006) 389–397www.elsevier.com/locate/matbio

Matrix Biology 25 (2

Forced chondrocyte expression of sonic hedgehog impairs joint formationaffecting proliferation and apoptosis

S. Tavella b,⁎, R. Biticchi a, R. Morello c, P. Castagnola a, V. Musante b, D. Costa b,R. Cancedda a,b, S. Garofalo a,b

a Istituto Nazionale per la Ricerca sul Cancro, Genova, Italyb Dipartimento di Oncologia, Biologia e Genetica, Universita' di Genova, Genova, Italy

c Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, TX, USA

Received 28 February 2006; received in revised form 3 July 2006; accepted 18 July 2006

Abstract

Proliferation and apoptosis are two fundamental processes that occur during limb development, and in particular in joint formation. To studythe role of hedgehog proteins in limbs, we have misexpressed Sonic Hedgehog specifically in chondrocytes.

We found that the appendicular skeleton was severely misshapen while pelvic and shoulder girdles developed normally. In particular, wedetected fusion of the elbow/knee joint, no definite carpal/tarsal, metacarpal/metatarsal bones and absence of distinct phalanges, fused in acontinuous cartilaginous rod. Molecular markers of joints, such as Gdf5 and sFrp2 were absent at presumptive joint sites and Tenascin C, amolecule associated with joint formation and expressed in permanent cartilage, was expressed in a wider region in transgenic animals as comparedto the wild type. The ratio of proliferating to non-proliferating chondrocytes was about two times higher in transgenic developing cartilage ascompared to the wild type. Accordingly, the proapoptotic gene Bax was barely detectable in the growth plate of transgenic mice and Tunel assayshowed the absence of apoptosis in presumptive joints at E15.5.

Taken together, these results suggest that misexpression of Sonic Hedgehog causes apoptosis and proliferation defects leading to the lack ofjoint cavity and fusion of selected limb skeletal elements.© 2006 Elsevier B.V./International Society of Matrix Biology. All rights reserved.

Keywords: Sonic hedgehog; Chondrocytes; Transgenic mice; Joint formation; Proliferation; Apoptosis

1. Introduction

The limb starts to develop as the mesenchyme condenses inthe limb bud. Mesenchymal cells proliferate and begin todifferentiate into chondrocytes, producing an abundant extracel-lular matrix, that will eventually surround them and form thecartilaginous template rich in collagen type II and proteoglycans.As development proceeds each bone element acquires its ownidentity, with specific boundaries, after a series of events thatculminate in joint formation. The separation of single skeletalelements begins with the branching of the mesenchyme of thefuture femur/humerus to give rise to the radius-ulna/tibia-fibula.

⁎ Corresponding author. Lab. Medicina Rigenerativa, DOBIG, Largo RosannaBenzi 10, 16132 Genoa, Italy. Tel.: +39 0105737241; fax: +39 0105737257.

E-mail address: sara.tavella@unige.it (S. Tavella).

0945-053X/$ - see front matter © 2006 Elsevier B.V./International Society of Matrdoi:10.1016/j.matbio.2006.07.005

Subsequently, the distal region of these cartilage templatesbranches further to form carpals/tarsals, metacarpals/metatarsalsand then the phalanges of fingers/toes (Capdevila and IzpisuaBelmonte, 2001). Within the splitting mesenchyme, the regionsof the future joints are marked by the so-called “interzone”. Thisregion is a specialized area of higher cell density that subse-quently evolves into a three-layered area consisting of two outerregions of higher cell density and a central region of lower celldensity (Archer et al., 2003). Cells, in this central region, firstlose properties of chondrogenic cells, then become flattened, stopto produce cartilage extracellular matrix components and even-tually die according to some authors (Niedermaier et al., 2005;Spitz and Duboule, 2001) or become incorporated in the outerlayers according to others (Wu et al., 2002). These processes leadto the joint cavity formation. At a molecular level the “interzone”is defined by the expression of Gdf5, a member of the BMP

ix Biology. All rights reserved.

390 S. Tavella et al. / Matrix Biology 25 (2006) 389–397

family. This molecule plays a positive role during chondrocytematuration (Coleman and Tuan, 2003; Storm and Kingsley,1999) and it is one of the most frequently used molecular markersof developing joints (Merino et al., 1999; Francis-West et al.,1996; Brunet et al., 1998; Hartmann and Tabin, 2001). Loss offunction mutations in this gene result in aberrant formation orabsence of joints both in mice (Storm et al., 1994) and in humans(Polinkovsky et al., 1997; Settle et al., 2003).

Several molecules are involved in controlling endochondralossification and joint formation. Among them Hedgehog pro-teins play an important role in limb development. The vertebratehedgehog gene family consists of three members: Sonic (Shh),Indian (Ihh) and Desert (Dhh) hedgehog.

Shh is involved in embryo patterning. In the limb bud, it isexpressed in the polarizing zone where it defines the antero-posterior axis along which the limb will develop and later inpostnatal growth plate chondrocytes (Wu et al., 2002).

The other member, Ihh is expressed in a specific region of thegrowth plate of long bones, the prehypertrophic region, where,together with PTHrP, generates a regulatory loop controlling therate of chondrocyte proliferation and differentiation (Lanskeet al., 1996; Vortkamp et al., 1996; Long et al., 2001).

Several evidences from loss- and gain-of-function mice sug-gest an important role of Hhs molecules in skeletal developmentand joint formation. In particular Ihh is critical for chondrocyteproliferation and differentiation since homozygous gene inac-tivation in mice leads to a lethal form of short-limbed dwarfism(St-Jacques et al., 1999). In transgenic mice overexpressing Ihhspecifically in chondrocyte proliferation and differentiation ofthese cells are altered and joints are missing or severely affected(Long et al., 2001; Minina et al., 2001). Similarly , digital raysof the chick autopod infected with Shh or Ihh expressing ret-roviruses, lack interphalangeal joints and Gdf5 expression (Me-rino et al., 1999).

Our previous report showed that SHH misexpression inchondrocytes causes a lethal phenotype at birth with a severecranio-rachischysis involving the entire axis. In these mice,SHH maintains chondrocytes in an immature stage of develop-ment, upregulating Sox9 levels and enhancing type II collagenexpression. Transgenic cartilage lacks of hypertrophic chon-drocytes and type X collagen is barely detectable (Tavella et al.,2004). Here we investigate the effect of the chondrocyte SHHmisexpression in joint formation showing that excessive SHHsignaling in cartilage leads to a higher proliferative rate anddisruption of apoptotic signals in transgenic chondrocytes im-pairing joint cavitation.

Fig. 1. Transgene construct (A): splicing acceptor, SA; Bovine growth hormoneexpression in the diaphysis of an E15.5 transgenic embryo (B).

2. Results

2.1. Chondrocyte overexpression of SHH affects joint devel-opment and causes joint fusion

To obtain transgenic mice expressing SHH in chondrocytes,the full-length cDNA of human SHH was placed under thetranscriptional control of the tissue- and cell lineage-specificregulatory region of the murine α1 (II) procollagen geneCol2a1 (Fig. 1A) (Tavella et al., 2004). This region has beensuccessfully used to target the expression of several cDNAs inchondrocytes (Garofalo et al., 1999; Zhou et al., 1995). Trans-genic mice were identified by PCR and Southern blot (notshown). Transgene expression was evaluated by both RT-PCR(Tavella et al., 2004) and immunohistochemical analysis. Usingan antibody specific for the SHH amino terminal domain, it wasdetected, in transgenic mice, mainly associated with chondro-cytes rather than with the extracellular matrix (Fig. 1B). Thisdatum is consistent with a previous report where the amino-terminal fragment of Shh was described as predominantly at-tached to cell membranes (Peters et al., 2004).

We focused our attention on limb development and in par-ticular on joint formation in these transgenics.

Alizarin Alcian Blue staining of E18.5 embryos revealedprofound changes in skeletal elements. In addition to the alreadydescribed persistence of cartilage in the diaphysis of long bonesof transgenic animals (Tavella et al., 2004), defects of jointswere evident, in particular in those formed by mesenchymalsplitting (Fig. 2A, B).

Elbow/knee joint failed to form properly and the longbone epiphyses were dramatically misshapen. The olecranonprocess of ulna expanded posteriorly and the distal epiphysisof the humerus did not articulate with the radius and ulna asthe WT (Fig. 2E). In the knee of transgenic mice the distalpart of the femur failed to split from the proximal part of thetibia and included the patella in a continuous cartilaginousrod (Fig. 2F). The hand and foot (autopod) of both the fore-and the hindlimb of transgenic mice showed an undefinedarticular surface between the carpal/tarsal bones, fusion ofthe metacarpal/metatarsal bones with the phalanges and mis-sing interphalangeal joints (Fig. 2G, H). While these jointswere affected, pelvic and shoulder girdles developed nor-mally (Fig. 2C, D).

Histological staining with Toluidine Blue of E15.5 limbsshowed a persistent fibrous tissue in the elbow between thehumerus and ulna (Fig. 3B) and, in the autopod, lack of

polyadenylation addition region, BpA. Immunohistochemistry showing shh

Fig. 2. Skeletal analysis. Alizarin red/Alcian blue staining of forelimb (A, C, E, G) and hindlimb (B, D, F, H) of E18.5 embryos. WT limbs are shown on the left andCol2-SHH are on the right. In both transgenic forelimbs and hindlimbs, Alcian blue staining extends deep into the diaphysis of long bones. Proximal joints are normal(A–D). Arrowheads indicate elbow and knee joint severely impaired with fusion among some different skeletal elements (A, B, E, F). Arrow indicates an extendedprocess of the proximal end of transgenic ulna (E). Hand and foot of transgenic mice show fusion of phalanges (sindactylia) with lack of a clear definition of bothcarpal/tarsal metacarpal/metatarsal bones compared to WT (G, H).

391S. Tavella et al. / Matrix Biology 25 (2006) 389–397

cavitation between the carpal bones with the presence of acontinuous element, representing the metacarpal and the pha-langes rudiments, both at E15.5 and E18.5 (Fig. 3C, D). On theother hand, articular surfaces of the scapula and humeruswere normally shaped and separated by the synovial cavity(Fig. 2A).

Fig. 3. Forelimb histology. Toluidine blue staining of the forelimb at E15.5 (A–C) andof the limb. Proximal epiphyses of both the humerus and the femur of the transgenialterations at E15.5 of the transgenic forelimb (right) lacking a well-defined separatioas continuous cartilaginous anlagen without joint cavitation (C, D 5×).

2.2. Defective expression pattern of Gdf5, sFrp2 and TenascinC (TNC) genes in joints of transgenic mice

To investigate the process of joint formation in transgenicmice, we analyzed the expression pattern of several genes spa-tially restricted to developing joints. Among them, Gdf5, a

E18.5 (D) shows the morphology of the embryonic epiphysis of different bonescs exhibit distinct borders that define the joint cavitation (A 10×). Evidence ofn between humerus and ulna (B 10×). Phalangi of both hand and foot remaining

392 S. Tavella et al. / Matrix Biology 25 (2006) 389–397

member of the BMP family, is one of the best describedmolecular regulators of joint development (Brunet et al., 1998;Francis-West et al., 1996; Hartmann and Tabin, 2001; Merinoet al., 1999; Storm and Kingsley, 1999). Mutations in this geneand in its human ortholog, CDMP1, are responsible for defectsaffecting the appendicular skeleton such as reduced length of thelong bones, joint abnormalities and a reduced number of bonesin digits II–V (Gruneberg and Lee, 1973; Storm and Kingsley,1996).

sFrp2 is a secreted protein with a putative Wnt antagonistfunction expressed in joints (Leimeister et al., 1998; Dreyeret al., 2004). While at E13.5 days of development sFrp2expression is detectable in the perichondrium surrounding thecartilage condensation in the distal limb bud, at E15.5 its ex-pression becomes restricted to the joints and the mesenchymaltissue around them (R. Morello, personal communication).

In situ hybridization analysis was performed at E15.5 tocompare the expression pattern of both Gdf5 and sFrp2 intransgenic forelimb autopods versus the WT mice. Gdf5mRNAwas detected as expected in those regions where joint formationoccurs in the WT mice (Fig. 4B), but its expression was almostabsent in the presumptive interphalangeal and carpal joints ofthe transgenic mice (Fig. 4F). The sFrp2 mRNA expressionpattern was also altered: it failed to localize in the joints while itremained expressed in the perichondrium of the autopod bones,which is typical of an earlier developmental stage (Fig. 4G).This confirms the lack of proper articular joint formation and it

Fig. 4. Analysis of Joint marker expression. Sections of E15.5 forelimb hybridized witoluidine blue to define the area of the hand shown with radioactive in situ hybri(arrowhead) following the segmentation of the bones (B). Gdf5 expression is nearlyjoint formation in the embryo both in interphalangeal joints (⁎) and carpals (arrowheathe bones without defining the absent joint (G). Immunoperoxidase staining for Tenarticular chondrocytes and in the perichondrium of the bones. The expression pattern ielbow (arrowheads) (I).

is in agreement with the alizarin and the toluidine blue analysisof the transgenic mice (Fig. 2).

A further gene expressed at the articular surface is TenascinC (TNC), a large extracellular molecule transiently expressedduring development. Its expression is enriched in the chondro-genic cell condensation region of the chick limb bud at day 4 inthe embryo, becoming progressively more restricted in thearticular cap at later stages of development and maintained inthis region during postnatal life (Pacifici, 1995).

To evaluate whether TNC expression is altered in transgeniclong bones we performed an immunohistological staining witha specific rabbit antiserum against TNC. TNC was widelyexpressed in WT long bone of E15.5 (data not shown) andbecame restricted at E18.5 to the perichondrium and to thearticular surface. In E18.5 transgenic mice instead, althoughTNC restricted its expression towards the bone periphery, itremained expressed in a wider area of the bone rudiment respectto the WT mice (Fig. 4H, I). This is consistent with the effect ofSHH in blocking the differentiation of immature chondrocytes.

2.3. Shh increases chondrocyte proliferation in transgenics

Because in Col2-SHH mice the splitting of mesenchymeprimordia was affected, we examined how different cell beha-viors could influence the process of joint cavitation.

Several lines of evidence indicate that overexpression of Ihhin cartilage regulates chondrocytes proliferation both directly

th antisense probes for Gdf5 (B, F) and sFrp2 (C, G). Serial sections stained withdization (A, D). Gdf5 expression in the interphalangeal (⁎) and carpals jointsabsent in the interphalangeal joints (F). Similar to Gdf5, sFrp2 is present duringd) (C); in transgenics its expression remains confined to the perichondrium (⁎) ofascin C at the elbow of E18.5 embryos (H, I 10×). Tenascin C is present in thes slightly wider in the perichondrium and in the articular surface of the transgenic

393S. Tavella et al. / Matrix Biology 25 (2006) 389–397

and indirectly by upregulating PTHrP (Long et al., 2001;Minina et al., 2001; Vortkamp et al., 1996).

In this scenario, we first studied cell proliferation in Col2-SHH mice. We assessed chondrocyte proliferation in vivo byintraperitoneal injection of BrdU in pregnant females and mea-surement of BrdU incorporation into replicating cells. Accord-ing to Long et al. (2001) we considered for our analysis theproliferative zone close to the epiphyseal region of the growthplate. Immunoperoxidase staining for BrdU showed an in-creased number of BrdU positive chondrocytes in this area (Fig.5A, B); in particular, BrdU labeling was about two times moreabundant in Col2-SHH chondrocytes compared to the WT (Fig.5C). These experiments demonstrate a positive role of Shh incontrolling chondrocyte proliferation. Furthermore, while in thearticular region of normal mice, cells with BrdU incorporationwere absent, in the presumptive transgenic joints, some positivecells were found, showing that in this area, not all of the cellswent to quiescence prior to the start of the cavitation process(Fig. 5D, E).

Other genes cooperate with Ihh to regulate cell proliferationduring endochondral ossification. Among them, FGFR3 is an

Fig. 5. Proliferation, apoptosis assay. Immunoperoxidase staining with a monoclonaincorporation rate in transgenic chondrocytes (B) when compared to the WT (A). Blittermates (9.7%) (C). Very few BrdU labelled chondrocytes are present in the artipresumptive joint site (⁎) (E). Immunoperoxidase staining for FGFR3 in femur of E182-SHH mice (G) compared to the WT (F). Tunel assay of forelimbs at E15.5 (H, I 1apoptosis. Cells at the articular surface of the distal epiphysis of the humerus (arrowhother hand, the elbow of transgenic mouse is abnormal where joint fusion occurs, cellof the humeral epiphysis is detected (I). Immunoperoxidase staining for Bax at 15.5Eof the WT humerus (L), while it is absent in the transgenic counterpart (M).

important negative regulator of bone growth (Xu et al., 1999).Recent studies have also suggested that FGFR3 can down-regulate chondrocyte proliferation suppressing Ihh expression(Naski et al., 1998). To evaluate a possible role of this gene inCol2-SHH transgenic mice, we performed immunoperoxidasestaining at E18.5. Using a specific antiserum for FGFr3, we showthat its expression was severely reduced in transgenic chon-drocytes of the humerus, while it was strongly expressed in thepre-hypertrophic chondrocytes of WT growth plate (Fig. 5F, G).

Furthermore, we showed previously a peculiar organization oflong bone chondrocytes in these transgenic mice. Chondrocytesdid not form the typical mammal growth plate along the longi-tudinal axis of the stylopod, but they were dispersed in theextracellular matrix with some cells arranged in columns orientedfrom the center toward the periphery of the bone. In addition,these mice showed a low type X collagen expression in somechondrocytes close to the perichondrium (Tavella et al., 2004).The immunostaining for FGFR3 also showed that the very fewFGFR3 positive cells were scattered in transgenic cartilage notparticipating in the peculiar columnar organized collagen Xexpressing cells in proximity to the periosteum (Fig. 5G).

l antibody anti-BrdU of E15.5 humerus (A, B 20×). The staining reveals higherrdU labeling index of Col 2-SHH chondrocyte (22.5%) is 2.3 times that of WTcular surface of WT mice (D arrowheads) compared to those in the transgenic.5 forelimbs (F, G) shows few scattered positive cells for this receptor in the Col0×). Descriptive histology of the transgenic is shown in Fig. 3B (right) to studyead) undergo apoptosis defining the joint with the ulna in WT mouse (H). On thes failed to show evidence of apoptosis (arrowhead) and no clear border definition(L, M 20×). Bax protein is expressed in the hypertrophic cells of the metaphysis

394 S. Tavella et al. / Matrix Biology 25 (2006) 389–397

2.4. Chondrocyte over-expression of SHH reduces apoptosis

Recently, it was shown that Shh can block apoptosis, inducedby Ptc1, during chick neural tube development (Thibert et al.,2003).

Since Col2-SHH mice demonstrated the absence of limbjoints, we tested the possibility that the misexpression of SHHin chondrocytes could block the splitting of the mesenchyme,and affect joint formation.

In this context, we performed Tunel assays on E15.5 em-bryonic limb sections to evaluate whether apoptosis occursduring joint formation.

In both WT forelimb and hindlimb, Tunel assay revealed celldeath along the entire articular regions. In particular, a positivestaining was observed in the elbow cavity between the humerusand ulna (Fig. 5H), and in metacarpal and phalangeal rudiments(data not shown). Contrastingly, transgenic forelimbs, at dayE15.5, showed a reduced or absence of cell death in the cartilag-inous template, impairing joint cavitation. Therefore, presump-tive joint cannot assume a well-defined border (Fig. 5I).

In endochondral ossification, apoptosis is involved, besidesjoint formation, in programmed cell death of hypertrophicchondrocytes. Cytoplasmatic Bax is one of the proteins that cellstypically express during this process. When overexpressed, Baxcounters the antiapoptotic effects of Bcl-2 causing acceleratedcell death (Oltvai et al., 1993; Yin et al., 1994). Therefore, tobetter study the antiapoptotic role of Shh in our transgenic mice,we also evaluated Bax gene expression. Immunoperoxidasestaining on WT E18.5 limbs revealed that Bax was expressedonly in hypertrophic chondrocytes, cells that will subsequentlyundergo programmed cell death (Fig. 5L). On the contrary, Baxwas barely detectable or absent from the entire transgenic bonerudiments at the same stage (Fig. 5M).

Recently, it was found that Shh can also upregulate Bcl-2transcription (Bigelow et al., 2003). We have analyzed Bcl-2mRNA expression by in situ hybridization to determine whetherthis survival gene is expressed during endochondral ossification.

Normal proliferating chondrocytes express high levels of Bcl-2 until they become hypertrophic and begin to die. At this stage,Bcl-2 expression is switched off and replaced with the expressionof genes implicated in programmed cell death (Amling et al.,1997). In Col2-SHH mice, Bcl-2 expression was maintainedthroughout the skeletal element and persisted, without anydecrement, during limb development (data not shown).

3. Discussion

Limb formation is the result of sequential and tightly con-trolled events. In this scenario the hedgehog gene family playscrucial roles orchestrating different stages of skeletal develop-ment. SHH is expressed in the polarizing region of the limb budat a very early stage of development, where it determines theantero-posterior axis along which the different limb elementsdevelop (Johnson et al., 1994; Riddle et al., 1993; Riddle andTabin, 1999). In the growth plate both Sonic and Indian Hedge-hog are expressed by chondrocytes where they regulate the rateof chondrocyte differentiation and proliferation together with

other molecules (Wu et al., 2002). Furthermore, as misexpres-sion of both Shh and Ihh in the anterior region of the limb budcauses a mirror-image duplications of posterior limb structures,it was proposed that these two proteins have similar biologicalproperties (Vortkamp et al., 1996). This study focused on theconsequence of SHH chondrocyte overexpression during jointformation.

The analysis of the skeleton of these transgenic mice showedalterations in joints development indicating a defect in mes-enchymal splitting: elbow/knee joints were characterized by acontinuous cartilaginous tissue, several autopod componentswere absent and the phalangy were fused. Strikingly, however,pelvic and shoulder girdles developed normally and the jointsbetween ilium/femur and scapula/humerus were normal. Theseresults suggest that these joints develop through a distinct mo-lecular signaling. Similarly, the analysis of Emx2 homozygousmutants showed agenesia of both the scapula and ilium, whiledistal limb skeletal elements were normal, leading to the hypo-thesis that different signaling systems pattern the proximal andthe distal limb along the three axes (Pellegrini et al., 2001).

The fused phalangy phenotype was also shown in radiation-induced mouse mutant, called short digit (Dsh), with an inver-sion on chromosome 5 that resulted in a biphasic regulation ofShh expression (Niedermaier et al., 2005). These mice, in addi-tion to fused phalangy, also displayed short digits, a phenotypethat resembles human brachydactyly type A1. In Col2-SHHhowever, phalangy are not shorter than WT, most likely becausethe overexpression of Shh, under the control of type II collagenpromoter has already occurred at E9.5 (Zhou et al., 1995),whereas in the Dsh mice Shh overexpression was observed atlater stages, E13.5–E14.5 (Niedermaier et al., 2005).

During development, limb joint regions become evident whena signal specifies the presumptive joint position in the continuousmesenchymal rod. This happens around E12.5 in the mouse, by alocalized production of Wnt14, a secreted signaling moleculeexpressed prior to mesenchymal segmentation (Hartmann andTabin, 2001). Overexpression of Wnt14 in chick micromassculture showed that Wnt14 prevented the differentiation of cellaggregates into cartilage nodules indicating that Wnt14 acts as anantagonizer of chondrogenesis. In these aggregates Wnt14 wasinstructing prechondrogenic cells to become interzone cells, in-ducing Gdf5 expression.

We show that our transgenic mice failed to express jointmarkers such as sFrp2 and Gdf5 and, in addition, we failed todetect either Wnt14, the specific negative chondrogenic regulator(data not shown). This correlates with the fusion of skeletalelements observed in the genetic ablation of β-catenin/Wnt14signaling in early differentiating chondrocytes (Guo et al., 2004).

The absence of Wnt14, in inhibiting chondrogenesis in thepresumptive joints, could also allow a wider expression of Tncin the transgenic respect to WT bones typical of permanentcartilage (Pacifici, 1995). This pattern of Tnc expression in theepiphysis and perichondrium of our transgenic mice is con-sistent with the maintenance of type II collagen expressingchondrocytes induced by Shh. Therefore, we hypothesize thatShh acts by inhibiting the expression of Gdf5 and maintainingthe expression of chondrogenic markers, whose repression is

395S. Tavella et al. / Matrix Biology 25 (2006) 389–397

necessary to initiate joint formation, leading to defective jointcavitation.

Ihh is also relevant in the control of growth plate chon-drocyte proliferation. Long et al. have identified two differentzones in the growth plate with different proliferative rates:Zone 1, closer to the articular surface, incorporates less BrdUthan Zone 2, closer to the prehypertrophic zone (Long et al.,2001). Using the UAS-Gal4 system, Ihh was overexpressed incartilage and this increased, in Zone 1, of 2.5-fold the in-corporation of BrdU, leaving the labeling index in Zone 2largely unchanged.

In Col2-SHH mice, these two zones are not distinguishablebecause the stylopod is made only of immature cartilage,without the particular organization of the growth plate along thebone longitudinal axis. Here, like in Ihh overespression, Shhincreases the proliferative rate of chondrocytes in the presump-tive zone 1 compared to the WT. Furthermore, we also show thepresence of some positive cells in the regions were the jointsshould develop, indicating that Shh also alters the chondrocyteproliferative capability within the would-be “interzone.”

FGFR3 is one of the molecules that controls proliferation inthe growth plate. Depending from the developmental stage, itcan either promote or inhibit proliferation in the embryo (Iwataet al., 2001). Previous studies have also indicated that FGFR3downregulates chondrocyte proliferation suppressing Ihh ex-pression (Naski et al., 1998). The absence of FGFR3 in Col2-SHH mice suggests the existence of a negative regulatory loopbetween these two genes in regulating cell proliferation in thegrowth plate.

The involvement of apoptosis during joint development iscontroversial. During rat joint formation, although cavitationoccurred within the interzone, no apoptosis was observed, but arearrangement of this area with incorporation into the outerlayers of the central cells (Ito and Kida, 2000). Nonetheless,apoptosis was evidenced during murine interphalangeal jointformation at E14.5 (Niedermaier et al., 2005).

The joint between the stylopod and zeugopod is alreadyformed at E15.5 of development and we observed apoptosissculpting the borders of bones in normal mice, as shown byTunel assays. On the contrary, apoptosis is markedly reduced inthe cartilage of Col2-SHH mice demonstrating an antiapoptoticrole of shh.

Furthermore, in the growth plate, as hypertrophic chondro-cytes were absent, bax expression was undetectable in thetransgenic diaphysis of the stylopod. The whole stylopod oftransgenics instead expresses the antiapoptotic factor Bcl-2mRNA (data not shown). The ability of Shh to induce endo-genous Bcl-2 expression, through Gli1 transcription factor, waspreviously shown in keratinocytes (Bigelow et al., 2003).Amling et al. demonstrated that Bcl-2 is also downstream of thePTHrP signaling during skeletal development (Amling et al.,1998). It is well known that hedgehog proteins induce PTHrPin the articular region (Vortkamp et al., 1996). Via this mech-anism, SHH could enhance Bcl-2 expression via Gli1 or viaPTHrP or both, maintaining chondrocyte survival and contrib-uting to a lack of cavitation in transgenic animals. It should benoticed that the absence of a proper cavitation in Col2-SHH

mice cannot be due to an exclusive effect of PTHrP, since theCol2-PTHrP transgenic mice did not show joint alterations(Weir et al., 1996).

Lastly, it should be pointed out that in WT mice the cells ofthe future joint region are not in proximity to the Ihh expressionregion located in the growth plate of long bones. Instead, in ourtransgenics, Shh is unconfined in the prehypertrophic region.Therefore, Shh should be able to interact directly with cells inthe presumptive joint regions, altering their proliferation andapoptosis rate and sustaining the survival of type II collagenexpressing chondrocytes. This mechanism results in the inhibi-tion of joint cavitation and morphogenesis.

4. Experimental procedures

4.1. Generation of transgenic mice

Generation of transgenic mice and transgene expressionwere performed as previously described (Tavella et al., 2004).Briefly, the expression of the cDNA coding for the entire humanSHH was driven by the tissue-specific cis-element of murine α1(II) procollagen promoter/enhancer.

Microinjection was performed according to an establishedprocedure (Hogan et al., 1994) using the FVB mouse strain.Transgenic embryos were identified by both PCR and Southernblot of EcoRI digested genomic DNA hybridized with a trans-gene-specific probe (not shown).

4.2. Histology, immunohistochemistry and immunofluorescence

Embryos were removed by cesarean section at E15.5 andE18.5. The skeletons of E18.5 embryos were cleared in 1%KOH and stained with Alizarin red S and Alcian Blue (Sigma)as previously described (McLeod, 1980).

For histology, embryos were fixed in 4% formaldehydeovernight, dehydrated with increasing ethanol concentrationsolutions and embedded in paraffin. Five-micrometer-thick sec-tions were cut, deparaffinized and stained by Toluidine Blue.

For immunohistochemistry, deparaffinized sections weretreated with methanol and 3% H2O2 to inhibit endogenousperoxidase activity. Primary antibodies were a goat polyclonalantibody N-19 against murine Shh (Santa Cruz), a rabbit anti-TNC, (Kalembey et al., 1997), a mouse monoclonal anti-Bax(Santa Cruz) and a rabbit anti-FGFR3 (Santa Cruz). NormalIgGs (Santa Cruz) of the same species and at the same con-centration were used as controls for each antibody and nostaining was observed (not shown).

4.3. Proliferation and apoptosis assay

BrdU incorporation was performed by intraperitoneal injec-tion of pregnant mice with 50 μg BrdU/gram body weight 90 minbefore sacrifice. Embryos were collected by cesarean section andlimbs were dissected and embedded as described above. BrdUwas detected by immunohistochemistry with a monoclonal anti-BrdU antibody, according to the manufacturer's protocol(Roche), and expressed as percentage of incorporating cells.

396 S. Tavella et al. / Matrix Biology 25 (2006) 389–397

For Tunel assays, E15.5 embryos were collected, briefly fixedwith 4% formaldehyde/sucrose and embedded in OCT as pre-viously described (Barthel and Raymond, 1990). Five-microm-eter-thick sections were treated with the In Situ Cell Deathdetection kit, AP (Roche).

4.4. RNA probes and in situ hybridization

E15.5 wild-type (WT) and transgenic mouse embryos wereharvested. Both fore- and hind-limbs were collected, washed incold PBS and fixed in 4% paraformaldehyde/PBS at 4 °C for12 h. They were then dehydrated through an ethanol series,embedded in paraffin and 6-μm sections along both the A/P andD/Vaxis were generated. In situ hybridization was performed asdescribed (Albrecht et al., 1998). Antisense and sense ribop-robes were synthesized with T7, T3 or SP6 RNA polymerase inthe presence of [35S]UTP (1250 Ci/mmol; ICN). The mouseGdf5 cDNA probe corresponds to nucleotides 1735–2266 ofGenBank accession No. U08337 and was generated by RT-PCRamplification of NIH-3T3 cells RNA. Primers were sense, 5′-TTCATCGACTCTGCCAAC-3′; antisense, 5′-CATACTCTTCTCTTCACCCC-3′. The sFrp2 specific probe was generatedfrom a mouse EST clone (IMAGE clone I.D. 536389, kindlyprovided by C. Leimeister, Wuerzburg). Tissues were visualizedby fluorescence of Hoechst dye-stained nuclei and silver grainsby darkfield microscopy (Zeiss Axioskop). Digital images werecaptured with a SPOT-CCD camera.

Acknowledgments

We thank R. Arbicò for technical assistance, K. Imanaka-Yoshida for antitenascin antibody. This work was partiallysupported by the European (ESA-ERISTO) and Italian (ASI)Space Agencies (RC). We thank dott. William Giannobile fromthe University of Michigan, Ann Arbor, MI, USA, for editingthe manuscript.

References

Albrecht, U., Lu, H., Revelli, J., Xu, X., Lotan, R., Eichele, G., 1998. Studyinggene expression on tissue sections using in situ hybridization. In: Adolph, K.(Ed.), Human Genome Methods. CRC Press, Boca Raton, pp. 93–120.

Amling, M., Neff, L., Tanaka, S., Inoue, D., Kuida, K., Weir, E., Philbrick, W.M.,Broadus, A.E., Baron, R., 1997. Bcl-2 lies downstream of parathyroidhormone-related peptide in a signaling pathway that regulates chondrocytematuration during skeletal development. J. Cell Biol. 136, 205–213.

Amling, M., Posl, M., Hentz, M.W., Priemel, M., Delling, G., 1998. PTHrP andBcl-2: Essential regulatory molecules in chondrocyte differentiation andchondrogenic tumors. Verh. Dtsch. Ges. Pathol. 82, 160–169.

Archer, C.W., Dowthwaite, G.P., Francis-West, P., 2003. Development ofsynovial joints. Birth Defects Res. Part C Embryo Today 69, 144–155.

Barthel, L.K., Raymond, P.A., 1990. Improved method for obtaining 3-micronscryosections for immunocytochemistry. J. Histochem. Cytochem. 38,1383–1388.

Bigelow, R.L., Chari, N.S., Unden,A.B., Spurgers, K.B., Roop, D.R., Toftgard, R.,McDonnell, T.J., 2003. Transcriptional regulation of bcl-2 mediated by thesonic hedgehog signaling pathway through gli-1. J. Biol. Chem.

Brunet, L.J., McMahon, J.A., McMahon, A.P., Harland, R.M., 1998. Noggin,cartilage morphogenesis, and joint formation in the mammalian skeleton.Science 280, 1455–1457.

Capdevila, J., Izpisua Belmonte, J.C., 2001. Patterning mechanisms controllingvertebrate limb development. Annu. Rev. Cell Dev. Biol. 17, 87–132.

Coleman, C.M., Tuan, R.S., 2003. Growth/differentiation factor 5 enhanceschondrocyte maturation. Dev. Dyn. 228, 208–216.

Dreyer, S.D., Naruse, T., Morello, R., Zabel, B., Winterpacht, A., Johnson, R.L.,Lee, B., Oberg, K.C., 2004. Lmx1b expression during joint and tendonformation: localization and evaluation of potential downstream targets. GeneExpr. Patterns 4, 397–405.

Francis-West, P.H., Richardson,M.K., Bell, E., Chen, P., Luyten, F., Adelfattah,A.,Barlow, A.J., Brickell, P.M., Wolpert, L., Archer, C.W., 1996. The effect ofoverexpression of BMPs andGDF-5 on the development of chick limb skeletalelements. Ann. N.Y. Acad. Sci. 785, 254–255.

Garofalo, S., Kliger-Spatz, M., Cooke, J.L., Wolstin, O., Lunstrum, G.P.,Moshkovitz, S.M., Horton, W.A., Yayon, A., 1999. Skeletal dysplasia anddefective chondrocyte differentiation by targeted overexpression of fibroblastgrowth factor 9 in transgenic mice. J. Bone Miner. Res. 14, 1909–1915.

Gruneberg, H., Lee, A.J., 1973. The anatomy and development of brachypodismin the mouse. J. Embryol. Exp. Morphol. 30, 119–141.

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 for-mation. Genes Dev. 18, 2404–2417.

Hartmann, C., Tabin, C.J., 2001. Wnt-14 plays a pivotal role in inducingsynovial joint formation in the developing appendicular skeleton. Cell 104,341–351.

Hogan, B., Beddington, R., Costantini, F., Lacy, E., 1994. Manipulating themouse embryo: a laboratory manual, 2nd edition. Cold Spring HarborLaboratory Press, Cold Spring Harbor, NY, USA.

Ito, M.M., Kida, M.Y., 2000. Morphological and biochemical re-evaluation ofthe process of cavitation in the rat knee joint: cellular and cell strataalterations in the interzone. J. Anat. 197, 659–679.

Iwata, T., Li, C.L., Deng, C.X., Francomano, C.A., 2001. Highly activated Fgfr3with the K644M mutation causes prolonged survival in severe dwarf mice.Hum. Mol. Genet. 10, 1255–1264.

Johnson, R.L., Riddle, R.D., Laufer, E., Tabin, C., 1994. Sonic hedgehog: a keymediator of anterior–posterior patterning of the limb and dorso-ventralpatterning of axial embryonic structures. Biochem. Soc. Trans. 22, 569–574.

Kalembey, I., Yoshida, T., Iriyama, K., Sakakura, T., 1997. Analysis of tenascinmRNA expression in the murine mammary gland from embryogenesis tocarcinogenesis: An in situ hybridization study. Int. J. Dev. Biol. 41, 569–573.

Lanske, B., Karaplis, A.C., Lee, K., Luz, A., Vortkamp, A., Pirro, A., Karperien,M., Defize, L.H., Ho, C., Mulligan, R.C., Abou-Samra, A.B., Juppner, H.,Segre, G.V., Kronenberg, H.M., 1996. PTH/PTHrP receptor in earlydevelopment and Indian hedgehog-regulated bone growth. Science 273,663–666.

Leimeister, C., Bach, A., Gessler, M., 1998. Developmental expression patternsof mouse sFRP genes encoding members of the secreted frizzled relatedprotein family. Mech. Dev. 75, 29–42.

Long, F., Zhang, X.M., Karp, S., Yang, Y., McMahon, A.P., 2001. Geneticmanipulation of hedgehog signaling in the endochondral skeleton reveals adirect role in the regulation of chondrocyte proliferation. Development 128,5099–5108.

McLeod, M.J., 1980. Differential staining of cartilage and bone in whole mousefetuses by Alcian blue and alizarin red S. Teratology 22, 299–301.

Merino, R., Macias, D., Ganan, Y., Economides, A.N., Wang, X., Wu, Q., Stahl,N., Sampath, K.T., Varona, P., Hurle, J.M., 1999. Expression and function ofGdf-5 during digit skeletogenesis in the embryonic chick leg bud. Dev. Biol.206, 33–45.

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.

Naski, M.C., Colvin, J.S., Coffin, J.D., Ornitz, D.M., 1998. Repression ofhedgehog signaling and BMP4 expression in growth plate cartilage byfibroblast growth factor receptor 3. Development 125, 4977–4988.

Niedermaier, M., Schwabe, G.C., Fees, S., Helmrich, A., Brieske, N., Seemann,P., Hecht, J., Seitz, V., Stricker, S., Leschik, G., Schrock, E., Selby, P.B.,Mundlos, S., 2005. An inversion involving the mouse Shh locus results inbrachydactyly through dysregulation of Shh expression. J. Clin. Invest. 115,900–909.

397S. Tavella et al. / Matrix Biology 25 (2006) 389–397

Oltvai, Z.N., Milliman, C.L., Korsmeyer, S.J., 1993. Bcl-2 heterodimerizes invivo with a conserved homolog, Bax, that accelerates programmed celldeath. Cell 74, 609–619.

Pacifici, M., 1995. Tenascin-C and the development of articular cartilage. MatrixBiol 14, 689–698.

Pellegrini,M., Pantano, S., Fumi,M.P., Lucchini, F., Forabosco, A., 2001. Agenesisof the scapula in Emx2 homozygous mutants. Dev. Biol. 232, 149–156.

Peters, C., Wolf, A., Wagner, M., Kuhlmann, J., Waldmann, H., 2004. Thecholesterolmembrane anchor of theHedgehog protein confers stablemembraneassociation to lipid-modified proteins. Proc. Natl. Acad. Sci. U. S. A. 101,8531–8536.

Polinkovsky, A., Robin, N.H., Thomas, J.T., Irons, M., Lynn, A., Goodman, F.R.,Reardon, W., Kant, S.G., Brunner, H.G., van der Burgt, I., Chitayat, D.,McGaughran, J., Donnai, D., Luyten, F.P., Warman, M.L., 1997. Mutations inCDMP1 cause autosomal dominant brachydactyly type C. Nat. Genet. 17,18–19.

Riddle, R.D., Tabin, C., 1999. How limbs develop. Sci. Am. 280, 74–79.Riddle, R.D., Johnson, R.L., Laufer, E., Tabin, C., 1993. Sonic hedgehog

mediates the polarizing activity of the ZPA. Cell 75, 1401–1416.Settle Jr., S.H., Rountree, R.B., Sinha, A., Thacker, A., Higgins, K., Kingsley, D.M.,

2003.Multiple joint and skeletal patterning defects caused by single and doublemutations in the mouse Gdf6 and Gdf5 genes. Dev. Biol. 254, 116–130.

Spitz, F., Duboule, D., 2001. Development. The art of making a joint. Science29, 11713–11714.

St-Jacques, B., Hammerschmidt, M., McMahon, A.P., 1999. Indian hedgehogsignaling regulates proliferation and differentiation of chondrocytes and isessential for bone formation. Genes Dev. 13, 2072–2086.

Storm, E.E., Kingsley, D.M., 1996. Joint patterning defects caused by single anddouble mutations in members of the bone morphogenetic protein (BMP)family. Development 122, 3969–3979.

Storm, E.E., Kingsley, D.M., 1999. GDF5 coordinates bone and joint formationduring digit development. Dev. Biol. 209, 11–27.

Storm, E.E., Huynh, T.V., Copeland, N.G., Jenkins, N.A., Kingsley, D.M., Lee,S.J., 1994. Limb alterations in brachypodism mice due to mutations in a newmember of the TGF beta-superfamily. Nature 368, 639–643.

Tavella, S., Biticchi, R., Schito, A., Minina, E., Di Martino, D., Pagano, A.,Vortkamp, A., Horton, W.A., Cancedda, R., Garofalo, S., 2004. Targetedexpression of SHH affects chondrocyte differentiation, growth plateorganization, and Sox9 expression. J. Bone Miner. Res. 19, 1678–1688.

Thibert, C., Teillet, M.A., Lapointe, F., Mazelin, L., Le Douarin, N.M., Mehlen,P., 2003. Inhibition of neuroepithelial patched-induced apoptosis by sonichedgehog. Science 301, 843–846.

Vortkamp, A., Lee, K., Lanske, B., Segre, G.V., Kronenberg, H.M., Tabin, C.J.,1996. Regulation of rate of cartilage differentiation by Indian hedgehog andPTH-related protein. Science 273, 613–622.

Weir, E.C., Philbrick, W.M., Amling, M., Neff, L.A., Baron, R., Broadus, A.E.,1996. Targeted overexpression of parathyroid hormone-related peptide inchondrocytes causes chondrodysplasia and delayed endochondral boneformation. Proc. Natl. Acad. Sci. U. S. A. 93, 10240–10245.

Wu, L.N., Lu, M., Genge, B.R., Guo, G.Y., Nie, D., Wuthier, R.E., 2002.Discovery of sonic hedgehog expression in postnatal growth plate chon-drocytes: differential regulation of sonic and Indian hedgehog by retinoicacid. J. Cell. Biochem. 87, 173–187.

Xu, X., Weinstein, M., Li, C., Deng, C., 1999. Fibroblast growth factor receptors(FGFRs) and their roles in limb development. Cell Tissue Res. 296, 33–43.

Yin, X.M., Oltvai, Z.N., Korsmeyer, S.J., 1994. BH1 and BH2 domains of Bcl-2are required for inhibition of apoptosis and heterodimerization with Bax.Nature 369, 321–323.

Zhou, G., Garofalo, S., Mukhopadhyay, K., Lefebvre, V., Smith, C.N.,Eberspaecher, H., de Crombrugghe, B., 1995. A 182 bp fragment of themouse pro alpha 1(II) collagen gene is sufficient to direct chondrocyteexpression in transgenic mice. J. Cell Sci. 108 (Pt 12), 3677–3684.