Piezo1, F-Actin Remodeling, and Podocyte Survival and Regeneration

Visual Abstract

Abstract

Key Points

• Piezo1 regulates perinuclear actin remodeling, and its loss induces nuclear envelope defects and accumulation of DNA damage.

• Piezo1 loss in podocytes leads to mitotic catastrophe.

• Podocyte-specific and podocyte progenitor–specific knockout of Piezo1 in mice result in severe albuminuria after adriamycin nephropathy.

Background

Podocytes and podocyte progenitors are interdependent components of the kidney's glomerular structure, with podocytes forming the glomerular filtration barrier and progenitors being key players in podocyte regeneration during pathophysiologic processes. Both cell types are subjected to constant mechanical forces, whose alterations can initiate podocytopathy and worsen glomerular injury. Despite this, the specific mechanosensors and mechanotransduction pathways involved in their response to mechanical cues remain only partially explored.

Methods

We used transcriptomics, immunofluorescence, and silencing experiments on human primary podocyte progenitor cell cultures to demonstrate the expression and function of Piezo1 channels. We generated inducible podocyte-specific and podocyte progenitor–specific Piezo1 knockout (KO) mice to evaluate the effects of Piezo1 loss in the context of adriamycin nephropathy and over 10 months of aging.

Results

Silencing of Piezo1 in progenitors triggered F-actin remodeling, which induced cell shape modification and nuclear envelope defects with accumulation of DNA damage that led to mitotic catastrophe in differentiated podocytes. Podocyte-specific KO of Piezo1 induced higher susceptibility to podocyte injury in adriamycin nephropathy and led to accumulation of DNA damage and mild albuminuria starting from adult age. Podocyte progenitor–specific KO of Piezo1 in mouse resulted in severe albuminuria during adriamycin nephropathy, leading to the generation of defective podocytes.

Conclusions

These results demonstrated that Piezo1, thanks to its role in F-actin cytoskeleton maintenance, is essential for the survival of podocytes exposed to mechanical stress conditions and for their correct regeneration.

Introduction

Several types of physical forces are present in the glomerulus even under normal conditions, e.g., fluid shear stress, circumferential stretch, osmotic stress, stiffness, and elastic resistance of the basement membrane.^1,2 Modifications of these forces are observed in medical conditions^3,4 with a consequent increase in podocyte stress, detachment, and loss, leading to progressive glomerulosclerosis and to CKD,^5–7 a major health care problem affecting approximately 15% of the population.⁸ Podocyte loss triggers a rapid response in the parietal epithelial cells of the Bowman capsule,^9–11 a heterogeneous epithelium including a population of progenitors capable of regenerating podocytes.^11–13 These cells maintain homeostasis, fulfill endogenous regenerative processes after injury, and can contribute to pathologic conditions through defective or abnormal responses.^9,11,13–16 Although many studies have provided evidence of the effects of different biochemical stimuli on progenitors,^{10,12,17–19} the molecular mechanisms by which they perceive the physical stimulus have been only partially elucidated.²⁰ Mechanosensitive channels are deputed to convert physical stimuli into electrical and calcium signals, activating intracellular signal cascades.²¹ Groundbreaking developments in this area include the discovery of the mechanosensitive channels Piezo1 and Piezo2.^22,23 The Piezo2 channel largely functions in neuronal cells, while Piezo1 exhibits a broader expression pattern with context-dependent functional roles.²³ In the kidney, Piezo1 has been reported in the rodent nephron and collecting duct,^24–27 mediating stretch sensitivity in proximal tubular cells,²⁶ and regulating urinary osmoregulation in principal cells of the collecting duct.²⁷ Few studies reported the expression of Piezo1 in mouse glomeruli, opening new perspectives on the hypothesis that it may have a role in glomerular disease and represent a new target for drugs.^28,29 Indeed, in the kidney, targeting mechanobiology is an interesting topic that could improve therapeutic precision and efficacy.³⁰ We investigated the expression of Piezo1 in podocytes and podocyte progenitors in vitro and in vivo and explored its function in glomerular physiopathology.

Methods

Cell Cultures

Human CD133⁺CD24⁺ podocyte progenitors were obtained in agreement with the Ethical Committee on Human Experimentation of the Azienda Ospedaliero-Universitaria Careggi, Florence, Italy (project n. 17155/CAM_BIO). All participants gave their informed consent for inclusion before they participated in the study.

Animal Studies

The Pax2.rtTA; TetO.Cre; R26.mT/mG.Piezo1^floxed (Pax2-Piezo1-knockout [KO]) and NPHS2.iCreER^T2;R26.mT/mG.Piezo1^floxed (NPHS2-Piezo1-KO) mice were developed on a C57Bl/6 background by crossing the Pax2.rtTA; TetO.Cre; R26.mT/mG or the NPHS2iCre; R26.mT/mG. mouse previously generated,³⁰ with Piezo1^tm2.1Apat (Piezo1^floxed) (The Jackson Laboratory, Bar Harbor, ME). Pax2.rtTA; TetO.Cre; R26.mT/mG (Pax2-WT) and NPHS2iCre; R26.mT/mG (NPHS2-WT) mice were used as wild-type (WT) controls. Pax2.rtTA; TetO.Cre; R26.Confetti mice were used for single-cell RNA sequencing (scRNAseq) experiments. Homozygous mice were used.

Statistical Analysis

The results were expressed as mean±SEM or as median (interquartile range). Comparison between groups was performed by Mann–Whitney U test, t test, or through the ANOVA for multiple comparisons (ANOVA for repeated measures) with Bonferroni post hoc analysis. Pearson correlation coefficient was used to measure linear correlation between two sets of data. A P value < 0.05 was considered statistically significant. Statistical analysis was performed using OriginPro software.

Results

Piezo1 Was Necessary for F-Actin Cytoskeleton Organization in Podocyte Progenitors

In human healthy glomeruli, we revealed Piezo1 expression in progenitors, identified as cells expressing CD133 in the Bowman capsule (Figure 1A and Supplemental Figure 1A), and in podocytes, identified by staining with synaptopodin (Figure 1B and Supplemental Figure 1B) and nephrin (Figure 1, C and D, and Supplemental Figure 1C). The almost complete colocalization of Piezo1 with nephrin inside the glomerulus allowed us to exclude its expression in mesangial and endothelial cells, as further demonstrated by the absence of colocalization with CD31 (Supplemental Figure 1D).

Figure 1.

Piezo1 orchestrates F-actin cytoskeleton and nuclear shape in podocyte progenitors. (A) Confocal image of a human glomerulus of a healthy participant showing Piezo1 expression (green) in PECs expressing the progenitor marker CD133 (red). DAPI stains nuclei. Bar=25 μm. (B) Confocal image of a human glomerulus of a healthy participant showing Piezo1 expression (green) in podocytes expressing synaptopodin (syn, red). DAPI stains nuclei. Bar=25 μm. (C) Double staining immunofluorescence for Piezo1 (green) and nephrin (red) in human glomerulus of a healthy participant. DAPI stains nuclei. Bar=25 µm. (D) High-magnification image showing colocalization of Piezo1 and nephrin signal. Bar=5 µm. (E) Piezo1 mRNA expression in podocyte progenitors cultured in static condition and following application of shear stress. Data are expressed as percentage over static culture condition (n=4). (F) Piezo1 mRNA expression in podocyte progenitors cultured on substrates with increasing stiffness. Data are expressed as percentage over 0.5 kPa. (n=4). (G) Piezo1 protein expression in podocyte progenitors cultured on substrate of 2 and 50 kPa elastic module. At least 40 cells for each condition were analyzed. Quantification was performed through the assessment of Piezo1 fluorescence staining. (H) Representative confocal microscopy showing the organization of the actin cap in scramble (top) and Piezo1-silenced (bottom) podocyte progenitors. F-actin was stained using phalloidin, and nuclear DNA was stained using DAPI. Bars=10 µm. (I) Cross-sectional views from the xy image showing perinuclear actin (arrow) in scramble podocyte progenitors (top) and its absence in Piezo1-silenced cells (bottom). Bars=10 µm. (J) Quantification of perinuclear actin expression in scramble-treated and Piezo1-silenced podocyte progenitors. At least 20 cells for each condition were analyzed in two independent experiments. (K) Confocal images showing the morphology of nuclear envelope in scramble (left) and Piezo1-silenced podocyte progenitors (right). Lamin-A is showed in green; nuclear DNA was stained using DAPI. Bars=7.5 µm. (L) Quantification of the percentage of abnormal nuclei in scramble-treated and Piezo1-silenced podocyte progenitors. At least 40 cells for each condition were analyzed from two independent experiments. (M) Nuclei area of scramble-treated and Piezo1-silenced podocyte progenitors. At least 15 cells for each condition were analyzed in two independent experiments. (N) Confocal images of nuclei of scramble (left) and Piezo1-silenced podocyte progenitors (right) showing γH2AX staining (green). DAPI stains nuclei. Bars=25 µm. (O) Quantification of the number of γH2AX foci per nucleus in scramble-treated and Piezo1-silenced podocyte progenitors. At least 30 cells for each condition were analyzed in two independent experiments. (P) Flow cytometry analysis of γH2AX stained scramble-treated and Piezo1-silenced podocytes (n=4). Box-and-whisker plots: line=median, box=25%–75%, whiskers=10%–90%. Mann–Whitney U test for (E), (G), (J), (L), (M), and (O); one-way ANOVA followed by Bonferroni post hoc test for (F); two sample t test for (P). A.U., arbitrary units; DAPI, 4′,6-diamidino-2-phenylindole; LNA, locked nucleic acid; PEC, parietal epithelial cells; γH2AX, phosphorylated histone protein H2AX.

scRNAseq on three primary cell cultures of CD133⁺CD24⁺ human progenitor cells (podocyte progenitors) showed that Piezo1 was the mechanosensitive channel expressed at the highest level in the cells composing the clusters (Supplemental Figure 1, E–K). Its expression and functionality were confirmed by immunofluorescence (Supplemental Figure 2A) and intracellular calcium measurements in response to the Piezo1 agonist Yoda1³¹ performed using confocal microscopy (Supplemental Figure 2, B and C) and flow cytometry (Supplemental Figure 2D). Downregulation of Piezo1 expression with a 16-mer locked nucleic acid (LNA)-gapmer (LNA-Piezo1, Piezo1-silenced cells), verified by quantitative RT-PCR and immunofluorescence (Supplemental Figure 2, E–G), blunted the increase in intracellular calcium induced by Yoda1 (Supplemental Figure 2H), validating its specificity for Piezo1. Furthermore, Piezo1 expression was upregulated in response to shear stress and increasing substrate stiffness (Figure 1, E–G).

Gene microarray analysis in LNA-scramble and LNA-Piezo1 primary cell cultures demonstrated segregation of the samples into two distinct groups based on transcript expression levels (Supplemental Figure 2I). The Piezo1 transcript was significantly downregulated in the Piezo1-silenced cells (P = 0.006). Gene set enrichment analysis through annotation with the Kyoto Encyclopedia of Genes and Genomics pathways showed that the “Regulation of actin cytoskeleton” gene set was underexpressed in Piezo1-silenced cells (Supplemental Figure 2J), with the leading-edge subset including genes involved in the regulation of Rho-GPTase activity (ARHGEF12, DOCK1)³² and encoding Rho-GTPase effectors (WASP, DIAPH2, and DIAPH3).³²

Morphologically, Piezo1-silenced podocyte progenitors acquired a more symmetric shape compared with scramble cells (Supplemental Figure 3A). Similarly, Piezo1 pharmacologic inhibition with GsMTx4 or activation with Yoda1 induced acquisition of a regular multisided polygonal shape or a spindle shape, respectively (Supplemental Figure 3B). Morphometric analysis of circularity showed higher cell circularity with a loss of correlation between cell and nucleus shape in Piezo1-silenced cells (Supplemental Figure 3, C and D) that also exhibited higher average nuclear shape factor with a narrow distribution (Supplemental Figure 3E). Piezo1-silenced cells showed an overall reduction in phalloidin staining (Supplemental Figure 3F) and F-actin cytoskeleton reorganization, with a prevalence of peripheral actin arcs compared with the presence of multiple parallel lines of stress fibers in scramble cells (Supplemental Figure 3G). GsMTx4 induced modifications similar to Piezo1 silencing, while Yoda1 increased stress fiber formation (Supplemental Figure 3H). The modifications in F-actin induced by Yoda1 were abolished in the presence of the Rho inhibitor (Supplemental Figure 3, I–K), and those induced by Piezo1 silencing were reverted by the Rho activator II (Supplemental Figure 3, L–N), suggesting the involvement of RhoA in the regulation of actin cytoskeletal dynamic after Piezo1 signaling. Indeed, pulldown assays demonstrated reduced levels of active RhoA and Rac1 in Piezo1-silenced progenitors (Supplemental Figure 3, O and P).

The perinuclear actin is a structure forming a cap above the apical surface of the nucleus, responsible for the regulation of nuclear shape and its coupling to cell shape.³³ Confocal cross-sectional images of F-actin staining in podocyte progenitors revealed a less organized perinuclear actin cap in Piezo1-silenced cells (Figure 1, H–J). Moreover, Piezo1-silenced cells presented profound aberrations in nuclear morphology (Figure 1, K and L), with wrinkling of the nuclear envelope and nuclear invaginations being the prevalent deformations, associated with a reduction in nuclear area (Figure 1M). Nuclear envelope deformation and rupture can lead to DNA damage.³⁴ Accordingly, Piezo1-silenced podocyte progenitors showed an increase in nuclei presenting C-terminus phosphorylated histone protein H2AX (γH2AX) foci and in the number of foci per nucleus (Figure 1, N–P), demonstrating induction of DNA double-strand breaks after Piezo1 silencing.

Piezo1 Protected Podocytes from Mitotic Catastrophe

An association between glomerular accumulation of DNA double-strand breaks, specifically in podocytes, and glomerular diseases has been reported.^35–37 We investigated whether Piezo1 silencing induced F-actin reorganization and DNA double-strand breaks also in podocytes, obtained by 48 hours of treatment of podocyte progenitors with all-trans retinoic acid.³⁸ Piezo1 silencing did not hamper the differentiation of the progenitor cells, evaluated through quantification of nephrin expression (Figure 2A), but induced alterations in the distribution of F-actin (Figure 2, B and C). These modifications were reverted in the presence of the Rho activator II (Figure 2D), suggesting the involvement of RhoA after Piezo1 signaling also in podocytes. In addition, Piezo1-silenced podocytes presented alterations in nuclear shape (Figure 2, C and F) and an increase in the number of nuclei with γH2AX foci and in the number of foci per nucleus (Figure 2, E–H). More specifically, Piezo1-silenced podocytes frequently presented multiple micronuclei (Figure 2I), multiple nuclei (Figure 2J), and chromosome bridges (Figure 2K), phenotypic characteristics of mitotic catastrophe, a cell death modality observed in podocyte.^39–41 Interestingly, Piezo1-silenced podocytes exposed to cell culture conditions that artificially reproduce the shear stress associated with hyperfiltration underwent apoptosis (Figure 2L).

Figure 2.

Piezo1 protects podocyte from DNA damage accumulation, thus preventing mitotic catastrophe. (A) Assessment of nephrin mRNA expression in scramble-treated and Piezo1-silenced podocyte progenitors differentiated toward podocytes for 48 hours with 100 μM ATRA (n=3 experiments). Results are reported as fold higher over untreated cells. (B–D) Representative confocal images showing F-actin (red) in scramble (B), Piezo1-silenced podocytes (C), and Piezo-silenced podocytes in presence of Rho activator II (D). DAPI stains nuclei. Bars=25 μm. (E and F) γH2AX foci (green) in scramble (E) and Piezo1-silenced podocytes (F). DAPI stains nuclei. Bars=25 µm. (G) Quantification of the number of γH2AX foci per nucleus in scramble-treated and Piezo1-silenced podocytes. At least 20 cells for each condition were analyzed in two independent experiments. (H) Flow cytometry analysis of γH2AX stained scramble-treated and Piezo1-silenced podocytes (n=4 experiments). (I) Confocal images showing micronuclei (arrows) and multiple micronuclei (arrowhead) in Piezo1-silenced podocytes showing γH2AX foci (green). DAPI stains nuclei. Bars=7.5 µm. (J) Confocal images showing multinucleated Piezo1-silenced podocytes (asterisks). Bars=7.5 µm. (K) Representative image of a chromosome bridge observed during the last stages of mitosis in Piezo1-silenced podocyte (red, γ-tubulin, white DAPI). Bar=5 µm. (L) Annexin V/propidium iodide flow cytometry in scramble-treated and Piezo1-silenced podocytes cultured in standard static condition or subjected to shear stress. Box-and-whisker plots: line=median, box=25%–75%, whiskers=10%–90%. Mann–Whitney U test for (A) and (G); two sample t test for (H). ATRA, all-trans retinoic acid.

Piezo1 Was Expressed in Podocytes and Podocyte Progenitors in Mouse Glomeruli

In healthy mice, the expression of Piezo1 in podocyte progenitors and podocytes was demonstrated by immunofluorescence on kidneys of Pax2.rtTA; TetO.Cre; R26.mT/mG (Pax2-WT) and NPHS2iCre; R26.mT/mG (NPHS2-WT) mice in which, after transgene induction, green fluorescent protein (GFP) was expressed in progenitors or in podocytes, respectively (Figure 3, A and B). Piezo1 was almost absent in mesangial and endothelial cells, identified as PDGFRβ- or CD31-expressing cells (Supplemental Figure 4, A and B). RNAscope in situ hybridization assay confirmed Piezo1 expression in the murine glomerulus at the mRNA level (Supplemental Figure 4C). Piezo1 expression in the mouse kidney was further assessed by scRNAseq of glomeruli isolated from Pax2-Confetti mice using the Dynabeads method. In these mice, induction with doxycycline causes Pax2+ progenitors to randomly express one of four fluorescent proteins (yellow/YFP, cyano/CFP, red/RFP, and green/GFP), allowing their identification and tracking.^11,42 Unsupervised clustering of the scRNAseq data enabled the identification of different cell types, which include podocytes and podocyte progenitors (Figure 3C and Supplemental Figure 4, D–F).¹³ The expression of Pax2, Claudin 1, and the fluorophores encoded by the Confetti reporter validated the assignment of cells to the progenitor cluster (Figure 3D). Piezo1 was expressed in approximately 30%–40% of the cells in podocyte and podocyte progenitor clusters (Figure 3, E and F). Gene set enrichment analysis in Piezo1+ versus Piezo1− cells showed enrichment in “Integrin signaling” and “Focal adhesion (FA) assembly” gene sets in the Piezo1+ progenitor cells (Figure 3G), and in “Rho matrix” gene set (Figure 3H) and in “Signaling through Rho-GTPases” (Supplemental Figure 4G) in Piezo1+ podocytes. Comparison of the transcriptomes of progenitors versus podocytes highlighted upregulation of a panel of genes involved in DNA repair in progenitors (Supplemental Figure 4H). To determine whether these modulations of gene sets, which are canonical representations of a biologic process compiled by domain experts,⁴³ corresponded to the involvement of Piezo1 in those specific biologic processes, we analyzed the effects of Piezo1 silencing on cell adhesion and FAs. Piezo1 silencing in podocyte progenitors reduced cell adhesion (Figure 3I) and FA dimension (Figure 3J) and orientation (Figure 3, K and L), with FA oriented along the major axis of cells in scramble (higher percentage of FA showing angles between 0° and 30°), and randomly distributed in Piezo1-silenced podocyte progenitors. Interestingly, exposure of Piezo1-silenced podocytes to the focal adhesion kinase activator ZINC40099027 or the Rho Activator II reduced the apoptosis induced by shear stress (Figure 3M), suggesting that Piezo1 signaling maintains podocyte survival by regulating the integrin signaling through focal adhesion kinase and the RhoA activity.

Figure 3.

Piezo1 regulates FA elongation and orientation. (A and B) Confocal microscopy images showing expression of Piezo1 (white) in Pax2+ podocyte progenitors (green, A) and in podocytes (green, B) assessed in Pax2-WT mouse and in the NPHS2-WT mouse, respectively. DAPI stains nuclei. Bar=25 µm. (C) UMAP showing podocyte and podocyte progenitor mouse clusters. (D) Violin plot showing expression of the podocyte progenitor–specific markers Cldn1, Pax2, and Fluo in the progenitor cluster and their absence in the podocyte cluster. (E) Dot plot showing expression of Piezo1 in podocyte and podocyte progenitor clusters. Color scale indicates average expression level; circle size indicates the percentage of cells expressing the gene. (F) UMAP of podocytes and progenitors showing Piezo1 expression. (G) Matrix plot showing enrichment of genes in the integrin signaling (top) and FA assembly (bottom) gene set in Piezo1+ podocyte progenitors. (H) Matrix plot showing enrichment of genes in the Rho matrix gene set in Piezo1+ podocytes. (I) Analysis of cell adhesion in scramble-treated and Piezo1-silenced podocyte progenitors. Data are expressed as percentage over scramble. (J) Quantification of FA length. At least 200 FA for each condition were analyzed. (K) Confocal images showing elongated, polarized, and oriented FA in podocyte progenitors 48 hours after transfection with scramble (top) or punctate type and randomly distributed FA in Piezo1-silenced podocyte progenitors (bottom). Bars=25 µm. (L) Rose plot of the angles of adhesions respective to the cell major axis in scramble-treated and Piezo1-silenced podocyte progenitors. At least 150 FA for each condition were analyzed. (M) Apoptotic cells assessed by annexin V/propidium iodide flow cytometry in scramble-treated and Piezo1-silenced podocytes subjected to shear stress in presence or absence of ZINC40099027 or Rho Activator II. Mann–Whitney U test for (I), (J), and (M). FA, focal adhesion; Fluo, fluorophores; GFP, green fluorescent protein; UMAP, uniform manifold approximation and projection; WT, wildtype.

Podocyte-Specific Piezo1 KO Aggravated Albuminuria after Injury and during 10 Months of Aging

We then evaluated the role of Piezo1 in podocytes in vivo in adriamycin nephropathy, a rodent model of podocyte damage associated with attempts of regeneration operated by progenitors.^17,19 This was studied in both NPHS2-WT mice and the newly generated NPHS2-Piezo1 KO mice, in which tamoxifen treatment induced the loss of Piezo1 expression in podocytes, as assessed by immunofluorescence and Western blot performed on isolated glomeruli (Supplemental Figure 5, A–C). Quantitative RT-PCR and Western blot analysis showed upregulation of Piezo1 in kidney cortex fragments of adriamycin-treated NPHS2-WT but not NPHS2-Piezo1-KO mice (Supplemental Figure 5, D–F). However, Piezo1 was absent in glomeruli of adriamycin-treated NPHS2-Piezo1-KO mice (Supplemental Figure 5, G and H), with podocytes as the sole glomerular cells expressing Piezo1 in WT mice (Supplemental Figure 5, I and J). Thus, the upregulation of Piezo1 in NPHS2-WT mice after adriamycin was due to an increased expression in podocytes. The NPHS2-Piezo1-KO mice exhibited an albumin:creatinine ratio higher than that of the WT mice (Figure 4A) and worsened glomerular injury (Figure 4, B and C). Interestingly, the albumin:creatinine ratio was higher in the KO mice starting from day 14 and showed a partially remitting course, suggestive of podocyte regeneration mediated by progenitors expressing Piezo1. The number of podocytes per glomerular section, assessed through counting of nephrin+ cells in the NPHS2-WT and NPHS2-Piezo1-KO mice, as well as the podocyte density, were similar (Supplemental Figure 5, K–M). Optical clearing of kidney tissue, immunofluorescent staining for nephrin, and 3D reconstruction revealed reduced foot process coverage (Figure 4, D–G). Cross-sectional images of glomerular capillaries stained with wheat germ agglutinin and collagen 4 showed both foot process effacement as well as a thickened, more heterogeneous, “bumpy” glomerular basement membrane (GBM) in NPHS2-Piezo1-KO mice in comparison with NPHS2-WT mice (Figure 4, H–K). Finally, we observed lower RhoA and phospho-RhoA (Thr100) expression in glomeruli of adriamycin-treated NPHS2-Piezo1-KO mice compared with WT mice that support the involvement of Rho-GTPase in Piezo1 signaling also in vivo (Supplemental Figure 5, N and O).

Figure 4.

Loss of Piezo1 in NPHS2-Piezo1-KO mice results in worsening of proteinuria, effacement of foot processes, and alterations to the GBM in adriamycin nephropathy. (A) Albumin:creatinine ratio in urine from NPHS2-WT and NPHS2-Piezo1-KO mice with and without adriamycin nephropathy (n=7 male mice for each group). (B) Percentage of glomeruli with sclerosis in NPHS2-WT (n=6) and NPHS2-Piezo1-KO (n=6) mice at day 28. (C) Representative PAS-stained kidney sections from NPHS2-WT and NPHS2-Piezo1-KO mice at day 28 after adriamycin nephropathy induction. Bars=100 µm. (D–F) Representative 3D reconstructions of whole glomeruli after optical tissue clearing in NPHS2-WT healthy mice (D), NPHS2-WT mice at day 28 after adriamycin treatment (E), and in NPHS2-Piezo1-KO mice at day 28 after adriamycin treatment (F). Nephrin expression is shown with a depth coding profile to preserve z-information. Bars=10 μm. In the boxed area, representative images of podocyte foot processes by using STED super-resolution microscopy applied to a selected area of the whole glomeruli. Nephrin is stained in blue. Bars=2 μm. (G) Corresponding density of foot processes assessed using STED super-resolution microscopy in nephrin-stained kidney sections. Quantification was performed for five randomly selected areas for each glomerulus in at least five glomeruli per mouse (n=4 mice per group). Each dot in the graph represents a glomerulus. (H–J) Cross-sectional images of glomerular capillaries showing both effacement (wider foot processes, asterisks), as well as a thickened, “bumpy” (marked with arrows) GBM as visualized by collagen 4 staining (green) in NPHS2-Piezo1-KO mice. WGA stains the apical side of podocyte foot processes (glycocalyx). Bars=2 μm. (K) Measurements of GBM thickness in the four groups of mice showing both a thickened and more heterogeneous (larger spread of the data) glomerular basement membrane in NPHS2-Piezo1-KO mice. Data are presented as individual GBM measurements in capillaries (at least 80 measurements) from at least five different glomeruli from healthy NPHS2-WT (n=2), NPHS2-WT with adriamycin nephropathy (n=3), healthy NPHS2-Piezo1-KO (n=2), and NPHS2-Piezo1-KO with adriamycin nephropathy mice (n=4). Two-way ANOVA with Bonferroni correction for (A). Mann–Whitney test for (B), (G), and (K). ADN, adriamycin nephropathy; FP, foot process; GBM, glomerular basement membrane; KO, knockout; PAS, periodic acid–Schiff; STED, stimulated emission depletion; WGA, wheat germ agglutinin.

We also investigated whether Piezo1 KO in podocytes induced age-related effects by analyzing the mice from 5 weeks to 10 months of age (Supplemental Figure 6, A and B). In the NPHS2-Piezo1-KO group, we observed mild histologic alterations (Supplemental Figure 6C) in a higher percentage of glomeruli compared with same-aged WT animals (Supplemental Figure 6D). Accordingly, in the NPHS2-Piezo1-KO mice, the albumin:creatinine ratio increased from 5 months to middle age (10 months, Supplemental Figure 6E) with a concomitant reduction in the slit diaphragm density in comparison with same-aged WT mice (Supplemental Figure 6, F and G) indicating foot process effacement. A higher number of podocytes positive for γH2AX was also observed in middle-aged NPHS2-Piezo1-KO mice (Supplemental Figure 6, H and I).

Piezo1 KO in Podocyte Progenitors Aggravated Proteinuria by Impairing Their Differentiation

We then evaluated the role of Piezo1 in podocyte progenitors in vivo in adriamycin nephropathy induced in Pax2-WT mice and in the newly generated podocyte progenitor–specific Piezo1 KO mice (Pax2-Piezo1-KO). Piezo1 KO was induced by doxycycline treatment, and the loss of Piezo1 expression in Pax2+ progenitors was confirmed by immunofluorescence (Supplemental Figure 7, A and B). We did not observe alterations of Piezo1 expression after adriamycin treatment in Pax2-positive progenitors in both WT and Pax2-Piezo1-KO mice (Supplemental Figure 7, C–E). Pax2-Piezo1-KO mice showed higher proteinuria than the Pax2-WT mice, with onset at day 21 (Figure 5A), associated with worse glomerular damage (Figure 5, B and C) and a higher percentage of glomeruli with hyperplastic lesions originating from the Pax2+ cells (Figure 5, D and E). We evaluated podocyte regeneration by quantifying the GFP+ cells inside the glomerular tuft showing morphologic characteristics of podocytes. At day 28, the percentage of glomeruli with GFP+ cells within the tuft (Supplemental Figure 7F) and the number of GFP+ cells/glomerular section (approximately one cell) were similar in both groups. However, the GFP+ cells inside the glomeruli of Pax2-WT mice exhibited morphologic features of podocytes (Figure 5F), whereas GFP+ cells in the glomeruli of Pax2-Piezo1-KO mice had smaller cell body and a simplified morphology with less finely arborized protrusions (Figure 5G), although they expressed nephrin (Figure 5H). Interestingly, we observed a higher percentage of glomeruli in the Pax2-Piezo1-KO mice containing fragmented GFP+ structures, reminiscent of cell debris (Figure 5, I and J). Finally, at day 28, the number of podocytes per glomerular section in both groups was similar (Supplemental Figure 7G), but the Pax2-Piezo1-KO mice presented a wider glomerular area (Figure 5K) and, consequently, reduced podocyte density (Supplemental Figure 7H). A significant reduction in foot process coverage (Figure 6, A–D) as well as a thickened, more heterogeneous, “bumpy” GBM (Figure 6, E–H) was observed at day 28 in Pax2-Piezo1-KO mice.

Figure 5.

Loss of Piezo1 in Pax2-Piezo1-KO results in worsening of proteinuria during adriamycin nephropathy. (A) Albumin:creatinine ratio in urine from Pax2-WT and Pax2-Piezo1-KO mice with and without adriamycin nephropathy (n=8 for each group, four male and four female). (B) Percentage of sclerotic glomeruli per slice in Pax2-WT (n=7) and Pax2-Piezo1-KO mice (n=7) mice with or without adriamycin nephropathy. (C) Representative image of PAS-stained kidney sections from Pax2-WT and Pax2-Piezo1-KO mice at day 28. Bars=100 µm. (D) Representative images of a kidney cortex of Pax2-Piezo1-KO mouse showing glomerulus with hyperplastic lesion. Bar=25 µm. (E) Percentage of glomeruli presenting hyperplastic lesions in Pax2-WT (n=4) and Pax2-Piezo1-KO mice (n=4) mice with or without adriamycin nephropathy. (F and G) Representative image of a Pax2-WT mouse glomerulus (F) showing GFP+ cell with complex interdigitated foot processes and a representative glomerulus of a Pax2-Piezo1-KO mouse (G) showing less arborized morphology at day 28 after adriamycin nephropathy induction. Bar=10 µm. (H) Confocal image of a glomerulus of Pax2-Piezo1-KO mouse showing nephrin expression in GFP+ cell. Bars=25 µm. (I) Representative confocal image showing fragmented GFP+ cells in the glomerular tuft. (J) Quantification of glomeruli with fragmented GFP+ cells in the glomerular tuft in Pax2-WT (n=4) and Pax2-Piezo1-KO mice (n=4) with or without adriamycin nephropathy. (K) Evaluation of glomerular area of Pax2-WT (n=4) and Pax2-Piezo1-KO mice (n=4) with or without adriamycin nephropathy. Two-way ANOVA with Bonferroni correction for (A). Mann–Whitney U test for (B), (E), (J), and (K).

Figure 6.

Loss of Piezo1 in Pax2 cells induces effacement of foot processes as well as alterations to the GBM in adriamycin nephropathy. (A–C) Representative 3D reconstructions of whole glomeruli after optical tissue clearing in Pax2-WT healthy mice (A), Pax2-WT mice at day 28 after adriamycin treatment (B), and in Pax2-Piezo1-KO mice at day 28 after adriamycin treatment (C). Nephrin expression is shown with a depth coding profile to preserve z-information. Bars=10 μm. In the boxed area, representative images of podocyte foot processes by using STED super-resolution microscopy applied to a selected area of the whole glomeruli. Nephrin is stained in blue. Bars=2 μm. (D) Corresponding density of foot processes assessed using STED super-resolution microscopy in nephrin-stained kidney sections. Quantification was performed for five randomly selected areas for each glomerulus in at least five glomeruli per mouse (n=4 mice per group). Each dot in the graph represents a glomerulus. (E–G) Cross-sectional images of glomerular capillaries showing both effacement (wider foot processes, asterisks), as well as a thickened, “bumpy” (marked with arrows) GBM as visualized by collagen 4 staining (green). WGA stains the apical side of podocyte foot processes (glycocalyx). Bars=2 μm. (H) Measurements of GBM thickness in the four groups of mice showing both a thickened and more heterogeneous (larger spread of the data) glomerular basement membrane in Pax2-Piezo1-KO mice. Data are presented as individual GBM measurements in capillaries (at least 80 measurements from at least five different glomeruli from healthy Pax2-WT (n=2), Pax2-WT with adriamycin nephropathy (n=3), healthy Pax2-Piezo1-KO (n=2), and Pax2-Piezo1-KO with adriamycin nephropathy (n=4) mice. Mann–Whitney U test for (D) and (H).

Discussion

Podocytopathies are disorders of the glomerular epithelial compartment whose prognosis is determined by the balance between the nature, duration, and severity of podocyte injury and the efficiency of the repair response provided by podocyte progenitors.⁴⁴ Understanding the biochemical/biophysical stimuli and the molecular pathways involved in maintaining this balance is a fundamental research objective, with the potential to unveil new druggable targets for these pathologies that are major contributors to kidney failure. This study reveals the crucial role of the mechanosensitive channel Piezo1 in podocyte and podocyte progenitor physiology.

In vitro studies revealed that Piezo1 silencing altered the F-actin cytoskeleton, particularly the perinuclear actin assembly, causing defects in nuclear envelope morphology and integrity. Perinuclear actin is a contractile structure that provides physical continuity with the apical surface of the interphase nucleus⁴⁵ ensuring rapid transmission of mechanical signals from the plasma membrane to the chromatin.^46,47 The molecular basis of its formation is poorly understood, with some researchers suggesting a role of the Rho-GTPase proteins.⁴⁷ In our study, we provided evidence linking Piezo1 signaling and the Rho-GTPase pathway, particularly RhoA. Different contrasting roles have been reported for Rho-GTPases in glomerular disease pathogenesis, as either excessive or inadequate activity of Rho-GTPases can be detrimental to podocyte health.^48–51 Since Piezo1 regulates Rho-GTPases, its expression in glomerular cells is crucial for maintaining health, with both high and low levels, or timing of activation/repression, potentially influencing kidney disease pathogenesis, as seen in the cardiovascular system,⁵² and recently in Drosophila nephrocytes, the podocyte homologue in the fly.^53,54 These critical points could also explain the conflicting results on the role of Piezo1 in podocytes observed in different models.⁵⁵

Loss of nuclear envelope integrity and accumulation of DNA damage after Piezo1 silencing in podocytes and podocyte progenitors further supports its pathophysiologic role in kidney disease, similar to the pathogenic role for nuclear envelope rupture events resulting in DNA damage in response to mechanical stimuli described in muscle, heart, neurons, and mammary duct epithelial cells.^56–62 As consequence of DNA damage, Piezo1 silencing in podocytes induced morphologic changes characteristic of mitotic catastrophe^41,63–67 and increased susceptibility to cell death under mechanical stress conditions, such as shear stress, thus adding a new regulator to the growing list of mechanisms already reported to induce mitotic catastrophe in podocytes.^{16,40,67–71} Notably, pharmacologic activation of Rho-GTPase reduced the death of Piezo1-silenced podocytes under shear stress conditions, highlighting the intricate interplay between Piezo1, actin cytoskeleton, Rho-GTPase signaling, DNA repair processes, and podocyte survival. Piezo1 loss could induce mitotic catastrophe also for its role during mitosis. Indeed, mechanical forces are exerted during cytokinesis,⁷² and Piezo1 is required for an effective cytokinesis, its loss generating chromosomal instability and multinucleated cells.⁷³ In our study, Piezo1 loss in vitro induced mitotic catastrophe in podocytes but not in podocyte progenitors. In other epithelial cells, Piezo1 regulates cell numbers by controlling cell division and death,⁷⁴ with differences linked to mechanical force sensed,⁷⁴ cell geometry DNA repair efficiency, or to the cellular differentiation status.^75–77 Our observation of higher expression of DNA repair genes in progenitors versus podocytes may explain their reduced susceptibility to mitotic catastrophe after Piezo1 loss.

The generation of inducible podocyte-specific and in podocyte progenitor–specific KO mice models allowed us to reveal the role of Piezo1 after glomerular injury. Both KO groups showed a worsening of the albumin:creatinine ratio, but with differences in onset of proteinuria and its progression. Indeed, in the NPHS2-Piezo1-KO mice, we observed a peak at day 14 with a partially remitting course, while in the Pax2-Piezo1-KO mice, we observed a later onset of proteinuria (at 21 days) that did not decrease over 28 days. At day 28, no differences were found between the two models in podocyte count, foot process effacement, and GBM thickness, except for lower podocyte density in the Pax2-Piezo1-KO. We acknowledge that transmission electron microscopy is the gold standard for evaluating podocyte damage, and its absence may limit our study. However, the alternative method we used has proven effective in detecting filtration barrier alterations. Elucidating the precise mechanism by which loss of Piezo1 channel in podocytes and progenitors causes different phenotypes appears challenging, likely due to subtle differences between the two mouse strains in terms of podocyte loss, podocyte regeneration, and the “quality” of regenerated podocytes. In the absence of earlier time point analysis revealing the extent of podocyte loss after damage and longer time points assessment of kidney function, we can only speculate that two distinct events could account for the different phenotypes. The earlier onset of proteinuria in the NPHS2-Piezo1-KO mice may be due to higher susceptibility of KO podocytes. However, in the NPHS2-Piezo1-KO animals, podocyte loss could be partially compensated by the generation of new podocytes carrying “WT” Piezo1 from the progenitors. On the contrary, in the Pax2-Piezo1-KO mice, newly generated podocytes would be defective, failing to compensate for the damage (resulting in glomerular hypertrophy) and preferentially leading to the generation of hyperplastic lesions. The observation of a less arborized morphology of the podocytes generated in Pax2-Piezo1-KO mice supports this hypothesis, along with the more frequent presence of GFP+ fragmented structures inside the glomerular tuft, likely generated by the separation of the distal parts of the foot processes.

Our findings on the role of Piezo1 in podocytes conflict with recent studies.^55,78 However, as mentioned above, both activation and repression of Piezo1 affect podocytes. Moreover, discrepancies may arise from differences in the promoter for podocyte-specific KO (nephrin or podocin), substrain/genetic background of mice and related susceptibility to adriamycin nephropathy, age at induction, or type of podocytopathy (genetic, infectious, immunologic, or toxic), which warrant further investigation.

Evidence that targeting mechanosensing and mechanotransduction could ameliorate organ damage has been provided in the lung,⁷⁷ and treating glomerular diseases by targeting podocyte mechanobiology is an increasingly interesting topic as it could improve therapeutic precision and efficacy.¹⁷ Systemic targeting of Piezo1 or myeloid cell–specific deletion alleviated kidney fibrosis,^79,80 proposing Piezo1 as a target for the treatment of fibrotic diseases.⁸¹ Its targeting in glomerular diseases seems complicated by the fact that both Piezo1 depletion and activation/overexpression preserve podocyte morphology and function, making the stringent regulation of its expression level the therapeutic focus.^{29,53–55,78} However, the clinical relevance of Piezo1 in human diseases remains to be addressed, which represents a limitation of this study.

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/F140.

Funding

None.

Author Contributions

Conceptualization: Laura Lasagni, Maria Elena Melica, Anna Julie Peired.

Data curation: Giulia Antonelli, Laura Lasagni, Maria Elena Melica.

Formal analysis: Laura Lasagni.

Funding acquisition: Laura Lasagni.

Investigation: Maria Lucia Angelotti, Giulia Antonelli, Daniela Buonvicino, Giulia Carangelo, Giuseppina Comito, Carolina Conte, Tommaso Dafichi, Letizia De Chiara, Camilla Fantini, Gilda La Regina, Samuela Landini, Laura Maggi, Maria Elena Melica, Alice Molli, Valentina Raglianti, Roberto Semeraro.

Methodology: Maria Elena Melica.

Software: Benedetta Mazzinghi.

Supervision: Carolina Conte, Laura Lasagni, Anna Julie Peired, Paola Romagnani.

Writing – original draft: Laura Lasagni, Maria Elena Melica.

Writing – review & editing: Letizia De Chiara, Elena Lazzeri, Anna Julie Peired, Paola Romagnani.

Data Sharing Statement

Data related to transcriptomic, proteomic, or metabolomic data. Experimental Data. Gene Expression Omnibus. GSE273761, GSE273762, GSE291924.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/JSN/F141.

Supplemental Methods

Supplemental Figure 1. Piezo1 is expressed in human podocyte and podocyte progenitors.

Supplemental Figure 2. Piezo1 mediates Ca2⁺ flux in podocyte progenitors.

Supplemental Figure 3. Piezo1 regulates F-actin distribution through Rho GTPase.

Supplemental Figure 4. Piezo1 is expressed in mouse podocytes and podocyte progenitors.

Supplemental Figure 5. Generation of NPHS2-Piezo1-KO mice and adriamycin nephropathy induction.

Supplemental Figure 6. Loss of Piezo1 in NPHS2-Piezo1-KO mice results in worsening of proteinuria during 10 months of aging.

Supplemental Figure 7. Generation of Pax2-Piezo1-KO mice and adriamycin nephropathy induction.

Supplemental Figure 8. Uncropped images from Western blots.