Maternal‐Fetal Interplay During Mid‐gestation: Inflammatory Biomarker Levels in Maternal Sera, Cord Sera and Placenta

ABSTRACT

Problem

Maternal inflammation is linked to adverse maternal and fetal outcomes, but how maternal inflammation affects fetal inflammation during mid‐gestation is not known.

Method of Study

We measured 20 biomarkers of inflammation, including chemokines, pro‐and anti‐inflammatory cytokines, and immunomodulatory cytokines, in matched maternal and cord sera and placental tissue collected during the second trimester from a diverse cohort (N = 106). We examined trends in biomarkers across the second trimester using a linear trend test, and compared distributions across sociodemographic groups using a Kruskal‐Wallis test. We calculated Spearman correlation coefficients and stratified the correlation matrices by gestational age at sample collection (<20 vs. ≥ 20 gestational weeks).

Results

Most inflammatory biomarkers did not exhibit clear trends with gestational age, except for cord levels of CCL17 and CCL22; both Th2‐associated chemokines significantly increased across the second trimester. There were no consistent patterns in biomarker levels across sociodemographic groups. Overall, correlations among maternal, placental and cord biomarkers were stronger before 20 weeks’ gestation than in later pregnancy. While correlations between maternal and placental CCL17 declined after 20 weeks, the correlation between placental and cord CCL17 increased during this same period.

Conclusions

Maternal and fetal inflammatory responses are more tightly coordinated earlier in the second trimester.

1. Introduction

During pregnancy, adaptation of the maternal immune system allows for the fetus to grow, despite the presence of paternally‐derived antigens that would initiate rejection under other circumstances [1]. Furthermore, fetal tolerance of maternal antigens is also required for successful pregnancy [2, 3]. Our biological understanding of this coordination is growing.

Evidence suggests that on the maternal side, very early pregnancy is associated with a pro‐inflammatory state, which helps the trophoblast invade the uterine endometrium and recruits leukocytes to the maternal‐fetal interface, where they excrete anti‐inflammatory cytokines like IL‐10 and TGF‐$\beta$ and IL‐1 receptor agonist [4]. This change in the cytokine environment affects the balance of T‐cells, with a relative suppression of Th1 and Th17 and promotion of T2 and Tregs [1]. A large number of natural killer (NK) cells arrive in the maternal decidua, at the maternal‐fetal interface, and help promote immune tolerance by reducing Th17‐mediated inflammation [5]. However, as pregnancy progresses, the number of uterine NK cells decreases while the proportions of other cells, such as T cells and immature dendritic cells, increase [6]. In particular, immune profiling across pregnancy showed that NK cells and Tregs begin to decrease starting around week 20 of pregnancy [3, 6]. The timing of these changes affects the likelihood of gestational success and the risk of pregnancy complications [7, 8].

During this time, the fetal immune system is undergoing development. During the first trimester, hematopoietic precursor cells develop in the fetal bone marrow, and immature neutrophils and T cell develop [9]. B lymphocytes mature more rapidly, and by the end of the first trimester they can make IgM [10]. Fetal dendritic and NK cells have also been detected in the first trimester [11].

It is less clear how inflammation is regulated by both the mother and fetus during pregnancy, and how this regulation changes across the course of gestation. In particular, it is not well understood when the fetal immune system develops the capacity to independently regulate inflammation, particularly during mid‐gestation. Since inflammation during pregnancy is linked to adverse outcomes for both the mother and newborn [12, 13], this gap limits our understanding of the pathogenesis of many adverse perinatal and neonatal health outcomes, with origins rooted during pregnancy. Furthermore, there is increasing evidence that there are specific periods of pregnancy, also called critical windows, during which exposure to chemical and non‐chemical stressors is most harmful for fetal development [14, 15]. There is an incomplete understanding, however, of the biologic mechanisms, including dynamics within the maternal‐fetal unit, underlying these differences across gestation. To address this, we quantified levels of inflammatory markers in maternal sera, placental tissue, and cord sera to explore relationships between the mother and fetus across the second trimester, a critical and understudied period of gestation.

2. Material and Methods

Our study population consisted of a cohort of 106 pregnant participants who received elective terminations between gestational week 15–24 at UCSF‐affiliated clinics between 2010 and 2016. Participants were residents of Northern California and the Central Valley, who presented with no known pregnancy complications, were at least 18 years of age, English‐ or Spanish‐speaking, and received a dilation and evacuation (D&E) procedure. Pregnancy dating was established by last menstrual period and confirmed by ultrasound. We collected matched placental tissue, maternal serum, and cord serum, and administered surveys to all consented participants. The Institutional Review Board at UCSF approved the study protocol, and all participants provided informed consent upon enrollment (IRB #10‐02246).

Maternal blood samples were collected one day prior to the D&E procedure. Placentas and cord blood samples were collected on the day of the procedure. Maternal and cord blood samples were allowed to clot at room temperature, centrifuged at 3000 rpm for 10 min at 4°C, and then serum was aliquoted into amber vials. Full‐thickness placental biopsies were collected by removing the amniotic membrane, and dissecting tissue containing floating villi, portions of anchoring villi, and their uterine attachment sites. Biopsies included a section of the basal plate representing the maternal‐fetal interface, including chorionic villi and decidua. Samples were washed in phosphate buffered saline, blotted dry, and flash frozen in liquid nitrogen. All biospecimen samples were stored at ‐80°C. Two cord blood samples were excluded due to insufficient volume.

2.1. Measurement of Inflammatory Markers in Serum

We measured the concentrations (in pg/mL) of 20 biomarkers of inflammation─Interleukin (IL)‐1β, ‐2, ‐4, ‐6, ‐7, ‐8, ‐10, ‐12, ‐28A; C‐C Motif Ligand (CCL)‐2, ‐3, ‐4, ‐17, and ‐22; C‐X‐C Motif Chemokine Ligand (CXCL)‐9 and ‐10; Interferon (IFN)‐α, ‐γ; and Tumor Necrosis Factor (TNF)‐α, and ‐R1─using the custom human premixed multiplex magnetic Luminex assay (Bio‐Techne, Minneapolis, MN). The panel included chemokines, pro‐and anti‐inflammatory cytokines, and immunomodulatory cytokines [16, 17, 18, 19]. We added 25 µL of maternal sera, 50 µL of cord sera, or 25 µL of placental protein extraction to each well of a 96‐well plate and mixed with prepared magnetic microparticle beads. Plates were incubated at room temperature on a shaker at 800 rpm for 2 h, followed by incubation with a biotin‐antibody for 1 h and then with streptavidin‐PE for 30 min. Plates were washed with wash buffer 3 times after each incubation. Following resuspension of the microparticle beads into wash buffer, the plates were read on BioPlex‐200 equipment (Bio‐Rad). Each sample was measured in technical duplicate, with one replicate run on each of two independent assay plates. Standards (6 concentrations) and blanks were included on each plate for quality control. Median fluorescence intensities (MFIs) were acquired for each cytokine and corrected by subtracting the background (blanks). MFIs below the background were set to zero. MFIs were interpolated into concentrations (pg/mL) using the standard curve for each analyte; any MFI above or below the detection range was imputed by dividing the lowest interpolated value by $\sqrt{2}$ or multiplying the highest interpolated value by $\sqrt{2}$, respectively.

2.2. Placental Protein Extraction and Quantification

Frozen placental biopsies stored in glass tubes were thawed on ice, and approximately 25 mg of tissue was transferred to a sterilized microfuge tube. Halt Protease Inhibitor (Thermo Scientific) and Tissue Extraction Reagent I (Invitrogen) were mixed at 1:100, and 10 µL of the mixture was added into the Eppendorf tube per 1 mg of the tissue sample. The samples were homogenized by passing through a 20G needle multiple times, vortexed, and then centrifuged at 10,000 rpm for 5 min to pellet tissue debris. The supernatant was collected, and protein concentrations were quantified using the Pierce BCA Protein Assay Kit (Thermo Scientific) according to the manufacturer's protocol. Protein extracts were normalized to a concentration of 1 µg/µL and stored at ‐80°C for until further analysis.

2.3. Statistical Analysis

We log‐transformed biomarker measures and examined trends across the second trimester using a linear trend test. Using a combination of medical record abstraction and self‐report, we created the following covariates: 1) insurance type (Medicaid, Private or HMO, or self‐pay); 2) patient age; 3) patient education (less than high school or GED, some college, college degree or more); 4) parity (no prior pregnancies >20‐24 weeks, one prior pregnancy >20‐24 weeks, two or more pregnancies >20‐24 weeks); and 5) race/ethnicity (Latina, Black, White, Asian or Pacific Islander, or Other). We compared distributions across sociodemographic groups using a Kruskal‐Wallis test. For both the linear trend test and the tests for differences across groups, we adjusted for multiple comparisons using a Bonferroni correction.

We calculated Spearman correlation coefficients of all biomarker concentrations across maternal, fetal, and placental samples and stratified the correlation matrices by gestational age at sample collection (<20 vs. ≥ 20 gestational weeks). Changes in correlation coefficients between gestational age groups (∆r) were summarized to highlight biomarker pairs exhibiting the largest shifts across gestation. As sensitivity analyses, we also evaluated differences using 19 and 21 weeks of gestation as the cutoff, rather than 20 weeks. To evaluate whether correlations were robust to covariate adjustment, we compared the associations of maternal and cord biomarkers using linear regressions with log‐transformed concentrations. We stratified regressions by gestational age at sample collection and controlled for covariates.

3. Results and Discussion

Table 1 shows the demographics of study participants overall and by gestational age of sample collection. Participants receiving care after gestational week 20 were more likely to be age 25–34, have at least a college degree, and to report White race/ethnicity compared to those receiving care before gestational week 20.

TABLE 1.

Descriptive statistics of participants overall and by gestational age at sample collection (N = 106).

	Total	Less than 20 weeks	20 weeks or more
	N (%)	N (%)	N (%)
Age
Less than 25	58 (54.7)	21 (63.6)	37 (50.7)
25 to less than 35	38 (35.8)	9 (27.3)	29 (39.7)
35 or older	10 (9.4)	3 (9.1)	7 (9.6)
Gestational age at sample collection
Less than 20 weeks	33 (31.1)	33 (100)	0 (0)
20 or more weeks	73 (68.9)	0 (0)	73 (100)
Insurance type
Medicaid	79 (74.5)	27 (81.8)	52 (71.2)
Private or HMO	5 (4.7)	1 (3.0)	4 (5.5)
Self‐pay	21 (19.8)	5 (15.2)	16 (21.9)
Missing	1 (0.9)	0 (0)	1 (1.4)
Education
High school or GED	52 (49.1)	18 (54.5)	34 (46.6)
Some college	36 (34.0)	12 (36.4)	24 (32.9)
College or more	18 (17.0)	3 (9.1)	15 (20.5)
Parity
No children	36 (34.0)	8 (24.2)	28 (38.4)
One child	33 (31.1)	13 (39.4)	20 (27.4)
Two or more children	37 (34.9)	12 (36.4)	25 (34.2)
Race/ethnicity
Latina	34 (32.1)	16 (48.5)	18 (24.7)
Black	24 (22.6)	7 (21.2)	17 (23.3)
White	21 (19.8)	3 (9.1)	18 (24.7)
Asian or Pacific Islander	9 (8.5)	6 (18.2)	3 (4.1)
Other	7 (6.6)	1 (3.0)	6 (8.2)
Missing	11 (10.4)	0 (0)	11 (15.1)

When examining trends in concentrations across gestational weeks 15–24, we observed significant increases in cord serum of the T‐helper 2 (Th2)‐responsive chemokines CCL17 and CCL22 (Figures 1 and S1), both which share the same receptor, C‐C Motif Chemokine Receptor 4 (CCR4) [20]. We also observed increases in maternal IL‐2 (Figure S2) and IL‐8 (Figure S3), placental IL‐10 (Figure S4), maternal IFN‐$\gamma$ (Figure S5) and maternal TNF‐R1 (Figure S6), although the trends were not significant after adjustment for multiple testing. The remaining cytokines exhibited less robust time‐dependent differences (Figures S7–S19).

FIGURE 1.

Log concentrations (in pg/mL) of CCL17 in maternal serum, placental tissue, and cord serum across gestational age. Note: We compared differences in biomarker concentrations across gestational age using a linear trend test, controlling for participant age, educational attainment, insurance type, race/ethnicity, and parity. Any significant differences with Bonferroni‐corrected p‐values < 0.05 are marked with *** in the upper left‐hand corner of the plot. Dots in a horizontal line at the bottom of plots represent the values at the lower limit of detection.

We did not observe any differences in biomarker levels across sociodemographic groups (Figures S20–S39). However, consistent with the linear trend differences we noted above, levels of cord CCL17 and CCL22 were significantly higher in samples collected after 20 weeks compared to before (Figures S32 and S33).

Correlations among inflammation biomarkers between second trimester maternal and cord sera were strongest before 20 gestational weeks compared to afterward (Figure 2). Differences in the correlation structure were specific to this gestational time point, as the trends were not as pronounced when using gestation week 19 or 21 as the cutoff (Figures S40–S41). This pattern of stronger relationships between maternal and cord biomarkers before vs. after 20 weeks gestation remained robust after adjusting for sociodemographic factors linked to timing of sample collection (Figure S42). Placental correlations were generally weaker in magnitude than those observed between maternal and cord serum; however, similar to maternal‐cord associations, correlations between the placenta and maternal or cord sera tended to be stronger prior to 20 gestational weeks.

FIGURE 2.

Correlation matrices of log concentrations (in pg/mL) of inflammation biomarkers across maternal, placental, and cord samples stratified by gestational age at sample collection. Note: Biomarkers are grouped according to sample type. For instance, those with the “M” suffix in blue text come from maternal serum, those with the “P” suffix in red text are from placental tissue, and those with the “C” suffix in orange text come from cord serum.

While most biomarker associations weakened in later gestation, a subset of specific correlations, mostly involving the placenta, demonstrated a marked strengthening after 20 weeks. Among biomarker pairs with the largest changes in correlation (∆r>0.4), correlations between placental CCL17 and cord CCL17, placental IFN‐α, or placental IL‐1β increased the most (Figure 3). Similarly, correlations between cord CCL2 and placental IL‐10, placental IL‐1β, placental TNF‐α and maternal IL‐7 also increased by 0.4 or more. The only other biomarker‐pair with correlations that exceeded this threshold were placental IL‐28A and maternal IL‐1β.

FIGURE 3.

Differences in correlation between log concentrations (in pg/mL) of maternal, placental, and cord biomarkers of inflammation before versus after 20 weeks gestation. Note: Biomarkers are labeled according to sample type. For instance, those with the “M” suffix are from maternal serum, those with the “P” suffix are from placental tissue, and those with the “C” suffix are from cord serum.

In contrast, the broader maternal‐fetal immune structure showed a significant decoupling. Overall, 95 biomarker pairs exhibited decreases in correlation of ∆r > 0.4 between the two gestational age groups, of which 10 pairs decreased by ∆r > 0.55. Those with the most pronounced decreases (∆r ≥ 0.55) involved maternal CCL17 with placental CXCL10, placental IL‐6, and cord IFN‐α, IL‐10, IL‐12, or IL‐7. Correlations between placental IL‐6 and maternal IFN‐α, as well as placental IL‐6 and placental IL‐10, decreased by ≥ 0.55. Finally, correlations between maternal IL‐8 and cord IL‐6, and maternal TNF‐RI and cord IL‐2 also followed this downward trend.

In this study, we characterized correlations among inflammation biomarkers in matched maternal, placental, and cord biospecimens during the second trimester, a period marking the transition from placental development to functional maturation. We observed that correlations among biomarkers, especially those between maternal and cord serum, were strongest prior to gestational week 20. These relationships were robust to adjustment for sociodemographic factors linked to timing of care, suggesting 20 weeks may serve as a critical inflection point in the maternal‐fetal inflammatory relationship, although longitudinal data and mechanistic evidence are needed to confirm this.

A prominent feature of this shift involved the chemokines CCL17 and CCL22. While maternal CCL17 levels remained stable, cord CCL17 levels increased significantly across the second trimester. Furthermore, the strength of correlation between placental and cord CCL17 increased after 20 weeks’ gestation, suggesting a developmental transition from maternal dependency to greater endogenous fetal production around this time. In addition to cord CCL17, cord CCL22 levels also increased significantly across 15–24 weeks gestation. Both chemokines share the receptor CCR4, which is expressed on Th2 cells, a subset of CD4+ T lymphocytes that promote anti‐inflammatory responses [20]. The pregnancy‐associated shift from Th1 to Th2 is hypothesized to facilitate maternal tolerance of fetal alloantigens, particularly at the maternal‐fetal interface [21]. Therefore, the observed increases in fetal CCL17 and CCL22 across the second trimester may reflect a key transition in fetal immune maturation. Dysregulation of this process may lead to increased risk of preterm delivery or neurodevelopmental delay, as higher maternal CCL17 levels have been found among women experiencing preterm labor [22] and CCL17 plays an important role in neuroimmune coordination [23] and implicated in the development of allergies and autoimmune diseases [24, 25].

While cytokines are not known to cross the placental barrier, they can stimulate the release of other factors that may cross the barrier and activate inflammatory signaling pathways [26]. As the placenta matures in the second trimester, molecular transfer becomes more restricted, and the fetus increasingly develops and regulates its own immune system. Our findings indicate that maternal and fetal inflammatory responses are more coordinated earlier in the second trimester, emphasizing the importance of gestational age in profiling inflammation biomarkers, stress and development during pregnancy. Furthermore, this work suggests that the heightened coordination between maternal and fetal inflammatory states during early gestation may provide a mechanistic basis for the specific vulnerability of early gestational periods to environmental and social stressors [14, 27, 28, 29].

Funding

This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (Grant R21HD109723) and the National Institute of Environmental Health Sciences (Grants P30ES030284, P01ES022841).

Conflicts of Interest

The authors report no conflicts of interest.

Supporting information

Supporting File: aji70324‐sup‐0001‐SuppMat.docx

Goin D. E., Li L., Gaw S. L., et al. “Maternal‐Fetal Interplay During Mid‐gestation: Inflammatory Biomarker Levels in Maternal Sera, Cord Sera and Placenta.” American Journal of Reproductive Immunology 96, no. 3 (2026): e70324. 10.1111/aji.70324