The high failure rate of therapeutics showing promise in mouse models to translate to patients is a pressing challenge in biomedical science. Though retrospective studies have examined the fidelity of mouse models to their respective human conditions, approaches for prospective translation of insights from mouse models to patients remain relatively unexplored. Here, we develop a semi-supervised learning approach for inference of disease-associated human differentially expressed genes and pathways from mouse model experiments. We examined 36 transcriptomic case studies where comparable phenotypes were available for mouse and human inflammatory diseases and assessed multiple computational approaches for inferring human biology from mouse datasets. We found that semi-supervised training of a neural network identified significantly more true human biological associations than interpreting mouse experiments directly. Evaluating the experimental design of mouse experiments where our model was most successful revealed principles of experimental design that may improve translational performance. Our study shows that when prospectively evaluating biological associations in mouse studies, semi-supervised learning approaches, combining mouse and human data for biological inference, provide the most accurate assessment of human in vivo disease processes. Finally, we proffer a delineation of four categories of model system-to-human “Translation Problems” defined by the resolution and coverage of the datasets available for molecular insight translation and suggest that the task of translating insights from model systems to human disease contexts may be better accomplished by a combination of translation-minded experimental design and computational approaches. Empirical comparison of genomic responses in mouse models and human disease contexts is not sufficient for addressing the challenge of prospective translation from mouse models to human disease contexts. We address this challenge by developing a semi-supervised machine learning approach that combines supervised modeling of mouse datasets with unsupervised modeling of human disease-context datasets to predict human in vivo differentially expressed genes and enriched pathways. Semi-supervised training of a feed forward neural network was the most efficacious model for translating experimentally derived mouse biological associations to the human in vivo disease context. We find that computational generalization of signaling insights substantially improves upon direct generalization of mouse experimental insights and argue that such approaches can facilitate more clinically impactful translation of insights from preclinical studies in model systems to patients.
Gamma delta T cells play a crucial role in anti-tumor immunity due to their cytotoxic properties. However, the role and extent of γδ T cells in production of pro-tumorigenic interleukin-17 (IL-17) within the tumor microenvironment of colorectal cancer (CRC) remains controversial. In this study, we re-analyzed nine published human CRC whole-tissue single-cell RNA sequencing datasets, identifying 18,483 γδ T cells out of 951,785 total cells, in the neoplastic or adjacent normal tissue of 165 human CRC patients. Our results confirm that tumor-infiltrating γδ T cells exhibit high cytotoxicity-related transcription in both tumor and adjacent normal tissues, but critically, none of the γδ T cell clusters showed IL-17 production potential. We also identified various γδ T cell subsets, including poised effector-like T cells, tissue-resident memory T cells, progenitor exhausted-like T cells, and exhausted T cells, and noted an increased expression of cytotoxic molecules in tumor-infiltrating γδ T cells compared to their normal area counterparts. We proposed anti-tumor γδ T effector cells may arise from tissue-resident progenitor cells based on the trajectory analysis. Our work demonstrates that γδ T cells in CRC primarily function as cytotoxic effector cells rather than IL-17 producers, mitigating the concerns about their potential pro-tumorigenic roles in CRC, highlighting the importance of accurately characterizing these cells for cancer immunotherapy research and the unneglectable cross-species discrepancy between the mouse and human immune system in the study of cancer immunology.
The co-culture of two adult human colorectal cancer cell lines, Caco-2 and HT29, on Transwell is commonly used as an in vitro gut mimic, yet the translatability of insights from such a system to adult human physiological contexts is not fully characterized. Here, we used single-cell RNA sequencing on the co-culture to obtain a detailed survey of cell type heterogeneity in the system and conducted a holistic comparison with human physiology. We identified the intestinal stem cell-, transit amplifying-, enterocyte-, goblet cell-, and enteroendocrine-like cells in the system. In general, the co-culture was fetal intestine-like, with less variety of gene expression compared to the adult human gut. Transporters for major types of nutrients were found in the majority of the enterocytes-like cells in the system. TLR 4 was not expressed in the sample, indicating that the co-culture model is incapable of mimicking the innate immune aspect of the human epithelium.
Type 2 diabetes (T2D) is implicated as a risk factor for Alzheimer’s disease (AD), the most common form of dementia. In this work, we investigated neuroinflammatory responses of primary neurons to potentially circulating, blood–brain barrier (BBB) permeable metabolites associated with AD, T2D, or both. We identified nine metabolites associated with protective or detrimental properties of AD and T2D in literature (lauric acid, asparagine, fructose, arachidonic acid, aminoadipic acid, sorbitol, retinol, tryptophan, niacinamide) and stimulated primary mouse neuron cultures with each metabolite before quantifying cytokine secretion via Luminex. We employed unsupervised clustering, inferential statistics, and partial least squares discriminant analysis to identify relationships between cytokine concentration and disease-associations of metabolites. We identified MCP-1, a cytokine associated with monocyte recruitment, as differentially abundant between neurons stimulated by metabolites associated with protective and detrimental properties of AD and T2D. We also identified IL-9, a cytokine that promotes mast cell growth, to be differentially associated with T2D. Indeed, cytokines, such as MCP-1 and IL-9, released from neurons in response to BBB-permeable metabolites associated with T2D may contribute to AD development by downstream effects of neuroinflammation.