Primary degeneration of motor neurons by toxic lectins conveyed from the peripheral nerve

Journal of the Neurological Sciences, 1985, 70:327-337

327

Elsevier JNS 2558

Primary Degeneration of Motor Neurons by Toxic Lectins Conveyed from the Peripheral Nerve Teiji Y a m a m o t o , Y u z o lwasaki, H i d e h i k o K o n n o and H i r o k o K u d o Department of Neuropathology, Institute of Brain Diseases, Tohoku University School of Medicine, Sendai, Miyagi 980 (Japan)

(Received 11 April, 1985) (Revised, received 6 June, 1985) (Accepted 6 June, 1985)

SUMMARY In attempts to degenerate motor neurons experimentally by way of retrograde axoplasmic transport, ricinus communis agglutinin (RCA), a potent protein inhibitor, was intraneurally injected into the rat sciatic nerve. ImmunohistochemicaUy, RCA was shown to be intra-axonaUy carded up to motor neuronal soma and to the dorsal root ganglia of L4_6. Within a few days, these dorsal root ganglion cells and large motor neurons giving rise to sciatic nerve efferents in the lumbar spinal cord degenerated, whereas small internuncial neurons and glia remained unaffected. The degeneration of motor neurons was characterized by a profound diffuse chromatolysis and subsequent dissolution, after which a mild gliosis remained. The retrograde axoplasmic flow of neurotoxic substance and motor neuron degeneration observed here may be a phenomenon implicated in the pathogenesis of human motor neuron diseases.

Key words: Lectin - M o t o r neuron disease - Retrograde axoplasmic transport - Ricin

INTRODUCTION In the past decade, ample evidence has been accumulated to indicate that a wide variety of substances can be transported by retrograde axoplasmic flow (Dumas et al. This study was in part supported by grants from the Educational Ministry (Project 60480219). Correspondenceshould be addressed to: Teiji Yamamoto, M.D., Department of Neuropathology, Institute of Brain Diseases, Tohoku UniversitySchoolof Medicine, 1-1 Seiryo-Machi,Sendai,Miyagi980, Japan. 0022-510X/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)

328 1979; Lasek 1980). Although the physiological significance of retrograde axoplasmic transport, particularly of exogenous substances, largely remains to be elucidated, the role of the axoplasmic transportation system in the etiology of motor neuron diseases, such as amyotrophic lateral sclerosis, is emphasized in a recent theory of motor neuron trophic factor deficiency (Appel 1981). In order to evaluate the roles of retrogradely transported substances in motor neuron activities of the spinal cord, we initiated studies of motor neuron changes induced by exogenous neurotoxins applied to the peripheral nerve. When doxorubicin, an antineoplastic antibiotic with DNA-directed RNA inhibiting action, was injected into the rat sciatic nerve, intranuclear accumulation of this agent in parental motor neurons resulted in neuronal degeneration and ultimate loss of motor neurons from corresponding segments of the spinal cord (Yamamoto et al. 1984). In this communication, we report a series of pathological changes in rat spinal cord motor neurons after intraneural injection of toxic lectins, potent inhibitors of protein synthesis, into the sciatic nerve. The pathology induced by these lectins is distinct from that of doxorubicin, and we hope that studies on the comparative pathology of motor neurons induced by various neurotoxins with different modes of actions will throw light on the pathogenesis of motor neuron diseases. MATERIALS AND METHODS Eighty Wistar strain rats of both sexes (body wt. 140-200 g) were used. Under intraperitoneal sodium pentobarbital anesthesia (40 mg/kg body wt.), the sciatic nerve was exposed posterior to the hip joint under an operating microscope and was freed from the adjacent soft tissue. Then, one of the ricinus communis agglutinins (either RCA 60 or RCA 120, EY Lab. Cal) was slowly injected intranenrally into the nerve with a microsyringe. The dose ranged from 2 to 8 #g (0.1 ~ in saline) for light- and electron-microscopic studies. A larger dose of 20-30/,g of RCA 120 (3 ~o in saline solution) was injected in a similar fashion for immunohistochemical demonstration of retrograde transport of the toxic lectin to neuronal soma. In several additional animals, the brachial plexus was surgically exposed and 3 ~ RCA 120 was intraneurally injected into the primary rami in order to demonstrate the retrograde transport of RCA into motor neurons of the cervical segments. In several animals in each series, the contralateral sciatic nerve was transected in order to compare the effect of simple nerve severence to that of lectin injection. For immunohistochemistry, animals injected with RCA 120 were transcardially perfused with 4 ~ paraformaldehyde in Millonig's phosphate buffer (4 ° C, pH 7.3) 18 h to 2 days after RCA injections. The L4_ 6 or C4_ 8 cord segments and corresponding dorsal root ganglia were dissected out and quenched in isopentane/liquid nitrogen. The frozen sections were cut at a thickness of 20/~m and mounted on gelatinized slides. These sections were first treated with 1~o H202 for 30 min, after which they were incubated with 1~o normal goat serum for 30 min. Then, they were incubated with the first antiserum, rabbit anti-RCA 120 antibody (EY Lab. Cal), diluted 600 × with PBS containing 0.3~o Triton-X, for 48 h in a moist chamber (4°C). Next, the second

329 antibody, biotinylated anti-rabbit IgG (goat, Vector Co. Cal) in PBS-Triton-X was applied for 40 rain. After rinsing, the slides were covered with avidin-biotin-peroxidase complex (Vector Co. Cal) in PB S-Triton-X for I h. The sections were then preincubated with a solution (200 ml) containing 100 nag of DAB, 5 ml of 1% CoG12and 4 ml of 1~o nickel ammonium sulfate in Millonig's phosphate buffer for 15 min, after which 0.6 ml of 3~o H202 was added and they were incubated for 6-10 rain. For the light microscopic study, 24 h to 3 months following RCA injection, animals were transcardially perfused with 10% neutral buffered formalin. The lumbar spinal cord segments (L4_6), with the roots and spinal ganglia attached were embedded in paraffm. The histological sections were stained with HE, by the Kltlver-Barrera and Bodian methods. Glial fibrillary acidic protein immunostainings were also utilized. For ultrastructural studies, animals were perfused with a 4~o paraformaldehyde, 0.5 ~o glutaraldehyde mixture in Millonig's phosphate buffer (4 oC, pH 7.3). The lumbar cord and dorsal root ganglia were dissected out and postfixed with 1~o OsOa. After dehydration with a series of graded ethanol, the tissue was embedded in epoxy resin. Ultrathin sections were doubly stained with uranyl acetate and lead citrate, and observed with a Hitachi H-600 electron microscope. RESULTS

(A) Immunohistochemical demonstration of RCA in neuronal soma RCA-immunoreactive neurons were found in the dorsal root ganglion cells of the L4_6 cord segments as early as 19 h after RCA injection into the sciatic nerve. After injection of RCA into the brachial plexus, the ipsilateral C4_8 spinal ganglia contained many RCA-immunoreactive ganglion cells. In these neurons reaction products were dispersed in the perikaryon as f'me granules, whereas the nucleus was free of immunoreactivity. Although somatic motor neurons in the lumbar ventral horn remained negative for RCA, many RCA-immunoreactive neurons could already be identified in the ipsilateral C4_8 ventral horn in the brachial plexus injection cases (Fig. 1). Between 24 h and 2 days after RCA injection into the sciatic nerve trunk, the ipsilateral ventral horn revealed a small number of immunoreactive somatic motor neurons, several per case, but the reaction was not as intense as that of cervical cases. There were no immunoreactive glial cells within the spinal cord, indicating that the retrograde uptake was achieved solely by neurons. The distribution of immunoreactive neurons was confined to the motor neuron cell columns of sciatic or brachial nerve efferents as demonstrated by the HRP retrograde tracer method in our preliminary experiments. The contralateral dorsal root ganglia or the ventral horn were free of immunoreactive neurons. (11) Histological observations The f'n'st morphological changes of somatic motor neurons were detected within 2 days. These were found exclusively in areas of the ventral horn which were known to be the territory of the sciatic nerve efferent neurons (L4_6). These cells, chiefly located within the ventrolateral and dorsolateral motor neuron columns, showed subtle but

330

Fig. 1. a: Several anterior horn cells at the C7 spinal cord level were immunoreactive for RCA 24 h after a intraneural microinjection into the brachial plexus. × 35. b: Reaction products are granular in appearance and restricted to perikarya, sparing the nucleus, The avidin-biotin complex method with the primary antibody against RCA 120 was used for indirect immunohistocbemistry. × 80.

Fig. 2. a: Motor neurons at the L 5 cord level undergoing degeneration, 48 h after R C A microinjection into the sciatic nerve. Perikarya of these motor neurons are diffusely chromatolytic and somewhat swollen, but the nucleus remains unchanged. No alteration is seen in such supporting elements as elias and blood vessels. × 140. b: Three days after R C A treatment, motor neurons appear quite pale due to the loss of most of their Nissl substance as if they are 'albino' neurons. Notice clumpings of Nissl-like substances in perikarya. Toluidine blue staining, x 450. c: Normal toluidine-blue stained motor neurons as a comparison, × 450.

331 diffuse discoloration of the Nissl substance (Fig. 2a). However, the nucleus and nucleolus appeared unaffected. In 3 days, however, Nissl substance within the motor neurons became almost indiscernible, leaving a bleached and somewhat swollen cytoplasmic profile. This 'albino'-like change of neurons was most striking with toluidineblue stained semithin epon-embedded sections (Fig. 2b and c). At this stage the nucleolus of motor neurons appeared somewhat smaller and also discolored. The supporting elements, such as glias and capillaries, remained unaffected. The L4_6 dorsal root ganglia of the injected side contained many necrotic neurons and empty baskets. Within 1 week, many neurons giving rise to sciatic nerve efferents had rapidly dissolved, and the remaining ones were extremely discolored. Profiles of fragmented neuronal cell bodies and their dendrites were often observed. Neuronophagia, however, was not encountered. Astrocytes appeared to be slightly increased in number. The ipsilateral posterior column contained ascending axons undergoing Wallerian degeneration due to a loss of the ipsilateral dorsal root ganglia. After 2 weeks, motor neurons in the sciatic motor areas had been almost totally wiped out, leaving mild gliosis (Fig. 3a). In contrast, small and medium-sized neurons in and around the depopulated motoneuronal columns, possibly of internuncial type, appeared preserved (Fig. 3b). The axonal swellings were not observed. The contralateral ventral horn appeared unaffected; the motor neurons showed mild central chromatolysis as seen in cases with sciatic nerve transection, but neuronal loss was not evident.

(C) Electron microscopic observations Ultrastructural changes in motor neurons of the RCA-injected side first became discernible within 48 h. Clusters of ERs began to disintegrate into small blocks or pieces and scatter in a disorderly fashion (Fig. 4a) and polyribosomes were dissociated. In more severely affected neurons mono- and polyribosomes thus detached were randomly distributed in the perikaryal cytoplasm (Fig. 4b). Some but not all mitochondria also showed swelling with loss of their cristae. There were no apparent changes, however, in the nucleus or nucleolus. Seventy-two hours later, motor neurons underwent profound degenerative changes. An extreme degree of ER disintegration was noted whereby the cisternal walls were often left totally devoid of polyribosomes and fragmented into pieces (Fig. 4c). In some areas ribosomes were clumped together, forming masses of f'mely granular material (Fig. 4d). At the same time, Golgi cisterns appeared to dilate in an irregular fashion, giving the impression of vacuole formations. Perikarya appeared to contain an increased number of neurofilaments but neurotubuli were only rarely encountered. Mitochondria were swollen and often disrupted. The nucleus contained coarse flocculent material and clumps of heterochromatins. Interesting at this stage of degeneration was the dissociation of axo-somatic synapses in that the postsynaptic membrane with its characteristic density remained identifiable while terminal bouton to motor neuronal soma had been detached (Fig. 5). There appeared to be thin processes of astrocytes interposed between the bouton and motor neuronal cell membrane. Degenerative changes in the glial cells or vascular endothelium were not, however, encountered.

332

m

8 .l

| i'

~tt

4 o ~11

J

•

Fig. 3. a: The L 5 spinal cord level 2 weeks after RCA microinjection into the sciatic nerve. The anterior horn of the right side is deprived of large somatic motor neurons except for a few remaining ones. The contralateral side is, however, normal-looking, b: A larger magnification of the anterior horn area (bottom arrow) of a. The anterior horn is almost totally deprived of motor neurons, accompanied by mild glial proliferations. No apparent round cell infiltrations or perivascular cuffings were seen. Notice that small neurons probably of internuncial character still remain unaffected, x 250. c: Notice also that the dorsolateral part of the ipsilateral posterior column (the area pointed by the arrow of a) contains degenerating fibers due to degeneration of dorsal root ganglion cells, x 125.

333 DISCUSSION The motor neuron degeneration by way of retrograde transport of toxic lectins observed here confirmed earlier reports that neurons may incorporate via this route various substances, some of which could even be noxious to them (Harper et al. 1980; Wiley et al. 1982). RCA, one of the most potent biological toxins, acts by interfering with protein synthesis: inhibition of 60s ribosomal functions (Olsnes et al. 1974). The toxicity of RCA is not, however, specific to motor neurons. In previous papers, we demonstrated that sensory ganglion cells were equally vulnerable to this toxin when similarly injected into the sensory or mixed peripheral nerve (Yamamoto et al. 1983). Nevertheless, the motor neuronal degeneration observed here is of particular significance since by this pathway exogenous substances may reach motor neurons, circumventing the blood-brain barrier. In other words, it is only by this route that many substances may gain access to motor neurons and hence to the CNS, even when the brain is not accessible to them (Broadwell and Brightman 1976). This mode of entry to the CNS has long been known to be present in several viral infections such as rabies and herpes simplex (Kristensson et al. 1978), and is also implicated in the symptomatologies of certain bacterial infections, e.g. dramatic motor manifestations of tetanus (Dimpfel 1980). Harper et al. (1980) first demonstrated that autonomic postganglionic neurons may take up, by way of retrograde intraaxonal transport, substances which may be noxious to them and ultimately kill themselves, and the term "suicide transport" was coined for this phenomenon (Wiley et al. 1982). Doxorubicin, an antineoplastic antibiotic, may be effectivelytransported to somatic motor neurons and dorsal root ganglion cells when injected into the sciatic nerve (Yamamoto et al. 1984) or into the muscles (Bigotte and Olsson 1982, 1983). The target motor neurons start to degenerate within a week. In contrast to RCA, doxorubicin acts on double-helicoidal DNA strands by intercalation and hence impairs cellular RNA synthesis. However, degenerative sequences observed in RCA cases were quite distinct from those caused by doxorubicin. In particular, rapid, diffuse chromatolytic change is striking and unique to RCA, whereas bizarre nuclear heterochromatinization is an early event in doxorubicininduced motor neuronal degeneration (Yamamoto et al. 1984). These contrasting findings certainly reflect the difference in mechanism of action of these two substances. The pathology of motor neuron diseases includes various degenerative profiles; e.g. simple atrophy, central and diffuse chromatolysis, spheroid formations, neurofibrillary accumulation, Bunina bodies and intracytoplasmic hyaline inclusions (Iwata and Hirano 1979). To date, however, none of these observations can be linked to the pathogenesis of this disorder. The experimentally induced motor neuron degeneration shown here is certainly reminiscent of this particular disease and would be of importance in disclosing evidence of the primary death of motor neurons without affecting supporting cells when noxae is introduced through the intraaxonal route. This aspect of pathology is in accordance with observations of human motor neuron disease in which non-motor neuronal components of the anterior horn are preserved, whereas in acute poliomyelitis even small neurons in the anterior horn are wiped out (Iwata and

i

L~ 4~

335

Fig. 4. Ultrastruetures of motor neurons undergoing degeneration, a: Very early stage of degeneration, 24 h following RCA injection into the sciatic nerve. The neuronal perikaryon appears somewhat electron-lucent and granular ERs have disintegrated into small blocks and pieces. Also, manY but not all mitochondria are swollen. The nucleus and its nucleolus, however, appear unremarkable, x 5 000. b-d: Progressive degradation of neuronal pedkaryon, b: 48 h al~er RCA treatment, the perikaryon contains free floating monoribosomes, dispersed fragments of ERs, swollen, lucent mitochondria and empty vacuoles. Notice that boutons attached to the soma, however, appear unaffected (arrows). x 12800. c: Three days after RCA treatment. ER fragments appear denuded due to total loss of their ribosomes, which are now clumping together, x 12000. d: Clumpings of ribosomes now form large masses. Vacuolations possibly by swollen Golgi apparatus and mitochondria are identified. Neurofilaments appear increased in number at this stage. The nucleus contains clumps of heterochromatin and diffuse chromatin also appears flocculent, x 9000.

Hirano 1979). Involvement of the corticospinal tract was not observed in this experimental form. There could be several explanations for this. Transsynaptic transfer of toxic lectin may not occur with a magnitude sufficient to degenerate the upper motor neurons, or phylogenetic differences in the corticospinal tract, the termination of which in the rat is not directly onto motor neurons but rather onto internuncials, may be the cause (Brown 1971). There have been many hypotheses for the etiology and pathogenesis of motor neuron diseases (Bradley and Krasin 1982; Rowland 1984). None of these, however, explains with satisfaction the selective vulnerability of motor neurons in this particular disease. Therefore, the recently proposed theory by Appel (1981; Smith and Appel 1983) regarding the deficiency of motor neuron trophic hormone is an attractive one. Here, the presence of motor neuron-specific trophic substances produced in striated muscle or in the neuromuscular junction and retrogradely transported to motor neuronal soma is proposed. The roles played by the retrograde axoplasmic transport are much less understood than those of orthograde flow. Nerve growth factor is one of those substances known to be transported in a retrograde fashion (Levi-Montalcini 1982;

336

Fig. 5. Detachment of axosomatic synapses from soma. In the advanced stage of neuronal degeneration (3 days after RCA treatment), boutons having formed axosomatic synapses detach from the somatic membrane (arrows). Notice that postsynaptic density and subsurface cistern are still identifiable. Between the separated synapses, thin sheets of astrocytic processes intervene, x 9200. Barde et al. 1983). The present study together with other reports implies that some neurotoxic agents m a y also be taken up in the peripheral nerve (Baruah et al. 1981) and ultimately destroy the parental neurons ( H a r p e r et al. 1980; Wiley et al. 1982; Y a m a m o t o et al. 1983, 1984). W e hypothesize that if there is an exogenous neurotoxin with specific affinity to somatic m o t o r axonal m e m b r a n e s and with the ability to flow with retrograde intraaxonal stream, it could be a c a n d i d a t e for the etiology o f m o t o r neuron diseases. Investigation of such substances a p p e a r s to be one o f indispensible a p p r o a c h e s for the elucidation o f the pathogenesis o f these diseases. REFERENCES Appel, S.H. (1981) Unifying hypothesis for the cause of amyotrophic lateral sclerosis, Parkinsonism, and Alzheimer disease, Ann. Neurol., 6: 499-505. Barde, Y.-A., D. Edgar and H. Thoenen (1983) New neurotrophic factors, Ann. Rev. Physiol., 45: 601-612. Baruah, J.K., Rasool, C.G., Bradley, W.G. and T.L. Munsat (1981) Retrograde axonal transport of lead in rat sciatic nerve, Neurology (NIT), 31: 612-616. Bigotte, L. and Y. Olsson (1982) Retrograde transport of doxorubicin (Adriamycin) in peripheral nerves of mice. Neurosci. Lett.. 32: 217-221.

337 Bigotte, L. and Y. Olsson (1983) Cytotoxic effects of Adriamycin on mouse hypoglossal neurons following retrograde axonal transport from the tongue, Acta Neuropath. (Bed.), 61: 161-168. Bradley, W.G. and F.K. Krasin (1982) A new hypothesis of the etiology of amyotrophic lateral sclerosis - - The DNA hypothesis, Arch. Neurol. (Chic.), 39: 677-680. Broadwell, R. D. and M.W. Brightman (1976) Entry ofperoxidase into neurons of the central and peripheral nervous systems from extracerebral and cerebral blood, J. Comp. Neurol., 166: 257-284. Brown, L T. (197 i) Projections and termination of the corticospinal tract in rodents, Exp. Brain Res., 13: 432-450. Dimpfel, W. (1980) Pathogenic actions of tetanus toxin, Trends Neurosci., 3: 80-82. Dumas, M., M.E. Schwab and H. Thoenen (1979) Retrograde axonal transport of specific macromolecules as a tool for characterizing nerve terminal membranes, J. Neurobiol., 10: 179-197. Harper, C. G., J. O. Gonatas, T. Mizutani and N. K. Gonatas (1980) Retrograde transport and effects of toxic ricin in the autonomic nervous system, Lab. Invest., 42: 296-404. Iwata, M. and A. Hirano (1979) Current problems in the pathology of amyotrophic lateral sclerosis. In: H. Zimmerman (Ed.), Progress in Neuropathology, Vol. 4, Raven Press, New York, pp. 277-298. Kristensson, K., A. Yahlne, L.A. Persson and E. Lyche (1978) Neural spread of herpes simplex virus type 1 and 2 in mice after corneal or subcutaneous (footpad) inoculation, J. Neurol. Sci., 35: 331-340. Lasek, R.J. (1980) Axonal transport: A dynamic view of neuronal structures, Trends Neurosci., 3: 87-91. Levi-Montalcini, R. (1982) Developmental neurobiology and the natural history of nerve growth factor, Ann. Rev. Neurosci., 5: 341-362. Olsnes, S., K. Refsnes and A. Pihl (1974) Mechanisms of action of the toxic lectins abrin and ricin, Nature (Lond.), 249: 627-631. Rowland, L.P. (1984) Motor neuron diseases and amyotrophic lateral sclerosis, Trends Neurosci., 7: 110-112. Smith, R. G. and S.H. Appel (1983) Extracts of skeletal muscle increase neurite outgrowth and cholinergic activity of fetal spinal motor neurons, Science, 219: 1079-1081. Yamamoto, T., Y. Iwasaki and H. Konno (1983) Experimental sensory ganglionectomy by way of suicide axoplasmic transport, J. Neurosurg., 60:108-114. Yamamoto, T., Y. Iwasaki and H. Konno (1984) Retrograde axoplasmic transport of adriamycin - - An experimental form of motor neuron disease? Neurology (Cleveland), 34: 1299-1304. Wiley, R. G., W. W. Blessing and D. J. Reis (I 982) Suicide transport-- Destruction of neurons by retrograde transport of ricin, abrin and modeccin, Science, 216: 889-890.