Ubiquitin-specific protease 2a promotes hepatocellular carcinoma progression via deubiquitination and stabilization of RAB1A

Abstract

Purpose

Deubiquitination, the inverse process of ubiquitination, is catalyzed by deubiquitinases (DUBs) that remove ubiquitin from target proteins and subsequently prevent their degradation by proteasomes. Previously, deubiquitination has been found to be involved in hepatocellular carcinoma (HCC) progression. As yet, however, little is known about the exact role of deubiquitination in the development and/or progression of this type of cancer.

Methods

HCC tissues and tissue microarrays were used to detect expression of the DUB ubiquitin-specific protease 2a (USP2a). The critical role of USP2a in HCC development and progression was assessed in both in vitro cell and in vivo animal models. LC-MS/MS analyses were performed to identify potential targets of USP2a in HCC cells, after which regulation of target protein stability and ubiquitin status by USP2a were investigated.

Results

We found that USP2a was significantly upregulated in HCC tissues, and that a high expression was positively associated with a poor prognosis. Subsequently, we found that USP2a silencing resulted in inhibition of HCC cell proliferation, migration and invasion, whereas exogenous USP2a overexpression resulted in the opposite effects, both in vitro and in vivo. Mechanistically, LC-MS/MS analysis revealed that RAB1A, a key regulator of the ER and Golgi vesicular transport system, serves as a potential target of USP2a in HCC cells. In addition, we found that USP2a can deubiquitinate and stabilize RAB1A and prevent its degradation, and that this process is required for inducing HCC progression by USP2a.

Conclusions

Our data indicate that USP2a can promote HCC progression via deubiquitination and stabilization of RAB1A. This observation indicates that DUB targeting may serve as a novel approach to improve the treatment of HCC.

Introduction

Primary liver cancer (PLC) is estimated to be the sixth most commonly diagnosed cancer and the fourth leading cause of cancer-related death worldwide [1]. Hepatocellular carcinoma (HCC) is the main pathological subtype of PLC (> 80%), and represents a heavy medical burden [2]. Patients presenting with early stage HCC may profit from satisfactory treatment effects through surgical resection, liver transplantation and/or arterial embolization. As yet, however, only a small percentage of patients can be fully cured, whereas most patients develop advanced stages [3]. At present, there is no effective treatment for advanced HCC, although some molecular-targeting drugs have been shown to delay HCC progression and to improve the survival rates of these patients [4]. Due to the acquisition of drug resistance and other factors, however, the overall median survival rate of patients with HCC treated with these specialized drugs is only one year [4, 5]. Therefore, there is a pressing need to develop new therapies for advanced HCC in order to improve the overall survival of these patients. Continued exploration of the molecular mechanism(s) underlying HCC development and/or progression is an important basis for achieving this goal.

Dysregulated ubiquitination leads to abnormal expression of proteins and irregular regulation of signaling pathways, which may result in multiple diseases, including cancer [6]. Protein ubiquitination is both a dynamic and reversible process. Deubiquitination is catalyzed mainly by deubiquitinases (DUBs), which can inhibit the degradation of target proteins through the ubiquitin-proteasome pathway, or affect endocytosis, transport and activity of target proteins by removing ubiquitin chain modification(s) of target proteins [7]. Recent studies have revealed that aberrant deubiquitination mediated by DUBs can lead to an abnormal stability and activity of many essential proteins that may be involved in DNA damage repair, proliferation, apoptosis, migration, invasion and metastasis [8]. Ubiquitin-specific proteases (USPs) represent the largest subgroup of the DUB family, comprising more than 60 members, and they have been extensively studied in various cancers [9]. USP22 can, for example, maintain gastric cancer stem cell properties and promote gastric cancer progression by stabilizing BMI1 (B cell-specific Moloney murine leukemia virus integration site 1) [10]. At present, little is known about abnormal deubiquitination mediated by USPs in HCC development and/or progression. So far, it has been reported that the expression of USP21 [11] and USP7 [12] is significantly increased in HCC tissues, and that these proteases can promote the proliferation, migration and invasion of HCC cells via deubiquitination and stabilization of some important cancer-related proteins. Our preliminary research revealed that USP4 is upregulated in HCC cells and promotes their metastasis by increasing TGF-β signaling-induced epithelial-mesenchymal transition [13].

USP2a is another USP that was first found to be involved in cancer development through its abnormal expression in prostate cancer and its binding and deubiquitination of fatty acid synthase (FASN), resulting in increased FASN levels that, in turn, prevent apoptosis [14]. In recent years, through progress in biochemical technologies, several key cancer-related proteins such as MDM2 (mouse double minute 2 homolog) [15], EGFR (epidermal growth factor receptor) [16] and β-catenin [17], have been identified as targets of USP2a, indicating that USP2a may play an important role in cancer development and/or progression. In the current study, we aim to assess the role of USP2a in HCC.

Materials and methods

Reagents

Antibodies directed against human USP2a were purchased from Abgent (San Diego, CA, USA), whereas those directed against human RAB1A, Flag-Tag, V5-Tag, K48-linkage-specific polyubiquitin and β-actin were purchased from Cell Signaling Technology (Danvers, MA, USA). Protein A magnetic beads were obtained from Merck Millipore (Billerica, MA, USA). Cycloheximide (CHX) was purchased from Abmole (Shanghai, China) and MG132 was purchased from MCE (Monmouth Junction, NJ, USA). Cell Counting Kit-8 was purchased from Dojindo Molecular Technologies (Rockville, MD, USA).

HCC tissues and tissue microarray construction

HCC tissues and matching adjacent normal control tissues were collected between 2014 and 2016 from patients treated with primary surgery at the Second Affiliated Hospital Surgery Department of Chongqing Medical University (Chongqing, China), and were further processed into a tissue microarray using standard procedures. None of the patients received radiotherapy or chemotherapy prior to surgery, and all tissues obtained were evaluated and subjected to histological diagnosis by pathologists. All patients involved in this study provided informed consent, i.e., that their tissues could be retained and analyzed for research purposes only. The study was approved by the Human Research Ethics Committee of the Second Affiliated Hospital of Chongqing Medical University.

Cell lines and culture

The human normal hepatocyte cell line L02 and human HCC cell lines SK-Hep1, Huh7 SMMC-7721 and MHCC97-H were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The human HCC cell line HepG2 was purchased from the American Type Culture Collection (ATCC, VA, USA). All cells were maintained in high glucose DMEM medium (Hyclone Laboratories, UT, USA) containing 10% fetal bovine serum (Gibco, NY, USA), and cultured in an incubator at 37 °C containing 5% CO₂.

Lentivirus infection and stable cell line selection

Lentiviruses carrying a USP2a small hairpin RNA (shRNA) or a USP2a overexpression sequence were synthesized by Hanbio Biotechnology (Shanghai, China). Lentiviruses carrying a RAB1A shRNA or a RAB1A overexpression sequence were purchased from GeneCopoeia (MD,USA). Cells in their exponential growth phase were seeded into six-well plates and cultured overnight. On the next day, the cells were infected with the above mentioned lentiviruses according to the manufacturer’s instructions. 48 h later, puromycin was added to the medium for one week to obtain stable cell lines. The human USP2a siRNA-specific target sequences used were [18]: siRNA1, 5’-CAGATTGTGGTTACTGTTCTA-3’; siRNA2, 5’-CAGGAGAATGGCACACTTTCA-3’; siRNA3, 5’-CCGCGCTTTGTTGGCTATAAT-3’. The negative control siRNA used was: 5’-TTCTCCGAACGTGTCACGTAA-3’.

Quantitative RT-PCR assay

Total RNA from HCC cells was extracted using RNAiso Plus reagent (Takara, Dalian, China). Next, the RNA was converted into cDNA using a PrimeScript™ RT Master Mix (Takara), according to the manufacturer’s instructions. qRT-PCR assays were performed using a SYBR Green Master Mix Kit (Takara) and β-actin was used as an internal control. The 2^−ΔΔCq method was employed to analyze relative gene expression levels. The primer sequences used were: USP2a forward: 5’-TCCGAACCAGCAAGCTCACA-3’, reverse: 5’-CCCTGGACTGCGACAGTAGG-3’; RAB1A forward: 5’-CCCGGAACAGCCTATCTCAT-3’, reverse:5’-TCGGGATTCATGCTGGACAT-3’; GAPDH forward:5’-CATGAGAAGTATGACAACAGCCT-3’, reverse: 5’-AGTCCTTCCACGATACCAAAGT-3’.

Cycloheximide chase, immunoprecipitation and ubiquitination assays

A cycloheximide (CHX) chase assay was performed to examine the half-life of RAB1A. In brief, cells were treated with the protein synthesis inhibitor CHX (100 µg/ml) for varying periods. Next, proteins were extracted from the treated cells and Western blotting was performed to detect RAB1A expression.

For immunoprecipitation (IP) assays, total protein from HCC cells was extracted using an IP lysis buffer (Thermo Fisher Scientific, MA, USA). Prior to IP, protein A magnetic beads were added to the cell lysates to remove non-specific proteins. Next, pure lysates were immunoprecipitated using target antibodies at 4 °C overnight. On the next day, protein A magnetic beads were added to the lysates containing the antigen-antibody complexes and incubated at 4 °C for 2 h. Subsequently, the supernatants were discarded and the remaining magnetic beads-antigen-antibody complexes were retained and suspended with 1x loading buffer, followed by heating at 100 °C for 10 min. Next, the magnetic beads were discarded and the remaining immunoprecipitated protein complexes were subjected to Western blotting for identification.

To detect ubiquitination of RAB1A, cells were treated with the proteasome inhibitor MG132 (10 µM) for 6 h, after which proteins were extracted and IP assays were performed.

Immunofluorescence staining

Cells were seeded in a 35-mm dish and cultured overnight. Next, the medium was discarded and the cells were sequentially fixed with polyformaldehyde for 10 min, treated with X-Triton for 5 min and then with hydrogen peroxide for 10 min, and blocked with goat serum at 37 °C for 15 min. After this, the cells were incubated with primary antibodies at 4 °C overnight. On the following day, the cells were incubated with appropriate secondary fluorescent antibodies (Cell Signaling Technology) at room temperature for 2 h. Nuclei were stained with DAPI. Finally, protein fluorescence was evaluated using confocal microscopy.

In vivo tumor growth and metastasis assays

Subcutaneous tumor growth assays were employed to assess tumor growth in vivo. Briefly, 100 µl medium containing 5 × 10⁶ cells was subcutaneously injected into nude mice (BALB/c, SPF grade, 4 weeks old). Tumor sizes were measured every week. Four weeks later, the mice were sacrificed and the tumors were removed and weighed. All tumors were subsequently fixed with formalin and subjected to hematoxylin and eosin as well as immunohistochemical (IHC) staining. In addition, the abdominal cavities of the mice were dissected to monitor whether the subcutaneous tumors had metastasized, which may reflect metastatic capacity.

LC-MS/MS analysis

Protein samples from HepG2 and SK-Hep1 cells were collected via IP. Next, 10 mM dithiothreitol (final concentration) was added to the samples to reduce proteins, followed by the addition of a final concentration of 55 mM ammonium iodide acetate and, lastly, 1 mg of trypsin to hydrolyze proteins overnight (8–16 h). The resulting hydrolyzed polypeptides were desalted using a C18 column and, after the desalted polypeptides were dried, they were dissolved in a loading buffer containing 0.1% formic acid and 3% acetonitrile. Finally, the polypeptides were analyzed on a LC-MS/MS instrument (ekspertTM nanoLC; AB Sciex TripleTOF 5600-plus) and the original data directly submitted to Proteinpilot software (AB Sciex) for database retrieval.

CCK-8 cell proliferation, protein expression, migration and invasion assays

Cells in the exponential growth phase were seeded in a 96-well plate at a density of 5 × 10³ cells/well and cultured overnight. Next, at the same time interval, CCK-8 reagent (Cell Counting Kit-8) was added to the culture medium according to the concentration recommended in the manual, and incubated at 37 °C for 2 h. Optical densities of the cells were subsequently determined using a Universal Microplate Spectrophotometer and plotted as cell proliferation curves.

Immunohistochemistry (IHC), Western blotting, colony formation, scratch wound-healing, transwell and matrigel assays were all performed as described before [13, 19].

Statistical analysis

Data are presented as mean ± standard deviation. Each assay was repeated three times within independent experiments. Statistical analyses were performed using the SPSS 22.0 software (Chicago, IL, USA). Student’s t-tests or non-parametric tests were used to compare two groups of independent samples. Variance tests, followed by Dunnett or Bonferroni post-hoc tests were used to analyze differences of multiple comparisons. A chi-squared test was used to assess correlations between USP2a expression and clinicopathological features. Kaplan-Meier survival analysis was used to evaluate the correlation between USP2a expression and patient prognosis. Spearman rank correlation tests were used to analyze the correlation between USP2a and RAB1A expression. P < 0.05 was considered statistically significant.

Results

USP2a expression in HCC is significantly increased and positively associated with poor pathological grading, lymphatic metastasis and a poor prognosis

USP2a expression was first evaluated on a tissue microarray encompassing 100 pairs of HCC and matched adjacent tissues. IHC staining revealed that the expression of USP2a in the HCC tissues was significantly higher than that in the matched control adjacent tissues (Fig. 1A and B). Some cases were further evaluated using Western blotting, confirming that USP2a was significantly increased in HCC tissues compared to controls (Fig. 1C and D).

Fig. 1.

USP2a expression in HCC is significantly increased and positively associated with a poor prognosis. (A) Representative images of IHC staining of USP2a in HCC tissues and matched adjacent normal control tissues. (B) USP2a expression scores in HCC tissues and control tissues, ** p < 0.001. (C) USP2a expression in HCC and matched adjacent normal control tissues detected by Western blotting. N: matched adjacent control tissues; T: HCC tissues. (D) Quantitative analysis of Western blot assays, with relative expression of USP2a defined as the ratio of USP2a to β-actin. (E) Correlation between USP2a expression and overall survival of HCC patients assessed by Kaplan-Meier assay with log-rank testing

Next, we evaluated correlations between USP2a expression and several clinicopathological features. We found that USP2a expression was positively associated with a pathological grading of poor differentiation and liver portal lymph node metastasis, whereas no significant correlations were noted between USP2a expression and other clinicopathological features, such as patient gender, age, tumor size or HBV infection (Table 1). Kaplan-Meier survival analysis indicated that patients exhibiting a high USP2a expression had a poor prognosis (Fig. 1E). Follow-up data indicated that the median survival time of HCC patients with a low USP2a expression was 25 months, whereas that of patients with a high USP2a expression was only 10.5 months. These data suggest that USP2a may be involved in the malignant progression of HCC.

Table 1.

Correlation between USP2a expression and clinicopathological features of HCC patients

		USP2a
Clinicopathological features	Cases	Low expression	High expression	P value
Sex
male	86	62	24	0.959
female	14	10	4
Age
≤ 50	44	33	11	0.553
> 50	56	39	17
Tumor size (cm)
≤ 5	56	44	12	0.084
> 5	40	25	15
N/A	4	3	1
Pathological grading
I	11	9	2	0.036*
II	58	46	12
III	23	12	11
N/A	8	5	3
Liver portal lymph node metastasis
No	83	65	18	0.005**
Yes	14	6	8
N/A	3	1	2
HBV infection
No	14	9	5	0.530
Yes	80	58	22
N/A	6	5	1
AFP (ng/mL)
≤ 400	59	45	14	0.277
> 400	32	21	11
N/A	9	6	3

N/A: unknown data, *p < 0.05, **p < 0.01

USP2a significantly accelerates HCC cell growth

To assess the effect of USP2a on the biology of HCC cells, we first set out to examine the expression of USP2a in several HCC cell lines. Using Western blotting, we found that compared to the normal human hepatocyte cell line L02, USP2a was highly expressed in the HCC cell lines HepG2, SK-Hep1 and SMMC-7721, and that there was no significant difference with the Huh7 and MHCC97-H cell lines (Fig. 2A). Based on these data, we next utilized a shRNA-carrying lentivirus to knock down USP2a expression in HepG2 and SK-Hep1 cells, and a lentivirus carrying a USP2a overexpression sequence to exogenously overexpress USP2a in Huh7 cells. The effect of the respective expression interventions was verified at both the mRNA and protein levels (Fig. 2B and C). We found that the growth rate of USP2a silenced HepG2 cells was decreased compared to that of the negative control cells, indicating that USP2a may affect HCC cell growth (Fig. 2D). Next, CCK-8 assays were performed to determine the impact of USP2a on HCC cell proliferation. We found that the proliferation rates of USP2a silenced HepG2 and SK-Hep1 cells were significantly lower compared to those of control cells, and that the proliferation of USP2a overexpressing Huh7 cells was significantly increased compared to control cells (Fig. 2E). Additional colony formation assays showed that the ability of HepG2 and SK-Hep1 cells to form colonies was significantly impaired after USP2a silencing, and that the ability of Huh7 cells to form colonies was significantly enhanced after exogenous USP2a overexpression (Fig. 2F). Together, these data indicate that USP2a can promote HCC cell growth.

Fig. 2.

USP2a significantly accelerates HCC cell growth. (A) USP2a expression in normal human liver and HCC cell lines detected by Western blotting, *** p < 0.001. (B) Efficiency of USP2a silencing or overexpression in HCC cells evaluated by qRT-PCR, *** p < 0.001. (C) Confirmation of USP2a silencing or overexpression in HCC cells by Western blotting. (D) Growth state of USP2a silenced HepG2 cells or control cells observed under an inverted microscope at 200 × magnification. (E) Effect of USP2a silencing or overexpression on the proliferation of HCC cells evaluated using CCK-8 assays, *** p < 0.001. (F) Effect of USP2a silencing or overexpression on colony formation of HCC cells evaluated using colony-formation assays, *** p < 0.001

USP2a significantly promotes HCC cell migration and invasion

SK-Hep1 cells are considered to possess mesenchymal characteristics, leading to a more metastatic potential [20]. Here, we noted that the distribution of SK-Hep1 cells was scattered, but that these cells began to cluster after USP2a silencing (Fig. 3A). This observation suggests that USP2a may be involved in HCC cell metastasis. In accordance with this notion we found, using scratch wound-healing assays, that the healing (i.e., migration) capability of SK-Hep1 cells following USP2a silencing was significantly decreased compared to that of negative control cells (Fig. 3B), whereas the healing capacity of Huh7 cell following exogenous USP2a overexpression was significantly enhanced compared to that of control cells (Fig. 3C). Subsequent transwell migration assays confirmed that USP2a silencing significantly reduced the migratory ability of HepG2 and SK-Hep1 cells (Fig. 3D), while exogenous USP2a overexpression significantly promoted this ability in Huh7 cells (Fig. 3E). Since migration is the basis of invasion, we next investigated the effect of USP2a expression on HCC cell invasion. Using matrigel invasion assays, we found that USP2a silenced HepG2 and SK-Hep1 cells exhibited a lower invasive ability than control cells (Fig. 3F), whereas USP2a overexpressing Huh7 cells exhibited a higher invasive ability than control cells (Fig. 3G). These data indicate that USP2a can promote HCC cell migration and invasion.

Fig. 3.

USP2a significantly promotes HCC cell migration and invasion. (A) Distribution of USP2a silenced SK-Hep1 cells and control cells observed under an inverted microscope at 100 × magnification. (B) Effect of USP2a silencing on the migration of SK-Hep1 cells evaluated using scratch wound-healing assays (magnification, 100×). (C) Effect of USP2a overexpression on the migration of Huh7 cells evaluated using scratch wound-healing assays (magnification, 100×), *** p < 0.001. (D) Migration abilities of USP2a silenced HepG2 and SK-Hep1 cells evaluated using transwell migration assays (magnification, 100×), *** p < 0.001. (E) Migration ability of USP2a overexpressing Huh7 cells evaluated using transwell migration assays (magnification, 100×), ** p < 0.01. (F) Effect of USP2a silencing on the invasive ability of HepG2 and SK-Hep1 cells evaluated using matrigel invasion assays (magnification, 100×), ** p < 0.01, *** p < 0.001. (G) Effect of USP2a overexpression on the invasive ability of Huh7 cells evaluated using matrigel invasion assays (magnification, 100×), ** p < 0.01

USP2a enhances HCC cell growth and metastasis in vivo

To further confirm the effects of USP2a on HCC cell growth in vivo, subcutaneous tumor assays were performed. Identical numbers of SK-Hep1 cells with USP2a silencing and negative control cells were subcutaneously injected into nude mice, after which tumor formation was monitored and measured every week. Five weeks after inoculation, the nude mice were sacrificed, and tumors were removed and weighed. We found that the sizes and weights of tumors derived from USP2a silenced SK-Hep1 cells were significantly smaller than those derived from negative control cells (Fig. 4A). Similarly, identical numbers of Huh7 cells with USP2a overexpression and negative control cells were subcutaneously injected into nude mice. The sizes and weights of tumors derived from USP2a overexpressing Huh7 cells were significantly larger than those from negative control cells (Fig. 4B). These data indicate that USP2a can promote HCC cell growth in vivo.

Fig. 4.

USP2a promotes HCC cell growth and metastasis in vivo. (A) Sizes and weights of subcutaneous tumors derived from SK-Hep1 cells exhibiting USP2a silencing and negative control cells, * p < 0.05, ** p < 0.01, *** p < 0.001. (B) Sizes and weights of subcutaneous tumors derived from Huh7 cells exhibiting USP2a overexpression and negative control cells, * p < 0.05, *** p < 0.001. (C) USP2a silencing significantly inhibits SK-Hep1 cell metastasis in vivo. (D) H&E and IHC staining confirming the pathological characteristics of subcutaneous tumors derived from the SK-Hep1 cell treatment groups, and USP2a expression within these tumors. (E) H&E and IHC staining confirming the pathological characteristics and USP2a expression of subcutaneous tumors derived from USP2a overexpressing Huh7 cells versus negative control cells

To test whether the subcutaneous tumors exhibited metastatic potential, we dissected the abdominal cavities of the aforementioned nude mice to evaluate whether the subcutaneous tumors had metastasized. We found that in the SK-Hep1 negative control group large numbers of metastatic nodules were present in the liver and connecting portal vein, and that a small number of metastatic nodules was present in other parts of the abdominal cavity. In contrast, no metastatic nodules were found in the abdominal cavities of the mice inoculated with USP2a silenced SK-Hep1 cells (Fig. 4C). This result indicates that USP2a silencing significantly inhibits HCC metastasis in vivo.

All tumors recovered from the mice were fixed with formalin and subjected to paraffin sectioning. Subsequent H&E staining confirmed that the pathological characteristics of all the tumors were consistent with malignant disease. USP2a expression in the tumors from the different treatment groups was confirmed by IHC staining (Fig. 4D and E).

USP2a interacts with RAB1A in HCC cells

Although previous studies have shown that USP2a may be involved in cancer progression via targeting oncoproteins, the mechanism underlying USP2a promotion of HCC progression has remained unclear. To investigate this potential mechanism, we performed immunoprecipitation (IP) assays using an anti-Flag-USP2a antibody to precipitate potential USP2a target proteins in HCC cells. The precipitated proteins were subsequently identified by LC-MS/MS analysis. In total, 210 proteins from HepG2 cells and 279 proteins from SK-Hep1 cells were identified, of which 121 proteins were coincident. Among those overlapping proteins, we found that RAB1A might serve as a downstream target of USP2a. (Fig. 5A). RAB1A is a member of the Rab small GTPase family and is known to mediate dynamic membrane trafficking between the endoplasmic reticulum (ER) and Golgi vesicular transport system. This transport system involves various important signaling transduction processes, including those governing cell growth, lipid metabolism and autophagy. In addition, it has been reported that RAB1A may promote cancer cell proliferation, migration and invasion. Therefore, we reasoned that USP2a may promote malignant HCC progression by regulating RAB1A. To test this hypothesis, Flag-USP2a and V5-RAB1A were co-expressed in HepG2 or SK-Hep1 cells, after which co-immunoprecipitation (co-IP) was carried out. By doing so, we found that V5-RAB1A immunoprecipitated with Flag-USP2a using an anti-Flag antibody. Similarly, we found that Flag-USP2a immunoprecipitated with V5-RAB1A using an anti-V5 antibody (Fig. 5B). In addition, endogenous IP was performed in SK-Hep1 cells, and we found that endogenous RAB1A can form a complex with endogenous USP2a (Fig. 5C). Besides, immunofluorescence staining showed that USP2a and RAB1A are co-localized in the cytoplasm of HepG2 and SK-Hep1 cells (Fig. 5D). Together, these results indicate that USP2a can interact with RAB1A in HCC cells.

Fig. 5.

USP2a interacts with RAB1A in HCC cells. (A) IP assays were performed in HepG2 and SK-Hep1 cells to precipitate proteins that may bind directly to USP2a. The Venn diagram shows the number of binding partners for USP2a in the two groups, which were identified by LC-MS/MS analysis. The top ten overlapping proteins are presented. (B) Co-IP assays of lysates from SK-Hep1 cells co-expressing Flag-USP2a and V5-RAB1A were performed to validate USP2a/RAB1A complex formation at the exogenous level. (C) IP assays of lysates from SK-Hep1 cells were performed to confirm USP2a/RAB1A complex formation at the endogenous level. (D) Flag-USP2a and V5-RAB1A were co-expressed in HepG2 and SK-Hep1 cells, and ICC assays were employed to observe their colocalization

USP2a stabilizes RAB1A and protects it from degradation

Although we found that USP2a can interact with RAB1A, it remains to be determined whether it can regulate the fate of RAB1A. To address this question, we first investigated the effect of USP2a on RAB1A expression at the transcriptional level using qRT-PCR. We found that at the transcriptional level RAB1A was unaffected after USP2a silencing or exogenous overexpression in HCC cells (Fig. 6A). Next, Western blotting was used to assess the effect of USP2a on RAB1A expression at the protein level. We found that RAB1A expression was significantly reduced after USP2a silencing in HepG2 and SK-Hep1 cells, and that it was markedly increased in Huh7 cells exhibiting USP2a overexpression (Fig. 6B). These results indicate that USP2a may affect RAB1A expression at the protein level, but not at the transcript level.

Fig. 6.

USP2a deubiquitinates and stabilizes RAB1A in HCC cells. (A) Effect of USP2a on RAB1A expression at the transcriptional level in HCC cells detected by qRT-PCR, *** p < 0.001, NS, not significant. (B) Effect of USP2a on RAB1A expression at the protein level in HCC cells detected by Western blotting. (C) CHX chase assays were performed to monitor the half-life of the RAB1A protein in USP2a silenced SK-Hep1 cells and negative control cells, * p < 0.05, ** p < 0.01. Each group of cells was treated with CHX (100 µg/ml) for a fixed period. Next, proteins were extracted from these cells after which Western blotting was used to detect RAB1A expression. (D) CHX chase assays were performed to monitor the half-life of the RAB1A protein in USP2a overexpressing Huh7 cells and negative control cells, ** p < 0.01. (E) Impact of USP2a silencing on RAB1A ubiquitination. Lysates of SK-Hep1 cells were immunoprecipitated using an antibody directed against V5-Tag and analyzed using Western blotting to detect K48-linkage-specific polyubiquitin of RAB1A. (F) Impact of USP2a overexpression on RAB1A ubiquitination in Huh7 cells

As mentioned above, USP2a can inhibit the degradation of target proteins through the ubiquitin-proteasome pathway. Therefore, we next assessed the effect of USP2a on the stability of the RAB1A protein. Using CHX chase assays we found that the half-life of RAB1A in SK-Hep1 cells was decreased after USP2a silencing (Fig. 6C), whereas exogenous USP2a overexpression increased the half-life of RAB1A in Huh7 cells (Fig. 6D), indicating that USP2a can stabilize RAB1A in HCC cells. Concomitantly, the effect of USP2a on the ubiquitination of RAB1A was monitored. To this end, SK-Hep1 cells were co-infected with lentiviruses carrying USP2a/shRNA1 or RAB1A and, subsequently, treated with MG132 for 6 h. After this, proteins were extracted and IP assays were performed. By doing so, we found that the K48-linkage-specific polyubiquitination of RAB1A was significantly increased in USP2a silenced cells compared to that in control cells (Fig. 6E). Similarly, Huh7 cells were co-infected with USP2a and RAB1A, after which IP assays indicated that USP2a overexpression led to opposite results (Fig. 6F). These data indicate that USP2a can stabilize RAB1A and protect it from degradation through deubiquitination.

RAB1A is required for the oncogenic role of USP2a in HCC cells

Since it has amply been shown that RAB1A is involved in tumor cell proliferation, migration and invasion, we next hypothesized that RAB1A may affect the oncogenic effect of USP2a in HCC. To corroborate this hypothesis, we performed a functional rescue test. Using CCK-8 assays, we found that USP2a silencing significantly inhibited HCC cell growth, which could be reversed by exogenous overexpression of RAB1A (Fig. 7A). In contrast, we found that the pro-proliferative effect of USP2a overexpression could be restricted by RAB1A knockdown (Fig. 7B). Subsequent migration assays revealed that USP2a silencing led to decreased HCC cell migration, which could be reversed by RAB1A overexpression (Fig. 7C). Similarly, RAB1A knockdown abrogated USP2a-induced HCC cell migration (Fig. 7D). Therefore, we conclude that RAB1A is required for USP2a-propagated HCC cell growth and migration.

Fig. 7.

RAB1A is required for USP2a to promote HCC cell growth and migration. (A) USP2a silencing inhibits SK-Hep1 cell growth, which is reversed by exogenous RAB1A overexpression, ** p < 0.01, *** p < 0.001. (B) USP2a overexpression promotes Huh7 cell growth, which was restricted by RAB1A knockdown, * p < 0.05, *** p < 0.001. (C) Inhibition by USP2a silencing of SK-Hep1 cell migration is recovered by exogenous RAB1A overexpression, * p < 0.05, ** p < 0.01, *** p < 0.001. (D) Enhancement of Huh7 cell migration by USP2a overexpression is restricted by RAB1A knockdown, * p < 0.05, ** p < 0.01, *** p < 0.001

Discussion

Biochemical analyses have shown that the downstream targets of USP2a include several relevant oncoproteins, such as FASN [14], EGFR[16] and Cyclin D1 [18], which indicates that abnormal expression of USP2a may play a role in cancer development. So far, however, it has only been found in prostate cancer [21, 22] and non-small-cell lung carcinoma (NSCLC) [16] that the expression of USP2a is significantly higher than that in its corresponding normal tissues. In addition, it has been found that the expression of USP2a in high-grade gliomas is significantly higher than that in low-grade gliomas [23]. Here, USP2a expression in HCC was first evaluated using a tissue microarray, and we found that its expression in HCC tissues was significantly increased compared to that in matched normal adjacent tissues. Previously, it has been reported that USP2a expression may be regulated by the Akt signaling pathway, since USP2a protein expression was increased significantly after treatment of cancer cells with an activator of this pathway [24]. It is generally accepted that activation of the Akt signaling pathway is a common event in HCC [25], and this could be the reason for the increased expression of USP2a in HCC. Additional studies have revealed that the anomalous expression of USPs in cancer cells may be related to dysregulation of upstream microRNAs. Increased expression of USP4 in HCC was, for example, found to result from dysregulation of miRNA-148a [26], providing a hint for investigating USP2a expression in HCC.

It has been reported that USP2a can promote malignant proliferation of prostate cancer [14] and NSCLC [16] cells, and promote the migration and invasion of breast cancer cells [27]. To explore the biological function of USP2a in HCC, we examined USP2a expression in several HCC cell lines compared to the human normal hepatocyte cell line L02. USP2a was found to be highly expressed in most of the HCC cell lines tested, which was consistent with the results from the primary HCC tissues. Next, using a lentivirus technology, we obtained stable cell lines exhibiting USP2a silencing or overexpression. Using both in vitro and in vivo experiments, we found that USP2a silencing led to a significant inhibition of the proliferation, migration and invasion of HCC cells, and that exogenous USP2a overexpression led to opposite effects. Combining these results with the observation that USP2a expression in HCC was significantly increased and positively associated with a poor prognosis, we conclude that USP2a acts a oncoprotein in HCC, similar to its role in prostate cancer, NSCLC and breast cancer. It has, however, also been reported that USP2a may play an apoptotic role in glioblastomas and inhibit their growth [28], suggesting that USP2a may also elicit anti-cancer effects. Thus, USP2a appears to play different roles in different cancers and, thus, should be differentiated according to the specific type of cancer or even its different stages. USP2a as a deubiquitinase has many downstream targets. When its targets are oncoproteins, it may play a role mainly in cancer promotion, but when its targets are anti-oncoprotein(s), it may play a role mainly in cancer suppression. Notably, most of the USP2a downstream targets revealed so far are key oncoproteins, so its role appears to be most commonly to promote cancer.

As already stated above, RAB1A is known to mediate dynamic vesicular transport between the ER and Golgi complex [29]. As these systems play major roles in the biosynthesis/transport of proteins and lipids, the vesicular transport is highly dynamic and rapidly changes in response to nutrient conditions and metabolic demands [30]. In addition, this transport system has been found to be involved in the synthesis, modification and transport of receptors on the cell surface. Therefore, it also acts as the main center of cellular signal regulation and transmission in response to extracellular and intracellular stimuli, such as nutrient availability and ER stress [31]. Through the catalysis of guanine nucleotide exchange factors, RAB1A changes from its inactive GDP-bound form to its active GTP-bound form, after which it recruits effectors - including, but not limited to, tethering factors, motor proteins and SNAREs - to regulate vesicular targeting and fusion to the Golgi complex [32]. An increasing number of studies has shown that RAB1A can lead to abnormal intracellular signal transduction and, thereby, result in the development of various diseases, including cancer [33]. RAB1A has, for example, been found to be highly expressed in colorectal cancer and to promote the proliferation and invasion of cancer cells by activating the amino acid (AA)-mTORC1 signaling pathway [34]. In addition, RAB1A has been found to be overexpressed in different histological subtypes of lung cancer compared to normal lung tissue, and the expression level of RAB1A has been found to be related to the size and stage of lung cancer [35]. It is noteworthy that RAB1A has also been reported to be highly expressed in HCC and to promote the proliferation and metastasis of HCC cells through the AA-mTORC1 signaling pathway [36]. Although the reason behind anomalous RAB1A expression in cancer remains poorly understood, recent studies have suggested that it may be related to abnormal regulation by long non-coding RNAs or microRNAs at the transcriptional level [37–40], but whether there is a relation to abnormal posttranslational modification has so far not been reported. Here, we performed IP assays combined with LC-MS/MS and found that USP2a can interact with RAB1A in HCC cells. Further proteomics studies revealed that USP2a can deubiquitinate RAB1A and increase its stability. Thereby, we provide a novel mechanism for abnormal expression of RAB1A in cancer. As mentioned above, increased RAB1A expression promotes tumor proliferation, migration and invasion. Therefore, we hypothesized that RAB1A may play a crucial role in USP2a-induced HCC progression. Function rescue experiments revealed that USP2a silencing significantly inhibited HCC cell growth, migration and invasion, which could be reversed by RAB1A overexpression. Accordingly, we found that RAB1A knockdown abrogated the functional consequences of USP2a in HCC cells. These data indicate that RAB1A is required for the USP2a-induced malignant behavior of HCC cells. Notably, this restoration or block of functions by RAB1A was not complete, indicating that USP2a may also promote HCC progression through other targets. In this respect, the mass spectrometry experiments performed here may provide important clues and, therefore, we will continue to identify new targets of USP2a in HCC.

In conclusion, we identified USP2a as a key protein in facilitating HCC progression. A critical mechanism underlying the role of USP2a in HCC appears to be deubiquitination and stabilization of RAB1A. Our findings may provide new avenues for the treatment of HCC.

Compliance with ethical standards

Conflict of interest

The authors declare no conflict of interest.