Aberrant Expression of Folate Metabolism Enzymes and Its Diagnosis and Survival Prediction in Ovarian Carcinoma

This study was to validate changes in the levels of folate receptor-α (FOLR1), dihydrofolate reductase (DHFR), and methionine synthase reductase (MTRR) in the tissue of OC patients. The expression of FOLR1, DHFR, and MTRR was evaluated in 80 cases of primary OC, 50 cases of benign ovarian tumors, and 30 normal ovarian tissues. Associations between protein expression and clinicopathological characters were assessed, and diagnostic and prognostic evaluation of FOLR1, DHFR, and MTRR was performed. Results showed that upregulated FOLR1 and MTRR and downregulated DHFR were detected in OC. Patients with abnormality of FOLR1, DHFR, and MTRR tend to have a higher percentage of platinum resistance. Moreover, the areas under receiver operating characteristic curves (AUCs-ROC) for FOLR1, DHFR, and MTRR were 0.723, 0.717, and 0.714, respectively. The combination of FOLR1, DHFR, and MTRR could produce an area of 0.864 under the receiver-operating characteristic curve in distinguishing platinum-resistant patients from platinum-sensitive patients (P < 0.0001). Correlations were present between the expression of FOLR1, DHFR, and MTRR. Furthermore, Kaplan-Meier curves indicated that the patients with overexpressed MTRR had a poorer overall survival time compared to those with low expression (P < 0.05). Thus, folate metabolic enzymes could provide a potential promising biomarker for diagnosis platinum-resistant in OC.


Introduction
Ovarian cancer is the fourth most common cancer in women worldwide, and it has the most noteworthy lethal rate around gynecologic malignancies. Two most critical barriers to treatment of ovarian malignancy are absence of early diagnostic markers and advancement of drug resistance after therapy, especially in advanced stages. Various epigenetic changes have been recognized in ovarian cancer. Recent progresses in our understanding of molecular pathogenesis of ovarian malignancy have dramatically provided potential new targets for molecularly targeted therapies. There thus is a critical need for improved biological markers and therapies for ovarian carcinoma, which will come from a better understanding of the biology of the disease.
Folate is an essential component in DNA synthesis, replication and repair, protein synthesis, and methylation reactions. This is especially true for rapidly dividing cells [1]. Folate receptor 1 (FOLR1) internalizes folates by means of receptor-mediated endocytosis and reduced folate carrier (RFC) uses a bidirectional anion-exchange mechanism to transport folates into cytoplasm [2,3]. Dihydrofolate reductase (DHFR) catalyzes the reduction of dihydrofolate (DHF) to tetrahydrofolate (THF), which plays a vital role in cellular metabolism and cell growth [4]. Methionine synthase reductase (MTRR) is an enzyme controlling the activity of MTR in folate metabolism by transferring the methyl group of methyltetrahydrofolate to homocysteine via the methionine synthase, which is responsible for DNA methylation [5].
In the previous research in our lab, ovarian cancerresistant cell lines were established to screening drugresistant genes [6]. Most of them are associated with metabolism, especially folate metabolism. Meanwhile, we have reported that overexpression of MTRR plays an important role in cisplatin resistance, and silencing MTRR expression partially reverses cisplatin-resistant phenotype [7]. In the present study, the expression levels of FOLR1, DHFR, and MTRR were examined in OC tissues, and the prognostic ability of these three proteins was investigated and compared.
The aim of this study was to validate changes in the levels of FOLR, DHFR, and MTRR in the tissue of OC patients. We sought to determine whether folate metabolism enzymes could serve as a novel biomarker for early diagnosis and prognosis of platinum-resistant OC patients, as well as their clinical significance in OC. 2.2. Western Blotting Analysis. 160 fresh specimens were sonicated with an ultrasonic tissue disrupter in lysis buffer for 30 min. The tissue debris was pelleted by centrifugation, and supernatants were collected. After measuring the protein concentration by BCA protein assay, proteins were subjected to SDS-PAGE and then transferred onto PVDF membranes. After blocking, the membranes was treated with 5% (w/v) BSA in PBST (PBS, pH 7.5, containing 0.1% Tween-20) and then incubated with primary antibodies MTRR (1 : 1000; Santa Cruz Biotechnology), FOLR1 (1 : 1200; Abcam), and DHFR (1 : 1000; Santa Cruz Biotechnology) overnight at 4°C. These PVDF membranes were subsequently treated with PBST and incubated with peroxidase-conjugated secondary antibody (1 : 1000) (Santa Cruz Biotechnology) for 1 h. Visualization was detected by using a chemiluminescence system (Pierce, USA) according to the manufacturer's instructions. The band intensities were quantified using the ImageQuant software (Molecular Dynamics, Sunnyvale, CA, USA). Then, the membranes were stripped and reincubated with anti-GAPDH (1 : 1000; Santa Cruz Biotechnology) for normalization.

Statistical Analysis.
Nominal variables were compared using the χ 2 test, and ordinal categorical variables were evaluated by a nonparametric Spearman's rank test. Receiver-operating characteristic (ROC) curves were established to evaluate the diagnostic value of FOLR1, MTRR, and DHFR for differentiating benign and malignant, and the cutoff values were also calculated. According to the cutoff value of the ROC curve, we defined the result of protein below the cutoff value as low expression and above the cutoff value as high expression. OS curves were plotted by the Kaplan-Meier method and compared by log-rank test. The assessment of correlation between survival time and multiple clinicopathological variables was carried out by the Cox proportional hazards regression model. Univariate and multivariate Cox proportional hazard models were used to identify variables associated with OS in the group of OC patients. When the significant variables associated with OS were obtained by univariate analysis, multivariate analysis was used for evaluating which variables were the most important in prediction of OS. A P value less than 0.05 was considered as statistically significant. All of the statistical calculations were performed using the SPSS software (19.0, Chicago, IL, USA), and GraphPad Prism 5.0 (GraphPad Software Inc., CA) was used to generate graphs. All of the P values <0.05 were considered to be statistically significant.

Results
3.1. Expression of FOLR, DHFR, and MTRR. FOLR, DHFR, and MTRR expression in tissues of 160 patients was assessed by Western blot analysis ( Figure 1). Quantitative analysis of Western blotting analysis showed that FOLR1 and MTRR have the highest expression in OC tissues, while DHFR has the highest expression in benign tumor tissues (Figure 1(a)). DHFR and MTRR expression was clearly elevated in platinum-resistant OC compared with platinumsensitive OC, while FOLR1 expression in platinum-resistant OC was lower than in platinum-sensitive OC (Figure 1(b)). The correlation between the expression of FOLR, DHFR, MTRR and the kinds of ovarian tissue is summarized in Table 1. The expression of FOLR, DHFR, and MTRR was significantly correlated with the kinds of ovarian tissue (P < 0 05, respectively).
In order to detect the location, FOLR, DHFR, and MTRR expression in clinical samples was assessed by immunohistochemical staining of sections isolated from 10 OC, 10 patients with benign tumors OC, and 10 patients with normal ovaries. Representative examples of staining are shown in Figure 2. Immunohistochemistry showed that FOLR-, DHFR-, and MTRR-positive cells had brown granules in the cytoplasm. FOLR1 and MTRR had moderate to strong expression in OC tissues, whereas DHFR demonstrated little or no immunoreactivity in OC tissues (data not shown).  Table 2. The results showed that the expression of the FOLR and MTRR proteins did not correlate with omentum metastasis and ascites. However, the expression of DHFR was significantly correlated with the omentum metastasis and outcome (P < 0 05, respectively).
3.4. Correlation between the Expression of FOLR, DHFR, and MTRR. FOLR, DHFR, and MTRR were located and coexpressed in the cytoplasm of the OC tissue ( Figure 2). The correlation analysis between the expression of FOLR, DHFR, and MTRR in the OC tissues is summarized in Table 4. The results showed that positive MTRR expression was significantly associated with positive FOLR1 and DHFR expression (P < 0 001 and <0.01, respectively). Correlation was also found between positive FOLR1 and positive DHFR (P < 0 05).

Cox Proportional Hazard Regression Models of Risk
Factors Associated with OS among OC Patients. The results of univariate and multivariate analysis are presented in Table 5. Histology, node status, omentum metastasis, organ metastasis, FOLR1, and DHFR are not significant predictive factors for the prognosis of OC patients as determined by univariate analysis (P > 0 05). However, FIGO Stage, grade, ascites, platinum resistance, and high expression of MTRR were significant predictive factors for the prognosis of OC patients (P < 0 05 for all, Table 5). In the multivariate Cox analysis of OS, FIGO Stage, grade, ascites, platinum resistance, and high expression of MTRR were not independent predictive risk factors for the prognosis of OC patients (P > 0 05).

Survival Analysis and Prognostic
Significance of the Expression of FOLR, DHFR, and MTRR. To establish survival curves, continuous expression levels of FOLR1, DHFR, and MTRR were converted to a dichotomous variable, using their cutoff values from ROC curve analyses as a threshold,   Notes: * : metastasis to any one of liver, lung, brain and spleen.

Discussion
The folate metabolic pathway comprises a cycle mediated by FOLR1, DHFR, and MTRR ( Figure 5). Aberrant promoter methylation has been linked to the development of OC poor prognosis or clinical severity in several cancer types, including platinum resistance [8][9][10]. Our results indicated that the abnormal folate metabolic pathway could offer advantages for platinum resistance and proliferation in OC. FOLR1 overexpresses in a wide range of epithelial malignant cancers [11][12][13][14]. Some studies have developed that serum FOLR1 is a biomarker for ovarian cancer with implications for diagnosis, prognosis, and prediction of OC [11,15,16]. Song et al. demonstrated that suppression   of FOLR1 reversed taxol resistance in nasopharyngeal carcinoma cell lines [17]. Moreover, in neuroendocrine tumor, low FOLR1 expression was also identified as a marker for more aggression and associated with shorter OS and PFS [18]. In this study, we found that FOLR1 was upregulated in OC tissues compared with the normal ovarian tissues and benign ovarian tumor as well. However, high levels of FOLR1 were not associated with shorter OS of OC patients.
Interfering with the expression of DHFR is an approach in improving pharmacokinetics and reversing drug resistance in OC [19,20]. In NSCLC, osteosarcoma, and lymphoblastic leukemia, DHFR expression or polymorphism also has been associated with sensitivity to drug resistance [21][22][23][24]. Additionally, non-small cell lung carcinoma patients with low DHFR expression had a longer median PFS and OS compared with patients with a higher DHFR expression. However, the difference was not statistically significant [25]. Our data also demonstrated that overexpression of DHFR is associated with platinum resistance in OC. Similarly, overexpression of DHFR was not associated with shorter OS of OC patients.
MTRR (rs1801394) has been linked to many cancers [26][27][28]. But early studies reported no significant association between polymorphisms in the MTRR genes and ovarian cancer risk [29][30][31]. In our previous study [7], we found that the overexpression of MTRR is related to cisplatin resistance, most probably because of inducing apoptosis and reducing autophagy in OC cells. In this study, MTRR's actual impact on clinical outcome of OC patients is still very scarce and incomplete. In this study, we confirmed the presence of increased MTRR expression in the majority of our platinum-resistant patients. The OS difference was found between patients with high and low MTRR expression.
Our results showed that the combined ROC analysis revealed an AUC value of 0.832 in discriminating platinum-resistant patients from platinum-sensitive patients. This analysis provided information for an effective combined prognostic approach for drug resistance. To our knowledge, this is the first report showing the relationship between platinum resistance and expressions of the three folate metabolism enzymes in OC. The limitations of this study need to be presented. First is the lack of other types of histology disease and a small sample size of OC patients, which could lead to the lack of power and the consequent imprecision. Second, although the biological functions of the three folate metabolism enzymes have been inferred by previous gene functional analysis, the mechanisms behind the predictive values of these three folate metabolism enzymes in OC are still not clear, and their functional roles should be further explored in experimental studies.
In summary, the evaluation of FOLR1, DHFR, and MTRR expression may provide useful information for doctors to make optimal clinical decisions. MTRR is a better factor for determining the prognosis of OC. Combining the three enzymes diagnosis might contribute to the identification of patients who are likely to develop chemotherapy resistance and may be a novel potential target for OC therapy, which will require analysis by further validation studies.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
We declare that we have no financial or personal relationships with other people or organizations that can inappropriately influence our work and that there is no professional or other personal interest of any nature in any product, service, and/or company that could be construed as influencing the manuscript.