Background:  Mutations conferring resistance to ART compounds carry a selection advantage for HIV. To investigate whether drug escape is also associated with either immune escape or de novo immune responses, we investigated CD8⁺ effector T cells reactive against either wild type or NRTI- or NNRTI-resistant HIV mutants.

Methods:  In 5 HIV-infected subjects selected from the Hamburg cohort according to HLA-A2 positivity and development of RT resistance mutations during HAART, plasma virus was genotyped at different time-points (T1 and T2). Subject 1 had an identical mutational pattern at both time-points, whereas in subjects 2 through 5, new RT mutations arose at T2. Subjects 4 and 5 were ART-naive at T1 and had wild type virus. Mutation peptides (RT AA substitutions M41L, Y181C, M184V, M184I, or L210W), as well as their corresponding wt peptides were selected according to high affinity score predictions based on the individual HLA class I haplotype, using 2 prediction algorithms (SYFPEITHI, BIMAS). Optimal HIV-1 peptides and mutation/wild type peptides with low affinity scores served as positive/negative controls. For subjects 3 to 5 wild type and mutation peptides were also synthesized according to the autologous RT sequence. CD8⁺ T cell responses were determined by interferon-g (IFN-g) ELISpot.

Results:  Clear responses against wild type and mutation peptides were detected in subjects 3 to 5, mostly directed against peptides containing M41L, M184V and I, and the corresponding wild type. De novo responses against the M41L mutation clearly emerged in subjects 3 and 4, both of whom had developed the M41L mutation at T2. They responded to 5 of 5 or 3 of 5 mutation peptides at T2, respectively, in contrast to only 1 of 5 or no mutation peptides at T1. Both subjects also recognized the wild type at T2, albeit at lower amplitudes, with a response to the wild type at baseline (T1) in subject 3 only. Responses to the M184V occurred together with those against M184I and wild type. Subject 4 responded to all 3 peptides at both time-points, independent of the actual presence of the M184V mutation, whereas subject 5 had no response at baseline, but at T2, when the mutation was actually present.

Conclusions:  Emergence of drug resistance mutations can lead to enhancement of cytotoxic T lymphocyte recognition. Clearly, in subjects acquiring the M41L mutation, new responses directed against this mutation occur. After emergence of M184V, new responses against this mutation can be detected as well. Their occurrence in conjunction with responses against M184I and wild type suggests epitope-sharing.