In many cases, pathogens (disease-causing organisms) are identified by a common series of steps, known as Koch's postulates. Robert Koch described these steps in 1876 when he used them to prove that Bacillus anthracis was the cause of anthrax. During the past century, his steps have been used successfully many times.
Koch's steps are summarized in the image below. In short, if a pathogen is causing a disease, we should be able to find it in an individual who is suffering from that disease. Once we find it, we grow it, and separate it from everything else so that we obtain a pure culture. A pure culture contains only one type of bacteria (or virus, etc.).
If that microbe really is the culprit, and caused the disease once, it should be able to cause the disease again.
To prove this, we inoculate a healthy animal with the microbe. If that animal develops the disease, then we know that the microbe we cultured is the guilty party.
Of course, following the steps isn't as easy as it looks in my diagram and sometimes there are technical reasons that make Koch's postulates difficult, if not impossible, to meet.
For example, if we are to meet Koch's criteria for proof, we have to be able to:
- Culture the pathogen
- Find an animal (plant, fish, etc.) that gets the disease with the same kinds of symptoms.
But what if we can't culture the bacteria?
Or what if the disease is caused by different kinds of bacteria who work together?
What if it's hard to find an animal that gets the disease? (we call them "animal models")
These cases present the biggest challenges to would-be microbe hunters. Finding an animal (besides a human) that gets the disease isn't always easy and some bacteria have never been cultured.
What do we do in those cases?
What do you do if you don't have a good animal model?
I don't have good answers for solving this problem, but I can describe things that people have done. In some cases, researchers have tried to create mice with "human" immune systems, in others, they have searched for many years to find the right animal model, and sometimes they've used themselves.
Barry Marshall, co-winner of the Nobel prize for medicine in 2005, described his experience in his autobiography. He began studying a possible connection between stomach ulcers and the bacteria, Helicobacter pylori, in 1981. In 1984, frustrated by the inability to find an animal model, and skepticism on the part of the scientific community, he tested the bacteria on himself. In retrospect, this was a dangerous experiment. Marshall made himself severely ill and ended up hospitalized with stomach damage.
Quoting from his autobiography:
I was driven to get this theory proven quickly to provide curative treatment for the millions of people suffering with ulcers around the world.
Fortunately, he survived and proved his point. I'm sure that all the ulcer patients, who were told that their stomach ailments were a result of their type A personalities, were greatly relieved to find out they could be cured by taking antibiotics.
Identifying uncultured or unculturable bacteria
What do we do when someone has a disease and we can't isolate and grow a virus, bacterium, or other microbe?
We can look for other clues. Sometimes we look for antibodies. The causative agent for Burkitt's lymphoma was discovered because people with this condition have antibodies to Epstein Barr virus.
Now, we also look for nucleic acids such as DNA or RNA and we use the sequences to find out what kinds of microorganisms are present. Tara presented a wonderful description of using metagenomics to sample and evaluate our microbial communities.
In the 80's one of the most amazing stories of identifying and culturing a pathogen, involved HIV, the causative agent of AIDS. HIV was quite difficult to grow in culture. For many years, it was known that people with AIDS also had antibodies to HIV, but it was difficult to prove that HIV caused AIDS until it was possible to grow HIV in culture and evaluate Koch's postulates. There is an interesting story about how researchers were finally able to grow HIV.
But that's a post for another time.