Post-mortem lung tissue: the fossil record of the pathophysiology and immunopathology of severe COVID-19

Abstract

The lungs are the main site that is affected in severe COVID-19, and post-mortem lung tissue provides crucial insights into the pathophysiology of severe disease. From basic histology to state-of-the-art multiparameter digital pathology technologies, post-mortem lung tissue provides snapshots of tissue architecture, and resident and inflammatory cell phenotypes and composition at the time of death. Contrary to early assumptions that COVID-19 in the lungs is a uniform disease, post-mortem findings have established a high degree of disease heterogeneity. Classic diffuse alveolar damage represents just one phenotype, with disease divisible by early and late progression as well as by pathophysiological process. A distinct lung tissue state occurs with secondary infection; extrapulmonary causes of death might also originate from a pathological process in the lungs linked to microthrombosis. This heterogeneity of COVID-19 lung disease must be recognised in the management of patients and in the development of novel treatment strategies.

Figure 1

Haematoxylin and eosin staining of post-mortem lung sections from patients who died with COVID-19 (A) Exudative diffuse alveolar disease featuring prominent hyaline membranes (arrow) with mild interstitial infiltration by mononuclear cells, alveolar wall congestion, and cellular debris (original magnification ×400). (B) Perivascular lymphocytic infiltrate (arrow; ×400). (C) Organising diffuse alveolar disease with myofibroblast proliferation (arrowhead), fibrinous exudates, hyaline membrane remnants (arrow), and substantial alveolar architectural disruption (×200). (D) Acute bronchopneumonia featuring prominent intra-alveolar neutrophilic infiltration (×200). (E) Pulmonary cardiac oedema, characterised by eosinophilic fluid within the alveolar airspace (arrowhead), alveolar wall congestion (arrow), and absence of cellular damage (×200). (F) Platelet-rich thrombus (arrow) causing expansion of an intra-acinar vessel (×200).

Figure 2

Application of imaging mass cytometry to lung tissue from patients who died with COVID-19 Lung sections are stained with a panel of rare earth metal-labelled antibodies, laser-ablated into an ionised plume, and spatially identified on the basis of particle mass. (A) Multiplexed images resembling immunofluorescence images are generated, with each colour representing a unique antibody target. (B) Analysis software is used for single-cell segmentation based on antibody binding, resulting in each cell being covered by a distinct mask. (C) Once segmented, analyses include cell clustering analysis according to expression profiles (top left panel), heatmap analysis based on differential expression (top right panel), and spatial analysis of neighbouring cell–cell interactions (middle and bottom panels) and avoidances (ie, the apparent absence of interactions between cells that usually interact).

Figure 3

Application of GeoMx digital spatial profiling to lung tissue from patients who died with COVID-19 Lung sections were stained with the morphology markers pan-cytokeratin (green) and CD68 (red), and with a nuclear stain (blue). (A) Representative section showing pathologist-guided ROIs (small purple circles or green squares). Cells within ROIs can be analysed in their entirety for protein expression or mRNA abundance. (B) A single ROI (yellow box in part A) enlarged to show segmentation analysis. Isolating cell populations are defined by morphology markers (CD68 segmentation mask, pink; excluded cells, purple). (C) CD68 and pan-cytokeratin expression in CD68⁺ segment (top) and CD68⁻ segment (bottom). (D) Heatmap showing inter-patient heterogeneity in expression of 45 immune-associated protein targets. Each column represents a single ROI and eight ROIs were examined for each for the nine patients studied (colour coded as above). ROI=region of interest.

Figure 4

Outline of a workflow combining conventional and advanced pathology approaches Post-mortem lung tissue sections from patients who died with COVID-19 have shown diffuse alveolar disease, bronchopneumonia, or tissue thrombosis (alone or in combination) using conventional pathology techniques. Advanced tissue pathology technologies are used to generate multiplexed images that are amenable to deep bioinformatics analyses, including cell clustering analysis, heatmap analysis showing expression patterns, and interactional analyses. As the investigation progresses, the cells present at the time of death are identified and characterised to determine the cell types, their functional states, and interactions and avoidances. Validation of findings might be sought using conventional pathology techniques. The ultimate goal is to identify clues (ie, mechanisms of action and targets) that could inform future treatment strategies. Some figure components were created with BioRender.com.