Wnt5a Is Necessary for Normal Kidney Development in Zebrafish and Mice

Congenital kidney and urinary tract abnormalities have been described in patients with Robinow syndrome [1, 2], but the molecular mechanisms are unclear. WNT5A mutations were found in patients with Robinow syndrome and Wnt5a, and its receptor Ror2 knockout mice show phenotypes similar to these patients [3]. Wnt5a is a noncanonical glycoprotein of the Wnt family that regulates a wide range of developmental processes. It is known that Wnt5a is essential for proper skeletal, urogenital and gonad development [4, 5], but the role of Wnt5a in kidney development is unknown.

The mammalian kidney originates from intermediate mesoderm (IM). The metanephros, which develops into the permanent kidney, forms when the ureteric bud (UB) grows out of the nephric (wolffian) duct (ND) and contacts the metanephric mesenchyme (MM). The UB then elongates and branches to form the collecting duct system. The metanephric mesenchymal cells condense around the tip of the UB, aggregate, epithelialize and differentiate to podocytes and proximal and distal tubular cells.

Our goal in this study was to determine the role of Wnt5a in kidney development. We used two models, zebrafish and mice. The simple nature of the zebrafish pronephric kidney makes it a suitable system to study the early developmental events that lay the foundation for genesis of more complex kidneys, while mouse metanephric kidney development is similar to human kidney development.

Wild-type and Tg(wt1b:GFP) (green fluorescent protein) transgenic zebrafish [6]were reared and maintained as described [7]. Embryos were collected after natural spawn, kept at 28.5°C, and staged as described [8]. Embryos from 48-hpf Tg(wt1b:GFP)zebrafish were anesthetized in 0.016% tricaine solution and embedded in 3% methyl cellulose with dorsal side facing up and imaged using a fluorescence microscope (Leica M205C).

Wnt5a morpholino oligonucleotides (MOs) were generated by Gene Tools, LLC (Philomath, Oreg., USA). MOs were injected into embryos at the one-to-four-cell stage as described [9]. Capped mouse Wnt5a full-length mRNA was synthesized using the mMessage mMachine T7 kit (AM1344, Ambion). For the rescue experiments, 40 pg of Wnt5a mRNA was coinjected with the AUG-MOs into one-to-four-cell stage embryos.

Wnt5a^-/- null mice were generated by mating CMV-Cre mice (Jackson Lab stock No. 006054) with Wnt5a^fl/fl mice [10] and were similar in phenotype to the commercially available Wnt5a global knockout mice [11]. The day of vaginal plug was considered E0.5. The Wnt5a^-/- embryos were identified visually by a reduced size phenotype, and the absence of Wnt5a was confirmed by PCR as described [10, 11]. IM from E10.5, E11.5, and metanephros from E13.5 and E15.5 were dissected out and fixed for whole-mount in situ hybridization (wmISH). E11.5-18.5 metanephros were dissected out, and images were taken with a Leica M205C dissecting microscope. MRI of E16.5 Wnt5a^-/- mouse embryos was performed at the Penn Small Animal Imaging Facility using a 9.4-tesla (400 MHz) vertical bore spectrometer with 55-mm gradients.

Mouse Wnt5a was subcloned from MSCV2.2Wnt5aIRES-GPF (kindly provided by Dr. Stephen Jones at University of Massachusetts Medical School) into the pEGM-T Easy Vector (Promega). The plasmid for the mouse Ret gene was kindly provided by Dr. Mendelsohn, Columbia University. The zebrafish wnt5a template was prepared by TA cloning. Digoxigenin-labeled RNA probes were prepared by in vitro transcription (DIG RNA Labeling kit, Roche). wmISH in mouse or zebrafish embryos was performed as previously described [5, 12].

Phenotypes for the rescue experiments were classified as abnormal after MO injection by the presence of any combination of the following features: edema, short body size and curled tails. For comparison of means, a t-test was performed using SPSS software (v.15.0; SPSS Inc., Chicago, Ill., USA). For all tests, p < 0.05 was considered to represent statistically significant differences.

wmISH of wild-type zebrafish embryos at 72 h after fertilization (hpf) confirmed wnt5a expression in the pronephric kidney (fig. 1a, b). The wt1a probe served as a positional marker to confirm wnt5a expression in the pronephric glomerulus (fig. 1c).

We knocked down wnt5a with a translation blocking morpholino (AUG-MO), which targets the start codon and affects both maternal and zygotic wnt5a mRNA, and an exon/intron border splice morpholino (splice-MO), which targets the third splice donor site and affects only the zygotic transcript of wnt5a (fig. 2a). Both the AUG and the splice morphants phenocopied each other, with multiple defects at 72 hpf, including: reduced axis length, curly tail down body axis, and pericardial edema (fig. 2b). HE staining of the transverse histological sections of morphants revealed disrupted and dilated glomerular structure and dilated proximal tubules (fig. 3a). The Tg(wt1b:GFP) transgenic zebrafish [7], in which GFP is driven by the wt1b promotor so that GFP recapitulates endogenous expression of wt1b, was used to examine the pronephric structure. The Tg(wt1b:GFP) fish showed glomerular cyst formation after wnt5a knockdown (fig. 3b). Cilia staining showed disordered pronephric cilia in wnt5a knockdown zebrafish compared to the injection controls (fig. 3c).

To determine whether Wnt5a plays a role in mouse metanephric kidney development, we performed wmISH of Wnt5a in E10.5 and E11.5 IM, and E13.5 and E15.5 metanephros. wmISH of the Ret gene expression was used to identify the ND and UB structure (fig. 4a-d). At E10.5, Wnt5a mRNA was diffusely expressed in the entire IM with higher expression level caudally (fig. 4e, g). At E11.5 when the UB starts to branch, Wnt5a expression increased significantly in the gonad and mesonephric region, however, not in the metanephric region where the UB starts to initiate (fig. 4f, h). At E13.5 and E15.5, Wnt5a was expressed in the metanephros (fig. 4i, j) with a pattern similar to Ret expression (fig. 4c, d), indicating uretic bud expression of Wnt5a.

Wnt5a^-/- mice were used to assess the loss of Wnt5a expression on kidney development. Since Wnt5a^-/- mice are perinatal lethal, an MRI was performed on E16.5 embryos, and showed kidneys in the wild-type mice (fig. 5a), but not in the Wnt5a^-/- embryo (fig. 5b). We then dissected multiple wild-type and Wnt5a^-/- embryos and identified several different, but severe, phenotypes (fig. 5c-f, online suppl. fig. 1; for all online suppl. material, see www.karger.com/doi/10.1159/000368411, and table 1). Histology (fig. 5g) of the Wnt5a^-/- kidneys showed markedly reduced kidney development.

wmISH of the Ret gene was used to identify the ND and UB structure. In wild-type embryos, UB initiated at the caudal region of the IM, and started to branch at E11.5 (fig. 6a, d). The mutant embryos had truncated IM and abnormal UB initiation, such as medial dislocation of the outgrowth sites (fig. 6e), delayed UB branching (fig. 6e) and outgrowth at more than two sites (fig. 6c, f). By E15.5, the Ret expression was localized at the branched UB (fig. 6h); unilateral hypoplasia and duplex kidney and ureter were seen in Wnt5a^-/- mutant embryos (fig. 6i, j).

Pleotropic kidney defects, including cystic kidney disease [2], hypoplastic kidney [13], hydronephrosis and nephrocalcinosis [14] have been reported in patients with Robinow syndrome. Wnt5a has been found to play important roles in these patients' craniofacial and skeletal abnormalities. Our study utilized two different animal models, zebrafish and mouse, to demonstrate the role of Wnt5a during normal kidney development, and to help identify the causes of kidney abnormalities in Robinow syndrome patients.

Wnt5a knockdown in zebrafish resulted in glomerular cyst formation. Wnt5a functions in planar cell polarity (PCP) regulation in mice [15], and disruption in the PCP pathway has been shown to result in kidney cyst formation [16]. The abnormal body axis curvature with downward-curving tail in our wnt5a morphants is also seen in zebrafish with defects in cilia motility [17, 18] or assembly [19]. Wnt5a morphants had disordered pronephric cilia, suggesting that wnt5a might work through ciliogenesis and PCP to control normal kidney development. It is also possible that the pronephric tubular dilation could have played a role in the ciliary phenotype. Further investigations will be conducted to establish the relationship among wnt5a, cilia and PCP.

We found pleiotropic abnormal renal phenotypes in Wnt5a^-/- mutant mice (table 1). There are many examples of pleiotropic renal phenotypes in knockout mice. For instance, only one third of newborn Ret knockout mice showed complete absence of ureters and kidneys, and one tenth had bilateral kidney rudiments [20].

The cause of the duplex kidney and ureter was likely due to Wnt5a controlling IM extension [21] and regulating the interaction of ND and MM [22]. Besides duplex ureter/kidney, our data also demonstrated bilateral and unilateral kidney agenesis in Wnt5a knockout mice. Our Wnt5a knockout mice had truncated IM, medial dislocation of the UB outgrowth sites and delayed UB branching. Dislocated UB outgrow sites can cause a dysregulated interaction between ND and MM and result in kidney agenesis. Delayed UB branching results in kidney hypoplasia. It is interesting that Wnt5a strongly expresses at the mesonephros, but does not strongly express in the UB outgrowth area, indicating an important role of the mesonephros in the development of metanephros. Wnt5a may play a role in UB branching morphogenesis, but is likely not involved in the mesenchymal epithelial transformation or S-shaped body cell differentiation because these structures develop relatively normally in the Wnt5a null kidneys. Another notable phenotype of Wnt5a global knockout mice is fused kidneys. The fused kidney is secondary to defects in capsule formation, which prevents the kidneys from detaching from the body wall. The reason why loss of Wnt5a expression results in fused kidneys is not clear, though the medial dislocation of the UB outgrowth sites might be one of the causes of fused kidneys. FoxD1, a transcription factor expressed in the stromal mesenchymal region of the developing kidney, plays a role in controlling cellularity in the renal capsule [23]. Whether Wnt5a works through a Fox signaling pathway or is downstream of FoxD1 signaling remains to be seen.

Wnt5a regulates skeletal morphogenesis, and MRI of the mouse embryo showed significant skeletal abnormalities. Our previous work showed that Wnt5a^-/- mice had significant urogenital abnormalities, which may be the reason for the hydronephrosis phenotype seen in the kidney. It is difficult to determine cyst formation in this Wnt5a mutant model because of perinatal lethality and urogenital abnormalities; therefore, kidney-specific Wnt5a knockout mice are needed, which we are currently generating.

In summary, the present study establishes a key role for Wnt5a during early kidney development, and allows it to join the large group of Wnts that are involved in different stages of metanephrogenesis. Whether Wnt5a is involved in ciliogenesis and PCP signaling needs further investigation. Our finding helps us to understand the mechanism of kidney and urinary anomalies in patients with Robinow syndrome [1, 3]. Interestingly, defective wnt5a signaling causes renal cyst formation in our zebrafish model and cystic kidney disease has also been reported in patients with Robinow syndrome [24]. Our future efforts will be directed towards studying kidney-specific Wnt5a knockout mice, which should allow the mice to survive long enough to fully manifest renal defects.

This study was funded by Veteran Affair Merit Award to J.H.L, NIH (DK069909 and DK047757) to J.H.L, DK093625 to L.H, Satellite Healthcare (Norman S. Coplon Extramural Research Grant) to J.H.L, and University of Pennsylvania Translational Medicine Institute (Pilot Grant) to J.H.L.

The University of Pennsylvania Biomedical Imaging Core Facility of the Cancer Center is acknowledged for providing imaging services, and the University of Pennsylvania Zebrafish Core for providing essential services such as maintenance and breeding of fish. Dr. F. Perry Wilson is gratefully acknowledged for assistance with statistical analyses. Mrs. Andrea Wecker and Dr. Michael Solhaug are gratefully acknowledged for reviewing the paper and offering helpful comments and suggestions.

There are no competing financial interests.