Identification of a Transmission Chain of HIV Type 1 Containing Drug Resistance-Associated Mutations
AIDS Res Hum Retroviruses. 2003 May;19(5):353-61.
Taylor S, Cane P, Hue S, Xu L, Wrin T, Lie Y, Hellmann N, Petropoulos C, Workman J, Ratcliffe D, Choudhury B, Pillay D.
Public Health Laboratory Service Antiviral Susceptibility Reference Unit, Division of Immunity and Infection, University of Birmingham, Birmingham B9 5SS, UK.
"....the long-term impact and significance of drug-resistant transmissions on the incidence of resistance within the whole infected population will depend on the stability of these viruses through multiple generations of transmission events.....Although the replicative capacity of this virus is low compared with nonresistant clinical isolates, it appears that fitness is not sufficiently compromised to lead to the loss of resistance-associated mutations....single mutations required to lead the virus back to a traditional "wild-type" genotype may indeed further compromise replicative capacity, thus maintaining the founder virus, albeit somewhat compromised compared with "wild type." The significance of such a hypothesis is that widespread use of antiretroviral drugs may lead to irreversible changes in the apparent "wild-type" circulating strain...... a variety of resistance-associated mutations have persisted for up to 2 years in the absence of therapy.....The novelty of our present findings are that the position 215 variants are not only stable over time, but can then be subsequently transmitted. Thus, the presence of this virus in patient C does not represent transmission of the T215Y-containing virus from patient B, but rather from the previous individual within the transmission chain.... with the transmitted virus in patient C matching most closely the semen virus of patient B. Although predictable, this observation has implications for the study of HIV transmission events..... additional studies are required to assess the unique evolutionary paths taken by HIV in various tissue compartments, and the impact of these different pathways on the transmissibility of HIV.....phenotyping does not capture the potential dynamic
evolution events consequent on genotypic alterations such as these (in this study; I refer you to the report of John Mellors' study at Resistance Workshop where he finds NNRTI-naive and experienced patients with NNRTI resistance not detected by commercial tests: http://www.natap.org/2003/resistance/day5.htm>). Thus we recommend that both genotypic and phenotypic analyses be required to identify transmission of drug-resistant virus..., the keys to prevention of further spread of such viruses are ....reaffirmation of healthy sexual practices within all sectors of the population".
We have investigated a potential transmission chain of HIV-1 with drug resistance-associated mutations between three individuals over a period of 5 years by use of cloning and sequencing of viral genes, and phenotypic characterization. Viruses containing reverse transcriptase drug resistance-associated mutations were transmitted sequentially between three homosexual men (A, B, and C), and persisted in one individual for at least 4 years, despite intermittent therapy and reduced viral replicative capacity compared with wild-type strains. Clonal analysis of the envelope gene from semen and blood virus showed that the virus transmitted to patient C was more closely related to virus from the semen than the blood of patient B. Our data suggest that
HIV variants with drug resistance-associated mutations can persist following primary infection, despite intervening antiretroviral therapy, and subsequently sexually transmitted. We provide "proof of principle" that such mutations can therefore become "fixed" within the circulating virus pool.
THE PREVALENCE of drug-resistant HIV-1 in newly infected individuals is increasing, reaching 20% in some studies. The overall impact of these data at the community level is de-pendent not only on such primary transmission, but also on the capacity for these viruses to be further transmitted to others.6-8
Mathematical modeling of this phenomenon suggests that the contribution of transmitted resistance to the overall prevalence of resistance within treated and untreated populations will probably remain minor, because of the reduced "fitness" associated with drug-resistant viruses, although this has been ques-tioned more recently. Such impaired fitness may manifest as reduced transmissibility and/or reversion of resistance-associated mutations to wild type following transmission, in the process restoring full replicative capacity. However, it has been shown that partial reversion of reverse transcriptase position 215 mutations (from Y to C/D/S) within transmitted drug-resistant
virus can restore viral replicative capacity, with the resulting virus more prone to reemergence of zidovudine (ZDV) resistance compared with wild-type virus. This may be a mechanism by which resistant virus can establish itself within
the circulating virus pool. Further, drug-resistant viruses are increasingly characterized by multiple mutations that may minimize the fitness deficit associated with specific mutants, and that may make reversion to wild type more difficult. Therefore, continuous monitoring for transmission and per-sistence of these variants is essential, and the demonstration of a transmission chain of such resistant virus would provide "a proof of principle" for spread within the community.
Three homosexual men (patients A, B, and C) were identified through routine epidemiological contact tracing.
Resistance assays were initially undertaken on plasma samples from patient C at the time of seroconversion, in 1998, and his virus was found to contain the mutations M41L, E44D, L210W, and T215D in the RT gene, and the polymorphisms L10I and I93L in the PR gene. These results were consistent
with transmission of virus containing resistance-associated mutations.
Retrospective testing of stored samples from patient A and patient B were then undertaken. Two plasma samples from patient A, obtained during 1994 and 1995, demonstrated an accumulation of nucleoside analog resistance mutations in RT, with the presence of mutations M41L, E44A/E, D67N/D, T69MIX, V118I/V, L210W, and T215MIX (mixtures at two nucleotide positions at codons 69 and 215 make a number of amino acid changes possible) shortly before his death. Unfortunately, a stored sample from patient A before initiating ther-apy, from the late 1980s, was not identified. Four sequential plasma samples from patient B were tested, the first obtained during 1994 between the dates of the samples tested for patient A. This initial sample showed virus with the
RT mutations M41L, E44D, V118I, L210W, and T215D, compatible with transmission from patient A. Of note, the T215D mutations was present, which has been described as a reversion mutation associated with primary infection with ZDV-resistant virus.
The subsequent three plasma samples from patient B, spanning 4 years, demonstrated identical mutations in RT. These samples included one obtained during viral load rebound following treatment interruption, suggesting that this viral variant represented the founder virus at the time of initial infection.
Virus from a semen sample from patient B obtained at about the time of presumed transmission to patient C also demonstrated the same RT and PR genotypes. Interpretation of the reverse transcriptase mutations observed
in viruses from all patients was made by use of at least one soft-ware
interpretation system. The presence of M41L, L210W, and T215C/D with or without additional changes at positions 44 and 118 was identified as conferring ZDV resistance by the Stanford database and GeneSeq genotype interpretation (Viro-Logic), but was identified as fully susceptible to ZDV by the VirtualPhenotype (Virco Belgium). This discrepancy is probably due to the fact that the VirtualPhenotype interpretation did not register the T215D mutation as a resistance-associated mutation.
Viruses from patients B and C, from about the time of trans-mission
between these two individuals (samples B4 and C1), were assessed for drug susceptibility and replicative capacity by the single cycle recombinant virus assay (PhenoSense). Of interest, they both appeared fully susceptible to all antiretroviral drugs. Replicative capacity was 30 and 27%, respectively, com-pared with the wild-type laboratory virus control. These values are at the lower end of the normal variation observed for clinical isolates without drug resistance-associated mutations (median value, 68%; n 5112), but fall around the median values derived from viruses with reduced phenotypic drug susceptibility (28%, n 5321; ViroLogic, N. Hellman, unpublished data).
The emergence of HIV drug resistance in treated individuals is well established, as is transmission of these viruses. However, the long-term impact and significance of such transmissions on the incidence of resistance within the
whole infected population will depend on the stability of these viruses through multiple generations of transmission events. The reduced virus "fitness" associated with drug-resistant viruses may minimize such a potential problem. Our observations, of transmission of virus with multiple resistance-asso-ciated
mutations through three individuals, are therefore important. A number of issues are worth highlighting. First, the viruses identified in these individuals contain the novel amino acid change from T/Y to D at position 215 of reverse transcriptase. The 215C/S/D changes represent partial revertants from the
ZDV resistance T215Y mutation, characteristic of the evolution of resistant virus (T215Y) in the absence of therapy, which restores fitness, and therefore confers greater selective advan-tage (and possibly greater likelihood of transmission).
In the cases we report, genetic mixtures at position 215 were present in patient A at about the time of transmission to patient B, presumably because of variable adherence to ZDV mono-therapy at the time. The aspartate (D) at this position was sub-sequently maintained through to patient C. Second, the genotype of plasma virus in patient B did not change over a period of 4 years, despite two different treatment regimens. Of particular note, the virus that reemerged in both blood and semen at the time of treatment interruption in 1998 had the same genotype as the original infecting virus, as did the virus present contemporaneously with the transmission to patient C. It is well documented that treatment interruption leads to outgrowth of wild-type virus archived from pretherapy. Our data demonstrate that the transmitted resistant virus does, in fact, represent the founder, archived virus within this individual. Although the replicative capacity of this virus is low compared with nonresistant clinical isolates, it appears that fitness is not sufficiently compromised to lead to the loss of resistance-associated mutations. Further, the array of such mutations within this virus may make such evolution more difficult. In other words, single mutations required to lead the virus back to a traditional "wild-type" genotype may indeed further compromise replicative capacity, thus maintaining the founder virus, albeit somewhat compromised compared with "wild type." The significance of such a hypothesis is that widespread use of antiretroviral drugs may lead to irreversible changes in the apparent "wild-type" circulating strain.
Patient B within our transmission chain received nucleoside analogs during the period 1994-1998, and it is possible that these drugs maintained the selective pressure for continuation of the resistant virus. However, it is noteworthy that even during some prolonged periods of treatment interruption, the original
infecting resistant virus emerged. The persistence of transmitted multidrug-resistant HIV-1, despite impaired replicative competence, has also been observed by others, and our own surveillance of transmitted drug resistance has identified four patients in whom a variety of resistance-associated mutations have persisted for up to 2 years in the absence of therapy (our
unpublished data). The novelty of our present findings are that the position 215 variants are not only stable over time, but can then be subsequently transmitted. Thus, the presence of this virus in patient C does not represent transmission of the T215Y-containing virus from patient B, but rather from the previous individual within the transmission chain. Unfortunately, stored samples from patient A at the time of his first diagnosis, in the late 1980s, were unavailable, precluding a confirmation that he was originally infected with wild-type virus, without reduced replicative capacity.
Study of HIV within the genital tract is important because this is the source of transmitted virus for the majority of infections worldwide. Genetic differences between virus in blood and semen have been documented previously, reflecting
the potential for separate evolutionary paths within these compartments.
Despite similarities in drug resistance-associated mutations in PR and RT between blood and semen in patient B, clonal analysis of the envelope gene identified clear differences at about the time of transmission, with the transmitted virus in patient C matching most closely the semen virus of pa-tient B. Although predictable, this observation has implications for the study of HIV transmission events. Compartmentalization between blood and semen must be considered in the process of assessing genetic distances between plasma virus of linked individuals. Further work is required to identify precisely
how selective pressures impact on virus in the genital tract in the era of antiviral therapy, as it is known that different drugs can penetrate the male genital tract to varying degrees. Also, additional studies are required to assess the unique evolutionary paths taken by HIV in various tissue compartments, and the impact of these different pathways on the transmissibility
It is interesting that phenotypic analyses of these viruses did not identify any reduced drug susceptibilities, despite the presence of the RT mutations M41L, E44D, L210W, and 215D. The L210W change increases resistance to ZDV in the context of T215Y and M41L; however, it is apparent from our results that position 215 partial reversion to aspartate (D) obviates this resistance, at least when measured in the single-cycle recombinant assay. Of note, borderline ZDV resistance was noted for a virus containing the mutations M41L, L210W, and T215C.15 By contrast, the genotype from patients B and C was interpreted as ZDV resistant when the gene sequence was submitted to two of three established software systems. Limitations in currently available interpretations systems, and phenotypic testing, are highlighted with respect to the position
215 revertants. Although it is clear that the T215C/S/D amino acid changes alone do not alter drug susceptibility in vitro, they represent markers for a more rapid development of resistant virus under an appropriate drug-selective pressure. In other words, phenotyping does not capture the potential dynamic
evolution events consequent on genotypic alterations such as these. Thus we recommend that both genotypic and phenotypic analyses be required to identify transmission of drug-resistant virus.
Our findings of discordance between interpretations systems also highlight the dangers of using the outputs from such laboratory and software approaches (i.e., drug x, resistance; drug y, susceptible) to infer possible transmission events involving drug-resistant viruses. Comprehensive phylogenetic analyses
remain the standard for estimating the likelihood of transmission of HIV.
Finally, the data presented here, as well as by others, demonstrate the potential for drug-resistant variants to be maintained in the absence of drug-selective pressure, despite their reduced fitness, and this must be considered in assessing the risk of spread of resistance within the population as a whole.
On the other hand, despite the presence of such viruses, patients B and C in our transmission chain did manage to successfully suppress viral load to undetectable levels with appropriate combination therapy. Thus, the keys to prevention of further spread of such viruses are appropriate clinical management of patients identified with drug-resistant HIV, and reaffirmation of healthy sexual practices within all sectors of the population.