3rd International Workshop on HIV Drug Resistance 2-5 August 1994, Kauai, Hawaii, USA

INTRODUCTORY TALK: MODELLING HIV GENETIC VARIATION

Int Wkshop HIV Drug Res 1994 Aug 2-5;3:1 (abstract no. P1)

John M. Coffin
Department of Moleculr Biology & Microbiology, Tufts University School of Medicine, Boston, MA, USA

One of the most striking features of my and related lentiviruses is the extent of genetic variation seen in virus populations in vivo. Comparison with a model system for retrovirus variation that we previously developed has led to the conclusion that these populations behave (to a first approximation) as quasispecies subject to Darwinian evolution. Simple modelling of mutant frequency in such populations as a function of number of replication cycles, mutation rate per cycle, and selective forces lead to the following conclusions.

1) At population sizes and numbers of replication cycles likely for my in vivo. subtle selection is the most important force determining, mutant frequency. Differences in fitness of less than 1% have a very strong effect whereas 10-fold differences in mutation rate have very little effect.

2) Under these conditions, no mutations can be considered to be truly neutral and assumptions (such as a linear rate of accumulation of mutants) and analysis (like synonymous/non-synonymous mutation ratios) based on neutral theory are completely inappropriate.

3) In cases where the contributions to fitness of different mutations are simply additive, they will be unlinked to each other; that is each will be randomly distributed across a background of all others at a frequency determined by its own selective advantage or disadvantage.

4) At high numbers of replication cycles, the frequency of all mutations approaches a steady state in which the appearance rate (by mutation) matches the loss rate (by counterselection).

These last two points are of particular importance for drug resistant variants which must be somewhat deleterious in the absence of the drug, since they define the frequency of preexisting resistant mutants at the time treatment is started. In order to fully understand these processes, it will clearly be necessary to develop a dynamic model of the my infection process in vivo. In its simplest form, such a model will relate the average rate of virus replication, the frequency of infection of susceptible cells, the effect of the infection in the life span of the cell, the replenishment rates of target cells, the relative latent and productive periods of individual cells and the transfer of cells from the site of replication to the site of detection. This is obviously a complex task, and we have far too little information to draw such a model at present. Most analyses (such as the in situ PCR) provide only snapshots of the dynamic processes. However, the rapid and dramatic effects of non-nucleoside RT inhibitors on HIV populations in vivo may well provide the key information necessary to resolve this problem, and approaches to this kind of analysis will be presented.

Download PDF of this abstract.

1994-08-02
P1

Copyright © 1994 - International Medical Press Ltd. Reproduction of this abstract (other than one copy for personal reference) must be cleared through the International Medical Press Ltd. 2-4 Idol Lane, London EC3R 5DD UK.