A spatiotemporal single-cell transcriptomic atlas of mouse placentation (E9.5–E18.5).

(A) Schematic illustration of the placenta samples collected for spatial transcriptomics profiling. (B) Uniform manifold approximation and projection (UMAP) of snRNA-seq clustering showing 35 cell types. GC, glycogen cell; JZP, junctional zone progenitor; LaTP, labyrinth trophoblast progenitor; P-TGC, parietal trophoblast giant cell; S-TGC, sinusoid trophoblast giant cell; SpT, spongiotrophoblast cell; SynT, syncytiotrophoblast cell; DSC, decidual stromal cell; EC, endothelial cell; DC, dendritic cell; NK, natural killer cell. (C) Spatial visualization of all cell distributions within placentas at different developmental stages from E9.5 to E18.5 using Stereo-seq data. Cells are colored by their annotations. Scale bars, 500 μm. (D) Cell type distribution and quantification in placenta sections from E9.5 to E18.5. Anatomical regions (left) were identified as Labyrinth zone, Junctional zone (JZ), Maternal decidua, and Myometrium. Regions are colored based on anatomical region annotations. Cells are colored based on cell type annotations. Stacked area plots showing cell type proportions in the corresponding regions. Scale bars, 500 μm.

Molecular, spatial, and developmental characterization of GC subtypes.

(A) UMAP visualization showing the reclustering of trophoblast cells. (B) Spatial visualization of all cell distributions in the E14.5 placenta section. Two magnified fields of view highlight the distinct spatial distribution of the two GC subclusters. Scale bars, 500 μm. (C) Heatmap (left) showing the DEGs of GC subclusters, with significantly enriched GOs shown by bubble plot (right). Shared genes and GOs are outlined in red, GC-1-specific genes and GOs are outlined in yellow, and GC-2-specific genes and GOs are outlined in pink. (D) Heatmap (left) showing the SCENIC transcription factor (TF) regulon scores in GC-1 and GC-2 based on regulon activity. Spatial visualization of selected regulons, Anxa11 and Prmd1, are shown respectively in the E14.5 placenta section (right). (E) Line chart (left) showing the percentage of the two GC subclusters in the maternal decidua and JZ region from E12.5 to E18.5. The spatial visualization of the two GC subclusters at E13.5, E15.5, E16.5 and E18.5 is shown on the right. Cells are colored by their annotations. Scale bars, 500 μm. (F) UMAP visualization showing cell types across GC differentiation, including junctional zone progenitor (JZP), GC precursor, and two GC subclusters. Cells are colored by cell type annotation. (G) Pseudotime trajectory of cell types across GC differentiation, analyzed using Monocle3 and CytoTRACE. Cells are colored by pseudotime. (H) UMAP visualization showing the origin of GC based on the time point. (I) UMAP visualization of cell types across GC differentiation. Cells are colored by cell type annotation.

All lethal mouse knockout lines according to previous reports.

All genes are categorized according to the time of lethality. They are roughly divided into postnatal lethality and lethality at various stages of embryonic development. The table also provides information on whether placental phenotype defects have been reported, along with the corresponding references. Lethal, no homozygotes are recovered. Subviable, homozygotes recovered but at less or equal to 13% (≤13%). Viable, homozygotes recovered more than 13% (>13%).

Ano6 deficiency disrupts placental development and increases GC abundance.

(A) Representation of the curated list of embryonic lethal mouse mutant genes (orange denotes abnormality detected). The selected mutant genes are categorized by lethality stage and marked with corresponding placental phenotypes. (B) Pseudo-temporal expression dynamics of specific representative genes along the GC developmental trajectory. Genes previously identified to be associated with lineage development are labeled in black, while potential regulators whose loss leads to embryonic lethality are highlighted in red. (C) Detection of Ano6 and Tpbpa protein synthesis by RIBOseq in E13.5 placenta sections. Dotted lines encircle the same region. Scale bars, 500 μm. (D) Schematic illustration of the experimental design. (E) Representative images of placentas from WT, HET, and KO mice at embryonic day 18.5 (E18.5). Black arrows point to the phenotypic defects in the KO placenta. Ano6 KO placentas are smaller than WT and HET, and exhibit prominent white plaques on the fetal side, indicative of vascular structural defects. (F) Quantification of the weight and diameter of the WT, HET, and KO placenta. WT (n=7), HET (n=11), KO (n=7). One-way analysis of variance (ANOVA). *p < 0.05. All data represent means ± SEM. (G) UMAP representation of all cell types in E18.5 WT, HET, and KO placentas. (H) UMAP visualization of snRNA-seq clustering showing the 25 cell types in E18.5 WT, HET, and KO placenta samples. (I) Spatial visualization of all cell distributions in Stereo-seq data. Cells are colored by their annotations. Scale bars, 500 μm. (J) Bubble plot showing the percentage of GC subclusters in E18.5 WT, HET, and KO placentas. (K) Spatial visualization of macrophages in E18.5 WT, HET, and KO placentas. Cells are colored by their annotations. Scale bars, 500 μm. (L) Bar plot showing the percentage of macrophages in the labyrinth of E18.5 WT, HET, and KO placentas. (M) F4/80 immunofluorescence staining of E18.5 WT, HET, and KO placentas. Cross sections of the entire placentas and enlarged views are shown, respectively. Scale bars, 500 μm (overall shape) and 20 μm (high-magnification view). (N) The proportion of F4/80-positive cells in the labyrinth is quantified in WT and KO placentas. WT (n=5), HET (n=5), and KO (n=8). One-way analysis of variance (ANOVA). **p < 0.01, ***p < 0.001. All data represent means ± SEM.

Genotype distribution of Ano6 knockout mice.

GC persistence and excessive glycogen accumulation in Ano6-null placentas.

(A) Overall shape and high-magnification images of the PAS-stained WT, HET and KO placentas. Scale bars, 500 μm (overall shape) and 10 μm (high-magnification view). Quantification was performed on GCs located in both the junctional zone (JZ) and decidua. (B) Percentage of PAS-stained positive area relative to the total tissue area analyzed using ImageJ. WT (n=6), HET (n=6), KO (n=6). One-way analysis of variance (ANOVA). *p < 0.05, **p < 0.01. All data represent means ± SEM. (C) Schematic illustration of glycogen cells scanned by transmission electron microscope (TEM). (D) Electron micrographs of WT, HET, and KO placentas at E18.5. The outline of glycogen cells is circled with white dashed lines. GC, glycogen cells; N, nucleus; Gly, glycogen granules. Scale bars, 500 μm (column 1), 200 μm (column 2), 500 nm (column 3), and 200 nm (column 4). (E) Glycogen granule density in WT, HET, and KO placentas at E18.5. Glycogen granule number per μm2 was calculated. WT (n=6), HET (n=6), KO (n=6). One-way analysis of variance (ANOVA). ***p < 0.001. All data represent means ± SEM.(F) Total placental glycogen content (mg) in WT, HET, and KO placentas at E18.5. WT (n=6), HET (n=6), KO (n=6). One-way analysis of variance (ANOVA). *p < 0.05. All data represent means ± SEM. (G) Placental glycogen expressed as a percentage of placental weight in WT, HET, and KO placentas. WT (n=6), HET (n=6), KO (n=6). One-way analysis of variance (ANOVA). **p < 0.01. All data represent means ± SEM.

Defective GC glycogen degradation reduces fetal energy supply and survival.

(A) Schematic of glycogenolysis pathway. (B) Schematic workflow for metabolites analysis in placenta and fetal liver by LC-MS. (C and D) Relative abundance of Glucose, G1P and G6P in placental (C) and fetal liver tissues (D) measured by targeted metabolomics. Data are presented as mean ± SEM from four independent experiments. Statistical significance was assessed using One-way analysis of variance (ANOVA). **p < 0.01, ***p < 0.001. (E) Schematic of maternal glucose supplementation strategy. (F) Significant increase survival of Ano6-/- mice after maternal glucose supplementation. * p < 0.05, Fisher’s exact test. (G and H) Relative abundance of Glucose, G1P and G6P in placental (G) and fetal liver tissues (H) after glucose supplementation measured by targeted metabolomics. Data are presented as mean ± SEM from four independent experiments. Statistical significance was assessed using One-way analysis of variance (ANOVA). *p < 0.05, **p < 0.01, ***p < 0.001.

Genotype distribution of Ano6 knockout mice after glucose supplementation with pregnant mice.

Secondary macrophage accumulation and immune activation in the labyrinth.

(A) Heatmap showing the upregulated genes in macrophages within the labyrinth of KO placenta compared with macrophages within the maternal region. (B) Bar plot showing the enriched GO terms of differentially upregulated genes in macrophages within the labyrinth of KO placentas. (C) Spatial visualization of representative genes in enriched GO terms in the KO placenta section. Scale bars, 500 μm. (D) Co-immunostaining of F4/80 and TGFβ on KO placentas. Scale bars, 1000 μm (overall shape) and 20 μm (high-magnification view). (E) Spatial visualization of C3ar1 in the WT, HET and KO placenta sections. Scale bars, 500 μm.

Experimental design and quality control analysis of snRNA-seq and Stereo-seq dataset, related to

Figure 1 (A) Bright field images of samples from different developmental stages collected for sequencing experiments. White dashed lines indicate the collected placenta samples. (B) Violin plots showing the number of reads and genes, and percentage of mitochondrial genes in the snRNA-seq data. (C) Quality control of Stereo-seq data on cell number, UMI count and gene count. Stereo-seq spot overlay (cellbin) showing number of reads and genes. Scale bars, 500 μm. (D) Heatmap showing top gene expression of all identified cell types. (E) Integration of snRNA-seq data in this study with published datasets.

Anatomical regions and marker gene expression in Stereo-seq, related to

Figure 1 (A) Anatomical regions were identified. Regions were colored based on anatomical region annotation. Scale bars, 500 μm. (B) UMAP displaying snRNA-seq data from different time points. (C) Heatmap showing marker gene expression of all identified cell types using Stereo-seq dataset. (D) Schematic diagram of annotated placental regions based on our scStereto-seq data analysis.

Molecular differences between two GC subclusters, related to

Figure 2 (A) Violin plot showing top genes of GC-1 and GC-2 in all trophoblast cells. (B) Spatial expression of top genes, Aldh1a3 and Prl7b1, in GC-1 and GC-2 on an E14.5 placental section, respectively. Scale bars, 500 μm. Dotted lines encircle the regional boundaries. Scale bars, 500 μm. (C) Detection of Aldh1a3 and Prl7b1 protein synthesis by RIBOseq in E14.5 and E18.5 placental sections. Dotted lines encircle the regional boundaries. Scale bars, 500 μm. (D) Spatial visualization of selected regulons in GC-1 and GC-2 on E14.5 placental sections, respectively. (E) Chord (left) diagram showing the interactions between GC-1 (sender) and other cell types (receiver). Heatmap (right) showing the selected ligand-receptor interactions. (F) Chord (left) diagram showing the interactions between GC-2 (sender) and other cell types (receiver). Heatmap (right) showing the selected ligand-receptor interactions. (G) The spatial visualization of the two GC subclusters from E12.5 to E18.5. Cells are colored by their annotation. Scale bars, 500 μm. Inner dotted lines represent the boundary of decidua and JZ. Scale bars, 500 μm.

Cellular trajectory analysis identifies genes with temporally regulated expression patterns along pseudotime, related to

Figure 3 (A) Cellular trajectory reconstruction of trophoblast cells using the Monocle3 and CytoTRACE. (B) UMAP visualization showing the origin of trophoblasts based on the time point. (C) Z-score heatmap of gene expression in different branches, where rows are genes and columns are cells ranked by pseudotime value. Genes were first fit with Moran’s I test with ranked pseudotime as independent variable. Genes with the most significant time dependent model (q_value=0 and morans_I>0.25) were extracted and clustered hierarchical clustering. Cells were ordered according to scaled pseudotime value from 0 to 1. Genes previously identified to be associated with lineage development are labeled in black, while potential regulators whose loss leads to embryonic lethality are highlighted in red.

Expression pattern of embryonic lethal genes in placental cells, related to

Figure 3 (A) Heatmap showing the expression patterns of 151 genes listed in Table 1 in annotated cell types across E9.5-18.5 placentas (snRNA-seq data). The Ano6 gene is highlighted with a red box. (B) Heatmap showing the normalized expression patterns of top 98 genes selected from Table 1 (based on the expression threshold) in annotated cell types across E9.5-18.5 placentas (snRNA-seq data). The mutant genes are colored with categories by the timing of lethality. The mutants represented in blue are postnatal lethal; the mutants represented in pink are lethal before E8.5; the mutants represented in green are lethal between E9.5-E14.5; the mutants represented in brown are perinatal lethal; and the mutants represented in black are lethal with uncertain embryonic stage. The boxes indicated selected enriched patterns, and the corresponding specific cell types of placentas are also encircled below. The Ano6 gene is highlighted with a red box.

Characterization of Ano6 WT, HET and KO placentas, related to

Figure 3 (A) The expression changes of Ano6 gene in mouse embryos from E5.5 to E18.5 examined using quantitative PCR (qPCR), separately analyzing the expression in the fetal part and the placental part after E8.5. One-way analysis of variance (ANOVA). *p < 0.05, **p < 0.01, ***p < 0.001. All data represent means ± SEM. (B) Volcano plot of differentially expressed genes (DEGs) identified between GC and other trophoblast cells. The orange dots denote up-regulated gene expression, the blue dots denote down-regulated gene expression, and the gray dots denote gene expression without marked differences. Several up-regulated genes in GC related to embryonic lethality are highlighted. (C) Ano6 expression in placental cells across different developmental stages, as revealed by snRNA data. The dashed circle highlights the GC lineage. (D) Genotyping results of WT, HET and KO Ano6 placentas. (E) qPCR analysis of Ano6 gene in WT/HET/KO placentas at E18.5. One-way analysis of variance (ANOVA). ***p < 0.001. All data represent means ± SEM. (F) Representative images of placentas from WT, HET, and KO mice at E15.5, E16.5, and E18.5. Black arrows point to the phenotypic defects in the KO placenta. Scale bars, 2000 μm. (G) Quantification of white plaque area (mm²) on the fetal surface of the placenta across E14.5 to E18.5 in WT, HET, and KO mice. One-way analysis of variance (ANOVA). ***p < 0.001. Data are presented as mean ± SEM. (H) H&E staining of WT, HET, and KO placental sections at E15.5, E16.5, and E18.5. The black arrow indicates the location where phenotypic abnormalities appear. Scale bars, 500 μm. The dashed lines indicate the boundary of the labyrinth region. (I) H&E staining of WT/HET/KO placentas at E18.5. Scale bars, 500 μm (overall shape) and 100 μm (high-magnification view). (J) Statistics on the proportion of the labyrinth region in the total area of the placenta. One-way analysis of variance (ANOVA). ***p < 0.001. All data represent means ± SEM. (K) H&E staining of WT/HET/KO placentas (left, the labyrinth region was encircled), and statistical results of cell quantity and cell density in the labyrinth region of WT, HET, and KO placentas (right). Cell density was calculated as the number of nuclei divided by the labyrinth area. Scale bars, 500 μm. (L) CD31 immunofluorescence staining of the WT/HET/KO placentas at E18.5. CD31 labels fetal blood vessels. Scale bars, 500 μm (overall shape) and 20 μm (high-magnification view). (M) Percentage of fetal blood vessels in the labyrinth region. WT: n=7, HET: n=7, KO: n=7. One-way analysis of variance (ANOVA). ***p < 0.001. All data represent means ± SEM.

Quality control analysis and characterization of snRNA-seq and Stereo-seq data for E18.5 WT, HET and KO placentas, related to

Figure 3 (A) Violin plots showing the mitochondrial percentage, UMIs and genes in WT/HET/KO placentas (top) and in each cell type (bottom). (B) Quality check of Stereo-seq data. Stereo-seq spot overlay (cellbin) showing number of genes and reads. Scale bars, 500 μm. (C) Identification of anatomical regions on WT/HET/KO placentas. Regions are colored based on anatomical region annotation. Scale bars, 500 μm. (D) Bubble plot showing the percentage of all cell types in the E18.5 WT, HET, and KO placentas using Stereo-seq data. (E) Cell type distribution in E18.5 WT, HET, and KO placenta sections. Differences in Ano6 KO placenta compared with the WT/HET placentas are marked red. Cells were colored based on cell type annotation. Bar plots showing cell subcluster proportions in the corresponding regions. Scale bars, 500 μm.

Time-course analysis of placental glycogen content in the defective placenta, related to

Figure 4 (A) H&E staining of WT, HET and KO placentas. The black dashed line indicates the boundary between the JZ and decidua. The glycogen cells in JZ and decidua are characterized by vacuolated, glycogen-rich cytoplasm and appear as compact cell islets. Scale bars, 500 μm (overall shape) and 10 μm (high-magnification view). Quantification was performed on GCs located in both the junctional zone (JZ) and decidua. (B) The glycogen cell number in JZ and decidua per mm2. WT (n=4), HET (n=4), KO (n=4). One-way analysis of variance (ANOVA). **p < 0.01. All data represent means ± SEM. (C) Total placental glycogen content (mg) in WT, HET, and KO placentas across E13.5 to E18.5. One-way analysis of variance (ANOVA). *p < 0.05. Data are presented as mean ± SEM (n = 3-5 per group). (D) Glycogen content expressed as a percentage of total tissue weight (%) in placental tissues across the same developmental stages. *p < 0.05. (E) Percentage of PAS-stained positive area relative to the total tissue area analyzed using ImageJ. **p < 0.01.

The expression of GC markers and genes encoding key glycogenolytic enzymes by snRNA-seq of E18.5 placentas, related to

Figure 5 (A) Relative expression of GC-specific marker genes (Pcdh12, Aldh1a3, and Gjb3) in WT, HET, and KO placentas across development. One-way analysis of variance (ANOVA). **p < 0.01. Data are presented as mean ± SEM. (B) Schematic of glycogenolysis pathway enzymes and the genes that encode them. (C) The expression of genes encoding key enzymes at E18.5 WT, HET, and KO placentas by snRNA-seq data. The blue circles indicate the GC population.

Spatial distribution of differentially expressed genes in the labyrinth between KO and WT placentas, related to

Figure 6 (A) Volcano plot showing the differentially expressed genes (DEGs) in the labyrinth of KO placentas compared with WT placentas. (B) Spatial visualization of representative upregulated genes in enriched GO terms on the WT, HET and KO placental sections. Scale bars, 1 mm. (C) Spatial visualization of representative downregulated genes in enriched GO terms in the WT, HET and KO placenta sections. Scale bars, 1 mm. (D) Two-sided bar plot showing the enriched GO terms of differentially upregulated and downregulated genes in the labyrinth of KO placentas.

Spatial visualization of unregulated genes in the labyrinth of KO placentas compared with WT placentas, related to

Figure 6 Spatial visualization of representative upregulated genes in enriched GO terms on the WT, HET and KO placental sections. Scale bars, 1 mm.

Spatial visualization of downregulated genes in the labyrinth of KO placentas compared with WT placentas, related to

Figure 6 Spatial visualization of representative downregulated genes in enriched GO terms on the WT, HET and KO placental sections. Scale bars, 1 mm.

Aberrantly localized macrophages are closely linked to the emergence of vascular abnormalities, related to

Figure 6 (A) Co-immunostaining of F4/80 and anti-inflammatory cytokine (IL-10), inflammatory cytokines (IL-12 and TNF-α). Scale bars, 1000 μm (overall shape) and 20 μm (high-magnification view). (B) Immunostaining of CD31 and F4/80 on adjacent placental sections. Scale bars, 1000 μm (overall shape) and 50 μm (high-magnification view).