Divergent distribution of HIV-1 drug-resistant variants on and off antiretroviral therapy

Antiviral Therapy 7:245-250 Divergent distribution of HIV-1 drug-resistant variants on and off antiretroviral therapy Giulietta Venturi 1, Laura Romano 1, Tiziana Carli 2, Paola Corsi 3, Luigi Pippi 4, Pier E Valensin 1 and Maurizio Zazzi 1 * 1 Sezione di Microbiologia, Dipartimento di Biologia Molecolare, Università di Siena, Siena, Italy 2 UO Malattie Infettive Grosseto, Grosseto, Italy 3 UO Malattie Infettive Careggi, Firenze, Italy 4 UO Malattie Infettive Siena, Siena, Italy *Corresponding author: Tel: +39 0577 233 850; Fax: +39 0577 233 870; E-mail: zazzi@unisi.it Objective: To investigate the distribution of drug-resistant HIV-1 variants in plasma RNA and peripheral blood monuclear cell (PBMC) DNA at treatment failure while on therapy and after stopping therapy. Design: Fifty-eight patients failing their first highly active antiretroviral treatment while on therapy and 50 patients after a median of 18.6 weeks after treatment interruption following multiple treatment failures were analysed. Methods: Paired plasma HIV-1 RNA and PBMC HIV-1 DNA were used for genotypic antiretroviral resistance testing. Drug resistance was computed by using the Stanford on-line genotype interpretation system. Results: The extent of drug resistance was larger in plasma RNA than in PBMC DNA in the on-therapy group (P=0.004) and in PBMC DNA than in plasma RNA in the off-therapy group (P=0.04). Interpretation of genotype based on PBMC DNA and plasma RNA in the on-therapy and in the off-therapy group would have missed detection of resistance to one or more drugs in 21 and 22% of the patients, respectively, compared to interpretation based on the other blood compartment. In a subset of 27 patients for whom a sample before stopping therapy was available, there was a significant decrease in the number of RNA (mean 8.1 to 5.3, P=0.004), but not DNA (mean 6.8 to 5.7, P=0.143), resistance mutations following treatment interruption. Conclusion: While plasma RNA is the material of choice for early detection of drug resistance while on therapy, analysis of PBMC DNA may additionally support and possibly improve sensitivity of resistance testing in the absence of therapy. Introduction Development of drug resistance is a major cause for treatment failure in patients adherent to antiretroviral therapy. Indeed, most heavily treated viraemic patients harbour HIV-1 variants resistant to multiple drugs. Recent reports have indicated that discontinuation of antiretroviral therapy favours the overgrowth of the wild-type virus [1,2]. This raises the possibility that structured treatment interruptions allow effective recycling of previously used drugs against a virus population that has been replaced by a predominantly drug-sensitive quasi-species during the washout period [3 5]. Antiretroviral drug resistance testing is commonly performed on HIV-1 RNA circulating in plasma, taken as representative of the actively replicating virus population. However, ongoing replication of different HIV-1 variants also generates a dynamic pool of latently and productively infected cells harbouring a comprehensive HIV-1 genetic archive. This may have important implications in terms of interpretation of drug resistance test results. To investigate the distribution of drug-resistant genomes in cell-associated DNA and plasma RNA in different contexts, we have analysed plasma RNA and peripheral blood mononuclear cells (PBMC) DNA for the presence of drug resistance mutations in 58 patients at their first HAART failure while on therapy and in another 50 patients after treatment interruption. Materials and methods Patients The samples used in this study derived from routine genotypic resistance testing at the HIV Monitoring Service of the University of Siena, where a genotyping service has been established since 1996 and PBMC DNA, in addition to plasma, has been stored for all samples. The study was approved by the Institutional Review Board. Two distinct groups of patients were studied by using all samples received in 2002 International Medical Press 1359-6535/02/$17.00 245

G Venturi et al. the same 6-month period and meeting the following criteria. The first group included 50 subjects who discontinued antiretroviral therapy after at least two (median 4, range 2 11) treatment regimens due to therapy failure. A blood sample was available for drug resistance testing at a median of 18.6 weeks (range 1.0 47.9 weeks) after treatment interruption. Median HIV-1 RNA and CD4 counts were 5.16 log copies/ml (range 2.93 6.69 log copies/ml) and 300 cells 10 6 /l (range 26 861 cells 10 6 /l), respectively. Another sample obtained while on therapy a median of 11.4 weeks (range 1.9 31.1 weeks) before stopping therapy was available for a subset of 27 patients. The second group included 58 previously drug-naive patients experiencing a rebound to >5000 HIV-1 RNA copies/ml after reaching and maintaining undetectable viraemia for at least 6 months with their first combination of two nucleoside reverse transcriptase (RT) inhibitors plus a protease (PR) inhibitor. A single blood sample for each subject was analysed at the time of virological treatment failure while on treatment. Median HIV-1 RNA and CD4 counts for this group were 4.29 log copies/ml (range 3.77 6.20 log copies/ml) and 319 cells 10 6 /l (range 60 1350 cells 10 6 /l), respectively. HIV-1 genotyping and resistance computation PBMCs and plasma were obtained from each available blood sample by standard gradient centrifugation. PBMC DNA was prepared by a salting out procedure and plasma RNA was extracted by using the QIAmp Viral RNA kit (Qiagen, Hilden, Germany). The HIV-1 sequence encoding for the whole PR and for RT amino acids 1 230 was determined from PBMC DNA and from plasma RNA by an in-house method described previously [6], with amplification and sequencing primers modified in order to generate a single 1263 bp fragment and use a dual laser Licor sequencer. Briefly, RNA was reverse transcribed and amplified using Superscript RT for long templates (Life Technologies Italia, S Giuliano Milanese, Italy) with primers PRO1 (5 -AAAAGGGCTGTTGGAAATGTG-3 ; position 2017 2038 on HIV-1 HXB2) and P87 (5 -CCTGSA- TAAATCTGACTTGCCCA-3 ; 3355 3367). The same primer pair was used in the first amplification round for PBMC DNA. Nested PCR was then performed with primers P93 (5 -CTGARAGACAGGC- TAATTTTTTAGG-3 ; 2068 2093) and P75 (5 -CTAAYTTCTGTATRTCATTGACAGTCCA-3 ; 3303 3330) using 1/25 of the first amplification product as the template. The crude PCR product was used directly as the template for a bidirectional cycle sequencing reaction (DYEnamic Direct Cycle Seq kit, Amersham Pharmacia Biotech Italia, Cologno Monzese, Italy) directed by the IRD800-labelled primer IR4 (antisense; 5 -TAGGCTGTACTGTC- CATTTATCAGG-3 ; 3255 3280) and the IRD700- labelled primer IR27 (sense; 5 -ACAACYCCCTCTCMGAAGCAGGA-3 ; 2197 2219). Unpurified termination products were resolved on a Licor automated sequencer (model IR2) and analysed by the E-seq software. Drug susceptibility was inferred by using the on-line drug resistance interpretation system available at the Stanford HIV resistance database website [7]. In order to compare DNA and RNA genotype, resistance was recorded for those drugs labelled with intermediate resistance or high-level resistance. In addition, in order to provide a quantitative estimate more respectful of the different levels of resistance indicated by the Stanford interpretation system, a resistance score was calculated for each genotype by adding 1, 2 and 3 points for each drug labelled with low-level, intermediate and highlevel resistance, respectively. At the time of this analysis, the system predicted susceptibility to the six nucleoside RT inhibitors zidovudine, didanosine, zalcitabine, stavudine, lamivudine and abacavir, the three non-nucleoside RT inhibitors nevirapine, delavirdine and efavirenz, and the five PR inhibitors saquinavir, ritonavir, indinavir, nelfinavir and amprenavir. HIV-1 PR and RT sequences have been submitted to GenBank under accession numbers AF517253 AF517522. Statistics The Wilcoxon signed rank sum test was used to analyse paired differences between continuous variables. The Fisher exact test was used to compare the two groups (subjects on and off therapy) for the number of patients with evidence of drug resistance in HIV-1 RNA but not in HIV-1 DNA and vice versa. The analyses were performed with SPSS 8.0. A P<0.05 was considered significant. Results Of the 58 patients analysed at the time of virological failure while on treatment, 41 (70.7%) harboured virus with significantly reduced susceptibility to at least one drug, based on DNA and/or RNA genotype. Of these, 33 (56.9%) had such an evidence of resistance both in HIV-1 DNA and RNA but eight (13.8%) had it in RNA only. The resistance mutations most frequently occurring in plasma RNA in the absence of their counterpart in PBMC DNA were M184V (eight cases), K70R (four cases), M41L, D67N/E and T215F/Y (three cases each) in RT and D30N (three cases), V82A and L90M (two cases each) in PR. Overall, there was a higher number of resistance mutations in the sequences derived from plasma RNA than in those derived from PBMC DNA (Table 1). This 246 2002 International Medical Press

HIV-1 resistance on and off therapy translated into a higher number of predictably inactive drugs and a higher resistance score. For example, there were 12 (20.7%) patients where interpretation of DNA genotype would have missed resistance to one or more drugs (mean 3.4; 95% CI: 1.9 5.0 drugs) with respect to interpretation of RNA genotype (Figure 1A and Table 2). Of the 50 subjects analysed after treatment interruption, 33 (66.0%) had evidence of drug resistance in viral DNA and/or RNA. In contrast with the results obtained on the former group, there were five (10.0%) patients with evidence of drug resistance in PBMC DNA but not in plasma RNA and only one (2.0%) subject with the opposite picture. The prevalence of patients with evidence of drug resistance only in viral DNA and only in viral RNA was significantly different between the on-therapy and the off-therapy group (P=0.019 and 0.036, respectively). Thymidine analogue mutations were those most often present in RT DNA only (M41L, five cases; K70R, four cases; T215F/Y and K219N/Q/E, three cases), followed by M184I/V and T69N (two cases each). There was only one occurrence of a primary PR inhibitor resistance mutation (V82A) in PBMC DNA but not in plasma RNA. Although the number of DNA and RNA mutations was not statistically different in the whole off-therapy group, there was a significantly larger extent of drug resistance based on interpretation of viral DNA with respect to viral RNA (Table 1). In 11 (22.0%) patients, interpretation of drug susceptibility based on plasma RNA would have missed resistance to at least one drug (mean 2.7; 95% CI: 1.4 4.0 drugs) with respect to PBMC DNA (Figure 1B and Table 2). However, there were also four (8.0%) patients where the opposite pattern was detected (mean 1.7 drugs, 95% CI: 0.2 3.3 drugs), mainly regarding non-nucleoside RT inhibitors. A blood sample obtained 10 37 weeks before therapy interruption for three of these patients indicated that the mutations present only in HIV-1 RNA in the off-therapy samples were not present in HIV-1 DNA in the on-therapy sample (data not shown). In the subset of 27 patients for whom a sample before stopping therapy was available, there was a significant decrease in the number of RNA (mean 8.1 [95% CI: 6.2 10.1] to 5.3 [95% CI: 3.2 7.3], P=0.004), but not DNA (mean 6.8 [95% CI: 4.8 8.7] to 5.7 [95% CI: 3.8 7.5], P=0.143) resistance mutations following treatment interruption. In this small group of patients the mutations most often maintained in PBMC DNA but lost in plasma RNA were M41L, E44D, L210W in RT and L90M in PR, while other mutations showed a comparable decline (D67N, K103N, M184V, T215Y in RT; M46L, I54V, V82A in PR) or durability (K70R in RT) in both compartments. This more prolonged maintenance of mutations in PBMC DNA with respect to plasma RNA was consistent with the temporal trend in the difference between the numbers of DNA and RNA mutations in the whole off-therapy group: there was an average 0.13, 0.22 and 0.81 more DNA than RNA mutations at <12 weeks (n=16), 12 24 weeks (n=18) and >24 weeks (n=16), respectively, after stopping treatment. Discussion The results obtained in the on-therapy group corroborate clearly the use of plasma RNA for resistance testing at treatment failure. While earlier appearance of Table 1. HIV-1 drug resistance in plasma RNA and PBMC DNA for 58 subjects on therapy and 50 subjects off therapy. Values are presented as mean with 95% confidence intervals given in parentheses. Both primary and secondary mutations are considered Group and parameter Plasma RNA PBMC DNA P-value On-therapy group (n=58) PR inhibitor resistance mutations 2.3 (1.9 2.8) 1.9 (1.4 2.3) 0.002 RT inhibitor resistance mutations 2.0 (1.5 2.5) 1.5 (1.0 1.9) 0.001 Total drug resistance mutations 4.3 (3.6 5.1) 3.4 (2.6 4.1) <0.001 Drugs to which resistance was detected * 2.3 (1.6 3.0) 1.7 (1.1 2.3) 0.016 Resistance score 7.7 (5.8 9.6) 5.9 (4.1 7.6) 0.004 Off-therapy group (n=50) PR inhibitor resistance mutations 2.7 (1.9 3.4) 2.8 (2.1 3.6) 0.464 RT inhibitor resistance mutations 2.4 (1.7 3.2) 2.7 (2.0 3.4) 0.123 Total drug resistance mutations 5.2 (3.8 6.5) 5.6 (4.3 6.9) 0.124 Drugs to which resistance was detected * 2.7 (1.7 3.7) 3.1 (2.1 4.1) 0.050 Resistance score 8.1 (5.3 11.0) 9.5 (6.7 12.3) 0.040 * Resistance was recorded for the drugs included in the two top resistance categories ( intermediate-level and high-level resistance) by the on-line Stanford genotype interpretation system. For each drug, 1, 2 or 3 points were assigned for low-level, intermediate-level and high-level resistance, respectively, as indicated by the on-line Stanford genotype interpretation system. Antiviral Therapy 7:4 247

G Venturi et al. Figure 1. Number of patients harbouring virus resistant to individual drugs in plasma RNA (grey bars) and PBMC DNA (white bars) in the on-therapy (A) and off-therapy (B) group 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 A B AZT DDI DDC D4T 3TC ABC NVP DLV EFV SQV RTV IDV NFV APV AZT DDI DDC D4T 3TC ABC NVP DLV EFV SQV RTV IDV NFV APV Resistance was recorded for the intermediate-level and high-level resistance categories according to the on-line Stanford genotype interpretation system. AZT, zidovudine; DDC, zalcitabine; DDI, didanosine; D4T, stavudine; 3TC, lamivudine; ABC, abacavir; NVP, nevirapine; DLV, delavirdine; EFV, efavirenz; SQV, saquinavir; RTV, ritonavir; IDV, indinavir; NFV, nelfinavir; APV, amprenavir. RT drug resistance mutations in plasma RNA than in PBMC DNA was initially suggested by anecdotal reports [8 10], some recent studies have indicated a substantial agreement between the resistance data obtained from analysis of the two blood compartments [11,12]. This report is the first plain evidence that using PBMC DNA for resistance testing at treatment failure entails the possibility of missing relevant resistance mutations that are concomitantly detectable in plasma RNA. It is likely that this discrepancy concerns particularly the initial development of resistance when the circulating resistant virus population has been not yet extensively archived as proviral DNA. Maintenance of drug pressure is then expected to result in homogenisation of resistance mutations in the two compartments. Thus, the time elapsed between selection of drug resistance mutations and resistance testing may play a major role in determining the rate of agreement between PBMC and plasma resistance data. On the other hand, there a was a trend towards detection of a higher number of drug resistance mutations in PBMC DNA compared to plasma RNA in the group of patients analysed after stopping antiretroviral therapy. Although this difference did not reach statistical significance, interpretation of the RNA genotype failed to predict resistance to at least one drug in more than a fifth of the patients with respect to the DNA genotype. As expected, this increased sensitivity was more evident as more time passed since treatment interruption. Recent reports have indicated a rapid replacement of drug-resistant variants by the wild-type virus in plasma RNA after treatment interruptions [1,2], feeding the hypothesis that recycling of drugs to which resistance was previously detected can be successful after a washout period. Indeed, short-term responses to re-initiation of therapy have been documented after a treatment interruption [3 5]. However, our parallel plasma RNA and PBMC DNA analysis indicates that drug resistance mutations may persist longer in the genetic archive than in circulating free viral particles, possibly hampering long-term response. It must be noted that our analysis did not differentiate between integrated and unintegrated HIV DNA in PBMC, nor can we exclude that at least some drug resistance mutations were in defective proviruses unable to sustain replication. Moreover, given the present sensitivity limitations of genotypic resistance testing [13], there is no warranty that resistant quasispecies are no more present even when they are not detected by direct sequencing of PBMC DNA [14]. A further possible implication of these findings is that resistance testing in drug-naive subjects is also expected to be more sensitive when performed on PBMC DNA rather than on plasma RNA, since transmission of a mixture of wild-type and resistant viruses could resemble a treatment interruption in the presence of drug resistance. However, the case should be different if a totally resistant virus population is transmitted, since true reversion to wild-type would take longer than replacement of the resistant variant(s) by the co-transmitted wild-type virus. Given the relatively low rate of transmission of drug-resistant HIV-1, parallel PBMC and plasma analysis of a large number of drug-naive subjects is required to address this issue. It must be noted that a minority of samples had a larger number of drug resistance mutations in plasma RNA than in PBMC DNA after treatment interruption. As suggested by the lack of these additional mutations also in PBMC DNA obtained before stopping therapy, it is likely that the mutations present in RNA only had not yet been archived in PBMC DNA at the time of 248 2002 International Medical Press

HIV-1 resistance on and off therapy Table 2. Comparison between drug resistance scores derived from plasma RNA and PBMC DNA HIV genotype for individual drugs in 58 subjects on therapy and 50 subjects off therapy On-therapy group No. of cases with resistance score Off-therapy group No. of cases with resistance score Drug RNA>DNA DNA>RNA RNA=DNA RNA>DNA DNA>RNA RNA=DNA Zidovudine 9 2 47 2 8 40 Didanosine 5 1 52 2 3 45 Zalcitabine 5 1 52 4 4 42 Stavudine 5 1 52 4 4 42 Lamivudine 9 0 49 1 2 47 Abacavir 6 1 51 3 4 43 Nevirapine 3 1 54 3 2 45 Delavirdine 1 1 56 3 2 45 Efavirenz 2 1 55 3 3 44 Saquinavir 3 0 55 0 2 48 Ritonavir 3 0 55 0 2 48 Indinavir 3 0 55 1 2 47 Nelfinavir 7 0 51 0 2 48 Amprenavir 1 0 57 1 1 48 Drug resistance score is defined in Materials and methods. treatment interruption. The time required for this archiving could be in general proportional to the fitness defects of the resistant variants, similar to the rate of replacement of the mutated quasispecies by the wild-type virus following treatment interruption [15,16]. In summary, while plasma RNA is the material of choice for early detection of drug resistance, PBMC DNA merits further investigation at least as an additional source for detecting persistent drug resistance mutations in the absence of therapy. However, since the effects of drug resistance on viral fitness are likely to play an important role in vivo, analysis of both compartments may help design appropriate strategies for managing drug resistance in clinical practice. Ackowledgements This work was supported by Istituto Superiore di Sanità, Ministero della Sanità, Rome, Italy (Grants 30C.79, 30B.15) and Fondazione Monte dei Paschi di Siena, Siena, Italy (Grant Diagnostica microbiologica diretta mediante tecniche biomolecolari ). References 1. Devereux HL, Youle M, Johnson MA & Loveday C. Rapid decline in detectability of HIV-1 drug resistance mutations after stopping therapy. AIDS 1999; 13:F123 F127. 2. Verhofstede C, Wanzeele FV, Van Der Gucht B, De Cabooter N & Plum J. Interruption of reverse transcriptase inhibitors or a switch from reverse transcriptase to protease inhibitors resulted in a fast reappearance of virus strains with a reverse transcriptase inhibitor-sensitive genotype. AIDS 1999; 13:2541 2546. 3. Miller V, Sabin C, Hertogs K, Bloor S, Martinez-Picado J, D Aquila R, Larder B, Lutz T, Gute P, Weidmann E, Rabenau H, Phillips A & Staszewski S. Virological and immunological effects of treatment interruptions in HIV-1 infected patients with treatment failure. AIDS 2000; 14:2857 2867. 4. Izopet J, Massip P, Souyris C, Sandres K, Puissant B, Obadia M, Pasquier C, Bonnet E, Marchou B & Puel J. Shift in HIV resistance genotype after treatment interruption and short-term antiviral effect following a new salvage regimen. AIDS 2000; 14:2247 2255. 5. Deeks SG, Wrin T, Liegler T, Hoh R, Hayden M, Barbour JD, Hellmann NS, Petropoulos CJ, McCune JM, Hellerstein MK & Grant RM. Virologic and immunologic consequences of discontinuing combination antiretroviraldrug therapy in HIV-infected patients with detectable viremia. New England Journal of Medicine 2001; 344:472 480. 6. Zazzi M, Riccio ML, Venturi G, Catucci M, Romano L, De Milito A & Valensin PE. Long-read direct infrared sequencing of crude PCR products for prediction of resistance to HIV-1 reverse transcriptase and protease inhibitors. Molecular Biotechnology 1998; 10:1 8. 7. Shafer RW, Jung DR, Betts BJ, Xi Y & Gonzales MJ. Human immunodeficiency virus reverse transcriptase and protease sequence database. Nucleic Acids Research 2000; 28:346 348. 8. Smith MS, Koerber KL & Pagano JS. Zidovudine-resistant human immunodeficiency virus type 1 genomes detected in plasma distinct from viral genomes in peripheral blood mononuclear cells. Journal of Infectious Diseases 1993; 167:445 448. 9. Kozal MJ, Shafer RW, Winters MA, Katzenstein DA & Merigan TC. A mutation in human immunodeficiency virus reverse transcriptase and decline in CD4 lymphocyte numbers in long-term zidovudine recipients. Journal of Infectious Diseases 1993; 167:526 532. 10. Kaye S, Comber E, Tenant-Flowers M & Loveday C. The appearance of drug resistance-associated point mutations in HIV type 1 plasma RNA precedes their appearance in proviral DNA. AIDS Research and Human Retroviruses 1995; 11:1221 1225. 11. Gunthard HF, Wong JK, Ignacio CC, Guatelli JC, Riggs NL, Havlir DV & Richman DD. Human immunodeficiency virus replication and genotypic resistance in blood and lymph nodes after a year of potent antiretroviral therapy. Journal of Virology 1998; 72:2422 2428. Antiviral Therapy 7:4 249

G Venturi et al. 12. Devereux HL, Loveday C, Youle M, Sabin CA, Burke A & Johnson M. Substantial correlation between HIV type 1 drug-associated resistance mutations in plasma and peripheral blood mononuclear cells in treatment-experienced patients. AIDS Research & Human Retroviruses 2000; 16:1025 1030. 13. Hance AJ, Lemiale V, Izopet J, Lecossier D, Joly V, Massip P, Mammano F, Descamps D, Brun-Vezinet F & Clavel F. Changes in human immunodeficiency virus type 1 populations after treatment interruption in patients failing antiretroviral therapy. Journal of Virology 2001; 75:6410 6417. 14. Albrecht D, Zollner B, Feucht HH, Lorenzen T, Laufs R, Stoehr A & Plettenberg A. Reappearance of HIV multidrug-resistance in plasma and circulating lymphocytes after reintroduction of antiretroviral therapy. Journal of Clinical Virology 2002; 24:93 98. 15. Devereux HL, Emery VC, Johnson MA & Loveday C. Replicative fitness in vivo of HIV-1 variants with multiple drug resistance-associated mutations. Journal of Medical Virology 2001; 65:218 224. 16. Birk M, Svedhem V & Sonnerborg A. Kinetics of HIV-1 RNA and resistance-associated mutations after cessation of antiretroviral combination therapy. AIDS 2001; 15:1359 1368. Received 23 January 2002; accepted 16 July 2002 250 2002 International Medical Press