Discordant outcomes following failure of antiretroviral therapy are associated with substantial differences in human immunodeficiency virus-specific cellular immunity

Abstract

Many individuals chronically infected with human immunodeficiency virus type 1 (HIV-1) experience a recrudescence of plasma virus during continuous combination antiretroviral therapy (ART) due either to the emergence of drug-resistant viruses or to poor compliance. In most cases, virologic failure on ART is associated with a coincident decline in CD4(+) T lymphocyte levels. However, a proportion of discordant individuals retain a stable or even increasing CD4(+) T lymphocyte count despite virological failure. In order to address the nature of these different outcomes, we evaluated virologic and immunologic variables in a prospective, single-blinded, nonrandomized cohort of 53 subjects with chronic HIV-1 infection who had been treated with continuous ART and monitored intensively over a period of 19 months. In all individuals with detectable viremia on ART, multiple drug resistance mutations with similar impacts on viral growth kinetics were detected in the pol gene of circulating plasma virus. Further, C2V3 env gene analysis demonstrated sequences indicative of CCR5 coreceptor usage in the majority of those with detectable plasma viremia. In contrast to this homogeneous virologic pattern, comprehensive screening with a range of antigens derived from HIV-1 revealed substantial immunologic differences. Discordant subjects with stable CD4(+) T lymphocyte counts in the presence of recrudescent virus demonstrated potent virus-specific CD4(+) and CD8(+) T lymphocyte responses. In contrast, subjects with virologic failure associated with declining CD4(+) T lymphocyte counts had substantially weaker HIV-specific CD4(+) T lymphocyte responses and exhibited a trend towards weaker HIV-specific CD8(+) T lymphocyte responses. Importantly the CD4(+) response was sustained over periods as long as 11 months, confirming the stability of the phenomenon. These correlative data lead to the testable hypothesis that the consequences of viral recrudescence during continuous ART are modulated by the HIV-specific cellular immune response.

FIG. 1.

Characterization of the three patient groups (S, D, and F) by longitudinal analysis of pVL and total CD4⁺ T lymphocyte count. The three clinical groups were derived following close clinical and laboratory monitoring of the 53 patients over a period of 19 months. The time of the sampling for this study is defined as month 0. The median pVL (log copies/ml) and median CD4 (10⁶ cells/liter) counts are shown over the 19 months prior to the date of sampling. The regression slopes for median CD4 counts show that group S (n = 24) had a positive CD4 slope and undetectable pVL over this period; group F (n = 8) had detectable pVL and a negative CD4 slope. Group D (n = 21) had detectable pVL and a positive CD4 slope over this period.

FIG. 2.

pVL and total CD4⁺ and CD8⁺ T lymphocyte counts for the three groups, F, D, and S, at the time of analysis. pVL on continuous ART at the time of analysis is shown in the upper left panel. Total CD4⁺ T lymphocyte counts before ART are shown in the lower left panel, and total CD4⁺ T lymphocyte counts on continuous ART at the time of analysis are shown in the lower right panel. Total CD8 counts on continuous ART at the time of analysis are shown in the upper right panel. Each circle represents a single patient. Bars indicate median values.

FIG. 3.

Recombinant virus growth kinetics. Recombinant viruses containing PR and RT genes from patient isolates were generated as described above. (A and B) Growth kinetics, determined by quantification of p24 Gag in culture supernatants, are shown for five such recombinants produced from group F patient virus isolates (A) and seven group D patient virus isolates (B); each symbol represents results for an individual recombinant. (C) The mean production at each time point from each group of viruses described for panels A and B is shown, together with bars for the standard error of the mean.

FIG. 4.

Antigen-specific CD4⁺ and CD8⁺ T lymphocyte responses. HIV-specific CD4⁺ T lymphocyte responses determined by IFN-γ ELISPOT analysis with CD8-depleted PBL are shown in the upper two panels. In the upper left panel, HIV p24 Gag-specific frequencies in CD8⁺ T lymphocyte-depleted PBL are shown. For each patient, the sum of CD4⁺ T lymphocyte frequencies specific for all HIV-derived antigens tested within CD8⁺ T lymphocyte-depleted PBL is shown in the upper right panel. In the lower left panel, frequencies of CD4⁺ T lymphocytes specific for the HIV-unrelated antigens SK/SD within CD8⁺ T cell-depleted PBL are shown. The lower right panel shows frequencies of HIV-specific CD8⁺ T cells as determined by IFN-γ ELISPOT analysis. Pooled overlapping peptides spanning HIV Env, RT, p24 Gag, p17 Gag, Nef, Tat, and Rev were used for stimulation of PBL, and for each patient, the sum of the responses to the individual peptide pools is represented by a circle. This approach strictly measures both HIV-specific CD4⁺ and CD8⁺ T lymphocyte responses; however, the HIV-specific CD4⁺ T lymphocyte responses observed in this study were more than 1 order of magnitude lower than the corresponding HIV-specific CD8⁺ T lymphocyte responses and did not impact on the statistical analysis presented (Table 6). Each circle represents a mean of two measurements from an individual patient. Bars indicate median values. Lines represent the detection limits of the experiments.