Interruption of antiretroviral therapy to augment immune control of chronic HIV-1 infection: Risk without reward

There is no doubt that continuous treatment of HIV-infected persons with combinations of three or more antiretroviral drugs reduces morbidity and mortality from HIV-1 infection (1, 2). Nevertheless, growing concern about the long-term toxicity of antiretroviral drugs, the prevalence of HIV-1 drug resistance in patients receiving therapy,† and the substantial cost of continuous treatment has focused interest on postponing therapy or administering it only intermittently (3). Potential benefits of intermittent therapy include (i) augmentation of HIV-specific immunity through “autovaccination” (4) from recrudescent viremia that occurs during treatment interruptions; (ii) lowering of drug exposure to decrease toxicity; and (iii) reduction in the cost of antiretroviral therapy (5). Data in humans (6) and animal models (4) support “autovaccination” in acute HIV-1 infection (i.e., before serologic conversion), but this strategy is controversial in chronic HIV-1 infection because of limited and conflicting data. An important report in this issue of PNAS by Oxenius et al. (7) describes the disappointing results of sequential, structured treatment interruptions (STIs) on virologic and immunologic parameters in a large prospective trial of STI in chronic HIV-1 infection.

Oxenius et al. describe the disappointing results of sequential, structured treatment interruptions on virologic and immunologic parameters.

HIV-1 infection and the global pandemic of AIDS continue to pose a pressing challenge for humanity. After two decades of research, there are more questions than answers about immune control of HIV-1 infection. The precise pathogenic mechanisms whereby HIV-1 consistently evades the host immune response have not been elucidated, and the immune correlates of HIV-1 control are still incompletely defined. Cellular immunity is known to play a critical role in the control of chronic viral diseases, including HIV-1 infection. CD4⁺ T cell function is required for sustained CD8⁺ cytotoxic T lymphocyte (CTL) responses to HIV-1 (8). HIV-specific CD4⁺ T cell proliferative responses are inversely correlated with viral load (9) and are lost during the acute phase of HIV-1 infection. It is possible that HIV-specific CD4⁺ T cells persist but are anergic during chronic infection (10). Whatever the exact mechanism, loss of HIV-specific proliferative responses compromises the effectiveness of the immune response in chronic HIV-1 infection; this defect is not reversed by antiretroviral therapy (11), as evidenced by prompt rebound of viremia after cessation of antiretroviral therapy in chronically infected persons (12). By contrast, antiretroviral therapy initiated during the earliest stages of acute HIV-1 infection can preserve HIV-specific CD4⁺ T cell proliferative responses and results in enhanced immune control of HIV-1 after sequential STIs (6). This important finding led to a flurry of studies on STIs as a therapeutic strategy in chronic HIV-1 infection. Data emerging from these studies have been difficult to interpret, however, because of their small and nonrandomized design, differences in the STI schedule used, and variation in the methods of data analysis (13–20).

In this regard, the report by Oxenius et al. (7) is based on an analysis of 97 patients with chronic HIV infection enrolled in the Swiss-Spanish Intermittent Therapy Trial and is a welcome addition. All enrollees were on continuous antiretroviral therapy with a CD4 count >300 cells per mm³ and a plasma HIV-1 RNA <50 copies per ml for at least 6 months. After four cycles of STI (2 weeks off and 8 weeks on antiretroviral therapy), therapy was discontinued for 12 weeks. Detailed virologic and immunologic analyses were performed, including serial quantification of plasma HIV-1 RNA and HIV-1 peptide-elicited IFN-γ production from peripheral blood lymphocytes (PBLs) and CD8-depleted PBLs. Virologic analyses showed that sequential STIs did not substantially alter the viral load set-point established before the initiation of antiretroviral therapy. The average viral load after sequential STIs and cessation of antiretroviral therapy was ≈0.4 log₁₀ HIV-1 RNA copies per ml lower than the average viral load before initiation of therapy. This small difference is difficult to interpret because of the short duration of follow-up and the potential impact of reinitiating continuous therapy that occurred in a subset of patients but did seem to be accounted for in the comparison.

Immunologic analyses showed that sequential STIs did not increase the breadth or magnitude of the CD8⁺ T cell responses to a panel of HLA class I-restricted HIV-1 CTL epitopes. In fact, the CD8⁺ T cell responses after sequential STIs were inferior to those observed before the initiation of antiretroviral therapy among the subset of patients studied at both time points. This new finding demonstrates that sequential STIs partially restore the CD8⁺ T cell responses that existed before the initiation of antiretroviral therapy rather than enhancing responses to new levels. In addition, no correlation could be found between the breadth and magnitude of HIV-specific CD8⁺ T cell responses and the viral load observed after cessation of therapy. This indicates that immune control of HIV-1 is more complex than that characterized by CD8⁺ T cell responses to a panel of CTL epitopes. Other components of the immune system, including humoral immunity (21), antigen-presenting cells, natural killer cells, and soluble factors (22), must be accounted for to understand better the determinants of the virologic set point and to identify more promising strategies to enhance immune control of HIV-1 replication. The “autovaccination” strategy alone appears incapable of accomplishing this goal in patients with chronic HIV-1 infection. Autovaccination appears to be more effective in acute HIV-1 infection possibly because of an intact immune system and lesser genetic diversity of the HIV-1 population.‡ A major problem, however, that prevents the application of the STI strategy in this favorable setting is that acute HIV-1 is usually not recognized clinically.

The report by Oxenius et al. (7) does not weigh the potential risks of sequential treatment interruptions. These risks include symptomatic acute retroviral syndrome, reseeding of viral reservoirs, CD4⁺ T cell decline, greater risk of HIV-1 transmission, selection of HIV-1 drug resistance, and poor adherence to subsequent continuous therapy. All of these risks have been reported in the context of treatment interruption (15, 17, 23–25).§ Of these, the risk of drug resistance is probably the easiest to quantify. Several recent reports suggest that this risk is substantial. Zala et al.¶ reported that six of nine patients initiating therapy with a nevirapine-containing (a nonnucleoside reverse transcriptase inhibitor) treatment regimen developed nevirapine resistance mutations during the first STI. Similarly, other investigators have reported the selection of lamivudine (3TC) resistance during STIs.‖**‡‡ Among a subset of patients participating in the Swiss-Spanish Intermittent Therapy Trial, 8 of 21 (38%) leaving the trial before completing all four STIs developed 3TC resistance mutations.** In addition, minor virus populations with mutations at L90M and M184V were found.‡‡ Mathematical models have also predicted increased risk of drug resistance with STI (26, 27). The risk is more pronounced for drugs with a long half-life or low genetic barrier to resistance, such as nonnucleoside reverse transcriptase inhibitors and 3TC.

This risk of drug resistance must be carefully considered when exploring other intermittent therapy strategies aimed at reducing drug exposure to limit drug toxicity and cost. In a 10-patient pilot study, Dybul et al. (5) recently showed that a 7-day on/7-day off structured intermittent therapy (SIT) maintained effective viral suppression for 32–68 weeks without a significant change in CD4⁺ T cell counts, immune response parameters, or HIV-1 drug susceptibility and was associated with improved serum lipid profiles, which suggest less drug toxicity. The short 7-day cycle of treatment interruption in this study may have been critical in preventing detectable viral rebound and emergence of drug resistance; longer periods of treatment interruption (10–21 days) were associated with viral rebound (5). Despite the appeal of these initial results, larger studies with longer follow-up are needed to assess the risk of resistance and the comparative efficacy of this short-cycle strategy.

The clinical and research implications of the report by Oxenius et al. (7) are clear. STIs should not be performed in patients with chronic HIV-1 infection with the expectation of immunologic benefit, and the risk of drug resistance must be carefully weighed against the potential benefits of treatment interruptions. Risk should not be taken without reward.