Phosphatidylinositol metabolism of the renal proximal tubule S3 segment is disturbed in response to diabetes

Abstract

Diabetes is a main risk factor for kidney disease, causing diabetic nephropathy in close to half of all patients with diabetes. Metabolism has recently been identified to be decisive in cell fate decisions and repair. Here we used mass spectrometry imaging (MSI) to identify tissue specific metabolic dysregulation, in order to better understand early diabetes-induced metabolic changes of renal cell types. In our experimental diabetes mouse model, early glomerular glycocalyx barrier loss and systemic metabolic changes were observed. In addition, MSI targeted at small molecule metabolites and glycero(phospho)lipids exposed distinct changes upon diabetes in downstream nephron segments. Interestingly, the outer stripe of the outer medullar proximal tubular segment (PT_S3) demonstrated the most distinct response compared to other segments. Furthermore, phosphatidylinositol lipid metabolism was altered specifically in PT_S3, with one of the phosphatidylinositol fatty acid tails being exchanged from longer unsaturated fatty acids to shorter, more saturated fatty acids. In acute kidney injury, the PT_S3 segment and its metabolism are already recognized as important factors in kidney repair processes. The current study exposes early diabetes-induced changes in membrane lipid composition in this PT_S3 segment as a hitherto unrecognized culprit in the early renal response to diabetes.

Figure 1

Diabetes in ApoE-KO mice manifests in systemic metabolic alterations without clear signs of DN. Representative images of control and diabetic mouse kidney sections stained with (A) periodic acid-Schiff (PAS) and (B) kidney injury molecule-1 (KIM-1), with insets showing more detailed images of the cortex, outer stripe of outer medulla (OSOM) and inner stripe of outer medulla (ISOM) (scale bar = 200 µm and for inset = 50 µm). Representative images of direct glycocalyx staining using fluorescent labeled (C) lectin Lycopersicon esculentum (LEA-FITC) and anti-CD31 antibodies for endothelial cell detection or (D) neurocan (Ncan-dsRed) and anti-CD31 antibodies; scale bar 50 μm. (E & F) Reduction of the luminal glycocalyx (LEA or Ncan-dsRed) thickness and luminal glycocalyx (LEA or Ncan-dsRed) coverage, assessed in a subset of n = 3 control and diabetic ApoE-KO mice. Systemic metabolic measurements comparing the (G) energy expenditure, (H) respiratory quotient, (I) fat oxidation rate and (J) carbohydrate oxidation rate at night of control and diabetic mice (n = 4/group used for subsequent MSI analysis). HA = hyaluronan.

Figure 2

Metabolic histology of control and diabetes renal tissue. (A) Experimental workflow for in situ metabolic histology analysis of control and diabetic mouse. (B) Metabolic heterogeneity in the mouse kidney, visualized in a Uniform Manifold Approximation and Projection (UMAP) plot of MALDI-MSI data (n = 4/group). (C) Metabolic histology of control and diabetic renal tissue (example), identifying ten different renal cell types. (D) Visualization of the cluster centroids of the metabolic histological identified renal cell types, showing similarity between the control and diabetic renal clusters. (E) Relative contribution of the various renal cell types to the total pixel population in the control and diabetic mouse kidneys.

Figure 3

Changes in metabolic profile of specific renal cell types. (A) Schematic of a nephron, with the regions of interest for further data interrogation of metabolic characterization highlighted. (B) Venn diagram displaying the number of metabolites that are changed with a fold change larger or smaller than 0.5 or -0.5 respectively in the nephron segments. (C) Volcano plots of the nephron regions of interest, illustrating the changed metabolites in the glomeruli, PT_S1/S2, PT_S3 and DT of the diabetic mice compared to control.

Figure 4

Lipid alterations in specifically PT_S3 cells upon diabetes. (A) Ion marker distribution of m/z 762.5 and m/z 909.5 clearly show the distinction between PT_S1/S2 and PT_S3, (B) confirmed by LTL⁺ staining and morphology. (C) Changes in phosphatidylinositol (PI) lipid species upon diabetes in the four nephron segments of interest: Glomeruli, PT_S1/S2, PT_S3 and DT. (D) In situ distribution and quantification of renal specific PS feature and PT_S3 specific PI features that are differentiating between diabetes and control. When tandem MS spectra are available, fatty acid tail composition is indicated and schematic drawings of the lipids are displayed (scale bar = 500 µm and for inset = 200 µm, *indicates author’s interpretation since tandem MS spectra were not available).