Single-cell transcriptomics showed that maternal PCB exposure dysregulated ER stress-mediated cell type-specific responses in the liver of female offspring

Abstract

Polychlorinated biphenyls (PCBs) are persistent environmental toxicants that bioaccumulate in the food chain and readily cross the placenta, raising concerns for developmental toxicity. While PCB exposure has been associated with metabolic and neurodevelopmental disorders, its cell type-specific effects on liver development remain poorly understood. This study aimed to investigate how maternal exposure to an environmentally relevant Fox River PCB mixture affects liver development in female offspring at single-cell resolution. We hypothesized that early-life PCB exposure disrupts hepatic metabolic and immune function in a cell type-specific manner. Using single-cell RNA sequencing (scRNA-seq) on liver tissue from postnatal day 28 female mice perinatally exposed to PCBs, we identified major hepatic and immune cell populations and assessed cell-specific transcriptional responses. PCB exposure significantly altered the proportions of endothelial cells and Kupffer cells and reduced neutrophil abundance. Transcriptomic analysis revealed that PCBs dysregulated key pathways in hepatocytes and non-parenchymal cells, including ER stress responses, drug metabolism, and glucose/insulin signaling. Notably, hepatocytes exhibited upregulation of phase-I drug-metabolizing enzymes and uptake transporters, but downregulation of phase-II enzymes and efflux transporters. Kupffer cells and endothelial cells had altered immune and metabolic gene expression, and intercellular communication analysis predicted disrupted fibronectin, collagen, and chemokine signaling due to PCB exposure. RT-qPCR validation confirmed increased hepatic ER stress marker expression. Together these findings demonstrate that perinatal PCB exposure induces persistent, cell type-specific transcriptomic reprogramming in the liver, impairing metabolic and immune functions. This study highlights the utility of single-cell transcriptomics for revealing toxicant effects with cellular precision during critical windows of development.

Figure 1.

Experimental design of the study. Dams were exposed to peanut oil in peanut butter (vehicle) or the Fox River PCB mixture (6 mg/kg body weight) once daily until pups were weaned at PND 21. Using the livers of pups at PND 28, scRNA-seq (n = 3) was performed to profile the cell type-specific responses in liver or RT-qPCR (n = 10) to investigate whole liver mRNA expression changes.

Figure 2.

Clustering and visualization of marker genes in each cell type of the liver. A. Visualization of cell clusters with cell types identified using marker gene expression. B. Representation of key marker genes for cell type labeling. C. Changes in proportions of liver cell types following early life exposure to PCBs. Y-axis shows the fold change in each cell type of the PCB exposed group over that of the vehicle exposed group. Asterisks represent p-value (two-way t-test with unequal variance assumption). D. Total number of differentially expressed genes (Bonferroni-adjusted p-value < 0.05) for each liver cell type from perinatal exposure to PCBs.

Figure 3.

Top 5 up- and down-regulated gene ontology enrichment results in Kupffer cell (A), hepatocyte (B), endothelial cell (C), and myofibroblast (D) from perinatal exposure to PCBs. Dotted lines represent -log10 FDR-adjusted p-value at 0.05.

Figure 4.

Dysregulated expression signatures of drug-processing genes in resident hepatic cell populations following perinatal exposure to PCBs. A. Violin plots that describe the distribution of expression of prototypical target genes of aryl hydrocarbon receptor (AHR), pregnane X receptor (PXR), constitutive androstane receptor (CAR), and peroxisome proliferator-activated receptor alpha (PPARα) from perinatal exposure to PCBs. Asterisks represent Bonferroni-adjusted p-value < 0.05. B. Phase-I drug-processing enzymes that were differentially expressed (rows) in at least one of the major hepatic resident cell types (columns) with xenobiotic biotransformation capabilities are shown in a heatmap (left side). The colors of the heatmap represent the log2 fold change of liver genes of the PCB exposed pups as compared to the vehicle control. “Direction” indicates whether a gene is up- (orange) or down- (green) regulated from perinatal exposure to PCBs (Bonferroni-adjusted p-value < 0.05). C. Phase-II enzymes and transporters involved in xenobiotic metabolism processes that were differentially expressed (rows) in at least one of the major hepatic resident cell types (columns) with xenobiotic biotransformation capabilities are shown in a heatmap (left side). The colors of the heatmap represent the log2 fold change of liver genes of the PCB exposed pups as compared to the vehicle control. “Direction” indicates whether a gene is up- (orange) or down- (green) regulated from perinatal exposure to PCBs (Bonferroni-adjusted p-value < 0.05).

Figure 5.

A. Representation of up-regulated cell type-specific gene ontology terms related to insulin signaling and glucose homeostasis from perinatal PCB exposure. Red and blue show up- and down-regulated gene ontology terms, respectively. Color gradient indicates -log10 transformed FDR-adjusted p-value. B. Number of up- and down-regulated genes related to glucose and insulin signaling in hepatocytes. C. Down-regulated expression of glucose and insulin signaling markers in hepatocytes. Grey and red indicate low and high expression, respectively.

Figure 6.

A. Representation of up-regulated cell type-specific gene ontology terms related to insulin signaling and glucose homeostasis from perinatal PCB exposure. Red and blue show up- and down-regulated gene ontology terms, respectively. Color gradient indicates -log10 transformed FDR-adjusted p-value. B. Average expression of endoplasmic reticulum stress markers in hepatic cell types following perinatal PCB exposure. Red and blue colors represent up- and down-regulation, respectively. Vehicle and PCB-exposed groups are shown as circles and triangles, respectively. Asterisks indicate differential expression (Bonferroni-adjusted p-value < 0.05). C. Up-regulated mRNA expression of endoplasmic reticulum stress markers by RT-qPCR in livers perinatally exposed to PCBs. Asterisks show p-value < 0.05 (two-way t test with assumption of unequal variance).

Figure 7.

A. Significantly altered predicted cell-cell signaling pathways from perinatal PCB exposure (permutation test p-value < 0.05). Black and red highlights indicate significant predicted signaling pathways enriched in the vehicle and PCB exposed group, respectively. Predicted decrease in fibronectin (B) and increase in protease-activated receipts (PARs) (C) signaling in liver cell types from perinatal exposure to PCBs.

Figure 8.

Summary of key findings.