Faculty of Biological Sciences, University of Leeds

These pages have been left in this location as a service to the numerous websites around the world which link to this content. The original authors are no longer at the University of Leeds, and the former Centre for Human Biology became the School of Biomedical Sciences which is now part of the Faculty of Biological Sciences.


Evolutionary Developmental Biology

Lecture 15 How many toes, and would sir like them webbed?(OVERHEAD)

If the limb bud is too small from the start, as opposed to growing less than it should, digits simply do not appear in the foot. When digits are simply removed like this it is commoner to lose postaxial digits. This is unexpected, because if the ZPA organises from the region of digit V and we might guess that anterior toes might be more vulnerable. However, there are plenty of examples: (OVERHEAD) postaxial hemimelia, described by Searle, has a postaxial mesodermal deficiency and loses posterior digits.

Another set of mutations seem to have the opposite effect: they increase the size of the limb bud and the number of digits. Once again polydactyly may be pre or postaxial, but more commonly preaxial. Preaxial polydactyly is often associated with an apparently contradictory loss of more proximal preaxial elements. As an example lets look at one of the best described mutants Carter's luxate (lx, OVERHEAD). This is a semidominant gene i.e. in the heterozygotes the hind feet only are mildly affected, the homozygotes more so. In heterozygotes the limb may be normal, or the hallux may have three phalanges (a common indication of low grade polydactyly) or be joined by a pre-hallux. In the homozygote there may be up to seven toes. The polydactyly is accompanied in homozygotes by a progressive reduction of the preaxial side of the limb skeleton: the tibia is first reduced then replaced by a ligament, the fibula is thickened: in more extreme cases the femur and even the pelvis are affected. (OVERHEAD). In the most extreme individuals this is accompanied not by polydactyly but by loss of preaxial digits down to four or three.

The limb buds here are reduced in size at 11 days, with a small AER. At 12.5 days the AER is larger than normal preaxially (OVERHEAD). A little later the tibial blastema is seen to be reduced, with concomitant extra digital rays, or absent, in which case preaxial digits are reduced.

Several similar mutations are known, all showing the same features, an initially narrow bud with later preaxial excess. Tibial reduction followed by digital excess can be explained like this. The initial limb is small, probably due to delay or absence of the anterior part of the AER. If the AER does not recover from this initial setback then the tibia is reduced and there are no extra preaxial digits - in fact there may be losses. Frequently in these mutants, however, we see a recovery of the preaxial part of the AER, which is present for the normal length of time: but because it starts late it also finishes late. Mesoderm perhaps normally destined for the tibia is generated at a time when digits are being formed, and preaxial polydactyly ensues.

If extra digits are added to the row what will they be? How does the limb bud control system cater for this sort of event? In polydactylous limbs (OVERHEAD) it is quite clear that we often have examples of mirror image duplications, just like those produced by graftiing an additional anterior ZPA (OVERHEAD). If we allow that simple preaxial polydactyly is a low grade of mirror image duplication (as it certainly is in luxate and many other cases) then we know of a couple of dozen separate genes which insert an anterior ZPA. This seems rather a lot. Why should all these genes, on different chromosomes and with different effects all produce mirror image duplication?

Interestingly early limbed animals had more than five toes, six, seven or eight (OVERHEAD). The tetrapod limb is often referred to as the pentadactyl limb because it has five digits, and we assumed that five was the primitive number. But is it? We know that many species (horses, sheep, birds) have less, some (whales, dolphins) have more. Are we biased because we have five, or do most species have five? Certainly the retention of the 'primitive' five fingers is regarded as a prerequisite for tool usage.

Now the five fingered limb doesn't exist in fish, because fins are made rather differently, broad based and symmetrical: it is first seen in early land dwelling tetrapods. The earliest of these were found in Greenland inn 1929 by a Danish expedition. They are Devonian, 390-340mya. On the death of their discoverer Jarvik took over the study of the fossils in the 50s. Although no specimens of the earliest, Ichthyostega and Actanthostega had a full hand or foot Jarvik, of course, reconstructed them with a 'primitive' pentadactyl limb.

All was well until 1984 when another early fossil Tulerpeton was shown to have six digits. Then in 1990 Coates and Clack showed that Acanthostega had seven digits and Ichtyosega seven although the upper parts of the limbs coincided closely with Jarvik's material. So the primitive condition was not five: three intact feet are known, non with five toes, one six, one seven, one eight.

Both Acanthostrega and Ichthyostega have two groups of digits, a main hand and a small group of digits in the thumb region. i.e. they have mirror image polydactyly. Could it be that early tetrapods had another ZPA anteriorly as well acting variably to produce some more digits? And could it be that a single developmental step that cuts out this second ZPA stabilises the pattern at the five digits that develop from it, instead of the 6,7, or 8 that develop from the pair? This scheme is attractive because the change from 5 to 6 or 6+ is not just quantitative, it is qualitative as well. Five digits may have chosen itself as a maximum - perhaps because a signal from the ZPA would only reach that far: and perhaps the next ZPA along was for another of those seven pairs of limbs/fins we mentioned.

Yet some contrary animals do normally have six fingers: but if so the sixth is always developed as a different entity from the bunch of five, not as a continuation. In six toed frogs the extra digit arises as an extension of the normally non-branching radius or tibia. The panda's thumb is really an extension of a wrist bone: so is that of the mole.

A sense of proportion

Most of the evolutionary change that we see works on the five toed limb by changing its pattern in various ways. The most obvious of these is by obliterating or adding digits by various means, but we have seen that proportions and growth can also change for power or speed. Lets look at mutations covering these as well.

Disproportionate change

The mouse mutant brachypod has a normal axial skeleton and short legs. The interest in this particular mutation is that the internal proportions of the legs are changed (OVERHEAD). The ulna and tibia are mildly shortened, the femur and humerus rather more and the manus and pes more still. Obviously this is not a simple proximo-distal gradient.

The brachypod limb bud is of normal size until the digital contours and blastemal condensations arise at 13 days. At this point the digits are thin, and the M/P joint too proximal.

This misallocation of material may give us a clue as to how non-proportional changes in limbs might be brought about. What is brachypod doing? We are not sure, but interestingly one of the earliest effects of the gene is known to be on cell adhesion . If this is the cause, presumably by an effect on blastemal shape and size, this might be the factor which determines the relative length of the various segments of our limbs - an important difference between, say, man and gorilla or dog and badger.

Proportionate change.

This is more common: many mutants are known which affect all parts of the limb and result in short limbs attached to a relatively normal body. We have already talked about achodroplasia, as in the circus dwarf, which actually affects all cartilage replacement bones.

In nature we might expect this not to be very useful: but we should remember that evolutionary potential is exploited by man. In the eighteenth century the Ancon sheep (OVERHEAD) was developed. These have short legs and are less able to jump over fences, so became very popular. Dogs, such as basset hounds and dachshunds are selected for short legs, known to be due to a single gene. The rest of the body is unaffected in these animals.

Changes in proportion between different species was one of the problems tackled by D'Arcy Thompson (OVERHEAD) in his classic On Growth and Form. He saw that it was possible to apply a Cartesian transformation to bone outlines Cartesian transformation means basically that you draw a grid over a bone outline then deform it in a particular way to produce another bone outline. The cannon bones of giraffe, sheep and ox are thus related to each other by simple proportion: width changes but length doesn't.

The same technique will not work on the whole limb: we need to apply a different degree of distortion or stretch to each segment, proportionally more distally than proximally to mimic the change. This could, of course be achieved by varying

the conditions for outgrowth in a non linear way: if this happens we are looking at a more sophisticated change than that seen in Ancorn sheep or beagles.

Cell death

We have discussed many of the processes which shape the limb bud: there is however one more, not yet mentioned, which is important not only in the limbs but elsewhere in the body, both as a morphological shape determinant and in a wider context.

Almost as important as growth in the developing embryo is cell death or apoptosis (OVERHEAD). All cells, of course, die sooner or later. But which is it to be? Once germ cells have been separated from somatic cells the latter rapidly split into one or other of two groups. Some cells, nerve cells, muscle fibres, liver parenchyma, endothelium, fibrocytes slow down the process of division almost to the point of stopping. In most cases, but not all, these cells can be reactivated if the population is depleted by accident or surgery. Others, like epithelial cells in the gut, or blood forming cells have a short life in fully differentiated form, and so are continually replaced from a population of stem cells. Stem cells are potentially immortal: those from the haemopoitic system of mice can be labelled and grafted into other mice: this experiment was discontinued, with the cells still healthy after five times the normal mouse lifespan. Stem cells, and incidentally cancer cells show no sign of ageing or death.

The cell deaths which interest us happen during embryological and fetal development. The idea that cell death might be planned arose in the first half of the present century, because groups of dying cells or areas of debris were observed at specific sites within the embryo. In the last thirty years cell death has been identified in the implanting mouse embryo, the limb, the palate, the nervous system, the heart and the eye.

Cell death certainly varies between classes of vertebrates, but also between species: the control of programmed cell death may therefore be used as a developmental strategy, to change the shape or extent of a morphological feature.

Glucksmann (OVERHEAD) considered cell death to be of three kinds.

morphogenetic: involved in the shaping of organs e.g. neural tube closure, palatal fusion, limb shaping.

hisogenetic: associated with the differentiation of tissues e.g. the degeneration of Wolffian and Mullerian ducts in the development of the reproductive system

phylogenetic: regression of structures in higher vertebrates which have a definite function in lower vertebrates, such as the ductus arteriosus, pronephros and mesonephros.

These categories clearly overlap, and talk of higher is hardly PC. Truman gave a rather more sensible assessment, saying that cell death was necessary to match the size of a population of cells (such as nerves) to the tissues with which they must interact, that it may be necessary as a modifier of phenotype and that it may act as a waste disposal system for excess cells.

Saunders introduced other questions. Is cell death assassination or suicide? Since some areas of cell death are highly predictable is it essential for those processes.

In any case we can distinguish this cell death from necrosis, which happens when cells are exposed to non physiological conditions: here nuclei swell and appear to burst. In apoptosis the cells shrink or condense and may separate from their neighbours before being phagocytosed or fragmenting. Also necrosis often leads to an area of tissue damage: apoptosis never does.

The obvious place to study cell death is nematodes. Remember that we know all about cell lineages in these animals and that 111 of the cells in the male die. In fact sexual dimorphism is brought about by cell death. Nematodes are a bit kinky in that there are two sexes male and hermaphrodite: at about 470 hours of development in males hermaphrodite specific neurones die, and in hermaphrodites male specific neurones do. Two mutations are known which affect cell death: in one phagocytosis is blocked, but cells die on time anyway. In the other cell death is inhibited and the escaped cells go on to differentiate into recognisable cell types.

In leeches development includes a process of formation of parallel columns of cells, 32 of which will make up a leech segment. Cell death occurs in columns 33 and above, at the rear of the embryo. If you swap the cells around, putting anterior ones to the back columns 33 et seq still die - so its not mitotic order or age which determines their fate.

The best examples of cell death in vertebrates come, perhaps predictably, from our test system, the limb. Limb bud shaping by differential cell death occurs in both ectoderm and mesoderm of developing limb buds in most amniotes, although it seems to be absent in amphibia. Mesodermal death is more extensive and differs between fore and hindlimb, between species and between normal and mutant individuals.

Four loci of cell death have been regularly seen:(OVERHEAD) the anterior and posterior necrotic zones (ANZ,PNZ) which shape the limb, the central opaque patch and later the interdigital necrotic zones (INZ). all these are present in the chick and are easily mapped by vital staining with Nile blue or neutral red, which are picked up by macrophages which are inevitably present in areas with significant cell death.

In the chick limb the ANZ is seen as a proximal wave of death followed by a distal one, whilst the posterior necrotic zone seems to be a single entity. The opaque patch perhaps has something to do with separation of the radius and ulna.

The extent and presence of these zones of cell death varies between species of birds and is characteristically absent in rats and mice, which have more digits, and where there is only a small necrotic area adjacent to digit 1.. The ANZ and PNZ may therefore limit the amount of distal mesenchyme which is available for digit formation in these species with a reduced set. However Saunders found that if the ANZ is suppressed experimentally in the chick no extra digits are formed. Against this the small necrotic area found in mouse and rat is absent in mole, which has a rather extraordinary prehallux or sixth digit anteriorly. In mutants we find that talpid, which has up to eight digits has no necrotic zones, whilst wingless has a very large ANZ.

The opaque patch is situated at a point which corresponds to the proximal end of the radius and ulna: we already know that chondrogenesis depends on increased cell adhesion: dead cells don't stick together - they round up. Might this be something to do with the switch from one skeletal element to two?

A little later in development further areas of cell death (OVERHEAD) appear in the footplate in what will be the interdigital areas. These follow a well defined timetable. They have been found in all amniotes studied including man, rat, mouse, birds turtles. The INZs are suppressed in web footed birds, and in a number of mutants and teratological conditions where feet are webbed.

The idea of cell death as a morphological mechanism is unusual: we usually think of specific production of a protein or a peptide as an indicator of differentiation, not death. So how does a cell prepare for its own death? Scattered cells seem to be destined for death, and become vacuolated, condense their chromatin and round up. Adjacent cells phagocytose these victims and in fact convert to macrophages, containing the remains of up to 10-12 of their neighbours.

The dying cells contain an active Golgi apparatus which produces additional lysosomal enzymes. So perhaps our ideas of differentiation being the result of novel or amplified gene products are not far off the mark. Since only some cells in a field die it seems likely that they are pre-programmed in some way. Do they form a pattern? Do they have positional information? Have they undergone a fixed number of mitotic cell divisions and then died as some cells do in culture?


This page is maintained by Steve Paxton