Hsu, Li-Yang MRCP*; Subramaniam, Ravathi RN*; Bacheler, Lee†; Paton, Nicholas I. MD*
The development of drug resistance is the main factor limiting long-term success of antiretroviral therapy. An understanding of the pattern of development of resistance mutations is crucial for making rational decisions about treatment regimens, both for the individual patient and for populations. Virtually all published data on HIV drug resistance have been obtained from studies of subtype B virus, which is the dominant subtype in Western countries. However, subtype B virus accounts for only a minority of the HIV-infected individuals worldwide, and there is emerging evidence that other subtypes may differ in the pattern of development of drug resistance mutations.1 With the advent of major international initiatives to expand drug access, antiretroviral therapy will become increasingly used by HIV-infected patients living in developing countries. Socioeconomic factors or limitation in the development of health care infrastructure may result in use of suboptimal drug combinations or poor long-term treatment adherence. These are potent factors promoting the development of drug resistance, and resistance is likely to be a major concern in the scale-up of treatment programs. Hence, there is an urgent need to address the deficiency of knowledge about drug resistance in non-B viral subtypes to provide a rational basis for development of country- or region-specific treatment policies.
CRF01_AE virus is the predominant subtype in Southeast Asia where it has been estimated to account for 63% of new infections.2 We performed this study to determine the pattern of resistance mutations and polymorphisms in CRF01_AE virus isolates from treatment-exposed and -naive patients in Singapore.
We recruited consecutive eligible patients attending the outpatient clinics at the Communicable Disease Center, Tan Tock Seng Hospital (the national referral center for HIV infection in Singapore) between February 2002 and May 2003. Patients were eligible for study inclusion if they were currently receiving treatment with antiretroviral therapy and had 1 of the following: a detectable viral load during treatment of >400 copies/mL; if viral load measurement was not available, a progressive decline in CD4 cell count during consecutive tests; or if neither viral load nor consecutive CD4 cell count measurements were available, received treatment with a suboptimal (non-highly active antiretroviral) treatment regimen for at least 1 year. We also recruited consecutive newly diagnosed HIV-infected patients presenting to our clinic between March and May 2003 as an antiretroviral-naive comparison group.
All patients gave written informed consent, and the Tan Tock Seng Hospital Ethics Committee approved the study.
Case records were reviewed to abstract relevant clinical and demographic data, including risk factors for acquisition of HIV infection, detailed antiretroviral treatment history, and current CD4 cell count and viral load where available.
A 5-mL blood sample was obtained from each patient in an EDTA tube, and the plasma was separated and stored at −20°C. Samples were transported in batches to Virco (Mechelen, Belgium) for genotypic analysis. In brief, the analysis involved the extraction of viral RNA from 200 μL of plasma (QIAamp Viral RNA Extraction Kit, Qiagen, Hilden, Germany), amplification of reverse transcriptase (RT) and protease sequences using RT-PCR analysis, and automated dye terminator sequencing of the entire protease and up to codon 400 of the RT genes. Amino acid differences from a reference HXB2 genotype were determined. Characterization of HIV-1 subtypes was done by calculating an overall similarity score for the protease and RT sequences of the isolate in comparison with reference sequences for other subtypes and recombinant viruses. (Nucleotide and amino acid RT and protease sequences of CRF01_AE virus isolates in this study were submitted to GenBank.)
The frequency of mutations in CRF01_AE virus isolates was compared with reference data obtained from the Stanford database for treatment-exposed subtype B virus isolates3 and with the frequency of mutations in our CRF01_AE virus isolates from treatment-naive patients. For the purposes of classification of the frequency of mutations by drug class, isolates were included if they were from patients who had been exposed to the drug class in the past, even if they were not currently receiving drugs from that class at the time the sample was collected.
Drug resistance mutations were defined as mutations known to cause drug resistance based on published experimental data.4 CRF01_AE virus-specific polymorphisms were defined as mutations that occurred at significantly higher frequencies in CRF01_AE virus isolates from treatment-exposed patients than in subtype B virus isolates but were present in similar frequencies among treatment-exposed and treatment-naive patients.
A sample size of ∼100 treatment-experienced patients and 50 treatment-naive patients was selected based on considerations of obtaining representative data together with limitations of feasibility and cost. Frequencies were compared using the χ2 test or, if a cell value was <5, Fisher exact test. Statistical analysis was conducted using Analyze-it for Microsoft Excel version 1.68. In view of the large number of comparisons performed, we considered P < 0.01 to be indicative of statistical significance.
We obtained samples from 97 treatment-exposed patients for whom resistance was suspected. These samples yielded 69 CRF01_AE virus isolates, 18 subtype B virus isolates, and 2 subtype C virus isolates; viral amplification failed for 8 samples. Viral load testing was performed retrospectively on the samples that failed genotyping and showed values of <1000 copies/mL in all cases. Forty-five samples were obtained from treatment-naive patients. These samples yielded 35 CRF01_AE virus isolates, 6 subtype B virus isolates, 1 subtype C virus isolate, and 1 subtype AG virus isolate; viral amplification failed for 2 samples. The viral load for the 2 samples that failed genotyping was >100,000 copies/mL, and hence, failure most likely represents sample degradation. We present data for 69 treated patients and 35 treatment-naive patients infected with CRF01_AE virus because the other groups were too small for meaningful comparison.
The demographic and clinical data for the patients are shown in Table 1. Most patients were male, were of Chinese ethnicity, and reported heterosexual contact with high-risk sexual partners as the risk factor for transmission. Nucleosides were used at some time by 94% of patients, and 30% and 49% had been exposed to treatment with nonnucleoside RT inhibitor (NNRTI) and protease inhibitor (PI) drugs, respectively. At the time the samples were collected, 16 patients were receiving highly active antiretroviral therapy with a PI, 13 were receiving highly active antiretroviral therapy with an NNRTI, 16 were taking dual nucleoside RT inhibitors (NRTIs) only, 7 were taking dual PIs only, 8 were taking a combination including both NNRTIs and PIs, and the remainder were taking other combinations.
Drug Resistance Mutations in Treatment-Naive Patients
Mutations at sites of known drug resistance were uncommon in treatment-naive patients. There were no major mutations found at NRTI resistance sites. Three treatment-naive patients (9%) had the V106I mutation, and 7 (20%) had the V179I mutation; however, there were no patients with K103N or other major NNRTI mutations. At sites of protease resistance in treatment-naive patients, M36I occurred in 100%, K20R, in 18%, and L10I, L63P, and I93L, each in 11%.
Drug Resistance Mutations in NRTI Treatment-Exposed Patients
Sequences were available for 65 patients infected with CRF01_AE virus who had received treatment with ≥1 NRTIs. Frequencies of mutations at known sites of NRTI drug resistance are shown in Table 2, with frequencies of mutations obtained from the Stanford database for subtype B virus isolates from patients exposed to ≥1 NRTIs shown for comparison. The overall frequency of mutations was similar for treatment-exposed patients infected with CRF01_AE virus and those infected with subtype B virus. The only statistically significant differences were that D67N was more common in CRF01_AE virus and M41L and T215Y were less common in CRF01_AE virus than in subtype B virus. Although the overall frequency of mutations at position 210 was similar for CRF01_AE and subtype B viruses, L210F was significantly more common in CRF01_AE virus (P < 0.001), and L210W tended toward being more common in subtype B virus (P = 0.060).
Drug Resistance Mutations in NNRTI Treatment-Exposed Patients
Sequences were available for 21 patients infected with CRF01_AE virus who had been exposed to treatment with ≥1 NNRTIs. Frequencies of mutations at known sites of NNRTI drug resistance are shown in Table 3, with subtype B virus isolates from the Stanford database shown for comparison. Frequencies of mutations K101E, V106M, V179I/D, and G190A/S/E in CRF01_AE virus were higher than frequencies of those in subtype B virus. There were no mutations that were significantly more common in subtype B virus than in CRF01_AE virus.
Drug Resistance Mutations in PI Treatment-Exposed Patients
Sequences were available for 34 CRF01_AE virus-infected patients who had been exposed to ≥1 PIs. Frequencies of mutations at known sites of PI drug resistance are shown in Table 4, with subtype B virus isolates from the Stanford database shown for comparison. There was considerable variability between CRF01_AE and subtype B viruses in the frequency of mutations at many positions. There was a significantly higher frequency of mutations in CRF01_AE virus than in subtype B virus at positions K20I, L33F, M36I, and G48V. The M36I mutation was also seen with high frequency in treatment-naive patients infected with CRF01_AE virus and therefore represents a polymorphism. Frequencies of mutations at positions M46I/L, L63P, A71V, V77I, V82A, and I93L were significantly lower in CRF01_AE virus than in subtype B virus.
Mutations at Non-Drug Resistance Sites
At positions not conventionally associated with drug resistance, mutations occurred significantly more often in NRTI treatment-exposed patients infected with CRF01_AE virus than in treatment-exposed patients infected with subtype B virus at positions 6, 11, 35, 39, 43, 122, 123, 173, 174, 177, 178, 200, 207, 211, 238, 245, 272, 286, 291, 292, 293, 312, 326, 329, 335, 357, 359, 366, 371, 390, and 395 of RT. In PI-treated patients infected with CRF01_AE virus, mutations were more common at positions 13, 16, 35, 36, 41, 69, 70, and 89. All of these mutations were seen with high frequency in treatment-naive patients infected with CRF01_AE virus and, therefore, likely represent polymorphisms rather than treatment-related mutations.
Two mutations were found that were significantly more common in PI-treated patients infected with CRF01_AE virus than in PI-treated patients infected with subtype B virus but were not present in treatment-naive patients infected with CRF01_AE virus. Seven treated patients infected with CRF01_AE virus (20%) had mutations at position 74, with an A substitution in 2 (6%) and an S substitution in 5 (14%). In treated patients infected with subtype B virus, these occurred at a frequency of 1% and 5%, respectively. Six of the 7 patients had received ritonavir as part of their treatment (usually with saquinavir). Four patients (12%) had an N83D mutation that occurred in 0.3% of patients infected with subtype B virus. Three of these patients had been treated with ritonavir together with saquinavir.
We found that CRF01_AE virus is the predominant HIV subtype in Singapore, representing a shift from the earlier years of the epidemic when subtype B virus was more common. This is consistent with epidemiological data showing a change from homosexual transmission to heterosexual transmission through contact with high-risk sexual partners in Thailand and other countries in Southeast Asia where CRF01_AE virus is common.5,6 We found only a few isolated drug resistance mutations in the treatment-naive group and none that would result in clinically important drug resistance. This contrasts with observations in some Western countries, where the prevalence of significant resistance mutations among treatment-naive patients can approach 10%.7 The treatment-naive patients had resistance testing performed at the time of first presentation, but this was in most cases likely to be many years after HIV infection was acquired. Therefore, any transmitted drug-resistant strains may have been replaced with wild-type virus. The absence of resistance in treatment-naive patients may also reflect the low level of antiretroviral use in the population at the time that these patients likely became infected. NRTIs have been available in Singapore since the early 1990s, and PIs and NNRTIs have been available since 1997; however, financial considerations have limited the uptake of highly active antiretroviral therapy in Singapore and other countries in Southeast Asia. This situation may change as therapy becomes more widely used in the region, and there is already data from some sites in Asia indicating an emerging problem with resistance.8,9
The few mutations that were observed occurred at secondary resistance sites and likely represent polymorphisms rather than transmitted treatment-related resistance mutations. The pattern of mutations was consistent with findings of previous studies of treatment-naive patients infected with CRF01_AE virus and other non-subtype B viruses.8,10 In treatment-naive patients, the NNRTI mutation at position V179I was seen in 20%, and the V106I mutation was seen in 9%. The V179I mutation is common to subtype A virus and is also seen in ∼6% of treatment-naive patients infected with subtype C virus.11,12 The M36I polymorphism, which is considered a minor resistance mutation to PIs, was present in 100% of our isolates from treatment-naive patients. This polymorphism has also been found to be common in previous studies of isolates from treatment-naive patients infected with non-subtype B viruses.8,13,14
Our study, which to our knowledge is the largest to date of treatment-related mutations in CRF01_AE virus, showed that the overall pattern and frequency of NRTI resistance mutations were similar to those for subtype B virus, although there were differences in the frequency of several thymidine analogue mutations (D67N was more common in CRF01_AE virus, and M41L and T215F/Y were less common in CRF01_AE virus). These differences may result in part from some differences in the local pattern of nucleoside use, although the combinations of NRTIs used in Singapore resemble Western practice and several key mutations (M184V, K70R, and K219Q/R) occurred with almost identical frequency in the two subtypes. The differences in frequency may also represent an element of biologic variation in the pathway to development of resistance. Although some studies have concluded that there are no subtype-specific differences15, a number of studies have suggested that there may be differences in mutation frequency between subtypes. A study of CRF01_AE virus in Japan found differences in mutations from those in subtype B virus (T69N and V75M were more common in CRF01_AE virus) that were not completely accounted for by differences in drug use.16 One mutation seen in our study that cannot be explained by differences in drug use is the L210F mutation, which occurred in 9% of CRF01_AE virus-infected patients but is rare in subtype B virus. The relative impact on resistance of the F mutation compared with the 210W mutation (which tended to occur more commonly in subtype B virus) is not known. A study comparing subtype F1 virus with subtype B virus in patients taking similar nucleoside analogue regimens in Brazil showed a significantly lower prevalence of the L210W mutation in subtype F1 virus. This was explained by differences in genetic polymorphisms at this site creating a higher genetic barrier to mutation in subtype F1 virus.17 A comparison of subtype C virus isolates and subtype B virus isolates from patients receiving similar antiretroviral therapy in Israel showed a significantly lower prevalence of M41L, D67N, K70R, L210W, and T215Y mutations in subtype C virus, which is consistent with some of our observations of the sites of intersubtype variability.13
The most common mutation seen in our NNRTI-treated patients infected with CRF01_AE virus was K103N, which is also the dominant mutation in NNRTI-treated patients infected with subtype B or C virus.12 K103N was also found to be the most common mutation to develop in women infected with CRF01_AE virus who were treated with a single dose of nevirapine in pregnancy.18 The V106M mutation that was seen in 3 patients infected with CRF01_AE virus (14%) is of particular note, because it does not occur in subtype B virus and has been described previously only in clinical subtype C virus isolates.12 This mutation confers high-level resistance to both efavirenz and nevirapine, whereas the 106A mutation (which is seen in subtype B virus) confers high-level resistance to nevirapine but low-level resistance to efavirenz.19 All 3 of our patients who developed the V106M mutation had been treated with efavirenz, and a strong association with this drug was noted in subtype C virus isolates with this mutation.12 Furthermore, this mutation was not seen in a group of Thai women exposed to intrapartum nevirapine.18 The K101E and G190A mutations that were more common in CRF01_AE virus are both associated with resistance to efavirenz and nevirapine.20 The V179D mutation contributes to NNRTI resistance.3
We also observed a number of differences in the frequency of mutations between CRF01_AE virus and subtype B virus in the protease region. Protease mutations are classified as "major" (tend to occur early and by themselves have a major effect on resistance) or "minor" (tend to occur later and do not have a major effect by themselves on phenotype).4 Of the treatment-related mutations that were significantly more common in CRF01_AE virus, only G48V is a major mutation. This is selected by saquinavir use and is associated with resistance to saquinavir and with lesser degrees of resistance to other PIs. A higher frequency of the G48V mutation was not seen in a smaller study of CRF01_AE virus isolates in Japan, which instead found that L10F, K20I, L33I, and N88S mutations were seen more commonly in CRF01_AE virus infection than in subtype B virus infection.16 Our study only confirmed that the K20I mutation and the F mutation at L33 were more common in CRF01_AE virus. The major protease mutation V82A was less common in CRF01_AE virus than in subtype B virus. This mutation is selected by indinavir and ritonavir and confers high-level resistance to other PIs. Subtype-specific variation has previously been noted at position 82, with the V82I mutation noted to be more common in subtype C virus than expected for subtype B virus.21 A significantly lower propensity to develop the D30N mutation has been demonstrated in subtype C virus compared with subtype B virus.22 None of our isolates had a D30N mutation, although this may reflect the fact that nelfinavir was not commonly used in our population of patients. These variations in mutation frequency may result from variation in the pattern of drug use in Singapore in comparison with that in Western countries. For example, for economical reasons, the combination of ritonavir and saquinavir is used often either alone as dual therapy or in combination with a single nucleoside, and lopinavir is rarely used. It is possible that the differences in frequency of mutations result, at least in part, from biologic differences in the way the virus evolves in the face of drug pressure, as has been shown in studies of other viral subtypes.22
In addition to mutations at known sites of resistance, we identified 2 treatment-related mutations in the protease at T74A/S and N83D that were present at higher levels in CRF01_AE virus than in subtype B virus. A high rate of mutation at position 74 has previously been noted in both subtype B virus and subtype C virus isolates from treated patients, although the significance is unclear.21 Further investigation is necessary to determine the impact of these mutations on drug resistance.
In conclusion, the pattern of drug resistance mutations that are selected by antiretroviral therapy in patients infected with CRF01_AE virus is similar to that of drug resistance mutations in subtype B virus. However, there are a few mutations that appear to differ in frequency, and this finding, taken together with findings of studies on other subtypes, supports the existence of subtype-specific mutational pathways that may in a few cases result in clinically relevant differences in the development of drug resistance. The differences in mutational pathways may result from pretreatment differences in nucleotide substitutions at sites of resistance that may be silent or may manifest as amino acid polymorphisms.11,17 It is important to understand these subtype-specific differences, given the impending scale-up of antiretroviral therapy in regions such as Southeast Asia where non-subtype B virus predominates.
The authors thank Dr. Bernard Peperstraete, Ms. Liang Fang Teo, and Maragaret Lee of the Infectious Disease Research Centre, Tan Tock Seng Hospital, Singapore, and Mr. Martin Tuohy of Virco, Mechelen, Belgium, for administrative support.
1. Kantor R, Katzenstein D. Drug resistance in non-subtype B HIV-1. J Clin Virol.
2. Osmanov S, Pattou C, Walker N, et al. Estimated global distribution and regional spread of HIV-1 genetic subtypes in the year 2000. J Acquir Immune Defic Syndr.
3. Stanford HIV Drug Resistance Database. Vol 2004.
4. Johnson VA, Brun-Vezinet F, Clotet B, et al. Drug resistance mutations in HIV-1. Top HIV Med.
5. Kalish ML, Korber BT, Pillai S, et al. The sequential introduction of HIV-1 subtype B and CRF01AE in Singapore by sexual transmission: accelerated V3 region evolution in a subpopulation of Asian CRF01 viruses. Virology.
6. Chew SK. Trends in human immunodeficiency virus infection: epidemiology in Singapore. Ann Acad Med Singapore.
7. Weinstock HS, Zaidi I, Heneine W, et al. The epidemiology of antiretroviral drug resistance among drug-naive HIV-1-infected persons in 10 US cities. J Infect Dis.
8. Lan NT, Recordon-Pinson P, Hung PV, et al. HIV type 1 isolates from 200 untreated individuals in Ho Chi Minh City (Vietnam): ANRS 1257 Study. Large predominance of CRF01_AE and presence of major resistance mutations to antiretroviral drugs. AIDS Res Hum Retroviruses.
9. Zhong P, Kang L, Pan Q, et al. Identification and distribution of HIV type 1 genetic diversity and protease inhibitor resistance-associated mutations in Shanghai, P. R. China. J Acquir Immune Defic Syndr.
10. Brindeiro PA, Brindeiro RM, Mortensen C, et al. Testing genotypic and phenotypic resistance in human immunodeficiency virus type 1 isolates of clade B and other clades from children failing antiretroviral therapy. J Clin Microbiol.
11. Turner D, Brenner B, Moisi D, et al. Nucleotide and amino acid polymorphisms at drug resistance sites in non-B-subtype variants of human immunodeficiency virus type 1. Antimicrob Agents Chemother.
12. Grossman Z, Istomin V, Averbuch D, et al. Genetic variation at NNRTI resistance-associated positions in patients infected with HIV-1 subtype C. AIDS.
13. Grossman Z, Vardinon N, Chemtob D, et al. Genotypic variation of HIV-1 reverse transcriptase and protease: comparative analysis of clade C and clade B. AIDS.
14. Pieniazek D, Rayfield M, Hu DJ, et al. Protease sequences from HIV-1 group M subtypes A-H reveal distinct amino acid mutation patterns associated with protease resistance in protease inhibitor-naive individuals worldwide. HIV Variant Working Group. AIDS.
15. Pillay D, Walker AS, Gibb DM, et al. Impact of human immunodeficiency virus type 1 subtypes on virologic response and emergence of drug resistance among children in the Paediatric European Network for Treatment of AIDS (PENTA) 5 Trial. J Infect Dis.
16. Ariyoshi K, Matsuda M, Miura H, et al. Patterns of point mutations associated with antiretroviral drug treatment failure in CRF01_AE (subtype E) infection differ from subtype B infection. J Acquir Immune Defic Syndr.
17. Dumans AT, Soares MA, Machado ES, et al. Synonymous genetic polymorphisms within Brazilian human immunodeficiency virus type 1 subtypes may influence mutational routes to drug resistance. J Infect Dis.
18. Jourdain G, Ngo-Giang-Huong N, Le Coeur S, et al. Intrapartum exposure to nevirapine and subsequent maternal responses to nevirapine-based antiretroviral therapy. N Engl J Med.
19. Brenner B, Turner D, Oliveira M, et al. AV106M mutation in HIV-1 clade C viruses exposed to efavirenz confers cross-resistance to non-nucleoside reverse transcriptase inhibitors. AIDS.
20. Huang W, Gamarnik A, Limoli K, et al. Amino acid substitutions at position 190 of human immunodeficiency virus type 1 reverse transcriptase increase susceptibility to delavirdine and impair virus replication. J Virol.
21. Kantor R, Zijenah LS, Shafer RW, et al. HIV-1 subtype C reverse transcriptase and protease genotypes in Zimbabwean patients failing antiretroviral therapy. AIDS Res Hum Retroviruses.
22. Grossman Z, Paxinos EE, Averbuch D, et al. Mutation D30N is not preferentially selected by human immunodeficiency virus type 1 subtype C in the development of resistance to nelfinavir. Antimicrob Agents Chemother.
© 2005 Lippincott Williams & Wilkins, Inc.