Unbiased single cell spatial analysis localises inflammatory clusters of immature neutrophils-CD8 T cells to alveolar progenitor cells in fatal COVID-19 lungs

Abstract

Single cell spatial interrogation of the immune-structural interactions in COVID -19 lungs is challenging, mainly because of the marked cellular infiltrate and architecturally distorted microstructure. To address this, we developed a suite of mathematical tools to search for statistically significant co-locations amongst immune and structural cells identified using 37-plex imaging mass cytometry. This unbiased method revealed a cellular map interleaved with an inflammatory network of immature neutrophils, cytotoxic CD8 T cells, megakaryocytes and monocytes co-located with regenerating alveolar progenitors and endothelium. Of note, a highly active cluster of immature neutrophils and cytotoxic CD8 T cells, was found spatially linked with alveolar progenitor cells, and temporally with the diffuse alveolar damage stage. These findings provide new insights into how immune cells interact in the lungs of severe COVID-19 disease. We provide our pipeline [Spatial Omics Oxford Pipeline (SpOOx)] and visual-analytical tool, Multi-Dimensional Viewer (MDV) software, as a resource for spatial analysis.

Disease trajectories in hospitalized COVID-19 patients are predicted by clinical and peripheral blood signatures representing distinct lung pathologies

“…In contrast, our data indicate that much of the underlying pathology of the chronic lesions in the post-mortem lung in the late death progression might be due to an inadequate early response to clear the virus, then persistence of high amounts SARS-CoV-2 antigens in the epithelial compartment (even after 20 days of disease). This might lead to high levels of apoptosis, inefficient/dysregulated alveolar epithelial and stromal repair responses, resulting in increased fibrosis (Melms et al, 2021;Olajuyin et al, 2019;Praveen Weeratunga, 2022). Importantly, our analyses cannot determine whether virus positive cells reflect an active (chronic) infection or simply reflect the presence of high levels of residual viral antigen in the absence of virus replication, although the latter seems more likely.…”

“…Differential expression is determined by a spatial permutation test followed by adjust for multiple hypothesis testing using a background null distribution reshuffling the cells within the same cell type (Dries et al, 2021). The SpOOx pipeline was also applied to further validate the statistically significant spatial enrichment, which applies a 3-step spatial association analysis, quadrat correlation matrices (QCMs), cross-pair correlation functions (cross-PCFs) and adjacency cell network (ACN) (Praveen Weeratunga, 2022). The first step, QCM, identifies statistically significant cell co-occurrences (FDR < 0.05) based on correlations of the absolute numbers of cell pairs within square quadrats of up to 100uM.…”

Integrating diverse statistical methods to analyse stage-discriminatory cell interactions in colorectal neoplasia

“…The Quadrat Correlation Matrix (QCM) describes correlations between the counts of different cell types within squares, or 'quadrats', of a specified edge length. First defined in an ecological context (Morueta-Holme et al 2016), it has since been used to analyse correlations between cell types in digital pathology images of colorectal cancer multiplex immunohistochemistry slides (Gatenbee et al 2022) and imaging mass cytometry images of Covid-19 lung sections (Weeratunga et al 2023). The QCM identifies statistically significant co-occurrences within a ROI between cells of different types, comparing the strength of the observed correlation between a given pair of cell types against the expected correlation that would be observed if cell labels were assigned randomly.…”

“…While these spatial statistics describe correlation between points across length scales, other approaches describe correlation within a fixed length scale. For example, ecological methods like the Morisita-Horn index (Hagos et al 2022) and quadrat correlation matrix (QCM) have been used to characterise cell-cell relationships (Morueta-Holme et al 2016;Gatenbee et al 2022;Weeratunga et al 2023) and the cellular composition of distinct tissue regions (Phillips et al 2021;Schürch et al 2020). Network approaches have also been used to examine spatial relationships between vascular and neuronal densities in the brain (Wu et al 2022) and to quantify the numbers of interactions within a connectivity network (Weeratunga et al 2023;.…”

Systematic analysis of YFP traps reveals common mRNA/protein discordance in neural tissues

“…Post-transcriptional regulation is very common across all of the nervous systems, acting hand in hand with transcriptional regulation to create a complex tapestry of protein distribution in time and space. We present our data as a resource that is easily browsable in the context of a rich landscape of genomics, functional, and bioinformatics data using Multi-Dimensional Data Viewer (MDV), an open-source and flexible software platform ( Weeratunga et al, 2022 Preprint ).…”