| |
The Impact of HIV Tropism on Decreases in CD4 Cell Count, Clinical Progression, and Subsequent Response to a First Antiretroviral Therapy Regimen.
|
| |
| |
[authors suggest: Tropism Testing May Identify Patienta At Risk For Accelerated Disease Progression and the Need For Earlier HAART.]
Clinical Infectious Diseases May 15, 2008;46:1617-1623
Laura Waters, Sundhiya Mandalia, Paul Randell, Adrian Wildfire, Brian Gazzard, and Graeme Moyle
St. Stephens AIDS Trust, Department of Genitourinary/HIV Medicine, Chelsea and Westminster Hospital, London, United Kingdom
ABSTRACT
Background. Human immunodeficiency virus (HIV) uses 2 distinct chemokine receptors, CCR5 (R5) or CXCR4 (X4), during entry. Viruses may be R5 tropic, X4 tropic, or dual/mixed (D/M) tropic. R5-tropic virus predominates at high CD4 cell counts, with the number of X4-tropic strains increasing as CD4 cell count decreases.
Methods. We investigated the relationship between tropism and decreases in CD4 cell count before antiretroviral therapy initiation, the frequency of clinical events, and responses to antiretroviral therapy in a cohort of treatment-naive patients.
Results.
Four hundred two treatment-naive patients underwent tropism determination; 326 harbored R5-tropic virus, and 76 harbored X4- or D/M-tropic virus.
After adjustment for baseline characteristics, the rate of decrease in CD4 cell count was significantly greater in patients infected with X4- or D/M-tropic virus at 12 months (p=.026). [Animal models have shown that reduced HIV-specific immunity may result in a tropism switch [14], supporting the hypothesis that the emergence of X4-tropic virus may be triggered by, rather than a cause of, immunological decline.]---- [Tropism testing could play a role in predicting the rate of disease progression in treatment-naive patients and could assist in deciding when to commence ART.]
Two hundred twenty-nine individuals infected with R5-tropic virus and 60 individuals infected with X4- or D/M-tropic virus commenced antiretroviral therapy between tropism testing and the time of data analysis.
Time to viral suppression and the proportion of patients achieving viral suppression were similar at 6, 12, and 24 months. CD4 cell count increases were similar.
Clinical events were significantly more common in the group infected with X4- or D/M-tropic virus.
Multivariate analysis demonstrated a relative risk of experiencing a clinical event of 2.56 (95% confidence interval, 1.37-4.76; p=.003) among patients infected with X4- or D/M-tropic virus.
Conclusions. The presence of D/M- or X4-tropic virus has a deleterious effect on CD4 cell count decrease and risk of clinical disease. Response to standard antiretroviral therapy is not affected by viral tropism.
BACKGROUND
HIV entry into target cells is a stepwise process that includes attachment to CD4 cell surface, interaction with cell surface coreceptors, and fusion of the viral envelope and host cell membranes. Viral gp120 binds to CD4 cells to expose a binding site allowing gp120 to interact with 1 of 2 cell-surface chemokine receptors, CCR5 (R5) or CXCR4 (X4). This coreceptor binding results in conformational changes in gp41, bringing the viral and host cell membranes together to facilitate fusion and the release of the viral core into the host cell cytoplasm. Virus can use R5 or X4 individually or can be dual tropic (virus that can bind to R5 and X4). Current rapid screening technology does not distinguish between dual-tropic virus and mixtures of R5- and X4-using virus; thus, this virus is referred to as dual/mixed (D/M) virus. Although a number of other chemokine receptors can be used by HIV in vitro, there is no evidence to date that these play a role in vivo. Genetic sequences within viral gp120, particularly the highly variable and flexible gV3 looph region, determine coreceptor use [1].
Sexually transmitted HIV is predominantly R5 tropic; transmission of X4-using strains appears to be constrained [2, 3]. The presence of X4-tropic virus is consistently associated with low CD4 cell counts in both treatment-naive and treatment-experienced individuals; however, R5-tropic virus and D/M-tropic populations are found in individuals at all CD4 cell count strata [4-9]. The emergence of X4-tropic virus is suggested to be associated with a more rapid decrease in CD4 cell count and accelerated disease progression [10-13], although it is unclear whether this is causative. Animal models have shown that reduced HIV-specific immunity may result in a tropism switch [14], supporting the hypothesis that the emergence of X4-tropic virus may be triggered by, rather than a cause of, immunological decline.
A number of rapid, reliable, high-throughput recombinant coreceptor phenotype assays have been developed. These assays use virus synthetically created with patient-derived HIV envelope sequences to infect cell lines engineered to express CCR5 or CXCR4, and infection of cell lines is detected using a reporter gene (such as luciferase in the Monogram Biosciences Trofile assay [15]). Results are reported as R5, X4, or D/M virus. With the Trofile assay, minority subpopulations with alternate tropisms are reliably detected when present at 5%-10%; at 10%, the sensitivity of the assay was 100%. The plasma viral load above which samples can be reliably amplified in the Trofile assay is 1000 copies/mL [15].
The development and application of these assays have been instrumental in advancing the evaluation and study of coreceptor antagonists in vitro and in clinical trials and have also been used to reevaluate the epidemiology of coreceptor use in large populations of antiretroviral therapy (ART)-naive and -experienced individuals [4-10]. Tropism testing could play a role in predicting the rate of disease progression in treatment-naive patients and could assist in deciding when to commence ART.
Discussion
In our cohort of treatment-naive patients, we observed a significantly greater decrease in CD4 cell count over 12 months and significantly more clinical events in those harboring X4- or D/M-tropic HIV, compared with those harboring R5-tropic virus. However, the immunological and virological responses to ART were similar over 2 years, regardless of baseline tropism, suggesting that the presence of X4-using virus, either alone or as D/M virus, does not impact subsequent response to therapy. This is important not only in relation to the evolution toward X4 tropism with advancing disease, but also because viremia during ART is associated with genetic evolution and tropism shifts [16]. Because of the ongoing debate with regard to when to start ART and the trend toward earlier treatment [17], tropism testing could play a role in contributing to decisions regarding the commencement of therapy. Earlier initiation of ART must be balanced against the costs and risks of longer-term drug exposure, and identifying patients at the greatest risk of disease progression could help target earlier therapy toward those most at risk of progression. Of note, differences in decrease in CD4 cell count emerged over time, indicating that studies evaluating the effects of tropism changes under the influence of coreceptor antagonist drugs require prolonged (>1 year) follow-up. vIt remains unclear whether the impact of tropism in treatment-naive individuals on decrease in CD4 cell count differs from the observation of D/M- or X4-tropic virus during coreceptor antagonist therapy.
More patients in the group with R5-tropic virus than in the group with X4- or D/M-tropic virus received NNRTI-based, first-line therapy (85.8% vs. 71.7%). This was likely to be secondary to the fact that individuals infected with X4- or D/M-tropic virus tended to start treatment sooner (as demonstrated by the number of patients receiving therapy at 12 months), and the increasing use of NNRTI-based HAART in the group with R5-tropic virus reflects changes in UK prescribing practices over time. In addition, the fact that individuals harboring X4-tropic virus had higher baseline viral loads and lower CD4 cell counts may have resulted in more PI use in this group because of the bias toward selecting PIs for patients with advanced disease. Because this was a nonrandomized study, differences in the prescribing patterns between the groups were to be expected, but the majority of individuals received an NNRTI, regardless of tropism, and large cohort analyses have demonstrated little difference between boosted PI- and NNRTI-based regimens.
We observed an increased rate of clinical events during the 12 months after tropism testing among individuals infected with X4- or D/M-tropic virus after adjustment for baseline CD4 cell count and viral load. Because the majority of patients infected with X4- or D/M-tropic virus commenced therapy within this 12-month period, they were not censored at the start of HAART. Other studies have yielded similar findings [10-13] when viral tropism was determined by specific assay [10, 11] or estimated by analyzing the ability of the virus to induce syncytium formation in culture [12, 13]. A longitudinal study involving 68 individuals from the Multicenter AIDS Cohort Study demonstrated that emergence of X4-tropic virus was more common in patients who experienced progression of HIV infection than in those who did not [18]. In addition, over one-half of the participants in this study who experienced a sudden decrease in CD4 cell count exhibited a switch to X4 tropism before the decrease occurred. Finally, a recently presented cohort analysis also found that X4 tropism was associated with accelerated disease progression when adjusted for baseline viral load and CD4 cell count, consistent with our results [19]. This may be secondary to the fact that advancing HIV infection is associated with immune dysfunction beyond CD4 cell depletion that could not be adjusted for in our analysis. X4-using viruses can infect a wider range of target cells, compared with HIV strains limited to CCR5-mediated entry, such as eosinophils [20] and CD8 cells [21]. Target cell susceptibility may account, in part, for different rates and types of opportunistic events.
In this study, we analyzed outcomes from a tropism evaluation performed at a single time; in the absence of longitudinal tropism testing, we cannot exclude the possibility that some individuals classified as being infected with R5-tropic or D/M-tropic virus for analysis purposes would have experienced a tropism shift over time. Shifts in tropism results between samples obtained as little as 4-6 weeks apart have been reported in 8% of individuals included in clinical trials with CCR5 antagonists [22, 23]. In a longitudinal study of HIV-infected patients who received stable ART and had detectable viral loads, shifts in tropism occurred in both directions [24]. In this study, 72 patients underwent a median of 3 tropism tests over 9 months; 12% of patients infected with R5-tropic virus at baseline shifted to D/M-tropic virus by 1 year, and 11% infected with D/M-tropic virus at baseline shifted to R5- or X4-tropic virus over the same period. These apparent tropism shifts may represent oscillations in the amount of X4-tropic virus around the limit of assay detection. The bidirectional nature of the tropism changes suggests that any changes in our population should have been balanced. In addition, because our patients were treatment naive, there may have been less evolutionary pressure towards tropism change.
In summary, we demonstrated that the presence of D/M-tropic HIV is associated with accelerated immunological and clinical progression. Individuals infected with X4-tropic virus only were analyzed in the group infected with D/M-tropic virus, but because the numbers were very small (3 patients infected with X4-tropic virus only in the whole cohort), these findings do not necessarily apply to pure X4-tropic virus. Viral tropism, however, does not impact the response of CD4 cell count to ART, time to viral suppression, or rates of viral suppression over a 2-year period.
Results
A total of 402 eligible treatment-naive patients underwent tropism testing; 326 of these patients harbored R5-tropic virus, 73 harbored D/M-tropic virus, and 3 harbored X4-tropic virus only. At the time of analysis (August 2006), 340 of 402 individuals had commenced ART; 51 of 340 patients were excluded (the duration of follow-up was insufficient for 28 patients, and 23 were prescribed an excluded regimen [see Methods]), leaving 229 patients infected with R5-tropic virus and 60 patients infected with X4- or D/M-tropic virus.
Baseline characteristics are described in table 1; only baseline CD4 cell count and viral load differed between the 2 groups. The baseline CD4 cell count was significantly lower and the viral load was significantly higher in the patients infected with X4-tropic or D/M-tropic virus, compared with the patients infected with R5-tropic virus. There was no difference between the 2 groups in terms of the number of baseline genotypic resistance mutations (for nucleoside reverse-transcriptase inhibitor mutations, p=.826; for NNRTI mutations, p=.893; for PI mutations, p=550) (data not shown).
Decreases in CD4 cell count before the commencement of therapy are shown in figure 1. At 12 months, the mean decrease in CD4 cell count in the group infected with X4- or D/M-tropic virus was statistically significantly greater than that in the group infected with R5-tropic virus (p=.026). Thereafter, a decrease in the number of patients in the group infected with X4- or D/M-tropic virus was seen; thus, statistical analyses were likely to yield uncertain results.
Figure 1. Decrease in CD4 cell count from baseline (time of tropism assay) in antiretroviral therapy (ART)-naive patients. All data are from before the initiation of ART and were censored at the initiation of ART. Error bars indicate 95% CIs. Time-weighted differences in mean values from baseline were estimated using a linear MIXED model. (note from Jules: although the authors say the decrease in the number of patients precludes analysis, the decline in CD4 count increases over time in the X4 or D/M-tropic virus group from about -180 at 12 months to -220 at 24 months, so further study would be interesting to see if these patients experiencing increasing CD4 declines over time because if that is so earlier HAART would be important).
The first prescribed ART regimen is shown in table 2. Overall, there was a statistically significant difference between the 2 groups in terms of regimen type (NNRTI-, unboosted PI-, or boosted PI-based regimen), with more patients in the group infected with R5-tropic virus commencing an NNRTI, and more patients in the group infected with X4- or D/M-tropic virus initiating a boosted or unboosted PI. Because of the differences in regimen type prescribed, this was adjusted for in the response to ART analysis.
The immunological response to ART is outlined in figure 2. Over 24 months, the increase in CD4 cell count in the 2 groups did not significantly differ. At 12 months, the increase in CD4 cell count was 185 cells/mm3 (95% CI, 166-204 cells/mm3) and 182 cells/mm3 (95% CI, 145-219 cells/mm3) in the groups infected with R5-tropic virus and X4- or D/M-tropic virus, respectively (p=.812). At 24 months, individuals infected with R5-tropic virus experienced an increase in CD4 cell count of 247 cells/mm3 (95% CI, 227-267 cells/mm3), and patients in the group infected with X4- or D/M-tropic virus had an increase in CD4 cell count of 292 cells/mm3 (95% CI, 254-330 cells/mm3; p=.482).
Rates of virological suppression in the 2 groups were comparable; 71.2%, 73.4%, and 72.5% of patients infected with R5-tropic virus had a viral load <50 copies/mL at 6, 12, and 24 months, respectively. Rates of virologocal suppression among those infected with X4- or D/M-tropic virus were 75.0%, 78.3%, and 68.3% at the same times, respectively (at 6 months, =.637; at 12 months, p=509; at 24 months, p=.670). Time to viral suppression was also similar in the 2 groups (data not shown).
In the group infected with R5-tropic virus, 27 (8.3%) of 326 patients started ART during the 12-month follow-up period, which was significantly fewer than the 25 (32.9%) of 76 patients infected with X4- or D/M-tropic virus who initiated therapy during the same period (p<.001). The frequency of clinical events was significantly greater in the group infected with X4- or D/M-tropic virus after adjustment for CD4 cell count, viral load, and ART regimen. A total of 23 (7.1%) of 326 patients infected with R5-tropic virus experienced a clinical event, compared with 17 (22.4%) of 76 patients infected with X4- or D/M-tropic strains (p<.001), after adjustment for baseline CD4 cell count and viral load. Using logistic regression analysis adjusted for baseline immunological and virological parameters, the relative risk of a clinical event occurring in patients harboring R5-tropic virus, compared with those harboring X4- or D/M-tropic virus, was 0.46 (95% CI, 0.22-0.98; p=.044). The clinical events diagnosed are outlined in table 3.
Methods
The Chelsea and Westminster Hospital (London, United Kingdom) has an active cohort of 6000 HIV-infected individuals. Routine investigations (including hematological, biochemical, immunological, and virological investigations) are performed every 3-4 months, and the results of these investigations, in addition to treatment history and clinical events, are prospectively entered onto a large clinic database. Plasma samples have been collected and stored routinely since 1996; the Riverside Research Ethics Committee supervises their use.
Among a cohort of HIV-infected patients who underwent tropism testing once as part of a cross-sectional epidemiology study [4], we identified all patients who were treatment naive at the time of tropism determination. For the epidemiological study, samples were selected on the basis of availability of a date-matched viral genotype (Virco) to provide information on any possible association with pol gene mutations and viral clade. Samples were also required to have a viral load >1000 copies/mL to ensure PCR amplification. Data on patients who met the above criteria and had successful determination of coreceptor phenotype by the Phenosense assay were used to construct the database. Characteristics of samples that failed coreceptor phenotype determination were also assessed to confirm that the typed sample set was representative of the sampled population.
Coreceptor tropism was determined using the Monogram Biosciences Trofile assay according to the methodology described elsewhere [4]. Patients were divided into 2 groups: patients infected with R5-tropic virus and patients infected with D/M- or X4-tropic virus.
We used the database to record data regarding CD4 cell count and viral load before and after commencing first-line ART. For the purposes of this study, ART was defined as at least 2 nucleoside analogues in combination with a nonnucleoside reverse-transcriptase inhibitor (NNRTI) or a protease inhibitor (PI), regardless of ritonavir boosting. In addition, clinical events (AIDS-defining illness or death) occurring after the initiation of ART were recorded. Patients prescribed dual- or monotherapy regimens or who had <6 months of follow-up after tropism testing were excluded from analysis; data were censored at 96 weeks, the end of follow-up, or at the time of switching therapy because of virological failure.
Because longitudinal data were available for CD4 cell count, a linear MIXED model was used to derive time-adjusted decrease in CD4 cell count from baseline (defined as before initiation of therapy) and was adjusted for viral load at baseline. The MIXED procedure in SAS (SAS Institute) was used by fitting CD4 cell count as a dependent variable. Independent variables included the fixed effects of tropism and the duration of tropism (grouped into 3-month intervals) and tropism interaction by time. An unstructured covariance matrix was used to model the within-patient errors.
Estimates of change in CD4 cell count from baseline were obtained from intervention by time interaction. Estimates of trends over time are presented as point estimates with 95% CIs. Further multivariable analyses were used to adjust for other covariables assumed to have potential confounding or residual effect on trend in CD4 cell count over time; mainly, immunological response to ART was adjusted for baseline viral load and ART regimen.
The χ2 analysis was used to test the difference between the proportions of patients in each group who achieved viral suppression (viral load, <50 copies/mL) after commencing therapy and time to reach <50 copies/mL. Survival analyses were adjusted for baseline viral load and use of ART. Univariate and multivariable logistic regression analysis was performed using the GENMOD procedure in SAS (SAS Institute), with logit link and binomial error distribution to identify factors associated with the likelihood of R5 tropism. All statistical analyses were performed using SAS, version 9.1 (SAS Institute), and all presented P values are 2-tailed.
|
|
| |
| |
|
|
|
