Loss of Epithelial Membrane Protein 2 Aggravates Podocyte Injury via Upregulation of Caveolin-1

Abstract

Nephrotic syndrome is a CKD defined by proteinuria with subsequent hypoalbuminemia, hyperlipidemia, and edema caused by impaired renal glomerular filtration barrier function. We previously identified mutations in epithelial membrane protein 2 (EMP2) as a monogenic cause of this disease. Here, we generated an emp2-knockout zebrafish model using transcription activator-like effector nuclease–based genome editing. We found that loss of emp2 in zebrafish upregulated caveolin-1 (cav1), a major component of caveolae, in embryos and adult mesonephric glomeruli and exacerbated podocyte injury. This phenotype was partially rescued by glucocorticoids. Furthermore, overexpression of cav1 in zebrafish podocytes was sufficient to induce the same phenotype observed in emp2 homozygous mutants, which was also treatable with glucocorticoids. Similarly, knockdown of EMP2 in cultured human podocytes resulted in increased CAV1 expression and decreased podocyte survival in the presence of puromycin aminonucleoside, whereas glucocorticoid treatment ameliorated this phenotype. Taken together, we have established excessive CAV1 as a mediator of the predisposition to podocyte injury because of loss of EMP2, suggesting CAV1 could be a novel therapeutic target in nephrotic syndrome and podocyte injury.

Nephrotic syndrome (NS) is a CKD defined by nephrotic-range proteinuria caused by disruption of the renal glomerular filtration barrier, resulting in hypoalbuminemia, hyperlipidemia, and edema. Childhood NS, if untreated, is associated with increased risk of life-threatening infections, thromboembolism, lipid abnormalities, and malnutrition.¹ NS is classified by response or nonresponse to a standardized glucocorticosteroid therapy as steroid-sensitive nephrotic syndrome (SSNS) and steroid-resistant nephrotic syndrome (SRNS), respectively. Recent studies of SRNS have identified monogenic causes of SRNS to reside in podocyte-specific genes, which if mutated, lead to podocyte dysfunction and defective glomerular filtration barrier.²

Using the combined approach of homozygosity mapping and whole-exome sequencing, we have previously identified mutations in epithelial membrane protein 2 (EMP2) as a novel monogenic cause for NS.³ EMP2 was initially identified by its homology with the growth-arrest-specific 3/peripheral myelin protein 22 family⁴ and negatively regulates the mRNA and protein levels of CAV1, a major protein component of plasma membrane microdomain caveolae.^5,6

To explore the pathogenic mechanism underlying EMP2 mutations, we have now generated emp2 deletion mutant zebrafish and caveolin 1 (cav1)-overexpressing transgenic zebrafish. Using an inducible podocyte injury model and a transgene-based proteinuria assay that we previously established,⁷ we demonstrated that either homozygous deletion of emp2 or podocyte-specific overexpression of cav1 predisposed zebrafish podocytes to injury, which was ameliorated by glucocorticoid treatment. Consistently, homozygous deletion of emp2 in zebrafish upregulated cav1 transcript level, and glucocorticoids could downregulate cav1 in cultured human podocytes and zebrafish, reducing podocyte injury. Therefore, our data provide a pathogenic model for NS caused by EMP2 mutations and furthermore suggests that CAV1 may serve as a novel therapeutic target for NS and podocyte injury.

Results

Quantitation of Podocyte Injury in a Zebrafish Model

Previously we established a transgenic zebrafish model of inducible podocyte injury, in which the bacterial nitroreductase (NTR) is expressed specifically in podocytes under the control of zebrafish podocin (pod) promoter.⁷ The prodrug metronidazole (MTZ) is converted into a cytotoxin only in podocytes of this transgenic line, resulting in podocyte damage.^8,9 To refine this model of podocyte injury, we generated a new transgenic line Tg(pod:Gal4) to express the transcription factor Gal4 specifically in podocytes of zebrafish. When Tg(pod:Gal4) was combined with Tg(UAS:NTR-mCherry), a stronger expression of NTR-mCherry was achieved specifically in podocytes (Figure 1A). This significantly enhanced the efficiency of podocyte injury induced by MTZ. After 48 hours of MTZ treatment, disruption of podocyte integrity was observed as indicated by reabsorbed mCherry in proximal tubule, and podocytes underwent apoptosis by immunostaining of cleaved caspase-3 in 5-day-old double-transgenic larvae (Figure 1, A and B). These larvae manifested periorbital edema (POE), resembling the edema symptom seen in children with NS (Figure 1C). Extensive foot process effacement was revealed in pronephric glomeruli of edematous fish by transmission electron microscopy (TEM), confirming the disruption of GFB (Supplemental Figure 1A). Using the ratio between pupil spacing distance (PS) and body length (BL) of the larvae as a metric of the severity of POE (Supplemental Figure 1B), we determined that PS/BL in normal control fish with no observable POE is 0.147±0.007 (n=30), and fish with PS/BL≤0.17 does not show observable POE (P<0.001) (Figure 1D, Supplemental Figure 1, C–F). In those fish with severe POE, yolk sac edema and mild pericardial effusion were also observed (Supplemental Figure 2, A–D). The heart rate of edematous fish was significantly higher than that of normal controls (Supplemental Figure 2E), reminiscent of a similar symptom caused by intravascular depletion in nephrotic patients with edema. Taking advantage of this unique POE phenotype as an indicator of podocyte injury, we quantitated the percentage of larvae exhibiting POE and determined the dosage response curve to MTZ concentrations (Figure 1E). This has established a convenient method to assess the susceptibility of podocytes to injury in zebrafish.

Figure 1.

Zebrafish model of MTZ-induced podocyte injury. (A) Epifluorescence image of pronephric glomeruli of Tg(pod:Gal4; UAS:NTR-mCherry) larvae. In control fish without MTZ treatment, mCherry fluorescence in podocytes is strong, but after MTZ treatment, mCherry fluorescence in podocytes is reduced and appears in proximal tubules (white arrowheads), indicating podocyte injury and podocyte loss. (B) Confocal image of pronephric glomerulus from transgenic larvae Tg(pod:Gal4; UAS:NTR-mCherry). MTZ induces podocyte apoptosis as shown by immunostaining against cleaved caspase-3 (green). (C) Representative dorsal view of 5-day postfertilization larvae treated with various concentrations of MTZ (0, 60, 80, 100, and 200 μM) for 48 hours that exhibit POE (red arrow) and yolk sac edema (blue arrow). (D) Distribution of the ratio between PS and BL for Tg(pod:Gal4; UAS:NTR-mCherry) larvae treated with MTZ. (E) Quantitation of the percentage of Tg(pod:Gal4; UAS:NTR-mCherry). Magnification in Panel A and D, ×9.2; Panel B ×400. DMSO, dimethyl sulfoxide.

emp2 Deletion in Zebrafish Predisposes Podocytes to Injury, Which Is Treatable with Glucocorticoids

To investigate the pathogenic mechanism underlying the NS caused by EMP2 mutations, we generated emp2 deletion mutant zebrafish using two customized TAL-like nucleases (TALENs). These introduced a double-strand DNA break in exon 4 of zebrafish ortholog of EMP2 and established two independent mutant lines of emp2 deletion. Subsequent sequencing of the genomic DNA and the cDNA of emp2 confirmed that one mutant allele contained a 145 base-pair (bp) deletion and 26 bp insertion, whereas the other included a 181 bp deletion (Figure 2A). Both alleles removed the obligatory splice junction between intron 3 and exon 4, revealing a cryptic splice site within exon 4 (Figure 2B). This resulted in a mutated transcript for emp2 predicted to encode a novel protein that shares only the first 59 amino acids with the wild-type emp2 (Supplemental Figure 3, A and B). The mutant emp2 has no homology with any known protein in the National Center for Biotechnology Information database. Because the wild-type emp2 contains four transmembrane domains, whereas the mutant has kept only one, we deduced that the mutant is likely to be a functional null. Indeed, overexpression of the wild-type zebrafish emp2, but not the mutant, could suppress cav1 expression in zebrafish embryos and in cultured human podocytes (Supplemental Figure 3, C and D). This suggests that the mutant is potentially a loss-of-function allele. To minimize the potential genetic background variation and off-targeting of TALENs, we performed the phenotypic characterization in both alleles.

Figure 2.

emp2 deletion exacerbates MTZ-induced podocyte injury. (A) TALENs are designed to target emp2 coding sequence in exon 4 (red bar); homozygous deletion on genomic DNA and cDNA is confirmed by PCR. (B) Schematic graph of wild-type and mutated emp2 alleles. Dotted box indicates the deleted region, which includes an obligatory splicing acceptor site. (C and D) Quantitation of the percentage of fish with POE in two alleles of emp2 deletion mutants after MTZ treatment. Bars in the dot plot indicate the mean values of three triplicate tests. (E) Electron micrographs of podocyte foot processes in pronephric glomeruli of 5-day postfertilization larvae. The wild type (WT) fish serves as a control.

The homozygous mutants appeared morphologically normal and were viable and fertile with no obvious edema (Supplemental Figure 4, A and B). However, when podocyte injury was induced with MTZ, the homozygotes displayed an increased sensitivity to MTZ, showing a higher percentage of fish with POE compared with the heterozygous siblings when treated with the same dosage of MTZ (Figure 2, C and D). In the wild-type control fish without POE, foot processes of podocytes appeared normal in TEM, whereas massive loss of foot processes was observed in pronephric glomeruli of emp2^−/− with POE in TEM (Figure 2E), indicating that the edema phenotype is associated with the loss of GFB. Therefore, emp2 deletion predisposed the podocytes to MTZ-induced injury in our zebrafish model.

Given the report of EMP2 mutations associated with SSNS,³ we tested glucocorticoid treatment on emp2 mutant fish. We found that the percentage of fish with MTZ-induced POE for both emp2 alleles could be significantly lowered by treatment of either dexamethasone or prednisone (Figure 3, A and B). This indicated that glucocorticoids could significantly reduce the MTZ-induced podocyte injury in emp2 mutants.

Figure 3.

Glucocorticoid treatment reduces the MTZ-induced podocyte injury in zebrafish emp2 mutants. Quantitation of the percentage of fish POE phenotype induced by MTZ (80 μM) in zebrafish larvae of emp2^{−/−12d145a26} (A) and emp2^{−/−45d181} (B) after dexamethasone (50 nM) and prednisone (10 μM) treatment. Dex, dexamethasone; Prdn, prednisone. DMSO, dimethyl sulfoxide. ***P<0.001 (on the basis of three triplicate tests).

Inhibition of EMP2 in Cultured Podocytes Increases Puromycin Aminonucleoside-induced Podocyte Injury, Which Is Treatable with Glucocorticoids

To verify our findings in emp2 mutant zebrafish, we characterized podocyte injury in the cultured human podocyte model of EMP2 knockdown. Puromycin aminonucleoside (PAN) is known to induce podocyte apoptosis in vitro¹⁰; therefore, we treated the cultured human podocytes with PAN after shRNA-mediated knockdown of EMP2 and found that EMP2 knockdown significantly increased the level of cleaved CASPASE-3 compared with the scrambled shRNA control in the presence of PAN (Figure 4A). This was confirmed by the quantitative assays of CASPASE-3 activity when podocyte apoptosis was induced by 30 or 60 mg/ml PAN (Figure 4B). This indicated that more PAN-induced podocyte apoptosis occurred when EMP2 was inhibited. More importantly, 10 μM dexamethasone treatment negated such an increase of CASPASE-3 activity because of EMP2 knockdown (Figure 4B). Therefore, consistent with our zebrafish model of emp2 deletion, the in vitro model of EMP2 knockdown revealed that inhibition of EMP2 exacerbated podocyte injury, which was responsive to glucocorticoid treatment.

Figure 4.

EMP2 depletion in cultured human podocytes exacerbates PAN-induced podocyte apoptosis, which is alleviated by dexamethasone treatment. (A) Western blot of CASPASE-3 for cultured human podocytes transfected with scramble shRNA (Scr) or shRNA against EMP2 (shrNA#1) followed by puromycine aminonucleoside (puro) treatment. The sequence of shRNA#1 was previously reported.³ Arrowhead indicates the cleaved CASPASE-3. (B) Quantitation of CASPASE-3 activity in cultured human podocytes. CASPASE-3 activity was induced by puro treatment, and dexamethasone (Dex) treatment reduced this activity significantly. *P<0.05; ***P<0.001; N.S, no significance.

Caveolin-1 Is Upregulated by Inhibition of emp2 and Downregulated by Glucocorticoids in Zebrafish and Cultured Human Podocytes

Previously we demonstrated that knockdown of EMP2 in cultured human podocytes elevated CAV1 expression, here we sought to determine whether emp2 negatively regulates cav1 expression in vivo. By quantitative RT-PCR, we found that the transcript level of cav1 was significantly higher in emp2 homozygotes than emp2 heterozygotes at 1 and 4 days post-fertilization (Figure 5A). This was true as well in the isolated mesonephric glomeruli from adult zebrafish (Figure 5A), confirming that emp2 deletion elevated cav1 expression in vivo.

Figure 5.

Glucocorticoids reduce CAV1 expression due to loss of EMP2 in cultured human podocytes and zebrafish. (A) Quantitation of cav1 expression in zebrafish larvae and isolated mesonephric glomerulus. (B) Western blot against EMP2, CAV1, phosphorylated-CAV1, VEGFA, ACTIN for cultured human podocytes transfected with scramble shRNA (Scr), or shRNA against EMP2 (shRNA#1) accompanied with dexamethasone treatment. (C) Quantitation of cav1 expression in both alleles of zebrafish emp2 mutant larvae treated with DMSO, dexamethasone, or prednisone. ef1a was used as an internal control of quantitative RT-PCR. ***P<0.001 (on the basis of three triplicate tests). DMSO, dimethyl sulfoxide.

Moriyama et al. reported that CAV1 expression was significantly decreased in glomerular diseases treated with steroids¹¹; therefore, we tested whether glucocorticoids may decrease cav1 expression in our model of emp2 deletion. We found that dexamethasone could reduce the protein levels of CAV1 and phosphorylated CAV1 in cultured human podocytes with EMP2 knockdown (Figure 5B). Consistently, dexamethasone and prednisone could significantly decrease the transcript level of cav1 in emp2 mutant zebrafish (Figure 5C). These data suggested that the increase of CAV1 expression caused by EMP2 mutations may mediate the podocyte dysfunction in this disease and that the beneficial effect of glucocorticoids may act through the downregulation of CAV1.

Overexpression of cav1 in Zebrafish Podocytes Causes Podocyte Injury and Proteinuria, Which Is Treatable with Glucocorticoids

To test this hypothesis, we generated a transgenic zebrafish model of podocyte-specific cav1 overexpression. Because zebrafish have two alternatively spliced transcripts of cav1 (cav1a and cav1b), we generated two transgenic lines using the UAS sequence to drive the expression of each isoform, respectively (Supplemental Figure 5A). We used the Tol2 gateway system to assemble the transgenic constructs and incorporated the cmlc2:EGFP transgene to facilitate the screening of the transgene carrier such that the transgenic embryos could be easily identified by a green fluorescent heart (Supplemental Figure 5B). When these transgenic fish were crossed with Tg(pod:Gal4), strong expression of cav1a or cav1b in pronephric podocytes could be detected by in situ hybridization with transcript-specific probes (Supplemental Figure 5C). The double transgenic fish Tg(pod:Gal4;UAS:cav1a/b,cmlc2:EGFP) appeared morphologically normal and were viable and fertile.

We therefore tested whether the cav1 overexpression augmented the susceptibility to MTZ-induced podocyte injury. Tg(pod:Gal4;UAS:cav1a/b,cmlc2:EGFP;UAS:NTR-mCherry), when subjected to treatments with various concentrations of MTZ, showed a higher percentage of fish with POE compared with the control group without cav1 overexpression at 48 hours post-MTZ treatment (Figure 6, A and B). This indicated an increased susceptibility to MTZ-induced podocyte injury because of the elevated cav1 expression. Furthermore, we measured the proteinuria in Tg(pod:Gal4;UAS:cav1a/b,cmlc2:EGFP;UAS:NTR-mCherry;lfabp:vdbp-gfp) using ELISA against GFP after MTZ treatment and found that cav1-overexpressing fish produced significantly higher proteinuria compared with the control group Tg(pod:Gal4;UAS:NTR-mCherry;lfabp:vdbp-gfp) treated with the same dosage of MTZ at 24 hours post-treatment (Figure 6C, Supplemental Table 1), indicating that podocyte-specific overexpression of either cav1 transcript predisposed podocytes to MTZ-induced injury. Because the proximal tubules reabsorb the VDBP-GFP protein through megalin-mediated endocytosis,¹² we used a morpholino oligo to inhibit the zebrafish ortholog of megalin (lrp2a) and block such reabsorption.¹³ This enabled us to detect a milder protein leakage through the glomerular filtration. In fact, when lrp2a was inhibited, both cav1a and cav1b transgenic fish manifested significantly higher GFP leakage than the control group (Figure 6D). These data suggested that excessive cav1a or cav1b in podocytes had already impaired the glomerular filtration barrier even without any induction of podocyte injury. However, most plasma protein leaking through GFB was presumably recovered through tubular reabsorption.

Figure 6.

Overexpression of cav1a or cav1b in zebrafish podocytes exacerbates MTZ-induced podocyte injury and proteinuria. (A) Representative dorsal views of Tg(pod:Gal4;UAS:NTR-mCherry;UAS:cav1a/b,cmlc2:EGFP) larvae with POE after being treated with MTZ (60 µM) for 48 hours. (B) Quantitation of the percentage of fish with POE after MTZ treatment. The nonfluorescence fish were used as negative controls and had no observable POE phenotypes after being treated with MTZ. (C) Measurement of proteinuria after MTZ treatment. *P<0.05; **P<0.01; ***P<0.001 (on the basis of three triplicate tests). (D) Measurement of proteinuria for Tg(pod:Gal4;UAS:NTR-mCherry;UAS:cav1a/b,cmlc2:EGFP; lfabp:vdbp-gfp) without MTZ induction of podocyte injury. lrp2a MO was used to inhibit the megalin-mediated reabsorption in proximal tubules. NTR-mCherry, Tg(pod:Gal4; UAS:NTR-mCherry); NTR-mCherry and Cav1a, Tg(pod:Gal4; UAS:NTR-mCherry; UAS:cav1a,cmlc2:EGFP); NTR-mCherry and Cav1b, Tg(pod:Gal4; UAS:NTR-mCherry; UAS:cav1b,cmlc2:EGFP).

We further tested whether glucocorticoids may protect the podocytes in cav1 overexpressing transgenic fish. We treated these transgenic fish with 80 μM of MTZ after a 5-hour pretreatment of dexamethasone or prednisone. In the presence of an increasing dosage of dexamethasone or prednisone, the percentage of POE in cav1a or cav1b transgenic fish was decreased to a level comparable with that of Tg(pod:Gal4;UAS:NTR-mCherry) control group (Figure 7, A and B, Supplemental Table 2). Consistently, we found the proteinuria induced by 80 μM of MTZ was also significantly reduced by cotreatment of 50 nM of dexamethasone or 10 μM of prednisone in cav1a or cav1b transgenic fish (Figure 7C). Therefore, glucocorticoid treatment could effectively reduce the percentage of larvae that developed POE and proteinuria, suggesting the podocyte injury was ameliorated by glucocorticoids in cav1 transgenic fish.

Figure 7.

Glucocorticoids ameliorate MTZ-induced podocyte injury and proteinuria in cav1 overexpressing transgenic zebrafish. (A and B) Quantitation of the percentage of fish with POE induced by 80 µM of MTZ in Tg(pod:Gal4; UAS:NTR-mCherry) and Tg(pod:Gal4; UAS:NTR-mCherry; UAS:cav1a/b, cmlc2:EGFP), with various dosages of dexamethasone or prednisone treatment. (C) Measurement of proteinuria for Tg(pod:Gal4; UAS:NTR-mCherry; UAS:cav1a/b, cmlc2:EGFP; lfabp:vdbp-gfp) after 80 µM MTZ and glucocorticoid treatments.

Taken together, we conclude that overexpression of cav1 in podocytes is detrimental to podocyte function and renders the podocyte more susceptible to MTZ-induced injury, whereas glucocorticoids could rescue such defects through downregulation of cav1.

Discussion

The disruption of the glomerular filtration barrier, primarily because of podocyte dysfunction, plays a pivotal role in the pathogenesis of NS. During the last decade, mutations in more than a dozen of genes have been identified to be the monogenic causes of SRNS, strongly indicating that SRNS is a genetic disease. However, whether SSNS, which constitutes most NS patients, has any genetic basis remains debatable. Previously we identified detrimental mutations in EMP2 associated with SSNS and SRNS. In this study, we have now established a pathogenic model of EMP2 mutations in which loss of EMP2 elevates CAV1 expression, subsequently leading to podocyte dysfunction. In this model, glucocorticoids downregulate CAV1 expression and protect podocytes from injury.

EMP2 has been shown to negatively regulate CAV1 and CAV2 expression in epithelial cells.^5,6 Our data also demonstrated such a function of EMP2 in podocytes in vitro and in vivo; however, the mechanism underlying the regulation of CAV1 by EMP2 remains unclear. CAV1 is a major protein component of caveolae, special lipid raft domains on the plasma membrane, where various cell signaling molecules are located.¹⁴ In glomeruli, CAV1 expression is found in both podocytes and endothelial cells and interacts with proteins essential for the slit diaphragm, such as nephrin and CD2AP.¹⁵ However, Cav1-deficient mice, which do not form caveolae, appear to have no glomerular defect¹⁵; therefore, CAV1 and caveolae are dispensable for normal glomerular function. On the other hand, an increased expression of CAV1 was found in renal glomeruli of patients with glomerular diseases,¹¹ suggesting that excessive CAV1 may be contributing to the pathogenesis of glomerular diseases.

In this study, we focused on the pathogenic role of excessive CAV1 in podocytes and used transgenic zebrafish models to show that overexpression of CAV1 is sufficient to impair normal podocyte function and exacerbate podocyte injury. Consistently, Ren et al. have shown that Cav1 overexpression significantly enhances podocyte apoptosis and nephrin dephosphorylation induced by angiotensin-II in cultured murine podocytes.¹⁶ Therefore, it is a high level of CAV1, rather than a lower level, that negatively affects podocyte health after injury. Intriguingly, caveolae proteins have been implicated in the regulation of p53-dependent cell senescence because Cav1 interacts with and sequesters Mdm2,¹⁷ a negative regulator of p53 pathway.^18–20 Indeed, it has been reported that Mdm2 regulates p53-mediated podocyte apoptosis.^21,22 Therefore, it is possible that excessive CAV1 in podocytes will deregulate the p53 pathway and affect podocyte survival. CAV1 has also been implicated in the regulation of Ca²⁺-permeable channels, such as TRPC²³ and CRAC.^24,25 Therefore, it is likely that excessive CAV1 may alter the Ca²⁺ homeostasis in podocytes and deregulate the downstream transcriptional pathways.

Steroids are known to be clinically effective in treating glomerular diseases, particularly SSNS, but the molecular mechanism underlying the beneficial effects of steroids is largely unknown.² Because other immunosuppressants are also used successfully to treat glomerular diseases, it is thought that steroid therapy is mediated through immunosuppression. However, this notion has recently been challenged because a growing collection of evidence suggests that these drugs exert therapeutic effects directly on podocytes without alteration of immune cells.^26–28 Our findings produced a novel hypothesis regarding the therapeutic effect of steroids on podocytes, in which steroids downregulate CAV1. Cav1-deficient mice are viable, but they have a reduced lifespan because of progressive development of pulmonary, cardiac, and other diseases.²⁹ Notably, CAV1 deficiency in mice protects glomeruli from diabetic nephropathy³⁰; therefore, lowering CAV1 level may be protective to glomerular podocytes under certain disease conditions. However, little is known about how CAV1 expression is regulated.

Being a member of the tetraspan protein superfamily, EMP2 is known to interact with and regulate a number of important cell adhesion and cell signaling molecules, such as β1 integrin,⁴ FAK, and Src,^31,32 which are also known to play important roles in podocytes. The mechanisms by which expression and activity of these proteins are affected by EMP2 deletion and how it subsequently influences podocyte function and health deserve further investigation.

Zebrafish has been recognized as a new model system for glomerular research and podocyte biology. With the advent of genome editing tools, such as TALEN, zebrafish now have emerged as a new model organism for kidney diseases, in addition to its conventional usage in developmental biology studies of the kidney. With genetically modified zebrafish, we have established a new animal model of monogenic NS treatable with glucocorticoids. Such a disease model not only has aided in the study of pathogenic mechanism, but it also holds potential for the discovery of novel therapeutic alternatives to the conventional steroid regimen.

Concise Methods

Zebrafish Maintenance

Wild-type (AB strain) zebrafish, transgenic zebrafish, and larvae were maintained in the zebrafish facility according to the University Committee on Use and Care of Animals standards. All procedures were approved by the University Committee on Use and Care of Animals at the University of Michigan. Eggs were collected in petri dishes and raised in incubators with a constant temperature of 28.5°C.

Mutants

TALEN constructs TAL3246 (#42709, Addgene) and TAL3247 (#42710, Addgene), which target emp2 on exon 4, were obtained from the laboratory of Dr. Keith Joung at Massachusetts General Hospital. Capped RNA of TALENs were synthesized using mMESSAGE mMACHINE T7 Transcription Kit (Ambion) and microinjected into one cell–stage embryos as described.³³ Positive founders were identified by detecting heterozygous offsprings after outcrossing to wild type. Homozygous mutants were obtained by in-crossing the F1 heterozygous individuals with the same mutant allele and confirmed by sequencing. Mutated emp2 cDNA was PCR amplified after RNA isolation and cDNA synthesis followed by sequencing. Primers for emp2 mutant genotyping were listed in Supplemental Table 3.

Transgenic Zebrafish

Gal4 coding sequence was amplified from SAGVG³⁴ and subcloned into pTol2-pod:GFP^7,35 by replacing GFP with AgeI and ClaI to generate pTol2-pod:Gal4 construct. Zebrafish cav1a and cav1b were cloned into the pME-MCS followed by recombination with p5E-UAS, p3E-polyA, and pDestTol2pA7 using the multisite gateway cloning method³⁶ to generate Tg(UAS:cav1a,cmlc2:EGFP) and Tg(UAS:cav1b,cmlc2:EGFP). The transgenic constructs were coinjected with Tol2 transpotase RNA into one cell–stage embryos for genomic integration. Tg(pod:Gal4) stable transgenic lines were screened by PCR verification of Gal4 transgene and further defined by pronephric expression of mCherry in the progenies when crossed to Tg(UAS:NTR-mCherry) [also termed Tg(UAS:Eco.NfsB-mCherry)] transgenic fish, which were obtained from Zebrafish International Resource Center (ZFIN ID: ZDB-TGCONSTRCT-110215–5). Tg(UAS:cav1a,cmlc2:EGFP) and Tg(UAS:cav1b,cmlc2:EGFP) were identified by presence of EGFP expression in the myocardium (Figure 1A). Triple transgenic zebrafish Tg(pod:Gal4;UAS:NTR-mCherry;UAS:cav1a/cav1b,cmlc2:EGFP) were obtained by multiple rounds of crossing the transgenic zebrafish. Quadruple transgenic zebrafish embryos Tg(pod:Gal4; UAS:NTR-mCherry;UAS:cav1a/cav1b,cmlc2:EGFP;lfabp:vdbp-egfp) were obtained by crossing the triple transgenic zebrafish previously mentioned to Tg(lfabp:vbp-gfp).⁷ emp2 mutants at Tg(pod:Gal4; UAS:NTR-mCherry) background were obtained by several rounds of crossing Tg(pod:Gal4; UAS:NTR-mCherry) to emp2 mutants followed by genotyping. Primers for cloning the Gal4, cav1a, and cav1b were listed in Supplemental Table 3.

Characterization of Pero-orbital Edema Phenotypes

Transgenic embryos Tg(pod:Gal4; UAS:NTR-mCherry) and Tg(pod:Gal4; UAS:NTR-mCherry; UAS:cav1a/cav1b,cmlc2:EGFP) were sorted according to mCherry expression in the pronephros and GFP in the myocardium, respectively, and treated with MTZ at 80 hours postfertilization (hpf) for 48 hours. Transgenic zebrafish Tg(pod:Gal4; UAS:NTR-mCherry) in emp2 mutant background were harvested and treated with MTZ at 96 hpf for 48 hours. PS and BL were measured using the CellSens software (Olympus), and specific values of PS/BL were calculated. The normality of the PS/BL data in wild-type control group was tested using Shapiro–Wilk test, and the normal distribution was confirmed (P=0.88). The PS/BL of 0.17 defines the 99.9 percentile of the distribution. There were 25–30 embryos were pooled together in groups for MTZ treatment. Each experiment was repeated independently for at least three times.

Glucocorticoid Treatment

Transgenic embryos were pretreated with glucocorticoids for 5 hours and subsequently subjected to MTZ treatment along with either dexamethasone or prednisone administration at indicated concentration. Emp2 mutant embryos were treated with glucocorticoid at 96 hpf for 48 hours to detect cav1 expression. There were 30 embryos put together as each cohort. All experiments were repeated at least three times independently.

Proteinuria Assay

Transgenic emrbyos Tg(pod:Gal4; UAS:NTR-mCherry; lfabp:vbp-gfp) and Tg(pod:Gal4; UAS:NTR-mCherry; UAS:cav1a/cav1b,cmlc2:EGFP; lfabp:vdbp-gfp) were sorted on the basis of fluorescence, and three embryos were placed in each well of a 96-well plate in 150 µl of E3 medium as a cohort. MTZ administration was performed at 96 hpf. E3 medium in each cohort was collected and subjected to proteinuria assay using ELISA⁷ at 120 hpf. Transgenic embryos were injected with 0.1 mM lrp2a morpholino(5_-AATCAGTGCTTGTGGTTTACCTGGG-3)¹³at one cell stage and sorted out at 72 hpf by fluorescence. Ten positive embryos were put together in a 24-well plate with 500 µl of E3 medium as each cohort. ELISA was performed at 96 hpf.

Real-Time PCR

Glomerulus from zebrafish larvae or adults were dissected under fluorescence microscopy. Total RNA were extracted and cleaned (Zymo Research) followed by first-strand cDNA synthesis (Invitrogen). Real-time PCR was performed using the SYBR Green Kit on the MyiQ real-time PCR machine (BIORAD). The primers used are listed in Supplemental Table 3.

Whole-Mount In Situ Hybridization

Partial coding sequences for cav1a and cav1b³⁷ were PCR amplified and cloned into pCMV-Script, and in situ riboprobes were synthesized using DIG RNA Labeling Kit (Roche). Embryos were collected and fixed in 4% paraformaldehyde in PBS. In situ hybridization was carried out following standard protocol.³⁸

Immunostaining and Confocal Microscopy

Transgenic embryos Tg(pod:Gal4; UAS:NTR-mCherry) were fixed in 4% paraformaldehyde in PBS. Immunostaining was carried out in PBST (0.1% TritonX-100, 5% bovine serum albumin with anti-cleaved Caspase3 (cat 559565, 1:1000; BD Pharmingen). Images were taken under Leica SP5 confocal microscopy (Leica).

Cell Culture and Transfection

The immortalized human podocytes were maintained at the permissive temperature of 33°C in RPMI + GlutaMAX-I supplemented with 10% fetal bovine serum, penicillin (50 IU/ml)/streptomycin (50 μg/ml), and insulin-transferrin-selenium-X. To differentiate podocytes, they were cultured at 37°C for 14 days. All cell culture materials were purchased from Life Technologies. Podocytes stably transfected with scrambled or EMP2 shRNAs, and myc-tagged EMP2 expression constructs (myc-EMP2,myc-EMP2^Q62*, myc-emp2(zebrafish), myc-emp2^MT(zebrafish emp2 mutant)) were previously described.³ The target sequence of shRNAs is GCGTGAAGACATTCACGACAA (shRNA #1). The shRNA-stable podocytes were selected and maintained with 8 μg/ml puromycin. Dexamethasone (Sigma-Aldrich) were dissolved in culture medium at concentration of 10 μM and treated for 14 days during differentiation of podocytes. Dexamethasone-containing or control medium was supplied on alternate days. To induce apoptosis in podocytes, PAN was added to the medium at the concentrations ranging from 0 to 100 μg/ml for 24 hours.

Immunoblotting

Immunoblotting was performed as described previously.³⁹ Then 50 μg of cell lysates was loaded into each lane. Anti-EMP2 (Abgent), anti-VEGFA, anti–β-actin (Abcam, Inc.), anti-CAV1, anti–phospho-CAV1, and anti-CASPASE3 (Cell Signaling Technology) were purchased from commercial sources.

Caspase-3 Assay

The activity of caspase-3 was measured using the Caspase 3 Assay Kit (cat# ab39401; Abcam, Inc.) according to the manufacturer’s instructions.

TEM

TEM was carried out according to published procedure.⁴⁰

Disclosures

None.