Discrete placental gene expression signatures of diabetes during pregnancy

BACKGROUND: Although gestational diabetes affects up to 10% of all pregnancies, there is ambiguity as to whether the disease subtype classifications are biologically significant at the maternal-fetal interface, or more reflective of an extended spectrum of normal pregnancy endocrine physiology. OBJECTIVE: Bulk RNA-sequencing (RNA-seq), single-cell RNA-sequencing (scRNA-seq), and spatial transcriptomics have been integrated to reveal gene signatures of disease in subsets of cells and microenvironments. We aimed to combine these high-resolution technologies with rigorous classification of diabetes subtypes in pregnancy. We hypothesized that differences between pre-existing and gestational diabetes subtypes would be associated with altered gene expression profiles in specific placental cell populations. STUDY DESIGN: In a two-phase case-cohort design, we clinically validated gestational diabetes mellitus type 1 (GDMA1), GDMA2, and type 2 diabetes (T2DM) cases within a cohort of placentae and compared them to healthy controls by bulk RNA-seq (n=54). Quantitative analysis with RT-qPCR of presumptive genes of significant interest were undertaken in an independent and non-overlapping validation cohort of similarly well-characterized cases and controls (n=122). Additional integrated analyses of term placenta single-cell, single-nuclei, and spatial transcriptomics data to determine the cellular subpopulations and niches that aligned with the GDMA1, GDMA2, and T2DM gene expression signatures. RESULTS: Dimension reduction of the bulk RNA-seq data revealed that the most common source of placental gene expression variation was diabetic disease subtype. Relative to controls, we found 2,052 unique significantly differentially expressed transcripts (-2<Log2(fold-change)>2 threshold; q<0.05 Wald Test) among GDMA1 placental specimens, 267 among GDMA2, and 1,520 among T2DM. Several candidate marker genes (CSH1, PER1, PIK3CB, FOXO1, EGFR, IL2RB, SOD3, DOCK5, and SOGA1) were validated in an independent and non-overlapping validation cohort (q<0.05 Tukey). Functional enrichment revealed the pathways and genes most impacted for each diabetes subtype, and degree of proximal similarity to other subclassifications. Surprisingly, GDMA1 and T2DM were more proximal by virtue of increased expression of chromatin remodeling and epigenetic regulation genes, while albumin was the top marker for GDMA2 with increased expression of genes in the wound healing pathway. Assessment of these gene signatures in the single-cell, single-nuclei, and spatial transcriptomics data revealed expression of these genes were highly variable by placental cell and microarchitecture type. For example, at the cellular and spatial (e.g., microarchitectural) levels, distinguishing features were observed in in extravillous trophoblasts (GDMA1) and macrophages (GDMA2). Lastly, we utilized these data to generate machine learning models to predict participants diabetes status and observed greater proximity of placental gene expression among GDMA1 and T2DM participants relative to GDMA2. CONCLUSION: Consistent with their distinct risks of perinatal outcomes, placentae from GDMA1, GDMA2, and T2DM affected pregnancies harbor gene signatures which can be further distinguished by placental microarchitecture and cellular subtypes.