Engineer Sets Aside One Distinguished Career to Plunge Into Cancer Research

People often say that a book changed their lives—for some it might be "To Kill a Mockingbird," for others it’s Plato's "The Republic." For Dimitris Anastassiou, it was a biology textbook.

Andrea Retzky
March 22, 2012

A dozen years ago, after nearly 20 years in electrical engineering at Columbia, Anastassiou decided he was ready for a change. His successful patents in international digital television standards, used in all Blu-ray discs and TV receivers, generated significant revenues to Columbia and to his lab, which gave him the flexibility to explore research in an unexpected area—computational biology.

“I abandoned my first career—it made me feel young to try something totally new,” said the Greek-born Anastassiou, the Charles Batchelor Professor of Engineering. “And there was a part of me that always wanted to be a scientist, too.”

He started his new career in computational biology after spending a year-long sabbatical teaching himself molecular biology. In 2006, he came across "The Biology of Cancer" by MIT professor Robert Weinberg, a pioneer in cancer research. That book inspired him to pursue a new direction. Having lost both his parents to the disease, he found the opportunity to make a difference all the more compelling.

The explosion of “big data”—and new means of managing and mining the massive amounts of information drawn from data-intensive technologies—offered rich sources of information related to cancer. Most of it is available on the Web for free, creating a unique opportunity for computational scientists like Anastassiou to analyze and identify patterns and associations, shedding new light on the biological mechanisms that cause cancer.

Anastassiou was particularly interested in the genetic signatures of invasive tumors because most cancer deaths result from cancers that eventually metastasize. Weinberg, the textbook’s author, had described a process called epithelial-mesenchymal transition (EMT) by which cancer cells undergo changes that make them migratory and invasive. While research results were always assumed to be specific to each cancer type, Anastassiou wondered if something more universal could be happening.

Through the application of his pattern-finding algorithms, he found a precise signature of EMT‑associated genes that was present in two cancers he examined—ovarian and colon—when a particular stage of invasiveness was reached. He then found that precisely the same signature was present in all kinds of solid cancers: breast, prostate, kidney, melanoma, head and neck, among others. These computational results were published in the journal "BMC Medical Genomics."

Anastassiou then partnered with biologists at Columbia's College of Physicians & Surgeons to conduct research on mice. That research led to the discovery that most of the genes of the signature were expressed by the tumor itself, and not by the surrounding tissue. This proved that, indeed, the invasive cancer cells themselves had undergone this radical transformation inside the living animal. It also proved that this phenomenon applies to even more cancer cell types than was believed to be the case. The journal BMC Cancer recently published these findings.

In more recent research, just published in "PLoS ONE," Anastassiou again looked at publicly available data, this time about glioblastoma, a kind of brain tumor. He examined the time it took for tumors to recur and found that all patients who went for longer periods until recurrence had low levels of the same signature. Anastassiou says that these results are consistent with Weinberg’s theory that this mechanism causes cancer cells to act like stem cells—both self-renewing as well as tumor-seeding cells—therefore making them more resistant to therapy. Patients who don’t have the signature would have fewer such cancer stem cells and therefore a smaller chance of recurrence.

“We have an opportunity to try to inhibit cancer invasiveness or recurrence by targeting the genes that could be causing this mechanism and interfering with their process. Now we want to collaborate with a pharmaceutical company to find or develop compounds that can do so,” said Anastassiou.

“I think this could be much more significant than the work I did in my first career,” he said, adding, “Human health is far more important than consumer electronics.”