Background: Calcineurin inhibitors (CNIs) Cyclosporine A (CsA) and Tacrolimus (TAC) are key immunosuppressive agents in organ transplantation but limited by their intrinsic nephrotoxicity. Understanding the mechanism of nephrotoxicity is critical to developing strategies to maximize the success rate of organ transplantation. We explored the potential inflammatory response following CNI use to further elucidate the potential role of immune-epithelial niches in CNI nephrotoxicity.
Methods: Eight-week-old C57Bl/6 male mice were fed a low-salt diet and treated with 90 or 120 mg/Kg of CsA daily for two weeks. Kidneys were harvested and frozen in OCT media for spatial transcriptomics (ST). Here, high quality tissue sections (RIN 6-8) were processed for use on the 10x Visium platform including imaging and library generation followed by next-generation sequencing.  FASTQ and TIFF file processing were performed with CellRanger.  Expression matrices and images were imported into Seurat (v4.0) for analysis. Cell-type distribution within Visium spots were determined with SPOTLight (v0.1.7) using a scRNASeq reference library generated from mouse kidneys. Kidney proximal tubular epithelial cells (PTECs) were treated in vitro, with 30 μM TAC for 24hr and analyzed by LISA, Seahorse XF Bioanalyzer, Western Blot and mRNA transcripts of inflammatory markers.
Results: Cortical and medullary markers of the murine kidney were identified by ST and correlated with histologically defined regions of the kidney. The injury markers, Havcr1 and Lcn2, were significantly increased with CNI treatment. Using reference cell types from scRNASeq of mouse kidney, Visium spots were deconvolved, identifying epithelial cells that localized to expected regions of the kidney. Immune cells mapped to Visium spots and associated with specific tubular subsegments (e.g. cortical vs. medullary-thick ascending limb) in a CNI dose-dependent manner (Figure).  Furthermore, analysis using ReactomePA and GOterm databases showed differential upregulation of profibrotic genes and chemo-attractants as well as metabolic pathways with downregulation of respiratory function and kidney development-related genes in the kidneys with CNI treatment compared to control. In vitro, PTECs treated with TAC (versus vehicle) secreted more HMGB1 (1490 pg/mL vs 1272 pg/mL, p=0.008), TNF-α and TGF-β, with IL-6 trending upwards, indicating a possible paracrine effect of CNI-mediated HMGB1 release. Mitochondrial dysfunction was further quantitated and confirmed by Seahorse Analyzer in TAC-treated cells. 
Conclusions: Our findings suggest that ST can uncover molecular profiles and cellular niches of CNI nephrotoxicity. Further, increasing doses of CNI induced a shift in the transcriptional state of the kidney and changes in the cellular niches of toxicity. Altogether, these data suggest a paradoxical impact of CNI use with suppression of alloimmunity with ongoing inflammatory injury within the kidney.