The 28-year wait: Robert Langer on rejection, polymers, and a thousand students

Lex Fridman opens with a light question: Langer has a love of magic, and Fridman wants to know whether he sees a connection to science. Langer does. Magic can surprise you, and science can surprise you, and there is something magical about making discoveries. Fridman offers a duality — the magician knows how the universe of the trick works, while the observer, like a scientist staring at nature, initially knows nothing. Langer's version is simpler: fascination. When he watches magic he is always fascinated, sometimes treating it as a puzzle, but also just registering the sheer fact that something you never thought could happen does happen.

He thinks about that in science too. Sometimes it is something you might dream about discovering, and then you do, in some way or form. His favorite trick is the invisible pack. The spectator is asked to shuffle an imaginary deck, pick a card, turn it upside down from every other card, and shuffle again. Langer then pulls a real deck from his back pocket; one card is reversed, and it is the card they named. The trick is a small model of the career that follows: an idea that looks impossible from the outside, then a physical object that makes it real.

Fridman recites the numbers: more than 295,000 citations, an h-index of 269, one of the most cited people in history and the most cited engineer. Then he asks about failure. Langer's first big work with his postdoctoral advisor, Judah Folkman, was two interrelated discoveries. One was isolating substances that could stop blood vessels from growing, which nobody had done. The other was a way to study that, which required slowly releasing those substances for months, not a day — also never done. They sent one paper to Nature and were rejected, revised it, and Science took it. The other paper they sent to Science, were rejected, and Nature took it.

The rejections were depressing. Langer thought they had done really good work, and Folkman thought so too. Early grant rejections felt the same. The emotions stacked: sad, upset, a little angry because the reviewers did not get it. Then the thought process shifted. Maybe he just had not explained it well enough. "You go through stages," he says, "and so you say well okay I'll explain it better next time." Reviews tell you what they did not like or did not understand; you incorporate that into the next version. That habit, not a claim of indifference, is what he carried forward.

He tells students to do something big, something that can change the world rather than something incremental. How he himself found those ideas was not a rigorous process. It was more spontaneous: exposure, and watching people. Folkman was very good at having big ideas; Langer also met a lot of people who were not. You can spot an idea with broad implications when you see it, while others keep doing derivative work. He has never done it systematically. Some ideas are new creations, like engineering tissues from scratch; others are deep understanding that later leads to technology. The "aha" is rarely instant. Even ideas he really liked took a long time to go from the thought that they might work to knowing that they did.

Fridman, coming from computer science and robotics, finds the human body exceptionally complicated and is amazed anyone can figure anything out. Langer agrees: we are still scratching the surface. There are no great, reliable knobs across populations, though every new drug or device is a small attempt at control, and a biological pathway you can affect might become one. The example he walks through is the postdoc project with Folkman. They wanted drugs that would stop blood vessels from growing, or make them grow. At the time, people did not even believe those things could happen.

A blood vessel — artery, vein, or capillary — provides oxygen and nutrients and gets rid of waste. In cancer, vessels grow into the tumor, which lets it get bigger and metastasize, spreading through the body and ultimately killing someone. The hope was to find substances that could stop that. First they had to develop a bioassay to study blood-vessel growth; none existed. Vessels grow slowly, over months, so they needed polymer systems — plastics or rubber — that could slowly diffuse a molecule at a controlled rate for a long enough time that the biology could actually happen. After they had the polymer and the assay, Langer isolated many molecules from cartilage. Almost all of them did not work. They were fortunate to find one, not yet purified, that did.

That 1976 *Science* paper was the isolation of some of the first angiogenesis inhibitors. Over the next 28 years, other people followed in their footsteps. Different growth factors were isolated, sometimes using techniques Langer's group had developed, and then ways to stop those factors. It took billions of dollars and work by many companies, including Genentech. In 2004, 28 years after they started, Avastin was approved by the FDA. It became one of the top-selling biotech drugs in history, approved for all kinds of cancers and for eye diseases with abnormal vessel growth, such as macular degeneration. Cut off the blood supply and you starve the tumor; it becomes small, and chemotherapy or immunotherapy has a much easier time. In retrospect the idea looks obvious. When Folkman first proposed it, a lot of people thought it was pretty crazy.

Langer was in the dark in more than one sense: he did not know a lot of biology or biochemistry, and much of the work was blind experimentation. Fridman analogizes to quantum mechanics, a theory that works incredibly well without a settled picture of its underlying nature. Are drugs like that? It varies, Langer says. Sometimes people have figured out the pathways they are interfering with; other times it is shooting in the dark; sometimes people make serendipitous discoveries and do not even realize what they did. There is no single way to go about drug discovery. For the blood-vessel case the steps were Folkman's theories, techniques that could study and at least semi-quantitate growth, finding substances that stopped it, then trying to purify them. Before you have an effective drug you still have to show it is safe and effective, starting in animals and then in patients, through multiple kinds of clinical trials.

From that early work Langer helped turn drug delivery into its own problem. The definition is compact: getting a drug to go where you want it, at the level you want it, in a safe way. The hard parts are specific. Could you target just the cancer cell and no other cell? Could you ever give insulin orally, or passively through the skin? Can you get drugs across the blood-brain barrier? Can you make smart systems that respond to physiologic signals in the body? One example: if a person with diabetes has more glucose, could you get more insulin? His lab encapsulated beta cells, the insulin-producing cells, so they are protected; glucose goes in, the cells make insulin. That is a closed loop of materials and cell biology, not a metaphor.

Fridman asks how close these systems are to robots, and how much room there is to add intelligence. Langer is cautious about bells and whistles because cost goes up. They have made intelligent microchips that can release a drug in response to a signal, and he thinks systems like that could someday take an instruction such as "when there is more glucose, deliver more insulin." Robotic devices roaming the body long-term? Someday, he says, but we are not very close. That is nanotechnology, which would mean even miniaturizing what he just described. Fridman also asks where data-driven AI fits in the chain of designing and delivering drugs. Langer sees a role when you have a lot of information. Suppose you have a fast test for disease X, you screen 10,000 chemical substances, a couple work, most do not. The right kind of AI could look at the structures that work and those that do not, find commonalities, and project the next generation of things to test.

On society's relationship with pharmaceutical drugs, Langer does not take a philosophical swipe at over-prescription. Life expectancy and quality have increased tremendously, which he thinks is a really good thing. What would also be good is extending that more to people in the developing world, work his lab has been doing with the Gates Foundation. Reducing costs is a real issue. So is making better drugs for cancer, diabetes, rare diseases, and many other situations where it would be great to do better and help more people.

Tissue engineering and regenerative medicine, as Langer and Jay Vacanti framed them years ago, mean building an organ or tissue from scratch — someday a liver, or new cartilage. The same ideas enable organs on a chip, which might lead to better drug testing and less testing on animals. The structure has two big pieces. One is a scaffold, which Fridman analogizes to a canvas; Langer likes the phrase and says he will use it. On that scaffold you might put cartilage cells, bone cells, stem cells that could differentiate, or more than one cell type. The canvas does not have to live in the body. It can be a structure outside, which is the organ-on-a-chip. Fridman asks whether you could weave in a computational component to sense and control growth. Langer says yes; right now people are mostly validating that the chips work as effectively as putting something in the body, but sensing and control would certainly be possible someday.

Skin has already been made and approved by the FDA for burn victims and patients with diabetic skin ulcers. Phase 3 trials are complete or near completion for making new blood vessels; former student Laura Niklason led a lot of that. Other work is at different stages: helping patients hear better, patients who might be paralyzed, eye problems, liver, kidneys. Neural tissue and even the brain have people working on them; Langer's group has done a little, and collaborator Evan Snyder has done spinal-cord work with neural stem cells. Rejection is a real problem. One path, which Langer and former postdoc Dan Anderson used for a painted pancreas, is encapsulating cells so immune cells and antibodies cannot get in. Other strategies include making cells less immunogenic by masking or gene editing, using the patient's own cells or a close relative's, and immunosuppressive drugs — currently what is done for a liver transplant.

Will we one day regenerate any organ? Langer is an optimist, and he flags the timeframe: "someday" could be hundreds of years. He thinks we will be able to regenerate many things, using cells themselves or molecules that help regenerate cells. On longevity ten to twenty years out, he does not expect a radical increase. In certain areas people will live better lives, and maybe somewhat longer lives.

Asked for the most beautiful scientific idea in biology or engineering he has come across, Langer names CRISPR. He is careful to say it was not his idea. What he finds interesting is that people capitalized on a mechanism by which bacteria destroy viruses; understanding that led to machinery that can cut and paste genes and fix a cell. Copy-and-paste in a body this complicated is exceptionally difficult, which does not mean it will not be done. A lot of companies and people are trying, and in some areas it will be.

One way to lower the bar is not to fix the whole body directly. You could take a cell that might be useful and give it cancer-killing capabilities — a CAR-T cell — and maybe make that better. Medicine has moved stepwise. Fixing a brain would be very hard; maybe someday that will happen too. Fridman asks whether medicine and bioengineering proceed by big leaps every decade or by a lot of incremental work. Langer thinks there is both. Every so often a new technique comes out — genetic engineering, CRISPR — that makes a big difference. To actually make a new drug or device, there is still a lot of work that needs to be done.

Langer has more than 1,100 current or pending patents, licensed or sublicensed to over 300 companies. For the most part he sees strengths in the patent system. If you did not have patents, especially in medicine, you would never get the funding it takes to make a new drug or device. According to Tufts, making a new drug costs over $2 billion, and nobody would come close to giving you that money if anybody else could then copy it. That is also the drawback. Once somebody has a very successful drug, you want to make it available to everybody, and the same system that allowed it to happen in the first place may impede that afterward, at least for some people or some companies.

The most expensive part of that $2 billion is clinical trials, by far, in money; pain is harder to score. Proving that something new is safe and effective in people is almost always the biggest expense. The process, in general, starts with cells, then a couple of animal models that are not perfect for humans, then at least three sets of clinical trials: phase one for safety in a small number of patients, phase two for effectiveness in a small number, phase three for safety and effectiveness in a large number — hundreds or thousands — in carefully controlled studies. You have to manufacture the drug, watch the patients, and show that it treats the disease better than whatever the gold standard was. "First do no harm," Fridman says. "First do no harm," Langer agrees. Show that it is safe first, then that it is effective.

In practice at MIT, they usually do a certain amount of research, make findings, prove a hypothesis or invent a technique, and write a disclosure for the technology-transfer office. That office gives it to patent attorneys, who talk to the inventors and draft a patent. Then you go back and forth with the USPTO. They may not allow it the first, second, or third time, but they will tell you why, and you may adjust. Maybe you eventually get it, and maybe you will not.

Langer has been part of launching 40 companies, together worth an estimated $23 billion at the time of the conversation. Asked for a formula, he breaks it into categories and then undercuts the scientist's instinct. From the science standpoint he will go over the checklist, but he actually thinks the most important thing is probably the business people he works with. When he looks back at the companies that have done well, it is because they had great business people; when they have not done as well, they have not had as good business people.

On the science side he wants a discovery that is almost a platform: you could use it for drug A, B, C, D, E, and so forth, the way the drug-delivery systems were. He wants it taken far enough that they have written at least one really good paper in a top journal, reduced it to practice in animal models, and filed patents — maybe issued patents — with what he calls very good and broad claims, which is the key in a patent. A lot of times someone in the lab, a postdoc or graduate student, spent a big part of their life on it and wants to work at the company because they have a passion to see something they did make a difference in people's lives.

The business instincts that matter: you have to be a good judge of people so you hire the right people; you have to be strategic, because medical research is so expensive and a platform that could do many things still has to be sequenced first, second, third. You need an FDA regulatory and clinical-trial strategy. You have to be able to raise money, incredibly. You have to be a good manager of people. Deciding which indication to do first is not just science. Langer lists something like six or seven criteria: is the market big or small; are there good animal models so you are not waiting fifty years; can the clinical trials be set up with clear endpoints; what is the competition; can you get reimbursed; are there manufacturing issues. Scientists are part of that process. They are not the whole of it.

Langer leads one of the largest academic labs in the world, with over $10 million in annual grants and over 100 researchers, probably over a thousand since the lab began. Fridman notes that researchers can be individualistic and eccentric. Langer's management philosophy sounds, he admits, very simplistic: he wants people in the lab to be happy, doing things that will make the world a better place, working on science that can make the world a better place. If they can do that, it kind of runs itself. What makes a researcher happy is equally plain. If people feel they are working on something really important that can affect many other people's lives, and they are making some progress, they will feel good about themselves and they will be happy.

Big questions arrive from many directions. Sometimes Langer thinks of them, or somebody in the lab does — a new technique, or a desire to understand something better. The Gates Foundation and the Juvenile Diabetes Foundation have also come and asked whether the lab could help. Inspiration, for him, is talking to people, going over their ideas and their progress, and being honest about his own setbacks. People who know him know he had them. You just try to keep pushing. On the larger landscape of science, he thinks going to industry is good and being in academia is good; lots of his students have had great careers doing both. The biggest academic concern at a place like MIT is funding, particularly funding that is not super directed, so you can do basic research. That is probably the number one thing.

The last ten or fifteen years the lab has done pretty well on funding, but he always worries. They operate on more soft money than hard money. The Gates Foundation, the Juvenile Diabetes Foundation, some companies, and NIH grants have all mattered. He just views the worry as part of the process. Fridman puts him in the shoes of a philanthropist with $100 billion that cannot go to his own research. Langer's first move is a concrete vision. David Koch was interested in cancer research, in part because he had prostate cancer; many donors follow a disease they or someone they loved had. Bill Gates, once he had his fortune, asked how he could have the greatest impact and landed on people in the developing world, medicines, vaccines. Once you have the vision, you ask who in the world does the best work toward that goal. He adds a tension that is easy to miss: you also want to help young people, and that may be at odds with funding an older lab that is very good at the problem. Government, disease foundations, and other foundations coming at it from different angles — seeking people out, or putting out a request for proposals — is the healthy mix. Koch and Gates did both.

Langer's own early faculty search was hard because he did not fit a bucket: not quite chemical engineering, not quite biology. How do you make room for people who are brilliant but outside last decade's disciplines? Department heads have to have a vision, and some of them do. MIT's mechanical engineering department hired Giovanni Traverso, a molecular biologist and gastroenterologist who had also done great mechanical engineering designing new pills. Langer gives them a lot of credit; he thinks the department will be the richer for it. The Media Lab hired people like Ed Boyden who work on very different things. That, he says, is part of the vision of the leadership.

Will we one day cure cancer? Langer thinks so. He does not know how long it will take, and he does not think it will come soon. Very few things are solvable in a few years, and "all cancers" is pretty tough. If a cure arrives, he thinks it will be a combination of biology and engineering: biology to understand the right genetic mechanisms, and maybe the right immunological ones; engineering to produce the molecules, develop the right delivery systems, and target them. That is a concise description of how he has spent a career — not as a biologist or a chemist, but as someone who brought engineering methods into medical problems.

On a lighter topic, Langer loves chocolate. Fridman had been staring at pictures of Soho Globs from Rosie's Bakery in Chestnut Hill at 3 a.m. instead of finishing a paper, and he asks whether we will one day engineer the perfect cookie with chemistry and a bit of data-driven AI, or whether cookies are more art than engineering. Langer thinks there is some of both, and that engineering will probably help someday. Same for chocolate. He points to former postdoc David Edwards, a Harvard professor who started Café ArtScience and companies that look at fragrances in new ways. He expects that someday AI and engineering will play a role in almost everything, including creating the perfect cookie.

Looking back at a life of scientific and engineering impact, what is he most proud of? "My students." Close to a thousand have gone through the lab. He thinks 18 are in the National Academy of Engineering, 16 in the National Academy of Medicine; they have been CEOs of companies and presidents of universities; eight or so are faculty at MIT, maybe about 12 at Harvard. They are not his children, but they are close to his children in a way. It makes him feel really good to see them have such great lives, do so much good, and be happy. The citations, the patents, the companies, and the market caps are the public scoreboard. What he puts last, and first, is the people.