Acta Neuropathologica
Acta Neuropathol. (Berl.) 46, 1- 10 (1979)
9 Springer-Verlag1979
Original Works
Early Events in Canine Distemper Demyelinating Encephalomyelitis B. A. Summers, H. A. Greisen, and M. J. G. Appel Department of Pathology and James A. Baker Institute for Animal Health, New York State Collegeof Veterinary Medicine, Cornell University, Ithaca, NY 14853, USA
Summary. The early neuropathological development of demyelinating Canine Distemper Encephalomyelitis (CDE) was studied in SPF dogs. Neural tissues were examined up to 30 days post infection (PI). Three phases of activity were observed. The primary event (first observed 8 days PI) was a nonsuppurative encephalomyelitis associated with the initiation of central nervous system (CNS) infection by virus-laden lymphocytes. At 24 days PI noninflammatory demyelination occurred in well defined, subependymal loci. Cell fusion and syncytia formation accompanied this early demyelination. The third phase, found at day 30 PI in one dog showing signs of recovery, was a second wave of nonsuppurative inflammation. The initial encephalomyelitis was widely disseminated throughout the CNS but subsequent demyelination appeared to be initiated from within the ventricular system. Myelin was phagocytosed by endogeneous CNS macrophages often infected with Canine Distemper Virus (CDV). The possible importance of viral induced cell fusion as welt as immune factors in the mechanism of demyelination are discussed.
Key words: Dog -
Canine distemper virus - Cell fusion - Demyelination - Encephalomyelitis Immunofluorescence - Electron microscopy
(MS) and Subacute Sclerosing Panencephalitis (SSPE) (Koestner, 1975). O f particular interest in C D E is the mechanism of demyelination. As for many of the h u m a n demyelinating diseases, two possibilities are entertained - either a direct effect or an immune mediated mechanism. Although several authors have identified CDV within lesions of C D E (Appel, 1969; Wisniewski et al., 1972) its role in demyelination is unclear. Recently Vandevelde and Kristensen (1977), in an immunofluorescent study of C D E concluded that proportionately little viral antigen was present in areas showing demyelination. Immunological mechanisms have been supported by the studies of K r a k o w k a et al. (1973) who demonstrated that sera from dogs with C D E contain antimyelin antibody. Much of the data pertaining to C D E has been obtained from spontaneous, clinical disease or terminal experimental infection. We have performed a sequential investigation following experimental exposure to a biotype of CDV designated Cornell A75-17 which had been isolated from a field case of CDE. We were particularly interested in the temporal relationsl~ips of viral entry, inflammation, and demyelination in the CNS.
Materials and Methods Canine Distemper (CD) is a worldwide disease of Canidae caused by a paramyxovirus (Genus: Morbillivirus) closely related to Measles virus of man and to Rinderpest virus of cattle (Appel and Gillespie, 1972; Imagawa, 1968). The virus first infects lymphoid organs and then spreads to epithelial and nervous tissue (Appel, 1969). The nonsuppurative demyelinating component of C D is of comparative medical interest and has been proposed as a model of Multiple Sclerosis Offprint requests to: B. A. Summers, M.D. (address see above)
Twelve specificpathogen-free beagle dogs between 10 and 12 weeks of age were inoculated intranasally with a standard dose of 1 x 10-~ TCtDso of CDV strain Cornelt A75-17. Dogs were euthanized, in pairs, 8,10,15,20,24, and 30 days PI by deep barbiturate anesthesia and then exsanguinated from the femoral artery. The CNS was removed and portions were either fixed in 10 % buffered formalin for histopathology or frozen for immunofluorescence (IF) or immersion fixed in glutaraldehyde for electron microscopy(EM). One dog (24 d PI) was deeply anesthetized with heparinized thiopentone (500 units heparin) and perfused with 41 of 1% Karnovsky's fixative at a pressure of 110 mm Hg. Central nervoussystemtissue fixed in 10 % buffered formalin was routinely processed for light microscopy and stained with ]~ematoxylin and eosin. Sections were taken just rostral to the corpus callosum, through the intertbalamic adhesion, at the rostral aspect of
0001-6322/79/0046/0001/$2.00
2 the rostral colliculus, obliquely through the optic cortex and medially through the cerebellum and medulla. Transverse and longitudinal sections of the spinal cord were taken from the cervical intumescence. For IF, sections of cerebrum (2), midbrain, cerebellum and cervical spinal cord were cut on a cryostat at 20~C and mounted on untreated glass slides. Sections were fixed in acetone for I0 rain and air dried. Fluorescein-labeled anti-CDV conjugate with auraminerhodamine counterstain was applied for 20 min in a moisture chamber at room temperature followed by 20 min washing in phosphate buffered saline (PBS). Slides were then dipped in distilled water, mounted in buffered glycerol and examined. For EM, specimens were taken from the cerebrum, midbrain, cerebellum and cervical spinal cord. Tissues were prepared by conventional methods. Most samples were fixed in 3 ~ glutaraldehyde in cacodylate buffer. Tissues from the perfused dog were removed and placed in Karnovsky's fixative for an additional 3 h. All specimens were post-fixed in 1 ~ OsO4 in cacodylate buffer, treated in 0.5 ~ uranyl acetate, dehydrated in ethanol and propylene oxide and embedded in Epon-Araldite. Sections l ~t thick were stained with toluidine blue for light microscopy. Areas selected for thin sectioning were cut with a diamond knife, stained with uranyl acetate and lead citrate (Reynolds 1963) and viewed with a Philips EM-300. Tissues from a normal, age matched dog served as controls for all procedures.
Results
Clinicopathological Findings Following inoculation, dogs showed a classical biphasic temperature response starting 4 - 5 days PI. Concomitantly a circulating lymphopenia developed and there was a severe depression of PHA blastogenesis (details not reported here). By day 10 weight loss was evident and dogs showed clinical depression, this being variable from day to day. Mucoid ocular and nasal discharge was common. By PI day 30, one dog had begun to show evidence of recovery, characterized by an improving attitude, elevation of lymphocyte counts, and some PHA blastogenesis activity. None of the animals showed clinical evidence of CNS disease.
Gross and Microscopic Pathology No gross pathological changes were observed in the CNS. Histopathological findings in dogs euthanized 8 days PI were of occasional to moderately frequent lymphocytic cuffs 1 - 2 cells thick around capillaries and venules (Fig. l a). Inflammatory changes were commonly found in one dog but were more sparse in the other. Cuffing was found in gray and white matter areas at all levels of the CNS examined. A few meningeal vessels had cuffs also. Sometimes focal gliosis was observed in the neuropil in the vicinity of cuffed vessels and some "glial nodules" contained lymphocytes derived from cuffs. Occasional parenchymal vessels showed a, thickening and hypercellularity of their adventitia generally mixed with inflammatory cells. At 10 days PI findings were similar but more pronounced. In both dogs relatively more capillaries and venules
Acta Neuropathol. (Berl.) 46 (1979) showed cuffing, sometimes reaching 3 cells thick. Occasionally, lymphocytes were seen migrating from cuffed vessels into the adjacent parenchyma. By 15 days inflammatory changes were declining. L y m p h o c y t i c cuffs and glial foci were now infrequently found. A new observation was of sporadic pale eosinophilic intracytoplasmic viral inclusions in ependymal cells. At 20 days PI cuffing and gliosis were virtually absent; rarely a vessel with proliferative changes could be found. Sporadic ependymal cells contained intracytoplasmic inclusions. The medial vestibular nucleus revealed vacuolation of the neuropil and m a n y chromatolytic neurones which contained intranuclear and intracytoplasmic eosinophilic inclusions, the latter often being arranged peripherally in the somata (Fig. 1 b). At 24 days PI findings were o f demyelination in the hippocampal fornix, caudal commissure, optic tract, rostral medullary velum and cerebellar white matter. Demyelination was characterized by myelin pallor and vesiculation (Fig. 1 c) and was accompanied by hypert r o p h y and proliferation o f subependymal glia, mainly astroglia. In the medullary velum large multinucleated astroglial syncytia were formed and myelin loss was severe (Fig. 1 d and e). The ventricular lining showed some h y p e r t r o p h y and apparent fusion o f ependymal cells overlying areas of demyelination. Occasional hypertrophic astroglia were f o u n d in normal periventricular white matter. In contrast to the cerebrum and brain stem, demyelination in the spinal cord was sparse. Generally at 24 days PI demyelination was not accompanied by an influx o f hematogeneous inflamm a t o r y cells. However, rare cuffed vessels and associated small glial loci were found elsewhere in the C N S at this time, for example, in the thalamus and spinal cord. By 30 days demyelination was more m a r k e d and followed the distribution outlined before. Small foci o f demyelination were also found within the cerebellar folia, sometimes adjacent to the granule cell layer. In one dog h e m a t o g e n o u s inflammation was lacking whereas in the other, which was clinically recovering, diffuse perivascular cuffs o f lymphocytes and plasma cells were found, occasionally forming thick collars, most p r o n o u n c e d in the cerebellum (Fig. If). The extent o f demyelination in b o t h dogs was comparable. Intracytoplasmic inclusion bodies in ependymal cells were c o m m o n .
Immunofluorescence Eight days PI, I F was m a r k e d in one dog and mild in the other. Positive areas were found in all levels o f the
B, A. Summers et al.: Early Events ill Canine Distemper Demyelinating Encephalomyelitis
3
Fig. 1 a--f. Light microscopy, a Cerebral white matter. Perivascular lymphocytic cuffing. PI day 8. H & E, • 455. h Medial vestibular nucleus. Neuron contains a single intranuclear and many intracytoplasmic inclusions. PI day 20. H & E, x 727. e Hippocampus. Subependymal zone of demyelination delineated by arrows. PI day 24. H & E, x 121. d Rostral medullary velum. Diffuse demyelination and astrogliosis. PI day 24. H & E, x 182. e Rostral medullary velum from Id. Syncytial astroglial cells within the area of early demyelination. PI day 24. H & E, x 300. f Cerebellar medulla. Extensive demyelination, cuffing, and diffuse gliosis. PI day 30. H & E, x 152
4
Acta NeuropathoI. (Berl.) 46 (1979)
Fig.2 a--d. Fluorescent microscopy. a Brain stem. Fluorescence in a perivascula: cuff. PI day 10, x 556, b Cerebellum. Fluorescence in the cerebellar cortex, predominately granular cell layer. PI day 30, x 172. e Cerebellum. White matter fluorescence in a folium. Granular cell layer is at upper left. PI day 30, x 354. d Spinal cord. Fluorescence of ependymal cells (central canal C) and periependymal cells. PI day 30, x354
neuraxis. M o s t fluorescence was f o u n d in perivascular m o n o n u c l e a r cells (Fig. 2a); occasionally positive cells were f o u n d within the vascular lumen or in the vessel wall. Perivascular I F was seen as a finely granular to globular deposit within the cytoplasm of m o n o n u c l e a r cells. In the meninges it sometimes was more granular and not cell associated, following the c o n t o u r o f meningeal membranes. Fluorescence o f individual cells within the gray or white matter was sporadic and resembled that seen in a perivascular location. By 10 days fluorescence was heavier but had the same distribution overall, being mostly related to the vasculature. In one dog a striking finding was a nest o f meningeal m o n o n u c l e a r cells showing I F which could
be seen extending into the adjacent brainstem, forming a wedge-shaped infiltrate with its base on the pia. Sporadic positive cells were seen in the neuroparenc h y m a below the pia in other sections also. A few o f these latter cells were larger than the free, perivascular m o n o n u c l e a r cells suggesting that there was some neuronal infection by this time. Other findings were a very light granular fluorescence of the granule layer o f the cerebellar cortex and an occasional positive ependymal cell in the spinal cord. A t 15 days, I F was still present but fewer vessels were cuffed with fluorescing m o n o n u c l e a r cells. A n occasional parenchymal cell suggestive o f a neurone, and short linear deposits in white matter tracts were seen.
B. A. Summers et al. : Early Events in Canine Distemper Demyelina6ng Encephalomyelitis
5
Fig.3 a--c. Electron microscopy (EM). a Brain stem. Lymphocyte from a perivascular cuff contains two loci of CDV nucleocapsids (arrows). PI day 8, x 11,900. b Cerebellum. Capillary endothelial cell contains an aggregate of tubuloreticular inclusions. PI day 8, • 23,600. e Cerebellum. Glial process contains CDV nucleocapsids. PI day 20, x 23,600
Sometimes points of fluorescence within adventitia in the vessel wall were evident. Fluorescence in the cerebellar cortex was light. At 20 days findings did not differ appreciably apart from the occasional positive cell in gray or white matter. The latter, in short rows, presumably reflected interfascicular oligodendroglia. More meningeal fluorescence was seen. At 24 days, IF studies were performed on the dog which was processed routinely; the perfused dog was
not satisfactory for IF. Fluorescence of cells in gray and white matter occurred in sporadic foci but only the granular layer of the cerebellum could be identified specifically with any confidence. Light perivascular fluorescence was still present. Several vessels showed fluorescence within their walls some of which could have been confused with endothelial fluorescence. By 30 days IF was more sporadic in a perivascular location and involved rather neurons, glia, and their processes. More profiles resembling neurons were seen
6
Acta NeuroDathol. (Berl.) 46 (1979)
with globular cytoplasmic areas of fluorescence. Sometimes clusters of fluorescing cells of neuronal size and smaller were seen in the gray matter. Very rarely cerebellar Purkinje cells fluoresced, but areas of strong granular layer fluorescence were consistently present (Fig. 2b). Light diffuse granular foci of fluorescence were seen within white fiber tracts at all levels, sometimes in a linear pattern; it was heaviest in the cerebellum (Fig. 2c) and spinal cord. Occasionally ependymal cells showed IF, most readily appreciated in the spinal cord (Fig. 2d).
Electron Microscopy At 8 days PI areas of lymphocytic cuffing were identified in 1 ~ sections from one dog. Corresponding thin sections revealed aggregates ofCDV nucleocapsids within the cytoplasm of perivascular lymphocytes (Fig. 3 a). One infected astrocyte foot process was identified. Very little cuffing was found in thick sections from the second dog as was anticipated from light microscopic observations. No evidence of virus was found. However, a focus of crystalline tubuloreticular inclusions in parallel array was observed in an endothelial cell (Fig. 3b). These inclusions were similar to, but distinct from, viral nucleocapsids on the basis of arrangement and location within rough endoplasmic reticulum. In both dogs at 10 days, perivascular lymphocytes containing nucleocapsid were identified in the brain, spinal cord, and choroid plexus. A few pericytes contained nucleocapsid; none was found in the endothelia. Occasionally tubuloreticular inclusions were found in endothelia. Material from dogs euthanized 15 days PI was not examined ultrastructurally. At 20 days findings were few, again corresponding with sparse light microscopic observations. Examinations of many thin sections revealed occasional aggregations of nucleocapsids within glial processes (Fig. 3c) often close to intact myelin sheaths. These processes had moderately electron-dense cytoplasm with long, stringy, rough endoplasmic reticulum, a few lysosomes and sometimes pseudopodia, as seen in microglia. Tubuloreticular inclusions, but not viral nucleocapsid, were found in endothelial cells and a few pericytes. At day 24 many areas of periventricular demyelination were found. Overlying areas of myelin loss, fusion of surface ependymal cells was observed (Fig. 4a), sometimes with remnants of plasmalemmae evident between fused cells. Some ependyma contained CDV nucleopcapsid. Multinucleated ependymal cells were found at or immediately below the ventricular surface. Deeper within the velum, multinucleated cells formed by fused astrocytes were found (Fig. 4b); they too were infected. Many of these large astroglial
Fig.4a and b. EM.a Rostralmedullaryvelum.Syncytiumof four fused ependymal cells. PI day 24, x 7,660. b Rostral medullary velum. Binucleate astroglial cell; one nucleus shows severe vacuolation. Astroglial processes containingbundles of filamentsare numerous. PI day 24, x 2,670
syncytia showed marked nuclear vacuolation. Within these areas of syncytia formation many demyelinated axons were found, sometimes with large macrophages containing myelin debris lying adjacent to the denuded axon (Fig. 5). Macrophages containing cytoplasmic nucleocapsids were found consistently in areas of demyelination which contained both intact and naked fibers. Astrocytic processes containing large bundles of cytofilaments were numerous; they sometimes were apposed to naked axons. Oligodendroglial cells were rarely encountered. The extracellular space was enlarged (edema) in both immersion and perfusion-fixed tissue. In areas of more extensive tissue necrosis ballooning of myelin occurred around swollen degenerate axons (spheroids) which contained numerous electron-dense bodies and mitochondria.
B. A. Summers et al.: Early Events in Canine Distemper Demyelinating Encephalomyelitis
7
Fig. 5. EM. Rostral medullary velum. A demyelinated axon (A) is bordered by macrophages (M) which contain engulfed myelin. Astroglial processes (As) lie apposed to the axon and throughout the tissue. One spheroid (S) - swollen axon - is seen. The extracellular space (E) is enlarged. PI day 24, x 6,650
At 30 days, demyelination was more widespread and large macrophages containing myelin remnants and CDV nucleocapsids were plentiful (Fig. 6). One phagocytic fibrous astrocyte with engulfed cytoplasmic myelin was found and infected astroglia were common. In areas of primary demyelination small electron dense ceils, probably microglia, containing nucleocapsids were seen often closely apposed to normal myelin sheaths (Fig. 7). Granule cells in the cerebellar cortex showed extensive cell necrosis and contained abundant cytoplasmic nucleocapsids. Tubuloreticular inclusions within the endothelia were less common.
Discussion
By performing sequential studies at known times after experimental infection, we were able to establish the sequence of early events in CDE for this biotype of CDV. Auxiliary studies performed in this investigation (virus isolation, serum antibody titration, lymphocyte blastogenesis, etc.) are not reported here. However, we would indicate that unlike SSPE in which condition the virus is defective and not readily retrieved (Meulen et al., 1972), CDV can be isolated from CDE by culture of washed, trypsinizcd cells. The initial events were inflammatory associated with the seeding of virus into the CNS. As this declined,
8
Acta Neuropathol. (Berl.) 46 (1979)
Fig. 6. EM. Cerebellum.Two macrophages (M) are seen. One on the right contains a large loop of myelin. CytoplasmicCDV nucleocapsid (NC) is evident. Adjacent myelin sheaths are intact. PI day 30, x 15,900
Fig. 7. EM. Cerebellum.A glial cell containing CDV nucleocapsid is closely apposed to a normal myelin sheath. PI day 30, x 13,400
probably as a result of extreme lymphopenia, findings in the CNS were few, apart from occasional aggregates of C D V nucleocapsid in glial processes and of one focus of neuronal infection. This brief quiescent period was followed by the initiation of demyelination in several constant areas, adjacent to the ventricular system. Thirdly, in one dog showing evidence of recovery, a further wave of mononuclear cell infiltration followed. The distribution of IF correlated with early lesion development. Initial histological findings were of widespread perivascular inflammation and most fluorescence was associated with mononuclear cells around meningeal and parenchymal blood vessels at 8 and 10 days PI. As the initial inflammatory phase declined, IF did likewise and became quite spotty. Later on (24
and 30 days P I ) s t r o n g IF was seen within the neuroparenchyma, both in gray and white matter areas. Some fluorescence of blood vessels and meningeal membranes was observed and electron microscopy revealed that pericytes, astrocyte processes, and meningeal fibroblasts did contain nucleocapsid. Nucleocapsid was not detected within endothelial cells although tubuloreticular structures, which have been described in a wide variety of inflammatory and neoplastic conditions (Grimley and Schaff, 1976) and said to be nonspecific for CD (Baringer, 1971; Blinzinger et al., 1972), were abundant. They were also seen in macrophages and astrocytes, although less commonly. Tubuloreticular inclusions were recognized by their localization within rough endoplasmic reticulum and their tendency to form crystalline arrange-
B. A. Summerset al.: Early Events in Canine Distemper DemyelinatingEncephalomyelitis ments (Fig. 3b). Nucleocapsids were not seen in cells containing tubular structures. Previous studies of CDE (McCullough et al., 1974; Raine, 1976) argued that the nonsuppurative inflammation follows myelin destruction. We identified an earlier cycle of widespread perivascular inflammatory changes, preceding demyelination, and associated with it the seeding of virus into the CNS. As we have recently reported (Summers et al., 1978) ultrastructural examination of the initial inflammatory lesions revealed virus infected perivascular lymphocytes throughout the CNS. Inflammatory changes and IF were mild in one dog 8 days PI and moderate in the other, suggesting that this is about the time of earliest CNS infection. Primary CDV replication occurs in lymphoid tissue (Appel, 1969) and it seems that dissemination to other tissues, including the CNS, occurs by infected lymphocytes, probably in a nonspecific fashion. Why virusinfected lymphocytes migrate from the bloodstream into the CNS and other tissues is not known. Probably alterations to plasma membranes of infected lymphocytes are important in changing normal trafficking of these cells. Prineas and Wright (1978) have suggested that there is a normal migration of small lymphocytes through the CNS. The sequestration of these infected lymphocytes throughout the body may explain in part the lymphopenia characteristic of CD. The early inflammatory changes were widespread whereas the earliest demyelination occurred in a limited number of fiber tracts close to the ventricular system. Hence one could argue that demyelination is initiated by some factor, possibly CDV, derived from within the CSF. It is possible that viral diffusion and spread is impeded in white matter and that the earliest exposure of myelinated fibers to virus comes from infected cells in the CSF. Fluorescing mononuclear cells have been demonstrated in the CSF in CD (Appel, 1969) and ependymal infection was common. The formation of hypertrophic and multinucleated astroglia and syncytial ependyma accompanied the earliest demyelination. Raine (1976) also identified astroglial giant cells in early lesions of CDE. Our findings confirm his contention that this giant cell formation is associated with the earliest demyelination in CDE; furthermore we identified occasional areas of astroglial swelling in normally myelinated tissue as has been described recently in acute MS (Prineas and Raine, 1976) suggesting that astroglial activity precedes myelin breakdown. The initial demyelination was largely noninflammatory; we did occasionally find small cuffs in other areas of the CNS at the stage of early myelin loss which were presumably remnants of the primary wave. The source of macrophages in areas of demyelination was closely sought. Significantly, many con-
9
rained cytoplasmic CDV nucleocapsids but little in the way of identifying nuclear or cytoplasmic characteristics. Their presence was not heralded by an infiltrate of blood monocytes and so we conclude that they are endogeneous to the CNS. Several possibilities exist and these are not mutually exclusive. Astrocytes were infected and have recently been implicated to, some extent in myelin removal in CDE (Raine, 1976). Rarely a macrophage was found with bundles of cytoplasmic filaments allowing characterization of that cell as an astrocyte. Pericytes were found to be infected and may become active and phagocytic under suitable circumstances (Baron and Gallego, 1972). Possibly resting microglia were involved also as some glial processes with nucleocapsid, identified at day 20, may have been microglial. Autophagy of myelin by oligodendroglia is another, although less likely, possibility. The mechanism of demyelination in CDE is unknown. Wallerian-type degeneration would not account for selective myelin loss and oligodendroglial infection has rarely been demonstrated (Campbell, 1957; Raine, 1976). Consequently, the possibility of immune mediated damage to myelin in CDE has received attention of late (Koestner et al., 1974). Krakowka et al. (1973) demonstrated specific antimyelin antibody in both spontaneous and experimental CDE, but whether this antibody is operative in vivo in the demyelinating process is yet to be demonstrated. Our studies of the early stages of demyelination incriminate CDV. We often found glial cells containing nucleocapsids closely apposed to intact myelin sheaths within areas of active demyelination. It appeared that glial cells became infected and subsequently effected demyelination. Some of the known viral agents which will produce demyelinating encephalomyelitis (CDV, visna in sheep, mouse corona, measles in SSPE) are syncytia-forming viruses and so cell fusion may be involved in the mechanism of demyelination, as was alluded to by Choppin (1968). We observed that fusion of infected ependymal cells and astroglia accompanied the early demyelination in CDE. In an earlier study of experimental SSPE in the hamster, Raine et al. (1974) found polykaryocytes, some apparently derived from different cell types within brain lesions. Myelin disruption in CDE may result from the fusion of glial cell membranes from infected astroglia or microglia with myelin sheaths. Following fusion of plasmalemmae, neutral proteases from activated glial macrophages could degrade myelin (Cammer et al., 1978) leading to its phagocytosis. The factor in CDV which induces cell fusion is probably contained within envelope glycoproteins (Fisher and Bussell, 1977). Neutralization of these viral antigens may inhibit demyelination because Krakowka et al. (1975) associated reduced levels of anti-envelope antibody with demyelination in CDE.
10 Several authors have proposed that demyelination in C D E c a n n o t be a s c r i b e d to viral effects a l o n e ( K o e s t n e r et al., 1974; V a n d e v e l d e a n d K r i s t e n s e n , 1977). F o l l o w i n g s e n s i t i z a t i o n to m y e l i n o r viral antigens, f u r t h e r p r o g r e s s i v e loss o f m y e l i n m a y result f r o m an i m m u n e m e d i a t e d p r o c e s s , e i t h e r h u m o r a l , cellular, o r b o t h . T h i s m a y be c l a r i f i e d it it c a n be e s t a b l i s h e d w h e t h e r l y m p h o i d cells e n t e r i n g the C N S f o l l o w i n g the p r i m a r y d e m y e l i n a t i o n are r e s p o n d i n g in a p r o t e c t i v e i m m u n e c a p a c i t y o r to p a r t i c i p a t e in a n d a m p l i f y the process of myelin destruction.
Acknowledgements. We thank Dr. J. Cummings for helpful discussions and Dr. L. Krook for assistance with photography. The assistance of Mary Beth Metzgar and Ann Signore is greatly appreciated. This work was supported by the Whitehall and Richard King Mellon Foundations.
References Appel, M. J. G. : Pathogenesis of canine distemper. Am. J. Vet. Res. 30, 1167 1182 (1969) Appel, M. J. G., Gillespie, J. H.: Canine distemper virus. Virology monographs. Vol. 11. Wien, New York: Springer 1972 Baringer, J. R.: Tubular aggregates in endoplasmic reticulum in herpes-simplex encephalitis. N. Engl. J. Med. 285, 943-945 (1971) Baron, M., Gallego, A.: The relation of the microglia with the pericytes in the cat cerbral cortex. Z. Zellforsch. 128, 4 2 - 5 7 (1972) Blinzinger, K., Anzil, A. P., Deutschlfinder, N. : Nature of tubular aggregates. N. Engl. J. Med. 286, 157-158 (1972) Camer, W., Bloom, B. R., Norton, W. T., Gordon, S. : Degradation of basic protein in myelin by neutral proteases secreted by stimulated macrophages. A possible mechanism of inflammatory demyelination. Proc. Natl. Acad. Sci. USA 75, 15541558 (1978) Campbell, R. S. F. : Encephalitis in canine distemper. Br. Vet. J. 113, 143 161 (1957) Choppin, P. W. : Pathogenic immune reactions in virus infection. In : Textbook of immunopathology, (eds. P. A. Miescher, A. J. Mfiller-Eberhard). VoL 1. New York: Grune and Stratton 1968 Fisher, L. A., Bussell, R. H.: Cell fusion by canine destemper virusinfected cells and their plasma membranes. Intervirology 8, 218 - 225 (1977) Grimley, P. M., Schaff, Z. : Significance of tubuloreticular inclusions in the pathobiology of human diseases. In : Pathobiology annual,
Acta Neuropathol. (Berl.) 46 (1979) (ed. H. k. Joachim). Vol. 6. New York: Appleton-CenturyCrofts 1976 Imagawa, D. T. : Relationships among measles, canine distemper and rinderpest viruses. Prog. Med. Virol. 10, 160-193 (1968) Koestner, A. : Animal model - distemper-associated demyelinating encephalomyelitis. Am. J. Pathol. 78, 361 364 (1975) Koestner, A., McCullough, B., Krakowka, G. S., Long, J. F., Olsen, R. G.: Canine distemper: a virus-induced demyelinating encephalomyelitis. In : Slow virus diseases (eds. W. Zeman, E. H. Lennette). pp. 86-101. Baltimore: Williams and Wilkins 1974 Krakowka, S., McCullough, B., Koestner, A., Olsen, R.: Myelinspecific autoantibodies associated with central nervous system demyelination in canine distemper virus infection. Infect. Immun. 8, 819-829 (1973) Krakowka, S., Olsen, R., Confer, A., Koestner, A., McCullough, B. : Serologic response to canine destemper viral antigens in gnotobiotic dogs infected with canine distemper virus. J. Infect. Dis. 132, 384-392 (1975) McCullough, B., Krakowka, S., Koestner, A. : Experimental canine distemper virus-induced demyelination. Lab. Invest. 31, 216 222 (1974) ter Meulen, V., Katz, M., Mfiller, D.: Subacute sclerosing panencephalitis: A review. Curr. Top. Microbiol. Immunol. 57, 1 - 3 8 (1972) Prineas, J. W., Raine, C. S. : Electron microscopy and immunoperoxidase studies of early multiple sclerosis lesions. Neurology (Minneap.) 26, 29-32 (1976) Prineas, J. W., Wright, R. G.: Macrophages, lymphocytes, and plasma cells in the perivascular compartment in chronic multiple sclerosis. Lab. Invest. 38, 409 421 (t978) Rathe, C. S. : On the development of CNS lesions in natural canine distemper encephalomyelitis. J. Neurol. Sci. 30, t3 28 (1976) Raine, C. S., Byington, D. P., Johnson, K. P. : Experimental subacute sclerosing panencephalitis in the hamster. Ultrastructure of the chronic disease. Lab. Invest. 31, 355- 368 (1974) Reynolds, E. S. : The use of lead citrate at high pH as an electronopaque stain in electron microscopy. J. Cell. Biol. 17, 208-212 (1963) Summers, B. A., Greisen, H. A., Appel, M. J. G. : Possible initiation of viral encephalomyelitis in dogs by migrating lymphocytes infected with distemper virus. Lancet II, 187 190 (1978) Vandevelde, M., Kristensen, B. : Observations on the distribution of canine distemper virus in the central nervous system of dogs with demyelinating encephalitis. Acta Neuropathol. (Berl.) 40, 233 236 (1977) Wisniewski, 14., Raine, C. S., Kay, W. J.: Observations on viral demyelinating encephalomyelitis. Canine distemper. Lab. Invest. 26, 589-599 (1972) Received July 5, 1978/Accepted December 19, 1978
Acta Neuropathol. (Berl.) 46, 1- 10 (1979)
9 Springer-Verlag1979
Original Works
Early Events in Canine Distemper Demyelinating Encephalomyelitis B. A. Summers, H. A. Greisen, and M. J. G. Appel Department of Pathology and James A. Baker Institute for Animal Health, New York State Collegeof Veterinary Medicine, Cornell University, Ithaca, NY 14853, USA
Summary. The early neuropathological development of demyelinating Canine Distemper Encephalomyelitis (CDE) was studied in SPF dogs. Neural tissues were examined up to 30 days post infection (PI). Three phases of activity were observed. The primary event (first observed 8 days PI) was a nonsuppurative encephalomyelitis associated with the initiation of central nervous system (CNS) infection by virus-laden lymphocytes. At 24 days PI noninflammatory demyelination occurred in well defined, subependymal loci. Cell fusion and syncytia formation accompanied this early demyelination. The third phase, found at day 30 PI in one dog showing signs of recovery, was a second wave of nonsuppurative inflammation. The initial encephalomyelitis was widely disseminated throughout the CNS but subsequent demyelination appeared to be initiated from within the ventricular system. Myelin was phagocytosed by endogeneous CNS macrophages often infected with Canine Distemper Virus (CDV). The possible importance of viral induced cell fusion as welt as immune factors in the mechanism of demyelination are discussed.
Key words: Dog -
Canine distemper virus - Cell fusion - Demyelination - Encephalomyelitis Immunofluorescence - Electron microscopy
(MS) and Subacute Sclerosing Panencephalitis (SSPE) (Koestner, 1975). O f particular interest in C D E is the mechanism of demyelination. As for many of the h u m a n demyelinating diseases, two possibilities are entertained - either a direct effect or an immune mediated mechanism. Although several authors have identified CDV within lesions of C D E (Appel, 1969; Wisniewski et al., 1972) its role in demyelination is unclear. Recently Vandevelde and Kristensen (1977), in an immunofluorescent study of C D E concluded that proportionately little viral antigen was present in areas showing demyelination. Immunological mechanisms have been supported by the studies of K r a k o w k a et al. (1973) who demonstrated that sera from dogs with C D E contain antimyelin antibody. Much of the data pertaining to C D E has been obtained from spontaneous, clinical disease or terminal experimental infection. We have performed a sequential investigation following experimental exposure to a biotype of CDV designated Cornell A75-17 which had been isolated from a field case of CDE. We were particularly interested in the temporal relationsl~ips of viral entry, inflammation, and demyelination in the CNS.
Materials and Methods Canine Distemper (CD) is a worldwide disease of Canidae caused by a paramyxovirus (Genus: Morbillivirus) closely related to Measles virus of man and to Rinderpest virus of cattle (Appel and Gillespie, 1972; Imagawa, 1968). The virus first infects lymphoid organs and then spreads to epithelial and nervous tissue (Appel, 1969). The nonsuppurative demyelinating component of C D is of comparative medical interest and has been proposed as a model of Multiple Sclerosis Offprint requests to: B. A. Summers, M.D. (address see above)
Twelve specificpathogen-free beagle dogs between 10 and 12 weeks of age were inoculated intranasally with a standard dose of 1 x 10-~ TCtDso of CDV strain Cornelt A75-17. Dogs were euthanized, in pairs, 8,10,15,20,24, and 30 days PI by deep barbiturate anesthesia and then exsanguinated from the femoral artery. The CNS was removed and portions were either fixed in 10 % buffered formalin for histopathology or frozen for immunofluorescence (IF) or immersion fixed in glutaraldehyde for electron microscopy(EM). One dog (24 d PI) was deeply anesthetized with heparinized thiopentone (500 units heparin) and perfused with 41 of 1% Karnovsky's fixative at a pressure of 110 mm Hg. Central nervoussystemtissue fixed in 10 % buffered formalin was routinely processed for light microscopy and stained with ]~ematoxylin and eosin. Sections were taken just rostral to the corpus callosum, through the intertbalamic adhesion, at the rostral aspect of
0001-6322/79/0046/0001/$2.00
2 the rostral colliculus, obliquely through the optic cortex and medially through the cerebellum and medulla. Transverse and longitudinal sections of the spinal cord were taken from the cervical intumescence. For IF, sections of cerebrum (2), midbrain, cerebellum and cervical spinal cord were cut on a cryostat at 20~C and mounted on untreated glass slides. Sections were fixed in acetone for I0 rain and air dried. Fluorescein-labeled anti-CDV conjugate with auraminerhodamine counterstain was applied for 20 min in a moisture chamber at room temperature followed by 20 min washing in phosphate buffered saline (PBS). Slides were then dipped in distilled water, mounted in buffered glycerol and examined. For EM, specimens were taken from the cerebrum, midbrain, cerebellum and cervical spinal cord. Tissues were prepared by conventional methods. Most samples were fixed in 3 ~ glutaraldehyde in cacodylate buffer. Tissues from the perfused dog were removed and placed in Karnovsky's fixative for an additional 3 h. All specimens were post-fixed in 1 ~ OsO4 in cacodylate buffer, treated in 0.5 ~ uranyl acetate, dehydrated in ethanol and propylene oxide and embedded in Epon-Araldite. Sections l ~t thick were stained with toluidine blue for light microscopy. Areas selected for thin sectioning were cut with a diamond knife, stained with uranyl acetate and lead citrate (Reynolds 1963) and viewed with a Philips EM-300. Tissues from a normal, age matched dog served as controls for all procedures.
Results
Clinicopathological Findings Following inoculation, dogs showed a classical biphasic temperature response starting 4 - 5 days PI. Concomitantly a circulating lymphopenia developed and there was a severe depression of PHA blastogenesis (details not reported here). By day 10 weight loss was evident and dogs showed clinical depression, this being variable from day to day. Mucoid ocular and nasal discharge was common. By PI day 30, one dog had begun to show evidence of recovery, characterized by an improving attitude, elevation of lymphocyte counts, and some PHA blastogenesis activity. None of the animals showed clinical evidence of CNS disease.
Gross and Microscopic Pathology No gross pathological changes were observed in the CNS. Histopathological findings in dogs euthanized 8 days PI were of occasional to moderately frequent lymphocytic cuffs 1 - 2 cells thick around capillaries and venules (Fig. l a). Inflammatory changes were commonly found in one dog but were more sparse in the other. Cuffing was found in gray and white matter areas at all levels of the CNS examined. A few meningeal vessels had cuffs also. Sometimes focal gliosis was observed in the neuropil in the vicinity of cuffed vessels and some "glial nodules" contained lymphocytes derived from cuffs. Occasional parenchymal vessels showed a, thickening and hypercellularity of their adventitia generally mixed with inflammatory cells. At 10 days PI findings were similar but more pronounced. In both dogs relatively more capillaries and venules
Acta Neuropathol. (Berl.) 46 (1979) showed cuffing, sometimes reaching 3 cells thick. Occasionally, lymphocytes were seen migrating from cuffed vessels into the adjacent parenchyma. By 15 days inflammatory changes were declining. L y m p h o c y t i c cuffs and glial foci were now infrequently found. A new observation was of sporadic pale eosinophilic intracytoplasmic viral inclusions in ependymal cells. At 20 days PI cuffing and gliosis were virtually absent; rarely a vessel with proliferative changes could be found. Sporadic ependymal cells contained intracytoplasmic inclusions. The medial vestibular nucleus revealed vacuolation of the neuropil and m a n y chromatolytic neurones which contained intranuclear and intracytoplasmic eosinophilic inclusions, the latter often being arranged peripherally in the somata (Fig. 1 b). At 24 days PI findings were o f demyelination in the hippocampal fornix, caudal commissure, optic tract, rostral medullary velum and cerebellar white matter. Demyelination was characterized by myelin pallor and vesiculation (Fig. 1 c) and was accompanied by hypert r o p h y and proliferation o f subependymal glia, mainly astroglia. In the medullary velum large multinucleated astroglial syncytia were formed and myelin loss was severe (Fig. 1 d and e). The ventricular lining showed some h y p e r t r o p h y and apparent fusion o f ependymal cells overlying areas of demyelination. Occasional hypertrophic astroglia were f o u n d in normal periventricular white matter. In contrast to the cerebrum and brain stem, demyelination in the spinal cord was sparse. Generally at 24 days PI demyelination was not accompanied by an influx o f hematogeneous inflamm a t o r y cells. However, rare cuffed vessels and associated small glial loci were found elsewhere in the C N S at this time, for example, in the thalamus and spinal cord. By 30 days demyelination was more m a r k e d and followed the distribution outlined before. Small foci o f demyelination were also found within the cerebellar folia, sometimes adjacent to the granule cell layer. In one dog h e m a t o g e n o u s inflammation was lacking whereas in the other, which was clinically recovering, diffuse perivascular cuffs o f lymphocytes and plasma cells were found, occasionally forming thick collars, most p r o n o u n c e d in the cerebellum (Fig. If). The extent o f demyelination in b o t h dogs was comparable. Intracytoplasmic inclusion bodies in ependymal cells were c o m m o n .
Immunofluorescence Eight days PI, I F was m a r k e d in one dog and mild in the other. Positive areas were found in all levels o f the
B, A. Summers et al.: Early Events ill Canine Distemper Demyelinating Encephalomyelitis
3
Fig. 1 a--f. Light microscopy, a Cerebral white matter. Perivascular lymphocytic cuffing. PI day 8. H & E, • 455. h Medial vestibular nucleus. Neuron contains a single intranuclear and many intracytoplasmic inclusions. PI day 20. H & E, x 727. e Hippocampus. Subependymal zone of demyelination delineated by arrows. PI day 24. H & E, x 121. d Rostral medullary velum. Diffuse demyelination and astrogliosis. PI day 24. H & E, x 182. e Rostral medullary velum from Id. Syncytial astroglial cells within the area of early demyelination. PI day 24. H & E, x 300. f Cerebellar medulla. Extensive demyelination, cuffing, and diffuse gliosis. PI day 30. H & E, x 152
4
Acta NeuropathoI. (Berl.) 46 (1979)
Fig.2 a--d. Fluorescent microscopy. a Brain stem. Fluorescence in a perivascula: cuff. PI day 10, x 556, b Cerebellum. Fluorescence in the cerebellar cortex, predominately granular cell layer. PI day 30, x 172. e Cerebellum. White matter fluorescence in a folium. Granular cell layer is at upper left. PI day 30, x 354. d Spinal cord. Fluorescence of ependymal cells (central canal C) and periependymal cells. PI day 30, x354
neuraxis. M o s t fluorescence was f o u n d in perivascular m o n o n u c l e a r cells (Fig. 2a); occasionally positive cells were f o u n d within the vascular lumen or in the vessel wall. Perivascular I F was seen as a finely granular to globular deposit within the cytoplasm of m o n o n u c l e a r cells. In the meninges it sometimes was more granular and not cell associated, following the c o n t o u r o f meningeal membranes. Fluorescence o f individual cells within the gray or white matter was sporadic and resembled that seen in a perivascular location. By 10 days fluorescence was heavier but had the same distribution overall, being mostly related to the vasculature. In one dog a striking finding was a nest o f meningeal m o n o n u c l e a r cells showing I F which could
be seen extending into the adjacent brainstem, forming a wedge-shaped infiltrate with its base on the pia. Sporadic positive cells were seen in the neuroparenc h y m a below the pia in other sections also. A few o f these latter cells were larger than the free, perivascular m o n o n u c l e a r cells suggesting that there was some neuronal infection by this time. Other findings were a very light granular fluorescence of the granule layer o f the cerebellar cortex and an occasional positive ependymal cell in the spinal cord. A t 15 days, I F was still present but fewer vessels were cuffed with fluorescing m o n o n u c l e a r cells. A n occasional parenchymal cell suggestive o f a neurone, and short linear deposits in white matter tracts were seen.
B. A. Summers et al. : Early Events in Canine Distemper Demyelina6ng Encephalomyelitis
5
Fig.3 a--c. Electron microscopy (EM). a Brain stem. Lymphocyte from a perivascular cuff contains two loci of CDV nucleocapsids (arrows). PI day 8, x 11,900. b Cerebellum. Capillary endothelial cell contains an aggregate of tubuloreticular inclusions. PI day 8, • 23,600. e Cerebellum. Glial process contains CDV nucleocapsids. PI day 20, x 23,600
Sometimes points of fluorescence within adventitia in the vessel wall were evident. Fluorescence in the cerebellar cortex was light. At 20 days findings did not differ appreciably apart from the occasional positive cell in gray or white matter. The latter, in short rows, presumably reflected interfascicular oligodendroglia. More meningeal fluorescence was seen. At 24 days, IF studies were performed on the dog which was processed routinely; the perfused dog was
not satisfactory for IF. Fluorescence of cells in gray and white matter occurred in sporadic foci but only the granular layer of the cerebellum could be identified specifically with any confidence. Light perivascular fluorescence was still present. Several vessels showed fluorescence within their walls some of which could have been confused with endothelial fluorescence. By 30 days IF was more sporadic in a perivascular location and involved rather neurons, glia, and their processes. More profiles resembling neurons were seen
6
Acta NeuroDathol. (Berl.) 46 (1979)
with globular cytoplasmic areas of fluorescence. Sometimes clusters of fluorescing cells of neuronal size and smaller were seen in the gray matter. Very rarely cerebellar Purkinje cells fluoresced, but areas of strong granular layer fluorescence were consistently present (Fig. 2b). Light diffuse granular foci of fluorescence were seen within white fiber tracts at all levels, sometimes in a linear pattern; it was heaviest in the cerebellum (Fig. 2c) and spinal cord. Occasionally ependymal cells showed IF, most readily appreciated in the spinal cord (Fig. 2d).
Electron Microscopy At 8 days PI areas of lymphocytic cuffing were identified in 1 ~ sections from one dog. Corresponding thin sections revealed aggregates ofCDV nucleocapsids within the cytoplasm of perivascular lymphocytes (Fig. 3 a). One infected astrocyte foot process was identified. Very little cuffing was found in thick sections from the second dog as was anticipated from light microscopic observations. No evidence of virus was found. However, a focus of crystalline tubuloreticular inclusions in parallel array was observed in an endothelial cell (Fig. 3b). These inclusions were similar to, but distinct from, viral nucleocapsids on the basis of arrangement and location within rough endoplasmic reticulum. In both dogs at 10 days, perivascular lymphocytes containing nucleocapsid were identified in the brain, spinal cord, and choroid plexus. A few pericytes contained nucleocapsid; none was found in the endothelia. Occasionally tubuloreticular inclusions were found in endothelia. Material from dogs euthanized 15 days PI was not examined ultrastructurally. At 20 days findings were few, again corresponding with sparse light microscopic observations. Examinations of many thin sections revealed occasional aggregations of nucleocapsids within glial processes (Fig. 3c) often close to intact myelin sheaths. These processes had moderately electron-dense cytoplasm with long, stringy, rough endoplasmic reticulum, a few lysosomes and sometimes pseudopodia, as seen in microglia. Tubuloreticular inclusions, but not viral nucleocapsid, were found in endothelial cells and a few pericytes. At day 24 many areas of periventricular demyelination were found. Overlying areas of myelin loss, fusion of surface ependymal cells was observed (Fig. 4a), sometimes with remnants of plasmalemmae evident between fused cells. Some ependyma contained CDV nucleopcapsid. Multinucleated ependymal cells were found at or immediately below the ventricular surface. Deeper within the velum, multinucleated cells formed by fused astrocytes were found (Fig. 4b); they too were infected. Many of these large astroglial
Fig.4a and b. EM.a Rostralmedullaryvelum.Syncytiumof four fused ependymal cells. PI day 24, x 7,660. b Rostral medullary velum. Binucleate astroglial cell; one nucleus shows severe vacuolation. Astroglial processes containingbundles of filamentsare numerous. PI day 24, x 2,670
syncytia showed marked nuclear vacuolation. Within these areas of syncytia formation many demyelinated axons were found, sometimes with large macrophages containing myelin debris lying adjacent to the denuded axon (Fig. 5). Macrophages containing cytoplasmic nucleocapsids were found consistently in areas of demyelination which contained both intact and naked fibers. Astrocytic processes containing large bundles of cytofilaments were numerous; they sometimes were apposed to naked axons. Oligodendroglial cells were rarely encountered. The extracellular space was enlarged (edema) in both immersion and perfusion-fixed tissue. In areas of more extensive tissue necrosis ballooning of myelin occurred around swollen degenerate axons (spheroids) which contained numerous electron-dense bodies and mitochondria.
B. A. Summers et al.: Early Events in Canine Distemper Demyelinating Encephalomyelitis
7
Fig. 5. EM. Rostral medullary velum. A demyelinated axon (A) is bordered by macrophages (M) which contain engulfed myelin. Astroglial processes (As) lie apposed to the axon and throughout the tissue. One spheroid (S) - swollen axon - is seen. The extracellular space (E) is enlarged. PI day 24, x 6,650
At 30 days, demyelination was more widespread and large macrophages containing myelin remnants and CDV nucleocapsids were plentiful (Fig. 6). One phagocytic fibrous astrocyte with engulfed cytoplasmic myelin was found and infected astroglia were common. In areas of primary demyelination small electron dense ceils, probably microglia, containing nucleocapsids were seen often closely apposed to normal myelin sheaths (Fig. 7). Granule cells in the cerebellar cortex showed extensive cell necrosis and contained abundant cytoplasmic nucleocapsids. Tubuloreticular inclusions within the endothelia were less common.
Discussion
By performing sequential studies at known times after experimental infection, we were able to establish the sequence of early events in CDE for this biotype of CDV. Auxiliary studies performed in this investigation (virus isolation, serum antibody titration, lymphocyte blastogenesis, etc.) are not reported here. However, we would indicate that unlike SSPE in which condition the virus is defective and not readily retrieved (Meulen et al., 1972), CDV can be isolated from CDE by culture of washed, trypsinizcd cells. The initial events were inflammatory associated with the seeding of virus into the CNS. As this declined,
8
Acta Neuropathol. (Berl.) 46 (1979)
Fig. 6. EM. Cerebellum.Two macrophages (M) are seen. One on the right contains a large loop of myelin. CytoplasmicCDV nucleocapsid (NC) is evident. Adjacent myelin sheaths are intact. PI day 30, x 15,900
Fig. 7. EM. Cerebellum.A glial cell containing CDV nucleocapsid is closely apposed to a normal myelin sheath. PI day 30, x 13,400
probably as a result of extreme lymphopenia, findings in the CNS were few, apart from occasional aggregates of C D V nucleocapsid in glial processes and of one focus of neuronal infection. This brief quiescent period was followed by the initiation of demyelination in several constant areas, adjacent to the ventricular system. Thirdly, in one dog showing evidence of recovery, a further wave of mononuclear cell infiltration followed. The distribution of IF correlated with early lesion development. Initial histological findings were of widespread perivascular inflammation and most fluorescence was associated with mononuclear cells around meningeal and parenchymal blood vessels at 8 and 10 days PI. As the initial inflammatory phase declined, IF did likewise and became quite spotty. Later on (24
and 30 days P I ) s t r o n g IF was seen within the neuroparenchyma, both in gray and white matter areas. Some fluorescence of blood vessels and meningeal membranes was observed and electron microscopy revealed that pericytes, astrocyte processes, and meningeal fibroblasts did contain nucleocapsid. Nucleocapsid was not detected within endothelial cells although tubuloreticular structures, which have been described in a wide variety of inflammatory and neoplastic conditions (Grimley and Schaff, 1976) and said to be nonspecific for CD (Baringer, 1971; Blinzinger et al., 1972), were abundant. They were also seen in macrophages and astrocytes, although less commonly. Tubuloreticular inclusions were recognized by their localization within rough endoplasmic reticulum and their tendency to form crystalline arrange-
B. A. Summerset al.: Early Events in Canine Distemper DemyelinatingEncephalomyelitis ments (Fig. 3b). Nucleocapsids were not seen in cells containing tubular structures. Previous studies of CDE (McCullough et al., 1974; Raine, 1976) argued that the nonsuppurative inflammation follows myelin destruction. We identified an earlier cycle of widespread perivascular inflammatory changes, preceding demyelination, and associated with it the seeding of virus into the CNS. As we have recently reported (Summers et al., 1978) ultrastructural examination of the initial inflammatory lesions revealed virus infected perivascular lymphocytes throughout the CNS. Inflammatory changes and IF were mild in one dog 8 days PI and moderate in the other, suggesting that this is about the time of earliest CNS infection. Primary CDV replication occurs in lymphoid tissue (Appel, 1969) and it seems that dissemination to other tissues, including the CNS, occurs by infected lymphocytes, probably in a nonspecific fashion. Why virusinfected lymphocytes migrate from the bloodstream into the CNS and other tissues is not known. Probably alterations to plasma membranes of infected lymphocytes are important in changing normal trafficking of these cells. Prineas and Wright (1978) have suggested that there is a normal migration of small lymphocytes through the CNS. The sequestration of these infected lymphocytes throughout the body may explain in part the lymphopenia characteristic of CD. The early inflammatory changes were widespread whereas the earliest demyelination occurred in a limited number of fiber tracts close to the ventricular system. Hence one could argue that demyelination is initiated by some factor, possibly CDV, derived from within the CSF. It is possible that viral diffusion and spread is impeded in white matter and that the earliest exposure of myelinated fibers to virus comes from infected cells in the CSF. Fluorescing mononuclear cells have been demonstrated in the CSF in CD (Appel, 1969) and ependymal infection was common. The formation of hypertrophic and multinucleated astroglia and syncytial ependyma accompanied the earliest demyelination. Raine (1976) also identified astroglial giant cells in early lesions of CDE. Our findings confirm his contention that this giant cell formation is associated with the earliest demyelination in CDE; furthermore we identified occasional areas of astroglial swelling in normally myelinated tissue as has been described recently in acute MS (Prineas and Raine, 1976) suggesting that astroglial activity precedes myelin breakdown. The initial demyelination was largely noninflammatory; we did occasionally find small cuffs in other areas of the CNS at the stage of early myelin loss which were presumably remnants of the primary wave. The source of macrophages in areas of demyelination was closely sought. Significantly, many con-
9
rained cytoplasmic CDV nucleocapsids but little in the way of identifying nuclear or cytoplasmic characteristics. Their presence was not heralded by an infiltrate of blood monocytes and so we conclude that they are endogeneous to the CNS. Several possibilities exist and these are not mutually exclusive. Astrocytes were infected and have recently been implicated to, some extent in myelin removal in CDE (Raine, 1976). Rarely a macrophage was found with bundles of cytoplasmic filaments allowing characterization of that cell as an astrocyte. Pericytes were found to be infected and may become active and phagocytic under suitable circumstances (Baron and Gallego, 1972). Possibly resting microglia were involved also as some glial processes with nucleocapsid, identified at day 20, may have been microglial. Autophagy of myelin by oligodendroglia is another, although less likely, possibility. The mechanism of demyelination in CDE is unknown. Wallerian-type degeneration would not account for selective myelin loss and oligodendroglial infection has rarely been demonstrated (Campbell, 1957; Raine, 1976). Consequently, the possibility of immune mediated damage to myelin in CDE has received attention of late (Koestner et al., 1974). Krakowka et al. (1973) demonstrated specific antimyelin antibody in both spontaneous and experimental CDE, but whether this antibody is operative in vivo in the demyelinating process is yet to be demonstrated. Our studies of the early stages of demyelination incriminate CDV. We often found glial cells containing nucleocapsids closely apposed to intact myelin sheaths within areas of active demyelination. It appeared that glial cells became infected and subsequently effected demyelination. Some of the known viral agents which will produce demyelinating encephalomyelitis (CDV, visna in sheep, mouse corona, measles in SSPE) are syncytia-forming viruses and so cell fusion may be involved in the mechanism of demyelination, as was alluded to by Choppin (1968). We observed that fusion of infected ependymal cells and astroglia accompanied the early demyelination in CDE. In an earlier study of experimental SSPE in the hamster, Raine et al. (1974) found polykaryocytes, some apparently derived from different cell types within brain lesions. Myelin disruption in CDE may result from the fusion of glial cell membranes from infected astroglia or microglia with myelin sheaths. Following fusion of plasmalemmae, neutral proteases from activated glial macrophages could degrade myelin (Cammer et al., 1978) leading to its phagocytosis. The factor in CDV which induces cell fusion is probably contained within envelope glycoproteins (Fisher and Bussell, 1977). Neutralization of these viral antigens may inhibit demyelination because Krakowka et al. (1975) associated reduced levels of anti-envelope antibody with demyelination in CDE.
10 Several authors have proposed that demyelination in C D E c a n n o t be a s c r i b e d to viral effects a l o n e ( K o e s t n e r et al., 1974; V a n d e v e l d e a n d K r i s t e n s e n , 1977). F o l l o w i n g s e n s i t i z a t i o n to m y e l i n o r viral antigens, f u r t h e r p r o g r e s s i v e loss o f m y e l i n m a y result f r o m an i m m u n e m e d i a t e d p r o c e s s , e i t h e r h u m o r a l , cellular, o r b o t h . T h i s m a y be c l a r i f i e d it it c a n be e s t a b l i s h e d w h e t h e r l y m p h o i d cells e n t e r i n g the C N S f o l l o w i n g the p r i m a r y d e m y e l i n a t i o n are r e s p o n d i n g in a p r o t e c t i v e i m m u n e c a p a c i t y o r to p a r t i c i p a t e in a n d a m p l i f y the process of myelin destruction.
Acknowledgements. We thank Dr. J. Cummings for helpful discussions and Dr. L. Krook for assistance with photography. The assistance of Mary Beth Metzgar and Ann Signore is greatly appreciated. This work was supported by the Whitehall and Richard King Mellon Foundations.
References Appel, M. J. G. : Pathogenesis of canine distemper. Am. J. Vet. Res. 30, 1167 1182 (1969) Appel, M. J. G., Gillespie, J. H.: Canine distemper virus. Virology monographs. Vol. 11. Wien, New York: Springer 1972 Baringer, J. R.: Tubular aggregates in endoplasmic reticulum in herpes-simplex encephalitis. N. Engl. J. Med. 285, 943-945 (1971) Baron, M., Gallego, A.: The relation of the microglia with the pericytes in the cat cerbral cortex. Z. Zellforsch. 128, 4 2 - 5 7 (1972) Blinzinger, K., Anzil, A. P., Deutschlfinder, N. : Nature of tubular aggregates. N. Engl. J. Med. 286, 157-158 (1972) Camer, W., Bloom, B. R., Norton, W. T., Gordon, S. : Degradation of basic protein in myelin by neutral proteases secreted by stimulated macrophages. A possible mechanism of inflammatory demyelination. Proc. Natl. Acad. Sci. USA 75, 15541558 (1978) Campbell, R. S. F. : Encephalitis in canine distemper. Br. Vet. J. 113, 143 161 (1957) Choppin, P. W. : Pathogenic immune reactions in virus infection. In : Textbook of immunopathology, (eds. P. A. Miescher, A. J. Mfiller-Eberhard). VoL 1. New York: Grune and Stratton 1968 Fisher, L. A., Bussell, R. H.: Cell fusion by canine destemper virusinfected cells and their plasma membranes. Intervirology 8, 218 - 225 (1977) Grimley, P. M., Schaff, Z. : Significance of tubuloreticular inclusions in the pathobiology of human diseases. In : Pathobiology annual,
Acta Neuropathol. (Berl.) 46 (1979) (ed. H. k. Joachim). Vol. 6. New York: Appleton-CenturyCrofts 1976 Imagawa, D. T. : Relationships among measles, canine distemper and rinderpest viruses. Prog. Med. Virol. 10, 160-193 (1968) Koestner, A. : Animal model - distemper-associated demyelinating encephalomyelitis. Am. J. Pathol. 78, 361 364 (1975) Koestner, A., McCullough, B., Krakowka, G. S., Long, J. F., Olsen, R. G.: Canine distemper: a virus-induced demyelinating encephalomyelitis. In : Slow virus diseases (eds. W. Zeman, E. H. Lennette). pp. 86-101. Baltimore: Williams and Wilkins 1974 Krakowka, S., McCullough, B., Koestner, A., Olsen, R.: Myelinspecific autoantibodies associated with central nervous system demyelination in canine distemper virus infection. Infect. Immun. 8, 819-829 (1973) Krakowka, S., Olsen, R., Confer, A., Koestner, A., McCullough, B. : Serologic response to canine destemper viral antigens in gnotobiotic dogs infected with canine distemper virus. J. Infect. Dis. 132, 384-392 (1975) McCullough, B., Krakowka, S., Koestner, A. : Experimental canine distemper virus-induced demyelination. Lab. Invest. 31, 216 222 (1974) ter Meulen, V., Katz, M., Mfiller, D.: Subacute sclerosing panencephalitis: A review. Curr. Top. Microbiol. Immunol. 57, 1 - 3 8 (1972) Prineas, J. W., Raine, C. S. : Electron microscopy and immunoperoxidase studies of early multiple sclerosis lesions. Neurology (Minneap.) 26, 29-32 (1976) Prineas, J. W., Wright, R. G.: Macrophages, lymphocytes, and plasma cells in the perivascular compartment in chronic multiple sclerosis. Lab. Invest. 38, 409 421 (t978) Rathe, C. S. : On the development of CNS lesions in natural canine distemper encephalomyelitis. J. Neurol. Sci. 30, t3 28 (1976) Raine, C. S., Byington, D. P., Johnson, K. P. : Experimental subacute sclerosing panencephalitis in the hamster. Ultrastructure of the chronic disease. Lab. Invest. 31, 355- 368 (1974) Reynolds, E. S. : The use of lead citrate at high pH as an electronopaque stain in electron microscopy. J. Cell. Biol. 17, 208-212 (1963) Summers, B. A., Greisen, H. A., Appel, M. J. G. : Possible initiation of viral encephalomyelitis in dogs by migrating lymphocytes infected with distemper virus. Lancet II, 187 190 (1978) Vandevelde, M., Kristensen, B. : Observations on the distribution of canine distemper virus in the central nervous system of dogs with demyelinating encephalitis. Acta Neuropathol. (Berl.) 40, 233 236 (1977) Wisniewski, 14., Raine, C. S., Kay, W. J.: Observations on viral demyelinating encephalomyelitis. Canine distemper. Lab. Invest. 26, 589-599 (1972) Received July 5, 1978/Accepted December 19, 1978
