- Email: [email protected]
Effect of ionising radiation on the axon reaction of mouse anterior horn motor neurons
Journal of the Neurological Sciences, 1985, 67:1-14 Elsevier
1
JNS 02438
Effect of Ionising Radiation on the Axon Reaction of Mouse Anterior Horn Motor Neurons A Histological and Immunocytochemical Study Using a Monoclonal Antibody to Neurofilament Protein T . H . M o s s 1 a n d S.J. Lewkowicz 2 1Department of Neuropathology, Frenchay Hospital, Bristol, and 2M. R. C. Radiobiology Unit, HarweU, Didcot (U. K.) (Received 8 June, 1984) (Revised, received 31 July, 1984) (Accepted 5 August, 1984)
SUMMARY
The changes taking place in irradiated central nervous tissue prior to the onset of delayed radionecrosis are poorly understood, but functional abnormalities occurring during the latent interval after irradiation are likely to be of importance. In order to investigate functional disturbances in neurones during this period, unilateral sciatic nerve crush was performed in mice following sub-lethal X-irradiation of the lumbar spinal cord. Alterations in the axon reaction of anterior horn cells were studied using a monoclonal antibody to neurof'dament protein. With irradiation immediately prior to crush, the normal, well-defined increase in perikaryal neurofflament protein was significantly diminished, although there was no concurrent radiation necrosis and no alterations were seen in contralateral neurones with intact distal axon processes. The effect was more marked in neurones irradiated one month prior to nerve crush, and in the non-irradiated nerve crush region regeneration was delayed, with diminished neurofilament protein in the regenerating axons. These observations indicate that ionising radiation can progressively impair the ability ofneurones to synthesise neurofilament protein during distal axon regeneration. This may result from inadequate repair of radiation induced DNA strand-breaks, but may also follow more generalised damage to protein transcription enzymes and RNA metabolism.
Address for correspondence: T.H. Moss, Department of Neuropathology, Frenchay Hospital, Frenchay, Bristol BS16 1LE, U.K. 0022-410X/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)
Key words:
Axon
reaction -
Irradiation
-
Motor
neurone
-
Neurofilament
protein
-
Regeneration
INTRODUCTION
Delayed radionecrosis in the central nervous system is a well recognised phenomenon in both man (Pennybacker and Russell 1948; Crompton and Layton 1961) and animals (Haymaker et al. 1968; Goffmet et al. 1976), and may be preceded by a clinicallysilentintervalof up to severalyears duration, depending on species sensitivity and radiation dose (Zeman 1968). Although abnormalities may be detected in nervous tissue prior to the onset of delayed necrosis (Zeman 1968; Hamberger et al. 1970), the events occurring during the latentintervalare rather poorly understood (Caveness et al. 1963; Dorfman et al. 1982) and in particular, littleis known about alterations to neuronal perikarya during thisperiod. Minor changes to nodal m y ~ ( M a s t ~ a et al. 1976) and denervation changes in skeletalmuscle (Fewings ctal. 1977) have provided some histologicalevidence of neuronal abnormality during the latentinterval,but purely functional disturbances are also likelyto be important immediately followingirradiation (Olkowski ct al. 1972). Following trauma, for example, irradiatedcentralnervous tissue shows evidence of an impaired regenerativecapacity,with diminished glialproliferation and remyelination (Cavanagh 1968a; Blakemore and Patterson 1978). A similar functionalimpairment has not been previouslyshown to involve neuronal perikarya,but would only bc detectable following a suitable regenerative stimulus, such as that produced by distal axon trauma. This study describes an altered regenerative response in mouse anterior horn ncuroncs, occurring afterspinalirradiationbut before the onset of delayed degenerative changes. The irradiated neurones were stimulated by crushing their peripheral proccsses, and the resultingaxon reaction assessed by using a monoclonal antibody marker for neurofilament protein, RT97 (Anderton ct al. 1982). MATERIALS AND METHODS Materials
A total of 90 male, adult mice, with ases between I0 and 14 weeks, were taken from the inbred C B A / C a H strain maintained at the M. R. C. Radiobiology Unit, Hat-,veil. Animals in the main part of the study were divided into three groups: (i) nerve crush only, (ii) irradiation less than 4 h before nerve crush, and (iii) irradiatm'n 1 month before nerve crush. In each group, material was examined from at least 3 animals at 2, 5, 8,
I I, 14, 21 and 28 days afternerve crush. In addition,material from at least2 irradiated animals, which had not undergone nerve crush, was examined at monthly intervalsfrom 3 to 10months after irradiation, together with material from non-irradiated, agematched control animals.
Me~o~ (a) Nerve crush procedure Mice were anaesthetised by inhalation of 3% halothane (May and Baker) in oxygen. The sciatic nerve was exposed in the upper part of the right leg and crushed at the level of the sciatic notch using watchmaker's forceps, applied for 15 s. The incision was closed with 4-0 silk sutures and the animals allowed to recover.
(b) Irradiation procedure Animals were again anaesthetised With 3 % halothane (May and Baker) in oxygen. The maximum duration of anaesthesia did not exceed 25 rain. Radiation was given as a single dose of 25 Gy (2500 rads), using 250 kV HVT 1.1 mm Cu X-rays With an exposure rate of 3.21 Gy/min (321 rads/min, measured at the estimated position of the 2nd and 3rd lumbar spines in an equivalent tissue mass). Animals were positioned with the right lateral side facing the beam and protected by a lead shield. The irradiated field was def'med by a rectangular window in the shield, 1.0 x 1.5 cm in size and exposing the T13 to L4 vertebral spines (L2 to L6 spinal cord segments). The sciatic nerve in the region of the crush site was thus outside the irradiated field (Fig. 1).
Fig. 1. Radiograph of an anaesthetised animal positioned for radiation treatment. The area irradiated is defined by the aperture in a lead screen, seen here as a dark rectangle over the lumbar spine region. The site of sciatic nerve crush at the sciatic notch (arrow) is clearly outside the irradiated zone. 2 x life size.
(c) Histologicalpreparation Mice were anaesthetised with 0.6 mg/g intraperitoneal pentobarbitone and fixed by cardiac perfusion with Bouin's fluid. A posterior laminectomy was performed and the anatomical segments of the lumbar spinal cord were identified. The most cranial spinal rootlets were taken as delineating the upper extent of each cord segment. The motor and sensory roots at L4 were found to comprise the major anatomical contribution to the sciatic trunk in the lumbo-sacral plexus, and this cord segment was taken from all animals. In addition, a 5 mm segment of sciatic nerve was removed from both sides at the level of the sciatic notch. All material was prepared for light microscopy using standard methods. Serial 5 #m transverse sections of the wax-embedded cord segments were cut from the cranial end of each specimen. Every 10th section was mounted consecutively for immunoperoxidase staining (see below). Samples of intervening sections were stained with cresyl violet stain for Nissl substance, haematoxylin and eosin, or combined cresyl violet and luxol fast blue stain for myelin. Longitudinal sections of sciatic nerve taken through the point of widest diameter were stained with immunoperoxidase stain (see below), Palmgren's silver stain for axon cylinders, or Van Gieson's stain.
(d) Immunoperoxidase stain for neurofilament antibody A monoclonal antibody with mammalian neurofflament specificity (RT97) was used. The antibody had been raised in mice against the 3 neurofilma~mt triplet proteins, purified from rat brain. Details of the preparation and characterisatic~a of the antibody are described elsewhere (Wood and Anderton 1981). Sections were ~ by the indirect immunoperoxidase technique, using the antibody as a first layer and a rabbit anti-mouse peroxidase (DAKO) as second layer, both incubated for 30 rain at room temperature. The reaction product was precipitated with d i n m l n o ~ and 1~o H202, and the sections lightly counterstained with Harris's haematoxylin for 12 s.
(e) Quantificationof immunoperoxidase-stained sections All neurones staining positively in the spinal cord showed the morphological characteristics of anterior horn motor neurones. These positive staining neurones were counted in a minimum of 12 consecutive sections from each animal. The results were expressed as the mean number of positive cells per 10 sections. A response curve was obtained by plotting the results graphically against time aflar nerve crush.
09 Examination for delayed effects of radiation Neurological function in the re~naining ~nimals was ~ s s ~ l each week using a hanging reflex (Gofrmet et al. 1976). The mice were suspended from the rim of a tall glass beaker by their hind limbs a~! the time interval ~ which they could remain hanging unaided was recorded. For histological examination, both irradiated and control animals were anaesthetised and fixed by cardiac perfusion as described above (c). The entire thoracic and lumbar spinal column, including the spinal cord and nerve roots, was removed and decalcified for 3 weeks in buffered ethylene diaminetetraacetic acid (EDTA, 75 g/1 in 2 ~ formalin at pH 7.2). The tissue was then trans-
ferred to 70~ methanol, processed and wax embedded using standard methods. Sample transverse 10-pan sections of both lumbar and non-irradiated thoracic regions were stained with haematoxylin and eosin, Van-Gieson's stain, or combined luxol fast blue and periodic acid-Schiff stains for myelin and blood vessel architecture. RESULTSj
(I) Observations on lumbar spinal cord (a) Nerve crush without irradiation The histological and immunocytological changes observed in the non-irradiated, control animals were confined to perikarya of anterior horn motor neurones on the side of nerve crush. The changes were similar to those reported previously (see Discussion) and will be only briefly described here, for purposes of comparison. Cresyl violet stain for Nissl substance. Changes in motor neurones undergoing the axon reaction were not marked. The only obvious alteration on Nissl staining was flocculation of the normally distinctly clumped Nissl substance, giving the cytoplasm a more evenly granular appearance. This change was ill-defined and short-lived in the control animals, and only prominent at 2 and 5 days postoperatively. By 11 days after crush, the anterior horn cells had regained a more normal appearance with re-clumping of Nissl substance. Immunoperoxidase stain for neurofilamentprotein. On the contralateral side of the cord to the nerve crush, all neuronal perikarya, including anterior horn motor neurones, remained unstained with the neurofilament antibody (Fig. 2a). On the operated side, a rapidly increasing number of anterior horn motor neurones showed positive cytoplasmic staining following crush (Fig. 2b). This response was well-defmed and short-lived, with a maximum number of positively staining cells at 11 days and virtually no remaining positive neurones by 21 days after crush (Fig. 3). (b) Nerve crush combined with lumbar cord irradiation In both groups of irradiated animals, the histological and immunocytological changes were again entirely confined to the perikarya of anterior horn motor neurones on the side of the nerve crush. During the period of these changes, no evidence of necrosis or other degenerative change resulting from the radiation treatment was detected in either spinal cord tissue or nerve roots. Irradiation immediately prior to nerve crush. With cresyl violet staining, the changes in anterior horn motor neurones on the operated side appeared similar to those in non-irradiated control animals, and no qualitative alteration or prolongation of the perikaryal response to nerve crush could be detected. Using the immunoperoxidase stain for neurofilament antibody, the positive staining response of anterior horn motor neurones following crush was still present, and followed a similar, short-lived time course to that in the control animals. Quantitatively, however, the normal peak of the response was diminished (Fig. 3), with a reduction in the maximum number of positively staining neurones at 11 days after crush and a highly significant alteration in the overall shape of the response curve (P < 0.0001 using the chi-square test).
Fig. 2. Anterior horn motor neurones from a non-irradiated, control animal 11 days post crush, stained by immunoperoxidase for the antibody to neurofilament protein. Immunoperoxidase stain for RT97 with a light haematoxylin counterstain; x 750. a: Normal neurones from the unoperated side, showing negatively staining cell bodies outlined by the positively staining axon meshwork, b: Neurones from the operated side. Many cells present show uniformly positively staining cytoplasm. POSITIVE ANT. HORN CELLS/IO SECTIONS
T
50NON-IRRADIATED CONTROLS 40-
It" I ~ I .... ! ..... I T."'f "f" ~
30-
l
;I ,,/j -'~;
20-
~"."/"
10-
........
i"...t.. ' I~ "
.............. IRRADIATION ONE MONTH PRIOR TO CRUSH
!
'... .....
............
"~ ",
12
0
IRRADIATED IMMEDIATELY PRIOR TO CRUSH
"'T'. " .... -I-
......"I
I
I
I
I
I
5
10
15
20
25
30
DAYS AFTER CRUSH
Fig. 3. Graphs showing the response of anterior horn motor neurones to sciatic nerve crush alone (solid line), irradiation immediately prior to crush (dashes) and irradiation I month prior to nerve crush (dotted line), when stained with antl~ody to neurolUament protein. In non-irradiated, Control animals there is a short-lived, well-defined response with a maximum number of positive stming ceils at 11 days post e ~ h : In animals irradiated immediately prior to crush, the peak of the response at this postoperative age is diminished. In animals irradiated I month prior to nerve crush, the positive st~nlng response is more severely diminished and less well defined, with prolonged positive staining of many neurones at 14 and 21 days postoperatively. The vertical bar represents the standard error of the mean at each age.
Irradiation I monthprior to nerve crush. In this group of animals, alterations in the
axon reaction of anterior horn motor neurones following nerve crush were much more marked than in animals irradiated immediately prior to crush. On cresyl violet staining, neuronal perikarya showed the expected flocculation of Nissl substance, but this change was prolonged relative to that in the non-irradiated, control animals. Many neurones still appeared abnormal 14 days postoperatively (Fig. 4) and re-clumping of Nissl substance did not become comparable with control animals until after 21 days post crush. With the immunoperoxidase stain for neurofilament protein, the peak of the positive staining response at 11 days post crush was not only markedly diminished, but appeared poorly defined compared to that in control animals (Fig. 3). As with the changes in Nissl substance, the antibody staining response was also abnormally prolonged, with many cells still showing a positive reaction 14 days after crush and some examples remaining positively stained at 28 days (Fig. 3). The alteration of the overall shape of the response curve was again highly statistically significant (P < 0.0001 using the chi-square test). (2) Observations on sciatic nerve
Sciatic nerve was examined in the crush region, recognisable in each case by the localised fibrous scarfing of the perineurium. This part of the nerve was outside the
O
t
4a
b
Fig. 4. Anterior horn motor neurones 14 days post crush, stained for Nissl substance. Cresyl violet stain; x 750. a: Nerve crush without irradiation, showing a relatively normal appearance with re-clumping of cytoplasmic Nissl substance, b: Irradiation 1 month prior to nerve crush, showing neuronal swelling and persistence of granular dispersal of Nissl substance.
irradiated lumbar cord area (see Methods) and there was no histological evidence of radiation change on either the operated or contralateral side. The expected increase in cellularity following nerve crush did not appear diminished in irradiated animals as compared to control animals. (a) Nerve crush without lumbar cord irradiation Using both the silver stain and the antibody stain for neurofilament protein, the appearances following crush were similar to those described previously (Moss and Lewkowicz 1983) and only a few details important for comparative purposes will be mentioned here. Using the silver stain, thin longitudinally aligned regenerating axons were in'st observed running across the crush region after 5 days, and had become abundant in the same region by 14 days. With the antibody to neurofflament protein, the normal positive staining of these axons was relatively slow to re-appear, with positive staining processes still scanty at 11 days after crush and not becoming abundant until 28 days (Fig. 5a). (b) Nerve crush combined with lumbar cord irradiation Irradiation immediately prior to nerve crush. At each post-operative age examined, regenerating axons at the crush site showed a similar appearance on silver staining to that in the non-irradiated, control animals. The later increase in positive staining of regenerating axons for neurofilament antibody was again similar to that in the control animals. Irradiation 1 month prior to nerve crush. Both staining techniques suggested that regeneration of axons through the nerve crush region was delayed compared to the non-irradiated control animals. With the silver stain, regenerating axons were less abundant during the first 3 weeks after crush, with the most marked difference at 14 days. The appearances became comparable to the control animals by 28 days postoperatively. Positive staining of regenerating axons for n ~ ~ p r o t m also showed a greater delay than in control animals, with virtually no po~tively Staining axons present at 14 days post-operatively, and with a marked reduction i n staining persisting at least until 28 days (Fig. 5). (3) Late effects of lumbar cord irradiation (a) Neurological observations When suspended from the rim of a tall glass beaker, non-irradiated control animals remained hanging by their hind limbs for an indefinite period. In irradiated animals, this ability was preserved during the first 4 months after treatment and no neurological abnormalities were observed during this time. During the 5th month, many irradiated animals showed progressive weakness of the hanging rctlzx and also developed abnormal flexion of the hind limbs when suspendegl freely by the taft. By 6 months after treatment, none of the irradiated mice could support themselves by hanging from their hind limbs for more than 2 s. These abnormalities pzrsisted in animals allowed to survive up to 10 months after irradiation.
Fig. 5. Longitudinalsections of sciatic nerve from the crush site at 28 days after crush. Immunoperoxidase for the antibody to neurofllament protein, RT97, with a light haematoxylincounterstain; x 470. a: Nerve crush only,showingabundant regenerating axons staining positivelyfor neurofilamentprotein, b: Irradiation 1 month prior to nerve crush, showing fewer positively staining regenerating axons in the same region.
(b) Histological observations No abnormalities were seen in the spinal cord or roots of irradiated animals examined during the first 6 months after exposure. Histological evidence of delayed radiation change was first observed in one animal 7 months following irradiation and was present in all animals examined at later times up to 10 months. The changes were similar in all affected animals and consisted of bilateral, asymmetric areas of necrosis in the intraspinal part of posterior lumbar spinal roots (Fig. 6). Changes were not found outside the irradiated lumbar region, nor in age-matched control animals. The necrotic lesions varied in size, with smaller foci affecting only a few myelinated fibres and larger ones involving virtually the entire root at some levels. Affected areas showed fragmentation and loss of myelin sheaths and axons, with occasional small foci of haemorrhage, but no detectable cellular reaction. No degenerative changes of blood vessel walls were seen, and both white and grey matter of the spinal cord appeared preserved. In particular, large anterior horn motor neurones showed no visible abnormality at any age examined.
10
Fig. 6. A typical example of delayed radiation change in the L4 region of an animal allowed to survive 8 months after lumbar spine irradiation. White and grey matter of the spinal cord appears preserved, but there are bilateral, asymmetric areas of necrosis in the posterior spinal roots. Combined luxol fast blue and periodic acid-Schiff stain; × 80.
DISCUSSION
The results of this study have shown an altered response of irradiated but histologically intact anterior horn cells following sciatic nerve crush. Since the changes in neurofflament staining only occurred in neurones receiving the stimulus of distal axon trauma, the present observations suggest that irradiation in this dose range initially has a purely functional effect on central neurones, not detectable in normal, non-regenerating perikarya. An analogous effect has been observed in the peripheral nervous system, where the response of peripheral nerve elements to crush is known to be diminished following irradiation of the nerve tissue itsdf(Cavanagh 1968b; Love 1983). In the central nervous system, irradiation is known to diminish the proliferation of neurogha following trauma (Hopewell and Wright 1967; Cavanagh 1968a), but alteration in the regenerative response of neuronal perikarya following traumatic stimuli has not been previously demonstrated. The precise nature of the altered neuronal response shown here is uncertain, but is unlikely to be related to the delayed radiation necrosis eventually observed in longer surviving animals. These delayed degenerative changes were confined to spinal nerve roots, as described in the lumbar region of rats after similar radiation doses (Van der
11 K0gel 1979), and no damage to cord tissue or anterior horn cells was observed. Moreover, at the dose used here, the altered regenerative response of the anterior horn neurones occurred several months before the end of the latent interval, thus suggesting that the radiation had an earlier and more direct effect on the functional integrity of the perikarya. In non-regenerating neurones, a similar early functional effect has also been found after much lower doses of radiation, insufficient to produce delayed radionecrosis (Olkowski et al. 1972), and as in the present experiment, these early changes are likely to involve protein synthesis mechanisms. The increased staining of neuronal perikarya for neurofilament protein has been previously described in non-irradiated neurones reacting to distal axon trauma (Moss and Lewkowicz 1983), and appears to reflect the generalised increase in perikaryal protein synthesis which forms a normal part of the axon reaction (Torvick 1976; Grafstein and McQuattie 1978). The results of the present study thus indicate that ionising radiation may directly impair the ability of regenerating neurones to synthesise neurofilament protein, perhaps as a result of unrepaired damage to DNA strands. Interference with DNA synthesis is one of the most important effects of ionising radiation (Gray 1953; Nakazawa et al. 1965) and single or double DNA strand-breaks constitute a major mechanism of damage (Grosch and Hopwood 1979; Singh and Singh 1982). Neurones are a non-dividing cell population and thus have been considered relatively radioresistant (Cavanagh 1968a), but in addition to inhibiting mitosis, damage to DNA strands may also affect non-dividing cells due to impairment of DNAdependent RNA transcription, and hence also of protein synthesis (Shungskaya et al. 1966; Hanson and Zhivotovsky 1976). In the present study, the impairment of protein synthesis persisted at least 1 month after irradiation, thus suggesting that the underlying DNA damage may not have been adequately repaired. DNA strand-breaks can be rapidly corrected under favourable circumstances (Elkind and Sutton 1959; Zeman et al. 1968), but the efficiency of repair mechanisms varies with a number of factors including cell type and animal age (Wheeler et al. 1981; Singh and Singh 1982). Moreover, certain types of DNA damage are probably not amenable to repair (Painter 1974) and there is some evidence that faulty repair can occur, with persistent abnormality of DNA strands (Teebor et al. 1984). In addition to producing DNA strand-breaks, ionising radiation is known to affect other cell components (Singh and Singh 1982), and it is also possible that more generalised damage to nucleoprotein metabolism and cytoplasmic protein synthesis may have occurred in the neurones studied here. For example, there is in vitro evidence that radiation may directly inhibit the activity of RNA transcription enzymes (Borsa et al. 1979) and may also affect the functioning of both messenger RNA (Ponta et al. 1979) and transfer RNA (Walden and Farkas 1981 ; Singh and Vadasz 1983). In support of such mechanisms, other in vivo studies using low dose radiation have demonstrated impairment of both cytoplasmic RNA production and related enzyme activity in irradiated but non-regenerating central neurones of a variety of animal species (Yamamoto et al. 1964; Hamberger et al. 1970; Olkowski 1972). Whatever the mechanism involved in the neuronal perikarya, the effects of irradiation in this study were also associated with delayed distal axon regeneration at the non-irradiated crush site, indicating a possible impairment of axon transport. The
12 diminished neurofilament staining of the regenerating axons, for example, may have been due to disturbance oforthograde transport, with delayed bulk flow ofneurofflament protein out of the irradiated perikarya and down to the regenerating end of the axons (Lasek and Hoffman 1976). It is also possible that the abnormally late positive staining of many pre-irradiated perikarya resulted from retarded retrograde transport, with consequent delay in the return of degraded neurofflaments from the damaged ends of the axons back to the cell bodies (Moss and Lewkowicz 1983). In either case, it seems unlikely that any alterations in axon transport were the result of direct radiation damage to the exposed, proximal parts of the motor spinal roots, as these showed no histological degenerative changes, even in the longer surviving animals. It is possible, however, that axon transport may have been affected either by impaired energy production due to radiation induced mitochondrial damage (Grosch and Hopwood 1979; Ghadia and Shah 1983), or as a secondary result of the diminished synthesis of structural proteins in the neuronal perikarya. In conclusion, these experiments have shown that moderate doses of radiation have an early, deleterious effect on the regenerating capacity of anterior horn cells. This effect is probably related to faulty or inadequate repair of D N A strand-breaks, since it is more marked with time following X-ray exposure, but there may also be direct effects on RNA metabolism and protein transcription, or on energy-dependent axon transport mechanisms. While such functional impairment of irradiated neurones is unlikely to be related to delayed radionecrosis, it may affect the ability of the neuronal perikarya to repair the continuous incidental, minor damage known to occur in adult nerves with increasing age (Lubinska 1958; Sung and Mastri 1983). ACKNOWLEDGEMENTS The authors wish to thank Dr, B. Anderton for supplying his monoclonal antibody to neurofdament protein, RT97, without which this work would not have been possible, and also Dr. B. Browner for her encouragement and support. The technical ~sistance of Mrs. E. W. Young and Mr. A. Churchill is gratefully acknowledged. The manuscripts were typed by Mrs. P. Penfold. REFERENCES Anderton, B.H., D. Breinburg, M.J. Dowries, P.J. Green, B.E. Tomlinson,J. Ulrich, J.N. Wood and J. Kahn, (1982) Monoclonal antibodies show that neurofibriUarytangles and neurofllaments share antigenic determinants, Nature (Lend.), 298: 84-86. Blakemore,W. F. and R. C. Patterson (1978)Suppressionofremyelinationin the CNS byX-irradiation,Acta Neuropath. (Berl.), 42: 105-113. Borsa, J., M.D. Sargent and P. A. Lievaart(1979)Ionisingradiation perturbs the switch-onof transcriptase in a model transcription complexin vitro, Int. J. Radlat. biol., 35: 459-472. Cavanagh, J. B. (1968a)EffectsofprevionsX-irradiationon the cellularresponseof nervoustissue to injury, Nature (Lend.), 219: 626-627. Cavanagh, J.B. (1968b)Prior X-irradiation and the cellularresponse to nerve crush -- Duration of effect, Exp. Neurol., 22: 253-258. Caveness, W.F., L. Roizen, L R.M. Innes and A. Carsten (1963) Delayed effectsof X-irradiation on the central nervous system of the monkey. In: T.J. Haley and R.S. Snider (Eds.), Second International
13 Symposium on the Response of the Nervous System to lonising Radiation, Little, Brown and Co., Boston, pp. 448-475. Crompton, M.R. and D.D. Layton (1961) Delayed radionecrosis of the brain following therapeutic X-irradiation of the pituitary, Brain, 84: 85-101. Dorfman, L.J., S.S. Donaldson, P.R. Gupta and T.M. Bosley (1982) Electrophysiologic evidence of subclinical injury to the posterior columns of the human spinal cord after therapeutic radiation, Cancer, 50: 2815-2819. Elkind, M.M. and H. Sutton (1959) X-ray damage and recovery in mammalian cells in culture, Nature (Lond.), 184: 1293-1295. Fewings, J. P., J. B. Harris, M. A. Johnson and W. G. Bradley (1977) Chronic denervation of muscle by spinal cord irradiation, Brain, 100: 157-183. Ghadia, P.K. and U.C. Shah (1983) Radiation-induced alterations in succinate dehydrogenase activity in the muscle of pidgeon, J. Radiat. Res., 24: 203-208. Goffinet, D. R., G. W. Marsa and J. M. Brown (1976) The effects of single and multifraction radiation courses on the mouse spinal cord, Radiology, 119: 709-713. Grafstein, B. and I. G. McQuattie (1978) Role of the nerve cell body in axonal regeneration. In: C. V. Cotman (Ed.), Neuronal Plasticity, Raven Press, New York, pp. 155-195. Gray, L. H. (1953) The initiation and development of cellular damage by ionising radiations, Brit. J. Radiol. N. S., 26: 609-618. Grosch, D. S. and L.E. Hopwood (1979) Biological Effects of Radiations, 2nd edition, Academic Press Inc., London. Hamberger, A., C. Blomstrand and B. Rosengsen (1970) Effect of X-irradiation on respiration and protein synthesis in neuronal and neuroglial cell fractions, Exp. Neurol., 26: 509-517. Hanson, K.P. and B.D. Zhivotovsky (1976) Effects of X-irradiation on the hybridization of rat thymus nuclear RNA with repeated and unique DNA sequences, Int. J. Radiat. Biol., 30: 129-139. Haymaker, W., M.Z.M. Ibrahim, J. Miquel, N. Call and A.J. Riopelle (1968) Delayed radiation effects in the brains of monkeys exposed to X and yrays, J. Neuropath. Exp. Neurol., 27: 50-79. Hopewell, J.W. and E.A. Wright (1967) Effect of previous X-irradiation on the response of the brain to injury, Nature (Lond.), 215: 1405. Lasek, R. and P. Hoffman (1976) The neuronal cytoskeleton, axonal transport and axon growth. In: R.D. Goldman and T.D. Pollard (Eds.), Cell Motility, Vol. 3, Cold Spring Harbour Laboratories, pp. 1021-1049. Love, S. (1983) An experimental study of peripheral nerve regeneration after X-irradiation, Brain, 106: 39-55. Lubinska, L. (1958) Intercalated internodes in nerve fibres, Nature (Lond.), 181: 957-958. Mastaglia, F.L., W.I. McDonald, J.V. Watson and K. Yogerdran (1976) Effects of X-irradiation on the spinal cord - - An experimental study of the morphological changes in central nerve fibres, Brain, 99: 101-122. Moss, T.H. and S.J. Lewkowicz (1983) The axon reaction in motor and sensory neurones of mice studied by a monoclonal antibody marker of neurofilament protein, J. Neurol. Sci., 60: 267-280. Nakazawa, S., T. Hara and K. Ueki (1965) The effects of X-irradiation on protein metabolism of the brain, Can. J. Biochem., 43: 1091-1098. Olkowski, Z. (1972) Radioautographic studies on the [3H]Uridine incorporation into the motor neurones of the spinal cord of X-irradiated mice, Radiat. Res., 51: 280-292. Olkowski, Z., S.L. Manocha and G.H. Bourne (1972) Response of motorneurones of the spinal cord to gamma radiation - - A cytochemicai study, Strahlentherapie, 143: 202-210. Painter, R.B. (1974) DNA damage and repair in eukaryotic cells, Genetics, 78: 139-148. Pennybacker, J. and D.S. Russell (1948) Necrosis of the brain due to radiation therapy - - Clinical and pathological observations, J. Neurol. Neurosurg. Psychiat. , 11: 183-198. Ponta, H., M.-L. Pfennig-Yeh, E.F. Wagner, M. Schweiger and P. Herrlich (1979) Radiation sensitivity of messenger RNA, Mol. Gen. Genet., 175: 13-17. Shungskaya, U. E., E. N. Zarkh and N. D. Nikoleava (1966) Investigation of the amount of RNA in the cells of the spinal cord during gamma irradiation, Radiobiologia, 6: 23-27. Singh, A. and H. Singh (1982) Time-scale and nature of radiation-biological damage - - Approaches to modern radiation protection and post-induction therapy, Progr. Biophys. Mol. Biol., 39: 69-107. Singh, H. and J.A. Vadasz (1983) Effect of gamma radiation on E. coli ribosomes, to RNA and aminoacyl-tRNA synthetases, Int. J. Radiat. Biol., 43: 587-597. Sung, J.H. and A.R. Mastri (1983) Aberrant peripheral nerves and micro neuromas in otherwise normal medullas, J. Neuropath. Exp. Neurol., 42: 522-528.
14 Teebor, G.W., K. Frenkel and M. S. Goldstein (1984) Ionising radiation and tritium transmutation both cause formation of 5-hydroxymethyl-2'-deoxyuridine in cellular DNA Proc. Nat. Acad. Sci. (USA), 81 : 318-321. Torvick, A. (1976) Central chromatolysis and the axon reaction - - A re-appraisal, Neuropath. Appl. Neurobiol., 2: 423-432. Van der KSgel, A.J. (1979) Late Effects of Radiation on the Spinal Cord - - Dose-Effect Relationships and Pathogenesis, Radiobiological Institute of the Organisation for Health Research, TNO, Rijswijk, Netherlands. Walden, T. and W.R. Farkas (1981) The effect of X-irradiation on the capacity of transfer RNA to act as a substrate for guanine-tRNA transferase Radiat. Res., 87: 350-359. Wheeler, K.T., J.V. Wierowski and P. Ritter (1981) Are inducible components involved in the repair of irradiated neuronal and brain tumour DNA?, Int. J. Radiat. Biol., 40: 293-296. Wood, J. N. and B. H. Anderton (1981 ) Monoclonal antibodies to mammalian neurofilaments, Biosci. Rep., 1: 263-268. Yamamoto, Y.L., L.E. Feinendegen and V.P. Bond (1964) Effect of radiation on the RNA metabolism of the central nervous system, Radiat. Res., 21 : 36-45. Zeman, W. (1968) Histologic events during the latent interval in radiation injury. In: O.T. Bailey and D.E. Smith (Eds.), The Central Nervous System - - Some Experimental Models of Neurological Disease, Williams and Wilkins, Baltimore, pp. 184-200. Zeman, W., J.M. Ordy and T. Samorajski (1968) Modification of acute radiation effect on cerebellar neurones of mice by actinomycin D, Exp. Neurol., 21: 52-57.
1
JNS 02438
Effect of Ionising Radiation on the Axon Reaction of Mouse Anterior Horn Motor Neurons A Histological and Immunocytochemical Study Using a Monoclonal Antibody to Neurofilament Protein T . H . M o s s 1 a n d S.J. Lewkowicz 2 1Department of Neuropathology, Frenchay Hospital, Bristol, and 2M. R. C. Radiobiology Unit, HarweU, Didcot (U. K.) (Received 8 June, 1984) (Revised, received 31 July, 1984) (Accepted 5 August, 1984)
SUMMARY
The changes taking place in irradiated central nervous tissue prior to the onset of delayed radionecrosis are poorly understood, but functional abnormalities occurring during the latent interval after irradiation are likely to be of importance. In order to investigate functional disturbances in neurones during this period, unilateral sciatic nerve crush was performed in mice following sub-lethal X-irradiation of the lumbar spinal cord. Alterations in the axon reaction of anterior horn cells were studied using a monoclonal antibody to neurof'dament protein. With irradiation immediately prior to crush, the normal, well-defined increase in perikaryal neurofflament protein was significantly diminished, although there was no concurrent radiation necrosis and no alterations were seen in contralateral neurones with intact distal axon processes. The effect was more marked in neurones irradiated one month prior to nerve crush, and in the non-irradiated nerve crush region regeneration was delayed, with diminished neurofilament protein in the regenerating axons. These observations indicate that ionising radiation can progressively impair the ability ofneurones to synthesise neurofilament protein during distal axon regeneration. This may result from inadequate repair of radiation induced DNA strand-breaks, but may also follow more generalised damage to protein transcription enzymes and RNA metabolism.
Address for correspondence: T.H. Moss, Department of Neuropathology, Frenchay Hospital, Frenchay, Bristol BS16 1LE, U.K. 0022-410X/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)
Key words:
Axon
reaction -
Irradiation
-
Motor
neurone
-
Neurofilament
protein
-
Regeneration
INTRODUCTION
Delayed radionecrosis in the central nervous system is a well recognised phenomenon in both man (Pennybacker and Russell 1948; Crompton and Layton 1961) and animals (Haymaker et al. 1968; Goffmet et al. 1976), and may be preceded by a clinicallysilentintervalof up to severalyears duration, depending on species sensitivity and radiation dose (Zeman 1968). Although abnormalities may be detected in nervous tissue prior to the onset of delayed necrosis (Zeman 1968; Hamberger et al. 1970), the events occurring during the latentintervalare rather poorly understood (Caveness et al. 1963; Dorfman et al. 1982) and in particular, littleis known about alterations to neuronal perikarya during thisperiod. Minor changes to nodal m y ~ ( M a s t ~ a et al. 1976) and denervation changes in skeletalmuscle (Fewings ctal. 1977) have provided some histologicalevidence of neuronal abnormality during the latentinterval,but purely functional disturbances are also likelyto be important immediately followingirradiation (Olkowski ct al. 1972). Following trauma, for example, irradiatedcentralnervous tissue shows evidence of an impaired regenerativecapacity,with diminished glialproliferation and remyelination (Cavanagh 1968a; Blakemore and Patterson 1978). A similar functionalimpairment has not been previouslyshown to involve neuronal perikarya,but would only bc detectable following a suitable regenerative stimulus, such as that produced by distal axon trauma. This study describes an altered regenerative response in mouse anterior horn ncuroncs, occurring afterspinalirradiationbut before the onset of delayed degenerative changes. The irradiated neurones were stimulated by crushing their peripheral proccsses, and the resultingaxon reaction assessed by using a monoclonal antibody marker for neurofilament protein, RT97 (Anderton ct al. 1982). MATERIALS AND METHODS Materials
A total of 90 male, adult mice, with ases between I0 and 14 weeks, were taken from the inbred C B A / C a H strain maintained at the M. R. C. Radiobiology Unit, Hat-,veil. Animals in the main part of the study were divided into three groups: (i) nerve crush only, (ii) irradiation less than 4 h before nerve crush, and (iii) irradiatm'n 1 month before nerve crush. In each group, material was examined from at least 3 animals at 2, 5, 8,
I I, 14, 21 and 28 days afternerve crush. In addition,material from at least2 irradiated animals, which had not undergone nerve crush, was examined at monthly intervalsfrom 3 to 10months after irradiation, together with material from non-irradiated, agematched control animals.
Me~o~ (a) Nerve crush procedure Mice were anaesthetised by inhalation of 3% halothane (May and Baker) in oxygen. The sciatic nerve was exposed in the upper part of the right leg and crushed at the level of the sciatic notch using watchmaker's forceps, applied for 15 s. The incision was closed with 4-0 silk sutures and the animals allowed to recover.
(b) Irradiation procedure Animals were again anaesthetised With 3 % halothane (May and Baker) in oxygen. The maximum duration of anaesthesia did not exceed 25 rain. Radiation was given as a single dose of 25 Gy (2500 rads), using 250 kV HVT 1.1 mm Cu X-rays With an exposure rate of 3.21 Gy/min (321 rads/min, measured at the estimated position of the 2nd and 3rd lumbar spines in an equivalent tissue mass). Animals were positioned with the right lateral side facing the beam and protected by a lead shield. The irradiated field was def'med by a rectangular window in the shield, 1.0 x 1.5 cm in size and exposing the T13 to L4 vertebral spines (L2 to L6 spinal cord segments). The sciatic nerve in the region of the crush site was thus outside the irradiated field (Fig. 1).
Fig. 1. Radiograph of an anaesthetised animal positioned for radiation treatment. The area irradiated is defined by the aperture in a lead screen, seen here as a dark rectangle over the lumbar spine region. The site of sciatic nerve crush at the sciatic notch (arrow) is clearly outside the irradiated zone. 2 x life size.
(c) Histologicalpreparation Mice were anaesthetised with 0.6 mg/g intraperitoneal pentobarbitone and fixed by cardiac perfusion with Bouin's fluid. A posterior laminectomy was performed and the anatomical segments of the lumbar spinal cord were identified. The most cranial spinal rootlets were taken as delineating the upper extent of each cord segment. The motor and sensory roots at L4 were found to comprise the major anatomical contribution to the sciatic trunk in the lumbo-sacral plexus, and this cord segment was taken from all animals. In addition, a 5 mm segment of sciatic nerve was removed from both sides at the level of the sciatic notch. All material was prepared for light microscopy using standard methods. Serial 5 #m transverse sections of the wax-embedded cord segments were cut from the cranial end of each specimen. Every 10th section was mounted consecutively for immunoperoxidase staining (see below). Samples of intervening sections were stained with cresyl violet stain for Nissl substance, haematoxylin and eosin, or combined cresyl violet and luxol fast blue stain for myelin. Longitudinal sections of sciatic nerve taken through the point of widest diameter were stained with immunoperoxidase stain (see below), Palmgren's silver stain for axon cylinders, or Van Gieson's stain.
(d) Immunoperoxidase stain for neurofilament antibody A monoclonal antibody with mammalian neurofflament specificity (RT97) was used. The antibody had been raised in mice against the 3 neurofilma~mt triplet proteins, purified from rat brain. Details of the preparation and characterisatic~a of the antibody are described elsewhere (Wood and Anderton 1981). Sections were ~ by the indirect immunoperoxidase technique, using the antibody as a first layer and a rabbit anti-mouse peroxidase (DAKO) as second layer, both incubated for 30 rain at room temperature. The reaction product was precipitated with d i n m l n o ~ and 1~o H202, and the sections lightly counterstained with Harris's haematoxylin for 12 s.
(e) Quantificationof immunoperoxidase-stained sections All neurones staining positively in the spinal cord showed the morphological characteristics of anterior horn motor neurones. These positive staining neurones were counted in a minimum of 12 consecutive sections from each animal. The results were expressed as the mean number of positive cells per 10 sections. A response curve was obtained by plotting the results graphically against time aflar nerve crush.
09 Examination for delayed effects of radiation Neurological function in the re~naining ~nimals was ~ s s ~ l each week using a hanging reflex (Gofrmet et al. 1976). The mice were suspended from the rim of a tall glass beaker by their hind limbs a~! the time interval ~ which they could remain hanging unaided was recorded. For histological examination, both irradiated and control animals were anaesthetised and fixed by cardiac perfusion as described above (c). The entire thoracic and lumbar spinal column, including the spinal cord and nerve roots, was removed and decalcified for 3 weeks in buffered ethylene diaminetetraacetic acid (EDTA, 75 g/1 in 2 ~ formalin at pH 7.2). The tissue was then trans-
ferred to 70~ methanol, processed and wax embedded using standard methods. Sample transverse 10-pan sections of both lumbar and non-irradiated thoracic regions were stained with haematoxylin and eosin, Van-Gieson's stain, or combined luxol fast blue and periodic acid-Schiff stains for myelin and blood vessel architecture. RESULTSj
(I) Observations on lumbar spinal cord (a) Nerve crush without irradiation The histological and immunocytological changes observed in the non-irradiated, control animals were confined to perikarya of anterior horn motor neurones on the side of nerve crush. The changes were similar to those reported previously (see Discussion) and will be only briefly described here, for purposes of comparison. Cresyl violet stain for Nissl substance. Changes in motor neurones undergoing the axon reaction were not marked. The only obvious alteration on Nissl staining was flocculation of the normally distinctly clumped Nissl substance, giving the cytoplasm a more evenly granular appearance. This change was ill-defined and short-lived in the control animals, and only prominent at 2 and 5 days postoperatively. By 11 days after crush, the anterior horn cells had regained a more normal appearance with re-clumping of Nissl substance. Immunoperoxidase stain for neurofilamentprotein. On the contralateral side of the cord to the nerve crush, all neuronal perikarya, including anterior horn motor neurones, remained unstained with the neurofilament antibody (Fig. 2a). On the operated side, a rapidly increasing number of anterior horn motor neurones showed positive cytoplasmic staining following crush (Fig. 2b). This response was well-defmed and short-lived, with a maximum number of positively staining cells at 11 days and virtually no remaining positive neurones by 21 days after crush (Fig. 3). (b) Nerve crush combined with lumbar cord irradiation In both groups of irradiated animals, the histological and immunocytological changes were again entirely confined to the perikarya of anterior horn motor neurones on the side of the nerve crush. During the period of these changes, no evidence of necrosis or other degenerative change resulting from the radiation treatment was detected in either spinal cord tissue or nerve roots. Irradiation immediately prior to nerve crush. With cresyl violet staining, the changes in anterior horn motor neurones on the operated side appeared similar to those in non-irradiated control animals, and no qualitative alteration or prolongation of the perikaryal response to nerve crush could be detected. Using the immunoperoxidase stain for neurofilament antibody, the positive staining response of anterior horn motor neurones following crush was still present, and followed a similar, short-lived time course to that in the control animals. Quantitatively, however, the normal peak of the response was diminished (Fig. 3), with a reduction in the maximum number of positively staining neurones at 11 days after crush and a highly significant alteration in the overall shape of the response curve (P < 0.0001 using the chi-square test).
Fig. 2. Anterior horn motor neurones from a non-irradiated, control animal 11 days post crush, stained by immunoperoxidase for the antibody to neurofilament protein. Immunoperoxidase stain for RT97 with a light haematoxylin counterstain; x 750. a: Normal neurones from the unoperated side, showing negatively staining cell bodies outlined by the positively staining axon meshwork, b: Neurones from the operated side. Many cells present show uniformly positively staining cytoplasm. POSITIVE ANT. HORN CELLS/IO SECTIONS
T
50NON-IRRADIATED CONTROLS 40-
It" I ~ I .... ! ..... I T."'f "f" ~
30-
l
;I ,,/j -'~;
20-
~"."/"
10-
........
i"...t.. ' I~ "
.............. IRRADIATION ONE MONTH PRIOR TO CRUSH
!
'... .....
............
"~ ",
12
0
IRRADIATED IMMEDIATELY PRIOR TO CRUSH
"'T'. " .... -I-
......"I
I
I
I
I
I
5
10
15
20
25
30
DAYS AFTER CRUSH
Fig. 3. Graphs showing the response of anterior horn motor neurones to sciatic nerve crush alone (solid line), irradiation immediately prior to crush (dashes) and irradiation I month prior to nerve crush (dotted line), when stained with antl~ody to neurolUament protein. In non-irradiated, Control animals there is a short-lived, well-defined response with a maximum number of positive stming ceils at 11 days post e ~ h : In animals irradiated immediately prior to crush, the peak of the response at this postoperative age is diminished. In animals irradiated I month prior to nerve crush, the positive st~nlng response is more severely diminished and less well defined, with prolonged positive staining of many neurones at 14 and 21 days postoperatively. The vertical bar represents the standard error of the mean at each age.
Irradiation I monthprior to nerve crush. In this group of animals, alterations in the
axon reaction of anterior horn motor neurones following nerve crush were much more marked than in animals irradiated immediately prior to crush. On cresyl violet staining, neuronal perikarya showed the expected flocculation of Nissl substance, but this change was prolonged relative to that in the non-irradiated, control animals. Many neurones still appeared abnormal 14 days postoperatively (Fig. 4) and re-clumping of Nissl substance did not become comparable with control animals until after 21 days post crush. With the immunoperoxidase stain for neurofilament protein, the peak of the positive staining response at 11 days post crush was not only markedly diminished, but appeared poorly defined compared to that in control animals (Fig. 3). As with the changes in Nissl substance, the antibody staining response was also abnormally prolonged, with many cells still showing a positive reaction 14 days after crush and some examples remaining positively stained at 28 days (Fig. 3). The alteration of the overall shape of the response curve was again highly statistically significant (P < 0.0001 using the chi-square test). (2) Observations on sciatic nerve
Sciatic nerve was examined in the crush region, recognisable in each case by the localised fibrous scarfing of the perineurium. This part of the nerve was outside the
O
t
4a
b
Fig. 4. Anterior horn motor neurones 14 days post crush, stained for Nissl substance. Cresyl violet stain; x 750. a: Nerve crush without irradiation, showing a relatively normal appearance with re-clumping of cytoplasmic Nissl substance, b: Irradiation 1 month prior to nerve crush, showing neuronal swelling and persistence of granular dispersal of Nissl substance.
irradiated lumbar cord area (see Methods) and there was no histological evidence of radiation change on either the operated or contralateral side. The expected increase in cellularity following nerve crush did not appear diminished in irradiated animals as compared to control animals. (a) Nerve crush without lumbar cord irradiation Using both the silver stain and the antibody stain for neurofilament protein, the appearances following crush were similar to those described previously (Moss and Lewkowicz 1983) and only a few details important for comparative purposes will be mentioned here. Using the silver stain, thin longitudinally aligned regenerating axons were in'st observed running across the crush region after 5 days, and had become abundant in the same region by 14 days. With the antibody to neurofflament protein, the normal positive staining of these axons was relatively slow to re-appear, with positive staining processes still scanty at 11 days after crush and not becoming abundant until 28 days (Fig. 5a). (b) Nerve crush combined with lumbar cord irradiation Irradiation immediately prior to nerve crush. At each post-operative age examined, regenerating axons at the crush site showed a similar appearance on silver staining to that in the non-irradiated, control animals. The later increase in positive staining of regenerating axons for neurofilament antibody was again similar to that in the control animals. Irradiation 1 month prior to nerve crush. Both staining techniques suggested that regeneration of axons through the nerve crush region was delayed compared to the non-irradiated control animals. With the silver stain, regenerating axons were less abundant during the first 3 weeks after crush, with the most marked difference at 14 days. The appearances became comparable to the control animals by 28 days postoperatively. Positive staining of regenerating axons for n ~ ~ p r o t m also showed a greater delay than in control animals, with virtually no po~tively Staining axons present at 14 days post-operatively, and with a marked reduction i n staining persisting at least until 28 days (Fig. 5). (3) Late effects of lumbar cord irradiation (a) Neurological observations When suspended from the rim of a tall glass beaker, non-irradiated control animals remained hanging by their hind limbs for an indefinite period. In irradiated animals, this ability was preserved during the first 4 months after treatment and no neurological abnormalities were observed during this time. During the 5th month, many irradiated animals showed progressive weakness of the hanging rctlzx and also developed abnormal flexion of the hind limbs when suspendegl freely by the taft. By 6 months after treatment, none of the irradiated mice could support themselves by hanging from their hind limbs for more than 2 s. These abnormalities pzrsisted in animals allowed to survive up to 10 months after irradiation.
Fig. 5. Longitudinalsections of sciatic nerve from the crush site at 28 days after crush. Immunoperoxidase for the antibody to neurofllament protein, RT97, with a light haematoxylincounterstain; x 470. a: Nerve crush only,showingabundant regenerating axons staining positivelyfor neurofilamentprotein, b: Irradiation 1 month prior to nerve crush, showing fewer positively staining regenerating axons in the same region.
(b) Histological observations No abnormalities were seen in the spinal cord or roots of irradiated animals examined during the first 6 months after exposure. Histological evidence of delayed radiation change was first observed in one animal 7 months following irradiation and was present in all animals examined at later times up to 10 months. The changes were similar in all affected animals and consisted of bilateral, asymmetric areas of necrosis in the intraspinal part of posterior lumbar spinal roots (Fig. 6). Changes were not found outside the irradiated lumbar region, nor in age-matched control animals. The necrotic lesions varied in size, with smaller foci affecting only a few myelinated fibres and larger ones involving virtually the entire root at some levels. Affected areas showed fragmentation and loss of myelin sheaths and axons, with occasional small foci of haemorrhage, but no detectable cellular reaction. No degenerative changes of blood vessel walls were seen, and both white and grey matter of the spinal cord appeared preserved. In particular, large anterior horn motor neurones showed no visible abnormality at any age examined.
10
Fig. 6. A typical example of delayed radiation change in the L4 region of an animal allowed to survive 8 months after lumbar spine irradiation. White and grey matter of the spinal cord appears preserved, but there are bilateral, asymmetric areas of necrosis in the posterior spinal roots. Combined luxol fast blue and periodic acid-Schiff stain; × 80.
DISCUSSION
The results of this study have shown an altered response of irradiated but histologically intact anterior horn cells following sciatic nerve crush. Since the changes in neurofflament staining only occurred in neurones receiving the stimulus of distal axon trauma, the present observations suggest that irradiation in this dose range initially has a purely functional effect on central neurones, not detectable in normal, non-regenerating perikarya. An analogous effect has been observed in the peripheral nervous system, where the response of peripheral nerve elements to crush is known to be diminished following irradiation of the nerve tissue itsdf(Cavanagh 1968b; Love 1983). In the central nervous system, irradiation is known to diminish the proliferation of neurogha following trauma (Hopewell and Wright 1967; Cavanagh 1968a), but alteration in the regenerative response of neuronal perikarya following traumatic stimuli has not been previously demonstrated. The precise nature of the altered neuronal response shown here is uncertain, but is unlikely to be related to the delayed radiation necrosis eventually observed in longer surviving animals. These delayed degenerative changes were confined to spinal nerve roots, as described in the lumbar region of rats after similar radiation doses (Van der
11 K0gel 1979), and no damage to cord tissue or anterior horn cells was observed. Moreover, at the dose used here, the altered regenerative response of the anterior horn neurones occurred several months before the end of the latent interval, thus suggesting that the radiation had an earlier and more direct effect on the functional integrity of the perikarya. In non-regenerating neurones, a similar early functional effect has also been found after much lower doses of radiation, insufficient to produce delayed radionecrosis (Olkowski et al. 1972), and as in the present experiment, these early changes are likely to involve protein synthesis mechanisms. The increased staining of neuronal perikarya for neurofilament protein has been previously described in non-irradiated neurones reacting to distal axon trauma (Moss and Lewkowicz 1983), and appears to reflect the generalised increase in perikaryal protein synthesis which forms a normal part of the axon reaction (Torvick 1976; Grafstein and McQuattie 1978). The results of the present study thus indicate that ionising radiation may directly impair the ability of regenerating neurones to synthesise neurofilament protein, perhaps as a result of unrepaired damage to DNA strands. Interference with DNA synthesis is one of the most important effects of ionising radiation (Gray 1953; Nakazawa et al. 1965) and single or double DNA strand-breaks constitute a major mechanism of damage (Grosch and Hopwood 1979; Singh and Singh 1982). Neurones are a non-dividing cell population and thus have been considered relatively radioresistant (Cavanagh 1968a), but in addition to inhibiting mitosis, damage to DNA strands may also affect non-dividing cells due to impairment of DNAdependent RNA transcription, and hence also of protein synthesis (Shungskaya et al. 1966; Hanson and Zhivotovsky 1976). In the present study, the impairment of protein synthesis persisted at least 1 month after irradiation, thus suggesting that the underlying DNA damage may not have been adequately repaired. DNA strand-breaks can be rapidly corrected under favourable circumstances (Elkind and Sutton 1959; Zeman et al. 1968), but the efficiency of repair mechanisms varies with a number of factors including cell type and animal age (Wheeler et al. 1981; Singh and Singh 1982). Moreover, certain types of DNA damage are probably not amenable to repair (Painter 1974) and there is some evidence that faulty repair can occur, with persistent abnormality of DNA strands (Teebor et al. 1984). In addition to producing DNA strand-breaks, ionising radiation is known to affect other cell components (Singh and Singh 1982), and it is also possible that more generalised damage to nucleoprotein metabolism and cytoplasmic protein synthesis may have occurred in the neurones studied here. For example, there is in vitro evidence that radiation may directly inhibit the activity of RNA transcription enzymes (Borsa et al. 1979) and may also affect the functioning of both messenger RNA (Ponta et al. 1979) and transfer RNA (Walden and Farkas 1981 ; Singh and Vadasz 1983). In support of such mechanisms, other in vivo studies using low dose radiation have demonstrated impairment of both cytoplasmic RNA production and related enzyme activity in irradiated but non-regenerating central neurones of a variety of animal species (Yamamoto et al. 1964; Hamberger et al. 1970; Olkowski 1972). Whatever the mechanism involved in the neuronal perikarya, the effects of irradiation in this study were also associated with delayed distal axon regeneration at the non-irradiated crush site, indicating a possible impairment of axon transport. The
12 diminished neurofilament staining of the regenerating axons, for example, may have been due to disturbance oforthograde transport, with delayed bulk flow ofneurofflament protein out of the irradiated perikarya and down to the regenerating end of the axons (Lasek and Hoffman 1976). It is also possible that the abnormally late positive staining of many pre-irradiated perikarya resulted from retarded retrograde transport, with consequent delay in the return of degraded neurofflaments from the damaged ends of the axons back to the cell bodies (Moss and Lewkowicz 1983). In either case, it seems unlikely that any alterations in axon transport were the result of direct radiation damage to the exposed, proximal parts of the motor spinal roots, as these showed no histological degenerative changes, even in the longer surviving animals. It is possible, however, that axon transport may have been affected either by impaired energy production due to radiation induced mitochondrial damage (Grosch and Hopwood 1979; Ghadia and Shah 1983), or as a secondary result of the diminished synthesis of structural proteins in the neuronal perikarya. In conclusion, these experiments have shown that moderate doses of radiation have an early, deleterious effect on the regenerating capacity of anterior horn cells. This effect is probably related to faulty or inadequate repair of D N A strand-breaks, since it is more marked with time following X-ray exposure, but there may also be direct effects on RNA metabolism and protein transcription, or on energy-dependent axon transport mechanisms. While such functional impairment of irradiated neurones is unlikely to be related to delayed radionecrosis, it may affect the ability of the neuronal perikarya to repair the continuous incidental, minor damage known to occur in adult nerves with increasing age (Lubinska 1958; Sung and Mastri 1983). ACKNOWLEDGEMENTS The authors wish to thank Dr, B. Anderton for supplying his monoclonal antibody to neurofdament protein, RT97, without which this work would not have been possible, and also Dr. B. Browner for her encouragement and support. The technical ~sistance of Mrs. E. W. Young and Mr. A. Churchill is gratefully acknowledged. The manuscripts were typed by Mrs. P. Penfold. REFERENCES Anderton, B.H., D. Breinburg, M.J. Dowries, P.J. Green, B.E. Tomlinson,J. Ulrich, J.N. Wood and J. Kahn, (1982) Monoclonal antibodies show that neurofibriUarytangles and neurofllaments share antigenic determinants, Nature (Lend.), 298: 84-86. Blakemore,W. F. and R. C. Patterson (1978)Suppressionofremyelinationin the CNS byX-irradiation,Acta Neuropath. (Berl.), 42: 105-113. Borsa, J., M.D. Sargent and P. A. Lievaart(1979)Ionisingradiation perturbs the switch-onof transcriptase in a model transcription complexin vitro, Int. J. Radlat. biol., 35: 459-472. Cavanagh, J. B. (1968a)EffectsofprevionsX-irradiationon the cellularresponseof nervoustissue to injury, Nature (Lend.), 219: 626-627. Cavanagh, J.B. (1968b)Prior X-irradiation and the cellularresponse to nerve crush -- Duration of effect, Exp. Neurol., 22: 253-258. Caveness, W.F., L. Roizen, L R.M. Innes and A. Carsten (1963) Delayed effectsof X-irradiation on the central nervous system of the monkey. In: T.J. Haley and R.S. Snider (Eds.), Second International
13 Symposium on the Response of the Nervous System to lonising Radiation, Little, Brown and Co., Boston, pp. 448-475. Crompton, M.R. and D.D. Layton (1961) Delayed radionecrosis of the brain following therapeutic X-irradiation of the pituitary, Brain, 84: 85-101. Dorfman, L.J., S.S. Donaldson, P.R. Gupta and T.M. Bosley (1982) Electrophysiologic evidence of subclinical injury to the posterior columns of the human spinal cord after therapeutic radiation, Cancer, 50: 2815-2819. Elkind, M.M. and H. Sutton (1959) X-ray damage and recovery in mammalian cells in culture, Nature (Lond.), 184: 1293-1295. Fewings, J. P., J. B. Harris, M. A. Johnson and W. G. Bradley (1977) Chronic denervation of muscle by spinal cord irradiation, Brain, 100: 157-183. Ghadia, P.K. and U.C. Shah (1983) Radiation-induced alterations in succinate dehydrogenase activity in the muscle of pidgeon, J. Radiat. Res., 24: 203-208. Goffinet, D. R., G. W. Marsa and J. M. Brown (1976) The effects of single and multifraction radiation courses on the mouse spinal cord, Radiology, 119: 709-713. Grafstein, B. and I. G. McQuattie (1978) Role of the nerve cell body in axonal regeneration. In: C. V. Cotman (Ed.), Neuronal Plasticity, Raven Press, New York, pp. 155-195. Gray, L. H. (1953) The initiation and development of cellular damage by ionising radiations, Brit. J. Radiol. N. S., 26: 609-618. Grosch, D. S. and L.E. Hopwood (1979) Biological Effects of Radiations, 2nd edition, Academic Press Inc., London. Hamberger, A., C. Blomstrand and B. Rosengsen (1970) Effect of X-irradiation on respiration and protein synthesis in neuronal and neuroglial cell fractions, Exp. Neurol., 26: 509-517. Hanson, K.P. and B.D. Zhivotovsky (1976) Effects of X-irradiation on the hybridization of rat thymus nuclear RNA with repeated and unique DNA sequences, Int. J. Radiat. Biol., 30: 129-139. Haymaker, W., M.Z.M. Ibrahim, J. Miquel, N. Call and A.J. Riopelle (1968) Delayed radiation effects in the brains of monkeys exposed to X and yrays, J. Neuropath. Exp. Neurol., 27: 50-79. Hopewell, J.W. and E.A. Wright (1967) Effect of previous X-irradiation on the response of the brain to injury, Nature (Lond.), 215: 1405. Lasek, R. and P. Hoffman (1976) The neuronal cytoskeleton, axonal transport and axon growth. In: R.D. Goldman and T.D. Pollard (Eds.), Cell Motility, Vol. 3, Cold Spring Harbour Laboratories, pp. 1021-1049. Love, S. (1983) An experimental study of peripheral nerve regeneration after X-irradiation, Brain, 106: 39-55. Lubinska, L. (1958) Intercalated internodes in nerve fibres, Nature (Lond.), 181: 957-958. Mastaglia, F.L., W.I. McDonald, J.V. Watson and K. Yogerdran (1976) Effects of X-irradiation on the spinal cord - - An experimental study of the morphological changes in central nerve fibres, Brain, 99: 101-122. Moss, T.H. and S.J. Lewkowicz (1983) The axon reaction in motor and sensory neurones of mice studied by a monoclonal antibody marker of neurofilament protein, J. Neurol. Sci., 60: 267-280. Nakazawa, S., T. Hara and K. Ueki (1965) The effects of X-irradiation on protein metabolism of the brain, Can. J. Biochem., 43: 1091-1098. Olkowski, Z. (1972) Radioautographic studies on the [3H]Uridine incorporation into the motor neurones of the spinal cord of X-irradiated mice, Radiat. Res., 51: 280-292. Olkowski, Z., S.L. Manocha and G.H. Bourne (1972) Response of motorneurones of the spinal cord to gamma radiation - - A cytochemicai study, Strahlentherapie, 143: 202-210. Painter, R.B. (1974) DNA damage and repair in eukaryotic cells, Genetics, 78: 139-148. Pennybacker, J. and D.S. Russell (1948) Necrosis of the brain due to radiation therapy - - Clinical and pathological observations, J. Neurol. Neurosurg. Psychiat. , 11: 183-198. Ponta, H., M.-L. Pfennig-Yeh, E.F. Wagner, M. Schweiger and P. Herrlich (1979) Radiation sensitivity of messenger RNA, Mol. Gen. Genet., 175: 13-17. Shungskaya, U. E., E. N. Zarkh and N. D. Nikoleava (1966) Investigation of the amount of RNA in the cells of the spinal cord during gamma irradiation, Radiobiologia, 6: 23-27. Singh, A. and H. Singh (1982) Time-scale and nature of radiation-biological damage - - Approaches to modern radiation protection and post-induction therapy, Progr. Biophys. Mol. Biol., 39: 69-107. Singh, H. and J.A. Vadasz (1983) Effect of gamma radiation on E. coli ribosomes, to RNA and aminoacyl-tRNA synthetases, Int. J. Radiat. Biol., 43: 587-597. Sung, J.H. and A.R. Mastri (1983) Aberrant peripheral nerves and micro neuromas in otherwise normal medullas, J. Neuropath. Exp. Neurol., 42: 522-528.
14 Teebor, G.W., K. Frenkel and M. S. Goldstein (1984) Ionising radiation and tritium transmutation both cause formation of 5-hydroxymethyl-2'-deoxyuridine in cellular DNA Proc. Nat. Acad. Sci. (USA), 81 : 318-321. Torvick, A. (1976) Central chromatolysis and the axon reaction - - A re-appraisal, Neuropath. Appl. Neurobiol., 2: 423-432. Van der KSgel, A.J. (1979) Late Effects of Radiation on the Spinal Cord - - Dose-Effect Relationships and Pathogenesis, Radiobiological Institute of the Organisation for Health Research, TNO, Rijswijk, Netherlands. Walden, T. and W.R. Farkas (1981) The effect of X-irradiation on the capacity of transfer RNA to act as a substrate for guanine-tRNA transferase Radiat. Res., 87: 350-359. Wheeler, K.T., J.V. Wierowski and P. Ritter (1981) Are inducible components involved in the repair of irradiated neuronal and brain tumour DNA?, Int. J. Radiat. Biol., 40: 293-296. Wood, J. N. and B. H. Anderton (1981 ) Monoclonal antibodies to mammalian neurofilaments, Biosci. Rep., 1: 263-268. Yamamoto, Y.L., L.E. Feinendegen and V.P. Bond (1964) Effect of radiation on the RNA metabolism of the central nervous system, Radiat. Res., 21 : 36-45. Zeman, W. (1968) Histologic events during the latent interval in radiation injury. In: O.T. Bailey and D.E. Smith (Eds.), The Central Nervous System - - Some Experimental Models of Neurological Disease, Williams and Wilkins, Baltimore, pp. 184-200. Zeman, W., J.M. Ordy and T. Samorajski (1968) Modification of acute radiation effect on cerebellar neurones of mice by actinomycin D, Exp. Neurol., 21: 52-57.