James Bradner, M.D., is President of Novartis’ Institutes for BioMedical Research, one of the  world’s largest pharmaceutical research organizations. He developed the development of the JQ1 molecule, which inhibits proteins associated with certain cancers, and his decision to share the molecule freely with researchers all over the world, including a guide on how to synthesize and use it. Throughout his career, he has advocated for open-source science, which promotes free dissemination of scientific data, methodology and literature.

You are a major proponent for open-source science. Can you explain what open-source science looks like and how you think it will change the biomedical world?

I have tried to exemplify the principles of open source, but applied to the typically very private world of therapeutic science [and] drug discovery. In training as a cancer doctor, it occurred to me that a potential roadblock between innovation and the patient is extraordinary secrecy. Working at a cancer laboratory, I had a great opportunity to apply principles of open source to drug discovery. Like with any experiment, we implemented a design principle — we would download the best practices of the software industry and adapt them to biomedical research. Fast forward six years, and we have witnessed a dramatic acceleration in the pace of discovery research and, moreover, have seen an acceleration in the understanding of the fundamental biology.

The cost of medical treatment is increasingly becoming a prohibitive factor for many patients in the United States. Will open-source science affect healthcare costs?

There is no question that the rate of healthcare expenditure in the United States — if not worldwide — is rising at an unsustainable clip. Drug costs are an important contributor to that, but there are many inefficiencies in the administration of healthcare, even beyond medicines, which are, I have learned, a minority contributor to costs. Open source can, in the fullness of time, improve the efficiency of healthcare delivery, and perhaps even bring down drug prices. An unexpected learning of our experience with the JQ1 molecule was how quickly pharmaceutical and biotech companies would jump on this idea [of open source]. Within four years of publishing the source code for the JQ1 molecule, we found 79 patents from 29 different groups. People should pursue original ideas, not low-hanging fruit, but, for patients, this may be a good thing. Having more than one drug on the market very likely will contribute to lower drug prices.

Since sharing JQ1 and moving into industry research, what advances have you made in your research that excite you?

I transitioned from the Dana-Farber Cancer Institute and Harvard to Novartis two years ago, so it’s early days, but I have already seen remarkable achievements in therapeutic science since joining Novartis. Last year, we had the very first approval for a brand new type of medicine, called the CAR T-Cell, for children with acute lymphoblastic leukemia. This medicine is very creative: It’s a T-cell isolated from the patient into which a new gene is inserted that turns that T-cell against the patient’s cancer. This is a living medicine that is infused, alive, back into the patient’s vein, where it multiplies logarithmically, attacking cancer as it goes. Though I didn’t remotely contribute to inventing this medicine, it was a real privilege to bring the medicine forward for these patients first, and, behind it, to build a platform for new, disruptive therapeutic technologies for other types of B-cell cancers and, in the fullness of time, solid tumors, as well.

The experience of delivering CAR T-cell therapy to patients was an early, powerful experience in my time in pharma. I spend most of my time at Novartis dreaming up the next great therapeutic technology, [but] I also lead a research effort to invent a new type of drug that does not just bind its protein target — but that binds and destroys its protein target. We call this targeted protein degradation. This thread of science is new, but now, just two years in, we have degraded or destroyed 50 very compelling new therapeutic targets.

How long does it take to fully develop a drug, from basic research all the way to human clinical trials?

The delivery of a disruptive new therapeutic idea, all the way to FDA approval and broad use throughout the healthcare community, is slow, expensive and highly variable. On average, successful drug discovery and development projects take 10 to 12 years and cost one to two billion dollars. Those are astonishing numbers, still, to me, coming from an academic, NIH-funded environment, but they are accurate. We have some projects that may take two decades to mature, and others that could move much more swiftly, on the order of three to five years. You have to kiss a lot of frogs in this line of work, because the research is so uncertain. Industrywide, maybe 10 percent of projects initiated will ultimately benefit patients. The most ambitious projects are the ones that, regrettably but realistically, are the most uncertain.

You were quoted in Nature saying that any distinction between academic and industry research is a false distinction. You said this just before you joined Novartis. Have your views changed now that you have worked extensively in both academic and industry contexts?

I believe the distinction between academia and industry, in biomedical research, is a hand overplayed by many. As scientists, we cannot but help ourselves to identify distinctive phenotypes and apply tests of significance and tease out the differences, and indeed there are differences. But the commonalities are truly overwhelming. Academic biomedical research has transitioned to be much more translational: There’s still a foundation of basic science, but NIH grant requests require a page and a half on significance to humanity. This has moved the pendulum strongly toward understanding human biology. This shift has been momentous, even over the past 10 years. When I was training as a chemist, the faculty at Harvard chemistry couldn’t be bothered to make a therapeutically-relevant molecule. Now, storied academic institutes have drug discovery centers. I regard this as a good thing because science no longer has to be irrelevant to be basic. You can ask very fundamental questions in yeast, but now with CRISPR/Cas9 and sophisticated small molecules you can ask these same questions in pluripotent stem cell-derived cortical neurons. On the flipside, pharmaceutical companies have become much more interested in basic science. As academia swings more translational and we swing more basic, we meet in the middle.

This interview has been edited for length and clarity.