Prognostic impact of Dynamin related protein 1 (Drp1) in epithelial ovarian cancer

The mitochondrial fission protein, Dynamin related protein 1 (Drp1), and its upstream protein calcium/calmodulin–dependent protein kinase I (CaMKI) play a critical role in chemoresistance in ovarian cancer (OVCA). Thus, we examined the expression of Drp1, CaMKI and their phosphorylated forms and their prognostic impact in epithelial OVCA patients. Expression analysis was performed by immunohistochemistry (IHC) of paraffin-embedded tumor samples from 49 patients with epithelial OVCA. Staining intensity and the percentage of positively stained tumor cells were used to calculate an immunoreactive score (IRS) of 0–12. The expression scores calculated were correlated with clinicopathological parameters and patient survival. High immunoreactivity of phospho-Drp1Ser637 was significantly correlated with high-grade serous carcinoma (HGSC) (p = 0.034), residual postoperative tumor of > 1 cm (p = 0.006), and non-responders to adjuvant chemotherapy (p = 0.007), whereas high expression of CaMKI was significantly correlated with stage III/IV [International Federation of Gynecologists and Obstetricians (FIGO)] (p = 0.011) and platinum-resistant recurrence (p = 0.030). ROC curve analysis showed that Drp1, phospho-Drp1Ser637 and CaMKI could significantly detect tumor progression with 0.710, 0.779, and 0.686 of area under the curve (AUC), respectively. The Kaplan-Meier survival curve showed that patients with high Drp1, phospho-Drp1Ser637 and CaMKI levels had significantly poorer progression free survival (PFS) (p = 0.003, p < 0.001 and p = 0.017, respectively). Using multivariate analyses, phospho-Drp1Ser637 was significantly associated with PFS [p = 0.043, hazard ratio (HR) 3.151, 95% confidence interval (CI) 1.039–9.561]. Drp1 and CaMKI are novel potential candidates for the detection and prognosis of epithelial OVCA and as such further studies should be performed to exploit their therapeutic significance.


Background
Ovarian cancer (OVCA) is the most lethal gynecological malignancy, and ranks fifth as the cause of cancer death among women. The standard treatment is cytoreductive surgery coupled with the treatment of first-line chemotherapy with paclitaxel and carboplatin [1]. Regardless, more than half of the patients treated experience disease recurrence within 2 years, irrespective of the effectiveness of first-line chemotherapy and is associated with poor prognosis. Reliable prognostic biomarkers are therefore needed to aid in patient differential diagnosis and tailored therapeutic alternatives to improve patient survival.
Plasma tumor markers such as carbohydrate antigen 125 (CA125) is widely used for differential diagnosis of ovarian tumor and prognosis, tumor recurrence and the prediction of treatment response [2,3]. However, CA125 is unstable and their levels are affected by histological subtypes, FIGO stage or physiological conditions; thus, making their utilization questionable [4]. Circulating plasma gelsolin (pGSN) has recently been shown to be effective in detecting early stage OVCA and predicting residual disease compared with CA125; however, a large patient cohort is needed to substantiate these findings [5]. The combination of pGSN and CA125 provided a 100% sensitivity in detecting early stage OVCA [5] thus, providing an evidence that combining multiple tumor markers on a panel could increase OVCA diagnosis and revolutionize treatment. In terms of the prediction of treatment response or prognosis, various genomic, transcriptomic and proteomic biomarkers have been reported [6]. Moreover, the usefulness of imaging modalities such as 18 F-fluorodeoxyglucose positron emission tomography ( 18 F-FDG/PET) which reflect cellular glycolytic metabolism has also been reported, and is believed to be a more accurate prediction tool of chemotherapeutic response than CA-125 [7]. However, the identification of reliable biomarkers applied for all patient is urgently needed.
Mitochondria are highly dynamic organelles, and their fission and fusion fulfill mitochondrial function, including respiration, calcium buffering, apoptosis, and autophagy. Dynamin-related protein 1 (Drp1) is the master regulator of mitochondrial fission. Drp1 is mainly present in the cytoplasm but is translocated into the mitochondrial outer membrane and binds to its partner fission proteins such as mitochondrial fission factor (MFF) or mitochondrial fission 1 protein (Fis1) during mitochondrial fission [8]. Drp1 controls the balance between fission and fusion by their phosphorylation at two distinct serine moieties. Phosphorylation of Ser616 activates Drp1 and induces mitochondrial fission, whereas Drp1 is inactivated via Ser637 phosphorylation, resulting in mitochondrial fusion [9]. The role of Drp1-dependent mitochondrial fission and fusion in apoptotic progression and chemoresistance has been reported in different cancer studies [10]. Few studies have reported the relationship between Drp1 and chemoresistance in OVCA although most of them were in vitro studies using OVCA cells with no reports on their clinical relevance [11][12][13][14][15][16].
We have previously reported that OVCA cells expressed phospho-Drp1 Ser637 and are prone to form highly interconnected networks [17]. A calcium mobilizing agent, Saikosaponin-d, suppresses phospho-Drp1 Ser637 content and calcium/calmodulin-dependent protein kinase I (CaMKI) phosphorylation -which has also been reported to up-regulate Drp1 -leading to mitochondrial fission and subsequently apoptosis [17].
Extending from these in vitro findings, we have examined in this current study the clinical relevance and prognostic impact of Drp1, CaMKI and their phosphorylated forms in epithelial OVCA. Our results could assist in the development of targeted treatment options related to mitochondrial dynamics and calcium signaling in epithelial OVCA patients.

Patients and treatment
This study included 49 patients with primary epithelial OVCA treated between 2012 and 2017 at the Department of Obstetrics and Gynecology, University of Fukui. Formalin-fixed, paraffin-embedded tissue samples for all patients were obtained and analyzed retrospectively. Clinical and pathological factors were evaluated by reviewing medical charts and pathology records. The patients with histologically confirmed epithelial OVCA were included and the definitive histopathological diagnosis was performed by 2 certified pathologists based on the World Health Organization (WHO) classification. The patients' treatment included a combination of debulking surgery and adjuvant chemotherapy according to the clinical guidelines of the Japan Society of Gynecologic Oncology. Patients were followed-up for at least 24 months after the date of their first visit or until death. The study protocol has been approved by the institutional review board of the University of Fukui Hospital (IRB Number:20180150).

Tissue samples and immunohistochemistry
Formalin-fixed, paraffin-embedded, 2.5 μm sections were obtained from the samples collected. IHC staining was performed with the avidin-biotin-peroxidase complex method as previously described [18]. Antibodies used in IHC staining are shown in supplementary Table 1. To determine the stain intensity, the stroma was used as internal negative control and the vascular endothelium cells used as internal positive control (supplementary figure 1) [19][20][21]. The intensity and distribution of the Drp1, phospho-Drp1 Ser637 , CaMKI, and phospho-CaMKI Thr177 immunohistochemical staining reaction was evaluated using a semi-quantitative method (IRSscore) as described previously [18]. IRS-score was calculated as follows: IRS = SI x PP, where SI is the optical stain intensity graded as 0 = no, 1 = weak, 2 = moderate, and 3 = strong staining, and PP is the degree of positively stained cells defined as 0 = no staining, 1 = < 10%, 2 = 11-50%, 3 = 51-80%, and 4 = > 81%). Immunohistochemistry staining was scored by 2 independent observers. Representative images of immunostaining are shown in Fig. 1.

Statistical analysis
The outcome measures were progression-free survival (PFS) and overall survival (OS). PFS was assessed from the date of debulking surgery, whereas OS was from the date of first visit. Tumor progression was confirmed by either tissue biopsy or serial imaging showing evidence of progressive disease. The sample size calculation was performed by the statistical software EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [22] based on the results of CA125 as the predictive marker for PFS [23]. The Mann-Whitney U test was used to analyze relationships between clinical characteristics and IRS-score of each protein.
Receiver operating characteristic (ROC) curve analysis was performed to determine optimal cut-off values for discrimination with high accuracy based on the area under the curve (AUC) s for each protein. The Kaplan-Meier curve was used to assess the relationship between tissue markers and PFS and OS; log-rank test was used to calculate the statistical significance. The correlative studies using Pearson were used to determine the correlation among proteins or clinical characteristics. Cox proportional hazards regression modeling was used for univariate and multivariate analyses. Significance was defined as p < 0.05 (2-sided testing). All statistical analyses were performed using SPSS Statistics version 24 (IBM, Armonk, NY).

Patient characteristics
Clinical information of the 49 patients with their FIGO stages (I, 19; II, 3; III, 20; IV, 7) is summarized in Table 1.
The median age at diagnosis was 57.6 years (range, 31-82 years). Twenty patients were < 55 years old and 29 patients were ≥ 55 years old. Histopathological subtypes included high-grade serous carcinoma (HGSC) (n = 21), clear cell carcinoma (n = 13), endometrioid carcinoma (n = 6), mucinous carcinoma (n = 6), low-grade serous carcinoma (LGSC) (n = 2) and malignant Brenner tumor (n = 1). Twenty-nine patients received complete surgery (no postoperative residual tumor), 10 patients received optimal surgery (postoperative residual tumor of less than 1 cm), and 10 patients received suboptimal surgery (postoperative residual tumor of greater than to equal to 1 cm). Forty-three patients (87.8%) received adjuvant chemotherapy with paclitaxel and carboplatin. If the patients had severe neurotoxicity by paclitaxel, docetaxel was used instead of paclitaxel. Six patients (12.2%) did not received adjuvant chemotherapy because of the early stage of disease or the patients' demand. Based on the findings of computed tomography after last cycle of adjuvant chemotherapy, patients were evaluated for treatment response, which was classified into four categories: complete response when there was resolution of all evidence of disease for at least 1 month; partial response when there was a decrease of ≥50% in the product of the diameters (maximum and minimum) of all measurable lesions without the development of new lesions for at least 1 month; stable disease if there was a decrease of < 50% or an increase of < 25% in the product of the diameters of all measurable lesion; and progressive disease if there was an increase of ≥25% in the product of the diameters of all measurable In the present study, 32 (74.4%) patients who had complete response were considered as responders, whereas 11 (25.6%) patients in the other three categories (partial response, stable disease, and progressive disease) were considered as nonresponders [24,25]. The median follow-up period was 43.5 months (range, 11.7-80.7 months). Twenty-two patients (44.9%) had tumor progression including platinumsensitive (greater than to equal to 6 months) (16.3%) and -resistant (less than 6 months) (28.6%) during the followup period, and 10 patients (20.4%) died.
Prognostic effect of Drp1, phospho-Drp1 Ser637 , and CaMKI expression We excluded phospho-CaMK Thr177 in the subsequent analysis because ROC curve analysis failed to show any significant effect on the tumor progression and survival. Kaplan-Meier survival curves showed that the patients with high expression of Drp1, phospho-Drp1 Ser637 , and CaMKI showed significantly poor PFS (p = 0.003, p < 0.001 and p = 0.017, respectively) compared with patients with low expression. Moreover, patients with high expression of CaMKI showed significantly poorer OS (p = 0.030) than those with low expressions (Fig. 3). Univariate analysis showed that high IRS-score of Drp1, phospho-Drp1 Ser637 , and CaMKI were significantly associated with poor PFS (p = 0.018, 0.001, and 0.022, respectively). ≥55 years old, FIGO stage III-IV, HGSC, suboptimal surgery, non-responders, platinumresistant recurrence, and high values of CA125 were also significantly associated with poor PFS (p = 0.041, p < 0.001, p = 0.013, p < 0.001, p < 0.001, p < 0.001, and p = 0.021, respectively; Table 3). In multivariate analysis, since Drp1 and phospho-Drp1 Ser637 were related variables using Pearson correlative studies (p = 0.001), two different models including Drp1 and phospho-Drp1 Ser637 separately were used. FIGO stage, the completeness of surgical reduction, treatment response to adjuvant chemotherapy, and tumor progression were also related variables (p = 0.001). Therefore, FIGO stage was used for multivariate analysis. FIGO stage III-IV (p = 0.009) and phospho-Drp1 Ser637 (p = 0.043) were significantly associated with poor PFS and independent prognostic factors for PFS (Table 3).

Discussion
In the present study, we have demonstrated that the expressions of Drp1 and CaMKI had the most significant prognostic correlations in the patients with epithelial OVCA. Combined with our previous in vitro studies, both Drp1 and CaMKI could serve as potential target proteins for therapeutic purposes as well as possible prognostic biomarkers.
Mitochondria fusion and fission are essential to maintaining healthy mitochondrial function cells and important for many physiological functions, including energy generation, metabolism, calcium signaling and cell death. Drp1 is a member of the dynamin family of guanosine triphosphatases (GTPases) and plays a critical role in the mitochondrial dynamics. Drp1 has also been reported to be associated with the development of cancers by regulating various cellular processes such as cell death, metabolic reprogramming or cell cycle. Although many  [27]. In the present studies, the expression of Drp1 is mainly observed in the cancerous lesions compared with the healthy ovarian tissues; this also correlated with tumor progression suggesting that Drp1 plays a pivotal role in the progression of various cancers including ovarian cancer. Drp1 controls the balance between fission and fusion by phosphorylation at two distinct serine moieties. Phosphorylation of Ser616 activates Drp1 and induces mitochondrial fission whereas Drp1 is inactivated via Ser637 phosphorylation, resulting in mitochondrial fusion [9]. In some cancers such as lung cancer, breast cancer or melanoma, increased levels of phospho-Drp1 Ser616 and mitochondrial fission are associated with cancer progression [26][27][28]. Han Y et al. also reported in ovarian cancer cells that hypoxia promoted mitochondrial fission and cisplatin resistance through down-regulation of phospho-Drp1 Ser616 [11]. In contrast, we observed that chemoresistant uterine cervical and ovarian cancer cells exhibit highly interconnected mitochondrial networks and that mitochondrial fusion may contribute to chemoresistance [16,17,29]. Moreover, we have previously shown that the calcium mobilizing agent Saikosaponin-d suppresses phospho-Drp1 Ser637 content and CaMKI phosphorylation, leading to mitochondrial fission and subsequent apoptosis [17]. These results are consistent with the study by Yu Y et al. which reported that the inhibitor of anti-apoptotic BCL2 family protein increased the levels of Drp1, mitochondrial fission and apoptosis in cisplatin resistant ovarian cancer cells [13]. Thus, Drp1-dependent mitochondrial dynamics may confer chemosensitivity or resistance depending on the cancer type and in a cell line-specific manner. Therefore, we explored the expression of Drp1, CaMKI and their phosphorylated-form; these were related with clinical and prognostic effects using immunohistochemical analysis of tissue samples collected from patients with epithelial OVCA.  We have explored the correlation of Drp1, phospho-Drp1 Ser637 , CaMKI and phospho-CaMKI Thr177 expressions as well as serum CA125 levels with clinicopathological parameters. Age was not correlated with any parameter although most ovarian cancers develop after menopause. High levels of CA125 was significantly associated with advanced FIGO stage and HGSC and these results are consistent with previous report [4]. High expression of phospho-Drp1 Ser637 but not Drp1 was associated with HGSC, suboptimal surgery and non-responders for adjuvant chemotherapy. High expression of CaMKI was significantly associated with advanced FIGO stage and platinum-resistant recurrence whereas the expression of phospho-CaMKI Thr177 was not observed in the present study. In our previous in vitro study, CaMKI was phosphorylated by calcium mobilizing agent, leading to mitochondrial fission. Thus, phospho-CaMKI Thr177 was not expressed because the patients were newly diagnosed without any exposure to treatment. Interestingly, phospho-Drp1 Ser637 and CaMKI were more associated with clinicopathological parameters than Drp1, suggesting that these patients could be stratified and selected for targeted therapy such as mobilizing agents as previously described [17]. However, our results are not consistent with the interactive open-access database which was published in 2017 [19] where CaMKI was not expressed in OVCA tissues. The reason (s) for this apparent difference is not known, it is possible that this could partly be due to the differences of antibodies used for the immunohistological analysis as we used monoclonal antibody for CaMKI detection against their polyclonal antibody. Thus, further studies are needed to investigate the expression patterns of these tissue markers using more reliable antibodies. In terms of patients' outcome, the expression of Drp1, phospho-Drp1 Ser637 and CaMKI as well as CA125 were significantly associated with PFS. In particular, phospho-Drp1 Ser637 emerged as an independent prognostic factor for PFS but not OS. PFS provides insight into the time frame for tumor recurrence and thus plays a key role in chemoresistance. These results are conceivable since high expression of phospho-Drp1 Ser637 was also associated with suboptimal surgery and non-responders for adjuvant chemotherapy. These findings are consistent with that of Meshach A-W et al, where pGSN mRNA expression was associated with PFS but not OS in ovarian cancer patients [30]. Conversely, none of these proteins presented as independent prognostic factors for OS, a phenomena consistent with other reports [19]. Thus, further investigations are needed to explore novel biomarkers predictive of patient survival.
Although the findings from this study are promising, we also acknowledge some associated limitations. Our study is retrospective and monocentric. Therefore, further prospective and multicenter studies are needed to determine the prognostic value of Drp1, CaMKI and their phosphorylated-forms. We also look forward to validating these findings in larger patient cohorts with diverse histological subtypes.

Conclusion
We have for the first time provided new insight into the clinical prognostic impact of Drp1 and CaMKI in epithelial OVCA patients. Phospho-Drp1 Ser637 emerged as an independent prognostic factor for PFS which serves as the first report to explain the role of Drp1 and its related proteins in comprehensive patients' cohort using immunohistochemistry. These findings are promising and provide important insights into developing novel prognostic marker and targeted therapy in the patients with epithelial OVCA.