Share this article on:

Influence of Major HIV-1 Protease Inhibitor Resistance Mutations on CTL Recognition

Mueller, Sandra M PhD*; Spriewald, Bernd M MD; Bergmann, Silke MTA*; Eismann, Kathrin MTA*; Leykauf, Melanie MTA*; Korn, Klaus MD; Walter, Hauke MD; Schmidt, Barbara MD; Arnold, Marie-Luise MTA*; Harrer, Ellen G MD*; Kaiser, Rolf PhD§; Schweitzer, Finja MSc§; Braun, Patrick MSc; Reuter, Stefan MD; Jaeger, Hans MD#; Wolf, Eva MSc**; Brockmeyer, Norbert H MD††; Jansen, Klaus MSc††; Michalik, Claudia MA‡‡; Harrer, Thomas MD*the German Competence Network for HIV/AIDS

JAIDS Journal of Acquired Immune Deficiency Syndromes: February 1st, 2011 - Volume 56 - Issue 2 - p 109-117
doi: 10.1097/QAI.0b013e3181fe946e
Basic and Translational Science

Background: HIV-1 protease is subjected to dual selection pressure exerted by protease inhibitors (PIs) and cytotoxic T lymphocytes (CTL). Recently, we identified KMIGGIGGF (KF9) as a HLA-B*1501-restricted CTL epitope, including several major PI resistance mutations (M46I/L, I47A/V, G48V, I50V). To assess potential interactions between KF9-specific CTL and emergence of these important resistance mutations, we studied CTL recognition of the mutations and analyzed protease sequences in an HLA-I-typed patient cohort.

Methods: CTL recognition of KF9 and resistance mutations in KF9 were studied in 38 HLA-B*1501-positive HIV-1-infected patients using variant KF9 peptides in interferon-γ enzyme-linked immunospot assays. Protease sequences were analyzed in 302 HLA-I-typed HIV-1-infected patients.

Results: G48V abolished KF9 recognition by CTL in all patients. Furthermore, M46I, I47A, and I50V could impair or abolish CTL recognition in many patients. In contrast, M46L and I47V showed good CTL recognition in nearly all patients. HIV-1 protease sequence analysis showed no statistical correlation between the occurrence of resistance mutations in KF9 and HLA-B*1501. Viral load in patients failing therapy with KF9 mutations was significantly lower in HLA-B*1501-positive patients in comparison with HLA-B*1501-negative patients.

Conclusions: PI mutations, G48V, M46I, and I47A, can abrogate CTL recognition, indicating potential interactions between development of drug resistance and CTL response. However, we could not find evidence that development of these PI mutations is influenced by KF9-specific CTL.

From the *Department of Internal Medicine 3, Institute of Clinical Immunology; and †Department of Internal Medicine 5, University Hospital Erlangen, University of Erlangen-Nuremberg, Erlangen, Germany; ‡Institute of Clinical and Molecular Virology, German National Reference Center for Retroviruses, University of Erlangen-Nuremberg, Erlangen, Germany; §Institute of Virology, University of Cologne, Cologne, Germany; ∥Praxis Dr. med Heribert Knechten, Aachen, Germany; ¶Department of Gastroenterology, University of Dusseldorf, Dusseldorf, Germany; #MVZ Karlsplatz, HIV Research and Clinical Care Centre, Munich, Germany; **MUC Research, Munich, Germany; ††Department of Dermatology, Venereology and Allergology, Ruhr-University Bochum, Bochum, Germany; and ‡‡ ZKS, Clinical Trial Centre, Cologne, Germany.

Received for publication June 4, 2010; accepted September 24, 2010.

Supported by the Deutsche Forschungsgemeinschaft (DFG) (DFG grant HA 2331/2-1), Interdisziplinäres Zentrum für Klinische Forschung (IZKF) Erlangen (project A27), Hector Foundation, Erlanger Leistungsbezogene Anschubfinanzierung und Nachwuchsförderung, (ELAN-Fond), and the Bundesministerium für Bildung und Forschung (BMBF) supporting the patient cohort of the Competence Network for HIV/AIDS (grant number 01 KI 0501).

Partly presented at the Conference on Retroviruses and Opportunistic Infections (CROI) 2009, February 8-11, 2009, Montreal, Canada.

The authors have no conflicts of interest to disclose.

Correspondence to: Sandra M. Mueller, PhD, Department of Internal Medicine 3, University Hospital Erlangen, Krankenhausstraße 12, 91054 Erlangen, Germany (e-mail:

Back to Top | Article Outline


HIV-1 protease (PR) is an important target for antiretroviral drug therapy (ART) in HIV-1-infected patients. Being one of the major drug targets, the PR is subjected to a strong selection pressure exerted by ART. Resistance to the individual protease inhibitor (PI) is characterized by the accumulation of defined patterns of major and minor drug resistance mutations.1-7

Furthermore, several studies provide evidence that immunological pressure exerted by HLA class I-specific cytotoxic T lymphocytes (CTL) shapes the sequence of HIV-1 by selection of CTL escape mutations.8-11 Others and we described a significant interaction between the CTL-mediated immune response and ART.11-20

The HLA-B*1501-restricted epitope KMIGGIGGF (KF9, amino acids 45-53 of HXB2)18 that was recently defined by our group is of special interest because it comprises several major drug resistance mutations (M46I, M46L, I47A, I47V, G48V, and I50V). The KF9 epitope is highly conserved in all HIV-1 subtypes.21 Nevertheless, PI-based therapy can induce distinct resistance mutations in KF9. The mutations M46I/L, I47A/V, G48V, and I50V have been described as major drug resistance mutations for indinavir (IDV; M46I/L), lopinavir (LPV; I47A/V), saquinavir (SQV; G48V), fosamprenavir (FPV; I50V), darunavir (I47V, I50V), and tipranavir (TPV, I47V) and as minor mutations for atazanavir (ATV) and nelfinavir (NFV).5

The aim of this study was to investigate the interaction between the KF9-specific CTL response and the development of drug resistance in the functionally important KF9 epitope. For that purpose, we analyzed the influence of drug resistance mutations on recognition by KF9-specific CTL and determined the recognition patterns of KF9-specific CTL in a cohort of HLA-B*1501-positive patients comprising therapy-naive patients and patients on ART with and without PI. We restricted our analysis to HLA-B*1501-positive patients as this HLA-B15 subtype constitutes the majority of the serologically typed HLA-B62-positive patients in Europe.22,23 To assess potential immune selection of drug resistance mutations by KF9-specific CTL, we analyzed the viral PR sequence in a cohort of HLA-typed HIV-1-infected patients (n = 302).

Back to Top | Article Outline


Patients Analyzed for KF9 Recognition

PR-specific CTL responses were analyzed by γ-interferon enzyme-linked immunospot (ELISPOT) assays in 38 HIV-1-infected HLA-B*1501-positive subjects of the Erlangen HIV-1 cohort. The analyses were restricted to HLA-B*1501 because HLA-B*1501 is the most frequent HLA-B15 allele coding for the serologically determined HLA-B62 in Europe. The median time of documented HIV-1 infection was 8 years (range, 1 month to 25 years since HIV-1 diagnosis). At the time of their last visit during the study, 71% of the patients had undetectable viral loads (VL) (<20 copies/mL). The median HIV-1 VL of the remaining patients was 290 copies per milliliter (range, 20-38,000 copies/mL). The median CD4 cell count was 595 cells per microliter (range, 27 to 1858 cells/μL). At the time of ELISPOT analysis, 27 of 38 patients were taking PIs or had taken PIs in the past. Among the 11 PI-naive patients, 2 were without ART and 9 had been exposed only to PI-sparing therapies. The study was approved by the Ethics Committee of the Medical Faculty of the University of Erlangen-Nuremberg, and the patients gave informed consent. For 34 of these 38 patients, sequence analyses of HIV-1 PR were available. For 28 patients PR sequences were obtained in the context of resistance testing. On average, genotypic resistance testing was performed 2 years before the ELISPOT analyses were done in the course of this study (median time, 694 days; range, 27 days to 6 years). Because of low VL under efficient ART, PR sequences for 6 patients were obtained from proviral DNA of peripheral blood mononuclear cells (PBMCs) instead of plasma RNA.

Back to Top | Article Outline

Patients Analyzed for Correlations Between PR Mutations and HLA Class I Alleles

For statistical analysis, PR sequences from 47 HLA-B*1501-positive patients were compared with PR sequences from 255 HLA-B*1501-negative patients from the Erlangen HIV-1 cohort and the cohort of the German Competence Network for HIV/AIDS.24,25 The 47 HLA-B*1501-positive patients included 34 patients tested in ELISPOT analysis and additional 13 HLA-B*1501-positive patients who could not be tested by ELISPOT due to loss of follow up or due to lack of suitable samples.

In 9 of the 47 HLA-B*1501-positive patients, PR sequences were derived before the start of ART. In the other 38 patients, PR was sequenced in the context of resistance testing either during virologic failure on ART or during treatment interruptions. Of these 38 patients, 34 patients where on PI-containing therapies at the time of resistance testing and 4 patients were on PI-free regimens. Of the 255 HLA-B*1501-negative patients, PR was sequenced in 100 patients before the start of ART and in 155 patients on failing therapies or during therapy interruption. One hundred forty three of the HLA-B*1501-negative patients on ART had previously received PI treatment or were currently on therapies including PIs. The remaining 12 patients were PI naive.

Back to Top | Article Outline

Analysis of Viral PR Sequences

Viral RNA was isolated from plasma using the QIAamp Viral RNA Kit (Qiagen, Hilden, Germany) and was reverse transcribed by random hexamers or sequence-specific primers and Superscript III reverse transcriptase (Invitrogen, Karlsruhe, Germany). For sequence analysis, the relevant part of the pol gene including the PR was amplified from patients' plasma (41/47 HLA-B*1501-positive patients and 242/255 HLA-B*1501-negative patients) and was sequenced using the ViroSeq HIV-1 Genotyping System (Abbott Diagnostics, Wiesbaden, Germany). In some patients with low or undetectable VL, HIV-1 pol sequences were obtained from PBMC proviral DNA (6 of 47 HLA-B*1501-positive patients and 13 of 255 HLA-B*1501-negative patients). The estimated detection limit for minority species was 25%. HXB2 was used as reference virus to determine mutations and polymorphisms in the patients' samples.

Back to Top | Article Outline

HLA-I Typing

HLA class I typing was performed using standard serologic techniques (Biotest AG, Dreieich, Germany) or genotypic analyses (enzyme-linked probe hybridization assay Biotest ELPHA; Biotest AG). HLA-Cw loci were determined by DNA-based typing techniques (Dynal AllSet SSP; Dynal Biotech, Karlsruhe, Germany). HLA-B15 subtyping was done using the S4 HLA-B* single allele and locus-specific sequencing system from Protrans (Hockenheim, Germany) and the SBTEngine software (V 2.10.00; Genome Diagnostics, Utrecht, the Netherlands).

Back to Top | Article Outline

Cells and Culture Media

PBMCs were obtained by Ficoll-Hypaque (Biotest AG) density gradient centrifugation. The protocol for generation of HIV-1-specific T cell lines has been described in detail elsewhere20 and is briefly summarized here. PBMCs were stimulated with KF9 or variant peptides at a final concentration of 6 μg/mL in 1 mL of RPMI 1640 medium containing 10% (vol/vol) heat-inactivated fetal bovine serum, l-glutamine (4 mmol/L), penicillin (50 U/mL), streptomycin (50 μg/mL), Hepes (10 mmol/L), and 10 U/mL recombinant interleukin-2 (Proleukin; Chiron BV, Amsterdam, Netherlands). After 10 days, outgrowing cells were tested for recognition or cross-recognition of the KF9 peptides. Epstein-Barr virus-transformed B lymphoblastoid cell lines (B-LCL) were generated as described previously.26

Back to Top | Article Outline

Analysis of HIV-1-Specific T Cells by ELISPOT Assays

The protocol utilized for interferon-γ (IFN-γ) ELISPOT assays has been described in detail by Schmitt et al. 2000.20 Peptides were synthesized as C-terminal carboxamides (EMC Microcollections, Tübingen, Germany). For the assessment of the functional avidity of peptides by peptide titration assays, peptides were added in serial dilutions ranging from 20 μg/mL to 1 ng/mL to ELISPOT plates and incubated with CD8+ T cell lines for 16h. Half maximal lysis was calculated using the theorem on intersecting lines and calculations were conducted using Microsoft Office Excel 2003.

Back to Top | Article Outline

Cytotoxicity Assay

B-LCL were sensitized with synthetic peptides and tested in a 4-hour chromium-release assay as described previously.20

Back to Top | Article Outline

Determination of HLA Restriction

CD8+ T-cell lines specific for KF9 or variant peptides were seeded in 96-well plates in duplicates. A panel of allogeneic B-LCL sharing only 1 HLA allele with the CD8+ T cells was used as antigen-presenting cells. B cells were incubated with antigenic peptides (6 μg/mL) for 1 hour. Subsequently, the antigen-presenting cells were washed with phosphate-buffered saline and added to the CD8+ T cells. T cell recognition of peptide-loaded B-LCL was verified by IFN-γ ELISPOT assays.

Back to Top | Article Outline

Statistical Analysis

Correlations between HLA alleles and the occurrence of resistance mutations were evaluated by χ2 test and Fisher exact test. Statistical correlations between presence of HLA-B*1501 and VL at the time of resistance testing were evaluated by Mann-Whitney U test.

Back to Top | Article Outline


KF9 Is a Frequently Recognized Epitope in HLA-B*1501-Positive Patients

The interaction between CTL and drug resistance mutations in the HIV-1 PR epitope KF9 (KMIGGIGGF, amino acids 45-53)18 was evaluated in samples of 38 HIV-1-infected HLA-B*1501-positive patients. CTL responses to the KF9 wild-type peptide and mutant peptides comprising the major drug resistance mutations M46I, I47A, I47V, G48V, or I50V (Table 1) were analyzed by ELISPOT assays. At the time of ELISPOT analyses, the majority of patients included in this study were on an efficient ART, which is usually associated with a strong decline of the frequency of HIV-1-specific CTL in the peripheral blood.27-29 Therefore, we assessed the prevalence of KF9 recognition using ELISPOT analyses of PBMCs stimulated with KF9 and variant peptides. Using this method, we detected KF9-specific CD8+ T cells in the majority (31/38 = 80%) of the patients, indicating that KF9 is a very immunogenic and frequently recognized epitope in HLA-B*1501-positive patients. Chromium release assays using peptide-pulsed HLA-B*1501-matched B-LCL proved that KF9-stimulated T cells were indeed cytotoxic T cells (data not shown). In comparison with the HLA-B*1501-restricted KF9 epitope, the previously described HLA-A2 epitope KI10 (KMIGGIGGFI, amino acids 45-54),30 which covers the same region in the PR was recognized by only 40% of the analyzed patients (22/54). KF9 showed a much better peptide-sensitizing capacity in ELISPOT assays in HLA-B*1501-positive patients with a median half maximal peptide-sensitizing concentration of 575 ng/mL (range, 56 ng/mL to 3642 ng/mL; n = 26) than KI10 in HLA-A2-positive patients (median half maximal peptide-sensitizing concentration, 11,895 ng/mL; range 10,918 ng/mL to 12,872 ng/mL, n = 2). Values for half maximal lysis were calculated using the theorem on intersecting lines.



Back to Top | Article Outline

CTL Recognition of Drug Resistance Mutations in KF9

To assess the recognition of the major drug mutations occurring within the KF9 epitope (M46I, M46L, I47A, I47V, G48V, and I50V), PBMCs from HLA-B*1501-positive HIV-1-infected patients were separately stimulated with KF9 and the 5 KF9 variants (KF9-M2I, KF9-M2L, KF9-I3A, KF9-I3V, and KF9-I6V) (Table 1) comprising the drug resistance mutations M46I, M46L, I47A, I47V, G48V, and I50V. Outgrowing cell lines were analyzed by ELISPOT for cross-reactivity to the peptides.

The patients recognizing KF9 (n = 31) can be divided into 3 groups according to their recognition pattern of KF9 and the KF9 variants. Group 1 (n = 12) includes all patient samples showing strong recognition of all tested mutations apart from the SQV resistance mutation G48V that led to a complete loss of CTL recognition in all patients (example shown in Fig. 1A). However, the recognition of variants containing M46I or I47A was moderately attenuated in comparison with the recognition of the wild-type peptide. Representative graphs of peptide titration experiments analyzing the functional avidity of the wild-type KF9 peptide and the mutant peptides are shown in Figures 2A and B. Patients in this group differed regarding the recognition of the KF9-M2I peptide (comprising mutation M46I). Although 75% of the patients in this group showed good recognition of the KF9-M2I peptide (example shown in Fig. 2B), the recognition of peptides including the M46I mutation was attenuated in 25% of the patients (example shown in Fig. 2A). In the second group (n = 11), both the mutation G48V and the mutation I47A acted as escape mutations (example shown in Figs. 1B, C). In the third group (n = 8), recognition of peptides comprising the mutations G48V, I47A, and M46I was lost or strongly reduced (example shown in Figs. 1C, 2D).





No patient recognized a KF9 variant without recognition of the KF9 wild-type peptide. Roughly 90% (28/31) of the patients with KF9-specific immune responses also showed recognition of peptides comprising the M46I mutation (KF9-M2I, major drug resistance mutation for IDV, examples shown in Figs. 2A-C, patients 2, 4, and 5). In 12 of these 28 patients, recognition of KF9-M2I was attenuated in comparison with the recognition of the wild-type peptide (example shown in Fig. 2A patient 4). In 3 patients, the M46I mutation abrogated recognition completely (example shown in Fig. 2D, patient 6).

In contrast, peptides containing the M46L mutation (KF9-M2L) showed a good recognition in all patients tested so far (n = 13) (examples shown in Figs. 2A, C, D). In 61% of the patients recognizing KF9, recognition of the I47A (KF9-I3A) mutation was abrogated (examples shown in Figs. 2C, D: patients 2 and 6) or strongly attenuated in comparison with the wild-type KF9 peptide (examples shown in Figs. 2A, B, patients 4 and 5). In contrast, peptides containing the I47V mutation (KF9-I3V) were recognized in all tested samples (n = 26). In most of these patients (n = 22), recognition of peptides comprising the I47V mutation was comparable to the recognition of the wild-type peptide (examples shown in Figs. 2A, B). Only in 4 patients, recognition of the I47V mutation was reduced in comparison with the wild-type peptide (examples shown in Figs. 2C, D). The I50V mutation (KF9-I6V) was recognized in all but 2 samples (19/21), but in most tested samples, the functional avidity in peptide titration assays was reduced in comparison with the wild-type peptide (examples shown in Figs. 2A-D). Figure 2 shows the cross-recognition of variant peptides by KF9-stimulated CTL lines. Additional peptide titration assays were conducted using CTL lines generated by stimulation with peptides comprising the M46I, M46L, I47A, I47V, or I50V mutations. The obtained CTL lines showed similar recognition patterns in peptide titration assays as the KF9 stimulated CTL lines (data not shown).

The recognition pattern of the KF9 epitope was not influenced by the therapeutic regimen as each group included patients on PI and patients on PI-sparing therapies (data not shown). Recognition of KF9 and its variants by freshly isolated PBMC was analyzed by ELISPOT in 2 untreated PI-naive patients and in 13 patients on ART. Both untreated patients showed good recognition of KF9 (patient 2: 460 SFU per 1 × 106 cells; patient 5: 135 SFU per 1 × 106 cells) (Fig. 3). The analysis of the recognition of viral variants mirrored exactly the different recognition patterns, which were found in stimulated CTL lines (example given for patient 2 in Figs. 1B, 3B). Only in 2 of 13 patients on ART recognition of KF9 by freshly isolated PBMCs could be detected (70 SFU/1 × 106 cells and 150 SFU/1 × 106 cells, respectively).



Back to Top | Article Outline

Determination of HLA Restriction of Mutant KF9 Peptides

To rule out that mutations in the KF9 epitope did lead to the formation of a neoepitope binding to other HLA alleles, ELISPOT analyses were carried out using peptide-pulsed HLA-matched allogeneic B-LCL as target cells. These analyses verified that CTL lines stimulated with KF9-M2I, KF9-M2L, KF9-I3V, and KF9-I6V were also HLA-B*1501 restricted (data not shown). Restriction analysis for KF9-I3A could not be carried out due to the weak activity of cells stimulated with KF9-I3A.

Back to Top | Article Outline

Sequence Analysis of the PR

HIV-1 PR sequence analyses were carried out in 47 HLA-B*1501-positive patients and in 255 HLA-B*1501-negative patients. Our analysis showed that the frequency of patients with at least 1 of the potential CTL escape mutations (M46I, I47A, and G48V) in the KF9 epitope was very similar in HLA-B*1501-negative patients (19.6%) and HLA-B*1501-positive patients (19.1%); this was also true if each mutation was analyzed individually (see Table 2). As the HLA-A2 restricted epitope KI10 overlaps with the KF9 epitope, we performed an additional analysis, taking HLA-A2 into account. However, there was no evidence for selection of the CTL escape mutations by HLA-B*1501 and/or HLA-A2 (Table 2). Subanalysis of different patients groups, therapy-naive patients (n = 109; 9 patients HLA-B*1501 positive and 100 patients HLA-B*1501 negative), patients on therapies excluding PIs (n = 16), and patients on PI-containing therapies (n = 177; 34 patients HLA-B*1501 positive and 143 patients HLA-B*1501 negative) clearly showed that the potential CTL escape mutations are only associated with the usage of the respective PIs. Mutations in the KF9 epitope were only found in patients with current or previous PI-containing regimens. None of the therapy-naive patients showed amino acid substitutions in the KF9 epitope.

Table 2

Table 2

Considering these results, further analyses were restricted to the patients who had received PI treatment in the past. Protease sequence analysis in these patients revealed a slightly higher percentage of patients with at least 1 of the potential CTL escape mutations (M46I, I47A, and G48V) in HLA-B*1501-negative patients (35%; 50/143) than in the HLA-B*1501-positive patients (26.5%; 9/34). These differences were not statistically significant (Fisher exact test, P = 0.4). Numbers and percentages for the individual mutations are given in Table 2.

To investigate whether HLA-B*1501-restricted immune responses may influence viral replication in the context of KF9 resistance mutations, we analyzed VL at the time point of resistance testing in all 71 patients with mutations (M46I/L, I47A/V, G48V, I50V) in the KF9 epitope. All these 71 patients had a detectable VL ranging from 40 copies per milliliter to 1,120,000 copies per milliliter. Seventy of these patients were on a failing ART, and 1 patient (HLA-B*1501 positive; VL of 2562 copies/mL) was on a treatment interruption. VLs were significantly lower in the 11 HLA-*B1501-positive patients (median VL, 1015 copies/mL; range, 40-131,000 copies/mL) than in the 60 HLA-B*1501-negative patients (median VL, 9118 copies/mL; range, 378-1,120,000 copies/mL; Mann-Whitney U test, 2-tailed, P = 0.012). Lower VLs in HLA-B*1501-positive patients in comparison with the HLA-B*1501-negative patients could not be explained by the patterns of mutations within KF9 (data not shown).

Back to Top | Article Outline


Drug targets like the HIV-1 PR are exposed to both pharmacological selection pressure and immune selection exerted by CTL. In the past, several studies have provided evidence that selection pressure by CTL can interfere with the development of drug resistance mutations.12,14,18,19,20,31 To assess the interaction between pharmacological and immunological selection pressure, we performed a detailed study of CTL recognition of the HLA-B*1501-restricted KF9 epitope,18 which comprises several major PI resistance mutations (M46I/L, I47A/V, G48V, and I50V).5 The epitope was recognized by 80% of the HLA-B*1501-positive patients, and peptide stimulation induced a vigorous outgrowth of KF9-specific CTL lines in most of the patients, indicating that KF9 is a dominant CTL epitope in HLA-B*1501-positive patients. ELISPOT analyses using freshly isolated PBMCs demonstrated good recognition of KF9 and variant peptides in untreated patients, whereas patients on ART showed no or only weak recognition of the epitope. This is in accordance with earlier studies showing that the frequency of CTL decays after initiation of ART.27,28,29

Our analyses show that CTL of HLA-B*1501-positive patients can target important drug resistance mutations in the KF9 epitope, such as the M46I, M46L, I47A, I47V, and I50V, although there were strong differences between individual patients regarding the recognition of M46I and I47A. Interestingly, all patients failed to recognize the G48V mutation, which is a major mutation for SQV resistance. Despite the potential of the M46I, I47A, and G48V mutations for immune escape, we could not find evidence for CTL-mediated selection of these mutations in our cohort. We assume that the lack of immune selection of G48V in our cohort could be explained by the strong negative impact of this mutation on viral infectivity that has been described by Mammano et al.32

The second important mutation negatively affecting recognition by KF9-specific CTL was the LPV resistance associated I47A mutation. In contrast to the I47A mutation, I47V, also a major resistance mutation for LPV, was well recognized by KF9-specific CTL. Both mutations were only infrequently found in our cohort (I47V n = 9; I47A n = 4). In the HLA-B*1501-positive cohort, both mutations were only found in a single patient. This is arguing against a strong influence of HLA-B*1501 on the selection of the I47A mutation. The I47V mutation is induced by a single nucleotide A to G substitution, whereas the I47A mutation requires an additional U to C substitution. Thus, the I47V mutation is a necessary intermediate mutation in the virus for the development of the I47A mutation. As the I47V mutation is well recognized by KF9-specific CTL, the 2 nucleotide exchanges needed for the potential CTL escape mutation I47A could be a significant obstacle for a CTL-mediated immune selection of I47A.

The M46I mutation (major mutation for IDV; minor mutation for ATV, FPV, LPV, NFV), located at the P2 anchor position of the epitope, showed an unexpected recognition pattern. M46I abrogated (n = 3) or attenuated (n = 12) CTL recognition in about half of the HLA-B*1501-positive patients with recognition of the KF9 peptide (n = 31), but it was equally well recognized as KF9 in the other half of patients (n = 16). The decreased or abolished recognition of the M46I mutation in a subgroup of patients only does not seem to be the consequence of a lower peptide-binding affinity of the mutated peptide, given that in the majority of patients, peptide titration curves showed a similar recognition of the M46I-containing peptide and the wild-type KF9. Both a methionine and an isoleucine can serve as anchor for the HLA-B*1501 allele, and as all the tested patients carried the HLA-B*1501 allele, HLA-B15 polymorphism can be ruled out as a cause of weaker recognition of the M46I mutation. In addition, HLA restriction analyses of the mutated peptides demonstrated that the M46I-comprising peptide was actually presented by HLA-B*1501, thus excluding cross-presentation of the M46I mutation by other HLA alleles as the cause for the observed differences in the recognition of M46I. Therefore, we hypothesize that despite of its location in the P2-anchor groove of the HLA-B*1501 allele, the M46I mutation could induce changes in the conformation of the HLA-peptide complex, which may affect T-cell receptor (TCR) recognition depending on the patients' individual TCR. In contrast to the M46I mutation, M46L, which is a major mutation for IDV and a minor mutation for ATV, FPV, LPV, NFV, and TPV, was well recognized in all tested samples. Peptide titration curves showed a similar recognition of peptides including M46I or M46L. This could be a further hint that recognition of peptides with these mutations is dependent on the individual TCR.

The I50V is a major resistance mutation for FPV and darunavir and a minor resistance mutation for ATV, TPV, and NFV. Peptides comprising this mutation were well recognized by the majority of patients, although the I50V-containing peptide exhibited a lower peptide-sensitizing capacity than the wild-type peptide.

The HLA-B*1501-restricted epitope KF9 is located within the known HLA-A2-restricted CTL epitope KMIGGIGGFI.30 In contrast to the strong response against KF9 in HLA-B*1501-positive patients, KI10 was recognized only by 40% of HLA-A2-positive patients, and the KI10 peptide usually possessed only a weak stimulatory capacity for the induction of KI10-specific CTL in PBMCs from HLA-A2-positive donors in comparison with the good stimulatory capacity of the KF9-peptide in PBMCs from HLA-B*1501-positive patients. This demonstrates that there are strong functional differences between different HLA alleles regarding the capacity to mount CTL responses against epitopes that are located in the same sequence area. The observed stronger stimulatory capacity of the HLA-B*1501-restricted KF9 peptide in comparison with the HLA-A2-restricted KI10 peptide is in accordance with an earlier study from Kiepiela et al8 reporting on a substantially greater induction of CTL by HLA-B alleles than by HLA-A alleles.

Despite strong CTL activity against the KF9 epitope, sequencing of HIV-1 PR genes failed to provide clear evidence for CTL-mediated immune selection within the KF9 epitope in HLA-B*1501-positive patients in our cohort. None of the mutations, which acted as CTL escape mutations in the ELISPOT analysis (M46I, I47A, G48V), occurred at higher frequency in HLA-B*1501-positive patients than in HLA-B*1501-negative patients. Although the M46I, I47A, and G48V mutations can strongly impair CTL recognition, the KF9 is a highly conserved CTL epitope and especially mutations at position 47 and 48 occur only infrequently. We assume that the rarity of mutations is due to strong functional constraints for mutations within the KF9 epitope as it is located in a region of the HIV-1 PR with important functional activity. The amino acids comprising KF9 are all located within the flap region of the PR. Several studies have shown that the flexible flaps participate in the binding of the substrate and therefore are strongly conserved. Amino acids 45, 46, and 47 are located in the flap tips of the PR, and amino acids 48 and 50 are substrate cleft residues.33-38 There is a high genetic barrier for most of the major drug resistance mutations within the KF9 epitope. In the past, several studies have indicated that major drug resistance mutations in functionally important regions of the PR require compensatory mutations in other regions of the PR or of Gag, which compensate for the functional impairments induced by the major resistance mutations.39-42

All patients in our cohort with mutations in the KF9 epitope were on antiretroviral regimens selecting for the respective mutations, indicating that PIs exert greater selection pressure than HLA-B*1501-restricted CTL responses. Nevertheless, the frequency of resistance mutations within KF9 was low even in viremic patients on PI treatment.

Together with several other rare HLA-B15 subtypes, HLA-B*1501 represents the serologically defined HLA-B62 group. In an earlier study by Hentges et al,43 HLA-B62 was associated with prolonged clinically asymptomatic HIV-1 infection. So far, it is unknown, whether and to what extent CTL recognition of the KF9 epitope has a beneficial effect on the course of HIV-1 infection in HLA-B*1501-positive patients. Although there are a number of HLA-B*1501-restricted CTL epitopes in various parts of the HIV-1 genome, we conclude from our study that the KF9 epitope is an important epitope for HLA-B*1501-positive patients that should be considered in the design of preventive and therapeutic HIV-1 vaccines. From our sequence analysis, we could find no evidence that HLA-B*1501 contributes to the development of drug resistance mutations in the KF9 epitope. Rather, the frequency of mutations in KF9 was slightly higher in HLA-B*1501-negative patients (35%) than in HLA-B*1501-positive patients (26.5%) on PI therapy (not statistically significant). Therefore, we hypothesize that the cross-recognition of drug resistance mutations by KF9-specific CTL could inhibit the emergence of drug resistance mutations in HLA-B*1501-positive patients on PI therapy. This would be in line with our observation that HLA-B*1501 was associated with lower VLs in patients failing PI therapy with mutations in the KF9 epitope.

Prospective studies in larger cohorts of patients on PI therapy are needed to assess the role of HLA-B*1501-restricted CTL responses for the success or failure of PI therapy.

Back to Top | Article Outline


The authors thank all patients participating in this study. They also thank the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, for providing the HIV-1 Consensus B Pol (15-mer) peptides, complete set.

Back to Top | Article Outline


1. Borman AM, Paulous S, Clavel F. Resistance of human immunodeficiency virus type 1 to protease inhibitors: selection of resistance mutations in the presence and absence of the drug. J Gen Virol. 1996;77(pt 3):419-426.
2. Condra JH, Holder DJ, Schleif WA, et al. Genetic correlates of in vivo viral resistance to indinavir, a human immunodeficiency virus type 1 protease inhibitor. J Virol. 1996;70:8270-8276.
3. Condra JH, Schleif WA, Blahy OM, et al. In vivo emergence of HIV-1 variants resistant to multiple protease inhibitors. Nature. 1995;374:569-571.
4. Croteau G, Doyon L, Thibeault D, et al. Impaired fitness of human immunodeficiency virus type 1 variants with high-level resistance to protease inhibitors. J Virol. 1997;71:1089-1096.
5. Johnson VA, Brun-Vezinet F, Clotet B, et al. Update of the drug resistance mutations in HIV-1: December 2009. Top HIV Med. 2009;17:138-145.
6. Martinez-Picado J, Savara AV, Sutton L, et al. Replicative fitness of protease inhibitor-resistant mutants of human immunodeficiency virus type 1. J Virol. 1999;73:3744-3752.
7. Molla A, Korneyeva M, Gao Q, et al. Ordered accumulation of mutations in HIV protease confers resistance to ritonavir. Nat Med. 1996;2:760-766.
8. Kiepiela P, Leslie AJ, Honeyborne I, et al. Dominant influence of HLA-B in mediating the potential co-evolution of HIV and HLA. Nature. 2004;432:769-775.
9. Leslie A, Kavanagh D, Honeyborne I, et al. Transmission and accumulation of CTL escape variants drive negative associations between HIV polymorphisms and HLA. J Exp Med. 2005;201:891-902.
10. Leslie AJ, Pfafferott KJ, Chetty P, et al. HIV evolution: CTL escape mutation and reversion after transmission. Nat Med. 2004;10:282-289.
11. Moore CB, John M, James IR, et al. Evidence of HIV-1 adaptation to HLA-restricted immune responses at a population level. Science. 2002;296:1439-1443.
12. Gatanaga H, Ode H, Hachiya A, et al. Impact of human leukocyte antigen-B*51-restricted cytotoxic T-lymphocyte pressure on mutation patterns of nonnucleoside reverse transcriptase inhibitor resistance. AIDS. 2010;24:F15-F22.
13. John M, Moore CB, James IR, et al. Interactive selective pressures of HLA-restricted immune responses and antiretroviral drugs on HIV-1. Antivir Ther. 2005;10:551-555.
14. Karlsson AC, Deeks SG, Barbour JD, et al. Dual pressure from antiretroviral therapy and cell-mediated immune response on the human immunodeficiency virus type 1 protease gene. J Virol. 2003;77:6743-6752.
15. Mahnke L, Clifford D. Cytotoxic T cell recognition of an HIV-1 reverse transcriptase variant peptide incorporating the K103N drug resistance mutation. AIDS Res Ther. 2006;3:21.
16. Manosuthi W, Butler DM, Pérez-Santiago J, et al. Protease polymorphisms in HIV-1 subtype CRF01_AE represent selection by antiretroviral therapy and host immune pressure. AIDS. 2010;24:411-416.
17. Mason RD, Bowmer MI, Howley CM, et al. Antiretroviral drug resistance mutations sustain or enhance CTL recognition of common HIV-1 Pol epitopes. J Immunol. 2004;172:7212-7219.
18. Mueller SM, Schaetz B, Eismann K, et al. Dual selection pressure by drugs and HLA class I-restricted immune responses on HIV-1 protease. J Virol. 2007;81:2887-2898.
19. Samri A, Haas G, Duntze J, et al. Immunogenicity of mutations induced by nucleoside reverse transcriptase inhibitors for human immunodeficiency virus type 1-specific cytotoxic T cells. J Virol. 2000;74:9306-9312.
20. Schmitt M, Harrer E, Goldwich A, et al. Specific recognition of lamivudine-resistant HIV-1 by cytotoxic T lymphocytes. AIDS. 2000;14:653-658.
21. Kuiken C, Leitner, T., Foley, B., et al, eds. HIV Sequence Compendium 2009. Los Alamos: Theoretical Biology and Biophysics Group, Los Alamos National Laboratory, NM; 2009. LA-UR 06-0680.
22. Doxiadis I, Haasnoot G, Korfagen N. Virtual PRA calculation. Available at: Accessed May 31, 2010.
23. Middleton D, Menchaca L, Rood H, et al. New allele frequency database. Tissue Antigens. 2003;61:403-407.
24. Jansen K, Brockmeyer NH, Hahn M, et al. Epidemiological composition, clinical and treatment characteristics of the patient cohort of the German Competence Network for HIV/AIDS. Eur J Med Res. 2009;14:415-425.
25. Jansen K, Michalik C, Hahn M, et al. The Patient Cohort of the German Competence Network for HIV/AIDS (KompNet): a profile. Eur J Med Res. 2009;14:323-331.
26. Walker BD, Flexner C, Birch-Limberger K, et al. Long-term culture and fine specificity of human cytotoxic T-lymphocyte clones reactive with human immunodeficiency virus type 1. Proc Natl Acad Sci U S A. 1989;86:9514-9518.
27. Casazza JP, Betts MR, Picker LJ, et al. Decay kinetics of human immunodeficiency virus-specific CD8+ T cells in peripheral blood after initiation of highly active antiretroviral therapy. J Virol. 2001;75:6508-6516.
28. Kalams SA, Goulder PJ, Shea AK, et al. Levels of human immunodeficiency virus type 1-specific cytotoxic T-lymphocyte effector and memory responses decline after suppression of viremia with highly active antiretroviral therapy. J Virol. 1999;73:6721-6728.
29. Ogg GS, Jin X, Bonhoeffer S, et al. Decay kinetics of human immunodeficiency virus-specific effector cytotoxic T lymphocytes after combination antiretroviral therapy. J Virol. 1999;73:797-800.
30. Propato A, Schiaffella E, Vicenzi E, et al. Spreading of HIV-specific CD8+ T-cell repertoire in long-term nonprogressors and its role in the control of viral load and disease activity. Hum Immunol. 2001;62:561-576.
31. Casazza JP, Betts MR, Hill BJ, et al. Immunologic pressure within class I-restricted cognate human immunodeficiency virus epitopes during highly active antiretroviral therapy. J Virol. 2005;79:3653-3663.
32. Mammano F, Trouplin V, Zennou V, et al. Retracing the evolutionary pathways of human immunodeficiency virus type 1 resistance to protease inhibitors: virus fitness in the absence and in the presence of drug. J Virol. 2000;74:8524-8531.
33. Gustchina A, Weber IT. Comparison of inhibitor binding in HIV-1 protease and in non-viral aspartic proteases: the role of the flap. FEBS Lett. 1990;269:269-272.
34. Ishima R, Freedberg DI, Wang YX, et al. Flap opening and dimer-interface flexibility in the free and inhibitor-bound HIV protease, and their implications for function. Structure. 1999;7:1047-1055.
35. Miller M, Schneider J, Sathyanarayana BK, et al. Structure of complex of synthetic HIV-1 protease with a substrate-based inhibitor at 2.3 A resolution. Science. 1989;246:1149-1152.
36. Scott WR, Schiffer CA. Curling of flap tips in HIV-1 protease as a mechanism for substrate entry and tolerance of drug resistance. Structure. 2000;8:1259-1265.
37. Shao W, Everitt L, Manchester M, et al. Sequence requirements of the HIV-1 protease flap region determined by saturation mutagenesis and kinetic analysis of flap mutants. Proc Natl Acad Sci U S A. 1997;94:2243-2248.
38. Wu TD, Schiffer CA, Gonzales MJ, et al. Mutation patterns and structural correlates in human immunodeficiency virus type 1 protease following different protease inhibitor treatments. J Virol. 2003;77:4836-4847.
39. Bally F, Martinez R, Peters S, et al. Polymorphism of HIV type 1 gag p7/p1 and p1/p6 cleavage sites: clinical significance and implications for resistance to protease inhibitors. AIDS Res Hum Retroviruses. 2000;16:1209-1213.
40. Berkhout B. HIV-1 evolution under pressure of protease inhibitors: climbing the stairs of viral fitness. J Biomed Sci. 1999;6:298-305.
41. Cote HC, Brumme ZL, Harrigan PR. Human immunodeficiency virus type 1 protease cleavage site mutations associated with protease inhibitor cross-resistance selected by indinavir, ritonavir, and/or saquinavir. J Virol. 2001;75:589-594.
42. Verheyen J, Litau E, Sing T, et al. Compensatory mutations at the HIV cleavage sites p7/p1 and p1/p6-gag in therapy-naive and therapy-experienced patients. Antivir Ther. 2006;11:879-887.
43. Hentges F, Hoffmann A, Oliveira de Araujo F, et al. Prolonged clinically asymptomatic evolution after HIV-1 infection is marked by the absence of complement C4 null alleles at the MHC. Clin Exp Immunol. 1992;88:237-242.

CTL; drug resistance; protease inhibitor

© 2011 Lippincott Williams & Wilkins, Inc.