HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wilson, D. B. Articles by Houghten, R. A. Search for Related Content PubMed PubMed Citation Articles by Wilson, D. B. Articles by Houghten, R. A.

Immunogenicity. I. Use of Peptide Libraries to Identify Epitopes That Activate Clonotypic CD4⁺ T Cells and Induce T Cell Responses to Native Peptide Ligands¹

Darcy B. Wilson^2,*, Clemencia Pinilla^*, Dianne H. Wilson^*, Kim Schroder^*, César Boggiano^*, Valeria Judkowski^*, Jonathan Kaye, Bernhard Hemmer^3,, Roland Martin and Richard A. Houghten^*

^* Torrey Pines Institute for Molecular Studies, San Diego, CA 92121; Scripps Research Institute, La Jolla, CA 92037; and National Institute of Neurological Diseases and Stroke, National Institutes of Health, Bethesda, MD 20892

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Recent studies have demonstrated the utility of synthetic combinatorial libraries for the rapid identification of peptide ligands that stimulate clonotypic populations of T cells. Here we screen a decapeptide combinatorial library arranged in a positional scanning format with two different clonotypic populations of CD4⁺ T cells to identify peptide epitopes that stimulate proliferative responses by these T cells in vitro. An extensive collection of mimic peptide sequences was synthesized and used to explore the fine specificity of TCR/peptide/MHC interactions. We also demonstrate that many of these deduced ligands are not only effective immunogens in vivo, but are capable of inducing T cell responses to the original native ligands used to generate the clones. These results have significant implications for considerations of T cell specificity and the design of peptide vaccines for infectious disease and cancer using clinically relevant T cell clones of unknown specificity.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Several recent studies have reported the use of combinatorial peptide libraries to identify sequences that stimulate clonotypic populations of autoreactive CD4⁺ (1, 2, 3) and tumor-specific (4) or alloreactive (5, 6) CD8⁺ T cells. For autoreactive CD4⁺ clones, the results of these in vitro studies indicate, in general, that 1) Ag recognition is degenerate, 2) a large number of different ligands exist having a range of different stimulatory potencies, 3) the native autoantigen used to establish the clone is usually suboptimal compared with many of the peptides deduced from library scans, and 4) it is possible to identify peptide ligands from self and infectious pathogens that are significantly more potent than the native ligand (1). What is less clear is whether peptide ligands deduced from library scans are effective immunogens in vivo and whether they induce immune responses to the original peptide ligand.

In the present study, we use a synthetic decapeptide positional scanning library composed of L-amino acids to identify from two clonotypic populations of CD4⁺ T cells an extensive series of peptide epitopes that stimulate proliferative responses by these T cell clones. One of these clonotypic populations consists of T cells derived from transgenic mice expressing TCR ß-chains specific for the 88–104 peptide fragment of pigeon cytochrome c (PCC)⁴ and the E^k class II MHC molecule (7). The second is a human DR2a-restricted T cell clone, TL3A6, reactive to myelin basic protein (MBP) peptide 86–96 (8). We report here an extensive series of peptide epitopes that, despite multiple different L-amino acid substitutions, are able to stimulate these clones in culture more effectively than the native Ags used to generate these clones. In addition, we demonstrate that many of these "superagonist" epitopes, even some synthesized as D-enantiomer peptide or as retro-inverso- (Dri-) peptidomimetic analogues that contain reversed peptide bonds between each residue along the D-amino acid sequence, are potent immunogens that provoke effective T cell-mediated immune responses in vivo. More importantly, these immune responses are directed not only to the immunizing Ag, but also to original native Ag epitopes. Finally, with the PCC system, we show how these agonist peptides can be used to dissect the fine specificity of peptide/MHC/TCR interactions at the clonal level. These results have significant implications for considerations of T cell specificity, the degeneracy of Ag recognition by TCR, and the design of peptide vaccines for potential use in prevention of infectious diseases and immunotherapy of cancer using clinically relevant T cell clones of unknown specificity.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Strategy

Optimal peptide ligand sequences for the two clonotypic T cell populations were defined by a three-step deconvolution strategy involving 1) stimulation of the T cell population with peptide library mixtures in four independent experiments to determine a series of peptide sequences, 2) synthesis of these predicted peptide sequences, and 3) EC₅₀ assessments of individual peptides defined in terms of peptide concentration that stimulates a half-maximal proliferative response by the relevant T cell population.

Library and peptides

A synthetic N-acetylated, C-amide L-amino acid combinatorial decapeptide library arranged in a positional scanning format (PCL 97-4) was prepared at Multiple Peptide Systems (San Diego, CA) as described previously (9). It consists of 200 mixtures in the OX₉ format where O represents one each of the 20 natural L-amino acids in a defined position and X represents all of the natural amino acids, with the exception of cysteine, in each of the remaining positions. For example, the first mixture has alanine (A) in position 1 (A₁X₉), while mixture number 200 has tyrosine (Y) in position 10 (X₉Y₁₀). Each OX₉ mixture consists of 3.2 x 10¹¹ (19⁹) different decamer peptides in approximate equimolar concentration, and the total X₁₀ library consists of 6.4 x 10¹² (20 x 19⁹) different peptides. Assuming an average m.w. of 1200 for a decapeptide mixture and a concentration of 100 µg/ml (83 µM), the concentration of each individual decapeptide is 2.6 x 10^-16 M.

Individual peptides were synthesized by the simultaneous multiple peptide synthesis method (10). Purity and identity of each peptide were characterized using a electrospray mass spectrometer interfaced with a liquid chromatography system.

T cells

Splenic T cells were obtained from (BAND x B10.BR) F₁ mice and depleted of red cells by hypotonic lysis. BAND mice were derived from homozygous AND TCR transgenic mice bred onto an H-2^b background. T cells from AND mice are predominately CD4⁺ and express the TCR ß genes V11 J84 and Vß3 Jß1.2, endowing them with specificity for the 17-mer peptide of PCC_88–104 KAERADLIAYLKQATAK and the E^k class II MHC molecule (7).

TL3A6 clone is a CD4⁺, DRB5*0101 (DR2a)-restricted MBP peptide (86–96)-specific Th1 clone, which was established from an HLA-DRB1*1501/DRB5*0101-positive multiple sclerosis (MS) patient and characterized in detail previously (8). This clone was maintained by weekly stimulation with irradiated (3000 rad) autologous or HLA-DRB5*0101-matched PMBC (1 x 10⁷ PBMC; 1 x 10⁶ cells of the clone), Ag (10 µg/ml MBP peptide), and IL-2 (human rIL-2, 20 U/ml, Tecin by Hoffmann La Roche; kindly provided by the National Cancer Institute, National Institutes of Health, Bethesda, MD).

Culture conditions for stimulation with library mixtures

Whole spleen cells, depleted of erythrocytes, from (BAND x B10.BR) F₁ mice were cultured (200–300 x 10³ cells per well) in flat-bottom microtiter plates in standard T cell medium containing library mixtures at 100 µg/ml (100 µM). Cultures were harvested at 72 h following overnight exposure to 0.5 µC [³H]TdR (6.7 Ci/mM), and incorporated radioactivity was assessed by scintillation counting. The medium used for mouse T cells consists of RPMI 1640 (Fischer Scientific, Pittsburgh, PA) supplemented with 8% FCS (J.R. Scientific, Woodland, CA), HEPES buffer (10 mM; Sigma, St. Louis, MO), 2-ME (50 µM 2-ME; Bio-Rad Laboratories, Richmond, CA), penicillin-streptomycin (5 U/ml and 50 µg/ml, respectively; M.A. Bioproducts, Walkersville, MD), and glutamine (2 mM; Sigma).

For stimulation of the human TL3A6 T cell clone with peptide library mixtures, cells were rested 8–12 days, washed, and resuspended at 1 x 10⁵ cells/ml in complete medium (IMDM containing 5% human serum, 1% penicillin/streptomycin, 0.2% gentamicin; BioWhitaker, Gaithersburg, MD). Then, 100 µl of this cell suspension was added to each well of 96-well U-bottom plates containing 5 x 10⁴ irradiated (3000 rad) PBMC and the various peptide library mixtures (100 µg/ml). Cells were cultured for 72 h at 37°C. During the last 8 h of culture, 1 µCi [³H]thymidine was added to each well. Cells were then harvested, and incorporated radioactivity was determined.

Determination of peptide EC₅₀ values

T cell populations were cultured using conditions described above with varying dilutions of peptides. The peptide concentration causing a half-maximal proliferative response was determined by curve-fitting using a scientific graphics software program (GraphPad Prism; Graph Pad Software, San Diego, CA).

Immunizations and proliferation assays

Young adult B10.BR mice were immunized at the base of the tail and in the inguinal region with 100 µg of selected PCC mimic peptides in CFA. Nylon wool purified T cell suspensions were prepared from draining lymph nodes 15–18 days after immunization and stimulated (300 x 10³ cells per well) in flat-bottom microwell cultures with various dilutions of the immunizing peptide or the native PCC decamer peptide in the presence of irradiated syngeneic spleen cells (700 x 10³ cells per well; 3000 rad) for 3 days. Then, 0.5 µCi [³H]thymidine was added to the wells for 16 h before harvest.

Young adult female LEW rats were immunized in one rear footpad with 25–200 µg of selected MPB peptide mimics in CFA. Draining lymph nodes were recovered 18 days after immunization, and nylon wool-enriched T cell populations were prepared and stimulated (100 x 10³ cells per well) with varying dilutions of the immunizing peptide or the native MBP_87–99 peptide sequence as above, except that irradiated syngeneic thymocytes (10⁶ per flat-bottom well or 200 x 10³ per round-bottom well) were used as a source of APC.

Experimental allergic encephalomyelitis (EAE) clinical scores

Standard scoring methods were used to assess clinical disease on a graded scale based on the following symptoms: 1, flaccid tail; 2, tail and mild hind quarter paresis; 3, severe hind quarter paralysis and incontinence; and 4, tetraparesis, morbidity, and death. Mean duration of disease and mean clinical scores were calculated as the average for each group of animals. The disease index is a composite score calculated as the product of mean duration x mean maximal clinical score x incidence. The first evidence of active EAE generally was evident around 10 days after immunization.

All animal studies received prior approval from the Institutional Animal Care and Use Committee of The Torrey Pines Institute for Molecular Studies.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Library scan on PCC T cells

A decapeptide library was used to scan a PCC-reactive clonotypic T cell population to identify a series of possible candidate sequences that best stimulate proliferative T cell responses. This clonotypic population consists of T cells derived from transgenic mice expressing TCR ß-chains specific for the 88–104 peptide fragment of PCC and the E^k class II MHC molecule (7). Spleen cells from adult transgenic mice were stimulated in duplicate cultures (200–300 x 10³ cells/well) with each of the 200 mixtures of the library at a concentration of 20 µg per well. Control cultures were left unstimulated or were stimulated with the 17-mer (PCC17) peptide 88–104 KAERADLIAYLKQATAK.

Four experiments were conducted to determine which library mixtures in the positional scanning OX₉ format caused the most proliferation; the results of one such experiment are presented in Fig. 1, and they show the following. First, one or more mixtures at positions 2, 4, 5, 7, 8, and 9 caused unambiguous (3 x background) responses in all four experiments; these included A in position 2 (A2), F4, K5, A7, T8, and T9, while mixture-induced responses at the other positions were less clear. Mixtures with I1, Y/A3, P4, I5, S/P7, K9, and F10 were active in two or more of the four separate scans (data not shown). Second, the defined amino acids in the most active mixtures at positions 2, 5, 7, and 8 matched the correct sequence of the 10 C-terminal amino acids of the 17-mer native peptide. This identifies the T cell epitope within the 17-mer sequence for the PCC clone and suggests that a truncated 10-mer peptide may be able to act as an effective stimulator for these cells.

View larger version (51K): [in this window] [in a new window]   FIGURE 1. Library scan of PCC T cells. A decapeptide library consisting of 200 mixtures arranged in a positional scanning format was used to deconvolute the specificity of splenic lymphocytes derived from (BAND x B10.BR)F1 transgenic mice that express the TCR ß-chains specific for PCC (7 ). The x-axis represents no stimulation (&) or stimulation with library mixtures having a defined amino acid (single letter code A, C, D, etc.) at an indicated position. The y-axis represents proliferative activity (cpm) following 3 days of culture (see Materials and Methods for details).

Based on this information, a panel of 48 different peptide sequences (TPI-799), representing all of the possible combinations of the library predictions that may mimic PCC, was synthesized. These were then used in serial dilutions to determine the EC₅₀ for each peptide, i.e., that concentration of peptide causing half-maximal proliferative response of the PCC T cells. The EC₅₀ values of these peptides along with three control peptides are shown in Table I. These control peptides included the smaller 10-mer fragment (PCC10) 95–104 IAYLKQATAK, representing the C-terminal sequence of PCC17 predicted from the library screen, a nonamer cytochrome c fragment from moth (MCC9) 95–103 IAYLKQATK; and a decamer fragment from human (HCC10) 95–104 IAYLKKATNE.

These data indicate three important features of this collection of PCC mimic peptides. First, 25% of the peptides deduced from the library scan were more effective than the native PCC peptides in stimulating proliferative responses by the PCC transgenic T cells. The EC₅₀ value for the native 10-mer PCC peptide was in the 100 nM range; however, 12 of the peptides deduced from the library scans were more effective than this ligand, some of them having EC₅₀ values in the 1–10 nM range. Seven of these peptides were more effective than MCC9, among the most effective of the natural ligands for this TCR ß-chain combination (11, 12). Second, it seems clear that no correlation exists between the number of substitutions and the proliferative activity of the peptides.

The final conclusion relates to the individual amino acid usage for the E^k MHC-binding residues (95, 98, 100, and 103 corresponding to position 1, 4, 6, and 9, respectively) defined from biochemical data (13, 14) and high-resolution x-ray crystal structures (15) and the presumed TCR contact residues (residues 97, 99, and 102 corresponding to peptide positions 3, 5, and 8, respectively) for PCC (12). Comparison of the most effective peptides with nanomolar activity and the least effective peptides with micromolar activity indicates that there is no difference in the E^k-binding residues I, F/P, A, and K at peptide positions 1, 4, 6, and 9 taken individually, respectively, for example for the four peptides p28, 30, 32, 36, and 36. Therefore, it might be presumed that these two groups of peptides have approximately similar binding affinities to the E^k molecule, but that the residues at the TCR contact positions should be different. Contrary to expectations, the most effective peptides use the individual amino acids Y/A, I/K, and T at positions 3, 5, and 8, respectively, the same residues present in the least effective peptides. This finding implies that the contribution of individual amino acids in specific positions of a peptide, i.e., as MHC anchor residues or TCR-binding positions, is greatly outweighed by the combination of amino acids comprising a specific binding motif in the overall peptide sequence.

The specific combination of amino acids critical for effective E^k- and TCR-binding motifs can be deduced from the most active peptides shown in Table I. Three of the most common motifs for E^k and TCR binding are listed in Table II, along with those peptides among the 48 within the TPI-799 collection that have these motifs. The peptides indicated in bold have EC₅₀ values <100 nM. From this list, it is possible to rank the motifs in order of their effectiveness based on the frequency of active peptides using them. For example, 7 of 12 of the most active peptides have I, F, A, and K at anchor positions 1, 4, 6, and 9, respectively, and this is ranked as the most optimal E^k motif, while only 3 of 12 peptides have I, P, A, and K, and this is ranked as a suboptimal motif. In order of their frequency of usage among the active peptides the most common TCR motifs at positions 3, 5, and 8 were YKT, YIT, and AKT, respectively. It should be noted that the most effective peptide (p12) uses the optimal E^k motif and a suboptimal TCR motif; the next most active peptide (p16) uses the opposite pattern. Only three peptide sequences (p2, 4, and 6) include the combination of both optimal TCR and E^k motifs, and these three are among the most active.

Data in Table III also indicate an interesting feature of position 7 for activity of some of the PCC mimic peptides. In peptide 4, for example, a single YA substitution at position 3 results in an analogue with no activity (p28). However, double A substitutions at positions 3 and 7 restore activity for an analogue (p26) of this peptide. In the context of peptide binding to E^k, position 7 is considered to be exposed to solvent (15).

Immunogenicity of PCC mimic peptides

An important issue is whether these PCC mimic peptides, defined from library scans of a clonotypic T cell population, are effective immunogens and, in particular, whether they provoke T cell responses against the original PCC native ligand. Several of these peptides having EC₅₀ values in the 1–10-nM, the 500-nM, and 10,000-nM range were emulsified in CFA and used to immunize normal B10.BR mice. Splenic T cells were recovered from these animals 15–18 days after immunization, purified by passage through nylon wool, and stimulated in microwell cultures with various dilutions of the immunizing peptides and with the native PCC 10-mer peptide ligand. Fig. 2 shows the results of proliferative responses by lymph node T cells from mice immunized with two of the PCC mimic peptides. One, peptide 12 with a suboptimal TCR motif (YIT) and the optimal E^k binding motif (IPAK), has the most effective EC₅₀ (1 nM). The other, peptide 15 with an optimal TCR motif (YKT) and one of the most suboptimal E^k motifs (IPAT), has one of the least effective EC₅₀ values (>10,000 nM). Both peptides proved to be immunogenic, generating T cell populations that respond well in culture to the immunizing peptide, and it should be noted that T cells from mice immunized with the high-potency peptide 12 responded equally well to the native PCC peptide 49. However, contrary to expectations, mice immunized with the low-potency peptide 15 also responded as well, if not better, to the native peptide 49.

View larger version (24K): [in this window] [in a new window]   FIGURE 2. Immunogenicity and cross-reactivity of PCC mimic peptides. Peptides 12 (IAYFIAPTKF), with an EC50 in the nanomolar range and peptide 15 (IAYPKASTTF), with an EC50 in the 10 µM range, were emulsified in CFA and used to immunize normal B10.BR mice. Nylon wool-passaged T cell populations were prepared from draining lymph nodes 15–18 days later. These were stimulated (300 x 103 cells per well) for 3 days with various dilutions of the immunizing peptides and peptide 49, the native 10-mer PCC sequence (IAYLKQATAK), in the presence of irradiated syngeneic spleen cells.

The same decapeptide library was used at 100 µg/ml in a series of four replicate experiments to deduce candidate peptide sequences that best stimulate proliferative responses by the human CD4⁺ T cell clone TL3A6 reactive to the 11-mer fragment 86–96 of MBP. From the library predictions indicated in Table V, a series of 36 peptide mimics (K38) having acetylated N and amide C termini was synthesized and used for additional experiments. Eight of these peptides in the K38 series along with their EC₅₀ values on the TL3A6 clone and the number of amino acid substitutions are indicated in Table V. Several of these peptides were two orders of magnitude more effective than the native MBP sequence in stimulating the TL3A6 clone.

The immunogenicity of the eight MBP analogues deduced from the library scans of the human TL3A6 clone, and the question whether these peptides are able to induce immune responses effective against the native MBP epitopes, was explored in two ways: 1) whether they can provoke active EAE in vivo, and 2) whether they stimulate T cell proliferation.

EAE is a T cell-mediated neuroinflammatory disease of the CNS that can be induced in LEW rats following active immunization with MBP or with either of two encephalitogenic peptide fragments of MBP in CFA. MBP_68–86 of guinea pig origin (YGSLPQKSQRSQDENPVVH) is highly encephalitogenic; paralytic disease generally develops 9–11 days after immunization and lasts for 1 wk. The rat 68–86 peptide sequence differs by a single amino acid, threonine for serine, at position 78 (TS) and is weakly immunogenic as an autoantigen. The second encephalitogenic fragment of MBP, 87–99, is highly conserved, having an identical amino acid sequence (VHFFKNIVTPRTP) in humans, guinea pigs, and rats, but it is also only weakly encephalitogenic; disease onset is slightly later and the extent and duration of disease is significantly less. If any of the peptide mimics of the MBP_87–99 sequence provoke EAE, it can be presumed that the T cell response to the immunizing peptide cross-reacts with a natural MBP ligand present in the CNS.

This strategy explores the question whether peptide mimics deduced from a library scan on a human T cell clone reactive to an epitope that is encephalitogenic in animals will also induce autoimmune disease in the animal model. If so, this cross-species result would strongly support the relevance of this myelin peptide epitope in human disease. LEW rats were immunized with several candidate sequences deduced from the library scan on the human TL3A6 clone having a range of different EC₅₀ values. Two control peptides, the native 13-mer fragment 87–99 VHFFKNIVTPRTP and the native 10-mer fragment 89–98 FFKNIVTPRT (peptide K38-37) were also included. The day of onset, the duration, and clinical extent of paralytic EAE was recorded for these animals.

Table VI shows the results of a series of different experiments using several of the K38 L-amino acid peptide sequences and, in addition, analogues of the peptide 23 sequence synthesized with D-amino acids and as a retro-inverso (Dri-) peptidomimetic containing reversed peptide bonds between each residue along the D-amino acid sequence. Several conclusions can be drawn from these data. First, somewhat surprisingly, of the eight different K38 L-amino acid peptide mimics tested, three (p9, p23, and p31) induced severe EAE with disease indices in the range of 20–30. A score of 30 is about as severe a morbidity score as can be for EAE without being fatal. One other peptide, K38-1, was capable of inducing EAE, but with a weak to moderate disease severity score (disease index, 10) more typical of that caused by immunization with the native 87–99 peptide (data not shown). Four other peptides (p19, p21, p33, and p35) caused very little, if any, disease. Second, in general, the extent of disease induced with these peptides was somewhat less as the immunizing dose of Ag was increased. Third, there appears to be no correlation between EC₅₀ values of the various peptides as stimulators of the TL3A6 human T cell clone and their capacity to induce EAE in rats. Finally, one finding was unexpected: the D- and retro-inverso analogues of p23, especially at lower immunization doses, proved to be severely encephalitogenic. These results indicate that several of the MBP mimic peptides deduced from the library scan on a human T cell clone are capable of inducing EAE with significant disease severity in the LEW rat, and two of these peptides synthesized with unnatural amino acids cause near maximal disease.

View larger version (15K): [in this window] [in a new window]   FIGURE 3. Immunogenicity of MBP peptide mimics. LEW rats were immunized with peptide 23, an L-amino acid analogue of peptide 37 (MBP89–98, see Table V) or with a D-retro-inverso analogue of p23. Draining lymph nodes were recovered 18 days later, and nylon wool-purified T cells were prepared and cultured for 3 days with serial dilutions of the immunogen and the native peptide 37 in the presence of irradiated syngeneic thymocytes.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

TCR degeneracy

It has been generally accepted that T cell recognition of peptide/MHC molecules is exquisitely specific and that the potential to respond to a vast number of different potential antigenic peptides is achieved by the diversity of TCR chain genes and the process whereby these are rearranged during the assembly of mature TCRß heterodimers to generate an extensive repertoire of T cell specificities (17). In large part, this notion, combined with the process of negative selection during T cell ontogeny in the thymus, appeared to be a solution to the problem of avoiding cross-reactive immune responses that might also result in autoimmunity.

However, one problem with this scenario is that the numbers of possible different T cell peptide epitopes are several orders of mag- nitude greater than the numbers of T cells comprising the lymphon of an individual. This leads to the notion that the expressed T cell repertoire of an individual can respond to only a small fraction of peptides from potentially pathogenic organisms. The number of different possible MHC class I-associated nonamer L-amino acid peptides available for recognition by CD8⁺ T cells is large (20⁹ = 5 x 10¹¹), and, because of their greater length, the number of possible peptide sequences, for example 12- to 16-mers, available for presentation to CD4⁺ T cells by MHC class II molecules is significantly larger (20¹² to 20¹⁶ = 4 x 10¹⁵ to 6.5 x 10²⁰). But the total number of T cells in a mouse is of the order 10⁸, for humans 10¹¹, and the number of different clonotypic specificities in these T cell repertoires is less. Because of this discrepancy between the number of different potentially antigenic peptides, and the number of different T cell clonotypes available for immune reactivity, it seems clear that Mason’s argument in a recent article (18) provides a solution to this paradox: namely, T cell specificity must be highly degenerate to allow for adequate protective immune responses to any T cell epitope.

We show here and others (1, 2, 3, 5, 6, 16) have recently demonstrated that a synthetic combinatorial decapeptide library used at a concentration of 100 µg/ml is capable of stimulating activation and function of selected T cell clones in microwell cultures. This decamer library consists of all possible combinations of L-amino acid peptide ligands in approximate equimolar concentration systematically arranged in a positional scanning format with 200 mixtures. Mixture 89 (K5X₉), for example, stimulates a significant proliferative response by PCC T cells in culture that is 4-fold above background stimulation levels (see Fig. 1). This mixture has 3.2 x 10¹¹ (19⁹) different decamer peptides and, at a library concentration of 100 µg/ml, each different peptide is present at a concentration of 3 x 10^-10 µM. Typically, a CD4⁺ T cell clone requires a concentration of native peptide ligand in the range of 10^-2 µM for detectable stimulation, a concentration 30 x 10⁶-fold greater than that of a single peptide in the library mixture. Thus, finding that the K5X₉ mixture causes a detectable clonal response implies that as many as 30 million different peptides present in the library mixture contributed to this process.

The finding that T cell specificity may be so promiscuous raises the paradoxical problem of why cross-reactive autoimmune responses are not as prevalent as might be expected for such an extent of TCR degeneracy. For activated CD4⁺ T cells that recognize peptide/MHC class II complexes, part of the answer may be that expression of class II molecules is limited to relatively few cells of an individual, a circumstance that would diminish opportunistic autoimmune cross-reactivity. For CD8⁺ T cells, in contrast, MHC class I molecules are expressed almost ubiquitously, and the possibility of autoimmune cross-reactivity would seem to be significant. Studies currently in progress suggest that TCR degeneracy of the CD8⁺ subset may be several orders of magnitude less than for CD4⁺ T cells, thus providing one possible solution to this apparent paradox.

Deduced PCC mimic peptides

We opted to explore the use of libraries to identify peptide mimics in the PCC system because much is known of the MHC and TCR contact residues from a variety of sources including the crystal structure of the peptide/E^k complex (13, 14, 15). Initial library predictions (See Table I) generated from these scans showed that amino acids indicated for 6 of the 10 positions corresponded to the correct PCC sequence and identified an epitope recognized by these clonotypic T cells at the C terminal of the 17-mer PCC sequence. This demonstrates that such library scans can be used with T cells of unknown but relevant clinical specificity to identify native protein immunogens from appropriate protein database searches as well as the involved T cell epitopes themselves. It seems likely that this strategy might provide new approaches for identification of T cell ligands, an essential part of the problem of vaccine design.

Forty-eight different PCC analogues predicted from the library scan of PCC-reactive splenic T cells (see Fig. 1) were synthesized and assessed for their potency in causing proliferative responses by this clonotypic T cell population; 26 of these are indicated in Table I ranked by their EC₅₀ values. There are several conclusions to be drawn from this data. First, when assessed in terms of their proliferative activity, a significant fraction (12/48 = 25%) of these deduced sequences had EC₅₀ values substantially (5–500 fold) better than the native 17-mer PCC peptide. Many of the most effective of these decapeptides had substitutions in as many as 6 of the 10 positions and bear little resemblance to the native ligand. It is of interest that the number of amino acid substitutions is not appreciably different for the most effective peptides with activity in the 1–10 nM range and the less effective peptides having activity in the 10 µM range. It is of further interest that many of the peptides most active on PCC T cells, for example peptide 12, have as many as 6 different substitutions compared with the native sequence, most of which involve amino acids having different chemical character, i.e., nonconservative changes. This finding demonstrates the importance of the deconvolution strategy used in this and previous studies of T cell specificity, wherein data derived from library scans is used to synthesize peptide sequences representing the combinations of the most active amino acids (5, 6, 8) rather than the approach of using library scan data to alter the native ligand in the search for more active ligands (1, 16).

Second, the finding that the most and least effective PCC mimics share the same amino acids at individual anchor positions presumed to be involved in MHC binding and in other positions thought to function as TCR contact residues indicates that no amino acids are strictly required for these positions. This is difficult to reconcile with notions of TCR/peptide/MHC interactions based on strict involvement of selected amino acids at defined positions. The present data extend previous views of TCR peptide recognition based on the finding that specificity of an alloreactive T cell hybridoma is highly flexible and that such cells are capable of recognizing numerous different MHC-associated peptide epitopes (19). Taken together, these data indicate that the total combination of amino acids in the peptide sequence is a critical feature in the recognition of Ag by T cells, a notion in agreement with previous predictions (3, 8). Table II shows that MHC and TCR binding motifs can be deduced and ranked in terms of their contribution to the stimulatory activity of a peptide from the frequency of active peptides that incorporate these motifs within their sequence. An active peptide (EC₅₀ <100 nM) requires that either the TCR or MHC motif be an optimal one; the other can be second or third rank. But, it does not follow that a third-rank motif with an optimal one will necessarily generate an active peptide. Actual binding MHC and TCR binding affinities for these peptides are currently being assessed to determine how such binding data correlates with ranking obtained with proliferative activity.

A finding of some interest is that several of the PCC mimic peptides least effective as agonists and having optimal MHC motifs and suboptimal TCR motifs are highly effective as antagonists in culture; they cause a significant inhibition of responses by T cells prepulsed with superagonist peptides (D.B.W. et al., manuscript in preparation).

Third, the results reported in Table III with four different PCC mimic peptide pairs extend this notion one step further and suggest that two peptides differing by a single amino acid and having similar activity in stimulating a T cell clone may interact with MHC and TCR molecules in different ways. If two peptides, for example p26 and p2, with an alanine to tyrosine substitution at position 3 (3A3Y) have similar EC₅₀ values (50 nM), one might assume that the A and Y residues interact in a similar way in MHC and TCR binding; these two amino acids might have a major role, a minimal role, or no interaction with MHC and TCR, but the entire chain of both peptides binds similarly. If this were so, it would be expected that identical changes in these two peptides would yield similar results. The data in Table III fail to support this simplistic model. Identical single and double substitutions at positions 4, 5, and 7 in these two peptides failed to alter activity of a series of different 3Y peptides but resulted in complete loss of activity in the series of 3A peptides. It follows from this that peptides p26 and p2 with identical amino acid sequences (other than 3A and 3Y) must either be using different contact residues for MHC/TCR binding or the same residues binding in different ways to different side chains in the MHC and TCR molecules. It seems significant that the most "immutable" PCC mimic peptides 2, 4, and 6 are the ones that include the both of most optimal TCR and E^k binding motifs within their sequences. These finding support recent conclusions from structural studies concerning extensive plasticity in T cell recognition of the peptide/MHC complex (20).

Immunogenicity of peptide mimics for native peptide ligands

Our principal objective in this study was to determine whether peptide mimics deduced from library scans of clonotypic T cells would be immunogenic and induce T cell responses to the native peptide ligands used to generate these T cells. The results with the PCC_95–104 peptide mimics (Fig. 2 and Table IV) and the MBP_89–98 mimics (Fig. 3 and Table VI) were similar. As expected, the various mimic peptides in these two systems proved to be potent immunogens that stimulated strong T cell-mediated immune responses against the immunizing peptides. All six of the PCC decapeptide mimics synthesized with L-amino acids (Fig. 2 and Table IV) and one of the MBP mimics that was tested (p23; Fig. 3) generated good T cell responses against themselves in proliferation assays. This finding demonstrates that peptide sequences deduced from library scans having a range of EC₅₀ values, for example 1 nM to 10 µM on the PCC T cell population, have the appropriate amino acid combinations required for immunogenicity.

Four other findings concerning the immunogenicity of these peptide mimics are worthy of further comment. First, while these peptide mimics induced good immune responses against themselves, they were also surprisingly effective in inducing T cell responses against the original native PCC and MBP ligands. We used two different models for assessing immunogenicity to native ligands: T cell proliferation assays and whether mimics of an autoimmune peptide would cause active autoimmune disease. In the PCC system, five peptide mimics generated strong T cell proliferative responses to the native PCC peptide 95–104 (799–49 in Table IV), and p23 in the MBP model generated good proliferative responses against the native MBP ligand (p37; Fig. 3). Also in the MBP model, five of the eight MBP mimics synthesized with L-amino acids proved not only to be good immunogens, but were significantly encephalitogenic in rats (Table VI). This finding shows that numerous different mimics of an autoantigen peptide epitope having a variety of multiple amino acid substitutions can cause potent immune responses directed to that autoantigen, and it further demonstrates the extensive cross-reactivity of T cell responses.

Second, the effectiveness of the peptide mimics in generating T cell responses to the native ligands appears to be independent of the activity of these peptides for the clonotypic T cell population used in their selection from library scans (Fig. 2 and Table IV). PCC mimic peptides p11, p15, and p36, having EC₅₀ values in the 10,000-nM range, were as effective in stimulating T cell responses to the native ligand as peptide p12, which is four orders of magnitude more active on PCC T cells. Peptide 12 has a second-rank TCR motif and the optimal E^k motif in terms of the frequency of active peptides (Table II). Peptides 15 and 36 use either an optimal TCR or E^k motif with the other being suboptimal, but peptide 11 uses a second-rank TCR motif and a third-rank E^k motif. These findings indicate that the requirements for in vivo immunogenicity may be less stringent than for in vitro stimulation, and, again, they support the conclusion concerning extensive TCR degeneracy. The observation that peptides with low EC₅₀ activity on a clonotypic population and suboptimal MHC motif are effective at inducing responses to the native peptide ligand is a somewhat surprising one. Currently, we are exploring the immunogenicity of these peptides in terms of their ability to stimulate transgenic T cells in vivo.

Third, peptide analogues synthesized with D-amino acids or as retro-inverso peptidomimetics with reversed peptide bonds between component D-amino acids (Dri-) are surprisingly effective at inducing immune responses to their native L-amino acid counterparts. This was demonstrated both for T cell proliferative responses in the PCC and MBP models and in their ability to induce paralytic EAE. The immunogenicity of D- and Dri-peptide analogues for T cell responses remains controversial (21, 22, 23), especially concerning their ability to be processed and presented for MHC binding. Given their resistance to proteolysis and their longer half-lives in vivo resulting in more stable bioactive analogues, D- and Dri-peptidomimetics might ultimately be used as effective immunogens. Why the unnatural Dri-p23 peptidomimetic and the L-amino acid p23 analogues of MBP apparently cause less severe EAE at higher immunizing doses is not clear. One possibility currently under consideration is that these analogues may provoke Ab responses that partially inhibit codevelopment of pathogenic T cell responses.

Finally, the finding that a series of peptides that are highly encephalitogenic in rats can be deduced from library scans on a human T cell clone reactive to a suspected autoantigenic epitope in MS may be a useful one. It suggests the possibility that some T cell epitopes suspected of underlying human autoimmune disease can be assessed directly in an animal model. It also implies that the human class II MHC DR2a molecule associated with MS and the class II RT1.D^l gene product of the LEW rat may have similar binding affinities for T cell epitopes, a suggestion that has been made before (24).

In summary, we show that most of the peptide mimics deduced from peptide library scans of a T cell clone are good immunogens for inducing T cell responses to native peptide ligands. This provides an important strategy for identifying from T cell clones of relevant clinical specificity a variety of agonist and superagonist ligands that can be expected to be potent immunogens and therefore of importance in the design of vaccines for future use in the prevention and treatment of human disease.

	Acknowledgments

 
We thank our colleagues at Multiple Peptide Systems (San Diego, CA) for making the decamer peptide library PCL 97-4 available to us.

	Footnotes

 
¹ This work was supported by United States Public Health Grants AI-22519 and CA-78040 (D.B.W.), AI-31231 (J.K.), and National Science Foundation Grant CHE-9520142 (R.A.H.). B.H. was a fellow of the Deutsche Forschungsgemeinschaft (He 2386/1-2).

² Address correspondence and reprint requests to Dr. Darcy B. Wilson, Torrey Pines Institute for Molecular Studies, 3550 General Atomics Court, San Diego, CA 92121-1122. E-mail address:

³ Current address: Neurologische Klinik mit Poliklinik, Rudolf-Bultmann-Strasse 8, 35039 Marburg, Germany.

⁴ Abbreviations used in this paper: PCC, pigeon cytochrome c; MBP, myelin basic protein; Dri-, retro-inverso-peptidomimetic analogue that contains reversed peptide bonds between each residue along a D-amino acid sequence; EAE, experimental allergic encephalomyletitis; MS, multiple sclerosis.

Received for publication March 16, 1999. Accepted for publication September 23, 1999.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Hemmer, B., B. Fleckenstein, M. Vergelli, G. Jung, H. McFarland, R. Martin, K.-H. Wiesmüller. 1997. Identification of high potency microbial and self ligands for a human autoreactive class II-restricted T cell clone. J. Exp. Med. 185:1651.[Abstract/Free Full Text]
2. Hemmer, B., C. Pinilla, J. Appel, J. Pascal, R. Houghten, R. Martin. 1998. The use of soluble synthetic peptide combinatorial libraries to determine antigen recognition of T cells. J. Peptide Res. 52:338.[Medline]
3. Hemmer, B., M. Vergelli, C. Pinilla, R. Houghten, R. Martin. 1998. Probing degeneracy in T-cell recognition using peptide combinatorial libraries. Immunol. Today 19:163.[Medline]
4. Blake, J., J. V. Johnston, K. E. Hellström, H. Marquardt, L. P. Chen. 1996. Use of combinatorial peptide libraries to construct functional mimics of tumor epitopes recognized by MHC class I-restricted cytolytic T lymphocytes. J. Exp. Med. 184:121.[Abstract/Free Full Text]
5. Gundlach, B. R., K.-H. Wiesmüller, T. Junt, S. Kienle, G. Jung, P. Walden. 1996. Specificity and degeneracy of minor histocompatibility antigen-specific MHC-restricted CTL. J. Immunol. 156:3645.[Abstract]
6. Udaka, K., K.-H. Wiesmüller, S. Kienle, G. Jung, P. Walden. 1996. Self-MHC-restricted peptides recognized by an alloreactive T lymphocyte clone. J. Immunol. 157:670.[Abstract]
7. Kaye, J., M.-L. Hsu, M.-E. Sauron, S.C. Jameson, N. R. J. Gascoigne, S. M. Hedrick. 1989. Selective development of CD4⁺ T cells in transgenic mice expressing a class II MHC-restriced antigen receptor. Nature 341:746.[Medline]
8. Hemmer, B., M. Vergelli, B. Gran, N. Ling, P. Conlon, C. Pinilla, R. Houghten, H. F. McFarland, R. Martin. 1998. Cutting edge: predictable TCR antigen recognition based on peptide scans leads to the identification of agonist ligands with no sequence homology. J. Immunol. 160:3631.[Abstract/Free Full Text]
9. Pinilla, C., J. R. Appel, R. A. Houghten. 1994. Investigation of antigen-antibody interactions using a soluble nonsupport-bound synthetic decapeptide library composed of four trillion sequences. Biochem. J. 301:847.
10. Houghten, R. A.. 1985. General method for the rapid solid-phase synthesis of large numbers of peptides: specificity of antigen-antibody interaction at the level of individual amino acids. Proc. Natl. Acad. Sci. USA 82:5131.[Abstract/Free Full Text]
11. Kaye, J., N. J. Vasquez, S. M. Hedrick. 1992. Involvement of the same region of the T-cell antigen receptor in thymic selection and foreign peptide recognition. J. Immunol. 148:3342.[Abstract]
12. Reay, P. A., R. M. Kantor, M. M. Davis. 1994. Use of global amino acid replacements to define the requirements for MHC binding and T cell recognition of moth cytochrome c (93–103). J. Immunol. 152:3946.[Abstract]
13. Leighton, J., A. Sette, J. Sidney, E. Appella, C. Ehrhardt, S. Fuchs, L. Adorini. 1991. Comparison of structural requirements for interaction of the same peptide with I-E^k abd I-E^d molecules in the activation of MHC class II-restricted T-cells. J. Immunol. 147:198.[Abstract]
14. Boniface, J. J., N. L. Allbritton, P. A. Reay, R. M. Kantor, L. Stryer, M. M. Davis. 1993. pH affects both the mechanism and the specificity of peptide binding to a class II major histocompatibility complex molecule. Biochemistry 32:11761.[Medline]
15. Fremont, D. H., W. A. Hendrickson, P. Marrack, and J. Kappler. 96 A.D. Structures of an MHC class II molecule with covalently bound single peptide. Science 272:1001.
16. Gundlach, B. R., K.-H. Wiesmüller, T. Junt, S. Kienle, G. Jung, P. Walden. 1996. Determination of T cell epitopes with random peptide libraries. J. Immunol. Methods 192:149.[Medline]
17. Kisielow, P., H. von Boehmer. 1995. Development and selection of T cells: facts and puzzles. Adv. Immunol. 58:87.[Medline]
18. Mason, D.. 1998. A very high level of crossreactivity is an essential feature of the T-cell receptor. Immunol. Today 19:395.[Medline]
19. Daniel, C., S. Horvath, P. M. Allen. 1998. A basis for alloreactivity: MHC helical residues broaden peptide recognition by the TCR. Immunity 8:543.[Medline]
20. Garcia, K. C., M. Degano, L. R. Pease, M. Huang, P. A. Peterson, L. Teyton, I. A. Wilson. 1998. Structural basis of plasticity in T cell receptor recognition of a self peptide-MHC antigen. Science 279:1166.[Abstract/Free Full Text]
21. Bartnes, K., K. Hannestad, G. Guichard, J. P. Briand. 1997. A retro-inverso analog mimicks the cognate peptide epitope of a CD4⁺ T cell clone. Eur. J. Immunol. 27:1387.[Medline]
22. Meziere, C., M. Viguier, H. Dumortier, R. Lo-Man, C. LeClerc, J.-G. Guillet, J.-P. Briand, S. Muller. 1997. In vivo T helper cell respnse to retro-inverso peptidomimetics. J. Immunol. 159:3230.[Abstract]
23. Hervé, M., B. Maillére, G. Mourier, C. Texier, S. Leroy, A. Ménez. 1997. On the immunogenic properties of retro-inverso peptides. Total retro-inversion of T-cell epitopes causes a loss of binding to MHC II molecules. Mol. Immunol. 34:157.[Medline]
24. Reizis, B., M. Eisenstein, F. Mor, I. R. Cohen. 1998. The peptide-binding strategy of the MHC class II I-A molecules. Immunol. Today 19:212.[Medline]

This article has been cited by other articles:

				P. Mana, M. Goodyear, C. Bernard, R. Tomioka, M. Freire-Garabal, and D. Linares Tolerance induction by molecular mimicry: prevention and suppression of experimental autoimmune encephalomyelitis with the milk protein butyrophilin Int. Immunol., March 1, 2004; 16(3): 489 - 499. [Abstract] [Full Text] [PDF]

				A. W. Purcell and J. J. Gorman Immunoproteomics: Mass Spectrometry-based Methods to Study the Targets of the Immune Response Mol. Cell. Proteomics, March 1, 2004; 3(3): 193 - 208. [Abstract] [Full Text] [PDF]

				Y. Uemura, S. Senju, K. Maenaka, L. K. Iwai, S. Fujii, H. Tabata, H. Tsukamoto, S. Hirata, Y.-Z. Chen, and Y. Nishimura Systematic Analysis of the Combinatorial Nature of Epitopes Recognized by TCR Leads to Identification of Mimicry Epitopes for Glutamic Acid Decarboxylase 65-Specific TCRs J. Immunol., January 15, 2003; 170(2): 947 - 960. [Abstract] [Full Text] [PDF]

				C. A. Lawendowski, G. M. Giurleo, Y. Y. Huang, G. J. Franklin, J. M. Kaplan, B. L. Roberts, and C. A. Nicolette Solid-Phase Epitope Recovery: A High Throughput Method for Antigen Identification and Epitope Optimization J. Immunol., September 1, 2002; 169(5): 2414 - 2421. [Abstract] [Full Text] [PDF]

				J. A. M. Borghans and R. J. De Boer Memorizing innate instructions requires a sufficiently specific adaptive immune system Int. Immunol., May 1, 2002; 14(5): 525 - 532. [Abstract] [Full Text] [PDF]

				V. Rubio-Godoy, C. Pinilla, V. Dutoit, E. Borras, R. Simon, Y. Zhao, J.-C. Cerottini, P. Romero, R. Houghten, and D. Valmori Toward Synthetic Combinatorial Peptide Libraries in Positional Scanning Format (PS-SCL)-based Identification of CD8+ Tumor-reactive T-Cell Ligands: A Comparative Analysis of PS-SCL Recognition by a Single Tumor-reactive CD8+ Cytolytic T-Lymphocyte Clone Cancer Res., April 1, 2002; 62(7): 2058 - 2063. [Abstract] [Full Text] [PDF]

				Y. Zhao, B. Gran23, C. Pinilla, S. Markovic-Plese, B. Hemmer, A. Tzou, L. W. Whitney, W. E. Biddison, R. Martin, and R. Simon Combinatorial Peptide Libraries and Biometric Score Matrices Permit the Quantitative Analysis of Specific and Degenerate Interactions Between Clonotypic TCR and MHC Peptide Ligands J. Immunol., August 15, 2001; 167(4): 2130 - 2141. [Abstract] [Full Text] [PDF]

				A. Amrani, P. Serra, J. Yamanouchi, J. D. Trudeau, R. Tan, J. F. Elliott, and P. Santamaria Expansion of the Antigenic Repertoire of a Single T Cell Receptor upon T Cell Activation J. Immunol., July 15, 2001; 167(2): 655 - 666. [Abstract] [Full Text] [PDF]

				C. Trollmo, A. L. Meyer, A. C. Steere, D. A. Hafler, and B. T. Huber Molecular Mimicry in Lyme Arthritis Demonstrated at the Single Cell Level: LFA-1{{alpha}}L Is a Partial Agonist for Outer Surface Protein A-Reactive T Cells J. Immunol., April 15, 2001; 166(8): 5286 - 5291. [Abstract] [Full Text] [PDF]