- Email: [email protected]
Neural spread of herpes simplex virus types 1 and 2 in mice after corneal or subcutaneous (footpad) inoculation
Journal of the Neurological Sciences, 1978, 35:331-340 © Elsevier Scientific Publishing Company, Amsterdam- Printed in The Netherlands
331
NEURAL SPREAD OF HERPES SIMPLEX VIRUS TYPES I A N D 2 IN MICE AFTER CORNEAL OR SUBCUTANEOUS (FOOTPAD) INOCULATION
K. KRISTENSSON, A. V A H L N E , L. A. PERSSON and E. LYCKE
Department of Virology, Institute of Medical Microbiology, University of Gbteborg, Gbteborg, and Neuropathological Laboratory, Department of Pathology II, University of Link6ping, Linkgping (Sweden) (Received 28 July, 1977)
SUMMARY
Twelve herpes simplex virus (HSV) strains, 6 of each type, were inoculated subcutaneously into the left hind foot and into the cornea of the right eye of 12-day-old Swiss albino mice. The neural spread of virus to trigeminal and spinal ganglia and to brain and spinal cord was studied by demonstration of infective virus, histology and electron microscopy. Type 1 and type 2 infections seemed to spread equally well by intra-axonal transport. Using a protein tracer (horseradish peroxidase) injected into the same site as the virus, transport of the tracer to neurons corresponding to those infected with virus was observed. The extensive destruction of CNS tissue in the transitional region of the trigeminal root where CNS and the peripheral nervous system meet is discussed with reference to the pathogenesis of HSV encephalitis.
INTRODUCTION
Studies on base sequences of herpes simplex virus (HSV) type 1 and 2 genomes have revealed that the two subtypes differ by as much as 50 ~ (Kieff, Hoyer, Bachenheimer and Roizman 1972; Ludwig, Biswal and Benyesh-Melnick 1972). In spite of the genetic differences, pronounced biological subtype characteristics are difficult to find. Both types seem to cause latent reactivable infections in ganglion cells (Stevens and Cook 1973; Baringer and Swoveland 1974); and may thus invade the central nervous system by intra-axonal transport (Kristensson, Lycke and Sj6strand 1971; Cook and Stevens 1973). However, in adults HSV encephalitis is prevailingly caused by type 1 infections (Baringer 1974). The present study reports observations on transport of both types of HSV to trigeminal and spinal ganglia.
332 MATERIAL AND METHODS
Virus Altogether 12 strains, 6 of each type of HSV, were studied. Of these 3 (F, MP and G) were well-known strains supplied by Dr. Bernhard Roizman, Chicago, 111. Another 5 were thoroughly characterized in our laboratory using serology, thymidine kinase stability, sensitivity to thymidine and silver nitrate, and ability of strains to induce liver necrosis in mice as the discriminating tests (Vahine, Blomberg, Olofsson and Lycke 1975). These 8 strains were all passed in mice before use. Four strains were freshly isolated from patients and not passed in mice before testing. They were typed by their ability to induce liver necrosis in mice and by immunoelectro-osmophoresis against type-specific antiserum (Jeansson 1972). Batches of each of the strains were produced in green monkey kidney AH-1 (GMK) cells, dispersed in 1 ml aliquots and kept frozen a t - - 7 0 °C until used. Each batch of virus was titrated in G M K cells and by intracerebral inoculation of 3-week-old Swiss albino mice of our own laboratory breed. The pfu/LDs0 ratios were determined. These ratios were about 1 for all the strains used. Before inoculation the strains were diluted in Hanks' BSS so that about 1000 pfu were inoculated of each type.
Inoculation procedure Of each of the 12 virus strains 0.02 ml was injected subcutaneously into the left hind footpad of groups of 12 12-day-old anesthetized suckling Swiss albino mice. From these groups 2 mice were taken 2, 3 and 4 days after the virus inoculation for virus titrations and 2 mice after 2 and 3.5 days for histological examination. Samples were taken from the left foot, left sciatic nerve, left and right lumbar spinal ganglia, lumbar spinal cord and left cerebral hemisphere. In addition, in 6 groups of 12 mice, a drop of virus suspension from 3 type 1 and 3 type 2 strains was placed on to the cornea of the right eye which was scratched several times with an injection needle. The trigeminal ganglia on both sides and pieces from the brain stem and left cerebral hemisphere were sampled from 2 mice after 2 and 3 days for virus titrations and from 2 mice after 2 and 3.5 days for histological examinations. Blood samples were also taken for virus titration. Virus titrations For titration of virus, each specimen was ground with a glass pestle and suspended in 1 ml cell culture medium and inoculated undiluted and in 10-fold dilutions, respectively, into 2 cultures of G M K cells, 0.5 ml per culture. The cells were cultivated in 5 cm plastic dishes using Eagle's MEM supplemented with 2 ~o bovine serum and antibiotics. The medium of the inoculated cultures was replaced by cell culture medium containing 1 ~ methyl cellulose. After inoculation at 37 °C for 4 days the number of plaques was counted. Histology and electron microscopy For morphological observations, the mice were perfused with 2.5 ~ glutaraldehyde in phosphate buffer or with 5 ~ glutaraldehyde in Millonig's buffer with 2.7
333 low molecular dextran and 6.85 % sucrose to provide optimal fixation of central myelin sheaths (Carlstedt 1977). The specimens were embedded in E p o n , 1/zm thick sections were cut and stained with toluidine blue. Thin sections from selected areas were stained with uranyl acetate and lead citrate and examined in an Elmiskop. For demonstration of direct pathways from the periphery to the nerve cell bodies the protein tracer horseradish peroxidase (HRP) (type II, Sigma) was used. In 1 group of mice 0.02 ml of 50% H R P in Ringer solution was injected into the left footpad. In another group of mice, the right cornea was scratched several times with a needle, after which 0.2 mg H R P in dry powder was directly applied to the scarified cornea and the eyelid closed with a suture. Twenty-four hours later the mice were perfused with 2.5 % glutaraldehyde in phosphate buffer. Frozen sections, 30--40/~m thick, from lumbar spinal ganglia and cord, and from trigeminal ganglia were cut and incubated with 3,3'-diaminobenzidine and hydrogen peroxide for demonstration of peroxidase activity (Graham and Karnovsky 1966). RESULTS
Course of infection The spread of virus was judged on the appearance of symptoms of infection of the CNS and spinal cord, on recovery of virus from specimens collected and on morphological findings. After corneal inoculation all mice developed symptoms of ipsilateral conjunctivitis. By the third day neurological signs and deaths were observed and 4 days after inoculation all animals remaining, except those inoculated with type 1 strain F, were moribund. Mice inoculated into the left hind foot developed on days 2 and 3 after inoculation left sided pareses followed by parapareses. All animals became moribund and died before day 5, except those inoculated with strain F in which onset and progress of disease were delayed by 1 day. Otherwise there were no differences between the type 1 and 2 strains with regard to length of incubation period (time of inoculation to appearence of neurological symptoms), symptomatology and time of death.
Virological observations After infection by the corneal route infective virus was demonstrable in the ipsilateral trigeminal ganglion, and in the brain stem but not from corresponding specimens of the uninoculated side. One group of animals infected with a type 2 strain had virus in the ganglion but not in the brain stem and no virus was found in animals inoculated with strain F (type 1). In mice inoculated into the left foot infective virus was demonstrable in the inoculated foot of all the animals, but from the sciatic nerve in only 2 groups of type 1 inoculated mice and in 1 inoculated with type 2. Four of the 6 type 1 strains were demonstrable in the left spinal ganglia while this was the case with 2 of the type 2 strains. From all mice, virus isolations were made from spinal cord. On the other hand, no infective virus was demonstrated in the right spinal ganglion or in the blood specimens. Thus, there were indications of neural spread of virus to trigeminal and spinal
334 Ioa no pfu -10 6-
I
4-
2-
ot 2
3
t, days
Fig. 1. Amounts of virus in spinal cords of mice inoculated subcutaneously in left hind footpad. The spinal cord specimens were examined on days 2, 3 and 4 after virus inoculation. Means in log no. pfu per spinal cord of HSV type 1 (open circles) and type 2 inoculated mice (filled circles) are plotted against time in days. Bars denote ranges of titers. ganglia as well as to brain stem and spinal cord. One type 1 strain seemed less neuroinvasive than the others, but as a whole there were no significant differences in spread of the type 1 and type 2 strains. Varying amounts of virus were detected, which might partly be explained by differences in the amounts of tissue available of the different specimens examined. In Fig. 1 the virus titers obtained with the spinal cord specimens of footpad inoculated mice are plotted against time in days after inoculation. It was also observed that with the spinal cord specimens - - which did not markedly differ in size - - the amounts of virus demonstrable varied considerably. High virus titers were observed in both type 1 and type 2 virus-inoculated mice.
Morphological observations After corneal inoculation infected nerve cells were seen in the ipsilateral trigeminal ganglion 2 days after inoculation of 2 of the strains. The neurons showed intranuclear inclusion bodies of Cowdry type A or had a swollen homogenous nucleus and displayed signs of chromatolysis or degeneration. Such neurons were scattered within the medial part of the ganglion (Fig. 2). After corneal application of HRP, scattered neurons containing the reaction product were seen within the same area (Fig. 3P). This area corresponds to neurons innervating the cornea (Arvidsson 1975). After 3.5 days, when several of the mice were moribund, many infected neurons were seen in the medial part of the ganglion. The altered neurons were confined to this area and there was no spread into areas corresponding to the maxillar and mandibular branches of the ganglion (Fig. 4). After 3.5 days there was also extensive degeneration of the trigeminal nuclei in the brain stem and widespread tissue destruction. Interesting findings were made at the transitional region of the trigeminal root, where tissue from the CNS meets the peripheral nervous system (PNS). The PNS part of the ~oots was well preserved and showed no light-microscopic changes. The CNS tissue, on the contrary, showed extensive changes beginning directly at the border. There was a marked sponginess and destruction of the CNS myelin, degeneration of glial cells and infiltration of mononuclear inflammatory cells (Fig. 5). Ultrastructurally, the transition from PNS myelin to CNS myelin generally occurs at a node of Ranvier. Often a nerve fiber myelinated in the PNS part was naked in the CNS (Fig~ 6), but it
335
Fig. 2_ Chromatolytic neurons with swollen, homogenous nucleus in ipsilateral trigeminal ganglion 2 days after inoculation to scarified cornea, x 735. Fig. 3. HRP-containing neurons in the ipsilateral trigeminal ganglion after application to scarified cornea. Darkfield condenser, x 775. Fig. 4. Extensive infection of most neurons in medial part of trigeminal ganglion (M) 3.5 days after injection. Lateral part (L) is not involved, x 375.
336
Fig_ 5. Trigeminal root 3.5 days after HSV inoculation into the Jpsdateral cornea. Extensive changes in CNS part, while the PNS looks normal. ' 1520. Fig. 6. Nerve fiber at the transitional region of the trigemmal root showing well-developed PNS myelin, while the CNS part of the axon is naked. Note also glial cells with herpes nucleocapsids in their nuclei. "< 7980.
337
Fig. 7. Nerve fiber at the trigeminal transitional region with splitting and degeneration of the CNS myelin while the PNS myelin is intact, x 7980. Fig. 8. Schwann cell close to the transitional region containing herpes virus particles. The myelin sheaths and axons are well preserved, x /8,616_
should be noted that at this stage o f development the contralateral r o o t still contains a few well-myelinated P N S fibers which are transformed into non-myelinated C N S fibers. In several nerve fibers morphologically well-preserved P N S myelin could be seen directly bordering the C N S myelin, which showed vacuolation due to separation from the axolemma or splitting at interperiod lines (Fig. 7). Virus particles were f o u n d both in Schwann cells in the PNS part o f the zone and in astrocytes and oligodendroglial-like cells in the C N S part (Figs. 6 and 8). In the trigeminal ganglion herpes virus
338 particles were found in neurons seen by light microscopy to be abnormal and in satellite cells. After injection into the left foot infected neurons were seen in the ipsilateral spinal ganglia after 2 days in 1 HSV type 1 and in l HSV type 2 strain. The infected neurons were scattered in the ganglia. In the spinal cord extensive destruction occurred in mice examined 3.5 days after injection. The destruction involved particularly the ipsilateral anterior and posterior horns and the posterior column, The contralateral spinal ganglia showed no changes and no lesions were seen in the left sciatic nerve. In sections from the foot extensive inflammation occurred in the skin, subcutaneous tissue and muscles. After injection of H R P into the left foot, numerous scattered neurons were labeled with H R P in the ipsilateral spinal ganglia in a similar distribution to that in the infected neurons. Also in the spinal cord HRP-containing neurons occurred in the anterior horn on the ipsilateral side. Animals in all groups of mice inoculated by the corneal route could, on morphological grounds, be suspected of being infected. Of those inoculated into the footpad 4 out of 6 virus strains of both types caused cellular changes suggesting virus infection. Thus, no morphological findings indicating differences in neural spread of the type 1 and type 2 strains studied were obtained. In general there was a good agreement between virological and morphological findings. Comparing samples taken from the ganglia, spinal cords or brain stems, both histological and virological examination demonstrated virus infection in 19 groups of specimens. In 9 virus was isolated but no morphological evidence of virus infection was observed, and in 7 the morphological findings suggested that the tissue was infected but no virus was isolated from these mice. DISCUSSION Previously, it has been stated that one of the biological differences between herpes simplex virus type 1 and type 2 is that the type 2 strains should be more neurotropic following inoculation of mice (Nahmias and Roizman 1973). No such consistent difference between the types was found in the present study. All strains examined showed a relatively rapid invasion of the nervous system. No difference between strains passed several times by intracerebral inoculation and freshly isolated HSV type 1 and type 2 strains were noted. These strains spread in a similar manner both to the trigeminal and spinal ganglia as suggested by observations of symptomatology, demonstration of virus and morphological observations. In a previous study we lound that the most likely route for the spread of HSV in peripheral nerves is via the axons (Kristensson et al. 1971). That such a retrograde intra-axonal transport of both viruses and macromolecules exists, has been shown in many investigations during the last years (Kristensson 1975). This also correlates well with findings of the present study where altered neurons were found in the ipsilateral ganglia. These neurons corresponded to those which had accumulated H R P in the perikarya after the protein tracer had been injected into the point of virus inoculation. Such an accumulation is probably the result of an uptake of H R P in axonal endings at the site
339 of injection followed by a retrograde axonal transport to the nerve cell body. It is therefore plausible that HSV is spread through the capacity of the neurons for uptake and axonal transport. In this way a localized infection in the ganglia can be established with, in the trigeminal ganglion, altered neurons restricted to the medial part (Knotts, Cook and Stevens 1974). Except for ganglion cells showing signs of infection there was not much tissue destruction and the inflammatory cell infiltration was rather sparse. It is possible that the immune response of the animals may restrict further spread of the virus within the ganglion, once it has reached the nerve cell bodies (Walz, Yamamoto and Notkins 1976). It should also be noted that vessels both in spinal and trigeminal ganglia are permeable to protein which should facilitate the passage o f h u m o r a l antibodies to the areas around neurons in the ganglia. This contrasts with the CNS where the so-called blood-brain barrier prevents such a passage (Brightman, Klatzo, Olsson and Reese 1970). This is of importance since in animals which recover from an infection, the virus may remain latent in the ganglion. In such experiments virus can no longer be demonstrated by fluorescent antibody or routine viral culture techniques, but in explant cultures of ganglia cytopathic effects can be produced. Apparently, the altered milieu of the in vitro culture is sufficient and essential for this effect (Baringer and Swoveland 1974). When activated in the living animal or man the virus must be transported in the orthograde direction of the axon to the periphery to cause eruptions in the innervated areas (Kristensson, Ghetti and Wi~niewski 1974), but it is not clear whether the virus is not also transported into the brain stem via the axonal branch in the root during these circumstances. In the present experiments the PNS was relatively well preserved but an extensive tissue destruction occurred after spread to the CNS. This could be especially well studied in the transitional region of the trigeminal root where CNS and PNS tissue directly meet. Such a well-preserved PNS directly bordering upon damaged CNS has also been noted previously by Townsend and Baringer (1976) in HSV-infected rabbits. The reason for this difference in reactivity is not known. The blood-brain barrier which is present in the CNS may prevent circulating antibodies from reducing extracellular virus spread early during infection. In the CNS a number of astrocytes are infected, which may react more severely to the virus. Infection of 1 oligodendroglial cell may cause more widespread changes than infection of 1 Schwann cell, since the former myelinate several internodes in different axons. Furthermore, CNS myelin appears to be much more sensitive to a variety of noxious agents, which in CNS myelin can cause cytotoxic edema but not in PNS (Scheinberg, Taylor, Herzog and Mandell 1966). These studies therefore show that CNS tissue may respond quite differently from that of the PNS to a virus infection. This may be a factor of importance in understanding how a virus such as HSV, causing inapparent and latent infections of peripheral nerves and ganglia when reaching the CNS, can cause a very severe necrotizing infection in man.
340 ACKNOWLEDGEMENTS T h i s s t u d y was s u p p o r t e d by g r a n t s f r o m the S w e d i s h M e d i c a l R e s e a r c h C o u n c i l N o . B 7 8 - 1 2 X - 0 4 4 8 0 - 0 4 A a n d -4514 and by R i k s f 6 r e n i n g e n f6r T r a f i k - o c h Polioskadade.
REFERENCES Arvidsson, J. (1975) Location of cat trigeminal ganglion cells innervating dental pulp of upper and lower canines studied by retrograde transport of horseradish peroxidase, Brain Res., 99 : 135-139. Baringer, J. R. (1974) Human herpes simplex virus infections, Advanc. Neurol., 6 : 41-51. Baringer, J. R. and P. Swoveland (1974) Persistent herpes simplex virus infection in rabbit trigemmal ganglia, Lab. Invest., 30: 230-240. Brightman, M. W., I. Klatzo, Y. Olsson and T_ S. Reese (1970) The blood-brain barrier to proteins under normal and pathological conditions, J. neurol. Sci., 10: 215-239. Carlstedt, T. (1977) Observations on the morphology at the transition between the peripheral and central nervous system in the cat. I. A preparative procedure useful for electron mtcroscopy oF the lumbosacral dorsal rootlets, Acta physiol, scand., Suppl. 446: 5-21. Cook, M. L. and J. G. Stevens (1973) Pathogenesis of herpetic neuritis and ganglionitis in mice: evidence for intraaxonal transport of infection, Inject. lmmunoL, 7: 272-288. Graham, R. C. and M. J. Karnovsky (1966) The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney_ Ultrastructural correlates by a new technique, J. Histochem. Cytochern., 14: 291-302. Jeansson, S. (1972) Differentiation between herpes simplex virus type 1 and type 2 strains by Jmmunoelectroosmophoresis, Appl. Microbiol., 24: 96-100. Kieff, E., B. Hoyer, S. Bachenheimer and B. Roizman (1972) Genetic relatedness of type 1 and type 2 herpes simplex viruses, J. Virol_, 9: 738-745. Knotts, F. B., M. L. Cook and J. G. Stevens (1974) Pathogenesis of herpetic encephalitis in mice after ophthalmic inoculation, J. infect. Dis., 130: 16-27. Kristensson, K. (1975) Retrograde axonal transport of protein tracers. In: W. M. Cowan and M. Cu6nod (Eds.), The Use of Axonal Transport for Studies of Neuronal Connectivity, Elsevier, Amsterdam, pp. 69-82. Kristensson, K., E. Lycke and J. Sj6strand (1971) Spread of herpes simplex wrus in peripheral nerves, Acta neuropath. (BerL), 17: 44-53. Kristensson, K., B. Ghetti and H. M. Wi~niewski (1974) Study of the propagation of herpes simplex virus (type 2) into the brain after intraocular injection, Brain Res., 69: 189-201. Ludwig, H. O., N. Biswal and M. Benyesh-Melnick (1972) Studies on the relatedness of herpes viruses through D N A - D N A hybridization, Virology, 49: 95-101. Nahmias, A. J. and B. Roizman (1973) Infection with herpes simplex viruses 1 and 2, New Engl. J. Med., 289: 667-674. Scheinberg, L. C., J. M. Taylor, I. Herzog and S. Mandell (1966) Optic and peripheral nerve response to triethyltin intoxication in the rabbit - Biochemical and ultrastructural studies, J. Neuropath. exp. Neurol., 25 : 202-213. Stevens, J. G. and M. L. Cook (1973) Latent herpes simplex virus in sensory ganglia, Perspect. ViroL, 8:171-187 Townsend, J. J. and J. R. Baringer (1976) Comparative vulnerability of peripheral and central nervous tissue to herpes simplex virus, J. Neuropath. exp. Neurol., 35 : 100. Vahlne, A., J. Blomberg, S. Olofsson and E. Lycke (1975) Subtyping of herpes simplex virus, Acta path. rnicrobiol, scand., 83: 506-512. Walz, M. A., H. Yamamoto and A. L. Notkins (1976) Immunological response restricts number of ceils in sensory ganglia infected with herpes simplex virus, Nature (Lond.), 264: 554-556.
331
NEURAL SPREAD OF HERPES SIMPLEX VIRUS TYPES I A N D 2 IN MICE AFTER CORNEAL OR SUBCUTANEOUS (FOOTPAD) INOCULATION
K. KRISTENSSON, A. V A H L N E , L. A. PERSSON and E. LYCKE
Department of Virology, Institute of Medical Microbiology, University of Gbteborg, Gbteborg, and Neuropathological Laboratory, Department of Pathology II, University of Link6ping, Linkgping (Sweden) (Received 28 July, 1977)
SUMMARY
Twelve herpes simplex virus (HSV) strains, 6 of each type, were inoculated subcutaneously into the left hind foot and into the cornea of the right eye of 12-day-old Swiss albino mice. The neural spread of virus to trigeminal and spinal ganglia and to brain and spinal cord was studied by demonstration of infective virus, histology and electron microscopy. Type 1 and type 2 infections seemed to spread equally well by intra-axonal transport. Using a protein tracer (horseradish peroxidase) injected into the same site as the virus, transport of the tracer to neurons corresponding to those infected with virus was observed. The extensive destruction of CNS tissue in the transitional region of the trigeminal root where CNS and the peripheral nervous system meet is discussed with reference to the pathogenesis of HSV encephalitis.
INTRODUCTION
Studies on base sequences of herpes simplex virus (HSV) type 1 and 2 genomes have revealed that the two subtypes differ by as much as 50 ~ (Kieff, Hoyer, Bachenheimer and Roizman 1972; Ludwig, Biswal and Benyesh-Melnick 1972). In spite of the genetic differences, pronounced biological subtype characteristics are difficult to find. Both types seem to cause latent reactivable infections in ganglion cells (Stevens and Cook 1973; Baringer and Swoveland 1974); and may thus invade the central nervous system by intra-axonal transport (Kristensson, Lycke and Sj6strand 1971; Cook and Stevens 1973). However, in adults HSV encephalitis is prevailingly caused by type 1 infections (Baringer 1974). The present study reports observations on transport of both types of HSV to trigeminal and spinal ganglia.
332 MATERIAL AND METHODS
Virus Altogether 12 strains, 6 of each type of HSV, were studied. Of these 3 (F, MP and G) were well-known strains supplied by Dr. Bernhard Roizman, Chicago, 111. Another 5 were thoroughly characterized in our laboratory using serology, thymidine kinase stability, sensitivity to thymidine and silver nitrate, and ability of strains to induce liver necrosis in mice as the discriminating tests (Vahine, Blomberg, Olofsson and Lycke 1975). These 8 strains were all passed in mice before use. Four strains were freshly isolated from patients and not passed in mice before testing. They were typed by their ability to induce liver necrosis in mice and by immunoelectro-osmophoresis against type-specific antiserum (Jeansson 1972). Batches of each of the strains were produced in green monkey kidney AH-1 (GMK) cells, dispersed in 1 ml aliquots and kept frozen a t - - 7 0 °C until used. Each batch of virus was titrated in G M K cells and by intracerebral inoculation of 3-week-old Swiss albino mice of our own laboratory breed. The pfu/LDs0 ratios were determined. These ratios were about 1 for all the strains used. Before inoculation the strains were diluted in Hanks' BSS so that about 1000 pfu were inoculated of each type.
Inoculation procedure Of each of the 12 virus strains 0.02 ml was injected subcutaneously into the left hind footpad of groups of 12 12-day-old anesthetized suckling Swiss albino mice. From these groups 2 mice were taken 2, 3 and 4 days after the virus inoculation for virus titrations and 2 mice after 2 and 3.5 days for histological examination. Samples were taken from the left foot, left sciatic nerve, left and right lumbar spinal ganglia, lumbar spinal cord and left cerebral hemisphere. In addition, in 6 groups of 12 mice, a drop of virus suspension from 3 type 1 and 3 type 2 strains was placed on to the cornea of the right eye which was scratched several times with an injection needle. The trigeminal ganglia on both sides and pieces from the brain stem and left cerebral hemisphere were sampled from 2 mice after 2 and 3 days for virus titrations and from 2 mice after 2 and 3.5 days for histological examinations. Blood samples were also taken for virus titration. Virus titrations For titration of virus, each specimen was ground with a glass pestle and suspended in 1 ml cell culture medium and inoculated undiluted and in 10-fold dilutions, respectively, into 2 cultures of G M K cells, 0.5 ml per culture. The cells were cultivated in 5 cm plastic dishes using Eagle's MEM supplemented with 2 ~o bovine serum and antibiotics. The medium of the inoculated cultures was replaced by cell culture medium containing 1 ~ methyl cellulose. After inoculation at 37 °C for 4 days the number of plaques was counted. Histology and electron microscopy For morphological observations, the mice were perfused with 2.5 ~ glutaraldehyde in phosphate buffer or with 5 ~ glutaraldehyde in Millonig's buffer with 2.7
333 low molecular dextran and 6.85 % sucrose to provide optimal fixation of central myelin sheaths (Carlstedt 1977). The specimens were embedded in E p o n , 1/zm thick sections were cut and stained with toluidine blue. Thin sections from selected areas were stained with uranyl acetate and lead citrate and examined in an Elmiskop. For demonstration of direct pathways from the periphery to the nerve cell bodies the protein tracer horseradish peroxidase (HRP) (type II, Sigma) was used. In 1 group of mice 0.02 ml of 50% H R P in Ringer solution was injected into the left footpad. In another group of mice, the right cornea was scratched several times with a needle, after which 0.2 mg H R P in dry powder was directly applied to the scarified cornea and the eyelid closed with a suture. Twenty-four hours later the mice were perfused with 2.5 % glutaraldehyde in phosphate buffer. Frozen sections, 30--40/~m thick, from lumbar spinal ganglia and cord, and from trigeminal ganglia were cut and incubated with 3,3'-diaminobenzidine and hydrogen peroxide for demonstration of peroxidase activity (Graham and Karnovsky 1966). RESULTS
Course of infection The spread of virus was judged on the appearance of symptoms of infection of the CNS and spinal cord, on recovery of virus from specimens collected and on morphological findings. After corneal inoculation all mice developed symptoms of ipsilateral conjunctivitis. By the third day neurological signs and deaths were observed and 4 days after inoculation all animals remaining, except those inoculated with type 1 strain F, were moribund. Mice inoculated into the left hind foot developed on days 2 and 3 after inoculation left sided pareses followed by parapareses. All animals became moribund and died before day 5, except those inoculated with strain F in which onset and progress of disease were delayed by 1 day. Otherwise there were no differences between the type 1 and 2 strains with regard to length of incubation period (time of inoculation to appearence of neurological symptoms), symptomatology and time of death.
Virological observations After infection by the corneal route infective virus was demonstrable in the ipsilateral trigeminal ganglion, and in the brain stem but not from corresponding specimens of the uninoculated side. One group of animals infected with a type 2 strain had virus in the ganglion but not in the brain stem and no virus was found in animals inoculated with strain F (type 1). In mice inoculated into the left foot infective virus was demonstrable in the inoculated foot of all the animals, but from the sciatic nerve in only 2 groups of type 1 inoculated mice and in 1 inoculated with type 2. Four of the 6 type 1 strains were demonstrable in the left spinal ganglia while this was the case with 2 of the type 2 strains. From all mice, virus isolations were made from spinal cord. On the other hand, no infective virus was demonstrated in the right spinal ganglion or in the blood specimens. Thus, there were indications of neural spread of virus to trigeminal and spinal
334 Ioa no pfu -10 6-
I
4-
2-
ot 2
3
t, days
Fig. 1. Amounts of virus in spinal cords of mice inoculated subcutaneously in left hind footpad. The spinal cord specimens were examined on days 2, 3 and 4 after virus inoculation. Means in log no. pfu per spinal cord of HSV type 1 (open circles) and type 2 inoculated mice (filled circles) are plotted against time in days. Bars denote ranges of titers. ganglia as well as to brain stem and spinal cord. One type 1 strain seemed less neuroinvasive than the others, but as a whole there were no significant differences in spread of the type 1 and type 2 strains. Varying amounts of virus were detected, which might partly be explained by differences in the amounts of tissue available of the different specimens examined. In Fig. 1 the virus titers obtained with the spinal cord specimens of footpad inoculated mice are plotted against time in days after inoculation. It was also observed that with the spinal cord specimens - - which did not markedly differ in size - - the amounts of virus demonstrable varied considerably. High virus titers were observed in both type 1 and type 2 virus-inoculated mice.
Morphological observations After corneal inoculation infected nerve cells were seen in the ipsilateral trigeminal ganglion 2 days after inoculation of 2 of the strains. The neurons showed intranuclear inclusion bodies of Cowdry type A or had a swollen homogenous nucleus and displayed signs of chromatolysis or degeneration. Such neurons were scattered within the medial part of the ganglion (Fig. 2). After corneal application of HRP, scattered neurons containing the reaction product were seen within the same area (Fig. 3P). This area corresponds to neurons innervating the cornea (Arvidsson 1975). After 3.5 days, when several of the mice were moribund, many infected neurons were seen in the medial part of the ganglion. The altered neurons were confined to this area and there was no spread into areas corresponding to the maxillar and mandibular branches of the ganglion (Fig. 4). After 3.5 days there was also extensive degeneration of the trigeminal nuclei in the brain stem and widespread tissue destruction. Interesting findings were made at the transitional region of the trigeminal root, where tissue from the CNS meets the peripheral nervous system (PNS). The PNS part of the ~oots was well preserved and showed no light-microscopic changes. The CNS tissue, on the contrary, showed extensive changes beginning directly at the border. There was a marked sponginess and destruction of the CNS myelin, degeneration of glial cells and infiltration of mononuclear inflammatory cells (Fig. 5). Ultrastructurally, the transition from PNS myelin to CNS myelin generally occurs at a node of Ranvier. Often a nerve fiber myelinated in the PNS part was naked in the CNS (Fig~ 6), but it
335
Fig. 2_ Chromatolytic neurons with swollen, homogenous nucleus in ipsilateral trigeminal ganglion 2 days after inoculation to scarified cornea, x 735. Fig. 3. HRP-containing neurons in the ipsilateral trigeminal ganglion after application to scarified cornea. Darkfield condenser, x 775. Fig. 4. Extensive infection of most neurons in medial part of trigeminal ganglion (M) 3.5 days after injection. Lateral part (L) is not involved, x 375.
336
Fig_ 5. Trigeminal root 3.5 days after HSV inoculation into the Jpsdateral cornea. Extensive changes in CNS part, while the PNS looks normal. ' 1520. Fig. 6. Nerve fiber at the transitional region of the trigemmal root showing well-developed PNS myelin, while the CNS part of the axon is naked. Note also glial cells with herpes nucleocapsids in their nuclei. "< 7980.
337
Fig. 7. Nerve fiber at the trigeminal transitional region with splitting and degeneration of the CNS myelin while the PNS myelin is intact, x 7980. Fig. 8. Schwann cell close to the transitional region containing herpes virus particles. The myelin sheaths and axons are well preserved, x /8,616_
should be noted that at this stage o f development the contralateral r o o t still contains a few well-myelinated P N S fibers which are transformed into non-myelinated C N S fibers. In several nerve fibers morphologically well-preserved P N S myelin could be seen directly bordering the C N S myelin, which showed vacuolation due to separation from the axolemma or splitting at interperiod lines (Fig. 7). Virus particles were f o u n d both in Schwann cells in the PNS part o f the zone and in astrocytes and oligodendroglial-like cells in the C N S part (Figs. 6 and 8). In the trigeminal ganglion herpes virus
338 particles were found in neurons seen by light microscopy to be abnormal and in satellite cells. After injection into the left foot infected neurons were seen in the ipsilateral spinal ganglia after 2 days in 1 HSV type 1 and in l HSV type 2 strain. The infected neurons were scattered in the ganglia. In the spinal cord extensive destruction occurred in mice examined 3.5 days after injection. The destruction involved particularly the ipsilateral anterior and posterior horns and the posterior column, The contralateral spinal ganglia showed no changes and no lesions were seen in the left sciatic nerve. In sections from the foot extensive inflammation occurred in the skin, subcutaneous tissue and muscles. After injection of H R P into the left foot, numerous scattered neurons were labeled with H R P in the ipsilateral spinal ganglia in a similar distribution to that in the infected neurons. Also in the spinal cord HRP-containing neurons occurred in the anterior horn on the ipsilateral side. Animals in all groups of mice inoculated by the corneal route could, on morphological grounds, be suspected of being infected. Of those inoculated into the footpad 4 out of 6 virus strains of both types caused cellular changes suggesting virus infection. Thus, no morphological findings indicating differences in neural spread of the type 1 and type 2 strains studied were obtained. In general there was a good agreement between virological and morphological findings. Comparing samples taken from the ganglia, spinal cords or brain stems, both histological and virological examination demonstrated virus infection in 19 groups of specimens. In 9 virus was isolated but no morphological evidence of virus infection was observed, and in 7 the morphological findings suggested that the tissue was infected but no virus was isolated from these mice. DISCUSSION Previously, it has been stated that one of the biological differences between herpes simplex virus type 1 and type 2 is that the type 2 strains should be more neurotropic following inoculation of mice (Nahmias and Roizman 1973). No such consistent difference between the types was found in the present study. All strains examined showed a relatively rapid invasion of the nervous system. No difference between strains passed several times by intracerebral inoculation and freshly isolated HSV type 1 and type 2 strains were noted. These strains spread in a similar manner both to the trigeminal and spinal ganglia as suggested by observations of symptomatology, demonstration of virus and morphological observations. In a previous study we lound that the most likely route for the spread of HSV in peripheral nerves is via the axons (Kristensson et al. 1971). That such a retrograde intra-axonal transport of both viruses and macromolecules exists, has been shown in many investigations during the last years (Kristensson 1975). This also correlates well with findings of the present study where altered neurons were found in the ipsilateral ganglia. These neurons corresponded to those which had accumulated H R P in the perikarya after the protein tracer had been injected into the point of virus inoculation. Such an accumulation is probably the result of an uptake of H R P in axonal endings at the site
339 of injection followed by a retrograde axonal transport to the nerve cell body. It is therefore plausible that HSV is spread through the capacity of the neurons for uptake and axonal transport. In this way a localized infection in the ganglia can be established with, in the trigeminal ganglion, altered neurons restricted to the medial part (Knotts, Cook and Stevens 1974). Except for ganglion cells showing signs of infection there was not much tissue destruction and the inflammatory cell infiltration was rather sparse. It is possible that the immune response of the animals may restrict further spread of the virus within the ganglion, once it has reached the nerve cell bodies (Walz, Yamamoto and Notkins 1976). It should also be noted that vessels both in spinal and trigeminal ganglia are permeable to protein which should facilitate the passage o f h u m o r a l antibodies to the areas around neurons in the ganglia. This contrasts with the CNS where the so-called blood-brain barrier prevents such a passage (Brightman, Klatzo, Olsson and Reese 1970). This is of importance since in animals which recover from an infection, the virus may remain latent in the ganglion. In such experiments virus can no longer be demonstrated by fluorescent antibody or routine viral culture techniques, but in explant cultures of ganglia cytopathic effects can be produced. Apparently, the altered milieu of the in vitro culture is sufficient and essential for this effect (Baringer and Swoveland 1974). When activated in the living animal or man the virus must be transported in the orthograde direction of the axon to the periphery to cause eruptions in the innervated areas (Kristensson, Ghetti and Wi~niewski 1974), but it is not clear whether the virus is not also transported into the brain stem via the axonal branch in the root during these circumstances. In the present experiments the PNS was relatively well preserved but an extensive tissue destruction occurred after spread to the CNS. This could be especially well studied in the transitional region of the trigeminal root where CNS and PNS tissue directly meet. Such a well-preserved PNS directly bordering upon damaged CNS has also been noted previously by Townsend and Baringer (1976) in HSV-infected rabbits. The reason for this difference in reactivity is not known. The blood-brain barrier which is present in the CNS may prevent circulating antibodies from reducing extracellular virus spread early during infection. In the CNS a number of astrocytes are infected, which may react more severely to the virus. Infection of 1 oligodendroglial cell may cause more widespread changes than infection of 1 Schwann cell, since the former myelinate several internodes in different axons. Furthermore, CNS myelin appears to be much more sensitive to a variety of noxious agents, which in CNS myelin can cause cytotoxic edema but not in PNS (Scheinberg, Taylor, Herzog and Mandell 1966). These studies therefore show that CNS tissue may respond quite differently from that of the PNS to a virus infection. This may be a factor of importance in understanding how a virus such as HSV, causing inapparent and latent infections of peripheral nerves and ganglia when reaching the CNS, can cause a very severe necrotizing infection in man.
340 ACKNOWLEDGEMENTS T h i s s t u d y was s u p p o r t e d by g r a n t s f r o m the S w e d i s h M e d i c a l R e s e a r c h C o u n c i l N o . B 7 8 - 1 2 X - 0 4 4 8 0 - 0 4 A a n d -4514 and by R i k s f 6 r e n i n g e n f6r T r a f i k - o c h Polioskadade.
REFERENCES Arvidsson, J. (1975) Location of cat trigeminal ganglion cells innervating dental pulp of upper and lower canines studied by retrograde transport of horseradish peroxidase, Brain Res., 99 : 135-139. Baringer, J. R. (1974) Human herpes simplex virus infections, Advanc. Neurol., 6 : 41-51. Baringer, J. R. and P. Swoveland (1974) Persistent herpes simplex virus infection in rabbit trigemmal ganglia, Lab. Invest., 30: 230-240. Brightman, M. W., I. Klatzo, Y. Olsson and T_ S. Reese (1970) The blood-brain barrier to proteins under normal and pathological conditions, J. neurol. Sci., 10: 215-239. Carlstedt, T. (1977) Observations on the morphology at the transition between the peripheral and central nervous system in the cat. I. A preparative procedure useful for electron mtcroscopy oF the lumbosacral dorsal rootlets, Acta physiol, scand., Suppl. 446: 5-21. Cook, M. L. and J. G. Stevens (1973) Pathogenesis of herpetic neuritis and ganglionitis in mice: evidence for intraaxonal transport of infection, Inject. lmmunoL, 7: 272-288. Graham, R. C. and M. J. Karnovsky (1966) The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney_ Ultrastructural correlates by a new technique, J. Histochem. Cytochern., 14: 291-302. Jeansson, S. (1972) Differentiation between herpes simplex virus type 1 and type 2 strains by Jmmunoelectroosmophoresis, Appl. Microbiol., 24: 96-100. Kieff, E., B. Hoyer, S. Bachenheimer and B. Roizman (1972) Genetic relatedness of type 1 and type 2 herpes simplex viruses, J. Virol_, 9: 738-745. Knotts, F. B., M. L. Cook and J. G. Stevens (1974) Pathogenesis of herpetic encephalitis in mice after ophthalmic inoculation, J. infect. Dis., 130: 16-27. Kristensson, K. (1975) Retrograde axonal transport of protein tracers. In: W. M. Cowan and M. Cu6nod (Eds.), The Use of Axonal Transport for Studies of Neuronal Connectivity, Elsevier, Amsterdam, pp. 69-82. Kristensson, K., E. Lycke and J. Sj6strand (1971) Spread of herpes simplex wrus in peripheral nerves, Acta neuropath. (BerL), 17: 44-53. Kristensson, K., B. Ghetti and H. M. Wi~niewski (1974) Study of the propagation of herpes simplex virus (type 2) into the brain after intraocular injection, Brain Res., 69: 189-201. Ludwig, H. O., N. Biswal and M. Benyesh-Melnick (1972) Studies on the relatedness of herpes viruses through D N A - D N A hybridization, Virology, 49: 95-101. Nahmias, A. J. and B. Roizman (1973) Infection with herpes simplex viruses 1 and 2, New Engl. J. Med., 289: 667-674. Scheinberg, L. C., J. M. Taylor, I. Herzog and S. Mandell (1966) Optic and peripheral nerve response to triethyltin intoxication in the rabbit - Biochemical and ultrastructural studies, J. Neuropath. exp. Neurol., 25 : 202-213. Stevens, J. G. and M. L. Cook (1973) Latent herpes simplex virus in sensory ganglia, Perspect. ViroL, 8:171-187 Townsend, J. J. and J. R. Baringer (1976) Comparative vulnerability of peripheral and central nervous tissue to herpes simplex virus, J. Neuropath. exp. Neurol., 35 : 100. Vahlne, A., J. Blomberg, S. Olofsson and E. Lycke (1975) Subtyping of herpes simplex virus, Acta path. rnicrobiol, scand., 83: 506-512. Walz, M. A., H. Yamamoto and A. L. Notkins (1976) Immunological response restricts number of ceils in sensory ganglia infected with herpes simplex virus, Nature (Lond.), 264: 554-556.