Faculty of Biological Sciences, University of Leeds

Last time I suggested that if we wanted to explain the variation that we saw in limb proportions and the reduction in digits that we see in things like horses and cows we must investigate changes in limb embryology. The same thing, of course, goes for the more drastic changes that we see outside the mammals (OVERHEAD).

This time I want to look at the sorts of changes in limb morphology that we see in mutations, because it is likely that any system, like the mouse or chick limb where we can make a close experimental study will carry within it at least the seeds of an evolutionary mechanism.

But aren't mutations oddities, mistakes, monsters - nothing to do with normal development? In a sense yes they are, but one of the commonest sorts of mutation that is described, in all sorts of animals is the so called atavism - a mutation that takes an animal back to a supposed ancestral state. A cat with seven toes (OVERHEAD) it is a monster, because we don't have evidence of that sort of cat from fossils. But when we see a guinea pig with 5 toes instead of the normal three that is an atavism, a throwback, a possible ancestor which does resemble fossil forms.

But both these animals have originated by exactly the same mechanism, so atavisms and mutations are similar in being possible variations on the theme.

When we left the developing limb it had (up to) five digits preformed as mesenchymal blastemata. Remember we dealt with, some time ago, the way in which mesenchyme changed into cartilage. After the digits are formed in cartilage they are recast again as bone. And at a fairly late stage they are separated into distinct entities (unless we are looking at something like a duck) by a process we haven't talked about much, controlled cell death.

At any stage from the outgrowth of the limb bud to the separation of the fingers modifications can be brought about by changing the conditions slightly. In practice most of the mutants we know of either change the size of the limb bud, or change the pattern, or mess up growth by systematically affecting cartilage or bone growth.

Modifying the size of the limb bud - downwards

Any change in the size of the limb bud is likely to affect distal rather than proximal structures, and anterior rather than posterior ones. Remember the ZPA (OVERHEAD) sits posteriorly in the limb bud and directs operations. Lets draw the analogy of God as pastrycook. If there is a shortage of dough then distal and preaxial structures are likely to suffer because they have yet to be formed when we run out. The reverse is also true: if there is any dough left over we can have a few extra fingers, probably at the preaxial margin.

This process is illustrated by removing the AER at various stages during limb development experimentally at various stages (OVERHEAD). The resulting limb simply looks as if it has been amputated at various points along its length according to the time of the operation. Early reduction in the size of the limb bud is often due to a failure in the interaction between ectodermal ridge and underlying mesoderm, as we see (OVERHEAD) in the wingless chick.

Decreased limb bud size at a slightly later stage, probably due to reduced growth, may have two consequences. Either digits are normal in number, but too closely squashed together , in which case we produce syndactyly. Or we have the right spacing, and therefore not enough digits formed - oligodactyly. Both these conditions are demonstrated by a series of mouse mutants investigated by Grneberg in the 50s and 60s.

Shaker with syndactylism is so named because it has unusual behaviour and unusual feet. Digits 2,3 and 4 of the hindlimb are fused at the phalangeal (though not at the metatarsal) level (OVERHEAD). During development the footplate of the limb bud is narrow distally and its periphery is reduced: within the reduced mesenchymal area the condensations of the affected phalanges are fused from their initial appearance.

Oligosyndactylism (OVERHEAD) has, at first glance, similar effects. Here the fusion of the digits resembles that seen in shaker with syndactylism, but in more badly affected individuals there may also be loss of digit 2, including its metatarsal. The mesodermal deficiency here is recognisably preaxial. The phalangeal blastemata of digits 2 and 3 are too close together, and later fuse. Those of digit 2 is often small and missing.

So here is a rather more complex effect, bringing to our notice two more ways of changing pattern: as well as appearing as a single blastema (as in shaker) blastemata may appear separately and fuse, or appear as a small entity and then disappear. Experiments with chick suggest that about 40 blastemal cells is the minimal size for survival.

Since limb bones pass through both a mesenchymal , a cartilaginous and a bony stage an element which is to be suppressed or fused can theoretically be lost at one of many stages.

This process has a counterpart in normal development. In many fossil reptiles (but not in modern ones) the proximal row of wrist cartilages was made up of three elements, radiale, intermedium and ulnare. In modern lizards like Calotes (OVERHEAD) the adult form has no intermedium, but its mesenchymal blastema is present in the embryo.. In the hind limb the analogous centrale also appears in mesenchyme and becomes part of a single fused element.

In the chick wing however the only mesenchymal carpal blastemata present are those which chondrify - but one of these later becomes pycnotic and does not ossify.

So in two examples we see three ways of shedding blastemata:

1. absence of mesodermal condensation

2. mesodermal condensation present but regresses

3. cartilaginous element regresses.

Presumably the total absence of an element represents an earlier loss than a regression. Reduction in the size of a blastema may produce delay in chondrification, delay in histogenesis of cartilage, delay in ossification. Reduction of digits by fusion may take place at different stages. There are many ways to kill a cat.

In animals where there are marked differences in eventual size between adjacent elements (like tibia and fibula) the blastemata from which they form are often indistinguishable. This being so what governs the differential growth?

Ewart (1894) was probably the first to be interested in this problem, especially in relation to the horse, where effectively only one digit survives. In horse embryos he found that three metacarpal cartilages were present, of which only the central one survives as bone. All were approximately the same length, but the central one was almost twice the diameter of the others. In later embryos the growth rate of the central metacarpal was much faster. Ewart also identified the transient phalanges of the second and fourth digits. The second digit initially had three phalanges separated by poorly developed joints: these later fused to together and to the distal end of the metacarpal.

The chick tibia and fibula show a similar effect., and are rather easier to study. At 5 days of development the blastemata of tibia and fibula are the same length, although the tibia is always a little stouter. But growth of the tibia is much faster, and 3 days later it twice as long as the fibula.

Wolff and his co-workers suggested that tibia and fibula were in competition. In culture it is possible to add or remove mesenchyme from an otherwise normal leg bud. When they did this they found that the fibula was less stable than the tibia. Loss of mesenchyme produced a normal tibia and thin fibula: excess produced a normal tibia and a wide fibula. Inserting a mica plate between tibia and fibula makes the fibula grow much larger than normal. Presumably mesenchyme cells lying between the two rudiments are added to one or other blastema, normally in the ratio of some attractiveness factor. The mica barrier presumably gives the less attractive fibula a better chance.

What is this attraction? Other work suggests that the growth rate of a blastema is determined at the time of its formation. We can demonstrate this in culture. Hicks (OVERHEAD) isolated tibial and fibular cartilages complete with their perichondria and was able to demonstrate that in culture they grew at different rates. The tibial elements, four times the size of the fibula when explanted were twelve times as big after six days in culture.

Since the question of differential growth is an important one, and there is evidence that it is genetically controlled, lets stay with Hicks who went on to ask how the growth rates differed. By what cellular mechanism does the tibia grow much faster than the fibula?

There are a number of possibilities (OVERHEAD).

1. differential recruitment by condensations and later by chondrogenic blastemata of surrounding mesenchyme cells.

2. different cell division rates within condensations or chondrogenic rudiment.

3. different rates of synthesis of intracellular matrix

4. changes in cell volume during cartilage hypertrophy.

We have already looked at the first of these. Growth rate was already determined by the time the explants were made, and each grew on at the predetermined rate. Hicks looked at competition by growing two elements together in all possible combinations In no case did the presence of one element affect the other.

Cell division rates were assessed by counting dividing cells labelled with tritiated thymidine, incorporated into DNA at mitosis. Although there were regional differences there was no significant difference between tibia and fibula.

Hicks looked at matrix synthesis by measuring the incorporation of 35S into chondroitin sulphate. Over the two day measuring period the tibia was making more matrix than the fibula., but only because of its size: the rate per unit volume was very similar.

Hicks also counted cell numbers per unit area in various regions, using thin resin sections. Tibial cells seemed to be initially smaller than fibula cells, and became larger during hypertrophy.

So the only measurable changes are in cell volume during hypertrophy: this still doesn't quite answer the question, since there is differential growth before any hypertrophy can be identified. Perhaps the answer lies in the finding that tibial cells are initially smaller: if this is so then there must be more of them making up the tibial blastema.

One clearly important factor is the absence of a distal epiphysis in the chick: but again differential growth is seen before the epiphyses develop.

Elsewhere, notably in mouse chondrodystrophic mutants, cell kinetics have also been found to be abnormal. In mammals cartilage cells are arranged in columns in the ends of the diaphysis of longbones (OVERHEAD). The growth of a cartilaginous rudiment may thus be divided into two components:-

a. The increase in length of a cartilage column.

This is rather complex . It is essentially a product of the length of the column of hypertrophied cells (which, in practice, is nearly constant) and the rate of chondrocyte production, which varies widely from bone to bone, and according to age. The growth rate of the human femur, for instance, is the same at either end for the first year, but subsequently 70% of the growth is at the lower end. The maximal growth rate is in the order of 5 - 10 new cells per column per day.

b. The increase in the number of columns.

This is simpler: the number of columns is increased by appositional growth beneath the perichondrium.

Both mouse and chick chondrodystrophies are a rather mixed lot and often affect matrix synthesis. This is likely to have a knock on event on cell division and hypertrophy.

In a series of papers in mouse Thurston et al found various defects in different mutations. In achondroplasia itself they found reduced thymidine labelling (i.e. reduced cell division) and a reduction in hypertrophic cell height. The more affected individuals had no measurable cell division in the growth plates (but plenty in adjacent tissues) and reductions in cell numbers at all growth plate levels.

In another mutation, stumpy, cell division rate was again reduced, (OVERHEAD) cell divisions confined to the upper part of the column and hypertrophic cell height reduced throughout. So it looks as if all the factors that we deduced might affect cartilage growth in fact do so, but that the way in which these factors operate and the links between them are far from clear.

The cartilage in our skeletons is mainly replaced by bone. This process is also subject to modification. Talpid chicks have many problems: we have already mentioned their failure to produce distinct blastemata and their increased cell adhesion. Hinchliffe noted the additional failure of cartilage replacement bone in talpid embryos. The chick shoulder girdle (like our clavicle) contains both membrane and cartilage replacement bone. If a normal shoulder is explanted onto a chick chorioallantoic membrane both types of bone develop: talpid shoulders develop good membrane bone but no cartilage replacement bone, even a week after it should have been seen. The problem seems to lie in the periochondrium, where no osteoblasts are formed.

There are other mutations, which I shall not consider in detail, which lead to either defective bone matrix, making for rather fragile bones, or excessive and misplaced calcification, which may lead to ossification of hair follicles, joints and many other tissues. But one other class of mutation is worth mentioning, because in at least one animal the mutant phenotype has been adopted as the norm.

Because of the way in which the bony skeleton grows, by surface apposition, it is necessary to remove surplus bone to reshape elements during growth. The calvarium is a good example of this and the inner surface of the calvarium is a good source of the cells which do this, osteoclasts (OVERHEAD).

Microphthalmia in mouse is traditionally considered to be a gene whose chief skeletal effect lies in failure of bone to resorb. So what are microphthalmia osteoclasts like? They were reported by light microscopists as being small and scarce: this was confirmed under the e.m.: osteoclasts released from bone normally had up to 10 nuclei, in mi only one had more than one nucleus. The ruffled border, a characteristic sign of activity in osteoclasts, was missing (OVERHEAD). However above erupting tooth tips we were able to find osteoclasts with ruffled borders (OVERHEAD).

Osteoporosis in mi can be cured by parabiosis with normal littermates, or injection of spleen or liver cells into irradiated mi hosts. Conversely, irradiated normal mi hosts become osteopetrotic when injected with mi cells. So clearly the osteoclasts are dodgy: perhaps they are deaf. The signal to resorb bone is probably loudest over erupting teeth - perhaps just a few cells here get the message and respond.

In microphthalmia and other similar osteopetroses bones are thick and clumsy because they have not remodelled. Marrow cavities are often full of bone spicules (OVERHEAD).

This (OVERHEAD) is a manatee. Manatees and the similar dugongs belong to the order Sirenia, sea cows and probably separated a long time ago from a common stock, since one is Atlantic and the other Pacific. JZ Young states that the bones have a characteristic structure 'probably due to lack of stressing' Fawcett had other ideas. The long bones of the Florida manatee are characteristically clumsy and heavy because they have no marrow cavity: (OVERHEAD) they also have few osteoclasts, and thus resemble osteopetrotic mice.

What is the benefit of this to the manatee? How did it come about? Well this kind of dense bone is characteristic of marine fossil reptiles too, so perhaps being marine is the clue. Having heavy bones is certainly no disadvantage in water, in fact it is an advantage to a bottom grazing animal which must suffer from too much buoyancy. How did it come about? We can only guess, but the presence of this type of bone in so many different lines argues against repeated mutation. Sea cows eat seaweed: seaweed is full of iodine. Iodine excess interferes with thyroid function. Rare congenital absence of the thyroid and cretinism both produce odd bones. Osteopetrotic mice often have thyroid or parathyroid involvement. Impaired thyroid function also lowers metabolic rate; good in a sub aquatic browser.

So if we can't see how we can see a couple of whys. Iodine in the diet might slow you down, a benefit in a browser underwater. It might also mess up resorption - good for decreasing buoyancy and saving energy. As to whether it is genetic or environmental - no one has, to my knowledge, tried to raise manatees (which weigh about 650lb and eat a lot) on an iodine free diet.