Extended Interview

An Interview With Francis Collins of the NIH

In this full transcript from a 53-minute interview, Collins touches on proposals to restructure parts of the National Institutes of Health, why new therapies cost so much, the future of human genomics, and why it's so hard to recruit institute directors.

[With Science staffers Jocelyn Kaiser, Eliot Marshall, and Laura Zahn. NIH Associate Director for Communications and Public Liaison John Burklow also present.]

J.K.: What has it been like going from directing the genome institute to running the whole NIH?
F.C.: It's pretty amazing the number of issues that occupy your attention on any given day. It's a long and complex but interesting list. When I ran the genome institute I felt I could pretty much get my head around most of the major issues that we needed to attend to on any given day. Here it's quite a stretch and I really have to depend on all of the other incredibly bright minds around me to assist with that. And to bring things to my attention they have reached the point where the director had better get involved.

The people here are unbelievably dedicated and smart and hardworking, and that makes the job a joy. The people who run those 27 institutes and centers are phenomenal. My own sort of senior staff including, John [Burklow], have been just terrific teammates. I'm really pleased that Kathy Hudson accepted my offer to come and be my chief of staff because that's been a great advantage because of her consummate skills both in terms of science and science policy.

It's a lot of fun. But every day is different. The hours are long, I'm probably working 100 hours a week an average week trying to be sure that I am doing all of the things the NIH director should be doing. One of the biggest challenges is to keep your focus on the big picture and the vision and where you want to take this organization in the course of a 2- or 3-year timetable and not just get swamped with the dailiness of what's coming across your desk. Because there's lots of that stuff.

And sometimes that's a new scientific development that's exciting and you want to find out as quickly as possible what that was and what you might do next about it. Sometimes it's a request from the Hill for information that you have to have by that afternoon and you better be sure you send it out in a way you know it's right. And sometimes it's a press call or it's a budget issue that you have to straighten out as quickly as possible.

And oftentimes one of the great advantages of this job is you have incredible convening power. If there's a topic that you want to work on you can ask people to drop what they're doing and show up in a fairly short period of time and they do. We had a meeting 10 days ago called the "big think." And this was an effort to try to bring together about 60 people that I thought were most capable of sort of rising out of their own specific scientific interest and looking at the whole landscape of biomedical research and brainstorming about, "What are we not doing that we should be?" This was particularly relevant for trying to generate ideas that might fit into the Common Fund because the Common Fund is one of my primary responsibilities.

So we just had them here for a day, and we worked them pretty hard, and broke them up into breakout groups actually focused on the first three of the five themes in that Science paper. And charged them to come back after 4 or 5 hours of debate and say what they thought were some of the unattended opportunities that we should start to focus on. It was a lot of fun. And again, the people had maybe less than a month's notice and I think our acceptance rate was about 95%. Because they wanted to be here to have a chance to be part of something like that. Just today I've been in all day today and most of yesterday in this Scientific Management Review Board [SMRB] meeting. So what's that?

J.K.: I know what that is and I have a question for you about it.
F.C.: This was created by the NIH Reauthorization Act and is now a body that's supposed to look at NIH structure, organizational structure, and advise about whether there are changes that could be made to make us better. And we have a fascinating discussion today about the Clinical Center which then grew into a broader discussion about clinical research in general. And that tied into the legislation that's in health care reform called the Cures Acceleration Network (CAN). And it was a really wonderfully rich conversation with some very substantive people. This group is chaired by Norm Augustine. That's all part of my job.

J.K.: Since you mention that management board, part of what they're supposed to be doing is looking at whether to merge institutes. I think I asked you about this a couple years ago and you seemed to think that wasn't going to happen. So is this just sort of an exercise or—that is what they're supposed to be advising you on, right?
F.C.: And they have a subgroup that is supposed to be focused on particularly addiction research and whether the National Institute on Drug Abuse and the National Institute on Alcohol Abuse and Alcoholism should be merged into one entity. And they had a very tough time. That was most of yesterday's meeting was discussing that and hearing from a lot of advocates, from researchers who'd been funded by one or the other of those institutes, by both of the current institute directors. And frankly the SMRB were not able to come to a consensus. Which is interesting because I guess most people would have said, "Well yeah, of course." But when you look at the details, and you recognize that the science of alcohol addiction and drug addiction, while there's some overlap, there's also some clear differences. That the alcohol institute not only looks at addiction but also at metabolic aspects, particularly liver disease which doesn't overlap with other drugs. And you consider that alcohol is after all a legal substance and 90% of us at some point in our lives are comfortable with taking it in while the drug abuse institute is largely focused on drugs that are not legal. So there's a personality of the institute issue here that people thought might be important to preserve, others thought would be good not to preserve. On the other hand, it goes to show.

J.K.: Is that the end of it then?
F.C.: Ultimately they're advisory to me so the end of it will be when I decide what to do.

J.K.: But if they couldn't decide?

E.M.: That's an indication of how difficult—if that is the one that looked like the easy one, in a sense.
F.C.: Right. So what does that say about others that might be even more complicated?

On the other hand, this was maybe not so much of a perceived problem from the inside as a perceived problem from the outside. I can imagine other things and maybe even involving clinical research where from the inside it might seem to us like it's time to think about whether we should be structured. So the clinical center, which currently is this wonderful jewel, this research hospital, 240 beds here on the campus, all patients there are on research protocols, is chronically in financial trouble. Because the costs of medical care doesn't decide to go flat just because the Congress decided our budget would go flat for 5 years. And so the clinical center is increasingly squeezed and what should we do about that?

So we had a big discussion about whether the Clinical Center ought to open its doors to outside investigators to come and run studies. And whether the Clinical Center's budget actually ought to be moved out of where it currently is in the intramural program and put somewhere else in NIH, maybe in its own budget line, maybe in the director's office. And it looks likely that the SMRB is going to recommend that. So that's a pretty substantive change and a pretty important part of what we're doing.

So there are possible things that can be done. It's not always, no we want to keep things just the way they are.

J.K.: So we wanted to talk about the budget and the stimulus cliff which you've been talking about on the Hill lately. You've said that you expect very low success rates, 15%. What's this going to mean? Is it going to shut labs down? We've heard people talking about postdocs being out of work. What's it going to mean if that's what you get?
F.C.: I don't think anybody can be exactly clear on that. First of all, success rates are a function of several things. One is, what is the budget that you actually have to spend. We don't know yet. We don't know what FY '11 will turn into and obviously Congress is wrestling with that.

Yesterday's primary result was a big disappointment, I think to many people. We were hoping to see Senator [Arlen] Specter [(D–PA)] reelected because he's been such a strong champion for medical research, as you all know. And I am personally grieving at that outcome because of what it may mean for medical research in the congressional appropriations process. And because Senator Specter as a human being has been remarkably devoted to making the case for the importance of this kind of work. And he's a courageous survivor of cancer himself.

So we don't know what the budget will end up being. Obviously the signs are not particularly good that the Congress will do better than the president's budget. Some noises might even indicate that they'll do worse with all the other pressures upon them from education needs and so on.

The other question we don't know the answer to is what will be the number of grant submissions in FY '11. I will tell you, there were some concerns that we would see a huge deluge already in the January submissions because even though they would be submitted in January 2010 because of the cycle time, they wouldn't be funded until early '11. And many people were wondering if all those Challenge Grant applicants, those 20,000 that came in during the stimulus and only 800 of those got funded, are the other 19,200 going to come back? And the January submission, because that's the first point at which we allowed the 12 page application, that's what the Challenge Grants were, so just put a new date and send it back in. That did not happen. We didn't see a particularly big bump.

We are of course trying to do everything we can to spread out the cliff, make it more of a slope. And that is including letting grantees know that they are going to be listened to if they ask for a no cost extension of their 2-year Recovery Act grant. Because many of those PIs [principal investigators] didn't actually get funded into well into that first year. And they may very well not be able to do all of the things they wanted to do in a year and a half and want to have more time. And that would of course delay, but not reduce ultimately, the problem.

We will undoubtedly have to look at draconian things like downward negotiations, which means cutting the budgets of approved grants in order to free up dollars to approve more grants. We are trying to protect certain parts of the enterprise. Postdoc training slots, for instance, are protected and actually for the first time in several years postdoc stipends, as you probably noticed, are scheduled to go up by 6% after falling progressively behind in terms of keeping up with where the Congress told us we should be.

Early stage investigators who we have been doing everything we can to try to encourage, we will continue that process by making sure that they are put into a pool. Their applications are in there competing against each other and not against the more established investigators who maybe had more preliminary data and more experience.

But I'm sorry, I can't come up with a magic solution here that is going to reduce all of the pain. There's going to be pain in the circumstance where the mismatch between science opportunity and science resources—it's going to be fairly stark. And that's the thing that keeps me awake up and night and certainly lots of people are brainstorming about potential ways to lighten the load, but you can only lighten it so much.

J.K.: And what about coming years, is NIH going to be facing more flat funding?
F.C.: Who knows. We've gone through boom and bust cycles before. And I suspect that trend is not going to stop at this moment or probably will stop in the foreseeable future. If I was king of the world, and could put into place long term budgeting where you had a stable trajectory over multiple years that could be counted on, it would certainly be a better way to support science. But that's not our system.

Our system is every year is a brand new year and you're very much at the mercy of what's happening with the economy and the political system. And we have to roll with that and do the best we can. And we should be clear that we are also fortunate to be in a country where biomedical research is supported to the tune of $32 billion, I mean that's a lot of money. It will obligate all of us who are science managers in a circumstance like what we're facing in FY '11 to really look hard at whether there are less productive areas of science that we can't continue to afford to push forward just because we always have.

That of course creates winners and losers, which can be very unpopular, but I don't think you can be a good science manager and not include that in any year and especially in a year where things are really tight.

EM: It does make it much more difficult to start something brand new and fun and big. The scope of your Common Fund, that gives you some leeway?
F.C.: It does. And I would argue the worst time to back off from doing something new is when your budget is tight, cause then you have no reason to argue that it ought to get bigger in the future. I mean if the science opportunities are as good as we all say they are, and I really believe that they are, to put that off would seem irresponsible. What you may not be able to do is to go as aggressively into those new territories as you'd like, but you'd better do everything you can.

So for instance in putting together the president's budget for FY '11, those five themes that you see there were very much factored into the way in which the advocacy for this billion dollars that ultimately emerged from the process came forth. And if you look at the spread among the 27 institutes and notice that they're not all in lock step at 3.2%, it's because of their track record and the way in which they have invested in those five themes which we thought ought to be featured a bit in what we're planning to do going forward.

The Common Fund is, I think, a good example of an innovation space that should always be used for high-risk, high-reward purposes. That's certainly my goal. The Common Fund although it is now $550 million, most of what is being funded this year are multiyear projects which are not going to turn over next year. So the amount of new money available in FY '10 was only about $20 million and the amount that will be available in 2011 is about that same amount.

But that doesn't mean that we can't start some interesting things and I did start seven new common fund projects in February. And we're right now in the business of trying to figure out where we want to start next year because it's not too soon to put some ideas together, get some RFAs drafted and out on the street.

J.K.: You mentioned the president's budget proposal. So I think some researchers are worried that this emphasis on translational research and these big goals will mean cutting back on investigator-initiated grants. Are they wrong?
F.C.: [Laughs.] I don't think they should be very worried. Everybody's going to be stressed so it will be tempting if your grant didn't get funded to look around for some reason other than the fact that it was a tough year budgetarily. So I think it's going to be very important to get the numbers straight in front of people's minds.

The amount of additional funds that might go into focusing on translation are going to be maybe 1% or thereabouts of the overall NIH effort. That shouldn't have a very big effect.

J.K.: Does that mean 1% less for investigator initiated research?
F.C.: This is an interesting sort of question. Does—first of all, is there an interaction between new initiatives and the opportunity to be appropriated at a decent level? I would argue that if NIH simply said, "We're going to keep doing what we've been doing all along," we're not in a very good position then to ask the administration or the Congress to give us more resources. Our argument for more resources almost always has to be tied to some goal, some new initiative, some new opportunity.

The translational goals, I must tell you, get a lot of traction with the Congress, with the public. They should. I mean, we're the National Institutes of Health. We are supposed to be invested and, when the opportunity arises, coming up with ways to prevent and treat disease. And if we have a chance to do that now, we should be, I think, willing to put forward some bold initiatives of that sort and to make it clear that these are opportunities that won't go very far unless they have some resources.

I mean that's pretty much what's happening right now. When you no doubt listened to the hearings in the House and the Senate, a lot of interest about translation, about the Cures Acceleration Network and whether it's going to get appropriated and not just authorized. And I think that puts us in a pretty good position to say, we really are at a critical juncture as far as moving basic science discoveries into the clinic in ways that we couldn't have 5 or 6 years ago.

All that being said, I totally want to protect the basic science foundation of everything we do because that is our future. And I would certainly say to any R01 investigator who's worried about this that the vast majority of the discoveries that are going to matter are going to come from those hypothesis-driven, investigator-initiated efforts.

But I also want to empower those investigators with resources that make their job easier, which is what a lot of the Common Fund projects do by providing common databases with free data that you don't have to generate yourself, mouse knockouts or whatever.

But I also want to inspire individuals who may have thought of themselves lifelong as basic scientists to think translationally if the opportunity arises and not to feel like that's off limits. And that's not for everybody but I can certainly point to a few basic scientists who found that pretty exhilarating. Why do people go into biomedical research? Curiosity, sure, it's fascinating, it's intellectually stimulating, it's not very profitable if you're in the academic sector. But also because you have a dream that you're going to do something, that you're going to discover something that wasn't known before and you're going to help somebody. And if we in this effort to emphasize translation are giving a few more basic scientists a chance to see that part of their dream become a little more real, I don't think that's bad as long as we're not seen as putting down the basic science part of their careers and many others like them. We can't do that.

J.K.: As part of this you're investing more and more in drug discovery and development. I know about the specific programs, like TRND [Therapeutics for Rare and Neglected Diseases]. And CAN, although that's not funded yet. Isn't this expensive and risky? Why should we be using public money, public funds for it?
F.C.: Let's talk about TRND. In terms of dollars it's pretty puny. FY '10, $24 million. What's in the president's budget for FY '11 is $50 million. That on the scale of 31 to 32 billion, it is a pretty tiny blip.

It is high risk for sure. Why should we do it at all? Well, if you're talking about rare diseases, or diseases of the developing world that may be not rare at all but don't have much of an economic market, if the government doesn't get invested in those therapeutic efforts, they're not going to happen. There's going to be no incentive on the basis of economics to try to find a cure.

So if we've arrived at the point scientifically, and I think we have, where that opportunity is becoming real because we've discovered the molecular basis of hundreds of diseases and because we have technologies now in the hands of academic investigators to allow them to take a basis science discovery and turn that into a high throughput screen that yields up a small molecule that looks like it might be promising, that seems like a good thing.

And what TRND offers is to even go one step beyond that into the preclinical phase to take some of those promising compounds and see whether you could get them all the way to FDA [U.S. Food and Drug Administration] approval and a clinical trial.

So examples. Because right now TRND has three pilots that they're trying to test out their system. And of course they're just getting started.

One is schistosomiasis. So this is not a rare disease, this affects 250 million people. There's been no new drug for this disease in 50 years and the drug that's available while it seems effective, resistance is starting to pop up. And here's a circumstance where more basic investigations led to an idea about a new drug target, which then led to a high throughput screen, which led to the identification of a compound ,which has already been shown to be curative in the mouse model.

But to get it all the way to a clinical trial, you've got more work to do. So TRND is going to take that on, putting it through the standard toxicology tests, figuring out what the pharmacokinetics and pharmacodynamics are and the tissue distribution and so on.

A second one is Niemann-Pick Type C. Very rare disease but if it's your kid who has it, you're not really that taken by the fact that it's rare, therefore, it doesn't matter. So somebody should be working on this.

And even Niemann-Pick Type C, the gene for which was discovered here at NIH, I had a small role in that, is clearly a disorder of lipid metabolism leading to—it's a storage disorder. So there is now an interesting, small molecule that in a cell-based assay seems to cure Niemann-Pick C cells. But before you would think of putting that in a patient, there's a lot of work to do.

And the third one is an even rarer disease called hereditary inclusion body myopathy, so it's one of the many myopathies. That's one where there's already data in a mouse model that a small molecule can be quite effective. But in order to get FDA approval for a clinical trial, it's got to go through the standard toxicology testing, so TRND will provide that.

So projects can enter this at various stages depending on how much is known. And actually they can leave the pipeline if there's a company that says, "Okay you've gone far enough, I believe you, even though there's a small market here the risk have also been greatly reduced by NIH having done this investment, We'll take it from here." And NIH would be totally ready to license those out at that point.

And by the way, expect a royalty which would then result, if all goes well, in some reimbursement to the government for the work the government did to try to get something to the public that the public needs. Seems like a pretty good model.

J.B.: Dr. Collins, is there any connection between that and the NIH-FDA [initiative]?
F.C.: Absolutely. So that's another thing which is very modest in terms of dollar expenditure but really important and significant as far as science regulatory policy. So [FDA Commissioner Margaret] Peggy Hamburg and I started meeting even before I was fully in place here and had many conversations about how can we get FDA and NIH better connected? So that in an era where the science is changing so quickly, FDA will have a chance to learn from NIH and we will from them, so we don't design pathways towards therapeutics that end up getting rejected in review because we didn't realize what review was needed.

If we're empowering academic scientists to get more involved in drug development, most of them have no experience with what it takes to get an IND [investigational new drug]. So this has led to this leadership that Peggy and I are co-chairing and this will have a public meeting on June 2nd to see what the various stakeholders think the leadership council should focus on particularly.

And it's led to a research program funded by the Common Fund, one of the seven things we started this year, on regulatory science. Which is an effort to try to provide FDA with an engine of discovery—what was it somebody today called it today, their jet propulsion laboratory, to enable them to have the kind of science that they need to guide their review process in more creative ways.

E.M.: They do have a research program, right?
F.C.: Very, very limited and they certainly have not had anything like this. And so we—most of this is actually funded by NIH because their budget is so tight, so we volunteered because it seems like a great idea that will cover most of the costs. We put out an RFA [request for applications] and we got 59 letters of intent. I've read all of those and they're really interesting. In fact we're going to be challenged when the review goes through, we may not have enough money to fund the best stuff and have to go scrabbling around to fund more.

So all of that I think is very defensible. I mean why didn't we do it 10 years ago? The science wasn't ready for it. Why don't we wait 10 years? Huh? When you've got people right now suffering from these rare and neglected diseases for whom we might be able to offer some therapeutic?

J.B.: Did you talk about the subset of rare diseases being a subset of common diseases?

J.K.: Yes, I wonder how far this is going to go. I know that's in your [Science] commentary. You talked about how common diseases are splintering into rare diseases as you understand them genetically I guess. So how far does this go then? Does NIH become the drug development?
F.C.: NIH should only get engaged in this kind of activity in instances where there is no private sector interest. If you're talking about a target for a common disease where a pharma or a biotech is all over it, then they should be all over it and we should not. But if there's a target or a disease that is being neglected because it's not seen as economically attractive but it has some chance of turning out to be valuable to somebody somewhere, then we should think about stepping in and taking that on.

We're basically trying to fill the gap. Because there is a big gap right now. And it is about rare diseases and neglected diseases, but it's also about neglected targets.

So let's talk about common diseases and genome wide association studies [GWAS]. So GWAS. You go to the Web based database that's curated every week, we're approaching a thousand now of validated genetic variations associated with a common disease. Most of them with very small odds ratios. So probably not much use for making predictions about future risk of illness with some interesting exceptions, but most of them would probably not be much benefited by knowing your status, although I found out my status, such as it is, and actually found it interesting.

But very important point. I see no reason why there should be any connection at all between the strength of a variant as far as its risk on future illness and whether it is pointing you towards a target that could be very valuable for new drug development. So I think we have a thousand new drug targets sitting in front of us.

J.K.: And for how many have you actually found the target?
F.C.: Oh, almost none. I mean there are targets, but how many of them will turn out to be druggable?

J.K.: Or you don't even know the gene yet in most cases.
F.C.: Well, you can probably say it's one or two or three in this vicinity. And for many of them now as you begin to get a measure on gene expression, this is turning out to be fairly useful in winnowing through there. You can say, okay, the SNP that seems to be carrying the highest risk is next to these three genes, we'll go to relevant tissue and we'll look at whether there's an allele-specific effect on gene expression. Ah, well it's one of those three and guess what, it's actually overexpressed in the presence of risk. That means it's a good target for drug development, I mean an antagonist against that protein might be valuable. That's not true of a thousand of them, it's probably true of few dozen, but the data's growing.

And you take diabetes, which my lab over there in building 50 is working on furiously. Eighteen published loci for type 2 diabetes. Papers that are now in press will more than double that. And if you look at those 18, on that list is KCNJ11. What's that? That's the drug target for sulfonylureas. Also on that list—PPaR gamma. What's that? That's the drug target for thiazolidinediones. Those are the two major classes of oral anti-diabetic drugs. And they popped up in this genome-wide association screen across the whole genome. We didn't stack the deck, they just popped up.

Okay. If two of the 18 are sort of your positive controls, they're the known drug targets, are you going to tell me the other 16 don't have a chance of being interesting, too? Seems to me like there's a good likelihood there. Some of them have biological features, like a zinc transporter that's only present in the islet, that make them sound like they could be fruitful for investigation.

So I get a little bit frustrated when people say, "Oh, GWAS has been a bust because everything you found had a small odds ratio so it doesn't matter." It may be true that most of them didn't matter very much for personal prediction of illness, but they can be gold for pointing us towards new targets for therapies.

L.Z.: I did have a question regarding the databases and the access to it. Has that been an issue with the confidentiality on the basis of the Homer paper*?
F.C.: That was an interesting juncture, wasn't it? Basically what we have now is a system which works ,although it does require investigators and people overseeing the process to do a fair amount of work to be sure that we're protecting privacy. So what we now have in dbGap, which is the database of genotypes and phenotypes, is a deposit there of all of the data from NIH-funded genome wide association studies. I've lost track of how many hundreds of thousands of DNA samples are represented there, but it's huge.

And those are, however, even though they don't have personal identifiers attached to them, you don't get to look at that data unless you have submitted an application to a data access committee who assesses, "What are you asking for this data for, and let's make sure it's consistent with the consent that the participants gave when they gave their blood samples and their phenotypes. And let's also be sure that if you're getting this data, that you are prohibited from doing certain things such as identifying who these people are."

That has worked pretty well. Let's be clear: There's also other things in there because this is a database where data has to be deposited immediately and there's a need to protect the rights of investigators who did the work to publish before being scooped by others who've done nothing. There was one instance in PNAS [Proceedings of the National Academy of Sciences] that didn't go so well, but it's amazing actually how well that has been adhered to.

I think it's a pretty good system. It is a lot of work and I think people worrying about, can we sustain, especially for the data access committees that have to go through these proposals week after week with a quick turnaround. But I don't think we have a lot of other choices if we're serious about protecting privacy.

J.K.: We wanted to ask you what's coming next with human genomics. What's the next big thing after GWAS and how will you measure the benefits? The payoff for human health?
F.C.: Well, obviously complete genome sequencing is the next big thing and it's coming awfully fast. We're sort of right now in the phase of whole exomes, and that has been pretty exciting too as you have published in Science. People've been publishing all over the place. I was just at Cold Spring Harbor [Laboratory in New York state] last week where if you didn't have at least six or seven published genomes, you didn't get on the program. It's setting the bar pretty high now for what you have to accomplish.

Clearly that is turning up examples of less common variants of large effect beginning to give a hint about where the missing heritability could be lurking. Clearly that's being applied to cancer in incredibly exciting ways as part of the Cancer Genome Atlas. And I think that is an area of perhaps the most obvious and immediate consequences because it is already allowing us to take apart diagnoses that we used to think of as sort of monolithic like glioblastoma and recognize that there are subsets within those groups that are very clearly defined by the molecular nature that have different prognoses and some of them at least different responses to therapy.

So I think the implications here for genomics for clinical applications for cancer are coming fairly quickly. For other diseases—will be coming in the next 3 or 4 years as we find the dark matter of the genome and figure out where the heritability has been hiding, then the predictability for future personalized medicine will start to get more compelling.

Because I do think it's in there somewhere and if we discover it and more and more, people will be interested for preventable diseases, what their situation is. We have a lot of work to do to figure out exactly how that will play out in reality in terms of how do people decide about altering their health behaviors. I'm an example of a person who has altered his behavior, that's an example of n=1. We've proved the principle. [Laughs.]

E.M.: I know you've spoken this. As these diseases fracture into lots of more complex diseases, or more specifically, you get into personalized medicine. But is this going to be more expensive medicine? So this kind of clashes with the goal of trying to control costs. How do you work that out in your mind?
F.C.: I don't think anybody is able to say with total consequences how this is going to play off, but there are competing factors that might actually make this actually less expensive, and which is going to win out.

So if you are trying to develop therapeutics and you have a chance, you have a targeted therapy that you know is going to be based—its success is likely to be based on the individual's genome, then you can run a clinical trial that needs to be much smaller than the thousands of patients to see an effect. Because if you don't see a beneficial effect in a small group of carefully chosen genotyped individuals, then you know it isn't going to work.

You also know in that process of that kind of individualized clinical trial design rescue compounds that traditionally fail. Why do compounds fail? They fail because of toxicity, and you won't be able to stop that, but most of them fail because they weren't effective or they weren't more effective than the standard approach. But they might have been effective for a subset of people, and in the past you had no way of knowing that, so all of your hundreds of millions of dollars just went out the window.

If you can rescue compounds that were otherwise doomed by recognizing that there is a subset of people who will respond, then you greatly reduce the cost of drug development. You can make that drug available more cheaply, and you can actually make it available at all.

So how that's going to play out, I don't know.

E.M.: At the moment, it looks expensive. We just had this prostate cancer vaccine that was approved which I read costs $93,000 for a patient's course of therapy and they may get 4 months added to their life.
F.C.: But do you think that $93,000 had anything really to do with the cost of developing that? No, that was a market analysis where somebody who had no connection with the research that led to the product tried to figure out what will the market bear. So that's another issue, of course. The market can't bear $93,000 for compounds across the board that give you 4 months of additional quality of life. I mean, that's just not going to be sustainable when we have hundreds of those. And I think everybody knows that's coming.

But right now, the people who are doing that analysis figure, okay, we can get it, we'll ask for it. I mean, do you think Gleevec needs to cost almost $50,000 a year? No, but the market will sustain that, it will tolerate it, so that's where the price got set.

J.K.: A couple questions about the cancer institute. You have a director now.
F.C.:Yes! It's so good to be able to say so after all these months of, oh well, I can't really talk about that.

J.K.: So first, there seems to be this paranoia that you and Dr. Varmus want to take away the special status of the NCI. Or if not ask Congress to undo it, pull the NCI under your wing or something.

E.M.: Harold Varmus has talked about the need to make all of the institutes part of the NIH.
F.C.: And the IOM [Institute of Medicine] in their report in 2003 concluded that the special status of the NCI has been more a negative than a positive. I mean, they were pretty blunt about that. So what is it that people are worried about here? Are they worried that somehow not having the ability to submit the bypass budget is going to have a big effect on cancer research? As far as I can tell, the bypass budget has no effect on anything.

Are they worried that not having a presidentially appointed institute director is going to do damage to the leadership of the institute? Well, look back over time and you decide whether that presidential appointment, which means it becomes a political issue, has been a good thing or not. The IOM thought it wasn't such a good thing. I'll leave you to their opinion.

So I don't know what the anxiety would really be. I do think frankly that the NCI over the course of the last 30 or 40 years has at times at least been less connected to the rest of NIH than it might have been for its own food and for the good of the rest of us. And now with the science increasingly drawing connections between cancer and other things, there's even more of an argument why we should be working together.

Take therapeutics. NCI has been the most engaged over the last 10 or 20 years in the development of new therapeutic approaches to cancer, so if we're trying to expand that to other diseases, we need to be connected in that experience. But I think it would also be fair to say that NCI's approach hasn't been 100% successful, either, and if we're going to have a broader NIH effort in therapeutics, they might want to be connected to that, too.

So I think Harold and I have the same view here, that this is a big place with a lot of smart people, and the best outcomes are generally when you don't have walls between parts of the organization that prevent people from learning from each other.

J.K.: We did wonder, were you in contact with President Obama much before the election? And are you now?
F.C.: I joined the Obama science committee after I resigned from the genome institute in August of 2008, so I was part of that deliberation. I had no personal interactions with the president before the election.

J.K.: With his staff?
F.C.: Yeah. And then after the election, I was asked to serve on the transition team; I was asked by Bill Corr, who's now the [HHS] deputy secretary, because he was running the transition team for HHS. In fact, the people on the transition team assigned to NIH were two—me and Harold Varmus [laughs]. And we had a really interesting and very intense 3-month experience working 24/7 trying to get the Administration ready to step in to this fast-moving territory.

E.M.: Was it hard to persuade [Varmus] to come back [to NIH as NCI director]?
F.C.: [Laughs.] You know, initially Harold was helping with the search process. Which had to be informal because this was a presidential appointment, but he obviously knew the field and so he and a few others were helping me with the trolling of the landscape. And it was only after a couple, three, months of that that it became more and more clear to all of us that there was a really perfect candidate right there in our midst. I had not realized until about December that Harold was planning to step down from his Sloan-Kettering job. Which he had planned to do all along but he hadn't said much about. He did his 10 years, and he always said he would do 10 years. ... So at that point, then, it really became a real possibility.

Obviously, this is a big investment for him to come back to NIH, for a guy who loves New York City and to take this on and all that's with that. But he's got a lot of ideas, a lot of fire in the belly. It's going to be wonderful having him here.

J.K.: We wondered why it was so hard to find someone. Is there something unattractive about NIH institute director positions?
F.C.: Most of the people that you would probably put on your short list are currently at the peak of their scientific and research careers. Coming to the NCI would undoubtedly make it rather difficult for them to continue that kind of intense focus on their own research. It's a big job. You're going to have to put most of your time into running the institute.

Harold, by the way, is bringing a small lab with him. He'll continue to do research in a way but in a way that he knows is limited.

So yeah, it would be, I think, not surprising then that a lot of the people that you would have on your list are not really in a position to do this without great damage to their own scientific careers.

And frankly, the government doesn't pay very well. That's a reality that I can testify to. And we would never be able to compete, I think, on a purely financial basis with the top-notch cancer researchers in the country, plus they would be asked, in fact required, to give up all consulting relationships, all outside activities that involve remuneration. You can't do that and be director of an institute. And I understand why, and I support that.

So there's a lot of issues there. You have to really be devoted to public service to do a job at this [unclear].

E.M.: How much time do you get in your own lab?
F.C.: It varies from week to week depending on what else is going on. I have a lab of very independent-minded postdocs, and that's good. They don't want me looking over their shoulder. So I would say I try to be over there on Monday mornings. I have about 10 people in my lab, three projects: diabetes, progeria, and asthma.

It's great for me, it's like a wonderful respite from dealing with lots of other issues to actually look at data and talk to young scientists about where they're going.

So I think it's—And I would argue it also makes me a little bit more effective as an NIH director to be anchored in the reality of what science is all about instead of just reading it in journals.It works out. It wouldn't work out if I wasn't willing to work 100 hours a week.

J.K.: Is that more than you worked as genome institute director?
F.C.:A little bit [laughs].

*N. Homer et al. "Resolving Individuals Contributing Trace Amounts of DNA to Highly Complex Mixtures Using High-Density SNP Genotyping Microarrays." PLoS Genetics 4 (8), (2008): e1000167. doi:10.1371/journal.pgen.1000167