Do resistance mutations disappear off therapy?

The mechanism for reversion from mutant to wild type following a treatment interruption is thought to be possible through either overgrowth by competitive advantage or genetic reversal of mutations.

Two studies at this meeting seemed to point to the former being the more important and likely mechanism, and concluded that a prerequisite for reversal to occur must therefore include a residual pool of wild-type virus when treatment is withdrawn, and a relatively less fit mutant virus. Additionally, the hope that mutated virus is less replicatively competent, if shown in vivo, may offer benefit from continuing on a failing treatment when no other options are available.

A study presented by Clive Loveday, from the Royal Free Hospital, London provided important information on the relative fitness of multi-drug resistance-associated mutations in vivo. The relevance of viral fitness cannot studies on viral fitness are in vitro studies. [1]

11 patients (all gay males) previously treated with mean of 5 previous ARVs (range 3-8), whose current treatment was failing virologically (mean viral load 4.88 log, range 3.26 – 6.06 log10 copies/ml) had RT-PCR-based resistance assays performed prior to, and a mean of 3 times (range 3-5) subsequent to, stopping ARV therapy.

All codons in protease and reverse transcriptase genes were assessed for changes and the proportion of wild-type and mutant viruses present were determined by measuring electropherogram peak heights.

Viral fitness was determined by comparing change in viral fitness following stopping treatment as

s = 1/t × 1n × [ q(t)p(0)/p(t)q(0) ]

s = viral fitness; q = fittest virus (assumed to be W-T when off therapy); p = least fit virus; t = time since stopping therapy; 0 = last time patient received therapy

Two clear patterns emerged amongst mutant codons associated with resistance:

  1. decline in proportion (majority of primary RT and PR mutations)
  2. little or no change in proportion (majority of secondary PR mutations)

The primary mutations in protease at 30N, 46I/L and in RT at 184I/V were associated with substantial reduction in viral fitness (approx 12%, 21% and 11.5% loss in fitness respectively compared to wild-type). By contrast 10I/V, 20R, 36I and 63P mutations in protease persisted after stopping treatment and remained fit compared to wild-type (0%, 0%, 2.2%, 2.2% respectively). In the case of the detailed response given to two patients, the virus remained 100% mutant at codons 10I and 63P out to 135 and 43 days respectively.

The clinical relevance of detecting these secondary mutations following a treatment interruption may be that it leads to an increase in fitness of mutant virus if protease treatment is re-introduced. However, the study concluded that the re-emergence of resistant population may be very slow if the new treatment is itself able to significantly reduce HIV replication. No information was presented about restarted treatment in these patients.

The likelihood of reversal to wild-type virus, and the associated risks from a treatment interruption may be more important if it results in shaking off multi-drug resistant mutations. Two case studies presented by Elodie Fontaine, from CRP-SantŽ, Luxembourg, looked at individual examples of the MDR resistant mutations Q151M and the 69-insertion, both of which are associated with broad nucleoside class resistance, following a treatment interruption. [2]

Each of the patients in this study had regular RT sequencing from plasma and cellular DNA (PBMC) before and after the detection of these mutations whilst on treatment, during a period of treatment interruption, and, in one patient, for 5 months following re-starting treatment. The 69-insertion was detected in a patient on failing combination ZDV/ddI/RTV of (viral load >500,000 copies/ml) together with changes at codons 70 and 215 and persisted despite switching background nucleosides to d4T/3TC, and later ritonavir to nevirapine.

Four months after stopping all treatment the 69-SSS mutation disappeared (or changed to a mixed populations) in both plasma and PBMC. When cultured in vitro, a recombinant 69-sss virus continued to grow in the absence of drug for over four months, although in a mixture with wild-type (in vitro) showed weaker fitness and was outgrown by wild-type virus within three months. No information was provided about response to restating treatment in this patient (and therefore whether selective drug pressure would have resulted in reappearance of this mutation. The Q151M mutation appeared in the second patient while on dual therapy with ZDV/ddC (VL >500,000 copies/ml), first in plasma and then in PBMC. During an 18 month period without treatment, a very slow reversion at codon 151 to wild-type occurred in plasma but even after this time a mixed population persisted in proviral DNA. Following reintroduction of treatment (ZDV/3TC/EFV) the Q151M mutation rapidly returned (within a month) in both plasma and PBMC.

The long in vivo persistence of these mutations during treatment interruption in each of these cases led the investigator to conclude that both 151 and 69-insertion mutation have good replication fitness. Also, that although stopping treatment will not lead to a return to wild-type, continued treatment with compounds resistant to these mutations, offered no further benefit to these patients.


It would appear unlikely that treatment breaks have any longer term impact on resistance mutations. There does remain, however, a beneficial effect when resistant to wild type switches are seen off treatment.

This switch is associated with an enhanced virological response when therapy is reintroduced. It remains to be seen if this benefit persists over time. The loss of CD4 cells often seen during treatment interruptions must also be balanced against this observed virological benefit.


  1. C. Loveday et al – Replicative Fitness in vivo of HIV Variants with Multiple Drug Resistance-Associated Mutations. 4th Workshop on Drug Resistance & Treatment Strategies. June 2000, Sitges. Abstract 136.
  2. E. Fontaine et al – Slow Disappearance of Muti-Drug Resistance Mutations in Plasma and PBMC after Treatment Interruption. 4th Workshop on Drug Resistance & Treatment Strategies. June 2000, Sitges. Abstract 159.

Links to other websites are current at date of posting but not maintained.