Pregnancy-associated systemic gene expression compared to a pre-pregnancy baseline, among healthy women with term pregnancies

Background

Pregnancy is known to induce extensive biological changes in the healthy mother. Little is known, however, about what these changes are at the molecular level. We have examined systemic expression changes in protein-coding genes and long non-coding (lnc) RNAs during and after pregnancy, compared to before pregnancy, among healthy women with term pregnancies.

Methods

Blood samples were collected from 14 healthy women enrolled in our prospective pregnancy cohort at 7 time-points (before, during and after pregnancy). Total RNA from frozen whole blood was used for RNA sequencing. Following raw read alignment and assembly, gene-level counts were obtained for protein-coding genes and long non-coding RNAs. At each time-point, cell type proportions were estimated using deconvolution. To examine associations between pregnancy status and gene expression over time, Generalized Estimating Equation (GEE) models were fitted, adjusting for age at conception, and with and without adjusting for changes in cell type proportions. Fold-changes in expression at each trimester were examined relative to the pre-pregnancy baseline.

Results

Numerous immune-related genes demonstrated pregnancy-associated expression, in a time-dependent manner. The genes that demonstrated the largest changes in expression included several that were neutrophil-related (over-expressed) and numerous immunoglobulin genes (under-expressed). Estimated cell proportions revealed a marked increase in neutrophils, and less so of activated CD4 memory T cells, during pregnancy, while most other cell type proportions decreased or remained unchanged. Adjusting for cell type proportions in our model revealed that although most of the expression changes were due to changes in cell type proportions in the bloodstream, transcriptional regulation was also involved, especially in down-regulating expression of type I interferon inducible genes.

Conclusion

Compared to a pre-pregnancy baseline, there were extensive systemic changes in cell type proportions, gene expression and biological pathways associated with different stages of pregnancy and postpartum among healthy women. Some were due to changes in cell type proportions and some due to gene regulation. In addition to providing insight into term pregnancy among healthy women, these findings also provide a "normal" reference for abnormal pregnancies and for autoimmune diseases that improve or worsen during pregnancy, to assess deviations from normal.

Enhancer Associated Long Non-coding RNA Transcription and Gene Regulation in Experimental Models of Rickettsial Infection

Recent discovery that much of the mammalian genome does not encode protein-coding genes (PCGs) has brought widespread attention to long noncoding RNAs (lncRNAs) as a novel layer of biological regulation. Enhancer lnc (elnc) RNAs from the enhancer regions of the genome carry the capacity to regulate PCGs in cis or in trans. Spotted fever rickettsioses represent the consequence of host infection with Gram-negative, obligate intracellular bacteria in the Genus Rickettsia. Despite being implicated in the pathways of infection and inflammation, the roles of lncRNAs in host response to Rickettsia species have remained a mystery. We have profiled the expression of host lncRNAs during infection of susceptible mice with R. conorii as a model closely mimicking the pathogenesis of human spotted fever rickettsioses. RNA sequencing on the lungs of infected hosts yielded reads mapping to 74,964 non-coding RNAs, 206 and 277 of which were determined to be significantly up- and down-regulated, respectively, in comparison to uninfected controls. Following removal of short non-coding RNAs and ambiguous transcripts, remaining transcripts underwent in-depth analysis of mouse lung epigenetic signatures H3K4Me1 and H3K4Me3, active transcript markers (POLR2A, p300, CTCF), and DNaseI hypersensitivity sites to identify two potentially active and highly up-regulated elncRNAs NONMMUT013718 and NONMMUT024103. Using Hi-3C sequencing resource, we further determined that genomic loci of NONMMUT013718 and NONMMUT024103 might interact with and regulate the expression of nearby PCGs, namely Id2 (inhibitor of DNA binding 2) and Apol10b (apolipoprotein 10b), respectively. Heterologous reporter assays confirmed the activity of elncRNAs as the inducers of their predicted PCGs. In the lungs of infected mice, expression of both elncRNAs and their targets was significantly higher than mock-infected controls. Induced expression of NONMMUT013718/Id2 in murine macrophages and NONMMUT024103/Apol10b in endothelial cells was also clearly evident during R. conorii infection in vitro. Finally, shRNA mediated knock-down of NONMMUT013718 and NONMMUT024103 elncRNAs resulted in reduced expression of endogenous Id2 and Apl10b, demonstrating the regulatory roles of these elncRNAs on their target PCGs. Our results provide very first experimental evidence suggesting altered expression of pulmonary lncRNAs and elncRNA-mediated regulation of PCGs involved in immunity and during host interactions with pathogenic rickettsiae.