Abstract
Background Podocytes are critical to maintaining the glomerular filtration barrier, and mutations in nephrotic syndrome genes are known to affect podocyte calcium signaling. However, the role of calcium signaling during podocyte development remains unknown.
Methods We undertook live imaging of calcium signaling in developing podocytes, using zebrafish larvae and human kidney organoids. To evaluate calcium signaling during development and in response to channel blockers and genetic defects, the calcium biosensor GCaMP6s was expressed in zebrafish podocytes. We used electron microscopy to evaluate filtration barrier formation in zebrafish, and Fluo-4 to detect calcium signals in differentiating podocytes in human kidney organoids.
Results Immature zebrafish podocytes (2.5 days postfertilization) generated calcium transients that correlated with interactions with forming glomerular capillaries. Calcium transients persisted until 4 days postfertilization, and were absent after glomerular barrier formation was complete. We detected similar calcium transients in maturing human organoid glomeruli, suggesting a conserved mechanism. In both models, inhibitors of SERCA or IP3 receptor calcium-release channels blocked calcium transients in podocytes, whereas lanthanum was ineffective, indicating the calcium source is from intracellular podocyte endoplasmic-reticulum stores. Calcium transients were not affected by blocking heartbeat or by blocking development of endothelium or endoderm, and they persisted in isolated glomeruli, suggesting podocyte-autonomous calcium release. Inhibition of expression of phospholipase C-γ1, but not nephrin or phospholipase C-ε1, led to significantly decreased calcium activity. Finally, blocking calcium release affected glomerular shape and podocyte foot process formation, supporting the critical role of calcium signaling in glomerular morphogenesis.
Conclusions These findings establish podocyte cell–autonomous calcium signaling as a prominent and evolutionarily conserved feature of podocyte differentiation and demonstrate its requirement for podocyte foot process formation.
Functioning of the glomerular filtration barrier is critically dependent on the morphology of podocyte foot processes and specialized slit diaphragm cell junctions.1,2 Podocyte disruption leads to foot process effacement, loss of slit diaphragms, and proteinuria, and is often the first step in progressive kidney disease.3⇓–5 Podocyte morphology and underlying cytoskeletal regulation in disease states is known to be exquisitely controlled via intracellular calcium levels. Notably, mutations in nephrotic syndrome genes, including transient receptor potential C6 (TRPC6)6⇓⇓–9 and phospholipase C-ε1 (PLCe1),10,11 affect podocyte calcium signaling and are required for normal podocyte structure. In addition, ectopic expression of the NPHS1 gene, encoding Nephrin, has been shown to activate PLCG, leading to intracellular calcium release.12 The TRP channels TRPC5 and TRPC6 are key calcium-influx pathways in podocytes.5⇓–7
Of the >40 genes implicated in monogenic causes of glomerular disease, many also play key roles in podocyte development.13 Recent high-resolution morphologic studies of podocyte differentiation reveal a sequential process of podocyte junctional migration, maturation, and basal foot process formation as the glomerulus forms.14 How these morphologic changes are regulated and the sequence of events linking genes required for podocyte function to morphologic differentiation is less well understood. Notably, although it is likely that calcium signaling and subsequent cytoskeletal remodeling are as important in initial podocyte differentiation as in disease states,5 it is currently not known how calcium signaling affects the initial formation of the glomerular filtration barrier. Assessment of podocyte calcium as a developmental signal has been hampered by the difficulty of studying glomerular development in vivo, where real-time imaging of mammalian embryonic glomeruli poses a significant challenge.
Glomerular development in the zebrafish pronephros, on the other hand, provides an optically and genetically accessible vertebrate model of early podocyte differentiation.15⇓⇓⇓⇓⇓–21 The larval pronephric glomerulus accurately replicates development of a size-discriminating glomerular filtration barrier and the effect of human glomerular disease gene mutations.2,11,22 In vitro modeling of human kidney development has been advanced with induced pluripotent stem cell (iPSc)–derived human kidney organoids. We and others have shown that kidney organoids represent a relevant model of the developing glomerulus at the anatomic and transcriptional level.23,24 Human kidney organoids express protein components associated with a maturing glomerular basement membrane (GBM) and show a transcriptional profile aligned with the mature human podocyte.23 We have also shown the presence of organoid podocyte clusters with evidence of slit diaphragm formation at the level of electron microscopy (EM), and the appropriate polarization and colocalization of slit diaphragm proteins.23 Several kidney-organoid differentiation protocols have successfully generated vascularized kidney organoids,23,25,26 illustrating their relevance to studies of glomerular morphogenesis. Here, we report live imaging of calcium signaling in developing podocytes during glomerular morphogenesis using both zebrafish larvae in vivo and human podocytes from kidney organoids in vitro. Our results reveal podocyte-autonomous calcium release from intracellular stores is a prominent and evolutionarily conserved feature of glomerular morphogenesis that is required for foot process formation.
Methods
Fish Care
Zebrafish (Danio rerio) adults and zebrafish larvae were maintained and raised according to established protocols.27 Fish lines used in this study are referenced in Table 1. All embryos and larvae were anesthetized in 0.02% tricaine methanesulfonate (MS 222; Sigma-Aldrich) and euthanized in 0.2% tricaine methanesulfonate. All studies adhere to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and were approved by the Massachusetts General Hospital Institutional Animal Care and use Committee.
Name of the zebrafish transgenic lines and mutants used in this study
Generation of Zebrafish Transgenic Lines
To generate the Tg(wt1b:Gal4ERT) zebrafish line expressing the endoxifen-inducible ERT domain fused to Gal4 in wt1b-expressing cells, the wt1b promoter was amplified from the pCRII-TOPO-zfwt1b plasmid (kind gift from the Englert laboratory) into a p5E vector, and then assembled into the final expression vector in a three-fragment gateway reaction (Invitrogen) using p5E-wt1b, pME-Gal4-ERT-VP16,32 p3E-ploy(A), and pDestTol2a-crystallin mCherry.
ERT Induction by Endoxifen Treatment
The induction of Tg(wt1b:Gal4ERT) expression was performed according to previous reports,32 with the difference that (E/Z)-endoxifen hydrochloride hydrate (E8284; Sigma) was used instead of 4-hydroxytamoxifen. Endoxifen was dissolved in ethanol at a stock concentration of 10 mM, and then added to fish water at 12 hours postfertilization (hpf) at a final concentration of 10 μM until the embryos were imaged (around 24 hpf).
Calcium Imaging of Zebrafish Larvae and Glomeruli
Embryos and larvae were staged according to Kimmel et al.,33 dechorionated manually, anesthetized in 0.02% tricaine methanesulfonate, and then mounted in glass-bottomed dishes filled with 1.5% low-melting point agarose. Calcium imaging was performed at 1 Hz with a Zeiss LSM 510 confocal microscope (calcium imaging; Carl Zeiss) equipped with a 40× water objective for about 8 minutes (500 seconds). Fiji and MATLAB were used for image analysis as previously described.34 Regions of interest were manually selected on the basis of an SD Z-projection. ΔF/F and normalized integral of activity were calculated with custom scripts written in MATLAB (available at https://github.com/lydiadjenoune/calcium_analysis_podocytes). The fluorescence change associated to GCAMP6s signals was defined as the Δ function ΔF/F. ΔF/F for each region of interest was calculated as ΔF/F=(F[t]−F0)/F0, where F0 is a manually selected baseline after correcting for photobleaching, and F(t) is the fluorescence intensity at a given time. Baseline was corrected by subtraction using a second-degree polynomial fitted to thresholded minima on the ΔF/F trace, and then integral ʃΔF/F was calculated on the basis of this corrected ΔF/F trace using a trapezoidal integral approximation. For imaging involving several experimental conditions, all solutions, unless otherwise noted, were bath applied. On all scatter dot plots, one dot represents one cell. On raw traces, pre- and post-traces represent the activity of a given cell before and after drug treatment. Note that, for all drug treatments, the activity of the exact same cells have been analyzed before and after drug application.
Generation of iPSC-Derived Kidney Organoids and Glomeruli Isolation
Kidney organoids were generated using the previously described iPSC MAFBmTagBFP2/+ reporter line35 by following our previously published directed-differentiation protocol.23,36 Briefly, iPSCs were cultured in a monolayer and differentiated into metanephric mesenchyme for 7 days. On day 7 (D7), cells were dissociated and bioprinted into three-dimensional structures to form an organoid.37 Organoids were then cultured for up to 19 days (D7+19) before being analyzed as intact organoids or processed using mechanical dissociation and sieving for isolation of glomeruli (isolated organoid glomeruli; OrgGloms) as previously described.23 Briefly, organoids were enzymatically and mechanically dissociated by gentle pipetting and incubation with TrypLE Select Enzyme (Thermo Fisher Scientific). The homogeneous cell suspension was then passed through a series of strainers to isolate intact glomerular cell aggregates.
Whole Mount Organoid Immunostaining
Fixed kidney organoids were immunostained as previously described36 with the following primary and secondary antibodies: LAMA5 (Abcam 1:300, cat# ab77175), Nephrin (NPHS1 1:300, Bioscientific, cat# AF4269), Claudin-1 (CLDN1 1:100, Thermo Fisher Scientific, cat# 71-7800), Alexa fluor 405 donkey anti-mouse (Abcam, cat# ab175659), Alexa fluor 488 donkey anti-goat (Molecular Probes, cat# A11055) and Alexa fluor 568 donkey anti-rabbit (Life Technologies, cat# A10042).
Calcium Imaging of Kidney Organoids and OrgGloms
Both kidney organoids and OrgGloms were incubated with calcium indicator Fluo-4, AM, cell permeant (Fluo-4; 5 µM; F14201; Thermo Fisher Scientific) diluted in Essential 6 Medium (05946; Stem Cell Technologies) for 1 hour and 10 minutes, respectively, at 37°C. OrgGloms were treated with either 100 µM lanthanum (La3+), 20 µM cyclopiazonic acid (CPA), 1 µM thapsigargin, and 30 µM 2-aminoethyl diphenylborinate (2-APB) or vehicle (DMSO) for 30 minutes before imaging. Calcium-flux recordings were acquired at 1 Hz using a Dragonfly Spinning Disc Confocal Microscope (Andor Technology) equipped with a 20× air objective for 2–4 minutes. Calcium-flux recordings were done under a controlled temperature of 37°C and 5% carbon dioxide. Regions of interest restricted to MAFBmTagBFP2/+-expressing podocytes were identified and image analysis was performed as described for zebrafish samples. Experiments were performed on three biologic replicates, where each biologic replicate is an independent directed-differentiation experiment. Statistical analysis was performed by one-way ANOVA with the Dunnett multiple comparison test.
Drug Treatments in Zebrafish
To test whether podocyte calcium transients were generated after a release from intracellular stores or from the extracellular compartment, we used the calcium channel blocker La3+ (100 µM; 298182; Sigma), the sarcoplasmic reticulum calcium-pump inhibitors CPA (20 µM; C1530; Sigma) and thapsigargin (1 µM; T9033; Sigma), and the IP3 receptor inhibitor 2-APB (30 µM; D9754; Sigma) on 3 days postfertilization (dpf) Tg(podocin:Gal4;UAS:GCaMP6s) zebrafish larvae. To test the contribution of the angiotensin pathway, we used the selective angiotensin AT1 receptor antagonist losartan (100 µM; 3798; Tocris). To test the contribution of vascular pressure, we used the heartbeat blocker amperozide (30 µM; AOB5283; Aobious). To induce podocyte injury, we injected puromycin aminonucleoside (PAN; 45 mg/ml; P7130; Sigma), which induces oxidant injury via the xanthine oxidase pathway, into the cardinal vein or pericardial sac of 5 dpf Tg(podocin:Gal4;UAS:GCaMP6s) larvae, and imaged them at 7 dpf.38⇓–40
Morpholino Injections
Morpholinos (Gene Tools) were resuspended in sterile water and stored at −20°C as 2 mM stock solutions. For injection, morpholino oligonucleotides were diluted in 0.1 M potassium chloride, 0.01M HEPES, and 0.2% phenol red, and injections were performed using a Nanoliter 2000 microinjector (World Precision Instruments). Tg(podocin:Gal4;UAS:GCaMP6s) zebrafish embryos were injected at the one-cell stage. Morpholinos used in this study are referenced in Table 2. Only morphants displaying the expected phenotype, as previously described,2,11,41⇓–43 were kept for imaging at 3 dpf and subsequent analysis, and validation of knockdown efficiency was determined by RT-PCR using previously published primers (see references in Table 2).
Name and sequences of morpholinos used in this study
Dissociation and Isolation of Zebrafish Glomeruli
To image calcium activity in podocytes of dissociated glomeruli, Tg(podocin:Gal4;UAS:GCaMP6s)-expressing larvae were first suspended in HBSS (10-527F; Lonza). Dithiothreitol (10 mM; BP172-5; Fisher Scientific) in HBSS was then added for 30 minutes at room temperature, with gentle mixing, before washing three times with HBSS. Larvae were then treated with 10 mg/ml of collagenase type 2 (Worthington) in HBSS, for 30 minutes at 28.5°C, and homogenized three times by pipetting. The reaction was stopped by adding HBSS in excess. Isolated glomeruli were then screened (manually selected) under a Leica fluorescent stereoscope and carefully mounted in 1.5% agarose covered in HBSS.
EM
Larvae were anesthetized in 0.02% tricaine methanesulfonate, and then euthanized in 0.2% tricaine methanesulfonate before fixation overnight at 4°C in a solution containing 1.5% glutaraldehyde, 1% paraformaldehyde, 70 mM sodium phosphate (pH 7.2), 3% sucrose, and 1% tannic acid. Preparation of the larvae for transmission EM was performed as previously described.15 After embedding in EMbed 812, semithin (300 nm) and ultrathin (90 nm) sectioning was performed with a 3× Ultracut E (Reichert-Jung; Nussloch, Germany), and sections were transferred onto copper slit grids (Electron Microscopy Sciences, Hatfield, PA). Grids were then stained with uranyl acetate and lead citrate, and imaged using a Morgagni electron microscope (FEI, Eindhoven, NL) operated at 80 Kv.15 Statistical comparisons were analyzed by Brown–Forsythe and Welch ANOVA tests with the Dunnett T3 multiple comparison test.
Statistical Analyses
t tests were used for the comparison of cell calcium activity before and after application of drug treatment or glomeruli processing, unless specified otherwise. The level of significance was P<0.05 for all datasets. P values are represented as follows: *P<0.05, **P<0.01, ***P<0.001, and ****P<0.0001. All quantifications are represented as medians with 95% confidence intervals. All graphs were prepared and statistical analyses performed using Prism8 (GraphPad).
Results
Zebrafish Podocytes Generate Calcium Transients during Glomerular Morphogenesis
To examine the dynamics of glomerular morphogenesis, we imaged the Tg(wt1b:GFP) and Tg(flk1:mCherry) zebrafish transgenic lines that label the anterior pronephros (green) and endothelial cells (red) (Figure 1, A and B). Live imaging between 30 and 72 hpf revealed early stages of the formation of the glomerular filtration barrier, when podocytes are motile, migrate toward the midline aorta, and interact with forming capillaries (Figure 1, C–G, Supplemental Video 1). To explore the role of calcium activity during glomerular morphogenesis, we performed live imaging of the genetically encoded calcium biosensor GCaMP6s in differentiating podocytes using the Tg(podocin:Gal4;UAS:GCaMP6s) or Tg(wt1b:Gal4ERT;UAS:GCaMP6s) (for time points before 36 hpf2) transgenic lines between 1 and 7 dpf (Figure 1, H–J). We observed that migrating podocytes did not exhibit significant calcium activity (from 1 to 2 dpf; Figure 1J). However, frequent intracellular calcium transients were detectable once midline convergence of podocyte primordia was complete (3 dpf; Figure 1, I and J, Supplemental Video 2). Podocyte cytosolic calcium transients showed a rapid spike and decay behavior and had an average duration of approximately 40 seconds (Supplemental Figure 1, A and C). Amplitude, represented in arbitrary units of ΔF/F, was relatively constant (Figure 1I, Supplemental Figure 1B). Podocytes exhibited calcium transients while they were differentiating (between 3 and 4 dpf; Figure 1J) and returned to a silent state at a stage when formation of a size-discriminating glomerular filtration barrier was complete (5–7 dpf; Figure 1J, Supplemental Video 2 2). Changes in integrated values of calcium signaling (Figure 1J) reflect changes in transient frequency at a relatively constant amplitude (Figure 1I, Supplemental Figure 1A). The results demonstrate a previously undetected calcium-signaling activity during the maturation of the glomerular filtration barrier in zebrafish.
Calcium signaling in zebrafish podocytes correlates with glomerular morphogenesis. (A) Lateral and (B) dorsal views of a Tg(wt1b:GFP;flk1:mCherry) zebrafish larva showing vasculature (red) and pronephric glomeruli (green, bracket) at 3 dpf. Scale bars, 250 µm. (C–G) Dorsal views of Tg(wt1b:GFP;flk1:mCherry) zebrafish embryos and larvae from (C) 30 hpf up to (G) 72 hpf, showing pronephric glomeruli (green) migrate toward the midline to reach the dorsal aorta and enwrap glomerular capillaries (red) during the differentiation of the glomerular filtration barrier. Scale bars, 20 µm. (H) Dorsal view showing the expression of GCaMP6s in podocytes in a Tg(podocin:Gal4;UAS:GCaMP6s) larva at 3 dpf. Example of region of interest selection (purple) for the analysis of podocyte calcium activity. (I) Calcium-imaging traces from podocytes in 3-dpf Tg(podocin:Gal4;UAS:GCaMP6s) larvae. Traces represent cells with median integral values taken from all analyzed cells. Horizontal scale bar, 60 s; vertical scale bar, 100% ΔF/F. (J) Integral quantification of zebrafish podocyte calcium activity from 1- to 7-dpf Tg(wt1b:Gal4ERT;UAS:GCaMP6s) (at 1 dpf) and Tg(podocin:Gal4;UAS:GCaMP6s) zebrafish embryos and larvae showing increased podocyte calcium activity during foot processes differentiation and interaction with glomerular capillaries (3 and 4 dpf). Data from 58 cells from seven embryos for 1 dpf, 81 cells from seven embryos for 2 dpf, 163 cells from 12 larvae for 3 dpf, 91 cells from eight larvae for 4 dpf, 98 cells from eight larvae for 5 dpf, 103 cells from eight larvae for 6 dpf, and 105 cells from 13 larvae for 7 dpf. Data from cells from different animals were pooled and statistics were done per cell.
Differentiating Podocytes Display Calcium Transients in Human Kidney Organoids
To determine whether calcium signaling during podocyte maturation was a shared feature among vertebrates, we investigated calcium dynamics within glomeruli forming in iPSC-derived human kidney organoids (Figure 2). Kidney organoids were generated from a MAFBmTagBFP2/+ iPSC reporter line,35 resulting in fluorescent labeling of podocytes (blue) as they arose in the organoids. This facilitated definitive identification of glomeruli within organoids, allowing us to image their development in real time (Figure 2B). MAFB+ podocytes could be identified as early as 10 days after nephrogenesis induction (D7+10), with these maturing across the subsequent 9 days to form defined glomeruli, characterized by polarized NPHS1+ podocytes sitting on a GBM, and surrounded by CLDN1+ parietal epithelial cells (Figure 2C). Calcium activity was imaged by labeling kidney organoids (Figure 2D) or OrgGloms (Figure 2E) with the calcium indicator Fluo-4 at two stages of organoid culture termed “early” and “late” time points (D7+12/14 and D7+18/19, respectively; Figure 2A). Interestingly, podocyte calcium activity was limited in early podocytes, whereas it was significantly increased in podocytes imaged at the late time point (Figure 2H). Of note, the diffusion of the Fluo-4 calcium probe was improved in OrgGloms (Supplemental Figure 2, A and B). Generally, we were able to demonstrate reliable and reproducible calcium transients in maturing podocytes imaged in intact kidney organoids and OrgGloms (Figure 2, F and G, respectively, Supplemental Video 3) at the late time point. These results suggest increased calcium signaling is a conserved feature of glomerular morphogenesis.
Human podocytes display calcium transients when differentiating in kidney organoids. (A) Schematic diagram of the protocol to generate kidney organoids from iPSCs. (B) Example of a MAFBmTagBFP2/+ organoid cultured over time, showing an increasing number of glomeruli and MAFB-BFP expression. Scale bars, 100 μm. (C) Kidney organoid stained for GBM (LAMA5, blue), parietal epithelial cells (CLDN1, green), and podocytes (NPHS1, white and red) showing glomeruli. Scale bars, 50 µm. (D) Intact kidney organoid and (E) OrgGloms labeled with the calcium probe Fluo-4. Scale bars, 50 µm. (F and G) Example of calcium transients imaged in (F) intact organoid and (G) OrgGloms. Scale bars, 25 µm. Arrowheads indicate pulsing cells. (H) Sample traces of calcium activity in podocytes from organoids at early (D7+12/14, yellow) and late (D7+18/19, orange) time points. Traces represent cells with the top 10 integral values taken from all analyzed cells. Horizontal scale bar, 25 s; vertical scale bar, 100% ΔF/F. Integral quantification of human organoid podocyte calcium activity shows differentiating podocytes (D7+18/19, 178 cells from 16 glomeruli) are highly active when immature ones (D7+12/14, 175 cells from 27 glomeruli) are not. Data were acquired from three independent directed-differentiation experiments. ****P<0.0001.
Podocyte Calcium Transients Are Generated by Release from Intracellular Stores via IP3 Receptors and Not Extracellular Calcium Influx
To identify pathways underlying podocyte calcium activity during glomerular morphogenesis, we investigated the contribution of the vasoactive hormone angiotensin II (Ang II). Indeed, an increase of cytosolic calcium activity by Ang II has been observed in rat podocytes in culture.44 However, in our system, the Ang-II receptor blocker losartan did not have a significant effect on differentiating podocyte calcium activity (Figure 3A), suggesting developing podocytes rely on a different calcium-signaling pathway. Because membrane calcium channels, including TRPC5/6, have been identified in podocytes,5⇓–7 we next tested the contribution of extracellular calcium. We treated zebrafish embryos with the calcium channel blocker La3+ and did not observe a significant reduction of calcium activity in zebrafish nor in human organoid developing podocytes (Figure 3, B and D, respectively, Supplemental Videos 4 and 5). Although TRPC5 has been reported to be stimulated by La3+ and affects podocyte structure in adult mice under pathologic conditions,45 expression of this channel is not observed in developing zebrafish glomeruli46 or enriched in human iPSc-derived OrgGloms.23 To determine whether calcium was released from intracellular stores, we used the SERCA pump inhibitors, CPA and thapsigargin, and the IP3 receptor blocker, 2-APB. Using these compounds, we observed a significant reduction of calcium activity in both zebrafish and human organoid podocytes (Figure 3, C and D, respectively, Supplemental Videos 4 and 5). These data suggest the calcium activity observed during both human and zebrafish podocyte differentiation relies on release from intracellular stores via IP3 receptors and is not mediated by cell membrane calcium channels.
Podocyte calcium during transients’ glomerular morphogenesis is generated by release of intracellular calcium stores via IP3 receptors. (A) Calcium-imaging traces from podocytes in 3-dpf Tg(podocin:Gal4;UAS:GCaMP6s) larvae before (gray) and after (brown) 40 minutes of losartan treatment (20 µM). The “pre” traces represent cells with top integral values taken from all analyzed cells. The “post” traces represent the same cells after losartan treatment. Integral quantification of zebrafish podocyte calcium activity before and after 40 minutes of DMSO (0.1%, 31 cells from three fish) and losartan (86 cells from nine fish) treatment. (B) Podocyte calcium-imaging traces and integral quantification of zebrafish podocyte calcium activity before (gray) and after (blue) 50 minutes of La3+ treatment (100 µM, 72 cells from seven fish). The “post” traces represent cells with top integral values taken from all analyzed cells. The “pre” traces represent the same cells before La3+ treatment. (C) Calcium-imaging traces from podocytes before (gray) and after 40 minutes of CPA (20 µM, light green, 46 cells from four fish), thapsigargin (1 µM, green, 106 cells from eight fish), and 2-APB (30 µM, dark green, 49 cells from four fish) treatments. The “pre” traces represent cells with top integral values taken from all analyzed cells. The “post” traces represent the same cells after drug treatment. Integral quantification of zebrafish podocyte calcium activity before and after 40 minutes of DMSO (0.1%, 91 cells from seven fish; and 0.06%, 45 cells from four fish) and drug treatments. (D) Calcium-imaging traces from Fluo-4–labeled podocytes in D7+18 isolated OrgGloms untreated (light gray, 226 cells from 41 glomeruli) and after application of DMSO (0.001%, light gray, 237 cells from 52 glomeruli), La3+ (100 µM, blue, 252 cells from 60 glomeruli), CPA (20 µM, light green, 238 cells from 41 glomeruli), thapsigargin (1 µM, green, 270 cells from 47 glomeruli), and 2-APB (30 µM, dark green, 238 cells from 52 glomeruli) for 30 minutes. Traces represent cells with top integral values taken from all analyzed cells. Integral quantification of organoid podocyte calcium activity after application of calcium inhibitors. Statistical analysis was performed by one-way ANOVA with the Dunnett multiple comparison test. Data were acquired from three independent directed-differentiation experiments. **P<0.01, ****P<0.0001.
Kidney Podocytes Generate Autonomous Calcium Transients via Plcg1
To further explore the signaling pathways triggering calcium release in developing podocytes, we tested whether mechanical stretch of podocytes induced by BP could cause calcium transients. We blocked heartbeat in zebrafish by morpholino knockdown of troponin T type 2 (tnnt2, also called silent heart) and pharmacologically using the small molecule amperozide (Figure 4, A and B). Neither of these treatments caused a significant reduction in calcium transient activity, suggesting podocyte developmental calcium transients are not induced by vascular pressure. Because podocytes are in direct contact with endothelial cells, we tested whether signals from the endothelium may be required to initiate podocyte calcium signals by examining the cloche mutant, in which endothelial cell development is blocked, resulting in mutant embryos devoid of endothelial cells.31 Calcium transients were still observed in cloche mutants, despite failed podocyte midline migration (Figure 4C, Supplemental Video 6).18 We tested the contribution of endoderm signaling by blocking all endoderm development using sox32 morpholino injection. No reduction of podocyte calcium activity was observed in embryos lacking endoderm (Figure 4D). These results indicate podocyte calcium transients are not induced by cues from either endothelial cells or the endoderm during glomerular differentiation, and suggest an autocrine source for the stimulus leading to intracellular calcium release.
Podocytes generate autonomous calcium transients via the phospholipase C Plcg1. (A) Calcium-imaging traces from podocytes in noninjected (NI, gray) and tnnt2 morphants (light brown). Traces represent cells with top integral values taken from all analyzed cells. Integral quantification of zebrafish podocyte calcium activity in tnnt2 morphants (69 cells from eight morphants) and noninjected siblings (62 cells from five fish). (B) Podocyte calcium-imaging traces before (gray) and after heartbeats were blocked by amperozide treatment (20 µM, dark brown). The “post” traces represent cells with top integral values taken from all analyzed cells. The “pre” traces represent the same cells before amperozide treatment. Integral quantification of zebrafish podocyte calcium activity before and after amperozide treatment (71 cells from six fish). Data from 59 cells from five larvae for DMSO (0.1%). (C) Calcium-imaging traces from podocytes in wild type (WT; cloche +/+ and cloche +/− siblings, gray, 58 cells from four fish) and cloche −/− mutants (dark blue, 81 cells from six fish). Traces represent cells with top integral values taken from all analyzed cells. Integral quantification of zebrafish podocyte calcium activity in cloche mutants and WT siblings. (D) Calcium-imaging traces from podocytes in foxD1 (used as control, gray) and sox32 morphants (light blue). Traces represent cells with median integral values taken from all analyzed cells. Integral quantification of zebrafish podocyte calcium activity in foxD1 (15 cells from three larvae) and sox32 morphants (30 cells from four larvae). (E) Calcium-imaging traces from podocytes in vivo (gray) and from dissociated glomeruli (orange) from 3-dpf (dark orange) and 7-dpf (light orange) larvae. Traces represent cells with top integral values taken from all analyzed cells. Integral quantification of zebrafish podocyte calcium activity from in vivo (165 cells from 12 larvae at 3 dpf, and 105 cells from eight larvae at 7 dpf) and dissociated glomeruli (158 cells from 11 glomeruli at 3 dpf, and 89 cells from eight glomeruli at 7 dpf). (F) Calcium-imaging traces from podocytes in noninjected (gray), nephrin (purple), plce1 (jam), and plcg1 (magenta) morphants. Traces represent cells with median integral values taken from all analyzed cells. Integral quantification of zebrafish podocyte calcium activity in these morphants. Data from 105 cells from 10 noninjected larvae, 87 cells from nine nephrin morphants, 48 cells from five plce1 morphants, and 102 cells from 10 plcg1 morphants. ****P<0.0001.
To determine whether podocytes generate calcium signals autonomously, we dissociated glomeruli and plated them for imaging in HBSS. Even under dissociated conditions with no contact with adjacent tissue, freshly isolated zebrafish podocytes still exhibited calcium transients (Figure 4E, Supplemental Video 7). Consistent with what we observed in vivo, podocytes isolated from older, 7-dpf larvae were silent ex vivo with respect to calcium transients, indicating calcium signaling is developmentally regulated (Figure 4E, Supplemental Video 7). This also indicates that calcium transients observed in 3-dpf isolated glomeruli were not due to injury incurred by cell dissociation. Similarly, human kidney-organoid podocyte calcium transients were not affected by glomerular isolation. Neither average amplitude nor average duration of transients were markedly affected in zebrafish or organoid podocytes by isolation (Figure 2, Supplemental Figure 1, B and C). Note the La3+ treatments of dissociated podocytes in zebrafish did not affect their calcium activity, further confirming the intracellular release as the source of calcium (Supplemental Figure 1D).
IP3-regulated calcium release from intracellular stores relies on the activity of PLC and the generation of IP3 from phosphatidylinositol 4,5-bisphosphate.47,48 Two PLC enzymes have been shown to have a role in podocyte biology: PLCe, in which mutations cause proteinuric kidney disease in humans,11 and PLCg, a broadly expressed PLC isotype.49⇓⇓–52 In addition, phosphorylation of the slit diaphragm cell adhesion molecule Nephrin has been shown to stimulate calcium release when ectopically expressed.12 To test the involvement of these candidate regulators, we transiently knocked them down with antisense morpholino oligos and characterized the resulting calcium activity. Whereas knockdown of PLCe and Nephrin had no significant effect on calcium activity, knockdown of PLCg strongly inhibited calcium release in podocytes (Figure 4F, Supplemental Video 8). Given reports of ectopically expressed Nephrin causing calcium release,12 we confirmed that, in embryos completely lacking wild-type nephrin mRNA, calcium transients were still present (Supplemental Figure 3). Our results suggest podocytes signal cell autonomously, either by cell contact or a locally produced ligand, to stimulate PLCg and IP3 production, leading to transient release of calcium from intracellular stores.
Podocyte Calcium Signaling Is Required for Foot Process Formation
Podocyte calcium transients correlate developmentally with the formation of the glomerular capillary loop and formation of the filtration barrier. The glomerular filtration barrier is defined by podocyte foot process interdigitation and establishment of specialized cell adhesions, the slit diaphragms. Final maturation of the filtration barrier in zebrafish occurs between 3.5 and 4.5 dpf.15 We tested whether calcium transients were required for formation of the filtration barrier by inhibiting calcium release in zebrafish larvae and examining ultrastructure of the filtration barrier by EM. To avoid affecting the entire embryo, we established low doses of the SERCA pump inhibitors CPA or thapsigargin, which blocked podocyte calcium release but did not affect heartbeat. Control larvae treated with DMSO exhibited normal foot process and filtration barrier formation, with podocytes adhering to the midline aorta (Figure 5, A and B). In contrast, foot process formation in larvae treated with 20 µM CPA or 1 µM thapsigargin from 3.5 to 4.5 dpf was arrested or inhibited, with no evidence of morphologically normal foot processes and the appearance of large, effaced podocyte processes (Figure 5B). Quantification of cellular processes in contact with the basement membrane revealed a significant decrease in the number of defined cell processes per micron GBM in the presence of endoplasmic reticulum calcium-pump inhibitors, as compared with control (Figure 5C). No loss of cells during treatment with endoplasmic reticulum calcium-pump inhibitors was observed (Figure 5D). The results suggest autonomous calcium signaling is required for morphologic maturation of the filtration barrier and podocyte foot process formation.
Blocking podocyte calcium release affects glomerular shape and foot process formation. (A) Dorsal views of 3-dpf Tg(wt1b:GFP;flk1:mCherry) zebrafish larvae after treatment with DMSO (0.1%), thapsigargin (1 µM), and CPA (20 µM) between 60 and 72 hpf show that blocking calcium release affects glomerular shape. (B) Transmission EM of larvae treated with thapsigargin and CPA between 3.5 and 4.5 dpf shows podocyte developmental calcium signaling is critical for podocyte foot process differentiation. Red overlay, podocyte; yellow overlay, GBM. Scale bars, 1 µm. (C) Quantification of the number of podocyte foot processes per micrometer of GBM per section (DMSO, two sections, 46.4 µm GBM; thapsigargin, five sections, 42.8 µm GBM; CPA, seven sections, 41.7 µm GBM) shows that the number of foot processes is greatly reduced when calcium release is impaired. Statistical comparison was analyzed by Brown–Forsythe and Welch ANOVA tests with the Dunnett T3 multiple comparison test. (D) Quantification of the total number of podocytes per larva after thapsigargin and CPA treatments between 60 and 72 hpf. Data from three larvae for each condition. Statistical comparison was analyzed by Brown–Forsythe and Welch ANOVA tests with the Dunnett T3 multiple comparison test. (E) Podocyte calcium-imaging traces from 7-dpf Tg(podocin:Gal4;UAS:GCaMP6s) larvae 42 hours postinjection of water (gray) or PAN (45 ml/ml, red). Traces represent cells with median integral values taken from all analyzed cells. Integral quantification of zebrafish podocyte calcium activity from PAN-injected larvae shows calcium transients are reinitiated in mature podocytes after injury. Data from 64 cells from six water-injected larvae, and 70 cells from six PAN-injected larvae. *P<0.05, **P<0.01, ****P<0.0001.
Podocyte Injury Reinitiates Calcium Transients in Mature Glomeruli
Calcium signaling in podocytes is associated with a highly dynamic stage of foot process interdigitation and cytoskeletal rearrangements that were blocked by calcium depletion. At later stages, when the filtration barrier functionally discriminates blood solute on the basis of molecular size, calcium transients (Figure 1J) and active cell rearrangements cease.53 After podocyte injury in vivo, foot processes retract and efface, leading to protein leakage into the filtrate or proteinuria. To determine if podocyte injury correlated with reinitiation of calcium transients, we treated 7-dpf zebrafish larvae with PAN38⇓–40,54 and assayed calcium activity. PAN treatment resulted in a significant increase in calcium transients (Figure 5E, Supplemental Video 9), suggesting calcium signaling may be required in dynamic states of cell rearrangement, both developmentally and as a component of an injury response.
Discussion
In this study, we report calcium signaling as an essential and evolutionarily conserved feature of developing podocytes during glomerular morphogenesis. Importantly, we show that this activity is podocyte autonomous, critical for proper foot process differentiation, and conserved between zebrafish and human organoid kidney glomeruli. On the basis of these findings, we propose that autonomously regulated, PLCg/IP3-mediated calcium release from intracellular stores triggers essential processes leading to podocyte foot process differentiation and formation of the glomerular filtration barrier.
Our conclusion that podocyte calcium transients contribute to the differentiation of podocyte foot processes rests on two primary observations. Calcium-signaling activity peaks right before zebrafish podocyte terminal differentiation, establishing a temporal link with the formation of interdigitating foot processes between 3.5 and 4.5 dpf.2 In addition, in mature podocytes, calcium transients can be observed only after injury, when effacement and remodeling of foot processes occurs, further suggesting that this activity underlies signaling pathways leading to foot process reorganization. Importantly, blocking calcium release during podocyte terminal differentiation impairs foot process formation, leading to an effaced, disorganized cell-projection phenotype. Podocyte calcium transients are unlikely to contribute to midline cell migration because this was not blocked by SERCA inhibitors (data not shown) and because injured podocytes, which actively signal with calcium, do not migrate.55
Human organoid glomeruli represent a unique three-dimensional model of the human glomerulus with respect to appropriate gene and protein expression.23,24 Podocytes within organoid glomeruli show polarization and slit diaphragm and foot process formation.23 The accuracy with which they model the human developing glomerulus is evidenced by their ability to replicate the glomerular defects present in several inherited forms of congenital nephrotic syndrome, including NPHS1 (NPHS1 mutation)23,56 and NOS1AP-associated nephrotic syndrome,57 and reflect nephrotoxic responses, including PAN nephrosis58 and gentamycin treatment.23 However, it is evident that these remain immature with respect to GBM maturation, with the majority of organoid glomeruli lacking glomerular capillaries.23 Data from the zebrafish would predict that the latter is not a barrier to slit diaphragm formation, hence organoid glomeruli prove to be a reliable model for studying the role of calcium in cytoskeletal dynamics during human glomerular maturation.
The TRP channels TRPC5 and TRPC6 have been identified as calcium-influx pathways mediating calcium ion conductance and cytoskeletal changes in podocytes.5⇓–7 Nonetheless, we found that cell membrane calcium channels do not contribute to developmental podocyte calcium signaling, indicating podocytes rely on different calcium-release pathways in early stages of differentiation. The finding that cell membrane calcium channels do not participate in early calcium signals is consistent with reports that most patients with gain-of-function TRPC6 mutations only show a late onset of FSGS,6 well after kidney development is complete. Also, TRPC6 knockout in mice does not lead to early glomerular morphogenesis defects or to changes in urinary albumin excretion/proteinuria.59 Similarly, overexpression or activation of TRPC5 does not cause kidney barrier injury or aggravate podocyte injury under pathologic conditions.60 These results suggest TRPC6 and TRPC5 act later in adult kidney function but not in differentiating podocytes.
Podocyte calcium transients were generated by release from intracellular stores via IP3 receptors in both zebrafish and human kidney organoids. This type of calcium release relies on the activity of PLC and the generation of IP3 from phosphatidylinositol 4,5-bisphosphate.47,48 Perhaps surprisingly, our zebrafish-knockdown screen revealed PLCg1, and not the FSGS-associated PLCe1,11 as the key PLC isotype involved. PLCg is broadly expressed49⇓⇓–52 and has been linked to cytoskeletal remodeling.50 Mouse PLCg1 chimeric mutants exhibit multicystic renal dysplasia with sclerotic glomeruli,61 consistent with a role in glomerular maturation or function. Diacylglycerol, the other major product of PLCg1, activates protein kinase C and promotes Nephrin removal from podocyte cell membrane,62 suggesting dynamic turnover of slit diaphragm proteins may be a feature of podocyte foot process formation. Nephrin phosphorylation has been reported to trigger calcium signaling by recruitment and activation of PLCg.12 However, our results show that Nephrin was not required for in vivo podocyte calcium activity, suggesting other podocyte proteins, for instance the tyrosine kinase substrates Neph proteins,63 may compensate for Nephrin deficiency, as they are proposed to do in chicken that lack a Nephrin gene.64
Podocyte calcium release during zebrafish glomerular morphogenesis was autonomous in isolated groups of podocytes in vitro and did not require signals from neighboring endothelial cells or the underlying endoderm in vivo. This conclusion is based, in part, on the phenotype of the zebrafish mutant cloche, where mutations in the neuronal PAS domain-protein 4-like transcription factor prevent differentiation of all endothelial and hematopoietic cell lineages.65,66 The persistence of calcium signaling in cloche mutant podocytes, and the requirement for calcium signaling in foot process formation we observed in SERCA inhibitor experiments, is consistent with EM findings of podocyte foot process formation in cloche mutants, despite the complete lack of endothelial cells.18 Similarly, in human kidney organoids, nearly all clusters of MAFBmTagBFP2/+-positive podocytes in OrgGloms showed calcium signaling, although the presence of endothelial cells in OrgGloms was rare.23 Although we cannot definitively rule out the presence of endothelial cells, most OrgGloms are avascular, supporting observations in zebrafish that endothelial cells are not required for podocyte calcium transients. More broadly, however, glomerular morphogenesis in zebrafish and maintenance of the structure and selective permeability of the glomerular filtration barrier in mammals requires a complex interplay of endothelial, mesangial, and podocyte cell signals.15,20,67⇓⇓–70 Nonetheless, developmental podocyte calcium signaling and initial formation of foot processes appears to be podocyte cell autonomous.
Calcium signaling was arrested as zebrafish glomeruli matured, suggesting either the loss of an activating signal or active suppression by a glomerular maturation signal. Cessation of calcium signaling in mature glomeruli is required for stabilization of glomerular function, as suggested by the occurrence of FSGS caused by gain-of-function TRPC6 mutations6 and the reappearance of calcium transients in injured podocytes (Figure 5E).71,72 We found that, similar to the initiation of developmental calcium transients, cessation of calcium transients occurred independently of endothelial cells, because 5- and 6-dpf cloche mutants were silent with respect to calcium signaling, similar to wild-type embryos of the same age (data not shown). Taken together with our finding that podocytes from 7-dpf zebrafish maintained ex vivo did not display calcium transients, these data indicate both initiation and cessation of calcium signaling may be mediated by podocyte-autonomous mechanisms. Further studies of autonomous cell-signaling systems, such as Neph slit diaphragm protein interactions or autocrine secreted signals, may reveal podocyte-specific inducers of intracellular calcium release. Alternatively, podocytes may signal completely autonomously by calcium-induced calcium release, as occurs in the renal microvasculature.73
Podocyte injury leads to foot process effacement that involves rearrangements of the actin cytoskeleton.4,74⇓⇓⇓⇓–79 Consistent with our results, other studies have reported an increase of intracellular calcium after podocyte injury,71,72,80,81 which is consistent with the role of calcium in actin cytoskeleton remodeling82⇓–84 and foot process effacement. Although short-duration calcium transients may not be sufficient to cause proteinuria,85 stabilizing podocyte intracellular calcium in the long term, and in the context of nephrotic syndrome, may be critical to prevent further glomerular damage.86 Our study establishes the larval zebrafish and human kidney organoids as models for glomerular diseases caused by impaired calcium signaling, and as a screening platform to identify new substances to modulate podocyte morphogenesis in glomerular disorders.
Disclosures
I.A. Drummond reports having ownership interest in Abbott Laboratories, CRISPR Therapeutics, and Editas; being a scientific advisor for, or member of, Imagine Institute Paris; and receiving honoraria from the NIH. A. Hidalgo-Gonzalez reports having ownership interest in AstraZeneca. M. Little reports serving as an editorial board member of Cell Stem Cell, Development, Developmental Biology, JASN, and Kidney International; serving as the president elect of the International Society for Stem Cell Research; having patents and inventions with Murdoch Children’s Research Institute and Uniquest Pty Ltd.; serving as an advisory board member for Nature Reviews Nephrology; and receiving research funding from Novo Nordisk. All remaining authors have nothing to disclose.
Funding
L. Djenoune, R. Tomar, and I. A. Drummond were supported by National Institute of Diabetes and Digestive and Kidney Diseases grants R01DK053093, UH3DK107372, UC2DK126021, and R01DK071041. M. Little is a National Health and Medical Research Council Senior Principal Research Fellow, through grant GNT1136085. This work was supported by the National Health and Medical Research Council grant GNT1098654, the Royal Children’s Hospital Foundation, and the Stafford Fox Medical Research Foundation.
Acknowledgments
The authors thank Dr. Christophe Englert (Leibniz Institute on Aging, Jena, Germany) for the zebrafish wt1b promoter, Dr. Urs Lucas Böhm (Harvard University) for providing the Tg(UAS:GCaMP6s) zebrafish, Dr. Alex Vasilyev for information on use of amperozide, and Dr. Hermann Haller (Hannover Medical School) for input on experiments.
L. Djenoune and R. Tomar performed calcium-imaging experiments in zebrafish; I. Ghobrial and A. Dorison generated kidney organoids and isolated glomeruli; A. Dorison and A. Hidalgo-Gonzalez acquired the calcium-imaging human kidney organoid dataset; H. Schenk, J. Hegermann, and L. Beverly-Staggs performed EM analysis; L. Djenoune analyzed the data; L. Djenoune, R. Tomar, A. Dorison, M. Little, A. Hidalgo-Gonzalez, and I. A. Drummond designed the experiments; M. Little and I. A. Drummond supervised the research; and L. Djenoune and I. A. Drummond wrote the manuscript, with input from all authors.
Supplemental Material
This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2020101525/-/DCSupplemental.
Supplemental Figure 1. Podocytes generate calcium transients during glomerular morphogenesis of similar features in both intact and dissociated conditions (related to Figures 1⇑–3).
Supplemental Figure 2. Human kidney organoids enable to investigate calcium dynamics in developing human podocytes during glomerular morphogenesis (related to Figure 2).
Supplemental Figure 3. nephrin and plcg1 morpholinos induce cardiac edema and a knock-down of their respective transcripts (related to Figure 4).
Supplemental Video 1. Migration of the pronephric glomeruli (labeled by the Tg(wt1b:GFP) transgenic line, green) toward the midline to reach the dorsal aorta (visualized by the Tg(flk1:mCherry) transgenic line, red) in zebrafish between 31 and 38 hpf, Related to Figure 1. The scale bar represents 20 μm.
Supplemental Video 2. Differentiating podocytes in zebrafish larvae exhibit calcium transients (3 dpf) while mature one are silent (7 dpf), related to Figure 1. Activity recorded at 1 Hz, displayed here at 80 frames per second (fps). The scale bar represents 10 μm.
Supplemental Video 3. Differentiating podocytes (D7+18) in intact human kidney organoids are highly active in regard to their calcium activity while differentiating ones (D7+13) are not, Related to Figure 2. Activity recorded at 1 Hz, displayed here at 80 fps. The scale bar represents 20 μm.
Supplemental Video 4. Calcium transients in differentiating podocytes in zebrafish are generated via release from intracellular reticulum endoplasmic stores, Related to Figure 3. Activity recorded at 1 Hz, displayed here at 80 fps. The scale bar represents 10 μm.
Supplemental Video 5. Human podocyte calcium activity during glomerular morphogenesis is generated by release from intracellular stores via IP3 receptors, Related to Figure 3. Activity recorded at 1 Hz, displayed here at 80 fps. The scale bar represents 40 μm.
Supplemental Video 6. Endothelial cells do not contribute to podocyte calcium activity during their differentiation, Related to Figure 4. Activity recorded at 1 Hz, displayed here at 80 fps. The scale bar represents 10 μm.
Supplemental Video 7. Podocytes generate autonomous calcium transients while they differentiate, Related to Figure 4. Activity recorded at 1 Hz, displayed here at 80 fps. The scale bar represents 10 μm.
Supplemental Video 8. Podocyte developmental calcium activity is sustained via the phospholipase C isoform PLCg1, Related to Figure 4. Activity recorded at 1 Hz, displayed here at 80 fps. The scale bar represents 10 μm.
Supplemental Video 9. Podocyte injury re-initiates calcium transients in mature glomeruli, Related to Figure 5. Activity recorded at 1 Hz, displayed here at 80 fps. The scale bar represents 10 μm.
Footnotes
Published online ahead of print. Publication date available at www.jasn.org.
- Copyright © 2021 by the American Society of Nephrology
References
In this issue
Jump to section
More in this TOC Section
Cited By...
- No citing articles found.