Phenotyping the virulence of SARS-CoV-2 variants in hamsters by digital pathology and machine learning

Meehan, G. R.; Herder, V.; Allan, J.; Huang, X.; Kerr, K.; Mendonca, D. C.; Ilia, G.; Wright, D.; Nomikou, K.; Gu, Q.; Arias, S. M.; De Lorenzo, G.; Cowton, V.; Upfold, N.; Palmalux, N.; Brown, J. C.; Barclay, W.; Filipe, A. D. S.; Furnon, W.; Patel, A.; Palmarini, M.

doi:10.1101/2023.08.01.551417

Home/Archive/bioRxiv

bioRxiv · 10.1101/2023.08.01.551417

Phenotyping the virulence of SARS-CoV-2 variants in hamsters by digital pathology and machine learning

View original source ↗Export metadata ↓

Abstract

SARS-CoV-2 has continued to evolve throughout the COVID-19 pandemic, giving rise to multiple variants of concern (VOCs) with different biological properties. As the pandemic progresses, it will be essential to test in near real time the potential of any new emerging variant to cause severe disease. BA.1 (Omicron) was shown to be attenuated compared to the previous VOCs like Delta, but it is possible that newly emerging variants may regain a virulent phenotype. Hamsters have been proven to be an exceedingly good model for SARS-CoV-2 pathogenesis. Here, we aimed to develop robust quantitative pipelines to assess the virulence of SARS-CoV-2 variants in hamsters. We used various approaches including RNAseq, RNA in situ hybridization, immunohistochemistry, and digital pathology, including software assisted whole section imaging and downstream automatic analyses enhanced by machine learning, to develop methods to assess and quantify virus-induced pulmonary lesions in an unbiased manner. Initially, we used Delta and Omicron to develop our experimental pipelines. We then assessed the virulence of recent Omicron sub-lineages including BA.5, XBB, BQ.1.18, BA.2 and BA.2.75. We show that in experimentally infected hamsters, accurate quantification of alveolar epithelial hyperplasia and macrophage infiltrates represent robust markers for assessing the extent of virus-induced pulmonary pathology, and hence virus virulence. In addition, using these pipelines, we could reveal how some Omicron sub-lineages (e.g., BA.2.75) have regained virulence compared to the original BA.1. Finally, to maximise the utility of the digital pathology pipelines reported in our study, we developed an online repository containing representative whole organ histopathology sections that can be visualised at variable magnifications (https://covid-atlas.cvr.gla.ac.uk). Overall, this pipeline can provide unbiased and invaluable data for rapidly assessing newly emerging variants and their potential to cause severe disease.

Source connections

Published article ↗

Explore related subjects

microbiology Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Meehan, G. R., Herder, V., Allan, J., Huang, X., Kerr, K., Mendonca, D. C., Ilia, G., Wright, D., Nomikou, K., Gu, Q., Arias, S. M., De Lorenzo, G., Cowton, V., Upfold, N., Palmalux, N., Brown, J. C., Barclay, W., Filipe, A. D. S., Furnon, W., Patel, A., Palmarini, M.. 2023-08-01. Phenotyping the virulence of SARS-CoV-2 variants in hamsters by digital pathology and machine learning. https://doi.org/10.1101/2023.08.01.551417

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Preclinical Development of Enterobacter ludwigii LS57, a Human-Derived Swarming Strain, as a Live Biotherapeutic Product for Inflammatory Bowel Disease

Inflammatory Bowel Disease (IBD) remains an area of major unmet need, and the undefined composition and manufacturing burden of fecal microbiota transplantation have motivated a shift toward single-strain Live Biotherapeutic Products (LBPs). Swarming bacteria are enriched in the stressed gut and limit mucosal damage, but no human-derived swarming strain has been developed as an LBP. We therefore report the complete preclinical development of Enterobacter ludwigii LS57, a swarming strain isolated from human feces. Chemistry, Manufacturing and Controls (CMC) covered a three-tier cell bank, 30 L fed-batch fermentation and a colon-targeted enteric-coated capsule, enabling consistent pilot-scale production. No treatment-related findings emerged in safety pharmacology (Sprague-Dawley rats) or in single-dose and Good Laboratory Practice (GLP)-compliant 1-week repeat-dose toxicity studies (Sprague-Dawley rats and Beagle dogs). LS57 was identified as E. ludwigii by 16S rRNA and whole-genome sequencing and produced metabolites with reported anti-inflammatory activity in culture. In Dextran Sulfate Sodium (DSS)-induced colitis, LS57 reduced weight loss, preserved colon length, lowered colonic pro-inflammatory cytokines and remodeled the gut microbiota. It also attenuated Trinitrobenzene Sulfonic Acid (TNBS)-induced colitis and enhanced the effect of ciclosporin A. In vitro, LS57 was non-hemolytic, non-cytotoxic, and acid-labile, formed no biofilm, and its resistance to aminopenicillins, cephamycins and imipenem was not transferable by conjugation. In vivo, LS57 colonized the inflamed colon without translocation in wild-type and interleukin-10-deficient mice, and experimental bacteremia was treatable with gentamicin or levofloxacin. LS57 combines efficacy across two mechanistically distinct colitis models with a safety package that addresses the principal regulatory concerns for LBPs, and a scalable manufacturing process, compatible with current good manufacturing practice, in which enteric coating overcomes the strain's poor gastric survival. These data support clinical development of LS57 in IBD.

microbiology↗

bioRxiv2026

Characterization of Lactobacillus rhamnosus-Derived Extracellular Vesicles and Their Cellular Cytoprotective Effects Against Trimethylamine-Induced Oxidative Stress in Caco-2 Cells

Dietary trimethylamine (TMA) underwent gut microbiota-mediated catabolism and subsequent hepatic oxidation to generate trimethylamine N-oxide (TMAO), a key metabolic factor linked to vascular and colorectal diseases. While the systemic effects of circulating TMAO were heavily investigated, the direct localized cellular interactions of its precursor, TMA, at the intestinal epithelial level required further elucidation. This study characterized extracellular vesicles isolated from Lactobacillus rhamnosus (Lr-EVs) and evaluated their cytoprotective potential in an in vitro human intestinal Caco-2 cell model under TMA-induced stress. Exposure to a high concentration of TMA (20 mM) significantly reduced Caco-2 viability to 78.8%, triggered intracellular reactive oxygen species (ROS) accumulation, and compromised tight junction integrity as observed via the disruption of Zonula Occludens-1 (ZO-1) alignment. Co-treatment with Lr-EVs (100-300 g/mL) counteracted this cellular toxicity, enhancing cell survival up to 172%. Quantitative PCR analysis revealed that Lr-EVs suppressed the transcriptional upregulation of the pro-inflammatory gene IL-1{beta}, restored the expression of the metabolic regulator SIRT1, and stabilized transcript levels of key mitochondrial-encoded electron transport chain genes (ND6, CYTB, CO1, and ATP6). Additionally, Lr-EVs exhibited intrinsic radical scavenging activities in cell-free ABTS and DPPH assays. These preliminary in vitro findings demonstrated that Lr-EVs mitigated TMA-induced cellular injury, oxidative stress, and junctional disruption at the transcriptomic level, supporting further investigation of probiotic-derived EVs as prospective cell-free functional ingredients.

microbiology↗

bioRxiv2026

Metagenomic Insights into the Fecal Microbiome, Resistome, and Mobilome of Beef Cattle in Cow-Calf Operations in Florida, United States

Food-producing animals are a focus of One Health approaches to antimicrobial resistance (AMR) as reservoirs of antimicrobial-resistant bacteria (ARB) and antimicrobial resistance genes (ARGs). We hypothesized that microbial composition, ARGs profiles, and mobile genetic element (MGE) profiles would differ by operation size and animal age group. Fecal samples from 162 cows and calves were combined into 54 composite samples, each comprising three animals of the same age group. Across operation sizes, Bacillota (51.4-64.9%), Bacteroidota (12.0-17.3%), Pseudomonadota (10.3-16.5%), and Actinomycetota (3.3-10.9%) were predominant phyla. Alpha diversity did not differ among operation sizes; however, cows exhibited greater observed richness, Shannon diversity, and Simpson diversity than calves (all P < 0.001). Microbial composition did not differ significantly among operation sizes (PERMANOVA, R{superscript 2} = 0.065, P = 0.609) but differed significantly between cows and calves (R{superscript 2} = 0.200, P < 0.001). Seventy-five ARGs representing 14 resistance classes were identified. Normalized ARG hit rates were higher in small operations (0.444) than in medium (0.278) and large operations (0.222; P = 0.088), whereas calves exhibited a higher estimated marginal mean normalized ARG hit rate than cows (0.383 vs. 0.247 hits/million reads; P < 0.001). Of 1,054 curated ARG hits, 23.9% occurred on predicted plasmid contigs, 0.9% on predicted viral contigs, and 7.1% on contigs containing insertion sequences. Overall findings demonstrate differences in fecal microbiome diversity and ARG abundance between cows and calves and indicate that predicted plasmids are associated with a notable proportion of ARG hits in Florida beef cow-calf operations.

microbiology↗