The Wnt family of signaling proteins in humans has 19 identified members, with the major subgroup of Wnts (WNT1, WNT3A, WNT8) signaling through activation of β-catenin dependent transcription (Nusse,2005). This canonical, or β-catenin–dependent, Wnt signaling pathway functions primarily to activate cell proliferation and cell fate change during development. A noncanonical, β-catenin–independent, signal transduction pathway that controls cell polarity or movement has also been identified (Heisenberg et al.,2000). Wnt5a has been shown to signal in a noncanonical manner, modulating cellular movements independent of β-catenin (Slusarski et al.,1997; Heisenberg et al.,2000). Noncanonical Wnt signaling is necessary for the directional cell migration of pancreatic islet cell progenitors during pancreas formation in zebrafish and mice (Kim et al.,2005). In addition, noncanonical Wnt5a signaling regulates directional cell migration necessary for secondary palate fusion during mouse development (He et al.,2008). Noncanonical Wnt signaling is an area of active research in developmental biology, and it may involve multiple downstream pathways. Evidence exists for a Wnt5a signaling cascade involving the intracellular activation of calcium/calmodulin-dependent protein kinase and protein kinase C by means of the Frizzled2 transmembrane receptor, resulting in Ca2+ fluxes (Kohn and Moon,2005). Other studies suggest that Wnt5a can signal through the orphan tyrosine kinase receptor, Ror2, but there may be multiple downstream mediators of this ligand-receptor complex (Oishi et al.,2003; Mikels and Nusse,2006; Schambony and Wedlich,2007). In the work presented here, we demonstrate that mutations in WNT5A are associated with human phenotypes similar to those found with loss-of-function ROR2 mutations (Afzal et al.,2000; van Bokhoven et al.,2000), suggesting a role for this newly identified pathway in human development and disease.
In 1969, Meinhard Robinow and colleagues described a human syndrome characterized by short stature, mesomelic limb shortening, hypertelorism, mandibular hypoplasia, irregular dental alignment, and hypoplastic external genitalia (Robinow et al.,1969). Based on the initial pedigree, Robinow syndrome was recognized as an autosomal dominant inherited syndrome [MIM 180700] with high penetrance. Over 100 patients with Robinow syndrome have since been identified in families with both autosomal dominant and autosomal recessive inheritance patterns (Patton and Afzal,2002).
The autosomal recessive form of Robinow syndrome [MIM 268310], which is characterized by more severe skeletal, vertebral, and craniofacial abnormalities (Patton and Afzal,2002; Mazzeu et al.,2007b) is often caused by loss-of-function mutations in the gene encoding the tyrosine kinase-like orphan receptor 2, ROR2 (van Bokhoven et al.,2000; Afzal et al.,2000). Recently, Ror2 has been identified as a putative receptor for Wnt5a (Mikels and Nusse,2006; Schambony and Wedlich,2007). Wnt5a and Ror2 are expressed in adjacent and partially overlapping domains during mouse embryogenesis, and Wnt5a can directly bind to the extracellular cysteine-rich domain of Ror2 (Nomi et al.,2001; Oishi et al.,2003; Schleiffarth et al.,2007). Wnt5a null mice and Ror2 null mice exhibit phenotypes grossly similar to those found in Robinow syndrome patients, including shortening of the anterior–posterior axis, facial dysmorphism, genital hypoplasia, and cardiac defects (Yamaguchi et al.,1999; DeChiara et al.,2000; Oishi et al.,2003; Schleiffarth et al.,2007). Wnt5a null mice have a more pronounced phenotype compared with Ror2 null mice (Oishi et al.,2003), but these differences can be explained by functional compensation by the related gene Ror1 in Ror2 nulls (Nomi et al.,2001). Ror1/Ror2 double mutant mice exhibit a Robinow syndrome-like phenotype that more closely resembles Wnt5a null mice (Yamaguchi et al.,1999; Nomi et al.,2001; Oishi et al.,2003). Wnt5a null mice exhibit a phenotype that is more severe than human dominant Robinow syndrome patients with perinatal lethality, 100% penetrance of cardiac defects, and presence of rib fusions (Yamaguchi et al.,1999; Schleiffarth et al.,2007). Heterozygous mutations in ROR2 that cause brachydactyly type B [MIM 113000] do not result in simple loss-of-function like those ROR2 mutations described in recessive Robinow syndrome. Rather, the ROR2 mutations reported in patients with brachydactyly type B are thought to result in dominant-negative mutations (Oldridge et al.,2000; Schwabe et al.,2000; Stricker et al.,2006). These studies suggest that mutations in other genes in a ROR2 signal transduction pathway may be associated with dominant Robinow syndrome. We tested the hypothesis that mutations in WNT5A are associated with autosomal dominant Robinow syndrome, thereby identifying an important role for this signaling pathway in human craniofacial and skeletal development.
Wnt5a null mice and Ror2 null mice exhibit phenotypes similar to those found in Robinow syndrome patients, including shortening of the anterior–posterior axis, facial dysmorphism, genital hypoplasia, and cardiac defects (Robinow et al.,1969; Afzal et al.,2000; van Bokhoven et al.,2000; Oishi et al.,2003; Schwabe et al.,2004; Mazzeu et al.,2007b). Figure 1 illustrates the phenotypic characteristics of this well-described mouse model and patients with a clinical diagnosis of dominant Robinow syndrome (Robinow et al.,1969; Yamaguchi et al.,1999; Oishi et al.,2003; Mazzeu et al.,2007b). These observations led us to screen for WNT5A mutations in patients with dominant Robinow syndrome.
Our initial screening for WNT5A mutations in patients with dominant Robinow syndrome focused on the family in which the syndrome was initially described (Robinow et al.,1969). In that report, the skeletal, genital, and craniofacial abnormalities of four living affected family members were described in detail. We obtained DNA from the original four patients initially described by Dr. Robinow and colleagues, and from all additional living family members both affected and unaffected (Fig. 2C). Family members not previously reported (Generation V), who are offspring of the affected female sibling of the proband described in the original report (Robinow et al.,1969), manifested mesomelic limb shortening, short stature, and characteristic facial features. The affected male family members have no children.
Polymerase chain reaction (PCR) amplification (Table 1) and sequencing of the coding regions of WNT5A in this family revealed heterozygous missense mutations in two sequential base pairs in exon 4 of WNT5A. This mutation of WNT5A, 544-545CT→TC (Fig. 3D) results in an amino acid substitution, C182R, at a highly conserved cysteine residue (Fig. 3B). This mutation is present in all seven affected family members and not present in the only living unaffected family member, (III-3) (Fig. 2C and data not shown). This mutation is in a highly conserved region of the WNT5A protein (Fig. 3D) and was not identified in 196 unaffected control samples by ARMS PCR or sequencing, or by review of human single nucleotide polymorphism (SNP) databases (http: //hapmap.org/, GeneCards genome: http://www.stanford.edu/cgi-bin/gene cards/index.shtml, and UCSC Genome Browser http://www.genome.ucsc. edu/), which showed no human SNPs at WNT5A 544-545. No other mutations in the coding regions of WNT5A were identified in this family (Fig. 2C).
Table 1. WNT5A Primer Sequences
Forward Primer (5′-3′)
Reverse Primer (5′-3′)
Linkage analysis for this family (see pedigree Fig. 2) was performed concurrent with WNT5A exon sequencing by using a total of 382 fluorescently labeled short tandem repeat (STR) polymorphic markers (ABI PRISM Linkage Mapping Set MD10.A). All of the affected family members (Fig. 2) showed linkage with a logarithm of odds (LOD) score of 1.78 at chromosome 3 in a region containing the human WNT5A locus. A LOD score of 1.78 is the maximum predicted score achievable with the number of meioses (individuals) available for this pedigree. Importantly, linkage to other candidate genes implicated in WNT5A signaling, including ROR2 and the WNT5A receptor FRIZZLED2, was excluded by this analysis.
We identified a second heterozygous missense WNT5A mutation 248G→C in an unrelated adult patient with a sporadic dominant Robinow syndrome phenotype using the same methods. This mutation predicts an amino acid substitution (C83S; Fig. 3F) affecting a second highly conserved cysteine residue (Fig. 3A). This patient's phenotype included marked hypertelorism, short nose, short stature, and mesomelic shortening of the limbs. The parents were unaffected clinically; however, no parental DNA is available for testing. Sequence data from 173 control DNA samples did not show the 248G→C missense mutation in exon 3, nor was it found in the SNP databases as described above. Coding regions of WNT5A were sequenced in 23 additional unrelated patients with clinical diagnosis of dominant Robinow syndrome, and no WNT5A mutations resulting in amino acid substitutions were identified. This suggests that mutations in the coding regions of WNT5A cause only a subset of this syndrome as defined by current clinical criteria (Mazzeu et al.,2007b).
Mutations of cysteines in Drosophila wingless (wg), and mammalian Wnts negatively affect Wnt function in multiple assay systems (McMahon and Moon,1989; Mason et al.,1992; van den Heuvel et al.,1993; Willert et al.,2003). We hypothesize that mutating any of the 24 conserved cysteines might inhibit WNT5A function by altering protein folding, secretion, palmitoylation, or receptor binding. PolyPhen (Polymorphism Phenotyping) protein structure modeling (http://genetics.bwh.harvard.edu/pph/) supports our prediction that the WNT5A mutations identified in our study affect WNT5A function. Based on this modeling, the calculated PSIC software score difference of 3.151 and 4.513 for the WNT5AC83S and WNT 5AC182R mutations, respectively, is characterized as “probably damaging.” Therefore, both the WNT5AC83S and WNT5AC182R mutations are predicted to, with a high level of confidence, affect protein function.
The functional relevance of the WNT5A mutations C182R and C83S was tested in vivo by overexpression of mRNA in zebrafish embryos. We have previously shown that wnt5 signaling is required for normal pancreatic islet cell coalescence in zebrafish and mice and that this process requires a gradient of signal along the anterior–posterior axis in zebrafish (Kim et al.,2005). Disruption of this signaling gradient by mRNA overexpression of human WNT5A results in a failure of insulin-expressing cell coalescence into a pancreatic islet (Fig. 4D,F) when compared with control embryos (Fig. 4B,E). Overexpression of WNT5AC182R mRNA (Fig. 4G,H) and WNT5AC83S mRNA (not shown) are both less effective at disruption of islet cell coalescence than overexpression of wild-type WNT5A mRNA in this assay of Wnt signaling function (WNT5AC182R vs control, P = 0.00005; WNT5AC83S vs. control, P = 0.002; Fig. 4K). These data suggest that both the C182R and C83S mutations create phenotypes due to differential protein stability or expression of less active (hypomorphic) forms of WNT5A.
WNT5A can precociously activate canonical, β-catenin–dependent Wnt signaling during early axis formation in frogs (He et al.,1997). Although this scientific scenario is an artificial test for WNT5A because WNT5A does not perform this function in vivo, it is nevertheless a test of biochemical potential for activation of canonical Wnt signaling. We tested the ability of both wild-type WNT5A and WNT 5AC182R mRNAs to induce duplicate axis formation upon injection into zebrafish embryos, and we noted a similar incidence of duplicated embryonic axes when assayed at 24 hours postfertilization (hpf; data not shown). Interestingly, and in sharp contrast to the similarly reduced signaling ability in a noncell fate Wnt pathway, injection of WNT5AC83S mRNA exhibited a significantly lower incidence of duplicated embryonic axes compared with wild-type injected embryos (data not shown). This unexpected result suggests a possible change in the biochemical potential for canonical Wnt signaling in this WNT5A sequence variant. As we do not yet understand the rules whereby a cell distinguishes between the canonical and noncanonical signaling potential of Wnt proteins, this observation suggests a loss of a single cysteine residue can serve as a major delineating structural feature.
To specifically test the hypothesis that these human variants generate antimorphic proteins, we compared the injections of these human variants with injections of mRNA encoding a dominant-negative WNT5A protein form. Injections with a C-terminal truncation, dominant-negative mutant, (dn) WNT5A (Sen et al.,2001), did not cause axis duplication, but instead produced phenotypes similar to the Wnt5 loss of function embryos (pipetail; Hammerschmidt et al.,1996) (Fig. 4J). These data suggest that the C83S and C182R mutations are not acting as dominant-negative inhibitors of Wnt signaling, but they are instead acting as partial loss of function or hypomorphic alleles of WNT5A.
Additional studies were performed to test the function of these mutations in altering convergent-extension movements during Xenopus gastrulation (Moon et al.,1993) and to place them in a novel ROR2-mediated noncanonical Wnt pathway (Schambony and Wedlich,2007). In1993, Moon et al showed that Xwnt5A blocks activin-mediated elongation of blastula stage animal caps without altering mesoderm formation (Moon et al.,1993). To determine the effect of the WNT5AC182R and WNT5AC83S mutations on cell movement in Xenopus embryos, we compared control animal caps isolated at prebastula stages with control animal caps exposed to activin protein, and to animal caps harvested from embryos that had been injected at the two-cell stage with either moderate dose (30 pg) or high dose (200 pg) HWNT5A, or WNT5AC182R, and WNT5AC83S mutant constructs (Fig. 5) and then exposed to activin at blastula stages. Nonactivin treated uninjected embryos (Fig. 5A) and embryos injected with 200 pg of WNT5AC182R mRNA (Fig. 5C) are also shown. These experiments demonstrated reduced ability of the mutant forms of WNT5A to block activin-mediated cell movement during gastrulation at moderate doses (Fig. 5D–F), however, at higher doses all three forms were effective at blocking activin-mediated cell movement and animal cap elongation (Fig. 5G–I). This observation confirms that while both the WNT5AC182R and WNT5AC83S mutations have reduced activity in this assay system, some functional ability to activate WNT5A signaling is retained and supports our hypothesis that these mutations represent an alteration of WNT5A function due to a hypomorphic allele, and not a dominant-negative mechanism of action.
Recent studies have demonstrated Wnt5A/ROR2 mediated activation of noncanonical Wnt signaling pathways and of Xenopus paraxial protocadherin (XPAPC) (Schambony and Wedlich,2007). We used a simple assay of XPAPC activation as evidenced by induction of mRNA expression to determine if the mutations described in patients with Robinow syndrome acted through the WNT5a/ROR2 signaling pathway. We found that in this assay, the WNT5AC83S mutation had a significantly reduced ability to induce XPAPC expression when compared with wild-type HWNT5A (Fig. 6) and that the WNT5AC182R mutation trended to reduced ability to alter XPAPC expression when compared with wild-type mRNA, without reaching statistical significance (data not shown).
The overlapping phenotypes displayed by the dominant and recessive forms of Robinow syndrome implicate WNT5A as a ligand for the tyrosine kinase receptor ROR2. Ror2 function in noncanonical Wnt signaling is newly described during Xenopus gastrulation (Schambony and Wedlich,2007), as no prior experimentation has demonstrated a role for the highly conserved kinase domain of Ror2 in any known Wnt-dependent process (Kim and Forrester,2003; Oishi et al.,2003; Mikels and Nusse,2006). The array of mutations in ROR2 associated with recessive Robinow syndrome demonstrates a role for the conserved kinase-like domain of ROR2 in normal human development (Schwabe et al.,2000). This work, describing WNT5A mutations in dominant Robinow syndrome, supports a noncanonical signaling model in which a Wnt ligand signals by means of a tyrosine kinase receptor, and implicates the WNT5A/ROR2 pathway in human craniofacial, skeletal and genital development.
In this work, we describe WNT5A coding sequence variations in the original family in which Robinow syndrome was described and a second unrelated patient. Robinow syndrome characteristics include mesomelic shortening of the limbs, a flat facial profile, prominent forehead, and hypertelorism (Robinow et al.,1969). Recently, Wnt signaling has been shown to be a major factor controlling differences in facial development between chicken and mice (Brugmann et al.,2007). Comparisons of the dominant and recessive forms of Robinow syndrome suggest that, during normal development, the concentration of ligand governs the degree of limb outgrowth and continued development of facial structures. Humans with heterozygous ROR2 loss-of-function mutations are unaffected carriers (Afzal et al.,2000; van Bokhoven et al.,2000). Heterozygous Ror2 null mice are also phenotypically normal (DeChiara et al.,2000; Takeuchi et al.,2000; Schwabe et al.,2004). Loss-of-function mutations in both copies of ROR2 are necessary for the more severe phenotypic manifestation of recessive Robinow syndrome (Afzal et al.,2000; van Bokhoven et al.,2000). One normal functioning copy and one copy of WNT5A with reduced function are associated with dominant Robinow syndrome. Despite this phenotype arising from one partially functioning WNT5A allele in humans, external examination of heterozygous mice in late gestation revealed none of the characteristic craniofacial, skeletal, or growth abnormalities associated with the homozygous Wnt5a−/− mutation. There is certainly precedent for a disease-causing heterozygous mutation in humans to be grossly undetectable in mice, and for mouse phenotype to vary with genetic background, as demonstrated in DiGeorge Syndrome (22q11 deletion/TBX1 mutation; Jerome and Papaioannou,2001) and holoprosencephaly (SHH mutation; Belloni et al.,1996; Roessler et al.,1996).
The molecular abnormalities described here in dominant Robinow syndrome suggest a model by which the WNT5A/ROR2 signal transduction pathway is regulated by the local concentration of the ligand and the receptor concentration is not rate limiting. In this report, we describe attenuation of WNT5A function associated with defects in limb outgrowth and abnormal craniofacial morphogenesis. This suggests that modest activity changes in WNT5A could be a mechanism for some normal variation of limb and facial development in humans.
Several reasons may explain why WNT5A mutations were detected in only a subset of patients. First, because we used direct sequencing of the WNT5A coding regions, our analysis would not detect mutations in the promoter or other regulatory elements, or other large deletions or duplications. Subsequent to our sequence analysis, an alternatively spliced form of exon 1 has been identified and confirmation and future sequencing of this exon would be appropriate in patients with clinically diagnosed dominant Robinow syndrome (Katoh,2009). Second, it is likely that dominant Robinow syndrome is genetically heterogeneous, and mutations in other components of the WNT5A/ROR2 signaling or regulatory pathways may also be responsible for dominant Robinow syndrome. Recently, Mazzeu et al. have identified abnormalities of chromosome 1 associated with limb, skeletal, genital, and craniofacial characteristics commonly found in Robinow syndrome (Mazzeu et al.,2007a). Potential genetic heterogeneity is confounded by difficulties in accurate clinical diagnosis of craniofacial and skeletal syndromes. Recessive Robinow syndrome has unique characteristic rib and vertebral anomalies and pronounced limb and craniofacial abnormalities that improve clinical diagnostic accuracy and are well correlated with ROR2 mutations (Mazzeu et al.,2007a). In contrast, the phenotype of dominant Robinow syndrome is reported to be more subtle, with less severe skeletal and craniofacial changes when compared with the recessive phenotype. Identification of a functional mutation in WNT5A in the initially described family with dominant Robinow syndrome sets a phenotypic standard with molecular correlation that will help us begin to sort out genotype/phenotype correlations and identify molecular etiologies in Robinow syndrome. The knowledge developed through the characterization of ROR2 mutations in patients with recessive Robinow syndrome has been limited by the paucity of information about the role of this molecule in development. The identification of WNT5A mutations in dominant Robinow syndrome patients places this WNT5A/ROR2 signaling pathway in an important role in craniofacial and skeletal development and disease.
Approval for this study was obtained from the Institutional Review Boards of the participating institutions (University of Minnesota IRB Code #0603 M83366 and Radboud University Nijmegen Medical Center). Patients with Robinow syndrome were identified based on clinical phenotype (Mazzeu et al.,2007b) and pedigree consistent with autosomal dominant transmission by one of our investigators (S.B., J.F.M., H.G.B.). All living members of the original family described by Dr. Robinow were enrolled. The patient with the 248G→C mutation is a female with a typical phenotypic presentation of the autosomal dominant Robinow syndrome phenotype. Informed consent was obtained from the patient, parent, or guardian in all cases, and 5 ml of blood were obtained by venipuncture. DNA was isolated with the VersaGene Genomic DNA Purification Kit (Gentra Systems, Inc., Minneapolis, MN) using a standard protocol for DNA extraction from whole blood. Control DNA for complete WNT5A coding sequencing was obtained from phenotypically normal unrelated individuals who consented to participate in this study and was isolated from whole blood as above.
Wnt5a Genomic PCR
A total of 100 ng of DNA were used in PCR reactions performed with intronic WNT5A primer pairs (Table 1) spanning across each individual exon for all patient samples and controls. Negative control reactions for each primer set were run without DNA template to assay for DNA contamination (Table 1).
PCR products were separated on a 1.5% agarose gel, and amplicons were gel extracted using the S.N.A.P. Gel Purification Kit (Invitrogen, Carlsbad, CA). A total of 30 ng of purified DNA was then sequenced by our on site facility. A DNA ABI PRISM 3730xl DNA Analyzer was used in sequencing reactions of PCR products using dye-labeled terminator chemistry. Sequencing was conducted in both 5′ and 3′ directions using the same primers used for the PCR reactions (Table 1). Sequence variants were confirmed by re-sequencing in both direction three times, and compared with verified sequence of an additional 173 controls for exon 3 obtained from phenotypically normal individuals (J.L.L. and H.G.B.). One hundred ninety-six controls from normal banked DNA were also analyzed for the mutation identified in exon 4 by ARMS-PCR (179 controls; Perrey et al.,1999) or sequencing (17 controls). Primers used for ARMS-PCR recognized wild-type WNT5A exon 4: (5′-GGACTGGC TCTGGGGCGGCT-3′, 5′-GAGGAGAG GACGGAGCTACA-3′) and the C182R mutation in WNT5A exon 4: (5′-GG ACTGGCTCTGGGGCG GTC-3′,5′-GA GGAGAGGACGGAGCT ACA-3′).
A total of 40 ng of genomic DNA was used in ARMS-PCR analysis. PCR parameters were 94°C for 30 sec, 62°C for 60 sec, 72°C for 60 sec, loop 30 times, 72°C for 10 min.
Linkage analysis was performed using a total of 382 fluorescently labeled STR polymorphic markers (ABI PRISM Linkage Mapping Set MD10.A) with easyLINKAGE Plus pairwise and multipoint comparison. Data were analyzed using GeneHunter v2lr5 software, with parametric analysis for a dominant model of inheritance. We assumed complete penetrance, and a frequency of the disease allele of 1:10,000.
Genotyping of Mice
Embryos used in this study were obtained by intercrossing Wnt5a−/− mice (Yamaguchi et al.,1999). Wild-type littermates were used as controls. Genomic DNA was PCR amplified as described (Kim et al.,2005). The study has been approved by the University of Minnesota Animal Care and Use Committee.
Skeletal Preparation and Histology
The skeletons of wild-type and Wnt5a−/− embryos were stained with Alizarin red and Alcian blue as described (Hogan et al.,1986).
WNT5A, WNT5AC83S, WNT5AC18 2R, and dnWNT5A were PCR engineered from cDNA clone (catalog no. MHS1010-9204179, Open Biosystems, Huntsville, AL) incorporating a 5′ BglII site, a Kozak translation site (GCCACC), start codon, coding sequence, stop codon, and a 3′ SpeI restriction site. These constructs were ligated into pT3TS (Hyatt and Ekker,1999) using BglII and SpeI engineered restriction sites. To make the dnWNT5A, a stop codon was inserted after the SPDYC motif as described (Sen et al.,2001). The WNT5A mutant constructs were generated by PCR using overlapping primers containing the mutated bases (base pairs 544-545, CT/TC) or (base pair 248, G/C) to incorporate the mutations into full-length constructs of WNT5AC 182R and WNT5AC83S, respectively. All clones where sequence-verified from both directions with primers flanking the mutation site (not shown). All four clones (pT3TS WNT 5A, pT3TS WNT5AC182R, pT3TS WNT5AC83S, and pT3TS dnWNT5A) were linearized with XbaI and 5′ (7-methyl guanosine) capped mRNA was transcribed with T3 RNA polymerase (T3 Message Machine, Ambion, Austin, TX).
Zebrafish Maintenance and Injections
Wild-type Danio rerio (Segrest Farms, Gibsonton, FL) and insulin-enhanced green fluorescent protein (eGFP; Huang et al.,2001) embryos were raised at 30°C and spawning was performed as previously described (Kimmel et al.,1995). WNT5A, WNT 5AC83S, WNT5AC182R, and dnWN T5A mRNAs were injected into one-cell embryos. Axis duplication was scored at 24 hpf in wild-type embryos. Insulin-eGFP embryos were injected with mRNAs and scored for islet formation defects at 30 hpf.
Whole-Mount In Situ Hybridization
pCRIITOPO Danio rerio insulin clone containing base pairs 38-369 NCBI accession no. NM_131056 was linearized with NotI and digoxigenin-labeled riboprobe was synthesized with T7 RNA polymerase (Roche, Mannheim, Germany). Whole-mount in situ hybridization was performed as described (Jowett,1999).
Xenopus Embryos and Injections
Xenopus embryos were obtained by in vitro fertilization after induction of ovulation in female frogs using human chorionic gonadotropin (Sigma). Animal cap experiments were performed as described in (Moon et al.,1993). Briefly, the animal poles of both blastomeres were injected with 30 or 200 pg of HWNT5A, WNT5AC83S mRNA or WNT5AC182RmRNA at the two-cell stage. The upper one-fourth of the embryo was harvested using an eyebrow knife at blastula stage (stage 8) and cultured in 1/3 strength modified Ringer's solution and 5 ng/ml activin protein (R&D Systems, Minneapolis, MN). Embryos were assayed for elongation when control embryos were at stage 40. For XPAPC induction, the two dorsal blastomeres were injected with 100 pg of mRNA at the four-cell stage and embryos rapidly frozen at stage 10.5 (early gastrulation) for RNA isolation, cDNA synthesis and quantitative PCR. Methods were adapted from Schambony and Wedlich (2007).
XPAPC Induction by Quantitative PCR
XPAPC induction in injected Xenopus embryos was assessed by quantitative PCR (qPCR). Total RNA was isolated from embryos using TRIzol reagent (Invitrogen Life Technologies). cDNA was synthesized and amplifications were performed in triplicate with 1× RT2 Real-Time SYBR green mix (SuperArray) on a Stratagene Mx 3000P Real-Time PCR system with the following parameters: 45 cycles, 95°C for 30 s, 55°C for 60 s, and 72°C for 60 s. qPCR data were normalized using ornithine decarboxylase (ODC) as a control. Primer sequences (Schambony and Wedlich,2007) for XPAPC were as follows: 5′-cccagtcgg tctcttcttctttg-3′ (forward), 5′-ttgctgatg ctgctcttggttag-3′ (reverse), and ODC 5′-gccattgtgaagactctctccattc-3′ (forward), 5′-ttcgggtgattccttgccac-3′ (reverse). Data are presented as the mean ± SE using GraphPad Prizm4.
The transgenic insulin-eGFP experimental values are the individual scores from each embryo. The transgenic insulin-eGFP–positive cell coalescence experimental mean was then calculated for each experiment. The mean value of the means of each individual experiment was calculated [(sum of individual experiment means) / number of individual experiment means = mean of the means]. The mean variance of the means was also calculated. The percentage of embryos injected with wild-type WNT5A mRNA in pancreatic islet coalescence and axis duplication assays was normalized to 100% activity and the percentage activity of WNT5AC182R and WNT5AC83S mRNA-injected embryos was compared with wild-type WNT5A mRNA injected activity.
We used Student's t-test to compare inhibition of insulin-eGFP cell coalescence in embryos injected with WNT5A mRNA compared with embryos injected with WNT5AC83S, or WNT5AC182R mRNA. P values < 0.05 were considered significant.
We dedicate this manuscript to the memory of Dr. Ian N. Jongewaard. We thank our patients and their families for their participation in and encouragement of this study. We thank The Robinow Syndrome Foundation for assistance with this research effort and Drs. Harry Orr, Saulius Sumanas, Hyon Kim, and William Oetting for their scientific comments. This research was supported by Minnesota Medical Foundation Research Grants to A.P. and J.R.S.; Howard Hughes Fellowship to J.R.S.; a gift from the Sit Investment Associates Foundation to J.L.L.; and FAPESP grant support for J.F.M. Drs. Anthony Person and Mara Robu were supported by the MinnCResT Training Program and C.J.B. was supported by the Musculoskeletal Training Grant NIH-NIAMS.
Accession numbers and URLs for data presented herein are as follows:
Study design by S.C.E., S.B., J.L.L., and L.A.S. Human subjects approval obtained by J.L.L. and H.G.B. Sample retrieval and phenotyping done by J.H., S.B., J.F.M., J.L.L., and H.G.B., A.D.P., C.M.S., A.N.N., S.H. were involved in PCR reactions and sequencing of PCR products. Generation of constructs pT3TSWNT5A, pT3TSWNT5AC182R, pT3TSWNT5A C83S, and pT3TSdnWNT5A by ADP. Zebrafish experiments done by ADP and MER. Xenopus experiments and qPCR were done by A.N.N., C.M.S., C.J.B., and J.L.L. Mouse data and Figure 1 by J.R.S., A.P., and M.E.R. Linkage analysis done by H.G.B. and HVB. Manuscript written by A.D.P., S.C.E., J.L.L., H.G.B., H.V.B., L.A.S., and J.F.M. All authors have agreed to all content in the manuscript, including the data presented. J.R.S. was a Howard Hughes Medical Institute Medical Student Research Training Fellow, C.J.B. was funded by the Medical Scientist Training Program of the University of Minnesota, and J.F.M. was funded by the Robinow Syndrome Foundation, Andover, Minnesota. The authors declare that they have no competing financial interests.