Autoimmune Valvular Carditis Requires Endothelial Cell TNFR1 Expression

Abstract

Background: Inflammation is a key driver of cardiovascular pathology, and many systemic autoimmune/rheumatic diseases are accompanied by increased cardiac risk. In the K/B.g7 mouse model of coexisting systemic autoantibody-mediated arthritis and valvular carditis, valve inflammation depends on macrophage production of TNF (tumor necrosis factor) and IL-6 (interleukin-6). Here, we sought to determine if other canonical inflammatory pathways participate and to determine whether TNF signaling through TNFR1 (tumor necrosis factor receptor 1) on endothelial cells is required for valvular carditis.

Methods: We first asked if type 1, 2, or 3 inflammatory cytokine systems (typified by IFNγ, IL-4, and IL-17, respectively) were critical for valvular carditis in K/B.g7 mice, using a combination of in vivo monoclonal antibody blockade and targeted genetic ablation studies. To define the key cellular targets of TNF, we conditionally deleted its main proinflammatory receptor, TNFR1, in endothelial cells. We analyzed how the absence of endothelial cell TNFR1 affected valve inflammation, lymphangiogenesis, and the expression of proinflammatory genes and molecules.

Results: We found that typical type 1, 2, and 3 inflammatory cytokine systems were not required for valvular carditis, apart from a known initial requirement of IL-4 for autoantibody production. Despite expression of TNFR1 on a wide variety of cell types in the cardiac valve, deleting TNFR1 specifically on endothelial cells protected K/B.g7 mice from valvular carditis. This protection was accompanied by reduced expression of VCAM-1 (vascular cell adhesion molecule), fewer valve-infiltrating macrophages, reduced pathogenic lymphangiogenesis, and diminished proinflammatory gene expression.

Conclusions: TNF and IL-6 are the main cytokines driving valvular carditis in K/B.g7 mice. The interaction of TNF with TNFR1 specifically on endothelial cells promotes cardiovascular pathology in the setting of systemic autoimmune/rheumatic disease, suggesting that therapeutic targeting of the TNF:TNFR1 interaction could be beneficial in this clinical context.

Keywords: autoimmunity; cardiac valves; endothelial cells; macrophages; rheumatic heart disease; tumor necrosis factor-alpha.

Figure 1.. Canonical activators of type 1 and type 3 inflammation are not required for valvular carditis in K/B.g7 mice.

Top panels, H&E stained sections of representative mitral valve (MV) sections from K/B.g7 mice with indicated genotypes. Bottom panels, MV thickness measurements in the same mice. Orientation of left atrium (LA) and left ventricle (LV) across all panels is the same as denoted in A; all scale bars represent 50 μm. Data were analyzed using a two-tailed Mann-Whitney test. Data are represented as mean ± SD; p-values are displayed in each panel.

Figure 2.. Type 2 inflammatory pathways are not required for valvular carditis in K/B.g7 mice.

Top panels, H&E staining of K/B.g7 MVs from mice that received indicated antibody blockade treatment and corresponding isotype control antibody mouse IgG (mIgG) or rat IgG (rIgG) or indicated Cx3Cr1-Cre mediated receptor deletion. Bottom panels, MV thickness measurements in the same mice. Orientation of left atrium (LA) and left ventricle (LV) across all panels is the same as denoted in A; all scale bars represent 50 μm. Data were analyzed using a two-tailed Mann-Whitney test. Data are represented as mean ± SD; p-values are displayed in each panel.

Figure 3.. Single-cell RNA sequencing shows heterogenous expression of TNFR1 and confined expression of TNF.

A, Uniform Manifold Approximation and Projection (UMAP) visualization of scRNA sequenced K/B.g7 and B.g7 mitral valves (MVs). Cells were obtained from 3-, 8-, and 25-week-old mice (n=2 per age group from K/B.g7 and B.g7) MVs. Cell types characterizing each cluster are listed to the right of UMAP plot. Three endothelial cell clusters are distinguished by expression of Vwf (3), Prox1 (7), Hapln1 (11). Cluster 2 “SMCs” denote smooth muscle cells, cluster 9 “RBCs”; red blood cells. B, Transcript expression of Cdh5 (VE Cadherin or CD144) overlayed on UMAP plot. C, Tnfrsf1a (encodes TNFR1, top) and Tnfrsf1b (encodes TNFR2, bottom). D, Expression of Ptprc (CD45, top) and Fcgr1 (CD64, bottom). E, Expression of cytokines Tnf (TNF, top) and Il6 (IL-6, bottom).

Figure 4.. Endothelial deletion of TNFR1 prevents and reverses mitral valve disease.

A, qRT-PCR analysis of Tnfr1 mRNA levels in TNFR1^fl/fl (closed circle) and TNFR1^ΔCdh5 (open circle) mitral valve (MV) tissue from sorted CD45⁻CD31⁻ cells, sorted hematopoietic (CD45⁺) cells only, or sorted endothelial cells (CD31+) only. B, Schematic depiction of intraperitoneal (IP) tamoxifen (TAM) injections and end-point analysis for K/B.g7 TNFR1^fl/fl and TNFR1^ΔCdh5 mice. Cre expression was induced by administration of tamoxifen (TAM) at 3 weeks (preventative) or 6 weeks of age (therapeutic). Hearts were analyzed 5 weeks after initiation of TAM. C-D, Trichrome stain of representative MV section from preventative (C) and therapeutic (D) Cre induction in TNFR1^fl/fl (top image) and TNFR1^ΔCdh5 (bottom image) mice. Scale bars represent 100 μm. Left atrium (LA) and left ventricle (LV) are indicated, surrounding the mural leaflet. MV thickness measurements from non-inflamed (KRN-neg) and inflamed (TNFR1^fl/fl) littermate control mice are reported alongside TNFR1^ΔCdh5 mice (right). E-F, Average change in ankle thickness (left) and anti-GPI IgG titers (right) from preventative Cre (E) and therapeutic Cre (F). Anti-GPI IgG titers were read at an absorbance (A) of 415 nm and reported in arbitrary units (AU). MV thickness were analyzed with a non-parametric Kruskal-Wallis followed by Dunn’s multiple comparisons. Anti-GPI titers were analyzed with a non-parametric two-tailed Mann-Whitney test. Change in ankle thickness was analyzed with a 2-way repeated measures ANOVA with Sidak’s multiple comparisons. Data are represented as mean ± SD; p-values are displayed in each panel.

Figure 5.. TNFR1^ΔCdh5 mice have reduced endothelial activation, immune cell infiltration, apoptotic cells, and lymphatic vessel development.

Representative immunofluorescent staining in MVs (white outline) from control TNFR1^fl/fl (left image) and TNFR1^ΔCdh5 (right image) from therapeutic Cre induction. A, VCAM-1 (yellow) expression in mitral valve (MV). Area of positive VCAM1 signal as a percent of total MV area is quantified at right. Total images analyzed for preventative TNFR1^fl/fl, TNFR1^ΔCdh5 and therapeutic TNFR1^fl/fl, TNFR1^ΔCdh5 as follows: n=6, n=8, n=9, n=12. B, CD64 (red) and DAPI (blue) staining in MV. Total CD64 cells as a percent of total cell nuclei quantified. Total images analyzed as follows (n=12, n=11, n=11, n=11). C, VEGFR3 (red) and LYVE1 (green) staining. Colocalized signal (yellow) of both markers are lymphatic vessels (white arrows). Total number of lymphatic vessels is reported. Total images analyzed as follows (n=8, n=12, n=9, n=10). D, Cleaved caspase-3 (green) staining for apoptotic cells (white arrows). Total caspase-3+ cell count reported. Total images analyzed as follows (n=12, n=14, n=12, n=10). Data points may represent averaged image quantification from same specimen. Exposure set using a species-matched isotype control. Scale bars represent 100 μm. Left atrium (LA), left ventricle (LV), mural leaflet (ML), and septal leaflet (SL) are indicated. Orientation in A applies to all panels unless indicated otherwise. Data were analyzed using a two-tailed Mann-Whitney test. Data are represented as mean ± SD; p-values are displayed in each panel.

Figure 6.. Deletion of endothelial TNFR1 reduces expression of genes associated with inflammatory and lymphangiogenic processes.

MVs from preventative Cre-induced TNFR1^ΔChd5 and control mice were analyzed via qRT-PCR for gene transcripts generally associated with A, Extracellular matrix remodeling; B, inflammation; C, angio- and lymphangiogenesis. Expression normalized to B-actin and 2^ΔΔCT values are displayed. Statistical analyses performed with unpaired, nonparametric Mann-Whitney test. Red bar indicates mean; p-values are displayed in each panel. Correction for multiple testing was not performed.