In rat dorsal root ganglia, histochemical staining of carbonic anhydrase (CA) and cholinesterase (CE) yields a reciprocal pattern of activity: Sensory processes are CA positive and CE negative, whereas motor processes are CA negative and CE positive. In rat infraorbital nerve (a sensory peripheral nerve), we saw extensive CA staining of nearly 100% of the myelinated axons. Although CE reactivity in myelinated axons was extremely rare, we did observe CE staining of unmyelinated autonomic fibers. Four weeks after transection of infraorbital nerves, CA-stained longitudinal sections of the proximal stump demonstrated 3 distinct morphological zones. A fraction of the viable axons retained CA activity to within 2 mm of the distal extent of the stump, and the stain is capable of resolving growth sprouts being regenerated from these fibers. Staining of unmyelinated autonomic fibers in serial sections shows that CE activity was not retained as far distally as is the CA sensory staining. MICROSURGERY 13:39-44 1992
HISTOCHEMICAL DISCRIMINATION OF FIBERS IN REGENERATING RAT INFRAORBITAL NERVE RUSSELL A. WILKE, DANNY A. RILEY, Ph.D., and JAMES R. SANGER, M.D.
Post-injury regrowth of peripheral nerves is currently receiving attention at all levels of investigation, from molecular’ to microanat~mical~-~ to It is known that after a peripheral nerve is severed the distal stump eventually degenerates, leaving behind Schwann cells and the connective tissue network of the original endoneuria. Although the Schwann cell morphology changes somewhat, these endoneurial tubes remain basically intact and provide a scaffolding upon which regrowth can be directed. Growth cones sprouting from the proximal stump are guided along this scaffolding, and the success of reinnervation depends at least partly upon association of these sprouts (either sensory or motor) with the appropriate endoneurial sheaths. Traumatized patients whose newly generated sprouts fail to successfully reinnervate the appropriate end organs will see limited (if any) recovery of clinical function. In such cases, histochemically-guided microsurgical reconstruction of the nerve may be indicated. In the past, investigators have been able to identify with s2
’
From the Department of Cellular Biology and Anatomy (R.A.W., D.A.R.) and the Department of Plastic and Reconstructive Surgery (J.R.S.), Medical College of Wisconsin, Milwaukee, WI. Address reprint requests to Danny A. Riley, Ph.D.. 8701 Watertown Plank Road, Milwaukee, WI 53226. Acknowledgments: This research was supported in part by NASA grant NAG 2-552 to D.A.R. R.A.W. is supported by the Medical Scientist Training Program at the Medical College of Wisconsin. The authors greatly appreciate the excellent assistant of Mr. G.R. Slocum with the photomicroscopy and figure preparation and S.Tjepkema-Burrows with plate reproduction. Received May 13, 1991; revision accepted August 9, 1991. 0 1992 Wiley-Liss, Inc.
a high degree of resolution the position of myelinated sensory axons within a mixed peripheral nerve fascicle by histochemically staining for axoplasmic carbonic anhydrase (CA).9-’1 Other workers have demonstrated the usefulness of staining for cholinesterase (CE) activity, in the localization of myelinated motor axons. 12-14 In recent years, Riley and coworkers have used a modification of this CE staining procedure in combination with the CA method to identify the relative positions of motor versus sensory processes in serial sections of peripheral nerve. 15*16 In the present study, we performed controlled microsurgical injury of rat infraorbital nerves and applied the CAKE tandem staining technique after 4 weeks of regrowth. Our goals were to further demonstrate the reciprocal nature of the stains, to highlight the morphological changes seen in a sensory nerve after 4 weeks of degeneratiodregeneration, and to address some current issues in the literature regarding the usefulness of the CE technique for labelling autonomic fibers. Herein, we also report further improvements in the CE staining method which were made in an effort to abbreviate the overall time requirements of the procedure. MATERIALS AND METHODS Animal Surgery and Specimen Preparation Nine male Sprague-Dawley rats (383 f. 25 g. b. wt.)
were prepared for surgery with sterile instruments and aseptic procedures (as outlined in NIH “Guide for the Care and Use of Laboratory Animals,” Pub. 85-23, and the Medical College of Wisconsin “Policy on Surgery and Post-Surgical Care”). The rats were anesthetized with a loading dose of
40
Wilke et al.
Figure 1. Reciprmiity of carbonic anhydrase (CA) and cholinesterase (CE) staining is demonstrated for myelinated axons in a lower lumbar spinal ganglion of a normal rat. A: A 10 pm section stained for CA activity. Approximately 50% of the myelinated sensory axons of the dorsal root (OR) are darkly reactive. Many of the neuronal cell
bodies exhibit moderate to dark staining. 6:A 20 p n section stained for C E activity. About 75% of the myelinated motor axons of the ventral root (VR) show reactive axoplasm. These motor axons do not contain CA activity, and the sensory axons do not possess CE activity above background. x 92.
50 mg/kg of intraperitoneal sodium pentobarbital (Nembutal) and also received 0.05 mg/kg of subcutaneous atropine. Intraperitoneal maintenance doses (0.5 ml) of sodium pentobarbital (6.5 mglml) were delivered as necessary. The left infraorbital nerve of each animal was exposed using a lateral skin incision, inferior to the caudal-most vibrissae, and isolated by blunt dissection under a binocular dissecting microscope. Care was taken not to include the buccal branch of the facial motor nerve, which lies immediately superficial to the ventral 1/3 of the infraorbital nerve. To contain later regrowth, the isolated infraorbital nerve was wrapped with sterile “Saran Wrap” brand plastic film (from Dow Chemicals; 0.5 mm polyvinylidene chloride) to form an impervious sheath of approximately 3 mm in length. This sheath was then wrapped with one loop of 6-0 silk suture, tied tightly, and transected immediately distal to the ligature, severing the nerve. The wounds were closed and the rats were housed on soft sawdust. Wounds were inspected daily for infection. Following these precautions, and those listed in our institution’s “Animal Use Bulletin on Post-Operative Care in Rodents,” no problems developed that required elimination of rats from the study. Four weeks later, the animals were sacrificed and the proximal stump of each left infraorbital nerve was removed by transection approximately 0.5 cm proximal to the ligature. Right infraorbital nerves were harvested at similar levels and used as controls. Excised nerve segments were immediately transferred to 20 ml vials for immersion fixation in 2.5% glutaraldehyde, 0.1 M phosphate buffer, pH 7.4 at room temperature. Preliminary tests of fixation by 11% paraformaldehyde, both with and without 0.2% glutaraldehyde, yielded unsatisfactory results with the CE stain; sections showed diffuse background staining and poor resolution of myelinated axons. After 15 minutes, the vials were
transferred to a 5°C refrigerator for 1 hour 45 minutes. Following fixation, the specimens were rinsed 3 times in buffer and subjected to the following sucrose cryoprotection series: 10% sucrose for 20 minutes, 20% sucrose for 40 minutes, 30% sucrose overnight-all in 0.1 M phosphate buffer, pH 7.4. The next morning, tissues were attached to index cards, quick-frozen in freon, and stored under liquid nitrogen as described previously. l 5 Lower lumbar dorsal root ganglia and the accompanying dorsal and ventral roots were also removed from 3 of the animals following euthanasia. These tissues were fixed, cryoprotected, and frozen as described above; they were used in preliminary studies aimed at further refinement of the CE staining protocol. Histochemical Staining for Carbonic Anhydrase and Cholinesterase Activities
After removal from storage in liquid nitrogen, control and injured nerve segments were sectioned either longitudinally or transversely with a cryostat microtome at -22°C. Staining was found to be optimal if the thickness of the sections was alternated, 10 pm for CA and 20 pm for CE. As in previous work,15 the distance sectioned from the free end of the transected proximal stumps was carefully recorded so that any observed morphological changes could be accurately located along the length of the injured nerve. Serial sections were collected on 2% gelatin-coated glass slides and stained histochemically for either CA or CE activity. Staining for CA activity was conducted following the improved method of Riley and coworkers. 10711 In this method, tissue carbonic anhydrase converts CO, in the incubation medium to H,CO,, changing the pH in the microenvironment of the enzyme and ultimately {causinglocal
lnfraorbital Nerve Fiber Types
41
Figure 2. Serial cross-sections of fascicles in a normal infraorbital nerve. A: Every fascicle is filled with CA reactive myelinated sensory axons. B: The fascicles contain differing amounts of CE reactive structures. x 68.
precipitation of cobalt s a k 7 It should also be noted that while this study was completely light microscopic, the procedures for fixation and staining are compatible with electron microscopy." Staining for CE activity was conducted by a modification of the original method of Karnovsky and Roots. l 2 This method supplies acetylthiocholine as a substrate for tissue cholinesterases. The thiocholine released participates in a reduction that ultimately deposits copper ferrocyanide at the site of the enzyme. Details of our previous modifications have been published earlier.15In the present study, we further enhanced the relatively weak CE staining by elevating the temperature of the 1-hour tissue incubation from 22°C to 46°C (suggested in reference 14) and by increasing section thickness from 10 pm to 20 pm. Following incubation, the tissue was rinsed in distilled water and covered with 1% AgNO,. Subsequent exposure to bright illumination under a fiber optic light (100 W halogen bulb) promotes the reduction of silver over the copper ferrocyanide precipitate to intensify the staining. Finally, the CA- and CE-stained sections were rinsed, dehydrated, and mounted permanently using pre- 1978 Fisher Permount. Post- 1978 Permount oxidizes and rapidly fades the CA reaction product unless retarded by an added antioxidant.l 1 We have also found that Canadian balsam is a satisfactory nonoxidizing substitute medium.
ganglia, approximately half of the myelinated sensory axons of the dorsal nerve root stained intensely CA positive (Fig. la). Although the sensory cell bodies also react, a smaller fraction of them are strongly positive, and they exhibit greater variability in the intensity of staining than do their processes. More notably, however, none of the large myelinated motor axons of the ventral root demonstrate CA activity (Fig. la). This staining pattern is consistent with past observations.l5 In alternate serial sections stained for CE activity (Fig. lb), approximately 75% of the myelinated axons in the ventral motor root react positively. Twenty micron section thickness yielded improved visualization over 10 pm sections, in which about 50% of myelinated motor axons stain positively.15 The most likely explanation for this is as follows: Most peripheral nerve processes, whether sensory or motor, contain cholinesterases to some degree. In order to label motor axons preferentially, fixation and reaction conditions have traditionally been manipulated such that only that subpopulation of myelinated motor processes expressing the greatest amount of enzyme will react. Under these conditions, the low levels of CE contained in sensory processes are not detected, and the specificity of the stain is therefore maintained. Unfortunately, these conditions also exclude that fraction of motor axons with lower levels of CE, and many of the motor axons that do react do so quite weakly. Hence, the sensitivity has been compromised in an RESULTS AND DISCUSSION effort to maximize specificity of the stain for motor processes. Increasing section thickness from 10 pm to 20 pm Staining of Normal Dorsal Root Ganglia and Spinal Roots has somewhat compensated for this by decreasing the abilRat lumbar spinal nerve roots provide an excellent ity of the reaction product to diffuse away from the location model for demonstrating the reciprocal nature of the CAKE of the enzyme and by doubling the depth of the reaction tandem staining technique." Using this model, we at- product as viewed perpendicular to the plane of the section. tempted to improve definition of CE staining. Within the One concern we had about increasing section thickness was
42
Wilke et at.
Figure 3. Near serial sections of a fascicle in the infraorbital sensory nerve shown in Figure 2, A: The CA activity is clearly within the axoplasm of myelinated axons. 6: CE staining is present in the unmyelinated autonomic fibers which lie between the myelinated axons. The myelinated axons do not stain above background. x 280.
the possibility that background staining might become intensified and obscure our ability to accurately distinguish reactive from non-reactive axons. Comparison of panels a and b of Figure 1 shows that this is not the case. In those regions where the CA stain (Fig. la) has demonstrated the existence of only sensory fibers (within the dorsal root as discussed above), we see no CE reactivity on the serial section (Fig. lb). Both the increased section thickness and the increased reaction temperature have improved our CE staining results, enabling us to reduce the incubation time necessary for the development of a discernible reaction product to less than 1 hour. Although this study employed an overnight fixation technique, the improvements outlined above can generate adequate CE reactivity even when combined with a much shorter fixation regimen. Such a combination is compatible with the time constraints of surgery, and it has recently been used at this institution as an intraoperative aid to free muscle transplantation" and human nerve reconstruction (Matloub HS, Riley DA, Sanger JR, Yousif NJ, Bain JL: Intraoperative use of cholinesterase and carbonic anhydrase histochemical staining in nerve grafting and delayed repair, submitted for publication). To avoid alternating the thickness of serial sections, we considered increasing the section thickness for the CA procedure to 20 pm as well. We have found, however (unpublished observations), that CA staining efficiency is greatly reduced when the tissue is sectioned thicker than 10 pm. This is consistent with the fact that the reaction which lo-
Figure 4. Longitudinal section (10 pm) through the proximal stump of a regenerating infraorbitalnerve that had been transected 4 weeks previously. In this CA reacted section, two morphologically distinct zones are present: 1) Swollenzone. The nerve fibers in the proximal half (left) are intensely stained, but their myelinated sheaths show swelling and granular degeneration.2) Fragmented zone (right). Axons terminate as fragmented nonstainedprocesses in the distal enlargement. (Note also that a third zone, containing normal axon morphology, cannot be seen on this photomicrograph.It would lie further proximally, to the extreme left.) x290.
calizes the enzyme requires a liquid/gas interface, so that CO, can readily access the carbonic anhydrase and drive the staining reaction." Thus, we have concluded that the benefits attained from alternating section thickness (10 pm for
lnfraorbital Nerve Fiber Types
43
Figure 5. A higher magnification of the swollen zone of a fascicle in the regenerating nerve in Figure 4. This oblique section (10 pm) extends even more proximally (left) into a zone of somewhat normal morphology. CA reactive growth sprouts (arrows) are evident, arising from axons in the swollen zone. X 320.
CA, 20 Fm for CE) far outweigh the minor technical inconvenience encountered in doing so. Staining of Normal lnfraorbital Nerves
The infraorbital nerve is a branch of the maxillary portion of the trigeminal cranial nerve serving a purely sensory function. Staining for CA activity in sections of control infraorbital nerve verifies this (Figs. 2a, 3a). Nearly 100% of the myelinated fibers in each individual fascicle stains intensely CA positive, whereas very few of the myelinated processes are labelled when stained for CE reactivity (Fig. 3a,b). The fact that the infraorbital nerve contains an occasional CE-reactive myelinated axon indicates that at the level of our dissection, this nerve has already received some alpha motor fibers from the adjacent buccal branch of the facial motor nerve. In order to obtain a peripheral nerve specimen which is completely free from anastomosis with any motor processes whatsoever, it would have been necessary to isolate the nerve at a more proximal level. Previous work in our lab has shown the infraorbital nerve to be
exclusively sensory when excised at the level of its foramen. In our model, the majority of CE staining seen above background in the infraorbital nerve is that of unmyelinated fibers dispersed among the myelinated sensory processes (Figs. 2b, 3b). These have been shown previou~ly'~ to be sympathetic postganglionic fibers. Their presence in the infraorbital nerve is not surprising, since we would expect to find an extensive supply of both sensory and autonomic innervation to the area of the vibrissae in rats. In 1988, Yunshao and Shizhen demonstrated that human peripheral sensory nerves often display diffuse CE staining of unmyelinated autonornics to a greater degree than do their motor counterparts.l4 This has led some investigators to encourage the use of CE staining alone to distinguish sensory from motor fibers, based upon the assumption that myelinated CE reactivity identifies motor fibers, while diffuse unmyelinated CE reactivity identifies a fascicle which is principally sensory. Some words of caution are warranted regarding this approach. First, as discussed above, many
44
Wilke et al.
myelinated sensory processes also contain some minimal cholinesterase activity and can be falsely labelled positive if conditions are inappropriately controlled. Secondly, if resolution of the reaction product is so poor that no myelinated axons are labelled at all, some investigators would apply Yunshao’s results and suggest that any fascicles with dense unmyelinated reactivity could be assumed to be primarily sensory. Our results dispute this. While many of our infraorbital sensory fascicles do contain diffuse autonomic CE reactivity, it should be pointed out that some of them do not (Fig. 2b). Hence, we have shown that the degree of unmyelinated reactivity can vary among different sensory fascicles. Therefore, absence of unmyelinated reactivity does not, in and of itself, indicate that a fascicle is primarily motor. We conclude, then, that a tandem staining technique using both CA and CE provides much more reliable information about the identity of fiber types than the CE staining alone. Staining of Transected lnfraorbital Nerves and Their Regenerating Processes
Four weeks after transection, longitudinal sections of the regenerating proximal stump stained for CA activity demonstrate 3 distinct morphological zones: proximal intact, middle swollen, and distal fragmented (Figs. 4, 5). This is consistent with past observations. 16,20 The middle zone is approximately 2 to 3 mm in length and contains extensive swelling, disruption of the myelin sheaths, and phagocytic removal of degenerating axonal debris (Figs. 4, 5); yet, there is still a significant number of sensory processes positive for CA activity. In fact, many regenerating growth sprouts are present within this middle section (Fig. 5). This indicates that in the infraorbital model, sensory processes are metabolically functional and capable of regrowing across a site of surgical repair within 1 month of transection. Interestingly, unmyelinated (autonomic) CE reactivity was preserved only to the distal edge of the zone of normal axon morphology. It seems to be obscured within the middle zone by the swelling and degradation of the myelin, so that it is not stainable as far distally as the CApositive sensory regrowth sprouts. Hence, the CA stain allows us to visualize growing sensory processes (albeit sparsely) at a level where CE staining alone would seem to indicate complete absence of viable nerve fibers. This provides further evidence in support of our claim that more accurate information is attained by the CAKE staining combination than by the use of CE staining alone, and it extends our conclusions to encompass the staining of injured nerves. The present study demonstrates improved resolution of both nerve fiber type and viability when tandem CA and CE histochemical staining is performed on intact and regener-
ating peripheral nerves. For this reason, we anticipate an increase in the intraoperative use of this (and similar) tandem staining procedures during human surgical cases involving nerve reconstruction.
REFERENCES 1. Dodd J, Jessell TM: Axon guidance and the patterning of neuronal projections in vertebrates. Science 242:692-699, 1988. 2. Rush RA: Immunohistochemical localization of endogenous nerve growth factor. Nature 312:364-366, 1984. 3. Brushart TM, Seiler WA: Selective reinnervation of distal motor stumps by peripheral motor axons. Exp Neurol 97:289-300, 1987. 4. Sunderland S: The anatomical basis of nerve repair, in Jewett DL, McCarroll HR (ed): Nerve Repair and Regeneration. St. Louis, The C.V. Mosby Co., 1980, pp 14-35. 5 . Gruber H, Freilinger G, Holle J, Mandl, H: Identification of motor and sensory funiculi in cut nerves and their selective reunion. Br J Plastic Surg 29:70-73, 1976. 6. Levinthal R, Brown WJ, Rand RW: Comparison of fascicular, interfascicular, and epineurial suture techniques in the repair of simple nerve laceration. J . Neurosurg 47:744-750, 1977. 7. Donaso RS, Ballantyne JP, Hansen S: Regeneration of sutured human peripheral nerves. An electrophysiological study. J Neurol Neurosurg Psychiatry 42:97-106, 1979. 8. Suematsu N: Tubulation for peripheral nerve gap; its history and possibility. Microsurgery 10:7 1-74, 1989. 9. Riley DA, Lang DH: Carbonic anhydrase activity of human peripheral nerves: A possible histochemical aid to nerve repair. J Hand Surg 9A:112-120, 1984. 10. Riley DA, Ellis S, Bain JL: Carbonic anhydrase activity in skeletal muscle fiber types, axons, spindles, and capillaries of rat soleus and extensor digitorum longus muscles. J Histochem Cytochem 30: 12751288, 1982. 11. Oswald T, Riley DA: Peripheral nerve carbonic anhydrase activity and chronic acetazolamide treatment of rats. Brain Research 406:379384, 1987. 12. Karnovsky MJ, Roots L: A “direct-coloring’’ thiocholine method for cholinesterases. J Histochem Cytochem 12:219-221, 1964. 13. Gruber H, Zenker W: Acetylcholinesterase: Histochemical differentiation between motor and sensory nerve fibers. Brain Res 51:207-214, 1973. 14. Yunshao H, Shizhen Z: Acetylcholinesterase: A histochemical identification of motor and sensory fascicles in human peripheral nerve and its use during operation. Plast Reconstr Surg 82:125-130, 1988. 15. Riley DA, Sanger JR, Matloub HS, Yousif NJ, Bain JL, Moore GH: Identifying motor and sensory myelinated axons in rabbit peripheral nerves by histochemical staining for carbonic anhydrase and cholinesterase activities. Brain Res 453:79-88, 1988. 16. Sanger JR, Riley DA, Matloub HS, Yousif NJ, Bain JL, Moore GH: Effects of axotorny on the cholinesterase and carbonic anhydrase activities of axons in the proximal and distal stumps of rabbit sciatic nerves: A temporal study. Plast Reconstr Surg 87:7:26-738, 1991. 17. Lonnerholm G: Histochemical localization of carbonic anhydrase in mammalian tissue, in Tashian RE, Hewett-Emmett D (ed): Biology and Chemistry ofthe Carbonic Anhydrases. New York, N. Y. Academy of Science, 1984, pp 369-380. 18. Riley DA, Ellis S, Bain JL: Ultrastructural cytocheniical localization of carbonic anhydrase activity in rat peripheral sensory and motor nerves, dorsal root ganglia, and dorsal column nucllei. Neuroscience 13:189-206, 1984. 19. Sanger JR, Riley DA, Yousif NJ, Matloub HS, Bain JL: Histochemical staining of nerve endings as an aid to free muscle transplantation. Microsurgery 12:361-366, 1991. 20. Sunderland S : Nerves and Nerve Injury, 2nd ed. Edinburgh, Churchill-Livingstone, 1978, p 191.
HISTOCHEMICAL DISCRIMINATION OF FIBERS IN REGENERATING RAT INFRAORBITAL NERVE RUSSELL A. WILKE, DANNY A. RILEY, Ph.D., and JAMES R. SANGER, M.D.
Post-injury regrowth of peripheral nerves is currently receiving attention at all levels of investigation, from molecular’ to microanat~mical~-~ to It is known that after a peripheral nerve is severed the distal stump eventually degenerates, leaving behind Schwann cells and the connective tissue network of the original endoneuria. Although the Schwann cell morphology changes somewhat, these endoneurial tubes remain basically intact and provide a scaffolding upon which regrowth can be directed. Growth cones sprouting from the proximal stump are guided along this scaffolding, and the success of reinnervation depends at least partly upon association of these sprouts (either sensory or motor) with the appropriate endoneurial sheaths. Traumatized patients whose newly generated sprouts fail to successfully reinnervate the appropriate end organs will see limited (if any) recovery of clinical function. In such cases, histochemically-guided microsurgical reconstruction of the nerve may be indicated. In the past, investigators have been able to identify with s2
’
From the Department of Cellular Biology and Anatomy (R.A.W., D.A.R.) and the Department of Plastic and Reconstructive Surgery (J.R.S.), Medical College of Wisconsin, Milwaukee, WI. Address reprint requests to Danny A. Riley, Ph.D.. 8701 Watertown Plank Road, Milwaukee, WI 53226. Acknowledgments: This research was supported in part by NASA grant NAG 2-552 to D.A.R. R.A.W. is supported by the Medical Scientist Training Program at the Medical College of Wisconsin. The authors greatly appreciate the excellent assistant of Mr. G.R. Slocum with the photomicroscopy and figure preparation and S.Tjepkema-Burrows with plate reproduction. Received May 13, 1991; revision accepted August 9, 1991. 0 1992 Wiley-Liss, Inc.
a high degree of resolution the position of myelinated sensory axons within a mixed peripheral nerve fascicle by histochemically staining for axoplasmic carbonic anhydrase (CA).9-’1 Other workers have demonstrated the usefulness of staining for cholinesterase (CE) activity, in the localization of myelinated motor axons. 12-14 In recent years, Riley and coworkers have used a modification of this CE staining procedure in combination with the CA method to identify the relative positions of motor versus sensory processes in serial sections of peripheral nerve. 15*16 In the present study, we performed controlled microsurgical injury of rat infraorbital nerves and applied the CAKE tandem staining technique after 4 weeks of regrowth. Our goals were to further demonstrate the reciprocal nature of the stains, to highlight the morphological changes seen in a sensory nerve after 4 weeks of degeneratiodregeneration, and to address some current issues in the literature regarding the usefulness of the CE technique for labelling autonomic fibers. Herein, we also report further improvements in the CE staining method which were made in an effort to abbreviate the overall time requirements of the procedure. MATERIALS AND METHODS Animal Surgery and Specimen Preparation Nine male Sprague-Dawley rats (383 f. 25 g. b. wt.)
were prepared for surgery with sterile instruments and aseptic procedures (as outlined in NIH “Guide for the Care and Use of Laboratory Animals,” Pub. 85-23, and the Medical College of Wisconsin “Policy on Surgery and Post-Surgical Care”). The rats were anesthetized with a loading dose of
40
Wilke et al.
Figure 1. Reciprmiity of carbonic anhydrase (CA) and cholinesterase (CE) staining is demonstrated for myelinated axons in a lower lumbar spinal ganglion of a normal rat. A: A 10 pm section stained for CA activity. Approximately 50% of the myelinated sensory axons of the dorsal root (OR) are darkly reactive. Many of the neuronal cell
bodies exhibit moderate to dark staining. 6:A 20 p n section stained for C E activity. About 75% of the myelinated motor axons of the ventral root (VR) show reactive axoplasm. These motor axons do not contain CA activity, and the sensory axons do not possess CE activity above background. x 92.
50 mg/kg of intraperitoneal sodium pentobarbital (Nembutal) and also received 0.05 mg/kg of subcutaneous atropine. Intraperitoneal maintenance doses (0.5 ml) of sodium pentobarbital (6.5 mglml) were delivered as necessary. The left infraorbital nerve of each animal was exposed using a lateral skin incision, inferior to the caudal-most vibrissae, and isolated by blunt dissection under a binocular dissecting microscope. Care was taken not to include the buccal branch of the facial motor nerve, which lies immediately superficial to the ventral 1/3 of the infraorbital nerve. To contain later regrowth, the isolated infraorbital nerve was wrapped with sterile “Saran Wrap” brand plastic film (from Dow Chemicals; 0.5 mm polyvinylidene chloride) to form an impervious sheath of approximately 3 mm in length. This sheath was then wrapped with one loop of 6-0 silk suture, tied tightly, and transected immediately distal to the ligature, severing the nerve. The wounds were closed and the rats were housed on soft sawdust. Wounds were inspected daily for infection. Following these precautions, and those listed in our institution’s “Animal Use Bulletin on Post-Operative Care in Rodents,” no problems developed that required elimination of rats from the study. Four weeks later, the animals were sacrificed and the proximal stump of each left infraorbital nerve was removed by transection approximately 0.5 cm proximal to the ligature. Right infraorbital nerves were harvested at similar levels and used as controls. Excised nerve segments were immediately transferred to 20 ml vials for immersion fixation in 2.5% glutaraldehyde, 0.1 M phosphate buffer, pH 7.4 at room temperature. Preliminary tests of fixation by 11% paraformaldehyde, both with and without 0.2% glutaraldehyde, yielded unsatisfactory results with the CE stain; sections showed diffuse background staining and poor resolution of myelinated axons. After 15 minutes, the vials were
transferred to a 5°C refrigerator for 1 hour 45 minutes. Following fixation, the specimens were rinsed 3 times in buffer and subjected to the following sucrose cryoprotection series: 10% sucrose for 20 minutes, 20% sucrose for 40 minutes, 30% sucrose overnight-all in 0.1 M phosphate buffer, pH 7.4. The next morning, tissues were attached to index cards, quick-frozen in freon, and stored under liquid nitrogen as described previously. l 5 Lower lumbar dorsal root ganglia and the accompanying dorsal and ventral roots were also removed from 3 of the animals following euthanasia. These tissues were fixed, cryoprotected, and frozen as described above; they were used in preliminary studies aimed at further refinement of the CE staining protocol. Histochemical Staining for Carbonic Anhydrase and Cholinesterase Activities
After removal from storage in liquid nitrogen, control and injured nerve segments were sectioned either longitudinally or transversely with a cryostat microtome at -22°C. Staining was found to be optimal if the thickness of the sections was alternated, 10 pm for CA and 20 pm for CE. As in previous work,15 the distance sectioned from the free end of the transected proximal stumps was carefully recorded so that any observed morphological changes could be accurately located along the length of the injured nerve. Serial sections were collected on 2% gelatin-coated glass slides and stained histochemically for either CA or CE activity. Staining for CA activity was conducted following the improved method of Riley and coworkers. 10711 In this method, tissue carbonic anhydrase converts CO, in the incubation medium to H,CO,, changing the pH in the microenvironment of the enzyme and ultimately {causinglocal
lnfraorbital Nerve Fiber Types
41
Figure 2. Serial cross-sections of fascicles in a normal infraorbital nerve. A: Every fascicle is filled with CA reactive myelinated sensory axons. B: The fascicles contain differing amounts of CE reactive structures. x 68.
precipitation of cobalt s a k 7 It should also be noted that while this study was completely light microscopic, the procedures for fixation and staining are compatible with electron microscopy." Staining for CE activity was conducted by a modification of the original method of Karnovsky and Roots. l 2 This method supplies acetylthiocholine as a substrate for tissue cholinesterases. The thiocholine released participates in a reduction that ultimately deposits copper ferrocyanide at the site of the enzyme. Details of our previous modifications have been published earlier.15In the present study, we further enhanced the relatively weak CE staining by elevating the temperature of the 1-hour tissue incubation from 22°C to 46°C (suggested in reference 14) and by increasing section thickness from 10 pm to 20 pm. Following incubation, the tissue was rinsed in distilled water and covered with 1% AgNO,. Subsequent exposure to bright illumination under a fiber optic light (100 W halogen bulb) promotes the reduction of silver over the copper ferrocyanide precipitate to intensify the staining. Finally, the CA- and CE-stained sections were rinsed, dehydrated, and mounted permanently using pre- 1978 Fisher Permount. Post- 1978 Permount oxidizes and rapidly fades the CA reaction product unless retarded by an added antioxidant.l 1 We have also found that Canadian balsam is a satisfactory nonoxidizing substitute medium.
ganglia, approximately half of the myelinated sensory axons of the dorsal nerve root stained intensely CA positive (Fig. la). Although the sensory cell bodies also react, a smaller fraction of them are strongly positive, and they exhibit greater variability in the intensity of staining than do their processes. More notably, however, none of the large myelinated motor axons of the ventral root demonstrate CA activity (Fig. la). This staining pattern is consistent with past observations.l5 In alternate serial sections stained for CE activity (Fig. lb), approximately 75% of the myelinated axons in the ventral motor root react positively. Twenty micron section thickness yielded improved visualization over 10 pm sections, in which about 50% of myelinated motor axons stain positively.15 The most likely explanation for this is as follows: Most peripheral nerve processes, whether sensory or motor, contain cholinesterases to some degree. In order to label motor axons preferentially, fixation and reaction conditions have traditionally been manipulated such that only that subpopulation of myelinated motor processes expressing the greatest amount of enzyme will react. Under these conditions, the low levels of CE contained in sensory processes are not detected, and the specificity of the stain is therefore maintained. Unfortunately, these conditions also exclude that fraction of motor axons with lower levels of CE, and many of the motor axons that do react do so quite weakly. Hence, the sensitivity has been compromised in an RESULTS AND DISCUSSION effort to maximize specificity of the stain for motor processes. Increasing section thickness from 10 pm to 20 pm Staining of Normal Dorsal Root Ganglia and Spinal Roots has somewhat compensated for this by decreasing the abilRat lumbar spinal nerve roots provide an excellent ity of the reaction product to diffuse away from the location model for demonstrating the reciprocal nature of the CAKE of the enzyme and by doubling the depth of the reaction tandem staining technique." Using this model, we at- product as viewed perpendicular to the plane of the section. tempted to improve definition of CE staining. Within the One concern we had about increasing section thickness was
42
Wilke et at.
Figure 3. Near serial sections of a fascicle in the infraorbital sensory nerve shown in Figure 2, A: The CA activity is clearly within the axoplasm of myelinated axons. 6: CE staining is present in the unmyelinated autonomic fibers which lie between the myelinated axons. The myelinated axons do not stain above background. x 280.
the possibility that background staining might become intensified and obscure our ability to accurately distinguish reactive from non-reactive axons. Comparison of panels a and b of Figure 1 shows that this is not the case. In those regions where the CA stain (Fig. la) has demonstrated the existence of only sensory fibers (within the dorsal root as discussed above), we see no CE reactivity on the serial section (Fig. lb). Both the increased section thickness and the increased reaction temperature have improved our CE staining results, enabling us to reduce the incubation time necessary for the development of a discernible reaction product to less than 1 hour. Although this study employed an overnight fixation technique, the improvements outlined above can generate adequate CE reactivity even when combined with a much shorter fixation regimen. Such a combination is compatible with the time constraints of surgery, and it has recently been used at this institution as an intraoperative aid to free muscle transplantation" and human nerve reconstruction (Matloub HS, Riley DA, Sanger JR, Yousif NJ, Bain JL: Intraoperative use of cholinesterase and carbonic anhydrase histochemical staining in nerve grafting and delayed repair, submitted for publication). To avoid alternating the thickness of serial sections, we considered increasing the section thickness for the CA procedure to 20 pm as well. We have found, however (unpublished observations), that CA staining efficiency is greatly reduced when the tissue is sectioned thicker than 10 pm. This is consistent with the fact that the reaction which lo-
Figure 4. Longitudinal section (10 pm) through the proximal stump of a regenerating infraorbitalnerve that had been transected 4 weeks previously. In this CA reacted section, two morphologically distinct zones are present: 1) Swollenzone. The nerve fibers in the proximal half (left) are intensely stained, but their myelinated sheaths show swelling and granular degeneration.2) Fragmented zone (right). Axons terminate as fragmented nonstainedprocesses in the distal enlargement. (Note also that a third zone, containing normal axon morphology, cannot be seen on this photomicrograph.It would lie further proximally, to the extreme left.) x290.
calizes the enzyme requires a liquid/gas interface, so that CO, can readily access the carbonic anhydrase and drive the staining reaction." Thus, we have concluded that the benefits attained from alternating section thickness (10 pm for
lnfraorbital Nerve Fiber Types
43
Figure 5. A higher magnification of the swollen zone of a fascicle in the regenerating nerve in Figure 4. This oblique section (10 pm) extends even more proximally (left) into a zone of somewhat normal morphology. CA reactive growth sprouts (arrows) are evident, arising from axons in the swollen zone. X 320.
CA, 20 Fm for CE) far outweigh the minor technical inconvenience encountered in doing so. Staining of Normal lnfraorbital Nerves
The infraorbital nerve is a branch of the maxillary portion of the trigeminal cranial nerve serving a purely sensory function. Staining for CA activity in sections of control infraorbital nerve verifies this (Figs. 2a, 3a). Nearly 100% of the myelinated fibers in each individual fascicle stains intensely CA positive, whereas very few of the myelinated processes are labelled when stained for CE reactivity (Fig. 3a,b). The fact that the infraorbital nerve contains an occasional CE-reactive myelinated axon indicates that at the level of our dissection, this nerve has already received some alpha motor fibers from the adjacent buccal branch of the facial motor nerve. In order to obtain a peripheral nerve specimen which is completely free from anastomosis with any motor processes whatsoever, it would have been necessary to isolate the nerve at a more proximal level. Previous work in our lab has shown the infraorbital nerve to be
exclusively sensory when excised at the level of its foramen. In our model, the majority of CE staining seen above background in the infraorbital nerve is that of unmyelinated fibers dispersed among the myelinated sensory processes (Figs. 2b, 3b). These have been shown previou~ly'~ to be sympathetic postganglionic fibers. Their presence in the infraorbital nerve is not surprising, since we would expect to find an extensive supply of both sensory and autonomic innervation to the area of the vibrissae in rats. In 1988, Yunshao and Shizhen demonstrated that human peripheral sensory nerves often display diffuse CE staining of unmyelinated autonornics to a greater degree than do their motor counterparts.l4 This has led some investigators to encourage the use of CE staining alone to distinguish sensory from motor fibers, based upon the assumption that myelinated CE reactivity identifies motor fibers, while diffuse unmyelinated CE reactivity identifies a fascicle which is principally sensory. Some words of caution are warranted regarding this approach. First, as discussed above, many
44
Wilke et al.
myelinated sensory processes also contain some minimal cholinesterase activity and can be falsely labelled positive if conditions are inappropriately controlled. Secondly, if resolution of the reaction product is so poor that no myelinated axons are labelled at all, some investigators would apply Yunshao’s results and suggest that any fascicles with dense unmyelinated reactivity could be assumed to be primarily sensory. Our results dispute this. While many of our infraorbital sensory fascicles do contain diffuse autonomic CE reactivity, it should be pointed out that some of them do not (Fig. 2b). Hence, we have shown that the degree of unmyelinated reactivity can vary among different sensory fascicles. Therefore, absence of unmyelinated reactivity does not, in and of itself, indicate that a fascicle is primarily motor. We conclude, then, that a tandem staining technique using both CA and CE provides much more reliable information about the identity of fiber types than the CE staining alone. Staining of Transected lnfraorbital Nerves and Their Regenerating Processes
Four weeks after transection, longitudinal sections of the regenerating proximal stump stained for CA activity demonstrate 3 distinct morphological zones: proximal intact, middle swollen, and distal fragmented (Figs. 4, 5). This is consistent with past observations. 16,20 The middle zone is approximately 2 to 3 mm in length and contains extensive swelling, disruption of the myelin sheaths, and phagocytic removal of degenerating axonal debris (Figs. 4, 5); yet, there is still a significant number of sensory processes positive for CA activity. In fact, many regenerating growth sprouts are present within this middle section (Fig. 5). This indicates that in the infraorbital model, sensory processes are metabolically functional and capable of regrowing across a site of surgical repair within 1 month of transection. Interestingly, unmyelinated (autonomic) CE reactivity was preserved only to the distal edge of the zone of normal axon morphology. It seems to be obscured within the middle zone by the swelling and degradation of the myelin, so that it is not stainable as far distally as the CApositive sensory regrowth sprouts. Hence, the CA stain allows us to visualize growing sensory processes (albeit sparsely) at a level where CE staining alone would seem to indicate complete absence of viable nerve fibers. This provides further evidence in support of our claim that more accurate information is attained by the CAKE staining combination than by the use of CE staining alone, and it extends our conclusions to encompass the staining of injured nerves. The present study demonstrates improved resolution of both nerve fiber type and viability when tandem CA and CE histochemical staining is performed on intact and regener-
ating peripheral nerves. For this reason, we anticipate an increase in the intraoperative use of this (and similar) tandem staining procedures during human surgical cases involving nerve reconstruction.
REFERENCES 1. Dodd J, Jessell TM: Axon guidance and the patterning of neuronal projections in vertebrates. Science 242:692-699, 1988. 2. Rush RA: Immunohistochemical localization of endogenous nerve growth factor. Nature 312:364-366, 1984. 3. Brushart TM, Seiler WA: Selective reinnervation of distal motor stumps by peripheral motor axons. Exp Neurol 97:289-300, 1987. 4. Sunderland S: The anatomical basis of nerve repair, in Jewett DL, McCarroll HR (ed): Nerve Repair and Regeneration. St. Louis, The C.V. Mosby Co., 1980, pp 14-35. 5 . Gruber H, Freilinger G, Holle J, Mandl, H: Identification of motor and sensory funiculi in cut nerves and their selective reunion. Br J Plastic Surg 29:70-73, 1976. 6. Levinthal R, Brown WJ, Rand RW: Comparison of fascicular, interfascicular, and epineurial suture techniques in the repair of simple nerve laceration. J . Neurosurg 47:744-750, 1977. 7. Donaso RS, Ballantyne JP, Hansen S: Regeneration of sutured human peripheral nerves. An electrophysiological study. J Neurol Neurosurg Psychiatry 42:97-106, 1979. 8. Suematsu N: Tubulation for peripheral nerve gap; its history and possibility. Microsurgery 10:7 1-74, 1989. 9. Riley DA, Lang DH: Carbonic anhydrase activity of human peripheral nerves: A possible histochemical aid to nerve repair. J Hand Surg 9A:112-120, 1984. 10. Riley DA, Ellis S, Bain JL: Carbonic anhydrase activity in skeletal muscle fiber types, axons, spindles, and capillaries of rat soleus and extensor digitorum longus muscles. J Histochem Cytochem 30: 12751288, 1982. 11. Oswald T, Riley DA: Peripheral nerve carbonic anhydrase activity and chronic acetazolamide treatment of rats. Brain Research 406:379384, 1987. 12. Karnovsky MJ, Roots L: A “direct-coloring’’ thiocholine method for cholinesterases. J Histochem Cytochem 12:219-221, 1964. 13. Gruber H, Zenker W: Acetylcholinesterase: Histochemical differentiation between motor and sensory nerve fibers. Brain Res 51:207-214, 1973. 14. Yunshao H, Shizhen Z: Acetylcholinesterase: A histochemical identification of motor and sensory fascicles in human peripheral nerve and its use during operation. Plast Reconstr Surg 82:125-130, 1988. 15. Riley DA, Sanger JR, Matloub HS, Yousif NJ, Bain JL, Moore GH: Identifying motor and sensory myelinated axons in rabbit peripheral nerves by histochemical staining for carbonic anhydrase and cholinesterase activities. Brain Res 453:79-88, 1988. 16. Sanger JR, Riley DA, Matloub HS, Yousif NJ, Bain JL, Moore GH: Effects of axotorny on the cholinesterase and carbonic anhydrase activities of axons in the proximal and distal stumps of rabbit sciatic nerves: A temporal study. Plast Reconstr Surg 87:7:26-738, 1991. 17. Lonnerholm G: Histochemical localization of carbonic anhydrase in mammalian tissue, in Tashian RE, Hewett-Emmett D (ed): Biology and Chemistry ofthe Carbonic Anhydrases. New York, N. Y. Academy of Science, 1984, pp 369-380. 18. Riley DA, Ellis S, Bain JL: Ultrastructural cytocheniical localization of carbonic anhydrase activity in rat peripheral sensory and motor nerves, dorsal root ganglia, and dorsal column nucllei. Neuroscience 13:189-206, 1984. 19. Sanger JR, Riley DA, Yousif NJ, Matloub HS, Bain JL: Histochemical staining of nerve endings as an aid to free muscle transplantation. Microsurgery 12:361-366, 1991. 20. Sunderland S : Nerves and Nerve Injury, 2nd ed. Edinburgh, Churchill-Livingstone, 1978, p 191.
