CASK Silence Enhances Chemosensitivity of Hepatocellular Carcinoma Through activating Apoptosis, Inducing JNK/c-Jun-Mediated Autophagic Cell Death and Decreasing ABCG2

Background: Advanced hepatocellular carcinoma (HCC) patients usually fail to be treated because of drug resistance, including sorafenib. Methods: The expression and prognostic role of calcium/calmodulin-dependent serine protein kinase (CASK) in HCC were assessed by combination of bioinformatic analysis and experimental validation. The effects of CASK in regulating proliferation, apoptosis and drug resistance of HCC cells in vitro and in vivo were investigated using gain- or loss-of-function strategies by performing lots of specic methods including Cell Counting kit-8 (CCK8), colony formation assay, ow cytometry, transmission electron microscopy, immunouorescent confocal laser microscopy and tumor xenograft experiments, immunohistochemistry staining. Moreover, the underlying molecular mechanisms responsible for CASK’s functions in HCC were also explored. Results: Currently, we discovered that CASK was positively associated with sorafenib resistance of HCC in vitro and in vivo, and was signicantly related with poor prognosis in HCC. Moreover, inhibition of CASK can increase the effect of sorafenib partially by promoting apoptosis and autophagy, while CASK overexpression presented the opposite results. Besides, all the pan-caspase inhibitor Z-VAD-FMK, autophagy inhibitor 3-Methyladenine (3-MA) and small interfering RNA (siRNA) of LC3B reversed CASK knockout-induced effects with sorafenib treatment, suggesting that both apoptosis and autophagy were involved in CASK-mediated above functions and autophagy played a pro-death role in this research. Intriguingly, similar results were observed in vivo. In molecular level, CASK knockout activated the c-Jun N-terminal kinase (JNK) pathway, and treatment with JNK inhibitor SP600125 or transiently transfected with si-JNK signicantly attenuated CASK knockout-mediated autophagic cell death. Besides, knockout of CASK dramatically inhibited the expression of ATP binding cassette subfamily G member 2 (ABCG2) and reversed of multidrug-resistance (MDR) of HCC. Conclusions: Collectively, all these results together indicated that centrifuged to remove the A BCA protein assay kit used to measure concentrations of proteins. Equal amount of protein sample was subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to PVDF membrane. The membrane was blocked with 5% non-fat milk at room temperature for 1 hour, and incubated with primary antibodies overnight at 4°C. After that, the membrane was incubated with secondary antibody for 1h at room temperature. The results were visualized and analyzed by ECL detection solution (Thermo Scientic™) and Image Lab software (Bio-Rad), respectively. the positive quantied kit-8; 3-MA: 3-Methyladenine; 5-Aza: decitabine; MDR: multidrug-resistance; EMT: Epithelial-mesenchymal transition; EGFR: Epidermal growth factor receptor; MAGUK: membrane-associated guanylate kinase; 5-FU: Fluorouracil; ABC: ATP-binding cassette; ABCG2: ATP binding cassette 2; ABCC3: ATP binding cassette C 3; TEM: Transmission electron microscopy; DMSO: Dimethyl sulfoxide; DAPI: 4′,6-diamidino-2-phenylindole; H&E: Hematoxylin and eosin; immunohistochemistry; mRNA: messenger RNA; TCGA: Cancer Genome HPA: human siRNA: qRT-PCR: quantitative GSEA: enrichment


Plasmids construction and transfection
CASK expression was knocked out using CASK sgRNA CRISPR/Cas9 system (target sequences: 5-CGACGACGACGTGCTGTTCG-3) and stable expression cell lines were selected using puromycin.
pcDNA3.1 and pcDNA-CASK plasmids were designed by Repbio Co., Ltd (Hangzhou, China) and stable expression cell lines were selected using G418. All these transfections were performed according to the manufacturer's instruction.
Cell viability assay and colony formation assay Cell viability assays were performed by the CCK8 assays. Cells were plated into 96-well plates with 3000 cells per well. After 12 hours, cells were transfected. Next, cells were exposed with drug treatment and cultured for 72h. Subsequently, 10 ul of CCK8 was added to each well for 4 h, then the OD value was measured at 450 nm. HCC cells were seeded into 6-well plates for colony formation assay (1000 cells/well). After maintained at 37°C and 5% CO 2 in incubator for 2 weeks, the cells were stained with Wright's-Giemsa stain according to the manufacturer's instructions. The number of colonies more than 50 cells/colony were counted.

RNA extraction and quantitative real-time PCR (qRT-PCR)
Total RNA was extracted from tissues and cells using RNAiso plus Reagent (TaKaRa, Kusatsu, Japan) and reverse transcribed to cDNA using PrimeScript RT Reagent Kit (TaKaRa, RR0037A). cDNA was analyzed by qRT-PCR using SYBR Premix Ex Taq (TaKaRa, RR420A). GAPDH was used as an endogenous control. The relative expression values were calculated using 2 −ΔΔCT method. Gene-speci c primers were listed in Table S1.

Western blot analysis
Total proteins were lysed with RIPA buffer containing with protease inhibitor and phosphatase inhibitor on ice and centrifuged at 14000 rpm to remove the debris. A BCA protein assay kit (Beyotime Biotec, China) was used to measure concentrations of proteins. Equal amount of protein sample was subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to PVDF membrane. The membrane was blocked with 5% non-fat milk at room temperature for 1 hour, and incubated with primary antibodies overnight at 4°C. After that, the membrane was incubated with secondary antibody for 1h at room temperature. The results were visualized and analyzed by ECL detection solution (Thermo Scienti c™) and Image Lab software (Bio-Rad), respectively.
Transmission electron microscopy (TEM) About 5×10 6 si-NC or si-CASK-3 transfected cells treated with or without 12.5 µM sorafenib were collected and xed in 2.5% glutaraldehyde at least for 4h. All the samples were treated with 1% osmium tetroxide and dehydrated in graded concentrations of ethanol and acetone, nally samples were embedded in Durcupan resin. Ultrathin sections (70 nm) were examined under a JEM-1230 electron microscope (JEOL, Japan) at 80 kV.

Tumor xenograft experiments
All the experiments were approved by the Institutional Animal Care and Use Committee of Zhejiang University. Male nude athymic mice (3-5 weeks old) were purchased from SLAC Laboratory Animal CO., Ltd (Shanghai, China). A stable CASK knockout cell line of SMMC-7721-sora was established (SMMC-7721-sora sg-CASK). In order to determine the effect of CASK knockout on sorafenib resistance in vivo, a total of 5×10 6 cells were injected into the right ank of each immune-de cient nude mice (control group or sg-CASK group, N = 10 per group). 6 days later, all of the animals in each group were divided into another two groups (Dimethyl sulfoxide (DMSO) treated or sorafenib treated, 10mg/kg) and administered every 3 days. Additionally, to explore the role of CASK regulates autophagy and augments sorafenib sensitivity in vivo, a total of 5×10 6 cells were injected into the immune-de cient nude mice. 6 days later, all of the animals were divided into four groups (no treated or 3-MA treated or sorafenib treated or combination of 3-MA and sorafenib treated, 3-MA: 20mg/kg, sorafenib: 10mg/kg, N = 4 per group) and administered every 3 days. Mice body weight and tumor volume (1/2×length×width 2 ) were measured every 3 days, continue to day 24. Besides, after the animals were terminated, tumor tissues were separated and weighted.
Immuno uorescent confocal laser microscopy SMMC-7721-sora control cells or SMMC-7721-sora sg-CASK cells were grown on coverslips and transfected with a GFP-LC3 plasmid overnight and pretreated with or without 20 µM SP600125 for 2h. Then, cells were treated with 12.5 µM sorafenib or not treated for another 36h. After that, cells were xed with 4% paraformaldehyde for 15 min and stained with 4′,6-diamidino-2-phenylindole (DAPI) for 5 min. Images were taken under confocal uorescence microscopy (Nikon AIR confocal microscope).

Immunohistochemistry (IHC) staining
Hematoxylin and eosin (H&E) staining and immunohistochemistry (IHC) staining were performed to gain the expression of CASK, LC3B, p-JNK and p-c-Jun in 4-µM-thick para n-embedded sections from tumor xenografts. Slides were examined using an optical microscope (Olympus). At least 7 randomly selected 40x elds of Ki67 staining were visualized and the percentage of positive nuclei were quanti ed using Image J software (NIH, Bethesda, USA).

Statistics
Experimental data were performed using GraphPad Prism software (version 7.0.3) and expressed as mean ± standard deviation (SD) of at least three independent experiments. Differences between two groups were analyzed using unpaired Student's t-test and p value less than 0.05 was indicated statistically signi cant.

Results
Hypomethylation-associated upregulation of CASK expression in HCC was positively correlated with poor prognosis To explore the potential function of CASK in HCC, we rst analyzed the expression of CASK through bioinformatic analysis and experimental validation. The results showed that the messenger RNA (mRNA) expression of CASK was signi cantly upregulated in HCC tissues compared with normal liver tissues in The Cancer Genome Atlas (TCGA) (https://cancergenome.nih.gov/) (Fig. 1A), and IHC staining performed by the human protein atlas (HPA) database (https://www.proteinatlas.org/) showed strongly positive CASK staining in HCC cancer tissue, but weakly positive staining in normal tissue (Fig. 1B). Receiver operating characteristic (ROC) curve analysis indicated that the expression of CASK could effectively distinguish HCC from normal tissues (Fig. 1C). And upregulation of CASK expression was associated with advanced stage in HCC (Fig. 1D). Furthermore, we compared the expression of CASK in 60 paired HCC samples and corresponding normal samples and found that CASK was markedly upregulated in HCC cancer samples (Fig. 1E). Besides, the Kaplan-Meier analysis (http://kmplot.com/analysis/) indicated that HCC patients with high-CASK expression demonstrated poor overall survival rate, and consistent result was observed in HCC patients treated with sorafenib ( Fig. 1F and Fig. 1G). Promoter DNA methylation is closely related with gene expression [30,31]. Thus, a possible link between promoter hypomethylation and upregulation of CASK expression in HCC was investigated. According to the analysis from UALCAN database (http://ualcan.path.uab.edu/), we found that promoter methylation level of CASK was remarkably decreased in HCC tumor tissues compared with normal tissues (Fig. 1H). Moreover, SMMC-7721 cells treated with 5 µM or 10 µM demethylation agent, 5-AZA, signi cantly increased CASK mRNA levels ( Fig. 1I). Collectively, all the data suggested that promoter hypomethylationassociated CASK upregulation in human HCC was positively related with sorafenib resistance and poor prognosis.
Knockdown of CASK inhibited HCC proliferation, promoted HCC apoptosis and enhanced the sorafenib sensitivity in vitro SMMC-7721-sora and SMMC-7721 cells were transfected with speci c siRNAs targeting CASK (si-CASK-1, si-CASK-2, si-CASK-3) and negative control si-NC, the qRT-PCR assay indicated that si-CASK-3 with the best effect in inhibiting CASK expression ( Figure S1A and Figure S1B). Additionally, the western blotting analyses were further performed to verify the effect of si-CASK-3 in SMMC-7721-sora and SMMC-7721 cells ( Fig. 2A). And CCK8 assay showed that after decreasing expression of CASK, the sorafenib treatment was signi cantly sensitive in various concentrations in SMMC-7721-sora and SMMC-7721 cells ( Fig. 2B and Fig. 2C). Besides, overexpression of CASK showed opposite consequences in SMMC-7721sora and SMMC-7721 cells ( Figure S1C and Figure S1D). For colony information assay, HCC cells were treated with si-CASK-3 or si-NC for 12 h, then cultured with sorafenib (sorafenib concentrations: 8 µM for SMMC-7721-sora, 2 µM for SMMC-7721) for 14 days. The assay further indicated that downregulation of CASK signi cantly decreased clonogenicity of SMMC-7721-sora and SMMC-7721 cells, and the effect was more pronounced when co-treated with sorafenib ( Fig. 2D and Fig. 2E). Furthermore, CASK upregulation combined with sorafenib treated notably increased clonogenicity of SMMC-7721-sora and SMMC-7721 cells compared with control groups ( Figure S1E and Figure S1F). To further de ne the mechanism of CASK-induced sorafenib chemoresistance, cells transfected with si-CASK-3 or si-NC cultured with varying concentrations of sorafenib or DMSO were analyzed for the apoptotic marker changes by the ow cytometric analyses. Markedly, a signi cant increase in apoptosis was observed in CASK-knockdown cells compared with control cells, and the effect was more noticeable when combined with sorafenib ( Fig. 2F and Fig. 2G). To further con rm these consequences, cleaved PARP and cleaved caspase-7, two markers of apoptosis, were monitored by western blotting analysis. As expected, the levels of cleaved PARP and cleaved caspase-7 were increased upon inhibition of CASK, with or without sorafenib treatment ( Fig. 2H and Fig. 2I). In contrast, overexpression of CASK in SMMC-7721-sora and SMMC-7721 cells had the opposite effects, leading to inhibit sorafenib-induced apoptosis ( Figure S1G and Figure S1H).

Silence of CASK increased autophagy of HCC cells and activated the JNK/c-Jun signaling pathway
In order to further discover the potential molecular mechanism underlying the increased drug sensitivity induced by decreased CASK expression, we performed the gene set enrichment analysis (GSEA) using the data from TCGA. We interestingly found that CASK expression was closely related with regulation of autophagy (Fig. 3A). Since autophagy is a signi cant regulatory progress in maintaining the cellular homeostasis, and plays a dual role in chemoresistance, we hypothesized that regulation of autophagy may participated in CASK-mediated sorafenib resistance. To test this, we investigated the levels of LC3B, p62 and Beclin-1 in SMMC-7721-sora and SMMC-7721 cells. Western blotting results showed that CASK knockdown signi cantly increased the expression of autophagic marker LC3B-II and Beclin-1, and decreased SQSTM1/p62 expression, with or without sorafenib treatment, whereas CASK overexpression decreased the expression of those proteins when treated with sorafenib ( Fig. 3B and Figure S2A). The TEM results revealed that CASK-knockdown cells contained more autophagosomes in the cytoplasm compared with control cells, with or without sorafenib treated (Fig. 3C).
Increasing studies have con rmed that JNK pathway activation plays a pivotal role in regulating autophagy, and is closely related with chemoresistance and tumor progression [32][33][34]. To further investigate the mechanism of CASK knockdown-induced autophagy, the effect of CASK knockdown on the JNK pathway was detected. Western blotting analysis indicated that CASK knockout indeed led to the increase of the protein levels of phosphorylated-JNK (p-JNK) and phosphorylated-c-Jun (p-c-Jun) in SMMC-7721-sora and SMMC-7721 cells with or without sorafenib treated (Fig. 3D).
Knockout of CASK inhibited HCC cell tumorigenesis and increased the effect of sorafenib in vivo SMMC-7721-sora and SMMC-7721 cells with stable knockout of CASK using CRISPR/Cas9 were screened out for the next research, and western blotting assay was performed to validate the effective knockout of CASK ( Fig. 4A and Fig. 4B). Then, SMMC-7721-sora (control cells) and SMMC-7721-sora sg-CASK cells were selected to conduct in vivo analysis. As shown in the diagram (Fig. 4C), a total of 5×10 6 control cells or sg-CASK cells in 100 ul PBS were injected into the mice, respectively. 6 days later, each group mice were further injected with 10 mg/kg of sorafenib every 3 days or DMSO every 3 days, and tumor volumes were measured. We found that the tumors derived from control cells grew evidently faster than those from sg-CASK cells, and the difference was more obvious when treated with sorafenib ( Fig. 4D). Likewise, tumor weight of xenografts derived from CASK suppression demonstrated a superior response to sorafenib compared to controls (Fig. 4E). Consistently, the mean volume of tumors in CASK knockout groups showed markedly smaller than in controls groups, especially in combination of sorafenib treated (Fig. 4F). H&E staining and Ki67 staining further indicated that knockout of CASK signi cantly inhibited proliferation (Fig. 4G). Together, these results suggested that depletion of CASK inhibited HCC cell tumorigenesis, increased apoptosis and enhanced the therapeutic effect of sorafenib in vivo.
CASK depletion modulated autophagic cell death-mediated sorafenib sensitization through activating JNK/c-Jun signaling pathway In the next step, we want to gure out whether CASK depletion-triggered autophagy showed a pro-survival or pro-death role, and whether it is mediated by JNK/c-Jun signaling pathway. Hence, a pan-caspase inhibitor (Z-VAD-FMK) was applied to CASK knockout treatment at rst. And the CCK-8 assay indicated that Z-VAD-FMK treatment partially reversed CASK knockout-induced cell death (Fig. 5A). It suggested that CASK was involved in the regulation of sorafenib resistance by regulating apoptosis, but nonapoptotic form of cell death might exist. Next, an autophagy inhibitor (3-MA) and the siRNA of LC3B were applied to inhibit autophagy. Western blotting was performed to detected the effect of 3-MA and si-LC3B ( Figure S2B). The CCK-8 assay indicated that 3-MA and si-LC3B treatment noticeably suppressed CASK knockout-induced cell death ( Fig. 5B and Fig. 5C). Then, to determine whether CASK knockout activated autophagic cell death to sensitize HCC cells of sorafenib in vivo, xenograft tumor models of SMMC-7721sora sg-CASK cells were generated. A total of 16 male nude athymic mice were randomly divided into 4 groups, including the control group, 3-MA group, sorafenib group, and sorafenib plus 3-MA group. As shown in Fig. 5D and Fig. 5E, the tumor size of 3-MA group was signi cantly larger than that without 3-MA treated group, especially in combination with sorafenib treated. Taken together, all these data support that inhibition of autophagy attenuates CASK knockout-induced cell death.
To determine whether CASK knockout-induced autophagy was dependent on JNK pathway activation, we speci cally inhibited JNK signaling pathway pretreated with JNK-speci c inhibitor SP600125, and CCK8 assay showed that inhibition of JNK/c-Jun signaling pathway with SP600125 attenuated the cytotoxicity activity of sorafenib in CASK knockout cells (Fig. 5F). Western blotting data further showed that SP600125-mediated inhibition of JNK signi cantly decreased the expression levels of JNK phosphorylation and LC3B-II in CASK knockout cells (Fig. 5G). Similar consequence was occurred when JNK/c-Jun pathway was inhibited by siRNA of JNK ( Fig. 5H and Fig. 5I). In accordance with these results, laser confocal images indicated that the number of LC3 positive puncta in CASK knockout cells were obviously increased than in control cells, and SP600125 pre-treatment signi cantly inhibited CASK knockout-induced LC3 positive puncta numbers (Fig. 5J). In addition, immunohistochemistry analyses of mice tumor tissues also showed that LC3B-II expression, p-JNK expression and p-c-Jun expression were higher in sg-CASK tissues than that in the control tissues with or without sorafenib treated ( Figure S2C). These data illustrated that JNK/c-Jun signaling pathway was involved in CASK-mediated autophagy.

ABCG2 was involved in CASK-regulated multi-drug resistance of HCC cells
In addition, we found that the drug-sensitizing effect could be observed in other anticancer drugs, including doxorubicin and daunorubicin ( Fig. 6A and Fig. 6B). By pumping the drugs outside from cancer cells and attenuate the potency of chemotherapeutics, ATP-binding cassette (ABC) transporters superfamily often involves in chemoresistance [35][36][37]. At rst, we studied the expression changes of those drug e ux pump proteins closely related with drug resistance in HCC. qRT-PCR indicated that the mRNA expressions of ATP binding cassette subfamily C member 3 (ABCC3) and ABCG2 were signi cantly decreased when knockdown of CASK ( Fig. 6C and Fig. 6D), and positive correlation were found between CASK and ABCC3 or ABCG2 from GEPIA (http://gepia.cancer-pku.cn/) and starbase (http://starbase.sysu.edu.cn) database ( Fig. 6E and Fig. 6F). While knockout of CASK only signi cantly downregulated the protein level of ABCG2 when treated with sorafenib, but not MRP3 ( Fig. 6G and Fig. 6H). All results con rmed that ABCG2 was involved in CASK-regulated multi-drug resistance of HCC cells.

Discussion
Chemoresistance is one of the major obstacles to improve the life quality and survival time of HCC patients. Elucidation the mechanism of drug resistance will help to identify potential and effective therapeutic targets to reverse drug resistance of HCC. This study illustrated that CASK was important for the sorafenib resistance of HCC cells in vitro and in vivo and further explored the underlying mechanism of CASK in HCC pathogenesis and progression.
Firstly, we found that CASK expression was signi cantly upregulated in HCC and was closely related with poor prognosis for HCC patients, which was regulated by promoter hypomethylation. More importantly, our data rst showed that CASK depletion-mediated sorafenib sensitization in vitro and in vivo mainly through increasing apoptosis, autophagy and decreasing the expression of ABCG2. It's well to known that the cytotoxic effect of chemotherapeutic drugs relies on their ability to induce apoptosis, also known as programmed cell death. Importantly, evading apoptosis is a common and key characteristic of cancer cells and is responsible for chemoresistance [38,39]. In the current study, the experimental data revealed that CASK downregulation increased HCC cell apoptosis through enhancing cleaved PARP and cleaved caspase 7 activation. Although apoptosis is the most widely studied programmed cell death, recent analyses have highlighted the signi cance of additional forms of cell death, like autophagic cell death [40][41][42]. In this research, we found that CASK knockout-induced autophagy per se enhanced its cell death effect. Expect apoptosis and autophagy, necrosis is another major mechanism explore for mammalian cell death, and we will demonstrate the relationship between CASK and necrosis in our future study.
Previous researches have indicated that JNK/c-Jun signaling pathway that belongs to mitogen-activated protein kinase (MAPK) pathway, has vital function in regulating autophagic cell death. For example, Bai, Y. et al have reported that PDIA6 knockdown suppressed NSCLC cell proliferation and increased cisplatininduced autophagic cell death via interacting with MAP4K1 to activate the JNK/c-Jun signaling pathway [43]; Hu, L. et al have proved that SNX-2112, the Hsp90 inhibitor, enhanced TRAIL-induced apoptosis and autophagy of cervical cancer cells through activating the ROS-regulated JNK-p53-autophagy-DR5 pathway [32]; Zhu, Q. et al indicated that irinotecan (IRI) stimulated the reactive oxygen species (ROS)related JNK-and p38-MAPK signaling pathways to increase autophagy-dependent apoptosis and inhibit growth of gastric cancer cells [44]. Therefore, the active status of the JNK/c-Jun signaling pathway was detected under CASK knockdown condition in the present study. As expected, we indeed observed that phosphorylation of JNK/c-Jun was signi cantly increased when CASK knockdown with the presence or absence of sorafenib treatment. In addition, inhibition of JNK via SP600125 or siRNA markedly suppressed sorafenib induced-autophagy in CASK knockdown SMMC-7721-sora cells. The consequences indicated that the JNK/c-Jun pathway was partially responsible for CASK knockout-medicated autophagic cell death.
An active e ux mechanism is one of the main reasons for MDR in cancer. Recently, mounting studies have highlighted the critical role of ABCG2 in mediating multidrug resistance of HCC cells [45,46]. Our study rst revealed that ABCG2 was involved in CASK-regulated MDR, providing a novel insight into how CASK regulated sorafenib resistance in HCC. Moreover, it's surprising to found that the mRNA and protein expression of CASK in parental cells and sorafenib-resistance cells were on the contrary (data not show), more efforts need to be put to explore this phenomenon. Besides, in order to further improve the clinical value of our experimental study, the clinical signi cance of CASK detection in sorafenib sensitivity or resistance HCC tissues in predicting the chemotherapy response or survival rate is also worth study in the future.
In conclusion, our study demonstrated that hypomethylation-induced upregulation of CASK in HCC is associated with poor prognosis for HCC patients. Furthermore, activation of apoptosis and JNK/c-Jun signaling pathway mediated autophagic cell death, and downregulation of ABCG2 produced by CASK downregulation, which reinforces sorafenib's effect in HCC cells (Fig. 7). Thus, CASK may serve as a potential novel prognostic indicator in HCC, and targeting CASK may be a promising strategy for HCC patients, especially for sorafenib resistant HCC patients.

Conclusions
In summary, our analysis for the rst time showed that hypomethylation-mediated high expression of CASK in hepatocellular carcinoma is associated with poor prognosis for hepatocellular carcinoma patients, and depletion of CASK enhances the sorafenib sensitivity in vitro and in vivo through activating apoptosis and autophagic cell death and decreasing ABCG2 expression.     partly attenuates CASK knockout-mediated sorafenib-induced HCC cell death by CCK8 assay. SMMC-7721-sora cell was pretreated with or without Z-VAD-FMK (50 μM) for 2h followed by exposure to 10μM sorafenib for 24h. (b) 3-MA partly attenuates CASK knockout-mediated sorafenib-induced SMMC-7721sora cell death by CCK8 assay. Cells pretreated with or without 3-MA (2.5 mM) for 2h, and following treated with 10μM sorafenib for another 24h. (c) si-LC3B partly reversed CASK knockout combine with sorafenib treatment-induced cell death by CCK8 assay. Cells pretreated with or without transiently transfected with si-LC3B for 12h, and incubated with 10 μM sorafenib for another 72h. (d) Representative image of CASK knockout xenograft tumors. (e) Tumor volume in each group. (f) SP600125 partly decreased CASK knockout-mediated sorafenib-induced SMMC-7721-sora cell death by CCK8 assay. Cells pretreated with or without 20 μM SP600125 for 2h, and following treated with 10 μM sorafenib for another 24h. (g) Western blotting were performed to detect the expression of p-JNK and LC3B following the treatment of SP600125 in SMMC-7721-sora sg-CASK cell. (h) si-JNK partly decreased CASK knockout-mediated sorafenib-induced SMMC-7721-sora cell death by CCK8 assay. Cells pretreated with or without transiently transfected with si-NC/si-JNK for 12h and following treated with 10 μM sorafenib for another 72h. (i) Western blotting were performed to detect the expression of p-JNK and LC3B following the treatment of si-JNK in SMMC-7721-sora sg-CASK cell. (j) CASK knockout or control cells transiently transfected with GFP-LC3 plasmid were pretreated with or without 20 μM SP600125 for 2h.