761
CONGENITAL MALFORMATIONS AND THE NEURAL CREST
JOHN CAMERON
JANET MCCREDIE
ROSE SHOOBRIDGE
University of Sydney, Sydney, New South
Wales 2006,
Australia
IN response
to comments on
the
hypothesisl-3
that
thalidomide embryopathy and similar congenital malformations are the expression of embryonic neuropathy neural-crest injury,4-6 we performed two experiments which might provide support for the hypothesis. It is widely believed that there are no nerves in the early limb bud of the vertebrate embryo. 1 2 7 This dictum seems to be based upon observations by Bardeen and Lewis8 (1901), who were unable to demonstrate nerve-fibres in limb buds of early embryos from the Carnegie collection, using methods of tissue preparation in use at the turn of the century. In view of the delicate structure of immature axons before myelination and the inadequacy of older methods of fixation and embedding for accurate neurohistological examination, we suspected that nerve axons, if present, would have been destroyed during tissue preparation. Thus, while we accepted the observations of Bardeen and Lewis as correct, we questioned the conclusion that the limb bud contains no nerve-fibres. or
FIRST EXPERIMENT
Our first
experiment was designed to examine this conclusion by the use of modern methods of tissue preparation for neurohistological examination, which cause minimal shrinkage of neural tissue. embryos were harvested by laparotomy at 290 hours’ gestation (12th day)-that is, during the thalidomide-sensitive period (7-14 days in rabbits) but before the appearance of carRabbit
Fig. 1-Longitudinal
Fig. 2-High-power view of nerve-fibres (NF) in limb bud (area arrowed in fig. 1) (x250). in the forelimb bud at 320 hours (13th day). Forelimb buds and adjacent neural tube were excised, fixed in glutaraldehyde, blocked in ’Araldite’, and sectioned in a plane either longitudinal or transverse to the axis of the limb bud. Light microscopy of longitudinal sections stained with toluidine blue revealed bundles of nerve-fibres entering the limb bud (figs. 1 and 2). The distal ramifications of these fibre bundles could be traced within the mesenchyme to a point approximately one-third of the way down the limb bud. The possibility of more distal branches could not be excluded, since they may have been too small to be seen by light microscopy. Electron microscopy of transverse sections through the proximal half of the forelimb buds showed aggregates of nerve axons, confirmed by the presence of neurotubules (fig. 3).
tilage
Discussion This evidence
counters the argument that the neuralfails because there are no nerves in the hypothesis limb bud. The early presence of nerve axons as the first differentiated tissue within the limb bud, at least highly 30 hours before cartilage formation, is contrary to the viewl that "proximodistal differentiation of the limbs consistently precedes the growth of nerves into the limbs". The presence of the nerve before any other differentiation at least suggests that it may play a role in limb growth, a possibility analogous to limb regeneration in newts, which has been shown to be nerve-depencrest
section of rabbit limb bud at 290 hours’
gestation: bundle of nerve fibres entering limb bud (x40) Arrows indicate major nerve trunk in proximal third of limb bud while mesenchyme is still undifferentiated.
3-Electron micrograph of transverse section of limb bud, showing neurotubles (x 15 000).
Fig.
nerve
in
762
"definite, well-developed groups of cells on each side of the spinal cord". Langman 10 says that "spinal nerves are
Fig. 4-Deformity resulting from ablation of left absence of distal leg and foot. dent.9 We do
not
not
to
neural crest:
believe that newts are a special case, this applicable argument.Many biologists regard the newt as an adult form which has retained an embryonic capacity to regenerate limbs, a capacity lost in higher vertebrates. It is not impossible that the neurotrophic mechanism underlying limb regeneration in newts might apply to limb formation in all vertebrate embryos, bearing in mind the common ancestry. Our experimental findings in rabbits agree with data from human embryological and clinical studies. Bardeen and Lewis8 described the inception of human limb buds as thickenings of the wolffian ridge during the second half of the third embryonic week, when spinal ganglia
penetrate into the limbs as soon as the buds are formed". Lewis11 sketched a nerve crossing the base line into a forelimb bud in a human embryo of about 4 weeks. Hamilton and Mossman 12 present a photomicrograph, similar to fig. 1, showing a large nerve trunk traversing the proximal half of the forelimb bud of a 10 mm human embryo (approximately 35 days’ gestation), with early mesenchymal condensation adjacent to the nerve. Thus, there is evidence that nerve axons are penetrating into the human forelimb bud between 21 and 35 days’ gestation. Clinical studies of thalidomide embryopathy showed that most of the major malformations of the arms resulted from drug ingestion between 24 and ’36 days’ gestation.13 The coincidence of this, "sensitive period" and the ingrowth of nerve fibres into undifferentiated mesenchyme provides further circumstantial evidence in favour of the neural-crest hypothesis. The "aneurigenic" limbs of amblystoma larvar have been cited2as evidence against the neural-crest hypothesis. The concept of aneurigenic limbs is, however, itself open to question, since Egar et al. 14 re-examined these "aneurigenic" limbs by electron microscopy and discovered nerve-fibres in over 60% of specimens. Our experiment and that of Egar et al. illustrate the need for some caution in accepting conclusions about unmyelinated nerve fibres, when such conclusions are based upon superseded methodology. This implies no criticism of past contributors, who worked within the technical limitations of their own era, as we do today. However, there is room for critical appraisal of old methods and old conclusions in the light of technical advances over the past 40 years. Some of the classical grafting experiments quoted by Stephensls can be viewed in this light and could usefully be repeated. Stephens 15 apparently has not appreciated some important features of neural-crest migration. Our own studies of neural crest agree with those of Weston 16 and Tennyson,17 who have observed that axonal outgrowths appear within a few hours of the onset of migration. Cytoplasm (including axons) is not labelled by tritiated thymidine, which is permanently incorporated into nuclear D.N.A.18 and marks only the nucleus for autoradiography. Weston states that these autoradiographic preparations are unsuitable for assessment of nerve-fibre outgrowth.’9 Autoradiographic absence of labelled crest nuclei in limb buds of stage 21-23 chick embryos, quoted by Stephens, does not exclude the presence of axons and cannot be interpreted to mean that the limb bud is devoid of neural-crest influence. SECOND EXPERIMENT
experiment designed to test whether injury to the crest produced limb malformations required surgical access to the early embryo at a stage before limbbud formation and before complete closure of neural tube, so that neural crest is still superficial, unfused, and An
neural
accessible at its caudal end. Fig. S-Residual mass of dorsal root (with standard errors). C=controls; AND=ablated, L=left.
not
ganglia
in
16-day
chicks
deformed; D=deformed; R=right;
The chick embryo at 48 hours’ incubation (stages 11-13 of Hamburger and Hamilton) fulfilled these requirements. A unilateral segment of caudal neural crest, at presumptive somite levels 26-30, was ablated by diathermy, and the eggs were returned to the incubator until day 16. Matched controls were
763 included for each stage of the ablation procedure. A total of 475 fertilised eggs of White Leghorn Austral strain were used. The operative mortality was high: of 143 embryos submitted to diathermy ablation, 17 survived. 10 chicks were not deformed, and 7 others had leg deformities, 2 bilaterally. 5 had similar unilateral deficiencies of the lower leg on the side of the neural-crest ablation (fig. 4). Analysis of residual tissue showed no measurable reduction in spinal cord or’mesodermal mass. However, dorsal-root ganglion mass was significantly reduced on the side of the ablation in the 5 chicks with unilateral deformities (fig. 5).
tributes to the pathogenesis of congenital limb malformations. The first experiment was supported by a grant from the Ramaciotti Foundations, New South Wales, and the second experiment by the N.H. and M.R.C. of Australia. We wish to acknowledge the help of Miss Virginia Best and of the departments of surgery, medicine, pathology, and biology of the University of Sydney. Details of both experiments are to be published in full elsewhere. Address for correspondence: J. M., Department of Surgery, University of Sydney, Sydney, N.S.W. 2006. Australia.
B
Discussion
REFERENCES
This experiment produced deformities in 41% of survivors. Reduction of dorsal-root ganglion mass on the side of the deformity is interpreted as evidence of neuralcrest damage. It should be emphasised that ablation of neural crest was carried out before limb buds had formed, and our results cannot be construed as neuralcrest depletion secondary to limb-bud damage.20 There was variable but minor damage to the unsegmented mesenchyme adjacent to the crest, the effect of which is being assessed at present. The ten chicks which survived diathermy ablation without limb deformity are thought to illustrate the known ability of neural crest to regulate and repair deficiencies. The two bilateral deformities may be due to diathermy current crossing the midline. Our results suggest that primary neural-crest injury con-
Preliminary
Communication
DIAGNOSIS OF PROTEIN-LOSING ENTEROPATHY BY GASTROINTESTINAL CLEARANCE OF ALPHA1-ANTITRYPSIN
J. J. BERNIER* CH. DESMAZURES†
CH. FLORENT* CH. AYMES† CH. L’HIRONDEL*
*
Clinique
de
Gastro-enterologie
and †Laboratoire de
Biochimie, Hôpital Saint-Lazare, 107 Rue du Faubourg Saint-Denis, 75010 Paris, France
Alpha1-antitrypsin was assessed, in 10 patients with protein-losing enteropathy and 13 control subjects, as an endogenous marker of plasma-protein loss into the gastrointestinal tract. Both
1. Gardner, E., O’Rahilly, R. Lancet, 1976, i, 635. 2. Wolpert, L. Br. med. Bull. 1976, 32, 1,65. 3. Poswillo, D. ibid. p. 59. 4. McCredie, J. Lancet, 1973, ii, 1058. 5. McCredie, J.Med.J.Aust. 1974, i, 159. 6. McCredie, J. J. neurol. Sci. 1976, 28, 373. 7. O’Rahilly, R., Gardner, E. Anat Embryol. 1975, 148, 1. 8. Bardeen, R. B., Lewis, W. H. Am. J. Anat. 1901, 1, 1. 9. Singer, exp. Zool. 1943, 92, 297.
M. J.
10. Langman, J. Medical Embryology. Baltimore, 1969. 11. Lewis, W. H. Am. J. Anat. 1902, 1, 145. 12. Hamilton, W. J., Mossman, H. W. Human Embryology. Cambridge, 1972. 13. Lenz, W., Knapp, K. Dt. med. Wschr. 1962, 87, 1232. 14. Egar, M., Yntema, C. L., Singer, M. ibid. 1973, 186, 91. 15. Stephens, T. D. Lancet, 1978, ii, 434. 16. Weston, J. A. Personal communication. 17. Tennyson, V. M.J. comp. Neurol. 1965,124, 267. 18. Weston, J. A. in Methods in Developmental Biology (edited by F. Wild and N. Wessels); p. 723. New York, 1967. 19. Weston, J. A. Devl Biol. 1963, 6, 279. 20. Stephens, T. D. Lancet, 1978, i, 282.
coprotein is not degraded by pancreatic enzymes and is readily detectable in the faeces of newborn infants, children, and adults.l-3 We have tried o-A.T. as an endogenous marker of protein loss. PATIENTS AND METHODS
23 adult inpatients were studied. 13 had no apparent digestive disease and acted as controls. 10 fulfilled the criteria of protein-losing enteropathy: either they had excessive loss of intravenously administered radiolabelled chromium chloride into the fxces or they had evidence of active chronic inflammatory bowel disease (see table). Blood-samples were taken on days 1, 5, and 10 and all stools were collected for 10 days. The clearance of serum «1-A.T. by the gastrointestinal tract was determined for each day by the formula FxV C=———— p
Summary
where F is fsecal o-A.T.
PATIENTS WITH PROTEIN-LOSING ENTEROPATHY
fæcal &agr;1-A.T. concentrations and fæcal &agr;1-A.T. clearance were significantly higher in patients than in controls. With clearance there was no overlap between the groups. Over 10 days the normal gastrointestinal clearance of &agr;1-A.T. was 3·07±2·25 (S.D.) ml/day. Measurement of &agr;1-A.T. clearance is easy and requires no radio-
isotopes. INTRODUCTION
of plasma-protein can be measerum-protein, her-labelled albumin, 67Cu-labelled cxruloplasmin, and other labelled macromolecules. These methods are expensive, imprecise, and cumbersome. Alphal-antitrypsin (o-A.T.), a protease inhibitor synthesised by the liver, is present in serum at a concentration of 2-5 g/1. This gly-
concentration, V is daily faecal volume,
GASTROINTESTINAL loss sured with ’31I-labelled
N.D.=not done.
CONGENITAL MALFORMATIONS AND THE NEURAL CREST
JOHN CAMERON
JANET MCCREDIE
ROSE SHOOBRIDGE
University of Sydney, Sydney, New South
Wales 2006,
Australia
IN response
to comments on
the
hypothesisl-3
that
thalidomide embryopathy and similar congenital malformations are the expression of embryonic neuropathy neural-crest injury,4-6 we performed two experiments which might provide support for the hypothesis. It is widely believed that there are no nerves in the early limb bud of the vertebrate embryo. 1 2 7 This dictum seems to be based upon observations by Bardeen and Lewis8 (1901), who were unable to demonstrate nerve-fibres in limb buds of early embryos from the Carnegie collection, using methods of tissue preparation in use at the turn of the century. In view of the delicate structure of immature axons before myelination and the inadequacy of older methods of fixation and embedding for accurate neurohistological examination, we suspected that nerve axons, if present, would have been destroyed during tissue preparation. Thus, while we accepted the observations of Bardeen and Lewis as correct, we questioned the conclusion that the limb bud contains no nerve-fibres. or
FIRST EXPERIMENT
Our first
experiment was designed to examine this conclusion by the use of modern methods of tissue preparation for neurohistological examination, which cause minimal shrinkage of neural tissue. embryos were harvested by laparotomy at 290 hours’ gestation (12th day)-that is, during the thalidomide-sensitive period (7-14 days in rabbits) but before the appearance of carRabbit
Fig. 1-Longitudinal
Fig. 2-High-power view of nerve-fibres (NF) in limb bud (area arrowed in fig. 1) (x250). in the forelimb bud at 320 hours (13th day). Forelimb buds and adjacent neural tube were excised, fixed in glutaraldehyde, blocked in ’Araldite’, and sectioned in a plane either longitudinal or transverse to the axis of the limb bud. Light microscopy of longitudinal sections stained with toluidine blue revealed bundles of nerve-fibres entering the limb bud (figs. 1 and 2). The distal ramifications of these fibre bundles could be traced within the mesenchyme to a point approximately one-third of the way down the limb bud. The possibility of more distal branches could not be excluded, since they may have been too small to be seen by light microscopy. Electron microscopy of transverse sections through the proximal half of the forelimb buds showed aggregates of nerve axons, confirmed by the presence of neurotubules (fig. 3).
tilage
Discussion This evidence
counters the argument that the neuralfails because there are no nerves in the hypothesis limb bud. The early presence of nerve axons as the first differentiated tissue within the limb bud, at least highly 30 hours before cartilage formation, is contrary to the viewl that "proximodistal differentiation of the limbs consistently precedes the growth of nerves into the limbs". The presence of the nerve before any other differentiation at least suggests that it may play a role in limb growth, a possibility analogous to limb regeneration in newts, which has been shown to be nerve-depencrest
section of rabbit limb bud at 290 hours’
gestation: bundle of nerve fibres entering limb bud (x40) Arrows indicate major nerve trunk in proximal third of limb bud while mesenchyme is still undifferentiated.
3-Electron micrograph of transverse section of limb bud, showing neurotubles (x 15 000).
Fig.
nerve
in
762
"definite, well-developed groups of cells on each side of the spinal cord". Langman 10 says that "spinal nerves are
Fig. 4-Deformity resulting from ablation of left absence of distal leg and foot. dent.9 We do
not
not
to
neural crest:
believe that newts are a special case, this applicable argument.Many biologists regard the newt as an adult form which has retained an embryonic capacity to regenerate limbs, a capacity lost in higher vertebrates. It is not impossible that the neurotrophic mechanism underlying limb regeneration in newts might apply to limb formation in all vertebrate embryos, bearing in mind the common ancestry. Our experimental findings in rabbits agree with data from human embryological and clinical studies. Bardeen and Lewis8 described the inception of human limb buds as thickenings of the wolffian ridge during the second half of the third embryonic week, when spinal ganglia
penetrate into the limbs as soon as the buds are formed". Lewis11 sketched a nerve crossing the base line into a forelimb bud in a human embryo of about 4 weeks. Hamilton and Mossman 12 present a photomicrograph, similar to fig. 1, showing a large nerve trunk traversing the proximal half of the forelimb bud of a 10 mm human embryo (approximately 35 days’ gestation), with early mesenchymal condensation adjacent to the nerve. Thus, there is evidence that nerve axons are penetrating into the human forelimb bud between 21 and 35 days’ gestation. Clinical studies of thalidomide embryopathy showed that most of the major malformations of the arms resulted from drug ingestion between 24 and ’36 days’ gestation.13 The coincidence of this, "sensitive period" and the ingrowth of nerve fibres into undifferentiated mesenchyme provides further circumstantial evidence in favour of the neural-crest hypothesis. The "aneurigenic" limbs of amblystoma larvar have been cited2as evidence against the neural-crest hypothesis. The concept of aneurigenic limbs is, however, itself open to question, since Egar et al. 14 re-examined these "aneurigenic" limbs by electron microscopy and discovered nerve-fibres in over 60% of specimens. Our experiment and that of Egar et al. illustrate the need for some caution in accepting conclusions about unmyelinated nerve fibres, when such conclusions are based upon superseded methodology. This implies no criticism of past contributors, who worked within the technical limitations of their own era, as we do today. However, there is room for critical appraisal of old methods and old conclusions in the light of technical advances over the past 40 years. Some of the classical grafting experiments quoted by Stephensls can be viewed in this light and could usefully be repeated. Stephens 15 apparently has not appreciated some important features of neural-crest migration. Our own studies of neural crest agree with those of Weston 16 and Tennyson,17 who have observed that axonal outgrowths appear within a few hours of the onset of migration. Cytoplasm (including axons) is not labelled by tritiated thymidine, which is permanently incorporated into nuclear D.N.A.18 and marks only the nucleus for autoradiography. Weston states that these autoradiographic preparations are unsuitable for assessment of nerve-fibre outgrowth.’9 Autoradiographic absence of labelled crest nuclei in limb buds of stage 21-23 chick embryos, quoted by Stephens, does not exclude the presence of axons and cannot be interpreted to mean that the limb bud is devoid of neural-crest influence. SECOND EXPERIMENT
experiment designed to test whether injury to the crest produced limb malformations required surgical access to the early embryo at a stage before limbbud formation and before complete closure of neural tube, so that neural crest is still superficial, unfused, and An
neural
accessible at its caudal end. Fig. S-Residual mass of dorsal root (with standard errors). C=controls; AND=ablated, L=left.
not
ganglia
in
16-day
chicks
deformed; D=deformed; R=right;
The chick embryo at 48 hours’ incubation (stages 11-13 of Hamburger and Hamilton) fulfilled these requirements. A unilateral segment of caudal neural crest, at presumptive somite levels 26-30, was ablated by diathermy, and the eggs were returned to the incubator until day 16. Matched controls were
763 included for each stage of the ablation procedure. A total of 475 fertilised eggs of White Leghorn Austral strain were used. The operative mortality was high: of 143 embryos submitted to diathermy ablation, 17 survived. 10 chicks were not deformed, and 7 others had leg deformities, 2 bilaterally. 5 had similar unilateral deficiencies of the lower leg on the side of the neural-crest ablation (fig. 4). Analysis of residual tissue showed no measurable reduction in spinal cord or’mesodermal mass. However, dorsal-root ganglion mass was significantly reduced on the side of the ablation in the 5 chicks with unilateral deformities (fig. 5).
tributes to the pathogenesis of congenital limb malformations. The first experiment was supported by a grant from the Ramaciotti Foundations, New South Wales, and the second experiment by the N.H. and M.R.C. of Australia. We wish to acknowledge the help of Miss Virginia Best and of the departments of surgery, medicine, pathology, and biology of the University of Sydney. Details of both experiments are to be published in full elsewhere. Address for correspondence: J. M., Department of Surgery, University of Sydney, Sydney, N.S.W. 2006. Australia.
B
Discussion
REFERENCES
This experiment produced deformities in 41% of survivors. Reduction of dorsal-root ganglion mass on the side of the deformity is interpreted as evidence of neuralcrest damage. It should be emphasised that ablation of neural crest was carried out before limb buds had formed, and our results cannot be construed as neuralcrest depletion secondary to limb-bud damage.20 There was variable but minor damage to the unsegmented mesenchyme adjacent to the crest, the effect of which is being assessed at present. The ten chicks which survived diathermy ablation without limb deformity are thought to illustrate the known ability of neural crest to regulate and repair deficiencies. The two bilateral deformities may be due to diathermy current crossing the midline. Our results suggest that primary neural-crest injury con-
Preliminary
Communication
DIAGNOSIS OF PROTEIN-LOSING ENTEROPATHY BY GASTROINTESTINAL CLEARANCE OF ALPHA1-ANTITRYPSIN
J. J. BERNIER* CH. DESMAZURES†
CH. FLORENT* CH. AYMES† CH. L’HIRONDEL*
*
Clinique
de
Gastro-enterologie
and †Laboratoire de
Biochimie, Hôpital Saint-Lazare, 107 Rue du Faubourg Saint-Denis, 75010 Paris, France
Alpha1-antitrypsin was assessed, in 10 patients with protein-losing enteropathy and 13 control subjects, as an endogenous marker of plasma-protein loss into the gastrointestinal tract. Both
1. Gardner, E., O’Rahilly, R. Lancet, 1976, i, 635. 2. Wolpert, L. Br. med. Bull. 1976, 32, 1,65. 3. Poswillo, D. ibid. p. 59. 4. McCredie, J. Lancet, 1973, ii, 1058. 5. McCredie, J.Med.J.Aust. 1974, i, 159. 6. McCredie, J. J. neurol. Sci. 1976, 28, 373. 7. O’Rahilly, R., Gardner, E. Anat Embryol. 1975, 148, 1. 8. Bardeen, R. B., Lewis, W. H. Am. J. Anat. 1901, 1, 1. 9. Singer, exp. Zool. 1943, 92, 297.
M. J.
10. Langman, J. Medical Embryology. Baltimore, 1969. 11. Lewis, W. H. Am. J. Anat. 1902, 1, 145. 12. Hamilton, W. J., Mossman, H. W. Human Embryology. Cambridge, 1972. 13. Lenz, W., Knapp, K. Dt. med. Wschr. 1962, 87, 1232. 14. Egar, M., Yntema, C. L., Singer, M. ibid. 1973, 186, 91. 15. Stephens, T. D. Lancet, 1978, ii, 434. 16. Weston, J. A. Personal communication. 17. Tennyson, V. M.J. comp. Neurol. 1965,124, 267. 18. Weston, J. A. in Methods in Developmental Biology (edited by F. Wild and N. Wessels); p. 723. New York, 1967. 19. Weston, J. A. Devl Biol. 1963, 6, 279. 20. Stephens, T. D. Lancet, 1978, i, 282.
coprotein is not degraded by pancreatic enzymes and is readily detectable in the faeces of newborn infants, children, and adults.l-3 We have tried o-A.T. as an endogenous marker of protein loss. PATIENTS AND METHODS
23 adult inpatients were studied. 13 had no apparent digestive disease and acted as controls. 10 fulfilled the criteria of protein-losing enteropathy: either they had excessive loss of intravenously administered radiolabelled chromium chloride into the fxces or they had evidence of active chronic inflammatory bowel disease (see table). Blood-samples were taken on days 1, 5, and 10 and all stools were collected for 10 days. The clearance of serum «1-A.T. by the gastrointestinal tract was determined for each day by the formula FxV C=———— p
Summary
where F is fsecal o-A.T.
PATIENTS WITH PROTEIN-LOSING ENTEROPATHY
fæcal &agr;1-A.T. concentrations and fæcal &agr;1-A.T. clearance were significantly higher in patients than in controls. With clearance there was no overlap between the groups. Over 10 days the normal gastrointestinal clearance of &agr;1-A.T. was 3·07±2·25 (S.D.) ml/day. Measurement of &agr;1-A.T. clearance is easy and requires no radio-
isotopes. INTRODUCTION
of plasma-protein can be measerum-protein, her-labelled albumin, 67Cu-labelled cxruloplasmin, and other labelled macromolecules. These methods are expensive, imprecise, and cumbersome. Alphal-antitrypsin (o-A.T.), a protease inhibitor synthesised by the liver, is present in serum at a concentration of 2-5 g/1. This gly-
concentration, V is daily faecal volume,
GASTROINTESTINAL loss sured with ’31I-labelled
N.D.=not done.
