Heterogeneity in inflammatory responses to endotoxin at the fetomaternal interface

Abstract

The fetomaternal interface (FMi), comprising fetal chorionic trophoblast cells (CTCs) and maternal decidual stromal cells (DECs), plays a critical role in providing immune tolerance during pregnancy. Intrauterine inflammation is major trigger of adverse outcomes such as preterm birth, yet the cell-specific inflammatory responses at the FMi -; poorly defined. We investigated differential inflammatory responses of fetal and maternal cell populations at the FMi upon exposure to endotoxin (lipopolysaccharide [LPS]). Primary CTCs and DECs were isolated from human term fetal membrane tissues. Cells were treated with LPS (100 ng/mL, 48 h. Transcriptomic profiling, multiplex immunoassays and western blotting was performed. Regulatory network analysis identified upstream drivers of cell-specific responses. LPS induced a pronounced inflammatory response in DECs, marked by high expression of proinflammatory chemokines, and prostaglandin enzymes associated with adverse pregnancy outcomes. In contrast, CTCs exhibited attenuated response, characterized by selective induction of stress-associated genes with minimal activation of inflammatory pathways. Network analysis revealed distinct cell type–specific regulatory hubs, including STAT1 and IRF7 in DECs and RELA and MYD88 in CTCs. DECs also activated anti-inflammatory signaling and pyroptosis-related pathways, which were largely absent in CTCs, indicating compartmentalized immune regulation Our findings demonstrate fundamental heterogeneity in inflammatory responses at the FMi, with maternal cells exhibiting greater sensitivity to endotoxin-induced activation compared with fetal CTCs. These differential responses may protect the fetus from excessive inflammation. Understanding cell-specific responses provides a foundation for understanding tolerance and mediators of intolerance at the FMi and identifying potential targets for therapeutic strategies in inflammation-associated pregnancy complications.

Introduction

The fetomaternal interface (FMi) is a uniquely regulated immunological environment where maternal decidual tissues and fetal-derived trophoblasts coordinate to support a successful pregnancy.^1–3 This interface must balance tolerance to the semi-allogeneic fetus with the capacity to mount effective immune responses against invading pathogens. Disruption of this balance, particularly through infection and non–infection-induced inflammation, has been closely linked to adverse outcomes such as preterm birth, fetal growth restriction, and pregnancy loss.^4–7 Ascending bacterial infections targeting the decidua and fetal membranes are a common pathological trigger, activating robust inflammatory cascades that compromise fetal and maternal tissue integrity.^8,9

Ascending bacterial infection is a major factor associated with adverse pregnancy outcomes. The presence of lipopolysaccharide (LPS) in the amniotic fluid and fetal inflammatory response, even in the absence of a culturable infectious agent, has been reported.^10–12 A key molecular mediator of infection-associated inflammation at the FMi is LPS, a potent Toll-like receptor 4 (TLR4) agonist derived from Gram-negative bacteria. LPS stimulation leads to rapid activation of nuclear factor κB (NF-κB) signaling, inflammasome assembly, and the release of proinflammatory cytokines such as interleukin (IL)-1β, IL-6, and tumor necrosis factor α (TNF-α), hallmarks of labor-associated and pathological inflammation.^13–17 Although the signaling pathways activated by LPS are well characterized, how different cell populations at the FMi contribute to the overall inflammatory response remains incompletely understood.

The immune defense mechanism of pyroptosis through gasdermin D (GSDMD) and caspase activation has become essential for protecting tissues from injury during pregnancy-associated inflammation. Research studies have shown that inflammasome activation through pyroptotic signaling (NLRP3-CASP1-IL1β axis) leads to preterm birth and damages fetal membrane structures.^18,19 Given this evidence, we sought to determine whether LPS-induced inflammatory responses at the FMi engage pyroptotic pathways differentially between decidual stromal cells (DECs) and chorionic trophoblasts.

The trophectoderm gives rise to placental and chorionic trophoblasts that function as separate fetal compartments. The villous placenta develops from placental trophoblast cells that include cyto- and syncytiotrophoblasts to perform nutrient exchange and maternal-fetal gas transfer. The chorionic trophoblast cells (CTCs) make up the trophoblast layer of the chorion laeve, which exists as the fetal portion of extraplacental membranes that touch the maternal decidua. The placental trophoblast cells differ from CTCs because they lack invasive properties and produce HLA-G to prevent immune rejection and progesterone and immunomodulatory cytokines to sustain fetal membrane stability. Research by Choudhury et al²⁰ and Lintao et al²¹ shows that CTCs maintain separate gene expression patterns and operational profiles than placental trophoblasts because they show reduced inflammatory reactions and stronger protective mechanisms for barriers. Therefore, throughout this study, CTCs specifically refers to trophoblasts derived from the chorion layer of the fetal membrane, distinct from placental trophoblasts of the villous placenta. The maternal decidua, composed of stromal and immune cells, is a dynamic immune tissue capable of both promoting inflammation and maintaining tolerance. DECs play key roles in modulating immune responses through cytokine and prostaglandin production, leukocyte recruitment, and trophoblast interaction.^22–24 CTCs, which form the outermost layer of the fetal membranes, are classically viewed as structural components.^20,25–27 However, they also express innate immune receptors and secrete immunomodulatory molecules, suggesting potential for active participation in infection responses.^{21,24,26,28–33} It has been hypothesized that DECs and CTCs have evolved complementary immune roles: maternal cells act as the primary immune sensors and amplifiers of inflammation and fetal cells function to preserve barrier integrity and regulate downstream responses. Although individual responses of DECs and CTCs to inflammatory triggers have been studied, a direct, systematic comparison of their responses to endotoxin challenge has not been performed. Such comparisons are critical to understanding how compartment-specific inflammatory programs may shape overall outcomes at the FMi during infection.

In this study, we used primary human DECs, and CTCs isolated from term fetal membranes and stimulated them with LPS to compare their transcriptional and functional inflammatory responses. Through transcriptomic profiling, cytokine quantification, and regulatory network analysis, we show that DECs mount a robust, canonical inflammatory response, whereas CTCs exhibit a markedly attenuated and heterogeneous profile. These findings highlight the cell type–specific heterogeneity in immune responsiveness at the FMi and provide insight into how localized inflammatory signals may be differentially interpreted by maternal versus fetal tissues. Understanding this compartmentalization of inflammation may inform the development of targeted therapeutic approaches to manage infection-driven pregnancy complications.

Methods

Tissue collection and cell isolation

Placental DECs and CTCs were obtained from healthy human term pregnancies after elective cesarean delivery (with informed consent and institutional approval). DECs were isolated by gently scraping and enzymatically digesting the maternal decidua basalis from the fetal membranes. CTCs were isolated from the chorion laeve (the trophoblast layer of the fetal membrane) by trypsin/collagenase digestion and density gradient enrichment. Cells were cultured in Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum and antibiotics. Primary cells from each donor were used at passage 0 or 1. For LPS stimulation, DECs and CTCs were plated and grown to ∼80% confluence, then treated with 100 ng/mL ultra-pure LPS (Escherichia coli 0111: B4; InvivoGen) or an equivalent volume of phosphate-buffered saline (PBS) as a vehicle control. After 6 h of stimulation, cells were harvested for RNA extraction using a Qiagen RNeasy kit. Each experiment was performed in biological triplicate for each cell type (cells from 3 different donors).

RNA sequencing and differential expression analysis

RNA samples with high integrity (RNA Integrity Number = 9.9–10) were used to prepare sequencing libraries (Illumina TruSeq Stranded mRNA) following the manufacturer’s protocol. Libraries were sequenced on an Illumina platform to generate 75 bp paired-end reads. Raw sequencing reads were quality-checked (FastQC) and aligned to the human reference genome (GRCh38) using the STAR aligner. Gene-level counts were quantified with featureCounts. Differential gene expression analysis was performed using Ingenuity Pathway analysis (Qiagen) as well as Omics Playground software which uses DESeq2, Limma and other R-based packages. Genes with an adjusted P value < 0.05 (Benjamini-Hochberg correction) were considered significantly differentially expressed. We applied a log₂ fold-change threshold of 1 (2-fold linear change) to focus on biologically meaningful expression changes. Principal component analysis and Uniform Manifold Approximation and Projection (UMAP) were conducted on variance-stabilized expression values to visualize sample-level transcriptomic differences. All sequencing data have been deposited in the Gene Expression Omnibus under accession number GSE316086.

Pathway enrichment and module analysis

To interpret the gene expression changes, we carried out functional enrichment analyses. Up- and downregulated gene lists from each cell type were analyzed for overrepresentation of Gene Ontology (GO) terms, Kyoto Encyclopedia of Genes and Genomes pathways, Reactome pathways, and WikiPathways using the Omics playground analysis differential expression package. Significantly enriched pathways (false discovery rate q < 0.05) were further visualized using Enrichment Map using Omics playground enrichment map module, which groups related pathways by shared gene members. We built a network in which nodes represent enriched GO terms (biological processes) and edges represent gene overlap between terms; an unsupervised community detection algorithm (Louvain method) was then applied to identify clusters of highly related terms. These clusters correspond to major functional modules in the data. For select pathways of interest (e.g. inflammasome/pyroptosis, PPAR signaling), we consulted curated pathway databases (like Ingenuity Pathway Analysis [IPA]) to map the expression changes onto known signaling diagrams.

Network inference (prize-collecting Steiner forest) analysis

We employed the prize-collecting Steiner forest algorithm to reconstruct putative signaling networks that connect the LPS receptor to downstream differentially expressed genes in DECs and CTCs. Each significantly upregulated gene was assigned a “prize” proportional to its magnitude of change. We used the Omics Integrator implementation of prize-collecting Steiner forest algorithm with a curated human protein–protein interaction network using string database as the interactome background. TLR4 was set as the root node of the network, and the algorithm was run with parameters tuned to favor inclusion of high-prize nodes and a small number of connecting nodes (Steiner nodes). The output network was a set of interacting proteins forming a forest that links TLR4 to the DE gene products. Driver genes appeared as hubs in the network computed using a PageRank either by the centrality or by the fold-change measure. We interpreted these networks to identify central hubs (nodes with high connectivity or that appeared in multiple optimal forests across different parameter runs). The Kamada-Kawai algorithm was used for graph layout. It is based on a physical spring model that produces a static, force-directed arrangement of nodes. It provided the hierarchical layout that arranged nodes in a tree-like structure. The DECs and CTC networks were analyzed for key driver differences and visualized Omics playground network viewer. This analysis highlighted candidate master regulators in each cell type’s LPS response.

Data and statistical analysis

Unless otherwise specified, all data analyses were performed using BigOmics software backed up by R (4.1; R Foundation for Statistical Computing) or an earlier version, and the graphs were generated. Heatmaps were produced with the cluster package from Omics playground package. For the cytokine and chemokine gene panels, normalized read counts were transformed to z scores for visualization across samples. Pearson correlation was used to compare fold-change profiles between DECs and CTCs. Throughout, n refers to independent biological replicates. Statistical significance for gene expression differences was derived from the DESeq2 differential expression model (Wald test), or edge R(QLF) and limma (trend) and adjusted P values are reported. Additional details on experimental procedures and analysis scripts are available upon request.

Results

Global transcriptomic landscape reveals distinct inflammatory signatures in maternal and fetal cells upon endotoxin challenge

To assess the differential transcriptional responses at the FMi following endotoxin exposure, we performed transcriptomic profiling of DECs and CTCs stimulated with LPS. UMAP analysis revealed a clear clustering distinction between LPS-treated and PBS-treated groups in both DECs and CTC populations, indicating cell type–specific transcriptomic alterations in response to endotoxin stimulation (Fig. 1A, B). Notably, the separation along UMAP dimensions 1 and 2 suggests high intergroup variance with minimal overlap between conditions.

Figure 1.

Differential transcriptomic responses to LPS stimulation in DECs and CTCs. (A) UMAP plots of differentially expressed genes in DECs (left) and CTCs (right), showing distribution and density of LPS-responsive genes across cell populations. (B) Principal component analysis 3D plots demonstrating sample clustering in DECs (left) and CTCs (right) under control and LPS-treated conditions. (C, D) Volcano plots of differentially expressed genes in DECs (A) and CTCs (B) following LPS treatment as determined by 4 statistical pipelines: DESeq2, edgeR, trend-limma, and voom-limma. Upregulated genes are marked in red and downregulated genes in blue.

Principal component analysis further confirmed this separation, with LPS-treated samples showing marked divergence from their respective controls across the first 3 principal components (Fig. 1C). This demonstrates that LPS induces robust and consistent transcriptomic shifts in both maternal and fetal compartments.

To comprehensively map differentially expressed genes, we applied 4 statistical methods DESeq2, edgeR, limma-trend, and voom-limma to both cell types. DEC samples exhibited a pronounced upregulation of interferon-stimulated genes such as IFI6, MX2, OAS2, and CXCL8 across all pipelines (Fig. 1E, left), reflecting a robust activation of innate immune signaling. Notably, genes including TNFRSF10D and ZMAT3 were consistently downregulated. In contrast, CTCs displayed a distinct differentially expressed gene profile characterized by upregulation of MYH16, RBMS, PLXND1, and ARX, alongside downregulation of TSPAN6 and GCLC, pointing toward divergent regulatory programs in fetal-derived cells (Fig. 1F, right).

Functional module analysis reveals divergent pathway activation in maternal and fetal cells

Network-based enrichment was used to delineate the biological functions of LPS induced transcriptomic alterations for functional enrichment and pathway module analyses. The SystemsBio module of Omics playground identified highly interconnected modules in DECs enriched for interferon signaling, chemokine-mediated inflammation, and antigen processing (Fig. 2A). In contrast, CTCs exhibited distinct modules involving oxidative phosphorylation, transmembrane transporter activity, and stress-related apoptotic pathways (Fig. 2B), suggesting a metabolically biased and stress-adaptive response.

Figure 2.

Systems-level analysis reveals divergent inflammatory network responses and pathway activation in DECs and CTCs upon LPS stimulation. (A, B) Enrichment maps showing GO pathway clustering of differentially expressed genes in DECs (A) and CTCs (B) following LPS exposure. Nodes represent enriched GO terms and edges indicate shared genes. DEC networks are dominated by immune activation (e.g. interferon signaling, chemokine activity), while CTC networks are characterized by stress response, oxidoreductase activity, and transporter function. Hierarchical clustering reveals distinct and coordinated transcriptional response to endotoxin, characterized by strong induction of inflammatory mediators and immune-response genes and suppression of homeostatic pathways. (C) Heatmap of top differentially expressed genes in DECs clustered by expression pattern under PBS and LPS conditions. (D) Correlation plots showing enrichment of hallmark pathways across 6 signature clusters (S1–S6) in DECs under PBS and LPS treatments. Each pathway is color-coded by cluster and ranked by enrichment score (R). Cluster-specific patterns indicate coordinated activation of inflammatory, interferon, and cell death–related pathways alongside modulation of metabolic and structural programs, highlighting a pronounced maternal inflammatory response. (E) Heatmap of differentially expressed genes in CTCs under PBS and LPS conditions.Rows represent genes and columns represent samples, with hierarchical clustering that demonstrates a more selective transcriptional response relative to DECs, with induction of stress-associated genes and limited activation of canonical inflammatory pathways. (F) Corresponding enrichment correlation plots for hallmark pathway activation in CTCs across 6 distinct clusters (S1–S6), highlighting differences in metabolic, structural, and immune-related signaling compared with CTCs. Each cluster displays coordinated activation or suppression of biological pathways, revealing heterogeneity in metabolic, structural, and immune signaling responses to LPS and highlighting compartment specific regulatory programs in CTCs.

Unsupervised hierarchical clustering coupled with cell cycle phase annotation revealed robust segregation of cell states in both DECs and CTC populations. In DECs, 5 transcriptionally distinct clusters emerged, with cluster C1 predominantly enriched for proinflammatory and interferon-related transcripts, particularly in the LPS condition (Fig. 2C). Gene set enrichment correlation plots showed strong positive association of cluster C1 with inflammatory signaling pathways such as IL6-JAK-STAT3, TNF-α via NF-κB, and interferon γ response, while control clusters aligned with basal metabolic programs (Fig. 2D).

Six clusters were identified in CTCs, with C1 and C2 showing moderate activation under LPS treatment. These clusters were not enriched for classical inflammatory genes but aligned with stress-responsive and metabolic signaling signatures, such as unfolded protein response, MYC targets, and oxidative phosphorylation (Fig. 2E). Pathway correlation further revealed CTC-specific enrichment in metabolic adaptation, unfolded protein response, and estrogen signaling pathways (Fig. 2F), distinct from the inflammation-driven programs in DECs.

DECs exhibit robust IL-mediated inflammatory responses to endotoxin compared with CTCs

To further validate the transcriptomic findings, we assessed inflammatory cytokine production and IL gene expression in DECs and CTCs following LPS stimulation. Multiplex cytokine quantification revealed that DECs mounted a robust inflammatory response, with significantly elevated secretion of IL-6, IL-1β, and TNF-α compared with PBS controls (Fig. 3A). In contrast, CTCs exhibited only minimal cytokine secretion upon LPS exposure, indicating an attenuated inflammatory output. Several machine learning algorithm (LASSO, elastic nets, random forests and extreme gradient boosting)–based analysis using BigOmics identified predicted IL biomarkers from the transcriptomic dataset, summarized as heatmaps (Fig. 3B, D). These predicted biomarkers showed strong enrichment of IL-6, IL-1β, IL-23A, IL-24, IL-32, and IL-33 in DECs under LPS challenge, consistent with a robust cytokine-driven response. Corresponding boxplots (Fig. 3C) illustrate significantly elevated expression of these proinflammatory mediators. Proteomic validation further supported these predictions, demonstrating concordant increases in IL-6– and IL-8–associated pathways within the DEC proteome. Conversely, CTCs displayed selective and limited transcriptional activation, with modest upregulation of IL-32, IL-17RB, and IL-1R1, whereas most other IL transcripts were unchanged or downregulated (Fig. 3E). These results underscore a cell type–specific divergence in inflammatory potential, in which maternal decidual cells demonstrate a heightened and coordinated IL-driven response to endotoxin. At the same time, chorion trophoblasts maintain a more restrained expression profile.

Figure 3.

Differential cytokine and interleukin expression profiles in DECs and CTCs under LPS-induced inflammatory conditions. (A) Bar plots comparing IL-6, IL-8, and TNF-α levels across control (PBS) and LPS-treated DEC and CTC conditions, showing heightened cytokine release in DECs. (B, C) Heatmap (B) and boxplots (C) of significantly differentially expressed IL genes in DECs under LPS versus PBS treatment. Expression values are shown as normalized z scores. The top upregulated and downregulated ILs are highlighted. (D, E) Heatmap (D) and corresponding boxplots (E) of IL expression profiles in CTCs under LPS and PBS conditions. Compared with DECs, CTCs display a muted and heterogeneous IL response, with select ILs (e.g. IL1A, IL32) upregulated and others unaffected. Statistical significance was assessed using 1-way analysis of variance or Student’s t test. Data are presented as mean ± SEM. Significance is indicated as ***P < 0.001.

NF-κB pathway activation diverges between maternal and fetal cells at the FMi

To identify central regulators of the inflammatory response, we performed upstream pathway analysis using IPA interpret on LPS-stimulated DECs and CTCs. In DECs, the network topology revealed STAT1, IRF7, and TNF as major hubs activating downstream targets including IL1β, IL6, CSF2, and IFNAR2. This network was tightly linked to functional categories such as leukocyte activation, type I interferon signaling, and antimicrobial responses (Fig. 4A). In contrast, CTCs demonstrated a distinct hub-centric response driven by RELA (NF-κB p65), MYD88, and CD40L, with connections to IL1A, IL1B, TNF, and TLR family members (TLR2, TLR4, TLR9), suggesting a more NF-κB–centric axis focused on adhesion and immune modulation (Fig. 4B). These pathway-level differences suggest differential upstream sensing and integration of endotoxin signaling across cell types.

Figure 4.

Differential NF-κB signaling pathway activation in DECs and CTCs upon LPS stimulation. (A, B) IPA network maps illustrating predicted upstream regulators and downstream functional annotations of LPS-induced genes in DECs (A) and CTCs (B). Orange nodes indicate activated genes/pathways; blue nodes represent predicted inhibited targets. (C, D) Western blot images assessing NF-κB (p65/RELA) protein levels in DECs (C) and CTCs (D) under PBS and LPS treatment. Protein bands confirm activation patterns consistent with transcriptomic data. (E, F) UMAP visualization of key inflammatory and NF-κB–associated gene expression across LPS-treated DECs (E) and CTCs (F). Red and blue gradients indicate gene expression upregulation and downregulation, respectively. Statistical evaluations were conducted via differential expression analysis. Western blots are representative of 3 biological replicates.

While Western blot analysis confirmed NF-κB activation (p65 phosphorylation) in both DECs and CTCs (Fig. 4C, D), gene expression–based UMAP visualization further confirmed cell type–specific distribution of differentially expressed transcripts, including inflammatory and stress-associated genes (e.g. TNF, IL6, IRF7) upon LPS exposure (Fig. 4E), whereas in CTCs, RELA, CD40, IL1A, and downstream targets exhibited spatially distinct expression patterns (Fig. 4F). In response to 48-hr LPS stimulation, phosphorylated STAT3 levels were not significantly altered in CTCs. However, DECs demonstrated a significant increase in pSTAT3/total STAT3 ratio following LPS exposure, indicating cell type–specific activation of STAT3 signaling (Fig. S2). Collectively, these data suggest that while both cell types activate inflammatory signaling, DECs mount a STAT1/IRF7-driven antiviral response. In contrast, CTCs rely more on NF–κB–mediated adhesion and cell communication signaling.

Chemokine and cytokine profiles reveal cell type–specific inflammatory signatures in response to endotoxin

Chemokine and cytokine signaling networks are modulated at the FMi in response to inflammation. Volcano plots revealed that DECs displayed a robust upregulation of multiple chemokines, including CXCL1, CXCL2, CXCL3, CXCL8, CCL2, and IL6, while only a few genes were significantly downregulated (Fig. 5A). In contrast, CTCs exhibited a more restricted transcriptional response, with modest upregulation of CXCL10, TNFSF10, and CCL20 and minimal downregulation (Fig. 5B).

Figure 5.

Distinct chemokine and cytokine gene expression patterns in DECs and CTCs in response to LPS stimulation. (A, B) Volcano plots showing differentially expressed genes related to chemokine signaling in DECs (A) and CTCs (B) after LPS treatment compared with PBS controls. Significantly upregulated (red) and downregulated (blue) genes are annotated. (C) Heatmap illustrating chemokine gene expression profiles in DECs under LPS and PBS conditions, showing robust upregulation of multiple CCL and CXCL family members upon LPS stimulation. (D) Boxplots highlighting significantly induced chemokine genes in DECs, including CXCL8, CXCL1, CXCL4, CXCL5, CXCL2, CCL10, CCL5, and CCL20. (E) Heatmap showing chemokine expression profiles in CTCs under LPS and PBS conditions, indicating an overall attenuated response with selective gene induction. (F) Boxplots depicting key differentially expressed chemokines and cytokines in CTCs, including CXCL1, ATMACC3, CCL20, CCR2, CXCR4, CCL12, CCR5, and CXCL11. Gene expression data were obtained from transcriptomic analysis of 3 biological replicates per group. Statistical significance was determined using DESeq2 with adjusted P < 0.05.

Heatmaps demonstrated a broad induction of chemokine ligands and their receptors in DECs exposed to LPS (Fig. 5C). Genes such as CXCL1, CXCL2, CXCL3, CXCL5, CXCL6, CXCL8, and CCL20 were highly upregulated compared with PBS controls. Boxplots confirmed significant increases in these transcripts, supporting a robust inflammatory cascade in maternal cells (Fig. 5D). Notably, CXCL1, CXCL5, and CCL2 showed the highest fold change.

In CTCs, LPS stimulation induced more moderate changes in chemokine gene expression (Fig. 5E). Although some genes, including CXCL10, CXCL11, and CCL20, were upregulated, the magnitude and breadth of response were more limited than in DECs. Expression of VCAM1, CCR5, and CXCR4 also increased, suggesting activation of specific migratory and immune recruitment pathways. Boxplot visualization confirmed statistically significant but relatively modest changes in these key chemokines and receptors (Fig. 5F).

TNF and IL gene families exhibit distinct inflammatory signatures in DECs and CTCs

To further dissect the inflammatory signaling landscape at the FMi, we analyzed differential expressions of TNF and IL gene families. Volcano plot analysis and log2 fold change ranking in DECs revealed a strong induction of multiple IL family members, including IL6, IL1B, IL1A, IL32, IL27, and IL17 (Fig. 6A). Notably, IL6 showed the highest fold change and significance, confirming its pivotal role in driving maternal inflammation. In contrast, the CTCs exhibited only modest downregulation of a few IL-related transcripts (e.g. IL1R1, IL6, and IL32) with minimal statistical impact (Fig. 6B).

Figure 6.

Divergent cytokine and TNF-superfamily responses to endotoxin in maternal and fetal cell types. (A, B) Volcano plots showing differentially expressed interleukins in DECs (A) and CTCs (B) following LPS stimulation. Annotated genes indicate top significantly upregulated cytokines. (C, D) Summary tables listing IL family members differentially expressed in DECs (C) and CTCs (D) with log₂ fold change (log2FC) and statistical significance (adjusted P values; **P < 0.01, ***P < 0.001). (E) Boxplots depicting key LPS-induced TNF superfamily genes in DECs, including TNFAIP3, TNFRSF1B, TNFSF10, and TNFRSF25. (F) Heatmap showing global TNF superfamily gene expression in DECs under LPS and PBS conditions. (G) Boxplots illustrating LPS-induced TNF pathway genes in CTCs, such as TNFRSF1A, TNFRSF21, TNFSF10, and TNFRSF25. (H) Heatmap visualizing global expression of TNF superfamily genes in CTCs with LPS versus PBS exposure, indicating a more selective activation pattern compared with DECs. Transcriptomic profiling was performed using RNA sequencing on primary DEC and CTC samples. Data represent normalized gene expression values across 3 biological replicates per group.

Analysis of TNF family genes in DECs revealed significant upregulation of TNFSF10 (TRAIL), TNFSF18, and TNFRSF9, alongside several TNF receptors such as TNFRSF1B, TNFRSF4, and TNFRSF8 (Fig. 6C, D). CTCs also exhibited differential expressions in select TNF superfamily genes, including TNFRSF10C, TNFRSF10A, and TNFSF13B, though with generally lower fold changes. Heatmap comparison showed that while both DECs and CTCs respond to LPS by activating TNF-related transcripts, DECs display a broader and more pronounced pattern (Fig. 6E, F).

Distinct GO and pathway enrichment profiles reveal divergent immune and metabolic programming in DECs and CTCs

To uncover the functional implications of transcriptional changes induced by LPS at the FMi, we conducted GO and pathway enrichment analyses in DECs and CTCs.

GO enrichment

In LPS-treated DECs, GO biological process analysis revealed significant enrichment of immune-related pathways, including type I interferon-mediated signaling, interferon β production, neutrophil chemotaxis, T cell activation, and antigen presentation via major histocompatibility complex class II (Fig. 7A). These findings highlight robust activation of antiviral and inflammatory mechanisms in maternal tissues. Conversely, CTCs showed enrichment in cellular metabolism, steroid biosynthesis, RNA methyltransferase activity, and amino acid transmembrane transport (Fig. 7B), indicating a shift toward metabolic reprogramming and cellular maintenance rather than active inflammation.

Figure 7.

Pathway enrichment analysis reveals distinct biological processes activated by LPS in maternal and fetal compartments. (A, B) GO enrichment heatmaps showing upregulated pathways in DECs (A) and CTCs (B) following LPS stimulation versus PBS. DECs displayed robust activation of immune response and interferon pathways, while CTCs showed enrichment of metabolic and epigenetic regulatory processes. (C, D) GO tree graphs visualize hierarchical clustering of enriched biological processes in DECs (C) and CTCs (D). Color gradient reflects significance and direction of regulation (red = upregulated, blue = downregulated). (E, F) Kyoto Encyclopedia of Genes and Genomes pathway enrichment heatmaps depicting upregulated canonical signaling pathways in DECs (E) and CTCs (F). DECs showed dominant activation of interferon, IL, and viral response pathways, while CTCs was enriched for gluconeogenesis, WNT signaling, and cholesterol metabolism. Pathway analyses were performed on differentially expressed genes using GO biological processes and Kyoto Encyclopedia of Genes and Genomes databases with a false discovery rate threshold <0.05. Results emphasize the compartment-specific molecular programs triggered by inflammatory stimulation at the FMi.

GO tree visualizations

GO term tree diagrams further illustrated the dichotomy: DECs showed upregulation of innate immune defense and antiviral signaling cascades (Fig. 7C), while CTCs displayed functional clustering around RNA regulation, epithelial development, and transmembrane transport processes (Fig. 7D). Notably, immune-related terms were largely absent in the fetal trophoblast tree, reinforcing their relatively restrained immune profile. Pathways in CTCs are indicative of cellular reorganizational activities in response to the treatment and maintaining intracellular homeostasis rather than responding or exaggerating the inflammatory response.

Reactome pathway analysis

Reactome-based enrichment in DECs confirmed strong upregulation of pathways, including interferon α/β signaling, TLR cascade, IL signaling, and NOD2 pathway activation (Fig. 7E). These reflect a comprehensive inflammatory and antigen-presenting profile in response to endotoxin. In contrast, CTCs showed mild enrichment of metabolic and biosynthetic pathways, including glucagon signaling, fatty acid elongation, and WNT ligand biogenesis (Fig. 7F), indicating preferential investment in structural and developmental functions.

Pathway mapping highlights distinct STAT signaling polarization between CTCs and DECs in response to LPS

To further elucidate signaling differences underlying the divergent inflammatory responses observed between DECs and CTCs, we investigated pathway enrichment using IPA focused on the highly enriched JAK-STAT signaling cascade (Kyoto Encyclopedia of Genes and Genomes: hsa04630).

DECs exhibited a pronounced proinflammatory profile. Upon LPS exposure, significant upregulation was observed in canonical inflammatory mediators, including IL6, TNF, STAT1, and STAT2, along with several downstream effectors such as CXCL9, MAPK8, and IFIT1 (Fig. 8A). Enhanced STAT1/2 signaling and MAPK activation reflect a robust type I interferon and NF-κB–driven immune response in the maternal compartment.

Figure 8.

IPA pathway analysis of LPS-induced gene expression in DECs and CTCs reveals divergent regulation of JAK-STAT and inflammatory signaling networks. Pathway maps display differentially expressed genes between LPS and PBS treatments in DECs (top) and CTCs (bottom), overlaid on the JAK-STAT signaling pathway (Kyoto Encyclopedia of Genes and Genomes: hsa04630). Red boxes indicate upregulated genes; blue boxes indicate downregulated genes; green diamonds denote transcription factors or pathway regulators. In CTCs, LPS stimulation leads to dominant activation of STAT3/IL-10 signaling (anti-inflammatory), while in DECs, upregulation of IL6, TNF, and STAT1/STAT2 suggests a robust proinflammatory and interferon-mediated cascade.

Conversely, in CTCs, LPS treatment primarily activated anti-inflammatory signaling nodes, notably STAT3, IL10, and suppressors of cytokine signaling (SOCS3) (Fig. 8B). Downstream targets of STAT3, including genes involved in immune suppression and epithelial repair (e.g. BCL3, CDKN1A, SOCS1, MRAS), were upregulated, while proinflammatory intermediates such as IL6, TNF, and STAT1 remained largely unaltered or downregulated. This suggests a tightly regulated or dampened inflammatory response in fetal-derived cells, possibly to maintain immune tolerance at the FMi.

Pyroptosis pathway analysis reveals cell type–specific inflammasome activation in response to LPS

To further dissect the mechanisms of LPS-induced cell death and inflammation at the FMi, we analyzed the pyroptosis signaling pathway using IPA.

DECs demonstrated broader pathway engagement. In addition to upregulated IL1B, IL18, and GSDMD, maternal cells exhibited marked induction of upstream mediators including TLR, CASP1, IRF2, and FOXO3 (Fig. 9A). This implies that DECs not only activate canonical NLRP3 inflammasome signaling, but also engage alternative pyroptotic routes through TLR activation and transcriptional priming. The heightened inflammatory output in maternal cells may reflect their sentinel role in detecting infection and initiating immune responses.

Figure 9.

IPA pathway mapping of pyroptosis signaling in maternal and fetal cells in response to LPS stimulation. Pathway maps represent gene expression changes in DECs and CTCs treated with LPS versus PBS, overlaid on the pyroptosis pathway. Red boxes indicate upregulated genes, green boxes indicate downregulated genes, and blue outlines highlight key regulatory nodes. DECs exhibit broader activation of upstream TLR signaling and caspase-1 pathways, with upregulation of CASP1, GSDMD, and IRF2, suggesting a more inflammatory pyroptotic cascade. In contrast, in CTCs, significant upregulation of IL1B, IL18, NLRP3, and GSDMD reflects inflammasome activation and canonical pyroptosis.

In contrast, in CTCs, LPS treatment resulted in selective activation of key inflammasome components, including NLRP3, GSDMD, IL1B, and IL18 and highly expressed IL1A (Fig. 9B) which was not observed in DECs. Notably, these events were coordinated with mild upregulation of CASP1, but limited engagement of upstream pattern recognition receptors (PRRs) or gasdermin-independent mechanisms. This suggests that fetal-derived cells undergo a controlled, inflammasome-driven pyroptosis, potentially acting as an immune signal without triggering extensive cytotoxicity.

Discussion

The chorio-decidual FMi plays a pivotal role in maintaining pregnancy, orchestrating parturition, and contributing to the pathobiological mechanisms underlying preterm birth. This interface supports pregnancy primarily through immune tolerance while also serving as a physical and mechanical barrier. Within the vascular decidua parietalis, stromal cells act as key amplifiers of the inflammatory response, owing to their intrinsic immunological activity and the high density of immune cells (∼35%), particularly enriched with natural killer (NK) cells. In contrast, chorion trophoblasts—despite lining up next to this immunologically active decidual layer and lacking intervening matrix or structural barriers remain largely refractory to these immune stimulators, highlighting a unique aspect of immune privilege at the FMi. While inflammation is a key driver of both term and preterm parturition, the responses of fetal membrane interface (FMi) cells to endogenous or exogenous stimuli remain poorly understood. The observed refractoriness of chorionic trophoblast cells (CTCs) to inflammatory insults²¹ led us to hypothesize that FMi and CTCs exhibit distinct immunologic responses, potentially due to cellular heterogeneity. This study provides a comprehensive transcriptomic dissection of maternal DECs and fetal CTCs in response to LPS-induced inflammation, revealing critical insights into the cell type–specific inflammatory architecture at the FMi. Our data indicate that CTCs do not simply mirror the inflammatory escalation seen in DECs; instead, they appear to enact an immunoregulatory response that may protect the fetus. DECs consistently demonstrated upregulation of canonical proinflammatory genes, supporting the previous report.^34–36 Conversely, CTCs displayed limited induction of classical inflammatory mediators involved in tissue remodeling, redox regulation, and structural maintenance. These patterns suggest that CTCs, rather than amplifying inflammation, may shift toward adaptive responses that favor tissue protection and homeostasis. This implies that trophoblast cells contribute to immune tolerance and minimize fetal exposure to inflammatory damage, supporting their specialized role within the FMi.^37–40

In DECs, LPS exposure precipitated a strong activation of immune-related gene sets, including those associated with interferon signaling, antigen presentation, and inflammatory apoptosis, highlighting a shift toward an immunologically active and potentially cytotoxic phenotype.^35,41,42 However, CTCs maintained expression of pathways associated with stress adaptation, protein folding, and metabolic reprogramming, including activation of oxidative stress modulators and estrogen-responsive network elements known to support tissue integrity and immune tolerance under inflammatory stress.^21,26 Notably, clusters enriched in CTCs showed signatures linked to steroid biosynthesis and glutathione metabolism, aligning with their known capacity to contribute to local hormone (P4) production and redox buffering at the FMi.^43,44 These molecular distinctions emphasize the unique, compensatory role of these fetal-derived cells adopt under inflammatory stress, contrasting the immune-alert phenotype in DECs. This aligns with growing evidence that fetal trophoblasts actively contribute to maintaining fetal tolerance and shield and blunt an inflammatory response to minimize the impact of maternal inflammation. For instance, trophoblasts express pattern recognition receptors and secrete cytokines that modulate maternal immune responses without promoting excessive inflammation.^37,45–47 This may be coming from trophoblast-derived exosomes that contain immunomodulatory microRNAs that dampen maternal TLR4 signaling, reducing proinflammatory cytokine release.⁴⁸

In DECs, the robust induction of classical proinflammatory cytokines (IL-1β, IL-6, IL-8, and IL-32) is consistent with prior reports describing decidual tissues as immunologically vigilant in sensing and amplifying pathogen-associated molecular patterns. IL-1β and IL-6 in particular are key drivers of labor-associated inflammation,^49,50 and their exaggerated expression in DECs aligns with their established roles in promoting prostaglandin synthesis, labor initiation, and leukocyte infiltration.^51,52 The elevated IL-32 expression further underscores a sustained NF-κB activation state in gestational tissues.^51,53 In contrast, CTCs demonstrated an attenuated inflammatory transcriptomic signature, with relatively lower expression of canonical proinflammatory cytokines and modest elevation in regulatory mediators such as IL-11, IL-24, and IL-20RA. These cytokines have been implicated in promoting epithelial integrity and tissue remodeling while dampening immune hyperactivation, particularly in mucosal and placental tissues.^54–56 Such expression patterns in CTCs parallel prior findings that fetal trophoblasts adopt a tolerogenic and anti-inflammatory posture to preserve fetal allograft survival, while maintaining barrier competence under inflammatory stress.^38,57 These findings reinforce the functional distinction between chorionic trophoblasts, which serve as noninvasive, barrier-protective cells within the fetal membranes, and placental trophoblasts, which are specialized for invasion and nutrient exchange at the villous interface.

Activation of proinflammatory signaling via the JAK/STAT and NF-κB axes evidenced by upregulation of IL6, IL1B, and TNF.^58–60 This aligns with previous reports describing robust STAT3 and NF-κB (RELA/MYD88) activation in maternal decidua during infection-induced preterm birth models.^61–64 In contrast, CTCs exhibited a pronounced activation of anti-inflammatory IL-10 signaling, with limited engagement of NF-κB targets, reflecting their dampened inflammatory state and possible immunoregulatory phenotype.^59,64–66 While IL-10 signaling was present in both cell types, DECs showed parallel activation of proinflammatory feedback loops, particularly via STAT1/3 (IRF7) and MAPK pathways, suggesting a cytokine-driven amplification mechanism, minimizing an anti-inflammatory response.^67–70 CTCs retained elevated expression of SOCS proteins and IL10R components, which are known to suppress TLR-mediated inflammation, indicating a tight regulatory control over their immune responsiveness.^71–74 The chorion layer expresses high levels of IL-10 and TGF-β, cytokines implicated in promoting an anti-inflammatory microenvironment at the FMi.^75,76

DECs upon LPS exposure, including upregulation of NLRP3, CASP1, and IL1B, consistent with classical pyroptotic pathways previously reported in decidual inflammation models.^18,19,77 The marked upregulation of GSDMD and ASC (apoptosis-associated speck-like protein) in DECs suggests active execution of pyroptosis, corroborating earlier studies linking inflammasome activation to preterm labor pathophysiology.^78,79 Conversely, CTCs showed limited induction of CASP1 and restricted IL1B and IL18 expression, aligning with reports that fetal-derived trophoblasts maintain an immune-tolerant state to protect fetal integrity.^79,80 This indicates a decoupling of TLR-mediated signaling from pyroptotic execution, possibly via differential regulation of miR-155 or oxidative stress pathways.^81,82The downregulation of proapoptotic mediators such as BAX and CASP3 in CTCs further supports a survival-oriented cellular response, which has been previously characterized in trophoblast immunobiology.⁸³ These observations reflect cell type–specific divergence in programmed inflammatory cell death pathways that may determine susceptibility or resistance to intrauterine inflammation. Although pyroptosis was not the primary hypothesis at study inception, the consistent upregulation of canonical inflammasome components (NLRP3, CASP1, IL1B, GSDMD) in decidual cells indicated that this pathway might be actively engaged in maternal compartments following LPS exposure. However, in fetal CTCs, pyroptosis-related gene induction was limited, suggesting a restrained inflammasome activation that aligns with their immune-tolerant phenotype. Therefore, we propose pyroptosis as a key secondary mechanism amplifying inflammation in maternal tissues but attenuated in fetal cells to preserve barrier integrity.

As we reported previously, the chorion layer of the fetal membrane functions as a “great wall” at the FMi. Our data demonstrate immunological protective roles, whereas the inflammatory responses are blunted in the chorion compared with their amplification in the decidua. In addition to the immunological inertness exhibited by the chorion, it is one of the sources of progesterone at this FMi. Chorion trophoblasts, unlike placental trophoblasts, also express HLA-G, which negates the invasion of the activated NK cells from the decidua, all of which strengthens the function of the great wall barrier. However, it is to be noted that endotoxin and some microbes can still invade the chorion layer without a localized immune activation and reach the amnion or amniotic fluid. Unfortunately, this invasion is detrimental as the amnion epithelial layer is highly vulnerable and can trigger fetal inflammatory response. This feedback response from the companion layer in the fetal membranes can weaken the defense of the chorion trophoblast. Hence, it is the fetal inflammatory response from the fetus and amnion layers, and not the maternal inflammatory response, that is more detrimental to the pregnancy. Stopping this response is critical in preventing adverse outcomes.

Conclusion

In summary, this study provides comprehensive evidence demonstrating the cellular and molecular heterogeneity of maternal and fetal components of the fetal membranes in response to endotoxin-induced inflammation. We revealed that DECs mount a classical proinflammatory response, whereas CTCs adopt a tightly regulated immune posture, favoring anti-inflammatory and homeostatic pathways. Our findings highlight the dynamic, nonredundant contributions of fetal membranes to intrauterine immune regulation and offer a novel perspective on how cellular compartmentalization may influence the pathophysiology of pregnancy-related disorders such as preterm birth. This mechanistic insight lays the foundation for future translational strategies targeting fetal membrane-derived pathways for therapeutic modulation of inflammation during pregnancy.

Author contributions

A.K.K. and R.M. conceived the project. A.K.K. and M.T. designed experiments. M.T., T.J.T., I.M., and L.R. collected and prepared tissue samples. V.M. performed bioinformatic analyses. V.M., M.T., L.R., A.K.K., and R.M. interpreted the data and wrote the manuscript. All authors contributed to editing the manuscript and approved the final version.

V.M.(Data curation [Equal], Formal analysis [Lead], Software [Lead], Writing—original draft [Equal]), M.T. (Data curation [Equal], Formal analysis [Equal], Investigation [Equal], Writing—review & editing [Equal]), T.J.T. (Data curation [Equal], Formal analysis [Equal]), E.A. (Methodology [Equal], Writing—review & editing [Equal]), I.M. (Data curation [Equal], Formal analysis [Equal], Writing—review & editing [Equal]), L.R. (Formal analysis [Equal], Validation [Equal], Visualization [Equal], Writing—review & editing [Equal]), R.M. (Conceptualization [Equal], Funding acquisition [Equal], Project administration [Equal], Resources [Equal], Supervision [Equal], Writing—original draft [Equal], Writing—review & editing [Equal]), and A.K.K. (Conceptualization [Equal], Data curation [Equal], Formal analysis [Equal], Funding acquisition [Equal], Resources [Equal], Supervision [Equal], Writing—original draft [Lead], Writing—review & editing [Equal])

Supplementary material

Supplementary material is available at The Journal of Immunology online.

Funding

This study is funded by National Institutes of Health/National Center for Advancing Translational Sciences funds to R.M. (1U2CTR004868-01) and Eunice Kennedy Shriver National Institute of Child Health and Human Development A.K.K. (5R01HD113193-03).

Conflicts of interest

The authors declare no conflict of interest.

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author (A.K.K) upon reasonable request.