SPATIAL IMMUNE CELL PROFILING DETECTS DISEASE-SPECIFIC IMMUNE MICROENVIRONMENT IN IGG4-RELATED KIDNEY DISEASE

IgG4-related kidney disease (IgG4-RKD) is a fibroinflammatory condition characterized by lymphoplasmacytic tubulointerstitial nephritis. The spatial organization within the immune microenvironment of IgG4-RKD remains poorly defined.

We applied high-dimensional confocal microscopy using co-detection by indexing with a 54-marker antibody panel to kidney biopsy specimens from patients with IgG4-RKD (n = 7), ANCA-associated vasculitis (AAV, n = 9), and control kidneys (n = 4). Cells were segmented with a fine-tuned Cellpose 2.0 pipeline, and cell classification was performed by pseudo-spectral angle mapping (pSAM) with 59 membrane and 16 nucleus classes. pSAM cell prediction performance was also assessed using 2,947 ground-truth-annotated cells from IgG4-RKD. Structural segmentation based on a fine-tuned U-Net and Omnipose was also applied to determine the location of each cell in the kidney structure. Immune cell proportions and densities were compared among groups, followed by correlation and neighborhood analyses to identify disease-specific immune microenvironments.

A total of 889,919 cells were segmented from whole-kidney tissues across all samples. pSAM accurately classified immune cell types with weighted-average F1 scores of 0.88. Structural segmentation confirmed that almost all of the immune cells in IgG4-RKD were in the interstitial compartment, while up to 20% of those in AAV were in the tubular compartment. Thus, we focused on 369,949 cells in interstitial areas for downstream analyses. Myeloid cells were the most frequent immune population in both diseases, but CD8⁺T cells were significantly enriched in IgG4-RKD. Distinct immune subsets in IgG4-RKD included HLA-II⁺granzyme K⁺ICOS+PD-1+CD4+T follicular helper (Tfh) cells, HLA-II⁺CD27⁺CD4⁺Foxp3+regulatory T cells (Tregs), CD8⁺Foxp3+Tregs, CD11c⁺B cells, and MerTK⁺CD68+M1-like macrophages. In contrast, AAV was characterized by TCF1⁺TOX⁻ stem-like CD4⁺T/CD8⁺T cells, CD68+CD163+M2-like macrophages, and CD16-CD56+immature natural killer cells. Spearman’s correlation analysis demonstrated broader interlineage associations among CD4⁺T, CD8⁺T, humoral, myeloid, and myofibroblasts in IgG4-RKD, whereas correlations in AAV were largely confined to CD4⁺ and CD8⁺ T cell compartments. Density-based spatial clustering of applications with noise, followed by Louvain community detection, identified 20 unique immune/myofibroblast niches. IgG4-RKD–enriched niches included: (1) CD69+CD4⁺/CD8⁺ tissue-resident memory T cells with ICOS+PD-1+CD4+Tfh cells and CD11c-CD27-CD20+T-bet+double-negative type 3 B cells; (2) granzyme B⁺CD8⁺cytotoxic T lymphocytes with CD8⁺Foxp3+Tregs and γδT cells; and (3) glycoprotein-NMB (GPNMB)⁺CD68⁺M1-like and GPNMB+CD68+CD163+M2-like macrophages. AAV-enriched niches included: (1) CD4⁺Foxp3+Tregs, Th1 cells, and TIGIT⁺CD8⁺T cells; (2) CD68+CD163+M2-like macrophages, CD90⁺αSMA+myofibroblasts, and CD68⁺CD163⁺αSMA+macrophage–mesenchymal transition cells; and (3) CD11b+CD16+myeloperoxidase (MPO)+neutrophils and MPO⁺CD68⁺macrophages.

Spatial immune profiling reveals distinct immune microenvironments in IgG4-RKD, characterized by unique CD4⁺/CD8⁺T cell, B cell, and macrophage subsets, along with broader intercellular correlations. These findings provide novel insights into disease pathogenesis and may inform future therapeutic strategies.