Eosinophil role and activation state in hypereosinophilia-associated heart disease

Usman Sunusi¹Immaculeta Osuji¹Benjamin Ziegelmeyer¹Mario Medvedovic¹Haley Todd¹Joe Abou-Khalil²Nives Zimmermann^1*

• ¹University of Cincinnati, Cincinnati, United States
• ²Yale University, New Haven, United States

Background: Cardiac complications in patients with hypereosinophilia cause significant morbidity and mortality. However, mechanisms of how eosinophilic inflammation causes heart damage are poorly understood. Methods: We developed a model of hypereosinophilia-associated heart disease by challenging hypereosinophilic mice with peptide from the cardiac myosin heavy chain. Disease outcomes were measured by histology, immunohistochemistry, flow cytometry, and measurement of cells and biomarkers in peripheral blood. Eosinophil dependence was determined by using eosinophil-deficient mice (DdblGATA). Single cells from the heart were subjected to single cell RNA sequencing to assess cell composition, activation states and expression profiles. In vitro studies used bone marrow derived eosinophils (BMDeos), and stimulated them with cytokines and pathogen-associated molecular patterns, followed by assessment of activation markers by flow cytometry. Results: Mice challenged with myocarditic and control peptide had peripheral blood leukocytosis, but only those challenged with myocarditic peptide had heart inflammation. Heart tissue was infiltrated by eosinophil-rich inflammatory infiltrates associated with cardiomyocyte damage. Disease penetrance and severity were decreased in eosinophil-deficient mice. Single cell RNA sequencing showed enrichment of myeloid cells, T-cells and granulocytes (neutrophils and eosinophils) in the myocarditic mice. Focusing on eosinophils, there was increased expression of genes associated with type 1 cell activation (such as CD274/PDL1), complement activation and pathogen-associated molecular pattern recognition. To verify findings generated by single cell RNA sequencing on a protein level, we performed flow cytometry analysis and assessed the level of type 1 and type 2 biomarkers CD274 and CD101 respectively. The proportion of cells expressing surface CD274 increased on both neutrophils and eosinophils, particularly in mice which showed inflammation by histology. There was no significant increase in expression of CD101. Finally, we assessed whether activation markers can be induced on eosinophils in vitro. IFNg markedly increased expression of CD274, consistent with type 1 polarization. Furthermore, BMDeos stimulated with LPS showed a concentration-dependent increase in the level of CD274 expression. Conclusion: Eosinophils are required for heart damage in hypereosinophilia-associated heart disease. Heart infiltrating eosinophils in an inflammatory condition show type 1 activation, which can be recapitulated in vitro.