The endoplasmic reticulum, or ER, is the cellular organelle responsible for folding and quality-controlling newly synthesized proteins. When protein folding demand outstrips capacity, misfolded proteins accumulate in the ER lumen, triggering a coordinated adaptive program known as the unfolded protein response. This response is orchestrated by three principal sensors embedded in the ER membrane, one of which is PERK, or protein kinase RNA-like ER kinase. Upon activation, PERK phosphorylates the translation initiation factor eIF2α, which globally dampens protein synthesis to relieve the folding burden while selectively upregulating ATF-4, a transcription factor that reprograms cells toward amino acid metabolism, antioxidant defense, and autophagy. Because the PERK–ATF-4 arm of the unfolded protein response is intimately tied to nutrient sensing and metabolic adaptation, researchers have increasingly suspected it might be relevant to hyperemesis gravidarum, a condition defined in part by starvation-like physiology.

The research team, led by Bahriye Oznur Oguçbilek of Darende Hulusi Efendi State Hospital in Malatya, Turkey, together with colleagues from Konya City Hospital and Selçuk University, designed a prospective case-control study conducted between August and November 2024. They enrolled 85 pregnant women diagnosed with hyperemesis gravidarum within the first 16 weeks of gestation, alongside 85 healthy pregnant women serving as controls. Ethical approval came from the Karatay University Scientific Research Ethics Committee, and all participants provided written informed consent. The investigators measured circulating levels of PERK and ATF-4 in serum using enzyme-linked immunosorbent assays, then compared the two groups and explored how the markers related to clinical and laboratory characteristics, including urinary ketone status, a bedside indicator of the fasting and fat-metabolism state that often accompanies severe vomiting.

The headline result was a clear and statistically robust difference in ATF-4. Women with hyperemesis gravidarum had significantly lower serum ATF-4 concentrations than controls, with a median of 4.55 nanograms per milliliter in the patient group versus 6.16 nanograms per milliliter in healthy pregnant women, a difference that reached significance at p equals 0.002. Strikingly, PERK levels themselves did not differ between the groups, with medians of 3.73 and 3.29 nanograms per milliliter respectively and a p-value of 0.965. This dissociation is intriguing from a mechanistic standpoint. If the PERK–ATF-4 axis were simply activated as part of a classical stress response, one might expect coordinated changes in both proteins. The selective reduction of ATF-4 without a corresponding PERK shift hints that the integrated stress response in hyperemesis gravidarum may be modulated at a level downstream of PERK, or that circulating ATF-4 reflects contributions from tissues responding to nutritional stress in ways that do not track with PERK abundance in blood.

Correlation analysis added another layer of complexity. Across the hyperemesis group, the control group, and the overall study population, ATF-4 and PERK levels were strongly and positively correlated, with Spearman correlation coefficients of 0.837, 0.709, and 0.765 respectively, all highly significant. This tight coupling suggests that the two proteins are physiologically linked in pregnancy, consistent with their partnership in the canonical stress pathway, even though the disease state shifts the overall set point of ATF-4 downward. In other words, the relationship between the two markers is preserved, but the baseline against which that relationship plays out appears to be altered in women experiencing severe nausea and vomiting.

The researchers also asked whether ATF-4 levels differed according to urinary ketone positivity within the hyperemesis group, reasoning that ketosis marks the most metabolically stressed patients. The answer was no: ATF-4 concentrations did not vary significantly with ketone status, with a p-value of 0.282. This null result suggests that the ATF-4 signal is not simply a readout of acute starvation severity as captured by a single urine dipstick measurement. It may instead reflect a broader or more chronic adaptation to the hyperemetic state, or it may be influenced by factors such as hydration status, treatment, or individual variation in stress pathway regulation that a ketone test cannot capture.

To assess clinical utility, the team performed receiver operating characteristic analysis, asking whether serum ATF-4 could discriminate hyperemesis patients from healthy pregnant women. The answer was, at best, modestly. The area under the curve was 0.638, statistically significant at p equals 0.001, but corresponding to limited discriminatory power. At the chosen threshold, sensitivity was 45.9 percent, meaning more than half of true cases would be missed, while specificity reached 77.6 percent. In practical terms, ATF-4 alone is not ready to serve as a diagnostic biomarker for hyperemesis gravidarum. However, the statistically significant separation between groups indicates the marker carries real biological information, and the authors frame their findings as a foundation for further molecular and longitudinal investigation of the PERK–ATF-4 axis rather than as the basis for an immediate clinical test.

The study sits within a rapidly evolving research landscape. Hyperemesis gravidarum has historically been under-studied and sometimes dismissed as a psychological or trivial complaint, but the 2018 identification of associations between the placental and appetite genes GDF15 and IGFBP7 and the condition, followed by the landmark 2024 Nature study linking GDF15 to maternal risk of nausea and vomiting of pregnancy, has transformed the field. GDF15 is now understood to act as an endocrine signal of nutritional and cellular stress, acting on hindbrain circuits that suppress appetite and induce nausea. The new ATF-4 findings dovetail conceptually with this work: the integrated stress response is precisely the intracellular machinery through which cells sense amino acid deprivation and metabolic duress, and ATF-4 is the transcriptional executor of that response. A lowered circulating ATF-4 in hyperemesis patients may therefore represent another fingerprint of the same stress physiology that elevated GDF15 signals, though the direction of the change and the causal chain remain to be established.

Endoplasmic reticulum stress is also well documented in other pregnancy complications, lending plausibility to its involvement in hyperemesis. Increased ER stress has been reported in decidual tissue from pregnancies complicated by fetal growth restriction and preeclampsia, and ER stress signaling has been shown to mediate proinflammatory cytokine inhibition of trophoblast invasion. The PERK–ATF-4–CHOP pathway has been implicated in hypoxia-induced trophoblast apoptosis in preeclampsia, and the same research group has previously examined ER stress parameters in gestational diabetes mellitus. Placental ER stress is often described as a double-edged sword, with moderate activation supporting adaptation while excessive or prolonged activation contributes to pathology. Against this backdrop, the observation that hyperemesis gravidarum is associated with reduced rather than elevated circulating ATF-4 is a reminder that systemic markers may not mirror tissue-level stress responses, and that the placenta, liver, and other metabolically active organs may each contribute differently to the blood pool of these proteins.

The authors are careful about the limitations inherent in their design. As a cross-sectional case-control study, it demonstrates association rather than causation, and it cannot determine whether reduced ATF-4 precedes, accompanies, or results from hyperemesis gravidarum. ELISA measurements of circulating proteins capture only what reaches the bloodstream, not the intracellular activation state of the pathway in placental or hypothalamic tissue where the biology likely unfolds. The study period was confined to a few months in late 2024, and the cohort was drawn from Turkish hospital settings, raising questions about generalizability that only multi-center and longitudinal work can resolve. Still, the strong correlations, the significant group difference, and the explicit call for longitudinal follow-up give the field a concrete next step. If future studies confirm and extend these findings, the PERK–ATF-4 axis could become a mechanistic bridge connecting the hormonal triggers of hyperemesis gravidarum, such as GDF15, to the cellular stress physiology that produces its debilitating symptoms, and perhaps eventually to targeted interventions for the roughly one to three percent of pregnancies it complicates worldwide.

Subject of Research: Endoplasmic reticulum stress markers PERK and ATF-4 in hyperemesis gravidarum

Article Title: Investigation of Endoplasmic Reticulum Stress in Pregnant Women with Hyperemesis Gravidarum

Article References: Investigation of Endoplasmic Reticulum Stress in Pregnant Women with Hyperemesis Gravidarum. (n.d.). https://doi.org/10.1007/s43032-026-02228-y

Image Credits: AI Generated

DOI: 10.1007/s43032-026-02228-y

Keywords: hyperemesis gravidarum, endoplasmic reticulum stress, ATF-4, PERK, unfolded protein response, pregnancy, GDF15, integrated stress response, ketosis, ELISA, placenta, Reproductive Sciences