ABSTRACT
Severe lung damage in COVID-19 is known to involve complex interactions between diverse populations of immune and stromal cells. In this study, we applied a spatial transcriptomics approach to better delineate the cells, pathways and genes responsible for promoting and perpetuating severe tissue pathology in COVID-19 pneumonitis. Guided by tissue histology and immunohistochemistry we performed a targeted sampling of dozens of regions representing a spectrum of diffuse alveolar damage from the post-mortem lung of three COVID-19 patients. Application of a combination of differential gene expression, weighted gene correlation network, pathway and spatial deconvolution analysis stratified the sampled regions into five distinct groups according to degree of alveolar damage, levels of cytotoxic inflammation and innate activation, epithelial reorganization, and fibrosis. Integrative network analysis of the identified groups revealed the presence of proliferating CD8 T and NK cells in severely damaged areas along with signatures of cytotoxicity, interferon signalling and high expression of immune cell chemoattractants (including CXCL9/10/11 and CCL2). Areas of milder damage were marked by innate immune signalling (including TLR response, IL-1, IL-6) together with signatures of antigen presentation, and fibrosis. Based on these data we present a cellular model of tissue damage in terminal COVID-19 that confirms previous observations and highlights novel opportunities for therapeutic intervention.