Placental dysfunction influences fetal monocyte subpopulation gene expression in preterm birth

Abstract

The placenta is the primary organ for immune regulation, nutrient delivery, gas exchange, protection against environmental toxins, and physiologic perturbations during pregnancy. Placental inflammation and vascular dysfunction during pregnancy are associated with a growing list of prematurity-related complications. The goal of this study was to identify differences in gene expression profiles in fetal monocytes - cells that persist and differentiate postnatally - according to distinct placental histologic domains. Here, by using bulk RNA-Seq, we report that placental lesions are associated with gene expression changes in fetal monocyte subsets. Specifically, we found that fetal monocytes exposed to acute placental inflammation upregulate biological processes related to monocyte activation, monocyte chemotaxis, and platelet function, while monocytes exposed to maternal vascular malperfusion lesions downregulate these processes. Additionally, we show that intermediate monocytes might be a source of mitogens, such as HBEGF, NRG1, and VEGFA, implicated in different outcomes related to prematurity. This is the first study to our knowledge to show that placental lesions are associated with unique changes in fetal monocytes and monocyte subsets. As fetal monocytes persist and differentiate into various phagocytic cells following birth, our study may provide insight into morbidity related to prematurity and ultimately potential therapeutic targets.

Keywords: Adaptive immunity; Chemokines; Monocytes; Reproductive Biology; Vascular Biology.

Figure 1. Hierarchical clustering on flow cytometry data revealed 2 major clusters.

Group A was characterized by prevalence T cells and B cells, and Group B was characterized by increased abundance of neutrophils.

Figure 2. Cell composition by placental pathology domains.

(A–F) There were no significant differences in cell composition when comparing neutrophils, eosinophils, NK Cells, CD34⁺, B cells, or T cells between different domains of placental pathology (none = 15 cases, AI = 13 cases, CI = 7 cases, MVM = 33 cases, FVM = 2 cases).

Figure 3. Monocyte subpopulations by placental pathology domains.

(A–D) There were no significant differences in total population size when comparing different placental domains (none = 15 cases, AI = 13 cases, CI = 7 cases, MVM = 33 cases, FVM = 2 cases). For a total of 70 samples, we analyzed monocyte subtype composition and only found significant differences in the nonclassical subtype. These differences were between the AI to MVM (*P = 0.004), AI to FVM (**P = 0.005), and None to FVM (**P = 0.03) based on Kruskall-Wallis rank sum test.

Figure 4. Principal component analysis of intermediate and classical cord blood monocytes global gene expression.

A principal component analysis (PCA) of all samples (n = 130) illustrates GA (PC1, 15% variance) and cell type (PC2, 11% variance) and explains major variability within the data set.

Figure 5. Classical monocyte gene expression clusters by GA, clinical phenotypes, and placental domains.

Classical monocyte (n = 64) hierarchical clustering shows clustering primarily based on GA, placental inflammation, and placental vascular lesions. Cluster 1 included genes related to leukocyte activation, Cluster 2 contained genes involved in biosynthetic processes, and Cluster 3 contained genes related to cell adhesion and locomotion. There were 498 differentially expressed genes based on DeSeq2 using FDR < 0.05 and 2-fold change.

Figure 6. WGCNA of classical monocytes reveals gene clusters based on GA and mode of delivery.

WGCNA of n = 64 samples revealed ontologies related to immune cell activation, migration, and chemokine response were unregulated in term vaginal deliveries (Module 9 [M9]). Genes related to inflammation and inflammatory response were upregulated in term vaginal and preterm vaginal deliveries and were relatively downregulated in preterm cesarean deliveries (M6 and M7). Genes related to neutrophil activity and granulocytes were upregulated in term preterm vaginal deliveries and downregulated in term vaginal deliveries (M2). Classical monocytes from preterm vaginal deliveries upregulated genes associated with DC interaction compared with monocytes isolated from preterm caesarean deliveries (M3).

Figure 7. Placental domains and their associated clinical phenotypes lead to clustering of classical monocyte gene expression.

Classical monocytes samples (n = 34) based on placental domain with the associated clinical phenotype, broken down into 3 hierarchical clusters. The groups comprised primarily of term infants (Group A), preterm infants with acute chorioamnionitis and placental inflammation (Group B), and preterm infants with placental vascular lesions and preeclampsia (Group C). There were 2219 differentially expressed genes based on an FDR < 0.05 and 2-fold change.

Figure 8. WGCNA of placental domains and their associated clinical phenotypes reveal unique cluster of classical monocyte gene expression.

WGCNA of classical monocyte samples (n = 34) with available placental pathology and pregnancies complicated by clinical findings associated with that placental pathology reveals modules that are unique to each group. Genes related to platelet interactions were present in Module 1 (M1) and highly expressed in the inflammation and term groups. M2 contained genes related to inflammatory cytokines based on GO biological processes that were relatively upregulated in the preterm group. Genes related to antigen presentation were found in M3 and were relatively upregulated in the term group. Genes related to processes like artery development, such as VEGFA, were expressed in term neonates but were lower in preterm neonates exposed to placental pathology.

Figure 9. Gene expression of classical monocytes from preterm dyads cluster based on associated clinical findings and placental pathology.

Classical monocytes (n = 23) with known placental pathology and associated clinical features were divided into 2 hierarchical clusters. Group A was composed of preterm deliveries with chorioamnionitis and placental inflammation. Group B was composed of preterm deliveries complicated by preeclampsia MVM lesions. Genes related to granulocyte function were upregulated in Group A, while genes related to cell proliferation were upregulated in Group B. Based on an FDR < 0.05 and 2-fold change, there were 598 differentially expressed genes using DeSeq2.

Figure 10. Intermediate monocyte gene expression clusters by GA, clinical phenotypes, and placental domains.

Similar to classical monocytes, intermediate monocyte samples (n = 66) show hierarchical clustering based primarily based on GA. Cluster 1 included genes related to angiogenesis and cytokine production. Cluster 2 contained genes involved in biosynthetic processes, and Cluster 3 contained genes related to leukocyte activation, chemotaxis, and defense response. Differentially expressed genes were assessed with DeSeq2 using FDR < 0.05 and 2-fold change.

Figure 11. WGCNA of intermediate monocytes reveals gene clusters based on GA and mode of delivery.

WGCNA analysis of intermediate monocyte gene expression, like classical monocytes, showed that genes related to chemokine signaling were upregulated in term deliveries (M1). Genes related to immune cell migration were highly expressed the vaginal delivery group (M2). The gene ontologies in this module were like those in classical monocytes obtained from vaginal deliveries (Figure 6; M7) and were most highly expressed in the preterm vaginal delivery group.

Figure 12. Placental domains and their associated clinical phenotypes lead to clustering of intermediate monocyte gene expression.

There were 3 major groups based on hierarchical clustering of intermediate monocyte gene expression (n = 66). The groups were composed primarily of term infants (Group B), preterm infants with acute chorioamnionitis and placental inflammation (Group C), and preterm infants with MVM and preeclampsia (Group A). There were 3549 differentially expressed genes based on DeSeq2, using FDR < 0.05 and 2-fold change. The genes in Cluster 1 were involved in inflammatory biological processes and cell signaling. The genes in Cluster 2 were involved in cell proliferation and cell adhesion. Cluster 3 contained genes related to cytokine signaling and leukocyte degranulation. Cluster 4 and Cluster 5 contained genes involved in transcription and translational processes.

Figure 13. WGCNA of placental domains and their associated clinical phenotypes reveal unique cluster of intermediate monocyte gene expression.

WGCNA of intermediate monocytes samples (n = 34) with available placental pathology and pregnancies complicated by clinical findings associated with that placental pathology reveals 5 modules. The genes Module 1 (M1), M2, and M3 were highly expressed in the inflammation group. The genes in this group were associated with biological processes involved in monocyte chemotaxis, cell adhesion, and hemostasis. M4 had higher levels of expression in intermediate monocytes from the preterm group with vascular lesions and healthy premature infants. These genes were related to IFN signaling and signal transduction.

Figure 14. Chronic placental inflammation with preeclampsia leads to unique clustering of intermediate monocyte gene expression.

Intermediate monocytes (n = 23) with known placental pathology and associated clinical features also divided into 3 hierarchical clusters. Group A was composed of preterm deliveries complicated by preeclampsia and placental vascular lesions. Group B was composed of preterm deliveries complicated by preeclampsia and placental vascular lesions, and this group’s patients also had high degree chronic inflammatory lesions. Group C was composed of preterm deliveries with chorioamnionitis and placental inflammation. Genes related to leukocyte activity, function, migration, and degranulation were upregulated in Group C. Genes related to lipid metabolism and organic cyclic compound processing were upregulated in Groups A and B. There were 1852 differentially expressed genes based on an FDR < 0.05 and 2-fold change using DeSeq2.

Figure 15. Nonclassical monocyte gene expression clusters by GA.

Nonclassical monocytes (n = 24) divided into 2 major hierarchical cluster. One cluster was composed primarily of samples obtained from preterm deliveries with MVM on placental pathology (Group A). The second cluster contained samples from term deliveries with presumed normal placental pathology (Group B). Based on DeSeq2, 1938 genes were differentially expressed using FDR < 0.05 and 2-fold change. There were 2 major gene clusters based on k-means clustering. The genes in Cluster B were upregulated in Group A, and these genes were related to leukocyte adhesion to vascular endothelial cells and leukocyte tethering.

Figure 16. VEGFA gene expression levels positively correlate with VEGFA plasma serum levels and GA.

Plasma VEGFA and monocyte gene expression of VEGFA was found to be significantly increased in samples obtained from term deliveries relative to samples obtained from preterm deliveries exposed to MVM (n = 24). Plasma VEGFA and monocyte expression of VEGFA was also significantly higher in samples exposed to inflammation compared with MVM. P < 0.002 using Bonferroni’s correction was used to determine significance for serum VEGFA concentration.