Fine epitope mapping within the pathogenic thyroglobulin peptide 2340–2359: minimal epitopes retaining antigenicity across various MHC haplotypes are not necessarily immunogenic

Abstract

We have previously reported that the 20-mer peptide p2340 (amino acids 2340–2359), of human thyroglobulin (Tg) has the unique feature that it causes experimental autoimmune thyroiditis (EAT) in mouse strains bearing high-responder (HR) or low-responder (LR) MHC haplotypes in Tg-induced EAT. In this study, we have employed fine epitope mapping to examine whether this property of p2340 is the result of recognition of distinct or shared minimal T-cell epitopes in the context of HR or LR MHC class II molecules. Use of overlapping peptides showed that a core minimal 9-mer epitope (LTWVQTHIR, amino acids 2344–2352) was recognized by p2340-primed T cells from both HR (H2^k,s) and LR (H2^b,d) strains, whereas a second 9-mer epitope (HIRGFGGDP, amino acids 2350–2358) was antigenic only in H2^s hosts. Truncation analysis of LTWVQTHIR and HIRGFGGDP peptides delineated them as the minimal epitopes recognized by p2340-primed T cells from the above strains. Subcutaneous challenge of all mouse strains with the 9-mer core peptide LTWVQTHIR in adjuvant elicited specific lymph node cell proliferative responses and mild EAT only in HR hosts, highlighting this sequence as a minimal pathogenic Tg peptide in EAT. The 9-mer peptide HIRGFGGDP was not found to be immunogenic in H2^s hosts. These data demonstrate that minimal T-cell epitopes, defined as autoantigenic in hosts of various MHC haplotypes, are not intrinsically immunogenic. Activation of naive autoreactive T cells may require contributions from flanking residues within longer peptide sequences encompassing these epitopes.

Introduction

Experimental autoimmune thyroiditis (EAT) develops in mice following administration of thyroglobulin (Tg) in adjuvant and exhibits many of the characteristics of Hashimoto’s thyroiditis in humans: mononuclear cell infiltration and destruction of the thyroid follicles, autoantibody production and T-cell proliferative response to Tg.¹ Incidence and severity of EAT and anti-Tg antibody titres differ among inbred mouse strains and are under the influence of MHC genes, separating strains into high-responder (HR, susceptible) and low-responder (LR, resistant) haplotypes.^2,3 Non-MHC genes, such as the Tg gene, may also be responsible for differential EAT development^4,5 but their role is secondary and few data are available concerning their nature and mechanism of action.

Over the last 20 years, Tg T-cell epitopes have been mapped and used as model antigens in the study of immunoregulation of EAT. Mapping efforts used mostly HR strains (particularly H2k), and among the 35 Tg peptides examined, 25 peptides were administered in HR strains,^6–14 six peptides in LR strains¹⁰ and three peptides in both HR and LR strains.^15,16 The outcome of these studies was the mapping of 14 Tg peptides that induce EAT in HR strains,^{7–9,11,14–17} whereas no Tg peptide was found to cause thyroiditis exclusively in LR mice. In this regard, we have previously reported that the 20-mer human Tg peptide p2340 [amino acids (aa) 2340–2359], exhibiting 80% identity with its mouse homologue, has the unique feature that it causes EAT both in HR (H2k,s) and LR (H2b,d) hosts and additionally in DR3-transgenic mice.^18–20 In this study, we proceeded to map the minimal T-cell epitope(s) within the p2340 sequence in order to investigate whether the immunopathogenic properties of this sequence are the result of recognition of one or more distinct epitopes presented in the context of HR or LR MHC class II genes. Such knowledge would expand the study of Tg-peptide-induced EAT to genetic backgrounds (e.g. B6 mice) that harbour non-MHC alleles and are useful for immunoregulation studies.

Materials and methods

Animals and antigens

Six- to eight-week-old CBA/J (H2k), SJL/J (H2s), BALB/c (H2d) and C57BL/6 (H2b) female mice were bred in the animal facility of the Hellenic Pasteur Institute and treated following the European requirements on animal experimentation. All procedures were approved by the Prefecture of Athens, Department of Veterinary Applications and Animal Protection. The human Tg peptide p2340 QVAALTWVQTHIRGFGGDPR and mapping peptides: QVAALTWVQ, AALTWVQTH, ALTWVQTHI, LTWVQTHIR, LTWVQTHI, TWVQTHIR, TWVQTHIRG, WVQTHIRGF, QTHIRGFGG, THIRGFGGD, HIRGFGGDP, IRGFGGDPR, IRGFGGDP, RGFGGDPR were synthesized by JPT Peptide Technologies GmbH (Berlin, Germany). The human Tg peptide PVEFSRKVPTFATPWPDFVP (p2652) (2652–2671), by Genosys Biotechnologies (Cambridge, UK) was used as control. All peptides were blocked with an N-terminal acetyl group and a C-terminal amide group, and used in experiments at > 80% purity. The amino acid coordinates of peptides have been assigned according to the human Tg sequence data by Van de Graaf et al.²¹

T-cell proliferation assays

Lymph node cell (LNC) proliferation assays were performed as previously described.²² Briefly, mice were immunized subcutaneously at the tail base and on the back with 100 nmol intact p2340 or 100 or 200 nmol of minimal epitopes (LTWVQTHIR) and (HIRGFGGDP) emulsified in complete Freund’s adjuvant (CFA; Sigma, Taufkirchen, Germany). Ten days later, inguinal and axillary lymph nodes (pooled from three mice) were collected aseptically and single cell suspensions were prepared in Dulbecco’s modified Eagle’s minimal essential medium (Biochrom, Cambridge, UK) which was supplemented with 10% heat-inactivated fetal bovine serum (Gibco, Scotland, UK), 20 mm HEPES buffer (Biochrom), 2 mm l-glutamine, 0·5 mm sodium pyruvate, 0·2% sodium bicarbonate (Gibco), 5 × 10⁵ IU streptomycin (KOPER, Athens, Greece), 5 × 10⁵ IU penicillin (KOPER), 5 × 10⁻⁵ mβ-mercaptoethanol (Sigma). After centrifugation and washing, cells (4 × 10⁵ cells per well, in triplicate wells) were cultured in the presence or absence of antigen, in flat-bottomed 96-well plates (Orange Scientific, Braine-l’Allend, Belgium) and incubated for 3 days at 37°C in a 5% CO₂, 95% air humidified incubator. During the last 18 hr of culture, 1 μCi [³H]thymidine (25 Ci/mmol) (Amersham Pharmacia, Stockholm, Sweden) was added to each well. The cells were harvested using a semiautomatic harvester (Skatron Inc, Bath, UK) and incorporated radioactivity was measured using the 1450 microbeta Trilux counter (Wallac, Turku, Finland). Stimulation index (S.I.) was defined as: counts per minute in the presence of peptide/ counts per minute in the absence of peptide.

Induction of EAT

Mice were challenged (subcutaneously) at the base of the tail with 100 nmol (LTWVQTHIR) in CFA emulsion and boosted 21 days later with 50 nmol peptide in incomplete Freund’s adjuvant (Sigma). Thirty-five days after the initial challenge, the thyroids were removed for EAT assessment. Thyroid glands were fixed in 10% formalin buffer overnight at room temperature. They were then dehydrated in ethanol solutions with increasing concentrations, cleaned in xylene (Surgipath, Richmond, UK) and finally embedded in paraffin (Paraplast, Tyco Healthcare Group LP, Mansfield, MA) using a fluid-transfer tissue processor (Leica, Solms, Germany). Serial 4-μm sections were cut with the microtome RM2135 (Leica) and stained with haematoxylin Harris and eosin (BDH, Poole, UK). The extent of thyroid mononuclear infiltration was determined by histological evaluation of 200 sections throughout both thyroid lobes. Mononuclear cell infiltration index (I.I.) was scored as follows: 0 = no infiltration, 1 = interstitial accumulation of cells between two or three follicles, 2 = one or two foci of cells at least the size of one follicle, 3 = extensive infiltration, 10–40% of total area, 4 = extensive infiltration, 40–80% of total area and 5 = extensive infiltration > 80% of total area.

Results

The 9-mer epitope (aa 2344–2352) is antigenic in both HR and LR strains

Six nonamer and one octamer peptides spanning the p2340 sequence, and overlapping by seven amino acids, were tested in vitro for their capacity to induce proliferation of LNC from p2340-primed CBA/J, SJL/J, BALB/c and C57BL/6 mice. Dose–response curves were obtained for each peptide within the 0·6–40 μm range and an arbitrary cut-off was assigned so that a given peptide was considered to be ‘stimulatory’ when it could consistently induce LNC proliferation (S.I. ≥ 3·0) at 2·5 μm in three independent experiments. This analysis clearly showed that only the core nonamer peptide LTWVQTHIR (aa 2344–2352) stimulated a potent recall response in p2340-primed T cells both in HR CBA/J, SJL/J strains (S.I. values 10·0 ± 0·8 and 6·0 ± 0·7, respectively) (truncated peptide 3, Table 1) and LR BALB/c, C57BL/6 strains (S.I. values 3·6 ± 0·3 and 22·0 ± 0·3, respectively) (truncated peptide 3, Table 1). A second nonamer peptide (HIRGFGGDP, aa 2350–2358) induced significant proliferation (S.I. = 8·5 ± 0·9) only in LNC from SJL/J mice (truncated peptide 6, Table 1). LNC from all strains responded strongly to p2340 (S.I. values 13·0–42·0), confirming earlier results¹⁹ as well as the efficacy of the immunization protocol. Background responses were observed against the control peptide p2652. These data delineated the LTWVQTHIR sequence as a Tg epitope targeted by T cells from both HR and LR strains that develop p2340-induced EAT.

Table 1.

T-cell epitope mapping within the p2340 sequence

																						S.I. (mean ± SD)1 of p2340-primed LNC2 from
																						CBA/J	SJL/J	BALB/c	C57BL/6
Amino acid coordinates3 p2340	23-	40	41	42	43	44	45	46	47	48	49	50	51	52	53	54	55	56	57	58	59
		Q	V	A	A	L	T	W	V	Q	T	H	I	R	G	F	G	G	D	P	R	13·0 ± 0·84	38·0 ± 0·14	42·0 ± 0·34	31·0 ± 1·04
Truncated peptides
1		Q	V	A	A	L	T	W	V	Q												1·6 ± 0·6	0·7 ± 0·3	2·3 ± 0·3	1·2 ± 0·8
2				A	A	L	T	W	V	Q	T	H										1·9 ± 0·9	2·4 ± 0·5	1·2 ± 0·5	1·2 ± 0·8
3						L	T	W	V	Q	T	H	I	R								10·0 ± 0·84	6·0 ± 0·74	3·6 ± 0·34	22·0 ± 0·34
4								W	V	Q	T	H	I	R	G	F						1·2 ± 1·0	1·5 ± 0·4	1·6 ± 0·9	2·4 ± 0·3
5										Q	T	H	I	R	G	F	G	G				1·1 ± 0·5	1·0 ± 0·6	0·8 ± 0·6	1·1 ± 0·7
6												H	I	R	G	F	G	G	D	P		1·4 ± 0·6	8·5 ± 0·94	0·7 ± 0·8	1·0 ± 0·5
7														R	G	F	G	G	D	P	R	1·3 ± 0·7	0·9 ± 1·0	0·8 ± 0·7	1·3 ± 0·3
Amino acid coordinates3	26-	52	53	54	55	56	57	58	59	60	61	62	63	64	65	66	67	68	69	70	71
Control peptide		P	V	E	F	S	R	K	V	P	T	F	A	T	P	W	P	D	F	V	P	1·2 ± 0·9	1·5 ± 0·3	1·8 ± 0·4	1·1 ± 0·2

¹Stimulation index (S.I.) [mean ± standard deviation (SD) values of triplicate wells] of lymph node cells responding in vitro to 2·5 μm of the peptides shown. Values are extrapolated from titration curves obtained within the 0·6–40 μm range. Background counts per minute were: 2739 (SJL/J), 6731 (CBA/J), 1654 (BALB/c) and 1231 (C57BL/6). Results are representative of three separate experiments.

²Lymph node cells were obtained from mice challenged with 100 nmol p2340 in complete Freund’s adjuvant at the base of the tail and on the back, 10 days earlier.

³Amino acid coordinates are assigned according to the thyroglobulin sequence data by Van de Graaf et al.²¹ and do not include the 20 amino acid leader sequence.

⁴S.I. ≥ 3 in the presence of 2·5 μm of peptides is the cut-off for the characterization of stimulatory peptides.

Peptide (aa 2344–2352) is a minimal T-cell epitope within p2340

To examine whether the LTWVQTHIR sequence (aa 2344–2352) comprised a minimal epitope, we performed truncation analysis to delineate its boundaries. Mice were primed again with p2340 and 10 days later, their inguinal and axillary LNC were tested against peptides truncated by one amino acid at the N-terminal or C-terminal end of the LTWVQTHIR sequence. As previously, consistent significant stimulatory activity (S.I. ≥ 3·0) at 2·5 μm–extrapolated from titration curves in the 0·6–40 μm range –in at least three independent experiments was used as a criterion of antigenicity. Removal of the C-terminal R2352 residue abrogated the antigenicity of the epitope in all strains tested (truncated peptide 1 versus 2, Table 2). Removal of the N-terminal L2344 residue similarly abolished antigenicity in CBA/J, SJL/J and BALB/c mice but, surprisingly, the octamer TWVQTHIR peptide elicited responses in cultures of C57BL/6 LNC (S.I. value 9·0 ± 2·0, truncated peptide 3, Table 2). These results demonstrated that the LTWVQTHIR peptide retains antigenicity in all mouse strains tested only when the N-terminal L2344 and C-terminal R2352 residues are integral parts of this minimal epitope.

Table 2.

Truncation analysis of the LTWVQTHIR sequence

																						S.I. (mean ± SD)1 of p2340-primed LNC2 from
																						CBA/J	SJL/J	BALB/c	C57BL/6
Amino acid coordinates3	23-	40	41	42	43	44	45	46	47	48	49	50	51	52	53	54	55	56	57	58	59
p2340		Q	V	A	A	L	T	W	V	Q	T	H	I	R	G	F	G	G	D	P	R	28·0 ± 0·04	72·0 ± 0·44	35·0 ± 0·84	24·0 ± 0·14
Truncated peptides
1						L	T	W	V	Q	T	H	I	R								14·7 ± 0·44	3·0 ± 0·94	3·3 ± 0·34	26·6 ± 1·04
2						L	T	W	V	Q	T	H	I									2·0 ± 0·2	0·9 ± 0·0	0·9 ± 0·1	0·7 ± 0·2
3							T	W	V	Q	T	H	I	R								1·3 ± 0·1	1·6 ± 0·2	1·2 ± 0·4	9·0 ± 2·04
Amino acid coordinates3	26-	52	53	54	55	56	57	58	59	60	61	62	63	64	65	66	67	68	69	70	71
Control peptide		P	V	E	F	S	R	K	V	P	T	F	A	T	P	W	P	D	F	V	P	0·4 ± 0·2	0·5 ± 0·1	1·0 ± 0·1	0·5 ± 0·2

¹Stimulation index (S.I.) [mean ± standard deviation (SD) values of triplicate wells] of lymph node cells responding in vitro to 2·5 μm of the peptides shown. Values are extrapolated from titration curves obtained within the 0·6–40 μm range. Background counts per minute were: SJL/J, 5537; CBA/J, 2979; BALB/c, 4661 and C57BL/6, 1192. Results are representative of three separate experiments.

²Lymph node cells were obtained from mice challenged with 100 nmol p2340 in complete Freund’s adjuvant at the base of the tail and on the back, 10 days earlier.

³Amino acid coordinates are assigned according to the thyroglobulin sequence data of Van de Graaf et al.²¹ and do not include the 20 amino acid leader sequence.

⁴S.I. ≥ 3 in the presence of 2·5 μm of peptides is the cut-off for the characterization of stimulatory peptides.

The 9-mer minimal epitope (aa 2344–2352) is immunogenic and pathogenic in HR strains

To examine the immunogenic properties of the nonamer LTWVQTHIR peptide (aa 2344–2352), mice from CBA/J, SJL/J, BALB/c and C57BL/6 strains were challenged subcutaneously with 100 nmol of this peptide. Ten days later, the proliferative responses of their LNC were examined in the presence of varying concentrations (0·3–20 μm) of the immunizing peptide. LTWVQTHIR at 20 μm induced significant and specific proliferation in LNC from CBA/J mice (S.I. = 4·0 ± 0·5) and SJL/J mice (S.I. = 5·0 ± 0·1), whereas background responses were observed from LNC in both LR BALB/c mice (S.I. = 1·8 ± 0·2) and C57BL/6 mice (S.I. = 1·8 ± 0·2) (Fig. 1). The lack of immunogenicity of LTWVQTHIR in these LR strains was not the result of a suboptimal immunizing dose because priming even with 200 nmol peptide failed to stimulate a response in these mice (data not shown). Direct challenge of CBA/J (n = 8) and SJL/J (n = 6) with the LTWVQTHIR peptide led to the induction of mild EAT in both strains (I.I. = 0·5 ± 0·5 and 0·5 ± 0·8 for CBA/J and SJL/J mice, respectively). In terms of disease severity, focal mononuclear cell infiltration of the thyroid was seen in four out of eight CBA/J mice and one out of six SJL/J mice and more severe infiltration in one out of six SJL/J mice (Figs. 2 and 3). These data demonstrated that the LTWVQTHIR peptide was both immunogenic and pathogenic only in HR mice.

Figure 1.

Immunogenicity of LTWVQTHIR peptide in CBA/J, SJL/J, BALB/c and C57BL/6 mouse strains. Mice were immunized with 100 nmol peptide and 10 days later, lymph node cell (LNC) responses [mean stimulation index (S.I.) of triplicate samples] against the immunizing and control peptide (0·3–20 μm) were obtained. Background counts per minute were: 9901 (CBA/J), 4000 (SJL/J), 7273 (BALB/c) and 1859 (C57BL/6). Error bars represent standard deviations of counts per minute means from triplicate wells. Results are representative of three independent experiments.

Figure 2.

Pathogenicity of LTWVQTHIR peptide in high-responder mouse strains. SJL/J and CBA/J mice were immunized with 100 nmol peptide and 3 weeks later received a boosting dose of 50 nmol. Experimental autoimmune thyroiditis was assessed 5 weeks after the initial challenge and sections of the thyroid glands were scored for infiltration index (I.I.) as described in the Materials and methods. Normal gland: I.I. = 0, interstitial accumulation of inflammatory cells: I.I. = 1, one or two foci of inflammatory cells: I.I. = 2.

Figure 3.

Histological appearance of mononuclear cell infiltration in mouse thyroids after administration of the LTWVQTHIR peptide. Normal gland: infiltration index (I.I.) = 0 (a), interstitial accumulation of inflammatory cells: I.I. = 1 (b), one or two foci of inflammatory cells: I.I. = 2 (c). Magnification: × 125.

The 9-mer peptide (aa 2350–2358) comprises a minimal antigenic but not immunogenic T-cell epitope

The minimal epitope included in the HIRGFGGDP sequence (aa 2350–2358) (peptide recognized only by SJL/J T cells) was determined by truncation analysis. The LNC from SJL/J mice immunized with p2340 were cultured in vitro in the presence of serial dilutions (0·6–40 μm) of peptides truncated by one amino acid at the N-terminal or C-terminal end of the HIRGFGGDP sequence. The characterization of stimulatory peptides was based on the same arbitrary cut-off as previously [consistent stimulatory activity (S.I. ≥ 3·0) at 2·5 μm]. Removal of either the N-terminal H2350 residue or the C-terminal P2358 residue abrogated the antigenicity of the epitope (truncated peptides 2 and 3 versus 1, Table 3) and defined the boundaries of the HIRGFGGDP epitope at these residues. To investigate the immunogenicity of this minimal epitope, SJL/J mice were challenged subcutaneously with 100 or 200 nmol peptide and 10 days later, LNC were collected and cultured in the presence of varying concentrations (0·3–20 μm) of the immunizing peptide. Surprisingly, no response was detected against HIRGFGGDP regardless of the priming dose used (data not shown). These findings suggested that although HIRGFGGDP is clearly antigenic, it could not elicit a response among naive T cells in SJL/J hosts.

Table 3.

Truncation analysis of the HIRGFGGDP sequence

																						S.I. (mean±SD)1 of SJL/J p2340-primed LNC2 Peptide concentration
																						40 μm	10 μm	2·5 μm	0·6 μm
Amino acid coordinates3	23-	40	41	42	43	44	45	46	47	48	49	50	51	52	53	54	55	56	57	58	59
p2340		Q	V	A	A	L	T	W	V	Q	T	H	I	R	G	F	G	G	D	P	R4	30·0 ± 0·2	28·0 ± 0·3	28·0 ± 0·4	22·0 ± 1·0
Truncated peptides
1												H	I	R	G	F	G	G	D	P4		7·4 ± 0·4	8·6 ± 0·5	8·3 ± 0·3	5·5 ± 0·6
2												H	I	R	G	F	G	G	D			2·6 ± 0·6	2·0 ± 0·5	1·7 ± 0·3	1·2 ± 0·6
3													I	R	G	F	G	G	D	P		0·9 ± 0·2	1·3 ± 0·1	1·3 ± 0·0	1·1 ± 0·0
Amino acid coordinates3	26-	52	53	54	55	56	57	58	59	60	61	62	63	64	65	66	67	68	69	70	71
Control peptide		P	V	E	F	S	R	K	V	P	T	F	A	T	P	W	P	D	F	V	P	1·1 ± 0·2	1·0 ± 0·1	0·5 ± 0·1	0·8 ± 0·0

¹Mean stimulation index (S.I.) and standard deviation (SD) values of triplicate wells of lymph node cells responding in vitro to titration curves within the 0·6–40 μm range of the peptides shown. Background counts per minute were: 5537. Results are representative of three separate experiments.

²Lymph node cells were obtained from mice challenged with 100 nmol p2340 in complete Freund’s adjuvant, 10 days earlier.

³Amino acid coordinates are assigned according to the thyroglobulin sequence data by Van de Graaf et al.²¹ and do not include the 20 amino acid leader sequence.

⁴S.I. ≥ 3 in the presence of 2·5 μm of peptides is the cut-off for the characterization of stimulatory peptides.

Lack of synergistic effect between LNC responding to the minimal epitopes (aa 2344–2352) and (aa 2350–2358)

To investigate a potential synergistic effect between LNC simultaneously responding to the LTWVQTHIR (aa 2344–2352) and HIRGFGGDP (aa 2350–2358) epitopes, both peptides were added at equimolar concentrations (0·6–40 μm) in cultures of LNC from p2340-primed SJL/J mice. Proliferation of LNC was only slightly increased when both peptides were present in culture versus the presence of each peptide alone (S.I. = 5·5 ± 0·2 versus 4·1 ± 0·4 and 3·1 ± 0·1, respectively, at 2·5 μm) (Fig. 4). Nevertheless, neither peptide mediated recall responses analogous to those seen against the intact p2340 (S.I. = 9·0 ± 0·4 at 2·5 μm) (Fig. 4). These data did not provide supporting evidence for synergistic effects between p2340-primed T cells responding in vitro to the LTWVQTHIR and HIRGFGGDP epitopes.

Figure 4.

Proliferation (mean stimulation index of triplicate samples) of p2340-primed lymph node cells (LNC) from SJL/J mice in the presence of both LTWVQTHIR and HIRGFGGDP peptides (LTWVQTHIR + HIRGFGGDP) compared with the effect of each individual peptide (LTWVQTHIR, HIRGFGGDP) or of the intact peptide p2340 (0·6–20 μm, each). Background counts per minute were 5573. Error bars represent standard deviations of counts per minute means from triplicate wells. Data shown are representative results from three independent experiments.

Discussion

In the field of Tg-induced mouse EAT, mapping of pathogenic T-cell epitopes has mainly relied on the use of algorithmic approaches searching for candidate MHC-binding peptides within the large Tg sequence.^{8–10,18,23–25} Most studies have employed the use of HR MHC haplotypes, and in particular H2k, resulting in the discovery of several (14 out of 29 tested) A^k-binding or E^k-binding pathogenic epitopes, two of which are known to bind also to A^s molecules.^{6–9,14,16,17,24} This has left open the question of whether the Tg-induced EAT observed in mice with LR MHC haplotypes involves recognition of some of the already characterized epitopes that can ‘promiscuously’ bind to various MHC or recognition of distinct Tg peptides that comprise a non-overlapping set. The answer to this question may afford a perspective that is currently lacking, namely the knowledge of the relative value of the known Tg epitopic map in genetically diverse strains and the implications that this might have when extrapolated to an outbred population as occurs with the human disease.

To further explore this issue we have used, in this study, the 20-mer human Tg peptide p2340 as a model antigen because of its unique feature, among the known pathogenic Tg peptides, to induce EAT in both HR and LR strains.¹⁹ We sought to determine whether p2340 pathogenicity can be attributed to recognition of shared or distinct minimal T-cell epitopes in HR and LR hosts. Using p2340-primed LNC as a tool and antigenicity as a criterion, we found via overlapping peptides and truncation analysis that the core 9-mer LTWVQTHIR (aa 2344–2352) is a minimal T-cell epitope recognized by effector cells in both HR (CBA/J, SJL/J) and LR (BALB/c, C57BL/6) mouse strains. This finding provides support for the notion that, in mouse strains of diverse genetic backgrounds, EAT may be mediated by autoreactive T-cell clones recognizing the same Tg epitope(s) in the context of various MHC class II molecules. These data also extend previous observations that minimal Tg T-cell epitopes, such as p2496, can be presented even in the context of non-isotypic (E^k and A^s) molecules from different MHC haplotypes.²² If applied to human disease, such ‘promiscuous’ recognition of shared Tg T-cell epitopes may account in part for the lack of strong HLA associations with human disease.²⁶

When the immunopathogenicity of the core T-cell epitope LTWVQTHIR (aa 2344–2352) was examined, we observed that LNC proliferative responses and direct EAT were elicited in HR and not in LR strains. This observation suggests possible involvement of flanking residues for the activation of LR naive T cells specific for this minimal epitope. Amino acid residues flanking the minimal epitope may enhance the stability of MHC–peptide complexes and so extend their lifespan, prolonging T-cell stimulation. Such a role has been seen by other groups studying the binding of peptides from conalbumin (Ca)²⁷ and hen egg lysozyme (HEL52-61)²⁸ to A^k molecules or of Ea chain peptide (Ea52-68) to A^b molecules.²⁹ Flanking residues may also mediate direct binding to the T-cell receptor.²⁸ The indispensable presence of flanking residues for the activation of LR T cells has been shown in another study concerning the T-cell epitope mapping of a diabetes-associated peptide in the glutamic acid decarboxylase 65 (GAD65) molecule, p524–543. It was shown that a minimal T-cell epitope missing N-terminal flanking residues 528 and 529 was stimulatory in non-obese diabetic (NOD) mice but not in MHC-matched NOD-resistant mice (NOR), suggesting that a protective response in the resistant mice may require T-cell recognition of one or more of the N-terminal flanking residues.³⁰ These data suggest that HR and LR MHC molecules may present the same epitopes but another mechanism such as processing or trimming of T-cell epitopes may account for the protective responses in LR.

An additional mechanism that could account for the lack of immunogenicity of LTWVQTHIR in LR hosts might involve discrepancies in sequence homology. The mouse analogue of LTWVQTHIR possesses serine instead of threonine (conservative substitution) at position 2349 and glycine instead of arginine (non-conservative substitution) at position 2352. Similar considerations could be applied to the second minimal T-cell epitope within p2340 [HIRGFGGDP (aa 2350–2358)], which was recognized in vitro only by p2340-primed T lymphocytes from SJL/J mice but was not immunogenic in these mice. The mouse analogue of HIRGFGGDP possesses glycine instead of arginine (non-conservative substitution) at position 2351 and alanine instead of glycine (non-conservative substitution) at position 2352. These changes have no apparent effect on antigenicity but in the absence of stabilizing effects by flanking residues they may influence activation of peptide-specific naive T cells, which is known to be more stringent than that of effector cells. Indeed, another study concerning the polylipoprotein peptide PLP139–151, which induced experimental autoimmune encephalomyelitis, has shown that the introduction of some conservative and non-conservative amino acid substitutions at T-cell receptor and MHC contact residues significantly diminished the ability of peptide analogues to induce active disease, but these substitutions had little or no effect on the ability to activate PLP139–151-primed T cells for proliferation or disease transfer.³¹

Lastly, the magnitude of T-cell responses and the incidence and severity of disease induced by the minimal epitope LTWVQTHIR was different from that observed following challenge with p2340 (CBA/J: mean I.I. 0·5 with LTWVQTHIR and 0·3 with p2340, SJL/J: mean I.I. = 0·5 with LTWVQTHIR and mean I.I. = 1·8 with p2340).¹⁹ Apart from the stabilizing effects of flanking residues as well as discrepancies in sequence homologies that may be responsible also for these differences, it remains possible that the superior immunopathogenicity of p2340 in SJL/J mice, when compared with that of the mapped minimal epitopes, can be attributed to activation of cytotoxic T cells by HIRGFGGDP, which was not assessed in this study. This was the case for a peptide of mouse hepatitis virus glycoprotein S, which contained epitopes that did not induce proliferation of T-cell clones but promoted their lytic action.³² Moreover, it must be taken into account that p2340 induces the production of specific antibodies recognizing minimal B-cell epitopes LTWVQTHIR (aa 2344–2352) and RGFGGDPR (aa 2352–2359), whereas LTWVQTHIR does not (unpublished data).

To conclude, this report delineates the presence of a core minimal T-cell epitope LTWVQTHIR (aa 2344–2352) within the immunopathogenic peptide p2340, that is a target of proliferative T cells both in HR and LR hosts. The p2340-induced EAT in LR strains, such as C57BL/6 mice, provides further opportunities to investigate genetic influences on EAT via the use of recombinant inbred strains, H2 congenic strains or transgenic mice expressing proteins useful in the study of the immunoregulation of EAT.

Glossary

aa

amino acids

CFA

complete Freund’s adjuvant

cpm

counts per minute

EAT

experimental autoimmune thyroiditis

HR

high responder

I.I.

infiltration index

LNC

lymph node cell

LR

low responder

s.c.

subcutaneously

S.I.

stimulation index

SD

standard deviation

TCR

T-cell receptor

Tg

thyroglobulin

Disclosures

The authors have no competing interests.