Relative activities of distinct isotypes of murine and human major histocompatibility complex class II molecules in binding toxic shock syndrome toxin 1 and determination of CD antigens expressed on T cells generated upon stimulation by the toxin

Relative abilities of murine and human major histocompatibility complex class II molecules to bind toxic shock syndrome toxin 1 (TSST-1) and T-cell subsets activated by the toxin were investigated. TSST-1 binding was observed in L cells transfected with I-Ab, I-Ak, DR2, and DQw1 genes. Scatchard plot analysis showed similar Kd values (1 x 10(-8) to 3 x 10(-8) M) for these cells. By comparison, binding was not detected in L cells transfected with I-Ek, DPw4, and DP(Cp63) genes. All of the transfectants supported TSST-1-induced proliferative response and interleukin-2 production by murine and human T cells. Levels of accessory activity were lower in the I-Ek transfectants and the DPw4 and DP(Cp63) transfectants than in the I-Ab and I-Ak transfectants and the DR2 and DQw1 transfectants, respectively. The results indicate that I-A, DR, and DQ molecules bind TSST-1 with similar affinities, whereas I-E and DP molecules bind it with fairly low affinity. TSST-1-activated T cells consisted of both CD4+ and CD8+ T cells, indicating that TSST-1 activates these two T-cell subsets.     

Structural analysis of human anti-mouse H-2E xenorecognition: T cell receptor bias and impaired CD4 interaction contribute to weak xenoresponses.

T lymphocyte responses to the MHC of an evolutionarily distant species are known to be weak compared with responses against allogeneic MHC products within a species. This fact was used to examine the regions of human MHC class II molecules required for the stimulation of strong primary immune responses against MHC alloantigens. A panel of mouse DAP.3 transfectants expressing the products of wild-type and recombinant DR1/H-2Ek MHC class II genes paired to either DR alpha or H-2E alpha genes was generated, and tested as stimulator cells for purified human T cells. A strong proliferative response to DAP.3 transfectants expressing allogeneic HLA-DR molecules was seen. In contrast, weak or absent responses were recorded against DAP.3 cells expressing H-2E molecules. Substitution of the DR1 beta chain with H-2E beta k led to a dramatic loss of recognition; alpha chain substitution had a less marked effect. Furthermore, replacement of the beta 2 domain of DR1 with H-2E sequence caused 90% inhibition, whereas introduction of the beta 2 domain of DR1 into H-2Ek led to a 10-fold increase in T cell response. These results are most readily explained if the beta 2 domain contributes to the interaction site for the CD4 molecule. Substitution of either half of the beta 1 domain led to a marked loss of response. This was more impressive following substitution of the TCR-contacting alpha-helical region of the domain.(ABSTRACT TRUNCATED AT 250 WORDS)     

Definition of sites on HLA-DR1 involved in the T cell response to staphylococcal enterotoxins E and C2

We have exploited the relative inefficiency of interaction between staphylococcal enterotoxins, SEE or SEC2, and H-2Ek compared to HLA-DR1 molecules to deduce which regions of the major histocompatibility complex (MHC) class II molecule are involved in the T cell response to these superantigens. Transfectants expressing hybrid DR/H-2E MHC class II molecules were used to present SEE to the T cell receptor Vβ8.1-expressing Jurkat cell line, and SEC2 to human peripheral blood T cells. For SEE, the critical region of the class II molecule for T cell reactivity and for binding was the β1 domain α-helix. The functional data were corroborated by measurements of direct binding. Sequence comparison between DR and H-2E raised the possibility that the glutamic acid at position 84 in the β chain of H-2Ek, in place of glycine was responsible for the observed functional effects. This suggestion was supported by the finding that DQw2 (glutamine at 84) transfectants supported the SEE response much more efficiently than DQw6 that has glutamic acid at this position. In addition, amino acid substitutions at either position 36 or 39 in the DRα1 domain abolished T cell reactivity without any obvious alteration in binding. For SEC2, use of transfectants expressing exon-shuffled α and β chain genes showed that replacement of the α1, α2 and β1 domains with H-2E sequence inhibited the presentation of SEC2. Similarly, the substitutions at positions 36 and 39 in the α1 domain abolished the Tcell response to SEC2. Taken together, these data may be best explained by a model in which these two toxins have primary binding sites on the β1 domain (SEE) and the α1 and α2 domains (SEC2), but by virtue of a secondary binding site on the opposite surface of the class II molecule, cross-link two adjacent DR molecules. Such cross-linking may be important in the induction of T cell reactivity.     

Major histocompatibility complex class II-associated peptides determine the binding of the superantigen toxic shock syndrome toxin-1

Superantigens bind to major histocompatibility complex (MHC) class II proteins and interact with variable parts of the T cell antigen receptor (TCR) β-chain. Cross-linking the TCR with MHC class II molecules on the antigen-presenting cell by the superantigen leads to T cell activation that plays an essential role in pathogenesis. Recent crystallographic data have resolved the structure of the complexes between HLA-DR1 and staphylococcal enterotoxin B (SEB) and toxic shock syndrome toxin-1 (TSST-1), respectively. For TSST-1, these studies have revealed possible contact sites between the superantigen and the HLA-DR1-bound peptide. Here, we show that TSST-1 binding is dependent on the MHC-II-associated peptides by employing variants of T2 mutant cells deficient in loading of peptides to MHC class II molecules as superantigen-presenting cells. On HLA-DR3-transfected T2 cells, presentation of TSST-1, but not SEB, was dependent on HLA-DR3-associated peptides. Thus, although these superantigens can be recognized in the context of multiple MHC class II alleles and isotypes, they clearly bind to specific subsets of MHC molecules displaying appropriate peptides.     

The peptide-binding motif of HLA-DR8 shares important structural features with other type 1 diabetes-associated alleles

The objective of this study was to characterize the peptide-binding motif of the major histocompatibility complex (MHC) class II HLA-DR8 molecule included in the type 1 diabetes-associated haplotype DRB1(*)0801-DQA1(*)0401/DQB1(*)0402 (DR8-DQ4), and compare it with that of other diabetes-associated MHC class II alleles; DR8-bound peptides were eluted from an HLA-DR homozygous lymphoblastoid cell line. The repertoire was characterized by peptide sequencing using a LTQ ion trap mass spectrometer coupled to a multidimensional liquid chromatography system. After validation of the spectra identification, the definition of the HLA-DR8 peptide-binding motif was achieved from the analysis of 486 natural ligands, based on serial alignments of all possible HLA-DR-binding cores. The DR8 motif showed a strong similarity with the peptide-binding motifs of other MHC class II diabetes-associated alleles, HLA-DQ8 and H-2 I-A(g7). Similar to HLA-DQ8 and H-2 I-A(g7), HLA-DR8 preferentially binds peptides with an acidic residue at position P9 of the binding core, indicating that DR8 is the susceptibility component of the DR8-DQ4 haplotype. Indeed, some DR8 peptides were identical to peptides previously identified as DQ8- or I-A(g7) ligands, and several diabetes-specific peptides associated with DQ8 or I-A(g7) could theoretically bind to HLA-DR8. These data further strengthen the association of HLA-DR8 with type I diabetes.     

Hb(64-76) epitope binds in different registers and lengths to I-Ek and I-Ak

The nature of peptide binding to MHC molecules is intrinsically degenerate, in what, one given MHC molecule can accommodate numerous peptides which are structurally diverse, and one given peptide can bind to different alleles. The structure of the MHC class II molecules allows peptides to extend out of the binding groove at both ends and these residues can potentially influence the stability and persistence of peptide/class II complexes. We have previously shown that both I-E(k) and I-A(k)-restricted T cell hybridomas could be generated against the Hb(64-76) epitope. In this study, we characterized the binding register of the Hb(64-76) epitope to I-A(k), and showed that it was shifted by one residue in comparison to its binding to I-E(k), and did not use a dominant anchor residue at P1. This conclusion was further supported by the modeling of the Hb(64-76) epitope bound to I-A(k), which revealed that all of its putative anchor residues fit into their corresponding pockets. We identified the naturally processed Hb epitopes presented by both I-E(k) and I-A(k), and found that they consisted of different species. Those associated with I-A(k) being 20-22 residues long, whereas, those found to I-E(k) contained 14-16 residues. These findings suggested that the lack of a dominant P1 anchor could be compensated by the selection of longer peptides. Overall, these studies revealed the Hb(64-76) epitope bound to I-E(k) and I-A(k) in distinct registers and lengths, demonstrating the plasticity MHC molecules have in generating distinct TCR ligands from the same amino acid sequence.     

B-B cell interaction involved in polyclonal B cell activation is restricted by I-A but not by I-E molecules

We have previously demonstrated that the class II MHC restricted B-B cell interaction is involved in the polyclonal differentiation of unprimed murine B cells into IgM-producing cells induced by a T cell-derived lymphokine B151-TRF2 or bacterial LPS. The present study has addressed the question of whether I-A and/or I-E molecules function as restriction elements for the B-B cell interaction. The results revealed that (B10 x B10.BR)F1(H-2b/k) B cells could be separated into I-Ab- and 1-Ak-restricted subpopulations by their ability to bind to B10(H-2b) or B10.BR(H-2k) B cell monolayers, whereas an I-E-restricted F1 B cell population was not obtained. Moreover, B10-derived B cells isolated from (B10 + B10.BR) - (B10 x B10.BR)F1 but not from B10 - (B10 x B10.BR)F1 radiation-induced bone marrow chimeras acquired newly the ability to co-operate with mitomycin C-treated auxilary B cells expressing I-Ak but not I-Ek molecules. Thus, these results indicate that I-E molecules, unlike I-A molecules, do not serve as restriction elements for the B-B cell interaction, and that I-A and I-E molecules on B cells play functionally disparate roles in the activation of polyclonal B cells.     

H-2 I-E molecules isolated from MIs1a stimulatory cells do not activate Mls1a-responsive T cells but do present exogenous staphylococcal enterotoxins

The T cell response to allogeneic murine Mls determinants is not H-2 restricted but is dependent on H-2 class II molecules on the Mls-expressing stimulator cells. We have tested planar membranes containing H-2 class II I-E molecules alone or with I-A molecules for their ability to activate a panel of Mls1^a-specific T hybrids. Despite the ability of the planar membranes to activate an alloreactive T hybrid and to present staphylococcal enterotoxins or an antigenic peptide to appropriately responsive T hybrids, they failed to stimulate the Mls1^a-specific T hybrids. These findings, in the light of the various controls demonstrating sufficiency of the I-E molecules in the planar membranes, indicate that Mls1^a determinants are not covalently bound to I-E molecules; the two molecular species are thus either not physically associated or are linked by a relatively weak interaction. In addition, our experiments show that isolated I-E molecules but not I-A molecules present staphylococcal enterotoxins A and B to two independently derived T hybrids expressing T cell receptor Vβ1, Vβ8.2 and Vβ6 elements.     

Cytokine Induction by Mycoplasma arthritidis-Derived Superantigen (MAS), but not by TSST-1 or SEC-3, is Correlated to Certain HLA-DR Types

Superantigens bind to major histocompatibility complex (MHC) class II molecules on antigen presenting cells and T cells in a Vβ-restricted manner. Both cell types are activated resulting in cytokine production. Although the MHC-II binding site for superantigens has been well described, little is known as to whether this binding complex has an influence on cytokine induction. In order to assess superantigen induced cytokine production and its correlation to HLA-DR types, the authors stimulated peripheral blood from 40 subjects with superantigens toxic shock syndrome toxin-1 (TSST-1), staphylococcal enterotoxin C-3 (SEC-3) and Mycoplasma arthritidis -derived superantigen (MAS), and measured cytokine levels thereafter. The HLA-DR type was determined in each subject. A statistical evaluation was carried out between the highest superantigen cytokine induction and the presence of certain HLA-DR types. Whereas MAS presented a statistical association between the highest cytokine production with HLA-DR4, DR7 and DR12, no such associations were observed for TSST-1 and SEC-3. These results demonstrate that T cell stimulation, and consequently its cytokine production by MAS but not by TSST-1 and SEC-3, depends on the presenting HLA-DR type. Because the diverse HLA-DR specificities are given according to the variability of the β chain of the HLA-DR molecule, the data suggest the participation of the human MHC-II β chain in the MAS/MHC-II binding.     

Replacement of the membrane proximal region of I-A(d) MHC class II molecule with I-E-derived sequences promotes production of an active and stable soluble heterodimer without altering peptide-binding specificity

The MHC class II molecule I-A is the murine homologue of HLA-DQ in humans. The I-A and DQ heterodimers display considerable heterodimer instability compared with their I-E and HLA-DR counterparts. This isotype-specific behavior makes the production of soluble I-A and DQ molecules very difficult. We have developed a strategy for production of soluble I-A^d molecules involving expression of I-A^d as a glycosil phosphatidyl inositol (PI) anchored chimera in Chinese Hamster Ovary (CHO) cells. The regions comprising the membrane proximal segments of I-A^d alpha and beta chains were substituted for the corresponding regions of I-E, and the derived constructs were expressed in CHO cells. Procedures for purification of the soluble class II molecules were optimized and the WT and chimeric molecule were compared for structure, biochemical stability and functionality. Our analysis revealed that the substitutions in the membrane proximal domains improved cell surface expression and thermal stability of I-A^d without altering the peptide binding specificity of the class II molecule. The results suggest that similar strategies could be used to increase the stability of other unstable class II molecules for in vitro studies.     

Functional reconstitution of class II major histocompatibility complex molecules, I-A alpha k and I-A beta k

Genomic DNA fragments carrying the genes, I-A alpha k and I-A beta k, encoding the class II major histocompatibility complex (MHC) molecule I-Ak were inserted into a transducing vector pRSVneo. This vector was introduced into a B lymphoma line expressing the class II molecules (H-2d). The resultant transformants expressed I-Ak molecules on the membrane surface. These transformants were capable of presenting antigen to I-Ak restricted helper T (Th) cells and were susceptible to lysis by class II specific killer T cells. The vector pR-A alpha kA beta k which contains both I-A alpha k and I-A beta k genes will provide a useful tool to analyze function of Ia molecules in cellular interaction, recognition, regulations and stimulation.     

Characterization of the interaction of a TCR alpha chain variable domain with MHC II I-A molecules.

The alphabeta TCR recognizes peptides bound to MHC molecules. In the present study, we analyzed the interaction of a soluble TCR alpha chain variable domain (Valpha4.2-Jalpha40; abbreviated to Valpha4.2) with the MHC class II molecule I-Au. Valpha4.2 bound specifically to I-Au expressed on the surface of a transfected thymoma cell line. Modifications in the amino acid residues located within the three complementarity-determining regions (CDRs) of the Valpha domain did not markedly affect this interaction. However, mutation of glutamic acid to alanine at position 69 of the fourth hypervariable region (HV4alpha) significantly increased the binding. Antibody inhibition studies suggested that the binding site was partly contributed by a region of the beta chain of I-Au. Furthermore, the binding of Valpha4.2 to the MHC molecule was dependent on the nature of the peptide bound in the groove. Soluble Valpha4.2 specifically inhibited the activation of TCR transfectants by I-Au-expressing cells pulsed with an N-terminal peptide of myelin basic protein. Valpha4.2 also bound to MHC class II-expressing spleen cell populations from mice of the H-2(u) and H-2(d) haplotypes. The binding of Valpha4.2 to I-A molecules might explain the immunoregulatory effects reported previously for TCR alpha chains. This Valpha4.2 interaction may also be relevant to models of antigen presentation involving the binding of intact proteins to MHC class II molecules followed by their processing to generate epitopes suitable for T cell recognition.     

Site of nonrestrictive binding of SEA to class II MHC antigens

We have used the synthetic peptide approach to show that the N-terminal 45-amino acids of staphylococcal enterotoxin A (SEA), SEA(1-45), constitute an important part of its binding site on class II major histocompatibility complex (MHC) molecules. SEA(1-45) and to a lesser extent SEA(1-27) were able to displace SEA from HLA-DR on Raji cells as assessed by flow cytometry and to compete with radiolabeled SEA for interaction with HLA-DR in a direct binding assay. Specific binding of SEA to Ia on murine A-20 cells could be inhibited by the same peptides [i.e. SEA(1-45) greater than SEA(1-27)] that blocked binding to HLA-DR. Therefore, different class II MHC molecules associate with the same functional site on SEA. Further, an ELISA system was used to demonstrate that SEA(1-45) is able to directly bind to a mouse synthetic I-A beta b peptide, I-A beta b (65-85), which contains a binding site of the class II MHC molecule involved in SEA presentation to T cells. Thus, we have localized a site on SEA that is involved in selective surface association with class II MHC antigens and identified the region on the class II MHC antigen to which that site binds.     

Activation of human T cells by toxic shock syndrome toxin-1: the toxin-binding structures expressed on human lymphoid cells acting as accessory cells are HLA class II molecules

Toxic shock syndrome toxin-1 (TSST-1)-binding assay using 125I-labeled TSST-1 showed the presence of specific TSST-1 binding in a B cell fraction of human peripheral blood mononuclear cells and L cells transfected with DR2 genes or DR4 genes but not in a T cell fraction and control L cells. Fixation with paraformaldehyde, an inhibitor of antigen processing, did not remove TSST-1-binding activity of the transfectants. Binding of 125I-labeled TSST-1 to the transfectants was reduced by an anti-DR monoclonal antibody. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed the presence of a single band with TSST-1-binding activity and the same migration pattern as DR heterodimers. TSST-1-induced T cell responses, proliferation and interleukin 2 (IL2) production were observed in the presence of the transfectants but not in the presence of control L cells, while concanavalin A-induced IL2 production was observed in the presence of either the transfectants or control L cells. Presence of an anti-DR monoclonal antibody inhibited the TSST-1-induced responses. Paraformaldehyde-fixed Daudi cells were effective in supporting TSST-1-induced IL2 production by T cells. These results indicate that HLA class II molecules directly bind intact TSST-1 and perform an essential role as the TSST-1-binding structures on accessory cells in T cell activation by the toxin.     

Characterization of defective I-A surface expression in a mixed isotype expressing murine B cell lymphoma: continued expression of E alpha d A beta d despite competition from restored A alpha d A beta d pairs.

The BALB/c-derived mouse B cell lymphoma line, 2PK3, expresses mixed isotype E alpha dA beta d and classical I-E class II molecules on its surface, but normal surface I-A expression is not detectable. Northern blot analysis showed comparable amounts of A alpha mRNA in 2PK3 as compared to another Iad expressing B cell lymphoma, A20, which predominantly expresses I-A and I-E. Sequence analysis of 2PK3 A alpha cDNA revealed a single nucleotide difference in the signal sequence that would result in a proline for leucine substitution at position - 12. In vitro translation of 2PK3 A alpha mRNA gave results suggesting that the signal peptide mutation prevented translocation of the A alpha protein across the rough endoplasmic reticulum which would provide an explanation for the lack of I-A expression in 2PK3. I-A expression was restored by transfecting a functional A alpha d gene into 2PK3. Although I-A was expressed at high levels in some transfectants, in all cases significant levels of mixed isotype were still detected. Functional studies performed using antigen-specific I-A(d)-restricted and E alpha d-A beta d-specific T cell hybridomas confirmed the levels of expression of I-A(d) and E alpha dA beta d respectively on the transfectants and showed that these molecules were functional. An interesting observation from this study is the continued expression of significant levels of E alpha dA beta d in spite of competition from restored expression of I-A(d).     

Functional evidence for the recognition of endogenous peptides by autoreactive T cell clones

The fine specificity of two human T cell clones responding to autologous HLA-DR1 expressing antigen-presenting cells (APC) in the absence of nominal antigen has been investigated using Epstein-Barr virus-transformed B cells (BCL) of known DR beta 1 domain sequence. It was found that responsiveness was markedly affected by changes in a limited number of residues in this domain. Substitution of the DR1 beta sequence at one residue, position 74, even conservatively, was found to be particularly significant. Located on the beta 1 domain alpha-helix, this residue is predicted to point into the antigen-binding groove and is therefore unlikely to make contact with the T cell receptor. This finding suggests that these T cells are specific for a bound endogenous peptide within the autologous major histocompatibility (MHC) binding groove. The autospecific T cell clones also responded to murine L cell transfectants expressing DR alpha DR1 beta as well as to transfectants expressing the mouse/human hybrid MHC molecule I-E alpha DR1 beta but not to the reciprocal combination DR alpha I-E beta, thus confirming the importance of the beta 1 domain to T cell recognition. In contrast to the autocytotoxicity observed with BCL, cytolysis of the murine L cells expressing the HLA-DR1 molecule was slight and only found at high effector-target ratios. In addition, although fixation enhanced the recognition of BCL, capacity of the murine L cells bearing the HLA-DR1 molecule to stimulate T cell clone proliferation was markedly reduced by aldehyde fixation. When taken together, these results suggest that the endogenous peptides recognized by these autoreactive T cells are of human origin.     

The peptide-binding strategy of the MHC class II I-A molecules

Despite the importance of murine major histocompatibility complex (MHC) class II I-A molecules for immunological research, the overall peptide-binding specificities of I-A and the homologous human HLA-DQ molecules remain unresolved. Here, Boris Reizis and colleagues review current evidence suggesting that DQ/I-A molecules bind peptides with a different hierarchy of anchor positions than has been found in the well-characterized DR/I-E proteins.     

A recombinant single-chain human class II MHC molecule (HLA-DR1) as a covalently linked heterotrimer of α chain, β chain,and antigenic peptide,with immunogenicity in vitro and reduced affinity for bacterial superantigens

Major histocompatibility complex (MHC) class II molecules bind to numerous peptides and display these on the cell surface for T cell recognition. In a given immune response, receptors on T cells recognize antigenic peptides that are a minor population of MHC class II-bound peptides. To control which peptides are presented to T cells, it may be desirable to use recombinant MHC molecules with covalently bound antigenic peptides. To study T cell responses to such homogenous peptide-MHC complexes, we engineered an HLA-DR1 cDNA coding for influenza hemagglutinin, influenza matrix, or HIV p24 gag peptides covalently attached via a peptide spacer to the N terminus of the DR1 β chain. Co-transfection with DR α cDNA into mouse L cells resulted in surface expression of HLA-DR1 molecules that reacted with monoclonal antibodies (mAb) specific for correctly folded HLA-DR epitopes. This suggested that the spacer and peptide did not alter expression or folding of the molecule. We then engineered an additional peptide spacer between the C terminus of a truncated β chain (without transmembrane or cytoplasmic domains) and the N terminus of full-length DR α chain. Transfection of this cDNA into mouse L cells resulted in surface expression of the entire covalently linked heterotrimer of peptide, β chain, and α chain with the expected molecular mass of approximately 66 kDa. These single-chain HLA-DR1 molecules reacted with mAb specific for correctly folded HLA-DR epitopes, and identified one mAb with [MHC + peptide] specificity. Affinity-purified soluble secreted single-chain molecules with truncated α chain moved in electrophoresis as compact class II MHC dimers. Cell surface two-chain or single-chain HLA-DR1 molecules with a covalent HA peptide stimulated HLA-DR1-restricted HA-specific T cells. They were immunogenic in vitro for peripheral blood mononuclear cells. The two-chain and single-chain HLA-DR1 molecules with covalent HA peptide had reduced binding for the bacterial superantigens staphylococcal enterotoxin A and B and almost no binding for toxic shock syndrome toxin-1. The unique properties of these engineered HLA-DR1 molecules may facilitate our understanding of the complex nature of antigen recognition and aid in the development of novel vaccines with reduced superantigen binding.     

Synergistic effect between CD40 and class II signals overcome the requirement for class II dimerization in superantigen-induced cytokine gene expression

Although staphylococcal enterotoxin A (SEA), B (SEB), and toxic shock syndrome toxin 1 (TSST-1) bind to major histocompatibility complex (MHC) class II molecules, they differ in their mode of binding. Signaling induced by these toxins via MHC class II molecules seems to be largely mediated by their mode of interaction. In the present study, we have demonstrated that contrary to SEA, stimulation of the human monocytic cell line THP-1 with SEB or TSST-1 failed to induce interleukin-1β or tumor necrosis factor-α gene expression. Treatment of THP-1 cells with interferon-γ increased the level of MHC class II expression but did not enhance the SEB and TSST-1 response. However, cross-linking of SEB or TSST-1 bound to MHC class II molecules with specific antibodies leads to cytokine gene expression, indicating that dimerization of class II molecules is a requirement for this superantigen-induced response. The presence of anti-CD40 antibodies in the course of SEB or TSST-1 stimulation overcomes this requirement, indicating that certain signal(s) induced via CD40 molecules can replace those induced by dimerization of class II molecules. Pretreatment with anti-lymphocyte functional antigen-1 (LFA-1) antibodies completely inhibited SEA-induced response as well as that induced by SEB or TSST-1 in the presence of CD40 antibodies, supporting the involvement of LFA-1 intercellular adhesion molecule system in these responses. The entirety of these results demonstrate clearly that dimerization of class II molecules is a prerequisite for superantigen-induced T cell-independent cytokine gene expression which can be replaced by signaling via CD40 in an LFA-1-dependent system.