Quiescence-Inducing and Antiapoptotic Activities of IL-15 Enhance Secondary CD4⁺ T Cell Responsiveness to Antigen¹

Interleukin 2 is produced transiently by T lymphocytes in response to an antigenic stimulation and is a central regulator of the acute phase of the immune response. This is reflected by its dual activity on TCR-activated T cells. IL-2 acts as a strong growth factor, promoting expansion of the activated T cell population. More recently, however, IL-2 has also been described as a cytokine that renders activated T cells susceptible to cell death induced by repeated TCR engagement (reviewed in 1 . This idea was originally proposed by Lenardo (2) and has since been confirmed by several studies using mice deficient in IL-2 signaling pathway (3, 4, 5). Although thymic and peripheral T cell development is apparently normal in IL-2, IL-2Rα, and IL-2Rβ knockout mice, these animals suffer from severe lymphoproliferative diseases and autoimmunity. This phenotype is a consequence of the inability to activate the Fas/Fas ligand (FasL)³ death pathway (3), which is the main effector mechanism for maintenance of peripheral lymphoid homeostasis (reviewed in 6 . In addition to TCR-induced death, apoptosis induced by deprivation from growth factor is a second general mechanism for restoration of cellular homeostasis after subsidence of an immune response (7). The transient nature of IL-2 production by activated T cells, critically dependent on growth factor for their survival, adds to the deletion of superfluous effector cells once the Ag has been cleared successfully. Therefore, IL-2 contributes to both the development and the conclusion of a primary immune response.

IL-15 is a cytokine that was cloned from CV-1/EBNA (8), a simian kidney epithelial cell line, and from the human T cell leukemia cell line HUT-102 (9). Although IL-15 does not show sequence homology with IL-2, both cytokines share many biologic functions. IL-15 induces proliferation of the CD8⁺ T cell clone CTLL-2 and of phytohemagglutinin-activated CD4⁺ and CD8⁺ human peripheral blood T lymphocytes (8). IL-15, like IL-2, promotes differentiation and growth of human B cells (10), induces development of NK cells (11), activates cytotoxic activity of NK cells (12), and is a chemoattractant for T lymphocytes (13). These overlapping activities are not surprising, since both cytokines use the same IL-2Rβ and γ_c-chain for binding and signal transduction (14). Nevertheless, the composition of the high-affinity receptor of both cytokines differs, because IL-15 uses a specific IL-15R α-chain that is structurally similar to IL-2Rα, but does not bind IL-2 (15). As a consequence, a differential regulation of both α-chains, as reported by Kumaki et al. (16), may determine the reactivity to either cytokine. Also, IL-15 and IL-2 differ in their cellular source of production. IL-15 mRNA is most abundantly found in fibroblast and epithelial cell lines, placenta, skeletal muscle, and activated peripheral blood monocytes (8). Activated T lymphocytes, however, do not produce any IL-15 protein (17) but are the exclusive source of IL-2. The transient and local nature of IL-2 production, as opposed to the presumed persistent and systemic availability of IL-15, provides a mechanism in vivo for the distinct activities of these cytokines. Furthermore, it has recently been demonstrated that IL-15, but not IL-2, protects against Fas-mediated apoptosis in the liver, spleen, and thymus of mice treated with an IL-15 IgG2b fusion protein (18). This result indicates that both cytokines also exert different functions. Therefore, it is reasonable to hypothesize that IL-15 has a distinct place in the regulation of T cell responses.

In the present study, we followed the pro- and antiapoptotic as well as growth-inducing activities of IL-2 and IL-15 during the course of an in vitro CD4⁺ T cell response. The fate of the T cells was followed during TCR activation, after subsidence of activation when the cells became devoid of autocrine growth factor, and finally, during rechallenge with Ag.

Female C57BL/6 (H-2^b) mice were purchased from the Broekman Instituut (Eindhoven, The Netherlands). All mice were used at the age of 9 to 14 wk.

The influenza A/H3 hemagglutinin (HA)-specific and H-2^b-restricted CD4⁺ murine T cell clone T-HA was developed in our laboratory by initial immunization of C57BL/6 mice with 10 μg bromelain-cleaved hemagglutinin (BHA) and 0.1 ml Ribi adjuvant (Ribi Immunochem Research, Hamilton, MT) and a second immunization with 3 μg BHA after 3 wk. Five days after this boost immunization, lymph nodes were isolated, and 3 × 10⁷ cells were stimulated in vitro with 0.5 μg/ml BHA in 25-cm² culture flasks. On day 4, 10 U/ml murine IL-2 (from PMA-stimulated EL4.IL-2 cells) was added to the cultures. After 2 additional biweekly restimulations with 0.5 μg/ml BHA and APC, a pool of optimally HA-reactive T lymphocytes was obtained. T-HA cells were maintained long term in vitro by biweekly restimulation in 25-cm² culture flasks with 10 ng/ml BHA and 7 × 10⁷ syngenic spleen cells from C57BL/6 mice (3000 rad gamma-irradiated). On day 2, 30 IU/ml of human IL-2 was added, after which T cells were further cultured and expanded by medium renewal and addition of IL-2 every 4 days. T-HA cells were cultured in 12.5 mM HEPES-buffered RPMI 1640 (Life Technologies, Paisley, Scotland) supplemented with 10% FCS, 2 mM GlutaMAX-I (Life Technologies), 100 U/ml penicillin, 100 μg/ml streptomycin, 1 mM sodium pyruvate, and 5 × 10⁻⁵ M 2-ME.

Human rIL-15 was purchased from PeproTech (London, U.K.) and had a sp. act. of 2 × 10⁶ U/mg. Human rIL-2 was produced in our laboratory and had a sp. act. of 1.3 × 10⁷ IU/mg as determined in a CTLL-2 assay (1 IU corresponds to 77 pg). These cytokine preparations were used throughout this study.

T-HA cells were harvested from cultures by incubation in nonenzymatic cell dissociation buffer (Sigma Chemical, St. Louis, MO). Dead cells were removed by centrifugation on a Histopaque-1077 density gradient (Sigma-Aldrich, Irvine, U.K.) for 25 min at 2000 rpm. Routinely, cultures contained less then 2% dead cells after density gradient centrifugation. Recovered cells were washed three times before further use. For IL-2 or IL-15 pretreatment, 2–5 × 10⁵ viable T-HA cells were cultured for 48 h in 24-well flat-bottom tissue culture plates in the presence of variable concentrations of IL-2 or IL-15.

Viable cell numbers were determined in a hemocytometer on the basis of trypan blue exclusion. Apoptosis was analyzed by addition of 30 μM propidium iodide (PI; ICN Pharmaceuticals, Costa Mesa, CA) to harvested cells; the percentage of PI⁺ cells was measured with an EPICS 753 flow cytometer (Coulter Electronics, Luton, U.K.), equipped with an argon-ion laser emitting at 488 nm, after gating out cell debris. PI fluorescence was detected at 610 to 630 nm. Additionally, the percentage of apoptotic cells was also determined by forward light scatter analysis (not shown). In all experiments, data obtained by the latter method correlated well with the PI dye uptake data.

For immunofluorescence, rat anti-mouse CD25 (clone PC 61), rat anti-mouse CD71 (clone R217 17.1.3, kindly provided by Dr. G. Leclercq, Ghent, Belgium), and hamster anti-mouse CD95 (clone Jo2; PharMingen, San Diego, CA) were used as primary Abs. Anti-CD25 and anti-CD71 binding was detected with a FITC-conjugated goat anti-rat IgG (Sera-Lab, Crawley Down, U.K.). FITC-conjugated anti-hamster IgG (clone G70-204; PharMingen) was used as secondary Ab for anti-CD95. Purified anti-CD3 mAb (145-2C11; kindly provided by Dr. G. Leclercq) was used at a concentration of 10 μg/ml in PBS to coat flat-bottom microwells (30 μl/well) for 2 h at 37°C. Unbound Ab was removed before adding cells. The mitochondrial transmembrane potential was measured by addition of 1 μM rhodamine 123 (Molecular Probes, Eugene, OR) to the cells for 30 min and subsequent flow cytometric analysis of the fluorescence intensity.

Cells, cultured under the conditions indicated, were harvested and washed three times to remove cytokines. Cytokine-induced proliferation was measured by incubating 1 × 10⁴ T-HA cells with serial dilutions of IL-2 or IL-15. [³H]TdR (Amersham Life Science, Amersham, U.K.) was added at 0.5 μCi/well for the last 8 to 12 h of incubation. Cells were harvested on glass fiber filters, and [³H]TdR incorporation was measured by liquid scintillation in a TopCount (Packard Instrument, Meriden, CT). All results are means of triplicate cultures. Ag-induced proliferation was determined with 200 ng/ml BHA and 2 × 10⁵ irradiated C57BL/6 spleen cells (as a source of APC) in 96-well flat-bottom microtiter plates. Cultures were pulsed with [³H]TdR for the last 12 h of an 84-h assay period. Results shown are the means of triplicate wells. Cocultures of T cells and APC without Ag were always included in the experiments as a control on the Ag dependency of the response. Proliferation of these cultures never exceeded 1000 cpm (not shown).

T-HA cells were harvested, washed once in cold PBS, and lysed in Krishan’s reagent (0.05 mg/ml PI, 0.02 mg/ml ribonuclease A, 0.3% Nonidet P-40, 0.1% sodium citrate). Cell nuclei were analyzed for DNA content by flow cytometry; the distribution of cells along the distinct stages of the cell cycle was calculated with Para1 software (Coulter Electronics).

mRNA isolation was conducted with a MicroFastTrack kit (Invitrogen, San Diego, CA). cDNA was synthesized in the presence of RNase block (Stratagene Cloning Systems, La Jolla, CA) after the addition of oligo(dT) primer (Boehringer, Mannheim, Germany) and incubation at 37°C with Superscript II reverse transcriptase (Life Technologies). The primers used for PCR amplification were 5′-CAGCTCTTCCACCTGCAGAAGG-3′ and 5′-CAATATTCCTGGTGCCCATGAT-3′ (murine FasL, 597-bp fragment), as well as 5′-TGGAATCCTGTGGCATCCATGAAAC-3′ and 5′-TAAAACGCAGCTCAGTAACAGTCCG-3′ (murine β-actin, 348-bp fragment). The PCR reaction mixture contained 1.5 mM MgCl₂, 0.4 mM dNTP, 200 nM primers, and 0.5 U of Goldstar Taq polymerase (Eurogentec, Seraing, Belgium). Samples were amplified during 35 cycles (FasL) or 30 cycles (β-actin) (1 min denaturation at 94°C, 2 min annealing at 58°C, and 1 min extension at 72°C) in a Peltier Thermal Cycler-200 (MJ Research, Watertown, MA). In each PCR, water was included as a negative control. For semiquantitative RT-PCR, four twofold dilutions of each cDNA sample were amplified. PCR products were analyzed on a 2% agarose gel and visualized by ethidium bromide staining.

T-HA cells were harvested and washed twice in serum-free medium. Cells (1 × 10⁶–10⁷) were resuspended in 1 ml of diluent A and stained with the membrane stain PKH2-GL (2 μM; Sigma Chemical) following the manufacturer’s instructions. Stained cells were washed twice with serum-containing medium and were incubated overnight in their culture medium to allow dissociation of excess dye from the membrane. In mixed cultures of PKH-2GL-stained T-HA cells and splenocytes, percentages of viable and apoptotic T-HA cells were obtained by flow cytometric analysis of PI⁻ and PI⁺ cells, respectively, emitting green fluorescence (525 nm).

Spleen cells (8 × 10⁸) were prepared from the spleens of naive, 8-wk-old C57BL/6 mice and were activated in 25-cm² tissue culture flasks with 1 μg/ml soluble anti-CD3 mAb (145-2C11). After 24 h, excess Ab was removed, and cells were further cultured for 72 h without addition of exogenous cytokine. Following this stimulation period, cultures were harvested, and CD4⁺ T cells were isolated by immunomagnetic cell sorting. A negative selection procedure, using an Ab mixture designed for the enrichment of murine CD4⁺ T cells (StemSep; Stem Cell Technologies, Vancouver, Canada), was followed according to the manufacturer’s instructions. Recovered cells (7.5 × 10⁶) were further cultured for 10 days and supplemented (every fourth day) with their respective cytokines (none, 10 ng/ml IL-15, or 10 ng/ml IL-2). Viable cell numbers were determined on day 14, based on trypan blue dye exclusion. For restimulation, we used 1 μg/ml soluble anti-CD3 mAb and the immortalized macrophage cell line Mf4/4 (19). Before use, Mf4/4 cells were activated for 24 h with 400 U/ml IFN-γ to enhance expression of costimulatory molecules. The cells were then treated for 90 min with 30 μg/ml mitomycin C (Duchefa, Haarlem, The Netherlands) to block their proliferation, thus avoiding interference with proliferation measurements from the restimulated lymphocytes. Alternatively, for determination of susceptibility to anti-CD3-induced death, freshly isolated, unsorted spleen cells were activated for 72 h in 24-well plates with 1 μg/ml soluble anti-CD3 mAb (145-2C11) without exogenous cytokine and were supplemented on day 3 with 10 ng/ml IL-15 or IL-2. After an additional 8-day culture period, viable cells were isolated on a Histopaque density gradient and restimulated with plate-bound anti-CD3 mAb (10 μg/ml). Apoptotic cell numbers were determined after 24 h by PI dye uptake. CD4:CD8 ratios were determined by labeling 1 × 10⁵ cells with 0.5 μg PE-conjugated rat anti-mouse CD4 mAb (PharMingen) and 0.5 μg/ml FITC-labeled rat anti-mouse CD8 mAb (clone 53-6.7, kindly provided by Dr. G. Leclercq) and, after gating out dead cells and debris, analysis of stained populations on a FACScalibur flow cytometer (Becton Dickinson (Sunnyvale, CA)). Absolute numbers of CD4⁺ T cells in the respective cultures were calculated from the percentages obtained and total viable cell countings by trypan blue dye exclusion.

The T-HA helper T cell clone was routinely cultured by biweekly antigenic stimulation followed by addition of exogenous IL-2. This culture condition has become the standard procedure to propagate Ag-specific T cell clones for prolonged periods in vitro. Addition of IL-2 at the time when autocrine production ceases is necessary to ensure further survival of the activated T lymphocytes in the periods between a repeated challenge with Ag. However, this IL-2 not only promotes survival but also supports further expansion of the T cells, thus keeping the T cells in a semiactivated, proliferative state, not representative of the in vivo situation in which it is believed that Ag-stimulated T cells persist as small, resting lymphocytes once the Ag has been cleared (20). In our study, we replaced IL-2 with IL-15 in the periods between antigenic stimulation and examined the evolution of viable cell numbers and occurrence of cell death. T-HA cells, harvested 4 days after stimulation with Ag/APC, were cultured in the presence of decreasing concentrations of IL-15 (6.6–0.03 ng/ml; 460.0–2.0 pM) or IL-2 (9.0–0.03 ng/ml; 585.0–2.0 pM). After 3 days of treatment, the absolute numbers of viable cells and the percentage of apoptotic cells in the various cultures were determined (Fig. 1). As expected, incubation with IL-2 resulted in a dose-dependent increase in cell numbers (Fig. 1,A). In contrast, treatment with IL-15 kept the number of viable cells stable at ∼30,000, which is slightly above the input of 20,000 (Fig. 1 B). Numbers of viable cells dropped dramatically when IL-15 or IL-2 was omitted from the cultures. Concomitantly, extensive cell death was observed as a consequence of growth factor deprivation. Addition of IL-15 reduced cell death to background levels, comparable to IL-2. Even with concentrations as low as 0.08 ng/ml (6 pM) IL-15, no significant cell death was observed. It may be noted that the minimal concentration of IL-2 required for a similar full protection was ∼30- to 40-fold higher, namely 3 ng/ml (200 pM). From these data, we conclude that IL-15 induces a survival signal in Ag-primed CD4⁺ T lymphocytes that, contrarily to IL-2, is not accompanied by an increase in cell number.

IL-15 has been described previously as a factor that induces and sustains the growth of mitogen-stimulated CD8⁺ T cells and also CD4⁺ T cells, much in the same way as IL-2 does (8). However, the observed stabilization of cell numbers without an apparent occurrence of cell death suggests that IL-15 silences DNA synthesis after TCR-induced stimulatory signals have subsided. To verify whether these stable cell numbers in IL-15-treated cultures indeed reflect an IL-15-induced growth arrest, T-HA cells derived from standard IL-2 cultures were assayed for proliferation in response to increasing concentrations of IL-15 or IL-2. From the results shown in Figure 2,A, it is clear that IL-15, even in concentrations as high as 200 ng/ml (14 nM), does not induce DNA synthesis. Furthermore, starting from an actively dividing population, IL-15 induced a gradual transition of T cells to a nondividing condition, while IL-2 further supported cell proliferation (Fig. 2,B). These data, together with the results from Figure 1 B, demonstrate that IL-15 is a survival factor but not a growth factor for CD4⁺ T lymphocytes when TCR aggregation is absent.

It is thought that after conclusion of a primary immune response, a fraction of activated effector cells reverts to a resting state and persists in the animal as a population of small lymphocytes, ready for a “memory” response in case of reemergence of their Ag (20). We wondered whether T-HA lymphocytes surviving with IL-15 without cycling could be phenotyped as small, resting lymphocytes. Therefore, a number of features generally recognized as parameters for lymphocyte quiescence were studied. We determined whether the observed growth arrest took place in a specific phase of the cell cycle. Cell cycle analysis by flow cytometry revealed that IL-15-treated cells accumulated in the G₀/G₁ phase (Fig. 3), indicative of the induction by IL-15 of an arrest in cell cycle entry. Thus, cycling cells treated with IL-15 are arrested neither immediately nor randomly, which in fact would be apoptosis inducing, but proceed with their cycle until they reach G₀/G₁ and then exit cell cycle progression in an orderly manner without triggering programmed cell death (PCD). Additionally, cell size, expression of activation markers, and the mitochondrial transmembrane potential as indicators of the metabolic state of the cells were evaluated. IL-15-treated T-HA cells exhibited all the hallmarks of resting cells: the cells were small, expressed low levels of the CD25 (IL-2Rα) and CD71 (transferrin receptor) activation markers, and had a low mitochondrial transmembrane potential (Fig. 3). In contrast, IL-2-cultured cells were large blastoid cells with high CD25 and CD71 expression levels and a high oxidative metabolism, as indicated by the increased mitochondrial transmembrane potential. Thus, the IL-15-induced arrest in the G₀/G₁ phase of T-HA cells is accompanied by acquisition of a typical quiescent phenotype.

IL-2, secreted by TCR-activated T lymphocytes, regulates TCR-induced death by sensitizing T cells to execution triggered by Fas/FasL interactions (1, 6). This mechanism for TCR-induced cell death has primarily been demonstrated by pretreating CD4⁺ T cells with IL-2 and subsequently activating them with plate-bound anti-CD3 mAb. We compared IL-2 and IL-15 in their ability to sensitize to TCR-induced death. In a first type of experiment, T-HA cells were pretreated for 48 h with either IL-15 or IL-2, then assayed for their sensitivity to cell death induced by immobilized anti-CD3 mAb. As described above, cell cultures differed in their survival and cell cycle status at the moment viable cells were isolated for CD3 triggering, according to the cytokine added during pretreatment: IL-15 kept the T-HA cells fully viable but in a growth-arrested state, high-dose IL-2 (10 ng/ml; 650 pM) induced vigorous cell cycling, and low-dose IL-2 (0.1 ng/ml; 6.5 pM) resulted in poor viability (50% survival). As shown in Figure 4,A, T-HA cells pretreated with high or low IL-2 concentrations were susceptible to anti-CD3-induced death. This cell death, shown as percentages of cells that had taken up the exclusion dye PI, showed all of the typical features of apoptosis, namely membrane blebbing, nuclear condensation, hypoploidy, and disintegration of the cells into apoptotic bodies (not shown). In contrast to IL-2, IL-15 pretreatment resulted in a strong resistance to anti-CD3-induced death (Fig. 4,A). As Fas/FasL interaction is believed to be the actual trigger of TCR-induced death in mature T lymphocytes (21), we evaluated whether altered levels of Fas or FasL expression in IL-2- or IL-15-treated T-HA cells underlaid the differential susceptibility to anti-CD3-induced death. Although Fas expression before activation was slightly lower in IL-15- than in IL-2-cultured cells, in agreement with the resting vs activated state of the respective populations, both showed a vigorous and similar up-regulation of Fas in response to CD3 triggering (Fig. 4,B). Also, IL-2 or IL-15 treatment did not significantly affect FasL mRNA levels before and after activation (Fig. 4 C). As a control, β-actin was amplified to make sure that equivalent amounts of cDNA were present in all samples. These data demonstrated that protection against anti-CD3-induced death by IL-15 was not the consequence of an IL-15-induced impairment to express Fas or FasL after TCR activation. They suggest, rather, that IL-15-mediated protection is based on interference with the Fas/FasL-signaling pathway. It has recently been documented that IL-15 protects CD8⁺ T cells and B cells against Fas-induced apoptosis (18). Our results confirm these data and add that CD4⁺ T cells also can be protected against TCR-induced apoptosis by IL-15.

Cell death also occurs in response to appropriate T cell activation (1, 6), namely TCR ligation in the presence of costimulatory ligands. Therefore, we evaluated the protective effect of IL-15 on cell death triggered by appropriate TCR stimulation with Ag. T-HA cells were treated for 48 h with IL-15 (1 ng/ml; 70 pM) or IL-2 (10 ng/ml (650 pM) or 0.1 ng/ml (6.5 pM)) before antigenic restimulation. The latter was performed in the absence of exogenous cytokines. To follow the onset of cell death after Ag activation, the IL-15- or IL-2-treated T-HA cells were labeled with the green fluorescent dye PKH2-GL, allowing them to be discriminated from APC during flow cytometric analysis of percentages of apoptotic cells. The results, shown in Figure 5, clearly demonstrate that pretreatment with 10 ng/ml of IL-2 resulted in considerable cell death 48 to 72 h after activation with Ag/APC. T-HA cells precultured with IL-15, or those that had survived low dose (0.1 ng/ml) IL-2 pretreatment, on the contrary, showed no increase or even a slight decrease in cell death as compared with initial background levels at 24 h. These data clearly demonstrate that T cells cultured with IL-15 are desensitized to cell death triggered by appropriate stimulation with Ag. Also, the fraction of cells that survived in the cultures with low dose IL-2 (50%) were resistant to TCR Ag-induced death. In agreement with previous reports (1, 6), high levels of IL-2 raise the susceptibility of the cells.

Resistance against TCR Ag-induced cell death is expected to result in an enhanced proliferative response of T-HA lymphocytes upon exposure to Ag. To verify this, we pretreated T-HA cells with IL-15 or high-dose IL-2 and stimulated the cells with their Ag presented by appropriate spleen APC. These Ag-stimulated cultures were set up either without exogenous cytokine or with the addition of IL-2 or IL-15. [³H]TdR was added to the cultures after 72 or 120 h. The data shown in Figure 6,A demonstrate that pretreatment with IL-15 resulted in enhanced proliferation to Ag as compared with IL-2 pretreated cells, whether or not IL-2 or IL-15 was added during the restimulation period. This indicates that the status of resistance to TCR Ag-induced death caused by IL-15 enhances the potential of T-HA cells to respond to a renewed Ag challenge. Addition of IL-15 to the cultures during the assay period further supported the proliferative response. Although this effect was limited at 72 h, probably due to competition with autocrine IL-2 for IL-2Rβ and γ-chains, it became spectacular when the cells were cultured for 120 h (Fig. 6 B). At this time point, proliferation in response to Ag subsided. The availability of IL-15, but not of IL-2, at this time point not only prolonged but also further increased the response. Clearly, IL-15 not only acts as a survival factor but, in these conditions, also shows a pronounced growth factor activity. Correlating these results with the conclusions drawn before, it may be concluded that IL-15 either induces quiescence or supports growth, depending on whether TCR cross-linking occurs or not. Furthermore, this concept reconciles the quiescence-inducing activity that we have demonstrated herein with previous reports describing IL-15 as a cytokine with growth factor activity for PHA-activated CD4⁺ T lymphocytes (8).

Fresh, unsorted spleen cells from naive C57BL/6 mice were isolated and polyclonally stimulated in vitro. The stimulus consisted of soluble anti-CD3 mAb (1 μg/ml), which in the presence of costimulation by spleen APC, polyclonally activates naive T cells (22). After 24 h, the remaining anti-CD3 mAb was removed, and the activated cells were further cultured in the absence of exogenous cytokine. To confirm that activation occurred, anti-CD3-activated and unstimulated cells were pulsed with [³H]TdR. Soluble anti-CD3 mAb induced a strong proliferative response: 25,304 cpm as opposed to 2,581 cpm for unstimulated cells. On day 4, CD4⁺ cells were isolated by immunomagnetic cell sorting and further cultured without cytokine or in the presence of IL-15 (1 ng/ml; 70 pM). After 10 days of culture in the absence of exogenous cytokine, viable cell numbers had dropped to 15% of the cell input, while IL-15 maintained cell numbers at ∼60% of cell input (Fig. 7,A). Cells surviving with IL-15 (10 ng/ml; 700 pM) appeared as small resting lymphocytes and did not reveal DNA synthesis (59 cpm), whereas proliferation could be induced with IL-2 (4,184 cpm with 10 ng/ml (650 pM)). Hence, for freshly isolated and TCR-activated CD4⁺ T cells as well, IL-15 acts as a survival factor and induces quiescence. Next, we investigated the resistance to TCR-induced cell death, triggered by immobilized anti-CD3 mAb, in these polyclonally activated T cell cultures. The CD4⁺ T cell population maintained throughout with IL-15 was largely resistant, whereas cells cultured with IL-2 showed extensive cell death (Fig. 7,B). Finally, CD4⁺ T cells residing in an IL-15-induced resting state proliferated in response to renewed stimulation with soluble anti-CD3 and APC, while cells maintained with IL-2 did not (Fig. 7 C). Also, the addition of IL-15 to the IL-15 pretreated cultures further increased the proliferative response, thus confirming the growth-promoting activity of IL-15 in the presence of TCR aggregation. These experiments demonstrate that the characteristics induced by IL-15 in the clonal CD4⁺ T cell T-HA, namely long term survival as a resting population, resistance to apoptosis, and increased responsiveness to TCR restimulation are also acquired by freshly isolated CD4⁺ T cells treated with IL-15.

To verify the significance for long term T cell responses of the combined activities exerted by IL-15, a comprehensive experiment was conducted (Table I). T-HA cells, primed either by IL-15 or by high dose or low dose IL-2, were given a first stimulation with Ag/APC and, after subsidence of the response, were further cultured with the respective cytokine concentrations for 8 additional days. Next, the cultures were analyzed for their quantitative and qualitative secondary response potential. Starting from a fixed number of IL-15-treated T-HA cells (1 × 10⁵), the combination of optimal proliferation in response to Ag/APC stimulation and subsequent long term persistence of the generated effector cells with IL-15 resulted on day 12 in a 16-fold increase of T cells available for a renewed Ag/APC response (Table I, Expt. 1). A similar treatment schedule with high dose IL-2 or low dose IL-2 raised T cell numbers 7.6- and 1.2-fold, respectively. A comparable result was obtained in another independent experiment (Table I, Expt. 2). Next, the various cultures were harvested, and equal cell numbers were examined for their functionality upon a second antigenic restimulation. As expected from the above experiments, cells cultured with IL-15 or low dose IL-2 expanded vigorously, resulting in the accumulation of high numbers of immune effector cells on day 16, as measured by both cell counting (Table I, Expt. 1) and [³H]TdR incorporation (Table I, Expt. 2). These different yields and this differential responsiveness to Ag/APC of the respective T cell populations were combined in a recovery and reactivity index, indicative of the strenght of the secondary immune response. As shown in Table I, these indices are dramatically higher for the cell populations kept in IL-15. Obviously, the enhancement of availability and response potential of CD4⁺ T cells by IL-15 has a cumulative effect, resulting in strongly enhanced secondary responses. These features could not be achieved by either dose of IL-2. These results clearly demonstrate that IL-15, but not IL-2, has the properties required for generating an efficient secondary T cell response, thus providing a strong survival signal that allows the long term persistence of immune effectors in a quiescent state as well as simultaneously priming these cells for an optimal response when Ag exposure reoccurs.

In this report, we have compared the antiapoptotic and growth-inducing properties of IL-15 and IL-2, not only on TCR-activated CD4⁺ T lymphocytes but, importantly, also on T cells in which TCR signals had subsided. Most of our experiments were performed with the long term CD4⁺ T cell clone T-HA, which is dependent for survival on regular restimulation with Ag (influenza hemagglutinin). We believe that prolonged in vitro culture did not affect pathways involved in the regulation of cell death nor TCR responsiveness in this T-HA clone. This is supported by the fact that T-HA cells die in the absence of growth factor (IL-2) and are susceptible for TCR-induced death in the presence of IL-2. Furthermore, the basic observations made in experiments with T-HA cells were confirmed for freshly isolated CD4⁺ T cells.

Perhaps the most important new finding of our study is the ability of IL-15 to keep Ag-experienced T cells in a quiescent condition for prolonged periods by providing the necessary signals for survival in the absence of TCR engagement. Persistence of Ag-experienced T lymphocytes for prolonged periods—according to some studies, even for the life span—after elimination of the pathogen involved, likely requires survival factors from the microenvironment to maintain these T cells, devoid of autocrine growth factors, in a primed but resting state (23). Candidates that have recently been proposed for this helper function are an unidentified, >30-kDa factor secreted by stromal cells (24), which promotes T cell survival without inducing proliferation, and TGFβ (25). TGFβ can reverse the effector T cell population into a more resting state and synergizes with IL-2 to prevent the induction of apoptosis in T lymphocytes of the Th2 type. Also, IL-4, IL-7, and IL-15, initially described as T cell growth and differentiation factors (8, 26, 27, 28), induce a strong survival signal in T cells deprived of endogenous IL-2 (29). However, this protection was always accompanied by the onset of cell cycle progression, also when IL-15 was used to maintain the cells alive. The fact that the authors used T lymphocytes shortly after activation with PHA possibly explains the observed mitogenic activity of IL-15 in their system. Furthermore, it is unlikely that, after subsidence of an immune response, the levels of IL-2, IL-4, or IL-7 in the animal would be sufficient to support long term survival of Ag-experienced T cells. On the other hand, the previously described widespread distribution of IL-15 mRNA in the placenta, skeletal muscle, kidney, lung, heart, fibroblasts, epithelial cells, and monocytes, but not in T cells (8, 17), agrees with the notion of IL-15 as a microenvironmental factor that also remains available when the immune response has been terminated. This persistent presence of IL-15, combined with our data showing that IL-15 induces quiescence when TCR engagement is absent and simultaneously permits survival in the absence of autocrine growth factor, suggests that IL-15 could be an important cytokine for the survival of descendants of activated CD4⁺ T lymphocytes as resting memory cells. Although the continuous presence of IL-15 was required to maintain the state of resistance to PCD (data not shown), our data nevertheless demonstrate that very low levels (0.08 ng/ml (6 pM) or less) of the cytokine, likely to be available in different tissues, are sufficient for T cell survival. Recent studies suggesting the importance of IL-4, IL-6, and IL-7 for survival of virgin T cells (30, 31) further support the idea that cytokines play a pivotal role in the long term maintenance of T cells in vivo.

Several authors have demonstrated that IL-2, transiently secreted during immune reactivity, regulates the immune response in a bivalent way by promoting T cell clonal expansion as well as sensitizing the cells to TCR-induced cell death triggered by Fas and/or TNF-R55 (reviewed in 1 . The prevention of TCR-induced cell death that we observed in CD4⁺ T lymphocytes exposed to IL-15 provides an escape mechanism from cell death for Ag-primed T cells. The mechanism by which IL-2 sensitizes T lymphocytes to TCR-induced cell death remains unclear. Signals that affect cell division are known also to affect the cell death program. Overexpression of the survival factor Bcl-2 retards transition from G₀ to S phase and represses TCR-induced death, whereas the opposite effects are observed in T lymphocytes from Bcl-2-deficient mice (32, 33). Also, mature T cells expressing a proapoptotic baxα transgene show accelerated S phase entry in response to IL-2 (34). However, the correlation between the proapoptotic property of IL-2 and its ability to drive T lymphocytes into the S phase of the cell cycle remains controversial (35, 36, 37). Our observation that IL-15-treated CD4⁺ T lymphocytes are blocked in the G₀/G₁ phase and are desensitized to TCR-induced death supports the hypothesis whereby this cell death is correlated with IL-2-driven cell cycling. Interestingly, our results show that T cells, once made resistant to TCR-induced cell death by IL-15, remain protected even when cycling in response to autocrine IL-2 (Fig. 5), indicating that in addition to growth arrest, protection also involves induction of antiapoptotic proteins or down-regulation of proapoptotic proteins. In this respect, Bulfone-Paus et al. (18) recently provided evidence that IL-15-mediated suppression of anti-Fas induced T cell apoptosis is strictly dependent on RNA synthesis. Regarding the growth-supporting activity of IL-15 upon TCR triggering, this observation can be explained by the protective effect of the cytokine against TCR-induced death, thus abolishing the negative feedback on autocrine IL-2-driven proliferation. Also, it was recently described how IL-15 potentiates Con A-induced IL-2 secretion in human T lymphoblasts (38). We are currently investigating whether a similar mechanism exists in TCR-activated murine T cells. However, it cannot be excluded that IL-15 also exerts its growth-promoting activity in a direct way. A TCR-induced up-regulation (or down-regulation) of IL-15Rα as a mechanism for IL-15-induced growth or quiescence seems unlikely considering that both activities were obtained at similar, low IL-15 concentrations, hence implicating the high affinity IL-15Rα in both functions. Also, both differential activities could be elicited with simian IL-15 (data not shown), which can bind on murine cells only when IL-15Rα is present but not when only the dimeric IL-2Rβγ_c is available (15). An involvement of IL-15RX, a recently described new type of IL-15R that is present on mast cells (39), in the observed IL-15 activities is contradicted by the inhibitory effect of anti-IL-2Rβ mAb (data not shown). This leaves open the possibility that by some intricate mechanism the signaling pathway activated by the trimeric IL-15R complex differs in the absence or presence of TCR cross-linking.

Collectively, our results show that IL-2 and IL-15 differ dramatically in their antiapoptotic and growth-inducing properties. Also, with regard to secondary CD4⁺ T cell responses, IL-15 elicited an enhancing activity that could not be mimicked by high or low doses of IL-2. These results support the view that IL-15 is an important regulator of CD4⁺ T cell responses during and after TCR triggering, as distinct from IL-2. Based on the nature of the IL-15 activities we describe herein, we propose that this regulatory role of IL-15 consists of promoting the generation of resting, long-lived CD4⁺ memory T lymphocytes in vivo. Evidence to support this hypothesis has to come from future in vivo studies evaluating immune memory formation in IL-15-treated animals.

We thank D. Ginneberge for technical assistance.