Neural Network-Assisted Humanization of COVID-19 Hamster scRNAseq Data Reveals Matching Severity States in Human Disease

Friedrich, V. D.; Pennitz, P.; Wyler, E.; Adler, J. M.; Postmus, D.; Alves, L. G. T.; Prigann, J.; Pott, F.; Vladimirova, D.; Hoefler, T.; Goekeri, C.; Landthaler, M.; Goffinet, C.; Saliba, A.-E.; Scholz, M.; Witzenrath, M.; Trimpert, J.; Kirsten, H.; Nouailles, G.

doi:10.1101/2024.01.11.574849

bioRxiv · 10.1101/2024.01.11.574849

Neural Network-Assisted Humanization of COVID-19 Hamster scRNAseq Data Reveals Matching Severity States in Human Disease

Friedrich, V. D.·Pennitz, P.·Wyler, E.·Adler, J. M.·Postmus, D.·Alves, L. G. T.·Prigann, J.·Pott, F.·Vladimirova, D.·Hoefler, T.·Goekeri, C.·Landthaler, M.·Goffinet, C.·Saliba, A.-E.·Scholz, M.·Witzenrath, M.·Trimpert, J.·Kirsten, H.·Nouailles, G.

View original source ↗Export metadata ↓

Abstract

Translating findings from animal models to human disease is essential for dissecting disease mechanisms, developing and testing precise therapeutic strategies. The coronavirus disease 2019 (COVID-19) pandemic has highlighted this need, particularly for models showing disease severity-dependent immune responses. Single-cell transcriptomics (scRNAseq) is well poised to reveal similarities and differences between species at the molecular and cellular level with unprecedented resolution. However, computational methods enabling detailed matching are still scarce. Here, we provide a structured scRNAseq-based approach that we applied to scRNAseq from blood leukocytes originating from humans and hamsters affected with moderate or severe COVID-19. Integration of COVID-19 patient data with two hamster models that develop moderate (Syrian hamster, Mesocricetus auratus) or severe (Roborovski hamster, Phodopus roborovskii) disease revealed that most cellular states are shared across species. A neural network-based analysis using variational autoencoders quantified the overall transcriptomic similarity across species and severity levels, showing highest similarity between neutrophils of Roborovski hamsters and severe COVID-19 patients, while Syrian hamsters better matched patients with moderate disease, particularly in classical monocytes. We further used transcriptome-wide differential expression analysis to identify which disease stages and cell types display strongest transcriptional changes. Consistently, hamsters response to COVID-19 was most similar to humans in monocytes and neutrophils. Disease-linked pathways found in all species specifically related to interferon response or inhibition of viral replication. Analysis of candidate genes and signatures supported the results. Our structured neural network-supported workflow could be applied to other diseases, allowing better identification of suitable animal models with similar pathomechanisms across species. Key PointsO_LINeural networks can successfully match disease states between animal models and humans using single-cell data as shown for COVID-19 C_LIO_LIModerately diseased patients best matched Syrian hamster cells; severely diseased patients best matched Roborovski hamster neutrophils C_LI

Explore related subjects

immunology Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Friedrich, V. D., Pennitz, P., Wyler, E., Adler, J. M., Postmus, D., Alves, L. G. T., Prigann, J., Pott, F., Vladimirova, D., Hoefler, T., Goekeri, C., Landthaler, M., Goffinet, C., Saliba, A.-E., Scholz, M., Witzenrath, M., Trimpert, J., Kirsten, H., Nouailles, G.. 2024-01-12. Neural Network-Assisted Humanization of COVID-19 Hamster scRNAseq Data Reveals Matching Severity States in Human Disease. https://doi.org/10.1101/2024.01.11.574849

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

KEEP EXPLORING

Related preprints

bioRxiv2026

Systemic TLR7 agonism moderates disease severity in acute DSS-induced colitis in mice

Toll-like receptor 7 (TLR7) is an endosomal sensor of single-stranded RNA, studied for its antiviral and pro-inflammatory functions. Its role in intestinal inflammation is undefined, and no approved or clinical-stage therapy for inflammatory bowel disease (IBD) engages this pathway. Here we report that systemic administration of the reference TLR7 agonist GS-9620 (vesatolimod) moderates disease severity in acute dextran sodium sulfate (DSS)-induced colitis in mice. GS-9620-treated mice lost less weight, had lower Disease Activity Index scores, and were protected from DSS-induced colon shortening. Immunophenotyping of the spleen showed reduced Th17 and Tr17 populations and an increased proportion of regulatory T cells. These findings indicate that pharmacological TLR7 agonism is not merely tolerated in the inflamed intestine but is associated with a shift of the systemic immune compartment toward a regulatory phenotype. The results identify TLR7 as a candidate disease-moderating target in IBD and motivate direct tests of causality, route and chronic-disease relevance.

immunology↗

bioRxiv2026

ATF3 Preserves Macrophage Cholesterol Homeostasis during Metabolic Stress

Background: Macrophages must coordinate immune activation with maintenance of cellular cholesterol homeostasis, yet the transcriptional mechanisms integrating these processes remain incompletely understood. Activating transcription factor 3 (ATF3) suppresses inflammatory and interferon responses and has been implicated in macrophage lipid metabolism, suggesting a role in coordinating immune and sterol-responsive transcriptional programs. Methods: Wild-type and ATF3-deficient bone marrow-derived macrophages were analyzed by bulk RNA sequencing following Toll-like receptor stimulation and under conditions of sterol loading or depletion. Cholesterol content, cytokine production, and sterol-responsive gene expression were quantified. Human metabolic dysfunction-associated steatotic liver disease (MASLD) datasets and Kupffer cell-specific ATF3 knockout mice were used to assess disease relevance in vivo. Results: ATF3 deficiency enhanced inflammatory and interferon-responsive transcriptional programs while suppressing a sterol-responsive gene module enriched for SREBP2 target genes. This phenotype was conserved across diverse Toll-like receptor stimuli and accompanied by persistent Ch25h upregulation. Despite exhibiting elevated cellular cholesterol, ATF3-deficient macrophages retained transcriptional responsiveness to sterol loading and depletion while demonstrating reduced induction of SREBP2-dependent genes across conditions. IFNAR blockade reduced interferon-responsive gene expression but did not normalize Ch25h or sterol-responsive gene expression. In contrast, CRISPR-mediated reduction of Ch25h partially restored sterol-responsive gene expression in inflammatory ATF3-deficient macrophages. In human MASLD, hepatic ATF3 expression positively correlated with disease severity. In mice, ATF3 was preferentially induced in Kupffer cells during metabolic stress, and Kupffer cell-specific deletion of ATF3 increased liver injury and Kupffer cell lipid accumulation. Conclusions: ATF3 supports sterol-responsive transcriptional programs while restraining inflammatory activation in macrophages. ATF3 loss results in elevated cellular cholesterol, impaired SREBP2-dependent gene expression, and persistent derepression of Ch25h, which contributes to but does not fully explain the sterol phenotype. These findings identify ATF3 as a coordinator of macrophage adaptation to inflammatory and metabolic stress and support a protective role for Kupffer cell ATF3 during metabolic liver disease.

immunology↗

bioRxiv2026

Circulating CD4 regulatory T cells are broadly preserved in Parkinson's disease despite altered HELIOS expression

Background Regulatory T cells (Tregs) are an important component of adaptive immune regulation and have been implicated in Parkinson's disease (PD), with previous studies suggesting altered Treg proportions, phenotype and, in some cases, function. However, existing studies have been limited by variable phenotyping approaches and a lack of detailed functional assessment. Given the growing interest in Tregs as potential therapies in PD, better understanding of their phenotype and function in the disease is important. Methods Circulating FOXP3+ Tregs were characterised in individuals with PD and age-matched healthy controls (HC) using a 24-colour spectral flow cytometry panel. Treg suppressive function was assessed by their ability to inhibit proliferation and cytokine production by polyclonally stimulated autologous T effector cells in vitro. Treg lineage stability was assessed by analysis of FOXP3 TSDR demethylation. Results The proportions and phenotypic characteristics of circulating CD4 Tregs were broadly similar between individuals with PD and HCs, as were Treg suppressive function and FOXP3 TSDR demethylation. However, naive CD4 Tregs from individuals with PD showed reduced HELIOS expression, with a similar but non-significant trend in FOXP3 expression. Reduced HELIOS expression was also observed in the CD8 naive Treg compartment. Conclusions Major quantitative or functional abnormalities in circulating Tregs do not appear to be a dominant feature PD. However, reduced HELIOS expression in naive Tregs suggests more subtle phenotypic alterations that warrant further investigation. The broadly preserved suppressive function is reassuring for therapeutic approaches aimed at increasing Treg numbers, although it will be important to establish that expanded Tregs retain a stable phenotype and suppressive function in inflammatory and tissue environments.

immunology↗