CD4+ T cells enhance tumor destruction by CD8+ T cells. One benefit that underlies CD4+ T cell help is enhanced clonal expansion of newly activated CD8+ cells. In addition, tumor-specific CD4+ help is also associated with the accumulation of greater numbers of CD8+ T cells within the tumor. Whether this too is attributable to the effects of help delivered to the CD8+ cells during priming within secondary lymphoid tissues, or alternatively is due to the action of CD4+ cells within the tumor environment has not been examined. In this study, we have evaluated separately the benefits of CD4+ T cell help accrued during priming of tumor-specific CD8+ T cells with a vaccine, as opposed to the benefits delivered by the presence of cognate CD4+ cells within the tumor. The presence of CD4+ T cell help during priming increased clonal expansion of tumor-specific CD8+ T cells in secondary lymphoid tissue; however, CD8+ T cells that have low avidity for tumor Ag were inefficient in tumor invasion. CD4+ T cells that recognized tumor Ag were required to facilitate accumulation of CD8+ T cells within the tumor and enhance tumor lysis during the acute phase of the response. These experiments highlight the ability of tumor-specific CD4+ T cells to render the tumor microenvironment receptive for CD8+ T cell immunotherapy, by facilitating the accumulation of all activated CD8+ T cells, including low-avidity tumor-specific and noncognate cells.
Cancer vaccines and T cell immunotherapy face numerous hurdles. Because many tumor Ags are also self Ags, central and peripheral tolerance mechanisms, such as anergy, deletion, and immune regulation, limit the function and numbers of available tumor-reactive T cells (1, 2, 3, 4). However, even when large numbers of tumor-reactive CTLs from the endogenous repertoire are successfully expanded by a suitable vaccine, or when such CTLs are provided by adoptive immunotherapy, this does not always result in an effective antitumor response (5, 6). The discrepancy between large systemic numbers of tumor-specific CTLs and low rates of tumor clearance may, in part, be due to impaired trafficking into tumors (7, 8). Extravasation of CD8+ T cells is greatly enhanced at sites of inflammation, and methods that stimulate inflammation at the site of tumor growth, such as irradiation (9), instillation of bacillus Calmette-Guerin (10), or provision of TLR ligands (11), can improve the efficacy of tumor destruction by the immune system.
Because most solid tumors express MHC class I, but not class II, it has been assumed that class I-restricted CD8+ T cells are the primary means by which T cells destroy tumors. However, CD4+ T cells may play a critical role in orchestrating the antitumor response. CD4+ T cells can exert antitumor activity that is independent of CD8+ T cells by recruiting and activating macrophages and eosinophils, which produce tumor-destroying free radicals (12, 13). IFN-γ secreted by CD4+ Th1 cells may have antitumor (12, 14) or antiangiogenic effects (15). In contrast, CD4+ T cells also enhance the activity of CD8+ T cells against tumors (16, 17, 18). This aspect of CD4-CD8 cooperation (19) can be conceptually divided into priming events that occur within secondary lymphoid organs and postpriming events in the tissue parenchyma. Within lymph nodes that drain the tumor, dendritic cells (DCs)3 present cognate Ag to CD8+ T cells. Whether this interaction leads to the activation of a primary CD8+ T cell response depends on functional changes that occur in maturing DCs (20). DC maturation can result from the recognition of pathogen-associated molecular patterns (21), or immune danger signals (22), such as those provided by a cancer vaccine. Additionally, activated CD4+ T cells can also deliver maturational signals to DCs presenting cognate Ag (23), such as those delivered through CD40-CD40L interactions (24). However, even though CD8+ T cells primed with a stimulus that provides a potent source of danger signals acquire functionally similar characteristics regardless whether CD4 help is also provided during the acute, primary, phase of the response, “unhelped” cells later develop functional deficits (19). In contrast, CD8+ T cells previously primed in the presence of cognate CD4+ T cell help expand vigorously (25, 26, 27), produce larger cytokine responses, and show greater resistance to cell death (28) upon secondary encounter with Ag during the recall response. In the setting of a solid tumor, these mechanisms could significantly enhance tumor lytic activity.
Whether tumor-specific CD4+ T cells also deliver postpriming benefits to CD8+ T cells within the tumor environment has received less scrutiny. It has been reported that CD4+ T cells augment the recruitment of tumor-specific CD8+ T cells into tumors (17, 18). But it was not clear whether the observed increase in CD8+ T cell numbers within the tumor was primarily due to CD4+ T cell help delivered during the priming of CD8+ T cells within secondary lymphoid organs causing enhanced clonal expansion and a systemic increase in CD8+ T cell numbers, or alternatively, could be attributed to the activity of CD4+ cells at the site of the tumor.
To distinguish between these two possibilities, we studied the antitumor CD8+ T cell response primed with a cancer vaccine in a murine model of spontaneous tumor formation. Upon provision of tumor-specific CD4+ T cell help, vaccine-primed CD8+ T cells underwent enhanced clonal expansion (29), both CD4+ and CD8+ cells accumulated in large numbers within the tumor, and tumor destruction was enhanced. In contrast, although nontumor-specific CD4+ cells could deliver efficacious help during CD8+ priming as assessed by greatly expanded numbers of CD8+ cells, the CD8+ cells did not accumulate within the tumor and did not promote efficient tumor destruction. Our results suggest that in addition to providing help for priming within secondary lymphoid tissues, the inflammation associated with tumor-specific CD4+ T cells enhances the accumulation of CD8+ T cells within the tumor to promote tumor eradication.
Materials and Methods
B10.D2 rat insulin promotor (RIP)-Tag2-hemagglutinin (HA) and InsHA mice on a B10.D2 background (B10.D2 InsHA mice) have been previously described (30), and were used at 8 wk of age. All B10.D2 RIP-Tag2-HA mice used as recipients were generated by breeding B10.D2 RIP-Tag2+/− mice with B10.D2 HA+/+ mice, and were SV40 T Ag+/− and HA+/−. B10.D2, clone 1 TCR mice, bred with the congenic markers Thy1.1 or Ly5.1, were used between 8 and 14 wk of age. These mice express a TCR specific for HA518–526 (IYSTVASSL) in the context of H-2Kd. SFE mice (also known as 6.5 TCR mice) express a TCR that recognizes HA110–119 (SFERFEIFPK) in the context of I-Ed. B10.D2 SFE and B10.D2 Thy1.1+ SFE mice were used between 6 and 8 wk of age. B10.D2 DO11.10 mice express a TCR that recognizes OVA323–339 in the context of I-Ad, and were used when they were between 6 and 8 wk of age. All animal experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee of the Scripps Research Institute.
Recombinant vaccinia viruses were amplified using HeLa cells, and titered on 143 thymidine kinase-deficient cells. Vaccinia Ova-HA2 was designed to express the full-length chicken OVA open reading frame (ORF), fused in frame with a Gly-Gly-Ser linker and the HA2 domain of the HA ORF from Influenza A/PR8/1934 (starting with the residues GFFGAI and ending with the residues QCRICI). The following PCR primers were used to generate this construct: OVA 5′ primer, 5′-ATTATAGGTACCATGGGCTCCATCGGCGCAG-3′; OVA 3′-HA2 primer, 5′-CAAATAGACCAGATCCTCCAGGGGAAACACATCTGCCAAAGAAG-3′; OVA-HA2 5′ primer, 5′-CTTCTTTGGCAGATGTGTTTCCCCTGGAGGATCTGGTCTATTTGGAGCCATTGCCG-3′; and HA2 3′ primer, 5′-TTATTAACTAGTCAGATGCATATTCTACACTGC-3′. The plasmids pAC-neo-OVA, contributed by M. Bevan (University of Washington, Seattle, WA), and pKG10, contributed by D. Morgan (University of Bristol, Bristol, U.K.), were used as the templates for the PCR of the OVA and HA domains, respectively. The OVA-HA2 ORF was cloned into the shuttle vector, pSC11, to generate pSC11 OVA-HA2, and this was used to generate vaccinia OVA-HA2 by homologous recombination, as previously described (31
Vaccinia SFE was similarly constructed by using PCR to clone the HA1 domain (residues MKANLL to IPSIQS, followed by the insertion of a premature stop codon) of the HA ORF from Influenza A/PR8/1934 into pSC11 to generate the shuttle vector, pSC11 SFE. The following PCR primers were used: PR8HA1 5′ primer, 5′-ATAATCGGTACCATGAAGGCAAACCTACTGGTCCTG-3′; PR8HA1 3′ primer, 5′-TAAAATACTAGTTCAGGATTGAATGGACGGAATGTTC-3′. pSC11 SFE was subsequently used to generate vaccinia SFE, as described above.
Vaccinia OVA was a gift from T. Shin and D. Pardoll (Johns Hopkins School of Medicine, Baltimore, MD), who had originally obtained the virus from J. Yewdell (National Institutes of Health, Bethesda, MD).
Abs and flow cytometry
CD4+ or CD8+ cells were purified from the spleens and lymph nodes of mice by negative selection on a magnetic column using the respective untouched cell purification reagents from Miltenyi Biotec, following the manufacturer’s protocols.
Preparation and adoptive transfer of naive TCR transgenic T cells
Single-cell suspensions were prepared from the lymph nodes and spleens of B10.D2 clone 1 TCR mice and stained for CD8a and either Ly5.1 or Thy1.1, and the concentration of clone 1 cells was derived by FACS analysis.
CD4+ cells were purified from the lymph nodes and spleens of B10.D2 SFE mice, stained with clonotypic anti-SFE TCR Ab, goat anti-rat IgG secondary Ab, and anti-CD4 Ab. The concentration of SFE cells was determined by FACS analysis.
The indicated amounts of T cells and vaccinia virus were mixed in HBSS and injected through the tail vein.
Blood glucose monitoring
Mice were monitored for glucose levels at the indicated time points with an Accu-Check Advantage glucometer and Accu-Check comfort curve test strips (Roche), using blood samples obtained from retro-orbital eye bleeds. Mice were anesthetized with isoflurane during the procedure. Upper limit of glucose detection was 600 mg/dl; lower limit of detection was 10 mg/dl.
Mice were labeled in vivo by i.p. injection with 1 mg of buffered BrdU solution. Cells were processed and stained using the FITC BrdU flow kit from BD Biosciences.
Confocal immunofluorescence microscopy was performed on a Bio-Rad MRC1024 laser-scanning confocal microscope (Zeiss) housed in the microscopy core facility at the Scripps Research Institute. Images were analyzed using ImagePro 3DS version 5.1 (Media Cybernetics).
In vitro analysis of cells from the pancreas
6 cells was incubated overnight with 2 μg/ml HA518–526 peptide and 1 μg/ml Golgiplug, fixed, permeabilized, stained, and analyzed by FACS. Cells in control wells were cultured in the absence of HA peptide.
CD4+ cells were first purified from the spleen of donor animals by negative selection on a Miltenyi column. These cells were subsequently stained with anti-SFE TCR hybridoma supernatant, allophycocyanin goat anti-rat IgG Ab (BD Biosciences), PE anti-mouse CD4 (BD Biosciences), FITC anti-mouse CD8a (BD Biosciences), and FITC anti-B220 (BD Biosciences). Cell sorting was performed on either a FACSAria or FACSVantage (BD Biosciences). FITC-positive cells were assigned to a dump gate, and PE-positive cells were sorted on the basis of TCR expression into CD4+ SFE TCR+ and CD4+ SFE TCR− populations.
The provision of HA-specific CD4+ T cells augments the antitumor activity of clone 1 CD8+ T cells
B10.D2 RIP-Tag2-HA mice express both the HA protein of PR8 and a transforming agent, the large SV40 T Ag, as unlinked transgenes under the control of the RIP (30). As a result, the mice spontaneously develop insulinomas that express HA as a tumor-associated self Ag. The dysregulated secretion of insulin eventually leads to death from complications associated with hypoglycemia. Conversely, because tumor destruction is associated with loss of insulin-producing cells and the development of hyperglycemia, blood glucose levels can be used as a surrogate marker of tumor destruction.
The CD8+ T cells used in these studies were obtained from the clone 1 TCR transgenic mouse that expresses an HA-specific TCR with relatively low avidity for this tumor-expressed Ag (32). We reported previously that following the transfer of CD8+ T cells from HA-specific clone 1 TCR mice into B10.D2 RIP-Tag2-HA mice, detectable tumor ablation occurred only if the cells were first activated in vivo by immunizing the mice with PR8. The extent and duration of tumor destruction were much less than observed with high-avidity clone 4 TCR cells, which resulted in complete tumor ablation. The efficacy of tumor lysis by clone 1 cells was augmented greatly by the cotransfer of HA-specific SFE CD4+ T cells (32). We now show that similar results can be obtained using a recombinant vaccinia virus expressing the HA protein (vacHA) instead of PR8 (Fig. 1⇓). B10.D2 RIP-Tag2-HA mice that received clone 1 cells and vacHA (Fig. 1⇓A) were compared with recipient mice that received clone 1, SFE CD4+ T cells and vacHA (Fig. 1⇓B). More mice in the latter group responded with sufficient tumor destruction to develop frank hyperglycemia (3 of 6 vs 6 of 6 mice), and a more durable response was achieved in the presence of SFE CD4+ T cells (median areas under the curve from weeks 0 to 10: 1552 vs 5629 mg wk/dl). Moreover, we noted an early kinetic difference in the glucose response that was not explored in our earlier studies. Among the animals in Fig. 1⇓A that developed hyperglycemia, blood glucose levels only became elevated during the second week following activation of clone 1 cells with vacHA; in contrast, all of the animals that also received SFE CD4+ cells were hyperglycemic within the first week (insets to Fig. 1⇓). This difference was associated with the provision CD4+ T cell help for the antitumor activity of clone 1 cells. B10.D2 RIP-Tag2-HA mice given SFE cells and virus, but not clone 1, do not develop hyperglycemia (data not shown) (32). The underlying basis for CD4+ T cell enhancement of tumor destruction by clone 1 cells within the first week of immunotherapy was further investigated.
Comparison of clone 1 cells activated in the presence or absence of SFE
Because the enhanced tumor lysis associated with the cotransfer of SFE cells is apparent within the first week following T cell activation, we looked for qualitative differences in the clone 1 cells at 4 days following T cell activation by vacHA that could explain the impact of SFE cells. We found that in the absence or presence of SFE cells, clone 1 cells developed into effector T cells that displayed similar changes in acute activation markers both in the spleen and pancreas (Fig. 2⇓A). There was comparable down-regulation of CD62L and up-regulation of CD25 and CD44. Functionally, SFE-helped and SFE-unhelped clone 1 cells in the spleen (Fig. 2⇓B) and pancreas (Fig. 2⇓C) produced similar levels of IFN-γ following ex vivo peptide stimulation.
Because there did not appear to be a qualitative basis for the enhanced tumor lytic effect of clone 1 cells primed in the presence of SFE cells, we next considered the possibility of a quantitative effect. The priming of clone 1 cells by vacHA in the presence of SFE was associated with a small, but consistent increase in clone 1 cells as a percentage of total CD8+ T cells in the spleen (Fig. 3⇓A). Because the absolute CD8+ T cell numbers had increased in response to vaccinia virus, this represented a 2- to 3-fold increase in the total number of clone 1 cells in the spleen at the peak of the response on day 4 (Fig. 3⇓B). To determine whether this enhancement in numbers was due to increased clonal expansion, BrdU was administered to recipient mice that had been given clone 1 and vacHA, with or without SFE cells. A somewhat larger proportion of clone 1 cells incorporated BrdU in the group that received SFE cells (65 vs 57%), suggesting that cell turnover was slightly enhanced in this group of mice (Fig. 3⇓C).
SFE cells enhance the accumulation of clone 1 cells in the tumor
To determine whether the increase in clone 1 numbers observed in the spleen also reflected the situation in the pancreas, pancreatic sections obtained from B10.D2 RIP-Tag2-HA mice given Ly5.1+ clone 1 cells and vacHA, with or without SFE cells, were examined by confocal immunofluorescence microscopy. In the former group of mice, we found a prominent infiltrate of Ly5.1+ cells centered round the insulin-expressing pancreatic islets (Fig. 4⇓A, and measured as described in Fig. 4⇓B). Because the infiltrate was too dense to obtain an accurate count, we repeated the experiment using fewer clone 1 cells. At all input clone 1 numbers, we detected more clone 1 cells in the pancreas of mice that received SFE cells as early as 4 days after transfer of cells and vacHA. To obtain a more quantitative measurement of the enhanced recruitment of clone 1 cells into the pancreas, the numbers of CD8+ Ly5.1+ clone 1 cells present in spleen and pancreatic homogenates were assessed by flow cytometry (Fig. 4⇓C). The level of enhancement in clone 1 numbers associated with the cotransfer of SFE cells was far greater in the pancreas (1,888 vs 17 events; Fig. 4⇓C) than in the spleen of recipient mice (2- to 3-fold increase; Fig. 3⇑B). There was also a large increase in the numbers of endogenous Ly5.1− CD8+ T cells recruited to the pancreas (27,659 vs 823 events; Fig. 4⇓C). The endogenous CD8+ T cells in the pancreas displayed phenotypic surface markers typical of effector T cells, i.e., they were CD62L low, CD43 high, and CD44 high (data not shown), suggesting these may have been activated by the viral infection.
The increase in pancreatic CD8+ T cell numbers 4 days after T cell priming could be due to increased recruitment into, or decreased efflux from the pancreatic parenchyma. Alternatively, it could be due to increased proliferation of T cells or decreased T cell death within this tissue. We compared the turnover of clone 1 cells in the pancreas in the absence or presence of SFE help by labeling recipient mice on day 4 with a brief pulse of BrdU. The presence of SFE help made little difference in the percentage of either clone 1 or endogenous CD8+ T cells that incorporated BrdU, suggesting it was unlikely that differences in the rate of CD8+ T cell turnover were responsible for the observed accumulation of CD8+ cells in the parenchyma at this time point (Fig. 4⇑D).
CD4+ T cells must recognize tumor Ag to enhance accumulation of CD8+ T cells within the tumor
The presence of CD4+ SFE cells increased the numbers of CD8+ T cells in the pancreas far more than the increase observed in secondary lymphoid tissue. This suggested that the major role of SFE help in this tumor model may occur in the tumor rather than during priming. We hypothesized that if this were the case, the provision of CD4+ T cell help would be of less value if it was provided only during priming, and not within the tumor. To address this question, we made use of CD4+ T cells derived from the OVA-specific TCR transgenic mouse, DO11.10. To ensure that DO11.10 cells could deliver CD4+ T cell help to clone 1 during priming, we physically linked the epitopes recognized by clone 1 and DO11.10 so that they could be presented on the same APC (33) by producing a recombinant vaccinia vector, vaccinia OVA-HA2 (vacOVA-HA2), which expresses full-length OVA fused to the HA2 domain from influenza A HA (Fig. 5⇓A). When clone 1 cells were primed in the presence of DO11.10 CD4+ T cells and vacOVA-HA2 recombinant virus, the clone 1 cells expanded even more than when they were primed with vacHA in the presence of SFE (6.4 vs 3.0%; Fig. 5⇓B, top row, middle and right panels). However, 10-fold fewer clone 1 cells were recovered from the pancreas of mice that received DO11.10 and vacOVA-HA2 compared with mice that received clone 1 cells together with SFE and vacHA (94 vs 1028 events; Fig. 5⇓B). Furthermore, in contrast to the enhanced tumor destruction by clone 1 cells observed when mice received SFE CD4+ T cells, DO11.10 cells were unable to assist in tumor destruction (Fig. 5⇓C). Thus, provision of help during priming using CD4+ T cells that are not specific for a tumor Ag did not promote the early accumulation of clone 1 cells in the tumor. This is most likely due to the fact that SFE cells accumulated in large numbers in the pancreas, whereas DO11.10 cells did not (Fig. 5⇓B). In a control experiment, we infected B10.D2 RIP-Tag2-HA mice with vaccinia that expressed enhanced GFP, and could not detect expression of GFP in the pancreas of these animals, suggesting that infection with vaccinia vectors did not lead to significant expression of vaccinia Ags in the pancreas (data not shown).
Activated HA-specific CD4+ T cells were necessary and sufficient for the accumulation of effector CD8+ T cells within the tumor site
In the previous experiments, we observed large numbers of endogenous CD8+ T cells with an effector phenotype in the tumors of mice that received clone 1 and SFE cells, and which were primed with vacHA. This suggested that the presence of tumor-specific CD4+ T cells also enhanced the accumulation of endogenous CD8+ T cells within the tumor. We wanted to evaluate whether adoptive transfer of activated SFE cells was necessary and sufficient for the accumulation of CD8+ T cells within the tumor-bearing pancreas regardless of their Ag specificity. SFE cells were adoptively transferred into B10.D2 InsHA hosts and activated in vivo with vaccinia SFE (vacSFE). CD4+ cells from these mice were sorted into CD4+ SFE TCR+ and CD4+ SFE TCR− populations, each of which was injected into different recipient B10.D2 RIP-Tag2-HA mice. To produce a population of nontumor-specific CD8+ cells, we made use of vacSC8, which expresses β-galactosidase, to prime a polyclonal source of effector CD8+ T cells that were subsequently transferred into the recipient mice (Fig. 6⇓A). We used B10.D2 InsHA mice as the source of the donor CD8+ T cells because these mice have no detectable high-avidity HA-responsive CD4+ or CD8+ T cells remaining within their repertoire (34). When the pancreas of recipient mice was analyzed by flow cytometry to detect CD8+ cells, we found that the pancreas of mice that received previously activated SFE TCR+ cells contained 57-fold more polyclonal effector CD8+ T cells than the pancreas of mice that received SFE TCR− cells (Fig. 6⇓B). This suggests that the observed accumulation of effector CD8+ T cells within the tumor could be solely attributed to events set in motion by activated tumor-specific CD4+ T cells. Moreover, because SFE and polyclonal CD8+ T cells were primed in separate donor animals, this experiment also suggested that unlike CD4+ T cell help for the priming of naive CD8+ T cells (23, 33), this mechanism of augmentation of CD8+ T cell tumor lytic activity does not require linkage between MHC class I and II epitopes.
We have evaluated how CD4+ T cells promote tumor destruction by tumor Ag-specific clone 1 cells within the first week following activation with a potent tumor vaccine. Two mechanisms are operating within this time frame. Tumor-specific CD4+ T cells promote clonal expansion of clone 1 cells by licensing DCs for the cross-presentation of tumor Ag (23). Tumor-infiltrating CD4+ T cells also increase the accumulation of CD8+ T cells in the tumor (17). We designed experiments to distinguish between the relative contributions of CD4+ help for the priming of naive CD8+ T cells in lymph nodes, and the impact of CD4+ T cells on the tumor environment. Our results indicate that CD4+ T cell infiltration into the tumor parenchyma is associated with increased accumulation of CD8+ T cells, and that this is the dominant mechanism by which tumor-specific CD4+ T cells enhance tumor ablation by vaccine-primed CD8+ T cells during this early phase of the antitumor response.
The clone 1 TCR was originally derived from an InsHA mouse infected with influenza A PR8 (32). Due to mechanisms of self-tolerance, all HA-specific CD8+ T cells obtained from these mice, including clone 1, demonstrate low avidity for HA. The clone 1 TCR transgenic mouse was specifically produced to evaluate how the efficacy of tumor immunotherapy using CD8+ T cells with low avidity for Ag could be optimized. CD8+ T cells from clone 1 mice possess properties consistent with the expression of a TCR that has low affinity for HA. It is possible that the observed inefficiency of tumor infiltration by clone 1 cells in the absence of CD4 help may be related to low avidity for tumor Ag. This may restrict infiltration and/or enhance egress due to poor lytic capability and weak association with Ag-bearing tumor, or result in the inefficient production of inflammatory cytokines that may further enhance infiltration (35).
Because the B10.D2 RIP-Tag2-HA recipient mice used in our experiments are replete in the CD4+ T cell compartment, clone 1 cells stimulated with vacHA in the absence of SFE cells could have received help during priming from endogenous CD4+ T cells specific for vaccinia Ags. However, due to the large numbers of clone 1 cells injected, such endogenous CD4+ T cell help might be insufficient to promote optimal activation of clone 1 cells. Even though the published literature suggests that a potent source of immune danger signals (36), like vaccinia virus, can effectively prime primary CD8+ T cell responses in the absence of CD4+ T cell help (37, 38), it is possible that low-avidity clone 1 cells might have more stringent helper requirements for T cell activation (39) that SFE cells can help overcome. However, we did not observe functional deficits in clone 1 cells primed in the absence of SFE cells at the time points assayed (Fig. 2⇑). In contrast, we observed quantitative differences that could account for the disparate biological outcomes depicted in Fig. 1⇑. By day 4 following T cell priming, there was a 10- to 100-fold enhancement in the numbers of CD8+ T cells in the tumor-laden pancreas of mice that received HA-specific CD4+ T cells. The observed accumulation of CD8+ cells in the pancreas could be due to several mechanisms, as follows: increased recruitment, decreased efflux or increased retention, increased lymphocyte division within the pancreas, or decreased cell death. Our data showing BrdU uptake by CD8+ T cells in Fig. 4⇑D suggested that the different extents of CD8+ T cell accumulation were not likely due to a disparity in the rates of cellular turnover. Instead, the observed accumulation of CD8+ T cells is more plausibly due to a net effect of SFE cells on the rates of lymphocyte recruitment to, or egress from the pancreas. However, our data do not distinguish between these latter two possibilities. There was increased accumulation of both HA-specific clone 1 and endogenous CD8+ T cells with an effector phenotype. Because the endogenous CD8+ T cells in RIP-Tag2-HA mice have been tolerized to HA (32), it is unlikely that these cells accumulated in the pancreas on the basis of their reactivity to HA. In addition, there was no important role for the endogenous CD8+ T cells that could have been induced against neo-Ags created by the transforming activity of SV40 large T Ag, because we also observed increased accumulation of both endogenous CD8+ T cells and clone 1 cells in the presence of SFE cells in nontumor-bearing B10.D2 InsHA hosts (data not shown). In fact, despite the specificity of the clone 1 TCR for HA, clone 1 cells did not demonstrate a marked preferential ability to accumulate in the pancreas compared with endogenous CD8+ T cells. In contrast to the large increase in absolute CD8+ T cell numbers in the pancreas of recipient mice that received SFE cells, in all groups the ratio of clone 1 to endogenous CD8+ T cells in the pancreas remained close to the ratio found in the spleen (Figs. 4⇑C and 5⇑B). This HA-nonspecific accumulation of activated CD8+ T cells within tissue parenchyma in which HA-specific SFE cells recognize their cognate Ag suggests an underlying inflammation-driven process initiated by CD4+ T cell activity within the tumor.
Several studies have reported that the tumor microenvironment, including RIP-Tag tumors, can be resistant to lymphocytic infiltration (8, 40, 41, 42). Previous studies have shown that the provision of large numbers of tumor-specific CTLs may not guarantee tumor clearance, and feeble responses have been correlated with poor trafficking of tumor-specific CD8+ T cells into the tumor (7, 8). Although the activation of T cells generally increases the tropism of these cells for nonlymphoid tissue (43), this is usually associated with tissue inflammation and endothelial activation that can help to guide cells through production of chemokines and inflammatory mediators. In the absence of such guidance, TCR specificity alone does not necessarily assure access to parenchymal tissue (reviewed in Ref. 44). Previously, it was shown that in situations wherein tumor-reactive circulating CD8+ T cells appear to ignore Ag-expressing tumor cells, inflammation brought about either by irradiation or the provision of TLR ligands could result in T cell infiltration of tumors (8, 9, 10, 11, 45, 46, 47, 48).
Our studies suggest that cognate, tumor-infiltrating CD4+ T lymphocytes can generate an inflammatory environment within the pancreas that attracts and/or retains CD8+ T lymphocytes regardless of their TCR reactivity. Previous work has implicated IFN-γ and IFN-γ-inducible chemokines as important mediators of lymphocyte trafficking into the pancreas (49, 50, 51, 52). It should be noted that pancreatic islets do not express detectable amounts of MHC class II molecules on the cell surface (53) and cannot directly present Ag to CD4+ T cells. Therefore, SFE recognition of HA within the tumor probably occurs when tissue-resident MHC class II-positive APCs process and present tumor-derived HA to CD4+ T cells. We propose that productive CD4+ T cell-APC interactions in the tissue sites of Ag expression might lead to the elaboration of IFN-γ by CD4+ T cells, which contributes to the recruitment of effector CD8+ T cells to these tissue sites. This effect might be mediated by an IFN-γ-inducible chemokine produced by the tissue, e.g., CXCL-10, operating on activated CXCR3+ CD8+ T cells (50). Other cytokines not necessarily induced by IFN-γ may be also involved in the recruitment of CD8+ T cells to tissues, e.g., CCL2 and CCL5 (54). CD40 signaling can also play a pivotal role in the generation of effective antitumor immunity (55, 56, 57, 58). Although the ability of CD40 agonists or CD154-expressing CD4+ T cells to augment the activity of CD8+ T cells is thought to occur primarily through the activation of an APC intermediary, it is conceivable that direct CD4-CD8 CD154-CD40 interactions (59) within the pancreas might also be a mechanism that allows SFE cells to augment the accumulation of activated CD8+ T cells there.
In summary, this study defines a previously underappreciated mechanism underlying the importance of tumor-specific CD4+ T cells in augmenting the antitumor CD8+ T cell response. This mechanism is distinct from CD4+ T cell help for the licensing of DCs to cross-prime naive CD8+ T cells. CTLs only remain activated for a finite period following priming. Programmed contraction occurs even if Ag persists (60); and with contraction also comes reversion to quiescence (61). Remaining tumor-specific T cells are susceptible to tolerization by persistent tumor Ag (3, 30, 62, 63). Therefore, there is clear advantage to facilitating swift access of CD8+ T cells to tumor. We have shown that tumor-specific CD4+ T cells can facilitate this access by rendering the tumor microenvironment receptive for CD8+ T cell immunotherapy, with important implications for the rational design of cancer vaccines and the optimization of adoptive immunotherapy for tumors. Our results suggest that distinct vaccine and therapeutic strategies designed to optimize CD8+ T cell priming or T cell accumulation within the tumor can be combined in concert to maximize the antitumor CTL response.
We acknowledge the assistance of Alan Saluk, Cheryl Kim, Alex Ilic, and Quyen Nguyen from the FACS Core Facility of the Scripps Research Institute. We also thank Bill Kiosses of the Core Microscopy Center and Diane Kubitz of the Ab Production Core Facility at the Scripps Research Institute. We thank Judith A. Biggs, Kristi Marquardt, and Rebecca Trenney for breeding mice used in these experiments, and technical assistance with immunohistochemistry.
The authors have no financial conflict of interest.
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 This work was supported by Grant R01 CA57855 from National Institutes of Health. S.B.J.W. was funded by a fellowship from the National University of Singapore. R.B. was funded by a Netherlands Organization for Scientific Research Rubicon Grant from The Netherlands.
↵2 Address correspondence and reprint requests to Dr. Linda A. Sherman, Department of Immunology, The Scripps Research Institute, IMM-15, 10550 North Torrey Pines Road, La Jolla, CA 92037. E-mail address:
↵3 Abbreviations used in this paper: DC, dendritic cell; HA, hemagglutinin; ORF, open reading frame; RIP, rat insulin promotor; vacHA, vaccinia virus expressing the HA protein; vacOVA-HA2, vaccinia OVA-HA2; vacSFE, vaccinia SFE.
- Received August 20, 2007.
- Accepted December 26, 2007.
- Copyright © 2008 by The American Association of Immunologists