Mapping of Helper Epitopes to HPA-1a in Neonatal Alloimmune Thrombocytopenia with T-Cell Clones

Background: The major cause of severe neonatal allo-immune thrombocytopenia (NAITP) in Caucasians is fetal-maternal incompatibility for the human platelet antigen-1 (HPA-1), which is determined by the dimorphism Leu (HPA-1a) or Pro (HPA-1b) at amino acid position 33 of β3 integrin glycoprotein IIIa. Maternal responsiveness to HPA-1a shows a very strong association with DR52a (DRB3*0101) and 33Leu is thought to create a binding motif for this MHC class II molecule, thereby conferring helper T-cell responsiveness. At present, there are no preventative measures, reliable predictors of severity or screening procedures to identify women at risk of HPA-1a alloimmunization. Since the T-cells that help HPA-1a antibody production represent an attractive target for both screening assays and specific therapy, the aim was to define the epitope(s) they recognize in detail. Methods: From the peripheral blood of three HLA DR52a positive mothers with anti-HPA-1a antibodies and affected babies, we generated a total of six stable long-term CD4+ T-cell clones that respond specifically to the HPA-1a+ glycoprotein sequence. They have enabled us to characterize, for the first time, the fine specificity and restriction of the HPA-1a helper epitope. The core epitope was mapped by testing the responsiveness (proliferation and cytokine production) of the clones to panels of synthetic peptides spanning the HPA-1a 33Leu polymorphism, including sequences of different lengths, with selected single amino acid substitutions, and with the polymorphic residue located at different positions. Restriction was defined using antigen-presenting cells sharing HLA-DR and by flow cytometric analysis of staining with peptide-DR52a tetramers. The results, together with structural analyses, were used to model the interactions between MHC class II, HPA-1a peptide and specific helper T-cell receptor. Results: The 6 Th clones showed clear specificity for their HPA-1a epitope, even when naturally processed from whole platelets; they recognized only GPIIIa peptides (or platelets) with 33Leu and not 33Pro. The results of screening panels of linear peptides with 33Leu at different positions are consistent with a “core” epitope of 25WCSDEALPL33. The clones also specifically bound tetramerized DR52a complexed with a peptide spanning these residues. Together, the results show that 25Trp, 28Asp and 33Leu of HPA-1a are each important for anchoring respectively in pockets 1, 4 and 9 in DR52a, whereas 33Pro of HPA-1b sterically hinders docking in pocket 9. Extra residues 34Gly-35Ser did not affect T-cell recognition, but certain clones preferred N-terminal 24Ala or 23Cys-24Ala extensions. It has been previously reported that T-cells from alloimunized women with anti-HPA-1a recognize a 33Leu peptide cyclized by disulfide bridges (circular-Leu; 26C*SDEALPLGSPRC*38) but not to the circularized 33Pro equivalent. Our T-cell clones also responded moderately to this circular-33Leu peptide, demonstrating that 25Trp anchor at pocket 1 may not be essential when anchoring in pocket 4 and 9 are strong. Conclusions: Characterization of the HPA-1a specific Th clones reveals that they all recognize the core epitope GPIIIa 25Trp-33Leu, with the polymorphic 33Leu selectively anchoring it in the strongly predisposing HLA-DR52a. Although predicted to lie outside the peptide-binding groove of DR52a, extra N-terminal sequences promote optimal recognition by some T-cells, but are no longer required if the sequence is cyclized. Despite differences in TCR gene usage, the clones show remarkable consistency in fine specificity, supporting our previous evidence from polyclonal T-cell responses from women with anti-HPA-1a antibodies. The identification of a single “core” epitope, and the uniformity of restriction by DR52a in alloimmunized women, opens the way to both diagnostic and therapeutic exploitation in NAITP due to anti-HPA-1a.