Characterizing immune variation and diagnostic indicators of preeclampsia by single-cell RNA sequencing and machine learning

Preeclampsia is a multifactorial and heterogeneous complication of pregnancy. Here, we utilize single-cell RNA sequencing to dissect the involvement of circulating immune cells in preeclampsia. Our findings reveal downregulation of immune response in lymphocyte subsets in preeclampsia, such as reduction in natural killer cells and cytotoxic genes expression, and expansion of regulatory T cells. But the activation of naïve T cell and monocyte subsets, as well as increased MHC-II-mediated pathway in antigen-presenting cells were still observed in preeclampsia. Notably, we identified key monocyte subsets in preeclampsia, with significantly increased expression of angiogenesis pathways and pro-inflammatory S100 family genes in VCAN + monocytes and IFN + non-classical monocytes. Furthermore, four cell-type-specific machine-learning models have been developed to identify potential diagnostic indicators of preeclampsia. Collectively, our study demonstrates transcriptomic alternations of circulating immune cells and identifies immune components that could be involved in pathophysiology of preeclampsia. A study profiled transcriptomic characteristic of circulating immune cells in preeclampsia pregnancy by scRNA-seq, then develop cell-type-specific models to identify potential diagnostic indicators of preeclampsia by machine learning.