The analysis of receptor gene expression changes by specific cell type reveals both common and cell-type-specific changes in associated molecular functions (
Figure 63B–I). Cardiomyocytes demonstrated the decreases in molecular functions associated with the aggregates of all cell types for each condition, as well as additional decreases in insulin receptor substrate binding, insulin-like growth factor binding, insulin-like growth factor I binding, insulin-like growth factor receptor binding, protein tyrosine kinase activity, protein-hormone receptor activity, transmembrane receptor kinase activity, and transmembrane receptor protein tyrosine kinase activity in both nonobstructive and obstructive HCM. Cardiomyocytes also showed increases in calcium channel activity, calcium ion transporter activity, calcium-release channel activity, calmodulin binding, cation channel activity, channel activity, divalent inorganic cation transmembrane transporter activity, gated channel activity, intracellular ligand-gated ion channel activity, ion channel activity, ion channel binding, ligand-gated calcium channel activity, ligand-gated cation channel activity, ligand-gated channel activity, ligand-gated ion channel activity, metal ion transmembrane transporter activity, passive transmembrane transporter activity, protein kinase A catalytic subunit binding, protein kinase A regulatory subunit binding, and sulfur compound binding as noted in the aggregate, but these changes occurred in cardiomyocytes from both nonobstructive and obstructive HCM (
Figure 63B), which differ in that the aggregate nuclei population only showed these changes in nonobstructive HCM (
Figure 63A). Cardiomyocytes in nonobstructive HCM showed unique increases in 3′,5-cyclic-GMP phosphodiesterase activity, 3′,5′-cyclic nucleotide phosphodiesterase activity, cyclic nucleotide phosphodiesterase activity, and phosphoric ester hydrolase activity. Fibroblasts (
Figure 63C), dendritic cells (
Figure 63F), leukocytes (
Figure 63G), and smooth muscle cells (
Figure 63H) demonstrate receptor gene expression patterns similar to those shown for the general aggregate population of nuclei. Endothelial cell receptor gene expression profiles generally paralleled the general population, except that for the various ion and calcium channel activities the increases occurred in both nonobstructive HCM and obstructive HCM, as was seen in cardiomyocytes, and for ammonium transmembrane transporter activity, cyclic nucleotide-gated ion channel activity, intracellular cyclic nucleotide activated cation channel activity, polyol transmembrane transporter activity, water channel activity, and water transmembrane transporter activity, which were uniquely increased in obstructive HCM (
Figure 63D). Pericytes showed receptor gene expression patterns similar to those of the general population, except that the differences between nonobstructive and obstructive HCM were less pronounced (
Figure 63E). Neurons also showed receptor gene expression patterns similar to those of the general population but shared some features with cardiomyocytes (increased 3′,5′ cyclic GMP and nucleotide phosphodiesterase and less pronounced differences between nonobstructive and obstructive HCM for other molecular functions (
Figure 63I)). Overall, these findings indicate that there are changes in ligand and receptor gene expression that are cell-specific and either common or unique to each type of HCM that may provide unique, personalized, and cell-specific therapeutic targets for future therapy.
4. Fibroblast and Cardiomyocyte Subtypes Communicate Differently in Nonobstructive and Obstructive HCM
In prior studies, fibroblast subtypes in both obstructive and nonobstructive HCM developed distinctive interaction profiles with other cells
[13,14][10][11]. To compare the fibroblast interactome in both nonobstructive and obstructive HCM, we analyzed the L–R pair gene expression for all fibroblast subtypes (composed of five clusters) in conjunction with all other cell types (
Figure 74). As expected, the fibroblast subtype intercellular communication networks for normal (2157), nonobstructive HCM (1291), and obstructive HCM (838) were qualitatively similar to the general cell interaction network in that fibroblast subtypes from nonobstructive HCM demonstrated reduced interactions compared to normal tissue and interactions were reduced to a greater extent in obstructive HCM tissue (
Figure 74A). The greatest reduction in L–R pairs was found in ligands broadcasted from fibroblast clusters 2 and 3 to receptors on fibroblast clusters 1, 4, and 5 and leukocytes in obstructive HCM compared to nonobstructive HCM (
Figure 74C,D). This reduction was due to the loss of ITGB1 receptor expression in the recipient clusters in obstructive HCM. There are no cases of increased L–R interaction in obstructive HCM compared to nonobstructive HCM.
Figure 74. Cell–cell communication networks between fibroblast subtypes and other heart cell types in normal control, nonobstructive HCM, and obstructive HCM conditions. (A) Comparison of normal (left), nonobstructive (middle), and obstructive HCM (right) communication networks. Line color indicates ligand broadcast by the cell population with the same color. Lines connect to cell types that expressed cognate receptors. Line thickness is proportional to the number of uniquely expressed ligand–receptor pairs. Loops indicate communication within a cell type. (B) Quantity of ligands and receptors in expressed ligand–receptor pairs described by cell type and condition (normal, nonobstructive HCM, or obstructive HCM). (C–E) Cell–cell communication networks broken down by cell type and fibroblast cluster in normal control (C), nonobstructive (D), and obstructive (E) conditions. Figure formatting follows panel A. Numbers indicate the quantity of uniquely expressed ligand–receptor pairs between the broadcasting cell type (expressing ligand) and receiving cell type (expressing receptor).
Cardiomyocyte subtypes also demonstrate diversity in their intercellular interactions in normal and HCM tissue, as previously reported
[13,14][10][11]. To understand how these interactions compare and contrast in nonobstructive and obstructive HCM, we analyzed the L–R pair gene expression for all cardiomyocyte subtypes in conjunction with all other cell types (
Figure 85). As with the general population and the fibroblast subtype interaction networks, in the cardiomyocyte subtype interactome there was a stepwise decrease in L–R pairs from normal (
n = 4854) to nonobstructive (
n = 3836) to obstructive (
n = 2294) HCM. Cardiomyocyte clusters 1, 2, 3, 7, and 8 did not follow this trend, however, expressing higher numbers of ligands in nonobstructive HCM compared to both normal and obstructive HCM tissue. Cardiomyocyte clusters 3 and 7 also demonstrated higher than normal numbers of ligands in obstructive HCM, but to a far lesser extent (
Figure 85B–E). Cardiomyocyte clusters 3, 5, and 7 and dendritic cells expressed higher numbers of receptors than normal tissue in nonobstructive HCM. Cardiomyocyte clusters 3 and 7 also showed higher numbers of receptors in obstructive HCM compared to normal tissue, but less than those in nonobstructive HCM (
Figure 85B–E). Comparing L–R pairs more directly in HCM samples indicates that cardiomyocyte cluster 10 demonstrates the greatest reduction in L–R pairs in obstructive HCM compared to nonobstructive HCM, while cardiomyocyte cluster 9 shows the greatest increase (
Figure 85D,E). The reduction in cardiomyocyte cluster 10 communication to dendritic cells; leukocytes; and cardiomyocyte clusters 4, 5, 8, 10, 12, 13, and 14 in obstructive HCM is driven by a loss of ITGB1 receptor expression in conjunction with a reduction in the expression of ITGB1 ligands. The increase in cardiomyocyte cluster 9 communication involves cardiomyocyte clusters 7, 9, and 13 and is driven by the increased expression of the ITGB1 ligands COL6A2, LUM, and VEGFA and the ligand CALM2.
Figure 85. Cell–cell communication networks between cardiomyocyte subtypes and other heart cells in normal control, nonobstructive HCM, and obstructive HCM conditions. (A) Comparison of normal (left), nonobstructive HCM (middle), and obstructive HCM (right) communication networks. Line color indicates ligand broadcast by the cell population with the same color. Lines connect to cell types that expressed cognate receptors. Line thickness is proportional to the number of uniquely expressed ligand–receptor pairs. Loops indicate communication within a cell type. (B) Quantity of ligands and receptors in expressed ligand–receptor pairs described by cell type and condition (normal, nonobstructive HCM, or obstructive HCM). (C–E) Cell–cell communication networks broken down by cell type and cardiomyocyte cluster in normal control (C), nonobstructive HCM (D), and obstructive HCM (E) conditions. Figure formatting follows panel A. Numbers indicate the quantity of uniquely expressed ligand–receptor pairs between the broadcasting cell type (expressing ligand) and receiving cell type (expressing receptor).
The cardiomyocyte–fibroblast interaction network has also been postulated to play an important role in the pathogenesis of HCM through direct cellular interactions and indirect interactions with the extracellular matrix, and prior studies have shown HCM-associated alterations in cardiomyocyte–fibroblast interactions [5,13,14]. The analysis of L–R pairs among the cardiomyocytes and fibroblasts reveals the same trend seen in the general population of cells, with the greatest number of potential interactions occurring in normal cells (n = 4500) followed by nonobstructive HCM cells (n = 3509) then obstructive HCM cells (n = 1933; Figure 9A,B). Cardiomyocyte clusters 1, 2, 3, 7, and 14 do not follow this trend by expressing higher numbers of ligands in nonobstructive HCM compared to both normal and obstructive tissue. Cardiomyocyte clusters 3 and 7 also express higher numbers of ligand in obstructive HCM than normal tissue, although the numbers are less than those seen in nonobstructive tissue (Figure 9B). Cardiomyocyte clusters 8, 10, and 12 and fibroblast clusters 1 and 4 express higher levels of receptors than normal cells in nonobstructive HCM. Cardiomyocyte clusters 8 and 12 also express higher levels of receptors than normal cells, but to a lesser degree than nonobstructive HCM cells (Figure 9B–E). A more direct comparison of nonobstructive HCM and HCM reveals that the greatest difference in L–R pair number occurs in cardiomyocyte cluster 10, with reduced communication with cardiomyocyte clusters 5, 8, 10, 12, 13, and 14 and with fibroblast clusters 1, 4, and 5, largely due to the loss of the ITGB1 receptor and many ITGB1 ligands (Figure 9D,E). The greatest increase in communication in the cardiomyocyte–fibroblast interaction network occurs between fibroblast cluster 2 and itself and cardiomyocyte cluster 9 and itself. The increase in communication in fibroblast cluster 2 is driven by the increased expression of ITGB1 ligands such as COL1A1, FBLN1, FBN1, HPG2, LAMC1, LGALS38P, and the LRP1 receptor. The increase in communication in cardiomyocyte cluster 9 is driven by the expression of the CALM2 ligand; the ITGB1 receptor ligands COL6A2, LUM, and VEGF-A; the receptor INSR; and at least one component of the L–R pair TIMP1 and CD63.
Figure 96. Cell–cell communication networks between cardiac fibroblast and cardiomyocyte subtypes in normal control, nonobstructive HCM, and obstructive HCM conditions. (A) Overall communication networks between normal, nonobstructive, and obstructive HCM cardiomyocytes and fibroblasts. Line color indicates ligand broadcast by the cell population with the same color. Lines connect to cell types which expressed cognate receptors. Line thickness is proportional to the number of uniquely expressed ligand–receptor pairs. Loops indicate communication within a cell type. (B) Quantity of ligands and receptors in expressed ligand–receptor pairs described by cell type and condition (normal, nonobstructive HCM, or obstructive HCM). (C–E) Cell–cell communication networks broken down by cardiomyocyte cluster and fibroblast cluster in normal control (C), nonobstructive HCM (D), and obstructive HCM (E) conditions. Figure formatting follows panel A. Numbers indicate the quantity of uniquely expressed ligand–receptor pairs between the broadcasting cell type (expressing ligand) and receiving cell type (expressing receptor).