Elevated expression of Erbin destabilizes ERα protein and promotes tumorigenesis in hepatocellular carcinoma

Accepted Manuscript Elevated expression of Erbin destabilizes ERα protein and promotes tumorigenesis in hepatocellular carcinoma Hua Wu, Su Yao, Shen Zhang, Jing-Ru Wang, Peng-Da Guo, Xiu-Ming Li, Wen-Juan Gan, Lin Mei, Tian-Ming Gao, Jian-Ming Li PII: DOI: Reference:

S0168-8278(17)30067-3 http://dx.doi.org/10.1016/j.jhep.2017.01.030 JHEPAT 6414

To appear in:

Journal of Hepatology

Received Date: Revised Date: Accepted Date:

14 October 2016 23 January 2017 29 January 2017

Please cite this article as: Wu, H., Yao, S., Zhang, S., Wang, J-R., Guo, P-D., Li, X-M., Gan, W-J., Mei, L., Gao, T-M., Li, J-M., Elevated expression of Erbin destabilizes ERα protein and promotes tumorigenesis in hepatocellular carcinoma, Journal of Hepatology (2017), doi: http://dx.doi.org/10.1016/j.jhep.2017.01.030

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Elevated expression of Erbin destabilizes ERα protein and promotes tumorigenesis in hepatocellular carcinoma Hua Wu1, 2#, Su Yao3, 4#, Shen Zhang2, Jing-Ru Wang2, Peng-Da Guo 2, Xiu-Ming Li2, Wen-Juan Gan5, Lin Mei6 , Tian-Ming Gao7,* and Jian-Ming Li1,2,3* 1

Department of Pathology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou

510120, China; 2

Department of Pathology, Soochow University, Suzhou 215123, China;

3

Department of Pathology, Nanfang Hospital, Southern Medical University, Guangzhou 510515,

China 4

Department of Pathology, Guangdong General Hospital and Guangdong Academy of Medical

Sciences, Guangzhou 510080, China; 5

Department of Pathology, The First Affiliated Hospital of Soochow University, Suzhou 215006,

China; 6

Department of Neuroscience and Regenerative Medicine, Georgia Regents University, Augusta,

GA 30912, USA 7 Department of Neurobiology, School of Basic Medical Sciences, Southern Medical University, Guangzhou 510515, China #

These authors contributed equally to this work.

*

Correspondence to:

Jian-Ming Li, M.D., Ph.D. Department of Pathology Sun Yat-sen Memorial Hospital Sun Yat-sen University Guangzhou 510120, China Phone: +86 512 65882673 Fax: +86 512 65882673 E-mail: [email protected]

Tian-Ming Gao M.D., Ph.D. Department of Neurobiology School of Basic Medical Sciences Southern Medical University Guangzhou 510515, China Phone: +86-20-62789362 Fax: 86-20-62789362 E-mail: [email protected] 1

Key Words: Hepatocellular Carcinoma, Tumorigenesis, ERα Signaling, Tamoxifen therapy, Mouse Model Abbreviations: HCC (hepatocellular carcinoma), ERα (estrogen receptor-α), IHC (immunohistochemistry),

DEN

(diethylnitrosamine),

MTT

(3-(4,

5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide ) Electronic word count: 7494 Number of figures: 7 Conflict of interest statement: The authors have no conflict of interest to disclose. Financial support statement: This work was supported by National Natural Science Foundation of China (81525020, 31570753, 81272300 to J-M.L., 81372574, 31300630, 31540036 to H.W., and 81401942 to W-J.G), and Natural Science Foundation of Jiangsu Province (BK20130337 to H.W. and BE2016666 to J-M.L). This work was also supported by a project funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions. Authors contributions: HW and JML designed and supervised the study; HW, SY, SZ, JRW, PDG and XML jointly performed the experiments; XML contributed to the clinical sample collection and pathological analysis; JRW contributed to immunohistochemistry analysis in mice; SZ and PDG helped with the animal study; LM and TMG contributed to critical material support; HW, SY, TMG and JML analyzed, interpreted the data and drafted the manuscript.

2

Abstract Background & Aims: Aberrant estrogen receptor-α (ERα) expression and signaling are implicated in the development of hepatocellular carcinoma (HCC), but its regulation in HCC remains enigmatic. Herein, we aimed to identify a new mechanism by which ERα signaling is regulated in HCC, which may lead to a potential new strategy for HCC therapy. Methods: Expression levels of` Erbin and ERα in human HCC samples were evaluated by immunohistochemistry. In vitro and in vivo experiments were used to assess the effect of Erbin and ERα signaling on HCC cells growth. Crosstalk between Erbin and ERα signaling was analyzed by molecular methods. Animal models of diethylnitrosamine (DEN) or DEN/CCl4-induced HCC in wild-type Erbin +/+ and mutant Erbin∆C/∆C mice were observed. The regulatory effects of Erbin on tamoxifen treatment of HCC were evaluated in vitro and in vivo. Results: Erbin inactivated ERα signaling to drive tumorigenesis of HCC, acting to enhance binding of CHIP to ERα via its interaction with ERα and thereby promoting ubiquitination and degradation of ERα. Deletion of the PDZ domain of Erbin in Erbin ∆C/∆C mice, disrupted the interaction of CHIP and ERα, increased the stability of ERα protein, and thus inhibited tumorigenesis of HCC. Silencing of Erbin effectively sensitized the response of HCC after tamoxifen treatment in vitro and in vivo. Conclusions: Our data uncovered an important role of Erbin in regulating HCC tumorigenesis through inactivating ERα-mediated tumor-suppressive signaling, suggesting a new strategy for tamoxifen therapy in HCC by targeting Erbin/ERα 3

signaling axis. Lay summary: Erbin expression is significantly elevated in human hepatocellular carcinoma (HCC) tissue. This elevated expression of Erbin contributes to tumorigenesis of HCC by negatively regulating ERα signaling. However, restoring ERα signaling by inhibiting Erbin expression enhances the sensitivity of HCC cells to tamoxifen treatment, providing a new approach for tamoxifen treatment in HCC.

4

Introduction Hepatocellular carcinoma (HCC) is one of the most prevalent cancers in the world. Epidemiologically, the incidence of HCC in males is 2- to 11-fold higher than in females [1], and the male patients with HCC usually have a poorer prognosis than female patients with HCC [2], suggesting the roles of sex hormones and their receptors in HCC. Estrogen receptor-α (ERα), a member of the nuclear hormone receptor subfamily, has been implicated in the development of HCC. The elevation in the activity of the ERα signaling contributes to the lower viral loads and reduces the risk of HCC by repressing transcription of HBV genes [3, 4]. It is also noteworthy that administration of estradiol (E2) in wild-type (WT) male mice, but not in ERα-deficient (ERα-/-) male mice, exerts significant protection against diethylnitrosamine (DEN)-induced hepatocarcinogenesis [5], indicating that the ERα signaling plays a protective role in attenuating HCC development. The anti-estrogenic compound tamoxifen has been shown to inhibit hepatocyte proliferation in vitro and in vivo [6], and several clinical trials have shown that tamoxifen therapy can improve the survival of advanced HCC patients [7, 8]. Unfortunately, in a large clinical study, tamoxifen therapy did not improve the survival of patients with advanced HCC [9]. All these controversial studies suggest that better understanding of the key regulator and mechanisms responsible for ERα signaling in HCC is necessary. Erbin, a member of the leucine-rich repeat and PDZ domain protein, was originally described as Erbb2-interacting protein (ERBB2IP) [10]. Accumulating 5

studies have suggested that Erbin plays a critical role in regulating cancer cells proliferation, apoptosis, and inflammatory response [11-15]. The role of Erbin in cancer is very complex, and is dependent on specific cell types and contexts. On the one hand, Erbin can function as a tumor suppressor by negatively regulating TGFβ or ERK signaling [16-18]. Loss of Erbin expression in cervical cancer contributes to anoikis resistance of cervical cancer cells by inhibiting STAT3 signaling [19]. On the other hand, Erbin can act as a tumor promoter. Silencing of Erbin expression inhibits the formation of multicellular tumor spheroids in HT-29 colon carcinoma cells [20]. Notably, our recent data revealed that Erbin was required for ErbB2-dependent breast tumor formation and progression [21] or EGFR-dependent colorectal tumorigenesis [13]. Nevertheless, the role of Erbin in HCC development remains unknown, and it is further unclear whether Erbin is involved in regulating ERα signaling in HCC. Herein, we demonstrate that Erbin exerts oncogenic effects in the tumorigenesis process of HCC. Mechanistically, Erbin inactivates ERα-mediated tumor-suppressive signaling via its interaction with ERα, which facilitates Chip binding to, and ubiquitination of ERα, and thereby promoting ubiquitination and degradation of ERα. Importantly, our study suggests a new strategy by targeting Erbin/ERα signaling axis and sensitizing the response of tamoxifen treatment in HCC.

6

Materials and Methods Animal maintenance and treatments All mice were housed in a pathogen-free facility under 12 hour light-dark cycle with free access to food and water. To evaluate in vivo tumor growth, 1.5× 106 cells were subcutaneously injected into nude mice (BALB/c, male, 16-18 g, 4-5-week old). Body weight and tumor sizes were measured weekly. The mice were sacrificed after 6 weeks, and their tumors were removed for assessments. For drug treatment, mice (n = 6 per group) were treated intraperitoneally after 1 week of xenografts with olive oil or tamoxifen (10 mg/kg) once every other day (eleven injections in total). Tumor sizes of mice were measured every four days. One month later, mice were killed and the tumors removed for assessments. To evaluate hepatocarcinogenesis, wild-type (Erbin +/+) and Erbin ∆C/∆C mice (male, n = 8 per group) were injected intraperitoneally with 25 mg/kg of dithylnitrosamine (DEN, Sigma) at 14 days of age. The mice were killed at 9 months of age, and liver tissues were harvested for tumor evaluation and standard histopathological studies. For DEN/CCl4-induced hepatocarcinogenesis, WT (Erbin +/+) and Erbin ∆C/∆C mice (male, n = 8 per group) as indicated were injected intraperitoneally with either DEN (25 mg/kg) or CCl4 (1.2 mg/kg), and killed at the indicated time. All animal experiments were approved by the Animal Care and Use Committee of Soochow University. Human HCC samples Human HCC samples were obtained from Outdo Biotech Co., Ltd (Shanghai, China). These samples were collected from 2002 to 2009. The clinical characteristics of all 7

patients are listed in Supplementary Table 1. This study was approved by Soochow University for Biomedical Research Ethics Committee, and all of the patients provided informed consent. Statistical analysis The relationships between Erbin expression and clinicopathological factors were analyzed using Pearson’s chi-square test, and the correlations between the expression levels of Erbin and ERα, C-myc were calculated using Spearman's rank Correlation test. The Kaplan-Meier survival analysis was used to illustrate the prognostic relevance of Erbin in univariate analysis. Each assay was performed in three independent experiments. All of the data were presented as mean ± standard error of the mean. Statistical analyses were conducted using SPSS package (version 18.0). Statistical comparisons between two groups were performed using a Student's t test. One-way ANOVA was used to compare the differences among more than two groups. p < 0.05 was considered statistically significant. Protocols for other procedures are provided in the Supplementary methods.

8

Results Erbin is highly expressed in HCC and correlates with tumor size, differentiation and poor survival of HCC patients. To determine the role of Erbin in the development of HCC, we first investigated the expression profiles of Erbin in human HCC cell lines and clinical HCC samples. Western blotting analysis showed that Erbin displayed a high expression level in most HCC cell lines (Fig. 1A). In the human HCC samples, we found higher level of Erbin transcripts and protein expression in primary HCC tumors than matched surrounding tissues (Supplementary Fig. 1A, B). Immunohistochemistry (IHC) analysis further confirmed that Erbin was highly expressed in most HCC tissues compared with the matched surrounding tissues of HCC (Fig. 1B, C). Moreover, Erbin expression was increased with the advance of clinical grades of HCC (Fig. 1D, E). In line with these findings, increased Erbin protein levels in HCC tumors were significantly correlated with the clinicopathological parameters such as tumor size, AJCC stage, and differentiation (Supplementary Table 1), thereby indicating a potential critical role of Erbin in the development of HCC. We also assessed the correlation between Erbin expression and HCC patients’ prognosis. Tissue microarrays from 131 patients with HCC were examined by IHC with Erbin antibody, and the results revealed that patients with high Erbin expression had a worse overall survival (OS) than those with low Erbin expression (Fig. 1F). The median OS time of HCC patients with high Erbin expression was approximately 7 months, which was markedly shorter than those with low Erbin expression (~ 24 months). These results suggested that Erbin expression may be a valuable prediction factor for HCC prognosis. Erbin promotes HCC cells proliferation in vitro and tumor growth in vivo. 9

To determine the significance of the above clinical findings, we sought to generate cells with stable knockdown of Erbin protein using a lentiviral shRNA technique in SMMC-7721 and BEL-7402 cells which express high levels of Erbin protein.

The

knockdown

efficiency was

confirmed

by Western

blotting

(Supplementary Fig. 2). The ability of cell proliferation was firstly monitored by Cell Counting Kit-8 (CCK-8) assay, and the results showed that silencing of Erbin expression greatly inhibited the proliferative ability of HCC cells (Fig. 2A). In agreement with this result, colony formation assays demonstrated that knockdown of Erbin expression dramatically repressed the ability of SMMC-7721 and BEL-7402 cells to form colonies (Fig. 2B). We also injected the same number of shRNA/Control or shRNA/Erbin cells into different flanks of nude mice and examined tumor growth in vivo by monitoring tumor sizes every week. Results showed that depletion of Erbin expression significantly impaired tumor growth indicated by the reduced size and weight of tumors derived from shRNA/Erbin cells than shRNA/Control cells (Fig. 2C, D). We further performed a gain-of-function study and investigated the effects of Erbin overexpression on the proliferation and growth of HCC cells. Huh-7 cells that express a low level of Erbin protein was used in this study. The overexpression efficiency was confirmed by Western blotting (Supplementary Fig. 3A). In colony formation assays, we found that ectopic expression of Erbin greatly enhanced the ability of Huh-7 cells to form foci (Supplementary Fig. 3B, C). Consistent with these results, our in vivo observations showed that overexpression of Erbin remarkably promoted the growth of xenograft tumors indicated by the increased growth rate of Huh-7/Erbin tumor cells than control cells (Supplementary Fig. 3D), and the sizes and weight of tumors derived from Huh-7/Erbin cells were significantly increased than the 10

control cells (Supplementary Fig. 3E). Together, these results suggested that overexpression of Erbin in HCC cells contributed to the pro-tumorigenic effect of HCC as indicated by increased proliferation and growth. To investigate the mechanism responsible for Erbin-mediated pro-proliferation effect in HCC, cell cycle distribution was evaluated using flow cytometry assay. The results showed that siRNA/Erbin-transfected SMMC-7721 cells resulted in redistribution of cell cycle progression with an increased proportion of cells in G1 arrest compared to siRNA/control-transfected cells (Supplementary Fig. 4). These results suggested that Erbin promoted cell growth by enhancing cell cycle progression. Erbin regulates ERα protein expression in HCC cells. Aberrant ERα signaling is involved in the development of HCC [22, 23]. To explore the possible effects of Erbin on ERα signaling, we first investigated whether Erbin affected ERα expression. Interestingly, silencing of Erbin expression significantly enhanced endogenous ERα protein expression in SMMC-7721 and BEL-7402 cells (Fig. 3A). Inversely, overexpression of Erbin in Huh-7 cells greatly inhibited endogenous ERα protein expression (Fig. 3B). Similarly, ectopic Erbin expression in HEK293T cells markedly decreased exogenous ERα protein expression in a dose-dependent manner (Fig. 3C). These results were further confirmed by in vivo studies showing that Erbin protein expression negatively correlated with ERα protein expression (Supplementary Fig. 5). Given that Erbin could regulate ERα expression, we next sought to determine whether Erbin-mediated ERα expression was functionally associated with the proliferation of HCC cells. Colony formation assays revealed that silencing endogenous Erbin with siRNA in SMMC-7721 cells significantly reduced cell 11

proliferation by approximately 25%, which could be reversed by knockdown of ERα expression (Supplementary Fig. 6A, B). Together, these results revealed functional significance of Erbin-mediated ERα signaling in HCC development. Erbin inhibits ERα signaling through destabilizing ERα protein expression. Our next task addressed the manner in which Erbin regulated ERα expression. HCC cells were transfected with siRNA to decrease Erbin expression or Erbin-expressing vector to increase Erbin expression, and then ERα mRNA levels were examined by qPCR. However, we did not observe any significant change in the ERα mRNA levels (Supplementary Fig. 7A), indicating that Erbin may modulate the expression of ERα at the post-transcriptional level. Indeed, Erbin-induced ERα protein reduction was greatly reversed when HCC cells were treated with MG132, an inhibitor of proteasome pathway (Supplementary Fig. 7B). To examine whether Erbin affected the turnover of ERα protein, we treated HCC cells with cycloheximide (CHX), an inhibitor of protein synthesis, and monitored ERα expression by Western blotting at various intervals. The results showed that overexpression of Erbin in Huh-7 cells largely decreased the half-life of ERα (Fig. 3D), while silencing of Erbin in SMMC-7721 cells significantly increased the half-life of ERα (Fig. 3E). These results indicated that Erbin promoted the turnover of ERα protein. We next determined whether Erbin was involved in regulating ERα transactivation. Through a ERE-responsive luciferase reporter assay, we found that overexpression of Erbin led to a decrease of ERα transactivation, while silencing of Erbin resulted in the opposite (Fig. 3F). Therefore, these results suggested that Erbin-mediated decrease of ERα expression resulted in a reduction in ERα nuclear activity as a transcriptional factor. Erbin interacts with ERα. 12

To study the underlying mechanisms by which Erbin destabilizes ERα protein, we examined the possibility whether there was an interaction between Erbin and ERα using a co-immunoprecipitation assay. Our results showed that an interaction between endogenous Erbin and ERα was observed in Huh-7 cells (Fig. 4A). In agreement with this result, ectopically expressed Myc-Erbin could interact with GFP-ERα in HEK293T cells (Fig. 4B). Furthermore, confocal microscopy revealed that Erbin and ERα were colocalized in the membrane and cytoplasm of HCC cells (Fig. 4C). These data suggested that Erbin could interact with ERα in the cytomembrane and cytoplasm. To further explore the molecular basis of the interaction between Erbin and ERα, we sought to identify which domain of Erbin was required for its interaction with ERα. Interestingly, we found that upon the deletion of the C-terminal PDZ domain, Erbin lost the ability to bind with ERα (Fig. 4D), suggesting that the PDZ domain of Erbin was required for its interaction with ERα. Erbin promotes ERα ubiquitination. Given our above findings that Erbin could interact with and destabilize ERα protein through the proteasome pathway, we next sought to determine whether Erbin affected ERα ubiquitination. To address this, Huh-7 cells were transfected with expression plasmids encoding Myc-Erbin and treated with MG132 to block the degradation of ERα. As expected, overexpression of Erbin dramatically increased the ubiquitination of ERα (Fig. 4E). Consistent with this finding, knockdown of Erbin expression resulted in the decrease of ERα ubiquitination in the presence or absence of MG132 (Fig. 4F). Because our above evidence showed that the PDZ domain of Erbin was indispensable for its interaction with ERα, we speculated whether this domain also contributed to Erbin-mediated ERα ubiquitination. Indeed, WT Erbin greatly 13

promoted ERα ubiquitination, whereas deletion of the C-terminal PDZ domain of Erbin exerted little effect on ERα ubiquitination (Fig. 4G), indicating that the PDZ domain of Erbin was responsible for its driving effect on ERα ubiquitination. We further tested whether the PDZ domain of Erbin was involved in Erbin-driven growth of HCC cells by colony formation assays. The results showed that WT Erbin markedly enhanced the proliferative ability of HCC cells. However, the Erbin mutant with deletion of C-terminal PDZ domain did not significantly affect the growth of HCC cells (Supplementary Fig. 8A, B). Together, these results suggested that the regulatory effects of Erbin on ERα stability and the HCC cells growth were dependent on the PDZ domain of Erbin. Erbin promotes ERα cytoplasmic retention and facilitates Chip binding to and ubiquitination of ERα. To further explore how Erbin affected ERα stability, we studied the effect of Erbin on the localization of ERα by immunofluorescent staining. Surprisingly, we found that silencing Erbin expression in SMMC-7721 cells greatly increased the levels of nuclear ERα compared with the control cells (Fig. 5A), which was further confirmed by cell fractionation assays (Fig. 5B). These results suggested that Erbin promoted cytoplasmic retention of ERα. It was unlikely that Erbin served as an ubiquitination enzyme to directly regulate ERα ubiquitination. Thus, we speculated that some ubiquitination enzymes may be involved in this biological process. Previous studies have revealed that Chip (Carboxyl terminus of Hsc70-interacting protein), an E3 ubiquitin ligase, links ERα for ubiquitination and proteasomal degradation[24]. We therefore tested whether this regulatory mechanism was affected by Erbin. Interestingly, our data showed that the interaction between Chip and ERα was dramatically enhanced by ectopic expression 14

of Erbin in Huh-7 cells (Fig. 5C). Inversely, knockdown of Erbin expression in SMMC-7721 cells significantly impaired their interaction (Fig. 5D). These results suggested a critical role of Erbin in regulation of the interaction between Chip and ERα in HCC cells. We also determined whether Erbin was involved in Chip-mediated ERα ubiquitination and found that the expression of Erbin significantly enhanced Chip-mediated ERα ubiquitination (Fig. 5E). Depletion of Erbin markedly impaired the ubiquitination of ERα (Fig. 5F), indicating that Erbin had a key role in regulating ERα ubiquitination through modulation of the interaction between Chip and ERα. We next evaluated the role of the PDZ domain of Erbin in the regulation of the interaction between Chip and ERα. We found that the co-expression of Erbin significantly enhanced the interaction of Chip and ERα, whereas deletion of the PDZ domain of Erbin did not exert any effect on their interaction (Fig. 5G). Consistently, the ubiquitination of ERα was greatly enhanced by WT Erbin in a dose-dependent way. However, the deletion of the PDZ domain fully disrupted the ability of Erbin to regulate Chip-mediated ERα ubiquitination (Fig. 5H). Therefore, our data indicated that the PDZ domain was required for Erbin-mediated regulation of ERα ubiquitination facilitated by Chip. The PDZ domain of Erbin is indispensable for its oncogenic activity in mouse models of DEN or DEN/CCl4-induced hepatocarcinogenesis. To verify the in vivo regulatory roles of Erbin-mediated ERα signaling in HCC development, we established diethylnitrosamine (DEN) or DEN/CCl4-induced hepatocarcinogenesis model using Erbin ∆C/∆C mice in which the PDZ domain of Erbin required for the interaction between Erbin and ERα had been deleted. In DEN-induced hepatocarconogenesis model, male WT (Erbin+/+) and Erbin mutant 15

(Erbin ∆C/∆C) mice were injected with a single dose of DEN on postnatal day 14, and animals were dissected after 9 months to examine HCC incidences (Fig. 6A). As shown in Fig. 6B, Erbin +/+ and Erbin∆C/∆C mice developed visible hepatic tumor foci. A marked reduction in HCC incidence and dramatic decreases in the number and sizes of liver tumors were seen in Erbin ∆C/∆C mice (Fig. 6C). We further detected the number of proliferating cells using IHC with Ki67 antibody. The result revealed that the number of proliferating cells in Erbin ∆C/∆C mice was significantly reduced as compared with Erbin +/+ mice (Fig. 6D). Similarly, in DEN/CCl4-induced hepatocarcinogenesis model of Erbin +/+ and Erbin∆C/∆C mice injected intraperitoneally with either DEN or CCl4 (Supplemental Fig. 9A), all animals in both Erbin+/+ and Erbin ∆C/∆C groups displayed visible hepatic tumor foci (Supplemental Fig. 9B). However, the number and sizes of liver tumors and Ki67-positive cells in Erbin∆C/∆C mice were significantly decreased compared with Erbin+/+ mice (Supplementary Fig. 9C, D). Together, these results indicated that the PDZ domain of Erbin was required for HCC elicited by chemical carcinogen. Finally, the molecular mechanism by which Erbin regulated HCC development was examined in these mouse models. As shown in Fig. 6E, there were elevated ERα levels in the tumor and surrounding tissues in Erbin∆C/∆C livers, as compared to controls (Fig. 6E), suggesting that ERα signaling was implicated in Erbin-regulated HCC development. Moreover, an endogenous interaction between Chip and ERα was greatly inhibited in the liver samples from the Erbin∆C/∆C mice (Fig. 6F; Supplementary Fig. 9E). We also determined the status of ERα ubiquitination in these mouse models, and found that the ubiquitination of ERα was markedly decreased in the liver samples from the Erbin∆C/∆C mice (Fig. 6G; Supplementary Fig. 9F). Taken together, the above results establish a functional relationship between Erbin and ERα 16

signaling in vivo. Erbin-mediated ERα signaling is involved in regulating gender disparity in the development of HCC We next determined whether Erbin-mediated ERα signaling played a key role in the gender disparity in HCC. In DEN-induced hepatocarcinogenesis model, we found a marked reduction in HCC incidence in WT females (~ 25%), while all male WT mice developed HCC (Supplementary Fig. 10A, B). Interestingly, our results further showed that there was a difference in both Erbin and ERα expression between male and female mice. As shown in Supplementary Fig. 10C, the expression of Erbin was significantly higher in the liver tissues from male WT mice compared with the liver tissues from the female WT mice. However, ERα expression was lower in the liver tissues of male WT mice than female WT mice, which was consistent with the previous report that showed that ERα expression was decreased in human male HCC [25]. We further evaluated the interaction between Chip and ERα in WT male and female mice using a co-immunoprecipitation assay. As shown in Supplementary Fig. 10D, an enhanced interaction between Chip and ERα was observed in the liver tissues prepared from WT male mice compared with WT female mice. Consistently, the status of ERα ubiquitination was greatly increased in the liver samples from the WT male mice (Supplementary Fig. 10E). Together, these results indicate that Erbin-mediated ERα signaling may be involved in regulating gender disparity in the development of HCC. Inverse correlation of Erbin and ERα expression in clinical HCC samples To determine whether our findings had clinical relevance, we examined the expression of Erbin and ERα in clinical HCC samples by IHC analysis. Our results showed that high Erbin expression was associated with low ERα expression in Case 1. 17

Inversely, the low levels of Erbin correlated with the high levels of ERα in Case 2 (Fig. 7A). Interestingly, IHC studies showed that ERα in HCC tissues was predominantly cytoplasmic. The specificity of ERα antibody was confirmed by IHC analysis in breast cancer and ovarian cancer wherein ERα primarily resided in the nucleus (Supplementary Fig. 11), which was consistent with previous observation [26, 27]. Spearman's rank Correlation analysis further revealed that Erbin had a significant inverse correlation with ERα expression (Fig. 7B). Moreover, we examined gene expression data from TCGA (The Cancer Genome Atlas), and found that there was no significant correlation between Erbin and ERα at the mRNA levels. However, the expression of Erbin was correlated with c-myc expression (Fig. 7C), which is a target gene of ERα. Together, these results that showed the negatively correlated expression between Erbin and ERα protein expression, but not mRNA transcript levels, strengthen our findings that Erbin regulates ERα expression occur at the protein level. Silencing of Erbin restores the sensitivity of HCC cells to tamoxifen treatment. Tamoxifen, an estrogen receptor blocker, has previously been used for HCC therapy in several clinical trials. However, the effects of tamoxifen on the treatment of HCC patients is not satisfactory[28]. Given that Erbin has a critical role in regulating ERα protein stability in HCC, we sought to examine whether the expression of Erbin affected the effect of tamoxifen therapy. To achieve this, HCC cells including Huh-7, QGY-7703, SMMC-7721, and HepG2 cells were treated with tamoxifen and then the proliferative ability of those cells were examined by colony formation assays. Interestingly, we found that tamoxifen treatment significantly suppressed the proliferation of Huh-7 and QGY-7703 cells, which express a low level of Erbin protein. However, the ability of proliferation was not significantly affected by tamoxifen treatment in SMMC-7721 and HepG2 cells, which express a high level of 18

Erbin protein (Supplementary Fig. 12A, B). Thus, we speculate that the expression of Erbin may be involved in the sensitivity of HCC cells to tamoxifen treatment. To address this hypothesis, we treated SMMC-7721 cells that stably express shRNA/Erbin

(SMMC-7721/shRNA/Erbin)

or

shRNA

vector

(SMMC-7721/shRNA/Control) with tamoxifen, and surprisingly found that tamoxifen treatment markedly inhibited the proliferation of SMMC-7721/shRNA/Erbin cells, while SMMC-7721/shRNA/Control cells exhibited reduced sensitivity to tamoxifen treatment (Fig. 7D). To further confirm this, we performed xenograft in vivo experiments, and found that tamoxifen effectively inhibited the tumor growth in nude mice with Erbin-silencing tumors but not in the nude mice with control tumors (Fig. 7E, F). Thus, these results suggested that Erbin overexpression confers the resistance of HCC cells to tamoxifen treatment. We next determined whether ERα downregulation played a critical role in Erbin-mediated resistance to tamoxifen. Our results from CCK-8 assay showed that knockdown of Erbin enhanced HCC cells sensitivity to tamoxifen treatment. However, the effect of knockdown of Erbin on the sensitivity of the cells to tamoxifen was greatly impaired when the expression of ERα was knocked down (Supplementary Fig. 13). Thus, these results indicated that the downregulation of ERα is involved in Erbin-mediated tamoxifen resistance.

19

Discussion Primary liver cancer is the third most common cause of cancer-related death in the world, and the second in China [29, 30]. However, few therapeutic strategies increase survival rates for patients with advanced disease. Therefore, to elucidate the molecular mechanism underlying HCC tumorigenicity is vital to the development of effective HCC therapies. In this study, we demonstrate for the first time that Erbin is critically involved in human and mouse liver cancers. Erbin was overexpressed in HCC cell lines and clinical HCC samples. Analyzing the association of Erbin expression with pathological characteristics in clinical HCC specimens revealed a significant correlation of Erbin expression with tumor size, AJCC stage, and differentiation. The Kaplan-Meier analysis showed that overexpression of Erbin was significantly correlated with poor survival. These data suggested that Erbin overexpression may be involved in the development of HCC. Indeed, in vitro and in vivo experiments revealed that silencing Erbin expression in HCC cells markedly inhibited cancer cells proliferation and growth, indicating the potential oncogenic function for Erbin in regulating HCC development. Consistent with these results, recent findings, including those from our group, show that Erbin functions as tumor promoter to drive breast, colorectal and skin tumorigenesis [12, 13, 21]. Together, these observations establish Erbin as a promoter for tumorigenesis of HCC. Herein, we identified Erbin as a novel and key modulator for ERα signaling in HCC. ERα signaling plays an important role in the regulation of cells proliferation 20

and growth. Dysregulation of this signaling pathway, such as ERα loss or reduction, leads to human cancers, including HCC [31, 32]. ERα deficiency increases HCC development in ERα-/- mice [5]. Inversely, overexpression of ERα inhibits HCC cells growth [33], indicating that ERα signaling plays a protective role in attenuating HCC development. One of our key findings is that we identified Erbin as a novel modulator to downregulate ERα protein through the ubiquitin-proteasome pathway. Erbin functions through its interaction with, and degradation of ERα, to drive HCC cells growth. The interaction between Erbin and ERα results in ERα cytoplasmic retention, which prevents ERα nuclear translocation and function as a transcriptional factor in the nucleus. Silencing Erbin expression in HCC cells promotes ERα nuclear translocation and enhances its nuclear activity, thereby inhibiting HCC cells growth. Recent studies indicate that FoxA1 and FoxA2, two members of the vertebrate forkhead box A (Foxa) gene family of transcription factors, cooperate with estrogen signaling to protect against hepatocarcinogenesis [34]. However, FoxA1 and FoxA2 mRNA and protein levels were not influenced by overexpression or knockdown of Erbin. Coversely, Knockdown of FoxA1 and FoxA2 in HCC cell lines also did not influence the levels of Erbin (Data not shown). These results suggested that the regulation of Erbin to ERα signaling may be the critical mechanism involved in hepatocarcinogenesis, which is independent on regulatory effects of FoxA1 and FoxA2. Although Erbin does not have the activity of ubiquitinating enzyme, Erbin has the ability to promote ubiquitination and binding of Chip to ERα via its cytoplasmic interaction with ERα, which finally leads to ERα degradation through proteasome 21

pathway ( Fig. 5I). In DEN-induced hepatocarcinogenesis model, the expression of Erbin and ERα exhibited HCC gender disparity. Compared with the female mice, the expression of Erbin was significantly higher in liver tissue of male mice, and an enhanced interaction of Chip with ERα and increased ERα ubiquitination was observed in the male mice. Thus, Erbin-mediated ERα signaling may be involved in regulating gender disparity in the development of HCC. The regulatory mechanism may help explain why males are prone to liver cancer than females. Our findings further demonstrate that the PDZ domain of Erbin was responsible for Chip-mediated ERα ubiquitination and degradation at the molecular level, suggesting that the PDZ domain of Erbin may be a potential target for HCC therapy. Accumulating studies have evidenced the role of PDZ domain of Erbin as crucial for epithelial homeostasis and cancer development. For example, Erbin interacts with ErbB2 in a PDZ domain-dependent manner to regulate apical-basolateral polarity of epithelial cells and ErbB2-dependent breast tumorigenesis [10, 21]. The Erbin PDZ domain binds with p0071, which is required for maintaining the integrity of epithelial cell monolayer [35]. We recently reported that the interaction between Erbin and C-cbl mediated by the PDZ domain of Erbin was indispensable for preventing c-cbl-mediated ubiquitination and degradation of EGFR, which contributes to colorectal tumorigenesis [13]. Consistent with these studies, our observations suggested that the PDZ domain of Erbin was required for HCC development. Erbin physically interacts with ERα through its PDZ domain in the cytoplasm, which may confer Chip more opportunity to bind to, and ubiquitinate ERα. Disruption of this 22

interaction by deleting the PDZ domain of Erbin potently impairs the ability of Chip to bind to and ubiquitinate ERα, and subsequently inhibits HCC cells growth. Here we also provide in vivo evidence that the PDZ domain-containing protein Erbin is responsible for the stability of ERα protein. In DEN or DEN/CCl4-induced hepatocarcinogenesis model, Erbin∆C/∆C mice showed a reduction in HCC tumorigenesis with the low levels of the interaction between Chip and ERα, and decreased ERα ubiquitination in liver tissues, as compared to Erbin+/+ mice, suggesting a critical role of the PDZ domain of Erbin in physiologically regulating ERα signaling in HCC development. Our in vitro and in vivo experiments indicate that knockdown of Erbin expression in HCC cells is likely to benefit from tamoxifen treatment. It has been suggested that the growth of HCC is regulated by estrogens, providing the rationale for evaluating the efficacy of tamoxifen treatment by estrogen receptor blockage. However, the efficacy of tamoxifen therapy in HCC patients is controversial [7, 8, 28]. Considering the functional role of tamoxifen in ER activity, it is easy to speculate that tamoxifen therapy may be beneficial for HCC patients with high ERα expression. Indeed, in some HCC patients, the amount of ERα protein is often downregulated [31, 32]. Even in those patients with high ERα expression, the location of ERα is also very important. In our studies, we found that ERα shows a significant cytoplasmic retention in HCC, which may potentially be a major barrier to tamoxifen therapy in HCC patients. Strikingly, the results presented here show that knockdown of Erbin expression could increase ERα expression and nuclear translocation, which enhanced 23

HCC cells sensitivity to tamoxifen treatment. Thus, we speculate that overexpression of Erbin may be the primary cause of decreased ERα expression or increased ERα cytoplasmic retention in some HCC, which ultimately results in the resistance of HCC cells to tamoxifen therapy. Based on this speculation, restoring ERα expression or nuclear localization by inhibiting Erbin expression may provide a new entry point for restoring tamoxifen sensitivity. In summary, we report an oncogenic role for Erbin as a novel regulator of ERα signaling in HCC. The regulation of this molecular process may provide new approaches of tamoxifen treatment for inhibiting HCC development by restoring ERα expression or nuclear localization.

24

References [1] Yeh SH, Chen PJ. Gender disparity of hepatocellular carcinoma: the roles of sex hormones. Oncology 2010;78 Suppl 1:172-179. [2] El-Serag HB, Mason AC, Key C. Trends in survival of patients with hepatocellular carcinoma between 1977 and 1996 in the United States. Hepatology 2001;33:62-65. [3] Yu MW, Chang HC, Chang SC, Liaw YF, Lin SM, Liu CJ, et al. Role of reproductive factors in hepatocellular carcinoma: Impact on hepatitis B- and C-related risk. Hepatology 2003;38:1393-1400. [4] Wang SH, Yeh SH, Lin WH, Yeh KH, Yuan Q, Xia NS, et al. Estrogen receptor alpha represses transcription of HBV genes via interaction with hepatocyte nuclear factor 4alpha. Gastroenterology 2012;142:989-998 e984. [5] Naugler WE, Sakurai T, Kim S, Maeda S, Kim K, Elsharkawy AM, et al. Gender disparity in liver cancer due to sex differences in MyD88-dependent IL-6 production. Science 2007;317:121-124. [6] Francavilla A, Polimeno L, DiLeo A, Barone M, Ove P, Coetzee M, et al. The effect of estrogen and tamoxifen on hepatocyte proliferation in vivo and in vitro. Hepatology 1989;9:614-620. [7] Farinati F, Salvagnini M, de Maria N, Fornasiero A, Chiaramonte M, Rossaro L, et al. Unresectable hepatocellular carcinoma: a prospective controlled trial with tamoxifen. J Hepatol 1990;11:297-301. [8] Martinez Cerezo FJ, Tomas A, Donoso L, Enriquez J, Guarner C, Balanzo J, et al. Controlled trial of tamoxifen in patients with advanced hepatocellular carcinoma. J Hepatol 1994;20:702-706. [9] Barbare JC, Bouche O, Bonnetain F, Raoul JL, Rougier P, Abergel A, et al. Randomized controlled trial of tamoxifen in advanced hepatocellular carcinoma. J Clin Oncol 2005;23:4338-4346. [10] Borg JP, Marchetto S, Le Bivic A, Ollendorff V, Jaulin-Bastard F, Saito H, et al. ERBIN: a basolateral PDZ protein that interacts with the mammalian ERBB2/HER2 receptor. Nat Cell Biol 2000;2:407-414. [11] Huang H, Song Y, Wu Y, Guo N, Ma Y, Qian L. Erbin loss promotes cancer cell proliferation through feedback activation of Akt-Skp2-p27 signaling. Biochem Biophys Res Commun 2015;463:370-376. [12] Xie CM, Wei D, Zhao L, Marchetto S, Mei L, Borg JP, et al. Erbin is a novel substrate of the Sag-betaTrCP E3 ligase that regulates KrasG12D-induced skin tumorigenesis. The Journal of cell biology 2015;209:721-737. 25

[13] Yao S, Zheng P, Wu H, Song LM, Ying XF, Xing C, et al. Erbin interacts with c-Cbl and promotes tumourigenesis and tumour growth in colorectal cancer by preventing c-Cbl-mediated ubiquitination and down-regulation of EGFR. The Journal of pathology 2015;236:65-77. [14] Liu N, Zhang J, Liu S, Liu Y, Zheng D. Erbin-regulated sensitivity of MCF-7 breast cancer cells to TRAIL via ErbB2/AKT/NF-kappaB pathway. J Biochem 2008;143:793-801. [15] McDonald C, Chen FF, Ollendorff V, Ogura Y, Marchetto S, Lecine P, et al. A role for Erbin in the regulation of Nod2-dependent NF-kappaB signaling. The Journal of biological chemistry 2005;280:40301-40309. [16] Dai F, Chang C, Lin X, Dai P, Mei L, Feng XH. Erbin inhibits transforming growth factor beta signaling through a novel Smad-interacting domain. Mol Cell Biol 2007;27:6183-6194. [17] Dai P, Xiong WC, Mei L. Erbin inhibits RAF activation by disrupting the sur-8-Ras-Raf complex. The Journal of biological chemistry 2006;281:927-933. [18] Huang YZ, Zang M, Xiong WC, Luo Z, Mei L. Erbin suppresses the MAP kinase pathway. The Journal of biological chemistry 2003;278:1108-1114. [19] Hu Y, Chen H, Duan C, Liu D, Qian L, Yang Z, et al. Deficiency of Erbin induces resistance of cervical cancer cells to anoikis in a STAT3-dependent manner. Oncogenesis 2013;2:e52. [20] Dardousis K, Voolstra C, Roengvoraphoj M, Sekandarzad A, Mesghenna S, Winkler J, et al. Identification of differentially expressed genes involved in the formation of multicellular tumor spheroids by HT-29 colon carcinoma cells. Mol Ther 2007;15:94-102. [21] Tao Y, Shen C, Luo S, Traore W, Marchetto S, Santoni MJ, et al. Role of Erbin in ErbB2-dependent breast tumor growth. Proceedings of the National Academy of Sciences of the United States of America 2014;111:E4429-4438. [22] Baldissera VD, Alves AF, Almeida S, Porawski M, Giovenardi M. Hepatocellular carcinoma and estrogen receptors: Polymorphisms and isoforms relations and implications. Med Hypotheses 2016;86:67-70. [23] De Maria N, Manno M, Villa E. Sex hormones and liver cancer. Mol Cell Endocrinol 2002;193:59-63. [24] Fan M, Park A, Nephew KP. CHIP (carboxyl terminus of Hsc70-interacting protein) promotes basal and geldanamycin-induced degradation of estrogen receptor-alpha. Mol Endocrinol 2005;19:2901-2914. 26

[25] Jiang R, Deng L, Zhao L, Li X, Zhang F, Xia Y, et al. miR-22 promotes HBV-related hepatocellular carcinoma development in males. Clinical cancer research : an official journal of the American Association for Cancer Research 2011;17:5593-5603. [26] Pan X, Zhou T, Tai YH, Wang C, Zhao J, Cao Y, et al. Elevated expression of CUEDC2 protein confers endocrine resistance in breast cancer. Nat Med 2011;17:708-714. [27] Halon A, Materna V, Drag-Zalesinska M, Nowak-Markwitz E, Gansukh T, Donizy P, et al. Estrogen receptor alpha expression in ovarian cancer predicts longer overall survival. Pathol Oncol Res 2011;17:511-518. [28] Castells A, Bruix J, Bru C, Ayuso C, Roca M, Boix L, et al. Treatment of hepatocellular carcinoma with tamoxifen: a double-blind placebo-controlled trial in 120 patients. Gastroenterology 1995;109:917-922. [29] Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA: a cancer journal for clinicians 2005;55:74-108. [30] He J, Gu D, Wu X, Reynolds K, Duan X, Yao C, et al. Major causes of death among men and women in China. N Engl J Med 2005;353:1124-1134. [31] Chen JJ, Tang YS, Huang SF, Ai JG, Wang HX, Zhang LP. HBx protein-induced upregulation of microRNA-221 promotes aberrant proliferation in HBVrelated hepatocellular carcinoma by targeting estrogen receptor-alpha. Oncol Rep 2015;33:792-798. [32] Liu WH, Yeh SH, Lu CC, Yu SL, Chen HY, Lin CY, et al. MicroRNA-18a prevents estrogen receptor-alpha expression, promoting proliferation of hepatocellular carcinoma cells. Gastroenterology 2009;136:683-693. [33] Deng L, Yang H, Tang J, Lin Z, Yin A, Gao Y, et al. Inhibition of MTA1 by ERalpha contributes to protection hepatocellular carcinoma from tumor proliferation and metastasis. J Exp Clin Cancer Res 2015;34:128. [34] Li Z, Tuteja G, Schug J, Kaestner KH. Foxa1 and Foxa2 are essential for sexual dimorphism in liver cancer. Cell 2012;148:72-83. [35] Jaulin-Bastard F, Arsanto JP, Le Bivic A, Navarro C, Vely F, Saito H, et al. Interaction between Erbin and a Catenin-related protein in epithelial cells. The Journal of biological chemistry 2002;277:2869-2875.

Author names in bold designate shared co-first authorship 27

Figure Legend Fig. 1. Erbin is frequently upregulated in HCC cell lines and clinical HCC tissues. (A) The expression of Erbin protein in HCC cell lines was determined by Western blotting. (B) Representative images from immunohistochemical staining of Erbin in HCC tissues and the surrounding tissues of HCC. (C) Scatter plot analysis of Erbin levels in 214 HCC tissue samples and their surrounding tissues. Data were analyzed using a two-tailed, paired Student’s t test. **p < 0.01. (D) Representative images from immunohistochemical staining of Erbin expression in HCC tissues with different histological grades (Ⅰ-Ⅲ). (E) Scatter plot analysis of Erbin levels in HCC tissues with different histological grades (Ⅰ - Ⅲ ). Data were analyzed using one-way ANOVA.

**p

< 0.01. (F) The correlation between Erbin expression and overall

survival in HCC patients was assessed by Kaplan-Meier survival analysis. Fig. 2. Silencing Erbin suppresses HCC cells proliferation and growth in vitro and in vivo. (A) The cell proliferation was determined by CCK-8 assays in control or Erbin-silenced HCC cells. (B) Effects of Erbin knockdown in HCC cells on colony formation. The representative images are presented (top), and the relative number of colonies was counted (bottom). (C and D) Effects of Erbin knockdown in SMMC-7721 (C) or BEL-7402 (D) cells on growth in nude mice. The representative images of tumors each group are shown (top), and the volumes of the tumors were measured weekly during the indicated period (bottom). Data were analyzed using a two-tailed, unpaired Student’s t test. *p < 0.05, **p < 0.01. Fig. 3. Erbin downregulates ERα expression in HCC cells. (A) Western blotting showing the expression of endogenous ERα in SMMC-7721 or BEL-7402 cells that transiently express siRNA/Erbin or siRNA/Control. (B) The plasmid of Myc-Erbin was transfected into Huh-7 cells for 36 h, and cell lysates were prepared and analyzed 28

by Western blot using the indicated antibodies (C) Western blotting of exogenous ERα in HEK293T cells co-expressing Erbin and ERα. (D and E) Erbin destabilized endogenous ERα protein. The expression of ERα was detected by Western blotting in Erbin overexpressed Huh-7 cells (D, top) or shRNA-transduced SMMC-7721 cells (E, top) that treated with cyclohexamide (CHX) (100 µg/ml) for the indicated time intervals. The intensity of endogenous ERα expression for each time-point was quantified by densitometry (D and E, bottom). (F) Luciferase reporter assay for measuring ERα transactivation in Erbin overexpressed Huh-7 cells (top) or siRNA-transduced SMMC-7721 cells (bottom). Data were analyzed using a two-tailed, unpaired Student’s t test. *p < 0.05, **p < 0.01. Fig. 4. Erbin interacts with ERα and promotes ERα ubiquitination. (A) Immunoprecipitation assay for the endogenous interaction between Erbin and ERα in Huh-7 cells. (B) Immunoprecipitation assay for the exogenous interaction between Erbin and ERα in HEK293T cells transfected with Myc-Erbin and GFP-ERα. (C) Confocal microscopy showing colocalization of endogenous Erbin and ERα in BEL-7402 and SMMC-7721 cells. (D) The PDZ domain of Erbin is required for its interaction with ERα. A schematic representation of the Erbin deletion mutants (top); the interaction of wild-type Erbin or mutant Erbin (Erbin△PDZ) and ERα was evaluated with immunoprecipitation assay (bottom). (E and F) Ubiquitination assays of endogenous ERα in the lysates from Huh-7 cells transfected with Myc-Erbin (E) or SMMC-7721 cells stably expressing Erbin shRNA (shRNA/Erbin) or shRNA (shRNA/Control) (F). The cells were treated with or without MG132 (20µM) before harvesting and then immunoprecipitation with anti-ERα antibody. (G) Ubiquitination assays of exogenous ERα in the lysates from HEK293T cells cotransfected with GFP-ERα, HA-Ub and Myc-Erbin or Myc-Erbin△PDZ. 29

Fig. 5. Erbin promotes ERα cytoplasmic retention and enhances Chip interacting with and ubiquitination of ERα. (A) Immunofluorescent staining of ERα (red) in SMMC-7721 cells stably expressing Erbin shRNA (shRNA/Erbin) or shRNA (shRNA/Control). Nuclei were stained with DAPI (blue). (B) Subcellular fractionation analysis of Erbin and ERα expression in SMMC-7721 cells stably expressing

Erbin

shRNA

(shRNA/Erbin)

or

shRNA

(shRNA/Control).

Immunoblotting of GAPDH and Lamin B1 served as controls for the purity of cytoplasmic

(C)

and

nuclear

(N)

fractions,

respectively.

(C

and

D)

Immunoprecipitation assay for the endogenous interaction between Chip and ERα in Huh-7 cells transfected with Myc-Erbin (C), or SMMC-7721 cells stably expressing Erbin shRNA (shRNA/Erbin) or shRNA (shRNA/Control) (D). (E and F) Ubiquitination assays of endogenous ERα in the lysates from Huh-7 cells cotransfected with Myc-Erbin and myc-Chip (E) or Chip overexpressed SMMC-7721 cells stably expressing Erbin shRNA (shRNA/Erbin) or shRNA (shRNA/Control) (F). (G) Immunoprecipitation assay for the exogenous interaction between Chip and ERα in HEK293T cells cotransfected with Myc-chip, GFP-ERα and Myc-Erbin or Myc-Erbin△PDZ. (H) Ubiquitination assays of exogenous ERα in the lysates from HEK293T cells cotransfected with GFP-ERα, HA-Ub, Myc-Chip and Myc-Erbin or Myc-Erbin△PDZ. The cells were treated with MG132 (20µM) before harvest and immunoprecipitation with anti-GFP antibody. (I) A model for Erbin action in HCC development through regulating ERα signaling. Fig. 6. The deletion of PDZ domain impairs DEN-induced hepatocarcinogenesis and enhances ERα signaling in Erbin∆C/∆C livers. (A) Mouse model of DEN-induced hepatocarcinogenesis. The male mice were injected intraperitoneally 30

with dithylnitrosamine (DEN, 25 mg/kg) at 14 days of age. The mice were killed at 9 months of age, and then the liver tissues were collected. (B) Gross appearances of representative livers with tumors in Erbin+/+ and Erbin∆C/∆C mice (n = 8 per group). (C) Tumor incidence, tumor size and the number of tumor were determined in in Erbin +/+ and Erbin∆C/∆C mice. (D and E) Immunostaining of Ki-67 (D) and ERα (E) were performed in the liver tissues from Erbin+/+ and Erbin ∆C/∆C mice. (F) Immunoprecipitation assay for the endogenous interaction between Chip and ERα in the liver tissues from Erbin+/+ and Erbin ∆C/∆C mice. (G) Ubiquitination assays of endogenous ERα in the lysates from liver tissues of Erbin +/+ and Erbin∆C/∆C mice. Data were analyzed using a two-tailed, unpaired Student’s t test. *p < 0.05, **p < 0.01. Fig. 7. Inverse correlation between Erbin and ERα in HCC, and silencing Erbin increases sensitivity of HCC cells to tamoxifen. (A) Representative images from immunohistochemical staining of Erbin and ERα in tumors of HCC. (B) Spearman correlation analysis between Erbin and ERα in 120 cases of HCC tissues. (C) Dot plot correlates the mRNA levels of Erbin and ERα (top) or Erbin and c-myc (bottom) in HCC. Statistical significance was assessed by the Spearman rank correlation test. (D) Tamoxifen-mediated cell growth inhibition in control or Erbin-silencing SMMC-7721 cells was evaluated by colony formation assay (left), and statistical analysis is shown (right). (E and F) Effects of tamoxifen on HCC cells growth in nude mice. The volumes of the tumors were measured every 4 days during the indicated period (E), and the representative images of tumors and the average tumor mass of each group were shown (F). Data were analyzed using a two-tailed, unpaired Student’s t test. *p < 0.05, **p < 0.01.

31

Figure 1 B Surrounding T Tissue

A

Erbin

β-actin

Grade Ⅰ

Carcinoma T Tissue

D

100×

C

Grade Ⅲ

Erbin Score

Grade Ⅱ

15

**

10

5

0 Surrounding Tissue (N = 214) 100×

E 15

400×

**

1.0 10

10

5

0 Grade Ⅰ (N = 26)

Grade Ⅱ (N = 130)

Carcinoma Tissue (N = 214)

F

Cum mulative survival

Erbin Score

400×

Grade Ⅲ (N = 58)

0.8 0.6 0.4 0.2 0.0 0.00

20.00

40.00

60.00

Time (Months)

80.00

100.00

Figure 2 A

B

3

** **

1 0.5

**

SMMC-7721 Cells

1.5

** ns

**

1

2

BEL-7402

ntrol shRNA/Con

OD450 Value

2

SMMC-7721

shRNA/Control shRNA/Control shRNA/Erbin shRNA/Erbin

2.5

0

5

6

**

shRNA/Errbin

4

6

Colony forming number

3 Days

3.5

shRNA/Control shRNA/Control shRNA/Erbin

2.5

BEL-7402 Cells

**

2

**

1.5 1 0.5

** ns

*

1

2

0 3

4

5

Days

**

300 200 100 0 shRNA/ Control

C

Colony forming number

OD D450 Value

3

200

**

150 100 50 0

shRNA/ Erbin

shRNA/ Control

D BEL-7402

SMMC-7721 shRNA/ Control

shRNA/ Control

shRNA/ Erbin

250

**

200

**

150 100

**

50

**

0 1

2

3 4 5 6 Weeks shRNA/Control ( n = 8) shRNA/Erbin (n = 8)

Tumor Volume (mm3)

Tum mor Volume (mm3)

shRNA/ Erbin

800 700 600 500 400 300 200 100 0

** ** ** ** 1

2

3 4 5 6 Weeks shRNA/Control ( n = 5) shRNA/Erbin (n = 5)

shRNA/ Erbin

Figure 3 A

SMMC-7721 siRNA/Control siRNA/Erbin #1 siRNA/Erbin #2

+ -

+ -

D

BEL-7402

+

+ -

+ -

+

Myc-Erbin

-

-

+

+

+

+

CHX (Hours） 0

3

6

9

0

3

6

9

ERα

ERα

β-actin

β-actin 1.0

2.2 1.4

1.0

1.4 1.6 Control 系列1

C

B

HEK293T cells

Huh-7 cells -

Myc-Erbin GFP-ERα +

+

Myc-Erbin

+ +

++ +

Myc-Erbin

ERα

GFP-ERα

β-actin

β-actin 1.0

ERα/β-actin

0.3

1.0

0.7

100

**

**

80 60

**

40 20

0 CHX(h）

0.4

E

Erbin 系列2

120 % of ERα protein remaining

ERα/β-actin

ERα/β-actin

-

M E bi Myc-Erbin

Erbin

Myc-Erbin

-

0 1

3 2

36

49

F +

+

+

+

-

-

-

-

shRNA/Erbin

-

-

-

-

+

+

+

+

CHX (Hours）

0

3

6

9

0

3

6

9

Erbin ERα β-actin

shRNA/Control 系列1

shRNA/Erbin 系列2

** Relative ERα trransactivation activity

shRNA/Control

**

20

**

15 10 5 0

ERα Erbin

1-

2+

-

-

*

100 80

* **

60 40 20

0 CHX(h）

10

3 2

63

94

Relative ERα transactivatio on activity

% of ERα α protein rema aining

120

+ 3

4+

++

+

**

1.5 1.2 0.9 09 0.6 0.3 0

siRNA/Control siRNA/Erbin

+1 -

2+

Figure 4 A

B Input

Ip:IgG

Ip:ERα

Input Myc-Erbin GFP-ERα

+ -

IP: GFP + +

+ -

Input GFP-ERα Myc-Erbin

+ +

Erbin

Myc-Erbin

GFP-ERα

ERα

GFP-ERα

Myc-Erbin

C

+ -

IP: Myc + +

+ -

+ +

D BEL-7402 Erbin

1

ERα

DAPI

Merge

391

1279 1371

Erbin Erbin△PDZ Input GFP-ERα Myc-Erbin Myc-Erbin△PDZ Myc Erbin△PDZ

SMMC-7721 Erbin

ERα

DAPI

Merge

+ -

+ + -

IP: Myc + +

+ -

+ + -

+ +

GFP-ERα

Myc

F

E

G

MG132 IP: Myc-Erbin

-

ERα -

shRNA/Control shRNA/Erbin MG132

ERα +

+ +

GFP-ERα HA-Ub Myc-Erbin Myc-Erbin△PDZ

+ +

HA-ub

GFP-ERα

Myc-Erbin

Myc

ERα

GFP-ERα

+ + -

+ + + -

+ + +

Inputt

ERα

Input

ERα

Inputt

Myc-Erbin

+ -

IP:G GFP

ERα

Poly-ub

+ -

IP P:ERα

IP P:ERα

Poly-ub

IgG

Figure 5 B

A shRNA /Erbin

C

C

N

Myc-Erbin

Erbin

IB:ERα

ERα

IB:Chip

Merrge

+

+ +

+ -

GFP-ERα HA-Ub Myc-Chip Myc-Erbin Myc-Erbin△PDZ MG132

+ +

+ +

+ + + +

+ + + + +

+ + + ++ +

IB:HA-Ub

IB:ERα

IB:Erbin

IB:Myc

IB:Myc-Erbin

IB:Myc-Chip Input

Input

IB:GFP-ERα

Inputt

IB:ERα

IB:Myc-Chip

+ + +

IP:GFP-ERα

IP:ERα

IB:poly-Ub

I IP:ERα

IB:poly-Ub

IB:Myc-Chip IB:GFP-ERα

IB:ERα

IB:ERα

I

G

IB:Myc-Erbin IB:GFP-ERα

+ + -

+ + + -

Stimulatory Modification Inhibitory Modification Ubiquitination

+ + +

IB:Myc-Chip IB:Myc Chip

IB:GFP-ERα

Input

IB:Myc-Erbin and Myc-Erbin△PDZ

Tumorigenesis

Degradation

IP:GFP-ERα

IB:Myc-Chip

+ -

Erbin n Expression Level

Myc-Chip GFP-ERα Myc-Erbin Myc-Erbin△PDZ

+ + + ++ +

Input

-

+ -

+ + + + +

IB:Erbin IB:ERα

H shRNA/Control shRNA/Erbin Myc-Chip

+

IB:ERα

IB:ERα

F

+ -

IB:Chip

IB:Erbin

GAPDH

Myc-Erbin Myc-Chip

shRNA/Control shRNA/Erbin IgG IB:Chip

+

IgG IB:Chip

N

Lamin B1

E

-

Input

DAPI

shRNA /Erbin

IP:E ERα

shRNA /Control

D

IP:E ERα

ERα

shRNA /Control

C

Normal Erbin ERα Chip

p Chip ERα

ERα

Erbin

ERα nuclear activity

ERα nuclear activity

Hepatocellular Carcinoma Development

Figure 6 A

B Harvest

DEN

0 14d

Erbin △C/ △C

Erbin +/+

9M

100 80 60

*

40 20 0 Erbin +/+

D

Erbin △C/ △C

Ki-67

25

8.0

Number of tum mors

Maximal tumor size e (mm)

Tumor incidenc ce (%)

C 60 6.0 4.0

**

2.0

20 15

5 0

0.0 Erbin

+/+

Erbin

Erbin +/+

△C/ △C

Erbin △C/ △C

2#

IB:Chip

N T

1#

2# IP:ERα

Erbin +/+

1#

IB: ERα

IB:Chip

N

Input

Erbin △C/ △C

Erbin △C/ △C

F

Ki-67

Erbin +/+

T IB: ERα

100 ×

400 ×

E

Erbin +/+

G HE

ERα

T

T

N

ERα

N

T

N

IB:poly-Ub

T

N

IB: ERα

40×

200×

Input

IB: ERα

40×

Rα IP:ER

Erbin △C/ △C

**

10

Figure 7 A

B

ERα

Case 1 (High h Erbin)

Erbin

Low ERα Expression High ERα Expression R = - 0.288 P = 0.001

Case 2 (Low w Erbin)

%o of Specimens

100%

76.25

60% 40%

52.5

20%

23.75 High

Low 2

1

Erbin Expression

D Vehicle

3000

shRNA/Erbin

er Colony forming numbe

shRNA/Control

P = 0.077

2500 2000 1500

Tamoxifen

1000 500 0 0

1000

2000

3000

4000

Erbin expression (RNA seq)

10000

E shRNA/Control + Vehicle 系列1 shRNA/Erbin + Vehicle 系列2 shRNA/Control + Tamoxifen 系列3 shRNA/Erbin + Tamoxifen 系列4

100

Tumor Volume (mm3)

6000 4000 2000 0 0

1000

2000

3000

Erbin expression (RNA seq)

4000

200 150 100 50 0

shRNA/Control

shRNA/Control + Tamoxifen

shRNA/Erbin

shRNA/Erbin + Tamoxifen

2

3

4

F

P = 0.001

8000

** ns

250

1

120 R = 0.273 N = 140

300

ns

80 60

*

40

Tumor Weigh ht (mg)

R = 0.550 N = 140

3500

ERα expression (RNA seq)

47.5

0%

C

n C-myc expression (RNA seq)

80%

180 160 140 120 100 80 60 40 20 0

* ns

1

2

3

4

20 0 13

27

11 3

15 19 4 5 Days

23 6

27 7 Vehicle

Tamoxifen

Stimulatory Modification Inhibitory Modification Ubiquitination

Tumorigenesis

Erbin Exp pression Lev vel

Degradation

Normal Erbin ERα Chip

Chip ERα

ERα

Erbin

ERα nuclear activity

ERα nuclear activity

Hepatocellular Carcinoma Development