Overexpression of miR-125a-5p Inhibits Hepatocyte Proliferation through the STAT3 Regulation In Vivo and In Vitro

microRNAs (miRNAs) are critically involved in liver regeneration (LR). miR-125a-5p (miR-125a) is a tumor-suppressing miRNA, but its role in LR has not been studied. Our previous studies have proved that miR-125a was related to LR at the initiation phase, while the mechanism hepatocyte proliferation triggered by miR-125a in LR has been rarely evaluated. Herein, we mainly studied the molecular mechanism of miR-125a in triggering hepatocyte proliferation and the proliferation stage of LR. Firstly, a striking reduction of miR-125a was found at 24 h as well as 30 h following partial hepatectomy (PH) in rat liver tissue by miRNAs expression profiles as well as qRT-PCR analysis. Furthermore, in vitro, upregulation of miR-125a decreased proliferation as well as G1/S conversion, which promoted hepatocytes apoptosis. STAT3 was the target of miR-125a. In vivo, upregulation of miR-125a by tail vein injection of agomir inhibited LR index. Upregulation of miR-125a inhibited LR index and hepatocytes proliferation by STAT3/p-STAT3/JUN/BCL2 axis. In summary, these current discoveries indicated that miR-125a inhibited hepatocytes proliferation as well as LR by targeting STAT3 and via acting on the STAT3/p-STAT3/JUN/BCL2 axis.


Introduction
Liver has a remarkable regenerative capacity when it is damaged by injury or surgical resection [1][2][3]. 2/3 PH, also known as 70% PH, refers to the removal of the middle and left lobes of liver that account for 70% of the total liver [4]. After 2/3 PH, hepatocytes instantly enter G 1 phase, and soon traverse to S phase, this moment the DNA synthesis is up to a peak at 24-48 h after PH [5][6][7]. LR is a synchronized multistep process consisting of initiation (2 to 6 h), proliferation (12 to 72 h) and termination (120 to 168 h) [8,9]. However, the regeneration potential of liver is often impaired by pathological factors, such as liver failure and liver cirrhosis [10,11], thus it is worth emphasizing the necessity of finding novel channels to improve liver's regenerative potential.
miRNAs are series of small non-coding RNAs with approximately 20 nucleotides or so, which lead to translational repression or mRNA degradation via negatively regulating their target genes at post-transcriptional level [12,13]. miRNAs take on vital roles in multiple biological fields, in addition to common cell proliferation and apoptosis, as well as metabolism and even carcinogenesis [14][15][16]. Our previous study found some miRNAs were associated with LR at the initiation phase; among them, miR-125a showed a significant up-regulative level at 2 h post-2/3 PH [17]. Recently, several studies have indicated miRNAs take on key roles during the LR [16,18]. For instance, overexpression of miR-21 facilitated liver cell proliferation trough targeting PELI1 [19], FASLG [20] and PTEN [21]; However, downregulation of miR-378 [22], miR-26a [23] and miR-127 [24] contributed to hepatocytes proliferation during LR. miR-378 was involved in epithelial-mesenchymal transformation (EMT) in Hedgehog-driven regenerative liver cells by targeting GLI-Kruppel family member 3 (Gli3) [22]. In addition, downregulation of miR-23b inhibited cell growth as well as promoting cell apoptosis during the termination phase of LR by activating the TGF-b1/Smad3 pathway [25]; moreover, miR-20a inhibited mouse hepatocyte proliferation and liver regeneration via the TCF4/CDC2/CDC6 axis [26]. These effects of miRNAs on hepatocyte proliferation are relatively basic, and most of them do not involve in vivo mechanism studies. Therefore, it is necessary to excavate underlying mechanisms of other miRNAs in LR.
In the study, we discovered that miR-125a was markedly decreased in the rat liver tissue at 24 as well as 30 h post-2/3 PH. Subsequently, we illustrated that miR-125a inhibited hepatocytes proliferation, as well as G 1 /S conversion, via the target gene of STAT3, resulting in the inhibition of STAT3/p-STAT3/JUN/BCL2 axis in vivo and in vitro. Conjointly, our study identified miR-125a could act as a suppressor of hepatocytes proliferation as well as LR through regulating of the STAT3/p-STAT3/JUN/BCL2 axis, which might provide a novel target for LR after injury in future.

miR-125a Was Associated with Rat Liver Regeneration
It has previously demonstrated that the time 24 h post-PH was a peak of rat hepatocytes proliferation representing via the value of AST as well as BrdU positive cells [27,28]. Here, we established a 2/3 PH model, and analyzed the corresponding biochemical indicators serum aspartate aminotransferase (ALT) and alanine aminotransferase (AST). Compared with the control group. AST and ALT did not increase significantly after PH 24 and PH 30 h (Table S1), which may be related to the recovery of liver function. miRNA expression profiles analysis was used to compare miRNA expression between 24/30 h (post-PH) and 0 h. As a result, miR-29b-2, miR-21, miR-200c, miR-152, miR-132, miR-222 were found to be upregulated, while miR-3585, miR-503, miR-34b, miR-144, miR-145, miR-125a and let-7c-1 downregulated at 24 h and 30 h after PH ( Figure 1A). Subsequently, qRT-PCR was utilized to verify the miRNA expression profiles results. As shown in Figure 1B, miR-125a (previous IDs: miR-125a-5p) was sharply decreased at 24/30 h (post-PH). Additionally, miR-125a was also reduced on the 3rd as well as 4th days in hepatocytes post-PH ( Figure 1C), which was the hepatocytes logarithmic phase representing via PCNA expression ( Figure 1D). Accordingly, research was needed into the function of miR-125a in the proliferation of LR as well as hepatocytes proliferation.

miR-125a Reduced Hepatocytes Proliferation
miR-125a mimics (100 nM) and its control NC mimics were transfected to BRL-3A cells to survey the effect of miR-125a on rat hepatocytes growth as well as proliferation. miR-125a level was increased after the treatment of miR-125a mimics (Figure 2A). Both MTT and EdU assays demonstrated that miR-125a mimics inhibited BRL-3A cells proliferation ( Figure 2B,D). However, the opposite results were observed in treatment with an miR-125a inhibitor ( Figure S1). Additionally, a cell cycle assay indicated that miR-125a overexpression decreased the transition of G 1 -to-S for BRL-3A cells ( Figure 2C and Table S2). Therefore, miR-125a was confirmed to inhibit the BRL-3A cell growth and the proliferation.

MiR-125a Reduced Hepatocytes Proliferation
miR-125a mimics (100 nM) and its control NC mimics were transfected to BRL-3A cells to survey the effect of miR-125a on rat hepatocytes growth as well as proliferation. miR-125a level was increased after the treatment of miR-125a mimics ( Figure 2A). Both MTT and EdU assays demonstrated that miR-125a mimics inhibited BRL-3A cells proliferation ( Figure 2B,D). However, the opposite results were observed in treatment with an miR-125a inhibitor ( Figure S1). Additionally, a cell cycle assay indicated that miR-125a overexpression decreased the transition of G1-to-S for BRL-3A cells ( Figure 2C and Table S2). Therefore, miR-125a was confirmed to inhibit the BRL-3A cell growth and the proliferation. . Cell proliferation was examined through EdU (red) assay. Data are shown as mean ±SEM, * p < 0.05, ** p < 0.01.

MiR-125a Induced Hepatocytes Apoptosis
To survey the effect of miR-125a on apoptosis, miR-125a mimics and mimics NCs were transfected into BRL-3A cells. As Figure 3A,B demonstrate, the apoptotic rate was increased in mimic groups when compared with NCs group (p < 0.05), indicating that miR-125a induced BRL-3A apoptosis. Apoptotic cells were showed in areas of J4 (FITC positive dots showing early apoptosis cells) as well as J2 (V-FITC along with PI double

miR-125a Induced Hepatocytes Apoptosis
To survey the effect of miR-125a on apoptosis, miR-125a mimics and mimics NCs were transfected into BRL-3A cells. As Figure 3A,B demonstrate, the apoptotic rate was increased in mimic groups when compared with NCs group (p < 0.05), indicating that miR-125a induced BRL-3A apoptosis. Apoptotic cells were showed in areas of J4 (FITC positive dots showing early apoptosis cells) as well as J2 (V-FITC along with PI double positive dots showing late apoptosis cells) in Figure 3A, suggesting that miR-125a induced hepatocytes apoptosis. Figure 2. The effect of miR-125a on hepatocytes proliferation (A). miR-125a levels were examined through qRT-PCR following transfection of miR-125a mimics (miR-125a-m), and its control NC mimics in BRL-3A cells. (B). Cell viability was examined through MTT. (C). The role of miR-125a in BRL-3A cell cycles was detected by FACScan. (D). Cell proliferation was examined through EdU (red) assay. Data are shown as mean ±SEM, * p < 0.05, ** p < 0.01.

MiR-125a Induced Hepatocytes Apoptosis
To survey the effect of miR-125a on apoptosis, miR-125a mimics and mimics NCs were transfected into BRL-3A cells. As Figure 3A,B demonstrate, the apoptotic rate was increased in mimic groups when compared with NCs group (p < 0.05), indicating that miR-125a induced BRL-3A apoptosis. Apoptotic cells were showed in areas of J4 (FITC positive dots showing early apoptosis cells) as well as J2 (V-FITC along with PI double positive dots showing late apoptosis cells) in Figure 3A, suggesting that miR-125a induced hepatocytes apoptosis.

miR-125a Was Negatively Correlated with STAT3 Level In Vivo and Negatively Regulated STAT3 Level In Vitro
Previous studies have reported that miR-125a exercises its function in hepatic glycolipid metabolism by targeting STAT3 [29]. Here, the STAT3 expression decreased after treatment with miR-125a mimics in BRL-3A cells ( Figure 4A,B). Meanwhile, an opposite trend was found between the STAT3 level and miR-125a level at 12, 24 and 30 h post-PH compared with the control in rat liver ( Figure 4C,D), suggesting that there was a possible negative relation between miR-125a and STAT3 during rat LR proliferative stage, STAT3 may be a target gene of miR-125a in rat hepatocytes.

STAT3 Was the Direct Target of miR-125a in Rat Hepatocytes
To reveal the effect of miR-125a on rat hepatocytes, GO annotation and KEGG pathway were utilized to categorize the target genes predicted by miRWalk, miRanda, miRDB, miRMap, miRNAMap, RNAhybrid and Targetscan. 2421 putative targets were identified through the above database (Table S3). The JAK/STAT3 signaling pathway was one of the most significantly enriched pathways indirectly related with miR-125a (Table 1), and the target genes were mainly enriched in cell proliferation, death, and apoptosis ( Table 2 and S4). The activity of the STAT3 3 -UTR luciferase reporter was sharply decreased after treatment of the miR-125a mimic by dual-luciferase assay, but no significant changes were found in mutated STAT3 3 -UTR in the same treatment ( Figure 5A,B), suggesting that STAT3 was a direct target of miR-125a in rat hepatocytes. Furthermore, BRL-3A cell proliferation was inhibited after treatment with siRNA of STAT3 ( Figure S2).

STAT3 Level In Vitro
Previous studies have reported that miR-125a exercises its function in hepatic glycolipid metabolism by targeting STAT3 [29]. Here, the STAT3 expression decreased after treatment with miR-125a mimics in BRL-3A cells ( Figure 4A,B). Meanwhile, an opposite trend was found between the STAT3 level and miR-125a level at 12, 24 and 30 h post-PH compared with the control in rat liver ( Figure 4C,D), suggesting that there was a possible negative relation between miR-125a and STAT3 during rat LR proliferative stage, STAT3 may be a target gene of miR-125a in rat hepatocytes.

STAT3 Was the Direct Target of miR-125a in Rat Hepatocytes
To reveal the effect of miR-125a on rat hepatocytes, GO annotation and KEGG pathway were utilized to categorize the target genes predicted by miRWalk, miRanda, miRDB, miRMap, miRNAMap, RNAhybrid and Targetscan. 2421 putative targets were identified through the above database (Table S3). The JAK/STAT3 signaling pathway was one of the most significantly enriched pathways indirectly related with miR-125a (Table  1), and the target genes were mainly enriched in cell proliferation, death, and apoptosis (Tables 2 and S4). The activity of the STAT3 3′-UTR luciferase reporter was sharply decreased after treatment of the miR-125a mimic by dual-luciferase assay, but no significant changes were found in mutated STAT3 3′-UTR in the same treatment ( Figure  5A,B), suggesting that STAT3 was a direct target of miR-125a in rat hepatocytes. Furthermore, BRL-3A cell proliferation was inhibited after treatment with siRNA of STAT3 ( Figure S2). Table 1. Pathway enrichment analysis of the predicted target genes of miR-125a.   Table 2. Functional enrichment analysis of the predicted target genes of miR-125a.

MiR-125a Inhibited Hepatocytes Proliferation through STAT3/P-STAT3/JUN/BCL2 Axis
Several key genes downstream of STAT3 were examined using qRT-PCR as well as WB for further evaluating the potential mechanism of miR-125a in rat hepatocytes proliferation. Levels of STAT3, p-STAT3, JUN and BCL2 were decreased, and the CASPASE3 level was increased following transfection of miR-125a mimics compared with controls ( Figure 6A,B). Collectively, these findings demonstrated that miRNA-125a

miR-125a Inhibited Hepatocytes Proliferation through STAT3/P-STAT3/JUN/BCL2 Axis
Several key genes downstream of STAT3 were examined using qRT-PCR as well as WB for further evaluating the potential mechanism of miR-125a in rat hepatocytes proliferation. Levels of STAT3, p-STAT3, JUN and BCL2 were decreased, and the CASPASE3 level was increased following transfection of miR-125a mimics compared with controls ( Figure 6A,B). Collectively, these findings demonstrated that miRNA-125a restrained hepatocytes proliferation by acting on the axis of STAT3/P-STAT3/JUN/BCL2 in vitro.

Expression of miR-125a Agomir in Mouse Liver
To investigate the role of miR-125a in the hepatocytes proliferation of LR, miR-125a agomir and its control NC were injected into mice by tail vein injection 12 h and 24 h before 2/3 PH. 48 h post-2/3 PH, the samples were collected, and qRT-PCR and fluorescence observation were used to examine the expression of miR-125a in mouse liver. Cy5-labeled cells were found in miR-125a agomir and its control NC groups ( Figure 7A). The miR-125a level was increased following the injection of miR-125a agomir compared with controls ( Figure 7B).

Expression of miR-125a Agomir in Mouse Liver
To investigate the role of miR-125a in the hepatocytes proliferation of LR, miR-125a agomir and its control NC were injected into mice by tail vein injection 12 h and 24 h before 2/3 PH. 48 h post-2/3 PH, the samples were collected, and qRT-PCR and fluorescence observation were used to examine the expression of miR-125a in mouse liver. Cy5-labeled cells were found in miR-125a agomir and its control NC groups ( Figure 7A). The miR-125a level was increased following the injection of miR-125a agomir compared with controls ( Figure 7B).
To investigate the role of miR-125a in the hepatocytes proliferation of LR, miR-125a agomir and its control NC were injected into mice by tail vein injection 12 h and 24 h before 2/3 PH. 48 h post-2/3 PH, the samples were collected, and qRT-PCR and fluorescence observation were used to examine the expression of miR-125a in mouse liver. Cy5-labeled cells were found in miR-125a agomir and its control NC groups ( Figure 7A). The miR-125a level was increased following the injection of miR-125a agomir compared with controls ( Figure 7B).  Figure 7. Efficiency detection of miR-125a agomir in mouse liver. (A) Cy5-labeled cells were found in miR-125a agomir and its control NC groups by fluorescence microscope. (B) miR-125a level was examined through qRT-PCR analysis following injection of miR-125a agomir compared with controls. Data are shown as mean ± SEM. Nuclei were stained with DAPI (blue). Scale bar, 100 µm, ** p < 0.01.

miR-125a Inhibited LR through STAT3/p-STAT3/JUN/BCL2 Axis
Liver index and PCNA-labeled cells were decreased in injection of the miR-125a agomir group ( Figure 8A,B) when compared with controls groups; meanwhile, the levels of key genes downstream of STAT3 including p-STAT3, JUN and BCL2 were decreased, and CASPASE3 level was increased in miR-125a agomir groups ( Figure 8C,D). Collectively, these findings demonstrated that miRNA-125a restrained the proliferation process of LR through STAT3/p-STAT3/JUN/BCL2 axis. Liver index and PCNA-labeled cells were decreased in injection of the miR-125a agomir group ( Figure 8A,B) when compared with controls groups; meanwhile, the levels of key genes downstream of STAT3 including p-STAT3, JUN and BCL2 were decreased, and CASPASE3 level was increased in miR-125a agomir groups ( Figure 8C,D). Collectively, these findings demonstrated that miRNA-125a restrained the proliferation process of LR through STAT3/p-STAT3/JUN/BCL2 axis.

Discussion
2/3 PH offers a unique model to study the mechanism of LR [9]. LR post-PH is primarily aroused via hepatocytes proliferation, which is affected by miRNAs expression [22,25]. At present, the possible mechanisms remain largely unclear. We discovered that some miRNAs, including miR-125a and miR-145, were sharply decreased at 24 and 30 h

Discussion
2/3 PH offers a unique model to study the mechanism of LR [9]. LR post-PH is primarily aroused via hepatocytes proliferation, which is affected by miRNAs expression [22,25]. At present, the possible mechanisms remain largely unclear. We discovered that some miRNAs, including miR-125a and miR-145, were sharply decreased at 24 and 30 h post-2/3 PH in rat liver tissue through miRNA expression profiles analysis. Herein, we mainly discuss the role of miR-125a in LR and hepatocytes proliferation. The upregulation of miR-125a restrained proliferation as well as G 1 /S transition in hepatocytes. Additionally, miR-125a negatively regulated STAT3 in vitro/vivo. The upregulation of miR-125a restrained mouse liver regeneration in vivo. The following findings illustrated that miR-125a restrained hepatocytes proliferation and LR at the proliferation stage by effecting the axis of STAT3/P-STAT3/JUN/BCL2.
Multiple studies have illustrated that miR-125a played an inhibitory role in several kinds of tumors, such as glioblastoma [30], non-small cell lung carcinoma [31][32][33], hepatocellular cancer [34], retinoblastoma [35], osteosarcoma [36], cervical cancer [37] and breast cancer [38]. In liver, the studies of miR-125a have focused on liver virus infection and hepatocellular carcinoma. It has been reported that miR-125a could interfere with viral replication via binding with surface antigen encoded by its own transcript [39]. Up-regulating miR-125a significantly restrained HCC proliferation as well as metastasis through regulating MMP11 as well as VEGF-A [40]. Additionally, miR-125a-5p along with miR-125b induced the cell cycle blocked at p21-dependent G 1 phase via suppressing SIRT7 as well as CCND1 level in HCC [41]. Downregulation of miR-125a-5p might protect against isoflurane-induced liver injury by regulating hepatocyte proliferation and apoptosis [42]. miR-125a-5p improved hepatic glucose and lipid metabolism disorders in patients with type 2 diabetes by targeting STAT3 [29]. The above studies suggested that miR-125a may play a significant role in the liver. Here, we discovered miR-125a was sharply downregulated at 24 as well as 30 h post-2/3 PH through miRNA high-throughput sequencing along with qRT-PCR in liver tissues. Previously studies demonstrated that the time of 24 h post-PH was the peak of hepatocytes proliferation, which was shown via a value of serum AST as well as BrdU labeled cells [7]. Moreover, a low level of miR-125a was detected at the logarithmic growth phase of cultivated hepatocytes in the 3rd and 4th days. This indicated that miR-125 may be involved in the proliferation stage of LR. Our subsequent findings also proved that miR-125a restrained cell proliferation, G 1 /S transition, and induced apoptosis in hepatocytes.
It has been proved that the proliferative effect of miRNAs on LR depends on particular target genes of their own [16][17][18][19][20]. Therefore, to further survey the potential regulative mechanism of miR-125a in LR. We screened out cell proliferation-related signaling pathways by GO annotation as well as KEGG pathway assays, such as JAK-STAT3. Interestingly, it was found that STAT3 had an opposite tendency with miR-125a in LR and hepatocytes. Subsequently, the dual-luciferase system along with WB analysis further identified that STAT3 was one of the targets of miR-125a.
STAT3 is a potential transcription factor that can be activated by a series of cytokines and growth factors [43,44]. Activation of STAT3 signal pathway played vital effects on a series of complex life activities including apoptosis, proliferation, invasion, metastasis, differentiation as well as angiogenesis [45][46][47][48]. Additionally, activated STAT3 pathway scould induce abnormal proliferation as well as malignant transformation. Herein, STAT3 has been defined as an oncogene [49]. IL-6 is a pleiotropic cytokine that promotes liver regeneration through the activation of STAT3, and responses to liver injury. IL-6 mediates acute phase responses and induces cytoprotective and mitotic functions [50,51]. In general, IL-6 binds to the interleukin-6 receptor (IL-6R), and the IL-6/IL-6R complex initiates glycoprotein 130 (gp130) for the activation of JAK/STAT, MAPK and PI3K/AKT, which is essential for the early onset of liver disease and the progression and maintenance of the regenerative process [52,53]. STAT3, one of the targets of miR-125a, suppressed tumor invasion as well as metastasis in cervical carcinoma [37]. C-Jun is a major regulator of hepatocyte survival [54] and hepatocyte proliferation during regeneration [55]. It has been illustrated that synergistic activity of STAT3 and c-JUN were observed in human cancer specimens [56,57]. In addition, c-JUN or BCL2 was a downstream target gene of STAT3 in tumorigenesis [58][59][60]. However, whether the STAT3/p-STAT3/JUN/BCL2 axis is administered by miR-125a in hepatocytes and LR at the proliferation stage is unclear. Our findings demonstrated that up-regulation of miR-125a inhibited STAT3, p-STAT3, JUN and BCL2 expression in cultivated hepatocytes and liver tissue.
In conclusion, our current findings showed that miR-125a was presented a low level in the proliferative stage of LR and inhibited the proliferative efficiency of hepatocytes via acting on the STAT3/p-STAT3/JUN/BCL2 axis. Therefore, miR-125a could be served as a potential and novel promising target and regulate the development of LR as well as liver carcinoma.

PH Model Preparation and Tail Vein Injection
Adult Sprague Dawley (SD) rats (male, weighing 230 ± 20 g) were provided by the laboratory animal control office of Henan Normal University. The operation of animal experiments was allowed through Animal Care as well as the Use Committee at the university (License No: HNSD-2020-02-17), and executed sternly in the light of the Animal Protection Law of China. The 2/3 PH models' operations were executed according to the described method [61]. In short, 30 rats were randomly divided into 5 groups: 2/3 PH groups (PH 24 h) (n = 6), 2/3 PH groups (PH 30 h) sham-operated groups (SO 24 h), shamoperated groups (SO 30 h) (n = 6), and one control group (0h) (n = 6). These were used in miRNA high-throughput sequencing. In addition, another 12 rats were used in miR-125 and STAT3 validation experiments at PH/SO 12 h. Both the 0 h group and sham group were used as controls for miRNA expression profile data analysis as well as miR-125a and STAT3 validation experiments.
For the vivo studies of miR-125a, 6-8 week Balb/c mice (male, 25 ± 2 g) were purchased from the experimental animal center of Zhengzhou University. The mouse 2/3 PH model was executed according to the method of rat model described above. In short, 12 mice were randomly divided into two groups: the PH-miR-125a agomir group (n = 6), and the PH-NC agomir group (n = 6). miR-125a agomir (5 nM) and its control agomir NC were injected via tail vein 24 h and 12 h before 2/3 PH operation, respectively. When sampling, the models were sacrificed, and the liver was removed; part of the samples were fixed within 4% paraformaldehyde for immunohistochemistry; part of the samples were embedded with frozen section embedding agent for fluorescence observation; the rest of the samples were stored at −80 • C or liquid azote until further experiment. The agomir was chemically modified and labeled with cy5 for easier cell membrane penetration and fluorescence observation and was purchased from the company (Ribobio, Guangzhou, China).

Fluorescence Observation
The 5-µm-thick slices were obtained from tissue embedded in frozen section embedding agent. After 30 min at room temperature, slices were fixed with acetone and then stained with DAPI for 10 min. Fluorescence microscopy (Axio Imager D2, Carl Zeiss, Germany) was used for photography.

Immunohistochemistry
After deparaffinization/hydration, the 7-µm-thick slices were blocked with endogenous peroxidase blocking solution. They were then treated with 10% normal goat serum (CWBIO, Beijing, China), followed by incubation with the anti-PCNA antibody (1:500, Cell Signaling Technology Cat# 13,110) at 4 • C overnight, and then incubated with Biotin labeled secondary antibody and HRP labeled with streptavidin respectively (CWBIO, China). Finally, slides were stained with 3, 3 -diaminobenzidine and images were captured with a microscope.

Biochemical Index Analysis
After the mice fasted for 12 h, blood was taken from eyelid, 4 • C overnight; blood was centrifuged at 2500 rpm, and then serum was taken and stored at −80 • C. The activity of ALT and AST were executed by ALT/GPT and AST/GOT kits (Wanleibio, Shenyang, China) according to manufacturer's instruction. Briefly, a 5 µL sample was added into the 20 µL matrix liquid buffer, incubated at 37 • C for 1/2 h. Then, 20 µL of 2, 4-dinitrophenylhydrazine, thoroughly mixed, incubated at 37 • C for 1/3 h was added. Then 200 µL of 0.4 mol/L NaOH, gently shaken, incubated at room temperature for 1/4 h was added. Finally, OD value was determined through a microplate reader at 510 nm.

miRNA High-Throughput Sequencing and Analysis
Total RNA extraction from liver tissue, and miRNA high-throughput sequencing and analysis were conducted as previously described [17,61]. Library preparation and Illumina sequencing were carried out according to Illumina small RNA sample preparation protocol outlined by Shanghai Biotechnology Company (Shanghai, China) [62]. Briefly, firstly, the total liver tissue RNAs were checked and quantified using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Qualified RNA was purified and further linked to the 5 linker and 3 linker by T4 RNA ligase, which was then reverse-transcribed into cDNA by the SuperScript II Reverse Transcription Kit (Invitrogen, Carlsbad, CA, USA), which is then amplified and purified. Finally, the purified cDNA library was quantified with a qubit fluorometer (Invitrogen, Carlsbad, CA, USA), and then analyzed by Illumina Genome Analyzer IIx for cluster generation and 36 nt single-end sequencing. The raw data were analyzed as previously described [17]. The difference in miRNA level between PH groups and the control group was considered to be significant at a fold change ≥2 as well as a p-value ≤ 0.05.

Cell Culture and Transfection
Treatments of this part were conducted as our previously described [61]. Briefly, BRL-3A cells were obtained from cell bank of the School of Basic Medicine of Peking Union Medical College (Beijing, China). The cells had been kept in a high glucose DMEM complete medium, supplemented with 10% fetal bovine serum (FBS), 1% penicillin and 1%streptomycin in a concentration of 5% CO 2 incubator at 37 • C. miR-125a mimics (Ribobio, Guangzhou, China), miR-125a inhibitor (Ribobio, China), or their negative controls handled BRL-3A for 48 h, respectively. To further investigate the role of STAT3 in hepatocyte proliferation, cells were transfected with siRNA-STAT3 (siRNA1, 2, 3, Ribobio, China) or negative control for 48 h, respectively.

MTT Assay
After the cells were transfected with miR-125a mimics and their control, NC mimics, MTT reagent (0.5 mg/mL) was added. After four hours, DMSO was added and dissolved the crystallization; the microplate reader was measured, with OD value at 490 nm. The treatment was operated three times.

EdU Proliferation Assay
After the cells were transfected with miR-125a mimics and its control NC mimics, respectively, EdU (50 µmol/L, Ribobio, Guangzhou, China) was added and incubated for 2 h at 37 • C. Subsequently, 4% paraformaldehyde was used for fixing for 1/2 h at 4 • C. Subsequently Triton X-100 (0.5%) was used for permeabilization for 10 min. Finally, after 1 × Apollo ® reaction cocktail treating for 1/2 h, DAPI was utilized for staining for 1/2 h. Fluorescence microscopy was used for photography.

Cell Cycle Analysis
After the cells were transfected with miR-125a mimics and its control NC mimics for 48 h, respectively, the transfected BRL-3A cells were harvested. After fixing overnight at least, PI (20 µg, Sigma, Saint Louis, MI, USA) and RNase A (50 µg, Sigma, Saint Louis, MI, USA) in 1 mL PBS solution were used for treating without light for 0.5 h at 37 • C. Then DNA content was measured by FACSCan.

Cell Apoptosis
After the cells were transfected with miR-125a mimics and its control NC mimics for 48 h, respectively, the transfected BRL-3A cells were harvested. 1× Binding Buffer was used for washing and resuspending cells. Next, Annexin-V-FITC (3.5 µL, BD, Franklin Lakes, NJ, USA) and PI (3.5 µL, BD, USA) in 100 µL Binding Buffer was added and cells were incubated without light for 1/2 h. Finally, 1× Binding Buffer (400 µL) was added to stop the reaction, followed by FACSCan as quickly as possible.

Luciferase Vector Acquisition and Detection
Firstly, 3 UTR of STAT3 containing the recognition site by miR-125a was amplified and inserted into psiCHECK-2 (Promega, Madison, WI, USA). Meanwhile, the corresponding mutant 3 UTR of STAT3 without a corresponding recognition site was amplified and inserted into the same empty vector. Next, BRL-3A cells were plated overnight, then co-treated with miR-125a reagents as well as the recombinant WT/Mut-vector. Finally, a dual-luciferase kit (Promega, Madison, WI, USA) was used to obtain luciferase activities.

RNA Acquisition and qRT-PCR
RNA acquisition as well as cDNA synthesis was performed in RNA isolation reagent (Invitrogen, Carlsbad, CA, USA) and cDNA synthesis kit (Promega, Madison, WI, USA), respectively. The detail qRT-PCR instruction was described in our previous study [61]. Information of the primers used is summarized in Table 3. U6 or GAPDH was used as the internal reference to normalize miRNA and total mRNA level. Finally, the relative quantitative method of 2 −∆∆Ct was applied to calculate the levels of genes under test.

Data Analysis
SPSS software version 18.0 (SPSS Inc, Chicago, IL, USA ) was applied to deal with data. Mean ± SEM were used to show data. The difference between groups was analyzed by an independent t-test or an ANOVA-containing post hoc test. The value of p < 0.05 showed statistical significance.

Data Availability Statement:
All data supporting the findings of this study appear in the submitted manuscript or are available from the corresponding author upon reasonable request.