Regulation of cell polarity in the cartilage growth plate andperichondrium of metacarpal elements by HOXD13 and WNT5A

Pia Kuss a,b,1, Katerina Kraft a,b, Jürgen Stumma, Daniel Ibrahim a,b,c,Pedro Vallecillo-Garcia a, Stefan Mundlos a,b,c,n, Sigmar Stricker a,b,nn

a Max Planck Institute for Molecular Genetics, Ihnestrasse 63-73, 14195 Berlin, Germanyb Institute for Medical and Human Genetics, Charité University Medicine, Augustenburger Platz 1, 13353 Berlin, Germanyc Berlin-Brandenburg Center for Regenerative Therapies (BCRT), Charité University Medicine, Augustenburger Platz 1, 13353 Berlin, Germany

a r t i c l e i n f o

Article history:Received 24 June 2013Received in revised form11 October 2013Accepted 14 October 2013Available online 23 October 2013

Keywords:CartilageGrowth platePerichondriumCell polarityHoxd13Wnt5a

a b s t r a c t

The morphology of bones is genetically determined, but the molecular mechanisms that control shape,size and the overall gestalt of bones remain unclear. We previously showed that metacarpals in thesynpolydactyly homolog (spdh) mouse, which carries a mutation in Hoxd13 similar to the humancondition synpolydactyly (SPD), were transformed to carpal-like bones with cuboid shape that lackcortical bone and a perichondrium and are surrounded by a joint surface. Here we provide evidence thatspdh metacarpal growth plates have a defect in cell polarization with a random instead of linearorientation. In parallel prospective perichondral cells failed to adopt the characteristic flattened cellshape. We observed a similar cell polarity defect in metacarpals of Wnt5a�/� mice. Wnt5a and theclosely related Wnt5b were downregulated in spdh handplates, and HOXD13 induced expression of bothgenes in vitro. Concomitant we observed mislocalization of core planar cell polarity (PCP) componentsDVL2 and PRICKLE1 in spdh metacarpals indicating a defect in the WNT/PCP pathway. Conversely theWNT/β-CATENIN pathway, a hallmark of joint cells lining carpal bones, was upregulated in theperichondral region. Finally, providing spdh limb explant cultures with cells expressing either HOXD13or WNT5A led to a non-cell autonomous partial rescue of cell polarity the perichondral region andrestored the expression of perichondral markers. This study provides a so far unrecognized link betweenHOX proteins and cell polarity in the perichondrium and the growth plate, a failure of which leads totransformation of metacarpals to carpal-like structures.

& 2013 Elsevier Inc. All rights reserved.

Introduction

During endochondral ossification mesenchymal progenitorcells condense to form a cartilaginous template (anlage) of thefuture bones. Concomitantly, cells surrounding the condensationform the perichondrium consisting of flattened cells. At either endof the anlage a growth plate is established driving longitudinaloutgrowth of the skeletal element. Here small, round chondro-cytes (resting or reserve chondrocytes) differentiate to proliferat-ing chondrocytes, which form stacks of discoid cells. Finally calls

undergo prehypertrophic differentiation characterized by theexpression of Indian hedgehog (Ihh) and subsequently becomehypertrophic. At the same time cells from the perichondrium formthe first bone collar (cortical bone) around the proliferating/hypertrophic chondrocytes. This stack-like arrangement of prolif-erative cells and its restriction by the perichondrium/bone isconsidered the main mechanical force enabling longitudinalgrowth of long bones as compared to spherical growth of, forexample, wrist bones. The transition of spherical to discoid cellshape is achieved by cell polarization concomitant with lateralintercalation into a column in a process reminiscent of convergentextension movements seen in vertebrate gastrulation (Li andDudley, 2009).

In the past years signaling by components of the WNT family ofsecreted factors have emerged as key players in this event(Romereim and Dudley, 2011). Classically, WNT pathways aredivided into the “canonical” branch, acting via stabilization ofβ-CATENIN, and several alternative pathways (Angers and Moon,2009) such as the WNT/planar cell polarity (PCP) pathway. WNT/PCP was demonstrated to regulate oriented cell divisions and cell

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/developmentalbiology

Developmental Biology

0012-1606/$ - see front matter & 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.ydbio.2013.10.013

n Corresponding author at: Institute for Medical Genetics, Charité UniversityMedicine Berlin, Charité University Medicine Berlin, Augustenburger Platz 1, 13353Berlin, Germany. Fax: þ49 30 450569915.

nn Corresponding author at: Max Planck Institute for Molecular Genetics, Devel-opment and Disease Group, Ihnestrasse 63-73, 14195 Berlin, Germany.Fax: þ49 30 84131385.

E-mail addresses: stefan.mundlos@charite.de (S. Mundlos),strick_s@molgen.mpg.de (S. Stricker).

1 Present address: Sanford-Burnham Medical Research Institute, La Jolla,CA, USA.

Developmental Biology 385 (2014) 83–93

polarity in the chicken growth plate (Li and Dudley, 2009). WNT5Ais a major ligand activating the PCP pathway in the mouse (Kikuchiet al., 2012; Komiya and Habas, 2008). Targeted disruption ofWnt5a leads to severely impaired skeletogenesis (Yamaguchi et al.,1999) with altered growth plate morphology (Yang et al., 2003).Moreover, it was recently demonstrated that WNT5TA controls cellshape during condensation of cartilage in mouse digits via a PCPpathway (Gao et al., 2011).

Hox genes control patterning, differentiation and morphogen-esis during development. HOXD13 is a master regulator of autopodskeletal morphogenesis. In previous studies we examined therole of Hox genes in cartilage and bone formation in the limbsmainly using a Hoxd13 mutant called synpolydactyly homolog(Hoxd13spdh/spdh, hereafter termed spdh) which carries a seven-alanine expansion in HOXD13. In humans identical mutations inHOXD13 lead to synpolydactyly (SPD) (Muragaki et al., 1996). Themutation results in the degradation of HOXD13 protein and thus aloss of function. In addition, mutant HOXD13 interacts with otherco-expressed HOX proteins that contain poly-alanine stretchessuch as HOXA13 and inactivates them resulting in a combinedinactivation of HOX proteins in the autopod (Bruneau et al., 2001;Villavicencio-Lorini et al., 2010). Spdh mice exhibit a transforma-tion of metacarpals (long bones) to carpals (cuboid bones). Spdhmetacarpals have a variable shape that ranges from a longitudinalappearance to almost fully round. Mutant metacarpals invariantlyhave no perichondrium and fail to form any cortical bone. Insteadthey are surrounded by joint-like structures and undergo second-ary ossification similar to the ossification scheme observed incarpal bones (Villavicencio-Lorini et al., 2010).

Here we propose that the underlying cause for this phenom-enon is a defect in polarization and orientation of spdh metacarpalgrowth plate and perichondral cells. Wnt5a�/� mice have asimilar phenotype and we provide evidence that Hoxd13 isupstream of Wnt5a. The defects seen in spdh mice are accom-panied by dysregulation of both the canonical WNT/β-CATENINpathway as well as the WNT/PCP pathway. These data indicate thata local Hox code might influence shaping of skeletal elements byinfluencing cell shape via local induction of factors suchas WNT5A.

Materials and methods

Mice

Spdh mice (Johnson et al., 1998) were obtained from TheJackson Laboratory (Bar Harbor, ME, USA), Wnt5a� /� mice(Yamaguchi et al., 1999) from E. Arenas (Karolinska Institute,Stockholm, Sweden) and Axin2LacZ reporter mice (Lustig et al.,2002) from W. Birchmeier (Max Delbrück Centre, Berlin,Germany). Genotyping was performed as described previously(Albrecht et al., 2002; Lustig et al., 2002; Yamaguchi et al., 1999).

Skeletal preparation, in-situ hybridization, immunolabeling

Skeletal preparation and in-situ hybridization was performedas described previously (Witte et al., 2010). Probes for Hoxd13,Wnt5a and Wnt5b were described in (Albrecht et al., 2002; Witteet al., 2009). For immunolabeling, antigen retrieval was performedusing citrate buffer or high-pH buffer (Dako). After permeabiliza-tion (0.2% Triton X-100 for 15 min) and blocking (5% NGS for 1–5 h) primary antibodies (see Supplementary material) wereapplied at 4 1C overnight and detected with fluorescence-conjugated secondary antibodies (Molecular Probes, Eugene, OR,USA; 1:500) at room temperature (RT) for 1 h. Phalloidin-AlexaFluor-488 staining (Molecular Probes) was done 1:200 for

45 min at RT. Nuclei were stained with 4′,6-diamidino-2-phenyl-indole (DAPI; Sigma-Aldrich, St. Louis, MO, USA). For anti-RCASstaining, additional signal amplification using the Tyramide SignalAmplification system (Perkin-Elmer) was performed according tothe manufacturer's protocol.

Polarity index, golgi apparatus orientation

After phalloidin labeling or differential interference contrastimaging, the maximum and minimum diameter of cells wasdetermined using the AutMess tool integrated in the Zeiss AxioVi-sion software (Zeiss, Oberkochen, Germany). The ratio of theseparameters was used as an index for cell polarization. For Golgiorientation, the angle between the Nucleus–Golgi axis (Golgiapparatus stained via GM130 antibody) and the longitudinal axisof the cartilage element was determined using the AutMess tool(Zeiss). In each case at least three specimens were analyzed.

Real-time PCR

DF1 cells were infected with concentrated viral supernatantscontaining RCAS(A)–Hoxd13WT, RCAS(A)�Hoxd13þ7Ala andRCAS(A)–GFP and grown for 3 or 7 days in DMEM supplementedwith 10% FCS and 2% chicken serum. RNA was extracted withpeqGold Trifast (PEQLAB, Erlangen, Germany). Real-time PCR forWnt5a was performed as described before (Hecht et al., 2007)using three independent repeats for each condition.

Explant cultures

DF1 cells were transfected with RCAS(A)–Hoxd13 or RCAS(A)–Wnt5a together with RCAS(B)–GFP for visualization, or with RCAS(A)–GFP alone as control. Cells were grown for 7 days in DMEMsupplemented with 10% FCS and 2% chicken serum. For implanta-tion, DF1 cells were trypsinized and pelleted. Cells from the pelletwere transferred to a mouth pipet by gentle aspiration. Embryoswere harvested at E13.5 or E14.5. The limb culture protocol wasmodified from (Minina et al., 2001). Forelimbs were dissected andthe explants were placed on cell strainer nets in limb culturemedium (BGJb, 1% BSA, 1% L-Glutamine, 0.2 mg/ml ascorbic acid,50 U/ml PenStrep, 0.05 mg/ml Gentamycin, 0.3 μg/ml Amphoter-icin B) at the liquid–air interface. The epidermis was slit with atungsten needle and DF1 were implanted next to metacarpalcondensations using a mouth pipet. Explants were incubated for4 or 3 days, respectively. After incubation, placement of theimplants was controlled via GFP visualization. Only specimensshowing a strong GFP signal adjacent to metacarpal condensationswere used for further analysis. To localize the implant on sections3C2 antibody labeling and detection with the ABC staining kit(Vector Labs, Burlingame, CA, USA) was used. In-situ hybridizationor quantification of cell polarization was performed in regions nextto explants either on the same or on adjacent sections. In each caseat least three specimens were analyzed.

Results

Hoxd13spdh/spdh and Wnt5a� /� metacarpals show lateralenlargement of condensations, lack of a defined border andcartilaginous fusion

We have shown before that postnatal spdh metacarpals showspherical growth and lack perichondrium and cortical bone(Villavicencio-Lorini et al., 2010). We thus re-analyzed the histo-logical appearance of spdh metacarpal condensations duringdevelopment. This confirmed the lack of perichondrium formation

P. Kuss et al. / Developmental Biology 385 (2014) 83–9384

and was concomitant with a lack of a clearly defined border andbeginning lateral growth of the cartilaginous condensation atE14.5 (Fig. 1A). This in turn led to cartilaginous fusion of con-densations at E16.5 (Fig. 1B). In addition, spdh metacarpals did notform an organized growth plate, at E16.5 chondrocytes throughoutthe condensation were small and appeared roundish (Fig. 1B). Thelateral expansion of cartilage elements and disorganization of thegrowth plate was reminiscent of Wnt5a mutant mice (Yamaguchi

et al., 1999; Yang et al., 2003), so we analyzed both mutants incomparison. An overview of the skeletal phenotypes is depicted inFig. 1C. Spdh mice show appearance of additional condensations,irregular shaping of condensations and proximal joint fusions inthe autopod. In Wnt5a mutants not only the autopod is affectedleading to an overall severe shortening of the limb. In the autopodthe lateral expansion of metacarpal condensations is obvious, notethat Wnt5a� /� mice do not form phalangeal condensations while

Fig. 1. Abnormal expanded shape and lack of lateral restriction in spdh and Wnt5a�/� metacarpals. (A) and (B) Forelimb sections at E14.5 and E16.5 stained withHematoxylin/Eosin and Alcian blue for cartilage matrix. Boxed areas are shown as magnifications. (A) At E14.5 the metacarpal condensations show an apparent lateralexpansion, magnifications reveal the lack of a clearly defined border and a perichondrium in both mutants. (B) At E16.5 condensations in spdh and Wnt5a�/� mutants showcartilaginous fusions; the magnifications also show a complete lack of growth plate organization within those structures. (C) Skeletal preparations of new born spdh andWnt5a�/� mice stained with Alcian blue (cartilage) and Alizarin red (bone). Skeletal preparations are depicted at the same scale, the Wnt5a mutant is additionally shown intwo-fold magnification in the insert. Note that in spdh mice metacarpals are mostly fused to the phalanges. In Wnt5a�/� mice only metacarpal condensations form, thephalanges are missing.

P. Kuss et al. / Developmental Biology 385 (2014) 83–93 85

spdh mice do. Analyzing tissue sections indicated a definedphenotypic overlap specifically in the metacarpal condensations.In E14.5 Wnt5a�/� metacarpals the roundish shape of the con-densations is already obvious (Fig. 1A). Comparable to the spdhmouse no perichondrium is forming and the condensation showsan ill-defined border (Fig. 1A, insets). Moreover, at E16.5 themetacarpals of Wnt5a�/� mice show cartilaginous fusion(Fig. 1B). Magnification of metacarpal condensation also showeda complete lack of regular growth plate cartilage organization(Fig. 1B, insets).

Defective perichondral and growth plate cell polarity inHoxd13spdh/spdh and Wnt5a� /� metacarpals

WNT5A is required for polarization of cells in chondrogeniccondensations (Gao et al., 2011). To better appreciate cell shape inmetacarpal condensations we analyzed hand plate sections withdifferential interference contrast (DIC). In spdh and Wnt5a� /�

metacarpals cells in the circumference of the condensation (i.e.cells that should form the perichondrium) failed to adopt aflattened elongated shape at E14.5 (Fig. 2A), which is the typicalhallmark of perichondrium formation. In addition, we observeda clear defect in cell polarization in the cartilage growth plate.

In E14.5 wild type (wt) metacarpals, resting zone chondrocyteshave a roundish appearance while chondrocytes towards thecenter of the condensation already show clearly visible flatteningand polarization perpendicular to the longitudinal axis of thecondensation (Fig. 2B). Spdh and Wnt5a� /� metacarpals on theother hand exhibited a uniform roundish appearance of chondro-cytes throughout (Fig. 2B).

Autopod skeletal elements in spdh mice show a developmentaldelay (Albrecht et al., 2002). The perichondrium and growth platephenotypes in spdh embryos, however, were persistent in new-born animals, excluding a mere reflection of a developmentaldelay (Supplementary Fig. S1). Moreover, at E13.5 and E14.5 cellsin the center of the anlage express Ihh, a marker for prehyper-trophic chondrocytes, albeit strongly decreased and in a less-confined area than in the wt (Fig. S2). There is also abundantexpression of the IHH target Ptc1 already at E13.5 (Fig. S2). Hencechondrocytes in spdh anlagen have undergone prehypertrophicdifferentiation by E14.5, which suggests that the lack of polariza-tion is not simply a consequence of a failure to differentiatebeyond the resting zone stage.

To quantify the defect in growth plate chondrocyte polariza-tion, phalloidin staining and measurement of a quotient of celllength and height at embryonic day (E) 16.5 was used (Fig. 3A).

Fig. 2. Defective growth plate and perichondral cell morphology in spdh and Wnt5a� /� metacarpals. Differential interference contrast (DIC) images of wt, spdh andWnt5a�/� E14.5 hand plates demonstrate cell morphology; the metacarpal condensations are outlined by a dashed line. (A) Close-up inserts of indicated metacarpal regionsshow the perichondrium (pc) consisting of densely packed flattened cells in the wild type. This array of cells is not present in the spdh or Wnt5a�/� metacarpals. The borderof the cartilaginous condensation (cart) is indicated by a dashed line. (B) The metacarpal condensation of the third digit is indicated by a dashed line in each overview panel.Magnifications show the reserve zone left and the central region right. Note the failure of chondrocytes in the central anlage of spdh and Wnt5a� /� metacarpals to adopt anelongated flattened shape.

P. Kuss et al. / Developmental Biology 385 (2014) 83–9386

Wild type proliferating chondrocytes showed a marked polariza-tion perpendicular to the longitudinal axis of the cartilage elementwhereas resting chondrocytes did not show signs of polarization,as did chondrocytes in the carpal elements. In contrast, chondro-cytes in growth plates of spdh or Wnt5a� /� mice exhibited thesame quotient as wt resting chondrocytes or metacarpal chon-drocytes. Positioning of the Golgi apparatus was used as alter-native marker for cell polarization (Fig. 3B). In wt proliferatingchondrocytes the Golgi apparatus is located predominantly in anangle of 701–901 to the longitudinal axis of the metacarpal, whilein resting chondrocytes orientation is random. Consistent withphalloidin labeling, Golgi localization demonstrated perturbedpolarization in spdh and Wnt5a�/� mutants.

Thus, spdh and Wnt5a mutants display an apparent cellpolarity defect in both growth plate and perichondrium. Thissuggests that the function of Hox genes as well as Wnt5a functionis required in metacarpal condensations and adjacent mesench-yme for the establishment of cell polarity as a prerequisite ofgrowth plate and perichondrium architecture.

Partial coexpression of Hoxd13 and Wnt5a

Given the overlapping phenotypes between Hoxd13 and Wnt5amutant mice we tested if Hoxd13 might be involved in regulatingthe expression of Wnt5a. To this end we first compared theexpression of Hoxd13 and Wnt5a. Section in situ hybridization(ISH) at E13.5 and E14.5 demonstrated partially overlappingexpression patterns of Hoxd13 and Wnt5a in the autopodmesenchyme surrounding metacarpals and digits (Fig. 4A). Intri-guingly, both genes are hardly expressed in the carpal region(Figs. 4A and 5A). To substantiate that WNT5A protein might beproduced from HOXD13 positive cells, we performed immunola-beling for both proteins. HOXD13 showed highest expression inthe circumference of metacarpals at E13.5 (Fig. 4B) and was alsopresent in interdigital cells. WNT5A protein was detected pre-dominantly in the prehypertrophic region of the growth plate(Fig. 4B, arrows) as well as in the perichondral area. Magnificationof the perichondral region demonstrates intense WNT5A labelingsurrounding cells positive for nuclear HOXD13 (Fig. 4B) suggestingproduction of WNT5A protein by these cells. This shows thatHoxd13/HOXD13 and Wnt5a/WNT5A show partially overlappingdomains of expression, especially in the region encompassing themetacarpal condensations, i.e. the future perichondrium.

Fig. 3. Disrupted chondrocyte polarity in spdh and Wnt5a�/� metacarpal growthplate chondrocytes. (A) Phalloidin staining at E16.5. Note the appearance offlattened stack-forming cells in the region of proliferating chondrocytes (PC) incontrast to the roundish reserve chondrocytes (RC) or carpal chondrocytes in thewt. Chondrocytes in growth plates of spdh and Wnt5a�/� mutants exhibit thesame spherical morphology as wt RC or carpal chondrocytes. A polarity index wascalculated as a quotient of maximal length and height, shown as a box whisker plotbelow. (B) Golgi apparatus visualized by staining for GM130 at E16.5. In wt PC theGolgi apparatus was localized laterally, hence the Golgi–nucleus axis is orientedapproximately perpendicular to the longitudinal axis of the growth plate. In wt RCthe Golgi apparatus was distributed randomly. The same random orientation wasobserved in all metacarpal chondrocytes of spdh or Wnt5a�/� embryos. Forquantification, the angle between the Golgi–nucleus axis and the growth plate'slongitudinal axis was measured (see schematic image bottom right). Quantificationshows the number of cells for angles between 01 and 901; diagrams showcomparison between spdh, Wnt5a�/� and WT proliferating chondrocytes (left)or between spdh, Wnt5a� /� and WT reserve chondrocytes (right).

Fig. 4. Partial coexpression of Hoxd13 and Wnt5a. (A) Comparison of Hoxd13 and Wnt5a mRNA expression by in-situ hybridization on E13.5 and E14.5 hand plate sections.Note the overlapping expression patterns of Hoxd13 and Wnt5a in the perichondral area of metaracpal condensations and low expression of both genes in the carpal region(c). (B) Immunolabeling for HOXD13 andWNT5A on E13.5 hand plate sections. Overview pictures above, magnification of boxed areas below. Note predominant expression ofHOXD13 in cells surrounding the condensations (digit condensations 2–5 are indicated by a dashed line and denominated as d2–d5) but also in interdigital cells. WNT5Aprotein can be detected in the perichondrium but also in the prehypertrophic zone of metacarpal condensations (arrows). Lower panel: inserts show high-power imagesfrom the perichondral region (boxed). Cells expressing high amounts of HOXD13 (nuclear) also show abundant signal for WNT5A protein in their surrounding.

P. Kuss et al. / Developmental Biology 385 (2014) 83–93 87

Hoxd13 is upstream of Wnt5a and Wnt5b expression

Next we analyzed spdh mutants for Wnt5a expression. Spdhembryos showed a specific loss of Wnt5a expression in theproximal autopod mesenchyme encompassing the metacarpalcondensations at E13.5 and E14.5 (Fig. 5A). In the distal mesench-yme (where Hoxd13 is only weakly expressed), the expression ofWnt5a appeared normal. Downregulation of Wnt5a in spdhautopods was corroborated by quantitative real-time PCR(Fig. 5C) from whole hand plates at E14.5 and E16.5 indicating apersistent effect. The tissue we used included the distal mesench-yme where Wnt5a expression appeared normal in the ISH; none-theless the downregulation was clearly detectable.

Wnt5b is closely related to Wnt5a and is expressed in partiallyoverlapping domains in the autopod mesenchyme (Witte et al.,2009). Moreover, WNT5B can activate non-canonical pathways(Hardy et al., 2008; Lin et al., 2010; Roszko et al., 2009) and bothWNT5A and WNT5B were recently involved in cell polarization ina cartilage in vitro model (Randall et al., 2012). We therefore alsoanalyzed Wnt5b expression in spdh mice and found a comparabledownregulation of Wnt5b in metacarpal condensations and insurrounding mesenchyme (Fig. 5B), which was corroborated byquantitative real-time PCR (Fig. 5C).

We have previously shown that Hoxd13 overexpression wasable to induce perichondral markers, including Wnt5a, in chickenmicromass cultures (Villavicencio-Lorini et al., 2010). In this cellsystem, however, one cannot unambiguously distinguish betweena primary induction of Wnt5a expression and a secondary upre-gulation due to induction of perichondral cell differentiation.To test whether Hoxd13 is able to induceWnt5a orWnt5b expressionin a system that does not involve cell differentiation, chicken DF1fibroblasts were transfected with replication-competent retrovirusescarrying wild type or mutant (plus 7 alanine) Hoxd13. Wild type

HOXD13, but not HOXD13þ7Ala led to a significant upregulation ofWnt5a and Wnt5b expression (Fig. 5D).

In addition, we tested the Hoxd13spdh and Wnt5anull alleles forgenetic interaction. As readout we employed metacarpal lengthmeasurement in skeletal preparations of 1 week old mice (Fig. 6).Interestingly heterozygous spdh mice, which appear normal, havealready a small but significant phenotype in this measurement,while Wnt5aþ /� mice do not show a heterozygous phenotype.Double heterozygous mice show a small but significant additionaldecrease in metacarpal length, indicating genetic interaction.

Impaired planar cell polarity signaling in growth plate andperichondrium of Hoxd13spdh/spdh mice

WNT5A activates non-canonical Wnt pathways in mammals(Kikuchi et al., 2012), which have been involved in growth platearchitecture (Romereim and Dudley, 2011). Importantly WNT5Aappears to be a major factor influencing cartilage cell polarity (Gaoet al., 2011; Yang et al., 2003). There is compelling evidence for acrucial role of WNT/PCP signaling in chondrocyte polarization,with WNT5A being one possible ligand (Gao et al., 2011; Li andDudley, 2009; Randall et al., 2012). Since we found a similarpolarity defect in spdh and Wnt5a� /� metacarpals and a down-regulation of Wnt5a in spdh hand plates, we analyzed thedistribution of PCP core components Disheveled-2 (DVL2) andPRICKLE1 in spdh metacarpals via immunolabeling. In wt prolif-erating chondrocytes, DVL2 was localized specifically at the lateralends of columnar chondrocytes (Fig. 7A, arrows). In spdh mice, incontrast, this lateral localization of DVL2 was lost with most cellsexhibiting a more uniform distribution (Fig. 7A). PRICKLE1 equallyshowed a disrupted distribution in growth plate chondrocytes ofspdh mice. In wt metacarpals PRICKLE1 was, like DVL2, localized tolateral edges of the chondrocytes, albeit displaying a broader

Fig. 5. Misregulation of Wnt5a in spdh mutants. (A) and (B) Section in-situ hybridization for Wnt5a and Wnt5b on wt and spdh handplates. Stages as indicated. Note lowWnt5a and Wnt5b expression specifically in the wt carpal (c) area. Border between carpal area and metacarpals (m) is indicated by a dashed line; magnification of carpal andmetacarpal areas is shown right. In spdh handplates Wnt5a is downregulated not only in the carpal area but also in mesenchyme encompassing the metacarpalcondensations; a similar downregulation can be seen for Wnt5b. (C) Quantification of Wnt5a and Wnt5b expression in E14.5 and E16.5 autopodes by quantitative real-timeqPCR confirmes reduction of Wnt5a and Wnt5b mRNAs in spdh animals compared to wild type controls. The wt expression level was set as 100%. (D) Induction of Wnt5a andWnt5b expression in DF1 cells retrovirally transfected with either Hoxd13 or Hoxd13þ7Ala (7 alanine expansion). Real-time qPCR performed 3 and 7 days after transfectionshows a strong upregulation of Wnt5a and Wnt5b expression by Hoxd13, but not by Hoxd13þ7Ala.

P. Kuss et al. / Developmental Biology 385 (2014) 83–9388

signal than DVL2. This specific localization was not present in spdhmetacarpals (Fig. 7B). Interestingly, the lateral edges of chondro-cytes in spdh metacarpals were devoid of either DVL2 or PRICKLE1staining. Thus, the localization seen in mutants might be char-acterized as a reversal of the wild type distribution. This couldindicate a reversal of growth plate orientation; however nomorphological indication for this was seen (compare Figs. 1 and2, S1).

Importantly, in the wt perichondrium PRICKLE1 was distribu-ted to the apical and basal sides of the cells (Fig. 7B). PRICKLE1localization in spdh cells adjacent to cartilage (i.e. the positionwhere a perichondrium should form) showed randomizedPRICKLE1 distribution with an apparent reduction in PRICKLE1protein (Fig. 7B), indicating that PCP signaling could be involved inthe establishment or maintenance of polarity in perichondral cells.

Non-canonical WNT pathways are involved in the establish-ment of the primary cilium (Wallingford and Mitchell, 2011) and adefect in primary cilium function leads to disrupted cell shape andcolumnar orientation in growth plates (Haycraft et al., 2007;Koyama et al., 2007; Song et al., 2007). We therefore comparedprimary cilia in Wnt5a�/� and spdh metacarpals by immunohis-tochemistry against acetylated tubulin. Wt resting chondrocytesshow an average cilium length of 2.5 μm confirming previousreports (de Andrea et al., 2010). We found a reduction of primarycilium length of approx. 40% in spdh metacarpals. A similarreduction was observed in Wnt5a� /� mice (Fig. S3). Althoughthe functional consequences of this perturbation are unclear, itprovides a further phenotypic parallel between spdh and Wnt5amutants.

Misregulation of WNT/β-CATENIN signaling in Hoxd13spdh/spdh mice

WNT5A is known to negatively regulate the canonical WNTpathway in vitro and in vivo via diverse mechanisms (Kikuchiet al., 2012; Mikels and Nusse, 2006; Topol et al., 2003). Usingthe Axin2::LacZ reporter (Lustig et al., 2002) we observed strongβ-CATENIN pathway activity in wt joint surface and in osteoblasts

adjacent to the bone collar, while the perichondrium was negative(Fig. 8A). In contrast, spdh mice showed intense staining sur-rounding the complete cartilage anlagen (Fig. 8A) similar to carpalcondensations and joint surfaces in wt mice. This result wascorroborated by immunolabeling for β-CATENIN, which showedincreased abundance in the perichondral region of spdh metacar-pals (Fig. 8B). Altogether this is in line with the metacarpal-to-carpal transformation and the acquirement of a joint-like identityby spdh perichondral cells (Villavicencio-Lorini et al., 2010). Theupregulation of β-CATENIN signaling in the perichondrium of spdhmice also is in agreement with the observed downregulation ofWnt5a.

Partial rescue of perichondrium in Hoxd13spdh/spdh mice by exogenousHOXD13 or WNT5A

Our results indicate that HOXD13 expression in mesenchymesurrounding the metacarpal condensation and its perichondriumis necessary for perichondral cell specification concomitant withtheir polarization. This presumably can be achieved by cell-autonomous effects by HOXD13 in perichondral precursors or viathe regulation of secreted factors such as WNT5A, which in turncan act in a paracrine and/or autocrine fashion. To substantiate thehypothesis that HOXD13 can regulate the polarity of adjacent cellsin a non-cell autonomous manner we performed a rescue experi-ment using spdh limb explant cultures. For this purpose chickenDF1 cells that had been transfected with RCAS(A) retrovirusesexpressing HOXD13 were implanted into the interdigital mesench-yme next to metacarpals at E13.5 or E14.5. In a parallel experimentcells expressing WNT5A were implanted. Implantation of RCAS(A)–GFP transfected cells was used as a control. For directvisualization of the implants after incubation (i.e. to select forsuccessful implantation), DF1 cells were co-transfected with RCAS(B) expressing GFP (examples shown in Fig. S4). The limb explantswere cultured for 3 or 4 days and analyzed. The explants weresectioned and the localization of the cell implant was visualized byantibody staining for a RCAS-specific surface antigen. Cell polarity

Fig. 6. Genetic interaction of the Hoxd13spdh and Wnt5anull alleles. Length of metacarpals (m2, m3, m4) at postnatal day 7 (p7) were used as readout; in each case the wildtype (WT) metacarpal length was set as 1. Wnt5aþ /� metacarpals have normal or slightly increased length, Hoxd13þ /spdh metacarpals show a significant shortening. Double-heterozygotes (Wnt5aþ /�; Hoxd13þ /spdh) show a small but significant increase in shortening indicating genetic interaction. For each genotype, five animals were analyzed.Error bars represent SEM. nn: po0.005; nnn: po0.001. Representative images of skeletal preparations are shown below.

P. Kuss et al. / Developmental Biology 385 (2014) 83–93 89

in the vicinity of the implant was analyzed on the same sections;in-situ hybridization was performed on adjacent sections. We firstanalyzed the effect of cell implants on putative perichondral cellshape in spdh mutants. Implanting GFP-expressing control cellsdid not change the morphology of wt presumptive perichondralcells nor did it rescue the appearance of the unpolarized cell layerssurrounding the cartilaginous condensations in spdh mutants. Incontrast, implantation of HOXD13 or WNT5A expressing cellsinduced the formation of flattened cell morphology in the peri-chondral region of spdh metacarpals (Fig. 9A and B). Next weanalyzed whether HOXD13 might be able to non-cell autono-mously induce perichondrium-specific gene expression in adja-cent cells. Wild type perichondral cells express several marker

genes, amongst them Crabp1 and Tsp (Fig. 9C, upper panel). Asshown previously, in spdh mice cells adjacent to metacarpalcondensations do not express these markers (Fig. 9C middlepanel). Implantation of HOXD13-expressing DF1 cells betweenspdh metacarpal condensations induced the expression of peri-chondral markers Crabp1 and Tsp in their vicinity (Fig. 9C).

Discussion

The skeletal elements of the limb are laid down in diverseshapes; however most of them are longitudinal while only thecarpal bones are cuboidal. Carpal bones do not form a growthplate, undergo secondary ossification and do not have corticalbone. Instead, they are surrounded by a joint surface. Our previousstudies have indicated a phenotypic switch of metacarpal long-itudinal bones to carpal like elements in Hoxd13-mutant spdhanimals and SPD patients (Villavicencio-Lorini et al., 2010). Impor-tantly, this phenotype does not occur in Hoxd13 null mice (Dolleet al., 1993), which is explained by a dominant-negative effect ofthe SPD/spdh HOXD13 protein leading to a combined inactivationof autopod-expressed poly-alanine containing HOX proteins(Bruneau et al., 2001; Villavicencio-Lorini et al., 2010). Hoxd11-13and Hoxa13 have well-established roles in autopod patterning (seee.g. Delpretti et al., 2012; Sheth et al., 2012; Spitz, 2010; Tschoppand Duboule, 2011; Zakany and Duboule, 2007). However, ourknowledge about the actual translation of patterning informationinto locally controlled differentiation and morphogenesis is lim-ited. Recently, a key role in this process has been attributed to cellpolarity regulated by non-canonical WNT pathways (Gao et al.,2011). We here propose that Hox genes are involved in theregulation of cell polarity during the formation of cartilageelements, thereby influencing the condensation's shape.

Hox genes apparently have several roles in this process. First,the concerted action of the autopod-expressed Hox genes isrequired for the formation of a bona-fide perichondrium. Second,these Hox genes are involved in the regulation of polarity in thecartilaginous growth plate. This process is a key event in chon-drogenesis, as it is a prerequisite for the formation of columnarchondrocytes, and thus shapes the skeletal element giving it itslongitudinal appearance. Our results indicate a model where thedistinction between a cuboidal and longitudinal skeletal elementis determined by at least 2 factors: the formation of a perichon-drium that supports and initiates the formation of cortical bonebut also restricts lateral growth, and the implementation ofcellular polarization in the growth plate along the longitudinalaxis of the skeletal element. If no functional perichondrium ispresent, as it is the case in carpal bones and spdh metacarpals, ajoint surface is formed and the characteristic polarization ofchondrocytes is absent. Polarization results in directed growthalong the longitudinal axis, whereas unpolarized cells grow asspheres. Thus, the perichondrium and its signaling capacity maybe a main determinant conferring the ability of a growth plate toperform growth in the longitudinal direction as opposed tospherical growth. A role for Hoxd13 and Wnt5a in this process issupported by the very low Hoxd13 or Wnt5a expression in thecarpal region (see Figs. 4A and 5A). It was proposed before that thedeveloping wrist region (the mesopodium) develops from a regionexcluded from Hoxd gene expression and therefore presumablyexhibits an overall lower HOX protein content (Woltering andDuboule, 2010 and references therein). Recently Andrey et al.(2013) have shown that this is due to a regulatory switch betweendifferent domains of the Hoxd cluster enabling this separation, andthus allowing for the emergence of an articulation betweenzeugopod and autopod. Thus we propose that in spdh metacarpals,where the combined function of Hox genes/proteins is impaired

Fig. 7. Defective PCP signaling in spdh metacarpal chondrocytes. Localization of thecore PCP components Disheveled-2 (DVL2) (A) and PRICKLE1 (B) in metacarpals ofE14.5 wt and spdh mice. Note lateral distribution of DVL2 and PRICKLE1 proteins(arrows) in wt proliferative chondrocytes in contrast to a more randomizeddistribution in spdh chondrocytes. The longitudinal axis of the growth plate isindicated by a double-headed arrow in magnifications. In the wt perichondrium(pc) PRICKLE1 is distributed to apical and basal parts of perichondral cells while inspdh mice PRICKLE1 shows random distribution. Note lower PRICKLE1 stainingintensity in the spdh perichondrium.

P. Kuss et al. / Developmental Biology 385 (2014) 83–9390

(Bruneau et al., 2001; Villavicencio-Lorini et al., 2010), the meta-carpal region is shifted from its original autopod domain with theresult that metacarpals develop as wrist-like elements.

Metacarpal condensations of Wnt5a�/� mice exhibited aphenotype that in part overlapped with features seen in spdhmice. Several lines of evidence indicate involvement of autopodHOX proteins in Wnt5a regulation. First, Wnt5a is downregulatedin spdh mutants (Fig. 5A, C), second, HOXD13 induces Wnt5aexpression in vitro (Fig. 5D). Third, Hoxd13 and Wnt5a expressiondomains partially overlap, specifically in the perichondral regionof metacarpals, where extracellular WNT5A protein was detectedsurrounding HOXD13 positive cells (Fig. 4A and B). In line with thisit has previously been shown that HOXA13, the function of whichis also affected in spdh animals, controls Wnt5a expression in themouse (Rinn et al., 2008) providing further indication for apossible regulation of Wnt5a expression by autopod-expressedHOX transcription factors. However at this point we cannotconclude whether this regulation is direct or indirect.

Concomitant with the downregulation of Wnt5a we find adefect in the expression level (in the perichondrium) and theintracellular distribution (perichondrium and growth plate) of keyPCP components. This further strengthens the hypothesis that aPCP-driven convergent extension like mechanism is involved inpolarization and arrangement of chondrocytes, as other studieshave suggested (Ahrens et al., 2009, 2012; Gao et al., 2011; Li andDudley, 2009). Moreover this pathway appears to be active inperichondral cells as well.

The defect in Wnt5a expression observed in the spdh mice wasspecific for the developing metacarpals, however Wnt5a appearsto have similar functions in other parts of the skeleton sinceWnt5a� /� animals, contrary to spdh, exhibit a polarity defect also

in zeugopod and stylopod elements (Yang et al., 2003 and data notshown).

The activity of HOXD13 in this respect appears to applyspecifically to the forming metacarpal elements of the autopod.Why is the loss of Wnt5a expression in spdh mutants restricted tothe metacarpal region? In chicken micromass cultures and limbbuds, recombinant FGF enhanced Wnt5a expression (Chimal-Monroy et al., 2002), and a recent report indicated that MEK1/2(which are downstream components of the FGF pathway) posi-tively regulate the Wnt5a promoter (Katula et al., 2012). Thereforeit is conceivable that distal expression of Wnt5a is under control ofe.g. FGF signals from the apical ectodermal ridge (AER). Conver-sely, in more proximal regions (i.e. where cells have left theinfluence sphere of the AER) mesenchymal HOX proteins maytake over the regulation. This might also explain why the phalan-geal condensations, albeit clearly malformed, do not grow asspherical elements in spdh mice, since those condensations mightexperience an appropriate dosage of WNT5A expression duringtheir condensation phase. This might also be an explanation whyspdh limb buds are not expanded along the dorso–ventral axis, asare Wnt5a� /� limb buds. It has been shown that morphogenesisof the early limb bud is controlled via cell rearrangements by theinteraction of AER/FGF and WNT5A signaling (Gros et al., 2010;Wyngaarden et al., 2010), a process that is likely not to be affectedby the localized expression defect observed here.

We found that besides Wnt5a also expression of the closelyrelated Wnt5b gene was affected. WNT5A and WNT5B both belongto the group of “non-canonical” Wnt ligands (Kikuchi et al., 2012),which means that they predominantly do not signal via theβ-CATENIN pathway, but via alternative pathways. In the courseof growth plate chondrocyte differentiation, WNT5A and WNT5B

Fig. 8. Misregulation of WNT/β-CATENIN signaling in spdh metacarpals. (A) LacZ staining on cryosections of wt or spdh E14.5 embryos heterozygous for the Axin2::LacZreporter; magnifications of boxed areas below. LacZ staining shows a strong signal in lateral hypertrophic regions and cortical osteoblasts in the wt. In spdh mice this stainingis absent in line with the failure of hypertrophic differentiation and osteoblast formation. Perichondral cells are devoid of LacZ staining in the wt, which is only visible in jointareas. In spdh mice intensified LacZ staining in cells adjacent to the cartilage surrounding the complete anlage can be seen in a fashion resembling wt carpals (see insertright). (B) Immunolabeling for β-CATENIN on paraffin sections of E14.5 metacarpals, counterstaining with DAPI. In spdh metacarpals a stronger β-CATENIN signal is seen inthe interdigital mesenchyme in line with Axin2::LacZ staining.

P. Kuss et al. / Developmental Biology 385 (2014) 83–93 91

were shown to have opposing functions (Yang et al., 2003),however in vitro experiments suggest that both have overlappingfunctions with respect to cartilage cell polarity (Bradley and Drissi,2011; Randall et al., 2012). It is therefore possible that WNT5A andWNT5B cooperate in induction of cartilage cell polarity.

The WNT/β-CATENIN pathway is a central regulator of jointformation and pathway activity is high in joint cells (Guo et al.,2004; Hill et al., 2005). We observed such a pattern of β-CATENINpathway activity in cells surrounding carpal bones, as well as cellssurrounding the spdh but not the wt metacarpal condensationssubstantiating the metacarpal-to-carpal conversion proposedbefore (Villavicencio-Lorini et al., 2010).

Conclusion

A role for Hox genes in local morphogenesis

Hox genes are continuously expressed in autopod mesenchymeafter the digit patterning phase and we previously showed thatHOXD13 is required here for the repression of interdigital chon-drogenesis (Kuss et al., 2009). This study puts forward the ideathat 5′ Hox genes are also involved in local morphogenesis ofskeletal elements by the establishment of the perichondrium,which in concert with the arrangement of columnar chondrocytesin the growth plate shapes longitudinal skeletal elements anddiscerns them from cuboidal/spherical bones. A failure of thiscoordinative effort in spdh mice leads to (i) a defective perichon-drium with finally a joint-like appearance (including upregulated

β-CATENIN signaling) and (ii) a lack of the polarized growth platearchitecture resulting in spherical growth.

Acknowledgments

This work was funded by a Grant from the German ResearchFoundation (DFG) to S. M. J. S., P. V-G. and S. S were funded by theGerman Research Foundation (DFG; Grant GK1631), French-German University (UFA-DFH; Grant CDFA-06–11) and the Asso-ciation Française contre les Myopathies (AFM) as part of theMyoGrad International Research Training Group for Myology. Wethank Nicole Rösener and Asita Stiege for excellent technicalassistance. We thank P. Francis-West (London, UK) for the RCAS-Wnt5a and Wnt5b constructs.

Appendix A. Supplementary material

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.ydbio.2013.10.013.

References

Ahrens, M.J., Li, Y., Jiang, H., Dudley, A.T., 2009. Convergent extension movements ingrowth plate chondrocytes require gpi-anchored cell surface proteins. Devel-opment 136, 3463–3474.

Albrecht, A.N., Schwabe, G.C., Stricker, S., Boddrich, A., Wanker, E.E., Mundlos, S.,2002. The synpolydactyly homolog (spdh) mutation in the mouse—a defect inpatterning and growth of limb cartilage elements. Mech. Dev. 112, 53–67.

Fig. 9. Implantation of Hoxd13- or Wnt5a-expressing cells rescues the perichondrium in spdh limb explants. (A) Sections of limb explants (metacarpal region shown);brown staining demarcates implanted DF1 cells expressing GFP (control), Hoxd13 or Wnt5a. Implantation of cells expressing Hoxd13 or Wnt5a led to a partial rescue offlattened perichondral cell morphology in the vicinity of the implant in spdh limbs (see inserts). Implantation of Hoxd13 or Wnt5a-expressing cells did not alter cellmorphology in wt explants. (B) Quantification of perichondral cell morphology in the vicinity of implanted cells shows increased polarity index (quotient of maximal length /height of cells) in spdh specimens implanted with Hoxd13- or Wnt5a-expressing cells. Error bars represent SEM. At least three independent samples were analyzed in eachcase. T-test: nn indicate po0,005. (C) In situ hybridization with specific probes against perichondral markers Crabp1 and Tsp on E14.5 wt and spdh metacarpal sections showsrescue of perichondral marker expression in spdh limbs implanted with Hoxd13-expressing cells. Implants were localized on adjacent sections by immunolabeling (seeinsert, arrows); the corresponding regions on the ISH sections are indicated by asterisks.

P. Kuss et al. / Developmental Biology 385 (2014) 83–9392

Andre, P., Wang, Q., Wang, N., Gao, B., Schilit, A., Halford, M.M., Stacker, S.A., Zhang, X.,Yang, Y., 2012. The wnt coreceptor ryk regulates wnt/planar cell polarity bymodulating the degradation of the core planar cell polarity component vangl2. J.Biol. Chem. 287, 44518–44525.

Andrey, G., Montavon, T., Mascrez, B., Gonzalez, F., Noordermeer, D., Leleu, M.,Trono, D., Spitz, F., Duboule, D., 2013. A switch between topological domainsunderlies HoxD genes collinearity in mouse limbs. Science 340, 1234167.

Angers, S., Moon, R.T., 2009. Proximal events in Wnt signal transduction. Nat. Rev.Mol. Cell Biol. 10, 468–477.

Bradley, E.W., Drissi, M.H., 2011. Wnt5b regulates mesenchymal cell aggregationand chondrocyte differentiation through the planar cell polarity pathway. J. CellPhysiol. 226, 1683–1693.

Bruneau, S., Johnson, K.R., Yamamoto, M., Kuroiwa, A., Duboule, D., 2001. Themouse Hoxd13(spdh) mutation, a polyalanine expansion similar to human typeII synpolydactyly (SPD), disrupts the function but not the expression of otherHoxd genes. Dev. Biol. 237, 345–353.

Chimal-Monroy, J., Montero, J.A., Ganan, Y., Macias, D., Garcia-Porrero, J.A., Hurle, J.M.,2002. Comparative analysis of the expression and regulation of Wnt5a, Fz4,and Frzb1 during digit formation and in micromass cultures. Dev. Dyn. 224,314–320.

de Andrea, C.E., Wiweger, M., Prins, F., Bovee, J.V., Romeo, S., Hogendoorn, P.C.,2010. Primary cilia organization reflects polarity in the growth plateand implies loss of polarity and mosaicism in osteochondroma. Lab. Investig.J. Tech. Methods Pathol. 90, 1091–1101.

Delpretti, S., Zakany, J., Duboule, D., 2012. A function for all posterior Hoxd genesduring digit development? Dev. Dyn. 241, 792–802.

Dolle, P., Dierich, A., LeMeur, M., Schimmang, T., Schuhbaur, B., Chambon, P.,Duboule, D., 1993. Disruption of the Hoxd-13 gene induces localized hetero-chrony leading to mice with neotenic limbs. Cell 75, 431–441.

Gao, B., Song, H., Bishop, K., Elliot, G., Garrett, L., English, M.A., Andre, P., Robinson, J.,Sood, R., Minami, Y., Economides, A.N., Yang, Y., 2011. Wnt signaling gradientsestablish planar cell polarity by inducing Vangl2 phosphorylation through Ror2.Dev. Cell 20, 163–176.

Gros, J., Hu, J.K., Vinegoni, C., Feruglio, P.F., Weissleder, R., Tabin, C.J., 2010. WNT5A/JNK and FGF/MAPK pathways regulate the cellular events shaping the verte-brate limb bud. Curr. Biol. 20, 1993–2002.

Guo, X., Day, T.F., Jiang, X., Garrett-Beal, L., Topol, L., Yang, Y., 2004. Wnt/beta-catenin signaling is sufficient and necessary for synovial joint formation. GenesDev. 18, 2404–2417.

Hardy, K.M., Garriock, R.J., Yatskievych, T.A., D'Agostino, S.L., Antin, P.B., Krieg, P.A.,2008. Non-canonical Wnt signaling through Wnt5a/b and a novel Wnt11 gene,Wnt11b, regulates cell migration during avian gastrulation. Dev. Biol. 320,391–401.

Haycraft, C.J., Zhang, Q., Song, B., Jackson, W.S., Detloff, P.J., Serra, R., Yoder, B.K.,2007. Intraflagellar transport is essential for endochondral bone formation.Development 134, 307–316.

Hecht, J., Seitz, V., Urban, M., Wagner, F., Robinson, P.N., Stiege, A., Dieterich, C.,Kornak, U., Wilkening, U., Brieske, N., Zwingman, C., Kidess, A., Stricker, S.,Mundlos, S., 2007. Detection of novel skeletogenesis target genes by compre-hensive analysis of a Runx2(-/-) mouse model. Gene Expr. Patterns 7, 102–112.

Hill, T.P., Spater, D., Taketo, M.M., Birchmeier, W., Hartmann, C., 2005. CanonicalWnt/beta-catenin signaling prevents osteoblasts from differentiating intochondrocytes. Dev. Cell 8, 727–738.

Johnson, K.R., Sweet, H.O., Donahue, L.R., Ward-Bailey, P., Bronson, R.T., Davisson, M.T.,1998. A new spontaneous mouse mutation of Hoxd13 with a polyalanineexpansion and phenotype similar to human synpolydactyly. Hum. Mol. Genet.7, 1033–1038.

Katula, K.S., Joyner-Powell, N.B., Hsu, C.C., Kuk, A., 2012. Differential regulation ofthe mouse and human Wnt5a alternative promoters A and B. DNA Cell Biol. 31,1585–1597.

Kikuchi, A., Yamamoto, H., Sato, A., Matsumoto, S., 2012. Wnt5a: its signalling,functions and implication in diseases. Acta Physiol. (Oxf) 204, 17–33.

Komiya, Y., Habas, R., 2008. Wnt signal transduction pathways. Organogenesis 4,68–75.

Koyama, E., Young, B., Nagayama, M., Shibukawa, Y., Enomoto-Iwamoto, M.,Iwamoto, M., Maeda, Y., Lanske, B., Song, B., Serra, R., Pacifici, M., 2007.Conditional Kif3a ablation causes abnormal hedgehog signaling topography,growth plate dysfunction, and excessive bone and cartilage formation duringmouse skeletogenesis. Development 134, 2159–2169.

Kuss, P., Villavicencio-Lorini, P., Witte, F., Klose, J., Albrecht, A.N., Seemann, P., Hecht,J., Mundlos, S., 2009. Mutant Hoxd13 induces extra digits in a mouse model of

synpolydactyly directly and by decreasing retinoic acid synthesis. J. Clin. Invest.119, 146–156.

Li, Y., Dudley, A.T., 2009. Noncanonical frizzled signaling regulates cell polarity ofgrowth plate chondrocytes. Development 136, 1083–1092.

Lin, S., Baye, L.M., Westfall, T.A., Slusarski, D.C., 2010. Wnt5b-Ryk pathway providesdirectional signals to regulate gastrulation movement. J. Cell Biol. 190, 263–278.

Lustig, B., Jerchow, B., Sachs, M., Weiler, S., Pietsch, T., Karsten, U., van de Wetering,M., Clevers, H., Schlag, P.M., Birchmeier, W., Behrens, J., 2002. Negative feedbackloop of Wnt signaling through upregulation of conductin/axin2 in colorectaland liver tumors. Mol. Cell Biol. 22, 1184–1193.

Mikels, A.J., Nusse, R., 2006. Purified Wnt5a protein activates or inhibits beta-catenin-TCF signaling depending on receptor context. PLoS Biol. 4, e115.

Minina, E., Wenzel, H.M., Kreschel, C., Karp, S., Gaffield, W., McMahon, A.P.,Vortkamp, A., 2001. BMP and Ihh/PTHrP signaling interact to coordinatechondrocyte proliferation and differentiation. Development 128, 4523–4534.

Muragaki, Y., Mundlos, S., Upton, J., Olsen, B.R., 1996. Altered growth and branchingpatterns in synpolydactyly caused by mutations in HOXD13. Science 272,548–551.

Randall, R.M., Shao, Y.Y., Wang, L., Ballock, R.T., 2012. Activation of Wnt PlanarCell Polarity (PCP) signaling promotes growth plate column formation in vitro.J. Orthop. Res. 30, 1906–1914.

Rinn, J.L., Wang, J.K., Allen, N., Brugmann, S.A., Mikels, A.J., Liu, H., Ridky, T.W.,Stadler, H.S., Nusse, R., Helms, J.A., Chang, H.Y., 2008. A dermal HOX transcrip-tional program regulates site-specific epidermal fate. Genes Dev. 22, 303–307.

Romereim, S.M., Dudley, A.T., 2011. Cell polarity: the missing link in skeletalmorphogenesis? Organogenesis 7, 217–228.

Roszko, I., Sawada, A., Solnica-Krezel, L., 2009. Regulation of convergence andextension movements during vertebrate gastrulation by the Wnt/PCP pathway.Semin. Cell Dev. Biol. 20, 986–997.

Sheth, R., Marcon, L., Bastida, M.F., Junco, M., Quintana, L., Dahn, R., Kmita, M.,Sharpe, J., Ros, M.A., 2012. Hox genes regulate digit patterning by controllingthe wavelength of a Turing-type mechanism. Science 338, 1476–1480.

Song, B., Haycraft, C.J., Seo, H.S., Yoder, B.K., Serra, R., 2007. Development of thepost-natal growth plate requires intraflagellar transport proteins. Dev. Biol. 305,202–216.

Spitz, F., 2010. Control of vertebrate Hox clusters by remote and global cis-actingregulatory sequences. Adv. Exp. Med. Biol. 689, 63–78.

Topol, L., Jiang, X., Choi, H., Garrett-Beal, L., Carolan, P.J., Yang, Y., 2003. Wnt-5ainhibits the canonical Wnt pathway by promoting GSK-3-independent beta-catenin degradation. J. Cell Biol. 162, 899–908.

Tschopp, P., Duboule, D., 2011. A genetic approach to the transcriptional regulationof Hox gene clusters. Annu. Rev. Genet. 45, 145–166.

Villavicencio-Lorini, P., Kuss, P., Friedrich, J., Haupt, J., Farooq, M., Turkmen, S.,Duboule, D., Hecht, J., Mundlos, S., 2010. Homeobox genes d11-d13 and a13control mouse autopod cortical bone and joint formation. J. Clin. Invest. 120,1994–2004.

Wallingford, J.B., Mitchell, B., 2011. Strange as it may seem: the many links betweenWnt signaling, planar cell polarity, and cilia. Genes Dev. 25, 201–213.

Witte, F., Chan, D., Economides, A.N., Mundlos, S., Stricker, S., 2010. Receptortyrosine kinase-like orphan receptor 2 (ROR2) and Indian hedgehog regulatedigit outgrowth mediated by the phalanx-forming region. Proc. Natl. Acad. Sci.USA 107, 14211–14216.

Witte, F., Dokas, J., Neuendorf, F., Mundlos, S., Stricker, S., 2009. Comprehensiveexpression analysis of all Wnt genes and their major secreted antagonistsduring mouse limb development and cartilage differentiation. Gene Expr.Patterns 9, 215–223.

Woltering, J.M., Duboule, D., 2010. The origin of digits: expression patterns versusregulatory mechanisms. Dev. Cell 18, 526–532.

Wyngaarden, L.A., Vogeli, K.M., Ciruna, B.G., Wells, M., Hadjantonakis, A.K., Hopyan,S., 2010. Oriented cell motility and division underlie early limb bud morpho-genesis. Development 137, 2551–2558.

Yamaguchi, T.P., Bradley, A., McMahon, A.P., Jones, S., 1999. A Wnt5a pathwayunderlies outgrowth of multiple structures in the vertebrate embryo. Devel-opment 126, 1211–1223.

Yang, Y., Topol, L., Lee, H., Wu, J., 2003. Wnt5a and Wnt5b exhibit distinct activitiesin coordinating chondrocyte proliferation and differentiation. Development130, 1003–1015.

Zakany, J., Duboule, D., 2007. The role of Hox genes during vertebrate limbdevelopment. Curr. Opin. Genet. Dev. 17, 359–366.

P. Kuss et al. / Developmental Biology 385 (2014) 83–93 93