Facial development in the mouse; a comparison between normal and mutant (amputated) mouse embryos

Facial development involves the complex and co-ordinated outgrowth and morphogenesis of a number of related facial structures; the facial arches and the naso-frontal process (Figs. 1–3). Patten (1961) has described facial develop-ment in the human embryo. In this paper Patten suggests that the originally distinct facial arches grow out and coalesce, not by a process of fusion as between the palatal shelves but by mergence. Mergence describes the expansion of each facial arch and the naso-frontal process as the adjacent mesenchyme proliferates, forcing the covering epithelium to lose its original shape, like a rubber glove being blown up so that finally no trace of the fingers remains. This description has been recently confirmed by Waterman & Meller (1973) in a study of hamster facial development.

Patten (1961) suggested that localized regions of active mesenchymal cell proliferation would be found to account for the details of facial morphogenesis. Minkoff & Kuntz (1977) have tested this hypothesis in the chick. They compared tritated thymidine labelling in the actively outgrowing naso-frontal region and adjacent control areas. The interesting result was that cell proliferation did not increase in the morphogenetically active region during the early stages of facial outgrowth, it stayed the same. But cell proliferation in the control region decreased. There was therefore a relative increase in cell proliferation in the naso-frontal region. Minkoff & Kuntz (1977) explained the outgrowth of this region and most probably the other facial regions in terms of this relative increase.

In the present paper facial development is analysed in normal and mutant mouse embryos. The mutant amputated a single, autosomal recessive mutation (Flint, 1976, 1977a) is characterized by a number of pleiotropic effects including taillessness, shortened limbs, fused ribs and vertebrae and a strongly fore-shortened snout. In addition to the important role of cell proliferation in facial development suggested by Minkoff & Kuntz (1977), an additional mechanism is revealed by this work. Three hypotheses are checked to account for the reduced outgrowth of the mutant snout. First, that the number of cells contributing to facial development is less in the mutant -the mutant has a poor start -secondly, cell proliferation is less in the mutant and, thirdly, that cell contact and adhesion are greater in the mutant restricting the expansion of the proliferating mesenchyme. Investigation of the third hypothesis not only confirms that cell contact is greater in the mutant facial mesenchyme as it is in mutant somite sclerotomal mesenchyme (Flint & Ede, 1978), but also shows that expansion of the facial mesenchyme, probably as a result of the secretion of extracellular matrix material, plays an important role in normal facial development. Part of this work has already appeared as an abstract (Flint, 1977b).

Mice from a CBA/101 hybrid stock carrying the amputated gene were mated and pregnant females opened 9, 10, 12 and 16 days after the first observation of a vaginal plug. Time of conception was taken to correspond with an ovulation at the middle of the preceding dark period (Snell, Fekete, Hummel & Law, 1940). Embryos at 10·5, 12·5 and 16·5 days were fixed in formol-alcohol taken to water and stained overnight in luxol-fast blue, then prepared as wholemounts in benzyl benzoate. A group of 16·5-day embryos was stained for bone with alizarine red (see Humason, 1972, p. 195). Some 9·5-and 10·5-day embryos were fixed in Bouin, wax embedded and sectioned longitudinally for counts of metaphase nuclei in the embryonic facial mesenchyme. Others were fixed and prepared for 1 μm plastic sections and for scanning and transmission electron microscopy after the method described by Flint & Ede (1978). Quantitative analysis of cell contacts and cell surface area also followed the techniques described by Flint & Ede (1978). Statistical analysis followed the standard analysis of variance described by Sokal & Rohlf (1969) or Snedecor & Cochran (1967). Wherever tests showed this to be necessary, data was normalized by square root transformation. A probability of P < 0·05 was taken to indicate a significant difference between groups of results. The numbers of embryos analysed and cells counted can be found in Table 1.

Facial morphology does not appear to be significantly different at 10·5 days in normal and amputated embryos. By 12·5-days growth of the naso-frontal region in amputated is retarded though no difference can be seen between mandibular regions. A difference between mandibular regions becomes clear in the 16·6-day embryos. The size and shape of cranial regions posterior to the naso-frontal/mandibular region are not affected in the mutant. Alizarine red bone clearances of 16·5-day embryos (Fig. 4) also confirm that structures in the head posterior to the snout region are normally formed in the mutant, but that the bones associated with the snout, though present, are greatly telescoped together.

No clear difference can be seen between mutant and normal heads at 9·5 days of development either in shape or size (Fig. 2). By 10·5 days the first differences can be observed (Fig. 3), differ-ences that could not be observed in wholemounts under the light microscope. Though there is no clear difference in the size of the facial structures in mutant and normal, in amputated the olfactory pit is not so deeply invaginated as it is in the normal embryo.

The naso-frontal region is the region first affected in the mutant. The first differences between mutant and normal can be detected at 10·5 days. None can be seen at 9·5 days. These two stages were therefore further analysed with particular reference to the naso-frontal region to determine the etiology of the abnormality at the cellular level. The differences observed would then most probably be associated with the onset of facial foreshortening rather than secondary effects caused by the abnormality.

Much, though not all, of the mesenchyme contributing to the outgrowth of the facial structures is derived from the cranial neural crest (Johnston, 1966; Weston, 1970; Noden, 1975; Le Douarin, 1976). A small contribution is made by the primary mesenchyme formed after immigration through the primitive streak. For example, Le Douarin (1976) and Noden (1975) found that muscle associated with the facial arches did not contain labelled cells of crest origin after grafting labelled neural crest to host chick embryos. Homozygous amputated embryos cannot be detected before 8·5–9·0 days of development because differences of axial length between mutant and normal embryos resulting in a shorter mutant embryo (Flint, 1976; Flint, Ede, Wilby & Proctor, 1978) are not yet apparent. Consequently a direct test, possibly involving the grafting of the earlier formed neural crest to chick hosts, was not possible. There is, however, some indirect evidence. Neural crest derivatives in the mutant embryo do not appear to be generally deficient. For example, the dorsal root ganglia achieve the same size and number as in the normal embryo (Flint, 1976, 1977a) even though the axis is much shorter. The dermal skeleton of the skull, including the tympanic ring in the ear, is not deficient (Fig. 4) and the dermis is of neural crest origin (Le Douarin, 1976). The mandibular, maxillary and neasal skeleton is shortened but not missing. It is not possible to determine whether the shortening is a secondary consequence of facial shortening or primarily caused by a neural crest deficiency. However, if there were a neural crest deficiency, one would expect it to affect all neural crest derivatives rather than to be localized in this manner.

There is another line of argument. At 9·5 days of development there is no differences of size or shape between the heads of normal and mutant embryos, nor (see Tables 3, 4) is there any significant difference in the cell density found in the naso-frontal mesenchyme (normal: 8·54±1·02 cells per 10³ μm²; amputated: 8·95 ±2·56 cells per 10³ μm²; P > 0·25 -data from Tables 3, 4, plastic sections) or in the mandibular regions (normal: 8·31 ±0·76 cells per 10³ μm²; amputated: 7·89 ±0·38 cells per 10³ μm²; P > 0·10 -data from 22 wax sections of two normal and two amputated embryos). Whatever the proportional contributions made by the primary mesenchyme or the cranial neural crest the total number of cranial mesenchyme cells in the facial region is therefore approxi-mately the same in normal and in amputated embryos at 9·5 days of development.

The evidence suggests that amputated does not begin facial development at a disadvantage because there are fewer mesenchymal cells available at the time the facial anomaly is about to appear. Nor does there appear to be any anomalous development of the neural crest cells and their derivatives which cannot on this evidence be accounted for as other than secondary to the general anomaly of facial development in the mutant.

Cell proliferation is given as mitotic index, the number of dividing nuclei as a percentage of the total number of cells scored. Mitotic index was scored not only in the naso-frontal region, but in the other important region for snout outgrowth, the mandibular arch. A third control region, latero-ventral and just posterior to the fore-brain, not involved in facial outgrowth was also scored. Average values, standard deviations and a statistical analysis can be found in Table 2.

In none of the sections of mandibular or naso-frontal mesenchyme did there appear to be any anatomically asymmetrical distribution of mitotic cells which might contribute to facial outgrowth. Within any of the three regions, whatever the age or the embryo type, there is no significant difference in the levels of mitotic index. Nor is there any difference of mitotic index between the naso-frontal regions and the mandibular regions. But there is a very significant difference (P < 0·001) between the control region and the two regions of active facial outgrowth; the control region having the lower mitotic index overall. In other words, there is no temporal change in cell proliferation in any region, but mesenchyme of the early snout region has a slightly higher level of cell proliferation than is found elsewhere in the head mesenchyme. These results agree with those of Minkoff & Kuntz (1977). The mouse embryos at 9·5 and 10·5 days were equivalent to chick stages at 18–22 or the very earliest chick stages that Minkoff & Kuntz (1977) analysed. At this stage they found a differ-ence of thymidine labelling index (labelled nuclei as a% of total) between control and naso-frontal regions of 2·9%, equivalent to the very small difference found here, which is statistically significant because of the very large number of nuclei counted (see Table 1). It was at later developmental stages in the chick that Minkoff & Kuntz (1977) found an increasing difference of labelling index between naso-frontal and control regions in the chick. These stages in the normal and mutant mouse were not analysed because differences of naso-frontal outgrowth (see above) are already apparent in the 10·5-day embryo and therefore if mitotic index were important there should already be a difference between mutant and normal at this stage. In fact, there is no significant difference between mutant and normal (Table 2) within any of the three regions. We can conclude that differences of cell proliferation rate do not play an important role in reducing facial outgrowth in the amputated embryo.

Somite cells of the amputated embryo, both in vivo (Flint & Ede, 1978) and in vitro (Flint, 1977 and in preparation), have greater areas of cell-cell contact. In culture the increased areas of cell contact reduce the speed of cell movement. Increased cell-cell contact is therefore probably to be equated with increased strength of cell adhesion. An inhibition of cell locomotion by increased strength of cell-cell adhesion is all that is required to explain the reduced axial length of amputated embryos (Flint et al. 1978). Increased cell-cell adhesion may also be involved in the production of the fused and distorted vertebral chondrogenic condensations which are found in amputated (Flint, 1977; Flint & Ede, 1978). In another mutant, the talpid³ chick, an increase of cell adhesion has been demonstrated (Ede & Flint, 1975a, b) and accounts for many of the features of anomalous limb development in the mutant (Ede, 1976). Changes of cell con-tact and cell mobility are also involved in the expression of other mutant phenotypes, t⁹ (Spiegelman & Bennett, 1974; Spiegelman, 1976; Yanigasawa & Fujimoto, 1977) and brachypodia (Elmer & Selleck, 1975; Duke & Elmer, 1977) in the mouse. On this evidence it is a good prediction that increased areas of cell-cell contact will be found in amputated facial mesenchyme. If this is the case, how can it account for anomalous mutant facial development?

General appearance. 1 μm sections were cut transversely to the naso-frontal region and blocks were trimmed to obtain EM sections of the naso-frontal mesenchyme. At 9·5 days there are small areas of cell contact between most normal cells in the closely packed mesenchyme. Much more extensive regions of contact are observed between amputated cells (Fig. 5 A, B). There are no other superficial differences. The cytoplasm of both mutant and normal cells is full of ribosomes and there is occasional rough endoplasmic reticulum.

By 10-5 days the cell density in both normal and amputated mesenchyme has apparently fallen. The differences between mutant and normal are now more striking. Normal cells have developed long and tenuous filopodia by means of which narrow regions of contact are maintained (Fig. 5C). In the mutant, areas of cell contact are less than they were at 9·5 days, but elongate filopodia are rare (Fig. 5D). There are no cytoplasmic differences between amputated cells at 10·5 days and amputated and normal cells at 9·5 days. But in normal cells at 10·5 days, the rough endoplasmic reticulum is much more extensive and mitochondria more frequent (Fig. 5C).

Higher resolution electron micrographs confirm that normal cells have mainly narrow regions of contact as opposed to the generally extensive contacts between amputated cells (Fig. 6 A, B). Further differences are now apparent. As in the presomitic mesenchyme of the 9-5-day amputated embryo (Flint & Ede, 1978), filopodia are sometimes trapped between adjacent mutant cells but rarely, if ever, between normal cells (Fig. 6A, B). Desmosome-like junctions are very rare in normal facial mesenchyme but in the mutant they are frequent and extensive (Fig. 6C–F). Occasionally sections of mutant, but not of normal, mesenchyme cut through, knotted bundles of interdigitating cellular processes (Fig. 6F). Similar ‘knots’ are observed in the sclerotomal mesenchyme in the mutant (Flint & Ede, 1978). The cytoplasm in the knotted cell processes (Fig. 6F) appears to be full of microfibrils. Polysomes are frequent in both mutant and normal cytoplasm.

Narrow junctions made by normal cells at 10·5 days are associated with a localized, very dense and granular cytoplasm (Fig. 7A, C), and occasionally with orientated microfibils. This is also true in amputated where narrow junctions occur (Fig. 7B). But unlike the normal filopod, the amputated. filopod buries itself in a well-like pit formed by the walls of the adjacent cell (Fig. 7B). Very long junctions of the desmosome type are found throughout amputated, but not normal facial mesenchyme at 10·5 days (Fig. 7D).

At 9·5 days there is no significant difference of cell density in normal and amputated facial mesenchyme (Tables 3, 4), but by 10·5 days there has been a significant reduction in both, though not so great in the mutant.

At 9·5 days there is no difference in the average measured length of each cell’s perimeter in mutant and normal mesenchyme (Tables 3, 4) but by 10·5 days there is a significant expansion of cell surface in both, though not so great in the mutant. This confirms the EM observations.

Cell contact length per cell, whether measured as an absolute value or as a percentage of the cell perimeter is always greater in mutant than in normal facial mesenchyme (Tables 3, 4). At 9·5 days nearly three times as much mutant cell surface is involved in cell-cell contacts as in the normal mesenchyme. By 10·5 days the difference is less, approximately twice as much in the mutant, but it is still significant. In both mutant and normal there is a reduction of cell contact area between 9·5 and 10·5 days.

The number of cells in contact in any of the fields photographed (5·8 × 10³ μm²) is not significantly different in normal cells at 9·5 and 10.5 days, but in the mutant approximately three times as many cells are in contact at 9·5 than at 10·5 days (Table 3, 4). It follows that three times as many mutant as normal cells are in contact at 9·5 days. Though far fewer mutant cells are in contact at 10·5 days than at 9·5 days, there are still sig-nificantly more than in normal facial mesenchyme. This may be because cell density is slightly less in normal than amputated at 10·5 days. The number of cells in contact is cell density dependent (Flint & Ede, 1978).

The purely observational, morphological data suggests that contacts are larger and adhesions are stronger between mutant than between normal cells. The evidence for this is: (i) the more extensive gap junctions, (ii) more and more extensive desmosome-like junctions, (iii) contacted cell surfaces tending to wrap round the ends of contacting filopodia, and (iv) the tangled knots that form between numbers of contacting filopodia. The quantitative data confirms this. First, at 9·5 days three times as many amputated cells are in contact with one another as normal cells in an equivalent area. Secondly, at the same stage, nearly three times as great a per-centage of the amputated as the normal cell surface is involved in contacts with other cells. At 10·5 days, though cell density and areas of contact have reduced in both normal and amputated still twice as great a percentage of mutant as normal cell surface is involved in cell contacts.

In a similar quantitative analysis of the somite sclerotome in amputated and normal mice, Flint & Ede (1978) also found that greater percentages of mutant cell surface were involved in cell contacts and that more mutant cells were in contact at equivalent cell densities. Electron micrographs suggest that this is also true of the presomitic primary mesenchyme (Flint & Ede, 1978). This difference between mutant and normal appears to be general for mutant mesenchyme cells. It is highly likely that the mutant gene is causing some change at the cell surface which alters cell contact behaviour. We cannot however rule out changes in the cytoskeleton, though these have not appeared in our electron micrographs or changes in the extracellular matrix which can alter cell contact behaviour and cell mobility (Toole, Jackson & Gross, 1972; Fisher & Solursh, 1977). These questions are currently being investigated.

There is a reduction of cell density between 9·5 and 10·5 days. Cells move apart. But the number of cells in contact does not change because of the extension of elongate filopodia, involving an increase in cell surface area, through the newly formed large intercellular spaces. These spaces cannot be empty, but must be full of inter-cellular matrix, most probably glycosaminoglycans (Strudel, 1976; Fisher & Solursh, 1977). There has therefore been a massive secretion of intercellular materials between 9·5 and 10.5 days and a corresponding differentiation of cell morphology (also confirmed by the cytoplasmic changes described). At this stage cell proliferation is not an important factor in facial outgrowth. Our own results and those of Minkoff & Kuntz (1977) show that there is only a very small difference between cell proliferation in the naso-frontal region and adjacent control areas. Yet this must be an important stage in facial morphogenesis because the first signs of retarded facial outgrowth are visible in the amputated embryo at 10-5 days. A most significant factor in facial outgrowth at this early stage must therefore be the secretion of extracellular matrix and the consequent increase in extracellular volume forcing the cells to move apart.

In amputated there is also an increase in extracellular volume between 9·5 and 10·5 days, but cell morphology does not change significantly. Cells are forced to move apart but do not maintain contact by the extension of long filopidia (there is a much smaller increase in cell surface area with reduction of cell density in amputated) and so a threefold reduction in the number of cells in contact follows. But at 10·5 days the percentage of cell surface area involved in cell contacts, though less than at 9·5 days, is still twice as much as in normal mesenchyme. Cell density does not fall so far in amputated as in normal facial mesenchyme and the evidence suggests that the increase in extracellular volume is retarded by the stronger and more extensive contacts formed between the cells. We conclude that facial outgrowth in amputated is retarded in the early stages of facial morphogenesis by the mechanism suggested and that this slowing up of outgrowth must be the primary cause for the anomalies of facial morpho-genesis that have been described.

Facial morphogenesis is retarded in the snout region of the amputated mouse mutant. This anomaly cannot be accounted for in terms of a reduced cell population, or abnormally low proliferation rates. There is a reduction of cell density, an increase of tissue volume at the early stage of facial morphogenesis between 9·5 and 10·5 days of development which is a significant factor in early facial outgrowth. But this increase in tissue volume is retarded in the mutant by the tendency of mutant cells to clump together as a result of stronger and more extensive intercellular contacts. It is becoming apparent that whenever the mutant cells are presented with a situation in which they have to move, either passively as here, or actively as in the sclerotome (Flint & Ede, 1978), in culture (Flint, 1977 a) or in the regression of the streak (Flint et al. 1978) an anomaly of development ensues. It is interesting from this viewpoint that limb outgrowth is also retarded in the mutant (Flint, 1976). Ede & Law (1969) and Ede & Flint (1975 a) have stressed the importance of cell movement in limb outgrowth.

The authors would like to thank the SRC for support in this work.