HARVARD GAZETTE ARCHIVES

Todd Golub led a research team that uncovered a connection between cancer and molecules that influence normal growth and development. (Staff photo Rose Lincoln/Harvard News Office)

When miRNAs are present in insufficient numbers, it is believed that cells may divide without proper regulation, a hallmark of cancer. In other words, miRNAs may play a role in keeping cells on a normal growth track.

"Our unexpected findings include the large amount of diagnostic information encoded in a relatively small number of miRNAs," reports Jun Lu, a researcher at the Broad Institute of the Massachusetts Institute of Technology (MIT) and Harvard in Cambridge, Mass. "The implication is that a modest number of miRNAs, about 200, might be sufficient to classify all human cancers."

Lu is the lead author of a report in the June 9 issue of the scientific journal Nature, co-written with colleagues at the Broad Institute, Harvard Medical School, MIT, Dana-Farber Cancer Institute, and the Howard Hughes Medical Institute. Two other reports in the same issue describe discoveries about the potential diagnostic uses of miRNAs. One found that a particular set of miRNAs might be involved in causing a blood cancer known as B-cell lymphoma. A second describes evidence that a well-known cancer-promoting gene activates miRNAs.

Jun Lu (right), Gad Getz, and colleagues worked with small, genetic molecules that they found can distinguish healthy cells from tumors and one type of cancer from others. (Staff photo Kris Snibbe/Harvard News Office)

According to a commentary in the journal by Paul Meltzer of the National Human Genome Research Institute, "These three studies change the landscape in cancer genetics by establishing the specific miRNAs expressed in most common cancers."

RNAs are genetic materials that perform various vital tasks in human cells, including translating instructions coded in our genes (DNA) into the proteins we need to function physically and mentally. Over the past few years, scientists have found that miRNAs exert a key influence on normal human growth and development. Specifically, they seem to regulate the translation of DNA so that cells grow and form organs, such as hearts and brains, in a normal way. When cells don't have the right amount of miRNA, they proliferate in an abnormal, or cancerous way.

This finding raises the possibility that treatments may be developed to interfere with miRNAs that aid tumor development. There is even evidence that some of these molecules may suppress tumor growth.

Colored beads and long nights

Until the Harvard-MIT experiments, Lu notes, "the potential for miRNA expression to inform cancer diagnosis had not been systematically explored." That's what he and his colleagues, who include Gad Getz of the Broad Institute, Eric Miska of MIT, and Todd Golub of Harvard Medical School, set out to do.

"When we started," Lu recalls, "we didn't have much of a detailed idea about what we were going to find (although we did have faith that we were going to find something). It was fully possible that the expression of miRNAs is nothing more than pure randomness, given that there are only about 200 of these little guys known to exist. Indeed, when we started, fellow researchers questioned if we were going to find anything interesting. Happily, things turned out to be just the opposite."

The first task was to develop a quick and accurate way to profile, or determine, the ingredients of all miRNAs. That's a challenging task because miRNAs are so small and similar to each other. (They each involve only 20-25 "letters" of a genetic code that is billions of letters long.) The researchers succeeded by using microbeads coated with colored chemicals that adhere to specific miRNAs. These beads boast the advantage that additional colors can be added to allow detection of newly discovered miRNAs.

"We then had to determine the expression patterns of all known miRNAs across a large panel of samples representing diverse human tissues and tumor types," Lu explains. "This was especially difficult because the samples we had varied widely in quality. It required many nights of work and much computer time to mine out the meaning behind all the numbers we obtained."

The effort was well worth it. As things turned out, samples from colon, liver, pancreas, and stomach tissues reflected their common origin in embryonic stem cells that develop into the stomach and intestines. "This suggests that such clustering of miRNAs is driven by human developmental history," Lu says.

The next thing to determine was whether the miRNAs could actually distinguish tumors from normal tissues. These experiments led to the conclusion that miRNA expression is tied to the state of cell differentiation, the process by which unspecialized cells or tissues change into more specialized types. Examples would be stem cells progressively developing into muscle, blood, or bone cells. From here, it was an easy intellectual jump to the idea that changes in miRNAs associate with abnormal differentiation that researchers call "a hallmark of all human cancers."

Better patient treatments

The investigators next took on the problem of questionable cancers, those 2 to 4 percent of tumors that cannot be clearly identified. MicroRNAs gave the correct answers with far greater accuracy than other methods of diagnosis.

"Even if microRNA-based diagnosis doesn't turn out to beat many routine methods now used to diagnose cancers, we believe that miRNAs have great potential for these difficult and uncertain cases," Lu notes. "Right now, physicians often have to guess about the identity of such tumors, so it's not surprising that the patients have poor outcomes. Our success with miRNAs could change this. It also may help reveal subtypes of certain cancers, potentially leading to more personalized and thus better treatments."

Because of its accuracy and ease of use, the miRNA method could be ideal for use in hospitals and doctors' offices worldwide. "Many current cancer diagnoses depend heavily on the trained eyes of specialists, and, in some cases, even the best-trained specialists will disagree," Lu points out. "This raises the possibility of harmful diagnoses in parts of the world where medical training is not as rigorous as in the United States. We believe that miRNAs could avoid ambiguity and improve accuracy in such cases."

Its low price also makes this technique a candidate for worldwide application. "MiRNA profiling cost less than $15 per sample, at least in our lab," Lu says. "Further experiments that we are planning may bring the price down even more."

Having said all that, don't expect to be able to obtain a cancer test at your doctor's office for $15 any time soon. "Our work is still preliminary," Lu cautions. "The robustness of our results needs more testing by us and other groups on many more samples. The process must be streamlined and further tested at different locations by different hands and on different machines. We hope that these results are as happy as they have been in our lab."