Science, Medicine, Technology

  Will Kirk/Johns Hopkins University

Eye, robot

Sometimes it’s not what doctors know, but how deftly they can handle surgical tools that determines the limits of medicine.

If surgeons’ hands were perfectly still, for instance, they could inject clot-busting drugs into the eyeball’s tiniest veins, which can clog as we age, obscuring our vision. And if their tools could twist into an S-shape, they might navigate more nimbly the confined regions of the throat, where the removal of a tumor can easily damage vocal chords.

To help doctors perform these kinds of delicate operations safely, Nabil Simaan, a 35-yearold Columbia assistant professor of mechanical engineering, has developed small, snakelike robots to handle surgical tools in tight quarters. “They’re made for spaces where doctors have trouble moving their hands and where lots of tiny anatomical parts can easily get damaged, such as inside the eye, ear, and throat,” says Simaan.

The robots are at least two years away from testing on animals or cadavers, but they’re so promising that Simaan and Columbia University recently launched a start-up company, Auris Technologies, to raise funds to perfect the technology and eventually bring it to market. The Massachusettsbased biotech investment firm Medical Capital Advisors has joined the venture, one of about 10 start-ups launched last year by Columbia’s technology transfer office, Science and Technology Ventures.

The robots’ arms are made of nickel titanium, a flexible alloy, and they’re tiny — ranging in width from one half a millimeter to four millimeters, depending on the surgical application. They can be equipped with lasers, tiny needles, grippers for tying sutures, sensors for determining how much resistance tissues exert, lights, and cameras. Doctors can control the arms using a joystick while viewing a three-dimensional image of the operating site on a computer workstation.

“Today, even high-tech laser surgeries are performed with handheld tools that are inflexible,” says Simaan, who directs the Advanced Robotics & Mechanism Applications Research Laboratory at Columbia’s engineering school. “That makes it difficult for doctors to work in certain areas of the body. Doctors are also constrained by the innate tremors in their hands, and they typically can do the most high-precision tasks with only one of their hands. With our robots, doctors will be able to perform high-precision manipulations ambidextrously, using more than one tool at a time.”

The robots’ arms are incredibly nimble because a computer fine-tunes their positioning 1000 times per second. “Our goal is to achieve precision to the nearest five microns, or about 20 times more stable than a surgeon’s hand,” says Howard Fine, a Columbia ophthalmologist and assistant professor who’s helping Simaan design a robot for eye surgery. “Doctors could then inject drugs into smaller veins than ever before. That would have implications for vascular microsurgery throughout the body, not just in the eye.”

Simaan is also developing robots that could guide cochlear implants into the ear and help doctors remove tumors near the vocal chords. The robot for throat surgery, which is pictured at left, is being built with computer science professor Russell H. Taylor at John Hopkins University, where Simaan began designing his robot as a postdoctoral researcher in 2002. Simaan is also working with Columbia professor Dennis Fowler, a pioneer of minimally invasive surgery, on a robot for abdominal procedures.

“The robot’s small size and dexterity will allow us to enter the body through a single incision in a natural orifice, whereas minimally invasive surgeries today require at least three access ports,” says Fowler, who’s vice president and medical director for perioperative services at Columbia University Medical Center. He made national headlines last year by removing a woman’s gallbladder through her vagina.

“We could use the technology to remove an appendix, a kidney, or a gallbladder,” Fowler says. “And it will mean less pain, less trauma to the body, and easier recovery.”

To learn more, visit


Closing in on breast cancer

Columbia oncologists have demonstrated for the first time how a mutated gene associated with breast cancer actually contributes to runaway cell growth, a finding that could lead to new treatments. Scientists first linked the gene BRCA1 to breast cancer in the 1990s, but until now they haven’t understood how it works.

The oncologists, led by Columbia professor Ramon Parsons, showed recently that BRCA1, whose job is to fix routine DNA damage in other genes, often fails to repair damage in a gene called PTEN when BRCA1 itself mutates. The gene PTEN, which Parsons discovered in 1997, is known to increase dangerously the activity of proteins that trigger cell growth.

“Ever since the link was established between BRCA1 and breast cancer, we’ve been frustrated by our lack of understanding about how its mutations work,” says Parsons, who directs the Avon Foundation Breast Cancer Research Laboratory and the Breast Cancer Program of the Herbert Irving Comprehensive Cancer Center at Columbia University Medical Center. “Now that we know it acts on PTEN, we have a target for therapy.”

Breast cancers associated with the BRCA1 gene account for about 10 percent of all cases and tend to be aggressive and tricky to diagnose. Parsons estimates that PTEN will be found to be involved in about half of BRCA1 breast cancers once a complete chromosomal analysis is done.

Columbia oncology student Lao Saal ’07GSAS, ’09PS was lead author of the paper, which appeared in the January issue of Nature Genetics. Other collaborators included the oncologist Åke Borg at Sweden’s Lund University.

  Steve Gschmeissner/Photo Researchers, Inc.
  Yeast cells reproduce by elongating and then growing a dense ring of protein molecules that eventually contracts, pinching each cell in two.

Divisions of labor

For nearly 100 years, scientists have been trying to understand the reproductive mechanisms of individual cells. A breakthrough came in the 1970s, when cells were shown to form a ring of proteins around their membrane to squeeze themselves in two, using a cinching motion like that of an old-fashioned purse string.

But how do cells create the so-called contractile ring? Columbia engineers and physicists have shed light on this mystery, working with cell biologists at Yale and Lehigh universities. The biologists recently observed a ring’s formation by illuminating yeast cells with fluorescent agents and then monitoring the subcellular activity on video. They watched as proteins around the cells’ equators sprouted long filaments, which appeared to form solid connections with other proteins, pulling the pairs together. The process occurred countless times over a few minutes, forming a dense protein ring around each cell.

But when Columbia researchers tried to simulate the construction process with computer modeling in order to test its physical feasibility, the results were a flop: The rings came out crude and dysfunctional. That inspired the “aha moment”: Maybe filaments don’t connect proteins permanently, but pull the proteins toward each other only to quickly release them, allowing them to swap partners in a group dance that creates a ring of ever-denser mass. The Columbia team ran computer simulations of this scenario and found that a connection time of 20 seconds was optimal to build a ring, at least theoretically.

“That’s the power of computer modeling,” says Ben O’Shaughnessy, a chemical engineering professor and trained physicist who led the Columbia researchers. “You can test lots of different mechanisms with relative ease and speed, and then suggest new directions for realworld experiments.”

The biologists then zeroed in on the precise movements of the proteins and found that, indeed, they shifted directions. More detailed video observations corroborated the suspicion: The filaments broke their connections every 20 seconds. “It’s as if evolution has fine-tuned the mechanism to use reaction rates nearly identical to what is ideal,” says O’Shaughnessy.

“Future work will involve testing the concepts learned from yeast in other cell types to learn if the mechanism is universal,” says Yale biologist Thomas Pollard, a coauthor of the paper, which appeared in the January issue of Science. “Other cells, including human cells, depend on similar proteins for cell division, so it’s entirely possible that they use the same strategy.”