This article is mentioned in:

Introduction

Granulosa cell tumors (GCTs) are malignant neoplasms, which originate from the sex cord. GCT is a relatively rare disease, but commonly occurs in the ovary. GCTs frequently occur in young women, and the majority of GCT cases are diagnosed in the late stages due to early stage GCT exhibiting few clinical symptoms, leading to a poor prognosis (1,2). Due to the current lack of understanding of this particular disease, the clinical management of GCTs is similar to that of other types of epithelial ovarian cancer. However, unlike other types of ovarian cancer, the prognostic assessment for GCT predominantly relies on clinicopathological variables, including stage and grade, although these do not provide biological insight into the disease (3). Thus, early diagnosis is critical for the successful management of numerous types of human cancer, including ovarian GCTs.

The identification of molecular or epigenetic markers may provide biological insight into GCT and a serve as a critical tool in successful treatment of the disease. Current knowledge indicates that the development and progression of cancer are driven by the accumulation of genetic abnormalities and epigenetic alterations, which include gene mutations and silencing, and epigenetic modifications of genomic DNA, including methylation of DNA CpG islands or covalent modification of histone tails (4,5). Gene promoter methylation silences the expression of tumor suppressor genes, which is one of the key mechanisms in tumor development (6,7). Aberrant gene methylation is also one of the earliest molecular alterations occurring in tumorigenesis, and is considered a biomarker for early tumor detection or a potential treatment strategy (8–10). Histone modification also regulates genetic programs in normal cells, but is altered in cancer cells (11). Therefore, the pathogenesis of GCT may follow a similar trend to other types of human cancer.

In the present study, the methylation status of three putative tumor suppressor genes, cadherin 13 (CDH13), dickkopf WNT signaling pathway inhibitor 3 (DKK3) and forkhead box L2 (FOXL2), and expression of enhancer of zeste homolog 2 (EZH2) were assessed in GCT tissue samples and compared with follicular cyst tissues. The aim of the present study was to screen and identify tumor markers for the early detection of GCT. As GCTs are rare and reports are limited, the present study focused on CDH13, DKK3 and FOXL2, as these are putative tumor suppressor genes, whose functions and expression are associated with the ovary, including ovary development, function maintenance. The present study aimed to provide novel information to improve current understanding of the development of GCT, and to potentially identify biomarkers for the early detection of GCT.

Materials and methods

Patients and samples

In the present study, 31 patients with GCTs were recruited from Shandong University Qilu Hospital (Jinan, China) and Shanxian Central Hospital (Heze, China) between 2010 and 2013. All patients were diagnosed histologically with GCT, and all tissue specimens were reconfirmed by pathologists in the Department of Pathology, Qilu Hospital, which resulted in 30 cases being available for use in the study. The present study also included tissues from 30 patients with follicular cysts, which were selected as a control. Clinicopathological data from each patient, including age, tumor size, Federation of Gynecology and Obstetrics (FIGO) stage and postoperative recurrence, were collected from medical records. The present study was approved by the Ethics Committee of Shandong University School of Medicine (Jinan, China) and the patients or their guardians provided signed informed consent prior to involvement in the investigation. Written informed consent was obtained from patients.

Methylation-specific polymerase chain reaction (MSP)

A total of 31 patients with GCTs were recruited from Shandong University Qilu Hospital (Jinan, China) and Shanxian Central Hospital (Heze, China) between 2010 and 2013. All patients were diagnosed histologically with GCT, and all tissue specimens were reconfirmed by pathologists in the Department of Pathology, Qilu Hospital. One tissue wax block was not large enough, which resulted in only 30 cases being available for use in the present study. For MSP, two 8-µm tissue sections were prepared from the 30 paraffin-embedded tissue blocks and deparaffinized in xylene, following which tumor cells were dissected from sections for genomic DNA extraction. Specifically, genomic DNA was extracted using a Genomic DNA Purification kit (Qiagen, Hilden, Germany) and subjected to bisulfite conversion using a CpGenome DNA Modification kit (Intergen Co., Purchase, NY, USA), according to the manufacturer's protocols. Subsequently, 2 µl of the modified DNA (50 ng) was subjected to PCR amplification in a 50 µl volume reaction [0.25 µl Taq polymerase, 5 µl 10X PCR buffer; 4 µl dNTP mix (2.5 mM); 0.5 µl forward and reverse primers; 50 ng DNA template made up to the volume in Milli Q water] under the following conditions: 45 cycles at an annealing temperature of 58°C for 45 sec (CDH13), 60°C for 45 sec (DKK3) or 58°C for 45 sec (FOXL2), and primer extension at 72°C. All PCR amplifications were performed with positive controls for unmethylated and methylated alleles, and DNA-free empty controls. The PCR amplification kit was purchased from Eppendorf AG (Hamburg, Germany). The primers (Sangon Biotech., Co., Ltd., Shanghai, China) for each gene promoter methylation were designed according to previous reports (12–14) by first identifying the methylated- and unmethylated-specific sequences, respectively, and subsequent synthesis for MSP amplification of the CDH13, DKK3 and FOXL2 genes (Table I). The PCR products were then separated on 3% agarose gels and visualized using ethidium bromide staining under an UV-3000 ultraviolet light box.

Table I Primer sequences for PCR amplification of multiple tumor suppressor genes and methylation-specific PCR analysis of gene promoter methylation.

Table I

Primer sequences for PCR amplification of multiple tumor suppressor genes and methylation-specific PCR analysis of gene promoter methylation.

Gene	Sequence	Amplicon (bp)	Temperature (°C)
CDH13-M-F	5′-TCGCGGGGTTCGTTTTTCGC-3′	243	58
CDH13-M-R	5′-GACGTTTTCATTCATACACGCG-3′
CDH13-U-F	5′-TTGTGGGGTTTGTTTTTTGT-3′	242
CDH13-U-R	5′-AACTTTTCATTCATACACACA-3′
DKK3-M-F	5′-GGGGCGGGCGGCGGGGC-3′	120	60
DKK3-M-R	5′-ACATCTCCGCTCTACGCCCG-3′
DKK3-U-F	5′-TTAGGGGTGGGTGGTGGGGT-3′	126
DKK3-U-R	5′-CTACATCTCCACTCTACACCCA-3′
FOXL2-M-F	5′-GTTATAATATTTTTTCGGTTGTTC G-3′	211	58
FOXL2-M-R	5′-CTAACTCCACGACCTATACTCGAT-3′
FOXL2-U-F	5′-AGGTTATAATATTTTTTTGGTTGTTTG-3′	214
FOXL2-U-R	5′-CCTAACTCCACAACCTATACTCAAT-3′

[i] PCR, polymerase chain reaction; CDH13, cadherin 13; DKK3, dickkopf WNT signaling pathway inhibitor 3; FOXL2, forkhead box L2; M, methylated; U, unmethylated; F, forward; R, reverse.

Immunohistochemistry

Immunohistochemistry was performed to detect the protein expression of EZH2 in the tissue samples. In brief, 3-µm thick tissue sections were prepared from the paraffin-embedded tissue blocks, deparaffinized in xylene and rehydrated in a series of ethanol. The sections were then incubated with 0.5% TritonX-100 for 30 min at room temperature to ensure that the antibody entered the nuclei. The sections then underwent epitope retrieval in a steam cooker in 0.01 M citric buffer (pH 6.0) for 15 min at 100°C. The slides were subsequently washed with phosphate-buffered saline (pH 7.4) three times for 5 min. The slides were immersed in 3% H2O2 methanol solution (freshly prepared) for 10 min. The slides were subsequently washed as above. Following being blocked in normal serum for 30 min, the sections were incubated with anti-EZH2 antibody at a dilution of 1:100 (cat. no. 5246; Cell Signaling Technology, Inc., Danvers, MA, USA) at 4°C overnight. The following day, the sections were washed with phosphate-buffered saline (PBS) three times and further incubated with a rabbit anti-human secondary antibody Dako (Glostrup, Denmark; cat. no. K5007) at room temperature for 30 min. The positive signal was visualized using diaminobenzidine as the chromogen. Breast cancer tissue sections (Shanxian Central Hospital; patient was diagnosed as infiltrating ductal carcinoma) were used as a positive control, and PBS was used as a negative control. The stained sections were reviewed and scored under a light microscope (BX43; Olympus, Tokyo, Japan) by two investigators for staining intensity and percentage of staining (15). The staining intensity score was recorded as follows: Absent, 0; weak, 1; moderate, 2; strong, 3. The percentage of positive cells was recorded as follows: Absent, 0; ≤10%, 1; 11–50%, 2; 51-≤80%, 3; >80 %, 4. These two scores were then multiplied to obtain a staining index. If the staining index was ≥3, the case was considered positive.

Statistical analysis

All statistical analyses were performed using SAS version 9.1 software (SAS Institute, Inc., Cary, NC, USA). The frequencies of methylation were compared using χ2 test or Fisher's exact test. P<0.05 was considered to indicate a statistically significant difference. All tests were two-sided.

Results

Differential methylation status of CDH13, DKK3 and FOXL2 promoters in GCTs

In the present study, the methylation status of the CDH13, DKK3 and FOXL2 promoters in the GCT tissue specimens were assessed and compared with those in follicular cyst specimens. Table II summarizes the methylation rates of these three genes in the GCTs, compared with the follicular cysts. Representative examples of the MSP data are shown in Fig. 1. Significant differences were found in the methylation of the CDH13, DKK3 and FOXL2 promoters in the GCT tissues, compared with the follicular cyst tissues (P<0.001). The associations between gene methylation and clinicopathological parameters, including age, tumor size, FIGO stage and postoperative recurrence, were also analyzed (Table III), however, no significant associations were observed.

Figure 1 MSP analysis of CDH13, DKK3 and FOXL2 promoter methylation in GCT and follicular cyst tissues. The MSP products in the M lanes indicate the presence of methylated alleles, whereas those in the U lanes indicate the presence of unmethylated alleles. MSP, methylation-specific polymerase chain reaction; T, GCT; N, follicular cyst; CDH13, cadherin 13; DKK3, dickkopf WNT signaling pathway inhibitor 3; FOXL2, forkhead box L2; U, unmethylated; M, methylated.

Table II Comparison of methylation rates of CDH13, CKK3 and FOXL-2 promoters between GCT and follicular cyst tissues.

Table II

Comparison of methylation rates of CDH13, CKK3 and FOXL-2 promoters between GCT and follicular cyst tissues.

Gene	GCT (n=30)	Follicular cyst (n=30)	χ2	P-valuea
CDH13	86.67 (26)	23.33 (7)	21.70	<0.001
DKK3	80 (24)	26.67 (8)	17.14	<0.001
FOXL2	70 (21)	20 (6)	17.38	<0.001

a χ² test compared with normal endometrium. GCT, granulosa cell tumor; CDH13, cadherin 13; DKK3, dickkopf WNT signaling pathway inhibitor 3; FOXL2, forkhead box L2.

Table III Association between CDH13, DKK3 and FOXL2 methylation and clinicopathological data from patients with granulosa cell tumors.

Table III

Association between CDH13, DKK3 and FOXL2 methylation and clinicopathological data from patients with granulosa cell tumors.

		DKK3 (M)		CDH13(M)		FOXL2(M)		EZH2 (M)
Clinical feature	n	n (%)	P-valuea	n (%)	P-valuea	n (%)	P-valuea	n (%)	P-valuea
Age (years)			1.00		0.55		1.00		1.00
<40	8	7 (87.5)		8 (100.0)		6 (75.0)		3 (37.5)
≥40	22	17 (77.3)		18 (81.8)		15 (68.2)		8 (36.4)
Tumor size (cm)			0.37		1.00		0.68		0.69
≥10	19	14 (73.7)		16 (84.2)		14 (73.3)		6 (31.6)
<10	11	10 (90.9)		10 (90.9)		7 (63.6)		5 (45.5)
FIGO stage			0.57		0.55		1.00		0.64
I	24	20 (83.3)		20 (83.3)		17 (70.8)		8 (33.3)
II–III	6	4 (66.7)		6 (100.0)		4 (66.7)		3 (50.0)
Recurrence			0.50		1.00		0.53		0.53
Yes	3	2 (66.7)		3 (100.0)		3 (100.0)		2 (44.4)
No	27	22 (81.5)		23 (85.2)		18 (66.7)		9 (33.3)

a Fisher's exact test. CDH13, cadherin 13; DKK3, dickkopf WNT signaling pathway inhibitor 3; FOXL2, forkhead box L2; (M), methylated; EZH2, enhancer of zeste homolog 2; FIGO, Federation of Gynecology and Obstetrics.

Differential protein expression of EZH2 in GCT tissues

The immunohistochemical staining showed that EZH2 protein was localized in the nuclei of the positive tumor cells. Representative examples of the immunohistochemical data are shown in Fig. 2. It was found that the expression of EZH2 was higher in the GCT tissues, compared with the follicular cysts (Table IV), however, no associations were found between the expression of EZH2 and the clinicopathological parameters of the patients with GCT with respect to age, tumor size, FIGO stage and metastasis (Table II). The present study then examined the associations between the expression of EZH2 and methylation of the CDH13, DKK3 and FOXL2 gene promoters, however, no positive associations were found (Table V).

Figure 2 Immunohistochemical analysis of the protein expression of enhancer of zeste homolog 2 in (A) granulosa cell tumor (positive) and (B) follicular cyst tissues (negative) (magnification, ×200).

Table IV Comparison of the expression of enhancer of zeste homolog 2 between GCT and follicular cyst tissues.

Table IV

Comparison of the expression of enhancer of zeste homolog 2 between GCT and follicular cyst tissues.

	EZH2
Tissue	+ (n)	− (n)	%	χ2	P-valuea
GCT	11	19	36.7
Follicular cyst	0	30	0	13.5	<0.001

a χ² test compared with normal endometrium. GCT, granulosa cell tumor; EZH2, enhancer of zeste homolog 2.

Table V Association between the expression of EZH2 and methylation of the CDH13, DKK3 and FOXL2 promoters.

Table V

Association between the expression of EZH2 and methylation of the CDH13, DKK3 and FOXL2 promoters.

	DKK3			CDH13			FOXL2
EZH2	M	U	P-valuea	M	U	P-valuea	M	U	P-valuea
Positive (n)	9	2		9	2		9	2
Negative (n)	15	4	1.00	17	2	0.61	12	7	0.41

a Fisher's exact test of the association between gene methylation and the protein expression of EZH2. M, methylated; U, unmethylated; EZH2, enhancer of zeste homolog 2; CDH13, cadherin 13; DKK3, dickkopf WNT signaling pathway inhibitor 3; FOXL2, forkhead box L2.

Discussion

GCTs are a relatively rare type of malignancy in the ovary, and are inconsistent in size, ranging between small spots and large masses, with an average diameter of 10 cm. Although the clinical appearance, symptoms and management are similar to those of epithelial ovarian tumors, the mechanism underlying the development and progression of GCT may be different from other types of ovarian cancer (1). Thus, an improved understanding of the mechanism underlying the development and progression of GCT may lead to improved options for the early detection, prevention and treatment of GCT clinically (16). To date, clinical prognostic indicators rely predominantly on clinicopathological variables, including patient age, tumor stage and grade. Thus, the identification of molecular or epigenetic markers may provide biological insight and serve as a critical tool in the successful treatment of GCT. The present study assessed the methylation status of the CDH13, DKK3 and FOXL2 promoters, and found that the promoters of these three genes were significantly hypermethylated in the GCT tissues, compared with the follicular cyst tissues. Expression of the EZH2 protein was also high in the GCT tissues, however, no associations were found between these alterations and the clinicopathological data from the patients with GCT. Further investigations with a larger sample size are required to confirm these findings. As GCT is a relatively rare type of tumor, a consortium of different cancer centers or hospitals may be required to obtain sufficient numbers of tissue samples to facilitate further molecular investigations.

Although the cause and pathogenesis of GCTs remain to be elucidated, GCTs are similar to the majority of other types of human cancer, the development of which involves gene mutation and promoter methylation, and epigenetic modifications of genomic DNA. The accumulation of genetic abnormalities and epigenetic alterations lead to the malignant transformation of normal cells. To date, the known primary human epigenetic modifications include alterations of DNA methylation status in CpG islands and covalent modifications of histone tails. Accumulating evidence suggests that more genes are affected by aberrant epigenetic alterations than genetic mutations in human carcinogenesis (6,7,17). Thus, promoter methylation-induced silencing of tumor suppressor genes has been suggested as a key mechanism in the development of several types of human cancer. Aberrant gene methylation is also one of the earliest molecular alterations occurring during tumorigenesis and may be used as a marker for early tumor detection. As methylation of the promoter region is a reversible process, the detection of gene methylation levels may provide guidance for individualized chemotherapy. The present study showed for the first time, to the best of our knowledge, methylation of the CDH13, DKK3 and FOXL2 gene promoters in GCT tissue samples, with no similar findings reported previously. Friedrichs et al (15) first reported a specific pattern of CpG island hypermethylation in different types of human cancer. In GCTs, the detection of different gene promoter methylation has been shown in cell lines and in a limited number of tumor tissues, with the most frequently methylated gene promoters being p16 and ER-α (40%), BRCA1 and RASSF1A (36%), MGMT (32.5%) and hMLH1 and FHIT (28%) (16,18,19).

CDH13 is a cell adherence protein of a unique cadherin superfamily member and functions to mediate intracellular signaling in vascular cells. Emerging evidence indicates that CDH13 is a candidate tumor suppressor in several types of human cancer, including breast and lung cancer (12,20–26), colorectal cancer (21,27), hepatocellular carcinoma (28), bladder cancer (29), cervical cancer (30) and ovarian cancer (31–34). Previous studies have showed that CDH13 promoter methylation is a frequent event in cancer, which is associated with unfavorable tumor features, increased risk of recurrence and poorer survival rates, and has been suggested as an independent predictor for tumor recurrence and progression (26,29). DKK3 is a secreted protein, which is involved in embryonic development through its interactions with the Wnt signaling pathway. The expression of DKK3 is reduced in a variety of cancer cell lines and may function as a tumor suppressor gene by antagonizing Wnt signaling (35–37). Epigenetic silencing of DKK3 has been observed to disrupt normal Wnt/β-catenin signaling and apoptosis regulation (38). DKK3 methylation has been frequently detected in a broad range of types of cancer and appears to be important in tumor development (13,37–42). FOXL2 is a member of the forkhead transcription factor family and functions as an essential regulator of ovarian maintenance. FOXL2 protein is expressed in the pituitary gonadotrope, thyrotrope cells and ovarian granulosa cells, and is required for commitment to ovary differentiation (43,44). FOXL2 mutations are associated with syndromic and non-syndromic ovarian failure, and occurs in ovary GSTs with a mutation rate at FOXL2 (402 C->G) of 97% in adult GCT (45–50). Tran et al showed that the CpG island of the murine FOXL2 proximal promoter was differentially methylated in primary and immortalized cells (51). The FOXL2 promoter was also abnormally methylated in non-small cell lung cancer (52). In the present study, the methylation statuses of three gene promoters in GCT tissues were detected, and the results demonstrated that promoter methylation was associated with the development of GCT, but not with its progression. Further investigations aim to investigate the underlying molecular mechanism for silencing the expression of these three genes in GCT.

Histone modifications are considered to regulate genetic programs in normal cells, but are altered in cancer cells. The methylation of histone H3 at lysine 27 silences gene expression, which induces transcriptional repression and is thus involved in controlling gene expression patterns (53,54). EZH2 is a methyltransferase and a component of the polycomb repressive complex 2, which is essential in the epigenetic maintenance of the H3K27me3 repressive chromatin mark (55). The abnormal expression of EZH2 has been associated with aggressive tumor subgroups, disease-free survival rates and overall survival rates in patients with cutaneous melanoma, and in cancer of the endometrium, prostate, breast, colorectal and ovary (11,56–59). The present study was the first, to the best of our knowledge, to detect the expression of EZH2 in GCT and found 11 positive cases in 30 GCT tissue samples (36.7%), compared with follicular cyst tissue samples, in which no positive cases were found. In addition, a previous study linked EZH2 to gene silencing in association with the maintenance of DNA methylation (58). EZH2 may affect DNA methylation by direct interaction with DNA methyltransferases, however, the majority of H3K27me3-marked genes lack DNA methylation in embryonic stem cells, indicating that EZH2 recruitment may not be sufficient to promote DNA methylation (60). In the present study, no associations were found between the expression of EZH2 and methylation of the CDH13, DKK3 and FOXL2 promoters. Thus, further investigation is required to assess the functions of the EZH2 protein in GCTs.

The results of the present study were proof-of-principle, and future investigations with a larger sample size are required to verify the findings. Future investigations aim to assess how the methylation of the CDH13, DKK3 and FOXL2 gene promoters affects the expression of their proteins, and how these proteins contribute to the development of GCT. Whether the altered methylation status of these gene promoters can be detected as biomarkers for the early detection of GCT also requires investigation.

Acknowledgments

The authors would like to thank Medjaden Bioscience Ltd. (Hong Kong, China) for their assistance with manuscript editing.

References