Imaging of podocyte foot processes by fluorescence microscopy

Abstract

Visualizing podocyte foot processes requires electron microscopy, a technique that depends on special equipment, requires immunogold for colabeling, and does not take advantage of the growing number of in vivo fluorophores available. To address these limitations, we developed a genetic strategy to allow detailed visualization of single podocytes and their foot processes by conventional fluorescence microscopy. We generated a transgenic mouse line expressing a GFP-Cre-ERT2 fusion protein under the control of the collagen α1(I) promoter with strong podocyte expression. Administration of submaximal tamoxifen allowed genetic labeling of single podocytes when crossed with a Cre-reporter line. Of three different reporter systems that we evaluated for the ability to reveal fine structural details of podocytes, bigenic Coll1α1GCE;Gt(ROSA)26Sor(tm9(CAG-tdTomato)) mice allowed podocyte labeling with a strong and homogeneous reporter signal that was easily observed by epifluorescence. We could easily detect anatomic features of podocytes down to tertiary foot processes, and we were able to visualize and quantitate ultrastructural changes to foot processes after podocyte injury. In summary, using this method of genetic labeling and conventional fluorescence microscopy to visualize podocyte foot processes will complement electron microscopy and facilitate the analysis of podocytes and their precursors in vivo.

Figure 1.

Administration of tamoxifen/4-OHT to bigenic Coll1α1GCE⁺;reporter⁺ mice fate-labels podocytes in a dose-dependent fashion. (A) Experimental approach for conditional labeling of podocytes in vivo. Injection of tamoxifen or its active metabolite 4-OHT into bigenic Coll1α1GCE⁺;R26LacZ⁺, Coll1α1GCE⁺;mT/mG⁺, or Coll1α1GCE⁺;R26Tomato⁺ mice enables temporary translocation of the GFP-Cre-ERT2 fusion protein into the cell nucleus, removing the LoxP-flanked DNA sequence of the reporter allele and leading to permanent expression of LacZ, membrane-targeted EGFP, and tdTomato, respectively. (B–D) Validation of the labeling strategy. Only in adult kidneys of tamoxifen/4-OHT–injected bigenic animals is there robust visceral epithelial expression of LacZ, membrane-targeted EGFP, or tdTomato. Insets (D, right) show a labeled pericyte (red) in the medulla of a Coll1α1GCE⁺;R26Tomato⁺ kidney, and antilaminin staining (green) highlights its tubulointerstitial localization. Scale bars, 50 μm. (E–G) Semiquantitative assessment of recombination efficiency in glomeruli in relation to time point and dose of tamoxifen/4-OHT administration. Animals injected only one time at P0 exhibited no evidence of recombination in glomeruli of the outer cortex, suggesting a lack of collagen1α1 promoter activity in podocyte progenitors of the nephrogenic zone at the time point of injection. Single injection at P3 resulted in a more homogenous but mostly partial labeling of glomeruli throughout the cortex. By contrast, repeated tamoxifen/4-OHT injections (P0 and P3) labeled the majority of podocytes in almost all glomeruli. Quantitation was performed in Coll1α1GCE⁺;mT/mG⁺ and Coll1α1GCE⁺;R26Tomato⁺ mice, respectively. Data are given as mean ± SEM; n=3–4 for each data point.

Figure 2.

Validation of podocyte-specific fate-labeling in glomeruli of bigenic Coll1α1GCE⁺; R26Tomato⁺ mice. (A–C) Immunostaining for podocyte-specific marker nephrin (green) colabels cells that express the fate marker tdTomato (red). By contrast, immunostaining for endothelial (CD31) or mesangial cells (PDGFRβ) shows no overlap with fate-labeled cells. Scale bars, 10 μm.

Figure 3.

Imaging of structural details of sporadically labeled podocytes in uninjured Coll1α1GCE⁺;R26Tomato⁺ kidneys by conventional fluorescence microscopy. (A–C) Whole antilaminin-stained glomerulus (green) highlighting close association of tdTomato+ network of podocyte processes and laminin+ glomerular basement membrane. Several regions of interest are illustrated. (D) A region with interdigitating foot processes from two neighboring podocytes that are both fate-labeled is shown. Of note, resolution is insufficient to delineate details beyond the nucleated cell bodies (*) and main processes (α). (E–G) By contrast, these cutouts highlight areas in which only every other foot process is genetically labeled, and neighboring pedicles remain optically silent. Resolution is adequate to assess morphology and extension of single tertiary foot processes. Note that the lateral branching of tertiary foot processes out from main processes resembles the typical arrangement of fish bones. (H–K) Antilaminin-stained glomerular capillary loop (green) covered by a network of genetically labeled podocyte foot processes (red). (K) The three-dimensional view illustrates how foot processes conform to the ridge-like elevation of the underlying capillary (top to bottom=2.2 μm). (L) Example of a labeled podocyte with classic podocyte architecture: a nucleated cell body (*) and a primary process (α) branching into several secondary processes (β), which again branch into many minor, tertiary foot processes at regular intervals. (M) Example of tertiary foot process branching in very close proximity to the nucleated cell body (*). Scale bars, 10 μm in A–C; 2 μm in D–M.

Figure 4.

Analysis of foot process changes in PS-perfused Coll1α1GCE⁺;R26Tomato⁺ kidneys by fluorescence microscopy. (A and B) Micrographs of whole glomerulus show the normal overall arrangement of labeled podocytes after PS perfusion and show no noticeable difference of main cell bodies compared with uninjured control. (C–K) Closer inspection, however, reveals ultrastructural abnormalities of processes, comprising the whole spectrum of morphologic changes associated with foot process effacement, including clubbing (arrowheads), lateral spreading and flattening (*), and retraction and shortening. Scale bars, 10 μm in A and B; 2 μm in C–K. (L–P) Comparative analysis of morphometric features of labeled podocytes from PS-perfused and uninjured control kidneys. Although main cell body areas are comparable between groups (52.2±1.1 versus 53.0±1.1 μm²), the data show a significant increase in width of primary (0.93±0.02 versus 1.10±0.04 μm), secondary (0.50±0.01 versus 0.67±0.02 μm), and tertiary foot processes (0.29±0.001 versus 0.32±0.002 μm) as well as a reduction of tertiary foot process length (1.79±0.02 versus 1.64±0.01 μm) in the PS-perfused group. Dot plots show individual measurements, and box plots represent the 25th (bottom) and 75th (top) percentiles. The bands within the plots indicate the medians, and the solid squares indicate the means. The whiskers represent minimum and maximum values; n=4 for each group. *P<0.0001; ns, not significant.