*Institute of Infectious and Tropical Diseases, University of Brescia, Brescia, Italy, †Institute of Infectious Diseases, University of Bari, Bari, Italy, ‡Biostatistics Unit, IRCCS Policlinico S. Matteo, Pavia, Italy, §Infectious Diseases Section, Department of Medical and Occupational Medicine, University of Foggia, Foggia, Italy, and ||Infectious Diseases Department, S. M. Annunziata Hospital, ASL Firenze, Firenze, Italy
C. T. and E. Q. R. contributed equally to the work.
Reprints: Carlo Torti, Institute of Infectious and Tropical Diseases University of Brescia, P.le Spedali Civili, 1 25123 Brescia, Italy (e-mail: firstname.lastname@example.org).
To the Editor:
Discordance between rules-based interpretation systems and results of phenotyping for the evaluation of resistance to antiretroviral drugs exists, 1,2 particularly for dideoxynucleoside reverse transcriptase inhibitors. The inefficient conversion of these drugs to the active compound ddNTP in stimulated lymphocytes in vitro 3 might be responsible for underestimation of HIV-1 phenotypic resistance to dideoxynucleoside, thus explaining the low levels of phenotypic resistance detected even for patients heavily pretreated with these drugs. However, the actual mechanisms of this event are still unknown.
The objective of the present study was to analyze genotypic changes in the HIV-1 reverse transcriptase gene that might explain low-level phenotypic resistance to dideoxynucleoside.
Two clinical cohorts of patients enrolled in 2 prospective trials of real versus virtual phenotyping were merged. 4,5 Resistance testing was performed at a central laboratory (Virco, Mechelen, Belgium). For this substudy, only the results of real phenotyping were considered. Mutations were defined as amino acid changes from consensus B of the reverse transcriptase sequence (HXB2) and ranked into 3 groups: those reported by IAS in 2003 6 and referred to as “IAS mutations”; those occurring at the same positions as in the IAS list but involving different amino acid substitutions (“IAS substitutions”); and any other naturally occurring amino acid change (“polymorphisms”). Univariate and multivariate logistic regression analyses were performed to test for the association between the fold resistance increase for dideoxynucleoside and the prevalence of amino acid changes, entering in the multivariate model only variables associated with P ≤ 0.2 in the univariate model. The median value for fold resistance increase for each dideoxynucleoside in our study was considered as the cutoff value. Analyses were performed with STATA (Stata Statistical Software release 7.0; StataCorp 2000, College Station, TX) and with STATISTICA for Windows (computer program manual; StatSoft, Inc. 2000, Tulsa, OK). P < 0.05 was considered to indicate statistical significance.
Two hundred fifty-five patients for whom HIV complete sequences of the reverse transcriptase gene and real phenotyping results were available were selected for this study. The mean age ± SD of the patients was 39.5 ± 7.5 years. The mean CD4+ T-cell count ± SD was 319 ± 208. The mean HIV RNA level ± SD was 4.5 ± 0.8 log10 copies/mL. Patients had experience with a median of 4 (interquartile range [IQR], 3–5) nucleoside reverse transcriptase inhibitors, 1 (IQR, 0–1) non-nucleoside reverse transcriptase inhibitor, and 2 (IQR, 1–3) protease inhibitors. HIV isolates harbored a median of 4 (IQR, 2–5) IAS mutations, 0 (IQR, 0–1) IAS substitution, and 16 (IQR, 14–19) polymorphisms.
First, descriptive analysis of the phenotypic fold resistance increase for each dideoxynucleoside was performed. The median with respect to the wild-type reference strain was 1.1-fold resistance for either zalcitabine or stavudine and 1.2-fold resistance for didanosine. Second, possible amino acid changes associated with low-level phenotypic resistance were investigated by logistic regression analysis, using the above cutoff values. Multivariate analysis results are illustrated in Figure 1. Polymorphisms at position 200 (T200A/I/E/K/R/S/V) were inversely associated with an increase of >1.1-fold resistance to stavudine (odds ratio [OR], 0.36; 95% confidence interval [CI], 0.17–0.76; P = 0.007), while the presence of L214F (OR, 3.20; 95% CI, 1.24–8.21; P = 0.016) and polymorphisms at positions 377 (T377L/S/Q/M/R/I/K/A/V) (OR, 3.08; 95% CI, 1.22–7.74; P = 0.017) and 386 (T386I/A/V/S/M) (OR, 2.28; 95% CI, 1.10–4.73; P = 0.02) were directly associated with >1.1-fold resistance to this drug (Fig. 1a). With regard to didanosine, polymorphisms at position 200 were protective (OR, 0.47; 95% CI, 0.23–0.97; P = 0.04). By contrast, L74VI mutations appeared to confer resistance (OR, 3.56; 95% CI, 1.34–9.46; P = 0.01) (Fig. 1b). Last, an increase of >1.1-fold resistance to zalcitabine was directly related to polymorphisms involving positions 162 (S162A/N/G/Y/T/C/D/H/F) (OR, 3.02; 95% CI, 1.07–8.48; P = 0.03) and 377 (T377L/S/Q/M/R/I/K/A/V) (OR, 3.23; 95% CI, 1.60–7.84; P = 0.01). Changes at positions 70 (K70R or K70E/G/R) (OR, 0.33; 95% CI, 0.12–0.90; P = 0.03), 181 (Y181C/I or Y181V) (OR, 0.31; 95% CI, 0.12–0.78; P = 0.01), and 200 (OR, 0.47; 95% CI, 0.22–0.99; P = 0.04) were associated with a ≤1.1-fold resistance to this drug (Fig. 1c). These associations were independent of the type of previous or current reverse transcriptase inhibitor (either nucleoside or non-nucleoside) included in the antiretroviral regimen (data not shown).
As more studies are conducted, new mutations gain a role in modulating phenotypic susceptibility, as recently demonstrated by a broader spectrum of mutations in the last update of the IAS list. 6 There is much debate on the issue of mutations associated with the variability of phenotypic resistance to dideoxynucleoside. On one hand, technical reasons have been suggested to play a role in this phenomenon 3; on the other hand, it is conceivable that molecular mechanisms marked by specific mutations, which have not yet been ascertained, might be implicated in phenotypic resistance to dideoxynucleoside. Our data suggest that amino acid substitutions at position 200 are constantly correlated with low-level resistance and that substitutions at newly recognized positions might modulate phenotypic resistance to individual dideoxynucleosides.
Several limitations of the present study need to be recognized. First, data were obtained using arbitrary cutoff values. However, such cutoff values have been related to clinical response in some studies 7–9 and were lower than the biologic cutoff values currently used (updated from Harrigan et al 10). Second, even if multivariate analysis suggested an independent effect of each of these mutations, we were unable to assess specific resistance patterns due to the small patient sample. Third, because of the retrospective nature of our analysis, the hypothesis herein suggests that the merits should be tested in prospective studies, using clinical end points as outcome measures.
In conclusion, these data suggest that phenotypic susceptibility to dideoxynucleoside is influenced not only by IAS mutations but also by a constellation of other amino acid changes. Although some of these changes may increase phenotypic resistance, others (such as those at position 200) appear to increase phenotypic susceptibility. Further studies conducted with clinical end points and site-directed mutagenesis experiments are needed to demonstrate whether the genetic changes described are merely natural genetic polymorphisms or mutations that have a role in dideoxynucleoside drug resistance or hypersusceptibility.
Sergio Lo Caputo||
for the GenPheRex and PhenGen Study Groups of the MASTER Cohort
*Institute of Infectious and Tropical Diseases, University of Brescia, Brescia, Italy, †Institute of Infectious Diseases, University of Bari, Bari, Italy, ‡Biostatistics Unit, IRCCS Policlinico S. Matteo, Pavia, Italy, §Infectious Diseases Section, Department of Medical and Occupational Medicine, University of Foggia, Foggia, Italy, ||Infectious Diseases Department, S. M. Annunziata Hospital, ASL Firenze, Firenze, Italy
1. Torti C, Quiros-Roldan E, Keulen W, et al. Comparison between rules-based HIV-1 genotype interpretations and real-phenotype or virtual-phenotype: concordance analysis and correlation with clinical outcome in heavily treated patients. J Infect Dis
2. De Luca A, Cingolani A, Di Giambenedetto S, et al. Variable prediction of antiretroviral treatment outcome by different systems for interpreting genotypic HIV-1 drug resistance. J Infect Dis
3. Shafer RW. Genotypic testing for HIV type 1 drug resistance. Clin Microbiol Rev
. 2002;15: 247–277.
4. Mazzotta F, Lo Caputo S, Torti C, et al. Real versus virtual phenotype to guide treatment in heavily pretreated patients: 48-week follow-up of the Genotipo-Fenotipo di Resistenza (Gen-PheRex) trial. J Acquir Immune Defic Syndr
5. Saracino A, Angarano G, PhenGen Study Group. Change of antiretroviral therapy guided by genotypic or phenotypic resistance testing: an open label, randomized, multicentre study (PhenGen). HIV Med
6. Johnson VA, Brun-Vezinet F, Clotet B, et al. Drug resistance mutations in HIV-1. Top HIV Med
7. Katzenstein D, Bosch RT, Wang N, et al. Baseline phenotypic susceptibility and virologic failure over 144 weeks among nucleoside RT inhibitor-experienced subjects in ACTG 364. 9th Conference on Retroviruses and Opportunistic Infections, Seattle, February 2002.
8. Swanstrom R, Cheng H, Hellmann N, et al. Alternative phenotypic sensitivity scores and outcome measures increase detection of the impact of drug resistance on response to salvage therapy—results from ACTG 359. 9th Conference on Retroviruses and Opportunistic Infections, Seattle, February 2002.
9. Shulman NS, Hughes MD, Winters MA, et al. Subtle decreases in stavudine phenotypic susceptibility predict poor virologic response to stavudine monotherapy in zidovudine experienced patients. J Acquir Immune Defic Syndr
10. Harrigan PR, Montaner JS, Wegner SA, et al. World-wide variation in HIV-1 phenotypic susceptibility in untreated individuals: biologically relevant values for resistance testing. AIDS