|Home | About | Journals | Submit | Contact Us | Français|
Although untreated human immunodeficiency virus (HIV)–infected patients maintaining undetectable plasma HIV RNA levels (elite controllers) have high HIV-specific immune responses, it is unclear whether they experience abnormal levels of T cell activation, potentially contributing to immunodeficiency.
We compared percentages of activated (CD38+HLA-DR+) T cells between 30 elite controllers, 47 HIV-uninfected individuals, 187 HIV-infected individuals with undetectable viremia receiving antiretroviral therapy (antiretroviral therapy suppressed), and 66 untreated HIV-infected individuals with detectable viremia. Because mucosal translocation of bacterial products may contribute to T cell activation in HIV infection, we also measured plasma lipopolysaccharide (LPS) levels.
Although the median CD4+ cell count in controllers was 727 cells/mm3, 3 (10%) had CD4+ cell counts <350 cells/mm3 and 2 (7%) had acquired immunodeficiency syndrome. Controllers had higher CD4+ and CD8+ cell activation levels (P < .001 for both) than HIV-negative subjects and higher CD8+ cell activation levels than the antiretroviral therapy suppressed (P = .048). In controllers, higher CD4+ and CD8+ T cell activation was associated with lower CD4+ cell counts (P = .009 and P = .047). Controllers had higher LPS levels than HIV-negative subjects (P < .001), and in controllers higher LPS level was associated with higher CD8+ T cell activation (P = .039).
HIV controllers have abnormally high T cell activation levels, which may contribute to progressive CD4+ T cell loss even without measurable viremia.
Fewer than 1% of chronically HIV-infected individuals are capable of maintaining clinically undetectable plasma HIV RNA levels (<75 copies/mL) in the absence of antiretroviral medications [1–4]. We and others have been systematically recruiting these “elite controllers” to characterize fully the correlates of protective immunity. For example, controllers have much stronger polyfunctional HIV-specific CD8+ T cell responses than chronically HIV-infected patients maintaining antiretroviral treatment–mediated viral suppression , and they are enriched for protective class I HLA haplotypes associated with delayed disease progression [3, 6–11]. Furthermore, controllers maintain high frequencies of CD4+ T cells that secrete interleukin-2 and proliferate in response to HIV peptides [12, 13]. These observations have led to the consensus that controllers are capable of containing HIV replication at very low levels with strong and durable HIV-specific immune responses.
Much less attention has been paid to the non–HIV-specific immune function of these individuals. Indeed, although lower plasma HIV RNA levels in the absence of antiretroviral therapy predict slower rates of clinical progression , the extent of viral replication may explain less than half of the variability in the rates of subsequent CD4+ T cell decline and progression to AIDS [15, 16], and some patients maintaining undetectable or nearly undetectable plasma HIV RNA levels in the absence of antiretroviral therapy have experienced clinical progression [4, 17]. These observations suggest that factors other than the level of viral replication contribute to immunodeficiency in HIV infection. One such factor is the ability of HIV to cause generalized T cell activation. We and others have demonstrated that higher CD8+ T cell activation levels are associated with more-rapid clinical progression and CD4+ T cell decline in untreated patients [18–23] as well as fewer antiretroviral therapy–mediated CD4+ T cell gains [24–27], independent of plasma HIV RNA level. T cell activation may distinguish pathogenic from nonpathogenic simian immunodeficiency virus (SIV) infection , and generalized T cell activation in the absence of viral infection appears to be capable of inducing CD4+ T cell depletion and immunodeficiency in a murine model . To our knowledge, no study to date has specifically addressed whether controllers have abnormal levels of generalized T cell activation, potentially contributing to immunodeficiency.
To address these issues, we compared T cell activation levels between controllers and 3 comparator groups: HIV-uninfected patients at risk for HIV infection from a study of nonoccupational postexposure prophylaxis, HIV-infected patients maintaining undetectable plasma HIV RNA levels as a consequence of antiretroviral therapy, and untreated HIV-infected patients with detectable viremia. We also assessed the relationship between T cell activation and CD4+ T cell counts among controllers. Finally, we explored indirect mechanisms by which HIV might cause T cell activation despite clinically undetectable plasma HIV RNA levels. Specifically, because mucosal translocation of bacterial products may contribute to T cell activation in HIV infection , we assessed the relationship between T cell activation and plasma lipopolysaccharide (LPS) levels among controllers.
HIV-infected controllers were recruited from the San Francisco Bay Area into the Study of the Consequences of the Protease Inhibitor Era (SCOPE), a clinic-based cohort of >700 chronically HIV-infected individuals based at San Francisco General Hospital and the San Francisco Veterans Affairs Medical Center. Participants are followed every 4 months with detailed questionnaires, clinical laboratory monitoring, and biological specimen banking. Controllers were defined as having a positive HIV-1 serologic finding (including confirmatory Western blot) and an undetectable plasma HIV RNA level with clinically available assays (i.e., <75 copies/mL) without antiretroviral therapy in the preceding 12 months. Chronically HIV-infected patients with undetectable plasma HIV RNA levels during antiretroviral therapy and untreated patients with detectable viremia were also sampled from the SCOPE cohort. T cell activation was measured in a consecutive sample of SCOPE participants; a subset of these treated patients have been described elsewhere . At-risk but HIV-uninfected individuals were sampled from a previously described trial of nonoccupational postexposure prophylaxis in San Francisco . These HIV-uninfected individuals had reported a sexual or needle-sharing exposure to an individual with known or suspected HIV infection within 72 h of presentation, subsequently received ≥2 antiretroviral medications for 30 days, and had continued to have negative HIV-1 serologic results after 12 weeks. A subset of these participants had T cell activation measured at week 12 and were included in this analysis.
Plasma HIV RNA levels were determined by means of the bDNA assay (Quantiplex HIV RNA, version 3.0; Chiron Corporation) on the date of immunophenotyping. Hepatitis C virus (HCV) serostatus was determined by the Ortho HCV Enzyme Immunoassay Test System (version 3.0; Ortho Diagnostic Systems) on the day of immunophenotyping for most participants, but 10 of the controllers had HCV testing performed in the clinical laboratory before the date of immunophenotyping. The nadir CD4+ T cell count was the lowest self-reported value before the date of immunophenotyping.
T cell activation was measured in freshly collected, EDTA-anticoagulated whole blood and analyzed by 4-color flow cytometry performed with a Beckman Coulter Epics XL flow cytometer. Blood was stained on a Beckman Coulter PrepPlus and lysed on a Beckman Coulter TQ-Prep. Activated (CD38+HLADR+) T cells were identified with fluorescein isothiocyanate–conjugated anti-HLA-DR, phycoerythrin (PE)–conjugated anti-CD38, PE–cyanin red 5.1–conjugated anti-CD3 (to exclude monocyte and natural killer cells), and PE–Texas red (energy-coupled dye)–conjugated anti-CD4 or CD8. The activation markers CD38 and HLA-DR were gated from the CD4+ and CD8+ cells on a 2-dimensional dot plot in which quadrant gates, set on an isotype control, were used to define positive and negative populations. T cell activation levels were reported as the percentages of CD4+ and CD8+ T cells expressing both HLA-DR and CD38.
Plasma LPS levels were measured with the limulus amoebocyte assay (Cambrex) in thawed plasma samples diluted to 20% in endotoxin-free water and have been reported elsewhere . Plasma LPS measurements were included in the current analysis if they were obtained within 1 year of the T cell activation measurement.
Wilcoxon rank sum and Fisher's exact tests were used for unadjusted comparisons between groups. Adjusted differences between groups were assessed with linear regression, with appropriate transformation of continuous variables and calculation of SEs with heteroscedasticity-consistent covariance matrix estimators when necessary to satisfy model assumptions . All factors associated with T cell activation in unadjusted analyses (P < .10) were considered as potential confounders in multivariable models but were removed in a stepwise manner if their inclusion changed the β coefficient of the primary predictor (controllers vs. control group) by <10%.
The analysis included 30 untreated HIV-infected controllers, 47 HIV-uninfected participants, 187 chronically HIV-infected participants maintaining treatment-mediated viral suppression, and 66 untreated HIV-infected participants with detectable viremia. Most of the HIV-infected participants were men between the ages of 40 and 50 years, but, compared with the other groups, controllers were more likely to be female and coinfected with HCV (table 1). Compared with treated participants maintaining viral suppression, controllers had both higher median current CD4+ T cell counts (727 vs. 442 cells/mm3; P < .001) and self-reported nadir CD4+ T cell counts (496 vs. 114 cells/mm3; P < .001). However, compared with HIV-uninfected participants, controllers tended to have lower CD4+ T cell counts (median, 727 vs. 943 cells/mm3; P = .059). Notably, 3 (10%) of 30 controllers had a CD4+ T cell count <350 cells/mm3, in a range where antiretroviral therapy should be considered according to current US Department of Health and Human Services guidelines . Furthermore, 2 (7%) of 30 controllers met the clinical definition of AIDS: 1 had a CD4+ T cell count <200 cells/mm3, and 1 had recently developed biopsy-confirmed cutaneous Kaposi sarcoma despite a CD4 count of 630 cells/mm3.
As expected, compared with HIV-infected participants with detectable viremia, controllers had lower median percentages of activated CD4+ T cells (3.8% vs. 7.7%; P < .001) and CD8+ T cells (15.5% vs. 30.8%; P < .001) (figure 1). However, compared with HIV-uninfected participants, controllers had higher median percentages of activated CD4+ T cells (3.8% vs. 2.2%; P < .001) and CD8+ T cells (15.5% vs. 5.1%; P < .001). In unadjusted analyses among all participants, higher CD4+ T cell activation levels were associated with HCV coinfection (P < .001), female sex (P = .023), and older age (P = .005), and higher CD8+ T cell activation levels were associated with HCV coinfection (P < .001) and female sex (P = .004). Controllers continued to have a mean 1.6-fold higher percentage of activated CD4+ T cells (P = .008) than HIV-uninfected participants after adjustment for age and continued to have higher CD4+ T cell activation levels than HIV-uninfected participants when the analysis was restricted to HCV-seronegative men (P = .036). Similarly, compared with HIV-uninfected participants, the controllers continued to have higher CD8+ T cell activation levels when the analysis was restricted to HCV-seronegative men (P = .015). Even though the lower observed CD4+ T cell counts in controllers may be a consequence of T cell activation, rather than a true confounder, controllers continued to have a mean 1.3-fold higher CD4+ T cell activation level (P = .064) than HIV-uninfected participants and a mean 2.3-fold higher CD8+ T cell activation level (P < .001) after adjustment for CD4+ T cell count.
Controllers also had a higher median percentage of activated CD8+ T cells than treated participants maintaining undetectable plasma HIV RNA levels during antiretroviral therapy (15.5% vs. 11%; P = .051). After adjustment for HCV serostatus and CD4+ T cell count, controllers continued to have a mean 1.3-fold greater percentage of activated CD8+ T cells than treated patients maintaining viral suppression (P = .055), and there was no evidence for confounding by age or sex. Conversely, there was no evidence for a difference in CD4+ T cell activation levels between controllers and treated patients maintaining viral suppression in either unadjusted analyses (P = .17) or analyses adjusted for CD4+ T cell count, HCV serostatus, sex, and age (P = .78).
Among controllers, lower CD4+ T cell counts were associated with higher levels of activated CD4+ T cells (Spearman's ρ = -0.47; P = .009) and CD8+ T cells (Spearman's ρ = -0.37; P = .047) (figure 2). These relationships were similar to those observed among the HIV-infected participants maintaining treatment-mediated viral suppression (for the comparison between CD4+ T cell count and activated CD4+ T cells, Spearman's ρ = -0.56 and P < .001; for the comparison between CD4+ T cell count and activated CD8+ T cells, Spearman's ρ = -0.36 and P < .001). Conversely, there was no evidence for a relationship between CD4+ T cell count and the percentage of either activated CD4+ T cells (Spearman's ρ = 0.11; P = .64) or CD8+ T cells (Spearman's ρ = 0.23; P = .30) among HIV-uninfected participants. Among controllers, there was no evidence for a relationship between CD4+ T cell count and either sex or HCV serostatus (P = .14 for both). Among controllers, each 2-fold increase in the percentage of activated CD4+ T cells was associated with a mean of 165 fewer CD4+ T cells/mm3 (P = .016). Similarly, each 10% increase in activated CD8+ T cells was associated with a mean of 101 fewer CD4+ T cells/mm3 (P = .010). Also of note, the controller who was recently given a diagnosis of cutaneous Kaposi sarcoma despite a CD4+ T cell count of 630 cells/mm3 had the second highest CD4+ T cell activation level (10.9%) and the third highest CD8+ T cell activation level (35.8%) in the cohort.
Because controllers had abnormally high T cell activation levels despite maintaining undetectable plasma HIV RNA levels, we considered other mechanisms by which HIV disease can cause T cell activation. HIV infection is associated with disruption of the gut mucosal barrier and persistent translocation of bacterial products, which in turn may cause systemic immune activation . To address this potential mechanism in our controllers, we compared plasma LPS levels between 14 controllers, 31 HIV-uninfected participants, and 33 untreated HIV-infected participants with detectable viremia. Controllers had significantly higher median plasma LPS levels than uninfected participants (61 vs. 28 pg/mL; P < .001) (figure 3A). However, there was no evidence of a difference in plasma LPS levels between controllers and untreated patients with detectable viremia (61 vs. 71 pg/mL; P = .80). There was also no evidence of a relationship between plasma HIV RNA levels and LPS levels among untreated participants with detectable viremia (Spearman's ρ = -0.21; P = .24). Among controllers, higher plasma LPS levels were associated with higher CD8+ T cell activation levels (Spearman's ρ = 0.56; P = .039) (figure 3B) but not CD4+ T cell activation levels (Spearman's ρ = 0.26; P = .37).
Although many studies have characterized the mechanisms by which elite controllers suppress HIV replication, to our knowledge no study has fully characterized the extent of generalized T cell activation in these individuals. In the present study, we demonstrated that controllers have significantly higher levels of CD4+ and CD8+ T cell activation than HIV-uninfected individuals and higher CD8+ T cell activation levels than HIV-infected patients maintaining antiretroviral treatment–mediated viral suppression. Furthermore, higher T cell activation levels among controllers was associated with lower CD4+ T cell counts, consistent with the hypothesis that T cell activation may promote HIV disease progression independent of HIV replication. Finally, controllers had abnormally high plasma LPS levels, which were associated with higher CD8+ T cell activation levels, suggesting mucosal translocation of bacterial products as either a cause or a consequence of T cell activation in this setting.
It is important to note that most of the controllers in our study exhibited no clinical signs of immunodeficiency and continued to maintain normal CD4+ T cell counts during observation. There is irrefutable evidence that among untreated individuals low plasma HIV RNA levels are associated with slower rates of CD4+ T cell decline and clinical progression . However, plasma HIV RNA levels do not fully explain the variability in the rate of CD4+ T cell decline [15, 16]. For example, although patients maintaining plasma HIV RNA levels <500 copies/mL in the absence of antiretroviral therapy maintained stable CD4+ T cell counts and experienced much slower progression to AIDS and death than those with higher plasma HIV RNA levels in an international study of recently HIV-infected patients (CASCADE), 10 (7%) of 145 virologic controllers eventually experienced an AIDS-defining illness, and 5 experienced clear evidence of continued CD4+ T cell decline while plasma HIV RNA levels remained <500 copies/mL. Although many of the virologic controllers in the CASCADE study may have had detectable plasma HIV RNA levels between 75 and 500 copies/mL, these results are consistent with our data reported here.
Our observation that T cell activation is associated with CD4+ T cell decline even in the absence of measurable virus replication suggests a mechanistic link between chronic inflammation and progressive immunodeficiency and is supported by a number of other observations. For example, transgenic mice with a constitutive generalized T cell activation phenotype experience dramatic CD4+ T cell decline, thymic depletion, and death from Pneumocystis pneumonia, all without viral infection . The observation that T cell activation distinguishes pathogenic from nonpathogenic SIV infection further suggests that T cell activation is a significant mediator of immunodeficiency in HIV infection . Clinical data from our group and other groups further demonstrate that higher CD8+ T cell activation levels are associated with more rapid CD4+ T cell decline and clinical progression among untreated patients [18–23] and fewer treatment-mediated CD4+ T cell gains at any given level of viral replication [24–27]. Finally, among individuals who are infected with the typically less pathogenic HIV-2, T cell activation is more closely associated with CD4+ T cell depletion than the extent of viral replication . Although these longitudinal studies suggest that CD8+ T cell activation markers are better predictors of subsequent clinical progression, we found a stronger correlation between CD4+ T cell activation levels and CD4+ T cell counts in our current cross-sectional analysis. This apparent discrepancy may be partly explained by CD4+ cell lymphopenia causing homeostatic CD4+ T cell activation and proliferation, an effect primarily observed in cross-sectional analyses.
We also found that controllers had higher CD8+ T cell activation levels than HIV-infected participants maintaining treatment-mediated viral suppression, a difference only partially explained by the higher prevalence of HCV coinfection among controllers. This is unlikely to be due to relatively higher plasma HIV replication levels <75 copies/mL in the controllers, because controllers appear to have a lower viral burden than patients maintaining antiretroviral therapy–mediated viral suppression . Alternatively, controllers may be enriched for polymorphisms that lead to stronger innate immune responses or lower regulatory T cell activity in response to HIV infection, both factors that might contribute to more-potent suppression of HIV replication at the expense of higher generalized T cell activation levels [35–42].
Our finding of an association between plasma LPS levels and CD8+ T cell activation among controllers suggests another indirect mechanism by which HIV may cause T cell activation. Most HIV-infected individuals experience a dramatic and early loss of CD4+ T cells from the lamina propria of the gastrointestinal tract [43–45], physical disruption of the mucosal barrier [46, 47], and mucosal repair gene expression defects . These perturbations allow for translocation of bacterial products, including LPS, which may contribute to generalized T cell activation by stimulating innate immune responses . Contrary to an earlier report of 3 controllers who maintained normal to high CD4+ T cell density in gut-associated lymphoid tissue and preserved expression of mucosal repair genes , our finding of abnormally high LPS levels among a much larger cohort suggests that mucosal barrier defects persist in the majority of controllers, potentially contributing to abnormal CD8+ T cell activation levels. It remains unclear whether mucosal translocation of bacterial products is primarily a cause of immune activation in HIV infection or simply a consequence of HIV-mediated disruption of the mucosal barrier and noncausally associated with HIV-induced T cell activation. However, the absence of a relationship between plasma LPS and HIV RNA levels among HIV-infected individuals supports a direct relationship between microbial translocation and T cell activation.
Our study has several limitations that deserve comment. First, our sample of controllers may have been subject to a referral bias, skewed toward the more uncommon individuals who experience clinical progression despite low plasma HIV RNA levels. However, the proportion of controllers in our study who progressed to AIDS is remarkably similar to a population-based estimate among untreated patients with comparably low plasma HIV RNA levels from the CASCADE study . Second, because we did not perform high-sensitivity measurements of plasma HIV RNA levels (2–75 copies/mL) and assess viral replication in other compartments in our participants, we cannot exclude the possibility that the association between T cell activation and CD4+ T cell count observed among controllers is simply mediated by differences in viral replication below the detection limits of clinically available assays. However, a recent study suggests that controllers typically have lower levels of residual viremia than patients maintaining antiretroviral treatment–mediated viral suppression . Finally, because our study was cross-sectional, we cannot definitively establish that T cell activation is a cause rather than a consequence of CD4+ T cell depletion in controllers. However, because T cell activation declines dramatically after initiation of antiretroviral therapy, long before significant restoration of CD4+ T cells in peripheral blood or tissues, T cell activation is unlikely to be solely a consequence of CD4+ T cell depletion in HIV infection [44, 45, 49].
In summary, we have demonstrated that controllers have abnormal levels of T cell activation despite maintaining clinically undetectable levels of viral replication and higher CD8+ T cell activation levels than HIV-infected patients with antiretroviral therapy–mediated viral suppression. Furthermore, higher T cell activation levels were associated with lower CD4+ T cell counts, consistent with the hypothesis that T cell activation mediates T cell count decline and immunodeficiency even in the absence of clinically detectable viremia. Because T cell activation might not decline significantly with antiretroviral therapy in patients experiencing clinical progression in the setting of clinically undetectable plasma HIV RNA levels  and because higher T cell activation levels during antiretroviral therapy are associated with fewer treatment-mediated CD4+ T cell count gains , it will be important to develop interventions that specifically target the indirect mechanisms by which HIV causes T cell activation.
We thank the Cleveland Immunopathogenesis Consortium for advice and discussion.
Financial support: Universitywide AIDS Research Program (grant CC99-SF-001); University of California, San Francisco–Gladstone Institute of Virology and Immunology Center for AIDS Research (grants P30 AI27763 and P30 MH59037); National Institutes of Health (NIH; grants R01 AI 52745, R37 AI40312, NS 37660, and K23 AI65244); General Clinical Research Center at San Francisco General Hospital (grant 5-MO1-RR00083–37); Center for AIDS Prevention Studies (grant P30 MH62246). J.M.M. is a recipient of the Burroughs Wellcome Fund Clinical Scientist Award in Translational Research and the NIH Director's Pioneer Award Program, part of the NIH Roadmap for Medical Research (grant DPI OD00329).
Potential conflicts of interest: none reported.
Presented in part: Keystone Symposium on Molecular and Cellular Determinants of HIV Pathogenesis, Whistler, British Columbia, Canada, 25–29 March 2007 (poster 265).