Age at Menopause

Age at natural menopause varies substantially between women and is one of the more heritable common traits studied in human genetics. Three studies, spanning European and Japanese cohorts and both ordinary and clinically early menopause, converge on a real biological story rather than a scattered gene list.

DNA repair, showing up again and again

Stolk et al. 2012 meta-analysed 22 genome-wide association studies in 38,968 women of European descent, replicating in up to 14,435 more, and found 13 new loci beyond four already known. The candidate genes at these loci cluster strikingly around two biological themes: DNA repair (rs1635501 in EXO1, rs10183486 in TLK1, among others) and immune function (rs12461110 in NLRP11). Pathway analysis pointed specifically to exonuclease activity, NF-κB signalling and mitochondrial dysfunction. Seven of this page's variants come from here.

Horikoshi et al. 2018 studied up to 67,029 women of Japanese ancestry — most prior menopause genetics had been European-only, a real gap this study closed. It found 26 loci for reproductive ageing overall, highlighting GNRH1, the gene for the master reproductive hormone that governs the hypothalamic-pituitary-gonadal axis, as a menopause-timing candidate — evidence that the brain's own reproductive hormone signalling plays a direct role, not just the ovary ageing passively on its own. One gene, MCM8 — part of the cellular machinery that licenses DNA to replicate — turned up independently in both this Japanese study and the 2012 European one above: the same gene, a different ancestry, the same finding. Seven of this page's variants come from here.

The DNA-repair theme is not a coincidence of two papers happening to name similar genes. A plausible mechanism connects it directly to biology: oocytes are formed once, before birth, and must survive and remain capable of accurate DNA repair for decades until they are used or lost. Genetic variation that shifts how well that repair machinery works, over that much time, is exactly the kind of thing that would show up as variation in how long the ovarian reserve lasts.

Day et al. 2015 pushed the DNA-repair connection further, in a dual study of common and rare protein-coding variants across roughly 70,000 women. It found 44 regions with common variants (plus 2 more carrying rare, large-effect missense alleles), enrichment near genes involved in the onset of puberty — the first molecular link between the start and end of reproductive life — and, most strikingly, the first common coding variant in BRCA1 ever associated with any complex trait. Using Mendelian randomization, the same study found evidence that later menopause causally raises breast cancer risk by roughly 6% per year of delay, likely through prolonged exposure to sex hormones rather than through the DNA-repair mechanism itself. rs5762534, on this page, sits in CHEK2 — itself a well-known DNA-damage-response and breast-cancer-susceptibility gene, in the same family of biology as BRCA1. Twenty of this page's variants come from this study.

The same genetics, at its earliest extreme

Perry et al. 2013 asked a related but distinct question: is early menopause — occurring before age 40, affecting up to 10% of women and a leading cause of infertility — genetically separate from ordinary variation in timing, or the same architecture pushed further? Meta-analysing 3,493 early-menopause cases and 13,598 controls across 10 studies, the study found no new loci of its own, but confirmed that the 17 variants already known for normal age at menopause were also associated with early menopause and primary ovarian insufficiency. One of this page's variants, rs1046089 in PRRC2A, comes from this study. The same genetic architecture underlies both the ordinary range of variation and its clinically significant early extreme — not two separate causes.