icon star paper   HIV Articles  
Back grey_arrow_rt.gif
Long-Term Persistence of Primary Genotypic Resistance After HIV-1 Seroconversion
  JAIDS Journal of Acquired Immune Deficiency Syndromes: Volume 37(5) 15 December 2004
Pao, David MRCP*; Andrady, Ushan FRCP†; Clarke, Janette FRCP‡; Dean, Gillian FRCP*; Drake, Susan FRCP§; Fisher, Martin FRCP*; Green, Tanya MRCP‖; Kumar, Siva FRCP¶; Murphy, Maurice FRCP#; Tang, Alan FRCP**; Taylor, Stephen MRCP§; White, David FRCP§; Underhill, Gillian FRCPath††; Pillay, Deenan FRCPath‡‡; Cane, Patricia PhD§§
From the *Royal Sussex County Hospital, Brighton, United Kingdom; †Ys-byty Gwynedd, Bangor, United Kingdom; ‡Leeds General Infirmary, Leeds, United Kingdom; §Heartlands Hospital, Birmingham, United Kingdom; ‖Pinderfields Hospital, Wakefield, United Kingdom; ¶Queen Elizabeth Hospital, Kings Lynn, United Kingdom; #St. Bartholomew's Hospital, London, United Kingdom; **Royal Berkshire Hospital, Reading, United Kingdom; ††Portsmouth Hospitals National Health Service Trust, Portsmouth, United Kingdom; ‡‡University College, London, United Kingdom; and §§Health Protection Agency Antiviral Susceptibility Reference Unit, Birmingham, United Kingdom.
Primary infection with drug-resistant HIV-1 is well documented. We have followed up patients infected with such viruses to determine the stability of resistance-associated mutations.
Fourteen patients who experienced primary infection with genotypic evidence of resistance were followed for up to 3 years. Drug resistance-associated mutations persisted over time in most patients studied. In particular, M41L, T69N, K103N, and T215 variants within reverse transcriptase (RT) and multidrug resistance demonstrated little reversion to wild-type virus. By contrast, Y181C and K219Q in RT, occurring alone, disappeared within 25 and 9 months, respectively.
Multidrug resistance in 2 patients was found to be stable for up to 18 months, the maximum period studied.
We conclude that certain resistance-associated mutations are highly stable and these data support the recommendation that all new HIV diagnoses in areas where primary resistance may occur should undergo genotyping irrespective of whether the date of seroconversion is known.
This article describes the duration of persistence of drug resistance-associated mutations after transmission. Resistance mutations that arise in wild-type virus after treatment may rapidly become undetectable when treatment is stopped. This may mainly be a result of overgrowth of wild-type virus originating in the viral reservoirs rather than true reversion of the mutant virus per se.
As previously described, virus from patients infected with AZT-resistant virus showing mutations at codon 215 of RT frequently demonstrated onward mutation at this codon resulting in T215S/D/C. These variants were shown in this study to be quite stable. Most other mutations were also found to persist over the period of study. The exceptions were A62V, Y181C, and K219Q when these mutations occurred singly. The disappearance of mutations could be caused by reversion and selection of a fitter virus or, alternatively, by overgrowth of wild-type virus that was present in the original infecting inoculum as a minority population. In the case of Y181C, it is unlikely that this mutation is highly unstable per se, because we have observed this in a new diagnosis of a patient with AIDS (data not shown).
Samples from patients with MDR virus showed no change over the period studied. The slow evolution of MDR virus to fitter wild-type virus may be a consequence of the multiple mutations needing replacement in these viruses. Mutation is a stochastic process, and it is likely that replacement of each site leads to only a small increase in replicative capacity; thus, replacement with fitter virus lacking all mutations is a stepwise process, possibly with a contribution by recombination, and may take a long time. Further, because MDR viruses may already contain fitness-compensating mutations, the route to reversion to wild type may require the virus to travel through a fitness trough.
The observation that some patients infected with MDR virus have unusually low viral loads without treatment has been made before. Salomon et al5 described 2 patients with many resistance mutations that were preserved for 1 and 4 months, but 1 patient had a viral load <50 copies/mL at 4 months, whereas the other had a viral load of 279,360 copies/mL 1 month later. At diagnosis, patients D and L had moderate levels of viral load. A third patient infected with MDR virus, who was not included in this report, had a high viral load on diagnosis. It thus seems likely that these MDR viruses are able to replicate to high levels during the initial phase of infection. It is known that the replicative capacity of MDR virus may be reduced relative to wild-type virus, however, so it is possible that the replicative capacity of MDR viruses may be sufficiently compromised that the immune response is able to exert greater control than that seen in most patients infected with wild-type virus. This illustrates the fine balance between viral fitness and the immune response in controlling the level of viral load in patients. It will be interesting to determine whether these patients infected with MDR virus become long-term nonprogressors.
Thus, we have demonstrated that drug resistance-associated mutations may often persist for a considerable time after primary infection. This observation correlates with the findings that the prevalence of resistance in new diagnoses is similar to that observed in sero-converters. An implication of this finding is to highlight the potential for onward transmission of resistant virus from untreated patients. Therefore, even if most treated patients are virologically suppressed, primary resistance may increase if it is introduced into a community with high risk/rates of onward transmission. It has previously been observed that transmitted resistant virus can be passed on to other individuals.
Finally, the persistence of resistance-associated mutations for prolonged periods after infection underscores the importance of performing a resistance test at baseline, even when evidence of recent infection is not present. Care must be taken with interpretation algorithms, however, particularly with respect to mutations at codon 215, where the classic resistance mutations T215Y and T215F are usually rapidly replaced with variants restoring fitness to the virus. Such variants do not confer resistance per se but are indicators of primary infection with resistant virus, and some interpretation algorithms may not highlight such mutations; thus, there may be an overestimation of potential response to a drug regimen containing AZT or stavudine. It may be that algorithms need some modification for analysis of samples from untreated patients. Testing of patients at diagnosis may result in some underestimation of transmitted resistance because of loss of some mutations with time. Nevertheless, it can certainly provide some benefit for the guidance of treatment options and should be considered for all at-risk patient groups.
The use of antiretroviral therapy (ARV) has dramatically improved the life expectancy of individuals infected with HIV-1. Nevertheless, virus with reduced susceptibility to available drugs emerges in many individuals on therapy; consequently, there is an ever-increasing pool of potential transmitters of resistant virus.
Primary infection with HIV-1 that carries resistance-associated mutations has been reported from many countries. Initially, only extremely low levels of transmitted drug resistance were observed in the 1990s, but these had increased to up to 25% of new cases of primary infection in some populations by 2001. Estimates of primary infection with resistant virus are difficult to compare with each other, however, because of different criteria being used to define resistance. For example, phenotypic testing may underestimate the prevalence of transmitted zidovudine (AZT) resistance because of rapid replacement of the key resistance mutation T215Y with fitter variants (see below). Estimates of transmitted resistance using genotypic tests are also difficult to compare because of differences in weight given to particular mutations and the different algorithms used to interpret genotypic data.
With regard to surveillance, most patients are diagnosed some time after primary infection, and other studies have examined new diagnoses to determine the prevalence of resistance in treatment-naive patients. Variable levels of resistance were found as in the primary infection patients, and, again, comparisons between studies are confounded by inconsistencies in criteria for defining resistance.
It has been well documented that the key AZT resistance mutations, T215Y/F, seldom persist long term after transmission in an untreated patient. Instead, 215D, -S, -N, or -C replaces 215Y/F within a few months, but these mutations may then persist for up to 3 years.13,14 These changes restore fitness to the virus as well as susceptibility to AZT but are more able to revert to 215Y/F quickly under drug pressure because of the shorter mutational distance required. K70R, which also confers reduced susceptibility to AZT, has been observed in transmitted virus and shown to be stable for >1 year. Brenner et al reported variable stability of multidrug-resistant virus (MDR) ranging up to 5 years. Infection with virus that has genotypic resistance has been shown to prolong the period required for virologic suppression. Consequently, it is important to consider the resistance profile of a new infection to optimize initial therapy, as recommended by current guidelines.
The purpose of this study was to monitor the persistence of resistance-associated mutations after transmission in untreated patients. The study included patients with primary infection with virus showing nucleoside reverse transcriptase inhibitor (NRTI) or nonnucleoside reverse transcriptase inhibitor (NNRTI) resistance or multidrug resistance, and the follow-up periods were for up to 3 years.
Study Samples

Patients were identified as having primary HIV-1 infection through having had a negative antibody test <18 months previously, laboratory evidence of acute seroconversion illness such as an evolving antibody response, or the serologic testing algorithm for recent HIV seroconversion (STARHS) along with clinical markers indicating recent infection.18 The STARHS assay used was the BioMerieux Vironostika HIV-1 microELISA system with a cutoff of a standardized optical density of 1.0, which is associated with a window period of 170 days (95% confidence interval: 162-183). Follow-up samples were obtained from 1 to 36 months after diagnosis of primary infection. Details of periods between initial and subsequent samples are shown in Table 1.
Sequence was obtained by in-house methods from the entire protease gene and first 230 codons of reverse transcriptase (RT) as previously described. Briefly, RNA was extracted from pelleted virus, and a nested RT-polymerase chain reaction (PCR) assay was performed. Sequencing was carried out using a CEQ 2000 sequencer (Beckman Coulter), and sequences were edited using Sequencher software (Gene Codes Corporation). Differences from the subtype B consensus sequence were derived using the Stanford database.
Drug resistance-associated mutations were observed in 14 patients with evidence of primary infection for whom follow-up samples were available, with the patients remaining untreated. The mutations observed in the initial and follow-up samples are summarized in Table 1, together with the time intervals between samples. All patients were infected with subtype B virus.
Nucleoside Reverse Transcriptase Mutations
M41L was detected in 3 patients and was still present in the last samples tested from all these patients, which were obtained between 7 and 33 months later. A62V was observed alone in 1 patient. This mutation is usually associated with the multi-NRTI resistance complex based on Q151M. It is unlikely that this mutation alone confers resistance, and its presence may not represent transmitted resistance. Nevertheless, the mutation rapidly disappeared, becoming undetectable within 2 months. T69N was observed in the first sample from 4 patients, including 1 patient with MDR virus. This mutation was still present in all subsequent samples between 15 and 32 months later. One of the patients showed V118I in addition (patient M), which was also unchanged after 16 months.
Codon 215 variants were observed in 5 patients at diagnosis. Patient C had T215L along with M41L, and no change was observed after 7 months. Patient D had T215Y in conjunction with MDR virus, and no change was observed after 17 months. Patient E had T215Y together with M41L, and this was replaced with T215C in a sample taken 21 months later. Patients K and N showed T215D alone at diagnosis, and this was still present 11 and 13 months later, respectively.
K219Q was observed in 2 patients, and K219R was seen in 1 patient. One patient with 219Q and 1 patient with 219R, both with many other resistance mutations, showed retention of these mutations after 18 and 17 months, respectively. One patient with 219Q alone showed loss of the mutation within 9 months; further testing of samples at 24 and 36 months showed continued absence of this mutation.
Nonnucleoside Reverse Transcriptase Inhibitor and Protease Inhibitor Mutations
Two of the patients studied had NNRTI resistance alone: patient A showed only Y181C, which had disappeared 25 months later. The second patient (patient F) showed K103N, which was still present after 23 months, although a mixture at codon 108 (I/V) became fully wild type during this period. Two patients were infected with MDR virus, including K103N and V179L (patient D) and A98G, V106I, and Y188L (patient L), and these mutations were unchanged after 17 and 18 months, respectively.
No patients in the study had protease inhibitor (PI) resistance alone. The 2 MDR patients showed L10I, L24F, L33F, I54V, L63P, A71V, G73S, V77I, V82A, and L90M (patient D) and L10I, I54L, L63P, A71V, G73T, I84V, L90M, and I93L (patient L) in protease, and as with the NNRTI mutations in these patients, these PI mutations were unchanged 17 and 18 months after diagnosis, respectively. In general, little or no change occurred in secondary PI mutations during the course of these observations.
Multidrug Resistance and Viral Load
As described previously, 2 patients (D and L) were infected with MDR virus. The viral loads and CD4 counts for these patients are shown in Figure 1. Viral loads remained low off treatment for some time for both patients.
Patient D's viral load on diagnosis was 2500 copies/mL, fell to 150 copies/mL after 2 months (no treatment), and has remained below 1000 copies/mL for the subsequent 18 months. His CD4 count has declined from a peak of 1100 to 670 cells/μL during this period. No change in resistance-associated mutations was observed in this period.
Patient L's viral load was 4140 copies/mL at diagnosis and 456 copies/mL 18 months later, although his CD4 count showed little change during this period (from 685 to 527 cells/μL). As with patient D, no change was observed in the resistance-associated mutations during this period.
  icon paper stack View Older Articles   Back to Top   www.natap.org