To our knowledge, this study provides the most comprehensive information on HIV–1 drug resistance-associated natural polymorphisms and the acquired DRM profile of CRF01_AE in China to date using a large dataset from a long-term ART cohort in Shenyang.CRF01_AE isolates in this study demonstrated high levels of polymorphisms at both DRM sites and other sites, with several lineage-specific characteristics. However, the little variation in polymorphisms between TF and TS patients implied little impact of CRF01_AE polymorphisms on the development of DRMs. Moreover, this study found that the most common NRTIs and NNRTIs-associated DRMs among CRF01_AE patients who experienced TF were similar to those DRMs among subtype B patients. However, several potential new DRMs selected during ART might be CRF01 _AE-specific DRMs. Among these DRMs, L228R occurred simultaneously or following the appearance of Y181C, and it might be an accessory mutation to Y181C.
In this study, 31 CRF01_AE-specific polymorphism sites (including five known DRMs) were detected, which is even higher than the number of polymorphism sites in CRF01_AE strains mainly from Southeast Asia in a global study on non-B HIV–1 over 10 years ago [25]. Moreover, the two CRF01_AE lineages in this study corresponded to the two lineages epidemic mainly among men who have sex with men (MSM) in China [26, 27], and they demonstrated both common and lineage-specific polymorphisms. A similar situation also occurred for gag and env regions of different lineages of CRF01_AE [28], suggesting that other lineages of CRF01_AE in China might have distinct polymorphisms, which might further complicate DRM development and drug resistance genotype interpretation [29].
In this study, the polymorphisms at five known drug resistance-associated sites (V179I/D, V118I, K103R K238R, and E40Q) were polymorphic accessory mutations or other mutations that did not independently decrease drug sensitivity. None of them were associated with TF. Only V75L, a low-frequency mutation, was associated with virologic failure, implying that most polymorphisms in CRF01_AE seldom lead to TF. A study from a London cohort found that different baseline polymorphisms, including V90I, A98S, and K103R, were associated with virologic failure [30], but their effects could not be differentiated from the impacts of the different treatment regimens and HIV strains.
At present, two NRTIs plus an integrase strand transfer inhibitor (INSTI) are recommended as a first-line ART regimen for adults in developed countries while, in developing countries like China, two NRTIs plus an NNRTI are still recommended as a first-line ART regimen. In this study, we evaluated the DRM profile of CRF01_AE after TDF/3TC/EFV TF. The most common acquired DRMs among CRF01_AE were K65R, M184V, G190S/C Y181C, and K103R, all of which are also common among subtype B and other subtypes. Despite some differences in the mutation type and rate, we cannot confirm that the differences were caused by the various subtypes [31, 32].
More importantly, we detected two mutations with significant increases but without annotation in the Stanford HIVdb algorithm, V75L and L228R. Site 75 is a drug resistance-associated site but no explanation for V75L is provided in the Stanford HIVdb algorithm. The V75L mutation has been reported to provide a selective advantage by allowing escape from the host immune responses [33] and it is believed to be a TDF-associated mutation [21]. The L228R mutation has been reported to be related to the treatment of non-B HIV–1 subtypes in several studies [34, 35], and its phenotype has not yet been described. In this study, for the first time, it was suggested that L228R might be associated with the known DRM Y181C and it might act as an accessory mutation to Y181C based on a co-variation analysis and longitudinal evolution study. These results implied that more unannotated mutations in non-B HIV–1 during TF might be accessory mutations associated with drug resistance. Therefore, more studies are needed to strengthen the phenotypic research on drug resistance in non-B HIV–1 [36], and to provide more evidence for drug resistance interpretation for non-B HIV–1.
Several limitations exist in the present study. First, due to the high success rate of ART in this cohort, only a small number of TF patients could be included in the acquired DRM analysis. Second, the impact of the L228R mutation with or without Y181C needs further validation using virus growth competition and drug resistance phenotype assays. Nevertheless, this study provided more evidence of polymorphisms and DRMs in the non-B HIV–1 strain CRF01_AE.