BRCA1 promoter hypermethylation in human placenta: a hidden link with β-hCG expression

Abstract

Gestational trophoblastic diseases (GTD) are group of pregnancy-related tumors characterized by abnormal levels of ‘β-hCG’ with higher incidence in South-East Asia, especially India. Our laboratory has reported that wild-type BRCA1 transcriptionally regulates β-hCG in triple negative breast cancers (TNBCs). These factors culminated into analysis of BRCA1 status in GTD, which would emanate into elucidation of BRCA1- β-hCG relationship and unraveling etio-pathology of GTD. BRCA1 level in GTD is down-regulated due to the over-expression of DNMT3b and subsequent promoter hypermethylation, when compared to the normal placentae accompanied with its shift in localization. There is an inverse correlation of serum β-hCG levels with BRCA1 mRNA expression. The effects of methotrexate (MTX), which is the first-line chemotherapeutic used for GTD treatment, when analyzed in comparison with plumbagin (PB) revealed that PB alone is efficient than MTX alone or MTX-PB in combination, in showing selective cytotoxicity against GTD. Interestingly, PB increases BRCA1 levels post-treatment, altering DNMT3b levels and resultant BRCA1 promoter methylation. Also, cohort study analyzed the incidence of GTD at Sree Avittom Thirunal (SAT) Hospital, Thiruvananthapuram, which points out that 11.5% of gestational trophoblastic neoplasia (GTN) cases were referred to Regional Cancer Centre, Thiruvananthapuram, for examination of breast lumps. This has lend clues to supervene the risk of GTD patients towards BRCA1-associated diseases and unveil novel therapeutic for GTD, a plant-derived naphthoquinone, PB, already reported as selectively cytotoxic against BRCA1 defective tumors.

The manuscript discusses the down-regulation of BRCA1 in GTD, owing to its promoter hypermethylation as a result of DNMT3b up-regulation, in comparison with the normal placentae, along with BRCA1-β-hCG correlation and potential therapeutic strategies for reversal of BRCA1 promoter hypermethylation.

Introduction

Gestational trophoblastic diseases (GTD) are a cluster of pregnancy-associated disorders, which arise due to fertilization defects, resulting in the over-proliferation of the trophoblasts. GTD includes benign gestational trophoblastic tumors (GTT), commonly appearing as hydatidiform moles or vesicular moles (VM) and malignant gestational trophoblastic neoplasia (GTN). They are characterized by higher levels of human chorionic gonadotrophin (hCG) (1000–30,00,000 mIU/ml) than the normal pregnancy range (5–2,88,000 mIU/ml), as they are secreted by the over-proliferating trophoblasts. hCG is a glycoprotein hormone with ‘α’ and ‘β’ subunits and β-hCG is debated for exhibiting both tumorigenic and anti-tumorigenic roles (1,2).

The GTD incidence is very high in Kerala, India, at about 4.8 per 1000 deliveries as per the statistics of GTD patients taken during 2006–2010, when compared with global incidence rates like that in the USA, which was 0.7–1 per 1000 pregnancies (3,4). The GTN incidence was 20.4% of GTD patients in 2008, while it increased to 22.03% in 2013 in Kerala (4,5).

BRCA1 mutations predispose women to hereditary breast ovarian cancers. Recent findings from our group have proved the role of β-hCG in promoting BRCA1 defective breast tumorigenesis (6,7). The current study evaluates GTD as the ideal system to analyze the inverse correlation between β-hCG and BRCA1, which we have observed in BRCA1-deficient breast tumors, since β-hCG peaks up in GTD than any other known physiological levels. Thus, we aimed to analyze the probable risk of BRCA1-associated tumors in GTD patients.

Methotrexate (MTX) is the first-line chemotherapeutic drug for the treatment of GTD that is practiced in Kerala. The increase in incidence of GTN and its relapse and amidst of MTX therapy, along with its longer therapeutic duration and side effects like nausea, dizziness, etc. calls for a better therapeutic agent for GTD (8). In the present study, plumbagin (PB), which is a plant-derived naphthoquinone, known for its anti-proliferative activity against BRCA1-related tumors, has been compared with MTX, for its effects against GTD. This study would help reveal the role of BRCA1 in GTD and the possible therapeutic approaches for better management of GTD.

Methods

Cell lines

The study involved choriocarcinoma cell lines, JEG-3 and JAR, normal trophoblast cell line, HTR-8/SV neo and triple negative breast adenocarcinomas cell line HCC1937 and its isogenic cell line HCC1937/wt BRCA1, details of which are provided as Supplementary Information, available at Carcinogenesis Online. Choriocarcinoma cell lines were purchased from ATCC (VA, USA) while breast cancer cell lines used in the study were authenticated by STR profiling.

Human tissue samples

GTD tissues were obtained after evacuation, while first-trimester age-matched normal placentae was obtained after medical termination of pregnancy from the healthy subjects from Sree Avittom Thirunal (SAT) Hospital, Government Medical College, Thiruvananthapuram, Kerala, India and the PRS Hospital, Karamana, Thiruvananthapuram, Kerala, India, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

Quantitative real-time PCR (qRT-PCR)

The mRNA expression of relevant genes was carried out by qRT-PCR. Total RNA was isolated and converted to cDNA. Relative and absolute quantification of the gene expression was done. The list of primers used has been provided as Supplementary Table S1, available at Carcinogenesis Online and further details are provided in the Supplementary Information, available at Carcinogenesis Online.

Immunoblotting analysis

To analyze the protein expression, immunoblotting was carried out for different proteins in the whole cell lysate, details of which are provided in the Supplementary Information, available at Carcinogenesis Online. The antibodies used in the study have been detailed in Supplementary Table S2, available at Carcinogenesis Online.

Enzyme-linked immunosorbent assay (ELISA)

The analysis of secretary β-hCG in the spent media of cells was carried out by ELISA, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

siRNA knockdown studies

To analyze the correlation between β-hCG and BRCA1, siRNA knockdown was carried out for both these genes independently, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

Immunofluorescence analysis (IFA)

The protein localization studies were carried out in cell lines by IFA, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

Immunohistochemical analysis (IHC)

Protein localization in the tissue samples was carried out by IHC, details of which are provided in the Supplementary Information, available at Carcinogenesis Online. Hematoxylin and Eosin staining was also performed with tissue samples to analyze their morphology.

Proximity ligation assay (PLA)

PLA was carried out to check for any interaction between β-hCG and BRCA1 in the normal as well as tumor placental cells and tissues, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

Mutation analysis for BRCA1 and BRCA2

GTD and normal placental cell lines were analyzed for any mutations in BRCA1/2 by next-generation sequencing, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

Analysis of methylation in BRCA1 promoter

Cellular DNA was isolated and bisulfite converted DNA was analyzed for methylation at the BRCA1 promoter using semi-quantitative PCR using a set of methylation and unmethylation specific primers, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

Cell proliferation assay or MTT assay

Cell proliferation or viability analysis was carried out by 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

Hormone assay

The levels of β-hCG, estrogen and progesterone in untreated and drug-treated cells were analyzed by chemiluminiscence immuno assay, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

Cohort study for follow-up analysis involving GTD patients

A cohort study was carried out to analyze the GTD incidence at SAT Hospital, Government Medical College, Thiruvananthapuram, Kerala, India, from January 2012 to February 2016. The data was further classified and analyzed against the similar data, which was collected from the normal cohorts, details of which are provided in the Supplementary Information, available at Carcinogenesis Online.

Statistical analysis

Statistical analysis was performed by one-way analysis of variance and two-tailed Student’s T-test for all expression analysis in the experiments. All the experiments were performed in technical triplicates, expressed as mean ± SD from at least three independent experiments and the data shown in SD ; P < 0.05 was considered statistically significant. Error bars are given on the basis of calculated SD values. Both Pearson and Spearman analysis was used to determine significance of inverse correlation between β-hCG and BRCA1. Statistical significance was defined as (*) P ≤ 0.05 and (**) P ≤ 0.005 and P ≤ 0.001 as (***).

Results

BRCA1 expression is down-regulated in GTD

The study was carried out to establish the inverse correlation between β-hCG and BRCA1. β-hCG expression is reported to be higher in GTD (1); thus, BRCA1 expression was suspected to be lower. As reported earlier, β-hCG expression is higher in GTD in comparison with normal placentae, in both cell lines (HTR-8/SV neo, JEG-3 and JAR) and tissue samples (n = 10 each for GTD and normal placentae) at both mRNA (Supplementary Figure S1A–B, available at Carcinogenesis Online) and protein (Supplementary Figure S1C–E, available at Carcinogenesis Online) levels. Also, greater degree of β-hCG glycosylation in the GTD tissues than in normal placental control is evident from the higher degree of streaking in the immunoblot analysis (Supplementary Figure S1F, available at Carcinogenesis Online).

BRCA1 mRNA expression is significantly lower; almost 5-fold reduced, in GTD cells (JEG-3 and JAR) than normal placental cell line, HTR-8/SV neo and about 30-fold reduction in GTD tissue (VM 1–6) than normal placentae (FT 1–6) (Figure 1A and B). At protein levels too, BRCA1 is lower; almost 3-fold, in GTD than normal placental counterparts (Figure 1C–E). Thus, BRCA1 expression is lower in GTD in comparison with the normal placentae.

Figure 1.

BRCA1 expression and correlation with β-hCG levels. (A) BRCA1 mRNA expression by qRT-PCR analysis in cell lines. Expression was normalized to HTR-8/SVneo. (B) BRCA1 mRNA expression by qRT-PCR analysis in tissue samples (VM, n = 10; FT, n = 10 and P value is calculated between the average values of expression in FT and VM). (C) BRCA1 protein expression in GTD and normal placental cell lines and tissue samples by western blot analysis. Quantification of BRCA1 protein expression by western blot analysis in (D) tissue samples (VM, n = 6 and FT, n = 6) and (E) cell lines. (E) BRCA1 - β-hCG correlation graph plotting serum β-hCG levels (mIU/ml) on x-axis against BRCA1 mRNA levels on y-axis, analyzed from the tissue samples (VM, n = 10; FT, n = 10; P value is calculated between the average values of expression in FT and VM). BRCA1 - β-hCG correlation graph plotting secretary β-hCG levels (ng/ml) on x-axis against BRCA1 mRNA levels on y-axis, analyzed from the cell lines at (F) 16 h time point. (G) Twenty-four hours time point. (H) Forty-eight hours time point. Expression is normalized to HTR-8/SVneo at respective time points. [HCC represents triple negative breast cancer cell line HCC1937; HCCR represents HCC1937/wtBRCA1 i.e. HCC1937 reconstituted with full length BRCA1; HTR represents HTR-8/SVneo; FT represents first-trimester age-matched normal placenta; VM represents Vesicular Mole/GTD tissue samples; statistical significance was defined as (*) P ≤ 0.05, (**) P ≤ 0.005, (***) P ≤ 0.001 and NS represents non-significant].

BRCA1 expression is inversely correlated to serum β-hCG levels in GTD

Further, β-hCG expression either in serum (for human tissue samples) or in cell line supernatant (for cell lines) was analyzed by ELISA. Cell line supernatant was analyzed at different time points of 16, 24 and 48 h and correlated with BRCA1 mRNA levels in the respective samples. β-hCG levels (in mIU/ml) exhibits an inverse correlation with BRCA1 mRNA levels in the respective tissues and cell lines (Figure 1F–I).

BRCA1 reduction and parallel β-hCG over-expression was analyzed by siRNA silencing of BRCA1 and β-hCG, separately, in cell lines. Silencing was confirmed by qRT-PCR. β-hCG mRNA expression increases about 6- to 10-fold in normal placentae and about 4-fold in GTD cell lines, when BRCA1 is silenced (Figure 2A–C). However, there is not any significant change in BRCA1, in either cell lines, when β-hCG is silenced (Figure 2D–F). This indicates possibility of some kind of regulatory role of BRCA1 over β-hCG, but not vice versa.

Figure 2.

BRCA1 and β-hCG siRNA silencing studies and BRCA1 localization. Expression of BRCA1 and β-hCG in BRCA1 silenced (A) HTR-8/SVneo, (B) JEG-3 and (C) JAR cells by qRT-PCR. Expression is normalized to scrambled siRNA transfected control of the respective cell lines. Two different siRNAs, siBRCA1 1 and siBRCA1 2 were used. Expression of β-hCG and BRCA1 in β-hCG silenced (D) HTR-8/SVneo, (E) JEG-3 and (F) JAR cells by qRT-PCR. Expression is normalized to scrambled siRNA transfected control of the respective cell lines. (G) Immunofluorescence analysis of BRCA1 in cell lines. (H) Immunohistochemical analysis of BRCA1 and β-hCG in GTD and normal placental tissue samples (GTD, n = 8 and normal placenta, n = 8). The larger images are 10× magnification, while the images in inset are of 40× magnification. [Statistical significance was defined as (*) P ≤ 0.05, (**) P ≤ 0.005, (***) P ≤ 0.001 and NS represents non-significant].

BRCA1 is localized to the cytoplasmic membrane in GTD

As reported earlier, in the present study also, IFA of cell lines and IHC of placental tissue samples, reveals that β-hCG expression is localized mainly to the cytoplasmic membrane, cytoplasm and nuclear membrane and scarcely to the nucleus (Supplementary Figure S1G, available at Carcinogenesis Online). Cellular localization of BRCA1 is mainly to cytoplasmic membrane and scarcely to cytoplasm of GTD cells and tissues, while in normal placental counterparts, it is localized to the nucleus and cytoplasm (Figure 2G–H). Additionally, BRCA1 protein expression is comparatively lower in both GTD cell lines and tissues. Thus, the shift in BRCA1 localization in GTD might be contributing to its tumorigenicity.

BRCA1 interacts with β-hCG at the protein level in GTD but not in normal placentae

As analyzed by in situ PLA, there is an interaction between BRCA1 and β-hCG, localized at the cytoplasm, in GTD cells and tissues, systems where BRCA1 is down-regulated (Figure 3A and B). Further, in GTD tissues, the interaction signals are localized to syncytiotrophoblasts and not cytotrophoblasts. However, in normal placental counterparts, there is not any direct interaction between these two proteins (Figure 3B). The exact reason behind this interaction, in a system where BRCA1 is down-regulated, has yet to be elucidated.

Figure 3.

Interaction of BRCA1 and β-hCG and mutation analysis for BRCA1 and BRCA2. In situ PLA analysis for BRCA1 and β-hCG interaction in (A) cell lines and (B) tissue samples (GTD, n = 8, normal placenta, n = 8). Red spots, but not red background, are characteristic of BRCA1 and β-hCG interaction in PLA analysis. (C) Protocol and outcome of mutation analysis for BRCA1 and BRCA2 in cell lines.

BRCA1/2 mutation analysis in GTD and placental cell lines

BRCA1/2 mutation analysis by the Illumina MiSeq/next-generation sequencing technologies revealed that there is lack of any mutation in either BRCA1/2 genes, in all the three cell lines analyzed. Hence, the possibility of BRCA1 down-regulation in GTD, owing to mutation, was ruled out (Figure 3C).

Epigenetic silencing counts for BRCA1 down-regulation in GTD

Methylation PCR was carried out to analyze if BRCA1 down-regulation in GTD was due to hypermethylation of BRCA1 promoter. Hypermethylation of BRCA1 promoter is observed in GTD cells and tissues, but not in normal placentae (Figure 4A and B). mRNA expression of the key methylation enzymes, DNMT1/3a/3b/3L, was analyzed in the cell lines, which reveals that, DNMT3b, but not DNMT1/3a/3L is up-regulated in GTD than normal placentae (Figure 4C–F). DNMT3b protein over-expression is seen in GTD cells and tissues than normal placental counterparts; but, without any changes in DNMT3a levels (Figure 4G–I). Hence, elevated DNMT3b could be the reason for BRCA1 promoter hypermethylation and its resultant down-regulation in GTD.

Figure 4.

Analysis for BRCA1 promoter methylation status and DNMT levels. Methylation PCR for BRCA1 promoter in (A) cell lines and (B) tissue samples (VM, n = 9 and FT, n = 9). mRNA expression of (C) DNMT1, (D) DNMT3a, (E) DNMT3b and (F) DNMT3L by qRT-PCR in cell lines. Expression is normalized to HTR-8/SVneo. (G) Proteins level expression of DNMT3a and DNMT3b in cell lines and tissue samples by western blot analysis. Quantification of DNMT3a and DNMT3b protein expression by western blot analysis in (H) cell lines and tissue samples (VM, n = 6 and FT, n = 6). [H and HTR represents HTR-8/SVneo; G represents JEG-3; R represents JAR; U represents amplification with Unmethylated primers; M represents amplification with Methylated primers; FT represents first-trimester age-matched normal placenta; VM represents Vesicular Mole/GTD tissue samples; statistical significance was defined as (*) P ≤ 0.05, (**) P ≤ 0.005, (***) P ≤ 0.001 and NS represents non-significant].

GTD is associated with over-expression of Ezrin-Radixin-Moesin (ERM) proteins

ERM proteins can localize BRCA1 to cytoplasmic membrane; this led us to analyze their status in GTD. ERM expression is higher in GTD cells (JEG-3 and JAR) and tissues when compared with the normal placental counterparts (HTR-8/SVneo) at both protein and mRNA levels (Figure 5A–E). Phospho-Ezrin levels are higher in GTD cells and tissues, in comparison with normal placentae, indicating an augmented Ezrin phosphorylation and activation in GTD (Figure 5F–H). IFA reveals ERM localization to cytoplasm in GTD cells and to cell membrane and faintly to cytoplasm in normal placentae (Figure 5I). IHC reveals a shift in localization of ERM proteins from apical cell surface, as seen in normal placentae, to cytoplasm and plasma membrane in GTD tissues (Figure 5J). Hence, over-expression of ERM in GTD, along with its shift in localization, could possibly correlate with BRCA1 localization at the cytoplasmic membrane and resultant increase in tumor progression.

Figure 5.

Expression and localization of ERM. (A) ERM protein expression in GTD and normal placental cell lines and tissue samples by western blot analysis. Quantification of ERM protein expression by western blot analysis in (B) cell lines and (C) tissue samples (VM, n = 6 and FT, n = 6). mRNA expression of ERM genes by qRT-PCR analysis in (D) cell lines (Expression is normalized to HTR-8/SVneo) and (E) tissue samples (VM, n = 10; FT, n = 10; P value is calculated between the average values of expression in FT and VM). (F) Phospho-Ezrin protein expression in cell lines and tissue samples by western blot analysis. Fold activation of Phospho-Ezrin over Ezrin protein expression in (G) tissue samples (VM, n = 6 and FT, n = 6) and (H) cell lines (Expression is normalized to HTR-8/SVneo). (I) Immunofluorescence analysis of ERM proteins in cell lines. J) Immunohistochemical analysis of ERM proteins in GTD and normal placental tissue samples (GTD, n = 8 and normal placenta, n = 8). The images are 10× magnification. [HTR represents HTR-8/SVneo; FT represents first-trimester age-matched normal placenta; VM represents Vesicular Mole/GTD tissue samples; statistical significance was defined as (*) P ≤ 0.05, (**) P ≤ 0.005, (***) P ≤ 0.001 and NS represents non-significant].

GTD exhibits greater sensitivity to PB than MTX

PB was compared against MTX for its cytotoxic effects by cell proliferation MTT assay in GTD. The IC₅₀ of MTX and PB in JEG-3 is 83 and 13 µM, in JAR is 94 and 11 µM while in HTR-8/SVneo, it is 160 and 32 µM, respectively (Figure 6A and B). PB causes cytotoxicity in GTD cells, JEG-3 and JAR, at a lower concentration, which is nearly 1/8th the IC₅₀ of MTX in these cells. Also, in normal placental cells, HTR-8/SVneo, PB has an IC₅₀, which is nearly 1/5th that of MTX. Thus, PB is effectively cytotoxic in GTD at about 2.8-fold lower concentration than normal placentae while MTX is effective at 1.7-fold lower concentration in GTD in comparison to normal placentae. The concentration of MTX and PB used for further experiments was half the IC₅₀ for GTD cells, which were 45 µM of MTX and 6µM of PB.

Figure 6.

Cell proliferation MTT assay and the effects of MTX and PB in cell lines. Cell proliferation MTT assay to analyze the effect of (A) MTX, (B) PB and (C) MTX-PB combinations in the cell lines. Chemiluminiscence assay to analyze the secretary β-hCG levels post-drug treatment with MTX and PB individually and in combination in (D) HTR-8/SVneo, (E) JEG-3 and (F) JAR cells at three different time points. Chemiluminescence assay to analyze the secretary estradiol levels post-drug treatment with MTX and PB individually and in combination in (G) HTR-8/SVneo, (H) JEG-3 and (I) JAR cells at three different time points. Chemiluminescence assay to analyze the secretary Progesterone levels post-drug treatment with MTX and PB individually and in combination in (J) HTR-8/SVneo, (K) JEG-3 and (L) JAR cells at three different time points. qRT-PCR to analyze the BRCA1 mRNA levels post-48 h of drug treatment with MTX and PB individually and in combination in (M) HTR-8/SVneo, (N) JEG-3 and (O) JAR cells. [MTX and M represents Methotrexate; PB and P represents Plumbagin; 45, 30, 6 and 4 are the concentrations of respective drugs in µM; statistical significance was defined as (*) P ≤ 0.05, (**) P ≤ 0.005, (***) P ≤ 0.001 and NS represents non-significant].

Further, the combinatorial cytotoxicity of PB and MTX was studied at various combinations with IC_50/2 and IC_50/3 of each MTX and PB for GTD cells. IC_50/2 combination of each PB (6 µM) and MTX (45 µM) is the most cytotoxic one, showing a cell death of only 5% in normal placentae, while about 45% in the GTD cells and this combination was selected for further combinatorial studies (Figure 6C).

PB down-regulates β-hCG, estrogen and progesterone levels

The activity of PB, in comparison with MTX, was studied by analyzing the levels of β-hCG, estrogen and progesterone by chemiluminiscence immuno assay, in the cell line supernatant post-48 h treatment. Post-MTX treatment, β-hCG, estrogen and progesterone levels are reduced not only in GTD (JEG-3 and JAR), but also in normal placental cells (HTR-8/SV neo), which indicates that it is toxic to the latter also. However, post-PB alone or PB-MTX combination treatment, β-hCG, estrogen and progesterone levels are reduced in GTD cells more efficiently, while comparatively retaining or minimally reducing them in normal placentae, than post-MTX treatment (Figure 6D–L). Therefore, PB alone or PB-MTX combination is more effective than MTX alone, remaining less toxic to normal placental cells.

PB up-regulates BRCA1 expression

Since BRCA1 is down-regulated in GTD in comparison with the normal placentae, BRCA1 mRNA was analyzed by qRT-PCR, post-drug treatment with MTX alone, PB alone and PB-MTX combination, to check if the drugs altered them. BRCA1 mRNA levels post-MTX alone treatment is reduced in GTD as well as normal placental cells. However, post-PB treatment (PB alone and PB-MTX combination), BRCA1 expression is increased in GTD, without significant changes in the normal placentae (Figure 6M–O). Thus, on analysis of the levels of β-hCG, estrogen, progesterone and BRCA1 mRNA, it was observed that PB could be a better therapeutic than MTX for GTD.

PB alters the epigenetic regulation of BRCA1 in GTD

Since BRCA1 mRNA levels were altered, BRCA1 promoter methylation status was also analyzed post-drug treatment. Post-MTX treatment, hypermethylation of BRCA1 promoter appeared in normal placentae, while there was not any significant change in GTD cells. On the contrary, without changing the methylation pattern of BRCA1 promoter in normal placental cells, PB treatment reversed the hypermethylated status of the BRCA1 promoter in the GTD cells. Thus, MTX either retained the methylation status or induced hypermethylation in the cells, while PB possessed a selective hypomethylation activity in GTD cells, which resulted in altered BRCA1 expression post-treatment. However, post-PB-MTX combinatorial treatment, there was not much alteration of BRCA1 promoter methylation in normal placentae, but this could not reverse the BRCA1 promoter methylation status in GTD cells effectively in comparison with PB alone (Figure 7A–D).

Figure 7.

Analysis of BRCA1 promoter methylation levels post-drug treatment. Methylation PCR for BRCA1 promoter post-drug treatment in the cell lines with MTX and PB (A) individually and (B) in combination. Quantification of degree of methylation in methylation PCR analysis in the cell lines post-drug treatment with MTX and PB (C) individually and (D) in combination. (E) DNMT3b protein levels post-drug treatment with MTX and PB individually and in combination in cell lines by western blot analysis. Quantification of DNMT3b protein levels by western blot analysis post-drug treatment in (F) HTR-8/SVneo, (G) JEG-3 and (H) JAR cells. [HTR represents HTR-8/SVneo; JEG represents JEG-3; MTX represents Methotrexate; PB represents Plumbagin; U represents PCR products with Unmethylated primers; M represents PCR products with methylated primers; Combi represents MTX and PB in combination; statistical significance was defined as (*) P ≤ 0.05, (**) P ≤ 0.005, (***) P ≤ 0.001 and NS represents non-significant].

DNMT3b protein level was also analyzed post-drug treatment. Post-MTX treatment, there is an increase in DNMT3b in both normal placental and GTD cells, while post-PB treatment (PB alone and PB-MTX combination), DNMT3b is significantly reduced in GTD cells and in a meagre amount in the normal placentae. All these data corroborate that PB could be a more potent mono-therapeutic drug than MTX for GTD treatment (Figure 7E–H).

Cohort analysis for the follow-up of GTD patients for breast cancer incidence

Epidemiological details were analyzed for 89 GTD patients whose tissue samples were analyzed in the study. It was observed that 85% of the GTD patients were below 25 years of age, 11% of the patients were in the age group of 25–30 years and 4% of the patients were above 30 years of age (Supplementary Table S3A, available at Carcinogenesis Online).

GTD incidence rate was found to be 5.03 per 1000 deliveries at the SAT Hospital, Government Medical College, Thiruvananthapuram, Kerala, India, between January 2011 and February 2016. 9.3% of GTD cases progressed to GTN from January 2012 to February 2016. All the GTN patients who came across in the study were treated with MTX, as the first-line single chemotherapeutic agent (Supplementary Table S4, available at Carcinogenesis Online).

Discussion

β-hCG plays a pivotal role in tumorigenesis as its expression has been reported to be elevated not only in the tumors of trophoblasts and germ cells, but also the non-trophoblastic tumors, which includes cancers of the kidney, bladder, head and neck, ovary, prostate, lungs, breast, endometrium and cervix (9). We had already reported that the wild type, but not mutant BRCA1 transcriptionally inhibits β-hCG and thus, β-hCG levels are elevated in breast cancer cells having mutated BRCA1. Most of the BRCA1-mutated breast cancers are TNBCs. However, we have proved that β-hCG expression is regulated by BRCA1 status and not linked to TNBCs, as evidenced by the results obtained with BRCA1 wild-type TNBC cell line, MDA-MB 231 versus BRCA1 mutant TNBC cell lines, HCC 1937 and SUM 149 (6). It has been reported by Li et al., 2009 that wild-type BRCA1 and Smad3 interaction might contribute to the inhibition of oxidative stress within the cancer cells and mutation of BRCA1 impairs TGFβ signaling (10). However, we had shown that β-hCG secreted by the BRCA1 mutant breast cancer cells interacts with TGFβRII, thereby, triggering the Smad3-induced cell signaling, resulting in cell proliferation. Furthermore, the TGFβRII inhibition in BRCA1 mutant breast cancer cells was found to hamper the cell proliferation. This could be the reason for selective tumorigenesis of β-hCG in BRCA1 mutant breast cancer cells and not TNBCs as a whole (6,7,11). This study analyzes molecular etio-pathology of GTD, where β-hCG is highly over-expressed, focusing on BRCA1 status and BRCA1 - β-hCG relationship in placentae and placental tumors. This is the first report unveiling BRCA1 status in GTD.

BRCA1 down-regulation or mutations has been known to be associated with hereditary breast ovarian cancers (12). The present study shows BRCA1 to be down-regulated in GTD when compared with the normal placentae, which might contribute to an inefficient DNA repair mechanism, leading to augmented tumorigenicity in placentae. There is an inverse relationship between β-hCG and BRCA1 expression in BRCA1 wild-type breast cancer cells (6,7). Even though β-hCG level increases when BRCA1 is silenced, BRCA1 expression did not change when β-hCG is silenced in either of the cells. Hence, BRCA1 might exert a direct or indirect regulatory role over β-hCG, but not vice versa. This regulation might be either transcriptional as reported earlier in breast cancer cells or by ubiquitination action of BRCA1 or by indirect mechanisms through other proteins, which has to be elucidated further (6). However, reports from our laboratory have shown that BRCA1 expression increases while β-hCG is silenced in BRCA1-mutated TNBC cells. This proves regulation of β-hCG over BRCA1 in TNBCs, but not in trophoblasts. Further, it has to be reinstated that wild-type BRCA1 but not mutant BRCA1 exerts transcriptional inhibition on β-hCG, by binding to its promoter in TNBCs (6,7).

BRCA1 protein mostly localizes to nucleus (13), but, in GTD, it is localized to cytoplasmic membrane and scarcely to cytoplasm. This mis-localization might contribute to dysregulated cellular BRCA1 functions, contributing to pathogenesis in GTD. A direct interaction between BRCA1 and β-hCG is found localizing to syncytiotrophoblasts and not cytotrophoblasts, in GTD, but not normal placentae. Numerous reasons might exist for this interaction including ubiquitination activity of BRCA1 on β-hCG, but has to be elucidated further. However, results are intriguing as interaction occurs in GTD, where BRCA1 is down-regulated, but not in normal placenta where BRCA1 levels are comparatively higher.

The next step was in identification of the possible reason for down-regulation of BRCA1 in GTD. Mutation analysis for BRCA1/2 shows absence of any mutations; thus, possibility of BRCA1 mutations leading to its down-regulation was ruled out, which strengthened the possibility for BRCA1 promoter hypermethylation in GTD. About 30–40% of sporadic breast tumors have BRCA1 down-regulation due to its promoter hypermethylation (14). We have identified that BRCA1 promoter is hypermethylated in GTD, leading to its down-regulation. Even though DNMT1/3a/3b/3L are known to be key regulators affecting the de novo methylation in cells, DNMT3b has been reported to regulate BRCA1 promoter hypermethylation in breast and ovarian cancers (15,16). We have seen that DNMT3b is up-regulated in GTD. Thus, the up-regulation of DNMT3b might contribute to promoter hypermethylation and subsequent down-regulation of BRCA1 in GTD.

ERM proteins are over-expressed in cancers which regulate cellular adhesion and motility (17,18). BRCA1 interacts with ERM proteins, localizing BRCA1 to cytoplasmic membrane (19). ERM proteins and Phospho-Ezrin, the activated form of Ezrin, are up-regulated in GTD than in normal placentae. Also, localization shift of ERM proteins from apical cell surface to cytoplasmic membrane and cytoplasm in GTD are indicative of activated EMT process and poor prognosis, as reported in breast cancers (17–20). Over-expression of ERM, its activation and shift in localization, together shed light towards their possible role, in the absence of BRCA1 at the cytoplasmic membrane, in inducing cell motility, adhesion, migration and invasion; thus, contributing to tumor progression in GTD.

The increasing GTN incidence and GTD relapse, which along with the side effects of MTX, including fertility issues and its longer administration duration for therapy, calls for a better therapeutic. PB, a plant-derived naphthoquinone isolated from the roots of Plumbago zeylanica, has already been reported to be cytotoxic at considerably lower concentrations in BRCA1 defective cancer cells in comparison to other chemotherapeutic agents like doxorubicin or platinum-based drugs like cisplatin and carboplatin, and remains to be relatively less cytotoxic to the normal cells, as reported both in vitro and in vivo. We had shown that PB has targeted cytotoxic activity against BRCA1 defective cancers than the wild-type cancers (21–27). In BRCA1-deficient ovarian cancer cells, the cytotoxicity was in the order of PB > doxorubicin > cisplatin, as reported earlier (21,27). Only at least about 5-fold higher concentration, at 10 mg/kg b.w., PB has been reported of eliciting fertility toxicity (28). IC₅₀ values of MTX are very high for GTD cells and almost 1.7-fold higher in the normal placental cells. However, IC₅₀ of PB was lower for GTD cells and almost 3-fold higher in the normal placental cells. Thus, PB exhibits more cytotoxicity on GTD cells at lower concentrations than MTX. Further, combinatorial studies identified that IC_50/2 combinations of PB-MTX (6 µM of PB and 45 µM of MTX) was exhibiting cell death of about 45% in GTD cells, while only about 5% in the normal placentae.

An ideal drug for GTD should lower β-hCG in GTD cells, without affecting normal placental cells to a greater extent. Estrogen and progesterone are other two significant elevated steroid hormones which help in successful implantation and sustaining the pregnancy until parturition (29). Their levels are also higher during GTD, leading to accelerated cell proliferation, invasion and migration (30). Treatment regimes for GTD are supposed to down-regulate levels of all the three steroid hormones. MTX reduces β-hCG, estrogen and progesterone levels not only in GTD but also in the normal placental cells. However, PB alone and PB-MTX combination efficiently reduces β-hCG, estrogen and progesterone levels in GTD, without significantly affecting the normal placentae; this proves that PB could be a better drug than MTX alone for GTD therapy, as MTX is comparatively toxic to the normal placentae.

Since BRCA1 was down-regulated in GTD, its level post-drug treatment was analyzed. Upon MTX treatment, BRCA1 mRNA is reduced in both normal placental and GTD cells. However, with PB alone or PB-MTX combination, it did not change much in normal placentae, but either remained same or increased in the GTD cells. The elevated BRCA1 mRNA levels post-PB treatment in these patients is beneficial as it would be protective either by inducing high-fidelity DNA damage repair or apoptosis. We also observed that MTX resulted in hypermethylation of BRCA1 promoter in normal placentae without making changes in GTD cells; however, PB alone could efficiently revert back the hypermethylation in the GTD cells to the normal state, while retaining the methylation levels in the normal placental cells. Further, post-MTX treatment, DNMT3b levels increased in normal and tumor placental cells, but with PB alone or PB-MTX combination, DNMT3b levels decreased in GTD cells, without any significant changes in normal placentae. Thus, in terms of the changes created in BRCA1 promoter methylation status, along with DNMT3b levels, PB alone works to be beneficial than MTX alone or PB-MTX combination, with much lesser toxicity to the normal placental cells.

As MTX is a DHFR antagonist, it has no influence on DNA damage (31). Unlike the platinum-based drugs, acting via DNA adduct formation or doxorubicin, acting mainly as a Topoisomerase II inhibitor, PB has shown to exhibit higher cytotoxicity against BRCA1 defective cancers owing to its high reactive oxygen species (ROS)-inducing potential. Thus, with PB treatment, cells will elicit ROS-mediated DNA damage response (25). This might be reverting hypermethylation status and thereby BRCA1 expression. In normal placental cells, available BRCA1 protein would be sufficient to elicit DNA damage repair. In GTD, since BRCA1 is hypermethylated, the ROS-mediated DNA damage induced by PB would result in triggering apoptosis, though cells will try to re-express BRCA1 by reverting its hypermethylation. Probably this could be the mechanism of action by PB in GTD and normal placentae. Considering all the parameters analyzed post-drug treatment, inclusive of the β-hCG, estradiol, progesterone and BRCA1 mRNA levels along with the methylation status of the BRCA1 promoter and DNMT3b levels, it could be inferred that PB alone is efficient than MTX alone and MTX-PB combination in terms of the effectiveness in GTD treatment (Supplementary Figure S2, available at Carcinogenesis Online). The decreased availability of successful mouse models for GTD restricts the confirmation of drug activity in vivo.

The current study also included analysis of epidemiological parameters of the patients enrolled. A greater risk of GTD incidence has been reported in the age group less than 20 years and above 35 years (1). However, in the current study, 85.2% of GTD patients were in between 20 and 30 years of age. The lack of proper registry or database for GTD in India has been a hindrance to the access its global incidence rate. The GTD incidence rate is 5.03 per 1000 deliveries (data collected from SAT Hospital, Thiruvananthapuram, India). 9.3% of GTD cases have progressed to GTN. This is the first study of its kind which tries to link BRCA1 expression in GTD, which raises the possible risk of BRCA1-related diseases in GTD patients (32).

Conclusion

GTD is associated with BRCA1 down-regulation, owing to the promoter hypermethylation resultant of DNMT 3b up-regulation, when compared with the normal placentae. BRCA1 is localized to the cytoplasmic membrane in GTD, possibly linked to the ERM expression, thereby contributing to the tumorigenesis in GTD. The up-regulated β-hCG could be resultant of BRCA1 down-regulation in GTD. It has also been shown that PB alone is efficient than MTX alone or PB-MTX combination, in terms of the reduction of β-hCG/estradiol and progesterone levels in GTD. Interestingly, PB increases the BRCA1 levels post-treatment in comparison with MTX, which reduces the BRCA1 levels, both owing to the alterations in the DNMT3b levels and BRCA1 promoter methylation. Thus, BRCA1 down-regulation in GTD patients could be a possible risk towards BRCA1-associated diseases in future.

Funding

This work was supported by intramural grant from Rajiv Gandhi Centre for Biotechnology, Kerala State Council for Science Technology and Environment, Government of Kerala (016/SRSHS/2011/CSTE), grant-in-aid from the Board of Research in Nuclear Sciences, Government of India (No. 2009/37/5/BRNS/1620 and No. 37(1)/14/16/2014), Indian Council for Medical Research, Government of India (No. 53/20/2012-BMS), Department of Science and Technology, Government of India (EMR/2017/002222) and Department of Biotechnology, Government of India (BT/IN/Indo-US/Foldscope/39/2015), to P.S.

Glossary

Abbreviations

ELISA

enzyme-linked immunosorbent assay

ERM

Ezrin-Radixin-Moesin

GTD

gestational trophoblastic diseases

GTN

gestational trophoblastic neoplasia

GTT

gestational trophoblastic tumors

IFA

immunofluorescence analysis

IHC

immunohistochemical analysis

MTT

3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyl tetrazolium bromide

MTX

methotrexate

PB

plumbagin

PLA

proximity ligation assay

ROS

reactive oxygen species

SAT

Sree Avittom Thirunal

TNBC

triple negative breast cancers

VM

vesicular moles

Author contributions

R.N. carried out the experiments, analyzed the data and prepared the manuscript; S.K.S., S.K.H., A.R. and N.R.L. helped in analyzing the data; S.Y. and K.R.S. helped in immunohistochemical analysis; A.B.V. and S.A. helped in cohort study; S.P.S. and J.D.S.U. helped in tissue sample collection and processing; G.R.V. and R.T. helped in reviewing the manuscript; S.S. and A.T.V. helped in analyzing the GTD and normal placental tissue samples; J.V.V., N.C., S.S. and A.P.V. were the collaborators at the medical care centers for enabling the GTD and normal placental tissue collection and P.S. conceived and designed the study and approved the manuscript.

Conflict of Interest Statement: None declared.