Fig. 3: SAR progression significantly influences maternal–fetal interface composition.

a, Characteristics of SAR stages 1–5 manually assessed. b, MIBI colour overlay of manually assessed stage 1 arteries. Representative image of n = 70 FOVs. c, Inset of b showing stage 1 arteries. d, MIBI colour overlay of manually assessed stage 2 arteries. Representative image of n = 98 FOVs. e, Inset of d. Arrowhead indicates swollen endothelial cells. f, MIBI colour overlay of manually assessed stage 3 arteries. Representative image of n = 29 FOVs. g, Inset of f. Arrowhead indicates substantial loss of smooth muscle. h, MIBI colour overlay of one manually assessed stage 4 artery. Representative image of n = 21 FOVs. i, Inset of h. Arrowheads indicate intravascular EVTs. j, MIBI colour overlay of one manually assessed stage 5 artery. Representative image of n = 20 FOVs. k, Inset of j. Arrowhead indicates endothelialized EVTs lining the spiral artery lumen. l, Distribution of SAR manually assessed stages by GA. GA in days is binned to weeks for visualization. m, Schematic of calculating the continuous SAR remodelling score (δ). n, Volcano plot distinguishing GA-driven from SAR (δ)-driven cell-type frequencies. x axis, log₂ ratio of R² derived from linear regression against SAR (δ) and GA. y axis, –log₁₀ of the P value for the better-fitting regression model. Points are colour-coded by the trend size observed in the better-fitting regression model. o, Left, proportion of genes in artery (2,932 in total) tissue where expression changes significantly correlate with SAR (δ) (1,785), GA (517) or both (Sync; 633). Centre, SAR (δ)-correlated genes in artery tissue showing mean normalized expression (Z-score) by SAR (δ) stage. Right, two SAR (δ)-trending gene ontology pathways showing normalized expression of genes in the GO pathway by SAR (δ). Data are presented as the mean gene expression ± s.e.m.

Source Data