Storm Toxic Shock: Not a Simple Th1 Cytokine the Mechanism of Superantigen-Mediated

The Mechanism of Superantigen-Mediated Toxic Shock: Not a Simple Th1 Cytokine Storm

This information is current as Lee Faulkner, Anneli Cooper, Cristina Fantino, Daniel M. of September 30, 2021. Altmann and Shiranee Sriskandan J Immunol 2005; 175:6870-6877; ; doi: 10.4049/jimmunol.175.10.6870 http://www.jimmunol.org/content/175/10/6870 Downloaded from

References This article cites 50 articles, 19 of which you can access for free at: http://www.jimmunol.org/content/175/10/6870.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 30, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

The Mechanism of Superantigen-Mediated Toxic Shock: Not a Simple Th1 Cytokine Storm1

Lee Faulkner, Anneli Cooper, Cristina Fantino, Daniel M. Altmann, and Shiranee Sriskandan2

The profound clinical consequences of Gram-positive toxic shock are hypothesized to stem from excessive Th1 responses to superantigens. We used a new superantigen-sensitive transgenic model to explore the role of TCR␣␤ T cells in responses to staphylococcal enterotoxin B (SEB) in vitro and in two different in vivo models. The proliferative and cytokine responses of HLA-DR1 spleen cells were 100-fold more sensitive than controls and were entirely dependent on TCR␣␤ T cells. HLA-DR1 mice showed greater sensitivity in vivo to two doses of SEB with higher mortality and serum cytokines than controls. When D- galactosamine was used as a sensitizing agent with a single dose of SEB, HLA-DR1 mice died of toxic shock whereas controls did not. In this sensitized model of toxic shock there was a biphasic release of cytokines, including TNF-␣, at 2 h and before death at ␣␤ ␣ 7 h. In both models, mortality and cytokine release at both time points were dependent on TCR T cells. Anti-TNF- pretreat- Downloaded from ment was protective against shock whereas anti-IFN ␥ pretreatment and delayed anti-TNF-␣ treatment were not. Importantly, anti-TNF-␣ pretreatment inhibited the early TNF-␣ response but did not inhibit the later TNF-␣ burst, to which mortality has previously been attributed. Splenic T cells were shown definitively to be the major source of TNF-␣ during the acute cytokine response. Our results demonstrate unequivocally that TCR␣␤ T cells are critical for lethality in toxic shock but it is the early TNF-␣ response and not the later cytokine surge that mediates lethal shock. The Journal of Immunology, 2005, 175: 6870–6877. http://www.jimmunol.org/ evere sepsis leading to shock remains a leading cause of release of proinflammatory cytokines, including IL-1, TNF-␣, mortality and morbidity, with a rising incidence due to TNF␤, and IFN-␥ (8–11) with rapid expansion of T cells followed S increasing numbers of immunosuppressed individuals and by deletion or anergy (12). increasing age (1, 2). Infections due to Gram-positive pathogens, Animal models have contributed significantly to our understand- including Staphylococcus aureus and Streptococcus pyogenes, ac- ing of the mechanisms involved in toxic shock, however, mice are count for roughly half of severe sepsis cases. The immune re- much less sensitive to superantigen-mediated effects than humans sponse to Gram-positive infection may be exacerbated by the hu- (13, 14). The differential sensitivity resides in the lower binding man pathophysiological reaction to bacterial superantigens, affinity of superantigens to murine MHC class II than to human proteins with profound immunological potency that may lead to HLA class II (15). HLA class II transgenic mice thus provide an by guest on September 30, 2021 toxic shock (3–5). opportunity to study toxic shock in a readily manipulated animal Superantigen activation of T cells is dependent on HLA class II model with similar sensitivity to superantigens as humans (16– and TCR binding but is distinct from conventional Ag-specific 19). To enhance the effects of toxins in mice, some researchers responses which require Ag processing, presentation in the class II have used a “double dose” regimen in which high doses of supe- groove, and clonotypic T cell recognition (6). Superantigens bind rantigen are given twice, one essentially a priming dose and the to MHC class II molecules on APCs outside the peptide-binding other a challenge, or have coadministered the hepatotoxic drug, groove and to the TCR variable chains on the T cell. Different D-galactosamine. This drug increases sensitivity by up to 1000- superantigens have distinct affinities for different HLA molecules, fold through an effect on liver transcription (20, 21), although un- e.g., staphylococcal enterotoxin B (SEB)3 displays differential der the conditions used it has no lethal or immunostimulatory ef- binding for HLA-DRϾDQϾDP (7). Specific superantigens also fect alone (11, 22). favor interactions with particular V␤ families (3). Thus, in the case Superantigens have a marked ability to stimulate T cells in vitro of a large V␤ family, such as V␤8 receptors, which account for and during toxic shock trigger profound hypotension and multior- ϳ20% of the murine T cell repertoire, activation with an appro- gan failure. The proposed causal link between in vitro and in vivo priate superantigen may affect an enormous sector of the total rep- observations is the cumulative Th1 cytokine storm elicited by ex- ertoire, far greater than the proportion involved in a given peptide- cessive T cell activation by superantigens (23, 24). In support of a specific response. Superantigenic stimulation results in a systemic simple T cell activation model, early studies from Marrack et al. (25), using SEB-induced weight loss as a correlate of shock, Department of Infectious Diseases, Imperial College London, London, United King- showed that nude or cyclosporine-treated mice were relatively pro- dom tected from this effect (11, 25). Bette et al. (26) showed that early Received for publication April 28, 2005. Accepted for publication August 18, 2005. cytokine mRNA production in the spleen was restricted to the T The costs of publication of this article were defrayed in part by the payment of page cell-dependent area of the periarteriolar lymphatic sheets (PALS) charges. This article must therefore be hereby marked advertisement in accordance following SEB injection and was not affected by macrophage de- with 18 U.S.C. Section 1734 solely to indicate this fact. pletion. In addition, Tsytsykova and Goldfeld (27) showed that 1 This work was supported by Defense Advanced Research Projects Agency/Army mice deficient in NFATp, a critical T cell transcription factor for Reserve Office. TNF-␣, are resistant to SEB-induced toxic shock. However, there 2 Address correspondence and reprint requests to Dr. Shiranee Sriskandan, Depart- ment of Infectious Diseases, Imperial College, Hammersmith Hospital, Du Cane is evidence that superantigens can in certain circumstances elicit Road, London, W12 0NN, U.K. E-mail address: [email protected] inflammation in a non-T cell-dependent manner. Some recombi- 3 Abbreviation used in this paper: SEB, staphylococcal enterotoxin B. nant mutant superantigens, which lack residues necessary for T

cell activation, still retain lethality in rabbit models of toxic shock where 150 ␮l/well of medium was removed after 48 or 72 h of culture and (28). Conversely, some mutant superantigen constructs that have stored at Ϫ20°C until analysis. lost lethality have unaltered ability to superantigenically stimulate Tissue and intracellular TNF-␣ T cells (29). The evidence supporting the importance of T cells in toxic HLA-DR1 mice were treated with 20 mg of D-galactosamine and 20 ␮gof ␣ shock led to the expectation that T cell-derived cytokines would SEB. For measurement of tissue TNF- , tissue samples were taken after 0, ␥ 2, and 7 h. One-hundred milligram spleen and liver samples were homog- play a key role in the pathogenesis of toxic shock. IFN- ,asthe enized in 500 ␮l of RPMI 1640 medium containing Protease Inhibitor hallmark cytokine of Th1 activation, is implicated by the ﬁnding Cocktail III (Calbiochem) and 1 ml of the same medium was used for that IFN-␥R knockout mice are resistant to endotoxin or SEB- peritoneal lavage. The samples were freeze-thawed three times and spun at induced shock (30). However, in other studies, treatment with anti- 2000 ϫ g for 20 min to remove tissue debris. ␣ IFN-␥ Abs appears only partially protective (10, 31, 32). Thus, the For measurement of intracellular TNF- , tissue samples were taken after 0, 1.5, and 6.5 h. Peritoneal cells were removed by lavage using 5 ml aim of the current study was, using a highly superantigen-sensitive of RPMI 1640, 10 U/ml heparin and the spleen was homogenized to a transgenic model, to dissect the events leading to toxic shock, ask- single cell suspension. Cells were incubated with 50% FCS in PBS for 15 ing which elements of immune activation are most evidently min and washed in 2% FCS in PBS, incubated with anti-CD3-FITC, anti- linked to the lethal effects of the bacterial toxin. We have charac- B220-FITC, anti-Mac-1-FITC, and isotype-matched control Abs for 20 min and washed, ﬁxed for 15 min in Cell Fix and washed, incubated in terized the response of HLA-DR1 transgenic mice to SEB both in 0.2% saponin in PBS for 10 min and washed, incubated with anti-TNF- vitro and in vivo, and compared this response to control mice. Our ␣-PE and isotype-matched control Ab for 30 min and washed in 0.2% objectives were to investigate which cytokines and which cell pop- saponin in PBS. All incubations were carried at 4°C and all reagents were

ulations are critical for lethality in SEB-induced toxic shock. To supplied by BD Biosciences. Samples were analyzed on a FACSCalibur Downloaded from this end, we have compared responses of HLA-DR1 mice with and and 10,000 events were acquired for each sample. Data analysis was per- formed using CellQuest software. without an intact TCR␣␤ T cell population. Cytokine detection Materials and Methods IL-6, IL-12, TNF-␣, and IFN-␥ were detected by ELISA using matched Ab Mice pairs and according to the manufacturer’s instructions (R&D Systems). FVB/N wild-type controls, HLA-DR1, and HLA-DR1 TCR␣Ϫ/Ϫ mice http://www.jimmunol.org/ were bred and maintained in accordance with Home Office guidelines. Statistics HLA-DR1 transgenic mice were generated from FVB/N mice by insertion Data were analyzed using Student’s t test. of HLA-DRAI*0101 and HLA-DRB1*0101 genes as described in Ref. 33. TCR transgenic mice carry a targeted disruption in the TCR ␣ gene as described in Ref. 34. They have no TCR␣␤-positive T cells but have nor- Results mal TCR␥␦ T cells. These mice were then bred with HLA-DR1 mice and HLA-DR1 transgenic mice show enhanced sensitivity to SEB in ϩ Ϫ a breeding colony maintained as HLA-DR1 TCR␣ / and HLA-DR1 vitro TCR␣Ϫ/Ϫ mice. The presence of TCR␣␤-positive T cells in HLA-DR1 TCR␣ϩ/Ϫ and their absence in HLA-DR1 TCR␣Ϫ/Ϫ mice was confirmed To establish a model of toxic shock in HLA-DR1 mice we first Ϫ/Ϫ by FACS. HLA-DR1 TCR␣ mice had similar or proportionally higher investigated the sensitivity of spleen cells to SEB in vitro. The by guest on September 30, 2021 ␥␦ levels of B cells, monocytes, and TCR cells in spleen and lymph nodes proliferative response of HLA-DR1 spleen cells was at least 100- compared with HLA-DR1 TCR␣ϩ/Ϫ mice (data not shown). The genotype of the transgenic mice was routinely assessed by PCR. fold greater than control mice when exposed to 1–100 ng/ml SEB (Fig. 1A). Over the same concentration range of SEB, HLA-DR1 Treatment groups spleen cells also released significantly higher amounts of TNF-␣, ␥ Groups of mice received injections i.p. with 20 mg of D-galactosamine IL-6, and IFN- compared with control spleen cells (Fig. 1, B–D). (Sigma-Aldrich) in 0.2 ml of saline and 0.2–200 ␮g of SEB (Toxin Tech- Thus, HLA-DR1 mice show enhanced sensitivity to SEB in vitro nology) in 0.2 ml of saline. Mice rarely showed any adverse symptoms for as we and others have already shown for other HLA class II trans- the first 6 h following treatment. Thereafter mice suffering from toxic genic mice (16, 18, 19, 37). shock showed prostration, piloerection, weight loss, dehydration, and liver necrosis (21, 22), usually dying 7–8 h after treatment. Alternatively, ␣␤ groups of mice received injections i.p. with 2 doses of 50–100 ␮gofSEB The presence of TCR T cells is an absolute requirement for in 0.2 ml of saline, 48 h apart. Survival was monitored 8–24 h after in vitro responses to SEB last dose. Some groups of mice received injections i.p. with anti-cytokine Abs in IL-6 is an acute phase cytokine released by a wide range of cell addition to challenge with 20 mg of D-galactosamine and 20 ␮g of SEB. types and not generally associated with the specific activation of Mice received injections with 500 ␮g/mouse of hamster anti-TNF-␣ Ab, Th1 cells. To investigate whether this response can be elicited in TN3-19.12 (35) or the isotype-matched control, L2-3D9 (CellTech) 1 h the absence of TCR␣␤ cells, we backcrossed HLA-DR1 transgenic ␮ ␥ before or 4 h after challenge and 400 g/mouse of purified rat anti-IFN- mice onto a homozygous deletion for the TCR␣ locus. Use of Ab R4-6A2 (36) (provided by Dr. A. Annenkov, Kennedy Institute of ␣Ϫ/Ϫ ␥␦ Rheumatology, Imperial College, London, U.K.) or the isotype control, TCR mice was also of particular interest given that TCR T 337.217.7 1 h before challenge. Ab efficiency in removing cytokine from cells are reported to respond to bacterial superantigen (38, 39). the serum was established by ELISA of serum samples (data not shown). Spleen cells from HLA-DR1 TCR␣ϩ/Ϫ mice, which have a nor- Blood samples were taken by tail bleed or cardiac puncture and serum mal T cell repertoire, proliferated in response to SEB following a was stored at Ϫ20°C until analysis. Only female mice were used in ex- periments. Mice were age and weight matched. similar dose response to that seen for HLA- DR1 spleen cells (Figs. 1A and 2A). In contrast, spleen cells from HLA-DR1 Proliferation assays TCR␣Ϫ/Ϫ mice, which lack TCR␣␤ T cells, did not proliferate at ␮ A single cell suspension of spleen cells in RPMI 1640 medium containing any concentration of SEB, even at 10 g/ml SEB which was ca- 10% FCS, 2 mM glutamine, 50 U/ml penicillin, 50 ␮g/ml streptomycin pable of inducing a marked response in control mice. Spleen cells was plated at 2 ϫ 105 cells/well in triplicate in 96 round-bottom plates. The from HLA-DR1 TCR␣ϩ/Ϫ mice released significant amounts of ␮ cells were cultured at 5% CO2, 37°C for 72 h in the presence of 0–1 g/ml IL-6, TNF-␣, and IFN-␥ in response to SEB whereas spleen cells ␮ ␮ SEB or 2 g/ml Con A (Sigma-Aldrich) in a final volume of 200 l. The ␣Ϫ/Ϫ cells were pulsed with 1␮Ci/well [3H]thymidine (Amersham Biosciences) from HLA-DR1 TCR mice did not (Fig. 2, B–D). Therefore, for the last 18 h of culture. The incorporated radioactivity was measured by the proliferative and cytokine responses of spleen cells to SEB are liquid scintillation. Duplicate cultures were set up for cytokine analysis entirely dependent on the presence of TCR␣␤ T cells. TCR␥␦ T 6872 MECHANISMS OF SUPERANTIGEN-MEDIATED TOXIC SHOCK Downloaded from

FIGURE 2. Response to SEB in vitro is T cell dependent. Spleen cells

from HLA-DR1 TCR mice were exposed to 0–10 ␮g/ml SEB for 3 days. http://www.jimmunol.org/ A, Proliferation measured by [3H]thymidine uptake. Proliferation in the presence of 2 ␮g/ml Con A was 124,221 Ϯ 49,264 for HLA-DR1 FIGURE 1. Spleen cells from HLA-DR1 mice are more responsive to TCR␣ϩ/Ϫ and 6,293 Ϯ 6,421 for HLA-DR1 TCR␣Ϫ/Ϫ. Cytokine release SEB in vitro. Spleen cells from control FVB/N and HLA-DR1 mice were after 48 h: B, TNF-␣; C, IL-6; D, IFN-␥. Data shown are means of values ␮ exposed to 0–10 g/ml SEB for 3 days. A, Proliferation was measured by from three individual mice Ϯ SD. All values except medium alone were 3 ␮ [ H]thymidine uptake after 72 h. Proliferation in the presence of 2 g/ml p Ͻ 0.001 by t test. Con A was 245,547 Ϯ 9,764 for controls and 238,015 Ϯ 20,716 for HLA- DR1. Cytokine release after 72 h: B, TNF-␣; C, IL-6; D, IFN-␥. Data .p Ͻ 0.005 by t test ,ء .shown are means from three individual mice Ϯ SD by guest on September 30, 2021 Serum cytokines produced in response to SEB are T cell dependent cells present in HLA-DR1 TCR␣Ϫ/Ϫ spleen cells were unable to The role of HLA class II and TCR␣␤ T cells in cytokine release in substitute for the lack of TCR␣␤ T cells. vivo was investigated by measuring serum cytokines after two doses of SEB. HLA-DR1 mice had significantly higher levels of serum IL-6, TNF-␣, and IFN-␥ 7 h after the second SEB dose Sensitivity of HLA-DR1 mice to SEB in vivo is also T cell compared with control mice (Fig. 3A). Ϫ Ϫ dependent HLA-DR1 TCR␣ / mice had significantly lower levels of serum IL-6, TNF-␣, IL-12, and IFN-␥ 7 h after the second SEB dose Next, we investigated whether the enhanced sensitivity to SEB we ϩ Ϫ compared with HLA-DR1 TCR␣ / mice (Fig. 3B) demonstrating had seen in vitro was mirrored by a similar sensitivity in vivo. that production of these cytokines was largely, though not com- Although a single dose of superantigen is rarely fatal to unsensi- pletely, T cell driven. To our knowledge this is the first demon- tized mice, treatment with two doses can cause mortality (11, 32). stration that superantigen induced cytokine responses in vivo de- Control and HLA-DR1 mice were given two doses of SEB 48 h pend upon an intact TCR␣␤ T cell repertoire. apart and their survival monitored for a further 24 h. All control mice survived (5 of 5) this treatment whereas 40% of DR1 mice Lethal effects of D-galactosamine and SEB in HLA-DR1 mice died (2 of 5 mortality). Control mice gained weight over the course are T cell-dependent Ϯ Ͻ of the experiment (gain 1.0 0.5 g from 0 to 72 h, p 0.005 by The sensitivity of mice to SEB can be further enhanced by coad- t test), whereas HLA-DR1 mice lost weight (lost 1.2 Ϯ 0.6 g from ministration of D-galactosamine such that lethal toxic shock is in- Ͻ 0to72h,p 0.005 by t test). duced within 8 h. HLA-DR1 mice were highly susceptible to toxic The role of T cells in the response to SEB in vivo was investi- shock induced by D-galactosamine and SEB. In contrast, control ␣␤ gated by giving two doses of SEB to mice without TCR T cells. mice showed no adverse effects to this treatment and survived even ␣Ϫ/Ϫ All HLA-DR1 TCR mice survived (5 of 5) whereas only 40% the highest dose of SEB given. Neither D-galactosamine alone nor ϩ/Ϫ of HLA-DR1 TCR mice survived (2 of 5) two doses of SEB. 200 ␮g of SEB alone was lethal to HLA-DR1 mice (Table I). ϩ Ϫ HLA-DR1 TCR␣ / mice lost more weight over the course of the Next we investigated whether TCR␣␤ T cells were responsible Ϫ Ϫ experiment (2.7 Ϯ 1.52 g) than HLA-DR1 TCR␣ / mice (1.9 Ϯ for lethality in this sensitized model of toxic shock. HLA-DR1 0.60 g) but this difference was not statistically significant. Thus TCR␣ϩ/Ϫ mice showed a similar sensitivity to the lethal effects of HLA-DR1 mice show enhanced sensitivity to SEB in vivo com- 20 ␮g of SEB in the presence of D-galactosamine, as HLA-DR1 pared with control mice and mortality to SEB appeared to be de- mice (Table I). In contrast, all HLA-DR1 TCR␣Ϫ/Ϫ mice survived pendent on TCR␣␤ T cells. without ill effects, clearly demonstrating for the first time that The Journal of Immunology 6873

showed a much stronger serum cytokine response to SEB than control mice (Fig. 4, A–D). This difference was not due to differing sensitivity to D-galactosamine, because D-galactosamine alone produced little or no cytokines. HLA-DR1 mice showed an early burst of TNF-␣, IL-2, and IL-6 release 2 h after treatment followed by a drop in levels and a further rise in TNF-␣ and IL-6 just before death from toxic shock (Fig. 4, E and F). In contrast, serum IL-12 and IFN-␥ levels in HLA-DR1 mice showed a gradual increase over 7 h coinciding with the second phase of TNF-␣. IL-12 levels peaked at 6 h and preceded the rise in serum IFN-␥ (Fig. 4G). Thus, the response to SEB in this model can be broadly considered to encompass a rapid, early IL-2, IL-6, and TNF-␣ response, followed some hours later by build up of a more classically Th1, IL-12-driven IFN-␥ response.

Early but not late release of TNF-␣ is a critical mediator in lethality due to SEB Downloaded from To determine which of these cytokines might be key mediators of lethality in the D-galactosamine model, HLA-DR1 mice were pre- treated with Abs to remove systemic TNF-␣ or IFN-␥ before challenge with D-galactosamine and SEB. Pretreatment with anti-TNF-␣ Ab was protective (Table II) and

intriguingly led to a marked reduction in the early but not later http://www.jimmunol.org/ phase of cytokine response. Both IL-6 and TNF-␣ levels were FIGURE 3. Increased serum cytokines in HLA-DR1 mice after two reduced 1 h after D-galactosamine and SEB in HLA-DR1 mice ␣ doses of SEB are T cell dependent. Mice received injections i.p. with treated with anti-TNF- Ab compared with mice treated with con- 50–100 ␮g of SEB 48 h apart. Serum cytokines were analyzed 7 h after the trol Ab. However, 7 h after challenge both control and treatment second dose. A, Control FVB/N and HLA-DR1 mice. B, HLA-DR1 TCR groups had similar levels of IL-6 and TNF-␣ (Fig. 5, A and B). p Ͻ Furthermore, the later surge in IFN-␥ levels was unaffected by ,ء .mice. Data are means of values from ﬁve individual mice Ϯ SD ϩ Ͻ 0.001; , p 0.01 by t test. treatment with anti-TNF-␣ Ab (Fig. 5C). This suggested that the

early cytokine response is largely responsible for lethality in the by guest on September 30, 2021 TCR␣␤ T cells are critical to the lethal effects of toxic shock in this D-galactosamine model as opposed to the later, conventional Th1 model. response involving IFN-␥. Both groups of mice lost the same amount of weight over the 8 h ϩ Ϫ To test this hypothesis, administration of anti-TNF-␣ Ab was of the experiment (HLA-DR1 TCR␣ / lost 1.3 Ϯ 0.4 g, HLA- Ϫ Ϫ delayed until after the early cytokine response but before the later DR1 TCR␣ / lost 1.3 Ϯ 0.5 g; both p Ͻ 0.001). Previously, ␣ Marrack et al. (25) found that weight loss correlated with lethality rise in TNF- (administration 4 h after D-galactosamine and SEB). ␣ from a single superantigen dose and that nude mice or mice treated As expected, anti-TNF- Ab administered at 4 h reduced serum ␣ ␥ with cyclosporine were protected from weight loss. In this in- TNF- and IL-6 just before death (7 h) although IFN- and IL-12 ␣ stance, mice were sensitized with D-galactosamine which may ac- levels were unaffected (data not shown). Delaying the anti-TNF- count for the difference in results. However, the weight loss in Ab until after early cytokine response proved critical as HLA-DR1 HLA-DR1 TCR␣Ϫ/Ϫ mice was transient and mice returned to their mice were not protected from toxic shock and mortality was the starting weights overnight (data not shown). same in both control and treatment groups (Table II). Abs to IFN-␥ were not protective against toxic shock and had no HLA-DR1 mice show an early and late burst of serum cytokines effect on TNF-␣ production (Table II). Serum TNF-␣ levels were after D-galactosamine and SEB not signiﬁcantly different between anti-IFN-␥-treated mice (2940.5 Next we investigated the systemic release of cytokines in the D- ng/ml, SD 928.7) and control mice (3824.2 ng/ml, SD 902.4) galactosamine-sensitized model. HLA-DR1 transgenic mice at7h.

a Table I. Mortality of mice from toxic shock induced by D-galactosamine and SEB

Mortality

Treatment Control HLA-DR1 HLA-DR1 TCR␣Ϫ/Ϫ HLA-DR1 TCR␣ϩ/Ϫ

20 mg of Dgal nd 0/3 200 ␮g SEB nd 0/3 20 mg of Dgal ϩ 0.2 ␮g of SEB 0/5 0/5 20 mg of Dgal ϩ 2 ␮g of SEB 0/5 4/5 20 mg of Dgal ϩ 20 ␮g of SEB 0/3 5/5 20 mg of Dgal ϩ 200 ␮g of SEB 0/3 nd 20 mg of Dgal ϩ 20 ␮g of SEB 0/10 10/10

a Control FVB/N, HLA-DR1, and HLA-DR1 TCR mice were treated i.p. with D-galactosamine (Dgal), SEB, or both. Survival was monitored over 8 h. nd, Not done 6874 MECHANISMS OF SUPERANTIGEN-MEDIATED TOXIC SHOCK Downloaded from http://www.jimmunol.org/

FIGURE 4. Serum cytokines are higher in HLA-DR1 mice after treatment with D-galactosamine and SEB. Mice received injections i.p. with 20 mg of D-galactosamine with and without 20 ␮g of SEB. Serum cytokines at 7 h for control FVB/N and HLA-DR1 mice: A, TNF-␣; B, IL-6; C, IFN-␥; D, IL-12. Serum cytokine time course for HLA-DR1 mice: E, IL-2; F, TNF-␣ and IL-6; G, IL-12 and IFN-␥. Data are means of values from three individual mice .p Ͻ 0.01; ϩ, p Ͻ 0.05 by t test ,ء .for D-galactosamine treatment and ﬁve individual mice for D-galactosamine and SEB treatment Ϯ SD by guest on September 30, 2021

TCR␣␤ T cells facilitate both the early acute phase and later HLA-DR1 TCR␣Ϫ/Ϫ mice at 7 h (Fig. 6D). Thus the presence of Th1 cytokine response to SEB in D-galactosamine-sensitized TCR␣␤ T cells enhanced both the acute phase and late Th1 cyto- mice kine response to SEB in the presence of D-galactosamine. We have shown that SEB-induced lethality is dependent on the acute phase TNF-␣ response in D-galactosamine-sensitized mice, T cells from the spleen are a source of early TNF-␣ in response and that T cells are critical to this lethal effect. To assess the con- to SEB in D-galactosamine-sensitized mice ␣␤ tribution of TCR T cells to early (acute phase) and late (clas- To identify the source of TNF-␣ in SEB-induced toxic shock, sical Th1) cytokine production in toxic shock, HLA-DR1 HLA-DR1 mice received injections with D-galactosamine and SEB ␣ϩ/Ϫ ␣Ϫ/Ϫ TCR and HLA-DR1 TCR mice were treated with SEB and tissue homogenates of spleen, liver, and peritoneal cells were in the presence of D-galactosamine. Cytokines were measured after subjected to cytokine analysis. There was a significant increase in 1 h to assess the acute phase response and at7htoassess the later TNF-␣ in spleen tissue and to a lesser extent peritoneal cells at 2 cytokine rise found before death. Although there was some evi- and 7 h but no such increase in the liver (Table III). Next we ␣␤ ␥ dence of cytokine response in mice lacking TCR cells, IFN- , analyzed individual spleen and peritoneal cells to identify the spe- ␣ TNF- , and IL-6 were significantly reduced in HLA-DR1 cific cell type responsible for TNF-␣ production in mice injected ␣Ϫ/Ϫ ␣ϩ/Ϫ TCR mice compared with HLA-DR1 TCR mice (Fig. 6, with D-galactosamine and SEB. A–C) at both time points. IL-12 was only significantly reduced in Cells were labeled for surface CD3, B220, and Mac-1 expression, fixed, permeabilized, and then labeled for intracellular TNF-␣. Intracellular TNF-␣ expression was not detected in peri- Table II. Mortality of mice treated with anti-cytokine Abs, a D-galactosamine, and SEB toneal cells (data not shown) at either time point. However, a dis- crete but definite rise in TNF-␣-positive cells was detected in the Mortality spleen at 1.5 h. Intracellular staining demonstrated that CD3-positive cells accounted for the majority of the rise in TNF-␣-positive Cytokine Time Given Control Ab Anti-cytokine Ab cells (Table IV). IFN-␥ 1 h before 5/5 5/5 TNF-␣ 1 h before 3/5 0/5 TNF-␣ 4 h after 4/5 4/5 Discussion The aim of the current study was to investigate which elements of a HLA-DR1 mice were treated with anti-cytokine Abs 1 h before or 4 h after treatment with D-galactosamine and 20 ␮g of SEB. Survival was monitored over immune activation are critical for the lethal effects of staphylococ- 8–24 h. cal superantigens. In doing this, we have brought together several The Journal of Immunology 6875 Downloaded from http://www.jimmunol.org/ FIGURE 6. Both early acute phase and late cytokine responses after treatment with D-galactosamine and SEB are T cell dependent. HLA-DR1 TCR mice received injections i.p. with 20 mg of D-galactosamine and 20 ␮g of SEB. Serum cytokines were analyzed at 1 and 7 h: A, IL-6; B, TNF-␣; C, IFN-␥; D, IL-12. Data are means of values from five individual .p Ͻ 0.005; ϩ, p Ͻ 0.05 by t test ,ء .mice Ϯ SD by guest on September 30, 2021 single dose of SEB with the sensitizing agent D-galactosamine. We have established a model system using HLA-DR1 transgenic mice to enhance the sensitivity to superantigens based on the observa- tion that HLA-DR1 has a higher affinity for SEB than murine FIGURE 5. Early acute phase cytokine response is inhibited by anti- MHC class II. We have also investigated the specific role of TNF-␣ Ab. HLA-DR1 mice received injections i.p. with 500 ␮g of anti- TCR␣␤ cells in SEB responses and toxic shock by comparing ␣ TNF- Ab or an isotype-matched control 1 h before challenge with 20 mg HLA-DR1 mice with and without an intact TCR␣␤ cell popula- ␮ of D-galactosamine and 20 g of SEB. Serum cytokines were measured at tion. Previous studies conducted in nude, SCID and cyclosporine- 1and7h:A, TNF-␣; B, IL-6; C, IFN-␥. Data are means of values from five treated mice (11, 25) are more difficult to interpret due to the broad .p Ͻ 0.001 by t test ,ء .individual mice Ϯ SD immune defects found or induced in these mice. Finally, we have used flow cytometry to directly and positively identify the cells features necessary for a relevant model but which have not nec- which produce the lethal burst of TNF-␣. essarily been considered in previous studies: an optimum model We have shown that HLA-DR1 mice are more sensitive to SEB needs to encompass lethality, high sensitivity, and the ability to both in vitro and in vivo than control mice. This is consistent with precisely analyze the role of T cell responses. Many previous stud- previous reports on streptococcal superantigens and HLA-DQ8 ies have investigated the effect of superantigens in vivo after a mice (16, 17) and reports on staphylococcal superantigens and single dose or under conditions that are not harmful (32, 40, 41). HLA-DR3, HLA-DQ6 and HLA-DQ8 mice (18, 19, 42). In our We have concentrated specifically on protocols which induce le- study, proliferative and cytokine responses to SEB in vitro were thal toxic shock either by giving two doses of SEB or by giving a entirely dependent on the presence of TCR␣␤ cells. Although

Table III. Tissue source of TNF␣a

0h 2h 7h

TNF-␣ pg/ml Mean SD Mean SD Mean SD

Spleen 111.3 46.1 1490.2* 303.0 561.1* 121.5 Liver 1228.4 95.7 1307.6 139.6 1373.3 217.9 PC 10.0 13.1 127.5* 63.1 213.1* 107.2

a HLA-DR1 mice were treated with D-galactosamine and 20 ␮g of SEB. Spleen (100 mg), liver (100 mg), and total peritoneal cells (PC) were analyzed for tissue TNF␣ (n ϭ 4). Values marked with an asterisk (*), appearing in bold, are signiﬁcantly different from concentration at time0h(p Ͻ 0.005 by t test). 6876 MECHANISMS OF SUPERANTIGEN-MEDIATED TOXIC SHOCK

Table IV. Cellular source of TNF␣a

0 h 1.5 h

Mean SD Mean SD

% of all spleen cells expressing TNF␣ 0.82 0.29 3.05* 0.91 % TNF-␣ positive cells also CD3 positive 13.1 12.5 81.1* 4.2 % TNF-␣ positive cells also B220 positive 51.7 31.0 17.7 5.0 % TNF-␣ positive cells also Mac-1 positive 35.2 30.6 1.2 2.1

a HLA-DR1 mice were treated with D-galactosamine and 20 ␮g of SEB. Spleen cells were stained for surface expression of CD3, B220, and Mac-1 and intracellular expression of TNF␣ (n ϭ 3). Values marked with an asterisk (*), appearing in bold, are signiﬁcantly greater than time0h(p Ͻ 0.001 by t test).

TCR␥␦ cells have been shown to respond to superantigens in vivo duced Th1 cytokines. Administration of anti-cytokine Abs to re- (38, 39) it seems likely that the number of these cells in the sys- move IL-2, IL-12, and IFN-␥ has not proved effective in protecting temic circulation was too low to mount a detectable or biologically mice in various models of toxic shock (10, 30–32, 49). We also signiﬁcant response. found that Abs to IL-12 (data not shown) or IFN-␥ were not pro- Elevated levels of IL-10, IL-12, and IFN-␥ have previously been tective in our lethal model of toxic shock. associated with lethality in the double dose model of toxic shock In contrast to classical Th1 cytokines, we and others have shown Downloaded from (32, 40). Florquin et al. (32) also found that anti-IFN-␥ Abs were that Abs against TNF-␣ are protective when given just before SEB protective, strongly implicating T cells as critical for lethality. In and D-galactosamine (11, 31, 49). Importantly, the protective effect this study, we have conclusively shown that lethality was depen- of anti-TNF-␣ pretreatment was associated with inhibition of the dent on the presence of TCR␣␤ cells because HLA-DR1 early TNF-␣ and IL-6 cytokine response in HLA-DR1 mice but TCR␣Ϫ/Ϫ mice were protected from toxic shock when given two not the later burst of TNF-␣ which was associated with rising

doses of SEB. This also suggests that TCR␥␦ cells do not play a IL-12 and IFN-␥. When administration of anti-TNF-␣ Abs was http://www.jimmunol.org/ significant role in the response to SEB in vivo. Cytokine produc- delayed until after the early cytokine response, no protective effect tion in vivo was clearly dependent on TCR␣␤ cells because HLA- was found despite the fact that the late rises in TNF-␣ and IL-6 DR1 TCR␣Ϫ/Ϫ mice had significantly lower serum IL-6, TNF-␣, were markedly reduced. This strongly suggests that it is the early and IFN-␥ than HLA-DR1 TCR␣ϩ/Ϫ mice. cytokine burst which is responsible for lethality in toxic shock When D-galactosamine is used as a sensitizing agent then supe- rather than the later rise in Th1 cytokines. The key mechanism of rantigens can induce toxic shock within hours. This is character- lethality is probably TNF-␣-induced liver damage (22, 49, 50). ized by elevated serum TNF-␣, TNF␤, IFN-␥, IL-2, IL-10, and Next, we addressed the tissue and cellular origin of the early IL-12 (10, 11). T cell involvement in this process has been dem- serum TNF-␣ seen in our model. The predominant cell types first onstrated by investigating cytokine release from purified splenic T exposed to SEB in this model are monocytes and macrophages in by guest on September 30, 2021 cells (43) and PBMCs (9, 24, 44), by measuring expansion of the peritoneum. SEB is rapidly absorbed into the circulation where specific TCR V␤ T cell subsets (11, 45, 46), by using nude or it can induce cytokine production in both lymphoid and nonlym- SCID mice (11, 25, 46), by disrupting T cell function (11, 25, 45, phoid tissue such as the spleen and liver (49). In a nonlethal model 46), and by disrupting the HLA class II-superantigen-T cell inter- of toxic shock, Bette et al. (26) identified early TNF-␣ mRNA action (47, 48). From this work has emerged the hypothesis that production in the PALS area of the spleen, which is predomi- toxic shock develops as a result of superantigen-induced T cell nantly, although not exclusively a T cell area. We also found that activation resulting in uncontrolled Th1 cytokine release culminat- the spleen is a major source of early TNF-␣ in our lethal model of ing in fatal liver and tissue damage. In our model, we found cy- toxic shock. Some TNF-␣ was also detected in the peritoneum but tokine release was biphasic with an early burst of IL-2, TNF-␣, the cellular source could not be identified. Importantly, we showed and IL-6 followed by a gradual increase in serum IFN-␥, preceded that the increase in TNF-␣ in the spleen originated from CD3 by a rise in IL-12, consistent with an ongoing Th1 response and the positive cells, providing direct evidence for the first time that T timing of death is concomitant with high serum levels of IFN-␥. cells are responsible for the very early burst of TNF-␣ seen during To directly address whether the Th1 response was central to toxic shock. The rapidity of TNF-␣ release in vivo may result from lethality in the D-galactosamine-sensitized model of toxic shock, the proximity between HLA class II-expressing and TCR-express- we compared HLA-DR1 with and without an intact TCR␣␤ rep- ing cells in a solid organ such as the spleen which contrasts with ertoire. Protection of HLA-DR1 TCR␣Ϫ/Ϫ mice from toxic shock the cumulative release of TNF-␣ and classical Th1 cytokines seen correlated with lower serum levels of IL-6, IL-12, TNF-␣, and in vitro following superantigen stimulation. IFN-␥. However, it was noteworthy that lack of functional ␣␤ T In summary, using a novel transgenic model we have shown that cells markedly affected the early TNF-␣ and IL-6 response, in mortality and serum cytokines in toxic shock are critically depen- addition to the expected effects on the later Th1 cytokine surge. dent on the presence of TCR␣␤ T cells. The accumulation of Th1 Previous reports have found that weight loss correlated with su- cytokines before death is not the decisive factor in lethality; le- perantigen dose and lethality (10, 25) so that mice protected from thality is crucially dependent on an early burst of TNF-␣ that orig- toxic shock did not lose weight. In contrast, we found that weight inates from TCR␣␤ T cells in the spleen. loss was not entirely controlled by TCR␣␤ T cells. Both IL-6 and TNF-␣ have been implicated in weight loss (10) and it may be than Disclosures the levels of these cytokines in HLA-DR1 TCR␣Ϫ/Ϫ mice are The authors have no financial conflict of interest. sufficient to cause transient weight loss. There are, however, several pieces of evidence which do not References 1. Martin, G. S., D. M. Mannino, S. Eaton, and M. Moss. 2003. The epidemiology support the assumption that toxic shock in the D-galactosamine of sepsis in the United States from 1979 through 2000. N. Engl. J. Med. 348: sensitized model is caused by accumulation of superantigen-in- 1546–1554. The Journal of Immunology 6877

2. Benjamim, C. F., C. M. Hogaboam, and S. L. Kunkel. 2004. The chronic con- 28. Roggiani, M., J. A. Stoehr, B. A. Leonard, and P. M. Schlievert. 1997. Analysis sequences of severe sepsis. J. Leukocyte Biol. 75: 408–412. of toxicity of streptococcal pyrogenic exotoxin A mutants. Infect. Immun. 65: 3. Proft, T., and J. D. Fraser. 2003. Bacterial superantigens. Clin. Exp. Immunol. 2868–2875. 133: 299–306. 29. Earhart, C. A., D. T. Mitchell, D. L. Murray, D. M. Pinheiro, M. Matsumura, 4. Schlievert, P. M. 1993. Role of superantigens in human disease. J. Infect. Dis. P. M. Schlievert, and D. H. Ohlendorf. 1998. Structures of five mutants of toxic 167: 997–1002. shock syndrome toxin-1 with reduced biological activity. Biochemistry 37: 5. Azuma, K., K. Koike, T. Kobayashi, T. Mochizuki, K. Mashiko, and 7194–7202. Y. Yamamoto. 2004. Detection of circulating superantigens in an intensive care 30. Car, B. D., V. M. Eng, B. Schnyder, L. Ozmen, S. Huang, P. Gallay, unit population. Int. J. Infect. Dis. 8: 292–298. D. Heumann, M. Aguet, and B. Ryffel. 1994. Interferon ␥ receptor deficient mice 6. Irwin, M. J., K. R. Hudson, J. D. Fraser, and N. R. J. Gascoigne. 1992. Entero- are resistant to endotoxic shock. J. Exp. Med. 179: 1437–1444. ␤ toxin residues determining T-cell receptor V binding specificity. Nature 359: 31. Mountz, J. D., T. J. Baker, D. R. Borcherding, H. Bluethmann, T. Zhou, and 841–843. C. K. Edwards. 1995. Increased susceptibility of fas mutant MRL-lpr/lpr mice to 7. Ulrich, R. G., S. Bavari, and M. A. Olson. 1995. Staphylococcal enterotoxins A staphylococcal enterotoxin B-induced septic shock. J. Immunol. 155: and B share a common structural motif for binding class II major histocompat- 4829–4837. ibility complex molecules. Nat. Struct. Biol. 2: 554–560. 32. Florquin, S., Z. Amraoui, and M. Goldman. 1995. T cells made deficient in 8. Fast, D. J., P. M. Schlievert, and R. D. Nelson. 1989. Toxic shock syndrome- interleukin-2 production by exposure to staphylococcal enterotoxin B in vivo are associated staphylococcal and streptococcal pyrogenic toxins are potent inducers primed for interferon-␥ and interleukin-10 secretion. Eur. J. Immunol. 25: of tumour necrosis factor production. Infect. Immun. 57: 291–294. 1148–1153. 9. Hackett, S. P., and D. L. Stevens. 1993. Superantigens associated with staphylococcal and streptococcal toxic shock syndrome are potent inducers of tumour 33. Altmann, D. M., D. Douek, A. J. Frater, C. M. Hetherington, H. Inoko, and necrosis factor-␤ synthesis. J. Infect. Dis. 168: 232–235. J. I. Elliott. 1995. The T cell response of HLA-DR transgenic mice to human 10. Matthys, P., T. Mitera, H. Heremans, J. Van Damme, and A. Billiau. 1995. Anti-␥ myelin basic protein and other antigens in the presence and absence of human interferon and anti-interleukin-6 antibodies affect staphylococcal enterotoxin B- CD4. J. Exp. Med. 181: 867–875. induced weight loss, hypoglycaemia, and cytokine release in D-galactosamine- 34. Philpott, K. L., J. L. Viney, G. Kay, S. Rastan, E. M. Gardiner, S. Chae, sensitized and unsensitized mice. Infect. Immun. 63: 1158–1164. A. C. Hayday, and M. J. Owen. 1992. Lymphoid development in mice congen- 11. Miethke, T., C. Wahl, K. Heeg, B. Echtenacher, P. H. Krammer, and H. Wagner. itally lacking T cell receptor ␣␤-expressing cells. Science 256: 1448–1452. Downloaded from 1992. T cell-mediated lethal shock triggered in mice by the superantigen staph- 35. Sheehan, K. C. F., N. H. Ruddle, and R. D. Schreiber. 1989. Generation and ylococcal enterotoxin B: critical role of tumour necrosis factor. J. Exp. Med. 175: characterisation of hamster monoclonal antibodies that neutralise murine tumour 91–98. necrosis factors. J. Immunol. 142: 3884–3893. 12. MacDonald, H. R., R. K. Lees, S. Baschieri, T. Herrmann, and A. R. Lussow. 36. Spitalny, G. L., and E. A. Havell. 1984. Monoclonal antibody to murine ␥ in- 1993. Peripheral T-cell reactivity to bacterial superantigens in vivo: the response/ terferon inhibits lymphokine-induced antiviral and macrophage tumoricidal ac- anergy paradox. Immunol. Rev. 133: 105–117. tivities. J. Exp. Med. 159: 1560–1565. 13. Peavy, D. L., W. H. Adler, and R. T. Smith. 1970. The mitogenic effects of 37. Welcher, B. C., J. H. Carra, L. DaSilva, J. Hanson, C. S. David, M. J. Aman, and

endotoxin and staphylococcal enterotoxin B on mouse spleen cells and human S. Bavari. 2002. Lethal shock induced by streptococcal pyrogenic exotoxin A in http://www.jimmunol.org/ peripheral lymphocytes. J. Immunol. 105: 1453–1458. mice transgenic for human leukocyte antigen-DQ8 and human CD4 receptors: 14. Bavari, S., R. E. Hunt, and R. G. Ulrich. 1995. Divergence of human and non- implications for development of vaccines and therapeutics. J. Infect. Dis. 186: human primate lymphocyte responses to bacterial superantigens. Clin. Immunol. 501–510. Immunopathol. 76: 248–254. 38. Rust, C., D. Orsini, Y. Kooy, and F. Koning. 1993. Reactivity of human ␥␦ T 15. Herman, A., G. Croteau, R. P. Sekaly, J. Kappler, and P. Marrack. 1990. cells to staphylococcal enterotoxins: a restricted reaction pattern mediated by two HLA-DR alleles differ in their ability to present staphylococcal enterotoxins to T distinct recognition pathways. Scand. J. Immunol. 38: 89–94. cells. J. Exp. Med. 172: 709–717. 39. Morita, C. T., H. Li, J. G. Lamphear, R. R. Rich, J. D. Fraser, R. A. Mariuzza, 16. Sriskandan, S., M. Unnikrishnan, T. Krausz, H. Dewchand, S. Van Noorden, and H. K. Lee. 2001. Superantigen recognition by ␥␦ T cells: SEA recognition J. Cohen, and D. M. Altmann. 2001. Enhanced susceptibility to superantigen- site for human V␥2 T cell receptors. Immunity 14: 331–344. associated streptococcal sepsis in human leukocyte antigen-DQ transgenic mice. J. Infect. Dis. 184: 166–173. 40. Muraille, E., B. Pajak, J. Urbain, M. Moser, and O. Leo. 1999. Role and regu- lation of IL-12 in the in vivo response to staphylococcal enterotoxin B. Int. 17. Unnikrishnan, M., D. M. Altmann, T. Proft, F. Wahid, J. Cohen, J. D. Fraser, and by guest on September 30, 2021 S. Sriskandan. 2002. The bacterial superantigen streptococcal mitogenic exotoxin Immunol. 11: 1403–1410. Z is the major immunoactive agent of Streptococcus pyogenes. J. Immunol. 169: 41. McCormack, J. E., J. E. Callahan, J. Kappler, and P. C. Marrack. 1993. Profound 2561–2569. deletion of mature T cells in vivo by chronic exposure to exogenous superantigen. 18. DaSilva, L., B. C. Welcher, R. G. Ulrich, M. J. Aman, C. S. David, and S. Bavari. J. Immunol. 150: 3785–3792. 2002. Humanlike immune responses of human HLA-DR3 transgenic mice to 42. Rajagopalan, G., M. K. Smart, E. V. Marietta, and C. S. David. 2002. Staphy- staphylococcal enterotoxins. J. Infect. Dis. 185: 1754–1760. lococcal enterotoxin B-induced activation and concomitant resistance to cell 19. Yeung, R. S., J. M. Penninger, T. Kundig, W. Khoo, P. S. Ohashi, G. Kroemer, death in CD28-deﬁcient HLA-DQ8 transgenic mice. Int. Immunol. 14: 801–812. and T. W. Mak. 1996. Human CD4 and human major histocompatibility complex 43. Hoiden, I., and G. Moller. 1996. CD8ϩ cells are the main producers of IL10 and class II (DQ6) transgenic mice: supersensitivity to superantigen-induced septic IFN ␥ after superantigen stimulation. Scand. J. Immunol. 44: 501–505. shock. Eur. J. Immunol. 26: 1074–1082. 44. Cardell, S., I. Hoiden, and G. Moller. 1993. Manipulation of the superantigen- 20. Decker, K., and D. Keppler. 1972. Galactosamine induced liver injury. Prog. induced lymphokine response: selective induction of interleukin-10 or interfer- Liver Dis. 4: 183–199. on-␥ synthesis in small resting CD4ϩ T cells. Eur. J. Immunol. 23: 523–529. 21. Decker, K., and D. Keppler. 1974. Galactosamine hepatitis: key role of the nu- 45. Heeg, K., S. Bendigs, T. Miethke, and H. Wagner. 1993. Induction of unrespon- cleotide deﬁciency period in the pathogenesis of cell injury and cell death. Rev. siveness to the superantigen staphylococcal enterotoxin B: cyclosporin A resis- Physiol. Biochem. Pharmacol. 71: 77–106. tant split unresponsiveness unfolds in vivo without preceding clonal expansion. 22. Gantner, F., M. Leist, S. Jilg, P. G. Germann, M. A. Freudenberg, and G. Tiegs. Int. Immunol. 5: 929–937. 1995. Tumour necrosis factor-induced hepatic DNA fragmentation as an early 46. Miethke, T., C. Wahl, K. Heeg, and H. Wagner. 1993. Acquired resistance to marker of T cell-dependent liver injury in mice. Gastroenterology 109: 166–176. superantigen-induced T cell shock: V␤ selective T cell unresponsiveness unfolds 23. Bisno, A. L., M. O. Brito, and C. M. Collins. 2003. Molecular basis of group A directly from a transient state of hyperreactivity. J. Immunol. 150: 3776–3784. streptococcal virulence. Lancet Infect. Dis. 3: 191–200. 24. Arad, G., R. Levy, and R. Kaempfer. 2002. Superantigen concomitantly induces 47. Saha, B., D. M. Harlan, K. P. Lee, C. H. June, and R. Abe. 1996. Protection Th1 cytokine genes and the ability to shut off their expression on re-exposure to against lethal toxic shock by targeted disruption of the CD28 gene. J. Exp. Med. superantigen. Immunol. Lett. 82: 75–78. 183: 2675–2680. 25. Marrack, P., M. Blackman, E. Kushnir, and J. Kappler. 1990. The toxicity of 48. Diehl, S., and M. Rincon. 2002. The two faces of IL-6 on Th1/Th2 differentiation. staphylococcal enterotoxin B in mice is mediated by T cells. J. Exp. Med. 171: Mol. Immunol. 39: 531–536. 455–464. 49. Nagaki, M., Y. Muto, H. Ohnishi, S. Yasuda, K. Sano, T. Naito, T. Maeda, 26. Bette, M., M. K. H. Schafer, N. van Rooijen, E. Weihe, and B. Fleischer. 1993. T. Yamada, and H. Moriwaki. 1994. Hepatic injury and lethal shock in galac- Distribution and kinetics of superantigen-induced cytokine gene expression in tosamine-sensitized mice induced by the superantigen staphylococcal enterotoxin mouse spleen. J. Exp. Med. 178: 1531–1539. B. Gastroenterology 106: 450–458. 27. Tsytsykova, A. V., and A. E. Goldfeld. 2000. Nuclear factor of activated T cells 50. Schumann, J., H. Bluethmann, and G. Tiegs. 2000. Synergism of Pseudomonas transcription factor NFATp controls superantigen-induced lethal shock. J. Exp. aeruginosa exotoxin A with endotoxin, superantigen, or TNF results in TNFR1- Med. 192: 581–586. and TNFR2-dependent liver toxicity in mice. Immunol. Lett. 74: 165–172.