Intra-host evolutionary dynamics of the hepatitis C virus among people who inject drugs

Abstract

Most individuals chronically infected with hepatitis C virus (HCV) are asymptomatic during the initial stages of infection and therefore the precise timing of infection is often unknown. Retrospective estimation of infection duration would improve existing surveillance data and help guide treatment. While intra-host viral diversity quantifications such as Shannon entropy have previously been utilized for estimating duration of infection, these studies characterize the viral population from only a relatively short segment of the HCV genome. In this study intra-host diversities were examined across the HCV genome in order to identify the region most reflective of time and the degree to which these estimates are influenced by high-risk activities including those associated with HCV acquisition. Shannon diversities were calculated for all regions of HCV from 78 longitudinally sampled individuals with known seroconversion timeframes. While the region of the HCV genome most accurately reflecting time resided within the NS3 gene, the gene region with the highest capacity to differentiate acute from chronic infections was identified within the NS5b region. Multivariate models predicting duration of infection from viral diversity significantly improved upon incorporation of variables associated with recent public, unsupervised drug use. These results could assist the development of strategic population treatment guidelines for high-risk individuals infected with HCV and offer insights into variables associated with a likelihood of transmission.

Figure 1

Phylogenetic tree of HCV consensus sequences and genotype counts. Maximum likelihood phylogenetic tree of HCV WGS. Reference genomes for each genotype are shown with black points and different coloured tips signify those included in this study. This tree was rooted on a genotype 7 reference sequence. Stars indicate mixed infections which were removed from subsequent analyses, red for sample 33 and blue for sample 983 (two longitudinal samples). Inset bar chart depicts counts of genotypes for each sample (including mixed and genotype switches) separated by time point.

Figure 2

Shannon diversity across the HCV genome and over time. Read depth normalized Shannon diversities for each gene grouped by timeframe of infection: less than 6 months (< 6 mo), between 6 months and 1 year (< 1 yr), between 1 and 2 years (< 2 yr), between 1 and 3 years (< 3 yr), between 3 and 4 years (< 4 yr), and greater than 5 years (> 5 yrs).

Figure 3

Measuring the time to infection prediction capacity of diversities from each gene/region. Shown are the root mean square error (RMSE) values used to measure the capacity of diversities from each region to predict duration of infection. The top 17 models determined by their median RMSE values are shown for all regions tested.