Th2 lymphoproliferative disorder of LatY136F mutant mice unfolds independently of TCR-MHC engagement and is insensitive to the action of Foxp3+ regulatory T cells

Mutant mice where tyrosine 136 of linker for activation of T cells (LAT) was replaced with a phenylalanine (Lat(Y136F) mice) develop a fast-onset lymphoproliferative disorder involving polyclonal CD4 T cells that produce massive amounts of Th2 cytokines and trigger severe inflammation and autoantibodies. We analyzed whether the Lat(Y136F) pathology constitutes a bona fide autoimmune disorder dependent on TCR specificity. Using adoptive transfer experiments, we demonstrated that the expansion and uncontrolled Th2-effector function of Lat(Y136F) CD4 cells are not triggered by an MHC class II-driven, autoreactive process. Using Foxp3EGFP reporter mice, we further showed that nonfunctional Foxp3(+) regulatory T cells are present in Lat(Y136F) mice and that pathogenic Lat(Y136F) CD4 T cells were capable of escaping the control of infused wild-type Foxp3(+) regulatory T cells. These results argue against a scenario where the Lat(Y136F) pathology is primarily due to a lack of functional Foxp3(+) regulatory T cells and suggest that a defect intrinsic to Lat(Y136F) CD4 T cells leads to a state of TCR-independent hyperactivity. This abnormal status confers Lat(Y136F) CD4 T cells with the ability to trigger the production of Abs and of autoantibodies in a TCR-independent, quasi-mitogenic fashion. Therefore, despite the presence of autoantibodies causative of severe systemic disease, the pathological conditions observed in Lat(Y136F) mice unfold in an Ag-independent manner and thus do not qualify as a genuine autoimmune disorder.     

The type 1 cysteinyl leukotriene receptor triggers calcium influx and chemotaxis in mouse alpha beta- and gamma delta effector T cells

Linker for activation of T cells (LAT) is essential for T cell activation. Mice with mutations of distinct LAT tyrosine residues (LatY136F and Lat3YF) develop lymphoproliferative disorders involving TCR alphabeta or gammadelta T cells that trigger symptoms resembling allergic inflammation. We analyzed whether these T cells share a pattern of gene expression that may account for their pathogenic properties. Both LatY136F alphabeta and Lat3YF gammadelta T cells expressed high levels of the type 1 cysteinyl leukotriene receptor (CysLT(1)). Upon binding to the 5(S)-hydroxy-6(R)-S-cysteinylglycyl-7,9-trans-11,14-cis-eicosatetraenoic acid (LTD(4)) cysteinyl leukotriene, CysLT(1) induced Ca(2+) flux and caused chemotaxis in both LatY136F alphabeta and Lat3YF gammadelta T cells. Wild-type in vitro-activated T cells, but not resting T cells, also migrated toward LTD(4) however with a lower magnitude than T cells freshly isolated from LatY136F and Lat3YF mice. These results suggest that CysLT(1) is likely involved in the recruitment of activated alphabeta and gammadelta T cells to inflamed tissues.     

Rapid loss of IgM expression by normal murine B cells undergoing IgG1 and IgE class switching after in vivo immunization

Injection of mice with polyclonal goat anti-mouse IgD antibody (G alpha M delta) stimulates a potent T cell-dependent immune response characterized by large increases in serum IgG1 and IgE concentrations and by the generation of substantial numbers of membrane (m)IgG1+ B cells. The onset of this response occurs 6 days after G alpha M delta injection and peaks by day 7 to 8. Utilizing two color fluorescence analysis and cell sorting we demonstrate that most mIgG1-expressing B cells lack mIgM during the period of onset of Ig isotype switching (day 6). Both IgG1 and IgE are produced predominantly by mIgM- cells. On day 6, IgG1 and IgE are secreted predominantly by cells expressing mIgG1 and mIgE, respectively. By day 8, a majority of the IgG1 secretion occurs among the mIgG1- cells but virtually all IgE secretion continues to come from the mIgE+ population. B cells that strongly express mIgG1 secrete little IgM or IgE. Freshly harvested B cells expressing mIgG1, 6 days after G alpha M delta injection, have undergone substantial deletion of CH mu-specific DNA in contrast to their mIgG1- counterparts. Hence, the great majority of B cells that switch to the IgG1 or IgE isotypes in vivo rapidly lose their expression of IgM.     

Humoral immune functions in IL-4 transgenic mice

We have analyzed mice expressing IL-4 as a transgene, and found that expression of this lymphokine has profound effects on B cell function. B cells from transgenic mice exhibit phenotypic changes, including an increase in size and elevated expression of class II MHC. IL-4 increases the quantity of IgE produced by transgenic-derived B cells in response to LPS stimulation. In vivo, IL-4 markedly affects the serum Ig isotype repertoire. Serum levels of IgG1 and IgE are elevated, and levels of IgG2a, IgG2b, and IgG3 are depressed in IL-4 transgenic mice. Ag-specific antibody responses to immunization with hapten-carrier conjugates are also affected by IL-4. Transgenic mice show increased anti-hapten IgE and IgG1 and reduced anti-hapten IgG2a, IgG2b, and IgG3, compared with wild-type mice. Ag-specific IgE is substantially induced by T cell-dependent Ag, but not T cell-independent Ag, suggesting that cognate T-B interactions in addition to IL-4 are required for generating IgE responses in vivo. In vivo treatment with the anti-IL-4 mAb 11B11 reverses many of the isotype alterations in the transgenic mice, indicating that these changes arise as a direct consequence of IL-4 secretion.     

Role of antigen-specific T cell help in the generation of in vivo antibody responses. I. Antigen-specific T cell help is required to generate a polyclonal IgG1 response in anti-IgD antibody-injected mice.

A system in which injection of mice with an antibody to mouse IgD that they recognize as foreign stimulates a large, T cell-dependent IgG response was used to study whether Ag-specific T cell help is required to stimulate polyclonal (non-Ag-specific) IgG production in vivo. Igha x Ighb allotype heterozygous mice were injected with a conjugate of a foreign Ag coupled to a mAb specific for one of the two IgD allotypes expressed in these mice. This conjugate cross-links mIgD on B cells that express the recognized allotype. These cells process the conjugate and present the foreign Ag to Ag-specific T lymphocytes, which become activated. Thus, B cells of the recognized allotype can be stimulated by cross-linking of their mIgD, Ag-specific T cell help, non-Ag-specific cytokines, and non-Ag-specific contact with activated T cells. In contrast, B cells that express the Igh allotype not recognized by the Ag-anti-IgD antibody conjugate (bystander B cells) can be stimulated in this system only by non-Ag-specific cytokines and non-Ag-specific contact with activated T cells. Although both recognized and bystander B cells in conjugate-injected mice demonstrated substantial increases in size and Ia expression, only the recognized B cells were induced to synthesize DNA and to make a substantial polyclonal Ig response. Bystander B cells still failed to secrete IgG when mice were injected with an anti-IgD-Ag conjugate specific for the other Igh allotype as well as a mAb that cross-linked IgD of the bystander B cell allotype. These observations demonstrate that although non-Ag-specific cytokine and contact-mediated T cell help are sufficient to induce B cells to increase in size and Ia expression in anti-IgD antibody-injected mice, Ag-specific T cell help is required to stimulate the generation of an IgG response in these mice.     

Polyclonal activation of the murine immune system by a goat antibody to mouse IgD. IX. Induction of a polyclonal IgE response

The possibility that injection of mice with an affinity-purified goat antibody to mouse IgD (GaM delta) that stimulates polyclonal IgG1 secretion might also stimulate differentiation of B cells into IgE-secreting cells was suggested by the observation that such treatment induces T cells from those mice to secrete a lymphokine, B cell stimulatory factor 1 (BSF-1), that can stimulate both IgG1 and IgE secretion in vitro. Studies described in this paper show that injection of BALB/c mice with 200 to 3200 micrograms of GaM delta greatly increased the quantity of splenic epsilon chain-encoding mRNA, the number of spleen cells with cytoplasmic IgE, and the concentration of serum IgE 7 days after injection. Serum IgE levels obtained in these mice were approximately 100 times baseline levels and were comparable with those found in mice infected with the nematode parasite Nippostrongylus brasiliensis, but were approximately 2000-fold less than the peak serum IgG1 levels induced by GaM delta injection. Both IgE and IgG1 secretion in GaM delta injected mice were T dependent (blocked by anti-L3T4 antibody). These observations are consistent with the hypothesis that BSF-1 may play a role in the in vivo stimulation of IgE secretion and provide an easy to apply model for the investigation of in vivo regulation of IgE responses.     

A novel B cell-mediated transport of IgE-immune complexes to the follicle of the spleen

Ag administered i.v. to mice along with specific IgE or IgG2a induces higher Ab- and CD4(+) T cell responses than Ag administered alone. The IgE effect is completely dependent on the low-affinity receptor for IgE, CD23, whereas the IgG2a effect depends on activating FcgammaRs. In vitro studies suggest that IgE/Ag is presented more efficiently than Ag alone to CD4(+) T cells by CD23(+) B cells and that IgG2a/Ag is presented by FcgammaR(+) dendritic cells (DCs). In this study, we investigate in vivo the early events leading to IgE- and IgG2a-mediated enhancement of immune responses. OVA administered i.v. in PBS in combination with specific IgE binds circulating B cells after 5 min and is found in B cell follicles bound to follicular B cells (CD23(high)) after 30 min. This novel B cell-dependent route of entry is specific for IgE because IgG2a-Ag complexes were trapped in the marginal zone. OVA-specific CD4(+) T cells were found at the T-B border in the T cell zones 12 h after immunization both with IgE/OVA or IgG2a/OVA and proliferated vigorously after 3 days. The findings suggest that IgE- and IgG2a-immune complexes are efficient stimulators of early CD4(+) T cell responses and that Ag bound to IgE has a specific route for transportation into follicles.     

miRNA Signature of Mouse Helper T Cell Hyper-Proliferation

Helper T cells from a mutant mouse model, LAT Y136F, hyper-proliferate and cause a severe lymphoproliferative disease that kills the mice by six months of age. LAT Y136F mice carry a tyrosine to phenylalanine mutation in the Linker for Activation of T cells (LAT) gene. This mutation leads to a number of changes in T cells that result in altered cytokine production including increased IL-4 production, increased proliferation, and decreased apoptosis. Hyper-proliferation of the mutant T cells contributes to lymphadenopathy, splenomegaly, and multi-organ T cell infiltration. miRNAs are short non-coding RNAs that regulate expression of cohorts of genes. This study investigates which miRNAs are expressed in LAT Y136F T cells and compares these to miRNAs expressed in wild type T cells that are undergoing proliferation in two other settings. The first setting is homeostatic proliferation, which was modeled by adoptive transfer of wild type T cells into T cell-deficient mice. The second setting is proliferation in response to infection, which was modeled by infection of wild type mice with the nematode H. polygyrus. By comparing miRNA expression in these three proliferative states (LAT Y136F hyper-proliferation, homeostatic proliferation and proliferation in response to H. polygyrus infection) to expression in wild type naïve CD4⁺ T cells, we found miRNAs that were highly regulated in all three proliferative states (miR-21 and miR-146a) and some that were more specific to individual settings of proliferation such as those more specific for LAT Y136F lymphoproliferative disease (miR-669f, miR-155 and miR-466a/b). Future experiments that modulate levels of the miRNAs identified in this study may reveal the roles of these miRNAs in T cell proliferation and/or lymphoproliferative disease.     

Preferential role for NF-kappa B/Rel signaling in the type 1 but not type 2 T cell-dependent immune response in vivo.

T cell function is a critical determinant of immune responses as well as susceptibility to allergic diseases. Activated T cells can differentiate into effectors whose cytokine profile is limited to type 1 (IFN-gamma-dominant) or type 2 (IL-4-, IL-5-dominant) patterns. To investigate mechanisms that connect extracellular stimuli with the regulation of effector T cell function, we have measured immune responses of transgenic mice whose NF-kappa B/Rel signaling pathway is inhibited in T cells. Surprisingly, these mice developed type 2 T cell-dependent responses (IgE and eosinophil recruitment) in a model of allergic pulmonary inflammation. In contrast, type 1 T cell responses were severely impaired, as evidenced by markedly diminished delayed-type hypersensitivity responses, IFN-gamma production, and Ag-specific IgG2a levels. Taken together, these data indicate that inhibition of NF-kappa B can lead to preferential impairment of type 1 as compared with type 2 T cell-dependent responses.     

Polyclonal activation of xid B cells by auto-Ia-reactive T cell clones

The mechanism of T cell-dependent activation of xid B cells into Ig-producing cells was studied by employing H-2-restricted, antigen-specific T cell clones. Helper factors (B cell stimulatory factors, BSF) released from KLH-specific T cell lines could induce polyclonal Ig production in B cells from (CBA/N X BALB/c)F1 (NBF1) female mice but not from CBA/N or NBF1 male mice. Direct addition of helper T cell lines induced Ig production in xid B cells from CBA/N or NBF1 male mice. A T cell clone, MK6, which was derived from NBF1 male mice and specific against Iad determinant, could activate NBF1 male but not CBA/N B cells. Another clone, CK4, derived from CBA/N mice and having specificity against KLH plus I-Ak determinant could activate both CBA/N and NBF1 male B cells into IgM- and IgG-producing cells in the absence of KLH, and monoclonal anti-I-Ak antibody specifically blocked such activation. These results suggest that xid B cells are able to be activated by the signal provided by the recognition of Ia molecules on B cells by auto-Ia-reactive T cells. Xid B cells from CBA/N mice that had been co-cultured with a T cell line specific against I-Ak determinant for 24 hr became reactive to BSF and capable of differentiating into Ig-producing cells in the presence of BSF. The results showed that even xid B cells could be responsive to BSF if they were in a certain activation stage.     

Macrophages as effector cells of protective immunity in murine schistosomiasis. VI. T cell-dependent, lymphokine-mediated, activation of macrophages in response to Schistosoma mansoni antigens

Macrophages from Schistosoma mansoni-infected mice kill significant numbers of skin stage schistosomula and murine fibrosarcoma cells in vitro. In order to determine whether the macrophage tumoricidal and larvicidal activation observed in mice as a result of S. mansoni infection are mediated through T cell-dependent (lymphokine) or B cell-dependent (antibody or immune complex) mechanisms, the development of macrophage populations with cytotoxic activity against schistosome larvae or tumor cells was monitored in S. mansoni-infected nude or mu-suppressed mice. Whereas peritoneal cells from S. mansoni-infected congenitally athymic mice had no activity in either assay, cells from mu-suppressed S. mansoni-infected mice showed cytotoxic activity equivalent to that of cells from untreated S. mansoni-infected counterparts. Cells from mu-suppressed uninfected mice were not activated. The mu-suppressed animals had no detectable nonspecific IgM or specific antischistosome IgM, IgG, or IgE antibodies and showed a 90% reduction in numbers of splenic IgM+ cells upon fluorescence activated cell sorter analysis. These results indicate that antibody is not required for in vivo activation of macrophages during S. mansoni infection. Further experiments showed that lymphoid cells from S. mansoni infected mice respond in culture with various specific antigens (such as living or dead whole schistosomula or soluble adult worm antigens) by production of factors capable of activating macrophages from uninfected control mice to kill schistosomula or tumor cells in vitro. Macrophage-activating factors were produced by T cell-enriched, but not T cell-depleted or B cell-enriched, populations from spleens of schistosome-infected mice in response to schistosome antigen. Similar lymphokines may be responsible for the macrophage activation observed during chronic murine schistosomiasis. These observations emphasize the potential contribution of T cell-mediated immune mechanisms in resistance to S. mansoni infection.     

Cellular basis of in vitro anti-DNA antibody production: evidence for T cell dependence of IgG-class anti-DNA antibody synthesis in the (NZB X NZW)F1 hybrid

An in vitro system was designed to measure anti-DNA antibody synthesis, and the cellular basis of this autoantibody production in NZB X NZW (B/W)F1 (B/W F1) mice was analyzed. The spleen cells from old B/W F1 mice contained a number of B cells that spontaneously produced anti-DNA antibodies of both IgM and IgG classes in the absence of stimulants, thereby demonstrating that these B cells had been activated in vivo. These activated B cells could be removed by Sephadex G-10 column (G-10) filtration. Such G-10-passed, homogeneously small B cells were activated by the stimulant lipopolysaccharide (LPS) and produced both IgM and IgG class anti-DNA antibodies. The G-10-passed cells contained both B and T cells, and the cytotoxic treatment of the cells with monoclonal antibodies to T cells, anti-Thy-1 and anti-L3T4, abolished the LPS-induced IgG class, but not IgM class, anti-DNA antibody syntheses. Thus, the LPS-induced production of IgG class anti-DNA antibodies in B/W F1 mice is regulated by T cells. Reconstitution experiments revealed the requirement of T-B cell contact but not of the proliferative response of T cells. Moreover, there was no apparent adherent cell requirement. Such IgG class anti-DNA antibodies were produced only by spleen cells from old B/W F1 mice, but not from young B/W F1, NZB, NZW, and C57BL/6 mice. Like IgM class anti-DNA antibodies, LPS-induced synthesis of polyclonal IgM was T cell-independent. Only a slight reduction in the polyclonal IgG synthesis was observed after the G-10-passed cells had been treated with anti-Thy-1 antibody plus complement. This study should facilitate investigation of cell to cell interactions in the formation of autoantibodies and their correlations to immunologic abnormalities in autoimmune disease.     

Suppression of polyclonal antibody production in Trypanosoma cruzi-infected mice by treatment with anti-L3T4 antibodies

Acute Trypanosoma cruzi infection of mice results in a very marked polyclonal activation of B and T lymphocytes, accompanied by high numbers of Ig-secreting PFC and lectin-dependent effector CTL. Treatment of mice with monoclonal anti-L3T4 antibodies from the time of infection (days 0, 4, and 8) completely suppresses the polyclonal PFC response and CTL generation. Treatment of nude mice with antibody does not alter the lipopolysaccharide-induced polyclonal PFC response, and it only modulates the isotypic profile of the PFC response to T. cruzi infection, without reducing its magnitude. Furthermore, antibody-treated, T. cruzi-infected euthymic mice do not develop the typical B cell blastogenic response, but show high numbers of activated Lyt-2+ lymphoblasts in the spleen. These results indicate that effector cell generation in T. cruzi-infected mice is predominantly helper T cell-dependent.     

IgE production by normal human B cells induced by alloreactive T cell clones is mediated by IL-4 and suppressed by IFN-gamma

Seven T cell clones were established from mixed leukocyte cultures in which PBMC from two healthy donors and from one patient suffering from the hyper-IgE syndrome were stimulated by the irradiated EBV-transformed B cell lines JY or UD53. Five of seven T cell clones, after activation by co-cultivation with JY or UD53 cells, induced a low degree of IgE production by normal blood B cells. In one experiment in which the normal B cells could activate the T cell clones directly, IgE production was also observed in the absence of the specific stimulator cells. IgE production was also obtained with supernatants of the T cell clones collected 4 to 5 days after activation by their specific stimulator cells. In addition, the supernatants induced IgG, IgA, and IgM synthesis. All seven clones produced variable concentrations of IL-4 and IFN-gamma. The clones FA-28 and BG-39, which failed to induce IgE synthesis, produced, compared with the other clones tested, relatively high quantities of IFN-gamma (4700 and 2500 pg/ml, respectively). These high levels of IFN-gamma accounted for the lack of induction of IgE synthesis, because in the presence of a polyclonal anti-IFN-gamma antiserum, supernatants of FA-10 and BG-39 induced significant IgE production. In addition, the low degree of IgE production induced by supernatants of two other T cell clones (FA28 and BG24) was 15- and 3-fold enhanced, respectively, in the presence of the anti-IFN-gamma antiserum. IgE synthesis by normal B cells was also induced by rIL-4, indicating that IL-4 present in T cell clone supernatants was responsible for induction of IgE production. This notion was supported by the finding that IgE production induced by supernatant of BG-24 was strongly inhibited by a polyclonal anti-IL-4 antiserum. In contrast, IgG and IgA production induced by supernatant of BG-24 were not significantly affected by the anti-IL-4 antiserum. Only a slight inhibition of IgM synthesis was observed. Collectively, our results indicate that both recombinant and naturally produced IL-4 induce normal human B cells to synthesize IgE. However, final IgE production induced by T cell clone supernatants is the net result of the inducing and suppressive effects of IL-4 and IFN-gamma respectively, that are secreted simultaneously by the T cell clones upon activation.     

Polyclonal activation of the murine immune system by an antibody to IgD. XI. Contribution of membrane IgD cross-linking to the generation of an in vivo polyclonal antibody response

The injection of mice with a foreign, polyclonal antibody to IgD sequentially induces: 1) activation of B cells by cross-linking of their cell membrane (m) IgD; 2) B cell processing and presentation of the bound anti-IgD antibody to T cells; 3) activation of these T cells; and 4) T-dependent stimulation of B cell differentiation into IgG1 secreting cells. To determine whether the cross-linking of B cell membrane IgD and/or the resulting B cell activation that follows contribute to the generation of the polyclonal IgG1 response, we examined the abilities of three sets of anti-delta mAb or mAb fragments to stimulate polyclonal IgG1 production. Within each set mAb were matched for species and Ig isotypic determinants, but differed in avidity for IgD or in ability to cross-link IgD. In addition, experiments were performed to determine whether the anti-delta mAb had to be foreign to the immunized mouse to stimulate an IgG1 response. Results of these experiments indicate that: 1) recognition of the injected anti-delta antibody as foreign is required for the induction of a polyclonal IgG1 response; 2) the cross-linking of B cell membrane Ig, which directly activates B cells, can contribute considerably to the generation of in vivo IgG1 production; and 3) that even relatively weak cross-linking of membrane Ig by ligands that bind it with low avidity can make this contribution.     

Analysis of the virus-specific and nonspecific B cell response to a persistent B-lymphotropic gammaherpesvirus

Respiratory challenge of mice with murine gammaherpesvirus 68 (gammaHV68) results in acute replication in respiratory epithelial cells and persistent, latent infection of B cells and macrophages. gammaHV68 elicits virus-specific Ab, and also nonspecifically activates B cells to Ab production through a CD4+ T cell-dependent process. The current analysis characterizes virus-specific and nonspecific Ab production at the single cell level and investigates the requirements and nature of the nonspecific response. Virus-specific Ab-forming cell (AFC) numbers were dwarfed by the increase in total AFC in all sites examined, indicating substantial nonspecific Ab production. Clear increases and decreases in specific and total AFC numbers occurred in the lymph nodes draining the respiratory tract and the spleen, but AFC numbers in the bone marrow (BM) increased to a plateau and remained constant. The longevity of the BM response was reflected in a sustained increase in virus-specific and total serum Ab levels. Generally, the IgG2a and IgG2b isotypes predominated. Analysis of cytokine-deficient mice, CD40 ligand-deficient mice, and radiation BM chimeras lacking MHC class II expression specifically on B cells indicated that nonspecific Ab production is independent of IL-6 or IFN-gamma, and dependent on cognate CD4+ T cell help. Several observations were consistent with polyclonal B cell activation by gammaHV68, including the induction of durable serum levels of IgG reactive with mammalian dsDNA and murine type II collagen. Our findings indicate new directions for studies of this valuable model of gamma-herpesvirus pathogenesis.     

Immunoregulation in the Peyer's patch microenvironment. Cellular basis for the enhanced responses by the B cells of X-linked immunodeficient CBA/N mice

The Peyer's patches (PP) of X-linked immunodeficient (xid) CBA/N and hemizygous (CBA/N X DBA/2)F1 (CDF1) male mice contain a B cell subpopulation that expresses the Lyb-5 maturational marker and is responsive to type 2 and T cell-dependent antigens in vitro, a B cell phenotype which is absent from the spleens of xid mice. Experiments reported here show that xid spleen B cells co-cultured with B cell-depleted PP cells from xid mice differentiated into specific plaque-forming cells in response to trinitrophenyl-Ficoll (type 2) and sheep erythrocytes (T cell-dependent). Two cell types were involved in this normalization of xid B cell responses. An accessory cell activity present in the PP, but not the spleens, of both CDF1 male (xid) and CDF1 female (normal) mice was required for the response to either the type 2 or T cell-dependent antigens. In the presence of this PP accessory cell, T cells from the PP of either xid or normal mice supported responses to both classes of antigens. In contrast, T cells from the spleens of xid mice did not support the response to trinitrophenyl-Ficoll, although the splenic T cells from normal mice did synergize with PP accessory cells in allowing plaque-forming cell development by xid B cells to this type 2 antigen. The xid PP T cell activity required for the type 2 response by xid B cells was present in the Ly-1+, Lyt-2- subpopulation, and the xid PP accessory cell activity was provided by an enriched population of dendritic accessory cells. These results demonstrate the the lymphoreticular cells comprising the PP microenvironment provide effective support for the differentiation of xid B cells in response to type 2 and T cell-dependent antigens.