Functional Connectivity of theTransected Brachial Plexus After

Intercostal Neurotization in Monkeys

H. CHENG,1,2* H.M. SHOUNG,2 Z.A. WU,2 K.C. CHEN,3 AND L.S. LEE2

1Department of Neuroscience, Karolinska Institute, S-171 77 Stockholm, Sweden2Neurologic Institute, Veterans General Hospital-Taipei, Taiwan3Division of Anatomy, National Yang-Ming University, Taiwan

ABSTRACTMicrosurgical reconstructions of brachial plexuses were performed on twelve monkeys by

using ipsilateral intercostal nerves (T3–9). Reinnervation in individual nerves was evaluatedmonthly by observations of neuromuscular and electromyographic improvements. Theelectromyographic studies revealed reappearance of motor unit potentials. According to amotor scale ranging from 0 to 4, the mean muscle power 6 months after operation improved to2.75 in the deltoid muscles, 2 in the biceps muscles, 1.22 in the triceps muscles, 1.13 in theflexor carpi radialis muscles, and 1.6 in the intrinsic muscles of the hands. Retrogradetransport of horseradish peroxidase (HRP) from the neuromuscular junctions of the recon-structed musculocutaneous nerves 6 months after complete brachial plexus lesion in fouranimals demonstrated HRP-labeled neurons in the anterior horns, spinal ganglia andsympathetic ganglia of the thoracic spinal cords. It suggested that the regenerated afferentand efferent circuits in the thoracic cords innervating the transected brachial plexuses wereable to generate the movements in the paralyzed upper limbs. However, as evidenced by thebehavior patterns and the fact that retrograde-labeled neurons were all found in the thoraciccords, the novel movements observed in the reconstructed brachial plexuses were insynchrony with respiration. These results suggested that the plasticity of central neuralnetworks is limited between two widely separated areas, such as between the midcervical andmidthoracic motor cortical areas in the present studies, and therefore, the efforts toreconstruct neural networks, both centrally and peripherally, should aim at rebuildingsituations as nearly to the original status as possible. J. Comp. Neurol. 380:155–163, 1997.r 1997 Wiley-Liss, Inc.

Indexing terms: spinal nerve; nerve regeneration; behavior; horseradish peroxidase;

electromyogram

Root avulsion of the brachial plexus due to traumaticstretching of the plexus poses a difficult surgical reconstruc-tion problem. This type of nerve injury is not treated todaybecause it is considered a type of central nervous systeminjury not amenable to surgery (Carlstedt, 1995). Thedivided rootlets at the point of connection with the spinalcord, i.e., the ventral rootlet entry zone (VREZ) and thedorsal rootlet entry zone (DREZ), may withdraw to a largeextent (before the decision to operate has been madeseveral weeks after the injury), making direct repair byrepositioning difficult (Narrakas, 1987). Many previousstudies (in humans, nonhuman primates, and other mam-mals) have shown that sectioned peripheral axons of onenerve can regenerate through foreign nerves to reinner-vate different motor or sensory fields. In clinical practice,reinnervation of the distal portion of injured brachial

plexus is possible by anastomosis with neighboring nerves.Such reconstruction was first carried out by Seddon, whoin 1961 reinnervated the biceps and the brachialis musclesby anastomosing the distal part of the musculocutaneousnerve with the second, third and fourth intercostal nerves(Yeoman and Seddon, 1961). This ‘‘intercostal neurotiza-tion’’ was further investigated by several authors (Allieu,1977; Millesi, 1977; Sedel, 1982; Narakas, 1987). Neuroti-zation with other nerves, including spinal accessory nerves

Contract grant sponsor: Taiwan Chin-Lin Medical Research Fund;Contract grant number: NSC77-0412-B075-17.*Correspondence to: Henrich Cheng, M.D., Ph.D., Department of Neuro-

science, Karolinska Institute, S-171 77 Stockholm, Sweden.E-mail: Henrich.Cheng@neuro.ki.seReceived 7 May 1996; Revised 30August 1996; Accepted 27 October 1996

THE JOURNAL OF COMPARATIVE NEUROLOGY 380:155–163 (1997)

r 1997 WILEY-LISS, INC.

and anterior nerves of the cervical plexus, has also beenreported to lead to some positive clinical results (Kotani etal., 1972; Brunelli, 1980). However, independent voluntaryelbow flexion was rarely reported even after long-termfollow-up. Despite experience accumulated over manyyears, these types of surgery are still in an experimentalphase and suffer from lack of fundamental knowledge(Narakas, 1987). The overall results are inconclusive and auniversally adopted code of management is lacking. It isthus necessary to explore possible technical improvementsand additional therapeutic modalities, as well as to gainmore knowledge regarding the central integration of func-tions restored in the upper limb by the adopted neuralcircuits which, in this case, belong to the intercostalnerves.In the present studies, functional connectivity of the

reconstructed brachial plexus after intercostal neurotiza-tion was studied by behavioral and electrophysiologicalmonitoring (Hallin et al., 1981; Raimbault, 1987; Rowland,1991), as well as by using horeradish peroxidase (HRP)retrograde tracing (Halperin, 1975; Geisert, 1976; Old-field, 1977; Heimer, 1981, 1989; Mesulam, 1982). We thusattempted to generate a neuroanatomical background forintercostal neurotization which might help develop betterfunctional recovery with additional treatments in thefuture. Application of a fascicular suture technique on allintercostal-brachial nerve anastomoses was also tried toobtain more precise point-to-point approximation. Finally,the reestablished neural network will be discussed.

MATERIALS AND METHODS

Animals

Twelve adult monkeys (Macaca mulatta, 3–6 kg) ofeither sex were used. The animals were intubated andanesthetized with an intravenous injection of pentobarbi-tone sodium (30mg/kg) after premedication with ketamine(25 mg/kg) subcutaneously. If pneumothorax occurredduring the dissection of intercostal nerves, the lungs wereinflated followed by removal of the air in the thoraciccavity by using a Fr. 10 nasogastric tube. The animalswere operated on a heating pad. Rectal temperature wasmonitored and maintained no less than 3°C below normal.Bipolar electrocauterization was used to minimize thetotal amount of bleeding. Arterial blood gas analyses wereperformed before and after the operation.Antibiotics (Oxa-cillin 50 mg/kg i.m. and Gentamicin 2 mg/kg i.m.) weregiven twice daily for one week; infections were not seen.Animals were kept in individual cages with food and waterad lib. Monkeys were chosen for these experiments due tothe greater feasibility of carrying out the microsurgicalanastomosis with intercostal nerves and the possibilitiesfor carrying out detailed functional evaluation of hands.The experiments have been approved by the NationalCommittee for the Use of Experimental Animals for Medi-cal Purposes in NSC, Taipei.

Intercostal neurotization of the transectedbrachial plexus

Under aseptic conditions, the left brachial plexus wasexposed with an incision along the lateral margin of thepectoris major muscle. The individual nerves of the bra-chial plexus were located, identified and assured by intra-

operative motor nerve stimulation and recording. Thebrachial plexus was transected at the proximal junctionwith the lateral, posterior and medial cords, respectively.Right brachial plexuses were preserved as controls. Theintercostal muscles were incised parallel to the ribs for 5–6cm with meticulous hemostasis. Dissection was performedalong the plane above the parietal pleura to free as long alength of the intercostal nerves as was possible. Understereomicroscopical guidance, the distal nerves of thetransected brachial plexus were anastomosed by finesutures (10-0 prolene) with several of the intercostalnerves, using a fascicular suture technique (Fig. 1).Six monkeys underwent complete brachial plexus tran-

section and intercostal neurotization. The other six under-went transection in one or several nerves in the brachialplexus and intercostal neurotization. In terms of indi-vidual nerves, 8 axillary, 8 musculocutaneous, 9 radial, 8median and 10 ulnar nerves underwent microsurgicalreconstruction with the intercostal nerves.The distances from the axilla to the base of the digit

were 12–15 cm in our animals. Given normal regenerationrates of about 1–2 mm/day, it was decided to follow theoutcome for up to 6 months after operation to allowpossible reinnervation of the hand muscles.

Behavioral studies

Neurologic behavior after operation was evaluatedmonthly by recording muscle power. A hanging bananawas introduced into the cage after having fastened theanimal’s right upper limb (control side) to the trunk.Movement of the left upper limb was then evaluated. Foranimals subjected to incomplete brachial plexus lesion,data from the muscles relevant to the injuried nerves wererecorded. Muscle power of each muscle was graded as: 4 5normal; 3 5 mild weakness; 2 5 moderate weakness, ableto oppose gravity; 1 5 severe weakness, unable to opposegravity but able to move horizontally; and 0 5 no move-ment. Functions of the intrinsic muscles of the hands weremonitored based on the grip attempts. Opponent move-ment (opponens pollicis; adductor pollicis) and abduction(abductor pollicis brevis) of the thumbs as well as control ofradial and ulnar deviation of the index fingers (ensured bythe concerted action of all three intrinsic muscles: firstdorsal interosseous, first palmar interosseous and firstlumbrical muscles) were recorded as positive performanceof intrinsic muscles. Simple flexion-extension movementsof the fingers in the grip attempts were regarded asfunctions of extrinsic muscles (Long et al., 1970; Chao etal., 1976; Maier and Hepp-Reymond, 1995).

Electromyographic (EMG) studies

Under ketamine sedation (50 mg/kg, i.m.), conventionalEMG recordings were performed before the operation andmonthly after the operation for 6months. Inmotor conduc-tion studies, stimulation was given at the nerve cords ofthe brachial plexus (before operation) or at the intercostalnerves above anastomosis (after operation) with bipolarsurface electrodes (bar electrode, Medelec, UK); recordingwas performed in these muscles by using subcutaneousneedle electrodes (E2 subdermal electrode, Grass, MA)with belly-tendon arrangement. Needle EMG was alsoperformed with a concentric needle (Medelec, UK). Ani-mals were kept sedated and in a recumbent position. Theoccurrence of abnormal spontaneous activities, such asfibrillation and a positive sharp wave, were observed when

156 H. CHENG ET AL.

the animals were relaxed. Motor unit potentials wererecorded when the animals were in less sedation withsome voluntary movements. To obtain a measure of thefunctional motor recovery, the amplitudes and durations ofthemotor unit potentials weremeasured. The needle EMGresults were quantified by the following scale: 0: electricalsilence; 1: appearance of simplified monophasic and low-amplitude responses with long latency; 2: polyphasic waveswith higher amplitude than 1; 3: the motor unit patternsappear within the polyphasic background activities; 4: the

EMG is normalized, demonstrating only all-or-none motorunit patterns (cf. Terzis et al., 1987).

Retrograde tracing with horseradishperoxidase (HRP)

Six months after the operation, four animals with com-plete transection of the left brachial plexuswere reanesthe-tized to explore the site of anastomosis. Gentle dissectionwas performed distally along the peripheral nerve. The

Fig. 1. The operative procedures using intercostal nerves to recon-struct the transected brachial plexus.A: Explored left brachial plexus.a, axillary nerve; m, musculocutaneous nerve; u, ulnar nerve; r, radialnerve; b, branchial artery. B: Intercostal nerves were dissected for

microsurgical reconstruction. CW, chest wall; n, intercostal nerve. C:The distal part of the severed peripheral nerves were anastomosedunder microscopical guidance with several intercostal nerves by usinga fascicular suture technique.

RECONSTRUCTED BRACHIAL PLEXUS IN PRIMATE 157

musculocutaneous nerve was severed near the neuromus-cular junctions. HRP (1,000 units) soaked in gelfoam with100 µl normal saline was applied to the cut ends of theperipheral nerves in a sealed plastic tube. Seventy-twohours after the application, the animals were deeplyanesthetized with pentobarbitone sodium (after premedi-cation with ketamine) and perfused intracardically withheparin (2,000 units) in 800 cc normal saline, followed by1,500 cc of 2% phosphate-buffered glutaraldehyde and 4%paraformaldehyde at 4°C for 15 minutes, and 1,500 cc of10% sucrose buffer at 4°C for 30 minutes. The spinal cords,spinal ganglia and sympathetic ganglia were removed forcryostat sectioning. The cervical and thoracic spinal cordswere cut transversely into serial 60-µm sections. Thespinal ganglia and sympathetic ganglia were cut longitudi-nally into 60-µm sections. The sections were incubatedwith tetramethylbenzidine to reveal HRP-containing struc-tures. In one of the four animals, the right musculocutane-ous nervewas also tracedwithHRPas control. Photomicro-graphs of the sections were either printed or scanned(SprintScan, Power Macintosh 9500) and, unless other-wise stated in the figure legends, processed as for photo-prints regarding brightness and contrast (Photoshop 3.0).A printer was used to print the figures (Kodak XLS8600PS).

RESULTS

Neurological recovery after microsurgicalreconstruction of the brachial plexus

The degree of recovery in the reinnervated muscles issummarized in Table 1. Among eight cases with reinner-vated deltoid muscle (shoulder elevation, representingaxillary nerves), seven returned a score of 3 or better, whileone returned a score of 1–2. Among eight cases withreinnervated biceps muscle (elbow flexion, representingmusculocutaneous nerves), six returned a score of 3 orbetter and the other two returned a score of 1–2. Amongnine cases with reinnervated triceps muscles (elbow exten-sion, representing radial nerves), two returned a score of 3and seven returned a score of 1–2. Among eight cases withreinnervated flexor carpi radialis muscles (wrist flexion,representing median nerve), two returned a score of 3, fivereturned a score of 1–2, while one failed to recover. Amongten cases with reinnervated intrinsic muscles of hands(opponent movement or adduction/abduction of the thumbas well as radial/ulnar deviation of the index, representingulnar nerves), only two showed minimal opponent actionsof thumb and index fingers during grip attempts, while theother eight had no function at all (these hands demon-strated severe wasting of the muscles and an inability togrip the hanging banana). Improvement ofmuscle strengthin reinnervated muscles of the left upper limbs accordingto recovery time is demonstrated in Figure 2.

The recovery of motor function in all cases was alsomonitored by EMG examinations (Figs. 3, 4). After dener-vation, the needle EMG showed gradually decreasingfibrillations and positive sharp waves if regenerative mo-tor unit potentials occurred. The motor conduction studiesdemonstrated the regenerated motor unit potentials. Gen-erally, the proximal muscle groups innervated by axillaryand musculocutaneous nerves demonstrated the first mo-tor unit potentials 2–3 months after operation. The tricepsmuscle showed responses around 3–5 months after opera-tion while the distal muscle innervated by median nerveshowed potentials after 4–6 months. The EMG responsesin intrinsicmuscles of the handswereminimal, correspond-ing to the poor recovery of hand grip function. In general,the appearance of motor unit potentials was well corre-lated with the recovery of motor function (cf. Figs. 2 and 4,correlation between the means of muscle power and EMGgrading: r 5 0.89).

Retrograde tracing with HRP

The reinnervated brachial nerves showed transport ofHRP to the motorneurons of the thoracic spinal cord wherethe proximal intercostal nerves arose (Fig. 5). In a monkey6 months after complete transection of the left brachialplexus and complete intercostal nerve reconstruction, 69labeledmotorneurons from the reinnervated bicepsmuscle

TABLE 1. Recovery of the Reinnervated Muscles 6 Months After Operation

MuscleNumberof cases Function Nerve

Muscle power, 6 months after operation

Grade 3–4 Grade 1–2 Grade 0

Deltoid 8 shoulder elevation axillary 7 1 0Biceps 8 elbow flexion musculo-cutaneous 6 2 0Triceps 9 elbow extension radial 2 7 0Flexor carpi radialis 8 wrist flexion median 2 5 1Intrinsic muscles of hand 10 hand grip1 ulnar 0 2 8

1The grip function of the intrinsic muscles of the hand includes opponent movement and abduction of the thumb as well as radial and ulnar deviation of the index finger.

Fig. 2. Improvement of muscle strength in reinnervated muscles ofleft upper limbs. S.E., shoulder elevation, deltoid muscles; E.F., elbowflexion, biceps muscles; E.E., elbow extension, triceps muscles; W.F.,wrist flexion: flexor carpi muscles; F.A., finger adduction and abduc-tion, intrinsic muscles of hand (focused on grip action).

158 H. CHENG ET AL.

were all found in the left T4 segment with a thoracicpattern of distribution. On the other hand, 224 labeledmotorneurons from the normal biceps muscle distributedin the right C5–7 segments (Fig. 6). HRP was also trans-ported to the spinal and sympathetic ganglia at thecorresponding thoracic levels (Fig. 5).

DISCUSSION

For the patient with complete avulsion of the brachialplexus and a totally paralytic and anesthetic limb, onepossibility of regaining function, albeit limited, is to pro-vide an alternative source of axons. Intercostal neurotiza-tion is one such possible method to treat traumatic rootavulsion of the brachial plexus. In reviewing publishedresults, one does not arrive at a definite code of manage-ment leading to convincing functional recovery (Yeomanand Seddon, 1961; Kotani et al., 1972; Millesi, 1977;Allieu,

1977; Sedel, 1982; Narakas, 1987). After neurotization ofthe musculocutaneous nerve, for example, the elbow flex-ionwas initially tied to respiration and could occur involun-tarily with coughing and sneezing. However, whether suchrestoration of elbow flexion actually enhances the habitualfunctional use of the limb is questionable. To improve theoutcome of such operations, further work, including techni-cal modifications, physiotherapy and other possible benefi-cial modalities such as neurostimulation may be manda-tory.To define the neuroanatomical pathways in the current

experiments, we used HRP to trace the reconstructedneuronal pathways. We proved that the reinnervatedbrachial nerves could transport HRP to those motorneu-rons of the thoracic spinal cord that give rise to theintercostal nerves. The retrograde tracer was also found inthe corresponding spinal and sympathetic ganglia. Thesefacts indicate that intercostal neurotization can lead to

Fig. 3. Serial nerve conduction studies of the left upper limbs in amonkey subjected to complete transection of the left brachial plexuswhich was subsequently reconstructed completely with intercostalnerves. 1, deltoid (representing axillary nerve); 2, biceps (representingmusculocutaneous nerve); 3, triceps (representing radial nerve); 4,flexor carpi radialis (representingmedian nerve), 5, intrinsic muscle ofhand (first interossei of index finger, representing ulnar nerve).Stimulation was given at the brachial plexus using bipolar surfaceelectrodes and recording was performed in these muscles usingsubcutaneous needle electrodeswith belly-tendon arrangement. Arrow-head in the nerve conduction studies indicates stimulus given at the

nerve cords (before operation) or at the intercostal nerves aboveanastomosis (after operation).A:Preoperative stage: themotor conduc-tion studies of the above five muscles. Note the all-or-none motor unitpotentials in all muscles.B:Motor conduction studies of the above fivemuscles 2 months after operation. There was no muscle response tothe stimulation of intercostal nerves. C: Motor conduction studies 3months after operation. Recovery of the motor conduction was noted inthe three proximal muscles (1, 2, 3). D: Motor conduction studies 6months after operation. Recovery of the motor conduction progressedin the proximal muscles (1, 2, 3) and flexor carpi radialis (4). Nomuscleresponse was noted in the intrinsic muscle of hand (5).

RECONSTRUCTED BRACHIAL PLEXUS IN PRIMATE 159

sensory, autonomic and motor restoration and confirm theprinciple that a sectioned peripheral nerve can regeneratethrough foreign nerves (see Jessell, 1991). Intercostalnerves are apparently a good source of different kinds ofaxons which are able to regenerate. The essential compo-nents in the local circuit of thoracic spinal cord may thusbe restored. Therefore, methods designed to enhance cen-tral pathway control are also worthy of further investiga-tion.The muscles innervated by the anastomosed brachial

plexus showed additional improvement of strength alsoafter the EMG improvement (about one month later).However, the tests using food stimulation did not showsignificant purposeful movements, such as grasping by thereconstructed upper limbs. This might be the result ofinadequate regeneration in the distal muscles, especiallythe intrinsic muscles of hands needed for skillful move-ments. On the other hand, limitation of central plasticitymight also contribute. Movements of the reconstructedupper limbs were found in synchrony with respiration.Hence, the thoracic segments used to reinnervate thebrachial plexus has led to the formation of novel neuralcircuits that receive input signals from the pyramidal tractand reticular formation serving to initiate and executemotor patterns. However, as evidenced by the behavioralpatterns in these studies and the fact that the retrograde-labeled neurons were all in the thoracic cords, thesepatterns of novel motility were basically of a thoracic type.Thus, our results are in accord with observations in aseries of rodent studies that abnormal reinnervation by

cross-anastomosis causes abnormal motor behaviors(Sperry, 1941, 1942, 1943, 1945). Recent evidence suggeststhat the cortical working memory for performing tasks isactually point-to-point among elements of the integratedcircuitry (Wilson et al., 1993; Goldman-Rakic, 1995). Theplasticity of these somatotopically arranged neural net-works is supposed to be limited between two widelyseparated areas, such as between the midcervical andmidthoracic motor cortical areas of concern in the presentstudies. A compensatory take-over phenomenon can beobserved in the rodent vibrissae-representing fields of thetrigeminal ganglion or thalamus (i.e., expansion of adja-cent barrels after removal of vibrissae follicles at birth:Jeanmonod et al., 1981; Johansson and Arvidsson, 1994).However, this plasticity is generated by rewiring fromnearby neurons and occurs in neonates. Alterations ofmovement representations in primary motor cortex ofadult monkeys by task training or peripheral nerve lesionsin rats have been shown (Donoghue and Sanes, 1987;Donoghue et al., 1990, 1992; Sanes et al., 1990; Nudo andMilliken, 1996). Nevertheless, the output reorganizationappeared to be restricted to cortically adjacent areas(Donoghue and Sanes, 1987; Donoghue et al., 1990, 1992;Sanes et al., 1990). Under the digit-task training inmonkeys, use-dependent reorganization of the motor cor-tex was noted primarily in single-movement digit catego-ries or in dual-response categories involving the combina-tion of a digit movement and another distal forelimbmovement (Nudo and Milliken, 1996). That is to say, thereorganization occurs mainly within the same or closelyassociated target motor neuron pools. In our studies, wehave no evidence to support remote compensatory mecha-nisms such as between midcervical and midthoracic corti-cal motor areas. It is likely that there is much less chanceto have synaptic rearrangement, activation of normallysilent synapses, or axonal growth between two widelyseparate areas in brain (Raisman and Field, 1973;Dostrovsky, 1976; Metzler and Marks, 1979; Nelson et al.,1979). Another possible mechanism is perhaps that theremaining cervical representation prevented a thoracic‘‘invasion’’ of cervical cortex (Lane et al., 1995). Therefore,efforts to bring the reconstructed neural networks as closeto the original status as possible cannot be overemphasized.In the present studies, the intrinsic muscles of the hand,

which are crucial in carrying out certain hand tasks, didnot regain significant reinnervation and function in mostcases. The reason why these muscles improved much lessthan the more proximal muscles of the upper limb mightrelate to the distance from the lesion and the severewasting conditions leading to inadequate survival ofmusclefibers to be reinnervated. Perhaps providing some trophicsupport for these distal crucial muscles during the waitingperiod might increase the degree of functional reinnerva-tion.Recently, direct reconstruction of connectivity between

the spinal cord and the nerves after spinal nerve rootinjury has also been demonstrated (Cullheim et al., 1989;Carlstedt et al., 1990, 1993; Smith and Kodama, 1991).This kind of approach is more capable of bringing thereconstructed neural networks closer to the original sta-tus. However, in clinical practice, retracted ends of theavulsed roots are difficult to find and approximate afterthe 2–3 months of neurological evaluation, which is theminimum time necessary to verify a real neurotemesis(Leffert, 1983). In usual cases, this hinders the attempt to

Fig. 4. Stages of nerve regeneration as indicated by the changes ofneedle electromyographic (EMG) activities in the muscles of left upperlimbs. Note EMG recovery of the proximal group of muscles (deltoid,biceps, and triceps) preceding the recovery of the distal group ofmuscles. EMG grading scale: 0, electrical silence; 1, appearance ofsimplified monophasic and low amplitude responses with long latency;2, polyphasic waves with higher amplitude than 1; 3, the motor unitpatterns appear within the polyphasic background activities; 4, theEMG is normalized, demonstrating only all-or-none motor unit pat-terns. D, deltoid, axillary nerve; B, biceps, musculocutaneous nerve; T,triceps, radial nerve; FC, flexor carpi radialis, median nerve; IH,intrinsic muscles of hand (first interossei of index fingers), ulnarnerve.

160 H. CHENG ET AL.

Fig. 5. Labeled thoracic neurons (at the level of T4) containinghorseradish peroxidase (HRP) transported retrogradely from thereinnervated left biceps muscle 6 months after operation. A: Trans-verse section of the thoracic cord showing HRP-labeled cells in theventral horn (boxed area). The right side spinal cord was marked by adorsal cut.B:Magnification of the box inA. Note labeledmotorneuronsin the ventral horn (asterisks). C: Longitudinal section of the left T4

dorsal root ganglia (DRG). Many small and large DRG neurons areHRP-labeled. D: Close-up of the box in C. Note both small and largelabeled DRG neurons (asterisks).E: Longitudinal section of the left T4sympathetic ganglion. F: Close-up of the box in E. Note the HRP-labeled ganglion cells (asterisks). Scale bars 5 50 µm for B, D, F; 0.6mm forA; 250 µm for C, E.

RECONSTRUCTED BRACHIAL PLEXUS IN PRIMATE 161

reinsert the avulsed roots to the spinal cord. For earlierdiagnosis, applying functional neuroimaging techniquesmight become helpful. Recently, we have found that theuse of trophic factors in fibrin glue can facilitate nerveregeneration in central nervous system (CNS) or periph-eral nerves in CNS tissues (Cheng et al., 1995a,b, 1996).Results of the present studies suggest, in case diagnosticand operative procedures related to brachial plexus repaircan be improved, that a combination of parts of themethodology reported here and nerve root reimplantation,with or without free nerve grafts as well as trophic factorsin fibrin glue, might lead to a better clinical outcome in thefuture.

ACKNOWLEDGMENTS

The study was supported by the Taiwan Chin-Lin Medi-cal Research Fund and NSC77-0412-B075-17. We thankLars Olson for valuable advice and assistance as well asIda Engqvist for excellent editorial assistance.

LITERATURE CITED

Allieu, Y. (1977) Exploration et traitement direct des lesions nerveusesdans les paralysis traumatiques par elongation du plexus brachial chezl’adulte. Rev. Chir. Orthop. 63:107–122.

Brunelli, G. (1980) Neurotization of avulsed roots of the brachial plexus(preliminary report). Int. J. Microsurg. 2:55–58.

Carlstedt, T. (1995) Spinal nerve root injuries in brachial plexus lesions:Basic science and clinical application of new surgical strategies. Areview. Microsurgery 16:13–16.

Carlstedt, T., M. Risling, H. Lindå, S. Cullheim, and B. Ulfhake (1990)Regeneration after spinal nerve root injury. Restor. Neurol. Neurosci.1:289–295.

Carlstedt, T., R. Halin, K. Hedstrom, and I. Nilsson-Remahl (1993)Functional recovery in primates with brachial plexus injury after spinalcord implantation of avulsed ventral roots. J. Neurol. Neurosurg.Psychi. 56:649–654.

Chao, E., J. Opgrande, and F. Axmear (1976) Three-dimensional forceanalysis of finger joints in selected isometric hand functions. J. Bio-mech. 9:387–396.

Cheng, H., S. Almstrom, and L. Olson (1995a) Fibrin glue used as anadhesive agent in CNS tissues. J. Neural Transp. & Plast. 5:233–243.

Cheng, H., B. Hoffer, I. Stromberg, D. Russell, and L. Olson (1995b) Theeffect of glial cell line-derived neurotrophic factor in fibrin glue ondeveloping dopamine neurons. Exp. Brain Res. 104:199–206.

Cheng, H., Y. Cao, L. Olson (1996) Spinal cord repair in adult paraplegicrats: Partial restoration of hindlimb function. Science 273:510–513.

Cullheim, S., T. Carlstedt, H. Linda, M. Risling, and B. Ulfhake (1989)Motorneurons reinnervate skeletal muscle after ventral root implanta-tion into the spinal cord of the cat. Neuroscience 29:725–733.

Donoghue, J.P., and J.N. Sanes (1987) Peripheral nerve injury in develop-ing rats reorganizes representation pattern in motor cortex. Proc. Natl.Acad. Sci. USA 84:1123–1126.

Donoghue, J.P., S. Suner, and J.N. Sanes (1990) Dynamic organization ofprimary motor cortex output to target muscle in adult rats: II. rapidreorganization following motor nerve lesions. Exp. Brain Res. 79:492–503.

Donoghue, J.P., S. Leibovic, and J.N. Sanes (1992) Organization of theforelimb area in squirrel monkey motor cortex: Representation of digit,wrist, and elbow muscles. Exp. Brain Res. 89:1–19.

Dostrovsky, J., J. Millar, and P. Wall (1976) The immediate shift of afferentdrive of dorsal columnnucleus cells following deafferentation:Acompari-son of acute and chronic deafferentation in gracile nucleus and spinalcord. Exp. Neurol. 52:480–495.

Geisert, E. (1976) The use of tritriated HRP for defining neuronal path-ways: A new application. Brain Res. 100:253–269.

Goldman-Rakic, P. (1995) Cellular basis of working memory. Neuron14:477–485.

Hallin, R., Z. Wiesenfeld, and H. Lugnegård (1981) Neurophysiologicalstudies of peripheral nerve functions after neural regeneration follow-ing nerve suture in man. Int. Rehab. Med. 3:187–192.

Halperin, J. (1975) A study of the dynamics of retrograde transport andaccumulation of HRP in injuried axons. Brain Res. 100:253–269.

Heimer, L. (1981) Neuroanatomical Tract-Tracing Methods Vol. 1. NewYork: Plenum Press.

Heimer, L. (1989) Neuroanatomical Tract-Tracing Methods Vol. 2. NewYork: Plenum Press.

Jeanmond, D., F. Rice, and H.V. Loos (1981) Mouse somatosensory cortex:Alterations in the barrelfield following receptor injury at different earlypostnatal ages. Neuroscience 6:1503–1535.

Jessell, T. (1991) Reactons of neurons to injury: Peripheral nerve graftspromote the growth of central axons. In E.R. Kandel, J.H. Schwartz,and T.M. Jessell (eds): Principles of Neural Science. NewYork: Elsevier,pp. 259–269.

Johansson, K., and J. Arvidsson (1994) Central plasticity in rat trigeminalprimary sensory neurons innervating vibrissae after neonatal periph-eral nerve injury. Neurosci. Lett. 167:187–190.

Katoni, P., H. Matsuda, and T. Suzuki (1972) Trial surgical procedure ofnerve transfers to avulsion injuries of plexus brachialis. Excerpta med.Int. 12th Congress Series. Series 291 (SICOT Congress, Tel Aviv, 1972),pp. 348–350.

Lane, R.D., C. Bennett-Clarke, N. Chiaia, H. Killackey, and R. Rhoades(1995) Lesion-induced reorganization in the brainstem is not com-pletely expressed in somatosensory cortex. Proc. Nat. Acad. Sci. USA92:4264–4268.

Fig. 6. Distribution of HRP-labeled motorneurons retrogradelytraced from the biceps muscle in a monkey 6 months after completetransection of the left brachial plexus and complete intercostal nervereconstruction (Left side: the musculocutaneous nerve was subjectedto cutting and reconstruction with T4 intercostal nerve; rightside: untreated control). The labeled motorneurons from the reinner-vated biceps muscle were all found in the left T4 segment with athoracic pattern of distribution. The labeled motorneurons from thenormal biceps muscle are distributed in the right C5–C7 segments.Numbers under the figures indicate counts of HRP-labeled motorneu-rons.

162 H. CHENG ET AL.

Leffert, R. (1983) Surgery of the peripheral nerves and brachial plexus. InH.H. Schmidek and W.H. Sweet (eds): Operative Neurosurgical Tech-niques. NewYork: Grune & Stratton, pp. 1495–1540.

Long, C., P. Conrad, E. Hall, and S. Furler (1970) Intrinsic-extrinsic musclecontrol of the hand in power grip and precision handling. J. Bone. JointSurg. Am. 52:853–867.

Maier, M., and M.C. Hepp-Reymond (1995) EMG activation patternsduring force production in precision grip. Exp. Brain Res. 103:108–122.

Mesulam, M. (1982) Principles of HRP neurohistochemistry and theirapplications for tracing neural pathways-axonal transport, enzymehistochemistry and light microscopic analysis. In M.M. Mesulam (ed):Tracing Neural Connections with Horseradish Peroxidase. New York:JohnWiley & Sons, pp. 1–151.

Metzler, J., and P. Marks (1979) Functional changes in cat somaticsensory-motor cortex during short term reversible epidermal blocks.Brain Res. 177:379–383.

Millesi, H. (1977) Surgical management of brachial plexus injuries. J. HandSurg. 2:367–379.

Narrakas, A. (1987) Thoughts on neurotization or nerve transfers inirreparable nerve lesions. In J.K. Terzis (ed): Microreconstruction ofNerve Injuries. Philadelphia: WB Saunders, pp. 447–454.

Nelson, S., T. Collatos, A. Niechaj, and L. Mendell (1979) Immediateincrease in Ia-motorneuron synaptic transmission caudal to spinal cordtransection. J. Neurophysiol. 42:655–664.

Nudo, R., and G. Milliken (1996) Use-dependent alterations of movementrepresentations in primary motor cortex of adult squirrel monkeys. J.Neurosci. 16:785–807.

Oldfield, B. (1977) Uptake and retrograde transport of HRP by axon ofintact and damaged peripheral nerve trunks. Neuroscience 6:135–141.

Raimbault, J. (1987) Electrical assessment of muscle denervation. In J.K.Terzis (ed): Microreconstruction of Nerve Injuries. Philadelphia: WBSaunders, pp. 211–221.

Raisman, G., and P. Field (1973) A quantitative investigation of thedevelopment of collateral reinnervation after partial deafferentation ofthe septal nuclei. Brain Res. 50:241–264.

Rowland, L. (1991) Diseases of the motor unit. In E.R. Kandel, J.H.Schwartz, and T.M. Jessell (eds): Principles of Neural Science. NewYork: Elsevier, pp. 244–257.

Sanes, J.N., S. Suner, and J.P. Donoghue (1990) Dynamic organization ofprimarymotor cortex output to target muscle in adult rats: I. Long-termpatterns of reorganization following motor or mixed peripheral nervelesions. Exp. Brain Res. 79:479–491.

Sedel, L. (1982) The results of surgical repair of brachial plexus injuries. J.Bone Joint Surg. 64B:54–66.

Smith, K., and R. Kodama (1991) Reinnervation of denervated skeletalmuscle by central neuron regenerating via ventral roots implanted intothe spinal cord. Brain Res. Bull. 30:447–451.

Sperry, R.W. (1941) Effect of crossing nerves to antagonistic muscles in hindlimb of rat. J. Comp. Neurol. 75:1–19.

Sperry, R.W. (1942) Transposition of motor nerves and muscles in forelimbof rat. J. Comp. Neurol. 76:283–321.

Sperry, R.W. (1943) Functional results of crossing sensory nerves in rat. J.Comp. Neurol. 78:59–90.

Sperry, R.W. (1945) Problems of central nervous reorganization after nerveregeneration and muscle transposition. Quart. Rev. Biol. 20:311–369.

Terzis, J.K., W.T. Liberson, and H.A. Maragh (1987) Motorcycle brachialplexopathy. In J.K. Terzis (ed): Microreconstruction of Nerve Injuries.Philadelphia: W.B. Saunders, pp. 361–384.

Wilson, F., S.O. Scalaidhe, and P. Goldman-Rakic (1993) Dissociation ofobject and spatial processing domains in primate prefrontal cortex.Science 260:1955–1958.

Yeoman, P., and H. Seddon (1961) Brachial plexus injuries: Treatment ofthe flail arm. J. Bone Joint Surg. 43B:493–499.

RECONSTRUCTED BRACHIAL PLEXUS IN PRIMATE 163