Involvement and Targeted Intervention of Mortalin-Regulated Proteome Phosphorylated-Modification in Hepatocellular Carcinoma

Objectives To reveal the mechanisms of the effects of mortalin in hepatocellular carcinoma (HCC) and to identify potential novel chemical inhibitors of mortalin. Materials and Methods For the experiments, three HCC cell lines (HepG2 cells, Hep3B cells, and sorafenib-resistant HuH7 cells) and xenografted nude mice were used. For the clinical analysis, cohorts of 126 patients with HCC and 34 patients with advanced recurrent HCC receiving sorafenib therapy were examined. Results Mortalin regulated the phosphorylation-modification of cancer-associated proteins and also regulated angiogenesis-related secretome to cause angiogenesis and sorafenib resistance in HCC cells. Two molecular mechanisms were identified. In one, via phosphatidylinositol 3-kinase (PI3K)/Akt signaling, mortalin regulated nuclear factor (NF)-κB and then activated vascular endothelial growth factor (VEGF)/vascular endothelial growth factor receptor (VEGFR)2 and granulocyte-macrophage colony-stimulating factor (GM-CSF), leading to neovascularization. In the other, mortalin regulated PI3K/Akt/β-catenin and then regulated Bcl-XL and Bcl-2, leading to the antiapoptosis effect of HCC. Treatment of the sorafenib-resistant xenografts with sorafenib in combination with mortalin knockdown facilitated the sorafenib-mediated inhibition of tumor growth and angiogenesis and increased apoptosis. Mortalin was a potential risk factor for HCC, predicting poor prognosis and sorafenib resistance. Finally, we showed that caffeic acid (C9H8O4) could bind to and induce the ubiquitination-mediated degradation of mortalin, which in turn blocked the abovementioned signaling pathways, leading to the inhibition of angiogenesis and the reversal of sorafenib resistance. Conclusions Mortalin, which regulates the phosphorylation of cancer-associated proteins, caused angiogenesis and sorafenib resistance, and was a competitive risk factor for HCC. Caffeic acid can therefore be considered a novel chemical inhibitor that targets the action of mortalin and a potential treatment for HCC.


INTRODUCTION
Hepatocellular carcinoma (HCC) is one of the most common solid tumors, and the fourth leading cause of cancer-related mortality worldwide (1). At present, the long-term outcome of patients with HCC is poor because of frequent recurrence and metastasis (2). Hematogenous metastasis represents a major shift in HCC biology: indeed, tumor angiogenesis, which plays a critical role in providing oxygen and nutrients, is necessary for HCC growth, intrahepatic and extrahepatic metastasis, and postsurgical recurrence (3). Thus, in addition to the curative therapies, resection, or transplantation (for the treatment of early-stage HCC, minority), antiangiogenesis strategies, such as sorafenib, have become the main approach for the treatment of advanced and recurrent HCC (which forms the majority of cases) (4). However, the therapeutic effects of these treatments are less than satisfactory. This is largely due to angiogenesis and the development of sorafenib resistance (5,6). Consequently, the identification of novel therapeutic strategies for HCC is a priority.
Mortalin is a stress-inducible molecular chaperone that belongs to the family of heat shock proteins (7). The overexpression of mortalin is associated with increased malignancy, angiogenesis, and metastasis in HCC and early recurrence (8,9). In the classical mechanism, mortalin sequestrates p53 in the cytoplasm, leading to the inactivation of p53 function and the suppression of apoptosis (10). However, unlike patients from Western countries, who have a low frequency (between 10% and 20%) of p53 mutations, patients with HCC from Southeast Asia (e.g., China and Japan) harbor a high frequency (up to 60%) of p53 mutations owing to chronic hepatitis B infection and aflatoxin B1 exposure (11,12). Furthermore, the increased expression of mortalin is mainly closely associated with cell protection, which promotes the survival of cells under lethal conditions and acts as a "guardian" to combat stress and apoptosis (13). Collectively, in addition to the classical repression of p53 activity, the molecular mechanisms underlying the cancer-promoting effects of mortalin in HCC are very complicated.
In our present study, we focused on describing a p53independent mechanism of mortalin that leads to angiogenesis and sorafenib resistance in HCC both in vitro and in vivo. Furthermore, we also identified the novel potential chemical inhibitors of mortalin via computer docking simulations and verified them using molecular biology techniques.

Xenografts in Nude Mice and Immunohistochemistry
The in vivo protocols were approved by the Nanjing Medical University Institutional Animal Care and Use Committee (2017-KY010). BALB/c nude mice were obtained from the SLRC Laboratory Animal Center (Shanghai, China) and were usually kept in a specific pathogen-free and temperature-controlled environment (20°C-22°C) under a 12-h light-dark cycle, with free access to drinking water and chow, as described previously (14). To establish the xenograft study, 2 × 10 6 cells in 100 ml of Matrigel were injected subcutaneously into the right armpit of each mouse, and mice were fed normally for 3 weeks. The dose of sorafenib (Selleckchem, Houston, TX, USA) administered was 60 mg/kg body weight (BW) via gavage, with negative control (NC)-siRNA or mortalin-targeted siRNA (100 nM, Santa Cruz Biotechnology, CA, USA; Supplementary Table S1) administered via intratumoral injection every 3 days. The tumor volume was calculated using the formula: V = ½ (width 2 × length). After 18 days, the mice were killed, and the tumor tissues were removed for further investigation. For immunohistochemistry (IHC), the sections were mounted on silanized slides, dewaxed in xylene, dehydrated in ethanol, boiled in 0.01 M citrate buffer (pH 6.0) for 20 min in a microwave oven, and then incubated with 3% hydrogen peroxide for 5 min. The sections were washed in PBS, incubated in 10% normal bovine serum albumin for 5 min, and then incubated with the primary antibody at 4°C overnight. The slides were then incubated with a horseradish peroxidase-conjugated secondary antibody at room temperature for 30 min. The samples were then visualized using diaminobenzidine, dehydrated, cleared, mounted, and photographed using a panoramic-scan digital slice scanning system (3DHISTECH Co. Ltd., Budapest, Hungary). The graphs were analyzed using Image-Pro Plus 6.0 software, as described previously (14,15). The criteria determining the immunostaining score are listed in Supplementary Table S2. and S4. The tissue microarray was constructed by Shanghai Zhuoli Biotechnology Co. Ltd. (Zhuoli Biotechnology Co., Shanghai, China). In each case, 1-2-mm-thick sections were cut from paraffin tissue blocks, dewaxed, pretreated, and placed on glass slides using an adhesive tape transfer system in order to avoid ultraviolet cross-linking. All reactions were performed using an automated staining device (Roche, Ventana Medical Systems, Oro Valley, AZ, USA). The immunostaining was quantified by two independent researchers who were blinded to the patient details described above.

Phosphoprotein-Specific Microarray Analysis
The phosphoprotein profiling was designed and manufactured using the phospho-explorer antibody array (PEX100) from Wayen Biotechnology (Shanghai, China). The samples were processed by protein extraction, lysate and marker buffer replacement, protein quantification, biotin labeling, and other operations according to the manufacturer's recommendations. Then, array closure, incubation, protein and array hybridization, and streptavidin combined with Cy3 detection were performed. The Surescan Dx Microarray Scanner was used to scan the array and obtain the data. The microarray contains 1,318 antibodies; each phosphorylation site was detected by two different types of antibodies to identify the phosphorylation status. Two replicates of each antibody were present and were located in two symmetrical blocks. The data for each phosphorylation site are presented as the mean of two biological replicates located in the two symmetrical blocks, partly narrowed the margin of error, with data correction performed at the same time. GenePix Pro 6.0 software was used to read the original data, including the fluorescence signal and background. A ratio was used to compute the extent of protein phosphorylation from the equation: Phosphorylation ratio = Phosphorylated value/unphosphorylated value. The data were shown in Supplementary Datasheet S1.

Angiogenic Antibody Array
A Quantibody ® Human Angiogenesis Array manufactured by RayBiotech (Guangzhou, China) was used to detect angiogenic factors in accordance with the manufacturer's instructions. Briefly, freshly frozen samples that were stored at -80°C were thawed and dissolved in lysis buffer to disintegrate the protein, and the protein concentrations were measured. The glass chip was air-dried at room temperature for 1-2 h, and standard dilutions of cytokine solutions were prepared. Then, 100 ml of sample diluent was added into each well on the glass chip and incubated at room temperature for 30 min to block nonspecific binding to the slides. The buffer was decanted from each well, and 100 ml of standard cytokines or sample was added to each well; then, the arrays were incubated at room temperature for 1-2 h. The glass chip was washed using Thermo Scientific Wellwash Versa Chip washer, incubated with the antibody detection cocktail and Cy3 equivalent dye-Streptavidin. Finally, an InnoScan 300 Microarray Scanner (wavelength, 532 nm; resolution, 10 µm) was used to measure the fluorescence, GenePix Pro 6.0 software was used to extract the data, and Quantibody ® Q-Analyzer software was used to analyze the data. The data for each of the angiogenic factors were presented as the mean of two biological replicates located in two symmetrical blocks, and the data were shown in Supplementary Datasheet S2.

Analysis of Biological Functions
The STRING database (Search Tool for the Retrieval of Interacting Genes; https://string-db.org) was employed to construct the protein-protein interaction network. The DAVID database (Database for Annotation, Visualization and Integrated Discovery; http://david.ncifcrf.gov) was employed to process the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. A p-value of <0.05 was set as the cutoff criterion for significant enrichment.

Determination of the Binding Status of Mortalin With Caffeic Acid or Ubiquitin
Huh7 sr Cells were pretreated with 0 or 20 mM MG-132 for 2 h and then exposed to 20 mM CaA for 6 h. Then, the cells were extracted with immunoprecipitation (IP) lysis buffer (Beyotime) for 30 min. The preparations were centrifuged, and 100 mg of total protein was incubated with anti-mortalin antibody at 4°C overnight. The protein-antibody complexes were incubated with IgG Sepharose beads (Beyotime) at 4°C for another 12 h. Subsequently, the supernatants were removed (positive control), and the beads were washed three times (the residual supernatants served as a negative control), boiled, and centrifuged. Then, the test substance was dissolved in water and analyzed by high-performance liquid chromatography-mass spectrometry (HPLC-MS) using the Thermo Fisher Q Exactive Plus LC-MS system with the CaA standard as the positive control. To determine the extent of protein ubiquitination, a co-immunoprecipitation assay was performed as described previously (16). Briefly, HuH7 SR cells were treated and extracted, and the IP complex was prepared as described above. Then, the samples were analyzed by performing a Western blotting assay with ubiquitin antibody.

Functional Assays
The other functional assays are presented in the Supplementary Materials and Methods section.

Statistical Analysis
The datasets were compared using GraphPad 8.0 (GraphPad Software, Inc., CA, USA), and the data were presented as mean ± SD. The differences were analyzed using Student's t-test, one-way analysis of variance followed by Dunnett's t-test, or two-way analysis of variance followed by Sidak's multiple comparison test, or linear regression. Survival curves were estimated using the Kaplan-Meier method, and differences in the survival distributions were evaluated by the log-rank test. A p-value of <0.05 was considered statistically significant.

Mortalin Caused Angiogenesis and Sorafenib Resistance in Hepatocellular Carcinoma Cells
First, we found that mortalin expression in HuH7 SR cells was significantly higher than that in HepG2 and Hep3B cells (Supplementary Figure S1). Then, the mortalin-plasmids were transfected to HepG2 and Hep3B cells, while the mortalin-siRNAs were transfected to HuH7 SR cells. We next evaluated the effects of mortalin on angiogenesis in the three cell lines and found that the tube formation ability of mortalin-plasmidtransfected cells were significantly increased, whereas knockdown of mortalin resulted in the opposite effect (Figures 1A-D and Supplementary Figure S1). Next, we investigated the effects of mortalin on sorafenib resistance. Similarly, the knockdown of mortalin increased the efficiency of sorafenib, whereas the overexpression of mortalin induced the opposite effect. The IC 50 s of sorafenib in mortalin-plasmid-or mortalin-siRNA-transfected cells with respective control cells were as follows: 12.33 vs. 7.682 for HepG2 cells, 12.6 vs. 6.362 for Hep3B cells, and 10.29 vs. 20.14 for HuH7 SR cells ( Figures 1E, F). Collectively, these results indicated that mortalin may be an effective factor that could enhance angiogenesis and contribute to sorafenib resistance in HCC cells. Moreover, mortalin may cause the abovementioned effect in both a p53-dependent and a p53independent manner.

Mortalin Regulated Phosphorylation in Cancer-Associated Proteins
In addition to the classic role of mortalin in the inactivation of p53, it also improves cancer cell survival and enhances tumor progression (especially in drug-resistant tumors) through other methods, such as synergistic effects with human telomerase reverse transcriptase and heterogeneous nuclear ribonucleoprotein K (17). First, we employed the STRING database to predict the potential mortalinregulated downstream signal transduction mechanisms. With mortalin employed as the central molecule, a protein-protein interaction network was generated for the 20 most frequently altered neighbor interactors and the 30 surrounding indirect interactors (Supplementary Figure S2). Next, we employed the DAVID database to conduct KEGG pathway analysis on the 50 factors identified. The enrichment analysis identified eight pathways related to the functions of mortalin, and the highest ranked function was oxidative phosphorylation (Figure 2A). Previous studies have shown that mortalin was involved in tumor development through its impact on the phosphorylation of several key molecules (18,19). Therefore, to clarify the effects of mortalin on protein phosphorylation, control and mortalin knockdown HuH7 SR cells were analyzed using a phospho-antibody microarray. We identified a spectrum of proteins with a change of more than 1.5-fold (i.e., a 50% increase or a 33% decrease) in phosphorylation compared with the mock-transfected control ( Figure 2B). In the KEGG enrichment analysis of the array, the PI3K/Akt pathway was ranked most highly ( Figure 2C and Supplementary Figure S3). Furthermore, many of these phosphorylated proteins are vital to tumor progression, and the knockdown of mortalin resulted in the extensive suppression of protein phosphorylation. As shown in Figure 2D, the phosphorylation of vascular endothelial growth factor receptor (VEGFR), insulin-like growth factor receptor (IGFR), fibroblast growth factor receptor (FGFR), and epidermal growth factor receptor (EGFR), which are receptors with extracellular signal transduction functions, is clearly declined. The results also revealed a reduction in the phosphorylation of many key components crucial to signaling pathways, such as PI3K/Akt (E, F) After the three HCC cell types were transfected with mortalin-siRNA or mortalin-plasmids, they were treated with different concentrations of sorafenib for 24 h. The cell viability was determined in triplicate, and IC 50 values were calculated. The differences were analyzed using two-way analysis of variance followed by Sidak's multiple comparison test.
these results indicated that mortalin could regulate the phosphorylation of cancer-associated proteins, and that PI3K/Akt might be a key downstream pathway, regulated by mortalin, in highly angiogenic and sorafenib-resistant HCC cells.

Identification of PI3K/Akt as an Important Downstream Factor of Mortalin, Regulating NF-Kb and GSK3/B-Catenin in Hepatocellular Carcinoma Cells
We further investigated the underlying mechanisms using GO analysis based on the proteins enriched in the PI3K/Akt signaling pathway. As shown in Figure 3A, the top-ranked biological process was negative regulation of apoptotic process. Next, we analyzed the phosphoproteins with the most significant changes in the PI3K/Akt signaling pathway and listed those downregulated or upregulated at least 2-fold compared with the control HuH7 SR cells. Here, in mortalin knockdown cells, the phosphorylation of many important factors including AKT1, VEGFR2, NF-kB, and Bcl-XL was significantly decreased, whereas the phosphorylation of p27kiP1 and glycogen synthase kinase (GSK)3b was significantly increased ( Figure 3B). Then, we further investigated the differential phosphorylation of proteins enriched in vascular endothelial growth factor (VEGF)/VEGFR, GSK3/b-catenin, and antiapoptosis signaling and confirmed that the abovementioned three signaling pathways were all inactivated in mortalin knockdown cells ( Figure 3C). Previous studies revealed that VEGFR2 would be targeted by sorafenib; therefore, to sensitize cells to sorafenib, another pathway (VEGF-independent manner) might be involved (20,21). Therefore, it is important to analyze the angiogenic proteome in HCC cells. Here, we employed an angiogenesis array to detect angiogenic factors in the HCC cells in mortalin-overexpressing HepG2 cells. The results showed that, among all the factors that could be detected, the levels of granulocyte-macrophage colony-stimulating factor (GM-CSF) and VEGF were markedly increased by the transfection of mortalinplasmids (Supplementary Datasheet S2 and Figure 3D). Moreover, NF-kB, which is regulated by the PI3K/Akt signaling pathway, was associated with VEGF expression. In addition, NF-kB activity and VEGF expression were also correlated with microvessel density in tissues and cell lines during in vitro angiogenesis, confirming NF-kB as one of the critical upstream regulators of VEGF (22,23). Meanwhile, GM-CSF is also one of the classic factors regulating angiogenesis that is regulated by NF-kB signaling (24,25). The evidence indicated that the activation of GM-CSF via NF-kB might be another way that is regulated by mortalin of HCC  to promote angiogenesis besides VEGF signaling. Based on these results, we proposed two potential methods. In one, PI3K/Akt regulated VEGF/VEGFR and GM-CSF via NF-kB signaling, leading to neovascularization; in the other, PI3K/Akt regulated GSK3/b-catenin, leading to an antiapoptosis effect. The two main neoplastic biological processes described are both involved in mortalin-induced angiogenesis and sorafenib resistance in HCC ( Figure 3E).

Verification of Microarray and Bioinformatics Results In Vitro and In Vivo
To verify the abovementioned hypothesis, we first determined the regulatory effect of mortalin on the PI3K/Akt, VEGF/VEGFR, and b-catenin/anti-apoptosis pathways, as determined by the level of phosphorylation of PI3K-p85 (Tyr458), Akt (Ser473), NF-kB p65 (Ser536), VEGFR2 (Tyr951), GSK3b (Tyr216), b-catenin (Thr41/ Ser 45), Bcl-XL, and the extent of apoptosis. As shown in Figure 4A, the knockdown of mortalin decreased the phosphorylation (activation) of PI3K/Akt/NF-kB/VEGFR2 but increased the phosphorylation of GSK3b, which in turn inactivated the bcatenin/Bcl-XL pathway. The knockdown of mortalin in HuH7 SR cells attenuated the secretion of VEGF and GM-CSF, whereas the overexpression of mortalin in HepG2 and Hep3B cells increased the secretion; however, the inhibition of PI3K/Akt led to the opposite effects ( Figure 4B). In addition, the knockdown of mortalin enhanced apoptosis, but the opposing trend was observed when mortalin was overexpressed, although apoptosis was significantly elevated when the cells were pretreated with LY294002 ( Figure 4C). LY294002 was only used as a pretreatment; it was removed after treatment and not present in the conditioning medium. Here, compared with the mortalin-plasmid-transfected cells, the tube formation ability of the LY294002-pretreated cells was markedly reduced ( Figure 4D). Furthermore, the IC 50 of sorafenib was also decreased in both HepG2 and Hep3B cells subjected to the LY294002 pretreatment ( Figure 4E). We also performed in vivo experiments to confirm our results. As shown in Figures 4F, G,  treatment of the xenografts with sorafenib alone or knockdown of mortalin inhibited tumor growth and angiogenesis. In combination, the knockdown of mortalin facilitated the sorafenib-mediated inhibition of tumor growth and angiogenesis and subsequently enhanced apoptosis. Collectively, these results suggested that the knockdown of mortalin blocked the PI3K/Akt pathway, which in turn inactivated the VEGF/VEGFR, GM-CSF, and b-catenin/ antiapoptosis signaling pathways, leading to the attenuation of angiogenesis and sorafenib resistance in HCC cells.

The Clinical Significance of Mortalin in Hepatocellular Carcinoma
We then evaluated the expression of mortalin in patients with HCC. As shown in Figures 5A, B, compared with the adjacent non-tumor liver tissues, a considerable increase in mortalin expression was observed in HCC tissues. To investigate the association between mortalin expression and angiogenesis, the HCC specimens were divided into "mortalin-low" vs. "mortalinhigh" groups according to the IHC-Q-Scores. The number of microvessels in the specimens in the mortalin-high group was significantly higher than those in the mortalin-low group ( Figure 5C). Moreover, a significant positive correlation was found between mortalin expression and the number of microvessels ( Figure 5D). We also found that there was a positive correlation between mortalin and 14-3-3h [an inducer of angiogenesis and sorafenib resistance described in our previous study (14,15)] ( Figure 5E). Finally, Kaplan-Meier survival analysis showed that patients in the mortalin-high group had poorer overall survival and recurrence-free survival than those in the mortalin-low group ( Figure 5F). The survival analysis of 34 patients with advanced recurrent HCC (who had received combined sorafenib treatment and transarterial chemoembolization therapy) revealed the same trend ( Figure 5G, left). In addition, we further defined the median survival of the cohort, 434 days, as the criterion for response to sorafenib: patients who survived less than 434 days were defined as sorafenib nonresponders, whereas patients who survived longer than 434 days were defined as sorafenib responders.
There was a significant difference between the sorafenib responders and nonresponders ( Figure 5G, right). These results indicated that mortalin may be regarded as a potential risk factor for HCC, predicting poor prognosis and sorafenib resistance. (G) Immunohistochemistry (IHC) staining of the mortalin, ki67, CD34, and terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) staining. The differences were analyzed using Student's t-test and one-way analysis of variance followed by Dunnett's t-test.

Identification of a Novel Potential Chemical Inhibitor of Mortalin: Caffeic Acid
MKT-077 (C 21 H 22 ClN 3 OS 2 ) is a classic chemical inhibitor of mortalin, which functions by eliminating mortalin-p53 interactions but does not alter the expression of mortalin (26). Previous studies revealed that caffeic acid phenethyl ester (CAPE; C 17 H 16 O 4 ), a specific inhibitor of NF-kB activation, was also able to induce the disruption of the mortalin-p53 complex, leading to nuclear translocation and the activation of p53, and was accompanied by a decrease in the expression of mortalin via transcriptional inactivation (27). However, the effects of these chemicals on the stability of the mortalin protein remain unclear.
Our previous study revealed that, similar to CAPE, CaA was also an effective NF-kB inhibitor (28), and that NF-kB was an upstream transcriptional regulator of mortalin (29). Moreover, via computational docking studies using SYBYL-X software, we found that the basic skeleton structure of CAPE, CaA (C 9 H 8 O 4 ), could bind to mortalin, and that the binding region of CaA to mortalin was the same as that of MKT-077 ( Figure 6A). Thus, we adopted an innovative approach, combining both IP and liquid chromatography-mass spectrometry (LC-MS) techniques, to reveal the interaction between CaA and mortalin. Mortalin was immunoprecipitated by a specific antibody, and its CaA-binding status was determined by LC-MS to investigate if the IP complex contained CaA. As shown in Figure 6B, the retention time of CaA was 3.29 min, and the area of CaA in the positive control group (total protein, supernatant) was 9.6 × 10 5 . The area of CaA in the immunoprecipitated complex (experimental group, sample) was 7.5 × 10 5 . CaA was almost undetectable in the negative control groups (lysis buffer and residual supernatant). Moreover, CaA treatment increased the ubiquitination of mortalin in HuH7 SR cells ( Figure 6C), decreased the expression/phosphorylation of mortalin, PI3K-p85 (Tyr458), Akt (Ser473), NF-kB(Ser536), VEGFR2 (Tyr951), b-catenin (Thr41/Ser 45), and Bcl-XL ( Figure 6D). We also used a previously established MHCC97H xenograft model and applied CaA (30,31) to confirm the effects of CaA on the expression/ phosphorylation of the abovementioned factors in vivo and found that CaA treatment inhibited mortalin, PI3K/Akt, NF-kB, VEGFR2, and b-catenin/Bcl-XL (Supplementary Figure S4). Moreover, when we treated HuH7 SR cells with CaA, the tube formation ability and the IC 50 of sorafenib were lower than those in the control group ( Figures 6E, F). Collectively, these results indicated that CaA inhibited the activity of mortalin by both transcriptional (indirectly, via blocking NF-kB) and posttranscriptional (directly, via targeting and inducing the ubiquitin-mediated degradation of mortalin) modifications in HCC cells.

DISCUSSION
Mortalin is an abundant mitochondrial protein also found in the mitochondria, endoplasmic reticulum, plasma membrane, and cytoplasmic vesicles. It participates in various biological processes, including proliferation, functional maintenance, and stress response in cancer cells and promotes tumor progression (7,32). Studies have also revealed that mortalin has an antiapoptotic effect (13). The loss of mortalin induces mitochondrial dysfunction and is always accompanied by a decrease in ATP and an increase in reactive oxygen species (ROS) generation (33). The abnormal phosphorylation of several molecules and the regulation of a series of key downstream signal molecules increase the antioxidant capacity of tumor cells, which are also essential requirements for cells to obtain and maintain the redox balance. In the subset of drug-resistant tumor cells, two biological behaviors (disorder of the redox balance and abnormal phosphorylation) occur in parallel; it is inevitable that these behaviors will mutually coexist and have a synergic effect. There is no inconsistency between oxidative phosphorylation and the classical role of mortalin in the regulation of ROS generation. However, the regulatory role of mortalin in protein phosphorylation remains unconfirmed. Here, we found that mortalin caused angiogenesis and sorafenib resistance via the regulation of phosphorylated cancer-associated proteins in HCC. The 14-3-3h is an isoform of the 14-3-3 family of proteins, which are phosphoserine/threonine-binding proteins (34). We previously showed that the 14-3-3h isoform was a novel neoplastic factor of HCC, inducing growth, angiogenesis, and multidrug resistance (MDR) properties (14,15). Here, a positive correlation between the expression of mortalin and 14-3-3h was found, suggesting that the mortalin-induced phosphorylation of the cancer-associated proteins may be mediated by 14-3-3h. This hypothesis requires further confirmation. Mechanically, PI3K/Akt is a key signaling pathway of particular relevance in cancer progression, and the hyperphosphorylation of PI3K and the activation of Akt are considered crucial factors that contribute to perturbation of the antiapoptotic, stability, and normal functions (35). The enhancement of signal transduction by the PI3K/Akt pathway is a key contributor to neovascularization (36). Currently, VEGF has been considered the most potential angiogenesis-inducing factor. In the VEGF/VEGFR pathway, VEGFR2 is the most important VEGFR. In our study, we found that VEGFR2 was remarkably downregulated, including at several classic phosphorylation sites, such as Tyr951 and Tyr1059, which may promote the neovascularization and regeneration of tumor cells, contributing to the development of cancer (37,38). In addition, GSK3 functions as an important downstream regulatory switch for numerous signaling pathways involved in the formation of tumors, including the regulation of the Wnt/bcatenin pathway (39,40). In the pathway, activated AKT could bind to the GSK3b complex and then phosphorylate GSK3b at Ser9 and increase the b-catenin levels (41, 42) but inhibit GSK3b phosphorylation at Tyr216, also causing the activation of bcatenin (43). Here, we found that the phosphorylation of GSK3b at Ser9 was downregulated, whereas the phosphorylation of GSK3b at Tyr216 was upregulated in mortalin knockdown cells, which inhibited the phosphorylation of b-catenin at several main sites, such as Ser37, Thr41, and Tyr654. This downregulation of phosphorylation could inhibit b-catenin separation from the cell membrane, preventing it from entering the nucleus and performing its function and promoting the progression of cancer (44,45). In addition, the downstream apoptosis process-related genes, such as Bcl-XL, Bcl-2, and BAD, were also affected and induced the apoptosis process, consistent with their functions (46). It has been shown that mortalin has an irreplaceable role in the regulation of protein phosphorylation and causes angiogenesis and sorafenib resistance in HCC. CaA is a phenolic compound synthesized abundantly by plants that is found in the basic skeleton of many phytochemicals, including CAPE (47,48). For the antioxidant activity, antiatherosclerotic activity, antiproliferative activity, and other various pharmacologic actions of CaA and its derivatives, they have been widely employed in clinical treatment so far and were recognized as antithrombotic, antihypertensive, antidiabetic, anticancer, and anti-inflammatory agents (49,50). In addition, it was reported that the expression level of mortalin showed significant correlation with HCC progression, and there was a high level of mortalin autoantibody in serum of patients with liver cirrhosis, suggesting that mortalin could be a promising serological marker for diagnosis (51). The relationship between mortalin, HCC, and CaA treatment remains to be further explored. Previously, we found that CaA may alter the progression of HCC through various means. To inhibit angiogenesis, CaA attenuated the activation of hypoxia-inducible factor (HIF)-1a by reducing c-Jun N-terminal kinase (JNK)1 activation and reducing HIF-1a stabilization and decreased p-STAT-3, attenuating the recruitment of HIF-1a and p-STAT-3 to the VEGF promoter (30). We also found that CaA attenuated cancer stem cell-like properties through the epigenetic regulation of the transforming growth factor (TGF)b-mothers against decapentaplegic homolog 2 (SMAD2) signaling pathway (31). In addition, CaA could block the expression/secretion of endogenous interleukin (IL)-6 by attenuation of the NF-kB-IL-6-STAT-3 feedback loop in HCC cells (28). Nevertheless, these effects induced by CaA are all dependent on an indirect means of signal transduction. However, our present study revealed that CaA also exhibited targeted intervention to mortalin by inducing the ubiquitin-mediated degradation of mortalin in HCC. Combined with the above evidence, we infer that the application of CaA may be a promising clinical targeted therapy for HCC in the future.

CONCLUSIONS
Our present study revealed that mortalin, which regulated the phosphorylation of cancer-associated proteins and angiogenesisrelated secretome, caused angiogenesis and sorafenib resistance in HCC. Among them, PI3K/Akt regulated NF-kB and b-catenin signaling, two neoplastic biological processes that play important roles in angiogenesis and sorafenib resistance of HCC. Moreover, mortalin was a competitive risk factor for HCC, predicting poor prognosis and the poor curative effect of sorafenib. Furthermore, we found that CaA was a novel chemical inhibitor targeting mortalin in HCC (Figure 7). Our present study not only reveals the key process of mortalin-induced HCC progression and the anticancer functions induced by CaA and other phytochemicals with a similar structure but also provides a new theoretical basis for further research into the targeted intervention of HCC.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Medical Ethics Committee of the Affiliated