Single-Cell Imaging Maps Inflammatory Cell Subsets to Pulmonary Arterial Hypertension Vasculopathy

Abstract

Rationale: Unraveling immune-driven vascular pathology in pulmonary arterial hypertension (PAH) requires a comprehensive understanding of the immune cell landscape. Although patients with hereditary (H)PAH and bone morphogenetic protein receptor type 2 (BMPR2) mutations have more severe pulmonary vascular pathology, it is not known whether this is related to specific immune cell subsets. Objectives: This study aims to elucidate immune-driven vascular pathology by identifying immune cell subtypes linked to severity of pulmonary arterial lesions in PAH. Methods: We used cutting-edge multiplexed ion beam imaging by time of flight to compare pulmonary arteries (PAs) and adjacent tissue in PAH lungs (idiopathic [I]PAH and HPAH) with unused donor lungs, as controls. Measurements and Main Results: We quantified immune cells' proximity and abundance, focusing on those features linked to vascular pathology, and evaluated their impact on pulmonary arterial smooth muscle cells (SMCs) and endothelial cells. Distinct immune infiltration patterns emerged between PAH subtypes, with intramural involvement independently linked to PA occlusive changes. Notably, we identified monocyte-derived dendritic cells within PA subendothelial and adventitial regions, influencing vascular remodeling by promoting SMC proliferation and suppressing endothelial gene expression across PAH subtypes. In patients with HPAH, pronounced immune dysregulation encircled PA walls, characterized by heightened perivascular inflammation involving T cell immunoglobulin and mucin domain-3 (TIM-3)⁺ T cells. This correlated with an expanded DC subset expressing indoleamine 2,3-dioxygenase 1, TIM-3, and SAM and HD domain-containing deoxynucleoside triphosphate triphosphohydrolase 1, alongside increased neutrophils, SMCs, and alpha-smooth muscle actin (ACTA2)⁺ endothelial cells, reinforcing the heightened severity of pulmonary vascular lesions. Conclusions: This study presents the first architectural map of PAH lungs, connecting immune subsets not only with specific PA lesions but also with heightened severity in HPAH compared with IPAH. Our findings emphasize the therapeutic potential of targeting monocyte-derived dendritic cells, neutrophils, cellular interactions, and immune responses to alleviate severe vascular pathology in IPAH and HPAH.

Keywords: PAH; TIM-3+ T cells; monocyte-derived dendritic cells; multiplexed ion beam imaging by time of flight; neutrophils.

Figure 1.

Multiplexed imaging interrogation of pulmonary arteries (PAs) from patients with pulmonary arterial hypertension and donor control (Con) subjects. (A) Schematic depicting the workflow for the study. Representative PAs from 10 Con and 12 patients with PAH of whom 6 patients were diagnosed with idiopathic (I)PAH and 6 with hereditary (H)PAH, were selected and embedded in a tissue microarray. Ninety PAs were scored on H&E in relation to the severity of the lesion, and subsequent tissue sections were obtained to conduct multiplexed ion beam imaging by time of flight (MIBI-TOF) analyses. The image was created with BioRender.com. (B) Cell lineage assignments based on mean normalized expression of lineage markers (heatmap columns) hierarchically clustered (euclidean distance, average linkage). Rows are ordered by cell lineage (bottom) and immune cell breakdown (top). The absolute abundance of each cell type is displayed (left). (C) Representative MIBI-TOF image overlay from a PA section showing cell identity by color, as defined in B. Scale bars, 100 μm. Six zoomed insets showing MIBI-TOF overlays across diverse regions displaying neutrophil cells (region of interest [ROI] 1), DCs and macrophages (ROI 2), T cells, together with epithelial and SAM domain and HD domain-containing protein 1 (SAMHD1)⁺ cells (ROI 3). Scale bars, 50 μm. (D) Difference in endothelial cell (EC) and smooth muscle cell (SMC) counts between Con and patients with PAH, together with difference in alpha-smooth muscle actin (ACTA2)⁺ ECs and vimentin⁺ SMCs. (E) Difference in immune cell frequency (mean and SD) between Con and patients with PAH. (F) Difference in immune cell subsets frequency (median and interquartile range) between Con and patients with PAH. (G) Relationship (correlation) between immune cell infiltration and PA score in patients with PAH. Spearman’s rank correlation (r_S) was used for the correlation, with a 95% confidence interval and two-tailed test. P values were calculated with unpaired t test with Welch’s correction or a Wilcoxon rank-sum test, where *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. DCs = dendritic cells; Epi = epithelial cell; H&E = hematoxylin and eosin; HH3 = histone H3; Macro = macrophage; Mese = mesenchymal cell; Mono = monocyte; MPO = myeloperoxidase; Neutro = neutrophil; NK = natural killer cell; Pan-CK = pan-cytokeratin; PL = plexogenic lesion; Treg = regulatory CD3⁺ CD4⁺ FOXP3⁺ cell; Vim = vimentin.

Figure 2.

Intramural immune cells relate to exacerbated vascular pathology. (A) Binary masks were generated to localize immune cells with respect to the vascular architecture. Alpha-smooth muscle actin (ACTA2) staining from the multiplexed ion beam imaging by time-of-flight scan was used to extrapolate the intramural region (maroon). A region within 100 pixels (∼50 μm) of the vessel was defined as perivascular (dark orange). Any region beyond that was considered nonvessel (gray). (B) The relationship (correlation) between pulmonary artery (PA) score and the immune cell frequency across vascular regions in patients with PAH. (C) The immune cell composition across vascular regions, with IM representing intramural, PV representing perivascular, and NV representing nonvessel regions. (D) Heatmap displaying the relationship (correlation) between immune cell counts across vascular regions (heatmap rows) and PA score (heatmap columns). The heatmap includes total cell counts (regardless of vascular region) as well as counts specific to IM, PV, and NV regions. The P values are indicated by white asterisks. Correlations were performed using Spearman’s rank correlation (r_S) with 95% confidence interval and two-tailed test, where *P < 0.05; **P < 0.01; and ***P < 0.001. DC = dendritic cell; HH3 = histone H3; Macro = macrophage; Mono = monocyte; Neutro = neutrophil; NK = natural killer cell; Treg = regulatory CD3⁺ CD4⁺ FOXP3⁺ cell.

Figure 3.

The spatial location of dendritic cells (DCs) and ACTA2+ endothelial cells relate to vessel score. (A) Representative image of a pulmonary artery (PA) displaying how cell types are grouped into local microenvironments (MEs). The corresponding hematoxylin and eosin image provides context for the PA tree architecture. Ten MEs are displayed in the image corresponding to different combinations of proximal cell types, as defined in B. Scale bars, 100 μm. (B) Cell composition of the 10 tissue MEs identified: immune and smooth muscle cell enriched (ME-1); immune enriched (ME-2); epithelial associated (ME-3); smooth muscle cell enriched (ME-4); endothelial enriched (ME-5); CD8 T cell enriched (ME-6); mesenchymal associated (ME-7); endothelial and epithelial enriched (ME-8); immune and mesenchymal enriched (ME-9); and B cell enriched (ME-10). (C) Heatmap showing the functional profile of each ME. The mean normalized expression of hierarchically clustered (euclidean distance, average linkage) functional markers across MEs is reported. (D) Representative images of PAs showing their architecture with hematoxylin and eosin staining and the cell composition of specific MEs. Black arrows highlight the ME locations within the images. Both ME-1 and ME-4 are intramural, with ME-1 in the subendothelial and adventitial regions and ME-4 in the medial regions of the PA. ME-10, predominantly B cell enriched, is adjacent to the PA wall and localizes in tertiary lymphoid structures. ME-2 and ME-6, enriched with immune cells, surround the PA wall. Both ME-5 and ME-8 are localized within endothelial-enriched regions, with ME-5 in the intima layer and ME-8 in extramural vessels surrounding the PA, enriched with neutrophils and natural killer cells. Scale bars, 100 μm. (E) Relationship (linear regression) between the frequency of DCs or alpha-smooth muscle actin (ACTA2)⁺ endothelial cells within ME-1 and the patient vessel score in PAH. EC = endothelial cell; Epi = epithelial cell; GrnzB = granzyme B; Macro = macrophage; Mese = mesenchymal cell; Mono = monocyte; MPO = myeloperoxidase; Neutro = neutrophil; NK = natural killer cell; SMC = smooth muscle cell.

Figure 4.

The spatial location of monocyte-derived dendritic cells (mo-DCs) dictates their involvement in vascular remodeling and endothelial dysfunction. (A) DC composition within the subendothelial and adventitial regions (microenvironment [ME]-1) of pulmonary arteries. Highlighted in bold are CD11c-negative DC3s (mo-DC–like cells), the predominant DC subset. (B) Representative image of a pulmonary artery showing the spatial localization of distinct MEs, denoted by color as in Figure 3A. A zoomed inset of a region including ME-1 was chosen to examine the presence of CD11c-negative DC3s using CD14, CD209, and CD163 marker expression. ACTA2 was displayed to highlight the pulmonary arterial tissue. ME map scale bar, 100 μm; zoomed insets scale bars, 50 μm. (C) Heatmap displaying the association (correlation) between DC subsets within ME-1 and patient vessel score or proliferating smooth muscle cells (Ki-67⁺ SMC). The P values (white asterisk) are reported. The absolute abundance of each cell type is displayed (right). Correlations were performed using Spearman’s rank correlation (r_S) with 95% confidence interval and two-tailed test. (D) Schematic illustration depicting the workflow implemented to obtain and culture mo-DCs with pulmonary arterial smooth muscle cells (PASMCs) from donor control subjects and patients with PAH. Undifferentiated CD14⁺ cells were isolated from PBMCs. CD14⁺ cells were incubated with differentiation media to obtain immature and mature mo-DCs. Each mo-DC maturation stage was cocultured with PASMCs in Transwells at a ratio of 1:1. A separate well with only PASMCs served as a control. The image was created with BioRender.com. (E) PASMC proliferation (median and interquartile range) across coculture conditions (immature mo-DC–PASMC; mature mo-DC–PASMC; PASMC-PASMC) and study groups (donor control subjects and patients with PAH). (F) Differences among culture conditions involving immature mo-DCs and PASMCs for PECAM-1 and CDH5 gene expression measured in pulmonary arterial endothelial cells (PAECs). Five technical replicates are displayed per condition. Before evaluating coculture differences, data were normalized to PAEC alone control. Coculture significance was assessed via a mixed-effects model, accounting for repeated measures as a random effect, followed by Tukey-adjusted pairwise comparisons. Significance levels are denoted as follows: *P < 0.05; **P < 0.01; ***P < 0.001. B2M = beta-2 microglobulin; CDH5 = cadherin-5; Con = donor control; GM-CSF = granulocyte-macrophage colony-stimulating factor; PBMCs = peripheral blood mononuclear cells; PECAM1 = platelet endothelial cell adhesion molecule; TNF-α = tumor necrosis factor-alpha.

Figure 5.

Immune dysregulation in patients with HPAH. (A) Difference (median and interquartile range [IQR]) in SMC counts (top) and ACTA2⁺ endothelial cells (ECs) within microenvironment (ME)-1 (bottom) between patients with hereditary (H)PAH and patients with idiopathic (I)PAH. (B) Schematic image depicting the minimum distance of extramural cell types to the intramural region (left panel). Difference in minimum distance (median, line; IQR, dotted lines) to the vessel wall of extramural immune cell subsets between PAH subtypes (right panel). (C) Difference (median and IQR) in TIM-3⁺ (CD4 and CD8) T cells within MEs surrounding the PA wall (ME-2) between PAH subtypes. (D) Difference (median and IQR) in the frequency of dendritic cells coexpressing IDO-1, TIM-3, and SAMHD1 between PAH subtypes. (E) Relationship (linear regression) between IFN-γ⁺ CD4 T cells and IDO-1⁺ TIM-3⁺ SAMHD1⁺ DCs in patients with PAH. Blue dots depict patients with HPAH, and magenta dots depict patients with IPAH. (F) Schematic illustration summarizing the results. ACTA2 = alpha-smooth muscle actin; SMC = smooth muscle cell and fibroblast; EC = endothelial cell; IDO-1 = indoleamine 2,3-dioxygenase 1; IFN-γ = Interferon-gamma; SAMHD1 = SAM Domain and HD Domain-containing Protein 1; TIM-3 = T-cell immunoglobulin and mucin Domain-3. P values were calculated with a Wilcoxon rank-sum test, where *P < 0.05; **P < 0.01; and ****P < 0.0001. The image was created with BioRender.com.

Figure 6.

Neutrophil enrichment in HPAH pulmonary artery endothelial regions. (A) Difference (median and interquartile range) in the frequency of neutrophils within extramural endothelial-enriched microenvironments (ME-8) between patients with HPAH and patients with idiopathic (I)PAH (top), together with corresponding representative images (bottom). Neutrophils are highlighted in the hematoxylin and eosin zoomed insets (black arrows) and confirmed by signature markers (CD11b, CD15, MPO) using multiplexed ion beam imaging by time of flight (MIBI-TOF). Cell overlays from the MIBI-TOF images depict neutrophils in blue. Images scale bars, 100 μm; zoomed insets scale bars, 50 μm. P values were calculated with unpaired t test with Welch’s correction, where **P < 0.01. (B) Representative image of a pulmonary artery showing the spatial localization of distinct MEs on the left, together with the frequency of neutrophils across MEs on the right in PAH. Black arrows in the image indicate where neutrophils preferentially accumulate. The text highlighted in bold and underlined specifies the cell composition of these MEs. The corresponding hematoxylin and eosin image provides context for the pulmonary artery tree architecture. Images scale bars, 100 μm. (C) The relationship (correlation) between neutrophils within intramural endothelial-enriched microenvironments (ME-5) and the frequency of nitric oxide synthase (iNOS)⁺ or HLA class I⁺ ECs in patients with PAH. Correlations were performed using Spearman’s rank correlation (r_S) with 95% confidence interval and two-tailed test, where P values are reported as **P < 0.01. ECs = endothelial cells; HH3 = histone H3; MPO = myeloperoxidase.