A study that looks at the genetic makeup of archived samples of drug-resistant HIV in the UK has found evidence that some drug-resistant strains of HIV are very long-lived and are persisting in the population. However, they appear to be passed on by people who are undiagnosed or not on antiretroviral treatment (ART) and are not arising anew in people on treatment. Some resistance mutations are ‘fossil mutations’ that arose to drugs that are no longer in wide use.
In contrast, resistance mutations that commonly arise on treatment now, and do have an impact on choice of drug regimen, are not being commonly transmitted. This is probably because these mutations tend to make HIV less ‘fit’ and less easily transmitted. These include the K65R and M184V/I mutations, which confer resistance to the two drugs used in pre-exposure prophylaxis (PrEP).
The researchers observed that “there is discordance between the mutation patterns observed in HIV-infected patients failing therapy and those seen in TDR” (transmitted drug resistance), and therefore wished to investigate the persistence of mutations commonly seen in transmitted drug resistance.
The prevalence of transmitted drug resistance in the UK peaked at 15.5% in 2002 and had fallen to 11% by 2009. Even in people who experience HIV treatment failure, only 30-40% now prove to have drug resistance, as drugs with more ‘forgiving’ profiles that only develop resistance slowly in response to poor adherence are increasingly used.
The study looked at 1140 samples of HIV from the UK HIV Drug Resistance Database, sampled between 1997 and 2011, with the majority taken from 2004 onwards. All samples had HIV drug resistance.
The researchers only sampled HIV from people who had not taken ART and looked only at the most common primary mutations seen in transmitted virus against three classes of drugs. Between them, these represented about half the viruses in the database. These were:
- T215Y and related mutations: these cause resistance to the thymidine analogue drugs in the class of nucleoside reverse transcriptase inhibitors (NRTIs). These, the drugs stavudine (d4T) and zidovudine (AZT), are rarely used these days in the UK.
- K103N: this is the most common resistance mutation to the non-nucleoside reverse transcriptase inhibitor drugs (NNRTIs). This still has clinical relevance as it causes failure of the first-generation NNRTIs efavirenz (also in Atripla) and nevirapine. It does not cause resistance to the second-generation drugs etravirine and rilpivirine unless there are additional drug resistance mutations.
- L90M: this is the most commonly found primary transmitted resistance mutation to protease inhibitor (PI) drugs. It only causes resistance to the PIs saquinavir and nelfinavir, which are very rarely used now.
These were present in 58%, 38% and 13% of samples respectively, mostly as the sole mutation: only 7% had more than one mutation and only 1% all three.
The researchers used phylogenetic analysis, which links together viruses that have a pre-defined degree of genetic similarity. Viruses that are closely similar are treated as ‘clusters’ that may be connected by transmission. Dating when individual samples appeared within clusters can show how persistent the transmission of a particular drug resistance mutation is over time.
The analysis found 193 clusters, of which 78 were isolated pairs and 115 had more than two members.
The oldest cluster found was a group of nine viruses with the T215Y mutation whose first member was sampled in about 2000 and whose latest around 2008; the youngest was also in nine viruses with T215Y which appeared around 2006 and was still persisting in 2010. A cluster of ten viruses with K103N arose in about 2002 and was still persisting in 2010. The largest cluster, of 15 viruses with the L90M mutation, appeared in 2004 and was still present in several samples collected in 2011.
The researchers were able to determine from their analysis how commonly the viruses within clusters were transmitted. They calculated the ‘reproduction number’, which is the average number of times HIV gets transmitted from one individual. These varied from 1.3, implying that most of the time one person with the virus only transmitted it to one other, so forming a linear chain, to 2.8, implying that on average they passed it on to nearly three people.
The researchers say that it is not possible to know whether all the transmitted drug resistance has been transmitted while the source patient (who may not be on the database) was still untreated, especially with the K103N mutation, which still arises in response to currently used drugs. However, they comment, “since 82% of new infections are estimated to be derived from the undiagnosed (and therefore untreated) and while the proportion of patients failing treatment with resistance is small, it seems likely that the continuing major source of TDR in the UK is from untreated patients.”
The researchers forecast that although the transmitted mutations are persistent, they will decay slowly over time, with a decay time of 13 years being observed with T215Y. L90M is now only observed in 1% of newly diagnosed people.
The authors comment: “It is reassuring to note that much TDR in the UK is likely to be derived from untreated and likely undiagnosed individuals and that it is mostly unlikely to have considerable impact on therapy outcomes.
“However, the same cannot be presumed for developing countries where therapy options and laboratory monitoring are much more limited.”
Mbisa JL et al. Evidence of self-sustaining drug resistant HIV-1 lineages among untreated patients in the UK. Clinical Infectious Diseases, advance online publication. doi: 10.1093/cid/civ393. 2015.