Family history of menstrual irregularity or diabetes mellitus enhances the susceptibility to polycystic ovary syndrome among subjects harboring rs7903146 genetic variant of TCF7L2

Abstract

Purpose

TCF7L2 mediated Wnt signaling cascade regulates glucose homeostasis by orchestrating expression, processing, and hepatic clearance of insulin. Type 2 diabetes mellitus (T2DM) and polycystic ovary syndrome (PCOS) significantly overlap in pathophysiological features with insulin resistance as a central driver. While TCF7L2 is the most potent T2DM locus, studies on the association of TCF7L2 with PCOS are limited and inconclusive. Therefore, in addition to expression profiling, the association of TCF7L2 polymorphic variant rs7903146 with PCOS was evaluated.

Methods

Using Rotterdam-2003 criteria for the diagnosis, 120 PCOS cases, and 120 age-matched controls were recruited. Subjects underwent clinical, biochemical, and hormonal assessment, followed by genotyping for rs7903146, carried out by PCR–RFLP and TCFL2 expression profiling by qRT-PCR. Genotype–phenotype correlation analysis was performed to evaluate any such associations. Odds ratios (ORs) with 95% confidence intervals (95% CIs) were computed by conditional logistic regression.

Results

Higher odds of developing PCOS were observed in the women having a family history (FH) of either T2DM (OR = 3.86, 95% CI:1.90 – 7.83), hirsutism (OR = 4.74. 95%CI: 1.91 – 17.21) or menstrual irregularities (MI) (OR = 3.07, 95%CI: 1.61 – 8.54). The genotypes of rs7903146 did not confer any risk for developing PCOS (OR = 0.46;95%CI: 0.15 – 2.03). However, the elevated risk was seen in the subjects who harbored the variant allele and had FH of either T2DM (OR = 6.71; 95%CI: 1.89 – 23.78) or MI (OR = 9.71; 95% CI:1.89 – 23.78).

Conclusion

TCF7L2 polymorphic variant rs7903146 is not independently linked to PCOS risk, but modulates the risk in the subjects having a family history of either T2DM or MI.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40200-022-01050-y.

Introduction

Among various cellular signaling cascades, Wnt signaling transduction system represents a multitude of networks triggered by the interaction of Wnt ligand and its receptor. This signaling cascade orchestrates cellular functioning in several ways including polarity, proliferation, movement, differentiation, survival, and self-renewal [1]. Despite the complexity in functions, Wnt signals act through the interaction of only two mediators, β-catenin and transcription factor 7-like 2 (TCF7L2), to regulate the expression of multiple genes, post formation of β-cat/TCF complex in the nucleus [2]. The downstream function of this signaling cascade includes synthesis, secretion, and degradation of insulin, glucagon-like peptide-1 (GLP-1), glucose-dependent insulin-tropic peptide (GIP), growth regeneration, regulation of insulin-like growth factor-1 (IGF-1), and functionality of pancreatic beta islets [3]. Mediated via several target genes and downstream regulatory networks, TCF7L2 plays a pivotal role in orchestrating the regulation of the expression and processing of mature insulin. Additionally, TCF7L2 also influences peripheral insulin sensitivity and its hepatic clearance [4], making TCF7L2 an important modulator of glucose homeostasis.

In women, Type 2 diabetes mellitus (T2DM) and insulin resistance (IR) are two hallmarks of a common female disorder: polycystic ovary syndrome (PCOS) that shares genetic features with T2DM. PCOS is emerging as the most pervasive endocrine–metabolic disorder in young women [5], clinically represented by chronic anovulation, hyperandrogenism, and polycystic ovaries [6, 7]. In 2017, for PCOS, the age-standardized incidence rate has now elevated to ~ 83/10⁵ women of reproductive age, globally [8]. Similar to the global scenario, the prevalence of PCOS in India is also on the rise and is reported to range between 3.7 to 26% [9]. However, in Kashmir, an ethnically distinct, northern Indian part, its reported prevalence is as high as ~ 29% and ~ 34% by NIH and AE-PCOS criteria respectively [10]. Given the rapidly increasing global prevalence of PCOS, its crosstalk with IR and T2DM, and shared genetic susceptibility factors including TCF7L2 indicates the latter as an important candidate in the yet poorly understood aetiology of PCOS. Moreover, these scientific evidences suggest that IR, T2DM and PCOS must be an outcome of the dysregulation of common pathway/s. The TCF7L2 regulates GLP-1, which in turn stimulates secretion and biosynthesis of insulin and inhibits glucagon release, highlighting the biological candidature of TCF7L2 as a PCOS susceptibility locus [11]. Moreover, in vitro depletion of TCF7L2 in human pancreatic β-cells has been shown to enhance apoptosis and decrease β-cell proliferation in addition to diminishing glucose-stimulated insulin secretion [12]. In subjects with impaired glucose tolerance, TCF7L2 confers the elevated risk and hastens their transition of developing T2DM [13]. Given that PCOS is associated with IR and poses a seven-fold risk for developing T2DM, the pivotal role of TCF7L2 in glucose homeostasis makes later a strong candidate for PCOS pathogenesis [13].

The clinical presentation, complex, and poorly understood pathogenesis of PCOS, is determined by genetic and epigenetic factors [14, 15]. Limited studies from Kashmir have evaluated the genetic determinants of PCOS presenting varied results [16–18]. The TCF7L2 gene is an essential marker for T2DM, and its genetic variations have been persistently associated with T2DM [19, 20]. However, its association studies on PCOS have shown mixed results [21, 22]. Therefore, we carried out a case–control study to determine the comparative TCF7L2 expression profile, the association of its polymorphic variant (rs7903146) with risk of PCOS, and its effect modification on clinical and metabolic variables of PCOS.

Material and methods

Subject recruitment

The study enrolled 240 unrelated subjects, comprising 120 PCOS subjects and 120 ethnically matched healthy controls of Kashmir ethnicity. Institutional review board approval for this study was obtained from Institute Ethical Committee (No. IEC 19/16). Full informed consent was taken and documented in all subjects electing to participate. Complete subject history was recorded by administering a predesigned questionnaire. Subjects were interviewed if their first-degree relatives had either T2DM, irregular menstrual cyclicity, acne and hirsutism. Cases included women diagnosed by Rotterdam criteria from the Endocrinology clinic of a tertiary care hospital in North India, who further underwent a detailed clinical examination followed by subsequent biochemical and hormonal tests during their respective follicular phases. The control group consisted of apparently healthy women, age-matched (± 2 years) for respective cases, recruited from the community and various educational institutes as a part of screening-cum-community outreach programs periodically conducted by the Departments of Clinical Research, and Endocrinology. All the recruited controls underwent similar clinical and laboratory evaluation to that of the cases. The general anthropometrical assessment included measurements of height, weight, and waist circumference. The biochemical parameters assessed included oral glucose tolerance test, indicators of lipid profile, kidney function test, and hormone profiling was done as reported previously [10].

DNA Isolation

Genomic DNA was extracted from the peripheral blood using the standard phenol–chloroform method [23]. The purity and integrity of the genomic DNA were checked on Nanodrop and agarose gel electrophoresis.

Genotyping of TCF7L2 rs7903146

Amplification of DNA for rs7903146 was carried out using already reported primer sequences (sense:5`‑ATTAGAGAGCTAAGCACTTTTAGGTA‑3` and antisense:5`‑AAGCTTCTCAGTCACACAGG‑3` by polymerase chain reaction (PCR) [24], using a thermal cycler (Agilent Master Cycler gradient). In a 25 μl PCR reaction mixture of ~ 100-150 ng of gDNA, 0.25 μM of each primer, Taq DNA polymerase (0.25 U), and deoxynucleotide triphosphate (0.5 mM each), (Fermantas, MBI, Vilnius, Lithuania) were added. The PCR was carried out for 35 cycles (3 min at 95 °C: initial denaturation; 45 s at 95 °C: repeating denaturation, 50 s at 60 °C: annealing step; 1 min at 72 °C: extension step, followed by 10 min at 72 °C as last extension step. Amplification by PCR resulted in a 176 base pairs (bp) product that was subjected to electrophoresis on ethidium bromide containing agarose gel, followed by visualization in a Gel documentation system (Alpha Innotech Corporation, CA, USA.). The PCR product was digested using Rsa1 (1U) for 2 h at 37 °C yielding fragments of 27 and 149 bp, respectively. The wild-type genotype lacked the restriction site and would resolve as a single band of 176 bp.

Expression profiling of TCF7L2

For quantification of TCF7L2 mRNA in PCOS cases vs. controls, total RNA was isolated from fresh blood samples using QIAamp RNA-easy Kit (Qiagen Inc USA) by following the manufacturer's protocol. RNA extraction was done within two hours of procurement of samples. Using the cDNA Synthesis Kit (ThermoFisher, Scientific, Wilmington, USA), 2 μg of total RNA was subjected to cDNA synthesis and subsequent quantification by real-time PCR. The 10 μl qRT-PCR reaction was set in 1 × of SYBR green master mix containing 0.25 μM of each primer in addition to cDNA (1:10). The PCR was set up at 95 °C for 3 min, 94 °C for 20 s, 72 °C for 20 s followed by a final extension of 3 min for 40 cycles using thermocycler (Rotor-Gene Q, Qiagen, USA). We used β-Actin as the normalizing gene for the analysis. The primer sequences used were TCF7L2: Forward: 5’-AAGAGCAAGCGAAATACTAC-3’; Revere: 5’- CTTCTTTCCATAGTTATCCCG-3’;

β-Actin: Forward: 5’-TCACCATGGATGATGATATCGC-3’, Revere: 5’-ATAGGAATCCTTCTGACCCATGC-3’. All the samples were run in duplicates.

Statistical analysis

The categorical variables were presented as numbers and percentages, while as the continuous variables were calculated as mean $\pm$ standard deviation. Unpaired student t-test was used to compare different clinical, biochemical, and hormonal parameters between PCOS and controls. To calculate odds ratios (ORs) and corresponding 95% confidence intervals (CIs) for assessing the association of TCF7L2 polymorphic variant with the PCOS risk, we used conditional logistic regression models. Besides effect modification by the studied genetic variation (gene – clinical phenotype interaction) was also assessed. Results were adjusted for age, levels of testosterone, luteinizing hormone (LH), follicle-stimulating hormone (FSH), family history of T2DM and hirsutism, mean Ferriman-Gallway (FG) score, age of menarche, and body mass index (BMI) to avoid any possible confounding. Statistical analysis was carried out using STATA (STATA Corp., TX, USA, Version 16). Two-sided P-values of the magnitude < 0.05 were considered statistically significant. Comparative 2 ^−ΔΔCT method using β-actin as a reference gene was used to analyze the data to quantify TCF7L2 mRNA.

Results

A total of 240 subjects including an equal number of cases and age-matched controls were considered who underwent a detailed physical examination before their enrolment in the current study. Unlike controls, the women having PCOS reported less number of menstrual cycles in a year (7.7(± 3.23) vs. 11.96(± 0.20); p > 0.001). The BMI, mean FG score, levels of serum total testosterone, and LH were significantly elevated in cases than respective controls (p > 0.001). The PCOS cases had higher serum levels of fasting blood glucose, insulin, and indices of lipid profile than controls (Table 1). Moreover, higher odds (> threefold) of developing PCOS was observed in the subjects who had family histories of either T2DM, hirsutism or menstrual irregularities. (OR = 3.86, 95% CI:1.90 – 7.83; OR = 4.74. 95% CI: 1.91 – 17.21; OR = 3.07, 95%CI: 1.61 – 8.54, respectively; Table 2).

Table 1.

Anthropometric, hormonal, metabolic, and biochemical parameters in PCOS cases and controls recruited in the study

Parameters	PCOS (N = 120) Mean ($\pm SD$)	Control (N = 120) Mean ($\pm SD$)	p-value
Mean age (years)	24.2 ($\pm$4.09)	23.37$(\pm$3.03)	0.077
Age at menarche (years)	12.97 ($\pm 1.31)$	13.49 ($\pm$1.23)	0.002
Menstrual cycles/year	7.7($\pm$3.23)	11.96($\pm$0.20)	0.000
Mean FG score	11.32 ($\pm$4.40)	5.97 ($\pm$0.99)	0.000
BMI (Kg/m2)	25.58 ($\pm 4$.33)	21.99 ($\pm 3$.72)	0.000
Waist circumference (cm)	89.68 ($\pm$10.50)	82.33 ($\pm 8.49)$	0.000
Serum fasting blood glucose (mg/dl)	88.18 ($\pm 10.05)$	83.67 ($\pm 9.23)$	0.000
Serum total T4 (ug/dL)	8.59 ($\pm 2.01)$	8.13 ($\pm 1.70)$	0.094
Serum TSH (ug/dL)	3.51 ($\pm 1.84)$	2.79 ($\pm 1.24)$	0.000
Serum LH (mIU/ml)	8.49 ($\pm 4.83)$	6.09 ($\pm 3.24)$	0.000
Serum FSH (mIU/ml)	6.49 ($\pm 1.79)$	6.67 ($\pm 2.50)$	0.553
Serum total testosterone (ng/mL)	63.53 ($\pm 25.62)$	24.79 ($\pm 9.37)$	0.000
Serum prolactin (ng/dL)	16.73 ($\pm 8.00)$	12.55 ($\pm 6.74)$	0.000
Serum fasting insulin (mIU/ml)	19.16 ($\pm 13.82)$	6.26 ($\pm 3.25$)	0.000
HOMA IR	4.03($\pm 2.92)$	1.31 ($\pm 0.71)$	0.000
QUICKI	0.32 ($\pm 0$.04)	0.38 ($\pm 0$.03)	0.000
FGIR	6.59 ($\pm 5$.18)	17.13 ($\pm 8$.93)	0.000

PCOS Polycystic ovary syndrome; FG Ferrimen-gallway score; BMI Body mass index; TSH Thyroid-stimulating hormone; LH Luteinising hormone; FSH Follicle stimulating hormone; IU International units; HOMA IR Homeostasis model assessment estimated insulin resistance; QUICKI Qualitative insulin sensitivity check index; FGIR Fasting glucose-insulin ratio. A student t-test was sued to compare the means between the groups

Statistically significant p values (<0.05) are writen in bold letters

Table 2.

Association of various family histories with the risk of PCOS

	Cases (%)	Controls (%)	Crude OR (95%CI)	Adjusted OR (95%CI)	p-value
Family history of type 2 diabetes mellitus (T2DM)
Absent	48 (40.68)	89 (74.79)	Referent	Referent	-
Present	70 (59.32)	30 (25.21)	4.61 (2.69 -13.35)	3.86 (1.90 – 7.83)	0.000
Family history of hirsutism
Absent	94 (78.33)	113(94.17)	Referent	Referent	-
Present	26 (21.67)	7 (5.83)	4.79 (1.83- 12.58)	4.74 (1.91 – 17.21)	0.002
Family history of menstrual irregularities
Absent	81 (67.50)	102 (85.00)	Referent	Referent	-
Present	39 (32.50)	18 (15.00)	3.10 (1.52—6.32)	3.07 (1.61—8.54)	0.002

OR odds ratio; CI confidence interval. The results in the adjusted model were adjusted for age, levels of testosterone, LH, FSH, family history of T2DM and hirsutism, FG score, age of menarche, and BMI

Statistically significant p values (<0.05) are writen in bold letters

The distribution of the two alleles (C and T) between cases and controls was nearly equal. The presence of either heterozygous (CT) or mutant (TT) genotypes did not confer any susceptibility for developing PCOS (OR = 0.66, 95%CI: 0.11 – 3.39); p > 0.05). On further evaluation of the risk modification by the variant genotypes in the subjects who had family histories of T2DM, hirsutism, or menstrual irregularities, we found enhanced risk in the subject that harbored the variant genotype of rs7903146 and family history of T2DM (OR = 6.71. 5%CI: 1.89 – 23.78) and menstrual irregularities (OR = 9.71, 95%CI: 1.89 – 23.78; albeit p interaction > 0.05, Tables 3 and 4). However, we did not observe any effect modification in the subjects who harboured the variant genotype and had a family history of hirsutism, possibly because of low numbers in the model. Moreover, a significant over-representation of cases with menstrual irregularities and harboured the variant allele of rs7903146 was also observed. However, odds ratio was not calculated due to very low numbers in the model (Table 3).

Table 3.

Association of polymorphic variant (rs7903146) of TCF7L2 gene with PCOS and its effect modification in subjects having family histories of PCOS phenotypes

Variable	Cases (N = 120)	Controls (N = 120)	Unadjusted OR (95%CI)	Adjusted OR (95%CI)	p-value
C	215 (89.58)	218 (90.83)	referent	referent
T	25 (10.42)	22 (9.17)	1.15 (0.62 – 2.08)	1.14 (0.50 – 2.99)	0.759
rs7903146
CC (Wildtype)	97 (80.83)	100 (83.33)	Referent	Referent
CT (Heterozygous)	21 (17.50)	18 (15.00)	1.18 (0.62 – 2.24)	0.66 (0.11 – 3.99)	0.622
TT (Mutant)	2 (1.67)	2 (1.67)	1 (0.01 – 7.09)	-	1.000
CT + TT (Variant)	23 (19.17)	20 (16.67)	1.15 (0.64 – 2.14)	0.46 (0.15 – 2.03)	0.307
Menstrual Irregularities (MI) of the subjects
No MI + wildtype	39 (32.50)	100 (83.33)	Referent	Referent
No MI + Variant	58 (48.33)	0 (0.00)	-	-	< 0.0001
MI + wildtype	5 (4.17)	20 (16.67)	0.71 (0.23 – 2.25)	0.68 (0.19 – 2.20)	0.4715
MI + Variant	18 (15.00)	0 (0.00)	-	-	< 0.0001
Family history of Diabetes mellitus (FH DM)
No FH of DM + wildtype	42 (35.59)	73 (61.34)	Referent	Referent
FH of DM + wildtype	53 (44.92)	26 (21.85)	3.35 (1.73 – 6.48)	2.81 (1.29—6.11)	0.009
No FH of DM + variant	6 (5.08)	16 (13.45)	0.61 (0.21 – 1.83)	0.48 (0.13—1.73)	0.265
FH DM + variant	12(12.63)	4 (3.36)	6.93(2.14 – 22.45)	6.71 (1.89—23.78)	0.003
Family history (FH) of hirsutism
No FH of hirsutism + Wildtype	76 (63.33)	93 (77.50)	Referent	Referent	-
FH of hirsutism + Wild type	21 (17.50)	7 (5.83)	3.84(1.42 – 10.40)	0.99(0.37–2.69)	0.990
No FH of hirsutism + Variant	18 (15.00)	20 (16.67)	1.05 (0.53—2.07)	3.60 (1.11–11.66)	0.032
FH of hirsutism + Variant	5 (4.17)	0 (0.00)	-	-	0.051
Family history (FH) of menstrual irregularity (MI)
No FH (MI) + Wildtype	65 (54.17)	83 (69.17)	Referent	Referent
Yes FH (MI) + Variant	32 (26.67)	17 (14.17)	2.84 (1.34 – 6.03)	2.81 (1.29 – 6.11)	0.009
No FH (MI) + Wild type	16 (13.33)	19 (15.83)	1.14 (0.56 – 2.32)	0.48 (0.13 – 1.73)	0.265
Yes FH (MI) + Variant	7 (5.83)	1 (0.83)	11.58 (1.32 – 101.60)	9.71 (1.89 – 23.78)	0.003

OR odds ratio; CI confidence interval; N number of individuals; F/H Family history; T2DM type 2 diabetes mellitus; Variant CT + TT. The results were adjusted for age, levels of testosterone, LH, FSH, family history of T2DM and hirsutism, FG score, age of menarche, and BMI

Statistically significant p values (<0.05) are writen in bold letters

Table 4.

Gene–gene and gene-environment interactions

Interaction between	Standard error	OR (95% CI)	p-interaction
rs7903146 x FHDM	2.80	3.35 (0 .65 – 17.22)	0.147
rs7903146 x FH hirsutism	-	-	0.999
rs7903146 x FH MI	4.12	3.56 (0.37 – 34.37)	0.273

CI confidence interval; OR odds ratio; FHDM Family history of T2DM; FHMI Family history of menstrual irregularity

Quantification of the mRNA levels of TCF7L2 among PCOS cases and controls was done using qRT-PCR. No significant differences in the expression profile between cases and controls. (p > 0.05; (Fig. 1) were observed.

Fig. 1.

Expression profile of TCF7L2 in PCOS case vs control. Expression of TCF7L2 gene was checked using qRT-PCR between cases and controls. Student t-test was used to calculate p-value (p > 0.05)

Discussion

In the current study, strong risk of developing PCOS was observed in the women who had family histories of either T2DM, hirsutism, or menstrual irregularities. Although genetic variant rs7903146 was not linked with the development of PCOS, however, an elevated risk in the subjects who harboured the variant allele and had a family history of T2DM or menstrual irregularities was observed. Our data indicates an important role of yet to be unveiled gene–gene interaction of heritable genetic factors in the pathogenesis of PCOS. Moreover, our data indicates that despite etio-pathophysiological and epidemiological overlap between T2DM and PCOS, there are other factors involved in the ovarian response to IR that possibly orchestrate the progression of PCOS.

Unlike controls, the BMI, levels of serum testosterone, mean FG score, and LH in subjects with PCOS were significantly elevated. The observations are common and have been reported in diverse ethnic populations as well [25]. However, some of these variations are influenced by ethnicity as well. As an example, unlike black women, Asian women having PCOS are rarely hirsute. Moreover, in the current study, unlike earlier reports, there was no significant difference in the levels of FSH between cases and controls [17, 26].

Our study showed that the presence of either heterozygous (CT) or mutant (TT) genotype did not confer any susceptibility to develop PCOS. TCF7L2 is a strong candidate gene in T2DM, and the T allele of rs7903146 has been hypothesized to influence the glycemic control of the subject [27, 28]. However, the exact underlying molecular mechanisms of rs7903146 mediated impact on glucose homeostasis is not understood yet but it is speculated that rs7903146 polymorphic variant of TCF7L2 is likely to modulate the TCF7L2 gene expression or alter its protein structure, affecting the biological role of TCF7L2 and the subsequent effect on glucose homeostasis[29]. The dysregulated glycemic control is a hallmark of PCOS and studies have persistently associated rs7903146 with T2DM [29]. However, studies on PCOS have failed to demonstrate any conclusive results [30]. A meta-analysis by Wang et al. did not find any association while an earlier meta-analysis carried out by Shen et. al. (2015)[21] reported a marginal association with PCOS. The heterogenic results for linking rs7903146 with PCOS risk suggest that TCF7L2 possibly modulates the PCOS risk via some yet-to-be-discovered gene–gene interactions.

The mode of familial transmission of PCOS remains poorly understood. Although recent attempts have been made to unveil the causative genetic loci for PCOS. Previous studies are suggestive of a robust genetic predisposition for PCOS, which is indicated by familial clustering of PCOS symptoms such as hyperandrogenism, menstrual irregularities, and metabolic syndrome [31–34]. The family history (FH) of T2DM, menstrual irregularities and hirsutism has been linked with the enhanced risk of developing PCOS among offspring [35].

T2DM and PCOS are known to share genetic features. Moreover, strong evidence has accumulated suggesting that positive FH of T2DM likely enhances the T2DM risk in women with PCOS [36–39]. In the present study, it was found that there is a more than threefold risk of developing PCOS in the subjects who had positive family histories of T2DM, menstrual irregularity and hirsutism, suggesting them to be important determinants of the risk and heritability of PCOS disease phenotypes. Our findings are consistent with published studies reporting that women having PCOS and a positive FH of T2DM have an adverse metabolic profile as well [40].

On further stratifying the subjects based on the histories of T2DM, menstrual irregularity, and hirsutism, it was found an elevated risk in the subjects who had FH of T2DM and menstrual irregularity and harboured the variant genotype of rs7903146. Thus, our findings are in line with the earlier reports suggesting the role of genetic factors in metabolic derangements among PCOS women [40]. Our findings suggest that the FH assessment is likely to facilitate stratification of PCOS women based on their risk, which is important, given the high prevalence of PCOS. Moreover, the risk modification in the subjects who had FH of hirsutism was not be assessed possibly because of the low number in the model. Besides, a significant over-representation of cases with menstrual irregularities that harboured the variant allele of rs7903146 was observed. However, the odds ratio was not calculated due to very low numbers in the model. It is established that PCOS poses an elevated risk for metabolic aberrations in first-degree relatives [41].

Studies have demonstrated that the inheritance of risk variant of rs7903146 (T allele) may alter the TCF7L2 expression in the human pancreas. Compared to nondiabetic cadaveric donors, the TCF7L2 was fivefold upregulated in the pancreas of T2DM patients. The TCF7L2 is highly expressed in the islets of nondiabetic TT carriers. Similar reports were elucidated in murine T2DM models as well. A study revealed a 2.3-fold surge in insulin gene expression with simultaneous aberrant insulin secretion from normal human islet cells overexpressing the TCF7L2 gene [42]. Decreased expression of TCF7L2 is also reported in ovarian cumulus cells of lean PCOS women[43]. While an earlier study has reported a ~ two fold upregulation of TCF7L2 in the granulosa cells of non-IR women with PCOS[37], we did not observe any differences in the expression profile of TCF7L2 between PCOS cases and controls. The lack of association of the expression profile can be attributed to the fact that the profiling was done in blood, unlike the earlier study that more aptly looked for such differences in the granulosa cells. Pertinent to mention, while it would be interesting to replicate such findings, but the availability of such tissues is also a challenge to researchers.

Adjustment of the results for potential confounders, diagnosis of PCOS cases by limited staff, and estimation of correlation of TCF7L2 variant with multiple risk determinants are among the major strengths of the study. However, smaller sample size is likely to be a reason for not detecting any association in the gene-environment interaction analysis.

Conclusion

A family history of T2DM, hirsutism, or menstrual irregularities is a strong risk factor for women to develop PCOS. The polymorphic variant rs7903146 of TCF7L2 is not associated with the PCOS risk independently but likely modulates the risk in subjects with a family history of T2DM, hirsutism or menstrual irregularities. More studies employing a mechanistic approach are warranted to substantiate our findings on the effect modification.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

MAG conceived the study; MAG, BAG and MG designed the study. RR collected the data and performed experiments. IAS and MMA analyzed and interpreted the data. RR, IAS, BAG and MAG wrote the first draft of the manuscript. All the authors reviewed and approved the final draft of the manuscript.

Declarations

Conflicting interests

All the authors declare that they do not have any competing financial and/or non-financial or conflicting interests concerning the work described.