How do chronic infections change over time? This is the broad question addressed in recent research published in the Journal of Virology. In their study, a team of scientists headed by Fabio Luciani investigated a hepatitis B virus infection over a course of 15 years.
Hepatitis B virus (HBV) is a DNA virus that targets liver cells and can cause chronic, life-long infection. Infected cells carry the viral genome in the nucleus, where it generates the mRNAs that will become viral proteins. The virus doesn’t use host cellular DNA replication machinery to make copies of its genome, however; it uses the RNA synthesis machinery to generate a long RNA intermediate, which is turned into DNA by a viral reverse transcriptase (rt).
This complicated replication process is a bit faulty: the reverse transcriptase doesn’t have proofreading capabilities, so misincorporated nucleotides are passed onto progeny virions. This generates viral genomes with mutations that allow immune or drug avoidance, and selection forces such as drug therapies, can lead to a new, dominant viral population. The research team wanted to investigate how the viral population changed over time in a single patient.
Liver and blood samples were collected four times over a 15-year period, from a patient who had undergone liver transplantation (see figure, right). Viruses from the samples were sequenced using Pacific Bioscience single-molecule sequencing techonology, which allow high-throughput long sequence reads compared to other next-generation sequencing technologies. The full-length HBV genomes could be read in one read using this technology, and detect multiple genome sequences more accurately than the short reads that require post-sequencing assembly.
This last point is important, because not only does HBV generate single nucleotide variants in new genomes, but it also generates variants missing large chunks of the parental genome. The most common are the preS variants, which are missing a section of the ENV (envelope protein gene) open reading frame, and spliced variants – genomes that are missing large chunks in the middle. These splice variant HBV genomes, or spHBVs, are important to HBV pathogenesis because they can enhance wild-type HBV replication, but estimating the proportion of spHBVs can be difficult, depending on sequence read length.
This method revealed that at all times, over a third of the aligned reads contained deletions over 500 nucleotides in length – potential spHBVs. Of the 95 distinct spHBVs detected, 83 were observed only in one sample, while 12 were observed at multiple time points. Nine of these observed in multiple times were first seen in the liver explant sample, and were also observed in later blood samples. After validating the results through additional sequencing methods, the authors concluded the spHBV genome diversity is greater than previously measured, confirming that the single molecule sequencing is a more sensitive method.
In addition to spHBVs, single nucleotide variants increased over time. Viral diversity, measured by Shannon index, increased from T0 to T1, decreased between T1 and T2, and increased again between T2 and T3 (see the pink line in the above graph). Most of the fixation mutants at all time points were non-synonymous mutations in the rt genomic region, which is a region known to contain drug-resistant variants. As one would predict, the evolution of resistant variants changed as the therapeutic regimen changed, with resistant variants selected for by the presence of antiviral drugs. Regions outside of the rt region were also observed to change with the treatment regimen, though the authors were unsure if these were selected by the treatment.
The viral genomics from a single patient over time aren’t immediately applicable to other patients, but this data does emphasize that an infection is a dynamic event that changes over time. Last week, researchers described a chronic bacterial infection they had tracked for 20 years in a patient with cystic fibrosis. This bacterial infection underwent genomic variation and selection in a very similar fashion, illustrating that this is a general microbial phenomenon. The more sensitive our measurements of these changes, and the more infections measured over time, the better able physicians will be to predict therapeutic outcomes.
-- Julie Wolf