Mapping thyroid peroxidase epitopes using recombinant protein fragments

Abstract

Tonacchera M, Cetani F, Costagliola S, Alcalde L, Uibo R, Vassart G, Ludgate M, Mapping thyroid peroxidase epitopes using recombinant protein fragments. Eur J Endocrinol 1995,132:53-61. ISSN 0804–4643 The identification of autoantibody epitopes is important to the understanding of autoimmune thyroid diseases. In the case of thyroid peroxidase antibodies (TPO-ab), recent reports have disagreed about the number and type of autoantibody epitopes found in human TPO. In order to clarify the nature of these epitopes, we used an approach that provides recombinant human TPO produced by bacterial cells. The cDNA of four slightly overlapping fragments of human TPO—TPO 1(Glu 17–Ser 227), TPO 2(Tyr 226–Thr 476), TPO 3(Glu 471–Ser 720) and TPO 4(Phe 709–Leu 993)—were amplified by polymerase chain reaction and subcloned into the expression vector pMAL. In addition, a TPO 3 species for an alternatively spliced form of TPO of 876 amino acids was constructed (TPO 3M). Each of these constructs encodes a fusion protein, in which the amino terminal portion is maltose-binding protein, followed by the sequence of the fragment of human TPO. The plasmid constructs were transformed in Escherichia coli and, after growth, bacterial cells were harvested, lysed and the lysate was passed over an amylose affinity column and eluted with maltose. Western blots were performed using 33 sera from patients with autoimmune thyroid disease (group 1) and 17 sera from patients with nodular goiter and focal thyroiditis (group 2), all positive for TPO-ab measured by radioimmunoassay; sera from 10 healthy people with no clinical evidence of thyroiditis and positive for TPO-ab measured by radioimmunoassay (group 3) and sera from 30 patients with antigastric parietal cell antibodies without signs or symptoms of thyroiditis, 16 negative for TPO-ab (group 4a) and 14 positive for TPO-ab (group 4b), were included in the study. The four TPO fragments and the alternatively spliced form of TPO were used as antigens. Our results show that in the first group 91% of sera recognized TPO 3, 33% recognized TPO 4 while 21% and 18% reacted with TPO 1 and TPO 2, respectively. In the second group of patients 76% of sera recognized TPO 3, 35% recognized TPO 4, 29% reacted with TPO 1 and 29% with TPO 2. In group 3 (normal healthy people) the individual peptides were recognized at a similar frequency compared to the other groups. In group 4a, 75% and in group 4b, 93% of sera reacted with one of the fusion recombinant proteins. Immunoreactivity with TPO 3M was the same as that of TPO 3. In conclusion: (i) we have optimized the production of TPO peptides in bacterial cells; (ii) TPO-ab are able to bind many amino acid sequences of the protein with a hot spot in TPO 3 (known to contain the linear epitopes C2(590–622) and C21(709–721); (iii) TPO-ab recognize other epitopes in the amino and carboxyl portions of the protein; (iv) no difference is observed when comparing the levels of antibodies and the number or type of peptide fragments recognized; (v) there are no disease-specific epitopes, as sera from normal healthy subjects with TPO-ab recognize the same epitopes in a similar percentage to those recognized by patients; (vi) the immunoreactivity of TPO 3 is not changed when the fragment expressed is the alternatively spliced form; (vii) we confirm that antigastric parietal cell antibodies recognize epitopes on the TPO molecule even when TPO-ab are negative by radioimmunoassay, suggesting a shared epitope between TPO and the gastric parietal cell. Massimo Tonacchera, Institut de Recherche Interdisciplinaire, Université Libre de Bruxelles, Campus Erasme, 808 Route de Lennik, 1070 Bruxelles, Belgium