Objective: To evaluate the recognition of computationally designed, centralized HIV-1 antigens derived from clade B, C and group M sequences by individuals infected with HIV-1-M clades B and C.
Methods: Three centralized sequences have been described - consensus, ancestor and center-of-tree - each of which attempts to minimize the genetic distance to circulating viruses. It is unclear whether any of these sequences affords an advantage for T cell recognition. The ability of centralized clade B and C and group M peptides to be targeted in ELISpot assays was assessed using samples from the United States, Peru, Barbados and South Africa.
Results: Each of the clade-specific centralized peptide sets was equally powerful in detecting cytotoxic T cell (CTL) responses. Importantly, combination of these sets detected significantly broader responses. Although broad responses were observed in populations from which few sequences informed the design of these centralized sequences, the genetic distance between local sequences and the respective test set was inversely associated with response rates. Furthermore, the CTL reactivity in clade C-infected subjects using clade B peptides was reduced relative to within-clade peptide responses, while responses to group M peptides were comparable to within-clade peptide responses in these individuals.
Conclusions: All tested centralized antigens provided a similarly potent set of antigenic peptides. However, the significantly broader responses detected using the combination of sets highlight the importance of maximizing coverage of HIV-1 sequence diversity in vaccine preparations, as well as in the evaluation of CTL responses in HIV-1-infected individuals and those vaccinated.