THE JOURNAL OF COMPARATIVE NEUROLOGY 323~219-237 (1992)

Effects of Advancing Age on Peripheral Nerve Regeneration DEBORAH WHITTAKER VAUGHAN Department of Anatomy and Neurobiology, Boston University School of Medicine, Boston, Massachusetts 02118

ABSTRACT Following axotomy, the regrowth of peripheral axons takes longer in older individuals than in young ones. The present study compares the crush-induced process of degeneration and regeneration in the buccal branch of the facial motor nerve in groups of rats aged 3 months and 15 months. Observations are based on qualitative and quantitative analyses of the nerve 20 mm from the site of injury in rats 1,2, 4,16,21, 28, and 56 days after crush. The buccal branch is purely motor and contains a unimodal population of about 1,600 axons commonly in a single fascicle. During the first 28 days post crush (dpc) in the 3-month animals, the progression of myelin and axon degeneration, myelin clearance, regrowth of axon sprouts, and axon maturation are relatively synchronized and uniform. In the older rats, the degeneration of myelin and axons, myelin clearance, and the appearance of axon sprouts at the site of sample are all delayed. In the younger animals, axon sprouts increase in numbers from their first appearance at 4 dpc through the 2 weeks examined following the restoration of whisking behavior. The numbers of regenerating older axons increase at a rate comparable to that in the younger animals through the time that bilaterally symmetrical whisking behavior is evident, but afterwards the number of axon sprouts decreases. At 2 months after crush the young animals have 30%more fibers in the buccal branch than control nerves, while the older animals have fewer than control numbers. In the 3-month regenerated nerve, 2 months post crush, 30% of the regenerated fibers are of very small caliber, less than 3 pm2 in cross sectional area, and typically these small axons have unusually thick myelin sheaths; the older nerves do not have such a skewed distribution of axon areas. The older regenerated axons at 2 months post crush have an unusually high density of microtubules compared to the younger regenerated ones (and controls), and the ratio of neurofilaments to microtubules is very low. The conclusions are that motor neurons in older animals regenerate damaged axons after a delay not apparent in the young; the strong regenerative response apparent initially in animals of both age groups is not maintained in the older animals; and the relationship between the numerical density of cytoskeletal elements and the axon cross-sectional area deviates from normal in the regenerated axons of the older animals. D 1992 WiIey-Liss, Inc. Key words: microtubules, neurofilaments, axon

The process of axotomy-induceddegeneration and regeneration in the peripheral nervous system has been characterized in great detail since Waller first described the process that bears his name (1850). When crushed, the distal portion of an axon degenerates and the neuron subsequently regrows an axon. Regeneration experiments can provide valuable insight about the effects of advancing age on nerve cells. Previous studies have shown that in the aging animal, peripheral regeneration is slower (McMartin, '83; Pestronk et al., '80; Black and Lasek, '79; Drahota and Gutmann, '61) and produces fewer axons (Tanaka and Webster, '91; Moyer et al., '60) than in the young adult, but details of these differences are limited. This present study is part of a series of studies analyzing the effects of advancing

o 1992 WILEY-LISS, INC.

age on peripheral nerve regeneration, to ascertain what a substantial metabolic challenge, such as regrowing an axon, can reveal about how advancing age affects the neuron. The subject for these studies is the facial motor system in the albino rat, and the responses of young adult animals to crushing the facial nerve have been compared to those of middle-aged animals. Although the distance through which axons must regenerate is equivalent in the two age groups, recovery of function is delayed in the older animals. The morphometric studies of axotomy-inducedresponses in the facial motor neuron cell bodies demonstrate that during the period of axon outgrowth the chromatolytic response in the Accepted May 15,1992.

220

D.W. VAUGHAN TABLE 1. Animals Examined ~

Sumval

times Animal dgr 3months 15months

~~~~~

1

2

4

10

16

21

28

dpc'

dpc

dpc

dpc

dpc

dpc

dpc

Zmonths postcrush

2

4 5

4

5 5

5

4

4

4

3 4

4

2

5

2

'dpc, days post crush

older animals is temporally delayed, although it is not qualitatively different in the two age groups. However, during the period of axon maturation, after the nasolabialis muscle has been reinnervated, the central responses reveal age-related differences of a more substantial nature (Vaughan, '90a). Enzyme cytochemical analyses of the facial motor nucleus also show that while the young respond to reinnervation with a rapid return to preaxotomy levels of enzyme activities, the older animals recover more gradually (Vaughan, '90b). Because of the importance of the muscle to the regenerative efforts of the neuron, studies have also been conducted on the nasolabialis muscle that moves the large vibrissae (morphometric analysis: White and Vaughan, '91a; enzyme cytochemical analysis: White and Vaughan, '91b). These studies have shown that recovery of this target muscle is successful in both age groups, and so it can be assumed that the muscle has become fully reinnervated and the trophic functions of the older neurons have not become severely compromised with advancing age. Given that the axon is the anatomical expression of the principal synthetic efforts of the neuron during regeneration, the aim of this study is to follow the axotomy-induced events in a peripheral nerve containing its axons during the period of degeneration and regeneration in young adult and middle-aged animals. The goal is to identify age-related deficiencies that may extend beyond a simple retardation of response.

Fig. 1. Two-micrometer thick transverse sections of the buccal branch of the facial nerve, 20 mm distal to the exit of the facial nerve from the stylomastoid foramen (and site of crush).The larger fascicle is the branch of the facial nerve and the smaller one, marked with anasterisk, is a branch of the sensory trigeminal. a: Nerve from the right side of the head, the control side. b: Nerve from the left side of a 3-month-old animal, 2 dpc, in which all fibers within the larger fascicle are degenerating. x 115.

animal's eye, and a second point 2 mm rostra1 to the first. An incision about 1 mm into the muscle along the ventral aspect of this nerve forms the wide end of the wedge-shaped block of tissue that is subsequently removed. The tissue block thus consists of the nerve upon the surface of a small block of the masseter muscle cut in a shape that allows the METHODS tissue to be processed and embedded in Araldite such that Animal preparation the nerve will be sectioned in its transverse plane, at its A total of 29 three-month-old and 33 fifteen-month-old caudal end. In animals of both age groups, this point lies 20 Sprague-Dawley derived male rats were used in this study. mm from the site of axon crush. The buccal branch of the facial nerve travels in a broad In all of the animals the left facial nerve was crushed between the tips of a Dumont No. 5 forceps at its exit from epineurium as one of two major fascicles of nerve fibers the stylomastoid foramen in a manner described previously illustrated in Figure 1. The smaller fascicle is of the (Vaughan, '90a). For surgery the animals were anesthe- maxillary branch of the trigeminal nerve and after the tized with Ketamine HC1 (75 mgikg) and Xylazine (10 facial nerve is crushed all of the nerve fibers within the mg/kg) delivered by intramuscular injection. Following larger fascicle degenerate, while in most animals, none of surgery the animals were allowed to survive for periods that the fibers in the smaller fascicle degenerate (Fig. lb). All of ranged from 1 day to 56 days. The numbers of animals in the qualitative and quantitative analyses of this investigaeach sample are listed in Table 1. At the selected survival tion were conducted on the fibers of the larger fascicle. The process of degeneration and regeneration was foltimes animals were anesthetized and fixed by a two-stage vascular perfusion of glutaraldehyde and formaldehyde in lowed in light and electron microscope preparations. Con0.1M cacodylate buffer according to procedures described trol material from animals of both ages was examined to previously (Vaughan, '90a). Blocks of tissue were removed determine the extent to which normal aging produces on the day following perfusion, after fixation had become differences in the number, morphology, and cytoskeleton of axons. complete. The buccal branch of the facial nerve contains motor Terminology axons responsible for moving the animals' large vibrissae This account of regeneration uses the following terms, (e.g., Semba and Eggar, '86). This nerve runs along the surface of the masseter muscle, passing beneath the eye and which are defined at this point to aid in understanding the toward the nose and is exposed by removing the skin from following sections. Axon sprouts. Axons that arise from the proximal tip of the cheek area of the animal's head. A tissue block is removed by cutting the nerve bundle first perpendicular to the interrupted axon, as well as from the regenerating its long axis at a point in line with the caudal extent of the branches within the distal axon; included in this term are

AGE AND PERIPHERAL NERVE REGENERATION

22 1

small axon profiles containing microtubules, neurofila- and three 15-month control and regenerated buccal ments and small mitochondria, and growth cones recogniz- branches. Measurement of axon caliber. The distribution of axon able by the presence of a variety of organelles, including small membranous vesicles, membranous tubules, and cross-sectional areas in the right and left nerves of the saccules resembling smooth endoplasmic reticulum with young and the older 2-month survivors was determined by atypically electron-dense cisternae, neurofilaments, micro- morphometric analysis of electron micrographs taken at a . areas were meafilaments, and small mitochondria. Examples of axon primary magnification of ~ 2 , 0 0 0 Axon sured with the aid of a graphics tablet and software for the sprouts are shown in Figure 3. Regenerating unit. A group of axon sprouts and Apple IIe (Curcio and Sloan, '861, calibrated to exact Schwann cell processes, circumscribed by a single basal magnification on an image of the electron microscope lamina; the group is believed to be derived from a single diffraction grating. For each of the injured nerves between transected myelinated axon (Morris et al., '72). Immature 240 and 320 axons were measured, and among the control regenerating units are illustrated in figure 3, while more- nerves of each of these five animals 160 to 200 axon areas were measured. mature regenerating units are illustrated in Figure 9. Myelin thickness. The thickness of the myelin was Schwann tube. The space and structures enclosed by the persisting basal lamina of the Schwann cells distal t o measured on the same material as axon areas, with the the site of injury; the basal lamina remains intact following same instrumentation and calibration described above. Numerical density of microtubules: Two months post axotomy and serves as scaffolding directing the path of regrowing axons. crush. Microtubules were counted on images of well-fixed Band of Bungner. The long chain of cells and degenerat- axons sectioned precisely perpendicular to the axon's long ing myelin in the distal nerve produced by hypertrophy and axis, so that axon area would not be exaggerated by a proliferation of Schwann cells within the Schwann tube slightly oblique plane of section. Photographic negatives of (see Cajal, '28; also Bardosi, '89; Roytta et al., '87); bands of axons from control and regenerated nerves taken at a Bungner include both the regenerating unit and the react- magnification of X4,500 were examined at a further magniing Schwann cells to which axons have not yet grown. The fication of x 13 in an Aus Jena Documentor which projects term fiber refers to an axon ensheathed in myelin; the term an image with no distortion at a final magnification of nerue refers to the named fascicle, the buccal branch. x 58,500.For each axon selected, microtubules were counted and the axon area traced. The axon area was subsequently determined with the aid of the computerized system deQuantitative methods scribed above, and the numerical density of microtubules Analysis of regenerating units. An analysis of the calculated. The axons thus examined had areas measuring number of axon sprouts in the regenerating units was less than 20 km2 (diameter about 5 p,m) and so among the carried out on material from the 4 through 28 dpc survivors animals the samples constitute a relatively uniform populain both age groups of animals. In electron micrographs tion encompassing most of the range of axon diameters. printed at magnifications ranging from x 7,500 to x 16,800, Ratio of microtubules to neurofilaments. The numbers regenerating units were categorized as having zero, 1, 2, 3, of microtubules and neurofilaments were quantified in 4, or 5-or-more axon sprouts. Only units that were comaxon profiles directly from the ~ 4 , 5 0 negatives, 0 magnified pletely within the micrographs were counted; the number x 13, as described above. evaluated per animal varied from 32 to 101, with the total Statistical analysis. Analysis of the total numbers of number examined per sample no less than 200. Theoretiaxon sprouts per 100 regenerating units calculated for each cally, a regenerating unit includes a t least one axon sprout, animal required square root transformations of the numbut for the purpose of this study profiles of reacting bers prior to subjecting the data to a 2-way ANOVA to Schwann cells to which axons had not yet grown were determine the significance of the effects of animal age, the included in the counts. Thus all cellular profiles enveloped effects of the survival period, and the interaction of age and by a basal lamina were categorized and counted, and then survival period (Zar, '84). For the quantified data describfor each animal the percent of regenerating units, categoing axon area and myelin thickness in the 15-month-old rized by number of axon sprouts, was calculated. This study and others (e.g., Horch and Lisney, '81; Haftek and Tho- 2-month survivors, the numbers were log transformed mas, '68) indicate that after a crush injury all the original (because of lack of normalcy; Zar, '84) and paired t tests were applied; as there were only two 3-month animals a t axons initially give rise to regenerating axon sprouts that test could not be used. Graphic representation of these data grow within the basal lamina of the original fiber. is provided for the inspection and interpretation of these Total numbers of axon sprouts. After the percent distriresults. Since the numerical density of microtubules varies bution of regenerating units was determined as described with axon caliber, simple regression analyses were perabove, the total number of axon sprouts per 100 regenerating units in each sample was calculated for the purpose of formed on these data and interpretation based on the 95% confidence intervals of the regression lines. comparing the rates of growth in the young and old animals. Total number of axons in the buccal branch: Two months RESULTS post crush. In 2 km thick, toluidine blue-stained plastic Functional recovery sections, profiles of myelinated fibers were traced with a camera lucida attachment to a light microscope. All myeliAmong the animals of both age groups in this study, the nated fibers in the motor nerve fascicles were traced with a crushed facial nerves regenerate and the animals within 100 x oil immersion objective, and the profiles subsequently both age groups uniformly resume bilaterally symmetrical counted. Nine control (right) side buccal branches, includ- whisking. The recovery requires more time in the older ing those from animals sacrificed earlier in the study, were animals: the first signs of voluntary control of whisking analyzed as well as the two 3-month regenerated (left side) activity occur in the 3-month animals at 12 days post crush

222

D.W. VAUGHAN

(dpc), while the 15-month animals begin recovery 4 days later. The period between first signs of whisking movement and attaining full range of whisker swing is also slightly longer in the older animals: 3 days for the 3-month animals, compared to 4 or 5 days for the 15-month animals. Although a simple behavior like whisking can appear restored before bhe nerve is fully reconstituted at the target muscle (Duncan and Jarvis, ’43; Gutmann and Sanders, ’431, for this study we consider this point of “full” recovery an unambiguous and equivalent stage in both animal age groups, and a valid demonstration of the delayed recovery occurring in the older animals.

Morphological observations: Temporal patterns The following section describes the chronology of events and the age-related differences that occur during degeneration and regeneration at a locus 20 mm from the site of crush. One day post crush. Twenty-four hours after the axons are crushed in the 3-month animals the myelin ultrastructure is not disrupted and axon swelling is rare. Nearly 50% of the axon profiles still contain recognizable microtubules and neurofilaments, but among the remaining ones there are no recognizable cytoskeletal elements and the axons contain a dispersed amorphous flocculent material. In the 15-month animals, 24 hours after crush there is no morphological disruption of the axons or their myelin. Two days post crush. By 2 dpc the crushed axons of the 3-month animals are all deteriorating. As can be seen in Figure Ib, some fibers are enlarged and swollen, and others have collapsed, producing irregular profiles. Occasionally, recognizable but disrupted cytoskeletal material exists in the axoplasm, but most axons appear empty or they are filled with amorphous flocculent material which is often coagulated into granulated denser masses beneath the axolemma, like that shown in Figure 2. Mitotic figures among the Schwann cells indicate the bands of Bungner are forming. A few simple rounded Schwann cell profiles containing no myelin, like those illustrated in Figure 2b, are encountered. No new axon sprouts are evident. In the 15-month animals the crushed axons exhibit the same morphological features as those of the younger animals: a few axons appear empty, others are collapsed flat, and some contain the amorphous flocculent material. As with the younger animals, scattered among the many degenerating profiles are a few simple rounded profiles of Schwann cells containing no myelin debris. Four days post crush. At 4 dpc in the 3-month animals there is great cellular activity in the buccal branch. Most of the crushed axons are collapsed and the myelin is dense and compressed, although amorphous material is still evident within some of the axons (Fig. 2a). The Schwann cells appear to be solely responsible for myelin resorption: the internalized myelin debris is condensing and breaking up into dense membranous fragments. No sign of myelin appears within about 10%of the total Schwann cell profiles; these cells represent portions of the Bungner bands that are either the products of mitotic division or hypertrophied regions from which the myelin has retracted. Schwann cells (both those with myelin debris and those without) are transforming into irregular-shaped cells causing the appearance of multiple cytoplasmic profiles within the encircling basal lamina.

At this stage about 50% of the total regenerating units observed in the fascicle bear one to four small pale profiles (0.5 to 2 pm) on their surface within the basal lamina like those illustrated in Figure 2a. These pale profiles are clearly axon sprouts. The smaller ones usually contain microtubules and small mitochondria while the larger ones, considerably more prevalent at 4 dpc than the smaller ones, contain a variety of organelles including small membranous vesicles, a maze of tubules and membranous sacks, neuroflaments, microfilaments, and small mitochondria (Figs. :!a and 3). In the 15-month animals, the situation is different. There is widespread evidence that the axon fibers have not degenerated to the extent that they have in the younger animals, and there is a virtual absence of the small pale axon sprouts growing along the outside of the Schwann cells. Only 1 2 8 of the bands of Bungner profiles have a:n axon sprout in the older animals, although small open gutters appear along the surfaces of some Schwann cells, like those illustrated in figure 2b, as if providing a path for the growing axons. Within the nerve fascicle there still remain swollen or collapsed axon profiles, either empty 0.r with amorphous contents. Ten days post crush. In the %month animal, 10 days after axotomy, relatively little recognizable myelin debris remains within the Schwann cells. The most common profile within the nerve fascicle at this stage is the regener-. ating unit, consisting of axonal and Schwann cell processes, commonly 4 to 20 separate profiles, enveloped in a single basal lamina (Figs. 3 and 4) that frequently appears to be larger than necessary with excess material forming evaginated folds and ruffles as shown in Figure 3b. The Schwann cell components of the unit are usually more electron dense than the axon sprouts and often irregular in shape (Fig. 3). Although the Schwann cell cytoplasm commonly has a uniform, medium electron density, regional changes in the cytoplasmic composition creates small areas of pale cytoplasm with regularly spaced microtubules which, as noted also by Friede and Birschhausen (’801, may confound the identification of axonal sprouts (see area indicated with an asterisk in Figure 2a). However, characteristically among the inclusions within the cytoplasm of the Schwann cell are cisternae of rough endoplasmic reticulum, and pale lipid droplets and smaller homogeneous dense bodies, which are remnants of the cells’ digestive efforts (Fig. 4a). It is not uncommon for the regenerating unit to include in section the Schwann cell nucleus, and frequently profiles of more than one perikaryon are located a t a given transverse level. The frequency with which these nuclei appear in section reflect the fact that the newly formed Schwann cells are considerably shorter and more numerous than those in the control and long-term regenerated nerves. At 10 dpc only 2% of the Schwann cell profiles in the fascicle have no axon sprouts associated with them (Fig. 5). In 45%of the regenerating units there is a single large axon profile that is 1to 2 pm in diameter and recognizable by its content of transversely sectioned microtubules, neurofilaments, small mitochondria, and occasional small membranous cisternae. Growth cones are also part of many of the regenerating units, especially those already accommodating one or two axon sprouts (Fig. 3). Evidence of dystrophic change is occasionally found among the axon sprouts. Their morphology is varied: most commonly, several small electron dense lamellar structures like those illustrated in Figure 8 appear within the axo-

Fig. 2. Four days post crush. a, 3-month; b, 15-month. Degenerating myelin and axoplasm (A)dominate the fascicle at both ages. S marks some of the Schwann cells. In a, new axon sprouts (arrowheads) accompany the Schwann cells that contain degenerating myelin as well as the Schwann cells that are free of debris. The asterisk marks a region

of Schwann cell cytoplasm whose pale appearance can be confused with axoplasm. In b, asterisks mark open gutters along the surface of a Schwann cell. The macrophage (M) has no basal lamina, nor do the attenuated fibroblast processes (0.x9,500. Note that the same magnification is used in Figures 2 , 4 , 6 , and 9.

224

D.W. VAUGHAN

plasm; rarely, axon swelling or shrinkage is evident. The amorphous cytoskeletal deterioration that occurs imrnediately after the crush is not observed. The size of the dystrophic axons is variable: they not only may be among the smallest axons, but also may be the dominant axon in a regenerating unit. The dystrophic profiles are usually Burrounded by Schwann cells that contain axons of nor,mal morphology and so it is unlikely they are artifacts of fixation. Whereas the buccal branch in the younger animals is characterized by signs of growth and regeneration at 10 dpc, the nerves of the 15-month animals are still undergoing extensive degeneration. Profiles of recognizable degenerating axons, some collapsed, in which the myelin has not begun to condense and fragment, still persist. Regenerating units are present, but they tend to be smaller, composed of fewer processes, than those of the younger animals and there is usually no “dominant” axon larger than the others. Rather the axonal processes are all small or are only growing tips (Fig. 4b). As reflected in the charts of Figure 5, about 20% of the regenerating units at 10 dpc consist entirely of Schwann cell processes, with no apparent axons. Sixteen days post crush. In the 3-month animals, all regenerating units have axon sprouts. As illustrated in Figure 5, only 34% of the regenerating units have just one axon, while the balance have two or more axons. The dominant axon is 2 to 3 km in diameter and, when present, additional profiles are either smaller (0.5 bm diameter) than the dominant one, or they are growth cones. The Schwann cells at this stage have begun to form myelin sheaths and synthesize new basal laminae (Fig. 6a). In one half of the regenerating units the dominant axon is wrapped with a thin myelin sheath. As is apparent in Figure 6a, the axons that are myelinated are among the largest ones; no regenerating unit at this stage supports more than one myelinated axon, and not all of the largest axons are myelinated. The remnants of the debris of degeneration contained within the cytoplasm of Schwann cells is either in the form of lipid spheroids or electron dense bodies. While Schwann cells bearing the multiple electron-lucent lipid droplets can provide myelin for axons, those with dense myelin debris usually do not form myelin for the axons they enclose. At 16 dpc most of the regenerating units are surrounded by a single continuous basal lamina. However, among some of those that appear destined to separate into individual fibers, a basal lamina is forming at the interface between the internal apposing membranes. Some of the regenerating units have two separate new basal laminae, one for each cell surface (Fig. 6a), while some have only a single shared lamina that may or may not extend the full length of the interface. Double circumferential basal laminae are not yet observed. Debris of degeneration still dominate the buccal branch of the older animals (Fig. 6b). A substantial number of swollen axons with flocculent contents remain. Some myelin remnants are collapsed and retain their lamination, and do not appear to have begun to deteriorate. All these Fig. 3. Regenerating units, 10 days post crush, composed of numer- debris-congestedSchwann cells are found within regeneratous Schwann cell processes and one or more axon sprouts, enveloped by ing units that contain one or more axon sprouts. a basal lamina. a: A small axon sprout containing microtubules and a All regenerating units in the 15-month animals contain small mitochondria is indicated with a single arrow. The larger profile axon sprouts at 16 dpc. In general the units are smaller, filled with vesicles (double arrows) is a growth cone. b: The axon sprout in the center of the regenerating unit contains many filaments (encir- comprised of fewer cell processes than those of the younger cled) among the few microtubules (smallarrow) and the small mitochon- animals, and perhaps as a reflection of their small size they continue to exhibit excess evaginated folds of basal lamina dria; a membrane-filled bud appears at the arrowhead. ~ 2 1 , 0 0 0 .

Fig. 4. Ten days post crush. a, 3-month; b, 15-month. Regenerating units (at arrows in both figures) in the 15-month animal appear less complex than those of the 3-month animals. Endoneurial cells include the fibroblast with its extensive rough endoplasmic reticulum (F), and macrophages with pale and electron dense debris (MI. At the arrowhead in a,a macrophage has penetrated the basal lamina of the regenerating unit. ~ 9 , 5 0 0 .

D.W. VAUCHAN

226 3 months, 10 dpc

15 months, 10 dpc

3 months, 16 dpc

15 months, 16 dpc

3 months, 21 dpc

15 months, 21 dpc

3 months, 28 dpc

15 months, 28 dpc

Fig. 5. Graphic representation of the percent distribution of regenerating units, by number of axon sprouts, within the nerve fascicles of animal groups as labeled. The shading code used in all the pie charts is indicated in the 3-month, 10 and 21 days post crush (dpc) and the 15-month, 10 dpc.

like those shown in Figure 6b. In 52% of the regenerating units there is only one axon profile; in fact, the percent distribution of the regenerating units is very similar to that of the 3-month animals at 10 dpc. Fewer than 1%of all the regenerating units have a myelinated axon and these axons are usually only very thinly wrapped with a few loose lamellae. Basal laminae have not yet formed at the cellular interfaces within the regenerating units. Twenty-one days post crush. In the 3-month old buccal branch over 90% of the regenerating units have at least one axon that is myelinated; 5% have two myelinated axons. The axons that were once part of single, compact regenerating units at earlier stages are beginning to separate from each other and become independent, so that they are now individually and completely bound by a basal lamina. Despite the dispersion of the maturing fibers, the original units form defined groups that are surrounded at least in part by the original basal lamina, like that shown in Figure 9a. Almost 20% of these maturing regenerating units have five or more smaller axons as part of the group, with nearly half of those units containing 8 to 14 smaller axons. Pale lipid inclusions have become less frequent in the Schwann cell cytoplasm, but they now commonly appear within the cytoplasm of endoneurial cells (Fig. 7a). Both myelinated and nonmyelinated fibers can appear dystrophic, and although smaller axons are more commonly among the dystrophic profiles, they are not exclusively the ones found to bear the small dense membranous inclusions like those in Figure 8. Occasionally, several members of a regenerating unit appear to be dystrophic. In the 15-month animals at 21 dpc only one out of every three of the regenerating units has an axon that is myelinated, in contrast to the nine out of ten observed in the younger animals at 21 dpc. In another significant contrast with the young, as illustrated in Figure 5, it is seen that the percent distribution of regenerating units is very similar to that of the 3-month animals at 16 dpc. Comparisons between Figure 7a,b show that the older nerve fascicle

retains more myelin debris than that of its younger counterpart. Debris of degeneration is still found within cells ofthe regenerating units, although most remnants of degeneration now appear within cells of the endoneurium. Twenty-eight days post crush. In the 3-month animals, fewer than 2% of the regenerating units have axons with no myelin; some original regenerating units surrounded in part by an original basal lamina have as many as four myelinated axons, as shown in Figure 9a. Collagen fibrils, synthesized by the Schwann cells (e.g., Thomas '641, appear for the first time among the separating components of the regenerating units as well as between the inner and outer layers of circumferential basal laminae. Lipid inclusions remain as evidence of degeneration, but they are now contained principally within the irregular processes of the endoneurial cells. The regenerated buccal branch of the 15-month animal is morphologically similar to that of the 3-month animals except the regenerating units are individually smaller and simpler, and as a population appear more uniform in caliber (Figs. 7d, 9b). About 97% of the regenerating units have myelinated axons. As illustrated in Figure 5, only 17% of the regenerating units have more than three axon sprouts; fewer than are present in the 3-month animals at either 21 or 28 dpc. Encircling basal laminae persist, and collagen fibrils synthesized by the Schwann cells appear within their boundaries (Fig. 9b). Two monthspost crush. The regenerated buccal branch in the 3-month animal differs in appearance from that of the 3-month control nerve in two prominent features: the presence of many small diameter axons (see below) and the presence of attenuated cellular processes among and often surrounding the regenerated axons. In the electron microscope, the attenuated cellular processes are found to be of two kinds. The first category includes endoneurial cell processes that contain occasional clumps of pale lipid inclusions and extend radially from the cell's perikaryon, stretching for several micrometers among the fibers. The

Fig. 6. Sixteen days post crush a: 3-month; b: 15-month. In the 3-month animals, axons are becoming myelinated, and components of the regenerating units are beginning to disperse. Basal laminae (at arrows) begin to form within the regenerating units. Endoneurial cells included fibroblasts with attenuated processes (Dinsinuated among the

regenerating units, and debris-laden macrophages (MIAbsorption of degenerating myelin and axoplasm (A) by the Schwann cells is delayed in the 15-month old relative to the 3-month old. A n asterisk marks a regenerating unit with excess basal lamina. X9.500.

Fig. 7. Light micrographs of buccal branch from 3-month animals (a,c,e)and 15-month animals (b,d,f) at 21 dpc (a,b), 28 dpc (c,d) and 2 months post crush (e,D. Retention of debris of degeneration (white d) characterizes the nerve of the older animals at later stages of regeneration. x 730.

AGE AND PERIPHERAL NERVE REGENERATION

229

Endoneurial cells and endoneurium. The endoneurium and the resident cells therein display no striking age-related differences, other than the apparent increased volume of extracellular material in the older regenerated nerves. Initially, at 1 and 2 dpc there is no indication that either the fibroblasts or the macrophages have become stimulated by the lesion, although by 4 dpc proliferation by mitotic cell division is apparent in animals of both age groups. By 10 dpc the fibroblasts appear to be very active, with extensive lengths of rough endoplasmic reticulum, often with expanded cisternae (Fig. 4a); the amorphous ground substance and fibrous matrix begins to accumulate among the developing fibers, especially in the older animals. Beyond 10 dpc the endoneurial macrophages come to contain more debris, initially as electron lucent lipid droplets, but later as dense and heterogenous vacuoles (Figs. 6b, 9a,b). As noted above, by 2 months post crush increased numbers of endoneurial cells persist in the nerves of both the young and the older animals.

Quantitative observations

Fig. 8. Dystrophic axons with dense membranous inclusions in 3-month, 21 dpc nerve. The encircling basal lamina (arrow), lying beyond the basal laminae of the Schwann cells (arrowheads), indicate the two fibers in the figure are derived from one unit. x 13,000.

second kind of attenuated processes apparently arise from the original regenerating units. These partially encircle one or more myelinated axons and have a basal lamina on both their convex and concave surfaces. Where the encircling cell process stops, the basal lamina of its outer surface often continues around the regenerated axon or group of axons. Some of these cells may be derived from intratubal macrophages that, as noted by Friede and Bischhausen ('801, differ from Schwann cells in their behavior toward axons. Others may be Schwann cells, for even 2 months after crush some envelop small pale axon profiles within their crescentshaped processes in the manner shown in Figure 9b. Two months after crush a small number of the myelinated axons in the fascicle appear dystrophic. Like those observed earlier, they contain multiple dense membranous inclusions dispersed within the axoplasm. Both large axons and smaller ones are found to be dystrophic. As noted above, the image of this change is not like that induced by the original crush injury. The 15-month regenerated nerves at 2 months post crush differ from the younger regenerated ones described above in several aspects: by the absence of the distinct subpopulation of small caliber axons, by an increased dispersion of the fibers with intervening extracellular matrix, and by persisting remnants of degeneration (Figs. 7e,f). The endoneurial cells, their arrangements, and their fine structure appear similar to those in the younger animals. Most of the myelinated axons in the 15-month animals do not share the original outer circumferential basal lamina with additional fibers inferring that most regenerating units have given rise to only one axon (see Fig. 9b). Degenerating axons with small dense membranous inclusions are present in these older animals with about the same frequency as in the young.

Growth of axon sprouts. Figure 10 illustrates the total number of axon sprouts per 100 regenerating units in animals between 4 and 28 dpc and it provides insight into the dynamic process occurring in the regenerating nerve. The graph suggests that the rates of growth for the young and the old are comparable, although the older animals are temporally delayed by a period of about 5 days. At the respective times when animals of both age groups recover bilaterally symmetrical whisking behavior (indicated on the graph with arrows), the number of axon sprouts in the regenerating nerves is similar. However, while the number of axon sprouts in the younger animals continues to increase for at least 2 weeks after whisking resumes, the total number of sprouts for the older animals does not increase after they resume whisking behavior. The ANOVA of the square-root transformed data indicates a highly significant effect of animal age ( P = ,0001) and a highly significant effect of survival time ( P = .0001). The P-value for the interaction of age and survival time (P = .0319) suggests that these two factors are each dependent on the other for the difference observed. Thus, with reference to the graph, the effect of animal age is reflected in the separation of the two curves, the effect of survival time is reflected in the slopes of the curves, and the interaction of these two effects is reflected in the deviation of the curves at the end. Total number of axons in the buccal branch. There are about 1,600 myelinated axons in the buccal branch at the locus sampled in this study. Figure 11 shows the mean numbers of fibers for the populations evaluated. The mean of the two control 3-month animals (actually, 5 months of age) is presented separate from that of the larger population of control animal means because it has been found that within-animal variation in total numbers of fibers in the buccal branch tends to be of smaller magnitude than between-animal variation (Mattox and Felix, '87). Only the younger animals have more fibers present in the buccal branch after regeneration: the two 3-month animals have 31% and 32% more fibers in their regenerated buccal branches than their respective control nerves, while the three 15-month animals have 20%, 7%, and 5% fewer fibers than their respective control nerve. The paired t test of the three 15-month animals indicates the difference between

230

D.W. VAUGHAN

Figure 9

AGE ANDPERIPHERALNERVEREGENERATION 400,

231

growth of axon sprouts

........................................................................................................

.........................................................................................................

I

Y 350 M

300

$

250

......................................................................... ...................................................................

.......................................................

...............................................

8 150 ..............

.......

2

............

6

10

I8

14

22

26

30

34

42

38

axon area (micrometers squared)

............. .- -. 15 m o n animals ~ 3...m.Or?th... 84 right

30

0

............................................................................

days post crush Fig. 10. Number of axon sprouts per 100 regenerating units during the 28 days following crush, in the 3-month and 15-month animals. Numbers are animal means 2S.E.M. Arrowheads mark the points at which the animals of the two age groups resume bilaterally symmetrical whisking.

...................................................................................

.................................................................... ................................................ 2

6

14

10

18

22

26

30

34

axon area (micrometers squared)

38

42

.............................................................................................................. 15 month ...............................................................................................

total numbers of fibers in buccal branch 2500

I

T

160 right

.................................................... ........................................................

160 left

......................................

axon area (micrometers squared)

301........... B....................................................................

............................................................. .............................................................

1s...m!mtnth c] 256 rigfit 256 left

........................................................ ........................................................

.-.--

J I T(

3month

15 month

Fig. 11. Histogram illustrating the mean total numbers (and S.E.M.) of fibers in the buccal branch of 3-month control nerves (N = 71, the 3-month control and regenerated nerves (N = Z), and the 15-month control and regenerated nerves (N = 3).

' '26' '30' '34 '38 axon area (micrometers squared)

L

30 .................................................................................

ls...ms!t!!

0

..... ..................................................................

1......................................................... Fig. 9. Twenty-eight days post crush. a: 3-month; b: 15-month.The common origin of the axons derived from regenerating units in the 3-month animals is evidenced by the persisting groups and/or the remnants of the original encircling basal lamina (arrows in both figures). Regenerating units of the 15-month animals, in contrast to those of the 3-month animals, contain fewer axons. Regenerating units in which no axon is myelinated (asterisk) are more common in the older animals than in the younger ones. Endoneurial macrophages (MI in b contain debris of degeneration. x9,500.

10

14

18

'

........

257right 237 left

n

..n..l k.1 I.-

42'

-

......................................................... 22

26

3 0 3 4

axon area (micrometers squared)

Fig. 12. Frequency histograms illustrating the distribution of axon cross-sectional areas in two 3-month and three 15-month animals, 2 month post crush. The right side (open bar) is the control side and the left is the regenerated side (solid bar); the numbers in the key refer to the animal.

D.W. VAUGHAN

232

the number of axons in the control and regenerated nerves is not significant ( P = .08). Axon calibers. The frequency histograms in Figure 12 show that advancing age, at least through 1 7 months of age (15 months plus 2 months survival), has little effect on the distribution and range of the axon cross-sectional areas. The buccal branch consists of a unimodal population of relatively small caliber axons whose mean cross-sectional area (2S.E.M.) is 12.8 pm2 ( 5 0 . 8 ) and 12.8 pm2 (20.5) for the young and older animals, respectively. The histograms from the two 3-month regenerated buccal branches (left side) graphically illustrate that more than 30% of these axons have areas of less than 3 pm2. In the three older animals the regenerated axons (left side) show more normal distributions than those of the younger animals, with considerably fewer very small axons. Furthermore, the range of sizes among the older regenerated axons is less than that of the younger regenerated axons. The mean area (5S.E.M.) for the entire population of regenerated young axons is 5.5 (2.04) pm2 and that of the regenerated older axons is about 5.2 (k.28) pm2. The paired t test of the log-transformed data indicate the difference between control and regenerated axon areas in the older animals is significant ( P = .010). Myelin thickness. Frequency histograms of myelin thickness, Figure 13, show there is no apparent affect of advancing age alone, with means (tS.E.M.) of 0.98 (2.03) pm and 1.09 (+.05) pm, respectively. Among the old and the young animals the thickness of the myelin associated with the regenerated axons is practically indistinguishable in range. This fact is surprising in view of the differing distributions of axon calibers in the young and older animals. The consequence is that in the young animals, most of the small axons have unusually thick myelin sheaths. The difference in myelin thickness between 15month control and regenerated axons is highly significant ( P = ,002). Numerical density of microtubules. The numerical density of microtubules in axoplasm is related to the axon diameter (e.g., Panesse et al., '84; Hoffman et al., '84). The scattergrams of Figure 14 demonstrate that there is no age difference in microtubule density between the two age groups of control axons. There is complete overlap of the intervals defined by the upper and lower 95% confidence limits of means (X,Y) and slopes of the two regression lines. For the regenerated fibers, the increased density of microtubules among the smallest axons is reflected by the increased slope of the regression lines. The older regenerated axons have by comparison even greater microtubule densities and this is reflected in the increased Y-intercept of the regression line. The regression analysis for the regenerated axons shows that while there is a small overlap in 95% confidence interval of X,Y means between the young regenerated and control axons, there is no overlap in the 95% confidence interval of the means between the 15-month regenerated and any of the other three groups. With regard to the slopes of the regression lines, there is a very small overlap in the 95% confidence intervals of the slope of the young regenerated and the control slopes. The 95% confidence interval for the slope of the 15-month regenerated axons, which overlaps considerably with that of the young regenerated axons, does not overlap with those of the controls. If the difference in microtubule density attributable to diameter is discounted by comparing only those axons

........................................................................................................

301::::::.

3 month

....................................................................

0 83 riglit

.........................

...........................................

0.08 0.24

0.4 0.56 0.72 018R 1.04

1.2 1.36 1.52 1.68 1.84

micrometers

-

30 .........................................................................................................

-

I

3 month 0 HI right

201............................................................................... ............

...................

rn

84 k f t

..................................

.......... 0.08 0 2 4

-

0.4 0.56 0.72 0:88 1.04

1 2 1.36 1.52 1.68 1.84

micrometers

.............................

15. mon,th ............................................................

20 .............................

..................................

0 I60 rip&

.............................

..................................

160 left

a 10 .........................

..................................

.......................

....................

-P-P-+,

0 0.08 0.24

0.4 0.56 0.72 0.88 1.04

1.2 1.36 1.52 1.68 1.84

micrometers

15 month

......................... .................................

0.08 0.24

0.4 0.56 0.72 0.88 1.04

1.2 1.36 1.52 1.68 1.84

micrometers

I

15 month

0 25:

rigiif

...................................... ............................

0.88 1.04

1.2 1.36 1.52 1.68 1.84

micrometers

Fig. 13. Frequency histograms illustrating the distribution of myelin thickness in the same animals as in Figure 12. Right side has the control fibers and left has the regenerate fibers.

AGE AND PERIPHERAL NERVE REGENERATION .................................................................................................

3 month control ................................................................................................

................................................................................................ 0

............................................................................................

..................0.............................................................................

0

0

0 0