Although early treatment of acute HIV-1 infection followed by treatment interruptions may enhance control of viremia [6
], the durability of this control remains unclear. Here we analyzed the long-term impact of initiation of antiviral therapy during acute HIV-1 infection followed by STIs in a cohort of 14 patients. Although initial control of viremia to less than 5,000 RNA copies/ml plasma was achieved in the majority of the individuals studied, a gradual increase in viremia and decline in CD4+ T cell counts was observed in most patients, even after a year or more of viral containment. Durable virologic control occurred infrequently, despite the presence of robust HIV-1-specific CD4+ and CD8+ T cell responses detected by standard assays. Moreover, even during periods of successful control of viremia, progressive loss of CD4+ T cells was frequently observed. These data indicate that although early treatment of acute and early infection is frequently associated with transient control of viremia after STI, ongoing low-level viral replication is associated with ultimate virologic breakthrough in most patients.
The standard immunologic assays and virologic assessments in this cohort revealed considerable heterogeneity among the study participants, and did not show a consistent pattern in duration of viremia control during successive treatment interruptions. Eleven of 14 patients (79%) were able to maintain a viral load of less than 5,000 copies/ml for at least 90 d, but progressive loss of control ensued in the majority of patients and only three patients (21%) were able to maintain control for more than 2 y. These three patients did so during the first (AC-10), the second (AC-02), and the third (AC-14) STI. Clinical, genetic, and immunological parameters did not distinguish these three individuals from the other 11 patients, nor did they predict the duration of control following treatment interruption. Indeed, the longer a patient was off therapy, the stronger and more broadly directed the CD8+ T cell responses became, but these were still not sufficient to maintain prolonged control in most patients. Although three patients did not complete the study as initially intended (patient AC-45 withdrew from the study after viral breakthrough on the first STI, AC-13 restarted therapy despite a viral load of less than 5,000 copies/ml during both the first and second STIs and then withdrew, and AC-05 restarted therapy prematurely during the second STI but then failed to control during the third STI), the results are not substantially different if these three are censored rather than considered to have failed to control.
Loss of viral control in this cohort occurred not only in the presence of strong CD8+ T cell responses, but in most cases also in the presence of virus-specific CD4+ T cell responses, although the CD4+ T cell responses often declined during periods of viremia. In addition, total CD4+ T cell numbers were also monitored and declined in most patients over time, including one of the three patients who were able to maintain low viral loads for at least 2 y. Mechanisms leading to rapid CD4+ T cell loss need to be further studied in future STI trials. Other parameters including chemokine receptor polymorphisms [26
] and GBV-C coinfection [27
] similarly failed to explain the different courses following treatment interruption. The only parameter found to be associated with longer control of viremia during the first treatment interruption was a lower viral load at time of institution of antiviral therapy. Given the multiplicity of comparisons made, the true significance of this finding is uncertain.
The reasons for progressive loss of control despite augmentation of virus-specific CD4+ and CD8+ T cell responses remain to be defined. In one individual (AC-06), HIV-1 superinfection in the setting of strong and broadly directed HIV-specific cellular immune responses was associated with the loss of viral control, as previously reported [29
]. No other cases of superinfection have been identified in these patients (data not shown). The immunologic studies performed failed to show an association between increases in viral load and loss of immune responses, but this may be due to the use of the current standard IFN-γ assays to quantify immune function. Numerous studies now indicate that IFN-γ production alone is not associated with viral load [19
] but rather that functional characteristics of CD4+ and CD8+ T cells may be better associated with viral control [32
]. Such studies will be important to pursue. In particular, even a low level of viremia correlates with a low or undetectable frequency of interleukin-2-producing HIV-1-specific memory CD4+ T cells endowed with proliferative capacity in vitro [36
], thus abrogating CD4+ T cell help crucial to maintain efficacy of CD8+ T cell functions. In an interim study of a subset of six of the 14 patients presented here (patients AC-02, AC-05, AC-14, AC-15, AC-25, and AC-46), a fully differentiated effector phenotype of HIV-1-specific CD8+ T cells for selected epitopes was found to be associated with better control of viremia [10
]. Other factors that may contribute include functional defects in antigen-specific cell-mediated immunity [35
], and progressive immune escape [43
]. HIV-1-specific humoral immunity can also affect viral control after treatment interruption [46
], and viral factors including viral fitness [47
] and infection with multiple viral variants [49
] can influence viral set point and the rate of disease progression. Virus sequencing studies currently in progress in this cohort indicate that viral breakthrough is associated with sequence changes within and outside known CTL epitopes (data not shown). Full evaluation of the relationship between immune escape and viral breakthrough will require extensive additional analyses, including detailed analysis of responses to autologous virus [50
]. Assessing the changes in CD4+ and CD8+ T cell functions over time as well as viral evolution under immune selection pressure will be important to evaluate immune correlates in this cohort.
These data are important in light of other recent data on treatment interruption in both acute and chronic infection. In chronic HIV-1 infection, STI studies showed only marginal, if any, improvements of HIV-1 viremia control following a number of treatment interruptions cycles, despite at least transient increases in HIV-1-specific CD8+ and CD4+ T cell responses [4
]. In the setting of infection with a multidrug-resistant virus, this strategy may even be deleterious [56
]. Other studies of STI after treated acute HIV-1 infection have shown limited benefits [9
], including recent trials such as the PrimSTOP trial [57
] and the QUEST study [58
]. However, little is known about the relationship between scheduling of HAART and treatment interruptions and the characteristics of viral rebound after therapy has been discontinued.
Although durable control of viremia was not achieved, it is noteworthy that the majority of patients were able to achieve transient relative containment of viremia, providing rationale for future studies aimed at further enhancing immune control. Early treatment alone should still be considered an important therapeutic option. Therapeutic vaccinations administered after treated acute HIV-1 infection and before cessation of therapy have given disappointing results thus far [9
], but the availability of new and more potent immunogens requires reassessment of this approach. Indeed, the ability to enhance CD4+ T helper cell responses in the chronic phase of infection has been demonstrated [59
], but whether this will enhance CD8+ T cell function requires additional studies. Some promising results have been obtained using immunomodulatory drugs, including cyclosporine [60
] and hydroxyurea [61
], in combination with antiviral therapy, presumably because of the limitation of T cell activation. Administration of granulocyte-macrophage colony-stimulating factor blunted the viral rebound following interruption of HAART, and largely prevented a decrease of CD4+ T cell counts in an STI trial in chronic HIV-1 infection [62
]. These additional therapeutic interventions deserve further investigation in future STI studies.
Although the present study shows progressive viral breakthrough, it was not designed to address whether there might be a change in set point viremia achieved or overall clinical benefit through transient early treatment of acute HIV infection. The definition of failure chosen for this study was a viral load of greater than 5,000 RNA copies/ml plasma, which at the time the study was initiated corresponded to the level of viremia at which treatment was recommended. Larger randomized trials will be needed to determine the potential clinical and virologic benefit of approaches based on STIs. In studies of untreated infection, there is only a 5-fold difference in viremia separating the quartile with the slowest disease progression from the quartile with the most rapid progression [63
], suggesting that small differences in steady-state viremia may influence clinical outcome. In the meantime, STI probably should be avoided outside the setting of controlled clinical trials. The data in this study may also be relevant to current efforts to develop a therapeutic AIDS vaccine designed to retard disease progression rather than prevent infection, since they suggest that durable maintenance of low-level viremia may be difficult to achieve.