| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Division of Immunology, and
Center for Immunotherapy of Cancer and Infectious Diseases, University of Connecticut Health Center, Farmington, CT 06032; and
Department of Surgery, and Emory Vaccine Center, Emory University School of Medicine, Atlanta, GA 30329
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The TNF superfamily of receptors contains a number of costimulatory molecules that are expressed pervasively throughout the immune system (5). Some of these molecules typically enhance activation resulting in protective immune responses (6, 7). For example, OX40 (CD134) has been shown to potently costimulate T cell effector function as measured by proliferation and cytokine production (8, 9, 10), while being capable of delivering a bidirectional signal into APCs via OX40 ligand (11). Moreover, OX40 was shown to reverse unresponsiveness in anergized T cells (12). Another example is CD40 stimulation, which directly induces APC activation, leading to numerous effects including enhanced proliferation and cytokine production (13, 14), as well as class switching of Abs (15, 16, 17). Recently, CD40 expression was shown to occur on T cells, which when ligated led to enhanced immunity (18), although this does not occur in all systems (19).
4-1BB (CD137) is another TNF family member with wide-spanning surface expression patterns and is also capable of costimulating T cells and APCs (20). A unique feature of 4-1BB, compared with OX40 or CD40, is its penchant for enhancing CD8 T cell responses over other cell types (21, 22). This has been particularly useful in controlling viral infections as well as enhancing antitumor immunity (23, 24, 25). Therefore, we investigated whether simultaneous costimulation through 4-1BB and either OX40 or CD40 would result in even better immune responses compared with activating one costimulator alone. Nevertheless, this was unlikely, because simultaneous stimulation of OX40 and CD40 does not significantly enhance T cell responses in vivo (26), suggesting that signaling through these TNF family members is at least partially redundant.
However, our results show that simultaneous dual costimulation through 4-1BB and OX40, but not with CD40, synergistically induces specific CD8 T cell clonal expansion in several in vivo models. The expanding T cells produced large amounts of effector cytokines like IFN-
and TNF when primed with dual costimulation, and effector function did not require T help. These functional outcomes were shown to be a result of enhanced effector T cell accumulation as opposed to increased entry into the cell cycle or increased rounds of cell division. Ultimately, we tested whether this approach could be used therapeutically to eliminate a murine sarcoma. Our data show that established tumors were eliminated by dual costimulation. Furthermore, the mechanism is CD8 T cell dependent and functions in the absence of T help. Therefore, enforced T cell dual costimulation by 4-1BB and OX40 may be an efficacious approach in fighting human cancer even under immunocompromising conditions.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B10.A and C57BL/6 mice were purchased from the National Cancer Institute (Frederick, MD) and The Jackson Laboratory (Bar Harbor, ME). BALB/c mice were purchased from Charles River (Frederick, MD). The OT-I transgenic RAG2-deficient (RAG–/–) mice were kindly provided by Dr. L. Lefrançois (University of Connecticut Health Center). All of the animals were maintained in accordance with federal guidelines at the University of Connecticut Health Center.
The staphylococcal enterotoxin A (SEA)3 was purchased from Sigma-Aldrich (St. Louis, MO) or from Toxin Tech (Sarasota, FL). OVA was purchased from Sigma-Aldrich, and SIINFEKL peptide was purchased from Invitrogen Life Technologies (Grand Island, NY). Tetramer forSIINFEKL-specific CD8 T cells was a kind gift from Dr. L. Lefrançois. The anti-4-1BB mAb (3H3 rat hybridoma; Ref. 21) is specific for murine 4-1BB. Anti-OX40 is an agonist rat mAb specific for murine OX40 (OX86 hybridoma; Ref. 27); and anti-CD40 (FGK45.5) is a rat mAb specific for murine CD40 which was a kind gift from T. Rolink (University of Basel, Basel, Switzerland) (28).
In all experiments, Ag, peptide, or superantigen (SAg) injection, as well as mAbs, were given by the i.p. route, and adoptive transfer of cells was by the i.v. route. To study endogenous T cell responses, 0.3 µg of SEA or 1 mg of OVA was injected into B10.A or C57BL/6 mice, respectively. For in vivo stimulation by enforced costimulation, mice were given anti-4-1BB, anti-OX40, and anti-CD40 alone or in various combinations at the same time as signal 1. For different experiments, the Ab dose changed due to batch-to-batch variation of each Ab preparation. This is based on careful titration experiments examining the in vivo effects of these mAbs on T cells. For the adoptive transfer model, OT-I RAG–/– splenocytes were transferred into C57BL/6 mice, and the day after, either 100 µg of SIINFEKL peptide or OVA diluted into balanced salt solution (BSS) was injected. The number of cells transferred is given in the figure, and on occasion, we found that establishment of OT-I cells was impaired, which was probably due to low transfer number or a poor take by the host.
Cell processing, staining, and flow cytometry
Spleen, peripheral lymph nodes (PLN; inguinal, axillary, brachial), and mesenteric lymph nodes (MLN) were crushed through nylon mesh cell strainers (Falcon; BD Biosciences, San Diego, CA), and then RBC in spleen were lysed with ammonium chloride. After several washes, the cells were counted using a Z1 particle counter (Beckman Coulter, Miami, FL). For nonlymphoid tissue, we followed published procedures (29). Briefly, liver tissue was perfused, crushed through a cell strainer, and resuspended in 35% Percoll (Sigma-Aldrich; or Amersham Biosciences, Piscataway, NJ). After centrifugation, pelleted cells were treated with ammonium chloride and then washed. Lung tissue was perfused, dissected, and digested in the presence of 1.3 mM EDTA and then incubated in collagenase (Invitrogen Life Technologies). After being crushed through a cell strainer, cells were resuspended in 44% Percoll and layered above a 67% Percoll cushion. After centrifugation, cells were isolated from the interface and washed.
For flow cytometry, cells were stained with primary Abs in the presence of a blocking solution containing 5% normal mouse serum (Sigma-Aldrich), 10 µg/ml human -globulin (Sigma-Aldrich), and 0.1% sodium azide in culture supernatant from the 2.4.G.2 hybridoma (anti-FcR; Ref. 30) for 30 min on ice and then washed in wash buffer (3% FBS and 0.1% sodium azide in BSS). For intracellular cytokine staining, 1 x 106 splenocytes were cultured with 1 µg of brefeldin A (Calbiochem, San Diego, CA) in the absence or presence of SIINFEKL peptide or SEA at 37°C for 5 h. The cells were stained with CD45.1-PE Cy5 (eBioscience, San Diego, CA) or CD8-allophycocyanin (BD Pharmingen, San Diego, CA) on ice for 30 min, and after a couple washes, the cells were fixed with 2% paraformaldehyde in BSS. The cells were placed in permeabilization buffer (0.25% saponin in wash buffer) and then incubated with anti-IFN- (eBioscience), anti-TNF (BD Pharmingen), or an isotype control rat IgG1 (eBioscience) for 20 min at room temperature (RT).
For intracellular propidium iodide staining, the cells were stained with anti-CD45.1-FITC (BD Pharmingen) on ice for 30 min. After several washes, the cells were fixed, and the pelleted cells were resuspended in a solution containing 5 mM EDTA, 5 µg/ml propidium iodide, 50 µg/ml RNase A, and 0.3% saponin in PBS. All stained cells were assayed on a FACSCalibur (BD Biosciences), and the data were analyzed with CellQuest (BD Biosciences) or FlowJo (Tree Star, San Carlos, CA) software.
Cell purification, depletion, and culturing
To study the survival of purified OT-I T cells, 2 x 106 splenocytes from OT-I RAG–/– mice were adoptively transferred into C57BL/6 mice, and 1 day later, OVA protein was injected with either anti-4-1BB alone, anti-OX40 alone, or the combination of anti-4-1BB and anti-OX40. After 60 h, spleens were crushed through a cell strainer and run through a nylon wool column (PerkinElmer Life and Analytical Sciences, Boston, MA) for bulk removal of APC. Cells were stained with anti-CD45.1-PE (BD Pharmingen), and then labeled with anti-PE microbead (Miltenyi Biotec, Auburn, CA) according to the manufacturer’s recommendation. CD45.1+ cells were positively selected using MACS columns (Miltenyi Biotec). After washing, purity of CD45.1 OT-I was measured by flow cytometry and was typically 95–98% pure as determined by SIINFEKL tetramer staining. The OT-I T cells were resuspended in CTM (MEM containing amino acids, salts, antibiotics, and FBS), and 100,000 cells were placed into a well of a 96-well plate and cultured at 37°C.
For CD4 T cell depletion in vivo, B10.A mice were treated with 200 µg of purified anti-mouse CD4 (L3/T4) mAb (Cedarlane, Ontario, Canada) on 6 and 2 days before immunization. This technique induced profound depletion of the CD4 T cells as determined by flow cytometry.
CFSE and BrdU analysis
For in vivo CFSE profile studies (31), spleen cells from OT-I RAG–/– mice were isolated, and pelleted cells were resuspended in TC solution (MEM, amino acids, and antibiotics). The cells were labeled with 15 µM CFSE (Molecular Probes, Eugene, OR) and incubated at 37°C for 10 min. The labeling procedure was stopped with cold CTM, and the cells were transferred into recipient mice. For in vitro CFSE profile studies, OT-I cells were transferred into C57BL/6 mice and immunized, and after 60 h, spleen cells were labeled with 1.5 µM CFSE at 37°C for 10 min, and the labeling reaction was halted with cold CTM. The washed cells were resuspended in CTM, and 1 x 106 cells were placed into a well of a 96-well plate. After 20 or 45 h, the cells were stained with anti-CD45.1 PE Ab.
For BrdU staining, we followed a previous report (32). Briefly, 10 µM BrdU (Sigma-Aldrich) was added at time 0 or 30 h of culture. After 15 h of BrdU, the cells were stained with anti-CD45.1-PE on ice for 30 min, and then dehydrated and fixed in ice-cold 95% ethanol. The cells were permeabilized with 1% paraformaldehyde, 0.01% Tween 20 in PBS, and then treated with 50 Kunitz U of DNase I (Sigma-Aldrich) at RT for 10 min. After several washes, the cells were stained with anti-BrdU-FITC (BD Pharmingen) at RT for 30 min, and then analyzed by flow cytometry.
Fibrosarcoma tumor model
MethA fibrosarcoma cells were maintained by weekly i.p. passage in BALB/c mice as done previously (33). Briefly, the effect of dual costimulation on the fibrosarcoma MethA tumor was studied by using an intradermal injection of 0.5–2.0 x 106 MethA tumor cells in the right flank next to the midline of 6- to 8-wk-old female BALB/c mice. Growth was monitored twice a week using calipers to measure tumor diameter. For CD4, CD8 depletion experiments, anti-CD4, anti-CD8, or rat IgG control Ab was injected into BALB/c mice 3 days before and 6 days after injection of 1 x 106 MethA cells. Lymphocyte depletion was confirmed by flow cytometry.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
SEA stimulates T cells in an oligoclonal fashion by activating TCR V
3-bearing CD4 or CD8 T cells in an MHC class II-restricted manner (34). SEA was previously shown to induce specific T cell clonal expansion followed by profound deletion, and sometimes anergy induction (35, 36). To test the effects of simultaneous dual costimulation, mice were injected with SEA plus control rat IgG, anti-4-1BB, -CD40, -OX40, or dual combinations of various costimulators (Fig. 1). On day 10 after injection, PLN, MLN, and spleen cells were analyzed ex vivo by flow cytometry for the presence of CD8 (Fig. 1a), or CD4 (b) T cells expressing TCR V3. Deletion of SEA-responsive T cells was observed in all tissues when SEA was given with control IgG. Consistent with previous data, 4-1BB costimulation rescued CD8, but not CD4, T cells from deletion (22). CD40 or OX40 costimulation did not block CD8 T cell deletion and had minimal effects on CD4 T cells at this time point. CD40 and 4-1BB costimulation induced additive effects on the SEA-responsive T cells. In contrast, 4-1BB and OX40 synergistically induced CD8 V3 responses in all tissues examined. This was not the case for CD4 T cells where the responses were additive at best. To stringently test the notion that 4-1BB and OX40 mAbs synergistically stimulate T cell responses, equivalent amounts of mAb were given as shown in Table I. The data show that the bulk amount of mAb did not determine the level of T cell response, because 75 µg of each individual mAb was not equivalent to 25 plus 50 µg of 4-1BB and OX40 mAbs, respectively.
|
|
3 T cells in the dual-costimulation group compared with the control IgG. Even in comparison to the 4-1BB and OX40 groups alone, the dual-costimulation group induced a >4-fold increase on days 4 and 7. Collectively, these data show that there is a massive increase in proportion and number of specific CD8 T cells during the effector phase after in vivo stimulation with dual costimulation by anti-4-1BB and -OX40.
|
|
The transfer of OT-I T cells into recipient mice has proven to be a useful model to study basic CD8 T cell activation (38), antitumor immunity, and other immunological responses (39, 40, 41). We used the CD45.1 congenic marker to track RAG–/– OT-I T cells after transfer into C57BL/6 mice. Recipient mice were immunized with OVA protein with anti-4-1BB, -OX40, or dual costimulation followed by monitoring the presence of peripheral blood CD8 T cells that are CD11ahigh, CD45.1 double-positive using flow cytometry (Fig. 2c). Dual costimulation, in comparison with the other treatments, was more effective at stimulating OT-I T cells in vivo. This was true for recipient mice, which received 50,000 OT-I cells (Fig. 2c, upper panel) where there was a >3-fold increase in the frequency of specific cells in blood. Even the 1,000 OT-I cell (Fig. 2c, lower panel) transfer showed an increase; however, these results were not synergistic and were more variable, potentially due to competition from endogenous SIINFEKL-specific cells. Thus, 4-1BB and OX40 dual costimulation, compared with treatment with either costimulator alone, synergistically enhances in vivo specific-CD8 T cell clonal expansion.
Characterizing effector function and fitness after dual costimulation
Our next goal was to determine whether the effects of dual costimulation were confined to lymphoid tissue. Sixty thousand OT-I cells were transferred into recipient mice, and the day after, they were treated with OVA protein and control rat IgG, anti-4-1BB, -OX40, or dual costimulation. On day 12, peripheral blood, PLN, MLN, spleen, and lung were analyzed for the percentage of CD8 T cells expressing high levels of CD11a and CD45.1 (Fig. 3a). In a separate experiment, 500,000 cells were transferred, and day 20 was examined (Fig. 3a, see bottom panel for liver data). In both experiments, dual costimulation led to a dramatic increase of OT-I cells not only in spleen and LNs but also in lung and liver tissue. Interestingly, the frequency of lung OT-I cells in the OX40 alone group was much higher than the 4-1BB alone group, but the opposite was true for day 20 liver; this may also be related to the difference in the number of OT-I cells transferred between the two experiments. Nevertheless, in both cases, the dual-costimulated mice contained the largest OT-I population. These data show that dual costimulation mediates spread of effector CD8 T cells throughout the body.
|
after restimulation and about one-third made TNF. Although the other individual costimulatory groups generated OT-I IFN- producers at both time points (data not shown), the amount of IFN- produced per OT-I cell was more when dual costimulation was used (data not shown). Therefore, the fact that many more OT-I T effector cells were present and produced greater amounts of IFN-, demonstrates that dual costimulation generates a robust boost to effector responses. Because of the increase in effector cells observed (Fig. 3a), it was possible that dual costimulation induced more proliferation or ability to undergo an increase in the number of cycles of division. To test this idea, we isolated OT-I cells, stained them with CFSE, and then transferred them into recipient mice. At 50 h after immunization with OVA and costimulation, dilution of CFSE in the OT-I population was analyzed (Fig. 4a). The data show that there was no difference in any of the groups examined. Nevertheless, it was possible that earlier time points may have revealed differences between the groups; thus, 28 or 32 h were examined but showed comparable dilution of CFSE between the groups (data not shown).
|
5-fold increase in the percentage of OT-I cells from the dual-costimulation group at 45 h; the other treatments generated a <2-fold response (Fig. 4b). The fold increase in the number of OT-I cells in these cultures from 0 to 45 h was 0.34 for the IgG control group, 2.64 for 4-1BB alone, 1.38 for OX40 alone, and 5.41 for dual costimulation. Thus, dual costimulation enhanced OT-I cell accumulation in vitro without added peptide, although it is possible that peptide was still being presented from the initial immunization. One possibility was that a greater proportion of OT-I T cells were initiating division. To test this idea, an identical culture was set up as in Fig. 4b, and BrdU was added either at the beginning or halfway through incubation. BrdU incorporation is a sensitive measurement of DNA synthesis, and under our test conditions, the percentage of OT-I cells incorporating BrdU would be attributable to their ability to initiate division. BrdU incorporation during the first 20 h of culture was higher for OT-I cells that received enforced in vivo costimulation compared with the IgG control group (Fig. 4c). Nevertheless, the type of costimulation did not affect the OT-I response, and all groups showed very similar levels of BrdU incorporation. When BrdU was added during the last 15 h of culture, all OT-I populations incorporated equivalent levels of BrdU. Therefore, dual costimulation did not induce a greater proportion of OT-I cells to incorporate BrdU as would be expected if these cells possessed an enhanced ability to initiate cell division.
To further test this idea, cells were taken from recipient mice as above, labeled with CFSE, and assayed for dye dilution at 20 and 45 h after culture. In all groups, the extent of CFSE dilution was similar, but cell counts between the groups were vastly different (Fig. 4d). At both time points, cells primed in dual-costimulation-treated mice induced a massive increase in cell counts compared with the other groups even though the extent of dye dilution was similar. Taken together, the data in Fig. 4 suggest that the OT-I CD8 T cells primed with dual costimulation are far more likely to remain viable after, and during, cell cycle traverse as opposed to treatment with either costimulator alone.
Perhaps the most stringent test of this idea is to assay pure OT-I T cells in the absence of contaminating populations that may provide survival signals in vitro. Thus, OT-I T cells were stimulated in vivo as in Fig. 4 and, at 60 h, purified using MACS columns with anti-CD45.1 mAb. Purity was 98% (CD8+CD45R.1+), and cells were placed in vitro without added APCs, other lymphocytes, or peptide. At 0 h of culture, the cells were analyzed for cell cycle status, and all groups possessed a substantial population of cells cycling through S phase, as measured by intracellular DNA content using propidium iodide (data not shown). After 18 h, the cells were analyzed for viability (Fig. 5, top panel) and cell cycle status (bottom panel) by flow cytometry. In vivo priming with OX40 mAb generated 439 viable OT-I cells/50,000 events (viable cells = cells from top panel multiplied by the percentage of OT-I cells from bottom panel), followed by 4-1BB with 2239 OT-I cells. Dual costimulation yielded the greatest number of viable OT-I cells at 4284. Collectively, these data show that dual costimulation does not induce more rounds of division (Fig. 4, a and d), nor a greater ability to initiate DNA synthesis (Fig. 4c), but does enhance accumulation (Fig. 4b), through efficient cell cycle traverse and survival (Fig. 5).
|
Using multiple in vivo models, we showed for the first time that dual costimulation by 4-1BB and OX40 synergistically induce specific CD8 T cell clonal expansion. We tested whether the responding CD8 T cells required T cell help for this effect. Thus, one group of mice were treated with a CD4-depleting mAb and the other given a control IgG, and depletion was confirmed throughout the experiment. Both groups were injected with SEA plus dual costimulation, and on days 2–5, peripheral blood lymphocytes were analyzed for the presence of responding CD8 V3 T cells (Fig. 6a). The data show that there was a profound increase in specific CD8 T cell clonal expansion whether or not CD4 T cells were present; however, in the presence of CD4 T cells, expansion was enhanced by 20%. Therefore, the vast majority of clonal expansion was CD4 independent. Multiple tissues were examined on day 11, and the data show that effector CD8 T cells from CD4-depleted mice did not accumulate to the same degree as cells in control mice (Fig. 6b). It is unlikely that migration can explain this difference, because accumulation was inhibited in nonlymphoid as well as lymphoid tissue. Collectively, this suggests that CD4 help may contribute to CD8 T cell survival in vivo.
|
5 h, and IFN- synthesis was evaluated by intracellular cytokine staining. Because TCR is down-regulated in this assay, we were only able to examine the percentage of CD8 T cells possessing IFN- (Fig. 6c). Thus, on day 5, there was little difference in the percentage of CD8 T cells producing IFN-. On day 12, there appeared to be weakened effector function, but this was not likely the case, because the proportion of V3 T cells in the CD4-depleted mice was lower on day 12 compared with CD4-intact mice (data not shown; but for a comparison, see Fig. 6b). Therefore, effector function is intact despite the fact that survival of the effector CD8 T cells is impaired in the absence of help. To test this idea without depleting cells, we stimulated OT-I recipient mice with SIINFEKL peptide instead of OVA. It is unlikely that CD4 T cells respond to SIINFEKL/MHC I complexes, and therefore, they have little opportunity for TCR ligation, especially because no adjuvant is used. Three groups of OT-I recipient C57BL/6 mice were injected with soluble SIINFEKL peptide plus either anti-4-1BB, -OX40, or dual costimulation. On day 5 after injection, the spleen (Fig. 6d, left panels) and PLN cells (right panels) were analyzed for the percentage of OT-I cells (top panels) and absolute OT-I number (bottom panels) by flow cytometry and carefully counting cell numbers. Differences in the percentages or numbers of OT-I cells between the 4-1BB and OX40 groups were very small regardless of how many cells were transferred. The dual-costimulation group generated massive amounts of OT-I T cells in spleen and PLNs as measured by absolute number or percentage. Although not as dramatic for 50–60,000 cell transfer groups, this trend was still very evident. To measure whether effector function was intact, cells from the same experiment were restimulated in vitro with SIINFEKL peptide or without (data not shown), and the data show that the dual-costimulation group yielded cells that either responded as well as the other groups or even better (Fig. 6e). Therefore, by factoring the proportion, number, and responsiveness of the effector cells, we conclude that dual costimulation provides a massive boost to effector function.
We tested whether dual costimulation would be useful in the rejection of established tumors. The MethA fibrosarcoma (MethA) is a stringent tumor model based on the fact that MethA is a progressive fibrosarcoma induced by methylcoanthrene in immunocompetent mice. A single injection of as few as 100,000 cells intradermally will lead to complete establishment of tumors and uniform mortality within 4 wk in normal BALB/c mice (42). The rejection of MethA induced by heat shock proteins was shown to be dependent on both CD4+ and CD8+ T cells (43, 44). Therefore, in the first experiment, we tested whether dual costimulation would hamper or enhance an established vaccine strategy in this model. In the first experiment, we demonstrated that 500,000 MethA cells expressing membrane-bound gp96 were rejected faster if they received dual costimulation (data not shown); however, dual-costimulation mAbs administered on days 3 and 6 completely protected mice from the control neo-transfected MethA tumor cell line (Fig. 7a). The control rat IgG did not protect the tumor-bearing mice.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To rigorously study dual costimulation by OX40 and 4-1BB, we conducted experiments in three different murine models where specific T cells were tracked in vivo. The data convincingly show that simultaneous dual costimulation by OX40 and 4-1BB mAbs synergistically stimulate SAg- or peptide-specific CD8 T cells to clonally expand. This included endogenous CD8 T cells like those specific to SAg or peptides, as well as cells that have been adoptively transferred into recipient mice and then immunized with whole protein (Fig. 2). Perhaps the most impressive data testing dual costimulation was that precursor frequency of the specific CD8 T cells was not restrictive. Throughout this study, it was demonstrated that high or low precursor frequency was not a limiting factor. Even when transferring as much as 500,000 or as little as 50,000 OT-I spleen cells, the synergistic effect was clearly evident, and at 1,000 OT-I cells, an enhancement was also observed, although this was complicated by the fact that competition by endogenous SIINFEKL-specific CD8 T cells was likely. This notion is consistent with a previous study demonstrating that endogenous cells can compete with transferred cells (45). Nonetheless, the endogenous cells responded synergistically by day 6 after immunization, demonstrating that precursor frequency did not limit the effects of dual costimulation (Fig. 2, b and c).
Recently, data from infectious disease models (29, 46) have demonstrated that responding primary and memory specific-CD8 T cells can migrate to nonlymphoid tissue, and likely take up residence (47). Perhaps this is an important mechanism to control future infections (48) but nevertheless is a trademark of a robust effector response. On day 12, OT-I cells primed with dual costimulation migrated to (non)lymphoid tissues with the highest frequency of OT-I cells detected in lung, which was the same for OX40 or 4-1BB alone (Fig. 3a). Interestingly, we detected a substantial population of activated CD11ahighCD45.1– endogenous cells. Presumably, these cells were responsive to OVA or possibly the agonist mAb itself. However, upon restimulation with SIINFEKL peptide, only the OT-I T cells responded by producing effector cytokines like TNF or IFN- (Fig. 3b). Therefore, dual costimulation did not alter migration, but maintained survival of the effector T cells and their function, although the mechanism of expansion was not clear.
To address this issue, we tested whether dual costimulation accelerated or increased rounds of cellular division, thereby leading to robust clonal expansion. The evidence suggested that dual costimulation did not accelerate division in vivo (Fig. 4a), although it was impossible to tell whether increased rounds of division occurred, because as the cells divided, they retained an undetectable level of CFSE. To address this issue, cells from dual-costimulated mice vs individual-costimulated mice were placed in vitro without any known stimulus and then enumerated 2 days after culture (Fig. 4b). Based on proportion or absolute number, there was a 5-fold increase in the dual-costimulatory group and a small increase in the others. This is even greater than the effect that LPS has on rescuing T cells from deletion (49). Therefore, dual costimulation increased clonal expansion by enhancing accumulation, but not by accelerating division.
Accumulation of activated T cells after immunization is a complicated biological process. This phenomenon is based on at least the ratio of T cell survival to death, vs proliferation. In one example of many, T cells may proliferate and die at the same rate and thereby do not accumulate beyond their initial numbers; however, populations that proliferate at the same rate but die less frequently will accumulate. Nonetheless, the half-life of a circulating T cell is very difficult to accurately measure, because techniques like BrdU incorporation are fraught with underappreciated toxicity issues (50). Thus, incorporation of BrdU can be minimally or reliably interpreted as an attempt to initiate cell cycle traverse. This is based on a short exposure to BrdU in a time span that permits BrdU incorporation. Our experiments show that costimulation priming provided an advantage in the uptake of BrdU in specific T cells at the beginning of culture, but not at the end (Fig. 4c). Nevertheless, T cells from dual-costimulated mice did not possess more or less of an ability to initiate cell division than the other costimulation groups. Therefore, even if peptide was being presented in this in vitro culture as a result of the initial immunization, these data show that it did not preferentially influence cells from the dual-costimulation group compared with the others.
To visualize survival and proliferation at the same time, the cells were labeled with CFSE and monitored in vitro as before (Fig. 4d). These data suggest that, although the ability to initiate division is the same between all costimulation groups, there is a greater propensity of OT-I cells to survive through the cell cycle. One possibility to explain how dual costimulation mediates survival is to examine whether the OT-I cells possess an intrinsic program allowing them to preferentially survive. Alternatively, nonspecific cells like dendritic cells, macrophages, or other lymphocytes may provide the OT-I cells with survival signals. Pure OT-I T cells were clearly more capable of surviving when they were derived from mice treated with dual costimulation through OX40 and 4-1BB as opposed to either alone (Fig. 5). The molecules involved in this process have not been identified. One possibility is a role for Bcl-xL, Bcl-2, and/or protein kinase B, which are reported to be essential for OX40 signaling in CD4 T cells (51, 52), although our data show little survival of CD8 T cells when costimulated through OX40 alone. From a cellular perspective, it is also possible that CD8 T cell survival is enhanced when CD4 T cells are present during priming of the CD8 T cells. Recently, CD4 T cells were shown to play a minimal role in helping the primary response of CD8 T cells, but in contrast influenced the secondary response of specific CD8 T cells (53, 54, 55). Therefore, we formulated the hypothesis that CD4 T cells may not play a role influencing CD8 T cells in our dual-costimulation model.
Depletion of CD4 T cells minimally affected CD8 T cell clonal expansion, but in contrast, survival of the effector cells was substantially impaired in all (non)lymphoid tissues examined (Fig. 6, a and b). These data complement and extend those reported in the pathogen models discussed above. However, perhaps the most important result was that effector function was maintained even at the later stages when specific T cell accumulation was impaired (Fig. 6c). These data demonstrate that survival and effector function are not always linked or dependent upon each other. In a second test of this idea, mice were primed with a CD8-specific peptide instead of whole protein to eliminate helper function, and we detected massive clonal expansion coupled to a robust effector response (Fig. 6, d and e). One caveat is that the mAbs themselves may have provided a CD4 TCR signal; however, the response of these potential CD4 cells would theoretically lag behind the CD8 response, because SIINFEKL peptide was already processed in contrast to the putative CD4 epitopes in the mAbs. Taken together, whether the life span of the effector CD8 T cell was impaired or not, or importantly whether CD4 T cells were present or present but not stimulated, CD8 effector cytokine production was robust. Therefore, we hypothesized that this strategy may be effective under recalcitrant conditions of immunity.
We tested whether dual costimulation could help fight immunologically resistant tumors. Initially, we showed that the combination of 4-1BB and OX40 efficiently induced rejection of an established fibrosarcoma in mice (Fig. 7a). To investigate the mechanism, we demonstrated that a single injection of mAb 7 days, as opposed to 3 days, after tumor establishment was sufficient to completely retard tumor growth (Fig. 7b). Two major aspects with these experiments require mention: first, total tumor rejection was observed with only two injections of mAb; and second, we used mAbs without any immunizing tumor Ag. In contrast, CTLA4 blockade without a tumor Ag vaccine targeting p53 was not very effective in MethA tumor rejection (56). It remains to be seen whether the combination of dual costimulation and a tumor Ag vaccine is most effective.
Ultimately, we examined antitumor effectiveness of dual costimulation under immunocompromising conditions such as in the absence of CD4 T cells. In this model, there is a substantial amount of evidence suggesting that CD4 T cells are essential for tumor rejection under vaccinating conditions or even during immunotherapy (42, 57). Dual costimulation was very effective in the presence or absence of CD4 T cells, but completely dependent upon CD8 T cells (Fig. 7c). Taken together, these data suggest that dual costimulation triggers CD8 effector function to eliminate tumor cells even in the absence of T cell help. Therefore, we conclude that dual costimulation may provide a fertile avenue for immunotherapy in humans suffering from cancer.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 This work was supported by in part by National Institutes of Health Grants AI42858A and AI52108 (to A.T.V.). L.M. is supported by National Institutes of Health Immunology Training Grant T32-AI07080. 
2 Address correspondence and reprint requests to Dr. Anthony T. Vella, Division of Immunology, MC1319, University of Connecticut Heath Center, 263 Farmington Avenue, Farmington, CT 06032. E-mail address: vella{at}uchc.edu
3 Abbreviations used in this paper: SEA, staphylococcal enterotoxin A; SAg, superantigen; BSS, balanced salt solution; PLN, peripheral lymph node; MLN, mesenteric lymph node; RT, room temperature.
Received for publication April 7, 2004. Accepted for publication June 25, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
heavy chain class switching. Immunity 5:319.[Medline]
. J. Exp. Med. 186:71.
11 integrin VLA-1 regulates CD8 T cell-mediated immune protection against heterologous influenza infection. Immunity 20:167.[Medline]
This article has been cited by other articles:
|
|
 |
|
  Y. H. Kim, B. K. Choi, W. J. Kang, K. H. Kim, S. W. Kang, A. L. Mellor, D. H. Munn, and B. S. Kwon IFN-{gamma}-indoleamine-2,3 dioxygenase acts as a major suppressive factor in 4-1BB-mediated immune suppression in vivo J. Leukoc. Biol., May 1, 2009; 85(5): 817 - 825. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. P. McAleer, R. J. Rossi, and A. T. Vella Lipopolysaccharide Potentiates Effector T Cell Accumulation into Nonlymphoid Tissues through TRIF J. Immunol., May 1, 2009; 182(9): 5322 - 5330. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-J. Lee, M. Long, A. J. Adler, R. S. Mittler, and A. T. Vella The IKK-neutralizing compound Bay11 kills supereffector CD8 T cells by altering caspase-dependent activation-induced cell death J. Leukoc. Biol., January 1, 2009; 85(1): 175 - 185. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Bandyopadhyay, M. Long, H. Z. Qui, A. T. Hagymasi, A. M. Slaiby, M. A. Mihalyo, H. L. Aguila, R. S. Mittler, A. T. Vella, and A. J. Adler Self-Antigen Prevents CD8 T Cell Effector Differentiation by CD134 and CD137 Dual Costimulation J. Immunol., December 1, 2008; 181(11): 7728 - 7737. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. Robertson, R. J. Messer, A. B. Carmody, R. S. Mittler, C. Burlak, and K. J. Hasenkrug CD137 Costimulation of CD8+ T Cells Confers Resistance to Suppression by Virus-Induced Regulatory T Cells J. Immunol., April 15, 2008; 180(8): 5267 - 5274. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. L. Redmond, M. J. Gough, B. Charbonneau, T. L. Ratliff, and A. D. Weinberg Defects in the Acquisition of CD8 T Cell Effector Function after Priming with Tumor or Soluble Antigen Can Be Overcome by the Addition of an OX40 Agonist J. Immunol., December 1, 2007; 179(11): 7244 - 7253. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-J. Lee, R. J. Rossi, S.-K. Lee, M. Croft, B. S. Kwon, R. S. Mittler, and A. T. Vella CD134 Costimulation Couples the CD137 Pathway to Induce Production of Supereffector CD8 T Cells That Become IL-7 Dependent J. Immunol., August 15, 2007; 179(4): 2203 - 2214. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Zhang, R. E. Sadun, R. S. Arias, M. L. Flanagan, S. M. Sachsman, Y.-C. Nien, L. A. Khawli, P. Hu, and A. L. Epstein Targeted and Untargeted CD137L Fusion Proteins for the Immunotherapy of Experimental Solid Tumors Clin. Cancer Res., May 1, 2007; 13(9): 2758 - 2767. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-W. Lee, Y. Park, A. Song, H. Cheroutre, B. S. Kwon, and M. Croft Functional Dichotomy between OX40 and 4-1BB in Modulating Effector CD8 T Cell Responses J. Immunol., October 1, 2006; 177(7): 4464 - 4472. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Lizee, L. G. Radvanyi, W. W. Overwijk, and P. Hwu Immunosuppression in melanoma immunotherapy: potential opportunities for intervention. Clin. Cancer Res., April 1, 2006; 12(7): 2359s - 2365s. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Murata, B. H. Ladle, P. S. Kim, E. R. Lutz, M. E. Wolpoe, S. E. Ivie, H. M. Smith, T. D. Armstrong, L. A. Emens, E. M. Jaffee, et al. OX40 Costimulation Synergizes with GM-CSF Whole-Cell Vaccination to Overcome Established CD8+ T Cell Tolerance to an Endogenous Tumor Antigen J. Immunol., January 15, 2006; 176(2): 974 - 983. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Serghides, J. Bukczynski, T. Wen, C. Wang, J.-P. Routy, M.-R. Boulassel, R.-P. Sekaly, M. Ostrowski, N. F. Bernard, and T. H. Watts Evaluation of OX40 Ligand as a Costimulator of Human Antiviral Memory CD8 T Cell Responses: Comparison with B7.1 and 4-1BBL J. Immunol., November 15, 2005; 175(10): 6368 - 6377. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Song, X. Tang, K. M. Harms, and M. Croft OX40 and Bcl-xL Promote the Persistence of CD8 T Cells to Recall Tumor-Associated Antigen J. Immunol., September 15, 2005; 175(6): 3534 - 3541. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Y. Ma, S. A. Mikolajczak, A. Danesh, K. A. Hosiawa, C. M. Cameron, A. Takaori-Kondo, T. Uchiyama, D. J. Kelvin, and A. Ochi The expression and the regulatory role of OX40 and 4-1BB heterodimer in activated human T cells Blood, September 15, 2005; 106(6): 2002 - 2010. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
); SEA plus 25 µg of anti-4-1BB (
); SEA plus 100 µg anti-OX40 (
); and SEA plus anti-4-1BB and anti-OX40 (
). Resident T cells in spleen were analyzed on days 2, 4, 7, 9, 11, and 14. The data show mean percentage ± SEM of CD8 T cells expressing V
20 h) or at 45 h (30
) or nondepleted (
), and the last received dual costimulation (
), or rat IgG control Ab (