Revisiting Cardiac Cellular Composition

Abstract

Rationale: Accurate knowledge of the cellular composition of the heart is essential to fully understand the changes that occur during pathogenesis and to devise strategies for tissue engineering and regeneration.

Objective: To examine the relative frequency of cardiac endothelial cells, hematopoietic-derived cells, and fibroblasts in the mouse and human heart.

Methods and results: Using a combination of genetic tools and cellular markers, we examined the occurrence of the most prominent cell types in the adult mouse heart. Immunohistochemistry revealed that endothelial cells constitute >60%, hematopoietic-derived cells 5% to 10%, and fibroblasts <20% of the nonmyocytes in the heart. A refined cell isolation protocol and an improved flow cytometry approach provided an independent means of determining the relative abundance of nonmyocytes. High-dimensional analysis and unsupervised clustering of cell populations confirmed that endothelial cells are the most abundant cell population. Interestingly, fibroblast numbers are smaller than previously estimated, and 2 commonly assigned fibroblast markers, Sca-1 and CD90, under-represent fibroblast numbers. We also describe an alternative fibroblast surface marker that more accurately identifies the resident cardiac fibroblast population.

Conclusions: This new perspective on the abundance of different cell types in the heart demonstrates that fibroblasts comprise a relatively minor population. By contrast, endothelial cells constitute the majority of noncardiomyocytes and are likely to play a greater role in physiological function and response to injury than previously appreciated.

Keywords: endothelial cells; fibroblasts; flow cytometry; heart; leukocytes.

Figure 1. Quantification of major cardiac cell types

(A) Representative confocal microscopy optical section of cardiac tissue stained from a 10 week old Cx₃cr1^GFP/+ mouse. (A’) Magnified view of image shown in A. Nuclei (DAPI); cardiomyocyte boundaries (wheat germ agglutinin, WGA); endothelial cells (isolectin B4, IB4); and leukocytes (Lin1 cocktail: GFP, CD45, Mrc1, MHCII and B220). Arrowheads indicate a representative cardiomyocyte (CM), endothelial cell (EC), leukocyte (Lin1⁺) and unstained cells (unst.). 0.969 µM optical sections. (B) Average proportions from all myocardial regions examined (48 micrographs from 4 hearts). (C) Representative fluorescence microscopy image of cardiac tissue from a PDGFRα^GFP/+ mouse stained for DACH1. Arrowheads indicate a representative endothelial cell (EC) and fibroblast (FB). (D) Alternative quantification of cells in mouse left ventricle (LV) with identification of VSMCs/pericytes and fibroblasts. Lin2 cocktail (CD5, CD11b, B220, 7-4, Gr-1, and Ter-119): leukocytes; PDGFRβ: VSMCs/pericytes; IB4: endothelial cells; DACH1: endothelial cell nuclei; α actinin 2 (ACTN2): cardiomyocytes; Col1a1-GFP or PDGFRα^GFP: fibroblasts. (E) Quantification of ACTN2⁺ cardiomyocytes; CD31⁺ endothelial cells and CD45⁺ leukocytes in human heart tissue. Data averaged from all quantified myocardial regions (ventricular septum, and left and right ventricles). Tissue from 3 independent, adult, healthy, human heart samples.

Figure 2. Flow cytometry analysis of cardiac composition

(A) Gating of nucleated (VDO⁺), viable (DAPI^lo/−), and metabolically active (calcein⁺) cells with subsequent distribution based on CD31 and CD45 expression. (B) Expression of cell surface markers. (C) Unsupervised cell clustering using SPADE analysis following gating on VDO⁺DAPI^lo/−calcein⁺ events. Cell clustering based on CD31, CD102, Sca-1, CD45, CD90 and CD11b expression. Dendrogram node color represents relative expression as indicated by heat map. (D) SPADE dendrogram with endothelial cells (ECs), leukocytes (Leuks), and resident mesenchymal cells (RMC) manually annotated. Clustering based on CD31, C102, Sca-1, CD45, CD90 and CD11b expression, following manual gating on nucleated, calcein⁺ DAPI^lo/− cells. (E) Proportions of cells within branches corresponding to ECs, Leuks or RMC identified in the SPADE dendrogram in D (n=8). Heat map indicates level of expression.

Figure 3. Major cardiac cell subsets

(A) Representative gating of cardiac CD31⁺CD45⁻ cells for SPADE analysis. (B) Unsupervised SPADE clustering of CD31⁺CD45⁻ cells. Cell clustering based on CD102, podoplanin, CD105 and CD90 expression. (C) Proportions of VECs and LECs of total CD31⁺CD45⁻ cells. (D) Gating of cardiac CD45⁺ cells for SPADE analysis. (E) Unsupervised SPADE clustering of CD45⁺ cells. Cell clustering based on CD11b, CD64, B220, IgM, CD3ε and CD90 expression. MΦs, macrophages. (F) Proportions of cardiac leukocytes based on surface marker expression. n=3–4 for 3A–F. (G) Representative gating of cardiac CD31⁻CD45⁻ cells (RMC) for SPADE analysis. (H) Percent of GFP⁺ cells in RMC from indicated genotypes (n=6–9). (I) Unsupervised SPADE clustering of RMC from Col1a1-GFP mouse hearts. Cell clustering based on MEFSK4, GFP, CD90, and Sca-1 expression. (J) Contour plots of MEFSK4 staining, CD90, or Sca-1 concomitant with GFP expression in cells from the indicated mice (CD31⁻CD45⁻ population). All above populations were gated following pre-gating on calcein⁺ singlet cells that were either DAPI⁻ or 7-AAD⁻. Heat maps indicate level of expression.

Figure 4. MEFSK4 staining and fibroblast GFP expression overlap in Tcf21 lineage fibroblasts

Flow cytometry contour plots displaying gating of Tcf21 lineage cells (tdTomato⁺) from (A) Tcf21^iCre/+;ROSA26R^tdTomato;PDGFRα^GFP/+ or (B) Tcf21^iCre/+; ROSA26R^tdTomato;Col1a1-GFP mice with subsequent analysis of MEFSK4 staining and GFP expression. Cell analysis performed a minimum of 2 months following tamoxifen induction. Representative contour plots.

Figure 5. Distribution of major non-myocyte cardiac cell types

Combined data demonstrating relative cell numbers determined by flow cytometry with exclusion of cardiomyocytes.