Injury to nerves and the initiation of amphibian limb regeneration

THE AMERICAN JOURNAL OF ANATOMY 169973-284 (1984)

Injury to Nerves and the Initiation of Amphibian Limb Regeneration ANTHONY L. MESCHER Medical Sciences Program, Anatomy Section, Indiana University School of Medicine, Bloomington, Indiana 47405

ABSTRACT The possibility has been investigated that nerves, which promote mitotic activity during the growth phase of limb regeneration, may also release factors upon injury to stimulate the cellular events during the initiation of regeneration. These events have been compared in control, innervated limb stumps and in limbs denervated for various periods before amputation. Wound closure occurred in a n essentially normal manner in limbs denervated a t the time of amputation but frequently occurred a t reduced rates or not a t all in limb stumps denervated 5 days before amputation. Tissue loss and various degrees of morphological regression were seen in many of the predenervated limb stumps, including all of those denervated for 6 days or more before amputation. Evidence is presented which suggests that regression may result from a combination of the denervated state and disturbance to the limb’s vascular system. Limbs predenervated 2-10 days before amputation were fixed 7 days after amputation and examined histologically. All limb stumps, including those involved in regression, showed tissue dissociation and cellular dedifferentiation. Autoradiography revealed that the dedifferentiation involved DNA synthesis in both denervated and innervated limbs. Seven days after amputation, DNA-labeling indices in limb stumps predenervated 2 or 4 days were similar to those of control limbs, but labeling indices were significantly reduced in limbs predenervated 6 or more days. The results are assessed in light of the state of nerve degeneration in the limbs at the time of amputation and are discussed in terms of what is known regarding nerve-derived growthpromoting substances. When limbs of most urodele amphibians are amputated, conditions are established which lead to extensive cell proliferation in the distal stump. This results in formation of a blastema resembling the embryonic limb bud. Cellular interactions leading to normal limb pattern formation take place within the growing blastema, resulting in the eventual development of a n anatomically complete appendage (Tank and Holder, 1981). Three factors are known with certainty to be required for the blastema-formation phase of limb regeneration (Thornton, 1968; Tassava and Mescher, 1975). First, tissue injury is necessary to initiate the process of “cellular dedifferentiation” by which mesenchymal cells arise from the internal tissues of the limb stump (Hay, 1974). Second, a “trophic” stimulus from the nerves is necessary to promote mitotic activity in the dedifferentiated cells (Singer, 1952, 1978; Mescher and Tassava, 1975). Third, a n epidermally derived epithe-

(c)

1984 ALAN R. LISS, INC.

lium, which closes the amputation wound and immediately overlies the injured internal tissues, is necesary to ensure continued proliferation and distal accumulation of the mesenchymal cells to form the blastema (Mescher, 1976; Tassava and Olsen, 1982). Of these three prerequisites, the role of injury has received perhaps the least attention from investigators of limb regeneration, and as a result factors that mediate the injury effect in this system and in other reparative processes are poorly understood. Amputation brings about a very complex sequence of events, which includes migration of the wound epithelium across the cut surface and local dissociation of muscle and connective tissue (Schmidt, 1968). Release of dedifferen-

Address reprint requests to Dr. A. L. Mescher, MSPIAnatomy, Indiana University, Bloomington, IN 47405. Received May 12,1983. Accepted September 20, 1983.

274

A.L. MESCHER

tiated cells from the dissociating tissues is accompanied by entry of these cells into the cell-proliferation cycle (Hay and Fischman, 1961). As suggested repeatedly in the older literature on wound healing (reviewed by Needham, 1952; Washburn, 19601, epithelial closure of a wound and renewed proliferative activity may be controlled by factors elicited locally by tissue damage. Substances capable of stimulating cell proliferation and cell migration during injury-induced growth are likely to appear from several sources. For example, components of the eroding connective-tissue intercellular matrix, in particular that of bone (Rath and Reddi, 19791, are known to be involved in the control of localized growth during tissue repair (reviewed by Shoshan, 1981). Moreover, various mitogenic substances are released at the site of injury during blood clotting and the inflammatory response, including platelet-derived growth factor (Ross and Vogel, 1978) and polypeptides from macrophages and lymphocytes (Oppenheim et al., 1982). Promoters of cell cycle reentry during dedifferentiation may also be released from nerves upon injury. As noted above, nerves are a major source of growth factors during blastema formation (Thornton, 1968; Singer, 1978). They are also known to release mitogenic agents that affect associated cells during the process of nerve degeneration following transection (reviewed by Abercrombie, 1957; Asbury, 1975). Using Ambystoma larvae, Karczmar (1946) showed that forelimb denervation 10 days before amputation greatly reduced the extent of dedifferentiation in stump tissues and concluded that nerves are somehow involved in promoting this process. More recent experiments in which denervation and amputation were performed on the same day (Tassava et al., 1974; Mescher and Tassava, 1975) indicated that intact nerves (i.e., axons in the limb stump still attached to perikarya) are not needed for epithelial closure, dedifferentiation, or cell cycle reentry. Those studies, however, could not rule out the possible involvement of factors released from degenerating nerves, because degeneration distal to the site of denervation was concurrent with the tissue response to amputation, which always includes retrograde nerve degeneration (Schmidt, 1968). The present study was designed to elucidate a stimulatory role for growth factors released from injured, degenerating nerves

in the cellular events that occur upon amputation of newt forelimbs. Scanning electron microscopy was used to assess the effect on wound closure of denervation performed either a t the time of amputation or 5 days earlier. DNA-labeling indices 1 week after amputation were determined in an attempt to quantify the effect of varying periods of denervation on cell cycle reentry by the distal mesodermal cells of amputated limbs. MATERIALS AND METHODS

All experiments involved the forelimbs of adult newts (Notophthalmus uiridescens), which were maintained in aerated tap water a t 25°C. All operations were done using tricaine methanesulfonate (Sigma) anesthesia. Denervations were unilateral and involved transecting brachial nerves 3, 4, and 5 just distal to the brachial plexus. Amputations were always bilateral and through the proximal one-third of the zeugopodia. Protruding ends of the radius and ulna, if present, were trimmed back t o the level of the cut soft tissues. The effect of limb denervation on the process of wound epithelium formation was examined. Twelve newts were denervated a t the time of amputation, and another 12 were denervated 5 days before amputation. Denervated and control (innervated) limbs from both groups of animals were fixed 10 hr after amputation for scanning electron microscopic examination of the epithelialization process. After overnight fixation in 2.5% glutaraldehyde in 0.1 M sodium phosphate buffer (pH 7.4), the specimens were rinsed in buffer, dehydrated in increasing concentrations of ethyl alcohol, critical-point dried with COz, mounted on specimen stubs, and vacuum coated with gold-palladium. The limbs were examined with a Jeol-35 scanning electron microscope at a beam voltage of 15 kV. TO study the effect of varying periods of denervation on the initiation of limb regeneration, denervation was performed on three newts on each of days 1 0 , 8 , 6 , 4 , and 2 prior to amputation. The limbs of three additional newts were denervated on the day of amputation. Limbs which were to be kept nerveless for 8 and 10 days before amputation were redenervated 6 days after the original nerve transection. One week after amputation each newt was injected intraperitoneally with [meth~l-~H thymidine ] (ICN Pharmaceuticals; 72 Cilmmol) a t a dose of 5 pCi/gm body

NERVE INJURY AND LIMB REGENERATION

275

weight. After a n incorporation period of 3 hr, forelimbs were removed a t the shoulder and fixed as pairs in Carnoy's fixative. The limbs were processed for paraffin histology and sectioned longitudinally a t 10pm. Sections from representative levels of each limb were prepared autoradiographically with Kodak NTB2 emulsion, using the method described previously (Yachnis and Mescher, 1982).The DNA-labeling index was determined for each limb; in several nonadjacent sections an average of 1000 mesodermal cells were sampled in an area approximately 2 mm proximal to the apical epithelium as previously described (Mescher and Tassava, 1975). A t test was used to determine the statistical significance of the DNA-labeling-index data. Other sections were stained to demonstrate the condition of the brachial nerves by a method based on that of Snodgress et al. (1961) for degenerating nerves. Briefly, slides were treated with 1%aqueous iron alum for 30-60 min, stained in Luxon fast blue (Eastman Chemicals solvent blue 38) for 4 hr a t 60°C, rinsed alternately in 70% ethanol and 0.05% Li2C03, and counterstained in 0.5% alcoholic erythrosin. For gross observations of the effect on limbs of disturbance to the arterial supply, forelimbs of 1 group of 10 newts were unilaterally devascularized. This was done by cutting

Wound closure Observations with a dissecting microscope indicated that limbs denervated on the day of amputation underwent wound contraction and epithelial closure of the amputation surface in a manner comparable to that of innervated control limb stumps. Scanning electron microscopy of the denervated limb stumps fixed 10 hr after amputation revealed in most cases complete or nearly complete closure of the amputation surface by a combination of skin contraction and epithelialization. Two cases in which wound closure was greatly delayed revealed a fibrous meshwork, apparently derived from clotted blood, across most of the wound surface with a n epithelial sheet migrating across this substrate from the basal layer(s) of the surrounding epidermis (Fig. 1). Closure of these limb stumps (and of

Fig. 1. Amputation surfaces of two limb stumps denervated at the time of amputation and fixed 10 hr later. Cut ends of bones and muscle are covered with a fibrin-

like meshwork 0across which is migrating an epithelial sheet (es) from the basal layers of the surrounding epidermis. Both x 100.

the brachial artery where it enters the limb and allowing the blood to clot before returning the animal to water. Forelimbs of two other similar groups were either denervated as before, with care taken not to disturb the vasculature, or denervated and devascularized. These operations were not repeated. Seven days after the operations, all the forelimbs were amputated, and the animals were observed daily for 5 weeks. RESULTS

276

A.L. MESCHER

were observed daily following amputation for effects of the denervation period on the healing process. Within 48 hr after amputation, two of the denervated limb stumps in the groups predenervated for 8 and 10 days showed evidence of regression at the distal tip of the stump. This was indicated primarily by protrusion of the cut ends of the radius andor ulna through the wound epithelium. Ends of these bones had been trimmed back to the same level as the cut ends of muscle and skin a t the time of amputation, and exposure of bone through the wound epithelium was not seen on any of the control, innervated limb stumps. By 4 days after amputation, all the limb stumps predenervated for 6 or more days, but fewer than half of the stumps predenervated for shorter periods, showed various degrees of tissue loss at the distal ends. The severity of limb regression varied considerably. Some limb stumps had exposed tips of bone like that seen a t 2 days; but in others skin, muscle, and periosteum had disappeared proximal to the amputation a s far as the elbow, leaving long projections of eroded bone exposed. By 7 days after amputation, when the limbs were fixed, no additional limb stumps had begun to regress, but in some cases tisLimb regression sue loss had continued beyond that seen 3 Limbs of the animals denervated unilater- days earlier. The denervated limb stumps of ally 10,8,6,4,2, or 0 days before amputation four animals had regressed proximally to a

the contralateral control limb stumps) resembled in all essential features the process described in detail by Repesh and Oberpriller (1978,1980). Limb stumps denervated 5 days before amputation showed more variability with regard to closure of the amputation wound. Healing was complete or nearly complete in only 50% of the denervated limbs by 10 hr after amputation, and in many cases the epithelial sheet was more irregular than that covering the contralateral limb stump (Fig. 2). The remaining denervated limb stumps in this group showed more drastic retardation with regard to skin contraction and wound epithelium formation. In two of these cases the smooth, fibrin-like meshwork covering the amputation surface was lost to a large extent, which may have contributed to the delayed epithelial migration. Figure 3a shows such a denervated limb stump, which still resembled a freshly amputated limb at a time when the control limb (Fig. 3b) was essentially healed. As described below, many limbs denervated 4 days or more before amputation did not close their cut surfaces in a normal manner as late as 1 week after amputation.

Fig. 2. Distal ends, 10 hr after amputation, of a limb stump denervated 5 days before amputation (a) and the contralateral innervated control (b).The cut surface of the denervated limb is almost completely covered by an

epithelial sheet, but the epithelium appears to be more attenuated and to have involved less contraction of the old skin than that covering the control limb stump. Both X75.


Fig. 3 Distal ends, 10 hr after amputation, of a limb stump denervated 5 days before amputation (a) and the contralateral innervated control 01).The cut surface of the control limb is completely closed, but that of the

midhumeral level causing complete loss of the elbows and humeral epiphyses. Histological analysis revealed that in all cases of extensive limb regression, the soft tissues in the distal stump tapered away from the exposed bone(s) and remained covered by wound epithelium of varying thickness. Longitudinal sections of limbs which underwent severe regression are shown in Figures 4-6. The cause of the soft tissue regression shown by the predenervated limbs is not clear. It is of interest, however, that the occurrence and extent of the phenomenon were greater in the groups of forelimbs denervated for longer periods, 6 days or more before amputation. It may be significant that in two of these groups, the limbs had undergone a second denervation operation. While care was taken not to damage major arteries when performing the denervations, it is possible that disturbances of the vasculature, combined with the prolonged absence of nerves, may have predisposed the limbs to the regression process that set in following amputation. To test this possibility, the following experiment was performed. Left forelimbs of 10 newts were denervated as described above; left forelimbs of 10 more newts had the brachial artery cut a t the scapular region; and left forelimbs of another 10 newts underwent both denervation and brachial artery transection. Within 2-3 days, the limbs with only the arteries cut were revascular-

277

denervated limb shows little sign of either epithelialization or skin contraction. Except for a partial covering by a fibrin-like meshwork (0, the denervated stump at this time resembles a freshly amputated limb. Both X75.

ized, as indicated by examination with a dissecting microscope. The limbs with both arteries and nerves cut were at best only partially revascularized. One week after the operations, 6 of the devascularizeddenervated limbs and 1limb that was only denervated had been lost just distal to the shoulder, apparently due to necrosis. The remaining limbs and the right, unoperated limbs of these animals were amputated at this time. One week later 1 of the 4 remaining devascularizeddenervated limbs showed severe regression. None of the remaining denervated-only or devascularized-only limbs, and none of the control limbs, were regressing. Table 1indicates that 5 weeks after amputation, when all the control limbs were early digit-stage regenerates [stages according to Singer, (195211, all 10 of the devascularizedonly limbs were palette-stage regenerates; 1 of the denervated-only limbs had regenerated to early palette stage and the rest had healed over but had not regressed; and of the 4 remaining devascularizeddenervated limbs, 1(as noted above) had regressed, 2 had merely healed, and 1 had regenerated to early palette stage. These results show that major tissue loss or regression occurs more readily when denervation is combined with injury to the limb’s arterial supply. It is thus possible that vascular damage happening inadvertently during the denervation procedure may be involved in the limb regression

278

A.L. MESCHER

Figs. 4-6 Longitudinal sections of three control limbs (4a, 5a, 6a) and the contralateral limbs denervated 6 days (4b, 5b) or 8 days (6b) before amputation. Each limb was amputated at the same level of the zeugopodium and fixed 7 days later. Control limb stumps show histological features typical of 7-day regenerates, including tissue dissociation and cellular dedifferentiation in distal areas. The predenervated limb stumps show severe regression, as indicated by complete loss of the zeugopodia, loss of the distal humeral epiphyses, and considerable resorption of soft tissues distally in the stylopodia.

(Figure 4b is a more lateral section and does not include the humerus.) Distal areas of the regressing limb stumps show histolysis and cellular dedifferentiation,which was apparently similar to that of regenerating limbs and appeared to extend through more proximal areas. As shown in the enlargements (insets)of the indicated areas, many of the dedifferentiated cells in both the regenerating and the regressing limbs have nuclei labeled by L3H]thymidine given 3 hr before fixation. H&E-stained DNA autoradiographs. All ~ 2 6insets ; x 132.

279


TABLE 1. Effects of denervation a n d o r deuascularization on amputated newt forelimbs Number of limbs 5 weeks after amputation showing Operation (1 wk before Number amputation) of limbs Tissue loss Healing only Regeneration Devascularization Denervation Devascularization and denervation

10

0 1

0 8

10 1

10

7

2

1

10

seen in some cases following amputation of previously denervated adult newt forelimbs.

Histology, DNA labeling, and mitotic activity When the predenervated limbs fixed 1 week after amputation were examined histologically, dissociation and dedifferentiation of muscle and connective tissue were seen to have occurred distally in all cases, regardless of the time between denervation and amputation (Figs. 4-6). Dedifferentiated cells with large nuclei were observed throughout the area 1-2 mm proximal to the apical wound epithelium, dispersed among erythrocytes and other blood cells. The extent of dedifferentiation in the denervated-amputated limbs was very difficult to quantify and was complicated in most cases by the occurrence of tissue regression. Most histological features of the limbs denervated on the day of amputation or 2 days earlier were similar to those of control limb stumps, a result described and discussed previously (Mescher and Tassava, 1975). Denervated limb stumps, in which regression of soft tissues occurred also showed histolysis and cellular dedifferentiation distally (Figs. 4-6).In all cases the muscle and connective tissues approximately 2 mm proximal to the distal epithelia of denervated regressing limb stumps were dissociated, and undifferentiated mononuclear cells were present along with extravasated blood cells and tissue debris. The histological appearance strongly suggested that. the release of undifferentiated cells during tissue dissociation accompanies the morphological regression of denervated-amputated adult newt forelimbs. In this respect, the regressing denervated newt limb resembles denervated limbs of larval urodeles, which invariably regress upon amputation (Butler and Schotte, 1941)or crush injury (Thornton, 1953; Mescher, 1982). Autoradiography revealed that in all cases the distal populations of dedifferentiated cells included cells undergoing DNA synthesis. [3H]-Thymidine-labeled cells were seen throughout the area of dedifferentiated tissue, both in control limb stumps and in denervated-amputated limbs, including those

that underwent regression (Figs. 4-6). The mesodermally derived cells in the area of dedifferentiation were counted, and the percentage of those labeled (DNA-labeling index) was determined for each of these limbs fixed 1 week after amputation. As shown in Fig. 7, limbs denervated on the day of amputation or up to 4 days before amputation had DNA-labeling indices not significantly different from those of control innervated limb stumps, a result similar to that presented earlier (Mescher and Tassava, 1975). However, limbs denervated 6 days or more before amputation had significantly reduced labeling indices (Fig. 7). These data are consistent with the results obtained by Maden (1978) using larval axolotl limbs denervated 6 days prior to or on the day of amputation. The percentage of labeled cells 1 week after amputation was further decreased with longer periods of denervation before amputation (Fig. 7). The labeling index of limbs denervated 10 days prior to amputation was very

Days donbrvkiad

bOtat8 ~ m p U t W l l ~ l l

Fig. 7. DNA-labeling indices (5.1.) of distal limb stump areas 7 days after amputation in limbs denervated for various periods before amputation. DNA-labeling index is defined as the percentage of mesodermal cells, excluding erythrocytes, with labeled nuclei. Each point represents the mean of three or four limbs f the standard error. The shaded bar represents the mean DNA labeling index ? standard error of three innervated control limb stumps 7 days after amputation. Only the limb stumps denervated 6 days or more before amputation had DNA-labeling indices significantly (P< 0.005) lower than those of the controls.

280

A.L. MESCHER

low, despite the fact that histolysis and cellular dedifferentiation were seen in these limbs. At the time these limbs were fixed, mitotic activity had begun in the control, innervated limb stumps. Occasional mitotic figures also occurred in some limb stumps denervated as long as 4 days prior to amputation. These, however, were in no case as numerous as in the contralateral innervated limbs, despite the similarities of labeling indices in the limb stumps of these animals. No mitotic figures whatsoever were found in any of the limbs denervated 6 days or more before amputation. Nerve degeneration Since DNA-labeling indices were lower in the limbs denervated for longer periods, the condition of the nerves 7 days after amputation was examined in both the control limbs and limbs predenervated for 0-10 days. As described by Schmidt (1968), nerves of control limbs were dissociated and disorganized in the general area of tissue dedifferentiation near the level of amputation. Histologically, this retrograde nerve degeneration resembled the reaction shown by the distal regions of nerves that had been cut near the brachial plexus on the day of amputation or 2 days earlier (Fig. 8). Myelin sheaths were irregularly swollen, distorted, and separated from one another along most of the nerve’s length. These structures in nerves transected 4 days before amputation were somewhat more swollen and more varicose (Fig. 9). By the 15th day of nerve degeneration (in limbs denervated 8 days before amputation), dissolution of the myelin sheaths was characterized by the formation of irregular spheres and ovoids. As shown in Fig. 10, these changes were accompanied by the apparent loss of some myelin-staining material. Nerves cut 10 days before amputation showed areas of variable size in which the myelin had almost completely degenerated, leaving the appearance of empty or debris-containing sheaths

Figs. 8-11. Longitudinal sections of degenerating nerve in limbs denervated 2-10 days before amputation and fixed 7 days after amputation. Total periods of nerve degeneration were 9 days for Figure 8,11 days for Figure 9, 15 days for Figure 10, and 17 days for Figure 11. Histological features at these times were typical of myelinated nerve degeneration (Allt, 1976): progressive swelling and varicosity of the myelin sheaths accompanied by formation of spherical or ovoid “chambers” of

myelin undergoing degradation (arrows, Fig. 10) and gradual loss of myelin-staining material, after which the sheaths appear to contain varying amounts of granular debris. Luxol fast blue stain; all X170.


(Fig. 11).The histological features and extent of nerve degeneration observed a t these times were consistent with the description of newt brachial nerve degeneration given in detail by Singer and Steinberg (1972) and by Schmidt (1968). DISCUSSION

Observations of epithelial migration across the amputation surface in this study confirm previous findings (Singer and Inoue, 1964; Mescher and Tassava, 1975)that this process is not drastically affected by denervation a t the time of amputation. When limbs were denervated 5 days prior to amputation, however, the rate of epithelialization was significantly delayed in approximately 50% of the cases. Results with this group were quite variable, ranging from essentially normal closure of the amputation surface to failure of any epithelial migration to occur by 10 hr after amputation. Singer and Inoue (1964) reported the finding, based on work with 3 limbs, that denervation 6 days before amputation did not reduce the thickness of the stump’s wound epithelium 1 week after amputation. The cause of the subnormal epithelialization seen here in half the limb stumps denervated 5 days before amputation is not clear, but may be related to a need for proliferative activity in the epidermis surrounding the amputation site, since this is involved in normal completion and thickening of the wound epithelium (Hay and Fischman, 1961; Schmidt, 1968). It is possible that a stimulus from nerves may be involved in this epidermal cell proliferation. Grafted nervous tissue, living or killed, exerts a mitogenic affect on larval Ambystoma epidermis (Overton, 1950), and denervation of larval forelimbs on the day of amputation abolishes the threefold increase in DNA-labeling index that normally occurs in the epidermis immediately proximal to the level of amputation (Tassava and Mescher, 1976). The results reported here are only suggestive of a role for nerve-derived factors in promoting the formation of the apical wound epithelium. Analysis is complicated by the dissolution of the fibrin-like material from many of the denervated limb stumps. Migration of a n epidermal sheet across or through clotted blood is the normal process for epithelial closure of amputated newt limbs (Schmidt, 1968; Repesh and Oberpriller, 1978, 1980). The presence of a normal clot is not, however, a prerequisite for epithelialization of the amputation surface, since freshly amputated, cultured limb stumps

281

usually form wound epithelia in a normal (Bromley and Angus, 1971) but slightly delayed (Mescher, unpublished observations) manner. Moreover, there is considerable evidence (reviewed by Repesh and Oberpriller, 1980)that other components of a wound exudate or extracellular matrix besides fibrin may contribute to the fibrous meshworks across which epithelial cells migrate. Dissolution of the clots, therefore, probably was not the primary cause of the delayed healing in those denervated limb stumps that failed to close normally. Internal tissues of the denervated limb stumps showed dissociation and dedifferentiation that appeared histologically to be comparable to that of controls. Those observations have been discussed previously (Mescher and Tassava, 1975) and are similar to the histological findings of others with larval limbs (Butler and Schotte, 1941). Moreover, limbs denervated a s long as 4 days before amputation were found to have essentially normal DNA-labeling indices but very low levels of mitotic activity 7 days after amputation, which is also consistent with the previous work (Mescher and Tassava, 1975). Such observations, together with similar results from cell-cycle analyses of denervated larval forelimbs following amputation (Tassava et al., 1974) or crush injury (Mescher, 1982), have suggested that the neurotrophic influence is not necessary for cell cycle reentry or DNA replication during the period of cellular dedifferentiation, but may be required for a normal Gz phase and mitosis (Tassava and Mescher, 1975; Tassava and McCullough, 1978). Evidence supporting the hypothesis that a neurotrophic effect is particularly important for passage of the dedifferentiated, preblastemal cells through Gz has come from studies on the stimulation of mitotic activity in cultured nerve-dependent blastemas. Carlone and Foret (1979) have reported that addition of newt brain extract to such blastemas 2 days after denervation by explantation results in a burst of mitotic activity 8 hr later, a time consistent with stimulation of cells out of Gz. Attempts to find evidence for a specific cell block by Feulgen microdensitometric studies have proven inconclusive. Results with denervated “cone stage” blastemas from larval Ambystoma forelimbs showing a n accumulation of cells in GI (Maden, 1979) cannot be taken as evidence against a G2 neurotrophic effect during the initiation of regeneration, because the blastemas used were not shown to be nerve-dependent. Cone-stage blastemas such

282

A.L. MESCHER

as those used by Maden (1979) are nerve independent and continue to develop after denervation (Schotte and Butler, 1944; Butler and Schotte, 1949). A GI accumulation of cells (as indicated by 2C DNA content) in such denervated blastemas is the expected result in that experiment, since the cells are destined to differentiate normally (Butler and Schotte, 1949). Results of a similar study, but using nerve-dependent blastemas, have been reported by Loyd and Connelly (19811, who found no significant accumulation of cells in either GI or G2. Cellular dedifferentiation and DNA synthesis occurred in distal areas of all limb stumps denervated prior to amputation, including those in which tissue resorption was extensive. Similar cellular events have been reported during larval limb regression induced by simultaneous denervation and amputation (Tassava et al., 1974)or crush injury (Mescher, 1982). Regression of denervated larval limbs has been interpreted as resulting from cellular dedifferentiation continuing unchecked in the absence of a neurotrophic influence to cause blastema formation, with subsequent resorption of material (Butler and Schotte, 1941; Singer, 1952). Regression of adult newt limbs when denervated for various times prior to amputation has not been described in the literature, but seems to resemble the situation in larval limbs in that a distal-to-proximal process of dedifferentiation and tissue resorption is apparently involved. With adult forelimbs, however, the occurrence and extent of forelimb regression following denervation and amputation is much more variable than in larvae and may depend on a n interplay of factors. Schotte and Karczmar (1944) have shown that the regression rate in denervated-amputated larval limbs is influenced by the limb’s size and the “physiological age” of its constituent tissues. The data in Table 1 suggest that temporary derangement of the blood supply may be involved in regression of denervated-amputated adult newt limbs, a possibility also raised by Rose (1948) referring to unpublished observations. Wertz (1982) has reported that local X-irradiation of intact newt forelimbs with doses sufficient to inhibit regeneration also resulted, after a latent period of 3 to 4 weeks, in distal-toproximal regression. The degree of tissue loss varied, but was apparently comparable in severity to that seen in the present study. Further study of factors involved in the pre-

cipitation of limb regression could provide insights into the control of the dedifferentiative events normally leading to formation of the blastema. Analysis of the mechanisms by which the amputation injury initiates regeneration have been complicated by the difficulty in assessing quantitatively the multifaceted process of tissue and cellular dedifferentiation. Since in every tissue that will contribute cells to the blastema, dedifferentiation is accompanied by DNA synthesis (Hay and Fischman, 1961), determination of the percentage of mesodermal cells synthesizing DNA 1 week after amputation provides a means of evaluating the extent of dedifferentiation and cell-cycle entry induced by the amputation injury. As shown in Figure 7, DNA-labeling indices in the area of dedifferentiation were consistently reduced in limb stumps denervated 6 days or more before amputation, which is consistent with Maden’s (1978) study on axolotl larval limbs. Limbs kept without nerves for the longest time (10 days before amputation) showed the lowest level of labeling (2% L.1.). These results are consistent with the earlier morphological and histological observations of Karczmar (1946), who reported a general retardation of dedifferentiation when larval Ambystoma forelimbs were denervated 10 days before amputation. Taken together, the data implicate the degenerating nerves in the limb as a source of factors that stimulate cycling of mesodermal cells after injury. When the nerves are transected at the brachial plexus and allowed to degenerate for 610 days, the stimulatory effect of limb amputation on the mesodermal tissues is reduced. It seems that in order for typical, injury-induced cell cycle entry to occur “the limb must not be rendered unresponsive by a prolonged denervation prior to its amputation” (Karczmar, 1946). Although denervation at any time prior to blastema formation inhibits regeneration due to the interruption of the supply of trophic factors promoting cell proliferation (Singer, 1952), denervation around the time of amputation allows dedifferentiation and cell-cycle entry to occur in a n approximately normal manner (Fig. 7) (Mescher and Tassava, 1975; Tassava and McCullough, 1978). Only when the period between denervation and amputation is prolonged is reduction observed in the extent of cell cycling induced by amputation. Such a period, during which the nerves are degen-


erating, would allow the elimination of factors that the nerves ordinarily release only when injured by the amputation. There are two possible mechanisms by which injured nerves could stimulate cellcycle entry by mesodermal cells at the amputation site, both of which are consistent with the data presented here. One involves the release of trophic factors from the degenerating axons, including the same factor(s) by which intact nerves normally promote mitotic activity in the dedifferentiated cells (Singer, 1978). The concept that transected nerves contain a limited and slowly released supply of trophic factor capable of stimulating the cellular events in denervated-amputated limbs has already been proposed by Maden (1978). Many axonal proteins are still present, apparently in granular endoneurial deposits, up to 5 weeks following sciatic nerve transection in rats (Schlaepfer and Hasler, 1979). Moreover, the trophic activity of soluble sciatic nerve proteins, as measured by the effect on protein synthesis in cultured muscle cells, is only slightly reduced in degenerating nerves 3 weeks after transection (Oh et al., 1980). Gradual depletion of axoplasmic factors or reduction in their rate of release following denervation would explain the reduced levels of cell-cycle entry in limb stumps predenervated for increasing lengths of time (Fig. 7). Another possible source of mitogenic factors from injured degenerating nerve consistent with the present data is the myelin sheath. The release of proliferation-promoting substances from transected degenerating mammalian nerve, with myelinated nerve consistently showing a much greater effect than nonmyelinated nerve, has been demonstrated by numerous in vivo studies (reviewed by Abercrombie and Santler, 1957, and by Allt, 1976)and by recent in vitro work (Salzer and Bunge, 1980; Yachnis and Mescher, 1982).Degradation of the myelin in brachial nerve of adult newts, which is quite extensive by 2 weeks after transection at the shoulder (Figs. 1 0 , l l ) (Singer and Steinberg, 1972), may therefore involve the loss of growth-promoting factors that are normally released only during the nerves’ degenerative reaction to the amputation injury. The possible existence in amphibian nerves of myelin-derived growth factors with a potential role in the cellular response to tissue damage is worthy of further investigation. One such substance may be fibroblast growth

283

factor, which is reportedly derived from a myelin protein (Westall et al., 1978; Gospodarowicz et al., 1982) and is known to be mitogenic for blastemal cells (reviewed by Mescher, 1983). In conclusion, results of the present study suggest the involvement of factors from injured nerves in the promotion of mesodermal cell cycling following limb amputation, a possibility that failed to emerge from previous work (Mescher and Tassava, 1975) with simultaneously amputated and denervated limbs. These findings, however, do not rule out the possible involvement of other factors in initiating the cellular events that follow amputation. Such substances may include, for example, hematogenous factors like insulin and other hormones in plasma (Vethamany-Globus et al., 1978), platelet-derived growth factor (Ross and Vogel, 19781, and mitogens released from macrophages (Greenburg and Hunt, 1978) or lymphocytes (Schmidt et al., 1982). Thus, while factors from degenerating nerves are apparently important, a complex interplay of other growthpromoting substances may also be involved in mediating the “injury effect” during the initiation of limb regeneration. ACKNOWLEDGMENTS

This work was supported by NIH Grants GM 27735 and GM 31080. The technical assistance of s. Turtil and R. Cohen is gratefully acknowledged. LITERATURE CITED Abercrombie, M. 1957 Localized formation of new tissue in an adult mammal. Symp. SOC.Exp. Biol., 11:235254. Abercrombie, M., and J.E. Santler 1957 An analysis of growth in nuclear population during Wallerian degeneration. J . Cell. Comp. Physiol., 50:429-450. Allt, G. 1976 Pathology of the peripheral nerve. In: The Peripheral Nerve. D.N. Landon, ed. J. Wiley & Sons, New York, pp. 666-739. Asbury, A.K. 1975 The biology of Schwann cells. In: Peripheral Neuropathy, Vol. I. P.J. Dyke, P.K. Thomas and E.H. Lamert eds. W.B. Saunders, Philadelphia, pp. 201-212. Bromley, S.C., and D.J. Angus 1971 Early regenerative responses of amputated limbs of the newt, Notophthalmus uiridescens, to culture conditions. Dev. Biol., 26:652-657. Butler. E.G., and O.E. Schotte 1941 Histological alterations in denervated non-regenerating limbs of urodele larvae. J. Exp. Zool., 88:307-341. Butler, E.G., and O.E. Schotte 1949 Effects of delayed1 denervation on regenerative activity in limbs of uro-’ dele larvae. J. Exp. Zool., 112:361-392. Carlone, R.L., and J.E. Foret 1979 Stimulation of mitosis in cultured limb blastemata of the newt, NotophthaG mus uiridescens. J. Exp. Zool., 210:245-252. Gospodarowicz, D., G. M. Lui, and J. Cheng 1982 Purifi-

284

A.L. MICSCHER

cation in high yield of brain fibroblast growth factor by preparative isoelectric focusing at pH 9.6. J. Biol. Chem., 257:12,266-12,276. Greenburg, G.B., and T.K. Hunt 1978 The proliferative response in vitro of vascular endothelial and smooth muscle cells exposed to wound fluids and macrophages. J. Cell. Physiol., 97t353-360. Hay, E.D. 1974 Cellular basis of regeneration. In: Concepts of Development. J. Lash and J.R. Whittaker, eds. Sinauer Associates, Stamford, CT. pp. 404-428. Hay, E. D., and D. A. Fischman 1961 Origin of the blastema in regenerating limbs of the newt Triturus uiridescens. Dev. Biol., 3:26-59. Karczmar, A. G. 1946 The role of amputation and nerve resection in the regressing limbs of urodele larvae. J. Exp. Zool., 103t401-427. Loyd, R.M., and T.G. Connelly 1981 Microdensitometric analysis of denervation effects on newt limb blastema cells. Experientia, 37t967-969. Maden, M. 1978 Neurotrophic control of the cell cycle during amphibian limb regeneration. J. Embryol. Exp. Morphol., 48t169-175. Maden, M. 1979 Neurotrophic and X-ray blocks in the blastemal cell cycle. J. Embryol. Exp. Morphol., 50t169173. Mescher, A.L. 1976 Effects on adult newt limb regeneration of partial and complete skin flaps over the amputation surface. J. Exp. Zool., 195t117-128 Mescher, A. L. 1982 Neurotrophic control of events in injured forelimbs of larval urodeles. J. Embryol. Exp. Morphol., 69: 183-192. Mescher, A. L. 1983 Growth factors from nerves and their roles during limb regeneration. In: Limb Development and Regeneration. Part A. J.F. Fallon and A. I. Caplan, eds. Alan R. Liss, New York, pp. 501-512. Mescher, A. L., and R. A. Tassava 1975 Denervation effects on DNA replication and mitosis during the initiation of limb regeneration in adult newts. Dev. Biol., 44:187-197. Needham, A. E. 1952 Regeneration and Wound Healing. Methuen, London. Oh, T. H., G. J. Markelonis, P.J. Reier, and A. A. Zalewski 1980 Persistence in degenerating sciatic nerve of substances having a trophic influence upon cultured muscle. Exp. Neurol., 67t646-654. Oppenheim, J. J., B. M. Stadler, R. P. Siraganian, M. Mage, and B. Mathieson 1982 Lymphokines: their role in lymphocyte responses. Properties of interleukin 1. Fed. Proc., 41257-262. Overton, J. 1950 Mitotic stimulation of amphibian epidermis by underlying grafts of central nervous tissue, J. Exp. Zool., 115t521-559. Rath, N. C., and A. €3. Reddi 1979 Collagenous bone matrix is a local mitogen. Nature, 278t855-857. Repesh, L. A., and J. C. Oberpriller 1978 Scanning electron microscopy of epidermal cell migration in wound healing during limb regeneration in the adult newt, Notophthalmus uiridescens. Am. J. Anat., 15lt539-556. Repesh, L. A,, and J. C. Oberpriller 1980 Ultrastructural studies on migrating epidermal cells during the wound healing stage of regeneration in the adult newt, Noto. phthalmus uiridescens. Am. J. Anat., 159:187-208. Rose, S. M. 1948 The role of nerves in amphibian limb regeneration. N.Y. Acad. Sci.. 49:818-833. Ross, R., and A. Vogel 1978 The platelet-derived growth factor. Cell. 14t203-210 Salzer, J. L. and R. P:Bunge 1980 Studies on Schwann cell proliferation. I. An analysis in tissue culture of proliferation during development, Wallerian degeneration, and direct injury. J. Cell Biol., 84,739-752. Schlaepfer, W. W., and M. B. Hasler 1979 The persistence and possible externalization of axonal debris during Wallerian degeneration. J. Neuropath. Exp. Neurol., 38:242-252. Schmidt, A. J. 1968 Cellular Basis of Vertebrate Regen-

eration and Repair. University of Chicago Press, Chicago. Schmidt, J. A,, S. B. Mizel, D. Cohen, and I. Green 1982 Interleukin 1, a potential regulator of fibroblast proliferation. J. Immunol., 128:2177-2182. Schotte, 0. E., and E. G. Butler 1944 Phases in regeneration of the urodele limb and their dependence upon the nervous system. J. Exp. Zool., 97:95-121. Schotte, 0. E., and A. G. Karczmar 1944 Limb parameters and regression rates in denervated amputated limbs of urodele larvae. J. Exp. Zool., 97t43-70. Shoshan, S. 1981 Wound healing. Int. Rev. Connect. Tissue Res., 9t1-26. Singer, M. 1952 The influence of the nerve in regeneration of the amphibian extremity. Q. Rev. Biol., 27t169200. Singer, M. 1978 On the nature of the neurotrophic phenomenon in urodele limb regeneration. Am. Zool., 18:829-841. Singer, M., and S. Inoue 1964 The nerve and the epidermal apical cap in regeneration of the forelimb of adult Triturus. J. Exp. Zool., 155r105-116. Singer, M., and M. C. Steinberg 1972 Wallerian degeneration: A re-evaluation based on transected and colchicine-poisoned nerves in the amphibian, Triturus. Am. J. Anat., 133:51-84. Snodgress, A. B., C. H. Dorsey, and L. B. Lacey 1961 Luxol fast blue staining of degenerating myelinated fibers. Anat. Rec., 140:83-89. Tank. P. W.. and N. Holder 1981 Pattern reeulation in the’regenerating limbs of urodele amphibiak. Q. Rev. Biol., 56t113-142. Tassava, R. A., and W. D. McCullough 1978 Neural control of cell cycle events in regenerating salamander limbs. Am. Zool., 18t843-854. Tassava, R. A., and A. L. Mescher 1975 The roles of injury, nerves and the wound epidermis during the initiation of amphibian limb regeneration. Differentiation, 24r23-24. Tassava, R. A,, and A. L. Mescher 1976 Mitotic activity and nucleic acid precursor incorporation in denervated and innervated limb stumps of axolotl larvae. J. Exp. Zool., 195:253-262. Tassava, R. A., and C. Olsen 1982 Higher vertebrates do not regenerate digits and legs because their wound epidermis is not functional. Differentiation, 22t151155. Tassava, R. A,, L. L. Bennett, and G. D. Zitnik 1974 DNA synthesis without mitosis in amputated denervated forelimbs of larval axolotls. J. Exp. Zool., 19Ot111116. Thornton, C. S. 1953 Histological modifications in denervated injured forelimbs of Amblystoma larvae. J. Exp. Zool., 122:119-149. Thornton, C. S. 1968 Amphibian limb regeneration. Adv. Morphogen., 7t205-250. Vethamany-Globus, S.,M. Globus, and B. Tomlinson 1978 Neural and hormonal stimulation of DNA and protein synthesis in cultured regeneration blastemata in the newt, Notophthalmus uirzdescens. Dev. Biol. 65t183192. Washburn, W. W. 1960 Wound healing as a problem of growth. In: Fundamental Aspects of Normal and Malignant Growth. w . w. Nowinski, ed. Elsevier, Amsterdam, pp. 664-711. Wertz, R. L. 1982 Cellular proliferation in the skin of Xrayed newt limbs (with a note on X-ray-induced limb regression). J. Exp. Zool., 221,337-343. Westall, F. C., V. A. Lennon, and D. Gospodarowicz 1978 Brain-derived fibroblast growth factor: Identity with a fragment of the bask protein of myelin. Proc. Natl. Acad. Sci. U.S.A., 75t4675-4678. Yachnis, A. T., and A. L. Mescher 1982 Stimulation of DNA synthesis in Balb/c 3T3 cells by peripheral nerve degenerating in vitro. Exp. Neurol., 76t139-149.