Evolution and Morphogenetic Rules in Cartilage Formation in the Vertebrate Limb

In Section 6.6 in the last chapter we showed how a mechanical model could generate the cartilage patterns in the vertebrate limb. There we proposed a simple set of general morphogenetic construction rules for how the major features of limb cartilage patterns are established. Here we use these results and draw on comparative studies of limb morphology and experimental embryological studies of the developing limb to support our general theory (which is essentially mechanism-independent) of limb morphogenesis. We then put the results in an evolutionary context. The following is mainly based on the work of Oster et al. (1988) which arose from discussions between George Oster, the late Pere Alberch and myself in 1985.

Since the limb is one of the most morphologically diversified of the vertebrate organs and one of the more easily studied developmental systems it is not surprising it is so important in both embryology and evolutionary biology. Coupled with this is a rich fossil record documenting the evolution of limb diversification (see, for example, Hinchliffe and Johnson 1980 for a comprehensive biology).

Although morphogenesis appears deterministic on a macroscopic scale, on a microscopic scale cellular activities during the formation of the limb involve considerable randomness. Order emerges as an average outcome with some high probability. We argued in Section 6.6 that some morphogenetic events are extremely unlikely, such as trifurcations from a single chondrogenic condensation. Mathematically, of course, they are not strictly forbidden by the pattern formation process, be it mechanochemical or reaction diffusion, but are highly unlikely since they correspond to a delicate choice of conditions and parameter tuning. This is an example of a 'developmental constraint' although the term 'developmental bias' would be more appropriate.

Let us recall the key results in Section 6.6 regarding the 'morphogenetic rules' for limb cartilage patterning. These are summarised in Figures 6.18(a)-(c) the key parts of which we reproduce for convenience in Figure 7.5.

The morphogenetic process starts with a uniform field of mesenchymal cells from which a precartilagenous focal condensation of mesenchymal cells forms in the proximal region of the limb bud. With the mechanical model discussed in the last chapter, this is the outcome of a model involving the cells, the extracellular matrix (ECM) and its displacement. With the model of Oster et al. (1985a,b), various mechanochemical processes are also involved. Subsequent differentiation of the mesenchymal cells is intimately tied to the process of condensation. It seems that differentiation and cartilage

Figure 7.5. Morphogenetic rules: the three basic cell condensation types, namely, a single or focal condensation, F, as in (a), a branching bifurcation, B, as in (b) and a segmental condensation, S, as in (c). More complicated patterns can be built up from a combination of these basic bifurcations; see Figures 7.8, 7.11 (cf. Figure 6.18(e)) and 7.12.

Figure 7.5. Morphogenetic rules: the three basic cell condensation types, namely, a single or focal condensation, F, as in (a), a branching bifurcation, B, as in (b) and a segmental condensation, S, as in (c). More complicated patterns can be built up from a combination of these basic bifurcations; see Figures 7.8, 7.11 (cf. Figure 6.18(e)) and 7.12.

morphogenesis are frequently interrelated phenomena. An alternative cell-chemotaxis model with cell differentiation whereby condensation and morphogenesis take place simultaneously was proposed by Oster and Murray (1989).

There is a zone of recruitment created around the chondrogenic focus. That is, an aggregation of cells autocatalytically enhances itself while depleting cells in the surrounding tissue. This is effectively setting up a lateral inhibitory field against further aggregation. Because nearby foci compete for cells this leads to almost cell-free regions between foci. In other words, a condensation focus establishes a 'zone of influence' within which other foci are inhibited from forming.

As the actual cartilagenous element develops, the cells seem to separate into two regions: the outer region consists of flattened cells concentrically arranged, while the cells in the inner region are rounded. The outer cells differentiate to form the perichondrium which sheaths the developing bone. As suggested by Archer et al. (1983) and Oster et al. (1985a,b), the perichondrium constrains the lateral growth of cartilage and forces its elongation. It also restricts the lateral recruitment of additional cells, so that cells are added to this initial condensation primarily by adding more mesenchymal cells at the distal end thus affecting linear growth as illustrated in Figure 7.6, which also shows the general features of the condensation process.

As we noted in Section 6.6, limb morphogenetic patterns are usually laid down sequentially, and not simultaneously over an entire tissue (Hinchliffe and Johnson 1980). The latter method would be rather unstable. Theoretical models show that sequential pattern generation is much more stable and reproducible. Recall the simulations associated with the formation of animal coat patterns in Chapter 3, Section 3.1, where the final pattern was dependent on the initial conditions, as compared with the robust formation of hexagonal feather germ and scale arrays in birds, discussed in Section 6.5 in the last chapter, and the supporting evidence from the model simulations by Perelson et al. (1986).

Although most of the pattern formation sequence proceeds in a proximo-distal direction, the differentiation of the digital arch (see Figure 7.11) occurs sequentially from anterior to posterior. The onset of the differentiation of the digital arch is correlated with the sudden broadening and flattening of the distal region of the limb bud into a paddlelike shape. From the typical dispersion relations for pattern generation mechanisms

Formation The Vertebrate Limb

Figure 7.6. Schematic illustration of the cell condensation process. (a) Cells aggregate initially into a central focus. (b) Development of the cartilaginous element restricts cell recruitment to the distal end of the condensation. (c) When conditions are appropriate the aggregation undergoes a F-bifurcation. (After Shubin and Alberch 1986)

Figure 7.6. Schematic illustration of the cell condensation process. (a) Cells aggregate initially into a central focus. (b) Development of the cartilaginous element restricts cell recruitment to the distal end of the condensation. (c) When conditions are appropriate the aggregation undergoes a F-bifurcation. (After Shubin and Alberch 1986)

(recall, for example, the detailed discussion in Chapter 2, Section 2.5) such a change in geometry can initiate independent patterns and is the key to understanding this apparent exception to the sequential development rule. Physically, this means that where the domain is large enough, an independent aggregation arises and is far enough away from the other aggregations that it can recruit cells to itself without being dominated by the attractant powers of its larger neighbours. Of course other model parameters are also important elements in the ultimate pattern and its sequential generation and initiation. The key point is that, irrespective of whether reaction diffusion or mechanochemical models create the chondrogenic condensations, the model parameters, which include the size and shape of the growing limb bud, are crucially important in controlling pattern. Experimental manipulations clearly confirm this importance.

Alberch and Gale (1983) treated a variety of limb buds with the mitotic inhibitor colchicine. This chemical reduces the dimensions of the limb by reducing cell proliferation. As we predicted, from our knowledge of pattern generation models and their dispersion relations, such a reduction in tissue size reduces the number of bifurcation events, as illustrated in Figure 7.7.

Note that a possibility that cannot be ruled out is that colchicine affects the timing and number of bifurcations by altering some other developmental parameter, such as cell traction or motility, in addition to the size of the recruitment domains. This alteration, of course, is still consistent with the theory. At this stage further experiments are required to differentiate between the various possibilities. The main point is that these experiments confirm the principle that alterations in developmental parameters (here tissue size) can change the normal sequence of bifurcation events, with concomitant changes in limb morphology that are significant.

Figure 7.7. Experimentally induced alterations in the foot of the salamander Ambystoma mexicanum and the frog Xenopus laevis through treatment of the limb bud with colchicine. (a) Normal right foot of the salamander and (b) the treated left foot. (c) Normal right foot of the frog with (d) the treated left foot. (From Alberch and Gale 1983; photographs courtesy of Dr. Pere Alberch)

Figure 7.7. Experimentally induced alterations in the foot of the salamander Ambystoma mexicanum and the frog Xenopus laevis through treatment of the limb bud with colchicine. (a) Normal right foot of the salamander and (b) the treated left foot. (c) Normal right foot of the frog with (d) the treated left foot. (From Alberch and Gale 1983; photographs courtesy of Dr. Pere Alberch)

Using the basic ideas of cartilage pattern formation in Oster et al. (1983), Shubin and Alberch (1986) carried out a series of comparative studies with amphibians, reptiles, birds and mammals, and confirmed the hypothesis that tetrapod limb development consists of iterations of the processes of focal condensation, segmentation and branching. Furthermore, they showed that the patterns of precartilage cell condensation display several striking regularities in the formation of the limb pattern. Figure 7.8 presents just some of these results; other examples are also given in Oster et al. (1988).

Was this article helpful?

0 0

Post a comment