A spatially resolved single-cell lung atlas integrated with clinical and blood signatures distinguishes COVID-19 disease trajectories

João Da Silva Filho,Vanessa Herder,Matthew P Gibbins,Monique Freire Dos Reis,Gisely Cardoso Melo,Michael J Haley,Carla Cristina Judice,Fernando Fonseca Almeida Val,Mayla Borba,Tatyana Almeida Tavella,Vanderson de Sousa Sampaio,Charalampos Attipa,Fiona McMonagle,Derek Wright,Marcus Vinicius Guimaraes de Lacerda,Fabio Trindade Maranhão Costa,Kevin N Couper,Wuelton Marcelo Monteiro,Luiz Carlos de Lima Ferreira,Christopher Alan Moxon,Massimo Palmarini,Matthias Marti

DOI: https://doi.org/10.1126/scitranslmed.adk9149

2024-09-11

Abstract:COVID-19 is characterized by a broad range of symptoms and disease trajectories. Understanding the correlation between clinical biomarkers and lung pathology during acute COVID-19 is necessary to understand its diverse pathogenesis and inform more effective treatments. Here, we present an integrated analysis of longitudinal clinical parameters, peripheral blood markers, and lung pathology in 142 Brazilian patients hospitalized with COVID-19. We identified core clinical and peripheral blood signatures differentiating disease progression between patients who recovered from severe disease compared with those who succumbed to the disease. Signatures were heterogeneous among fatal cases yet clustered into two patient groups: "early death" (<15 days until death) and "late death" (>15 days). Progression to early death was characterized systemically and in lung histopathological samples by rapid endothelial and myeloid activation and the presence of thrombi associated with SARS-CoV-2+ macrophages. In contrast, progression to late death was associated with fibrosis, apoptosis, and SARS-CoV-2+ epithelial cells in postmortem lung tissue. In late death cases, cytotoxicity, interferon, and T helper 17 (TH17) signatures were only detectable in the peripheral blood after 2 weeks of hospitalization. Progression to recovery was associated with higher lymphocyte counts, TH2 responses, and anti-inflammatory-mediated responses. By integrating antemortem longitudinal blood signatures and spatial single-cell lung signatures from postmortem lung samples, we defined clinical parameters that could be used to help predict COVID-19 outcomes.