Up-regulation of HDAC6 Results in Poor Prognosis and Chemoresistance in Patients With Advanced Ovarian High-grade Serous Carcinoma

Background: Ovarian high-grade serous carcinoma (HGSC) gradually acquires chemoresistance after recurrence. Our previous study on ovarian clear-cell carcinoma found histone deacetylase 6 (HDAC6) overexpression led to chemoresistance. This study aimed to evaluate HDAC6 as a predictor of chemoresistance and a therapeutic target for ovarian HGSC. Patients and Methods: The clinical significance of HDAC6 as a predictor of prognosis and chemoresistance in HGSC was immunohistochemically evaluated. In addition, expression of programmed cell death ligand-1 (PD-L1), and hypoxia-inducible factor-1α (HIF1α) were analyzed using clinical samples from 88 patients with ovarian HGSC, and their clinicopathological characteristics were reviewed. Results: Twenty-three patients had high HDAC6 expression, 10 positive PD-L1 expression, and 33 high HIF-1α expression. HDAC6 up-regulation was correlated with not undergoing interval debulking surgery (p<0.001), incomplete surgical resection (p=0.002), and frequent occurrence of stable disease/progressive disease according to the Response Evaluation Criteria in Solid Tumors (p=0.005) criteria. On Kaplan-Meier analysis, high HDAC6 expression was significantly associated with reduced progression-free (p=0.001) and overall (p=0.008) survival. On multivariate analysis, high HDAC6 expression (hazard ratio=1.65, 95% confidence interval 1.03-2.66; p=0.039) and surgery status were independent prognostic factors of progression-free survival. PD-L1 and HIF1α expression positively correlated with that of HDAC6. Conclusion: HDAC6 may become a potential therapeutic target in patients with ovarian HGSC since its up-regulation is considered to be associated with a poor prognosis in patients with this cancer.

Ovarian cancer is the leading cause of death owing to cancer of the female genital tract, with high-grade serous carcinoma (HGSC) being the most frequent histological type (1). The most important prognostic factor for patients with HGSC is the tumor stage; approximately 75-80% of patients have advanced-stage disease when they start showing symptoms, and <25% of cases with stage III/IV HGSC are curable using current therapies (2). Typically, HGSC shows a good response to a combination of platinum and taxane agents, which is a standard chemotherapy regimen for epithelial ovarian cancer. However, HGSCs recur frequently and gradually acquire resistance to these standard chemotherapy regimens (2). The most common molecular changes in HGSCs are alterations of tumor protein 53 (TP53) and inactivation due to germline or somatic mutation or promoter methylation of breast cancer susceptibility gene (BRCA)1 and BRCA2 in approximately 50% of HGSCs (3). TP53 activation was found to sensitize tumor cells to platinumbased chemotherapy, leading to cell-cycle arrest and apoptosis (4). Recently, new strategies have been devised for patients with ovarian cancer with BRCA mutations (5) or platinum-sensitive recurrence (6). However, treatments for patients with TP53 mutations or those with platinumresistant recurrence have yet to be developed.
Histone deacetylases (HDACs), which comprise 18 subtypes identified in humans, regulate tissue differentiation, apoptosis, migration, mitosis and angiogenesis by chromatinmodification via the deacetylation of histone or non-histone proteins (7). The inhibitors that target multiple HDACs exhibit cytotoxic effects in various types of cancer, including ovarian (8), but they are limited in their application for cancer treatment because of various toxicities (9). Therefore, more selective and effective HDAC inhibitors are required. Our previous study showed that high HDAC6 expression was an independent poor prognostic factor in epithelial ovarian cancer (10). HDAC6 was shown to increase the level of deacetylated α-tubulin, which up-regulated cancer cell growth by enhancing microtubule dynamics (11,12). HDAC6 down-regulation stabilizes p53 by increasing the levels of total p53 and p53 phosphorylation (13). In contrast, HDAC6 up-regulation leads to platinum resistance, whilst its down-regulation enhances platinum agent-induced DNA damage and apoptosis (13). Moreover, HDAC6 up-regulates several factors that cause chemoresistance, including hypoxia inducible factor-1α (HIF-1α) (14) and programmed death ligand-1 (PD-L1) (15). HDAC6-selective inhibitors were shown to be safe in clinical trials for multiple myeloma (16,17). HDAC6-selective inhibitors also suppressed the proliferation of AT-rich interactive domain 1A (ARID1A)-mutated ovarian clear-cell carcinoma and improved the survival of tumor-bearing mice (18).
In the current study, using clinical samples of ovarian HGSC, we immunohistochemically analyzed the association between HDAC6 expression and HGSC prognosis, and aimed to evaluate HDAC6 as a predictor of chemoresistance and as a therapeutic target for ovarian HGSC.

Patients and Methods
Patients and samples. All the methods, including the review of the electronic medical charts and pathological analysis, were performed in accordance with the 1975 Declaration of Helsinki after obtaining the approval of the Institutional Review Board (IRB number, 16-257) and informed consent (or with a formal waiver of consent). Altogether, 88 patients with ovarian HGSC that was surgically resected and pathologically confirmed at the Saitama Medical University International Medical Centre between 2007 and 2017 were recruited. The clinicopathological characteristics of these patients were reviewed, including age, recurrence, progression-free survival (PFS), overall survival (OS), International Federation of Obstetrics and Gynecology stage, treatment methods, surgical status (complete or incomplete resection), Response Evaluation Criteria in Solid Tumors (RECIST) status (19), and chemotherapy response score (CRS) (20,21). RECIST was used to stratify patients  according to the following responses: Complete response, partial response, stable disease (SD), and progressive disease (PD) using computed tomography before and after chemotherapy (19). Based on omental examination results, CRS was used to classify patients as follows: Patients with CRS of 3 had a complete/near complete response; those with CRS of 2 had a partial response; and those with CRS of 1 had minimal or no response (20,21).
Interpretation of immunohistochemical results. One gynecological oncologist (Mitsutake Yano) and one gynecological pathologist (Masanori Yasuda), both of who were blinded to the clinicopathological characteristics, evaluated the degree of immunohistochemical staining (Figure 1). The following scoring system was used: 0% stained cells indicated negative staining; 1-50% stained cells, mild staining; and 51-100% stained cells, marked staining. To optimize the differences in PFS and OS, the raw data were binarized for statistical analysis. For HDAC6 and HIF-1α, marked expression was considered to indicate high expression, while completely negative and mild expression was considered low expression. For PD-L1, mild and marked expression was defined as positive, while completely negative expression was considered negative.
Statistical analysis. Fisher's exact test or Pearson's chi-squared test was used to analyze the correlation between immunohistochemical expressions and the clinicopathological characteristics. Univariate survival analysis was performed by generating Kaplan-Meier curves, and differences between the groups were assessed using the log-rank statistic. The Cox proportional hazards model was used to perform univariate and multivariate survival analyses. All analyses were performed using SPSS v24.0 (SPSS Inc., IL, USA). p-Values of less than 0.05 were considered statistically significant.

. Kaplan-Meier survival analysis of progression-free survival (left) and overall survival (right) according to expression of histone deacetylase
L1 expression and the clinicopathological characteristics. HDAC6 expression was significantly positive correlated with PD-L1 (p=0.002) and HIF-1α (p=0.008) expression.

Correlation of HDAC6, PD-L1, and HIF-1α expression with survival.
On Kaplan-Meier survival curves, high HDAC6 expression was significantly associated with a decrease in both PFS (p=0.001) and OS (p=0.008) (Figure 2A). High HIF-1α expression was significantly associated with a decrease in PFS (p=0.006) and OS (p=0.033) ( Figure 2B). Positivity for PD-L1 expression was significantly associated with a decrease in PFS (p=0.029) and OS (p<0.05) ( Figure  2C). On univariate analysis using the Cox proportional hazards model, high HDAC6 expression, SD/PD RECIST status, CRS of 1 or 2, and incomplete resection were prognostic factors for PFS and OS (Table III) (Table III).

Discussion
HDAC6 up-regulation has been reported to be a poor prognostic factor in various cancer types, such as breast (22), esophageal (23), renal cell (24), and ovarian clear-cell (14). Our present study, however, is the first to report that HDAC6 expression is an independent prognostic factor for advanced ovarian HGSC. In this study, it was shown that high HDAC6 expression was an independent poor prognostic factor in patients with advanced ovarian HGSC. HDAC6 up-regulation was demonstrated to be significantly correlated with indicators of chemoresistance, such as surgical residual tumor and SD/PD according to RECIST criteria. Our study findingis consistent with the results of Wang et al. (13), who showed that HDAC6 up-regulation led to resistance to platinum agents, but its downregulation enhanced platinum agent-induced DNA damage and apoptosis. TP53 activation, the most common mutation in HGSC, sensitized cells to platinum agents and led to cell-cycle arrest and apoptosis (4). HDAC6 down-regulation stabilizes p53 by increasing the total p53 level and p53 phosphorylation (13). Moreover, the deacetylation of alpha-tubulin induced by HDAC6, reduces the effect of taxane microtubule-stabilizing agent ( Figure 3) (25). When HDAC6 was blocked, resistance to taxane agents was reversed in human epithelial ovarian cancer (25,26). These results suggest that HDAC6 upregulation is among the factors of recalcitrancy to standard platinum-based chemotherapy for HGSC. Optional treatment agents for ovarian cancer include bevacizumab [anti-vascular endothelial growth factor (VEGF)] and olaparib [poly(ADP-ribose) polymerase inhibitor] for BRCA-mutant or platinum-sensitive recurrent tumors (4,5), as well as pembrolizumab (PD-1/PD-L1 blockage) for microsatellite instability-high tumors (27). The current study showed that there was a positive correlation between HDAC6 and PD-L1 expression in ovarian HGSC, which has also been demonstrated in ovarian clear-cell carcinoma (14). HDAC6 inhibition has been reported to enhance the response to immunotherapy via PD1/PD-L1 (28,29). The anticancer effect of HDAC6 inhibitors was triggered by the G 2 /M cell-cycle arrest, apoptosis, and loss of mitochondrial membrane potential via the reduction of VEGF and poly (ADP-ribose) polymerase (30). In the present study, the expression of HDAC6 and HIF-1α, a factor upstream of VEGF, were significantly positively correlated ( Figure 3). Therefore, HDAC6 might be a potential therapeutic target for HGSC that is resistant to standard or optional chemotherapy. HDAC6-selective inhibitors have been shown to exhibit an antitumor effect in several cancer cell lines (16,(30)(31)(32), and are well tolerated with minimal toxicity from observations in clinical trials (16). The incidence of kidney failure (33) and peripheral neuropathy (34), which are common adverse effects caused by platinum and taxane agents, respectively, were reduced after using HDAC6-selective inhibitors. A phase Ib clinical trial of ricolinostat, an HDAC6-specific inhibitor, in combination with paclitaxel, demonstrated its safety in recurrent ovarian cancer (35). HDAC6 inhibition has potentially demonstrated antitumor effect (16,(30)(31)(32), synergistic antitumor activity with various chemotherapies (13,25,26,28,29), its safety (16,17,35), and furthermore efficacy in preventing or reversing chemotherapy-induced peripheral neuropathy (34,35) and kidney failure (33). Therefore, we suggest that HDAC6 is a potentially important and safe therapeutic target for ovarian HGSC.
As far as we are aware, the current study is the first to verify the clinicopathological correlation between HDAC6 and indicators of chemoresistance in clinical samples. But further confirmation is warranted via a large-scale population study. Secondly, the present study consisted solely of semiqualitative immunohistochemical analysis. Therefore, following a previous project of an experimental anticancer therapeutic model using ovarian clear-cell carcinoma cell lines (36), a quantitative analysis of HDAC6 protein and mRNA expression is planned to investigate the anticancer effect of HDAC6 inhibitors in serous carcinoma cell lines.

Conclusion
HDAC6 up-regulation resulted in a poor prognosis for patients with advanced ovarian HGSC because of chemoresistance. Therefore, HDAC6-selective inhibitors might become a promising therapeutic agent for ovarian HGSC that is resistant to the current standard/optional chemotherapy regimens.

Funding
This study was funded by Hidaka Research Projects in the Saitama Medical University (Grant numbers: 30-D-1-3) and Grants-in-Aid from the Ministry of Education, Science, Sports and Culture of Japan (Research Project Numbers: 18K06997).

Conflicts of Interest
The Authors declare no competing interests in regard to this study.