The Journal of Nutrition Vol. 128 No. 2 February 1998, pp. 471S-475S

The Role of Vitamin A in the Development of the Central Nervous System^1,2

Malcolm Maden³, Emily Gale, and Maija Zile^*

Developmental Biology Research Centre, King's College London, London WC2B 5RL U.K. and ^* Department of Food Science and Human Nutrition, Michigan State University, East Lansing, MI 48824

	ABSTRACT

Abstract Introduction Methods Results Discussion References

We describe here the defects that arise in the central nervous system (CNS) of quail embryos when they develop in the absence of vitamin A. It has been assumed that because of the effects of excess vitamin A and its metabolites, particularly retinoic acid (RA), on the CNS they are involved in various aspects of CNS development. We show that this is indeed the case, because these deficient quail embryos have three defects in their CNS. First, the posterior hindbrain fails to develop because the cells fated to form this part of the CNS in the very early embryo die by apoptosis. Second, the neural tube fails to extend neurites into the periphery both in vivo and in vitro. Third, the neural crest cells throughout the embryo die by apoptosis. These results demonstrate a crucial requirement for vitamin A in CNS development.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

It was shown in the 1930s using farm animals that embryos born to vitamin A-deficient females had many congenital defects, including abnormalities of the central nervous system (CNS) and neural crest derivatives (review, Kalter and Warkany 1959). Conversely, it was subsequently shown in the 1950s (Cohlan 1953) that too much vitamin A is equally harmful to the embryo and that a similar spectrum of defects arises in the CNS and neural crest derivatives. Most recently, this has also been shown to be true in humans with the appearance of cerebellar and craniofacial defects in babies born to mothers who had taken 13-cis-retinoic acid for the treatment of various skin problems (Lammer et al. 1985). It would be reasonable to conclude from this that the correct level of vitamin A is required for appropriate CNS development and that either too much or too little is equally harmful to the embryo.

In contrast to this emphasis on defective development that emerges from in vivo teratological studies of the CNS, in vitro studies have revealed very important stimulatory effects of vitamin A on neurite outgrowth in cultures of various types of neuronal cells and also on neuronal differentiation in pluripotent embryonal carcinoma (EC) cells. Thus the active metabolite of vitamin A, all-trans-retinoic acid (RA), induces EC cells to differentiate into a range of cell types, including neural cells, depending on the concentration applied (Edwards and McBurney 1983), and it induces neurites to form in neuroblastoma cells (Sidell 1982). In cells that are already neuronal, RA induces either neurite extension where there was none before or longer neurites if neurites were already present in the culture conditions. This has been shown in dissociated cultures or explanted tissue using embryonic dorsal root ganglia, spinal cord or sympathetic ganglia from chick, mouse, rat and human embryos (review, Maden and Holder 1992).

In the light of these positive results, one might have expected the in vivo effects of RA on the CNS to have been more stimulatory than the teratology described above. In fact, recent, more detailed studies have borne this expectation out. For example, in both mouse embryos (Leonard et al. 1995, Morriss-Kay et al. 1991) and zebrafish embryos (Holder and Hill 1991), when RA is administered at gastrulation a segment of the anterior hindbrain/posterior midbrain is missing. Most interestingly, if the dose of RA is slightly lowered, then instead of a deletion, a segment of the anterior hindbrain becomes respecified into another, more posterior segment (Hill et al. 1995, Marshall et al. 1992). The segments of the hindbrain are known as rhombomeres (see Fig. 1), and in these experiments rhombomere 2 takes on some of the characteristics of rhombomere 4. In the zebrafish experiments, the difference in doses between deletion and respecification was the difference between 0.15 µmol and 0.1 µmol, a remarkably small difference for such a strikingly different result.

View larger version (36K): [in this window] [in a new window]   Fig 1. A and B. Line drawings of the CNS of a normal (A) and an A (B) stage 12 quail embryo to show the segmented shape of the brain at this stage. The forebrain (fb) and midbrain (mb) are very similar in the normal and A embryos. But the hindbrain, shown as a series of seven rhombomeres in the normal embryo (numbers 1-7) followed by the spinal cord (sc) has only three rhombomeres (numbered 1-3). Next to each embryo are the expression domains of four genes that we have used to determine which rhombomeres are missing in the A embryo. Krox-20 is expressed in r3 and r5 in the normal embryo, and in the A embryo only one anterior stripe (the r3 stripe) is expressed. Fibroblast growth factor-3 (Fgf-3) is expressed in r4, r5 and r6 in the normal embryo and is not expressed in the A embryo. Hoxb-1 is expressed in r4 and then from r6 backwards in the normal embryo, and in the A embryo it is only expressed in the straight part of the neural tube. Hoxa-2 is expressed weakly in r2 (represented by stippling), strongly in r3, weakly in r4 and strongly from r5 backwards in the normal embryo, and in the A embryo it is expressed weakly in the anteriormost rhombomere, followed by strongly backwards. From these data, we deduce that the rhombomeres present in the A embryo are r1, r2 and r3. C: Patterns of apoptosis in a normal (left) and an A (right) embryo at the 7 somite stage drawn from wholemount TUNEL embryos. In the normal embryo there are no obvious apoptotic cells at this stage, but in the A embryo a band of apoptosis (stippling) is present in the neuroepithelium of the presumptive hindbrain. D: Drawings of neurofilament antibody immunreactivity in the neural tube of a normal (left) and an A (right) stage 19 embryo. The neurofilament expression is represented by stippling and in the normal embryo (left) is present in the lateral edges of the neural tube and within the dorsal root ganglia (drg) and sensory axons and emerging from the neural tube ventrally as motoneurons (vm). The notochord (n) is shown below the neural tube. In the A embryo (right), the neural tube is clearly underdeveloped and neurofilament expression only appears weakly around the periphery of the tube and in some cases crosses the dorsal mid-line. There is no staining in the periphery, suggesting that axons do not emerge from the A neural tube. E: Drawings of TUNEL wholemount embryos to show the distribution of apoptotic cells (stippling) in a stage 17 A embryo. Left: Section through the trunk showing two streams of apoptotic cells, one moving laterally underneath the ectoderm and one travelling ventrolaterally. These are the two directions in which neural crest cells travel from their origin in the dorsal neural tube. Right: Lateral view of the head showing a long stream of apoptotic cells travelling dorsal to the eye and a short stream that stops before it reaches the eye. These are the neural crest cells that would have formed the trigeminal galglion and the branches of the trigeminal nerve. Note also apoptotic cells in the first branchial arch.

In the experiments reported here, we are concerned with detailed studies on the other aspect of the teratology referred to above, namely what happens to the CNS in the absence of vitamin A.

It seems that the easiest experimental system for generating vitamin A-deficient (A) embryos are birds. The generation of A chick embryos was first described many years ago by Thompson et al. (1969), and he observed that in such embryos the cardiovascular system was most obviously affected. These observations have more recently been repeated and extended using quail embryos (Dersch and Zile 1993, Heine et al. 1985, Zile 1998).

From the point of view of CNS development, this A quail system is obviously extremely valuable for asking questions about the role of vitamin A that thus far have only been surmised from experiments described above on the effects of excess RA. There are three defects that we have characterized in the CNS and neural crest of these A embryos (Maden et al. 1996). One, the posterior hindbrain is completely missing. Two, the neural tube fails to extend neurites out into the periphery. Three, the neural crest cells die. Each of these observations will now be reviewed and further experiments reported.

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

Normal and vitamin A-deficient Japanese quail (Coturnix coturnix japonica) embryos were raised at the Michigan State University Poultry Research Farm as described in detail by Dersch and Zile (1993). Eggs were incubated at 38°C and the embryos staged according to the method of Hamburger and Hamilton (1951). For cultures of neural tubes from A quails, explants were taken from 2.5-d deficient embryos and cultured for 48 h in either basic medium, basic medium plus 25% fetal calf serum or basic medium plus 0.1 µmol all-trans-RA. Basic medium consisted of F-14 (Ham's nutrient mixture F-14 with L-glutamine) with 0.001 µmol nerve growth factor and serum supplement N2 (Bottenstein and Sato 1979).

	RESULTS

Abstract Introduction Methods Results Discussion References

The posterior hindbrain is missing.  During early development of the CNS in the normal vertebrate embryo, the initially smooth, cylindrical neural tube develops periodic undulations or neuromeres that characterize the various brain regions: the forebrain, the midbrain and the hindbrain. Within the hindbrain, known as the rhombencephalon, there are seven of these periodic undulations called rhombomeres (Fig. 1A). These rhombomeres have become a subject of intense interest in developmental neurobiology because they resemble the segments of insects in that they are lineage-restricted (Fraser et al. 1990), they have different adhesive properties (Guthrie and Lumsden 1991) and the rhombomere borders represent the anterior boundaries of expression of the 3 members of the Hox gene clusters (review, McGinnis and Krumlauf 1992).

By using the expression patterns of various Hox and other genes that are expressed in one or several rhombomeres, each rhombomere (r) can be individually identified in the embryo. For example, the gene Krox-20 is expressed in r3 and r5, Fgf-3 is expressed in r4, r5 and r6, Hoxa-2 is expressed in r2-r7 in levels of varying intensity, and Hoxb-1 is expressed in r4 and r7 backwards (Fig. 1A). In the A embryo, sections reveal that instead of the normal complement of seven rhombomeres there are only three, and the expression patterns of these genes revealed them to be r1, r2 and r3 (Maden et al. 1996) (Fig. 1B).

Our most recent studies (Maden et al. 1997) have shown that the loss of this tissue comes about very early in development, as the neural tube forms. Soon after gastrulation, at the 5 somite stage, we observed a band of cells in the mesoderm in the region of the first somite undergoing programmed cell death. This is followed 2 h later by a band of cell death in the neuroepithelium of the prospective hindbrain (Fig. 1C). Thus the absence of the posterior hindbrain is caused by a highly localized region of cell death occurring as the hindbrain becomes specified at the 7 somite stage. This is the time when the first segmentation event takes place in the neuroepithelium, and we presume that this loss of tissue in the A embryo is caused by an aberrant pattern of expression of the genes that are responsible for the regionalization of the anteroposterior axis of the embryo.

Two things are striking about these results. First, the posterior hindbrain region that is missing in these A embryos is one complete unit known as the myelencephalon, which stretches from r4 to r7. Secondly, excess RA causes defects in the anterior hindbrain (Leonard et al. 1995, Morriss-Kay et al. 1991), that is, anterior to the r3/4 border, as a result of the anterior expansion of Hox expression domains (Conlon and Rossant 1992), whereas a lack of RA causes defects in the posterior hindbrain, posterior to the r3/4 border, as a result of a posterior regression of Hox gene expression. It is highly likely therefore that Hox genes play a role in anteroposterior patterning and are under the control of RA in the early embryo. Furthermore, the r3/4 border seems to be an important boundary region for patterning the hindbrain.

Failure of neurite outgrowth from the neural tube.  The second defect in the CNS of the A quail embryo is that as differentiation of neurons in the neural tube begins, axons do not appear within the surrounding mesoderm as one expects and the axon trajectories within the neural tube itself are often unorganized and chaotic. At about stage 11, neurofilament antibody staining of normal quail embryos begins and over the subsequent stages develops into a typical pattern of peripheral localization within the neural tube and ventral motor and dorsal sensory axon tracts in the periphery (Fig. 1D).

In contrast, there is a delay of several stages in the onset of neurofilament protein expression in the neural tube of the A embryo, and the pattern within the neural tube never approaches that of normal. By stage 19, for example, no dorsal sensory axons or ventral motor axons can be seen in the periphery expressing neurofilament proteins, and within the neural tube itself there are considerably fewer expressing neurons than normal (Fig. 1D). Although neural crest cells, which show positive immunoreactivity to the antibody HNK-1, accumulate in the appropriate location of the dorsal root ganglia, they never begin to express neurofilament proteins as the controls do. Furthermore, some axons within the neural tube seem to display a chaotic trajectory, for example by crossing the dorsal mid-line, a phenomenon that never happens in the normal embryo.

We have confirmed that this failure to extend neurites into the periphery was not due to the presence of an inhibitory molecule in the sclerotome surrounding the neural tube in the following manner. We cultured the neural tubes of A embryos either in medium, medium with serum or medium with 0.1 µmol of tRA. Figure 2 A, B shows that explants of the neural tube from A embryos extend very few neurites into the medium, confirming the in vivo data. The cells were, however, healthy, because flat cells migrated out from the explant during the period of culture, and one of the explants produced one neurite (Fig. 2B). Adding tRA to the explants allowed good neurite outgrowth to occur (Fig. 2C, D), demonstrating that this single compound could rescue these explants. Similarly, adding fetal calf serum, which contains high levels of retinoids as well as other nutrients, also stimulated neurite outgrowth from A explants (Fig. 2E, F).

View larger version (146K): [in this window] [in a new window]   Fig 2. Top: Low power (A) and high power (B) micrographs of a neural tube explant from a 2.5-d A embryo cultured for 48 h in basic medium. The cells are still perfectly healthy and only one neurite has grown from the explant. The other explants gave no neurites but had flat cells migrating out from them. Middle: Low power (C) and high power (D) micrographs of an A neural tube explant with 0.1 µmol of all-trans-RA added to the basic medium. Neurites are clearly growing out well in these cultures. Bottom: Low power (E) and high power (F) micrographs of an A neural tube explant with 25% fetal calf serum added. Again, neurites are growing out well in these cultures. Magnification: low power =×100, high power =×400.

Apoptosis in the neural crest.  The reason why the dorsal root ganglia never express neurofilament antibodies as described above is that soon after the neural crest cells have ceased migrating and accumulate in the location of the ganglia, they begin to undergo cell death. This neural crest cell death begins at about stage 14 and affects all neural crest cells throughout the embryo, whether they have finished migrating or are still migrating. Our studies with the TUNEL technique have revealed streams of apoptotic cells that coincide with the known pathways of migration of neural crest cells. Figure 1E shows a cross-section of the trunk and a lateral view of the head of an A embryo with apoptotic cells marked. In both cases, the apoptotic cells are to be found in streams that coincide with streams of neural crest cells. In our previous study of this phenomenon (Maden et al. 1996), we showed that cells that are undergoing apoptosis in the dorsal root ganglia also stain with the HNK-1 antibody, supporting our contention that these are neural crest cells and that they need retinoids to survive.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

There can be no doubt that the embryonic CNS requires vitamin A and its derivatives, the retinoids, particularly RA, for its proper development. An excess of RA causes specific defects of the anterior hindbrain whereby anterior rhombomeres can be lost or respecified to more posterior ones (see Introduction) or it can cause a posteriorization of the whole CNS whereby forebrain structures are lost (Avantaggiato et al. 1996, Cunningham et al. 1994, Durston et al. 1989). These effects seem to be stage-specific. We describe here our experiments on the opposite situation, which concerns CNS development in the absence of RA, and we show that there are three severe defects present in the embryos.

First, a complete segment of the hindbrain, known as the myelencephalon, fails to develop because the cells fated to form this part of the posterior hindbrain die very early in development by apoptosis. We suggest that this arises because, in the absence of RA, anteroposterior specification of the early embryo after gastrulation is abnormal and the primary boundary in the hindbrain fails to be established (Conlon 1995).

Second, the neural tube fails to extend neurites into the periphery, and some of the neurites within the neural tube follow a chaotic trajectory. We demonstrated that this was not due to the presence of an inhibitory factor in the surrounding mesenchyme, but rather due to an intrinsic defect in the CNS, by culturing the neural tubes from A embryos. Such neural tubes still failed to extend neurites in culture, but on the addition of RA, neurites grew. This supports the many experiments using dissociated cultures or explants of neural tissue that have shown a role for RA in stimulating neurite outgrowth (review, Maden and Holder 1992). Furthermore, we have evidence that RA is chemotactic for neurites in the chick embryo neural tube (Maden, M. and Jones, G. E., unpublished results), and this would explain the disorganized appearance of neurites in the neurofilament stained A neural tube.

Third, the neural crest cells in these A embryos die by apoptosis. Interestingly, it has previously been suggested that neural crest cells need RA for survival in culture (Henion and Weston 1994), and we have clearly supported this idea with in vivo data. Excess RA, however, has a different effect on the neural crest, because it causes the streams of crest that emerge from the hindbrain to re-orientate and travel in aberrant directions (Gale et al. 1996).

Thus both excess RA and a deficiency of RA cause abnormal development of the CNS. It is important to emphasize that these are direct effects of RA because RA is an endogenous component of the developing CNS. High pressure liquid chromatography (Horton and Maden 1995), a transgenic reporter construct (Rossant et al. 1991) and a reporter cell line (Wagner et al. 1992) have all demonstrated that the embryonic CNS, particularly the spinal cord, is a site of high RA content. Clearly, the embryo must strictly regulate its synthesis, and studies on the enzymes involved will surely reveal important information about how the incredible complexity of the nervous system comes about.

	FOOTNOTES

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Avantaggiato V., Acampora D., Tuorto F., Simeone A. Retinoic acid induces stage-specific repatterning of the rostral central nervous system. Dev. Biol. 1996; 175:347-357 [Medline] [Medline]
• Bottenstein J. E., Sato G. H. Growth of a rat neuroblastoma cell line in serum-free supplemented medium. Proc. Natl. Acad Sci. U.S.A. 1979; 76:514-517 [Medline] [Abstract/Free Full Text]
• Cohlan S. Q. Excessive intake of vitamin A as a cause of congenital abnormalities in the rat. Science 1953; 117:535-536 [Free Full Text]
• Conlon R. A. Retinoic acid and pattern formation in vertebrates. Trends Genet. 1995; 11:314-319 [Medline] [Medline]
• Conlon R. A., Rossant J. Exogenous retinoic acid rapidly induces anterior ectopic expression of murine Hox-2 genes in vivo. Development 1992; 116:357-368 [Medline] [Medline]
• Cunningham M. L., Mac Auley, A., Mirkes P. E. From gastrulation to neurulation: transition in retinoic acid sensitivity identifies distinct stages of neural patterning in the rat. Dev. Dyn. 1994; 200:227-241 [Medline] [Medline]
• Dersch H., Zile M. H. Induction of normal cardiovascular development in the vitamin A-deprived quail embryo by natural retinoids. Dev. Biol. 1993; 160:424-433 [Medline] [Medline]
• Durston A. J., Timmermans J.P.M., Hage W. J., Hendricks H.F.J., de Vries N. J., Heideveld M., Nieuwkoop P. D. Retinoic acid causes an anteroposterior transformation in the developing nervous system. Nature (Lond.) 1989; 340:140-144 [Medline] [Medline]
• Edwards M.K.S., McBurney M. W. The concentration of retinoic acid determines the differentiated cell types formed by a teratocarcinoma cell line. Dev. Biol. 1983; 98:187-191 [Medline]
• Fraser S., Keynes R., Lumsden A. Segmentation in the chick embryo hindbrain is defined by cells lineage restrictions. Nature (Lond.) 1990; 344:431-435 [Medline] [Medline]
• Gale E., Prince V., Lumsden A., Clarke J., Holder N., Maden M. Late effects of retinoic acid on neural crest and aspects of rhombomere identity. Development 1996; 122:783-793 [Medline] [Abstract]
• Guthrie S., Lumsden A. Formation and regeneration of rhombomere boundaries in the developing chick hindbrain. Development 1991; 112:221-229 [Medline] [Abstract]
• Hamburger V., Hamilton H. L. A series of normal stages in the development of the chick embryo. J. Morphol. 1951; 88:49-92
• Heine U. I., Roberts A. B., Munoz E. F., Roche N. S., Sporn M. B. Effects of retinoid deficiency on the development of the heart and vascular system of the quail embryo. Virchow's Arch. B Cell Pathol. 1985; 50:135-152 [Medline]
• Henion P. D., Weston J. A. Retinoic acid selectively promotes the survival and proliferation of neurogenic precursors in cultured neural crest populations. Dev. Biol. 1994; 161:243-250 [Medline] [Medline]
• Hill J., Clarke J.D.W., Vargesson N., Jowett T., Holder N. Exogenous retinoic acid causes specific alterations in the development of the midbrain and hindbrain of the zebrafish embryo including positional respecification of the Mauthner neuron. Mech. Dev. 1995; 50:3-16 [Medline] [Medline]
• Holder N., Hill J. Retinoic acid modifies development of the midbrain-hindbrain border and affects cranial ganglion formation in zebrafish embryos. Development 1991; 113:1159-1170 [Medline] [Abstract]
• Horton C., Maden M. Endogenous distribution of retinoids during normal development and teratogenesis in the mouse embryo. Dev. Dyn. 1995; 202:312-323 [Medline] [Medline]
• Kalter H., Warkany J. Experimental production of congenital malformations in mammals by metabolic procedure. Physiol. Rev. 1959; 39:69-115 [Free Full Text]
• Lammer E. J., Chen D. T., Hoar R. M., Agnish A. D., Benke P. J., Braun J. T., Curry C. J., Fernhoff P. M., Grix A. W., Lott I. T., Richard J. M., Sun S. C. Retinoic acid embryopathy. N. Engl. J. Med. 1985; 313:837-841 [Medline] [Abstract]
• Leonard L., Horton C., Maden M., Pizzey J. A. Anteriorization of CRABP-I expression by retinoic acid in the developing mouse central nervous system and its relationship to teratogenesis. Dev. Biol. 1995; 168:514-528 [Medline] [Medline]
• Maden M., Gale E., Kostetskii I., Zile M. Vitamin A-deficient quail embryos have half a hindbrain and other neural defects. Current Biol. 1996; 6:417-426 [Medline] [Medline]
• Maden M., Graham A., Gale E., Zile M. Positional apoptosis during vertebrate CNS development in the absence of endogenous retinoids. Development 1997; 124:2799-2805 [Medline] [Abstract]
• Maden M., Holder N. Retinoic acid and development of the central nervous system. BioEssays 1992; 14:431-438 [Medline] [Medline]
• Marshall H., Nonchev S., Sham M. H., Muchamore I., Lumsden A., Krumlauf R. Retinoic acid alters hindbrain Hox code and induces transformation of rhombomeres 2/3 into a 4/5 identity. Nature (Lond.) 1992; 360:737-741 [Medline] [Medline]
• McGinnis W., Krumlauf R. Homeobox genes and axial patterning. Cell 1992; 68:283-302 [Medline] [Medline]
• Morriss-Kay G. M., Murphy P., Hill R. E., Davidson D. R. Effects of retinoic acid excess on expression of Hox-2.9 and Krox-20 and on morphological segementation in the hindbrain of mouse embryos. EMBO J 1991; 10:2985-2995 [Medline] [Medline]
• Rossant J., Zirngibl R., Cado D., Shago M., Giguere V. Expression of a retinoic acid response element-hsplacZ transgene defines specific domains of transcriptional activity during mouse embryogenesis. Genes & Dev. 1991; 5:1333-1344 [Medline] [Abstract/Free Full Text]
• Sidell N. Retinoic acid-induced growth inhibition and morphologic differentiation of human neuroblastoma cells in vitro. J. Natl. Cancer Inst. 1982; 68:589-596 [Medline]
• Thompson J. N., Howell J. M., Pitt G.A.J., McLaughlin C. I. The biological activity of retinoic acid in the domestic fowl and the effects of vitamin A deficiency on the chick embryo. Br. J. Nutr. 1969; 23:471-490 [Medline] [Medline]
• Wagner M., Han B., Jessell T. M. Regional differences in retinoid release from embryonic neural tissue detected by an in vitro reporter assay. Development 1992; 116:55-66 [Medline] [Abstract]
• Zile M. H. Vitamin A and embryonic development: an overview. J. Nutr. 1998; 128:455S-458S

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences