Maternal Autoantibodies That Sabotage a Key Inflammation Brake May Drive Preterm Birth

Preterm birth remains the leading cause of death in children under five worldwide, yet for most affected pregnancies no clear cause is ever identified. A new study published in Nature Immunology by Elze Rackaityte, Joseph DeRisi and colleagues at the University of California San Francisco, the Chan Zuckerberg Biohub and international collaborators now points to a surprising contributor: the mother’s own antibodies. By profiling the full landscape of autoantibodies circulating in thousands of pregnant women, the team found that pregnancy itself triggers a broad, shared program of self-reactivity, and that women who deliver early carry a distinct and expanded antibody signature that appears weeks or even months before labor begins.

The research hinged on a technology called phage immunoprecipitation sequencing, or PhIP-seq, which uses a programmable phage display library covering the human proteome to detect antibodies against nearly any human protein fragment in a single serum sample. The team applied the method to 3,101 serum samples drawn from 2,195 pregnant participants across eight cohorts in the United States, the United Kingdom and Italy, alongside 130 samples from never-pregnant controls. In total, 4,537 immunoprecipitations were performed, with rigorous quality control including technical replicates and a commercial antibody against the protein GFAP used as a positive control across plates and batches.

The first major finding was that normal pregnancy is accompanied by a striking remodeling of the maternal antibody repertoire. Using a cohort of women who conceived through frozen embryo transfer, the researchers measured each woman’s autoantibody profile against her own prepregnancy baseline and identified more than 3,000 pregnancy-specific peptide reactivities. A logistic regression classifier trained on these features could distinguish pregnant from nonpregnant sera with a mean area under the curve of 0.93, demonstrating that a large, shared set of autoreactive changes reliably accompanies gestation. The signature recapitulated known reactivities to histocompatibility molecules, pregnancy-associated proteins and sperm antigens, and revealed many targets not previously linked to pregnancy.

When the team compared women who delivered preterm with those who delivered at term, a second, more ominous pattern emerged. Preterm pregnancies carried significantly more reactivities within a preterm-specific autoantibody signature, encompassing 9,209 peptides from 3,594 proteins, and the breadth of this signature grew with advancing trimester, with divergence already visible in the first trimester. In logistic regression models, higher antibody burden predicted higher odds of preterm delivery, with an odds ratio of 1.54 per standard deviation increase, rising to 1.80 when the top 100 signature features were considered. Importantly, the antibody signal provided risk information largely independent of established clinical factors: women with both a prior preterm birth and elevated autoantibody levels had an odds ratio of 8.36 compared with women with neither risk factor.

The signature also showed biologically meaningful clinical correlations. It became progressively elevated during the 19 weeks preceding preterm delivery, was stronger in very early deliveries before 28 weeks than in moderate or late preterm births, and was amplified in women with four or more previous pregnancies, suggesting both a severity gradient and a cumulative pregnancy-dosage effect. To validate the findings independently, the researchers used sera from a subset of women to immunoprecipitate native proteins from placental extract and identified antibody-bound targets by mass spectrometry. Of 343 proteins recovered, 65 percent were also detected by PhIP-seq, and unsupervised clustering of the mass spectrometry data cleanly separated preterm from term samples.

Among the thousands of targets, one stood out for its mechanistic plausibility: interleukin-1 receptor antagonist, or IL1RA, the secreted brake on IL-1β, a cytokine long known to be capable of inducing labor when introduced into the intrauterine cavity of primates and rodents. Although anti-IL1RA antibodies were rare, detected in roughly 0.4 percent of pregnancies by PhIP-seq, orthogonal methods found higher prevalence: immunoprecipitation mass spectrometry recovered antibody-bound IL1RA in 16 percent of preterm samples, and a Luminex assay detected the antibodies in 6 percent of preterm pregnancies versus under 1.5 percent of term pregnancies. In a longitudinal cohort, early emergence of anti-IL1RA before 18 weeks was associated with a roughly sixfold increase in preterm risk, while antibodies appearing later in term pregnancies tended to lack neutralizing activity.

Epitope mapping showed that antibodies from multiple preterm patients converged on the same region of IL1RA, between amino acids 99 and 129, a highly charged stretch containing the protein’s only N-linked glycosylation site. In cell-based reporter assays, both commercial anti-IL1RA antibodies and anti-IL1RA-positive patient sera neutralized the antagonist, shifting the inhibitory concentration of IL1RA sixfold and licensing IL-1β signaling. The functional consequences were then tested in pregnant mice: dams injected intravenously with anti-IL1RA antibody during embryonic days 13.5 to 15.5, particularly in the context of IL-1β-driven inflammation, showed elevated maternal IL-1β, increased fetal resorption, pale and poorly perfused pups, reduced fetal weights, and placentas with constricted vessels, immune cell infiltration and elevated apoptosis on histological examination.

A final set of experiments revealed where the pathology might take hold in human pregnancy. Immunofluorescence of term human placentas showed that IL1RA, although classically considered a secreted protein, localizes prominently to CD31-positive fetal vessels, co-localizing with VE-cadherin at endothelial adherens junctions, structures central to vascular integrity. When placental sections were pre-blocked with anti-IL1RA-positive preterm sera, the characteristic junctional staining pattern was abolished, demonstrating that patient antibodies can recognize native, tissue-bound IL1RA in the placenta. This finding suggests a two-hit model in which dysregulated IL-1β signaling, whether preexisting or infection-induced, is amplified by antibodies that remove the antagonist, tipping the balance toward inflammation, placental vascular injury and early delivery.

The study’s implications extend beyond a single antibody. Rare genetic deficiency of IL1RA, known as DIRA, is itself associated with preterm birth and growth restriction, and IL1RA-deficient mice show reduced litter sizes and vasculitis, mirroring the antibody-induced phenotype. Because recombinant IL1RA, the drug anakinra, has been used clinically for more than two decades, the authors suggest that its use in pregnancy to prevent preterm labor, particularly in women stratified by anti-IL1RA antibody status, merits further evaluation. More broadly, the work reframes preterm birth as a condition with identifiable autoantibody-defined subtypes, shows that prior pregnancies may immunologically prime the maternal response to fetal and placental antigens, and offers a serological window into maternal-fetal tolerance that could ultimately yield earlier diagnostics and targeted immunomodulatory treatments for pregnancies at risk.

Subject of Research: Maternal autoantibodies, including anti-IL1RA antibodies, as contributors to the pathogenesis of preterm birth

Article Title: Evidence for maternal autoantibodies in the pathogenesis of preterm birth

Article References: Evidence for maternal autoantibodies in the pathogenesis of preterm birth. (n.d.). https://doi.org/10.1038/s41590-026-02689-y

Image Credits: AI Generated

DOI: 10.1038/s41590-026-02689-y

Keywords: preterm birth, autoantibodies, IL1RA, maternal-fetal tolerance, PhIP-seq, placenta, inflammation, IL-1 beta, pregnancy immunology, proteomics, anakinra, placental vasculature