Tankyrase inhibition aggravates kidney injury in the absence of CD2AP

Inappropriate activation of the Wnt/β-catenin pathway has been indicated in podocyte dysfunction and injury, and shown to contribute to the development and progression of nephropathy. Tankyrases, multifunctional poly(ADP-ribose) polymerase (PARP) superfamily members with features of both signaling and cytoskeletal proteins, antagonize Wnt/β-catenin signaling. We found that tankyrases interact with CD2-associated protein (CD2AP), a protein essential for kidney ultrafiltration as CD2AP-knockout (CD2AP−/−) mice die of kidney failure at the age of 6–7 weeks. We further observed that tankyrase-mediated total poly-(ADP-ribosyl)ation (PARylation), a post-translational modification implicated in kidney injury, was increased in mouse kidneys and cultured podocytes in the absence of CD2AP. The data revealed increased activity of β-catenin, and upregulation of lymphoid enhancer factor 1 (LEF1) (mediator of Wnt/β-catenin pathway) and fibronectin (downstream target of Wnt/β-catenin) in CD2AP−/− podocytes. Total PARylation and active β-catenin were reduced in CD2AP−/− podocytes by tankyrase inhibitor XAV939 treatment. However, instead of ameliorating podocyte injury, XAV939 further upregulated LEF1, failed to downregulate fibronectin and induced plasminogen activator inhibitor-1 (PAI-1) that associates with podocyte injury. In zebrafish, administration of XAV939 to CD2AP-depleted larvae aggravated kidney injury and increased mortality. Collectively, the data reveal sustained activation of the Wnt/β-catenin pathway in CD2AP−/− podocytes, contributing to podocyte injury. However, we observed that inhibition of the PARylation activity of tankyrases in the absence of CD2AP was deleterious to kidney function. This indicates that balance of the PARylation activity of tankyrases, maintained by CD2AP, is essential for normal kidney function. Furthermore, the data reveal that careful contemplation is required when targeting Wnt/β-catenin pathway to treat proteinuric kidney diseases associated with impaired CD2AP.

The glomerular filtration barrier consists of fenestrated endothelial cells, glomerular basement membrane and glomerular epithelial cells or podocytes. Disruption of the filtration barrier leads to proteinuria, which may progress to renal failure. Podocytes are terminally differentiated cells with a large cell body and interdigitating foot processes, which rest on the glomerular basement membrane. Podocyte injury has been shown to have an important role in the development of proteinuria as the loss of function of several podocyte proteins has been shown to lead to podocyte foot process effacement and impaired glomerular ultrafiltration. 1 Adapter protein CD2-associated protein (CD2AP) is essential for the maintenance of glomerular filtration. 2 CD2AP was originally characterized as an interaction partner of T-cell adhesion molecule CD2, 3 and as a protein upregulated during mesenchyme to epithelium differentiation in the developing kidney. 4 In the glomerulus, CD2AP interacts with nephrin, 5 the key component of the slit diaphragm interconnecting the neighboring podocyte foot processes. 6 CD2AP regulates cytoskeletal remodeling, 7 apoptosis 8 and vesicular trafficking. 9 Tankyrases 1 and 2 are members of the poly(ADP-ribose) (PAR) polymerase family. 10,11 They catalyze poly(ADP-ribosyl)ation (PARylation), a post-translational modification targeting tankyrases themselves or their interaction partners, subject the target proteins to ubiquitination and proteasomal degradation, 10 and thereby modulate diverse cellular processes. 12 During recent years, several new interaction partners and functions have been characterized for tankyrases; [13][14][15] however, the possible involvement of tankyrases in regulating podocyte function has remained unknown. PARylation activity of tankyrases is involved in regulating the activity of the canonical Wnt/β-catenin signaling pathway. 16 When the pathway is inactive, β-catenin is phosphorylated and targeted to degradation by the β-catenin destruction complex. Tankyrases PARylate AXIN in the destruction complex, stimulate its degradation, and thus enable dephosphorylation of β-catenin and its consequent translocation to the nucleus where it together with T-cell factor/lymphoid enhancer factor (TCF/LEF) family of co-activators drives Wnt-specific transcriptional programs. 17 In the adult kidney, Wnt/β-catenin pathway has been considered to be silent. Inappropriate activation of the pathway has been indicated in podocyte dysfunction and injury, 18 and in many kidney diseases such as diabetic nephropathy and adriamycininduced nephropathy. [19][20][21][22] Here, we show that tankyrases are novel interaction partners of CD2AP, and that CD2AP is a negative regulator of tankyrases. We demonstrate that activation of β-catenin accompanied by elevation of total PARylation associates with podocyte injury in the absence of CD2AP. However, inhibition of tankyrases in the absence of CD2AP enhances the expression of podocyte injury markers and aggravates kidney damage.

Results
Tankyrases 1 and 2 form a complex with CD2AP. We performed a yeast two-hybrid screening on a rat glomerular library to identify proteins that interact with the NH 2 terminus of CD2AP containing three SH3 domains. The screening identified a clone that covered amino acids 1-190 of tankyrase 2. Pull-down assays with the NH 2 Figure S1b). Also endogenous tankyrases were pulled down with the 2.SH3 domain of CD2AP from rat glomerular lysate ( Figure 1c). The interaction between CD2AP and tankyrase 2 was mediated by the ankyrin domains ( Figure 1d). Both fulllength tankyrases and the ankyrin domains of tankyrase 2 were co-immunoprecipitated with endogenous CD2AP (Figure 1e). Furthermore, CD2AP co-immunoprecipitated endogenous tankyrases from mouse kidney and isolated rat glomerular lysates (Figure 1f), confirming that CD2AP and tankyrases form a physiological complex.  Tankyrase-mediated total PARylation is increased in CD2AP − / − podocytes in vitro. Immunoblotting of lysates of WT and CD2AP − / − podocytes revealed multiple bands for both tankyrase 1 and tankyrase 2 ( Figure 3a). Tankyrase 1 is expected to migrate at 150 kDa and tankyrase 2 at 130 kDa. Quantification indicated that the total level of tankyrase 1 is decreased but its slower migrating, modified form is increased in CD2AP − / − podocytes compared with WT podocytes (Figure 3b). The total level of tankyrase 2 was similar between WT and CD2AP − / − podocytes; however, the modified form of tankyrase 2 was increased in the absence of CD2AP (Figure 3c). Tankyrase 1 is modified by phosphorylation, 23 and phosphatase (CIP) treatment of WT and CD2AP − / − cell lysates confirmed that the slower migrating form of tankyrase 1 is phosphorylated in podocytes (Supplementary Figure S2). Tankyrase 1 phosphorylation can enhance its PARylation activity 23 or autoPARylation. 16 We therefore examined whether total PARylation is increased in CD2AP − / − podocytes. Indeed, the absence of CD2AP increased total PARylation in podocytes compared with WT cells (Figure 3d). Treatment of CD2AP − / − podocytes with a tankyrase-specific inhibitor XAV939 reduced total PARylation (Figure 3e) whereas ABT-888 (veliparib), a PARP1 and PARP2 (poly(ADP-ribose) polymerase 1 and 2)-specific inhibitor, 24,25 had no effect (Figure 3f). The effectiveness of ABT-888 was confirmed by its ability to inhibit H 2 O 2 -induced PARP1-mediated PARylation (Supplementary Figure S3). These data confirm that the increase in total PARylation in the absence of CD2AP is due to the increased activity of tankyrases.
Tankyrases typically PARylate their interaction partners. 10,13,14,16,23,26 Pull down with PAR-affinity resin from WT podocyte lysates showed that CD2AP is not PARylated (Figure 3g). Interestingly, we detected less PARylated tankyrases in CD2AP − / − podocytes compared with WT podocytes (Figures 3g and h). One of the known PARylation targets of tankyrases is PARP1. 27 PARP1 is an enzyme involved in cell differentiation, survival and transformation, and its main function is to repair single-stranded DNA breaks. 28 Lack of CD2AP increased the PARylation of PARP1 in podocytes and tankyrase inhibitor XAV939 treatment reduced it (Figures 3g and i), indicating that PARylation of PARP1 in the absence of CD2AP is tankyrase mediated. Additionally, inhibitor treatment reduced PARylation of tankyrases in WT and CD2AP − / − podocytes confirming the effectiveness of the inhibitor (Figures 3g and h). Collectively, the absence of CD2AP increases tankyrasemediated PARylation, and one of the PARylation targets of tankyrases is PARP1.
Total PARylation is increased in CD2AP − / − podocytes in vivo. Shih et al. 2 showed that CD2AP-deficient mice develop high-grade albuminuria and exhibit lesions typical of focal segmental glomerulosclerosis (FSGS) by 3-4 weeks of age. Kidney failure results in the death of the animals by the age of 6-7 weeks. 2,8 We examined whether we observe an increase in total PARylation, shown to associate with kidney injury in diabetes, 29 in the kidneys of CD2AP-deficient mice at 3 weeks of age. Indeed, we observed that CD2AP deficiency leads to increased total PARylation in the kidneys compared with WT kidneys (Figure 4a). Double labeling of CD2AP − / − and WT kidneys with anti-PAR and anti-WT1 antibodies revealed that the number of podocytes with PAR-positive  Absence of CD2AP leads to the accumulation of active β-catenin, Lef1 and TCF -1 in podocytes. Owing to the observed overactivity of tankyrases, we hypothesized that one of the mechanisms contributing to podocyte injury in the absence of CD2AP is the activation of Wnt/β-catenin pathway. Indeed, active β-catenin was upregulated in CD2AP − / − podocytes compared with WT podocytes without exogenous Wnt stimulation (Figure 5a), and accumulated in the nucleus (Figures 5a-e). This was also observed in the glomerular podocytes of CD2AP − / − mouse kidneys when compared with WT (Figures 5f-k). TCF/LEF family of transcription factors are the major mediators of Wnt signaling and their expression is upregulated by stabilized β-catenin. 30 qRT-PCR indicated that the absence of CD2AP leads to upregulation of Lef1 and TCF-1 mRNAs in podocytes (Figure 5l), whereas TCF-3 and TCF-4 had similar expression. Interestingly, treatment with tankyrase inhibitor, Wnt/β-catenin pathway antagonist XAV939, had no effect on TCF-1, TCF-3 or TCF-4 expression; however, it significantly increased Lef1 expression in CD2AP − / − podocytes ( Figure 5l).
XAV939 inhibits tankyrases and stabilizes AXIN1, which enhances degradation of active β-catenin. 17 As expected, XAV939 downregulated active β-catenin and increased the levels of AXIN1 in both WT and CD2AP − / − podocytes (Figures 6a and b). LEF1 was upregulated in CD2AP − / − podocytes also at the protein level, and the upregulation was further enhanced by tankyrase inhibition (Figures 6a and b). Immunofluorescence staining confirmed that XAV939 treatment downregulates active β-catenin in both WT and CD2AP − / − podocytes and upregulates LEF1 in CD2AP − / − podocytes (Supplementary Figure S5). Active β-catenin binds to LEF1 to form transcription complexes to activate target gene expression. 31 In line with this, proximity ligation assay (PLA) revealed that there is more interaction between LEF1 and active β-catenin in CD2AP − / − podocytes compared with WT podocytes (Figures 6c, d and g). XAV939 treatment decreased the prevalence of the interaction in CD2AP − / − podocytes, possibly due to downregulation of active β-catenin (Figures 6e-g). Taken together, the absence of CD2AP leads to sustained activation of β-catenin in podocytes, which upregulates LEF1. Interestingly, XAV939 treatment further upregulated LEF1 in CD2AP − / − podocytes.
XAV939 treatment had no effect on MMP-9, Snail1 or FSP-1 expression in WT or CD2AP − / − podocytes (Figure 6h), and fibronectin expression remained elevated in CD2AP − / − podocytes (Figures 6i and k). Furthermore, XAV939 significantly upregulated PAI-1 in CD2AP − / − podocytes ( Figure 6h). As also previously shown, 8 the absence of CD2AP increased the expression of apoptotic p38 MAPK phosphorylation (Figures 6j and l). XAV939 treatment further increased p38 MAPK phosphorylation (Figures 6j and l). Also insulin-stimulated phosphorylation of anti-apoptotic AKT was reduced by XAV939 treatment in podocytes lacking CD2AP (Supplementary Figure S7). Additionally, we investigated the expression of apoptotic markers in CD2AP − / − podocytes with and without tankyrase inhibition. The absence of CD2AP increased the expression of pro-apoptotic Bcl-2-associated X protein (BAX) (Figures 6m and n), whereas the anti-apoptotic B-cell lymphoma 2 protein (Bcl-2), a dual regulator of apoptosis and autophagy, was downregulated. The autophagy signaling adaptor protein p62 (sequestosome-1), which in a context-dependent manner, has a role in the decision of the cells undergoing autophagy to either survive or die, was upregulated in the absence of CD2AP and XAV939 further increased its expression (Figures 6m and n). Full-length caspase-1 was downregulated in the absence of CD2AP (Figures 6m and n). XAV939 treatment downregulated full-length PARP1 in both WT and CD2AP − / − podocytes, and increased the expression of cleaved, apoptotic form of PARP1 in the absence of CD2AP (Figures 6m and n). The data indicate that inhibition of tankyrases in the absence of CD2AP aggravates podocyte injury by increasing pro-apoptotic signaling and upregulating LEF1 and PAI-1 (see cartoon in Figure 8).
Inhibition of tankyrases in cd2ap knockdown zebrafish aggravates kidney injury. To analyze whether inhibition of tankyrases increases podocyte injury in the absence of CD2AP in vivo, we knocked down cd2ap with morpholino antisense oligonucleotides (MOs) in zebrafish, previously shown to result in pronephric kidney dysfunction, 32 coupled with tankyrase inhibition. At 5 days post fertilization (dpf), 60-68% of cd2ap-MO-and cd2ap-MO and tankyrase inhibitor-treated larvae exhibited pericardial edema indicative of kidney dysfunction (Figures 7a and b,Supplementary  Figure S6). Notably, simultaneous suppression of cd2ap expression and tankyrase activity increased the number of larvae displaying both pericardial and yolk sac edema from 13 to 27% (Figure 7d,Supplementary Figure S6) and increased mortality from 3 to 17% compared with suppression of cd2ap expression alone (Figures 7b and d,  Supplementary Figure S6). Downregulation of cd2ap alone or XAV939 treatment of control-MO-injected larvae had no effect on viability (Supplementary Figure S6); and XAV939 treatment of control-MO-injected larvae caused no identifiable malformation (Figure 7c).
Control-MO-injected larvae without or with XAV939 treatment showed well-developed glomeruli and pronephric ducts (Figures 7e and g). The glomeruli of cd2ap-MO-treated larvae exhibited disturbed glomerular structures with distended capillary loops (Figure 7f), whereas the cd2ap knockdown larvae treated with XAV939 showed even more disturbed    (c and d) Inhibition of tankyrase activity with XAV939 in the cd2ap morphant between 3 and 5 dpf induces yolk sac edema and more severe pericardial edema (d, arrowhead) than knockdown of cd2ap alone (b), whereas XAV939 treatment has no effect on the control morphant (c). (e and g) Histological sections of 5 dpf control morphant (e) and control morphant treated with XAV939 (g) show normal glomeruli (arrowheads) and proximal convoluted tubules (arrows).
(f) cd2ap morphant shows a disturbed glomerular structure (arrowhead) with cysts (asterisk). (h) Inhibition of tankyrase activity with XAV939 in cd2ap morphants leads to even more destructed glomerular structure with a dilated (asterisk) and stretched septal glomerulus in the midline (arrowhead) and dilated proximal convoluted tubule (arrow).  (Figures 7i  and j), a protein essential for glomerular structure and function. 33 Notably, XAV939 treatment further downregulated podocin in cd2ap-MO-injected larvae, but had no effect in control-MO-injected larvae (Figures 7i and j). In line with the in vitro experiments, active β-catenin was upregulated in cd2ap-MO-injected larvae (Figures 7i and j). Additionally, suppression of cd2ap increased total PAR levels compared with controls ( Figure 7k). Collectively, the data indicate that suppression of cd2ap coupled with inhibition of tankyrases leads to more severe kidney damage than knockdown of cd2ap alone and increases mortality of the larvae.

Discussion
Canonical Wnt/β-catenin signaling mediates podocyte injury and the development of proteinuria, 18,20 and podocytespecific deletion of β-catenin protects from nephropathy. 19,20 Tankyrases have been shown to be efficient targets in inhibiting this signaling. 17 Here, we show that CD2AP interacts with tankyrase 1 and tankyrase 2, and that in the absence of CD2AP tankyrase-mediated total PARylation in podocytes is increased. Podocytes lacking CD2AP accumulate active β-catenin consequently leading to upregulation of LEF1 and fibronectin, both associated with podocyte injury. Tankyrase inhibitor treatment reduces total PARylation and β-catenin activation in cultured podocytes, but, surprisingly, worsens kidney injury and increases mortality of zebrafish in the absence of CD2AP.
Increased PARylation in podocytes has been shown to have a role in glomerular injury associated with type 2 diabetes and general PARP inhibitors ameliorate this pathological situation. 29 We found that tankyrases are phosporylated in podocytes and that the absence of CD2AP leads to increased total PARylation due to the activity of tankyrases. AutoPARylation of tankyrases was reduced in the absence of CD2AP, indicating that PARylation of tankyrase binding partners or unbound PAR would contribute to the overall increase in total PAR. In line with this, PARP1 PARylation was found modestly increased when CD2AP was depleted. The data also indicated that CD2AP is not PARylated. Very little is still known about the complex relationship between tankyrase phosphorylation, autoPARylation and PARylation activity and the data available on their relationship vary, apparently reflecting cell-type and interaction partner-dependent differences. [14][15][16]23,26,27 Most of the known interaction partners of tankyrases are acceptors of tankyrase PARylation. 14,15,23 CD2AP bears similarity with few proteins, Fanconi anemia protein FANCD2 that controls genomic integrity, 34 and GDP-mannose-4,6-dehydratase (GMD) that is required for fucose synthesis, 35 which bind tankyrases and inhibit their activity, but are not PARylated.
Our data indicate apparent dysregulation of the Wnt/β-catenin pathway in the absence of CD2AP. In cells deficient of CD2AP, active β-catenin is stabilized and translocated to the nucleus. This is coupled with upregulation of its downstream targets, LEF1 and fibronectin, whereas several other typical Wnt/β-catenin target genes associated with kidney injury were not affected. Notably, upregulation of LEF1 and fibronectin occurred without exogenic Wnt stimulation. Our data propose that the major enhancer of the pathway is enhanced tankyrase PARylation activity (Figure 8, CD2AP − / − ). Interestingly, MMP-9 was downregulated in the absence of CD2AP. Lack of CD2AP predisposes cells to apoptosis, 8 and therefore reduced level of MMP-9, shown to protect tubular cells against apoptosis in acute kidney injury, 36 could contribute to the susceptibility of CD2AP-deficient podocytes to apoptosis.
Inhibition of tankyrases has been shown to attenuate pathological conditions associated with activation of Wnt signaling, such as solid tumors 37,38 and fibrotic pulmonary diseases. 39,40 On the basis of this, targeting the Wnt/β-catenin pathway could also serve as a therapeutic strategy to treat a variety of proteinuric kidney diseases showing activation of the Wnt pathway, as also Wang et al. 18 suggested. We found that treatment of CD2AP − / − podocytes with tankyrase inhibitor reduced the activity of β-catenin. However, it did not significantly reduce the increased expression of fibronectin, further increased the expression of LEF1, upregulated PAI-1 ( Figure 8, CD2AP − / − +tankyrase inhibitor), a fibrogenic factor that has been associated with the development of nephropathy, 41 and increased pro-apoptotic signaling. Previously, it has been shown that upregulation of LEF1 coupled with accumulation of active β-catenin renders cancer cells unresponsive to tankyrase inhibition, which is due to LEF1 limiting the association of β-catenin with the destruction complex. 42 Tankyrase inhibition in these situations leads as a forward regulatory loop to further upregulate LEF1. A situation similar to this could explain our findings in podocytes lacking CD2AP. Moreover, LEF1 and TCF-1 can be transcriptionally active also independently of β-catenin, 43 and as PAI-1 gene promoter harbors a functional TCF/LEF site, 22 the observed PAI-1 upregulation can be a result of LEF1 activating the transcription of PAI-1. We thus propose a mechanism by which tankyrase inhibition aggravates podocyte injury in the absence of CD2AP: tankyrase inhibition leads to upregulation of LEF1, which in turn enhances transcription of PAI-1 independently of β-catenin ( Figure 8, CD2AP − / − +tankyrase inhibition). These events contribute to the detrimental consequences of tankyrase inhibition in the absence of CD2AP.
The harmful effects of inhibition of tankyrases in the absence of CD2AP were further supported in zebrafish larvae in vivo, in which knockdown of CD2AP coupled with inhibition of tankyrases increased the severity kidney injury and mortality. As both CD2AP and tankyrases are widely expressed, 2,44 may the phenotype reflect defects in the function of multiple tissues. The worsening of the kidney phenotype is a significant finding when considering treatment strategies for kidney diseases associated with impairment of CD2AP, such as FSGS. 45,46 Mutations in CD2AP observed in FSGS typically lead to reduced expression or absence of CD2AP or render CD2AP biologically defected. 45,46 This could lead to increased tankyrase activity in podocytes. Interestingly, the Wnt pathway has been shown to be activated in patients with primary or experimental FSGS, 47 tempting to speculate that Wnt inhibition would be a beneficial treatment strategy for FSGS. Our findings, however, point out that inhibiting the Wnt Tankyrase inhibition and kidney injury S Kuusela et al pathway with tankyrase inhibitors in the absence of CD2AP could be deleterious.
In summary, our data show a novel mechanism that contributes to kidney injury and dysfunction in the absence of CD2AP. We show that total PARylation and the activity of β-catenin are elevated in CD2AP − / − podocytes, and that both can be inhibited by tankyrase-specific small molecule inhibitor. Tankyrase inhibition, however, further upregulates proteins known to associate with kidney injury, and aggravates kidney dysfunction and increases mortality in CD2APdepleted zebrafish larvae. These findings highlight that inhibition of the Wnt/β-catenin pathway can have adverse effects and that tankyrase inhibitors are not a suitable approach to treat kidney diseases associated with lack of CD2AP.

Materials and Methods
Yeast two-hybrid screening. Construction of the rat glomerular yeast twohybrid library and screening of the library using CD2AP as a bait have been described previously. 48,49 Animals. Male Sprague-Dawley rats and male FVB mice were used to study tankyrases 1 and 2 expression in the kidney. Generation of CD2AP-knockout mouse has been described. 2 Kidneys of the knockout mice and controls were collected at 3 weeks of age and used for analyses as described below. Zebrafish embryos were obtained from the breeding line of Turku strain and raised as described. 50 Animal experiments were performed according to approved guidelines and were approved by the National Animal Experiment Board. Antibodies. Antibodies used were rabbit anti-tankyrase 1/2 and mouse anti-Bcl-2 (Santa Cruz Biotechnology, Dallas, TX, USA), rabbit anti-podocin and mouse anti-tubulin (Sigma-Aldrich), goat anti-tankyrase 2, rabbit anti-BAX and rabbit antifibronectin (Abcam, Cambridge, UK), rabbit anti-CD2AP 1774 or 1764, 4,51 mouse anti-PARP1 and rabbit anti-Poly(ADP-ribose) (Enzo Life Sciences, Farmingdale, NY, USA), mouse anti-active-β-catenin (Millipore, Billerica, MA, USA), rabbit anti-LEF1 and rabbit anti-phospho p38 MAPK (Cell Signaling Technology, Danvers, MA, USA) and mouse anti-WT1 (Upstate, Lake Placid, NY, USA), guinea pig anti-p62/SQSTM1 (Progen Biotechnik GmbH, Heidelberg, Germany) and mouse anti-Caspase-1 (Adipogen, San Diego, CA, USA).
Preparation of tissue and cell lysates and nuclear extracts. Glomerular and tubular fractions were isolated from rat kidney cortices using graded sieving. 52 Tissues and cells were lysed as described. 53 Pooled zebrafish larvae were sonicated in 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS supplemented with 1 mM PMSF and protease inhibitor cocktail (Roche, Basel, Switzerland). Nuclear extracts were prepared as described. 35,54 Pull-down assay, immunoprecipitation and immunoblotting. Pull-down assay with glutathione-S-transferase (GST)-tagged SH3 domains 55 or the NH 2 -terminus (amino acids 1-330) of mouse CD2AP was performed as described. 5 Pull-down assay for PARylated proteins was performed with a WWE Affinity Resin kit (Tulip Biolabs, West Point, PA, USA) according to the manufacturer's instructions, and immunoprecipitation as described. 5 Immunoblotting was performed as described. 51 Figure 8 Schematic illustration of the molecular mechanisms contributing to injury of CD2AP-deficient podocytes upon tankyrase inhibition. In WT podocytes, CD2AP interacts with tankyrases (TNKS) and inhibits their activity. In CD2AP − / − podocytes, the activity of tankyrases is increased, leading to inhibition of beta-catenin (BC) degradation. Active beta-catenin (ABC) accumulates and translocates to the nucleus, where it together with LEF1 upregulates fibronectin. Tankyrase inhibition in CD2AP − / − podocytes leads to accumulation of AXIN1 and degradation of active beta-catenin. However, tankyrase inhibition fails to downregulate fibronectin, further upregulates LEF1 and enhances PAI-1 expression. Additionally, phosphorylation of the pro-apoptotic p38 MAPK is enhanced. These events further aggravate injury in CD2AP-deficient podocytes Immunoperoxidase staining. Mouse kidneys were fixed with 10% formalin, dehydrated and embedded in paraffin. Immunoperoxidase staining was performed with a VectaStain Elite kit (Vector Laboratories, Burlingame, CA, USA) as described. 49 Sections were analyzed with a Nikon Eclipse 800 microscope (Nikon, Tokyo, Japan).
Immunofluorescence staining. Mouse kidney frozen sections (6 μm) were fixed with acetone and cultured podocytes were fixed with 4% paraformaldehyde (PFA) and permeabilized with 0.1% Triton X-100. The stainings were carried out as described. 53 AlexaFluor 594 donkey anti-rabbit and AlexaFluor 488 donkey anti-goat or donkey anti-mouse IgGs were used as secondary antibodies. Samples were analyzed with Leica TCS SP2 or SP8 confocal microscopes (Leica Microsystems, Wetzlar, Germany).
Quantitative RT-PCR. Quantitative PCR was performed as described in Wang et al. 56 Primer sequences for Snail1, MMP-9, FSP-1, PAI-1, LEF1, TCF-1, TCF-3 and TCF-4 are shown in Supplementary Table S1. The expression levels of the studied molecules were normalized to β-actin using the comparative C t method. Reverse transcriptions were performed with three individual RNA preparations and quantitative RT-PCRs were performed in triplicate.
Knockdown and tankyrase inhibitor treatment in zebrafish. MO blocking translation of cd2ap (5′-CATACTCCACCACCACCTCAACCAT-3′) 32 and a standard control MO were purchased from GeneTools, LLC (Philomath, OR, USA). Zebrafish embryos were injected with 4 nl of 100 μM MOs using a Narishige MN-153 microinjector (Tokyo, Japan). Embryos were allowed to develop until 3 days post fertilization (dpf) and then transdermally exposed to 5 μM XAV939 or an equivalent volume of DMSO (0.05%) in E3 medium for 2 days. The phenotype of larvae on 5 dpf was examined under a Leica dissecting microscope.
Histology of zebrafish larvae. Zebrafish larvae at 5 dpf were fixed with 4% PFA in PBS overnight at 4°C and embedded in JB-4 (Polysciences, Inc., Warrington, PA, USA). In all, 4 μm sections were stained with hematoxylin and eosin.
Statistical analysis. Statistical analysis was performed with one-way ANOVA. Results are presented as mean ± S.E.M.