In patients on highly active antiretroviral therapy (HAART), the time to virologic suppression is one of the important factors shown to influence CD4+ T-cell recovery and clinical outcomes.1 Minimizing the time to virologic suppression and attaining durable suppression, together with the early and overall patterns of viral load decay among virologically suppressed patients, lead to better immunologic and clinical responses to HAART.1 Recently, in a North American, HIV-positive, HAART-treated, adherent cohort of self-identified white (Caucasian) and black (African American) patients, who were followed for ≥6 months after initiation of HAART, we observed that the chemokine (C–C motif) receptor 5 (CCR5) −2459G>A genotype (rs1799987) had a strong association with time to achieve virologic success (TVLS) in black but not in white patients.2 The CCR5 −2459G/A allelic variants are also known as 59029G/A (based on GenBank accession no. U95626) and 303G/A,3 and the A allele–containing most prevalent haplotype is called HHE.3 In our study, black patients carrying the −2459G allele achieved virologic success significantly earlier (AA, GA, and GG: 131, 98, and 76 days, respectively [log-rank P = 0.04]), and this association remained significant in a Cox proportional hazards model that included a total of 10 covariates.2 None of the other genetic polymorphisms, including a 32-base pair deletion in the open reading frame of CCR5 (Δ32, rs333), CCR2 190G>A (Val64Ile, rs1799864), and a total of 5 single-nucleotide polymorphisms (SNPs) in 2 drug-metabolizing enzyme and 1 transporter genes, was significantly associated with TVLS in either racial group of patients. These observations suggested that the association between CCR5 −2459G>A genotype and TVLS of HAART was race specific in this cohort.2
Admixture is a type of gene flow that occurs when individuals from 2 or more parental populations that have been isolated for several generations form a new population.4 Admixed populations are common in North America, because the history of North America has been marked by the encounter of populations from different continents. Recent genetic association analyses in admixed populations, such as African Americans, have highlighted the complexities due to confounding by ancestry.5 In African Americans, systematic differences among individuals occur in the proportion of European ancestry,6 which can be accurately inferred using genome-wide ancestry informative markers (AIMs).7,8 Several recent epidemiological studies in African Americans have taken advantage of this, revealing that gene expression patterns and many disease-related phenotypes vary with the proportion of European ancestry.9–14 Herein, we applied this idea to measure the effects of varying African ancestry on CCR5 −2459G>A genotype-associated TVLS of HAART in a North American cohort.2 The major aims of the study were to determine (1) whether CCR5 −2459G>A genotype is still associated with TVLS of HAART when ancestry, not self-identified race, is considered and (2) whether the association between CCR5 −2459G>A genotype and TVLS of HAART is influenced by varying African ancestry.
MATERIALS AND METHODS
All subjects (n = 393) were adults with confirmed HIV infection, and were receiving care at the Special Immunology Unit of Case Western Reserve University/University Hospitals Case Medical Center, Cleveland, OH. The characteristics of this cohort (self-identified white patients, n = 175; black patients, n = 218), including HAART regimens, adherence, and previous antiretroviral exposure, as well as distribution of various allele frequencies, have been reported previously.2 All patients provided written informed consent for deidentified clinical data and specimen collection, storage, and usage in genetic and nongenetic studies. The data and specimen collection protocol was approved by the Institutional Review Board of University Hospitals Case Medical Center.
We used a set of 128 genome-wide AIMs (see Table S1, Supplemental Digital Content, http://links.lww.com/QAI/A509), which have been characterized and validated for determining continental origin and admixture proportions in common populations in North America.7,8 TaqMan SNP genotyping assays for these AIMs are commercially available (Life Technologies Corporation [Applied Biosystems], Carlsbad, CA). Assays were performed on the TaqMan OpenArray Genotyping System (Applied Biosystems), following the manufacturer's protocol. Each array contained several water blanks as negative controls, and 4 DNA samples (3 European and 1 African American) from the Coriell Cell Repositories (Coriell Institute for Medical Research, Camden, NJ) as positive controls.
The TaqMan Genotyper software v1.0.1 (Applied Biosystems) was used to analyze the primary data. The assay results were validated by re-running 20 of the 393 (5%) samples that were randomly selected. All SNPs with >10% missing data (6 SNPs excluded, 122 SNPs remaining), and all samples with >20% missing SNP data (67 samples excluded, 326 samples remaining) were removed.
Ancestry proportions in these 326 subjects were determined using the Bayesian clustering algorithms implemented in the program STRUCTURE v188.8.131.52,15,16 with the following run parameters: ancestral populations (k) = 3; burn-in iterations = 50,000; and subsequent iterations = 50,000. To have a better assessment of ancestry in these subjects, we spiked our data set with 502 HapMap phase III samples of known ancestry (Northern European, n = 165; Yoruban, n = 167; Japanese + Chinese, n = 170). Among the 326 subjects, those with ≥10% Southeast Asian ancestry were removed (16 samples excluded, 310 remaining).
Virologic success was defined as a viral load <400 copies per milliliter within 6 months of HAART initiation.2 Patients who had not achieved virologic success within this time were censored at 6 months. To describe the probability of achieving virologic success according to CCR5 −2459G>A genotype, Kaplan–Meier curves were plotted. The association between CCR5 −2459G>A genotype and TVLS was examined using Cox proportional hazards regression (Cox model).17 There was no significant interaction between the follow-up time and CCR5 −2459G>A genotype in the primary Cox model, which suggests that the proportional hazards assumption was appropriate. Covariates included in the analysis were age, gender, baseline CD4+ T-cell count, baseline viral load, and CCR2 190G>A genotype.2 All analyses were performed using SAS software (version 9.3, SAS Institute Inc., Cary, NC). A 2-sided P < 0.05 was considered statistically significant.
The characteristics of the whole cohort (n = 310) and various African ancestry proportion groups are presented in Table 1.
Ancestry Proportion Estimates in HapMap Samples and Study Cohort
The STRUCTURE program was run on all samples 3 times independently, and the European and African ancestry proportion estimates were highly concordant across runs (Pearson correlation coefficient, r > 0.999 in all pairwise comparisons). Similar to previous studies that have applied these 128 genome-wide AIMs,7,8 we observed very high proportion of corresponding ancestry in the HapMap samples (see Table S2, Supplemental Digital Content,http://links.lww.com/QAI/A509).
The distribution of ancestry in our final samples (n = 310) is shown in the Figure S1, Supplemental Digital Content,http://links.lww.com/QAI/A509. In these samples, the proportion of European ancestry ranged from 0.003–0.996, the proportion of African ancestry ranged from 0.002–0.993, and the proportion of Southeast Asian ancestry ranged from 0.002–0.096. No subject was characterized by a proportion of African ancestry between 0.125 and 0.452. All subjects with African ancestry ≤0.125 (n = 135) self-identified as white (Caucasian), and all but 2 with African ancestry ≥0.452 (n = 175) self-identified as black (African American) (Table 1). In the subjects with African ancestry ≥0.452, the mean and median proportions of African ancestry were 0.845 and 0.859, respectively; these values are consistent with previous AIMs-based estimates of African ancestry in African American communities in the nearby areas of the Northeast and Midwest regions of the United States.6
Effect of Ancestry on TVLS
We tested for association of ancestry alone with TVLS. Ancestry was not associated with TVLS (≤0.125 vs. ≥0.452, log-rank P = 0.72). We conducted a second set of analysis dividing the entire sample at the median proportion of African ancestry (<0.71 and ≥0.71, n = 155 each). Because it splits the sample evenly, it maximizes the power to detect differences between the 2 groups. Here also, ancestry was not associated with TVLS (<0.71 vs. ≥0.71, log-rank P = 0.90).
Interaction Effect of CCR5 −2459 Genotype and African Ancestry on TVLS
In our previous study, CCR5 −2459G>A genotype had race-specific association with TVLS.2 To examine this issue more thoroughly at the genomic level, we tested for effect modification by including a CCR5 −2459G>A genotype × African ancestry interaction term in a Cox model. When we modeled an interaction between CCR5 −2459G>A genotype and African ancestry ≤0.125 vs. ≥0.452, the effect of the GG genotype on TVLS, compared with the AA genotype (reference), was significantly different between the 2 ancestry proportion groups (P = 0.044). In contrast, the effect of the GA genotype on TVLS, compared with the AA genotype (reference), was not significantly different (P = 0.081).
In a Cox model including an interaction term between CCR5 −2459G>A genotype and African ancestry <0.71 vs. ≥0.71, the effect of the GG genotype on TVLS was stronger (P = 0.018), whereas the effect of the GA genotype was again not significantly different between the 2 ancestry proportion groups (P = 0.094). That the effect of the GA genotype, compared with the AA genotype, was not significantly different between 2 ancestry proportion groups suggests a dose effect of the −2459G allele on TVLS.
CCR5 −2459 Genotype–Associated TVLS Within African Ancestry Proportion Groups
Because we identified a significant interaction between CCR5 −2459G>A genotype and African ancestry, we conducted stratified Kaplan–Meier and Cox model analyses to estimate CCR5 −2459G>A genotype effects within each group of proportions of African ancestry. Kaplan–Meier analysis showed no significant difference in TVLS among CCR5 −2459G>A genotypes in patients with African ancestry ≤0.125 (log-rank P = 0.72). Marginally significant difference was observed in patients with African ancestry ≥0.452 (log-rank P = 0.057) (Figure 1A). In these patients, the association between the −2459G allele and TVLS was more significant after adjusting for age, gender, baseline CD4+ T-cell count, baseline viral load, and CCR2 190G>A genotype (Table 2).
No significant difference was observed in TVLS among CCR5 −2459G>A genotypes in patients with African ancestry <0.71 (log-rank P = 0.67). However, a significant difference was observed in patients with African ancestry ≥0.71 (log-rank P = 0.039) (Figure 1B). Again, the significance of the association between the −2459G allele and TVLS was enhanced with adjustment for age, gender, baseline CD4+ T-cell count, baseline viral load, and CCR2 190G>A genotype (Table 3).
Both Kaplan–Meier and Cox model analyses showed a stronger significant association between CCR5 −2459G>A genotype and TVLS in patients with African ancestry ≥0.71 (Figure 1B; Table 3) than in patients with African ancestry ≥0.452 (Figure 1A; Table 2).
African Ancestry Proportions Among CCR5 −2459 Genotypes
Although the allele frequencies of CCR5 −2459G and −2459A do not differ greatly among continental groups,2,18 we verified in our samples that the genotypes at this locus were not confounded with African ancestry. The mean proportions of African ancestry within a group, determined by a threshold of 0.452 or 0.71, did not differ significantly among CCR5 −2459G>A genotypes (Student t test, P > 0.1 for all comparisons).
In this study, using 128 genome-wide AIMs, we found that the interaction between CCR5 −2459G>A genotype and African ancestry, but not ancestry alone, was significantly associated with TVLS. Furthermore, when we performed analyses stratified by proportions of African ancestry, we found that the association between CCR5 −2459G>A genotype and TVLS was stronger in patients with African ancestry ≥0.71 than in patients with African ancestry ≥0.452. This result was intriguing, mainly because the sample size was reduced by 20 (≥0.452, n = 175; ≥0.71, n = 155). Among these 20 subjects, 19 were self-identified blacks, whereas 1 was self-identified white. The African ancestry proportions were 0.452, 0.49, and 0.534 for 3 of the 20 subjects, and >0.6 but <0.71 for the remaining 17 subjects. Conversely, the European ancestry proportions were 0.539, 0.5, and 0.429, respectively, for the 3 subjects, and ranged from 0.265 to 0.369 for the 17 subjects; these estimates are largely consistent with previous reports.6,19 Thus, these 20 subjects are more heterogeneous than the remaining 155 subjects in terms of their African ancestry proportions, and removing them creates a more homogeneous group with African ancestry proportions ≥0.71, showing a stronger association between CCR5 −2459G>A genotype and TVLS.
Studies investigating the association between genetic markers and HIV/AIDS outcomes have heavily relied on self-identified race classification. To the best of our knowledge, only recently have researchers begun to consider genetic ancestry into their analyses. In a multiethnic HIV-infected women cohort, Frasco et al20 showed evidence of substantial confounding of the association between cytochrome P450 2B6 enzyme polymorphisms and virologic response to efavirenz and nevirapine by genetic ancestry. In a multiethnic HIV-infected men cohort, Nicholaou et al21 showed that the effect of HIV+/HAART status on lipid levels, relative to HIV− status, varies by genetic ancestry (African/European admixed ancestry vs. European ancestry). However, in these studies, there was poor concordance between the self-identified race and genetic ancestry, and they did not elaborate on whether varying genetic ancestry among their subjects had any effect on the phenotypes investigated. Although self-identified race may serve as a good surrogate for genetic ancestry in the clinical setting, our results suggest the importance of considering quantitative assessment of ancestry to determine whether HIV/AIDS outcomes vary among African Americans with varying ancestry.
Our results draw support from several recent studies showing a significant relationship between African ancestry and a variety of disease-related phenotypes.9–11,13,14,22–24 In many of these studies, a specific genetic variant10,22 or multiple common SNPs in a single gene11 were found to be significantly associated with phenotype. Of particular interest, a quantitative assessment of ancestry revealed that the phenotypic effect increased significantly with stronger ancestry.11,13,22,24 Among male subjects, the relative rate of a severe asthma exacerbation was increased more than 4-fold for every 20% increase in the proportion of African ancestry.13 The risk of developing estrogen receptor–negative breast cancer, particularly triple-negative breast cancer, increased 3 times in women in the highest quintile of percent African ancestry relative to those in the lowest quintile.22 Odds of uterine leiomyomata were 34% lower among women in the highest quartile of percent European ancestry relative to those in the lowest quartile.24
An effect of CCR5 −2459G>A genotype on TVLS influenced by African ancestry implies the existence of one or more modifier genetic factors within the African ancestral genome. Although CCR5 is a major determinant, other chemokine receptor genes located on chromosome 3 near CCR5, and chemokine genes located elsewhere are also potentially important in HIV/AIDS pathogenesis.25–28 In addition, outside the chemokine receptor-ligand nexus, host genetic factors that are associated with viral load control have been identified by recent genome-wide association studies.29 Thus, the effect of CCR5 −2459G>A on virologic response to HAART might be due to strong linkage disequilibrium with SNP(s) in other gene(s) on the same chromosome,18 and/or epistatic interaction(s) with SNP(s) elsewhere in the genome. Also, in addition to commonly occurring polymorphisms in CCR5,26 a number of other polymorphisms, mostly non-synonymous, have been discovered in Caucasians and African Americans.30–32 Although these additional polymorphisms occur at low frequencies (<5%), most of them are functionally significant, showing profound effects on both HIV-1 co-receptor and specific ligand-induced functions.33,34 Therefore, by resequencing the CCR5 locus, the possibility should be explored that additional polymorphisms, in conjunction with −2459G>A, could offer insight regarding the race-specific difference in virologic response to HAART.
We acknowledge that our study has some limitations. For certain genotypic associations in African Americans, the effect of ancestry proximal to the gene of interest (local or cis genetic ancestry) is more significant than the effect of ancestry elsewhere in the genome (global or trans genetic ancestry).10–12 However, we were unable to perform a meaningful analysis of the local ancestry in our study, because we have only 9 markers located on chromosome 3, and the marker (rs9809104, C/T) closest to −2459G>A (rs1799987) is 7.25 Mb away (see Table S1, Supplemental Digital Content,http://links.lww.com/QAI/A509). This marker cannot distinguish local African vs. European ancestry with high accuracy (allele T frequency in 1000 Genomes Project African samples = 0.14, and in European samples = 0.79). For such analysis, a high-density African SNP panel providing higher resolution of global and local ancestry35 is needed, whereby the effect of ancestry proximal to CCR5 on virologic response to HAART can be evaluated, and can be compared with the effect of global ancestry. Regarding our results pertaining to African ancestry proportions ≥0.452 and ≥0.71, specifically the 20 subjects with more heterogeneous ancestry proportions, we cannot exclude the fact that residual confounding may exist because of unmeasured ethnic factors (environmental, social, cultural, or behavioral). A number of studies have reported correlation between these factors and adherence vis-à-vis response to HAART.36,37 However, a residual correlation of individual admixture with social or other environmental factors is less likely.38 Furthermore, using a subjective but clinically validated adherence assessment tool, we intentionally selected a stringently high level of adherence (≥90%) as an inclusion criterion for this cohort.2
In conclusion, we observed that stronger African ancestry influenced CCR5 −2459G>A genotype-associated TVLS of HAART in a North American cohort. Further studies to validate this observation in HAART-treated patients in other settings, ideally in a larger and more diverse cohort, and to explore the genomic mechanisms vs. ethnic factors contributing to this influence are important and highly relevant. Furthermore, given that HIV/AIDS continues to disproportionately affect African Americans,39 identifying and better understanding the genetic predictors of antiretroviral treatment response in this population is critical.
The authors thank Dr Sudha Iyengar, Dr Michael Lederman, Dave McNamara, Dr Carolyn Myers, and Alex Barron for helpful discussions and critical evaluation of this article; and Dr Martina Veigl and Debora Poruban (Gene Expression & Genotyping Core Facility) for performing the ancestry informative marker genotyping.
1. Marconi VC, Grandits G, Okulicz JF, et al.. Cumulative viral load and virologic decay patterns after antiretroviral therapy in HIV-infected subjects influence CD4 recovery and AIDS. PLoS One. 2011;6:e17956.
2. Mehlotra RK, Cheruvu VK, Blood Zikursh MJ, et al.. Chemokine (C-C motif) receptor 5 -2459 genotype in patients receiving highly active antiretroviral therapy: race-specific influence on virologic success
. J Infect Dis. 2011;204:291–298.
3. Gonzalez E, Bamshad M, Sato N, et al.. Race-specific HIV-1 disease-modifying effects associated with CCR5
haplotypes. Proc Natl Acad Sci U S A. 1999;96:12004–12009.
4. Chakraborty R. Gene admixture in human populations: models and predictions. Am J Phys Anthropol. 1986;29:1–43.
5. Deo RC, Reich D, Tandon A, et al.. Genetic differences between the determinants of lipid profile phenotypes in African and European Americans: the Jackson Heart Study. PLoS Genet. 2009;5:e1000342.
6. Parra EJ. Admixture in North America. In: Suarez-Kurtz G, ed. Pharmacogenomics in Admixed Populations. Austin, TX: Landes Bioscience; 2007:28–46.
7. Kidd JR, Friedlaender FR, Speed WC, et al.. Analyses of a set of 128 ancestry informative single-nucleotide polymorphisms in a global set of 119 population samples. Investig Genet. 2011;2:1.
8. Kosoy R, Nassir R, Tian C, et al.. Ancestry informative marker sets for determining continental origin and admixture proportions in common populations in America. Hum Mutat. 2009;30:69–78.
9. Cheng CY, Reich D, Haiman CA, et al.. African ancestry and its correlation to type 2 diabetes in African Americans: a genetic admixture analysis in three U.S. population cohorts. PLoS One. 2012;7:e32840.
10. Deo RC, Wilson JG, Xing C, et al.. Single-nucleotide polymorphisms in LPA
explain most of the ancestry-specific variation in Lp(a) levels in African Americans. PLoS One. 2011;6:e14581.
11. Kao WH, Klag MJ, Meoni LA, et al.. MYH9
is associated with nondiabetic end-stage renal disease in African Americans. Nat Genet. 2008;40:1185–1192.
12. Price AL, Patterson N, Hancks DC, et al.. Effects of cis
genetic ancestry on gene expression in African Americans. PLoS Genet. 2008;4:e1000294.
13. Rumpel JA, Ahmedani BK, Peterson EL, et al.. Genetic ancestry and its association with asthma exacerbations among African American
subjects with asthma. J Allergy Clin Immunol. 2012;130:1302–1306.
14. Schlesinger D, Grinberg LT, Alba JG, et al.. African ancestry protects against Alzheimer's disease-related neuropathology. Mol Psychiatry. 2013;18:79–85.
15. Falush D, Stephens M, Pritchard JK. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics. 2003;164:1567–1587.
16. Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics. 2000;155:945–959.
17. Cox DR. Regression models and life tables. J R Stat Soc [B]. 1972;34:187–220.
18. Clark VJ, Dean M. Haplotype structure and linkage disequilibrium in chemokine and chemokine receptor genes. Hum Genomics. 2004;1:255–273.
19. Zakharia F, Basu A, Absher D, et al.. Characterizing the admixed African ancestry of African Americans. Genome Biol. 2009;10:R141.
20. Frasco MA, Mack WJ, Van Den Berg D, et al.. Underlying genetic structure impacts the association between CYP2B6
polymorphisms and response to efavirenz and nevirapine. AIDS. 2012;26:2097–2106.
21. Nicholaou MJ, Martinson JJ, Abraham AG, et al.. HAART
-associated dyslipidemia varies by biogeographical ancestry in the multicenter AIDS cohort study. AIDS Res Hum Retroviruses. 2013;29:871–879.
22. Palmer JR, Ruiz-Narvaez EA, Rotimi CN, et al.. Genetic susceptibility loci for subtypes of breast cancer in an African American
population. Cancer Epidemiol Biomarkers Prev. 2013;22:127–134.
23. Qi L, Nassir R, Kosoy R, et al.. Relationship between hysterectomy and admixture in African American
women. Am J Obstet Gynecol. 2013;208:279.e1–279.e7.
24. Wise LA, Ruiz-Narvaez EA, Palmer JR, et al.. African ancestry and genetic risk for uterine leiomyomata. Am J Epidemiol. 2012;176:1159–1168.
25. An P, Li R, Wang JM, et al.. Role of exonic variation in chemokine receptor genes on AIDS: CCRL2 F167Y
association with pneumocystis pneumonia. PLoS Genet. 2011;7:e1002328.
26. Gonzalez E, Dhanda R, Bamshad M, et al.. Global survey of genetic variation in CCR5, RANTES
, and MIP-1alpha
: impact on the epidemiology of the HIV-1 pandemic. Proc Natl Acad Sci U S A. 2001;98:5199–5204.
27. Gonzalez E, Kulkarni H, Bolivar H, et al.. The influence of CCL3L1
gene-containing segmental duplications on HIV-1/AIDS susceptibility. Science. 2005;307:1434–1440.
28. McDermott DH, Beecroft MJ, Kleeberger CA, et al.. Chemokine RANTES promoter polymorphism affects risk of both HIV infection and disease progression in the Multicenter AIDS Cohort Study. AIDS. 2000;14:2671–2678.
29. van Manen D, van 't Wout AB, Schuitemaker H. Genome-wide association studies on HIV susceptibility, pathogenesis and pharmacogenomics. Retrovirology. 2012;9:70.
30. Ansari-Lari MA, Liu XM, Metzker ML, et al.. The extent of genetic variation in the CCR5
gene. Nat Genet. 1997;16:221–222.
31. Carrington M, Dean M, Martin MP, et al.. Genetics of HIV-1 infection: chemokine receptor CCR5
polymorphism and its consequences. Hum Mol Genet. 1999;8:1939–1945.
32. Carrington M, Kissner T, Gerrard B, et al.. Novel alleles of the chemokine-receptor gene CCR5
. Am J Hum Genet. 1997;61:1261–1267.
33. Blanpain C, Lee B, Tackoen M, et al.. Multiple nonfunctional alleles of CCR5
are frequent in various human populations. Blood. 2000;96:1638–1645.
34. Howard OM, Shirakawa AK, Turpin JA, et al.. Naturally occurring CCR5
extracellular and transmembrane domain variants affect HIV-1 co-receptor and ligand binding function. J Biol Chem. 1999;274:16228–16234.
35. Tandon A, Patterson N, Reich D. Ancestry informative marker panels for African Americans based on subsets of commercially available SNP arrays. Genet Epidemiol. 2011;35:80–83.
36. O'Connor JL, Gardner EM, Mannheimer SB, et al.. Factors associated with adherence amongst 5295 people receiving antiretroviral therapy as part of an international trial. J Infect Dis. 2013;208:40–49.
37. Oh DL, Sarafian F, Silvestre A, et al.. Evaluation of adherence and factors affecting adherence to combination antiretroviral therapy among White, Hispanic, and Black men in the MACS Cohort. J Acquir Immune Defic Syndr. 2009;52:290–293.
38. Mountain JL, Risch N. Assessing genetic contributions to phenotypic differences among “racial” and “ethnic” groups. Nat Genet. 2004;36:S48–S53.
39. Prejean J, Song R, Hernandez A, et al.. Estimated HIV incidence in the United States, 2006-2009. PLoS One. 2011;6:e17502.
African American; ancestry informative markers; CCR5; HAART; virologic success
Supplemental Digital Content
© 2014 by Lippincott Williams & Wilkins