Alloantigen bound to agarose beads and syngeneic carrier cells are capable of stimulating mouse cytolytic T lymphocytes in vitro.

Bead-bound antigen was prepared by coupling alloantigen covalently to agarose beads. Alloantigen-bearing syngeneic carrier cells were prepared by dilution of detergent solubilized alloantigen in the presence of syngeneic spleen cells. Both types of antigen were compared to spleen cells and reconstituted membrane fragments for the ability to stimulate cytolytic thymus-dependent lymphocytes in vitro. All these types of antigen could stimulate immune but not nonimmune spleen cells to form cytolytic T lymphocytes. The amount of lytic activity obtained with the bead-bound antigen was found to be only dependent upon the amount of H-2 antigen present in the culture and independent of the number of beads.     

Cytolytic T lymphocyte responses to metabolically inactivated stimulator cell. I. Metabolic inactivation impairs both CD and LD antigen signals

The effects of metabolic inactivation of spleen cells on antigen presentation to precursors of alloreactive cytolytic T lymphocytes (T_c) were examined. By serological methods, populations inactivated by ultraviolet irradiation, glutaraldehyde fixation or plasma membrane isolation were found to retain normal levels of H-2K/D and Ia antigens. However, comparison of the antigen doses required to stimulate secondary T_c responses in mixed leukocyte culture showed that the inactivated preparations were approximately 10-fold less immunogenic than X-irradiated spleen cells. Their total inability to stimulate primary cytolytic responses pointed to at least a 100-fold impairment of immunogenicity for unprimed T_c precursors in the case of uv-irradiated and glutaraldehyde-treated stimulator cells, and at least a 10-fold impairment for membrane fragments. Experiments showing that the capacity of cell monolayers to absorb precursor T_c from unprimed spleen populations was reduced following uv-irradiation or glutaraldehyde treatment provided direct evidence that this loss of immunogenicity was due in part to suboptimal antigen presentation to precursor T_c. It is concluded that, in addition to the traditional view that these treatments damage the “LD” signal to helper T lymphocytes, metabolic inactivation also impairs recognition of “CD” determinants by precursor T_c.     

Role of I-region gene products in T cell activation. I. Stimulation of T lymphocyte proliferative responses by subcellular membrane preparations containing Ia alloantigens

A model system has been developed for exploring the requirements for activation of T cells by subcellular forms of Ia alloantigen. Lymph node cells from mice recently primed subcutaneously with viable allogeneic cells show strong proliferative responses in vitro to membrane preparations derived from cells bearing the appropriate I-region-encoded glycoproteins. This stimulation shows kinetics characteristic of a secondary response, with a peak at 24 to 48 hr. Primary responses to alloantigen-bearing membranes are weak or absent under these conditions. The predominant cell type involved in the secondary response is the Lyt-1+ T lymphocyte, and the major antigenic stimulus is the I-A subregion-encoded Ia glycoprotein. Syngeneic Ia+ accessory cells do not appear necessary for activation to occur. Detergent solubilized reconstituted membrane vesicles also will stimulate primed T lymphocytes to respond by proliferation. The applications of this approach to the study of T cell recognition of antigen and the role of nonspecific lymphokines in T cell triggering are discussed.     

Cytolytic T lymphocyte responses to metabolically inactivated stimulator cells. II. Effect of a soluble "costimulator" factor(s) in primary and secondary mixed leukocyte culture

The ability of a concanavalin A-stimulated spleen cell supernatant (“costimulator”) to overcome the effects of impaired CD and LD antigen presentation by metabolically inactivated stimulator spleen cells was examined in the primary and secondary cytolytic T lymphocyte (T_c) response. (i) Cells inactivated by ultraviolet irradiation or mild glutaraldehyde treatment, which were unable to stimulate primary cytolytic activity on their own, generated near maximal responses in the presence of costimulator. The 30-fold lower efficiency of splenic membrane fragments as antigen in primary MLC with the supernatant indicated that the damage to immunogenicity caused by membrane isolation was not equivalent to that caused by uv light and glutaraldehyde, as has previously been assumed. (ii) Comparison of the relative effects of antigen and costimulator demonstrated that costimulator played the dominant regulatory role in primary MLC, increasing sensitivity to suboptimal antigen doses 10- to 30-fold; neither antigen nor the supernatant appeared preferentially to control the strength of the secondary response. (iii) Metabolically inactivated adult and untreated neonatal spleen cells failed to release costimulator activity in response to concanavalin A. However, the ability of the neonatal cells to induce a primary cytolytic response suggested that costimulator production by the stimulator cells themselves is not essential for primary T_c activation, and supports the hypothesis that the lack of primary immunogenicity of inactivated spleen cells reflects their failure to induce costimulator production by the responder population.     

Accessory cell stimulation of T cell proliferation requires active antigen processing, Ia-restricted antigen presentation, and a separate nonspecific 2nd signal

The roles of Ia+ accessory cells in H-2-restricted stimulation of antigen-specific T cell proliferation were explored in an in vitro model. L-glutamic acid60-L-alanine30-L-tyrosine10-(GAT) primed BALB/c nylon wool-passed T cells were depleted of Ia+ antigen-presenting cells (APC) by treatment with monoclonal anti-Ia antibody plus complement. Such cells failed to respond to soluble GAT, or to soluble GAT in the presence of phorbol myristic acetate (PMA), which is known to stimulate production of, or replace, IL-1 in vitro. Addition of gamma-irradiated syngeneic spleen cells reconstituted the response to soluble GAT, but addition of ultraviolet (UV) light-irradiated spleen cells did not, even in the presence of PMA. Preincubation of cells with GAT for 24 hr, followed by washing, then gamma irradiation, generated a cell population able to stimulate GAT-primed T cells to proliferate. The same pulsed cells exposed to UV irradiation failed to stimulate T cell responses unless PMA was added to the cultures. The relevant cells in this UV-irradiated population are Ia+. It is concluded that a finite period of time for interaction of metabolically intact APC with antigen is required before creation of an appropriate (Ia + antigen) signal recognized by the T cell. In addition to such Ia-restricted antigen presentation, however, a 2nd nonspecific signal, again requiring metabolically active APC for elaboration, is necessary for detectable T cell activation. These studies thus define 3 separable activities of APC during the process of H-2 restricted T cell activation.     

Cross reactive antigens of Acholeplasma laidlawii and L-form of Staphylococcus aureus]

We demonstrated that the membrane of Acholeplasma laidlawii PG8 and L-form of Staphylococcus aureus, both of which induce cellular immunity in BALB/c mice, were antigenically related each other. Foodpad responses of the mice immunized with a mixture of either antigen and Freund's complete adjuvant showed clearly a cross reaction when challenged with the other antigen. Cross responses to incorporate 3H-thymidine to the spleen lymphocytes of the mice immunized with either antigen occurred in the presence of the other antigen. Furthermore, the purified T cells, but not B cells, of the spleen were activated in the presence of antigen-presenting cells. These antigens existing in the membrane fractions of both microorganisms were purified by Razin's method. Finally, these membrane components of A. laidlawii and L-form of S. aureus were subjected to gel electrophoresis and transferring to nitrocellulose membrane and used to stimulate the spleen lymphocytes of the mice immunized with A. laidlawii or of non-immunized mice. The fractions representing molecular weights of approximately 45 kD, 25 kD, and 13 kD of both microorganisms consistently stimulated the lymphocytes of the immunized mice but not those of non-immunized mice.     

Inhibition of normal lymphocyte responses by cell membranes

Cell membranes bearing the appropriate antigen are known to stimulate a variety of cell-mediated immune responses. This report confirms that tumor cell membranes at doses of 2-5 micrograms protein/ml will stimulate in vitro generation of allogeneic cytotoxic T lymphocytes (CTL). However, higher doses (50-100 micrograms protein/ml) of the same membranes completely abrogate the generation of lytic activity. Responding lymphocytes are inhibited by membranes from either syngeneic or allogeneic cells. The inhibition appears to act at a proliferative or differentiation step in the generation of the CTL response, since membranes are known to have little direct effect on the lytic phase of CTL activity. Similar doses of membranes also inhibit LPS-induced B-cell proliferation. B-Cell proliferation is inhibited equally well by allogeneic and syngeneic membranes, and membranes from normal spleen cells are as inhibitory as tumor cell membranes. The inhibitory activity copurifies with the plasma membrane. The results raise important considerations regarding the use of subcellular forms of antigen in studies of lymphocyte recognition. In addition, these data suggest that cell-cell contacts might provide signals regulating the proliferation of lymphocytes.     

Papers to appear in forthcoming issues

Tolerance to the TNP haptenic determinant was induced by a single intravenous injection of trinitrophenylated syngeneic cells. Syngeneic spleen or thymus cells were capable of acting as carriers for tolerance induction while syngeneic bone marrow cells were not. Syngeneic spleen cells depleted of θ-positive and adherent cells were also suitable carriers for tolerance induction. Sonicated haptenated spleen cells, but not sonicated haptenated bone marrow cells induced tolerance. The ability of haptenated cells to induce tolerance was not correlated with their localization in lymphoid organs. Furthermore, cells recovered from the spleens of lethally irradiated animals reconstituted with bone marrow cells 1 week previously were incapable of inducing tolerance after hapten-modification. However, after 3 weeks, spleen cells from bone marrow-reconstituted mice had acquired the ability to induce tolerance. These results suggest that only certain types of syngeneic cells have the ability to act as carriers for tolerance induction; merely being syngeneic, and therefore presumably nonimmunogenic, is not sufficient to permit the cell to act as a carrier for tolerance induction.     

Pathogenetic mechanisms in immune polioencephalomyelitis: quantitative evaluation of protective and pathogenic effects of lymphoid cells.

Terminal dilution, adoptive cell transfer techniques were developed to quantify the protective effect of lymphoid cells in the pathogenesis of immune polioencephalomyelitis (IPE). The pathogenic effects of lymphoid cell populations were quantified by deleting the step of antigenic challenge. Regression curves were computer analyzed and PD50 values were compared. Immune spleen cells (ISC) from 4- to 6-week-old donors were more protective (PD50 = 4.9 +/- 1.3) than ISC from 12-month-old animals (PD50 greater than 7.0). The slopes of the regression curves also differed markedly (young mice, -0.24; old mice, -0.09). ISC were less protective in 12-month-indicator mice than in 5-month-old recipients (PD50 values of 5.2 +/- 0.8 and 3.7 +/- 0.8, respectively). When adoptive cell transfer tests were used to quantify the pathogenetic effects of donor cells it was found that ISC were pathogenetic at doses of 10(5) or less, but protective at higher doses. IPEC were pathogenetic at all test doses. When ISC were x-irradiated or sonicated the were only pathogenetic. Normal spleen or peritoneal exudate cells were neither protective nor pathogenetic. A model was developed in which mice were either thymectomized at birth (Tx), or Tx at birth and x-irradiated (500 R) 8 weeks later (Tx-XR). Sham Tx or Tx-XR mice served as controls. All of the mice were challenged with antigen (10(4) x-irradiated Ib cells). Only a portion (8/24) of the Tx mice developed IPE, indicating that resistance was T cell dependent but also involved a significant T cell independent component. The data also indicated that T cells were not pathogenetic effector cells in this model. Tx mice were not reconstituted by ISC (7/18 developed IPE), Tx-XR mice were partially reconstituted (3/12 developed IPE), but sham Tx-XR were fully restored (0/20 had IPE). Normal spleen cells did not reconstitute any of the mice.     

Separation of Adenosine Triphosphatase of HK and LK Sheep Red Cell Membranes by Density Gradient Centrifugation

Membrane fragments from high potassium (HK) and low potassium (LK) sheep red cells were separated by density gradient centrifugation. Three preparations were studied: (1) HK membranes sonicated for 20 minutes, (2) HK membranes sonicated for 3 minutes, and (3) LK membranes sonicated for 3 minutes. The adenosine triphosphatase (ATPase) activity in the maximally disrupted preparation (1) was not sensitive to Na + K and was recovered in relatively small but heavy (specific gravity 1.19) fragments which made up no more than 8 per cent of the total membrane. Both Na + K-sensitive (S) and Na + K-insensitive (I) ATPase activity were found in the more gently broken up preparations (2) and (3) but the ratio of S- to I-ATPase was much greater in HK than in LK membrane fragments. S-ATPase activity in preparation (2) was about 50 per cent that observed in HK membranes prior to sonication. S-ATPase activity was recovered from the density gradient in relatively large but light (specific gravity 1.10) fragments. As was the case with the maximally disrupted preparation (1), I-ATPase activity in both preparations (2) and (3) was recovered in small but heavy (specific gravity > 1.20) fragments. The possibility that sensitivity of sheep red cell membrane ATPase to Na + K depends on the association between units containing the enzyme(s) and large, light, phospholipid-containing components is discussed.     

Studies on the cellular site of action of macrophage RNA-antigen complexes.

Nonadherent spleen cell populations exposed in vitro to ribonucleic acid-rich preparations from mouse macrophages that had been incubated with human γ-globulin (RNA : HGG) were able to produce specific antibody, as measured by rosette-forming cells, in lethally irradiated (800 R), reconstituted, syngeneic mice. Exposure of the RNA to anti-HGG serum abrogated its ability to initiate antibody synthesis, as did monospecific anti-human γ chain and anti-human κ chain serum. Normal rabbit serum, anti-bovine albumin serum, anti-human μ chain or anti-human λ chain serum, when substituted for anti-HGG serum had no effect. Thus, the presence of both γ heavy chains and κ light chains of the antigen in the RNA moiety was indicated. Although both an adherent and a nonadherent cell were required by HGG to stimulate rosetteforming cells in irradiated mice, the need for the adherent cell was eliminated when RNA : HGG was substituted for HGG. In addition, anti-θ-treated bone marrow cells exposed to the RNA : HGG were capable of rosette-cell formation, suggesting that RNA attachment converted a T cell-dependent antigen to a T cell-independent antigen.     

Heterogeneity in the response of T cells to antigens presented by B lymphoma cells

Proliferation of antigen-specific T-cell populations was induced in cultures stimulated with antigen and a suitable source of antigen-presenting cells. Soluble (keyhole limpet hemocyanin) and particulate (horse red blood cells) antigens were presented by irradiated spleen cells and by a variety of B-lymphoma-cell lines, providing support for antigen-specific H-2-restricted T-cell responses. A marked heterogeneity was demonstrated, however, in the capacity of T-cell lines to proliferate in response to antigen presented by the B-lymphoma cells. T-cell populations were prepared from the lymph nodes of antigen-primed mice and restimulated in vitro in the presence of antigen and irradiated spleen cells. During the first six in vitro restimulations, these T-cell populations maintained the capacity to respond to antigen presented either by irradiated spleen cells or by B-lymphoma cells. Continued growth of these T-cell populations, again in the presence of antigen and irradiated spleen cells, resulted in the generation of T-cell lines which had lost the ability to respond to antigen presented by B-lymphoma cells. These lines however, fully retained the capacity to proliferate in the presence of antigen and irradiated spleen cells. T-cell clones derived from one of these lines were also unable to respond to antigen presented by B-lymphoma cells but again proliferated in the presence of antigen and irradiated spleen cells. Supernatants containing high levels of IL-1, IL-2, or IL-3 activity failed to reconstituted the antigen-specific response of T-cell lines which had lost the capacity to respond to antigen presented by B-lymphoma cells. Furthermore, titrated numbers of irradiated spleen cells, while having the capacity to support T-cell proliferation themselves, failed to synergize with B-lymphoma cells in the support of antigen-specific T-cell proliferation. Thus we have defined populations of antigen-specific, H-2-restricted T cells which do not recognize antigen presented by B-lymphoma cells and can therefore discriminate between different antigen-presenting cell types.     

Transfer of agammaglobulinemia in the chicken. I. Generation of suppressor activity by injection of bursa cells

Spleen cells from adult agammaglobulinemic (bursectomized) chickens taken 1 to 3 weeks after an injection of histocompatible bursa cells can inhibit the adoptive antibody response to B. abortus of normal spleen or bursa cells in irradiated recipients. Spleen cells from Aγ chickens not injected with bursa cells generally do not. Moreover, bursectomized chickens which have been reconstituted with spleen cells within the first week after hatching do not respond with suppressor cell formation upon bursa cell injection. This apparent “autoimmunization” with bursa cells induces suppressor T cells which are only minimally sensitive to treatment with mitomycin C or to 5000 R γ irradiation. The suppressor activity is neither induced nor potentiated by concanavalin A in vivo. It is much stronger in spleen than in thymus cells and appears to be macrophage independent and to require intact cells. The cell component which stimulates the suppressor activity is more pronounced on bursa than on spleen cells, and is at most present to a very limited extent on bone marrow, thymus, or peritoneal exudate cells. It is better represented in comparable cell numbers of Day 17 than of Day 14 embryonic bursa. The inducing cell component is present in the membrane fraction of disrupted bursa cells. Immunization with bursa cells from B locus histoincompatible chickens leads to suppressor activity against histocompatible bursa cells. Although the removal of Ig-bearing cells from bursa greatly diminishes its immunizing capacity, injection of serum IgM and IgG does not induce suppressor cells. It is suggested that tolerance to a B-cell antigen is lacking in adult Aγ chickens, resulting in an autoimmune response upon exposure to B cells. The B-cell antigen may be a cell surface-specific form of Ig, a complex of Ig and a membrane component, or a differentiation antigen which appears simultaneously with Ig during ontogeny.     

Induction of tolerance against the Mls antigen in mice no need for Mls-bearing cells

Infusion of CBA mice with spleen cells from the H-2-compatible, but Mis antigen-incompatible, strain C3H leads to a specific reduction of the MLC reactivity of the host's lymphocytes. One explanation of the reduced reactivity could be that the specifically Mls antigen-responsive CBA T cells become exhausted by intense antigen stimulation or that the infused cells actively neutralize specifically responsive cells. In this investigation, we have shown that depletion of membrane Ig-positive cells from C3H × CBA spleen cell preparations strongly reduces their capacity to stimulate CBA lymphocytes in the MLC, indicating that the Mls antigen is expressed on B but not on T cells. However, such B-cell depleted cell preparations were fully capable of reducing the MLC response of CBA hosts. Cell preparations of spleen and lymph nodes exhibited high stimulatory and inhibitory effects. Thymic cells lacked both these characteristics, whereas bone marrow cells were weak stimulators but relatively strong inhibitors. The results support the proposal that the observed reduction of MLC reactivity is due to an active process of the injected cells. The cell type which is working as an inhibitor has not been clearly defined yet.     

Immunochemically purified DR antigens in liposomes stimulate xenogeneic cytolytic T cells in secondary in vitro cultures

A monoclonal antibody directed against the human class II major histocompatibility antigen DR was generated. Use of this antibody, LB3.1, allowed isolation of large amounts of highly purified DR by immunoaffinity chromatography. The DR was reconstituted into liposomes and shown to stimulate secondary xenogeneic cytolytic T lymphocytes (CTL) specific for targets expressing DR antigens. DR digested with neuraminidase was equally as effective as native DR at stimulating CTL, while denatured DR and other purified membrane proteins were much less effective. The DR liposome-induced CTL lysed only target cells expressing class II antigens. Cytolysis of targets bearing class II antigens was blocked by DR-specific antisera.     

Evidence that the middle T antigen of polyomavirus interacts with the membrane skeleton.

The transforming protein of polyomavirus, middle T antigen, is associated with cellular membranes. We have examined the subcellular location of the middle T antigen in two different cell types by fractionation and detergent phase partitioning. Middle T antigen expressed in human cells by a recombinant adenovirus was detected primarily in the membrane skeleton. Sucrose gradient fractionation revealed that the middle T antigen was associated with complexes with molecular weights of 500,000 to 1,000,000. Several markers for cytoskeleton cofractionate with these complexes, including actin, tubulin, and vimentin. Electron micrographs of membrane skeleton prepared from cells expressing middle T antigen demonstrated that this material contained primarily fibrous structures and was clearly devoid of bilayer membranes. These structures were distinct from the filamentous structures observed in fractions enriched for cytoskeleton. Consistent with a role for membrane skeleton localization in transformation, middle T antigen was detected exclusively in fractions enriched for membrane skeleton in middle T antigen-transformed Rat-2 cells. Our results may resolve the apparent difference between middle T antigen localization as determined by immunomicroscopy and that determined by subcellular fractionation.     

Antigen presentation by neonatal murine spleen cells

The ability of murine neonatal spleen cells to present soluble antigen to T-helper cells and to produce growth factors in response to subsequent cellular interactions was studied. The T-helper-cell line (D10-G4.1) (D10), which is specific for the soluble antigen conalbumin presented on H-2-matched (H-2k) antigen-presenting cells, was used as cooperating and indicator cells in these cellular interactions. The D10 cells are TH2 T-helper cells which secrete the autocrine growth factor IL-4 and can also respond to exogenous IL-2 (T. R. Mosmann and R. L. Coffmann, Immunol. Today 8, 223, 1987). D10 cells require exogenous IL-1 for their proliferation and secrete, in addition to IL-4, IL-1 inducer factor and GM-CSF. The ability of neonatal spleen cells to present antigen and to stimulate D10 cells to produce IL-4 and proliferate is low. During antigen presentation there is an augmentation of IL-1 and IL-2 production by the antigen-presenting spleen cell population. However, neonatal spleen cells do not respond to the same levels as do adult spleen cells. The addition of exogenous IL-1 cannot repair the antigen presentation by neonatal cells. Experiments in which the antigen processing and presentation steps were separated from those requiring growth factor induction and secretion demonstrate that neonatal spleen cells are impaired in their ability to perform adequate antigen processing and presentation. The neonatal spleen cells are as competent as adult cells to cooperate with T-helper cells and secrete growth factors, provided antigen processing and presentation is performed by fully competent adult spleen cells. Experiments in which neonatal and adult antigen-presenting spleen cell populations were mixed, and others in which plastic adherent and nonadherent cells were separated, could not detect any suppressor mechanisms responsible for the low antigen presentation of neonatal cells. Thus, neonatal spleen cells are impaired in the initial stages of antigen processing and presentation. This impairment which leads to low levels of growth factor production is the major determinant in the ineffectual stimulation of T-helper cells by neonatal spleen cells.     

The induction in vitro of immunological tolerance in T lymphocytes: an inhibitory role of macrophages.

Hapten-reactive helper T cells were generated in the spleen of C57BL/6 mice primed with sulfanilated syngeneic IgG (S-MGG). Specific immunological tolerance was induced in vitro in these helper T cells, when spleen cell suspension was passed through Sephadex G-10 column to remove adherent cells and cultured in the presence of soluble S-MGG for 21 to 24 hours. On the other hand, tolerance was not inducible in unfractionated, primed spleen cells. When G-10-passed spleen cells were added to the culture dishes containing phagocytic, adherent cells of the spleen, tolerance was no more inducible in these reconstituted cell population. From these experimental results, it was concluded that macrophages played an interfering role in tolerance induction. The experimental data were also discussed in terms of macrophage function in the recognition of antigen by T lymphocytes.     

Failure of ovalbumin associated with allogeneic spleen ceels to stimulate proliferation of lymph node cells of immune mice.

Ovalbumin-pulsed spleen cells were found to stimulate thymidine uptake of lymph node cells of syngeneic mice immunized with ovalbumin in complete Freund's adjuvant after treatment of spleen cells with Mitomycin C but not after heating the spleen cells at 56degrees for 30 min. Ovalbumin-pulsed spleen cells of allogeneic mice failed to stimulate the immune lymph node cells more than unpulsed cells, although a net increase in the thymidine uptake above the allogeneic stimulation was observed when free ovalbumin was added to the mixed culture. To eliminate the high background of the mixed lymphocyte reaction, F1 mice were made chimeric with bone marrow of one of the parental strains. Using lymph node cells of the immunized chimeras, the stimulation by pulsed spleen cells was much greater when antigen was presented on cells of the parental strain used for bone marrow injection than when presented on cells of the other parental strain.     

Stimulation or tolerization of an anti-myelin basic protein T lymphocyte line with membrane fragments from antigen presenting cells.

We have investigated the abilities of a cell-free supernatant of splenocytes or thymocytes, which have been incubated with myelin basic protein (MBP), and of membranes prepared by lysing these cells, to stimulate proliferation of a Lewis rat anti-MBP T lymphocyte line in vitro. The supernatant fraction, obtained by low-speed centrifugation, is thought to contain shed membrane fragments bearing class II MHC protein (Ia) and processed antigen. Almost all of 67 preparations of supernatant fraction and about a third (26/70) of the membrane preparations stimulate proliferation of the line cells in the absence of other antigen-presenting cells and antigen. Some membrane preparations bearing the synthetic peptide S69 (residues 69-89 of MBP), containing the immunodominant encephalitogenic determinant for the Lewis rat, instead of processed MBP could also stimulate proliferation. Those membrane preparations bearing either processed MBP or synthetic S69, which do not stimulate proliferation, induce a state of unresponsiveness in which the cells do not proliferate but produce inositol phosphate. Stimulation of proliferation and induction of unresponsiveness were both inhibited by anti-Ia antibody. Addition of cyclosporin A prevents induction of unresponsiveness. Addition of allogeneic splenocytes or the cell-free supernatant fraction of syngeneic or allogeneic splenocytes or thymocytes, prevents induction of unresponsiveness by providing a necessary costimulatory signal. Further fractionation of the cell-free supernatant by high-speed ultracentrifugation showed that the costimulatory signal resided in a particulate fraction which sedimented and not in the supernatant. These results indicate that the encephalitogenic peptide can induce anergy in T cells when presented on class II MHC in the absence of the costimulatory signal. Tolerizing forms of the membrane preparations which lack the costimulatory signal may be useful for in vivo treatment of autoimmune response.