271
Evolution of HIV-1 CTL Epitopes: Balance between CTL Escape and Fitness Constraints
Yi Liu*1, H Zhao1, J McNevin2, I Genowati1, M McSweyn2, D Shriner3, M McElrath2, and J Mullins1
1Univ of Washington, Seattle, US; 2Fred Hutchinson Cancer Res Ctr, Seattle, WA, US; and 3Univ of Alabama at Birmingham, US
Background: CD8+ cytotoxic
T-lymphocyte (CTL) responses are associated with control of HIV-1 infection. CTL epitopes at which
escape mutants were detected,
as well as epitopes that persist at high frequencies
in HIV-1 infection have been reported. It is not clear why at
some epitopes escape mutants are detected while at
others they are not. However, viral polymorphisms
of high representation in the HIV Sequence Database have been correlated with higher
viral fitness.
Methods: We sequenced whole or multiple HIV-1 viral
genome segments from 1 person at 17 time points over the first 4 years of
infection, and characterized the recognition of epitopic
or mutant peptides spanning the entire proteome by interferon-gamma (IFN-g)-secreting T cells at 16 time points. For each amino
acid site associated with cytotoxic T lymphocyte (CTL)
escape, frequencies of the amino acid at the
respective positions in its epitope and escape mutant
forms were calculated and compared using the Wilcoxon
signed rank test. Furthermore, for each amino acid site in a recognized
CTL epitope, its Shannon
entropy from the HIV Sequence Database was calculated: lower entropy indicates lower variability,
suggesting a stronger fitness constraint. Entropies of amino acid sites
associated with CTL escape were compared with those of sites not associated
with escape using the Mann-Whitney test.
Results: We detected 25 CTL epitopes.
Mutations at 25 sites in 16 epitopes
were associated with CTL escape, 20 of which
showed lower database frequencies. Database frequencies of the mutant forms
were significantly lower than that of the epitope
forms (p = 0.0015). Compared with
amino acids in CTL epitopes not exhibiting escape mutations, amino acids
related to escape mutations were associated with higher Shannon entropy (p = 0.002) and we infer, weak functional
constraints. This association was further supported when all 42 reported escape
mutants of defined CTL epitopes in the HIV Immunology
Database were studied. Finally, we found that cross-reactivity of CTL responses
to a mutant form of an epitope was associated with a
failure to rise in frequency in the subject.
Conclusions: Although
CTL responses selected for escape mutants, many potentially have fitness costs.
The balance between the advantage of escaping
from immune responses and the disadvantage of losing replicative
fitness plays an important role in determining whether a CTL epitope will persist
or by replaced by escape mutants during infection.
|