Race and sex differences in placental lipid metabolism are associated with spontaneous early preterm birth^†

Abstract

Placental dysfunction is implicated in the pathogenesis of spontaneous preterm birth (SPTB). We investigated race (self-identified maternal race) and fetal sex differences in the placental metabolome and transcriptome associated with early SPTB (<32 weeks). Long-chain polyunsaturated fatty acids, acylcarnitines, acylglycerols, plasmalogens, and lysophospholipids were remarkably different between SPTB and Term placentas. These alterations were much more profound in Black than in White SPTB placentas. Mode of delivery and body mass index (BMI) had no effect on these differences. The lipid metabolic pathways disrupted in early SPTB placentas also exhibited fetal sex differences, particularly between Black male and Black female placentas. The expression of genes involved in multiple lipid metabolism regulating pathways (e.g., PI3K/AKT signaling and phospholipase activity), especially eicosanoid synthesis and secretion, was significantly altered in early SPTB placentas. The race- and sex-specific changes in lipid metabolites and gene expression were consistent with inflammation in SPTB placentas, which was further supported by dysregulation of various inflammation and immune response pathways. These findings reveal race and fetal sex differences in lipid metabolism and inflammation in SPTB placentas and suggest greater dysfunction and inflammation in Black compared to White SPTB placentas, which may explain mechanisms underlying early SPTB and the risk of SPTB in different populations.

Graphical Abstract

Open in new tabDownload slide

Introduction

Preterm birth (delivery before or at 37 weeks of gestation) is the leading cause of neonatal morbidity and mortality. In the United States, nearly 400 000 babies are born prematurely every year among the nearly 4 million births. Among them, ~64 000 are born early preterm (<32 weeks), which leads to approximately 12 000 infant deaths, representing 52% of all infant deaths [1]. Despite ongoing research, spontaneous preterm birth (SPTB) remains a poorly understood pregnancy complication. Multiple factors have been proposed to contribute to SPTB, including placental dysfunction, abnormal cervical remodeling, uterine distension, and chorioamnionitis [2, 3]. The clinical phenotypes of placental dysfunction include absent or reverse end-diastolic flow, placental vascular lesions or infarction, placental weight < 10% for gestational age, and birthweight <10% for gestational age and gender [2, 3].

Fetal sex, race, and social determinants of health that differentially impact pregnancy outcomes across racial/ethnic groups play important roles in pregnancy complications that frequently are attributed to placental dysfunction. For example, male fetal sex is a risk factor for preeclampsia and gestational diabetes in some populations [4, 5]. Similarly, a higher incidence of preterm birth is observed among women carrying male fetuses [5, 6]. Race (as self-identified) is a social construct and, when considered with other social determinants, can significantly impact pregnancy outcomes. In the United States, the overall preterm birth rate is 52% higher for Black women than for White women [7]. The incidence of other pregnancy complications—such as preeclampsia—is also higher among Black women, and research examining the role of racism on the burden of disease may identify effective strategies to reduce the incidence of preterm birth in Black women [8]. However, the mechanisms underlying race and fetal sex differences in the incidence of pregnancy complications are poorly understood.

A well-functioning placenta is crucial for normal gestation. Emerging evidence suggests that placental dysfunction is associated with a significant proportion of preterm births, especially early preterm births [9, 10]. The placenta has a high metabolic rate, and maternal nutrients are not only transferred to the fetus but also used to meet the energy requirements of the placenta. In our previous studies, we observed altered energy metabolism and mitochondrial function in SPTB placentas, suggesting an imbalance between substrate supply and metabolic demands [11, 12]. Although deficits in the capacity of the placenta to maintain bioenergetic and metabolic stability may ultimately result in SPTB, it is unclear whether race and fetal sex differences and race and sex interactions contribute to placental changes and dysfunction. The goal of this study was to elucidate molecular mechanisms underlying race and sex differences in placental function to better understand the etiology of prematurity and inform the development of race- and fetal sex-specific preventative treatments.

Materials and methods

Clinical characteristics

The majority of preterm and all term birth placental samples from non-Hispanic Black or White patients (self-identified maternal race) in the current study were selected from the larger Cellular Injury and Preterm Birth (CRIB 821376, NCT02441335) study at the University of Pennsylvania. CRIB enrollment criteria included patients aged 18–45 years with singleton pregnancies admitted to the hospital with either spontaneous labor (defined as regular contractions and cervical dilation) or premature rupture of membranes (PROM) occurring between 20 0/7 and 36 6/7 weeks of gestational age (preterm) or at 38 to 41 weeks of gestational age (term). CRIB exclusion criteria included multiple gestations, fetal chromosomal abnormalities, major fetal anomalies, intrauterine fetal demise, intrauterine growth restriction (IUGR), clinical chorioamnionitis, induction of labor, elective cesarean delivery, and gestational diabetes. Patients with preeclampsia—categorized as medically indicated preterm birth—were recruited in the CRIB study but excluded from the current study. All patients contributing placental samples presented in labor with either preterm PROM, PROM, or cervical dilation. Most of the patients with preterm labor and none of the patients who labored at term received betamethasone treatment prior to delivery (Table 1). The CRIB study was approved by the Institutional Review Board at the University of Pennsylvania (protocol #821376), and patients were enrolled after giving written informed consent. To increase the number of samples from White patients, six preterm birth placental samples (five males and one female) were from normotensive control pregnancies from the Placental Origins of Preeclampsia study approved by the Institutional Review Board at Stanford University (protocol #34745). Samples were from participants who approved the use of excess samples for other pregnancy studies.

All second-trimester placental samples from Black patients (self-identified maternal race) and 8 s-trimester placental samples from White patients (self-identified maternal race) with gestational age between 19 and 23 weeks were selected from the “Trophoblast Cells Isolation from Second and First Trimester Placenta” (TrISecT) study at the University of Pennsylvania and were utilized as gestational age controls for SPTB groups (gestational age between 21 and 32 weeks) in the current study. The enrollment criteria included patients aged 18–45 years receiving care at the hospital due to elective termination of a singleton pregnancy prior to 23 6/7 weeks gestational age. TrISecT exclusion criteria included multiple gestation, aneuploidy, and fetal congenital anomalies. The TrISecT study was approved by the Institutional Review Board at the University of Pennsylvania (protocol #827072), and patients were enrolled after giving written informed consent. To increase the number of samples from White patients, 11 second-trimester placental samples were provided from the Human Placental Development biobank approved by the Institutional Review Board at Stanford University (#31552). The race of these 11 samples was determined by Infinium Global Screening Array performed by the Center for Applied Genomics at Children’s Hospital of Philadelphia. The sex of all second-trimester placentas was determined by real-time quantitative polymerase chain reaction (q-PCR) using TaqMan probes for Xist as a positive control (Hs01079824_m1) and RPS4Y1 (Hs00606158_m1) as a marker of male sex (Thermo Fisher Scientific, Waltham, MA).

Placental sample collection

Preterm and term birth placental samples (placenta villi) were collected from the same site: the placenta near the cord insertion at 9 o’clock area along its longest length. Samples with the size of 0.5 × 0.5 × 0.5 cm were collected from the fetal side at the time of delivery, rinsed with 1x DPBS to remove maternal blood, and flash-frozen in liquid nitrogen. Second-trimester placental samples (100 mg) were collected from the placenta villi at the time of termination, rinsed with 1× DPBS to remove maternal blood, and flash-frozen in dry ice. All placental samples were stored at −80°C prior to metabolomic and transcriptomic studies.

Metabolomics and analysis

Placental metabolite extraction and primary metabolomic analysis were performed by Metabolon Inc. (Durham, NC, USA). The samples were prepared as previously described [11]. The extracted analytes were split into equal parts for analysis on complementary GC/MS (gas chromatography/mass spectrometry) and LC/MS (liquid chromatography/mass spectrometry) platforms as previously described [11]. Compounds were identified by comparison to Metabolon library entries of purified standards or recurrent unknown entities.

Following normalization to mass, log transformation, and imputation of missing values, if any, with the minimum observed value for each compound, ANOVA contrasts were used to identify compounds that differed significantly between experimental groups. A p < 0.05 was considered significant. An estimate of the false discovery rate (FDR) was calculated to correct for multiple comparisons and is described as a q-value.

Multivariable regression analyses were performed using Prism 9 (GraphPad, Boston, MA) to determine the association between the maternal BMI and levels of lipid metabolites.

Total RNA isolation and RNA-Seq library preparation

Total RNA was extracted from the placenta using TRIzol Reagent (Invitrogen), followed by Qiagen RNeasy Mini Columns following the manufacturer’s instructions. RNA integrity was determined using an Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA), and RNA integrity numbers greater than six were used for the RNA-seq study. RNA-Seq libraries were generated using the Agilent SureSelect Strand-Specific RNA Library Preparation Kit (Agilent, Santa Clara, CA).

RNA-Seq and gene expression analysis

RNA-Seq libraries were single-end sequenced to 100 bp on an Illumina NovaSeq platform in the Next-Generation Sequencing Core at the University of Pennsylvania. The RAVED pipeline (https://github.com/HimesGroup/raved) [13] was used to analyze RNA-Seq data. Adapter sequences in raw reads were trimmed out using Trimmomatic (v.0.32), and overall QC metrics were obtained using FastQC (v.0.11.7). Trimmed reads for each sample were aligned with the reference hg38 genome using STAR (v.2.5.2b). The Picard Tools (v.1.96; http://picard.sourceforge.net) RnaSeqMetrics function was used to compute the number of bases assigned to various classes of RNA. HTSeq (v.0.6.1, with the strand setting “-s reverse”) was used to quantify gene-level read counts. Gene-level differential expression analysis was performed under a negative binomial distribution model using DESeq2 (v.1.18.1) [14] while filtering out genes with a total read count <10 before differential expression analysis. The Benjamini–Hochberg approach was used to correct for multiple comparisons of genes, and an FDR (q-value) <0.05 was considered significant. Sequence data have been deposited in NCBI’s Gene Expression Omnibus (GEO) and are accessible through GEO Series accession number GSE251663. Functional analysis was conducted using QIAGEN’s Ingenuity Pathway Analysis (IPA). Core analyses were performed on genes with q < 0.05 and fold change ≥1.5-fold.

Integrated network analysis of the metabolome and transcriptome

Significantly altered metabolites and differentially expressed genes identified in each study group were analyzed using MetScape3.1 in Cytoscape (v3.8.0). The interactome networks were generated based on known protein–protein and protein–metabolite interactions. The metabolic pathways that were associated with protein–metabolite interactions were mapped onto each network.

Statistics

Patient clinical measurements are expressed as mean ± SD for continuous variables, or as a number (percent) for categorical variables. Two sample t-tests were used to compare means between the two groups. p < 0.05 was considered the threshold for significance for demographic comparisons. Statistical analyses for the levels of metabolites and gene expression are described above.

Study approval

Placentas were collected under protocols approved by the Institutional Review Boards at the University of Pennsylvania and Stanford University, and all patients gave their written consent for the use of their biological samples for research purposes.

Results

Clinical characteristics

The clinical data for the study cohort are presented in Table 1. The average gestational age for the SPTB group was between 25 and 29 weeks of gestation, 38 and 41 weeks for the term group, and 19 and 23 weeks for mid-gestation placentas, which were used as gestational age controls. There were no significant differences in maternal age or mode of delivery between term and preterm groups. BMI tended to be higher in the mid-gestation group and was significantly higher (p < 0.05 vs. Term Control) in Black patients with male fetuses (Table 1). Of note, none of the preterm or term placentas were from patients diagnosed with chronic hypertension, preeclampsia, or gestational diabetes, and none of the patients received low-dose aspirin for preeclampsia prevention. Most patients who delivered preterm received prenatal steroids, and roughly half of the preterm groups received antibiotics during labor (Table 1). The experimental design for the study is shown in Figure 1.

Global assessment of the placental metabolome

To identify mechanisms underlying early SPTB, an untargeted metabolomic analysis was performed on placentas from SPTB and term birth. Healthy mid-gestation placentas were used as gestational age controls. A total of 866 metabolites were identified. Overall, there were significant differences in the metabolome between SPTB and Term placentas, and most of these were not due to expected changes associated with gestational age other than steroid metabolism (p < 0.05) (Supplemental Table S1). Mode of delivery and the BMI had no effect on these differences. To assess the effects of race and fetal sex on placental metabolism and their relationship to SPTB, we compared Black and White (self-identified maternal race) and male and female placentas at three different time points (mid-gestation, SPTB, and Term). Surprisingly, maternal race was associated with profound differences in the metabolome at mid-gestation, with few differences between sexes (Supplemental Table S1). In SPTB placenta, sex was associated with a slightly larger difference than race, whereas in Term placenta, there were far fewer differences, and race had a much greater impact than sex (Supplemental Table S1, Supplemental Figure S1A).

Dysregulated lipid metabolism in early SPTB placenta

The most significant differences in metabolism between early SPTB and Term placenta were related to lipid metabolism. Changes in the levels of long-chain fatty acids (LCFAs) in SPTB vs. Term placentas exhibited profound race differences. Interestingly, there were no differences in male and female White SPTB vs. Term placentas. In contrast, the levels of LCFAs differed significantly between sexes in Black SPTB vs. Term placentas (Figure 2A & Supplemental Table S2). Of note, almost all long-chain polyunsaturated fatty acids (LC-PUFAs) were elevated in Black female SPTB placentas, including docosahexaenoate (DHA, ω-3 PUFA) and arachidonic acid (ARA, ω-6 PUFA). In contrast, multiple ω-3 PUFAs were significantly lower in Black male SPTB placentas, including eicosapentaenoate (EPA), docosapentaenoate, and DHA (p < 0.05, q = 0.14), suggesting that a lower ω-3 PUFA pool was available for the Black male fetus. These differences were not due to developmental changes (Supplemental Table S2). Further, despite a higher BMI in Black women carrying male fetuses, this was not associated with changes in lipid metabolism in Black male placentas.

LCFAs are delivered to mitochondria as acylcarnitines and undergo β-oxidation to generate acetyl-CoA and ATP [15]. Consistent with our previous findings [11], multiple acylcarnitines—particularly long-chain and hydroxy acylcarnitines—were elevated in SPTB vs. Term placenta with race and sex differences (Figure 2B and Supplemental Table S3).

Acylglycerols, generated from triacylglycerols, in addition to serving as an energy source, also act as lipid second messengers regulating proliferation, energy metabolism, and immune signaling [16–18]. Comparing early SPTB vs. Term placentas, multiple monoacylglycerols and diacylglycerols were significantly different in Black placentas, with modest differences in White placentas. Further, the sex differences in Black placentas were greater compared to White placentas (Figure 2C and Supplemental Table S4).

Both phosphatidylcholine (PC) and phosphatidylethanolamine (PE) can exist as plasmalogens, which have a vinyl ether bond at the sn-1 position. This vinyl ether bond is more susceptible to oxidation than the ester bond in other phospholipids and has been suggested to be protective against oxidative stress [19, 20]. Multiple plasmalogens and lysoplasmalogens were significantly altered in SPTB vs. term placentas with both race and sex effects (Figure 2D, Supplemental Table S5). Interestingly, two of the elevated plasmalogens contain ARA (P-16:0/20:4 and P-18:0/20:4), which have been shown to be higher in placentas of women with preeclampsia [21]. Consistent with increases in LC-PUFAs, lysophospholipid levels were significantly higher in Black female SPTB vs. Term placentas, but there were few changes in other groups (Figure 2E, Supplemental Table S6).

Steroid hormones including pregnenolone, progesterone, androgens, estrogens, and certain corticosteroids were significantly lower in all SPTB vs. Term placentas (Supplemental Table S7). These differences were expected since placental production of progesterone and placental metabolism of fetal adrenal cortex-derived androgens into estrogens rise with advancing gestational age. Metabolism of cortisol also increases with advancing gestational age.

We performed multivariable regression analyses and found no impact of maternal BMI on most lipid metabolites, except two acylcarnitines, eicosenoylcarnitine (C20:1) (p = 0.030) and docosatrienoylcarnitine (C22:3) (p = 0.033), and one plasmalogen PE, 1-(1-enyl-palmitoyl)-2-oleoyl-GPE (P-16:0/18:1) (p = 0.031). The levels of lipid metabolites were not affected by the mode of delivery (vaginal vs. C-section).

Global assessment of placental transcriptome

We performed RNA-Seq analysis on the same placentas used for the metabolomic study to identify differentially expressed genes and signaling pathways associated with lipid metabolism in early SPTB and to identify additional mechanisms that contribute to placental dysfunction in the setting of SPTB. Genes were considered differentially expressed with an FDR (q-value) ≤0.05 and fold change ≥1.5-fold. Comparing early SPTB vs. Term placentas combining race and fetal sex, 2189 differentially expressed genes were identified. After performing race and sex stratification, gene expression was profoundly different in Black male and female SPTB vs. Term placentas. Overall, 1748 and 1145 differentially expressed genes were identified in Black males and Black females, respectively. Interestingly, differences in the transcriptome between White SPTB vs. Term placentas were much smaller than those in Black placentas. Only 184 and 331 differentially expressed genes were identified in White males and White females, respectively. After removal of gestational age-associated genes and sex-linked genes, 288 autosomal genes were significantly different in Black SPTB males; 172 genes were significantly different in Black SPTB females; and 98 genes were significantly different in White SPTB females (Supplemental Table S8). Since only 39 genes with a q-value ≤0.05 and fold change ≥1.5 were identified in White SPTB males, we extended our analysis to 84 differentially expressed genes with q-values ≤0.15 and fold changes ≥1.5. The expression changes of these genes in early SPTB vs. Term placentas were significantly associated with preterm birth. Differences in gene expression directly comparing male vs. female SPTB placenta (32 and 8 autosomal differentially expressed genes in Black and White placenta, respectively) or between Black vs. White SPTB placentas (18 and 21 autosomal differentially expressed genes in male and female placenta, respectively) were more modest.

Venn diagrams were used to display overlapping and unique differentially expressed genes (Supplemental Figure S1B). There were few overlapping differentially expressed genes comparing Black and White or males and females, suggesting that molecular mechanisms underlying placental changes in early preterm birth differ between race and sex. The nine differentially expressed genes commonly identified in SPTB vs. Term placentas across all study groups included NID1, CKB, MPO, PLA2G4A, PRTN3, SLC22A3, SV2B, TGFBR2, and VEGFA (Supplemental Table S9). Interestingly, four genes (MPO, PLA2G4A, TGFBR2, and VEGFA) are involved in regulating lipid metabolism.

Interactome network analysis of the metabolome and transcriptome in early SPTB placenta

The interactome network model was generated by connecting protein–protein or protein–metabolite interactions to investigate the association between significantly altered metabolites and differentially expressed genes (Supplemental Figure S2). The analysis confirmed that multiple lipid metabolic pathways were altered in early SPTB placentas, including ARA metabolism, eicosanoid metabolism, fatty acid β-oxidation, phospholipid metabolism, and steroid hormone metabolism. The significantly altered metabolites and differentially expressed genes associated with each pathway are listed in Table 2.

Pathways modulating lipid metabolism and lipid signaling are dysregulated in early SPTB placenta

Signaling pathways that may contribute to placental alterations in early SPTB were identified by IPA. Similar to our metabolomic findings of dysregulated lipid metabolism in SPTB placentas, multiple pathways regulating lipid metabolism were significantly altered in early SPTB vs. Term placentas with race and sex differences (Table 3). Although alterations in genes and signaling pathways that regulate lipid metabolism exhibited race and sex differences, disrupted lipid metabolic processes were similar in all study groups. In addition to changes in expression of genes related to synthesis and release of fatty acids, multiple genes regulating eicosanoid metabolism were differentially expressed in early SPTB placentas, predicting activation of leukotriene, prostaglandin, and thromboxane synthesis (Table 2 and 4, Supplemental Table S10). Eicosanoids are derived from ARA and other polyunsaturated fatty acids. Multiple genes regulating ARA metabolism were consistently differentially expressed in SPTB placentas, particularly in Black placentas (Table 2 & 4, Supplemental Table S10). These gene expression patterns correlated with increased levels of ARA-containing lipid metabolites (e.g., arachidoylcarnitine, 1-arachidonylglycerol, linoleoyl-arachidonoyl-glycerol, and 1-arachidonoyl-GPC) in Black female but not Black male SPTB placenta (Supplemental Tables S3–S6) indicating sex differences. Interestingly, the expression of multiple genes involved in lipid metabolism also exhibited race and sex differences in Term placentas, including three genes (ALOX15B, GGT5, and CCL3L3) for eicosanoid metabolism (Supplemental Table S11).

Expression of genes regulating additional pathways of lipid metabolism differed between SPTB and Term placenta in a sex- and race-dependent manner. For example, multiple genes mediating PI3K/AKT signaling and leptin signaling (including GUCY1B1, NOTUM, and PDE3A) (Table 3) were differentially expressed in SPTB compared to Term placenta. Nuclear receptor-mediated signaling pathways—including pregnane X receptor (PXR), peroxisome proliferator-activated receptor α (PPARα), liver X receptor (LXR), and farnesoid X receptor (FXR)—were dysregulated in SPTB placentas with race and sex differences (Table 3). These nuclear receptors can act as sensors of fatty acids, cholesterol, and bile acids and regulate the expression of genes involved in lipid metabolism [22]. p38 MAPK signaling and ERK signaling were activated in SPTB placenta (Table 3), which can regulate sterol regulatory element-blinding protein (SREBP) and PPAR-γ, and modulate lipid metabolism [23, 24]. Moreover, eicosanoid signaling was dysregulated in early SPTB placentas with race and sex differences (Table 3). Eicosanoid signaling mediates inflammation and vascular permeability and is associated with preterm birth [25]. Finally, multiple genes encoding phospholipases were upregulated in Black male SPTB vs. Term placentas (Table 3), suggesting that phospholipases are more active in Black male SPTB placentas compared to the other groups.

Genes regulating mitochondrial and peroxisomal lipid metabolism are differentially expressed in early SPTB placenta in a race- and sex-specific manner

Multiple long-chain acylcarnitines were elevated in a race- and sex-specific manner in SPTB placenta vs. Term placenta (Figure 1B and Supplemental Table S3). By quantifying the flux of palmitate in placentas, we demonstrated a significantly reduced rate of palmitate oxidation in SPTB vs. Term placenta in our previous study [11], indicating possible defects in mitochondrial β-oxidation in SPTB placenta. Thus, we examined the expression of genes involved in mitochondrial β-oxidation and found that multiple genes were differentially expressed in SPTB vs. Term placenta (Supplemental Table S12). Our results suggest that Black male SPTB placentas have more efficient fatty acid β-oxidation compared with other SPTB groups, which is consistent with our findings in the metabolome, where acylcarnitines were not elevated in Black male SPTB vs. Term placentas.

Peroxisomes play critical roles in lipid metabolism and oxidative stress. They synthesize ether lipids (e.g., plasmalogens), catabolize the oxidation of very-long chain fatty acids (VLCFAs, ≥22 carbons), branched-chain fatty acids, and dicarboxylate fatty acids, and participate in bile acid metabolism [26]. In addition to alterations of plasmalogens and lysoplasmalogens in SPTB vs. Term placentas (Figure 1D, Supplemental Table S5), the levels of multiple branched-chain fatty acids, dicarboxylic fatty acids, and bile acids were significantly altered in SPTB vs. Term placentas (Supplemental Table S13), suggesting changes in peroxisome-mediated lipid metabolism in early SPTB placentas. Moreover, multiple genes regulating peroxisomal biogenesis and function [27, 28]–including PEX7, GNPAT, SLC27A2, and BRDT—were differentially expressed in SPTB vs. Term placentas with race and sex differences (Supplemental Table S13).

Additional pathways related to lipid metabolism are altered in early SPTB placenta with race and sex differences

Expression of genes that regulate platelet-activating factor (PAF) synthesis was dysregulated in SPTB placentas in a sex- and race-specific manner (Table 4, Supplemental Table S10). PAF is a phospholipid mediator of inflammation, causing platelet aggregation and vasodilation [29].

The endocannabinoid pathway was only inhibited in Black male SPTB placentas (z-score = −1.34, p = 0.023). Endocannabinoids are signaling molecules generated from phospholipids that are critical for trophoblast differentiation, placentation, and maternal immune tolerance [30]. Multiple genes comprising this signaling pathway were differentially expressed in Black male SPTB placenta, including PROK1 and NOS2. The levels of endocannabinoid metabolites were consistently lower in Black male SPTB vs. Term placentas, but in no other groups (Supplemental Table S14).

Pathways regulating angiogenesis and vascular remodeling are disrupted in early SPTB placenta

Several pathways regulating both lipid metabolism and vascular development/function were altered in SPTB placentas with race and sex differences, including vascular endothelial growth factor (VEGF), p38 MAPK, and hypoxia-inducible factor 1 (HIF-1) signaling (Supplemental Table S15). p38 MAPK not only regulates multiple aspects of lipid biology, but this protein also regulates VEGF function and increases vascular permeability [31, 32]. HIF-1 signaling—which mediates angiogenesis in response to hypoxic conditions—is associated with placental aging and preterm birth [33]. Further, Hif-1a promotes lipid uptake and synthesis and reduces lipid oxidation. Finally, multiple genes associated with vascular lesions and vaso-occlusion were differentially expressed in SPTB placentas in a sex- and race-specific manner (Supplemental Table S15).

Inflammatory and immune response signaling pathways are dysregulated in early SPTB placenta

Placental inflammation is highly associated with SPTB. Thus, we delved further into differentially expressed genes involved in modulating inflammatory and immune responses (Table 5). Signaling pathways including those involved in interleukin signaling, eicosanoid signaling, fibrosis signaling, and the STAT3 pathway—which regulate innate immunity, leukocyte, dendritic cell, and macrophage function—were significantly altered in early SPTB compared to Term placentas with race and sex differences. Importantly, the expression of these genes was not affected by the mode of delivery.

Discussion

In this study, we demonstrated marked differences in metabolites and expression of genes involved in lipid metabolism in early SPTB vs. Term placentas with race and fetal sex differences (highlighted in Figure 3 and summarized in Supplemental Table S16). These alterations were more profound in Black than in White SPTB placentas, suggesting that Black early SPTB placentas exhibited greater metabolic dysfunction than White SPTB placentas. We found that lipid metabolic pathways disrupted in early SPTB placentas were different, particularly between Black male and female groups, indicating that underlying mechanisms leading to early SPTB differ between women carrying male fetuses and those carrying female fetuses. Multiple other metabolic pathways—which have been previously reported to be altered in SPTB placenta [11, 34]—also differed in our study, including metabolism of branched-chain amino acids, polyamines, glutathione, and purines.

Lipids are critically important for normal placenta and fetal development [35, 36]. Fatty acids are actively transported into the placenta where they serve as energy-yielding substrates and play important roles in fetal energy utilization and placental function. However, accumulation of fatty acids in the placenta can induce inflammation leading to preterm labor, which is consistent with transcriptomic changes in SPTB placenta in our study. It is reported that maternal obesity alters placental lipid metabolism and transfer of LC-PUFAs to the fetal circulation, which were significantly higher in Black female SPTB placentas and lower in Black male SPTB placentas in our study [37, 38]. Although no significant differences were observed in average maternal BMI between preterm vs. term placentas in our study and we did not find an impact of maternal BMI on lipid metabolism, a study conducted at our institution showed that patients with SPTB have greater intake of dietary fat than patients who delivered at term [39], which may have a significant impact on lipid metabolism in placentas. This remains to be further explored.

Our study demonstrated that multiple classes of lipid metabolites significantly differed in early SPTB vs. Term placentas with race and sex differences. Almost all lipid metabolic pathways were markedly disrupted in Black female SPTB placentas, including LC-PUFAs, lysophospholipids, acylcarnitines, and acylglycerols. These changes suggest that placentas from Black females may be more vulnerable to additional stressors. Alternatively, these results could also be adaptive, and future studies need to be done to determine whether the changes in fatty acid metabolism are adaptive or deleterious.

Mitochondria are the powerhouse of the cell and play a central role in regulating placental function and pregnancy outcomes. They are critical in energy production; metabolism of glucose, amino acids, lipids, and nucleotides; redox homeostasis and oxidative stress; and inflammation. Acylcarnitines transport LCFAs into mitochondria for β-oxidation, and increased acylcarnitines in SPTB placentas may reflect mitochondrial dysfunction [37]. High concentrations of long-chain acylcarnitines can further inhibit mitochondrial oxidative phosphorylation, induce the formation of reactive oxygen species, and activate proinflammatory pathways [40]. This ongoing cycle can lead to severe placental dysfunction, ultimately resulting in preterm birth.

Peroxisomes are multifunctional organelles. In addition to hydrogen peroxide production and catabolism, peroxisomes are involved in several aspects of lipid metabolism [26]. Altered levels of plasmalogens, branched-chain fatty acids, dicarboxylic fatty acids, and bile acids and changes in the expression of genes regulating these processes suggest that changes in peroxisomal function, in addition to mitochondrial dysfunction, are key features of early SPTB placentas, which show race and sex differences. During pregnancy, triacylglycerols increase and are used as the main energy source. However, a high level of triacylglycerols during pregnancy is associated with an increased risk of preterm birth and preeclampsia [41, 42]. Increased mono- and di-acylglycerols together with saturated acylcarnitines in Black female SPTB placentas reflect increased utilization of triacylglycerols for energy production. In contrast, lower acylglycerols in Black male SPTB placentas suggest decreased triacylglycerol utilization. Both situations suggest the possibility of high triacylglycerol levels in Black women with SPTB. Additionally, elevated monoacylglycerols can activate PPAR signaling, which is critical in regulating lipid metabolism, energy storage, inflammation, and cytokine secretion [43]. Moreover, 2-AG (a monoacylglycerol)—which is an endocannabinoid and significantly higher in Black female SPTB vs. Term placenta—can promote mitochondrial dysfunction, endoplasmic reticulum (ER) stress, and apoptosis in trophoblast [44, 45]. Increased diacylglycerols can cause lipotoxicity in the placenta and are associated with insulin resistance and preeclampsia [46, 47].

There were significant race and sex differences in pathways regulating lipid metabolism. PI3K is a key regulator of lipid metabolism and oxidative stress by modulating mitochondrial oxidative phosphorylation and integrity [48]. Leptin is important for normal pregnancy; it plays a role in normal placentation and fetal-placental communication [49]. It also has a critical role in regulating β-oxidation and lipogenesis [50]. The p38 MAPK pathway can regulate lipid metabolism by modulating SREBP function [23]. Moreover, p38 MAPK activation—which was observed in early SPTB placenta in our study—can lead to senescence and premature aging of fetal tissues and is associated with inflammation capable of triggering preterm labor [51]. Together, our findings suggest that multiple pathways regulating lipid metabolism are dysregulated in SPTB placentas, and Black placentas exhibit the most profound changes.

A key finding in our transcriptomic study was the activation of eicosanoid synthesis in early SPTB placentas with profound race and modest sex differences. Increases in eicosanoids are associated with pregnancy complications, including preterm birth and preeclampsia [25, 52]. Aung et al. reported that eicosanoid levels in maternal blood or urine are significantly associated with SPTB, particularly lipoxygenase metabolites, which can be used as lipid biomarkers for preterm birth prediction [53]. In our study, the level of 15-HETE—which is generated by 15-lipoxygenase—was significantly higher (2.2-fold increase) in Black female SPTB vs. Term placentas. Although 15-HETE levels were not elevated in other SPTB groups, the expression of arachidonate 15-lipoxygenase type B (ALOX15B) was markedly higher (16-fold increase) in Black male SPTB vs. Term placenta, and arachidonate 5-lipoxygenase activating protein (ALOX5AP), which works with 5-lipoxygenase to generate 5-HETE and leukotriene A4, was increased 2-fold in White female SPTB placentas, consistent with increased eicosanoid synthesis in early SPTB placentas. We did not find changes in 12-HETE levels, and we did not detect other eicosanoids in our samples, possibly due to the threshold of detection and/or instability and short lifespan of eicosanoids. Further evaluation of eicosanoids with a more sensitive method may provide further insight into the roles of these bioactive lipids in pregnancy outcomes.

Inflammation is implicated as one mechanism underlying placental dysfunction, and lipid metabolism plays a critical role in modulating immune responses and inflammation. Conversely, immune cells—particularly resident macrophages—play a key role in lipid storage and utilization and regulate fatty acid and cholesterol metabolism [54]. Hofbauer cells are fetal-origin resident macrophages in the placenta, and malfunction of Hofbauer cells is associated with multiple pregnancy complications including preterm birth [55]. Our transcriptome data showed that many pathways regulating inflammation/immune responses were significantly altered in early SPTB placentas, which can cause tissue damage and alter vascularization. Interestingly, these immune pathways—including STAT3, IL-6, and IL-4 signaling—also indirectly regulate lipid metabolism [56, 57]. Moreover, leptin—which was identified as a common upstream regulator of the metabolome and transcriptome in our study—is an adipokine critical for regulating lipid metabolism. It also regulates immune responses including macrophage polarization, phagocytosis, and cytokine production [58]. Although labor is an inflammatory process, the levels of metabolites or gene expression in our datasets were not affected by the mode of delivery nor betamethasone treatment.

Profound race and sex differences in lipid metabolites and gene expression were demonstrated in early SPTB placentas. Social determinants affect maternal health and pregnancy outcomes. In fact, poor nutrition or exposure to environmental chemicals—such as endocrine disruptors—can alter lipid metabolism, possibly through epigenetic mechanisms [59]. DNA methylation is associated with lipid compositions and concentrations [60]. It is reported that DNA methylation patterns are altered in preterm birth placentas and associated with maternal socioeconomic status [61, 62]. Whether social determinants, which, in turn, induce DNA methylation/epigenetic changes, underlie the race differences observed in our study will need to be further investigated.

Sex hormones play important roles in numerous metabolic processes including lipid metabolism. The placenta synthesizes estrogens and is the major source of estrogens in pregnancy. Through estrogen receptors ERα, ERβ, and G protein-coupled estrogen receptor, estrogen can activate ERK, PI3K, and c-AMP signaling and regulate the expression of genes important for lipid metabolism [63–65]. The placenta also has a biological sex. The complement effects of sex chromosome and sex-linked genes—such as CHRDL1 and TNMD—also may regulate lipid metabolism [66–68]. Thus, both sex hormones and placental biological sex may underlie the sex differences in lipid metabolism observed in early SPTB placentas. Additionally, inflammation and mitochondrial dysfunction can lead to placental dysfunction and preterm birth. Multiple genes important for regulating immune function and inflammatory responses are X-chromosome encoded, including TLR7, TLR8, CD40LG, IL2RG, and IRAK1 [69]. Moreover, numerous X-linked genes critical for mitochondrial function and redox homeostasis—such as MAOA, NOX1, COX7B, and PRDX4—can escape from X-inactivation and show female-biased expression [70].

Our study is the first to demonstrate race and fetal sex differences in the metabolome and transcriptome in early SPTB placentas. Another strength of our study is the use of normal mid-gestation placentas as gestational age controls, which allows us to distinguish true alterations associated with preterm birth from changes due to placental development and growth. Additionally, our samples were collected from different studies, but the protocols used and the placental locations for biopsy sampling were the same between the two institutions. We did not observe significant differences in the placental metabolome and transcriptome between samples obtained from different institutions, further ensuring minimal variation in tissue sampling. Although we are unable to parse out cell type-specific differences using whole tissues in our study, we have identified marked differences in lipid metabolic pathways both in the transcriptome and metabolome, which may cause energy failure, inflammation, changes in angiogenesis, and early senescence leading to placental dysfunction and SPTB.

Our findings document that alterations in lipid metabolism are associated with adverse pregnancy outcomes [71, 72]. The observed race and fetal sex differences in placental lipid metabolism and gene expression provide new insights into the underlying mechanisms of disparities in pregnancy outcomes and development of race- and fetal sex-specific preventative treatments for preterm birth.

Acknowledgment

The authors would like to thank Dr. Charles E. Chalfant (University of Virginia) and Dr. Harry Ischiropoulos (Children’s Hospital of Philadelphia) for the discussion and insights into lipid metabolism, the clinical coordinators at the University of Pennsylvania in the collection of the placentas, and Ms. Aria Huang for sample preparation. Figure 3 was created in BioRender.com.

Author contributions

Y.-C.L., R.A.S., S.P., and J.S.III: conceptualization; Y.-C.L., R.L., V.D.W.: methodology; Y.-C.L., and R.A.S: validation; M.K.: formal transcriptome analysis; J.P.G.: metabolomic data organization; Y.-C.L., B.E.H., and R.A.S.: data curation; Y.-C.L.: writing—original draft preparation; R.A.S., S.P., and J.S.III: writing—review & editing; Y.-C.L., R.L., and R.A.S: project administration; R.A.S. and S.P: funding acquisition. All authors have read and agreed to the published version of the manuscript.

Conflict of interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding

Funding for the experiments on SPTB and Term birth placentas collected at delivery (part of the CRIB study) was provided (in part) by the March of Dimes to the Prematurity Research Center at the University of Pennsylvania (R.A.S. and S.P.). The TrISecT study was funded by the Hallam Hurt Endowed Professorship (R.A.S.), Funding for sample acquisition was also provided by March of Dimes Foundation Basil O’Conner Award (V.D.W) and R01HD60723 (V.D.W).

Data availability

RNA-seq data sets are deposited in the NCBI GEO (accession GSE251663). Metabolomic data sets are deposited in Figshare (10.6084/m9.figshare.27287829).

Footnotes

References