Abstract
We studied the size and avidity of primary and recall CD8+ T cell responses in vivo in mice immunized with dendritic cells presenting different densities of a MHC class I-restricted peptide. Increasing the epitope density on a fixed number of dendritic cells increased the size of the primary response, yet had no influence on the avidity of the effector cells. However, epitope density-based selection of cells with different avidities was evident in the subsequent memory population, and in recall responses. Additionally, mice primed with different peptide densities had similarly sized quiescent memory and recall responses. Our findings provide evidence for an important role for epitope density in the selection of T cells in vivo.
The development of higher avidity immunological effectors is known to provide enhanced protection against repeat exposure to the Ag (1, 2, 3). Despite the fact that TCRs do not undergo the somatic hypermutation process that results in high affinity B cell receptors, T cells exhibit increased avidity for Ag during the evolution of an immune response. The functional and structural avidities of recall T cells, measured as effector activity and binding to MHC-peptide multimers, respectively, are considerably greater than those of T cells in primary immune responses (4, 5, 6, 7). These results are consistent with the possibility that Ag selects T cells with progressively higher affinity TCRs over the course of the immune response (Ag-driven selection). However, it has also been shown in TCR-transgenic mice that avidity increases during the primary response (8, 9), and between primary and recall CD8+ T cells (10, 11, 12). Thus, avidity maturation can occur independently of the structure and affinity of the TCR, and may involve a developmentally programmed increased association of p56lck (8, 13, 14). However, while Ag initiates expansion and retention of higher avidity T cells, there is little substantial evidence that epitope density-based selection of T cells occurs in vivo.
If Ag-driven selection of T cells with different affinity TCRs occurs, then decreasing the density of epitope presented by dendritic cells (DC)3 should lead to the selective activation of T cells with higher affinity TCRs, and thus, higher avidity. Generation of CD8+ T cell lines by restimulation with low densities of peptide in vitro results in the development of considerably higher avidity populations than those maintained by restimulation with higher peptide densities (1, 2, 3), demonstrating that Ag density-based selection of T cells occurs in vitro. A single study demonstrated a very modest difference in structural avidity of recall T cells from mice immunized with two different doses of peptide in adjuvant (15). However, with the use of this immunogen, the time course, efficiency, and distribution of Ag delivery to lymphoid tissue are unknown. Thus, a systematic and direct evaluation of whether and how the density of epitope displayed on DC influences the avidity of T cells in vivo has not been performed. A second expectation of Ag-driven selection is that an increase in epitope density presented by DC should increase the number of activated T cells. Previously, using DC pulsed with an HLA-A*0201-restricted epitope derived from the melanocyte differentiation protein, tyrosinase (Tyr369(Y): YMDGTMSQV), we determined that graded increases in epitope density resulted in progressively larger numbers of activated primary CD8+ T cells, until supraoptimal levels of epitope density were used (16). It is currently unclear whether increases in the number of activated primary CD8+ T cells results from either reduced competition for access to the MHC-peptide complexes on the DC (5, 17), or the recruitment of lower avidity T cells into the response. In the present study, we have used DC that present controlled densities of epitope to evaluate the influence of epitope density on the functional and structural avidity, respectively, of primary and recall CD8+ T cells in vivo.
Materials and Methods
Animals
Transgenic mice containing a complete deletion of the tyrosinase gene and expressing a chimeric MHC class I composed of the α1 and α2 domains of HLA-A*0201 and the α3 domain of H2-Dd (AAD) (18, 19) were maintained in specific pathogen-free facilities and treated in accordance with guidelines of the University of Virginia Animal Use Committee (Charlottesville, VA).
Cell lines
C1R-AAD (18) was maintained in RPMI 1640 containing 5% FBS supplemented with SerXtend (Irvine Scientific, Santa Ana, CA) and 300 μg/ml G418 (Life Technologies, Grand Island, NY). The human melanomas DM93 (HLA-A*0201+ tyrosinase+) and DM331 (HLA-A*0201+ tyrosinase−) (20) were maintained in the same media without selection agent.
Peptides
Synthetic peptides were made in the University of Virginia Biomolecular Core Facility and purified to >98% purity by reverse-phase HPLC. Purity and identity were confirmed by mass spectrometry.
Dendritic cells
DCs were generated as described (21) with modifications (16). Immature DC were isolated on a StemSep column after incubation with a mixture of Abs that enrich for DC, then were activated at a 2:1 ratio overnight with irradiated CD40 ligand-transfected NIH-3T3 fibroblasts (a gift from Dr. R. Lapointe, National Cancer Institute, Bethesda, MD). Activated DCs expressed high levels of MHC class I, MHC class II, CD80, CD86, CD40, and demonstrated intracellular accumulation of IL-12 in the presence of brefeldin A (Sigma-Aldrich, St. Louis, MO).
Immunization
Mice were immunized i.v. with either 1 × 107 PFU of a recombinant vaccinia virus expressing full-length human tyrosinase (Tyr-vac) (22) or 105 DC that had been pulsed with peptide for 3–4 h at 37°C in HBSS containing 5% FBS and 5 μg/ml human β2 microglobulin (Calbiochem, La Jolla, CA), washed twice, and resuspended in HBSS.
Generation of peptide-specific HLA-A*0201-restricted CD8+ T cells
Spleens from primed mice were harvested at least 3 wk after immunization. Responder cells (1.5 × 107) were incubated in upright 25-cm2 tissue culture flasks (Costar, Cambridge, MA) with 7 × 106 autologous irradiated (2500 rad) spleen cells that had been pulsed with the indicated concentration of peptide for 3–4 h at 37°C. After culture for 6–7 days, cytokine production was assessed as described below.
Intracellular cytokine staining (ICS-IFN-γ)
IFN-γ expression was examined either ex vivo in freshly isolated spleen cells or in short-term cultures maintained in vitro. For primary ex vivo analysis, splenic CD8+ T cells were isolated 7 days postimmunization by incubation with a mixture of Abs to enrich for CD8 cells and passage over a StemSep column (StemCell Technologies, Vancouver, British Columbia, Canada). Preparations were consistently 85–95% CD8+ as assessed by flow cytometry. No enrichment of CD8+ T cells was performed for short-term cultures or recall responses. T cells were then incubated with peptide-pulsed C1R-AAD stimulator cells for 5 h at a ratio of 1:1 in medium supplemented with 50 U/ml IL-2 and 10 μg/ml brefeldin A. Stimulated cells were counterstained with either PE- or allophycocyanin-conjugated anti-CD8 (BD PharMingen, San Diego, CA), washed, then fixed, and permeabilized in PermWash/Fix (BD PharMingen), followed by staining with FITC-conjugated anti-IFN-γ (BD PharMingen). Flow cytometry was performed on a FACSCalibur using CellQuest software (BD Biosciences, San Jose, CA) and the number of cells stained for CD8 and IFN-γ was determined after subtraction of the background level of IFN-γ staining achieved with unpulsed C1R-AAD. Where indicated, normalized values were calculated by using the formula ((experimental value − background value for unpulsed stimulators)/(maximal value (using 100 μg/ml peptide-pulsed stimulators) − background value)) × 100.
Tetramer staining
HLA-A*0201 tetramers that had been folded around YMDGTMSQV (Tyr369(Y)) were produced by the National Institutes of Health Tetramer Facility (Emory University, Atlanta, GA). T cells were coincubated for 45 min at room temperature with the indicated concentration of tetramer and a 1/1000 dilution of anti-CD8, washed twice, and fixed in 1% paraformaldehyde. Staining was quantitated on a FACSCalibur using CellQuest software. For raw data, tetramer staining values of CD8+ T cells from mice immunized with irrelevant Ag were subtracted. Where indicated, normalized staining values were calculated by using the formula ((experimental value − background value for irrelevant Ag)/(maximal value (using 7.2 μg/ml tetramer) − background value)) × 100.
Results
The density of peptide presented by DC affects the number, but not the avidity, of primary CD8+ T cells
We sought to determine the impact of the density of epitope presented by DC on the characteristics of CD8+ T cells elicited during the primary immune response in vivo. AAD+ mice (that express a chimeric MHC class I molecule consisting of the α1 and α2 domains of HLA-A*0201 and the α3 domain of H2-Dd) were immunized with DC that had been pulsed with concentrations of Tyr369(Y) ranging between 0.001 and 10 μg/ml. Pulsing cells with higher concentrations of peptide results in an increase in the density of epitope presented at the surface of the pulsed cell, as judged by either MHC-peptide stabilization assays (3) or staining with MHC peptide-specific mAbs (data not shown). Splenocytes were harvested 7 days after immunization, and CD8+ T cells were enriched by magnetic bead-mediated selection. The expanded CD8+ T cell population was examined directly ex vivo by either staining with Tyr369(Y)-HLA-A*0201 tetramers or by staining for the intracellular accumulation of IFN-γ. The number of Ag-specific CD8+ T cells elicited by DC pulsed with 10 μg/ml Tyr369(Y) was on average 4-fold greater than that activated by DC pulsed with 0.01 μg/ml of this peptide (Fig. 1⇓a). Primary responses elicited by DC pulsed with 0.001 μg/ml Tyr369(Y) were generally below detectable levels for these two assay systems. These results are consistent with our previous observations (3), and confirm that increasing the concentration of Tyr369(Y) pulsed onto the immunizing DC resulted in a concordant increase in the size of the CD8+ T cell response.
Ex vivo analysis of the size and avidity of primary CD8+ T cells activated by DC pulsed with graded concentrations of Tyr369(Y). Cohorts of three AAD+ mice were primed with DC that had been pulsed with the indicated concentration of Tyr369(Y) and then washed. Seven days postimmunization, CD8+ T cells were enriched and either incubated with Tyr369(Y)-pulsed C1R-AAD before staining for ICS-IFN-γ or with Tyr369(Y)-tetramer. a, Maximal size of response for each CD8+ T cell population, defined by incubation with C1R-AAD pulsed with 100 μg/ml Tyr369(Y) and ICS-IFN-γ, or with 7.2 μg/ml tetramer. b, Functional avidity (SC50) of each CD8+ T cell population was determined by incubation with C1R-AAD that had been pulsed with the indicated concentrations of Tyr369(Y), followed by ICS-IFN-γ. SC50 values represent the concentration of Tyr369(Y) pulsed onto C1R-AAD cells required to activate 50% of the T cell population. c, Structural avidity (TC50) of each CD8+ T cell population was determined by incubating each population with serial dilutions of Tyr369(Y)-tetramer. TC50 values represent the concentration of tetramer required to stain 50% of the Tyr369(Y)-specific CD8+ T cells. Where indicated, data are normalized by setting the maximal response within each population to 100%, as described in Materials and Methods.
The primary CD8+ T cell populations were also assessed for differences in avidity. Functional avidity (the concentration of peptide required to activate 50% of the Tyr369(Y)-specific CD8+ T cells (SC50)) was determined by coculturing with stimulator cells that had been pulsed with graded concentrations of peptide, followed by staining for the accumulation of intracellular IFN-γ (ICS-IFN-γ). Structural avidity was determined as the concentration of Tyr369(Y)-HLA-A*0201 tetramer required to stain 50% of the maximum number of tetramer binding CD8+ T cells (TC50). Surprisingly, despite the difference in the number of primary CD8+ T cells activated by DC pulsed with different concentrations of peptide, over multiple experiments we found no discernible trend in the functional and structural avidities of these cells. For example, CD8+ T cells activated by DC pulsed with between 10 and 0.01 μg/ml had average functional avidities that differed by less than a factor of 2 (Fig. 1⇑b). Importantly, in this and other experiments, no systematic relationship between epitope density and the functional avidity of the primary T cell population was apparent. The functional avidity of primary CD8+ T cells generated by immunization with Tyr369(Y)-pulsed DC were very similar to those generated by immunization with Tyr-vac (data not shown). Staining with serial dilutions of Tyr369(Y)-tetramer revealed that all T cells expressed TCR with equivalent structural avidities (Fig. 1⇑c). Over several experiments there was no reproducible difference in the mean fluorescence intensity (MFI) of either IFN-γ or tetramer staining (not shown). The lack of differences observed in functional and structural avidities of these CD8+ T cells analyzed ex vivo at the peak of the primary response contrasts sharply with the 60-fold range of functional avidities that we had previously observed in Tyr369(Y)-specific CD8+ T cell lines maintained in vitro using DC pulsed with between 10 and 0.01 μg/ml peptide (3). Therefore, although changing the epitope density on DC did lead to a change in the number of responding CD8+ T cells in vivo, it did not lead to a change in avidity of the primary effector T cells.
The effect of the density of epitope presented during the primary immunization on the avidity and number of memory CD8+ T cells in vivo
The discrepancy between the effects of epitope density on the avidity of primary CD8+ T cells and T cells in short-term cultures led us to ask whether avidity selection occurred after the peak of the primary response. Therefore, we examined how primary immunization with DC pulsed with different concentrations of Tyr369(Y) affected the characteristics of recall responses measured ex vivo. Our initial approach of priming and boosting with peptide-pulsed DC led to Tyr369(Y)-specific CD8+ T cell recall responses that were not substantially higher than primary responses (Fig. 2⇓a). In contrast, using Tyr-vac for challenge of DC-primed mice resulted in Tyr369(Y)-specific recall responses that were substantially greater than primary responses, as would be expected of a recall response (Fig. 2⇓a). Therefore, we used Tyr-vac challenge to gauge the influence of peptide density used for priming on the avidity of recall responses.
Size and avidity of primary and recall T cells after immunization with DC pulsed with Tyr369(Y). a, Cohorts of four AAD+ mice were primed with either DC pulsed with 10 μg/ml Tyr369(Y) or Tyr-vac and challenged after 3 wk with either peptide-pulsed DC or Tyr-vac. Recall responses were evaluated 5 days postchallenge by ICS-IFN-γ and Tyr369(Y) tetramer. Naive (primary) or AAD+ mice that have been previously immunized with Tyr-vac (recall) were immunized with DC that had been pulsed with 10 μg/ml Tyr369(Y) and analyzed 5 days (recall) or 7 days (primary) post-DC immunization. The number of responding CD8+ T cells (b) was determined by ICS-IFN-γ using C1R-AAD cells that had been pulsed with 100 μg/ml Tyr369(Y), or by incubation with 7.2 μg/ml Tyr369(Y) tetramer. Functional (c) and structural (d) avidities of each response were determined as described in the legend for Fig. 1⇑. Where indicated, data are normalized by setting the maximal response within each population to 100%, as described in Materials and Methods.
Compared with primary responses, peak Tyr369(Y)-specific CD8+ T cell recall responses to Tyr-vac in mice that had been primed at least 21 days previously with 10 μg/ml Tyr369(Y)-pulsed DC was 10-fold greater in magnitude (Fig. 2⇑b) and had 13- and 11-fold higher functional and structural avidities, respectively (Fig. 2⇑, c and d). Similar increases in the number of and avidity of recall CD8+ T cells were found when Tyr-vac was used for the initial priming and DC for the challenge (data not shown). The increased avidity of recall CD8+ T cells was not due to the expression of higher levels of TCR (not shown). Importantly, there was an average 2-fold difference between the functional avidities of T cells elicited by DC pulsed with 10 and 0.01 μg/ml Tyr369(Y) (Fig. 3⇓a). An even larger (4-fold) difference was observed between the structural avidities of these two populations (Fig. 3⇓b). Interestingly, both of the T cell populations had higher functional avidity than an established T cell line, YMD0.01, that had been maintained with 0.01 μg/ml Tyr369(Y). No significant difference in the MFI of each population was observed for the intracellular cytokine staining assays (data not shown), indicating that on a per cell basis each population secreted similar amounts of IFN-γ. The MFI for tetramer bound to CD8+ T cells from the mice primed with 0.01 μg/ml Tyr369(Y) trended higher, but was not determined to be significant (not shown). Because the differences in avidity observed in these recall responses were a consequence of the density of Ag on DC used for priming, we conclude that epitope density-based selection occurs during the generation of memory CD8+ T cells.
Density of peptide on DC used for primary immunization affects subsequent avidity of recall CD8+ T cells. Three AAD+ mice per group were immunized with DC that had been pulsed with the indicated concentration of Tyr369(Y) and challenged 21 days later with Tyr-vac. The functional (a) and structural (b) avidities of each population were determined as described in the legend for Fig. 1⇑. The functional avidity of a previously described (3 ) T cell line, YMD0.01, which was generated by culture on 0.01 μg/ml Tyr369(Y)-pulsed splenocytes, is included for comparative purposes. Data represent one of three similar experiments. c, Relative size of CD8+ T cell recall responses after immunization with DC pulsed with 10 or 0.01 μg/ml Tyr369(Y). The maximal size of recall responses from AAD mice that had been immunized with DC pulsed with either 10 or 0.01 μg/ml Tyr369(Y) and challenged with Tyr-vac was determined by ICS-IFN-γ and tetramer. Data are derived from four independent experiments. For each experiment, the number of CD8+ T cells that respond to DC pulsed with 10 μg/ml Tyr369(Y) was set to 100%, and the number of CD8+ T cells that respond to 0.01 μg/ml Tyr369(Y) was proportioned accordingly and averaged.
We also evaluated the peak number of recall CD8+ T cells elicited by Tyr-vac challenge of mice primed with DC that had been pulsed with different concentrations of Tyr369(Y). Surprisingly, when averaged over four independent experiments, the number of recall CD8+ T cells elicited in mice primed with 0.01 μg/ml Tyr369(Y)-pulsed DC was only 20% smaller than that of mice primed with 10 μg/ml Tyr369(Y)-pulsed DC (Fig. 3⇑c), compared with 80% difference in the size of the primary response (see Fig. 1⇑a). This suggested either that the memory cells generated by priming with 0.01 μg/ml Tyr369(Y)-pulsed DC expanded more rapidly during the recall response than those generated by priming with 10 μg/ml Tyr369(Y)-pulsed DC, or that similar numbers of memory cells were generated from these two immunizations, despite the difference in size of the primary responses. To distinguish between these possibilities, we used a single cohort of animals primed with either 10 μg/ml or 0.01 μg/ml Tyr369(Y)-pulsed DC to compare the number of Tyr369(Y)-specific CD8+ T cells at the peak of the recall response, at the peak of the primary response, and in quiescent memory. Despite a 3-fold difference in the size of the primary responses of mice primed with 10 μg/ml Tyr369(Y) compared with mice primed with 0.01 μg/ml Tyr369(Y) (Fig. 4⇓a), we found that the number of Tyr369(Y)-specific CD8+ T cells in the quiescent memory was remarkably similar (Fig. 4⇓b), and consequentially the size of the recall response to Tyr-vac challenge was not significantly different between the cohorts (Fig. 4⇓c). Thus, these data demonstrate that while the CD8+ T cells primed with 10 μg/ml Tyr369(Y) undergo a 6.4-fold contraction to quiescent memory, the cells activated with 0.01 μg/ml Tyr369(Y) only contract by 2.7-fold. This difference leads to a more equivalent recall response upon Ag challenge.
Size of primary, memory, and recall CD8+ T cell populations specific for Tyr369(Y) in AAD+ mice. Cohorts of three AAD+ mice were immunized with DC pulsed with either 10 or 0.01 μg/ml Tyr369(Y). Tyr369(Y)-specific CD8+ T cells were enumerated on (a) day 7 (primary) or (b) day 21 (memory) by ICS-IFN-γ using 100 μg/ml Tyr369(Y)-pulsed C1R-AAD cells, or by staining with 7.2 μg/ml Tyr369(Y) tetramer. c, Recall responses to Tyr369(Y) were initiated on day 21 by challenge with 107 PFU Tyr-vac and were assayed for Tyr369(Y)-specific CD8+ T cells on day 26.
The epitope density used to challenge memory CD8+ T cells affects the avidity of the recall response
The above results showed that the density of epitope presented during priming did not affect the avidity of primary CD8+ T cell effectors and yet had a discernable influence on the avidity of subsequent recall responses. However, the avidity differences in these recall responses were still substantially less than observed in long-term T cell lines (3). A possible explanation for this discrepancy lay in the use of Tyr-vac to induce recall responses. High level epitope expression in Tyr-vac-infected APC might lead to the deletion of high avidity T cells, or a broad range of epitope densities in different infected APC might lead to a lack of selectivity in stimulating memory cells with different avidities. To circumvent this potential problem, splenocytes from mice that had been immunized with DC pulsed with either 10 or 0.1 μg/ml Tyr369(Y) 21 days previously were restimulated in vitro with either 10 or 0.1 μg/ml Tyr369(Y), and the functional avidities of the in vitro recall population were assessed 1 wk later. DC and splenocytes were pulsed with peptide for the same amount of time and washed before being used for priming or for restimulation. The avidities of T cells primed with 0.1 μg/ml Tyr369(Y)-pulsed DC and restimulated with 0.1 μg/ml peptide-pulsed splenocytes was much higher than those primed and restimulated with 10 μg/ml peptide-pulsed cells (Fig. 5⇓). However, the avidities of T cells restimulated in vitro with 0.1 μg/ml peptide-pulsed splenocytes were comparable, regardless of the epitope density on the DC used for priming. Avidities of T cells cultured in 10 μg/ml peptide were also similar, regardless of in vivo priming density. Thus, the density of epitope presented during the in vitro restimulation substantially outweighed any influence of priming dose on the avidity of the T cells. These data suggest either that use of Tyr-vac for in vivo challenge does blunt the selection of T cells with different avidities, or that T cells are selected differently in vitro and in vivo.
The influence of epitope density in vivo and in vitro on the avidity of short-term T cell cultures. AAD+ mice were immunized with DC that had been pulsed with 10 or 0.1 μg/ml Tyr369(Y) and rested for at least 21 days. Short-term cultures were established in vitro after pulsing splenocytes from immunized mice with either 10 or 0.1 μg/ml Tyr369(Y). After 6 days, the functional avidity of each population (a) was measured as described in the legend for Fig. 1⇑. b, Accumulated functional avidity data from three independent assays. Where indicated, data are normalized by setting the maximal response within each population to 100%, as described in Materials and Methods.
To determine directly whether avidity-based selection of recall T cells could occur in vivo, mice were primed with Tyr-vac, rested for 3 wk, and then challenged with DC that had been pulsed with graded concentrations of Tyr369(Y). After 5 days the size and functional avidity of each CD8+ T cell population was assessed directly ex vivo. We observed that the magnitude of the recall responses in Tyr-vac-primed mice was dependent upon the density of epitopes presented by DC, indicating that differences in the range used were relevant for recall responses (Fig. 6⇓a). DC pulsed with 0.01 μg/ml Tyr369(Y) elicited T cells whose functional avidity was 2-fold higher than that of T cells elicited by DC pulsed with 10 μg/ml of this peptide (Fig. 6⇓b). This difference is comparable to that observed in Fig. 3⇑, and shows that DC pulsed with these two concentrations of peptide elicited comparable differences in recall T cell avidity, whether used for priming or for challenge. This difference in avidity was also substantially less than achieved using 10 and 0.1 μg/ml peptide-pulsed cells in vitro. Thus, while avidity-based selection of recall T cells can occur in vivo, it is constrained relative to selection in vitro.
Density of Ag on DC used for in vivo restimulation affects the avidity of recall responses to Tyr369(Y). AAD+ mice were immunized with 107 PFU Tyr-vac and rested for 21 days. Cohorts of three immunized mice were challenged with DC pulsed with the indicated concentration of Tyr369(Y) and assayed 5 days postchallenge by ICS-IFN-γ to determine the size (a) and functional avidity (b), while structural avidity (c) was determined by staining with Tyr369(Y) tetramer. The functional avidity of a previously described (3 ) T cell line, YMD0.01, which was generated by culture on 0.01 μg/ml Tyr369(Y)-pulsed splenocytes, is included for comparative purposes. Where indicated, data are normalized by setting the maximal response within each population to 100%, as described in Materials and Methods. Data represent one of three similar experiments.
The relative parity in the number of recall T cells activated in Tyr-vac-primed mice by challenge with DC pulsed with 10 or 0.01 μg/ml Tyr369(Y) suggested that Tyr-vac priming had generated high avidity memory T cells. Confirming this, mice primed with Tyr-vac showed substantial recall responses to DC pulsed with 0.001 μg/ml Tyr369(Y), which were not immunogenic in naive mice (compare Figs. 6⇑a and 1⇑a). Importantly, the functional avidity of CD8+ T cells elicited using DC pulsed with 0.001 μg/ml Tyr369(Y) was 4-fold greater than that of T cells elicited using DC pulsed with 10 μg/ml of this peptide (Fig. 6⇑b), and the structural avidity was 11-fold greater (Fig. 6⇑c). Thus, memory T cells respond to a lower density of peptide Ag, and consequently across a broader range of Ag densities than primary T cells. This in turn enables T cells with more significant differences in avidity to be selected in vivo.
Avidity-based maturation and selection enhances recognition of tumor
A major reason for pursuing these studies was to be able to manipulate the avidity of T cells to better recognize low density Ags expressed on tumor cells. Thus, despite the relatively small differences in structural and functional avidity induced in vivo with DC pulsed with different concentrations of peptide, it was of interest to determine more directly whether these differences influenced tumor recognition. Therefore, we analyzed the proportion of each primary and recall Tyr369(Y)-specific CD8+ T cell population that recognized tyrosinase-expressing melanomas. Regardless of the Ag density used to immunize, primary CD8+ T cells failed to recognize DM93 melanoma cells, while a substantial fraction of recall CD8+ T cells (with a 6- to 12-fold higher functional avidity) did so (Fig. 7⇓a). In comparison to mice primed or challenged with 10 μg/ml peptide-pulsed DC, a slightly higher proportion of Tyr369(Y)-specific CD8+ T cells from mice primed or challenged with 0.01 μg/ml Tyr369(Y)-pulsed DC recognized DM93 (Figs. 7⇓, b and c). Thus, 2-fold differences in functional avidity in this system have a discernable but small effect on tumor recognition. Importantly, a substantially greater proportion of Tyr369(Y)-specific CD8+ T cells elicited by challenge using DC pulsed with 0.001 μg/ml peptide, which have the highest avidity of any population used in this study, were able to recognize DM93 (Fig. 7⇓c). These data indicated that immunization strategies that result in greater than 2-fold changes in functional avidity can have a significant impact on tumor recognition.
Recognition of melanoma by primary and recall CD8+ T cells. a, Naive (primary) or AAD+ mice that have been previously immunized with Tyr-vac (recall) were immunized with DC that had been pulsed with 10 μg/ml Tyr369(Y) and analyzed by ICS-IFN-γ 5 days (recall) or 7 days (primary) post-DC immunization. b, Recall CD8+ T cells from mice primed with DC pulsed with the indicated concentration of Tyr369(Y) and challenged with Tyr-vac. c, Recall CD8+ T cells from AAD+ mice primed with Tyr-vac and challenged with DC pulsed with the indicated concentrations of peptide. In each case, CD8+ T cells were incubated at a 1:1 ratio with DM93 (HLA-A2+, tyrosinase+) tumors and assayed for ICS-IFN-γ. Background levels of activation, determined by incubation with DM331 (HLA-A2+, tyrosinase−), have been subtracted.
Discussion
The experiments in this study analyze how the density of an individual peptide Ag presented by DC influences the size and avidity of the CD8+ T cell response at three distinct immunological stages. We have determined that, while epitope density directly affects the number of specific CD8+ T cells at the peak of the primary in vivo response, it has no discernable influence on their avidity. In contrast, the density of peptide presented during priming does influence the avidity of recall CD8+ T cells, but has a relatively small impact on the size of this response. Similarly, the density of peptide used to challenge preimmunized mice also influences the avidity of recall CD8+ T cells, and the increase in avidity found with memory T cells allows enrichment of a significantly higher avidity population of recall T cells in vivo.
Previous studies have amply demonstrated that variations in epitope density can be used to select in vitro T cell lines with different avidities. Our results have extended this by demonstrating that variations in epitope density can lead to the selection of recall CD8+ T cells with different avidities in vivo. Similar conclusions were reached in a previous study that examined only the structural avidity of recall CD4+ T cells after immunization with peptide in adjuvant (15). However, that study did not use processed Ag loaded onto a fixed number of DC, and thus could not distinguish epitope density, persistence, or distribution to different numbers of APC or lymphoid tissues as contributing factors. In the present study, we have also shown that T cells with higher functional avidity expressed TCR with higher structural avidity, suggesting that lower epitope densities displayed in vivo selected Ag-specific T cells with higher TCR affinities. In contrast, the differences in avidity obtained by epitope-density based selection in vivo were substantially smaller than those obtained in vitro (3). It is important to note that the highest avidity developed by selection of long-term cultures of T cells with low density peptide Ag was lower than the avidity of recall T cells elicited in vivo with the equivalent density of peptide. Thus, the greater range of avidities observed in vitro is mainly due to the enhanced outgrowth of lower avidity T cells. Previous work has established that higher avidity in vitro T cell lines are more vulnerable to cell death in the presence of high levels of Ag (23). Therefore, the broader avidity range observed in vitro may reflect the effects of prolonged activation in cell culture and increased sensitivity to Ag-induced cell death.
Although epitope density-based selection clearly occurs in the activation of recall CD8+ T cells, we found that the functional and structural avidities of primary CD8+ T cells were identical, regardless of the epitope density used for immunization. This is particularly striking because these same immunizations led to the epitope density-based selection of memory T cells with avidities that were discernibly different upon reactivation. Both the relatively short duration of epitope presentation by peptide-pulsed DC (24), and the recent report demonstrating Ag-independent contraction of primary CD8+ T cell responses (25), argues against epitope density-based selection of memory cells during the contraction phase of the primary response. Thus, the memory cells examined herein are apparently generated and selected based upon epitope density during early stages of the primary response. An alternative explanation for the differences in epitope density-based selection of primary and memory T cells is that the latter is derived from a distinct subset of activated naive T cells that led to primary effector cells, and were selected in a distinct manner. However, several groups have provided evidence for a direct lineage between primary effector cells and memory cells (26, 27, 28, 29). Even if memory cells are generated in parallel to primary effectors, it would not explain why epitope density-based selection of primary effector cells did not occur.
To reconcile these seemingly incompatible observations, we suggest that epitope density-based selection of naive T cells occurs during the primary response, but that differences in TCR affinity that should lead to differences in both structural and functional avidity of primary effector cells are minimized by compensatory mechanisms involving other adhesion or costimulatory molecules. These mechanisms might augment the avidity of effector T cells with low affinity TCR or blunt the avidity of those with very high affinity TCR. We further propose that these compensatory mechanisms become nonoperative in memory/recall T cells, leading to the expression of T cell avidities that are more directly correlated with TCR affinity. In support of this model, it has been shown that the structural or functional avidity of T cells expressing a single TCR can be altered (8, 9), and also that the structural avidity (measured by MHC-multimer staining) is dependent upon the organization of TCR in the membrane (9, 30). Thus, the structural and functional avidity of primary effector cells could be manipulated independent of the intrinsic affinity of the TCR. In addition, recall T cells have considerably higher functional and structural avidities than primary T cells (4, 5, 6, 7) (and our results), and also show reduced dependence on costimulatory pathways (13). It is possible that these changes supersede the effect of the compensatory mechanisms that operate to mask avidity differences in the primary response. Regardless of the exact details, the net effect would be to enable the expression of functional avidity differences in recall T cells that were not observable in their antecedents. It remains unclear whether the proposed compensatory mechanisms would be operative in naive cells, which could lead to activation of a broader selection of the T cell repertoire, or whether they only become operative in already activated cells.
The data presented in this study further reinforce our previous observations that higher epitope density generally results in an increase in the number of activated CD8+ T cells. We had previously hypothesized that more CD8+ T cells are activated by DC pulsed with higher concentrations of peptide because T cells with lower avidity were being recruited into the response. Although the absence of epitope density-based avidity differences in primary effector cells superficially argues against this interpretation, the existence of avidity differences in memory populations is consistent with it. Again, the model presented above provides a basis for reconciling these disparities through avidity-based selection of naive T cells. However, it is also possible that there is a minimum threshold density of Ag displayed by DC required to activate any T cells at all. A larger number of DC presenting higher densities of peptide Ag will display this minimum density initially, as well as at later points in time. Obviously, above the minimum density, the number of Ag complexes displayed may also dictate the number of T cells that can be productively engaged and activated.
Analyses of different epitope-specific CD8+ T cell populations during T cell responses to viruses or bacteria have demonstrated that the number of memory CD8+ T cells for each epitope was proportional to the number of epitope-specific primary effector cells (4, 31, 32). This suggests that extent of contraction of each epitope-specific T cell population is similar, regardless of its initial size. More recently, both the kinetics and the extent of CD8+ T cell contraction were shown to be the same for populations that had been primed with either high or low doses of Listeria monocytogenes (25), despite a 10-fold difference in the number of primary effector cells. In contrast, we found that the number of memory and recall T cells in mice primed with high epitope densities were remarkably similar to those in mice primed with low epitope densities. One explanation for this disparity is that the number of epitope-specific primary T cells ranged between 4 × 106 and 4 × 105 per spleen in the study by Badovinac (25), while in the present work, they were only 1.5 × 105 to 3 × 104. As the purpose of the contraction phase is to reduce the number of activated T cells to a density that minimizes immunopathology and enables T cell expansion during subsequent immune responses, smaller effector T cell populations need not contract to the same extent as larger ones. Our results demonstrate that immunizations using low epitope density to select for high avidity memory CD8+ T cells need not also lead to the generation of substantially smaller numbers of memory T cells. Furthermore, as we have previously demonstrated that the adoptive transfer of large numbers of low avidity CD8+ T cells does not result in substantial delay of tumor outgrowth (3), the increases in CD8+ T cell avidity obtained by immunizing with DC that present low peptide densities is likely to compensate for any reduction in number of CD8+ T cells generated by this vaccination strategy.
In summary, the results presented in this study demonstrate for the first time that CD8+ T cells undergo epitope density-based selection in vivo. The effect of epitope density-based selection is apparently masked in the primary effector populations, but becomes evident in the resulting memory populations. Importantly, our results also indicate that epitope density-based selection can also occur during the generation of recall responses, indicating an immunization strategy for the generation of high avidity T cell populations. Furthermore, we have found that priming with DC that present low epitope densities does not adversely affect the number of T cells elicited into a recall response, yet does select for higher avidity T cells. Together, these results suggest immunization protocols that avoid the use of excessive epitope densities are more likely to harness large numbers of high avidity T cells for effective immunotherapy.
Acknowledgments
We thank Janet Gorman and Sarah Tate for excellent technical assistance, and Dr. Rejean Lapointe for providing CD40 ligand-transduced NIH-3T3 cells.
Footnotes
-
↵1 This work was supported by U.S. Public Health Service Grants AI21393 and CA78400 (to V.H.E.). T.N.J.B. was the recipient of the Rudolph M. Montgelas Fellowship from the Cancer Research Institute. D.W.M. was supported by U.S. Public Health Service Training Grant AI07496 and is currently a fellow of the American Cancer Society.
-
↵2 Address correspondence and reprint requests to Dr. Victor H. Engelhard, Department of Microbiology and Carter Immunology Center, University of Virginia Health System, Room 4072C, MR4 Building, Charlottesville, VA 22908. E-mail address: vhe{at}virginia.edu
-
↵3 Abbreviations used in this paper: DC, dendritic cell; ICS, intracellular cytokine staining; AAD, α3 domain of H2-Dd; Tyr-vac, recombinant vaccinia virus expressing full-length human tyrosinase; Tyr369(Y), peptide encoded by amino acid 369-377 of human tyrosinase.
- Received October 3, 2002.
- Accepted December 10, 2002.
- Copyright © 2003 by The American Association of Immunologists
References
In this issue
