Transcriptomic comparison of human and mouse brain microvessels

Abstract

The brain vasculature maintains brain homeostasis by tightly regulating ionic, molecular, and cellular transport between the blood and the brain parenchyma. These blood-brain barrier (BBB) properties are impediments to brain drug delivery, and brain vascular dysfunction accompanies many neurological disorders. The molecular constituents of brain microvascular endothelial cells (BMECs) and pericytes, which share a basement membrane and comprise the microvessel structure, remain incompletely characterized, particularly in humans. To improve the molecular database of these cell types, we performed RNA sequencing on brain microvessel preparations isolated from snap-frozen human and mouse tissues by laser capture microdissection (LCM). The resulting transcriptome datasets from LCM microvessels were enriched in known brain endothelial and pericyte markers, and global comparison identified previously unknown microvessel-enriched genes. We used these datasets to identify mouse-human species differences in microvessel-associated gene expression that may have relevance to BBB regulation and drug delivery. Further, by comparison of human LCM microvessel data with existing human BMEC transcriptomic datasets, we identified novel putative markers of human brain pericytes. Together, these data improve the molecular definition of BMECs and brain pericytes, and are a resource for rational development of new brain-penetrant therapeutics and for advancing understanding of brain vascular function and dysfunction.

Figure 1

LCM of brain microvessels yields transcriptomic datasets enriched in endothelial and pericyte markers. (A,B) Representative lectin-stained mouse brain sections before (A) and after (B) LCM. Scale bars: 50 μm. (C,D) Transcript abundance [log₂(TPM + 1)] of endothelial, pericyte, and neural genes in mouse (C) and human (D) whole brain and LCM microvessel samples. Lines connect datapoints from matched LCM microvessel and whole brain samples. *P < 0.05 versus whole brain, DESeq2 Wald test with Benjamini–Hochberg correction. Exact P-values are provided in Supplementary Table S4.

Figure 2

Comparison of mouse LCM microvessel with existing mouse BMEC and pericyte transcriptomic datasets. (A) Comparison of transcript abundances [log₁₀(TPM)] in mouse LCM microvessels (y-axes; average of three biological replicates) versus the reference BMEC dataset or the reference pericyte dataset. Each point represents one gene with nonzero expression in both LCM microvessel and reference endothelial or pericyte datasets. The Pearson (r_p) and Spearman’s (r_s) correlation coefficients are inset and were calculated on raw TPM values prior to removal of undetected transcripts and log-transformation. Pseudocoloring indicates relative population density. (B) Pearson correlation coefficient (r_p) for the average of the biological replicates of mouse LCM microvessels calculated against reference datasets generated by combining brain pericyte and endothelial cell transcriptomes in different ratios (x-axis). (C) Comparison of transcript abundances [log₁₀(TPM)] in mouse LCM microvessels (y-axis; average of three biological replicates) versus the optimal combined reference dataset (42% pericyte transcript and 58% BMEC transcript weighting). Each point represents one gene with nonzero expression in both LCM microvessel and the reference dataset. The Pearson (r_p) and Spearman’s (r_s) correlation coefficients are inset and were calculated on raw TPM values prior to removal of undetected transcripts and log-transformation. Pseudocoloring indicates relative population density.

Figure 3

Differentially expressed genes in mouse and human LCM microvessels compared to whole brain. (A,B) Whole-transcriptome hierarchical clustering of mouse (A) and human (B) LCM microvessels and whole brain datasets. Color indicates expression that has been normalized within each gene (column). (C,D) Principal component analysis of mouse (C) and human (D) LCM microvessels and whole brain datasets. Data are plotted in the space of the first two principal components, with the percentage of variance explained by principal component 1 (PC1) and principal component 2 (PC2) shown in axis labels. Microvessel and whole brain datapoints of the same color are derived from matched samples. (E,F) Volcano plots illustrating genes differentially expressed between LCM microvessels and whole brain samples from mouse (E) and human (F). The number of LCM microvessel-enriched (Up) and depleted (Down) genes with adjusted P-values < 0.05 (from DESeq2) are shown in the legends. Full results of differential expression analysis are in Supplementary Table S4. (G,H) Heat maps illustrating transcript abundance in biological triplicates of LCM microvessels and whole brain for the 10 highest confidence LCM microvessel-enriched and the 10 highest confidence microvessel-depleted genes in mouse (G) and human (H). Color indicates expression that has been normalized within each gene (row). (I,J) Gene ontology (GO) terms for biological processes enriched in LCM microvessels compared to whole brain from mouse (I) and human (J). ES: enrichment score [− log₁₀(P)].

Figure 4

Mouse-human species differences in vasculature-associated gene expression. (A) Volcano plot illustrating genes differentially expressed between human LCM microvessels and mouse LCM microvessels. The number of human-enriched (Up) and depleted (Down) genes with adjusted P-values < 0.05 (from DESeq2) are shown in the legend. The complete list of mouse-human gene homology as used in this analysis, and full results of differential expression analysis, are in Supplementary Table S5. (B) Summary of filtering strategy used to identify human-mouse species differences potentially attributable to vascular gene expression. Of the 1,278 mouse-enriched transcripts identified in (A), Approach 1 selects the 142 genes also enriched in mouse microvessels versus whole brain (as determined in Fig. 3E). Of the 1,122 human-enriched transcripts identified in (A), Approach 1 selects the 211 genes also enriched in human microvessels versus whole brain (as determined in Fig. 3F). Complete filtered gene lists are in Supplementary Table S5. Alternative filtering strategies are shown in Supplementary Figure S5. (C,D) Heat maps illustrating transcript abundance in biological triplicates of mouse and human LCM microvessels and whole brain for the 20 highest confidence mouse-enriched (C) and human-enriched (D) microvessel genes from lists filtered by Approach 1 (genes also microvessel-enriched in the respective species). Color indicates expression that has been normalized within each gene (row).

Figure 5

Comparison of human LCM microvessel and brain endothelial cell transcriptomes. (A) Average transcript abundance of the 587 human microvessel-enriched genes (as determined in Fig. 3F) in human LCM microvessels (TPM_MV) versus reference human adult brain endothelial cells analyzed by single cell RNA-seq (TPM_EC). Vertical line at log₂(TPM_EC + 1) = 1 indicates the 1 TPM threshold employed to identify putative pericyte-derived transcripts in the microvessel samples. Endothelial genes (e.g. PECAM1, MFSD2A, SLC2A1, A2M) fall to the right of this line and are excluded, while known pericyte genes (e.g. NOTCH3, PDGFRB) fall to the left. Full results of this analysis are in Supplementary Table S6. (B) Summary of results of thresholding analysis described in (A). (C) Heat map illustrating transcript abundance in biological triplicates of human brain endothelial cells and LCM microvessels and whole brain for the 186 putative pericyte genes identified in (A). Known and putative novel pericyte genes are annotated. Color indicates expression that has been normalized within each gene (row). (D) Summary of filtering strategy used to identify putative human-specific pericyte genes. Of the 186 putative pericyte genes identified in (A), 48 are also human-enriched (as determined in Fig. 4B). (E) Heat map illustrating transcript abundance in biological triplicates of mouse LCM microvessels and whole brain, human brain endothelial cells, and human LCM microvessels and whole brain for the 48 putative human-enriched pericyte genes identified in (D). Genes are ranked by fold change in human versus mouse microvessels. Color indicates expression that has been normalized within each gene (row).