shorter and harbor fewer exons compared with protein-coding genes (Li et al., 2014; Shuai et al., 2014). To determine whether rose lncRNAs shared these features, all the 42,767 genes predicted in the genome of the rose ‘Old Blush’ were applied to analyze the distribution of length and exon number of the 13,957 lncRNAs. Figure 5A shows that ~55% of the lncRNAs ranged in size from 200 to 1000 nucleotides, with only 45% comprising >1000 nucleotides. In contrast, for the protein-coding transcripts, ~80% comprised >1000 nucleotides. Most (70%) of the genes encoding rose lncRNAs only contained ≤5 exons, while the number of exons for the protein-coding genes ranged from one to ≥20 ( Figure 5B ). All rose lncRNAs possessed ORFs with a length shorter than 100 amino acids, while the ORF lengths of protein-coding genes ranged from one to ≥1000 AA ( Figure 5C ). Collectively, these results indicated that most of the rose lncRNAs are relatively short and contain only a few exons compared to protein-coding genes.
The emission rate of floral-scent compounds changed among the flowering stages of rose. Therefore, it was hypothesized that floral-scent-related mRNAs and lncRNAs might be present in roses. A total of 9664 mRNAs were differentially expressed among the three flower-development stages of the rose ‘Tianmidemeng’ ( Supplementary Table 4 ). For lncRNAs, 534 of them were identified as differentially expressed. Among them, 109 lncRNAs were excluded as their expression levels were lower than 0.5 in all three stages, leaving 425 differentially expressed lncRNAs in the research ( Supplementary Table 5 ). From the early-flowering stage to the semi-flowering stage, 214 lncRNAs were differentially expressed, of which 140 and 74 were upregulated and downregulated, respectively. From the semi-flowering stage to the late-flowering stage, 83 lncRNAs were differentially expressed, of which 38 and 45 were upregulated and downregulated, respectively ( Figures 6A, B ). There were 19 lncRNAs expressed in both of these processes and these were deemed core candidate lncRNAs. Among the core lncRNAs, the expression levels of nine lncRNAs increased from bud to semi-flowering stage and then subsequently decreased, while nine lncRNAs decreased first and then increased ( Figure 6A ). This indicated that the nine and nine lncRNAs might have positive and negative roles, respectively, in floral scent synthesis. LncRNA is rich in biological functions and is involved in various important physiological processes. LncRNA can regulate the expression of target genes at transcriptional and post transcriptional levels (Schmitt and Chang, 2016). The two predominant mechanisms by which lncRNA regulates target gene are co-location, where the lncRNA may regulate the adjacent protein-coding genes, and co-expression, where the lncRNA regulates downstream genes through correlated expression. On the basis of these two mechanisms, 10,075 of the total 13,957 lncRNAs were predicted to possess a set of 29,622 target genes ( Supplementary Table 6 ). The above correlation prediction was also used to search for lncRNAs related to floral scent production. First, all genes involved in downstream synthesis pathways of floral scent compounds from were identified from the rose mRNAs in this research, and a total of 54 genes were obtained. Using the correlation prediction between lncRNA and target genes, a total of 849 corresponding lncRNAs were identified for the 54 genes, among which, 141 corresponding lncRNAs were differentially expressed ( Supplementary Table 7 ). Among the 141 lncRNAs, five were core lncRNAs, including TCONS_00007202, TCONS_00008447, TCONS_00117855, XR_002924185.1, and XR_002931444.1. The expression changes of upstream genes in floral scent synthesis pathways were usually irregular, hence their corresponding lncRNAs with irregular expression changes might also be candidates for floral scent. Accordingly, 15 upstream genes were selected and 87 differentially expressed lncRNAs corresponding to these genes were identified ( Supplementary Table 8 ). Among the 87 lncRNAs, three were core lncRNAs, and 84 were other potential lncRNAs. The three core lncRNAs—TCONS_00007202, TCONS_00008447, and TCONS_00117855—were also obtained in the analysis of downstream genes. To summarize, a total of 103 candidate lncRNAs for floral scent production were identified. Among them, the core 19 lncRNAs were identified according to gene expression changes during the flowering process, including five that were further validated by correlation analysis between lncRNAs and target genes of downstream syntheses in the floral scent synthesis pathway, and the other 84 lncRNAs were identified according to correlation analysis between lncRNAs and target genes of upstream syntheses in floral scent synthesis pathway. These lncRNAs are likely to be involved in floral scent production in rose. The expression profiles of some identified lncRNAs were confirmed by qRT-PCR. We randomly selected 10 lncRNAs—five with an up-down change and five with a down-up change—among the three flowering stages in sequencing results to conduct qRT–PCR validations ( Supplementary Figure 1 ). The fold changes in the lncRNA expression levels measured by qRT–PCR were closely correlated to that by RNA-Seq (R2 = 0.57, P<0.001) ( Figure 7 ), showing a good consistency between the qRT-PCR and RNA-seq results. It further suggested that the lncRNAs would have a role in floral scent production.
R2R3-MYBs and other TFs were reported to regulate floral scent synthesis in various plants (Muhlemann et al., 2014; Yeon and Kim, 2021). Based on previous publications about TFs for floral scent compounds, the sequences of all the TFs were collected—a total of 169 TFs, namely 139 MYBs, 7 bHLHs, 6 AP2/ERF, 4 WRKY, 5 NACs, 5 bZIPs, and 2 zinc finger-like, 1 ETHYLENE-INSENSITIVE3-like TFs—and used to screen for rose homolog transcripts. Among these TFs, 24 were differentially expressed, including 1 NAC, 18 MYBs, 2 ERFs, 1 bHLH, and 2 bZIPs, and all of them were potential candidate TFs for regulation of floral scent synthesis. Using the correlation prediction between lncRNAs and target genes, the 24 TFs were possible target genes of 208 lncRNAs. Among these 208 lncRNAs, 61 were differentially expressed and deemed as potential candidates ( Supplementary Table 9 ). For example, for PbbHLH4, a TF regulating floral scent production in Phalaenopsis, the homolog in rose was R. chinensis ICE1-like transcription factor gene (ID: 112175393), and the correlated lncRNA was TCONS_00111355, which was among the core lncRNAs. The results of this analysis indicated that these lncRNAs are likely to be involved in floral scent production via some TFs in rose.
WGCNA was employed to correlate lncRNAs with individual floral volatile compounds. As RNA-seq datas of the other three cultivars were obtained using non-strand-specific RNA sequencing method, less lncRNAs were identified from them compared to ‘Tianmidemeng’. Finally, 224 of the 425 differentially expressed lncRNAs in ‘Tianmidemeng’ were identified from the other three RNA-seq datas and clustered into 13 modules by WGCNA ( Figure 8 ). All lncRNAs in the modules are listed in Supplementary Table 10 . A correlation map between modules and compounds was generated ( Supplementary Figure 2 ) and the top three modules for each compound were selected according to the correlation rate. By this method, 11 modules were identified as related to 19 floral volatiles of the rose, including black, blue, green, greenyellow, magenta, pink, purple, red, tan, turquoise, and yellow ( Table 2 ). All the 11 modules were correlated with the 14 compounds of terpenoids, of which the green yellow module was most correlated to half of the terpenoid compounds, including geraniol, nerol, geranyl acetate, citronellyl_acetate, citral, β-Pinene, and dihydro-β-ionol, and secondly correlated to neral and neranyl acetate. Among the left 5 terpenoid compounds, trans-β-ocimene, 7,8-dihydro-β-ionone, and trans-β-ionone were most correlated to magenta module, while the most correlated module of β-copaene and aromandendrene was turquoise. Among phenylpropanoids/benzenoids, phenethyl alcohol, phenethyl acetate and DMT were all most correlated with lncRNAs in modules of purple. Another benzenoids methyleugenol was correlated with lncRNAs in modules of greenyellow, pink and green, while the fatty acid derivative 4-hexen-1-ol-acetate was correlated with modules of greenyellow, magenta and black. A total of 11 core lncRNAs were involved in six modules, of which five were related to floral scent production according to the above analysis, including green, greenyellow, purple, tan, and turquoise. The lncRNAs TCONS_00007202, XR_002924185.1 and XR_002931444.1, predicted to regulate floral scent synthase genes in the above correlation analysis, were involved in tan and greenyellow modules, respectively. Furthermore, as a potential regulator of the TF PbbHLH4 homolog for the synthesis of monoterpenes in the rose (ID: 112175393), the lncRNA TCONS_00111355 was involved in the green module, which was predicted as related to terpenoids production. The WGCNA results were consistent with the above correlation analysis and further validated the role of core lncRNAs in regulating floral scent production directly or via TFs.
One core lncRNA TCONS_00008447 was selected to be silenced in rose ‘Tineke’ using the VIGS method. Its expression pattern presented an up-down change and was validated by the qRT-PCR ( Supplementary Figure 1 ). After six days, flowers of rose ‘Tineke’ infiltrated with TRV-TCONS_00008447 ( Figure 9B ) showed a little unfolded and withered compared to TRV control flowers ( Figure 9A ). The infection of TRV- TCONS_00008447 into flowers aroused an emission increase of terpenoid 7,8-dihydro-β-ionone by 3.9 folds compared to TRV control flowers ( Figure 9D ). The qRT-PCR result revealed that the expression of TCONS_00008447 was decreased by 43% compared to the control flowers ( Figure 9C ). The result suggested that the TCONS_00008447 was involved in the regulation of floral scent production in the rose.
Floral scent production markedly changes during flower development, corresponding with its role in attracting pollinators to plants (Shalit et al., 2003; Boatright et al., 2004; Colquhoun et al., 2010; Rodriguez-Saona et al., 2011). In common, flowers did not emit fragrance until they arrived at a state similar to EF stage of rose in this paper, and emissions of most of their fragrance compounds peaked at a state that the petals were opened and stamens were exposed thoroughly, which could be deemed as full-flowering (FF) stage. However, some fatty acid derivatives decreased from the EF stage to the end of the flower development (Shi et al., 2018; Yang et al., 2021). R. hybrida was the hybrid progeny of Chinese and European roses, and inherited the complicated floral-scent profiles of the parents. During the flowering process of the rose, hundreds of volatile molecules could be obtained (Joichi et al., 2005; Bendahmane et al., 2013). Unlike other species, emissions of the most floral-scent compounds in the rose peaked at the SF but not FF stage (Guterman et al., 2002; Shalit et al., 2004; Feng et al., 2014; Yeon and Kim, 2020), indicating that the rose produced most of its scent compounds during the petal opening process. It was consistent with the second phase of petal development in the rose that petals grow rapidly resulting only from cell expansion, which was accompanied with the most production of floral scent (Guterman, 2002). In the research, emissions of the fatty acid derivatives peaked in SF stage as well and retained a higher level until the LF stage in ‘Tianmidemeng’, inconsistent with results in other species but consistent with former reports in roses (Shalit et al., 2004). Therefore, some different regulatory mechanisms may be involved in fatty acid derivatives production in rose. Whatever, the results also supported another finding that there was no direct relationship between fragrance synthesis and senescence of rose flowers (Borda et al., 2011).
Although many lncRNAs have been identified from numerous model plants, such as Arabidopsis, there are limited studies on lncRNAs in rose and further research is required in this area (Liu et al., 2022). In the present study, with the strand-specific RNA-Seq and a strict criteria pipeline widely used in previous studies in plants (Zhu et al., 2015; Zhang et al., 2016; Wang et al., 2017), a total of 13,957 lncRNAs were identified and classified in rose, a model plant for the study of floral scent ( Figure 3 ). Common transcriptome library construction and sequencing cannot separate the sense and antisense strands, which resulted in quite a missing of lncRNAs. With application of the strand-specific RNA-Seq, the strand orientation information of the lncRNAs were conserved, thus facilitating their identification and functionally analysis (Di et al., 2014; Shuai et al., 2014). The development of third-generation sequencing technology further promoted the study for lncRNAs, for it could catch the strand orientation information without strand-specific library and obtain longer lncRNAs (Cui et al., 2019; Teng et al., 2019). Therefore, the limitations of our rose lncRNA list still remained, including that the pair-end sequencing could not obtain complete sequences for all lncRNAs thorouly and the RNA-seq was not deep enough to explore rose lncRNAs fully. In summary, the specific sequencing and strict bioinformatics criteria of the current study generated a relatively reliable list of rose lncRNAs, which will potentially benefit to other researchers.
Remarkable progress has been made in elucidating important roles of lncRNAs in multiple of physiological processes in plants, including phosphate homeostasis, vernalization response, immune response, root development, seedling photomorphogenesis, gametophyte development, stress response, nitrate response, rice yield, leaf morphological development, disease resistance, pathogen infection, tomato ripening process, formation of root nodules, pollen development, male fertility, and so on (Wu et al., 2020). LncRNAs might be a general component of plant immune responses, for numerous differentially expressed lncRNAs were identified from the pathogen-infected plants including tomato, cotton, arabidopsis, rice, and mutants of some of them were shown to alter plant resistance to pathogens (Liu et al., 2022). For example, accumulation of a pathogen-responsive lncRNA ALEX1 could activate the (jasmonic acid) JA pathway in rice and enhance its resistance to bacterial blight (Yu et al., 2020), LincRNA CRIR1 regulated cold stress response of the cassava by modulating the expression of stress-responsive genes and increasing the translational yield (Li et al., 2022). It was found that lncRNAs could be transferred between Cuscuta Parasites and its host soybean plants, indicating their critical role as regulators to coordinate the host–dodder interaction (Wu et al., 2022). Moreover, lncRNAs could act as a switch in balancing plant defense and growth. In Arabidopsis thaliana, lcRNA SABC1 could repress plant immunity via decreasing transcription factor gene NAC3 and isochorismate synthase 1 (ICS1) transcriptions. However, upon pathogen infection, SABC1 was downregulated to depress plant resistance to bacteria and viruses (Liu et al., 2022). Due to the dual role in plant pollination and defense of floral scent, whether lncRNAs involved in its production functioned similarly was worth anticipating. In spite of various roles of lncRNAs in plant physiological processes, their functions in floral scent synthesis were absent in current researches. By a comprehensive approach combining methods of differential-expression analyses, co-location and co-expression prediction and WGCNA analysis, we predicted candidate lncRNAs for floral production in the rose. The results of the consequent VIGS experiment initially confirmed their regulator role in rose floral scent synthesis. However, function mechanisms of them would be further investigated. How lncRNAs regulate diverse biological processes is far from clear. It was found that lncRNAs with low expression tended to amplify their action by targeting transcription factors, while the cis-acting lncRNAs usually regulated the expression of their neighboring genes in the nucleus via epigenetic modifications (Gil and Ulitsky, 2020; Rinn and Chang, 2020). The trans-acting lncRNAs were usually indentified by a co-expressional grithm. In rice plants infected by rice black-streaked dwarf virus (RBSDV), a co-expression network of 56 differentially-expressed mRNAs and 20 differentially-expressed lncRNAs was construced, in which five mRNAs were verified to be regulated by three lncRNAs by the experiment conducted in rice calli (Zhang et al., 2020). Cis-acting LncRNAs functioned by recruiting DNA methyltransferases or demethylases to regulate the target gene transcription. In Arabidopsi, the lncRNA COLDAIR was generated from the intron of FLC and repress its expression by recruiting PRC2 via H3K27me3 (Heo and Sung, 2011).
In the past decade, several TFs were found to play important role in floral scent synthesis. Several TFs regulate gene expression of phenylpropanoid/benzenoid production in flowers, including four R2R3-type MYB TFs in petunia—ODO1 (Verdonk et al., 2005), EOBI (Van Moerkercke et al., 2011; Spitzer-Rimon et al., 2012), EOBII (Spitzer-Rimon et al., 2010; Colquhoun et al., 2011; Van Moerkercke et al., 2011), and PH4 (Cna'ani et al., 2015)—and two repressor TFs—PhMYB4 (Colquhoun et al., 2010) in petunia and MYB3 in Arabidopsis (Zhou et al., 2017). For terpene biosynthesis, two cases of TFs have been reported in floral organs. Firstly, in A. inflorescence, bHLH-like transcription factor AtMYC2 promoted the synthesis of sesquiterpene (E)-β-caryophyllene by binding to AtTPS11 and AtTPS21 promoters of the terpene synthetase gene (Hong et al., 2012). Secondly, in petals of Phalaenopsis bellina, five TFs—PbbHLH4, PbbHLH6, PbbZIP4, PbERF1, and PbNAC1—promoted the synthesis of floral terpene components, with PbbHLH 4 improving the expression of geranyl diphosphate synthase gene GDPS by combining with its promoter and enhancing the synthesis of monoterpenes in floral scent (Chuang et al., 2018). LncRNAs were supposed to target transcription factor genes to amplify their actions. A heat−inducible antisense lncRNA was involved in gametophyte development of A. thaliana by controlling the heat shock factor HSFB2a (Wunderlich et al., 2014). A novel ribonucleoprotein complex with lncRNA APOLO and the transcription factor WRKY42 forms a regulatory hub to trigger root hair cell expansion in response to cold by activating the master regulator RHD6 in Arabidopsis (Moison et al., 2021). In P. tomentosa, lncRNA PMAT interacted epistatically with PtoMYB46 promoted Pb2+ tolerance, uptake and plant growth of poplar by repressing PtoMATE and PtoARF2 (Chen et al., 2022). Interestingly, lncRNAs tended to target transcription factor genes nearby them, such as TWISTED LEAF, circular RNA (circRNA) SEP3, and SUF (Conn et al., 2017; Liu et al., 2018; Hisanaga et al., 2019). In the present study, homolog transcripts of 169 TFs reported to be involved in the production of floral volatile compounds in plants were obtained from rose RNA-seq data and were predicted to be target genes of 208 lncRNAs. Whether and how the lncRNAs regulate their expression would be further verified by biological experiments in the rose. LncRNAs could bind miRNAs as eTMs to regulate the expressions of target mRNAs. The LnRNA TCONS_00021861 could regulate YUCCA7 by sponging miR528-3p, to activate IAA biosynthetic pathway and confer resistance to drought stress in rice (Chen et al., 2021). Overexpression of lncRNA23468 in tomato significantly decreased expression of miR482b, and then increased the expression of its target genes NBS-LRRs, resulting in enhanced resistance to Phytophthora Infestans (Jiang et al., 2019). LncRNA regulated the expression of CSD1 by indirectly through competitively binding miR398 to improve cold resistance of winter wheat (Lu et al., 2020). In the rose, transcriptomic sequencing revealed the presence of a large number of ncRNAs, and the miR156 was proposed to be involved in synthesis of some terpenes in petals (Raymond et al., 2018). Although the prediction of target miRNAs for lncRNAs were lacked in this paper, it was essential to supplement for it in the future. Some protocols have been developed for miRNA–lncRNA interaction prediction in plants, such as an ensemble deep learning model based on multi-level information enhancement and greedy fuzzy decision (PmliPEMG), which could be applied to the cross-species prediction (Kang et al., 2021), an ensemble pruning protocol that for minining plant eTMs by predicting miRNA-lncRNA interactions based on dual-path parallel ensemble pruning method (Kang et al., 2022). By constructing the lncRNA-miRNA-mRNA regulatory network through biological experiment, the functions of potential eTMs could be further inferred through enrichment analysis.
An overview of the transcriptional regulation of floral scent production by lncRNAs in rose flowers was generated using a variety of techniques and analyses including genome-wide identification, characterization, differential expression, and co-expression network analysis of intergenic/intronic lncRNAs. As the first lncRNA research in rose, 13,957 lncRNAs were identified, including 10,887 annotated lncRNAs and 3070 novel lncRNAs, while 19 core lncRNAs were predicted to be candidates participating in floral scent synthesis. WGCNA suggested that expression of the 11 lncRNAs is highly enriched in co-expressed modules that are related to floral scent synthesis pathways, and function of one of them were confirmed by the VIGS experiment. Future research efforts will aim to elucidate the mechanism by which these lncRNAs regulate floral scent production. Overexpression, RNA interference, and promoter analysis are useful experimental approaches for characterizing lncRNA functions, which might provide valuable information for improving floral scent in rose.
The original contributions presented in the study are publicly available. This data can be found here: NCBI, PRJNA667625.
ZZ: conceived and designed the experiments. ZZ, SS, JW and XL: methodology. SS and SZ: experiments. SS: analysis of data. SS and SZ: writing original draft preparation. ZZ and SS: writing—review and editing. ZZ: supervision. All authors contributed to the article and approved the submitted version.
This work was funded by Shandong Provincial Natural Science Foundation (Grant number ZR2020QC159), Natural Science Foundation of Beijing Municipality (Grant number 6222030), National Natural Science Foundation of China (Grants numbers 32102430, 31501791, 31772344 and 31972444) and Innovation Project of Shandong Academy of Agricultural Sciences (Grants numbers CXGC2021B17, CXGC2022A06).
The authors acknowledge Pro. Mignfang Yi for her scientific suggestions, acknowledge Novogene Co., Ltd. for their technical support, and acknowledge Researcher Yumeng Huo for providing server for our bioinformatic analysis.
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 | false |
|
PMC9577378 | 36190397 | Jinbi Xie,Yong Ning,Lihang Zhang,Yuan Lin,Runsheng Guo,Shanjuan Wang | Overexpression of hsa_circ_0006470 inhibits the malignant behavior of gastric cancer cells via regulation of miR-1234/TP53I11 axis | 03-10-2022 | Gastric cancer,hsa_circ_0006470,miR-1234,TP53I11,cell viability | Gastric cancer (GC) is a subtype of a common malignant tumor found in the digestive system. Hsa_circ_0006470 is known to be closely associated with the development of GC. Nevertheless, the mechanism by which hsa_circ_0006470 regulates the tumorigenesis of GC has not been fully elucidated. To investigate the role of hsa_circ_0006470 in GC, its expression levels were assessed in GES-1, AGS, MKN45, and SNU5 cells by reverse transcription-quantitative PCR. Fluorescence in situ hybridization was used to evaluate the localization of hsa_circ_0006470 in AGS and MKN45 cells. In addition, cell counting kit-8 and 5-ethynyl- 2’-deoxyuridine assays were performed to evaluate the viability and proliferation of GC cells, respectively. The dual-luciferase reporter assay was used to explore the interaction among hsa_circ_0006470, microRNA (miR)- 1234, and TP53I11. The expression levels of TP53I11, Akt, p-Akt, forkhead box O1, and cyclin dependent kinase 2 in AGS cells were analyzed by Western blotting. The data indicated that hsa_circ_0006470 expression was downregulated in AGS cells. In addition, overexpression (OE) of hsa_circ_0006470 could inhibit the viability and proliferation of GC cells. Moreover, OE of hsa_circ_0006470 inhibited the migration of GC cells and induced G1 cell cycle phase arrest. Moreover, miR-1234 was bound to hsa_circ_0006470 and TP53I11 was targeted by miR-1234. Furthermore, OE of hsa_circ_0006470 inhibited the tumorigenesis of GC via the regulation of the miR-1234/TP53I11 axis. In summary, the present study demonstrated that OE of hsa_circ_0006470 notably inhibited the tumorigenesis of GC by regulating the miR-1234/TP53I11 axis. Therefore, the present study may provide a theoretical basis for exploring novel therapeutic strategies for the treatment of GC. | Overexpression of hsa_circ_0006470 inhibits the malignant behavior of gastric cancer cells via regulation of miR-1234/TP53I11 axis
Gastric cancer (GC) is a subtype of a common malignant tumor found in the digestive system. Hsa_circ_0006470 is known to be closely associated with the development of GC. Nevertheless, the mechanism by which hsa_circ_0006470 regulates the tumorigenesis of GC has not been fully elucidated. To investigate the role of hsa_circ_0006470 in GC, its expression levels were assessed in GES-1, AGS, MKN45, and SNU5 cells by reverse transcription-quantitative PCR. Fluorescence in situ hybridization was used to evaluate the localization of hsa_circ_0006470 in AGS and MKN45 cells. In addition, cell counting kit-8 and 5-ethynyl- 2’-deoxyuridine assays were performed to evaluate the viability and proliferation of GC cells, respectively. The dual-luciferase reporter assay was used to explore the interaction among hsa_circ_0006470, microRNA (miR)- 1234, and TP53I11. The expression levels of TP53I11, Akt, p-Akt, forkhead box O1, and cyclin dependent kinase 2 in AGS cells were analyzed by Western blotting. The data indicated that hsa_circ_0006470 expression was downregulated in AGS cells. In addition, overexpression (OE) of hsa_circ_0006470 could inhibit the viability and proliferation of GC cells. Moreover, OE of hsa_circ_0006470 inhibited the migration of GC cells and induced G1 cell cycle phase arrest. Moreover, miR-1234 was bound to hsa_circ_0006470 and TP53I11 was targeted by miR-1234. Furthermore, OE of hsa_circ_0006470 inhibited the tumorigenesis of GC via the regulation of the miR-1234/TP53I11 axis. In summary, the present study demonstrated that OE of hsa_circ_0006470 notably inhibited the tumorigenesis of GC by regulating the miR-1234/TP53I11 axis. Therefore, the present study may provide a theoretical basis for exploring novel therapeutic strategies for the treatment of GC.
Gastric cancer (GC) is a subtype of a common malignant tumor found in the digestive system. The main factors that induce GC are poor nutritional habits (long-term intake of preserved foods), Helicobacter pylori infection, chronic diseases and genetics. Early-stage GC is mostly asymptomatic or presents with mild symptoms. When the clinical symptoms are evident, the disease is already at an advanced stage. Patients with advanced GC suffer from great physical and psychological pain. The main treatment strategies for GC include drug and surgical treatments; however, the outcomes are not ideal. Therefore, it is imperative to explore novel therapeutic strategies against GC. Circular RNA (circRNA) is a subtype of closed circular RNA molecules formed by reverse splicing. The characteristics of circRNA molecules include high stability, biological evolutionary conservation and tissue expression specificity. It has been reported that circRNAs can exert various biological functions. For example, they can be used as a “sponge” of microRNAs (miRs) or as a competitive endogenous RNA of miRs. CircRNAs are differentially expressed in multiple tumors (GC, prostate cancer, and colorectal cancer). Moreover, hsa_circ_0004872 significantly inhibits the growth of GC cells by regulating miR-224 expression. In addition, overexpression (OE) of circCUL2 inhibits the malignant phenotype of cancer cells by targeting miR-142-3p. Moreover, hsa_circ_0006470 expression was downregulated in GC. Nevertheless, the mechanism by which hsa_circ_0006470 regulates the tumorigenesis of GC is still unclear. Based on the aforementioned evidence, the present research study aimed to investigate the function of hsa_circ_0006470 in GC. This investigation can aid the exploration of novel therapeutic strategies against GC.
The human gastric epithelial cell line GES-1 was purchased from Beyotime Institute of Biotechnology (Suzhou, China). The GC cell lines AGS and SNU5 were obtained from ATCC. The GC cell line MKN45 was provided by Procell Life Science & Technology Co., Ltd. (Wuhan, China). GES-1, AGS, MKN45, and SNU5 cells were cultured in DMEM supplemented with 100 U/mL penicillin/streptomycin and 10% FBS at 37°C in the presence of 5% CO2.
The isolation of total RNA from GES-1, AGS, MKN45, or SNU5 cells was achieved with TRIzol® reagent (Takara Bio, Inc., Kusatsu, Japan). Subsequently, the RNA was converted into cDNA using EntiLink™ 1st Strand cDNA Synthesis Kit (Takara Bio, Inc.). The StepOne™ Real-Time PCR System (Applied Biosystems 7500) was used to perform quantitative PCR. The primer sequences used were as follows: β-actin, forward, 5’-GTCCACCGCAAATGCTTCTA- 3’, reverse, 5’-TGCTGTCACCTTCACCGTTC- 3’; hsa-circ-0006470, forward, 5’-TTCGACTCATCATGGACTCCC- 3’, reverse, 5’-GACACAAAGAAGATGCGGTCC- 3’; hsa-liner-0006470, forward, 5’-CTGCTAAGGAGGTGCTCAACG- 3’, reverse, 5’-CGTGTGCTGCTCAAACTTG- 3’; TP53I11, forward, 5’-TGATGCGGTCTTTGATGGAG- 3’, reverse, 5’-CAGTGACCACCAAGAACTGGAC- 3’. The expression levels of the genes of interest were normalized to those of β-actin. The data were analyzed using the 2–ΔΔCq method.
FISH was applied to evaluate the localization of hsa_circ_0006470 in AGS and MKN45 cells. The same method was applied to identify the interaction of hsa_circ_0006470 and miR-1234 in AGS cells. The fluorescence-labeled oligonucleotide probes for hsa_circ_0006470 and miR-1234 were obtained from Guangzhou RiboBio Co., Ltd. (Guangzhou, China). Subsequently, AGS and MKN45 cells were incubated with the mixture containing the probes, and FITC-Avidin was used to stain the probe. In addition, the nuclei were stained using 4’,6-diamidino-2-phenylindole (DAPI). Subsequently, the localization of hsa_circ_0006470 in AGS and MKN45 cells was investigated using a fluorescence microscope (BX51TF, Olympus Corporation, Tokyo, Japan). The filter sets were as follows: blue excitation: 420~485 nm; green excitation: 460~550 nm; red excitation: 420~490 nm.
AGS and MKN45 cells were transfected with a pcDNA3.1 vector designed to induce OE of hsa_circ_0006470 (Guangzhou RiboBio Co., Ltd.) or a pcDNA3.1 control vector (pcDNA3.1-ctrl, Guangzhou RiboBio Co., Ltd.) using LipofectamineTM 2000 (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer’s protocol. AGS cells were transfected with miR-1234 mimics or miR-1234 mimics-control (mimics- ctrl, Guangzhou RiboBio Co., Ltd.) using LipofectamineTM 2000. AGS cells were transfected with small interfering RNA (siRNA) against TP53I11 (TP53I11 siRNA1, TP53I11 siRNA2, TP53I11 siRNA3 or siRNA-control (siRNA-ctrl) (Guangzhou RiboBio Co., Ltd.) using LipofectamineTM 2000. The sequences used for TP53I11 siRNAs and siRNA-ctrl were as follows: TP53I11 siRNA1: 5’-AGCAGTCAGTAGTTGGTCCTTTG-3’; TP53I11 siRNA2: 5’-CATCATCCCTGCCTCTACTGG-3’; TP53I11 siRNA3: 5’-CCATCAGTCCCGTCTTGAAAC-3’; siRNA-ctrl: 5’-GTGGGTGTCGCTGTTGAAGTC-3’.
CCK-8 was obtained from Beyotime Institute of Biotechnology. The cells were maintained in 96-well plates. Subsequently, AGS and MKN45 cells were incubated with 10 μL CCK-8 solution (Beyotime Institute of Biotechnology) at 37°C for 2 h following treatment. Subsequently, the viability of AGS and MKN45 cells was detected at 0, 12, 24, and 72 h using a microplate reader (Bio-Rad Laboratories, Inc., Hercules, CA, USA).
An EdU detection kit was obtained from Guangzhou RiboBio Co., Ltd. Following treatment, AGS and MKN45 cells were labeled with 100 μL 5-Ethynyl-2’- deoxyuridine (50 μM) for 1 h. The cells were washed with PBS twice for 5 min and incubated with 100 μL DAPI (1 μg/mL). Subsequently, the proliferation of AGS and MKN45 cells was detected using a fluorescence microscope (BX51TF, Olympus Corporation). Meanwhile, Apollo 567 was used to label the fluorochrome. The filter sets were as follows: blue excitation: 420~485 nm; red excitation: 420~490 nm.
DMEM with 10% serum was added to the basolateral chamber. Moreover, AGS and MKN45 cells were cultured in the apical chamber with 200 μL serum-free DMEM. Following incubation for 24 h, the medium in the basolateral chamber was removed and the cells were removed with PBS. Subsequently, the cells were fixed with paraformaldehyde for 20 min and stained with 0.1% crystal violet for 10 min. Finally, the migrated cells were detected using an optical microscope (BX53, Olympus Corporation).
AGS and MKN45 cells were collected using 0.25% trypsin. Subsequently, they were fixed with 70% ethanol and treated with propidium iodide (2 μg/mL)/RNase (10 mg/mL) staining buffer for 15 min in the dark at 4°C. The distribution of the cells was analyzed using a flow cytometer (FACScan™, BD Biosciences, Franklin Lakes, NJ, USA). The percentage of cell cycle phases was quantified using FlowJo software (FlowJo LLC, Ashland, OR, USA). Three measurement series per sample were recorded, and three replicates were performed in each group.
hsa_circ_0006470 wild-type/mutated-type (WT/MT) or TP53I11 (WT/MT) were synthesized by Sangon Biotech (Shanghai, China). The synthetic sequences were inserted into the pmiR-RB-REPORTTM vector. Subsequently, AGS cells were transfected with the recombinant vector of WT/MT containing miR- 1234 mimics or miR-1234 mimics-ctrl using LipofectamineTM 2000. The luciferase activity of AGS cells was assessed using a dual-luciferase reporter assay system.
AGS cells were lysed and the total protein was extracted using a protein lysis buffer. Subsequently, the protein concentration was determined by a bicinchoninic acid protein assay kit (Beyotime Institute of Biotechnology). The protein samples were separated by PAGE using 10% SDS gels (30 μg per lane). Subsequently, the proteins were transferred to polyvinylidene fluoride membranes (Beyotime Institute of Biotechnology). The following primary antibodies were used: TP53I11 (1:1,000, ab234860), Akt (1:1,000, ab8805), phosphorylated (p)-Akt (1:1,000, ab38449), forkhead box O (FOXO) 1 (1:1,000, ab179450), cyclin-dependent kinase (CDK) 2 (1:1,000, ab32147), β-actin (1:1,000, ab8226). A horseradish peroxidase-labeled secondary antibody (1:5,000, ab7090) was also used for western blotting. All antibodies were obtained from Abcam (Cambridge, UK). Finally, an ECL kit was employed to detect the protein expression. The expression levels of these proteins were normalized to those of β-actin.
The data were analyzed by GraphPad Prism. All data are presented as the mean ± standard deviation. The differences were assessed by one-way analysis of variance (ANOVA) and Tukey’s tests. A p-value <0.05 was considered to indicate a statistically significant difference.
It has been reported that the expression hsa_circ_0006470 is downregulated in GC. To confirm the role of hsa_circ_0006470 in GC cells, RT-qPCR was performed. As shown in Figure 1A, the levels of hsa_circ_0006470 in GC cells were lower compared with those of GES-1 cells. Since the expression levels of hsa_circ_0006470 were lower in AGS and MKN45 cells compared with those of SNU5 cells, these two cell lines were selected for subsequent experiments (Figure 1A). The results of FISH analysis indicated that hsa_circ_0006470 was mainly localized in the cytoplasm of GC cells (Figure 1B). Furthermore, RNase R did not affect the expression levels of hsa_circ_0006470 in AGS and MKN45 cells, while it significantly inhibited the expression levels of the linear mRNA molecule (Figure 1C). These data suggested that hsa_circ_0006470 exhibited a closed cyclic structure. Moreover, the data suggested that hsa_circ_0006470 expression was downregulated in AGS cells.
To explore the function of hsa_circ_0006470 in GC cells, they were transfected with a hsa_circ_0006470 OE vector. The results indicated that hsa_circ_0006470 OE caused a marked upregulation in the expression levels of hsa_circ_0006470 in GC cells (Figure 2A). In addition, hsa_circ_0006470 OE suppressed the viability of GC cells (Figure 2B). hsa_circ_0006470 upregulation inhibited the proliferation of GC cells (Figure 2 C,D). In summary, OE of hsa_circ_0006470 notably inhibited the proliferation of GC cells.
In order to explore the effects of hsa_circ_0006470 on the migration of GC cells, a Transwell assay was performed. As shown in Figure 3 A,B, pcDNA3.1-hsa_circ_0006470 decreased the migration of GC cells. OE of hsa_circ_0006470 significantly inhibited the migration of GC cells.
In order to evaluate the effects of hsa_circ_0006470 on the cell cycle distribution, flow cytometry was performed. As demonstrated in Figure 4 A,B, hsa_circ_0006470 OE caused a dramatic induction of G0-G1 phase arrest in GC cells.
In order to identify the downstream miRs of hsa_circ_0006470 involved in the development of GC, the circular RNA interactome (https://circinteractome.nia.nih.gov/api/v2/mirnasearch?circular_r na_query=hsa_circ_0006470&mirna_query=hsa-miR-1234&submit= miRNA+Target+Search) was used. The results indicated that hsa_circ_0006470 could bind to miR-1234 in GC cells (Figure 5A), and miR-1234 was shown to play a vital part in the oncogenesis of GC. Therefore, miR-1234 was selected in the present study. In addition, miR-1234 mimics significantly decreased the luciferase activity of WT-hsa_circ_0006470, while it did not affect the luciferase activity in MT-hsa_circ_0006470 (Figure 5B). hsa_circ_0006470 was found to be co-localized with miR-1234 in AGS cells (Figure 5C). To further explore the target of miR-1234, TargetScan was used. It was predicted that TP53I11 may be the downstream mRNA of miR-1234 (Figure 5D); moreover, TP53I11 has been shown to be a vital mediator of cell cycle progression. Therefore, TP53I11 was selected in the current study. In addition, miR-1234 mimics notably inhibited the luciferase activity of WT-TP53I11, whereas it had very limited effects on the luciferase activity of MTTP53I11 (Figure 5E). Moreover, hsa_circ_0006470 OE significantly increased the levels of TP53I11 in AGS cells (Figure 5 F,G). Taken together, the data indicated that hsa_circ_0006470 was bound to miR-1234 and TP53I11 was directly targeted by miR- 1234.
To further investigate the mechanism by which hsa_circ_0006470 regulates the tumorigenesis of GC, GC cells were treated with TP53I11 siRNA. The transfection efficiency was investigated. Transfection of the cells with TP53I11 siRNA notably decreased the expression levels of TP53I11 in AGS cells (Figure 6A). Since GC cells were more sensitive to the effects of TP53I11 siRNA1, the latter was selected for subsequent analysis (Figure 6A). In addition, hsa_circ_0006470 OE notably inhibited the viability of AGS cells, while this phenomenon was reversed by TP53I11 siRNA1 (Figure 6B). Moreover, upregulation of hsa_circ_0006470 expression significantly increased the expression levels of TP53I11 and FOXO1 and inhibited the expression levels of p-Akt and CDK2 in AGS cells; however, these phenomena were notably reversed in the presence of TP53I11 siRNA1 (Figure 6 C-G). In conclusion, OE of hsa_circ_0006470 notably suppressed the viability of GC cells by regulating the miR- 1234/TP53I11 axis.
GC seriously affects the physical and mental health of patients. The expression levels of specific circRNAs are closely associated with the neoplastic transformation of GC cells. In addition, it has been reported that hsa_circ_0006470 is involved in the progression of GC. For example, Cui et al found that hsa_circ_0006470 expression was significantly upregulated in GC cells and that it could regulate the proliferation, migration, and invasion of GC cells by binding to miR-27b-3p; Yao et al. indicated that the downregulation of the expression of hsa_circ_0006470 could predict tumor invasion of GC. In the current study, the data indicated that hsa_circ_0006470 OE could inhibit the malignant phenotype of GC cells and that hsa_circ_0006470 could bind to miR-1234. The current study explored the interaction between hsa_circ_0006470 and miR-1234 with regard to GC progression. Therefore, the data presented may provide novel evidence that can be used to investigate further the mechanisms underlying the function of circRNAs in GC. circRNAs can bind to certain miRs during the progression of GC. For example, OE of circ_SH3KBP1 binding protein 1 promoted the angiogenesis of GC cells by sponging miR-582-3p. Moreover, OE of circ-coiled-coil domain containing 9 suppressed the proliferation of GC cells by targeting miR-6792-3p. Similarly, miR-1234 was targeted by hsa_circ_0006470 in the present study. miR-6792-3p was verified to be a suppressor of cancer progression. Therefore, the similar function between miR-1234 and miR- 6792-3p in cancer may be used to explain the similar data presented between the current and previous research studies. In addition, OE of hsa_circ_0006470 induced cell cycle arrest of GC. Wang et al. indicated that TP53I11 is a cell cycle-related protein. Therefore, it may be suggested that OE of hsa_circ_0006470 inhibits the proliferation of GC cells by regulating the miR- 1234/TP53I11 axis. In addition, downregulation of TP53I11 expression promotes the AKT pathway in MCF10A cells. In the current study, upregulation of hsa_circ_0006470 expression significantly increased the expression levels of TP53I11, FOXO and inhibited the expression levels of p-Akt and CDK2 in AGS cells. The data reported in the present study were similar to those of previous reports, suggesting that TP53I11 could regulate the expression of AKT. Based on the aforementioned evidence, the novelty of the current study was as follows: i) The interaction between hsa_circ_0006470 and miR-1234 with regard to the progression of GC was explored for the first time; ii) the downstream mRNA of miR-1234 was investigated for the first time. In conclusion, OE of hsa_circ_0006470 notably inhibited the tumorigenesis of GC cells by regulating the miR-1234/TP53I11 axis. The findings may shed light on the discovery of novel therapeutic strategies against GC. | true | true | true |
Subsets and Splits