During endochondral ossification, the skeletal elements of vertebrate limbs form and elongate via coordinated control of chondrocyte and osteoblast differentiation and proliferation. The role of signaling by the ErbB family of receptor tyrosine kinases, which consists of ErbB1 (epidermal growth factor receptor or EGFR), ErbB2, ErbB3 and ErbB4, has been little studied during cartilage and bone development. Signaling by the ErbB network generates a diverse array of cellular responses via formation of ErbB dimers activated by distinct ligands that produce distinct signal outputs. Herstatin is a soluble ErbB2 receptor that acts in a dominant negative fashion to inhibit ErbB signaling by binding to endogenous ErbB receptors, preventing functional dimer formation. Here, we examine the effects of Herstatin on limb skeletal element development in transgenic mice, achieved via Prx1 promoter-driven expression in limb cartilage and bone. The limb skeletal elements of Prx1-Herstatin embryos are shortened, and chondrocyte maturation and osteoblast differentiation are delayed. In addition, proliferation by chondrocytes and periosteal cells of Prx1-Herstatin limb skeletal elements is markedly reduced. Our study identifies requirements for ErbB signaling in the maintenance of chondrocyte and osteoblast proliferation involved in the timely progression of chondrocyte maturation and periosteal osteoblast differentiation.
Vertebrate limb skeletal elements develop through a process known as endochondral ossification, in which cartilage model templates prefigure bone formation (de Crombrugghe et al. 2001; Wagner & Karsenty 2001; Provot & Schipani 2005). In the first step of this process, prechondrogenic mesenchymal cells aggregate and undergo overt chondrogenic differentiation to form cartilage condensations that undergo rapid longitudinal and appositional growth as a result of proliferation and matrix deposition to form the cartilage models. The chondrocytes of the cartilage models then undergo a coordinated program of maturation, whereupon they cease rapid proliferation, alter their gene expression profiles, and change their morphological appearance as they progress from proliferating to prehypertrophic to hypertrophic chondrocytes (Kronenberg 2003). Hypertrophic chondrocytes are required to signal the formation of the periosteal bony collar (Olsen et al. 2000) and the continued maturation of immature proliferating chondrocytes into hypertrophying chondrocytes and their removal to form the marrow space is essential for subsequent longitudinal growth of the long bones (de Crombrugghe et al. 2001; Wagner & Karsenty 2001).
The ErbB family of receptor tyrosine kinases consists of ErbB1 (also known as epidermal growth factor receptor or EGFR, also called Heregulin), three other receptors (ErbB2, ErbB3, and ErbB4), and multiple ligands including epidermal growth factor (EGF), transforming growth factor-α (TGF-α), and others (Yarden & Sliwkowski 2001; Jorissen et al. 2003; Citri & Yarden 2006). Following ligand binding, receptor homo- or hetero- dimer formation leads to ErbB auto- or trans-phosphorylation and recruitment of downstream signal effectors (Yarden & Sliwkowski 2001). Only ErbB1 and ErbB4 can form functional ligand-activated homodimers, as ErbB3 is kinase-inactive, and ErbB2 lacks a soluble ligand and a ligand binding pocket revealed by the crystal structure of its ectodomain (Garrett et al. 2003). Although ErbB heterodimers can form in all combinations, ErbB2 is the universally used and preferred heterodimer partner, and its participation in a heterodimer intensifies and modifies the resultant signal (Graus-Porta et al. 1997; Olayioye et al. 1998; Citri et al. 2003). Signals mediated by heterodimers are thought to account for most ErbB biological activities (Olayioye et al. 2000; Citri & Yarden 2006).
The dominant negative ErbB2 receptor known as Herstatin (Doherty et al. 1999) is a soluble ErbB2 receptor consisting of the first two subdomains of the extracellular domain of human ErbB2, as well as a unique ECDIIIa domain encoded by intron 8, which confers high affinity binding to ErbB receptor ectodomains (Doherty et al. 1999; Shamieh et al. 2004). Although Herstatin is an endogenously expressed human gene (Doherty et al. 1999), the unique peptide encoded by intron 8 is not conserved in non-primates such as mouse, rat or cow (Shamieh et al. 2004 and data not shown); suggesting Herstatin homologs in these species may not exist. The 1.4 kb human dnErbB2/Herstatin was cloned into the 2.4 kb mouse Prx1 promoter construct containing the enhancer elements that direct expression to the limbs and include a necessary 530 bp core element (Martin & Olson 2000). The Prx1 promoter directs transgene expression throughout the developing limb mesoderm and limb skeletal elements (Martin & Olson 2000; Akiyama et al. 2005). Prx1-dnErbB2 transgenic mice were generated by the University of Connecticut Health Center (UCHC) Gene Targeting and Transgenic Facility and genotyped using polymerase chain reaction (PCR). Expression of the Herstatin transgene in limb skeletal elements of transgenic mice was confirmed by in situ hybridization using a 238 bp human ErbB2 cDNA probe corresponding to the region of intron 8 that encodes the unique ECDIIIa domain of Herstatin, which is not conserved in mouse (Shamieh et al. 2004 and data not shown).
For whole mount skeletal staining, day 15.5, 16.5 and 18.5 littermate embryos were fixed in 95% ethanol, defatted in acetone and stained with 0.015% alcian blue. After destaining and initial clearing in KOH, embryos were stained with 0.01% alizarin red, cleared in KOH/glycerol and stored in 100% glycerol. Fore and hind limbs were removed and photographed. The lengths of the skeletal elements were directly measured on photographs taken at the same magnification and compared by independent groups t-test between means using statistical software (StatPac, Bloomington, MN, USA). For morphological analysis, limbs were subsequently dehydrated and embedded in paraffin, and adjacent sections were re-stained in alizarin red or hematoxylin and eosin.
For in situ hybridization including marker gene expression, limbs were fixed in 4% paraformaldehyde, embedded in paraffin and adjacent sections subjected to radioactive in situ hybridization using stringent hybridization conditions (55°C, 50% formamide) as described (Wang et al. 2004a). The cDNA probes used corresponded to the following: 345 bp of the intracellular domain of mouse ErbB2; 650 bp of mouse type X collagen; 1497 bp mouse Indian Hedgehog (Ihh); 750 bp mouse bone sialoprotein (BSP); and 430 bp mouse osteocalcin (OC).
To analyze cell proliferation, day 15.5 limbs were fixed in 4% paraformaldehyde, embedded in paraffin, and sections were subjected to immunohistochemistry with a polyclonal antiphospho-histone H3 antibody (p-histone H3, Upstate) (Fisher et al. 2005) using antigen-retrieval (Dako, Carpinteria, CA, USA). Cell proliferation was quantified as previously described (Fisher et al. 2005) using image analysis (Photoshop) in fixed areas of 20× digital photographs of skeletal tissue. For cartilage, a 60 × 100 mm area encompassing most of the proximal humerus, and a 40 × 100 mm area encompassing most of the distal humerus, was examined. For perichondrium/periosteum, a 10 × 50 mm area of the perichondrium/periosteum adjacent to the mid-diaphyseal region of the humerus (the region that contains the preosteoblasts and osteoblasts of the developing periosteal bony collar (Olsen et al. 2000) was examined. P-histone-labeled cells within the fixed area were automatically selected by color range, and a histogram was generated of the pixel area of the labeled cells. Eight to twenty sections each of proximal humerus, distal humerus or humerus perichondrium/periosteum were examined from the forelimbs of two different wild type or Prx1-Herstatin littermate embryos. The area of p-histone-labeled cells was normalized to the tissue area examined to obtain a value for cell proliferation per unit area of skeletal tissue, and these values were compared by independent groups t-test between means using statistical software (StatPac Inc.).
Endogenous ErbB2 and transgenic Herstatin expression in developing limb skeletal elements
The patterns of expression of endogenous ErbB2 and transgenic human Herstatin (truncated soluble ErbB2) in wild type and Prx1-Herstatin limb skeletal elements were examined by in situ hybridization (Fig. 1). In wild type limbs, expression of endogenous ErbB2 by developing cartilage elements was first seen in the digit rudiments at day 13.5 (Fig. 1A,B), and ErbB2 continued to be expressed at lower levels by all chondrocytes of the limb skeletal elements at day 15.5, including proliferative, prehypertrophic and hypertrophic chondrocytes (Fig. 1C,D). Endogenous ErbB2 continued to be expressed at low levels by the chondrocytes of the proliferative regions, and was also expressed at higher levels by the developing perichondrium/perisoteum at day 16.5 (Fig. 1E,F). Transgenes driven by the Prx1 promoter are expressed throughout the undifferentiated mesenchyme of the limb buds shortly after their formation (Martin & Olson 2000), and subsequently expressed the cartilage and bone of the limb skeletal elements (Akiyama et al. 2005). Consistent with this, transgenic human Herstatin was expressed by the chondrocytes as well as the developing perichondrium/periosteum of day 15.5 Prx1-Herstatin limb skeletal elements (Fig. 1G,H), in a domain encompassing the domain of expression of ErbB2 in wild type skeletal elements.
Limb skeletal elements of Prx1-Herstatin embryos are shortened, and chondrocyte maturation and periosteal osteoblast differentiation are delayed
Whole mount alcian blue/alizarin red staining indicated the skeletal elements of day 15.5 Prx1-Herstatin limbs were shortened compared to littermate wild type control limbs (Fig. 2A,B). The lengths of the limb skeletal elements (humerus and ulna for forelimbs, and femur and tibia for hindlimbs) of Prx1-Herstatin and wild type control littermate embryos were quantified by measurement and compared at days 15.5, 16.5 and 18.5 (Table 1). At day 15.5, the lengths of the limb skeletal elements of Prx1-Herstatin embryos were reduced compared to the lengths of the corresponding skeletal elements of littermate wild type control embryos, and this reduction was significant for the humerus, ulna and femur (Table 1). At day 16.5, the lengths of the limb skeletal elements of Prx1-Herstatin embryos were similarly reduced, but the reductions in length were only significant for the tibia (Table 1). By day 18.5, the lengths of the limb skeletal elements of Prx1-Herstatin embryos were comparable to littermate wild type controls (Table 1), indicating that the effects of Herstatin were overcome with time.
Table 1. The lengths of the skeletal elements of Prx-1 Herstatin transgenic limbs are reduced at day 15.5 and day 16.5 compared to control wild type littermate limbs, but not at day 18.5
Wild type (n = 8)
Prx1-Herstatin (n = 2)
Wild type (n = 10)
Prx1-Herstatin (n = 2)
Wild type (n = 12)
Prx1-Herstatin (n = 6)
Skeletal element lengths (mm) were measured from photographs of whole mount Alcian Blue Alizarin Red-stained limbs at 1.6× (day 15.5 and day 16.5) or 1× (day 18.5). Values shown (mean ± SEM) are the averaged lengths of the skeletal elements of each limb of littermate transgenic or control wild type embryos from one litter each at days 15.5 and 16.5. At day 18.5, data from three separate litters were combined. *P ≤ 0.1 versus control.
The histological appearance of the limb skeletal elements of Prx1-Herstatin and littermate wild type control embryos was also examined (Fig. 2C–F). Sectioning and alizarin red staining indicates that the hypertrophic chondrocytes located in the central portion of the diaphyses of the skeletal elements of wild type control fore- and hind- limbs had begun to secrete a mineralized matrix, and were being removed and replaced by marrow space and spicules of trabecular bone (Fig. 2C,E). In contrast, little or no mineralized matrix, marrow space and trabecular bone were present in the central portion of the diaphyses of the skeletal elements of littermate Prx1-Herstatin fore and hind limbs, rather, this region was occupied nearly completely by enlarged chondrocytes resembling hypertrophic chondrocytes (Fig. 2D,F). By day 18.5, the histological appearance of the limb skeletal elements of Prx1-Herstatin and littermate wild type control embryos was comparable (not shown).
To determine if molecular changes were apparent in Prx1-Herstatin limb skeletal elements, the patterns of expression of genes characteristic of chondrocyte maturation and bone formation in wild type control and littermate Prx1-Herstatin limb skeletal elements were examined by in situ hybridization (Figs 3, 4). In the diaphysis of wild type control limb skeletal elements, Indian Hedgehog (Ihh), a marker of prehypertrophic chondrocytes (St-Jacques et al. 1999), and type X collagen (Col x), a marker of hypertrophic chondrocytes (Apte et al. 1992; LuValle et al. 1992), were expressed in distinct domains, separated by the developing marrow space and trabecular bone (Fig. 3A,C). However, Ihh and Col × were expressed in single, virtually continuous domains in the central portions of Prx1-Herstatin limb skeletal elements (Fig. 3B,D). BSP is a marker of terminally differentiated chondrocytes and early osteoblasts (Young et al. 1994; Cowles et al. 1998). In wild type control limb skeletal elements, two distinct domains of low-level BSP expression in hypertrophic chondrocytes were present, which were separated by a high-level domain of BSP expression in osteoblasts of the developing marrow space and trabecular bone (Fig. 4A,C). However, in littermate Prx1-Herstatin limb skeletal elements, low-level BSP expression in hypertrophic chondrocytes occurred as a single, virtually continuous domain (Fig. 4B,D), similar to the single continuous domains of Ihh and Col × expression by the hypertrophic chondrocytes of Prx1-Herstatin skeletal elements shown in Figure 3, and the central domain of high-level BSP expression by osteoblasts of the trabecular bone was absent. In addition, expression of BSP by the early periosteal osteoblasts of wild type control limb skeletal elements was robust (Fig. 4A,C), but BSP expression by the early periosteal osteoblasts of littermate Prx1-Herstatin limb skeletal elements was weak (Fig. 4B,D). Osteocalcin (OC), a marker of later osteoblast differentiation (Rodan & Noda 1991), was expressed by the periosteal cells adjacent to the central portion of wild type control limb skeletal elements (Fig. 4E,G), but was not expressed by the corresponding region of the periosteum of littermate Prx1-Herstatin limb skeletal elements (Fig. 4F,H). By day 18.5, expression of BSP and OC was comparable in the trabecular bone and periosteum of the limb skeletal elements of wild type control and littermate Prx1-Herstatin embryos (not shown).
Proliferation by the chondrocytes and differentiating periosteal cells of Prx1-Herstatin humeri was markedly reduced
To determine if the gross, histological and molecular changes observed in day 15.5 Prx1-Herstatin limb skeletal elements were accompanied by changes at the cellular level, cell proliferation, determined by immunohistochemical detection of phosphorylated histone-labeling of dividing cells, was examined in wild type control and littermate Prx1-Herstatin humeri. Numerous proliferating chondrocytes were present in the proliferative zones of the proximal and distal epiphyses of wild type control humeri (Fig. 5A,C), but few were present in the proximal and distal proliferative zones of the littermate Prx1-Herstatin humeri (Fig. 5B,D). Similarly, abundant proliferating cells were observed in the periosteum of the central portion of wild type control humeri (Fig. 5E,G) but few in the periosteum of the central portion of littermate Prx1-Herstatin humeri (Fig. 5F,H). The numbers of proliferating cells per unit area of distal and proximal proliferating zones, and of periosteum, of wild type control and Prx1-Herstatin humeri were quantified. As shown in Table 2, proliferation per unit cellular area by the chondrocytes and by the differentiating periosteum (Fisher et al. 2005) was reduced by approximately 60% in Prx1-Herstatin humeri compared to littermate wild type control humeri.
Table 2. Cell proliferation by the chondrocytes of the proximal and distal proliferative zones, and by the cells of the periosteum, of day 15.5 Prx1-Herstatin humeri is reduced compared to control wild type littermate day 15.5 humeri
Wild type (n)
Cell proliferation per unit area (mm2) of skeletal tissue was quantified in fixed areas of the proximal proliferative zone, distal proliferative zone, and periosteum of 40× photographs of day 15.5 Prx1-Herstatin and control wild type littermate humeri. Values shown (mean ± SEM) are the averaged areas of p-histone-labeled cells normalized to the tissue area examined. (n) is the total number of sections examined from the humeri of two wild type and two littermate Prx1-Herstatin embryos. *P ≤ 0.001 versus control.
Involvement of ErbB2-containing ErbB receptor heterodimers during endochondral ossification
One mechanism of action of Herstatin in inhibiting ErbB function is to bind irreversibly to the extracellular domain of membrane-bound wild type ErbB2 (Doherty et al. 1999), and studies using confocal microscopy have directly demonstrated colocalization of transfected Herstatin with endogenous ErbB2 expressed by cells in culture, resulting in the formation of complexes comprised of Herstatin and wild type endogenous ErbB2 (Hu et al. 2005a, 2006). We have found that endogenous ErbB2 is expressed by immature and maturing chondrocytes as well as periosteal bone of developing mouse skeletal elements. In the developing skeletal elements of Prx1-Herstatin limbs, Herstatin is expressed in the cartilage and periosteum in a domain encompassing that of wild type endogenous ErbB2, demonstrating that the Herstatin transgene is expressed in the appropriate target tissue, which also expresses endogenous ErbB2. The end result of Herstatin binding to endogenous ErbB2 is to block its availability and/or ability to serve as a heterodimer partner for other ErbBs (Azios et al. 2001; Justman & Clinton 2002; Hu et al. 2006), inhibiting signaling by multiple ErbB receptor combinations. Expression of Herstatin in the developing skeletal elements of Prx1-Herstatin limbs could thus block formation of functional heterodimers comprised of endogenous ErbB2 in combination with another endogenously coexpressed ErbB. In this regard, developing mouse cartilage and bone cells and tissue also express ErbB1 and/or ErbB4 (Davideau et al. 1995; Nawachi et al. 2002), and roles for ErbB1 and/or ErbB4 signaling in skeletal development in vivo have been demonstrated in our own studies (Omi et al. 2005) and by others (Chan & Wong 2000; Chien et al. 2000; Nawachi et al. 2002; Sibilia et al. 2003; Wang et al. 2004b; Qin et al. 2005; and see below). Herstatin expression in the developing skeletal elements of Prx1-Herstatin limbs could also have direct effects on ErbB1/ErbB4 signals, as Herstatin has been shown to bind to and block ErbB1 homodimers (Azios et al. 2001) and potentially also ErbB4 homodimers (Shamieh et al. 2004). The results of our present study, while supporting known roles for ErbB1 and ErbB4 homodimers in endochondral ossification, also implicate for the first time ErbB2-containing heterodimers as endogenous ErbB signals in developing cartilage and bone.
ErbB2/ErbB signaling maintains the proliferation of maturing chondrocytes
Appropriate regulation of the rate of proliferation by the chondrocytes of the proliferative zones at the epiphyses of the developing limb skeletal elements is involved in the normal onset and progression of chondrocyte maturation to hypertrophy, which in turn drives elongation of the limb skeletal elements (Kronenberg 2003). Herstatin has been shown to block cell proliferation and/or growth in vitro (Doherty et al. 1999; Azios et al. 2001; Justman & Clinton 2002; Jhabvala-Romero et al. 2003; Hu et al. 2006) as well as in vivo (Staverosky et al. 2005). In this study, we have found that proliferation by the chondrocytes of the proliferative zones of the shortened Prx1-Herstatin limb skeletal elements is reduced by nearly 60%. We suggest that the mechanism by which Herstatin expression shortens the skeletal elements and delays chondrocyte maturation is to block ErbB-mediated proliferation by chondrocytes of the proliferative zones, reducing the pool of chondrocytes available to subsequently enter the maturation program. An endogenous role for ErbB2-containing heterodimers in chondrocyte proliferation is supported by our finding that ErbB2 expression in developing cartilage appears strongest in regions that contain abundant proliferating chondrocytes, such as the digit rudiments and the proliferative zones. Other ErbB signals may also be involved, as ErbB1 and/or ErbB4 are expressed by developing cartilage (Davideau et al. 1995; Nawachi et al. 2002), and chondrocyte proliferation is stimulated by ErbB1 and/or ErbB4 ligands (Ribault et al. 1997; Ishizeki et al. 2001; Ma & Lozanoff 2002; Qin et al. 2005). Although another study found no differences in the proliferative rate of the chondrocytes of the proliferative zones of the limb skeletal elements of ErbB1 null mice at birth (Sibilia et al. 2003), it is possible or even likely that simultaneous loss of function of more than one ErbB, such as is expected to occur as the result of Herstatin expression in this study, is required to effectively block chondrocyte proliferation. These observations are consistent with the possibility that ErbB signaling, perhaps using endogenously expressed ErbB2-containing heterodimers, is involved in mediating chondrocyte proliferation in vivo.
The ErbB signaling network is considered a paradigm for signaling diversity in part via its ability to integrate, antagonize and/or modify signals emanating from other receptor classes (Prenzel et al. 2001; Vivekanand & Rebay 2006). Our study demonstrates a positive role for endogenous ErbB2/ErbB signaling in maintaining proliferation of the chondrocytes of the proliferative zones of the developing limb skeletal elements. Other signals that positively regulate chondrocyte proliferation include signals mediated by the insulin-like growth factor (IGF), fibroblast growth factor (FGF) and Wnt families. Since ErbB signals are known to complement IGF (Hallak et al. 2002; Belaus et al. 2003), FGF (Schlessinger 2004) and Wnt (Musgrove 2004) signals, an intriguing possibility is that the ErbB receptor family may function in conjunction with one or more of these signaling networks to maintain chondrocyte proliferation during normal limb skeletal element development.
Role of ErbB2/ErbB signaling in periosteal proliferation and differentiation
The periosteal bony collar of the developing long bones forms when osteogenic precursors present in the perichondrium undergo proliferation and differentiation into periosteal osteoblasts (Olsen et al. 2000). Our findings that proliferation by the cells of the developing periosteum of Prx1-Herstatin limb skeletal elements is markedly reduced, and is accompanied by delayed expressed of osteocalcin, a marker of differentiated osteoblasts (Rodan & Noda 1991), suggests a requirement for endogenous ErbB2/ErbB signals in maintenance of proliferation and timely differentiation of periosteal osteoblasts. Studies show that a signal required for periosteal osteoblast differentiation is Ihh expressed by the adjacent prehypertrophic chondrocytes, as mice lacking Ihh expression in maturing chondrocytes do not form periosteal bone (Razzaque et al. 2005). Since the normal timing of the expansion of the domain of Ihh is delayed in Prx1-Herstatin limb skeletal elements, it is possible that the delayed differentiation and reduced proliferation of the periosteum of Prx1-Herstatin limb skeletal elements could be an indirect consequence of the effects of Herstatin on cartilage, which serve to delay the maturation process. However, other studies directly implicate ErbB signaling in stimulation of osteoblast proliferation and/or differentiation. ErbB1 and ErbB4 are expressed by primary osteoblast cells or osteoblast cell lines (Ng et al. 1983; Davideau et al. 1995; Qin et al. 2005), and the growth and proliferation of osteoblasts in vitro is stimulated by ligands that activate ErbB1 and/or ErbB4 (Chien et al. 2000; Qin et al. 2005) and are inhibited when ErbB1 signaling is blocked (Chien et al. 2000). Moreover, osteoblast growth and proliferation in vivo is stimulated by generalized overexpression of an ErbB1 ligand (Chan & Wong 2000) and impaired in ErbB1 null mice (Sibilia et al. 2003), and periosteal bone thickness is reduced in the skeletal elements of mice lacking amphiregulin, a ligand that activates ErbB1 and/or ErbB4 (Qin et al. 2005). These observations, along with our own finding that endogenous ErbB2 is itself highly expressed in developing periosteum, and that Prx1-driven transgenic Herstatin expression is particularly abundant in the periosteum of Prx1-Herstatin limb skeletal elements, suggests the possibility that the impaired periosteal proliferation and delayed differentiation of Prx1-Herstastin limb skeletal elements may be a direct effect of Herstatin expression and subsequent loss of ErbB function in the periosteal osteoblasts themselves.
Functional compensation may facilitate rescue of the Herstatin phenotype
The inhibitory effects of Herstatin expression on skeletal element elongation, chondrocyte maturation and periosteal bone formation are apparent at day 15.5, but not at day 18.5, indicating that the functional ErbB requirement blocked by Herstatin expression during endochondral ossification is compensated for over time. Studies suggest that functional compensation among coexpressed ErbB members may mask potential phenotypes that may otherwise be obtained following genetic loss of individual ErbBs (Xian et al. 2001; Wong 2003). Functional compensation between ErbB2 and ErbB1 has been directly demonstrated in vitro by the induction of ErbB1 expression and activation of ErbB1 signaling in response to loss of ErbB2 function (Hu et al. 2005b). Functional compensation has also been suggested to occur between ErbB1 and ErbB4, as developmental brain defects in mice caused by astrocyte-specific loss of function of ErbB4 or ErbB1 are exacerbated when ErbB4 and ErbB1 are inhibited together (Prevot et al. 2005). These observations suggest that ErbB signals mediated by ErbB1 or ErbB4 expressed by developing cartilage and bone (Davideau et al. 1995; Nawachi et al. 2002) may ultimately compensate for the loss of function of ErbB signaling during chondrocyte maturation and periosteal bone formation in Prx1-Herstatin limb skeletal elements, thus serving to restore normal development of the ErbB-deficient limb skeletal elements by day 18.5. Another possibility is that vascular invasion occurring after day E16.5 may accelerate the maturation process in mutant limbs such that compensation occurs by day E18.5. The restoration of normal skeletal element development in Prx1-Herstatin limbs by day 18.5 is consistent with the lack of reported skeletal defects in ErbB2 null mice genetically rescued from embryonic lethality by cardiac-specific ErbB2 expression (Morris et al. 1999; Woldeyesus et al. 1999).
This work was supported by NIH grant HD22610 (CND), NCI grant CA82503 (GMC) and NIH grant 79808 (NJM). We thank James Martin for the Prx1 promoter construct, and Joshua Sanes, Bjorn Olsen, Ken Muneoka, Alex Lichtler and Robert Kosher for providing cDNAs from which probes were made.