Combined CD137 (4-1BB) and adjuvant therapy generates a developing pool of peptide-specific CD8 memory T cells

Transcription

1 International Immunology, Vol. 18, No. 2, pp doi: /intimm/dxh371 ª The Japanese Society for Immunology All rights reserved. For permissions, please Combined CD137 (4-1BB) and adjuvant therapy generates a developing pool of peptide-specific CD8 memory T cells Lara Myers 1, Seung Woo Lee 2, Robert J. Rossi 1, Leo Lefrancois 1, Byoung S. Kwon 3, Robert S. Mittler 4, Michael Croft 2 and Anthony T. Vella 1 1 Department of Immunology, School of Medicine, University of Connecticut Health Center, 263 Farmington Avenue, Farmington, CT 06030, USA 2 Division of Molecular Immunology, La Jolla Institute for Allergy and Immunology, Science Center Drive, La Jolla, CA 92121, USA 3 Immunomodulation Research Center, University of Ulsan, Ulsan, Republic of Korea 4 Department of Surgery and Emory Vaccine Center, Emory University School of Medicine, Atlanta, GA 30329, USA Keywords: CD137, memory and adjuvant Abstract In practice, vaccines should induce lasting and efficacious T cell immunity without promoting deleterious pathological consequences. To accomplish this goal we immunized mice with ovalbumin peptide, polyinosinic polycytidylic and anti-cd137. Vaccinated mice retained a massive functional CD8 T cell memory pool in lymphoid and non-lymphoid tissues for >1 year. The memory T cells clonally expanded, produced substantial amounts of IFNc, and responded vigorously to vesicular stomatitis virus infection. To understand how the vaccine might function, we showed that the antigen-specific T cells must bear CD137 in order for optimal priming to occur. Thus, anti-cd137 agonist mab directly stimulated peptide-specific CD8 T cells and conditioned them to survive. In contrast, CD137-deficient CD8 T cells did not survive despite CD137 expression by antigen presenting cells. Taken together, the data indicate that CD137 and adjuvant combined therapy is an efficacious vaccine strategy for immunization with non-replicating inert antigen. Introduction Memory T cell subsets are detected in distinct locations in the body (1), with central memory cells restricted to lymphoid organs and effector memory cells dispersed throughout the non-lymphoid tissues such as the liver and lung (2, 3). The peripheral location and potent response of the effector memory cells allows them to be on the front line of defense after re-infection (4). An optimal vaccination strategy produces both subsets that are long-lasting populations and capable of functioning in aged animals. It has been demonstrated that as mice age the T cell compartment acquires a reduced ability to mount productive immune responses (5, 6), lending to the frustration of producing vaccines that provide protection years after immunization. Co-stimulation of T cells along with presentation of antigen in the context of MHC is required to generate a productive immune response. The compilation of data regarding co-stimulation suggest that co-stimulatory receptors share overlapping and unique functions, as well as distinct temporal and cell-type-specific expression patterns (7). This diversity provides multiple targets for immune conditioning through the exploitation of co-stimulatory signals during vaccination. Enforced co-stimulation can boost T cell survival (8), as well as enhance effector function of T cells during the priming phase of the immune response (9, 10). As an example, CD137 is a member of the tumor necrosis factor receptor family and has been shown to augment CD4 and CD8 T cell responses (11 16). CD137 is rapidly expressed on 24-h peptide stimulated CD4 and CD8 T cells (17), and at even earlier time points after in vivo superantigen stimulation (14). Importantly, CD137 co-stimulation substantially enhances CD8 T cell effector responses (7). This has been demonstrated in viral infection models of influenza (18) and lymphocytic choriomeningitis virus (19). Also, CD137 costimulation has been reported to expand populations of T cells Correspondence to: A. T. Vella; vella@uchc.edu Received 22 June 2005, accepted 18 November 2005 Transmitting editor: P. Ohashi Advance Access publication 22 December 2005

2 326 Combined CD137 (4-1BB) and adjuvant therapy specific to non-dominant viral epitopes (20). Lastly, therapeutic use of agonist anti-cd137 mabs as shown to induce elimination of established tumors (13, 21). We combined a therapeutic anti-cd137 mab with polyinosinic polycytidylic (PIC) to formulate a peptide-based vaccine. Peptide-specific CD8 T cells were detected in great abundance over a year later and, importantly, continued to possess potent Tc1-type effector function. To examine the mechanism of CD137 co-stimulation, we developed an in vivo model where the responding Tcells either expressed or lacked CD137. This is crucial since many cell populations besides activated T cells express CD137 (22 25). The data demonstrate that CD137-expressing T cells, as opposed to nonexpressing T cells, underwent substantial antigen-specific clonal expansion after triggering CD137 with the therapeutic mab. Taken together, CD137-enforced co-stimulation therapy may be a useful tool for developing vaccines toward nonreplicating inert antigen, without the need for live-attenuated or heat-killed pathogens. Methods Mice, reagents, mabs and injections C57BL/6, C57BL/6 CD45.1 and C57BL/6 Thy1.1 mice were purchased from the National Cancer Institute (Frederick, MD, USA) or Jackson Laboratory (Bar Harbor, ME, USA). The C57BL/6 CD OTI transgenc mice (26) were kindly provided by Zihai Li (University of Connecticut Health Center, Farmington, CT, USA), and the C57BL/6 CD137 ÿ/ÿ (27) and C57BL/6 OTI CD137 ÿ/ÿ mice were bred in our laboratories. All mice were maintained in accordance with federal guidelines. Adoptive transfers were performed by intravenous (i.v.) injection and all immunizations were through the intraperitoneal route. A total of 1 mg of ovalbumin peptide (OVA) (Sigma, St Louis, MO, USA) was injected in sterile balanced salts solution (BSS). The agonist anti-cd137 mab was purified from 3H3 hybridoma culture supernatant by protein-g agarose column (Invitrogen, Grand Island, NY, USA) and was injected at a dose range between 25 and 100 lg based on batch-tobatch potency determined by T cell clonal expansion in a titration assay (13). A rat IgG was used as the control (Sigma), and was given at an equivalent dose. PIC was administered at 150 lg and LPS at lg (both from Sigma). SIINFEKL peptide was produced by Invitrogen and injected at 100 lg for Tcell priming or at 200 lg for memory Tcell recall in congenic recipients (as noted in the figure legends). Cell processing, staining and flow cytometry Blood was collected by tail bleeding into 10 U ml ÿ1 heparinized BSS, followed by lysis of RBCs using ammonium chloride. For lung and liver analyses, mice were anesthetized with ketamine/xylazine and then perfused with heparinized PBS solution and tissues processed for lymphocyte removal as previously described (4). Briefly, the spleen and liver were crushed through nylon mesh cell strainers (Falcon; BD Biosciences, San Diego, CA, USA) and RBCs lysed with ammonium chloride. Liver hepatocytes were removed using a 35% percoll gradient and lung tissue digested in a 150-U ml ÿ1 collagenase solution followed by lymphocyte collection at the interface of a 44%/67% percoll gradient. Cells were washed with BSS, counted using a Z1 particle counter (Beckman Coulter, Miami, FL, USA) and placed into staining buffer (BSS, 3% FBS, 0.1% sodium azide). Non-specific binding was blocked using 5% heat-inactivated normal mouse serum, culture supernatant from the 2.4.G.2 hybridoma (28) and 10 lgml ÿ1 human c globulin (Sigma). The incubation with primary mabs was for ~30 min on ice. Anti-CD8, -CD11a, -CD45.1, -CD45.2, -CD90.1 and -CD90.2 mabs were purchased from either BD PharMingen (San Diego, CA, USA) or ebiosciences (San Diego, CA, USA) and were conjugated to FITC, PE, allophycocyanin, peridinin chlorophyll protein or PE Cy7. For tetramer staining, SIINFEKL-tetramer and anti- CD8 mab were incubated for 1.5 h at room temperature prior to primary antibodies. For intracellular cytokine staining, between and splenocytes were cultured in U-bottom 96-well plates (Falcon) for 5 h at 37 C and 5% CO 2 with 1 lg brefeldin A (Calbiochem). Cultures were placed in complete tumor media (MEM with amino acids, salts, antibiotics and FBS) with either nothing or 5 lgml ÿ1 SIINFEKL peptide (Invitrogen). Cells were then stained with anti-cd8 allophycocyanin (ebioscience) or anti-cd45.2 Cy5, fixed with 2% PFA BSS and permeabilized with 0.25% saponin (Sigma) in staining buffer. The cells were then incubated for 15 min at room temperature followed by 30 min on ice with anti-ifnc PE, anti-il-4 PE (data not shown) or an isotype control (ebioscience). Cells were analyzed on a FACSCalibur (BD Biosciences, Mountain View, CA, USA) or an LSR-II (Becton Dickinson) and data analysis was completed using CELLQUEST (BD Biosciences) or FlowJo (Tree Star, San Carlos, CA, USA) software. Cell proliferation To measure in vivo clonal expansion, to total cells from mice immunized previously were adoptively transferred into CD45.1 congenic recipients. In some instances, the splenocytes were enriched for lymphocytes by passing over a nylon wool column (29). The next day, 200 lg of SIINFEKL peptide was given intra-peritoneally (i.p.) and spleens of the recipients were analyzed 6 7 days later for the percent CD CD8 + SIINFEKL-tetramer + population. Also, without adoptive transfer, memory cells were directly challenged with 200 lg SIINFEKL. Mice were bled before and after recall immunization to determine the percent CD8 + CD11a hi SIINFEKL-tetramer + cells. Only mice that generated a priming response were analyzed. Proliferation of memory cells was also tested using infection by recombinant vesicular stomatitis virus expressing ovalbumin peptide (VSV-OVA) (30). Mice immunized either 2 or 17 months prior, or age-matched normal controls, were intravenously injected with 10 5 plaque-forming units (PFUs) and monitored for the percent of CD8 + T cells that were CD11a hi SIINFEKL-tetramer + post-infection. Results CD137 and adjuvant combined therapy generates long-lived CD8 + memory T cells Optimal adaptive immune responses are dependent on longterm surviving T cells that possess the ability to develop into

3 functional memory cells. To test the longevity potential of CD8 + T cells, mice were immunized with various combinations of OVA, anti-cd137 mab, control IgG, LPS and PIC or left as unimmunized controls and then analyzed after priming through over 1 year later. This approach permitted analysis of endogenous T cells thereby avoiding complications of analyzing transferred T cells. Table 1 lists the vaccine treatments, the percent of endogenous CD8 + T cells that were SIINFEKL-tetramer + during the priming phase (at 6 7 days) followed by the percent and number remaining at months post-immunization. As shown in Table 1, no immunization or immunization with OVA alone led to very little priming or memory. Even inclusion of PIC adjuvant or anti-cd137 was not sufficient to consistently induce lasting memory, although in some cases substantial priming was evident (Table 1, lines 9 and 10). However, combining PIC with anti-cd137 induced lasting T cell survival in response to OVA only when priming was excellent (Table 1, compare lines with 14 18). This result is magnified when analyzing absolute numbers especially considering that in unimmunized mice the specific cells are practically undetectable. This was also the case when analyzing adjuvant and anti- CD137 immunized mice at >2 years (Table 1, lines 21 30). Combined CD137 (4-1BB) and adjuvant therapy 327 A second test of the potency of this immunization protocol was a booster given a week after the primary immunization. An impressive >8-fold increase in number was detected with a single boost compared with the primary immunization alone (Table 1, lines 19 and 20). Also, a double booster of anti- CD137 and OVA significantly enhanced memory (Table 1, line 30). These data show that the combination of anti-cd137 and PIC synergistically induced lasting CD8 T cell memory in response to inert antigen. The CD8 + T cell memory pool disperses into peripheral tissues, acquire a survival advantage, produce effector cytokines and clonally expand An important feature of memory T cells is their ability to disperse into peripheral tissues. Therefore, we compared the frequency of specific T cells from spleen, lung and liver from 2-month immunized mice with month (Fig. 1a). The data make two important points with the first being that our immunization protocol robustly induces migration of memory T cells into lung and liver. Secondly, it is clear that as the mice aged the proportion of CD11a hi memory T cells increased as compared with the decline of CD11a lo CD8 T cells (compare Table 1. TLR and CD137 activation profoundly induced generation of peptide-specific memory CD8 T cells Line no. Treatment Immunization and assay dates: month/day/year Gated CD8 splenocytes Absolute number of tetramer + cells Start Finish Percent tetramer + (6 7 days) CD11a hi (10 29 months) 1 Nothing BG BG a n/a 2 Nothing BG BG n/a 3 OVA 8/15/03 12/10/ BG n/a 4 OVA/IgG 8/22/03 12/13/ BG n/a 5 OVA/IgG 8/22/03 12/13/ BG n/a 6 OVA/a-4-1BB 8/15/03 12/10/ BG n/a 7 OVA/a-4-1BB 8/22/03 12/13/ OVA/a-4-1BB 8/22/03 12/13/ BG n/a 9 OVA/IgG/PIC 8/22/03 12/13/ BG n/a 10 OVA/IgG/PIC 8/22/03 12/13/ OVA/a-4-1BB/PIC 9/18/04 8/25/ OVA/a-4-1BB/PIC 9/18/04 8/25/ OVA/a-4-1BB/PIC 9/18/04 8/25/ OVA/a-4-1BB/PIC 9/18/04 8/25/ OVA/a-4-1BB/PIC 10/26/04 8/25/ OVA/a-4-1BB/PIC 7/16/03 11/29/ OVA/a-4-1BB/PIC 8/15/03 12/10/ OVA/a-4-1BB/PIC 8/22/03 12/13/ OVA/a-4-1BB/PIC 32 7/29/03 11/10/ OVA/a-4-1BB/PIC 32 7/29/03 11/10/ OVA/IgG 6/14/02 11/16/04 BG BG n/a 22 OVA/IgG 7/30/02 11/03/04 BG BG n/a 23 OVA/IgG/PIC 6/14/02 12/07/ OVA/IgG/PIC/LPS 7/30/02 12/07/ BG n/a 25 OVA/IgG/PIC/LPS 7/30/02 12/07/ BG n/a 26 OVA/a-4-1BB 7/30/02 11/03/ OVA/a-4-1BB/PIC/LPS 8/23/02 11/03/ OVA/a-4-1BB/PIC/LPS 9/25/02 11/16/ OVA/a-4-1BB/PIC/LPS 9/25/02 12/07/ BG n/a 30 OVA/a-4-1BB 33 5/31/02 11/16/04 nd a BG: background, no definable population; n/a: not applicable and nd: no data. The left of column 4 is the percent CD8 + CD11a hi SIINFEKLtetramer + cells in PBL at 6 7 days post-immunization and the right side are results for spleen cells at months. Lines 19, 20 and 30 show results for boosters give a week apart after primary immunization.

4 328 Combined CD137 (4-1BB) and adjuvant therapy Fig. 1. Enforced CD137 co-stimulation induces dispersion of long-lasting CD8 + memory Tcells into lymphoid and non-lymphoid tissues. C57BL/6 mice were immunized with OVA/anti-CD137/PIC 2 months prior, or 16 months prior. (a) The percent of CD8 + cells being CD11a hi SIINFEKLtetramer + for spleen, lung and liver are shown. The spleen data for the control groups (untreated, OVA, OVA/IgG, OVA/IgG/PIC, OVA/anti-CD137) are given in Table 1. (b) The tabulated data of mice immunized 2 months previously with OVA/anti-CD137/PIC were compared with comparably 11- month primed mice. The left panel shows the percent of CD11a hi SIINFEKL-tetramer + CD8 gated cells at week one of priming, the middle panel shows the endpoint percentage and the right panel shows the absolute number. All data are mean 6 SEM from four comparably treated and primed mice. (c) The ability of the memory cells to produce IFNc in response to SIINFEKL peptide (right panels) was compared with media alone (left panels) in a 5-h in vitro intracellular cytokine assay. The upper number is the percent of CD8 + cells that produced IFNc and the number in parenthesis is the percent SIINFEKL-tetramer + of gated CD8 + cells in the spleen of the mouse prior to culture. top panels with bottom). To quantify changes within the OVAspecific population, we contrasted the priming phase and endpoint for 2 3 month versus >11-month immunized mice (Fig. 1b). We compared immunized mice that displayed comparable levels of excellent priming (Fig. 1b, left panel), and found that at 2 3 months the percent of the specific population was less than at >11 months, which was substantiated by absolute numbers (Fig. 1b, middle and right panels). These data suggest that a single immunization through CD137 and PIC, endows the SIINFEKL-specific T cells with an escalating survival advantage. Survival of memory T cells without function may actually be detrimental to the host since they would take up space and deplete resources without performing a protective role. Therefore, we tested if the memory cells were capable of producing IFNc in a 5-h re-stimulation assay followed by intracellular cytokine staining (Fig. 1c). Bulk splenocytes from unimmunized mice, mice immunized 2 or 16 months prior with OVA/anti-CD137/PIC were placed separately into culture and re-stimulated with nothing (media alone) or with SIINFEKL peptide. Since re-stimulation of T cells induces TCR downregulation thereby preventing TCR detection with tetramers, we determined the percent of tetramer-positive T cells before the 5-h culture (the data in parentheses), which is compared with the percent CD8 IFNc double-positive cells after stimulation (the data above the parentheses). No IL-4 was produced in response to SIINFEKL (data not shown). In contrast, CD8 T cells taken from the 2- or 16-month immunized mice produced substantial amounts of IFNc (Fig. 1c). By inference, nearly half of the 6.5% tetramer cells from the 16- month memory population made IFNc (Fig. 1c, bottom panel). To measure memory cell clonal expansion, we used young congenic recipients, so that T cell proliferation was tested without the complication of inefficient APC function in old mice (31). Following adoptive transfer into CD mice, a single dose of SIINFEKL peptide alone was given, and 6 7 days later, spleens were examined for changes in the percent of CD8 + SIINFEKL-tetramer + cells (Fig. 2a). SIINFEKL was used to stringently permit analysis of only the memory cells. For example, inclusion of adjuvants and/or co-stimulation would induce naive T cells and therefore create either help or competition for the target memory cells. To determine the starting population, gated CD8 T cells that were CD11a hi SIINFEKL-tetramer + was determined before recall at the time of transfer (left panels), and compared with the percent gated CD8 T cells bearing CD45.2 SIINFEKL-tetramer after recall (right panels) (Fig. 2b). The fold induction of clonal expansion was between 3.8 and 6.5 in this study (Fig. 2c).

5 Combined CD137 (4-1BB) and adjuvant therapy 329 Fig. 2. Aged CD8 + memory T cells mount excellent recall responses after transfer into young recipients. (a) Schematic representation of the memory transfer experiment using a CD45.1 congenic recipient. Between 1.5 and 3.5 million splenocytes from memory mice were adoptively transferred to CD45.1 recipients and the following day 200 lg SIINFEKL was given i.p. (b) Six or 7 days after recall, the percent CD8 + gated CD SIINFEKL-tetramer + population was determined (right panels) and for comparison the CD8 + gated SIINFEKL-tetramer + percentage before recall at the time of transfer is given (left panels). Data are from a 16-month normal control (top panels), a mouse immunized with OVA/anti- CD137/PIC 16 months prior (bottom panels). (c) The summary of fold clonal expansion after recall is shown for the mice listed in Table 1 on lines 16, 17 and 18. The data are representative of 27 mice that were unimmunized, immunized as shown or immunized in other ways. We also tested the ability of the memory cells to be recalled in intact mice without any adoptive transfer in three separate experiments. In experiment one, mice immunized 7 weeks prior with OVA/IgG/PIC or OVA/anti-CD137/PIC or unimmunized controls were bled on the day of recall to establish the basal level of SIINFEKL-specific T cells (Fig. 3a, day 0). The unimmunized mice did not generate a detectable immune response, but in mice primed with anti-cd137 mab, the percent of CD8 + SIINFEKL-tetramer + cells was more than doubled in blood by day 7 post-recall (Fig. 3a). In contrast, the mice primed without anti-cd137 co-stimulation had no significant increase in SIINFEKL-tetramer + cells following recall, although there was a detectable population both at time zero and on all the days the mice were analyzed. Therefore, these experiments showed that the memory cells were able to clonally expand in response to signal 1. Since in the previous experiments a detectable primary response in the normal control animals was not generated, we used a different experimental system that would allow us to compare a secondary response with a detectable primary response. To accomplish this goal, mice were immunized with OVA/anti-CD137/PIC 2 months (Fig. 3b) or 17 months (Fig. 3c) earlier and compared with age-matched unimmunized mice in response to infection with 10 5 PFU VSV-OVA (30). The primary response by the CD8 + SIINFEKL-tetramer + cells was detectable but the T cells in the memory mice generated a faster response and a higher magnitude of CD8 + SIINFEKLtetramer + cells. These data show the expected enhanced kinetics of memory cell responses and support the conclusion that this immunization protocol may be efficacious for vaccination against pathogens. Testing the direct effect of CD137 co-stimulation on specific T cells Unlike other co-stimulators, CD137 is expressed on many cell types thereby complicating the precise target of anti-cd137 mabs. Our goal was to examine the direct effects of CD137 ligation on T cells during priming to OVA. C57BL/6 (wt) and CD137 ÿ/ÿ mice were injected with OVA, anti-cd137 mab and PIC, and blood was taken on days 3 and 5, and the spleens were harvested 7 days post-injection to analyze the percent of CD8 + T cells being CD11a hi SIINFEKL-tetramer +. There were no detectable SIINFEKL-tetramer + cells 3 days after immunization in either the wt or CD137 ÿ/ÿ mice (Fig. 4b). By day 5 post-immunization, SIINFEKL-tetramer + populations increased in both types of mice. By day 7 the SIINFEKLtetramer + cells in the spleens of CD137 ÿ/ÿ mice had not increased from day 5 levels, but had increased by almost 3-fold in the wt mice. To directly test whether CD137 expression by T cells was required, wt OTI cells or OTI CD137 ÿ/ÿ cells were separately transferred into the recipient congenic Thy1.1 wt mice and a day later immunized with SIINFEKL, anti-cd137 mab and PIC. On day 5 after immunization, spleen or lymph node cells were analyzed for the absolute number of Thy1.1 ÿ OTI cells (Fig. 4a, top panel). The number of CD8 + T cells able to bind SIINFEKL-tetramer from peripheral lymph node was much greater compared with identically treated CD137 ÿ/ÿ OTI recipients. In a separate experiment we analyzed spleen cells using an identical experimental approach and found similar results (Fig. 4a, bottom panel). Importantly, the CD137 ÿ/ÿ OTI cells did expand on day 2 (data not shown), but clearly did not survive. Therefore, these data suggest that CD137 expression

6 330 Combined CD137 (4-1BB) and adjuvant therapy by peptide-specific T cells is necessary for optimal responsiveness to the anti-cd137 agonist mab. Discussion To avoid the classical problem of non-specific inflammatory responses such as those derived from heat-killed or liveattenuated pathogens, a vaccine formulation comprised of PIC, anti-cd137 therapeutic mab and an inert non-replicating antigen was tested. The results show that CD8 memory T cells were capable of robust effector function and were detected over 1 year later. In order for CD137 immunotherapy to be efficacious for vaccine development, the induction of lasting memory T cells with appropriate effector function is essential (7). In this context, survival of T cells is not sufficient for memory since survival coupled to unresponsiveness is likely to be detrimental. For example, surviving T cells unable to be recalled would tap the body for valuable resources and occupy space that could be used by productive memory cells. Thus, we set out to stringently test whether anti-cd137 immunotherapy would be efficacious in generating lasting, functional memory that would be retained in aged mice. Our data are comprehensively summarized in Table 1 showing various treatments, followed by measurements of primary responses, frequency and numbers of surviving peptide-specific T cells. Several important ideas are revealed by this data with one being that the combination of PIC with CD137 co-stimulation synergistically induced lasting survival of memory CD8 T cells (Table 1). The mechanism of this synergism is currently unknown, but perhaps it is dependent on one of these stimuli regulating the other. For example, PIC or other adjuvants may induce and maintain CD137 expression on the specific T cell allowing for more binding time with the antibody. On the other hand, it is possible that CD137 stimulation induces TLR3 expression on the stimulated T cells. This notion is consistent with previously published data showing that PIC can induce T cell survival (32). A second observation is that the magnitude of priming did not always determine the magnitude of memory. In fact, the booster immunization did not enhance priming compared with no booster, but boosting substantially enhanced the frequency and absolute number of memory cells by several fold (Table 1, compare lines with 14 15). This is an example of how clonal expansion or proliferation can be separated from survival, as we have suggested in the past (33). Nonetheless, excellent priming after anti-cd137/pic seemed to dictate the robustness of memory (Table 1, compare lines with 14 18). Ultimately, the surviving cells were dispersed throughout lymphoid and non-lymphoid tissues (Fig. 1). This is likely to be important for acquisition of functional memory since infection Fig. 3. Aged CD8 + memory T cells clonally expand in response to peptide or virus infection. (a) Unimmunized controls (open square) or immunized 7 weeks prior with OVA/IgG/PIC (open circle) or OVA/anti- CD137/PIC (filled circle) were bled for the starting SIINFEKL-tetramer + population then were given 200 lg SIINFEKL i.p. Peripheral blood was analyzed on days 3, 4, 5 and 7 after recall to determine the percent of CD8 + cells that were CD11a hi SIINFEKL-tetramer +. Data is the mean 6 SEM of three to four mice. (b) In experiment 2, C57BL/6 mice immunized 2 months prior with OVA/anti-CD137/PIC (filled circle) or unimmunized age-matched controls (open square) were given 10 5 PFU s of recombinant VSV-OVA i.v. route. Blood was taken on days 3, 4, 5, 6 and 8 post-infection and CD8 + cells analyzed for percent CD11a hi SIINFEKL-tetramer +. Data are the mean 6 SEM of four mice. (c) In the third experiment an identical VSV-OVA scheme as described in (b) was used, except using 17-month normal mice (open square) and mice immunized 17 months previously with OVA/anti-CD137/PIC (filled circle). Data are the mean 6 SEM of four mice.

7 Combined CD137 (4-1BB) and adjuvant therapy 331 may be localized to a specific peripheral organ (1). Additionally, we found suggestive evidence that the frequency of surviving cells appeared to increase over time after immunization (Fig. 1b). A possible explanation for the continued increase in frequency of specific T cells may be derived from the poor production of naive T cells from the thymii in aged mice as indicated by the reduction of CD11a low-expressing cells (Fig. 1a). Nevertheless, the absolute number of SIINFEKLspecific memory T cells also increased, suggesting additional processes like enhanced competition by surviving memory for resources and space (Fig. 1b). Ultimately, these data need to be extrapolated into other systems in order to form a broad conclusion regarding escalation of numbers of memory cells over time. Finally, we showed that the lasting memory cells were capable of potent recall responses. The fact that aged APCs may not be as functional as APCs from younger mice was examined using a transfer assay (31). The data show pronounced clonal expansion in response to SIINFEKL recall demonstrating that the lasting memory cells were very capable of clonal expansion (Fig. 2). Further testing showed that even recall in response to SIINFEKL peptide alone without any transfer induced expansion (Fig. 3a), but recall responsiveness to VSV-OVA infection was massive (Fig. 3b and c). Specifically, the former case may be more a stringent test but less likely to be physiological, whereas infection perhaps best mimics natural circumstances. As a measurement of efficacious effector function we chose to examine IFNc production in response to recall with SIINFEKL peptide. This is an important parameter since clonal expansion without acquisition of effector function would likely hamper vaccination efforts against pathogens. Peptidespecific T cells taken from the 16-month immunized mice produced substantial amounts of IFNc after in vitro recall (Fig. 1c). We estimated that over half of the SIINFEKL-tetramer + cells synthesized IFNc after in vitro recall, and this data were consistent with IFNc data from other mice immunized in the same fashion as well as the recipient mice shown in Fig. 2 (data not shown). Collectively, these results support our conclusion that enforcing stimulation of CD137 and TLR3 induces lasting functional memory CD8 T cells after only a single immunization with inert antigen. To better understand how the immunizing regimen was working, we examined the initiation of the response. During the primary response the peptide-specific T cells in the CD137 ÿ/ÿ mice exhibited weakened clonal expansion compared with T cells in wt mice (Fig. 4). One interpretation of this data suggests that non-t cells in the CD137 ÿ/ÿ mice lack CD137 Fig. 4. CD137 co-stimulation enhances the accumulation of antigenspecific CD8 + T cells. (a) One million wt OTI cells (filled square) or CD137 ÿ/ÿ OTI (open square) cells were separately transferred into Thy1.1 recipients and the following day 100 lg SIINFEKL/100 lg anti- CD137/150 lg PIC. The inguinal, axillary and brachial peripheral lymph nodes (top panel) from five recipients in each group were analyzed for the number of CD8 + Thy1.1 ÿ SIINFEKL-tetramer + cells 5 days post-immunization. In a separate experiment spleen cells were analyzed (lower panel). (b) C57BL/6 (filled bars) or CD137 ÿ/ÿ (open bars) mice were given 1 mg OVA/100 lg anti-cd137/150 lg PIC and peripheral blood lymphocytes was analyzed on days 3 and 5, and spleen cells on day 7 post-immunization for the percent of CD8 + being CD11a hi SIINFEKL-tetramer +. Data are the mean 6 SEM combined from two separate experiments analyzing n ¼ 4 C57BL/6 mice and n ¼ 6 CD137 ÿ/ÿ mice.

8 332 Combined CD137 (4-1BB) and adjuvant therapy and therefore could not provide all the appropriate signals from optimal expansion. This is particularly relevant since many different cell types like DCs, FDCs and NK cells can directly respond to CD137 stimulation (22 25). Alternatively, stimulation of CD137 on the peptide-specific T cells may be essential for this response. To test this idea, OTI and OTI CD137 ÿ/ÿ T cells were transferred into wt recipient mice and then immunized. The CD137-expressing OTI T cells accumulated to a much greater degree compared with the nonexpressing cells (Fig. 4a). Therefore, potential to express CD137 by T cells led to a greatly enhanced response and this data extend previous observations demonstrating that CD137 co-stimulation is crucial for the late primary response of peptide-specific CD8 T cells after influenza infection (17, 18, 34, 35). The mechanism of how the memory pool is maintained and how effector function is preserved is unknown, but perhaps direct CD137 signaling by the anti-cd137 therapeutic mab conditions the target cells to acquire a unique genetic program pre-disposed to survival and IFNc synthesis. For example, CD137 may induce expression of IL-7R or IL-15R, thereby providing a mechanism for maintenance of CD8 memory T cells (36, 37). Our study demonstrates that under the appropriate conditions, anti-cd137 therapeutic mab potently generates functional memory cells. This data may provide new insight into better vaccine practices, and perhaps more effective approaches for immunizing aged populations. Acknowledgements This work was supported by the National Institutes of Health (NIH) grants AI and AI (A.T.V.); AI 41576, DK and AI (L.L.); AI (M.C.); SRC Fund to Immunomodulation Research Center at University of Ulsan from KOSEF (B.S.K.) and L.M. was supported by NIH training grant T32-AI Abbreviations APC antigen presenting cells BSS balanced salts solution DC dendritic cell FDC follicular dendritic cells i.p. intra-peritoneally i.v. intravenous NIH National Institutes of Health OVA ovalbumin peptide PBL peripheral blood lymphocyte PFU plaque-forming unit PIC polyinosinic polycytidylic TLR toll-like receptor VSV-OVA vesicular stomatitis virus expressing ovalbumin peptide References 1 Kaech, S. M., Wherry, E. J. and Ahmed, R Effector and memory T-cell differentiation: implications for vaccine development. Nat. Rev. Immunol. 2: Reinhardt, R. L., Khoruts, A., Merica, R., Zell, T. and Jenkins, M. K Visualizing the generation of memory CD4 T cells in the whole body. Nature 410: Sallusto, F., Lenig, D., Forster, R., Lipp, M. and Lanzavecchia, A Two subsets of memory T lymphocytes with distinct homing potentials and effector functions. Nature 401: Masopust, D., Vezys, V., Marzo, A. L. and Lefrancois, L Preferential localization of effector memory cells in nonlymphoid tissue. Science 291: Linton, P. J., Haynes, L., Klinman, N. R. and Swain, S. L Antigen-independent changes in naive CD4 T cells with aging. J. Exp. Med. 184: Ginaldi, L., De Martinis, M., D Ostilio, A., Marini, L., Loreto, M. F., Corsi, M. P. and Quaglino, D Cell proliferation and apoptosis in the immune system in the elderly. Immunol. Res. 21:31. 7 Croft, M Co-stimulatory members of the TNFR family: keys to effective T-cell immunity? Nat. Rev. Immunol. 3: Weinberg, A. D., Vella, A. T. and Croft, M OX-40: life beyond the effector T cell stage. Semin. Immunol. 10: Vinay, D. S. and Kwon, B. S Role of 4-1BB in immune responses. Semin. Immunol. 10: Gramaglia, I., Weinberg, A. D., Lemon, M. and Croft, M Ox-40 ligand: a potent costimulatory molecule for sustaining primary CD4 T cell responses. J. Immunol. 161: Kwon, B. S. and Weissman, S. M cdna sequences of two inducible T-cell genes. Proc. Natl Acad. Sci. USA 86: Pollok, K. E., Kim, Y. J., Zhou, Z., Hurtado, J., Kim, K. K., Pickard, R. T. and Kwon, B. S Inducible Tcell antigen 4-1BB. Analysis of expression and function. J. Immunol. 150: Shuford, W. W., Klussman, K., Tritchler, D. D. et al BB costimulatory signals preferentially induce CD8+ Tcell proliferation and lead to the amplification in vivo of cytotoxic T cell responses. J. Exp. Med. 186: Takahashi, C., Mittler, R. S. and Vella, A. T Cutting edge: 4-1BB is a bona fide CD8 Tcell survival signal. J. Immunol. 162: Gramaglia, I., Cooper, D., Miner, K. T., Kwon, B. S. and Croft, M Co-stimulation of antigen-specific CD4 T cells by 4-1BB ligand. Eur. J. Immunol. 30: Cannons, J. L., Lau, P., Ghumman, B., DeBenedette, M. A., Yagita, H., Okumura, K. and Watts, T. H BB ligand induces cell division, sustains survival, and enhances effector function of CD4 and CD8 T cells with similar efficacy. J. Immunol. 167: Dawicki, W. and Watts, T. H Expression and function of 4-1BB during CD4 versus CD8 T cell responses in vivo. Eur. J. Immunol. 34: Bertram, E. M., Lau, P. and Watts, T. H Temporal segregation of 4-1BB versus CD28-mediated costimulation: 4-1BB ligand influences T cell numbers late in the primary response and regulates the size of the Tcell memory response following influenza infection. J. Immunol. 168: Tan, J. T., Whitmire, J. K., Ahmed, R., Pearson, T. C. and Larsen, C. P BB ligand, a member of the TNF family, is important for the generation of antiviral CD8 T cell responses. J. Immunol. 163: Halstead, E. S., Mueller, Y. M., Altman, J. D. and Katsikis, P. D In vivo stimulation of CD137 broadens primary antiviral CD8+ T cell responses. Nat. Immunol. 3: Melero, I., Shuford, W. W., Newby, S. A. et al Monoclonal antibodies against the 4-1BB T-cell activation molecule eradicate established tumors. Nat. Med. 3: Lindstedt, M., Johansson-Lindbom, B. and Borrebaeck, C. A Expression of CD137 (4-1BB) on human follicular dendritic cells. Scand. J. Immunol. 57: Wilcox, R. A., Chapoval, A. I., Gorski, K. S. et al Cutting edge: expression of functional CD137 receptor by dendritic cells. J. Immunol. 168: Futagawa, T., Akiba, H., Kodama, T., Takeda, K., Hosoda, Y., Yagita, H. and Okumura, K Expression and function of 4-1BB and 4-1BB ligand on murine dendritic cells. Int. Immunol. 14: Melero, I., Johnston, J. V., Shufford, W. W., Mittler, R. S. and Chen, L NK1.1 cells express 4-1BB (CDw137) costimulatory molecule and are required for tumor immunity elicited by anti-4-1bb monoclonal antibodies. Cell. Immunol. 190: Hogquist, K. A., Jameson, S. C., Heath, W. R., Howard, J. L., Bevan, M. J. and Carbone, F. R T cell receptor antagonist peptides induce positive selection. Cell 76: Kwon, B. S., Hurtado, J. C., Lee, Z. H. et al Immune responses in 4-1BB (CD137)-deficient mice. J. Immunol. 168:5483.

9 Combined CD137 (4-1BB) and adjuvant therapy Unkeless, J. C Characterization of a monoclonal antibody directed against mouse macrophage and lymphocyte F c receptors. J. Exp. Med. 150: Julius, M. H., Simpson, E. and Herzenberg, L A rapid method for the isolation of functional thymus-derived murine lymphocytes. Eur. J. Immunol. 3: Kim, S. K., Reed, D. S., Olson, S., Schnell, M. J., Rose, J. K., Morton, P. A. and Lefrancois, L Generation of mucosal cytotoxic T cells against soluble protein by tissue-specific environmental and costimulatory signals. Proc. Natl Acad. Sci. USA 95: Plowden, J., Renshaw-Hoelscher, M., Gangappa, S., Engleman, C., Katz, J. M. and Sambhara, S Impaired antigeninduced CD8+ T cell clonal expansion in aging is due to defects in antigen presenting cell function. Cell. Immunol. 229: Gelman, A. E., Zhang, J., Choi, Y. and Turka, L. A Toll-like receptor ligands directly promote activated CD4+ T cell survival. J. Immunol. 172: Maxwell, J. R., Ruby, C., Kerkvliet, N. I. and Vella, A. T Contrasting the roles of costimulation and the natural adjuvant lipopolysaccharide during the induction of T cell immunity. J. Immunol. 168: Bertram, E. M., Dawicki, W., Sedgmen, B., Bramson, J. L., Lynch, D. H. and Watts, T. H A switch in costimulation from CD28 to 4-1BB during primary versus secondary CD8 T cell response to influenza in vivo. J. Immunol. 172: Cooper, D., Bansal-Pakala, P. and Croft, M BB (CD137) controls the clonal expansion and survival of CD8 Tcells in vivo but does not contribute to the development of cytotoxicity. Eur. J. Immunol. 32: Schluns, K. S., Kieper, W. C., Jameson, S. C. and Lefrancois, L Interleukin-7 mediates the homeostasis of naive and memory CD8 T cells in vivo. Nat. Immunol. 1: Lodolce, J. P., Boone, D. L., Chai, S., Swain, R. E., Dassopoulos, T., Trettin, S. and Ma, A IL-15 receptor maintains lymphoid homeostasis by supporting lymphocyte homing and proliferation. Immunity 9:669.