The Role of Protein

Only a few investigators have attempted to map the actual epitopes on PF4-heparin complexes recognized by HIT antibodies. Horsewood et al. (1996) studied a total of 29 antibodies from patients with HIT that were positive in the PF4-heparin enzyme immunoassay (EIA) and in the SRA. Five of these antibodies also reacted with reduced alkylated PF4 in the presence of heparin. The same five antibodies also recognized a peptide containing the 19 COOH-terminal amino acid residues of the PF4 monomer, a region that encompasses a positively charged a-helical domain thought to be critical for heparin binding (Loscalzo et al., 1985). However, neither reduced PF4 nor the COOH-terminal peptide could inhibit binding of HIT antibodies to PF4-heparin complexes, even at high concentrations. Therefore, the clinical significance of the five antibodies is uncertain.

Amiral and coworkers (1996a) studied a subgroup of 15 patients thought to have HIT, whose antibodies were positive in a PAT but negative in PF4-heparin EIA. Nine of these patients had antibodies that recognized NAP-2 or IL-8, or both, two members of the CXC chemokine family that are homologous with PF4. These findings are of interest because five of the nine patients had thrombotic episodes. However, reactions of these antibodies against NAP-2 or IL-8 in their normal configurations (not immobilized on plastic) were not described, and their relation to antibodies that recognize PF4-heparin complexes is uncertain.

Ziporen et al. (1998) studied the binding of antibodies from 50 HIT patients to different constructs of PF4 containing a single amino acid substitution and chimeric proteins containing various portions of human PF4 and NAP-2. Mutation to alanine of three (K62, K65, K66) of the four lysine residues in the COOH-terminal a-helix had only minimal effect on the binding of HIT antibodies, and the K61 ~ A mutation reduced antibody binding by only about 50%, indicating that the COOH-terminal lysines of PF4 do not constitute the major antigenic site for HIT antibodies. NH2-terminal PF4-NAP-2 chimeras exhibited only slightly reduced antibody binding. In contrast, the PF4-NAP-2 chimera, in which the portion of PF4 lying between the third and fourth cysteine residue (amino acids 37-47) was substituted by the corresponding NAP-2 sequence, was almost totally nonreactive.

With a different approach, we found that, although human PF4 has 74% protein sequence identity to bovine and rat PF4, neither bovine nor rat PF4 complexed to heparin is recognized by HIT antibodies (Visentin, 1999). Yet, rat PF4 differs from its human counterpart at only six of its 47 COOH-terminal amino acids (Doi et al., 1987; Poncz et al., 1987) (Fig. 3). To characterize the binding sites for HIT antibodies on PF4-heparin, we constructed seven PF4 mutants in which the human sequence (reactive) was converted to the corresponding residues of rat PF4 (nonreactive) and determined the effect of each change on HIT antibody binding to the construct complexed with heparin. The PF4 constructs tested were comparable with wild-type PF4 in their avidity for heparin. Each of the 15 antibodies from HIT patients recognized PF4-heparin complexes containing PF4 constructs bearing mutations: E4 ~ S, L11 ~ V, and T16 ~ S at the NH2-terminus, or A57 ~ V at the COOH-terminus just as well as the wild-type human PF4-heparin complexes. In contrast, complexes containing other COOH-terminal mutants: P37 ~ A / T38 ~ V/A39 ~ P, R49 ~ S, and L55 ~ R exhibited varying degrees of reduced binding. The HIT antibodies tested recognized PF4 mutated at positions 49 and 55 only at a higher ratio of heparin to PF4 (0.8U/mL vs. 0.5 U/ mL). None of the 15 antibodies recognized peptides comprising the 26 or 15

COOH-terminal amino acid residues of the PF4 monomer or reduced alkylated human PF4 either in the presence or in the absence of heparin.

These results, together with the observations by Ziporen and associates (1998), point to the region of PF4 between the third and fourth cysteine residues as the major antigenic site for HIT antibody binding (Fig. 4). Li et al. (2002), using a series of mouse/human PF4 chimeras, identified another antigenic site on PF4-heparin that requires both P34 and an intact N-terminus (Fig. 4). The latter results, together with our studies utilizing biotin-labeled, affinity-purified HIT antibodies in a competitive inhibition assay (Suh et al., 1998), indicate that at least three dominant HIT antibody recognition sites can be distinguished and further support the idea that HIT antibodies recognize conformation-dependent "neoepitopes" formed on PF4 when it binds to heparin.

0 0

Post a comment