van Manen, Daniëlle; Kootstra, Neeltje A; Boeser-Nunnink, Brigitte; Handulle, Muna AM; van't Wout, Angélique B; Schuitemaker, Hanneke
Humans display a large variability in their susceptibility to HIV-1 infection and in subsequent disease course. In the absence of antiviral therapy, AIDS develops typically within 7–11 years after infection through homosexual contact [1–3], though very rapid (<2 years) or virtually no disease progressions are observed as well . The variable clinical course is determined by both viral and host factors. The emergence of HIV-1 variants that use coreceptor CXCR4 in the course of infection is associated with an accelerated CD4+ T-cell decline and more rapid progression to AIDS [5,6]. Other evidence for viral factors that may influence the clinical course of HIV-1 infection comes from a cohort of long-term nonprogressors (LTNPs) who were all infected with attenuated HIV-1 caused by a deletion in the viral nef gene . The human leukocyte antigen (HLA) type is a strong example of a host factor that is associated with HIV-1 disease course. HLA-B*5701 and HLA-B27 are more prevalent among LTNPs whereas HLA-B35 is associated with an accelerated progression to AIDS [8–10]. The HIV-1 coreceptors and the retroviral restriction factor Trim5α are other examples of host factors in which polymorphisms have been associated with the clinical course of HIV-1 infection [11–14].
Recently, Fellay et al.  published the first genome-wide association (GWA) study in HIV-1-infected individuals. They reported that the minor allelic variants of single-nucleotide polymorphisms (SNPs) rs9264942 and rs2395029 in the gene regions of the HLA-C and HLA complex P5 (HCP5), respectively, were associated with a significantly lower viral load at set point. In addition, they reported a significant association between seven polymorphisms (rs9261174, rs3869068, rs2074480, rs7758512, rs9261129, rs2301753, rs2074479) located in and near the ring finger protein 39 (RNF39) and zinc ribbon domain-containing 1 (ZNRD1) genes and HIV-1 disease progression as defined by CD4 T-cell depletion.
Here, we determined whether the association of these SNPs with viral load at set point could be replicated in the Amsterdam Cohort Studies on HIV infection and AIDS (ACS) in homosexual men with an accurately imputed seroconversion date. In addition, we studied in the same ACS whether these SNPs were associated with the clinical course of infection, also in relation to known progression markers.
We studied HIV-1-infected homosexual men who participate in the Amsterdam Cohort Studies on HIV infection and AIDS (ACS), were enrolled in the cohort between October 1984 and March 1986, and from whom long-term follow-up data are available (every 3 months: collection of clinical and epidemiological data and cryopreservation of serum and peripheral blood mononuclear cells). In the first serum sample taken at entry in the cohort, 728 men tested negative for HIV-1 antibodies and 238 men tested positive for HIV antibodies of whom four refused to participate further; 131 of the negative men subsequently seroconverted during active follow-up (until May 1996). For seroprevalent individuals, an imputed seroconversion date (on average, 18 months before entry into the ACS) was used . AIDS-free survival was similar for persons who seroconverted during the cohort study and seroprevalent persons at entry (log rank P value > 0.2), suggesting a good estimation of the seroconversion date in the latter group. The seroconverter cohort and the seroprevalent cohort both consisted of white men with homosexual contact as risk factor for HIV-1 infection. The mean age at (imputed) seroconversion, as well as viral load and CD4+ T-cell count at set point, was not different between both groups. Finally, heterozygosity for a 32 base-pair deletion in the CCR5 gene had a similar effect on AIDS-free survival in the two cohorts . Therefore, we here used the two cohorts as one study sample (n = 365).
Most seropositive men (n = 243 [67%]) did not receive any early treatment, 70 (19%) received zidovudine monotherapy, 10 (3%) received didanosine monotherapy, and 42 (11%) received other ineffective antiretroviral therapy before AIDS diagnosis. The mean age of participants at the time of (imputed) seroconversion was 34.5 years (range 19.5–57.7 years).
From 335 of these 365 cohort participants, a DNA sample was available for genotyping analysis and consequently these 335 individuals (205 seroprevalent cases and 130 seroconverters) were included in further analyses.
When AIDS according to the Centers for Disease Control and Prevention (CDC) 1993 definition  was used as an end point in Kaplan–Meier survival analysis, 235 individuals had an event, 57 were censored due to loss to follow-up, and 43 were censored because of initiation of HAART. When AIDS-related death, defined as death with AIDS-related malignancy, death with AIDS-opportunistic infections, or death with AIDS-related cause not specified by the treating physician was used as an end point, 180 individuals had an event, 81 were censored due to loss to follow-up, and 74 were censored at initiation of HAART. When a CD4+ T-cell count of 400 cells/μl was used as an end point, 281 had an event, 49 were censored due to loss to follow-up, and five were censored at initiation of HAART. For survival analysis after AIDS diagnosis, 235 individuals were included of whom 171 had an event, 31 were censored due to loss to follow-up, and 33 were censored at initiation of HAART.
The ACS has been conducted in accordance with the ethical principles set out in the declaration of Helsinki, and written informed consent is obtained prior to data collection. The study was approved by the Academic Medical Center institutional medical ethics committee.
We used SNP data for rs2395029, rs3869068, rs2074480, rs7758512, rs9261129, rs2301753 and rs2074479 that we generated in a GWA study in the ACS (van Manen et al. manuscript in preparation) using the Illumina Infinium HumanHap300 BeadChip (Illumina, San Diego, California, USA) . SNP rs9264942 (HLA-C) was absent on this chip and determined by PCR amplification of the relevant region using Taq DNA polymerase (Invitrogen, Carlsbad, California, USA) and primer pair rs9264942-FW (5′-CACAGTCCCAATTCCTTGATTCAG-3′) and rs9264942-RV (5′-CTGTGGAAGGCAGGCTGAGAC-3′). The following amplification cycles were used: 5 min 95°C; 35 cycles of 30 s 95°C, 30 s 50°C, 90 s 72°C; 5 min 72°C. PCR products were subjected to a restriction digest with AlwNI (2 h 37°C; New England Biolabs, Ipswich, Massachusetts, USA) and analyzed on a 2% agarose gel. A T at SNP position results in a undigested PCR product of 330 bp whereas a C at SNP position results in a 133 bp and a 197 bp product after AlwNI digestion. PCR-restriction digest results were confirmed by sequencing of five homozygous TT, five homozygous CC and five heterozygous TC samples, using the ABI prism BigDue Terminator kit V1.1 (Applied Biosystems, Foster City, California, USA), primers rs9264942-FW and rs9264942-RV, and an ABI 3130XL Genetic Analyzer for analysis.
Serum viral load was measured by using a quantitative HIV-1 RNA nucleic acid-based sequence amplification (Organon Teknika, Boxtel, The Netherlands) with electro-chemiluminescently labeled probes . Set point viral load data were available for 332 of 335 (99%) patients. For one participant, the RNA copy number in plasma was below the test threshold of quantification of the assay (50 copies/ml) and was therefore arbitrarily set at 50 copies/ml plasma. Viral load data were analyzed after log10 transformation.
Enumeration of CD4+ T cells was done using flow cytometry. CD4+ T-cell count was first measured at the first visit after entry in the ACS (for seroprevalent patients, this is ∼18 months after the imputed seroconversion date). Set point CD4+ T-cell count data were available for 325 of 335 (97%) patients.
For each of the three SNPs under study, the cohort was divided in homozygotes for the major allele (MAJ), heterozygotes, and homozygotes for the minor allele (MIN). The association of SNP genotypes with the clinical course of HIV-1 infection was tested in Kaplan–Meier survival analyses using AIDS according to the CDC 1993 definition , AIDS-related death, and CD4+ T-cell count below 400 cells/μl blood as end points. In addition, we studied time to death from the moment of AIDS diagnosis onwards. Log rank P value was used to determine significant differences in the clinical course of infection between genotypic groups for each SNP. Univariate and multivariate analyses in a model that included the minor allele of HLA-C rs9264942, the minor allele for HCP5 rs2395029, and the CCR5 Δ32 heterozygous genotype were done from seroconversion onwards. Multivariate analyses in a model that included, CD4+ T cells more than 500 cells/μl blood at approximately 2 years after imputed seroconversion and a viral RNA load less than 104.5 copies/ml plasma at approximately 2 years after imputed seroconversion were performed from 2 years after seroconversion onwards. For Kaplan–Meier survival analysis and the univariate and multivariate analyses, left truncation of follow-up time was performed for time between imputed seroconversion date and the first seropositive visit using S-PLUS 8 (Insightful Corporation, Seattle, Washington, USA).
The association of each of the three SNPs with viral load or CD4+ T-cell count at set point was tested using one-way analysis of variation (ANOVA) for SNPs with a MAJ, heterozygote, and MIN group (rs9264942 and rs2074479) or Student's t-test for the SNP with only a MAJ and heterozygote group (rs2395029) as implemented in GraphPad Prism 5 (GraphPad Software, La Jolla, California, USA).
Genotype distribution for rs9264942 in the HLA-C gene region, rs2395029 in the HCP5 gene region, and rs2074479 in the RNF39/ZNRD1 gene region in the Amsterdam Cohort Studies on HIV infection and AIDS
In our study population of 335 HIV-1-infected men who participate in the ACS, we first analyzed the prevalence of eight of the nine SNPs previously described to be associated with viral load at set point or the clinical course of infection . The eight SNPs under study are rs9264942 in the HLA-C gene region, rs2395029 in the HCP5 gene region, and rs3869068, rs2074480, rs7758512, rs9261129, and rs2301753, rs2074479 in the RNF39/ZNRD1 gene region. Considering the high linkage disequilibrium between these latter SNPs in the ACS (r2 = 0.87–1), only the results for rs2074479 are shown. Table 1 shows the allele frequencies and the percentage of individuals who were homozygous for the major allele (MAJ), heterozygous, or homozygous for the minor allele (MIN). No genotype data deviated from Hardy–Weinberg equilibrium. There was no difference in mean age at the imputed seroconversion date between individuals in the MAJ, heterozygote, and MIN groups for the three SNPs under study (data not shown).
Replication of the association of HLA-C rs9264942 and HCP5 rs2395029 genotypes with viral load at set point
The viral load at set point is considered to be established 18–24 months after seroconversion and is highly predictive for the clinical course of infection [20,21]. Viral load set point was one of the phenotypes used in the GWA analysis performed by Fellay et al. . To replicate their findings, we first evaluated the association between the selected SNPs and viral load at set point in the ACS. Indeed, the two SNPs reported to be associated with viral load at set point (HLA-C rs9264942 and HCP5 rs2395029) were also significantly associated with viral load at set point in the ACS (Fig. 1a), with P values of 2.94 × 10−2 and 5.59 × 10−3, respectively. The progression SNP rs2074479 in the RNF39/ZNRD1 gene region, for which a weaker association with viral load at set point was reported by Fellay et al.  (P =7.11 × 10−3), was not associated with viral load at set point in ACS participants (Fig. 1a).
Next, we analyzed whether these selected SNPs associated with the CD4+ T-cell count at 2 years after the imputed seroconversion date. In agreement with the association with viral load at set point, we observed significant associations between the minor alleles of HLA-C rs9264942 and HCP5 rs2395029 and a higher CD4+ T-cell count at 2 years after seroconversion, with P values of 5.66 × 10−8 and 6.35 × 10−3 (Fig. 1b). In line with the association of SNP rs2074479 with protection from HIV-1 disease progression as defined by CD4+ T-cell depletion, there was also a strong association between the minor allele of SNP rs2074479 in the RNF39/ZNRD1 gene and a higher CD4+ T-cell count at 2 years after seroconversion with a P value of 9.82 × 10−3 (Fig. 1b).
Association of HLA-C rs9264942 and HCP5 rs2395029 genotypes with clinical course of infection
Kaplan–Meier survival analysis in all 335 HIV-1-seropositive participants showed a significantly prolonged AIDS-free survival in carriers of the MIN and heterozygote genotype compared with carriers of the MAJ genotype of HLA-C rs9264942 [(log rank P = 7.64 × 10−4; median survival time in years: MAJ = 6.6 ± 0.3, heterozygote = 8.1 ± 0.8 and MIN = 8.7 ± 2.7); Fig. 2a]. A similar effect was observed when AIDS-related death was used as an end point in survival analysis (log rank P = 2.56 × 10−3; median survival time in years: MAJ = 9.4 ± 0.4, heterozygote = 12.6 ± 0.8 and MIN = 11.3 ± 1.7; Fig. 2a). When the first CD4+ T-cell count below 400 cells/μl blood was used as an end point, an even stronger prolonged survival was observed for individuals who carried the MIN or heterozygote genotype (log rank P = 5.67 × 10−5; median survival time in years: MAJ = 4.1 ± 0.2, heterozygote = 5.8 ± 0.6 and MIN = 6.3 ± 0.8; Fig. 2a), suggesting that this SNP genotype may have an effect mainly early in the clinical course of HIV infection. The minor allele of the SNP was not associated with prolonged survival after AIDS diagnosis (Fig. 2), though this may be due to insufficient power due to the small risk group for this analysis.
None of the ACS participants was homozygous for the minor allele of SNP rs2395029 in the HCP5 locus. Therefore, we compared individuals homozygous for the major allele with individuals heterozygous for this genotype in Kaplan–Meier survival analysis. Carriers of minor allele of SNP rs2395029 in the HCP5 gene had a prolonged AIDS-free survival and a prolonged time to AIDS-related death (log rank P = 3.17 × 10−3 and P = 3.91 × 10−3, respectively; median AIDS-free survival time in years: MAJ = 6.6 ± 0.3, heterozygote = 13.4 ± 3.2, median survival time in years MAJ = 10.2 ± 0.6, heterozygote > 15; Fig. 2b). The minor allele was also associated with a prolonged time to the first moment that the CD4+ T-cell count dropped below 400 cells/μl blood (log rank P = 1.97 × 10−4; median survival time in years: MAJ = 4.5 ± 0.2, heterozygote = 8.8 ± 2.6; Fig. 2b), suggesting that also this SNP genotype may have an effect early in the course of infection. The minor allele of this SNP was not associated with prolonged survival after AIDS diagnosis (Fig. 2) though here also this may be due to insufficient power.
SNP rs2074479 in the RNF39/ZNRD1 gene region was not associated with the clinical course of HIV-1 infection in the ACS (data not shown).
Finally, none of the SNPs was associated with time to the first appearance of CXCR4-using (X4) HIV-1 variants or the prevalence of X4 variants (data not shown).
Predictive value of HLA-C rs9264942 and HCP5 rs2395029 genotypes for HIV-1 disease course
To study the predictive value of the SNP genotypes, we performed univariate Cox proportional-hazard analyses at the time of seroconversion for HLA-C rs9264942, and separately for HCP5 rs2395029, as only these SNPs were significantly associated with disease course in the ACS. We observed a relative hazard for progression to AIDS of 0.78 [95% confidence interval (CI) 0.68–0.89] for carriers of the minor allele of HLA-C rs9264942 (Table 2). The minor allele of HCP5 rs2395029 provided a relative hazard of 0.38 (95% CI 0.20–0.74) for progression to AIDS. A similar protective effect of both minor alleles was observed when AIDS-related death was used as end point in the analysis (Table 2).
Multivariate analysis indicated that HLA-C rs9264942 and HCP5 rs2395029 were both independent predictors for delayed progression to AIDS or AIDS-related death (Table 2) from seroconversion onwards. Moreover, this predictive effect was also independent of the CCR5 Δ32 heterozygous genotype that has previously been demonstrated to be associated with protection from disease progression (Table 3) [11–13].
However, independent predictive values of both HLA-C rs9264942 and HCP5 rs2395029 were lost when viral load or CD4+ T-cell count or both at set point were included in a model analyzing the time from 2 years after seroconversion, in other words after the set points for CD4+ T-cell count and viral load were established (Table 4).
Cumulative effect of HLA-C rs9264942 and HCP5 rs2395029 genotypes on HIV-1 disease course
Using the Kaplan–Meier product-limit method, we analyzed whether clinical progression to AIDS was further delayed with an accumulating number of independent protective genotypes fulfilled (Fig. 3). Interestingly, 19 of 20 individuals, who had the minor allele for HCP5 rs2395029, also had the minor allele of HLA-C rs9264942. Therefore, the group of individuals with only one protective SNP genotype consisted of 112 individuals who had the protective SNP genotype in the HLA-C region and only one individual who had the minor allele for rs2395029 in the HCP5 gene region. This group had a relative hazard of 0.66 (95% CI 0.50–0.88) for progression to AIDS and a relative hazard of 0.67 (95% CI 0.48–0.92) for progression to AIDS-related death. This risk was even further decreased for individuals who had the minor alleles of both HLA-C rs9264942 and HCP5 rs2395029 [relative hazard 0.59 (95% CI 0.42–0.82) and 0.48 (95% CI 0.29–0.79)] for AIDS and AIDS-related death, respectively. Addition of the CCR5 Δ32 heterozygous genotype only slightly influenced the relative hazard of combined genotypes. As only three individuals had the minor allele for HLA-C rs9264942 and HCP5 rs2395029 in combination with a heterozygous CCR5 Δ32 genotype, these three individuals were included in the group who had two out of three protective genotypes. For this combined group, the relative hazards for AIDS and AIDS-related death were 0.56 (95% CI 0.44–0.47) and 0.47 (95% CI 0.34–0.64), respectively.
GWA studies allow assessment of genetic variability in relation to a phenotype without a priori assumptions about phenotype-related genes. The results of the HIV-1 host control study performed by Fellay et al.  is the first such example in the HIV-1 field. In this GWA study, SNPs in the gene regions of HLA-C and HCP5 were found to be significantly associated with viral load at set point. In addition, they observed SNPs in the RNF39/ZNRD1 gene region to be significantly associated with disease progression as defined by CD4+ T-cell depletion.
In our present study, we were able to confirm a significant association between viral load at set point for SNPs HLA-C rs9264942 and HCP5 rs2395029, respectively. In addition, we observed a strong association between these SNPs and other clinical end points. However, we did not observe an effect of rs3869068, rs2074480, rs7758512, rs9261129, rs2301753, and rs2074479 in RNF39/ZNRD1 on any of the clinical end points analyzed, except for CD4+ T-cell count at set point. Taking into account the relatively small number of study samples, the possibility remains that this is due to type 2 error. Additional studies with larger sample sizes maybe necessary to confirm the association between the RNF39/ZNRD1 gene region and HIV-1 disease course.
Moreover, our cohort consists of white men who are all infected with subtype B HIV-1 most likely transmitted through homosexual contact. It remains to be established whether the association of the SNPs in the RNF39/ZNRD1 gene region and HIV-1 disease course maybe replicated in cohorts with other ethnicities, HIV-1 subtypes, and HIV-1 risk factors.
The minor alleles of rs9264942 and rs2395029 in the HLA-C and HCP5 gene regions, respectively, were independent predictors of disease progression. However, their effect was noticeably reduced when viral load at set point was included as a covariate in multivariate analysis. This may imply that the effect of these two gene loci on HIV-1 disease course is partly mediated by influencing viral load, in line with their strong association with viral load at set point, though it cannot be excluded that a lack of power is responsible for the reduction of their effect in multivariate analysis.
It remains to be established how rs9264942 in the HLA-C gene region and rs2395029 in the HCP5 gene region influence HIV-1 disease course. The SNP in HLA-C is associated with HLA-C expression levels and may thus facilitate the host immune response to viral peptides. The SNP in HCP5 is in high linkage disequilibrium with HLA-B*5701 (r2 = 1 in our study population), which has been strongly associated with prolonged AIDS-free survival . Similar to our observation for HCP5 rs2395029, HLA-B*5701 has an effect early in the course of HIV-1 infection but not in the phase after AIDS has developed . It seems likely that the association of rs2395029 in the HCP5 gene region is in fact the already known effect of HLA-B*5701. However, further investigation is warranted to identify the most likely functional variants.
It may seem rather disappointing that a GWA study on HIV-1 disease course has delivered so few new host factors involved in the clinical course of disease. Indeed, the likely possibility that the effect of HCP5 rs2395029 is related to the effect of HLA-B*5701, and the absent replication of the effect of rs2074479 in RNF39/ZNRD1 in our cohort, may imply that the HLA-C gene region is the only novel host factor involved in HIV-1 disease course. However, the GWA study performed by Fellay et al.  was performed in a relatively small study group. Increasing the power of GWA studies by combining different cohorts of HIV-1-infected individuals may definitely reveal new host factors involved in the clinical course of infection.
The Amsterdam Cohort Studies on HIV infection and AIDS, a collaboration between the Amsterdam Health Service, the Academic Medical Center of the University of Amsterdam, Sanquin Research, and the University Medical Center Utrecht, are part of the Netherlands HIV Monitoring Foundation and financially supported by the Netherlands National Institute for Public Health and the Environment. We acknowledge funding from the Netherlands Organization for Scientific Research (TOP, registration number 9120.6046) and the European Union (Marie Curie International Reintegration Grant 029167).
Author contributions: A.B.W. and H.S. designed the study, B.B.N., M.A.M.H., and D.v.M. generated the data. D.v.M., N.A.K. and A.B.W. analyzed the data, D.v.M. and H.S. wrote the manuscript with helpful comments of N.A.K. and A.B.W.
There was no competing financial interests.
1. Munoz A, Sabin CA, Phillips AN. The incubation period of AIDS. AIDS 1997; 11(Suppl A):S69–S76.
2. Time from HIV-1 seroconversion to AIDS and death before widespread use of highly-active antiretroviral therapy: a collaborative re-analysis. Collaborative Group on AIDS Incubation and HIV Survival including the CASCADE EU Concerted Action. Concerted Action on SeroConversion to AIDS and Death in Europe. Lancet 2000; 355:1131–1137.
3. Veugelers PJ, Page KA, Tindall B, Schechter MT, Moss AR, Winkelstein W, et al. Determinants of HIV disease progression among homosexual men registered in the Tricontinental Seroconverter Study. Am J Epidemiol 1994; 140:747–758.
4. Klein MR, Miedema F. Long-term survivors of HIV-1 infection. Trends Microbiol 1995; 3:386–391.
5. Koot M, Keet IPM, Vos AHV, De Goede REY, Roos MThL, Coutinho RA, et al. Prognostic value of human immunodeficiency virus type 1 biological phenotype for rate of CD4+ cell depletion and progression to AIDS. Ann Intern Med 1993; 118:681–688.
6. Connor RI, Sheridan KE, Ceradini D, Choe S, Landau NR. Change in coreceptor use correlates with disease progression in HIV-1-infected individuals. J Exp Med 1997; 185:621–628.
7. Deacon NJ, Tsykin A, Solomon A, Smith K, Ludford-Menting M, Hooker DJ, et al. Genomic structure of an attenuated quasi species of HIV-1 from a blood transfusion donor and recipients. Science 1995; 270:988–991.
8. Navis M, Schellens I, van Baarle D, Borghans J, van Swieten P, Miedema F, et al. Viral replication capacity as a correlate of HLA B57/B5801-associated nonprogressive HIV-1 infection. J Immunol 2007; 179:3133–3143.
9. Migueles SA, Sabbaghian MS, Shupert WL, Bettinotti MP, Marincola FM, Martino L, et al. HLA B*5701 is highly associated with restriction of virus replication in a subgroup of HIV-infected long term nonprogressors. Proc Natl Acad Sci U S A 2000; 97:2709–2714.
10. Gao X, Bashirova A, Iversen AK, Phair J, Goedert JJ, Buchbinder S, et al. AIDS restriction HLA allotypes target distinct intervals of HIV-1 pathogenesis. Nat Med 2005; 11:1290–1292.
11. van Rij RP, De Roda Husman AM, Brouwer M, Goudsmit J, Coutinho RA, Schuitemaker H. Role of CCR2 genotype in the clinical course of syncytium-inducing (SI) or non-SI human immunodeficiency virus type 1 infection and in the time to conversion to SI virus variants. J Infect Dis 1998; 178:1806–1811.
12. De Roda Husman AM, Koot M, Cornelissen M, Brouwer M, Broersen SM, Bakker M, et al. Association between CCR5 genotype and the clinical course of HIV-1 infection. Ann Intern Med 1997; 127:882–890.
13. Ioannidis JP, Rosenberg PS, Goedert JJ, Ashton LJ, Benfield TL, Buchbinder SP, et al. Effects of CCR5-Delta32, CCR2-64I, and SDF-1 3′A alleles on HIV-1 disease progression: an international meta-analysis of individual-patient data. Ann Intern Med 2001; 135:782–795.
14. van Manen D, Rits MA, Beugeling C, van Dort KA, Schuitemaker H, Kootstra NA. The effect of Trim5 polymorphisms on the clinical course of HIV-1 infection. PLoS Pathog 2008; 4:e18.
15. Fellay J, Shianna KV, Ge D, Colombo S, Ledergerber B, Weale M, et al. A whole-genome association study of major determinants for host control of HIV-1. Science 2007; 317:944–947.
16. Van Griensven GJP, De Vroome EMM, Goudsmit J, Coutinho RA. Changes in sexual behaviour and the fall in incidence of HIV infection among homosexual men. BMJ 1989; 298:218–221.
17. Steemers FJ, Gunderson KL. Whole genome genotyping technologies on the BeadArray platform. Biotechnol J 2007; 2:41–49.
18. van Gemen B, van Beuningen R, Nabbe A, Van Strijp D, Jurriaans S, Lens P, et al. A one-tube quantitative HIV-1 RNA NASBA nucleic acid amplification assay using electrochemiluminescent (ECL) labelled probes. J Virol Methods 1994; 49:157–168.
19. Centers for disease control: 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. Morb Mortal Wkly Rep 1993; 41:1–19.
20. Mellors JW, Rinaldo CR Jr, Gupta P, White RM, Todd JA, Kingsley LA. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science 1996; 272:1167–1170.
21. De Wolf F, Spijkerman I, Schellekens PThA, Langendam M, Kuiken CL, Bakker M, et al. AIDS prognosis based on HIV-1 RNA, CD4+ T cell count and function: markers with reciprocal predictive value over time after seroconversion. AIDS 1997; 11:1799–1806.
© 2009 Lippincott Williams & Wilkins, Inc.