Mining expression and prognosis of FOLR1 in ovarian cancer by using Oncomine and Kaplan-Meier plotter

Abstract Objective: To further explore folate receptor 1 (FOLR1) gene expression in ovarian cancer and its association with patients’ prognosis by deep mining the Oncomine and Kaplan-Meier plotter databases. Methods: FOLR1 mRNA expression data of ovarian cancer were retrieved from the Oncomine database and further analyzed by comparing tumor to healthy tissue. The prognostic value of FOLR1 in ovarian cancer was analyzed by Kaplan-Meier Plotter, an online survival analysis database. Results A total of 439 studies were included in the Oncomine database in multiple types of cancers. Of the 439 studies, there were 54 with statistical differences for the expression of FOLR1, 19 with increased expression of FOLR1 and 35 with decreased expression comparing ovarian cancer to normal ovary tissue. After searching the Oncomine database, six datasets were discovered comparing the mRNA expression in ovarian tumor to healthy tissue. FOLR1 mRNA expression in ovarian tumor was significantly higher than that of normal ovarian tissue (all p<0.05). The Kaplan-Meier Plotter database analyzed the correlation between FOLR1 expression and ovarian cancer patient’s prognosis. A significant difference of progression-free survival between FOLR1 high and low expressing groups was found in ovarian cancer patients (HR=1.14, 95%CI: 1.00-1.29, p=0.043). However, the overall survival was not statistically different between high and low FOLR1 expressing patients (HR=0.95, 95%CI: 0.84-1.09, p=0.48). Conclusion FOLR1 mRNA was found to be highly expressed in ovarian tumor compared to normal ovarian tissue. Elevated FOLR1 mRNA expression was associated with the poor progression-free survival.


Introduction
Folate receptor 1 (FOLR1), also known as folate binding protein, is a glycosylphosphoinositol anchored membrane protein, which can mediate extracellular 5-methyltetrahydrofolate into cells and plays an important role in cell division, proliferation and tissue growth [1,2]. In normal tissue the gene expression level of FOLR1 is low, however, in many cancers such as lung [3,4], breast [5] and colorectal [6] the FOLR1 expression level is elevated. In ovarian cancer, the FOLR1 gene is over expressed in tumor compared to normal ovarian tissues [7]. However, most previous studies used small sample sizes with limited statistical power and weak clinical evidence. In our study, microarray FOLR1 mRNA expression data from the Oncomine database was used to determine prognosis.
The Oncomine database (https://www.oncomine. org/) is the largest oncogene chip databases with an integrated data-mining platform aimed at mining cancer gene information [8,9]. At present, 715 gene expression datasets and 86,733 samples of cancerous and normal tissues have been made available. Oncomine can provide differentially expressed genes between tumor and normal tissues for many cancer types and subtypes which can be analyzed based on clinical and pathological data, differentially expressed sorting and co-expression analyses, location of differentially expressed genes, identification of target genes. These analyses can then be used to determine research directions, not only to save time but also reducing the cost of scientific research.

Data extraction from Oncomine database and analysis
The Oncomine database is a web-based gene database and an integrated data mining platform, in which data can be screened and mined in accordance with requirements. This platform aimed to stimulate discoveries from genomewide expression analyses and compare the transcriptome data in most major cancer types. In our present work, folate receptor 1 (FOLR1) gene expression levels in ovarian cancer were compared to normal ovarian tissue and extracted and analyzed by the Oncomine database platform. The data search was as follows: (1) cancer type: ovarian cancer; (2) Gene: FOLR1; (3) Data type: mRNA; (4) Analysis type: Cancer vs normal; (5) Cut-off value: p<1E-4, fold change>2; gene rank=top 10%.

Survival analysis by Kaplan-Meier Plotter
The prognosis, in terms of overall survival (OS) and progression-free survival (PFS), was determined using tumor FOLR1 mRNA expression levels for ovarian cancer patients and analyzed by the Kaplan-Meier plotter (www. kmplot.com). This database provided online survival analysis of lung, ovarian, gastric and breast cancer patients. In order to evaluate the prognostic significance of FOLR1 expression on ovarian patient survival, the patients were divided into two cohorts according to median FOLR1 mRNA expression (high versus low expression). The OS and PFS of the ovarian cancer patients were compared between FOLR1 high and low expressing groups.

Statistical analysis
Data were analyzed by SPSS17.0 (http://www-01.ibm. com/software/analytics/spss/) and expressed by x̄±s. The between group comparison of the FOLR1 expression was conducted using a student t-test, and the counted data was expressed by rate. The between group comparison was made by χ 2 -test or Fisher's exact test. A proportional hazard ratio was used for survival analysis through a log-rank test and survival curve. P < 0.05 was taken as a statistically significant difference.

FOLR1 expression in multiple cancers
A total of 439 studies of multiple cancers types were included from the Oncomine database (Figure 1). Of the 439 studies, there were 54 with statistical differences for the expression of FOLR1, 19 with increased expression of FOLR1 and 35 with decreased expression when comparing ovarian tumors to normal ovarian tissues.

FOLR1 expression in ovarian cancer
After searching the Oncomine database, six datasets [10][11][12][13][14][15] were discovered after comparing the mRNA expression in tumor with normal tissue ( Table 1). FOLR1 was highly expressed in ovarian tumors when compared to normal ovarian tissue in all of the six series, Figure 2.

FOLR1 expression by microarray analysis
FOLR1 mRNA expression in ovarian tumor versus normal ovarian tissue by microarray analysis of the six data sets were shown in Figure 3. FOLR1 mRNA expression in ovarian tumor was significantly higher than that from normal ovary (all p<0.05).

FOLR1 expression and prognosis
The correlation between FOLR1 expression and ovarian cancer patients prognosis was analyzed by the Kaplan-Meier Plotter database [16]. A significant difference between PFS and FOLR1 high and low expressing groups was found in ovarian cancer patients (HR=1.14, 95%CI: 1.00-1.29, p=0.043). However, the OS was not statistically significantly different between high and low FOLR1 expressing patients (HR=0.95, 95%CI: 0.84-1.09, p=0.48), Figure 4. We further performed subgroup survival analysis of FOLR1 expression according to clinical stage and histology, Table 2. The prognostic significance of FOLR1 expression for subgroup analysis was demonstrated in Figure 5. The PFS was different between FOLR1 high and low expressing groups in stage 3 and 4 subgroups and    endometrioid histology subgroup (p<0.05). For OS, there was a statistical difference in stage 3 and 4 subgroups and serous histology subgroup (p<0.05).

Discussion
Ovarian tumors are one of the most common carcinomas in female genital organs [17]. The incidence of ovarian cancer ranks third, after cervical and uterine body cancers [18,19]. However, the mortality rate of epithelial ovarian cancer is the highest among all gynecological carcinomas [20]. The cancer epidemiological study from the United States in 2019, showed that there were 22,530 new cases of ovarian cancer and 19,380 deaths, and only ~40% of ovarian cancer patients were suitable for curative resection [17]. Epithelial cancer is the most common malignant ovarian tumors, followed by malignant germ cell tumors [21]. In patients with ovarian cancer, only 30% of the tumors were discovered to be confined to the ovary [22,23], the majority had spread to the uterus, bilateral appendages, omentum and pelvic organs [23,24]. The prognosis of patients with ovarian cancer is poor when diagnosed in advanced stages, and much more is needed to be done to improve early diagnostic methods. There are many factors which can affect the prognosis of ovarian cancer patients such as clinical stage, pathological classification, post-operative treatment, and surgical methods [25,26]. Studies have also shown that the expression of a variety of genes negatively correlates to patient prognoses, such as CA-125 [27,28] and FOLR1 [29].
FOLR1 is a member of the human folate-binding protein family (three isomers of folate-binding protein have been identified: FOLR1, FOLR2 and FOLR3). The gene is located on chromosome 11q13.3-13.5. Its protein is exposed to the outside of the cell and anchored by GPI on the cell membrane. FOLR1 transports extracellular 5-methyltetrahydrofolate into the cell and participates in DNA replication and replication as a coenzyme component. The expression level of FOLR1 is closely related to tumor progression and cell proliferation. Several studies have evaluated FOLR1 expression and ovarian cancer patient prognosis, but the results were inconclusive [29][30][31]. Kimberly et al. [32] examined FOLR1 protein expression in 186 primary ovarian cancer patients and 27 cases of recurrence disease by immunohistochemical assay. The authors found that the expression of FOLR1 did not correlate with the OS of the ovarian cancer patients. However, Fu et al. [33] evaluated the correlation between FOLR1 mRNA expression and prognosis in patients with serous ovarian carcinoma by RT-PCR. They found high expressing FOLR1 was correlated with poor prognosis, which was in accordance with our results.
In our present work, six comparisons of tumor versus normal tissue from ovarian cancer patients were included here. FOLR1 gene expression in ovarian cancer was analyzed and compared with normal ovarian tissue. Of the six data sets, three were from three studies which compared three different ovarian cancer subtypes. All six data sets confirmed FOLR1 gene expression levels from tumor were higher when compared to healthy control tissue. This may indicate that FOLR1 plays an important role in ovarian cancer development. We also found that elevated FOLR1 mRNA expression was associated with poor PFS. The correlation between elevated FOLR1 mRNA expression and poor prognosis indicated that FOLR1 mRNA could be used as a molecular biomarker for the prediction of ovarian cancer patient prognosis.

Conclusion
The deep mining of FOLR1 gene expression confirmed that the gene is highly expressed in ovarian tumors, and negatively correlates with ovarian cancer patient prognosis.