Better avian influenza surveillance

Avian influenza, especially in its highly pathogenic form, remains a serious threat to the poultry industry worldwide. By early June this year, the World Organization of Animal Health had reported detections of the disease in a whopping 25 countries, including Mexico but fortunately not the U.S. or Canada so far. The strategy for avian influenza virus (AIV) surveillance around the world, therefore, matters a great deal.

Wild birds, particularly waterfowl, are the natural reservoir for AIVs and spread them across the globe during migration. Current surveillance involves collecting samples from wild birds killed by hunters or that died due to natural causes. Live capture and release is very difficult.

The DNA/RNA in fecal swabs is tested for type of AI, to see what strains are where on the globe and if new strains have appeared. However, this strategy is both inefficient and ineffective, explains William Hsiao of the B.C. Centre for Disease Control (BCCDC), University of British Columbia (UBC) and Simon Fraser University.

Because wild birds usually do not suffer systemic infections from AI like commercial poultry do, Hsiao says surveillance involving random capture means that some or even many wild birds that carry the virus in its many forms could be missed. Testing larger numbers of wild birds is not feasible due to high costs, stress on the birds, ecological disruption and so on.

 Testing wetland sediments where wild bird feces accumulates, however, could be a way to not only determine what strains of virus are present in wild birds each year but to also establish whether new strains have appeared – all before an outbreak in a commercial flock occurs.

So, in 2015, after the major AI outbreak in B.C., the provincial government sought a novel surveillance approach. About 240,000 poultry birds died or were destroyed to control that outbreak, and over 48 million were lost through outbreaks in the U.S. around the same time.

A team involving Hsiao and partners such as the Canadian Wildlife Health Cooperative (CWHC) did a pilot study to analyze wetland samples. The results were very promising, with a 24 per cent rate of virus detection compared to less than a one per cent rate of detection through wild bird swab sampling.   

The project was supported by $2.5 million from partners such as Genome B.C., the Canadian Food Inspection Agency, Agriculture and Agri-Food Canada and the Sustainable Poultry Farming Group. The research is still going strong.

Hsiao is a co-lead investigator with Chelsea Himsworth of the B.C. ministry of agriculture and CWHC, Jane Pritchard of the B.C. ministry of agriculture, Natalie Prystajecky of BCCDC and UBC and Craig Stephen of CWHC.

Reflecting on this surveillance strategy in comparison to wild bird sampling, Himsworth notes that in a given wetland sediment sample, the feces of potentially several birds can be present. This makes the chances of getting a viral RNA sample much higher.

“In the proof-of-concept study from 2015 to 2018, we found AIV RNA in 20 per cent of samples – and none was found in wild birds prior to the HPAI outbreak in 2014/2015 in B.C.,” she explains.

“We found huge numbers of viral subtypes and strains in the sediment samples. There were up to 16 different AIV subtypes in a single sample. In a bird sample, if you can get any viral material at all, you might get one subtype that that particular bird happens to be carrying.”

“It’s a fairly novel approach to doing population-level sampling,” notes Prystajecky, “and a strategy, combined with genomics technology, that can really compliment surveillance globally and locally. We are talking to research groups in Canada, the U.S. and Europe and they are very interested to know how things are progressing. A group at the Animal and Health Agency of the U.K. also wants to collaborate.”

Research changes going forward
In the pilot study, the team used sampling technology from a B.C. firm called Fusion Genomics. However, because the firm’s probes are proprietary and limit the ability to add new probes as new AIV variants emerge, the team has developed its own probes.

“We would like our method to be what’s called ‘open source’ or freely available in order to enable this type of surveillance to move forward globally,” Himsworth explains. “We estimate that in two years, we will be able to release a research manual for others around the world, covering from how to best sample to how to make the probes to how to best do data analysis.”

PhD student Kevin Kuchinski of UBC and BCCDC says the team’s new probe approach works similarly to probes they’ve used before – a DNA sequence is placed on the probe, which is then exposed to the sample. It may contain matching DNA that automatically attaches to the probe sequence, making the helical ladder structure of the DNA molecule complete.

“Harnessing the way DNA matches up is used in countless technologies around the world,” Kuchinski says. “We send the sequence we want to a probe manufacturer and it comes back within a week or two. We tested our probes on a small number of samples this spring, and detected several AI strains commonly found in North American wild birds, as expected. We will now test hundreds more sediment samples to make sure the technology scales well.”

The ability to detect new AI strains that may appear in future has also been worked into the new probe design. “Some parts of the influenza genome never change, and we’ve included these sequences in our probe panel,” Kuchinski explains. “So, if we are detecting a lot of match-up from those probes, but not detecting the crucial H or N genes, that’s a red flag that there is something new out there and it’s time to design more probes.”

Best sample collection
Not only is the probe technology changing, but the group is also optimizing the way it collects wetland samples – and how it will recommend other groups around the world conduct sampling.

PhD student Michelle Coombe of UBC and the B.C. ministry of agriculture notes that the areas in a particular wetland where sediment samples could be taken is very numerous, and the number of wetlands even in a fairly small given region is vast, making the total sampling area extremely large.

“Where we sample is important, as collecting and analysis of samples takes time and costs money,” Coombe explains. “We want to get the most surveillance bang for our buck, so we need to find out if there are factors in the environment that help predict if there is a larger amount of, or any, AIV RNA in the sediment at a particular site.

“Does the pH or chemical composition of the sediment/water matter much, or are observations of wild bird flocks our best predictor? We are working on some mathematical models to help us determine this.”

Not only does the area where a sample is taken matter, but the type of sediment may also matter – that is, some sediments are better than others in terms of their willingness to give up pieces of AI viral RNA.

“We still don’t know whether one type of sediment is better than another at doing this,” Coombe says. “We know that viral RNA can have an electrostatic charge and some of it sticks to sediment particles and some doesn’t. While sticking to sediment protects RNA from the degradation that it usually experiences from floating freely in water, getting it away from sediment particles is an issue.”

In the end, it could be that the abiotic factors are more important, and the number of birds at a particular site may not matter very much in terms of optimized sampling. “We just want to make the sampling process as efficient and effective as possible,” says Coombe.

The fragility of RNA means that it’s always very likely fresh. “It seems that the virus doesn’t survive well year over year, based on preliminary data,” Himsworth explains. “The sediment, as it dries over the summer, goes through a self-sterilization. Birds come through in the spring on their migration and the wetlands then bake and dry up in the hot sun. The UV radiation is quite effective at destroying it.”

The scientists will also determine that a sample taken in the spring is not from last fall, which indicates if viruses in the sediment are a current reflection of viral strains circulating in wild ducks and, thus, how risky a particular season is with respect to potential transmission of AIV from wild birds into poultry.

“Our study will investigate this question,” she says, “by comparing the genetic relationship of AI viruses found in sediment to those found in wild ducks on the same wetlands and will observe if the viruses found in the sediment change, or don’t change.”  

The study will continue for two more years. In addition to finalizing probe design and optimizing the sampling strategy, the team, through doing simultaneous sampling of sediment and wild birds, will continue to determine how good a mirror sediment is to what wild birds are actually carrying.