Retrograde response to axotomy of motoneurons in the thoracic spinal cord of the aging cat

Neurobiology of Aging, Vol. 12, pp. 151-160. ~ Pergamon Press plc, 1991. Printed in the U.S.A.

0197-4580/91 $3.00 + .00

Retrograde Response to Axotomy of Motoneurons in the Thoracic Spinal Cord of the Aging Cat I. P. JOHNSON, T. A. S E A R S A N D A. S. H U N T E R Sobell Department o f Neurophysiology, Institute o f Neurology, Queen Square, London W C 1 N 3BG, UK R e c e i v e d 1 A u g u s t 1989; A c c e p t e d 4 S e p t e m b e r 1990

JOHNSON, I. P., T. A. SEARS AND A. S. HUNTER. Retrograde response to axotomy of motoneurons in the thoracic spinal cord of the aging cat. NEUROBIOL AGING 12(2) 151-160, 1991 .--The retrograde response of HRP-labelled intercostal motoneurons was compared in adult (1-2 years) and aging (10-15 years) cats, 64-68 days following crush of intercostal nerves or following nerve transection with proximal ligation. A comparison of the neuroglial response to these two lesions was also made. In both adult and aging cats, 64-68 days following nerve crush, most labelled motoneurons had a normal appearance. In contrast, 64-68 days following nerve transection and ligation the Nissl bodies of labelled motoneurons lacked the highly ordered ultrastructure characteristic of normal and control motoneurons. No axotomy-induced neuronal loss was found in aging cats. A three-fold increase in numbers of microglial cells was quantified in the ipsilateral ventral horn of aging cats following nerve transection and ligation. This increase was not seen following nerve crush in aging cats, nor following either type of nerve injury in adult cats. Numbers of astroglia and oligodendroglia were unaffected by axotomy in adult and aging animals. Aging

Axotomy

Motoneuron

Degeneration

Microglia

AS life expectancy increases, so does concern over age-related neurodegenerative diseases, such as motoneuron disease, Parkinson's disease and Alzheimer's disease. Central to many hypotheses on the aetiology of such diseases is the premise that aging neurons are more vulnerable to injury compared to those in the adult, due either to intrinsic mechanisms or to the loss of targetdependent trophic signals (4, 35, 58). Surprisingly, in relation to motoneuron disease, this important assumption does not appear to have been tested directly. While it is known that peripheral nerve section in the immediate postnatal period results in marked chromatolysis and massive motoneuronal degeneration (47,51), the situation in the adult animal regarding both neuronal loss and the significance of chromatolysis is less clear. Thus quantitative studies after peripheral nerve section in adult rats and cats have revealed that motoneurons are not lost (11,47) and even after two consecutive nerve lesions, only a small proportion of chromatolytic motoneurons die (48). Nevertheless, despite the weight of current experimental evidence indicating that chromatolysis is part of a regenerative response of neurons to injury (21,34), the older view that it heralds degeneration (23, 41, 65) still appears to persist in some areas of neuropathology and leads to the expectation or assumption that motoneurons in aging animals are more vulnerable to injury. In our ultrastructural studies of chromatolysis (30,53), we observed, contrary to previous studies in the light microscope, that all axotomised motoneurons in the adult cat thoracic spinal cord show an initial disruption of the normally highly ordered rough endoplasmic reticulum of their Nissl bodies. After nerve crush, the subsequent axonal regeneration and reinnervation of the tar-

get resulted in the reformation of highly ordered Nissl body ultrastructure. However, following nerve transection and ligation, preventing target reinnervation, Nissl body ultrastructure remained disorganised. As these target-dependent changes in Nissl body ultrastructure were not consistent with chromatolysis in the light microscope, the reliability of chromatolysis as an indicator of motoneuronal injury is again called into question. Similarly, although chromatolysis has been described postmortem in motoneuron disease (27), its relationship to motoneuronal degeneration is unclear. In this study, we have used the same experimental paradigm of nerve crush versus nerve transection and ligation in 10-15year-old cats to determine firstly if aging affects the target-dependent features of the retrograde response of motoneurons to axotomy and secondly to determine if aging increases the likelihood of motoneuronal death following axotomy. METHOD

Animals A total of 12 cats, including 2 normal adults and four aging cats, were used. The adults were 1-2 years old and the aging cats 10, 10, 15 and 15 years old. All animals were laboratory-bred and reared and in good general health throughout the study. The birthdates of the aging animals were reliably known. Anaesthesia was induced by sodium pentobarbitone (45 mg/kg, IP) prior to all surgery and perfusions.

Axotomy In 2 adult and 3 aging cats, the 7th and 9th internal and exter-

151

152

nal intercostal nerves (52) were exposed unilaterally at the level where the iliocostalis muscle traverses the intercostal spaces and the nerves either 1) crushed (T7) for 10 seconds using fine forceps or 2) transected (T9), the proximal stumps tightly ligated and 3-5 mm of the distal stumps removed; the axotomy sites (20-30 mm from the corresponding ventral horn of the spinal cord) were marked by a loose ligature around the nerves. Animals were perfused 64-68 days later, when, following the initial retrograde response, the target-dependent changes in the Nissl bodies of axotomised motoneurons are fully established (30). To ensure that analysis was confined to axotomised motoneurons, their cell bodies were identified by the presence of horseradish peroxidase (HRP) reaction product in histochemically processed slices of spinal cord following the retrograde axonal transport of HRP. One day prior to perfusion, the axotomy site was located and a region just proximal to this was freed and isolated from surrounding tissue using cotton wool impregnated with Vaseline. A small well was created around the nerves and the cotton wool secured to adjacent tissue with a few drops of 5% Agar. Nerves were newly lesioned and 10-20 ~1 of 40% HRP (Sigma type VI) in saline injected to surround the nerves. In all later experiments, newly transected proximal nerve stumps were drawn into a short upright polyethylene cannula, the base of which was secured to adjacent tissue with 5% Agar. The solution of HRP was then injected into the cannula to surround the nerves; both methods gave comparable results. In addition to the 2 normal adult cats, control data was available from 1) 1 adult cat in which HRP had been applied to the freshly cut proximal stumps of (previously uninjured) intercostal nerves 24 hours prior to perfusion and 2) 4 cats (3 adult and 1 aging) which had received injections of 10-20 Ixl of 40% HRP into either the levator costae and external intercostal muscles, or the internal and external intercostal muscles, in 5 ~1 aliquots using a glass microelectrode of 100 txm tip diameter.

Perfusion and Histochemistry Twenty-four hours after HRP labelling, control and experimental animals were perfused via the abdominal aorta with saline followed by a mixture of 2% glutaraldehyde and 1% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4. Segments T7 and T9 (axotomy animals) or T7-10 (normal and intramuscular HRP animals) of fixed spinal cord were sliced transversely at 70 ~m in buffer and processed to demonstrate peroxidase activity using 3,3'-diaminobenzidine with cobalt enhancement of the reaction product (2). After treatment with 1% osmium tetroxide, slices were dehydrated in an ascending series of ethanol solutions, equilibrated with Araldite via an intermediate stage in propylene oxide and fiat-embedded in Araldite between two polytetrafluorethylene-coated microscope slides (49). Although HRP labelling following retrograde transport was readily identified in the light microscope, the level achieved was insufficient to provide unambiguous evidence of labelling in the electron microscope when stained sections were viewed. Accordingly, as in previous studies from this laboratory (28,29), a two-stage procedure was adopted. Labelled motoneurons were first identified in the Araldite-embedded slices of spinal cord. The ventral horn containing the labelled motoneurons was mounted on an Araldite cylinder and transilluminated from below to facilitate trimming of the ventral horn to a suitably sized " m e s a " for ultrathin sectioning. This was followed by identification of labelled motoneurons in the light and electron microscope in adjacent semithin sections and ultrathin sections stained with toluidine blue or uranyl acetate and lead citrate, respectively. Analysis was restricted to retrogradely labelled motoneurons, except where stated.

JOHNSON, SEARS AND HUNTER

Quantitation of Neuroglia For each group of animals, 10 random toluidine blue-stained 0.5 ixm sections of the ventral horn were taken and as indicated in Fig. 1, five different areas each of 32,500 p,m 2 were photographed for each ventral horn, making a total of 50 prints per experimental group. These areas were printed at a final magnification of × 1230 and the numbers of astrocytes, oligodendrocytes and microglia counted without any prior reference to the experimental code. Neuroglia were classed solely on the basis of their nuclear morphology, as it was difficult in semithin toluidine bluestained sections to resolve their sparse and often lightly stained cytoplasm. In both adult and aging cats, oligodendrocytes had round or oval nuclei which were moderately stained and had a peripheral rim of chromatin; microglia had smaller, fusiform and much more densely stained nuclei with chromatin which predominantly formed a peripheral rim but also had stout projections into the central regions of the nucleus; astrocytes had the largest, most irregularly shaped and weakly stained nuclei (Fig. 2). As a check on the accuracy of the photographic method, the data from 2 normal adult cats was compared with the neuroglial counts obtained from direct observations in the light microscope at a magnification of × 1090 using a rectangular eyepiece graticule which delineated 32,500 Ixm2. This showed that the numbers of oligodendrocytes and microglia counted on the photographic prints were 94.5% and 100.5%, respectively, of the values obtained for these individual neuroglial cell types after direct counts were made using the light microscope on the same areas of the same sections. However, the numbers of astroglia counted in the photographic prints were only 37.5% of the number seen directly in the light microscope. This was attributed to the pale staining of astroglial nuclei causing some to be identified on the photographic prints as transversely sectioned dendrites, or other structures; none were confused with neurons which were much larger than astroglial nuclei and only examined if they contained a nucleus. Although a similar comparison of the photographic and direct counting method would have been desirable for astroglia in both the adult and aging experimental animals, time and resource limitations prevented this from being carried out. Nevertheless, as there was no evidence that astroglial nuclei differed in appearance in any of the experimental conditions, no correction has been made of the astroglial counts made on the photographs as the same error would be expected to apply to both control and experimental groups.

Quantitation of Neurons It was our intention to minimise the number of animals involved in this study by using the same material for the quantitation of neurons and neuroglial cells in the light microscope, as well as for electron microscopical studies. While large sections can be cut serially from tissue embedded in paraffin wax, thus facilitating neuronal counts, the classification of individual neuroglial cell types in such sections is unreliable and this approach precludes any subsequent analysis of the same tissue in the electron microscope. We therefore used the same Araldite-embedded material for all analyses. However, we found it very time-consuming to cut semithin transverse sections of cat thoracic spinal cord to include both the ipsilateral and contralateral ventral horns. For the neuronal counts, therefore, all neurons (retrogradely labelled and unlabelled) containing a nucleolus were counted by two independent investigators in all intact 70 ~m osmicated, Araldite-embedded sections through whole spinal cord segments. As these thick sections contained many neurons, all stained with osmium, this obviated the need to cut a great number of semithin

AXOTOMY OF AGING MOTONEURONS

153

13

FIG. 1. (a) Semithin toluidine blue-stained section of spinal cord at T9 from an adult cat. (b) Outline of (a) showing the area of ventral horn defined for neuronal counts (hatched) and the approximate positions within this area from which neuroglial counts were made. sections, the staining of which would not have been uniform. Moreover, as most of the 70 txm sections were intact transverse sections of the spinal cord, this facilitated comparison of ipsilateral and contralateral sides. For counts of HRP-labelled motoneurons alone (one cat), 56 random sections (T9) were quantified and for the counts of both labelled and unlabelled neurons (three cats), 10 sections (T9) were quantified. To define the area within which counts were made, the ventral horn and part of the intermediolateral horn of all sections was photographed and printed at a final magnification of × 1300. On these prints, a U-shaped clear plastic template with a flat base was overlaid on the ventral horn and the area selected for quantitation defined with a pen (Fig. 1). This area measured 330,000 i~m2 and from previous studies of the cat thoracic spinal cord (17,29) was known to contain HRP-labelled motoneurons. Within these boundaries, the positions of all neurons containing a nucleolus, identified by direct observation of 70 Ixm sections in the light microscope, were marked on the photograph. In this way, the spatial distribution of labelled and unlabelled neurons

in the ventral horn was recorded, as well as the distribution of neurons displaying any other features, such as eccentric nuclei. With the nucleolus in the plane of focus, the mean diameters (Max + Min/2) of these neurons were also determined. Using 70 txm sections and a 5 Ixm diameter nucleolus as the unit counted, the calculated Abercrombie correction factor (1) for the split cell error was 0.9333, which indicated that double counting of neurons was minimal. Almost all the neurons were > 10 Ixm in diameter, all had lightly stained nuclei and were only counted if a prominent nucleolus could be seen; as a result, it was considered highly unlikely that any would be confused with glial cells.

Statistical Analysis An analysis of variance showed that interanimal variability for the features quantified was not significant (p>0.05), so that pooled data was used where appropriate. The significance of differences between sample means was calculated using Student's t-test.

154

JOHNSON, SEARS AND HUNTER

FIG. 2. Light micrograph of the ventral horn at T9 of an adult cat to show the appearance of the nuclei of microglia (m), astroglia (a) and oligodendroglia (o). RESULTS Normal and Control Animals Neuroglia. The appearance of neuroglia in semithin sections has already been described (see the Method section, Fig. 2). In the electron microscope (Fig. 3), neuroglial cells of both adult and aging animals had a similar appearance, although neuroglial cells in aging cats contained more lipofuscin. This is in general agreement with previously published descriptions of neuroglia in adult and aging animals (8, 38, 39, 43, 56, 63, 64). Neurons. Attention was directed in this study towards the large (>40 ~m) and easily identified cell bodies of alpha motoneurons in the ventral horn. The cell bodies of nearby gamma motoneurons and interneurons have already been characterised ultrastructurally by us in adult cats (28, 30, 31) and were not examined in detail here. All the alpha motoneurons examined ultrastructurally contained a nucleus and in the control and experimental animals these motoneurons were also labelled by the retrograde axonal transport of HRP. In both adult and aging cats, normal and control alpha motoneurons were characterised by their large cell bodies and prominent Nissl bodies composed of regular arrays of rough endoplasmic reticulum (RER) between which rows of polyribosomes were interposed. Other aspects of the ultrastructure of alpha motoneurons and their synaptic terminals in the thoracic region of adult cats have been described previously and therefore not repeated here (29, 45, 46). Motoneurons in aging cats differed from those in adult cats only by their occasional greater content of lipofuscin (Fig. 4a and b). Axotomised Animals Neuroglia. In adult cats 64 days following both nerve transec-

tion and ligation and following nerve crush, microglia and oligodendroglia had a normal ultrastructure and no change from control values in the numerical density of the nuclei of any neuroglial cell type was quantified (Fig. 6). In adult cats, 64 days following nerve transection and ligation, but not 64 days following nerve crush, large stretches of the motoneuronal plasma membrane were occupied by overlapping, thin astroglial processes. These differences in astroglial cover, according to the type of nerve injury, were paralleled by corresponding changes in the percentage length of HRP-labelled alpha motoneuronal plasma membrane occupied by the presynaptic membrane of synaptic terminals (synaptic cover). After falling to 54% (crush) and 35% (transection) at 8-33 days, synaptic cover was restored to 68% of control values 64 days following nerve crush, but remained at 35%, 64 days following nerve transection with proximal ligation (28). In aging cats, the astroglial reaction was similar to that described for adult animals and no oligodendroglial reaction was seen (Fig. 6). However, following nerve transection and ligation, but not following nerve crush, increased numbers of microglial cells were seen in the ipsilateral ventral horn. In the electron microscope these microglia contained lipofuscin-like material to a much greater extent than the microglia of control aging cats (compare Figs. 3 and 5). Counts of neuroglial cells in the light micrographs (Fig. 6) revealed a statistically significant (p<0.001) increase in numbers of microglia 64-68 days following nerve transection and ligation in aging cats, with no significant change in the numbers of astroglia or oligodendroglia. Neurons. Aging motoneurons showed the same target dependent changes in Nissl body ultrastructure as previously described for adult motoneurons (30). Thus, in normal motoneurons, control motoneurons and those 6 4 ~ 8 days following nerve crush, Nissl bodies were composed of several RER lamellae, between which linear arrays of polyribosomes were interposed (Fig. 4c and d). However, in motoneurons 64-68 days following nerve section with ligation, which prevented target reinnervation, Nissl bodies were devoid of such orderliness; polyribosomes being associated with only short fragments of RER (Fig. 4e and f). HRPlabelled motoneurons from 4 additional adult cats used in related studies confirmed that these changes in Nissl body ultrastructure persisted for at least 305 days following nerve section with ligation. Because the method adopted allowed the light and electron microscopical appearance of individual Nissl bodies in labelled motoneurons to be compared in adjacent semithin and ultrathin sections, we were able to confirm our previous observations (30,32) that the presence of large Nissl bodies in the light microscope is neither a reliable indicator of their degree of ultrastructural orderliness nor of the success of axonal regeneration. In one 15-year-old cat, 68 days following nerve transection and proximal ligation, motoneurons were retrogradely labelled via the cut proximal stumps of both the ipsilateral and contralateral intercostal nerves 1 day prior to perfusion. In this animal, when the ipsilateral and contralateral ventral horns were compared in 56 intact sections, it was found that 1) the mean numbers of HRP-labelled motoneurons per 100,000 ixm2 of the ipsilateral (1.77±0.11) and contralateral (1.09±0.17) ventral horns were not significantly different (mean ± SEM). 2) The total numbers of neurons (labelled and unlabelled) per 100,000 ixm2 unit area of the ipsilateral (9.57 ± 0.99) and contralateral (9.71 ± 0.46) ventral horns were not significantly different (mean ± SEM); and 3) the mean diameters of labelled neurons in the ipsilateral (35.48 -~ 1.12 ~xm) and contralateral ( 3 2 . 8 0 ± 0 . 9 8 txm) ventral horns (mean ± SEM) were not significantly different. To obtain further evidence that motoneuronal loss had not occurred ipsilateral to the nerve transection and ligation at 6 ~ 6 8 days, data from the two aging cats in which intercostal motoneurons had been retrogradely labelled ipsilaterally 64 days following unilateral nerve

AXOTOMY OF AGING MOTONEURONS

155

FIG. 3. Electron micrographs of an astrocyte (a), oligodendrocyte (c) and microglial cell (e) in the ventral horn (T9) of a control aging cat. Astrocytic cytoplasm (b) is least electron-dense and contains bundles of intermediate filaments (arrow). Oligodendrocyte cytoplasm (d) is most electron-dense and contains short fragments of RER (arrow). Microglial cytoplasm (f) is of intermediate electron-density and contains long cisternae of RER (arrow).

transection and ligation was combined with data from the single aging cat where intercostal motoneurons had been retrogradely labelled bilaterally, 68 days after unilateral nerve transection and ligation. Since this involved pooling data where retrograde labelling had been performed either unilaterally or bilaterally, the

numbers and sizes of all neurons, labelled or not, in ten 70 txm sections containing both the ipsilateral and contralateral ventral horns were determined. For these three aging cats (Fig. 7), no significant difference in either the number of neurons per 100,000 p.m z in the ipsilateral (8.09 ± 0.91) and contralateral (8.16 ± 0.60)

56

JOHNSON, SEARS AND HUNTER

ADULT

AGED

FIG. 4. HRP-labelled control and axotomised motoneurons in adult and aging cats. Compared to control motoneurons in adult cats (a), control motoneurons in aging cats (b) are distinguished by their higher content of lipofuscin (arrow). Sixty-four days following nerve crush, highly ordered Nissl body ultrastructure is seen in both adult (c) and aging (d) motoneurons. Sixty-four days following nerve transection and ligation, disorganised Nissl body ultrastructure is seen in both adult (e) and aging (f) motoneurons.

ventral horn (mean : SEM), or their mean diameters ipsilaterally (33.36 -~ 0.75 i.Lm) or contralaterally (31.43 -+ 0.70 I.tm) was found (mean -+ SEM). Also, the spatial distribution of labelled motoneutons and of neurons noted as having eccentric nuclei was in accordance with the known location of the motor pools of the internal

and external intercostal motoneurons as established using HRP (17) or in this laboratory using retrograde chromatolysis (14). No marked clustering either of labelled motoneurons or motoneurons with eccentric nuclei, or regions of reduced neuronal density were observed when labelled motoneurons and large ( > 4 0 txm) unla-

AXOTOMY OF AGING MOTONEURONS

157

FIG. 5. Electron micrograph of the ipsilateral ventral horn of an aging cat 64 days following nerve transection and ligation showing a microglial cell containing an accumulation of lipofuscin-like material in its cytoplasm.

belled motoneurons with eccentric nuclei were plotted together within a standard outline o f the ventral horn. Increasing the number o f cats in the sample from 1 to 3 caused the ratio o f the number o f neurons in the ipsilateral:contralateral ventral horns to

b aging

a

adult

adult

aging

change only from 0.985 ( n = 1) to 0.992 ( n = 3 ) , but with the present sample size (3 aging cats) it remains possible that the null hypothesis may have been retained falsely as the relative standard error for the data obtained for neuronal numbers from 3 cats is 11% and 7.4% for the ipsilateral and contralateral ventral horns,

~=3o1

a

3

1 0

q

,q

'4

',1 c

im

£" im

t

c

t

c

t

c

2'0

30

4"0

50

60 pm

0

1"0

20

3"0

40

50

60 pm

c

d

mm t

10

0

c

im

0

20 ,o

im

t

c

t

c

FIG. 6. Quantitative data from adult and aging cats for (a) total neuroglia, (b) oligodendroglia, (c) astroglia and (d) microglia, expressed as numbers/10,000 ixm2 ventral horn area. Open columns, ipsilateral ventral horns; hatched columns, contralateral ventral horns. Note the increased numbers of microglia ipsilaterally in aging cats 64-68 days following nerve transection and ligation, im, intramuscular HRP; t, nerve transection and ligation; c, nerve crush. (mean ± SEM.)

diameter FIG. 7. Mean diameters of all neurons seen in 70 l~m slices of the ipsilateral (a) and contralateral (b) ventral horns of 3 aging cats 64-68 days following unilateral intercostal nerve transection and proximal ligation. The histograms have very similar shapes and means (arrow heads). Two hundred and sixty-three neurons measured ipsilaterally and 260 neurons measured contralaterally.

158

JOHNSON, SEARS AND HUNTER

respectively. In order to reduce the relative standard error to 5% for the ipsilateral ventral horn, the sample size here would need to be increased five-fold, which would be unrealistic with regard to the resources available and the timescale of the present study. DISCUSSION

There are three main findings of this study. Firstly, chronic interruption of peripheral target contact for motoneurons in both adult and aging animals results in Nissl bodies which lack the high degree of ultrastructural orderliness characteristic of normal motoneurons (Fig. 4). Secondly, 64-68 days following axotomy in aging cats there is no neuronal loss. Thirdly, in aging cats, the number of microglial cells increased in the ipsilateral ventral horn 6 4 ~ 8 days following nerve transection and ligation, but not following nerve crush (Fig. 6).

Neuroglial Response The retrograde response of motoneurons to peripheral axotomy is frequently accompanied by a reaction of nearby astroglia and microglia (41, 59, 61). Successful axonal regeneration is associated with the return of neuroglial morphology to normal, whereas unsuccessful axonal regeneration is associated with a persisting neuroglial reaction (18, 55, 60). In the present study, the only persisting neuroglial response following axotomy in adult cats was exhibited by astroglia. This involved, in the case of nerve transection and ligation, increased coverage of the motoneuronal plasma membrane at the expense of synaptic terminal coverage. Whether these astroglial changes are due to the retraction of intact synaptic terminals from axotomised motoneurons, the degeneration of synaptic terminals (50) or secondary to microglial activation (20,21) remains to be determined. Although microglia have been shown to be involved in the initial removal of synaptic terminals from the cell bodies of cranial motoneurons in rats and mice (9, 24, 61, 62), no longterm microglial response was detected in this study of adult cat thoracic motoneurons, nor in analogous studies on limb motoneurons (12). In contrast, a persisting microglial response was present at 64-68 days in aging cats, but only following nerve transection and ligation and then without evidence of neuronal loss. Thus the results of the present study of thoracic motoneurons differ from those obtained from studies of axotomised hypoglossal and vagal motoneurons in the adult cat where microglial proliferation was associated with a 15% and 28% loss of motoneurons, respectively, by 80 days (15).

Neuronal Response By 64 days following nerve crush, most labelled motoneurons in both adult and aging cats had a normal appearance, while following nerve transection with proximal ligation, motoneurons of all cats failed to reform Nissl bodies with the highly ordered ultrastructure characteristic of normal and control motoneurons. Previous studies in adult cats (30) have shown that these targetdependent changes in Nissl bodies persist for several months (limit of observations) following axotomy, but that subsequently the ultrastructural orderliness of Nissl bodies can be restored by providing motoneurons with a skeletal muscle target. It is known that axotomy causes changes both in the synthesis and intracellular distribution of proteins within the cell bodies of neurons (22, 34, 68). Thus the balance shifts from the normal one associated with the synthesis of proteins associated with neurotransmission to those associated with axonal growth (54,67). The restoration of the normal ultrastructural appearance of the RER comprising the Nissl bodies indicates that a normal pattern of protein synthesis

has been restored with successful axonal regeneration following nerve crush, but that the balance remains altered in the absence of peripheral target contact when such regeneration is prevented. This interpretation is corroborated by correlations made in nonneuronal cells that RER is associated with the synthesis of proteins destined for export from the cell and free polyribosomes are associated with the synthesis of proteins destined for internal use (42). Thus, in both adult and aging cats, contact with the peripheral target appears to regulate the pattern of protein synthesis in motoneurons. However, as no consequential loss of motoneutons was found following axotomy, peripheral target contact does not exert a major influence on motoneuronal viability in adult and aging cats. No evidence of axotomy-induced motoneuronal loss was found in aging cats as determined either by 1) counts of neurons viewed on 70 txm osmicated sections, or 2) counts of HRP-labelled motoneurons alone. Taken together, the results of these two approaches indicate that in aging cats there is no significant loss of axotomised motoneurons, motoneurons contributing to uninjured nerves or of interneurons. This result is seen in the context of our observations that chromatolytic alterations in axotomised cat thoracic motoneurons are maximal between 8-33 days, so that 6468 days falls well beyond what would be regarded as the period of chromatolytic neuronal degeneration for peripheral neurons. Although this interpretation of our quantitative data is in accord with the qualitative findings in both the light and electron microscope, we cannot discount the possibility that a small loss of neurons has remained undetected because of limits imposed by the sensitivity of the counting method. It is possible that survival beyond 68 days is required to reveal axotomy-induced loss of aging neurons, as may have been the case for the loss of motoneurons reported for two humans aged 54 and 60 who died 4.4 and 9 years, respectively, following lower limb amputation (33). While this is an interesting possibility, its assessment experimentally carries with it the problem of ensuring that the postoperative survival period lies within the natural life span of the animal; the mean life span of the domestic cat being 14 years (10) and old age in this animal considered to begin from 9 years (26). The survival of intercostal motoneurons is unlikely to be due to the persistence of "sustaining" axonal branches proximal to the site of nerve transection and ligation (19,36). The retrograde labelling method used ensured that only axotomised motoneurons were examined and there is no evidence that motor axons in the first filament from the internal intercostal nerve (52) which may have been spared proximal to the site of axotomy can send collateral branches within the main trunk of the intercostal nerve to innervate more distal portions of intercostal muscle. It has been shown (25) that medial gastrocnemius motoneurons in the adult cat twelve weeks after nerve transection and ligation produce supernumerary axons which synapse with other neuronal elements in the ventral horn. It is not known, however, if the additional axonal contacts thus formed help to sustain these limb motoneurons in the absence of peripheral target contact, nor whether such a phenomenon applies to thoracic motoneurons in the aging cat. Within the timescale of the present study, our results indicate that axotomy alone is not a sufficiently severe condition to provoke degeneration of aging motoneurons, particularly thoracic motoneurons. Experiments involving extended postoperative survival times, or other motoneuronal perturbations in addition to axotomy, may shed light on the conditions required, but are outside the scope of the present study. Consideration was given to the possibility that a prior axonal injury may alter the capacity of neurons to take up and retrogradely transport HRP (7,44). In this study, similar numbers of motoneurons in aging cats were retrogradely labelled by the ap-

AXOTOMY OF AGING MOTONEURONS

159

plication o f H R P to the freshly cut stumps o f intercostal nerves either 68 days following nerve transection and ligation, or to the previously intact, contralateral intercostal nerves. Thus axotomy performed up to 68 days previously does not appear to decrease the capacity o f newly lesioned axons to take up and retrogradely transport HRP. The present study has wider implications regarding the assumption that neurons develop an increased vulnerability with age. It has been suggested (4,5) that motoneuronal degeneration in human motoneuron disease reflects the failure o f aging muscle cells to release a neurotrophic hormone which, via retrograde axonal transport, acts on the motoneuronal cell body to sustain it.

It is further suggested that when the supply o f this neurotrophic hormone is reduced, any subsequent perturbation o f the system, such as trauma, toxins, ischaemia or infection, will increase the likelihood o f motoneuronal degeneration. The present study indicates that 6 4 - 6 8 days following axotomy, thoracic motoneurons of aging cats do not show any extra vulnerability to the loss o f their peripheral target.

ACKNOWLEDGEMENT Supported by the Medical Research Council.

REFERENCES 1. Abercrombie, M. Estimation of nuclear population from microtome sections. Anat. Rec. 94:239-247; 1946. 2. Adams, J. C. Technical considerations on the use of horseradish peroxidase as a neuronal marker. Neuroscience 2:141-145; 1977. 3. Aldskogius, H. Glial responses in the adult rabbit dorsal motor vagal nucleus during axon reaction. Neuropathol. Appl. Neurobiol. 8:341349; 1982. 4. Appel, S. H. A unifying hypothesis for the cause of amyotrophic lateral sclerosis, Parkinsonism and Alzheimer Disease. Ann. Neurol. 10:449-505; 1981. 5. Appel, S. H.; Smith, R. G.; Vaca, K.; McManaman, J. Neurotrophic factors and the pathogenesis of amyotrophic lateral sclerosis. In: Serratrice, G.; Cros, D.; Desnuelle, C.; Gastaut, J-L.; Pellisier, J-F.; Pouget, J.; Schiano, A., eds. Neuromuscular diseases. New York: Raven Press; 1984:347-351. 6. Barron, K. D. Ultrastructural changes in dendrites of central neurons during axon reaction. Adv. Neurol. 12:381-399; 1974. 7. Barron, K. D. Retrograde transport of injured neurons. Neurology 34:401-402; 1984. 8. Blakemore, W. F. Reactions of neuroglia and microglia to injury. Ph.D. Thesis. University of London; 1969. 9. Blinzinger, K.; Kreutzberg, G. W. Displacement of synaptic terminals from regenerating motoneurons by microglial ceils. Z. Zellforsch. Mikrosk. Anat. 85:145-157; 1969. 10. Brace, J. J. Theories of aging: An overview. Vet. Clin. North Am. 11:811-814; 1981. 11. Carlson, J.; Lais, A. C.; Dyck, P. J. Axonal atrophy from persistent peripheral axotomy in adult cats. J. Neuropathol. Exp. Neurol. 38: 578-585; 1979. 12. Chen, D. H. Qualitative and quantitative study of synaptic displacement in chromatolysed spinal motoneurons of the cat. J. Comp. Neurol. 177:635-644; 1978. 13. Clarke, P. G. H. Neuronal death in the development of the vertebrate nervous system. Trends Neurosci. 5:345-349; 1985. 14. Coffey, G. L. The distribution of respiratory motoneurones in the thoracic spinal cord of the cat. Ph.D. Thesis. University of London; 1972. 15. Cova, J. L.; Aldskogius, H. Effect of axotomy on perineuronal glial cells in the hypoglossal and dorsal motor vagal nuclei of the cat. Exp. Neurol. 93:662-667; 1986. 16. Cunningham, T. J. Naturally-occurring neuron death and its regulation by developing neural pathways. Int. Rev. Cytol. 74:163-186; 1982. 17. Fedorko, L. Localisation of respiratory motoneurone pools in the cat's spinal cord. J. Physiol. 332:28P; 1982. 18. Fernando, D. An electron microscopic study of the neuroglial reaction in the hypoglossal nuclei after transection of the hypoglossal nerve. Acta Neuropathol. (Berl.)86:1-14; 1973. 19. Fry, F. J.; Cowan, W. M. A study of retrograde cell degeneration in the lateral mammillary nucleus of the cat, with special reference to the role of axonal branching in the preservation of the cell. J. Comp. Neurol. 144:1-24; 1972. 20. Gage, F. H.; Olejniczak, P.; Armstrong, D. M. Astrocytes are important for sprouting in the septohippocampal circuit. Exp. Neurol. 102:2-13; 1988. 21. Giulian, D.; Baker, T. J. Peptides produced by ameboid microglia

regulate astroglial proliferation. J. Cell Biol. 101:2411-2415; 1985. 22. Grafstein, B.; McQuarrie, I. G. Role of the nerve cell body in axonal regeneration. In: Cotman, C., ed. Neuronal plasticity. New York: Raven Press; 1978:155-195. 23. Gudden, B. Experimentaluntersuchungen uber das peripherische und central Nervensystem. Arch. Psychiat. Nervenkrankh. 2:693-723; 1870. 24. Hamberger, A.; Hansson, H-A; Sjrstrand, J. Surface structure of isolated neurons. J. Cell Biol. 47:319-331; 1970. 25. Havton, L.; Kellerth, J-O. Regeneration by supernumerary axons with synaptic terminals in spinal motoneurons of cats. Nature 325:711714; 1987. 26. Henderson, G. N.; Coffey, D. J. Cat and cat care: An international encyclopaedia. London: David and Charles; 1973. 27. Hirano, A. Aspects of the ultrastructure of amyotrophic lateral sclerosis. In: Rowland, L. P., ed. Human motor neuron diseases. New York: Raven Press; 1982:65-82. 28. Johnson, I. P. Morphological correlates of altered protein synthesis: An ultrastructural analysis of axotomy and diphtheritic intoxication. Ph.D. Thesis. University of London; 1983. 29. Johnson, I. P. A quantitative ultrastructural analysis of alpha and gamma motoneurons in the thoracic region of the spinal cord of the adult cat. J. Anat. 147:55-72; 1986. 30. Johnson, I. P.; Pullen, A. H.; Sears, T. A. Target dependence of Nissl body ultrastructure in cat thoracic motoneurones. Neurosci. Lett. 61:201-205; 1985. 31. Johnson, I. P.; Sears, T. A. Ultrastructure of interneurons in the thoracic region of the spinal cord of the adult cat. J. Anat. 161:171185; 1988. 32. Johnson, I. P.; Sears, T. A. Organelle changes in cat thoracic ct- and ~-motoneurons following axotomy. Brain Res. 489:400-405; 1989. 33. Kawamura, Y.; Dyck, P. J. Permanent axotomy by amputation results in loss of motor neurons in man. J. Neuropathol. Exp. Neurol. 40:658-666; 1981. 34. Kreutzberg, G. W. Acute neuronal reaction to injury. In: Nicholls, J. G., ed. Repair and regeneration of the nervous system. Berlin: Springer-Verlag; 1982:57--69. 35. Langston, J. W. Aging, neurotoxins and neurodegenerative disease. Aging 32:149-164; 1988. 36. Lieberman, A. R. Some factors affecting the retrograde responses to axonal lesions. In: Bellairs, R.; Gray, E. G., eds. Essays on the nervous system. Oxford: Clarendon Press; 1974:71-105. 37. Ling, E. A.; Wong, W. C.; Yick, T. Y.; Leong, S. K. Ultrastrucrural changes in the dorsal motor nucleus of the monkey following bilateral cervical vagotomy. J. Neurocytol. 15:1-15; 1986. 38. Machado-Salas, J.; Scheibel, M. E.; Scheibel, A. B. Neuronal changes in the aging mouse: Spinal cord and lower brainstem. Exp. Neurol. 54:504-512; 1977. 39. Mori, S.; Leblond, C. P. Identification of microglia in light and electron microscopy. J. Comp. Neurol. 135:57-80; 1969. 40. Mori, S.; Leblond, C. P. Electron microscopic identification of three classes of oligodendrocyte and a preliminary study of their proliferative activity in the corpus callosum of young rats. J. Comp. Neurol. 139:1-30; 1970. 41. Nissl, F. Uber eine neue Untersuchungmethode des Centralorgans speciell zur Feststellung der Localisation der Nervenzellen. Zentral-

160

blatt fiir Nervenheikunde und Psychiatrie 17:337-354; 1894. 42. Palade, G. Intracellular aspects of the process of protein synthesis. Science 189:347-358; 1975. 43. Peters, A.; Palay, S. L.; Webster, H. DeF. The fine structure of the nervous system: The neurons and supporting cells. Philadelphia: W. B. Saunders Company; 1976. 44. Peyronnard, J. M.; Charron, L.; Lavoie, J.; Messier, J. P. Differences in horseradish peroxidase labelling of sensory, motor and sympathetic neurons following chronic axotomy of the rat sural nerve. Brain Res. 364:137-150; 1986. 45. Pullen, A. H.; Sears, T. A. Modification of 'C' synapses following partial deafferentation of thoracic motoneurons. Brain Res. 145:141146; 1978. 46. Pullen, A. H.; Sears, T. A. Trophism between C-type axon terminals and thoracic motoneurones in the cat. J. Physiol. 337:373-388; 1983. 47. Romanes, G. J. Motor localization and the effects of nerve injury on the ventral horn cells of the spinal cord. J. Anat. 80:117-131; 1946. 48. Romanes, G. J. The motor cell columns of the lumbosacral spinal cord of the cat. J. Comp. Neurol. 94:312-363; 1951. 49. Romanovicz, D. K.; Hanker, J. S. Wafer embedding: specimen selection in electron microscopic cytochemistry with osmiophilic polymers. Histochem. J. 9:317-327; 1977. 50. Rose, G.; Lych, G.; Cotman. C. W. Hypertrophy and redistribution of astrocytes in the deafferented dentate gyrus. Brain Res. Bull. 1:87-92; 1976. 51. Schmalbruch, H. Motoneuron death after sciatic nerve section in newborn rats. J. Comp. Neurol. 224:252-258; 1984. 52. Sears, T. A. The fibre calibre spectra of sensory and motor fibres in the intercostal nerves of the cat. J. Physiol. 172:150-161; 1964. 53. Sears, T. A. Structural changes in intercostal motoneurones following axotomy. J. Exp. Biol. 132:93-109; 1987. 54. Skene, J. H.; Willard, M. Axonally transported proteins associated with axon growth in rabbit central and peripheral nervous system. J. Cell Biol. 89:96-103; 1981. 55. Screide, A. J. Variations in the perineuronal glial changes after different types of nerve lesion: Light and electron microscopic investi-

JOHNSON, SEARS AND HUNTER

56.

57.

58. 59.

60.

61. 62. 63.

64.

65. 66. 67. 68.

gations on the facial nucleus of the rat. Neuropathol. Appl. Neurobiol. 7:195-204; 1981. Sturrock, R. R. A morphological study of the neostriatum of aged mice with particular reference to neuroglia. J. Hirnforsch. 5:505515; 1987. Sumner, B. E. H.; Watson, W. E. Retraction and expansion of the dendritic tree of motor neurones of adult rats induced in vivo. Nature 233:273-275; 1971. Tandan, R.; Bradley, W. G. Amyotrophic lateral sclerosis. Part 2. Etiopathogenesis. Ann. Neurol. 18:419-431; 1985. Tetzlaff, W.; Graeber, M. B.; Bisby, M. A.; Kreutzberg, G. W. Increased glial fibrillary acidic protein synthesis in astrocytes during retrograde reaction of the rat facial nucleus. Glia 1:90-95; 1988. Torvik, A.; SkjOrten, F. Electron microscopic observations of nerve cell regeneration and degeneration after axon lesions. II. Changes in the glial cells. Acta Neuropathol. (Berl.) 17:265-282; 1971. Torvik, A.; S0reide, A. J. Nerve cell regeneration after axon lesions in newborn rabbits. J. Neuropath. Exp. Neurol. 31:683-695; 1972. Torvik, A.; S0reide, A. J. The perineuronal gliat reaction after axotomy. Brain Res. 95:519-529; 1975. Vaughn, J. E.; Peters, A. The morphology and development of neuroglial cells. In: Pease, D. C., ed. Cellular aspects of neural growth and differentiation; UCLA Forum in Medical Sciences No. 14. Berkeley: University of California Press; 1971 : 103-144. Vaughn, J. E.; Peters, A. Neuroglial cells in the cerebral cortex of rats from young adulthood to old age: An electron microscopic study. J. Neurocytol. 3:405-429; 1974. Warrington, W. B. On the structural alterations observed in nerve cells. J. Physiol. 23:112-129; 1898. Watson, W. E. Observations on the nucleolar and total cell body nucleic acid of injured nerve cells. J. Physiol. 196:655-676; 1968. Watson, W. E. Cellular responses to axotomy and related procedures. Br. Med. Bull. 30:415-435; 1974. Whitnall, M. H.; Grafstein, B. Changes in perikaryal organelles during axonal regeneration in goldfish retinal ganglion cells: an analysis of protein synthesis and routing. Brain Res. 272:49-56; 1983.