Secreted phospholipase A(2) induces vascular endothelial cell migration

Secreted phospholipase A(2) (sPLA(2)) regulates a variety of cellular functions. The present investigation was undertaken to elucidate the potential role of sPLA(2) in endothelial cell (EC) migration. Bovine aortic endothelial cells (BAECs) exposed to sPLA(2) placed in the lower compartment of a modified Boyden chamber displayed increased migration compared to cells exposed to vehicle. The effect of sPLA(2) on EC migration was time and dose dependent. Migration of BAECs was observed at 30 minutes, increased over 1 to 2 hours, and declined thereafter. At 2 hours of stimulation, sPLA(2) (0.01-2 micromol/L) induced 1.2- to 3-fold increased cell migration compared with media alone. Among the different sPLA(2)s tested, bee venom, Naja naja, and porcine and human pancreatic PLA(2)s all evoked a migratory response in ECs. Moreover, human synovial fluid, obtained from patients with arthritis and containing sPLA(2) activity, induced EC migration. Migration of ECs was significantly reduced after exposure to a catalytic site mutant of pancreatic sPLA(2) with decreased lipolytic activity as compared to wild-type sPLA(2). Similarly, pretreatment of human synovial fluid with p-bromophenacyl bromide, an irreversible inhibitor of sPLA(2), markedly decreased the ability of human synovial fluid to stimulate EC migration. Moreover, migration of ECs was stimulated on exposure to hydrolytic products of sPLA(2) activity including arachidonic acid, lysophosphatidic acid, and lysophosphatidylcholine. These findings suggest that sPLA(2) plays a physiologic role in induction of EC migration. Moreover, the effects of sPLA(2) on EC migration are mediated, at least in part, by its catalytic activity. (Blood. 2000;96:3809-3815)

Flow cytometric analysis of T-independent antigen binding to dinitrophenyl-specific cells

Binding of Ag to membrane Ig (mIg) can lead to either activation or desensitization of the B cell. For thymus-independent (TI) Ags the nature and concentration of the Ag determines what type of signal is delivered to the cell. These Ags are capable of directly activating B lymphocytes and are an important model system for the study of mechanisms involved in B cell responses. In this study, we quantified TI Ag binding and B cell receptor involvement as functions of TI Ag structure, concentration, and epitope density. Various epitope densities of two structurally different TI Ags, DNP-polymerized flagellin (pol) and DNP-dextran (dex), were labeled with tetramethylrhodamine isothiocyanate (TRITC) and reacted with DNP-specific murine splenic B lymphocytes and with cells of a cloned DNP-specific cell line. The amount of Ag bound to the cell surface at various doses was measured directly by flow cytometry. For each Ag and dose, FITC-labeled DNP-L-papain was used to quantitate receptor sites not occupied by Ag. Approximately 5% receptor occupancy was observed for immunogenic doses of Ag. Higher Ag concentrations that can induce tolerance caused a substantial increase in the fraction of occupied receptors. This suggests that tolerogenic responses result from an overly restrictive cross-linking of surface receptors. By comparing these data to previously published data on biologic activity of the Ags, we are able to more clearly define those conditions of Ag binding that lead to B cell activation.

Flow cytometric analysis of T-independent antigen binding to dinitrophenyl-specific cells

Binding of Ag to membrane Ig (mIg) can lead to either activation or desensitization of the B cell. For thymus-independent (TI) Ags the nature and concentration of the Ag determines what type of signal is delivered to the cell. These Ags are capable of directly activating B lymphocytes and are an important model system for the study of mechanisms involved in B cell responses. In this study, we quantified TI Ag binding and B cell receptor involvement as functions of TI Ag structure, concentration, and epitope density. Various epitope densities of two structurally different TI Ags, DNP-polymerized flagellin (pol) and DNP-dextran (dex), were labeled with tetramethylrhodamine isothiocyanate (TRITC) and reacted with DNP-specific murine splenic B lymphocytes and with cells of a cloned DNP-specific cell line. The amount of Ag bound to the cell surface at various doses was measured directly by flow cytometry. For each Ag and dose, FITC-labeled DNP-L-papain was used to quantitate receptor sites not occupied by Ag. Approximately 5% receptor occupancy was observed for immunogenic doses of Ag. Higher Ag concentrations that can induce tolerance caused a substantial increase in the fraction of occupied receptors. This suggests that tolerogenic responses result from an overly restrictive cross-linking of surface receptors. By comparing these data to previously published data on biologic activity of the Ags, we are able to more clearly define those conditions of Ag binding that lead to B cell activation.

The association of alcohol consumption with self-reported illness in university students

Many reports over the years have indicated an association between alcohol consumption and infectious illness among chronic heavy drinkers; however, many patients in these studies have been chronically ill. Thus the question of whether alcohol can appreciably influence immunity in humans and affect the incidence of infectious diseases remains largely unanswered. For this study over 1,100 undergraduate students from a general education course at a large midwestern university were surveyed. Students were asked about their drinking habits and acute health problems. Analyses of their self-reports showed no increase in acute health problems or upper respiratory infections in students drinking between one and 21 drinks per week. However, students drinking 28 or more alcoholic drinks per week had significantly more health problems in the aggregate and those drinking more than 22 drinks per week had more upper respiratory infections than other students including nondrinkers. It was concluded that excessive alcohol intake increased the risk of respiratory infections and acute illnesses in these students, but more moderate alcohol consumption had little effect on the risk for these health problems.

Alcohol can inhibit effect of IL-4 on activated murine B cells

Ethanol addition to B lymphocytes in vitro inhibits antigen-induced antibody secretion and antigen-induced cell proliferation. We tested the hypothesis that ethanol can inhibit the effect of interleukins on B-cell proliferation by adding ethanol to in vitro functional assays for IL-2 and IL-4. Ethanol does not inhibit IL-2-induced B-cell differentiation, but it does inhibit IL-4-induced B-cell proliferation and IL-4-induced Ig class switching.

Ethanol inhibits production of messenger ribonucleic acid for kappa-chain in stimulated B lymphocytes

Alcohol consumption can adversely affect an individual's response to infection. We have shown that alcohol has a direct suppressive effect on numbers of antibody-secreting cells in antigen-activated B lymphocytes but does not suppress the early membrane and intracellular events that are associated with binding of antigens to specific receptors. The studies reported here were designed to determine whether alcohol inhibits immunoglobulin synthesis. When 150 mg/dl ethanol was added to anti-mu-stimulated purified B cells, proliferation was inhibited. Similar exposure to ethanol inhibited production of messenger RNA for kappa chain in anti-mu-stimulated B cells but did not affect total RNA production or messenger RNA for beta-actin. Thus alcohol inhibits both the proliferation of antigen-activated B lymphocytes and their synthesis of immunoglobulin.

Direct tolerance induction in mature B lymphocytes may resemble clonal anergy phenomenon

Several mechanisms have been postulated for direct induction of B-cell tolerance in mature B cells, such as the Bretscher and Cohn [5] hypothesis which states that an antigenic signal to B cells without the 'second signal' provided by T cells or lymphokines produces unresponsiveness. A second explanation is that tolerogens which cross-link immunoglobulin and Fc receptors abort the biochemical activation of the B cell via the phosphatidyl inositol pathway. Results of our studies are not consistent with either of these hypotheses. We speculated that DNP-MGG induces immunoglobulin receptor capping and internalization in mature B cells, but a suppressive signal is given to the B cell when DNP-MGG is present at the time of antigen receptor reexpression in a fashion similar to clonal anergy. The studies reported here test this hypothesis as a possible additional mechanism of direct B-cell tolerance. When pure DNP-specific B cells were incubated for 6 h with DNP-MGG in the presence of lymphokine-rich EL-4 supernatant, an immune response was induced; but 48 h preincubation with DNP-MGG in the presence of lymphokines induced tolerance. If B-cell cultures were preincubated with DNP-MGG for 6 h followed by a 24-hour incubation without conjugate and DNP-MGG was then added for the third preincubation period of 18 h (at a time when receptors are being reexpressed), tolerance was induced. Substitution of DNP-Ficoll antigen for DNP-MGG in either the first or third time period did not result in tolerance induction but substituting DNP-KLH during either of these two periods did result in tolerance.

B cell hyperactivity in autoimmune continuous B cell lines

Generalized increase in immunoglobulin secretion, which is a prominent feature of autoimmune diseases, may be due to abnormal T cell regulation, intrinsic abnormality of B cells, or both. To investigate this question we developed nonmalignant continuous B lymphocyte lines from 20-week-old BWF1 mice and compared their growth and immune response to that of BALB/c mice cell lines. The B cell lines contain less than 1% T cells and macrophages and require growth factors from phytohemagglutinin-stimulated EL-4 lymphoma (GF) or recombinant interleukin 4 for continuous growth. No antigens or mitogens are required for growth. In the presence of 20% GF (which is optimal for BALB/c cell growth and immune function) spontaneous growth of BWF1 B cells, and spontaneous entry into G1, was similar to that of BALB/c B cells. With concentrations of GF and anti-mu which were optimal for BALB/c, the growth and immune response of isolated BWF1 B cells are no different from those of BALB/c controls, but at suboptimal doses of GF there is a significant increase of both spontaneous immunoglobulin secretion and response to anti-mu in BWF1 B cells. Thus, these autoimmune B cells are more sensitive to the effects of both T cell factors and immunoglobulin receptors stimulation.

Immunologic tolerance to DNA in B cell lines from both normal and autoimmune mice

We examine whether B cell lines enriched for DNA specificity from either autoimmune (BWF1) or normal mice (Balb/c) can be rendered unresponsive to autoantigen in terms of the specific suppression of direct antibody-forming cells to DNA. These B cell lines were both Lyt-1 positive and negative. Preincubation with oligonucleotide, covalently linked to mouse gamma-globulin, specifically suppressed the antigen-driven response elicited by DNA horse red blood cells in B cell lines from both strains of mice. There is a 5-fold difference in susceptibility to DNA-specific tolerance induction between B cell lines of BWF1 and Balb/c mice. Thus, B cells from autoimmune mice do not appear to have an inherent absolute defect in being rendered tolerant to autoantigen, but are relatively less susceptible to DNA-specific tolerance than nonautoimmune cell lines.

Mechanisms of alcohol-induced suppression of B-cell response

Several investigators have shown that alcohol can suppress the production of antibodies by animals and humans. The studies reported here were designed to determine whether alcohol has a direct effect on the B-lymphocyte and to determine which stage of the B-cell response is inhibited by alcohol. B-lymphocyte lines specific for the antigen dinitrophenyl were used to study the effect of alcohol on the B-lymphocyte. As little as 100 mg% of alcohol inhibited the response of these pure B-cell lines to stimulation by either antigen (dinitrophenyl-Ficoll) or anti-mu antibody. Since no other cell types were present in the system the suppressive effect was on the B-cell itself. However, alcohol did not inhibit membrane depolarization induced by antigen crosslinking of immunoglobulin receptors, and it did not inhibit activation of the phosphatidyl inositol pathway by receptor crosslinking. When alcohol was added to antigen stimulated B-lymphocyte lines for varying periods during the immune response it was found that antibody production was inhibited if 150 mg% alcohol was present from 30 to 48 hr after the antigen was added to lymphocytes. Thirty-six to 48 hr is the time required for a stimulated mature B-lymphocyte to enter the proliferative phase of the immune response. These data raise the possibility that low doses of alcohol can inhibit antigen-induced B-cell proliferation.

Investigations of intrinsic abnormalities in DNA-specific B lymphocytes from autoimmune mice

In murine models of systemic lupus erythematosus and in many humans with SLE, antibodies against native DNA (dsDNA) are a major contributor to the pathogenesis of the disease. Loss of self-tolerance to the DNA antigen may be associated with B-cell defects or regulatory cell dysfunction. We have developed B-cell lines with specificity for the antigen DNA, from both the autoimmune BWF1 mouse strain and from the non-autoimmune BALB/c strain, to use in the investigation of inherent B-cell defects in autoimmunity. Six BWF1 cell lines and five BALB/c cell lines which are free of Thy1.2+ cells and esterase positive cells, and have between 35 and 89% rosetting with dsDNA-SRBC targets, have been propagated in vitro for 24-36 months. The cells are non-malignant, growth-factor dependent and have no antigen or mitogen in the growth medium. Lyt-1 positive cells are found in the cell lines, but Lyt-1 negative cells are also present. They respond to the antigen DNA-HRBC when EL-4 supernatant is present in culture, and the peak of the plaque-forming cell (PFC) response is the same for both strains. When cells from both strains are cultured with varying amounts of T-cell factors, there is no difference in spontaneous antibody-forming cell (AFC) formation or in response to anti-mu stimulation between BWF1 and BALB/c strains. BALB/c spleen cells do not respond to DNA-HRBC in this culture system, but BWF1 spleen cells, as well as cell line cells from both strains, respond to this antigen. T cells from non-responding BALB/c spleen and responding BWF1 spleen are able to suppress the immune response to DNA-HRBC of cell line B cells from both strains. Propagating B-cell lines in the presence of DNA for 2 weeks stimulates BWF1 cell line cells, but suppresses the response of BALB/c cell lines to antigen.

The role of cytoplasmic free calcium concentration in B-cell tolerance

Calcium is an important factor in the immune response. Extracellular calcium is required for antibody production by B lymphocytes. Several investigators have demonstrated that crosslinking of receptors on B lymphocytes by anti-mu antibody induces an increase in intracellular calcium. There are few data on the role of intracellular calcium mobilization or calcium influx in tolerance induction in B cells. We studied changes in free intracellular calcium concentration ([Ca+2]i) induced by exposure of dinitrophenyl (DNP)-specific B cells to the tolerance-inducing conjugate DNP-murine IgG2a (DNP-MGG). Splenic B cells enriched for DNP-specific cells and DNP-specific continuous B-cell lines were used for the studies. Exposure of B cells to the tolerogen DNP-MGG, the antigen DNP-keyhole limpet hemocyanin (DNP-KLH), or the antigen DNP-Ficoll induced an increase in free [Ca+2]i which was due to both mobilization of Ca+2 from endoplasmic reticulum (ER) and influx of extracellular Ca+2. This increase was DNP specific since no significant change was seen with carriers alone and no change was seen in cells that were not DNP specific. The DNP-MGG and DNP-Ficoll induced the same amount of Ca+2 release from ER but the release induced by DNP-KLH was higher. When B cells, which were made tolerant by in vitro incubation with DNP-MGG, were incubated with antigens, a mobilization of Ca+2 from endoplasmic reticulum occurred that was the same as that of nontolerant B cells. Since Ca+2 mobilization is associated with Ig receptor-dependent early B-cell activation, it is likely that the tolerant B cell can still receive an activation signal through the Ig receptors.

Membrane depolarization is induced in tolerant B lymphocytes by stimulation with antigen

Membrane depolarization is one of the earliest events in activation of cells by ligand receptor interaction. It is known that crosslinking of antigen-specific Ig receptors on B cells by antigen can induce membrane depolarization and subsequent Ia antigen expression on the cell surface. To determine whether a tolerance-inducing form of the antigen can also induce membrane depolarization after Ig receptor binding we used splenic B cells enriched for dinitrophenyl (DNP)-specific cells and determined relative membrane potential in these cells after binding of DNP-murine IgG2a (MGG) (tolerogen) or antigens (DNP-keyhole limpet hemocyanin (KLH) and DNP-Ficoll). Relative membrane potential was determined by loading the cells with the dye, 3.3-dipentyloxacarboxyanine (DiOC5(3)) after 2 hr incubation with ligand and determining relative fluorescence intensity on the fluorescence-activated cell sorter (FACS). Carriers alone did not depolarize these normal cell populations, but 100% of DNP-specific cells were depolarized by DNP-KLH and DNP-MGG while 85% were depolarized by DNP-Ficoll. To determine if tolerant B cells could be depolarized by antigen we induced tolerance in vitro or in vivo with DNP-MGG and measured the depolarization of DNP-specific B cells in response to antigens and tolerogen. DNP-specific B cells made tolerant by DNP-MGG underwent membrane depolarization when incubated with either DNP-KLH, DNP-MGG, or DNP-Ficoll but not with carriers alone. These data suggest that tolerogen induces membrane depolarization equally as well as antigen in normal cells. In addition, tolerant cells can be depolarized by Ig receptor crosslinking with either antigen or tolerogen. Thus, tolerance does not block the early membrane events induced by antigen in B cells.

Inositol phospholipid and intracellular calcium metabolism in B lymphocytes stimulated with antigen

Stimulation of B lymphocytes by anti mu antibody can activate the phosphatidylinositol pathway, but B cell activation by LPS does not involve this pathway. This study was done to determine if stimulation of B lymphocytes by their specific antigen involves this important activation pathway. We showed that levels of IP2 and IP3 increase while PIP and PIP2 decline when dinitrophenyl specific B lymphocytes are stimulated with the antigen DNP-Ficoll. Intracellular calcium concentration also increases with this stimulus. Thus, antigen stimulation of B lymphocytes is associated with activation of phosphatidylinositol pathway.

Defects in antigen-specific immune tolerance in continuous B cell lines from autoimmune mice

B cell hyperactivity and resistance to tolerance induction are well-recognized immunologic abnormalities associated with both human and murine models of systemic lupus erythematosus. Studies evaluating the role of B cells in these defects have been complicated by the difficulties of consistently isolating large numbers of B cells from T cells and other host-derived regulatory factors. We have recently developed continuous cell lines of B lymphocytes with a high degree of specificity for the antigen dinitrophenyl (DNP) from both New Zealand black times New Zealand white F1 hybrid (BWF1) and BALB/c mice, and we used them to study intrinsic B cell defects in autoimmunity. We found that the kinetics of the immune response to the antigen DNP-Ficoll of both the BWF1 and BALB/c B cell lines are not different. In addition, the BWF1 cell lines, like the BALB/c cell lines and normal B cells, require nonspecific T cell-derived factors as well as antigen to produce an immune response. Tolerance was tested in the BWF1 B cells by preincubating them with DNP-murine IgG2a (MGG), which can induce tolerance in BALB/c cell line lymphocytes. The BWF1 B cell lines were resistant to tolerance induction by DNP-MGG and required 50-fold higher dose of DNP-MGG than BALB/c cell lines for suppression. They were also relatively resistant to tolerance with trinitrophenyl-d-glutamyl lysine. Thus, DNP-specific B cells from autoimmune mice have an inherent defect in tolerance induction.

2,4-Dinitrophenyl (DNP)-specific continuous B cell lines as a model system for studying B cell activation and tolerance

Various model systems have been used to study isolated B cell response to receptor cross-linking and to lymphokines. Although each model is useful it is advantageous to have continuous cell lines of nonmalignant antigen-specific B lymphocytes to study antigen-induced B cell function. We further studied the characteristics of the 2,4-dinitrophenyl (DNP)-specific continuous B lymphocyte lines which we previously described (J. Exp. Med. 1983. 157:342). If the cell line lymphocytes are cultured with the antigen DNP-Ficoll without the presence of T cell factors or filler cells they do not produce an immune response above background, but the addition of supernatant from EL4 lymphoma and irradiated normal spleen filler cells results in a 7- to 10-fold increase in plaque-forming cells. The kinetics of the immune response is the same as that seen with normal B cells. Each cell line has a majority of cells which are small surface (s)IgM- lymphocytes which have cytoplasmic IgM and react with 14.8 antibody. There are also large sIgM+-bearing cells, which may be either in the resting or activated state. Some of the sIgM+ cells also bear IgD and Ia antigens but they do not bear IgG. From these studies we conclude that the continuously growing antigen-specific B cell lines can be a useful model to study B cell function.

The signal for tolerance in B cells is not transmitted through antigen-specific immunoglobulin receptors

Studies of B cell tolerance at the single-cell level require a ready source of antigen-specific B cells that are uncontaminated by T cells or accessory cells. We have isolated normal dinitrophenyl (DNP)-specific B cells from spleens of unprimed mice and propagated these cells in vitro. These B cells are uncontaminated by T cells or macrophages. Long-term cultures of these cell lines contain pre-B cells that are surface (s) IgM-, B cells with sIgM alone, and more mature B cells with sIgM, sIgD, and Ia antigens. Using the cell line lymphocytes we have shown that the early binding of the tolerogenic form of hapten to B cell receptor on mature B cells induces the same activation signal as antigen, and the negative signal induced by tolerogen occurs after B cell activation. Exposure of maturing B cells to DNP bound to murine IgG2a (MGG) for 30 days does not inhibit growth or receptor expression, but does induce tolerance that is reversible when DNP-MGG is removed. A 45-day exposure to DNP-MGG also induces a reversible tolerance.

Long-term growth of lines of murine dinitrophenyl-specific B lymphocytes in vitro

Studies of cellular events associated with antigen-induced triggering and differentiation of B cells would be greatly facilitated by the availability of homogeneous cell lines of antigen-specific lymphocytes that can be maintained in long-term culture. By combining the techniques of enrichment of lymphocytes for antigen-specific cells, cloning in soft agar, and long-term propagation of B cells we have been able to isolate, propagate, and maintain two lines of dinitrophenyl (DNP) -specific B lymphocytes. These cell lines are B lymphocytes that have 70% and greater than 80% DNP-specific rosette-forming cells, respectively. Both cell lines secrete small amounts of antibody spontaneously but can be stimulated by antigen in vitro in the presence of either supernatants from phytohemagglutinin-stimulated spleen cells or irradiated normal filler cells. Thus far these lines have been maintained in vitro for greater than 9 mo. They will be useful in studying factors associated with B cell response.

Arthritis associated with jejunoileal bypass: clinical and immunologic evaluation

Arthritis is a common complication of small bowel bypass, occurring in 5-20% of the postsurgical patients. Thirteen patients with arthritis related to jejunoileal bypass were studied. These patients had a symmetrical polyarthritis, and 8 also had extraarticular connective tissue disease manifestations. Immunologic evaluations were done on these patients and on a control group of 12 age- and sex-matched postintestinal bypass patients without arthritis. The incidence of positive ANA, rheumatoid factors, immune complexes, and antibodies to intestinal flora was the same in both groups. Patients in both groups showed similar changes in numbers of circulating T and B lymphocytes. More patients in the group with arthritis than in the control group had elevated IgA levels (38% versus 8%), but the difference was not significant (P less than 0.05). This study demonstrates that immunologic abnormalities occur after jejeunoileal bypass irrespective of the onset of arthritis or related symptoms. No specific immunologic abnormalities could be associated with the arthritis occurring after small bowel bypass.

Effect of anti-carrier antibody on carrier-determined tolerance

These experiments were originally designed to determine whether an anti-carrier antibody, e.g., anti-allotype could break hapten-specific tolerance in vivo. Tolerance to 2,4-dinitrophenyl (DNP) was induced in C57BL/6J mice using DNP-BALB/c IgG2a conjugate. When anti-allotype serum was injected in C57BL/6J mice one day after a single injection of DNP-IgG2a the mice were not tolerant. In contrast, when tolerance was induced by four weekly injections of tolerogen, the anti-allotype serum had no effect on the tolerant state. This effect was specific for tolerance-inducing carrier. Anti-carrier antibody injected in C57BL/6J mice one day after DNP-IgG2a produced a small but significant anti-DNP response without administration of the immunogen, whereas the tolerogen (DNP-IgG2a) by itself was not immunogenic. Similarly, despite multiple injections of DNP-IgG2a bearing the foreign allotype, only one out of 7 C57BL/6J mice showed a weak anti-carrier response. In contrast, a marked anti-carrier (IgG2a) response was obtained when the anti-allotype antibody was passively administered in C57BL/6J mice. In conclusion, these experiments suggest that tolerance to an antigenic determinant may be broken by an antibody directed not to this determinant, but to another on the same molecule. The significance of this finding in relationship to the mechanism of the carrier-determined tolerance and the breakdown of self-tolerance is discussed.

Loss of carrier-determined tolerance in vitro with loss of receptor blockade

Experiments were done to determine whether carrier-determined tolerance is reversible and whether the loss of tolerance is accompanied by the loss of receptor blockade. Spleen cells from mice made tolerant with DNP-isologous IgG remained tolerant when transferred to irradiated syngeneic mice. If these same tolerant spleen cells were incubated for 24 hr or more before transfer the tolerance was lost. Autoradiology was done on the tolerant cells with either 125I anti-DNP or 125DNP-KLH, before and after incubation in vitro. When the cells were tolerant the number of DNP ABC was decreased whereas cells having DNP on their surface were increased. When the cells lost tolerance after in vitro incubation, the hapten-bearing cells were no longer present although the number of cells free DNP receptors increased to normal. These data suggest that in carrier determined tolerance the reactivation of tolerant lymphocytes may involve reversible receptor blockade.

Carrier-determined tolerance in vitro

A primary immune response to normal BDF1 spleen cells was obtained in vitro to the T-independent antigen TNP-T4 coliphage. This anti-TNP response was suppressed by exposing the spleen cells for 6 hr to TNP bound either to isologous or heterologous gamma-globulin. The suppression was hapten specific. In contrast, TNP-albumin conjugates did not induce tolerance in vitro.