Combined inhibition of AURKA and HSF1 suppresses proliferation and promotes apoptosis in hepatocellular carcinoma by activating endoplasmic reticulum stress

In this study we aimed to assess the anti-tumor effect of co-inhibition of Aurora kinase A (AURKA) and heat shock transcription factor 1 (HSF1) on hepatocellular carcinoma (HCC), as well as to explore the mechanism involved. Expression of AURKA and HSF1 in primary HCC tissues and cell lines was detected by immunohistochemistry (IHC), qRT-PCR and Western blotting. AURKA was knocked down in HepG2 and BEL-7402 HCC cells using lentivirus-mediated RNA interference. Next, CCK-8, clone formation, transwell and flow cytometry assays were used to assess their viability, migration, invasion and apoptosis, respectively. The expression of proteins related to cell cycle progression, apoptosis and endoplasmic reticulum stress (ERS) was analyzed using Western blotting. In addition, in vivo tumor growth of HCC cells was assessed using a nude mouse xenograft model, and the resulting tumors were evaluated using HE staining and IHC. Both AURKA and HSF1 were highly expressed in HCC tissues and cells, while being negatively related to HCC prognosis. Knockdown of AURKA significantly inhibited the colony forming and migrating capacities of HCC cells. In addition, we found that treatment with an AURKA inhibitor (Danusertib) led to marked reductions in the proliferation and migration capacities of the HCC cells, and promoted their apoptosis. Notably, combined inhibition of AURKA and HSF1 induced HCC cell apoptosis, while increasing the expression of ERS-associated proteins, including p-eIF2α, ATF4 and CHOP. Finally, we found that co-inhibition of AURKA and HSF1 elicited an excellent in vivo antitumor effect in a HCC mouse model with a relatively low cytotoxicity. Combined inhibition of AURKA and HSF1 shows an excellent anti-tumor effect on HCC cells in vitro and in vivo, which may be mediated by ERS. These findings suggest that both AURKA and HSF1 may serve as targets for HCC treatment.


Introduction
Liver cancer is one of the most common malignancies, ranking sixth in incidence and fourth in mortality worldwide [1]. Hepatocellular carcinoma (HCC) is the main type of liver cancer, accounting for about 75-85 % of all cases [1]. HCC exhibits high fatality, high metastasis and high invasion rates [2]. Although significant progress has been made in recent years in its treatment, including surgical resection, transplantation and the application of cytotoxic drugs, the prognosis of HCC patients is still poor [3]. The pathogenesis of HCC is known to involve multiple signaling pathways, but the molecular mechanisms underlying its malignant progression have remained unclear. As such, it is of great clinical significance to identify more effective biomarkers for early diagnosis as well as novel targets for tailor-made therapies.
Aurora kinase A (AURKA), a cell cycle regulating enzyme, starts to accumulate in the S phase of the cell cycle and rises rapidly in the late G2 and M phases until the G1 phase of the following cycle [4]. During the process of chromosome separation, AURKA participates in the formation of microtubules and/or the stability of the spindle pole, promotes maturation of the centrosome, and ensures appropriate mitotic cell cycle progression [5]. It has been reported that the expression of AURKA is increased in various cancers, including breast cancer, colorectal cancer, bladder cancer, head and neck cancer and HCC, being related to clinical stage, local lymph node metastasis and distant metastasis [6][7][8][9][10][11]. AURKA can accelerate cell cycle progression and promote tumor metastasis by activating several signaling pathways [12]. Meanwhile, it can promote tumor progression by accelerating epithelialmesenchymal transition and regulating tumor angiogenesis [13,14]. Given the notion that AURKA may serve as a target for the treatment of HCC, further studies on the role of AURKA in HCC and its underlying mechanisms are warranted.
As an important transcriptional regulator, heat shock transcription factor 1 (HSF1) is essential for normal developmental processes, while its abnormal expression has been found to be closely related to the occurrence and severity of malignant tumors [15]. It has been found that the expression level and transcriptional activity of HSF1 in HCC are higher than those in normal liver tissues [16]. HSF1 regulates the expression of genes related to the growth and development of cancer cells by mediating the transcription of heat shock proteins [17]. In the meantime, it can promote drug resistance of HCC cells by regulating the level of autophagy [18]. Here, we aimed to investigate the effect of co-inhibition of AURKA and HSF1 on HCC cells in vitro and in vivo and to explore its underlying mechanism in order to, ultimately, develop a new therapy for HCC.

HCC specimens
Three pairs of HCC tissues and adjacent normal tissues were obtained from the Department of Pathology, Jinling Hospital, Medical School of Nanjing University. The patients did not receive any chemotherapy, immunotherapy or radiotherapy prior to specimen collection. This study was approved by the Ethics Committee of Jinling Hospital, and written informed consent was obtained from all the patients.

Cell culture
HepG2 cells and BEL-7402 cells were purchased from the Cell Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China), and maintained in Roswell Park Memorial Institute (RPMI)-1640 medium (Gibco, Scotland, UK) containing 10 % fetal bovine serum (FBS, Sigma, St. Louis, MO, USA), 100 U/ml penicillin, and 100 µg/ml streptomycin. All cells were cultured at 37°C with 5 % CO 2 .

Transfection assay
A small-interfering RNA (siRNA) targeting AURKA and a negative control were purchased from RIBOBIO (Suzhou, China). The target sequence of the AURKA siRNA was 5′-A T G C C C T G T C T T A C T G T C A -3 ′ , w h i l e t h e siN05815122147 NControl_05815 (standard) from RIBOBIO served as a siRNA control (Ctrl). HepG2 cells and BEL-7402 cells were transfected with AURKA-siRNA or Ctrl-siRNA using Lipofectamine® RNAiMAX Reagents (Thermo Fisher Scientific) according to the manufacture's protocol. After 48 h of transfection, the cells were collected and subjected to Western blotting.

Cell viability assay
Cell viability was measured using a Cell Counting Kit-8 (CCK-8) assay following the manufacturer's instructions. Cells were seeded into 96-well plates at a density of 5000 cells/well and treated with various concentrations of Danusertib. After 48 h of culture, 20 µl CCK-8 solution was added to each well for cell viability detection. Absorbance at 450 nm was measured using a microplate reader.

Clone formation assay
Cells were seeded into 6-well plates at a density of 500 cells/ well in RPMI-1640 serum-free medium. The growth medium of each well was carefully replaced with fresh medium every 2 days for two consecutive weeks. Next, the cells were washed with PBS and stained with crystal violet, after which the clones in each well were photographed and counted.

Cell migration and invasion assays
Cells in a logarithmic growth phase were seeded on a Transwell upper chamber or on a Matrigel-coated Transwell upper chamber in serum-free medium. The lower chambers contained DMEM medium with 10 % FBS. After 48 h, nonmigrated/invaded cells were gently removed from the upper chamber using cotton swabs, after which the migrated/ invaded cells were stained using 0.1 % crystal violet and counted under a microscope.

Cell cycle analysis
Cells were seeded in 6-well plates at a density of 4 × 10 5 cells/ well. After treatment with Danusertib, the cells were washed with PBS and incubated with RNase A solution (100 µl) for 30 min at 37°C. Next, 400 µl propidium iodide (PI) was added to the cells and, after an additional 30 min incubation at room temperature, the DNA content of the cells was measured using flow cytometry (Guava easyCyte HT; Millipore, USA) to determine cell cycle progression.

Cell apoptosis assay
Cells in a logarithmic growth phase were washed with PBS and resuspended at a density of 1 × 10 6 cells/ml. Next, cell apoptosis was determined using an eBioscience™ Annexin V Apoptosis Detection Kit (eBioscience, Thermo Fisher Scientific, Inc., USA) according to the manufacturer's protocol, in conjunction with flow cytometry.

Western blotting
Proteins from the indicated cells were extracted using RIPA lysis buffer (Beyotime, China), after which protein concentrations were determined using a BCA Protein Assay Kit (cat. no. P0010; Beyotime, China). Next, the extracted proteins were separated using SDS-PAGE and transferred onto PVDF membranes (Millipore, Bedford, MA, USA). The resulting membranes were incubated with the following primary antibodies overnight at 4°C GAPDH (Cell Signaling Technology, Inc.) served as a loading control. Optical densities of the bands were measured using an ECL detection system (Tanon, Shanghai, China), after which the bands were scanned and analyzed using Image J software (National Institutes of Health, Bethesda, USA).

Mouse xenograft model
Female BALB/c nude mice (6-week-old) were obtained from Hangzhou Ziyuan Experimental Animal Technology Co. LTD and housed in well-ventilated rooms under specific pathogen-free conditions. The animal experiments were approved by the Ethics Committee of Jinling Hospital (approval no: 2020JLHGKJDWLS-140).
A HCC xenograft model was generated by subcutaneous injection of HepG2 cells into the right flanks of the mice. After 12 days, the mice were randomly divided into 4 groups and treated with KRIBB11, Danusertib or KRIBB11 combined with Danusertib, respectively. The body weights and tumor volumes of the mice were recorded every two days. One month after inoculation, all the mice were sacrificed and the xenografts were recovered and weighted. In addition, heart, kidney and liver tissues were harvested. Xenograft tumor volumes were calculated using the formula (V) = 0.5 × l × w 2 , in which w and l represent the short diameter and long diameter, respectively. Finally, the tissues and xenografts were embedded in paraffin and stained with hematoxylin-eosin (HE) and for immunohistochemical (IHC) biomarkers.

HE staining
Paraffin-embedded sections were dewaxed, rehydrated and stained with hematoxylin. After being rinsed with running water, the sections were stained with eosin, dehydrated through an ethanol gradient, cleared with xylene and sealed with neutral gum. Next, the sections were evaluated by light microscopy (Nikon, Japan).

Immunohistochemistry
Paraffin-embedded sections were dewaxed with dimethyl benzene and rehydrated in dH 2 O. Next, the slides were immersed in 0.01 M citrate buffer (pH 6.0), heated under high pressure for 2 min and incubated with the following primary antibodies overnight at 4°C: anti-AURKA, anti-HSF1, anticleaved caspase-3, anti-Ki-67 (Cell Signaling Technology, Inc.) and anti-ATF4. After incubation with the primary antibodies, the slides were subjected to incubation with a secondary antibody and DAB staining (Beijing Zhongshan Golden Bridge Biotechnology Co, Ltd, Beijing, China). Finally, the slides were counterstained with hematoxylin, dehydrated and sealed in neutral gum. The staining results were evaluated by light microscopy (Nikon, Japan).

Statistical analysis
The data are presented as mean ± SD and were analyzed using SPSS 20.0 software. Group comparisons were performed using Student's t-test or one-way ANOVA with Tukey posttest. Correlation analyses were carried out using Pearson correlation test. P < 0.05 was considered statistically significant.

High AURKA expression is associated with a poor HCC prognosis
To investigate the correlation between AURKA expression and HCC prognosis, first 3 pairs of primary HCC tissues and adjacent normal tissues were tested for AURKA expression using qRT-PCR, Western blotting and IHC, respectively. We found that the expression level of AURKA in the HCC tissues was significantly higher than that in the adjacent normal counterparts (Fig. 1A-C). Subsequent analysis of data in the online Gepia database (gepia.cancer-pku.cn/detail.php?) confirmed that AURKA was highly expressed in HCC tissues compared with their adjacent normal samples (Fig. 1D). In addition, we found that the overall survival of HCC patients with a high AURKA expression was significantly shorter than that of those with a low AURKA expression (Fig. 1E). These data indicate that a high AURKA expression is related to a poor prognosis of HCC patients.

AURKA promotes the proliferation and migration/invasion of HCC cells
To test whether AURKA has a pro-oncogenic potential for HCC, we generated AURKA-knockdown HepG2 and Bel-7402 cells ( Fig. 2A) and subsequently assessed their clone forming and migrative/invasive capacities. We found that AURKA knockdown significantly decreased the clone forming ability of HepG2 and Bel-7402 cells, and markedly inhibited the migration and invasion of these cells (Fig. 2B-D). Together, the results suggest that AURKA may play an important role in regulating HCC cell proliferation and migration/invasion.

AURKA inhibition down-regulates proliferation and migration/invasion and promotes apoptosis of HCC cells
Next, we investigated the effect of AURKA inhibition on the biological behavior of HCC cells. For this purpose, HepG2 and Bel-7402 cells were treated with various concentrations of the AURKA inhibitor Danusertib. We found that Danusertib effectively inhibited the proliferation of HepG2 and Bel-7402 cells in a concentration-dependent manner (Fig. 3A). In addition, we found that the clone formation, migration and invasion capacities were significantly reduced in the Danusertib groups compared with the control groups (Fig. 3B-D), being consistent with the data obtained above with AURKAknockdown HepG2 and BEL-7402 cells. Subsequently, flow cytometry was used to determine the effects of Danusertib on the apoptosis of HepG2 and Bel-7402 cells. We found that HepG2 and Bel-7402 cells treated with Danusertib showed increased apoptotic rates, while being arrested in the G2/M phase of the cell cycle ( Fig. 4A and C). Additionally, we found that Danusertib treatment led to a significant decrease in the expression of CDC2 and cyclin B1, and a marked increase in the expression of cleaved PARP and cleaved caspase-3 (Fig. 4B, D). Taken together, these data indicate that inhibition of AURKA attenuates the proliferation and induces the apoptosis of HCC cells.

High HSF1 expression is associated with a poor HCC prognosis and positively correlates with AURKA expression
Next, we found that HSF1 was highly expressed in HCC tissues (Fig. 5A-D), and that the overall survival of HCC patients with a high HSF1 expression was significantly shorter than that of those with a low HSF1 expression (Fig. 1E). Subsequent correlation analysis revealed that HSF1 expression was significantly positively correlated with AURKA expression in HCC. These observations led us to speculate that simultaneous inhibition of HSF1 and AURKA may suppress the development of HCC.

Co-inhibition of HSF1 and AURKA induces apoptosis in HCC cells by activating endoplasmic reticulum stress
To explore the cancer-promoting mechanisms underlying HSF1 and AURKA, HepG2 and Bel-7402 cells were treated with the HSF1 inhibitor KRIBB11, Danusertib and KRIBB11 combined with Danusertib, respectively, after which Western blotting was used to detect the expression of proteins related to endoplasmic reticulum stress. We found that both ATF4 expression and EIF2α phosphorylation in HepG2 and Bel-7402 cells treated with KRIBB11 or Danusertib were increased, and that combination treatment with KRIBB11 and Danusertib led to marked increases in the expression of ATF4, CHOP and p-EIF2α (Fig. 6A). These data indicate that co-inhibition of AURKA and HSF1 activates endoplasmic reticulum stress in HCC cells. We next generated ATF4-knockdown HepG2 and Bel-7402 cells and treated these cells with or without KRIBB11 + Danusertib. We found that ATF4 knockdown was rescued and led to apoptosis after co-administration of KRIBB11 and Danusertib (Fig. 6B-C). Likewise, we found that the increased expression of ATF4 and apoptosis caused by the co-administration can be rescued by TUDCA, an inhibitor of endoplasmic reticulum stress (Fig. 6D-E). Collectively, these findings suggest that combined inhibition of AURKA and HSF1 induces apoptosis in HCC cells, presumably by activating endoplasmic reticulum stress.

Co-inhibition of HSF1 and AURKA attenuates in vivo tumorigenesis and tumor growth of HCC cells
Finally, we analyzed the in vivo pro-oncogenic activity of AURKA and HSF1 in HCC xenografted nude mice co-   administered with KRIBB11 and Danusertib. We found that there were no significant differences in body weights of the mice among the four groups (Fig. 7A). Additionally, we found that the xenografted tumor volumes in the four groups increased in a time-dependent manner, while the tumor growth of the xenografts in the KRIBB11 + Danusertib group was significantly slower than that in other three groups (Fig. 7B). Next, we removed and weighted the xenografted tumors from each mouse at the end of the experiment. Strikingly, we found that the xenografts in the KRIBB11 + Danusertib group were significantly smaller and lighter than those in the control (Ctrl) group (Fig. 7C). Subsequent HE staining revealed that there were no significant differences in pathological morphologies of heart, kidney and liver tissues among the four groups, suggestive of a low drug toxicity in the KRIBB11 + Danusertib group (Fig. 7D). Besides, immunohistochemical analyses showed that compared with the Ctrl group, the KRIBB11 + Danusertib group exhibited a reduced expression of Ki-67, as well as a markedly increased expression of cleaved caspase-3 and ATF4 (Fig. 7E). Together, these results indicate that coinhibition of AURKA and HSF1 attenuates the tumorigenesis of HCC cells.

Discussion
AURKA plays a key role in regulating mitosis, and the activity of AURKA significantly increases during transition from the G2 to the M phase of the cell cycle [4]. In addition, it has been reported that AURKA is closely related to the occurrence and development of various malignancies, such as endometrial cancer, colorectal adenocarcinoma and oral cancer [19][20][21]. Here, we found that AURKA is highly expressed in primary HCC tissues and cells, and is significantly related to a poor prognosis of HCC. In recent years, AURKA has been suggested to serve as an effective target for anti-cancer therapy due to its important role in cell cycle regulation and the availability of small molecule inhibitors [22]. Since the discovery of the first AURKA inhibitor (ZM447439) as a potential drug for targeted cancer therapy, more than 30 AURKA inhibitors have been introduced [22]. Here, we evaluated the effect of the AURKA inhibitor Danusertib on HCC cells, and found that Danusertib treatment led to a significant dose-dependent inhibition of their proliferation, migration and invasion. In addition, we found that Danusertib induced G2/M cell cycle arrest by activating the Cdc2/cyclin B1 pathway and promoted apoptosis by increasing the cleavage of PARP and caspase-3. Overall, we found that Danusertib exhibited marked antitumor effects on HCC cells, suggestive of a potential of AURKA as a target for HCC treatment.
HSF1 has been reported to be a potential therapeutic target for liver cancer. It can promote liver cancer cell proliferation, inhibit apoptosis and suppress anti-tumor immunity, and is involved in regulating cell cycle progression, cell growth and colony formation [23]. Here, we found that HSF1 is highly expressed in HCC tissues, and that there is a significant correlation between HSF1 and AURKA expression. Additional in vivo studies on co-inhibition of AURKA and HSF1 revealed that, compared to the control group, a HCC nude mouse model treated with both Danusertib and KRIBB11 displayed a marked inhibition in tumor growth with a relatively low drug-associated cytotoxicity. These results support a superior therapeutic efficacy of co-administration of AURKA and HSF1 inhibitors.
In order to additionally investigate the mechanism underlying HCC cell apoptosis induced by co-inhibition of AURKA and HSF1, we chose to analyze endoplasmic reticulum stress (ERS) in HCC cells treated with both AURKA and HSF1 inhibitors. The occurrence of HCC is closely associated with chronic inflammation and liver cirrhosis related to liver disease. These factors can induce ERS through oxidative stress, inflammation and the occurrence of gene mutations [24]. Early ERS can alleviate stress-induced cell damage by activating the unfolded protein response (UPR). It has been reported that HCC can adapt to various unfavorable environmental conditions by activating the UPR, thereby contributing to tumor cell survival [24]. Sustained ERS may, however, trigger apoptosis. Lin et al. [25] found that HepG2 cells may exhibit a stress response and an increase in apoptosis induced by tunicamycin, thereby linking ERS to the induction of HCC cell apoptosis. Jin et al. [26] noted that, while mice fed with hexavalent chromium showed increases in the expression of GRP78, ATF6 and CHOP in liver tissues, which subsequently promoted apoptosis, there was a correlation between the above effects and the dosage of hexavalent chromium. In addition, several reports have shown that the CHOP pathway of ER stress plays an important role in kaempferol-induced apoptosis of HCC cells [27]. Here, we found that co-inhibition of AURKA and HSF1 in HCC cells led to increased expression of ERS-related proteins, including ATF4, p-eIF2α and CHOP. Phosphorylation of eIF2α could theoretically cause selective translation of ATF4. Moreover, prolonged ERS may lead to the induction of several pro-apoptotic genes mediated by ATF4-CHOP, further promoting apoptosis [28]. The  above observations led us to propose that co-inhibition of AURKA and HSF1 may induce HCC cell apoptosis by activating endoplasmic reticulum stress, thereby inhibiting tumor progression.
Abbreviations AURKA, aurora kinase A; HSF1, heat shock transcription factor 1; HCC, hepatocellular carcinoma; IHC, immunohistochemistry; ERS, endoplasmic reticulum stress; FBS, fetal bovine serum; siRNA, small-interfering RNA; CCK-8, Cell Counting Kit-8; HE, hematoxylineosin; UPR, unfolded protein response Authors' contributions ZS and XH conceived and designed this study; LY, HZ, XJ and CJ performed the experiments and data analyses; ZS and XZ drafted the manuscript; XH provided critical comments, suggestions and revised the manuscript. All authors read and approved the final version of the manuscript. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations
Conflict of interest The authors declare that they have no competing interests.
Ethics approval and consent to participate This study was approved by the Ethics Committee of Jinling Hospital, and written informed consent was obtained from all participating patients. The animal experiments were approved by the Ethics Committee of Jinling Hospital (approval no: 2020JLHGKJDWLS-140).
Consent for publication Not applicable.