Journal of Orthopaedic Research 9 5 4 6 0 Raven Press, Ltd., New York 0 1991 Orthopaedic Research Society
Rerouting Peripheral Nerves for Spinal Cord Lesions “Victor M. Romano, *Sidney J. Blair, and ?$Robert D. Wurster Departments of *Orthopaedics and Rehabilitation and ?Physiology, Loyola University Medical Center, Maywood, and $Rehabilitation Research and Development Center, Hines Veterans Administration Medical Center, Hines, Illinois, U.S.A.
Summary: To provide control of paralyzed limb muscles following spinal cord lesions, peripheral nerves containing motor axons from motoneurons located above a spinal cord lesion could potentially be rerouted to nerves containing motor axons located below the spinal cord lesion. To test this hypothesis in rats, the distal end of a cut tibial nerve, innervated by the L4-6 spinal level, was anastomosed or rerouted to the central end of the cut femoral nerve, innervated by the L 3 4 spinal level. Appropriate controls were used. Recovery of lower hind limb motor function was followed at regular intervals, measuring the twitch tension of toe flexion (innervated by the tibial nerve) induced by transcutaneous stimulation of nerve rootlets exiting the spinal cord. After 4-6 months, 54% of motor function returned in the experimental group. Retrograde transport of horseradish peroxidase from the gastrocnemius muscles to spinal motoneuron cell bodies confirmed that the innervation of this group was at a higher level. Furthermore, after an L4 spinal transection, twitch tension responses to spinal cord outlet stimulation remained only in the experimental group. Therefore, a peripheral nerve containing motor axons from above the lesion was rerouted to a distal peripheral nerve to control muscles that would have otherwise been denervated. Key Words: Spinal cord lesion-Nerve anastomosis-Spinal cord-Peripheral nerve.
lesions (13,14,21-23). In the same manner, why cannot a nerve innervated above a spinal cord lesion be used to innervate an organ below the spinal cord lesion? The femoral nerve contains motor axons from motoneurons located at the L 3 4 spinal cord levels and innervates the quadriceps femoris muscles (10,19). A branch of the sciatic nerve, the tibial nerve contains motor axons from the L4-6 levels and innervates all of the muscles in the posterior compartment of the lower limb. An L4 spinal cord lesion will denervate the entire lower leg, which is innervated by the tibial and common peroneal branches of the sciatic nerve. By connecting the femoral nerve to the tibial nerve, an L4 spinal lesion theoretically can be partially “bypassed,” returning motor function to the posterior compartment of the lower hind limb. The purpose of the present study
Following complete spinal cord lesion, voluntary control of all functions innervated below the lesion is permanently lost. Many attempts to “bridge” the spinal cord above and below the lesion, including using peripheral nerves or fetal neural tissue, have met with limited success (1,4,7,9,16,17,20). Another approach is to reroute an intact peripheral nerve innervated above the lesion to a peripheral nerve innervated below the lesion, thus providing potential control to muscles that otherwise would have remained paralyzed. Rerouting of peripheral nerves has been used to correct peripheral nerve ~
~~~~
Received January 18, 1990; accepted May 31, 1990. Address correspondence and reprint requests to Dr. V. M. Romano at Loyola University Medical Center, Dept. of Orthopaedics & Rehabilitation, 2160 S. First Ave., Maywood, IL 60153, U.S.A.
54
PERIPHERAL REROUTING FOR SPINAL LESIONS
was to test the feasibility of this model, which later might be used for studies of voluntary control following spinal cord lesion.
divided randomly into four treatment groups (Fig. 1). As a sham surgical group (sham surg group, N = 12), one group had the two exposed nerves left intact. In the remaining three groups, the tibial nerve was transected proximally and the femoral nerve was transected just before entering the quadriceps femoris. A second group had a femoral to tibial anastomosis or “bypass” (fem-tib BP group, N = 12). Here, the distal tibial nerve was tunneled under the adductor muscles and anastomosed to the proximal femoral nerve. A third group had the tibial nerve reanastomosed to itself (tib reanast group, N = 12). All anastomoses were made using two or three 10-0 nylon perineural stitches. A fourth group left the tibial nerves cut (cut tib group, N = 12). At the end of the surgery, the skin edges were reapproximated with wound staples. Recovery of tibial nerve motor function was assessed at biweekly intervals by using a modification
METHODS Forty-eight adult, male Sprague-Dawley rats (weighing 225-300 g) were used in this study and housed using standard principles of laboratory animal care. Prior to surgery, the rats were anesthetized with sodium pentobarbital (35 mg/kg, intraperitoneal). Using standard aseptic microsurgical techniques, the ipsilateral femoral and tibial nerves of all rats were exposed. The contralateral limbs were not operated on and used as controls. In order to obtain an adequate length, the tibial nerve fascicle was separated proximally away from the sciatic nerve to the gluteal nerve branch, by removing the epineurium of the sciatic nerve. The rats were then
FIG. 1. Diagrams of surgery for each experimental group. In group A (sham surg), the femoral and tibial nerves were exposed on one side, but left intact. In three other groups, these two nerves were transected. In group B (fem-tib BP),the distal tibial nerve was anastomosed to the proximal femoral nerve. In group C (tib reanast), the tibial nerve was repaired. In group D (cut tib), the nerves were left divided.
55
GROUP A
GROUP C
S h a m Surg
Tib Reanast
femoral n
f e m o r a l n.
sciatic n.
s c i a t i c n.
t i b i a l n.
t i b i a l n.
to FHL
i
L4
t o FHL
GROUP B
GROUP D
F e m - T i b BP
Cut Tib L4
f e m o r a l n.
sciatic
L4 f e m o r a l n.
n.
t i b i a l n.
t i b i a l n. t o FHL
fl t o FHL
J Orthop Res, Vol. 9, No. 1 , 1991
V . M . ROMANO ET AL.
56
of the twitch-tension method (11) (Fig. 2). Easily reproducible, this method noninvasively measures the strength of toe flexion, which correlates well with reinnervation of the tibia1 nerve. After induction of sodium pentobarbital anesthesia, the rat was placed on a platform with its foot stabilized to a sliding post using a wide rubber band. The middle toe was attached to a force transducer by means of a silk suture. In order to be consistent when comparing results of each side, the initial tension was adjusted to 30 g by moving the sliding post by means of a rack and pinion positioning device. Next, the spinal root outlets were supramaximally stimulated with percutaneous 18 gauge needle electrodes placed along the lumbar vertebrae (60 V, 1.2 ms duration, and 1.5 Hz). The force transducer was connected to a bridge amplifier (Grass, Quincy, MA, U.S.A.), and then to an averaging computer (Nuclear Chicago, Chicago, IL, U.S.A.). Eight consecutive twitches are averaged and plotted on an X-Y plotter (Houston Instruments, Houston, TX, U.S.A.). This method allowed for repetitive assessment throughout the recovery period without impairing recovery. Following testing of the twitch responses on both intact and operated side, the recovery index was determined by expressing the amplitude of the operated side as a percent of the amplitude of the nonoperated side. The mean recovery index in each group of animals was calculated. Significant differences were determined within the groups using one-way analysis of variance (ANOVA). Post hoc analysis of differences between means of these groups was determined by the Student-Neuman-Keuls test. Statistically significant differences were defined as p < 0.05. To confirm that innervation to the foot flexors of OSCILLOSCOPE
'v'AVERAGER +PLOTTER
OSCILLOGAPH
M
UNIT
FIG. 2. Twitch tension method for functional recovery. See the text for a description.
J Orthop Res, Vol. 9, No. 1, 1991
the fem-tib BP group comes from a higher level in the spinal cord when compared to the remaining groups, the gastrocnemius muscles of one-half of the rats in each group were injected bilaterally with 15 pl of 30% horseradish peroxidase (HRP, Worthington, St. Louis, MO, U.S.A.). Allowing 3 days for retrograde transport of the HRP to the motoneuron cell bodies in the spinal cord, the rats were anesthetized with sodium pentobarbital and then sacrificed by a transcardiac perfusion of saline followed by 1% paraformaldehyde and 1.25% glutaraldehyde in 0.1 M phosphate buffer at pH 7.4. Afterwards, the spinal cords were removed and placed in phosphate-buffered 30% sucrose solution. Serial longitudinal sections (50 pm) of the spinal cord were cut using a cryostat. Tissues were processed for HRP reaction products according to the method of Mesulam: reacted with tetramethyl benzidine, mounted on slides, and coverslipped (15). Camera lucida drawings were made of the spinal cord sections, superimposing the location of the labeled cell bodies. In the remaining animals of each group, the spinal cord was transected at the L4 spinal level. Allowing 1 week for recovery from spinal shock, the twitch tension was then reassessed for each animal by directly stimulating the spinal cord above and below the lesion. The recovery index was then calculated and compared in a similar manner as above. RESULTS
The rats in all groups tolerated surgery well, without any evidence of autophagia of the limbs (i.e., no toe chewing). Figure 3 demonstrates the amplitude of toe flexion from the twitch-tension method for each group of rats 4 to 6 months following surgery. In the sham surgery group, toe flexion of the control side was similar to that of the sham-operated side. In the tib reanast and fem-tib BP groups, toe flexion was approximately 80% and 50%, respectively, of the control side. Finally, in the cut tib group, only a small toe flexion response was recorded on the experimental side. Figure 4 shows the mean recovery of motor function for each group of rats over time as indicated by the recovery index. After 2-3 weeks, a steady increase in recovery index occurred in the fem-tip BP as well as tib reanast groups, reaching a plateau by about 50 to 60 days. A small increase in recovery index of the cut tib group was noted. After 4-6 months, recovery indices of motor function of the
PERIPHERAL REROUTING FOR SPINAL LESIONS
-
sham surg, tib reanast, fem-tib BP, and cut tib groups were 99 4% (SEM), 72 3%, 54 ? 4%, and 31 9%, respectively. The sham surg group was significantly different from all other groups,
*
*
loo]
57
*
and the fern-tib BP and tib reanast groups were significantly different from the cut tib group. Figure 5 depicts the longitudinal distribution of motoneurons retrogradely labeled by HRP reaction
Fam-Tlb BP
-l-
40
-
20
FIG. 4. Recovery index over time. Motor function was tested biweekly using the twitch tension method and recorded as the percent of the control side.
Cut. Tib.
-
-
J Orthop Res, Vol. 9, No. I, 1991
V . M . ROMANO ET AL.
58 A.
B.
SHAM SURG.
1
FEW-TIB. BP.
-
mm.
1 mm.
" . :. z iy;r 0
S.P. L l
a -
C
;:.,?:r
$ I : *
.:r
-
S.P. L Z
D.
TIB. REANAST.
-
CUT TIB.
-
right
-4--8.p. L 1
a
S.P. L Z
-
1 mm.
1 mm.
L
S.P. LZ
S.P. L 1
-0
_c_~
v
-
FIG. 5. HRP studies of the lumbar spinal cord. These are examples of typical spinal cords in each group after retrograde labeling of the cell bodies via 30% HRP injected bilaterally into the gastrocnemius muscles. In group B, innervation to the operated lower hindlimb was higher in the spinal cord than in groups A and C and absent in group D.
products following injection into the gastrocnemius muscles bilaterally. In the sham surg group as well as in the control sides of the other groups, cell bodies were located 2-3 mm caudal and rostra1 to the L2 spinal process (s.P. L2). This corresponds to the L4-6 spinal level (6,14). A similar distribution was observed following tibial nerve reanastomosis (Fig. 5C, right side of spinal cord). Severing the tibial nerve without reanastomosis (cut tib) resulted in no cell bodies on the experimental side (Fig. 5D, right side of spinal cord). However, anastomosis of the distal cut tibial nerve to the proximal femoral nerve resulted in a shifting of most cell bodies innervating the tibial nerve to the level of the L1 spinal process (s.P. L1)-the location of femoral nerve motoneurons, L3-4 (6,14). Although no quantitative analysis was done, the number of labeled motoneurons was qualitatively less in the fern-tib group than the tib reanast group. This is because the relative size of the femoral nerve is one-third of the tibial nerve; thus, less motor axons were available for distal reattachment. Six months after surgery when nerve regeneration was completed, an L4 spinal transection was performed. An example of the results from direct stimulation of the spinal cord above the lesion is
J Orthop Res, Vol. 9,No. I , 1991
shown in Fig. 6. The top two traces demonstrate that, except for a small artifact, toe flexion was absent on both limbs of the tib anast group. This was also found in the sham surg and cut tib groups. On the other hand, in the fem-tib BP group (the bottom two traces), toe flexion remained only on the operated limb. DISCUSSION
These experiments demonstrate that the lower hindlimb of a rat may be reinnervated by a peripheral nerve receiving innervation from higher spinal cord levels that normally innervates the upper portions of the hindlimb. Here, the femoral nerve, innervated by the L3-4 spinal cord, was transferred successfully to the tibial nerve (L4-6) with 54% of functional recovery. This is about 75% of functional recovery that we and others have seen with primary nerve reanastomoses (11). The difference may be, in part, because the femoral nerve is one-third the size of the tibial nerve. After confirming nerve regeneration via HRP studies, the spinal cord was sharply divided at L4 and motor function was found to remain only in the lower hindlimb of the operated fem-tib BP group
59
PERIPHERAL REROUTING FOR SPINAL LESIONS
Experimental
FIG. 6. Twitch tension arnplitudes after L4 transection. This is an example of how motor function remained only in the operated side of bypass group after the spinal cord was transected at the L4 level.
Control
Experimental
Control
and absent in the nonoperated limb of that group and in both limbs of the remaining groups. With a hypothesis similar to the present study but requiring partial hemilaminectomy , Malik and coworkers anastomosed an intercostal nerve to an L6 or L7 lumbar root in dogs (14). Their only functional data presented were that some of the dogs 8-11 months later “lost the limp that they had postoperatively.” They histologically examined the ipsilateral hindlimb muscles and noted in a few dogs that no microscopic abnormalities were seen except that the muscle fiber types seem to be altered. Vorstman et al. have unilaterally anastomosed the lumbar (L7) to a sacral root (Sl) in cats while leaving the contralateral side intact (24). Stimulation of spinal roots exiting the vertebral column resulted in contraction of the urinary bladder. However, they did not test any functional use of this bladder following spinal cord lesion or without the contralateral side. Peripheral anastomoses of different nerves have been used clinically to restore function usually following a peripheral nerve lesion. In 1932, for treatment of facial paralysis, Ballance and Duel anastomosed the proximal end of an intact hypoglossal nerve to the distal end of a damaged facial nerve (2). Later, anastomosis to the intercostal nerves has been used to repair brachial plexus (22,23) and lumbar root injuries (14,21). Also, the phrenic nerve has been used to reinnervate the brachial nerve (13). A few clinical studies have attempted to utilize a similar approach as used in the present report, to provide control in patients with spinal cord pathologies. In 1907, Kilvington proposed crossing so-
matic nerves to the pelvic nerve as a treatment of neurogenic bladder dysfunctions (12). Chiasserini and later Carlsson and Sudin had limited success in restoring normal bladder function in a few paraplegic patients (4-6). Benassy and Robart reported in a C5 quadriplegic patient that anastomosis of an intact musculocutaneous nerve to a paralyzed median nerve returned some of the function of the median nerve (3). Epstein et al. reported some improved function in children with meningomyelocele following anastomosis of intercostal and motor nerves to the lower limb (8). A similar approach was taken by Patil in one patient with traumatic paraplegia (18). Almost all of these clinical studies utilized crossing of nerves within the spinal canal, whereas our present experiments involved anastomosing of peripheral nerves outside of the vertebral column. The direct clinical application of the present results are limited for at least five reasons: (a) The L4 spinal lesion does not represent the most common spinal lesion in humans. (b) The spinal cord lesion was made after the nerves were allowed to regenerate. (c) The use of more proximal nerves might be more clinically useful. (d) Although the anastomosis may be able to restore function to some muscle innervated by the tibia1 nerve, it denervated some muscles innervated by the femoral nerve. (e) Function is determined by supramaximally stimulating the nerves as they exit from the vertebral column but the ability to voluntarily control the hindlimb using the rerouted nerve was not tested. However, this model demonstrates that spinal cord lesions can be partially bypassed by rerouting peripheral nerves.
J Orthop Res, Vol. 9, No. 1. 1991
V. M.ROMANO ET AL.
60
Recovery of sensation and motor control in this model needs to be studied. If the central nervous system can learn to control a nerve innervating a different organ or limb than it previously had innervated, then peripheral nerve “rerouting” might become a useful technique to provide control for some spinal patients. This restored control would not be expected to be as fine as the normal control, but even a somewhat cruder control might be quite useful, e.g., to obtain sphincter control. In conclusion, an L4 spinal cord lesion can be “bypassed” by rerouting a peripheral nerve innervated above the lesion (femoral nerve, L2-3) to a peripheral nerve innervated below the lesion (tibia1 nerve, L4-6). REFERENCES 1 . Aguayo AJ, David S, Bray GM: Influences of the glial environment on the elongation of axons after injury: transplantation studies in adult rodents. J Exp Biol98:231-240, 1981 2. Ballance C, Duel AB: The operative treatment of facial palsy by the introduction of nerve grafts into the fallopian canal and by other intratemporal methods. Arch Otoluryngol 15:l70, 1952 3. Benassy J, Robart J: Un cas dC transposition du nerf musculocutanC sur le nerf mtdian. Ann Chir Plust 11:187-189, 1966 4. Carlsson CA, Sundin T: Reconstruction of efferent pathways to the urinary bladder in a paraplegic child. Rev Surg 24:73-76, 1967 5. Carlsson GA, Sundin T: Reconstruction of afferent and efferent nervous pathways to the urinary bladder in two paraplegic patients. Spine 5:3741, 1980 6. Chiasserini A: L’anastomose intercosto-radiculaire dans les traumatismes vertebraux avec section de la moelle lombaire. J Chir 46:5468, 1935 7. David S, Aguayo AJ: Axonal elongation into peripheral nervous system “bridges” after central nervous system injury in adult rats. Science 214:931-933, 1981 8. Epstein F, Spielholz N, Battista A, McCarthy J: Delayed
J Orthop Res, Vol. 9, No. 1, 1991
cauda equina reconstruction in meningomyelocele: preliminary report. Neurosurgery 6:540-541, 1980 9. Fernandez, Pallini R, Maira G, Rossi GF: Peripheral nerve autografts to the injured spinal cord of the rat: An experimental model for the study of spinal cord regeneration. Actu Neurochir 7 8 5 7 4 4 , 1985 10. Greene EC: Anatomy ofthe Rut, New York, Hafner, 1959 11. Kerns JM, Fakhouri AJ, Pavkovic IM: A twitch tension method to assess motor nerve function. J Neurosci Methods 191217-223, 1987 12. Kilvington B: An investigation in the regeneration of nerves with regard to surgical treatment of certain paralyses. Br Med J 1:988-990, 1907 13. Krieger AJ, Danetez I, Wu SZ, Spatola M, Sapru HN: Electrophrenic respiration following anastomosis of phrenic with brachial nerve in the cat. J Neurosurg 59:262-267, 1983 14. Malik HG, Buhr AJ: Intercostal nerve transfer to lumbar roots-Part I: Development of an animal model and cadaver studies. Spine 4:410-415, 1979 15. Mesulam M: Tetramethylbenzidine for horseradish peroxidase neurohistochemistry: A non-carcinogenic blue reaction product with superior sensitivity for visualizing neural afferents and efferents. J Histiochem 26:106-107, 1978 16. Munz M, Rasminsky M, Aguayo AJ, Vidal-sanz M, Devor M: Functional activity of the rat brainstem neurons regenerating axons along peripheral nerve grafts. Bruin Res 340: 115-125, 1985 17. Pate1 U, Bernstein JJ: Growth, differentiation, and viability of fetal rat cortical and spinal cord implants into adult rat spinal cord. J Neurosci Res 9:303-310, 1983 18. Patil A: Intercostal nerves to spinal nerve roots anastomosis (spinal cord bypass) and Hamngton rod fusion in traumatic paralgesia-technical note. Actu Neurochir 57:16-20, 1981 19. Peyronnard JM, Charron LF, Lavoie J, Messier JP: Motor, sympathetic and sensory innervation of rat skeletal muscles. Bruin Res 373:288-302, 1986 20. Reier PJ: Neural tissue grafts and repair of the injured spinal cord. Neuroputhol Appl Neurobiol 11:81-104, 1985 21. Sangaland VE, Buhr AJ, Malik HG: Intercostal nerve transfer to lumbar roots-Part 11: Neuropathological findings in the animal model. Spine 4:416-422, 1979 22. Seddon HJ: Nerve grafting. J Bone Joint Surg [Br] 45:447461, 1963 23. Tomita Y, Tsai T, Bums JT, Karaoguz A, Ogden L: Intercostal nerve transfer in brachial plexus injuries: An experimental study. Microsurgery 4:95-104, 1983 24. Vorstman B, Schlossberg SM, Kass L, Devine C: Urinary bladder reinnervation. J Urol 136:964-969, 1986
Rerouting Peripheral Nerves for Spinal Cord Lesions “Victor M. Romano, *Sidney J. Blair, and ?$Robert D. Wurster Departments of *Orthopaedics and Rehabilitation and ?Physiology, Loyola University Medical Center, Maywood, and $Rehabilitation Research and Development Center, Hines Veterans Administration Medical Center, Hines, Illinois, U.S.A.
Summary: To provide control of paralyzed limb muscles following spinal cord lesions, peripheral nerves containing motor axons from motoneurons located above a spinal cord lesion could potentially be rerouted to nerves containing motor axons located below the spinal cord lesion. To test this hypothesis in rats, the distal end of a cut tibial nerve, innervated by the L4-6 spinal level, was anastomosed or rerouted to the central end of the cut femoral nerve, innervated by the L 3 4 spinal level. Appropriate controls were used. Recovery of lower hind limb motor function was followed at regular intervals, measuring the twitch tension of toe flexion (innervated by the tibial nerve) induced by transcutaneous stimulation of nerve rootlets exiting the spinal cord. After 4-6 months, 54% of motor function returned in the experimental group. Retrograde transport of horseradish peroxidase from the gastrocnemius muscles to spinal motoneuron cell bodies confirmed that the innervation of this group was at a higher level. Furthermore, after an L4 spinal transection, twitch tension responses to spinal cord outlet stimulation remained only in the experimental group. Therefore, a peripheral nerve containing motor axons from above the lesion was rerouted to a distal peripheral nerve to control muscles that would have otherwise been denervated. Key Words: Spinal cord lesion-Nerve anastomosis-Spinal cord-Peripheral nerve.
lesions (13,14,21-23). In the same manner, why cannot a nerve innervated above a spinal cord lesion be used to innervate an organ below the spinal cord lesion? The femoral nerve contains motor axons from motoneurons located at the L 3 4 spinal cord levels and innervates the quadriceps femoris muscles (10,19). A branch of the sciatic nerve, the tibial nerve contains motor axons from the L4-6 levels and innervates all of the muscles in the posterior compartment of the lower limb. An L4 spinal cord lesion will denervate the entire lower leg, which is innervated by the tibial and common peroneal branches of the sciatic nerve. By connecting the femoral nerve to the tibial nerve, an L4 spinal lesion theoretically can be partially “bypassed,” returning motor function to the posterior compartment of the lower hind limb. The purpose of the present study
Following complete spinal cord lesion, voluntary control of all functions innervated below the lesion is permanently lost. Many attempts to “bridge” the spinal cord above and below the lesion, including using peripheral nerves or fetal neural tissue, have met with limited success (1,4,7,9,16,17,20). Another approach is to reroute an intact peripheral nerve innervated above the lesion to a peripheral nerve innervated below the lesion, thus providing potential control to muscles that otherwise would have remained paralyzed. Rerouting of peripheral nerves has been used to correct peripheral nerve ~
~~~~
Received January 18, 1990; accepted May 31, 1990. Address correspondence and reprint requests to Dr. V. M. Romano at Loyola University Medical Center, Dept. of Orthopaedics & Rehabilitation, 2160 S. First Ave., Maywood, IL 60153, U.S.A.
54
PERIPHERAL REROUTING FOR SPINAL LESIONS
was to test the feasibility of this model, which later might be used for studies of voluntary control following spinal cord lesion.
divided randomly into four treatment groups (Fig. 1). As a sham surgical group (sham surg group, N = 12), one group had the two exposed nerves left intact. In the remaining three groups, the tibial nerve was transected proximally and the femoral nerve was transected just before entering the quadriceps femoris. A second group had a femoral to tibial anastomosis or “bypass” (fem-tib BP group, N = 12). Here, the distal tibial nerve was tunneled under the adductor muscles and anastomosed to the proximal femoral nerve. A third group had the tibial nerve reanastomosed to itself (tib reanast group, N = 12). All anastomoses were made using two or three 10-0 nylon perineural stitches. A fourth group left the tibial nerves cut (cut tib group, N = 12). At the end of the surgery, the skin edges were reapproximated with wound staples. Recovery of tibial nerve motor function was assessed at biweekly intervals by using a modification
METHODS Forty-eight adult, male Sprague-Dawley rats (weighing 225-300 g) were used in this study and housed using standard principles of laboratory animal care. Prior to surgery, the rats were anesthetized with sodium pentobarbital (35 mg/kg, intraperitoneal). Using standard aseptic microsurgical techniques, the ipsilateral femoral and tibial nerves of all rats were exposed. The contralateral limbs were not operated on and used as controls. In order to obtain an adequate length, the tibial nerve fascicle was separated proximally away from the sciatic nerve to the gluteal nerve branch, by removing the epineurium of the sciatic nerve. The rats were then
FIG. 1. Diagrams of surgery for each experimental group. In group A (sham surg), the femoral and tibial nerves were exposed on one side, but left intact. In three other groups, these two nerves were transected. In group B (fem-tib BP),the distal tibial nerve was anastomosed to the proximal femoral nerve. In group C (tib reanast), the tibial nerve was repaired. In group D (cut tib), the nerves were left divided.
55
GROUP A
GROUP C
S h a m Surg
Tib Reanast
femoral n
f e m o r a l n.
sciatic n.
s c i a t i c n.
t i b i a l n.
t i b i a l n.
to FHL
i
L4
t o FHL
GROUP B
GROUP D
F e m - T i b BP
Cut Tib L4
f e m o r a l n.
sciatic
L4 f e m o r a l n.
n.
t i b i a l n.
t i b i a l n. t o FHL
fl t o FHL
J Orthop Res, Vol. 9, No. 1 , 1991
V . M . ROMANO ET AL.
56
of the twitch-tension method (11) (Fig. 2). Easily reproducible, this method noninvasively measures the strength of toe flexion, which correlates well with reinnervation of the tibia1 nerve. After induction of sodium pentobarbital anesthesia, the rat was placed on a platform with its foot stabilized to a sliding post using a wide rubber band. The middle toe was attached to a force transducer by means of a silk suture. In order to be consistent when comparing results of each side, the initial tension was adjusted to 30 g by moving the sliding post by means of a rack and pinion positioning device. Next, the spinal root outlets were supramaximally stimulated with percutaneous 18 gauge needle electrodes placed along the lumbar vertebrae (60 V, 1.2 ms duration, and 1.5 Hz). The force transducer was connected to a bridge amplifier (Grass, Quincy, MA, U.S.A.), and then to an averaging computer (Nuclear Chicago, Chicago, IL, U.S.A.). Eight consecutive twitches are averaged and plotted on an X-Y plotter (Houston Instruments, Houston, TX, U.S.A.). This method allowed for repetitive assessment throughout the recovery period without impairing recovery. Following testing of the twitch responses on both intact and operated side, the recovery index was determined by expressing the amplitude of the operated side as a percent of the amplitude of the nonoperated side. The mean recovery index in each group of animals was calculated. Significant differences were determined within the groups using one-way analysis of variance (ANOVA). Post hoc analysis of differences between means of these groups was determined by the Student-Neuman-Keuls test. Statistically significant differences were defined as p < 0.05. To confirm that innervation to the foot flexors of OSCILLOSCOPE
'v'AVERAGER +PLOTTER
OSCILLOGAPH
M
UNIT
FIG. 2. Twitch tension method for functional recovery. See the text for a description.
J Orthop Res, Vol. 9, No. 1, 1991
the fem-tib BP group comes from a higher level in the spinal cord when compared to the remaining groups, the gastrocnemius muscles of one-half of the rats in each group were injected bilaterally with 15 pl of 30% horseradish peroxidase (HRP, Worthington, St. Louis, MO, U.S.A.). Allowing 3 days for retrograde transport of the HRP to the motoneuron cell bodies in the spinal cord, the rats were anesthetized with sodium pentobarbital and then sacrificed by a transcardiac perfusion of saline followed by 1% paraformaldehyde and 1.25% glutaraldehyde in 0.1 M phosphate buffer at pH 7.4. Afterwards, the spinal cords were removed and placed in phosphate-buffered 30% sucrose solution. Serial longitudinal sections (50 pm) of the spinal cord were cut using a cryostat. Tissues were processed for HRP reaction products according to the method of Mesulam: reacted with tetramethyl benzidine, mounted on slides, and coverslipped (15). Camera lucida drawings were made of the spinal cord sections, superimposing the location of the labeled cell bodies. In the remaining animals of each group, the spinal cord was transected at the L4 spinal level. Allowing 1 week for recovery from spinal shock, the twitch tension was then reassessed for each animal by directly stimulating the spinal cord above and below the lesion. The recovery index was then calculated and compared in a similar manner as above. RESULTS
The rats in all groups tolerated surgery well, without any evidence of autophagia of the limbs (i.e., no toe chewing). Figure 3 demonstrates the amplitude of toe flexion from the twitch-tension method for each group of rats 4 to 6 months following surgery. In the sham surgery group, toe flexion of the control side was similar to that of the sham-operated side. In the tib reanast and fem-tib BP groups, toe flexion was approximately 80% and 50%, respectively, of the control side. Finally, in the cut tib group, only a small toe flexion response was recorded on the experimental side. Figure 4 shows the mean recovery of motor function for each group of rats over time as indicated by the recovery index. After 2-3 weeks, a steady increase in recovery index occurred in the fem-tip BP as well as tib reanast groups, reaching a plateau by about 50 to 60 days. A small increase in recovery index of the cut tib group was noted. After 4-6 months, recovery indices of motor function of the
PERIPHERAL REROUTING FOR SPINAL LESIONS
-
sham surg, tib reanast, fem-tib BP, and cut tib groups were 99 4% (SEM), 72 3%, 54 ? 4%, and 31 9%, respectively. The sham surg group was significantly different from all other groups,
*
*
loo]
57
*
and the fern-tib BP and tib reanast groups were significantly different from the cut tib group. Figure 5 depicts the longitudinal distribution of motoneurons retrogradely labeled by HRP reaction
Fam-Tlb BP
-l-
40
-
20
FIG. 4. Recovery index over time. Motor function was tested biweekly using the twitch tension method and recorded as the percent of the control side.
Cut. Tib.
-
-
J Orthop Res, Vol. 9, No. I, 1991
V . M . ROMANO ET AL.
58 A.
B.
SHAM SURG.
1
FEW-TIB. BP.
-
mm.
1 mm.
" . :. z iy;r 0
S.P. L l
a -
C
;:.,?:r
$ I : *
.:r
-
S.P. L Z
D.
TIB. REANAST.
-
CUT TIB.
-
right
-4--8.p. L 1
a
S.P. L Z
-
1 mm.
1 mm.
L
S.P. LZ
S.P. L 1
-0
_c_~
v
-
FIG. 5. HRP studies of the lumbar spinal cord. These are examples of typical spinal cords in each group after retrograde labeling of the cell bodies via 30% HRP injected bilaterally into the gastrocnemius muscles. In group B, innervation to the operated lower hindlimb was higher in the spinal cord than in groups A and C and absent in group D.
products following injection into the gastrocnemius muscles bilaterally. In the sham surg group as well as in the control sides of the other groups, cell bodies were located 2-3 mm caudal and rostra1 to the L2 spinal process (s.P. L2). This corresponds to the L4-6 spinal level (6,14). A similar distribution was observed following tibial nerve reanastomosis (Fig. 5C, right side of spinal cord). Severing the tibial nerve without reanastomosis (cut tib) resulted in no cell bodies on the experimental side (Fig. 5D, right side of spinal cord). However, anastomosis of the distal cut tibial nerve to the proximal femoral nerve resulted in a shifting of most cell bodies innervating the tibial nerve to the level of the L1 spinal process (s.P. L1)-the location of femoral nerve motoneurons, L3-4 (6,14). Although no quantitative analysis was done, the number of labeled motoneurons was qualitatively less in the fern-tib group than the tib reanast group. This is because the relative size of the femoral nerve is one-third of the tibial nerve; thus, less motor axons were available for distal reattachment. Six months after surgery when nerve regeneration was completed, an L4 spinal transection was performed. An example of the results from direct stimulation of the spinal cord above the lesion is
J Orthop Res, Vol. 9,No. I , 1991
shown in Fig. 6. The top two traces demonstrate that, except for a small artifact, toe flexion was absent on both limbs of the tib anast group. This was also found in the sham surg and cut tib groups. On the other hand, in the fem-tib BP group (the bottom two traces), toe flexion remained only on the operated limb. DISCUSSION
These experiments demonstrate that the lower hindlimb of a rat may be reinnervated by a peripheral nerve receiving innervation from higher spinal cord levels that normally innervates the upper portions of the hindlimb. Here, the femoral nerve, innervated by the L3-4 spinal cord, was transferred successfully to the tibial nerve (L4-6) with 54% of functional recovery. This is about 75% of functional recovery that we and others have seen with primary nerve reanastomoses (11). The difference may be, in part, because the femoral nerve is one-third the size of the tibial nerve. After confirming nerve regeneration via HRP studies, the spinal cord was sharply divided at L4 and motor function was found to remain only in the lower hindlimb of the operated fem-tib BP group
59
PERIPHERAL REROUTING FOR SPINAL LESIONS
Experimental
FIG. 6. Twitch tension arnplitudes after L4 transection. This is an example of how motor function remained only in the operated side of bypass group after the spinal cord was transected at the L4 level.
Control
Experimental
Control
and absent in the nonoperated limb of that group and in both limbs of the remaining groups. With a hypothesis similar to the present study but requiring partial hemilaminectomy , Malik and coworkers anastomosed an intercostal nerve to an L6 or L7 lumbar root in dogs (14). Their only functional data presented were that some of the dogs 8-11 months later “lost the limp that they had postoperatively.” They histologically examined the ipsilateral hindlimb muscles and noted in a few dogs that no microscopic abnormalities were seen except that the muscle fiber types seem to be altered. Vorstman et al. have unilaterally anastomosed the lumbar (L7) to a sacral root (Sl) in cats while leaving the contralateral side intact (24). Stimulation of spinal roots exiting the vertebral column resulted in contraction of the urinary bladder. However, they did not test any functional use of this bladder following spinal cord lesion or without the contralateral side. Peripheral anastomoses of different nerves have been used clinically to restore function usually following a peripheral nerve lesion. In 1932, for treatment of facial paralysis, Ballance and Duel anastomosed the proximal end of an intact hypoglossal nerve to the distal end of a damaged facial nerve (2). Later, anastomosis to the intercostal nerves has been used to repair brachial plexus (22,23) and lumbar root injuries (14,21). Also, the phrenic nerve has been used to reinnervate the brachial nerve (13). A few clinical studies have attempted to utilize a similar approach as used in the present report, to provide control in patients with spinal cord pathologies. In 1907, Kilvington proposed crossing so-
matic nerves to the pelvic nerve as a treatment of neurogenic bladder dysfunctions (12). Chiasserini and later Carlsson and Sudin had limited success in restoring normal bladder function in a few paraplegic patients (4-6). Benassy and Robart reported in a C5 quadriplegic patient that anastomosis of an intact musculocutaneous nerve to a paralyzed median nerve returned some of the function of the median nerve (3). Epstein et al. reported some improved function in children with meningomyelocele following anastomosis of intercostal and motor nerves to the lower limb (8). A similar approach was taken by Patil in one patient with traumatic paraplegia (18). Almost all of these clinical studies utilized crossing of nerves within the spinal canal, whereas our present experiments involved anastomosing of peripheral nerves outside of the vertebral column. The direct clinical application of the present results are limited for at least five reasons: (a) The L4 spinal lesion does not represent the most common spinal lesion in humans. (b) The spinal cord lesion was made after the nerves were allowed to regenerate. (c) The use of more proximal nerves might be more clinically useful. (d) Although the anastomosis may be able to restore function to some muscle innervated by the tibia1 nerve, it denervated some muscles innervated by the femoral nerve. (e) Function is determined by supramaximally stimulating the nerves as they exit from the vertebral column but the ability to voluntarily control the hindlimb using the rerouted nerve was not tested. However, this model demonstrates that spinal cord lesions can be partially bypassed by rerouting peripheral nerves.
J Orthop Res, Vol. 9, No. 1. 1991
V. M.ROMANO ET AL.
60
Recovery of sensation and motor control in this model needs to be studied. If the central nervous system can learn to control a nerve innervating a different organ or limb than it previously had innervated, then peripheral nerve “rerouting” might become a useful technique to provide control for some spinal patients. This restored control would not be expected to be as fine as the normal control, but even a somewhat cruder control might be quite useful, e.g., to obtain sphincter control. In conclusion, an L4 spinal cord lesion can be “bypassed” by rerouting a peripheral nerve innervated above the lesion (femoral nerve, L2-3) to a peripheral nerve innervated below the lesion (tibia1 nerve, L4-6). REFERENCES 1 . Aguayo AJ, David S, Bray GM: Influences of the glial environment on the elongation of axons after injury: transplantation studies in adult rodents. J Exp Biol98:231-240, 1981 2. Ballance C, Duel AB: The operative treatment of facial palsy by the introduction of nerve grafts into the fallopian canal and by other intratemporal methods. Arch Otoluryngol 15:l70, 1952 3. Benassy J, Robart J: Un cas dC transposition du nerf musculocutanC sur le nerf mtdian. Ann Chir Plust 11:187-189, 1966 4. Carlsson CA, Sundin T: Reconstruction of efferent pathways to the urinary bladder in a paraplegic child. Rev Surg 24:73-76, 1967 5. Carlsson GA, Sundin T: Reconstruction of afferent and efferent nervous pathways to the urinary bladder in two paraplegic patients. Spine 5:3741, 1980 6. Chiasserini A: L’anastomose intercosto-radiculaire dans les traumatismes vertebraux avec section de la moelle lombaire. J Chir 46:5468, 1935 7. David S, Aguayo AJ: Axonal elongation into peripheral nervous system “bridges” after central nervous system injury in adult rats. Science 214:931-933, 1981 8. Epstein F, Spielholz N, Battista A, McCarthy J: Delayed
J Orthop Res, Vol. 9, No. 1, 1991
cauda equina reconstruction in meningomyelocele: preliminary report. Neurosurgery 6:540-541, 1980 9. Fernandez, Pallini R, Maira G, Rossi GF: Peripheral nerve autografts to the injured spinal cord of the rat: An experimental model for the study of spinal cord regeneration. Actu Neurochir 7 8 5 7 4 4 , 1985 10. Greene EC: Anatomy ofthe Rut, New York, Hafner, 1959 11. Kerns JM, Fakhouri AJ, Pavkovic IM: A twitch tension method to assess motor nerve function. J Neurosci Methods 191217-223, 1987 12. Kilvington B: An investigation in the regeneration of nerves with regard to surgical treatment of certain paralyses. Br Med J 1:988-990, 1907 13. Krieger AJ, Danetez I, Wu SZ, Spatola M, Sapru HN: Electrophrenic respiration following anastomosis of phrenic with brachial nerve in the cat. J Neurosurg 59:262-267, 1983 14. Malik HG, Buhr AJ: Intercostal nerve transfer to lumbar roots-Part I: Development of an animal model and cadaver studies. Spine 4:410-415, 1979 15. Mesulam M: Tetramethylbenzidine for horseradish peroxidase neurohistochemistry: A non-carcinogenic blue reaction product with superior sensitivity for visualizing neural afferents and efferents. J Histiochem 26:106-107, 1978 16. Munz M, Rasminsky M, Aguayo AJ, Vidal-sanz M, Devor M: Functional activity of the rat brainstem neurons regenerating axons along peripheral nerve grafts. Bruin Res 340: 115-125, 1985 17. Pate1 U, Bernstein JJ: Growth, differentiation, and viability of fetal rat cortical and spinal cord implants into adult rat spinal cord. J Neurosci Res 9:303-310, 1983 18. Patil A: Intercostal nerves to spinal nerve roots anastomosis (spinal cord bypass) and Hamngton rod fusion in traumatic paralgesia-technical note. Actu Neurochir 57:16-20, 1981 19. Peyronnard JM, Charron LF, Lavoie J, Messier JP: Motor, sympathetic and sensory innervation of rat skeletal muscles. Bruin Res 373:288-302, 1986 20. Reier PJ: Neural tissue grafts and repair of the injured spinal cord. Neuroputhol Appl Neurobiol 11:81-104, 1985 21. Sangaland VE, Buhr AJ, Malik HG: Intercostal nerve transfer to lumbar roots-Part 11: Neuropathological findings in the animal model. Spine 4:416-422, 1979 22. Seddon HJ: Nerve grafting. J Bone Joint Surg [Br] 45:447461, 1963 23. Tomita Y, Tsai T, Bums JT, Karaoguz A, Ogden L: Intercostal nerve transfer in brachial plexus injuries: An experimental study. Microsurgery 4:95-104, 1983 24. Vorstman B, Schlossberg SM, Kass L, Devine C: Urinary bladder reinnervation. J Urol 136:964-969, 1986
