Microvascular immunity is organ-specific and remodeled after kidney injury in mice

Abstract

Many studies analyze tissue-resident or blood-borne leukocytes to monitor disease progression. We hypothesized that the microvasculature serves as a distinct site for immune cell activity. Here, we investigate microvascular leukocyte phenotypes before, during and after acute kidney injury (AKI) in mice, uncovering unique characteristics in the kidney, liver, and lung. Using single-cell sequencing, we identify several immune cells that were up to 100-fold expanded in the kidney vasculature, including macrophages, dendritic cells (DC), and B cells. Regeneration after AKI is characterized by sustained remodeling of the renal microvascular interface. Homeostatic microvascular C1q⁺ macrophages withdraw from the vascular barrier which is subsequently repopulated by new subsets, including CD11c⁺F480⁺ and CD11c⁺F480^- cells. These newly arrived macrophages exhibit enhanced phagocytic activity toward circulating bacteria and secretion of tumor necrosis factor, pointing to maladaptive repair mechanisms after AKI. These data suggest organ- and disease-specific microvascular immune dynamics which are not detectable through conventional blood and tissue analysis.

Fig. 1. Microvascular leukocytes are abundant, organ-specific and differ from peripheral blood.

a Experimental setup to detect microvascular leukocytes in mice, and compare them to tissue-resident and blood counterparts. Icons from the Servier Medical Art collection under CC BY 4.0. b The relation of microvascular and tissue-resident leukocytes in different healthy mouse organs. Ratios are shown in pie charts and absolute numbers as bar diagrams. Paired two sided t test. n = 7 independent experiments. Mean and SD are shown. c Visualization of microvascular leukocytes in healthy organs after intravenous injection of an anti-CD45 antibody. Scale bar 100 μm. d The combination of in vivo and ex vivo anti-CD45 staining highlights vascular and tissue-resident leukocytes. Intraluminal cell extensions (double positive) of perivascular cells (green) are predominantly found in the kidney. Scale bar 20 μm. Images are representative of 3 independent experiments. e Flow cytometry analysis of different intravascular leukocyte subsets in comparison to peripheral blood (red bar). Classical (CM) and nonclassical monocytes (NCM) are side-scatter (SSC) high and low, respectively. CD11b⁺ Ly6G⁻ SSC^hi cells were further analyzed using F480 and CD11c. Paired two sided t test of blood versus each organ. Mean and SD are shown. Each dot represents one independent experiment. Source data are provided as a Source Data file.

Fig. 2. Unsupervised phenotyping of microvascular leukocytes in the healthy mouse kidney.

a, b Single cell sequencing of leukocytes sorted from blood, the renal microvasculature (MV) and the renal tissue. UMAP visualization and annotation of 27 leukocyte subtypes. n = 4 independent experiments. c Abundance of cell cluster specific leukocytes in the three compartments. Blue is the lowest number and red the highest number for each column. The top bar graphs represent the total cell count of each cluster as mean ±+/− SD of n = 4 independent experiments. d Ranked fold change of microvascular versus peripheral blood (top), and microvascular versus tissue-resident (bottom) leukocytes for each cell cluster. e Selected cell clusters are shown in the three compartments as fraction of all detected leukocytes. Unpaired two sided t test between MV and blood, and MV and tissue. All clusters are shown in supplementary Fig. 4. Mean and SD are shown. Each dot represents one independent experiment. Source data are provided as a Source Data file.

Fig. 3. A transcriptional shift of microvascular B cells in glomerular capillaries.

a Reclustering of the B cell clusters shown in Fig. 2. Selected differentially expressed (DE) genes are annotated. b Demultiplexing of all B cells into their origins (peripheral blood, tissue or microvasculature, MV). Of all B cells in the kidney, 86% are found in the microcirculation. c Ranked fold change of microvascular versus peripheral blood B cells. d Absolute B cell numbers in all clusters with significant changes. Paired two sided t test. Mean and SD are shown. Each dot represents one independent experiment. Source data are provided as a Source Data file. e Slingshot analysis shows transcriptional phenotype trajectories from the main B cell cluster of the peripheral blood (cluster 0, significantly enriched in blood) to microvascular B cells of the kidney (clusters 7,4, and 6): unique upregulation of Irf4 and Egr1 in the microvascular B cell trajectories A and B, respectively. f Functional annotation of DE genes in peripheral blood B cells (cluster 0, 5; significantly enriched in blood) versus microvascular B cells of the kidney (clusters 1,4,6,7; significantly enriched in the microcirculation). g Intravascular anti-CD19 B cell staining in the healthy kidney identifies the glomerular capillaries as the main B cell residence. Scale bar 50μm. Each dot represents one microscopy image, n = 9 (tubulus) and n = 44 glomerulus, from n = 3 biological replicates. Mean and SD are shown.

Fig. 4. Disease-specific alterations of the microvascular immune landscape across different organs.

a Overview of the disease conditions. INF = inflammation by semi-sterile peritonitis. AKI = acute kidney injury. INF-reg/AKI-reg = full regeneration 12 days after surgery. UT = untreated. b Microvascular leukocyte subtypes in AKI and AKI-reg. Mean and SD are shown. n = 7 (UT), 3 (INF, INF-reg), 7 (AKI), 4 (AKI-reg). c The total number of CD45+ leukocytes in selected organs expressed as relative to untreated (UT, equals 1). One way Anova with Dunett’s multiple comparison showing significant changes compared to UT in each organ. Mean and SD are shown. d Microvascular neutrophils, monocytes and CD11b⁺ Ly6g⁻ SSC^hi cells in different organs and the peripheral blood expressed as relative to blood untreated (equals 1). Unpaired two sided t test showing significant changes comparing the organ microvasculature to the blood. n = 3–7, details are provided in the source data sheet. Mean and SD are shown. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Fig. 5. Long-term remodeling of the microvascular leukocyte landscape after recovery of kidney injury.

a Experimental setup to integrate AKI-reg (n = 4) and untreated (UT, n = 4) single cell sequencing data of the kidney. MV = microvascular. b Leukocyte cell clustering, UMAP visualization, subtype annotation and demultiplexing. c Heatmap visualizing relative abundance of each cell cluster. Blue is the lowest number and red the highest number for each column. The top bar graphs represent the total cell count of each cluster as the mean ± SD of n = 8 independent experiments. d Ranked fold change of each cell cluster in the blood (top), microvascular (center) and tissue (bottom) compartment. Negative and positive numbers show enrichment in the UT and AKI-reg condition, respectively. Cell cluster fractions are shown in supplementary Fig. 7. e Up- and downregulated cell clusters in AKI-reg compared to UT. Only cluster numbers with significant changes (unpaired two sided t test) are shown (gray boxes). Overlapping circles indicate that the same change can be detected in both (or all three) compartments. Source data are provided as a Source Data file.

Fig. 6. Microvascular macrophage and DC dynamics after recovery of kidney injury.

a Reclustering and UMAP visualization of renal macrophages and dendritic cells (DC) in the data set shown in Fig. 5 (untreated and AKI-reg conditions). Differentially expressed genes are annotated. b Heatmap of selected genes with hierarchical clustering and relative expression. c Demultiplexing into microvascular (left) and tissue leukocytes (right). Cells detected in the untreated (UT) and AKI-reg conditions are green and red, respectively. d The ratios of microvascular (MV) and tissue-resident macrophages/DC are shown. A cluster subanalysis of the MV cell fraction is shown on the right. e Fold change analysis of each cell cluster. Positive numbers indicate enrichment in AKI-reg. f Detailed analysis of clusters 0 and 2 as the most up- and downregulated microvascular clusters. Unpaired two sided t test. Mean and SD are shown. The number of dots represents independent biological replicates (n = 4). g Analysis of intravascular CD11c⁺ cells in UT and AKI-reg kidneys using immunofluorescence microscopy. The total (intra- and extravascular) number and microvascular cells were counted in 3 cortical regions of interests (ROI) from 3 independent biological replicates. Unpaired two sided t test. Mean and SD are shown. Scale bar 50μm. h Flow cytometry analysis of CD11b⁺ Ly6G⁻ CD11c⁺ F480^+/- leukocytes in the renal microvasculature (MV), tissue and peripheral blood. Unpaired two sided t test. Mean and SD are shown. The number of dots represents independent biological replicates (n = 3). Source data are provided as a Source Data file.

Fig. 7. Phagocytosis, M2 polarization and TNF secretion in microvascular macrophages and DC of the kidney.

a After iv-injection of pH-sensitive E.coli phagocytosis can be detected in renal intravascular CD11c⁺ cells within 5 min (arrows). Scale bar 10μm. b, c Flow cytometry analysis shows microvascular phagocytosis in healthy kidneys that increases significantly after recovery from kidney injury (AKI-reg). Unpaired two sided t test. Gating on microvascular CD45.2⁺CD11c⁺ leukocytes. Mean and SD are shown. d CD206 was analyzed in the microvascular (left bars) and tissue-resident (right bars) CD11c⁺ F480⁺ and CD11c⁺ F480⁻ leukocyte populations by flow cytometry. Unpaired two sided t test. Mean and SD are shown. e TNF secretion was determined by intracellular flow cytometry in microvascular and tissue resident CD11b⁺ CD11c⁺F480⁺ and CD11c⁺F480⁻ cell populations with and without ex vivo LPS stimulation. Mean and SD are shown. Paired two sided t test. The dots in all bar graphs represent independent biological replicates. Source data are provided as a Source Data file. f Main findings summarized as graphical abstract.