A sensitive phenotypic assay for the determination ... - Oxford Academic

J Antimicrob Chemother 2010; 65: 2493 – 2501 doi:10.1093/jac/dkq379 Advance Access publication 14 October 2010

A sensitive phenotypic assay for the determination of human immunodeficiency virus type 1 tropism Nuria Gonza´lez 1†, Mayte Pe´rez-Olmeda 1†, Elena Mateos 1, Almudena Cascajero 1, Amparo Alvarez 1, Sanne Spijkers 1, Javier Garcıá-Pe´rez 1, Sonsoles Sańchez-Palomino 2, Ezequiel Ruiz-Mateos 3, Manuel Leal 3 and Jose´ Alcami 1* 1

AIDS Immunopathology Unit, Centro Nacional de Microbiologıá, Instituto de Salud Carlos III, Majadahonda, Madrid, Spain; 2Hospital Clinic IDIBAPS, Barcelona, Spain; 3Servicio de Enfermedades Infecciosas, Hospitales Universitarios Virgen del Rocıó, Sevilla, Spain *Corresponding author. Tel: +34-91-822-3943; Fax: +34-91-509-7919; E-mail: [email protected] †The first two authors contributed equally to the work.

Received 7 April 2010; returned 2 August 2010; revised 25 August 2010; accepted 13 September 2010 Objectives: To develop a sensitive phenotypic assay based on recombinant viruses (RVs) for characterizing HIV-1 tropism. Methods: Viral tropism was assessed in 159 plasma samples. The env gene was amplified and ligated into pNL-lacZ/env-Ren, which carries a luciferase reporter gene. Resulting constructs were transfected into HEK293T cells to generate RVs. To assess co-receptor tropism, U87.CD4.CXCR4/CCR5 cells were infected and luciferase activity was measured. Results: RVs containing env from different HIV-1 subtypes were replication competent. Minor variants were detectable in 1% of the viral population. Tropism was determined in 65% of samples with a viral load of ,1000 copies/mL. The phenotypic assay described here was validated with the TrofileTM and TrofileTM ES assays. Considering the TrofileTM assay as a reference, the sensitivity for R5 and R5X4/X4 detection was 90% and 77%, and the specificity was 77% and 90%, respectively. Our assay was 86% concordant with TrofileTM (90% for R5 and 77% for R5X4/X4). When our system was compared with TrofileTM ES, the concordance was 89% (86% for R5 and 92% for R5X4/X4), the sensitivity for R5 was 86% and for R5X4/X4 was 92%, and the specificity for R5 was 92% and for R5X4/X4 was 86%. The phenotypic results were compared with those obtained using the following V3 genotypic prediction tools: position-specific scoring matrix; geno2pheno[coreceptor]; C4.5; C4.5 using positions 8 and 12; PART; support vector machines; and the charge rule. Conclusions: We describe a system to assess co-receptor tropism based on the generation of chimeric replication-competent viruses with high sensitivity in the detection of minor populations. A good correlation of our results with TrofileTM assays was found. Keywords: HIV co-receptors, phenotype, CCR5, tropism

Introduction Human immunodeficiency virus type 1 (HIV-1) entry into host cells is a process that includes interaction with different receptors, and fusion of the viral envelope and host cell membranes.1 – 4 The HIV-1 envelope glycoprotein gp120 binds to the CD4 molecule and two major chemokine receptors (CCR5 and/ or CXCR4).5,6 Accordingly, HIV-1 strains are classified as R5- or X4-tropic strains, depending on the use of CCR5 and CXCR4 receptors, respectively. In addition, R5X4- or dual-tropic variants correspond to HIV-1 strains that can enter into the cell through CCR5 and CXCR4.7 Actually, phenotypic assays cannot distinguish between dual tropic variants and mixtures of R5- and X4-using viruses, which are overall referred to as dual/mixed (DM).

Since the development of CCR5 antagonists,8,9 assessment of HIV tropism is required before initiating treatment with these drugs in HIV-infected patients. Both phenotypic10 – 18 and genotypic tropism assays19 – 25 have been described and proposed for use in the clinical setting. Although genotypic assays are fast, easy and cheap, their low sensitivity in the detection of minority quasispecies and the difficulty of establishing a genotypic– phenotypic correlation limit their use.26 Therefore, phenotypic methodologies are considered the gold standard for determining HIV-1 co-receptor usage. Two different commercially available phenotypic assays allow the determination of viral tropism, TrofileTM (Monogram Biosciences, Inc., USA)18 and the Tropism Recombinant Test (VIRalliance, France).12,16,27 One of the challenges in the development of new methods for tropism

# The Author 2010. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]

2493

Gonza´lez et al.

determination is to improve sensitivity in the detection of minority X4/DM. Minority quasispecies (,5%–10%) in the whole viral population are not detected with early commercial tests26 and are responsible for the failure of treatment with CCR5 antagonists.28 A more sensitive assay (TrofileTM ES) has been recently developed and can increase sensitivity in the detection of minority variants in the range of 0.3% –1%.29 In this study, we describe a novel HIV-1 tropism assay based on replication-competent recombinant viruses that can produce multiple cycles of infection. This strategy increases the sensitivity in the detection of minority viral populations in samples from patients. The results demonstrate a good correlation with TrofileTM and TrofileTM ES.

Materials and methods Study subjects and samples HIV tropism was assessed in plasma samples from HIV-1-infected patients selected in the expanded access to maraviroc protocol (n¼159) in Spain. Drug-experienced patients included in the expanded access protocol had treatment failure with different highly active antiretroviral therapy (HAART) regimens and met criteria to be treated with maraviroc as part of salvage therapy. Plasma was recovered using EDTA as anticoagulant and was preserved at 2808C for further analysis. Baseline plasma samples from these patients were previously phenotyped for HIV co-receptor usage with TrofileTM (n¼124) and TrofileTM ES (n ¼35) assays.

Cell lines Human embryonic kidney 293 (HEK293T) cells constitutively express the simian virus 40 (SV40) large T antigen.30 These cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM; BioWhittaker) supplemented with 10% fetal calf serum (FCS), penicillin (50 U/mL), streptomycin (50 U/mL) and L-glutamine (2 mM). GHOST cells [derived from human osteosarcoma cell lines (HOS)] stably express CD4 receptor and the chemokine receptor CXCR4 (GHOST.CD4.CXCR4) or CCR5 (GHOST.CD4.CCR5). The GHOST cells were grown in DMEM (BioWhittaker) supplemented with 10% FCS, penicillin (50 U/mL), streptomycin (50 U/mL) and L-glutamine (2 mM), and selected with G418 (500 mg/L), hygromycin (100 mg/L) and puromycin (1 mg/L). Astroglioma U87 cells that stably express the CD4 receptor (U87.CD4), or CD4 and either chemokine receptors CXCR4 (U87.CD4.CXCR4) or CCR5 (U87.CD4.CCR5) were maintained in medium containing G418 (300 mg/L). Moreover, the medium of the U87.CD4.CCR5 cells and the U87.CD4.CXCR4 cells was supplemented with puromycin (1 mg/L). The GHOST31 and U8732 cells lines were obtained from Dr Dan R. Littman through the US National Institutes of Health (NIH) AIDS Research and Reference Reagent Program. Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats of healthy blood donors by centrifugation through a Ficoll-Hypaque gradient (Pharmacia Corporation, North Peapack, NJ, USA). PBMCs were activated for 48 h with phytohaemagglutinin (PHA; 5 mg/L) plus interleukin-2 (IL-2; 300 IU/mL; Chiron, Emeryville, CA, USA) and maintained with IL-2.

Construction of proviral clones and plasmids The vector pNL4-3Ren33 was generated by replacing the gene nef of the HIV-1 proviral clone pNL4-334 (NIH AIDS Research and Reference Reagent Program, catalogue number 114) with the Renilla luciferase gene. To clone the Renilla luciferase gene, a NotI restriction site was introduced

2494

near the 5′ end of the nef coding region by PCR site-directed mutagenesis (QuickChange Site-Directed Mutagenesis Kit, Stratagene). The vector pNL-lacZ/env-Ren was generated by introducing an XbaI restriction site using PCR site-directed mutagenesis (QuickChange SiteDirected Mutagenesis Kit, Stratagene) in the 6114 position of pNL4-3Ren, followed by env deletion digesting with XbaI (6114) and NotI (8797), and env coding sequence replacement by the aminoterminal fragment of the lacZ gene. The lacZ gene was amplified from the pUC19 plasmid using the primers Xba-lacZ-U (5′ -AGCTCTAGAGCG CCCAATACGCAAACCGCCTC-3′ ) and Not-lacZ-D (5′ -TTTTGCGGCCGCG CCATTCGCCATTCAGGCTGCGC-3′ ), containing XbaI and NotI sites, respectively. The lacZ sequence had another recognition site for XbaI that was deleted using site-directed mutagenesis with the primers M-Xba-lacZ-U (5′ -GCCTGCAGGTCGACTCTGGAGGATCCCCGGGTACCG-3′ ) and M-Xba-lacZ-U (5′ -CGGTACCCGGGGATCCTCCAGAGTCGACCTGCAGGC-3′ ). This system allows replacement of the lacZ gene with the full-length HIV-1 env genes of different HIV strains or of patient samples, and avoids interference due to wild-type sequences. The JRCSF env gene was amplified by PCR from the pYK-JRCSF plasmid (NIH AIDS Research and Reference Reagent Program, catalogue number 2708), digested with XbaI and NotI, and inserted in the same sites in the pNL-lacZ/env-Ren resulting in vector pJRRen. The p89.6Ren plasmid, with an R5X4 env, was generated by inserting the Renilla luciferase gene into p89.6 vector35 between the sites NotI (8796) and XhoI (8895). The NotI restriction site was introduced in p89.6 by site-directed mutagenesis with the primers M-Not-89.6-U and (5′ -GCTATAAGATGGGAGGCGCGGCCGCAAAACGTAGGGCAGAGGG-3′ ) M-Not-89.6-D (5′ -CCCTCTGCCCTACGTTTTGCGGCCGCGCCTCCCATCTTATAG C-3′ ), and the XhoI restriction site at position 8382 was removed using the primers M-Xho-89.6-U (5′ -CATTATCGTTTCAGACCCTCCTCCCAGCCTACAAGGGACCCGACAGGCCCGAAG-3′ ) and M-Xho-89.6-D (5′ CTTCGGGCCTG TCGGGTCCCTTGTAGGCTGGGAGGAGGGTCTGAAACGATAATG 3′ ). pcDNA3-VSVG (a gift from Dr F. Arenzana, Pasteur Institute, Paris) encodes the vesicular stomatitis virus (VSV) G protein.

Viral RNA isolation and amplification of env gene from patient plasma samples Viral RNA was extracted from 500 mL of plasma using the ViroSeqTM HIV-1 Genotyping System (Applied Biosystems, Foster City, CA, USA), according to the manufacturer’s protocol. RT–PCR was performed with the SuperScriptTM One-Step RT–PCR with Platinumw Taq System (Invitrogen, Carlsbad, CA, USA) using the primers 30eu (5′ -TATGAAACTTACGGGGATACTTGGG-3′ ) and 11ed (5′ -CTGCCAATCAGGGAAGTAGCCTTGTGT-3′ ). The PCR product was used as a template for nested PCR amplification with the Expand High Fidelity PCR System (Roche, Mannheim, Germany), and the primers 53eu (5′ -GCTC TAGAGCTGTGGTCCATAGTAATCATAGAATATAGG-3′ ) and 52ed (5′ -TACTTTTTGC GGCCGCGCCACCCATCTTATAGC-3′ ) containing XbaI (TCTAGA) and NotI (GC GGCCGC) sites, respectively. The resulting 2.7 kb amplification products include the whole open reading frame of the HIV-1 gp160.

Cloning of env coding sequences PCR products were excised from agarose gels, purified and concentrated with the GFXTM PCR DNA and Gel Band Purification Kit (GE Healthcare BioSciences AB, Uppsala, Sweden). After digestion with XbaI and NotI, PCR amplification products were ligated into the pNL-lacZ/env-Ren vector. The resulting constructs were transformed into competent Escherichia coli cells (Gibco/BRL, Gaithersburg, USA) and only efficient reactions (as defined by the obtention of .200 clones and ,5% religated in a 1/10 volume of the bulk of transformed bacteria) were selected. Recombinant plasmids were purified with the Qiagen Plasmid Maxi Kit (Qiagen, Valencia, CA, USA).

JAC

Phenotypic assay to test HIV-1 tropism

To provide additional support to the diversity of the viral population amplified, a diluted sample (1/10) of extracted RNA was run in parallel with the RNA pool through reverse transcription and PCR procedures. Amplicons obtained from a given sample were considered representative of the viral population when the 1/10 dilution of the same sample could be amplified.

bioinformatic tools: position-specific scoring matrix (PSSM) (http://indra. mullins.microbiol.washington.edu/webpssm/);21 geno2pheno[coreceptor] (http://coreceptor.bioinf.mpi-inf.mpg.de/index.php); C4.5; C4.5 using positions 8 and 12; PART rule generator; support vector machines (SVM); and the charge rule (http://genomiac2.ucsd.edu:8080/wetcat/v3.html).23 R5X4 and X4 HIV-1 variants are classified as X4 by the bioinformatics tools.

Transfection and generation of virus stocks

Statistical analysis

HEK293T cells were transfected by calcium phosphate with 10 mg of the recombinant plasmids. The culture medium was replaced by fresh DMEM 8 h and 24 h after transfection. Cell supernatants were harvested 48 h after transfection.

Results were given as percentages. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated between the results obtained with tropism assays.

Results

Determination of co-receptor tropism GHOST or U87 cells were seeded at 2×104 cells per well in a 96-well culture plate and were infected in triplicate with 50 mL of virus stocks in a final volume of 200 mL. In order to assess co-receptor tropism, cells were lysed 48 h after infection and analysed for luciferase activity in a luminometer (Orion, Berthold Detection Systems, Pforzheim, Germany) using the Renilla Luciferase Assay System (Promega, Madison, USA). The results are shown as relative light units (RLU). Viruses were classified as X4 (GHOST.CD4.CXCR4 or U87.CD4.CXCR4), R5 (GHOST.CD4.CCR5 or U87.CD4.CCR5) or R5X4 (both of them), based on the production of luciferase in these cells.

Replication assays In replication assays, 1×106 activated PBMCs were infected with 5 ng of Gag p24 (quantified with the ELISA kit INNOTESTw HIV Antigen mAb, Innogenetics) of JRRen and 11525Ren. After 2 h of viral adsorption, cells were washed four times with PBS and cultured at 5×105 cells per well in a 24-well plate. Every 24 h, luciferase activity was measured in cell lysates. In addition, PBMCs were infected with supernatants from previous infections and cultured for 5 days. At day 5, luciferase activity was measured.

Replication assays A schematic representation of the plasmid pNL-lacZ/env-Ren is shown in Figure 1. Initially, tropism was assessed by the mean of recombinant viral clones derived from reference strains. NL4-3Ren (X4-tropic virus), JRRen (R5-tropic virus) and 89.6Ren (R5X4 tropism) were used to inoculate the U87 cell line (CD4, CD4.CCR5 and CD4.CXCR4) and PBMCs. Specific luciferase activity was observed in cell lines according to the expression of CCR5 or CXCR4 receptors (Table 1). Recombinant viral clones were also capable of infecting human PBMCs (Table 1). Replication kinetics of JRRen and the recombinant virus 11525Ren harbouring an R5 patient’s envelope was analysed. Both viruses were able to replicate in PBMCs, as detected by quantifying luciferase activity in cell lysates (Figure 2a). Viral replication of JRRen and 11525Ren was maintained after two consecutive passages in PBMCs (Figure 2b). In contrast, multiple cycle infection did not occur with a single cycle virus (NL4-3VSVRen) lacking the HIV envelope gene and pseudotyped with the envelope from VSV. These data demonstrate that this assay generates recombinant viruses that are replication competent and undergo several replication cycles.

Genotypic tropism testing A 1 kb fragment including V1/V2 and C2V3 regions of 82 plasmids was labelled using BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems). The labelling was carried out with an initial denaturation step of 968C for 1 min, 25 cycles of denaturation at 968C for 10 s, annealing at 508C for 10 s and elongation at 608C for 4 min. All samples were sequenced on an ABI 3730XL DNA Analyzer (Applied Biosystems). The V3 nucleic acid sequences were translated to amino acids with EditSeq 7.0 software (Lasergene, DNASTar). HIV-1 co-receptor usage was predicted from the V3 amino acid sequences according to the following

Assay performance To evaluate the efficacy of our system in the amplification of non-B subtype envelopes, recombinant viruses with reference subtype envelopes (A–E) were generated. The tropism of these viruses has already been described.36 The results obtained correlated with those previously described (Table 2). The reproducibility of the assay was measured using four different viral strains from patients (T601, 15214, 80013 and

env Renilla gag LTR

vif pol

nef

tat vpr

vpu

lacZ

LTR rev

Xba I

Xho I Not I

Figure 1. Schematic representation of the plasmid pNL-lacZ/env-Ren. This vector was generated by cloning the Renilla luciferase gene in the place of nef and replacing the env coding sequence by the amino-terminal fragment of the lacZ gene.

2495

Gonza´lez et al.

Table 1. Infection of different cell lines with the reference strains U87 Cell line NL4-3Ren JRRen 89.6Ren

PBMCs

CD4

CD4.CCR5

CD4.CXCR4

373000+64000 87300+2000 703000+56000

7960+440 7900+1 400 2507+97

7 190+620 1 420000+150000 34200+2 400

5 560000+670000 1 050+300 540000+140000

RLU values+SEM of triplicate assays are shown.

(a)

JRRen

11525Ren

RLU

100 000

10 000

1000 0

2

4

6

8

10

12

Days post-infection (b)

JRRen

11525Ren

NL4-3VSVRen

100 000

RLU

10 000

1000

100 0

1st Passage

2nd

Figure 2. (a) Replication kinetics of JRRen and 11525Ren. Viral replication was assessed for 10 days by quantifying luciferase activity in cell lysates. Mean+SEM of triplicate assays. (b) Measurement of Renilla luciferase activity after several viral cycles. PBMCs were infected with JRRen, 11525Ren or NL4-3VSVRen viruses and cultured for 5 days. Two consecutive passages with supernatant of previous infection were carried out. Luciferase activity of viruses was measured in cell lysates.

11476) (Figure 3a), which were generated in duplicate starting from different aliquots of plasma samples. All four samples infected GHOST.CD4.CCR5 and one of them also infected GHOST.CD4.CXCR4.

2496

Moreover, a total of 56 recombinant viruses were tested in two replicates on both GHOST CD4.CCR5 and CD4.CXCR4. Both the luciferase activity values and the tropism determination were concordant for the two replicates (Figure 3b). The sensitivity of amplification was evaluated using 227 plasma samples from patients presenting different viral loads. Samples were distributed according to ranges of viral load. The success rate for amplification was 100%, 91%, 87%, 89%, 76% and 56% for viral loads of .106, 105 –106, 104 –105, 103 –104, 102 –103 and 0 –102, respectively (Table 3). The sensitivity of detection of minority variants of this tropism assay was assessed using two different combinations of R5 and X4 HIV-1 clones (NP1525-X4:SF162-R5 and MN-X4:CM244-R5), in which viral RNAs were mixed at different ratios. The results obtained in these experiments showed that minority variants of X4 and R5 viruses were detectable in all mixed viral stock when they represented 1% of the whole population (Figure 4a and b).

Validation of phenotypic assay Tropism obtained using our recombinant virus system was compared with results obtained with the TrofileTM (Table 4) and TrofileTM ES (Table 5) assays. In the validation with TrofileTM , a total of 124 samples were analysed (Table 4). Reportable results were obtained for 100/ 124 samples (81%) and 103/124 samples (83%) with TrofileTM and our assay, respectively (Table 4). Of a total of 24 samples non-reportable (NR) with TrofileTM assay, 13 were NR in both systems and in 11 samples (7 R5, 3 DM and 1 X4) our system could define viral tropism (Table 4). Overall, in the patients in which viral tropism was reportable with both tests, concordance was 86% (79/92) (90% for R5, 71% for R5X4 and 100% for X4 viruses). From 103 patients in which tropism was reportable using the recombinant virus assay described here, 69 (67%) presented an R5 phenotype, 31 (30%) carried R5X4 viruses and, lastly, 3 (3%) patients displayed a pure X4-tropic population. Considering TrofileTM as the comparative standard for R5 detection, the sensitivity of our method was 90%, the specificity was 77%, the PPV was 89% and the NPV was 80%. Regarding R5X4/X4 viruses, the sensitivity of our method when compared with TrofileTM was 77%, the specificity was 90%, the PPV was 80% and the NPV was 89%. A modified version of our system was compared with TrofileTM ES in 35 samples. When our results were compared with Trofile second-generation tests (TrofileTM ES), the results of concordance (92% versus 76%), sensitivity (92% versus 77%) and NPV (95% versus 89%) for R5X4 variants improved (Table 5).

JAC


Table 2. Tropism characterization of recombinant viruses expressing the envelope from different HIV-1 subtypes Sample

env subtype

U87.CD4

U87.CD4.CCR5

U87.CD4.CXCR4

Tropism

A A B B C C D E E

750+170 113+22 53+21 62+11 42+3 49+2 49+10 79+2 33+4

27 800+8 000 213000+1 1000 450000+29000 100+32 14 630+730 74 600+9 700 75+13 206000+11000 77+14

23100+4 300 174+30 164+20 700000+130000 68+7 297+18 539000+58 000 111+8 540000+58 000

R5X4 R5 R5 X4 R5 R5 X4 R5 X4

92RW009 VI 191 SF162 MN DU174 92BR025 92UG024 CM244 NP1525

RLU values+SEM of triplicate assays are shown. (a)

Table 3. Evaluation of the amplification sensitivity GHOST CXCR4

GHOST CCR5

replicate 1 replicate 2 replicate 1 replicate 2 T601

3.25

3.15

5.59

5.17

15214

2.99

2.77

5.42

5.95

80013

2.99

3.30

5.07

5.59

11476

4.99

5.21

5.53

5.67

0 –102 102 –103 103 –104 104 –105 105 –106 .106

The mean luciferase background in GHOST.CD4 cells was 3.19. (b)

CCR5+ cells

Total

CXCR4+ cells

Replicate 2 (log RLU)

7.00 6.00 5.00 4.00 3.00 2.00 1.00 1.00

2.00

3.00

4.00

5.00

6.00

Number of samples

Amplified

Percentage success

27 21 65 94 11 9

15 16 58 82 10 9

56 76 89 87 91 100

227

190

84

mean concordance for the genotypic tools taking as a reference the phenotypic results was 79%. The best concordance was found with C4.5 and C4.5 using positions 8 and 12 only (85% in both cases), followed by PSSMX4R5 (84%) and the charge rule (83%). As shown in Table 6, for R5X4/X4 viruses the concordance between genotypic methods and the phenotypic assay was lower than for R5 strains (59% versus 86%). The sensitivities, specificities, PPVs and NPVs for detecting X4 variants with the different genotypic tools are summarized in Table 6. The sensitivity in X4 detection with the bioinformatics tools never exceeded 80%. The PART method showed the best sensitivity for X4 detection (80%), but its specificity was the worst (60%). These results were due to an overestimation of X4 viruses by this method.

8.00

0.00 0.0

Viral load, copies of RNA

7.00

8.00

Replicate 1 (log RLU)

Figure 3. Assay reproducibility. (a) Tropism determination was assessed in duplicate for separate aliquots of plasma samples in GHOST cells. RLU values (log) are shown for two replicates of four plasma samples. (b) Two replicates of 56 recombinant viruses were evaluated on both GHOST CCR5+ cells (open squares) and CXCR4+ cells (filled diamonds). Both the luciferase activity values and the tropism determination were concordant for the paired replicates.

Correlation between the tropism phenotypic assay and different genotypic methods For 82 samples, the phenotypic results were compared with the genotypic predictions obtained with different algorithms. The

Discussion CCR5 antagonists have been approved for the treatment of HIV-1-infected patients in therapeutic failure. However, these drugs can only be indicated in patients harbouring R5 viruses and, consequently, efficient methods for characterizing HIV-1 tropism are needed. The main objective of this study was to develop a system that allows the monitoring of HIV-1 tropism, and to improve the limitations of current genotypic and phenotypic assays. The recombinant virus system reported here shows good sensitivity, specificity and reproducibility in determining the co-receptor tropism of HIV-1. Although phenotypic prediction based on genotype methods is more accessible for

2497

Gonza´lez et al.

(a)

U87.CCR5

Mix NP1525: SF162

U87.CXCR4

1 000 000

RLU

100 000

10 000

1000

0: 10 0

0. 7: 99 .3 0. 3: 99 .7 0. 1: 99 .9

1: 99

5: 95

95 :5

99 :1

10 0: 0 99 .9 :0 .1 99 .7 :0 .3 99 .3 :0 .7

100

% X4 clone:% R5 clone (b)

U87.CCR5

Mix MN: CM244

U87.CXCR4

1 000 000

RLU

100 000

10 000

1000

0: 10 0

0. 7: 99 .3 0. 3: 99 .7 0. 1: 99 .9

1: 99

5: 95

95 :5

99 :1

99

.3

:0 .7

:0 .3 .7

:0 .1

99

.9 99

10 0: 0

100

% X4 clone:% R5 clone Figure 4. Assay sensitivity for the detection of minor variants in mixed populations. The following mixtures of viruses with envelopes derived from patients were analysed: (a) NP1525-X4:SF162-R5; and (b) MN-X4:CM244-R5. Mean+SEM of triplicate assays. Minority variants of X4 and R5 viruses were detectable in all mixed viral stocks when they represented 1% of the whole population.

Table 4. Comparison of results on viral tropism between our recombinant virus assay and the TrofileTM assay

Table 5. Comparison of results on viral tropism between our recombinant virus assay and the TrofileTM ES assay

New phenotypic assay

New phenotypic assay

TrofileTM

R5

DM

X4

NR

total

R5 DM X4 NR

55 7 0 7

6 22 0 3

0 0 2 1

5 3 0 13

66 32 2 24

Total

69

31

3

21

124

NR, not reported.

2498

TrofileTM ES

R5

DM

X4

total

R5 DM X4

19 1 0

3 12 0

0 0 0

22 13 0

Total

20

15

0

35

JAC


Table 6. Taking the phenotypic assay as a reference, concordance for different genotypic bioinformatic tools and sensitivities, specificities, PPVs and NPVs to predict X4 tropism with these algorithms

R5 concordance R5X4/X4 concordance Global concordance Sensitivity Specificity PPV NPV

PSSM (X4R5)

Geno2pheno[coreceptor]

C4.5

C4.5 (p8 and 12)

PART

SVM

Charge rule

Mean

92% 60% 84% 60% 92% 71% 88%

77% 70% 76% 70% 77% 50% 89%

98% 45% 85% 45% 98% 90% 85%

98% 45% 85% 45% 98% 90% 85%

60% 80% 65% 80% 60% 39% 90%

84% 60% 78% 60% 84% 55% 87%

94% 50% 83% 50% 94% 71% 85%

86% 59% 79% 59% 86% 67% 87%

C4.5, C4.5 decision tree generator; C4.5 (p8 and 12), C4.5 decision tree generator using positions 8 and 12; PART, part rule generator; SVM, support vector machines.

routine laboratories, they present important limitations. These methodologies are based on the principle that the V3 loop is the major determinant of co-receptor usage, but other env regions outside this region contribute to HIV-1 tropism.37 – 40 Besides, genotypic algorithms are less sensitive for detecting CXCR4-using viruses in non-B subtypes than in the B subtype.41 Taking the phenotypic assay described here as a reference, genotypic methods showed low sensitivity in X4 detection. A major limitation of genotypic methods is the detection of variant mixtures (CCR5/CXCR4) at levels ,10% –25%. The results of the study by Low et al.22 showed that currently default implemented co-receptor prediction algorithms are not yet adequate for predicting R5X4 strains in clinical samples. The main advantage of our recombinant virus assay with regard to the genotypic methodologies is that the complete env gene is amplified, which includes all determinants of viral tropism and allows an optimal representation of the viral population. Moreover, our amplification method allows us to assign tropism across diverse viral subtypes and to amplify up to 65% of samples with viral load of ,1000 copies/mL. This is important because maraviroc has been approved for use in treatment-experienced patients8,9 and, at present, a significant proportion of patients in treatment failure display low viraemia levels. In preliminary experiments, two different cell lines (GHOST and U87) were used. GHOST cells displayed higher luciferase activity than U87 when both cell types were infected with similar viral inputs. However, background levels of luciferase in control GHOST.CD4 cells were high when infected with X4 strains, due to the expression of low levels of CXCR4 in these control cells (data not shown). Based on these data, the final routine tropism test was performed using U87 as the target cell line. Unlike phenotypic methodologies based on homologous recombination and viral pseudotypes, the methodological strategy of our system is based on the direct cloning of env in a single vector; so allowing the efficient generation of recombinant viruses that represent viral heterogeneity in a patient’s plasma. A potentially relevant advantage of this system is that recombinant viruses display multiple cycles of replication, thus increasing the signal of luciferase activity and sensitivity of the assay. Taking into account the heterogeneous viral population present in clinical samples, an important aspect in the development of a new system is the sensitivity in detecting minority

clones. Our results have established a threshold sensitivity of 1%, which represents an advantage in the detection of minority X4 variants before initiating treatment with a CCR5 antagonist. Failure in the detection of minority X4 or DM viruses increases the risk of using an ineffective drug, reduces the likelihood of viral suppression and increases the possibility of developing antiretroviral resistance to all drugs used in combination therapy.42,43 However, it must be considered that the mixture of DNA from different viral clones has some limitations. Indeed, mixture experiments allow optimal amplification of envelope sequences, which will probably not be the case in patients with low viral loads. The sensitivity threshold of our test and other phenotypic tests should be confirmed through the mixing of plasmas with pure X4 and R5 viruses at low viral loads or by assessing the proportion of R5 and X4 strains directly in plasma by ultradeep sequencing. In comparing the results obtained using the described method and both TrofileTM and TrofileTM ES assays, there was good concordance (86% and 89%, respectively). The concordance for R5 variants was always good (90% and 86%, respectively) and concordance for R5X4/X4 variants improved from 77% to 92% when our results were compared with TrofileTM ES. In addition, it should be considered that the amplification of RNA sequences found in low proportions can be biased, because of the low efficacy of the retrotranscription process. Indeed, independent RT–PCR reactions should be performed to improve the probability of detecting RNA quasispecies present in low proportions. In summary, we have developed a novel system based on replicative competent recombinant viruses that allows us to determine viral tropism in patients who are candidates for treatment with CCR5 antagonists. Our results show important advantages compared with genotypic methods, and a good correlation with the TrofileTM and TrofileTM ES phenotypic systems.

Acknowledgements Part of this work was presented at the Seventeenth International HIV Drug Resistance Workshop, Sitges, Spain, 2008 (Abstract 109). We thank Cruz Ayerbe for advice on statistical analysis of the data. We would like to acknowledge Olga Palao for excellent secretarial assistance. We would also like to thank The Center of Blood Transfusions (Comunidad de Madrid) for supplying blood of healthy donors.

2499

Gonza´lez et al.

Funding This work was supported by a research grant from Pfizer. Partial support was obtained from the Spanish Health Ministry, Instituto de Salud Carlos III, AIDS Network ISCIII-RETIC RD06/0006/0037, FIS PI05/00017 and PI 080752, FIPSE Foundation 36630/07, EUROPRISE Network of Excellence of the EU (grant number LSHP CT-2006-037611) and Agence Nationale de la Recherche sur le SIDA (ANRS), France. N. G. was supported by a fellowship from the Spanish Ministry of Education and Science (FPI fellowship, SAF00/0028).

Transparency declarations None to declare.

References 1 Chan DC, Fass D, Berger JM et al. Core structure of gp41 from the HIV envelope glycoprotein. Cell 1997; 89: 263– 73. 2 Chen B, Vogan EM, Gong H et al. Structure of an unliganded simian immunodeficiency virus gp120 core. Nature 2005; 433: 834–41. 3 Kwong PD, Wyatt R, Robinson J et al. Structure of an HIV gp120 envelope glycoprotein in complex with the CD4 receptor and a neutralizing human antibody. Nature 1998; 393: 648–59. 4 Weissenhorn W, Dessen A, Harrison SC et al. Atomic structure of the ectodomain from HIV-1 gp41. Nature 1997; 387: 426– 30. 5 Deng H, Liu R, Ellmeier W et al. Identification of a major co-receptor for primary isolates of HIV-1. Nature 1996; 381: 661– 6. 6 Feng Y, Broder CC, Kennedy PE et al. HIV-1 entry cofactor: functional cDNA cloning of a seven-transmembrane, G protein-coupled receptor. Science 1996; 272: 872–77. 7 Berger EA, Doms RW, Fenyo EM et al. A new classification for HIV-1. Nature 1998; 391: 240. 8 Dorr P, Westby M, Dobbs S et al. Maraviroc (UK-427,857), a potent, orally bioavailable, and selective small-molecule inhibitor of chemokine receptor CCR5 with broad-spectrum anti-human immunodeficiency virus type 1 activity. Antimicrob Agents Chemother 2005; 49: 4721 –32. 9 Fatkenheuer G, Pozniak AL, Johnson MA et al. Efficacy of short-term monotherapy with maraviroc, a new CCR5 antagonist, in patients infected with HIV-1. Nat Med 2005; 11: 1170–72. 10 Gonza´lez N, Pe´rez-Olmeda M, Garcıá-Pe´rez J et al. Evaluation of HIV-1 tropism using a new and sensitive system based on recombinant viruses. Antivir Ther 2008; 13 Suppl 3: A119. 11 Weiser B, Philpott S, Klimkait T et al. HIV-1 coreceptor usage and CXCR4-specific viral load predict clinical disease progression during combination antiretroviral therapy. AIDS 2008; 22: 469– 79. 12 Li W, Webb E, Nary L et al. SensiTrop QT: a novel molecular diagnostic assay for the detection and quantification of HIV coreceptor tropism. In: Abstracts of the Fifteenth Conference on Retroviruses and Opportunistic Infections, Boston, 2008. Abstract 919, p. 412. Foundation for Retrovirology and Human Health, Alexandria, VA, USA. 13 Louder MK, Mascola JR. Determination of syncytium-inducing phenotype of primary HIV-1 isolates using MT-2 cells. In: Nelson M, Kim JH, eds. HIV Protocols. New York: Humana Press, 1999; 23–7. 14 Raymond S, Delobel P, Mavigner M et al. Development and performance of a new recombinant virus phenotypic entry assay to determine HIV-1 coreceptor usage. J Clin Virol 2010; 47: 126– 30. 15 Reeves JD, Coakley E, Petropoulos CJ et al. An enhanced-sensitivity TrofileTM HIV coreceptor tropism assay for selecting patients for therapy with entry inhibitors targeting CCR5: a review of analytical and clinical studies. J Viral Entry 2009; 3: 94 –102.

2500

16 Trouplin V, Salvatori F, Cappello F et al. Determination of coreceptor usage of human immunodeficiency virus type 1 from patient plasma samples by using a recombinant phenotypic assay. J Virol 2001; 75: 251–9. 17 van Baelen K, Vandenbroucke I, Rondelez E et al. HIV-1 coreceptor usage determination in clinical isolates using clonal and populationbased genotypic and phenotypic assays. J Virol Methods 2007; 146: 61– 73. 18 Whitcomb JM, Huang W, Fransen S et al. Development and characterization of a novel single-cycle recombinant-virus assay to determine human immunodeficiency virus type 1 coreceptor tropism. Antimicrob Agents Chemother 2007; 51: 566– 75. 19 Delobel P, Nugeyre MT, Cazabat M et al. Population-based sequencing of the V3 region of env for predicting the coreceptor usage of human immunodeficiency virus type 1 quasispecies. J Clin Microbiol 2007; 45: 1572– 80. 20 Hoffman NG, Seillier-Moiseiwitsch F, Ahn J et al. Variability in the human immunodeficiency virus type 1 gp120 Env protein linked to phenotype-associated changes in the V3 loop. J Virol 2002; 76: 3852 –64. 21 Jensen MA, Li FS, van ’t Wout AB et al. Improved coreceptor usage prediction and genotypic monitoring of R5-to-X4 transition by motif analysis of human immunodeficiency virus type 1 env V3 loop sequences. J Virol 2003; 77: 13376–88. 22 Low AJ, Dong W, Chan D et al. Current V3 genotyping algorithms are inadequate for predicting X4 co-receptor usage in clinical isolates. AIDS 2007; 21: F17– 24. 23 Pillai S, Good B, Richman D et al. A new perspective on V3 phenotype prediction. AIDS Res Hum Retrovir 2003; 19: 145–9. 24 Poveda E, Briz V, Roulet V et al. Correlation between a phenotypic assay and three bioinformatic tools for determining HIV co-receptor use. AIDS 2007; 21: 1487 –90. 25 Sing T, Low AJ, Beerenwinkel N et al. Predicting HIV coreceptor usage on the basis of genetic and clinical covariates. Antivir Ther 2007; 12: 1097– 6. 26 Vandenbroucke I, Eygen VV, Rondelez E et al. Minor variant detection at different template concentrations in HIV-1 phenotypic and genotypic tropism testing. Open Virol J 2008; 2: 8 –14. 27 Labrosse B, Labernardiere JL, Dam E et al. Baseline susceptibility of primary human immunodeficiency virus type 1 to entry inhibitors. J Virol 2003; 77: 1610– 3. 28 Fatkenheuer G, Nelson M, Lazzarin A et al. Subgroup analyses of maraviroc in previously treated R5 HIV-1 infection. N Engl J Med 2008; 359: 1442– 55. 29 Trinh L, Han D, Huang W et al. Technical validation of an enhanced sensitivity Trofile HIV coreceptor tropism assay for selecting patients for therapy with entry inhibitors targeting CCR5. Antivir Ther 2008; 13 Suppl 3: A128. 30 Pear WS, Nolan GP, Scott ML et al. Production of high-titer helper-free retroviruses by transient transfection. Proc Natl Acad Sci USA 1993; 90: 8392– 6. 31 Morner A, Bjorndal A, Albert J et al. Primary human immunodeficiency virus type 2 (HIV-2) isolates, like HIV-1 isolates, frequently use CCR5 but show promiscuity in coreceptor usage. J Virol 1999; 73: 2343– 9. 32 Bjorndal A, Deng H, Jansson M et al. Coreceptor usage of primary human immunodeficiency virus type 1 isolates varies according to biological phenotype. J Virol 1997; 71: 7478 –87. 33 Garcia-Perez J, Sanchez-Palomino S, Perez-Olmeda M et al. A new strategy based on recombinant viruses as a tool for assessing drug susceptibility of human immunodeficiency virus type 1. J Med Virol 2007; 79: 127–37. 34 Adachi A, Gendelman HE, Koenig S et al. Production of acquired immunodeficiency syndrome-associated retrovirus in human and


nonhuman cells transfected with an infectious molecular clone. J Virol 1986; 59: 284–91. 35 Collman R, Balliet JW, Gregory SA et al. An infectious molecular clone of an unusual macrophage-tropic and highly cytopathic strain of human immunodeficiency virus type 1. J Virol 1992; 66: 7517– 21. 36 Fenyo EM, Heath A, Dispinseri S et al. International network for comparison of HIV neutralization assays: the NeutNet report. PLoS ONE 2009; 4: e4505. 37 Boyd MT, Simpson GR, Cann AJ et al. A single amino acid substitution in the V1 loop of human immunodeficiency virus type 1 gp120 alters cellular tropism. J Virol 1993; 67: 3649– 52. 38 Carrillo A, Ratner L. Human immunodeficiency virus type 1 tropism for T-lymphoid cell lines: role of the V3 loop and C4 envelope determinants. J Virol 1996; 70: 1301– 9.

JAC 39 Groenink M, Fouchier RA, Broersen S et al. Relation of phenotype evolution of HIV-1 to envelope V2 configuration. Science 1993; 260: 1513–6. 40 Smyth RJ, Yi Y, Singh A et al. Determinants of entry cofactor utilization and tropism in a dualtropic human immunodeficiency virus type 1 primary isolate. J Virol 1998; 72: 4478 –84. 41 Garrido C, Roulet V, Chueca N et al. Evaluation of eight different bioinformatics tools to predict viral tropism in different human immunodeficiency virus type 1 subtypes. J Clin Microbiol 2008; 46: 887–91. 42 Cooper DA, Heera J, Goodrich J et al. Maraviroc versus efavirenz, both in combination with zidovudine –lamivudine, for the treatment of antiretroviral-naive subjects with CCR5-tropic HIV-1 infection. J Infect Dis 2010; 201: 803– 13. 43 Gulick RM, Lalezari J, Goodrich J et al. Maraviroc for previously treated patients with R5 HIV-1 infection. N Engl J Med 2008; 359: 1429– 41.

2501