welcome!

New to Close Concerns?

Subscribe

Already registered?

Log in

WHAT WE DO

consulting
strategic advice and comprehensive information for those interested in the diabetes industry

diabetes close up
our monthly newsletter

industry reports
periodically we take the long view on trends or on specific breakthroughs.

TEAM

about us

John Eng on the Discovery of Exendin-4

John Eng discovered Exenatide (exendin-4), which was developed by Amylin and Lilly and received FDA approval April 29, 2005. It is now known as Byetta. Recently, we had the opportunity to sit down with John to talk to him about the events leading up to the exendin-4 discovery. John, who has spent his career at the Veterans Administration Medical Center in the Bronx, New York and is now Director of Clinical Informatics there, spent a generous couple of hours with us. The following are his comments on a range of topics:

On being in endocrine: I’m a physician. I trained in internal medicine and clinical endocrinology and wanted to do endocrine research. I started out in Rosalyn S. Yalow’s lab at the Bronx VA and worked in her lab for the next 18 years until she retired. I learned that it is important to have a mentor in science. She was a wonderful mentor… She considered all of us, her fellows, as her scientific children. What I learned from her is not so much technical skills but rather a frame of mind to be persistent and tough during the search. Research is highly refined searching; search is after all part of the word research; and it's an endless search. As Thomas Edison said, science is 99% perspiration. I estimate that 99% of research findings are not earthshaking or new, but every stone still has to be turned over. It's the remaining 1% where you might find something interesting or new. I think that's what I really learned from Ros, the motivation to continue the search without becoming discouraged … She would challenge us in our weekly conferences by saying, “Think big thoughts. What have you learned?”

On one step forward and finding new hormones: “Okay, I'm going to take one small step forward” – in science there is a fair amount of that. There is a necessity to publish papers. Very often, the publications are small increments rather than really challenging discoveries, so it's important to encourage young people – like me, at the time – to think beyond science by smallest acceptable increments--to think big. My big thought was to say, “Since we deal with hormones in endocrinology, are there any new ones to be found?” I thought there are new hormones to be found.

About the time when I started working in Ros Yalow’s lab, a group in Sweden at the Karolinska Institute published, what to me were, very exciting papers. They described a new way to look for peptide hormones. Their line of thinking was this: peptide hormones often have certain chemical characteristics. If you look for those chemical characteristics you’ll find hormone candidates. It doesn’t assure you they are hormones, but the presence of those chemical characteristics greatly increased their chances. Victor Mutt and Kazuhiko Tatemoto in the Karolinska group purified, isolated, and published papers on a number of new hormones from both the intestine and the brain.

It turns out that there’s a commonality between hormones in the intestine and hormones in the brain, which has been called the brain-gut axis. Why is that? No one has really come up with a universal theory, but it’s been known for many years now.

The Karolinska group proved their concept by isolating new hormone candidates. In fact, one of the peptides they discovered is PYY, one that Amylin itself is hoping to develop.

The tool they used to isolate peptides is called a chemical assay. Ros Yalow was awarded the Nobel Prize in 1977 for discovering radioimmunoassay. However, radioimmunoassay will only detect what it was created originally to detect, unlike chemical assays, so the likelihood of finding something new by radioimmunoassay is small.

But chemical assay opened up the possibility of finding new things. In fact, Mutt and Tatemoto demonstrated that it can be used to find new hormones… They isolated a number of new peptides with unique chemical signatures. They were then faced with the need to prove the peptides are hormonal in nature – that’s one of the anticipated requirements if I were to use a chemical assay.

I said, “I think there are new hormones. I’m going to look for them with a chemical assay.”

[Interjection from Kelly Close] Wow, that’s a big thought!

Well, yeah! Ros said, “Think big thoughts,” but the corollary to that is the bigger your thoughts the harder you have to work!

Peptides are chains of amino acids. One end is the amino end, and the other is the carboxy end. Or in other terms, the N-terminal and the C-terminal part of the peptide chain.

Mutt and Tatemoto looked at the C-terminal part for C-terminal amide as a chemical signature. They had a well-developed assay, and it didn’t make sense for me to try to emulate them.

So I looked at the other end of the peptide, the N-terminal end. It turns out that there are many gut and pancreatic hormones that have the amino acid histidine at the N-terminus or the very first position of the peptide chain. Classical hormones like glucagon, VIP, secretin, all have histidine at the one position. So I said, “Hmm, that’s a good signal for a chemical assay.”

I set up an assay to look for histidine at the one position of peptides. The first radioimmunoassay that she [Ros Yalow] made with Solomon Berson was an assay for insulin. They found that in the guinea pig, insulin was very poorly detected by their assays… How to explain that? Either guinea pigs had little insulin or the guinea pigs actually had as much insulin as any other animal but their insulin was different enough that it reacted poorly in the assay, which turned out to be the case.

The second immunoassay developed was against glucagon. In the chinchilla, radioimmunoassay couldn’t detect any chinchilla glucagon at all. So you say, “Why chinchilla?”

Chinchillas are in the hystricomorph family, of which the guinea pig is also a member. We were interested in characterizing chinchilla glucagon. So we made a trip to a chinchilla farm [in Pennsylvania] and came back with chinchilla pancreases. I went through the usual steps: frozen pancreas is extracted by an alcohol and acid mixture, and assayed. Assay with a glucagon radioimmunoassay detected nothing. Nothing!

That raised the question: “Are chinchillas special?” Do they have no glucagon even though all known animals, mammals, and vertebrates have glucagon? Or is it their glucagon is sufficiently different in amino acid structure that the radioimmunoassay just couldn’t detect it?

At that point I said, “I have this His1 assay,” which is a chemical assay to look for histidine at the first position of the peptide chain. I applied it, and I found a fair amount of His1, the chemical signature… I used the assay to purify the peptide from the pancreatic extracts. At the very end, there was a single purified peak of peptide containing the His1 chemical signature. I sequenced the peptide, and sure enough, it was very close to human glucagon but not identical in sequence. There were amino acid differences, [which] explained why the glucagon radioimmunoassay could not detect it.

On confidence: The fact that this peptide sequence had some homology to glucagon gave me …confidence that I was on the right track.

I wanted to further validate this His1 assay with a more difficult test. I looked at a biological substance that I knew beforehand contained a His1-containing peptide but not in such great abundance that it would be a challenge to detect and purify. The reason for this added challenge is new hormones might be present in very low abundance; if they are, I would need a very sensitive chemical assay.

Around that time – so many years ago I forget which year it was – a report came out of the NIH in Jerry Gardner’s group, which is a GI group. They had noticed reports that…when animals or humans are envenomated sometimes pancreatitis develops. Being a GI group, they were interested what can cause pancreatitis. They screened a wide variety of venoms to look for compounds in venom that can stimulate the pancreas because one way to produce pancreatitis is by over-stimulating the pancreas.

Drum roll! On the Gila monster: They found activity in the venom of the Gila monster lizard that stimulated the pancreas.

Gila monsters are kept in serpentariums, a kind of farm, along with different venomous snakes. They are milked for their venom, which is then dried. People purchase the venom to create anti-venom.

If any of us were hiking out in the desert and got bitten by a snake we would go to a hospital, and they would administer anti-venom, which is antibody created to the venom, to neutralize the venom now circulating from the bite.

The NIH group screened venom from a variety of venomous reptiles, one of which was the Gila monster. I read their report and thought, “Wow, that’s great.” They named the peptide they found in Gila monster venom helospectin, based on the Gila monster’s scientific name, Heloderma suspectum. Helospectin has a histidine at the 1 position.

I ordered some Gila monster venom and put it into solution, because it comes in dried form, and assayed it. There was a lot of His1 activity by my chemical assay. I put it through the first purification step and observed two peaks, one of which was huge. The other one was very small. Looking at the pattern I said, “That’s what I expect,” because when you purify any peptide you’ll get a large peak of purified peptide and a much smaller peak of degraded peptide. We always see that; it’s very common. I said, “Clearly this major peak is helospectin and the minor peak, this small smidgeon peak, is degradation product.”

I went ahead and sequenced both peaks. It turns out that the very small peak was helospectin, but the major peak, when I sequenced it, was not helospectin. It gave a sequence never described before. But it did have a histidine at the 1 position.

When you find something new, you have to name it. I named this peptide Exendin. There are two different poisonous lizards. One is entirely black and lives mostly in Mexico. Its scientific name is Heloderma horridum. The other Gila monster lizard lives in the American Southwest. It’s banded. It’s actually very nice-looking. It has orange bands around the body. Its scientific name is Heloderma suspectum. The peptide I isolated from Heloderma horridum I named exendin-3 and the peptide from Heloderma suspectum I named exendin-4.

On exendin-4 and where it is: There was a very interesting coincidence. A person who worked at the NIH on isolating the earlier peptide, helospectin, was Jean-Pierre Raufman, a gastroenterologist. By that time he was working at the Downstate Medical School. I called him to say, “Hey, I found this peptide.” In talking with him, I said, “I’m naming this peptide Exendin….” but since the first peptide from the Gila monster was helospectin, that can be considered exendin-1. Because a Belgian group found a peptide corresponding to helospectin in the other Gila monster species, it can be considered exendin-2. That’s how I called the peptides that I isolated exendin-3 and exendin-4.

The name Exendin is actually very descriptive and practical. Basically, the venom secretion, which they now say is really a salivary secretion, is from an exocrine gland… Fluid flows to the outside of the body instead of into the bloodstream, so it’s an exocrine secretion that has endocrine function. Exocrine secretion with endocrine activity gives the name Exendin.

On what he thought at the time of discovery: Being a physician and doing research, a Holy Grail to is to make a discovery and bring that discovery from the bench to the bedside. Although I can’t personally develop the discovery to the point where it is used at the bedside, the concept is to make a discovery that later develops into something clinically useful. It can’t get any more gratifying than for that to happen.

On what he thought at the time of naming: The next step that had to be taken was to show it had biological activity. Otherwise, it’s just another peptide.

On timing and when exendin-4 was discovered: Exendin-4 was officially discovered in 1990. The first paper on exendin-3 was published that same year. Then the paper on exendin-4 came out two years later, in 1992.

On the next question: Peptides act on receptors. At this point, the question was: Does exendin-4 act on a receptor such as the glucagons receptor, the VIP receptor, the GLP-1 receptor, or some other receptor? Experiments showed that both peptides, exendin-3 from Heloderma horridum and exendin-4 from Heloderma suspectum, can act on the GLP-1 receptor. A difference between exendin-3 and exendin-4 is that exendin-3 also had a small amount of VIP activity, whereas both exendin-3 and exendin-4 act only on the GLP-1 receptor.

That is why exendin-4 was selected for development and not exendin-3, because of this additional activity of exendin-3.

[query from Kelly Close] And when was it actually selected for development?

That would be getting ahead of ourselves. I had just named it, and with Jean-Pierre Raufman, we found that it acts on the GLP-1 receptor. Then the question is, what’s so important about GLP-1? A scientist named Joel Habener at the Massachusetts General Hospital found GLP-1 to have a very unique, very valuable activity. GLP-1, like many hormones, is known to stimulate insulin secretion, but what was really unique about GLP-1 is it stimulates insulin secretion only in the presence of a high glucose level. Glucose dependent stimulation of insulin secretion is very desirable. What was not known at the time, but it became quickly known, was GLP-1 gets degraded quickly in the bloodstream. The bloodstream has enzymes that cleave GLP-1 into an inactive form. The first two amino acids are clipped off rendering it inactive.

On DPP-IV: DPP4 stands for dipeptidyl peptidase, type 4, which cleaves off the first two amino acids rendering GLP-1 inactive. If you inhibit this enzyme you will extend the lifetime of the endogenous—our own GLP1 secretion thereby making it active longer. That’s the basis for these new products that are being developed by other pharmaceutical companies.

On what else is being inhibited: I’m not knowledgeable about the science behind dipeptidyl peptidase enzymes. GLP-1 would have been the ideal therapeutic agent, but its biological half-life was too short. There was a rush to see what modifications can be made to GLP1 to make it last longer.

On the fork in the road: Coming back to me, it was clear at that time that if exendin 4 acts on GLP1 receptors it might be also be a therapeutic agent. I’ve worked in the VA my entire career; however, the VA decided not to patent my invention. In VA’s defense, at that time VA was only interested in patenting inventions specific to veterans such as spinal cord injury, loss of limbs or prostheses. That would be very much VA/veteran related and they would have been interested in patenting that kind of invention. Now, VA has changed, so they are prepared to patent any invention made in the VA. But back then it was different thinking in a different era. In fact, I would point to Rosalyn Yalow herself because Ros Yalow, along with Solomon Berson, invented radioimmunoassay. Ros herself said to us… many people thought radioimmunoassay should be patented. She refused. She said the work was done with government money; it’s for the public good, and it should be in the public domain. However, radioimmunoassay did not need patent protection to be commercially developed. Exendin-4 required patent protection before any pharmaceutical company would risk hundreds of millions of dollars to develop it. In the memo, the VA general counsel said, “VA declines to patent.” but went on to say, “You can patent it on your own.” Here I faced a fork in the road and like Yogi Berra I took it.

On patenting and the long roller coaster ride: I’m a physician, a scientist, what do I know about patenting? On the other hand, unlike Rosalyn Yalow, if I thought, “If I don’t patent it no one will touch it”. Bringing a drug to market costs hundreds of millions of dollars. No pharmaceutical company would ever think about initiating drug development without patent protection. My conclusion was, “I have to patent it. Without a patent it has no chance of being developed.” I found a patent lawyer in Chicago, and here I have to give my wife a lot of credit because the patenting process is like a roller coaster; a prolonged roller coaster ride. You don’t know when it will end; you don’t know what its prospects are for success. You apply for a patent, and it seems endless. Every month I received a bill from the patent law firm for expenses. My wife would say to me, “Why are you doing this?” I would say, “I have to.” And then she would say, “Well, when is it going to end?” “I don’t know.” This went on month after month and she finally said, “Why are you doing this?” I couldn’t think of a good answer so I said, “Well, it is my tuition.” She was fairly tolerant. How long did it take? Two years. And not a smooth two years either. After a year it was rejected. The patent attorney explained that the Patent Office routinely rejected everything the first time around, basically placing the burden of proof on the inventor to clearly demonstrate that persons of ordinary skills could not have come up with the invention. After the first rejection, I sent rebuttals and responses to the Patent Office. Another year passed before the patent was finally awarded.

On calling all the pharmaceutical companies: How did I feel when the patent was issued? Both elation and dread because I now had to market it. What did I know about marketing? I remember calling up all the major pharmaceutical companies and most were not interested. For example, Merck’s interest then was oral medications and not peptides, which have to be injected, like insulin. With other companies I talked to, exendin-4 represented something too risky. Large pharmaceutical companies face their own forks in the road. To commit to developing a drug is very expensive; phase 1 might be $15 or $30 million. Phase 2 is a whole lot more than that, maybe $50 to $100 million; phase 3 is tremendously expensive. They don’t commit to initiating drug development lightly. On the other hand, if a biotech firm took the risk and developed it to a point where risk was much reduced then risk is more tolerable. I think large pharmaceutical companies are willing to pay more money for partially developed compounds with reduced risk. Actually, I got very good advice from a person who was working at Eli Lilly at that time. Richard DiMarchi was in New York visiting Mt. Sinai where I asked him about the prospects for exendin-4. He gave me very good advice and I still remember it clearly. I offered to drive him to the airport. On the way to the airport I said, “Look, I have this compound - would Lilly be interested in it?” He said, “The marketplace asks how is your compound any better than what’s available?” Exendin-4 is a GLP1-like compound, so it had to be either more potent, or have fewer side effects or be longer acting than GLP-1 to have any value in the marketplace. I concluded that the competitive advantage, if exendin-4 had any, would have to be duration of activity.

On the excitement of testing diabetic mice: I set up an experiment with three groups of diabetic mice - one group was injected with saline or placebo, the other injected with GLP-1, and a third with exendin-4. Then I measured the glucoses in the three groups of mice over 24 hours. When I processed the data, the placebo group had high glucoses that remained very much elevated; it would be the equivalent of 300 mg/dl. In the GLP-1 group, the glucoses dropped dramatically to the mid-100s then started to rise at one hour, and by two to four hours it was back to their baseline elevated level. In the exendin-4 group, glucoses dropped dramatically at half an hour, stayed in the mid-100s and remained that way for hours. The animals varied, so some started to come back up within six or eight hours while a couple of animals had glucoses that remained in the mid-100s through to the next day.

On the VA: I can’t speak highly enough about the VA. I’m been here my entire career - part of our mission in VA is research, and that’s important. The VA’s mission is to care for veterans, to train healthcare workers, to do research, and to serve as backup for the Department of Defense. It’s estimated that 40% of all residents training in medicine have trained in the VA at some time in their career. The VA has a research career development program. I was part of that program; I entered the VA in the research career development program and spent nine years progressing through it, which is a very large investment in individuals. As part of this program, my being a physician, they said, “The expectation is that you spend at least a quarter of your time in clinical activity.” That’s very understandable because if you don’t engage in clinical activity you lose your skills are not exposed to clinical problems and needs. Three-quarters of the time I was in the lab doing experiments, the other quarter of the time I was seeing patients. I thought that was a good balance. And here I have to give credit to my wife for helping me to maintain balance in my life. The balance is: research, clinical activity. and family.

On the 1996 ADA meeting: I submitted an abstract to the annual ADA meeting on the experiment with diabetic mice comparing exendin-4 to GLP-1. It was accepted as a poster presentation at the ADA in the summer of 1996 in San Francisco. Andrew Young of Amylin came by. Richard DiMarchi at Lilly was also there. He invited me to give a talk at Lilly, which I did. It was a very interesting experience. Pharmaceutical companies have personalities. At Lilly, I was scheduled for meetings every half-hour for the entire day. I talked to everyone including people in chemistry, manufacturing, and physiology. I remember it was like a job interview. They wanted to fully understand exendin-4 and reduce their risk to the greatest degree possible – it was if they were looking for things that could go wrong. At the end of the day, I still recall, the final person I saw was Jose Caro. He said, “Thank you for coming.” I asked him, “Well, is there an interest?” And basically he said—I don’t really recall what he said—but basically he said no. Depressing. Exendin-4 didn’t clear the hurdle. But Andrew invited me to Amylin. They moved quickly – I gave a talk, they said they were still interested, and we negotiated a license. That was October 1st, 1996. I have to say this, one of the really wonderful things about the United States are biotechs willing to take risks that larger companies may not be willing to take and run with it. I don’t think the risk-taking culture exists in Europe so much; it’s one of those things that make the United States so great. The willingness to take large risks, and, of course, to have the rewards that come with success.

On the post-licensing trajectory: My trajectory after the licensing was this: as part the research process in academia we make research proposals. The proposals are evaluated, and either you get funded or you don’t. I had submitted a proposal that came back with the comment that this proposal has merit, but it is most appropriately done by a pharmaceutical company. I read it and thought, yes, that’s right. That really was correct. So I said at that point I’ll stop research and move on to a different career, which I’m still in, bringing electronic medical records into use to replace paper records. In 1996 the VA system rolled out an electronic medical record; and I became the manager. An entirely electronic medical record feeds into a database that in turn supports a better kind of healthcare delivery. For example, at the Bronx VA we have 25,000 active veterans. Querying the database shows there are 4800 diabetic patients. This group can be characterized in great clinical detail; what medications they are taking, etc. To me, electronic records have powerful potential for improving health care. For example, during the flu season, we can define risk groups and then track what percentage in each risk group had flu vaccine. One of the risk groups is transplant patients, 50 with transplants of any kind, five with heart transplants, four done at Columbia Presbyterian and one at Mt. Sinai. We still ask the same question: how many of the transplant patients had their flu vaccine? It’s a very powerful – this is where health care is headed in the near future. The VA has a program for improving health care by comparing itself to measurable benchmarks with other healthcare systems. Among diabetics, how many have had a hemoglobin A1c? How many diabetics have had their annual retinal exam; how many have had their feet checked for loss of sensation. These are all public or health care benchmarks where the VA can say, “This is how we compare to other health care systems.” And I’m part of VA’s movement to measure and establish quality markers in clinical care. I think it’s just great for health care, not only for individual patients, but for groups of patients – it helps ensure uniformity in health care. The aim is to assure that all patients receive equally good health care. We look at everyone in this system to determine which healthcare services each individual should be receiving. You could call it my second career in the VA, as Director of Clinical Informatics here.

On stopping research: I thought exendin-4 was in good hands at Amylin. I wasn’t interested in starting a completely new direction in research, so I reinvented myself in another emerging field, clinical informatics. I think it turned out right; if exendin-4 were a child, at a certain point of the child’s development you have to say, “I trust the teachers will take care of this child’s further development.” That’s the way I saw it.

On challenges of research: To me the hard part is finding something new. There are no guarantees to discovery. You explore, you take a certain path and follow up whatever that path brings. In terms of obstacles, the hardest part was doing many things well. To my endocrine fellows, I ask, “Can you do one thing well?” And they say, “Yes, yes we can do that.” And then I ask them, “Do you think you can do two things well?” And they say, “Yes, it’s a little harder, but we can do two things well.” “What about doing three things well?” Meaning research, clinical practice, and family—my life revolves around those areas and I suspect theirs will too. Doing three things well… doing three things well is very difficult. That’s where you have to make choices – without a supportive family, it is nearly impossible. I am a strong believer in having a balanced life. It’s a blessing to have an interesting life but sometimes very difficult!

On Exenatide’s clinical path: It’s really thrilling to follow Exenatide’s development. I’m an outsider because I’m not engaged in its development, but it’s wonderful to see. Again, I see its development like that of a child. As your child develops you have great pride in the child. You send it off to higher levels of school where the child develops in larger environments. In my case, the child and the development were guided flawlessly.

On watching development: Again, like any good parent, I have my fingers crossed and hope nothing bad happens. It really is like that. It’s not only the good things but also the absence of bad things. Maybe I’m taking the child analogy too far but we have hopes for children; we have hope that the child will not only do well but help mankind in some way. I know my parents had the same hopes for me; that I would do well and contribute to mankind in some way. This is so satisfying.

On whether he will go to ADA this year: I have to choose. I usually go to the Endocrine Society for continuing medical education credits. (Ed. Note – at this point, we tried to convince John he should attend ADA this year to see Byetta launched!)

On hypoglycemia: One of the more important problems that we face as clinicians is hypoglycemia. Exenatide is potentially one of the therapies that will reduce the incidence of hypoglycemia. From a clinician’s point of view that's an important problem and Exenatide a good solution. I had a patient on insulin trying to achieve tight control who was having hypoglycemic events. I said to him, “Look, you have to cut back. You can’t afford to have such frequent hypoglycemic events.” He said, “No, I can’t.” I asked him why? It’s not good for you. He said, “My mother had diabetes, she had two of her legs amputated and I don’t want it to happen to me.” Then a month later I learn he was hospitalized because he crashed his car; fractured his pelvis probably because of hypoglycemia. If it could prevent some of these things from happening it would be wonderful.

On the competitive landscape: I’m a biased party. I am biased for Exenatide, and I hope it does better than competing agents. However, I also see things from a patient’s point of view. With that in mind, really, what is best for the patient is, in the end, what counts. It won’t be a single compound. It will be multiple compounds. With Exenatide, it’s like software - there’s a version 1, a version 2—so there will be evolution there too. LAR? Oh, I can’t talk about that…

On education: A major portion of my time is spent educating patients. Patients make decisions about their care. To educate the patients I tell them that the better the diabetes control the less the chances of complications occurring. The goal is to prevent or delay the complications of diabetes. Education is always difficult, however, which is why the simple therapies that work are wonders.

On how more partnerships could be made: The United States has been an incredible engine for innovation and bringing new products to markets, essentially, transforming health care into something we could not have envisioned 10 years ago. How to do it even better requires innovation on the part of big pharmas, small pharmas, and biotech. Making a discovery is the first step, but then how do you translate--move that discovery from bench to bedside? There ought to be many models, not just one. If one doesn’t work there should be an ability to look to another model and yet another model after that. This is what we need to work on.

Editor’s final note – there was a slight lag with the interview and our final question - we were in touch with John briefly on Saturday, April 30 and had one final question for him.

On how John learned about the FDA approval of Byetta. [Amylin Chief Operating Officer] Dan Bradbury called me with news of FDA approval on Thursday April 28, 2005 at 9:30pm Eastern. I did not have much sleep the rest of the night from the excitement. My hope is Byetta will come to mean excellent diabetes control. That would be music to my ears.

--Kelly L. Close