ABSTRACT
This work reports the variability of human immunodeficiency virus type 1 (HIV-1) protease from treated and untreated patients infected with HIV-1 subtype C in the United Kingdom. The most common primary mutation observed in treated patients was L90M. D30N, M46I, V82A/F, and I84V were seen rarely. M36I and I93L mutations were observed in nearly all samples from both treated and untreated patients and so cannot be considered as resistance-associated mutations in this subtype.
Human immunodeficiency virus type 1 (HIV-1)-encoded protease (PR) is a key target of antiretroviral therapy. There are currently six protease inhibitors (PIs) licensed for use in clinical practice, namely, ritonavir (RTV), saquinavir (SQV), indinavir (IDV), amprenavir, nelfinavir (NFV), and lopinavir (LPV). A major limitation of long-term effectiveness of antiretroviral drugs is the development of resistance: it has been found that resistance to PIs is mediated by a variety of primary mutations, including D30N, M46I, G48V, I50V, V82A/F/T/S, I84V/A, and L90M (4). These primary mutations are often combined with accessory or secondary mutations that are probably mainly compensatory and which alone have little effect on drug resistance. However, accumulation of accessory mutations can increase the level of phenotypic resistance (16).
PR shows a high degree of polymorphism, with some accessory mutations occurring commonly in untreated patients (6). All published data on mutations associated with resistance to PIs have been derived from studies of subtype B HIV-1, the most prevalent subtype in the United States and Europe, where therapy is most often applied. However, worldwide the most prevalent subtypes are non-B, with subtype C being particularly common in Africa and Asia (5). Also, there is an increasing prevalence of non-B virus in the developed world, associated primarily with heterosexual transmission (1, 3, 14). Consequently, there is an increasing requirement for resistance data from the non-B subtypes. Reports on the characterization of non-B PR from untreated patients indicate that some of the accessory mutations occur frequently in untreated patients (8, 11, 13), although there is no evidence that these mutations compromise PI activity per se (11).
This report describes the variability of PR from treated and untreated patients infected with HIV-1 subtype C in the United Kingdom.
The samples analyzed were submitted for routine HIV-1 genotypic resistance testing. The reasons for the test included therapy failure, pretreatment analysis, and seroconversion. Treatment information was provided in most cases. HIV-1 RNA was extracted from plasma and the PR coding region was amplified by nested reverse transcriptase (RT-PCR) as previously described (9). Sequencing of PCR products was undertaken using either ABI 377 or Beckman CEQ2000 protocols. Sequences were analyzed and subtyped using the database maintained at Stanford University (12).
Sixty-two samples obtained from 58 patients provided complete PR sequences which were designated as subtype C on the basis of the PR sequences only. Subtype was confirmed by analysis of gag and/or env regions of the genome for 15 samples (2). Information on therapy was available for 60 samples. These included samples from 30 patients who had experienced PI therapy, of whom 10 were not receiving a PI at the time of sampling. For those receiving PI treatment at the time of sampling, the median number of PIs received in total was two. The PIs used in this study included NFV, SQV, IDV, and RTV. We also studied 25 patients who had received no PIs, of whom 12 had received no antiretroviral therapy at all. Four pairs of samples were repeats from the same patients: in the following analyses of prevalence of mutations only the later samples from pairs from the same patients are considered.
Primary mutations.The sequences were analyzed for the presence of primary mutations associated with resistance to PIs. These mutations include D30N, M46I, G48V, I50V, V82A/T/S/F, I84V/A, and L90M (4). Thirteen of 20 patients receiving PIs at the time of sample showed at least one of these primary mutations. The PI treatment histories of those patient showing primary PR mutations and the combinations of mutations observed are shown on Table1. The prevalence of these primary mutations in treated and untreated patients is shown in Table2. The most common primary mutation observed in PI-treated patients was L90M (observed in 13 of 30 treated patients, compared with 0 of 25 untreated patients). Of note, these patients included two who were SQV and NFV naı̈ve but had received IDV. The M46I mutation was seen in four patients, while V82A/F and I84V mutations were each observed in two treated patients, but these last two mutations were only seen in association with the L90M mutation. The prevalence of the L90M mutation is similar to that observed in patients infected with subtype B who failed combination therapy, but the prevalence of the V82A/T/F mutation (<10%) is much lower (15).
Summary of PI treatment histories for those samples showing primary mutations associated with PI resistance
Prevalence of primary mutations associated with PI resistance in patients infected with subtype C HIV-1
The D30N mutation was observed in only one patient, together with the N88D mutation, although NFV was widely used in this patient group. Experience with subtype B HIV-1-infected patients indicates that about 50% of patients receiving NFV in the short term fail therapy due to the development of the D30N mutation (7). It is possible that subtype C-specific genotypic determinants within PR may influence the route to drug resistance for this particular drug. Failure of NFV therapy predominantly due to the L90M mutation in this subtype might therefore compromise the use of other PIs.
G48V and I50V mutations were not observed in this study. No primary mutations associated with PI resistance were observed in the untreated patients. Of the 7 out of 20 patients receiving PI treatment who showed no primary PR mutations, 4 showed mutations in RT (data not shown); 2 showed no mutations in RT, indicating failure for a reason other than drug resistance; and no RT data were available for the final patient. Thus, failure of therapy including PIs in these subtype C HIV-1-infected patients showed an incidence of PI resistance similar to that observed in subtype B HIV-1-infected patients (10). Primary mutations associated with resistance to PIs were observed in three samples from patients who were no longer being treated with PIs at the time the sample was taken, indicating that rapid loss of these mutations in the absence of selective pressure does not occur. In one case, the period of no PI treatment was 9 months, but this information was not available for the other samples.
Secondary mutations.The prevalence of secondary mutations observed in the samples from patients with known treatment histories is summarized in Table 3. M36I/L and I93L mutations, which have been associated with resistance to PIs in subtype B, were present in samples from all patients, except one which did not show the M36 change, regardless of treatment status, indicating that within subtype C these represent the consensus. Mutations at codons 10, 33, 54, 73, and 88 were observed only in treated patients, but the numbers were too small to reach statistical significance (Table 3). The only secondary mutation that showed a statistically significant change in prevalence between treated and untreated patients was at codon 63 (P = 0.006) , where the proportion of mutants increased from 44% in untreated to 80% in treated patients (Table 3), levels similar to those observed with subtype B. The V77I mutation was observed in only six (11%) patients (treated or untreated) in contrast to subtype B, where it is seen in 20% of untreated patients, rising to 36% in multiple-PI-experienced patients (12). No other change relative to the subtype B consensus was observed to be more prevalent in treated than in untreated patients. Changes at codon 74 (T74A/S) usually occur in subtype B virus only in patients receiving treatment with PIs (12), though the effects of these mutations on resistance have not been examined. However, in these subtype C samples the T74S/A mutation was observed in both treated and untreated patients. Likewise, the D60E mutation was seen in both treated and untreated patients. Of interest, two samples came from infant twins, only one of whom had been treated with a PI (NFV), and the two samples differed only in that the sample from the PI-treated child showed the T74S mutation. In summary, this report demonstrates that the key primary mutation associated with resistance to PIs in subtype C HIV-1 is L90M. Some secondary mutations identified as significant in subtype B virus may be of lesser significance in subtype C. The effect of these mutations on the phenotypic profile of these viruses remains to be elucidated.
Prevalence of secondary mutations associated with PI resistance in patients infected with subtype C HIV-1
ACKNOWLEDGMENTS
We thank the clinicians (all from the United Kingdom), who have provided clinical information with respect to the samples described in this paper, namely, D. Natin, Stratford; S. Drake, Birmingham; M. McKendrick, Sheffield; J. Arumainagagan, Walsall; C. Bignell, Nottingham; J. Bendig, Epsom; P. Allan, Coventry; A. Palfreyman, Peterborough; A. Kilvert, Northhampton; M. McBride, Belfast; G. Downie, Kilmarnock; D. Carrington, Bristol; J. Mok, Edinburgh; G. Scott, Edinburgh; and S. O'Connell, Southampton. We also thank Judith Workman, Daina Ratcliffe, Clare Overton, and Susan Jackson for technical assistance.
FOOTNOTES
- Received 8 December 2000.
- Returned for modification 21 January 2001.
- Accepted 20 April 2001.
- Copyright © 2001 American Society for Microbiology
