Acta Neuropatholoslca

Transcription

1 Acta Neuropathol (Berl) (1987) 72:377-:388 Acta Neuropatholoslca 9 Springer-Verlag 1987 Virus spread and initial pathological changes in the nervous system in genital herpes simplex virus type 2 infection in mice A correlative immunohistochemical, light and electron microscopic study G. Georgsson*, J. R. Martin, G. L. Stoner, and H. de F. Webster Laboratory of Experimental Neuropatho][ogy, National Institute of Neurological and Communicative Diseases and Stroke, National Institutes of Health, Bethesda, MD 20892, USA Summary. Mice were infected by the vaginal route with the MS strain of herpes simplex virus type 2 (HSV-2). Serial vaginal cultures were used to confirm infection and to select mice for this study. Two mice were killed by perfusion on days 2-6 post infection (p.i.) and lumbar and sacral cord with cauda were fixed and embedded for electron microscopy. Semithin Epon-sections were stained for viral antigen using a rabbit anti-hsv-2 antiserum and the Avidin-Biotin (ABC) method. Thin sections from antigen-positive blocks were examined by electron microscopy, and the number and types of infected cells detected by these two methods were compared. A good correlation was found between detection of infected cells by these methods. Infected cells included neurons of dorsal root ganglia and spinal cord, satellite cells of dorsal root ganglia, non-myelinating Schwann cells, astrocytes, oligodendrocytes and arachnoidal cells. Infected cells were first detected in the cauda on day 3 p.i. and in the spinal cord on day 5; p.i. The temporal and spatial distribution of infected cells was consistent with neural spread to and within the CNS. The pathological lesions showed a good correlation with the distribution and number of infected cells and are probably due to a direct virus effect. The similar sensitivity of the Epon-ABC method to electron microscopy in detecting infected cells indicates that this method may have useful applications in both experimental and diagnostic work. Key words: Herpes simplex virus type 2 - Genital infection - Avidin-Biotin Method - Electron microscopy Offprint requests to: J. R. Martin (address see above) * Present address: Institute for Experimental Pathology, University of Iceland, Keldur, Box 8540, 128 Reykjavlk, Iceland Three of the key questions in studies of pathogenesis of central nervous system (CNS)viral infections are: how does the virus spread into and within the CNS, what cells become infected and how does infection correlate with pathological changes? The spread of herpes simplex viruses into the CNS has been extensively studied in various animal models following different routes of inoculation. The methods used for detection have been mainly virus isolation, immunofluorescence and electron microscopy or various combinations of these methods (reviewed by Baringer 1975). But, only one of these methods can be employed on a given tissue sample, a limitation recently overcome by the development of specific, sensitive immunostaining methods that can be used on tissue embedded in Epon and other resins (reviewed by Sternberger 1979). These methods now permit the determination of antigen distribution in semithin light microscopic sections and subsequent electron microscopic identification of the infected cells in thin sections cut from the same block. In each region examined, severity of infection now can be correlated with pathological changes. In the present work, we used this approach to study early events in the genital model of herpex simplex virus type 2 (HSV-2) infection in mice. The only method used previously to investigate spread of HSV-2 to the CNS after experimental genital infection in mice has been virus isolation (Lukas et al. 1975; Overall et al. 1975; Renis et al. 1976; Morahan et al. 1977; Stanberry et al. 1982). Besides our interest in the possible value of this methodological approach, we wanted to extend our earlier observations on localization of viral antigen and pathological lesions in the CNS in genital HSV-2 infection of mice (Martin and Stoner 1984) to changes found in the 1st week after infection, especially of the peripheral nervous system (PNS), as little attention

2 378 G. Georgsson et al.: Experimental genital HSV-2 infection in mice has been paid to damage of the PNS in HSV-2 infections. Our results on sections from the same block showed that electron microscopic examination and light microscopic immunocytochemistry were equally sensitive methods for virus detection, We also found that the distribution of infected cells correlated well with the observed pathology. The temporal and spatial distribution of HSV-2 was consistent with neural spread to and within the CNS. Materials and methods Virus preparation and isolation For details of virus preparation and virus isolation see Martin and Stoner (1984). In short, the MS strain of HSV-2 virus was used for inoculation of animals. The virus was grown in Vero cells. The tissue culture supernatant was harvested at the time of maximum cytopathic effect, centrifuged at 2000 rpm for 20 min and stored in aliquots at - 70 ~ C. The titer of the virus stock used in these experiments was TCIDso/ml. Animals and infection Outbread Swiss-Webster female mice (Buckberg Lab Animals, Thompkin's Cove, NY, USA) 12 weeks of age were used. Forty mice were inoculated; 20 served as uninfected controls. For inoculation, cotton pellets mm in diameter were saturated with 0.05 ml of undiluted virus stock suspension and inserted into the vagina using small curved forceps. To test if infection had been established, vaginal swabs were made on days 1, 2 and 3 after inoculation, cultured in Vero cells, and examined for cytopathic changes (Martin and Stoner 1984). Preparation of tissues Two infected mice were killed on each of days 2-6 post infection (p.i) for this experiment. The main guideline for selecting mice for this study were strongly positive cultures (Table 1). On days 4-6 p.i. mice were selected which, in addition to strongly positive cultures, had external genital lesions and/or hindleg weakness. These selection criteria were previously observed to be predictive of fulminant infection, severe CNS lesions, and death during the first 2 weeks of infection (Martin and Stoner 1984; Martin, unpublished observations). The mice were anesthetized with ether and perfused through the left ventricle with a mixture of 1% paraformaldehyde and 1% glutaraldehyde in 0.1 M phosphate buffei". The lower part of the vertebral column was then dissected out and put in the same fixative overnight. The lumbar and sacral cord, together with cauda equina, were removed and cut transversely in approximately 1-2-ram-thick blocks, 5-12 from each case. The blocks were postfixed in 2% OsO4 in 0.12 M phosphate buffer overnight and dehydrated in ethanol and embedded in Epon. One-micrometer-thick sections were cut from each block, one section stained with toluidine blue for light microscopic examination and other sections immunostained for viral antigens (see below). Blocks showing antigen-positive cells were trimmed and thin-sectioned. The sections were stained with uranyl acetate and lead citrate and examined in a Philips EM 400. In addition, several blocks negative by immunostaining were cut and examined by electron microscopy, with special emphasis on blocks from the first days after infection. Table 1. Virus isolations from vagina 1-3 days after vaginal infection of mice with HSV-2 a Killed day Virus isolations Day 1 Day 2 Day ND b ND Two mice killed on each day, both positive b ND: Not done Immunostaining A polyclonal antiserum raised against the 186 strain of HSV-2 in rabbits (Martin et al. 1982) was used for imunostaining of semithin sections by the Avidin-Biotin method. The sections were etched with sodium ethoxide (Erlandsen et al. 1979), 1 : 3 dilution of stock solution in absolute ethanol, for 15 min, washed in absolute ethanol, treated with 0.1% H202 in 0.1 M Tris buffer, ph 7.6 for 5 rain and washed in 0.5 M Tris buffer. Following incubation with blocking serum for 20 rain (Vectostain-kit, Vector Laboratories, Burlingame, CA, USA) the sections were incubated with rabbit anti-hsv-2 serum, usually in a dilution of 1:500 in 1% normal goat serum in 0.5 M Tris, overnight at 4~ This was followed by washing in 0.5 M Tris, incubation with biotinylated goat anti-rabbit IgG for 30 min and washing in 0.5 M Tris. Sections were then incubated with Avidin-Biotin complex for 30 min, washed in 0.5 M Tris buffer, treated for 7 rain with mixture of 0.05% diaminobenzidine (DAB) and 0.01% HzO2 in 0.1 M Tris. After washing in 0.1 M Tris buffer, the sections were treated with 2% OsO4 in 0.1 M phosphate buffer for 1-2rain, washed in 0.1 M Tris, dehydrated and mounted. Tissues of known antigen positivity and negativity and a pre-immune serum were used as controls. To plot the distribution of antigen-positive cells, all planes of sections showing positive iamnunostaining were photographed at a low magnification. Negatives were enlarged, projected onto paper and positions of positive cells were traced. Figures summarizing the locations of antigen-containing cells in lumbar and sacral cord levels and in nerve roots of 2 mice on each day p.i. were prepared by recording locations of positive cells on outlines of the cord at upper lumbar, lumbosacral junction, and lower sacral levels. For identification of antigen-positive cells, pairs of pictures, one of a toluidine blue-stained semithin section, and the other of an adjacent immunostained semithin section, were taken at a fixed magnification. Areas of all sections showing immunostaining were examined, and an attempt was made to trace every antigen-positive cell on the corresponding print of the toluidine blue-stained section. Results Correlative immunohistochemistry and electron microscopy Results of a comparison of virus detection in semithin Epon sections immunostained for viral antigen and in adjacent thin sections examined electron microscopi-

3 G. Georgsson et al. : Experimental genital HSV-2 infection in mice 379 Table 2. Proportion of blocks and number of cells showing positive immunostaining (Ag) and viral particles by electron microscopy (EM) a Killed Posterior roots Spinal ganglia Spinal cord Total positive day Ag EM Ag EM Ag EM Blocks Cells Ag EM Ag EM 2 0/13 Nd 0/2 0/1 0/12 ND /16 0/ 1 0/3 ND 0/12 ND /15 1/ 2 0/5 0/1 0/14 0/ /21 2/10 1/2 1/1 3/20 3/ /17 1/ 6 ND ND 8/18 7/ Number of positive blocks/total number of blocks examined Day 4 Day 5 Day 6 o Fig. 1. Distribution of antigen-positive cells in spinal cord and nerve roots on days 4-6 post infection (p.i.). Each dot represents one infected cell. The uppermost plane represents upper and middle levels of lumbar cord; the middle plane, lower lumbar and upper sacral levels, and the lowermost plane, lower sacral cord and cauda. Anterior and posterior roots and cauda are each shown as single structures at the anterior and posterior horns (see text) cally for virus particles are summarized in Table 2. There was in general a good correlation between the methods. In some instances, viral particles were not found by electron microscopy in sections showing viral antigens. Overall, in three out of 17 blocks positive by immunostaining, viral particles were not detected by electron microscopy. In nerve roots, infected cells could be detected by electron microscopy in only four out of ten nerve roots showing positive immunostaining. In seven out of eight spinal cords with positive immunostaining, viral particles could also be detected by ultrastructural analysis. When counts of infected cells in immunostained sections were compared to counts of infected cells in electron micrographs, a higher number was found on immunostained sections (Table 2). This discrepancy between these methods is due to sampling. Sections were first cut for antigen screening, then blocks were further trimmed for electron microscopy: thus, areas containing antigen-positive cells were not completely included in sections examined by electron microscopy. When cell counts were compared in precisely comparable areas, a very good correlation was found between the number of cells detected by

4 380 G. Georgsson et al.: Experimental genital HSV-2 infection in mice Table 3. Comparison of number and types of infected cells in peripheral nervous system as detected by immunostaining (Ag) and electron microscopy (EM) Day Neurons Schwann cells Satellite cells Arachnoidal cells Unidentified cells Ag EM Ag EM Ag EM Ag EM Ag EM Total Table 4. Comparison of number and types of infected cells in the central nervous system as detected by immunostaining (Ag) and electron microscopy (EM) Day Neurons Astrocytes Oligodendrocytes Glia Unidentified Ag EM Ag EM Ag EM Ag EM Ag EM Total The column headed glia includes cells that could be identified as glial cells where further characterization was uncertain both methods. Thus the same number of infected cells was found in five sections and a difference of only one to three cells in eight additional sections (data not shown). As shown in Table 2, very few infected cells were detected on the first few days after infection. However a substantial increase in number of cells occurred on day 5. The greater number of cells found on day 5 than 6 is because on day 5 a heavily infected spinal ganglion was included (32 infected cells were found), whereas in the sections from day 6, a spinal ganglion was not included. A comparison of the number of infected cells in the spinal cord showed an increased number of infected cells on day 6 (Fig. 1). In none of the 33 blocks antigen-negative by immunostaining were viral particles observed by electron microscopy. Localization of antigen-positive cells The distribution on antigen-positive cells on successive days p.i. is shown diagrammatically in Fig. 1. On the 3rd day after infection, the first antigen-positive cell, presumably a Schwann cell, was detected in a nerve fascicle of the cauda. On the 4th day, two antigenpositive cells were found, also in a nerve root of the cauda. By electron microscopy of the corresponding thin section, these cells could be located and identified as non-myelinating Schwann cells which contained viral nucleocapsids in their nuclei and cytoplasm. On the 5th day p.i., antigen-positive cells were found in a spinal ganglion (not shown), posterior nerve roots and, for the first time, in the spinal cord. Antigen in the cord was restricted to the posterior root entry zone and the adjacent part of the posterior horn of the gray matter at lower lumbar and upper sacral levels. On the 6th day p.i. there was a definite increase in the number of antigen-positive cells, and they were more widely distributed. The increase in number was mainly at lower lumbar and upper sacral levels of the spinal cord. In addition to an increased density of antigen-positive cells in the posterior root entry zone and posterior horn of the gray matter, cells were more widely dispersed in the gray matter and were also found at higher and lower levels. At higher levels the immunostained cells were mainly found in the gray commissure, but a few cells were also present in the posterior column. At lower levels, representing lower sacral cord, antigen-positive cells were mainly found in the same location as on the middle level, i.e., in the posterior root entry zone. Anterior nerve roots were always negative. Types of cells infected An effort was made to identify all infected cells on immunostained sections and in electron micrographs (Tables 3 and 4, Figs. 2-4). In the peripheral nervous system, the identity of infected dorsal root ganglion cells was readily established in immunostained

5 G. Georgsson et al.: Experimental genital HSV-2 infection in mice 381 Fig. 2A- C. Spinal ganglion. A Toluidine blue, B rabbit anti-hsv-2 serum 1 : 500, C control. Pre-immune rabbit serum 1 : 500. Neuron with cytoplasmic and nuclear swelling with positive immunostaining of nucleus and cytoplasm (arrows). Satellite cell with positive immunostaining (arrowheads). Several other neurons show positive immunostaining of nuclei or cytoplasm; x 250. D and E Nerve fascicle of cauda. D Toluidine blue, E rabbit anti-hsv-2 serum 1:250. Two antigen-positive cells (arrows); the lower swollen, the upper with condensation and margination ofchromatin, 420 sections (Fig. 2). The identity of infected cells in nerve roots in immunostained sections was frequently uncertain, even when seen in adjacent toluidine bluestained sections (Fig. 2). Ultrastructurally these cells were Schwann cells surrounding unmyelinated fibers (Fig. 4). Thus, the inability to resolve these fibers in semithin sections was an important limitation in the identification of infected cell types in the nerve roots by light microscopy. In the central nervous system, a comparison of irmnunostained sections and adjacent toluidine blue sections facilitated the identification of immunostained cells (Fig. 3), although precise identification of glial cell types was often difficult (Table 4).

6 382 G. Georgsson et al.: Experimental genital HSV-2 infection in mice Fig. 3A, B. Spinal cord. A Toluidine blue, B rabbit anti-hsv-2 serum 1:500. Several antigen-positive cells (some encircled) and same cells in toluidine blue section. 1-3 astrocytes; 4 and 5 oligodendrocytes; 6 necrotic cell. Besides necrotic cells, several glial ceils cannot be identified. 230

7 G. Georgsson et al.: Experimental genital HSV-2 infection in mice 383 Fig. 4. A Infected non-myelinating Schwann cell with nuclear swelling, clumping of chromatin and reduplication of nuclear membrane. Several nucleocapsids, hollow or with cores, in nucleus and cytoplasm. Some mature virions peripherally in nucleus. Axons intact x B Infected arachnoidal cell with several nucleocapsids in nucleus, x 27500

8 384 G. Georgsson et al.: Experimental genital HSV-2 infection in mice By electron microscopy, the identity of the infected cells could be ascertained with the exception of a few necrotic cells. In the peripheral nervous system, i.e., spinal ganglia and nerve roots, the cells most commonly infected were non-myelinating Schwann cells, which often showed a productive infection with formation of complete virions (Fig. 4). A reduplication of the nuclear membrane was frequently seen and usually the complete virions acquired their envelope from the nuclear membrane (Fig. 4). Occasionally, however, virions appeared to get the enveloping membrane from smooth membranes in the cytoplasm. In ganglia, neurons were frequently infected and viral nucleocapsids, with or without dense cores, and complete virions were observed. Satellite cells were often infected as well, and were occasionally seen to contain mature virions. A few of the outermost cells of the sheaths of the posterior roots contained viral nucleocapsids in their nucleus and cytoplasm (Fig. 4); no mature virions were found. These cells are thought to be arachnoidal in type, since they lack an enveloping basement membrane, and are joined to one another by desmosomes. In the CNS, astrocytes were most frequently found to harbor virus, followed by neurons and oligodendrocytes. In all these cell types, maturation to complete virions was found, frequently in neurons and astrocytes but less commonly in oligodendrocytes. Occasionally, extracellular virions were found both in the PNS and CNS. Intra-axonal virus particles were found on a few occasions both in unmyelinated and myelinated axons of the PNS. Pathology In general, the presence of pathological lesions correlated well with the presence of infected cells, progressing in size from the time of their first appearance (Fig. 1). Definite but minimal lesions, confined to nerve roots, were present on days 3 and 4, p.i. On days 5 and 6, spinal cord lesions appeared, and larger root lesions were seen. Lesions observed in sensory nerve roots on day 3 consisted of occasional myelinated fibers showing early degradation of myelin. On day 4, a minimal inflammatory component, consisting of lymphocytes, macrophages and polymorphonuclear leukocytes appeared, and increased in amount on subsequent days. On day 5, a few infected non-myelinating Schwann cells were found. Occasionally, an axon which was surrounded by an infected cell appeared vacuolated and was partly or totally disrupted. An occasional necrotic cell was found. On day 6, the lesions in the posterior nerve roots were more prominent. Only Schwann cells surround- ing unmyelinated fibers were seen to be infected, although partially or completely demyelinated axons of large diameter were occasionally found (Fig. 5). Damage of both myelinated and unmyelinated axons was common. Generally, unmyelinated axons which were enveloped by infected Schwann cells appeared normal (Fig. 4), but some showed vacuolization and sometimes complete disorganization with vesicular disruption (Fig. 5). Myelinated axons showed damage ranging from reactive changes with accumulation of mitochondria, neurofilaments and dense bodies to mild or severe degeneration, with vacuolization of axoplasm or complete necrosis (Fig. 5). A moderate degree of inflammation was evident, and phagocytes containing myelin breakdown products and remnants of necrotic cells were frequently seen. In a spinal ganglion from day 5, neurons were frequently swollen, showing karyolysis and chromatolysis (Fig. 2). Some infected neurons had vacuolated cytoplasm or accumulations of parallel stacks of smooth endoplasmic reticulum. An occasional neuron was necrotic, with numerous lipid droplets, irregular dense bodies, and vacuoles. Inflammation, like that found in the roots, was moderate in degree. There were, in addition, numerous phagocytes containing cell debris and membranous material in their cytoplasm. Lymphocytes were occasionally seen invading through the sheath of satellite cells. Several infected, non-myelinating Schwann cells (Fig. 4) and satellite cells (Fig. 2) had marginated and clumped chromatin, contained myelin-like membranous material in their cytoplasm, and some were necrotic. An occasional non-myelinated axon showed vacuolization of the axoplasm. Abnormalities in the spinal cord were found mainly in the meninges, posterior root entry zone, and adjacent gray matter of the posterior horn. They were of similar character on day 5 and 6, but a distinct progression which paralleled that of viral antigen distribution (Fig. 1) was seen. On day 5, a mild meningitis was present (Fig. 6), with scattered infiltration of lymphocytes, macrophages, and occasional plasma cells and polymorphonuclear leukocytes. Inflammation was most prominent at lower lumbar and upper sacral levels, diminishing cranially and caudally. Meningitis was moderate on day 6, extending to higher and lower cord levels, but did not reach the uppermost lumbar cord or lowest sacral levels. In the parenchyma, inflammation was practically absent on day 5, but on day 6, a moderate inflammatory reaction was present in the posterior root entry zones (Fig. 6). Inflammatory infiltrates consisted exclusively of mononuclear cells, lymphocytes, and macrophages which contained myelin degradation products. Vesicular disruption of neuropil (Fig. 3) was

9 G. Georgsson et al. : Experimental genital HSV-2 infection in mice 385 Fig. 5A, B. Posterior nerve root. A Infected non-myelinating Schwann cells (S) with margination and clumping of chromatin. Several phagocytes with myelin debris. Thinly mye, linated (arrow) and demyelinated axons (arrowhead); x B Infected non-myelinating Schwann cells (S) with clumping and ma,rgination of chromatin. Infiltration of lymphocytes (L) and polymorph leukocyte (P). Vacuolization and vesicular disruption of axons (arrows). Several myelinated axons with reactive or degenerative changes (arrowheads); 5530

10 386 G. Georgsson et al.: Experimental genital HSV-2 infection in mice Fig. 6. Spinal cord with posterior nerve root (upper right). Macrophages (M) in meninges and posterior root entry zone. Necrotic cell in meninges. Two swollen infected astrocytes (A). Some thinly myelinated and demyelinated axons. Swelling and vacuolization of several cytoplasmic processes; x 4300 present, which on the ultrastructural level appeared as swollen cytoplasmic processes with vacuolization or vesicular disruption, or as vacuolar widening of the extracellular space (Fig. 6). Axons frequently showed similar reactive or degenerative changes as in the posterior roots, and thinly myelinated or demyelinated axons were seen (Fig. 6). The infected cells were scattered throughout the lesions, and in nearby tissue of more normal appearance. Infected neurons and astrocytes were occasionally lytic or necrotic (Fig. 6). Infected cells which could be identified as oligodendrocytes showed few cytological alterations. A few necrotic cells were seen, and some of these appeared to be inflammatory in origin. Discussion This study combines, for the first time, HSV-antigen detection, and light and electron microscopy on adjacent sections of the same blocks of tissue. This approach thus overcomes uncertainties that occur when antigen screening and morphological studies are performed on different animals. Further, this rather simple method permits rapid screening of larger areas of Epon-embedded tissues that can usually be sectioned for electron microscopy, and also allows the detection of a few antigen-containing cells at early stages of infection. This antigen detection method appears to be similar in sensitivity to the detection of infected cells by electron microscopy, and gives results comparable to those obtained by virus isolation in earlier studies on spread of HSV-2 to the CNS in genital infection of mice (Overall et al. 1975; Renis et al. 1976; Morahan et al. 1977). In most previous animal models, entry of HSV into the CNS has been studied using HSV-1 in immature (Johnson 1964; Yamamoto et al. 1965, 1973; Severin and White 1968; Kristensson et al. 1971, 1978; Dillard et al. 1972; Townsend and Baringer 1976; Lascano and Berria 1980) or adult animals (Wildy 1967; Cook and Stevens 1973; Townsend and Baringer 1978; Townsend 1981 a, b; Kristensson et al. 1979) and by a variety of inoculation routes, including corneal scarification, cutaneous abrasion, or subcutaneous footpad inoculation. In the present study, we have established HSV-2 infections in adult animals using a vaginal route of inoculation, in an effort to more precisely mimic natural infections with this agent.

11 G. Georgsson et al.: Experimental genita]l HSV-2 infection in mice 387 In many respects, our findings in the PNS are in keeping with those studies in which HSV-1 was inoculated into adult mice by various peripheral routes. Evidence that these experimental conditions lead to spread of virus to the CNS via peripheral nerves is in accord with our observations of a PNS infection which precedes that in the spinal cord, and by the close correspondence of viral antigen-containing sensory roots to those spinal cord levels in which viral antigen first appears in sensory root entry zones. These observations are consistent with and extend our previous observations on HSV-2 infections in mice, which show that presence or absence of spinal cord disease correlates with presence or absence of sensory root lesions at lumbosacral levels (Martin and Stoner 1984). The suggestion that in adults HSV-2, unlike HSV-I, may spread to the CNS via the blood (Craig and Nahmias 1973) is not supported by our findings. The evidence that, in nerves, axons are the principal conduit for HSV passage from the periphery to the CNS (Wildy 1967; Kristensson et al. 1971; Hill et al. 1972; Cook and Stevens 197:3) has gained wide acceptance. The occasional observation of virus within axons in peripheral nerves in this study is consistent with this theory. In several respects, our observations in the PNS differ from those reported for HSV-1 infection in mice of a comparable age (Kristensson et al. 1979; Townsend 1981 a, b). We regularly find lesions in peripheral roots, although these lesions are usually not very large. Schwann cells, at least those which surround unmyelinated fibers, are infected, and may contain both nucleocapsids and mature virions, thus providing evidence that virus maturation in Schwann cells of adult animals is not arrested at the stage of nucleocapsid appearance, as has been suggested (Hill and Field 1973; Townsend 1981 a). Further, our findings of demyelinated and denuded axons in nerve root lesions, and of myelin debris in nearby macrophages has not previously been reported, and indicate that infection can result in Schwann cell destruction. Thus in a mature host, HSV may exhibit a greater potenital for PNS disease than has previously been recognized. In the CNS, a glial infection in root entry zones and a mixed neuronal and glial infection in the posterior gray horns were the chief early findings in this study. In HSV models, demyelination of central white matter at root entry zones with sparing of corresponding peripheral roots has been reported (Townsend and Baringer 1976; 19'78; Kristensson et al. 1978; Martin 1984) and in the present study, the disproportionate amplification of infection in spinal cord parenchyma compared to sensory roots on days 5 and 6 appears to be analogous to findings in previous reports. Several factors could limit HSV infection and disease in the PNS. First, virus maturation in Schwann cells could indeed tend to be abortive (Hill and Field 1973; Townsend 1981 a). Second, if Schwann cell infection reflects secondary spread from virus-containing axons, the extension of infection to neighboring Schwann cells could be impaired by the basal lamina which surrounds each Schwann cell, by the collagen matrix which fills the large extracellular space, and by the opportunity for the nascent immune response to block cell-to-cell virus spread due to large distances between Schwann cells of adjacent fibers. In the CNS, on the other hand, the numerous, extensive cytoplasmic processes of all cell types, their intimate contacts with one another, and the minimal extracellular space could all be expected to facilitate the spread of virus from one cell to the next, and the amplification of infection; the higher ratio of myelin sheaths to numbers of perikarya of myelin-forming cells in the CNS than in the PNS could provide relative amplification of CNS demyelination (Kristensson et al. 1978; Martin 1984). Acknowledgements. The authors are grateful to Mrs. Yoong Chang, Ms. Elizabeth Reed, Ms. Patricia McMahon and Ms. Eygl6 Gislad6ttir for technical help, and Mrs. Margr6t Kristinsd6ttir and Ms. Hrafnhildur Gudmundsd6ttir for secretarial help. References Baringer JR (1975) Herpes simplex virus infection of nervous tissues in animal and man. Prog Med Virol 30:1-26 Cook ML, Stevens JG (1973) Pathogenesis of herpetic neuritis and ganglionitis in mice. Evidence of intra-axonal transport of infection. Infect Immun 7: Craig CP, Nahmias AJ (1973) Different patterns of neurologic involvement with herpes simplex virus types 1 and 2: isolation of herpes simplex virus type 2 from the buffy coat of two adults with menigitis. J Infect Dis 127: Dillard SH, Cheatham W J, Moses HL (1972) Electron microscopy of zosteriform herpes simplex infection in the mouse. Lab Invest 26: Erlandsen SL, Parsons JA, Rodning CB (1979) Technical parameters of immunostaining of osmicated tissue in epoxy sections. J Histochem Cytochem 27: Hill T J, Field HJ (1973) The interaction of herpes simplex virus with cultures of peripheral nervous tissue: an electron microscopic study. J Gen Virol 21 : Hill TJ, Field HJ, Roome APC (1972) Intra-axonal location of herpes simplex virus particles. J Gen Virol 15: Johnson RT (1964) The pathogenesis of herpes virus encephalitis. I. Virus pathways to the nervous system of suckling mice demonstrated by fluorescent antibody staining. J Exp Med 119: Kristensson K, Lycke E, Sj6strand J (1971) Spread of herpes simplex virus in peripheral nerves. Acta Neuropathol (Bed) 17:44--53

12 388 G. Georgsson et al.: Experimental genital HSV-2 infection in mice Kristensson K, Vahlne A, Persson LA, Lycke E (1978) Neural spread of herpes simplex virus types 1 and 2 in mice after corneal or subcutaneous (footpad) inoculation. J Neurol Sci 35: Kristensson K, Svennerholm B, Persson L, Vahlne A, Lycke E (1979) Latent herpes simplex virus trigeminal ganglionic infection in mice and demyelination in the central nervous system. J Neurol Sci 43: Lascano EF, Berria MI (1980) Histological study of the progression of herpes simplex virus in mice. Arch Virol 64: Lukas B, Wiesendanger W, Schmidt-Ruppin KH (1975) Herpes genitalis in Guinea pigs. I. Kinetic study in infection with herpes-virus hominis type 2. Arch Viro149:1-11 Martin JR (1984) Intra-axonal virus in demyelinative lesions of experimental herpes simples type 2 infection. J Neurol Sci 63 : Martin JR, Stoner GL (1984) The nervous system in genital herpes simplex virus type 2 infections in mice. Lethal panmyelitis or nonlethal demyelinative myelitis or meningitis. Lab Invest 51: Martin JR, Stoner GL, Webster H de F (1982) Lethal encephalitis and non-lethal multifocal central nervous system demyelination in herpes simplex virus type 2 infections in mice. Br J Exp Pathol 63: Morahan PS, Breining MC, McGeorge MB (1977) Immune responses to vaginal or systemic infection of Balb/c mice with herpes simplex virus type 2. J Immunol 1i9: Overall JC Jr, Kern ER, Schlitzer RL, Friedman SB, Glasgow LA (1975) Genital herpes-virus hominis infection in mice. I. Development of an experimental model. Infect Immun 11: Renis HE, Eidson EE, Mathews J, Gray JF (1976) Pathogenesis of herpes simplex virus type 1 and 2 in mice after various routes of inoculation. Infect Immun 14: Severin M J, White RJ (1968) The neural transmission of herpes simplex virus in mice. Am J Pathol 53: Stanberry LR, Kern ER, Richards JT, Abbott TM, Overall JC Jr (1982) Genital herpes in guinea pigs. Pathogenesis of the primary infection and description of recurrent disease. J Infect Dis 146: Sternberger LA (1979) Immunocytochemistry, 2nd edn. John Wiley and Sons, New York Townsend JJ (1981 a) The relationship of astrocytes and macrophages to CNS demyelination after experimental herpes simplex infection. J Neuropathol Exp Neuro140: Townsend JJ (1981 b) The demyelinating effect of corneal H SV infections in normal and nude (athymic) mice. J Neurol Sci 50: Townsend JJ, Baringer JR (1976) Comparative vulnerability of peripheral and central nervous tissue to herpes simplex virus. J Neuropathol Exp Neurol 35:100 [abstr] Townsend J J, Baringer JR (1978) Central nervous susceptibility to herpes simplex infection. J Neuropathol Exp Neurol 37: Yamamoto T, Otani S, Shiraki H (1965) A study of the evolution of viral infection in experimental herpes simplex encephalitis and rabies by means of fluorescent antibody. Acta Neuropathol (Berl) 5: Yamamoto T, Otani S, Shiraki H (1973) Ultrastructure of herpes simplex virus infection of the nervous system of mice. Acta Neuropathol (Berl) 26: Wildy P (1967) The progression of herpes simplex virus to the central nervous system of the mouse. J Hyg (Lond) 65: Received June 2, 1986/Accepted October 13, 1986