rs2479106 — DENND1A

DENND1A rs2479106 — The Androgen Gene Switch Linked to PCOS Risk

Polycystic ovary syndrome affects 5–15% of women of reproductive age and is the leading cause of anovulatory infertility worldwide. One of its defining features is hyperandrogenism¹¹ hyperandrogenism
excess androgen production, typically from ovarian theca cells, causing irregular periods, hirsutism, and impaired follicle maturation. A landmark 2011 genome-wide association study identified a cluster of variants in the DENND1A gene at chromosome 9q33.3 as among the first confirmed PCOS susceptibility loci — with rs2479106 carrying an odds ratio of 1.34 (P=8.12×10⁻¹⁹) in over 10,000 Han Chinese women. Subsequent functional work has revealed why this locus matters: DENND1A directly controls the theca cell machinery that produces androgens.

The Mechanism

DENND1A encodes DENN domain-containing protein 1A²² DENN domain-containing protein 1A
a guanine nucleotide exchange factor involved in endosomal trafficking and membrane receptor recycling. The protein has an alternatively spliced isoform, DENND1A.V2, that is markedly overexpressed in the theca cells of women with PCOS. When researchers forced expression of DENND1A.V2 in normal theca cells, the cells acquired a PCOS-like phenotype: they upregulated CYP17A1³³ CYP17A1
steroid-17α-hydroxylase/17,20 lyase, the rate-limiting enzyme in androgen biosynthesis and CYP11A1 (cholesterol side-chain cleavage enzyme), producing excess androgens and progestins. Conversely, knocking out DENND1A.V2 reduced these enzymes and suppressed androgen output.

The rs2479106 variant sits in an intron of the DENND1A gene, spanning a region containing at least 38 candidate regulatory elements between introns 2 and 6⁴⁴ at least 38 candidate regulatory elements between introns 2 and 6
identified by ATAC-seq and ENCODE enhancer overlap in a 2024/2025 study. Epigenetic activation of these intronic enhancers using dCas9-P300 produced 1.7–3.2-fold increases in testosterone in adrenal cell models. The rs2479106 locus thus marks a region of the genome that regulates how much DENND1A is expressed in steroidogenic cells — and consequently how much androgen those cells produce.

The Evidence

The original discovery was made in a staged GWAS by Chen et al. 2011 in Nature Genetics⁵⁵ Chen et al. 2011 in Nature Genetics
discovery cohort: 744 PCOS cases/895 controls; two replication cohorts totaling 3,338 PCOS cases/5,792 controls; all Han Chinese. The G allele at rs2479106 conferred an odds ratio of 1.34 at a combined P-value of 8.12×10⁻¹⁹ — statistical confidence well beyond genome-wide significance.

A 2013 genotype-phenotype study of over 2,000 Han Chinese PCOS women found that carriers of the G allele (GG+AG genotype) had significantly elevated serum insulin levels 2 hours after a 75g oral glucose challenge⁶⁶ significantly elevated serum insulin levels 2 hours after a 75g oral glucose challenge
P=0.02, dominant model, suggesting impaired post-load insulin clearance or early insulin secretory dysfunction. A 2020 follow-up in 2,082 PCOS women refined this: the AA genotype (no G alleles) was actually associated with a higher rate of insulin resistance (53.6% vs 48.3%; OR 0.80 for GG+AG, P=0.036 after age/BMI adjustment), though this association disappeared when subjects with a family history of diabetes were excluded — suggesting complex confounding.

Meta-analyses confirm population-specific effects. Gao et al. 2016 (8 studies, 8,185 cases/28,675 controls)⁷⁷ Gao et al. 2016 (8 studies, 8,185 cases/28,675 controls) found a significant association in Asian populations (OR 1.32, 95% CI 1.25–1.39) but not in European populations (OR 1.01). Similarly, a 2012 replication study in Caucasian European cohorts (1,144 cases/17,635 controls) found OR 1.05 (P=0.45) for rs2479106, while the neighboring SNP rs10818854 replicated strongly with P=9.8×10⁻⁸. This suggests rs2479106 is a tag SNP that tracks the causal variant in Asian populations through linkage disequilibrium, but that LD pattern differs across ancestries.

A 2023 meta-analysis by Larsen et al. (10 studies, 3,627 cases/20,325 controls)⁸⁸ Larsen et al. (10 studies, 3,627 cases/20,325 controls)
including subgroup analyses by ancestry and genetic model found the Asian subgroup recessive model showed OR 1.84 (P=0.006), and the overall dominant model approached significance at OR 1.31 (P=0.05).

Separately, a 2025 Nature Communications study from Mount Sinai and Duke University⁹⁹ a 2025 Nature Communications study from Mount Sinai and Duke University
using ATAC-seq, allele-specific reporter assays, and dCas9-P300 epigenetic editing in human PCOS theca cell models identified 4 regulatory variants in the DENND1A locus with allele-specific activity, providing the first direct molecular evidence that inherited variants in this region can dysregulate DENND1A expression and drive testosterone overproduction.

Practical Actions

For women carrying the G allele — particularly those of East or Southeast Asian ancestry where this variant has the strongest population-level effect — rs2479106 adds to the evidence base for earlier reproductive endocrinology evaluation if PCOS symptoms are present. Specific monitoring for androgen excess (total and free testosterone, DHEAS, androstenedione) and post-load insulin dysregulation (2-h glucose tolerance test) is warranted, as these are the phenotypic features most consistently associated with G-allele carriage in published cohorts.

PCOS management strategies that specifically address androgen-driven ovulatory dysfunction include inositol supplementation (myo-inositol reduces androgens and improves ovulatory function through insulin-sensitizing pathways) and anti-androgen monitoring at reproductive transitions (puberty, pregnancy planning, perimenopause).

Interactions

rs10818854 and rs10986105 (DENND1A): These two neighboring DENND1A variants show stronger and more consistent PCOS association in European populations (OR 1.36 and 1.39 respectively in meta-analyses, P<0.001 in European cohorts). rs10818854 and rs10986105 are in high LD with each other (r²>0.65) but not strongly with rs2479106. The three variants capture different aspects of risk in different ancestral LD structures.

rs6166 (FSHR N680S): The FSH receptor sensitivity variant interacts with PCOS susceptibility at the ovarian level. Women with PCOS-associated genotypes at DENND1A (excess androgen production) and low FSH receptor sensitivity (FSHR GG) may face a double challenge: impaired follicle development from reduced FSH response AND hyperandrogenic suppression of folliculogenesis. Combined profiling of DENND1A + FSHR may better characterize ovarian response to stimulation in a PCOS context. This is a proposed compound interaction — no published clinical trial has formally tested this combination.

rs13405728 (LHCGR): Another PCOS GWAS locus, the LH/hCG receptor. Elevated LH in PCOS drives theca cell androgen production through LHCGR; carriers of risk alleles at both DENND1A and LHCGR may have a compounding androgenic phenotype. Published data from the Chen 2011 GWAS identified both loci simultaneously, but compound effects of carrying risk alleles at both have not been quantified in published studies.