What we can learn from Megaphage

Giant viruses are cool. Their genomes are bigger than those of many bacteria (especially endosymbionts), and they encode many housekeeping functions that are not obviously necessary for the viral lifestyle. They blur boundaries between the neat little categories we use to describe biological entities and thus remind us that those categories are conveniences, not reality. They demonstrate how little we understand living things, and that much of biology is still deeply mysterious.

Giant viruses are also a surprise. Phage biology managed to churn along for over a hundred years without finding them. Over this long history, many thousands of phages have been isolated, most from stool or sewage, most similar to already-known phages. None of these phages were megaphages. But it turns out that this was an oversight, not a reflection of biological reality. They are there. It’s just that nobody thought to look.

The authors of this paper in Nature Microbiology did look. Not by adding samples to bacteria and spreading the mixture on petri dishes to look for plaques, which is the traditional method of phage hunting. Instead they looked in gut metagenome sequences. This isn’t the first paper to identify new phages from sequence data (I think that would be this one), but it’s the first one to find megaphages. Complementary CRISPR sequences point to Prevotella as the likely hosts.

Nearly all phages used for phage therapy are tailed viruses with double-stranded DNA genomes of about 50,000 basepairs. Is that because these types of viruses are the most effective in phage therapy scenarios? Or is it because they are the easiest to find and grow? Is it possible that phage therapy has flailed and stagnated because the phage that were easiest to find were not the phage best-suited for therapy?

Phages are the most diverse organisms on Earth. Whenever we isolate new phages and sequence them, we find that most of their genes are completely novel, unrelated to any of the other millions of genes in our databases. These genes are the dark matter of genomics.

It’s a safe bet that many of these genes enable phage to exploit a variety of environmental niches. That’s what drives diversity and evolution in the rest of the biosphere. And the phage environment is bound to include their bacterial hosts. These hosts are themselves constantly adapting to the environment: responding to nutrient fluxes, fending off immune systems, alternately competing and collaborating with other bacteria.

Does it really make sense to expect that a phage which thrives in a monoculture of bacteria in a Petri dish to do the same in the bloodstream, or in a wound or in an infected lung? Especially if that phage is one that was isolated from a very different environment? Seems like wishful thinking to me.

Far too many phage therapy studies follow a familiar and unproductive pattern:

1. Screen sewage phage for pathogen host range under lab conditions (rich media, plenty of oxygen, no competing bacteria, no immune cells).
2. Spray/inject/gavage freshly infected mice with high doses of phage.
3. Record marginal therapeutic effects.
4. Publish a paper. Include a paragraph claiming that phage therapy has been long neglected but now there is “great interest” in PT as an alternative to antibiotics.

Hundreds of papers like this have been written. Zero phage therapeutics have been found safe and effective in humans.

The Prevotella megaphages are found not only in humans, but in other animal species. They must be reasonably abundant, but they escaped detection. Their discoverers were unable to isolate them from samples known (from sequence data) to contain them. Apparently they are well adapted to life in the gut, but poorly adapted to life on a Petri dish.

The dissonance between the number of phage that are out there, and the number we can culture–known as the Great Plaque Count Anomaly–has been apparent for over a decade. Students of soil phage biology estimate that perhaps 1% of soil phages can be cultivated. I would expect the number for gut phages to be higher, simply because we know how to cultivate most gut bacteria. But Prevotella megaphage shows us that the number is not 100% and is probably much less. And there is no reason to expect gut phages to thrive in a wound.

In the hunt for therapeutic phage, we don’t have to search for existing phage that are adapted to a particular therapeutic environment–we can create them. We may not know how to engineer phage that thrive in a wound or persist in the bloodstream. But we can select them. It’s possible to select phage that, for instance, not only persist in blood, but multiply when target bacteria are present. A recent paper shows that phage can be selected to persist a hundred times longer in the bloodstream:

Creating therapeutic phage is not rocket science. It’s biology (a much more complex science). If phage therapy is ever going to work, then phage therapists have to start thinking like biologists, seeing phage for the dark mysterious creatures that they are. It means making use of all that genomic dark matter to find or create phage that actually thrive in the role of therapeutic entities. It means that it’s no longer good enough to cure infections in Petri dishes or bogus mouse models and pretend you are doing something worthwhile.

Megaphage reminds us of just how diverse and strange bacteriophage are, and just how little we know about them. It’s time to start exploring that diversity and putting it to good use.