Upregulated UCA1 contributes to oxaliplatin resistance of hepatocellular carcinoma through inhibition of miR‐138‐5p and activation of AKT/mTOR signaling pathway

Abstract Hepatocellular carcinoma (HCC) inevitably developed oxaliplatin (OXA) resistance after long‐term treatment, but the mechanism remains unclear. Here, we found that LncRNA UCA1 was upregulated in most of OXA‐resistant HCC tissues and cells (HepG2/OXA and SMMC‐7721/OXA). Follow‐up analysis and online Kaplan–Meier Plotter revealed that HCC patients with high UCA1 level had a shorter survival compared with those with low expression. Overexpression of UCA1 increased OXA IC50 in HepG2 and SMMC‐7721 cells, whereas knockdown of UCA1 decreased OXA IC50 in resistant counterparts. Moreover, dual luciferase reporter assay showed that co‐transfection of UCA1‐WT plasmid with miR‐138‐5p mimics enhanced fluorescence signals, whereas co‐transfection of UCA1‐Mut plasmid and miR‐138‐5p mimics did not induce any changes. Consistently, UCA1 levels in HepG2/OXA and SMMC‐7721/OXA cells were downregulated after transfected with miR‐138‐5p mimics. UCA1 silencing or transfection of miR‐138‐5p mmics inhibited the activation of AKT and mTOR in HepG2/OXA and SMMC‐7721/OXA cells, whereas UCA1 overexpression increased the phosphorylated AKT and mTOR levels in parental counterparts. Rapamycin or miR‐138‐5p mimics similarly suppressed the activation of AKT and mTOR, whereas UCA1 overexpression exert opposite roles. Interestingly, administration of rapamycin or miR‐138‐5p mimics apparently antagonized the effects of UCA1 on AKT and mTOR activation. Besides, depletion of UCA1 triggered more dramatic regression of HepG2 xenografts than that of HepG2/OXA xenografts with OXA treatment and impaired the p‐AKT and p‐mTOR levels in vivo. In conclusion, our findings provide the evidence that UCA1 may contribute to OXA resistance via miR‐138‐5p‐mediated AK /mTOR activation, suggesting that UCA1 is a potential therapeutic target for HCC.


| INTRODUC TI ON
Liver cancer is the fourth leading cause of global mortality, and approximately 80% liver cancers are hepatocellular carcinoma (HCC). 1,2 The incidence of HCC in China is higher than that in other countries and the number of new cases and deaths in China account for nearly half 3,4 Guangxi province is a high-risk area for HCC in China where the mortality of HCC reaches 40%, mainly due to hepatitis B virus (HBV) infection, aflatoxin intake, and polluted drinking water. [5][6][7] HCC is always known as one of the most aggressive cancers that often progressing silently. Therefore, most HCC patients (∼80%) usually present with advanced stage and lose the chance of surgical resection, liver transplantation, or localized tumor ablation. 8 Systemic chemotherapy is commonly used in advanced HCC as a palliative therapy in order to relieve symptoms and improve quality of life, control tumor growth or spread and prolong overall survival (OS). 9,10 The first approved systemic agent to demonstrate a survival benefit in patients with advanced HCC was sorafenib, a multikinase inhibitor that inhibits tumor angiogenesis and growth. 11 Oxaliplatin (OXA) is a third-generation platinum agent with obvious advantages, including great tolerance, broad therapeutic window, as well as more cost-effectiveness than sorafenib. 12 OXA-containing regimens such FOLFOX4 (OXA plus infusional-fluorouracil (FU) and leucovorin (LV) (FOLFOX4)), XELOX (capecitabine plus OXA), and GEMOX (Gemcitabin plus OXA) have shown improved clinical activity against HCC. 13 FOLFOX4 was included in Chinese national clinical practice guidelines for HCC 14 and approved for the systemic treatment of locally advanced or metastatic HCC in China in 2013. 14,15 However, despite initial sensitivity to oxaliplatin, most HCC eventually develop OXA resistance which leads to treatment failure. 16 Therefore, OXA resistance has become a bottleneck that limits the clinical efficacy of HCC therapy.
Long noncoding RNAs (lncRNAs) are a class of heterogeneous nonprotein-coding RNAs with lengths ranging from 200 to 100,000 nt. lncRNAs regulate gene expression at multiple levels, including alternative splicing, chromatin modification, as well as protein localization and activity, due to their ability to bind to DNA, RNAs, and proteins. 17 Recently, lncRNAs have been discovered to play critical roles in the regulation of tumor proliferation, invasion, migration, and chemo-resistance. 18 Particularly, upregulated lncRNAs seem to exhibit tumor-promoting roles, whereas downregulated lncRNAs possess tumor-suppressive abilities. 19 For instance, lncRNA HOTAIR and MALAT-1 were upregulated in most HCC patient tissues, whereas inhibiting their expression could sensitize HCC cells to doxorubicin and cisplatin, respectively. 20,21 Conversely, L Wu et al. found that the decreased lncRNA KRAL is closely related to the resistance to 5-fluorouracil in HCC. 22 UCA1 (urothelial carcinoma associated 1) is a lncRNA that was firstly identified in human bladder carcinoma, 23 whose expression is also found to be elevated in many other cancers. 19,[24][25][26] UCA1 has been shown to facilitate the cancer cell growth, migration, invasion, metastasis and drug resistance by activating PI3K/ AKT, mTOR/STAT3, and other signaling pathways. 27,28 Moreover, UCA1 induced multidrug resistance to cisplatin and gemcitabine in bladder cancer cell by activating the transcription factor CREB after declining the expression of miR-196a-5p. 18 Silencing UCA1 upregulated cleaved PARP protein expression and inhibited the antiapoptosis protein Bcl-2, resulting in enhanced apoptosis and depressed chemotherapy resistance to adriamycin in gastric cancer cells. 29 Besides, UCA1 activated AKT/mTOR signaling, thus promoting NSCLC cells to undergo epithlial-mesenchymal transition (EMT) against Gefitinib and Erlotinib therapies. 30 However, the role of UCA1 in oxaliplatin resistance of HCC is still poorly understand.
In this study, we found that UCA1 was upregulated in OXAresistant HCC specimens and cells, and was closely associated with a shorter survival in HCC patients. Furthermore, miR-138-5p was shown to act as a potential target miRNA of UCA1 through bioinformatic analysis, dual-luciferase reporter assay, and qRT-PCR. In addition, overexpression of UCA1 induced the activation of AKT and mTOR and this effect was similarly inhibited by either administration of rapamycin or upregulation of miR-138-5p. In vivo model consistently showed that high UCA1 level maintained the growth of OXA-resistant exonegraft as well as elevated p-AKT and p-mTOR, which could be antagonized by UCA1 depletion. Together, our findings suggest that UCA1 may facilitate OXA resistance of HCC by miR-138-5p-mediated the activation of the AKT/mTOR signaling and act as a potential therapeutic target for HCC.

| Information of clinical sample and enrollment criteria
Electronic medical records from 2011 to 2017 were retrospectively studied to screen out 75 appropriate candidates as following criteria: (a) Written informed consents were obtained from patients. Patients were diagnosed with primary HCC according to "Guidelines that UCA1 may contribute to OXA resistance via miR-138-5p-mediated AK /mTOR activation, suggesting that UCA1 is a potential therapeutic target for HCC.

| Methyl thiazolyl tetrazolium (MTT) assay for detecting IC 50
As we previously reported, 34   would be able to targeted by UCA1. Therefore, the confirmation of the combination between UCA1 and miR-138-5p was conducted by using the dual-luciferase reporter assay system (Promega, USA).

| Colony formation assay
The miR-138-5p binding sites in UCA1 sequences (wild-type or mutant type) were embedded into the luciferase reporter vector pSI-

| Western blot
Western blot was performed as previously reported. 37 Whole cells or tissues extracts were prepared using 200ul RIPA lysis buffer.
Then 30 μg of each isolated protein was separated by 10% SDS-PAGE gel and transferred onto NC membrane. The whole blot membrane then was blocked using 5% skim milk. Membrane was probed with respective primary antibodies which are listed in supplementary Table S2. at 4°C overnight. Secondary antibody anti-rabbit IgG (1:30000, DyLight™ 800 4X PEG Conjugate, CST, US) or anti-mouse IgG (1:10000, Licor, USA) were used for visualization. Protein bands were then imaged using the infrared fluorescence imaging system Odyssey (LI-COR, USA). The intensity of each band was calculated using ImagJ Plus. The western blot assays were conducted at least three times. was calculated using a formula: V(mm 3 ) = (L × W 2 )/2. 38 The mice were further randomized to two treatment subgroups of either intraperitoneal injection of 10 mg/kg oxaliplatin or the same volume of saline once a week for 4 weeks. When the largest length of tumor exceed 20 mm, mice were sacrificed and the tumor was collected, weighted, and stored in ultra-low temperature freezer for further use.

| In situ hybridization (ISH) of UCA1
UCA1 expression in HCC tissues was measured using UCA1 ISH kit according to the manufacturer's protocol. All tissue sections were fixed by paraffin and then were deparaffinized with xylene and 100% ethanol. Subsequently, sections were rinsed in phosphatebuffered saline (PBS), fixed in 1% paraformaldehyde for 10 minutes at 25°C, then incubated in hybridization buffer for 4 hours at 42°C. The hybridization was performed using digoxin-labeled probe of UCA1 in the thermostat for reaction at 42℃overnight.

| Statistic analysis
As the results of ISH shown in Figure 1A (p < 0.05, Figure 1C). Moreover, online Kaplan-Meier Plotter created from the public databases including GEO, EGA, and TCGA 40 showed that Asian HCC patients with high expression of UCA1 had a shorter survival time compared to those with low UCA1 expression (p < 0.01, Figure 1D). To determine whether UCA1 is prognostically independent of the clinical variables it correlates with in Table S3, we also performed the univariate and multivariate Cox proportional hazards analysis. As shown in

| UCA1 expression is elevated in HCC cells resisted to OXA
To generate an in vitro cellular model of OXA resistance, we firstly tested the OXA efficacy on HCC cell lines HepG2, SMMC-7721, HCC-LM3, and SK-HEP-1. Figure 1E showed

| UCA1 induces OXA resistance in HCC cells
Since the enhanced expression of UCA1 was found in HepG2/ OXA and SMMC-7721/OXA, we hypothesized and tested that the elevated UCA1 could induce OXA resistance in HCC cells. As cells compared with their NC groups ( Figure 2B,D). Together, these results suggest that UCA1 contributes to OXA resistance and UCA1 knockdown enhances the sensitivity to OXA in HCC cells.

| UCA1 contains the binding sites of its inhibitor miR-138-5p
Next, we applied the bioinformatics prediction software DIANA tools lncBase Predicted v.2 and RNA22 version 2.0 to predict the potential targets of UCA1. The results suggested that miR-138-5p would be able to targeted by UCA1. Then, we generated mutations in UCA1 to explore whether miR-138-5p could potentially bind to the sequence of 3'-terminis" -ACCAGC-" of UCA1 ( Figure 2E). The results of dual-luciferase reporter assay showed that co-transfec-  Figure 2H). These data are in consistent with the result of DLR assay and further supports that miR-138-5p does in fact regulate UCA1 endogenously in HCC cells.

| UCA1 induces OXA resistance through suppression of miR-138-5p expression and activation of AKT/mTOR pathway
We previously proved that the activation of AKT/mTOR axis paly a predominant role in invasion and migration of HCC by inducing EMT transformation. 37 Thus, we were interested to dissect whether AKT/mTOR signaling might be involved in the regulatory role of UCA1 in OXA resistance of HCC cells. The results of western blot illustrated that the expression of p-AKT and p-mTOR was obviously increased (p < 0.001) in both HepG2 and SMMC-7721 cells with UCA1 overexpression (Figure 3A-C). Conversely, knockdown of UCA1 expression in HepG2/OXA and SMMC-7721/ OXA cells significantly decreased the expression of p-AKT and p-mTOR ( Figure 3D-F). Interestingly, miR-138-5p was assumed as a transcriptional suppressor of UCA1. 41 Thereby, we explored whether upregulation of miR-138-5p expression could cause the opposite effect on AKT/mTOR pathway activation induced by high UCA1 expression. As expected, we found that both UCA1 expression ( Figure 2H) and the level of phosphorylated AKT and mTOR in HepG2/OXA and SMMC-7721/OXA cells were greatly descended after transfection with miR-138-5p mimics (p < 0.01 or p < 0.05) ( Figure 3G-I). Furthermore, we used mTOR/AKT inhibitors in conjunction with UCA1 overexpression and/or miR-138-5p mimics to determine if the effect of UCA1 overexpression is mediated by miR-138-5p directly. As PI3 K inhibitor LY294002 has been reported to display its effect in low significance in HCC cells, whereas mTOR inhibitor rapamycin was found to robustly inhibit the activation of mTOR and AKT 42,43 and was more effective than LY294002 in HCC cells. 44 Thus, we used rapamycin as mTOR/ AKT inhibitor to perform validation experiments. We found that Bars are mean ± SD from 3 independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, compared to blank groups. J-M, Western blot analysis was used to measure the levels of AKT, mTOR and their phosphorylated forms in HepG2 and SMMC-7721 cells treated with AKT/mTOR inhibitor rapamycin (10 μmol/L, 24 h), or overexpressed UCA1, or transfected with miR-138-5p mimics, or combined administration of rapamycin or miR-138-5p mimics with UCA1. Bars are mean ± SD from 3 independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, compared to blank groups; # p < 0.05, ## p < 0.01, compared to overexpressing-UCA1 groups

| Knockdown of UCA1 re-sensitizes HCC cells to OXA treatment partially through inactivation of AKT/mTOR pathway in vivo
Combining the above results in clinical specimens and in vitro, we speculated that knockdown of UCA1 may re-sensitize HCC cells to OXA-shNC groups (p < 0.05 or p < 0.01, Figure 4F,G). The total amount of AKT and TOR was not significantly affected in both groups. All these data suggest that UCA1 knockdown re-sensitizes HCC to OXA therapy partially via activation of AKT/mTOR pathway in vivo.

| DISCUSS ION
HCC is notoriously refractory cancer which is resistant to conventional chemotherapy and radiotherapy, carrying a dismal prognosis.
LncRNAs are emerging as new players in the cancer paradigm demonstrating potential roles in both oncogenic and tumor-suppressive pathways. 45    On the other hand, activated AKT also suppresses the apoptotic initiator Bcl-2, but elevates the expression of Bax when HCC cells acquire OXA-resistance. 56 Also, microRNA-19a-3p promotes tumor metastasis and sorafenib resistance through the PTEN/Akt pathway in HCC. 57 Therefore, we dissected the molecular mechanism of UCA1 by focusing on detecting the activation level of AKT/mTOR pathway in our study, and found a positive correlation between AKT/mTOR activities and the OXA resistance in HCC.
Additionally, OXA has been administrated for advanced HCC patients due to the safety and economic applicability, 12 59 Additionally, UCA1 might function as an oncogene in tongue squamous cell carcinoma (TSCC) through regulating the miR-138-5p/ CC chemokine receptor (CCR7) axis, and the interaction between miR-138-5p and UCA1 or CCR7 has been identified. 60 In this study, we found that miR-138-5p could bind to UCA1 and upregulation of miR-138-5p expression exerts the opposite effect on AKT/mTOR pathway activation induced by high UCA1 expression.
Taken together, our results indicate that upregulated UCA1 contributes to OXA resistance in HCC via suppression of miR-138-5p expression and activation of AKT/mTOR pathway. These findings suggest that targeting UCA1/miR-138-5p/AKT/mTOR signaling axis potentially represents a novel therapeutic option for overcoming OXA resistance in HCC.

| CON CLUS ION
In summary, our study reports that UCA1 contributes to the OXA resistance of HCC in vitro and in vivo as well as in clinical patients.
Mechanistically, UCA1 downregulates miR-138-5p, thereby leading to activation of AKT/mTOR signaling pathway in HCC cells. These findings illustrate a novel mechanism for understanding OXA resistance in HCC, and suggest that UCA1/miR-138-5p/AKT/mTOR signaling axis has a potential therapeutic value for the treatment of OXA resistance in HCC patients.

ACK N OWLED G M ENT
We express our sincere gratitude to the research team of Prof. Hou and Prof. Li who provided experimental assistance for this project.

CO N FLI C T O F I NTE R E S T
None.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that supports the findings of this study are available on request from the corresponding author.