The results of work with protease inhibitors PIs indicate that the D30N mutation was not observed in CRF02_AG and CRF02_AE isolates in patients failing nelfinavir (NFV) therapy but rather that the N88S mutation emerged after NFV use in CRF01_AE and after indinavir [51
] use in subtype B [52
]. Although another study reported an absence of the D30N mutation in CRF01_AE, no information on the specific type of PIs received by the patients was available [54
]. A lower frequency of D30N was seen in subtype C isolates from Ethiopian immigrants to Israel after NFV usage than in subtype C viruses from Botswana [55
], suggesting that subtype C viruses from Ethiopia (the origin of the samples identified in Israel) and southern Africa might behave in different fashion. M89I/V mutations were observed in F, G, and C subtypes but not in other subtypes [26
], and the V82I natural polymorphism in subtype G led to the emergence of I82M/T/S in treatment failure [57
]. The L90M mutation is rare in subtype F but common in subtype B from Brazil [58
], and a recent paper suggests that polymorphisms at position 36 in PR may be important in determining the emergence of specific patterns of resistance mutations among viruses of different subtypes [59
To gain an understanding of the underlying mechanisms leading to the overall higher preponderance of D30N in subtype B relative to other subtypes, molecular dynamic simulations were performed. D30N appeared to selectively confer resistance to NFV in subtype B by increasing the flexibility of the protease (PR) flap region and destabilizing the PR inhibitor complex [60
]. In subtype C, D30N required the accessory N83T mutation to confer resistance and rescue fitness [61
Two comprehensive surveys reported differences in natural protease polymorphisms among non-B subtypes [62
] and positions less frequently mutated in non-B subtypes than in subtype B after exposure to ARVs. Residues of importance in subtype A in PR were at positions 10, 20, and 63, whereas, in subtype C, they were at residues 20, 53, 63, 74, and 82. Other differences were at residues 13 and 20 in subtype D, residues 10, 14, 20, and 77 in subtype F, residues 20, 67, 73, 82, and 88 in subtype G, residues 20, 63, 82, and 89 in CRF01_AE, and residue 20 in CRF02_AG [63
Higher rates of accumulation of NRTI and PI resistance mutations and equal rates of emergence of NNRTI mutations were also found in subtype B compared to C [64
]. A study from southern Brazil also showed a lower frequency of primary resistance to PIs in subtype C compared to subtype B, suggesting that PI mutations may be less well tolerated at the structural level in subtype C [65
However, HIV-1 subtype diversity has not limited the overall benefit of ART (). This notwithstanding there are subtype differences in the type and preference of pathways of resistance with some mutations emerging almost exclusively in some non-B subtypes, for example, the protease mutation 82
M in subtype G versus 82A/F/S in the others, 88D in subtype B versus 88S in subtypes C and CRF02_AG [66
]. Furthermore, HIV-2 has major mutations in regard to NRTIs, NNRTIs, and PIs, which contribute to innate NNRTI resistance and rapid development of multiclass drug resistance () [67
]. The V106M RT mutation in subtypes C and A versus V106A in subtype B is observed with resistance against NVP and EFV. Polymorphisms at RT residue 98, common in subtype G, are associated with NNRTI resistance in subtype B and may lower the resistance barrier and duration of efficacy of some NNRTIs [69
]. The frequency of some resistance mutations shared by B and non-B subtypes can vary after failure of first-line therapeutic regimens, as in the case of the K65R mutation. Differences in type and frequency of resistance mutations should not be underestimated. However, the TAM pathway 67N/70R/215Y found in subtype C in Botswana will probably be adequately detected by most resistance algorithms, since it does not involve new mutations.
A lower risk for accumulation of major (primary) resistance mutations in subtype C than B has been reported [64
]. The major mutations that emerged in both subtypes were the same. Since both subtypes B and C patients had similar profiles of virological failure after use of the same ART regimens, this rules out ancillary factors responsible for these differences. Minor mutations in subtype B PR may appear as frequent natural polymorphisms in several non-B subtypes (e.g., M36I, L89M) [58
]. The fact that the L89M polymorphism can lead to the M89I mutation that confers resistance to PIs suggests that there might be a lower accumulation of major mutations in C subtypes, if natural polymorphisms act similarly in subtype C as they do when present as secondary resistance mutations in subtype B.
The majority of non-B HIV-1 subtype isolates possess wild-type susceptibilities similar to those of subtype B wild-type isolates. Compared to B subtypes, diminished susceptibilities among wild-type isolates have been found for CRF02_AG recombinant viruses in three different studies in regard to ATV and NFV [63
]. No study has yet assigned statistical significance of drug susceptibility levels due to polymorphisms and small sample size. One analysis performed molecular modeling and suggested that distortions in the K26 pocket of A/G proteases appear to be responsible for a lower binding energy of NFV and hence lower susceptibility of A/G viruses to this drug [70
]. A/G isolates with lower susceptibilities to certain PIs (NFV and atazanavir (ATV)) have also been found. One study has detected an important proportion of WT isolates with lower susceptibilities to ATV [71
]. In most cases, phenotypes have been determined by commercial or in-house assays that were developed primarily to measure B-subtype drug susceptibilities based on the laboratory adapted strains NL4-3 or HXB2, through use of a modified clone of a laboratory strain that lacks both the terminal part of Gag and most of Pol. It should be recognized that most commercial assays do not monitor polymorphisms, and indeed sequences that lie within particular regions, such as the substrates of PR within gag or the RNaseH and connection domains within pol, can influence drug resistance in both B and non-B subtypes but may not be easily recognized. Although some work has been carried out in this field, it is clear that other studies are required [72
There are few data on the potential for cross-resistance to PIs among non-B subtypes in regard to NFV, although there is a tendency to select for the L90M pathway instead of D30N in subtype C. Competition fitness assays support the notion that subtype C viruses bearing D30N are impaired in replicative fitness, a finding that may explain the above results [61
Thermodynamic studies performed on target-inhibitor interactions in PR have specifically described a lower affinity of non-B subtype proteases for PIs and amplification of primary resistance mutations on the basis of polymorphisms that are present in background.
In addition to the foregoing, interesting results on polymorphisms that confer hypersusceptibility to some PIs have been recently reported [76
]. Some of these polymorphisms can potentially delay acquisition of drug resistance and may therefore enhance the long-term effectiveness of relevant drugs.