The onset of labor and/or childbirth prior to term is often preceded by the rupture of extraplacental, gestational (“fetal”) membranes (1–3). Preterm premature rupture of membranes (PPROM) is commonly associated with histological inflammation, a condition termed chorioamnionitis (CAM). Acute, neutrophilic CAM can be caused by infectious or non-infectious stimuli, but a significant infectious cause of CAM is the bacterium Group B Streptococcus (GBS) (4–9). Further consequences of GBS-associated CAM include maternal and fetal sepsis and stillbirth (10–16).

The gestational membranes differentiate along with the trophectoderm during early human pregnancy. Both these tissues contain multiple cell types derived from both the fetus and the mother. The placenta is largely made up of cytotrophoblasts (CTBs), a fetal cell type which differentiates into several distinct forms. The gestational membranes are comprised of an outermost layer, the decidua, dominated by maternal decidual stromal cells (DSCs); a central chorion layer of fetal CTBs abutting the decidua; an inner amnion comprised of fetal mesenchymal cells, and a single layer of fetal amnion epithelial cells (AECs) lining the innermost (fetal-facing) aspect of the membranes. The amnion has ample extracellular matrix and basement membrane for structural stability. Macrophages are the principle resident phagocyte within uninflamed gestational membranes and there are cells of the adaptive immune system there as well (17). The precise cause of membrane rupture during CAM remains to be determined, but it likely involves matrix metalloproteinases (MMPs) produced by structural and immune cells of the membranes, and possibly the bacteria themselves (18–20). Additionally, several pro-inflammatory cytokines such as IL-6, IL-1β, IL-8, and TNF-α have been associated with CAM and preterm birth (21, 22).

The interplay of saturated fatty acids, pregnancy, obesity, and inflammation is complex. Saturated fatty acids are associated with inflammation [reviewed in (23)]. In normal weight individuals consuming a western diet, the monounsaturated fatty acid oleate and the LCSFA palmitate comprise (on average) 31% and 27% of fatty acids in the blood stream, respectively. The composition of these fatty acids are changed during obesity, with palmitate being the most abundant fatty acid in circulation at 33%, representing approximately a 20% increase than in individuals of normal-weight (24, 25). Thus, palmitate and oleate are often used to model the fatty acid component of obesity in vitro. Furthermore, palmitate specifically has been associated with inflammation, adverse pregnancy outcomes, and in several other pregnancy-associated conditions, including gestational diabetes mellitus and preeclampsia (26–30). Pregnancy progressively increases the circulating concentrations of fatty acids to increase energy availability, among other effects (31). While we would expect the serum levels of palmitate to increase during pregnancies complicated with obesity similarly to non-pregnant conditions, the concentration of serum palmitate in obese and normal weight pregnancies is also impacted by fetal sex (32). For instance, the fatty acid profiles of maternal plasma in obese pregnant persons differs from normal weight controls in distinct ways depending on fetal sex: obese pregnancies with female fetuses have higher maternal plasma oleate levels relative to controls, while obese pregnancies with male fetuses have lower levels of oleate than controls (32). There are also differences in unsaturated fatty acids between male and female placentas (33).

An estimated 1 in 4 pregnancy complications can be tied to obesity (34, 35), and this number is expected to rise as obesity rates increase. During non-pregnant settings, individuals with obesity are at an increased risk of rectovaginal colonization with GBS (36–38) and thus at greater risk of GBS-related pregnancy complications including CAM, preterm prelabor rupture of membranes, and fetal sepsis (39, 40). The precise cause of greater GBS colonization in obese pregnant people is unknown; however, the contribution of LCSFAs to inflammation has been well established (41).

In this study, we sought to determine the impact of the LCSFA palmitate on GBS-induced inflammation in different compartments of the gestational membranes. We were particularly interested in the potential for synergy between GBS and palmitate in the induction of proinflammatory cytokines, chemokines, and MMP production throughout the gestational membranes, since these peptides play significant roles in regulating the cellular inflammatory response to infection and the integrity of the extracellular matrix. We found that the DSC and CTB cell types displayed synergy in IL-1β secretion when treated with GBS and palmitate together. In contrast to previous work in the absence of palmitate that showed suppression of macrophage activation by both DSCs and CTBs (42), neither cell type inhibited activation of primary placental macrophages or macrophages of the cell line THP-1 in response to infection when palmitate was present. The AEC layer was largely nonresponsive to GBS and palmitate co-incubation and may actively inhibit the inflammation in the choriodecidua when the membranes are intact. The differences of each structural cell type of the gestational membrane in response to GBS and palmitate suggest complex regulation and that paracrine signaling between layers of the gestational membrane will be necessary.

In this manuscript, we have shown in vitro and ex vivo that distinct layers of the gestational membranes respond differently to the saturated fat palmitate in the presence or absence of the common vaginal commensal bacterium GBS. We demonstrate that palmitate enhances GBS proliferation ( Figure 1 ), which may partially explain the increased complications due to GBS in people with obesity during pregnancy, due simply to the increased proliferation of GBS (52, 53). While independent layers in the choriodecidua respond to GBS and palmitate stimulation with synergy in cytokine and MMP induction, the amnion epithelium does not contribute to this phenotype and is largely unaffected by GBS and palmitate stimulation, both in isolated primary amnion epithelial cells and in gestational membrane punch biopsies, where the amnion has been separated from the choriodecidua. This was not surprising, as the broad dampening of inflammation by the amnion is well-established (54–56). Furthermore, we have shown that macrophages (either cell line or primary placental macrophages) co-cultured with structural cells of the choriodecidua in the presence of palmitate and GBS are not suppressed by the structural cells as can be seen in the absence of palmitate in previous work (42, 50). Different patterns of analyte induction in the presence or absence of GBS-palmitate synergy may be due to signaling and transcription factors controlling the particular analytes in a particular cell type; for instance, while IL-1β is reliably induced synergistically from GBS and palmitate in combination, MMP9 synergy is less reliable. IL-1β is controlled at many levels and processed from a pro-form to an active form, while MMP9 is largely controlled at the transcriptional level with the association of coactivators distinct from IL-1β (57). Synergy between palmitate and GBS stimulation likely depends entirely on the signaling pathways to, and regulation of, each particular analyte. Sensitization deriving from one stimulus then augmenting other stimuli has been shown during pregnancy before (58–62), although this has largely been shown for the interplay between viral stimulation or viral PAMPs and subsequent stimulation.

This synergistic production of cytokines and MMPs is not specific to live GBS, as shown by the ability of a TLR4 agonist (LPS) to induce synergy in cytokine expression when administered with palmitate ( Supplementary Figure S2 ). Because we observed cytokine augmentation when palmitate was combined with live GBS [a gram-positive bacterium) as well as a purified TLR4 agonist (LPS, a gram-negative pathogen-associated molecular pattern (PAMP)], this suggests broad applicability of our findings to different bacterial species and strains, both gram positive and gram negative. Further, it suggests that palmitate-induced alterations to GBS proliferation and metabolism are not the primary cause of increased inflammation in the gestational membranes, since the synergistic inflammation occurs with isolated PAMPs as well. We observed a relative inability of TLR2 agonist PAM3CSK4 to induce cytokines or MMPs on its own or in combination with palmitate when applied to DSCs. This lack of stimulation was recapitulated in CTBs as well ( Supplementary Figure S3 ), with the exception of GM-CSF, where PAM3CSK4 was able to induce GM-CSF alone and synergize with palmitate. However, PAM3CSK4 was able to induce several cytokines alone (IL-1β, CCL5, GM-CSF), but not MMP9, and synergize with palmitate in THP-1 macrophages similar to live GBS and purified LPS ( Supplementary Figure S4 ). Additional experimentation needs to be carried out to address the discrepancy between TLR2 activation and palmitate synergy in structural cells versus immune cells (such as macrophages) to determine the mechanism(s) at play, such as specific pathway activation or suppression, or receptor expression. Additional approaches could include using Pam2Cys or lipoteichotic acid (LTA), which utilize a different heterodimer of TLR2; while PAM3CSK4 utilizes a TLR1-TLR2 heterodimer, LTA and Pam2Cys utilize a TLR2-TLR6 heterodimer. The discrepancy between PAM3CSK4 and GBS-induced activation of cytokine and MMP may have to do with the activation of the TLR1/2 versus TLR2/6 heterodimers (63).

There are many varied fatty acids in the bloodstream, and one limitation of this study is the use of palmitate in isolation, or in combination with a single unsaturated fatty acid (oleate), instead of looking at the effects of a cocktail of saturated and unsaturated fatty acids in varying proportions on the cell types found within the gestational membrane. We acknowledge this shortcoming as a necessary oversimplification for the purposes of in vitro experimentation. Palmitate was chosen as it is the most abundant saturated fatty acid in circulation during obesity and can be the dominant circulating fatty acid. Furthermore, it is the most common saturated fatty acid used for in vitro experiments of obesity in the literature, and as such there is a large body of knowledge surrounding its mechanisms, making it an ideal candidate to study our models presented here in a reductionist manner. Continuing this line of inquiry using distinct mixtures of fatty acids, and the use of our previously-published metabolic cocktail (glucose, insulin, palmitate) (64) will more accurately reflect the milieu seen in vivo. However, defining basic patterns of synergy between infectious stimuli and a fatty acid using a reductionist approach gives us a base from which to determine the relative contributions of different fatty acids and sugars to the inflammatory state of the gestational membranes. The mechanism behind palmitate-induced inflammation remains unclear. There is much discrepancy in published work as to whether palmitate can signal through TLR4 (65–71) or not (72). As such, the mechanism behind synergy of GBS and palmitate is also unclear, but it is likely that palmitate stimulates a pathway or receptor independent of TLR4 signaling because palmitate treatment augments the inflammatory response to TLR4 agonists in macrophages, CTBs, and LPS ( Supplementary Figures S2 - S4 ). This includes the induction of: IL-1β, IL-8, CCL2, CCL5, and G-CSF in DSCs; TNFα, MMP9, and GM-CSF in CTBs; and MMP9 and GM-CSF in macrophages.

The condition of “diabesity”, a combination of type-2 diabetes and obesity, may more accurately reflect the reality of which pregnant individuals experience a greater rate of pregnancy complications than BMI alone. The inability of BMI to assess body fat distribution (abdominal versus distributed) or muscle mass adds heterogeneity into the system. Additionally, metabolic syndrome, a classification that encompasses the risk of developing type-2 diabetes and cardiovascular disease, can be found in individuals with any BMI (73). Access to patient electronic medical records containing bloodwork results such as circulating hemoglobin A1C or lipids, or a waist-to-hip ratio, could more accurately stratify our human gestational membrane punch data to reveal patterns that maternal BMI does not.

We expected to see an increase in inflammatory mediators upon co-stimulation of whole human gestational membrane punches with palmitate and GBS. However, for the analytes we selected for these experiments, we were unable to find consistent patterns or statistical significance between treatments. We had 11 patients and 24 membrane punches for each patient allowing for biological and technical replicates for each patient. As shown in Supplementary Table 1 , the BMI of most patients was between the overweight and morbidly obese categories by BMI. However, 1) there was not a direct relationship between BMI and magnitude of cytokine induction (data not shown), and 2) palmitate and GBS did synergize in IL-1β production from the membrane punches of roughly half the patients, while the other half of patients’ membrane punches showed no synergy or decreased production upon palmitate stimulation. We were similarly unable to show a direct correlation between BMI and these cytokine induction patterns. Sorting by fetal sex, maternal GBS status, ethnicity, or gestational age also did not uncover any correlation between magnitude or direction of cytokine induction. It is very possible that with a larger N value of patient samples, the impact of particular patient variables could be observed. Recent studies have found that male placentas from pregnancies of obese people had significantly increased levels of fetal placental macrophages, which could easily have an impact on inflammatory responses in the gestational membranes (74). The gestational membrane punch biopsies we used, even when separated, likely contained up to 13% immune cells (75). And fetal sex also has an impact on how many and what kind of circulating maternal fatty acids accumulate in the placenta (32, 33), which could add another confounder to data from human gestational membranes: the impact of increased saturated fatty acids (already pro-inflammatory) and unsaturated fatty acids (which may be anti-inflammatory) already present in the membrane punches.

Additionally, a better understanding of the metabolic conditions that drive the molecular underpinnings of infection and inflammation during pregnancy can inform immunological interventions to ameliorate the risk of adverse perinatal outcomes. IL-1β is a tantalizing target because this molecule has been implicated as an important biomarker associated with chorioamnionitis, PPROM, and preterm labor (21). IL-1β is a pro-inflammatory cytokine that perturbs maternal-fetal tolerance (76). Our work demonstrates that palmitate and GBS infection promote IL-1β production, underscoring the role both metabolic dysfunction and infection could play in pregnancy homeostasis. There is an association between absolute levels of IL-1β and pregnancy complications: as levels of IL-1β increase, the likelihood of preterm deliveries increases (77). Accordingly, there is likely biological relevance for the synergy of GBS and palmitate: it is not just presence or absence of IL-1β that can result in pregnancy complications, but the greater the IL-1β levels, the greater the likelihood of pregnancy complications, and thus the combination of palmitate and GBS inducing more IL-1β in than either one alone would predispose pregnant people with both stimuli to a greater likelihood of pregnancy complications.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving humans were approved by Vanderbilt University Institutional Review Board. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

JG: Funding acquisition, Methodology, Resources, Supervision, Writing – original draft, Writing – review & editing. RM: Investigation, Writing – original draft, Writing – review & editing. JL: Investigation, Writing – review & editing. LR: Investigation, Writing – review & editing. KN: Investigation, Writing – review & editing. AG: Formal analysis, Methodology, Writing – review & editing. DA: Funding acquisition, Methodology, Resources, Supervision, Writing – review & editing. DC: Funding acquisition, Resources, Supervision, Writing – review & editing. AE: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Visualization, Writing – original draft, Writing – review & editing.