Cutting Edge: Dendritic Cells Prime a High Avidity CTL Response Independent of the Level of Presented Antigen¹

Mature dendritic cells (DC)⁴ are distinguished by their high expression of costimulatory molecules and secretion of cytokines. In this state, DC are potent activators of naive T cells. Productive conjugate formation will result in naive T cell activation, proliferation, and the acquisition of effector functions. These CTL then enter the periphery where they exert lytic activity and release cytokine following encounter with infected cells.

Previous studies have demonstrated that, within the population of responding CTL, there are individual clones that encompass a wide spectrum of functional avidities (1). The term functional avidity refers to the number of peptide-MHC (pMHC) complexes required to activate a T cell. Low avidity CTL require a high number of pMHC complexes, whereas high avidity CTL can become activated when presented with a low number of pMHC complexes. Although a diverse clonal response is generated to a given pathogen, high avidity CTL appear to be the more important effector population. Adoptive transfer studies have demonstrated that high avidity CTL are more effective then their low avidity counterparts in viral clearance and tumor eradication (for examples, Refs. 1, 2, 3, 4, 5).

Although readily generated in vitro by stimulation with low vs high concentrations of peptide, the mechanism(s) by which high vs low avidity CTL are generated in vivo remains unclear. Several factors have been implicated in the regulation of functional avidity, i.e., the cytokines IL-12 and IL-15 (6, 7), CD8αβ expression (8, 9, 10), TCR affinity (3), and level of costimulatory molecules expressed by APC (11, 12). Therefore, it appears that multiple signals may be important in the development/regulation of functional avidity. As mature DC are characterized by their increased expression of costimulatory molecules and production of IL-12, it seemed possible that mature DC may skew the responding T cell response to be of higher avidity (13, 14).

The work presented here directly tested the ability of mature DC to generate both high and low avidity CTL. We found that mature DC presenting a high level of pMHC were poor inducers of low avidity CTL in contrast to other APC examined. However, a single encounter with Ag-bearing DC did not permanently fix avidity in the CTL, as they were capable of differentiating into a low avidity cell following encounter with a nonprofessional APC bearing a high level of peptide. These data indicate that regardless of pMHC level, mature DC prime a potent high avidity CTL response. Nonetheless these CTL retain plasticity that allows them to “fine tune” their peptide sensitivity as a result of subsequent Ag encounter.

C57BL/6 mice were obtained from the Frederick Cancer Research and Development Center (Frederick, MD). TgN(TCRP14LCMV) Rag2^tm1 mice (15) were purchased from Taconic Farms. All experiments in this study comply with the institutional guidelines approved by the Wake Forest University (Winston-Salem, NC) Animal Care and Use Committee. The lymphocytic choriomeningitis virus (LCMV) gp_33–41 and Ova_257–264 peptides were synthesized at the Wake Forest University School of Medicine Comprehensive Cancer Center Protein Analysis Core Laboratory.

The protocol used to generate BMDC was as previously reported by Pejawar et al. (16). On day 6 postculture, 200 ng of LPS was added to induce DC maturation.

TgN(TCRP14LCMV) Rag2^tm1 splenocytes (5 × 10⁵) were cocultured with 5 × 10⁶ C57BL/6 splenocytes, 2 × 10⁵ EL4 cells, or 5 × 10⁴ to 2 × 10⁵ DC previously pulsed with 10⁻⁴ M or 10⁻⁹ M gp_33–41 peptide in the presence of 10% T-stim (BD Biosciences). In some panels of lines, stimulators were irradiated before culture. Cultures were maintained by weekly stimulation.

Functional avidity of CTL lines was measured on day 7 following the third peptide stimulation by IFN-γ ELISA, as previously described (10).

LPS-matured BMDC and naive C57BL/6 splenocytes were left untreated or pulsed with titrated concentrations of OVA_257–264 (10⁻⁴ to 10⁻⁹ M) for 3 h followed by washing. Cells were stained with Fc block, the Ova_257–264/K^b-specific Ab 25.D1.16 (a gift of Dr. R. N. Germain, National Institutes of Health, Bethesda, MD), and anti-mouse IgG Alexa-Fluor 488. Samples were acquired on a FACSCalibur flow cytometer and analyzed using the CellQuest Pro software (BD Biosciences).

Splenocytes or LPS-matured BMDC (5 × 10⁵) that had been pulsed with graded concentrations of peptide were cocultured with CFSE-labeled TgN(TCRP14LCMV) Rag2^tm1 splenocytes (2 × 10⁵/well) for 5 h at 37°C. TCR surface expression was determined by staining with PE-labeled anti-TCR Ab (clone H57) followed by flow cytometric analysis.

Previously published data from our laboratory have demonstrated that both low and high avidity effector cells can be generated from TgN(TCRP14LCMV) Rag2^tm1 splenocytes specific for the LCMV gp_33–41 peptide (9, 10, 17). These CTL were generated by stimulation with peptide-pulsed naive splenocytes, a population with a very limited number of DC. Given that mature DC are the predominant activators of naive T cells in vivo, we determined whether stimulation with this APC would also give rise to T cells with distinct functional avidities. TgN(TCRP14LCMV) Rag2^tm1 CD8⁺ T cell lines were generated using our standard approach of stimulation with splenocytes or alternatively using LPS-matured BMDC, each pulsed with a high (10⁻⁴ M) vs low (10⁻⁹ M) concentration of gp_33–41 peptide. On day 7 after tertiary stimulation, functional avidity was examined by assessing the production of IFN-γ. As previously reported, CTL stimulated with 10⁻⁴ M peptide-pulsed splenocytes required ∼100-fold more peptide for half-maximal production of IFN-γ compared with the CTL stimulated with the 10⁻⁹ M peptide-pulsed splenocytes (Fig. 1). This indicated that, as expected, CTL of distinct functional avidities were generated by stimulation with this predominantly nonprofessional APC population. However, regardless of the priming Ag concentration used, CTL stimulated with DC exhibited a similar requirement for peptide. In fact, both CTL lines exhibited a characteristically high avidity phenotype, similar to the CTL generated by stimulation with 10⁻⁹ M peptide-pulsed splenocytes. This was the case at all times tested (i.e., following primary and secondary stimulation). Thus, for DC, the stimulatory peptide dose does not appear to control the quality of the T cell response generated, as both low and high concentrations of peptide led to the generation of high avidity CTL.

One possibility for explaining the failure to generate low avidity CTL was that the number of accessible MHC binding sites on the DC was very low, thus limiting the level of peptide loading in that cell type. We used two approaches to test this possibility: 1) examining the level of exogenously loaded Ova_257–263 peptide given that an Ova_257–264/K^b-specific Ab is available (18) (no such reagent exists to measure the presentation of gp_33–41); and 2) determining the degree of TCR internalization on naive TgN(TCRP14LCMV) Rag2^tm1 cells following gp_33–41 peptide encounter. For measuring Ova_257–264 loading, LPS-matured BMDC and C57BL/6 splenocytes were pulsed with the indicated concentration of peptide and stained with the Ova_257–264/K^b-specific Ab. As shown in Fig. 2,A, both APC exhibited similar levels of Ova_257–264/K^b complexes at each Ag concentration. Using TCR internalization as a probe for the level of presented peptide, we found that coculture with LPS-matured BMDC and splenocytes resulted in a comparable dose-dependent decrease in TCR cell surface expression (Fig. 2 B), consistent with TCR engagement of a similar number of pMHC molecules on the two APC. Together, these data suggest mature BMDC have accessible binding sites and that there is a similar dose-dependent loading of peptide onto the MHC molecules of splenocytes vs mature BMDC. Thus, it is highly unlikely that limiting pMHC would explain the failure of BMDC pulsed with the high concentration of peptide to generate lower avidity cells.

Another potential explanation for the inability of BMDC to modulate T cell avidity is that different numbers of splenocytes and BMDC were used to stimulate the TgN(TCRP14LCMV) Rag2^tm1 cells (5 × 10⁶ vs 5 × 10⁴). Initially we attempted to generate CTL lines by increasing the number of DC to 5 × 10⁶ or decreasing the number of splenocytes to 5 × 10⁴. However this resulted in either a high level of T cell death (with the BMDC) or poor recruitment of cells into the activated population (with the splenocytes). Therefore, we turned to EL4 cells, a thymoma cell line, as a non-DC stimulator. TgN(TCRP14LCMV) Rag2^tm1 cells were stimulated with 2 × 10⁵ EL4 cells or mature BMDC. CTL generated with BMDC again exhibited no difference in their peptide sensitivity. In contrast, EL4 cells presenting the high vs low concentration of Ag resulted in the generation of low vs high avidity CTL, respectively. The amount of peptide required for half-maximal IFN-γ production from multiple CTL lines generated in this fashion is shown in Fig. 3 B. The generation of both high and low avidity CTL using the matched number of EL4 cells and BMDC as APC suggested that the lower BMDC number was not responsible for the failure of DC to generate low avidity cells.

It is known that CTL clones with distinct functional avidities are generated in vivo during an immune response (1) and that naive T cells are activated by DC. Although these data may appear at odds with the findings presented here, in the context of a viral infection CTL would encounter Ag presented by various cell types, including DC in the lymphoid tissue and nonprofessional APC in the periphery, i.e., epithelial cells. We hypothesized that subsequent interactions with nonprofessional APC presenting a high level of peptide could induce the T cells to modulate their functional avidity. To address this possibility, we sought to determine whether naive T cells stimulated with mature DC exhibited a “fixed” peptide sensitivity or whether, following encounter with a nonprofessional APC, the responding CTL could “fine tune” their peptide sensitivity according to the stimulatory concentration of peptide. Naive TgN(TCRP14LCMV) Rag2^tm1 cells were stimulated with mature DC previously pulsed with the high or low concentration of peptide. After primary stimulation, the 10⁻⁴ M and 10⁻⁹ M peptide-stimulated CTL were divided and restimulated for two additional rounds with EL4 cells pulsed with either a high or low concentration of peptide. The functional avidity of the various cultures was then examined (Fig. 4). In support of the proposed hypothesis, differences in functional avidity could be detected when CTL initially stimulated with BMDC were restimulated with EL4 cells presenting a high vs low concentration of peptide. CTL stimulated with 10⁻⁴ M peptide-pulsed EL4 cells required a higher concentration of peptide for their half-maximal production of IFN-γ compared with the CTL stimulated with10⁻⁹ M pulsed EL4 cells. Therefore CTL initially activated by DC appear to maintain a window of plasticity that allows “tuning” of their peptide sensitivity following encounter with a nonprofessional APC.

Although the property of functional avidity is well documented, much remains to be learned with regard to the mechanism(s) that regulate this T cell property. In addition, the extrinsic factors that direct how high and low avidity CTL arise in vivo remain unclear. DC are considered to be the most potent APC for the activation of naive T cells due to their increased expression of costimulatory molecules and cytokines, i.e., IL-12 (reviewed in Refs. 13 and 14). Interestingly, high costimulatory molecule expression and IL-12 have been shown to promote higher avidity CTL responses (7, 11, 12). Therefore, we sought to determine whether mature DC presenting different concentrations of peptide were capable of generating high vs low avidity CTL or whether the signals provided by mature DC would inhibit the emergence of T cells with distinct functional avidities.

Our data demonstrated that the CD8⁺ T cells generated by stimulation with mature BMDC did not exhibit dose-dependent differences in peptide sensitivities; rather, the responding CTL were of high avidity regardless of pMHC levels. This is in stark contrast to the differences in functional avidity observed following stimulation with peptide-pulsed nonprofessional APC. The inability of mature DC to induce low avidity cells was also observed using splenic DC (data not shown).

In addition to differences in peptide sensitivity, we have previously found that CD8 is expressed differentially between high and low avidity TgN(TCRP14LCMV) Rag2^tm1 CTL, with low avidity lines exhibiting either lower absolute CD8αβ levels or lower β:α ratios, the latter suggesting higher expression of CD8αα homodimers (8, 9, 10, 19). Consistent with the failure of DC to induce CTL of low avidity, cells stimulated with mature DC pulsed with a high vs low concentration of peptide both exhibited high levels of CD8 (data not shown). Of note, the increased CD8 expression on high avidity CTL generated by stimulation with 10⁻⁴ M peptide-pulsed DC did not appear to solely account for the high avidity phenotype. Removing the contribution of CD8 (by the addition of blocking antibody) in the lines generated by stimulation with high peptide-pulsed DC vs splenocytes did not result in similar peptide sensitivities (data not shown). Thus, while increased CD8 expression likely contributes to high avidity in the DC generated line, it does not seem sufficient to explain it. In summary, our findings show that while the level of pMHC on nonprofessional APC drives the generation of high vs low avidity CTL, DC selectively promote generation of CTL with high avidity regardless of stimulatory pMHC levels.

Given the reported role for costimulatory molecules and IL-12 in promoting high avidity CTL (7), we tested whether the presence of these DC-derived signals might explain the selective generation of high avidity cells in this model. However, BMDC from CD80/86-, ICAM-1-, IL-15-, or p35-deficient animals all failed to generate CTL with differences in avidity (data not shown). Therefore, none of these molecules independently was responsible for the selective generation of high avidity cells by DC.

It has recently been shown that the immunological synapse formed by naive T cells following encounter with DC differs from the iconic bulls-eye pattern formed following interaction with B cells. T cells conjugated to DC preferentially form multifocal immunological synapses (for review see Ref. 20). Such changes have the potential to significantly impact the nature of the signal delivered, either quantitatively or qualitatively. Additionally, the kinetics of TCR signaling within T cells appears to be different upon conjugation with a professional vs nonprofessional APC. Delon et al. showed that naive T cells in conjugate with DC fluxed Ca⁺² with increased kinetics compared with resting B cells (21). Additionally, with regard to the length of time T cells are in conjugate with the APC, previous studies have shown that contacts between mature DC are more transient and dynamic compared to those with resting B cells (22). It is tempting to speculate that these transient contacts fail to result in the signal strength that may be required for the induction of low avidity or that the additional signaling through costimulatory molecules or cytokine receptors produces a unique signaling pattern protecting the cells from the consequences of encountering high levels of pMHC such as death or differentiation into low avidity CTL.

The results from our study shed light on a previously published report from Bullock et al. (23). In that study, DC pulsed with a high vs low concentration of peptide were used to immunize mice. No differences in avidity were detected in populations generated by immunization with DC pulsed with high vs low peptide levels. However, whether the failure to generate cells of disparate avidity was the result of in vitro vs in vivo differences or in the nature of the APC was unknown. Our data support the latter. The ability of DC to promote the generation of high avidity cells even in the presence of high levels of Ag could be of significant benefit, as the level of Ag presented by DC may not reflect that which would be encountered in the periphery. Such divergence in Ag levels could result from cell type-specific differences in permissivity to infection. The preferential generation of high avidity cells by DC would promote an optimally effective population as the default. In this scenario, if the level of presented peptide were significantly higher in the periphery, effectors would retain the capacity to down-modulate their avidity to protect against apoptosis (24). Such a system would promote optimal recognition and thus clearance by CTL.

We thank Dr. Ron Germain for provision of the 25.D1.16 hybridoma. We appreciate the helpful comments of Dr. Sharad Sharma and Dr. Marlena Wescott.

The authors have no financial conflict of interest.