Association between circulating inflammatory proteins and recurrent spontaneous abortion

To the Editor: Recurrent spontaneous abortion (RSA), defined as two or more consecutive pregnancy losses before 20 weeks of gestation with the same partner, affects 2%–5% of women of reproductive age worldwide. It not only disrupts the current pregnancy but also elevates the risk of subsequent complications, including intrauterine adhesions, endometriosis, and cervical insufficiency. The etiology of RSA is multifactorial, involving chromosomal abnormalities, uterine anomalies, autoimmune disorders, and endometrial dysfunction. Accumulating evidence points to immune tolerance imbalance as a key contributing factor. 1 Immune homeostasis during pregnancy relies on finely tuned interactions among maternal immune cells, cytokines, and their associated signaling pathways. Inflammatory proteins are critical mediators in this process, influencing trophoblast invasion, placental development, and maternal–fetal tolerance. A balanced ratio of T-helper 1 (Th1) and Th2 cytokines ensures regulated inflammatory responses, whereas the equilibrium between Th17 and regulatory T (Treg) cells is essential for maintaining tolerance at the maternal–fetal interface. Dysregulation of these pathways, along with aberrant signaling of key interface molecules such as interleukin-23 receptor (IL-23R) and T cell immunoglobulin and mucin domain-containing protein 3 (Tim-3) /galectin-9 (Gal-9), has been observed in RSA patients, underscoring the pivotal role of inflammatory mediators in sustaining successful gestation. Although observational studies have reported correlations between inflammatory proteins and RSA, causal relationships remain unclear. Mendelian randomization (MR) uses genetic variants, specifically single-nucleotide polymorphisms (SNPs), as instrumental variables to infer causality, mimicking the randomized controlled trial. 2 To evaluate whether circulating inflammatory proteins contribute causally to RSA, we performed a two-sample MR analysis using 91 inflammatory proteins exposure and RSA outcome data from an independent genome-wide association study (GWAS) cohort Supplementary Table 1, https: //links. lww. com/CM9/C848. The instrumental variables were selected based on genome-wide significance (P 10, confirming sufficient strength. We applied four MR methods: (1) inverse-variance weighted (IVW, primary method), (2) weighted median, (3) weighted mode, and (4) Mendelian randomization Egger (MR-Egger) regression. Pleiotropy was assessed and robustness confirmed using Mendelian Randomization Pleiotropy RESidual Sum and Outlier (MR-PRESSO), heterogeneity tests, and leave-one-out analyses. Colocalization analysis tested whether protein- and RSA-associated signals shared causal variants. To validate the MR findings at the protein level, circulating levels of implicated proteins were measured, and their tissue-specific expression was examined using RNA-seq datasets from decidua (GSE161969) and villus (GSE121950). Our MR analysis identified five circulating inflammatory proteins with significant, inverse causal associations with RSA risk Figure 1A and Supplementary Figure 1, https: //links. lww. com/CM9/C848. In IVW models, genetically predicted higher levels of programmed death-ligand 1 (PD-L1; odds ratio OR = 0. 588, 95% confidence interval CI = 0. 417–0. 829, P = 0. 002), urokinase-type plasminogen activator (uPA; OR = 0. 736, 95% CI = 0. 574–0. 942, P = 0. 015), C-C motif chemokine ligand 4 levels (CCL4; OR = 0. 802, 95% CI = 0. 661–0. 972, P = 0. 025), interleukin-20 receptor subunit alpha (IL20-RA; OR = 0. 640, 95% CI = 0. 437–0. 935, P = 0. 021), and neurotrophin-3 (NT-3; OR = 0. 664, 95% CI = 0. 463–0. 953, P = 0. 026) were consistently associated with reduced RSA risk. This suggests that lower circulating levels of these proteins may increase susceptibility to pregnancy loss. Sensitivity analyses confirmed the stability and reliability of these estimates, revealing no significant horizontal pleiotropy or outlier-driven effects Supplementary Figure 2 and Supplementary Table 2, https: //links. lww. com/CM9/C848. Among these proteins, CCL4 showed the most robust evidence, with consistent effects observed across all MR methods (MR-Egger, weighted median, weighted mode, and MR-PRESSO). Reverse MR analyses did not support a causal effect of RSA on circulating protein levels, indicating a unidirectional relationship. Figure 1: Identification and validation of inflammatory proteins associated with recurrent spontaneous abortion (RSA). (A) The forest plot shows the causal association between programmed cell death 1 ligand 1 (PD-L1), urokinase-type plasminogen activator (uPA), interleukin-20 receptor subunit alpha (IL20-RA), C-C motif chemokine 4 levels (CCL4), neurotrophin-3 (NT-3), and RSA. (B) ELISA validation of plasma levels of five proteins in RSA patients vs. healthy controls (HC) in early pregnancy. For all panels, unpaired t-tests was used for analyzing statistical significance; * P <0. 001. CI: Confidence interval; ELISA: Enzyme-linked immunosorbent assay; IVW: Inverse-variance weighted; MR-Egger: Mendelian randomization Egger regression; MR-PRESSO: Mendelian Randomization Pleiotropy RESidual Sum and Outlier; OR: Odds ratio; SNPs: Single nucleotide polymorphisms. Bayesian colocalization analysis further supported a shared genetic architecture between the inflammatory proteins and RSA. High posterior probabilities for a shared causal variant were observed for IL20-RA (rs11987533, Posterior Probability for Hypothesis 4 PPH4 = 0. 99), PD-L1 (rs822335, PPH4 = 0. 86), uPA (rs4251805, PPH4 = 0. 94), and CCL4 (rs113010081, PPH4 = 1. 00), indicating strong genetic coregulation with RSA Supplementary Figure 3, https: //links. lww. com/CM9/C848. In contrast, NT-3 showed no meaningful colocalization signal (PPH4 <0. 8), suggesting that MR association may not be driven by a shared causal SNP. It has been reported that CCL4 plays a central role in sustaining a supportive maternal–fetal environment by facilitating trophoblast migration, modulating immune responses at the fetal–maternal interface, and recruiting uterine natural killer cells. These effects are partly mediated through interactions with C-C chemokine receptor types 1 (CCR1) and 5 (CCR5) on immune cells, which contributes to immune tolerance during early pregnancy. 3 uPA has also been shown to promote extracellular matrix degradation and remodeling, thereby enhancing trophoblast invasiveness through interactions with its receptor (uPAR) and activation of downstream phosphatidylinositol 3-kinase (PI3K) /protein kinase B (Akt) and mitogen-activated protein kinase signaling pathways. Furthermore, uPA upregulates factors such as fucosyltransferase 4 (FUT4) and protein O-fucosyltransferase 1 (poFUT1), supporting trophoblast migration, uterine angiogenesis, and vascular remodeling. 4 IL-20RA can establish an immunosuppressive microenvironment through Janus kinase 1 (JAK1) -signal transducer and activator of transcription 3 (STAT3) signaling, influencing the recruitment and activity of anti-inflammatory immune cells. PD-L1 has also been widely implicated in maternal–fetal immune tolerance, with decreased expression linked to RSA in earlier studies. 5 NT-3, while supported by weaker evidence, may contribute to early embryonic development, implantation, and placental formation through neurotrophin-mediated signaling pathways. To validate these genetic findings at the protein level, serum concentrations of the five proteins were measured by enzyme-linked immunosorbent assay (ELISA) in RSA patients (n = 10) and healthy controls (women with uncomplicated early pregnancies, n = 9) during early pregnancy. All procedures were approved by the Human Research Ethics Committee of the Obstetrics and Gynecology Hospital of Fudan University (No. 2020-S218). The two groups were comparable in age, body mass index, and obstetric history. All participants were screened to exclude embryonic chromosomal abnormalities, maternal comorbidities, and structural reproductive anomalies. Consistent with the MR results, serum levels of all five proteins were higher in healthy controls than in patients with RSA. IL20-RA (P = 0. 0006) and CCL4 (P = 0. 0002) were markedly lower in women with RSA, whereas PD-L1, uPA, and NT-3 showed nonsignificant but directionally consistent reductions Figure 1B. During pregnancy, decidualization of the endometrium and its interaction with invading trophoblasts lead to the formation of two distinct placental structures: (1) the maternal-derived decidua and (2) the fetal-derived placental villus. While both tissues produce inflammatory proteins, their cellular origins differ; villous tissues primarily supply fetal-derived mediators. Transcriptomic analysis of RNA-seq datasets showed no significant differences in decidual tissues between RSA patients and healthy controls. In contrast, villous tissues from healthy pregnancies exhibited significantly higher expression of CCL4 and PLAU (encoding uPA), with CCL4 expression approximately fivefold higher (P <0. 01), supporting a trophoblast-derived origin for these proteins and aligning with the MR and ELISA findings Supplementary Figure 4, https: //links. lww. com/CM9/C848. Together, the integration of MR, colocalization, ELISA, and tissue-specific transcriptomic analyses indicates that these five proteins likely function as protective factors against RSA. Among them, CCL4 demonstrates the strongest and most consistent evidence across genetic, circulating protein, and villous expression data, highlighting its potential as a predictive biomarker and a candidate therapeutic target. The strengths of our study include the complementary integration of MR, colocalization, protein quantification, and tissue-specific transcriptomics, which together strengthen causal inference and mitigate risks of false-positive or false-negative findings. The use of large-scale GWAS data and validation in an independent Chinese cohort further enhances generalizability. However, several limitations must be acknowledged. Population heterogeneity arising from genetic and environmental differences may affect extrapolation. Other constraints include the exploratory nature of the MR analysis without multiple testing correction, variable strength of instrumental variables, and the inability to fully rule out residual pleiotropy or confounding. Despite these limitations, the consistent evidence across genetic, protein, and transcriptomic levels robustly supports protective roles for these inflammatory proteins, particularly CCL4 and IL20-RA. Future studies in diverse populations are needed to confirm these associations and elucidate their underlying mechanisms in RSA. In summary, our study provides integrated evidence for the causal and functional involvement of specific circulating inflammatory proteins in RSA. These findings advance the understanding of immune dysregulation in RSA, highlighting the contribution of trophoblast-derived factors to maternal–fetal tolerance. The results also offer a rationale for developing early diagnostic biomarkers and exploring targeted therapeutic strategies to improve pregnancy outcomes. Future research should aim to validate these proteins in larger, more diverse cohorts and to precisely delineate their roles in reproductive immunology. Funding This work was supported by grants from the National Key Research and Development Program of China (No. 2025YFC2708002) and National Natural Science Foundation of China (Nos. U23A20405, 82271713, 82101706, and 82502020) Data availability The datasets generated and/or analyzed during the current study are available in NHGRI-EBI GWAS catalog (https: //www. ebi. ac. uk/gwas/), FinnGenR10 (https: //r10. risteys. finngen. fi/), and NCBI (https: //www. ncbi. nlm. nih. gov/).