Heterotopic Dorsal-root Ganglion Cells Around the Spinal Cord in Children with Spina Bifida Aperta R. G. LENDON” J. L. E M E R Y t Introduction There have been isolated reports of heterotopic ganglion cells associated with cord lesions in children with myelomeningocele. In 1885, Marsh and colleagues described “ganglia aberrantia” within the myelomeningocele sac and large groups of heterotopic dorsal-root ganglion cells. Altschul (1929) described ectopic ganglion cells in a case of myelomeningocele with diplomyelia in the thoracic region; and Gagel (1936) also encountered a case with thoracic diplomyelia accompanied by a small ganglion posterior to the double cord. Lichtenstein (1941) illustrated a cord with ectopic ganglia associated with double central canals. A 20mm embryo investigated by Gruenwald (1941) had partial duplication of the cord cranial and caudal to the plaque, associated with heterotopic groups of ganglion cells in relation to the cord duplications. During our previous studies (Lendon 1968, Emery and Lendon 1973) we saw many groups of heterotopic ganglion cells and we now report their relative frequency at different levels of the cord and their relationship to the main types of cord lesion. Material and Methods Necropsy material came from 95 chiIdren with spina bifida aperta. 91 of these had been included previously in a study of cord lesions (Emery and Lendon 1973) and details of the necropsy technique are recorded in that paper. Serial paraffin sections were made of 25 cords and in the remaining 70 at least one ribbon of sections was cut from each neurological spinal segment. Sections were stained with gallocyanin, haematoxylin and eosin, or Masson’s trichrome. Each 50th section of the serially sectioned cords, and each slide from the remaining cords, was scanned microscopically and the presence or absence of heterotopic ganglion cells recorded, along with details of their position relative to any cord lesion. Twenty cords from apparently normal children were also sectioned, 10 serially and 10 taking at least one ribbon per segment, and were studied in the same way as the abnormal cords, centred on the lumbosacral region. Results Sixty-three of 95 cords from children with spina bifida aperta had heterotopic ganglion cells in nerve roots adjacent to their cords. They were usually located in dorsal nerve roots, either posterior, posterolateral or lateral to the cord (Figs. 1, 2). Rarely, heterotopic cells were found ventral to the cord (Fig. 3) or in the ventral nerve roots close to their point of exit from the cord or, in cases with diplomyelia, between the two hemicords. *Department of Anatomy, University of Manchester. ?Department of Pathology, Sheffield Children’s Hospital.
16
1. L. EMERY
11. G. LI N I N N
The heterotopic cells had an irregular distribution along the neuroaxis: in some cases they were found at only one level and in others at several levels. The highest concentration of cells was found in the upper sacral region (Fig. 4). The position of the heterotopic cells varied slightly from level to level: they tended to be more frequent lateral to the cord in the thoracic region and at S4, and posterior to the cord in the upper sacral region. Of the 20 normal cords, five were accompanied by heterotopic ganglion cells which were only found at a single neurological level-at L5 in two cases and at L I , L2 and S3 respectively in the remainder. Groups of one to three ectopic cells were found close to the point of entry of the dorsal nerve root into the cord in four of these five cords, and in the root lateral to the cord in one case. We have not recorded the number of heterotopic cells, but it was very clear that when they occurred they were more numerous in the spina bifida child than in the apparently normal child. The relationship between heterotopic ganglion cells and the principal primary cord lesions is shown in Table I. In 32 of the cords the main lesion cranial to, at, and below the plaque was complete or partial duplication of cord structure. In 30 of these there were heterotopic ganglion cells at one or more levels.
Fig. 1 (right). Mass of dorsal-root ganglion tissue lying dorsal o a cord at SI. (H & E 25.) Fig. 2 (luwar lrft). G r o ~ i pof ganglion cells lying lateral to spinal cord in a dorsal root. ( H & E > 100.) Fig. 3 ( l u w r right). Group of dorsal root ganglion cells lying ventral to spinal cord. ( H &
E
40.)
17
DEVELOPMENTAL MEDICINE AND CHILD NEUROLOGY.
1976, 18. Supp. 37
.
26
.. .
24
cn
i 22 Y
0 20
;
.
18
0
6
16
C
.
14
w
= 12
. ...
m . 8
i K 6
z 4 2
.. C B l i
2
3
4
5
6
7
8
9
1011112
LI
2
3
4
L5
si
2
3
4
SPINAL CORD LEVEL
Fig. 4. Chart showing distribution of heterotopic ganglion cells related to neuroaxis.
Forty-two cords exhibited an essentially flat plate of neural tissue at the plaque (myelocele). In eight of these there was some form of duplication caudal to the myelocele, while in 24 the cord was single in all cytoarchitectonic features. Cords with a tendency to duplication below a myelocele showed heterotopic cells in this region more often than those without this feature (Table I). The fourth category of lesions shown in Table I is a heterogeneous group described previously as ‘slit’ canal, ‘winged’ cords and syringomyeloceles (Emery and Lendon 1973). Of the 21 cords in this group, nine had heterotopic ganglion cells and in six these were found below the plaque where partial duplication of the cord was present. In 12 of 95 cords heterotopic ganglion cells were found cranial to either diplomyelia or a plaque lesion. In 10, the cells occurred several segments cranial to the major primary cord lesion adjacent t o a more-or-less normal cord, or one with hydromyelia. No heterotopic ganglion cells were seen at any level, as far as we could judge using our sampling method, in 32 of the 95 abnormal cords.
Discussion One of the major controversies concerning the embryogenesis of the cord lesions in spina bifida aperta is whether these result from primary non-closure of the neural plate or from secondary bursting open of the neural tube. During normal development, at the time the neural tube closes the neural crest is given off from the edges of the neural plate and migrates laterally and ventrally to form, amongst other things, dorsal root and autonomic ganglia. If the cord lesions were due to abnormal neurulation, one might expect anomalies to be found in derivatives of the neural crest. Thus, our observations on the distribution of heterotopic dorsal-root ganglion cells may be pertinent to this controversy. Heterotopic dorsal-root ganglion cells were found in nerve roots posterior to, and lateral to, the spinal cord in the great majority of children with spina bifida aperta. These cells were rare cranial to a major cord lesion (such as diplomyelia or a myelocele) and were most common when the cord was duplicated completely or had a tendency towards internal
R. G. L I N D O N
J. L. EMERY DIPLOMYELIA
NORMAL
N
I
-
I
0-
Hrteroroyic ~ ( j t i ~
Type of lesion Present
Diploinyelia Myelocele, duplex below Myelocele, single below Others I
I
Ahwrrt
30
7
14
4
10 9
14 12
Notochord
4 t
I
Fig. 5. Diagram showing possible mode of genesis of duplication of spinal cord with heterotopic dorsal-root ganglion tissue. (N = neural crest; HN = heterotopic neural crest.)
duplication (of the dorsal horns and, or central canal). This association between complete or partial duplication and heterotopic ganglion cells was most obvious in the upper sacral region, where over 20 per cent of the cords exhibited heterotopic ganglion cells at some level. The genesis of heterotopic cells in association with cord duplication can be explained, we believe, by a mechanism similar to that shown in Fig. 5. If one postulates that diplomyelia results from delayed, faulty closure of the neural tube, then an accompanying delay in migration of the neural crest might result, giving a combination of diplomyelia with heterotopic ganglion cells posterior to, or lateral to, the cord. If this is accepted, then the presence of heterotopic ganglion cells above the cord is not readily explained by theories which postulate bursting open of a previously closed tube ( e . g . Gardner 1966), because the neural crest would have migrated away before re-opening takes place. The presence of heterotopic cells i n nerve roots lateral to the myelocele type of lesion can also be interpreted in a similar fashion if it is assumed that a myelocele results from non-closure of a neural plate, with delayed neural crest migration. The occasional presence of cells anterior to the cord would be due to abnormal migration of neural crest cells around and beneath the neural tube. This interpretation was used by Sosa and Andrew (1947) to explain the occasional presence of ganglimi cells in the ventral ncrve roots of normal cords. They also proposed a similar theory to explain the presence of dorsal-root ganglion cells in the medial portion of the dorsal root in normal cords. The incidence of such heterotopic cells in normal cases is not known. Our observations on 20 cords suggest that they might occur in the lumbosacral region in up to 10 per cent of children at any one level; however, they appear to be less frequent than in the spina-bifida child. A further question is why larger numbers of heterotopic cells occur in the upper sacral region. The theory we have proposed to explain the presence of these cells (Fig. 5 ) is based on the assumption that a classical process of neurulation takes place, and indeed this could 19
DEVELOPMENTAL MEDICINE AND CHILD NEUROLOGY.
1976, 18. Supp. 37
explain the genesis of these cells in the thoracic and upper lumbar regions. However, the embryogenesis of the lumbosacral region in man is still in doubt. Holmdahl(l925) suggested that the neural tube in the lower trunk forms by canalisation of solid tail-bud tissue. If Holmdahl’s theory is correct, the dorsal-root ganglion cells presumably must develop from the dorsal part of the tail-bud tissue. Alternatively, perhaps the sacral ganglia come from cells which migrate out from the dorsal part of the neural tube, as was proposed by Orts Llorca (1934). Regardless of whether the lumbosacral cord forms by neurulation or canalisation, it appears from our observations that there is a marked tendency for heterotopic ganglion-cell formation in this region in the spina-bifida child, particularly when there is cord duplication. The high incidence of these cells may best be attributed to anomalous neural crest migration during abnormal neurulation, and could even represent an increased variability which is present in small degrees in normal individuals. SUM MARY
Sixty-three of 95 cords from children with spina bifida aperta had associated heterotopic ganglion cells, usually located in dorsal nerve roots posterior or lateral to the cord. Heterotopic cells were more common in cords with complete or partial diplomyelia, and the largest number was found in the upper sacral region. It is proposed that the heterotopic position of the cells is due to abnormal and delayed migration of neural crest cells during primary neurogenesis. The findings lend support to the concept that in children with myelomeningocele the cord lesion is not due to secondary rupture. RESUME
Cellules hktkrotopiques du ganglion des racines postkrieures dans la nioelle Ppini2re chez Ies enfants avec spina bifida aperta Les moelles tpiniires de 95 enfants prksentant un spina bifida aperta ont Cte examinees et il a Ctt trouvC dans 63 cas des cellules ganglionaires hetirotopiques. Ces cellules ttaient habituellement IocalisCes dans les racines nerveuses dorsales posttrieures ou laterales par rapport a la moelle et Ctaient plus communes dans les moelles avec diplomytlie complete ou partielle. Le plus grand nombre de cellules heterotopiques a t t t trouvri dans la region sacree suptrieure. II est suggCrC que la position hCtCrotopique des cellules est dCie a une migration anormale et retardte des cellules des cr&tes neurales durant la neurogtnese primaire. Ces decouvertes sont en faveur du fait que la ltsion mtdullaire dans le mytlom6ningocile n’est pas dfie une rupture secondaire. ZUSAMMENFASSUNG
Heterotop urn das Riickenmark gelegene Ganglionzellen der Hinterwurzeln bei Kindern init Spina bijida aperta Von 95 Kindern mit Spina bifida aperta wurde das Riickenmark untersucht, davon hatten 63 angrenzende heterotope Ganglienzellen. Diese Zellen waren gewohnlich an den Hinterstrangen posterior und lateral zum Mark angeordnet und sie wurden haufiger beim Ruckenmark mit kompletter oder partieller Diplomyelie beobachtet. Die groDte Anzahl heterotoper Zellpositionen fand sich im unteren Sacralbereich. Es wird angenommen, daD die heterotopen Zellpositionen durch abnorme oder verzogerte Migration der Zellen der Neuralleiste wahrend der primaren Neurogenese bedingt sind. Diese Befunde unterstutzen die Vorstellung, daD die Marklasion bei der Myelomeningocele nicht durch sekundare Ruptur entsteht. 20
I
16
1. L. EMERY
11. G. LI N I N N
The heterotopic cells had an irregular distribution along the neuroaxis: in some cases they were found at only one level and in others at several levels. The highest concentration of cells was found in the upper sacral region (Fig. 4). The position of the heterotopic cells varied slightly from level to level: they tended to be more frequent lateral to the cord in the thoracic region and at S4, and posterior to the cord in the upper sacral region. Of the 20 normal cords, five were accompanied by heterotopic ganglion cells which were only found at a single neurological level-at L5 in two cases and at L I , L2 and S3 respectively in the remainder. Groups of one to three ectopic cells were found close to the point of entry of the dorsal nerve root into the cord in four of these five cords, and in the root lateral to the cord in one case. We have not recorded the number of heterotopic cells, but it was very clear that when they occurred they were more numerous in the spina bifida child than in the apparently normal child. The relationship between heterotopic ganglion cells and the principal primary cord lesions is shown in Table I. In 32 of the cords the main lesion cranial to, at, and below the plaque was complete or partial duplication of cord structure. In 30 of these there were heterotopic ganglion cells at one or more levels.
Fig. 1 (right). Mass of dorsal-root ganglion tissue lying dorsal o a cord at SI. (H & E 25.) Fig. 2 (luwar lrft). G r o ~ i pof ganglion cells lying lateral to spinal cord in a dorsal root. ( H & E > 100.) Fig. 3 ( l u w r right). Group of dorsal root ganglion cells lying ventral to spinal cord. ( H &
E
40.)
17
DEVELOPMENTAL MEDICINE AND CHILD NEUROLOGY.
1976, 18. Supp. 37
.
26
.. .
24
cn
i 22 Y
0 20
;
.
18
0
6
16
C
.
14
w
= 12
. ...
m . 8
i K 6
z 4 2
.. C B l i
2
3
4
5
6
7
8
9
1011112
LI
2
3
4
L5
si
2
3
4
SPINAL CORD LEVEL
Fig. 4. Chart showing distribution of heterotopic ganglion cells related to neuroaxis.
Forty-two cords exhibited an essentially flat plate of neural tissue at the plaque (myelocele). In eight of these there was some form of duplication caudal to the myelocele, while in 24 the cord was single in all cytoarchitectonic features. Cords with a tendency to duplication below a myelocele showed heterotopic cells in this region more often than those without this feature (Table I). The fourth category of lesions shown in Table I is a heterogeneous group described previously as ‘slit’ canal, ‘winged’ cords and syringomyeloceles (Emery and Lendon 1973). Of the 21 cords in this group, nine had heterotopic ganglion cells and in six these were found below the plaque where partial duplication of the cord was present. In 12 of 95 cords heterotopic ganglion cells were found cranial to either diplomyelia or a plaque lesion. In 10, the cells occurred several segments cranial to the major primary cord lesion adjacent t o a more-or-less normal cord, or one with hydromyelia. No heterotopic ganglion cells were seen at any level, as far as we could judge using our sampling method, in 32 of the 95 abnormal cords.
Discussion One of the major controversies concerning the embryogenesis of the cord lesions in spina bifida aperta is whether these result from primary non-closure of the neural plate or from secondary bursting open of the neural tube. During normal development, at the time the neural tube closes the neural crest is given off from the edges of the neural plate and migrates laterally and ventrally to form, amongst other things, dorsal root and autonomic ganglia. If the cord lesions were due to abnormal neurulation, one might expect anomalies to be found in derivatives of the neural crest. Thus, our observations on the distribution of heterotopic dorsal-root ganglion cells may be pertinent to this controversy. Heterotopic dorsal-root ganglion cells were found in nerve roots posterior to, and lateral to, the spinal cord in the great majority of children with spina bifida aperta. These cells were rare cranial to a major cord lesion (such as diplomyelia or a myelocele) and were most common when the cord was duplicated completely or had a tendency towards internal
R. G. L I N D O N
J. L. EMERY DIPLOMYELIA
NORMAL
N
I
-
I
0-
Hrteroroyic ~ ( j t i ~
Type of lesion Present
Diploinyelia Myelocele, duplex below Myelocele, single below Others I
I
Ahwrrt
30
7
14
4
10 9
14 12
Notochord
4 t
I
Fig. 5. Diagram showing possible mode of genesis of duplication of spinal cord with heterotopic dorsal-root ganglion tissue. (N = neural crest; HN = heterotopic neural crest.)
duplication (of the dorsal horns and, or central canal). This association between complete or partial duplication and heterotopic ganglion cells was most obvious in the upper sacral region, where over 20 per cent of the cords exhibited heterotopic ganglion cells at some level. The genesis of heterotopic cells in association with cord duplication can be explained, we believe, by a mechanism similar to that shown in Fig. 5. If one postulates that diplomyelia results from delayed, faulty closure of the neural tube, then an accompanying delay in migration of the neural crest might result, giving a combination of diplomyelia with heterotopic ganglion cells posterior to, or lateral to, the cord. If this is accepted, then the presence of heterotopic ganglion cells above the cord is not readily explained by theories which postulate bursting open of a previously closed tube ( e . g . Gardner 1966), because the neural crest would have migrated away before re-opening takes place. The presence of heterotopic cells i n nerve roots lateral to the myelocele type of lesion can also be interpreted in a similar fashion if it is assumed that a myelocele results from non-closure of a neural plate, with delayed neural crest migration. The occasional presence of cells anterior to the cord would be due to abnormal migration of neural crest cells around and beneath the neural tube. This interpretation was used by Sosa and Andrew (1947) to explain the occasional presence of ganglimi cells in the ventral ncrve roots of normal cords. They also proposed a similar theory to explain the presence of dorsal-root ganglion cells in the medial portion of the dorsal root in normal cords. The incidence of such heterotopic cells in normal cases is not known. Our observations on 20 cords suggest that they might occur in the lumbosacral region in up to 10 per cent of children at any one level; however, they appear to be less frequent than in the spina-bifida child. A further question is why larger numbers of heterotopic cells occur in the upper sacral region. The theory we have proposed to explain the presence of these cells (Fig. 5 ) is based on the assumption that a classical process of neurulation takes place, and indeed this could 19
DEVELOPMENTAL MEDICINE AND CHILD NEUROLOGY.
1976, 18. Supp. 37
explain the genesis of these cells in the thoracic and upper lumbar regions. However, the embryogenesis of the lumbosacral region in man is still in doubt. Holmdahl(l925) suggested that the neural tube in the lower trunk forms by canalisation of solid tail-bud tissue. If Holmdahl’s theory is correct, the dorsal-root ganglion cells presumably must develop from the dorsal part of the tail-bud tissue. Alternatively, perhaps the sacral ganglia come from cells which migrate out from the dorsal part of the neural tube, as was proposed by Orts Llorca (1934). Regardless of whether the lumbosacral cord forms by neurulation or canalisation, it appears from our observations that there is a marked tendency for heterotopic ganglion-cell formation in this region in the spina-bifida child, particularly when there is cord duplication. The high incidence of these cells may best be attributed to anomalous neural crest migration during abnormal neurulation, and could even represent an increased variability which is present in small degrees in normal individuals. SUM MARY
Sixty-three of 95 cords from children with spina bifida aperta had associated heterotopic ganglion cells, usually located in dorsal nerve roots posterior or lateral to the cord. Heterotopic cells were more common in cords with complete or partial diplomyelia, and the largest number was found in the upper sacral region. It is proposed that the heterotopic position of the cells is due to abnormal and delayed migration of neural crest cells during primary neurogenesis. The findings lend support to the concept that in children with myelomeningocele the cord lesion is not due to secondary rupture. RESUME
Cellules hktkrotopiques du ganglion des racines postkrieures dans la nioelle Ppini2re chez Ies enfants avec spina bifida aperta Les moelles tpiniires de 95 enfants prksentant un spina bifida aperta ont Cte examinees et il a Ctt trouvC dans 63 cas des cellules ganglionaires hetirotopiques. Ces cellules ttaient habituellement IocalisCes dans les racines nerveuses dorsales posttrieures ou laterales par rapport a la moelle et Ctaient plus communes dans les moelles avec diplomytlie complete ou partielle. Le plus grand nombre de cellules heterotopiques a t t t trouvri dans la region sacree suptrieure. II est suggCrC que la position hCtCrotopique des cellules est dCie a une migration anormale et retardte des cellules des cr&tes neurales durant la neurogtnese primaire. Ces decouvertes sont en faveur du fait que la ltsion mtdullaire dans le mytlom6ningocile n’est pas dfie une rupture secondaire. ZUSAMMENFASSUNG
Heterotop urn das Riickenmark gelegene Ganglionzellen der Hinterwurzeln bei Kindern init Spina bijida aperta Von 95 Kindern mit Spina bifida aperta wurde das Riickenmark untersucht, davon hatten 63 angrenzende heterotope Ganglienzellen. Diese Zellen waren gewohnlich an den Hinterstrangen posterior und lateral zum Mark angeordnet und sie wurden haufiger beim Ruckenmark mit kompletter oder partieller Diplomyelie beobachtet. Die groDte Anzahl heterotoper Zellpositionen fand sich im unteren Sacralbereich. Es wird angenommen, daD die heterotopen Zellpositionen durch abnorme oder verzogerte Migration der Zellen der Neuralleiste wahrend der primaren Neurogenese bedingt sind. Diese Befunde unterstutzen die Vorstellung, daD die Marklasion bei der Myelomeningocele nicht durch sekundare Ruptur entsteht. 20
I
