MiR-21-3p Promotes Hepatocellular Carcinoma Progression Through Regulating Smad7/Yap1

While the relation between and We aimed to explore the inuence of miR-21-3p on HCC initiation and progression by targeting Smad7 and further facilitating the expression of MicroRNA (miRNA) microarray analysis was performed for miRNA screening. Dual-luciferase assay was adopted for target verifying. The expressions of miRNA and related genes were quantied by qRT-PCR, western blotting, and immunohistochemical staining. Flow cytometry and transwell assay were used to discover cell apoptosis, invasion and metastasis abilities. Rat models were established to explore the axis's role in hepatocarcinogenesis. Bioinformatics analysis was performed for analyzing clinical signicance.


Abstract
Background Hepatocellular carcinoma (HCC) remains the global burden due to its high prevalence and mortality.
Emerging evidence con rms that microRNAs (miRNAs) play a vital role in cancer initiation and progression. MiRNAs are widely involved in the regulation of signaling pathways by targeting downstream genes. MiR-21-3p as a liver-enriched miRNA has not been fully investigated. Abnormal activation of TGF-β transduction pathway promoted by deletion of Smad7 matters since HCC occurrence. While the relation between miR-21-3p and Smad7 has not yet been con rmed. We aimed to explore the in uence of miR-21-3p on HCC initiation and progression by targeting Smad7 and further facilitating the expression of Yap1.

Methods
MicroRNA (miRNA) microarray analysis was performed for miRNA screening. Dual-luciferase assay was adopted for target verifying. The expressions of miRNA and related genes were quanti ed by qRT-PCR, western blotting, and immunohistochemical staining. Flow cytometry and transwell assay were used to discover cell apoptosis, invasion and metastasis abilities. Rat models were established to explore the axis's role in hepatocarcinogenesis. Bioinformatics analysis was performed for analyzing clinical signi cance.

Conclusions
MiR-21-3p promotes HCC migration and invasion via directly inhibiting Smad7 and further improving the expression of Yap1.

Background
With a rising global incidence and the leading cause of cancer-related death status, hepatocellular carcinoma (HCC) is still lacking effective treatment (1). More than half of HCC patients are already in advanced stages of liver cancer when rst diagnosed (2). For those patients with unresectable HCC, sorafenib and lenvatinib as multi-kinases inhibitors are considered as systemic e ciency drugs (3), while their performance on prolonging patient survival rate is far from satis ed (4). It is time to explore speci c mechanisms of occurrence and progression of HCC to nd effective clinical treatments.
MicroRNAs (miRNAs), the small endogenous non-coding ribonucleic acid molecular family, are mainly involved in RNA silencing and regulation of gene expressions through complementary combination with the 3'-UTR region of genes (5). MiRNAs are considered to be biomarkers of diagnosis and prognosis since discovered (6). MiR-21-3p was demonstrated to inhibit cell apoptosis rate in liver cancer stem cells via directly regulating PTEN (7). Our miRNA microarray analysis results displayed that miR-21-3p is signi cantly upregulated in HCC. Previous studies con rmed that miR-21-3p could in uence hepatoma cell growth through directly down-regulating adenosyltransferases 2A and 2B (MAT2A/MAT2B) (8).
However, more important targets wait to be identi ed to elucidate the indispensable role of miR-21-3p in HCC.
Smad7 (I-Smad) is the negative regulator of the TGF-β signaling transduction pathway through binding TGF-β Receptor (TβR ) and interfering with the recruitment of Smad2/Smad3 (R-smad), blocking functional Smad complex from interacting with DNA in the nucleus (9, 10). As a "double-edged sword" involving in maintaining cell normal proliferation and differentiation, the TGF-β signaling pathway plays an irreplaceable role in tumor formation and progression (11). Dysfunction of Smad7 and excessive activation of the TGF-β signaling pathway were commonly seen in various kinds of cancers (12,13). T Feng's team has veri ed that the deletion of Smad7 could accelerate the progress of HCC in mice models through activating the STAT3 signaling pathway (14). Loss-of function of Smad7 promoted HCC cell proliferation, accelerated G1-S transition and reduced cell apoptosis in vivo (15). The role of Smad7's abnormal deletion in HCC needs more depth researches. Yap1 (Yes-associated protein 1) is the "star nuclear effector" of the Hippo pathway, which consists of a series of kinases (16). Yap1 is in hyperactive status and is considered as an oncogene in several kinds of solid tumors, including liver cancer (17). The rst discovery of the crosstalk between Hippo and the TGF-β signaling pathway was that cytoplasmic Yap1 protein helped inhibit part of the signal transduction of TGF-β in COS-7 cell ( a CV-1 African green monkey broblast cell line by transformation with a mutant strain of Simian Virus 40 (SV40)) via enhancing Smad7 binding to TβR , and Yap1 was demonstrated to be a novel Smad7-interacting protein (18). Nevertheless, the close relationship between Smad7 and Yap1 in HCC remains unclear.
The present study described the overexpression of miR-21-3p in human HCC. We rst proposed that MiR-21-3p could facilitate the expression of Yap1 and promote malignant phenotypes progression via directly targeting Smad7 in HCC.

Materials And Methods
Clinical tissues and cell culture Ten paired HCC tissues and corresponding background livers (BLs) were obtained from HCC patients undergoing partial hepatectomy between January 2019 and December 2020. After surgical resection, the specimens were immediately snap-frozen in liquid nitrogen. RNA-later (#AM7201, Thermo sher, USA) was added to the cryogenic vials to prevent RNA degradation for subsequent extraction. The collection and usage of these samples were followed ethical and institutional guidelines and implement it after organizing informed consent from donors. The study was approved by the local Research Ethics Committee at Zhongnan Hospital of Wuhan University (Approval No.2018078). Cell lines (L02, Huh7, HepG2, HCCL-M3) involved in this experiment were purchased from Stem Cell Bank, Chinese Academy of Sciences. Mixture of Dulbecco's Modi ed Eagle's Medium (Gibco, USA) and 10% fetal bovine serum (Gibco, Thermo Fisher, USA) and 1% penicillin-streptomycin (Gibco, USA) was used to culture cells. All cells were grown in humidi ed 5% CO 2 at 37℃.

MiRNA microarray analysis
The microarray data adopted for analysis have been described in detail before (19). Data have been uploaded in NCBI's Gene Expression Omnibus (GEO) public database http://www.ncbi.nlm.nih.gov/geo/, accession number, GSE20077).

RNA extraction, Reverse transcription and quantitative realtime PCR
Total RNA extraction from tissues and cells were using Triol Reagent (Invitrogen, USA) following the instructions. The quality and concentration of RNA were detected by Nanodrop 2000 (ThermoFisher, USA). cDNA was obtained through a reverse transcription reaction with kits (Toyobo, Japan). The expression of miRNAs and genes was examined using UltraSYBR Mixture (CWbio, Beijing, China) on the CFX96Touch Real-Time PCR system (Biorad, USA). The mRNA levels of miR-21-3p and genes were normalized to U6 (Ribobio, Guangzhou, China) and GAPDH, respectively. The relative expression ratio was calculated with the 2 −ΔΔCq method. Primers used listed here:

MiRNA related reagents, plasmids and cell transfection
The precursor and inhibitor of miR-21-3p, mimic-NC and inhibitor-NC were obtained from Ribobio (Guangzhou, China). Inhibitors of miR-21-3p were the complementary strand of miR-21-3p and was dedicated to compete the binding sites between miR-21-3p and target genes, but would not decrease the mRNA expression level of miRNAs. The nal concentration of these reagents was 50 nM. Smad7 expression vector (P-Smad7) was constructed using GV141 vector by Genechem (Shanghai, China) and plasmids-NC (P-NC) served as an experimental control. All transfection experiments were performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) following the guidance.
Fire y luciferase activity was normalized to renilla luciferase activity. Data were collected using Enspire 2300 (PerkinElmer). Six repetitions per group were calculated.
Cell apoptosis assay and ow cytometry A total of 4 × 10 5 cells were transfected with miR-21-3p mimics with/or Smad7 plasmids. After 24 h, collecting (including dead cells in the culture medium) and processing cells according to the instructions of Annexin V-FITC/PI Double stain apoptosis detection kit (#4101-2, Bestbio, Shanghai, China). Collecting the uorescence intensity on the ow cytometer to check the apoptotic rate (Cyto ex Beckman, China) immediately.

Cell invasion and metastasis assay
The procedure for processing cell Huh-7 was the same as the cell apoptosis experiment. Huh-7 was

Animal experiments
The establishment of Wistar rat models were detailed descripted before, and grouping of these rats were according to the Metavir score system (20,21). Here was brief introduction. Wistar rats ( male, 7-8 weeks old, 200-220 g) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. All animal handling and experimental procedures were approved by the Animal Care and Use Committee of Wuhan University following the Animal Experiment/Animal Biosafety Level-III Laboratory Guidelines. Rats were injected 40% carbon tetrachloride (CCl4) dissolved in maize oil (1.5 ml/kg) every twice a week. At 8, 14, 18 weeks, liver brosis, cirrhosis and liver cancer were formed, respectively. Hematoxylin-eosin (HE) staining and masson staing which is more conductive to observe the degree of ber crosslinking of rat models were displayed in Figure  Immuno uorescence (IF) The distribution of micro laments and cytoskeleton was clearly shown through uorescently labeled phalloidin staining. Cells were cultured in soft and still matrixes 24 h. Fixed with 4% paraformaldehyde 20 min at room temperature. Phalloidin dilution rate: 1:500. Incubation at 37℃ for 1 h. Nuclear staining with DAPI for 5 min. Images were processed with an inverted microscope (Olympus IX3).
Gene set variation analysis (GSVA) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis with miR-21-3p/Smad7 expression GSVA was a useful tool provided by R (3.5.2) to explore the potential biological process and enriched KEGG pathways of research objects according to its high or low expression. The results were presented in the form of intuitive volcano maps and speci c heat maps. Gene terms with |logFC| ≥ 0.

Statistical analysis
Data were presented here as mean ± standard deviation (SD) and analyzed by the Student's t-test or oneway ANOVA. Quantitative data were representative of three experiments. Wilcoxon signed-rank test and Wilcoxon rank-sum test was adopted to analyze gene expressions in paired and non-paired tissue samples. The relationships between disease stages and gene expressions was checked by Wilcoxon rank-sum test. Prognosis analysis was performed using Kaplan-Meier method and univariate Cox regression. All bioinformatics statistical analyses and plots were produced using R (v.3.5.2). P < 0.05 was considered statistically signi cant.

Results
MiR-21-3p was upregulated in HCC and its clinical signi cance in HCC based on bioinformatics analysis Based on our previous chip analysis results, miR-21-3p was signi cantly upregulated in HCC compared to normal liver tissues (NLTs), which had statistical signi cance (P=0.028) ( Table 1). To further verify the results of chip analysis, the mRNA expression of miR-21-3p was examined in human HCC tissue samples and Wistar rat models, respectively. Results showed that miR-21-3p was signi cantly enriched at HCC tissues compared to background livers (BLs) (Fig.1A (a)). Its upregulation was also observed in rats with late-stage brosis and cirrhosis ( Fig.1A (b)). Compared to human relative normal cell L02, miR-21-3p was increased in liver cancer cell lines (Huh-7, Hep-G2, HCCL-M3) (Fig.1A (c)). Considering the well expression consistency in human tissues, rat liver disease models and cell lines, miR-21-3p was selected for subsequent analysis. To explore the clinical signi cance of the miR-21-3p in human HCCs, the Cancer Genome Atlas (TCGA) data, including 376 HCC patients, whose survival status was available, were analyzed. We found that patients with high miR-21-3p expression were signi cantly correlated with advanced clinical stages (P = 0.029), especially in T staging (P = 0.026), while no noticeable difference was observed in age and gender. (Figure 1B). Higher miR-21-3p expression level was correlated to shorter ten-year overall survival (OS) time (log-rank P=0.026) (Fig. 1C). GSVA analysis of miR-21-3p displayed that high expression of miR-21-3p was mainly enriched at 20 pathways in HCC (Fig.1D). In particular, the top three gene sets related were nitrogen metabolism, fatty acid metabolism and primary bile acid biosynthesis, severally (Fig.1E). KEGG pathway analysis of miR-21-3p targets revealed that the downstream targets were strongly linked to metabolic pathways, Hippo signaling road and TGF-beta signaling pathway ( Table 2). The intersections of miR-21-3p potential targets from three authoritative databases (TargentScan, PicTar, miRDB) were Smad7, HBP1, FBXO11, respectively (Fig.1F). Of these, Smad7 earned the highest score (Table 3), while the potential relationship between miR-21-3p and Smad7 waited to be recti ed.

Smad7 as the direct target of miR-21-3p was decreased in HCC
To further verify our analysis, expressions of Smad7 in human tissue samples and rat models were checked. Compared to BLs, Smad7 was obviously decreased in HCCs (9/10) ( Fig.2A (a)). Also, the lighter brownish staining was observed in HCCs compared to BLs ( Fig.2B (a)). Along with the progressing of liver disease in rat models, a gradually descending trend was observed in Smad7's protein expression level ( Fig.2A (b)), and immunohistochemical staining showed the same trend ( Fig.2B (b)). Considering the apparent up-regulation of miR-21-3p and down-regulation of Smad7 in HCC, Dual-Luciferase assay was done in Huh-7 cells to con rm the linear regulation relation between miR-21-3p and Smad7. Prediction websites presented there were two binding sites in the 3'-UTR region of Smad7 with high scores. In the group of co-transfecting miR-21-3p and reporter vector containing wild-type (WT) Smad7, luciferase activity was signi cantly inhibited, while in the group containing mutation Smad7 reporter vectors (mutation-3 (Mut3) mutating two sites simultaneously), the inhibition e ciency was no statistical signi cance between miR-NC and miR-21-3p group (Fig.2C). The above results indicated that the two prediction sites might both exert the effect on inhibiting Smad7.
MiR-21-3p could enhance the expression of Yap1 partly via down-regulating Smad7 The KEGG pathway enrichment analysis displayed that miR-21-3p was signi cantly related to the Hippo signaling pathway (Table 2). Besides, a previous study con rmed that Yap1 could enhance the binding of Smad7 to TβR (18). To further investigate the potential link among, miR-21-3p, Smad7 and Yap1 in HCC, the expression of Yap1 was detected in HCC tissues and rat models. Compared to BLs, Yap1 was enriched at human HCC tissues (8/10) (Fig.5A), and advanced stages of liver brosis and cirrhosis compared to F0 (F: brosis) ( Fig.2A (b)). Th e darker brownish staining in immunohistochemistry images was observed in HCC tissues ( Fig.5B (a)) and the stage of F3 and F4 in rats compared to F0 (Fig.5B (b)).
The clinical signi cance of Smad7/Yap1 based on bioinformatics analysis Data about 376 HCC patients was downloaded from the Cancer Genome Atlas (TCGA) database.

Discussion
HCC as the second leading cause of cancer-related death worldwide lacks e cient treatments. Abnormal expression of miRNAs played a critical role in cancer occurrence and progress including hepatocarcinogenesis. Previous reports focused on introducing the guide strand miR-21-5p's function in HCC, few researches about the passenger strand miR-21-3p. Until recently, reports on the independent function of miR-21-3p gradually increased, which brought us a new understanding of this microRNA (23,24). The independent function of miR-21-3p includes it could be used as a signature in predicting the survival rate of triple-negative breast cancer (25). Besides, it's overexpression could facilitate the pulmonary metastasis by in uencing the junction between tumor and stromal cells (26). In our study, we investigated miR-21-3p's role in the occurrence and progression of HCC. MiR-21-3p was increased in HCC tissue samples compared to adjacent background livers (P<0.001). Further, based on the analysis of mRNA expression level of miR-21-3p in rat models, miR-21-3p may be involved in not only HCC progression but also hepatocarcinogenesis. KEGG pathway analysis of miR-21-3p's targets indicated that it was strongly related to TGF-β signaling transduction pathway and Hippo signaling pathway. Besides, the abnormal overexpression of miR-21-3p facilitated the malignant phenotypes of liver cancer cells.
Chronic liver diseases including liver brosis, cirrhosis are strongly associated with HCC (27). One in three patients with cirrhosis will develop HCC in their life (28). Dysregulation of TGF-β signaling transduction pathway was observed and testi ed in liver brosis and cirrhosis (29). Enhancing evidence demonstrated that miRNAs were involved in the process of liver brosis and cirrhosis through directly targeting Smad proteins (30). Our results highlighted the role of Smad7 in HCC. The absence of Smad7 was widely seen in HCC tissues compared to BLs, and had a gradually descending trend along with the progression of liver disease in rat models. Besides, deletion of Smad7 showed a good consistency in tissues and cells in HCC. We con rmed that Smad7 was the direct target of miR-21-3p in HCC. Restoring Smad7 could impair the migration and invasion abilities which were enhanced by miR-21-3p of liver cancer cell lines Huh-7 and Hep-G2, promote cell early apoptosis. During GSVA analysis, Smad7 was mainly involved in TGF-β signaling transduction pathway in HCC. Therefore, in the occurrence and development of HCC, the deletion of Smad7 really matters.
Previous gene sets analysis of miR-21-3p's targets con rmed that it had strong relation with Hippo signaling pathway. Yap1 was obviously enriched at HCC tissues compared to BLs. In advanced stages of liver disease in rat models, it's expression was higher than normal liver tissue. Our experiments explored that overexpression of miR-21-3p could facilitate the expression of Yap1. The ascending of Yap1 via overexpression of miR-21-3p could be partly reversed by transfecting Smad7 in the protein level. Inccreased Yap1 enhanced the expression of CTGF. Upstream regulator LATS2 was the core serine/threonine kinase of the Hippo signaling transduction pathway. Once the Hippo was activated, phosphorylation of LATS2 could be facilitated and then further phosphorylating Yap1, resulting in Yap1 inactivation, and there is a negative feedback between Yap1 and LATS2 (31). Expression of LATS2 in overexpression of Smad7 group was downregulated while surged through transfecting miR-21-3p. This indicated that the up-regualtion of Yap1 was not caused by inactivation of LATS2 but loss function of Smad7. Further, increased Yap1 weakened the negative in uence of LATS2. Vitro experiments effect in Huh-7 cells was more pronounced than Hep-G2 cells which was correlated to the basic expressions of miR-21-3p and Smad7 in cells (Fig.1A(a), S1B). We proposed that overexpression Smad7 could cut the expression of Yap1 in the protein level, not in the mRNA level. Here we concluded the complete axis in Figure 7.
Besides, recent studies suggested that not only Yap1 can be regulated by a series of microRNAs, but nuclear Yap1/TAZ (Yap1's paralog) could in uence the conversion of pri-miRNA into pre-miRNA (32).
Combining our results, we highly proposed that in the axis of miR-21-3p-Smad7/Yap1, there is a possibility that overexpression of nuclear Yap1 could in turn affect the expression of miR-21-3p. Previous studies con rmed that Yap1 as a mechano-transduction effector translocated from cytoplasm to nucleus when cells were shifted from soft to stiff matrices (33)(34)(35). Meanwhile, during the hepatocarcinogenesis, liver brosis and cirrhosis are typical examples of extracellular matrix (ECM) sclerosis (36), which are the "boosters" of tumor metastasis. Although there are great differences between people, the stiff of liver cancer tissue (55kpa) is almost about ten times that of normal liver tissue (4kpa) detected by Fibroscan (37). A reasonable and also surprising discovery was that about twofold up-regulation of miR-21-3p was found in three cell lines (Huh-7, Hep-G2, Hep-3B) when cells were cultured on stiffer substrates (40kpa) (Fig.S3A). Besides, the nuclear Yap1 in cells cultured on stiff gels was increased (Fig.S3B). Immuno uorescence results presented the deformation of cytoskeleton and changes of cell polarity in the stiffer matrixes (Fig.S3C). This indicated that there might be a positive loop between miR-21-3p and Yap1 during the progression of HCC. However, how did cells sense the mechanical signals and then transfer them into molecular signals, even changing the level miRNAs remained unclear. Our team believed that there was still an interesting story waiting to be gured out and we're working on it.

Conclusions
Our results revealed the oncogenic role of miR-21-3p in HCC. MiR-21-3p-Smad7/Yap1 axis was identi ed and might be potential therapeutic targets for HCC diagnose and treatments.

Consent for publication
All authors have been approved this publication.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Authors' contributions HYH designed and wrote this article. HYH and WD did the experiments. YML, WC and LJ revised this manuscript and gave useful suggestions. ZQ, LJ supervised this manuscript. CY funded this article and gave useful instructions. All authors read and approved the nal manuscript.