Long noncoding RNA HOTTIP mediates SRF expression through sponging miR‐150 in hepatic stellate cells

Abstract HOXA transcript at the distal tip (HOTTIP) has been shown to be up‐regulated in a variety of cancers and is identified as an oncogenic long noncoding RNA. However, the biological role of HOTTIP in liver fibrosis is unclear. Here, we reported that HOTTIP was specifically overexpressed in activated hepatic stellate cells (HSCs). HOTTIP knockdown suppressed the activation and proliferation of HSCs. Luciferase reporter assay showed that HOTTIP and serum response factor (SRF) were targets of miR‐150. RNA binding protein immunoprecipitation assay indicated the interaction between miR‐150 and HOTTIP. Further study revealed that HOTTIP increased SRF expression as a competing endogenous RNA for miR‐150, thus prompting HSC activation. Taken together, we provide a novel HOTTIP‐miR‐150‐SRF signalling cascade in liver fibrosis.


| INTRODUCTION
Liver fibrosis is a wound-healing response to chronic liver injury. 1 During chronic injury, the liver parenchyma is gradually replaced with excessive extracellular matrix, eventually causing disruption of liver architecture and the loss of liver function. Hepatic stellate cell (HSC) activation is regarded as the important event in the progression of liver fibrosis. 2,3 Recently, it has been reported that many noncoding RNAs have been linked to HSC activation. 4,5 However, the function and underlying mechanism of noncoding RNAs in liver fibrosis is far from being fully elucidated.
MicroRNAs (miRNAs) have been reported to be implicated in many cell functions, such as proliferation, differentiation, development, and apoptosis. 6,7 miRNAs are differentially expressed during the course of liver fibrosis. For instance, the increased expression of miR-125a and miR-34a as well as the decreased expression of miR-181, miR-200a, and miR-29b are observed in fibrotic livers. [8][9][10][11][12] miR-150 is reduced in activated HSCs and its mimics decreases HSC activation through inhibiting c-myb expression. 13 Our previous data also showed that miR-150 can hinder the activation of HSCs via inhibition of Sp1 and Col4A4. 14 Whether there are other targets of miR-150 remains largely unknown.
Increasing evidence indicates that HOTTIP is overexpressed in prostate cancer, lung cancer, and pancreatic cancer. [15][16][17] knockdown impedes cell viability, proliferation, invasion, and angiogenesis in human cancer cells, 18,19 therefore it is been identified as an oncogenic long noncoding RNA (lncRNA). In addition, HOT-TIP overexpression is in association with poor prognosis in cancers. 20,21 HOTTIP expression is regulated by miR-125, miR-192, and miR-204. 22,23 However, whether HOTTIP can modulate miR-NAs has not been investigated. Many studies have showed that lncRNAs can act as natural sponge to modulate target expression by sequestering miRNAs. 24,25 However, it is unclear whether HOTTIP may also have such a role to link miRNA and its target in liver fibrosis.
In the present study, our results showed that HOTTIP is specifically in activated HSCs. HOTTIP silencing can hinder activation and proliferation of HSCs. Further study revealed that HOT-TIP acts as a ceRNA to enhance serum response factor (SRF) expression by sequestering miR-150, thus promoting the activation of HSCs.

| Animals
Male C57BL/6J mice (20 ± 2 g) were bred in the animal house without specific pathogen. All experiments in this study were approved by the Animal Ethics Committee of Wenzhou Medical University. Mouse liver fibrosis was set up via a biweekly intraperitoneal injection with a 10% solution of carbon tetrachloride (CCl 4 , Sigma-Aldrich, St. Louis, MO, USA) diluted in olive oil for 8 weeks.

| Cell culture
Primary mouse HSCs were obtained by pronase/collagenase perfusion solution plus density gradient centrifugation, as previously detailed. 26 The purity of the isolated HSCs was evaluated based on immunocytochemical staining for a-SMA and the purity was in excess of 95%. Mouse hepatocytes were obtained using collagenase perfusion technique as previously described. 27 Cells were cultured in Dulbecco's modified Eagle's medium including 10% foetal bovine serum, streptomycin (100 g/mL), and penicillin (100 units/mL). The cells were incubated at 37°C in a humidified incubator of 5% CO 2 .
The constructs were validated by DNA sequencing.

| Immunofluorescence and Immunohistochemistry
Cells were fixed with an acetic acid:ethanol (1:3) solution, permeabilized in 0.1% PBS-Tween, and then blocked with 5% goat and horse serum/PBS. Then cells were treated with mouse primary antibodies against α-SMA and type I collagen (Abcam, Cambridge, UK) overnight at 4°C, followed by Alexa Fluor 568-labeled rabbit anti-mouse IgG (Life Technologies). Images were examined with a Carl Zeiss LSM710 confocal microscope (Carl Zeiss AG, Jena, Germany). After liver sections were dewaxed, dehydrated, and subjected to antigen retrieval, the samples were treated overnight at 4°C with mouse a-SMA antibody, followed by a biotinylated secondary antibody. The α-SMA expression was visualized by DAB staining.

| Quantitative real-time PCR (qRT-PCR)
The total RNA was isolated from HSCs and liver tissues using TRIzol

| RNA binding protein immunoprecipitation (RIP) assay
RIP experiment was performed using the EZ-Magna RIP kit (Millipore, Billerica, MA, USA). After HSCs were lysed with RIP lysis buffer, the cells were incubated with anti-Argonaute-2 (Ago2) antibody or Isotype-matched IgG. The samples were immunoprecipitated with proteinase k and RNA was isolated by TRIzol reagent. The quantification of HOTTIP and miR-150 was analysed by qRT-PCR.

| Cell proliferation assay
HSCs were transduced with Ad-shHOTTIP for 48 hours. Then the cells were incubated with 5-Ethyny-2′-deoxyuridine (EdU) for 12 hours. Cell proliferation was examined using an EdU detection kit (Beyotime Biotechnology, Jiangsu, China) in keeping with the manufacturer's instructions.

| Western blot
Proteins were isolated using a RIPA lysis buffer (Beyotime Biotechnology) and subjected to SDS-PAGE. After blocking, the membrane was incubated with the primary antibodies (β-actin, type I collagen, and α-SMA, Abcam) and the secondary antibodies (Li-Cor Biosciences Inc., Lincoln, NE, USA). Signals were examined using an Odyssey infrared scanner (Li-Cor Biosciences Inc.).

| Statistical analysis
The results were expressed as the mean ± SD. Differences between groups were compared using a Student's t test or one-way analysis of variance. P < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS 13.0 (IBM, Armonk, NY, USA).

| HOTTIP is up-regulated in activated HSCs
To explore the role of HOTTIP in liver fibrogenesis, we first analysed the expression of HOTTIP in HSCs. Primary quiescent HSCs were isolated from normal male C57BL/6J mice and cultured, which mimicks the in vivo activation process. qRT-PCR results showed that HOTTIP expression at day 10 was increased by 22.6-fold compared with that at day 2 ( Figure 1A). Then we further analysed whether the alteration of HOTTIP exists in a mouse liver fibrosis model. Masson staining was performed to evaluate the degree of liver fibrosis ( Figure 1B). The results of qRT-PCR indicated that the mRNA expression of Col1A1 and α-SMA were dramatically up-regulated in CCl 4 -treated mice compared with oil-treated mice ( Figure 1C).
Immunohistochemistry results showed that the protein expression of α-SMA in CCl 4 -treated mice was increased compared with oil-treated mice. However, the expression of HOTTIP was unaltered upon CCl 4 treatment ( Figure 1E). Based on the above findings, we deduced that HOTTIP may be specifically increased in HSCs during liver fibrosis. Therefore, we analysed the expression of HOTTIP in primary HSCs and hepatocytes isolated from CCl 4 -treated mice. Expectedly, HOTTIP expression was markedly up-regulated in HSCs isolated from CCl 4treated mice compared with those from oil-treated mice, whereas we failed to find a remarkable up-regulation of HOTTIP in hepatocytes ( Figure 1F). In conclusion, these results showed that HOTTIP is specifically dysregulated in HSCs during experimental fibrogenesis.

| Activation and proliferation of HSCs is negatively regulated by HOTTIP knockdown
The remarkable up-regulation of HOTTIP in HSCs urged us to determine the underlying biological role of HOTTIP knockdown in liver fibrogenesis. Knockdown of HOTTIP markedly reduced HOTTIP expression relative to the Ad-shCtrl (Figure 2A) were transduced with Ad-shHOTTIP. As indicated in Figure 2B,C, the mRNA levels of Col1A1 and α-SMA were decreased by 72% and 66% in Ad-shHOTTIP cells, respectively, compared with Ad-shCtrl cells. Consistently, the protein levels of type I collagen and α-SMA

| HOTTIP interacts with miR-150
Our previous study indicates that miR-150 is decreased in activated LX-2 cells. 14 Here, we detected the expression of miR-150 in mouse HSCs by qRT-PCR. There was a 66% reduction in the expression of miR-150 in primary mouse HSCs at day 10 compared with day 2 (Figure 3A), consistent with the previous report. 28 Recently, some miR-NAs have been reported to directly inhibit lncRNA expression. 29,30 Here, we explored whether miR-150 can suppress HOTTIP expression.
The online software RNA22 was employed to predict that HOTTIP contains a target of miR-150 ( Figure 3B). Next, we applied the luciferase assays to confirm whether miR-150 binds to HOTTIP. As shown in Figure 3C, miR-150 mimics effectively reduced the luciferase activity in pmirGLO-HOTTIP-wt compared with the miR-NC. However, mutating the seed regions for miR-150 almost abolished its luciferase activity. Then, we explored the effect of miR-150 on the expression of HOTTIP in HSCs. In comparison with the miR-NC, delivery of miR-150 mimics remarkably increased the level of miR-150 ( Figure 3D).
Consistently, miR-150 mimics resulted in the 58% reduction in HOT-TIP expression relative to the miR-NC ( Figure 3E). Taken together, our data indicates that HOTTIP is a target of miR-150.
We next set out to investigate whether HOTTIP binds with miR-150. It has been reported that miR-21 and GAS5 can inhibit each other via the RNA-mediated silencing mechanism. We inferred that HOTTIP

| HOTTIP controls SRF expression by competing for miR-150
It has been reported that some lncRNAs can serve as a sponge by binding miRNAs, thereby abrogating miRNA-induced inhibiting activity on the targets. Herein, we investigated whether HOTTIP can regulate target expression by competing for miR-150. Among the putative miR-150 targets, we selected SRF given that it is in control of transcription of α-SMA and Col1A1. We found that SRF 3′UTR harbours three target sites of miR-150 based on TargetScan software ( Figure 4A). As shown in Figure 4B, compared with the miR-NC, miR-150 mimics remarkably reduced the luciferase activity (55% inhibition) of the pmirGLO-SRF-wt reporter. However, miR-150 mimics did not decrease the luciferase activity in pmirGLO-SRF-mut, confirming that SRF is a direct target of miR-150. By contrast, exogenous expression of HOTTIP-mut did not restore the decreased luciferase activity by miR-150 ( Figure 4B). Then we verified the above results in primary HSCs. As indicated in Figure 4E,F, overexpression of Ad-HOTTIP-wt or Ad-HOTTIP-mut notedly increased the expression of HOTTIP relative to the Ad-Ctrl, respectively. Compared with the miR-NC, miR-150 decreased the mRNA expression of SRF ( Figure 4G). However, overexpression of Ad-HOTTIP-wt but not Ad-HOTTIP-mut rescued the down-regulation of SRF by miR-150. Our results support the conclusion that lncRNA can sequester miRNA, thus stopping it from repressing its target.

| HOTTIP promotes HSC activation through decreasing miR-150 and increasing SRF
Finally, we investigated whether miR-150 and SRF are required for HOTTIP-promoted HSC activation. As shown in Figure 5

| DISCUSSION
In this study, we show the pathological implication of HOTTIP in facilitating the activation and proliferation of HSCs. Mechanistically, HOTTIP can function as a regulator of SRF expression through miR-150 binding, which provides an intriguing approach for treating liver fibrosis.
Accumulating studies reveal that HOTTIP has been up-regulated in a multitude of diseases. [31][32][33] However, we failed to detect an aberrant expression of HOTTIP in fibrotic livers. It is likely that dysregulation of HOTTIP may occur particularly in a certain hepatic cell. As expected, the increased expression of HOTTIP was observed in activated HSCs, whereas no significant difference was found between hepatocytes isolated from oil-treated mice and CCl 4 -treated mice, suggesting that HOTTIP participates in the progression of liver fibrosis. Given that hepatocyte is the major cell type in the liver, we reason that invariability of HOTTIP in hepatocytes may conceal its alteration in HSCs.
HOTTIP has been reported to contribute to oncogenesis and tumour metastasis, thus identified as an oncogenic lncRNA. However, the role of HOTTIP in liver fibrosis has not been well- luciferase activity and mRNA expression of SRF by miR-150 was partially reversed by Ad-HOTTIP-wt but not Ad-HOTTIP-mut, which supports the conclusion that ceRNA can sequester miRNA, thus protecting its target mRNA against suppression.
Our previous study demonstrates that miR-150 can inhibit the activation of HSCs through targeting Sp1 and Col4A4. Here we identified SRF as the new target of miR-150, which enlarges the repertoire of miRNA targets. Further study showed that miR-150 and SRF are involved in the profibrogenic effect of HOTTIP, which implies that targeting the HOTTIP-miR-150-SRF axis may signify a new therapeutic application in liver fibrosis.
There are some limitations in the current study. HOTTIP may sequester many miRNAs simultaneously and one miRNA has many targets. Consequently, "multiple-to-multiple" ceRNA interactions exists in the real cellular context instead of this simple "one-to-one" model. In addition, the biological role of HOTTIP in vivo has not yet to be investigated.
In summary, our experimental data suggest that HOTTIP can serve as the endogenous sponge to bind miR-150 and then release its target SRF, thereby promoting the progression of liver fibrosis.

CONF LICT OF I NTERESTS
The authors confirm that there are no conflicts of interest.

AUTHORS' CONTRI BUTION
Fujun Yu and Zhiming huang designed the study; Jianjian Zheng and Yuqing Mao performed the research and wrote the manuscript; Peihong Dong analysed the data.