C1orf61 promotes hepatocellular carcinoma metastasis and affects the therapeutic response to sorafenib

C1orf61 is a specic transcriptional activator that is highly up-regulated during weeks 4–9 of human embryogenesis, the period in which most organs develop. We have previously demonstrated that C1orf61 acts as a tumor activator in human hepatocellular carcinoma (HCC) tumorigenesis and metastasis. However, the underlying molecular mechanisms of tumor initiation and progression in HCC remain obscure. In this study, we demonstrated that the pattern of C1orf61 expression was closely correlated with metastasis in liver cancer cells. Gene expression proling analysis indicated that C1orf61 regulated diverse genes related to cell growth, migration, invasion and epithelial-mesenchymal transition (EMT). Results showed that C1orf61 promotes hepatocellular carcinoma metastasis by inducing cellular EMT in vivo and in vitro. Moreover, C1orf61-induced cellular EMT and migration are involved in the activation of the STAT3 and Akt cascade pathways. We also found that C1orf61 was associated with HBV infection-induced cell migration in HCC. In addition, C1orf61 expression improved the ecacy of the anticancer therapy sorafenib in HCC patients. For the rst time, we report a regulatory pathway by which C1orf61 promoted cancer cell metastasis and regulated the therapeutic response to sorafenib.


Abstract Background
C1orf61 is a speci c transcriptional activator that is highly up-regulated during weeks 4-9 of human embryogenesis, the period in which most organs develop. We have previously demonstrated that C1orf61 acts as a tumor activator in human hepatocellular carcinoma (HCC) tumorigenesis and metastasis. However, the underlying molecular mechanisms of tumor initiation and progression in HCC remain obscure.

Methods
In this study, we demonstrated that the pattern of C1orf61 expression was closely correlated with metastasis in liver cancer cells. Gene expression pro ling analysis indicated that C1orf61 regulated diverse genes related to cell growth, migration, invasion and epithelial-mesenchymal transition (EMT).

Results
Results showed that C1orf61 promotes hepatocellular carcinoma metastasis by inducing cellular EMT in vivo and in vitro. Moreover, C1orf61-induced cellular EMT and migration are involved in the activation of the STAT3 and Akt cascade pathways. We also found that C1orf61 was associated with HBV infectioninduced cell migration in HCC. In addition, C1orf61 expression improved the e cacy of the anticancer therapy sorafenib in HCC patients. For the rst time, we report a regulatory pathway by which C1orf61 promoted cancer cell metastasis and regulated the therapeutic response to sorafenib.

Conclusions
These ndings increased our understanding of the molecular events that regulate metastasis and treatment in HCC.

Background
Hepatocellular carcinoma (HCC) is one of the most lethal cancers. Worldwide, HCC results in approximately 600,000 deaths each year. [1,2] In ltration into adjacent tissues and metastasis to distant organs are the causes of death in the majority of HCC patients. [3,4] Thus far, the detailed molecular mechanisms underlying the initiation and progression of tumor metastasis remain far from fully understood. As a result, medical prevention and treatment of HCC is disappointing.
HCC metastasis occurs via multiple steps, including neoangiogenesis, local invasion, intravasation, extravasation, and the colonization of a secondary anatomical site. [5] Many genes encoding cell surface receptors and secretory proteins cooperate are involved in a complex molecular network as part of this process. Some well-known components, such as metalloproteinase (MMPs), [6] Wnt, [7] transforming growth factor beta (TGF-β), [8] vascular endothelial growth factor (VEGF), [9] broblast growth factor (FGF) [10] and platelet-derived growth factor (PDGF), [11] play important and heterogeneous roles in cell survival, growth, proliferation, epithelial-mesenchymal transition (EMT) and metastasis. [12][13][14] Stromal components, including broblasts, vascular endothelia cells, nerve cells, in ammatory immune cells, mesenchymal stem cells (MSCs) and the surrounding extracellular matrix (ECM), play synergetic roles in maintaining tumor microenvironments. [15] This is an essential driver of HCC initiation, progression, invasion, and metastasis. [16] Some oncogenes may also play a more global role in regulating tumor metastasis. For example, gene expression signatures derived through transcriptional pro ling indicated that the transcription factor MYC is speci cally necessary for invasion and metastasis; MYC regulates downstream programs to regulate the expression of relevant genes or affects cancer cell EMT. [17] C1orf61 (chromosome 1 open reading frame 61) is a speci c transcriptional activator of c-fos that is widely expressed in proliferating and migrating cells of the developing brain. [18] We have previously demonstrated using gene expression pro ling that C1orf61 is highly up-regulated during weeks 4-9 of human embryogenesis, the critical period when most organs develop. [19] However, potential regulatory mechanisms and the mode of action of C1orf61 in inducing human liver cancer cell metastasis remains obscure.
In this study, we investigated the functions of C1orf61 in hepatocellular carcinoma metastasis. Our data showed that the level of C1orf61 expression was correlated with metastasis in liver cancer cells. C1orf61 promoted human hepatocellular cell invasion and migration in vitro and in vivo. In this process, C1orf61 activated STAT3 and Akt cascade pathways. In turn, these pathways promoted liver cancer cells towards EMT and eventually resulted in metastasis. Moreover, C1orf61 affected the response of liver cancer cells to sorafenib therapy. In this study, we present a novel mechanism by which C1orf61 contributes to metastasis in HCC. Our ndings identify C1orf61 as a new candidate for use in diagnosis, prognosis, and targeted therapy in HCC patients.

Materials And Methods
Cell lines and cell culture The following human liver cancer cell lines were used: (a) SMCC7721 and Hep3B cells were kindly provided by Dr. Zhiyong Mao (Tongji University), (b) HCCLM9 cells were obtained from the Wuhan University School of Medicine, (c) HepG2, HepG2.2.15, BEL7402, Huh7, FHCC98, and the immortalized non-malignant human normal liver cell line L02 were stored in our lab and (e) the stably transfected cell lines L02-C1orf61, BEL7402-C1orf61, HCCLM9 sh-C1orf61, Huh7 sh-C1orf61 were established by our lab.
HCCLM9 C1orf61-/-cell lines were established in our lab using the CRISPR/Cas system. Cells were all cultured in DMEM supplemented with 10% fetal bovine serum, penicillin (1000 U/ml), and streptomycin (1000 µg/ml) in a humidi ed chamber with 5% CO 2 . Cell culture dishes and plates were obtained from Wuxi NEST Biotechnology Co. Ltd. (Wuxi, China).

Microarrays and qRT-PCR
Total RNA was isolated using TRIzol reagent (Life technologies, Carlsbad, CA, US) according to the manufacturer's instructions. Samples were puri ed using the RNeasy Mini Kit (Qiagen, GmBH, Germany).
RNA samples from each group were then used to generate biotinylated cDNA targets for use in an Affymetrix GeneChip® Human Transcriptome Array 2.0. The biotinylated cDNA targets were then hybridized to the microarray. After hybridization, arrays were stained in a Fluidics Station 450 and scanned on an Affymetrix Scanner3000. The microarray experiments were performed following the protocol described by Affymetrix Inc. at Shanghai Biotechnology Corporation. The raw data were normalized using the SST-RMA method by an Expression console (Affymetrix). Ratios were calculated between L02-PHAGE, L02-C1orf61 and Huh7 cells. Genes with fold changes of at least 2 were selected for further analysis.

Plasmids and transfection
Two empty vectors, pHAGE.puro and PLKO.1.Sunny, were kindly provided by Dr. Zan Huang (Wuhan University). The empty vectors p-USE and p-USE-CA-Akt plasmids were purchased from Upstate Biotechnology (Lake Placid, NY, USA). shRNA targeting C1orf61 were purchased from GENECHEM Biotechnology (Shanghai, China). Cells were seeded in six-well plates, and plasmids were transfected for 48 hours with PEI (provided by Dr. Xiaodong Zhang).

Western Blot analysis and ELISA
After each treatment, cells were lysed and proteins were collected and their expressions were assessed by Western blot as previously described. Antibodies against C1orf61 was purchased from GL Biochem Ltd (Shanghai, China); the E-cadherin, N-cadherin, Vimentin, Occludin, Snail, p-Akt, Akt, STAT3, Caspase3, PARP antibodies were purchased from Cell Signaling Technology (Beverly, MA); the p-STAT3 (Y705F) antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA); and the β-actin, GAPDH antibodies were purchased from ProteinTech Group Inc (Chicago, IL, USA).
The levels of HBs and HBe in cultured supernatants were measured by ELISA to identify the transfection e ciency, following the instructions of Beyotime.
Immuno uorescence (IF) Brie y, cells were seeded on coverslips and xed with 4% paraformaldehyde for 20 minutes at room temperature. Cells were then washed twice with PBS and incubated with primary antibodies against Ecadherin, Vimentin, p-STAT3 (Y705F) at 4 °C overnight. After thorough washing, the cells were incubated with uorescence-conjugated secondary antibodies for 1 h. Nucleus were stained with DAPI (Beyotime). Fluorescent images were visualized using a confocal laser-scanning microscope (Fluoview FV1000; Olympus, Tokyo, Japan).
Wound-healing assay Cellular migration was determined by a scratch wound healing assay. Cells were seeded into a six-well plate and were allowed to reach con uence. A wound was created by scraping the monolayer of the well with a 200 µl pipette tip. The fresh complete medium was added after the oating cells were carefully removed. The cells were incubated at 37 °C for 24 h or 36 h. Images of the wound area was captured in three elds using an inverted light microscope (Olmpus).

Luciferase Reporter Assay
Cells were seeded at approximately 60% con uence in 24-well plates. 0.2 µg Akt luciferase reporters were transfected into L02-PHAGE, L02-C1orf61, HCCLM9 WT, and HCCLM9 sh-C1orf61 cells using PEI. The pRL-TK plasmid was used as an internal transfection control. After 36 hours, re y and Renilla luciferase activities were measured.

Cell proliferation and cell viability assays
Sorafenib was dissolved in DMSO before use. The effect of sorafenib on cell proliferation was characterized by cell counting. Cells were seeded at a density of 3 × 10 3 in a 96-well plate and cultured for 48 hours after treatment with the indicated doses of sorafenib. The cells were counted using a hemocytometer. Cell viability was measured using a trypan blue dye staining assay.

In vitro apoptosis assay
For the apoptosis assay, treated cells were harvested by 0.05% EDTA-trypsin, then washed with PBS, resuspended in 1 × binding buffer and stained with Annexin V-FITC and propidium iodide (PI) for 15 min at room temperature. Annexin V-FITC and PI uorescence were measured with a ow cytometer (Beckman-Counter).

Tumour xenograft assay
Five-week-old male BALB/C nude mice were purchased from the Model Animal Research Center (Changsha, China). Animals were handed according to the guidelines of the Laboratory Animal Center of Wuhan University. All experiments were conducted under approved procedures. HCCLM9 WT, HCCLM9 C1orf61-/-#1, HCCLM9 C1orf61-/-#2 cells were resuspend in 0.2 ml PBS and 5 × 10 6 cells were injected in the right ank or 1 × 10 6 cells were injected into tali vain. For ank tumor injection, tumor volumes and body weights were measured and recorded everyday. Tumor volumes were calculated as length × width 2 /2.

Tissue protein isolation and immunohistochemistry (IHC)
For ank tumor injection, mice were kept for 5 weeks and then sacri ced. For tail intravenous injection, mice were kept for 50 days and then sacri ced. Tumor tissues were extracted and washed in PBS and then stored at -80 °C. For western blot assay, the tumor tissue was lysed in RIPA buffer on ice and then centrifuged at 12000 g for 15 min at 4 °C to collect the supernate, and the proteins were subjected to Western bolt analysis, as described previously.
Lungs were harvested and immediately xed in 4% paraformaldehyde solution and subsequently para n embedded, sectioned and stained with H and E. Tumor tissues were xed in 4% paraformaldehyde solution and embedded in para n and sectioned (5 µm) for immunohistochemistry. The sections were incubated with the primary antibodies (E-cadherin, Vimetin, p-STAT3, p-Akt/Akt antibodies). The sections were further incubated with biotinylated goat anti-rabbit and goat anti-mouse antibodies. The speci c signals were then detected with streptavidin-conjugated horseradish peroxidase and with the use of diaminobenzidine as the chromogen.

Statistical analyses
Student's t test was used for statistical analyses. All data were expressed as the means ± SD. P < 0.05 was considered statistically signi cant.

Expression of C1orf61 promotes human liver cancer cell migration
To evaluate the effect of C1orf61 on cancer cell metastasis, we rst examined the relationship between the expression level of C1orf61 and HCC migration. Wound-healing assays and transwell assays showed that the expression level of C1orf61 positively correlated with cell migration in human liver cancer cells. HCCLM9 cells that expressed the highest level of C1orf61 possessed the strongest migration ability compared to other cell lines ( Fig. 1a and Fig. S1a, b). Furthermore, we found that the up-regulation of C1orf61 promoted migration in BEL7402 and L02 cells; in contrast, C1orf61 knock-down inhibited migration in HCCLM9 and Huh7 cells (Fig. 1b, c and Fig. S1c, d). Similarly, knock-out of C1orf61 using CRISPR/Cas9 also signi cantly inhibited cell migration in HCCLM9 cells ( Fig. 1d and Fig. S1e, f). Anoikis and soft agar colony formation suggested that multicellular survival and anchorage-independent growth occurred when cells were removed from inappropriate cell/ECM interactions; [20] these abilities were closely related to the metastasis potential of cells. Figure 1e and f show that C1orf61 expression inhibited anoikis in cells and promoted anchorage-independent growth without correct attachment in HCC. Taken together, these data suggest that the intracellular level of C1orf61 was associated with metastatic potential and the high expression of C1orf61 promoted cellular migration in human liver cancer cells.

C1orf61 regulates EMT-related gene expression and induces cellular EMT
To investigate the underlying molecular signaling events involved in C1orf61-mediated cell migration, we rst performed gene expression pro ling analysis using L02 (low C1orf61 expression), L02-C1orf61 (C1orf61 overexpression) and Huh7 cells (high C1orf61 expression). The results showed that C1orf61 regulated the expression of a diverse set of genes associated with biological function, cellular components and molecular activity ( Fig. 2a and Fig. S2a). Further comparison and analysis revealed that some genes were related to cell growth, migration, invasion and EMT (Fig. 2b). RT-PCR results for a number of genes were all consistent with the microarray data, suggesting that the gene expression pro ling array was correct (Fig. 2c). EMT is a key process in the metastatic cascade in tumors. It is regulated by multiple genes and complex interactions within the tumor microenvironment. Next, we examined whether C1orf61-induced cell migration was associated with EMT. Microscopy showed that the morphology of highly expressed C1orf61 cells changed from tightly packed colonies into spindle-shaped cells; the latter morphology is a typical feature of cells undergoing EMT (Fig. 2d). Moreover, isothiocyanate-conjugated phalloidin staining indicated that C1orf61 expression promoted BEL7402 and HCCLM9 cells to reorganize F-actin into parallel bundles and form lamellipodia. These ndings were consistent with the cellular morphology results (Fig. 2e). To further con rm above observation, we examined the status of epithelial and mesenchymal cell markers by Western blotting and immuno uorescence analyses. As shown in Fig. 2f and Fig. S2b, overexpression of C1orf61 in L02 and BEL7402 cells resulted in a decrease in the protein levels of the epithelial cell markers E-cadherin and occludin and the up-regulation of the mesenchymal cell markers N-cadherin, Vimentin and Snail. Correspondingly, C1orf61 knock-down in HCCLM9 and Huh7 cells resulted in a decrease in the protein levels of the mesenchymal cell markers and increased epithelial cell markers ( Fig. 2f and Fig. S2b).
Similar results were observed when C1orf61 was knocked-out in HCCLM9 (Fig. S2c, d). Therefore, we propose that C1orf61-induced cell migration is associated with the promotion of EMT in human liver cells.
Activating STAT3 is essential for C1orf61-induced cellular EMT and migration Signal transducer and activator of transcription 3 (STAT3) is a transcription factor involved in a major cascade responsive to stimulation by many growth factors and cytokines. [21] STAT3 is involved in the regulation of cell proliferation, apoptosis, cell invasion, angiogenesis and EMT processes. [22][23][24] Hence, we determined the levels of STAT3 and phosphorylated-STAT3. As shown in Fig. 3a, cells with high levels of C1orf61 protein expressed high levels of the phosphorylated-STAT3. These proteins were also translocated into the nucleus, where STAT3 regulates the expression of other genes ( Fig. 3b and Fig.  S3a). The STAT3-speci c inhibitor S3I-201 suppressed endogenous STAT3 phosphorylation and blocked the transfer of STAT3 into the nucleus (Fig. S3b). Correspondingly, the suppression of STAT3 with S3I-201 also impaired EMT in cells and attenuated cell migration in L02-C1orf61 but not L02 cells (Fig. 3c, d and Fig. S3c, d). Ectopic expression of STAT3 in L02 cells increased the STAT3 protein level and facilitated the transfer of STAT3 into the nucleus as expected (Fig. S3e). In this scenario, the protein levels of epithelial cell markers (E-cadherin and Occludin) were reduced and mesenchymal cell markers (N-cadherin, Vimentin) were up-regulated; cell migration was also promoted (Fig. 3e, f and Fig. S3f, g). The overexpression of STAT3-Y705F, which is an arti cially generated STAT3 containing a point mutation in tyrosine 705 resulting in a phenylalanine substitution, resulted in a pronounced dominant-negative effect on the activation of wild-type STAT3. STAT3-Y705F overexpression also inhibited cellular EMT and prevented migration (Fig. 3e, f and Fig. S3f, g). These results suggested that the activation of STAT3 has a signi cant promoting effect on C1orf61-induced cellular EMT and migration.

Akt activation is also involved in C1orf61-induced cellular EMT and migration
It has been shown that Akt is a key mediator of tumor metastasis and EMT induction in the progression of many human tumors. [25,26] Next, we evaluated the role of Akt in C1orf61-mediated cellular EMT and migration. As shown in Fig. 4a, ectopic C1orf61 expression in L02 and BEL7402 cells remarkably increased the level of phosphorylated Akt. C1orf61 silencing in HCCLM9 and Huh7 cells robustly suppressed Akt phosphorylation. A luciferase assay re ecting PI3K/Akt pathway activation also con rmed that high C1orf61 expression signi cantly enhanced the PI3K/Akt activity (Fig. 4b). This nding was consistent with the phosphorylation level. To examine whether Akt activation plays a role in C1orf61-induced cellular EMT and migration, we treated cells with the speci c PI3K/Akt inhibitor LY294002. The results showed that Akt inactivation by LY294002 increased E-cadherin expression and suppressed the level of N-cadherin in L02-C1orf61 cells. Moreover, wound-healing and transwell assays indicated that LY294002 signi cantly decreased cell migration in L02-C1orf61 cells; cell migration did not decrease in L02 cells where C1orf61 expression was low (Fig. 4c, d and Fig. S4a, b). In contrast, the ectopic expression of continually activated Akt (Akt CA) induced EMT in L02 cells and facilitated migration; the migration ability of these cells approached that of L02-C1orf61 cells (Fig. 4e, f and Fig.   S4c, d). These results revealed that Akt was involved in C1orf61-induced cellular EMT and migration.

C1orf61 Promoted Tumor Growth And Facilitated Metastasis In Vivo
To evaluate the effects of C1orf61 on tumor growth and metastasis in vivo, we established subcutaneous tumor xenograft models and an experimental metastasis assay (intravenous tumor cell inoculation) in athymic nude mice using HCCLM9 wild type and HCCLM9 C1orf61 knock-out cells, respectively. As shown in Fig. 5a and b, subcutaneous tumors containing HCCLM9 wild type cells grew dramatically faster than those containing C1orf61 knock-out cells after 28 days later. This was re ected in both tumor volume and weight. These observations revealed that C1orf61 facilitated tumor growth in vivo. To determine whether C1orf61 knock-out in HCCLM9 cells modulated the development of metastasis, we examined the formation of lung metastases in subcutaneous and intravenous tumor models. As shown in Fig. 5c and d, C1orf61 knock-out resulted in a signi cant reduction in metastatic animals and decreases in the number of metastatic nodules in the lung. Further examination of the expression of some related proteins by Western blot and immunohistochemistry indicated that C1orf61 knock-out inhibited cellular EMT, decreased STAT3 and Akt activity; in vivo results were consistent with in vitro ndings ( Fig. 5e and f). These data suggested that the up-regulation of C1orf61 promoted tumor growth and metastasis in vivo. The potential molecular mechanism may be related to the induction of cellar EMT, STAT3 and Akt regulation.

C1orf61 is involved in HBV infection-induced cell migration in HCC
Hepatitis B virus (HBV) is a major etiological factor for HCC and is closely associated with regulating liver cell malignancy, proliferation, metastasis and apoptosis. [27,28] Here, we determined whether C1orf61 modulated HBV infection-induced cell migration. HepG2.2.15 is a liver cancer cell line integrated with HBV; these cells exhibited increased EMT and migration compared to HepG2 cells without HBV infection. However, C1orf61 knock-down remarkably impaired HBV infection-induced cell migration (Fig. 6a, b and  Fig. S5a). To further con rm the role of C1orf61 in HBV-induced cell migration, L02-PHAGE cells were infected with HBV; EMT and migration potential were then examined. As shown in Fig. 6c and Fig. S5b, c, C1orf61 levels were positively correlation with HBV infection-induced EMT and migration. These ndings were consistent with that of HepG2 cells. The HBV genome encodes some primary proteins associated with infection and virus replication. These proteins include HBe, HBs, HBc, HBp and HBx, which regulate multiple HCC processes. Next, we determined which elements play critical roles in C1orf61-mediated tumor metastasis. Real-time PCR and Western blot analyses indicated that the transient expression of HBe and HBc remarkably increased C1orf61 mRNA and protein levels (Fig. 6d). Importantly, HBe and HBc induced C1orf61 expression in a transfection dose-dependent manner and facilitated EMT (Fig. 6e). Further wound-healing and transwell assays revealed that C1orf61-knockdown inhibited cell migration by HBe and HBc (Fig. 6f and Fig. S5d,e). These results showed that HBV, particularly the HBe and HBc proteins, promoted the migration of HCC cells via the up-regulation of C1orf61 expression.

C1orf61 improves the therapeutic response to sorafenib in HCC cells
Sorafenib, an oral small molecule multikinase inhibitor, has been recently approved for the treatment of hepatocellular carcinoma. Functionally, sorafenib inhibits tumor growth and angiogenesis, induces apoptosis and remodels the tumor microenvironment. [29,30] We next examined whether C1orf61 regulated cancer treatment in HCC. As shown in Fig. 7a and Fig. S6a, HCCLM9 cells highly expressing C1orf61 exhibited a more sensitive response to sorafenib treatment than other cells lines in regard to cell proliferation inhibition and induced cell death. BEL7402 cells, which express C1orf61 at a low level, showed relative resistance. We speculated that C1orf61 was involved in the regulation of sorafenib treatment. To further con rm this hypothesis, the same dose of sorafenib was used to treat BEL7402-PHAGE and BEL7402-C1orf61 or HCCLM9 and HCCLM9 C1orf61-/-cells. The results indicated that cells highly expressing C1orf61 performed better in response to sorafenib treatment (Fig. 7b). FACS and Western blot analyses of apoptosis revealed that the capacity of sorafenib to induce apoptosis was closely related the expression level of C1orf61 in human liver cells (Fig. 7c and Fig. S6b). Moreover, we also found that 4 µM sorafenib inhibited migration and repressed EMT in cells with a high level of C1orf61 but not in cells with low level of C1orf61 (Fig. 7d, e and Fig. S6c, d). Altogether, these ndings strongly suggest that C1orf61 improved the e cacy of sorafenib anticancer therapy against HCC.

Discussion
C1orf61 is located at human chromosome 1, band q22. Gene ampli cation in this region is enriched in hepatocellular carcinoma. Gene expression pro ling revealed that this gene is highly expressed during weeks 7-9 of human embryogenesis in particular. [18,19] Numerous developmental genes have been reported as associated with tumor progression and treatment. [31] Recent studies have demonstrated that C1orf61 is overexpressed in a population of liver cancer stem cells characterized by the expression of the membrane protein CD133 + . We have previously reported that C1orf61 is widely expressed in both proliferating and migrating cells during brain and bone development. Taken together, these ndings implied that C1orf61 plays a critical role in the initiation and progression of hepatocellular carcinoma.
Tumor metastasis is a clinical challenge that is the causative agent of the majority of cancer patient deaths. Although our understanding of the molecular events that regulate metastasis has universally improved, the processes of metastasis associated with C1orf61 remain unknown. The data presented here demonstrate that C1orf61 promoted human hepatocellular cell metastasis. In turn, the downstream of transcription factors STAT3 and Akt kinase are activated. This eventually induces EMT and cell migration. Moreover, C1orf61 is involved in regulating HBV infection-induced cell migration and affects therapeutic response to sorafenib in HCC cells (Fig. 7f).
In human hepatocellular carcinoma, C1orf61 exhibited controversial functions in tumor progression and cancer treatment. C1orf61 facilitated liver cancer initiation and metastasis, revealing a role in prooncogene activation and posing risk to healthy people when abnormally highly expressed. Conversely, C1orf61 improved the e cacy of sorafenib anticancer therapy. This indicated that C1orf61 expression bene ts HCC patients who are undergoing targeted chemotherapy. Whether C1orf61 contributes to the therapeutic e cacy of other treatments such as radiotherapy or chemotherapeutic agents requires further investigation. Therefore, our data implied that C1orf61 performed diverse functions in liver cancer; many of these underlying features are not yet fully understood. In fact, it genes with known con icting effects on tumor progression and treatment are very common. Many reports have shown that some proteins that maintain cell homeostasis, such as Akt, mitogen-activated protein kinase (MAPK), NF-κB and p21, directly contribute to tumor growth and the spread of metastases. [32][33][34] Such proteins with high activities can sensitize resistant cancer cells to the proapoptotic effects of some anti-cancer agents.
EMT is an important process in which epithelial cells change their phenotype from an apical-basal polarity to a spindle-shaped morphology; epithelial markers (Ecadherin and ccludin) are reduced and mesenchymal markers (N-cadherin, vimentin, and snail) are up-regulated. [35][36][37] These results in functional changes associated with the conversion of stationary cells to motile cells. In this study, woundhealing and transwell assays showed that cells expressing C1orf61 at high levels exhibited signi cant EMT potential and migration. Considering that C1orf61 has been described as a potential transcriptional activator that mediates the activation of the human c-Fos promoter, [18] we speculated that EMT in C1orf61-induced liver cancer cells might be the result of the transcriptional regulation of some related genes. Our gene expression pro ling analysis further con rmed our hypothesis that C1orf61 regulated the expression of diverse genes and was implicated in cell growth, migration, invasion and EMT (Fig. 2).
One limitation of our study is that we did not obtain transgenic mice despite many attempts. We speculate that C1orf61 plays an essential role in embryo development. To make our data more convincing, we instead performed subcutaneous tumor xenografts and intravenous tumor cell inoculations in athymic nude mice. Both metastasis assays were consistent with our in vitro observations; C1orf61 promoted liver cancer metastasis to the lung and regulated the downstream activation of Akt and STAT3 in vivo. Additionally, our prior gene expression analysis showed that C1orf61 is implicated in embryogenesis during the development of the skeletal system. C1orf61 expression was also found to be associated with low bone mineral density. Approximately 20% of HCC cases involve bone metastasis; [38] whether C1orf61 can promote the metastasis of liver cancer cells to bone is worth exploring, even if it the two issues seem unrelated.

Conclusions
In summary, we demonstrated here that C1orf61 acted as a tumor activator and promoted the metastasis of human hepatocellular cells in vitro and in vivo. Mechanistically, C1orf61 activated the STAT3 and Akt cascade pathways. These pathways induce cellular EMT and result in migration. Meanwhile, C1orf61 improved the therapeutic response to sorafenib in HCC cells. We report here a new regulatory pathway by which C1orf61 facilitates human liver cancer cells metastasis, which increases our understanding of the molecular events that regulate metastasis. C1orf61 may serve an effective candidate for diagnosis, prognosis, and targeted therapy in HCC patients.

Consent for publication
All authors agree to submit the article for publication.

Availability of data and materials
All data analyzed during this study are included in this manuscript and its supplementary information les.

Competing interests
The authors declare that they have no competing interests.  Overexpression of C1orf61 resulted in a morphological transformation from tightly packed colonies to a spindle-shaped, broblastic, dispersed morphology. (e) Stable C1orf61-overexpressing BEL7402 and stable sh-C1orf61 HCCLM9 stained for F-actin (phalloidin, red) and DNA (DAPI, blue) were visualized by confocal uorescence microscopy compared to BEL7402-PHAGE and HCCLM9-WT cells. Stress ber and lopodia were detected in HCCLM9 WT and BEL7402-C1orf61 cells, but not in control cells. (f) Expression of EMT markers (E-cadherin, Occludin, Vimentin, N-cadherin and Snail) were assessed by Western blot. βactin served as a loading control.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. supplement gure.docx