264 Nature Vol.

284 20 March 1980

13. Stouffer, R. L., Nixon, W. E .. Gulyas, B. J. & Hodgen, G. D. Endocrinology 100,506-511 avulsed dorsal root ganglia contained a mean of 5,850 myel-
(1977).
14. Williams, M. T., Roth, M. S., Marsh, J. M. & LeMaire, W. J. J. c/in. Endocr. Metab. 48, inated axons (s.e. = 510, n = 6); those in rats with intact spinal
437-440 (1979). roots and ganglia contained a mean of 10,950 myelinated axons
IS. Bernard, J. J. Reprod. Fert. 43,453-460 (1975). (s.e. = 1,660, n = 5). These results suggest that dorsal and
16. Goldenberg, R. L., Bridson, W. E. & Kohler, P. 0. Biochem. biophys. Res. Commun. 48,
101-108 (1972). ventral roots contributed to the innervation of the grafts but
17. Thanki, K. H. & Channing, C. P. Endocrinology 103, 74-80 (1978). were not the sole source of axons. They do not exclude the
18. Rolland, R., Gunsalus, G. L. & Hammond, J. M. Endocrinology 98, 1083-1091 (1976).
19. McNatty, K. P .. Hunter, W., McNeilly, A. S. & Sawers, R. S. J. Endocr. 64, 555-571 (1975). possibility that the grafts were entirely innervated by axons
20. Eiler, H. & Nalbandov, A. V. Endocrinology 100, 331-338 (1977). originating from dorsal root ganglia and motor neurones 12 .
21. Hammond, J. M. & Krall, E. Biochem. biophys. Res. Commun. 91, 284-288, 1979.
22. Gibori, G., Richards, J. S. & Keyes, P. L. in Ovarian Follicular and Corpus-Luteum A third group of experiments were carried out to identify the
Function, 112 (eds Channing, C. P., Marsh, J. & Sadler, W. A.) 53-69 (Plenum, New cellular origin of axons in the graft. In 15 rats, a segment of the
York, 1979). spinal cord at least 10 mm long was removed and a sciatic nerve
23. Thorneycroft, I. H. & Stone, S.C. Contraception 5, 129-136 (1972).
graft of equal length was inserted. The roots were left intact.
Three to four months later the operative site was re-exposed and
horseradish peroxidase (HRP) was injected into the spinal cord
immediately rostral or caudal to the graft. After 2 more days,
Axons from CNS neurones each animal was killed and labelled neurones were sought in a
regenerate into PNS grafts 10-mm segment of the non-injected spinal cord stump,
immediately adjacent to the graft (Fig. 1). After injection rostral
to the graft, the mean number of labelled neurones seen in the
P.M. Richardson, U. M. McGuinness & A. J. Aguayo caudal spinal cord stump was 48 (0, 8, 28, 40, 169); after
The Neurosciences Unit, Montreal General Hospital and injection caudal to the graft, the mean number in the rostral
McGill University, 1650 Cedar Avenue, Montreal, Canada H3G 1A4 stump was 6 (0, 0, 1, 6, 23). In five control animals, the graft was
crushed with jeweller's forceps immediately before HRP
injection and no HRP was seen 2 d later in spinal neurones on
Axons in the peripheral nervous system (PNS) and central the other side of the graft. We conclude that neurones in the
nervous system (CNS) form sprouts after injury1- 3 • Elongation experimental groups were not labelled spuriously because of
of regenerating axonal sprouts has been observed as the excep· interstitial diffusion through the graft, flow in the cerebrospinal
tion within the adult mammalian CNS but is the rule in the PNS fluid or extravasation into the systemic vasculature. Labelling is
of mammals as weD as in the CNS of some fish and amphibians 4 • thought to have resulted from incorporation of HRP into axon
The relative importance of intrinsic neuronal properties and terminals at or near the spinal cord-graft junction and retro-
axonal environment in determining the extent of axonal grade axonal transport across the graft to perikarya in the
regrowth is unknown5 • Neuroglial cells, nerve growth factor and adjacent spinal cord. In segments below the graft, labelled
target tissues such as smooth muscle are known to influence
neuronal responses to injury6 ' 7 , Here we have examined the
capacity of transected axons originating in the CNS to regrow
into nerve grafts containing Schwann cells.
The surgical technique used, a modification of that previously
described 8 ·9 , was carried out in young adult Sprague-Dawley
rats. In the first group of 16 animals, a 5-mm segment of the
midthoracic spinal cord was removed and an autologous sciatic
nerve graft was inserted sub-pially between the two stumps of
the spinal cord. One to four months later, the animals were
anaesthetised and perfused. The grafts and adjacent spinal cord
tissue were examined by light and electron microscopy. The
perineurium surrounding the common peroneal and posterior
tibial fascicles of the sciatic nerve was clearly identifiable in
cross-sections through the middle of the graft and contained
many myelinated or unmyelinated axons ensheathed by
Schwann cells. The ultrastructural appearance of each graft was
similar to that of a regenerated peripheral nerve. The grafts were
in gross continuity with the proximal and distal spinal cord Fig. 1 Left, 2-4 months after the original operation, HRP (20%
stumps. At the spinal cord-graft interfaces, small cysts were Sigma VI) was injected through a glass pipette into the spinal cord
frequently found but were rarely greater than 1 mm in diameter. immediately rostral (or caudal) to the sciatic nerve graft. The
In electron micrographs, the marginal spinal cord tissue was injection of small quantities of HRP (0.5 J..LI in 30 min) was designed
seen to contain many astrocytic processes and to be surrounded to minimise the possibility of a false positive result due to the
by a basal lamina. Small, irregular dome-like neuroglial struc- extravasation of enzyme and not to label every axon crossing the
tures were observed to protrude towards the graft at the ragged graft. Vaseline was spread extradurally and subdurally about the
edge of the spinal cord tissue. These protuberances resemble graft before injection, dry cotton batting was placed about the
pipette during the injection and mineral oil was injected through
structures seen at the interface of optic nerve grafts and the pipette as it was withdrawn. Two days later, the animal was
peripheral nerves 10 and also at the normal dorsal root entry perfused with glutaraldehyde. From a 10-mm segment of the spinal
zone 11 • Occasional fibres were found where the axon was sur- cord immediately across the graft from the injection site, specimens
rounded by peripheral myelin (periodicity 15.0 nm) and were removed and rinsed overnight in sucrose buffer. Approxi-
Schwann cell cytoplasm on one side of a node of Ranvier and mately 300 sections, each 20 J..Lm thick, were cut on a freezing
central myelin (periodicity 13.6 nm) and astrocytic processes on microtome, mounted, incubated with tetramethyl benzidine and
the other. From these morphological observations we conclude hydrogen peroxide and counterstained with neutral red 24 . Right,
that some axons crossed the border between the graft and spinal composite diagram illustrating the position of labelled nerve cell
cord, although the direction in which they grew could not be bodies in the spinal grey matter below the graft in four animals after
HRP injection above the graft. Each animal is represented by a
determined.
different symbol. Few neurones are labelled in the tips of the dorsal
Because of the potential ability of axons in the dorsal and horns and in the lateral portion of the anterior horns. Because of
ventral spinal roots to innervate such grafts, all roots crossing the the small amount of HRP injected and the short length of the spinal
pia-arachnoid at the graft site as well as the corresponding dorsal cord surveyed, it is assumed that more spinal neurones project into
root ganglia were avulsed in other rats. Grafts in rats with the graft than are shown here.

0028-{)836/80/ 120264--02$01.00 © 1980 Macmillan Journals Ltd

Nature Vol. 284 20 March 1980 265
This work was supported by the MRC of Canada, the Multiple
Sclerosis Society of Canada and the Montreal General Hospital
Research Institute.
Received 14 September 1979; accepted 21 January 1980.

1. Ramon Y Cajal, S. Degeneration and Regeneration of the Nervous System (ed. May, R. M.)
(Oxford University Press, London, 1928).
2. Liu, C. N. & Chambers, W. W. Archs Neurol. Psychiat. 79,46-61 (1958).
3. Raisman, G. Ann. Neurol. 3, 101-106 (1978).
4. Singer, M.. Nordlander, R. H. & Egar, M. J. comp. Neuro/. 185, 1-22 (1979).
5. Grafstein, B. & McQuarrie, I. G. in Neuronal Plasticity (ed. Cotman, C. W.) (Raven, New
York, 1978).
6. Stenevi, U., Bjerre, B., Bjorklund A. & Mobley, W. Brain Res. 69,217-234 (1974).
7. Varon, S. S. & Bunge, R. P. A Rev. Neurosci 1, 327-361 (1978).
8. Sugar, 0. & Gerard, R. W. J. Neurophysio/. 3, 1-19 (1940).
9. Kao, C. C., Chang, L. W. & Bloodworth, J. M. B. Exp/ Neuro/. 54,591-615 (1977).
10. Aguayo, A. J. eta/. Neurosci. Lett. 9, 97-104 (1978).
Diameter (pm) 11. Berthold, C. H. & Carlstedt, T. Acta physiol. scand. suppl. 446,23-42 (1977).
12. Bratten, B. & Hudson, A. Can. J. neurol. Sci. 6, 394-395 (1979).
13. Rexed, B. J. comp. Neurol. 100,297-379 (1954).
14. Adams, J. H., Daniel, P.M. & Prichard, M. M. L. J. comp. Neurol. 135, 121-144 (1969).
Fig. 2 Histogram showing the diameters as measured under the 15. Katzman, R., Broida, R. & Raine, C. S. Brain Res. 138,423-443 (1977).
light microscope, of perikarya in the grey matter of the lower 16. Svendgaard, N. A .. Bjorklund, A. & Stenevi, U. Adv. Anat. Embryo/. cell. Bioi. 51, pt 4,
thoracic cord (excluding the dorsal part of the dorsal horn and 7-77 (1975).
17. Svendgaard, N. A., Bjorklund, A. & Stenevi, U. Brain Res. 10Z, 1-22 (1976).
lateral part of the ventral horn). White columns represent neurones 18. Hancock, M. B. & Peveto, C. A. J. comp. Neuro/. 183, 65-72 (1979).
in a normal rat; dotted columns represent unlabelled neurones 19. Carlsson, A., Falck, B. & Hillarp, N. A. Acta physiol. scand. 60, 112-119 (1964).
below the graft; black columns represent labelled neurones below 20. Lampert, P. & Cressman, M. Lab. Inv.st. 13,825-839 (1964).
21. Weinberg, E. L. & Spencer, P. S. Brain Res. 162, 273-279 (1979).
the graft. Labelled neurones have a larger mean diameter than 22. Stensaas, L. J .. Burgess, P.R. & Horch, K. W. Soc. Neurosci. Abstr. 5, 684 (1979).
either control group. 23. Aguayo, A. J., Bray, G. M., Perkins, C. S. & Duncan, I. D. in Soc. Neurosci. Symp. IV (ed.J.
A. Ferrendelli) 361-383 (1979).
24. Mesulam, M. M. J. Histochem. Cytochem. 24, 1273-1280 (1978).

neurones were scattered through the central part of the grey

matter (laminae 4, 5, 7, 8, 10) 13 (Fig. 1). Both small and large
nerve cells were labelled and the diameter of HRP-containing
Response to stress of
neurones was slightly greater than that of unlabelled neurones in mesocortico-frontal dopaminergic
the same regions of the spinal grey matter (Fig. 2). Some anterior neurones in rats after long-term isolation
horn cells above the graft and some dorsal root ganglia at the
level of the graft and several segments below or above it also
contained labelled cells but the ganglia were not surveyed G. Blanc, D. Herve, H. Simon*, A. Lisoprawski,
systematically. The evidence demonstrates that some axons J. Glowinski & J.P. Tassin
originating from neurones within the spinal cord grew a distance
Groupe NB, INSERM U.114, College de France, 11, place Marcelin
of approximately 10 mm into a PNS graft. We do not know Berthelot, 75231 Paris Cedex 05, France
whether such axons then re-entered the spinal cord and formed
synapses. No functional improvement attributable to axonal
regeneration was observed in any animal. Recent studies suggest that the mesocortico-frontal
Mammalian central neurones with recognised capacity for dopaminergic neurones which originate in the ventral tegmental
axonal regrowth over a distance can be placed in two groups. In area (VTA) have an inhibitory role in locomotor activity 1 ' 1 •
the first group, which includes somatic motor neurones ~nd They are also markedly activated under stress 3 -s. This effect
autonomic preganglionic neurones, the axons project outs1de was shown in rats and mice subjected to electric foot-shocks
the CNS. Intrinsic CNS neurones with regenerative ability have by measuring either the rate of decline of dopamine (DA)
been found in the hypothalamus 14 and brain stem 15 ' 16 • The latter after a-methylparatyrosine treatmene or the changes in
neurones are cholinergic or monoaminergic and are thought to dihydroxyphenylacetic acid (dopac) levels and the dopac/DA
have thinly myelinated or unmyelinated axons 17 • Th~ findings ratio~. In rats, stress-induced activation of the dopaminergic
reported here indicate that there are other neurones m the rat neurones was prevented by benzodiazepines4 's, and studies in
spinal cord also capable of axonal elongation after injury. So~e BALB/c mice introduced for 2 min into an open field further
of the HRP-labelled perikarya probably represent autonomtc established the role of dopaminergic neurones in emotional
preganglionic neurones about the central canal 18 or in t~e responses 7 • These observations led us to examine the effects of
intermediolateral columns, but most of them were found m long-term isolation on the activity of the mesocortico-frontal
areas of spinal grey matter normally occupied by neither dopaminergic neurones in rats, some of which were subjected to
autonomic nor somatic motor neurones. None is likely to have a stressful situation. Indeed, several groups have reported that
been monoaminergic because cell bodies containing cate- long-term isolation in rodents induced behavioural disturbances
cholamines or serotonin have not been discovered in the spinal such as increased motor activity 8 ' 9 and aggression 9 ' 10 and hfper-
grey matter 19 • In short, the labelled cells do not belong. to a class reactivity to a new environment or stressful stimuli 10' 1 • As
of neurones in which axonal regrowth has been descnbed pre- measured by the changes in dopac levels or the dopac/DA ratio,
viously. It is unknown whether or not spinal axons with we report here that the activity of the mesocortico-frontal
regenerative potentiality have any specific anatomical or bio- dopaminergic neurones was reduced after isolation. This was
chemical property. not the case for the dopaminergic neurones projecting to the
Spinal axons such as those which grew into a peripheral nerve nucleus accumbens or the striatum, the rate of DA utilisation in
graft in these experiments have repeatedly demonstrated only these structures was even enhanced in isolated rats in which the
abortive sprouting after simple transection 1 ' 20 • This evidence activity of the mesocortico-frontal dopaminergic neurones was
that the regenerative response of similar axons differs in CNS markedly reduced. Finally, we will show that a 3-min electric
and PNS neuroglia, together with other experimental foot-shock session is more effective in enhancing dopac levels or
observations 10 •21 •22 , supports the hypothesis that Schwann cells the dopac/DA ratio in the frontal cortex of isolated than
are more conducive to axonal regeneration than central neuro- grouped rats.
glial cells 23 • The regenerative potentiality o~ CNS n~urones m~y
be expressed only when their neuroglial env1ronment 1s *Laboratoire de Neurobiologie du Comportement, Universite de Bordeaux II, Rue Le's
changed. Saignat, 33076 Bordeaux Cedex, France.

0028-0836/80/120265--{)3$01.00 © 1980 Macmillan Journals Ltd