In vitro stimulation of mouse splenocytes with hemagglutinin (HA) 173–190, a peptide derived from influenza virus hemagglutinin (A/JAP/305/57, H2N2), induces CTLs that are directed to the MHC class Ib molecule, H2-M3. M3 preferably binds peptides bearing an N-terminal formylmethionine. In this study, we show that several related nonformylated peptides can induce anti-HA CTLs in vitro: MLIIW (the minimal epitope), derived from HA186–190 at the C-terminal end of HA173–190; MLIIWG; MLIIWGV; and MLIIWGI, as well as formylated MLIIW. The heptamer peptides correspond to a polymorphism of HA192 in H2 strains of influenza; they have the highest relative affinities for M3 of the nonformylated peptides and higher affinities than some formylated mitochondrial peptides. Depending on the affinity of the peptide, a range of concentrations can be used to induce CTLs. One nanomolar of the high affinity f-MLIIW peptide can induce anti-HA CTLs, whereas 100-fold more of the lower affinity MLIIW peptide is needed. Lines induced with high concentrations (1 μM or greater) of f-MLIIW recognize Ag poorly, and the most efficient CTLs are induced with the lowest concentrations of peptide. Analysis with a panel of anti-TCRVβ Abs shows that different T cells respond to high vs low peptide; the repertoire of cells responding to higher concentrations is more diverse, consistent with the expansion of more, but less efficient, clones. Thus, peptide affinity and concentration should be considered together for generating efficient antipeptide CTLs in vitro.

Cytotoxic T lymphocytes recognize peptides bound in the groove of class I molecules. These peptides are derived from endogenous proteins synthesized within the cell that have been degraded by the proteasome or other proteolytic enzymes, and are loaded onto the class I molecule in the endoplasmic reticulum for display on the cell surface. Upon engagement of TCRs with the appropriate peptide/MHC molecule, CTLs are activated to lyse the cell presenting the Ag. By sampling the intracellular contents, class I molecules are beacons for CTLs to recognize and eliminate cancerous cells or those infected with intracellular pathogens.

In mice, more than 107 cells comprise the peripheral CD8+ T cell pool (1), and the frequency of CTLs specific for a given Ag is low. However, if the Ag is known, then the specific T cells can be expanded in vitro by stimulation with peptide. This approach has been useful for characterizing responses against nonimmunodominant epitopes (2, 3, 4), generating responses against Ags derived from pathogens that are not amenable to cell culture (5), and screening potential epitopes to determine immunogenicity (6, 7), and it holds promise for amplifying large numbers of Ag-specific T cells for immunotherapy against viral infections or cancers in humans (8, 9). In most cases, in vitro responses are generated with T cells obtained from immunized animals or infected humans, but examples of primary in vitro responses exist (10, 11, 12, 13). Because in vitro stimulation with peptide is an artificial way to induce T cell responses, generating highly effective CTLs can be elusive. It is not uncommon to generate T cells that capably lyse cells loaded with exogenous peptide, yet fail to lyse cells expressing the epitope endogenously (10, 12, 14). Alexander-Miller et al. have shown that the avidity of the CTLs for recognizing Ag is determined by the concentration of peptide used to induce the response (15). High concentrations of peptide induce low avidity T cells, whereas low concentrations induce high avidity T cells. Viola et al. have shown that high levels of peptide can down-regulate the levels of TCR on the cell surface, thus leading to decreased T cell sensitivity (16). And prolonged stimulation of TCR-transgenic T cells with peptide in vitro can lead to cell death by apoptosis (17). Nevertheless, in many instances, in vitro stimulation is the most viable option for studying Ag-specific responses.

Recently, we characterized an in vitro CTL response to a hemagglutinin (HA)3 peptide, HA173–190, derived from influenza A/JAP/305/57 (H2N2) and presented by the class Ib molecule H2-M3 (13). Within 1 wk of stimulation of naive C57BL/6 splenocytes with HA173–190 in vitro, anti-HA CTLs are generated that are M3 restricted and directed to a minimal pentamer peptide, MLIIW (HA186–190), present at the C-terminal end of the larger peptide. M3 is well known for its preference for binding peptides derived from bacteria and mitochondria and bearing an N-terminal formylmethionine (see Ref. 18 for review). Formylated peptides bind to M3 with at least 100-fold greater affinity than do nonformylated peptides, and sensitize target cells to lysis in the picomolar range, whereas nonformylated peptides, such as the MLIIW epitope, sensitize target cells in the nanomolar range (13, 19). These findings, along with the identification of M3-restricted responses to formylated peptides derived from Listeria monocytogenes, support the hypothesis that M3 is specialized for clearing infections with bacteria (20, 21, 22, 23). However, M3-restricted responses to nonformylated viral epitopes have not been studied. HA186–190 is perhaps the best candidate to date.

When the anti-HA response was first identified, Milligan et al. (24) reported that anti-HA CTLs were unable to lyse cells infected with influenza. We have suggested that this was due to the inability of the nonformylated HA epitope to compete with the endogenous formylated mitochondrial peptides for binding to M3 (13). However, it is also possible that their CTLs were inefficient at recognizing Ag, having been stimulated with high concentrations of peptide. Now that the minimal epitope for recognition by anti-HA CTLs has been identified, it is important to determine whether peptides shorter than HA173–190 can induce effective CTLs. In this study, we show that CTLs can be induced with shorter HA peptides, and we use the unique qualities of M3 presentation to show the linked effects of peptide affinity and concentration on the generation of anti-HA CTLs in vitro.

HA173–190 (VAKGSYNNTSGEQMLIIW), HA186–190 (MLIIW), HA186–191 (MLIIWG), HA186–192V (MLIIWGV), HA186–192I (MLIIWGI), and the formylated mitochondrial peptides f-ND1-6V (f-MFFINVLTL) and f-COI-3I (f-MFINRWLFS) were synthesized on a Rainin Symphony peptide synthesizer, using standard F-moc chemistry, as described previously (13, 19). Lyophilized peptides were dissolved in DMSO for 1–2 mM stock solutions.

Anti-HA CTLs were generated in vitro, as described previously (13, 24). Briefly, cell suspensions were prepared from the spleens of unimmunized female C57BL/6J (B6) mice (6–12 wk old) obtained from The Jackson Laboratory (Bar Harbor, ME) and housed in our conventional (not specific pathogen-free) facility. For lines generated against 5 μM peptide, 4–8 × 107 cells were incubated with 25 μM of peptide (from stock solutions) in 3–4 ml RP10 (RPMI 1640 supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, and 50 μM 2-ME) for 4 h at 37°C in upright 25-cm2 tissue culture flasks (430168; Corning, Corning, NY), then diluted 5-fold (to 15–20 ml final volume) and incubated for 1 wk. The lines against 1 μM of the various HA peptides were generated similarly using 5 μM peptide for initial 4-h incubation before diluting. For lines generated against lower concentrations of MLIIW and f-MLIIW, 1.75 × 107 spleen cells were γ-irradiated (3000 rad) and preincubated with the indicated concentration of peptide for 1 h in 1 ml of RP10, then washed for use as stimulators; peptide-pulsed stimulators were mixed with an equal number of unirradiated spleen cells and incubated in 12-well tissue culture plates (3 ml/well) for 1 wk at 37°C. All lines were restimulated weekly with syngeneic, irradiated (1500 rad) spleen cells either mixed with peptide (5 μM lines) or pulsed 1 h with appropriate concentration of peptide (all other lines). To generate long-term lines for use in cytotoxicity assays, responding cells were restimulated weekly in supplemented Mishell-Dutton medium plus IL-2, as previously reported (25), 1–2 × 106 responders to 6–8 × 106 stimulators per well in 12-well tissue culture plate.

The anti-f-ND1-6V line, B6 α anti-γ, has been described previously (26).

NZB/Icr-derived Pc11198 (H2d2) cells (27) were maintained in RP10 medium and were used as target cells in standard 51Cr release assays, as previously reported (13). CTLs were diluted in RP10 and added to a 96-well round-bottom microtiter plate (100 μl/well). Pc11198 cells were labeled with 51Cr for 1 h, washed, and resuspended to 105 cells/ml. Peptide was added to the target cells at the concentration indicated, and 100 μl of cells was added per well to the microtiter plate. For the competition assay, 106 target cells were incubated with 1 nM of f-ND1-6V peptide and the indicated concentration of competitor peptide in 1 ml of RP10 for 3 h, then labeled, washed, resuspended, and added to B6 α anti-γ effectors. Spontaneous and maximal release controls were prepared by adding 100 μl of target cells to 100 μl of RP10 or 1% Triton X-100 in water, respectively. Plates were centrifuged at 200 × g for 10 s and incubated for 4 h at 37°C. One hundred microliters of supernatant were harvested from each well, and radioactivity was measured in a gamma counter. Percent lysis represents the mean of duplicate or triplicate samples and was calculated as percent specific lysis = 100 × (experimental release − spontaneous release)/(maximal release) (28). Errors were <5% of maximal release, and spontaneous release varied between 5 and 12% of maximal release.

CTLs were harvested 5 to 6 days after restimulation, washed, and resuspended to 4 × 106 cells/ml in FACS staining buffer (PBS containing 1% BSA and 0.1% sodium azide). Fifty microliters of cells (2 × 105) were added per well to a 96-well microtiter plate and doubly stained with PE-labeled anti-CD8 and 1 of 13 different FITC-labeled anti-mouse TCRVβ Abs (2, 4, 5.1/2, 6, 7, 8.1/2, 8.3, 9, 10, 11, 12, 13, and 14 (all obtained from PharMingen, San Diego, CA)) for 20 min at 4°C. Free Ab was washed away, and cells were resuspended in cold staining buffer and analyzed on a FACScan flow cytometer (Becton Dickinson, Mountain View, CA). Vβ expression of CD8+ lymphocytes was determined with CellQuest software (Becton Dickinson) by counting the percentage of cells that stained higher than background (no anti-TCRVβ).

The peptide-binding cleft of H2-M3 is shorter than those of other class I molecules; the A pocket is blocked and the N-terminal formylmethionine side chain is buried in the B pocket instead (29). Therefore, peptides that are presented by M3 are typically five to seven amino acids long (18, 30). Previously, we showed that CTLs induced with HA173–190 recognized a minimal epitope, MLIIW, presented by M3 (13). This nonformylated epitope has at least 100-fold less affinity for M3 than the formylated form of the peptide, as judged by their ability to sensitize target cells for lysis. Because M3 can bind peptides longer than five amino acids, we sought to determine whether hexamer and heptamer peptides beginning at HA186 could bind M3 any better. A database search identified two peptides, MLIIWGV and MLIIWGI, derived from residues 186–192 of influenza A type 2 hemagglutinin that contain the minimal MLIIW epitope. These peptides differ at position 192, with a valine for A/JAP HA and an isoleucine for some avian strains of influenza.

Nonformylated MLIIW, MLIIWG, MLIIWGV, and MLIIWGI peptides were synthesized and used to generate CTL lines from the spleen cells of unimmunized B6 mice. All peptides induced CTLs that lysed Pc11198 cells loaded with peptide, but not untreated target cells (Fig. 1). (Pc11198 cells are ideal for measuring M3-restricted CTL responses, because they are H2d2 and eliminate the chance for H2-restricted or Qa1-restricted killing by the B6 CTLs (13).

FIGURE 1.

HA186–192 peptides are more potent than MLIIW for sensitizing target cells to lysis by anti-HA CTLs. Pc11198 (H2d2) cells were mixed with peptides at 1 μM or 1 nM and tested for lysis in a 4-h 51Cr release assay. Peptides were as follows: HA186–190, MLIIW; HA186–191, MLIIWG; HA186–192V, MLII WGV; and HA186–192I, MLIIWGI. CTLs were tested after 12 wk of restimulation with 1 μM of the respective peptides, and E:T ratio was 5:1. “None” is recognition of target cells in the absence of peptide, and was zero for most of the CTLs. Recognition was similar in serum-free conditions.

FIGURE 1.

HA186–192 peptides are more potent than MLIIW for sensitizing target cells to lysis by anti-HA CTLs. Pc11198 (H2d2) cells were mixed with peptides at 1 μM or 1 nM and tested for lysis in a 4-h 51Cr release assay. Peptides were as follows: HA186–190, MLIIW; HA186–191, MLIIWG; HA186–192V, MLII WGV; and HA186–192I, MLIIWGI. CTLs were tested after 12 wk of restimulation with 1 μM of the respective peptides, and E:T ratio was 5:1. “None” is recognition of target cells in the absence of peptide, and was zero for most of the CTLs. Recognition was similar in serum-free conditions.

Close modal

To determine the potency of the longer peptides, we tested the ability of CTLs to recognize cells incubated with 1 μM or 1 nM concentrations of each peptide. As shown in Fig. 1, 1 μM sensitized cells to lysis by all CTLs; at 1 nM, MLIIWG was the least potent for sensitizing target cells. Of the other peptides, MLIIW was recognized just above background, and MLIIWGV and MLIIWGI were the most potent. Recognition of the longer peptides was the same in serum-free conditions and was not due to proteolytic trimming by dipeptidases in the medium (31). Some CTLs were more specific for the peptide used for induction, e.g., anti-MLIIWG, and similar results have been obtained with other lines stimulated with MLIIWGI. In general, peptide-specific recognition is observed after prolonged restimulation; the anti-MLIIWG preference in Fig. 1 was seen after 12 wk of restimulation, whereas the specificity of the line was not apparent at 4 wk (data not shown).

To better determine the relative affinities of the nonformylated epitopes, we tested the ability of the most potent HA peptides to compete for binding to M3 in an inhibition assay using a CTL line, B6 α anti-γ, specific for the f-ND1-6V (f-MFFINVLTL) peptide presented by M3 (26). Target cells were incubated with 1 nM of f-ND1-6V peptide and increasing concentrations of MLIIW, MLIIWGV, or MLIIWGI or another high affinity mitochondrial peptide, f-COI-3I (32), and then tested for recognition by B6 α anti-γ CTLs. We chose 1 nM f-ND1-6V, because this concentration submaximally sensitizes Pc11198 cells to lysis and, therefore, is optimal for measuring inhibition of binding of other peptides to M3. As shown in Fig. 2, the relative affinities of the peptides correlated with the relative potency observed in Fig. 1. MLIIWGI had the highest relative affinity of the three HA peptides tested. Still, none of the peptides inhibited recognition to the level of f-COI-3I, because they lack the critical N-formyl group. Indeed, similar but separate experiments have shown that formylated MLIIW binds to M3 with an affinity equal to f-ND1-6A and slightly better than f-COI-3I (data not shown). Together, Figs. 1 and 2 indicate that the nonformylated 7-mer peptides bind to M3 better than the minimal epitope, MLIIW.

FIGURE 2.

Heptamer HA peptides have a higher relative affinity for M3. Pc11198 cells were incubated with 1 nM of f-ND1-6V peptide (f-MFFINVLTL) and the indicated concentrations of f-COI-3I (f-MFINRWLFS, ⋄), MLIIWGI (○), MLIIWGV (▵), or MLIIW (□) for 3 h, then tested for recognition by B6 α anti-γ CTLs, which are specific for f-ND1-6V/M3, in a 4-h 51Cr release assay. Values represent the percentage of inhibition of lysis by the CTL line in the presence of competitor peptide and are the means of two to four experiments for each peptide tested; bars show the SEs from the mean. Specific lysis of target cells + f-ND1-6V peptide (in the absence of competitor peptide) was 50–70% of maximal release in all experiments at an E:T ratio of 4:1 to 8:1. Formylated COI-3I and f-ND1-6V peptides represent the N termini of cytochrome oxidase I (COI) and NADH-dehydrogenase 1 (ND1), respectively, and function as minor histocompatibility Ags that bind to M3 with high affinity (1832 ).

FIGURE 2.

Heptamer HA peptides have a higher relative affinity for M3. Pc11198 cells were incubated with 1 nM of f-ND1-6V peptide (f-MFFINVLTL) and the indicated concentrations of f-COI-3I (f-MFINRWLFS, ⋄), MLIIWGI (○), MLIIWGV (▵), or MLIIW (□) for 3 h, then tested for recognition by B6 α anti-γ CTLs, which are specific for f-ND1-6V/M3, in a 4-h 51Cr release assay. Values represent the percentage of inhibition of lysis by the CTL line in the presence of competitor peptide and are the means of two to four experiments for each peptide tested; bars show the SEs from the mean. Specific lysis of target cells + f-ND1-6V peptide (in the absence of competitor peptide) was 50–70% of maximal release in all experiments at an E:T ratio of 4:1 to 8:1. Formylated COI-3I and f-ND1-6V peptides represent the N termini of cytochrome oxidase I (COI) and NADH-dehydrogenase 1 (ND1), respectively, and function as minor histocompatibility Ags that bind to M3 with high affinity (1832 ).

Close modal

CTLs generated against HA173–190 as well as MLIIW, MLIIWG, or MLIIWGV/I consistently recognize target cells sensitized with either nonformylated or formylated MLIIW. However, CTLs generated against 5 μM of the high affinity f-MLIIW peptide (5 μM is the concentration originally used for generating anti-HA CTLs (24)) lysed target cells incubated with 1 μM f-MLIIW, but not MLIIW (Table I, Expt. 1). Furthermore, these anti-f-MLIIW lines were more difficult to generate than the anti-HA173–190 lines and often died within a few weeks of induction. Recently, a number of reports have shown that the concentration of peptide used for generating CTLs in vitro can profoundly affect the sensitivity of the line generated (4, 12, 15). Our results in Table I show the same is true for the anti-HA response. In experiment 2, six lines from the same pool of spleen cells were stimulated with 1, 0.1, or 0.01 μM of MLIIW and f-MLIIW peptides and tested for recognition of target cells loaded with 1 μM of the respective peptides. Again, although background lysis was high, the line induced with the high concentration of f-MLIIW failed to distinguish cells incubated with the nonformylated epitope from untreated cells, but did recognize those incubated with f-MLIIW. By contrast, lines induced with 0.1 and 0.01 μM f-MLIIW recognized both peptides equally; by decreasing the concentration of f-MLIIW, CTLs can be induced that recognize both f-MLIIW and MLIIW peptides. With the lower affinity MLIIW peptide, 1 and 0.1 μM induced lines that recognized both peptides, yet 0.01 μM failed to induce any CTL correlating with the lower affinity of the nonformylated peptide.

Table I.

Lysis by CTLs induced with different concentrations of MLIIW and f-MLIIW peptides

CTLsE:T% Lysis Pc11198 + 1 μM Peptide
NoneMLIIWf-MLIIW
Expt. 1 (5 μM)a     
HA173-190 0.4 :1 35 34 
f-MLIIW 18 :1 −1 −1 44 
Expt. 2b     
MLIIW     
1 μM 7 :1 24 17 
0.1 μM 6 :1 51 36 
0.01 μM 12 :1 
f-MLIIW     
1 μM 6 :1 23 30 50 
0.1 μM 8 :1 58 52 
0.01 μM 5 :1 −1 19 13 
CTLsE:T% Lysis Pc11198 + 1 μM Peptide
NoneMLIIWf-MLIIW
Expt. 1 (5 μM)a     
HA173-190 0.4 :1 35 34 
f-MLIIW 18 :1 −1 −1 44 
Expt. 2b     
MLIIW     
1 μM 7 :1 24 17 
0.1 μM 6 :1 51 36 
0.01 μM 12 :1 
f-MLIIW     
1 μM 6 :1 23 30 50 
0.1 μM 8 :1 58 52 
0.01 μM 5 :1 −1 19 13 
a

Expt. 1 CTLs were induced with 5 μM peptide and derived from different mice and tested 5 wk (HA173-190) and 4 months (f-MLIIW) after induction.

b

Expt. 2 CTLs were induced with stimulator cells pulsed with peptide at the concentrations indicated (see Materials and Methods). CTLs were derived from the same pool of spleen cells from two mice and were tested 3 wk after induction.

The results in Table I show lysis of target cells incubated with 1 μM peptide, but f-MLIIW is at least 100-fold more potent than the low affinity MLIIW peptide for binding to M3 (13). Therefore, to determine the sensitivity of the CTLs induced with the high and low affinity peptides, target cells were incubated with various concentrations of f-MLIIW and MLIIW, and lysis was measured. Fig. 3 shows that the sensitivity of the CTLs was inversely correlated with the concentration of peptide used to induce the lines. For example, the 0.01 μM f-MLIIW line lysed target cells incubated with 1 pM f-MLIIW, whereas the 1 μM f-MLIIW line failed to lyse target cells incubated with 10,000-fold more f-MLIIW. The 0.1 μM f-MLIIW line was an intermediate between the two. Similarly, CTLs induced against 0.1 μM MLIIW were more sensitive than those induced with 10-fold more of the same peptide. Titration of the two peptides thus shows that the CTLs induced from the same spleen cell pool were vastly different in their ability to recognize Ag.

FIGURE 3.

CTLs induced with higher concentrations of peptide are less efficient at recognizing Ag. CTLs from Table I were tested 3 mo after induction for their abilities to recognize Pc11198 cells incubated with various concentrations of MLIIW and f-MLIIW in a 4-h 51Cr release assay. Anti-MLIIW lines are represented as circles, anti-f-MLIIW lines as squares, and concentrations used for stimulation as open (1 μM), crossed (0.1 μM), and closed (0.01 μM) symbols. E:T ratio was 7:1. The anti-1 μM f-MLIIW line had a high background lysis of target cells in the absence of peptide and lysed target cells incubated with 10 nM f-MLIIW the same, but lysis of target cells incubated with 1 μM f-MLIIW was at least 2-fold higher than background (Table I, and data not shown).

FIGURE 3.

CTLs induced with higher concentrations of peptide are less efficient at recognizing Ag. CTLs from Table I were tested 3 mo after induction for their abilities to recognize Pc11198 cells incubated with various concentrations of MLIIW and f-MLIIW in a 4-h 51Cr release assay. Anti-MLIIW lines are represented as circles, anti-f-MLIIW lines as squares, and concentrations used for stimulation as open (1 μM), crossed (0.1 μM), and closed (0.01 μM) symbols. E:T ratio was 7:1. The anti-1 μM f-MLIIW line had a high background lysis of target cells in the absence of peptide and lysed target cells incubated with 10 nM f-MLIIW the same, but lysis of target cells incubated with 1 μM f-MLIIW was at least 2-fold higher than background (Table I, and data not shown).

Close modal

None of the CTLs appeared to be specific for the peptide used for induction, as lines induced with f-MLIIW and MLIIW recognized both peptides similarly, except that 100-fold more nonformylated peptide was needed to sensitize lysis to an equivalent level. The inability of the 1 μM line to recognize the low affinity MLIIW peptide is most likely due to its poor ability to recognize Ag, although an f-MLIIW-specific component to recognition may exist. At least 1 μM of the high affinity f-MLIIW peptide was needed to sensitize target cells to lysis (Table I), so 100 μM MLIIW would be needed to sensitize at an equivalent level. However, incubating target cells with 100 μM MLIIW would require them to be exposed to 10% DMSO, which causes too high 51Cr release backgrounds. Whether or not the 1 μM f-MLIIW line is specific for the formylated peptide, its sensitivity to Ag is extremely poor. Together, the results in Table I and Fig. 3 demonstrate the importance of using the lowest concentration of peptide still capable of inducing CTLs to generate efficient lines in vitro.

When T cell clones are strongly stimulated via their receptors by high concentrations of peptide on APC, by anti-CD3 treatment, or by superantigen treatment, TCRs are internalized, rendering the T cells less responsive to Ag (16). These results suggest that our anti-HA CTLs induced with high concentrations of peptide could be less responsive to Ag due to decreased numbers of surface TCRs. Alternatively, poor recognition of Ag could be the result of differences in the affinity of TCRs for binding Ag.

To determine the reason for poor recognition by the CTLs induced with high peptide, we analyzed receptor expression by flow cytometry, using a panel of Abs to mouse TCRVβ-chains. Recently, a similar panel was used by Busch et al. to characterize the T cell repertoire in mice that respond to different L. monocytogenes epitopes (4); this panel covers >90% of the naive CD8+ T cell repertoire in B6 mice (unpublished data and (4)). Table II shows the Vβ expression of the lines, from Table I and Fig. 3, stimulated with the MLIIW and f-MLIIW peptides. The 1 μM f-MLIIW line expressed Vβ14 (>99%), while the others largely expressed Vβ4 (73% for 0.1 μM f-MLIIW, 96% for 0.01 μM f-MLIIW, 83% for 1 μM MLIIW, and >99% for 0.1 μM MLIIW). Because each line was stimulated from the same pool of spleen cells, these results show that differences in the concentration of peptide used for stimulation affect which T cells are amplified in vitro. The more efficient lines were Vβ4+, whereas the inefficient 1 μM f-MLIIW line was Vβ14+, yet each line began with the same distribution of Vβ4+ and Vβ14+ precursors.

Table II.

Vβ usage in anti-HA lines from Table I and Fig. 3 

CTLsa% CD8+ Cells Positive for TCRVβ
245.5678.58.391011121314
f-MLIIW              
1 μM 99.7 
0.1 μM 72.9 12.6 10.6 0.9 0.5 
0.01 μM 0.5 95.7 2.2 0.5 1.1 
MLIIW              
1 μM 1.0 82.5 0.9 0.5 0.2 0.3 1.4 0.4 2.8 10.4 
0.1 μM 99.7 
CTLsa% CD8+ Cells Positive for TCRVβ
245.5678.58.391011121314
f-MLIIW              
1 μM 99.7 
0.1 μM 72.9 12.6 10.6 0.9 0.5 
0.01 μM 0.5 95.7 2.2 0.5 1.1 
MLIIW              
1 μM 1.0 82.5 0.9 0.5 0.2 0.3 1.4 0.4 2.8 10.4 
0.1 μM 99.7 
a

MLIIW lines were stained (see Materials and Methods) at 10 wk and f-MLIIW lines at 12 wk after induction. Percentages >10% are boldfaced.

To get a better picture of the T cell response to high and low concentrations of peptide, we stimulated six lines from three individual B6 mice using 5 and 0.1 μM of the MLIIW peptide and examined Vβ expression after 5 wk (Fig. 4). All lines were directed to the HA peptide (data not shown), yet the T cell repertoire was diverse. For each mouse, the TCR usage differed in the CTLs induced with 5 μM vs 0.1 μM. Moreover, Vβ usage in the lines stimulated with the high concentration of peptide was more diverse than in the lines stimulated with the low concentration (although we cannot be certain for the 0.1 μM line from mouse B, because “other” includes all Vβ-chains not recognized by the Abs). Similar results have been observed in lines induced with other HA peptides as well (Table II and unpublished results).

FIGURE 4.

TCRVβ usage in lines induced with 5 and 0.1 μM MLIIW. Two lines were induced from the spleen cells of each of three mice by stimulating with unformylated MLIIW at 5 and 0.1 μM. Lines were restimulated weekly, for 4 wk, and harvested 5 days after restimulation for Vβ analysis, as indicated in Materials andMethods. “Other” refers to the percentage of CD8+ T cells that did not stain positive with any of the anti-Vβ Abs used, and does not necessarily represent a single Vβ-chain.

FIGURE 4.

TCRVβ usage in lines induced with 5 and 0.1 μM MLIIW. Two lines were induced from the spleen cells of each of three mice by stimulating with unformylated MLIIW at 5 and 0.1 μM. Lines were restimulated weekly, for 4 wk, and harvested 5 days after restimulation for Vβ analysis, as indicated in Materials andMethods. “Other” refers to the percentage of CD8+ T cells that did not stain positive with any of the anti-Vβ Abs used, and does not necessarily represent a single Vβ-chain.

Close modal

TCR expression level, as measured by anti-CD3ε staining by flow cytometry, was not strongly correlated with the responsiveness of the line (data not shown). Others have made similar observations (15). This is probably because the lines use different TCRs with different affinities for peptide/MHC.

The M3-restricted anti-HA response was first described based on the ability of an 18-mer peptide, HA173–190, to induce a CTL response in vitro. In this study, we have shown that the minimal epitope, MLIIW, as well as MLIIWG and MLIIWGV/I are capable of inducing CTLs (Fig. 1). CTLs induced with these peptides are not necessarily specific for the very peptide used for induction, although, with continued restimulation of the lines, some examples of specific recognition have been observed.

The heptamer peptides have the highest affinity for M3, and 10-fold less MLIIWGI is needed to inhibit recognition of f-ND1-6V, compared with MLIIW (Fig. 2). With a similar inhibition assay, we have compared the relative affinities of the N-formylated peptides derived from the 13 proteins encoded in the mitochondria (26). Formylated ND1 and f-COI have the highest affinities, followed by nine peptides of medium to low affinity and two peptides that do not inhibit recognition of f-ND1-6V at all. When the 13 peptides were tested for recognition by M3-alloreactive bulk cultures, the high affinity and some of the medium affinity peptides sensitized target cells to lysis. Our results show that MLIIW, MLIIWGV, and MLIIWGI can inhibit recognition of f-ND1-6V better than some formylated mitochondrial peptides. Moreover, with ∼10-fold less relative affinity than the f-COI-3I peptide, MLIIWGI could certainly be classified as a medium affinity peptide. These results suggest that the longer peptides could be naturally presented by M3 on cells infected with influenza virus.

CTLs can be highly sensitive to Ag. For some clones, fewer than 10 peptide/MHC complexes are needed to activate lysis of target cells (33); other CTLs require 103-fold more complexes for activation (34). TCR number and affinity for peptide/MHC as well as coreceptor interactions all contribute to the sensitivity of a T cell to Ag. In this study, we took advantage of the unique qualities of peptide presentation by M3 to show the linked effects of peptide concentration and affinity for inducing sensitive CTLs in vitro. Alexander-Miller et al. first showed that the concentration of peptide used to induce T cells in vitro affects the sensitivity of the resulting line (15), and this is true for M3-restricted anti-HA CTLs as well.

CTLs induced with 0.1 μM MLIIW are more sensitive than those generated with 1 μM. Even more dramatic, 0.01 μM f-MLIIW induces CTLs that require at least 105-fold less peptide for recognition than those generated with 1 μM f-MLIIW. Sensitivity is dose dependent: the less peptide used for induction, the more sensitive the resulting CTL line. However, the optimal concentration depends on the affinity of the peptide for MHC. By adding a formyl group to the MLIIW epitope, a peptide is created that is 100-fold more potent for sensitizing target cells to lysis, yet confers no discernible specificity for the formyl group by the resulting CTL line (Fig. 3). This difference in affinity correlates with the ability of peptide to induce CTLs in vitro, as summarized in Table III. One hundred nanomolar of the MLIIW peptide is the minimal concentration capable of inducing CTL, whereas 100-fold less f-MLIIW is needed. Consistent with this result, 10-fold less medium affinity MLIIWGI peptide can induce CTLs as well. Our results show that efficient CTLs, i.e., those capable of recognizing the nonformylated peptide, are generated by a narrow range of concentrations: too low, and the CTLs do not expand; too high, and the CTLs are inefficient.

Table III.

Summary: concentrations of HA peptides needed to induce anti-HA CTLs capable of recognizing nonformylated MLIIW

PeptideaLog10 M Peptide Used for Induction
−6−7−8−9−10
f-MLIIW b − 
MLIIWGI +/−c NT NT 
MLIIWGV − NT NT 
MLIIW − NT NT 
MLIIWG +/− NT NT NT 
PeptideaLog10 M Peptide Used for Induction
−6−7−8−9−10
f-MLIIW b − 
MLIIWGI +/−c NT NT 
MLIIWGV − NT NT 
MLIIW − NT NT 
MLIIWG +/− NT NT NT 
a

Peptides are listed in order of decreasing affinity for M3. Each concentration was tested at least twice for each peptide.

b

Results are scored as “+”, positive for recognition of 1 μM nonformylated MLIIW, “−”, negative for recognition, and “NT”, not tested, and are based on the lines in this report and others not shown. CTLs generated with high concentrations of f-MLIIW were specific for the f-MLIIW peptide, as indicated in Table I.

c

Stimulation with 10−8 MLIIWGI and 10−7 MLIIWG was not always successful.

So, what is the optimal concentration of peptide to use for inducing CTLs in vitro? For high affinity peptides that sensitize lysis in the picomolar range, data suggest 0.1–10 nM. The best anti-HA CTLs were generated with 1–10 nM f-MLIIW (Fig. 3); high avidity CTL lines against the I10 epitope derived from HIV-1 were generated with 0.1 nM (15); and two immunodominant epitopes from L. monocytogenes, LLO91–99 and p60217–225, presented by Kd, require 1 and 10 nM concentrations, respectively (4). By contrast, the lower affinity nonformylated HA epitopes require ∼100 nM (Table III), similar to two subdominant epitopes derived from L. monocytogenes (4).

Each T cell expresses a single TCRVβ-chain, so Vβ expression is a marker for detecting differences among the T cells that are amplified following peptide stimulation. All the lines in Table II arose from the same pool of precursor T cells. However, the line induced with 1 μM f-MLIIW expressed Vβ14, whereas the other four lines mainly expressed Vβ4. When stimulated with high concentrations of peptide, high avidity T cells are overstimulated and die by apoptosis (35). If Vβ4 were a marker for high avidity anti-HA T cells in these lines, it is probable that 1 μM f-MLIIW overstimulated the Vβ4 population and these cells died, allowing the less efficient Vβ14+ T cells to proliferate. Fig. 4 supports the point that stimulation with higher concentrations of Ag induces lower avidity T cells. In all three mice tested, the response to 5 μM peptide was more polyclonal than the response to 0.1 μM peptide. The cells responding to the low concentration of peptide were primarily composed of a single Vβ (with the possible exception of mouse B) and, in two of three mice (A and B), that Vβ was not observed in the line stimulated with the higher concentration. Less peptide induces a narrower repertoire of cells, probably because fewer T cells can respond to the lower concentrations of peptide.

In summary, we propose the following model. A number of different anti-HA/M3 receptors exist in the precursor T cell pool, and each T cell responds to peptide within an ideal concentration range. High concentrations of peptide overstimulate the high avidity T cells, which die by activation-induced cell death or apoptosis, allowing the more diverse low avidity T cells to proliferate. Stimulating with lower concentrations of peptide induces only the rarer, high avidity T cells to proliferate. The high avidity cells will be more adept at recognizing naturally presented Ag. As for anti-HA CTLs, these higher avidity cells will be useful for determining the ability of M3 to naturally present the nonformylated HA epitope on flu-infected cells.

We thank Annie Eubanks and Margaretan Stolfo for care of the mice, Angie Mobley for help with flow cytometry, and Jim Forman for helpful comments and discussion.

1

This work was supported by National Institutes of Health Grant AI37818 and the Medical Scientist Training Program (D.E.B.).

3

Abbreviation used in this paper: HA, hemagglutinin.

1
Tanchot, C., B. Rocha.
1995
. The peripheral T cell repertoire: independent homeostatic regulation of virgin and activated CD8+ T cell pools.
Eur. J. Immunol.
25
:
2127
2
Van der Most, R. G., K. Murali-Krishna, J. L. Whitton, C. Oseroff, J. Alexander, S. Southwood, J. Sidney, R. W. Chesnut, A. Sette, R. Ahmed.
1998
. Identification of Db- and Kb-restricted subdominant cytotoxic T-cell responses in lymphocytic choriomeningitis virus-infected mice.
Virology
240
:
158
3
Tsai, V., S. Southwood, J. Sidney, K. Sakaguchi, Y. Kawakami, E. Appella, A. Sette, E. Celis.
1997
. Identification of subdominant CTL epitopes of the GP100 melanoma-associated tumor antigen by primary in vitro immunization with peptide-pulsed dendritic cells.
J. Immunol.
158
:
1796
4
Busch, D. H., E. G. Pamer.
1998
. MHC class I/peptide stability: implications for immunodominance, in vitro proliferation, and diversity of responding CTL.
J. Immunol.
160
:
4441
5
Nair, S., J. S. Babu, R. G. Dunham, P. Kanda, R. L. Burke, B. T. Rouse.
1993
. Induction of primary, antiviral cytotoxic and proliferative responses with antigens administered via dendritic cells.
J. Virol.
67
:
4062
6
Stauss, H. J., H. Davies, E. Sadovnikova, B. Chain, N. Horowitz, C. Sinclair.
1992
. Induction of cytotoxic T lymphocytes with peptides in vitro: identification of candidate T-cell epitopes in human papilloma virus.
Proc. Natl. Acad. Sci. USA
89
:
7871
7
Ikeda-Moore, Y., H. Tomiyama, K. Miwa, S. Oka, A. Iwamoto, Y. Kaneko, M. Takiguchi.
1997
. Identification and characterization of multiple HLA-A24-restricted HIV-1 CTL epitopes: strong epitopes are derived from V regions of HIV-1.
J. Immunol.
159
:
6242
8
Rooney, C. M., C. A. Smith, C. Y. Ng, S. K. Loftin, J. W. Sixbey, Y. Gan, D. K. Srivastava, L. C. Bowman, R. A. Krance, M. K. Brenner, H. E. Heslop.
1998
. Infusion of cytotoxic T cells for the prevention and treatment of Epstein-Barr virus-induced lymphoma in allogeneic transplant recipients.
Blood
92
:
1549
9
Riddell, S. R., K. S. Watanabe, J. M. Goodrich, C. R. Li, M. E. Agha, P. D. Greenberg.
1992
. Restoration of viral immunity in immunodeficient humans by the adoptive transfer of T cell clones.
Science
257
:
238
10
Carbone, F. R., M. W. Moore, J. M. Sheil, M. J. Bevan.
1988
. Induction of cytotoxic T lymphocytes by primary in vitro stimulation with peptides.
J. Exp. Med.
167
:
1767
11
De Bruijn, M. L., T. N. Schumacher, J. D. Nieland, H. L. Ploegh, W. M. Kast, C. J. Melief.
1991
. Peptide loading of empty major histocompatibility complex molecules on RMA-S cells allows the induction of primary cytotoxic T lymphocyte responses.
Eur. J. Immunol.
21
:
2963
12
Wentworth, P. A., E. Celis, C. Crimi, S. Stitely, L. Hale, V. Tsai, H. M. Serra, M. F. Del Guercio, B. Livingston, D. Alazard.
1995
. In vitro induction of primary, antigen-specific CTL from human peripheral blood mononuclear cells stimulated with synthetic peptides.
Mol. Immunol.
32
:
603
13
Byers, D. E., K. Fischer Lindahl.
1998
. H2–M3 presents a nonformylated viral epitope to CTLs generated in vitro.
J. Immunol.
161
:
90
14
Staerz, U. D., F. Zepp, R. Schmid, M. Hill, J. Rothbard.
1989
. Recruitment of alloreactive cytotoxic T lymphocytes by an antigenic peptide.
Eur. J. Immunol.
19
:
2191
15
Alexander-Miller, M. A., G. R. Leggatt, J. A. Berzofsky.
1996
. Selective expansion of high- or low-avidity cytotoxic T lymphocytes and efficacy for adoptive immunotherapy.
Proc. Natl. Acad. Sci. USA
93
:
4102
16
Viola, A., A. Lanzavecchia.
1996
. T cell activation determined by T cell receptor number and tunable thresholds.
Science
273
:
104
17
Iezzi, G., K. Karjalainen, A. Lanzavecchia.
1998
. The duration of antigenic stimulation determines the fate of naive and effector T cells.
Immunity
8
:
89
18
Fischer Lindahl, K., D. E. Byers, V. M. Dabhi, R. Hovik, E. P. Jones, G. P. Smith, C. R. Wang, H. Xiao, M. Yoshino.
1997
. H2–M3, a full-service class Ib histocompatibility antigen.
Annu. Rev. Immunol.
15
:
851
19
Smith, G. P., V. M. Dabhi, E. G. Pamer, K. Fischer Lindahl.
1994
. Peptide presentation by the MHC class Ib molecule, H2–M3.
Int. Immunol.
6
:
1917
20
Gulden, P. H., P. Fischer, III, N. E. Sherman, W. Wang, V. H. Engelhard, J. Shabanowitz, D. F. Hunt, E. G. Pamer.
1996
. A Listeriamonocytogenes pentapeptide is presented to cytolytic T lymphocytes by the H2–M3 MHC class Ib molecule.
Immunity
5
:
73
21
Lenz, L. L., B. Dere, M. J. Bevan.
1996
. Identification of an H2–M3-restricted Listeria epitope: implications for antigen presentation by M3.
Immunity
5
:
63
22
Princiotta, M. F., L. L. Lenz, M. J. Bevan, U. D. Staerz.
1998
. H2–M3 restricted presentation of a Listeria-derived leader peptide.
J. Exp. Med.
187
:
1711
23
Pamer, E. G., M. J. Bevan, K. Fischer Lindahl.
1993
. Do nonclassical, class Ib MHC molecules present bacterial antigens to T cells?.
Trends Microbiol.
1
:
35
24
Milligan, G. N., L. Flaherty, V. L. Braciale, T. J. Braciale.
1991
. Nonconventional (TL-encoded) major histocompatibility complex molecules present processed viral antigen to cytotoxic T lymphocytes.
J. Exp. Med.
174
:
133
25
Dabhi, V. M., K. Fischer Lindahl.
1995
. MtDNA-encoded histocompatibility antigens.
Methods Enzymol.
260
:
466
26
Dabhi, V. M., R. Hovik, L. Van Kaer, K. Fischer Lindahl.
1998
. The alloreactive T cell response against the class Ib molecule H2–M3 is specific for high affinity peptides.
J. Immunol.
161
:
5171
27
Morse, H. C., R. Riblet, R. Asofsky, M. Weigert.
1978
. Plasmacytomas of the NZB mouse.
J. Immunol.
121
:
1969
28
Fischer Lindahl, K., B. Hausmann.
1980
. Expression of the I-E target antigen for T-cell killing requires two genes.
Immunogenetics
11
:
571
29
Wang, C.-R., A. R. Castaño, P. A. Peterson, C. Slaughter, K. Fischer Lindahl, J. Deisenhofer.
1995
. Nonclassical binding of formylated peptide in crystal structure of the MHC class Ib molecule H2–M3.
Cell
82
:
655
30
Dabhi, V. M., K. Fischer Lindahl.
1998
. Short peptides sensitize target cells to CTL specific for the MHC class Ib molecule, H2–M3.
Eur. J. Immunol.
28
:
3773
31
Byers, D. E. 1999. The cytotoxic T lymphocyte response to a viral peptide presented by H2–M3. Doctoral dissertation. University of Texas Southwestern Graduate School, Dallas, TX.
32
Morse, M. C., G. Bleau, V. M. Dabhi, F. Hétu, E. A. Drobetsky, K. Fischer Lindahl, C. Perreault.
1996
. The COI mitochondrial gene encodes a minor histocompatibility antigen presented by H2–M3.
J. Immunol.
156
:
3301
33
Sykulev, Y., M. Joo, I. Vturina, T. J. Tsomides, H. N. Eisen.
1996
. Evidence that a single peptide-MHC complex on a target cell can elicit a cytolytic T cell response.
Immunity
4
:
565
34
Kageyama, S., T. J. Tsomides, Y. Sykulev, H. N. Eisen.
1995
. Variations in the number of peptide-MHC class I complexes required to activate cytotoxic T cell responses.
J. Immunol.
154
:
567
35
Alexander-Miller, M. A., G. R. Leggatt, A. Sarin, J. A. Berzofsky.
1996
. Role of antigen, CD8, and cytotoxic T lymphocyte (CTL) avidity in high dose antigen induction of apoptosis of effector CTL.
J. Exp. Med.
184
:
485