Bi, Xiuqiong; Gatanaga, Hiroyuki; Ida, Setsuko; Tsuchiya, Kiyoto; Matsuoka-Aizawa, Saori; Kimura, Satoshi; Oka, Shinichi
From the AIDS Clinical Center, International Medical Center of Japan, Tokyo; and Graduate School of Medicine, University of Tokyo, Tokyo, Japan (Drs Bi and Kimura).
Received for publication February 23, 2003; accepted June 4, 2003.
This study was supported by a grant-in-aid for AIDS research from the Ministry of Health, Labor, and Welfare of Japan (H12-AIDS-001), by the Organization of Pharmaceutical Safety and Research (01-4), and by the Japanese Foundation for AIDS Prevention (X. B., K. T., and S. M.-A.).
Reprints: Shinichi Oka, Director, AIDS Clinical Center, International Medical Center of Japan, 1-21-1, Toyama, Shinjuku-ku, Tokyo 162–8655, Japan (e-mail: firstname.lastname@example.org).
It has been possible to control the viral load below detection limits in most HIV-1–infected individuals by highly active antiretroviral therapy (HAART). In a considerable number of patients, however, the viral load remains above the detection limits even during continuous administration of HAART. In these patients, drug-resistant viruses can emerge and invalidate the antiretroviral agents. 1–3 Therefore, early detection of resistant viruses is essential to change invalidated drugs to efficient ones before acceleration of the viral replication process.
Genotype assay using direct sequencing of plasma viruses has been applied for detection of resistant viruses, because plasma viruses turn over rapidly 4,5 and are considered to represent actively replicating HIV-1. In genotype assay of plasma viruses, it is often difficult to obtain results, however, due to the detection limit of gene amplification when the plasma viral load is less than 103 copies/mL. 6 Considering that replication-competent HIV-1 can be recovered from peripheral blood mononuclear cells (PBMCs) even from infected individuals whose plasma viral loads have remained below the detection limit for a long period, 7–11 analysis of HIV-1 proviral sequences in PBMCs during therapy may allow early detection of resistant viruses. Nevertheless, given that there is a genetic discordance between viruses in plasma and proviruses in PBMCs, 12–17 sequences of PBMC proviruses may yield different information.
In this study, we compared HIV-1 protease gene sequences in plasma viruses and PBMC proviruses in serial samples obtained during therapy (including protease inhibitors [PIs]), analyzed their genetic discordances, and assessed the spreading pattern of resistant viruses in plasma and PBMCs of infected individuals.
MATERIALS AND METHODS
In our clinic, the AIDS Clinical Center (ACC), International Medical Center of Japan (IMCJ), almost all HIV-1–infected patients agree to participate in retrospective clinical studies, and their plasma and PBMC stocks from residues of routine blood examinations for future studies are maintained in an ACC-IMCJ sample bank for varying intervals after obtaining signed informed consent.
Measurements of HIV-1 viremia (Amplicor HIV-Monitor, Roche, NJ) and CD4 and CD8 lymphocyte cell counts (monoclonal antibodies and flow cytometry) were performed at each blood sampling. For patients with viral loads persistently over 400 copies/mL in spite of continuous administration of HAART, genotypes of plasma HIV-1 were routinely assayed. In total, 22 HIV-1–infected patients who met the following criteria were enrolled in the present study (Table 1): (1) viruses in plasma must have acquired at least 1 PI resistance–associated mutation 6 and (2) PBMC samples taken at a minimum of 5 time points must be available in the ACC-IMCJ sample bank. The Institutional Ethics Committee approved this study (IMCJ-H13-80), and written informed consent for this study was obtained from each patient.
Sequence Analysis of Protease Gene of HIV-1
Plasma and PBMC samples (1.0–14.4 × 106 copies/sample) were obtained from a 7-mL EDTA-treated blood sample and stocked at −80°C. Total RNA extracted from 200 μL of each plasma sample was subjected to reverse transcriptase (RT) and the first polymerase chain reaction (PCR) with a primer pair of 01 and 02 using the One Step RNA PCR Kit (TaKaRa, Kyoto, Japan), followed by the second PCR with a primer pair of 03 and 04 to amplify HIV protease gene. 18,19 If the viral load was less than 1000 copies/mL or nested PCR could not amplify DNA sufficiently, an increased volume (500 μL) of plasma sample was used for RNA extraction. Total DNA extracted from a 10th of stocked PBMCs was also subjected to the nested PCR with the same sets of primers. Each PCR procedure consisted of 30 cycles of 94°C denaturing, 50°C elongation, and 72°C annealing. Primer sequences were as follows: sense primer 01 5′-CCA ACA GCC CCA CCA GA-3′ (2152–2168), antisense primer 02 5′-ATT TTC AGG CCC ATT TTT TGA-3′ (2711–2691), sense primer 03 5′-AGC AGG AGA CGA TAG ACA AGG-3′ (2213–2233), and antisense primer 04 5′-CTG GCT TTA ATT TTA CTG GTA-3′ (2592–2572). The numbering parallels the HIV-1MN sequence. Direct sequencing was performed bidirectionally with sense primer 03 and antisense primer 04 by an automatic sequencer (model 377; Applied Biosystems, Foster City, CA). A heterozygous base sequence was identified when the chromatogram showed a minor peak at >50% of the major peak.
Estimation of Time Lag Between Emergence of Mutations in Plasma Viruses and Peripheral Blood Mononuclear Cell Proviruses
The latency between the emergence of mutation in plasma viruses and PBMC proviruses was estimated as the difference between the sampling dates at which the mutation was first detected in plasma and PBMCs during the study period (34 ± 12 months [mean ± SD], range: 13–56 months). When the PI(s) was discontinued or changed to other PI(s) while genotypic discordance between plasma viruses and PBMC proviruses on the mutation was persistently noted, the time lag was estimated to have continued until the day of discontinuance or change of PI(s). Such genotypic lag times of all the mutations that had emerged in plasma viral genotype during the study period were analyzed.
Protease Inhibitor Resistance–Associated Mutations of Plasma Viruses Outnumber Those of Peripheral Blood Mononuclear Cell Proviruses
We analyzed 293 plasma and 211 PBMC samples from 22 HIV-1–infected patients enrolled in the study. All samples were collected during treatment that included 1 or more PIs. After the RNA extraction from the plasma samples, nested PCR following RT (RT-PCR) was performed for amplification of HIV-1 protease gene. RT-PCR failed to amplify HIV-1 protease gene from 18 plasma samples, however, all of which contained less than 400 copies/mL HIV-1 RNA (these samples are not included in Table 1). In our assay, when the plasma viral load was less than 400 copies/mL, the success ratio of RT-PCR was 5.3%. On the other hand, enough DNA was amplified by nested PCR from all PBMC samples, even when amplification from the plasma sample taken at the same time failed. Therefore, we analyzed 275 plasma-derived and 211 PBMC-derived HIV-1 samples by direct sequencing (see Table 1) and counted the number of observed PI resistance–associated mutations in each sequence (Table 2).
One hundred seven pairs of plasma-derived and PBMC-derived sequences were obtained from the same blood draws (see Table 1). In 70 of such 107 pairs (65.4%), the numbers of PI resistance–associated mutations were different between plasma-derived and PBMC-derived HIV-1 protease sequences, suggesting that the majority of the viruses in plasma were produced from a small portion of HIV-1–infected PBMCs or noncirculating cell(s).
The average number of such mutations in paired samples (plasma and PBMC, respectively) was calculated in each patient first, and such averages were then analyzed statistically to eliminate possible biasing of data due to the unequal number of samples available for each patient. In 17 of 22 patients, the average number of PI resistance–associated mutations was larger in plasma than in PBMCs, and in total of 22 patients, it was also larger in plasma than in PBMCs, with statistical significance (Table 3, bottom column). When the patients were divided according to geometric mean of plasma viral load during the collection period of paired samples, plasma samples still had significantly larger number of mutations than PBMC samples in both groups. The P value (Wilcoxon signed rank test) was smaller in the patients with a viral load <104 copies/mL than in the patients with a viral load >104 copies/mL (see Table 3, middle columns), however, which suggests that the difference between the number of PI resistance–associated mutations in plasma and that in PBMCs was larger in the patients with a lower viral load. To confirm the relation between the difference in the number of such mutations in 2 compartments and mean viral load, the ratio of average of mutation counts in PBMCs and that in plasma was plotted against mean viral load in each patient (Fig. 1). In the patients with a mean viral load >3 × 104 copies/mL (patients 3, 4, 5, 7, and 13), the ratios of mutation counts were 1.0 ± 0.1. In contrast, the ratios in the patients with a mean viral load <5 × 103 copies/mL (patients 6, 14, 17, 18, and 19) were less than 0.4. The ratio was significantly correlated with mean viral load (Spearman rank correlation coefficient, P = 0.0021). The number of mutations was smaller in PBMCs compared with that in plasma, especially in the patients with a lower viral load. This finding suggests that PI resistance–associated mutations appear in plasma-derived genotypic assay earlier than in PBMC-derived assay during therapy including PIs for patients with a low viral load.
Low Viral Load Is Associated with a Longer Time Lag
Plasma viruses had more PI resistance–associated mutations than PBMC proviruses during PI treatment, especially in the patients with a low viral load. To delineate how the difference in mutations emerged between plasma and PBMCs, we focused on the time lag of appearance of PI resistance–associated mutations in plasma viral and PBMC proviral genotypic assays. During the 34-month (1020-day) study period, a total of 58 PI resistance–associated mutations, including 27 (46.6%) primary mutations for some PIs, emerged in plasma viruses of 20 participants (see Table 2). Primary and secondary mutations emerged in 17 and 16 participants, respectively. The majority of such mutations (53 of 58 [91.4%]) appeared in plasma viral genotypic assay earlier than in PBMC proviral assay. The other mutations appeared in plasma-derived and PBMC-derived genotypic assays simultaneously. The estimated time lag between the emergence of the mutations in plasma viral and PBMC proviral genotypic assays ranged from 0 to 739 days (mean, 289 days).
To analyze the time lag statistically, the averages of time lag of primary and secondary mutations, respectively, were calculated in each participant. There was no significant difference in time lag between primary (mean, 319 days) and secondary (mean, 223 days) mutations (Mann-Whitney U test, P = 0.1444). When the patients were divided by mean viral load, however, the time lag of primary mutation was significantly longer in the patients with a viral load <104 copies/mL (mean, 425 days) than in the patients with a viral load >104 copies/mL (mean, 225 days) (Fig. 2A). The time lag of secondary mutations was also longer in the patients with a low viral load (mean, 294 days) than in the patients with a high viral load (mean, 152 days), although the difference was not significant (see Fig. 2B), probably because of small sample size.
To delineate the correlation between the time lag and viral load further, the mean time lag in each patient was plotted against the mean viral load during the time lag. In both primary and secondary mutations (Fig. 3), there was a significant correlation between the time lag and viral load. It can be said that PI resistance–associated mutations were first detected in plasma-derived genotypic assay and then in PBMC-derived assay and that the time lag was longer in the patients with a low viral load. That may be the reason why we observed more resistance-associated mutations in plasma genotype than in PBMC genotype during PI treatment, especially when the viral load was low.
The major findings of the present study are as follows: (1) more PI resistance–associated mutations observed in plasma viruses than in PBMC proviruses taken at the same time during PI therapy, (2) earlier appearance of such mutations in plasma genotype than in PBMC genotype, and (3) larger and longer genotypic discordance and the time lag of mutation appearance in the patients with a lower viral load.
Our analyses were based on direct sequencing, which could not detect minor genotypic populations. In this respect, Paolucci et al 20 analyzed direct and clonal HIV-1 sequences in plasma and PBMCs and found more resistance-associated mutations in plasma than in PBMCs of the patients during HAART by direct sequencing, which is compatible with our study. Furthermore, they also found multiple minor sequences with resistance-associated mutations in PBMCs by clonal sequences, which could not be detected by direct sequencing. These resistance-associated mutations were probably induced by previous therapy, and they still remained in PBMCs after disappearance from plasma viruses. Resistant genotypes rapidly disappear from plasma after interruption of antiretroviral treatment. 21 Devereux et al 22 detected more resistance-associated mutations in PBMCs than in plasma during the drug-off period of patients with a history of extensive antiretroviral treatment. Venturi et al 23 reported that the drug resistance was greater in plasma viruses than in PBMC proviruses in on-therapy patients and in PBMC proviruses than in plasma viruses in off-therapy patients. These observations, including ours, clearly demonstrate an evolutional change of PBMC proviral genotype later than that of plasma viral genotype.
We estimated the time lag between the emergence of mutations in plasma viral and PBMC proviral genotypes. The accuracy of such time lag is dependent on the frequency of blood sampling. Therefore, the exact time lag must be considered with caution. Nevertheless, patients in whom and the time at which resistance-associated mutations will emerge cannot be predicted; it is difficult to collect samples frequently at the appropriate time prospectively. In this study, we enrolled 22 patients who were already known to have PI-resistant HIV-1 from routine genotype assays and whose plasma and PBMC samples had been frequently stocked in our sample bank and analyzed their 275 plasma-derived and 211 PBMC-derived sequences retrospectively. Statistical analyses of such large samples should at least in part average out the inaccuracies that arise from each estimation of the time lag.
Our results showed that the emergence of primary PI resistance–associated mutations in plasma viruses preceded their emergence in PBMC proviruses by about 425 days when the plasma viral load was lower than 104 copies/mL and that the time lag between emergence of such mutations in plasma-derived and PBMC-derived genotypes correlated inversely with the plasma viral load. In this regard, Kaye et al 17 analyzed zidovudine (AZT) resistance–associated mutations in plasma viruses and PBMC proviruses in 10 patients receiving AZT therapy and reported that the mean time delay between plasma and PBMC mutations was 25 days in 1995 (before HAART was introduced). Considering that their patients were receiving only AZT and were classified as having Centers for Disease Control and Prevention (CDC) stage IV disease, their plasma viral loads might be around 105 copies/mL or higher. Therefore, our results are considered to be compatible with those of Kaye et al. 17
In summary, PBMC proviral genotype evolves later than plasma viral genotype, and the genetic discordance arising from this latency continues longer in the patients with a lower viral load. Plasma viruses should be the material of choice for early detection of drug resistance during antiretroviral treatment.
The authors thank Y. Hirabayashi for continuous discussions throughout this study and Y. Takahashi and F. Negishi for technical support.
1. Lorenzi P, Opravil M, Hirschel B, et al. Impact of drug resistance mutations on virologic response to salvage therapy. Swiss HIV Cohort Study. AIDS. 1999; 13(Suppl):F17–F21.
2. Matsuoka-Aizawa S, Sato H, Hachiya A, et al. Isolation and molecular characterization of a nelfinavir (NFV)-resistant human immunodeficiency virus type 1 that exhibits NFV-dependent enhancement of replication. J Virol. 2003; 77:318–327.
3. Tsuchiya K, Matsuoka-Aizawa S, Yasuoka A, et al. Primary nelfinavir (NFV)-associated resistant mutations during a follow-up period of 108 weeks in protease inhibitor naive patients treated with NFV-containing regimens in an HIV clinic cohort. J Clin Virol. 2003; 27:252–262.
4. Ho DD, Neumann AU, Perelson AS, et al. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature. 1995; 373:123–126.
5. Perelson AS, Neumann AU, Markowitz M, et al. HIV-1 dynamics in vivo
: virion clearance rate, infected cell life-span, and viral generation time. Science. 1996; 271:1582–1586.
6. Hirsch MS, Brun-Vezinet F, D'Aquila RT, et al. Antiretroviral drug resistance testing in adult HIV-1 infection: recommendations of an International AIDS Society-USA Panel. JAMA. 2000; 283:2417–2426.
7. Chun TW, Stuyver L, Mizell SB, et al. Presence of an inducible HIV-1 latent reservoir during highly active antiretroviral therapy. Proc Natl Acad Sci USA. 1999; 94:13193–13197.
8. Finzi D, Hermankova M, Pierson T, et al. Identification of a reservoir for HIV-1 in patients on highly active antiretroviral therapy. Science. 1997; 278:1295–1300.
9. Furtado MR, Callaway DS, Phair JP, et al. Persistence of HIV-1 transcriptase in peripheral-blood mononuclear cells in patients receiving potent antiretroviral therapy. N Engl J Med. 1999; 340:1614–1622.
10. Wong JK, Hezareh M, Günthard HF, et al. Recovery of replication-competent HIV despite prolonged suppression of plasma viremia. Science. 1997; 278:1291–1295.
11. Zhang L, Ramratnam B, Tenner-Racz K, et al. Quantifying residual HIV-1 replication in patients receiving combination antiretroviral therapy. N Engl J Med. 1999; 340:1605–1613.
12. Simmonds P, Zhang LQ, McOmish F, et al. Discontinuous sequence change of human immunodeficiency virus (HIV) type 1 env
sequences in plasma viral and lymphocyte associated proviral populations in vivo
: implications for models of HIV pathogenesis. J Virol. 1991; 65:6266–6276.
13. Michael NL, Chang G, Ehrenberg PK, et al. HIV-1 proviral genotypes from peripheral blood mononuclear cells of an infected patient are differentially presented in expressed sequences. J Acquir Immune Defic Syndr. 1993; 6:1073–1085.
14. 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. J Infect Dis. 1993; 167:445–458.
15. Shafer RW, Kozal MJ, Winters MA, et al. Combination therapy with zidovudine and didanosine selects for drug-resistant human immunodeficiency virus type 1 strains with unique patterns of pol
gene mutations. J Infect Dis. 1994; 169:722–729.
16. Zhang YM, Dawson SC, Landsman D, et al. Persistence of four related human immunodeficiency virus subtypes during the course of zidovudine therapy: relationship between virion RNA and proviral DNA. J Virol. 1994; 68:425–432.
17. Kaye S, Comber E, Tenant-Flowers M, et al. The appearance of drug resistance-associated point mutations in HIV type 1 plasma RNA precedes their appearance in proviral DNA. AIDS Res Hum Retroviruses. 1995; 11:1221–1225.
18. Ida S, Gatanaga H, Shioda T, et al. HIV type 1 V3 variation dynamics in vivo: long-term persistence of non-syncytium-inducing genotypes and transient presence of syncytium-inducing genotypes during the course of progressive AIDS. AIDS Res Hum Retroviruses. 1997; 13:1597–1609.
19. Gatanaga H, Aizawa S, Kikuchi Y, et al. Anti-HIV effect of saquinavir combined with ritonavir is limited by previous long-term therapy with protease inhibitors. AIDS Res Hum Retroviruses. 1999; 15:1493–1498.
20. Paolucci S, Baldanti F, Campanini G, et al. Analysis of HIV drug-resistant quasispecies in plasma, peripheral blood mononuclear cells and viral isolates from treatment-naïve and HAART patients. J Med Virol. 2001; 65:207–217.
21. Verhofstede C, Wanzeele FV, Van Der Gucht B, et al. 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.
22. Devereux HL, Loveday C, Youle M, et al. Substantial correlation between HIV type 1 drug-associated resistance mutations in plasma and peripheral blood mononuclear cells in treatment-experienced patients. AIDS Res Hum Retroviruses. 2000; 16:1025–1030.
23. Venturi G, Romano L, Carli T, et al. Divergent distribution of HIV-1 drug-resistant variants on and off antiretroviral therapy. Antiviral Ther. 2002; 7:245–250.