For the roughly one in ten women whose pregnancies are upended by hyperemesis gravidarum, the relentless vomiting that begins in the first trimester has long been treated as a mystery of hormones. Now a study combining whole-exome sequencing with a custom-built blood test points to a second player in the drama: GFRAL, the brainstem receptor through which the placental hormone GDF-15 triggers nausea and vomiting. The researchers report rare GFRAL variants found exclusively in women with the most severe form of pregnancy sickness, and they show for the first time that a cleaved, soluble fragment of the receptor circulates in the blood of pregnant women, where its timing and persistence track with disease. The findings, drawn from several partly small cohorts, remain exploratory, but they sketch a receptor-level mechanism for why some pregnancies desensitize to GDF-15 while others do not.

The biological backdrop is the GDF-15–GFRAL pathway. GDF-15 is a stress-inducible member of the TGF-beta superfamily whose highest physiological expression occurs in the placenta. From there it reaches the brainstem, where GFRAL, a glycosylphosphatidylinositol-anchored receptor expressed almost exclusively in neurons of the area postrema and the nucleus of the tractus solitarius, forms a signaling complex with the RET tyrosine kinase to transduce the hormone’s signal. Activation of this circuit produces the nausea and vomiting that make GDF-15 such a potent anorexigenic agent. Crucially, the receptor can be switched off by proteolytic cleavage: the enzyme MT1-MMP, also known as MMP14, cuts GFRAL at a juxtamembrane site, releasing a soluble ectodomain called sGFRAL and leaving behind a truncated receptor that can no longer be activated by GDF-15. In mice, haploinsufficiency or inhibition of MT1-MMP enhances GDF-15’s appetite-suppressing effects, confirming that cleavage inactivates the receptor.

Genetic evidence has already tied the pathway to hyperemesis gravidarum. Common variants in GDF15, the gene encoding the hormone, are among the strongest risk factors for the condition, and risk-associated GDF15 alleles are linked to lower baseline levels of circulating GDF-15 outside pregnancy. The prevailing model holds that women with low preconception GDF-15 remain hypersensitive to the hormone’s dramatic rise during pregnancy, whereas higher basal levels pre-condition the system and promote desensitization. But GDF15 is not the only gene implicated: common variants in GFRAL itself and in the progesterone receptor gene PGR also associate with risk. What remained unknown was whether rare GFRAL variants modulate susceptibility, and whether receptor cleavage actually occurs in human pregnancy.

To address the first question, the team sequenced the whole exomes of 926 women with hyperemesis gravidarum and 660 unaffected controls. They identified 14 rare GFRAL variants occurring exclusively in cases, including five missense variants, one of them novel and three predicted damaging by SIFT. Among these was a 5′ UTR variant, rs138536707, found at significantly elevated frequency in African American women with hyperemesis gravidarum, with an odds ratio of 2.6 in a comparison against a morning sickness clinic cohort. People of African ancestry are known to carry increased risk for the condition, making this variant of particular interest for follow-up. Another case, a woman whose illness was severe enough to require total parenteral nutrition, carried the p.Met370Thr variant at the GFRAL–RET51 interface; the authors speculate that replacing a bulky hydrophobic methionine with a polar threonine could subtly stabilize the GFRAL–RET signaling complex and slow its dissociation, though this remains hypothetical.

Variants found only in controls proved equally intriguing. Eleven rare variants absent from all cases included six missense changes, four intronic variants, and one frameshift elongation. The frameshift variant rs527905870 was identified in two controls who both reported pregnancies entirely free of nausea and vomiting, suggesting that a defective receptor might protect against pregnancy sickness. Three of the control-only missense variants, p.Pro218Ala, p.Ser222Asn, and p.Arg225His, sit at the intersection of the GDF-15 ligand-binding domain and the GDNF/GAS1 domain responsible for signal transduction, positions where altered function is plausible. The authors are careful to stress, however, that the study was not powered for quantitative enrichment analysis and that these case-exclusive and control-exclusive signals should be read as hypothesis-generating rather than as evidence of causality. GFRAL variations, they conclude, are far too infrequent to account for the majority of hyperemesis gravidarum or symptom-free pregnancies.

The study also examined whether a common non-coding GFRAL risk variant influences symptom duration. Nausea and vomiting generally resolve by the second trimester, but in hyperemesis gravidarum symptoms can persist well beyond 27 weeks’ gestation. Genotyping revealed that the risk allele rs7761177-T was associated with a 1.7-fold increased risk of prolonged disease, with an odds ratio of 1.74 and a marginal p value of 0.047 in a modest sample of 38 early-resolving and 87 prolonged cases. The team further sequenced rs10948901-T, recently identified as the strongest GFRAL risk variant in a large multi-ancestry study and used here as a proxy, but found no detectable correlation between the risk variants and circulating sGFRAL levels, leaving the mechanism linking the variant to prolonged symptoms unknown.

The serological arm of the study required building an in-house ELISA to measure sGFRAL in serum and plasma, something never before analyzed in pregnant women. The capture antibody, 3P10, binds GFRAL at the RET interaction interface centered on residue Arg294, within the MMP14-cleaved fragment. Any 3P10-reactive fragment detected in blood therefore implies a truncated membrane receptor left behind on brainstem neurons that is incapable of RET signaling, making sGFRAL a proxy for proteolytic receptor inactivation. Detection rates hovered between 37 and 47 percent across non-pregnant women and all three trimesters, without statistically significant differences. In the first trimester, however, detection was significantly higher in hyperemesis gravidarum cases, at 70 percent, than in women with ordinary nausea and vomiting of pregnancy, at 42 percent, or in symptom-free pregnancies, at 35 percent, with odds ratios of 3.35 and 4.38 respectively.

Longitudinal samples sharpened the picture. In ten women from an IVF cohort with normal pregnancy outcomes, sGFRAL was detectable in eight blood samples drawn between days 35 and 43 post-implantation, at a mean of 218.7 picograms per milliliter. Strikingly, by the second sampling between days 62 and 76, none of the eight analyzed women showed detectable sGFRAL any longer, a statistically significant drop. The only patient whose sGFRAL persisted was the one who went on sick leave because of nausea. In stark contrast, eight women receiving intravenous fluids for hyperemesis gravidarum-related dehydration showed no drop in sGFRAL between six and 12 weeks’ gestation, with means of 177.7 and 211.4 picograms per milliliter and no significant difference. Sustained receptor shedding in blood, the authors propose, may reflect persistent GFRAL synthesis in the brainstem of affected women, or more effective suppression of new synthesis in unaffected pregnancies. Notably, in the morning sickness clinic cohort, sGFRAL levels correlated inversely with early-pregnancy weight loss, consistent with a protective role of receptor cleavage.

A two-part regression model applied to the CGATE cohort at 12 and 24 weeks’ gestation added a mid-gestation dimension. Across both time points combined, women with hyperemesis gravidarum had significantly higher odds of detectable sGFRAL than symptom-free women, with an odds ratio of 3.03, while ordinary nausea and vomiting did not differ from symptom-free pregnancies. Yet among samples with detectable sGFRAL, concentrations were significantly lower in the hyperemesis gravidarum group, particularly at 24 weeks, where the fold change relative to symptom-free women was 0.339. Hyperemesis gravidarum thus differs from both comparison groups in the likelihood and in the magnitude of circulating sGFRAL, a pattern the authors interpret as altered dynamics of receptor shedding, synthesis, or both, while acknowledging that the present data cannot distinguish between the two.

The implications reach beyond pregnancy sickness. The GDF-15–GFRAL pathway is implicated in cancer cachexia and chemotherapy-induced nausea and vomiting, and pharmaceutical blockade is advancing rapidly: Pfizer’s anti-GDF-15 antibody ponesegromab safely increased weight and activity in cachexia patients, and CatalYm’s visugromab increased body weight in cancer patients with high GDF-15 levels. Clinical trials of GFRAL inhibition with the monoclonal antibody NGM120 are underway in the UK and Australia. If validated, sGFRAL analysis could serve as a novel biomarker for hyperemesis gravidarum diagnosis, risk stratification, and treatment response, and engineered sGFRAL mimetics acting as GDF-15 traps could become therapeutics for the condition and related disorders. The authors also note that in silico analysis predicts a GFRAL promoter rich in Sp1 sites with progesterone and non-canonical estrogen response elements, offering a possible explanation for the association between progesterone receptor variants and disease. Key limitations remain: low variant frequencies, small cohorts, missing ancestry data in several European samples, and the impossibility of directly measuring intact GFRAL on brainstem neurons in pregnant women. Rodents lack the vomiting reflex, making the disease difficult to model. Even so, the study opens a concrete, testable route into the receptor-level regulation of one of pregnancy’s most disabling conditions.

Subject of Research: Genetic variations in the GDF-15 receptor GFRAL and first measurements of circulating soluble GFRAL in hyperemesis gravidarum and normal pregnancy

Article Title: GFRAL genetic variations and circulating GFRAL in hyperemesis gravidarum

Article References: Haack, B., Neemuchwala, Z. H., DeVera, S., MacGibbon, K. W., Fischer, B., Höschle, F., Fasching, P. A., Housholder, A., Vanky, E., Friese, M. A., Bixo, M., Stiernman, L., Ekwurtzel, E., Nording, M., Vinnars, M.-T., Wischhusen, J., & Fejzo, M. S. (2026). GFRAL genetic variations and circulating GFRAL in hyperemesis gravidarum. iScience, 29(11), Article 117975. https://doi.org/10.1016/j.isci.2026.117975

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117975

Keywords: hyperemesis gravidarum, GDF-15, GFRAL, pregnancy, nausea and vomiting, genetics, biomarker, brainstem, RET signaling, MMP14, genetic, variations