Molecular Pathways and Cellular Subsets Associated with Adverse Clinical Outcomes in Overlapping Immune-Related Myocarditis and Myositis

Abstract

Immune checkpoint therapies (ICT) can induce life-threatening immune-related adverse events, including myocarditis and myositis, which are rare but often concurrent. The molecular pathways and immune subsets underlying these toxicities remain poorly understood. To address this need, we performed single-cell RNA sequencing of heart and skeletal muscle biopsies obtained from living patients with cancers treated with ICTs and admitted to the hospital with myocarditis and/or myositis (overlapping myocarditis plus myositis, n = 10; myocarditis-only, n = 1) or ICT-exposed patients ruled out for toxicity utilized as controls (n = 9). All biopsies were obtained within 96 hours of clinical presentation. Analyses of 58,523 cells revealed CD8+ T cells with a cytotoxic phenotype expressing activation/exhaustion markers in both myocarditis and myositis. Furthermore, the analyses identified a population of myeloid cells expressing tissue-resident signatures and FcγRIIIa (CD16a), which is known to bind IgG and regulate complement activation. Immunohistochemistry of affected cardiac and skeletal muscle tissues revealed protein expression of pan-IgG and complement product C4d, which were associated with the presence of high-titer serum autoantibodies against muscle antigens in a subset of patients. We further identified a population of inflammatory IL1B+TNF+ myeloid cells specifically enriched in myocarditis and associated with greater toxicity severity and poorer clinical outcomes. These results provide insight into the myeloid subsets present in human immune-related myocarditis and myositis tissues and nominate new targets for investigation into rational treatments to overcome these high-mortality toxicities. See related Spotlight by Fankhauser et al., p. 954.

Figure 1:. Overall cellular landscape of immune checkpoint therapy (ICT)-induced myocarditis and myositis.

A. Overview of enrolled patients and collected specimens. B. Uniform Manifold Approximation and Projection (UMAP) plot of live cells obtained from normal cardiac muscle tissue. C. UMAP plot of live cells from myocarditis tissue (downsampled to match cell count with control cardiac muscle samples). D. UMAP plot of live cells from normal skeletal muscle tissue. E. UMAP plot of live cells myositis tissue (downsampled to match cell count with control skeletal muscle samples). F. Top upregulated and downregulated differentially expressed genes in CD45⁺ cells in myocarditis versus myositis samples. G. Differentially expressed hallmark transcriptional pathways in total CD45⁺ immune cells between myocarditis and control cardiac muscle tissue. H. Differentially expressed hallmark transcriptional pathways in total CD45⁺ immune cells between myositis and control skeletal muscle tissue. I. Volcano plot of differentially expressed genes in CD45+ immune cells in cardiac muscle. J. Volcano plot of differentially expressed genes in CD45+ immune cells in skeletal muscle. K. Differentially expressed hallmark transcriptional pathways in patient-matched myocarditis and myositis tissue.

Figure 2:. Immune cell subsets in ICT-induced myocarditis and myositis.

A. UMAP projection of immune cell subclusters in cardiac muscle. B. Dot plot indicating expression of individual genes utilized to define subclusters in cardiac muscle. C. Comparison of individual cluster frequency in myocarditis versus control samples. D. UMAP projection of immune cell subclusters in skeletal muscle. E. Dot plot indicating expression of individual genes utilized to define subclusters in skeletal muscle. F. Comparison of individual cluster frequency in myositis versus control samples ranked by Log2FC. G. Cluster dendrogram of CD45⁺ immune cells in cardiac and skeletal muscle samples. H. UMAP projection of patient-matched myocarditis and myositis samples. I. Feature plot of patient-matched samples demonstrating relative abundance of cell clusters in myocarditis and myositis samples.

Figure 3.. Cytotoxic CD8⁺ T cells expressing activation markers are enriched in myocarditis and myositis.

A. UMAP projection of individual T cell clusters in cardiac muscle samples. B. Dot plot indicating expression of individual genes utilized to define T cell subclusters in cardiac muscle. C. Individual T cell subclusters present in differential frequencies in myocarditis versus control cardiac muscle specimens (red indicates patient with matched myocarditis and myositis sample available). D. UMAP projection of individual T cell clusters in skeletal muscle samples. E. Dot plot indicating expression of individual genes utilized to define T cell subclusters in skeletal muscle. F. Individual T cell subclusters present in differential frequencies in myositis versus control skeletal muscle specimens. G. Top differentially expressed transcriptional pathways within cytotoxic CD8⁺ T cell cluster in myocarditis versus control cardiac muscle tissue. H. Top differentially expressed transcriptional pathways within cytotoxic CD8⁺ T cell cluster in myocarditis versus control skeletal muscle tissue.

Figure 4.. Comparison of cytotoxic CD8⁺ T cell subset in patient-matched myocarditis and myositis tissues.

A. Feature plot of all immune cell subclusters in patient-matched myocarditis and myositis samples. B. Feature plot of cytotoxic CD8+ T cell subclusters in patient-matched myocarditis and myositis samples. C. Frequency of cytotoxic CD8+ T cell subpopulations in patient-matched myocarditis versus myositis samples. D. Differentially upregulated transcriptional pathways within cytotoxic CD8+ T cell population in patient-matched myocarditis vs. myositis.

Figure 5.. Clonally expanded cytotoxic CD8+ T cells expressing activation markers are shared in myocarditis and myositis.

A. TCR clonal overlap in matched cardiac and skeletal muscle in patient with myocarditis and myositis. B. T and NK cell clusters from the 5’ scRNA-seq dataset. C. Mapping of TCR clonotypes to specific T and NK cell clusters. D. Marker gene expressions in T and NK cell clusters from the 5’ scRNA-seq dataset.

Figure 6.. IL-1B⁺TNF⁺ myeloid cells are specifically enriched in ICT-induced myocarditis.

A. UMAP projection of individual myeloid cell clusters in cardiac muscle samples. B. Dot plot indicating expression of individual genes utilized to define myeloid cell subclusters in cardiac muscle. C. Frequency of individual myeloid cell subclusters in myocarditis versus control cardiac muscle specimens (red indicates patient with matched myocarditis and myositis sample available). D. Violin plot of complement associated gene expression in immune cell subsets in myocarditis samples. E. UMAP projection of individual myeloid cell clusters in skeletal muscle samples. F. Dot plot indicating expression of individual genes utilized to define myeloid cell subclusters in skeletal muscle. G. Frequency of individual myeloid cell subclusters in myositis versus control skeletal muscle specimens (red indicates patient with matched myocarditis and myositis sample available). H. Violin plot of complement-associated gene expression in immune cell subsets in myositis samples.

Figure 7.. IL-1B⁺TNF⁺ myeloid cells are associated with poorer clinical outcomes.

A. Feature plot of IL1B+TNF+ myeloid cell subclusters in patient-matched myocarditis and myositis samples (right), with all subclusters shown for reference (left). B. Frequency of IL1B+TNF+ myeloid cell subpopulations in patient-matched myocarditis versus myositis samples. C. Differentially upregulated transcriptional pathways within IL1B+TNF+ myeloid cell population in myocarditis versus control cardiac muscle. D. Differentially upregulated transcriptional pathways within IL1B+TNF+ myeloid cell population in myositis versus control skeletal muscle. E. Differentially upregulated transcriptional pathways within IL1B+TNF+ myeloid cell population in patient-matched myocarditis vs. myositis. F. Association of frequency of IL1B+TNF+ myeloid cells with serum troponin T. G. Association of frequency of IL1B+TNF+ myeloid cells with clinical outcomes. H. Overall survival stratified by frequency of IL1B+TNF+ myeloid cells.