Phenotypes and mechanisms of dysfunctional cardiac T-lymphocytes in dilated cardiomyopathy patients

Abstract

Background:

Dilated cardiomyopathy (DCM) is characterized by increased infiltration and activation of the innate immune system, including neutrophils, monocytes/macrophages, and dendritic cells. However, the phenotypic profile of cardiac CD3⁺ T-cells and its CD4⁺ and CD8⁺ subsets have not been characterized in DCM patients.

Methods:

We studied phenotypic signatures of T-cell subsets by analyzing publicly available single-cell and single-nuclear transcriptomic datasets from control and failing hearts of DCM patients.

Results:

Our analysis revealed increased cardiac infiltration of CD3⁺ T-cells in DCM patients with transcriptomic signatures indicating antigenic activation, T-cell exhaustion, diminished oxidative phosphorylation, and elevated TNF/NFκB and profibrotic TGF signaling. Among T-cell subsets, both CD4⁺ and CD8⁺ T-cells were found to be highly proliferative (increased G2M) and activated. Transcription profiling demonstrated four phenotypically different subsets for both CD8⁺ and CD4⁺ T-cells, however, only CD4⁺ T-cell subsets, regulatory T-cells and tissue resident memory (TRM) CD4⁺ T-cells, were significantly increased. Importantly, TRM cells displayed decreased expression of classical egress markers, such as CCR7, SELL, and MAL, and increased pro-inflammatory and pro-fibrotic signaling. We also observed increased estrogen receptor (ER)α expressing (with amplified ERα signaling) cardiac CD4⁺ T-cells which directly correlated with systolic dysfunction and mediated their pro-fibrotic effects in DCM patients.

Conclusion:

Here we demonstrate for the first time, an “activated phenotype” with increased pro-inflammatory and profibrotic signaling in cardiac CD3⁺ T-cells and its CD4⁺ helper T-cell subset in DCM hearts. Notably, increased ERα signaling provide novel avenues for targeted immunomodulatory therapies to modify DCM progression.

1. Introduction

Dilated cardiomyopathy (DCM) is characterized by eccentric hypertrophy, ventricular dilation, and systolic dysfunction [1]. While various factors contribute to the development of DCM, emerging clinical and preclinical studies have highlighted an important role for inflammatory innate and adaptive immune responses in this disease [2,3]. However, clinical trials aimed at immunosuppression through cytokine neutralization either failed to show any therapeutic benefits, as in RENEWAL [4] trial, or increased mortality and morbidity, as in ATTACH trial [5]. These paradoxical findings resulted in increased scrutiny regarding the role of immune cells that are direct targets of inflammatory cytokines to develop cell-specific immunomodulation strategies. Studies in rodents have shown that acute sterile cardiac injury, ischemic or pressure-overload mediated, initiates a rapid cascade of immune responses characterized by the activation and infiltration of innate (neutrophils, monocytes/macrophages and DCs) [6–9] as well as adaptive immune cells (CD4⁺ helper and CD8⁺ cytotoxic T-cells) [10–12] into the myocardium to clear dead and apoptotic cells and to promote fibrotic processes. In contrast, during cardiomyopathy, a key pathogenic role of antigen-specific CD4⁺ T-cells has also been reported in clinical [13,14] and preclinical [15] studies. Activated CD4⁺ T-cells during DCM express increased levels of pro-inflammatory and pro-fibrotic cytokines which not only provide critical cues to induce chemotaxis of other immune cells but also promote myofibroblast activation in a TGF-β dependent manner [16,17]. While experimental techniques to deplete CD4⁺ T-cells (antibody [16], DTR-mediated [11], knockdown of TCR [10], or MHC-II [18]) demonstrate ameliorated left-ventricular (LV) remodeling; adoptive transfer of activated T-cells from heart failure (HF) mice induce significant systolic dysfunction in naïve mice [16,19]. Recent studies have also shown the presence of antigen-specific T-cells in the circulation of HF patients such as against β1 adrenergic receptors (βADR1) [20] or myosin heavy chain α (MyHCα) [21], and preclinical rodent models [19] suggesting an auto-immune like mechanism of T-cell activation during DCM.

Despite prior studies showing presence of CD3⁺ T-cells in the hearts of patients with acute myocarditis and DCM [13], no studies have characterized the phenotype of cardiac infiltrated T-cells in human failing hearts and elucidated potential molecular mechanisms that mediate their pathological auto-immune like activation during DCM. To address this, we analyzed two publicly available single-cell/nuclei transcriptomic datasets [22,23] from the failing hearts of DCM patients and studied the activation status of cardiac infiltrated T-cells and characterized their phenotypic signatures at the single cell level. We are the first to categorize novel phenotypic signatures of activated T-cells during DCM in humans and identify targetable pathways to develop new therapeutics to subdue activated CD4⁺ T-cells to ameliorate disease progression.

2. Methods

2.1. Datasets and data processing

Two datasets (GSE183852 and GSE145154) containing either single-cell or single-nuclei RNA-seq samples from human hearts were analyzed. Data were obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) from previously published work [22,23]. In total, there were data from 48 hearts (20 with DCM while the remaining were control hearts).

Analysis was completed using the Seurat package (V4.3.0) [24] in R (V4.2.3). Cells were filtered based on their original publications criteria. After data normalization (NormalizeData function), variable gene selection (FindVariableFeatures), data scaling (ScaleData), and principal component analysis (RunPCA), the single-cell datasets were integrated together using canonical correlation analysis (CCA) integration (IntegrateLayers). Unsupervised Uniform Manifold Approximation and Projection (UMAP) dimensional reduction was computed using either CCA (single-cell) or PCA (single-nuclei) reduction. Clusters (FindNeighbors and FindClusters functions) and cluster markers (FindAllMarkers) were identified at a resolution of 3.0 (single-cell) or 1.0 (single-nuclei). The CellCycleScoring function was used to identify the cell cycle phase for each cell.

2.2. Differentially Expressed Genes (DEGs) and Gene Set Enrichment Analysis (GSEA)

To calculate Differentially Expressed Genes (DEGs) raw counts were aggregated by donor (AggregateExpression) for pseudobulk RNA-seq analysis. Genes with less than 10 counts were removed. The DESeq2 package (V1.38.3) [25] was used to detect DEGs using a threshold of P < 0.05, controlling for the dataset used during single-cell analysis. For Gene Set Enrichment Analysis (GSEA) ranked gene lists were created by ordering genes based on the Wald statistic. The resulting ordered gene lists were used for GSEA using the fGSEA package (V1.24.0). The Benjamini-Hochberg adjusted p-values (FDR) are reported.

2.3. Single cell gene set enrichment analysis (ssGSEA)

ssGSEA was performed using the escape package (V1.8.0) [26] using the UCell method [27] for single-cell gene signature scoring. The Wald’s test was utilized to test for differences in cell normalized enrichment scores between groups, and the Benjamini-Hochberg adjusted p-values are reported.

High Dimensionality Weighted Gene Co-expression Network Analysis (hdWGCNA).

hdWGCNA was performed on single-cell TRM-like CD4⁺ T-cells using the hdWGCNA package (V0.2.26) [28].Genes not found in at least 5 % of cells were excluded, and the number of cells aggregated was 10. A signed network was constructed using a soft threshold of 6. g:Profiler was used to explore the molecular function and biological process of each gene cluster [29].

2.4. Ingenuity pathway analysis (IPA)

The DEGs between activated TRM-like CD4⁺ T-cells and naive CD4⁺ T-cells in hearts with DCM were uploaded to Qiagen IPA (https://www.qiagenbioinformatics.com/products/ingenuitypathway-analysis) as an ‘expression analysis’ using log₂ fold change and p-value <0.05. ‘Graphical summary’, ‘Upstream analysis’, and ‘Target molecules’ with the ‘Diseases & Functions’ features were used to analyze the DEGs and produce the figures.

2.5. Immunofluorescence staining

Immunofluorescence staining was used to localize and quantify estrogen receptor (ER)α expressing CD4 T cells in the apical cardiac tissues collected from one formalin-fixed cadaver control heart and 9 DCM patients undergoing either heart transplant or LVAD implantation. Donor information is given in Supplement Tables 1, 2 and 3. All failing heart tissues were collected under an approved IRB#19928.

DCM tissues were fixed in 1 % (w/v) paraformaldehyde for 60 min, incubated serially in 15 % (w/v) and 30 % (w/v) sucrose solutions, each for 24 h, to prevent formation of ice-crystals upon freezing, and cryo-preserved in OCT medium. For immunostaining, 5 μm thick sections were cut and re-fixed with ice-cold acetone followed by washing with 0.01 % tween 20 in Tris-buffered saline (wash buffer). Sections were permeabilized with 0.1 % triton X100 (in Tris-buffered saline) for 5 min, blocked with 10 % (w/v) bovine serum albumin (in wash buffer) for 60 min, stained with wheat germ agglutinin (W11261; Invitrogen) conjugated to Alexa Fluor^™ 488 for 60 min followed by washing with wash buffer three times for 10 min each. Sections were then blocked with 10 % charcoal-stripped FBS for 60 mins at room temperature and labelled with rat anti-human CD4 (MA5-47810; Invitrogen) and mouse anti-human ERα antibodies (sc-8002; Santa Cruz) overnight at 4 °C, washed three times with the wash buffer and incubated with secondary antibodies (goat anti-rat Alexa Fluor^™ 633 and goat anti-mouse Alexa Fluor^™ 555, both from Invitrogen) for 60 min. After washing excess secondary antibody with wash buffer (three times,10 min each), nuclear staining was done using ProLong Glass Antifade with NucBlue Stain (P36981; Invitrogen). Images were acquired using a Leica DMi8 confocal microscope equipped with Violet, Blue, Green and Red lasers and LAS X 3.5.7.23225 software. The images were analyzed using LAS X office 3.7.6.25997 software for desktop.

2.6. Coculture experiments

Circulating CD4⁺ T-cells were isolated from patients with heart failure (NYHA class II and III) at the Pennsylvania State University Milton S. Hershey Medical Center (approved IRB#19928) using a MojoSort^™ Human CD4 T Cell Isolation Kit (BioLegend). T-cells were then cultured and treated for 24 h with either the estrogen receptor (ER)α agonist, propyl pyrazole triol (PPT; 10 nM; Tocris), or ER antagonist, Fulvestrant (100 nM; TCI Chemicals), or a vehicle control.

Following treatment, the CD4⁺ T-cells were washed and co-cultured with primary normal human cardiac fibroblasts (FC-0060; Lifeline Cell Technology) at a 2:1 ratio (200,000 T-cells to 100,000 fibroblasts) for an additional 48 h. Subsequently, the two cell populations were separated using a BD FACSMelody^™ Cell Sorter. RNA was then extracted from the cardiac fibroblasts for gene expression analysis. All treatment conditions were performed in duplicate for each patient.

3. Results

3.1. Cardiac CD3⁺ T-cells in DCM patients are activated and pro-inflammatory

Preclinical models of HF, pressure overload or ischemic, have shown an indispensable role of T-cells in mediating LV remodeling. Although there is some evidence of infiltrated CD3⁺ T-cells in the myocardium during acute myocarditis and DCM [13], detailed phenotypic characterization and activation mechanisms of T-cells have not been studied in human failing hearts. To answer these questions, we conducted bioinformatic analysis of cells present in the DCM hearts and found significantly higher proportions of CD45⁺ leukocytes (21.5 % ± 1.7 % vs 14.8 ± 1.1 %, p < 0.001) and CD3⁺ T-cells (3.7 % ± 0.8 % vs 1.3 ± 0.2 %, p = 0.009) than the control hearts (Figs. 1A & 1B), as has also been shown previously [13]. Analysis of cell transcription profiles showed that CD3⁺ T-cells infiltrated into the DCM hearts had a significantly higher fraction of activated/proliferating T-cells (G2/M phase, 29 ± 1 % vs 23 ± 1 %, P < 0.05) with fewer number of cells in the G1 phase (48 ± 3 % vs 62 ± 5 %, P < 0.05) as compared to the T-cells found in control hearts (Fig. 1C). The top differentially expressed genes that were significantly up- or down-regulated in DCM T-cells in all datasets are shown (Fig. 1D). GSEA pathway analysis of these genes demonstrated that gene-sets associated with increased IL-2 and IL-17 production (Fig. 1E), and T-cell activation, differentiation and proliferation (GOBP:0050798) were significantly up-regulated in cardiac T-cells from DCM patients (Fig. 1E and Fig. 1F) (enrichment score = 2.11, p <0.001). Using scRNA seq data from Koeing et al. 2022 [22], we also found that DCM hearts had significantly more T-cells exhibiting an ‘exhausted’ phenotype as evident by increased gene expression of PDCD1, LAG3, HAVCR2, CTLA4, TIGIT, CD160, LILRB4 (Fig. 1G). Using hallmark and gene ontology gene sets, we further determined that cardiac T-cells present in DCM patients (as compared to controls) exhibited decreased oxidative phosphorylation and fatty acid metabolism (Fig. 1E and Fig. 1H), but increased pro-inflammatory TNFα/NFκB, and pro-fibrotic TGFβ signaling (Fig. 1H). Similar changes in metabolic pathways were also observed in data derived from single-nuclei RNA sequencing (Supplement Fig. 1). Altered metabolic shifts characterized by decreased aerobic glycolysis and increased pro-inflammatory phenotype are consistent with previous preclinical studies showing changes in metabolism in activated and proliferative T-cells during chronic inflammatory conditions [30] and persistent antigenic stimulation [31].

Fig. 1.

A) Percentage of CD45⁺ leukocytes and B) CD3⁺ T-cells in each donor heart tissue derived from control (n = 27) and DCM patients (n = 18), p-values shown were determined using multiple linear regression after controlling for batch differences. C) Proportion of CD3⁺ T-cells in each phase of the cell cycle, predicted using CellCycleScoring, in the failing DCM vs non-failing control hearts. Differences between conditions were tested using a permutation test, *p < 0.05. D) Heatmap of differentially expressed genes and E) dot plot showing top activated and suppressed gene sets in cardiac CD3⁺ T-cells infiltrated in DCM hearts. F) Enrichment plot showing ranking of genes from the ‘Activated T-cell proliferation’ (GOBP:0050798) gene set comparing the gene expression in CD3⁺ T-cells present in the failing human hearts as compared to those present in the control hearts. Significance is reported as the False Discovery Rate (FDR). G) Ridge plots showing cell normalized enrichment scores for exhausted CD3⁺ T-cells (PDCD1, LAG3, HAVCR2, CTLA4, TIGIT, CD160, LILRB4) in the scRNA seq data. Significance is reported as the False Discovery Rate (FDR). H) Violin plots showing normalized enrichment scores for hallmark gene sets associated with oxidative phosphorylation, fatty acid metabolism, glycolysis, TNF/NFkB, and TGFβ signaling in scRNA seq data of CD3⁺ T-cells in DCM vs control hearts. Each dot represents an individual patient/control sample and p-values are listed in each figure. Significance is reported as the False Discovery Rate (FDR). *p (or q for FDR) <0.05 is considered significant.

3.2. DCM is characterized by distinct CD4⁺ helper and CD8⁺ cytotoxic T-cell populations within the heart

Previous preclinical studies have shown the presence of both CD4⁺ helper and CD8⁺ cytotoxic T-cells during MI and chronic HF [32,33]. Thus, using Uniform Manifold Approximation and Projection (UMAP) clustering on single-cell data set we separated all CD3⁺ T-cells into different subsets and found 8 phenotypically discrete T-cell sub-populations in failing DCM hearts (Fig. 2A). These subpopulations were categorized as either helper or cytotoxic T-cells based on the expression of CD4 (clusters 1, 2, 4 and 4) or CD8α (clusters 5, 6, 7 and 8), respectively, (Fig. 2B and Supplement Fig. 2). Phenotypically similar clusters were also observed with single nuclei data sets (Supplement Fig. 3 A, B and C). Importantly, significantly higher proportions of both CD4⁺ and CD8⁺ T-cells were in the actively dividing G2/M phase (Fig. 2C) with increased expression of T-cell activation and differentiation genes in DCM hearts (Fig. 2D). The top differentially up- and down-regulated genes in both CD4⁺ and CD8⁺ T-cells are shown in Supplement Fig. 4 A. While pathways associated with increased IL-2 production, NFκB signaling, Th cell activation, differentiation and proliferation, and Th-17 mediated immune responses were upregulated in CD4⁺ T-cells, pathways associated with protein acetylation and RNA splicing were seen in CD8⁺T-cells (Supplement Fig. 4B). Phenotypic analysis of the four CD4⁺ T-cell clusters identified using UMAP further showed that while cluster 1 consisted of naïve CD4⁺ T-cells, clusters 2 and 3 were activated CD4⁺ T-cells, and cluster 4 comprised of regulatory T-cells (FoxP3⁺ Tregs) (Fig. 2E). Among activated CD4⁺ T-cells, while cluster 2 demonstrated high expression (as compared to cluster 3) of activation genes like LTA, LTB, TNF, CD83, CD7, and IL-23, cluster 3 exhibited downregulation of egress marker genes like MAL, CCR7, and SELL with high expression of PDCD1, DUSP4, SAMSN1, and COTL1 suggesting that this cluster was enriched in tissue resident memory T-cells. Similar profiles were also seen in single-nuclei RNA seq data set (Supplement Fig. 3C). Relative proportions of different T-cell clusters are shown in Fig. 2F. DCM hearts showed higher proportions of active TRM-like CD4⁺ T-cells (cluster 3: 0.12 ± 0.08 vs 0.22 ± 0.02, p < 0.001) and FoxP3⁺ Tregs (cluster 4, 0.023 ± 0.008 vs 0.039 ± 0.010, p = 0.02). We also observed a lower proportion of T-cells within the active CD4⁺ cluster (cluster 2: 0.27 ± 0.08 vs 0.09 ± 0.01, p = 0.04) which could be due to their dynamic transition into the TRM phenotype by increasing expression of tissue-resident markers. None of the CD8⁺ T-cell subsets (Effector Memory, T cells expressing inhibitory killer cell immunoglobulin-like receptors [KIR], resting, or active) were found to be different between the DCM and controls hearts (Fig. 2F). The identified increased levels of CD4⁺ T-cells and Tregs are consistent with previous preclinical studies published by us [12,16] and others [10]. We have also shown a pro-inflammatory and dysfunctional phenotype of Tregs present in rodent failing hearts [12], however, due to low numbers of Tregs in the single-cell RNA seq data in control hearts (8 cells), detailed phenotypic characterization of Tregs could not be conducted.

Fig. 2.

A) Unsupervised Uniform Manifold Approximation and Projection (UMAP) dimensionality reduction plot of scRNA sequencing data of cardiac CD3⁺ T-cells. B) Violin plots depicting CD4 and CD8α expression in each cluster. C) Proportion of CD4⁺ helper and CD8⁺ cytotoxic T-cells in each phase of the cell cycle determined from the scRNA seq data of T-cells from control (n = 3) and failing DCM hearts (n = 8). Differences between conditions were tested using a permutation test, *p < 0.05. D) Enrichment plots showing ranking of genes upregulated within the ‘T-cell Differentiation’ (GO:0030217) and ‘T-cell Activation’ (GO:0042110) gene sets in scRNA seq of CD4⁺ and CD8⁺ T-cells in DCM hearts. Significance is reported as the False Discovery Rate (FDR) E) Heatmap of signature genes identifying different CD4⁺ T-cell clusters (naïve, activated, TRM and Tregs). F) Proportion of different CD4⁺ and CD8⁺ T-cell subsets in DCM (n = 20) vs control hearts (n = 28). All p-values were calculated using multiple linear regression controlling for batch differences. Data from cells is shown in panels A, B, C, D, and E while each dot represents an individual patient/control sample in panel F. *p (or q for FDR) <0.05 is considered significant.

To further investigate the relationship between cardiac T-cell infiltration and disease severity, we performed multivariate linear regression (controlling for age, sex, and source of data). We found that a higher percentage of T-cells, CD4⁺ T-cells, and TRM CD4⁺ T-cells in the heart corresponded to significantly lower ejection fraction (p = 0.02, p = 0.006, p < 0.001 respectively), as detailed in Supplement Fig. 5.

3.3. CD4⁺ helper T-cells within the DCM hearts are pro-inflammatory and pro-fibrotic

To further explore the transcriptional changes in cardiac CD4⁺ T-cells during DCM, we compared the gene expression of active TRM-like CD4⁺ T-cells (cluster 3) with naive CD4⁺ T-cells (cluster 1). Volcano plots showed downregulation of several genes associated with tissue egress such as CCR7, SELL, MAL, LEF1 and SATB1, and upregulation of exhaustion genes like DUSP4, PDCD1, SAMSN1, ARHGDIB, and CD2 (Fig. 3A) further supporting tissue-resident memory phenotype of cardiac T-cells during DCM [22]. GSEA pathway analysis further showed increased signaling of pathways associated with IFNα/γ response, hypoxia, TNFα/NFκB, IL-2/STAT5, TGFβ and auto-immune like allograft rejection in the active TRM CD4⁺ T-cells (Fig. 3B). Several cytokine/chemokine mediators of these pathways, including TGFβ, CCL4, IL2, and IL4 were upregulated in CD4⁺ T-cells in DCM, while others like LTB, LTA, and IL23A were found to be downregulated (Fig. 3C). Normalized cell counts for significantly up- or down-regulated cytokines and chemokines in active CD4⁺ T-cells are shown in Supplement Fig. 6. We also performed hdWGCNA analysis to identify potentially biologically relevant gene networks within the activated TRM-like CD4⁺ T-cells present in the DCM hearts (Fig. 3D). Four highly co-expressed gene clusters were identified (Fig. 3E), and top biological processes each gene cluster is most enriched in is shown in Fig. 3F. While cluster 1 was enriched in gene pathways related to ATP biosynthetic processes, cluster 2 was associated with promotion of chemotaxis of other immune cells such as dendritic cells and natural killer cells, cluster 3 was associated with increased Golgi lumen acidification, and cluster 4 was involved in the negative regulation of extracellular matrix (ECM). These upregulated gene clusters suggest a critical role of activated TRM-like T-cells in ECM remodeling and recruitment of other immune cells to the failing hearts.

Fig. 3.

A) Volcano plot for differentially expressed genes in the cardiac CD4⁺ T-cells in DCM patients (n = 8), and B) dot plot of the top activated gene sets from the hallmark gene sets. C) Volcano plot showing up-regulated and down-regulated cytokines in TRM-like T-cells when compared with ‘naïve’ CD4⁺ T-cells (clusters 3 and 1, respectively, from Fig. 2A) identified in scRNA seq of DCM hearts. D) Dendrogram, and E) identified gene clusters from high dimensionality Weighted Gene Co-expression Network Analysis (hdWGCNA) of the TRM-like CD4⁺ T-cells present in DCM hearts. F) Dot plot showing top three gene ontology biological processes associated with the four identified gene clusters. Data from cells is shown in all panels.

3.4. Activated CD4⁺ T-cells exhibit upregulated ESR1 signaling in DCM Hearts

We have previously shown that in mice estrogen receptor (ER)α (ESR1) mediated signaling is involved in pro-inflammatory activation and proliferation of CD4⁺ T-cells during HF [34]. Importantly, significantly upregulated DEGs in Fig. 3B also showed increase ‘early’ and ‘late’ estrogen responses. Thus, we utilized Ingenuity Pathway Analysis (IPA) to measure estrogen signaling in activated TRM-like CD4⁺ T-cells present in the failing human hearts. IPA analysis identified 548 genes associated with ER signaling that were differentially expressed (269 upregulated and 279 downregulated) in activated TRM-like CD4⁺ T-cells and, as shown in Fig. 4A, predicted activation of ESR1 (z-score: 1.592, p = 4.96E-8), and inhibition of ESR2 (ERβ; z-score: −2.106, p = 1.27E-10). IPA analysis also predicted that the altered genes which are regulated by ESR1 and ESR2 promote ‘cell migration’ and ‘fibrosis’ (p = 1.8E-25 and 4.7–25 respectively). Similar to our previously published preclinical studies [35], ER-regulated gene signatures in CD4⁺ T-cells present in DCM hearts also predicted T-cell activation, maturation and increased pro-inflammatory cytokine signaling (Fig. 4B). The top 20 upregulated ESR1 controlled genes and the top 20 downregulated ESR2 controlled genes found in TRM-like active CD4⁺ T-cells are shown in Fig. 4C.

Fig. 4.

A) Graphical summary of Ingenuity Pathway Analysis (IPA) of differentially expressed genes in the TRM CD4⁺ T-cells, as compared to naïve, in scRNA seq data of DCM hearts. The p-value of overlap was calculated using a right-tailed Fisher’s exact test B) IPA network plot of up- and down-regulated genes involved in ESR1 and ESR2 signaling and their predicted activation or inhibition functions in activated cardiac TRM CD4⁺ T-cells, as compared to ‘Naïve’ CD4⁺ T-cells, in DCM hearts. The reported p-value is calculated using a right-tailed Fisher’s exact test. C) Heatmaps of the top 20 upregulated genes in the ESR1 and top 20 downregulated genes in the ESR2 pathway. Data from cells is shown in all panels.

3.5. Cardiac levels of ERα + CD4⁺ T-cells correlate with cardiac dysfunction during DCM

To further validate ERα expression in cardiac CD4⁺ T-cells, we conducted immunohistochemical analysis from the left ventricular (LV) apex of 9 DCM patients (Age 58 ± 14 years, 3 (33 %) female, n = 3 each of ischemic, non-ischemic, and mixed CM etiology). As shown in Figs. 5A and 5B, 71 ± 12 % of infiltrated CD4⁺ T-cells exhibited ERα expression either in the cytoplasm or the nucleus. Control heart obtained from a cadaver did not show any CD4⁺ T-cells (Supplement Fig. 7 A). Importantly, the cardiac ERα⁺CD4⁺ T-cell fraction (of total CD4⁺ T-cells), varied inversely to the measures of LV systolic function (ejection fraction (EF) (r² = 0.50 and p = 0.033), LVOT VTI; r² = 0.485 and p = 0.037). ERα⁺CD4⁺ T-cell numbers also correlated directly with right ventricular systolic pressure (RVSP; r² = 0.648 and p = 0.009), consistent with compensatory RV functional changes that are known to occur in left heart disease [36–39]. These findings together with our bioinformatic data suggest a critical role for ERα-mediated signaling mechanisms in CD4⁺ T-cell activation, and/or their trafficking to the heart during the development and/or maintenance of the DCM phenotype.

Fig. 5.

A) Representative merged confocal microscopy image showing CD4 (red) and estrogen receptor (ER)α (yellow) expression in cells infiltrated into the failing hearts. Hoechst was used to label nuclei (blue) while wheat-germ agglutinin was used to label cell membranes (green). B) A small area was magnified to show individual channels and associated merged image is shown in C). D) Simple linear regression showing correlation between the ejection fraction (EF), left-ventricular outflow tract velocity time integral (LVOT VTI), and right ventricular systolic pressure (RVSP) with cardiac ERα expressing CD4⁺ T-cells (n = 9). E) Gene expression in cardiac fibroblasts after co-culture with CD4⁺ T-cells (n = 4) treated either with propyl pyrazole triol (PPT), an ERα agonist, or Fulvestrant, an ER antagonist. Each dot represents an individual patient sample and all p-values are listed in the figures. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.6. ERα signaling in CD4⁺ T-cells promote pro-fibrotic responses in cardiac fibroblasts

To confirm the role of CD4⁺ T-cell specific ERα signaling in mediating profibrotic cardiac remodeling, circulating CD4⁺ T-cells from HF patients (n = 4) were co-cultured with primary human cardiac fibroblasts. As shown in Fig. 5E, CD4⁺ T-cells treated with an ERα agonist, PPT, accentuated pro-fibrotic signaling in cardiac fibroblast and significantly increased the gene expression of ACTA1 and Collagen 1A1 (with increasing trends in TGFβ and CCN2). Alternatively, inhibition of ER signaling by Fulvestrant in CD4⁺ T-cells nullified PPT mediated increases in pro-fibrotic genes in fibroblasts. These results suggest that enhanced ERα signaling in CD4⁺ T-cells plays an important role in promoting a pro-fibrotic phenotype in cardiac fibroblasts.

4. Discussion

Preclinical models of ischemic and non-ischemic cardiac injury show pathological activation and pro-inflammatory phenotype of CD4⁺ T-cells as one of the key determinants for LV remodeling and interstitial fibrosis [10,12]. Although, recent studies have shown the presence of antigenically activated T-cells in the systemic circulation of patients post-MI [20,21], it is not known if increased levels of activated T-cells are also present in the chronically failing human hearts. Importantly, the molecular mechanisms that facilitate pathological activation of T-cells during cardiac remodeling and systolic dysfunction in humans are unknown. In this study, we show for the first time that human failing hearts demonstrate increased levels of activated and proliferating CD3⁺T-cells. These T-cells show a pro-inflammatory phenotype with reduced oxidative/aerobic metabolism consistent with mechanisms of T-cell activation in hypoxic and inflammatory environments [30,31]. Interestingly, T-cells present in failing human hearts have increased expression of genes associated with T-cell exhaustion, a state of dysfunction that arises in T-cells under chronic inflammatory conditions due to continuous antigenic stimulation [40]. Expression of inhibitory receptors is critical to avoid auto-immunity and effector T-cells often exhibit transient expression of inhibitory surface receptors such as PD-1, Lag-3 CTLA-4 and CD160 to maintain self-tolerance. However, persistent expression of multiple inhibitory receptors and downregulation of metabolic pathways like oxidative phosphorylation [41,42], as seen in T-cells present in DCM hearts, is a key feature of exhausted T-cells [40]. Exhausted T-cells have been commonly studied in the tumor microenvironment [43], autoimmune diseases [44], and chronic viral infections [41]. However, their role in mediating T-cell pathogenicity during HF remains unexplored. While this data is strictly observational, it presents promising avenue for future studies to delineate signaling mechanisms that mediate pathological CD4⁺ T-cell activation during HF.

Phenotypic characterization and activation of CD4⁺ helper and CD8⁺ cytotoxic T-cells in human HF has mostly been studied in circulating T-cells [13,45], and no studies have been conducted to characterize them in the failing human hearts. Human T-cells are largely comprised of non-tissue resident effector or central memory (TEM or TCM, respectively) or tissue-resident memory (TRM) T-cells with distinct phenotypes and surface expression patterns [46]. However, phenotypes of different T-cells in failing human hearts are not known. Our bioinformatic analysis identified several phenotypically different subsets of both helper (CD4⁺) and cytotoxic (CD8⁺) T-cells in the human failing hearts. While both CD4⁺ and CD8⁺ T-cells show an upregulation of genes involved in T-cell activation and proliferation, only TRM like CD4⁺ T-cells with reduced expression of tissue egress markers such as CCR7, SELL and MAL, and increased expression of PDCD1, DUSP4, COTL1 and SAMSN, and Tregs were found to be increased in DCM hearts. Whether changes in these gene patterns to promote TRM phenotype in cardiac CD4⁺ T-cells are a functional adaptation specific to the cardiac milieu to reflect antigenic stimulation, or merely a reflection of T-cell specific mechanisms activated to promote their transmigration with concomitant suppression of egress from the heart is unknown. Nonetheless, the transcription profile of cardiac CD4⁺ T-cells in DCM hearts suggested increased proliferative (as evident by increased proportion of cells in G2M phase), pro-fibrotic (TGFβ Signaling), pro-inflammatory (IFNα/γ response, and TNFα/NFκB signaling), chemotactic and ECM remodeling phenotype, as evidenced by changes in the expression of gene networks associated with proliferation, cytoskeletal, and cell-matrix interactions. Increased IFNα/γ and TNFα signaling in activated cardiac CD4⁺ T-cells is consistent with observations from preclinical studies showing increased infiltration of Th1 T-cells in failing rodent hearts [10,12] (pressure induced or ischemic). Although, our results suggest proinflammatory activation and proliferation of CD4⁺ T-cells, we could not ascertain the functional role of CD4⁺ T-cells present in the failing human hearts. However, considering preclinical evidence shows a pathological role of phenotypically similar and activated CD4⁺ T-cells, as demonstrated by their ability to induce systolic dysfunction upon adoptive transfer to naïve mice [12,19], similar inferences can be drawn for human T-cells as well.

Estrogen is a known cardioprotective hormone with significant epidemiological data showing decreased risk of cardiovascular disease in pre-menopausal women as compared to men or post-menopausal women [47]. Although, ERs are expressed in cardiomyocytes in humans [48] and rodents [49], a recent study has shown that cardiac ERα or ERβ expression is dispensable for cardioprotective effects seen with estrogen in rodents [50]. ERs are also widely expressed on immune cells with the highest expression observed in human CD4⁺ T-cells and heightened ERα signaling in females has been linked to increased antigenic CD4⁺ T-cell activation during autoimmune disease in humans and rodents [34,51]. However, it is not known if the estrogen mediated sex-bias in CVD could partly be due to its effects on T-cell activation and proliferation. Using preclinical mouse models of HF, we recently showed that ERα signaling is important for the activation, proliferation, and expression of pro-inflammatory cytokines in pathologically activated CD4⁺ T-cells during HF [34]. Importantly, activation of ERβ signaling to functionally antagonize ERα suppresses T-cell receptor mediated activation and proliferation of CD4⁺ T-cells and blunts progressive cardiac dysfunction and LV remodeling during HF in both sexes [35]. Expanding on these preclinical studies, we further show the presence of ERα⁺CD4⁺ T-cells in the DCM hearts and demonstrate that their levels are significantly associated with attenuated EF and LVOT VTI, and increase in RVSP during HF. Consistent with our published preclinical studies [35], our data show activation of ERα signaling in human CD4⁺ T-cells to promote pro-fibrotic gene expression in cardiac fibroblasts further supporting a pro-fibrotic and pathological role of ERα signaling mediated pathological transitioning of CD4⁺ T-cells during DCM. Importantly, our findings of increased ‘early’ and ‘late’ estrogen responses suggest continuous and sustained activation of genes responsible for acute (associated with early ERα mediated gene expression changes) as well as chronic estrogenic responses. This is also evident from clinical studies suggesting that very high and very low levels of estradiol are associated with increased mortality in men with systolic HF [52]. This coupled with our data showing upregulation of genes involved in ERα signaling in T-cells infiltrated into the failing human DCM hearts further validate our published preclinical studies and reinforce the importance of ERs in regulating pathological T-cell activation during HF. Thus, aberrant estrogen signaling within T-cells is a potential targetable mechanism for HF in humans and selective ERβ agonists, such as OSU-ERb-012 [35], could be developed as safe therapeutics to quell T-cell responses during HF.

5. Conclusions

Here, we validate unique T-cell mediated mechanisms found in preclinical studies by showing that cardiac T-cells in DCM patients are activated with a pro-inflammatory and pro-fibrotic phenotype. Due to limitations associated with human studies, we could not ascertain the functional role of these T-cells. However, there is sufficient evidence from preclinical studies supporting the pathologic nature of phenotypically similar and activated T-cells, supporting that T-cells present in human failing hearts may also be involved in DCM progression. We also show an increased TRM-like phenotype of cardiac T-cells with upregulated T-cell exhaustion and ERα signaling during DCM. Identification of potent mechanisms involved in pathological T-cell activation during HF can provide novel therapeutic targets to blunt progressive cardiac dysfunction and ameliorate LV remodeling during HF.

5.1. Limitations

Functional relevance of T-cells in DCM hearts could not be determined due to limitations associated with human studies, and only correlative findings are presented.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.yjmcc.2025.12.011.

Data availability

Data is available through the GEO database (GSE183852 and GSE145154).