Invited Discussion

In this report, the authors investigate the ability of sensory nerves to maintain the gastrocnemius muscle for delayed motor reinnervation, either with a proximal sensory to distal motor nerve repair or a direct sensory nerve to “aneural” muscle neurotization. Six months later, the common peroneal nerve was transected and a neurorrhaphy performed between its proximal end and the distal stump of the tibial nerve. An immediate peroneal nerve to tibial nerve neurorrhaphy served as a positive control, while tibial nerve transection with no procedure to preserve the distal nerve stump or muscle during a 6 month period of denervation served as the negative control. In all cases, analyses of the nerves and muscle were performed 3 months after the common peroneal to tibial nerve neurorrhaphy. The authors suggest that sensory nerves can protect the muscle independent of distal nerve sheath involvement, but that the distal nerve sheath may still play a role in enhancing nerve regeneration.

The greatest strength of this study is that it utilizes an interesting animal model to study a question of considerable clinical relevance - methods of preserving the neuromuscular junction - and suggests a techinique for baby-sitting denervated muscle that is eminently doable clinically without sparing a motor nerve in the process. Studies designed to protect the neuromuscular junction with neurally-induced embryonic stem cells instead of functioning host nerves are afoot,[1] [2] but procedures designed to protect or “baby-sit” denervated muscle with existing motor or sensory nerves have been successfully reported elsewhere,[3] [4] [5] and may offer the most promise for the immediate future. Despite this and other promising reports, however, there is also some data to suggest that repeated denervation-reinnervation cycles with the baby-sitter technique will impair skeletal muscle contractile function.[5]

While the authors should be commended for their use of multiple techniques for evaluating nerve regeneration, muscle reinnervation, architecture, and fiber distribution, a few methodologic changes may be in order as they continue to study this interesting model. First, one of the primary hypotheses of this study was to determine if sensory nerve fibers could protect denervated muscle. The investigators ligated the proximal stump of the tibial nerve and sutured it to the biceps femoris musculature to prevent regenerating tibial nerve fibers from contaminating the saphenous to distal tibial nerve neurorrhaphy. While this may have been sufficient to prevent contamination, it needs to be more rigorously validated. Utilizing transgenic mice whose axons constitutively express enhanced yellow or green fluorescent protein in their neurons, we have shown that regenerating fibers from the transected tibial nerve will travel along the surfaces of surrounding hamstring, gluteus, and gastrocnemius muscles and reach, in a very random fashion, the distal unrepaired nerve stump (Fig. [1]). The potentially contaminating effects of the proximal nerve stump, also reported in other rodent peripheral nerve injury models including end-to-side neurorrhaphy,[6] can be identified with additional staining techniques. While the authors suggest acetylcholinesterase to indentify motor, and carbonic anhydrase to identify sensory, fibers, a retrograde labeling technique such as horseradish peroxidase or Fast Blue may be more useful. Since the tibial nerve contains mixed motor and sensory fibers from spinal cord levels caudad to the purely sensory saphenous nerve, a retrograde technique could identify the modality and spinal cord origin of the fibers distal to the repair.

Figure 1 Randomly regenerating axons 3 weeks following epineurotomy of the mouse tibial nerve. Axons travel in a random fashion along biceps femoris and gastrocnemius muscle without the benefit of a nerve graft or neurorrhaphy. Image taken from a double transgenic mouse overexpressing enhanced yellow fluorescent protein (EYFP) in its peripheral nerves and glial-derived neurotrophic factor (GDNF) under the control of the glial fibrillary acidic protein promoter (GFAP) of the neuroglia. (Callibration bar = 5 mm; ^*= the epineurotomy site; > = the paths of randomly regenerating axons over a > 1 cm course in a living mouse using fluorescence microscopy, z-motor, CoolSnap Monochromic CCD, and MetaMorph version 6.2 software.)

Histomorphometric analysis was performed on only three representative samples of intramuscular tibial nerve per treatment group. There is no doubt that histomorphometric analysis is a tedious and difficult process, but to provide a more reliable quantitative assessment of nerve morphometry, all of the nerves in each group should have been similarly evaluated. In addition, the authors comment that the tibial nerve typically contains ∼1000 axons in its uninjured state. However, when evaluating every nerve in every group, we have identified as many as ∼6000 axons in the uninjured tibial nerves of male Lewis rats,[7] and ∼4000 axons in the mouse tibial nerve,[8] [9] [10] thus stressing the importance of thorough morphometric evaluation.

The authors describe an “aneural zone” and reference work performed by Brunelli,[11] utilizing a peroneal nerve to lateral gastrocnemius model in rabbits, rats, as well as direct neurotization of damaged muscles in people. In this reference, the anatomic description of the aneural zone is limited to qualitative electron microscopy of “newly formed motor end-plates” with no described control groups. In the rodent lateral and medial gastrocnemius heads, motor end-plates are reliably aligned as a transverse strip along the deep surface, and innervating terminal branches of the tibial nerve pass from proximal to distal. Thus, while we agree that the region they designated their aneural zone contained no motor end-plates, to rigorously confirm this, they could have used immunohistochemical techniques to establish the location of motor end-plates, and thus, by exclusion, confirmed their aneural zone. For example, acetylcholine receptors could have been labeled with bungarotoxin or other acetylcholine receptor (AchR) primary antibodies, and the terminal axons with a variety of neurofilament labels including the SMI monoclonals, or antibodies to synaptophysin or synaptobrevin. With fluorescence microscopy of transgenic mice that express enhanced yellow fluorescent protein (EYFP) in their axons, and rhodominated bungarotoxin to stain motor end plates, we demonstrate the uniform distribution of motor end-plates in living mice (Fig. [2]) Moreover, others have shown that regenerating terminal branches of axons travel along pre-existing paths to efficiently reinnervate their target end-plates and that disruption of these paths severely compromises reinnervation, rather than enabling reinnervation in a previoulsy aneural zone.[12] These data would suggest that new neuromuscular junctions are not formed in a previously aneural region.

Figure 2 Live imaging of the posterior surface of the lateral gastrocnemius muscle 3 weeks after denervation, and subsequent re-innervation. A transversely oriented strip of acetylcholine receptors (red) reappears as regenerating terminal axonal branches (green) reach them in a hThy-EYFP mouse whose axons are labeled with rhodominated-alpha-bungarotoxin. Acetylcholine receptors were not labeled in regions distant from this strip which is reliably found at the junction of the proximal and middle thirds of the posterior gastrocnemius. Similar strips of acetylcholine receptors are also reliably located in the peroneal and facial mimetic musculature (data not shown).

Based on the data presented in this interesting study, as well as other related work, [4] [5] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] denervated muscle can be preserved even with sensory nerves. We agree with the authors that preservation of the distal endoneurial tubes is important to maintain a favorable neuroregenerative enviroment, not only by preventing collagenation, but also by affecting the expression of neurotrophins to favorably benefit nerve regeneration in chronically denervated nerves. Also, regenerating axons that can travel a few centimeters to identify a distal target could very easily repopulate vacated endoneurial tubes, only a couple of millimeters away, following tibial nerve injury along the muscle surface. Given the presence of literally thousands of terminal tibial nerve branches traveling along the gastrocnemius, each intended for a specific acetylcholine receptor, we are thus inclined to think that even the direct implantation of the sural nerve into muscle actually affected a nerve-to-nerve coaptation rather than a true nerve to muscle connection, as suggested by the authors. Nonetheless, we are indebted to the authors for this important study and look forward to updated iterations of their model, as improved quantitative techniques become available.

REFERENCES

• 1 Myckatyn T M, Mackinnon S E. A review of research endeavors to optimize peripheral nerve reconstruction.  Neurol Res. 2004;  26 124-138
• 2 Myckatyn T M, Mackinnon S E, McDonald J W. Stem cell transplantation and other novel techniques for promoting recovery from spinal cord injury.  Transpl Immunol. 2004;  12 343-358
• 3 Endo T, Hata J, Nakayama Y. Variations on the “baby-sitter” procedure for reconstruction of facial paralysis.  J Reconstr Microsurg. 2000;  16 37-43
• 4 Mersa B, Tiangco D A, Terzis J K. Efficacy of the “baby-sitter” procedure after prolonged denervation.  J Reconstr Microsurg. 2000;  16 27-35
• 5 Yoshimura K, Asato H, Jejurikar S S, Cederna P S, Urbanchek M G, Kuzon Jr. W M. The effect of two episodes of denervation and reinnervation on skeletal muscle contractile function.  Plast Reconstr Surg. 2002;  109 212-219
• 6 Goheen-Robillard B, Myckatyn T M, Mackinnon S E, Hunter D A. End-to-side neurorrhaphy and lateral axonal sprouting in a long graft rat model.  Laryngoscope. 2002;  112 899-905
• 7 Myckatyn T M, Brenner M, Mackinnon S E, Chao C K, Hunter D A, Hussussian C J. Effects of external beam radiation in the rat tibial nerve after crush, transection and repair, or nerve isograft paradigms.  Laryngoscope. 2004;  114 931-938
• 8 Myckatyn T M, Ellis R A, Grand A G et al.. The effects of rapamycin in murine peripheral nerve isografts and allografts.  Plast Reconstr Surg. 2002;  109 2405-2417
• 9 Myckatyn T M, Hunter D A, Mackinnon S E. The effects of cold preservation and subimmunosuppressive doses of FK506 on axonal regeneration in murine peripheral nerve isografts.  Can J Plast Surg. 2003;  11 15-22
• 10 Brenner M J, Tung T H, Mackinnon S E, Myckatyn T M, Hunter D A, Mohanakumar T. Anti-CD40 ligand monoclonal antibody induces a permissive state, but not tolerance, for murine peripheral nerve allografts.  Exp Neurol. 2004;  186 59-69
• 11 Brunelli G. Direct neurotization of severely damaged muscles.  J Hand Surg. 1982;  7A 572-579
• 12 Nguyen Q T, Sanes J R, Lichtman J W. Pre-existing pathways promote precise projection patterns.  Nat Neurosci. 2002;  5 861-867
• 13 Sanes J R, Apel E D, Burgess R W et al.. Development of the neuromuscular junction: genetic analysis in mice.  J Physiol Paris. 1998;  92 167-172
• 14 Feng G, Mellor R H, Bernstein M et al.. Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP.  Neuron. 2000;  28 41-51
• 15 Fu S Y, Gordon T. Contributing factors to poor functional recovery after delayed nerve repair: prolonged axotomy.  J Neurosci. 1995;  15 3876-3885
• 16 Fu S Y, Gordon T. Contributing factors to poor functional recovery after delayed nerve repair: prolonged denervation.  J Neurosci. 1995;  15 3886-3895
• 17 Boyd J G, Gordon T. Glial cell line-derived neurotrophic factor and brain-derived neurotrophic factor sustain the axonal regeneration of chronically axotomized motoneurons in vivo.  Exp Neurol. 2003;  183 610-619
• 18 Boyd J G, Gordon T. Neurotrophic factors and their receptors in axonal regeneration and functional recovery after peripheral nerve injury.  Mol Neurobiol. 2003;  27 277-324
• 19 Gordon T, Sulaiman O, Boyd J G. Experimental strategies to promote functional recovery after peripheral nerve injuries.  J Peripher Nerv Syst. 2003;  8 236-250
• 20 Myckatyn T M, Mayckinnon S, Hunter D, Brakefield D, Parsadanian A. A novel model for the study of peripheral nerve regeneration following common nerve injury paradigms.  J Reconstr Microsurg. 2004;  20 533-544
• 21 Kalantarian B, Rice D C, Tiangco D A, Terzis J K. Gains and losses of the XII-VII component of the “baby-sitter” procedure: morphometric analysis.  J Reconstr Microsurg. 1998;  14 459-471
• 22 Kobayashi J, Mackinnon S E, Watanabe O et al.. The effect of duration of muscle denervation on functional recovery in the rat model.  Muscle Nerve. 1997;  20 858-866