Podocytes Respond to Mechanical Stress In Vitro

Capillaries in the renal glomerulus are formed by three cell types: mesangial cells, endothelial cells, and podocytes. All of these cells are exposed to mechanical load in vivo arising from capillary BP and flow (1). Podocytes possess primary processes and foot processes, which cover the outer aspect of the basement membrane of glomerular capillaries. The cytoskeleton of podocytes is well organized, and all major proteins, necessary to form contractile filaments, are expressed in foot processes (2,3). Recently, a role for podocytes to counteract capillary wall tension was proposed (4). Because increased glomerular capillary pressure and, hence, increased mechanical load are believed to result in glomerular damage (5), understanding the responses of glomerular cells to mechanical force is highly relevant.

With the development of special devices for studying the effect of mechanical force on cultured cells (6,7,8,9), it has become clear that mechanical forces modulate important cell functions. These include differentiation and proliferation, matrix and autacoid production, second messenger levels, and enzyme activities (10,11,12). The cytoskeleton plays a key role not only in providing mechanical stability to increased load but also in the signal transduction of mechanical force (13,14). In particular, cytoskeletal changes in podocytes have been implicated in the action of vasoactive hormones on the permeability of the glomerular filtration barrier (15,16). Formation of stress fibers and their alignment with respect to the force vector has been reported to represent the main alteration of the cytoskeleton in response to mechanical load for a variety of cells, including mesangial cells (9,17,18,19,20,21,22,23).

Concerning glomerular cells, the impact of mechanical stress has been studied only in cultured mesangial cells. In response to mechanical stress, mesangial cells elongate, align perpendicularly to the force vector, and reinforce transversal stress fibers (23). As a result of activation of the mitogen-activated protein (MAP) kinase pathway, mesangial cells exhibit a higher proliferation rate during mechanical stress (24). Furthermore, mechanical stress enhances expression and synthesis of prostaglandins, angiotensinogen, angiotensin-1 receptor, vascular endothelial growth factor, and transforming growth factor-β in mesangial cells, the last being responsible for increased deposition of matrix proteins (23,25,26,27). In contrast to mesangial cells, the response of podocytes to mechanical stress is unknown. In the present study, podocytes were subjected to mechanical stress in vitro, and their morphology and cytoskeleton was examined. Opposite to mesangial cells, the cell body size of podocytes is reduced in response to mechanical stress, and podocytes do not align in the force field. Most notably, mechanical stress induces a unique reorganization of the actin cytoskeleton in podocytes, featuring radial stress fibers and the formation of a special F-actin structure.

Materials and Methods

Results

Biaxial cyclic mechanical stress (0.5 Hz and 5% linear strain for 3 d) induced obvious changes in podocyte cell shape (Figure 1, A and B). Podocyte processes became thin and elongated in response to mechanical stress. The cell body was smaller in stressed as compared with unstressed podocytes. Quantitative analysis showed that cell body width was decreased by 50%, length was reduced by 33%, and, hence, cell body area was reduced by 68% (Figure 1C). Under control conditions, cell size and shape of podocytes were rather variable, whereas in mechanically stressed cells, they were more uniform. Although the strain of the silicone membranes is larger in the radial than in the circumferential direction (36), the orientation of mechanically stressed podocytes was indistinguishable from a random orientation (Figure 1D). The angle of orientation (-90° to 90°) between the longitudinal axis of mechanically stressed podocytes and an arbitrarily defined axis was 1.2° ± 4.9° (n = 106), a value not significantly different from 0°. Likewise, the SD of the orientation angles of mechanically stressed podocytes was 50.0° as compared with the expected value of 52.0° for randomly oriented cells. In addition, we did not note any differences in cell shape or size across the surface of membranes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Podocyte processes were thinner and more elongated in mechanically stressed cells (B) as compared with unstressed controls (A), whereas cell body size of stressed podocytes was reduced. Magnification, ×250. Quantified morphologic changes (cell body length, width, and area) of one experiment are presented in C (means ± SEM; 135 control cells and 106 stressed cells; ^*, P < 0.05). Similar results were obtained in two additional experiments. The frequency distribution of orientation angles (D) of mechanically stressed podocytes (n = 106 stressed cells) was indistinguishable from a random distribution (dashed line). Orientation angles (-90° to 90°) of mechanically stressed podocytes were grouped into intervals of 20°.

Next we examined the cytoskeletal organization of F-actin, microtubules (α-tubulin), and vimentin (Figure 2), which is the dominant intermediate filament in differentiated podocytes (2,28). The network of microtubules and vimentin in podocytes did not show major changes in response to mechanical stress (Figure 2, C through F). In contrast, F-actin organization was altered profoundly in response to mechanical stress (Figure 2, A and B). Transversal stress fibers, which were frequently observed in unstressed podocytes, were diminished in mechanically stressed podocytes and radial stress fibers emerged. The radial stress fibers converged into an ARC. The structure of the ARC in mechanically stressed podocytes was explored further by confocal microscopy. The three-dimensional reconstruction of the actin cytoskeleton from horizontal optical sections of mechanically stressed podocytes revealed that the ARC had an ellipsoidal shape and that the radial fibers were connected to the ARC at its equatorial plane (Figure 3, A and B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Mechanical stress—induced reorganization of the podocyte actin cytoskeleton (A and B; F-actin). Unstressed control cells (A) often exhibited transversal stress fibers (arrowheads). Stress fibers in mechanically stressed podocytes were radially organized (arrow-heads) and connected to an actin-rich center (ARC, arrow). Mechanical stress did not induce major changes of the microtubule and vimentin cytoskeleton (C through F; C and D, α-tubulin; E and F, vimentin). Magnification, ×160.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

The three-dimensional organization of F-actin in mechanically stressed podocytes was reconstructed with the aid of confocal microscopy (A, xy plane; B, xz plane). The ARC resembled an ellipsoid, and stress fibers were connected to the ARC at the equatorial plane. Three-dimensional reconstruction was done from 36 optical sections of 0.2 μm thickness. Magnification, ×1000. Quantitative analysis confirmed that the number of podocytes with transversal stress fibers was significantly diminished by mechanical stress, whereas the percentage of podocytes that possessed an ARC increased to 100% (C). Mechanical stress increased not only the number of ARC-positive cells but also the size of the ARC. The larger diameter of the slightly elliptical ARC is shown. For the frequency of transversal stress fibers and the ARC, 135 control cells and 106 stressed cells were classified. ARC diameter was measured in 30 control cells and 106 stressed cells. Data of one experiment are presented. Similar results were obtained in two additional experiments. ^*, P < 0.05.

The quantitative analysis of the actin cytoskeleton showed that under control conditions, approximately one third of podocytes had prominent transversal stress fibers (Figure 3C). After mechanical stress, transversal stress fibers were present in fewer than 10% of the cells. An ARC was found in 100 ± 0% of mechanically stressed podocytes but in only 31 ± 9% of control cells (three experiments with a total of 250 cells for each condition). The shape of the ARC was slightly elliptical with a larger diameter of 5.2 ± 0.2 μm and a smaller diameter of 3.7 ± 0.1 μm in mechanically stressed podocytes (n = 106). ARC in control cells, when encountered, were much smaller; both ARC diameters amounted to approximately 50% of those in mechanically stressed podocytes. The ARC was rarely found in the cell center (defined at 50% of cell body length and width) but had an excentric position, being found at 35% of length and 15% of width on the average. The number of radial stress fibers in mechanically stressed podocytes, which were connected to the ARC, was variable. Counted in three experiments, 2.3 ± 2.0 (n = 169), 1.5 ± 2.1 (n = 34), and 4.3 ± 3.4 (n = 33) radial filaments were connected to the ARC (mean ± SD).

The structure and composition of the ARC in mechanically stressed podocytes was confirmed further by electron microscopy. By transmission electron microscopy, the ARC appeared as a dense filamentous elliptical structure, with emanating fiber bundles attached to the cell membrane at focal adhesion contacts (Figure 4). The diameter of individual filaments was roughly 7 nm, consistent with actin filaments.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Transmission electron microscopy of mechanically stressed podocytes revealed the ellipsoidal and dense filamentous structure of the ARC. A fiber bundle (arrow) extends from the ARC to a focal contact (arrowhead) at the cell membrane (A). The single actin filaments, which are densely wound up in the ARC, can be resolved at higher magnification (B). The electron micrographs show a horizontal section through a podocyte. Magnifications: ×9500 in A; ×34,000 in B.

Because approximately one third of unstressed podocytes already possessed an ARC, one may argue that mechanical stress selected this subpopulation of podocytes than induced reorganization of the actin cytoskeleton. Because podocytes do not divide under nonpermissive conditions (28), selection can only occur through cell death. Indeed, cell number was lower for mechanically stressed podocytes as compared with unstressed controls. However, after application of mechanical stress for 1 and 3 d, respectively, cell number of mechanically stressed podocytes was only 33 ± 7% (n = 3) and 29 ± 3% (n = 4) lower as compared with unstressed controls. Because cells were lost within the first day of application of mechanical stress, mechanical load might have caused detachment or rapid disintegration of cells. Nuclear fragmentation, which is a hallmark of apoptosis, was observed only in a small fraction (<2%) of mechanically stressed podocytes after staining with Hoechst 33342, indicating that the large majority of mechanically stressed podocytes was viable (Figure 5, A and B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Mechanical stress did not induce apoptosis in stressed podocytes. DNA staining with Hoechst 33342 showed intact nuclei in podocytes, which were mechanically stressed fo r 1 d (B), and in unstressed controls (A). Mechanical stress did not induce an ARC or radial stress fibers in epithelial LLC-PK1 cells. F-actin was visualized in mechanically stressed LLC-PK1 cells (D) and unstressed controls (C). The F-actin cytoskeleton of unstressed primary cultured podocytes (E), which stained positively for nuclear Wilms' tumor-1 (WT-1; F), occasionally showed an ARC (arrow). Magnifications: ×180 in A and B; ×340 in C and D; ×230 in E and F.

In addition to podocytes, LLC-PK1 cells, the phenotype of which is similar to that of epithelial cells of the proximal tubule, and NIH 3T3 fibroblasts were mechanically stressed. Mechanical stress induced an ARC or radial stress fibers in neither LLC-PK1 cells (Figure 5, C and D) nor NIH 3T3 fibroblasts. Thus, the observed reorganization of the actin cytoskeleton in podocytes in response to mechanical stress seems to be a cell type-specific response of podocytes to mechanical stress. Furthermore, an ARC also could be observed in a fraction of unstressed primary cultured podocytes (Figure 5E), which had just grown out from isolated glomeruli and which stained positively for nuclear WT-1 (Figure 5F). The ARC therefore cannot be considered an artifact of the conditionally immortalized podocyte cell line.

The up- and downward motion of silicone membranes, which generates mechanical stress, is associated with a notable fluid motion of the cell culture medium. To confirm that the observed reorganization of the podocyte cytoskeleton was caused by mechanical stress and not by fluid shear stress, we rocked podocytes for 3 d. Fluid motion did not result in formation of radial stress fibers or an increased occurrence of ARC in rocked podocytes as compared with control cells. Further evidence that cytoskeletal changes were caused specifically by mechanical stress can be taken from the time course of reorganization of the actin cytoskeleton. After application of mechanical stress for 1 d, an ARC was found in approximately 50% of the cells (cf. Figure 6E). After 3 d, practically all mechanically stressed podocytes possessed an ARC (cf. Figure 3C). Furthermore, the reorganization of the actin cytoskeleton in response to mechanical stress was reversible within 1 wk after termination of stress.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Reorganization of the F-actin cytoskeleton in response to mechanical stress depended on Ca²⁺ influx and Rho kinase but not on Gd³⁺-sensitive cation channels. Representative images of mechanically stressed podocytes are shown for the different conditions: no treatment (A), 1 mM Ni²⁺ (B), 50 μM Gd³⁺ (C), and Rho kinase inhibition with 10 μM Y-27632 (D). The inhibitory effect of Ni²⁺ and Y-27632 on the stress-induced actin reorganization was quantified by counting of ARC-positive cells after application of mechanical stress for 1 d (E). Data are means ± SEM of three experiments examining 100 to 113 cells per experiment. ^*, P < 0.05. Magnification, ×230.

To gain a first insight into the mechanisms that may transduce the reorganization of the actin cytoskeleton in response to mechanical stress in podocytes, we mechanically stressed cells for 1 d in the presence of different inhibitors (Figure 6). Nonspecific blockade of Ca²⁺ influx with Ni²⁺ almost completely suppressed the formation of the ARC and of radial stress fibers; podocytes were large and broad, showing prominent transversal stress fibers (Figure 6B). The Ca²⁺ influx most likely did not occur through stretch-activated cation channels, because Gd³⁺ was without effect on the cytoskeletal reorganization in response to mechanical stress (Figure 6C). Rho kinase is known to regulate actin stress fiber formation (37). Inhibition of Rho kinase with the specific inhibitor Y-27632 (35) abolished stretch-induced actin reorganization and resulted in bizarre cell morphology (Figure 6D).

To clarify whether typical stress fiber proteins organized the F-actin in the ARC, we performed immunofluorescence with antibodies against the following actin-associated proteins: myosin II, α-actinin, and synaptopodin. None of these proteins was enriched in the ARC, but all of them were present in the radial stress fibers that arose from the ARC (Figure 7, A through C). Figure 7, D through F, shows the co-localization of F-actin and the podocyte-specific protein synaptopodin in detail. There was a broad overlap in F-actin and synaptopodin localization, especially in stress fibers (Figure 7F), as described previously (32). In contrast, the center of the ARC did not stain for synaptopodin (Figure 7D). Fluorescence intensity measurements by confocal microscopy showed that the intensity of F-actin was 20 to 100 times higher in the ARC than in radial stress fibers, whereas the intensities of myosin II, α-actinin, and synaptopodin were always lower in the ARC than in radial stress fibers. Thus, radial stress fibers and the ARC differ in protein composition.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Typical actin-associated proteins of stress fibers were not enriched in the ARC. Confocal sections through the ARC in the plane of stress fiber attachment revealed that the fluorescence intensity of myosin II (A), α-actinin (B), and synaptopodin (C) was considerably lower in the ARC (arrow) than in radial stress fibers. Co-localization of synaptopodin (D) and F-actin (E) in a mechanically stressed podocyte is shown in F. Whereas synaptopodin was co-localized with F-actin in stress fibers, it was practically absent from the center of the ARC. Co-localization was done using conventional fluorescence microscopy. Magnifications: ×720 in A through C; ×600 in D through F.

Finally, we tested whether the different protein composition of radial stress fibers and the ARC would manifest itself in the dynamic regulation of the F-actin. To this end, we treated podocytes, which had been mechanically stressed for 3 d (Figure 8A), with the specific Rho kinase inhibitor Y-27632. In response to Rho kinase inhibition, radial stress fibers disappeared within 20 min (Figure 8C). However, the ARC was not affected by Rho kinase inhibition and remained, even after prolonged treatment (2 h) of cells with the Rho kinase inhibitor. Treatment of mechanically stressed podocytes with cytochalasin D, which inhibits actin polymerization, completely disrupted the F-actin cytoskeleton including the ARC (Figure 8B). Only aster-like actin assemblies remained, which have been reported to be due to polarity sorting of actin filaments by myosin II (38). This finding demonstrates that the regulation of F-actin assembly differs between stress fibers and the ARC.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

Treatment with cytochalasin D completely disrupted the F-actin cytoskeleton (B) of mechanically stressed podocytes as compared with untreated cells (A). Radial stress fibers and the ARC disappeared; only aster-like F-actin assemblies remained (arrow). The Rho kinase inhibitor Y-27632 induced disassembly of radial stress fibers, whereas the ARC remained unaffected (C). Magnification, ×320.

Discussion

In the present study, we showed that podocytes are intrinsically sensitive to mechanical force. The responsiveness of podocytes to mechanical stress was detected by alterations in the cell morphology and the F-actin organization. Mechanical stress induced the formation of radial stress fibers, which converged to a focus, the ARC. Reorganization of the cytoskeleton in response to mechanical stress has been described for a variety of cells. Mechanical stress stimulates stress fiber formation and alignment of stress fibers perpendicular or at an oblique angle to the force vector in endothelial cells (9,17,19,21,22). Likewise, mesangial and smooth muscle cells respond to mechanical stress with an increased number and size of stress fibers, which orient perpendicular to the direction of strain (20,23). Increased formation and thickening of stress fibers was observed further in mechanically stressed osteoblasts (18). In contrast to the cell types studied so far, mechanical stress resulted in a completely different organization of stress fibers in podocytes. Concerning cell shape, endothelial, smooth muscle, and mesangial cells elongate in response to mechanical stress (9,17,19,20,22,23). In contrast, the area of podocyte cell bodies was reduced approximately threefold by mechanical stress. Thus, podocytes exhibit a unique response to mechanical stress as compared with other cell types, including mesangial cells, studied so far.

The microtubule and vimentin cytoskeleton of podocytes was not substantially altered by mechanical stress. Likewise, in endothelial cells, the distribution of tubulin and vimentin was not affected by mechanical strain (17). In smooth muscle cells, microtubule mass has been shown to be slightly altered during step increases of mechanical stress within the first hour (39), whereas the cellular content of actin, tubulin, and vimentin remained unchanged after application of cyclic mechanical stress for 2 wk (40). Using magnetic twisting cytometry, by which rotational mechanical stress is applied to integrin receptors with ferromagnetic beads, Wang (41) demonstrated that mechanical stiffness of endothelial cells is due mainly to actin filaments and that microtubules are responsible for viscous behavior, whereas disruption of intermediate filaments is without effect on mechanical parameters. Taken together, microtubules and vimentin seem to be less important for the cytoskeletal reorganization in response to mechanical stress than actin.

In addition to providing mechanical stability, the actin cytoskeleton and, hence, cytoskeletal reorganization serve a variety of other functions. Actin reorganization precedes and is necessary for the alignment of endothelial cells to mechanical force (17,42). Transduction of mechanical forces into changes of membrane potential, protein kinase activation, and gene expression rely on an intact actin cytoskeleton (10,43,44,45,46). It has been shown that the actin cytoskeleton is even involved in mRNA and ribosome movement in response to mechanical tension (47). In podocytes, reorganization of the actin cytoskeleton has been implicated in the action of vasoactive hormones on the permeability of the glomerular filtration barrier (15,16).

Mechanical stress is sensed by stretch-activated cation channels in many cells (10). Stretch-activated channels are thought to be linked both to the cytoskeleton and to extracellular matrix proteins, enabling deforming forces to be transmitted to the channel proteins (48). In general, stretch-activated cation channels can be blocked by micromolar concentrations of Gd³⁺. However, the F-actin reorganization in podocytes in response to mechanical stress was not affected by Gd³⁺ in our experiments. Nevertheless, Ca²⁺ influx seems to play an essential role in the stress-induced cytoskeletal reorganization in podocytes, because nonselective inhibition of Ca²⁺ influx by Ni²⁺ was highly effective in preventing F-actin reorganization. It is unlikely that Ca²⁺ influx in podocytes occurs through voltagegated Ca²⁺ channels, because they have not been detected in podocytes so far (16). Further studies are needed to characterize the channel that is responsible for Ca²⁺ influx and a possible connection to extracellular matrix proteins.

Formation of radial stress fibers, which converge to the ARC, is the most striking feature in the cytoskeletal reorganization of podocytes in response to mechanical stress. The ARC was found as a single structure in mechanically stressed podocytes with specific properties: (1) typical actin-associated proteins of stress fibers (myosin II, α-actinin, and synaptopodin) were not enriched in the ARC, and (2) in contrast to stress fibers, Rho kinase inhibition in mechanically stressed podocytes did not result in a rapid disassembly of the ARC. Thus, the protein composition and the dynamic regulation of the ARC seem to be distinct from that of stress fibers. It has been reported recently that transfection of epithelial cells with a dominant active mutant of the Rho kinase induces formation of radial stress fibers (49,50). The Rho kinase—induced radial stress fibers focused to a narrow area, forming a condensed F-actin structure, reminiscent of the ARC observed in our experiments. Thus, the radial stress fiber arrangement in podocytes could indicate that mechanical stress activates the Rho kinase in podocytes more strongly than in other cell types. Indeed, our results demonstrate that the Rho kinase is necessary for the F-actin reorganization in podocytes in response to mechanical stress, as application of mechanical stress for 1 d in the presence of the Rho kinase inhibitor Y-27632 prevented the formation of radial stress fibers and the ARC. Thus, initially, the formation of the ARC depends on the presence of stress fibers, which are sensitive to Rho kinase blockade. Later, when the ARC has been formed, the ARC is no longer affected by the rapid disassembly of stress fibers in response to Rho kinase inhibition.

The present study clearly demonstrated that podocytes respond to mechanical stress in vitro. This finding is likely to bear relevance for the in vivo situation. However, it is obvious that the cytoskeletal arrangement of cultured podocytes deviates from that of podocytes in vivo, where actin fibers are largely restricted to the foot processes. Furthermore, it is not fully evident how mechanical forces act on podocytes in vivo. Foot processes, which are attached to the capillary wall, are likely to experience capillary wall distention. Whether mechanical forces are transmitted to the cell body of podocytes via their primary processes is unclear. In addition, a dense F-actin structure, similar to the ARC, has not yet been observed in podocytes in vivo, neither in healthy glomeruli nor in conditions of increased mechanical stress, e.g., glomerular hypertension. Despite these limitations of the present study, future in vitro experiments may identify signal transduction cascades as well as cytoskeletal and other proteins involved in the response of podocytes to mechanical stress. These findings may well guide corresponding in vivo studies.

In summary, we have characterized the morphologic and cytoskeletal changes of podocytes in response to mechanical stress. Podocytes in vitro exhibit a unique Ca²⁺ influx— and Rho kinase—dependent reorganization of the actin cytoskeleton, consisting of radial stress fiber formation connected to an ARC. Thus, podocytes can be regarded as intrinsically mechanosensitive cells.