EXPERIMENTAL STUDIES AVULSED SPINAL NERVE
ON SURGICAL TREATMENT OF ROOTS IN BRACHIAL PLEXUS INJURY
T. CARLSTEDT From the Department of Hand Surgery, Sabbatsbergs Hospital, Stockholm, Sweden
This review summarises studies aiming at a surgical treatment of spinal nerve root avulsions from the spinal cord in bra&al plexus lesions. After dorsal root injury, regrowth of nerve fibres into the spinal cord occurs only in the immature animal. After ventral root avulsion and subsequent implantation into the spinal cord, neuroanatomical and neurophysiological data show that motoneurons are capable of producing new axons which enter the implanted root. Intra-neuronal physiological experiments demonstrate that new axons can conduct action potentials and elicit muscle responses. The neurons are reconnected in segmental spinal cord activity and respond to impulses in sensory nerve fibres. In primate experiments, implantation of avulsed ventral roots in the bra&al plexus resulted in functional restitution. These studies indicate the possibility of surgical treatment of ventral root avulsion injuries in bra&al plexus lesions in humans. Journal of Hand Surgery (British Volume, 1991) 16B: 477-482 Avulsion of nerve roots from the spinal cord makes the brachial plexus lesion the most complicated nerve injury. In my experience 65% of the adult and 75% of the obstetrical brachial plexus lesions that have been operated upon suffer from avulsion injuries. These lesions are not situated in the peripheral nervous system (P.N.S.), but confined to the central nervous system (C.N.S.), which has a very limited capacity for regeneration. Therefore, the avulsed roots are not used in the reconstruction of the plexus. The spinal cord segment that has suffered root avulsion injury will accordingly be permanently detached from the periphery. Today, the surgical strategy after root avulsion injury is palliative. Avulsed roots are neurotised with other roots within the plexus or to nerves outside the plexus. A considerably impaired function in the arm and hand is still the usual outcome of these severe injuries (Alnot, 1983). In order to compensate for the loss of neurons in the spinal cord segment that has suffered root avulsion, neurotisation to nerves like the accessory or intercostal nerves is performed (Narakas, 1984). This implies, however, that function is reduced in the ordinary territories of these nerves in order to gain activity in regions deprived of innervation. Thus, there is a decrease in trapezius muscle activity or vital capacity when using the accessory or intercostal nerves. Attempts to treat the root avulsion injuries have been performed. Thus, in a patient with avulsion of all roots in the plexus, Bonney and Jamieson at St Marys hospital in London, reconnected roots to the spinal cord (Jamieson and Eames, 1980). The outcome of the surgery was however disappointing. For regeneration to occur after dorsal root injury, sensory nerve fibres must elongate through C.N.S. tissue in order to reach the target regions in the dorsal horn of the spinal cord. After ventral root avulsion, regrowth of motoneuron axons must take place in the spinal cord VOL. 16B No. 5 DECEMBER
1991
before reaching the P.N.S. in the ventral root. Recently it was demonstrated that spinal cord motoneurons could regrow nerve fibres across a lesion site within the C.N.S. tissue of the spinal cord and reach the adjacent ventral root (Risling et al., 1983). The demonstration of this surprisingly strong regenerative ability of motoneurons initiated a series of animal experiments aiming at functional recovery after root avulsion. This review summarises these experiments. Ventral root injury After ventral root avulsion and subsequent implantation into the ventro-lateral aspect of the spinal cord (Fig. 1) of rats, regeneration of spinal cord motoneurons was investigated by means of retrograde horseradish peroxidase (H.R.P.) staining (for details see Carlstedt et al., 1986). A large number of retrograde H.R.P.-labelled neurons was found in the ventral horn of the spinal cord (Fig. 2). These neurons had the same size and location as motoneurons. Electrical stimulation of the implanted root gave rise to a widespread muscle twitch response (Carlstedt et al., 1986). Similar results have been reported elsewhere (Horvat et al., 1987; Smith et al., 1988; Hoffman et al., 1990). These initial results indicated that implantation of avulsed ventral roots may induce reinnervation of roots and peripheral nerves as well as skeletal muscles. To obtain conclusive evidence for motoneuron regeneration through implanted ventral roots and peripheral nerves, a study of individual regenerated neurons was performed in the cat (Cullheim et al., 1989). For technical reasons lumbo-sacral ventral roots were avulsed, and one root was implanted (Fig. 1). One year after surgery, intra-cellular neurophysiological experiments were performed. Neurons which could be activated antidromically via electrical stimulation of the implanted root were impaled with H.R.P.-filled glass 471
T. CARLSTEDT
Fig. 1
Fig. 2
After ventral
roots had been avulsed,
one root was implanted
The distribution of regenerated motoneurons in rat L4 spinal cord segment after ventral root avulsion and implantation (from Carlstedt et al., 1986, Acta Physiologica Scandinavica, 128 : 645-646).
micro-electrodes (Fig. 3) (Cullheim et al., 1989). These could record the electrophysiological activity in the neuron (Fig. 3). Electrical stimulation of the implanted ventral root elicited action potentials in neurons located in the motoneuron nucleus in the spinal cord. Electrical stimulation of sensory, dorsal root nerve fibres, which entered the same spinal cord segment, induced an action potential secondary to the generation of an excitatory post-synaptic potential (E.P.S.P.) (Fig. 4). This showed that the motoneuron had contact with afferent synapses, the activity of which could induce an action potential in the motoneuron. In some smaller motoneurons with the size of gamma motoneuron, the E.P.S.P. was followed by an 478
superficially
into the ventro-lateral
aspect of the spinal cord.
inhibitory post-synaptic potential I.P.S.P. (Fig. 4). Thus, these findings demonstratedthat axons which are able to conduct action potentials had regenerated into the avulsed and implanted ventral root and the motoneurons had been reconnected with afferent nerve fibres and participated in normal spinal cord segmental reflex activity. After the physiological experiments were concluded the neurons were injected via the intra-neuronal electrode and stained with H.R.P. for a subsequent morphological analysis (for details see Cullheim et al., 1989). Most stained neurons had the size of alpha motoneuTons, but also nerve cells with the size of gamma neurons occurred. The regenerated axons did not have the ordinary ventral direction in the spinal cord. Instead, they had a dorso-lateral course towards the implanted ventral root (Figs 4 and 5). Electron microscopical analysis showed that these axons had regenerated through C.N.S. tissue for a considerable distance before reaching the implanted root (Cullheim et al., 1989). Thus, a true C.N.S. regeneration seemed to be the first step towards recovery. Isometric muscle twitch responses could be elicited after electrical stimulation of the implanted root. In a muscle such as the soleus muscle, which normally has an extremely slow twitch behaviour (Burke et al., 1974), this specific property had been lost (Fig. 4). This indicated that the muscle had been denervated or lost contact with specific slow twitch soleus motoneurons in the spinal cord, and then became reinnervated non-specifically by predominantly fast twitch neurons. This is an expected outcome after this type of surgery (Miledi and Stefani, 1969 ; Brushart and Mesulam, 1980) and also gave strong support to the idea that the muscle had been reinnervated by newly formed axons via the implant. Thus, these results provided evidence that motoneurons had the capacity to regrow within the spinal cord, i.e. C.N.S. THE JOURNAL
OF HAND
SURGERY
EXPERIMENTAL
Fig. 3
STUDIES ON SURGICAL
TREATMENT
OF AVULSED SPINAL NERVE ROOTS IN BRACHIAL
PLEXUS INJURY
Schematic drawing showing an intra-cellular penetration of a motoneuron in a spinal cord segment. Electrical stimulation of the implanted root activated the neuron to produce action potential of normal size and time course. H.R.P. was injected through the electrode into the neuron for a subsequent morphological analysis (see Figs 4 and 5) (from Carlstedt et al., 1988, Progress in brain Research, 78: 225-229).
regeneration, and that this can lead to restoration of function. It has been speculated that a persisting defect in the blood-brain barrier as a consequence of a trauma, might promote regeneration by giving way for blood-borne trophic factors (Kiernan and Contestabile, 1980). Testing the patency of the blood-brain barrier after the avulsion trauma showed that there is a prolonged defect persisting up to 18-20 months after surgery (Risling et al., 1990). This long-term barrier dysfunction might be linked with the survival of axonal sprouts and support the regenerative capacity of the spinal cord motoneurons through the C.N.S. tissue. Implantation of a root to the lateral side of the cord will constrain the regenerating motor axons to regrow within the C.N.S. tissue in order to face the new location of the root. In recent experiments, alternative conduits for the regenerating motor axons were investigated. In wholemount preparations comprising the pia mater, the avulsion site and the implanted root, several axons could be followed in continuity between the avulsion site and the implanted root (Risling et al., 1991). The pia mater has a verified competence to promote the growth of regenerating axons (Risling et al., 1984; David, 1988). VOL. 16B No. 5 DECEMBER
1991
Thus, an alternative conduit for the regenerating motoaxons along the surface of the spinal cord, i.e. the pia mater, was found (Fig. 6). In order to optimise the conditions for regeneration it therefore seems logical to replant the ventral root as close to the avulsion site as possible. These encouraging results of neural regrowth point toward clinical application. A primate study is, however, inevitable before human application can take place. Therefore, an investigation has been initiated to see if regeneration after ventral root avulsion and implantation also occurs in primates (Carlstedt et al., in preparation). In cynomogulus monkeys, ventral roots C5-C7, connected to the brachial plexus were avulsed and concommittantly implanted into the ventro-lateral aspect of the spinal cord. The outcome of the surgery was followed clinically, neurophsyiologically as well as morphologically, and with H.R.P.-tracing techniques. With repeated E.M.G. investigations of the biceps muscle, denervation activity was noted four to six weeks after surgery. Electrophysiological signs of reinnervation started slightly before clinical activity, i.e. elbow flexion, was observed. After three months the function improved and the behaviour changed towards normal. Injection of H.R.P. in the biceps muscle showed a random distribu419
T. CARLSTEDT
tion of regenerated motoneurons in the ventral horn of the C6 spinal cord segment. Clinical investigations showed that side-effects such as leg spasticity due to injury of long spinal cord tracts during surgery did not occur. These results demonstrated that the ventral root avulsion injury in brachial plexus lesions could be managed surgically with functional return in primates.
Dorsal root injury
e
SOLEUS NORMAL
OPERATED 5@3
5009
After dorsal root avulsion and subsequent implantation into the dorsal horn of the spinal cord, regeneration of nerve fibres into the spinal cord did not occur (Fig. 7) (Jamieson and Eames, 1980; Carlstedt, 1985a). The conspicuous difference in nervous regenerative capacity in the P.N.S. versus the C.N.S. has been dramatically demonstrated at the P.N.S.-C.N.S. interface in dorsal roots (Reier et al., 1983; Carlstedt, 1985b). In adult mammals, post-lesional dorsal root nerve fibres regrow in the root but stopped growing abruptly at the P.N.S.C.N.S. interface, at the root-spinal cord junction (CarIstedt, 1985b, 1989). In the newborn animal, however, nerve fibres regrow across the root-spinal cord junction and further among the C.N.S. glial cells of the spinal cord (Carlstedt et al., 1987; Carlstedt, 1989). Regenerated dorsal root nerve fibres could be demonstrated in appropriate terminal regions in the spinal cord dorsal horn with H.R.P.-tracing technique (Fig. 8) (Carlstedt et al., 1987). With electron microscopy, it could be demonstrated that these regenerated dorsal root nerve fibres had developed to a normal size and elongated within C.N.S. tissue of the spinal cord (Carlstedt, 1989). Unfortunately this capacity to regrow from the dorsal root into the spinal cord is lost after the neonatal period is over (Carlstedt, 1989).
Comments Fig. 4
480
(a) A nemophysiologically tested and H.R.P.-tilled cat spinal cord motoneuron reconstructed from series of sections in a camera lucida drawing. The new axon has taken a dorsolateral course into the implanted root. Note the absence of an axon with the ordinary ventral projection. (b)-(d) Intra-cellular recordings. (b) Action potential after antidromic stimulation of the implanted root 15 mm. from the cord with a latency of 0.54.6ms. (c) Same neuron. Stimulation of afferent nerve fibres gave rise to a large E.P.S.P. which produced an action potential. (d) Occasionally, in some neurons, the intra-cellular electrode recorded I.P.S.P. Small arrow indicate time of stimulation. (Scales : vertical bar = 20 mV. ; horizontal bar = 2 ms. in (b) and (c) and 10 ms. in (d). (e) Isometric muscle twitch response of the soleus muscle in an operated and control animal. Note the large difference in time course between the normal twitch response and the soleus twitch in operated animals. The horizontal time scale in each recording gives the twitch contraction time (unit 20 ms.) (From Cullheim et al., 1989, Neuroscience, 29: 7255733).
Studies on root-spinal cord regeneration are of great neurobiological as well as clinical importance. After avulsion of ventral roots, new motor axons enter the ventral root implant. These axons take a new path and regrow for a considerable distance within the spinal cord or along the pia mater on the spinal cord surface, before entering the root. The new axons can conduct action potentials and elicit muscle responses. The motoneurons are reconnected in segmental spinal cord activity and respond to impulses in primary sensory nerve fibres. In primates, implantation of avulsed ventral roots of the brachial plexus can restore function. Instead of the presently performed palliative surgery after avulsion injury, there is experimental evidence that implantation of avulsed ventral roots can directly reconnect the detached spinal cord segments and restore THE JOURNAL
OF HAND
SURGERY
EXPERIMENTAL
Fie.
5
STUDIES
ON SURGICAL
Light microaranh of H.R.P.-filled Nkrroscienc~, 29: 725-733).
TREATMENT
motoneuron
OF AVULSED
after intra-cellular
function also in humans with severe brachial plexus injury. In contrast, dorsal root nerve fibres do not regrow into the spinal cord of the adult animal. There is thus a great discrepancy in regenerative propensity after ventral and dorsal root avulsion injury. This difference can not be easily explained today; motoneurons might be stimulated to regrow new axons in the spinal cord by the presence of denervated P.N.S.
SPINAL
NERVE
neurophysiological
ROOTS
IN BRACHIAL
PLEXUS
INJURY
tests. Scale bar 10 urn. (From Cullheim
et al.,
tissue in the implanted ventral root or by defects in the blood-brain barrier. Dorsal root axonal regeneration into the spinal cord of immature animals might depend on a greater regenerative capacity for immature than adult neurons, or immature glial cells might be more permissive to axonal elongation than in the adult. These are some of the neurobiological problems that should be explored in finding surgical treatment of dorsal root avulsion injuries.
Intramedullar , innervation
Fig. 6
VOL.
Schematic Neurology
drawings of possible and Neuroscience).
16B No. 5 DECEMBER
1991
routes for motoneurons
to reinnervate
the implanted
ventral
root (From
Risling
et al., 1991, Restorative
481
T. CARLSTEDT
References
Fig. 7
Light micrograph of a transverse section through the dorsal part of the spinal cord after dorsal root avulsion and implantation (arrows). Re-established connectivity from the root to the spinal cord cannot be demonstrated. Scale bar= 100 pm. (From Carlstedt, 1985, Neuroscience Letters, 55: 343348).
Fig. 8
Camera lucida drawing of H.R. P.-labelled dorsal root terminals in the superficial laminae of the dorsal horn after regeneration in the newborn rat. (From Carlstedt et al., 1987, Neuroscience Letters, 74: 1418).
Acknowledgements This review is based on studies supported by the Swedish Medical Research Council (project no. 6532) and funds from Karolinska Institutet. My sincere thanks to Dr HBkan LugnegBrd, Head of the Department of Handsurgery, Sabbatsbergs sjukhus, for support and criticism of this manuscript.
482
ALNOT, J. Y. (1983). Paralysie traumatique du plexus brachial de l’adulte. Encyclopedic Medico-Chirurgicale (Paris), 9: l-15. BRUSHART, T. M. and MESULAM, M. M. (1980). Alternation in connections between muscle and anterior horn motoneurons after peripheral nerve repair. Science, 208: 603-605. BURKE, R. E., LEVINE, D. N., SALCMAN, M. and TSAIRIS, P. (1974). Motor units in cat soleus muscle: physiological histochemical and morphological characteristics. Journal of Physiology, 238 : 503-514. CARLSTEDT, T. (1985a). Dorsal root innervation of spinal cord neurons after dorsal root implantation into the spinal cord of adult rats. Neuroscience Letters, 55 : 343-348. CARLSTEDT,T. (1985b). Regeneratingaxonsformnerveterminalsatastrocytes. Brain Research, 347: 188-191. CARLSTEDT, T. (1989). Reinnervation of the mammalian spinal cord after neonatal dorsal root crush. Journal of Neurocytology, 17: 335-350. CARLSTEDT, T., LINDA, H., CULLHEIM, S. and RISLING, M. (1986). Reinnervation of hindlimb muscles after ventral root avulsion and implantation in the lumbar spinal cord of the adult rat. Acta Physiologica Scandinavica, 128: 645-646. CARLSTEDT, T., DALSGAARD, C.-J. and MOLANDER, C. (1987). Regrowth of lesioned dorsal root nerve fibres into the spinal cord of neonatal rats. Neuroscience Letters, 74: 14-18. CULLHEIM, S., CARLSTEDT, T., LINDA, H., RISLING, M. and ULFHAKE, B. (1989). Motoneurons reinnervate skeletal muscle after ventral root implantation into the spinal cord of the cat. Neuroscience, 29: 725-733. DAVID, S. (1988). Neurite outgrowth from mammalian C.N.S. neurons on astrocytes in vitro may not be mediated primarily by laminin. Journal of Neurocytology, 17: 131-144. HORVAT, J-C., PECOT-DECHAVASSINE, M. and MIRA, J-C. (1987). Reinnervation fonctionelle d’un muscule squelettique du Rat adulte au moyen d’un greffon de nerf p&riph&que introduit dans la moelle Cpinibre par voi dorsale. Comptes rendus de I’Academie des Sciences (Paris), 304: 143-148. HOFFMANN, C. F. E., THOMEER, R. T. W. M. and MARANI, E. (1990). Ventral root avulsion of the cat spinal cord at the brachial plexus level (Cervical 7). European Journal of Morphology, 28: 418-429. JAMIESON, A. M. and EAMES, R. A. (1980). Reimplantation of avulsed brachial plexus rots: an experimental study in dogs. International Journal of Microsurgery, 2: 75-85. KIERNAN, J. A. and CONTESTABILE, A. (1980). Vascular permeability associated with axonal regeneration in the optic system of the goldfish. Acta Neuropathologica, 51: 39-45. MILEDI, R. and STEFANI, E. (1969). Nonselective re-innervation of slow and fast muscle fibres in the rat. Nature, 222: 569-571. NARAKAS, A. 0. (1984). Thoughts on neurotization or nerve transfers in irreparable nerve lesions. Clinics in Plastic Surgery, 11: 153-160. REIER, P. J., STENSAAS, L. J. and GUTH, L. The astrocytic scar as an impediment to regeneration in the central nervous system. In Cao, C. C., Bunge, R. P. and Reier, P. j. (Eds) Spinal Cord Reconstruction. Raven Press, NewYork, 1983: 163-195. RISLING, M., CULLHEIM, S. and HILDEBRAND, C. (1983). Reinnervation of the ventral root L7 from ventral horn neurons following intramedullary axotomy in adult cats. Brain Research, 280: 15-23. RISLING, M., DALSGAARD, C. J. and CUELLO, A. C. (1984). Invasion of lumbosacral ventral roots and spinal pia mater by substance P-immunoreactive axons after sciatic nerve lesion in kittens. Brain Research, 307: 351-354. RISLING, M., SJdGREN, LINDA, H., CULLHEIM, S., CARLSTEDT, T. and TELESTAM, M. Demonstration of a persistent defect in the bloodbrain barrier after ventral root implantation into the spinal cord using immunogold technique and staphylococcal alpha-toxin. In Johansson, B. B., Owman, C. and Widner, H. (Eds) Pathophysiology of the blood-brain barrier. Elsevier, 1990: 555-559. RISLING, M., CARLSTEDT, T., LINDA, H., and CULLHEIM, S. (1991). The pia mater-A conduit for regenerating axons after ventral root replant&ion. Restorative Neurology and Neuroscience, in press. SMITH, K. J., TERZIS, J. K., ERASMUS, M. and CARSON, K. A. Reinnervation of deneiiated skeletal muscle ‘by central nerve fibres regenerating along replanted ventral roots. In Gash, D. M. and Sladek Jr, J. R. (Eds) Transplantation into the Mammalian C.N.S. Progress in Brain Research, Vol78. Elsevier, 1988: 193-198. Thomas Carlstedt, M.D., Ph.D., Associate Professor, Department of Handsurgely, Sabbatsbergs Hospital, S-l 1382 Stockholm, Sweden. 0 1991 The British Society for Surgery of the Hand THE
JOURNAL
OF HAND
SURGERY
ON SURGICAL TREATMENT OF ROOTS IN BRACHIAL PLEXUS INJURY
T. CARLSTEDT From the Department of Hand Surgery, Sabbatsbergs Hospital, Stockholm, Sweden
This review summarises studies aiming at a surgical treatment of spinal nerve root avulsions from the spinal cord in bra&al plexus lesions. After dorsal root injury, regrowth of nerve fibres into the spinal cord occurs only in the immature animal. After ventral root avulsion and subsequent implantation into the spinal cord, neuroanatomical and neurophysiological data show that motoneurons are capable of producing new axons which enter the implanted root. Intra-neuronal physiological experiments demonstrate that new axons can conduct action potentials and elicit muscle responses. The neurons are reconnected in segmental spinal cord activity and respond to impulses in sensory nerve fibres. In primate experiments, implantation of avulsed ventral roots in the bra&al plexus resulted in functional restitution. These studies indicate the possibility of surgical treatment of ventral root avulsion injuries in bra&al plexus lesions in humans. Journal of Hand Surgery (British Volume, 1991) 16B: 477-482 Avulsion of nerve roots from the spinal cord makes the brachial plexus lesion the most complicated nerve injury. In my experience 65% of the adult and 75% of the obstetrical brachial plexus lesions that have been operated upon suffer from avulsion injuries. These lesions are not situated in the peripheral nervous system (P.N.S.), but confined to the central nervous system (C.N.S.), which has a very limited capacity for regeneration. Therefore, the avulsed roots are not used in the reconstruction of the plexus. The spinal cord segment that has suffered root avulsion injury will accordingly be permanently detached from the periphery. Today, the surgical strategy after root avulsion injury is palliative. Avulsed roots are neurotised with other roots within the plexus or to nerves outside the plexus. A considerably impaired function in the arm and hand is still the usual outcome of these severe injuries (Alnot, 1983). In order to compensate for the loss of neurons in the spinal cord segment that has suffered root avulsion, neurotisation to nerves like the accessory or intercostal nerves is performed (Narakas, 1984). This implies, however, that function is reduced in the ordinary territories of these nerves in order to gain activity in regions deprived of innervation. Thus, there is a decrease in trapezius muscle activity or vital capacity when using the accessory or intercostal nerves. Attempts to treat the root avulsion injuries have been performed. Thus, in a patient with avulsion of all roots in the plexus, Bonney and Jamieson at St Marys hospital in London, reconnected roots to the spinal cord (Jamieson and Eames, 1980). The outcome of the surgery was however disappointing. For regeneration to occur after dorsal root injury, sensory nerve fibres must elongate through C.N.S. tissue in order to reach the target regions in the dorsal horn of the spinal cord. After ventral root avulsion, regrowth of motoneuron axons must take place in the spinal cord VOL. 16B No. 5 DECEMBER
1991
before reaching the P.N.S. in the ventral root. Recently it was demonstrated that spinal cord motoneurons could regrow nerve fibres across a lesion site within the C.N.S. tissue of the spinal cord and reach the adjacent ventral root (Risling et al., 1983). The demonstration of this surprisingly strong regenerative ability of motoneurons initiated a series of animal experiments aiming at functional recovery after root avulsion. This review summarises these experiments. Ventral root injury After ventral root avulsion and subsequent implantation into the ventro-lateral aspect of the spinal cord (Fig. 1) of rats, regeneration of spinal cord motoneurons was investigated by means of retrograde horseradish peroxidase (H.R.P.) staining (for details see Carlstedt et al., 1986). A large number of retrograde H.R.P.-labelled neurons was found in the ventral horn of the spinal cord (Fig. 2). These neurons had the same size and location as motoneurons. Electrical stimulation of the implanted root gave rise to a widespread muscle twitch response (Carlstedt et al., 1986). Similar results have been reported elsewhere (Horvat et al., 1987; Smith et al., 1988; Hoffman et al., 1990). These initial results indicated that implantation of avulsed ventral roots may induce reinnervation of roots and peripheral nerves as well as skeletal muscles. To obtain conclusive evidence for motoneuron regeneration through implanted ventral roots and peripheral nerves, a study of individual regenerated neurons was performed in the cat (Cullheim et al., 1989). For technical reasons lumbo-sacral ventral roots were avulsed, and one root was implanted (Fig. 1). One year after surgery, intra-cellular neurophysiological experiments were performed. Neurons which could be activated antidromically via electrical stimulation of the implanted root were impaled with H.R.P.-filled glass 471
T. CARLSTEDT
Fig. 1
Fig. 2
After ventral
roots had been avulsed,
one root was implanted
The distribution of regenerated motoneurons in rat L4 spinal cord segment after ventral root avulsion and implantation (from Carlstedt et al., 1986, Acta Physiologica Scandinavica, 128 : 645-646).
micro-electrodes (Fig. 3) (Cullheim et al., 1989). These could record the electrophysiological activity in the neuron (Fig. 3). Electrical stimulation of the implanted ventral root elicited action potentials in neurons located in the motoneuron nucleus in the spinal cord. Electrical stimulation of sensory, dorsal root nerve fibres, which entered the same spinal cord segment, induced an action potential secondary to the generation of an excitatory post-synaptic potential (E.P.S.P.) (Fig. 4). This showed that the motoneuron had contact with afferent synapses, the activity of which could induce an action potential in the motoneuron. In some smaller motoneurons with the size of gamma motoneuron, the E.P.S.P. was followed by an 478
superficially
into the ventro-lateral
aspect of the spinal cord.
inhibitory post-synaptic potential I.P.S.P. (Fig. 4). Thus, these findings demonstratedthat axons which are able to conduct action potentials had regenerated into the avulsed and implanted ventral root and the motoneurons had been reconnected with afferent nerve fibres and participated in normal spinal cord segmental reflex activity. After the physiological experiments were concluded the neurons were injected via the intra-neuronal electrode and stained with H.R.P. for a subsequent morphological analysis (for details see Cullheim et al., 1989). Most stained neurons had the size of alpha motoneuTons, but also nerve cells with the size of gamma neurons occurred. The regenerated axons did not have the ordinary ventral direction in the spinal cord. Instead, they had a dorso-lateral course towards the implanted ventral root (Figs 4 and 5). Electron microscopical analysis showed that these axons had regenerated through C.N.S. tissue for a considerable distance before reaching the implanted root (Cullheim et al., 1989). Thus, a true C.N.S. regeneration seemed to be the first step towards recovery. Isometric muscle twitch responses could be elicited after electrical stimulation of the implanted root. In a muscle such as the soleus muscle, which normally has an extremely slow twitch behaviour (Burke et al., 1974), this specific property had been lost (Fig. 4). This indicated that the muscle had been denervated or lost contact with specific slow twitch soleus motoneurons in the spinal cord, and then became reinnervated non-specifically by predominantly fast twitch neurons. This is an expected outcome after this type of surgery (Miledi and Stefani, 1969 ; Brushart and Mesulam, 1980) and also gave strong support to the idea that the muscle had been reinnervated by newly formed axons via the implant. Thus, these results provided evidence that motoneurons had the capacity to regrow within the spinal cord, i.e. C.N.S. THE JOURNAL
OF HAND
SURGERY
EXPERIMENTAL
Fig. 3
STUDIES ON SURGICAL
TREATMENT
OF AVULSED SPINAL NERVE ROOTS IN BRACHIAL
PLEXUS INJURY
Schematic drawing showing an intra-cellular penetration of a motoneuron in a spinal cord segment. Electrical stimulation of the implanted root activated the neuron to produce action potential of normal size and time course. H.R.P. was injected through the electrode into the neuron for a subsequent morphological analysis (see Figs 4 and 5) (from Carlstedt et al., 1988, Progress in brain Research, 78: 225-229).
regeneration, and that this can lead to restoration of function. It has been speculated that a persisting defect in the blood-brain barrier as a consequence of a trauma, might promote regeneration by giving way for blood-borne trophic factors (Kiernan and Contestabile, 1980). Testing the patency of the blood-brain barrier after the avulsion trauma showed that there is a prolonged defect persisting up to 18-20 months after surgery (Risling et al., 1990). This long-term barrier dysfunction might be linked with the survival of axonal sprouts and support the regenerative capacity of the spinal cord motoneurons through the C.N.S. tissue. Implantation of a root to the lateral side of the cord will constrain the regenerating motor axons to regrow within the C.N.S. tissue in order to face the new location of the root. In recent experiments, alternative conduits for the regenerating motor axons were investigated. In wholemount preparations comprising the pia mater, the avulsion site and the implanted root, several axons could be followed in continuity between the avulsion site and the implanted root (Risling et al., 1991). The pia mater has a verified competence to promote the growth of regenerating axons (Risling et al., 1984; David, 1988). VOL. 16B No. 5 DECEMBER
1991
Thus, an alternative conduit for the regenerating motoaxons along the surface of the spinal cord, i.e. the pia mater, was found (Fig. 6). In order to optimise the conditions for regeneration it therefore seems logical to replant the ventral root as close to the avulsion site as possible. These encouraging results of neural regrowth point toward clinical application. A primate study is, however, inevitable before human application can take place. Therefore, an investigation has been initiated to see if regeneration after ventral root avulsion and implantation also occurs in primates (Carlstedt et al., in preparation). In cynomogulus monkeys, ventral roots C5-C7, connected to the brachial plexus were avulsed and concommittantly implanted into the ventro-lateral aspect of the spinal cord. The outcome of the surgery was followed clinically, neurophsyiologically as well as morphologically, and with H.R.P.-tracing techniques. With repeated E.M.G. investigations of the biceps muscle, denervation activity was noted four to six weeks after surgery. Electrophysiological signs of reinnervation started slightly before clinical activity, i.e. elbow flexion, was observed. After three months the function improved and the behaviour changed towards normal. Injection of H.R.P. in the biceps muscle showed a random distribu419
T. CARLSTEDT
tion of regenerated motoneurons in the ventral horn of the C6 spinal cord segment. Clinical investigations showed that side-effects such as leg spasticity due to injury of long spinal cord tracts during surgery did not occur. These results demonstrated that the ventral root avulsion injury in brachial plexus lesions could be managed surgically with functional return in primates.
Dorsal root injury
e
SOLEUS NORMAL
OPERATED 5@3
5009
After dorsal root avulsion and subsequent implantation into the dorsal horn of the spinal cord, regeneration of nerve fibres into the spinal cord did not occur (Fig. 7) (Jamieson and Eames, 1980; Carlstedt, 1985a). The conspicuous difference in nervous regenerative capacity in the P.N.S. versus the C.N.S. has been dramatically demonstrated at the P.N.S.-C.N.S. interface in dorsal roots (Reier et al., 1983; Carlstedt, 1985b). In adult mammals, post-lesional dorsal root nerve fibres regrow in the root but stopped growing abruptly at the P.N.S.C.N.S. interface, at the root-spinal cord junction (CarIstedt, 1985b, 1989). In the newborn animal, however, nerve fibres regrow across the root-spinal cord junction and further among the C.N.S. glial cells of the spinal cord (Carlstedt et al., 1987; Carlstedt, 1989). Regenerated dorsal root nerve fibres could be demonstrated in appropriate terminal regions in the spinal cord dorsal horn with H.R.P.-tracing technique (Fig. 8) (Carlstedt et al., 1987). With electron microscopy, it could be demonstrated that these regenerated dorsal root nerve fibres had developed to a normal size and elongated within C.N.S. tissue of the spinal cord (Carlstedt, 1989). Unfortunately this capacity to regrow from the dorsal root into the spinal cord is lost after the neonatal period is over (Carlstedt, 1989).
Comments Fig. 4
480
(a) A nemophysiologically tested and H.R.P.-tilled cat spinal cord motoneuron reconstructed from series of sections in a camera lucida drawing. The new axon has taken a dorsolateral course into the implanted root. Note the absence of an axon with the ordinary ventral projection. (b)-(d) Intra-cellular recordings. (b) Action potential after antidromic stimulation of the implanted root 15 mm. from the cord with a latency of 0.54.6ms. (c) Same neuron. Stimulation of afferent nerve fibres gave rise to a large E.P.S.P. which produced an action potential. (d) Occasionally, in some neurons, the intra-cellular electrode recorded I.P.S.P. Small arrow indicate time of stimulation. (Scales : vertical bar = 20 mV. ; horizontal bar = 2 ms. in (b) and (c) and 10 ms. in (d). (e) Isometric muscle twitch response of the soleus muscle in an operated and control animal. Note the large difference in time course between the normal twitch response and the soleus twitch in operated animals. The horizontal time scale in each recording gives the twitch contraction time (unit 20 ms.) (From Cullheim et al., 1989, Neuroscience, 29: 7255733).
Studies on root-spinal cord regeneration are of great neurobiological as well as clinical importance. After avulsion of ventral roots, new motor axons enter the ventral root implant. These axons take a new path and regrow for a considerable distance within the spinal cord or along the pia mater on the spinal cord surface, before entering the root. The new axons can conduct action potentials and elicit muscle responses. The motoneurons are reconnected in segmental spinal cord activity and respond to impulses in primary sensory nerve fibres. In primates, implantation of avulsed ventral roots of the brachial plexus can restore function. Instead of the presently performed palliative surgery after avulsion injury, there is experimental evidence that implantation of avulsed ventral roots can directly reconnect the detached spinal cord segments and restore THE JOURNAL
OF HAND
SURGERY
EXPERIMENTAL
Fie.
5
STUDIES
ON SURGICAL
Light microaranh of H.R.P.-filled Nkrroscienc~, 29: 725-733).
TREATMENT
motoneuron
OF AVULSED
after intra-cellular
function also in humans with severe brachial plexus injury. In contrast, dorsal root nerve fibres do not regrow into the spinal cord of the adult animal. There is thus a great discrepancy in regenerative propensity after ventral and dorsal root avulsion injury. This difference can not be easily explained today; motoneurons might be stimulated to regrow new axons in the spinal cord by the presence of denervated P.N.S.
SPINAL
NERVE
neurophysiological
ROOTS
IN BRACHIAL
PLEXUS
INJURY
tests. Scale bar 10 urn. (From Cullheim
et al.,
tissue in the implanted ventral root or by defects in the blood-brain barrier. Dorsal root axonal regeneration into the spinal cord of immature animals might depend on a greater regenerative capacity for immature than adult neurons, or immature glial cells might be more permissive to axonal elongation than in the adult. These are some of the neurobiological problems that should be explored in finding surgical treatment of dorsal root avulsion injuries.
Intramedullar , innervation
Fig. 6
VOL.
Schematic Neurology
drawings of possible and Neuroscience).
16B No. 5 DECEMBER
1991
routes for motoneurons
to reinnervate
the implanted
ventral
root (From
Risling
et al., 1991, Restorative
481
T. CARLSTEDT
References
Fig. 7
Light micrograph of a transverse section through the dorsal part of the spinal cord after dorsal root avulsion and implantation (arrows). Re-established connectivity from the root to the spinal cord cannot be demonstrated. Scale bar= 100 pm. (From Carlstedt, 1985, Neuroscience Letters, 55: 343348).
Fig. 8
Camera lucida drawing of H.R. P.-labelled dorsal root terminals in the superficial laminae of the dorsal horn after regeneration in the newborn rat. (From Carlstedt et al., 1987, Neuroscience Letters, 74: 1418).
Acknowledgements This review is based on studies supported by the Swedish Medical Research Council (project no. 6532) and funds from Karolinska Institutet. My sincere thanks to Dr HBkan LugnegBrd, Head of the Department of Handsurgery, Sabbatsbergs sjukhus, for support and criticism of this manuscript.
482
ALNOT, J. Y. (1983). Paralysie traumatique du plexus brachial de l’adulte. Encyclopedic Medico-Chirurgicale (Paris), 9: l-15. BRUSHART, T. M. and MESULAM, M. M. (1980). Alternation in connections between muscle and anterior horn motoneurons after peripheral nerve repair. Science, 208: 603-605. BURKE, R. E., LEVINE, D. N., SALCMAN, M. and TSAIRIS, P. (1974). Motor units in cat soleus muscle: physiological histochemical and morphological characteristics. Journal of Physiology, 238 : 503-514. CARLSTEDT, T. (1985a). Dorsal root innervation of spinal cord neurons after dorsal root implantation into the spinal cord of adult rats. Neuroscience Letters, 55 : 343-348. CARLSTEDT,T. (1985b). Regeneratingaxonsformnerveterminalsatastrocytes. Brain Research, 347: 188-191. CARLSTEDT, T. (1989). Reinnervation of the mammalian spinal cord after neonatal dorsal root crush. Journal of Neurocytology, 17: 335-350. CARLSTEDT, T., LINDA, H., CULLHEIM, S. and RISLING, M. (1986). Reinnervation of hindlimb muscles after ventral root avulsion and implantation in the lumbar spinal cord of the adult rat. Acta Physiologica Scandinavica, 128: 645-646. CARLSTEDT, T., DALSGAARD, C.-J. and MOLANDER, C. (1987). Regrowth of lesioned dorsal root nerve fibres into the spinal cord of neonatal rats. Neuroscience Letters, 74: 14-18. CULLHEIM, S., CARLSTEDT, T., LINDA, H., RISLING, M. and ULFHAKE, B. (1989). Motoneurons reinnervate skeletal muscle after ventral root implantation into the spinal cord of the cat. Neuroscience, 29: 725-733. DAVID, S. (1988). Neurite outgrowth from mammalian C.N.S. neurons on astrocytes in vitro may not be mediated primarily by laminin. Journal of Neurocytology, 17: 131-144. HORVAT, J-C., PECOT-DECHAVASSINE, M. and MIRA, J-C. (1987). Reinnervation fonctionelle d’un muscule squelettique du Rat adulte au moyen d’un greffon de nerf p&riph&que introduit dans la moelle Cpinibre par voi dorsale. Comptes rendus de I’Academie des Sciences (Paris), 304: 143-148. HOFFMANN, C. F. E., THOMEER, R. T. W. M. and MARANI, E. (1990). Ventral root avulsion of the cat spinal cord at the brachial plexus level (Cervical 7). European Journal of Morphology, 28: 418-429. JAMIESON, A. M. and EAMES, R. A. (1980). Reimplantation of avulsed brachial plexus rots: an experimental study in dogs. International Journal of Microsurgery, 2: 75-85. KIERNAN, J. A. and CONTESTABILE, A. (1980). Vascular permeability associated with axonal regeneration in the optic system of the goldfish. Acta Neuropathologica, 51: 39-45. MILEDI, R. and STEFANI, E. (1969). Nonselective re-innervation of slow and fast muscle fibres in the rat. Nature, 222: 569-571. NARAKAS, A. 0. (1984). Thoughts on neurotization or nerve transfers in irreparable nerve lesions. Clinics in Plastic Surgery, 11: 153-160. REIER, P. J., STENSAAS, L. J. and GUTH, L. The astrocytic scar as an impediment to regeneration in the central nervous system. In Cao, C. C., Bunge, R. P. and Reier, P. j. (Eds) Spinal Cord Reconstruction. Raven Press, NewYork, 1983: 163-195. RISLING, M., CULLHEIM, S. and HILDEBRAND, C. (1983). Reinnervation of the ventral root L7 from ventral horn neurons following intramedullary axotomy in adult cats. Brain Research, 280: 15-23. RISLING, M., DALSGAARD, C. J. and CUELLO, A. C. (1984). Invasion of lumbosacral ventral roots and spinal pia mater by substance P-immunoreactive axons after sciatic nerve lesion in kittens. Brain Research, 307: 351-354. RISLING, M., SJdGREN, LINDA, H., CULLHEIM, S., CARLSTEDT, T. and TELESTAM, M. Demonstration of a persistent defect in the bloodbrain barrier after ventral root implantation into the spinal cord using immunogold technique and staphylococcal alpha-toxin. In Johansson, B. B., Owman, C. and Widner, H. (Eds) Pathophysiology of the blood-brain barrier. Elsevier, 1990: 555-559. RISLING, M., CARLSTEDT, T., LINDA, H., and CULLHEIM, S. (1991). The pia mater-A conduit for regenerating axons after ventral root replant&ion. Restorative Neurology and Neuroscience, in press. SMITH, K. J., TERZIS, J. K., ERASMUS, M. and CARSON, K. A. Reinnervation of deneiiated skeletal muscle ‘by central nerve fibres regenerating along replanted ventral roots. In Gash, D. M. and Sladek Jr, J. R. (Eds) Transplantation into the Mammalian C.N.S. Progress in Brain Research, Vol78. Elsevier, 1988: 193-198. Thomas Carlstedt, M.D., Ph.D., Associate Professor, Department of Handsurgely, Sabbatsbergs Hospital, S-l 1382 Stockholm, Sweden. 0 1991 The British Society for Surgery of the Hand THE
JOURNAL
OF HAND
SURGERY
