Injury-induced actin cytoskeleton reorganization in podocytes revealed by super-resolution microscopy

Abstract

The architectural integrity of tissues requires complex interactions, both between cells and between cells and the extracellular matrix. Fundamental to cell and tissue homeostasis are the specific mechanical forces conveyed by the actomyosin cytoskeleton. Here we used super-resolution imaging methods to visualize the actin cytoskeleton in the kidney glomerulus, an organized collection of capillaries that filters the blood to make the primary urine. Our analysis of both mouse and human glomeruli reveals a network of myosin IIA-containing contractile actin cables within podocyte cell bodies and major processes at the outer aspects of the glomerular tuft. These likely exert force on an underlying network of myosin IIA-negative, noncontractile actin fibers present within podocyte foot processes that function to both anchor the cells to the glomerular basement membrane and stabilize the slit diaphragm against the pressure of fluid flow. After injuries that disrupt the kidney filtration barrier and cause foot process effacement, the podocyte's contractile actomyosin network relocates to the basolateral surface of the cell, manifesting as sarcomere-like structures juxtaposed to the basement membrane. Our findings suggest a new model of the podocyte actin cytoskeleton in health and disease and suggest the existence of novel mechanisms that regulate podocyte architecture.

Keywords: Cell Biology; Cytoskeleton; Nephrology.

Figure 1. STORM imaging of slit diaphragm proteins.

(A–C) Double-color STORM imaging of the glomerular basement membrane (GBM) protein agrin (blue) and slit diaphragm proteins nephrin (A, red), podocin (B, red), and CD2AP (C, red) showing how these proteins cluster near the subpodocyte aspect of the agrin-stained GBM. (D) Triple-color STORM imaging of agrin (blue), nephrin (red), and synaptopodin (green) shows that synaptopodin clusters are located between the individual nephrin clusters. (E) Triple-color STORM imaging of agrin (blue), synaptopodin (green), and α-actinin-4 (magenta) shows that synaptopodin and α-actinin-4 clusters colocalize. Scale bars: 200 nm.

Figure 2. STORM imaging of myosin IIA’s distribution in podocytes in healthy and diseased conditions.

(A) Triple-color STORM overview image of a WT glomerulus shows myosin IIA’s (blue) absence from foot processes in comparison with synaptopodin (red) and integrin β1 (green) labeling along the capillary wall. (B and C) Higher magnification images of the boxed area in A show that myosin IIA is present in the podocyte cell body and the primary processes (arrows in B) but not in the foot processes that contain synaptopodin (arrowheads in C). Note the transition between the foot processes and the primary processes (arrows in C). Synaptopodin patches were observed at the periphery of the podocyte cell body and primary processes (arrowheads in B). (D–I) Staining for actin (D and G), myosin IIA (E and H), and synaptopodin (F and I). (D–F) Confocal slices of the glomerular surface show the podocytes’ thick actin cables positive for myosin IIA and negative for synaptopodin (arrows). Note the diffracted en face image of the capillary wall stained positive for actin and synaptopodin and very little myosin IIA (arrowheads). The differential interference contrast (DIC) overlay image (F) shows the shadow of magnetic beads (4 μm) used to isolate glomeruli. (G–I) Deep confocal imaging shows the capillary wall (small arrows) positive for phalloidin (G) and synaptopodin (I) but negative for myosin IIA (H). Note the myosin IIA–positive mesangial cell inside the capillary borders (H). Scale bars: 200 nm (A–C) and 2,000 nm (D–I). Ca, capillary; Me, mesangial cell.

Figure 3. STORM and STORM-EM correlative imaging of slit diaphragm molecules in healthy and injured podocytes.

(A–F) Double-color STORM imaging of nephrin and agrin with EM correlation. (A) Low and (D) high magnification images show nephrin clusters in a periodic pattern (red) adjacent to the glomerular basement membrane (GBM) labeled with agrin (blue). (B and E) Deep-etch freeze fracture (DEFF) EM images of the same areas in A and D show individual foot processes on the capillary wall. (C and F) STORM-EM correlation shows that nephrin staining is between individual foot processes. (G–L) STORM (G and J), DEFF-EM (H and K), and STORM-EM correlation images of capillary walls of Cd2ap-KO (G–I) and Adriamycin nephropathy (J and L) mouse models. In contrast to the WT (A–C), nephrin (red) in both injury models was shifted apically away from the GBM (blue). (M–O) STORM images of WT (M), Cd2ap-KO (N), and Adriamycin-injured (O) glomeruli show an apical shift in podocin staining (red) away from the GBM (blue). (P–R) STORM images of WT (P), Cd2ap-KO (Q), and Adriamycin-injured (R) glomeruli show localization of CD2AP clusters (red) relative to agrin (blue). Note the lack of CD2AP staining in Cd2ap-KO (Q). Scale bars: 2,000 nm (A–C and G–L) and 200 nm (D–F and M–R). Ca, capillary; Po, podocyte.

Figure 4. Spatial distribution of synaptopodin after podocyte injury.

(A–F) Double-color STORM imaging of WT (A and B), Cd2ap-KO (C and D), and Adriamycin-injured (E and F) glomeruli stained for synaptopodin (red) and agrin (blue). In both injury models, synaptopodin was found in large clusters, closer to the glomerular basement membrane (GBM) than in WT kidneys. Note the single perpendicular synaptopodin cluster in the Cd2ap-KO glomerulus (arrowhead in D). (G and H) EM and STORM-EM correlative images, respectively, show that the area imaged is effaced. (I) Dot blot showing the sizes of synaptopodin clusters in the 4 models used for this study shows significant increases in the thickness of the clusters upon injury (*P < 0.00005 for all the genotypes by Student’s t test; the experiments and measurements were repeated 2 times). (J) Dot blot showing the distances between the adjacent synaptopodin clusters (edge to edge) shows significant increases in the Cd2ap-KO and in Adriamycin-induced nephropathy (*P < 0.00005 by Student’s t test) but not in the Lamb2-KO (the experiments and measurements were repeated 2 times). Scale bars: 2,000 nm (A, C, E, G, and H) and 200 nm (B, D, and F). Ca, capillary; Po, podocyte.

Figure 5. Contractile actin cables in injured podocytes.

Triple-color STORM overview images of Cd2ap-KO (A) and Adriamycin-treated (E) glomeruli show myosin IIA’s (blue) distribution in comparison with synaptopodin (red) and integrin β1 (green). (B–D and F–H) Higher magnification images of the boxed areas in A and E show the presence of myosin IIA near the glomerular basement membrane (GBM) in both Cd2ap-KO (B–D) and Adriamycin-injured podocytes (F–H) in sarcomere-like structures. (I) Quantification of the percentage of the GBM covered by sarcomere-like structures shows an increase in all injury models over WT. (J–L) Membrane-extracted isolated glomeruli stained for myosin IIA (blue) and synaptopodin (red) were imaged using Airyscan confocal microscopy to show the capillary wall of the WT (J), Cd2ap-KO (K), and Adriamycin-induced injury (L) models. While the WT did not show myosin IIA at the capillary wall (arrows in J), alternating clusters of synaptopodin and myosin IIA were apparent in Cd2ap-KO and Adriamycin nephropathy (arrows in K and L, respectively). Scale bars: 2,000 nm (A, E, and J–L) and 200 nm (B–D and F–H). Ca, capillary; Me, mesangial cell; Po, podocyte.

Figure 6. Direct en face imaging of podocyte foot processes.

En face views of the capillary wall of decapsulated glomeruli from WT (A–D), Cd2ap-KO (E–H), Adriamycin-induced injury (I–L), and Lamb2-KO (M–P) models stained for synaptopodin (red in A, E, I, and M) and myosin IIA (blue in B, F, J, and N). (A–D) Healthy podocyte foot processes are synaptopodin-positive (A and C) and myosin IIA–negative (B and C); myosin IIA–negative cables branch out of myosin IIA–positive cables (arrows and D). (E–P) Surface views of podocytes with foot process effacement show large synaptopodin-positive areas of effacement (arrows) that stain positive for myosin IIA. Scale bars: 4,700 nm (left 3 columns) and 2,350 nm (fourth column).

Figure 7. Myosin IIA is present in the human podocyte cell body and major processes but not in foot processes.

(A–C) Triple-color STORM images of a healthy human glomerulus stained for integrin β1 (green), synaptopodin (red), and myosin IIA (blue) show that while podocyte foot processes are labeled by synaptopodin (arrow in B), myosin IIA is primarily localized in the podocyte cell body and major processes (arrowheads in B and C). (D–F) A healthy human glomerulus stained for synaptopodin (red) and myosin IIA (blue) was imaged by Airyscan super-resolution microscopy. A capillary loop is covered by synaptopodin-positive processes (D) that are negative for myosin IIA (E) and (F). Myosin IIA is present in the podocyte cell body and the major processes (arrowheads), but very little myosin IIA staining overlaps with the finger-like synaptopodin staining (arrows). (G–I) Airyscan images of human kidney biopsies from patients with minimal change disease (G), focal segmental glomerulosclerosis (H), and diabetic nephropathy (I) labeled with antibodies against synaptopodin (red) and myosin IIA (blue) show the presence of sarcomere-like structures in injured podocytes (arrowheads). Scale bars: 2,000 nm (A, B, and D–I) and 200 nm (C). Po, podocyte; Ca, capillary; Me, mesangial cell; RBCs, red blood cells; MCD, minimal change disease; FSGS, focal segmental glomerulosclerosis; DN, diabetic nephropathy.

Figure 8. Model of the location of actin cables in podocytes before and after injury.

(A) WT healthy podocytes contain noncontractile actin cables in the foot processes (analogous to the dorsal stress fibers [DSFs]), and these are connected to the contractile actin cables located in the major processes and the cell body (analogous to the transverse arcs [TAs]). (B) Podocyte injury leads to the disassembly of the DSFs in the foot processes that support the contractile actin cables in the TAs, subsequently leading to the appearance of the sarcomere-like contractile actin cables (analogous to the ventral stress fibers [VSFs]) juxtaposed to the glomerular basement membrane (GBM).