LFA-1-mediated costimulation of CD8+ T cell proliferation requires phosphatidylinositol 3-kinase activity

LFA-1 binding to ICAM-I provides a costimulatory signal for CD8(+) T cell activation that results in increased IL-2 mRNA levels and protein production to support proliferation. CD28 binding to its B7 ligands has the same effect, and the two costimulatory receptors activate some of the same intracellular signaling events, including up-regulation of phosphatidylinositol (PI) 3-kinase activity. However, costimulation by LFA-1 depends upon the activity of this enzyme, whereas costimulation by CD28 does not, as evidenced by differential effects of specific inhibitors of PI 3-kinase. When cells are costimulated with ICAM-1 in the presence of the inhibitors wortmannin or LY294002, proliferation is blocked, but increases in IL-2 mRNA levels and protein production are not. Costimulation also results in increased surface expression of CD25, which is essential for formation of an active IL-2R. This is blocked by the PI 3-kinase inhibitors when costimulation is via LFA-1 but not when it is via CD28. Finally, IL-2-driven proliferation is not blocked by the inhibitors once CD25 surface expression has increased. Thus, the PI 3-kinase-dependent step in CD8 T cell costimulation by LFA-1 is up-regulation of IL-2R expression. In contrast, CD28 engagement also increases IL-2R surface expression, but the up-regulation does not depend upon PI 3-kinase activity.     

Production of lymphokine mRNA by CD45R+ and CD45R- helper T cells from human peripheral blood and by human CD4+ T cell clones

The expression of lymphokine mRNA by human CD4+CD45R+ and CD4+CD45R- Th cells was assessed after mitogen stimulation. These Ag have previously been shown to relate closely to virgin and primed T cells, respectively. CD4+CD45R+ (virgin) and CD4+CD45R- (primed) cell fractions were isolated by sorting double-labeled cells with a fluorescence-activated cell sorter. CD4+CD45R+ cells produced high levels of IL-2 mRNA when stimulated with either PMA together with calcium ionophore, or with PHA, but they expressed only trace quantities of mRNA for IL-4 or IFN-gamma. In contrast, CD4+CD45R- cells produced high levels of mRNA for IL-2, IL-4, and IFN-gamma. After 14 days of continuous culture, CD4+CD45R+ Th cells lost expression of the CD45R Ag, but gained high level expression of CDw29, such that they were indistinguishable from the cell population which originally expressed this Ag. At the same time, they acquired the ability to synthesize IL-4 mRNA. It seemed likely that the broad lymphokine profile of primed Th cells might mask clonal heterogeneity. Analysis of 122 CD4+ T cell clones showed that all of them synthesized IL-2 mRNA. One clone failed to express IL-4 mRNA, but did produce those for IL-2 and IFN-gamma. A total of 34 of the clones was investigated to determine expression of IFN-gamma mRNA; two of these clones were negative for IFN-gamma mRNA, and both expressed IL-2 and IL-4 message. These data suggest that while fresh virgin and primed peripheral blood T cells show a clear resolution of lymphokine production, a simple subdivision of human CD4+ T cell clones on the basis of their lymphokine production (such as that reported for mouse Th cell clones) is not possible.     

Altered IFN-gamma and IL-4 pattern lymphokine secretion in mice partially depleted of CD4 T cells by anti-CD4 monoclonal antibody.

Complete elimination of CD4 cells by in vivo treatment with anti-CD4 mAb may result in B cell polyclonal activation. Additionally, mice treated with doses of anti-CD4 that eliminate half the CD4 cells produced higher anti-SRBC antibody responses than controls. This suggests that partial CD4 depletion enhances Th2-like function. To test this hypothesis we examined Th1 and Th2 lymphokine potential in mice partially depleted of CD4 cells. We measured IL-4 and IFN-gamma secretion by stimulated unfractionated spleen cells and analyzed activated, purified CD4 cells by RNA in situ hybridization to determine the percentage of IFN-gamma- or IL-4-producing cells. Unfractionated splenocytes from partially CD4-depleted mice secreted more IL-4 and less IFN-gamma than splenocytes from control mice. In situ hybridization proved that CD4 cells from partially depleted mice contained a higher percentage of IL-4 and a lower percentage of IFN-gamma-producing cells than controls. These results indicate that treatment with a dose of mAb resulting in partial CD4 depletion may permit increased Th2-like lymphokine expression. This study also provides evidence that cells committed to Th2-like function exist in vivo in mice.     

Differential costimulatory effects of adhesion molecules B7, ICAM-1, LFA-3, and VCAM-1 on resting and antigen-primed CD4+ T lymphocytes.

Optimal proliferation of T cells although initiated via ligation of the CD3/TCR complex requires additional stimulation resulting from adhesive interactions between costimulatory receptors (R) on T cells and their counter-R on APC. At least four distinct adhesion molecules (counter-R) present on APC, B7, ICAM-1 (CD54), LFA-3 (CD58), and VCAM-1 have been individually shown to costimulate T cell activation. Because some of these molecules may be expressed simultaneously on APC, it has been difficult to examine relative contributions of individual counter-R during the induction of T cell proliferation. We have produced soluble IgC gamma 1 fusion chimeras (receptor globulins or Rg) of B7, ICAM-1, LFA-3, and VCAM-1 and compared their relative abilities to costimulate proliferation of resting or Ag-primed CD4+ T cells. When co-immobilized with mAb directed at TCR alpha beta or CD3 but not CD2 or CD28, each Rg induced proliferation of both resting and Ag-primed CD4+ cells. In contrast, similarly co-immobilized CD7 Rg or ELAM-1 Rg were ineffective. Resting CD4+ T cells produced more IL-2, expressed significantly higher levels of IL-2R alpha, and proliferated more efficiently when costimulated with either ICAM-1 Rg or VCAM-1 Rg than with B7 Rg or LFA-3 Rg. CD4+ CD45RO+ memory T cells proliferated more vigorously in response to the costimulation by each of the four Rg than CD4+ CD45RA+ naive T cells. In contrast with the behavior of resting CD4+ T cells, proliferation of Ag-preactivated CD4+ T cells was most efficient when costimulated by B7 Rg. The costimulatory effect of LFA-3 Rg on Ag-primed CD4+ T cells was weaker than that of B7 Rg but was significantly greater than that of either ICAM-1 Rg or VCAM-1 Rg. These results suggest that resting and Ag-primed CD4+ T cells preferentially respond by proliferation to different costimulatory counter-R. ICAM-1 and VCAM-1 may be involved in the initiation of proliferation of Ag-responsive T cells, and B7 and LFA-3 may facilitate sustained proliferation of Ag-primed T cells. The cumulative costimulation by the above counter-R may facilitate optimal expression of various regulatory and effector functions of T cells.     

Abnormalities of lymphokine gene expression in patients with common variable immunodeficiency

Common variable immunodeficiency (CVI) is a syndrome characterized by hypogammaglobulinemia, recurrent bacterial infections, and increased occurrence of both autoimmune disease and malignancy. In our study we examine the expression of lymphokine genes in mitogen-activated T cells from four patients with CVI. T cells from patients with CVI did not differ significantly from normals in total T cell number, CD4/CD8 ratio, CD45R expression, or proliferation in response to PHA. However, T cells from this group of patients did exhibit significant abnormalities of mitogen-induced lymphokine gene expression. T cells from patients exhibited significantly decreased expression of IL-2, IL-4, IL-5, and IFN-gamma when compared to normal controls. In contrast to these abnormal findings, mitogen-activated T cells from patients with CVI expressed normal amounts of IL-2R alpha and c-myc suggesting that these patients have a selective abnormality of T cell activation. Furthermore, it is likely that the deficient production of IFN-gamma by patient T cells is partially due to the abnormality of IL-2 production as the levels of IFN-gamma mRNA detected during the initial IL-2-independent phase of T cell activation were normal and the addition of exogenous rIL-2 was able to normalize IFN-gamma production by PHA-stimulated patient cells. Finally, supernatants from PHA-activated cultures of patients PBMC were deficient in their ability to support Ig secretion by Staphylococcus A Cowan's-activated normal B cells suggesting that these T cell abnormalities may contribute to the pathogenesis of this syndrome.     

Stimulation of IFN-gamma, TNF-alpha, and TNF-beta secretion in IL-2-activated T cells: costimulatory roles for LFA-1, LFA-2, CD44, and CD45 molecules.

Lymphokine-activated killer (LAK) cells are peripheral blood lymphocytes (PBLs) that possess the ability to kill target cells in a non-major histocompatibility complex (MHC)-restricted manner. Both NK and T cells can be stimulated with interleukin-2 (IL-2) to become LAK cells. We previously reported that the interaction of LAK cells with tumor cells also induces the secretion of interferon-gamma (IFN-gamma). The NK subset of LAK (LAK-NK) cells is stimulated by tumor cells to secrete IFN-gamma in a non-MHC-restricted manner while the T cell subset of LAK (LAK-T) cells is stimulated to secrete IFN-gamma upon cross-linking of the T cell receptor (TCR)-CD3 complex. We here report that LAK-T cells stimulated with anti-CD3 mAbs and tumor cells secrete two additional cytokines, tumor necrosis factor-alpha (TNF-alpha) and TNF-beta/lymphotoxin (TNF-beta). In addition, we demonstrate that at least four other structurally unrelated molecules, in addition to the TCR-CD3 complex, on LAK-T cells participate in the stimulation of IFN-gamma, TNF-alpha, and TNF-beta production. These molecules are the lymphocyte function associated antigen-1 (LFA-1), lymphocyte function associated antigen-2 (LFA-2), CD44, and CD45. LFA-1 is an integrin, LFA-2 is a member of the immunoglobulin supergene family, CD44 is homologous to the cartilage link proteins, and CD45 is a tyrosine phosphatase. Ligands to three of these molecules have been identified; ICAM-1, LFA-3, and hyaluronic acid binding to LFA-1, LFA-2, and CD44, respectively. LFA-1, LFA-2, and CD44 are reported to function both as adhesion molecules and as costimulators in resting T cells. Our data suggest that these three molecules enhance IFN-gamma, TNF-alpha, and TNF-beta production by augmenting LAK-T cell to tumor cell adhesion and also by functioning as costimulators.     

Distinct activation signals determine whether IL-21 induces B cell costimulation, growth arrest, or Bim-dependent apoptosis

IL-21 costimulates B cell proliferation and cooperatively with IL-4 promotes T cell-dependent Ab responses. Somewhat paradoxically, IL-21 also induces apoptosis of B cells. The present study was undertaken to more precisely define the expression of the IL-21R, using a novel mAb, and the circumstances by which IL-21 promotes B cell growth vs death. The IL-21R was first detected during T and B cell development, such that this receptor is expressed by all mature lymphocytes. The IL-21R was further up-regulated after B and T activation, with the highest expression by activated B cells. Functional studies demonstrated that IL-21 substantially inhibited proliferation and induced Bim-dependent apoptosis for LPS or CpG DNA-activated B cells. In contrast, IL-21 induced both costimulation and apoptosis for anti-CD40-stimulated B cells, whereas IL-21 primarily costimulated B cells activated by anti-IgM or anti-IgM plus anti-CD40. Upon blocking apoptosis using C57BL/6 Bim-deficient or Bcl-2 transgenic B cells, IL-21 readily costimulated responses to anti-CD40 while proliferation to LPS was still inhibited. Engagement of CD40 or the BCR plus CD40 prevented the inhibitory effect by IL-21 for LPS-activated B cells. Collectively, these data indicate that there are three separable outcomes for IL-21-stimulated B cells: apoptosis, growth arrest, or costimulation. We favor a model in which IL-21 promotes B cell maturation during a productive T cell-dependent B cell response, while favoring growth arrest and apoptosis for nonspecifically or inappropriately activated B cells.     

T cell activation induced by anti-CD3 antibodies requires prolonged stimulation of protein kinase C

Accessory cell-depleted T cells required the presence of a protein kinase C (PKC) stimulating phorbol ester, such as phorbol 12,13-dibutyrate (PDB), to be activated by soluble antibodies to the CD3 molecular complex. To determine the duration of PDB costimulation necessary to induce a proliferative response, highly purified T cells were pulsed with anti-CD3, incubated with PDB for limited periods of time, and then washed and recultured in the absence of PDB. T cells stimulated with anti-CD3 and PDB for 2 hr were unable to proliferate unless IL-2 or PDB was added to the second culture. With more prolonged exposure to PDB (4-18 hr), anti-CD3-pulsed cells exhibited an increased capacity to proliferate in the absence of additional PDB. Proliferation could be augmented by exogenous IL-2, but remained submaximal. Optimal DNA synthetic responses required the presence of PDB throughout the entire culture. Despite this, costimulation with anti-CD3 and PDB induced a significant number of cells to express IL-2 receptors and enter the cell cycle after 18 hr of costimulation with PDB. Moreover, T cells costimulated by anti-CD3 and PDB produced IL-2 within 4 hr. However, T cells that were stimulated with anti-CD3 and PDB for 4 hr, washed, and recultured rapidly lost the ability to continue to produce IL-2, which reflected a decrease in the content of mRNA encoding IL-2. This loss of IL-2 production was prevented by reculturing the cells with PDB. These studies therefore indicate that after initial T cell activation by anti-CD3, continued stimulation of PKC is necessary for ongoing IL-2 production. These results suggest a model of T cell activation in which sustained stimulation of PKC after cell cycle entry is required to maintain growth factor production and continued proliferation.     

CD28 costimulation induces delta opioid receptor expression during anti-CD3 activation of T cells

Previous studies have demonstrated that naive splenic mouse T cells express no or only very low levels of the delta-type opioid receptor (delta OR), but stimulation of mouse splenocytes with Con A results in induction of delta OR mRNA and protein. In this report we have shown that stimulation of highly purified populations of naive mouse T cells with anti-CD3 mAb alone results in T cell activation, as evidenced by sustained IL-2 secretion and cell proliferation, but fails to elicit delta OR expression. However, delta OR expression is induced by costimulation of these very pure T cells with anti-CD3 and anti-CD28 mAbs. The delta OR induction by anti-CD3 and anti-CD28 costimulation was completely blocked by inhibition of phosphatidylinositol 3-kinase with wortmannin. Because phosphatidylinositol 3-kinase activation in T cells is linked to costimulation, these results suggest that induction of delta OR expression during T cell activation is strictly dependent on costimulation. It also appears that costimulatory receptors other than CD28 can provide the signaling required for delta OR expression because delta OR mRNA was induced by Con A stimulation of splenocytes from CD28-deficient mice.     

EphA receptors inhibit anti-CD3-induced apoptosis in thymocytes

The EphA receptor tyrosine kinases interact with membrane-bound ligands of the ephrin-A subfamily. Interaction induces EphA receptor oligomerization, tyrosine phosphorylation, and, as a result, EphA receptor signaling. EphA receptors have been shown to regulate cell survival, migration, and cell-cell and cell-matrix interactions. However, their functions in lymphoid cells are only beginning to be described. We show in this study that functional EphA receptors are expressed by murine thymocytes, including CD4(+)CD8(+), CD4(+)CD8(-), and CD4(-)CD8(+) subpopulations. We demonstrate that activation of EphA receptors by the ephrin-A1 ligand inhibits the anti-CD3-induced apoptosis of CD4(+)CD8(+) double-positive thymocytes. Furthermore, ephrin-A1 costimulation suppresses up-regulation of both the IL-2R alpha-chain (CD25) and early activation Ag CD69 and can block IL-2 production by CD4(+) single-positive cells. In agreement, EphA receptor activation in thymocytes also inhibits TCR-induced activation of the Ras-MAPK pathway. Our findings suggest that EphA receptor activation is antithetical to TCR signaling in thymocytes, and that the level of engagement by ephrin-A proteins on thymic APCs regulates thymocyte selection.     

Cutting edge: stromal cell-derived factor-1 is a costimulator for CD4+ T cell activation

Stromal cell-derived factor (SDF)-1 is a chemoattractant for T cells, precursor B cells, monocytes, and neutrophils. SDF-1alpha was also found to up-regulate expression of early activation markers (CD69, CD25, and CD154) by anti-CD3-activated CD4+ T cells. In addition, SDF-1alpha costimulated proliferation of CD4+ T cells and production of IL-2, IFN-gamma, IL-4, and IL-10. Stimulation with SDF-1alpha alone did not induce activation marker expression, proliferation, or cytokine production by the CD4+ T cells. SDF-1alpha-mediated costimulation was blocked by anti-CXC chemokine receptor-4 mAb. RANTES also increased activation marker expression by anti-CD3-stimulated peripheral CD4+ T cells, but less effectively than SDF-1alpha did, and did not up-regulate IL-2 production and proliferation. These results indicate that SDF-1 and CXC chemokine receptor-4 interactions not only play a role in T cell migration but also provide potent costimulatory signals to Ag-stimulated T cells.     

Distinct regulation of lymphokine production is found in fresh versus in vitro primed murine helper T cells

The kinetics of lymphokine RNA induction and secretion of biologically active lymphokine from CD4-enriched splenic T cell populations was investigated. Cells stimulated immediately after isolation from murine spleen ("fresh" T cells) and cells restimulated after 4 days of in vitro culture ("primed" T cells) were compared. Northern blot analysis and bioassays were used to analyze and quantitate production of eight lymphokines and the IL-2R. Fresh T cells produced high levels of IL-2 and low to moderate levels of IL-3, granulocyte/macrophage-CSF, and IFN-gamma. In vitro primed T cells produced IL-2, IL-3, IL-4, IL-5, IL-6, granulocyte/macrophage-CSF, IFN-gamma, and high levels of IL-2R RNA. Comparison of RNA levels and bioassays of supernatants from these populations indicated that primed T cells produced at least 10-fold more of six of the lymphokines than fresh T cells. Only IL-2 was produced in near equal amounts by fresh and primed T cells. There were also marked differences in the kinetics of lymphokine production by fresh and primed CD4+ T cells. After restimulation with Con A and PMA, primed cells produced a short burst of lymphokine RNA that peaked between 7.5 and 13 h and declined after 18 h. Fresh T cells lagged in the initial production of lymphokine RNA, with levels peaking 18 to 44 h after mitogenic stimulation. Depletion of CD4+ cells indicated that cells of helper phenotype were responsible for the majority of lymphokine production from the primed cells. Thus different subpopulations of Th cells defined by their respective ability to respond either directly (fresh T cells) or only after culture and restimulation (primed T cells) show different patterns of lymphokine gene regulation. Other studies suggest that the activity of "fresh" Th cells is due to a population with a "memory" phenotype, while the cells which require culture have a "precursor" phenotype. These distinct patterns of lymphokine gene regulation in the two populations of Th cells may account in part for differences seen in the kinetics and magnitude of the naive and memory immune responses which are regulated by Th cells.     

The role of CD4 in antigen-independent activation of isolated single T lymphocytes

The membrane molecule CD4 (L3T4) is thought to facilitate activation of Class II H-2-restricted T cells by binding to Ia determinants on antigen-presenting cells. Recent reports suggest that CD4 can also contribute to antigen-independent activation by anti-T cell receptor (TCR) antibodies. An assay which measures the secretion of two lymphokines, granulocyte-macrophage colony-stimulating factor and interleukin 3 (IL-3), by single T cells activated with an anti-TCR antibody, F23.1, was used to analyze the effects of anti-CD4 antibodies on antigen-independent T cell activation. Single cells of a CD4+F23.1+ clone were micromanipulated into wells to which F23.1 had been immobilized, and their lymphokine secretion was measured 24 hr later. The frequency of lymphokine-secreting cells was consistently reduced up to 10-fold in the presence of soluble anti-CD4 antibody (GK1.5) but only up to 2.5-fold by an antibody to the cell adhesion molecule, LFA-1. In both bulk and single-cell cultures, responses to suboptimal concentrations of F23.1 were more susceptible to inhibition by GK1.5 than responses to optimal F23.1. The failure of GK1.5 to inhibit IL-2-stimulated lymphokine synthesis in bulk cultures suggested that CD4 ligation did not deliver a negative signal to the clone. By contrast, when either anti-CD4 or anti-LFA-1 was immobilized on the same surface as F23.1, the frequency of lymphokine-secreting cells could be increased up to 10-fold. It is concluded that anti-CD4 antibodies can act directly on the responding T cell to affect TCR-dependent activation, in the absence of interaction with antigen-presenting cells or any other cell type.     

Tumor-specific T cell activation by recombinant immunoreceptors: CD3 zeta signaling and CD28 costimulation are simultaneously required for efficient IL-2 secretion and can be integrated into one combined CD28/CD3 zeta signaling receptor molecule.

Recombinant immunoreceptors with specificity for the carcinoembryonic Ag (CEA) can redirect grafted T cells to a MHC/Ag-independent antitumor response. To analyze receptor-mediated cellular activation in the context of CD28 costimulation, we generated: 1) CEA+ colorectal tumor cells that express simultaneously B7-1 and B7-2, and 2) CEA-specific immunoreceptors that harbor intracellularly the signaling moities either of CD28 (BW431/26-scFv-Fc-CD28), CD3zeta (BW431/26-scFv-Fc-CD3zeta), or FcepsilonRIgamma (BW431/26-scFv-Fc-gamma). By retroviral gene transfer, we grafted activated T cells from the peripheral blood with these immunoreceptors. T cells that express the FcepsilonRIgamma or CD3zeta signaling receptor lysed specifically CEA+ tumor cells and secreted high amounts of IFN-gamma upon receptor cross-linking, whereas anti-CEA-CD28 receptor-grafted T cells did not, indicating that CD28 signaling alone is not sufficient for efficient T cell activation. CD28 costimulation did not affect cytolysis by T cells equipped with gamma- or zeta-signaling receptors, but enhanced both IFN-gamma secretion and proliferation. CD28 costimulation, however, was required for efficient IL-2 secretion of anti-CEA-gamma receptor-grafted T cells. Both purified CD4+ and CD8+ T cells grafted with immunoreceptors required CD28 costimulation for complete T cell activation. We integrated both CD28 and CD3zeta signaling domains into one combined immunoreceptor molecule (BW431/26-scFv-Fc-CD28/CD3zeta) with dual signaling properties. T cells grafted with the combined CD28/CD3zeta signaling receptor secreted high amounts of IL-2 upon Ag binding without exogenous B7/CD28 costimulation, demonstrating that both MHC-independent cellular activation and CD28 costimulation for complete T cell activation can be delivered by one recombinant receptor molecule.     

Characterization of accessory cell costimulation of Th1 cytokine synthesis.

We studied the capacity of macrophage and B cell lines to provide a costimulatory signal that enhances synthesis of IFN-gamma and IL-2 by mouse Th1 clones stimulated with suboptimal doses of immobilized anti-CD3 antibody. The J774 macrophage line and the CH27 B lymphoma line had the greatest costimulatory activity and routinely increased IL-2 production by 10-fold to 100-fold. Other macrophage and B cell lines had less activity and T cell lines were unable to costimulate. The J774 and CH27 lines did not costimulate IL-4 production by a Th2 clone and had only a small effect on IL-2 production by T cell hybridomas. The process of costimulation was fixation-sensitive, contact-dependent and did not involve stable cytokines present in the T cell/accessory cell conditioned media. Neutralizing antibodies for IL-1, IL-6, and TNF failed to inhibit costimulation. Antibodies to the LFA-1/ICAM-1 pair of adhesion molecules also failed to inhibit. Costimulation of IL-2 production by accessory cells was found to have a unidirectional species restriction: mouse accessory cells costimulated mouse and human IL-2-producing T cells, but human U937 cells induced with PMA were effective only for human T cells. The results indicate that accessory cells can significantly regulate Th1 effector function at the level of cytokine production.     

Frequency analysis of lymphokine-secreting CD4+ and CD8+ T cells activated in a graft-versus-host reaction

Lymphokine secretion by in vivo-activated T cells was analyzed at the population and single-cell levels in lymphocytes from mice undergoing an acute allogeneic graft-vs-host reaction (GVHR). Three observations were made. First, constitutive lymphokine production by these cells was very low but could be dramatically up-regulated by TCR ligation. Thus, even when harvested at the peak of the GVHR, fewer than 0.1% of lymphocytes secreted detectable granulocyte-macrophage (GM)-CSF, IFN-gamma, or IL-3 in the first 24 h in vitro, and average production of these lymphokines in bulk cultures was less than 10(-5) U/cell. However, when cultured for 24 h with anti-CD3 antibody under conditions which activated less than 0.1% of normal cells, about 30% of GVHR T cells secreted GM-CSF, IFN-gamma, and/or IL-3, and average production levels were increased by 10(3)- to 10(4)-fold. Together with evidence that host alloantigen-induced lymphokine secretion was 10 to 100 times lower than the anti-CD3 response, these data suggest that physiologic lymphokine synthesis by most T cells is low (less than 10(-18) mol of IL-3 per cell) but can be raised above the threshold of detection by TCR cross-linking. Second, individual GVHR lymphocytes varied markedly in their total and relative production of different lymphokines in response to anti-CD3 stimulation, with some cells secreting IL-3 alone, some secreting IL-3 accompanied by other lymphokines (GM-CSF and/or IFN-gamma), and some secreting other lymphokines without detectable IL-3. Finally, both CD4+ and CD8+ T cells from GVHR mice responded to anti-CD3 antibody by secreting IL-3 and other lymphokines: purified CD4+ cells contained an average of 16% and CD8+ cells an average of 10% anti-CD3-inducible lymphokine-secreting cells. By contrast, only 2 to 3% of cells of either subset formed clones in cultures with host allogeneic cells and IL-2, suggesting that clonogenic alloreactive cells were a minority of the T cells activated in the GVHR.