Skip Navigation

Effects of Human Leukocyte Antigen Class I Genetic Parameters on Clinical Outcomes and Survival after Initiation of Highly Active Antiretroviral Therapy

  1. Zabrina L. Brumme1,2,a,b,
  2. Chanson J. Brumme1,a,b,
  3. Celia Chui1,
  4. Theresa Mo1,
  5. Brian Wynhoven1,
  6. Conan K. Woods1,
  7. Bethany M. Henrick1,
  8. Robert S. Hogg1,2,
  9. Julio S. G. Montaner1,2 and
  10. P. Richard Harrigan1,2
  1. 1 British Columbia Centre for Excellence in HIV/AIDS, Canada
  2. 2 Department of Medicine, University of British Columbia, Vancouver, British Columbia, Canada
  1. Reprints or correspondence: Dr. P. Richard Harrigan, BC Centre for Excellence in HIV/AIDS, 603–1081 Burrard St., Vancouver, BC, Canada V6Z 1Y6 (prharrigan{at}cfenet.ubc.ca).

  1. Presented in part: XVIth International AIDS Conference, 13–18 August 2006, Toronto, Canada (abstract MOPE0027)

  • ↵a Z.L.B. and C.J.B. contributed equally to this work.

  • ↵a Present affiliation: Partners AIDS Research Center, Massachusetts General Hospital, Boston.

 
Next Section

Abstract

Background. Human leukocyte antigen (HLA) class I variation influences the progression of untreated human immunodeficiency virus (HIV) disease; however, it is not known whether HLA class I variation may influence clinical outcomes after initiation of highly active antiretroviral therapy (HAART).

Methods. Associations between HLA class I genotypes and pretherapy clinical parameters were investigated in a cohort of 765 antiretroviral$#x2013;naive adults initiating HAART. Cox proportional hazards regression was used to investigate the effects of HLA class I genotypes on time to suppression of the viral load to <500 HIV RNA copies/ mL, time to an increase in the CD4 cell count to >100 cells/mm3 above the baseline count, and time to nonaccidental death over a >5-year period after initiation of HAART.

Results. Homozygosity at any HLA class I locus and possession of common HLA alleles were associated with a higher pretherapy viral load (P < .05). In multivariate analyses controlling for sociodemographic and clinical parameters at baseline, HLA class I homozygosity was significantly associated with a poorer CD4 cell response (P = .008), whereas possession of uncommon HLA alleles was associated with slower virologic suppression after initiation of HAART (P > .02). We observed no significant association between HLA parameters and time to nonaccidental death after initiation of HAART (P > .05, univariate analysis).

Conclusion. HLA class I zygosity-dependent and frequency-dependent effects may influence short-term HAART outcomes, and, thus, they deserve further investigation. No effects of these HLA parameters on survival after initiation of HAART were observed.

The genes of the highly polymorphic HLA class I system help to define the strength and range of the cytotoxic (CD8+) T cell response [1, 2]. Longitudinal studies of the natural history of HIV infection have indicated that HLA class I genes influence the rate of disease progression [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]. Homozygosity at any class I locus is associated with more rapid disease progression during untreated infection [10, 11], likely because homozygosity may limit the ability of the immune system to respond to a diverse range of peptide antigens [14], allowing HIV to more easily escape the immune response [11]. Similarly, a hypothesis of “frequency-dependent selection” postulates that, on a population basis, the virus adapts to the most commonly encountered HLA-restricted immune responses, conferring a selective advantage on individuals expressing rare alleles [12, 13]. In addition, several specific alleles—most notably, those at the HLA-B locus [15]—are associated with the risk of progression of HIV disease [3, 4, 5, 6, 7, 8, 9]: HLA-B*5701 and B*27 are associated with protective effects [4] at different stages in the natural history of the disease [16], whereas HLA-B*35P(x) (comprising B*3502, B*3503, B*3504, and B*5301) [17] and the B22 serogroup (comprising B*54, B*55, and B*56) [5] are associated with accelerated disease progression.

Although a complete understanding of the contributions of HLA diversity to HIV disease has by no means been achieved, an area of equal interest has now emerged: the relevance of HLA genetic diversity in the era of highly active antiretroviral therapy (HAART). Although the introduction of HAART in the mid-1990s completely changed the clinical course of HIV infection among treated individuals, a wide range of interindividual differences in treatment responses continues to be observed. Because consistent evidence suggests that host genetic factors significantly influence the natural history of HIV/AIDS [18], it is reasonable to hypothesize that human genetic factors may continue to exert measurable effects after therapy is initiated. The major objective of the present study, therefore, was to investigate the contribution of HLA class I variation to virologic, immunologic, and survival outcomes in a large cohort of HIV-infected, antiretroviral-naive individuals initiating HAART. A secondary objective was to confirm known associations between HLA class I parameters [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13] and markers of progression of untreated disease in a pretherapy cross-sectional analysis of the present cohort.

Previous SectionNext Section

Materials and Methods

Study Subjects: The HAART Observational Medical Evaluation and Research (HOMER) Cohort

In British Columbia, Canada, antiretroviral agents are distributed free of charge to HIV-infected individuals, through a centralized HIV/AIDS drug treatment program [19]. Routine clinical monitoring of patients takes place at ∼3-month intervals, at which time plasma viral loads are determined (by use of the Roche Amplicor Monitor Assay) and CD4 cell counts are performed. Note that the ultrasensitive adaptation to the Roche viral load assay was not routinely available over the course of the study; for this reason, a constant lower assay limit of <500 HIV RNA copies/mL was used. The HOMER cohort is an open, treatment-based cohort that includes all HIV-infected, antiretroviral-naive adults ⩾18 years of age who have initiated HAART through the treatment program since August 1996 [19]. A subset of the HOMER cohort, comprising all HIV-infected individuals who initiated triple therapy (consisting of 2 nucleoside reverse-transcriptase inhibitors and either a protease inhibitor or a nonnucleoside reverse-transcriptase inhibitor) in British Columbia between August 1996 and September 1999 (n = 1191 ) has been described elsewhere [20, 21, 22]. The present study population represents a nonrandom subset of 765 of these 1191 subjects who were included in the current study on the basis of the availability of a blood sample for HLA typing. In the HOMER cohort, ∼97% of subjects have HIV-1 subtype B infections [23]. Ethical approval was obtained through the institutional ethics review board.

HLA Typing

HLA class I typing was performed on DNA extracted from blood by use of a validated “in-house” sequence-based typing procedure based on International Histocompatibility Working Group protocols. This procedure involves independent, locus-specific, nested polymerase chain reaction amplification of exons 2 and 3 of HLA-A, -B, and -C, followed by automated, bidirectional DNA sequencing. Allele interpretation was performed by comparing sequence data against all alleles in the ImMunoGeneTics (IMGT) HLA database [24] as of August 2005. Four-digit resolution was achieved for all specific alleles investigated in the present study (e.g., B*5701 and B*35P[x]). Where possible, ambiguous allele combinations were resolved through incorporation of known allele frequencies and/or haplotypes [25]. Homozygous combinations were defined as combinations exhibiting no nucleotide mixtures over exons 2 and 3 for each individual class I locus.

Statistical Analyses

Associations between HLA types and pretherapy HIV clinical parameters. For each subject, a single baseline viral load and CD4 cell count (representing the most recently collected measurement within the 180 days before initiation of HAART) was used as a clinical marker of HIV clinical status before treatment; 90% of viral load and CD4 cell measurements at baseline were collected in the 90 days preceding initiation of therapy. Associations between HLA parameters (including allele frequencies, zygosity, and possession of specific alleles) and pretherapy viral load and CD4 cell count were investigated using 1-way analysis of variance (ANOVA), as well as the Wilcoxon rank sum and Kruskal-Wallis tests.

Longitudinal analysis of the influence of HLA parameters on post-HAART outcomes. Associations between HLA types and post-HAART outcomes were investigated using Kaplan-Meier survival analysis and Cox proportional hazards regression. Virologic and immunologic end points were defined as follows: time to viral suppression was defined as the time elapsed between initiation of HAART and the time when the first of at least 2 consecutive viral loads of <500 HIV RNA copies/mL was achieved. Time to a CD4 cell count increase of >100 cells/mm3 was defined as the time elapsed between initiation of HAART and the time when the first of at least 2 consecutive CD4 cell counts >100 cells/mm3 above the baseline count was achieved. Event-free subjects were censored on the date when the last viral load or CD4 cell count was determined up to the study's end date of 30 September 2004. Deaths occurring to 30 September 2004 were identified through linkage with the British Columbia Department of Vital Statistics and were classified according to the International Classification of Diseases, 10th Revision. All nonaccidental deaths were considered to be potentially HIV related and were considered to be “events” in the survival analysis. Accidental deaths were considered to be nonevents, and subjects were censored at the date of death. The median length of follow-up was 5.8 years.

Cox proportional hazards regression was used to calculate univariate and multivariate hazard ratios (HRs) and 95% confidence intervals (CIs) for each end point. Baseline variables that were modeled were sex (male vs. female [reference group]), age (per 10-year increment), AIDS diagnosis (yes vs. no), plasma HIV RNA load (per log10 increment), CD4 cell count (per decrement of 100 cells/mm3), adherence to therapy during the first year after initiation (⩾95% adherent vs. <95% adherent) [24], type of therapy received at treatment initiation (therapy containing a protease inhibitor vs. therapy containing a nonnucleoside reverse-transcriptase inhibitor), history of injection drug use (yes vs. no), CCR5Δ32 genotype (Δ32/wt vs. wt/wt) [21], and HIV coreceptor phenotype (R5/X4 vs. R5) [20]. Variables for whichP < .1 in univariate analyses were included in multivariate analyses. All tests of significance were 2sided, withP < .05 denoting statistical significance. Statistical analyses were performed using SAS software (version 8.0; SAS Institute).

Previous SectionNext Section

Results

Study Population

Subjects represent a subset (64%) of the total population of all (n = 1191 ) treatment-naive adults who initiated triple-combination therapy in the province of British Columbia, Canada, between August 1996 and September 1999 and who were selected on the basis of the availability of a blood sample for HLA typing. Comparison of the pretherapy characteristics of subjects included in (n = 765 ) and excluded from (n = 426) the study reveals no significant differences in the pretherapy CD4 cell count (median, 280 cells/mm3). However, subjects who were included in the study had a slightly lower pretherapy viral load (median HIV RNA load, 5.07 vs. 5.15 log10 copies/ mL; P = .03); were, on average, older (median age, 37.2 vs. 36.5 years; P = .02); and were more likely to be male (percentage of subjects who were male, 88% vs. 77%; P < .0001 ) than were subjects who were excluded from the study. HLAA and HLA-B typing was completed for all 765 subjects, whereas HLA-C types were determined for 706 subjects. Although complete data on ethnicity are unavailable, the HLA allele frequencies noted were consistent with those expected in a predominantly North American white population.

Associations between HLA Types and Pretherapy Clinical Parameters

HLA zygosity effects. Homozygosity at any class I locus was associated with a significantly higher pretherapy viral load (median, 5.11 vs. 5.04 log10 HIV RNA copies/mL in subjects homozygous at any locus vs. subjects heterozygous at all loci; P = .03) (figure 1A). No additional effect was observed when homozygosity at ⩾2 loci was investigated (not shown). When class I loci were considered individually, HLA-C homozygosity was significantly associated with a higher pretherapy viral load (P = .03), and a trend was observed between HLA-B homozygosity and a higher viral load (P = .08) (figure 1A). Note that the results of analysis of the independent loci remained consistent in a secondary analysis that excluded individuals who were homozygous at multiple loci (data not shown). No association between HLA zygosity and pretherapy CD4 cell counts was observed (median CD4 cell count, 290 vs. 280 cells/mm3 in homozygous vs. heterozygous subjects, respectively; P = .9).

Table 1.
View larger version:
Table 1.

Associations between HLA class I genetic parameters (and other baseline variables) and time to nonaccidental death after initiation of highly active antiretroviral therapy (HAART).

Table 2.
View larger version:
Table 2.

Associations between HLA class I genetic parameters (and other baseline variables) and time to suppression of the viral load to <500 HIV RNA copies/mL after initiation of highly active antiretroviral therapy (HAART).

Table 3.
View larger version:
Table 3.

Associations between HLA class I genetic parameters (and other baseline variables) and time to an increase in the CD4 cell count to >100 cells/mm3 above the baseline count after initiation of highly active antiretroviral therapy (HAART).

Figure 1.
View larger version:
Figure 1.

A, Associations between HLA zygosity and the pretherapy plasma HIV RNA load. Diamonds and lines denote medians and interquartile ranges, respectively. The no. of subjects in each group (n) is shown below the graphs. Note that the different total n values for the “HLA-C” and “Any” categories reflect the fact that complete HLA-C types were available for only 706 of 765 subjects. B, Associations between HLA class I allele frequencies and the pretherapy HIV RNA load. Each subject was assigned an “HLA frequency score” reflecting the sum of their HLA allele frequencies at each locus. HLA frequency scores were stratified into quartiles, where a “rare” score denoted the first (i.e., lowest) quartile, a score “medium” denoted the second and third quartiles, and a “common” score denoted the fourth (i.e., uppermost) quartile. Diamonds and thin lines denote medians and interquartile ranges, respectively. The n values for each group are shown below the graphs. Note that different total n values for the “HLA-C” and “Any” categories reflect the fact that complete HLA-C types were available for only 706 of 765 subjects.

HLA frequency-dependent effects. Each subject was assigned a total of 4 “HLA frequency scores” reflecting the sum of cohort allele frequencies at each individual class I locus, as well the sum of allele frequencies across combined loci. Scores were stratified into the following categories: “rare” (first quartile), “medium” (second and third quartiles), and “common” (fourth quartile). Having rare HLA-B (but not HLA-A or HLAC) alleles was significantly associated with a lower pretherapy viral load (5.04 vs. 5.11 log10 HIV RNA copies/mL among subjects with rare versus common alleles, respectively; P = .01 ) (figure 1B). Similarly, a combined rare score was also associated with a lower pretherapy viral load (P = .01 ) (figure 1B).

HLA class I locus-specific and allele-specific effects. The pretherapy viral load varied significantly with specific HLA-B alleles (P = .02, 1-way ANOVA) (figure 2) but not with HLAA or HLA-C alleles (P > .1 [not shown]). HLA-B*5701 was associated with a lower pretherapy viral load (median, 4.6 vs. 5.1 log10 HIV RNA copies/mL for B*5701-positive vs. B*5701negative subjects, respectively; P < .001 ), whereas HLA-B*55 was associated with a higher pretherapy viral load (median, 5.4 vs. 5.0 log10 HIV RNA copies/mL for B*55-positive vs. B*55negative subjects, respectively; P = .017 ). Possession of HLAB*13 and B*42 alleles was also associated with a lower pretherapy viral load (P < .05). Of note, when the HLA-B*35– positive group was subdivided into HLA-B*35P(x)–positive and HLA-B*35P(y)–positive groups [17], no significant differences in the viral load at baseline were observed (P = .4). Analysis of HLA frequency-dependent and allele-specific associations with pretherapy CD4 cell counts revealed similar trends, although associations generally did not achieve statistical significance (not shown).

Figure 2.
View larger version:
Figure 2.

Contribution of individual HLA-B alleles to pretherapy HIV RNA load variation. Rectangular boxes display the median (vertical black line) and interquartile range (edges of box) of the log10 HIV RNA load in subjects possessing each HLA-B allele. Data are sorted by median viral load, from the lowest to the highest load. Whiskers denote the most extreme values observed within 2 SDs of the mean log10 HIV RNA load for each allele. Individual alleles significantly associated with higher or lower pretherapy plasma viral loads are denoted by an asterisk (P < .05, Wilcoxon rank sum test). The dotted line denotes the median log10 HIV RNA load in the population.

Effects of HLA Class I Genetic Parameters on Clinical Outcomes and Survival after Initiation of HAART

Effects of HLA class I parameters on survival after initiation of HAART. Over a median follow-up of 5.8 years, 119 deaths were recorded among 765 subjects, for a crude mortality rate of 15.6%. All nonaccidental deaths (n = 96 [80.7%]) were considered to be potentially HIV related and were classified as events in the survival analysis (note that, for 69 [71.9%] of these 96 nonaccidental deaths, HIV and/or AIDS was specified as the underlying cause; the next most common causes of death were cardiovascular disease [n = 11 {11.4%}] and cancers [n = 5 {5.2%}]). Accidental deaths (n = 23 [19.3%], the majority of which were due to overdoses) were censored on the date of death.

In univariate analyses, we observed no statistically significant association between homozygosity at any class I locus and time to nonaccidental death after initiation of HAART (HR, 0.83 [95% CI, 0.54–1.28], for homozygous vs. heterozygous subjects; P = .4) (figure 3 and table 1). Similarly, post-HAART survival did not differ significantly according to HLA frequency score (HR, 1.03 [95% CI, 0.59–1.80], for subjects with rare scores vs. the reference group with common scores; P = .9) (figure 3 and table 1).

Figure 3.
View larger version:
Figure 3.

A–F, Impact of HLA zygosity-dependent, allele frequency–dependent, and specific allele effects on the time to nonaccidental death after initiation of highly active antiretroviral therapy. +, positive; -, negative.

We investigated whether specific HLA alleles that were previously associated with differential rates of progression of untreated disease may also be associated with survival times after initiation of HAART. Of interest, possession of HLA-B*5701 was associated with a trend toward an increased risk of non-accidental death after HAART initiation; however, this trend did not reach statistical significance (HR, 1.61 [95% CI, 0.81– 3.19]; P = .2). Possession of B*27, B*35P(x), or B22 serogroup alleles was not significantly associated with post-HAART survival times, although trends were observed toward improved survival among B*27-positive subjects and poorer survival among B22 serogroup–positive subjects (P > .1, for all) (figure 3 and table 1).

Effects of HLA class I parameters on virologic and immunologic responses to HAART. We investigated whether HLA class I parameters were associated with initial virologic and immunologic responses to HAART, as defined by “time to achieve suppression of the viral load to <500 HIV RNA copies/ mL” and “time to achieve an increase in the CD4 cell count to >100 cells/mm3 above baseline,” respectively. Most (657 [85.9%] of 765) subjects achieved suppression of the viral load to <500 HIV RNA copies/mL; of these subjects, 522 (79.5%) achieved suppression within 1 year after HAART initiation. A total of 576 (75.3%) of 765 subjects achieved an increase in the CD4 cell count to >100 cells/mm3 above the baseline count; of these subjects, 336 (58.3%) achieved this increase within 1 year of HAART initiation.

In univariate analyses, we observed no significant association between HLA zygosity and time to virologic suppression (HR, 0.98 [95% CI, 0.83–1.15] for homozygous vs. heterozygous subjects; P = .8). In univariate analyses, a trend toward a longer time to achieve an increase of >100 cells/mm3 in the CD4 cell count was observed for homozygous subjects (HR, 0.84 [95% CI, 0.71–1.00]; P = .053). In multivariate analyses adjusting for demographic variables, clinical variables, adherence [26], and other host [21] and viral [20] genetic markers of HIV disease progression, this trend became statistically significant (HR, 0.79 [95% CI, 0.64–0.94]; P = .008) (tables 2 and 3).

In univariate analyses, we observed a significantly increased time to suppression of the viral load to <500 HIV RNA copies/ mL in subjects with rare allele scores (HR, 0.78 [95% CI, 0.62– 0.97]; P = .03), compared with that noted for subjects with common allele scores. This finding remained significant in multivariate analyses adjusting for demographic, clinical, and genetic variables at baseline (table 2). In univariate analyses, no significant association between HLA allele frequency scores and the CD4 cell response after HAART initiation was observed, although a trend toward a poorer immunologic response among subjects with rare allele scores was observed (table 3).

The HLA-B*5701, HLA-B*27, and B22 serogroup alleles were not significantly associated with differential short-term virologic or immunologic responses after HAART initiation (P > .1, univariate analysis) (tables 2 and 3). In univariate analysis, possession of B*35P(x) was associated with a more rapid time to virologic suppression to <500 HIV RNA copies/mL (table 2); however, this association did not remain significant in multivariate analyses adjusting for parameters at baseline (HR, 1.23 [95% CI, 0.87–1.75]; P = .2). No significant association between B*35P(x) and initial CD4 cell response was observed (table 3).

Secondary exploratory analyses. Results of exploratory analyses restricted to individuals infected with HIV-1 subtype B(∼97% of subjects) and individuals whose baseline CD4 cell count and viral load were measured in the 90 days preceding HAART initiation (∼90% of subjects) were consistent with the results of the original analysis.

Previous SectionNext Section

DISCUSSION

Associations between HLA genetic parameters and the status of untreated HIV disease were investigated using the pretherapy cross-section of the cohort in the present study. The observation that homozygosity at any class I locus was associated with a higher pretherapy viral load in our cohort is consistent with the reported “homozygous disadvantage” effect on rates of HIV disease progression observed in historic seroconverter cohort studies [1010, 11], and it suggests that this effect is strong enough to be observed even in cross-sectional analyses of chronically infected individuals with unknown lengths of infection. Similarly, the observation that possession of rare alleles at the HLAB (but not HLA-A or HLA-C) locus is associated with a lower pretherapy viral load supports frequency-dependent effects of class I alleles on untreated HIV disease [12, 13], and it suggests that the dominant locus driving this phenomenon is HLA-B [15]. Furthermore, the observation that the pretherapy viral load varied significantly according to possession of specific HLA-B (but not HLA-A or HLA-C) alleles is consistent with results from an independent analysis of chronically infected individuals that reported a profound influence of HLA-B on immune containment of HIV infection as estimated by viral load [15]. Finally, results suggest that the protective effects of HLA-B*5701 [4, 16] and the detrimental effects associated with B22 serogroup alleles [5] also persist into chronic infection (as evidenced by differences in pretherapy viral loads in these subjects). Of interest, no significant association was observed between HLA-B*27 and HLA-B*35P(x) alleles on pretherapy viral loads in the present study, in contrast to previous studies reporting protective [4] and detrimental [17] effects, respectively, of these alleles on rates of HIV disease progression.

Despite numerous statistically significant associations observed between HLA and pretherapy viral loads, it is important to note the limitations of this analysis. First, the cross-section of our cohort at baseline represents individuals in varying stages of chronic infection. HLA genotypes previously associated with rates of HIV disease progression in studies of seroconverter cohorts [4, 5, 10, 11, 12] or acute infection [3] may not necessarily be observed in the present cross-sectional analysis of pretherapy clinical parameters. In addition, no consistent significant associations were observed between HLA and pretherapy CD4 cell counts. This observation may partially be explained by the fact that this is an HIV treatment cohort, with individuals enrolled in the cohort at therapy initiation. Because the guidelines for HAART initiation are largely CD4 cell driven, a pretherapy cross-sectional analysis may underestimate associations between HLA and the CD4 cell count (and, to a lesser extent, the viral load). Despite these limitations, we were able to detect HLA frequency-dependent and allele-specific effects on HIV clinical parameters consistent with those noted in previous reports [3, 4, 5, 6, 7, 8, 9], suggesting that the effects of HLA class I on untreated HIV disease are substantial and persist into chronic infection.

The analysis of an HIV treatment cohort, however, is ideal for the investigation of the influence of HLA genetics on HAART outcomes. Although there is much evidence supporting the effects of HLA class I on rates of progression of untreated HIV disease [4, 5, 10, 11, 12], it is not known whether class I variation may still exert an effect on viral load responses, immunologic reconstitution, and/or survival once therapy is initiated. Overall, we observed no significant effects of HLA zygosity-dependent or frequency-dependent effects on survival in the >5-year period after initiation of HAART, suggesting that these parameters do not influence survival after therapy initiation.

In an analysis of specific alleles, we observed a very weak trend toward improved survival among B*27-positive subjects and poorer survival among B22 serogroup–positive subjects; however, these results did not achieve statistical significance (P > .1). Despite strong evidence identifying HLA-B35P(x) as a risk factor for disease progression in persons with untreated HIV infection [17], no association between B35P(x) and sur vival after HAART was observed in the present study. Given the identification of HLA-B*5701 as a risk factor for the lifethreatening abacavir hypersensitivity reaction [27, 28], it was interesting to note that HLA-B*5701 was associated with a weak trend toward poorer survival after HAART. Abacavir hypersensitivity was unlikely to explain this result, however. Of the 9 nonaccidental deaths of B*5701-positive subjects, only 2 involved subjects who had been prescribed abacavir (and abacavir prescription was not temporally correlated with death in either case). Alternatively, results may be attributable to a survivor bias where B*5701-positive subjects initiating therapy represent individuals who have been infected with HIV for a longer period (a hypothesis that we are unable to investigate, because dates of seroconversion are unknown). However, it should be noted that B*5701-positive individuals were not significantly older than B*5701-negative individuals in this cohort (data not shown).

Overall, results suggest that these specific HLA-B alleles do not significantly influence 5-year survival rates after HAART initiation. However, it is important to note that the numbers of observed events were small and, thus, larger cohorts and/or longer follow-up times may be necessary to detect effects, if present. As expected, the factors most strongly associated with survival after HAART were younger age, a higher CD4 cell count at baseline, and treatment adherence.

In addition, we wished to investigate whether HLA genetic parameters may influence shorter-term HAART outcomes— namely, the ability to achieve virologic suppression and/or immune reconstitution (as measured by an increase in the CD4 cell count) after initiation of therapy. In multivariate analyses, we observed a significant association between homozygosity at any class I locus and a poorer CD4 cell response after initiation of HAART, suggesting that the HLA “homozygous disadvantage” [10, 11] may persist even after initiation of therapy. Unfortunately, we were not able to address a recently reported association between the HLA*A1-B8-DR3-DQ2 haplotype and improved immunologic reconstitution after HAART [29], because neither haplotype analysis nor class II typing was performed in the current study.

Of interest, we observed a significantly increased time to virologic suppression in individuals with rare HLA alleles, compared with individuals with common HLA alleles, despite the fact that rare allele scores were associated with a lower pre-therapy viral load. A weak trend toward a poorer immunologic response to HAART was also observed among individuals in this group. These associations remained significant in an analysis restricted to subjects with HIV-1 subtype B infections, suggesting that this result cannot be explained by individuals who possess HLA alleles that are uncommon in North America and who harbor non–subtype B infections that may not respond as readily to HAART. Finally, to date, the strongest evidence supporting a role for HLA in the context of HAART outcomes relates to the association between HLA-B*5701 and class II allele DRB1*0101 and immune-mediated hypersensitivity reactions to abacavir [27, 28] and nevirapine [30], respectively. Because data on adverse events are incompletely available (and because class II typing was not performed), we were unable to investigate comprehensively associations between HLA and antiretroviral-associated hypersensitivity reactions in this cohort.

In conclusion, results from our cross-sectional pretherapy analysis suggest that previously characterized effects of HLA class I (including zygosity-dependent, frequency-dependent, and specific effects of HLA-B*5701) on untreated HIV disease may persist well into chronic untreated infection. HLA class I zygosity-and frequency-dependent effects may influence short-term immunologic and virologic HAART outcomes, respectively, and thus deserve further investigation in independent cohorts. Results do not support a significant continued influence of HLA class I zygosity-or frequency-dependent effects on survival in the 5-year period after initiation of HAART; however, results do not exclude a potential effect of HLA on survival in the longer term.

Previous SectionNext Section

Footnotes

  • Potential conflicts of interest: none reported.

  • Financial support: Michael Smith Foundation for Health Research (doctoral research awards to Z.L.B. and a Senior Scholar Award to R.S.H.); Canadian Institutes for Health Research (postdoctoral and doctoral fellowship award to Z.L.B.).

  • Received October 3, 2006.
  • Accepted December 18, 2006.
Previous Section
 

References

  1. 1.↵
    1. Klein J,
    2. Sato A
    . The HLA system: first of two parts. N Engl J Med 2000;343:702-9.
    CrossRefMedlineWeb of Science
  2. 2.↵
    1. Klein J,
    2. Sato A
    . The HLA system: second of two parts. N Engl J Med 2000;343:782-6.
    CrossRefMedlineWeb of Science
  3. 3.↵
    1. Altfeld M,
    2. Addo MM,
    3. Rosenberg ES,
    4. et al
    . Influence of HLA-B57 on clinical presentation and viral control during acute HIV-1 infection. AIDS 2003;17:2581-91.
    CrossRefMedlineWeb of Science
  4. 4.↵
    1. Kaslow RA,
    2. Carrington M,
    3. Apple R,
    4. et al
    . Influence of combinations of human major histocompatibility complex genes on the course of HIV-1 infection. Nat Med 1996;2:405-11.
    CrossRefMedlineWeb of Science
  5. 5.↵
    1. Dorak MT,
    2. Tang J,
    3. Tang S,
    4. et al
    . Influence of human leukocyte antigen- B22 alleles on the course of human immunodeficiency virus type 1 infection in 3 cohorts of white men. J Infect Dis 2003;188:856-63.
    Abstract/FREE Full Text
  6. 6.↵
    1. Hendel H,
    2. Caillat-Zucman S,
    3. Lebuanec H,
    4. et al
    . New class I and II HLA alleles strongly associated with opposite patterns of progression to AIDS. J Immunol 1999;162:6942-6.
    Abstract/FREE Full Text
  7. 7.↵
    1. MacDonald KS,
    2. Fowke KR,
    3. Kimani J,
    4. et al
    . Influence of HLA supertypes on susceptibility and resistance to human immunodeficiency virus type 1 infection. J Infect Dis 2000;181:1581-9.
    Abstract/FREE Full Text
  8. 8.↵
    1. Carrington M,
    2. O'Brien SJ
    . The influence of HLA genotype on AIDS. Annu Rev Med 2003;54:535-51.
    CrossRefMedline
  9. 9.↵
    1. Saah AJ,
    2. Hoover DR,
    3. Weng S,
    4. et al
    . Association of HLA profiles with early plasma viral load, CD4+ cell count and rate of progression to AIDS following acute HIV-1 infection. Multicenter AIDS Cohort Study. AIDS 1998;12:2107-13.
    CrossRefMedlineWeb of Science
  10. 10.↵
    1. Carrington M,
    2. Nelson GW,
    3. Martin MP,
    4. et al
    . HLA and HIV-1: heterozygote advantage and B*35-Cw*04 disadvantage. Science 1999;283:1748-52.
    Abstract/FREE Full Text
  11. 11.↵
    1. Tang J,
    2. Costello C,
    3. Keet IP,
    4. et al
    . HLA class I homozygosity accelerates disease progression in human immunodeficiency virus type 1 infection. AIDS Res Hum Retroviruses 1999;15:317-24.
    CrossRefMedlineWeb of Science
  12. 12.↵
    1. Trachtenberg E,
    2. Korber B,
    3. Sollars C,
    4. et al
    . Advantage of rare HLA supertype in HIV disease progression. Nat Med 2003;9:928-35.
    CrossRefMedlineWeb of Science
  13. 13.↵
    1. Nguyen L,
    2. Chaowanachan T,
    3. Vanichseni S,
    4. et al
    . Frequent human leukocyte antigen class I alleles are associated with higher viral load among HIV type 1 seroconverters in Thailand. J Acquir Immune Defic Syndr 2004;37:1318-23.
    CrossRefMedlineWeb of Science
  14. 14.↵
    1. Doherty PC,
    2. Zinkernagel RM
    . Enhanced immunological surveillance in mice heterozygous at the H-2 gene complex. Nature 1975;256:50-2.
    CrossRefMedline
  15. 15.↵
    1. Kiepiela P,
    2. Leslie AJ,
    3. Honeyborne I,
    4. et al
    . Dominant influence of HLAB in mediating the potential co-evolution of HIV and HLA. Nature 2004;432:769-75.
    CrossRefMedline
  16. 16.↵
    1. Gao X,
    2. Bashirova A,
    3. Iversen AK,
    4. et al
    . AIDS restriction HLA allotypes target distinct intervals of HIV-1 pathogenesis. Nat Med 2005;11:1290-2.
    CrossRefMedlineWeb of Science
  17. 17.↵
    1. Gao X,
    2. Nelson GW,
    3. Karacki P,
    4. et al
    . Effect of a single amino acid change in MHC class I molecules on the rate of progression to AIDS. N Engl J Med 2001;344:1668-75.
    CrossRefMedlineWeb of Science
  18. 18.↵
    1. Hogan CM,
    2. Hammer SM
    . Host determinants in HIV infection and disease. Part 2: genetic factors and implications for antiretroviral therapeutics. Ann Intern Med 2001;134:978-96.
    Abstract/FREE Full Text
  19. 19.↵
    1. Hogg RS,
    2. Yip B,
    3. Chan KJ,
    4. et al
    . Rates of disease progression by baseline CD4 cell count and viral load after initiating triple-drug therapy. JAMA 2001;286:2568-77.
    Abstract/FREE Full Text
  20. 20.↵
    1. Brumme ZL,
    2. Goodrich J,
    3. Mayer HB,
    4. et al
    . Molecular and clinical epidemiology of CXCR4-using HIV-1 in a large population of antiretroviral- naive individuals. J Infect Dis 2005;192:466-74.
    Abstract/FREE Full Text
  21. 21.↵
    1. Brumme ZL,
    2. Henrick BM,
    3. Brumme CJ,
    4. Hogg RS,
    5. Montaner JS,
    6. Harrigan PR
    . Association of the CCR5delta32 mutation with clinical response and >5-year survival following initiation of first triple antiretroviral regimen. Antivir Ther 2005;10:849-53.
    MedlineWeb of Science
  22. 22.↵
    1. Harrigan PR,
    2. Hogg RS,
    3. Dong WW,
    4. et al
    . Predictors of HIV drugresistance mutations in a large antiretroviral-naive cohort initiating triple antiretroviral therapy. J Infect Dis 2005;191:339-47.
    Abstract/FREE Full Text
  23. 23.↵
    1. Chui C,
    2. Cheung PK,
    3. Brumme CJ,
    4. et al
    . HIV VprR77Q mutation does not influence clinical response of individuals initiating highly active antiretroviral therapy. AIDS Res Hum Retroviruses 2006;22:615-8.
    CrossRefMedlineWeb of Science
  24. 24.↵
    ImMunoGeneTics (IMGT) HLA database ftp directory. [Accessed August 2005]. Available at: ftp://ftp.ebi.ac.uk/pub/databases/imgt/mhc/hla/.
  25. 25.↵
    ImMunoGeneTics (IMGT) HLA database ftp directory. [Accessed August 2005]. Available at: http://www.ncbi.nlm.nih.gov/projects/mhc.
  26. 26.↵
    1. Hogg RS,
    2. Heath K,
    3. Bangsberg D,
    4. et al
    . Intermittent use of triplecombination therapy is predictive of mortality at baseline and after 1 year of follow-up. AIDS 2002;16:1051-8.
    CrossRefMedlineWeb of Science
  27. 27.↵
    1. Hetherington S,
    2. Hughes AR,
    3. Mosteller M,
    4. et al
    . Genetic variations in HLA-B region and hypersensitivity reactions to abacavir. Lancet 2002;359:1121-2.
    CrossRefMedlineWeb of Science
  28. 28.↵
    1. Mallal S,
    2. Nolan D,
    3. Witt C,
    4. et al
    . Association between presence of HLAB* 5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse- transcriptase inhibitor abacavir. Lancet 2002;359:727-32.
    CrossRefMedlineWeb of Science
  29. 29.↵
    1. Fernandez S,
    2. Rosenow AA,
    3. James IR,
    4. et al
    . Recovery of CD4+ T cells in HIV patients with a stable virologic response to antiretroviral therapy is associated with polymorphisms of interleukin-6 and central major histocompatibility complex genes. J Acquir Immune Defic Syndr 2006;41:1-5.
    CrossRefMedlineWeb of Science
  30. 30.↵
    1. Martin AM,
    2. Nolan D,
    3. James I,
    4. et al
    . Predisposition to nevirapine hypersensitivity associated with HLA-DRB1*0101 and abrogated by low CD4 T-cell counts. AIDS 2005;19:97-9.
    MedlineWeb of Science
| Table of Contents

This Article

  1. J Infect Dis. (2007) 195 (11): 1694-1704. doi: 10.1086/516789
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 205 (5)
  1. Journal of Infectious Diseases
  1. Alert me to new issues

Most