Drug Resistance Mutations in HIV pol Sequences from Argentinean Patients Under Antiretroviral Treatment: Subtype,Gender,and Age Issues

Discussion

The patterns and frequencies of the DRMs described (Fig. 1) are consistent with present treatment strategies. The World Health Organization (WHO) recommends AZT or TDF plus 3TC together with EFV or NVP as a first-line antiretroviral regimen. ² The highly frequent M184V substitution is the first mutation that appears under 3TC or 3TC-containing regimens, resulting in complete resistance to 3TC. Mutations M41L, T215Y, D67N, L210W, and K70R are selected by thymidine analogs such as AZT, which has been heavily used in our country. These mutations belong to a group of substitutions known as thymidine analog mutations or TAMs, which decrease susceptibility to almost all nucleoside and nucleotide analogs. Furthermore, N348I and V118I are accessory mutations that, in combination with TAMs, also reduce susceptibility to most NRTIs. ⁸

Analogously, K103N and Y181C are the mutations most frequently selected by EFV and NVP, respectively. K103N confers high resistance to all first-generation NNRTIs. ⁹ K103N has the potential to persist for years, ¹⁰ and both K103N and Y181C have been shown to be present as minor variants prior to antiretroviral treatment and are correlated with the risk of virologic failure. ⁹ Furthermore, these two mutations, together with G190A, are known to be common in transmitted drug resistance. ¹¹ The WHO also recommends the inclusion of PIs in second-line regimens ² . DRMs L10I, I54V, A71V, and M46I are known to accumulate during failure of therapy with most PIs, causing gradual increases of resistance levels. ¹² Mutation L90M confers resistance to several Pis, ⁸ whereas V82A is selected by ritonavir and produces failure of therapy with most PIs. ⁸ Mutation L10V and other mutations at position 10 compensate for the loss of fitness associated with the major PI resistance mutations. ⁸ Interestingly, this mutation has been observed in antiretroviral-naive patients in three independent studies, ^{13 –15} which, together with the data described here, suggests that transmitted drug resistance could be important in our country.

Viral genotype was a good predictor of the frequencies of different DRMs (Tables 1, 2, and 3). These are not unexpected results, as it has been shown that non-B HIV-1 subtypes can present resistance profiles that differ from those observed in subtype B viruses, a fact that is attributed to differences in the number of mutations needed by viruses with different genetic backgrounds to achieve resistance (the so-called genetic barrier), which differs among viral subtypes. ^16,17 The high frequency of mutation T74S is unexpected and interesting, as it is rare among non-C subtype viruses. ⁸ This mutation was particularly frequent among the BF sequences studied here (Table 1), suggesting that recombination could drive the emergence of otherwise rare resistance mutations.

The presence of multiple DRMs was fairly common among the sequences studied here (Table 4). This condition is related basically to two facts: first, high level resistance and resistance to multiple antiretrovirals necessitate the accumulation of multiple amino acid substitutions; second, as most resistance mutations impair viral replication, the emergence of compensatory mutations that attenuate fitness loss usually follows the appearance of primary mutations. One of most remarkable features of our dataset was the unequal distribution of NRTI resistance mutations among adult and children, a situation that was also observed, though to a lesser extent, for PI resistance mutations (Table 4). A factor that could be responsible for this situation is vertical transmission, which results in the presence of primary resistance mutations to which more mutations may be readily added during suboptimal antiretroviral treatment. Another one is adherence, which is related to virological response and thus is crucial for the development of antiretroviral resistance. This is specially challenging in infants and children because adherence is frequently jeopardized by factors such as psychological issues, lack of pediatric formulations, poor palatability, high pill burden or liquid volume, frequent dosing requirements, and side effects. Thus, we think that PI and NRTI resistance mutations could have more opportunities to accumulate in children than in adults due to adherence issues. This is consistent with the fact that the number of NNRTI resistance mutations was similar among adult and children, as there is a genetic barrier of one or two mutations to NNRTI resistance, and thus the accumulation of further mutations would have minimal or no effect on antiretroviral resistance. ⁸

In general, sequences from men and women displayed different amounts of DRMs (Tables 5 and 6). In the case of NVP resistance ones, which were more prevalent among subtype B strains from adult women, a very plausible explanation is the use of this antiretroviral for preventing mother-to-child transmission. A single mutation can confer resistance to NVP. Furthermore, it has a long half-life, and drug levels persist for weeks after women receive single-dose NVP, a situation that constitutes the ideal scenario for the emergence of resistant viral variants. That is why this antiretroviral readily selects for NNRTI resistance in postpartum women and infants where transmission does occur. Thus, our results reinforce the idea that in order to decrease the incidence of NVP resistance, combinations and longer antiretroviral prophylaxis regimens should be preferred to single-dose NVP ones. ¹⁸ Also, these observations support current guidelines recommending PI-based, non-NVP-containing regimens for infants who do become infected despite single-dose NVP or extended NVP prophylaxis. ¹⁹ The fact that only subtype B sequences displayed statistically significant differences regarding the degree of NVP resistance in women compared to men points out that differences in the viral genetic background could be important when deciding which are the better treatment options.

The degrees of resistance to many PIs and NRTIs were higher among strains from men than among strains from women (Tables 5 and 6). Although we did not expect to observe such marked differences among sequences from men and women, this is not the first time that this situation has been observed, as a previous investigation performed in Puerto Rico also showed that the average number of resistance mutations was higher among viruses from men than among viruses from women. ²⁰ So far, some investigations have shown that gender could influence biological issues such as the characteristics of transmitted virus population, ^21,22 the levels of HIV RNA, ^{23 –25} and HIV-specific CD8⁺ T cell response. ²⁶ Nevertheless, these and other similar reports have led to intense debates without reaching a consensus on whether gender can determine the characteristics of HIV infection.

Although we do not completely discard this last possibility, we think that lurking epidemiological variables could better explain the differences in the number of PI and NRTI resistance mutations observed among viruses from men and women. For example, previous studies have shown that in Argentina, the incidence of HIV-1 genotypes varies among vulnerable groups such as drug users, sex workers, men who have sex with men, and heterosexuals, ^27,28 suggesting that viral populations circulating in Argentina are structured. Also, it is known that men are less adherent to treatment than women, which, as mentioned above, favors the nonsuppressive conditions that allow the emergence of resistant variants. Regardless of the preferred explanation for the gender-related differences observed here, our results indicate that there is a strong need for further epidemiological studies to understand the antiretroviral resistance epidemics in Argentina. Ideally, these studies must integrate large amounts of sequence data together with risk groups and demographic and clinical information.