High expressions of BCL6 and Lewis y antigen are correlated with high tumor burden and poor prognosis in epithelial ovarian cancer

Abstract

Aberrant regulation of BCL6 plays crucial oncogenic roles in various malignant tumors; howbeit, the function of BCL6 in tumorigenesis of ovarian cancer remains unclear. The aim of this study is to investigate the role of BCL6 in ovarian cancer. The methods of immunohistochemical staining, quantitative real-time polymerase chain reaction, immunocytochemical staining, and gene expression profile enrichment analysis were performed to identify the possible role of BCL6 in ovarian cancer. We observed that the expression of BCL6 was significantly higher in ovarian cancer tissues and correlated with higher tumor burden including advanced International Federation of Gynecology and Obstetrics stages, poor differentiation, Type II ovarian cancer, the presence of >1 cm residual tumor size, and appearance of recurrence or death (all p < 0.05). The expression patterns of Lewis y were similar to these of BCL6. Multivariate Cox analysis demonstrated that advanced International Federation of Gynecology and Obstetrics stage, lymph node metastasis, residual tumor size >1 cm, as well as high expressions of BCL6 and Lewis y antigen were independent factors of worse progression-free survival and overall survival (all p < 0.05). There was a positive correlation of the expressions of BCL6 and Lewis y antigen. The associated genes with BCL6 in response to Lewis y antigen were identified, including four upregulated genes (SOCS3, STAT1, PPARG, and GADD45A) and three downregulated genes (ACAN, E2F3, and ZBTB7B). In conclusion, the high expressions of BCL6 and Lewis y antigen are associated with development, high tumor burden, and worse prognosis of ovarian cancer and targeting BCL6 could be a novel therapeutic strategy for ovarian cancer treatment.

Keywords

Epithelial ovarian cancer BCL6 Lewis y antigen prognosis gene expression profile

Introduction

Ovarian cancer is the second cause of death among female reproductive malignancies and claims 140, 200 lives every year worldwide.¹ Most of the ovarian cancers are diagnosed at an advanced stage due to its innocuous symptoms such as abdominal distension and discomfort at the onset.² Thus, searching the early detector and prognostic predictor of ovarian carcinoma is vitally important to facilitate the early diagnosis or screen, surveillance of the reaction to the clinical therapy, and disease recurrence.

B-cell CLL/lymphoma 6 (BCL6), a highly conserved zinc finger (ZF) transcriptional factor, was first identified as an oncogene by gene expression profile in non-Hodgkin’s lymphoma in 2000;³ it exhibits pivotal roles in germinal center formation and regulation of lymphocyte function, differentiation, and survival.⁴ In recent years, more and more studies identified that BCL6 was overexpressed not only in B-cell lymphoma⁵ but also in gastric cancer,⁶ gallbladder cancer,⁷ breast cancer,⁸ colorectal cancer,⁹ head and neck cancer,¹⁰ and so on. However, there are quite a few studies on the function of BCL6 in ovarian cancer, and it remains unclear whether BCL6 may aberrantly express or be a potential prognosis marker in ovarian cancer. As an important component of glycoproteins and glycolipids, glycosyl antigen is widely expressed in the cell membrane,¹¹ and changes in the antigens are significantly associated with the variations of characteristics on tumor cell surface, resulting in the cell carceration, invasion, and migration.^11,12 Lewis y antigen, a tumor-associated carbohydrate antigen (TACA), expressed in over 75% of ovarian epithelial neoplasm.¹¹ α-1,2-Fucosyltransferase (FUT1) is a key enzyme in the synthesis of Lewis y.¹² In our previous studies, we have already transfected FUT1 gene into ovarian cancer cell line RMG-1 to establish RMG-1-hFUT cell line with high Lewis y expression, and we noted that the RMG-1-hFUT cells exhibit increased malignant biological behaviors, such as cell proliferation, adhesion, metastasis, and drug resistance.^13–15 Moreover, recently, we performed an item of whole genome microarray analysis to identify the alteration of gene expression profile in response to Lewis y antigen in ovarian cancer cells, and we discovered that BCL6 was the one upregulated gene that exhibited significantly differential expression.¹⁶ We speculated that BCL6 would exert an oncogene role in ovarian cancer.

In this study, we detected the expression of BCL6 and Lewis y antigen in ovarian cancer tissues and cells, and we also analyzed how the expressions of two markers correlate with clinical significance including prognosis; moreover, we conducted an enrichment analysis on basis of gene expression profile to find out the associated genes of BCL6 in ovarian cancer.

Materials and methods

Ethics statement

Samples were fully encoded to protect patient confidentially. The study was approved by the Ethical Committee of Shengjing Hospital of China Medical University (number of approval: 2010PS84K).

Patients and tissue samples

Selected paraffin-embedded samples (170 cases in total) were obtained from the operations performed from 2000 to 2012 in the Department of Gynecology of Shengjing Hospital of China Medical University. All tissue sections were examined by specialists to obtain a definitive diagnosis, in which 103 cases were malignant, 22 borderlines, 25 benign, and 20 normal ovarian tissues. A total of 20 normal ovarian samples were obtained from tissue excised in squamous cervical cancer operations. The mean age of overall patients was 50.2 (16–78) years. The median and age range of the ovarian cancer group was 53.0 (16–77) years; the borderline, benign, and normal ovarian group was 39.0 (19–78) years, 45.0 (28–68) years, and 46.0 (31–61) years, respectively (Table 1). There is no statistical difference among the ages of these groups (p > 0.05). Among the 25 benign ovarian tumors, 14 cases were mucinous and 11 cases were serous cystadenomas. Among the 22 borderline tumors, 12 cases were mucinous and 10 cases were serous cystadenomas. All the 103 ovarian malignant samples were obtained from primary debulking operations; no patient received chemotherapy or radiation therapy before operation. According to the 2010 International Federation of Gynecology and Obstetrics (FIGO) staging system for ovarian cancer, there were 33 patients in stage I, 11 patients in stage II, 55 patients in stage III, and 4 patients in stage IV. In total, 10 cases were well-differentiated, 32 were moderately differentiated, and 61 were poorly differentiated. By histological analyses,¹⁷ 45 patients were serous, 15 were mucinous, 8 were endometrioid, 10 were clear-cell, and 25 were undifferentiated carcinomas. According to dualism model classification,¹⁸ 54 patients were Type I ovarian cancer (including 21 cases of low-grade serous, 15 cases of mucinous, 8 cases of endometrioid, and 10 cases of clear-cell carcinomas) and 49 patients were Type II ovarian cancer (including 24 cases of high-grade serous and 25 cases of undifferentiated carcinomas). There were 17 patients who had lymph node metastasis, and 37 patients whose residual tumor size were more than 1 cm. General clinical and pathological information of patients was shown in Tables 1 and 3.

Table 1.

Ovarian tissue patient characteristics.

Variables	Overall	Malignant	Borderline	Benign	Normal
Patient samples	170	103	22	25	20
Age (mean years (SD))	50.2 ± 11.7	52.8 ± 10.3	42.6 ± 16.9	47.1 ± 10.1	48 ± 8.6
Age (median years (range))	50.0 (16–78)	53.0 (16–77)	39.0 (19–78)	45.0 (28–68)	46.0 (31–61)

SD: standard deviation.

Immunohistochemical staining and quantification

Consecutive paraffin-embedded sections of these cases were prepared and immunohistochemistry (IHC) was used to analyze the expression of BCL6 and Lewis y antigen in all these ovarian tissues. Streptavidin–biotin complex (SABC) and streptavidin–peroxidase (SP) kits were used and performed according to their manuals. Rabbit monoclonal anti-BCL6 antibody (EP529Y, ab33901; Abcam, USA) and mouse monoclonal anti-Lewis y antibody (F3, ab3359; Abcam, USA) were used, and the working dilutions were 1:150 and 1:200, respectively. Negative controls were performed by omission of the primary antibody or incubation with an isotype control antibody, whereas the positive controls were performed as follows: a normal tonsil tissue sample for BCL6 and a colon cancer sample for Lewis y antigen. The presence of brown-colored granules in the cells was taken as a positive staining and was classified according to color intensity as follows: not colored, light yellow, brown, and tan were scored as 0, 1, 2, and 3, respectively. Five high-power fields in series from each slice were observed; the mean percentages of the stained cells were recorded and scored as follows: a positive cell rate of less than 5% was 0, 5%–25% was 1, 26%–50% was 2, 51%–75% was 3, and more than 75% was 4. The final score was determined by multiplying the two values: 0–2 was considered negative (−), 3–4 was (+), 5–8 was (++), and 9–12 was (+++); − and + was considered as low expression, ++ and +++ as high expression. Two observers read the sections to control error.

Cell lines and cell culture

Human epithelial ovarian cancer cell line OVCAR3 was purchased from American Type Culture Collection (Manassas, VA, USA). α-1,2-Fucosyltransferase (FUT1) gene was transfected to OVCAR3 cells to establish cell line OVCAR3-H with high expression of Lewis y antigen, and OVCAR3-M was the cell line transfected with the empty plasmid alone. Transfection was carried out with a vector transfection kit according to the instructions that were shown in our previous studies.^19,20 All cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37°C, 5% CO₂ in humidified air.

RNA isolation and quantitative real-time polymerase chain reaction

Total messenger RNA (mRNA) was extracted from ovarian cancer cells using Trizol reagent. Complementary DNA (cDNA) was then synthesized using the real-time polymerase chain reaction (RT-PCR) kit (Invitrogen, China) according to manufacturer’s protocol. Quantitative RT-PCR was performed on Roche LightCycler 480 (Roche Diagnostics, Mannheim, Germany) sequence detection system. The PCR primer sequences were designed (Invitrogen) and chemically synthesized. The sequences of the primers for BCL6 was, sense, CTGCAGATGGAGCATGTTGT, and antisense, TCTTCACGAGGAGGCTTGAT; FUT1, sense, AGGTCATCCCTGAGCTGAAACGG, and antisense, CGCCTGCTTCACCACCTTCTTG. The housekeeping gene GAPDH (sense, AATCCCATCACCATCTTCCA, and antisense, TGGACTCCACGACGTACTCA) served as the constitutive control. PCR reactions of each sample were done in triplicate. Data were analyzed through the comparative threshold cycle (CT) method.

Immunocytochemistry

Cells cultured on slides were fixed with 4% paraformaldehyde for 30 min and then stained according to the instructions for the SABC kit. The working dilutions for BCL6 and Lewis y antigen were 1:150 and 1:50, respectively. The primary antibody was replaced by rabbit IgG for negative control. Images were obtained using a fluorescence microscope at magnification ×400.

Gene expression profile analysis and enrichment analysis

Using human whole genome oligonucleotide microarrays, gene expression profile alteration in response to Lewis y antigen was investigated and 215 differentially expressed genes (DEGs) were identified and validated in Caov-3 transfected with FUT1 compared with its empty-plasmid–transfected cells, in which 122 genes were upregulated and 93 genes were downregulated.¹⁶ BCL6 was among the upregulated genes in response Lewis y. Gene Ontology (GO) term enrichment analysis was performed ulteriorly in the 122 upregulated genes to find out the gene functions that BCL6 involved. Finally, the relationship between BCL6 and the other DEGs in response to Lewis y antigen in ovarian cancer was evaluated by uploading all these DEGs into the Qiagen’s Ingenuity^® Pathway Analysis IPA^®.

Statistical analysis

SPSS program (Version 22 for Mac; SPSS Inc., Chicago, IL, USA) and Graph Pad Prism 6 (version 6.0 h for Mac; Graph Pad Prism Software Inc., San Diego, CA, USA) were used for statistical analysis. Independent t test, chi-square (χ²) and Spearman correlation were employed. Quantitative data are presented as mean ± standard deviation (SD). Survival analysis was analyzed using Kaplan–Meier curves, and significant differences among different clinicopathological variants and immunomarkers were tested using the log-rank test. Cox proportional hazards regression models were used to control for confounding variables. Follow-up time was calculated from the date of surgery to the date of progression, death, and last visit or contact with the patient. Overall survival (OS) was defined as the time interval between the date of surgery and the date of death; progression-free survival (PFS) was defined as the time interval between the date of surgery and the date of identification of progressive disease (disease not treatable with curative intent). For all these end points, the last date of follow-up (September 2015) was used for censored subjects. A p < 0.05 was considered statistically significant.

Results

Expression patterns of BCL6 and Lewis y in overall ovarian tissues

BCL6 was mainly localized in the cell nucleus and cytoplasm (Figure 1), and the high expression rates in malignant, borderline, benign, and normal ovarian tissues were 69.9%, 59.1%, 28.0%, and 5%, respectively (Table 2). Malignant group showed the highest positive expression rate and significantly higher than that of the benign and normal groups (p < 0.001, Table 2). Similarly, the high expression rate in borderline was apparently higher than that in benign and normal groups, respectively (p < 0.05, Table 2). Lewis y antigen was mainly detected in the cell membrane and cytoplasm (Figure 1). Similar expression inclinations were also observed in Lewis y antigen, the high expression rates in malignant, borderline, benign, and normal ovarian tissue groups were 67.0%, 54.5%, 16.0%, and 5.0%, respectively (Table 2), and the malignant group maintained the highest high expression rate and obviously higher than that in benign and normal groups (p < 0.001); however, no significant difference was observed between the benign and normal groups (p = 0.243).

Figure 1.

Representative images of BCL6 and Lewis y antigen staining in ovarian tissues. Immunopositivity for BCL6 and Lewis y antigen is represented by brown staining. (a and e) malignant ovarian tumor, (b and f) borderline ovarian tumor, (c and g) benign ovarian tumor, (d and h): normal ovarian tissue. (a–d) BCL6 and (e–h) Lewis y antigen. Scale bar: 50 μm.

Table 2.

The expression of BCL6 and Lewis y in different ovarian tissues.

Groups	Cases	BCL6					Lewis y
Groups	Cases	−	+	++	+++	High (%)	−	+	++	+++	High (%)
Malignant	103	17	14	33	39	72 (69.9)^a	14	20	33	36	69 (67.0)^b
Borderline	22	6	3	10	3	13 (59.1)^c	5	5	9	3	12 (54.5)^d
Benign	25	8	10	6	1	7 (28.0)^e	18	3	4	0	4 (16.0)
Normal	20	12	7	1	0	1 (5.0)	16	3	1	0	1 (5.0)

Compared with the benign and normal group, p = 0.00011, p < 0.001, respectively.

Compared with the benign and normal group, p < 0.001, p < 0.001, respectively.

Compared with the benign and normal group, p = 0.031, p = 0.0002, respectively.

Compared with the benign and normal group, p = 0.005, p = 0.001, respectively.

Compared with the normal group, p = 0.045.

Correlation of BCL6 and Lewis y antigen expression with clinicopathological features of ovarian cancer

The high expression patterns of BCL6 were upregulated in tumors with a higher tumor burden, as defined by advanced FIGO stages (vs Stages I and II, p < 0.001), poor differentiation (vs moderate or well, p < 0.001), Type II ovarian cancer (vs Type I, p = 0.013), presence of >1 cm residual tumor size (vs ≤1 cm, p = 0.001), as well as the appearance of recurrence (p < 0.001) or death (p = 0.003). The high expression rate of BCL6 in serous carcinoma was significantly higher than that of in mucinous (86.7% vs 20%, p < 0.001); moreover, the endometrioid, clear-cell, and undifferentiated carcinomas maintained the high expression rates of BCL6. The high expression of BCL6 in lymphatic node group (88.2%) was higher than that of the non-lymphatic node metastasis group (66.3%), although the difference was not significant (p = 0.071). There was no obvious correlation between the expression of BCL6 and age (p > 0.05).

The high expression levels of Lewis y antigen were also upregulated in advanced FIGO stages (p = 0.020), the existence of recurrence (p = 0.027) or death (p = 0.017), as well as the presence of lymph node metastasis (p = 0.041). The high expression rate of Lewis y in serous carcinoma was significantly higher than that of in mucinous (75.6% vs 40.0%, p = 0.011). However, no obvious differences were observed in the high expression of Lewis y as to age, differentiation, dualism model classification, or residual tumor size, although high expression of antigen was observed in the >1 cm (78.4%) compared with ≤1 cm residual tumor size (60.6%) (Table 3).

Table 3.

Comparison of clinicopathological characteristics and expression of BCL6 and Lewis y of 103 cases epithelial ovarian cancer.

Characteristics	N	BCL6 expression		p value	Lewis y expression		p value
Characteristics	N	Low	High	p value	Low	High	p value
N	103	31	72		34	69
Age (years)
Mean ± SD	52.84 ± 10.29	51.39 ± 13.73	53.47 ± 8.44	0.438^a	52.65 ± 9.86	52.94 ± 10.57	0.621^a
Age group, n (%)
≤52.8	46	13 (28.3)	33 (71.7)	0.715^b	14 (30.4)	32 (69.6)	0.618^b
>52.8	57	18 (31.6)	39 (68.4)		20 (35.1)	37 (64.9)	0.618^b
FIGO stage, n (%)
I and II	44	27 (61.4)	17 (38.6)	<0.001^{*, b}	20 (45.5)	24 (54.5)	0.020^b
III and IV	59	4 (6.8)	55 (93.2)		14 (23.7)	45 (76.3)	0.020^b
Differentiation, n (%)
Well	10	9 (90.0)	1 (10.0)	<0.001^{*, b}	3 (30.0)	7 (70.0)	0.807^b
Moderate	32	9 (28.1)	23 (71.9)		12 (37.5)	20 (62.5)
Poor	61	13 (21.3)	48 (78.7)		19 (31.1)	42 (68.9)
Pathological subtype, n (%)
Serous	45	6 (13.3)	39 (86.7)	<0.001^{*, b}	11 (24.4)	34 (75.6)	0.092^b
Mucinous	15	12 (80.0)	3 (20.0)		9 (60.0)	6 (40.0)
Endometrioid	8	3 (37.5)	5 (62.5)		2 (25.0)	6 (75.0)
Clear cell	10	3 (30.0)	7 (70.0)		2 (20.0)	8 (80.0)
Undifferentiated	25	7 (28.0)	18 (72.0)		10 (40.0)	15 (60.0)
Dualism model
Type I	54	22 (40.7)	32 (59.3)	0.013^{*, b}	17 (31.5)	37 (68.5)	0.729^b
Type II	49	9 (18.4)	40 (81.6)		17 (34.7)	32 (65.3)	0.729^b
Lymph node metastasis, n (%)
No	86	29 (33.7)	57 (66.3)	0.071^b	32 (37.2)	54 (62.8)	0.041^{*, b}
Yes	17	2 (11.8)	15 (88.2)		2 (11.8)	15 (88.2)	0.041^{*, b}
Residual tumor size, n (%)
≤1 cm	66	27 (40.9)	39 (59.1)	0.001^{*, b}	26 (39.4)	40 (60.6)	0.066^b
>1 cm	37	4 (10.8)	33 (89.2)		8 (21.6)	29 (78.4)	0.066^b
Recurrence
No	60	27 (45.0)	33 (55.0)	<0.001^{*, b}	25 (41.7)	35 (58.3)	0.027^{*, b}
Yes	43	4 (9.3)	39 (90.7)		9 (20.9)	34 (79.1)	0.027^{*, b}
Death
No	72	28 (38.9)	44 (61.1)	0.003^{*, b}	29 (40.3)	43 (59.7)	0.017^{*, b}
Yes	31	3 (9.7)	28 (90.3)		6 (16.1)	26 (83.9)	0.017^{*, b}

FIGO: International Federation of Gynecology and Obstetrics; SD, standard deviation.

p < 0.05.

Independent t test.

Chi-square test.

Relevance of BCL6 and Lewis y expression in ovarian tissues

Totally, there were 54 and 16 cases simultaneously showing high and low expression patterns both in BCL6 and Lewis y antigen in all the 103 cases of ovarian cancer and Spearman correlation analysis revealed that these two markers positively correlated (r = 0.260, p = 0.008); moreover, when pooled all these ovarian tissues, there was 64 and 55 cases simultaneously showing high and low expression patterns in all the 170 cases of ovarian tissues and Spearman analysis further showed their positive correlation (r = 0.401, p < 0.001) (Table 4).

Table 4.

The relevance of BCL6 and Lewis y expression in ovarian tissues.

Lewis y	BCL6
Lewis y	Low	High	Total
	Ovarian cancer samples^a
Low	16	18	34
High	15	54	69
Total	31	72	103
	Pooled ovarian samples^b
Low	55	29	84
High	22	64	86
Total	77	93	170

Spearman correlation, r = 0.260, p = 0.008.

Spearman correlation, r = 0.401, p < 0.001.

Follow-up visit and prognostic factors analysis

During the period of follow-up for the ovarian cancer patients, 31 patients (30.1%) were dead, 12 patients (11.6%) were alive with recurrence of the disease, 52 patients (50.5%) were alive without evidence of disease, and 8 patients (7.8%) were lost. The median follow-up was 60.0 (range: 33.2–85.5) months, the 5-year PFS was 59.5% (Figure 2(a)), and mean survival time was 58.7 months (95% confidence interval (CI): 51.8–65.6); the 5-year OS was 68.7% (Figure 2(d)), and mean survival time was 70.0 months (95% CI: 64.5–75.6).

Figure 2.

Kaplan–Meier curves for progression-free survival (PFS) and overall survival (OS) in 103 ovarian cancer patients. (a) PFS curve and (d) OS curve for all the ovarian patients, the 5-year survival rate and mean survival time are shown in figures. PFS and OS according to the low and high expression of (b and e) BCL6 and (c and f) Lewis y.

Cox survival analyses indicated that among various prognostic factors, FIGO stages III and IV, lymph node metastasis, residual tumor size >1 cm, high expressions of BCL6, and Lewis y antigen were found to be significant risk factors that correlated with PFS (hazard ratio (HR): 8.466, 2.329, 4.889, 5.752, and 2.124, respectively, all p < 0.05, Table 5). As to OS, similarly, FIGO stages III and IV, lymph node metastasis, residual tumor size >1 cm, high expressions of BCL6, and Lewis y antigen were also found to be significant risk factors (HR: 6.342, 2.519, 3.414, 5.191, and 2.799, respectively, all p < 0.05, Table 5). Taken together, except the clinical pathological parameters, the high expressions of both BCL6 and Lewis y antigen were independent risk factors and significantly correlated with prognosis (Kaplan–Meier analysis, log rank, p < 0.001 and p = 0.002 as to PFS, Figure 2(b) and (e); p = 0.038 and p = 0.027 as to OS, Figure 2(c) and (f), respectively).

Table 5.

Cox survival analysis of factors affecting patient’s progression-free survival or overall survival for clinicopathologic variables and the IHC expression in 103 patients of EOC.

Variables	PFS			OS
Variables	B	HR (95% CI)	p	B	HR (95% CI)	p
Age (>52.8-year-old)	0.149	1.161 (0.635–2.123)	0.627	0.394	1.484 (0.718–3.065)	0.287
FIGO stage (III and IV)	2.136	8.466 (3.322–21.571)	<0.001*	1.847	6.342 (2.215–18.159)	0.001*
Differentiation (poor)	0.162	1.176 (0.738–1.875)	0.495	0.310	1.364 (0.755–2.461)	0.303
Pathological subtype (undifferentiated)	−0.031	0.969 (0.807–1.164)	0.739	−0.100	0.905 (0.724–1.131)	0.381
Dualism model (Type II)	0.074	1.077 (0.592–1.958)	0.809	0.177	1.194 (0.590–2.416)	0.622
Lymph node metastasis (Yes)	0.846	2.329 (1.171–4.634)	0.016*	0.924	2.519 (1.157–5.485)	0.02*
Residual tumor size (>1 cm)	1.587	4.889 (2.602–9.187)	<0.001*	1.228	3.414 (1.657–7.037)	0.001*
BCL6 expression (high)	1.749	5.752 (2.051–16.129)	0.001*	1.647	5.191 (1.576–17.097)	0.007*
Lewis y expression (high)	0.753	2.124 (1.018–4.433)	0.045*	1.029	2.799 (1.075–7.291)	0.035*

IHC: immunohistochemistry; EOC: epithelial ovarian cancer; PFS: progression-free survival; OS: overall survival; HR: hazard ratio; CI: confidence interval; FIGO: International Federation of Gynecology and Obstetrics.

p < 0.05.

The mRNA and protein expression levels of BCL6 in ovarian cancer cells after Lewis y overexpression

The mRNA expression levels of FUT1 and BCL6 were significantly elevated after FUT1 gene transfection (p < 0.001, p = 0.002, respectively), whereas there was no obvious change between OVCAR3 and OVCAR3-M cell lines (p >0.05, Figure 3). As to the protein expression level, both the Lewis y antigen and BCL6 showed the positive staining in ovarian cancer cells OVCAR3, Lewis y antigen was mainly localized in cytoplasm and membrane (Figure 4(a)), and BCL6 was mainly localized in the nucleus and cytoplasm (Figure 4(c)). In the FUT1-transfected cell line OVCAR3-H, the Lewis y antigen showed significant higher expression (Figure 4(b)), whereas the BCL6 was also obvious highly expressed (Figure 4(d)).

Figure 3.

The mRNA expression of FUT1 and BCL6 in OVCAR3 cells after FUT1 gene transfection. Quantitative real-time PCR revealed that the mRNA expression levels of FUT1 and BCL6 were significantly increased after the transfection of FUT1 (p <0.001, p =0.002, respectively), whereas there was no change between OVCAR3 and OVCAR3-M cell lines.

Figure 4.

Representative images of immunocytochemical staining of BCL6 and Lewis y antigen in OVCAR3 cells. Immunopositivity for BCL6 and Lewis y antigen is represented by brown staining, Lewis y antigen was mainly localized in cytoplasm and membrane, where BCL6 was predominantly localized in nucleus and cytoplasm. (a and c) OVCAR3 cells, (b and d) OVCAR3-H cells. (a and b) Lewis y antigen, (c and d) BCL6. Scale bar: 50 μm.

GO function analysis and interaction network of BCL6 involved in the DEGs in response to Lewis y

GO function enrichment analysis was performed in the upregulated DEGs based on molecular function, cellular component, and biological process (Table 6). The interaction network of genes associated with BCL6 was presented in Figure 5, including four upregulated genes (SOCS3, STAT1, PPARG, and GADD45A) and three downregulated genes (ACAN, E2F3, and ZBTB7B).

Table 6.

GO enrichment analysis that BCL6 involved for upregulated DEGs in response to Lewis y antigen in ovarian cancer cells.

GO term	Genes in overlap	p value
Biological process
Positive_regulation_of_cellular_process	13	9.97e−8	EIF5A, PLK2, BCL6, CAMKK2, BIRC5, IP6K2, RUNX1, VEGFA, EGR1, PPARG, MNAT1, CAPNS1, BTG1
Positive_regulation_of_biological_process	13	1.97e−7	BCL6, BIRC5, CAMKK2, EIF5A, PLK2, CAPNS1, BTG1, VEGFA, RUNX1, IP6K2, MNAT1, PPARG, EGR1
Apoptosis_go	10	6.59e−7	BIRC5, BCL6, DNAJB13, EIF5A, BTG1, GADD45A, STAT1, SOCS3, VEGFA, IP6K2
Programmed_cell_death	10	6.73e−7	SOCS3, VEGFA, IP6K2, BTG1, STAT1, GADD45A, EIF5A, BIRC5, DNAJB13, BCL6
Regulation_of_developmental_process	10	7.94e−7	STAT1, BTG1, RUNX1, IP6K2, VEGFA, PPARG, SOCS3, BCL6, BIRC5, EIF5A
Regulation_of_apoptosis	8	8.28e−6	SOCS3, VEGFA, IP6K2, BTG1, STAT1, EIF5A, BIRC5, BCL6
Regulation_of_programmed_cell_death	8	8.46e−6	EIF5A, BCL6, BIRC5, VEGFA, IP6K2, SOCS3, STAT1, BTG1
Cell_development	10	8.72e−6	IP6K2, VEGFA, SOCS3, GADD45A, STAT1, BTG1, EIF5A, BCL6, DNAJB13, BIRC5
Positive_regulation_of_developmental_process	6	4.8e−5	PPARG, EIF5A, IP6K2, RUNX1, BTG1, BCL6
Positive_regulation_of_cell_differentiation	3	5.44e−5	BTG1, PPARG, RUNX1
Cellular component
Intracellular_organelle_part	14	1.2e−5	IMP3, ARPC1B, BIRC5, PPOX, ABCF2, BCL6, EIF5A, KLC1, CETN3, ARPC2, SEC22A, BSCL2, PI4KA, TNNT2
Organelle_part	14	1.26e−5	PI4KA, BSCL2, TNNT2, ARPC2, SEC22A, CETN3, KLC1, EIF5A, PPOX, ARPC1B, BIRC5, IMP3, BCL6, ABCF2
Intracellular_non_membrane_bound_organelle	8	5.82e−4	ARPC2, TNNT2, BIRC5, ARPC1B, IMP3, BCL6, CETN3, KLC1
Non_membrane_bound_organelle	8	5.82e−4	ARPC2, TNNT2, BCL6, BIRC5, ARPC1B, IMP3, KLC1, CETN3
Nucleus	12	1.08e−3	BCL6, IMP3, RPSA, EIF5A, GADD45A, STAT1, LGALS3, BTG1, NR1D1, CEBPB, IP6K2, EGR1
Molecular function
No data

GO: Gene Ontology; DEGs: differentially expressed genes.

Figure 5.

Gene network generated by the IPA^® platform for BCL6-associated genes. Genes marked with red and green represent the upregulated and downregulated genes, respectively. Dotted lines indicate indirect interactions and solid lines indicate direct interaction between the genes.

Discussion

BCL6, a nuclear transcriptional repressor, can regulate transcription of various genes. Located on chromosome 3q27, the BCL6 gene encodes a 706 amino acid protein composed of an amino-terminal bric-a-brac, tramtrack, broad complex/Pox virus ZF (BTB/POZ) domain, and a carboxy-terminal six ZF domain,^4,21 these domains regulate transcription of target genes via distinct interactions. BCL6 exerts its pivotal roles for regulating germinal center development, proliferation/differentiation of lymphocytes, inflammation, and cell cycle/apoptosis.⁴ Initially, researchers found BCL6 mainly abundantly expressed in B cell and CD4⁺ T cell in germinal centers of normal lymphoid tissues,²² and the association of BCL6 with lymphoma has been widely investigated and well established,^3,5,22 especially in diffuse large B-cell lymphoma,²³ the pathogenesis mechanisms attribute to rearrangements or somatic mutations of BCL6, and so on.^4,24,25 BCL6 inhibits p53 expression in germinal center of B lymphocyte²⁶ and interacts with pro-oncogene c-myc to accelerate the carcinogenesis of lymphoma.²⁷ Based on these findings, BCL6 inhibitors (compound 79-6²⁸ or retroinverso BCL6 peptide inhibitor (RI-BPI)²⁹) have been developed and showed delightful potential application prospect as a therapeutic strategy for patients with hematological malignancies.^28,29 In recent years, overexpression of BCL6 was reported to be relevant to somatic malignant tumors, such as gastric cancer,⁶ gallbladder cancer,⁷ breast cancer,⁸ colorectal cancer,⁹ and head and neck cancer,¹⁰ suggesting that the tumorigenicity of BCL6 is not only limited to lymphomas.

BCL6 protein was highly expressed both in cell lines and tissue specimens of breast cancer⁸ and it enhanced the proliferation, growth, migration and survival of breast cancer cells through the regulation of epithelial–mesenchymal transition (EMT) by enhancing the expression of ZEB1 which bound to the E-cadherin promoter and repressing the E-cadherin transcription.³⁰ The ZEB1, a crucial transcriptional repressor, inhibits the transcription of E-cadherin to regulate the EMT response, which is also prominent in non–small-cell lung cancer (NSCLC)³¹ and colorectal cancer.³² At the gene level, BCL6 was a target gene of miR-339-5p^8,33,34 or miR-187-3p,³⁵ these microRNAs inhibit proliferation, migration, and invasion in breast cancer,³⁴ NSCLC,³⁵ or ovarian cancer cells.³³ In gallbladder cancer, BCL6 overexpression correlated with high tumor, node, and metastasis (TNM) stage, higher histological grade, vascular invasion, decreased p19^ARF expression, and worse survival.⁷ However, until now, there are quite a few studies on the potential biological roles of BCL6 in ovarian cancer. BCL6 was found to be bound to the cytoplasmic domain of heparin-binding epidermal growth factor–like growth factor (HB-EGF)³⁶ to suppress cell cycle,³⁷ and HB-EGF played a critical role in tumor formation and growth in patients with ovarian cancer.³⁸ Recently, BCL6 was identified to be one of the upregulated genes in a bioinformatics study designing to identify the molecular mechanisms of NSC319726 (a small-molecule anticancer drug functioning to the p53^R175 protein) in ovarian cancer,³⁹ which revealed that BCL6 may play an efficient role in ovarian cancer. One study reported that BCL6 is a negative prognostic factor and exerts pro-oncogenic characteristics in ovarian cancer,⁴⁰ high expression was tightly correlated with tumor grades, lymph node metastasis and recurrence, BCL6-stimulated proliferation, invasion, and migration in ovarian cancer cells. Our data further confirmed the tumorigenesis roles of BCL6 in ovarian tumor and demonstrated that BCL6 was correlated with high tumor burden, and the high expression was obviously increased along with the tumor carcinogenesis progression including late FIGO stages, poor differentiation, Type II ovarian cancer, >1 cm residual tumor size, and recurrence or death in ovarian cancer. BCL6 was a good and independent predictor both in PFS and OS. The discrepancy between expression pattern and tumor grading/differentiation in different pathological types suggests that the mechanism that relates BCL6 to carcinogenesis in ovarian cancer may be tumor-type specific.

As a kind of TACA, Lewis y antigen was upregulated at a different extent in more than 75% of epithelial ovarian cancer tissues, and patients with high expression of Lewis y usually have bad prognosis.¹⁵ In our previous studies, we have transfected FUT1 gene into human ovarian cancer cell line RMG-1 to construct cell line RMG-1-H with high expression of fucosyltransferase and Lewis y;^13,15 the overexpression of Lewis y antigen can aberrantly activate downstream signaling pathways, such as epidermal growth factor receptor (EGFR), PI3K/Akt, Ras/ERK, p38/MAPK, Smad,^{15,19,41–43} and anomalously regulate the expression of cellular factors, such as cell cycle proteins, autophagy-related factors, and apoptosis-related factors; thus enhance the malignant biological behaviors including proliferation, adhesion, apoptosis, drug resistance, autophagy, invasion, and metastasis; and ultimately promote the occurrence and development of ovarian cancer.^41–47 For further comprehending the fundamental molecular mechanism of Lewis y antigen in ovarian cancer, we have performed a gene expression profile analysis using human whole genome oligonucleotide microarrays in ovarian cancer cell line Caov-3 transfected with FUT1 compared with its empty-vector transfected cells and we identified 215 DEGs according to the selection criteria, in which 122 genes were upregulated and 93 genes were downregulated.¹⁶ BCL6 gene was among the upregulated DEGs in response to Lewis y. And here, in this study, we further demonstrate that the expression of BCL6 and Lewis y antigen is positively correlated with each other both at mRNA and protein levels; moreover, our enrichment analysis including GO term analysis and interaction network may provide crucial elements for successive concentrated investigation on BCL6 in ovarian cancer. Lewis y antigen could enhance the adhesion and metastasis capabilities of ovarian cancer cells by glycosylated modification of CD44 structure;⁴⁸ intriguingly, CD44s, one of CD44 splice isoforms, can activate the expression of ZEB1, resulting in a CD44s–ZEB1 regulatory loop to accelerate EMT process in cancer cells.⁴⁹ Likewise, BCL6 can increase the expression of ZEB1 to regulate the EMT process for enhancing the proliferation and migration in breast cancer cells;⁸ moreover, BCL6 can interact with HB-EGF by binding to its cytoplasmic domain.³⁶ Increased expression of Lewis y promotes cell proliferation and migration through activation of EGFR signal pathway.^15,50,51 Accordingly, there is a good possibility that Lewis y regulates the expression of CD44 via BCL6 by activation of EGFR signal to enhance the malignant behaviors in ovarian cancer, whereas the mechanisms are still warranted more in-depth and concentrated investigations.

Collectively, our results demonstrate that there is a positive correlation between the expression of BCL6 and Lewis y, and high expression of these two markers are significantly associated with adverse clinicopathological factors and bad prognosis, and BCL6 is an independent prognosticatory factor for ovarian cancer. Gene expression enrichment analysis shows the associated genes of BCL6, which can provide the fundamental data for in-depth investigation. These results suggest that BCL6 play a critical role in tumorigenesis of ovarian cancer, thus to provide new ideas for early detection and surveillance for ovarian cancer treatment.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China (no: 81172491, 81101527); the Research Fund for the Doctoral Program of Higher Education (no: 20112104110016, 20112104120019); Scientific Research Projects of Liaoning Province Department of Education (no: L2011129); and Outstanding Scientific Fund of Shengjing Hospital (no: 201303).

References

Jemal

Bray

Center

. Global cancer statistics. CA Cancer J Clin 2011; 61: 69–90.

Pliarchopoulou

Pectasides

Epithelial ovarian cancer: focus on targeted therapy. Crit Rev Oncol Hematol 2011; 79: 17–23.

BH.

BCL-6 in the pathogenesis of non-Hodgkin’s lymphoma. Cancer Invest 2000; 18: 356–365.

Jardin

Ruminy

Bastard

. The BCL6 proto-oncogene: a leading role during germinal center development and lymphomagenesis. Pathol Biol 2007; 55: 73–83.

Wang

. Triple-hit B-cell lymphoma with MYC, BCL2, and BCL6 translocations/rearrangements: clinicopathologic features of 11 cases. Am J Surg Pathol 2015; 39: 1132–1139.

Hirata

Ogasawara

Sasaki

. BCL6 degradation caused by the interaction with the C-terminus of pro-HB-EGF induces cyclin D2 expression in gastric cancers. Br J Cancer 2009; 100: 1320–1329.

Liang

Chen

. BCL6 overexpression is associated with decreased p19^ARF expression and confers an independent prognosticator in gallbladder carcinoma. Tumour Biol 2014; 35: 1417–1426.

Liu

Yan

. B-cell lymphoma 6 protein stimulates oncogenicity of human breast cancer cells. BMC Cancer 2014; 14: 418.

Sena

Mariani

Benincasa

. Morphological and quantitative analysis of BCL6 expression in human colorectal carcinogenesis. Oncol Rep 2014; 31: 103–110.

10.

Worsham

Chen

. Malignant and nonmalignant gene signatures in squamous head and neck cancer. J Oncol 2012; 2012: 752860.

11.

Miyoshi

Moriwaki

Nakagawa

Biological function of fucosylation in cancer biology. J Biochem 2008; 143: 725–729.

12.

Iwamori

Tanaka

Kubushiro

. Alterations in the glycolipid composition and cellular properties of ovarian carcinoma-derived RMG-1 cells on transfection of the α1,2-fucosyltransferase gene. Cancer Sci 2005; 96: 26–30.

13.

Zhang

Liu

Lin

. Increase in docetaxel-resistance of ovarian carcinoma-derived RMG-1 cells with enhanced expression of Lewis Y antigen. Int J Mol Sci 2011; 12: 7323–7334.

14.

Gao

. Elevated levels of Lewis Y and integrin α₅β₁ correlate with chemotherapeutic drug resistance in epithelial ovarian carcinoma. Int J Mol Sci 2012; 13: 15588–15600.

15.

Yan

Lin

Zhu

. Enhancement of the adhesive and spreading potentials of ovarian carcinoma RMG-1 cells due to increased expression of integrin α₅β₁ with the Lewis Y-structure on transfection of the α1,2-fucosyltransferase gene. Biochimie 2010; 92: 852–857.

16.

Gao

Zhu

Feng

. Gene expression profile analysis in response to α1,2-fucosyl transferase (fut1) gene transfection in epithelial ovarian carcinoma cells. Tumour Biol 2016; 37: 12251–12262.

17.

Kaku

Ogawa

Kawano

. Histological classification of ovarian cancer. Med Electron Microsc 2003; 36: 9–17.

18.

Koshiyama

Matsumura

Konishi

Recent concepts of ovarian carcinogenesis: type I and type II. Biomed Res Int 2014; 2014: 934261.

19.

Yan

Wang

Lin

. Lewis y enhances CAM-DR in ovarian cancer cells by activating the FAK signaling pathway and upregulating Bcl-2/Bcl-XL expression. Biochimie 2015; 113: 17–25.

20.

Liu

Lin

. Lewis y regulate cell cycle related factors in ovarian carcinoma cell RMG-I in vitro via ERK and Akt signaling pathways. Int J Mol Sci 2012; 13: 828–839.

21.

Dhordain

Albagli

Ansieau

. The BTB/POZ domain targets the LAZ3/BCL6 oncoprotein to nuclear dots and mediates homomerisation in vivo. Oncogene 1995; 11: 2689–2697.

22.

Bentebibel

Schmitt

Banchereau

. Human tonsil B-cell lymphoma 6 (BCL6)-expressing CD4⁺ T-cell subset specialized for B-cell help outside germinal centers. Proc Natl Acad Sci U S A 2011; 108: E488–E497.

23.

Wagner

Ahearne

Ko Ferrigno

The role of BCL6 in lymphomas and routes to therapy. Br J Haematol 2011; 152: 3–12.

24.

Jarosova

Kriegova

Schneiderova

. A novel non-immunoglobulin (non-Ig)/BCL6 translocation in diffuse large b-cell lymphoma involving chromosome 10q11.21 loci and review on clinical consequences of BCL6 rearrangements. Pathol Oncol Res 2016; 22: 233–243.

25.

Landsburg

DJ.

Management of patients with MYC-altered lymphomas. Curr Hematol Malig Rep 2016; 11: 208–217.

26.

Duy

Nahar

. BCL6 is critical for the development of a diverse primary B cell repertoire. J Exp Med 2010; 207: 1209–1221.

27.

Wang

Bossler

Schaffer

. A novel t(3;8)(q27;q24.1) simultaneously involving both the BCL6 and MYC genes in a diffuse large B-cell lymphoma. Cancer Genet Cytogenet 2007; 172: 45–53.

28.

Cerchietti

Ghetu

Zhu

. A small-molecule inhibitor of BCL6 kills DLBCL cells in vitro and in vivo. Cancer Cell 2010; 17: 400–411.

29.

Duy

Hurtz

Shojaee

. BCL6 enables Ph⁺ acute lymphoblastic leukaemia cells to survive BCR-ABL1 kinase inhibition. Nature 2011; 473: 384–388.

30.

Sun

Hua

. BCL6 induces EMT by promoting the ZEB1-mediated transcription repression of E-cadherin in breast cancer cells. Cancer Lett 2015; 365: 190–200.

31.

Gemmill

Roche

Potiron

. ZEB1-responsive genes in non-small cell lung cancer. Cancer Lett 2011; 300: 66–78.

32.

Ono

Imoto

Kozaki

. SIX1 promotes epithelial-mesenchymal transition in colorectal cancer through ZEB1 activation. Oncogene 2012; 31: 4923–4934.

33.

Shan

Bai

. MiR-339-5p inhibits migration and invasion in ovarian cancer cell lines by targeting NACC1 and BCL6. Tumour Biol 2016; 37: 5203–5211.

34.

Wang

. MiR-339-5p inhibits breast cancer cell migration and invasion in vitro and may be a potential biomarker for breast cancer prognosis. BMC Cancer 2010; 10: 542.

35.

Sun

Yang

. MicroRNA-187-3p mitigates non-small cell lung cancer (NSCLC) development through down-regulation of BCL6. Biochem Biophys Res Commun 2016; 471: 82–88.

36.

Dhordain

Albagli

Honore

. Colocalization and heteromerization between the two human oncogene POZ/zinc finger proteins, LAZ3 (BCL6) and PLZF. Oncogene 2000; 19: 6240–6250.

37.

Miyamoto

Hirata

Yamazaki

. Heparin-binding EGF-like growth factor is a promising target for ovarian cancer therapy. Cancer Res 2004; 64: 5720–5727.

38.

Kinugasa

Hieda

Hori

. The carboxyl-terminal fragment of pro-HB-EGF reverses Bcl6-mediated gene repression. J Biol Chem 2007; 282: 14797–14806.

39.

Xue

Yang

Ding

. Role of NSC319726 in ovarian cancer based on the bioinformatics analyses. Onco Targets Ther 2015; 8: 3757–3765.

40.

Wang

Weng

. BCL6 is a negative prognostic factor and exhibits pro-oncogenic activity in ovarian cancer. Am J Cancer Res 2015; 5: 255–266.

41.

Liu

Lin

Hao

. Lewis y antigen promotes the proliferation of ovarian carcinoma-derived RMG-I cells through the PI3K/Akt signaling pathway. J Exp Clin Cancer Res 2009; 28: 154.

42.

Liu

Wang

. The stimulation of IGF-1R expression by Lewis(y) antigen provides a powerful development mechanism of epithelial ovarian carcinoma. Int J Mol Sci 2011; 12: 6781–6795.

43.

Liu

. Lewis Y regulates signaling molecules of the transforming growth factor β pathway in ovarian carcinoma-derived RMG-I cells. Int J Oncol 2012; 40: 1196–1202.

44.

Gao

Zhang

. High expression of Lewis y antigen and CD44 is correlated with resistance to chemotherapy in epithelial ovarian cancers. PLoS ONE 2013; 8: e57250.

45.

Gao

Liu

. Expression of Lewis Y antigen and integrin αv, β3 in ovarian cancer and their relationship with chemotherapeutic drug resistance. J Exp Clin Cancer Res 2013; 32: 36.

46.

Zhu

Gao

. Membranous expressions of Lewis y and CAM-DR-related markers are independent factors of chemotherapy resistance and poor prognosis in epithelial ovarian cancer. Am J Cancer Res 2015; 5: 830–843.

47.

Zhu

Zhuang

Wang

. Overexpression of HE4 (human epididymis protein 4) enhances proliferation, invasion and metastasis of ovarian cancer. Oncotarget 2016; 7: 729–744.

48.

Gao

Yan

Lin

. Enhancive effects of Lewis y antigen on CD44-mediated adhesion and spreading of human ovarian cancer cell line RMG-I. J Exp Clin Cancer Res 2011; 30: 15.

49.

Preca

Bajdak

Mock

. A self-enforcing CD44s/ZEB1 feedback loop maintains EMT and stemness properties in cancer cells. Int J Cancer 2015; 137: 2566–2577.

50.

Shan

Aziz

Tian

. Ginsenoside Rg3-induced EGFR/MAPK pathway deactivation inhibits melanoma cell proliferation by decreasing FUT4/LeY expression. Int J Oncol 2015; 46: 1667–1676.

51.

Lin

Shi

. Lewisy promotes migration of oral cancer cells by glycosylation of epidermal growth factor receptor. PLoS ONE 2015; 10: e0120162.