Cross Reactivity of Heparin-Induced PF4 Antibodies with Heparin- Bound IL-8.

Background: Heparin-Induced Thrombocytopenia (HIT) is caused by polyclonal antibodies, formed to conformational neoepitopes exposed in platelet factor 4 (PF4) bound to heparin (H). HIT antibodies are commonly measured by one of several commercially available ELISAs. The literature reports cases of clinical HIT and/or positive HIT antibody-mediated platelet activation tests in patients without antibodies detectable in PF4 based ELISAs, and suggests that antibodies to other chemokines, such as interleukin-8 (IL-8), may be involved. The current study was undertaken to characterize the frequency of such antibodies in a population of consecutive serum specimens submitted for HIT diagnosis by ¹⁴C-serotonin release assay (SRA), and also in a group of retrospectively selected SRA-positive specimens.

Methods: Specimens (n=90) analyzed by SRA between March and July 2008 were subsequently tested by PF4 Enhanced ELISA (GTI, Waukesha, WI), and the Zymutest HIA ELISA (Hyphen BioMed, Neuville sur Oise, France). An additional ELISA was carried out on the specimens using the Zymutest heparin-bonded plate coated with purified, recombinant human interleukin-8 (IL-8) (72 aa form, Prospec, Rehovot, Israel). The second group consisted of archived specimens (n=32) that tested positive by SRA but negative by GTI HAT ELISA; these specimens were subsequently tested by the Zymutest HIA and the H:IL-8 ELISA.

Results: Recombinant IL-8 (1μg/well) added to heparin-bound ELISA plates was recognized by anti-IL-8 (BD Biosciences), but not anti-PF4 (Leinco Technologies, Inc.) antibodies and was not displaced from the plate by addition of PF4 from normal serum at the dilution (1:100) used in the ELISA. Of the 90 consecutive specimens, 21 (23.3%) tested positive in both the GTI and the Zymutest HIA tests; 15 of the 90 (16.6%) tested positive by one but not both of the PF4-based ELISAs (Table 1). In this group of 90 specimens, 17 (18.9%) had OD > 0.500 on the heparin-bound IL-8 plate. Only 3 of the specimens bound to the H:IL-8 exclusively, while 14 were also positive in the GTI and/or Zymutest HIA ELISA. Thus, of the 36 specimens that tested positive in one or both of the PF-4-based ELISAs, 14 (38.9%) had antibodies that also targeted heparin-bound IL-8. Ten of the 90 specimens tested positive in the SRA. Six of SRA-positives had PF4-specific antibodies that did not cross react with H:IL-8, 3 had H:IL-8 reactive antibodies in addition to PF4 reactive antibodies, and 1 SRA positive specimen tested negative in all ELISAs.

Table 1. Comparison of GTI HAT vs Zymutest HIA results: n = 90 consecutive patients

Numbers in parentheses indicate how many specimens in each subgroup contained antibodies that bound to H:IL-8.

Among the set of 32 specimens that tested positive by SRA and negative by GTI ELISA, 8 tested positive in the Zymutest ELISA, including 4 that had OD > 0.500 on the H:IL-8 plate. The 24 specimens that tested negative by both HIT ELISAs ranged in percent serotonin release (tested with 0.1 U/ml H) between 21–84% with a median of 45.6%, and did not contain antibodies that bound the H:IL-8 plate.

Conclusions: The data agree with other studies that show differences between results of commercially available HIT ELISA kits. The study also shows that over one third of specimens that test positive in the HIT ELISAs contain antibodies that are not specific for PF4, but also recognize heparin-bound IL-8. These results do not prove that the antibodies that bind PF4 targets are the same ones that bind IL-8. However, the significant degree of homology (40%) between the peptides suggests that the same antibodies may cross react with either target. The relevance of these antibodies is as yet unclear. Only 3% of specimens bound H:IL-8 plates but not PF4 plates, and these were not associated with SRA activity. Also HIT ELISA negative specimens with SRA activity, which showed only moderate percent serotonin release, did not contain H:IL-8-binding antibodies. The study did not find evidence of a role for H:IL-8 antibodies in HIT-related platelet activation tests.