Defining human skin immunity to cutaneous leishmaniasis via systems immunology approaches

National Institute Of Allergy And Infectious Diseases

Investigators

Abstract

In FY22, we established a collaboration with Dr. Elise OConnell (LPD) and the NIH Leishmaniasis Clinic to develop a preliminary single cell transcriptomic atlas of human CL. Over the past year, Dr. OConnell and colleagues obtained skin punch biopsies from the lesions of four CL patients on which we performed droplet-based single cell RNA-seq (scRNA-seq) using an adaptation of the protocol used for the Human Skin Cell Atlas. The causative species in one patient was L. aethiopica, whereas the other three patients were infected with L. braziliensis, of which two were relapsed cases and one was a new infection. The scRNA-seq analyses have been completed for the L. aethiopica case and the new case of L. braziliensis, composing a total of 140,000 sequenced cells which were analyzed via Seurat. Cell subsets were identified by initial mapping to the Human Skin Cell Atlas using SCINA, a machine learning algorithm, followed by manual curation. The largest subsets of immune cells in each lesion biopsy were T lymphocytes and macrophages, accompanied by smaller but prominent populations of plasma cells and NK cells. Skin stromal cells were readily identifiable in the dataset, including keratinocytes, fibroblasts, vascular endothelial cells, melanocytes, and pericytes. This preliminary single cell atlas yielded several novel insights into the roles of specific cell subsets that drive pathologic inflammation in CL lesions. The inflammatory cytokine CXCL8 derives largely from lesional macrophages, with lower expression across a variety of both immune and stromal cells. In contrast, IL-1-beta is produced predominantly by macrophages, while IL-6 is expressed mainly in fibroblasts and not immune cells. Similarly, CXCL10, a chemokine which has been validated as a marker of CL lesion severity, is expressed predominantly in a subset of keratinocytes, fibroblasts, and pericytes, but not in immune cells. Finally, CCL4, a chemokine whose expression predicts CL treatment failure in L. braziliensis, was expressed mainly in a subset of NK cells. These analyses demonstrate how the CL single cell atlas allows previously identified clinical markers of CL severity, progression, and treatment failure to now be attributed to specific cell subsets, generating novel mechanistic hypotheses behind how cutaneous immunity to CL leads to pathological tissue destruction versus protection. The development of skin microbiopsy devices allows serial sampling of the same skin site over time with minimal trauma and without requiring local anesthesia, which is not possible with conventional skin punch biopsies. Each microbiopsy collects 50 ug of skin tissue, yielding 1.4 ng of RNA for qPCR or bulk RNA-seq studies. In collaboration with Dr. Tarl Prow (University of South Australia) and Dr. OConnell, we performed microbiopsy sampling of CL lesions and normal skin from four CL patients over the past year. Three of the patients also had conventional punch biopsies collected for scRNA-seq (described above) one L. aethiopica patient, and both relapsed L. braziliensis patients. Additionally, we collected microbiopsies from a pediatric patient infected with L. tropica who was excluded from the punch biopsy cohort due to their age. To account for regional gene expression variability in the skin, three to five microbiopsies were collected per site (one lesion site and one site of normal skin at a similar anatomic location to the lesion). All patients tolerated the procedure well without complications and with normal healing of all the microbiopsy puncture sites.