Purified MHC Class I and Peptide Complexes Activate Naive CD8⁺ T Cells Independently of the CD28/B7 and LFA-1/ICAM-1 Costimulatory Interactions Free

Tcells recognize antigenic peptide fragments bound to MHC-encoded molecules (1, 2) through the TCR (3, 4). The generation of a signal through the TCR is modulated in multiple ways, including TCR fine specificity, MHC alleles, Ag processing (5), and alterations in peptide ligands (6). In addition to the complex TCR-ligand interaction, other signals may play a major role in the control of T cell responses. A model for this suggests that at least two signals (7, 8), one through the TCR and the second through another receptor, are required for T cell activation. This has been called the “two-signal model,” and the delivery of the second signal has been labeled costimulation (9, 10). TCR stimulation without costimulation decreases the proliferative and IL-2 response of CD4⁺ T cell clones (11, 12) and previously activated CD4⁺ cells (13, 14). These T cells become anergic or unresponsive to a subsequent TCR stimulation. Naive CD4⁺ cells have a greater reduction than memory T cells in IL-2 production when stimulated without a second signal (13, 14). IL-2 production is considered a key factor in the successful activation of T cells, since addition of exogenous IL-2 can reverse or prevent anergy in in vitro models (15). A number of studies suggest that the CD28 molecule plays a key role in the delivery of the second signal (16) with engagement of the the CD28 ligands B7-1 and B7-2 (17, 18, 19, 20). The functions of CD28 include maximization of IL-2 production (21), promotion of late cell cycle progression (22, 23, 24, 25), and regulation of apoptosis (26). LFA-1 is another molecule that has been shown to play an important role in T cell activation (27, 28).

Although the above studies have evaluated CD4⁺ T cells, the second signal can play a major role in CD8⁺ T cell activation as well. A CD8⁺ T cell clone was anergized for IL-2 production in the absence of a second signal (29). Purified resting CD8⁺ cells, without the addition of IL-2, were only activated to become functional cytotoxic lymphocytes if stimulated with APCs expressing the CD28 ligand B7 (30, 31). In addition, naive TCR transgenic (Tg)² CD8⁺ T cells were unable to produce IL-2 in response to peptide-MHC complexes presented by Drosophila cells in the absence of B7 or ICAM-1 molecules (32, 33). However, spleen cells with a high density of peptide-MHC complexes, through addition of exogenous peptide, can drive early CD8⁺ T cell proliferation, despite blockade of the CD28 pathway (24, 25). These peptide-loaded spleen cells were unable to drive late T cell proliferation without costimulatory signals. Thus, these studies lead to differing conclusions about the role of costimulatory molecules in CD8⁺ T cell activation.

Studies performed with naive CD8⁺ T cells have used either mAbs to stimulate the T cells (which bypass the physiologic ligand) or APCs. APCs provide a background of undefined molecular interactions, in addition to the interactions of CD28 and LFA-1, that may influence the T cell response. Nonmammalian APCs may have subtle effects on MHC folding and release mediators of lymphocyte responsiveness, such as DNA (32). Differences in some of the above-mentioned studies, which used the same Tg TCR T cells and similar peptides, are most likely due to differences in the APCs. To directly examine the requirements of costimulation in this Tg TCR system, we have developed an APC-free system.

This system allows for the evaluation of the activation requirements of naive CD8⁺ T cells when stimulated with the physiologic ligand (MHC class I and peptide) in the absence of APCs. TCR Tg mice provide a source of unprimed naive CD8⁺ T cells with known specificity. The 2C TCR recognizes p2C peptides in the context of an H-2L^d allo-MHC molecule. Using immobilized H-2L^d and p2C peptides, we have previously shown that naive CD8⁺ T cells proliferate, produce IL-2, and mature into cytotoxic effector cells at high T cell densities (34).

In this study, we demonstrate that optimally loaded purified MHC-peptide complexes can stimulate a high percentage of naive CD8⁺ T cells at low cell densities. In addition, H-2L^d-peptide-induced CD8⁺ T cell activation failed to be blocked with Abs to the LFA-1 or B7 molecules, thus ruling out T-T cell costimulation through these molecules. Finally, purified MHC-peptide complexes can generate a proliferative response that remained elevated at 96 h poststimulation.

Peptides were obtained from the Center for Biologics Evaluation and Research Facility for Biotechnology Resources (Bethesda, MD). Peptides were synthesized on an ABI 433 peptide synthesizer (Applied Biosystems, Foster City, CA), and purity was determined by capillary electrophoresis (P/ACE 5000, Beckman, Palo Alto, CA). The peptide purity was determined to be >99.5%. The following peptides were used in these assays: p2Ca, LSPFPDL; p2C-Y4, LSPYPDL; and p2C-QY5, QLSPYPFDL. Peptides were diluted in water and filter sterilized before use.

Generation of the soluble H-2L^d MHC class I molecule was previously published (34). Briefly, the soluble H-2L^d construct was generated by the ligation of an XbaI fragment of the genomic H-2L^d gene containing the α1 and α2 domains to the α3 domain of H-2D^d and the Q10b tail. The H-2L^d soluble construct was cotransfected with pSV₂neo into DAP3 cells using lipofectamine (Life Technologies, Grand Island, NY). High producer clones of soluble H-2L^d protein were expanded. H-2L^d was affinity purified on an anti-H-2L^d (30.5.7) (35) sepharose column from cell culture supernatants. The purified protein was dialyzed against PBS and filter sterilized before use. Variation in the H-2L^d preparations has been minimal, as determined by their similar binding to immobilized 30.5.7 by ELISA assay and the SDS-PAGE gel pattern.

Clonotypic 2C TCR Tg mice (36) specific for the endogenous peptide p2Ca derived from the α-ketoglutarate dehydrogenase protein (37) were a generous gift from Dr. Dennis Loh (Nippon Roache Research Center, Kamakura-sh, Japan). These mice were bred back to C57BL/6 mice for more than six generations in sterile microisolators and fed sterilized food and water ad libitum. At 3 wk old, mice were screened by PCR for the presence of the transgene from tail DNA treated with proteinase K (0.5 mg/ml). The presence of the transgene was further confirmed by flow cytometry with anti-Vβ8.1/8.2 mAb (MR5-2) (PharMingen, San Diego, CA) before each in vitro assay. All mice used in these studies were 8 to 12 wk old. Female B10.D2 mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and maintained in the same pathogen-free conditions. Sentinel mice were consistently negative for pathogens.

Commercially available columns (R & D Systems, Minneapolis, MN) were used to purify CD8⁺ spleen T cells by negative selection. Briefly, a single cell suspension of mouse spleens (2 × 10⁸ cells/column) was incubated with a mixture of mAb for 10 min at room temperature. After washing to remove unbound Ab, the cells were loaded onto the Ig-coated glass beads column and incubated 15 min. The cells were eluted, washed, and resuspended in complete medium. The negatively selected cell population obtained in this manner had the following phenotype: >90% V_β 8.1,8.2 TCR positive; 10 to 20% CD44^high; <3% Ia⁺; <5% CD16/32⁺; and <3% CD45R0⁺, as assessed by flow cytometry.

For sorting, the column-purified CD8⁺ T cells were stained with α-CD8 FITC (53-6.7) (PharMingen) vs α-CD44 phycoerythrin (PE; IM7) using 1 μg of Ab per 10⁶ cells. The sorted population (CD8⁺CD44⁻) was >99.5% pure as determined by postsort analysis. Cell sorting was performed with a FACStar^Plus cell sorter (Becton Dickinson, Mountain View, CA) equipped with an argon laser at 4880 A. Data analysis was performed with the Cell Quest software (version 1.12). The sorter was calibrated for fluorescence channels 1 and 2 with Immunocheck beads (Coulter Immunology, Hialeah, FL) and for forward and side scatter with Calibrate beads (Becton Dickinson) for peak separation. The column-purified CD8⁺ T cells were gated on forward vs side scatter for size and viability, and the cell population was collected based on fluorescence channel 1 vs 2.

H-2L^d (0.7 μg/well) was coated overnight on 96-well microtiter plates (Immulon 4, Dynatech, Chantilly, VA) at 4°C. The plates were washed three times with PBS and blocked with 2% BSA-PBS. After incubation at room temperature for 30 min the wells were washed three times with PBS. To enhance the unloading of endogenous peptide and the loading of the exogenous peptide, wells were incubated in 100 μl of citrate phosphate buffer (pH 6.5), p2C peptide (30 μM), and human β₂-microglobulin (hβ₂m) (Calbiochem, La Jolla, CA) (1 μg/well) (38). After 2 h at 37°C, the plates were washed three times with PBS. CD8⁺ T cells, in a final volume of 0.2 ml of complete medium (RPMI 1640 (BioWhittaker, Walkersville, MD) plus 10% FCS, 100 U/ml penicillin/streptomycin, nonessential amino acids, 2 μM l-glutamine, and 50 μM 2-ME), were added at 3 × 10⁴/well unless otherwise noted. Due to the rapid off rate of the p2C peptides, peptide and hβ₂m were added at 30 μM and 1 μg/well, respectively, at the beginning of the culture, unless otherwise noted. The CD8⁺ T cells were incubated at 37°C/5% CO₂. APC were obtained by irradiating (1500 rads) B10.D2 single-cell splenocyte suspensions. For determination of proliferation, cells were pulsed with 1 μCi/well of [³H]thymidine (DuPont/NEN Research, Boston, MA), which was added during the last 12 h of the culture period. The cells were harvested onto filter mats, and the incorporated radioactivity was measured in a Betaplate scintillation counter (LKB-Pharmacia, Piscataway, NJ).

All mAbs (FITC, PE, or unlabeled) used in our studies were purchased from PharMingen. In the blocking experiments, purified mAbs (no azide, low endotoxin) α-LFA-1 (M17/4), α-CD80 (1G10), α-CD86 (GL1), and rat IgG2a (isotype control) were added at 10 μg/ml at the beginning of the culture. The percentage of inhibition was calculated as follows: 1 − (sample + Ab cpm/sample − Ab cpm) × 100. The background cpm were less than 500 cpm.

Column-purified cells were stimulated on immobilized MHC and peptide. After 4 h of incubation at 37°C/5% CO₂, cells were washed two times with PBS and incubated for 15 min in 0.5 μM EDTA at 37°C/5% CO₂. Recovered cells were washed in complete medium and resuspended in 100 μl of FACS buffer (PBS, 5% FCS, and 0.1% NaN₃). Staining was performed using 1 μg/10⁶ cells and incubating with α-CD69 FITC (H1.2F3), α-CD8 PE (53.6.7), and α-CD3 FITC (145-2C11) for 30 min in ice. A total of 10⁶ events per sample were acquired using a FACScan (Becton Dickinson) flow cytometer. Data analysis was performed using the Cell Quest software. Fluorochrome-conjugated isotype-matched rat mAbs were used as negative controls.

For IL-2 determination, 100 μl of 48-h culture supernatants were assayed with a commercially available kit (Endogen, Boston, MA) as per manufacturer specifications. Horseradish peroxidase-conjugated streptavidin (Zymed, San Francisco, CA) and tetramethylbenzydine (DAKO, Carpinteria, CA) were used as developers. The absorbance was read on a Bio-Rad (Hercules, CA) model 3550 microplate reader at 655 nm or 450 nm after addition of 0.18 M sulfuric acid stop solution with background substraction.

Previous studies regarding the activation of naive CD8⁺ T cells have been hampered by the inability to examine the consequences of Ag-specific stimulation in the absence of APC interactions. We developed a system (34) in which we can induce activation of naive CD8⁺ T cells in response to purified MHC and peptide in the absence of APC. The 2C Tg naive CD8⁺ T cells respond specifically to purified H-2L^d and peptide by proliferating, secreting IL-2, and becoming lytic effectors in the absence of APCs. High cell densities (1–3 × 10⁵ cells/well) were required to detect a response. It was conceivable that the need for high cell density was due to T-T cell interactions. Although the 2C Tg T cells do not express the H-2L^d molecule, T-T cell costimulation in trans might have occurred. Therefore, it was important to exclude the possibility of these interactions and the potential presence of a rare contaminant syngeneic APC in the cultures. We suspected that the relatively low response observed in our APC-free system might have been due to low level recruitment of responder cells by the immobilized MHC-peptide complexes. This, in turn, might have been due to incomplete peptide and/or β₂m loading of the recombinant MHC molecules under the conditions used. On the basis of a recent study regarding the role of the pH and β₂m excess in facilitating MHC class I binding sites (38), we modified our strategy to increase the unloading of endogenous peptide and improve exogenous peptide loading. As shown in Fig. 1,A, citrate buffer treatment (pH 6.5) in the presence of exogenous peptide and hβ₂m produced a strong proliferative response even at 3 × 10⁴ cells/well when the sorted CD8⁺CD44⁻ T cells were stimulated with immobilized H-2L^d-p2C-QY5 complexes. At high cell densities (3 × 10⁵/well), naive CD8⁺ T cells proliferate with similar intensities independently of the H-2L^d pH treatment. Only the pH-treated H-2L^d molecules were able to support proliferation at lower cell densities. In addition, the new loading protocol decreased the peptide and MHC concentrations required for T cell activation (data not shown). We next compared the effect of the modified peptide loading protocol on the different p2C peptides in a proliferation assay (Fig. 1,B). Although at lower cell densities naive CD8⁺ T cells had a higher proliferative response to the more potent peptide (p2C-QY5), the weaker peptides, p2Ca and p2C-Y4, elicited strong proliferative responses when the H-2L^d was pH treated. We further evaluated the responses of serially diluted sorted CD8⁺ T cells after stimulation with either APCs or with plate-bound MHC class I + p2C-QY5. As seen in Fig. 1 C, the proliferative responses induced by either stimulus were similar. Interestingly, immobilized purified MHC class I-p2C-QY5 complexes were able to elicit a proliferative response even at the lower cell densities. In conclusion, the modified loading protocol, which presumably increases the number of peptide-MHC complexes available for TCR-CD3 engagement, allows for a response of the 2C naive CD8⁺ T cells at low cell densities. Since a proliferative response is detected at lower cell densities, the likelihood of costimulatory interactions between T cells or with a rare contaminant APC is diminished.

Even though the efficient peptide loading of the immobilized MHC molecules allowed for the activation of 2C CD8⁺ T cells at low cell densities, determining the percentage of the cell population that was activated is critical to the understanding of our system. Early activation events include the modulation of certain surface molecules. Among these molecules, the CD69 molecule (VEA) is up-regulated at the cell surface (39, 40), and it was recently shown (41) that down-regulation of the expression of the CD3 surface molecule follows CD8⁺ T cell TCR engagement and cell activation. Our previous studies had shown that at 48 h poststimulation, >95% of the naive 2C Tg CD8⁺ T cells up-regulated surface expression of the CD69 and the CD25 (IL-2Rα) molecules after activation with all the p2C peptides tested (34). We determined surface expression of CD69 and of CD3 at 4 h poststimulation to assess the percentage of cells that were receiving the activation signal and to rule out that a subpopulation of cells was being selected. Our results demonstrate (Fig. 2,A) that 40% of the gated CD8⁺ T cells up-regulate surface expression of the CD69 Ag and that >95% of the cells down-regulate the expression of the surface CD3 molecule. The reason that only 40% of the CD8⁺ T cells express CD69 is probably the early time point evaluated. The examination of the plates under the optical microscope (×100) verified the FACS results (Fig. 2 B). The majority of the cells showed an increase in size. These data corroborate that >95% of the cells recognized the Ag and had engaged their TCRs. Our results support the notion that a large percentage of the cells underwent an activation event and confirm that activation of naive CD8⁺ T cells with a potent stimulus, such as p2C-QY5, is not limited to a rare CD8⁺ T cell phenotype.

Although the activation of naive CD8⁺ T cells by purified Ag-MHC complexes can occur at low cell densities, we further examined whether costimulatory interactions were necessary by blocking two prominent costimulatory pathways: the CD28/B7.1-B7.2 and LFA-1/ICAM-1 pathways. We observed that when sorted CD8⁺CD44⁻ 2C T cells were stimulated with purified H-2L^d-p2C-QY5 complexes, blockade of the B7.1, B7.2, or LFA-1 molecules with specific mAbs was unable to decrease the proliferative response (Fig. 3). The costimulatory blockade seemed to slightly increase the response of the 2C T cells to purified MHC-peptide complexes. At the low cell density used in these experiments, this enhancement may be due to direct effects of the blocking Abs on the T cells. As expected, addition of anti-LFA-1, anti-B7.1, anti-B7.2, or the combination of anti-B7-1 and anti-B7-2 mAbs significantly decreased the proliferative response of naive CD8⁺ T cells to allogeneic APC stimulation. These results show that these two costimulatory pathways are not an absolute requirement in the activation of 2C Tg naive CD8⁺ T cells when a potent TCR stimulus is used. The inability of these Abs to block the proliferative response to purified MHC class I and peptide suggests that T-T cell interactions and/or potentially contaminating presenting cells do not participate in the response.

It has been shown that ligation of the CD28 costimulatory molecule stabilizes IL-2 mRNA (42) and increases IL-2 secretion (43) and CD8⁺ T cell survival (25). In Tg naive CD4⁺ T cells, IL-2 secretion was shown to be dependent on costimulation (14). To further demonstrate that a potent signal delivered through the TCR can drive T cells to full activation, we determined the levels of secreted IL-2 in the supernatants of 48-h cultures. Naive CD8⁺ T cells stimulated with either purified MHC and peptide or allogeneic APC (Fig. 4,A) produced nanogram amounts of IL-2 even at low cell densities. When cells were stimulated with H-2L^d and p2C-QY5, CD28 engagement was not an absolute requirement for induction of IL-2 production. The presence of IL-2 in the cultures of cells seeded at low cell densities indicates that cell-cell interaction was not a determinant factor in induction of IL-2 secretion. The lower amount of IL-2 generated by APC-stimulated T cells at high T cell densities may be due to a lower T cell:APC ratio with high T cell numbers. Furthermore, blockade of the B7.1, B7.2, or LFA-1 molecules (Fig. 4, B and C) decreased IL-2 secretion only when the cells were stimulated with APC. When cells were stimulated with H-2L^d and p2C-QY5 in the presence of costimulatory molecule blockade, the levels of IL-2 detected were found to be similar to that of control mAb. These results indicate that stimulation of naive CD8⁺ T cells with purified MHC class I and peptide can support IL-2 secretion independently of two major T cell costimulatory pathways.

Recently, studies of naive 2C-Tg T cells activated with allogeneic APCs demonstrated that the role of CD28 signaling is to sustain the late proliferative response and enhance long term cell survival (25, 33). Those studies demonstrated that 2C Tg T cells activated with a high TCR stimulus (addition of exogenous p2Ca to the APC) cannot overcome the need for CD28 costimulation. The APC-free system provides a unique method to discriminate the need for costimulatory molecules during T cell activation. We evaluated the requirements for costimulation during early and late proliferation in a time course study (Fig. 5). Naive CD8⁺ T cells were stimulated with APCs, using a high APC:CD8⁺ ratio (4:1) to ensure a high level of costimulation. The response was compared with that of cells stimulated with immobilized H-2L^d and p2C-QY5. The proliferative responses were assessed at 24, 48, 72, and 96 h. As expected, the proliferative response of 2C CD8⁺ T cells induced with APC was maximal at 72 h. MHC and peptide also induced a strong proliferative response at 48 and 72 h. Interestingly, the response of cells stimulated with MHC class I and p2C-QY5 complexes was not decreased at the 96-h time point. Our data show that naive CD8⁺ T cells stimulated with purified protein and peptide complexes can sustain proliferation up to 96 h and that a potent TCR signal can bypass the need for costimulation.

In our studies, we have used splenic CD8⁺ T cells sorted for the CD44⁻ population to study the costimulatory requirements of naive CD8⁺ T cells (44). In the present study, a strategy for efficient MHC class I-peptide loading (38) allowed us to observe significant proliferative and IL-2 responses at low cell densities (≤3 × 10⁴ T cells/well). The majority of these phenotypically naive T cells increased their size at 48 h, and a highly significant percentage of the T cells down-regulated the CD3 surface molecule at 4 h poststimulation. In addition, a large percentage of the CD8⁺ T cells showed early up-regulation of the surface CD69. This evidence strongly argues against a small subset of the sorted T cells being responsible for the observed costimulatory-independent activation. Furthermore, the responses of naive CD8⁺ 2C T cells to H-2L^d-p2C-QY5 complexes were not blocked by antibodies to the B7 or LFA-1 molecules. This clearly demonstrates the lack of an absolute requirement for B7 or LFA-1 molecules in the activation of naive CD8⁺ T cells.

Although we cannot absolutely exclude costimulatory signals involving molecules other than B7 and LFA-1, several facts suggest that costimulation is not an absolute requirement. First, the cell density at which we observe activation of a high percentage of the T cells is at 3 × 10⁴ cells/well, and at that density few of the added T cells are initially in contact with each other. Second, under physiologic conditions, adhesion/costimulatory molecules need to be present on the same cell surface as the TCR binding complex (45, 46). The 2C T cells used in the APC-free system do not have the antigenic allo-H-2L^d molecule, and we previously have shown that the presentation of peptide antigen on self-H-2^b haplotype molecules does not occur in our system (34). Therefore, effective costimulation from T-T cell interactions is unlikely. Finally, B7 molecules are up-regulated on activated 2C T cells (24), and ICAM-1 is expressed on T cells (47). Therefore, the ligands for the potent T cell costimulatory receptors CD28 and LFA-1 are present. We reasoned that if the TCR signal of the 2C T cells was insufficient for activation without cell-cell interactions, B7 molecules or LFA-1 would play a role. Our studies show that these molecules do not play a role, and consequently, it is highly unlikely that T-T cell costimulatory interactions or rare contaminant presenting cells are driving the T cell responses we observe.

Previous studies have had conflicting results regarding the need for ICAM-1 and B7 costimulation in the 2C system for induction of proliferation at 48 and 72 h (24, 25, 32, 33). Experiments in which transfected Drosophila cells are used as APCs for 2C T cell activation demonstrate a requirement for costimulatory molecules even in the presence of high affinity p2C peptides (32, 33). The same peptides with spleen cells or RMA-S.L^d cells as APCs demonstrate early 2C T cell activation without B7-CD28 or CD8 interactions (24, 25). The difference between these results is most likely due to the differences in the APC. Spleen cells express a number of cell surface molecules, other than B7, that can costimulate, and RMA-S cells have been shown to costimulate (48). Transfected Drosophila cells have only moderate expression of H-2L^d, the cells die at 37°C, and insect cell-expressed H-2L^d may have structural differences from mammalian cell-expressed H-2L^d. Thus, it is unclear whether the requirement for costimulatory molecules by the Drosophila cells is due to a weak TCR signal or due to the complete lack of any alternative mammalian costimulatory molecules. Our system, which is devoid of APC costimulatory molecules, demonstrates that under conditions of high Ag density, these molecules are not an absolute requirement. This suggests that the Drosophila cell results may be due to a less potent TCR signal or other effects by the dying Drosophila cells. Another factor that could increase the TCR signal in our system is the use of the extremely potent p2C-QY5 peptide.

In the activation of 2C T cells with potent p2C peptides, using either spleen cells or RMA-S.L^d-presenting cells, CD8 interactions are unnecessary (24). In our system, anti-CD8 can block 2C T cell activation (data not shown). The combination of high Ag density and affinity with APC costimulatory molecules may be more potent than high density Ag alone. It is certainly possible that addition of costimulatory signals to a very potent TCR signal may lead to CD8 independence. The addition of costimulatory signals to our system may decrease the need for CD8 interactions or increase the response in another way. Notwithstanding this, the response we observe, in the absence of costimulation, is a significant one.

In addition to early proliferation, we observe high levels of IL-2 production in the APC-free system. Previous investigators have observed that IL-2 secretion can be enhanced by the ligation of CD28 and that IL-2 is an important determinant in the outcome of T cell stimulation (43). It was also described that the lack of IL-2 in the cultures induced a state of anergy that could be reversed by the exogenous addition of this cytokine (15). In the APC-free system, the production of IL-2, which occurred in the absence of CD28 ligation, further demonstrates that the CD28/B7 pathway is not an absolute requirement for successful CD8⁺ T cell activation.

Some studies suggested that CD28 ligation may have less importance in initial stimulation of T cell proliferation and greater importance in prolonged T cell proliferation (22, 23, 24, 25). This effect on late proliferation may relate to the prevention of apoptosis by CD28 ligation (25, 26). These studies have noted greater effects of CD28 on T cell proliferation and survival at a later time point (96 h). Despite the lack of APC-derived costimulatory signals in our system, we observe a strong proliferative response of T cells at 96 h. One possible explanation is that a high affinity TCR ligand delivered at high density can generate late survival signals in the absence of costimulation. Alternatively, the sustained late proliferation we observed could be due to the lack of a negative APC-delivered signal. In such a model, costimulatory signals would be required to overcome APC-derived negative signals. These negative signals are absent in the APC-free system. The decreased late 2C T cell proliferation seen with splenic APC and p2C peptides (24, 25) can be interpreted as resulting from a negative APC signal. CD28-mediated costimulation is then required to overcome the inhibitory signal.

A similar system, using purified class II MHC and peptide to stimulate TCR Tg CD4⁺ T cells, was unable to show IL-2 production without presenting cells or addition of exogenous signals (49). Three reasons can be postulated to explain the differences between this system and ours: 1) It is possible that CD8⁺ and CD4⁺ cells differ in their requirements for costimulation (50). 2) Our system may achieve a higher ligand density. 3) The need for costimulation may be a function of the kinetic or equilibrium binding constants of the MHC-peptide complex for the TCR and the number of available MHC-peptide complexes. This is a reasonable hypothesis, since the measured affinity of certain p2C peptide-H-2L^d complexes for the 2C TCR is about 10⁻⁷ M (K_d) (51, 52, 53, 54), compared with an affinity of 5 × 10⁻⁵ M (K_d) for the class II MHC-peptide complex binding for the TCR (55, 56) studied in the purified class II MHC model.

The need for a second signal in T cell activation has implications for the maintenance of self-tolerance. It is widely accepted that the presence of an Ag is not sufficient to initiate an immune response and that a second component delivered by an APC is also necessary. Our results are compatible with the “signal intensity” hypothesis (57, 58), which postulates that the requirement for costimulation is dependent on the strength of the TCR signal rather than being an absolute requirement. A potent TCR signal such as the one used in our model system may be quite rare, and generally two signals are required for T cell activation. Nevertheless, rare high affinity-high density T cell ligands that can overcome the requirement for a second signal may play a role in the pathogenesis of autoimmune diseases and graft rejection. Reagents that quantitate peptide-MHC complexes (59, 60, 61, 62) will allow evaluation of Ag density effects on requirements for costimulation. Previous studies demonstrating the requirement for costimulation with tissue-specific expression of foreign Ags (63, 64) may not have had sufficient Ag density or affinity to bypass costimulation. The expression level of particular peptide-MHC complexes in physiologic and pathologic states may better define the likelihood of costimulatory independent costimulation. Even if T lymphocytes require costimulation under physiologic conditions, the ability to bypass costimulation with high Ag density will be a useful vaccination strategy for individuals treated with costimulatory blockade for autoimmune disease or graft rejection. Finally, the observation that a potent TCR signal can bypass the need for a costimulatory signal has important implications for the mechanisms of T cell signaling.

We thank Drs. Ezio Bonvini, David H. Margulies, Barbara Rellahan, Kathryn E. Stein, and Melanie S. Vacchio for review and discussion of the manuscript. We thank Dr. Dennis Loh for the 2C Tg mice and Mary Belcher for helping maintain the mouse colony. We acknowledge Drs. Nga Nguyen and Bhaskar Chandrasekhar of the Facility for Biotechnology Resources, Center for Biologics Evaluation and Research, for providing services in peptide synthesis.