Genetic and Immunologic Heterogeneity among Persons Who Control HIV Infection in the Absence of Therapy

MAJOR ARTICLE Genetic and Immunologic Heterogeneity among Persons Who Control HIV Infection in the Absence of Therapy Florencia Pereyra, 1,2 Marylyn M. Addo, 1 Daniel E. Kaufmann, 1 Yang Liu, 5 Toshiyuki Miura, 1 Almas Rathod, 1 Brett Baker, 1 Alicja Trocha, 1,4 Rachel Rosenberg, 1 Elizabeth Mackey, 1 Peggy Ueda, 1 Zhigang Lu, 1 Daniel Cohen, 3 Terri Wrin, 5 Christos J. Petropoulos, 5 Eric S. Rosenberg, 1 and Bruce D. Walker 1,4 1 Partners AIDS Research Center, Massachusetts General Hospital and Division of AIDS, Harvard Medical School, 2 Brigham and Women s Hospital, Division of Infectious Diseases, and 3 Fenway Community Health Care Center, Boston, Massachusetts; 4 Howard Hughes Medical Institute, Chevy Chase, Maryland; and 5 Monogram Biosciences, South San Francisco, California Background. Spontaneous control of human immunodeficiency virus (HIV) infection has been documented in a minority of HIV-infected individuals. The mechanisms behind this outcome remain largely unknown, and a better understanding of them will likely influence future vaccine strategies. Methods. HIV-specific T cell and antibody responses as well as host genetics were examined in untreated HIVinfected patients who maintain comparatively low plasma HIV RNA levels (hereafter, controllers), including those with levels of 50 RNA copies/ml (elite controllers, n 64), those with levels of 50 2000 copies/ml (viremic controllers, n 60); we also examined HIV-specific T cell and antibody responses as well as host genetics for patients with levels of 10,000 copies/ml (chronic progressors, n 30). Results. CD8 T cells from both controller groups preferentially target Gag over other proteins in the context of diverse HLA class I alleles, whereas responses are more broadly distributed in persons with progressive infection. Elite controllers represent a distinct group of individuals who have significantly more CD4 and CD8 T cells that secrete interferon- and interleukin-2 and lower levels of HIV-neutralizing antibodies. Individual responses were quite heterogeneous, and none of the parameters evaluated was uniquely associated with the ability to control viremia. Conclusions. Elite controllers are a distinct group, even when compared to persons with low level viremia, but they exhibit marked genetic and immunologic heterogeneity. Even low-level viremia among HIV controllers was associated with measurable T cell dysfunction, which has implications for current prophylactic vaccine strategies. Received 2 July 2007; accepted 16 August 2007; electronically published 5 February 2008. Potential conflicts of interest: none reported. Financial support: National Institutes of Health (to B.D.W, E.S.R, and D.E.K), the Howard Hughes Medical Institute (to B.D.W, F.P, and A.T) and a gift from the Mark and Lisa Schwartz Foundation. Reprints or correspondence: Bruce Walker, MD, Partners AIDS Research Center, Massachusetts General Hospital, 149 13th Street, Boston, MA 02129 (bwalker@partners.org). The Journal of Infectious Diseases 2008; 197:563 71 2008 by the Infectious Diseases Society of America. All rights reserved. 0022-1899/2008/19704-0011$15.00 DOI: 10.1086/526786 After more than 25 years of HIV infection, some persons remain asymptomatic with normal CD4 T cell counts and undetectable plasma viral loads, despite having never been treated with antiretroviral medications [1]. This extraordinary outcome provides hope that HIV infection may be durably controlled and that the development of a vaccine that prevents disease progression may be possible. Indeed, the accepted strategy for firstgeneration HIV vaccines is to provide protection from disease progression rather than to provide sterilizing immunity [2], because induction of broadly cross-reactive neutralizing antibodies (NAb) to prevent infection remains an elusive goal [3]. A major question for current vaccine strategies is the level of viremia that would be an acceptable goal following subsequent infection, to prevent transmission and disease progression. Insight into this question can be derived from HIV transmission studies [4 6]. Among a cohort of 415 discordant couples, none of the 51 subjects with serum viral loads of 1500 RNA copies/ml transmitted virus to their partners [7]. Viral load is the major independent risk factor for maternal-fetal transmission; a viral load of 1000 copies/ml is associated with a 12-fold increased risk of transmission. Viral load also clearly impacts disease progression; viral loads of less than 1000 2000 RNA copies/ml are associated with long-term AIDS- Spontaneous Control of HIV Infection JID 2008:197 (15 February) 563

free survival [8, 9]. These data suggest that controlling viral replication in this range would successfully limit the progression and the transmission of HIV infection. Although studies of persons who control HIV replication have been performed [10 14], interpretation has been limited by viral and host genetic heterogeneity in the small numbers included in most cohorts. Another limitation is that previous studies have addressed this subject largely through the study of long-term nonprogressors, a group that is usually defined by length of survival without AIDS, rather than on the basis of viral load. There have been no large cohort studies that have systematically examined adaptive immune responses and host genetics in persons who are able to control viremia, the study of whom has important implications for current vaccine strategies designed to induce a controller phenotype [15]. In the present study we examined cellular and humoral immune responses and host genetics in 126 persons who have been able to achieve control of HIV infection at plasma HIV RNA levels of 2000 copies/ml, all of whom had had documented infection for at least 1 year, with a median duration of infection of 16 years. We classified subjects into 2 controller subgroups: a group that maintains viral loads below the level of detection as measured by currently available ultrasensitive assays ( 50 or 75 copies/ml; hereafter, elite controllers ); and a second group that maintains viral loads between 50 and 2000 copies/ ml, whom we call viremic controllers. We compared these individuals to untreated persons with progressive infection, with the goal of defining immunologic and host genetic factors associated with control of HIV infection. METHODS Study subjects. HIV controllers were defined as described below and were recruited on the basis of viral load from outpatient clinics at local Boston hospitals; they were also referred from providers throughout the United States, after institutional review board approval and written informed consent. Elite controllers were defined as having without antiretroviral (ARV) therapy plasma HIV RNA levels below the level of detection for the available ultrasensitive assays (e.g., 75 copies/ml by bdna or 50 copies/ml by ultrasensitive PCR). Viremic controllers were defined as having without ARV therapy plasma HIV RNA levels between 50 and 2000 copies/ml; chronic progressors were recruited locally on the basis of viral load and were defined as having withouth ARV therapy plasma HIV RNA levels above 10,000 copies/ml (table 1). To qualify for the study, all subjects had to have a minimum of 3 determinations of plasma HIV RNA spanning at least a 12-month period. For elite controllers, isolated episodes of viremia with plasma levels up to 1,000 copies/ml were allowed if there were no consecutive viremia episodes and such episodes represented the minority of all available determinations. For viremic controllers, isolated episodes of viremia 2,000 copies/ml were allowed if such episodes represented the minority of all available determinations. For chronic progressors, episodes of viremia 10,000 copies were allowed as long as such episodes represented the minority of all available determinations. HLA typing and determination of CCR5 32 and CCR2-64I genotypes. Intermediate to high resolution HLA class I typing was performed for all patients by sequence-specific primer polymerase chain reaction as described elsewhere [16]. The presence of CCR5 32 and CCR2-64I was determined for all HIV controllers (i.e., both elite and viremic); reverse transcriptase polymerase chain reaction was performed as described elsewhere [17, 18] Assessment of HIV-specific CD8 T cell responses. Interferon (IFN)- enzyme-linked immunospot (Elispot) assays performed with peripheral blood mononuclear cells (PBMC) were performed as described elsewhere [16, 19]; these assays included 410 peptides 16 19 amino acids in length and overlapping by 10 amino acids, which spanned the entire HIV proteome. All positive responses were confirmed in a second assay. Responses were regarded as positive if they had at least 3 times the mean number of spot-forming cells (sfc) in the 3 negative control wells; positive responses also had to have 50 sfc/10 6 PBMCs. Responses to peptides had previously been shown to be largely mediated by CD8 T cells [19, 20]. Flow cytometric detection of antigen-induced intracellular IFN- and interleukin (IL)-2. Intracellular cytokine staining assays were performed as described elsewhere [21]. PBMC were incubated with HIV-1 Gag, Nef, Pol, and Env peptide pools and a combination pool of Tat, Rev, Vif, Vpr, and Vpu (TRVVV) peptides. For analysis, between 70,000 and 125,000 CD4 and CD8 cells were collected. Ten HIV-seronegative donors were screened for T cell responses by use of the same protocol. Among these individuals, 0.01% of CD4 or CD8 T cells responded to all HIV peptide pools (0.0092% for IL-2, 0.0014% for IL-2 IFN-, and 0.0028% for IFN- ). Responses were considered positive if they were 0.01% and at least 2 times the value of the unstimulated control. Neutralizing antibody assay. The antibody neutralization assay is a modification of the Monogram Biosciences HIV coreceptor tropism assay [22]. Briefly, HIV envelope sequences were amplified from plasma and transferred to an expression vector; pseudotyped viruses were generated by contransfection of HEK- 293 cells with the envelope expression vector and a genomic HIV vector which contains a luciferase gene in place of env. Virus infectivity was determined by measuring the amount of luciferase activity expressed in infected cells. Neutralizing activity was reported as the concentration or dilution of each plasma required to inhibit the virus by 50%. Statistical analysis. Statistical analyses and graphical presentations were performed using Prism (version 3.0, GraphPad). When the 3 groups were compared, determinations of significance were made on the basis of 1-way analysis of variance by use of a nonparametric Kruskal-Wallis test; when 2 groups were 564 JID 2008:197 (15 February) Pereyra et al.

Table 1. Patient characteristics and distribution of class I HLA alleles associated with disease progression. Parameter Elite controllers Viremic controllers Chronic progressors Subjects, no. 66 60 30 Age, mean (range), years 47 (22 75) 48 (32 67) 36 (18 70) Sex Male 42 (64) 47 (78) 24 (80) Female 24 (36) 13 (22) 6 (20) Race White 36 (54) 35 (58) 17 (57) Black 19 (29) 13 (22) 3 (10) Hispanic 9 (14) 5 (8) 8 (27) Other 2 (3) 7 (12) 2 (6) Plasma HIV RNA level, median (IQR), copies/ml Below detection 770 (348 1500) 152,000 (67,050 278,000) CD4 cell count, median (IQR), cells/mm 3 884 (641 1149) 602 (451 786) a 295 (203 455) Duration of HIV diagnosis, median (IQR), years 15 (9 22) 17 (13 25) 5.5 (1 17) CCR5 32 genotype WT 57 (86) 47 (78) NA HT 9 (14) 13 (22) CCR2-64I genotype WT 56 (85) 47 (78) NA HT 9 (14) 9 (15) HM 1 (1) 4 (7) HLA-B*57 29 (44) 20 (33) 3 (10) HLA-B*27 10 (15) 12 (20) 1 (3) All protective HLA alleles b 45 (68) 36 (60) 11 (37) Combined protective HLA alleles and chemokine receptors 48 (73) 43 (72) NA NOTE. Data are no. (%) of subjects, unless otherwise indicated. See Methods for details about the 3 study groups. WT, wild type; HT, heterozygous; HM, homozygous; NA, not applicable. a Elite controllers compared to viremic controllers, P.001 by nonparametric Mann-Whitney test. b Includes B*57, 27, 5801, 1503, 13 and 51. compared, a 2-tailed nonparametric Mann-Whitney test was used. A nonparametric Spearman test was used to calculate correlations. RESULTS Heterogeneity of host genetic factors associated with delayed disease progression among elite and viremic controllers. A total of 66 elite controllers and 60 viremic controllers were compared to 30 chronic progressors recruited over the same period (table 1). The absolute CD4 count differed among the 3 groups and was higher in elite controllers, compared with viremic controllers (median, 884 vs. 602 cells/mm 3, P.001). The frequency of HLA-B*57 alleles differed among the groups (P.01), but there was no difference in B*57 carriage between elite and viremic controllers (P.223; table 1), and only 44% of elite and 33% of viremic controller groups carried this allele, a frequency much less than that previously reported among smaller cohorts [13, 23]. Including other HLA alleles that have been associated with delayed disease progression [24, 25] in the analysis (B*5801, HLA-B*27, B*1503, B13, and B51), 68% of elite controllers and 60% of viremic controllers carry at least one HLA allele that has been associated with better prognosis, compared with 37% of chronic progressors (P.025), leaving almost one-third of HIV controllers without any known relatively protective HLA alleles. We also evaluated the frequency of CCR5 and CCR2 polymorphisms that are associated with slower disease progression (table 1) [17, 26]. No controllers, either elite or viremic, were homozygous for CCR5 32, whereas heterozygosity was present in 14% of elite controllers and 22% of viremic controllers (P.44), a frequency that is not significantly different from that reported among the population of individuals not infected with HIV [27]. Homozygosity for the CCR2-64I genotype also did not differ among the groups (P.84). These data indicate that although some HLA alleles are enriched in HIV controllers, neither HLA class I alleles nor well-characterized chemokine receptor polymorphisms are predictive of the control of viremia in the absence of therapy. Elite control is associated with the lowest breadth and magnitude of HIV-specific CD8 T cells. We next evaluated CD8 T cell responses in a randomly selected subgroup of 45 elite Spontaneous Control of HIV Infection JID 2008:197 (15 February) 565

Figure 1. Comparison of HIV-specific CD8 T cell responses to the entire expressed HIV-1 genome among 45 elite controllers (EC), 25 viremic controllers (VC), and 25 chronic progressors (CP). Black boxes, individual values for EC; white triangles, individual values for VC; gray circles, individual values for CP. Total magnitude (A) of the CD8 T cell responses measured by Elispot and given in spot-forming cells (sfc)/10 6 PBMC and total breadth (B) of the CD8 T cell responses given in number of responses against all HIV-1 protein subunits. After exclusion of B57 patients, total magnitude (C) and total breadth (D) of responses were compared for 25 EC, 17 VC, and 22 CP. Horizontal black bars, median numbers for each group. P values for comparison between 2 groups shown with horizontal black bars were calculated using a 2-tailed Mann-Whitney test. P value for comparison between 3 groups shown in black box was calculated using the Kruskal-Wallis nonparametric test. controllers, 25 viremic controllers, and 25 chronic progressors, by use of a panel of overlapping HIV peptides that spanned all viral proteins in an IFN- Elispot assay [28]. HIV-specific, IFN- producing T cells were detected in all subjects. The magnitude and breadth of T cell responses differed among elite controllers, viremic controllers, and chronic progressors (median magnitude, 5428, 6253, and 8300 sfc/10 6 PBMC, respectively; median breadth, 15, 19, and 27 responses, respectively; P.007 and.009, respectively) (figure 1A and 1B). Elite controllers demonstrated lower magnitude and breadth of responses, compared with chronic progressors (P.003 and.004, respectively), and there was a trend suggesting differences in magnitude and breadth between elite controllers and viremic controllers (P.053 and.067, respectively), but not between viremic controllers and chronic progressors. These relationships persisted even when individuals carrying HLA-B*57 were removed from the analysis (figure 1C and 1D). Remarkably, both the strongest and weakest HIV-specific T cell responses were detected in the elite controllers, with the magnitude of responses ranging from 500 to 25,000 sfc/10 6 PBMC, and breadth ranging from 2 peptides to 101 peptides. Comparing elite and viremic controllers with favorable and unfavorable HLA alleles and/or coreceptor polymorphisms, we found no difference in breadth (P.149)ormagnitude of responses (P.101)between the 2 groups (data not shown). Overall, these data show that although elite control is typically associated with less robust CD8 T cell responses as measured by IFN- Elispot assay, compared with those with even low-level viremia, there is marked heterogeneity in the breadth and magnitude of these responses among HIV controllers. Gag is preferentially targeted by CD8 T cell responses in HIV controllers. We next determined the relative contribution of responses to individual proteins to the total virus-specific CD8 T cell activity. These analyses demonstrated significant differences in the responses to the protein subunits among the groups (figure 2A and B). Both elite and viremic controller groups were distinguished from chronic progressors by preferential targeting of Gag. In contrast, there was comparable targeting of Gag, Pol, and Nef in persons with progressive infection, and significantly more targeting of Env. In elite and viremic controllers, Gag responses contributed most to the total magnitude (median, 42% and 48%, respectively) and total breadth (median, 41% in both groups), values that were significantly more than those observed in chronic progressors (Gag contribution to magnitude, 27%; P.01 and.003 respectively; Gag contribution to breadth, 27%; P.001 and 0.004, respectively). Similar results were also found when we analyzed just non-b*57 expressing individuals (figure 2 panels C and D). These data indicate that preferential targeting of Gag, rather than total magnitude or breadth of responses or specific HLA class I alleles, is 566 JID 2008:197 (15 February) Pereyra et al.

Figure 2. HIV-specific CD8 T cell responses targeting the protein subunits Gag, Nef, Pol, Env, and accessory proteins Tat, Rev, Vpr, Vif, and Vpu (TRVVV) among elite controllers (EC), viremic controllers (VC), and chronic progressors (CP). Comparison of the total magnitude of responses (A) and (B) total breadth of the CD8 T cell responses targeting each individual protein (color coded as shown). Magnitude (C) and breadth (D) after exclusion of B*57 patients. For both panels, the median and interquartile range is shown. Horizontal black bars, P values for comparison between all proteins in each study group, calculated by use of the Kruskal-Wallis nonparametric test. For each study group, asterisks are placed above Nef, Pol, Env, and TRVVV responses that are significantly lower than Gag responses (*P.05, **P.005, ***P.001). P values for comparison between 2 groups shown with asterisks were calculated using a 2-tailed Mann-Whitney test. associated with the control of HIV infection, which is similar to recent findings with respect to chronic untreated clade C infection [29]. Higher frequency of HIV-specific CD4 and CD8 T cells producing IFN- and IL-2 in elite controllers. We next analyzed the qualitative function of HIV specific CD4 and CD8 T cell responses as measured by their ability to secrete IFN- and/or IL-2. For this analysis, 14 elite controllers and 11 viremic controllers were randomly selected and compared with 12 chronic progressors who were selected on the basis of relatively preserved CD4 cell counts, to limit the potential bias that low peripheral CD4 counts might have on measurement of T cell function. Median CD4 counts were 884, 776, and 436 cells/mm 3, respectively; median viral load was 50, 985, and 86,000 copies, respectively. Although elite controllers had a higher percentage of circulating CD4 cells that secrete both IL-2 and IFN- (median, 0.06%; range, 0.01 1.2), compared with viremic controllers (median, 0.02%; range, 0 0.44; P.03) and chronic progressors (median, 0.01%; range, 0 0.38; P.02), there was marked heterogeneity among the subjects, with some entirely lacking these responses (figure 3A). No difference was found in CD4 cells that secrete IFN- alone among the 3 groups (data not shown). Elite controllers also had a higher percentage of circulating CD8 cells that secreted both IL-2 and IFN- (median, 0.05%; range, 0.01 0.23), compared with viremic controllers (median, 0%; range, 0 0.08; P.02) and progressors (median, 0%; range, 0 0.19; P.02) (figure 3B). In line with the Elispot data, there was a trend for a lower percentage of circulating CD8 cells that secrete IFN- alone in elite controllers, compared with viremic controllers and chronic progressors; this difference did not reach statistical significance in this smaller subset (median percentage, 0.75%, 1.7%, and 2.4%, respectively; P.06;data not shown). Thus, despite significantly lower CD8 T cell responses in elite controllers by Elispot assay, CD4 and CD8 T cell responses that produced both IL-2 and IFN- were higher, indicating enhanced T cell function in these persons. Paucity of broadly cross-reactive NAb in elite controllers. We next quantified autologous and/or heterologous NAb among 17 elite controllers, 17 viremic controllers, and 20 chronic progressors, randomly selected. Plasma was tested against 2 laboratory HIV-1 strains, 14 viruses isolated from the viremic controllers, and 15 viruses isolated from the chronic Spontaneous Control of HIV Infection JID 2008:197 (15 February) 567

progressors. Plasma from viremic controllers and chronic progressors exhibited extensive cross-neutralization of nearly all of the viruses from both viremic controllers and chronic progressors (figure 4) as well as high-level inhibition of JRCSF and laboratory strains NL43 and SF162 (data not shown). Despite the dramatic differences in viral load and CD4 counts between these 2 groups, there was no statistically significant difference between them in terms of NAb titers (P.50). In contrast, as a group, elite controllers had significantly lower-level NAb responses to primary viruses from viremic controllers and JRCSF, as well as lower-level responses to laboratory strains (P.001). However, once again there was marked heterogeneity among elite controllers, some individuals having broad NAb responses, and others having minimal or no NAb. Neutralization of autologous virus was consistently present and was equally strong in viremic controllers and chronic progressors. These data indicate that there is marked heterogeneity in not only cellular immune responses but also in humoral immune responses among elite and viremic controllers. DISCUSSION Figure 3. Proportion of total HIV-specific CD4 and CD8 T cells secreting both interferon (IFN)- and interleukin (IL)-2, as well as predominant HIV specific T cell population responses. Proportion is measured in each group by adding individual positive responses to Gag, Nef, Pol, Env, and Tat, Rev, Vpr, Vif, and Vpu (TRVVV) peptide pools. Black boxes, individual values for elite controllers (EC); white triangles, individual values for viremic controllers (VC); gray circles, individual values for chronic progressors (CP). A, Percentage of circulating polyfunctional CD4 cells; B, percentage of circulating polyfunctional CD8 T cells. For both A and B, horizontal black bars, median numbers for each group. P values for comparison between 2 groups shown with horizontal black bars were calculated using a 2-tailed Mann-Whitney test. In panels A and B, the P value for comparison between 3 groups shown in black box was calculated using the Kruskal-Wallis nonparametric test. There was no difference between the groups in terms of IFN- or IL-2 secretion, only CD4 or CD8 T cells (data not shown). C, Predominant HIV-specific T cell population responses is shown as the ratio of CD4 T cells to CD8 T cells secreting IL-2, IFN- or both to all tested protein pools (Gag, Nef, Pol, Env, TRVVV). The ratio is highest forec compared to VC and CP. P values for comparison between 2 groups shown with horizontal black bars were calculated using a 2-tailed t test. Quartiles and median numbers are shown in panel C; error bars (whiskers), range of values for each group. Current HIV vaccine strategies are focused on the induction of immune responses that will protect from disease progression rather than prevent infection, and thus it is critical to understand the immune parameters associated with persistent control of HIV. At a level of viremia of 2000 HIV RNA copies/ml, both transmission and progression would be expected to be markedly diminished, suggesting a target level for current vaccine efforts, and indicating a need to understand the immunologic and host genetic features associated with this level of control. Our results indicate that, as a group, persons who control viremia to 2000 RNA copies/ml in the absence of ARV therapy are distinct from persons with progressive infection, with preferential targeting of the HIV Gag protein and enrichment for specific HLA B alleles in both controller groups. Moreover, elite controllers are more likely than viremic controllers and chronic progressors to have CD4 and CD8 T cells that secrete IFN- and IL-2 and to have significantly less robust HIV-specific CD8 T cell responses by IFN- Elispot and less robust NAb. However, individual responses were quite heterogeneous, and none of these parameters was predictive of this phenotype. Moreover, responses were not simply driven by viral load, because the elite controller group included persons that had both the strongest and the weakest, as well as the broadest and the narrowest, CD8 T cell responses. Potential explanations for these differences might be the sequence of the infecting virus and immunologic escape due to viral mutations, which were not assessed in this study. Compared to previous smaller studies [30], one striking finding in the present study is that the majority of HIV controllers did not carry HLA-B*57; likewise specific chemokine polymorphisms associated with slower disease progression were not 568 JID 2008:197 (15 February) Pereyra et al.

Figure 4. Summary of antibody neutralization titers and relationship between breadth and potency of antibody responses and viral load among study groups. Pseudovirions isolated from viremic controllers (VC) and chronic progressors (CP) were used in neutralization assay against plasma from all study groups. The median concentration or dilution of each plasma required to inhibit the virus by 50% (IC 50 ) from each plasma tested with VC and CP viruses is shown in A. Elite controllers (EC) group had significantly lower neutralization of all viruses than viremic controllers (VC) and chronic progressors (CP) (*** P.001). P values shown with asterisks were calculated using a 2-tailed Mann-Whitney test; viruses from chronic progressors are equally susceptible to neutralization, compared with viruses from viremic controllers. Breadth and potency of antibody responses does not correlate directly with viral load. B, Viral load is graphed with the median neutralizing antibody IC 50 for each plasma tested with viruses from viremic controllers and chronic infected individuals; black circles, plasma from each group of patients. Viral load is graphed with the median IC 50 for plasma of chronic progressors (C) and viremic controllers (D). Correlations are calculated using a nonparametric Spearman test. more pronounced. If the carriage of HLA alleles reported to be protective are considered in combination with the presence of chemokine receptor polymorphisms associated with delayed disease progression, more than 25% of elite and viremic controllers lack these host genetic factors, which might be anticipated to at least partially contribute to their improved disease outcome. Moreover, the preferential targeting of Gag was independent of the presence of the HLA-B*57 allele, suggesting the key role of Gag targeting in durable immune control of HIV, perhaps as a result of selection for mutations that alter viral fitness [31] or of early targeting of infected cells due to immediate processing of preformed Gag protein on viral entry [32]. Given that Gag is the preferred antigen produced by the virus and incorporated into virions, one could hypothesize that increased levels of viremia result in an increased number of immune responses targeting non-gag proteins. This seems an unlikely explanation for our findings, given that the overall breadth and magnitude of responses, as well as the ability to secrete multiple cytokines, is not different between viremic controllers and chronic progressors, yet the former clearly target Gag preferentially. Spontaneous Control of HIV Infection JID 2008:197 (15 February) 569

Many recent studies have focused on identifying the functional characteristics of CD8 and CD4 T cell-mediated immune control [11, 14, 33 35], including the ability of T cells to proliferate as a result of stimulation and secrete multiple cytokines [33, 34, 36] and the presence of IL-2 secreting CD4 and CD8 T cells of a central memory phenotype [35, 37, 38]. Our data in the present study show that elite controllers, despite having lower overall IFN- secretion, have higher numbers of HIV-specific CD4 and CD8 T cells that secrete both IL-2 and IFN-, compared to both viremic controllers and chronic progressors. These responses may contribute to the enhanced ability of CD8 T cells from elite controllers to inhibit virus replication in vitro [39]. In addition, elite controllers have the highest ratio of functional CD4 to CD8 T cells and chronic progressors have the lowest ratio (figure 3C). These data confirm and extend the results of previously published studies [11, 14, 37] in which T cell polyfunctionality, as measured by secretion of multiple cytokines, was found to be higher among individuals who had lower levels of HIV replication, and suggest that even low-level viremia has a negative effect on T cell function. However, whether this association between T cell function and HIV control is cause or effect cannot be addressed by these or other human studies [11, 14]. In previous studies, HIV elite controllers were compared to noncontrollers and, therefore, the effects of chronic viremia on T cells could account for the differences observed [14]. Even when the effects of viremia are controlled in studies that compare elite controllers to successfully treated individuals [11], one cannot rule out that the differences seen are a consequence of immunologic damage during the viremic period that usually leads to initiation of ARV therapy. The role of NAb in the control of viral replication has been controversial. Some [10, 40 42], but not all [43], early studies showed higher titers of heterologous NAb in long-term nonprogressors, compared with individuals who had chronic progressive disease. More recently, in studies largely focused on persons who expressed HLA-B*57, their heterologous NAb levels were shown to be significantly lower than those observed in individuals with viremia 10,000 copies, but no difference was found in NAb titers against autologous, contemporaneous plasma viruses [44]. Our findings of lower NAb activity among elite controllers are in agreement with these data [44], which suggests that NAb do not play a major role in the maintenance of viral suppression in individuals who spontaneously control viral replication. Our findings among viremic controllers and chronic progressors suggest that any level of viremia maintains broad autologous and heterologous cross-neutralizing antibodies. In summary, this study of 126 persons who controlled HIV viremia to levels that may be expected to reduce transmission and disease progression indicates that, although elite and viremic controllers share some immunologic features and are distinct from persons with chronic progressive infection, the elite controllers are a distinct subgroup. However, even within these groups there is substantial heterogeneity in all of the parameters studied, which suggests that there are as-yet-undefined viral or host factors or combinations of factors that contribute to this remarkable phenotype. The lack of a definable immunologic marker for durable control and the limitations that small cohorts pose in defining host genetics provides a rationale for the use of newer approaches in defining this phenotype. Such studies, including extensive viral and host genetic analyses, are currently underway through the HIV Controller Consortium, which seeks to enroll 1000 elite controllers and 1000 viremic controllers in order to perform a whole-genome association study to define the genetic basis for this remarkable phenotype. This collaborative effort will extend the current study in the hope of defining disease pathways that not only predict durable control of infection but can also lead to new interventions to achieve this state in persons with otherwise-progressive HIV infection. References 1. Buchbinder SP, Katz MH, Hessol NA, O Malley PM, Holmberg SD. Long-term HIV-1 infection without immunologic progression. AIDS 1994; 8:1123 8. 2. Lambert M, Gannage M, Karras A, et al. Differences in the frequency and function of HHV8-specific CD8 T cells between asymptomatic HHV8 infection and Kaposi sarcoma. Blood 2006; 108:3871 80. 3. Haynes BF, Montefiori DC. Aiming to induce broadly reactive neutralizing antibody responses with HIV-1 vaccine candidates. Expert Rev Vaccines 2006; 5:347 63. 4. Quinn TC, Wawer MJ, Sewankambo N, et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group. N Engl J Med 2000; 342:921 9. 5. Krogstad P, Lee S, Johnson G, et al. Nucleoside-analogue reversetranscriptase inhibitors plus nevirapine, nelfinavir, or ritonavir for pretreated children infected with human immunodeficiency virus type 1. Clin Infect Dis 2002; 34:991 1001. 6. Study EC. Mother-to-child transmission of HIV infection in the era of highly active antiretroviral therapy. Clin Infect Dis 2005; 40:458 65. 7. Gray RH, Wawer MJ, Brookmeyer R, et al. Probability of HIV-1 transmission per coital act in monogamous, heterosexual, HIV-1-discordant couples in Rakai, Uganda. Lancet 2001; 357:1149 53. 8. Giorgi JV, Lyles RH, Matud JL, et al. Predictive value of immunologic and virologic markers after long or short duration of HIV-1 infection. J Acquir Immune Defic Syndr 2002; 29:346 55. 9. O Brien TR, Blattner WA, Waters D, et al. Serum HIV-1 RNA levels and time to development of AIDS in the Multicenter Hemophilia Cohort Study. Jama 1996; 276:105 10. 10. Cao Y, Qin L, Zhang L, Safrit J, Ho DD. Virologic and immunologic characterization of long-term survivors of human immunodeficiency virus type 1 infection. N Engl J Med 1995; 332:201 8. 11. Emu B, Sinclair E, Favre D, et al. Phenotypic, functional, and kinetic parameters associated with apparent T-cell control of human immunodeficiency virus replication in individuals with and without antiretroviral treatment. J Virol 2005; 79:14169 78. 12. Bailey JR, Williams TM, Siliciano RF, Blankson JN. Maintenance of viral suppression in HIV-1-infected HLA-B*57 elite suppressors despite CTL escape mutations. J Exp Med 2006; 203:1357 69. 13. Migueles SA, Sabbaghian MS, Shupert WL, et al. HLA-B*5701 is highly associated with restriction of virus replication in a subgroup of HIVinfected long term nonprogressors. Proc Natl Acad SciUSA2000; 97: 2709 14. 14. Betts MR, Nason MC, West SM, et al. HIV nonprogressors preferentially 570 JID 2008:197 (15 February) Pereyra et al.

maintain highly functional HIV-specific CD8 T cells. Blood 2006; 107: 4781 9. 15. Johnston MI, Fauci AS. An HIV vaccine evolving concepts. N Engl J Med 2007; 356:2073 81. 16. Frahm N, Korber BT, Adams CM, et al. Consistent cytotoxic-tlymphocyte targeting of immunodominant regions in human immunodeficiency virus across multiple ethnicities. J Virol 2004; 78:2187 200. 17. Anzala AO, Ball TB, Rostron T, O Brien SJ, Plummer FA, Rowland- Jones SL. CCR2 64I allele and genotype association with delayed AIDS progression in African women. University of Nairobi Collaboration for HIV Research. Lancet 1998; 351:1632 3. 18. Wu L, Paxton WA, Kassam N, et al. CCR5 levels and expression pattern correlate with infectability by macrophage-tropic HIV-1, in vitro. J Exp Med 1997; 185:1681 91. 19. Addo MM, Yu XG, Rathod A, et al. Comprehensive epitope analysis of human immunodeficiency virus type 1 (HIV-1)-specific T-cell responses directed against the entire expressed HIV-1 genome demonstrate broadly directed responses, but no correlation to viral load. J Virol 2003; 77:2081 92. 20. Feeney ME, Roosevelt KA, Tang Y, et al. Comprehensive screening reveals strong and broadly directed human immunodeficiency virus type 1-specific CD8 responses in perinatally infected children. J Virol 2003; 77:7492 501. 21. Betts MR, Ambrozak DR, Douek DC, et al. Analysis of total human immunodeficiency virus (HIV)-specific CD4( ) and CD8( ) T-cell responses: relationship to viral load in untreated HIV infection. J Virol 2001; 75:11983 91. 22. Richman DD, Wrin T, Little SJ, Petropoulos CJ. Rapid evolution of the neutralizing antibody response to HIV type 1 infection. Proc Natl Acad Sci U S A 2003; 100:4144 9. 23. Lambotte O, Boufassa F, Madec Y, et al. HIV controllers: a homogeneous group of HIV-1-infected patients with spontaneous control of viral replication. Clin Infect Dis 2005; 41:1053 6. 24. Frahm N, Kiepiela P, Adams S, et al. Control of human immunodeficiency virus replication by cytotoxic T lymphocytes targeting subdominant epitopes. Nat Immunol 2006; 7:173 8. 25. Honeyborne I, Prendergast A, Pereyra F, et al. Control of human immunodeficiency virus type 1 is associated with HLA-B*13 and targeting of multiple gag-specific CD8 T-cell epitopes. J Virol 2007; 81:3667 72. 26. Dragic T, Litwin V, Allaway GP, et al. HIV-1 entry into CD4 cells is mediated by the chemokine receptor CC-CKR-5. Nature 1996; 381: 667 73. 27. Lucotte G, Mercier G. Distribution of the CCR5 gene 32-bp deletion in Europe. J Acquir Immune Defic Syndr Hum Retrovirol 1998; 19:174 7. 28. Novitsky V, Cao H, Rybak N, et al. Magnitude and frequency of cytotoxic T-lymphocyte responses: identification of immunodominant regions of human immunodeficiency virus type 1 subtype C. J Virol 2002; 76:10155 68. 29. Kiepiela P, Leslie AJ, Honeyborne I, et al. Dominant influence of HLA-B in mediating the potential co-evolution of HIV and HLA. Nature 2004; 432:769 75. 30. Carrington M, Nelson GW, Martin MP, et al. HLA and HIV-1: heterozygote advantage and B*35-Cw*04 disadvantage. Science 1999; 283:1748 52. 31. Martinez-Picado J, Prado JG, Fry EE, et al. Fitness cost of escape mutations in p24 Gag in association with control of human immunodeficiency virus type 1. J Virol 2006; 80:3617 23. 32. Sacha JB, Chung C, Rakasz EG, et al. Gag-specific CD8 T lymphocytes recognize infected cells before AIDS-virus integration and viral protein expression. J Immunol 2007; 178:2746 54. 33. Migueles SA, Laborico AC, Shupert WL, et al. HIV-specific CD8 T cell proliferation is coupled to perforin expression and is maintained in nonprogressors. Nat Immunol 2002; 3:1061 8. 34. Lichterfeld M, Kaufmann DE, Yu XG, et al. Loss of HIV-1-specific CD8 T cell proliferation after acute HIV-1 infection and restoration by vaccine-induced HIV-1-specific CD4 T cells. J Exp Med 2004; 200: 701 12. 35. Zimmerli SC, Harari A, Cellerai C, Vallelian F, Bart PA, Pantaleo G. HIV-1-specific IFN- /IL-2-secreting CD8 T cells support CD4- independent proliferation of HIV-1-specific CD8 T cells. Proc Natl Acad Sci U S A 2005; 102:7239 44. 36. Rosenberg ES, Billingsley JM, Caliendo AM, et al. Vigorous HIV-1- specific CD4 T cell responses associated with control of viremia. Science 1997; 278:1447 50. 37. Harari A, Petitpierre S, Vallelian F, Pantaleo G. Skewed representation of functionally distinct populations of virus-specific CD4 T cells in HIV- 1-infected subjects with progressive disease: changes after antiretroviral therapy. Blood 2004; 103:966 72. 38. Younes SA, Yassine-Diab B, Dumont AR, et al. HIV-1 viremia prevents the establishment of interleukin 2-producing HIV-specific memory CD4 T cells endowed with proliferative capacity. J Exp Med 2003; 198: 1909 22. 39. Saez-Cirion A, Lacabaratz C, Lambotte O, et al. HIV controllers exhibit potent CD8 T cell capacity to suppress HIV infection ex vivo and peculiar cytotoxic T lymphocyte activation phenotype. Proc Natl Acad Sci USA2007; 104:6776 81. 40. Montefiori DC, Pantaleo G, Fink LM, et al. Neutralizing and infectionenhancing antibody responses to human immunodeficiency virus type 1 in long-term nonprogressors. J Infect Dis 1996; 173:60 7. 41. Pantaleo G, Menzo S, Vaccarezza M, et al. Studies in subjects with longterm nonprogressive human immunodeficiency virus infection. N Engl J Med 1995; 332:209 16. 42. Pilgrim AK, Pantaleo G, Cohen OJ, et al. Neutralizing antibody responses to human immunodeficiency virus type 1 in primary infection and long-term-nonprogressive infection. J Infect Dis 1997; 176:924 32. 43. Harrer T, Harrer E, Kalams SA, et al. Strong cytotoxic T cell and weak neutralizing antibody responses in a subset of persons with stable nonprogressing HIV type 1 infection. AIDS Res Hum Retroviruses 1996; 12: 585 92. 44. Bailey JR, Lassen KG, Yang HC, et al. Neutralizing antibodies do not mediate suppression of human immunodeficiency virus type 1 in elite suppressors or selection of plasma virus variants in patients on highly active antiretroviral therapy. J Virol 2006; 80:4758 70. Spontaneous Control of HIV Infection JID 2008:197 (15 February) 571