Direct Presentation Regulates the Magnitude of the CD8+ T Cell Response To Cell-Associated Antigen through Prolonged T Cell Proliferation

Abstract

The magnitude and complexity of antigen-specific CD8+ T cell responses is determined by both intrinsic properties of the immune system and extrinsic factors, such as vaccination. We evaluated mechanisms that regulate the CD8+ T cell response to two distinct determinants derived from the same protein antigen, SV40 T antigen (T Ag), following immunization of C57BL/6 mice with T Ag transformed cells. The results show that direct presentation of T cell determinants by T Ag transformed cells regulates the magnitude of the CD8+ T cell response in vivo, but not the immunodominance hierarchy. The immunodominance hierarchy was reversed in a dose-dependent manner by addition of excess naïve T cells targeting the subdominant determinant. However, T cell competition played only a minor role in limiting T cell accumulation under physiological conditions. We found that the magnitude of the T cell response was regulated by the ability of T Ag transformed cells to directly present the T Ag determinants. The hierarchy of the CD8+ T cell response was maintained when antigen presentation in vivo was restricted to cross-presentation, but the presence of T Ag transformed cells capable of direct presentation dramatically enhanced T cell accumulation at the peak of the response. This enhancement was due to a prolonged period of T cell proliferation, resulting in a delay in T cell contraction. Our findings reveal that direct presentation by non-professional antigen presenting cells can dramatically enhance accumulation of CD8+ T cells during the primary response, revealing a potential strategy to enhance vaccination approaches.

Introduction

CD8+ T cells are an important component of the immune response to both intracellular pathogens and tumors due to their ability to recognize and eliminate individual cells expressing the target antigen. The magnitude of the CD8+ T cell response has been correlated with control of acute infections (1, 2) and tumor progression (3-5) as well as the level of memory T cells that remain in the host following immunization or infection (6). Therefore, it is imperative to identify mechanisms that regulate the magnitude of the CD8+ T cell response in order to better understand the immune response to infection and to optimize vaccine strategies.

Previous studies have highlighted several factors that influence the magnitude of a given epitope-specific CD8+ T cell response, including recent results demonstrating a correlation between the frequency of naïve T cell precursors and the number of cells that accumulate at the peak of the response (7-10). In addition, CD8+ T cell competition for available APCs can limit the frequency of CD8+ T cells generated during an immune response (11-14), although this mechanism may not always contribute to the observed response (8, 15). The amount of antigen available to prime CD8+ T cells can also influence the magnitude of a CD8+ T cell response, with high antigen load shown to expand the size of the primary T cell response (16, 17), perhaps by prolonging the duration of antigen presentation in vivo (18).

Processing and presentation of antigen play important roles in initiating an effective immune response and can also contribute to immunodominance, a phenomenon in which the magnitude of T cells simultaneously targeting unique peptide/MHC class I complexes forms a predictable hierarchy (19). Both direct presentation and cross-presentation of antigen by professional APCs (pAPCs) can contribute to the induction of an effective CD8+ T cell response (20-23). In the case of tumor antigens, several reports indicate that direct presentation by tumor cells localized to the lymphoid organs can induce T cell activation directly (24-26). An additional role for direct antigen presentation by non-pAPCs may be to promote CD8+ T cell accumulation (27). Such a phenomenon has been observed following DNA vaccination where the combination of cross-presenting pAPCs and directly-presented antigen on transfected cells is capable of enhancing the magnitude of the CD8+ T cell response (28-30). A similar observation was made by Thomas et al. (31), in which antigen presentation by nonhemopoietic cells significantly contributed to the CD8+ T cell response towards LCMV, suggesting that nonhemopoietic cells can further drive proliferation during virus infection. More recently, Pavelic and colleagues found that direct presentation of the dominant LCMV epitope gp33 on tumor cells increased the frequency of responding CD8+ T cells relative to cross-presentation alone (32). The basis for this enhanced response remains unknown, however.

We investigated how the magnitude of the CD8+ T cell response is shaped toward the two most dominant determinants from SV40 T Ag following cellular immunization. The oncogenic protein SV40 T Ag has four well-defined H-2^b-restricted determinants designated sites I, II/III, IV and V. The CD8+ T cell response to sites I, II/III and V are H-2D^b-restricted, while the response to site IV is H-2K^b-restricted. Previously, Mylin et al. (33) showed that a hierarchal response to site IV>I>II/III is established 9 days post immunization of C57LB/6 mice with wild type T Ag-transformed cells and that this hierarchy is maintained at the T cell memory stage and after secondary booster immunization. Typically, no response against the immunorecessive site V determinant is detected when the dominant determinants are expressed (27). Here we demonstrate that the dominance of the CD8+ T cell response to site IV over site I is established early, and that the immunodominance hierarchy can be modulated by changing the naïve precursor frequency. We further show that the overall magnitude of response to each determinant is shaped by the absence or presence of direct presentation by the cells used for immunization and the associated effects on duration of T cell proliferation.

Materials and Methods

Mice

C57BL/6 (H-2^b) mice were purchased from the The Jackson Laboratory (Bar Harbor, ME) and maintained at the Milton S. Hershey Medical Center (Hershey, PA) animal facility under specific pathogen free conditions. B6.SJL mice were obtained from Taconic Farms (Germantown, NY). TCR-I mice on the C57BL/6 background have been previously described (34) and are available from The Jackson Laboratory as line B6.Cg-Tg(TcraY1,TcrbY1)416Tev/J. Transgene positive TCR-I progeny were identified by staining peripheral blood lymphocytes with FITC-labeled anti-Vβ7 antibody (BD Pharmingen). In some experiments, TCR-I females were bred with male C57BL/6-Tg(UBC-GFP)30Scha/J mice obtained from The Jackson Laboratory in order to obtain TCR-I T cells expressing green fluorescent protein. B6.FVB-Tg(Itgax-DTR/EGFP)57Lan/J (CD11cDTR mice) were obtained from The Jackson Laboratory at generation N5 on the C57BL/6 background and were backcrossed further to the N10 generation prior to performing experiments. All animal experiments were performed under approved protocols with guidelines established by the Pennsylvania State University College of Medicine Animal Care and Use Committee (Hershey, PA).

Cell lines and reagents

The B6/WT-19 (35) and TAP1^-/-/WT-Tag (27) cells, expressing full length wild type T Ag, have been previously described. B6/K-TagI cells (34) and B6/T122B1 cells (33) express full length T Ag variants in which sites II/III, IV and V or sites I, II/III, IV and V, respectively, have been inactivated by alanine substitutions at the MHC anchor positions. Cell line B6/TpLM249-15Bb (referred to as B6/Tag-IV-only) expresses a T Ag variant in which the sequence corresponding to sites I, II/III and V have been deleted (36). All cell lines were maintained in DMEM media as previously described (37). Lymphocyte suspensions were prepared in RPMI 1640 medium supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine, 10 mM HEPES, 50 μM 2-mercaptoethanol and 10% FCS. Peptides corresponding to influenza (Flu) nucleoprotein (NP) 366-374 (ASNENMETM), T Ag site I (₂₀₆SAINNYAQKL₂₁₅), T Ag site II/III (₂₂₃CKGVNKEYL₂₃₁), the T Ag site IV variant (₄₀₄VVYDFLKL_411L) and T Ag site V (₄₈₉QGINNLDNL₄₉₇) were synthesized in the Penn State Hershey Macromolecular Core Facility. Peptides were dissolved in DMSO and diluted to the appropriate concentration in RPMI 1640 medium prior to use.

Adoptive transfers and immunizations

TCR-I T cells were enriched from spleen and lymph nodes of TCR-I mice by AutoMACS sorting using CD8α+ microbeads (Miltenyi Biotec, Auburn, CA) according to the manufacturer's instructions. CD8-enriched cells were stained with anti-CD8 and Db/I tetramer to determine purity, which ranged between 85-90%. C57BL/6 mice received 10⁶ magnetically sorted TCR-I CD8+ T cells intravenously in 0.2 ml of HBSS unless otherwise noted. Where indicated, CD11c+ cells were depleted using CD11c microbeads (Miltenyi Biotec) and an AutoMACS sorter according to the manufacturer's instructions prior to CD8 enrichment. For immunizations, mice received an i.p. injection of 5×10⁷ live T Ag transformed cells.

Lymphocyte isolation and flow cytometry

Lymphocyte populations were isolated from spleens according to previously described protocols (38) and stained with PE-labeled MHC class I tetramers corresponding to the control Db/Flu (NP366-374), Db/I and Kb/IV at a 1:200 dilution in conjunction with anti-CD8 as previously described(4). Stained cells were fixed with 2% paraformaldehyde prior to flow cytometric analysis using a FACScan, FACSCalibur OR FACSCanto flow cytometer (BD Biosciences, San Jose, CA) housed within the Penn State Hershey Flow Cytometry Core Facility. Data were analyzed using FlowJo software (Tree Star, Ashland, OR). The percentage of epitope-specific CD8+ cells was determined by subtracting the percentage of cells that stained positive with the control Db/Flu tetramer from the percentage of cells that stained positive with the T Ag epitope tetramer. The following antibodies were purchased from eBioscience (San Diego, CA): PE-labeled anti-CD45.1 (clone A20), FITC- and eFluor450-labeled anti-CD8 (clone 53-6.7), FITC-labeled anti-CD11c (clone N418), FITC-labeled anti-IFN-γ (clone XMG1.2) and FITC-labeled anti-BrdU (clone PRB-1). PerCp-Cy5.5-labeled anti-CD8 (clone 53-6.7), PE-labeled anti-CD11c (clone HL3) and PE-labeled anti-VβTCR (clone H57-597) antibodies were purchased from BD Pharmingen.

Intracellular cytokine stain

One million spleen cells isolated from immunized mice were stimulated ex vivo with 1 μM of the indicated synthetic peptides or 5×10⁵ of the indicated T Ag transformed cells. Cultures were incubated at 37° C, 5% CO₂ for 6 hours, with 1 μg/ml brefeldin A added to each culture after 2 hours. Cells were then stained for CD8α and intracellular IFN-γ using the Cytofix/Cytoperm kit (BD Biosciences) per the manufacturer's instructions and analyzed by flow cytometry as above.

In vitro culture of T cells and ⁵¹Cr-release assay

Spleen cells from B6/K-TagI immunized mice were restimulated in vitro at 1×10⁷ cells/well of a 12-well plate with 5×10⁵ γ-irradiated (100 Gy) B6/K-TagI cells in 4 ml of complete RPMI-1640 medium. Cultures were passed 1:2 on day 7 and provided with fresh B6/K-TagI stimulators and 10 U/ml recombinant human IL-2 (kindly provided by Amgen, Thousand Oaks, CA). Cytotoxicity assays were performed on day 5-post restimulation and the percentage of specific target cell lysis determined using a previously described approach (36, 38).

In-vivo cytotoxicity assay

Splenocytes from sex-matched B6.SJL mice were incubated with 1 μM peptide for 1 hr at 37°C in complete RPMI medium. Cells were then washed three times and re-suspended at 10⁷/ml in PBS containing 0.1% BSA. Peptide pulsed targets were differentially labeled with the following concentrations of CFSE (Molecular Probes, Eugene, OR) for 10 min at 37°C: Flu NP pulsed (0.25 μM) site I-pulsed (0.5 μM) and site IV-pulsed (5 μM). Cells were washed twice and re-suspended at 4×10⁷/ml in PBS. Each mouse received 2×10⁶ cells of each target group in 150 μl by i.v. injection. The percentage of target cell elimination was assessed at 1, 4 or 16 hr post-adoptive transfer by gating on CD45.1⁺ cells. Percent in vivo cytotoxicity was determined using the following formula: [ratio of CFSE^hi (site IV) or CFSE^med (site I) cells to CFSE^lo cells in control unimmunized mice – ratio of CFSE^hi (site IV) or CFSE^med (site I) cells to CFSE^lo cells in experimental mice]/[ratio of CFSE^hi (site IV) or CFSE^med (site I) cells to CFSE^lo cells in control unimmunized mice].

In-vivo BrdU incorporation assay

Mice received two 1 mg doses of a 1 mg/ml BrdU solution (diluted in PBS) 12 hr apart by i.p. injection at the indicated times post-immunization. Splenocytes were harvested within 12 hrs of the last injection and stained for BrdU incorporation using a modified intracellular staining protocol (BD Biosciences). Briefly, lymphocytes were stained with PE-labeled tetramer and PerCp-Cy5.5-labeled anti-CD8 as described above. Cells were then re-suspended in 100 μl of Cytofix/Cytoperm (BD Pharmingen) and incubated for 30 min at room temperature. The cells were washed one time with 1× PermWash (BD Pharmingen), re-suspended again in 100 μl of Cytofix/Cytoperm and incubated for 10 min at room temperature. Cells were washed again and re-suspended in 100 μl of Cytofix/CytoPerm and incubated for 5 min at room temperature. After one wash, cells were incubated at 37°C for 1 hr with 100 μl of DNase I (300 μg/ml in Dulbecco's PBS) and washed once. Cells were then stained with 10 μl of FITC labeled anti-BrdU and 40 μl 1× PermWash per well for 20 min at room temperature. Cells were washed, fixed with 2% paraformaldehyde and analyzed by flow cytometry as described above.

In vivo Depletion of CD11c+ and NK1.1+ cells

CD11c-DTR mice received 2-4 ng/g body weight of diphtheria toxin (DT; D0564; Sigma-Aldrich, St. Louis, MO), i.p., suspended in PBS as previously described (22). C57BL/6 mice received 200 μg of anti-NK1.1 antibody (clone PK-136, BioXCell, West Lebanon, NH) diluted in PBS i.p. one day prior to and 3 days post-immunization. Control mice received an injection of rat IgG (Sigma-Aldrich).

Results

The early kinetics of in vivo cytotoxicity to site I and site IV correspond with the T Ag immunodominance hierarchy

We first investigated the relationship between the early kinetics of the T Ag-specific CD8+ T cell response and the immunodominance hierarchy previously observed at the peak of the CD8+ T cell response (33). Groups of C57BL/6 mice were immunized with B6/WT-19 cells, expressing wild type T Ag. Both in vivo cytotoxicity against peptide-pulsed target cells and ex vivo staining of lymphocytes with MHC tetramers were used to evaluate the presence of T Ag-specific T cells at early time points. SJL (CD45.1+)-derived splenocytes pulsed with peptides corresponding to site I, site IV or control Flu NP were injected at 2, 3, 4, 5 or 7 days post-immunization. The presence of CD45.1+ target cells in spleens was assessed after 16 hours and thus correspond to days 3, 4, 5, 6 and 8, respectively (Fig. 1A). No target cell killing was detected on day 3, but by day 4 68% of the site IV-pulsed targets had been eliminated and almost all of the site IV-pulsed targets were eliminated by day 5. In contrast, site I-specific activity lagged by 24 hours, with 48% of site I-pulsed targets eliminated by day 5 and complete elimination occurring only beyond 6 days post-immunization. MHC tetramer analysis revealed a small but detectable population of site IV-specific CD8+ T cells at day 6 post-immunization and this percentage increased dramatically by day 8 (Fig. 1B). Site I-specific CD8+ T cells were only detected above background levels by tetramer analysis at day 8. Thus, the in vivo cytotoxicity assay provided an approach for earlier detection of T Ag epitope-specific activity relative to MHC tetramer analysis. These results demonstrate that the kinetics of early detection correspond to the immunodominance hierarchy observed previously at the numeric peak of the response (33), with activity against site IV detected 24 hours prior to activity against site I.

Figure 1.

The immunodominance hierarchy to T Ag is established early following immunization. A, Groups of three C57BL/6 mice were immunized i.p. with wild type T Ag-transformed cells (B6/WT-19). A mixture of differentially CFSE-labeled and site I, site IV or Flu peptide-pulsed CD45.1⁺ target cells were transferred into the previously immunized mice at days 2, 3, 4, 5 and 7. After 16 hr, splenocytes were harvested and stained with anti-CD45.1 to determine the percentage of target cell killing. The data shown are from one mouse representative of three. The experiment was repeated twice with similar results obtained. B, Splenocytes from the mice in A were stained with anti-CD8 and Db/I or Kb/IV tetramer. Samples were gated on CD8+ cells and the percentage of tetramer positive cells is indicated. The error bars represent standard error. C, Groups of three B6 mice were immunized i.p. with B6/WT-19 cells. Seven days later mice received CFSE-labeled CD45.1 targets. Splenocytes were harvested 1, 4 and 16 hr post-adoptive transfer of targets. The splenocytes were then stained for CD45.1 and the percentage of target cell killing calculated. The data shown are from one mouse representative of three. The experiment was repeated twice with similar results obtained.

We observed complete elimination of site I- and site IV-pulsed targets by day 8 despite detection of dramatically different frequencies of CD8+ T cells by MHC tetramer staining at this time point (Fig. 1A and 1B). To determine whether the in-vivo killing assay could be modulated to better reflect this frequency difference, we determined the efficiency of target cell elimination at 1, 4 and 16 hours post-transfer of target cells into immunized mice. Approximately one hour post-transfer 71% of site IV-pulsed targets were eliminated, with 91% killed by 4 hours (Fig. 1C). In contrast, site I-specific activity was not detected until 4 hours post-transfer, with complete elimination of the targets occurring by 16 hrs. Thus, the efficiency of target cell killing is proportional to the frequency of T Ag epitope-specific CD8+ T cells. Overall these results establish that immunodominance of site IV is apparent very early following immunization.

Altering the naïve precursor frequency reverses the immunodominance hierarchy but does not block priming of endogenous T Ag-specific T cells

We next asked whether altering the initial precursor frequency of site I-specific CD8+ T cells would change the kinetics and/or magnitude of the T Ag-specific immunodominance hierarchy. One million CD8-enriched naïve site I-specific TCR transgenic (TCR-I) T cells were transferred into C57BL/6 mice followed by immunization with B6/WT-19 cells and injection of peptide pulsed CD45.1+ target cells at 3, 5 and 7 days post immunization. Four days post transfer, approximately 50% of the site I-pulsed targets were eliminated in mice that received TCR-I transfer, whereas no target cell killing was detected at this timepoint in mice that received immunization alone (Fig. 2A, left panels). Site I target cell elimination corresponded with the appearance of high levels of TCR-I T cells at early time points (Fig. 2A, right panels). The presence of TCR-I T cells did not significantly alter 16-hour in vivo cytotoxicity against site IV-pulsed target cells at any of the time points tested, but dramatically reduced the percentage of endogenous site IV-specific CD8+ T cells that accumulated in the spleens of immunized mice (Fig. 2A, lower panels). Consistent with this decrease in site IV-specific T cell accumulation, cytotoxicity was reduced to 52% on day 7 if the assay was harvested after only 4 hours (Fig. 2B). These results indicate that increasing the naïve CD8+ T cell precursor frequency i) accelerates the kinetics of the functional site I-specific response, ii) enhances total accumulation of site I-specific CD8+ T cells within the lymphoid organs and iii) limits accumulation of the endogenous site IV-specific CD8+ T cell response, effectively reversing the immunodominance hierarchy.

Figure 2.

Increasing the naïve precursor frequency of site I-specific CD8+ T cells reverses the immunodominance hierarchy. A, Groups of three mice were immunized with B6/WT-19 cells with and without adoptive transfer of 1×10⁶ TCR-I transgenic CD8+ T cells. At the indicated time points following immunization mice received CFSE-labeled CD45.1 target cells. After 16 hr splenocytes were harvested, stained for CD45.1 and the percentage of target cell killing determined (left panels). Imm only: mice receiving only immunization, TCR-I+Imm: mice receiving both TCR-I cells and immunization. Tetramer staining with Db/I and Kb/IV was performed on the same cell populations (right panels), with the percentage of CD8+ cells that were MHC tetramer+ positive indicated. This experiment was repeated twice with similar results obtained. B, Groups of three mice received either B6/WT-19 immunization or TCR-I cells and B6/WT-19 immunization. Seven days following immunization CFSE-labeled CD45.1+ targets were injected and after 4 hr isolated splenocytes were stained for CD45.1 and the percentage of killing was determined. C and D, Groups of three mice received either B6/WT-19 immunization and GFP+ TCR-I cells at the indicated doses or B6/WT-19 immunization alone (No AT). C, The frequency of site I- and IV-specific CD8+ T cells was determined by MHC tetramer staining of splenocytes 10 days post-immunization. Samples were gated on CD8+ cells and the percentage of tetramer positive CD8+ cells is indicated. D, The percentage of GFP+ and GFP- (host) site I-specific CD8+ T cells within the samples shown in C was determined. Error bars indicate standard error.

The above experiment revealed that the presence of excess site I-specific precursors inhibited accumulation of the endogenous site IV-specific CD8+ T cells, but did not completely block their initial activation. Thus, we determined the dose of TCR-I T cells required to inhibit accumulation of the endogenous site IV-specific CD8+ T cells. C57BL/6 mice received various doses of CD8-enriched GFP+ TCR-I T cells followed by immunization with B6/WT-19 cells. Doses of both 10⁵ and 10⁴ TCR-I T cells blocked accumulation of normal levels of endogenous site IV-specific CD8+ T cells, while normal levels of site IV-specific CD8+ T cells accumulated when only 10³ TCR-I cells were transferred (Fig. 2C). This result corresponded with a decreased accumulation of site I-specific T cells (Fig. 2C), and at the lowest dose also allowed the emergence of endogenous GFP negative site I-specific T cells (Fig. 2D). These results suggest that a large excess of site I-specific precursors is required to reverse the SV40 T Ag immunodominance hierarchy. Additionally, excess TCR-I cells block accumulation of the endogenous site I-specific CD8+ T cells more efficiently than the accumulation of endogenous site IV-specific CD8+ T cells.

Elimination of competing T Ag epitopes minimally alters accumulation of site I and IV-specific CD8+ T cells

Since the above experiments demonstrated that skewing the initial frequency of epitope-specific naïve CD8+ T cells reverses the immunodominance hierarchy, we investigated the potential role of competition among the physiologically low endogenous T Ag-specific CD8+ T cells in regulating T cell accumulation. C57BL/6 mice were immunized with either B6/K-TagI cells that express a T Ag variant in which sites II/III, IV and V have been inactivated or B6/Tag-IV-only cells that express a T Ag variant in which sites I, II/III and V have been eliminated. Control mice were immunized with B6/WT-19 cells expressing wild type T Ag. The frequency of T Ag epitope-specific CD8+ T cells was determined by MHC tetramer staining of splenocytes at 7, 10 and 14 days post-immunization. Site I-specific T cells showed a significant increase in accumulation at the peak of the response at day 10 following immunization with B6/K-TagI cells compared to mice immunized with B6/WT-19 cells (Fig. 3), but importantly did not expand to levels typically achieved by site IV-specific CD8+ T cells. A similar trend was observed for site IV-specific T cells following immunization with B6/Tag-IV-only cells, although this increase was not statistically significant. These results indicate that T cell competition has only a minor impact on the magnitude of the endogenous primary CD8+ T cell response to either determinant.

Figure 3.

T cell competition plays a minor role in limiting the magnitude of the endogenous T cell response. A, Groups of three C57BL/6 mice were immunized with either B6/WT-19, B6/K-TagI or B6/Tag-IV-only cells. At the indicated time points splenocytes were harvested and stained with anti-CD8 and Db/I or Kb/IV tetramers. The data represent the percentage of CD8+ cells that are tetramer positive. *p value = 0.0458 as determined by the Student's T test.

Inactivation of site IV in the T Ag variant expressed by B6/K-TagI cells could potentially trigger CD8+ T cells responding to novel immunodominant determinants following immunization, blocking further accumulation of site I-specific T cells. To examine this possibility, we evaluated the CD8+ T cell response of B6/K-TagI-immunized C57BL/6 mice toward cells expressing a T Ag variant, B6/T122B1, in which all four known H-2^b-restricted determinants were inactivated. Splenocytes from B6/K-TagI-immunized mice failed to produce significant amounts of IFN-γ in ex vivo intracellular cytokine assays following stimulation with B6/T122B1 cells but responded to both T Ag site I peptide and B6/K-TagI cells as expected (supplemental Fig. 1A). Similarly, in vitro stimulated splenocytes from the same mice failed to lyse B6/T122B1 cells in ⁵¹Cr-release assays (supplemental Fig. 1B). These results demonstrate that the inability of site I-specific CD8+ T cells to expand to the levels achieved by site IV-specific T cells is not due to the recruitment of CD8+ T cells targeting novel immunodominant determinants. The data rather support the conclusion that T cell competition plays only a minor role in establishment of the immunodominance hierarchy.

Direct presentation by T Ag transformed cells enhances the magnitude and extends the kinetics of the CD8+ T cell response

Previous studies from our laboratory suggested that both cross-presentation by host cells and direct presentation by the T Ag transformed cells contribute to the magnitude of the CD8+ T cell response toward the immunorecessive T Ag site V determinant (27). To evaluate the contribution of direct presentation toward the CD8+ T cell response to T Ag sites I and IV, we quantified the response to both determinants following immunization with wild type T Ag transformed TAP1-deficient (TAP1^-/-/WT-Tag) cells versus B6/WT-19 cells. The results demonstrate that both the kinetics and magnitude of T cell accumulation was altered in mice immunized with TAP1^-/-/WT-Tag cells relative to mice immunized with B6/WT-19 cells, with the peak of the response occurring at day 7 for the former (Fig. 4A). In contrast, site I- and site IV-specific CD8+ T cells continued to accumulate in the spleens of mice immunized with B6/WT-19 cells out to day 10. These results indicate that the magnitude and duration of the site I- and site IV-specific CD8+ T cell responses are reduced under conditions in which only cross-presentation of T Ag occurs, although the immunodominance hierarchy remains the same.

Figure 4.

The magnitude of T Ag-specific CD8+ T cells is reduced at late time points following immunization with TAP1^-/- cells. A, Groups of three mice were immunized with either B6/WT-19 or TAP1^-/-/WT-Tag (TAP1^-/-) cells. At the indicated days splenocytes were harvested and the frequency of site I- and site IV-specific CD8+ T cells was determined by staining with MHC tetramers. The data are representative of two independent experiments with similar results. B, Groups of three mice received either control IgG or anti-NK1.1 one day prior to and three days post-immunization with either B6/WT-19 (B6) or TAP^-/-/WT-Tag (TAP1^-/-) cells. At the indicated days splenocytes were harvested and the frequency of site I- and site IV-specific CD8+ T cells determined by staining with MHC tetramers. The data are representative of two independent experiments in which similar results were obtained. C, Splenocytes from B were stained with anti-NK1.1 to determine the efficiency of NK cell depletion. The gate represents the percentage of NK1.1 positive cells. Representative plots from the IgG and NK1.1 depleted groups are shown.

One possible explanation for these results is that the B6- and TAP1^-/--derived cells express different amounts of T Ag, leading to differential T cell triggering and expansion. However, we found that the TAP1^-/-/WT-Tag and B6/WT-19 cells express similar levels of T Ag by western blot analysis (data not shown), suggesting that the observed differences are explained by a lack of MHC surface expression rather than variation in the amount of antigen used for immunization. We also considered that the difference in T cell magnitude could be attributed to increased susceptibility of TAP1^-/-/WT-Tag cells to NK cells (39), resulting in accelerated removal of TAP1^-/-/WT-Tag cells. To examine this possibility we depleted B6 mice of NK cells by injection of anti-NK1.1 mAb prior to immunization and quantified the site IV-specific CD8+ T cell response at 7 and 10 days post-immunization. We chose to examine the more robust site IV-specific response in this experiment. Depletion of NK cells following administration of anti-NK1.1 was verified by flow cytometry (Fig. 4C). The results indicate that depletion of NK cells did not enhance the response to site IV following immunization with TAP1^-/-/WT-Tag cells on day 7 or 10 (Fig. 4B). Therefore, the minimal T Ag-specific response generated by TAP1^-/-/WT-Tag immunization cannot be attributed to enhanced NK cell mediated clearance of TAP1^-/- cells.

We next considered that the reduced accumulation of site I and site IV-specific CD8+ T cells following immunization with TAP1^-/-/WT-Tag cells might be explained by an earlier termination of CD8+ T cell proliferation. To determine the proportion of proliferating cells at various times following immunization, we utilized in vivo BrdU incorporation. Since total T Ag-specific T cell numbers are low at early time points post immunization, we examined the proliferative response of adoptively transferred TCR-I T cells. As outlined in Figure 5a, B6 mice received TCR-I T cells plus immunization with either B6/WT-19 or TAP1^-/-/WT-Tag cells the same day. Mice received BrdU injections at 12 and 24 hours prior to analysis at the indicated times post immunization. TCR-I cells accumulated to similar levels by day 5 following immunization with either TAP1^-/-- or B6-derived T Ag transformed cells (Fig. 5A). However, the proportion of cells that incorporated BrdU in the 24 hours preceding analysis was 5-fold higher in mice immunized with B6/WT-19 cells. By day 6, TCR-I T cells had declined to 3.5% in TAP1^-/-/WT-Tag immunized mice but continued to expand in B6/WT-19 immunized mice, with 46% still proliferating. To ensure that TCR-I T cells could proliferate in response to immunization with TAP1^-/-/WT-Tag cells, some mice were assessed at three days post immunization. At this early time point, a similar proportion of TCR-I T cells incorporated BrdU in response to both B6/WT-19 and TAP1^-/-/WT-Tag immunization (Figure 5B). These data indicate that the increased magnitude of T Ag-specific CD8+ T cells induced by immunization with B6/WT-19 cells is associated with a prolonged expansion phase and suggest that the proliferative response to cross-presented T Ag subsides prior to five days post immunization.

Figure 5.

TCR-I T cell proliferation is prolonged in response to directly presented antigen. A and B Groups of three mice received 1×10⁶ TCR-I CD8+ T cells and were immunized with either B6/WT-19 (B6) or TAP1^-/-/WT-Tag (TAP1-/-) cells. As indicated in the timeline mice received two injections of BrdU 12 hrs apart on days 5 or 6 (A) or day 3 (B). Splenocytes were harvested 12 hr after the last injection to determine the frequency of site I-specific CD8+ T cells (dot plots) and the proportion of epitope-specific BrdU+ cells (histograms). The percentage of MHC tetramer+ CD8+ cells and the percentage of Db/I+ CD8+ cells incorporating BrdU are indicated. C, Groups of three mice were immunized with B6/WT-19 or TAP1^-/-/WT-Tag cells. After 7 days, mice received 1×10⁶ CFSE-labeled TCR-I CD8+ T cells. After three days the frequency of CD8+ cells specific for site I was determined by MHC tetramer staining (dot plots) and the average percentage of proliferating Db/I+ CD8+ cells calculated (histograms). Standard deviation is indicated. This experiment was repeated with similar results.

We next investigated whether our results could be explained by differences in antigen duration in vivo following immunization with TAP1^-/-/WT-Tag cells versus B6/WT-19 cells. To examine this possibility, B6 mice were immunized with either TAP1^-/-/WT-Tag or B6/WT-19 cells and 7 days later received CD8-enriched CFSE-labeled naïve TCR-I T cells. Analysis of splenocytes isolated after three days showed that TCR-I T cells proliferated to a similar extent in both groups of mice (Figure 5C). These results indicate that cross-presentation of T Ag site I persists up to 7 days post immunization with either TAP-deficient or TAP-sufficient T Ag transformed cells. Since this length of time is beyond the point when TCR-I T cells showed decreased proliferation in response to TAP1^-/-/WT-Tag immunization, antigen duration is likely not a major limiting factor.

CD11c+ cells are not required at late times post immunization to promote T Ag-specific CD8+ T cell accumulation

CD11c+ dendritic cells, particularly the CD8α subset, have been shown to be important for cross-priming CD8+ T cells (40). To evaluate whether CD11c+ cells are needed to sustain T cell proliferation following immunization with TAP-sufficient B6/WT-19 cells, we determined whether elimination of CD11c+ cells following initial T cell expansion altered accumulation of TCR-I T cells at day 7. In this approach, CD11c-DTR mice, which express the high affinity simian diphtheria toxin (DT) receptor from the CD11c promoter (22), received 1×10⁶ CD8-enriched TCR-I T cells and immunization with B6/WT-19 cells. CD11c+ cells were depleted on day 4 post immunization by injection of DT and the frequency of site I-specific CD8+ T cells determined on day 7 (Fig. 6B). Injection of 2ng/g DT into CD11c-DTR mice resulted in the depletion of approximately 90% of CD11c+ cells (Fig. 6A). We note that similar to previous reports (22), GFP expression on CD11c+ cells was low in CD11c-DTR mice. Despite elimination of CD11c+ cells at day 4, site I-specific CD8+ T cells accumulated to equivalent levels in CD11c-depleted versus control treated mice (Fig. 6B).

Figure 6.

CD11c+ cells are not required late in the primary response to promote T cell accumulation. A, CD11c-DTR mice (3/group) received 2ng/g DT and splenocytes were harvested after 16 hours and stained for CD11c. The figure shows a representative mouse from the group. The percentage of CD1lc+ cells of total splenocytes is indicated. B, Groups of three CD11c-DTR mice received 1×10⁶ TCR-I CD8+ T cells and immunization with 5×10⁷ B6/WT-19 cells. As indicated in the timeline, mice received 2ng/g DT i.p. on day 4 and splenocytes were harvested on day 7. Splenocytes were stained with Db/I tetramer and anti-CD8 to determine the frequency of Db/I+ CD8+ cells. C, Groups of three CD11c-DTR mice received 1×10⁶ CFSE-labeled TCR-I CD8+ T cells. The next day, mice were injected with 2ng/g DT and after a further 12 hours were immunized with 5×10⁷ B6/WT-19 cells. Proliferation of CFSE-labeled TCR-I CD8+ T cells was determined after 3 days by co-staining with anti-CD8 and Db/I tetramer. Dot plots show the frequency of CD8+ cells specific for site I and histograms show the percentage of CD8+ Db/I tetramer+ cells that divided. The data are representative of mice from each group.

To ensure that cross-presentation by CD11c+ cells was reduced in DT treated mice, we determined the efficiency of cross-priming of naïve TCR-I T cells following transfer into freshly treated CD11c-DTR mice. CD11c-DTR mice received CFSE-labeled TCR-I T cells followed by injection of DT. Mice were then immunized with B6/WT-19 cells 12 hours following depletion and three days later the percentage of proliferating TCR-I T cells was determined. In the absence of CD11c+ cells, TCR-I T cells failed to proliferate by three days post immunization, while proliferating cells represented 93% of TCR-I T cells in non-depleted mice (Fig. 6C). Results similar to those presented in figure 6B were obtained when CD11c+ cells were depleted from TCR-I donor cells prior to CD8-enrichment and adoptive transfer into CD11c-DTR mice (supplemental Fig. 2). This result indicates that co-transfer of low numbers of contaminating CD11c+ CD8α+ cells, known to efficiently cross-prime naïve T cells in vivo (40, 41), is unlikely to contribute to the late proliferative response following immunization with B6/WT-19 cells. Taken together, the results indicate that CD11c+ cells are necessary during the initial priming of TCR-I T cells following immunization with B6/WT-19 cells, but are not required later in the primary response to promote maximum accumulation. This finding is consistent with the hypothesis that cross-presentation drives the early proliferative response, but direct presentation by T Ag transformed cells is required to promote extended CD8+ T cell proliferation.

TAP1-sufficient but not TAP1-deficient cells drive rapid accumulation of T Ag-specific CD8+ T cells late in the primary response

Since contraction of TCR-I T cells is initiated by day 6 following immunization with TAP1^-/-/WT-Tag cells (Fig. 5B), we asked whether provision of a second dose of TAP1^-/-/WT-Tag or B6/WT-19 cells at day 5 could promote further accumulation of TCR-I T cells in mice initially immunized with TAP1^-/-/WT-Tag cells. The frequency of TCR-I T cells was assessed four days following the second immunization (day 9 post primary immunization). Control mice received only a single immunization at day 0 or day 5. Mice that received immunization only at day 0 with TAP1^-/-/WT-Tag cells achieved only 6% TCR-I T cells at nine days post immunization compared to 42% in mice immunized with B6/WT-19 cells (Fig. 7). Mice that received a second dose of TAP1^-/-/WT-Tag cells showed no additional expansion of TCR-I T cells. In contrast, TAP1^-/-/WT-Tag-immunized mice that received B6/WT-19 cells for the second dose at day 5 had approximately 40% site I-specific CD8+ T cells by day nine. Importantly, day 5 immunization with B6/WT-19 cells alone yielded less than 10% site I-specific CD8+ cells. This result demonstrates that addition of cells directly presenting T Ag site I at a time coincident with initiation of T cell contraction fully restores T cell accumulation. Importantly, this prolonged response was not induced following immunization with cells that can only donate antigen for cross-presentation, indicating that antigen load is not a limiting factor in determining the magnitude of the T Ag-specific CD8+ T cell response.

Figure 7.

Provision of direct presentation at late times post T cell activation restores normal T cell accumulation. Groups of three mice received 1×10⁶ TCR-I CD8+ T cells and immunization on day 0 with either TAP1^-/-/WT-Tag (TAP) or B6/WT-19 (B6) cells. Five days after the initial immunization, groups of mice received a second immunization with either B6/WT-19 or TAP1^-/-/WT-Tag cells. Control groups were immunized with B6/WT-19 or TAP1^-/-/WT-Tag cells only at day 0 or day 5. On day 9, splenocytes were harvested and stained with anti-CD8 and Db/I tetramer to determine the frequency of site I specific T cells within the CD8+ population. Error bars indicate standard error.

Discussion

The results of this study demonstrate that the CD8+ T cell response to cell-associated antigen is dramatically enhanced by 1) increasing the naïve precursor frequency and 2) the presence of nonhemopoietic cells that can directly present the relevant determinants. While the former is regulated by development of the immune system (42), the latter can be manipulated during immunization to optimize the number of responder cells. Our results indicate that maximal T cell accumulation is achieved by prolonging T cell proliferation during the primary response in the presence of direct presentation. However, functional immunodominance is established early after immunization, even when antigen is only cross-presented.

Our finding that the in vivo kinetics of CD8+ T cell-mediated killing is staggered for T cells targeting two unique determinants within the same protein antigen is best explained by differences in the initial precursor frequency. Several studies have now shown a correlation between the naïve precursor frequency and the relative magnitude of the T cell response to a given determinant (8-10). Using a similar approach, we have acquired preliminary data indicating that the ratio of site IV:site I T cell dominance is maintained from the naïve state to the peak of the response (A. Tatum and T. Schell, unpublished observations). This observation is consistent with results from Obar et al. (10), demonstrating that the hierarchy of naïve CD8+ T cells targeting various virus-derived determinants is maintained following virus infection, indicating that T cell expansion is proportional over the course of the primary response. A recent elegant study by van Heijst et al (43) demonstrates that a high percentage of naïve T cells targeting a particular determinant are recruited into the response following virus infection, regardless of the infectious dose, suggesting that the basis for maintenance of the pre-immune hierarchy during the primary immune response is the nearly complete recruitment of the available clones. Thus, the early absence of site I- versus site IV-specific function observed following immunization is likely due to the subthreshold levels of site I-specific T cells that have accumulated rather than a lack of T cell priming.

An alternate explanation for the accelerated response of site IV-specific CD8+ T cells relative to site I-specific T cells is a potential difference in T cell avidity, allowing more efficient expansion of site IV-specific T cells following immunization. Such a mechanism is supported by recent findings from Zehn and colleagues in which T cells targeting low affinity ligands had a shorter duration of proliferation than T cells targeting higher affinity ligands, leading to lower T cell accumulation and more rapid contraction (44). While we have shown that detection of a functional site I-specific response is delayed in T Ag-immunized mice, our results indicate that the initiation of endogenous T cell contraction is similar for T cells responding to both site I and site IV (Fig. 3). We have not observed differences in the overall avidity of CD8+ T cell populations responding to sites I and IV (data not shown), but this may not represent the avidity of the early response to immunization since the population may be skewed toward higher avidity clones late in the response (44). We note that previous studies failed to define a clear relationship between immunodominance and T cell avidity/TCR affinity. Several studies revealed only a partial correlation (45-47), while others found no correlation (48) or an inverse correlation of T cell avidity with immunodominance (49). Overall these findings suggest that while T cell avidity may be important for prolonging proliferation of T cell clones within a population, a strict correlation with immunodominance has not been demonstrated.

Whether the staggered functional response toward site I and site IV has any biological significance remains unknown. An intriguing possibility is that naïve precursors are present at levels that provide the least interdeterminant interference while allowing optimal accumulation in response to antigenic challenge. For example, our results indicate that increasing the number of naïve site I-specific T cells was detrimental to the accumulation of site IV-specific T cells. However, when T cells are present at physiological levels, little to no competition was observed that limited T cell accumulation in response to wild type T Ag versus T Ag variants lacking three of the four determinants (Figure 3). Similar results have been observed in infectious systems using Listeria monocytogenes and LCMV (8, 15). Elimination of the dominant determinants in these pathogens did not change the hierarchy nor enhance the immune response to the subdominant determinants. Interestingly, Kotturi and colleagues found that there is an overall decrease in the total CD8+ T cell response following infection with a LCMV variant lacking the dominant determinants (8). In this case, CD8+ T cell responses to the subdominant determinants could only recover 30% of the lost dominant response. These results suggest that CD8+ T cell accumulation during the primary response is limited by intrinsic factors other than T cell competition, such as the naïve precursor frequency. Our results directly support this idea as the addition of naïve site I-specific T cells resulted in a dramatically enhanced response to site I.

That the site IV-specific CD8+ T cell response may be more biologically important than the response to site I is indicated from studies in SV40 T Ag transgenic mouse models in which T cells targeting the immunodominant site IV determinant can prevent the appearance of new pancreatic tumors (50) and are sufficient to promote the complete regression of established brain tumors (4). This latter result was associated with rapid accumulation of adoptively transferred site IV-specific CD8+ T cells within the brain followed by prolonged T cell persistence at the tumor site. In contrast, site I-specific CD8+ T cell entry into the brains of tumor-bearing mice is delayed despite the presence of a significant population of activated T cells among the donor lymphocytes, and their loss from the brain is rapid relative to site IV-specific CD8+ T cells (5). This difference may be partially explained by an increased susceptibility of site I-specific CD8+ T cells to tolerance induction (50). The characteristics of the site IV-specific CD8+ T cell response that promote more efficient control of tumors remain to be elucidated.

In the current study, the immunodominance hierarchy of site IV>site I was not altered following immunization with TAP1^-/- cells and a substantial number of T cells accumulated in response to cross-presentation as was found by Chen et al. in a previous study (23). However, the response to both determinants was increased in mice immunized with TAP-expressing cells due to an extended period of T cell proliferation. Thus when antigen could only be cross-presented, T cell accumulation at later time points was dramatically reduced compared to when antigen was available through both cross- and direct-presentation. This was despite nearly identical expansion kinetics prior to day 7. A recent study by Thomas et. al. (31) showed that during LCMV infection, virus-specific CD8+ T cell accumulation is maximized when antigen is presented on nonhemopoietic host cells in addition to pAPCs. However, accumulation of CD8+ T cells in response to VSV, vaccinia virus and Listeria monocytogenes infections was not dependent on presentation by nonhemopoietic host cells. The authors suggested this difference may be explained by increased survival of nonhemopoietic cells during LCMV infection or to differences in the inflammatory environments. In the experiments shown here, both forms of presentation likely contribute to the overall T Ag-specific immune response, with antigen presentation by CD11c- APCs driving an extended period of T cell proliferation.

Previous studies have shown that cross-presentation of antigen can be biased toward or against particular determinants (27, 51-53), raising the possibility that differences in the magnitude of the site I and site IV-specific CD8+ T cell responses could be explained by differences in the efficiency of cross-presentation of these two determinants. In particular, Wolkers and colleagues showed that determinants located within a functional signal peptide were inefficiently cross-presented in vivo, while relocation of the determinant within the same protein or inactivation of the signal peptide promoted efficient cross-priming of epitope-specific CD8+ T cells (51). Likewise Ma and coworkers found that the efficiency of cross-presentation of the SIINFEKL determinant is influenced by identity of the flanking amino acids (53). Previous studies with SV40 T Ag revealed that the immunorecessive site V determinant is weakly cross-presented in vivo relative to the immunodominant site I determinant (27), but exchange of site V and site I within T Ag did not enhance the immunogenicity of site V nor decrease the immunogenicity of site I (36), suggesting that protein context alone does not dramatically alter the immunogenicity of the T Ag determinants. While a similar study has not been performed for sites I and IV, immunization with recombinant vaccinia viruses expressing the individual determinants as either cytosolic or endoplasmic reticulum-targeted minigenes induces the same hierarchal CD8+ T cell response (site IV > site I) observed following immunization with full length T Ag (33). These results indicate that the observed immunodominance of CD8+ T cells responding to site IV versus site I is independent of the protein context. In addition, the current study reveals that the magnitude of the CD8+ T cell response to both sites I and IV is diminished when T Ag is only cross-presented, suggesting that cross-presentation is not biased in this instance.

Previous in vitro studies have shown that T cells undergo a programmed number of cell divisions following initial antigen encounter (54, 55), while in vivo studies have revealed that the magnitude of CD8+ T cell responses is proportional to the duration of antigen presentation in vivo (18, 43, 56-58). These studies particularly highlight the impact of antigen levels within the first few days following infection/immunization on T cell accumulation, as changes in antigen levels beyond the first few days had little impact on T cell accumulation. One exception was found using flank infections with HSV, in which a period of in vivo antigen presentation only beyond 4 days promoted optimal HSV-specific CD8+ T cell accumulation (58). This correlation was limited, however, to APCs that could prime naïve T cells, leaving in question any role of non-pAPCs in driving T cell accumulation. In the noninfectious model shown here, we found that in vivo antigen presentation capable of priming naïve T cells persisted at similar levels for up to 7 days, regardless of the cells use for immunization. In addition, CD11c+ cells were not required late in the primary response in order for T Ag-specific CD8+ T cells to accumulate to maximal levels following immunization with B6-derived cells. Taken together these results suggest that after initial T cell activation through cross-priming, interactions with non-CD11c+ APCs drive an extended period of T cell proliferation. While this source of antigen could involve direct presentation by the T Ag transformed cells used for immunization, our results do not rule out the possibility of antigen transfer to a non-CD11c+ cell population.

The question of when direct presentation contributes to extended T cell proliferation remains unknown. Our evidence demonstrates that directly presented antigen can be provided late in the primary response to allow recovery of maximum T cell accumulation. However, we cannot rule out that encounter with directly presented antigen at early time points induces a proliferative program of increased duration. The nature of antigen presentation by non-pAPCs, particularly at earlier time points, may lead to enhanced T cell proliferation due to the absence of negative regulatory signals, such as CTLA-4:B7 interactions (59). These questions will be important to address for optimal design of vaccination approaches targeting CD8+ T cells and to better understand the mechanisms that regulate the CD8+ T cell response following cell-based vaccination approaches.

Abbreviations used in this paper

B6

C57BL/6

T Ag

T antigen

SJL

B6.SJL

DT

diphtheria toxin

DTR

diphtheria toxin receptor

LCMV

lymphocytic choriomeningitis virus

VSV

vesicular stomatitis virus