Association of TGFA gene polymorphisms (rs11466297 and rs3732248) with polycystic ovary syndrome (PCOS): a case-control study

Abstract

Objectives

Polycystic ovary syndrome (PCOS) is a complicated endocrine condition that causes ovarian dysfunction, metabolic problems, and hormonal irregularities. Genetic factors profoundly affect its pathogenesis. This study investigates the association between polymorphisms in the Transforming Growth Factor Alpha (TGFA) gene, specifically rs11466297 and rs3732248, and the risk of PCOS.

Methods

We conducted a case-control study involving 200 individuals confirmed to have PCOS and 200 control subjects. We used polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) and Amplification Refractory Mutation System (ARMS-PCR) procedures to genotype the TGFA SNPs rs11466297 A/C and rs3732248 G/A. We performed statistical analyses, including odds ratios (OR) and 95% confidence intervals (CI), to evaluate the relationship between the polymorphisms and the risk of PCOS.

Results

Genotypic analysis revealed that the TGFA rs3732248 GA heterozygote was significantly associated with increased PCOS risk under the codominant model (OR = 2.16, 95% CI: 1.39–3.32, p < 0.001), dominant model (OR = 1.77, 95% CI: 1.18–2.65, p = 0.013), and overdominant model (OR = 2.26, 95% CI: 1.47–3.45, p = 0.001). All associations remained significant after Bonferroni correction (p < 0.025). For rs11466297, no significant association survived correction. PCOS patients had significantly higher BMI, waist circumference, total cholesterol, and triglycerides compared to controls (p < 0.001 for all). Haplotype analysis revealed no significant association between TGFA haplotypes and PCOS risk.

Conclusion

This study is the first to demonstrate that the TGFA rs3732248 polymorphism is significantly associated with PCOS susceptibility in an Iranian population. These findings provide novel genetic evidence supporting a role for TGFA in PCOS pathogenesis and warrant replication in larger independent cohorts.

Introduction

The combination of hyperandrogenism, ovulatory dysfunction, and polycystic ovarian morphology characterizes polycystic ovary syndrome (PCOS), one of the most prevalent endocrine disorders involving women of reproductive age [1–3]. PCOS is estimated to affect 5 to 20% of women, based upon how it is defined [4]. In Iran, the prevalence of PCOS has also been reported to be significant. Studies indicate that the prevalence of PCOS in the Iranian population ranges from 6% to 26% depending on the region and the criteria used for diagnosis [5]. The etiology of PCOS remains complex and multifactorial, involving a combination of genetic, environmental, and metabolic factors [6]. Recent studies have indicated that genetic predisposition plays a significant role in the pathogenesis of PCOS, highlighting the importance of identifying specific genetic variants associated with this condition [7–9]. While its etiology is multifactorial, genetics plays a significant role in its development and variation in phenotypes [10]. Among the genes implicated in the development of PCOS, the TGFA gene (transforming growth factor alpha) represents a compelling candidate. Although TGFA has not emerged as a genome-wide significant locus in large GWAS of European populations, two lines of evidence support its investigation in PCOS: (1) TGFA is expressed in the human ovary and activates the EGFR signaling pathway, which regulates granulosa cell proliferation and follicular development [11] and (2) polymorphisms in TGFA have been associated with metabolic disturbances and hyperandrogenism in preliminary studies. Given the ethnic-specific genetic architecture of PCOS, investigating TGFA variants in understudied populations such as our Iranian cohort may reveal novel associations not detectable in previous GWAS [12]. Preliminary studies have demonstrated a potential link between TGFA gene polymorphisms and clinical manifestations of PCOS-related diseases, including metabolic disturbances and hyperandrogenism. However, data remains limited, necessitating further exploration of these genetic variants in diverse populations. This study is the first to investigate the association between TGFA gene polymorphisms (rs11466297 and rs3732248) and PCOS risk in an Iranian population. To the best of our knowledge, no previous study has examined these two specific variants in the context of PCOS, and our findings provide novel preliminary evidence regarding the role of TGFA in PCOS susceptibility in a non‑European, understudied population.

Materials and methods

Study design

This study employed a case-control design to investigate the association between TGFA gene polymorphisms (rs11466297 and rs3732248) and PCOS. The study was conducted in compliance with ethical standards, and it was approved by the University of Sistan and Balouchestan’ ethics committee. The approval certificate can be found at https://ethics.research.ac.ir/IR.USB.REC.1402.037). This study adhered to the Declaration of Helsinki ethical guidelines (The 75th WMA General Assembly, Helsinki, Finland, October 2024, https://www.wma.net/policies-post/wma-declaration-of-helsinki/). Informed consent was obtained from all participants prior to inclusion in the study.

Participants

The study included 200 women diagnosed with PCOS and 200 healthy control participants matched for age. No specific ethnicity was considered for this study; the only inclusion criterion based on population was the geographical location of Zahedan city (All participants were recruited from the same geographical region (Zahedan, the capital of Sistan and Baluchestan Province, southeastern Iran). Although we did not systematically record the ethnic background of each participant, we ensured that cases and controls were drawn from the same source population by recruiting them from the same clinics and hospitals in Zahedan. To minimize potential population stratification, we verified through medical records and interviews that participants were long-term residents of this region (≥ 10 years) and that no specific ethnic group was overrepresented in either group. Nevertheless, we acknowledge that subtle population stratification cannot be completely ruled out in this single-center study). We based our sample size on previous studies [9, 13] that investigated similar research questions and utilized the same methodology. Sample size and power calculation: The sample size was estimated a priori using the Genetic Power Calculator. Assuming a minor allele frequency of 15% (based on previous studies in Iranian populations), an odds ratio of 1.8 for the high-risk genotype, an α level of 0.05, and a case-control ratio of 1:1, a minimum of 180 cases and 180 controls were required to achieve 80% power. Accordingly, we recruited 200 cases and 200 controls, which provided adequate power to detect moderate to large genetic effects (OR ≥ 1.8). However, we acknowledge that our study may have been underpowered to detect modest associations (OR < 1.5), which are common in complex traits such as PCOS. PCOS diagnosis was established according to the Rotterdam criteria, requiring at least two of the following three features: (1) oligo- or anovulation (defined as menstrual cycles > 35 days or < 8 cycles per year), (2) clinical hyperandrogenism (modified Ferriman–Gallwey score ≥ 8), and (3) polycystic ovarian morphology on transvaginal ultrasound (≥ 12 follicles measuring 2–9 mm in diameter and/or ovarian volume > 10 mL). All ultrasound examinations were performed by a single experienced radiologist to ensure consistency. Biochemical measurements of androgens (e.g., total testosterone, DHEA-S) and gonadotropins (LH/FSH) were not performed due to budget constraints. Therefore, hyperandrogenism was assessed solely on clinical grounds using the Ferriman–Gallwey scoring system. Individuals with thyroid disorders, congenital adrenal hyperplasia, hyperprolactinemia, or other endocrine conditions that could mimic PCOS were excluded based on clinical history and physical examination. Only women aged 18 to 35 years were included. All participants were premenopausal and had not received any hormonal treatments (including oral contraceptives, anti‑androgens, or insulin sensitizers) for at least three months prior to enrollment. Regarding ethnic background, no specific or segregated ethnic group was targeted; the only population‑based inclusion criterion was residence in the geographical region of Zahedan city [2]. Participants were recruited from Molud Zahedan Infertility Treatment Clinic, Ali Ibn Abi Talib Hospital, Zahedan, Iran.

Biochemical assessments

Pars Azmoon (Tehran, Iran) provided a commonly available spectrophotometric kit for the analysis of fasting blood sugar (FBS) and lipid indices, including high-density lipoprotein (HDL-c), low-density lipoprotein (LDL-c), triglycerides (TG), and total cholesterol (TC). We measured serum prolactin levels using an enzyme-linked immunosorbent assay (ELISA) assay kit (PishtazTeb Diagnostics^®, Tehran, Iran), as previously described [14]. While our current analysis focused on biochemical parameters such as lipid profiles, we view the hormonal profile as a critical area for further investigation.

Sample Collection

Peripheral blood samples (1 mL) were collected from each participant via venipuncture into EDTA anticoagulant tubes to prevent coagulation. Samples were immediately processed or stored at -20 °C until further use. Demographic data, including age, BMI, and clinical history, were collected through structured interviews and medical record reviews.

DNA extraction

Genomic DNA was extracted from whole blood using the Salting-Out method [15]. The quality and quantity of DNA (~ 90 ng/µL) were assessed using a Nanodrop Spectrophotometer.

SNP selection

Despite prior studies linking TGFA to diabetes [16, 17], the relationship between TGFA polymorphisms and PCOS remains largely unexplored. We focused on two specific SNPs within the TGFA gene: rs11466297, located in the 3’ untranslated region (3’ UTR), and rs3732248, located in an intron. These loci were selected because of their potential role in gene expression regulation, which is critical for ovarian cell function and follicular development.

Genotyping

Genotyping of TGFA SNPs was performed using polymerase chain reaction (PCR) followed by RFLP and ARMS analysis (Table 1 shows primer sequences used in our study. The primers were designed using Oligo7 software [18] and produced by Pishgam Co. (Tehran, Iran)).

Table 1.

Primer sequences used for genotyping TGFA polymorphisms

SNP	Function	Sequence	Genotyping method	RE	Amplified bands (bp)	Annealing Temp. (ºC)
rs11466297	3 Prime UTR Variant	F: GCCTGGCTTATTTGGGGATT R: AAGGGCAAGGAAACACAGG	PCR-RFLP	BamHI	C: 174 A: 120 + 54	60
rs3732248	Intron Variant	FO: CAGAGCCAATGTCACCAAGT RO: TGCCTTCCTTCTGCTATCACT F(A): GACCGCAACCTCTGACCTAA F(G): GACCGCAACCTCTGACCTAG	ARMS-PCR	-	Control: 166 A/G: 96	57

SNP Single nucleotide polymorphism, RE Restriction enzyme, bp Base pair, F forward, R reverse, PCR-RFLP Polymerase chain reaction-restriction fragment length polymorphism

PCR amplifications were performed in a final volume of 15 µL containing 7.5 µL of SinaClon PCR Master Mix (2X; SinaClon, Tehran, Iran), which contained 1.5 mM MgCl₂, 0.2 mM of each dNTP, and 0.05 U/µL of Taq DNA polymerase. Each reaction was supplemented with 0.4 µM of each forward and reverse primer (final concentration) and approximately 50–100 ng of genomic DNA template, adjusted to the final volume with nuclease-free water. Thermal cycling conditions consisted of an initial denaturation at 95 °C for 5 min, followed by 34 cycles of 94 °C for 30 s, annealing at 60 °C (for rs11466297) or 57 °C (for rs3732248) for 30 s, and extension at 72 °C for 30 s, with a final extension at 72 °C for 5 min. PCR products were resolved by electrophoresis on 2.5% agarose gels stained with Safe Stain (or ethidium bromide) and visualized under UV light (Fig. 1).

Fig. 1.

Gel electrophoretic separation of PCR products for genotyping TGFA rs11466297 (A) and rs3732248 (B) polymorphisms

To ensure genotyping reliability, 15% of randomly selected samples (n = 60; 30 cases and 30 controls) were re-genotyped in duplicate by two independent technicians who were blinded to the case–control status. The concordance rate between the initial and repeat genotyping was 100%, confirming the reproducibility of the results. Each PCR run included negative controls (nuclease-free water without template DNA) to exclude contamination, and positive controls (samples with known genotypes, previously confirmed by consistent banding patterns in preliminary experiments) to confirm the specificity of amplification and restriction enzyme digestion. Samples that repeatedly failed to amplify after three independent attempts, or that showed ambiguous or non-interpretable band patterns unresolved by two independent observers, were excluded from the final analysis. The overall genotyping call rates were 97.25% for rs11466297 (192/200 cases and 197/200 controls) and 97.75% for rs3732248 (193/200 cases and 198/200 controls), indicating high-quality, reliable genotyping data. Adjusting the data based on a significant demographic index (BMI) helped to validate our data.

Statistical analysis

We used SPSS Statistics version 27.0.1.0 (IBM Corp., Armonk, NY, USA) for statistical analysis. The genotype frequency distribution was evaluated for Hardy-Weinberg equilibrium with the Chi-square test. Genotypic and allelic frequency disparities between patients and controls were assessed using Chi-square or Fisher’s exact test. Odds ratios (ORs) and 95% confidence intervals (CIs) were computed by logistic regression analysis and adjusting with BMI. To adjust for multiple comparisons, a Bonferroni correction was performed after data analysis. Since two independent single nucleotide polymorphisms (rs11466297 and rs3732248) were tested, the significance level was adjusted from p < 0.05 to p < 0.025. Only associations with p-values below this corrected threshold were considered statistically significant. Haplotype analysis was performed using Haploview software (version 4.2). The two TGFA polymorphisms (rs11466297 and rs3732248) were analyzed for haplotype reconstruction using the Expectation-Maximization (EM) algorithm. Haplotypes with a frequency of less than 1% were excluded from the analysis. The most common haplotype was considered as the reference group. Linkage disequilibrium (LD) between the two SNPs was assessed using D’ and r² values. Global haplotype association was evaluated using a likelihood ratio test to compare haplotype frequency distributions between PCOS cases and controls.

Results

Demographic and clinical characteristics

The clinical and demographic characteristics of the PCOS cases (n = 200) and healthy controls (n = 200) are summarized in Table 2.

Table 2.

Clinical and demographic characteristics of PCOS patients and controls

Parameter evaluated	PCOS (mean ± SD)	Controls (mean ± SD)	p-value a
Age (year)	27.26 ± 2.06	27.54 ± 3.94	0.365
BMI (kg/m2)	28.74 ± 2.31	25.19 ± 1.42	< 0.001
Underweight (BMI < 18.5)	8	17	-
Ideal (18.5 < BMI < 24.9)	31	114	-
Overweight (25 < BMI < 30)	93	65	-
Obese (BMI > 30)	68	4	-
WC (cm)	101.63 ± 8.12	92.51 ± 7.49	< 0.001
FBS (mg/dl)	100.69 ± 13.23	96.80 ± 5.91	0.184
TC (mg/dl)	195.30 ± 9.56	180.10 ± 7.02	< 0.001
TG (mg/dl)	129.11 ± 19.23	119.21 ± 10.41	0.017
LDL (mg/dl)	108.20 ± 10.60	107.32 ± 7.51	0.211
HDL (mg/dl)	41.84 ± 7.43	44.16 ± 1.62	0.061

PCOS Polycystic ovarian syndrome, TC Total cholesterol, TG Triglyceride, HDL High-density lipoprotein, LDL Low-density lipoprotein, WC Waist circumference, BMI Body mass index. p < 0.05 was considered statistically significant; a, Mann–Whitney test

The two groups were well-matched for age, with no significant difference between PCOS patients (27.26 ± 2.06 years) and controls (27.54 ± 3.94 years; p = 0.365). However, significant differences were observed in several metabolic and anthropometric parameters. Body mass index (BMI) was significantly higher in PCOS patients (28.74 ± 2.31 kg/m²) compared to controls (25.19 ± 1.42 kg/m²; p < 0.001). Regarding BMI categories, the majority of PCOS patients were classified as overweight (25 < BMI < 30; n = 93, 46.5%) or obese (BMI > 30; n = 68, 34.0%), whereas most controls fell within the ideal BMI range (18.5 < BMI < 24.9; n = 114, 57.0%). Only 8 PCOS patients (4.0%) and 17 controls (8.5%) were underweight. Waist circumference (WC) was also significantly elevated in PCOS cases (101.63 ± 8.12 cm) compared with controls (92.51 ± 7.49 cm; p < 0.001), indicating a higher prevalence of central adiposity in the PCOS group. Regarding lipid profiles, TC and TG were significantly higher in PCOS patients than in controls (TC: 195.30 ± 9.56 vs. 180.10 ± 7.02 mg/dl, p < 0.001; TG: 129.11 ± 19.23 vs. 119.21 ± 10.41 mg/dl, p = 0.017). No significant differences were observed between the two groups for LDL: 108.20 ± 10.60 vs. 107.32 ± 7.51 mg/dl, p = 0.211 or HDL: 41.84 ± 7.43 vs. 44.16 ± 1.62 mg/dl, p = 0.061, although HDL levels tended to be lower in PCOS cases. FBS did not differ significantly between PCOS patients (100.69 ± 13.23 mg/dl) and controls (96.80 ± 5.91 mg/dl; p = 0.184). Taken together, these findings indicate that PCOS patients in our study population exhibited significantly higher BMI, waist circumference, total cholesterol, and triglyceride levels compared to healthy controls, consistent with the metabolic disturbances commonly associated with PCOS.

Genotyping of TGFA polymorphisms

The genotypic and allelic distributions of the two TGFA polymorphisms (rs11466297 and rs3732248) in PCOS cases and controls, along with the results of multiple genetic models, are presented in Table 3.

Table 3.

Allelic and genotypic distribution of TGFA polymorphisms

SNP	PCOS, n (%)	Control, n (%)	Genetic model	OR (95%CI) *	p-value*
rs11466297 A/C
AA	159 (82.8)	176 (89.3)		1 [ reference] -
AC	27 (14.1)	19 (9.6)	Codominant 1	1.59 (0.85–2.93)	0.155
CC	6 (3.1)	2 (1.0)	Codominant 2	3.35 (0.68–16.71)	0.125
HWE	0.001	0.085	Dominant	1.75 (0.98–3.16)	0.068
			Recessive	3.16 (0.64–15.80)	0.145
			Over dominant	1.55 (0.82–2.86)	0.179
A	345 (0.90)	371 (0.94)	Allelic	1 [ reference]
C	39 (0.10)	23 (0.06)	Allelic	1.85 (1.07–3.12)	0.029
rs3732248 G/A
GG	102 (52.8)	131 (66.2)		1 [ reference]
GA	83 (43.0)	50 (25.3)	Codominant 1	2.16 (1.39–3.32)	< 0.001
AA	8 (4.1)	17 (8.6)	Codominant 2	0.61 (0.26–1.47)	0.261
HWE	0.076	< 0.001	Dominant	1.77 (1.18–2.65)	0.013
			Recessive	0.49 (0.20–1.12)	0.081
			Over dominant	2.26 (1.47–3.45)	0.001
G	287 (0.74)	312 (0.79)	Allelic	1 [ reference]
A	99 (0.26)	84 (0.21)	Allelic	1.31 (0.93–1.81)	0.148

PCOS Polycystic ovarian syndrome, SNP Single-nucleotide polymorphism, CI confidence interval, OR odds ratio. HWE, Hardy Weinberg equilibrium; *p-value and OR (95%CI) were adjusted for BMI

Codominant 1 and Codominant 2 indicate the heterozygous and homozygous codominant models, respectively. p < 0.025 is considered statistically significant

All p‑values and odds ratios (ORs) were adjusted for body mass index (BMI), and the Bonferroni‑corrected significance threshold was set at p < 0.025 to account for testing two independent SNPs. For rs11466297 (A/C), the homozygous AA genotype was the most frequent in both groups. Under the codominant model, neither the heterozygous AC genotype (OR = 1.59, 95% CI: 0.85–2.93, p = 0.155) nor the homozygous CC genotype (OR = 3.35, 95% CI: 0.68–16.71, p = 0.125) reached statistical significance. Similarly, the dominant (AC + CC vs. AA), recessive (CC vs. AA + AC), and overdominant (AC vs. AA + CC) models showed no significant associations (all p > 0.05). Notably, under the allelic model, the minor C allele was significantly associated with an increased risk of PCOS (OR = 1.85, 95% CI: 1.07–3.12, p = 0.029), although this p‑value did not remain below the adjusted threshold of 0.025. For rs3732248 (G/A), the homozygous GG genotype was more common in controls, while the heterozygous GA genotype was more frequent in PCOS cases. Under the codominant model, the heterozygous GA genotype was associated with a significantly increased risk of PCOS compared to the GG genotype (OR = 2.16, 95% CI: 1.39–3.32, p < 0.001), and this association remained significant after Bonferroni correction (p < 0.025). In contrast, the homozygous AA genotype did not show a significant difference (OR = 0.61, 95% CI: 0.26–1.47, p = 0.261). Under the dominant model (GA + AA vs. GG), a significant association with PCOS risk was also observed (OR = 1.77, 95% CI: 1.18–2.65, p = 0.013), which remained below the corrected threshold. Furthermore, the overdominant model (GA vs. GG + AA) revealed the strongest effect, with heterozygous carriers having more than a two‑fold increased risk of PCOS (OR = 2.26, 95% CI: 1.47–3.45, p = 0.001). The recessive model (AA vs. GG + GA) was not significant (p = 0.081). Under the allelic model, the minor A allele was more frequent in PCOS cases than in controls (26% vs. 21%), but the association did not reach statistical significance (OR = 1.31, 95% CI: 0.93–1.81, p = 0.148). Taken together, these findings demonstrate that the rs3732248 GA heterozygote under the codominant and overdominant models, as well as the dominant model, is significantly associated with increased susceptibility to PCOS in our Iranian population, whereas rs11466297 showed no robust association after correction for multiple testing.

Haplotype and linkage disequilibrium analysis

Haplotype analysis identified four common haplotypes (frequency > 1%) for the two TGFA polymorphisms (rs11466297 and rs3732248): A–G (wild-type), A–A, C–G, and C–A. The most frequent haplotype was A–G, observed in 66.8% of PCOS cases and 74.8% of controls. None of the haplotypes showed a statistically significant association with PCOS risk (all p > 0.05). The C–G haplotype showed a non-significant trend toward increased risk (OR = 2.00, 95% CI: 0.82–4.94, p = 0.124). Linkage disequilibrium analysis revealed low LD between the two SNPs (D’ = 0.113, r² = 0.004) (Table 4).

Table 4.

Haplotype analysis of TGFA polymorphisms (rs11466297 and rs3732248) and their association with PCOS risk

Haplotype	PCOS, n (%)	Controls, n (%)	OR (95% CI)	p-value
A–G (Wild-type)	129 (66.8)	148 (74.8)	1.00 Reference	-
A–A	44 (22.9)	38 (19.2)	1.33 (0.81–2.18)	0.260
C–G	14 (7.2)	8 (4.0)	2.00 (0.82–4.94)	0.124
C–A	6 (3.1)	4 (2.0)	1.72 (0.47–6.23)	0.404

OR odds ratio, CI confidence interval, PCOS polycystic ovary syndrome, LD linkage disequilibrium. Haplotypes with frequency < 1% were excluded

The order of SNPs is rs11466297 followed by rs3732248. Values in parentheses are percentages of total haplotypes in each group (PCOS: n = 193; Controls: n = 198). Linkage disequilibrium measures (D’ and r²) between rs11466297 and rs3732248

Discussion

PCOS is a common and complex endocrine disorder affecting 6–26% of Iranian women [5] of reproductive age, characterized by hyperandrogenism, ovulatory dysfunction, and polycystic ovarian morphology [13]. Although genetic factors play a substantial role in PCOS susceptibility, the known loci identified by genome-wide association studies explain only a fraction of the heritability, particularly in non-European populations. The TGFA gene represents a compelling candidate, as it is expressed in ovarian granulosa and cumulus cells and activates the epidermal growth factor receptor (EGFR) signaling pathway, which regulates follicular development, steroidogenesis, and apoptosis [19]. In the present study, we investigated for the first time the association of two TGFA polymorphisms (rs11466297 and rs3732248) with PCOS risk in an Iranian population. A methodological concern that warrants careful consideration is the deviation from Hardy–Weinberg equilibrium (HWE) observed for rs3732248 in the control group (p < 0.001). It is important to distinguish between HWE deviations in cases versus controls. In the patient group, deviation from HWE (p = 0.076) is not statistically significant. However, in the control group, the significant deviation cannot be attributed to disease association and more likely reflects either subtle population stratification, a chance finding due to the modest sample size, or, less likely, genotyping artifacts although our rigorous quality control measures (15% duplicate genotyping with 100% concordance, inclusion of negative and positive controls, and high call rates of 97.25–97.75%) make systematic genotyping errors improbable. We acknowledge that the HWE deviation in controls is a limitation that could theoretically affect the validity of our findings. Nevertheless, the association between the rs3732248 GA genotype and PCOS risk remained highly significant under multiple genetic models (codominant, dominant, and overdominant) and survived Bonferroni correction for multiple testing. This consistency across models suggests that the observed association is unlikely to be an artifact of HWE deviation alone. However, we emphasize that our findings are preliminary and require replication in larger, independent cohorts with detailed ancestry information to definitively exclude population stratification as a confounding factor. The clinical and demographic characteristics of our study participants revealed that PCOS patients had significantly higher BMI, waist circumference, total cholesterol, and triglycerides compared to healthy controls. These findings confirm the well-established metabolic burden of PCOS, including central obesity and dyslipidemia. Importantly, all genetic associations were adjusted for BMI, reducing the likelihood that the observed genotype-phenotype relationships are confounded by obesity. The fact that rs3732248 remained significantly associated with PCOS after BMI adjustment suggests that this variant may influence PCOS susceptibility independently of obesity, possibly through direct effects on ovarian follicular development or androgen metabolism.

Our most striking finding was the significant association of rs3732248 with PCOS risk. Under the codominant model, the heterozygous GA genotype was associated with a more than two-fold increased risk of PCOS (OR = 2.16, 95% CI: 1.39–3.32, p < 0.001), and this association remained significant after Bonferroni correction. The overdominant model (GA vs. GG + AA) showed the strongest effect (OR = 2.26, 95% CI: 1.47–3.45, p = 0.001), suggesting a heterozygote advantage model of inheritance. Of note, rs3732248 has also been associated with non-syndromic cleft lip/palate in an independent study [20], suggesting potential pleiotropic effects, though the mechanisms differ from PCOS. To our knowledge, no previous study has investigated rs3732248 in PCOS, making our finding the first report of an association between this variant and PCOS susceptibility. For rs11466297, we observed a modest allelic association with PCOS (OR = 1.85, 95% CI: 1.07–3.12, p = 0.029), which did not survive Bonferroni correction. Therefore, we interpret this finding as preliminary and requiring replication in larger cohorts. A study by Avasthi and colleagues [21] in a North Indian population reported a significant association between rs11466297 and non-syndromic cleft lip and palate, suggesting that this variant may have context-dependent effects. Our genetic findings are strongly supported by existing biological evidence. A seminal study by Mason and colleagues [22] found no difference in TGFα protein levels between normal and polycystic ovaries, which at first glance appears to contradict our results. However, we propose that these observations are not mutually exclusive. Mason and colleagues measured protein levels, whereas we investigated genetic variants that may affect gene regulation, mRNA stability, or EGFR signaling efficiency without necessarily altering absolute TGFα concentrations. Furthermore, a comprehensive review by Homburg [23] proposed that hyperactivity of TGFα may block aromatase stimulation and attenuate follicular apoptosis, contributing to follicular arrest in PCOS. Our findings extend this line of evidence by demonstrating that specific genetic variants within TGFA are associated with PCOS risk, potentially explaining the functional hyperactivity described in the review. Finally, Singh and colleagues [24] demonstrated strong TGFA expression in porcine granulosa and cumulus cells across all stages of follicular development, with absence of expression in atretic follicles, providing direct biological evidence that TGFA plays an autocrine/paracrine role in healthy folliculogenesis.

We acknowledge several limitations. First, deviations from Hardy-Weinberg equilibrium were observed, and while we have provided justifications, we cannot rule out the possibility of subtle genotyping errors or population stratification. Second, our sample size provided only moderate power (approximately 45–65%) to detect modest genetic effects (OR < 1.5), which are typical in PCOS genetics. Third, we did not perform haplotype analysis, linkage disequilibrium assessment, or in silico functional predictions; these analyses were beyond the scope of the current study but represent important directions for future research. Fourth, our study population was recruited from a single geographical region, limiting generalizability to other ethnic groups. Fifth, we did not stratify PCOS patients by phenotypic subtypes or measure key hormonal parameters such as testosterone or LH/FSH ratio. Sixth, the genotyping methods used (PCR-RFLP and ARMS-PCR), while cost-effective, have higher error rates compared to modern techniques. Additionally, as highlighted by Vale-Fernandes and colleagues [25] in their comprehensive review on AMH as a diagnostic biomarker for PCOS, challenges such as variability in assay methods, absence of standardized cut-off values, and the influence of age, ethnicity, and obesity on biomarker levels limit universal applicability. Similarly, our genetic findings may be influenced by these population-specific factors, and future studies should incorporate robust biomarker measurements alongside genotyping to improve phenotype classification. Based on these limitations, we recommend replication studies in larger independent cohorts, functional studies including eQTL analysis and luciferase reporter assays, haplotype and LD analysis, in silico prediction of miRNA binding sites, and phenotype-stratified analyses across different ethnic populations. Although our findings are preliminary and require replication, they hold potential implications for personalized medicine. If validated in larger cohorts, the rs3732248 GA genotype could serve as a genetic risk marker for identifying women at increased risk of developing PCOS, particularly in the Iranian population. Furthermore, understanding the role of TGFA in PCOS pathogenesis could open new avenues for targeted therapeutic interventions modulating the EGFR signaling pathway. However, we emphasize that these applications remain speculative at this stage and require extensive validation through functional studies and clinical trials before any clinical translation. This study is the first to demonstrate a significant association between the TGFA polymorphism rs3732248 and PCOS susceptibility in an Iranian population. The association remained robust after Bonferroni correction and after adjustment for BMI. Our findings are supported by existing biological evidence showing TGFA expression in ovarian follicles and its role in folliculogenesis [23, 24]. Despite limitations including HWE deviations, moderate statistical power, and lack of functional validation, our study provides a strong foundation for future genetic and functional investigations of TGFA in PCOS.

Conclusion

In conclusion, this study is the first to demonstrate a significant association between the TGFA polymorphism rs3732248 and PCOS susceptibility in an Iranian population. The association remained robust after Bonferroni correction for multiple testing and after adjustment for BMI, suggesting that this variant may independently influence PCOS risk, possibly through effects on ovarian follicular development and EGFR signaling. While the rs11466297 variant showed only a modest allelic association that did not survive correction, it warrants further investigation in larger cohorts. Despite limitations including Hardy-Weinberg deviations, moderate statistical power, and lack of functional validation, our findings provide a strong foundation for future genetic and functional studies. Replication in independent populations and elucidation of the underlying molecular mechanisms are essential before any clinical translation, including potential applications in personalized risk assessment.

Abbreviations

PCOS

Polycystic ovary syndrome

TGFA

Transforming Growth Factor Alpha

SNP

Single nucleotide polymorphism

PCR-RFLP

Polymerase chain reaction-restriction fragment length polymorphism

ARMS-PCR

Amplification Refractory Mutation System

OR

Odds ratio

CI

Confidence interval

BMI

Body mass index

WC

Waist circumference

FBS

Fasting blood sugar

TC

Total cholesterol

TG

Triglyceride

HDL

High-density lipoprotein

LDL

Low-density lipoprotein

HWE

Hardy-Weinberg equilibrium

LD

Linkage disequilibrium

EM

Expectation-Maximization

EGFR

Epidermal growth factor receptor

GWAS

Genome-wide association study

DHEA-S

Dehydroepiandrosterone sulfate

LH

Luteinizing hormone

FSH

Follicle-stimulating hormone

ELISA

Enzyme-linked immunosorbent assay

EDTA

Ethylenediaminetetraacetic acid

UTR

Untranslated region

RE

Restriction enzyme

bp

Base pair

F

Forward primer

R

Reverse primer

DNA

Deoxyribonucleic acid

PCR

Polymerase chain reaction

mRNA

Messenger ribonucleic acid

eQTL

Expression quantitative trait locus

SPSS

Statistical Package for the Social Sciences

Authors’ contributions

MM, RS and MG Methodology and Design, FK, MM, RS and MG writing the draft, FK genotyping, RS data analysis and MG clinical patient assessment, MM and RS supervision and editing.

Funding

This work was supported by the University of Sistan and Baluchestan, Zahedan, Iran. The funding body had no role in the design of the study, collection, analysis, or interpretation of data, or in writing the manuscript.

Data availability

Data will be available upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the University of Sistan and Baluchestan Ethics Committee (approval code: IR.USB.REC.1402.037; available at: https://ethics.research.ac.ir/IR.USB.REC.1402.037). All procedures were performed in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki Declaration and its later amendments (The 75th WMA General Assembly, Helsinki, Finland, October 2024). Written informed consent was obtained from all participants prior to inclusion in the study.

Consent for publication

Not applicable. This manuscript does not contain any individual person’s data in any form (including individual details, images, or videos).

Competing interests

The authors declare no competing interests.