Distinct recognition by two subsets of T cells of an MHC class II-peptide complex

Abstract

We examine here the nature of the differential recognition by CD4⁺ T cells of a single peptide from hen-egg white lysozyme (HEL) presented by I-A^k class II MHC molecules. Two subsets of T cells (called A and B) interact with the same peptide, each in unique ways that reflect the nature of the complex of peptide and MHC. We show that the A and B set of T cells can be distinguished by their functional interaction with the three T cell receptor (TCR) contact residues of the bound peptide. The dominant peptide of HEL selected from processing is bound in a single register where a critical TCR contact residue is situated about the middle of the core segment of the peptide: all T cells establish functional contact with it. Three sets of T cells, however, can be distinguished by their differential recognition of two TCR contacts situated at the amino and carboxyl sides of the central TCR contact residue. Type A T cells, the conventional cells that see the peptide after processing of HEL, need to recognize all three TCR contact residues. In contrast, the type B T cells recognize the peptide given exogenously, but not when processed: these T cells recognize either one of the peripheral TCR contact residues, indicating a much more flexible interaction of peptide with I-A^k molecules. We discuss the mode of generation of the various T cells and their biological relevance.

We analyze here the nature of the interaction of the 48–63 peptide of hen-egg white lysozyme (HEL) with the I-A^k molecule that results in the generation of two distinctive and very unique T cells. Using a biochemical analysis, we have identified the major peptides presented by the MHC class II molecule I-A^k during processing of the model protein HEL (1). The most abundant epitope was presented as a family characterized by a 9-aa segment, from residues 52–60 (2), that occupied the P1 to P9 segment of the binding site. This peptide family may represent as much as 10% of the total peptides bound to I-A^k (3). In the HEL 52–60 segment (DYGILQINS), D52, I55, Q57, and S60 are the I-A^k anchoring residues whereas Y53, L56, and N59 are solvent exposed and function as the T cell receptor (TCR) contact amino acids (4). This segment is flanked by amino- and carboxyl-terminal residues, with the most frequently identified species starting at residue 48 and ending at either residue 62 or 63 (i.e., DGST-DYGILQINS-RW/W) (5, 6). A number of T cells induced by HEL immunization recognized the core peptide derived from endogenously processed HEL, as well as synthetic peptides offered exogenously to antigen-presenting cells (APCs). We refer to these traditional T cells as type A. We also identified another population of T cells that recognized the peptide given exogenously, but not the processed peptide from HEL. We refer to them as type B (7, 8). Because we ruled out that type B T cells did not recognize posttranslational or postsynthetic changes in the exogenous peptide (8, 9) and, moreover, because we relied on the chemical identity of the naturally processed peptide, we surmised that the exogenously offered peptide must be bound in a different conformation or in a different register to the I-A^k molecules.

The relevance of the type B T cells cannot be underestimated. First, these T cells represent a significant component of the T cell response to peptides (8). (This issue becomes important when using peptides as immunogens in vaccines. Thus a substantial number of T cells will not recognize the immunogen, defeating the purpose of the immunization.) Moreover and very critically is the fact that type B T cells escape negative selection in the thymus when the APCs present the protein—in our case HEL (8). Because these T cells peripheralize, they could be activated at sites of inflammation where there is proteolysis and release of peptides, and thus contribute to the tissue pathology. Finally, as we show below, type B T cells can recognize small peptides in a process that does not involve the auxiliary molecule H-2DM. Such entry of peptides into the presentation pathway may be important in the response to microbes in which there are numerous peptides released from the cell. This article reports on the unique specificity of each of the type A and B T cells that underscores how the core peptide of HEL 52–60 is assembled.

Materials and Methods

Antigen and Peptides.

HEL protein was purchased from Sigma. HEL 48–61 peptide (DGSTDYGILQINSR) and all of the variants were synthesized by using fluorenylmethoxycarbonyl chemistry on a Synergy 432A peptide synthesizer (Applied Biosystems) or Symphony/Multiplex peptide synthesizer (Protein Technologies, Tucson, AZ). The synthesized peptides were purified by C18 reverse-phase HPLC, and the sequence and purity of each peptide was verified by MS.

T Cells, APCs, and Assays.

Some of the T cell hybridomas used in this study have been described: 3A9 (10), ALV11 and DAV21 (7). Other T hybridomas were derived from the draining lymph nodes of mice immunized with HEL protein, or, in the case of type B cells, with HEL 48–61 peptide or the 46–61 tryptic peptide of HEL. All immunizations were done with 10 nmol of the antigen in 100 μl of complete Freund's adjuvant. After boosting ex vivo with HEL protein or peptide for 3 days, the draining lymph nodes cells were fused with the thymoma line BW5147α⁻β⁻. We present the results of the 28 T cells without any previous selection for a particular reactivity. We found that there was about equal representation of type A and B among T cells obtained by immunization with peptide.

The I-A^k-expressing B lymphoma line C3F6 (11) was used as the APC in many of the T cell assays. In some experiments, the C3F6 were fixed in 1% paraformaldehyde, as reported before (10). Chloroquine-treated C3F6 cells were prepared by incubating them in the presence of 0.5 mM chloroquine at 37°C. After 20 min, different doses of antigens were added and incubation was continued for another 4 h at 37°C. Cells were then washed and fixed as above. The T1 and T2 lines expressing the I-A^k molecules (12) were obtained from P. Cresswell (Yale Medical School, New Haven, CT) and tested for presentation to the various T cells. The T2-Ak line lacks the HLA-DM molecule and is defective in processing the 52–60 family of peptides. (Our biological evaluation of this point was reported previously in ref. 13.)

T cell hybridoma assays were performed in 96-well tissue culture plates containing a final volume of 200 μl of DMEM and 5% FCS. T cell hybridoma (1 × 10⁵ per well) were cultured with C3F6 (5 × 10⁴ per well) in the presence of indicated doses of antigens. Fixed or chloroquine-treated C3F6 were used at 1 × 10⁵ or 2 × 10⁵ per well, respectively. After incubation for 18 h at 37°C in 5% CO₂, 100-μl aliquots of the culture supernatant from each well were assayed for IL-2 production by using the IL-2-dependent cell line CTLL-2. Every T cell (Table 1) was tested in 2–6 independent experiments.

Table 1.

Summary of the response of type A and type B T cells to HEL peptides and their TCR analysis

T cell	HEL	48–61	48–60	48–59	48–58	52–61	52–58	Ala substitutions			TCR
								Tyr-53	Leu-56	Asn-59	Vα	Jα	Vβ
Type A
3A9	++++	+++	−	−	−	++	−	−	−	−	TRAV9D-4	TRAJ42	8.1/8.2
2H.3B1	++	++	−	−	−	−	−	−	−	−	N/D	N/D	8.3
HP.2C1	+++	+++	+	+	+	++	−	−	−	+++	TRAV14D-1	TRAJ31	8.1/8.2
P1.4A2	+	+	−	−	−	−	−	−	−	−	TRAV13	TRAJ34	4
2H.1D6	++	+++	−	−	−	+	−	−	+	−	TRAV4D-3	TRAJ12	N/D
2H.3C5	++	+++	+	+	−	++	−	−	−	−	TRAV6-7DV	TRAJ48	11
ALV11	++	++	−	−	−	++	−	−	−	−	TRAV12-03	TRAJ13	17
P1.1A1	++++	++++	+	+	−	+++	−	−	−	+	TRAV7D-4 and 8-1	TRAJ53 and 52	12
CP4.22	++	+++	+	+	−	+	−	−	−	−	TRAV1	TRAJ28	8.1/8.2
2H.1A2	+++	++++	++	−	−	+++	−	−	−	+	TRAV7D-4 and 6D-5	TRAJ12 and 112-2	10b
Type B long
MLA11.2	−	++	++	++	−	+	−	++	−	−	TRAV9D-4	TRAJ5	8.1/8.2
HP.1A5	−	++	++	+	−	+	−	++	−	−	TRAV12D-3	TRAJ47	2
HP.3A1	−	++	++	+	−	−	−	++	−	−	TRAV13D-1	TRAJ32	2
CP2.11	−	++	++	+	−	+	−	++	−	−	TRAV12D-3	TRAJ34	2
P2.4B2	−	++	++	++	−	+	−	+	−	−	N/D	N/D	5.1/5.2
P1.5C3	−	++	++	++	−	+	−	+	−	−	TRAV16	TRAJ22	11
P1.2B2	−	++	++	++	−	++	−	++	−	−	TRAV12D-2	TRAJ22	8.1/8.2
P2.2A2	−	++	++	+	−	+	−	++	−	−	TRAV13-2	TRAJ32	2
P2.5A4	−	++	++	++	−	+	−	++	−	−	TRAV13-2 and 8D	TRAJ56 and 112-2	4
Type B short
CP3.42	−	++	++	++	++	++++	++++	−	−	++	TRAV12D-3	TRAJ34	2
DAV21	−	++	++	++	++	++++	++++	−	−	++	TRAVD/DV11	TRAJ32	4
HP.1A2	−	++	++	++	++	++++	+++	−	−	++	TRAV6-3	TRAJ45	6
CP1.7	−	++	++	++	++	++++	++++	−	−	++	TRAV9D-4	TRAJ6	8.1/8.2
B10.5	−	+++	+++	+++	+++	+++++	+++++	−	−	+++	TRAV12	TRAJ32	8.1/8.2
A12.6	−	++	+	+	+	++++	++	−	−	++	TRAV6-7DV9	TRAJ6	6
A81.25	−	+	+	+	−	++	+	−	−	+	TRAV9D-4	TRAJ31	8.1/8.2
P1.2B4	−	+++	+	++	+	++++	++	−	−	+++	N/D	N/D	9
BL28.13	−	+++	+++	+++	++	++++	+++	−	−	+++	TRAV12-3 and 16/DV11	TRAJ13 and 21	7

On top is a ribbon diagram of the 48-61 peptide showing the TCR contact residues (taken from ref. 4). Shown are all the results of the functional assays. Every T cell was tested at least twice. The reactivity of each T cell to a given peptide is designated according to the antigen concentration that stimulates the half-maximal T cell response: +++++, <0.001 μM; ++++, 0.001 μM–0.01 μM; +++, 0.01 μM–0.1 μM; ++, 0.1 μM–1.0 μM; +, 1.0 μM–10 μM; −, >10 μM. The TCRα V- and J-, and TCRβ V-gene segment expression is given. N/D indicates that the gene segment could not be identified by PCR or FACS. 

In Vitro Peptide Binding Assay.

Binding assays were done by using baculovirus-generated preparations of I-A^k molecules containing a CLIP sequence tethered to a thrombin-sensitive linker (details are given in refs. 14 and 15). Binding was performed by a competitive binding assay and expressed as the 50% inhibitory dose, which is comparable to the K_D. For estimation of the time of dissociation of the complex, the peptides were labeled with ¹²⁵I and incubated with the I-A^k molecules for 24 h, after which the complex was isolated and incubated in the presence of a 1,000-fold excess of unlabeled 48–61 peptide.

Identification of the TCRs Expressed by Hybridomas.

TCRα genes were identified by reverse transcriptase–PCR (RT-PCR). RNA was isolated from 1 × 10⁷ hybridomas with the Qiagen RNeasy RNA extraction kit using the manufacturer's cytoplasmic RNA isolation protocol. RT-PCR was performed with the SuperScript One-Step RT-PCR System with Platinum Taq (Invitrogen) using standard manufacturer conditions and 250 ng of RNA per reaction. Primer sequences can be found as supporting information on the PNAS web site, www.pnas.org. PCR products were prepared for sequencing by treatment with exonuclease I and shrimp alkaline phosphatase for 30 min at 37°C followed by heat inactivation at 80°C for 15 min. Purified PCR products were sequenced directly by using the 10 pmol of the CαA primer and Big Dye Terminator Cycle Sequencing Ready Reaction (Applied Biosystems). Sequencing data were analyzed with vector nti software (Informax, Bethesda, MD) and the international ImMunoGeneTics database (IMGT, http://imgt.cines.fr) (16–19). Cell surface expression of TCRβ genes was determined by cell surface immunofluorescence staining with the TCRβ Screening Panel (PharMingen) under manufacturer's conditions. Immunofluorescence data were collected on a FACScalibur (Becton Dickinson) and analyzed with cellquest software (Becton Dickinson).

Results

Two Subsets of Type B Cells Segregate According to Their Preference for the Core Segment HEL 52–60.

The specificity of all of the type A and type B T cells to the segment 52–60 of HEL is summarized in Table 1. Type A T cells were obtained by immunizing with HEL or peptide, whereas the type B were obtained by peptide immunization. Fig. 1A shows the response of a typical type A T cell, which recognizes HEL, the 48–61 peptide, as well as the 52–61 peptide, which is recognized less efficiently than 48–61. Compared with 48–61, 52–61 binds less well to I-A^k, has a lower k_off and a shorter persistence in the APC, and forms complexes sensitive to denaturation in SDS/PAGE, an indicator of their weak interaction with I-A^k (20, 21).

Figure 1.

Two subsets of type B T cells according to their reactivity to HEL 52–61. T cell hybridomas and APC C3F6 were incubated for 18 h in the presence of different doses of HEL protein (■), 48–61 (●), or 52–61 (○) peptides. The released IL-2 was measured by the [³H]thymidine incorporation of IL-2-dependent cell line CTLL-2. Representative responses are shown for 3A9, type A T cell (A), MLA11.2, type B T cell (B), and DAV21, type B T cell (C).

Two representative type B T cells are shown in Fig. 1 B and C. The recognition of HEL is weak for both despite the extensive presentation of the chemically dominant 48–62/63 family from processed HEL (3, 22). The poor response to HEL is the criteria for classifying them as type B. Also note the difference in recognition between the two type B cells: MLA11.2, recognized 48–61 better than 52–61, whereas DAV21 recognized the shorter peptide 52–61 100-fold better than 48–61. At half-maximal responses, DAV21 needs more than 50 μM HEL, 0.5 μM of 48–61 peptide, but only 0.005 μM of the short peptide 52–61.

Among the 18 type B hybridomas examined, nine recognized the short peptide 52–61 at a 50- to 100-fold lower concentration when compared with 48–61 and are referred to as type B short. The remaining type B cells preferentially recognized the longer 48–61 peptide and are referred to as type B long. The 10 type A T cells and type B long T cells all recognized 48–61 better than 52–61 (Table 1). The type B state does not represent a transitory state in the binding of the peptide to I-A^k molecules, because APCs pulsed with peptides showed comparable responses by type A and type B T cells as a function of time (data not shown).

APCs Do Not Degrade 48–61 to 52–61 Intracellularly.

An explanation for our results with the type B short T cells is that 48–61 is recognized only when it is trimmed down to 52–61 by APCs. If this were the case, we would predict that the response of type B short T cells to 48–61 would be markedly impaired after fixation of APCs when compared with live APCs. Paraformaldehyde-fixed APCs present peptides on cell surface I-A^k molecules without intracellular antigen processing. As shown in Fig. 2A, fixed APCs did not present HEL to the type A T cells, because HEL needs to be transported to the deep endocytic compartments for processing (13, 23). Both fixed and live APCs presented 48–61 and 52–61 to type A and type B cells (Fig. 2 A and B). Like live APCs, fixed APCs presented 48–61 more efficiently than 52–61 to type A T cells. Also, like live APCs, 52–61 was better presented than 48–61 to the type B short T cells. Thus, for either type A or type B short T cells, their responses to 48–61 when presented by live and fixed APCs had similar pattern.

Figure 2.

HEL 48–61 is not degraded to 52–61 by APC. (A and B) C3F6 cells were prefixed by 1% paraformaldehyde. Fixed (1 × 10⁵) or live (5 × 10⁴) C3F6 cells and T cell hybridoma were incubated in the presence of HEL protein (squares), 48–61 (circles), or 52–61 (triangles) peptides. Filled symbols, fixed C3F6; open symbols, live C3F6. After 18 h, the released IL-2 was measured as described in Fig. 1. (C and D) C3F6 cells (2 × 10⁵ per well) were treated with or without 0.5 mM chloroquine at 37°C. After 20 min, different doses of HEL protein (squares) or 48–61 peptide (circles) were added and the incubation continued for another 4 h in the presence of chloroquine. Then, cells were washed and fixed by 1% paraformaldehyde. After three washes, T cells were added and the T cell assays were done as described in Fig. 1. Filled symbols, chloroquine-treated C3F6; open symbols, control C3F6 that not treated with chloroquine.

In further manipulations, APCs were treated with chloroquine to block vesicular acidification and thus reduce intracellular catabolism. As expected, chloroquine-treated APCs did not present HEL, but presented 48–61 peptide to type A T cells (Fig. 2C). Again, the response of type B short T cells to 48–61 presented by chloroquine-treated APCs was about the same as by the untreated APCs (Fig. 2D). Had 48–61 been trimmed to 52–61 in the live APCs, then the expectation was that in the presence of this drug the response of type B short T cells to 48–61 would decrease. This clearly was not the case; in fact, the opposite was found. Therefore, based on these results, we conclude that the 48–61 peptide is not degraded intracellularly to 52–61 by the APCs before recognition by T cells. Thus, exogenous 48–61 may adopt a different conformation in binding to I-A^k, which is recognized by type B short cells. This conformation is best mimicked by 52–61 interacting with I-A^k.

Type B T Cells Display Distinct TCR Contact Patterns.

The crystal structure of I-A^k complexed to HEL 46–62 revealed that the residues at P2, P5, and P8, which are Tyr-53, Leu-56, and Asn-59, respectively, were solvent exposed and could serve as T cell contact residues (4). Modification of the side chains of any of the three contact residues eliminates the response by type A T cells (ref. 24, see Table 1 and Fig. 3A). Most conventional CD4 T cells also interact with three TCR contact residues in MHC-bound peptides spatially distributed akin to 48–61 (14, 24–30). Therefore, we tested a panel of alanine substitutions of HEL 48–61 to compare the reactivities of the various type A and type B T cells (Fig. 3, summarized in Table 1).

Figure 3.

Distinct TCR contact patterns of type B T cells. The TCR contact residues Y53, L56, and N59 of the wild-type HEL 48–61 peptide were individually substituted with alanine. Peptides 48–61 (●), 48–61Y53A (□), L56A (▴), or N59A (○) were added to C3F6 cells, and the T cell assays were done as described in Fig. 1. The representative type A, type B long, and type B-short T cells are shown in A, B, and C, respectively.

None of the 28 T cells responded to an Ala-56 substitution, demonstrating the requirement of the centrally located Leu-56 as an obligate TCR contact residue for the 52–60-responsive I-A^k-restricted T cells. (One type A T cell, 2H.1D6, responded to the Ala-56 substitution but at about 100-fold higher concentration of peptide.) We emphasize that: first, extensive binding analysis on the Ala-substituted amino acids from 48–61 indicated that the major side chain that influenced the binding to I-A^k was Asp-52, which the crystal structure indicated to be at the P1 pocket (4, 31); and second, that the only register of the 48–63 peptide that is selected through the natural processing of HEL comprised the 52–60 segment in which the P1 pocket is occupied by Asp-52 and the three TCR contact residues mentioned above are solvent exposed. Other registers were not selected because of the unfavorable interactions of side chains that occupied the auxiliary pockets, i.e., P4, P6, P7, and P9 (15). Thus, the finding that Leu-56 is an obligatory residue confirms that the appropriate register for 52–60 has Asp-52 positioned at the P1 pocket.

Examination of Tyr-53 and Asn-59, the two TCR contact residues positioned at the ends of the core segment, confirmed that the majority of type A T cells established functional contact with them. Strikingly, this was not the case with any of the type B T cells, which interacted with one or the other, but not both. All of the type B long T cells examined recognized Asn-59 because Ala substitution of this residue abolished the response. Yet Ala substitution of Tyr-53 had no effect on their response. The reverse effect was found for all type B short T cells, which recognized Tyr-53 but not Asn-59.

We also examined the response of the various subsets to substitutions of the main MHC anchor residues. The results of these experiments did not establish any definite pattern. As we reported previously, substitution of Asp-52 for Ala impaired binding (31) and, as expected, had an effect on the response of most clones. Substitutions of the other residues showed minor effects and are not presented.

The interaction of the T cells with peptides truncated at the C terminus was examined. Fig. 4 shows that the type B short T cells also responded to a 7-mer peptide (HEL 52–58), which should bind from P1-P7. This response was as strong as the response to 52–61 and stronger than to 48–61. HEL 52–58 binds I-A^k with an IC₅₀ of 1.76 μM, which is about 8-fold less than 52–61. It shows a considerably shorter dissociation time (the half-time dissociation was 1 h, compared with 24 h for 52–61, and more than 96 h for 48–61). There are a few reports that peptides with 6–8 amino acids in length can bind to the MHC class II molecules (25, 32–35). In most of these cases, the shorter peptides stimulated T cells less efficiently when compared with the longer peptides. However, our study shows that a 7-mer peptide (HEL 52–58) stimulated the type B short set of T cells at a lower concentration than the longer peptide HEL 48–61. Finally, the responses of type B long and type A T cells to C-terminal-truncated peptides indicated that the type B long could recognize up to residue 59. However, the type A recognized the full-length peptide: they did not respond or responded poorly to 48–60 (Table 1).

Figure 4.

Type B short T cells respond efficiently to the truncated peptide HEL 52–58. Shown are the responses of two type B short T cells to peptides HEL 48–61 (●), 48–58 (■), 52–61 (○), or 52–58 (□) presented by C3F6 cells.

In summary, there is a very striking and major difference between the type A and the two subsets of type B T cells in their interaction with the TCR contact residues. The type A T cells recognized all three of the exposed TCR contact residues, whereas the type B long or B short T cells did not require either Tyr-53 or Asn-59, respectively. Furthermore, type B short T cells responded to both N- and C-termini peptide truncations, and to a peptide as short as a 7-mer. Thus, type A T cells recognize the entire peptide from P1-P10, i.e., 52–61, whereas the type B long T cells recognize the core peptide around P5-P8, 56–59, and the type B short T cells recognize the core peptide around P2-P5, 53–56.

APCs That Lack HLA-DM Can Present Peptide to Type B T Cells.

Our interpretation for the development of type A versus type B cells is based on the biology of HEL peptide processing and the interaction of peptide with I-A^k molecules: (i) HEL must be reduced in a lysosomal-like vesicle (23, 36); (ii) HEL must be assembled in a late vesicular compartment that receives nascent class II molecules bearing the invariant chain; and (iii) the processing of HEL requires the catalytic function of H-2DM (13, 37, 38). Fig. 5 demonstrates that an APC that expresses human HLA-DM (T1-A^k) presented HEL to type A T cells adequately, whereas the HLA-DM-deficient line T2-A^k presented HEL epitopes very poorly. (The murine I-A^k can interact with HLA-DM without any apparent problems. In unpublished experiments with Shirley Petzold, the amounts of the 52–60 family of peptides selected from HEL by I-A^k molecules was identical in C3.F6 and in T1-A^k cells.)

Figure 5.

APCs lacking HLA-DM can present peptide to type B T cells. T1A^k cell line and the corresponding HLA-DM-deficient cell line T2A^k were used as APCs. T cell hybridoma (1 × 10⁵) and T1A^k or T2A^k cell (5 × 10⁴) were incubated at 37°C in the presence of HEL protein (squares) or HEL 48–61 peptide (circles). After 18 h, the released IL-2 was measured by proliferation of the CTLL-2 line. Filled symbols, T1A^k cell; open symbols, T2A^k cell.

The results with peptide are in striking contrast. These are bound to I-A^k molecules in light vesicular compartments and/or plasma membrane using class II molecules that are capable of recycling peptides. Through biochemical analysis, we have found that when APCs are offered unfolded HEL or 48–61 peptide, the I-A^k molecules presented 48–61 independently of HLA-DM expression (13). Fig. 5 also shows that 48–61 offered to either T1-A^k or T2-A^k was presented to type A or B T cells.

Discussion

Our view is that the peptides generated from HEL are processed in a deep compartment at highly acidic pH and are bound to class II molecules, influenced by the catalytic function of H-2DM. In such an environment, H-2DM edits and selects for a very fixed conformer, and the flexibility to generate any other is limited. In contrast, in a recycling vesicle or on the plasma membrane (13), the exogenous peptides must bind by peptide exchange, at a higher pH and without the participation of H-2DM. The fact that exogenous peptides exchange with I-A^k molecules indicates that these I-A^k molecules are bound by peptides with a fast k_off, enabling the exchange to take place. Weak-binding peptides have side chains that are unfavorable and are likely to partially occupy or distort one of the pocket sites. Thus, the I-A^k at the site where exogenous peptides are loaded may have more flexibility and be less constrained, allowing peptides to adopt more than one conformational state. We do not know at this time whether the B conformers that gave rise to either B short or B long T cells represent fixed conformers or a dynamic state.

Regardless of how the peptide-MHC complex generates a type B reactivity, its uniqueness is striking, as evidenced by the distinct manner in which it responds to the 52–60 epitope. We discussed that all T cells studied so far interact with the three major TCR contact and so do our conventional type A T cells. However, type B do not follow such rules. It is important to understand that type B T cells are not a minor component of the T cell response; in the case of the 52–60 family of peptides, they represent a substantial contribution to the CD4 T cell repertoire. In normal mice about 30–50% of T cells from immunization with 48–61 are type B (8). Importantly in HEL transgenic mice, all are type B [i.e., the type A reactivity is 100% negatively selected (8)]. The type B T cells are heterogenous with respect to the Vα and Vβ gene segment utilization (Table 1), demonstrating that they do not use a restricted TCR repertoire. These findings are consistent with the notion that the type B T cells exist at significant frequencies within the naïve T cell repertoire and are not the product of the expansion of rare clones. Also, they are not restricted to special peptides. We have found type B T cells to all HEL epitopes examined thus far and to autologous peptides as well (data not shown). Finally, we believe that type B T cells can explain the findings of other researchers, including the reports of “cryptic” epitopes (39).

How are the type B T cells generated? It has been reported that inflammatory cytokines such as TNF-α, IL-1β, and IL-6, as well as others, promoted the proteolytic activity in dendritic cells and macrophages (40–42). Thus, at sites of inflammation and autoimmune reactions, proteolysis may lead to the generation of exogenous peptides, perhaps of a small size. The type B T cell population would be very sensitive to the exogenously generated peptides and could play a role in mediating the immune response to inflammation and autoimmunity. Our previous results that type B T cells against a mouse self-epitope could escape negative selection in the thymus (8) supports the idea that type B T cells may be involved in the autoimmune response, under conditions of tissue inflammation. Additionally, the mode by which type B cells are generated defines a unique pathway in presentation involving exogenous peptides that does not use the H-2 DM molecules.

Abbreviations

HEL

hen-egg white lysozyme

TCR

T cell receptor

APC

antigen-presenting cell