IL-4 regulates IL-2 induction of lymphokine-activated killer activity from human lymphocytes

IL-4 is a pluripotent lymphokine acting on various cell types. We investigated the role of human IL-4 on the generation of lymphokine-activated killer (LAK) activity. Human IL-4 alone did not induce LAK activity and inhibited IL-2 induction of LAK activity from unstimulated PBMC, peripheral blood null cells, spleen cells, and lymph node cells in a dose-dependent manner. IL-4 also inhibited several phenomena induced by IL-2 such as cell proliferation, augmentation of NK activity, increase of Leu-19+ cells, and expression of IL-2R(p55) on either CD3+ or Leu-19+ cells. IL-4, however, augmented cell proliferation with other T cell mitogens including PHA, Con A, PMA, or allo-MHC Ag with or without IL-2. In contrast to unstimulated cells, IL-4 alone induced marked cell proliferation and LAK activity as well as Leu-19+ cells from in vitro IL-2 preactivated PBMC or null cells, and did not inhibit IL-2 induced cell proliferation, LAK activity, Leu-19+ cells and IL-2R(p55) expression, but rather augmented them with low doses of IL-2. Although IL-4 alone induced LAK activity from peripheral blood of some patients previously given IL-2, IL-4 inhibited in vitro LAK generation with IL-2 from these cells in most cases. Therefore, IL-4 appears to directly inhibit the IL-2 activation pathway via IL-2R(p70) and prevent resting LAK precursors from proliferating and differentiating into final effector cells. However, once cells were sufficiently preactivated by IL-2, IL-4 induced LAK activity and did not inhibit IL-2 activation of these cells. These data suggest an immunoregulatory role of IL-4 on human null cells and T cells.     

Transforming growth factor-beta inhibits the in vitro generation of lymphokine-activated killer cells and cytotoxic T cells

Summary The effect(s) of purified transforming growth factor-beta (TGF-beta) and platelet-derived growth factor (PDGF) on the induction and function of lymphokine-activated killer (LAK) cells and cytotoxic T lymphocytes (CTL) was examined. The addition of TGF-beta, but not PDGF, to cultures containing fresh C57BL/6 mouse splenocytes or human peripheral blood lymphocytes plus recombinant interleukin-2 markedly inhibited the development of mouse and human LAK cell activity (measured after 3 days for cytotoxicity against cultured or fresh tumor targets in 4-h ⁵¹Cr release assays). The addition of TGF-beta, but not PDGF, to a one-way, C57BL/6 anti-DBA/2, mixed lymphocyte reaction effectively blocked the generation of allospecific CTL as well. However, TGF-beta did not inhibit the effector function of LAK cells or of allospecific CTL when added directly to the short-term cytolytic assay. A second form of homodimeric TGF-beta, type 2, was also found to be suppressive on the development of murine LAK cells and allospecific CTL. Collectively, these data demonstrate that the peptide TGF-beta is a potent inhibitor of LAK cell and CTL generation in vitro.     

Prerequisite for the induction of lymphokine-activated killer cells from T lymphocytes

Human blood mononuclear cells were separated into Leu-11+7-NK, Leu-11-7+, and Leu-11-7-T cells by means of a combination of the Percoll gradient method and C-mediated cytolysis using mAb. When purified Leu-11+7-NK, Leu-11-7+, and Leu-11-7-T cells were cultured with rIL 2 (500 U/ml) for 6 days in a medium supplemented with 10% FCS, Leu-11+7-NK cells responded at the maximum level and Leu-11-7+ cells responded moderately as shown by both cell-proliferation response and cytotoxic activity generated. On the other hand, Leu-11-7-T cells did not respond at all to rIL-2. However, when Leu-11-7-T cells were cultured with rIL-2 in a medium supplemented with 10% autologous serum, they showed considerable responsiveness to rIL-2. In addition, much greater response to Leu-11-7-T cells were produced by the addition of monocytes. Monocyte cytokines, neither IL 1, IFN-gamma, TNF, nor their combination were able to substitute for monocytes in the induction culture. In contrast, the response level of Leu-11+7- NK cells remained unchanged irrespective of supplementation with autologous serum to medium or the addition of monocytes to the culture. These results indicated that culture conditions in the experiments significantly affected the results as to determination of lymphokine-activated killer cell precursors, especially the result pertaining to the conversion of T lymphocytes to lymphokine-activated killer cells. Under appropriate conditions, not only NK cells but also T cells are important precursors of lymphokine-activated killer cells.     

Anti-Lyt-1 monoclonal antibody inhibits induction of anti-self-TNP but not of alloreactive cytotoxic T lymphocytes

The role of Lyt-1+ T cells in the induction of anti-self + TNP and anti-allogeneic CTL was investigated by the use of monoclonal antibodies directed against these molecules. When present during the 5 days of in vitro induction of CTL, anti-Lyt-1.1 mAb partially inhibited the induction of anti-self + TNP CTL but did not affect the induction of alloreactive CTL. The inhibitory effect was dose-dependent, Lyt-1 allotype-specific, and directed at the responding cells. The inhibition could be abolished by the addition of interleukin 2-containing supernatants. These results suggest that under some experimental conditions, Lyt-1-specific mAb can be shown to affect a T cell function, probably a helper function. Possible differences in the induction mechanisms of anti-self + TNP and alloreactive CTL are discussed.     

Depletion of human NK cells with monoclonal antibodies allows the generation of cytotoxic T lymphocytes without NK-like cells in mixed cultures

In vitro stimulation of human mononuclear cells with x-irradiated autologous lymphoblastoid cell line (LCL) or allogeneic normal cells in mixed leukocyte cultures (MLC) was previously shown to result in the generation of OKT3+ OKT8+ cytotoxic T lymphocytes (CTL) lytic for allogeneic and autologous LCLs and also of natural killer- (NK) like cells that are OKT3- and primarily OKT8- and are lytic for HLA- NK-sensitive K562 cells. The origin of the NK-like cells was not previously known because, although the majority of fresh human NK cells react with monoclonal antibodies OKM1 and B73.1, lymphocytes bearing these markers are not detected several days after the onset of MLC, when NK-like cells are present. In this study, experiments were undertaken to determine whether NK-like cells generated after stimulation with x-irradiated pooled allogeneic normal cells (poolx) or with autologous LCL are derived from cells expressing antigens reactive with monoclonal antibodies OKM1 or B73.1, which react with fresh NK cells. Mononuclear cells, depleted of monocytes, were stained with OKM1 or B73.1 and fluorescein-labeled goat anti-mouse IgG. Lymphocytes depleted of OKM1+ or B73.1+ cells, by fluorescence-activated cell sorting, and lymphocytes that were stained but not sorted were stimulated for 7 days with either poolx or autologous LCL. The generation of NK-like activity was decreased at least 90% after depletion of cells reactive with OKM1 or B73.1, whereas the generation of CTL against autologous and allogeneic LCL was minimally affected. These findings show that NK-like cells generated in MLC are derived from cells that express the phenotype of fresh NK cells (OKM1+ or B73.1+) and that CTL can be generated in cultures in which relatively little NK-like activity is concomitantly detected, by depleting NK cells with monoclonal antibodies before stimulation.     

Differential sensitivity of cytotoxic T lymphocytes and lymphokine-activated killer cells to inhibition by L-ornithine

The selective inhibition of murine cytotoxic T lymphocyte (CTL) differentiation in C57B1/6 (B6) anti-DBA/2 mixed leukocyte cultures (MLC) by the amino acid L-ornithine (Orn) could not be reversed by addition of up to 1000 U/ml IL-2. Analysis of the effects of Orn on induction of lymphokine-activated killer (LAK cells), using dosages of IL-2 from 10-1000 U/ml and measuring cytolytic activity against two tumor targets (P815 and YAC-1) over the course of 5 days, indicated that LAK cells were not suppressed by Orn. LAK precursors and effector cells were CD8- and ASGM1+, indicating that they were derived from natural killer (NK) cells. We also found that the growth and maintenance of cloned CTL lines were not sensitive to inhibition by Orn; nor was their acquisition of nonspecific cytolytic activity in the presence of high lymphokine concentrations. Thus, induction of naive CTL shows differential susceptibility to Orn inhibition relative to LAK and LAK-like activities by NK and cloned CTL lines in response to IL-2.     

Interleukin-6 does not support interleukin-2 induced generation of human lymphokine-activated killer cells

Summary The activity of lymphokine-activated killer (LAK) cells is supported by various cytokines. The objective of this study was to see if recombinant interleukin-6 (IL-6) either alone or in combination with interleukin-2 (IL-2) has any effect on the generation of LAK cells. Peripheral blood mononuclear cells of healthy donors were cultured for 4 or 6 days with both cytokines either alone or in combination. LAK activity against K562 and natural killer-resistant Daudi cells was assessed by a 4-h and an 18-h⁵¹Cr-release assay at various effector to target ratios. IL-6 alone in increasing concentrations did not induce LAK cell activity. Neither additive nor synergistic effects of IL-6 with IL-2 were observed. Immunofluorescence analysis with phycoerythrin-conjugated anti-CD56 antibody demonstrated that IL-6 could not maintain or increase the number of CD56-positive cells over a 6-day culture period. These results suggest that IL-6 does not support LAK cell generation by itself or increase LAK cell activity in combination with IL-2.     

IL-4 inhibits IL-2 synthesis and IL-2-induced up-regulation of IL-2R alpha but not IL-2R beta chain in CD4+ human T cells.

There is growing evidence to suggest a regulatory role of IL-4 in the immune system affecting both proliferation and lymphokine production. In the present work we have analyzed the effect of IL-4 on IL-2 and IFN-gamma synthesis by stimulating CD4+ human T cells (+10% accessory cells) with Con A in the presence of several doses (1 to 100 U/ml) of human rIL-4. The results showed an impaired IL-2 and IFN-gamma synthesis in the presence of IL-4. This inhibition was dose dependent and was evident only when IL-4 was added in the first 2 h of culture. Moreover, the external addition of IL-2 did not revert the inhibitory effect of IL-4 on IL-2 and IFN-gamma synthesis induced by Con A. We have also analyzed the effect of IL-4 on the expression of both alpha- and beta-chains of the IL-2R. Although the expression of IL-2R alpha mRNA was not modified after 6 h in culture in the presence of IL-4, a decrease was observed at 24 and 48 h. The addition of rIL-2 showed that the inhibition in IL-2R alpha expression could be explained by an impairment in the up-regulatory signal transmitted through the IL-2R. In addition to this, IL-4 did not modify the IL-2R beta mRNA expression at 6 and 24 h although a decreased expression was observed at 48 h which could be explained by the defective IL-2 production. The differential effect of IL-4 on the up-regulatory effect of IL-2 in the expression of IL-2R alpha and IL-2R beta suggest the existence of different regulatory mechanisms acting on the expression of both chains.     

Induction of the Epstein-Barr virus latent membrane protein 2 antigen-specific cytotoxic T lymphocytes using human leukocyte antigen tetramer-based artificial antigen-presenting cells

Cytotoxic T lymphocytes (CTLs) specific for the Epstein-Barr virus (EBV) latent membraneprotein 2 (LMP2) antigen are important reagents for the treatment of some EBV-associated malignancies,such as EBV-positive Hodgkin's disease and nasopharyngeal carcinoma.However,the therapeutic amount ofCTLs is often hampered by the limited supply of antigen-presenting cells.To address this issue,an artificialantigen-presenting cell (aAPC) was made by coating a human leukocyte antigen (HLA)-pLMP2 tetramericcomplex,anti-CD28 antibody and CD54 molecule to a cell-sized latex bead,which provided the dual signalsrequired for T cell activation.By co-culture of the HLA-A2-LMP2 bearing aAPC and peripheral bloodmononuclear cells from HLA-A2 positive healthy donors,LMP2 antigen-specific CTLs were induced andexpanded in vitro.The specificity of the aAPC-induced CTLs was demonstrated by both HLA-A2-LMP2tetramer staining and cytotoxicity against HLA-A2-LMP2 bearing T2 cell,the cytotoxicity was inhibited bythe anti-HLA class Ⅰ antibody (W6/32).These results showed that LMP2 antigen-specific CTLs could beinduced and expanded in vitro by the HLA-A2-LMP2-bearing aAPC.Thus,aAPCs coated with an HLA-pLMP2 complex,anti-CD28 and CD54 might be promising tools for the enrichment of LMP2-specificCTLs for adoptive immunotherapy.     

Role of IL-6, IL-2, and IL-4 in the in vitro induction of cytotoxic T cells.

The mode of IL-6-induced differentiation of Con A-stimulated CD4-CD8+ CTL-P was examined. Through application of neutralizing anti-IL-6, anti-IL-2, and anti-IL-4 mAb it was shown that IL-6 is an "early acting" factor for development of accessory cell-depleted thymocytes. IL-2 and IL-4 are obligatory "late acting" factors for this process. In accordance IL-4, but not IL-6, induced active CTL from CD4-CD8+ CTL-P. The increase of CD4-CD8+ CTL originates at least partially from CD4+CD8+ cells differentiating to active CTL in vitro. CTL development was paralleled by an increase in CD4-CD8+ cells and by a distinct increase in [3H]TdR uptake on day 2 of cultivation. Our data suggest that IL-6 induces Th cells to produce IL-2 and IL-4, the mediators for final differentiation of CD4-CD8+ cells.     

Accessory cells, dendritic cells, or monocytes, are required for the lymphokine-activated killer cell induction from resting T cell but not from natural killer cell precursors

In this study we have investigated the role of accessory cells in the development of lymphokine-activated killer cells (LAK) from highly purified human NK and small resting T cell progenitors. As accessory cells we used autologous, as well as allogeneic, monocytes, and dendritic cell enriched cells. Both NK and T cells were able to generate LAK activity, but their activation requirements were different. NK cells were activated merely by IL-2, and accessory cells did not enhance their lytic activity in the presence or absence of IL-2. Conversely, T cells were practically unresponsive to even high concentrations of IL-2 having a strict requirement for accessory cells for the development of lytic activity and proliferation. Accessory cells differed in their ability to activate T cells presumably depending on their ability to induce IL-2 synthesis, allogeneic dendritic cells being the most effective accessory cells and IL-2 synthesis stimulators. Allogeneic accessory cells could induce lytic activity in T cells even in the absence of exogenous IL-2. Thus, accessory cells play a central role in expanding the LAK effector cell population.     

Interleukin-6 enhances the induction of human lymphokine-activated killer cells

Summary Human peripheral blood mononuclear cells develop a powerful lytic capacity when cultured in vitro with interleukin-2 (IL-2), becoming lymphokine-activated killer cells (LAK cells). As part of an investigation into means of influencing this process, the effect of other cytokines has been examined. In this study we describe the ability of interleukin-6 (IL-6) to regulate the induction and function of human LAK cells. The results show that substitution of IL-6 for IL-2 did not lead to the development of functional LAK cells, nor was IL-6 able to alter the lytic capacity of established LAK cells. However, when IL-6 was included with IL-2 during the induction phase of the LAK cells, the resulting cells displayed considerably greater lytic activity than those prepared with IL-2 alone. This effect was IL-6 dose-related. These results indicate that LAK cell development may be positively regulated in vitro; the implications of this observation for the clinical usage of LAK cells are discussed.     

A novel negative regulator molecule, Cho-1, is involved in the cytotoxicity by human natural killer cells but not in cytotoxic T lymphocytes.

We previously reported the cytotoxic negative regulatory molecule, Cho-1, that was expressed on the cell surface of rat fetal fibroblast cells in the cytotoxicity by natural killer (NK) cells. This molecule was IFN-gamma-inducible, but appeared to be different from MHC class I. It was expressed on NK-resistant cells but not on NK-sensitive murine target cells such as YAC-1. In this paper, first we determined whether Cho-1 could also act as the negative regulatory molecule in a human NK-resistant HEPM line. Our data strongly suggested that Cho-1 could act as such a negative regulatory molecule in human NK cytotoxicity. The immunoprecipitates made with HEPM cell lysate and anti-MHC class I monoclonal antibody (mAb) did not react against anti-Cho-1 mAb, indicating that Cho-I was different from MHC class I. Second, an assessment was made as to whether or not this molecule is involved in the cytotoxicity of CD8 (+) cytotoxic T lymphocytes (CTL) against human autologous tumor cells. The data indicated that although this cell surface molecule was expressed on certain tumor lines, it was not involved in the cytotoxic mechanism of CTL. Thus, Cho-1 appeared to be the novel regulatory molecule in the NK cytotoxic mechanism.     

Effect of interleukin 2 on fetal thymocytes in organ cultures: generation of lymphokine-activated killer cells

Cells with cytolytic activity can be detected in mouse fetal thymic lobes cultured in the presence of interleukin 2 for 6 days. The lymphokine-activated killer cells from 14-day fetal thymic lobes are relatively resistant to treatment with anti-Ly-2 antibody and complement (CD8-) but sensitive to anti-Thy-1 and complement treatment (Thy-1+). They display major histocompatibility complex-unrestricted killing, lysing both syngeneic and allogeneic tumor cells, but will not lyse human xenogenic target cells. Low levels of cytotoxic activity can be detected in thymic lobes from Day 12-13 embryos and this activity increases with embryonic age. While the events which lead to the inhibition of normal maturation of fetal thymocytes by inclusion of IL-2 in fetal thymus organ cultures are unknown, the appearance of cytotoxic cells raises the question of whether they are involved in the normal intrathymic cell death process.     

Human neuroblastoma cell lines are susceptible to lysis by natural killer cells but not by cytotoxic T lymphocytes

The susceptibility of human neuroblastoma cells to direct cellular cytotoxicity has not been previously established. This is of particular interest because of their aggressive growth and low HLA expression. Neuroblastoma lines CHP 100 and CHP 126 were found to be excellent targets in 4-hr CML assays. Natural killer (NK) cells from fresh PBL and from an NK clone, 3.3, have high lytic activity against both cell lines. We also studied mixed lymphocyte culture-generated cytotoxic lines containing allo-specific cytotoxic T lymphocytes (CTL) directed against HLA antigens present on the neuroblastoma target cell lines. These lines did show excellent lytic activity, but cold target competition studies indicated that all of the lysis resulted from NK activity. This was verified by using inhibition studies with the use of monoclonal antibodies. OKT 3 and anti-HLA antibodies that block CTL function caused no reduction in kill. In contrast, anti-lymphocyte function antigen-1 (anti-LFA-1), which blocks both NK and CTL function, significantly inhibited lysis. These results serve as a functional confirmation of earlier findings of a very weak expression of HLA-A,B,C and beta 2-microglobulin on neuroblastoma cells.     

Monocyte-dependent,serum-borne suppressor of induction of lymphokine-activated killer cells in lymphocytes from melanoma patients

Summary Both phytohemagglutinin-induced cytotoxicity and recombinant-interleukin-2 (rIL-2)-induced lymphokine-activated killer (LAK) activity against noncultured melanoma cells were significantly reduced when peripheral blood mononuclear cells (PBMC) from patients with metastatic melanoma were incubated in RPMI medium 1640 and 10% autologous human serum instead of 10% fetal calf serum, while serum from either healthy donors or patients with primary melanoma did not affect the level of cytotoxicity. The serum-mediated suppression was not restricted by major histocompatibility complex and was time-dependent. Addition of 10% human serum from the patients with metastatic melanoma [HS-Pt(m)] to the culture of PBMC with rIL-2 at the same time or 1 day after incubation significantly inhibited LAK activity. However, addition of 10% HS-Pt(m) 2 or 3 days after incubation did not inhibit LAK activity. Incubation of PBMC for 2 h with a high dose (10⁴ U/ml) of rIL-2 in the presence of 10% HS-Pt(m), followed by incubation in the absence of either rIL-2 or HS-Pt(m), did not affect LAK cell activity. These results suggest that HS-Pt(m) inhibits the early stage of LAK cell differentiation, rather than the binding of rIL-2 to PBMC or a later stage in the differentiation. In contrast to PBMC, monocyte-depleted peripheral blood lymphocytes exhibited comparable levels of LAK activity when cultured with rIL-2 either in 10% fetal calf serum, 10% human serum from healthy donors or 10% HS-Pt(m). Addition of purified autologous monocytes to the culture of monocyte-depleted peripheral blood lymphocytes with rIL-2 suppressed LAK cell induction when 10% HS-Pt(m) was present. Thus serum-mediated suppression of LAK cell induction is largely dependent on the presence of monocytes, which may produce a secondary inhibitor that acts on lymphocytes. Addition of indomethacin to the culture did not reverse this monocyte-dependent serum-mediated suppression in a majority of cases, suggesting that prostaglandin E₂ does not have a major role in the suppression.This work was supported in part by NIH grant RR5511-25 and grants from The Council for Tobacco Research USA Inc., The Meadows Foundation, the Erwin Zaban Melanoma Research Foundation, and the Gillson-Longenbaugh Foundation     

Ultrastructural and functional studies of the interaction between IL-12 and IL-2 for the generation of lymphokine-activated killer cells

IL-12 promotes generation of LAK activity in short-term-cultured NK cells, but information on the structure and function of IL-12-induced LAK cells is not yet available. The latter issues have been here investigated with emphasis on interactions between IL-12 and IL-2. Peripheral blood mononuclear cells (MNC) exposed to IL-12 for 5-7 days displayed a decrease in the amount and density of the matrix of large granular lymphocyte (LGL)-associated granules. In cells cultured with IL-12 and IL-2 for 5-7 days, empty vacuoles were predominant and the electron-dense matrix was scanty. In MNC incubated with IL-2 for 5-7 days, most granules were loaded with electron-dense matrix. IL-12 and IL-2 displayed an additive effect on LAK cell cytotoxicity until approximately 48 h in culture which was followed by a sharp decline. Immunocytochemical and biochemical studies demonstrated that MNC cultured for 5-7 days with IL-12 and IL-2 displayed downregulated perforin expression and upregulated granzyme B expression. Fas ligand expression was virtually undetectable in MNC cultured for 5-7 days with or without cytokines. It appears that perforin downregulation plays a major role in the reduced cytotoxicity of MNC cultured with IL-12 and IL-2 for 5-7 days.     

IL-4 regulation of murine lymphokine-activated killer activity in vitro. Effects on the IL-2-induced expansion, cytotoxicity, and phenotype of lymphokine-activated killer effectors

The in vitro incubation of B6 splenocytes with purified, mouse rIL-4 for 4 to 5 days was sufficient to generate lymphokine-activated killer (LAK) activity. In addition, rIL-4 augmented LAK cytotoxic activity when combined with rIL-2, as measured in a 4 h 51Cr-release assay against fresh, syngeneic MCA-sarcoma (MCA-102 and MCA-105) cells. Interestingly, this augmentation was not observed against the cultured YAC-1 target. LAK generation and augmentation of cytotoxicity by rIL-4 was species-specific, because human rIL-4 (up to 20,000 U/ml) failed to elicit these effects in the mouse splenocyte cultures. When 5-day B6 LAK cells (splenocytes incubated in rIL-2 at 1000 U/ml for 5 days) were split and recultured in the combination of rIL-2 plus rIL-4 for 4 additional days at least a twofold greater expansion in cell number resulted compared to similar cells cultured in either rIL-2 or rIL-4 alone. Moreover, LAK cells expanded in rIL-2 plus rIL-4 exhibited substantial increases in in vitro cytolytic activity (on a per cell basis) against MCA-102 and MCA-105 sarcoma cells, but not against YAC-1 targets. FACS analysis or negative selection using Lyt-2 or NK-1.1 mAb plus C revealed no differences in effector phenotype(s) of LAK cells expanded in rIL-2 alone compared to rIL-2 plus rIL-4 to account for the differences observed in both expansion and cytolytic activity by rIL-4. The majority of cells was Thy-1+, Lyt-2+, T3+, and ASGM-1+. However, a marked increase in the granule-associated serine esterase, BLT-E, was found only in LAK cells expanded in the combination of both lymphokines. Collectively, these studies show that rIL-4 has potent regulatory activities on splenic LAK generation, expansion, and cytotoxic function in the mouse.     

IL-6/BSF-2 functions as a killer helper factor in the in vitro induction of cytotoxic T cells

rIL-6/B cell stimulatory factor 2 was found to augment CTL generation from mature as well as immature human T cells stimulated with UV-treated allogeneic cells. rIL-6 also acted on peanut agglutinin-positive murine thymocytes and Lyt-2-positive splenic T cells to give rise to CTL. rIL-6 alone could not induce CTL generation, the presence of IL-2 during the early phase of culture period was found to be essential for the IL-6 activity in the induction of CTL. The effect of rIL-6 was not mediated by the induction of IL-2 inasmuch as rIL-6 did not augment IL-2 production in MLC and anti-IL-2 antibody could not neutralize IL-6 activity. rIL-6 augmented CTL generation even when added 72 h after the initiation of cultures. The enhancing activity of rIL-6 could be neutralized with anti-IL-6 antibodies even when added 72 h after the initiation of cultures. The present data indicate that IL-6 acts in the late phase of CTL generation.