Aulthouse AM, Solursh M 1987 The detection of a precartilage, blastema-specific marker. Dev Biol 120:377-384
Brunet LJ, McMahon JA, McMahon AP, Harland RM 1998 Noggin, cartilage morphogenesis, and joint formation in the mammalian skeleton. Science 280:1455-1457
Capdevila J, Johnson RL 1998 Endogenous and ectopic expression of noggin suggests a conserved mechanism for regulation of BMP function during limb and somite patterning. Dev Biol 197:205-217
Chang C, Hemmati-Brivanlou A 1999 Xenopus GDF6, a new antagonist of noggin and a partner of BMPs. Development 126:3347-3357
Chen WS, Manova K, Weinstein DD et al 1994 Disruption of the HNF-4 gene, expressed in visceral endoderm, leads to cell death in embryonic ectoderm and impaired gastrulation of mouse embryos. Genes Dev 8:2466-2477
Dunn NR, Winnier GE, Hargett LK, Schrick J J, Fogo AB, Hogan BLM 1997 Haploinsufficient phenotypes in Bmp4 heterozygous null mice and modification by mutations in GH3 and A-lx4. Dev Biol 188:235-247
Erlacher L, McCartney J, Piek E et al 1998 Cartilage-derived morphogenetic proteins and osteogenic protein-1 differentially regulate osteogenesis. J Bone Miner Res 13:383-392 Fell HB, Canti RG 1934 Experiments on the development in vitro of the avian knee-joint. Proc R Soc 116:316-351
Francis-West PH, Abdelfattah A, Chen P et al 1999 Mechanisms of GDF-5 action during skeletal development. Development 126:1305-1315 Gañan Y, Macias D, Duterque-Coquillaud M, Ros MA, Hurle JM 1996 Role of TGFßs and BMPs as signals controlling the position of the digits and the areas of interdigital cell death in the developing chick limb autopod. Development 122:2349-2357 Grüneberg H, Lee AJ 1973 The anatomy and development of brachypodism in the mouse. J
Embryol Exp Morphol 30:119-141 Hall BK, Miyake T 1995 Divide, accumulate, differentiate: cell condensation in skeletal development revisited. Int J Dev Biol 39:881-893 Hamburger V, Hamilton HL 1992 A series of normal stages in the development of the chick embryo. Dev Dyn 195:231-272 Hogan BLM 1996 Bone morphogenetic proteins: multifunctional regulators of vertebrate development. Genes Dev 10:1580-1594 Johnson RL, Tabin CJ 1997 Molecular models for vertebrate limb development. Cell 90:979990
Kaplan FS, Delatycki M, Gannon FH, Rogers JG, Smith R, Shore EM 1998 Fybrodysplasia ossificans progressiva. In: Emery AEH (ed) Neuromuscular disorders: clinical and molecular genetics. Wiley, Chichester, p 289-321 Katagiri T, Boorla S, Frendo JL, Hogan BL, Karsenty G 1998 Skeletal abnormalities in doubly heterozygous Bmp4 andBmp7 mice. Dev Genet 22:340-348 Kent J, Wheatley SC, Andrews JE, Sinclair AH, Koopman P 1996 A male-specific role for
SOX9 in vertebrate sex determination. Development 122:2813-2822 Macias D, Gañan Y, Rodriguez-Leon J, Merino R, Hurle JM 1999 Regulation by members of the transforming growth factor beta superfamily of the digital and interdigital fates of the autopodial limb mesoderm. Cell Tissue Res 296:95-102 Merino R, Gañan Y, Macias D, Economides AN, Sampath KT, Hurle JM 1998 Morphogenesis of digits in the avian limb is controlled by FGFs, TGFßs, and noggin through BMP signaling. Dev Biol 200:35-45
Merino R, Macias D, Gañan Y et al 1999a Expression and function of Gdf-5 during digit skeletogenesis in the embryonic chick leg bud. Dev Biol 206:33-45 Merino R, Macias D, Gañan Y et al 1999b Control of digit formation by activin signalling.
Development 126:2161-2170 Nah H-D, Rodgers BJ, Kulyk WM, Kream BE, Kosher RA, Upholt WB 1988 In situ hybridization analysis of the expression of the type II collagen gene in the developing chicken limb bud. Coll Relat Res 8:277-294 Ng L-J, Wheatley S, Muscat GEO et al 1997 SOX9 binds DNA, activates transcription, and coexpresses with type II collagen during chondrogenesis in the mouse. Dev Biol 183:108-121 Pizette S, Niswander L 1999 BMPs negatively regulate structure and function of the limb apical ectodermal ridge. Development 126:883-894 Pizette S, Niswander L 2000 BMPs are required at two steps of limb chondrogenesis: formation of prechondrogenic condensations and their differentiation into chondrocytes. Dev Biol
Roark EF, Greer K 1994 Transforming growth factor-^ and bone morphogenetic protein-2 act by distinct mechanisms to promote chick limb cartilage differentiation in vitro. Dev Dyn 200:103—116
Shafritz AB, Shore EM, Gannon FH et al 1996 Overexpression of an osteogenic morphogen in fibrodysplasia ossificans progressiva. N Engl J Med 335:555—561 Storm EE, Kingsley DM 1996 Joint patterning defects caused by single and double mutations in members of the bone morphogenetic protein (BMP) family. Development 122:3969—3979 Storm EE, Kingsley DM 1999 GDF5 coordinates bone and joint formation during digit development. Dev Biol 209:11—27 Storm EE, Huynh TV, Copeland NG, Jenkins NA, Kingsley DM, Lee SJ 1994 Limb alterations in brachypodism mice due to mutations in a new member of the TGFjS-superfamily. Nature 368:639—643
Stringa E, Tuan RS 1996 Chondrogenic cell subpopulation of chick embryonic calvarium: isolation by peanut agglutinin affinity chromatography and in vitro characterization. Anat Embryol (Berl) 194:427—437 Zhang H, Bradley A 1996 Mice deficient for BMP2 are nonviable and have defects in amnion/
chorion and cardiac development. Development 122:2977—2986 Zimmerman LB, De Jesus-Escobar JM, Harland RM 1996 The Spemann organizer signal noggin binds and inactivates bone morphogenetic protein 4. Cell 86:599—606 Zou H, Wieser R, Massague J, Niswander L 1997 Distinct roles of type I bone morphogenetic protein receptors in the formation and differentiation of cartilage. Genes Dev 11:2191—2203
Hall: How is joint specification related to condensation? Do you think that we are getting a little closer to answering this?
Pipette: Not really. Some results that I did not present here suggest that the prechondrogenic condensations are necessary for the joints to form. This is based on the fact that in the autopod of embryos infected with the Noggin virus there are no condensations, no expression of the joint marker Gdf5, nor any morphological signs of joint formation or initiation. In addition, the expression of Gdf5 in a normal context is somewhat puzzling. Part of its expression pattern is clearly in transverse stripes in the prechondrogenic condensations, delineating the future joint regions. However in the zeugopod, Gdf5 expression is a little different. Although it is also expressed in a transverse stripe between the stylopodial and the zeugopodial condensations, there is this additional longitudinal expression between the radius and ulna and tibia and fibula condensations, which has not been described before. What was described was that Gdf5 is expressed throughout the whole condensation that will later give rise to radius and ulna and tibia and fibula. It is thus unclear to me what the relationship between the condensation and the joint forming region as marked by Gdf5 expression is in the zeugopod.
Hall: The radius and ulna begin as a single condensation, which then has to separate. It looked as if what you had with your Gdf5 expression was that single condensation in the middle that hadn't yet separated.
Pipette: That is possible. There is some interesting older work on joint formation in the knee of the chick (Fell & Canti 1934). In this paper they studied the region where the knee joint was coming from, looking at different stages. Early on, the region that will give rise to the knee joint maps the whole condensation, and then it reduces in size and approximately maps to what is called the opaque patch. There is a striking correlation between the expression of Gdf5 at these stages, the joint forming region and the opaque patch. Somehow, I feel that there may be some apoptosis going on. This might be the mechanism by which the initial condensation is split. To me it is not clear whether or not there is real branching to form the radius and ulna, or if in the beginning there is an initial mass and the expression of Gdf5 is restricted and maps the region of cell death, which would then help to split this condensation.
Blair: Is Gdf5 expression maintained later on after the beginning of epiphyseal bone formation?
Pipette: It is still expressed in the cavitating joint.
Kingsley: Most of the in situ hybridizations showed the expression dropping down at later stages. We have been doing some experiments using an approach that attempts to bridge early patterning events to these more local events that are defining skeletal elements by doing promoter analysis. If we can figure out why Gdf5 is turning-on in the stripes we might be able to link the stripy joint formation to earlier patterning events. One of the things we found in making various constructs that have Gdf5 regulatory elements hooked up to the lacZ reporter gene, is that we can detect expression along the surface of joints at much later stages than we were ever able to by in situ hybridization, including articular surfaces of joints way after birth and into adulthood.
Mundlos: Did you say that type II collagen is up-regulated in the overexpression experiments?
Mundlos: Then I have a problem: my understanding was that BMP is upstream of collagen type II and can initiate chondrogenesis. Inhibiting BMPs should down-regulate collagen type II expression. This doesn't make sense.
Pipette: It doesn't make sense at all, especially because I looked at the expression of Sox9, which has been shown to regulate collagen II expression, and it is down-regulated. This is pretty much in agreement with BMPs being upstream ofSox9 and collagen type II. If you compare expression of the BMP receptor 1B and collagen II, BMPR1B is slightly ahead in its expression in prechondrogenic condensations. I was expecting to see no or very little collagen II in the zeugopod region, because collagen II is expressed in prechondrogenic condensations. We were very surprised when we saw this huge up-regulation and ectopic expression. The next thing I want to do is to find out whether these regions that have ectopic collagen II expression are condensed or not. My initial feeling is that they are not.
Mundlos: If they express type II collagen, you should be able to see on a single histological section whether this is condensed.
Pipette: I haven't seen condensations on the histological sections, except for the region where Gdf5 is ectopically expressed. I'm now doing PNA (peanut agglutinin) staining.
Karsenty: Type II collagen is expressed before the condensation. There is no condensation at Day 7.5, when we start to see type II collagen.
Poole: Are we talking about collagen type IIa or IIb? This is an important distinction.
Newman: A peculiarity of collagen type II relates to its regulation at the post-transcriptional level. It seems to be almost constitutively expressed in early mesenchyme, and then it comes up a little bit more in the condensing regions. Unlike the proteoglycan core protein, for example, which seems to be stringently regulated, with no sign of the mRNA until differentiation is about to happen, type II collagen is expressed much more broadly and much earlier (Kosher et al 1986a,b).
Hall: Including in places such as basal membranes where you normally would not expect to find type II collagen.
Sandrine Pizette, have you had any of your embryos go much longer to see whether those cells that express collagen type II do anything?
Pipette: They don't form chondrocytes.
'Hall: But do they do anything else?
Pipette: I have no idea. On E5, in the limbs there is a huge domain of Gdf5 expression, but there is not much muscle because of a defect in muscle formation. There is some tendon, but this is all I can say. I let the embryos carry on to E6 or E7, but they usually die by this stage. Have you any ideas about what these cells might do?
Hall: There are a number of possibilities. A late phase of cartilage might appear; another possibility is that collagen II tends to be expressed quite ubiquitously, and it may later be turned off: those cells might then become soft tissue, or die and disappear.
Pipette: I haven't seen cell death at E4 or E5. 24 h after this up-regulation we looked to see whether the levels of collagen type II were still high. It is still expressed in the zeugopod in a strange pattern that has nothing to do with where we would expect to see condensation or even chondrocytes. Compared with wildtype embryos, the expression level is lower than what one would expect in chondrocytes at the same stage.
Kronenberg: You mentioned that motivation behind the Noggin experiments was to sort out how much of the action of BMP is on the differentiation of cells, how much is on proliferation and how much is on cell death. You showed us data that suggested that there is a bit of cell death here and there that might turn out be relevant. You also showed us a bromodeoxyuridine experiment that indicated that proliferation may have a role. But it looked to me as if most of what you were showing were things that couldn't easily be explained simply by changing the cell numbers. What was your conclusion?
Pipette: Because there is no cell death it looks like BMPs are not required as a survival signal. However, because there is this defect in proliferation, BMPs may well have a role in promoting proliferation ofcells. As I have shown you, once the condensation is established, the mitotic index decreases. If one wants to argue that BMPs are involved in stimulating proliferation, they would have to postulate that this is prior to condensation. It looks as ifBMPs and GDF have a role in promoting chondrocyte differentiation. In the zeugopod BMPs may have a different role. They may be needed to provide a survival signal for cells once they have condensed. The other possibility is that they may be involved in repressing the joint fate, maintaining a limited number of cells which are going to acquire a joint fate.
Reddi: Have you looked at endogenous expression of Noggin'? Does it play a role? It seems to me that it is very important to look at this.
Pipette: I forgot to mention the results of the Noggin knockout, which was done by Richard Harland (Brunet etal 1998). It results in the overgrowth of the cartilage elements. However, he has looked at the stage of the prechondrogenic condensed mesenchyme, and doesn't see any differences in the size of the condensations. This indicates that there is no effect at that stage. He also looked at proliferation of the growth plates, and didn't see any effect on proliferation either. The conclusion was that the BMPs may act at later stages through appositional growth and the recruitment of cells to the perichondria. In the embryo, Noggin is expressed in the prechondrogenic condensed mesenchyme, and also in immature and prehypertrophic chondrocytes. The striking thing is that if you compare the Noggin expression pattern with that of BMP receptor 1b in the prechondrogenic condensations of the digits, Noggin is expressed just in the centre of this domain of BMP receptor 1b expression, where it could potentially block all cells in the condensation from receiving BMP signalling. The knockout experiments suggested that Noggin was limiting the effect of BMPs. Here, if BMPs and GDFs are involved in forming the condensations, maybe Noggin would need to be expressed not in the middle of the condensation but at the edges to prevent more cells from being recruited. So the expression pattern we see is really puzzling.
Newman: This is one place where the dynamical models may be helpful, because in those models the source of the activator and the source of the inhibitor are in exactly the same place. Then it's just the relative levels of the proteins and their relative diffusion coefficients that determine what the spacing will be. This is the kind of pattern that you would expect. Now you just have to see the protein profiles to verify it.
Mundlos: Another possible explanation is that if you have expression of the inhibitor in the middle, it indicates that this expression is in the more mature cells. In contrast, those cells on the outside of the condensation are still recruiting and proliferating new chrondrocytes. The cells at the centre may have reached a certain level of differentiation and now this whole system is being switched off.
Pipette: I agree that this would make sense and you are right in that the cells in the centre of the condensations are the first cells that are going to start to form the chondrocytes. However this explanation conflicts with my results, since these results show that BMPs and GDFs are needed by the cells of the pre-chondrogenic condensed mesenchyme to make chondrocytes, those same cells which are in the centre of the condensation.
Mundlos: You misexpressed it where it is not usually expressed, meaning that you interfered at a position where it's not usually there, and then it does make sense.
Hall: Is the expression pattern of BMP the same as the pattern of expression of the receptor?
Pipette: Not at all: the BMPs are not expressed in the condensations, but instead in between them in the interdigit regions.
Burger: We should bear in mind that the skeleton is not just cartilage: it is also all the structures that link these cartilage elements. The continuity of the prechondrogenic condensation could also mean that the whole structure is laid down, possibly also including the muscles, tendons and ligaments, and it is only later that the foci of cartilage develop.
Hall: There are multiple roles of BMP in different tissues.
Pipette: There is a big effect in muscle and tendon formation.
Chen: Is the phenotype from your Noggin overexpression chicks the same as the dominant-negative BMP receptor chicks?
Pipette: Some features are comparable, but in the dominant-negative BMP receptor infected chicks it is mainly the distal tips of the digits that are missing. The phenotypes are pretty close; there is just a difference in efficiency.
Morriss-Kay: I would like to broaden this discussion to include another family of players — the retinoid receptors. I was struck by the similarity between the gene expression patterns illustrated by Sandrine and those of retinoic acid receptor (RAR)y, which is expressed in pre-chondrogenic condensations and the cartilaginous bone models, and RAR^, which is expressed in the interdigital regions. The complementary expression patterns shown by RARy and RAR^ is similar to the expression patterns that we have just seen for GDF5 and BMPs in the interdigital regions, and Noggin and the BMP receptor IB in the condensations. Is anything is known about the ways in which the RARs interact with the molecular elements that we've just heard about?
Underhill: It looks like RARy will support the formation of cartilage, whereas RARjS seems to inhibit this process. Some of these signalling pathways appear to be functioning downstream of BMP signalling in regulating chondroblast differentiation (Weston et al 2000).
Pipette: There is a paper showing that retinoic acid signalling is upstream of BMP in the interdigits of the chick embryo and that it is required for apoptosis (Rodriguez-Leon et al 1999).
Hall: One of the studies Juan Hurle et al (1989) did was to show that in the interdigital region of the chick limb, if you take off this piece of epithelium, you can induce a little ectopic area of cartilage between the digits. Epithelium is inhibiting chondrogenesis. I think they then did an experiment where the epithelium was left intact, a bead of BMP was implanted and they obtained ectopic chondrogenesis when the epithelium was still present (Ganan et al 1998, Merino et al 1998). This makes me wonder about the role of the epithelium in the BMP story, as a sort of global inhibitor of chondrogenesis around the developing limb bud.
Pipette: That is one of the main problems in the chick. The model that is used involves the interdigit region, and people tend to put beads soaked in different factors in the interdigital region to see whether this promotes ectopic cartilage formation. I don't know whether this is such a good model. As you said, simple removal of the epithelium will induce ectopic cartilage formation. If you just put in BMP beads without removing anything, this will not induce ectopic cartilage formation but more cell death. The system seems to be already a little biased in the sense that there are things in the interdigit region that seem to be able to respond to BMPs only as transducing a cell death signal.
My other point has to do with the still unpublished work by Richard Maas on the Msx1/Msx2 double knockout. BMPs are expressed in the ectoderm and in the mesenchyme. Expression in the ectoderm may regulate expression in the mesenchyme. In this double knockout it seems that Msx expression is lost in both compartments, as is BMP expression. Cell death is also lost. If you add back some BMP protein, you can still get apoptosis. The conclusion was that Msx expression is not needed for cell death, but just to maintain BMP expression. This goes back to the ectodermal control in the interdigital region where there is a signal in the ectoderm that tells the mesenchyme to die. When you put various factors into the interdigit region you might be competing in some way with this signalling.
Morriss-Kaj: Juan Hurle would give credit to Michael Solursh for the interpretation of his first experiments, which is based on the fact that the surface ectoderm is secreting hyaluronan, which inhibits chondrogenesis (Solhursh 1984).
Hall: He also made the important point that these cells in the interdigital region are not naive prechondrogenic mesenchyme; they are a separate population of cells. Whenever we are looking at these systems, we need to know something about cells that we're actually looking at, as they may respond quite differently to the same factors.
Tickle: What is the evidence that joint cells actually come from the pre-chondrogenic condensations? What is their origin?
Pi^ette: There is an old paper by Honor Fell on specification of the avian knee joint in vitro. The conclusion was that joint precursors are at least at some point present in the prechondrogenic condensations.
Kingslej: That is one of the classical studies. They essentially tried to culture the region where the joint was thought to form. They then asked whether it would still appear if they took the cells out, or transplanted the presumptive joint region into a new area to see whether a joint formed. The low resolution comes from the fact that the number of markers for identifying a presumptive joint region was small. All the transfers involved surgically removed cells, and it's not at the level of cell by cell resolution. There is always the question as to whether the source of the cells within the condensation is itselfheterogeneous. There is no way to resolve that in this sort of experiment. My own feeling is that the best hope for that comes from a system where there are fewer cells to keep track of. For example, in the zebrafish joint there are two orders of magnitude fewer cells involved than in the mouse and chick joints. Chuck Kimmel's lab is interested in doing a much more sophisticated cell-by-cell lineage tracing of where the cells are coming from.
Tickle: There is a similar problem with the muscle mass splitting: how the initial, muscle mass is split into discrete regions that then go and form individual muscles. It is not clear here whether there is cell death or cells coming in from the outside.
Was this article helpful?