Reciprocal regulatory effects of IFN-gamma and IL-4 on the in vitro development of human Th1 and Th2 clones.

The effects exerted on the in vitro development of purified protein derivative (PPD)-specific or Dermatophagoides pteronyssinus group I (Der p I)-specific T cell lines (TCL) and T cell clones (TCC) by IL-4 or IFN-gamma addition or neutralization in human PBMC cultures were examined. PBMC from two normal individuals, which were stimulated with PPD and then cultured in IL-2 alone, developed into PPD-specific TCL and TCC able to produce IFN-gamma and IL-2 but not IL-4 and IL-5 (Th1-like). IFN-gamma or anti-IL-4 antibody addition in bulk cultures before cloning did not influence the PPD-specific TCL profile of cytokine production. In contrast, the addition of IL-4 resulted in the development of PPD-specific TCL and TCC able to produce not only IFN-gamma and IL-2 but also IL-4 and IL-5. PBMC from one atopic Der p I-sensitive patient, which were stimulated with Der p I and then cultured in IL-2 alone, developed into Der p I-specific TCL and TCC able to produce IL-5 and large amounts of IL-4 but no IFN-gamma (Th2-like). The addition in bulk cultures, before cloning, of either IFN-gamma or anti-IL-4 antibody markedly inhibited the development of Der p I-specific T cells into IL-4- and IL-5-producing TCL. Accordingly, the development into Der p I-specific Th2-like TCC was significantly reduced by the addition of IFN-gamma in bulk culture and was virtually suppressed by the presence of both IFN-gamma and anti-IL-4 antibody. These data suggest that the presence or the absence of IL-4 and IFN-gamma in bulk cultures of PBMC before cloning may have strong regulatory effects on the in vitro development of human CD4+ T cells into Th1 or Th2 clones.     

Antigen receptor-mediated anergy in resting T lymphocytes and T cell clones. Correlation with lymphokine secretion patterns.

The goal of these studies was to define the stimuli and factors that control the induction of anergy in unimmunized resting T lymphocytes. Initial experiments, aimed at establishing the system, showed that exposure of Th1 but not Th2 clones to immobilized anti-CD3 leads to a block in autocrine growth factor production and proliferation upon subsequent restimulation with Ag+APC. Anergy is not prevented by accessory cells, suggesting that this model of T cell tolerance may be due to receptor-mediated inhibitory signals, independent of costimulatory molecules. Culture of small (resting) unimmunized T lymphocytes with anti-CD3 +/- IL-2 induces unresponsiveness to restimulation with anti-CD3, but culture with anti-CD3+IL-4, which stimulates the differentiation of resting cells into IL-4 producers, does not induce anergy. Thus, IL-4-producing clones and bulk populations of IL-4-producing T cells are resistant to Ag receptor-mediated inhibitory stimuli. These results provide experimental models for studying the mechanisms of anergy in normal, unselected, mature T cells, and demonstrate fundamental similarities between cloned cell lines and unimmunized T lymphocytes in the induction of anergy.     

IL-2/anti-IL-2 antibody complex enhances vaccine-mediated antigen-specific CD8+ T cell responses and increases the ratio of effector/memory CD8+ T cells to regulatory T cells

IL-2 is well described as a cytokine with two markedly distinct functionalities: as a necessary signal during CD4(+) and CD8(+) T cell activation/expansion and as an essential cytokine for the maintenance of CD4(+)CD25(+)FoxP3(+) T cells (regulatory T (T(REG)) cells) during homeostasis. In this study we demonstrate for the first time that, compared with the use of IL-2 alone, a complex of IL-2 and anti-IL-2 Ab (IL-2 complex) enhances the effectiveness of a viral vaccine in a mouse model with known Ag specificity. IL-2 complex led to an increase in the number of Ag-specific effector/memory CD8(+) T cells, cytokine production, and CTL lysis following Ag-specific restimulation in a vaccination setting. Our results further demonstrate that this effect is temporary and declines over the course of a few days after the IL-2 complex treatment cycle. Moreover, in contrast to the use of IL-2 alone, IL-2 complex greatly increased the ratio of effector/memory CD8(+) T cells to T(REG) cells. This phenomenon can thus potentially be used in the enhancement of immune responses to vaccination.     

IL-4 production by T cells from naive donors. IL-2 is required for IL-4 production

Utilizing a sensitive and selective assay for IL-4, it was shown that lymph node T cells from naive mice could produce small amounts of this lymphokine in response to anti-CD3 antibodies adsorbed to culture dishes. The capacity of these cells to produce IL-4 in response to plate-bound anti-CD3 was substantially enhanced by the addition of IL-2 to the culture and was strikingly inhibited by monoclonal anti-IL-2 antibody. Thus, IL-2 appears to be essential for IL-4 production by anti-CD3 antibody-stimulated T cells from naive mice. The effect of IL-2 was not mediated either by preferential proliferation or survival of precursors of IL-4 producing cells, indicating that IL-2 regulates T cell production of IL-4. IL-4 producing capacity of T cells from naive mice was found mainly among CD4+ T cells. Large T cells produced much more IL-4, on a per cell basis, than did small T cells. In contrast, small T cells appeared to be equal or superior to large T cells in producing IL-2. The superiority of large T cells in IL-4-producing capacity was not accounted for by a lack of an accessory cell population from the small T cells as addition of large spleen cells depleted of both B and T cells did not enhance IL-4 production by small lymph node T cells. These results suggest that the bulk of IL-4 production by T cell populations, from normal mice, in response to anti-CD3 depends upon cells that are already activated and that IL-2 is required for such production.     

Central role for TCR/CD3 ligation in the differentiation of CD4+ T cells toward A Th1 or Th2 functional phenotype.

Activated CD4+ T cells can be classified into distinct subsets; the most divergent among them may be considered to be the IL-2 and IFN-gamma-producing Th1 clones and the IL-4 and IL-5-producing Th2 clones. Because Th1 and Th2 clones can usually be detected only after several months of culture, we used conditions that modulate the IL-2 and IL-4 production in short term culture. Here we show that freshly isolated and subsequently in vitro-activated CD4+ T cells that were cultured for 11 days with rIL-2 and restimulated showed a IFN-gamma+ IL-2+ IL-3+ IL-4- IL-5- pattern. Because these cells were not capable of providing B cell help for IgG1, IgG2a, or IgE in an APC- and TCR-dependent T-B cell assay, they expressed a phenotype typical for most Th1 clones. In contrast, activated T cells that were cultured for 11 days with IL-2 plus a mAb to CD3 and then restimulated produced a IFN-gamma- IL-2- IL-3+ IL-4+ IL-5+ pattern. These cells were capable of providing B cell help for IgG1, IgG2a, and IgE synthesis and thus presented a phenotype typical for Th2 clones. Similar results were observed when mitogenic mAb to Thy-1.2 or to framework determinants of the alpha beta TCR were used. The induction of Th1- and Th2-like cells did not depend on the relative expression of CD44 or CD45 by the T cells before activation in vitro. Because the incubation of activated T cells with anti-CD3/TCR mAb induced high unrestricted lymphokine production, the latter might be responsible for the Th2-like lymphokine pattern observed after restimulation. To address this point, TCR V beta 8+ and V beta 8- T cell blasts were co-cultured in the presence of mAb to V beta 8. After restimulation, V beta 8+ cells had a IL-4high IL-2low phenotype and V beta 8- cells had a IL-4low IL-2high phenotype. This demonstrates that TCR ligation but not lymphokines alone are capable of inducing Th2-like cells, and this points out a central role for the TCR in the generation of T cell subsets.     

Dual mechanisms of potentiation of murine antigen-specific IgE production by cyclosporin A in vitro.

Concomitant administration of cyclosporin A (CsA) with Ag has been shown to augment the production of Ag-specific IgE in vivo. We demonstrate that addition of CsA also markedly potentiated Ag-specific IgE in vitro. Low doses of CsA (3 and 10 ng/ml) added at the time of culture initiation selectively enhanced Ag-specific IgE but not IgA or IgG1 production, whereas higher doses (30 ng/ml) suppressed production of all the isotypes. Augmented IgE production was found to correlate with enhanced production of IL-4 and diminished production of IFN-gamma. Delayed addition (after 2 days) of low doses of CsA to Ag-stimulated cultures did not potentiate IgE production, even though CsA differentially affected levels of IL-4 and IFN-gamma. CsA enhanced Ag-mediated cognate T/B interaction was not affected by neutralizing doses of anti-IL-4, suggesting Ag-mediated lymphocytic "synapses" may be inaccessible to anti-IL-4. The effect of CsA on Ag presentation was determined by pulsing peritoneal exudate cells, spleen cells, or primed B cells with Ag and low doses of CsA before incubation with primed splenocytes. Enhanced Ag-specific IgE responses were detected with no effect on IL-4 or IFN-gamma levels. Thus, our study indicates that CsA potentiation of Ag-specific IgE response is due to cumulative action of CsA on two independent pathways: first, CsA differentially modulates IL-4 and IFN-gamma levels during the early phase of cognate Th2/B cell interaction; and second, CsA directly affects APC and IgE isotype-specific amplifying cellular components without apparently affecting the secretory levels of IL-4 and IFN-gamma. Dual mechanisms of CsA-potentiated IgE production are consistent with the hypothesis of two-tiered T cell regulation of Ag-specific IgE responses.     

CD8+ T cells can be primed in vitro to produce IL-4.

IL-4 production by T lymphocytes from naive mice in response to stimulation by plate-bound anti-CD3 is concentrated among CD4+ T cells. In vitro stimulation of lymph node T cells with anti-CD3 plus IL-2 and IL-4 strikingly increases the frequency of cells that produce IL-4 in response to subsequent stimulation with anti-CD3 plus IL-2. Separation of these primed cell populations into CD4+ and CD8+ T cell by cell sorting reveals that the frequency of IL-4-producing cells in both population is similar. Verification that CD8+ T cells produce IL-4 is provided by the capacity of anti-IL-4 mAb to inhibit the response of the indicator cell line to the growth factor produced by the primed cells and by detection of IL-4 by an IL-4-specific ELISA. The in vitro "priming" of CD8+ T cells to produce IL-4 is not dependent on the presence of CD4+ T cells because highly purified CD8+ T cells can be stimulated to develop into cells capable of producing IL-4 by culture with plate-bound anti-CD3 plus IL-2 and IL-4.     

An adoptive transfer model of allergic lung inflammation in mice is mediated by CD4+CD62LlowCD25+ T cells

Animal models of allergic lung inflammation have provided important insight into the cellular and biochemical factors involved in the pathogenesis of human asthma. Herein, we describe an adoptive transfer model of OVA-specific eosinophilic lung inflammation in the mouse that is used to characterize the cells involved in mediating the pulmonary inflammatory response. We report that freshly isolated spleen cells from OVA-sensitized mice are unable to prime naive recipient mice to respond to a subsequent OVA aerosol challenge. Subjecting the spleen cells to short term restimulation with Ag in vitro, however, renders the cells competent to transfer activity. The magnitude and the kinetics of the eosinophilic pulmonary inflammation in the adoptive transfer recipients are nearly identical with those generated by a more conventional active sensitization/challenge protocol, with the notable exception of differential production of plasma IgE in the two models. Extensive negative and positive selection of splenocyte subtypes indicates that the transfer of Ag-primed CD4+ T cells is both necessary and sufficient to establish full responsiveness in the recipient mice. Additional phenotypic characterization of the transfer-reactive CD4+ T cells indicates that they are found within the CD62LlowCD25+ subset and secrete high levels of IL-5 in response to Ag stimulation. Limiting dilution analysis-derived minimal frequency estimates indicate that approximately 1 in 8500 of the sensitized, cultured spleen cells produces IL-5 in response to OVA stimulation in vitro, suggesting that eosinophilic lung inflammation can be induced in naive mice by the transfer of <1200 Ag-specific CD4+ T cells.     

Requirement for noncognate interaction with T cells for the activation of B cell immunoglobulin secretion by IL-2

The mechanism whereby noncognate contact with activated IL-2-producing Type 1 helper T cells (TH1) induces B cell activation was examined. Small resting B cells from C57B1/6 mice were cultured, in the absence of any ligand for surface Ig, with irradiated cells of the hapten-specific, CBA-derived, F23.1+ TH1 clone E9.D4 in F23.1 (anti-T cell receptor V-beta 8)-coated microwells. This induced polyclonal B cell activation to enter cell cycle (thymidine incorporation) at 2 days and to secrete immunoglobulin at 5 days. An anti-IL-2 mAb (S4B6) inhibited antibody production completely. Anti-IL-2 did not inhibit either LPS-induced B cell responses, or T cell activation (measured as IL-3 secretion). Anti-IL-2 receptor (anti-Tac) mAbs also inhibited T-dependent B cell responses, without affecting LPS responses. An anti-IFN-gamma mAb partially inhibited Ig secretion, without affecting entry into cycle. LPS responses or T cell activation. Other antibodies (anti-IL-3, IL-4, IL-5, Thy-1.2, CD5) were not inhibitory. After 2 days of culture with F23.1-activated T cells, B cells appeared to have become responsive to IL-2, in that they could be driven to immunoglobulin production by the addition of IL-2. Flow cytometry showed no expression by these B cells of 55-kDa (Tac) IL-2 receptors. Also, rigorous removal of T cells from 2-day cocultures prevented the response to IL-2, and readdition of T cells restored it. Because the reconstituted responses were inhibited both by anti-IL-2 and by anti-Tac, IL-2 must have acted indirectly, via the T cells that were present in these cultures. Continued contact with T cells was therefore necessary for the progression of B cells to antibody secretion.     

IL-4/B cell stimulatory factor 1 stimulates T cell growth by an IL-2-independent mechanism

Resting T cells are stimulated to synthesize DNA by IL-4 and phorbol myristate acetate (PMA). This response of T cells to IL-4 plus PMA is independent of the action of IL-2 as judged by 1) the lack of IL-2 in supernatants of stimulated cells, 2) the failure to detect IL-2 mRNA in stimulated cells by in situ hybridization, 3) the inability of anti-IL-2R antibody and of anti-IL-2 antibody to block responses to IL-4 plus PMA, and 4) the failure of cyclosporin A to block responses. T cells also respond to anti-CD3 antibodies and IL-4 in the presence of anti-IL-2R antibodies. IL-4 stimulation of growth of the long term T cell line HT-2 also appears to be independent of the action of IL-2. No IL-2 mRNA is found in IL-4-stimulated HT-2 cells by Northern blotting; the response of HT-2 cells to IL-4 is not blocked by anti-IL-2R antibodies; the response of HT-2 cells to IL-4 is not inhibited by cyclosporin A. Although IL-4 stimulation of T cells is independent of IL-2, IL-4 plus PMA treatment of resting T cells does cause enhanced expression of IL-2R and prepares cells to proliferate to IL-2 alone. In both these properties IL-4 resembles IL-2. These experiments lead us to conclude that IL-4 can act as an alternative to IL-2 as authentic T cell growth factor.     

The role of B cells in the development of CD4 effector T cells during a polarized Th2 immune response

Previous studies have suggested that B cells promote Th2 cell development by inhibiting Th1 cell differentiation. To examine whether B cells are directly required for the development of IL-4-producing T cells in the lymph node during a highly polarized Th2 response, B cell-deficient and wild-type mice were inoculated with the nematode parasite, Nippostrongylus brasiliensis. On day 7, in the absence of increased IFN-gamma, IL-4 protein and gene expression from CD4 T cells in the draining lymph nodes were markedly reduced in B cell-deficient mice and could not be restored by multiple immunizations. Using a DO11.10 T cell adoptive transfer system, OVA-specific T cell IL-4 production and cell cycle progression, but not cell surface expression of early activation markers, were impaired in B cell-deficient recipient mice following immunization with N. brasiliensis plus OVA. Laser capture microdissection and immunofluorescent staining showed that pronounced IL-4 mRNA and protein secretion by donor DO11.10 T cells first occurred in the T cell:B cell zone of the lymph node shortly after inoculation of IL-4-/- recipients, suggesting that this microenvironment is critical for initial Th2 cell development. Reconstitution of B cell-deficient mice with wild-type naive B cells, or IL-4-/- B cells, substantially restored Ag-specific T cell IL-4 production. However, reconstitution with B7-1/B7-2-deficient B cells failed to rescue the IL-4-producing DO11.10 T cells. These results suggest that B cells, expressing B7 costimulatory molecules, are required in the absence of an underlying IFN-gamma-mediated response for the development of a polarized primary Ag-specific Th2 response in vivo.     

Conventional, naive CD4+ T cells provide an initial source of IL-4 during Th2 differentiation

IL-4 is known to promote the differentiation of CD4+ T cells into IL-4-secreting Th2 cells. However, the cellular source of the early burst of IL-4 that drives Th2 responses in vivo has not been conclusively identified. Mice deficient for the IL-4 receptor alpha-chain (IL-4Ralpha-/-) retain the capacity to secrete IL-4 and can be used to identify those cell types that produce IL-4 without a requirement for prior IL-4-mediated stimulation. To address whether naive, conventional CD4+ T cells may act as initial producers of IL-4 in Ag-specific responses, we crossed the BALB/c IL-4Ralpha-/-mice to DO11.10/scid TCR transgenic mice. Lymph node cells from wild-type and IL-4Ralpha-/- DO11.10/scid mice secreted approximately 50 pg of IL-4 per10(6) cells within 48 h after peptide stimulation. This small amount of IL-4 was sufficient to cause the differentiation of wild-type CD4+ T cells into Th2 cells, particularly if IFN-gamma and IL-12 were neutralized during the priming cultures. CD4+ cells from the IL-4Ralpha-/- mice gave rise to a minor proportion (approximately 2%) of IL-4-producing cells upon stimulation in the presence of anti-IFN-gamma and anti-IL-12. These data show that conventional, naive CD4+ T cells may be considered as initial sources of IL-4 and, in the absence of IFN-gamma and IL-12, this IL-4 can induce Th2 polarization.     

A common precursor for CD4+ T cells producing IL-2 or IL-4.

To investigate whether CD4+ T cells are predetermined to produce a given pattern of lymphokines, we have used a culture system that allows the controlled induction of either IL-2- or IL-4-producing CD4+ T cells. Single, freshly isolated murine CD4+ T cells were activated with Con A, rIL-2, and APC; the developing clones were split and then cultured for an additional 14 days with either rIL-2 alone or with rIL-2 and anti-CD3 stimulation. Subclones expanded in the presence of rIL-2 alone produced predominantly IL-2, although subclones derived from the same precursor and expanded in the presence of rIL-2 and a mitogenic antibody to CD3 released predominantly IL-4. Subclones expanded for 2 wk in the presence of rIL-2 plus a mitogenic mAb to CD3 released up to 60 times more IL-4 but only 1/90 the amount of IL-2 released by subclones derived from the same precursor cell and expanded with rIL-2. Both phenotypes can be derived from IL-2-producing precursor cells. These results demonstrate that IL-2-producing clones can be derived from the same cells as IL-4-producing clones and are most consistent with the view that the IL-2-producing Th1 or the IL-4-producing Th2 phenotype of a T cell clone is acquired during T cell differentiation and is not secondary to the expansion of distinct subpopulations that are predetermined to produce a specific cytokine pattern.     

Antigen-induced suppression of the proliferative response of T cell clones

Proliferation of Ag-specific T cell clones can be inhibited by the addition of high concentrations of Ag at the beginning of culture. Under these conditions the cells produce lymphokines and express high affinity IL-2R but fail to divide, even after the addition of exogenous IL-2. This state results from restimulation of the cells with Ag-Ia molecule complexes approximately 20 h after initiation of culture. It can be prevented by addition of either an anti-Ia or an anti-L3T4 mAb at that time, but not by cyclosporin A, and mimicked in cultures containing low concentrations of antigen by addition at that time of high Ag concentrations or normal stimulatory concentrations of Con A. These observations suggest that restimulation of activated T cell clones 20 h after their initial stimulation prevents them from dividing.     

Long-term CD4+ memory T cells from the spleen lack MEL-14, the lymph node homing receptor.

We have characterized the surface phenotype and function of long-lived, Ag-specific memory CD4+ T cells generated in vivo by immunization with keyhole limpet hemocyanin (KLH). CD4+ T cells from the spleens of mice primed more than 2 mo previously with KLH, produced high levels of IL-2 and IL-3, and low levels of IL-4 and IFN-gamma in response to in vitro restimulation with specific Ag. The KLH-primed T cells mediated carrier-specific helper activity for the antibody production by NIP-primed B cells in secondary in vitro responses to NIP-KLH. Subsets of CD4+ T cells from KLH-primed mice were isolated on the basis of surface CD45RB (23G2) by magnetic separation and were examined for functional capacity in several assays of Ag-specific recall. Virtually all of the secretion of IL-2, IL-3, IL-4, and IFN-gamma in response to restimulation with Ag in vitro was associated with, and considerably enriched in, the CD45RB- subset of CD4+ T cells. Similarly, carrier-specific helper function and Ag-specific proliferation in vitro were also confined to the CD45RB-, CD4+ subset of T cells, confirming the previous association of this surface phenotype with memory Th cell activity. We also examined expression of the lymphocyte homing receptor, MEL-14 (gp90MEL), which is required for lymphocyte extravasation to peripheral lymph nodes and is present in high levels on naive T cells. MEL-14 positive and negative subsets of CD4+ T cells from long term KLH-primed mice were evaluated for Ag-specific memory function in terms of lymphokine production, Ag-induced proliferation, and helper activity. Each of these functions was associated exclusively with the MEL-14- subset of CD4+ T cells, which exhibited responses comparable to the CD45RB- subset. These data indicate that memory Th cell function in the spleen is contained within the MEL-14-, CD45RB- subset of CD4+ T cells and suggest that memory helper cells may have different patterns of recirculation from naive T cells.     

Regulatory roles of IL-2 and IL-4 in H4/inducible costimulator expression on activated CD4+ T cells during Th cell development

We found a tight correlation among the levels of H4/inducible costimulator (ICOS) expression, IL-4 production, and GATA-3 induction, using activated CD4(+) T cells obtained from six different murine strains. BALB/c-activated CD4(+) T cells expressed approximately 10-fold more H4/ICOS on their surfaces and produced approximately 10-fold more IL-4 upon restimulation than C57BL/6-activated CD4(+) T cells. BALB/c naive CD4(+) T cells were shown to produce much higher amounts of IL-2 and IL-4 upon primary stimulation than C57BL/6 naive CD4(+) T cells. Neutralization of IL-4 with mAbs in culture of BALB/c naive CD4(+) T cells strongly down-regulated both H4/ICOS expression on activated CD4(+) T cells and IL-4 production upon subsequent restimulation. Conversely, exogenous IL-4 added to the culture of BALB/c or C57BL/6 naive CD4(+) T cells up-regulated H4/ICOS expression and IL-4 production upon restimulation. In addition, retroviral expression of GATA-3 during the stimulation of naive CD4(+) T cells from C57BL/6 or IL-4(-/-) mice increased H4/ICOS expression on activated CD4(+) T cells. A similar effect of IL-2 in the primary culture of BALB/c naive CD4(+) T cells appeared to be mediated by IL-4, the production of which was regulated by IL-2. These data suggest that IL-4 induced by IL-2 is critical to the maintenance of high H4/ICOS expression on BALB/c-activated CD4(+) T cells.     

in vitro generation of IFN-γ-producing Listeria-specific T cells is dependent on IFN-γ production by non-NK cells

In vitro 5-day cultures of naive spleen cells with viable Listeria monocytogenes (VLM), but not heat-killed L. monocytogenes, induced CD4⁺ T cells that produced IFN-γ upon secondary antigen stimulation. The VLM-induced Listeria-specific T cells produced IFN-γ but lacked expression of IL-2 and IL-4. To study the role of IFN-γ in the induction of the IFN-γ-producing T cells, we added anti-IFN-γ mAb to the primary culture and analyzed IFN-γ production upon secondary antigen stimulation. Addition of anti-IFN-γ mAb to the culture suppressed generation of IFN-γ-producing CD4⁺ T cells, suggesting that IFN-γ is important in the induction of IFN-γ-producing CD4⁺ T cells. Furthermore, our results showed that depletion of NK cells from spleen cells by anti-asialo GM1 antibody plus complement before culture enhanced induction of IFN-γ-producing CD4⁺ T cells. Although NK cells are known to produce IFN-γ, the results indicate that NK cell-derived IFN-γ may not be important in induction of the Listeria-specific IFN-γ-producing CD4⁺ T cells in the culture system. In addition, we demonstrated that IFN-γ expression was high in CD4⁺ T cells from cultures of spleen cells with VLM at the primary culture level. These results suggest that IFN-γ derived from T cells may enhance production of IFN-γ by CD4⁺ T cells, while NK cells rather suppress the induction of IFN-γ-producing CD4⁺ T cells.     

Lectin-mediated induction of IL-4-producing CD4+ T cells.

Cultured murine CD4+ T cells have been shown to differentiate into IL-2 or IL-4-producing subsets. The factors responsible for the development of CD4+ T cells which produce IL-2 but not IL-4 and cells capable of producing IL-4 but not IL-2 are unknown. Here we describe a system that allows the controlled induction of IL-2- or IL-4-producing T cells after one single round of activation. Freshly isolated CD8-depleted T cells were activated with various polyclonal T cell activators for 48 h, washed, and then expanded under different conditions. IL-2 and IL-4 production were induced by restimulation of T cells and were measured with CTLL cells that respond to both cytokines and mAb to IL-2 and IL-4. T cells produced mainly IL-2 and small amounts of IL-4 when restimulated after expansion culture for 12 days with rIL-2 alone. However, after expansion for 12 days in the presence of rIL-2 plus Con A, we observed a 30- to 100-fold up-regulation of IL-4 activity and a 100-fold down-regulation of IL-2 when assessed by responses of CTLL cells incubated with the supernatant of restimulated T cells and by responses of CTLL cells cocultured with restimulated cells. An increase of IL-4 and decrease of IL-2 was also observed when the results were based on the cell numbers at the beginning of the expansion culture. The induction of IL-4 and the down-regulation of IL-2 1) were not reproduced with alpha-methyl-mannoside-treated supernatant of Con A-stimulated spleen cells, 2) were not dependent on the presence of large numbers of APC, 3) did not result from differential consumption of lymphokines after restimulation, 4) were not due to a difference in the time course of IL-2 or IL-4 release in either T cell population, and 5) were obtained regardless of the agents used to activate or to restimulate the T cells. Because Con A remained detectable on the T cell surface and because expansion of activated T cells with IL-2 plus Con A for several days was necessary, our results indicate that mainly IL-4-producing CD4+ T cells can be induced by prolonged engagement of T cell surface molecules.