HIV Controllers: A Homogeneous Group of HIV-1--Infected Patients with Spontaneous Control of Viral Replication

A small number of HIV-infected patients maintain high CD4+ T cell counts in the absence of antiretroviral therapy, despite a prolonged course of infection. These patients have been called long-term nonprogressors (LTNPs). The cause of the lack of disease progression in LTNPs is unclear, but long-term nonprogression seems to result from the interaction between factors linked to the virus (such as attenuation in HIV strains) or to the host [1]. Genetic susceptibility (e.g., the protective effect associated with the HLA B*5701 class I allele [2]) and the immune response (i.e., proliferation of HIV-specific CD8+ T cells and/or HIV-suppressive activity of CD8+ T cells) are involved [1, 3–5]. However, in contrast with data from basic science studies, demographic data for LTNPs are rare. Over time, the proportion of HIV-infected patients who are LTNPs progressively decreases, because of the progressive decrease of the CD4+ T cell count in some LTNPs [6]. The definition of LTNP, which is based on a high CD4+ T cell count, differs between studies and does not take into account the plasma HIV-1 RNA load [3, 4, 7–9]. For the majority of these patients, the plasma virus load is low, but the range is wide [3, 4, 9], and the decrease of the CD4+ T cells has been linked to high-level viremia [9]. In our clinical practice, we have observed patients in whom viral replication is spontaneously maintained below the lower limit of detection [10]. Therefore, we selected HIV-1—infected patients who had never received antiretroviral therapy and who had persistently maintained long-term control of HIV-1 viremia. The purpose of this study is to describe the frequency of this phenomenon among and the characteristics of those patients we have termed “HIV controllers.” Patients and methods. We defined an HIV controller as an HIV-infected patient who had been followed up for >10 years, who had received no antiretroviral treatment, and for whom >90% of the plasma HIV RNA load measurements were >400 copies/mL. CD4+ T cell count was not part of the definition. Among 1300 HIV-infected patients followed up in our department of infectious diseases (Centre Hospitalo—Universitaire de Bicêtre, Bicêtre, Assistance Publique—Hôpitaux de Paris), 8 met the above criteria (group A). We further extended our research to the French Agence Nationale de Recherche pour le SIDA et les Hépatites Virales SEROCO-HEMOCO cohort of 1551 HIV-infected patients that was initiated in 1988 [11]. In this cohort, 9 patients matched our definition of HIV controllers (group B), 2 of whom were followed up in our department and were included in group A. HIV DNA was quantified in PBMCs. One microgram of total DNA extracted from PBMCs was tested by means of a real-time PCR that targets a conserved region of the long terminal repeat region of the HIV-1 genome and had a sensitivity of 10 HIV-1 DNA copies per reaction [12]. Results were expressed as the number of HIV DNA copies per 106 PBMCs. HIV-specific CD8+ T cells were quantitated using an IFN-γ enzyme-linked immunospot (Elispot) assay, as described elsewhere [13]. Antigens consisted of peptides, distributed in pools, that corresponded to well-described optimal CD8+ T cell epitopes. The number of antigen-specific CD8+ T cells was expressed as the number of spot-forming cells (SFCs) per 106 PBMCs. To determine the HIV-specific CD4+ T cell response, lymphoproliferative responses to the p24 antigen were assessed by means of an IFN-γ Elispot assay [13]; data are specified as mean values ± SD. In addition, genotypic analysis of the gene encoding CCR-5 [11] and phenotypic analysis of the HLA class I alleles were performed. Results. Patients with HIV controller status were rare, comprising <1% of HIV-infected patients who were followed up in an infectious diseases department and/or were participating in the French prospective SEROCO-HEMOCO cohort, but the frequency of this condition was similar in each cohort. All patients had a positive HIV serologic titer. Their characteristics are given in table 1. All patients had stage A HIV infection, according to the Centers for Disease Control and Prevention classification, and were asymptomatic. Two patients had chronic hepatitis B, and 7 had chronic hepatitis C. Results of tests for detection of antiretroviral drugs in plasma specimens from patients in group A were negative. Characteristics of 15 patients in whom the plasma HIV RNA load has been undetectable for >10 years in the absence of antiretroviral treatment. The median CD4+ T cell count was 750 × 106 cells/L (range, 280–1398 cells/L) in group A and 750 × 106 cells/L (range, 332–1225 cells/L) in group B. Patients had a stable CD4+ T cell count over time: 3 had a CD4+ T cell count with a slope of 0 (mean change in the CD4+ T cell count, 0 ± 5 cells/year), 10 had a CD4+ T cell count with a slightly negative slope (mean decrease, 21 cells/year; range, 6–40 cells/year), and 2 had a CD4+ T cell count with positive slope (mean increase, 12 cells/year). The main characteristic common to these patients was the control of HIV replication. The HIV load was by definition nearly always <400 copies/mL during the follow-up period. Results of ultrasensitive tests with a lower limit of detection of 50 copies/mL revealed that 5 of 15 patients never had a detectable plasma HIV RNA load. The remaining 10 patients were found to have episodes of intermittent viremia (also known as “blips”): 38 (20%) of 178 virus load measurements were >50 copies/mL (median virus load, 165 copies/mL). For 2 patients, 1 (10%) of 10 and 1 (6.7%) of 15 virus load measurements were >400 copies/mL. The HIV DNA load in PBMCs was quantified (table 1). In all patients, the HIV DNA load in PBMCs was very low (mean HIV DNA load, 32 copies/106 PBMCs; range, 0–251 copies/106 PBMCs). Of interest, the 2 patients with the largest decrease in the CD4+ T cell count had higher HIV DNA levels in PBMCs. Of 7 patients from the SEROCO-HEMOCO cohort for whom sequential quantifications of HIV DNA were available, 6 had a stable HIV DNA load during the follow-up period. All patients in group A were infected with group M, subtype B HIV strains. We studied the immunological characteristics of patients in group A from whom fresh blood samples could be more easily obtained (table 1). Lymphoproliferative responses to p24 antigen were observed in all patients, with a mean stimulation index (±SD) of 10.8 ± 7 and a mean net counts per minute (±SD) of 33,913 ± 24,374. HIV-specific CD4+ T cells were detected in 7 of 8 patients (mean SFC count, 622 ± 715 cells/106 PBMCs) and large numbers of HIV-specific CD8+ T cells were observed in all patients (mean SFC count, 4800 ± 4217 cells/106 PBMCs) by means of an IFN-γ Elispot assay. We next investigated whether these patients had cellular features associated with a decreasing rate of HIV disease progression. No patient was homozygous for the CCR-5 δ32 deletion, and only 2 were heterozygous for this deletion. Some HLA phenotypes have been associated with LTNP status [2]. HLA phenotyping could only be performed for 6 patients from group A. Five patients tested positive for the HLA B*57 allele, and 3 tested positive for the HLA B*27 allele. Coinfection with GB virus C has been shown to be associated with a slower progression to AIDS [14]. We looked for GB virus C RNA in frozen plasma specimens from patients in group A, but only 1 patient had a specimen that tested positive for this pathogen. Discussion. We assembled a rare homogeneous group of patients (0.6% of patients in 2 cohorts) that we referred to as HIV controllers on the basis of a strict virological definition. These patients are characterized by spontaneous control of HIV-1 infection and by a prolonged follow-up period without progression to AIDS. Of interest, despite prolonged control of viremia, 3 of the 15 patients we described were not LTNPs, according to their CD4+ T cell counts. Excessive T cell activation may explain the decreased CD4+ T cell counts [9]. Patients selected on the basis of the same virus control criteria were described in 2001 [8], but no data were available about the frequency of such patients in an HIV-infected population or about the HIV DNA load in PBMCs. We found very low and very stable HIV DNA loads in PBMCs in all patients. These low HIV DNA loads contrast with the amounts observed in patients receiving prolonged and efficient HAART [15]. This finding helps explain the lack of disease progression despite prolonged infection, because the HIV DNA load has a significant and independent influence on the rate of disease progression [12]. The patients described in this report have been able to sustain a strong HIV-specific CD8+ T cell response (mean SFC count, 4800 ± 4217 cells/106 PBMCs, by the IFN-γ Elispot assay) despite a low level of viral antigens. This response is higher than, although not significantly different from, the response observed in viremic patients (mean SFC count, 3343 ± 2623 cells/106 PBMCs) [16]. However, when individual HLA-matched optimal peptides are analyzed instead of pools of peptides, the difference becomes significant (mean SFC count, 7716 ± 5404 vs. 4084 ± 3623 cells/106 PBMCs in HIV controllers and viremic patients, respectively; P = .004) (unpublished data). This point contrasts with recently published data [2, 5] showing no quantitative difference in HIV-specific CD8+ T cell responses between LTNPs and patients with progressing HIV disease. Moreover, these findings are markedly different from findings for patients with undetectable plasma HIV RNA loads during long-term HAART, in whom the HIV-specific CD8+ T cell count decreases as the control of viral replication is prolonged (mean SFC count, 621 ± 1295 cells/106 PBMCs; P < .0001) [13]. On the other hand, our results support previous data [17] showing that HIV-specific CD4+ T cell proliferation is correlated with restricted virus replication. We also found a bias in the detection of some HLA alleles, such as HLA B*57 and HLA B*27, which has been reported elsewhere [2]. Three non—mutually exclusive hypotheses that we are currently investigating may explain this status of HIV controllers. Attenuated virus strains may be involved, such as those with the nef deletion or those with the vpr mutation, which have been reported in some groups of LTNPs [1, 9]. Particular cellular phenotypes may result in a reduced susceptibility to HIV infection among CD4+ T cells, such as the expression of antiretroviral factors from the gene families encoding TRIM5-α or APOBEC3G [18]. Lastly, these patients may have a particularly efficient immune response that helps control HIV replication. Despite the rarity of HIV controllers, analysis of such persons offers a unique opportunity to improve our understanding of the pathogenesis of HIV infection and might provide new insights for vaccine development. We thank M. T. Rannou, A. Urrutia, the partipating nurses, and the patients, for their cooperation; Dr. I. Theodorou, for CCR-5 genotyping; and Prof. Laffont, for screening for GB virus C RNA. Financial support. Agence Nationale de Recherches sur le SIDA et les Hépatites Virales, Ensemble Contre le SIDA, INSERM, and Université Paris-Sud. Potential conflicts of interest. All authors: no conflicts.