miR‐3133 inhibits gastrointestinal cancer progression through activation of Hippo and p53 signalling pathways via multi‐targets

Abstract Background Malignant cell growth and chemoresistance, the main obstacles in treating gastrointestinal cancer (GIC), rely on the Hippo and p53 signalling pathways. However, the upstream regulatory mechanisms of these pathways remain complex and poorly understood. Methods Immunohistochemistry (IHC), western blot and RT‐qPCR were used to analyse the expression of RNF146, miR‐3133 and key components of Hippo and p53 pathway. CCK‐8, colony formation, drug sensitivity assays and murine xenograft models were used to investigate the effect of RNF146 and miR‐3133 in GIC. Further exploration of the upstream regulatory mechanism was performed using bioinformatics analysis, dual‐luciferase reporter gene, immunoprecipitation assays and bisulfite sequencing PCR (BSP). Results Clinical samples, in vitro and in vivo experiments demonstrated that RNF146 exerts oncogenic effects in GIC by regulating the Hippo pathway. Bioinformatics analysis identified a novel miRNA, miR‐3133, as an upstream regulatory factor of RNF146. fluorescence in situ hybridization and RT‐qPCR assays revealed that miR‐3133 was less expressed in gastrointestinal tumour tissues and was associated with adverse pathological features. Functional assays and animal models showed that miR‐3133 promoted the proliferation and chemotherapy sensitivity of GIC cells. miR‐3133 affected YAP1 protein expression by targeting RNF146, AGK and CUL4A, thus activating the Hippo pathway. miR‐3133 inhibited p53 protein degradation and extended p53's half‐life by targeting USP15, SPIN1. BSP experiments confirmed that miR‐3133 promoter methylation is an important reason for its low expression. Conclusion miR‐3133 inhibits GIC progression by activating the Hippo and p53 signalling pathways via multi‐targets, including RNF146, thereby providing prognostic factors and valuable potential therapeutic targets for GIC.


| INTRODUC TI ON
Gastric cancer and colorectal cancer are among the most prevalent and deadliest gastrointestinal cancer (GIC), contributing to substantial morbidity and mortality rates. 1 The complex pathogenesis of GIC involves intricate interactions between genetic, epigenetic and environmental factors, leading to the dysregulation of multiple signalling pathways and cellular processes. 2,3Understanding the molecular mechanisms underlying GIC is paramount for the development of effective diagnostic tools, prognostic markers and targeted therapies.
The Hippo and p53 signalling pathways play an important roles in physiological and pathological processes. 4,5The Hippo signalling pathway is highly conserved in mammals and regulates organ size and tumorigenesis. 6,7Its key components include a kinase cascade, mammalian ste20-like protein kinase 1/2 (MST1/2), a large tumour suppressor kinase 1/2 (LATS1/2) and two transcription coactivators, Yes-associated protein (YAP) and Tafazzin (TAZ). 8When the Hippo signalling pathway is activated, MST1/2-mediated phosphorylation and activation of LATS1/2 further phosphorylates and inactivates YAP/TAZ, causing its retention in the cytoplasm and degradation.In contrast, inactivation of the Hippo pathway leads to YAP/TAZ translocation to the nucleus and combination with transcription factors, such as TEADs.This promotes tumorigenesis via the activation of target genes, such as CTGF and CYR61. 9,10p53, an anticancer gene, causes cell cycle arrest, apoptosis and senescence by manipulating downstream targets. 11,12gradation of the p53 protein leads to the inactivation of the p53 signalling pathway and tumour development. 13,14MDM2 is the most widely studied ubiquitination mechanism of p53. 15As an E3 ubiquitin ligase, MDM2 targets p53 for proteasome degradation. 16In addition, it inhibits p53 transcription factor function by promoting the nuclear export of p53. 179][20] Despite a considerable amount of research effort, the upstream regulatory mechanisms of the p53 and Hippo signalling pathways require further investigation.
As an E3 ubiquitin ligase, RING finger protein 146 (RNF146) has a PAR-binding motif in the Trp-Trp-Glu (WWE) and RING domains, which can recognize poly (ADP-ribose) (PAR) and target them for proteasomal degradation. 21RNF146 is involved in several physiological and pathological processes such as bone dynamics, energy metabolism and oxidative stress. 22,23[26] However, the role of RNF146 in GIC, its interactions with other pathways, and upstream regulatory mechanisms remain unclear.
By simultaneously repressing a variety of target genes through binding to their 3′-untranslated regions (3′-UTRs), microRNAs (miR-NAs) can potentially affect multiple steps in cancer progression and tumorigenesis. 27MicroRNA-3133 (miR-3133), a novel miRNA, is located at the common fragile site (CFSs) and is defined as a cancer-related cytogenetic miRNA. 28Thus far, only five studies have reported the pathophysiological effects of miR-3133, and only two of which are tumour-related.These included retinoblastoma and clear cell renal cell carcinoma, in which miR-3133 plays a suppressive role in tumour progression.As such, the role of miR-3133 in tumours requires further investigation.
In this study, we primarily explored the upstream regulatory mechanisms of the Hippo and p53 pathways.First, we performed a series of experiments that analysed the role of RNF146 in GIC and its effect on the Hippo pathway.Second, we investigated the effect of miR-3133 on multiple target molecules including RNF146 and the role of miR-3133 in regulating the Hippo and p53 pathways.Finally, the expression and epigenetic status of miR-3133 in GIC were explored.Our findings aim to clarify whether miR-3133 can act as a promising prognostic factor and valuable therapeutic target for GIC.

| Tissue collection and ethics statement
A total of 160 formalin-fixed paraffin-embedded cancer samples, along with complete clinical and pathological data, were collected from gastric cancer (GC) patients who underwent surgical resection at the First Affiliated Hospital of Nanchang University, Nanchang, China, between 2016 and 2018.The clinicopathological features of the patients are provided in Table 1.Another 20 pairs of fresh GC tissues and 42 pairs of fresh colorectal cancer (CRC) tissues and their matched noncancerous tissues were obtained from the general surgery room between 2018 and 2021, immediately frozen, and stored in liquid nitrogen for subsequent experimentation.The protocols used in this study were approved by the ethics committee of the First Affiliated Hospital of Nanchang University.Written informed consent was obtained from all the patients.

| RNA extraction and RT-qPCR analyses
Total RNA was extracted from GC tissues, CRC tissues and paired adjacent noncancerous tissues.Further complementary DNA (cDNA) was obtained and Real-time PCR was performed as previous described. 29GAPDH served as an internal control.Relative mRNA expression of the indicated genes was calculated using the 2 −ΔΔCT method.The PCR primer sequences used are listed in Table S1.

| Western blotting
Protein samples were lysed in ice-cold radioimmunoprecipitation assay buffer (Beyotime, Shanghai, China).The lysates were separated using 10% SDS-PAGE and then transferred onto a nitrocellulose membrane (Millipore).The nitrocellulose membrane was blocked and incubated with the corresponding primary antibodies at 4°C overnight.
Afterwards, the membrane was washed thrice with TBST and incubated with goat anti-rabbit/mouse IgG (H + L) HRP secondary antibody (ZB-2305; ZB-2301; Zhongshan Golden Bridge, Beijing, China) at room temperature for 1 h.Finally, the proteins of interest were detected using an ECL reagent (WBULS0500; Merck Millipore, Darmstadt, Germany).The primary antibodies used are listed in Table S2.

| CCK-8 assay and colony formation assay
To assess cell viability, post-transfected cells were seeded in 96-well plates at 10 3 cells per well and cultured for indicated days.The CCK-8 assay was performed according to the manufacturer's instructions.Optical density values were measured at 450 nm absorbance using a microplate reader (Molecular Device, SpecrtraMax M5e).For the colony formation assay, GC or CRC cells (1000 per well) were seeded in 6-well plates 36 h post-transfection and cultured in RPMI -1640 medium containing 10% fetal bovine serum for 2 weeks.The cells were then fixed with 75% ethanol and stained with a 0.1% crystal violet solution.Only colonies with more than 50 cells per colony were counted.Experiments were performed in triplicate.

| Dual-luciferase reporter gene assay
Both the wild-type 3′-UTR sequence, containing an established complementary region, and the mutant 3′-UTR sequence, lacking the mutant complementary region, of the indicated genes were cloned
Subsequently, the levels of various proteins were analysed by western blot analysis and quantified using ImageJ software.

| Immunoprecipitation assays
An immunoprecipitation assay was performed to detect p53 ubiquitination.HCT116 p53 −/− cells were seeded into 10 cm dishes and transfected with miR-3133 inhibitors, miR-3133 mimics, HA-MDM2, His-Ub or p53 plasmid (10 μg each) for specific needs.The plasmids encoding HA-MDM2, p53 and His-Ub have been previously described. 20After 36 h, the cells were treated with MG132 (10 μM) for 6 h to inhibit proteasomal degradation and then harvested for cell extracts.Cell lysates were then immunoprecipitated with anti-p53 antibodies and protein A/G beads (#22202-20, Beaver Nano-Technologies) at 4°C overnight.The levels of ubiquitination of p53 were determined using an anti-Ub antibody by immunoblotting.

| Bisulfite sequencing PCR
The methylation status of miR-3133 promoter CpG islands in tissues was analysed using bisulfite sequencing PCR (BSP).According to the manufacturer's instructions, DNA was extracted from GC and adjacent tissues using a DNA extraction kit (Tiangen Biotech, DP304) and treated with sodium bisulfite using the Methylation-Gold Kit (D5005S (10), ZymoResearch).The treated DNA was then amplified by PCR using appropriate MSP primers.The primers were designed using MethPrimer (http://www.urogene.org/methp rimer), and the sequences were as follows: F: TTGTT ATT AGG TTG GGG TGT AGTGT and R: ACATA AAT CCA CAC AAA ACT TAT ACATAAAC.The PCR conditions were as follows: 95°C for 5 min; 34 cycles of 95°C for 30 s, 55°C for 30 s and 72°C for 30 s, with a final extension step at 72°C for 5 min.The PCR-amplified products were ligated to T-Vector PMD19 (6013, Takara Biotechnology, Dalian, China) and transformed into Escherichia coli DH5α cells.LB agar plates containing kanamycin (50 μg/mL) were used to select qualified colonies at 37°C temperature overnight.Colonies were selected and cultured in LB amp+ medium overnight, and the obtained plasmids were sequenced using ABI 3730XL (ABI).

| Immunohistochemistry staining and immunofluorescence
Immunohistochemical staining was performed as previously described. 30Staining results were evaluated by two pathologists based on the proportion of positively stained cells and staining intensity.

| RNA fluorescence in situ hybridization
Fresh GIC tissues were fixed with paraformaldehyde immediately after excision, and then 4 μm sections were prepared after dehydration and paraffin embedding.The sections were then dewaxed, dehydrated, digested with proteinase K, prehybridized and examined with an oligonucleotide-modified probe sequence (ATTGG GTT TTA AGA GTT CTTTA) for human miR-3133.

| Bioinformatics website
We assessed genomic and mRNA changes involving RNF146 in human cancers using the online database Cbioportal (http://www.cbiop ortal.org).Survival analysis of patients was performed using Kaplan-Meier plotter website data (https://kmplot.com/analysis/).

| Statistical analysis
All experiments were performed independently thrice, and all data are presented as mean ± SEM.Statistical analysis was performed using SPSS v.22.0 (SPSS, Chicago, IL, USA).Quantitative data are presented as the mean ± standard deviation.Student's t-test or chisquare test was used to analyse the statistical differences between the two groups.Survival curves were plotted using the Kaplan-Meier method and analysed using the log-rank test.Statistical significance was set at p < 0.05.

| RNF146 acts as an oncogene in GIC
To explore the role and expression of RNF146 in GIC, a series of biochemical experiments was conducted.Previous studies have shown that RNF146 is highly expressed and associated with poor prognosis in CRC. 26 Here, we focused on exploring the role of RNF146 in GC.Western blot assays showed that RNF146 protein expression increased in human GC tissue samples paired with adjacent non-tumour tissues (5/8, 62.5%, Figure 1A,B) as well as in most GC cell lines compared to that in the gastric normal epithelial mucosa cell line GES-1 (Figure 1C).In line with the above results, immunohistochemical staining of RNF146 was deeper in tumour tissues than in normal tissues and was positively correlated with the TNM stage (Figure S1A).
In addition, the relationship between the expression of RNF146 and the clinicopathological picture of patients was explored.As shown in   32 To verify this, we conducted RT-qPCR and western blot assays in xenograft tumours and GC cells.The results revealed that the expression of YAP1 and its downstream target genes was markedly decreased in RNF146-knockdown HGC-27 cells (Figure 1H; Figure S1F) and xenograft tumours (Figure 1L).Additionally, the knockdown efficiency of RNF146 in the xenograft tumours was further confirmed after the excision of the tumour mass (Figure S1G).Collectively, these data support that RNF146 acts as an oncogene in GIC.

| RNF146 is a direct target of miR-3133
To explore the upstream regulatory mechanism of RNF146, we first assessed genomic changes involving RNF146 in human cancers using the online database, Cbioportal (https://www.cbioportal.org).The results demonstrated that RNF146 genetic alterations were not significant in GC and CRC (Figure 2A).However, data from eight pairs of fresh GC samples and 42 pairs of fresh CRC samples showed that the expression of RNF146 mRNA was higher in GIC tissues than in adjacent noncancerous tissues (Figure 2B,C), suggesting that overexpression of RNF146 protein may be regulated at the post-transcriptional level in GIC.Furthermore, based on bioinformatic analysis tools (TargetScan; www.targe tscan.organd miRanda; www.micro rna. org), we speculated that miR-3133 might be a putative miRNA that regulates RNF146 (Figure 2D).KEGG pathway analyses revealed that miR-3133 may have a vital function in pathways involved in cancer, Hippo signalling, etc. (Figure 2E).Fluorescence in situ hybridization (FISH) assays were performed to verify the expression of miR-3133 in tumour tissues, and the results showed that the expression was decreased in the cytoplasm of fresh GIC tissues (Figure 2F).Next, we aimed to clarify whether miR-3133 and RNF146 mRNA can bind to each other.After transfection of miR-3133 mimics and inhibitors into HGC-27 and HCT-116 cells, we found that miR-3133 mimics markedly reduced the mRNA and protein expression of RNF146, whereas opposite trends were observed after transfection with miR-3133 inhibitor (Figure 2G,H).In addition, wild-type (WT) and mutant (Mut) miR-3133 were constructed (Figure 2D) and dual-luciferase assays were performed.miR-3133 mimics decreased, but miR-3133 inhibitors remarkably enhanced the luciferase activity of RNF146 in both transfected HGC-27 and HCT-116 cells (Figure 2I,J).Taken together, our data elucidated that RNF146 is a target of miR-3133 in GIC.

| miR-3133 inhibits proliferation and chemoresistance in GIC
As miR-3133 was found to be a negative regulator of RNF146, we hypothesized that miR-3133 may act as a tumour-suppressive miRNA in tumorigenesis and progression.To test this hypothesis, HGC-27 and  | 3099 recorded tumour formation and growth for 31 days.As expected, the weight of xenografts from lentiviral-miR-3133 mice grew at a slower rate compared to that in the control group (Figure 3J-L).Collectively, these data indicate that miR-3133 suppresses the proliferation and chemoresistance of cancer cells.

| miR-3133 activates the Hippo-YAP1 pathway
As our data showed that RNF146 affects the expression of YAP1 and downstream target genes, we speculated that miR-3133 could affect the Hippo-YAP1 pathway by targeting RNF146.As expected, overexpression of miR-3133 reduced RNF146 and total YAP1 protein levels but increased p-YAP protein levels in HGC-27 and HCT-116 cells (Figure 4A), and two related target genes CTGF and CYR61 were concomitantly decreased (Figure 4A,D).Immunofluorescence results confirmed that YAP1 was reduced as well as increased after transfection with miR-3133 mimics and inhibitors, respectively (Figure 4B,C; Figure S2A).Taken together, miR-3133 activated the Hippo pathway in GC and CRC cells.
Notably, the restoration of RNF146 only partially abrogated miR-3133-induced suppression of GC cell proliferation (Figure S2B), suggesting that, in addition to RNF146, other targets of miR-3133, to date unidentified, may also be involved.Based on the bioinformatic analysis, we found that certain Hippo pathway-related genes, such as CUL4A, AGK, CD44, LATS2 and TEAD1 are also potential targets of miR-3133 (Figure 4E).To further validate these results, RT-qPCR and western blotting assays were performed, and the results showed that upregulation of miR-3133 reduced the protein and mRNA expression levels of CUL4A, AGK and TEAD1, but not those of CD44 and LATS2 (Figure 4F,G).The luciferase reporter assay results indicated that upregulation of miR-3133 significantly repressed the luciferase activity of vectors carrying the WT 3'-UTR of CUL4A, AGK and TEAD1, but not the mutant (Figure 4H).Overall, these results indicated that miR-3133 activates the Hippo pathway by targeting RNF146, CUL4A, AGK and TEAD1.

| miR-3133 activates the p53 signalling pathway
The above data suggested that CUL4A was regulated by miR-3133; however, a previous study showed that CUL4A physically associates with MDM2 and participates in MDM2-mediated proteolysis of p53. 19Therefore, we conducted a series of experiments to explore whether miR-3133 regulates the p53 signalling pathway.Interestingly, as shown in Figure 5A,B, upregulation of miR-3133 dramatically elevated p53 protein levels in AGS (expressing wild-type p53) and HCT-116 (p53 +/+ ) cells without affecting mRNA levels.Both protein and mRNA levels of p53 target genes such as p21 and PUMA were increased in miR-3133-overexpressing cells.In contrast, opposite effects were observed in HCT116 p53 +/+ cells after transfection with the miR-3133 inhibitor (Figure 5C).To further clarify the effect of miR-3133 on p53, HCT116 p53 +/+ cells were treated with cycloheximide (CHX) and harvested in a time-dependent manner.A dramatic increase in the stability and longer half-life of p53 was observed in miR-3133-overexpressing cells, whereas a shorter half-life and decreased p53 activity were observed in miR-3133-knockdown cells (Figure 5D).In addition, we performed ubiquitination assays to investigate whether miR-3133 regulates p53 ubiquitination in HCT116 p53 −/− cells.As expected, MDM2 induced p53 ubiquitination, which was relieved by miR-3133 mimics or further enhanced by the miR-3133 inhibitor (Figure 5E).These results suggest that miR-3133 stabilizes the p53 protein by inhibiting MDM2-mediated ubiquitination.
Notably, bioinformatics analysis results showed that, in addition to CUL4A, miR-3133 could also bind USP15, MDM2, MDM4 and SPIN1 (Figure S2C), which have previously been reported to be involved in the regulation of the p53 pathway. 18,20,33We then investigated the expression of these genes in miR-3133-upregulated or -downregulated cells by western blotting and RT-qPCR and found that only the expression of USP15 and SPIN1 was reduced in miR-3133-transfected cells but increased in miR-3133 antagomirtransfected cells (Figure 5F,G).Luciferase reporter assays further demonstrated that miR-3133 mimics markedly decreased the luciferase activity of SPIN1 and USP15 3′-UTR plasmids in AGS and HCT-116 cells, but not that of SPIN1 and USP15 Mut-3′-UTR (Figure 5H).

| Low expression of miR-3133 induced by promoter methylation is associated with poor prognosis in patients
The above data indicated that miR-3133 was expressed at lower levels in tumour tissues (Figure 2F).To further verify these conclusions,  was significantly decreased in 13/20 (65%) GC and 26/42(61.9%)CRC tissues.In addition, the clinicopathological characteristics of the 42 patients were collected to explore their relationship with miR-3133 expression.As shown in Table 2, miR-3133 expression was strongly correlated with the depth of tumour size (p = 0.023), and Kaplan-Meier analysis (https://kmplot.com/)indicated that GC patients with low miR-3133 expression levels had worse overall survival than those with high miR-3133 expression levels, a phenomenon not seen in CRC patients (Figure 6C).Hypermethylation of DNA-induced epigenetic silencing of miRNAs is the primary mechanism for cancer progression. 34Therefore, we designed specific primers to conduct the BSP assay using six pairs of GC tissues.The results showed that the methylation level of CpG sites in the miR-3133 promoter was higher in tumour tissues than in adjacent normal tissues (97.433% ± 0.2319% vs. 92.9333%± 0.2319%, p < 0.001) (Figure 6E; Figure S2D).AGS cells were treated with 5-aza-dC (5-AZA, a DNA methyltransferase inhibitor) to verify whether miR-3133 expression is regulated by DNA methylation.As shown in Figure 6D, miR-3133 expression was significantly upregulated after 5-Aza treatment.Its target molecules, RNF146, AGK, CUL4A, TEAD1, USP15 and SPIN1, decreased concomitantly, and the key effectors of the Hippo pathway, CYR61, and the p53 downstream targets, p21 and PUMA, were altered accordingly.Collectively, our data indicated that low expression of miR-3133 is associated with tumour progression, and promoter methylation was identified as the cause of its reduced expression.

| miR-3133 affects cancer progression through multi-targets
The above data indicate that miR-3133 regulates the Hippo and p53 signalling pathways through multiple targets in GC and CRC cells.To further verify the relationship between miR-3133 and these targets, we used immunohistochemistry to detect the expression of target molecules in paraffin-embedded samples from 20 patients with GC and 42 patients with CRC.As shown in Figure 7A, in the tissues with low expression of miR-3133, the related target molecules were highly expressed.We then evaluated the IHC staining scores of the related target molecules and analysed their correlation with miR-3133.The results showed that the expression of miR-3133 was negatively correlated with SPIN1 (p = 0.008), TEAD1 (p = 0.016) and RNF146 (p = 0.044) (Table 3).Furthermore, the prognostic roles of RNF146, CUL4A, AGK, TEAD1, USP15 and SPIN1 in GC were evaluated using the KM plotter website (Figure 7B).The results showed that high expression of RNF146, TEAD1, YAP1 and SPIN1 indicated poor prognosis in GC and CRC.Overall, these results indicate that miR-3133 affects cancer progression through multiple targets.

| DISCUSS ION
Our study showed that the expression of RNF146 was elevated in GC and that its overexpression was associated with worse survival outcomes and more adverse clinicopathological features of GC.
Depletion of RNF146 reduced the proliferative ability of GIC cells by activating the Hippo pathway in vivo and in vitro.We also found that miR-3133, as an upstream regulator of RNF146, plays a tumoursuppressive role in GIC.miR-3133 was poorly expressed in GC and CRC tissues and negatively correlated with patient clinical features.
Mechanistically, we found that miR-3133 negatively regulates multiple targets, including RNF146, AGK, CUL4A, SPIN1, USP15 and TEAD1, thereby affecting the activity of the Hippo and p53 signalling pathways.However, miR-3133 is epigenetically regulated via DNA promoter methylation.Thus, this study not only provides novel insights into the upstream regulatory network of the Hippo and p53 signalling pathways but also provides the rationale for miR-3133 as a promising therapeutic target for GIC.
Abnormalities in signalling pathways, the hallmark of cancer, are frequently caused by the dysregulation of one or more molecules in these pathways. 35Uncontrolled cell proliferation is the main obstacle to cancer treatment, and both the Hippo and p53 signalling pathways are essential for cancer cell growth and chemoresistance. 33,36ing bioinformatics prediction, we found that miR-3133 can target and regulate the expression of some proteins that are known to affect the Hippo and p53 pathways as previous studies.For instance, CUL4A mediates the ubiquitination of LATS1 to regulate YAP activity in hepatocellular carcinoma, 37 AGK inhibits the activation of the Hippo pathway proteins in GC 38 and CD44 functions upstream of  the Hippo signalling pathway and attenuates its activation in glioblastoma. 39Some proteins, such as MDM2, MDM4 and TEAD1, are the key components of the Hippo and p53 signalling pathways.
Based on bioinformatic predictions, we further screened the target genes regulated by miR-3133 through experimental verification.Our data suggest that the upregulation of miR-3133 can downregulate these target genes and lead to the inactivation of YAP and decreased ubiquitination-mediated p53 degradation, thus changing the expression of downstream target genes and reducing the proliferation and drug resistance of tumour cells.This pattern of multiple targets coregulating signalling pathways provides a novel understanding of miRNA regulating signal transduction.
RNF146 is a target gene of miR-3133.Several studies have shown that RNF146 can promote the progression of various cancers by activating the Wnt signalling pathway. 25,26,40In addition, Wang Recently, the role of miRNAs in tumorigenesis has been validated by numerous studies.miRNAs may modulate tumorigenesis, proliferation, apoptosis and chemoresistance of cancers through multiple signalling pathways. 43,44In this study, we found that a newly discovered miRNA, named miR-3133, is involved in the proliferation of GIC cells.To date, the role of miR-3133 in cancer has been poorly investigated, except in retinoblastoma and clear cell renal cell carcinoma. 45,46In line with the above findings, our data indicate that miR-3133 plays a tumour-suppressive role in GIC.Previous studies have shown that miR-3133 is expressed at low levels in tumour tissues, but the mechanism of its downregulation remained unclear. 45,46Epigenetic mechanisms can alter gene expression patterns and genomic stability without altering DNA sequences, including DNA methylation, histone modification, chromatin remodelling and altered expression of noncoding RNAs. 47DNA methylation is an important part of the epigenetic regulatory mechanism, which refers to the process by which the C-5 position of the cytosine of a nucleotide is covalently modified to 5-methylation cytosine (5mC) by a methyl supplied by Sadenosine methionine through DNA methyltransferase. 48

TA B L E 3
The correlation of miR-3133 and its targets.

2. 9 |
In vivo animal studyFemale BALB/c mice (4 weeks old) were obtained from Zhejiang Weitong Lihua Laboratory Animal Technology Co., Ltd.(Zhejiang, China) and randomly grouped.5 × 10 6 GC cells stably transfected with lentivirus shRNF146-1, lentivirus miR-3133 or their controls were injected subcutaneously into the right armpit of nude mice (n = 6 per group).The tumour volumes were determined at the indicated time points (volume [mm 3 ] = [length × width 2 ]/2).Approximately 1 month after injection, nude mice were euthanized, and tumours were peeled off, weighed and photographed.This xenograft study was approved by the Ethics Committee of the First Affiliated Hospital of Nanchang University.
unfavourable clinical outcomes and poor survival of patients.The oncogenic effect of RNF146 was further investigated in vivo and in vitro.We transfected Scramble shRNA and RNF146 shRNA into HGC-27 and HCT-116 cells, respectively, and verified the transfection efficiency using immunoblotting and RT-qPCR (Figure S1D,E).As expected, both the CCK-8 and colony formation assays indicated that the downregulation of RNF146 suppressed cell proliferation in HGC-27 and HCT-116 cells (Figure 1D-G).In vivo experiments were performed by injecting RNF146-knockdown HGC-27 cells into the armpit of BALB/C-nude mice, and the xenograft tumours were measured for 28 days.The results showed that the average xenograft tumour weight and tumour volume derived from HGC-27 cells transduced with lentivirus shRNF146-1 were significantly decreased compared to those of the respective negative control (Figure 1I-K).Moreover, previous studies have shown that RNF146 is a critical upstream regulator of YAP1.

F I G U R E 1
RNF146 acts as an oncogene in gastrointestinal cancer (GIC).(A)and (B) Western blot images and quantification of RNF146 in gastric cancer (GC) tissues (T) and adjacent normal tissues.(C) The protein and mRNA levels of RNF146 in GC cells and normal epithelial mucosal cell lines GES-1 detected by western blotting and RT-qPCR.(D)-(G) Proliferation of GIC cells examined by CCK-8 assay and colony formation assay after transfection with RNF146 short hairpin RNAs (shRNAs) or the negative control.(H) YAP1 and its downstream molecule, CTGF, were detected after the downregulation of RNF146.(I)-(L) Xenograft tumour models derived from HGC-27 cells expressing LV-scramble shRNA (LV-NC) or LV-shRNF146.(I) Images of xenograft tumours harvested at the end of the experiments.(J) Growth curves of xenograft tumours.(K) Comparison of the average weights of tumours collected from the above two groups.(L) RT-qPCR of RNF146, CTGF and YAP1 in xenograft tumours (*p < 0.05, **p < 0.01 vs. corresponding control groups).

F I G U R E 2
RNF146 is a direct target of miR-3133.(A) The alteration frequency of RNF146 in various cancer tissues obtained from the cBioPortal for Cancer Genomics website (http://www.cbioportal.org/)derived from the TCGA database.(B) and (C) Relative RNF146 mRNA expression in eight pairs of fresh gastric cancer (GC) tissues and 42 pairs of fresh colorectal cancer (CRC) tissues.(D) Schematic diagram of the putative binding sites of miRNA-3133 in the RNF146 3′-UTR (red) and mutated binding sites of miRNA-3133 (blue).(E) KEGG pathway analysis of miRNA-3133 in miRPath (http://www.microrna.gr/miRPa thv3/).(F) Fluorescence in situ hybridization analysis of miRNA-3133 in GC and normal tissues, as well as in CRC and normal tissues.(G) and (H) RT-qPCR and Western blot analyses of RNF146 expression in HGC-27 and HCT-116 cells after transfection with miR-3133 mimics, miR-3133 inhibitor or negative control.(I) and (J) Luciferase activity of RNF146 3′-UTR in HGC-27 and HCT-116 cells transfected with wild-type miR-3133, mutant miR-3133, miR-3133 inhibitor or negative control (*p < 0.05, **p < 0.01 vs. corresponding control groups).HCT-116 cells were transfected with miR-3133 mimics, miR-3133 inhibitor or negative control.The results of CCK-8 assay revealed that the upregulation of miR-3133 suppressed the proliferative capacity of HGC-27 and HCT-116 cells, whereas the miR-3133 inhibitor showed the opposite effect (Figure 3A,B).In line with the above results, the colony formation assay and EdU assay further validated that the expression of miR-3133 truly affects the proliferation of cancer cells (Figure 3E-I).In addition, the IC 50 value of cancer cells decreased or increased significantly after transfection with miR-3133 mimics or miR-3133 inhibitors, respectively, indicating that miR-3133 affected the sensitivity of cancer cells to 5-FU treatment (Figure 3C,D).Furthermore, we determined the role of miR-3133 in the growth of GC xenografts in nude mice.We subcutaneously injected HGC-27 cells, with or without overexpressed miR-3133, into nude mice and F I G U R E 3 miR-3133 inhibits cell proliferation and chemoresistance in gastrointestinal cancer (GIC).(A)-(I) Cells were transfected with miR-3133 mimics, miR-3133 inhibitor and negative control and subjected to the CCK-8 assay (A) and (B), drug sensitivity experiments (C) and (D), colony formation assay (E) and (F) and EdU incorporation assay (G)-(I).(J)-(L) Xenograft tumour model derived from HGC-27 cells expressing LV-NC or LV-miR-3133.(J) Images of xenograft tumours harvested at the end of experiments; (K) Comparison of the average weight of collected tumours from the above two groups; (L) Growth curves of xenograft tumours (*p < 0.05, **p < 0.01 vs. corresponding control groups).F I G U R E 4 miR-3133 activates the Hippo-YAP1 pathway.(A) The levels of YAP1, p-YAP and their downstream target genes in HGC-27 cells and HCT-116 cells transfected with miR-3133 mimics were analysed using western blotting.(B) and (C) Immunofluorescence staining and quantitation of fluorescence intensity of YAP1 in HGC-27 cells after transfection with miR-3133 mimics and miR-3133 inhibitor.(D) RT-qPCR results for CTGF and CYR61 in HGC-27 and HCT-116 cells transfected with miR-3133 mimics.(E) Predicted miR-3133 target sequences in the 3′-UTRs of CUL4A, AGK, CD44, LATS2 and TEAD1 by TargetScan website (www.targetscan.org).(F) and (G) Western blot results and RT-qPCR results of CUl4A, AGK, CD44, LATS2 and TEAD1 expression levels in HGC-27 and HCT-116 cells transfected with miR-3133 mimics or negative control.(H) Luciferase activity of CUL4A 3′-UTR, AGK 3′-UTR TEADs 3′-UTR and their mutants were determined in HGC-27 cells transfected with miR-3133 mimics or negative control (*p < 0.05, **p < 0.01 vs. corresponding control groups).
RT-qPCR was performed on 20 pairs of fresh GC tissues and 42 pairs of fresh CRC tissues.As shown in Figure6A,B, miR-3133 expression F I G U R E 5 miR-3133 activates the p53 signalling pathway.(A) and (B) Western blotting and RT-qPCR results of p53, p21 and PUMA expression levels in AGS and HCT-116 p53 +/+ cells transfected with miR-3133 mimics.(C) Western blot and RT-qPCR results of p53, p21 and PUMA expression levels in HCT116 p53 +/+ cells transfected with the miR-3133 inhibitor.(D) The half-life of p53 was detected after transfection with miR-3133 mimics or miR-3133 inhibitor, treated with 50 μg/mL of cycloheximide (CHX), and harvested at different time points, as indicated.(E) MDM2-induced p53 ubiquitination in HCT116 p53 −/− cells transfected with the indicated constructions for 36 h and treated with MG132 for 6 h before being harvested was examined in an in vivo ubiquitination assay.(F) and (G) Western blot and RT-qPCR results of USP15, MDM2, MDM4 and SPIN1 expression levels in AGS cells and HCT116 p53 +/+ transfected with miR-3133 mimics or negative control.(H) Luciferase activity of USP15 3′-UTR, SPIN1 3′-UTR and their mutants in AGS and HCT-116 cells transfected with miR-3133 mimics and negative control (*p < 0.05, ** p < 0.01 vs. corresponding control groups).
et al. found that RNF146 could mediate NF2 regulation of the Hippo pathway,41 whereas Cao et al. showed that RNF146 is involved in the formation of the ubiquitin complex (Slug/GSK3β/RNF146) and induction of metastasis of GC.42However, the expression of RNF146 in GC and its upstream and downstream regulatory mechanisms remain unclear.Here, we verified for the first time that the expression of RNF146 is upregulated in GC and that the downregulation of RNF146 leads to the activation of the Hippo pathway and reduction of the proliferative ability of GC and CRC cells.
In this paper, methylation sequencing was used to verify the high methylation level of CpG islands in the miR-3133 promoter.Methylation inhibitors were used to restore the expression of miR-3133 in tumour cells, confirming that promoter methylation of miR-3133 was the cause of its low expression.Therefore, we explain that miR-3133 is expressed at low levels in GICs due to methylation, further improving the multi-target and dual-pathway regulatory network centred on miR-3133.A limitation of our study was that CRC patients with high miR-3133 did not show benefit with OS outcomes (p = 0.21, N = 160, Figure 6C), which may be attributed to the small sample size and/or the rectal location of the tumour in most of these patients.In summary, our data indicate that miR-3133 inhibits GC and CRC cell proliferation by activating the Hippo and p53 signalling pathways via multiple targets.Methylation of the miR-3133 promoter in tumour tissues downregulated miR-3133 expression.Therefore, this research suggested that miR-3133 may function as a potential target in anti-GC and -CRC treatment and provided a basis for further comprehensive follow-up studies.AUTH O R CO NTR I B UTI O N S Xiaojun Xiang: Conceptualization (lead); data curation (equal); funding acquisition (lead); writing -review and editing (lead).Ling Zhou: Investigation (lead); writing -original draft (lead).Hui Guo: Investigation (lead).Quan Liao: Data curation (lead); methodology (lead).Jianping Zou: Investigation (supporting).Yi Le: Formal analysis (supporting); investigation (supporting).Ziling Fang: Formal analysis (supporting); supervision (supporting); validation (supporting); writing -review and editing (supporting).Jianping Xiong: Resources (equal); supervision (equal); writing -review and editing (equal).Shanshan Huang: Conceptualization (equal); data curation (supporting); formal analysis (supporting); funding acquisition (equal); project administration (equal); writing -review and editing (equal).Jun Deng: Funding acquisition (equal); methodology (equal); supervision (equal); writing -review and editing (equal).

Table 1
The association between miR-3133 expression and clinicopathological factors in patients with colorectal cancer.
TA B L E 2Note: All the data were analysed by the chi-square test or Fisher exact test.Entries in bold font indicate statistically significant (p < 0.05).