- Email: [email protected]
Immunocytochemical evidence for Semliki Forest virus antigen persistence in mouse brain
Journal of the Neurological Sciences, 1988, 85:17-26
Elsevier
17
JNS 02988
Immunocytochemical evidence for Semliki Forest virus antigen persistence in mouse brain A. Khalili-Shirazi, N. Gregson and H . E . Webb The United Medical and Dental Schools of Guy's 'and St. Thomas" Hospital, London (U.K.)
(Received 31 August, 1987) (Revised, received23 December, 1987) (Accepted 23 December, 1987)
SUMMARY Semliki Forest virus (SFV) is neurotropic in mice. Mature virulent virus (strain L10) can be identified within the CNS by electron microscopy in adult mice. Inspite of high virus titres, avirulent SFV A7(74) cannot be visualised in the brain of adult mice. Immunocytochemical studies using monoclonal antibodies (MAbs) to A7(74) E 1 and E 2 proteins and viral envelope glycolipids, showed viral E1 to be labelled in the cerebral capillaries, the E2 and the putative envelope glycolipids were labelled in the cytoplasm of neurons, particularly in the hippocampal areas and glia in the cerebellum. By double labelling the presence of viral antigens in astrocytes and oligodendrocytes was demonstrated. Viral antigens were identified in the brain up to 183 days after infection. Paraffin sections from Bouin-fixed tissue were found to be the most suitable material for immunocytochemistry of SFV. The presence of life-long anti SFV antibody in the sera of animals after SFV infection, could be due to the persistence of viral antigens acting as constant stimuli to the immune system.
Key words: Semliki Forest virus persistence; Demyelination; Immunocytochemistry; Monoclonal antibodies
Correspondence to: Dr. A. Khalili-Shirazi,NeurovirologyUnit, The Rayne Institute, St. Thomas' Hospital, LondonSE1 7EH, U.K.
0022-510X/88/$03.50 © 1988Elsevier Science Publishers B.V.(BiomedicalDivision)
18 INTRODUCTION Semliki Forest virus (SFV) is an enveloped alphavirus of the Togaviridae. Two strains of this virus, used in this study, are L10 and A7(74). These 2 strains are serologicaUy identical. However, strain A7(74) kills young suckling mice aged 0-14 days but does not kill adult mice whereas L10 kills mice irrespective of age. EM studies have shown that the maturation of virus occurs in baby mice infected with either L10 or A7(74), whereas in adult mice only L10 appears to reach maturation (Pathak and Webb 1976). The clinical course following SFV infection in adult mice depends upon the strain of the virus used. Intraperitoneal (i.p.) infection of adult mice with A7(74) results in occasional paralysis, which is most apparent at post-infection day (PID) 7. Within 20 min of inoculation with A7(74), virus is detectable in the blood (Flemming 1977) but is later cleared by neutralising antibodies, which are detectable from PID 3 (Allner et al. 1974) or PID 4 (Jagelman et al. 1978). The anti-viral IgG antibodies have been measured in both serum and cerebrospinal fluid from PID 8 onwards (Suckling etal. 1982; Parsons and Webb 1984), and they remain high throughout the animal's life. The virus replicates to high titre in the central nervous system (CNS), with the maximum titre apparently at PID 4; after PID 12, no infectious virus could be isolated by conventional methods. However, if the animals were made immunodeficient by irradiation (Fazakerley and Webb 1987a) or if athymic immunodeficient nude mice were used (Fazakerley and Webb 1987b) infectious virus could be reisolated up to 3-4 weeks post-infection. Having raised a panel of MAbs to SFV A7(74) and partially characterised some of them in terms of antigenic specificity (Khalili-Shirazi et al. 1986), these MAbs were used in this study to demonstrate viral antigens within the brains of infected animals, in both athymic nude mice nu/nu and their immunocompetent nu/+ litter mates, up to 183 days post-infection. The problem posed for this study relates to a neurological disease. With A7(74) strain, there is demyelination by PID 14 when theoretically virus should have been cleared, and yet we fred antigen in a variety of cells. However, where antigen is seen does not necessarily correspond to full viral genome expression, but may be partial. Obviously use of RNA probes would to some extent answer this problem. This could then form the basis for continuation of the work to prepare a C-DNA library with probes to look for expression; this could also be used in looking at strain differences.
METHODS A group of 20 nu/nu (BALb/c origin) and nu/+ mice were inoculated i.p. with SFV 103.5 ICLDso. These mice were killed at days post-infection (p.i.) 6, 14, 21, 28, 68 and 183. The athymic nude mice were only studied for 40 days due to an unexpected high mortality rate. Their immunocompetcnt littermates were studied up to days 183 p.i. The brains were taken out and both fresh frozen and fixed paraffin sections were examined against a variety of fixatives to fred the most suitable one for this study. For the immunocytochemistry technique the method of Heyderman 0979)was
19 used, with the exception that after acid haematin was bleached by 6 ~ hydrogen peroxide, the endogenous peroxidase activity was not blocked by periodic acid. The addition of periodic acid, a deglycosilatingagent, resulted in decreased labelling. Unfixed frozen sections proved to be difficult to interpret due to too high a background. Frozen sections post-fixed with 10% formaldehyde followed by methanol, or post-fixed with acetone proved negative. Sections cut from paraffin blocks gave the best results, but the method of fixation used was crucial. Bouin's and to a lesser extent paraformaldehyde (and perhaps glutaraldehyde) were found to be the most suitable fLxatives, whereas formalin, Carnoy's and Methacon fixation resulted in poor and variable staining. To demonstrate the presence of antigen in oligodendrocytes and astrocytes double labelling of these cells was used. Sections were incubated with either rabbit anti-carbonic anhydrase (raised in New Zealand white rabbits against purified rat erythrocyte CA2) or rabbit anti-GFAP (Dako Ltd., Bucks, U.K.) followed by peroxidase conjugate and developed with DAB substrate. This was followed by incubation with either MAb 307 (anti-E2) or 308 (anti-envelope glycolipid) and subsequently an alkaline phosphatase conjugate, followed by development with Naphthol As-Mx phosphate and fast blue. To verify the specificity of the staining, irrelevant IgG and IgM as well as preimmune sera controls for infected brain, as compared to normal brain, were used.
RESULTS The antigen labelled with MAbs 307 (anti=E2) and 308 (anti-envelope glycolipid) appeared as clusters of granules of variable size, found in the cytoplasm of representative cells of all types scattered throughout the brain. This staining was specific, as the staining with 307 could be removed by pre-incubation of the antibody with A7(74) (Fig. 1). Absorption of 308 activity was not tried. However, two brain structures appeared to show most consistently the presence of antigen; a) the hippocampus, in which the pyramidal neurons of the cornu Ammonis frequently contained SFV antigen (Fig. 2); b) the cerebellar cortex where immunoperoxidase staining was most frequently seen in what appear to be the perikarya of the Bergmann glia (Fig. 3). The other cells of the cerebellum such as the Purkinje cells, were only occasionally stained and then most commonly in the early stages of the infection. During the early phase of the infection, when intracerebral inflammation could be found, SFV antigen was detected in perivascular inflammatorycells (Fig. 4). Some GFAP=positive astrocytes also showed particulate staining with MAbs 307 and 308 which could be found up to day 183 p.i. (Fig. 5). SFV antigen was occasionally seen in carbonic anhydrase-positive oligodendrocytes. No staining was observed with 308 in non-neural or non-CNS SFVinfected tissues, whereas the anti=E2 antibody 307, was positive (Illavia and Webb 1988). Labelling of viral antigens with mouse SFV ( + )re polyclonal antibody was also successful. The level of antigen remained high all along in the nude mice up to days 40 p.i. (Table 1). The same pattern of labelling was also observed in animals infected with L10 at days 3-4 p.i.
20
Fig. 1. (a) Demonstration of staining of E~ antigen with MAb 307 in the cerebellum. (h) No staining was observed if M A b 307 was absorbed with A7(74) prior to use. Monochromatic filter set in the range of 448-512 nm was used here. G = granule neurons; P = Purkinje cells; B = Bergmann glia.
21
Fig. 2. SFV antigen localised in hippocampal neurons. A774 mice 40 days post-inoculation. (a) Staining of CA neurons with MAb 308. (b) Lower power demonstration of the localisation of SFV E2 with MAb 307, predominantly in CA neurons (CA), rather than in the dentate (D).
22
Fig. 3. SFV localisation in Bergmann giia in the cerebellum. Consistent, and large amounts of viral antigen were only detected in what appear to be the perflcaryal regions of the Bergmann glia (B), The Purkinje cells (P) and granule neurons (G) only react with anti-SFV MAbs occasionally and early after infection. (a) MAb 308 staining of cerebellum 7 days p.i. with L10 strain of SFV. This is to show, similar envelope glycolipid staining of both L10 and A774 strain. (b) 40 days p.i. with A774 using MAb 308.
23
Fig. 4. Demonstration of SFV in perivascular cells during the inflammatory stage of the disease. Perivascular cuffing of a vessel in cerebellar white matter, 6 days p.i.. Localisation of SFV with MAb 308 and indirect immunoperoxidase. TABLE 1 SUMMARY OF PATHOLOGY, CIRCULATING ANTIBODY LEVELS AND IMMUNOCYTOCHEMISTRY RESULTS AFTER SFV INFECTION IN nu/nu AND nu/+ MICE Brains from immunodeficient athymic nude mice (nu/nu) and their immunocompetent littermates (nu/+ ) were taken at days post-infection (p.i.) indicated. Circulating antibody levels had previously been investigated (Suckling et al. 1982; Parsons and Webb 1984). These brains were examined for pathology and sections were taken for immunocytochemistry and examined against MAbs 308, 307, 302 and polyclonal mouse anti-SFV sera. Pre-immune sera and irrelevant MAbs were used as controls. Results obtained with Swiss A2G mice were comparable with those of nu/+ mice. Day p.i.
6
14 28 40 68 183
Mouse
Serum anti-SFV Igs (rag/100 ml serum)
nu/ + nu/nu
1 00
nu/+ nu/nu nu/+ nu/nu nu/+ nu/nu nu/+ nu/+
304.5 54.5 366 54 360 45 260 IgG + (ELISA) (actual levels not available)
Pathology
~" J
23
Immunocytochemistry (MAbs)
302,307,308 ( + ve)
+++ ++ + + -
307,308(+ve)
24
Fig. 5. Astrocyte (A), the only cell with cytoplasm G F A P ( + )ve staining, was also ( + )ve for SFV antigen (appearing as blue dots) with MAb 308.
DISCUSSION
In this study we have shown that SFV antigens persist in mouse brain for a very long time after infection. Their presence within the astrocytes particularly could be important, with regard to the A7(74) immune-mediated demyelination (Fazakerley and Webb 1987) and the report that astrocytes can act as antigen presenting cells within the CNS in context of an appropriate MHC expression (Massa et al. 1986). MAb 308 raised to SFV has previously been shown by immuno-thin layer chromatography to be directed to a CNS glycolipid which also occurs in the viral envelope (Khalili-Shirazi et al. 1986). However, with immunocytochemistry,both at the light and EM level, this antibody appears to label the glycolipid in the SFV envelope, but not in the normal host CNS cell membrane. This may be due to conformational or concentration differences in the glycolipid in the viral envelope, Also, there may be loss of the
25 lipid from the cell m e m b r a n e during processing. Viral antigens were labelled equally well in our study, irrespective o f using the virulent (L10) or avirulent (A774) strain o f SFV. In conclusion, with the use of 2 monoclonal antibodies, one reactive with the viral spike protein E 2 and the other with a glycolipid c o m p o n e n t o f the envelope, it has been found possible to stain intraceUular particles in the brains of mice infected with either A7(74) or the L10 strain o f SFV. These same antibodies failed to stain brain tissue from normal non-infected animals. Previously, it had not been demonstrated by immunocytochemistry that SFV antigens persist for as long as 3 months. After absorption of M A b 307 with virus-containing tissue, no staining o f virus-infected brain could be achieved. It is proposed that the staining seen represents the presence o f virus in certain cells and the continued expression of at least part o f its genome within those cells. The observation that the humoral immune response persists long after the apparent clearance of SFV (Suckling et al. 1982; Parsons and W e b b 1984) would also suggest that there is a persistent expression and release o f part or the whole o f the SFV genome.
ACKNOWLEDGEMENTS This work was supported by St. T h o m a s ' Charitable Funds, the Wellcome Trust, the Philip Fleming Charitable Trust and the Multiple Sclerosis Society o f Great Britain and Northern Ireland.
REFERENCES Allner, K., C.J. Bradish, R. Fitzgeorge and N. Nathanson (1974) Modifications by sodium-thiomalate of the expression of virulence in mice by defined strains of Semliki Forest virus, J. Gen. Virol., 24: 221-228. Atkinson, T., A.D.T. Barrett, A. MacKenzie and N.J. Dimmock (1986) Persistence of virulent Semliki Forest virus in mouse brain followingco-inoculation with defective interfering particles,,/. Gen. ViroL, 67:1189-1194. Fazakerley, J.K. and H.E. Webb (1987a) Semliki Forest virus induced, immune mediated demyelination. The effect of irradiation, Br. J. Exp. Pathol., 68: 101-113. Fazakerley, J.K. and H.E. Webb (1987b) Semliki Forest virus induced, immune mediated demyelination - adoptive transfer studies, and viral persistence in nude mice, J. Gen. Virol., 68: 377-385. Fleming, P. (1977) Age-dependent and strain related differences of virulence of Semliki Forest virus in mice, J. Gen. Virol., 37: 93-105. Gates, M.C,, B.J. Sheahan and M.A. O'Sullivan (1985) The pathogenicity of the A7, M9 and L10 strains of SFV for weanling mice and primary mouse cell cultures, J. Gen. Virol., 66 : 2365-2373. Heyderman, E. (1979) Immunoperoxidase technique in histopathology: applications, methods and controls, J. Clin. Pathol., 32: 971-978. Illavia, S.J. and H. E. Webb (1988) The effect of single and multiple inoculations on the peripheral nervous system of mice with the demyelinating strain of Semliki Forest virus, Acta Neurol. Scand., in press. Jagelman, A. J., A.J. Suckling, H. E. Webb and E. T. W. Bowen (1978) The pathogenesis of avirulent Semliki Forest virus infections in athymic nude mice, 3". Gen. Viral., 41: 599-607. Khalili-Shirazi, A., N. Gregson and H. E. Webb (1986) Immunologicalrelationship between a demyelinating RNA enveloped budding virus (Semliki Forest) and brain glycolipids,J. Neurol. Sci., 76: 91-103. Massa, P. T., R. Dorries and V. ter Meulen (1986) Viral particles induce Ia antigen expression on astrocytes, Nature, 320: 543-546.
26 Parsons, L.M. and H. E. Webb (1984) Specific immunoglobulin G in serum and cerebrospinal fluid of mice infected with the demyelinating strain of Semliki Forest virus, Microbios Lett., 25: 135-140. Pathak, S. and H.E. Webb (1976) An electron microscopic study of the development of virulent and avirulent strains of Semliki Forest virus in mouse brain, J. Neurol. Sci., 28: 289-300. Suckling, A.J., S.J. Jagelman and H.E. Webb (1982) Immunoglobulin synthesis in nude (nu/nu) nu + and reconstitued nu/nu mice infected with a demyelinating strain of Semliki Forest virus, Clin. Exp. Immunol., 47: 283-288.
Elsevier
17
JNS 02988
Immunocytochemical evidence for Semliki Forest virus antigen persistence in mouse brain A. Khalili-Shirazi, N. Gregson and H . E . Webb The United Medical and Dental Schools of Guy's 'and St. Thomas" Hospital, London (U.K.)
(Received 31 August, 1987) (Revised, received23 December, 1987) (Accepted 23 December, 1987)
SUMMARY Semliki Forest virus (SFV) is neurotropic in mice. Mature virulent virus (strain L10) can be identified within the CNS by electron microscopy in adult mice. Inspite of high virus titres, avirulent SFV A7(74) cannot be visualised in the brain of adult mice. Immunocytochemical studies using monoclonal antibodies (MAbs) to A7(74) E 1 and E 2 proteins and viral envelope glycolipids, showed viral E1 to be labelled in the cerebral capillaries, the E2 and the putative envelope glycolipids were labelled in the cytoplasm of neurons, particularly in the hippocampal areas and glia in the cerebellum. By double labelling the presence of viral antigens in astrocytes and oligodendrocytes was demonstrated. Viral antigens were identified in the brain up to 183 days after infection. Paraffin sections from Bouin-fixed tissue were found to be the most suitable material for immunocytochemistry of SFV. The presence of life-long anti SFV antibody in the sera of animals after SFV infection, could be due to the persistence of viral antigens acting as constant stimuli to the immune system.
Key words: Semliki Forest virus persistence; Demyelination; Immunocytochemistry; Monoclonal antibodies
Correspondence to: Dr. A. Khalili-Shirazi,NeurovirologyUnit, The Rayne Institute, St. Thomas' Hospital, LondonSE1 7EH, U.K.
0022-510X/88/$03.50 © 1988Elsevier Science Publishers B.V.(BiomedicalDivision)
18 INTRODUCTION Semliki Forest virus (SFV) is an enveloped alphavirus of the Togaviridae. Two strains of this virus, used in this study, are L10 and A7(74). These 2 strains are serologicaUy identical. However, strain A7(74) kills young suckling mice aged 0-14 days but does not kill adult mice whereas L10 kills mice irrespective of age. EM studies have shown that the maturation of virus occurs in baby mice infected with either L10 or A7(74), whereas in adult mice only L10 appears to reach maturation (Pathak and Webb 1976). The clinical course following SFV infection in adult mice depends upon the strain of the virus used. Intraperitoneal (i.p.) infection of adult mice with A7(74) results in occasional paralysis, which is most apparent at post-infection day (PID) 7. Within 20 min of inoculation with A7(74), virus is detectable in the blood (Flemming 1977) but is later cleared by neutralising antibodies, which are detectable from PID 3 (Allner et al. 1974) or PID 4 (Jagelman et al. 1978). The anti-viral IgG antibodies have been measured in both serum and cerebrospinal fluid from PID 8 onwards (Suckling etal. 1982; Parsons and Webb 1984), and they remain high throughout the animal's life. The virus replicates to high titre in the central nervous system (CNS), with the maximum titre apparently at PID 4; after PID 12, no infectious virus could be isolated by conventional methods. However, if the animals were made immunodeficient by irradiation (Fazakerley and Webb 1987a) or if athymic immunodeficient nude mice were used (Fazakerley and Webb 1987b) infectious virus could be reisolated up to 3-4 weeks post-infection. Having raised a panel of MAbs to SFV A7(74) and partially characterised some of them in terms of antigenic specificity (Khalili-Shirazi et al. 1986), these MAbs were used in this study to demonstrate viral antigens within the brains of infected animals, in both athymic nude mice nu/nu and their immunocompetent nu/+ litter mates, up to 183 days post-infection. The problem posed for this study relates to a neurological disease. With A7(74) strain, there is demyelination by PID 14 when theoretically virus should have been cleared, and yet we fred antigen in a variety of cells. However, where antigen is seen does not necessarily correspond to full viral genome expression, but may be partial. Obviously use of RNA probes would to some extent answer this problem. This could then form the basis for continuation of the work to prepare a C-DNA library with probes to look for expression; this could also be used in looking at strain differences.
METHODS A group of 20 nu/nu (BALb/c origin) and nu/+ mice were inoculated i.p. with SFV 103.5 ICLDso. These mice were killed at days post-infection (p.i.) 6, 14, 21, 28, 68 and 183. The athymic nude mice were only studied for 40 days due to an unexpected high mortality rate. Their immunocompetcnt littermates were studied up to days 183 p.i. The brains were taken out and both fresh frozen and fixed paraffin sections were examined against a variety of fixatives to fred the most suitable one for this study. For the immunocytochemistry technique the method of Heyderman 0979)was
19 used, with the exception that after acid haematin was bleached by 6 ~ hydrogen peroxide, the endogenous peroxidase activity was not blocked by periodic acid. The addition of periodic acid, a deglycosilatingagent, resulted in decreased labelling. Unfixed frozen sections proved to be difficult to interpret due to too high a background. Frozen sections post-fixed with 10% formaldehyde followed by methanol, or post-fixed with acetone proved negative. Sections cut from paraffin blocks gave the best results, but the method of fixation used was crucial. Bouin's and to a lesser extent paraformaldehyde (and perhaps glutaraldehyde) were found to be the most suitable fLxatives, whereas formalin, Carnoy's and Methacon fixation resulted in poor and variable staining. To demonstrate the presence of antigen in oligodendrocytes and astrocytes double labelling of these cells was used. Sections were incubated with either rabbit anti-carbonic anhydrase (raised in New Zealand white rabbits against purified rat erythrocyte CA2) or rabbit anti-GFAP (Dako Ltd., Bucks, U.K.) followed by peroxidase conjugate and developed with DAB substrate. This was followed by incubation with either MAb 307 (anti-E2) or 308 (anti-envelope glycolipid) and subsequently an alkaline phosphatase conjugate, followed by development with Naphthol As-Mx phosphate and fast blue. To verify the specificity of the staining, irrelevant IgG and IgM as well as preimmune sera controls for infected brain, as compared to normal brain, were used.
RESULTS The antigen labelled with MAbs 307 (anti=E2) and 308 (anti-envelope glycolipid) appeared as clusters of granules of variable size, found in the cytoplasm of representative cells of all types scattered throughout the brain. This staining was specific, as the staining with 307 could be removed by pre-incubation of the antibody with A7(74) (Fig. 1). Absorption of 308 activity was not tried. However, two brain structures appeared to show most consistently the presence of antigen; a) the hippocampus, in which the pyramidal neurons of the cornu Ammonis frequently contained SFV antigen (Fig. 2); b) the cerebellar cortex where immunoperoxidase staining was most frequently seen in what appear to be the perikarya of the Bergmann glia (Fig. 3). The other cells of the cerebellum such as the Purkinje cells, were only occasionally stained and then most commonly in the early stages of the infection. During the early phase of the infection, when intracerebral inflammation could be found, SFV antigen was detected in perivascular inflammatorycells (Fig. 4). Some GFAP=positive astrocytes also showed particulate staining with MAbs 307 and 308 which could be found up to day 183 p.i. (Fig. 5). SFV antigen was occasionally seen in carbonic anhydrase-positive oligodendrocytes. No staining was observed with 308 in non-neural or non-CNS SFVinfected tissues, whereas the anti=E2 antibody 307, was positive (Illavia and Webb 1988). Labelling of viral antigens with mouse SFV ( + )re polyclonal antibody was also successful. The level of antigen remained high all along in the nude mice up to days 40 p.i. (Table 1). The same pattern of labelling was also observed in animals infected with L10 at days 3-4 p.i.
20
Fig. 1. (a) Demonstration of staining of E~ antigen with MAb 307 in the cerebellum. (h) No staining was observed if M A b 307 was absorbed with A7(74) prior to use. Monochromatic filter set in the range of 448-512 nm was used here. G = granule neurons; P = Purkinje cells; B = Bergmann glia.
21
Fig. 2. SFV antigen localised in hippocampal neurons. A774 mice 40 days post-inoculation. (a) Staining of CA neurons with MAb 308. (b) Lower power demonstration of the localisation of SFV E2 with MAb 307, predominantly in CA neurons (CA), rather than in the dentate (D).
22
Fig. 3. SFV localisation in Bergmann giia in the cerebellum. Consistent, and large amounts of viral antigen were only detected in what appear to be the perflcaryal regions of the Bergmann glia (B), The Purkinje cells (P) and granule neurons (G) only react with anti-SFV MAbs occasionally and early after infection. (a) MAb 308 staining of cerebellum 7 days p.i. with L10 strain of SFV. This is to show, similar envelope glycolipid staining of both L10 and A774 strain. (b) 40 days p.i. with A774 using MAb 308.
23
Fig. 4. Demonstration of SFV in perivascular cells during the inflammatory stage of the disease. Perivascular cuffing of a vessel in cerebellar white matter, 6 days p.i.. Localisation of SFV with MAb 308 and indirect immunoperoxidase. TABLE 1 SUMMARY OF PATHOLOGY, CIRCULATING ANTIBODY LEVELS AND IMMUNOCYTOCHEMISTRY RESULTS AFTER SFV INFECTION IN nu/nu AND nu/+ MICE Brains from immunodeficient athymic nude mice (nu/nu) and their immunocompetent littermates (nu/+ ) were taken at days post-infection (p.i.) indicated. Circulating antibody levels had previously been investigated (Suckling et al. 1982; Parsons and Webb 1984). These brains were examined for pathology and sections were taken for immunocytochemistry and examined against MAbs 308, 307, 302 and polyclonal mouse anti-SFV sera. Pre-immune sera and irrelevant MAbs were used as controls. Results obtained with Swiss A2G mice were comparable with those of nu/+ mice. Day p.i.
6
14 28 40 68 183
Mouse
Serum anti-SFV Igs (rag/100 ml serum)
nu/ + nu/nu
1 00
nu/+ nu/nu nu/+ nu/nu nu/+ nu/nu nu/+ nu/+
304.5 54.5 366 54 360 45 260 IgG + (ELISA) (actual levels not available)
Pathology
~" J
23
Immunocytochemistry (MAbs)
302,307,308 ( + ve)
+++ ++ + + -
307,308(+ve)
24
Fig. 5. Astrocyte (A), the only cell with cytoplasm G F A P ( + )ve staining, was also ( + )ve for SFV antigen (appearing as blue dots) with MAb 308.
DISCUSSION
In this study we have shown that SFV antigens persist in mouse brain for a very long time after infection. Their presence within the astrocytes particularly could be important, with regard to the A7(74) immune-mediated demyelination (Fazakerley and Webb 1987) and the report that astrocytes can act as antigen presenting cells within the CNS in context of an appropriate MHC expression (Massa et al. 1986). MAb 308 raised to SFV has previously been shown by immuno-thin layer chromatography to be directed to a CNS glycolipid which also occurs in the viral envelope (Khalili-Shirazi et al. 1986). However, with immunocytochemistry,both at the light and EM level, this antibody appears to label the glycolipid in the SFV envelope, but not in the normal host CNS cell membrane. This may be due to conformational or concentration differences in the glycolipid in the viral envelope, Also, there may be loss of the
25 lipid from the cell m e m b r a n e during processing. Viral antigens were labelled equally well in our study, irrespective o f using the virulent (L10) or avirulent (A774) strain o f SFV. In conclusion, with the use of 2 monoclonal antibodies, one reactive with the viral spike protein E 2 and the other with a glycolipid c o m p o n e n t o f the envelope, it has been found possible to stain intraceUular particles in the brains of mice infected with either A7(74) or the L10 strain o f SFV. These same antibodies failed to stain brain tissue from normal non-infected animals. Previously, it had not been demonstrated by immunocytochemistry that SFV antigens persist for as long as 3 months. After absorption of M A b 307 with virus-containing tissue, no staining o f virus-infected brain could be achieved. It is proposed that the staining seen represents the presence o f virus in certain cells and the continued expression of at least part o f its genome within those cells. The observation that the humoral immune response persists long after the apparent clearance of SFV (Suckling et al. 1982; Parsons and W e b b 1984) would also suggest that there is a persistent expression and release o f part or the whole o f the SFV genome.
ACKNOWLEDGEMENTS This work was supported by St. T h o m a s ' Charitable Funds, the Wellcome Trust, the Philip Fleming Charitable Trust and the Multiple Sclerosis Society o f Great Britain and Northern Ireland.
REFERENCES Allner, K., C.J. Bradish, R. Fitzgeorge and N. Nathanson (1974) Modifications by sodium-thiomalate of the expression of virulence in mice by defined strains of Semliki Forest virus, J. Gen. Virol., 24: 221-228. Atkinson, T., A.D.T. Barrett, A. MacKenzie and N.J. Dimmock (1986) Persistence of virulent Semliki Forest virus in mouse brain followingco-inoculation with defective interfering particles,,/. Gen. ViroL, 67:1189-1194. Fazakerley, J.K. and H.E. Webb (1987a) Semliki Forest virus induced, immune mediated demyelination. The effect of irradiation, Br. J. Exp. Pathol., 68: 101-113. Fazakerley, J.K. and H.E. Webb (1987b) Semliki Forest virus induced, immune mediated demyelination - adoptive transfer studies, and viral persistence in nude mice, J. Gen. Virol., 68: 377-385. Fleming, P. (1977) Age-dependent and strain related differences of virulence of Semliki Forest virus in mice, J. Gen. Virol., 37: 93-105. Gates, M.C,, B.J. Sheahan and M.A. O'Sullivan (1985) The pathogenicity of the A7, M9 and L10 strains of SFV for weanling mice and primary mouse cell cultures, J. Gen. Virol., 66 : 2365-2373. Heyderman, E. (1979) Immunoperoxidase technique in histopathology: applications, methods and controls, J. Clin. Pathol., 32: 971-978. Illavia, S.J. and H. E. Webb (1988) The effect of single and multiple inoculations on the peripheral nervous system of mice with the demyelinating strain of Semliki Forest virus, Acta Neurol. Scand., in press. Jagelman, A. J., A.J. Suckling, H. E. Webb and E. T. W. Bowen (1978) The pathogenesis of avirulent Semliki Forest virus infections in athymic nude mice, 3". Gen. Viral., 41: 599-607. Khalili-Shirazi, A., N. Gregson and H. E. Webb (1986) Immunologicalrelationship between a demyelinating RNA enveloped budding virus (Semliki Forest) and brain glycolipids,J. Neurol. Sci., 76: 91-103. Massa, P. T., R. Dorries and V. ter Meulen (1986) Viral particles induce Ia antigen expression on astrocytes, Nature, 320: 543-546.
26 Parsons, L.M. and H. E. Webb (1984) Specific immunoglobulin G in serum and cerebrospinal fluid of mice infected with the demyelinating strain of Semliki Forest virus, Microbios Lett., 25: 135-140. Pathak, S. and H.E. Webb (1976) An electron microscopic study of the development of virulent and avirulent strains of Semliki Forest virus in mouse brain, J. Neurol. Sci., 28: 289-300. Suckling, A.J., S.J. Jagelman and H.E. Webb (1982) Immunoglobulin synthesis in nude (nu/nu) nu + and reconstitued nu/nu mice infected with a demyelinating strain of Semliki Forest virus, Clin. Exp. Immunol., 47: 283-288.