Revealing differential type I and type II IFN responses in different immune cells in various human diseases by a new IFNs transcriptomic atlas

Interferons (IFNs) are key cytokines mediating broad immunological responses in diverse processes including host defense, autoimmune diseases, and cancer. To gain single cell resolution in response to different IFNs in different human immune cells, we generated a new single-cell transcriptomic dataset in various human immune cell subsets by stimulating with IFN-α, IFN-β, IFN-γ, and IFN-λ1. We not only found classical IFN stimulated genes (ISGs) highly elevated across multiple immune cell subsets, but each immune cell subset also displayed unique gene sets in response to IFN-α or IFN-γ stimulation. Furthermore, expression fold changes of many ISGs shared between IFN-α and IFN-γ were different, which prompted us to utilize those unique features to generate a new scoring method. To obtain a comprehensive view of type I & II IFN responses from many different human diseases, we applied our tool to several published scRNAseq datasets from patients with known elevated IFN responses. We also applied this to a new cohort of newly diagnosed multiple myeloma (NDMM) patients and ran a cross-sectional and a longitudinal analysis to explore IFN responses in NDMM compared to healthy and NDMM after induction therapies, with or without daratumumab. We found higher circulating protein levels of IFN-α2A, IFN-β, and IFN-γ, and higher type I and type II IFN scores across all circulating immune cells in NDMM. Induction therapies only decreased type I IFN responses, especially in CD14 monocytes, and circulating IFN-α. Both known ISGs and novel cell-type specific ISGs from our ex vivo dataset could be recapitulated in in vivo NDMM. Our work provides insights into immune cell-type specific responses to IFNs and a new level of clarity to dissect different IFN responses from human diseases. With our cloud-based explorers, this new tool can be easily used by many others to further expand human IFN research.