Superantigen-Mediated Encephalitis

Encephalitis may result from the interaction between a pathogen and the host brain, such as in bacterial, viral or fungal infection of the CNS. There are, however, different states of aseptic encephalitis, which may be induced e.g. by way of (i) self-directed immune attacks as in experimental autoimmune encephalomyelitis or (ii) by certain substances like the copper chelating agent cuprizone (Torkildsen et al., 2008). In fact, in encephalitis, infectious and non-infectious processes do not mutually exclude each other. Many pathogens such as bacteria or viruses encode for immune stimulating peptides, better known as superantigens for their enormous potency to stimulate immune cells (Kappler et al., 1989; Fleischer, 1991). T-cell superantigens have been developed twice during the evolution, namely independently by bacteria and by viruses. Not all bacteria or viruses do, however, possess a superantigen. Superantigens act in a T-cell receptor V(beta) dependent manner (Figure 1). Thereby, up to 10% or even 20% of the T-cell repertoire may become activated, sometimes resulting in a fulminant inflammatory response. The latter depends also on the specific repertoire of the host’s antigen-detection apparatus, e.g., the human-leucocyte-antigen (HLA) molecules.


Introduction
Encephalitis may result from the interaction between a pathogen and the host brain, such as in bacterial, viral or fungal infection of the CNS. There are, however, different states of aseptic encephalitis, which may be induced e.g. by way of (i) self-directed immune attacks as in experimental autoimmune encephalomyelitis or (ii) by certain substances like the copper chelating agent cuprizone (Torkildsen et al., 2008). In fact, in encephalitis, infectious and non-infectious processes do not mutually exclude each other. Many pathogens such as bacteria or viruses encode for immune stimulating peptides, better known as superantigens for their enormous potency to stimulate immune cells (Kappler et al., 1989;Fleischer, 1991). T-cell superantigens have been developed twice during the evolution, namely independently by bacteria and by viruses. Not all bacteria or viruses do, however, possess a superantigen. -Superantigens act in a T-cell receptor V(beta) dependent manner ( Figure 1). Thereby, up to 10% or even 20% of the T-cell repertoire may become activated, sometimes resulting in a fulminant inflammatory response. The latter depends also on the specific repertoire of the host's antigen-detection apparatus, e.g., the human-leucocyte-antigen (HLA) molecules. Taken together, superantigens are expected to take part in the inflammatory events induced by their pathogenic source. In florid bacterial or viral encephalitis, the detrimental effects that may be attributed to the pathogen or to its superantigen may be hardly discerned, while both may contribute to the final outcome. However, in case of slowly progressing encephalitides of e.g. viral origin, the superantigenic stimulus might be responsible for the initial and leading symptoms while the consequences of viral degeneration could be compensated for for a considerable time period and may become symptomatic later during the disease course. This is what has been suggested to happen in multiple sclerosis (MS) (Kornhuber et al., 2002, Kornhuber 2006, Emmer et al., 2008. In this respect, it may be of significance that the initial events of MS plaque generation seem to develop in the absence of tissue inflammation (Filippi et al., 1998;Barnett and Prineas, 2004). Although the etiology for slow progression in MS remains to be established, it has been speculated on a possible role e.g. for human endogenous retroviruses such as MSRV (Garson et al., 1998;Perron and Lang, 2010;Antony et al., 2010). Our knowledge about the cerebral actions of T-cell superantigens, e.g. Staphylococcal enterotoxin A (SEA), relies on but a few experiments and, thus, is far from being comprehensive. Nevertheless, the results outlined below may be useful for future studies to further characterize the role of superantigens per se or in the context of bacterial or viral encephalitis, respectively.

Effects of intracerebral T-cell superantigen
Intracerebrally expressed superantigen induces a perivascular and periventricular inflammation (Kornhuber et al., 2002;Emmer et al., 2010). Fourty-microliter aliquots of superantigen or saline were slowly injected intracerebrally through a small burr hole in isoflurane-anesthetized male 300-g Lewis rats, 2.5 mm lateral from the midline at the bregma at a depth of 3.5 mm. Horizontal hematoxylin/eosin stained sections of the rat brains were investigated after fixing the brains with 4% buffered paraformaldehyde. Sections were obtained at the corpus callosum and at the level of the lateral ventricles. Cuffings of perivascular round cells were identified scattered around the injection canal. In the first 3 days, perivascular round cells could be observed in both hemispheres with a preponderance in the corpus callosum and the periventricular white matter. Thereafter perivascular round cells were confined to the injected hemisphere up to 12 days after SEA injection. Maximum response in the injected hemisphere was identified up to 8 days after injection (Fig. 1). We wondered why the response to superantigen was relatively variable and usually low. It is well known that relapses in MS are often precipitated by some nonspecific immune stress such as infection. Furthermore, it is known that only activated immune cells are capable to invade the CNS (Wekerle et al., 1986). For these reasons we tried to imitate the stress by loading the blood with activated immune cells.

Activated splenocytes amplify superantigen encephalitis
Activated syngeneic splenocytes were injected in volumes of 0.5 ml through the penis vein of 300 g male Lewis rats on the third day after intracerebral injection of the superantigen SEA (see above). Activation of splenocytes was achieved in the following way under sterile conditions: A syngeneic spleen was cut and the content passed through a sieve into isotonic NaCl-solution. Cells were washed three times and the pellet finally resuspended in RPMImedium with 5 % heat-inactivated fetal calf serum and with a final concentration of 2 µg/ml concanavalin A (ConA). Cells were harvested and washed after 3 days in culture when they were maximally stimulated. They were kept in NaCl on crushed ice for injection purposes. Usually 10 7 cells were injected i.v. under brief general isoflurane anesthesia. The time course of the tissue reaction to 1 mg/ml SEA was investigated (Kornhuber et al., 2002). In general, perivascular round cell infiltrates were more numerous and more reproducible than without adding activated splenocytes. In the first up to 3 days after splenocyte injection, reactive vessels could be identified in both cerebral hemispheres with a preponderance on the injected side. Thereafter, inflamed blood vessels were confined to the injected hemisphere. The response was short-lived and could last for further 3 up to 12 days. Thereafter, no reactive vessels could be identified. On day 5 after i.v.-injection, on average 18.5±11.4 vessels with round cell cuffs were observed. When compared with the corresponding numbers obtained without i.v.-injection of activated splenocytes, the difference was statistically significant (p < 0.05; two-sided U-test). When the amount of activated splenocytes was kept constant at 10 7 per animal, the number of reactive blood vessels increased linearly with the concentration of SEA. When the injected SEA was kept constant at a concentration of 1 mg/ml, the number of reactive blood vessels increased linearly with the number of intravenously injected activated splenocytes.

Immunohistochemical characterization of round cell cuffing
Immunohistochemical investigations of the SEA-encephalitis were performed using shockfrozen brains fixed at -80 °C. Six µm kryocut sections were made at -14 °C. Neighboring tissue sections that showed both, the cerebral ventricles and the stitch canal were taken for further evaluation. The avidin-streptavidin-biotin (ABC)-method was used throughout for immunohistochemical staining purposes. All used antibodies were ordered by BD Biosciences Pharmingen. After preincubation with goat serum for 20 min, incubation with the primary antibody (1:50, 1 h) was followed by incubation with the secondary antibody (1:50, 30 min). After 30 min in the pre-diluted streptavidin-horseradish-peroxidase (HRP) all tissue sections were finally incubated with diaminobenzidine (DAB) solution until the desired colour intensity was obtained. Sections were dehydrated three times on increasing grades of alcohol and covered with Roti-Histokit. Spleen tissue slices served as the positive controls. All used primary antibodies were highly specific for their target antigen. Negative controls included substitution of primary antibodies by antibodies of the same isotype with specificities against non-host antigens. 12 h after i.v. injection of ConA-activated spleen cells (i.e. 3 days after intracerebral injection of SEA), relatively high numbers of immunoreactive CD3+, CD4+ and CD8+ T-cells were present in a perivascular distribution and also scattered in the parenchyma around the stitch canal of the injected hemisphere (Fig. 3). The perivascular cuffs consisted of several layers of round cells. The amount of immunoreactive cells within the perivascular infiltrates, i.e. CD3+, CD4+ and CD8+ T-cells decreased gradually thereafter (Figs. 3 and 4) and amounted merely to usually 1 complete layer of immunopositive cells 3 days after i.v. injection of splenocytes and some loosely grouped perivascular cells after 5 days. Thus, the www.intechopen.com inflammatory response was less large than in the previous experiments (see above). Among the different T-lymphocyte subsets, CD8+ T-cells were generally the most numerous ones (Figs. 3 and 4). The number of CD3+ T-cells within slices of the injected right hemisphere made up only about one quarter of the sum of the T-cells found to express CD8 or CD4 (Figs. 3 and 4). The relative numbers of reactive blood vessels remained more or less constant within the investigated time period (not shown). In the non-injected left hemisphere, the inflammatory activity was generally less prominent (Fig. 4). In fact, notable numbers of CD4+ and CD8+ T-cells within perivascular cuffs were identified only 3 days after i.v. injection of ConA-activated spleen cells (Fig. 4). In the brains of control animals, Tcells expressing CD4, CD8 or CD3 were not detected except for isolated immunopositive cells in the area of the stitch canal 0.5 days after the i.v. injection of ConA-activated splenocytes (not shown). Five days after the i.v. injection, no stained T-cells were found in the investigated brain slices of both control animals. The cerebral inflammatory reaction was short-lived, presumably due to the rapid disappearance of the injected superantigen, e.g. by non-specific binding to cell surfaces. Differences in the duration of the inflammation in the order of several days may be due to different preparations of the superantigen, which may impact the immunostimulatory potency of the SEA reagent. The cerebral inflammation induced by SEA was most prominent within the injected hemisphere and consisted initially mainly of CD8+ T-cells, which made up about 65% of the perivascular round cell population (Fig. 4). As no similar inflammatory response could be identified in the brains of the control animals that had received saline intracerebrally, the results do not appear to be due to the stitch trauma. Furthermore, only relatively small numbers of T-cells were found within the non-injected left hemisphere 0.5 days after i.v. injection of the ConA-activated splenocytes. As the number of the perivascular round cells detected in the left hemisphere peaked after 3 days following the i.v. injection of the ConAactivated splenocytes (Fig. 4), migration from the injected right hemisphere via the corpus callosum is the presumable reason for their occurrence contralateral to the injection site as has been suspected previously (Kornhuber et al., 2002). How does the superantigen expressed in the brain tissue lead to local recruitment and activation of T-cells? Presumably, the unprocessed superantigen was presented by MHC molecules on the surface of perivascular cells, microglial cells or dendritic cells, which are known to express MHC-class II constitutively within the CNS (Sedgwick et al., 1993;Stoll, 2002). By way of contrast, MHC-class I is not present on cell surfaces in the cerebral parenchyma unless its expression is specifically induced (Sedgwick et al., 1993;Redwine et al., 2001). When the injected ConAactivated cells appear in the circulation in high numbers after i.v. injection, namely 3 days after the intracerebral SEA-injection, free superantigen seems unlikely to be present in the cerebral extracellular fluid. Therefore, direct binding of SEA to the T-cell receptor (TCR) of ConA-activated splenocytes that come to traverse the blood-brain barrier does not seem to play a major role for T-cell activation in the present case (Fleischer, 1991;Herrmann et al., 1990;Yagi et al., 1990). However, only relatively small numbers of T-cells migrate through the cerebral blood vessels as part of a surveillance process, unless specific stimuli force them to stay on the abluminal side (Wekerle et al., 1986). The persistence of T-cells within the parenchyma after intracerebral injection of SEA may be taken as evidence that a specific stimulus forced them to stay within the CNS, therefore. After local expression of SEA, the majority of the T-cells detected within the intracerebral perivascular infiltrates was CD8+, while a minority was CD4+ (Figs. 1 and 2). Although superantigens presented via MHC II are well known to stimulate CD4+ T-cells, it has been demonstrated that SEA may activate CD8+ T-cells in a TCR-dependent manner (Müller-Alouf et al., 2001;Stinissen et al., 1995). Results obtained by gene expression analysis for the SEA encephalitis are in line with a CD8+ T-cell driven immune response (see below; Emmer et al., 2008). At a first glance it seemed to be curious that the numbers of T-cells expressing CD8 or CD4 detected within the right hemisphere in sum outnumbered the CD3+ ones at each investigated time point. The immunostaining for CD3 like that for CD4 and CD8 was of sufficient quality to allow a clear distinction between positive and negative cells (Fig. 3). Therefore, it does not seem likely that the mismatch between the results for CD3 and CD8 was artificial in nature. Actually, a diminished expression of CD3 but not of CD4 or CD8 has been reported previously for Tcells that had been challenged by superantigen (Damle et al., 1993;Niedergang et al., 1995;Makida et al., 1996;Von Essen et al., 2004). Therefore, the finding of a lower expression of CD3 in comparison to CD8 on T-lymphocytes like in the present investigation can be taken as evidence for the presence of a previous superantigenic stimulus. Taken together, it has been demonstrated that the round cells that take part in perivascular cuffing of the encephalitis induced by the superantigen SEA are primarily composed of Tcells, especially of CD8+ T-cells. This result may be of importance with respect to the pathogenesis of inflammatory diseases of the central nervous system. The fact that upon the superantigenic stimulus T-cells become CD3-negative in significant numbers, leaving the expression of e.g. CD8 unaltered, may be used to demonstrate the involvement of a superantigenic stimulus in different states of encephalitis.

Gene expression profile of superantigen encephalitis
Of 5 male 300 g Lewis rats, two animals received 50 µl of 1 mg⁄ ml SEA and two animals were injected with saline into the right brain hemisphere during deep anaesthesia. Injections were placed 2.5 mm lateral to the midline and 2 mm behind the bregma. One rat was sham operated. Three days after this procedure, 1.5*10 7 ConA-activated splenocytes (see above), were injected into the penis vein of each animal. Eight days after the initial surgical procedure, brains were taken from all Lewis rats. A coronar disk (2 mm) including the injection channel was prepared and divided into an injected half and a non-injected half. The samples were snap frozen in isopropanol and stored at -80 °C. Microarray analysis was performed as described previously (Staege et al., 2004). Data analyses were performed by using Statistical Analysis of Microarrays (SAM) (Tusher et al., 2001). Results from the 4 rat brain hemispheres of the two 'SEA' animals were compared with the 4 hemispheres of the 2 'saline' animals. Furthermore, the differential gene expression after saline injection versus sham operation was calculated. Due to the small sample size, a relatively conservative approach was followed with Δ = 0.75, a false discovery rate of 0.099 and a minimum change factor of three. To be acceptable, the signal intensities had to be above 30. Validation of www.intechopen.com microarray data was done by real-time PCR. To validate the results of the microarray analysis, the transcriptional regulation of nine genes [CCL5 (RANTES), TIMP-1, osteopontin, CD74, RT1-Da, complement component 3, tenascin C, CD8 and CCL2 (MCP-1)] that showed significant differential expression in rat brains with SEA encephalitis was measured also using real-time PCR. Total RNA was extracted from brain tissue using Trizol (Invitrogen, Karlsruhe, Germany) or the RNeasy kit (Qiagen, Hilden, Germany). Quantitative RT-PCR (qRT-PCR) was performed by using the QuantiTect SYBR Green RT-PCR Kit (Qiagen) using the following conditions: 94 °C, 45 s; 62 °C, 45 s; 72 °C, 60 s. Each reaction was subjected to melting temperature analysis to confirm presence of the expected products. Specific gene amplification was normalized to GAPDH. Target genes and GAPDH were amplified with 40 cycles using a ROTOR GENE RG-3000 (Corbett Research, Sydney, Australia) and ROTOR GENE 6 software. The threshold cycle (CT) value was defined as the fraction cycle number and set at 10 times the standard deviation above the mean baseline fluorescence calculated from cycles 3 to 15. The fold changes in the target genes normalized to actin 22 (as house keeping gene) and relative expression of controls (1 uninjected rat brain) was calculated by using standard ΔΔCT method.
Of the 8800 investigated genes, 106 were at least 3-fold increased with SEA over saline, while 29 genes were decreased at least 3-fold. The respective microarray data of differentially overexpressed genes are summarized in Table 1. Genes with increased expression were grouped in the following order: antigen presentation, lymphocytes, chemokines ⁄ chemokine receptors, microglial reaction ⁄ macrophages, phagocytosis ⁄ opsonization, extracellular matrix ⁄ cell adhesion, anti-inflammatory reaction and miscellaneous/ compound to inflammation. Some of the genes with decreased expression (not shown) presumably belong to cerebral cell elements such as neurons or astrocytes, e.g. genes encoding for neurotransmitter receptors or ion channels. In fact, the expression for the genes encoding for CCL5 (RANTES), TIMP-1, osteopontin, CD74, RT1-Da, complement component 3, tenascin C, CD8 and CCL2 (MCP-1) in relation to the house keeping gene for actin 22 as measured by real-time PCR showed a high level of conformity in comparison with the results obtained by using microarrays. Differential gene expression after saline injection versus sham operation revealed at least 3-fold overexpression of six genes and underexpression of 40 genes (not shown). The relatively mild differences observed in the gene expression between both conditions may reflect the consequences of the injection trauma and are considered of minor relevance for the SEA encephalitis. When data were first analysed, it became obvious that after intracerebral SEA injection versus saline injection, expression of several genes was markedly increased in the injected hemisphere and also displayed considerable overexpression in the non-injected contralateral hemisphere as well. This finding might correspond to the bilateral perivascular inflammatory reaction observed by using histology in the first days of SEA encephalitis (see above). Due to this finding, it was decided to analyse both hemispheres together. This approach certainly reduces absolute values of differential gene expression and at the same time it might reduce detection of false-positive data, e.g. resulting from the small number of samples. The results are in conformity with the light microscopy findings of a perivascular inflammation.
Among the genes with elevated expression, there was a considerable number of genes encoding for MHC class II molecules, which are constitutively expressed on microglial cells in the brain. In a state of encephalitis, they may be detected on astrocytes as well.

www.intechopen.com
Superantigen is presented in the context of MHC class II. However, 8 days after intracerebral injection of the superantigen, it might be doubted whether the elevated expression of genes for MHC class II molecules is still a direct consequence of the superantigenic stimulus. Antigen-presenting cells present in the inflammatory area may comprise microglial elements, monocytes and astrocytes [elevated expression of genes for: Serping1, CD53 antigen, CCAAT⁄ enhancer binding protein (C ⁄ EBP) delta, glial fibrillary acidic protein (GFAP) and calcium binding protein S100A4]. T lymphocytes seem to play a major role among the hematogenous cellular infiltrates of the SEA encephalitis. While the genes for CD3 and CD8 were found to be significantly elevated, this was not the case with the gene for CD4. This fits to immunohistochemical results showing that the perivascular round cell cuffs are dominated by CD8+ T lymphocytes on days 3.5, 6 and 8 after intracerebral SEA injection (see above). This finding was unexpected as usually CD4+ T cells are activated by T-cell superantigen presented in the context of MHC class II molecules (Fields et al., 1996). By way of contrast, CD8+ T cells are predominantly stimulated in the context of MHC class I molecules (Jelonek et al., 1998). As the latter ones are not constitutively presented in the brain, it seems unlikely that these molecules play a major role in the induction of the SEA encephalitis. Rather SEA may have been presented in the context of MHC class II. Previously, a similar stimulation of CD8+ T cells via superantigen bound to MHC class II as found in the present investigation has been reported (Fraser, 1989). Of interest, there exists a parallel to MS, where CD8+ T cells have been reported to predominate among perivascular inflammatory infiltrates (Liu et al., 2007;Jilek et al., 2007). Further proteins involved in antigen presentation or in signalling cascades were significantly overexpressed with SEA comprise sialoprotein CD43, cathepsin C, and CD 72. Similar to other states of cerebral inflammation such as in MS or EAE, there was a profound increase in the expression of the genes for the following proteins involved in chemotaxis after SEA injection: RANTES (CCL5), osteopontin, MCP-1 (CCL2) and CXCL10. Furthermore, the gene for the receptor of MCP-1 (CCR2) showed a significantly increased expression. In contrast to the elevated chemokine levels, cytokines, such as interleukin-2, tumour necrosis factor alpha or interferon gamma, did not reveal significantly increased differential expression levels. As these cytokines belong to the group of substances which are released early in the course of an inflammation, it is quite plausible that the genes for these cytokines are not expressed any more differentially 8 days after injection of SEA. Intracerebral injection of SEA was followed by an enhanced expression of genes encoding different complement factors such as C3, C4a, C1q, B, D (adipsin) and serping 1. These factors may be released from microglial cells (Raivich and Banati, 2004) or from macrophages. Complement factors were suggested to play a role in opsonization and phagocytosis. Complement factors 1q, 3 and 4a showed a high expression in microarray studies in EAE and MS (Tajouri et al., 2003;Lock et al., 2002). Actually, increased expression was detected for genes that are also related to phagocytosis. These were Fc-gamma receptor, Vav1, galectin 3 (Wilkinson et al., 2006;Rotshenker, 2003). These genes were previously shown to display increased expression in EAE (Lock et al., 2002;Reichert and Rotshenker, 1999;Carmody et al., 2006). A number of genes with increased expression levels after SEA injection were related to the extracellular matrix. These were lysyl oxidase, tenascin C, alpha-1-collagen type III, syndecan 1, alpha-1-collagen, alpha-1-procollagen type I, vimentin, matrix-gla-protein, periostin, oxidized LDL-receptor-1 and alpha-tubulin. The www.intechopen.com gene for tenascin C was measured with elevated expression in EAE and in MS (Lock et al., 2002, Carmody et al., 2006, while a similar increase for the gene of alpha-tubulin was present in MS (Carmody et al., 2006). Not all of the above summarized genes that showed enhanced expression after intracerebral superantigen injection did so in EAE or MS. Furthermore, the gene for integrin alpha M was detected with elevated expression in the present study. Integrin alpha M mediates cellular adhesion to the extracellular matrix (Friedl and Brocker, 2000). It is also upregulated in EAE and in MS (Lock et al., 2002, Carmody et al., 2006. Whether the enhanced gene expression of components of the extracellular matrix reflects alterations in the context of the encephalitis or reflects reparative activity remains to be established. Increased expression of the following genes may be regarded as part of an anti-inflammatory tissue reaction: alpha-2-macroglobulin, metallothionein, heat shock protein 27 (HSP27), haeme oxygenase-1, C⁄ EBP-related transcription factor beta, coeruloplasmin and pleckstrin. The gene products take part in the inactivation of proteolytic enzymes (alpha-macroglobulin), in the reduction in oxidative stress (methallothionein, haeme oxygenase 1, coeruloplasmin) or in the apoptosis induction (HSP27). Several of these genes have been observed with increased expression in EAE or in MS (Table 3), such as metallothionein (Tajouri et al., 2003;Lock et al., 2002;Penkowa and Hidalgo, 2003;, haeme oxygenase 1 (Levine and Chakrabarty, 2004;Tan et al., 2004) and HSP27 (Tajouri et al., 2003). Furthermore, there was a significant increase for the genes of the metalloproteinase 9 (MMP9) and its inhibitor, the tissue inhibitor of metalloproteinase 1 (TIMP-1). Both genes were reported to be upregulated in EAE and in plaque tissue from patients with MS (Steinman, 1999;Pagenstecher et al., 1998). While MMP9 is e.g. required for the migration of lymphocytes through the basilar membrane and thus for invading the CNS, the much more pronounced upregulation of TIMP-1 may be regarded as an anti-inflammatory response.

B-Cell superantigens and oligoclonal bands
When the role of superantigens is considered with respect to encephalitis, B-cell superantigens have to be taken into consideration in addition T-cell superantigens. A prominent representative for B-cell superantigens is gp120, which forms part of the envelope of the human immune-deficiency virus (HIV) (Neshat et al., 2000;Patke and Shaerer, 2000;Zouali, 2007). Like T-cell superantigens, B-cell superantigens stimulate their target cells in a clonal manner (Müller and Köhler, 1997;Goodyear and Silverman, 2005). As more than 1 B-cell clone is expected to be stimulated by a B-cell superantigen, it may be speculated whether this type of stimulus would result in the presence of oligoclonal IgG www.intechopen.com bands on isoelectric focussing. Indeed, oligoclonal IgG bands have been identified in various encephalitic diseases in the cerebrospinal fluid (CSF), including e.g. different forms of viral or bacterial encephalitis. Usually, all the antibodies forming oligoclonal bands in these diseases are directed against proteins of the encephalitogenic pathogen. However, there are states of encephalitis like in MS, where the presence of oligoclonal bands cannot be attributed to a certain pathogen. In fact, the antigen specificities present in MS oligoclonal bands comprise almost any antigen that has been tested. Therefore, these oligoclonal antibodies in MS have been termed as "nonsense antibodies" (Mattson et al., 1980). Among this nonsense-spectrum of antigen specificities, frequently an intrathecal antibody sythesis against measles, rubella, varizella zoster virus, herpes simplex virus, Epstein-Barr virus, and Chlamydia pneumoniae have been found (Reiber et al., 1998;Skorstad et al., 2009;Franciotta et al., 2010;Fainardi et al., 2009). It may be interesting to mention here, that antibodies specific for myelin proteins form only a small part of the oligoclonal IgG antibodies in MS (Owens et al., 2009). If nonsense antibodies like in MS would be due to a B-cell superantigenic stimulus, experimental proof should be available. Therefore, we tested in vitro, whether B-cell superantigens were capable to induce the formation of oligoclonal IgG bands on isoelectric focussing. In fact, after stimulation of peripheral blood mononuclear cells in vitro with the B-cell superantigen gp120, we detected IgG-bands by isoelectric focussing of the supernatant ( Figure 5; Emmer et al., unpublished). This IgG-production Representative results obtained by isoelectric focussing after stimulation of peripheral blood mononuclear cells from 3 healthy human donors in vitro with gp120 (8 µg/ml) and without gp120 (control). The numbers underneath the images denote the different subjects.
depended on the concentration of the B-cell superantigen. The detected oligoclonal bands were quite similar to those found by isoelectric focussing in the cerebrospinal fluid of MSpatients. Our results suggest that B-cell superantigens may play a role in the pathogenesis of the inflammatory response of multiple sclerosis. The expression of oligoclonal IgG in the CSF of MS-patients per se could have a detrimental influence, e.g. by opsonization of central nervous system components and subsequent phagocytosis by macrophages. In fact, the progress of the disease has been reported to be unfavourable if multiple oligoclonal bands are detected in the CSF of MS-patients (Joseph et al., 2009).

Conclusion
The present review focusses on the encephalitogenic effects of the intracerebrally expressed T-cell superantigen SEA. It has been demonstrated that SEA is capable to induce a perivascular inflammatory response, which was short lived after a single intracerebral injection. In the context of a pathogen residing within the CNS, a T cell superantigen is, however, expected to be expressed for prolonged periods of time and could, therefore, induce a longer lasting inflammatory response. The latter might add to the noxious response of the pathogen itself. Furthermore it was demonstrated that B-cell superantigens are able to stimulate B-cells to produce IgG which is detected as oligoclonal bands by isoelectric focussing. These oligoclonal bands resemble those found in the CSF of MS-patients. Beside bacterial infections, the presented findings could be of special relevance for viral encephalitis and possibly for multiple sclerosis.

Acknowledgement
A.E. gratefully acknowledges the support obtained within the Wilhelm-Roux-grant (FKZ 21/22) by the Martin-Luther-University of Halle-Wittenberg.