So, our task was to look at how do we generate new targets for the pipeline. We identified basically two problems. That is, we need to identify and validate new targets, and this does require human tissue resources.

That is obvious for cancer in general, but for melanoma, this may be a little bit more different, and there are unique issues related to melanoma that make this a little more difficult, particularly the number of cases and the size of the lesions.

So, we had a lot of discussion first about the definition of a target and what constitutes a rational target.

TOP

This actually was a little bit controversial and took us a good hour to get through, sort of the approach of, is it all genes or does the microenvironment also play a role.

TOP

So, we came up with these sort of criteria for what would be considered a rational therapeutic target.

TOP

That would include that this particular target would be essential for melanoma maintenance in vivo.

That could be the result of a genetic abnormality, such as activation or inactivation, which could be mediated by a gene mutation, translocation, amplification, deletion, or an epigenetic mechanism.

There was a lot of discussion about the role of the microenvironment. So, other factors need to be considered, which could also be considered as a target, which would be non-genetic alterations such as the tumor compartment and the stroma component.

So, these are the sort of things that we need to see in developing or searching for potential targets.

TOP

The second issue we looked at, additional criteria for a therapeutic target would be the trackability. That is, what kind of target can we actually target right now. So, the drug-ability, this word that we have used a lot these last few days.

Some potential things that we could actually target would include cell surface receptors and ligands, catalytic domains, binding pockets, enzymes.

While we don't have the technology at this point to target transcription factors, in the future, that might be a possibility.

Whatever we were selecting should have a reasonable therapeutic window. After all, we are targeting essential pathways. So, there had to be some selectivity or some reasonable therapeutic window in terms of toxicity versus efficacy.

It would be very helpful to have biomarkers that we could measure the effect on a given target.

TOP

We identified two major impediments, which was basically knowledge and some technical issues.

TOP

Really, the key to identifying new targets in our field is to have a much better understanding of the genetics and the biology of both melanocytes and melanoma.

TOP

So, we need, really, to advance our understanding in key areas, including melanocyte biology, development, melanin biology, resistance to apoptosis.

This has been a recurrent theme about the role of apoptosis in melanocytes and melanoma, and a greater understanding of this, homeostasis.

In terms of melanoma, lots of different areas, but the area we focused on was in diagnostics, pathological diagnosis versus important genetic and genomic information, greater understanding of the stroma, gene-gene interactions, gene environment interactions, and the genotype phenotype correlation.

I mean, these are obvious things, but we still think we are a little bit behind in melanoma compared to some other fields, and greater emphasis in terms of the biology.

TOP

Then, there are currently some technical limitations. This is true for all fields, not just melanoma.

One is that, for a lot of the studies that we are going to discuss, the need for fresh and frozen tissue, again, a particularly difficult challenge in melanoma.

The need for high throughput cell based screens, and we are still lacking good models, particularly animal models, and we need better, faster, more robust models to help us identify new targets.

TOP

So, how do we generate a target pipeline? I divided this up into materials and methods.

TOP

Essentially, we thought, as everyone is doing, there is a genomic approach, and that would involve a genomic screen.

This is where we would take annotated clinical samples, meaning samples that we have good information about the history of a particular patient, or we would be able to collect it prospectively, and that this would include both dysplastic nevi and primary melanomas.

There would be extensive profiling. That would be array CGH, expression profiling and, in the future, perhaps proteomic profiling, but the group did recognize that at this point we wouldn't know what we would be looking for in terms of proteomics, but to set things up now to collect the relevant specimens so that later, when we had the knowledge and the technology, we could do that. Then, genome resequencing when appropriate.

TOP

Additional ways of looking for new targets would be genetic screens looking at in vivo and culture systems with growth and survival end points. Example approaches would be RNAi and mutagenesis.

Then also, once a pathway has been identified, the traditional way of comprehensive evaluation of that particular pathway.

It would be important to establish functional screens looking at cell lines, again, looking at high throughput screening and organotypic cultures, such as Meenhard's model.

Again, these are models that would be able to screen things in a more rapid way than maybe some of the models that we currently have.

TOP

This is where we think it is going to be critical in terms of the success of this kind of approach, which is looking at the materials.

It is going to be important that the samples that we analyze be selected very carefully and in really close collaboration with clinicians and pathologists.

So, what we put into an extensive array analysis is going to be worthwhile because we gave the basic scientist the appropriate relevant specimens.

So, this really requires very close collaboration between melanoma pathologists and clinicians and the basic scientists to work together to generate the materials and then the appropriate methods.

This will ultimately lead us to, we think, identify new targets. It could lead to better molecular classification of melanocytic neoplasms, and this kind of approach, class discovery, looking for new genes, and ultimately we would be able to identify appropriate patient populations for specific therapeutic approaches and to identify important prognostic factors.

TOP

So, we would see that such an approach would allow us to identify a relevant target. We would characterize that target, importantly, validate the targets as Frank talked about, and would lead to preclinical investigation and clinical investigation.

In some of these situations, the target is identified and there is already a drug. So, all of these steps would not necessarily be required.

TOP

In terms of identifying new targets and additional validation, we need better model systems to help validate and characterize the targets.

So, we need good models and we spent quite a bit of time discussing what would be a good animal model. We thought that those models should have, in terms of looking specifically at mouse models, good biological and pathological similarity to the human condition, that it should recapitulate the genetic evolution in human tumors and that, if one was looking at engineered mutations, it should mirror those in terms of human predisposition to the disease.

Acquired lesions during natural evolution should mirror changes seen in human melanomas, and looking at models with more of a natural microenvironment and evolution, so getting away from looking at B16 and tail vein injections as a not particularly relevant model to be looking at new targets and validation of targets.

TOP

So, we really only have one over-arching recommendation and a small minor recommendation, but we really feel strongly that an international consortium should be established to allow us to collect the relevant tissues to sort of get the field moving.

We really need a starting point to do the array analysis, and that only can be done with cooperation of participating institutions.

I would say that this is really an incredible recommendation and I hope this is something that does move forward, because the field cannot move forward unless we have the starting point to identify new genes and new targets.

Because of the need for fresh tissue for a lot of the expression profiling, this is not easy to do. So, if 10 samples of primary tumors frozen from each of us could be contributed to a consortium and then do the analysis, I think that would be critically important.

This data that was generated would be in an open database and it would be available to all investigators. So, if a gene was identified, then that particular lab with that interest could go with it. So, it would be a huge resource for this field, and for other tumor types as well.

TOP

Critically, we need the bioinformatic support to analyze the data that are generated. We would also see in parallel -- this would be sort of an example of what we would be looking for in terms of establishing a tissue bank, which would be to get 100 to 150 frozen primary melanomas for genomic experiments.

Again, this is why we need close collaboration with our pathologists, to get the appropriate specimens. Then look at maybe perhaps 1,000 melanocytic lesions from paraffin, but looking at all stages of tumor progression for tissue microarray studies and other types of studies.

Also, looking at various histological subtypes of melanoma. So, not just different stages of progression, but including the different histologic subtypes.

This is Boris' wish list, to have diagnostically ambiguous melanocytic lesions, also, to further characterize.

TOP

At the same time, it would be a collection of well characterized melanoma cell lines, which would also give us some additional information and hopefully identification of new targets.

These would be profiled and genotyped as for the frozen tumor specimens, and then additional assays could be performed to further validate and characterize a target.

So, our main recommendation is for the development of an international consortium to develop a tissue bank to be able to get the field moving, and then secondly, which has been echoed, I think, previously, is increased communication between the basic scientists, pathologists and clinicians.

Meetings such as these, I think, really move the field forward, and I think Meenhard's meeting in June will help facilitate that.

So, continued discussions, workshops, meetings are critical to identifying targets and new therapies for our patients. I will end there. Menasche or Pam or Meenhard, is that sort of the summary? Okay.

TOP

DON MORTON: I just want to make you aware that my tissue bank is available to the general research community on a collaborative basis.

If you will go on the Internet, NCI tissue, you will see all the details and the IRB hoops that we have to jump through.
If that is the problem, I have about 8,000 individual samples of melanomas frozen over the years.

DR. SCHUCTER: Do you have primaries frozen?

DON MORTON: I don't have primaries, but I think what kills people is not the primaries, it is the metastases.

PARTICIPANT: What do you mean by frozen primaries?

DR. SCHUCTER: To take it from the primary and getting, at that time, fresh tissue. So, sectioning a piece of the primary.

PARTICIPANT: Really, just also, to echo what Don has said, we have got a fairly extensive tissue bank as well, and we have had requests by various people in Australia and New Zealand for access to the tissue bank.

When we actually look at the people who are going to do the work, we haven't had enough confidence that they are actually competent to do the studies.

I suppose one of the things we would like to see before we give the tissues to anyone, that the studies are going to be done competently and well.

I just wonder, how are you going to address that particular issue? Are you going to offer a contract for someone to do this work and look at their --

DR. SCHUCHTER: We did not go the next step in sort of developing what the consortium would look like and how the funding would happen, but I think you are absolutely right, that it would have to be absolutely clear in terms of the expertise and the quality of the group doing the work. That is critical.

PARTICIPANT: I agree, that the B16 melanoma cell line is a pretty hopeless thing and it has been the death knell of many therapeutics.

I do think that the subcutaneous model does have some value. I think the problem with going to a wholly transgenic mouse-based system is that you will cut out people who don't have those sorts of facilities.

I think you also have to remember that we have much more relaxed systems for testing these things before we put them into patients.

If you had a system where only drugs that have been through this transgenic model will be tested in patients, you will fall behind the Europeans, who will simply carry on with the subcutaneous models, and they will beat you hands down with all the new drugs.

PARTICIPANT: There was a little bit of confusion there. It depends on the question that you are asking. Once the target is identified, yes, we are going to use the subcutaneous and the IV models to see, what is the contribution of this. So, we are not dismissing it. So, let me make it very clear.

PARTICIPANT: Just one addition to your excellent presentation. You know, I really think that, overall, academic biologists don't pay enough attention to chemistry and they sort of defer it to the pharmaceutical companies.

We all have fine chemistry departments. For example, on transcription factors, there is a whole field of inorganic mimetics that have been developed that people don't seem to be aware of.

That is probably true for every target you have listed up there, that your chemists are probably working on it and you are not even aware of it.

I think that enjoining academic biologists and chemists is really a very important thing to try to do, so that we don't have to always be primarily dependent on the pharmaceutical companies.

DR. SCHUCHTER: I think that is a good point.

DR. SOSMAN: Just a couple of things. First, I think the fact that Don and Peter just said that sort of gives a momentum right off the bat.

They really do have incredibly powerful -- even though it is not the primary lesions, they have clinical follow up.

If you could even get those together and set up a system in place so biologists could study those. I think that would be an incredibly positive effect.

I guess the other -- two other things. One is this whole issue of, I think all of us are sort of aghast at this idea of trying to get frozen tissue from a primary.

There are probably technologic advances, and whatever they are now, and whatever they become, it should be fairly rapid.

So, if that can be disseminated, whoever is trying to do them outside the two centimeter primary melanoma, which may not be really representative, and for the melanomas you want, or we talked about needing, those were not the cases that maybe would be that representative.

Then, I think the third, which could go under either of yours, I think these tumors modelers and clinicians have to get together to talk about therapeutic trials in these mice, and try to at least set those trials up right so we learn the markers, and so those of us who have been so frustrated by human trials that we can take some of that and improve some of the trial development in the mouse models.

PARTICIPANT: Lynn, I think everybody here would agree with your blueprint of collecting all stages of tumor progression as a means of studying melanoma.

I would want to reemphasize what Don said. If we are looking for therapeutic targets, if that is the problem, it is metastatic patients that kills patients.

I don't think I even saw that on your list as a stage of progression to focus on, and I think it is the phase that should be focused on.

I think distinguishing between progressive microscopic metastases and dormant micrometastases also ought to be a focus.

At least in the adjuvant group, we spent some time talking about the fact that dormant micrometastases might be different from progressive macrometastases, and yet, if we could control them, we could cure cancer, even if we couldn't actually destroy them. So, I would recommend adding an emphasis on metastatic disease.

Going back to the earlier disease, I don't think there is any question that it is ethically challenging to collect fresh tissue from early lesions of melanoma, because you really do need to identify those patients who do have some risk and distinguish them from those patients who have no risk.

At the moment, the only way to do that is really by total examination of the lesion by pathology. I am grateful for the fact that the pathologists haven't been picked on very much at this meeting, and for Peggy Tucker for pointing out the rarity of the truly advanced tumors that are relatively easy to collect.

It is a difficult issue to say -- and it may be impossible to say -- that you can truly collect a substantial amount of tissue from early primary melanomas.

That leads me to, I guess, follow up on Jeff's comment. New technology, I think the new technology that should probably be emphasized in this report is technology that enables us to address these questions in paraffin sections.

DR. SAXMAN: Let me just comment on that. I was going to comment on that in a minute, but since you raised that as an issue -- I am sorry for interrupting Don -- but I am not going to let you off the hook.

One of the unique problems with melanoma in terms of tissue banking, I don't want to get into the broad issues with tissue banking, but there is a unique aspect to melanoma, and it is just what you described. It is the struggle between the needs of the diagnostician and the needs of the researcher.

So, if we are going to get to some of the early genetic effects that occur in melanoma and dysplastic nevi and those sorts of things, we need you pathologists to figure out how we can do both of those things at the same time, the needs of the patient and the needs of the researcher.

So, how are you going to do that? How can that be done so that both of those things can be served in this disease.

I think for several reasons that is very unique to this disease. I mean, you can take out a pancreas and the pathologist is willing to give you a big old chunk of tumor, but the pathologists can't do that here in melanoma, as it relates to the primaries at least.

DR. ELDER: If I can just respond to that, I am glad you rephrased your initial comment, the needs of the diagnostician, to say the needs of the patient.

The reason that it is ethically challenging to take chunks of fresh tissues out of early lesions is that the patient needs to be assured that their lesion has been thoroughly examined, so that those patients who have some risk will be aware of it.

Yes, I believe that the pathologists, but also everybody else, needs to be working on methods of getting around this difficulty.

I would suggest that microdissection, amplification of paraffin embedded DNA and RNA is the way to go.

Boris Bastian has produced an enormous body of work, I think almost entirely out of paraffin sections. Now, DNA is easier, but I think we need to be working hard on RNA. Of course, proteins are fine out of paraffin sections as well.

PARTICIPANT: I think this is an important enough issue for moving ahead in melanoma research, and unique enough to melanoma that it might be worth having some sort of workshop where industry investigators, pathologists, patient advocates, are brought together and could discuss some of the strategies that people are using, investigating, maybe bring some new ideas to sampling or working with paraffin or some of the clever ways that some institutions have used to get these tissues, or potentially places where one could actually find some samples that are maybe stored somewhere that you didn't know about, that might be able to be used for this.

I think we could actually move the field forward if we could just figure out a way of getting 1,000 primary melanomas banked and classified.

PARTICIPANT: I think one of the biggest hurdles is going to end up being, especially in this country, is going to end up being the regulatory things that we are going to have to run into.

As somebody with a relatively large tumor bank of my own, trying to share these tumors with anybody else is horrifically difficult in my institution.

The biggest hurdle I think we have run into is with all the new regulations, both HIPAA and IRB regulations, that there is no consensus right now.

So, everybody's individual institution has their own lawyers who say, this is what we can do and can't do, and it may not be the same thing that the other institution down the road says.

So, one of the things that we probably need to do is get at least a national agreement as to what are the guidelines that you can do here, so that one person doesn't get burned in their own particular institution, while the other institution might not have the same problems.

DR. SONDAK: Actually, there are very clear guidelines that make it probably easier to share anonymous de-identified tissue outside your institution than within your institution.

This all came from the stem cell experience, actually, because you had to be able to certify that you knew stem cells didn't come from an aborted fetus.

Nonetheless, that evolved into guidance that says, if you are sharing information outside your institution, all you need is a written agreement that you will not share any identifiable data, and it is treated as exempt from ongoing IRB supervision. I think that is one of the little known facts, and we can talk about that off line.

PARTICIPANT: I know, but I think it is also that there are a lot of differences in IRBs, that everybody's IRB is their own little kingdom and they will make decisions they want to make.

You know, identifying information, if you have a relatively small group of patients, maybe 50 patients you see in your clinic, and half of them are women, so your identifying information is that she is a female in this age group with this kind of melanoma. Well, that narrows it down to 10 people. So, then she gets all upset that she has been identified. There are a lot of very tricky things here.

DR. SCHUCHTER: We are not going to solve that today. Is that okay?

PARTICIPANT: I know, I just think you have to address that, though.

DR. BASTIAN: I just wanted to comment on the frozen material and the ethical challenges. I agree that it is problematic to set material aside, but I think it is, in fact, clinical practice that, due to poor biopsy technique, which I think is becoming the standard of care, partial biopsies of melanoma, this is actually happening all the time.

I think it is a very common cause of misdiagnosis of melanocytic neoplasms, and it is a cause of misstating. If we can address this problem and actually get back to the point where we actually see, for the majority of the lesions, the entire lesion in paraffin or at least excised, I think there will be actually no loss of standard of care, if we set aside a little bit of material.

It is happening. The dermatologists are shaving the pigmented lesions. I think that becomes the predominant mode of dealing with them, and I think that it is important that we address that.

DR. BRADO: Victor Brado from Anderson. I just wanted to make two comments. One was in support of what David was saying. Unlike prostate, as you know from the discussion, we give 10 different prognostic factors.

You know, in prostate, you give the degree of differentiation and in pancreas it is the same, you are given the degree of differentiation.

In melanoma, you know, most of the staging, 99 percent of the staging, is just given by the histological analysis.

We pathologists, we do the most we can, which is to submit the entire specimen. I like what was done before, 10, 15, 20 years ago. We submitted everything. So, everything was in paraffin.

The problem is, you cannot simply decide, okay, I am going to put half of it for paraffin and the other half for frozen, because what about if the ulcerated area is in the frozen tissue? You never know.

As you know, this is upstaging, completely one more stage, just for the presence of ulceration. Therefore, I agree with what David was saying.

Definitely, we need to do more of our studies in paraffin sections. I know it is more difficult now, but Boris has shown how CGH can be done in paraffin sections.

Then, supporting what was said about the IRB, in our institution, if we wanted to provide any tissue to anybody, it has to be without age, without sex, without clinical information. So, you only know it is melanoma. That would be all the tissue that we could share in a tissue bank.

JOHN KIRKWOOD: I think, working closely with pathologists, our unit has put down hundreds of nevi, dozens of primaries. The problem still remains, even though you have a researcher, pathologist, clinician, all working together to get these frozen specimens without compromising the pathology, they are so small that even the research needs of the investigator who, with the pathologist and clinician decides what parts of the tissue to put down, are not satisfied.

So, the need to amplify, to reliably amplify the specimens, becomes critical. I think the technical issues need to be addressed.

MEENHARD HERLYN: Just a brief clarification on the metastases. The committee had certainly considered to study the different subtypes of metastases, but that was not considered to be a major problem in accrual, so it didn't show up prominently in the list.

DR. SCHUCHTER: That we could collect frozen metastases readily easily and you wouldn't need a consortium to get that.

DR. HALUSKA: I think we all agree that we need human tissue samples to validate things, but I would like to point out that none of the important melanoma genes were discovered in human samples.
p16 was discovered in a panel of cell lines at Murigenetics. Pam Pollock has shown, and Rich Murray have shown how BRAF was demonstrated from a cell line. Ras was cloned from a cell line, PTEN was cloned from a series of cell lines.

None of the targets actually came out of human samples. The most important step in their application to melanoma was the understanding that, in those samples, the distribution of melanoma suggested that they might be true in vivo, and all of that was validated.

So, discovery does not necessarily need human samples to proceed. So, that is just a point I would like to make.

The other question I wondered if you guys talked about, and we were hoping to get out of the committee is, one of the issues with regard to new target discoveries is, which of the technologies that are presently available are likely to be fruitful and which are the best applicable. Did you guys talk about that? Do you think that profiling is actually going to generate targets? Is high throughput sequencing something we should continue to pursue as a melanoma specific endeavor?

DR. BAR-ELI: First of all, Frank, this is exactly what we would like to change. Not cell lines, let's do it on the real specimens in real life.

Let's face it, all these discoveries in the cell lines, where did it get us right now? This is exactly what we would like to change. So, this is one comment.

The other thing is, we are not dismissing any technical new tools. If it would become available, we would like to use every one. We would like to start with CGH and expression profiling, but proteomics and protein protein interactions will be very useful and important, too.

DR. MORTON: I have a comment about animal models, B16 in particular. When I first went to the NCI in 1960, the research was being done on B16. I think, 43 years later, we are still working on B16.

I am reminded of something that Sir Peter Medawar said. When the surgeons showed you could transplant organs, and he had said, organ transplantation would never be successful. He said, we live under the tyranny of the skin graft, because they could not successfully bridge the transplantation barrier of skin grafts.

Well, B16 is very, very similar. Any similarity to true melanoma is purely coincidental. I really believe that our basic scientists should go back to the lab, generate new melanomas and ask the question, do newly induced, syngeneic melanomas behave the same way as B16.

I would be very surprised if there are not differences. So, I would throw a challenge out to the basic scientists.

I bet you still 90 percent of the animal tumor research is done with B16 melanoma. I used to say, when I was doing tumor immunotherapy experiments in animals is, if it has got a name, it is too far down the line to study tumor specific antigens. I just throw that challenge out.

MEENHARD: Just a quick comment. There are now at least three models in the mouse for melanoma, which could potentially be used for any drug screening and validation studies.

I think, Frank, we all agree that cell lines are a wonderful tool, but we would need fresh tissue for validation of whatever targets we are looking at. So, we cannot get around it and Menasche made the same point.

DR. SCHUCHTER: I mean, you can do these sort of in parallel, if we had fresh tissue and we do have paraffin specimens. We can do CGH and expression profiling, but the proteomics was not ready for prime time. Thank you.

TOP