As we celebrate our 20th anniversary, Lisa Maier shares the story of how one PLOS Biology paper set the course for her…
In this ‘behind the paper’ post, Megan Griffiths discusses the challenges of studying herpesvirus transmission by vampire bats in South and Central America.
My name is Megan Griffiths, and I’m in the fourth and final year of my PhD at the MRC-University of Glasgow Centre for Virus Research. I’m particularly interested in understanding and controlling zoonoses pre-outbreak, in their animal hosts. Bats have inadvertently become my primary host species of interest over the years, whilst rabies has always been one of my favourite viruses!
Our new article exploring a vampire bat betaherpesvirus as a transmissible vaccine vector candidate follows the theme of our research group; finding ways to reduce the burden of vampire-bat-transmitted rabies virus. From over a decade of research in South and Central America, it has become clear that current methods of control are not sufficient to reduce the public health burden, or stop the spread of rabies to new frontiers. Given the success of wildlife vaccination against rabies in other species (e.g., foxes and raccoons), vaccination of vampire bats seemed a logical approach to take. However, vaccine distribution to vampire bats faces unique challenges, as their diet prevents the use of food baits as a delivery mechanism. The discovery of betaherpesviruses in metagenomic sequencing of vampire bat saliva samples led us to consider the use of a virally vectored transmissible vaccine that would be able to self-spread between bats.
Previously, we’ve established that our vampire bat herpesvirus – DrBHV – has an extremely high prevalence, and the deep sequencing of just two samples showed that there might be multiple strains of DrBHV contributing to this. Since viral prevalence is an important indicator of how successful the transmissible vaccine might be, we decided that it was vital to determine how many strains of DrBHV were circulating in our bat populations, and at what frequency compared to total DrBHV prevalence.
In order to do this, we needed to deep-sequence saliva samples, collected from wild vampire bats over a 6-year period. One of the main challenges in this project was the initial sequencing process, from choosing the section of the genome to sequence, to the sequencing method itself. Given the low viral load in our field-collected samples, we needed to enrich our chosen section of genome by PCR. We then used Illumina short read sequencing – a method our group has used extensively for rabies sequencing in the past – on our samples.
As soon as we began our sequence analysis, it became clear that we had vastly underestimated the diversity of DrBHV across Peru – there were clearly far more than 2 strains present. Given this genetic diversity, we were unable to reconstruct large sections of sequence (contigs) from the short reads. In hindsight, a different sequencing method, such as long-read Nanopore sequencing, would have been more suitable for the dataset, and will most likely be implemented in the future. However, it was at this point that the pandemic reached the UK in full force, and lockdown prevented any further sequencing for nearly a year. As such, we persevered with the sequence data available, and (with a good deal of trial and error) adapted software originally designed for HIV sequence processing to pick out short read alignments to use for phylogenetic analysis and strain genotyping. With this method, we were able to identify not 2, but 11 distinct genotypes of DrBHV, and create geographically explicit prevalence profiles for each. Encouragingly, genotype-specific prevalence is still able to reach very high local levels, suggesting high rates of transmission.
Some of our most interesting results came from longitudinally collected samples, from 20 bats that were each captured 2-3 times over the sampling period. These samples showed that, not only could bats be infected with multiple genotypes of DrBHV, but that this infection could happen sequentially, meaning that there was no significant immune barrier to additional infection. This information is crucial to the success of DrBHV as a transmissible vaccine, as the vaccine needs to be able to transmit in our bat population, which we can see is already heavily infected.
Excitingly, the sequence data generated here can also be used to develop an epidemiological model of DrBHV transmission. As a recently discovered virus, its mode and rate of transmission are as yet uncharacterised; however, we plan to use the strain-specific prevalence data to inform model fitting. This means that we can simulate the spread of DrBHV as a vaccine vector and predict how effective it will be at reducing the impact of rabies virus in vampire bats – work that is currently underway. Additionally, with our results continuing to support DrBHV as a transmissible vaccine vector, the next big step of isolating and engineering the DrBHV can begin.