Finding disease by subtraction

The human body shelters a zoo of microbes – thousands of bacteria, viruses, fungi, and protozoa. Some are helpful, some are harmful, many are unknown. It’s a good bet that some of the unknowns provoke diseases whose causes remain a mystery.

How do we find those bad guys? Matthew Meyerson, a pathologist at Harvard Medical School, has come up with a surprisingly simple answer. The Human Genome Project is close to completing an inventory of all human genes. If you subtract this material from a sample of genetic sequences in a patient’s tissue, what is left will be from the germs.

“This is a very powerful method that we can use to find the causes of many diseases that have defied explanation for a long time,” says Meyerson, who works at the Harvard-affiliated Dana Farber Cancer Institute in Boston. “These include juvenile onset diabetes, rheumatoid arthritis, lupus, and even atherosclerosis. It also should be valuable for identifying the causes of new epidemics both natural, like AIDS, and terrorist-made.”

Meyerson came upon the idea by accident. Last year, he was looking for genetic sequences in human cells that are responsible for structures and functions similar to those in bacterial cells. He found three such fragments of DNA, and checked them against sequences mapped by the Human Genome Project. There was no match.

At first he thought this DNA hadn’t been mapped yet. But the odds of finding three pieces missed by the mappers seemed too great. Meyerson then realized that the DNA must be from bacteria present in the human cells he was studying. The way to be sure was to subtract out all the known human DNA from the cells he was working with. What’s left would be nonhuman, ergo would be from the bacteria.

It hit Meyerson that he had stumbled on a way to find all the nonhuman life in the human body. “It’s such a simple idea, I wondered why no one had thought of it before,” Meyerson recalls. “The first thing I did was call all the people who I thought might have thought of it.” No one had.

Finding unknown germs

To test the idea further, Meyerson and his colleagues checked libraries of human DNA and discovered that some of the DNA sequences came from bacteria, viruses, and fungi present in people who contributed tissue samples to the genome project. The microbes included viruses associated with hepatitis, liver cancer, and cervical cancer, a virus involved with skin tumors in AIDS patients, and a menagerie of other bacteria, fungi, and protozoa.

One analysis of 7,000 supposedly human sequences found 44 genes from nonhumans. Further filtering to remove traces of harmless stomach bacteria turned up two sequences that came from human papilloma virus-18, present in cervical cancer cells.

Meyerson points out that the presence of such microbes doesn’t necessarily mean that they cause the associated diseases. “It could be that having the disease makes you susceptible to infections by those micro-organisms,” he notes. “Other kinds of proof are needed to positively link the two.”

Does this mean you soon will be able to contribute a few cells to Meyerson and he could give you a list of all the dangerous microbes in your body? “I don’t see this as a method of screening people for disease,” he answers. “It might be a limit to my imagination, but I think there are faster and easier ways to look for organisms that cause known diseases. It’s the unknown ones we want to find.”

There are dozens of inflammatory diseases and immune system abnormalities for which no cause is known. Take Crohn’s disease for example. It’s a painful inflammation that affects the gastrointestinal tract from the mouth to anus. Finding the microbe responsible for Crohn’s would increase the likelihood of developing new diagnostic tests, and treatments for the disease.

Meyerson’s team has begun examining tissues from Crohn’s patients to determine if he can use his subtraction method to find the microbes responsible for this malady. The researchers are also trying to uncover the causes of multiple sclerosis and lymphoma, a lymph system cancer.

Germ census needed

Meyerson doesn’t know of anyone else doing this work, but he expects others to start soon. He and his colleagues have published a report on their work in the February issue of the scientific journal Nature Genetics.

The main reason he was first, Meyerson believes, was a prepared mind. “Since medical school, the thought of finding the unknown causes of disease has been bubbling in the back of my head,” he says.

Meyerson earned an M.D. in 1993 and a Ph.D. in 1994 from Harvard Medical School. His present position as an assistant professor of pathology and scientific director of the Belfer Center for Cancer Genomics at Dana Farber Cancer Institute gives him access to tissue samples of diseased patients and to the latest data from the Human Genome Project. Two students in Harvard’s M.D.-Ph.D. program, Griffin Weber and Jay Shendure, wrote the software that enables the team to subtract human DNA sequences from those in the blood and tissues of such patients..

In an article accompanying their report, David Relman of Stanford University agrees that “the human body is grossly contaminated with microbes,” but he points out that most of them are beneficial. Separating beneficial from harmful microbes will be a big problem, he notes.

Meyerson agrees that this is “a potential problem, but one that can be overcome.” Finishing the Human Genome Project will be a big help. More than 95 percent of the approximately 35,000 human genes have been sequenced, that is, their chemical identities are known. All the sequences should be complete in two to three years. “That will prevent us from confusing DNA from microbes with unknown DNA from humans,” Meyerson says, “and will make things much easier.”

Relman points out “our level of ignorance concerning the variability of microbial diversity within the human body,” and he cites the need “to embark on a comprehensive inventory” of microbes living in the body and their genes. Meyerson agrees that “the vast majority, perhaps 90 percent or more, of these commensal [harmless] and disease-causing organisms remain unidentified.”

While there is no national effort to take a census of the microbes within us, Meyerson notes that efforts to sequence the genes of various species are going on at a number of public and government laboratories. Also, the need to identify new “Andromeda strains” of natural and man-made organisms capable of starting widespread epidemics is growing in urgency.

Chance favors the prepared mind. – Pasteur