The survival and proliferation of CTL during the effector phase of the immune response is critical for the elimination of infectious agents and tumor cells. We report here that in an in vitro model system, the expansion and cytolytic function of tumor-reactive human CTL can be enhanced by CD4+ helper T lymphocytes through costimulatory signals that are mediated by cell surface molecules. The results presented here suggest that costimulatory receptors on CTL such as CD27, CD134 (4–1BB), and MHC class II are capable of directly interacting with the corresponding ligands on T-helper lymphocytes resulting in enhanced proliferation and survival of the CTL during the effector phase of antitumor immune responses. These findings underline the importance of antigen-specific helper T lymphocytes for the regulation and maintenance of CTL immunity, and have implications for the design of therapeutic vaccines for cancer.
Many issues remain unsolved regarding the possible mechanisms that regulate the longevity of cell-mediated T-cell responses during viral infections and antitumor immunity. Effector CTLs exert their cytolytic activity toward viral-infected or transformed target cells through the recognition of a foreign viral or tumor antigen in the form of MHC peptide complexes. Thus, disease progression is limited in an antigen-specific fashion. Because effector CTLs are terminally differentiated cells, they may have a limited capacity to propagate after TCR3 stimulation by antigen, even in the presence of growth-promoting lymphokines such as IL-2. Moreover, it is often that CTLs undergo AICD as a consequence of antigen stimulation, which may be enhanced by the IL-2 (1, 2, 3) . Despite this limitation, most viral infections are successfully cleared by effector CTLs, suggesting that under critical and perhaps life threatening circumstances, these cells may overcome AICD and are capable of proliferating to eliminate all of the virus-producing cells. Unfortunately, this does not appear to be the case in most antitumor immune responses, where early demise of CTLs and/or their lack of expansion at the tumor site may result in the unsuccessful elimination of disease.
HTLs are known to play an important role during the induction of new CTL responses, which usually takes place in secondary lymphoid organs. HTLs exercise this function by activating dendritic cells to present antigen to CTL precursors and through the production of lymphokines (e.g., IL-2) that stimulate CTL growth and induce their differentiation into effector CTL (4, 5, 6, 7, 8, 9) . Although the exact mechanism remains to be elucidated, it has been inferred that antigen-specific HTLs may also participate in the continuation of CTL responses in peripheral tissues (10, 11, 12, 13) . The persistence of adoptively transferred CTLs has been correlated with the presence of antigen-specific HTLs (14) , which may promote their survival and expansion in vivo through the production of cytokines that stimulate CTL growth and prevent AICD. Because it has been implied that HTLs may enhance CTL expansion through the production of IL-2, several clinical studies such as CTL peptide epitope vaccines and adoptive CTL immunotherapy are followed by the systemic administration of this lymphokine (15, 16, 17) . However, as exemplified here, it is likely that HTLs need to provide additional costimulatory signals to CTLs, some of which require direct cell-to-cell interaction, allowing the survival and expansion of CTLs during the effector phase of the immune response.
In general, human antigen-specific CTLs are difficult to propagate and maintain for long periods of time in tissue culture when stimulated with the appropriate APCs, regardless of the presence of IL-2 or other lymphokines. Nevertheless, some in vitro observations indicate that effector CTLs may have the capacity to overcome AICD and are capable of expanding after TCR stimulation if the appropriate costimulatory signals are provided (18, 19, 20, 21, 22, 23) . For example, an effective CTL-expansion procedure described by Riddell et al. (24 , 25) has allowed the production of large numbers of cells suitable for adoptive immunotherapy. This method uses large numbers of “feeder cells,” a TCR-stimulatory signal (anti-CD3 antibody) and IL-2, permitting the CTL to expand up to 1000-fold in ∼2 weeks. We have reported previously that antigen-specific stimulation (peptide-pulsed APC) can substitute for the use of anti-CD3 antibody in CTL expansion procedure (26) . However, in addition to the APC presenting the peptide, the presence of feeder cells was also required indicating that these cells play a necessary role for CTL expansion. Results presented herein indicate that the CD4+ T-lymphocyte component of the feeder cells plays a critical role in the TCR-mediated expansion of CTLs. Furthermore, direct contact between CTLs and HTLs through the interaction of known costimulatory molecules appears to provide costimulatory signals to the CTLs that enable these cells to proliferate while interacting with antigen-bearing tumor cells, enhancing their overall cytolytic activity. From these studies we conclude that effective immunotherapy for cancer will require the stimulation of both tumor antigen-specific CTLs and HTLs.
MATERIALS AND METHODS
Antigens and Antibodies.
The peptide corresponding to the HLA-A2-restricted epitope of gp100209–217 (ITDQVPFSV), derived from the sequence of the gp100 melanoma-associated antigen and the HLA-DR promiscuous peptide Pan DR Epitope (AKXVAAWTLKAAA; Ref. 27 ) were prepared by solid-phase synthesis and purified as described (28) . The following antibodies were purchased from PharMingen BD (San Diego, CA): anti-CD27 (clone M-T271), anti-CD28 (clone CD28.2), anti-CD30 (clone BerH8), anti-CD134 (OX40, clone ACT35), anti-CD137 (4–1BB, clone 4B4–1), anti-CD152 (CTLA-4, clone BNI3), normal mouse IgG1, and anti-MHC class II molecules (anti-DR+DP+DQ, clone TÜ39). Pharmaceutical grade anti-CD3 monoclonal antibody (OKT3, Orthoclone; Ortho Biotech, Inc., Raritan, NJ) was used to activate and induce T-cell proliferation.
Cells and Cell Lines.
The T2 cell line (homozygous for HLA-A2 and defective in endogenous antigen processing) was used as the target for CTL-mediated cytolysis to demonstrate peptide reactivity. Melanoma cell lines 624mel (HLA-A2+ and gp100+) and 888mel (HLA-A2- and gp100+) were provided by Dr. S. Rosenberg (National Cancer Institute, NIH, Bethesda, MD). All of the tumor cell lines were maintained in tissue culture using RPMI 1640 supplemented with 10% FBS (V/V), l-glutamine, nonessential amino acids, sodium pyruvate, and gentamicin (complete RPMI). All of the culture materials were purchased from Life Technologies, Inc. (Rockville, MD). The melanoma cells were treated with 100 units/ml IFN-γ for 48 h to increase the level of MHC class I expression before performing the CTL cytotoxicity assays. In vitro generation of tumor-reactive CTL lines and clones was performed using peptide-pulsed dendritic cells as described (28) . The CTLs were maintained in tissue culture by weekly stimulation with peptide-pulsed irradiated autologous adherent monocytes as APCs in the presence of 50 IU/ml IL-2 (28) . An antigen-specific HTL clone was prepared by weekly stimulation of purified CD4+ T lymphocytes with autologous irradiated APC and 10 μg/ml of PADRE peptide in the presence of IL-2 (50 IU/ml). After three to four rounds of stimulation, T-cell clones were isolated using limiting dilution. Purified CD4+, CD8+, and CD14+ cells were prepared from PBMCs from HLA-typed normal volunteers using immunomagnetic separation (Mitlenyi Biotec, Auburn, CA). The Institutional Review Board on Human Subjects (Mayo Foundation) approved this research and informed consent for blood donation was obtained from all of the volunteers.
CTL Expansion Assays.
The CTL expansion assays are based on the protocol as described by Riddell et al. (24 , 25) except that for the present studies we omitted the use of EBV-transformed lymphoblasts as part of the feeder cells. Expansion assays were performed either in upright T-25 culture flasks or in six-well tissue culture plates. Tumor reactive CTLs (5 × 104/T-25 flask or per well) were incubated with 2.5 × 107 irradiated (3300 rads) PBMCs, pooled from two to three normal donors (not HLA-matched), in the presence of 30 ng/ml OKT3. Beginning 24 h after the initiation of the assays, 50 IU/ml of IL-2 were added to the cultures every 2–3 days. For subsequent experiments, the basic assay was altered in several ways. The PBMC feeder cell population was modified by depleting (or adding) CD4+ or CD8+ T cells, which were prepared by immunomagnetic separation as described above. To determine the effect of direct cell contact on expansion, assays were performed in six-well plates with Transwell semipermeable membranes (Costar No. 3412; 0.4 μm pore size). In experiments where OKT3 was not used to provide antigenic stimulation, gp100 peptide pulsed irradiated (3300 rads) monocytes (HLA-A2+ CD14+ cells) or irradiated (6600 rads) 624mel cells were substituted as the source of antigen. In these experiments, unpulsed monocytes or 888mel cells were used as negative controls. In some experiments, recombinant IL-4 was used at 10 ng/ml and was added to the assays on day 1 and then every 2–3 days. On day 14, the cells from the CTL expansion assays were harvested, the total number of viable cells was determined, and CTL cytotoxicity assays were performed. Antibody blocking experiments were performed by the addition of 20 μg/ml of monoclonal anti-CD27, anti-CD137, anti-MHC class II antibodies, or normal mouse IgG to the cultures on days 0 and 3, and the cells were harvested on day 7 to estimate the number of viable CTLs.
Analysis of CTL Proliferation by CFSE Fluorescence Staining.
CTLs were washed once and resuspended at 2 × 107 cells/ml in serum-free RPMI 1640. An equal volume of 2 μm CFSE (Molecular Probes, Inc., Eugene, OR) was added to the cells, and they were mixed gently and incubated at 37°C for 10 min. After the incubation period, an equal volume of FBS was added to quench the reaction, and the cells were washed twice in medium containing 10% FBS. CFSE-labeled CTLs were incubated in 24-well tissue culture plates, at 5 × 104 cells/well with 5 × 105 melanoma tumor cells, with or without PADRE peptide-specific HTLs (1 × 105 cells/well) all in the presence of IL-2 (50 IU/ml). Where indicated, PADRE peptide was added at 3 μg/ml to activate the HTL clone. Cultures were fed with fresh medium containing IL-2 (50 IU/ml) two to three times per week. After 14 days in culture, the cells were harvested, washed, counted, stained with a phycoerythrin-labeled anti-CD8 monoclonal antibody (PharMingen). The cells were analyzed by flow cytofluorometry, and the percentage of undivided CTLs was estimated by counting the numbers of viable CD8-positive cells staining positive for CFSE on a gate established with nonstimulated CTLs.
CTL Cytotoxicity Assays.
Cytotoxic activity of CTLs was determined in a standard 4–6 h 51Cr release assay as described (28) . Peptide-pulsed T2 targets were prepared by incubating the cells with 10 μg/ml of peptide at 37 C° overnight. Adherent melanoma tumor cells were removed from the culture flask with trypsin-EDTA right before labeling. Target cells were labeled with 300 μCi 51Cr sodium chromate (Amersham Pharmacia Biotech, Piscataway, NJ) for 1–2 h at 37 C° in a water bath. Various numbers of effector cells were mixed with 2 × 104-labeled targets at different E:T ratios in 96-round-bottomed well plates at a final volume of 0.2 ml. After 4–6 h incubation at 37 C°, 30 μl of supernatant were collected from each well and the percentage of specific lysis was determined according to the formula: [(cpm of the test sample − cpm of spontaneous release)/(cpm of the maximal release − cpm of spontaneous release)] × 100. Overall lytic activity was calculated by dividing the total number of cells obtained in each culture by the LU50. Thus, the overall lytic activity values take into account both the cytolytic activity of the cultures and the number of cells recovered from the culture (i.e., the higher the overall lytic activity value, the higher cytotoxic potential of the cultures).
CTL Costimulation Assays.
Round-bottomed 96-well tissue culture plates were coated with a low concentration of TCR-stimulatory monoclonal anti-CD3 antibody (50 μl of OKT3 at 1 ng/ml in PBS) for 16 h. at 4 C°. The plates were washed twice with PBS, and the candidate costimulatory antibodies were coated at 10 μg/ml in PBS (50 μl/well) for 4 h at 37 C°, after which the plates were washed four times with PBS and once with culture medium. A total of 2 × 104 CTLs were added to each well (in triplicates), and the assays were incubated for 72 h at 37 C°. To measure costimulation, 0.5 μCi of 3[H] thymidine were added during the last 18 h of the culture, and the amount of radioactivity incorporated into DNA was determined using a scintillation counter after harvesting the cultures onto glass fiber filters.
HTL Effect on Anti-CD3-induced CTL Expansion.
We first examined the role of HTL in CTL expansion induced by anti-CD3 antibodies using a melanoma-specific HLA-A2-restricted human CTL clone that recognizes a peptide derived from tumor antigen gp100 (28) . CTLs stimulated with anti-CD3 antibodies in the presence of large numbers of feeder cells (irradiated PBMCs from several unrelated donors) expanded ∼800-fold in a 2-week period and displayed high levels of lytic activity (Fig. 1)⇓ . Interestingly, depletion of HTLs (CD4+ T cells) from the feeder cells resulted in a significant decrease of both expansion and cytolytic activity of the CTLs, even in the presence of IL-2 (Fig. 1)⇓ . Nevertheless, addition of purified (irradiated) CD4+ T cells to the HTL-depleted PBMCs restored the capacity of the feeder cells to stimulate the expansion and amplify the lytic activity of the CTLs. Removal of CD8+ T cells from the feeder cells increased the total cell number (the fold expansion of the CTLs) by ∼40%. However, the enhanced CTL expansion was accompanied by a substantial reduction in the lytic activity of the CTLs, reflected by an increase in the number of effector cells required to kill 50% of the targets (LU50). Thus, by eliminating the CD8 T cells from the feeder cells, the quality of these CTLs was diminished. No significant proliferation (or increased cytolytic activity) was detected if the CTL expansion cultures lacked either the TCR-stimulatory anti-CD3 antibodies or IL-2 (data not shown). It should be pointed out that depletion of CD4+ cells from the PBMCs did not affect significantly the numbers of monocytes (which express low levels of CD4 and are required for anti-CD3 activation of the CTLs), because the numbers of CD14+ cells remained unaltered (data not shown). Thus, it appears that CD4+ HTLs mediate the expansion and help maintain the cytolytic activity of CTLs after TCR stimulation.
Effect of Direct Cellular Interactions between CTL and HTL during Expansion.
The capacity of HTLs to mediate CTL expansion induced by TCR stimulation could be induced by soluble factors (i.e., lymphokines) and/or costimulation via direct cell-to-cell contact. To evaluate the requirement of direct cellular interactions between CTLs and HTLs, expansion cultures were performed in Transwell tissue culture plates, which separated the CD4+ T cells from the CTL by a semipermeable membrane (0.4 μm pore size), allowing the diffusion of lymphokines. In these experiments the addition of irradiated CD4+ T cells in direct contact with the CTLs resulted in a 1600-fold increase in the CTL numbers when anti-CD3 antibodies were used to trigger the cell proliferation (Fig. 2a⇓ , condition 2). In contrast, the absence of CD4+ T cells or the physical separation of these cells from the CTLs by the Transwell membrane resulted in a significant decreased expansion of the effector cells (Fig. 2a⇓ , conditions 1 and 3, respectively). To ensure that the HTLs were activated, anti-CD3 antibodies, APCs, and IL-2 were added to both sides of the membrane. Similar results were obtained when antigen (peptide-pulsed APCs) was used to stimulate the CTLs. In these studies, direct contact of the CD4+ T cells with the CTLs resulted in significantly higher CTL proliferation as compared with the absence of CD4+ T cells or with the addition of CD4+ T cells to the opposing side of the Transwell membrane (Fig. 2b)⇓ . The proliferation of the CTLs required antigen as seen by the lack of expansion in the absence of the peptide (Fig. 2b⇓ , black bars). The proliferation induced by peptide-pulsed APCs was approximately one-tenth of that obtained with anti-CD3 antibodies suggesting that activation of the CD4+ T cells may be important for facilitating the expansion of CTLs. Alternatively, anti-CD3 antibodies could be more effective than peptide-pulsed APCs in stimulating the TCRs and inducing the proliferative response of the CTLs. Altogether, these results indicate that HTLs are capable of mediating TCR-induced expansion of CTLs through a mechanism that involves direct contact between these cells.
HTL Effect on Tumor-induced CTL Expansion.
To directly assess the role of HTLs during the recognition of tumor cells by CTLs, melanoma cells were used in the expansion cultures instead of anti-CD3 antibodies or peptide-pulsed APCs. In these cultures, low numbers of CTLs were incubated with higher numbers (10-fold excess) of gp100+ melanoma cells both in the presence and absence of irradiated CD4+ T cells in medium containing IL-2. After 8 days, large numbers of adherent melanoma cells remained in the cultures that did not contain HTLs, and only a few surrounding CTLs could be observed (Fig. 3⇓ , condition I). In contrast, most of the adherent melanoma cells appeared to be absent in the cultures that contained CTLs with the irradiated HTLs (Fig. 3⇓ , condition II). At the same time, CTLs were clearly evident in these cultures indicating that these effector cells were able to proliferate and destroy the majority of the tumor cells. Interestingly, the addition of IL-4 together with IL-2 dramatically enhanced the expansion of the CTLs (Fig. 3⇓ , condition III). The potentiating activity of IL-4 could be because of the reported effects of this lymphokine in promoting CTL proliferation and survival (29, 30, 31) . Alternatively, because no mitogens (anti-CD3) were included in these cultures, IL-4 could be facilitating the expansion of the CTLs indirectly via an enhanced activation of the HTLs. Although the HTLs were irradiated to prevent the cells from dividing, these cells are still capable of responding to TCRs and cytokine signals by secreting lymphokines (data not presented). Without the addition of HTLs, the CTLs failed to expand and eliminate the adherent melanoma cells, even in the presence of IL-2 and IL-4 (Fig. 3⇓ , condition IV), indicating that other HTL-derived costimulatory signals besides these lymphokines play a role in this process. Most significantly, if the melanoma cells did not express the peptide/MHC complex recognized by the CTLs, the effector cells failed to proliferate and eliminate the adherent tumor cells, even in the presence of HTLs and lymphokines (Fig. 3⇓ , condition V). Lastly, even with the addition of lymphokines, the irradiated HTLs by themselves did not have any effect in the elimination of the melanoma cells (Fig. 3⇓ , condition VI). These results clearly illustrate that HTLs promote the antigen-driven expansion of CTLs, facilitating the elimination of the tumor cells.
To evaluate the CTL activity, these cultures were continued for a total of 14 days when the cells were harvested, counted, and tested for antigen-specific effector function in a 4-h 51Cr-release cytotoxicity assay against relevant and irrelevant target cells. High levels of antigen-specific cytotoxicity toward peptide-pulsed target cells were observed in all of the CTL cultures that were originally stimulated with the antigen-expressing melanoma cells (Fig. 4a)⇓ . On the other hand, the CTLs that were stimulated with melanoma cells lacking appropriate expression of antigen (HLA-A2 negative) but incubated in the presence of HTLs together with IL-2 and IL-4, exhibited a lower level of cytotoxicity (Fig. 4a⇓ , ♦). These findings demonstrate the requirement of antigen during the effector/expansion phase for the maintenance of cytotoxicity. As expected, the cultures that contained only irradiated HTLs and melanoma cells (Fig. 3⇓ , condition VI) did not yield viable cells, so they were not tested for cytotoxicity. In all of the cases, CTL-mediated lysis was antigen-specific. Even at high E:T ratios the CTLs failed to kill target cells not expressing the corresponding peptide epitope (<5%; data not shown). The CTL cultures that expanded well (conditions II and III from Fig. 3⇓ ) were also tested for their capacity to recognize and lyse melanoma cells that expressed the relevant HLA-A2-restricted antigen (gp100). As shown in Fig. 4b⇓ , the CTLs derived from cultures that were stimulated with the antigenic melanoma cells in the presence of HTLs, were also very effective in killing melanoma target cells expressing the corresponding peptide/MHC complexes. Moreover, these CTLs fail to kill melanoma cells that do not express the restricting HLA allele (data not shown). Taken together, these findings indicate that HTLs are required for CTL expansion induced by antigen-bearing tumor cells and that antigen-specificity and antitumor reactivity were adequately maintained during the 2-week expansion cultures.
Effect of HTLs on Overall Effector Activity.
Although ostensibly, the levels of cytotoxicity obtained in all of the cultures stimulated with the HLA-A2+ melanoma appear to be similar (Fig. 4, a and b)⇓ , the CTLs proliferated extensively only in those cultures that contained the irradiated HTLs (Fig. 4c⇓ , ▪ in conditions II and III). Therefore, when taking into consideration the degree of expansion of the CTLs in addition to the level of cytotoxicity, the presence of HTLs resulted in an increased overall effector activity as measured by the total number of lytic units per culture (Fig. 4c⇓ , ▨). Moreover, the use of IL-4 in addition to IL-2 additionally increased the expansion and overall lytic activity of the CTLs by ∼3-fold (Fig. 4c)⇓ . These results support our hypothesis that the presence of HTLs during the effector phase of antitumor immune responses increases the capacity of tumor-reactive CTLs to proliferate and maintain their effector function.
Antigen-specific HTLs Enhance TCR-induced CTL Expansion.
The experiments presented above used CD4+ cells that were purified using anti-CD4 immunomagnetic bead separation. Although approximately 95–98% of these cells represent T lymphocytes, the remaining contaminating cells could be responsible for providing the CTLs the costimulatory signals required for their expansion. Thus, we proceeded to evaluate the capacity of a CD4+, antigen-specific T-cell clone (that was maintained in tissue culture for more than a month in the absence of APCs) to enhance the proliferation of CTLs during their encounter with tumor cells. For these experiments, the gp100-specific CTL clone was labeled with the fluorescent dye CFSE to estimate the percentage of cells that undergo cell division in tissue culture under various experimental conditions. An HTL clone specific for the MHC class II promiscuous “PADRE” (AKXVAAWTLKAAA) peptide (27) was generated from the same donor and was expanded in culture in the absence of APCs for > 30 days to ensure that any costimulatory activity was derived from the CD4+ T cells and not from other accessory cells. In these experiments both the melanoma tumor cells and the HTLs were not irradiated, and the initial CTL:HTL:melanoma ratio was 1:2:10. In addition, all of the cultures contained an optimal concentration of IL-2 (50 IU/ml, added 2–3 times/week). After 2 weeks the cultures were harvested and analyzed by flow cytometry for the presence of remaining live tumor cells (Fig. 5⇓ , left panels) and for the percentage of CTLs that did not proliferate during this time period (Fig. 5⇓ , right panels). Observation of the cultures under an inverted microscope revealed similar results as those presented in Fig. 3⇓ , but in this experiment, stimulation of the HTLs by the addition of the PADRE peptide was necessary for the elimination of the majority of the melanoma cells, indicating that activation of the HTLs is required for the generation of optimal costimulatory function. Because both the CTL and HTL clones used in these experiments express surface MHC class II molecules (data not shown), antigen stimulation of the HTLs by the peptide could take place in the absence of additional APCs. Antigen presentation to the HTLs by the melanoma cells was unlikely, because these cells do not share the MHC class II alleles with the T cells and do not express surface MHC class II molecules (data not presented). The results of this experiment indicate that high numbers of live melanoma tumor cells (i.e., number of events in gate R2 of the scatter plots) remained in the cultures that contained CTLs alone (Fig. 5c⇓ , left panel). In contrast, in the presence of HTLs and the HTL-activating PADRE peptide, a close to 90% decrease of the number of viable tumor cells was observed (Fig. 5e⇓ , left panel). The reduction in the number of tumor cells was less evident in the cultures that contained HTLs in the absence of the PADRE peptide (Fig. 5d⇓ , left panel). As expected, the scatter plots of control cultures not containing melanoma cells did not exhibit any significant counts in gate R2 (Fig. 5, a and b⇓ , left panels). Importantly, there was no evident effect in the cell growth and numbers of live melanoma tumor cells when HTLs and PADRE were added to the cultures in the absence of CTLs, eliminating the possibility that lymphokines such as tumor necrosis factor produced by the peptide-stimulated HTLs had an antitumor effect (data not shown).
The percentage of nonproliferating (undivided) CTLs under the various experimental conditions was also assessed by flow cytometric analysis (Fig. 5⇓ , right panels). A control sample of CFSE-labeled CTLs was fixed and kept at 4 C° to set the gate for undivided cells (R3 in Fig. 5⇓ , right panels). The results of these experiments indicate that the CTLs divided more effectively (reduction of CFSE staining) when both HTLs and PADRE peptide were added to the culture (Fig. 5e⇓ , right panel). The CTLs divided less well when they were incubated with melanoma tumors alone or in the presence of HTLs but without the PADRE peptide (right panels of Fig. 5, c and d⇓ , respectively). In contrast, CTLs did not divide in the absence of melanoma cells (Fig. 5a⇓ , right panel) even if HTLs, PADRE peptide, and IL-2 were present (Fig. 5b⇓ , right panel), indicating that TCR stimulation of CTLs is required to induce cell division. It is important to note that loss of CFSE did not necessarily correlate with the cell numbers of CTLs that were obtained after the 2-week period. An increase in final CTL numbers (∼10-fold) was only observed in the presence of melanoma, HTLs, and PADRE, suggesting the possibility that many of the CTLs that proliferated to the melanoma antigens in the absence of activated HTLs could have undergone AICD.
Identification of Costimulatory Molecules That Mediate HTL Effect on CTL Expansion.
The data thus far presented suggest that HTLs are capable of providing costimulatory signals, possibly through cell surface molecules to CTLs, resulting in their expansion and increased overall cytolytic function. Several known costimulatory molecules are expressed on CTLs, and in many cases their corresponding ligands can be found on HTLs. For example, CD28 is constitutively expressed in most T cells, and its ligands, CD80 and CD86, have been observed on activated HTLs (20 , 21) . Similarly, other costimulatory molecules such as CD27, CD30, CD134 (OX-40), CD137 (4–1BB), and CD152 (CTLA-4) may be expressed on activated CTLs, and in some circumstances are able to provide costimulatory and/or antiapoptotic signals (32, 33, 34, 35, 36, 37, 38) . The presence of the costimulatory molecules and their corresponding ligands was examined on activated antigen-specific effector CTLs and HTLs, respectively. Cytofluorometric analysis was performed with fluorescent-labeled antibodies specific for these molecules. Activated CTLs, incubated overnight in anti-CD3 antibody-coated culture plates, expressed various levels of all of these costimulatory molecules mentioned above with the exception of CD28 (data not shown). The lack of expression of CD28 on our melanoma-specific CTLs was not surprising, because it has been reported that this molecule is decreased or absent in long-term cultured effector CTLs (39, 40, 41) . In agreement with these reports we have observed that most of the antigen-specific CTL lines and clones that are maintained for long periods of time in tissue culture cease to express CD28.4 Nevertheless, it is possible that recently activated CTLs will still express CD28. Of note, activated HTLs expressed most of the corresponding ligands (CD80, CD86, CD27L, CD30L, and CD137L) with the exception of CD134L (data not shown). These findings indicate that activated HTLs have the potential of providing CTLs with activation signals via one or several of the known costimulatory molecules, potentially resulting in their proliferation and survival. However, it is obvious that CD28-negative CTLs will not be able to receive costimulation by HTLs via CD80/CD86 and would have to rely on other costimulatory pathways.
In addition to the above, the possibility exists that CTLs and HTLs may interact in an antigen-specific manner. It is well known that activated T cells including CTLs express MHC class II molecules (42, 43, 44) , which could potentially interact with the TCRs of HTLs. Interestingly, cross-linking of MHC class II molecules using antibodies provides costimulatory signals to activated T cells (45, 46, 47) . Thus, in addition to CD27, CD30, CD137, and CD152, MHC class II molecules could participate in the interaction occurring between CTLs and HTLs. To evaluate the potential role of these costimulatory molecules in CTL activation and proliferation, monoclonal antibodies specific for these molecules were tested for their ability to enhance the proliferative responses of effector CTLs to TCR activation. These assays used a suboptimal concentration of anti-CD3 antibody and an optimal concentration of an individual anticostimulatory antibody, both coated onto the culture plates. The immobilization of the antibodies to potential costimulatory molecules onto the plastic surface was carried out to maximize cross-linking these molecules on the CTL surface, which would increase the costimulatory signals. The results presented in Fig. 6⇓ , obtained using two different antigen-specific CTL clones, show that antibodies specific to MHC class II, CD27s and CD137 enhanced the proliferative response of these cells to TCR activation. On the other hand, antibodies specific for anti-CD28, anti-CD30, anti-CD134, and anti-CD152 did not increase the proliferation of the CTLs to the suboptimal concentration of anti-CD3 antibody. In subsequent separate experiments, monoclonal antibodies specific for the individual MHC class II molecules, HLA-DR, -DP, and -DQ, were all capable of costimulating the antigen-specific CTLs (data not presented). However, not all of the anti-MHC class II monoclonal antibodies that we have tested demonstrate equivalent costimulatory activity. These results indicate that activated CTLs express on their surface a variety of costimulatory receptors of which the ligands are present on HTLs and that the interaction between one or several of these paired costimulatory molecules could enhance the expansion of CTLs in the presence of antigen.
To directly evaluate the role of some of these costimulatory receptors in the antigen-induced expansion of CTLs mediated by activated HTLs, we tested the capacity of various antibodies to block this interaction. The use of soluble antibodies in these experiments should inhibit the capacity of the ligands expressed on the HTLs to cross-link the costimulatory molecules on the CTL. As shown in Fig. 7⇓ , antibodies to CD70 (CD27L), CD137 (4–1BB), and MHC class II inhibited to a great extent the expansion of CTLs observed in the presence of antigen (melanoma tumor cells), HTLs, and PADRE peptide. The mixture of anti-CD70 and anti-CD137 slightly increased (∼50%) the inhibition of CTL expansion (data not shown), indicating that either the antibodies cannot completely block the access of the receptor to the ligand or that additional costimulatory molecules may be involved in the interaction between CTLs and HTLs.
The results presented here demonstrate that HTLs mediate the expansion and increase the overall lytic activity of antigen-specific CTLs during the effector phase of the immune response. In the absence of CD4+ T cells, CTLs expand poorly and exhibit low overall lytic activity (Fig. 1)⇓ . Interestingly, removing CD8+ T cells from the feeder cells used in the CTL expansion cultures resulted in a significant increase in the yield of cell numbers, but the lytic activity of these CTLs was diminished (Fig. 1)⇓ . At present we do not know by what mechanism the CD8+ feeder cells exert these opposing effects on the CTLs. We can only speculate that the CD8+ feeder cells may produce factors, some that inhibit the proliferation of the CTLs and others that may enhance the quality of their effector function. For example, transforming growth factor β, which could be produced by some of the anti-CD3 antibody-stimulated CD8+ feeder cells, would inhibit CTL proliferation, and at the same time other lymphokines such as IFN-γ and tumor necrosis factor-α would increase the expression of adhesion molecules on the CTLs, enhancing the lytic activity by facilitating their binding to the target cells.
The enhancement in CTL expansion and activity by HTLs appears to occur through direct cell-to-cell interactions (Fig. 2)⇓ , possibly via one or more costimulatory molecules. Proliferation signals in CTLs are likely to be triggered in part by the appropriate stimulation/cross-linking of the TCR, which requires an optimal density of specific peptide MHC complexes on the APCs. However, because tumor cells express much lower amounts of class I MHC molecules than professional APCs, proliferative responses in CTLs may not take place during the effector phase of antitumor immune responses. Nevertheless, many costimulatory signals tend to lower the TCR signaling threshold allowing T-cell proliferative responses to take place at suboptimal concentrations of antigen (i.e., low peptide/MHC density on tumor cells). Our results show that cross-linking of some CTL surface molecules such as MHC class II, CD27, and CD137 (4–1BB) increases CTL proliferation in conditions of suboptimal TCR stimulation (Fig. 6)⇓ . Interestingly, the ligands for these costimulatory CTL surface molecules (HTL TCR, CD70, and 4–1BBL) are found on activated HTLs. Because our results show that CTL expansion and effector function are augmented by direct contact with HTLs, we propose that HTLs play a critical role during the effector phase of CTL responses by providing strong costimulatory responses that may be critical for the elimination of tumor cells. The blocking effect of anti-CD70 (CD27L) and anti-CD137 (4–1BB) in the antigen-induced CTL expansion mediated by HTLs (Fig. 7)⇓ provides supporting evidence that these costimulatory pathways are involved in this phenomenon.
Our results also indicate that antigen-activated HTLs are required for the proliferation and expansion of CTLs resulting from their encounter with antigen. It was interesting to observe that although CTLs divide in the presence of antigen (and IL-2; Fig. 5c⇓ ), in the absence of antigen-activated HTLs the total number of these cells does not increase, suggesting that AICD occurs in many of the dividing cells. Thus, it is possible that HTLs may inhibit AICD of CTLs via the cross-linking of surface CD27 and CD137 (4–1BB). Notably, both of these costimulatory receptors are known to inhibit AICD in T cells (48 , 49) . In the model system used here, activated HTLs were able to costimulate the CTLs in the absence of APCs. Nevertheless, it is obvious that APCs will play a critical role at the tumor site for the appropriate activation of HTLs via the cross-presentation of antigens that are either released/secreted by live tumor cells or are derived from dead (apoptotic or necrotic) tumor cells.
It was interesting to observe that the cross-linking MHC class II molecules on CTLs by antibody resulted in strong costimulatory signals (Fig. 6)⇓ . However, there are already some reports in the literature indicating that MHC class II molecules on human CD4+ T lymphocytes can function as signal transduction elements (45, 46, 47) . The possibility that HTLs may exert their potentiating function on CTLs in part by interacting through their TCRs with MHC class II molecules expressed on the CTLs is tantalizing. Nevertheless, the nature of peptides complexed onto MHC class II molecules on CTLs that could be recognized by HTLs remains enigmatic. In our model system, the expansion cultures used either HTLs from unrelated blood donors (which would recognize allogeneic determinants present on the MHC class II molecules of the CTLs) or a peptide-specific HTL clone that recognized exogenously added peptide. At present, we can only speculate how HTLs and CTLs could interact in vivo in an antigen-specific fashion. One possible scenario is that when CTLs interact with professional APCs through the recognition of antigenic peptides presented by MHC class I, some of the MHC class II molecules on the APCs may be “snatched away” by the CTLs. A number of these MHC class II molecules would likely bear peptides derived from antigens expressed by the tumor cells. Although at first sight this possibility appears to be far fetched, it has been demonstrated recently that T cells can acquire MHC class I molecules and other cell surface proteins from APCs during the process of antigen recognition (50 , 51) . Moreover, in mice and rats, most of the MHC class II molecules expressed on T lymphocytes are derived from APCs during antigen-TCR-mediated interactions (52, 53, 54) , and the presence of passively acquired MHC class II molecules in human T cells has also been documented (55) . We have also observed that activated CTLs can incorporate MHC class II molecules from APCs when coincubated in culture.4 Accordingly, during their interaction with APCs in vivo, CTLs may acquire MHC class II molecules bearing antigenic peptides. Some of these peptides could be derived from the same pathogen (tumor cell or microorganism) that is recognized in the context of MHC class I by the CTLs. As a consequence, the newly acquired peptide/MHC class II complexes on CTLs would enable these cells to communicate with pathogen-specific HTLs and receive the appropriate costimulatory signals. We propose the following hypothetical model to explain the possible interactions between effector CTLs and HTLs at the tumor site, which may be necessary for the complete elimination of the tumor mass: (a) tumor-infiltrating HTLs become activated when recognizing MHC/peptide complexes on APCs that are present at the tumor site; the peptides that that stimulate these HTLs are derived from antigens expressed by dead/apoptotic tumor cells that are ingested by APCs and processed into the corresponding peptide/MHC complexes; (b) alternatively, HTLs may recognize MHC/peptide complexes directly on CTLs, when the CTLs passively acquire these complexes from APCs; (c) in either case, once activated the HTLs express costimulatory ligands such as CD70 and 4–1BBL that provide direct costimulation to the CTLs allowing these cells to divide and avoid AICD; and (d) the physical proximity of activated HTLs with the CTLs also facilitates the efficient local delivery of soluble costimulatory factors such as IL-2. Clearly, additional experimental data will be needed to support this model of antigen-specific interaction between CTLs and HTLs. Nonetheless, it is clear that HTLs play a critical function in regulating CTLs not only during the initiation of immune responses but also during the effector phase of the response, and this role is likely to be more complex than the simple production of soluble factors such as IL-2.
Our findings could have an important consequence for tumor immunotherapy, because these results emphasize the importance of antigen-specific T-helper cells for the augmentation of antitumor CTL responses. Thus, vaccines designed to induce antitumor CTLs should also trigger tumor-specific HTL responses to improve the clinical efficacy of antitumor immunization.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported by R01CA80782, R01CA82677, and RR-00585 from the NIH and by a research award (to R. L. G.) from the American College of Obstetricians and Gynecologists/Searle (1999 Research Award in Gynecologic Infections and Their Complications).
↵2 To whom requests for reprints should be addressed, at Department of Immunology, GU421A, Mayo Clinic, Rochester MN 55905. Phone: (505) 284-0124; Fax: (505) 266-5255; E-mail:
↵3 The abbreviations used are: TCR, T-cell receptor; IL, interleukin; AICD, activation-induced cell death; HTL, helper T lymphocyte; APC, antigen-presenting cell; FBS, fetal bovine serum; PBMC, peripheral blood mononuclear cell; CFSE, 5,6-carboxyfluorescein diacetate, succimidyl ester; LU50, number of effector cells required to produce 50% specific cytotoxicity.
↵4 Unpublished observations.
- Received August 30, 2001.
- Revision received October 29, 2001.
- Accepted November 1, 2001.