Melanoma 2003

SLIDES & TRANSCRIPTS
Friday, September 16, 2003

Session II: Relevant Molecular Markers and Targets

Andrew Godwin , Ph.D. (Fox-Chase Cancer Center)

Slide 1:

DR. SCHILDER: Now we will move on. Dr. Andrew Godwin, from Fox-Chase, in a few minutes, will tell us what we need to know about signal transduction.

DR. GODWIN: Thank you for allowing me to speak today. It is a little hard to cover signal transduction in this short a period of time, but I am going to try to cover it in a means that maybe everybody can understand, and then lead up to some discussion points.

I want to start off with this, by not raising any controversy but asking, is there going to be an imatinib for ovarian cancer?

It is not that I have any financial interests in Novartis. It is just that we also study gastrointestinal stromal tumors. So, I am kind of aware of the efforts of imatinib studies in other kinds of cancer.

I am going to go backwards a little bit and reframe some of the questions. What is molecular targeted therapy, and really, it is the idea of trying to link the concept of understanding what is wrong in a tumor to how we can better treat it.

Slide 2:

Really, some of the limitations up front that I am going to talk about, really, for most tumors, are multiple factors that drive us.

As George Coukos mentioned before, we are talking about oncogenic addiction, and sometimes it is hard to identify the driving oncogene that is essential in a tumor type that is so heterogeneous.

Therefore, the view, or some of the data that is really coming out is, is a single targeted therapy going to work.

Slide 3:

I use this as an example. On the left is a metastatic GIST. You can see, just from a karyotype standpoint, it is rather simple, whereas, compared to ovarian cancer, you see a large number of gene copies and hyperdiploid state.

If you then also look genetically, 90 percent or so of GIST have a single driving mutation, either in KIT or PDGF receptor, whereas there really isn't a comparable mutation that we can identify among all the different kinds of cancers that can be considered as the driving force, that you can easily target.

Slide 4:

There obviously been a lot in a variety of different kinds of cancers that have been looked at, and degrees of tyrosine kinases that are altered throughout a variety of different kinds of cancers. Most of the success, really, has come looking at GIST and CML in these cases.

Slide 5:

When we are looking at this -- and I am going to talk about the tyrosine kinase receptors and other kinase signaler molecules -- is that we look in general if they activate by ligand, whether growth factor or cytokine. You get dimerization, activation. And then signaling.

Slide 6:

The signaling can be kind of complex. This just shows two pathways that aren't even drawn out to completion. The idea here is that, if you can go in and target some of these, can you have a drag effect on oncogenesis and tumor growth?

Slide 7:

When we look at the drivers or the oncogenic changes that take place, sometimes we try to look for those with aberrant activation.

In this case, it is ligand independent. In some cases you might have mutations that lead to it.

Again, you have constitutive signaling in this case that is independent of ligand. This can occur from acquired mutation or amplification, sometimes, of the receptor we are looking at.

In this case we come in with the hopes of blocking all the downstream signaling by a targeted therapy.

Slide 8:

When we are looking at things like, for example, PGF receptor, EGF, a lot of the other kinds of tyrosine kinase receptors, there are a number of agents that are currently, in clinical trial, targeting whether it is EGF, HER-2, a dual HER-2NU EGF, PDGF and KIT as well as VEGF.

The goal is to try to block it either at the level of ligand binding or the kinase domain.

Slide 9:

You can see that some of these things -- this is, again, an example of preclinical work looking at GIST -- that some of these drugs, like imatinib, have a dramatic effect at very therapeutic relevant concentrations, activated KIT here, total KIT here. You can see, at very low concentrations of the drug, you get a remarkable downregulation of the target and downstream molecules.

Slide 10:

So, when we really try to look at it from the standpoint of preclinical studies, trying to identify what are good targets -- you kind of can make a very simple kind of outline of what you might want to look for, looking for what are the real drivers of tumorigenesis.

Is the receptor or the gene that you are looking at really activated through, maybe, amplification, and overexpression? Are there -- again, when we are talking about autocrine or pericrine loop kinds of mechanisms, is there an increase in the level of the ligand?

In some of the cases that have been highlighted more recently, are there mutations within there that lead to constitutive activation?

Again, the question is, how are we doing at identifying these aberrant signaling pathways that are important in the genesis of ovarian cancer.

Slide 11:

I bring this up, and this is in relationship to some trials that are ongoing with imatinib, for example. As I showed you before, in GIST, that 80 to 90 percent have a gain of function mutations in KIT or PDGF receptor.

If you look here, in ovarian cancer studies -- and they have actually gone from 1998 down to a more recent one in 2004 -- first of all, the number of ovarian tumors that actually express KIT now has dropped dramatically from the earlier studies to a much smaller percentage.

In none of those cases do you find a mutation, amplification or activation. So, it begs the question, why would you target this?

However, there are going to be some data coming out suggesting, even in these kinds of studies, there are some responses that are being seen with imatinib, and that suggests that maybe we don't know what the target is.

Maybe we have gone in -- because PDGF receptor is also questionable regarding whether it is every constituitively activated -- is maybe there are some other kinases we just don't appreciate in the case of ovarian cancer, showing us different pathways that we might want to look at.

Slide 12:

The other one that obviously everyone is well aware of is the family of ERB-B receptors. Those have been targeted quite substantially.

Slide 13:

They have been targeted at the level of ligand binding and at the kinase domain, as we mentioned before.

Slide 14:

There are a number of studies ongoing with this, one which we participated -- actually, two that we are actively participating in, looking at, again, small molecules and antibody-based therapies.

What you can see from this is not all the data is present. We have got our pertuxinid trial currently going on, but the response rates, in general, are quite low.

What we did in our particular study here is that we did see that we did have a partial response from a stable disease, and this is for unscreened patients.

Slide 15:

If we go actually and look at this, to begin with, is if you start to segregate our expression in regard to response, you can see a progression-free survival. It is substantially better if you express the target, as compared to not.

In all the four patients, even though this is a very small study, who had a long progression-free survival, all were EGFR-positive.

Slide 16:

I think this begs these other two studies that came out almost simultaneous to our report, looking at both non-small cell lung cancer and colon cancer, they have shown the same thing, is the expression and -- in these cases, due to copy number changes, they are a very good predictor of response to both antibody-based ones and small molecules.

Slide 17:

There are also other approaches taken. We are now ignoring the potential signaling upstream but starting to look at other signaling pathways downstream.

The most prevalent one studied obviously is that of the PI3 kinase AKT pathway, with a number of attempts to find molecules that will inhibit these particular pathways.

Slide 18:

So, why is AKT potentially a good pathway?

Well we first demonstrated with Geotesta, almost 13 years ago now, that AKT was actually amplified in ovarian cancer and overexpressed.

There have been a number of studies that have gone on since that, to show that basically both forms of AKT-1 and 2 are constituitively activated in many types of tumors, and that is due to many different mechanisms, mutations in the large subunit of PI3 kinase, loss of a phosphatase, overexpression of upstream signaling molecules and, in addition, some of the things that I talked about before with autocrine and pericrine types of simulations.

Slide 19:

So, what is the current status of those? Well, most of the inhibitors that I am aware of at this point are fairly toxic and haven't really made it into the clinic.

However, I think there are still some in the pipeline, and there are a lot of active efforts to try to find ones that are going to be effective, at the same time, less toxic.

I have actually reviewed a number of papers recently that will be coming out shortly, describing some other screens of new AKT inhibitors.

Slide 20:

Based on the fact that the inhibitors at this level are not working well, they have looked at other downstream kinds of ones in the stream, focusing on inhibitors of MTOR.

Slide 21:

What you can see from this data is that, when AKT is activated -- shown here, strongly activated -- you also get strong activation or phosphorylation of the MTOR. The same thing, is it weak, weak?

Slide 22:

So, there have been a lot of efforts in this, again, using rapamycin and some derivatives of that, to actually target the MTOR pathway, and these are currently also in clinical trials.

Slide 23:

The other kind of half of the pathway that everybody studies is the ras-raf pathway.

Slide 24:

Again, there are some efforts in this with some people in the room that can talk about these studies better than myself, who have been looking at specific inhibitors, including BAY 43-9006.

In these cases, they originally were targeted or used in the clinic based on the fact that they were a raf inhibitor, but have later found that they have a wider range of activity. These are currently in clinical trial and showing some response.

Slide 25:

Some of the areas that we are currently quite interested in, for example, are SARC, another downstream kinase in this case, that can be activated by a number of different ones, receptor tyrosine kinases, G protein coupled receptors, integrins. There are large efforts of trying to target this particular molecule.

Slide 26:

Currently, it is again a good molecule because, in this case, it is expressed and activated. I will emphasize activated in that case, and not just expressed, but activated in many late-stage ovarian cancers.

There are some molecules -- I am just highlighting this one, it is not to give favoritism to any particular people in the audience -- and it shows, again, a broad spectrum of activity here and, again, it is in preclinical studies that have shown to have an effect on even resistant CMLs at this point.

Slide 27:

Something we are studying actively in our lab, we are trying to look at, again, molecules that are targeting the JAK/STAT signaling pathway shown here.

The reason again for this is, and shown nicely from the paper that was published last year, is that there is an increased expression in STAT-3, one of the downstream activators of JAK-2, and that STAT-3 is also highly phosphorylated in late-stage tumors, shown here.

So, we have been studying a drug that was actually identified through a JNCI screen as a blocker of STAT-3 activity in ovarian cancer.

Slide 28:

Also, outside the realm of kinases, I put up a few examples of additional targets that are considered, and I would also like to indicate that one of the things that we can discuss is that there really haven't been approaches to look for ovarian-specific targets, but these, again, are other kinds of targets that look like they might have interesting activities in ovarian cancers, inhibitors of UPA, HDAC, which the idea behind that is to reactivate certain genes that might be suppressed during tumor development or during drug treatment, and try to get around resistance in some studies that are currently going on in other cancers, PARO inhibitors, as well as individual researchers' favorite ones.

I imagine everybody in this room who is doing lab-based research has identified their favorite ovarian target at this point.

I am going to try to wrap this up. The idea of molecular medicine, what are our expectations, and I think it really goes back to the same thing I talked about at the beginning, required for tumor maintenance, can we identify these molecules and target them effectively.

Slide 29:

In that sense, too, to have optimal effect with molecular targeted therapies, we do need to couple that with the characterization of the tumor itself. So, back to the thing, get a molecular definition of tumor, and you know what they are better likely to respond to.

Slide 30:

So, what I am going to do is go back and kind of highlight a few points potentially for discussion here. How do we identify and validate new targets, and I think we have been discussing that somewhat through some of the genomic approaches we have talked about before.

How do you actually take one of these new targets to therapy, which isn't very practical at very many institutes where you can identify something and actually come up with a drug.

Highly specific target drugs may have limitations, and this is a big question that I think has to be considered.

Should you inhibit one pathway? Should you inhibit multiple pathways using multiple drugs, and so forth?

Really, then we have to get back to the thing of, how many different blockages can you have without having toxicity?

Slide 31:

Finally, which targets are the most active in groups and how can you predict it? So, it goes back to the first session, which is, can we use genotyping which, in some cases, is actively being used in the clinic for GIST, looking at KIT mutations as well as EGFR mutations for non-small cell lung cancer prior treatment.

Eventually are we going to have gene profiling and protein expression profiling that will help us to better direct the therapies to patients.

How best to combine these therapies, this is really a question that is very difficult sometimes, that Bob brought up at the introduction.

We do need relevant models, and I think they have gone a long way since Tom Hamilton first reported the mouse model for ovarian cancer in addition to 3-D culture systems, to actually start to explore, and doing some mixing and matching in animals prior to taking them to therapy.

Slide 32:

Then my last point would be, what are some of the limitations when we are doing targeted therapy that I see from the clinic to the lab bench.

Really, again, it goes back to what we talked about before, available tissues to evaluate either pre- or post-drug treatments, and the quality of those samples that we can get to actually do some analyses in, because paraffin-embedded work is still limited, in what you can do.

Then the availability of these samples really does vary, in my experience, from institute to institute, and from protocol to protocol.

Slide 33:

No text is available for this slide.