αB-crystallin complexes with 14-3-3ζ to induce epithelial-mesenchymal transition and resistance to sorafenib in hepatocellular carcinoma

Authors

  • Xiao-Yong Huang,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Ai-Wu Ke,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Guo-Ming Shi,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Xin Zhang,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Chi Zhang,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Ying-Hong Shi,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Xiao-Ying Wang,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Zhen-Bin Ding,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Yong-Sheng Xiao,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Jun Yan,

    1. Department of Surgery, Fujian Provincial Tumor Hospital, Teaching Hospital of Fujian Medical University, Fuzhou, P.R. China
    Search for more papers by this author
  • Shuang-Jian Qiu,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Jia Fan,

    Corresponding author
    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    2. Cancer Center, Institutes of Biomedical Sciences, Fudan University, Shanghai P.R. China
    • Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, 200032, P.R. China
    Search for more papers by this author
  • Jian Zhou

    Corresponding author
    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, P.R. China
    2. Cancer Center, Institutes of Biomedical Sciences, Fudan University, Shanghai P.R. China
    3. Shanghai Key Laboratory of Organ Transplantation, Shanghai, P.R. China
    • Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion (Fudan University), Ministry of Education, Shanghai, 200032, P.R. China
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

Abstract

The overall survival of patients with hepatocellular carcinoma (HCC) remains poor, and the molecular pathogenesis remains incompletely defined in HCC. Here we report that increased expression of αB-Crystallin in human HCC predicts poor survival and disease recurrence after surgery. Multivariate analysis identifies αB-Crystallin expression as an independent predictor for postoperative recurrence and overall survival. We show that elevated expression of αB-Crystallin promotes HCC progression in vivo and in vitro. We demonstrate that αB-Crystallin overexpression fosters HCC progression by inducing epithelial-mesenchymal transition (EMT) in HCC cells through activation of the extracellular-regulated protein kinase (ERK) cascade, which can counteract the effect of sorafenib. αB-Crystallin complexes with and elevates 14-3-3ζ protein, leading to up-regulation of ERK1/2 activity. Moreover, overexpression of αB-Crystallin in HCC cells induces EMT progression through an ERK1/2/Fra-1/slug signaling pathway. Clinically, our data reveal that overexpression of both αB-Crystallin and 14-3-3ζ correlates with the HCC poorest survival outcomes, and sorafenib response is impaired in patients with αB-Crystallin overexpression. Conclusion: These data suggest that the αB-Crystallin-14-3-3ζ complex acts synergistically to promote HCC progression by constitutively activating ERK signaling. This study reveals αB-Crystallin as a potential therapeutic target for HCC and a biomarker for predicting sorafenib treatment response. (HEPATOLOGY 2013)

Hepatocellular carcinoma (HCC) is one of the major health problems worldwide.1 Because HCC is often diagnosed at an advanced stage, a large proportion of HCC patients display symptoms of intrahepatic metastases or postsurgical recurrence at the time of diagnosis. HCC is among the leading causes of cancer-related deaths in Asia, especially in China. Therefore, much hope is focused on obtaining a better understanding of the mechanism relevant to this disease in order to develop new preventive, diagnostic, and therapeutic options.

αB-Crystallin (Cryab), a member of the mammalian small heat shock protein (sHsp) superfamily, functions as a cytoprotective molecular chaperone by preventing stress-induced aggregation of denatured proteins and keeping aggregation-prone proteins in reservoirs of nonnative, refoldable intermediates within large, soluble, multimeric structures.2 The expression of Cryab in a temporally and spatially controlled fashion during development has led to the speculation that this sHsp may have important roles in the regulation of cellular physiology and growth.3 Indeed, ectopic expression of Cryab in diverse cell types has been demonstrated to confer protection against a broad range of apoptotic stimuli, including tumor necrosis factor alpha (TNF-α), TNF-related apoptosis-inducing ligand,4 oxidative stress,5 and exposure to anticancer drugs,6 while silencing Cryab expression by RNA interference (RNAi) sensitizes cells to apoptosis.7 Elevated expression of Cryab has been linked to cancer pathology.8 For example, Cryab overexpression has been clinically detected in a variety of human malignancies and is associated with poor prognosis for several different types of tumors, including renal, breast, hepatoma, and lung cancer.9-12 However, the mechanism remains unclear. Given the significant role of Cryab in tumor progression of human cancers, further investigation into the role and mechanism of Cryab in HCC is needed.

Most HCC deaths are due to metastasis, a multistep process that mediates the spread of tumor cells from primary tumors to distant sites.13 Although the molecular and genetic events underlying tumor metastasis are still not well understood, intense investigation into this process has led to the notion that molecules involved in epithelial-mesenchymal transition (EMT) are critical in tumor invasion and metastasis.14 On the other hand, recent studies have shown that a multikinase inhibitor, sorafenib, can inhibit serine/threonine kinases (c-RAF, and mutant and wildtype BRAF) and be used as a “gold” treatment for advanced HCC patients.15 However, the efficiency of sorafenib remains unstable. An investigation to discriminate patients with sorafenib treatment into discrete prognostic groups is lacking. Here, functional and genetic screens demonstrated that Cryab overexpression fosters tumor progression in HCC by inducing EMT by way of activation of extracellular-regulated protein kinase (ERK) signaling, which is subsequently shown to mediate this EMT and sorafenib resistance. Overexpression of αB-Crystallin in HCC cells induces EMT progression through an ERK1/2/Fra-1/slug signaling pathway. Remarkably, these studies demonstrated that Cryab complexes with and elevates 14-3-3ζ protein, leading to up-regulation of ERK1/2 activity. Moreover, high levels of Cryab are associated with impaired sorafenib response with respect to overall survival (OS).

Materials and Methods

Cell Lines, Tissues, and Transfection Experiment.

Five HCC cell lines used in this study, tissue collection, and the transfection experiment are described in the Supporting Materials and Methods. The target sequences are listed in Supporting Table S1.

Cell Matrigel Invasion Assays, MTT Assays, and In Vivo Metastasis Assays.

Matrigel assays and MTT assays were performed as described,16 with slight modification. The metastasis assay is described in the Supporting Materials and Methods.

RNA Expression Analysis, Western Blot Analysis, Immunofluorescence Assay, and Confocal Immunofluorescence.

Quantitative real-time polymerase chain reaction (qRT-PCR), semiquantitative PCR, immunofluorescence, and western blot were performed as described17 and are described in the Supporting Materials and Methods. The primers and antibodies in this study used are listed in Tables S2 and S3.

Construction of Tissue Microarrays and Immunohistochemistry (IHC).

Tissue microarray (TMA) was constructed as described in our earlier study.18 IHC staining for the target genes was carried out on sections of the formalin-fixed samples on the TMA.

Gene Microarrays.

Gene microassay analysis was done as described elsewhere.19 The log ratio of the red to green intensities for each signal was used for statistical analyses. We selected a fold change of 3 as the threshold for significant up-regulation.

Immunoprecipitation (IP) Assays and Immunoisolation of Cryab-Containing Complexes, In-Gel Tryptic Digestion, and Two-Dimensional Liquid Chromatography Coupled with Tandem Mass Spectrometry (2D-LC-MS/MS).

IP assays and immunoisolation of Cryab-containing complexes, in-gel tryptic digestion, and 2D-LC-MS/MS are described in the Supporting Materials and Methods.

Statistical Analysis.

Statistical analysis was performed with SPSS 15.0 software (Chicago, IL). All tests were two-tailed and P < 0.05 was considered statistically significant.

Abbreviations

α-SMA, alpha-smooth muscle actin; BCLC, Barcelona Clinic Liver Cancer; Cryab, αB-Crystallin; EMT, epithelial-mesenchymal transition; ERK, extracellular-regulated protein kinase; Fn 1, fibronectin 1; HCC, hepatocellular carcinoma; OS, overall survival; qRT-PCR, quantitative real-time polymerase chain reaction; RT-PCR, reverse transcription polymerase chain reaction; sHsp, small heat shock protein; shRNA, short hairpin RNA; siRNA, small interfering RNA; TMA, tissue microarray.

Results

Elevated Expression of Cryab Promotes Tumor Progression of HCC In Vivo and In Vitro.

We initially examined Cryab expression in a series of HCC cell lines with stepwise metastatic potential (HCCLM3, MHCC-97L, SMMC-7721, Bel-7402, and Hep3B). The trend of Cryab messenger RNA (mRNA) and protein expression in cancer cells was in line with their metastatic potential (Fig. 1A,B). We then generated a series of human isogenic cell line pairs in which Cryab expression was modified by RNA interference or complementary DNA (cDNA) transfection. We used HCC cell lines (termed HCCLM3-Mock/HCCLM3-vshCryab and Hep3B-Mock/Hep3B-Cryab) to evaluate the effect of Cryab expression on the invasion and motility of cancer cells (Fig. 1C). Matrigel invasion assays showed that decreased Cryab expression resulted in an impairment of the invasive ability of the HCC cells (Fig. 1D).

Figure 1.

High expression of Cryab promotes HCC metastasis and invasion both in vitro and in vivo. (A,B) qRT-PCR, western blot, and immunofluorescence analysis of the Cryab expression in HCC cell lines. (C) Cryab expression in HCCLM3 and Hep3B cells was modified by vshRNA interference and cDNA transfection. (D) The invasion of cancer cells was measured by transwell assays. (E) The volume of the tumors derived from HCC isogenic cell lines was measured in vivo for 6 weeks; serial sections from mouse lungs showed the metastatic ability of cancer cells expressing different levels of Cryab. (H) Survival and recurrence analysis of 403 HCC patients with respect to Cryab expression. Scale bars = 50 μm.

To further investigate the role of Cryab in HCC, we developed orthotopic HCC mouse models. We found that HCC cells with high expression of Cryab resulted in larger tumors and higher rates of lung metastasis (Fig. 1E; Supporting Fig. S1) when compared with cancer cells expressing low levels of Cryab. To elucidate the clinical relevance of Cryab in HCC patients, we analyzed Cryab expression in a TMA of 403 HCC specimens using IHC. The results revealed that Cryab expression was significantly correlated with poor prognosis (Fig. 1F). Cryabhigh accounts for 53.5% of HCC patients. Cryab overexpression was correlated significantly with vascular invasion (P < 0.001), absent tumor encapsulation (P = 0.009), and Barcelona Clinic Liver Cancer (BCLC) staging (P = 0.035) (Table S4). Multivariate analysis identified Cryab expression as an independent predictor for postoperative recurrence and OS (Table 1). Together, these results indicate that high Cryab expression promotes the invasive and metastatic potential of HCC cells.

Table 1. Univariate and Multivariate Analyses of Factors Associated with Survival and Recurrence
FactorsOSCumulative Recurrence
Univariate, PMultivariateUnivariate, PMultivariate
HR95% CIP valueHR95% CIP value
  1. OS, overall survival; NA, not adopted; NS, not significant; AFP, alpha-fetoprotein; HBsAg, hepatitis B surface antigen; 95%CI, 95% confidence interval; BCLC, Barcelona-Clinic Liver Cancer HR, hazard ratio; Cox proportional hazards regression model.

Sex (female vs. male)0.426  NA0.561  NA
Age (years) (≤50 vs. >50)0.563  NA0.520  NA
HBsAg (positive vs. negative)0.054  NA0.368  NA
HCVAb (positive vs. negative)0.708  NA0.962  NA
Child-Pugh classification (A vs. B)0.645  NA0.715  NA
Liver cirrhosis (yes vs. no)0.065  NA0.159  NA
Serum AFP, ng/mL (≤20 vs.>20)0.092  NA0.942  NA
Serum ALT, U/L (≤75 vs. >75)0.916  NA0.273  NA
Tumor size (diameter, cm) (>5 vs.≤5)<0.0011.6111.228-2.1140.0010.035  NS
Tumor encapsulation (absent vs. present)0.001  NS0.119  NA
Vascular invasion (yes vs. no)<0.0011.6251.200-2.1990.0020.001  NS
Tumor number (multiple vs. single)<0.0011.8311.345-2.491<0.001<0.0011.8921.344-2.665<0.001
Tumor differentiation (III/IV vs I/II.)0.404  NA0.877  NA
BCLC staging (0/A vs. B/C)<0.001  NA<0.001  NA
Cryab expression (high vs. low)0.0101.3431.018-1.7720.0370.0011.5751.165-2.1290.003

Overexpression of Cryab Promotes HCC Metastasis by Inducing EMT of Cancer Cells.

Differences in gene expression between cells with high and low Cryab expression were investigated using cDNA microarrays. Of the 41,000 mRNAs, 904 showed at least a 3-fold change in expression between the Hep3B-Cryab cells and the Hep3B-Mock cells (Fig. S2). Based on the association between Cryab expression and the development and progression of cancers in vivo and in vitro, and given that EMT is considered a striking feature of most cancers and plays a crucial role in cancer metastasis and invasion,20 we compared the expression of epithelial and mesenchymal markers as well as other molecules thought to induce EMT in cancer cells. As shown in Fig. 2A, Hep3B-Cryab cells expressed a lower level of the epithelial gene E-cadherin compared to Hep3B-Mock cells. The transcription factor slug and multiple mesenchymal genes (vimentin, fibronectin 1 [Fn 1], alpha-smooth muscle actin [α-SMA], and N-cadherin) were significantly up-regulated in Hep3B-Cryab cells compared with Hep3B-Mock cells. These results were further validated by reverse transcription PCR (RT-PCR) and western blot (Fig. 2B). HCCLM3 is a highly metastatic cell line that expresses a low level of E-cadherin and a high level of vimentin and is therefore thought to present a mesenchymal-like phenotype.21, 22 Interestingly, the level of E-cadherin was higher in HCCLM3-vshCryab than in HCCLM3-Mock, while multiple mesenchymal-associated genes (slug, vimentin, and N-cadherin) were down-regulated in HCCLM3-vshCryab cells (Fig. 2A,B).

Figure 2.

Cryab overexpression induces an EMT in HCC cells. (A) Different expression of epithelial and mesenchymal markers, as well as the transcription factors, were compared between cancer cells with high and low Cryab expression (Hep3B-Mock versus Hep3B-Cryab and HCCLM3-Mock versus HCCLM3-vshCryab cells). (B) RT-PCR and western blot verified the above results of gene scan. (C) Phase-contrast microscopy and immunofluorescent staining for Cryab, E-cadherin, F-actin, vimentin, and slug in Hep3B-Mock, Hep3B-Cryab, HCCLM3-Mock, and HCCLM3-vshCryab. DAPI stain (blue) was used to identify nuclei. (D) Representative examples of hematoxylin and eosin (H&E) and Cryab, E-cadherin, slug, and vimentin immunohistochemistry from Hep3B-Mock/Hep3B-Cryab and HCCLM3-Mock/HCCLM3-vshCryab tumors in nude mice. (E) Representative HCC cases in tissue microarrays (serial sections) were analyzed by H&E and immunohistochemical staining for Cryab, E-cadherin, slug, vimentin. Scale bars = 200 μm.

We further analyzed the morphology of HCC cells with different levels of Cryab expression. As shown in Fig. 2C, a distinct morphological difference was observed between Hep3B-Mock and HCCLM3-Mock cells and the corresponding cells with modified Cryab expression. Hep3B-Mock and HCCLM3-vshCryab cells presented the typical cobblestone-like appearance of normal epithelial cells, while Hep3B-Cryab and HCCLM3-Mock cells took on a spindle-like, fibroblastic morphology. We then performed immunofluorescence to detect the localization and intensity of Cryab and epithelial or mesenchymal marker expression (Fig. 2C). HCCLM3-Mock and Hep3B-Cryab cells revealed little or no detectable E-cadherin. Of note, we detected that tumor tissues derived from Hep3B-Cryab and HCCLM3-Mock cells-derived tumor cells had a fibroblast-like morphology and expressed a high level of slug and a low level of E-cadherin (Fig. 2D). More important, HCC cells overexpressing Cryab had a mesenchymal phenotype in HCC tissues (Fig. 2E). Thus, we conclude that the Cryab overexpression promotes HCC progression by inducing cancer cell EMT.

Cryab Overexpression Induces HCC EMT by Way of Hyperactivity of ERK Signaling and Promotes Resistance to Sorafenib.

We employed a multivariate approach for integrating genome-wide expression data and biological knowledge23 to search for pathways that are dysregulated as a consequence of Cryab overexpression. We used this method to search through functional pathways defined by KEGG and BioCarta in Hep3B-Mock versus Hep3B-Cryab cells. Pathways with significant P values at 0.01 are shown. We identified 28 gene sets in the BioCarta database and 67 gene sets in the KEGG database, with false discovery rates (Q values) < 0.05 by paired t tests between these populations (Table S5).

The magnitudes of MEK, FAK, Src, p38, ERK1/2, p65, and Akt phosphorylation in Hep3B-Mock and HCCLM3-Mock cells were compared to their corresponding control cells. As shown in Fig. 3A, markedly elevated levels of MEK, ERK1/2, and p38 phosphorylation were detected in HCCLM3-Mock and Hep3B-Cryab cells compared with the corresponding control cells, while consistent changes in Src, FAK, and p65 phosphorylation were not observed in the Hep3B-Cryab/Hep3B-Mock and HCCLM3-Mock/HCCLM3-vshCryab cells. Immunofluorescent staining showed that expression of Cryab in Hep3B-Cryab and HCCLM3-Mock cells correlated with high ERK1/2 phosphorylation (Fig. 3B).

Figure 3.

Cryab overexpression induces HCC EMT by way of hyperactivity of ERK signaling and resistance to sorafenib. (A) The phosphorylation of AKT, MEK, Src, p38, ERK1/2, FAK, and p65 were assessed in Hep3B-Mock, HCCLM3-Mock and their control cells. (B) Immunofluorescent staining showed that the expression of Cryab was in line with the activation of ERK1/2. (C) The inhibition of ERK1/2 phosphorylation by U0126 reversed EMT of HCC cells, while P38 inhibitor SD203580 did not. (D) Immunofluorescent staining for E-cadherin, F-actin, slug, and vimentin confirmed that EMT in Hep3B-Cryab and HCCLM3-Mock cells was reversed by U0126. (E) Cancer cells were incubated with various concentrations or with 10 μM of sorafenib for different lengths of time, and the phosphorylation of ERK1/2 was analyzed by western blot. (F) Comparison of overall survival curves between patients with high and low Cryab expression treated with sorafenib.

To identify the signaling pathways that might contribute to the observed phenotypic changes, we blocked MEK/ERK signaling using U0126 or P38 signal using SD203580. Upon ERK1/2 inhibition, the Hep3B-Cryab and HCCLM3-Mock cells presented an epithelial phenotype based on mRNA and protein expression (Fig. 3C) and cellular characteristics, such as decreased E-cadherin and increased Fn 1, vimentin, and F-actin when compared with the parental cells (Fig. 3D). These results indicate that hyperactivation of ERK1/2 appears to be crucial for the observed Cryab-mediated phenotypic characteristics of HCC cells.

Sorafenib, an oral multikinase inhibitor, offers hope for the clinical treatment of several advanced solid cancers by inhibiting intracellular signals in the ERK cascade and blocking receptor tyrosine kinases.17, 24 Here, we tried to assess whether sorafenib inhibited the activity of the ERK cascade induced by Cryab overexpression. As shown in Fig. S3, Hep3B-Mock cells were evidently inhibited compared with Hep3B-Cryab cells when the sorafenib concentration reached 10 or 20 μM (P < 0.05). We further evaluated the changes in the phosphorylation levels of ERK1/2 by western blot in both Hep3B-Mock/Hep3B-Cryab and HCCLM3-Mock/HCCLM3-vshCryab cells after treatment with sorafenib in a dose-dependent or time-dependent manner at the concentration of 10 μM. After treatment with 10 or 20 μM sorafenib, ERK phosphorylation was obviously decreased in both Hep3B-Mock and HCCLM3-vshCryab cells, but only slightly decreased in the Hep3B-Cryab and HCCLM3-Mock cells. Upon treatment with 10 μmol/L sorafenib, a decrease ERK phosphorylation in Hep3B-Mock and HCCLM3-vshCryab cells between 2 hours and 24 hours was seen, but the change was not obviously observed in the Hep3B-Cryab and HCCLM3-Mock cells (Fig. 3E).

Retrospective data from 33 advanced recurrent HCC patients receiving combined sorafenib treatment and transarterial chemoembolization therapy who had undergone liver resection from 2 to 51 months prior to the combined therapy were analyzed. Patient demographics (Table S6) and OS were recorded. Cryab expression was measured in the above 33 HCC tissues (Fig. 3F), and the Kaplan-Meier survival analysis showed that the OS probability of the Cryabhigh group was much lower than that of Cryablow group. Median OS was 9.0 months in the Cryabhigh group and 14.0 months in the Cryablow group (hazard ratio in Cryabhigh group, 3.001; 95% confidence interval, 1.223-7.364; P < 0.05). Thus, we conclude that a high level of Cryab leads to sorafenib resistance in HCC cells.

Proteomic Screen Maps Cryab Interacting Proteins and Identifies Cryab Complexing with 14-3-3ζ.

Signal transduction cascades involve multiple enzymes and are orchestrated by selective protein-protein interactions that are essential for the progression of intracellular signaling events.25, 26 To determine how Cryab activates the MEK/ERK signal, a combination of co-IP and MS was used to identify the interactome of Cryab in Hep3B-Cryab and HCCLM3-Mock cells expressing high levels of Cryab (Fig. 4A). Using this approach, 200 and 190 proteins were identified as interacting with Cryab in HCCLM3 and Hep3B-Cryab cells, respectively. Of these, 30 and 26 proteins identified in HCCLM3 and Hep3B-Cryab cells, respectively, were found to be related to the MEK/ERK signaling by way of WholePathwayScope software (a comprehensive pathway-based analysis tool for high-throughput data27) (Tables S7, S8; Fig. S4). In addition, 10 proteins (CYFIP1, FASN, GSTP1, HSP90, HSPB1, IQGAP1, PCNA, PRKDC, ACTN4, and 14-3-3ζ) overlapped in two different cell lines (Fig. 4B).

Figure 4.

Analysis of Cryab-interacting proteins. (A) Identification of the peptides from Cryab associated protein by mass spectrometry. (B) Venn diagram of the interacting protein associated with MEK/ERK signaling; the overlapping proteins are listed in the table. (C) The interference of the overlapping proteins confirm which protein relays the Cryab signal to ERK1/2. The decrease of 14-3-3ζ obviously down-regulates the phosphorylation of ERK1/2, while the decrease of HSP27 only slightly influences the phosphorylation of ERK1/2. (D,E) RT-PCR and western blot identified that the 14-3-3ζ interference reversed the EMT in Hep3B-Cryab and HCCLM3-Mock cells. (F) Phase-contrast microscopy and immunofluorescent staining for E-cadherin, slug, F-actin, and vimentin verified 14-3-3ζ interference reversed the EMT in Hep3B-Cryab and HCCLM3-Mock cells. Scale bars = 200 μm.

To determine which proteins relay the signal to activate ERK, we next inhibited the expression of the 10 aforementioned proteins by RNAi in Hep3B-Cryab cells. We determined that a decrease in 14-3-3ζ reduced the phosphorylation of ERK1/2, while a decrease in HSP27 only slightly influenced the phosphorylation of ERK1/2 (Fig. 4C). Furthermore, we found that reduced 14-3-3ζ expression up-regulated the expression of E-cadherin and down-regulated the expression of slug, Fn 1, and vimentin in Hep3B-Cryab and HCCLM3-Mock cells (Fig. 4D,E). Of note, Hep3B-Cryab-si14-3-3ζ and HCCLM3-Mock-si14-3-3ζ presented the typical cobblestone-like appearance of normal epithelial cells in phase-contrast photographs, while Hep3B-Cryab and HCCLM3-Mock cells took on a spindle-like, fibroblastic morphology (Fig. 4F). The results indicate that Cryab protein forms a complex with 14-3-3ζ that results in up-regulation of the 14-3-3ζ protein, which markedly induces ERK1/2 activity and confers the mesenchymal phenotype.

Cryab Inhibits 14-3-3ζ Degradation by Forming a Functional Complex with 14-3-3ζ.

We further verified the relationship between Cryab and 14-3-3ζ protein. As shown in Fig. 5A, Cryab formed a complex with 14-3-3ζ in Hep3B-Cryab and HCCLM3-Mock cells, and immunofluorescence showed that Cryab and 14-3-3ζ were colocalized in the cytoplasm of Hep3B-Cryab and HCCLM3-Mock cells (Fig. 5B). More important, we found that the up- or down-regulation of Cryab expression in the aforementioned cells resulted in a corresponding increase or decrease in the expression of 14-3-3ζ protein, respectively, but 14-3-3ζ mRNA did not change. Inhibition of 14-3-3ζ expression had little influence on Cryab expression at the level of both protein and mRNA (Fig. 5C,D). The phosphorylation of ERK1/2 conferred by Cryab overexpression was inhibited by 14-3-3ζ RNAi (Fig. 5E). We next determined the expression of Cryab and 14-3-3ζ protein in 30 HCC tissues and analyzed the relationship of both molecules (Fig. 5F). Correlation analysis revealed that the correlation coefficient between 14-3-3ζ and Cryab expression was 0.760 (P < 0.01) at the protein level.

Figure 5.

Cryab restrains 14-3-3ζ degradation. (A) Co-IP assays showed Cryab formed a complex with14-3-3ζ. (B) Colocalization of Cryab (green) and 14-3-3ζ (red) in Hep3B-Cryab and 95-C-Cryab cells by immunofluorescence (original magnification, ×1,000). (C) The forced expression of Cryab up-regulated 14-3-3ζ protein, but not 14-3-3ζ mRNA. (D) Inhibition of 14-3-3ζ expression showed little influence on Cryab expression. (E) The interference of 14-3-3ζ inhibited the activity of ERK1/2 in cells overexpressing Cryab. (F) Expression of Cryab and 14-3-3ζ protein and the phosphorylation of ERK1/2 were analyzed by western blot in HCC tissues; the relevance of Cryab and 14-3-3ζ expression in HCC tissues is illustrated.

Cryab Overexpression Induces EMT of Cancer Cells by Way of ERK1/2/Fra-1/slug Signaling.

Previous studies have reported that the translocation of activated ERK1/2 into nuclei can activate transcription factors, such as Fos and Jun. Fos (c-Fos, FosB, Fra-1, and Fra-2) proteins dimerize with Jun proteins (c-Jun, JunB, and JunD) to form activator protein-1 (AP-1), a transcription factor that binds to TRE/AP-1 elements and activates transcription.28 Therefore, we examined whether HCC cells expressing high Cryab showed characteristics of consistently activated expression of transcription factors. First, we compared the mRNA level of these transcription factors using the microarray gene expression profiles of HCCLM3-Mock/HCCLM3-vshCryab and Hep3B-Mock/Hep3B-Cryab cells. Interestingly, only the level of Fra-1 mRNA was markedly enhanced in HCCLM3-Mock and Hep3B-Cryab cells compared with that in HCCLM3-vshCryab and Hep3B-Mock cells. These findings were further validated by RT-PCR and western blot analysis (Fig. 6A,B). Taking into account the up-regulation of slug in cells expressing high levels of Cryab, we hypothesized that Fra-1 can regulate slug expression. Thus, we treated Hep3B-Cryab and HCCLM3-Mockcells with small interfering RNA (siRNA)-Fra-1 and assessed slug expression using western blot analysis. Slug expression was substantially inhibited after siRNA-Fra-1 treatment in both cell lines (Fig. 6C,D). Finally, we analyzed the effect of U0126-mediated ERK inhibition on slug expression in HCC cells. Importantly, Fra-1 and slug expression were markedly down-regulated in cancer cells treated with U0126 (Fig. 6E). These results indicate that Cryab induced EMT by way of Cryab/ERK/Fra-1/slug signaling in HCC cells.

Figure 6.

Cryab overexpression mediated HCC cell EMT through the ERK1/2/Fra-1/slug signal. (A) Heat map of AP-1 transcription factors was showed between high and low expression Cryab cells (Hep3B-Cryab versus Hep3B-Mock and HCCLM3-Mock versus HCCLM3-vshCryab). (B) RT-PCR and western blot verified the results of the gene expression analysis. (C,D) The interference of Fra-1 was verified by RT-PCR and western blot, and the interference of Fra-1 expression reversed the EMT in Hep3B-Cryab and HCCLM3-Mock. (E) U0126 inhibited the expression of Fra-1 and slug. (F) Working model of the hyperactivation of ERK signaling and cancer cells EMT induced by Cryab overexpression.

Overexpression of Both Cryab and 14-3-3ζ in HCC Is Associated with Increased Invasive Potential.

We examined Cryab and 14-3-3ζ expression in a cohort of 403 HCC patients. The results showed that both 14-3-3ζ and Cryab staining were located in the cytoplasm (Fig. 7A). We found that 168 of 403 HCC cases (41.7%) exhibited high levels of both Cryab and 14-3-3ζ. Strikingly, HCC patients expressing high levels of both Cryab and 14-3-3ζ showed the worst prognosis (Fig. 7B,C). These results indicate that overexpression of both Cryab and 14-3-3ζ promotes the progression of HCC.

Figure 7.

Overexpression of both Cryab and 14-3-3ζ is associated with the poorest survival of HCC patients. (A) Representation of coexpression of Cryab and 14-3-3ζ in HCC paracancerous and tumor tissues. The high expression of Cryab appears to associated with the high level of 14-3-3ζ protein. Scale bars = 200 μm. (B,C) Cryab and 14-3-3ζ co-overexpression correlates the poorest survival and highest recurrence rate of HCC patients.

Discussion

Here, the majority of our data reinforce the notion that Cryab is a positive regulator of HCC growth and aggressiveness. First, Cryab promoted HCC progression in vivo and in vitro. Second, functional and genetic screens demonstrated that Cryab overexpression fostered HCC progression by inducing EMT. We also demonstrate for the first time that Cryab complexed with 14-3-3ζ, and elevated expression of Cryab up-regulated 14-3-3ζ protein, which relayed the signal from Cryab to activate the ERK1/2. Clinically, we found that Cryab expression correlated with BCLC staging, patients' overall survival, and disease recurrence. Moreover, we demonstrated that Cryab overexpression activated the ERK1/2/Fra-1/slug signal to induce HCC cell EMT. The above results support the notion that Cryab does play an important role in the progression of HCC.

Based on a combination of co-IP with subsequent MS or western blot-based identification of binding partners, we demonstrated that Cryab physically complexes with 14-3-3ζ. Furthermore, our results showed that the forced expression of Cryab was accompanied by up-regulation of 14-3-3ζ protein, but not 14-3-3ζ mRNA, in HCC cells. In addition, the interference of 14-3-3ζ reverses the mesenchymal phenotype conferred by Cryab overexpression, suggesting that the Cryab can protect 14-3-3ζ protein from degradation. The correlation coefficient between the Cryab and 14-3-3ζ proteins reached 0.760 in HCC tissues, supporting the notion that the Cryab-14-3-3ζ complex functions as a cooperative unit in HCC cells. This notion was further supported by the observation that the patients with overexpression of both Cryab and 14-3-3ζ had the poorest prognosis. The 14-3-3 protein belongs to a family of conserved regulatory molecules expressed in all eukaryotic cells,29 and Cryab is the most abundant sHsp in heart and muscle.30 Because both Cryab and 14-3-3ζ regulate many important proteins that are essential for homeostasis,31, 32 directly targeting Cryab or 14-3-3ζ may be a challenge. Here, we failed to detect the Cryab and 14-3-3ζ complex in normal liver cells L02 (unpubl. data), which indicates that this complex may not exist in normal cells, or may only exist in very small amounts. Thus, our findings may provide an alternative molecular target for HCC therapies by promoting the dissociation of the Cryab and 14-3-3ζ complex.

By gene expression analysis, co-IP with MS, bioinformatics analysis, and step-by-step RNA interference, we demonstrated the Cryab overexpression-induced hyperactivity of the ERK signal by forming a complex with 14-3-3ζ. Specifically, this ERK signal hyperactivity was resistant to sorafenib. As one of the 14-3-3 proteins, 14-3-3ζ was first identified to be associated with Raf.33 Subsequent evidence suggested that the association of Raf with 14-3-3ζ protects p-Raf from dephosphorylation.34, 35 Alternatively, 14-3-3ζ may function either as a linker by assembling Raf and other signaling proteins into a complex, or as a chaperone by stabilizing Raf in a conformation that is accessible for activation.36 For example, the 14-3-3ζ protein acts as a scaffold in a side-to-side mode of Raf catalytic kinase dimerization,37, 38 consisting of c-Raf and the Raf-related pseudokinase KSR (kinase suppressor of Ras) or with other Raf molecules. This dimerization can drive Raf catalytic activation independent of Ras and lead to resistance to Raf inhibitors.35, 39-41 As one component of this complex, Cryab is a scaffold or pseudokinase. According to structure-function studies, some pseudokinases, such as KSR and ERBB3 (v-erb-b2 erythroblastic leukemia viral oncogene homolog 3), can serve as allosteric activators of their associated kinases in addition to their roles as scaffolds.42 Moreover, some pseudokinases still possess low kinase activity despite their lack of certain catalytic residues.43 Thus, it is possible that Cryab triggers the initial steps in the activation of the ERK pathway. Our data show that Cryab overexpression induces the hyperactivity of the ERK signal in serum-starved HCC cells, suggesting that the Cryab-14-3-3ζ complex may initiate the activation of the ERK cascade. Importantly, we found that high levels of Cryab and 14-3-3ζ associated with the activation of ERK1/2 in 30 HCC tissues. Thus, we suggest that the Cryab-14-3-3ζ complex may activate the ERK signal by inducing the side-to-side of dimerization of RAF catalytic kinase (Fig. 6F).

Sorafenib, a multikinase inhibitor, has been shown to block tumor cell proliferation and angiogenesis by inhibiting serine/threonine kinases (c-RAF, and mutant and wildtype BRAF), as well as receptor tyrosine kinases. Currently, sorafenib is approved for the treatment of advanced HCC cancer in the clinic. However, preliminary results show that the efficiency of sorafenib varies. Here, ectopic expression of Cryab in Hep3B cells reduced sorafenib-induced apoptosis. Accordingly, the phosphorylation of ERK1/2 was only slightly down-regulated by sorafenib in Hep3B-Cryab and HCCLM3-Mock compared with the corresponding control cells. Importantly, the OS probability of the Cryabhigh group of HCC patients was much lower than that of Cryablow group. In fact, recent studies have shown that treatment with Raf kinase inhibitors can paradoxically induce ERK cascade signaling by promoting dimerization of Raf family members. For example, Raf inhibitors can induce KSR1/B-Raf and C-Raf/B-Raf dimerization, which attenuates the effect of inhibitors on the ERK cascade.44, 45 As one component of the Cryab-14-3-3ζ complex, 14-3-3ζ was reported to enhance these dimerizations, and acts as a true bridging molecule that links Rafs in this scenario.37 Given that Cryab can not only complex with 14-3-3ζ, but also elevate the 14-3-3ζ protein, it is not difficult to understand the role of Cryab in sorafenib resistance. Thus, Cryab may be a promising biomarker for predicting prognosis and sorafenib response of HCC patients.

In conclusion, we provide insight into the biology of Cryab signaling in HCC and demonstrate that Cryab overexpression up-regulates ERK phosphorylation by complexing with 14-3-3ζ, leading to an increase in HCC invasion through EMT and resistance to sorafenib. Thus, our study implies an optimal therapeutic strategy would be to target the Cryab-14-3-3ζ complex in a subset of HCC and suggest that Cryab may be a biomarker for predicting response to sorafenib treatment.

Acknowledgements

We thank Professor Jack Liang (Brigham and Women's Hospital, Boston) and Yong-Ting Wang (Med-X Research Institute, Shanghai Jiao Tong University, Shanghai, China) for providing pAcGFP1-C1-Cryab cDNA plasmids. We acknowledge Professor Yujiang Geno Shi (Brigham and Women's Hospital, Boston) for insightful review of the article.

Ancillary