Automated classification of cellular expression in multiplexed imaging data with Nimbus.

Multiplexed imaging offers a powerful approach to characterize the spatial topography of tissues in both health and disease. To analyze such data, the specific combination of markers that are present in each cell must be enumerated to enable accurate phenotyping, a process that often relies on unsupervised clustering. We constructed the Pan-Multiplex (Pan-M) dataset containing 197 million distinct annotations of marker expression across 15 different cell types. We used Pan-M to create Nimbus, a deep learning model to predict marker positivity from multiplexed image data. Nimbus is a pre-trained model that uses the underlying images to classify marker expression across distinct cell types, from different tissues, acquired using different microscope platforms, without requiring any retraining. We demonstrate that Nimbus predictions capture the underlying staining patterns of the full diversity of markers present in Pan-M. We then show how Nimbus predictions can be integrated with downstream clustering algorithms to robustly identify cell subtypes in image data. We have open-sourced Nimbus and Pan-M to enable community use at https://github.com/angelolab/Nimbus-Inference.

Dimorphism in the T-cell receptor constant region affects T-cell function, phenotype and HIV outcome.

OBJECTIVES: CD8 T cells recognize human leukocyte antigen-peptide complex through the T-cell receptor. Although amino acid variation in T-cell receptor variable chains often affects antigen specificity, dimorphism in the beta chain constant region (TRBC1 and TRBC2) is not thought to affect T-cell function. A recent study suggested that adoptive transfer of TRBC1-specific chimeric antigen-receptor-T cells provided an option for T-cell leukemia therapy that preserved T-cell immunity in the TRBC2 subset. This raises an important question as to whether TRBC1T cells are qualitatively different from TRBC2T cells. DESIGN: Cross-sectional study. METHODS: Sixty-six antiretroviral therapy-naive HIV-infected individuals, including 19 viraemic controllers and 47 noncontrollers, were enrolled. Peripheral blood mononuclear cells were isolated for T-cell functional assays, tetramer analyses, TRBC1 staining and immunophenotyping. RESULTS: Viraemic controllers had a higher proportion of circulating TRBC1T cells than noncontrollers, raising the possibility that TRBC1T cells might be associated with HIV control. TRBC1T cells also showed more functional T-cell responses against both HIV and cytomegalovirus (P < 0.01). The immunophenotypes of TRBC1-bearing T cells were skewed towards naive and central memory phenotypes, whereas the majority of TRBC2-expressing T cells were terminally differentiated. Inverse correlations were observed between %TRBC1T cells and HIV plasma viral load, which was most pronounced for CD8 T cells (r = -0.7096, P = 0.00002357). CONCLUSION: These data suggest that TRBC1T-cell responses are of better quality than their TRBC2 counterparts, which should be considered in immunotherapeutic strategies for HIV infection. Conversely, depletion of TRBC1T cells as part of the treatment of TRBC1 T-cell malignancies may lead to compromised T-cell response quality.