Personalized cancer medicine has advanced from a distant hope to a clinical reality. Oncologists regularly individualize treatments to target a tumor’s unique genetic weaknesses. But because these personalized medicines reach healthy tissues and tumors alike, even the most targeted treatments can cause unwanted side-effects.

A new approach devised by nanotechnology experts at the Sloan Kettering Institute (SKI) at Memorial Sloan Kettering Cancer Center may improve the precision of personalized medicines by helping them avoid collateral damage. “We found a way to use machine-learning algorithms to design powerful nanomedicines that can deliver a stronger, safer, more personalized punch,” says Daniel Heller, PhD, a chemist in the molecular pharmacology program at SKI and an assistant professor at the Weill Cornell Graduate School of Medical Sciences.

Heller’s laboratory designs nanotechnologies that improve the detection and treatment of cancer. Their computer-aided method allows nearly any class of personalized drug to be packaged into a nanoparticle, an approach that could help make these drugs safer and more effective.

Particle Man: Memorial Sloan Kettering chemist Daniel Heller, PhD, with 3-D-printed models of his cancer-fighting nanoparticles. Credit: Memorial Sloan Kettering

Nanoparticles are objects with diameters less than one-thousandth of a human hair. Scientists have previously used nanoparticles to deliver drugs in the body. But so far, it has been difficult to design different nanoparticles to carry many of the newer types of cancer medicines because each new drug/particle combination requires a lot of work to develop at the lab bench. “Machine learning has been enlisted to do amazing things, like steer a self-driving car or help scientists find patterns in large amounts of human genome data,” Dr. Heller says. “So we thought it might help us design better nanomedicines.”

Using Chemical Structures to Make Predictions about Nanoparticles

Dr. Heller and a postdoctoral fellow in his lab, Yosi Shamay, PhD, observed that some drugs easily assemble into nanoparticles, while others do not. They wondered if they could determine which of a drug’s qualities influence its ability to form a particle and used machine-learning algorithms to churn through many types of data and identify common elements among good particle makers.

Yosi Shamay, PhD, postdoctoral fellow, Memorial Sloan Kettering. Credit: Memorial Sloan Kettering

“It turned out that we could predict with remarkable accuracy whether a particle would form, based on the structure of the drug,” Shamay says. He and Heller could also run the method backwards to design a drug molecule that more easily assembles into a nanoparticle. Previous nanoparticles mostly consisted of a polymer, which served as a kind of glue, with a little of the target drug. The new particles consist of up to 90% of the drug, making them much more efficient.

Helping Cancer Drugs Hit Their Molecular Mark

The SKI scientists demonstrated the effectiveness of their nanoparticles using drugs called MEK inhibitors. These precision medicines are an effective treatment for certain types of cancer, including melanoma and lung cancer, which share certain mutations. But MEK inhibitors can also harm the normal cells in the skin, causing a severe rash—limiting the amount that can be administered.

In 2018, Heller and Shamay published the results of a study done in collaboration with the laboratory of John Chodera, PhD, a computational chemist at SKI, and Scott Lowe, PhD, a cancer biologist and Chair of the Cancer Biology & Genetics Program at SKI. The team packaged MEK inhibitors into nanoparticles with an affinity for a particular structure found on tumor blood vessels. Using mouse models of cancer, they found that the nanoparticles homed to these sites, hitting their target without affecting the skin.

“If we target these new drugs directly to the tumors and away from the normal tissues, they will work a lot better, and the side-effects would not be so bad,” Heller says. “That’s what we hope this nanotechnology will enable.”

Monitoring Cancer Activity

Heller's lab also focuses on cancer detection. Along with colleagues, he has shown that nanotechnology can provide information on disease biomarkers, which are molecules that indicate biological activity, such as whether a tumor is growing.

Implantable nanosensors could offer a method to monitor the effectiveness of a cancer treatment without repeated biopsies and imaging scans. Heller uses sensors made from tiny, carbon-based rods called nanotubes that bind to particular biomarkers, emitting specific wavelengths of harmless infrared (IR) light as a result. IR can travel through several centimeters of tissue and still be readily detected and analyzed, so sensors could be implanted under the skin. A small device worn on the wrist could detect the sensor’s signal—acting as a kind of fitness tracker for disease.

The technique, which was reported last year in the journal Nature Biomedical Engineering, may reveal valuable information about disease activity in the body. The sensors could detect cancer in people at high risk, or in patients who have been successfully treated but have a strong chance of recurrence. In those who are undergoing treatment, the sensors could immediately report whether a biomarker is going up or down and—when needed—help a doctor know to switch from one drug to a second-line therapy.

Just recently, Dr. Heller and a postdoctoral fellow in the lab, Ryan Williams, PhD, along with colleagues from New York University, showed that a similar implantable nanosensor could detect biomarkers of ovarian cancer in animal models of the disease, findings they reported in the journal Science Advances.

“This may sound like science fiction, but we're working to make it reality,” Heller says. “We aim to develop some of our most promising advances and bring them to clinical trials.”