At A Glance
Name: Felix Frueh
Title: Associate Director for Genomics, FDA/CDER
Background: Founder and Managing Partner, Stepoutside Consulting — 2002-2004; Pharmacogenetics Research Director, Transgenomic — 2001-2002; Assistant Director, Protogene Laboratories — 1999-2000
Education: PhD in Biochemistry from the University of Basel, Switzerland — 1995; MS in Biology from the University of Basel, Switzerland — 1991
The voluntary pharmacogenomic data submission guideline document is “done … it’s been done for about four weeks,” said Lawrence Lesko, director of the US FDA’s Office of Clinical Pharmacology and Biopharmaceuticals, last week.
That document will be released alongside another that will spell out the operations of the Interdisciplinary Pharmacogenomic Review Group — the FDA organization responsible for reviewing sensitive genomic data, and giving drug sponsors a scientific helping hand.
At the same time, the IPRG is charged with keeping that data out of the realm of regulation — indeed, the group is being built to prevent its experts from crossing paths with information that might bias drug-reviewing work in other parts of the agency.
Felix Frueh is the head of the IPRG, and has been since May. His task is to assure that the group is able to perform the job of scientific advisor from within a regulatory agency, and part of that involves building up the group from the beginning. Frueh has pharmacogenomic experience from within industry and academia, he is a member of several professional societies, including the American Association of Pharmaceutical Scientists, where he served as Chair of the Pharmacogenetics and Pharmacogenomics Focus Group from 2004-2005.
How long has the IPRG been operating?
It didn’t have an official start date, as such. I think it’s been evolving out of the working group that was assembled to actually write the pharmacogenomics guidance in the first place. Certain individuals from that group migrated into the IPRG.
We are in the process of setting it up more formally, and all of that will be announced with the release of the final guidance. In addition to that — I think this has been publicized before — we’re going to have two more documents that are going to be published at the same time. They are what we call here in CDER — MaPPs [Manual of Policies and Procedures]. Essentially what it is — it’s a standard operating procedure or a document that spells out in more detail certain aspects that would not be appropriate to put in the guidance document.
The two documents that we’re talking about in the context of the pharmacogenomics guidance are addressing, I think, the two most-noted issues during the public comment period for the guidance. One is the responsibilities and the function of the IPRG. And the second is the process on how to submit voluntary submissions, and how these submissions are being reviewed within the agency.
Can you give me any insight into those?
Well, they are being looked at as we speak, so giving you much detail about that, I think, would be inappropriate. One of the big concerns about the industry was that there is a clear separation between the review process for regular submissions — such as INDs, NDAs, BLAs and so forth — and the voluntary submissions. And what we have created is an entity within the FDA that really operates along those lines. So, essentially, we have a review group that consists of individuals within the agency that have the ability to assign others so that the regular review process is not being jeopardized.
We’ve listened pretty carefully to that concern of the industry, because, of course, that is critical to the success of the voluntary submissions. So we really have a path by which these submissions are being kept independent of regular submissions.
By being handled by the IPRG?
Yes. So the IPRG is an entity that consists of people from all centers — from CDER, CBER, CDRH, as well as from NCTR — and they are assembled with two key issues in mind. One [issue] is to cover therapeutic areas that are most important to genomic data submissions as we see development currently — for example, we anticipate a lot in cardio-renal, as well as in oncology. The second [issue] is — the obvious one, of course — to have people that understand the genomic part well enough to make sense out of these submissions, and to assign specific reviewers to the task of in-depth review of these voluntary submissions.
How many people will be working permanently with the IPRG?
The IPRG is a review group, it’s not a full-time entity within the FDA. ‘Working group’ would be the wrong name also. It’s a group of people — including office directors, team leaders, division directors, and so forth — who have the oversight on submissions in a more product context. So they see the genomics aspect of it and can make judgment calls — how to evaluate that data. [People] who have policy experience. But [they can] also identify individuals, if there is the need, for in-depth analyses — and the key there is that those [experts] are not at the same time reviewing related NDAs or INDs.
So, members of the IPRG — there are full-time members, but that doesn’t mean that’s all they’re working on. They’re assigned full-time to the IPRG and depending on the submissions and the volume of submissions, their involvement is more or less.
So these people are affiliated through the IPRG but work in different segments of the FDA?
So they’ll be splitting their time between that work and the IPRG?
Yeah — we have a process that we call a ‘consult.’ When a clinical division is getting an IND or an NDA submission and requires additional expertise for an adept review of that submission, they require a consult from whomever in the agency that has the appropriate experience to review that part of the submission. And that’s kind of a similar situation — these people are essentially being consulted to review genomic data, genomic submissions.
Now, in addition to that, I’m located in the Office of Clinical Pharmacology and Biopharmaceutics, and within our office, we’re building a core group of genomics. All of us are members of the IPRG. I’m the one chairing the IPRG. So essentially, this is creating the core group for expertise in genomics. In addition to the function in the IPRG, another task that we [in OCPB] have — and to some extent also share with the IPRG, depending on the amount of work it is — are the consults that I was just talking about. So, for example, if the [OCPB] has an NDA that contains genomic data for review, they have the possibility to submit to [IPRG] the genomics part, which we help them evaluate. So what we do is, not giving out approval, non-approval, or approvable recommendations, but just an evaluation of the scientific content of the submission that they have in front of them.
How many submissions have you received so far?
We have received a number of voluntary submissions — numerous consults.
Since you came on board, has it increased?
Yes, absolutely. When I started [in May], I think we had one voluntary submission. Now we have … I can’t split it up between voluntary and just the consults, but it’s a total of a dozen or so, if you lump them together. That includes submissions that we’re requested to take a look at as consults.
All genomic data?
So, is the data that you are given going to be publicly available?
No. It’s no different from any other submission that we get, from that perspective. Now, in addition to that — because of the issues I was referring to before — we also have a mechanism by which the voluntary submission itself is not just broadly distributed within the agency. So it really goes to the IPRG, and that’s where it stays.
Now, what we learn from these submissions is a different story, of course, and that is being handled at different levels. There are two aspects to a submission like that. One is the result that has been presented to us. At least equally important is the approach to getting to the result. And in many ways, this is something that is more generalizable. Just the general approach of, for example, conducting a whole-genome scan for SNPs or a whole RNA-expression study and then the approach to narrow down to the identification of marker SNPs or marker genes together with the statistical measures that are taken into consideration — the underlying mechanistics and so forth — to come up with a handful or so of candidate SNPs or genes that are being taken forward, this is something of that is of general interest, and there are aspects to that that certainly aren’t proprietary. You can look at the identification of a pathway and say, ‘Based on the information that we found here, we’ve made these and these decisions.’ Then you have something on hand that we from the FDA have been exposed to and have looked at, and we can say, ‘You know what? We believe this makes sense.’ And I think that’s valuble information for a sponsor, in order to implement these strategies in their drug development process.
In essence, you give their approach some validation?
Right. And we know little about this today, and what turned out to be the case in meetings that we had so far with the sponsors that submitted voluntary data, both sides walked away with a clear lesson learned — having a good idea on what the next step should be.
Will the IPRG hang onto this data or throw it away?
We will be hanging onto it. But again, it’s not going to become part of the standard review process. But it might be — and we had this case — where the sponsor wants to take the next step, which is non-voluntary data submission, based on information that we provided during the first meeting. So of course, that would be silly to just discard what information we have at this point.
Will the data be used for any future studies or some other project?
It’s also conceivable that part of the information that’s being submitted under the voluntary process is becoming information that the sponsor then decides to use in their drug-development process to reach certain decisions, which essentially triggers a non-voluntary submission. So then the data will need to be resubmitted under the regular submission path.
And the data won’t be used by any other agency for anything?
No. The process that we are putting in place is basically the same process as a pre-IND.
In earlier reports, the IPRG was to sign on three individuals, one of whom was you — have either of the other two been decided on?
Initially, I think one of these people was to split time between oncology and the IPRG.
We have someone internally that has joined our group full time now. And we’re talking to a number of other people, depending on the funding situation and the budget. We definitely see the need for having more people dedicated to genomics.
You mentioned before a split position between oncology and the IPRG — that was one attempt. We’re sort of moving away from that because of two reasons: one, it might be not such a clever thing to do organizationally; second, we’ve seen since, such a tremendous increase in genomic data that we’re looking at, that it simply wouldn’t make a lot of sense any more. We essentially decided to have a core genomics group covering the variety of therapeutic areas as a group rather than individuals. But again, all of this is very much in flux. I’m hesitant to say we have an organizational structure that we have in place today.
Do you have an idea of the budget?
No, I do not.
How many people do you expect to need per submission?
It varies from one or two people working for a day or so answering questions asked by a sponsor, which requires very little review of the material that was submitted — to very extensive submissions, where we looked at statistical approaches for quite a long time, where we are even now getting additional information from a sponsor, to look at it more closely, and see whether or not we’re coming to the same conclusion and can reproduce how the sponsor has been looking at the data.
It’s almost impossible to give an answer. We specifically designed the voluntary submission process as a very open process in terms of not exactly specifying the type of information and the extent of information that we need to see. There are several reasons why we didn’t want to do that. For example, if we talk about data standards, I think it would be premature to put something in a guidance at this point because of course, most of these standards are evolving as we speak. So you might get one type today, another one tomorrow, and you know, maybe it’ll turn out that a third one is the one that actually makes the most sense. Also, we didn’t want to limit ourselves to the review of a certain type of genomic information.
So, for example, we’ve seen a submission where a sponsor — well, essentially what was submitted to us was the idea to conduct a very limited clinical trial based on the use of biomarkers in peripheral blood. Well, no experiments were done — so we were just talking about the feasibility of conducting such a study. If we specify any data submission requirements in terms of extent, we probably wouldn’t see something like that. It turned out to be a pretty interesting discussion that we had — on a purely scientific level — on whether or not that is feasible in the first place and second of all, what the requirements would be to conduct such a study and come up with something that makes clinical sense.
Is there a major bottleneck to the release of the voluntary submission guidelines?
You really can’t pinpoint it. It’s just that this is a process that involves people from the level of the center directors at CDER, CBER, and so forth — all the big policy people are involved. And I guess we might have underestimated the complexity of getting that document out within the time that we thought would be possible in the first place. It’s really not a single issue or anything major content-wise. I think it’s fair to say that the content of the guidance hasn’t changed in the sense that there is something big that has been changed between the draft and this one.
The two major issues were addressed by creating separate documents, because it wouldn’t have been appropriate to put that in the guidance. Essentially what we’ve been asked to do is clarify more specifically the definition for the biomarkers. The definition of the biomarkers has not changed, but what we added is that the biomarkers are being looked at in the context of their use.
One of the concerns also might have been that there are going to be changes to the decision trees that being part of the guidance, and those are not changed at all.
The documents are slated to come out at the same time?
Yes. What we’re planning to do is release everything at the same time. We’ll also then have a website — it’ll be the genomics at FDA website (I don’t know what the URL will be). Essentially, what we’re trying to do is put everything onto a website that will be a sort of one-stop solution for finding out about developments in genomics at the FDA. At that website will be the guidance, these two additional documents, additional information, references to papers that the FDA has published in Pharmacogenomics and other journals describing what’s going on at the agency, the current thinking. It’s also a way to update more frequently about recent developments and new insights and decisions that are being made at the agency.
And what are all three documents?
There is the guidance, there is the document that describes the IPRG and there is the document that describes the submission and review process for the voluntary data submissions.
Is it possible that genomic data could be used as a basis for the FDA to ask for toxicological tests to be performed?
Well, certainly. [In non-voluntary submissions.]
The way that I look at is that genomics — or pharmacogenomics — is not anything new in the sense of, ‘Now, all of a sudden, we have a means to solve all the problems and issues that we had with toxicity and efficacy of drugs.’ Essentially, what you want to look for in pharmacogenomics is an additional tool for doctors to assess whether somebody is a responder or non-responder, develops toxicity, or so forth — like any other biomarker that’s been used in the clinic. If you determine bilirubin levels for whatever the drug might be to assess whether or not the patient is going to respond, now we might have the possibility to look at the genotype and say, ‘OK, now we have more information and can make a more clinically informed decision.’
Who has oversight over IPRG?
It’s an agency-wide group. The first level of oversight for the IPRG is at the Center directors’ and commissioners’ level at the FDA. The IPRG is rooted within the agency, and is an inter-center group.
It just happens that I’m located in CDER and OCPB, but really that doesn’t mean that the IPRG is located there — it’s an agency-wide group.
At the same level as other centers, as far as oversight goes?
The individuals — members of the IPRG — are recruited from all the different centers.
Do members of the pharmaceutical industry seem satisfied with the solutions the agency has come up with?
Well, in the discussions that we’ve had so far, the answer is yes. Of course, we haven’t made the information publicly available, until the release of these two documents, but based on the experiences that we’ve had so far with sponsors of voluntary submissions, it’s been very welcome. And I think [it was] appreciated.
Again, I think that in all these meetings, one thing was clear — we are starting, hopefully, to understand a lot more about genomics and how genomics can be used during the drug-development process, but no one really has the final answer on what the best approach is on doing so. So really, it’s a way to seek the dialog with the industry, and the means by which we can do that without jeopardizing the process of INDs and NDAs really does work. So we were successful in creating a path that is separate from the regular review process. The way that I like to put it is that it allows for a scientific discussion rather than a discussion that has the regulatory decision as the mantra or the overhead.
Can you tell me how comments on the draft guidelines have been incorporated or rejected?
We, of course, looked carefully at all of them, and I don’t want to say that all of them have been addressed, but certainly a majority of them have been addressed. But again, a lot of them didn’t require a change in the content of the guidance — they were more questions asking us to clarify what we meant.
So, I think what you will find in the final version of the guidance — if you compare it to the draft version — is maybe clearer language. We were trying to be more precise in what we meant. And again the two major issues were addressed in the separate documents. But that’s really it — the comments were extremely helpful for us to actually think through the process more precisely, I would say, and go through a couple of scenarios, so that we could come up with more specific language.
Do any examples come to mind?
Well, for example, the definition of the biomarkers. We now define them in the context of their use. Before it was not clear that if we have a non-valid biomarker, if this has to be looked at every time a sponsor submits data. Now it’s clear that if the biomarker doesn’t have anything to do with the type of your assay, or the question that you’re asking in the therapeutic context, it makes no sense to look at it.
So, disease state, for example?
Right. It can also mean gene expression versus a SNP in a particular gene.