Move over, wastewater. Store-bought milk could be another way to track the bird flu outbreak in cows

Scientists from the University of Washington and the Fred Hutchinson Cancer Center have managed to generate a full genetic sequence of H5N1 virus from milk, a development they suggest means commercially purchased milk products could be used to monitor the progress of the bird flu outbreak in dairy cattle and to check for important changes in the virus over time.

With dairy farmers still reluctant to allow testing of their cattle, scientists trying to assess whether the outbreak is increasing or waning are in the dark. Likewise, their surveillance for important changes in the viruses — changes that would signal the virus is evolving to be better able to infect mammals — has been hampered by the limited data being shared by the U.S. Department of Agriculture.

Analyzing store-bought milk could provide a work-around, the University of Washington and Fred Hutchinson researchers suggested, similar to the efforts underway to analyze wastewater from around the country to check for presence of influenza A viruses. (H5N1 is a member of that large family of flu viruses.)

“I think the big thing is we’ve shown that this can be done, and it’s a good tool,” said Pavitra Roychoudhury, a research assistant professor in the department of laboratory medicine and pathology at the University of Washington School of Medicine. “And so if there’s resistance to having cows tested, this might be the next best option. And I think we can track for variant changes over time with this method.”

Since the outbreak of H5N1 in cows was first confirmed on March 25, a total of 51 herds in nine states have tested positive for the virus. One human infection has been detected, in a dairy farm worker in Texas who developed conjunctivitis.

Public health authorities and scientists believe H5N1 is more widely disseminated than those numbers would suggest, but are stymied in efforts to map the scale of the problem due to a lack of cooperation from dairy farmers.

Farmers have been reluctant to allow their cattle to be tested, seeing no upside and plenty of downside in being identified as having cows infected with the bird flu virus.

A U.S. Department of Agriculture regulation that went into effect on April 29 requires farmers to test lactating dairy cows due to be shipped across state lines. But the rule limits the number of cows that must be tested per shipment to 30, leaving the choice of the cows to be tested to the individual farmer. Only 15 new herds have been reported since the rule went into effect, leaving observers to wonder whether it can have its intended effect, as currently written.

“Whenever you have a regulation, someone will find a way around it,” said Keith Poulsen, director of the Wisconsin Veterinary Diagnostic Laboratory and a professor of large animal internal medicine at the University of Wisconsin-Madison.

Poulsen, who believes testing of milk in the bulk tanks on individual farms will eventually need to take place, agrees that monitoring milk for virus particles would provide useful data.

So does Richard Webby, director of the World Health Organization’s Collaborating Center for Studies on the Ecology of Influenza in Animals, located at St. Jude Children’s Research Hospital in Memphis, Tenn. Webby’s laboratory was involved in early testing to see if evidence of H5N1 RNA could be found in commercially purchased milk.

A large survey, conducted in mid-April by the Food and Drug Administration, found that one in five milk samples brought in 38 states was positive for H5N1 virus. It should be noted that the polymerase chain reaction (PCR) test that was used can detect both live viruses and fragments of dead ones, and cannot differentiate between the two. Efforts to grow live, infectious virus from positive milk samples have failed so far, buttressing the FDA’s claim that pasteurization kills H5N1 virus in milk.

Webby said there are some potential complications involved in using milk to monitor changes in H5N1; as the virus changes, it could be difficult to know at times what is being seen. For instance, if genetic sequencing showed evidence of more than one version of the virus in a sample, it might be hard to know whether changes observed in, say, the hemagglutinin and the neuraminidase proteins — the H and N in an influenza A virus’ name — occurred in all the viruses, or if some had one of the changes, some had the other, and some had both.

That’s not to suggest the work wouldn’t still provide important data, Webby said. “The fact that viruses with those two mutations are out there in itself is useful information. So I don’t want to pooh-pooh the whole idea at all. I think it’s absolutely a great way to follow the continued evolution of the virus. It’s just not going to be as tidy as doing single source samples.”

Peter Han, research manager of a laboratory run by Lea Starita at the Brotman Baty Institute for Precision Medicine at the University of Washington, said the group collected 40 samples of milk simply by asking people to bring in a test tube’s worth from the supply in their own fridges. Two of the 40 were positive for H5N1 fragments and the full genetic sequence was generated from one of those two samples. The two positive samples were from milk that was packaged in Colorado, which has reported outbreaks in herds.

Though inexpensive, the work wasn’t easy, Roychoudhury said. “Milk’s a really challenging sample to work with, because there’s a lot of fat,” she said. “So the fact that we were able to get a whole genome out of it was slightly surprising to us. It also means there was enough virus in there to pull a genome out, which to me was slightly — maybe not slightly — alarming.”

The genetic sequence of the virus was uploaded to Genbank, a sequence database run by the National Center for Biotechnology Information. Roychoudhury said the group struggled to figure out how to date the finding — genetic sequences are typically identified by where and when the virus sample was taken, and the species or substance from which the viral information was retrieved. With no clear idea when the milk was produced, the group identified the sequence simply by month and year.

The sequence has also been posted on Nextstrain.org, a website designed to track the evolution of pathogens like influenza and SARS-CoV-2, which causes Covid-19.

Trevor Bedford, a computational biologist in the vaccine and infectious disease division at Fred Hutchinson and one of the architects of Nextstrain, said the group behind the generation of the genetic sequence from milk has notified the various U.S. government agencies involved in the H5N1 response in the hopes that an organized effort to use this form of surveillance can be put in place.

“You could either imagine some systematic effort from something like USDA — I’m not sure who’s the exact right agency — or the academic groups could collaborate a bit more,” Bedford said, adding he thinks monitoring milk could provide key information that’s been missing up till now. “Separate from the sequencing for evolution [of the virus], which I think is very important, just even understanding prevalence through space and time, I think, would be really helpful and important.”