Hot models!

One of the things I love about science--but that can also be frustrating--is that every new piece of information leads to a new unanswered question. We've learned so much about microbiology and human disease since the time of Koch and Pasteur, but in many other ways, we're still at square one. One reason is because research over the last century has largely focused on disease-causing organisms--and within those, many studies have focused on identifying factors that allow these organisms to cause disease. This concentration has led to many breakthroughs (such as vaccine targets), but it has also left a number of notable shortcomings in our knowledge. For instance, I've written previously mentioning just how little we know about our own commensal flora. Another area where we have a surprising dearth of knowledge is the transmission of infectious agents.

To back up a bit, when we look at the components of an infectious disease, we often break it down into a few steps. (Note that I'm simplifying quite a bit here, but bear with me). The organism must first colonize the host, which means either being able to avoid or otherwise escape the host's intrinsic defenses. These defenses include "mechanical" barriers (such as the mucus layer) as well as more specific defenses, including the various cells and proteins of the immune system. They then must be able to replicate. Depending on the organism, this may require additional invasion into host tissues or the host cells themselves (such as viruses and intracellular bacteria). These organisms then may either persist in the host for years (such as the Pseudomonas I described here), or they may cause an acute infection and be quickly cleared by the host. In either case, the host will be a "dead end" unless the organism is then transmitted to another host.

Examining the phenomenon of transmission isn't as easy one may think. Generally, there are two ways we can approach it (and again, I'm simplifying). We can carry out epidemiological studies, looking at how an organism passes between individuals, and then working to correlate that to the biology of the organism (and/or the host). Occasionally, we can use human volunteers, but this frequently isn't possible. Alternatively, we can model this in animals. This is how typical studies of pathogen virulence have been carried out: inoculate a model animal with the organism of interest, and observe whatever parameters are of interest. Many permutations, of course, exist: the investigator can knock out a putative "virulence gene" in one strain of pathogen and inoculate them into one group of animals, while infecting a second group of animals with the wild-type strain to see whether the knocked out gene affects the course of disease. There are dozens of possibilities, but generally they involve an artificial administration of the pathogen (for example, by injection). This allows the investigator to control the dose of the organism the animal receives, but real life is obviously much more messy. These types of animal studies have taught us a lot about factors that contribute to colonization and virulence, but not as much about the other important part of disease development: transmission of the agent to a new host. Additionally, for transmission studies, it's difficult to directly extrapolate from animals living in cages to the human population. Therefore, while we can glean a great deal from these types of studies, they're certainly not perfect.

This gap has recently been highlighted by the outbreak of H5N1. We have a few clues suggesting why it's so virulent, but the million dollar question is: what will make it more easily transmissible between humans? This is when the trouble really starts: when it spreads among humans without any need for birds as a reservoir.

There has been at least one aspect of the host-virus interaction that has been proposed to play a role in effective transmission. As reviewed at Effect Measure, two groups of scientists have suggested that the specific types of sugars on cells of our lungs--and where they appear within the lobes of our lungs--play a role in the binding of H5N1 to mammalian cells, and therefore may effect how efficiently transmitted these viruses are. However, the basis of this is rather sketchy: we don't have enough information about other influenza viruses that were present in the human population but not effectively transmissible to know if this is a valid generalization or not. Additionally, animal models of influenza transmission have been less than optimal. Mice don't consistently transmit the virus from one animal to another, and influenza viruses generally have to be "mouse adapted" (for example, by serial passage) in order to efficiently infect the animal. (The 1918 "Spanish" influenza strain, and several recent H5N1 strains, have been notable exceptions). Additionally, even when using a mouse-adapted strain (WSN), it wasn't found to be transmitted between mice in an experimental setting.

The other animal model commonly used in studies of influenza--ferrets--are more naturally susceptible, but present other difficulties as far as acquisition and housing of animals. Therefore, a new study went back to the animal whose very name is synonymous with medical research--the guinea pig.

They found that this served as an improved model. Virus was transmitted between guinea pigs that were housed together, both in the same cage and when separated by a distance of ~3 feet (meaning you might want to stay a little farther away from those infected co-workers...) However, I don't think this paper was PNAS material. Yes, the guinea pig model seems to be less cumbersome than the ferret one, and perhaps more biologically similar to humans than a mouse one, but it's not a huge breakthrough, nor a particularly novel piece of research. Still, anything that draws more attention to this critical gap in our understanding of agent transmission is a good thing, IMO. If we had a better handle on the factors that caused an avian strain of influenza virus to be more efficiently transmitted among humans, then we could better focus our resources and know when to really sound the alarm--unlike now, when we're flying blind in many ways.


Lowen et al. 2006. The guinea pig as a transmission model for human influenza viruses. PNAS. 103:9988-92.

Image from

More like this

Ebola has long been associated with wildlife. From the early days, bats were viewed as a potential reservoir (though it wasn't confirmed that they actually harbored the virus until 2005). Contact with wild animals--particularly primates which were butchered for food--was also long thought to be a…
[Warning: this post is fairly long and has a reasonable geek factor. It explores the question whether the virulence of H5N1 "must" moderate as the virus evolves.] The high case fatality ratio of H5N1 (currently around 60%) is a reflection of how virulent this virus appears to be at the moment.…
Ebola has long been known to be a zoonotic virus--one which jumps between species. Though it took several decades to find evidence of Ebola virus in bats, these animals had previously been associated with human index cases of Ebola disease have worked in bat-infested warehouses or traveled to caves…
Time out for a bit of infectious disease terminology. The words pathogenic, virulent and transmissible get tossed around a lot when talking about the bird flu virus and the possibilities of a pandemic. They are sometimes used interchangeably. They aren't interchangeable, however, and their…

Nice post, Tara. What's especially interesting is your use of the word "model" here, which is the use biologists prefer. Models for a biologist are animal models, i.e., what Keller calls "stable targets for explanation," like a fruitfly, a mouse or a ferret. Animal models have several characteristics that make them useful to biologists. They can be "standardized," so that findings of one lab can be checked and compared with another; they can be manipulated experimentally, something we usually cannot do with the human animal (and hence provides employment for us observational epidemiologists); they retain all the relevant complexity of the real object of investigation (e.g., the mammalian cell, human physiology or biology). This last is one of the things that differentiates the biologist's models from those of the mathematician or physicist.

For those of us who are "modelers" a model is a stripped down bones-only version of what we are trying to understand. We try to get rid of just those complexities that are so important to biologists. This is one of the reason that mathematicians and physicists had such a hard time collaborating historically. The minute the physicist says to the biologist, "First assume your mouse is a perfect sphere," the biologist tunes out. The science studies scholar Evelyn Fox Keller at MIT has written a fascinating book about the history in the 20th century of mathematicians and physicists trying to work with developmental biologists. Keller is especially well equipped to discuss this as before she became one of the better known feminist science study specialists she was a mathematical biologist who worked with Lee Siegel on slime molds. Her book, Making Sense of Life was published in 2002 by Harvard Univ. Press and is now out in paper. It is a really fascinating read and I recommend it highly.

"research over the last century has largely focused on disease-causing organisms--and within those, many studies have focused on identifying factors that allow these organisms to cause disease"

That's very true. Another point that I find crucial and is most of the times misleading is not only the model used, but the strain. We think of pathogenic organisms as homogeneous entities that have been waiting for us to come during their evolutionary lifetime, and to which we are the tastiest thing.
With this common way of thinking we forget first of all, that in many cases most of the strains are non-pathogenic. That they might not be obligate pathogens.That they might not get as much as we think they can get from us ( there are many places where you can get food and you don't find such a "ferocious" immune system as ours ). And, finally, that what we call "pathogenicity factors" might have, in their original environment, a more useful function than to "cause disease".
Examples of this are found in the literature.

The outcome of this is that we have a lot of fascinating work to do, and many uncovered secrets to find!

Quoting revere:

The minute the physicist says to the biologist, "First assume your mouse is a perfect sphere," the biologist tunes out.

Try explaining "vector masking" to a programmer. Interdisciplinary language is sooo much fun!