Med student plays a role in pivotal disease research
by C.W. Wolff

Biomarkers are very small proteins in blood that can signal the early stages of a disease. Finding them, however, has been like seeking needles in haystacks. To make the search easier, scientists routinely remove what they consider “trash” proteins from blood samples.

Arpita I. Mehta, M05, and Dr. Lance A. Liotta, head of the Laboratory of Pathology for NIH. © Kay Chernush for HHMI

Apparently, though, the needles were being thrown out with the hay. Arpita I. Mehta, a third-year student at the School of Medicine, has helped to discover that there’s an ocean of thousands of previously unnoticed biomarkers bound to those trash proteins.

That research, published in February in Disease Markers, with Mehta listed as corresponding and lead author, promises a new and easier method for detecting diseases such as ovarian cancer. It also opens the door to a range of other fine-tuned, blood-based analyses, including the ability to stage and type various diseases.

Mehta interrupted her Tufts medical school studies during the 2002–03 academic year to work as a Howard Hughes Medical Institute/NIH research scholar at the sprawling, 1,500-lab National Institutes of Health complex in Bethesda, Md. She applied for the program at the last minute after hearing about it the day before applications were due. After she was accepted—one of only 40 medical and dental students chosen nationally for the program (Matthew Rand, a Tufts dental student, also was selected)—she was anything but haphazard.

Before medical school, Mehta was a mathematician for Motorola in Chicago, her hometown. She developed models for communication networks and scouted university graduate math programs for new ideas. She knew she wanted to use mathematics to solve biomedical problems, preferably problems that had strong clinical implications.

So she interviewed 35 NIH laboratory directors before meeting Dr. Lance A. Liotta, head of the Laboratory of Pathology for NIH and co-director, with the Food and Drug Administration’s Emanuel Petricoin, of a clinical proteomics program.

“I didn’t realize how ‘big’ Dr. Liotta was at the NIH when I interviewed,” says Mehta. “I liked that he had done some theoretical math. I also liked his straightforward honesty and his confidence in my ideas.”

It was a good choice for both of them. The year’s work in both biomarker physiology and a new approach to individualizing cancer therapies resulted in Mehta’s name gracing two patents, seven scientific papers (four published, three pending), a new approach to diagnosing ovarian cancer and a pending clinical trial.

Liotta and Petricoin had developed an alternative way to use blood to detect disease in its early stages. Instead of seeking a single, elusive biomarker, they used mass spectrometry to reflect the full serum proteome and then analyzed, compared and taught a computer to recognize different patterns.

While this approach was successful, the researchers were not completely sure what they were looking at in the patterns.

So Mehta settled down at a desk next to Liotta’s in his 200-person laboratory, and at his suggestion, developed a mathematical equation to determine if the peptides reflected in the mass spectroscopy graphs were actually biomarkers shed by diseased cells.

The equation calculating the journey of tiny biomarkers into and out of the blood worked when compared with previous studies, Mehta says, except for the biomarkers’ concentration levels. Those remained much higher than predicted.

“I realized the proteins might be bound to larger, carrier proteins that circulated longer in the blood,” says Mehta. “It seemed logical. It also seemed logical that they might be assuming the much longer half-lives of those larger proteins.”

That hunch proved true. It explained why free-floating biomarkers were so hard to find because the traditional method of seeking them included first purging the blood sample of big carrier proteins. The mass spectrometry analysis uses unmodified, whole serum to capture all of the concentrated, bound proteins.

To further evaluate the method, serum samples from ovarian cancer patients were tested. Only the proteins bound to the carrier protein, albumin, were analyzed. The results matched previously reported, full-blood patterns and proved 100 percent predictive of ovarian cancer in the samples tested.

Last fall, Mehta returned to the jam-packed third year of medical school, eager to work with patients and continue training to be a doctor. She stays in touch with her NIH colleagues and expects that some day, she will do more research. Right now, she says she is glad to be back in medical school: “I like variety.”