Despite the immune pressure from cytotoxic T lymphocytes (CTLs), most human immunodeficiency virus type 1 (HIV-1)-infected individuals experience high levels of ongoing virus replication. This lack of control is widely thought to be due, at least in part, to escape mutations in CTL epitopes that commonly develop during infection with HIV-1 [1–8]. There is increasing evidence that such adaptations can become apparent at the population level [9,10]. In a large association study among HIV-1-infected individuals in Australia, sequence polymorphism of HIV-1 was found to be associated with particular human leukocyte antigen (HLA) class I alleles [9,11]. In a more recent study involving multiple cohorts, a significant correlation was found between the prevalence of specific CTL escape mutations in circulating HIV strains and the frequency of the restricting HLA allele in the cohort. This phenomenon was not confined to individuals expressing the restricting HLA allele . This implies that the CTL escape mutations observed in viral strains from a single individual may not only reflect the CTL pressure experienced within the individual but also within the population. These observations have raised the question whether CTL escape mutations in HIV-1 are accumulating in the population over time, making the virus less susceptible to CTL responses. Cross-sectional analyses are unable to study the accumulation of escape mutations over time, because the escape mutations that are observed at one moment in time may later revert. Indeed, reversion of CTL escape mutations upon transmission to an HLA discordant recipient has been described in several studies [12,13].
Recently, Bunnik et al.  reported that HIV-1 has evolved towards a more neutralization resistant phenotype over calendar time and at a population level. Here we hypothesized that HIV-1 may also have accumulated adaptations to CTL pressure. We studied the change in the number of CTL epitopes in HIV-1 over time by comparing recently transmitted HIV-1 variants that were isolated from individuals who seroconverted at the beginning of the HIV-1 epidemic to HIV-1 strains isolated from individuals who seroconverted in more recent calendar time. Our data provide the first direct evidence that the number of CTL epitopes in HIV-1 has significantly decreased during 20 years of the HIV-1 epidemic. We also investigated which HLA alleles are the main drivers of HIV-1 adaptation at the population level, and in particular, whether HIV-1 is mostly adapting to HLA alleles that are frequent in the population, as previously suggested [15,16]. Remarkably, the accumulative adaptation was not driven by the most common HLA alleles in the Caucasian population, but by HLA-B alleles that are associated with a low relative hazard of HIV-1 disease progression.
Patients and methods
The study included 27 HIV-1-positive individuals from the Amsterdam Cohort Studies on HIV-1 infection and AIDS (ACS) with a known seroconversion date either in 1985 or in 2005/2006 (12 historical seroconverters and 15 contemporary seroconverters, respectively). Table 1 gives more details on the individual patient characteristics. Selection criteria for this study were infection with HIV-1 subtype B in 1985 or 2005/2006; Caucasian male; HAART-naive; and absence of all known resistance mutations in protease and RT. Although it was not a selection criterion, all patients included in our study were MSM. Historical and contemporary seroconverters did not differ in age [median (range) 34.6 (25.4–46.7) and 32.7 (23.4–48.3), respectively, P > 0.05], viral load [57500 (9800–250 000) and 19 128 (6915–98 204), respectively, P > 0.05] or CD4 cell counts [0.74 (0.25–1.00) and 0.52 (0.21–0.72), respectively, P > 0.05] at the moment of inclusion in the study. Two-digit HLA class I typing was performed for all individuals by sequence-specific primers (SSPs) PCR as described elsewhere . Patients 16 434 and 16 999 contained highly similar viruses (see Phylogenetic analysis section). Therefore HIV-1 sequences from these two patients were not treated as independent samples. In every statistical analysis, only data from either patient 16 434 or patient 16 999 (randomly chosen) were used.
HIV-1 RNA isolation from plasma, cDNA synthesis and sequencing
Viral RNA was isolated from plasma using the QIAgen Viral RNA Mini Kit and reverse transcribed into cDNA with Superscript II RnaseH Reverse Transcriptase (Invitrogen) with outer primers specific for Gag, Nef, and protease/reverse transcriptase (PR/RT). cDNA was amplified using the following primer combinations: Gag: (KVL-064) 5′-GTTGTGTTGTGACTCTGGTAACTAGAGATCCCTCAGA-3′ and (NCRev-2) 5′-CCTTCCTTTCCACATTTCCAACAG-3′ followed by (KVL-066) 5′-TCTCTAGCAGTGGCGCCCGAACAG-3′ and (NCRev-3) 5′-CTTTTTCCTAGGGGCCCTGCAATTT-3′; Nef: (Nef-1fw) 5′-AGCCATAGCAGTAGCTGAGG-3′ and (Nef-1rev) 5′-GCTTATATGCAGGATCTGAGG-3′ followed by (Nef-2fw) 5′-AGCTTGTAGAGCTATTCGCCACA-3′ and (Nef-2rev) 5′-AGCAAGCTCGATGTCAGCAG-3′; PR/RT: (PS1) 5′-TTTTTTAGGGAAAATTTGGCCTTC-3′ and (RTA9) 5′-TAAATTTAGGAGTCTTTCCCCATA-3′ followed by (PS2) 5′-TCCCTCAAATCACTCTTTGGCAAC-3′ and (RTA6subB) 5′-CCATTGGCCTTGCCCCTGCTTCTG-3′.
Bulk PCR products of Gag and Nef resulting from plasma RNA were cloned in the pGEM-Teasy Vector System (Promega) and transformed into DH5α competent cells. PR/RT samples were sequenced directly. In these sequences, mixed bases were detected if they were present in at least 25–30% of the total viral population and were depicted using the International Union of Biochemistry codes. Each RNA sample was PCR-amplified three independent times and each of the three independent PCR products was cloned into pGEM-Teasy separately to avoid resampling artefacts. Two inserts per PCR product were amplified using the nested primers KVL-066 and NCRev-3 (Gag) and 5′-GFP and 3′-GFP (Nef), PCR products were subsequently purified using EXOSAP-IT (USB, Cleveland, Ohio, USA) and sequenced using the ABI prism Big Dye Terminator v1.1/3.1 Cyclesequencing Kit (Applied Biosystems) using the nested PCR primers. Per patient, one to six sequences per protein were analyzed on an Applied Biosystems/Hitachi 3130 ×l Genetic Analyzer.
Phylogenetic trees were generated with the nucleotide sequences of recent and historical seroconverters and reference strains (Los Alamos, reference 2007) that are present in The Netherlands according to the HIV Database (last modified January 2009). Gag (HXB2 Nucleotide Sequence Numbering  790–1878), Pol (2253–3554) and Nef (8797–9417) were concatenated to obtain one long sequence. When more than one sequence per patient was available, we used the most ‘central’ sequence for phylogenetic analyses, that is the sequence with the minimum distance to all other sequences from the same patient [measured by dnadist in PHYLIP package (Felsenstein J. 1993. PHYLIP (Phylogeny Inference Package) version 3.5c., distributed by the author. Department of Genetics, University of Washington, Seattle, USA)]. The sequences were aligned using ClustalW, and the alignment was used to construct phylogenetic trees with PHYML  using the General Reversible Model (GTR) for nucleotide substitution [11,20].
The phylogenetic analysis showed that patients 16 434 and 16 999 contained highly similar viruses (on average 98 and 97% similarity was found for the DNA and protein sequences, respectively). This might be due to a (recent) transmission between the pair, or infection by a common donor, and therefore the data from this pair were not treated as independent samples.
The phylogenetic correction method PhyloD [11,20] was used to determine HIV-1 polymorphisms that are significantly over-represented in recently isolated viral strains compared to historical isolates. For each amino acid within Pol, Gag and Nef a P value representing the significance of the over-representation of a polymorphism in the recent/historical sequences was calculated. This analysis was performed using the online implementation of PhyloD with the Discrete Conditional Escape and Discrete Conditional Attraction  mode to identify possible escape mutations. The input tree to the program was constructed as described above. The set of positions identified using PhyloD was robust to the alternative models used to make the phylogenetic tree and the number of sequences used to generate the phylogenetic tree.
Prediction of cytotoxic T lymphocyte epitopes
Cytotoxic T lymphocyte epitopes were predicted using the proteasomal cleavage/transporter associated with antigen processing (TAP) transport/major histocompatibility complex (MHC) class I combined predictor using the Stabilized Matrix Method (SMM)  available at http://tools.immuneepitope.org. As alternative methods, we used NetChop (for prediction of epitope processing), available at http://www.cbs.dtu.dk/services/NetChop/ [22,23] and NetMHC (for the prediction of HLA-peptide binding) available at http://www.cbs.dtu.dk/services/NetMHC. For MHC binding predictions, the most abundant four-digit HLA type of each HLA serotype was used. CTL epitopes were predicted for the following HLA alleles: A*0101, A*0201, A*0301, A*1101, A*2301, A*2402, A*2601, A*2902, A*3002, A*3101, A*3201 (not available when using NetMHC), A*3301, A*6801, A*6901, B*0702, B*0801, B*1501, B*1801, B*2705, B*3501, B*4001, B*4403, B*4501, B*5101, B*5301, B*5701, B*5801. Cut-off values for processing predictions using SMM predictors were 1.135 for proteasomal cleavage, and −0.56 for TAP transport  or 1.25 when using the combined processing score (which corresponds to a processing rate of 30% in an independent test composed of one million randomly chosen bacterial peptides). For NetChop, a processing threshold of 0.5 was used as suggested originally [22,23]. SMM predictors for MHC binding and NetMHC yield a predicted IC50 value for a given peptide–HLA pair. For all HLA alleles, a predicted binding affinity higher than or equal to 500 nmol/l was used to distinguish binders from nonbinders. The numbers of predicted binders with this fixed threshold were in some cases very different. However, as we compared only early and late epitopes from the same HLA allele, instead of between alleles, the variation in the number of predicted epitopes per allele did not cause a bias in our analysis. Per patient, we predicted CTL epitopes in one to six different HIV-1 sequences (all isolated at the same time point). For every patient we used the mean number of predicted epitopes per HIV-1 sequence in further analyses. For the patients in whom Gag or Nef sequences were lacking (n = 5), the missing proteins were replaced by HXB2 sequences. In none of our tests did these patients differ from the other patients for the number of predicted CTL epitopes. To be able to exclude within-host evolution, we chose HXB2 as a fixed reference strain, and normalized the number of CTL epitopes restricted by any nonself HLA allele by the corresponding number of CTL epitopes (restricted by the same HLA alleles) in HXB2.
Data were analyzed using SPSS 15.0 software (SPSS, Chicago, Illinois, USA). Differences in the number of CTL epitopes between contemporary and historical HIV-1 isolates were analyzed using the Mann–Whitney U test. A P value less than 0.05 was considered statistically significant.
Accumulation of cytotoxic T lymphocyte escape mutations during the epidemic
To investigate whether the number of CTL epitopes in HIV-1 has decreased over time, we sequenced Gag, Nef, protease and RT from clonal HIV-1 variants isolated during the first year of HIV-1 infection from individuals who seroconverted early during the HIV-1 epidemic (in 1985, historical seroconverters) and individuals who seroconverted in more recent calendar time (in 2005/2006, contemporary seroconverters, see Table 1 for more details). In order to compare the number of CTL epitopes in these HIV-1 strains, we scanned all viral sequences for peptides that have been described in the Los Alamos HIV epitope database (http://www.hiv.lanl.gov/content/immunology) because of experimentally verified binding to specific HLA molecules, and peptides that are predicted to bind certain HLA molecules based on peptide prediction methods for proteasomal cleavage, TAP transport, and HLA binding (see Patients and methods section). The use of peptide predictions allowed us to study the evolution of CTL epitopes beyond the somewhat limited set of known HIV-1 epitopes, and thus avoided any possible biases in epitope databases.
As accumulation of HIV-1 adaptations is expected to occur mainly for common HLA alleles in the human population [10,15], we started off analyzing the number of CTL epitopes in these sequences restricted by three common HLA-A (A*0101, A*0201 and A*2402) and three common HLA-B (B*0702, B*0801 and B*4403) alleles in the Caucasian population. Clonal HIV-1 variants from contemporary seroconverters appeared to contain a similar number of CTL epitopes restricted by the most common HLA-A and HLA-B alleles under investigation as compared to HIV-1 variants from historical seroconverters (Fig. 1). Although HIV-1 may show adaptation to the most prevalent HLA type in the population in cross-sectional analyses , these adaptations apparently do not persist in the virus over time in the epidemic.
We also analyzed the number of CTL epitopes restricted by HLA-B*2705 and HLA–B*5701, since these two HLA alleles are strongly associated with relatively slow progression to AIDS , and are therefore thought to evoke strong immunological pressure on the virus. Interestingly, a significant decline in the number of CTL epitopes over time was evident for these two HLA-B alleles associated with slow progression to AIDS. To avoid the possibility that the observed adaptations were due to within-host evolution rather than a reflection of accumulated evolution during the epidemic, we repeated the analyses by excluding HIV-1 sequences from individuals who express the particular HLA allele under investigation, which did not influence the results (data not shown). Together these data suggested that the main drivers of the accumulation of CTL escape mutations over time may be the most protective rather than the most frequent HLA types in the human population.
Adaptation is driven by protective HLA-B alleles
To investigate whether the accumulation of CTL escape mutations during the epidemic is indeed related to the relative hazard of HIV-1 disease progression of the restricting HLA alleles, we extended the analyses by including all HLA-A and B alleles for which reliable prediction programs and the relative hazard  are available (n = 14 for HLA-A and n = 13 for HLA-B, see Patients and methods section). Independent of the method used, we observed a significant decrease during the HIV-1 epidemic in the total number of CTL epitopes presented by HLA-B alleles associated with slow HIV-1 disease progression (relative hazard <1; Fig. 2b; P = 0.003). In contrast, the number of CTL epitopes restricted by HLA-B alleles with a high relative hazard (>1) remained stable during the epidemic. The loss of CTL epitopes restricted by HLA-B alleles with a low relative hazard remained significant when restricting the analysis to only those alleles with a relative hazard less than 0.9 (P = 0.015, data not shown). No significant loss of epitopes was observed for CTL restricted by HLA-A alleles, independent of their relative hazard (Fig. 2a).
As some HIV-1 peptides included in these analyses have not yet been experimentally verified (see Supplementary Fig. S1, http://links.lww.com/QAD/A153), we studied whether the above results were also apparent when including only epitopes that have been described in the Los Alamos HIV database (http://www.hiv.lanl.gov/content/immunology). Indeed, these analyses yielded similar results as reported above for the complete set of known and predicted HIV-1 epitopes (Fig. 2c and d). Our results were also not due to within-host adaptation, because similar results were obtained after exclusion of the CTL epitopes restricted by the HLA alleles expressed by the individuals (Supplementary Fig. S2A and B, http://links.lww.com/QAD/A153, relative hazard <1, P = 0.007). Together these data show that the number of CTL epitopes restricted by protective HLA-B alleles has significantly decreased during the HIV-1 epidemic. In contrast, with none of the prediction methods used, nor when including only experimentally verified epitopes, did we find any significant loss of CTL epitopes restricted by common HLA-A or HLA-B alleles during the epidemic (Fig. 3; Supplementary Fig. S2C and D, http://links.lww.com/QAD/A153).
Our study population contained two contemporary seroconverters who seroconverted very shortly (2 and 23 days) before inclusion. To investigate whether these individuals influenced our conclusions, we repeated the analyses without these two individuals. Again, the number of CTL epitopes restricted by protective HLA-B alleles significantly decreased during the HIV-1 epidemic (P < 0.001), whereas the number of epitopes restricted by HLA-A alleles, HLA-B alleles with a high relative hazard, or common HLA-B alleles did not.
Phylogenetic correction for founder effects
Founder effects could have a large effect on viral polymorphisms associated with HLA molecules . To exclude a possible founder effect, a phylogenetic correction method [11,20] was used to determine HIV-1 polymorphisms that are significantly over-represented among viral variants obtained from contemporary seroconverters as compared to historical seroconverters. This method identified 23 HIV-1 polymorphisms that were either significantly more often or less often present in HIV-1 sequences from contemporary versus historical seroconverters (Supplementary Table 1, http://links.lww.com/QAD/A153, P < 0.05 for all the positions listed). All of the identified polymorphisms were found to lie within or flanking known CTL epitopes (LANL epitope summary tables, www.hiv.lanl.org). Thus, the above-described loss of CTL epitopes during the HIV-1 epidemic does not seem to be caused by founder effects.
Confirmation in an independent cohort
As an independent confirmation of our findings we repeated the analyses using HIV-1 sequences reported in the Los Alamos HIV Database (http://www.hiv.lanl.gov/content/sequence). Selection of all HIV-1 clade B sequences isolated from patients from the USA for which Gag, Nef, protease and RT had been sequenced yielded data from 21 historical seroconverters (sampled before 1988) and 21 contemporary seroconverters (sampled in 2005 or 2006). Again, the number of CTL epitopes restricted by protective HLA-B alleles turned out to have significantly decreased during the HIV-1 epidemic (P = 0.018), whereas the number of epitopes restricted by HLA-A alleles, HLA-B alleles with a high relative hazard, or common HLA-B alleles had not (Supplementary Fig. S3, http://links.lww.com/QAD/A153). These data provide important independent confirmation of our findings. Of note, the HIV-1 sequences selected from the Los Alamos Database came from patients at different stages of disease progression in which the effects of within-host evolution cannot be excluded. More specifically, most of the historical samples came from patients during end-stage HIV disease, whereas most of the recent samples were isolated during early HIV infection. It is all the more striking that a significant loss of CTL epitopes over time could nevertheless be observed in this independent cohort.
Taken together, our data demonstrate that despite the large degree of HLA polymorphism, CTL escape mutations have accumulated during the HIV-1 epidemic, leading to a significant reduction in the total number of CTL epitopes, especially those restricted by HLA-B alleles associated with slow HIV disease progression.
There is increasing evidence that HIV-1 is evolving, not only within the individual, but also at a population level during the HIV-1 epidemic. Recent studies have shown an increasing resistance of HIV-1 from neutralizing antibodies , and an increase in the replication capacity of HIV-1 over time . Although CTL escape mutations can clearly become apparent at the population level, it is still debated to what extent CTL pressure may drive the loss of CTL epitopes in HIV-1 over time [9–11,24,28–31]. Our data show that, despite the large degree of polymorphism of HLA molecules, also HIV adaptations to CTL responses are accumulating over time.
Our finding that HLA-B alleles are the main drivers of HIV-1 adaptation during the epidemic is in line with recent studies showing that HLA-B alleles have a stronger impact on the HIV-1 viral load set point than HLA-A alleles . We would, however, have expected to also find some degree of adaptation to protective HLA-A alleles. The most likely explanation for this lack of adaptation to HLA-A alleles is that the association between the presence of certain HLA-A alleles and protection against progression to AIDS is not as strong as observed for HLA-B alleles. For example, whereas HLA-B27 and HLA-B57 are significantly associated with slow disease progression (P = 0.001 and P = 0.04, respectively, ), none of the HLA-A alleles are significantly associated with slow disease progression .
The dominant role for HLA-B alleles associated with slow disease progression is, however, counterintuitive for two reasons. Firstly, although intuitively HIV-1 would be expected to have adapted most frequently to common HLA alleles , the HLA alleles for which we observed a significant loss of CTL epitopes over time are not common in the Caucasian population. Secondly, viral mutations that successfully escape CTL responses restricted by HLA-B27 and B57 have been shown to be associated with reduced viral fitness [33,34]. Hence such escape mutations are expected to revert to wild type upon transmission to a new host that does not express the specific HLA molecule . In line with this, we observed reversion of the well known HLA-B57 restricted mutation T242N in our cohort, indicative of the fact that certainly not all CTL escape mutations become fixed at the population level. The rate of reversion of CTL escape mutations upon transmission to a host without the restricting HLA background has been used to estimate the fitness cost of escape mutations . Escape mutations in CTL epitopes restricted by rare, protective HLA molecules would hence in fact be expected to be the last to become fixed at the population level. It has been shown, however, that HIV-1 variants that have been under large CTL pressure can accumulate accessory mutations to compensate for the loss of viral fitness [13,36–38]. We hypothesize that the presence of such compensatory mutations may explain why some CTL escape mutations in epitopes restricted by protective HLA-B alleles do not revert upon transmission to a new host (a phenomenon also known as compensatory fixation ), and are thereby accumulating over time. This has important implications for our understanding of HIV-1 evolution. Firstly, it implies that measuring the rate of reversion of CTL escape mutations upon transmission to HLA-disparate hosts is not an accurate measure of the viral fitness cost of a CTL escape mutation. In fact, mutations that revert slowly could be the very ones that caused the largest fitness cost to the virus, and therefore required most compensatory mutations. Secondly, if CTL escape mutations keep on accumulating throughout the HIV-1 epidemic, the protective HLA-B molecules by which they are restricted are expected to become less and less protective, as has been indicated for HLA-B51 in the Japanese population . In line with this, we observed that the average viral load set point, which was lower in HLA-B57-positive compared to HLA-B57-negative historical seroconverters, was no longer significantly different in contemporary seroconverters (van Manen and Schuitemaker, submitted), suggesting that HLA-B57 is indeed losing its protective effect. Moreover, Gras et al.  recently showed in 906 HIV-1-infected individuals that set point viral load levels have increased and CD4 T-cell counts at viral set point have decreased over the past decade of the HIV-1 epidemic in the Netherlands. The latter study included most patients of the current study.
Adaptation to the human immune system through loss of HLA-binding CTL epitopes is not restricted to HIV-1. It has previously been demonstrated that in Epstein–Barr virus (EBV) infection, an HLA-A11-restricted CTL epitope, which is immunodominant in HLA-A11 positive Caucasians, has been lost in viral strains isolated in New Guinea, where HLA-A11 is highly prevalent . Since EBV has been in the human population already for a long time, it is, however, difficult to determine if the CTL response directed against this epitope was ‘protective’. Additionally, herpesvirus proteins have been shown to contain far fewer HLA-binding CTL epitopes than would be expected based on their length and amino acid distribution, suggesting adaptation to CTL responses . These data suggest that evasion of immune detection via the loss of CTL epitopes is a widely used mechanism employed by different viruses. Our current results show that in HIV-1 infection, even 20 years of evolution were sufficient for this phenomenon to become apparent.
Collectively, these data emphasize the potential of HIV-1 to adapt to the human immune system, even in the relatively short period of the HIV-1 epidemic in men, and even despite the large degree of HLA polymorphism.
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 Blood Supply Foundation and the University Medical Center Utrecht, are part of the Netherlands HIV Monitoring Foundation. We would like to thank Margreet Bakker for providing patient data, Marco Brandt for his help with the analyses and Ilka Hoof and Philip Davies for critically reading the manuscript.
Author contributions: I.M.M.S., N.K., F.M., H.S., D.v.B., and J.A.M.B. designed the study; I.M.M.S., M.N., and B.B.-N. performed the experiments; B.B. provided data; I.M.M.S., M.N., H.W.M.v.D., N.K., F.M., C.K., H.S., D.v.B., and J.A.M.B. analyzed data; I.M.M.S., H.W.M.v.D., and C.K. performed bioinformatic analyses; I.M.M.S., H.W.M.v.D., F.M., C.K., H.S., D.v.B., and J.A.M.B. wrote the manuscript.
Conflicts of interest
This study was financially supported by the Landsteiner Foundation for Blood Transfusion Research (LSBR; grant 0317), and the Netherlands Organization for Scientific Research (NWO, grants 016.048.603 and 836.07.002).
1. Phillips RE, Rowland-Jones S, Nixon DF, Gotch FM, Edwards JP, Ogunlesi AO, et al. Human immunodeficiency virus genetic variation that can escape cytotoxic T cell recognition
2. Koenig S, Conley AJ, Brewah YA, Jones GM, Leath S, Boots LJ, et al. Transfer of HIV-1-specific cytotoxic T lymphocytes to an AIDS patient leads to selection for mutant HIV variants and subsequent disease progression
. Nat Med
3. Borrow P, Lewicki H, Wei X, Horwitz MS, Peffer N, Meyers H, et al. Antiviral pressure exerted by HIV-1-specific cytotoxic T lymphocytes (CTLs) during primary infection demonstrated by rapid selection of CTL escape virus
. Nat Med
4. Price DA, Goulder PJ, Klenerman P, Sewell AK, Easterbrook PJ, Troop M, et al. Positive selection of HIV-1 cytotoxic T lymphocyte escape variants during primary infection
. Proc Natl Acad Sci U S A
5. Goulder PJ, Phillips RE, Colbert RA, McAdam S, Ogg G, Nowak MA, et al. Late escape from an immunodominant cytotoxic T-lymphocyte response associated with progression to AIDS
. Nat Med
6. Jones NA, Wei X, Flower DR, Wong M, Michor F, Saag MS, et al. Determinants of human immunodeficiency virus type 1 escape from the primary CD8+ cytotoxic T lymphocyte response
. J Exp Med
7. Goonetilleke N, Liu MK, Salazar-Gonzalez JF, Ferrari G, Giorgi E, Ganusov VV, et al. The first T cell response to transmitted/founder virus contributes to the control of acute viremia in HIV-1 infection
. J Exp Med
8. Fischer W, Ganusov VV, Giorgi EE, Hraber PT, Keele BF, Leitner T, et al. Transmission of single HIV-1 genomes and dynamics of early immune escape revealed by ultra-deep sequencing
. PLoS ONE
9. Moore CB, John M, James IR, Christiansen FT, Witt CS, Mallal SA. Evidence of HIV-1 adaptation to HLA-restricted immune responses at a population level
10. Kawashima Y, Pfafferott K, Frater J, Matthews P, Payne R, Addo M, et al. Adaptation of HIV-1 to human leukocyte antigen class I
11. Bhattacharya T, Daniels M, Heckerman D, Foley B, Frahm N, Kadie C, et al. Founder effects in the assessment of HIV polymorphisms and HLA allele associations
12. Navis M, Matas DE, Rachinger A, Koning FA, van SP, Kootstra NA, et al. Molecular evolution of human immunodeficiency virus type 1 upon transmission between human leukocyte antigen disparate donor-recipient pairs
. PLoS ONE
13. Leslie AJ, Pfafferott KJ, Chetty P, Draenert R, Addo MM, Feeney M, et al. HIV evolution: CTL escape mutation and reversion after transmission
. Nat Med
14. Bunnik EM, Euler Z, Welkers MR, Boeser-Nunnink BD, Grijsen ML, Prins JM, et al. Adaptation of HIV-1 envelope gp120 to humoral immunity at a population level
. Nat Med
15. Trachtenberg E, Korber B, Sollars C, Kepler TB, Hraber PT, Hayes E, et al. Advantage of rare HLA supertype in HIV disease progression
. Nat Med
16. Scherer A, Frater J, Oxenius A, Agudelo J, Price DA, Gunthard HF, et al. Quantifiable cytotoxic T lymphocyte responses and HLA-related risk of progression to AIDS
. Proc Natl Acad Sci U S A
17. Cao K, Hollenbach J, Shi X, Shi W, Chopek M, Fernandez-Vina MA. Analysis of the frequencies of HLA-A, B, and C alleles and haplotypes in the five major ethnic groups of the United States reveals high levels of diversity in these loci and contrasting distribution patterns in these populations
. Hum Immunol
18. Korber B, Foley B, Kuiken C, Pillai S, Sodroski J. Numbering positions in HIV relative to HXB2CG. Human Retroviruses AIDS
19. Guindon S, Gascuel O. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood
. Syst Biol
20. Carlson J, Kadie C, Mallal S, Heckerman D. Leveraging hierarchical population structure in discrete association studies
. PLoS ONE
21. Tenzer S, Peters B, Bulik S, Schoor O, Lemmel C, Schatz MM, et al. Modeling the MHC class I pathway by combining predictions of proteasomal cleavage, TAP transport and MHC class I binding
. Cell Mol Life Sci
22. Kesmir C, Nussbaum AK, Schild H, Detours V, Brunak S. Prediction of proteasome cleavage motifs by neural networks
. Protein Eng
23. Nielsen M, Lundegaard C, Lund O, Kesmir C. The role of the proteasome in generating cytotoxic T-cell epitopes: insights obtained from improved predictions of proteasomal cleavage
24. Schmid BV, Kesmir C, de Boer RJ. The specificity and polymorphism of the MHC class I prevents the global adaptation of HIV-1 to the monomorphic proteasome and TAP
. PLoS ONE
25. O’Brien SJ, Gao X, Carrington M. HLA and AIDS: a cautionary tale
. Trends Mol Med
26. Gao X, Nelson GW, Karacki P, Martin MP, Phair J, Kaslow R, 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
27. Gali Y, Berkhout B, Vanham G, Bakker M, Back NK, Arien KK. Survey of the temporal changes in HIV-1 replicative fitness in the Amsterdam Cohort
28. Goulder PJ, Brander C, Tang Y, Tremblay C, Colbert RA, Addo MM, et al. Evolution and transmission of stable CTL escape mutations in HIV infection
29. Draenert R, Le GS, Pfafferott KJ, Leslie AJ, Chetty P, Brander C, et al. Immune selection for altered antigen processing leads to cytotoxic T lymphocyte escape in chronic HIV-1 infection
. J Exp Med
30. Leslie A, Kavanagh D, Honeyborne I, Pfafferott K, Edwards C, Pillay T, et al. Transmission and accumulation of CTL escape variants drive negative associations between HIV polymorphisms and HLA
. J Exp Med
31. Dilernia DA, Jones L, Rodriguez S, Turk G, Rubio AE, Pampuro S, et al. HLA-driven convergence of HIV-1 viral subtypes B and F toward the adaptation to immune responses in human populations
. PLoS ONE
32. Kiepiela P, Leslie AJ, Honeyborne I, Ramduth D, Thobakgale C, Chetty S, et al. Dominant influence of HLA-B in mediating the potential co-evolution of HIV and HLA
33. Martinez-Picado J, Prado JG, Fry EE, Pfafferott K, Leslie A, Chetty S, et al. Fitness cost of escape mutations in p24 Gag in association with control of human immunodeficiency virus type 1
. J Virol
34. Peyerl FW, Bazick HS, Newberg MH, Barouch DH, Sodroski J, Letvin NL. Fitness costs limit viral escape from cytotoxic T lymphocytes at a structurally constrained epitope
. J Virol
35. Asquith B, Edwards CT, Lipsitch M, McLean AR. Inefficient cytotoxic T lymphocyte-mediated killing of HIV-1-infected cells in vivo
. PLoS Biol
36. Brockman MA, Schneidewind A, Lahaie M, Schmidt A, Miura T, Desouza I, et al. Escape and compensation from early HLA-B57-mediated cytotoxic T-lymphocyte pressure on human immunodeficiency virus type 1 Gag alter capsid interactions with cyclophilin A
. J Virol
37. Crawford H, Prado JG, Leslie A, Hue S, Honeyborne I, Reddy S, et al. Compensatory mutation partially restores fitness and delays reversion of escape mutation within the immunodominant HLA-B*5703-restricted Gag epitope in chronic human immunodeficiency virus type 1 infection
. J Virol
38. Schneidewind A, Brumme ZL, Brumme CJ, Power KA, Reyor LL, O'Sullivan K, et al. Transmission and long-term stability of compensated CD8 escape mutations
. J Virol
39. van Maarseveen NM, Wensing AM, De Jong D, Taconis M, Borleffs JC, Boucher CA, et al. Persistence of HIV-1 variants with multiple protease inhibitor (PI)-resistance mutations in the absence of PI therapy can be explained by compensatory fixation
. J Infect Dis
40. Gras L, Jurriaans S, Bakker M, van SA, Bezemer D, Fraser C, et al. Viral load levels measured at set-point have risen over the last decade of the HIV epidemic in the Netherlands
. PLoS ONE
41. de Campos-Lima PO, Gavioli R, Zhang QJ, Wallace LE, Dolcetti R, Rowe M, et al. HLA-A11 epitope loss isolates of Epstein-Barr virus from a highly A11+ population
42. Vider-Shalit T, Fishbain V, Raffaeli S, Louzoun Y. Phase-dependent immune evasion of herpesviruses
. J Virol