Recent evidence in the literature suggests that molecular mimicry between viral and self antigens may be involved in the pathogenesis of autoimmune thyroid diseases in patients with chronic hepatitis C virus (HCV) infections [13].

Chronic HCV infection has been reported to be associated with thyroid autoimmunity and thyroid function disorders with a mean incidence of 10% and 3%, respectively [4, 5]. Alfa-IFN therapy may exacerbate or induce underlying latent thyroid disorders, increasing the incidence of thyroid autoimmunity and thyroid function disorders to 20% and 11%, respectively [4, 5].

In keeping with the tenets of the clonal selection theory of acquired immunity, an infectious agent may circumvent the deletion of anti-self lymphocytes activating clones with receptors sufficiently degenerated to respond to mimicking epitopes and host antigens [6].

A minimum of five to six amino acids are necessary to induce an immune response, and the probability of 20 amino acids occurring in six identical residues between two proteins is 206 (for each peptide, irrespective of the sequence) or 1 in 128 000 000 [7].

We performed the comparison between the amino acid sequence of the HCV polyprotein and five tissue-specific antigens of human thyroid, available in the database on www.ncbi.nlm.nih.gov/pubmed.

In particular, we examined the following HCV genotypes (with the respective NCBI sequence identification number): HCV1a (GI:130455), HCV1b (GI:130469), HCV1c (GI:385131), HCV2a (GI:130466), HCV2b (GI:130468), HCV2c (GI:555104), HCV3a (GI:514395), HCV3b (GI:676877), HCV4a (GI:402474), HCV5a (GI:2462303) and HCV6a (GI:2326455).

Regarding the thyroid gland, we examined the following tissue-specific antigens: the thyroglobulin (Tg) (GI:12644093), the thyroid peroxidase (TPO) (GI:129830), the thyrotropin receptor (TSHr) (GI:136448), the sodium/iodide symporter (NaIS) (GI:12643359) and Pendrin (GI:6174895).

Sequence alignments were carried out using the BLASTp, short nearly exact matches and BLASTp2 protein-protein comparison program (available at www.ncbi.nlm.nih.gov/BLAST).

Amino acid sequence homologies between the HCV polyprotein and the thyroid antigens are given in detail in Fig. 1, showing the presence of identical/conservative residues in the peptides. The following proteins of the HCV polyprotein have been examined: C (capsule, core protein), E1 (envelope glycoprotein 1), E2 (envelope glycoprotein 2, NS1), p7, NS2 (non-structural protein 2), NS3 (non-structural protein 3, protease/helicase), NS4a (non-structural protein 4a), NS4b (nonstructural protein 4b), NS5a (non-structural protein 5a) and NS5b (non-structural protein 5b, RNA polymerase).

Fig. 1
figure 1

Amino acid sequence homologies between the thyroid antigens and the HCV polyprotein (see text for abbreviations). *, Identical residues; +, conservative substitution. In each sequence, the first left-side residue inside the box corresponds to the number of the amino acid sequence. Amino acids in standard single letter code

The homologies between the thyroid and the viral peptides ranged from 62.5 % (five identical residues out of eight amino acids in the sequence) to 87.5% (seven identical residues out of eight amino acids in the sequence). The frequency of the homology increased up to 100%, when the conservative substitutions were included in the analysis (ten out of ten identical/conservative amino acids in the sequence, as indicated in Fig. 1 for NaIS444 and HCV1a-NS4a1665).

We found the presence of short peptides (eight to eleven amino acids) with a high degree of homology (62.5–100%) between the HCV polyprotein and five thyroid antigens (Tg, TPO, TSHr, NaIS and Pendrin).

The homology was not restricted to a single HCV genotype or to a single thyroid antigen.

The highest degree of homology was between the NaIs and the HCV1a-NS4a protein. The Tg antigen had the highest number of homologies with the different HCV genotypes.

Previous studies examining 20 amino acid-length peptides showed 41.7–58.3% sequence homologies between TPO and HCV-NS5a and HCV-NS2, increased to 75.0% when including conservative/identical residues [3].

We found mimicry between the TSH-r and the N-terminal hypervariable region 1 (HVR1) of E2 in HCV1a, which is well known to be involved in chronic HCV infection [8].

The length of the short peptides is consistent with the presentation of the self/viral antigens with the I class HLA molecules to CD8 positive lymphocytes, as the II class HLA molecules usually bind longer peptides, and the mimic peptides may be involved in the acceleration of autoimmune disorders occurring in chronic HCV infection [9]. In our examination, the RLGVRATRK-HCV2b-C43 sequence presented homology with the RLGVNVTWK-Tg1361 sequence; the same viral peptide has been recently identified as a HLA-A3 supertype-restricted cytotoxic T-lymphocyte epitope in patients with HCV infections [10].

The more frequent and earlier appearance of anti-Tg antibodies in the clinical course of the thyroid autoimmunity in HCV IFN-treated patients may be related to the high number of homologies between the Tg antigen and the HCV polyprotein, whereas the anti-TPO antibodies reflect a more advanced and aggressive autoimmune thyroid destruction [11].

Further studies are necessary in order to evaluate the clinical relevance of the presence of the molecular mimicry between the HCV and the thyroid antigens in the progression of autoimmune disease.