Molecular Abnormalities in Lung Cancer and the Potential for Novel Therapeutics by Ravi Salgia, MD, PhD

SLIDES & TRANSCRIPTS
Wednesday, June 14, 2000

Molecular Abnormalities in Lung Cancer and the Potential for Novel Therapeutics
Ravi Salgia, M.D., PhD

Slide 1:

DR. SALGIA: Thank you very much, Dr. Saxman, for inviting me. I am very honored to present to this distinguished panel. What I wanted to talk to you about is the molecular abnormalities and novel therapeutics in non-small cell lung cancer, and also use some examples towards the end, and also try to show you some time lapse video microscopy data that we have generated very recently, which is quite exciting.

The way I want to organize this talk is as follows, to give you just a general mechanisms overview. I apologize for not giving you everything that is available. These are exciting times, in the sense that a lot of data is being generated as we speak. I would also like to talk about chromosomal abnormalities, telomeres and telomerases specifically, as well as various tumor suppressor genes that are involved with cell cycle such as RB and p16INK4A, as well as certain oncogenes such as ras and tyrosine kinases.

Slide 2:

As a generalized overview, what we have done is put this in the context of non-small cell lung cancer. There are various abnormalities that can occur in cancerous cells at the nuclear level, at the cytoplasmic level, at the cell surface level, as well as the extracellular matrix level. It's important to realize that a lot of laboratories are working on trying to identify the chromosomal abnormalities which can give rise to tumor suppressor genes or deletion or aberration of tumor suppressor genes, such as the hot areas on 3p, 9p, as well as p53 and RB, and also dominant oncogenes such bcr-2, myc, ras, c-erbB-2, HER-2/neu. How these oncogenes, as well as tumor suppressor genes or the loss of the tumor suppressor genes, interacts with the cytoplasm are being identified, and a lot of novel therapeutic targets are being identified as well.

The e-mail I got from Dr. Saxman had a very nice review about the various therapeutic agents that are potentially important in targeting various proteins, as well as various genes. At the cell surface, it's important to remember, especially in non-small cell lung cancer, with the data that was recently presented at the ASCO meeting and the AACR meetings about the growth factor receptors such as EGFR, as well as autocrine, paracrine, or endocrine type effects that they might have in the growth of a cell, as well as integrins such as alpha/beta integrins which can interact with the extracellular matrix, and how matrix metalloproteinases may be important there as well.

Slide 3:

In order for a cell to be cancerous, certainly it has to have genetic mutations, as well as abnormalities in the cytoplasm or the cell surface. But metastatic potential is very important in non-small cell lung cancer as well. And this is just a simplistic diagram, but it relates to how complex things can be.

Here is the tumor cell as it occurs as a cluster. In order for it to metastasize to the liver or the bones, it has to do many things. But in a simplistic view it has to break off from its cluster, perhaps take along another cell in the cluster, crawl through the extracellular matrix, crawl through the blood vessel where angiogenesis is very important, roll as well as crawl in the bloodstream, and thereafter lodge onto the various cell surfaces in the various other organs to which it metastasizes, such as the liver as well as the bone. I won't discuss anti-angiogenesis agents, but I think it's important for us to realize that all of these are very important and novel therapeutic targets.

Slide 4:

This is a summary slide from our JCO article. It shows that there are a lot of mutations that can occur, as well as abnormal expression that can occur in the various oncogenes. For example, K-ras can be abnormal, with a point mutation at codon 12, and we'll talk about that a little later in adenocarcinomas, and 30% of the time in non-small cell lung cancer.

Myc amplification is more important in small cell lung cancer than it is in non-small cell lung cancer. C-erbB-2 and certainly these are important targets currently, can have increased expression in approximately 25%. And bcl-2 certainly, you can have an abnormal expression in approximately 25%.

And for every yin, there is a yang. So if you have a gain of function with oncogenes, you will have loss of function if you delete the tumor suppressor genes. And these are very hot areas, such as 3p deletions can occur in up to 50% of non-small cell lung cancers. RB B you can have deletion or altered protein expression in phosphorylation in 15%, as well as p53, where you can have deletion, point mutation, or over-expression in 50% of non-small cell lung cancer. And most recently described p16INK4A, where you can have expression of protein that is abnormal in approximately 60% of non-small cell lung cancer.

Slide 5:

Onto our first topic, which is chromosomal abnormalities. This is what was initially touted in small cell as well as in non-small cell lung cancer, helping us to determine what kind of abnormalities can occur. A lot of modeling has been going on with respect to leukemia, as well as lymphomas, where the models of leukemias and lymphomas were used to detect various abnormalities on the chromosomes. It is summarized in this slide, and certainly there is a lot of data available in the literature demonstrating that karyotypes of non-small cells are often complex. Alterations are detected on a variety of chromosomes, but commonly on 3p, 6q, 8p, 9p, and 9q, 13q, 17p, 18q, 19p, 21q, and 22q.

As an example, loss of heterozygosity has been detected on 3p, which everyone in this room knows about. For example, the FHIT gene on 3p14.2, as well as on 9p, for example, p16INK4A on 9p21, on 13q, on RB, on 13q14 and 17p, on p53 and on 17p13.1 B these are detected quite frequently in approximately 60-70% of the non-small cells. And tumor suppressor genes are being identified on the various deleted and altered chromosomes as well.

Slide 6:

As a review, [unintelligible] in 1999 presented this nice summary table where they listed various chromosomal abnormalities, as well as the various regions of deletion and the loss of heterozygosity that can occur. For example, in 1p you can have loss of heterozygosity on 1p31 region, and using these various chromosomal abnormalities people are beginning to identify more and more genes that might be quite relevant therapeutic targets.

Slide 7:

What is important about chromosomes is also at the ends of the chromosomes, because at the ends of the chromosomes are telomeres. Telomeres are believed to be quite important in our natural senescence. That is, as we age the telomeres shorten. In the telomere hypothesis for cancer generation, which various people have proposed, suggests that in cancerous cells telomeres do not shorten and the telomerase activity is quite activated.

As a review, telomeres are DNA protein complexes that cap and protect the ends of linear eukaryotic chromosomes. Telomeric DNA is composed of guanine rich repeats B TTA, GGG in the humans. Telomerase, which is an RNA-dependent DNA polymerase, synthesizes telomeric DNA in most eukaryotes, and long telomeres are present in germ cells and cancer cells via telomerase activity.

Slide 8:

This is a cartoon representation of how in a cancerous cell you can have telomerase activity, which is up regulated. That is, in normal cells the telomerase is off, and the telomeres here at the end of the chromosomes are not retained or extended. Whereas, in a cancerous cell a telomere continues to be reproduced in the various generations or various passages of cancer cells.

Slide 9:

So in non-small cell lung cancer there are a variety of laboratory groups that have shown importance in non-small cell lung cancer. Telomerase activity is increased in 60-95% in human tumor and that is a general overall number. Telomerase activity directly correlates with the malignant and metastatic phenotype. Generally, 80% of tumor tissue from lung cancer cells have telomerase activity. This would certainly serve as a potentially good therapeutic target, and certainly at the other sessions we can talk about this.

Slide 10:

As a continuation of chromosomal abnormalities, as well as tumor suppressor genes,

Slide 11:

this is a very hot area to identify various tumor suppressor genes that might be quite important in non-small cell lung cancer. As I presented initially, you can have abnormalities of p53, RB, p16 which are located on 17, 13, and 9 respectively in non-small cell lung cancer as well as in small cell lung cancer. It's also important to note that p16 is not really abnormal in small cell lung cancer. There might be a few case reports here and there. But there are other tumor suppressor genes such as PTEN. I don't really plan to talk about it, but I do want to mention this to you. We talk a lot about tyrosine kinase as oncogenes, but there are counter mechanisms such as phosphatases, which are quite important in cells proliferating. To use an example, the bcr/abl world, where if you have in CML, you have an activated tyrosine kinase oncogene such as bcr/abl, and you actually have loss of function of various phosphatases. So dual lipid phosphatases such as SHIP is downregulated in bcr/abl. I think that will come out to be true as well in non-small cell lung cancer. As an example, PTEN, which is located on 10q23, will have LOH as well as mutations. It has only been shown in less than 10% of non-small cell lung cancer, but I think we will find other phosphatases that are quite important as well.

Slide 12:

To go on to the classic tumor suppressor genes which affect the cell cycle -- Dr. Gandara presented this very nicely, so I will just give a brief overview with respect to this at this point -- basically there are two checkpoints which are quite important for a cell to grow from G1S, as well as for a cell to grow from G2M,

Slide 13:

and RB and RB phosphorylation is what is important for the cell to enter into S phase. So once RB is unphosphorylated, it is bound to E2F, whereas if it is phosphorylated, it releases E2F so transcription can occur and the cell can enter into G1S.

Slide 14:

A lot of mechanisms are available for the cell to make sure that it can proliferate, as well as go into the cell cycle. And there are a lot of checkpoints. Basically, there are a lot of proteins, as well as mechanisms available for the cell to make sure that if it wants to enter into the cell cycle, it can. What's important to realize is that RB de-phosphorylation is a critical event towards entrance into a cell cycle. Phosphorylation of RB occurs via these cyclin dependent kinases, which are then regulated by various other molecules as well.

As we discussed earlier, p16 has a negative regulatory element on CDK-2, CDK-4, p21, p27, and p57.

Slide 15:

But as an example of how one can therapeutically arrive at novel targets B this is from a Jeff Shapiro article most recently in the JCI B if you have an RB which gets phosphorylated by CDK-4, you can actually regulate this mechanism via various drugs such as cyclin D1, which regulates CDK-4. It can be down regulated by anti-estrogens, retinoic acid, rapamycin, and agents such as flavopirodol.

You can also have epigenetic phenomenon, such as methylation of p16 that occurs in non-small cell lung cancer, being regulated by demethylating agents, as well as histone deacetylase inhibitors. Or you can consider other therapeutic alternatives such as replacement of p16INK4A, such as derived peptides, as well as encoding p16INK4A by recombinant adenovirus, as well as other viral vectors that one can think about.

Slide 16:

That's all I want to say about cell cycle and the various proteins associated with the cell cycle, and how one should think about tumor suppressor genes. I want to go on to oncogenes, and spend a little bit of time with respect to discussing the various oncogenes.

As an example, ras has always been touted to be very important in non-small cell lung cancer, as many people here and elsewhere have shown. Indeed ras is a very good potential therapeutic target in non-small cell lung cancer. There are various dominant oncogenes such as H-, K-, and N-ras, and it's the K-ras that is important in non-small cell lung cancer. Ras is a 21 kilodalton molecular weight protein that is mutated in appropriately 30% of adenocarcinomas. It is commonly codon 12, which is an activating mutation, but some other abnormalities have been detected at codon 13, as well as codon 61. It is active in its GTP form, and it activates several downstream signal transduction events, and it undergoes several post-translational modifications, such as the farnesylation by FTPases to localize to the membrane.

Slide 17:

This is a simple cartoon figure that shows some very key issues. In order for ras GTP to go into a ras GDP model, ras GTP actually localizes to the membrane via this lipid membrane group. And we'll talk about that in a second. But it also can be dephosphorylated by ras gaps. These are actually very important targets therapeutically. ras GTP, once it is activated, can bind to agents or molecules such as raf, and then cause a whole signal transduction cascade, which can lead to activation of various transcription factors, as well as gene transcription.

Slide 18:

As shown recently in a [unintelligible] review article, I show that ras localization to the plasma membrane is very important for it to be active and for it to bind to agents such as raf and cause downstream signaling events. Ras has what is called CAAX box or CAAX motif, which gets farnesylated or modified by the farnesyltransferases to local it to the membrane to the lipid bilayer.

Slide 19:

So one can envision -- and this is again the review from [unintelligible] article -- which can have peptidomimetics such as CAAX mimetics. And these are the various drug designs that are available at this point, as well as farnesyl diphosphate analogs, as well as bisubstrate inhibition. And there is a plethora of compound libraries that have identified these various inhibitors against ras, which are in active clinical trials. I hope we can talk about that later today in the signal transduction session.

Slide 20:

For the final part of this talk I would actually like to emphasize tyrosine kinases in non-small cell lung cancer. A lot of data is being presented, especially at ASCO, as well as AACR with respect to EGF receptor, as to how you can block it with the blocking antibody or cause apoptosis with the blocking antibody, as well as using peptides and peptide analogs against EGF receptor, as well as the various compounds that are available that are being quite effective and potentially promising.

Slide 21:

Tyrosine kinase, or more specifically receptor tyrosine kinases, are part of a whole huge family which have a kinase domain on the side -- this is the extra cytoplasmic side, or the outer surface and the inner surface, and this is the plasma membrane. And inside in the cytoplasmic side you have a kinase domain. On the outside are the ligand binding domains. There are various family members such as the EGF receptor, or HER2, c-erbB-2, which are a part of the same family, as well as various other family members that have been identified.

Slide 22:

This is from a [unintelligible] article that has identified various tyrosine kinases, its expression, as well its association with prognosis and metastasis. If we look at this for non-small cell lung cancer, one has looked at tyrosine kinase expression, at least in the literature, and I'll show you some more examples now, you can have over-expression of HER2, HER3, c-MET, flt-1, or PDGF receptor. And some of these are also associated with prognosis or metastatic potential -- clearly HER2 and HER3, but also MET and flt-1. As a counter example, I show some small cell lung cancer, where not a lot of details are available, but c-kit, as Jeff Crystal has elegantly shown, is overexpressed and there is an autocrine loop with stem cell factor or steel factor, as well as [unintelligible]. But are they really associated with prognosis or a metastasis? It's not very clear at this point.

Slide 23:

I want to show you some recent data that we have been able to generate, and to use that as an example. We have more of this example in terms of bcr/abl, as well as in small cell lung cancer, but we are trying to investigate it at this point in non-small cell lung cancer as well. This is a drug from Novartis, STI571 that you will hear David Parkinson talk about. STI571 has been beautifully used in CML, as shown by Brian Drucker and his colleagues. We have tried to investigate its role in lung cancer.

STI571 is a tyrosine kinase inhibitor with specific activity against abl, abl fusion proteins such as bcr/abl, which we all know as CML, but [unintelligible], as well. PDGF receptor and c-kit are its other targets. It is quite effective, as shown in CML Phase I studies. But as an example, what we tried to do look at able activation in small cell and non-small cell lung cancer cells and we did not find it. But we knew that small cells, as well as non-small cells, may have c-kit and stem cell factor, and non-small cells may also have PDGF receptor, which is expressed. It is important to note that stem cell factor, which is the ligand for c-kit, is expressed in 70% of small cells and up to 30-40% of non-small cell.

We wanted to see if it was potentially useful in lung cancer.

Slide 24:

This is the data. We have determined the effects in a dose-responsive fashion on various cell lines. This is an example, a small cell lung cancer panel, but we are actually deriving this for non-small cell lung cancer as well. These are NCI-deriven 869 cell lines, H164 cell line, H249, H82, K562 which is a bcr/abl positive cell, [unintelligible], which a lymphoma cell line, which is bcr/abl negative.

Let me get to the positive control first in the sense that bcr/abl cells respond beautifully to this drug, as you can see. On the ordinate is the percent of control or viability on this graph, and the abscissa is the dose of STI571 in various micromolars. We believe that up to 10 micromolar is specific, but beyond 10 micromolar it becomes actually non-specific for CML, as well as lung cancer cells. But if you look at 869 and 864, we actually see some nice dose response of inhibition of these cell lines, even up to 10 micromolar, and also in H146 cells, with an IC50 being approximately 1-5 micromolar. There were two unresponsive cell lines, such as H249 and H82.

We then tried to figure out the mechanisms of why this should be. I will actually present the mechanism, but first I would like to present some relevant biological data. We determined that it had effects on cell cycle, with accumulation in G2M, but G1S was really not abrogated. We also showed that reactive oxygen species, which are important in non-small cell lung cancer, are reduced in the context of STI571.

Slide 25:

But the other thing we went on to show were various biological functions.

If I can ask Julio to turn on the videotape for me, I would also like to show you the data in a time-lapse video microscopy fashion. We took these small cell and non-small cell lung cancer cells and treated them with STI571, but we also used CML as our initial model to test it out.

So what you are going to see initially is a hematopoietic cell line called baf-3. What we have done is studied baf-3 for a long time. We have also studied -- baf-3 is a pre-B cell line, which is IL-3 dependent. We have studied this on various extracellular matrix components. We have studied it on things such as uncoated surface, as well as collagen-4 coated surface, collagen-1 surface, fibronectin surface. This actually represents a good model for us as to how cells may move, as well as metastasize either to the lymph nodes in non-small cell lung cancer, as well as to the distant organ sites.

This is a baf-3 cell line, which is pre-B cell line on a fibronectin-coated surface. I used that only as an example, but it's also true for other surfaces. These are nice round cells, with some projections. These projections are actually called microspikes, and they are actually dependent upon rac for their projections. In other important GTPases, which are contributing to this movement of these cells, it's actually secondary to rho, was well as ras, as well as PI3 kinase. So those are actually important therapeutic targets in the future as well, but nice cells.

In a time lapse video fashion, one minute of this tape actually equals 90 minutes of real time. We have been able to show that we can up regulate this kind of movement by adding on various tyrosine kinase oncogenes. So here is an example of bcr/abl induced baf-3 cell, which is actually quite elegant. It's almost like Pac Man, where you actually have the same cell line, baf-3 cells, which is activated by bcr/abl oncogene. You remember that's a very potent tyrosine kinase oncogene, which targets itself to the cytoskeleton. You can actually see nice ruffling, nice movement, directed movement, a lot of actin cytoskeletal projections.

These are actually very important biological surrogates that one can potentially measure. I think if we can use the example of CML, a CML-driven hypotheses, we can potentially apply this to non-small cell, as well as small cell lung cancers. But you can see a lot of filapodia formation. You can see a lot of membrane ruffling. You can see a lot of lamellapodia formation. You will also see [unintelligible], that is, the tails of these cells just being stuck. Then we took the same cells, bcr/abl transformed baf-3 cells, within three hours of treatment with STI571. So I think you can appreciate the dramatic difference in cell migration and cell motility in that here is an apoptotic cell, but you also have a very non-moving cell. Actually, that cell was moving because of the CO2 on the field.

It is very important to realize that we can study the effects of various drugs on various cells in terms of metastatic potential by looking at time lapse video microscopy, trans-well migration, adhesion assays. That is actually coming through the literature as you read through all of this.

Then we did the same thing, and I show you an example of small cell lung cancer. Here it is being videotaped at 240X. Notice this cell. It's actually blebing a lot, and we believe these are actually membrane ruffles for small cell lung cancer cell. I would have expected, knowing the clinical biology of small cell that they would be moving faster or differently, but they definitely move differently. Notice how this cell just attached itself to the other cluster. That is what small cells do, that is, they tend to group together in a cluster and they move together in a cluster. We can study again the migration of small cells as a cluster more than that of the individual small cells themselves. We're are now trying to evaluate how they move, and how various drugs have affects on these non-small cell lung cancer cells.

I wanted to show you these 869 cells. These same cells were treated with STI571, and this is between 3-6 hours of treatment. They lose their membrane blebing or membrane ruffling. They also lose their migratory capabilities.

So I think these novel therapeutics may have two separate roles. One of them being indeed that you can have cytotoxicity or cytostatic phenomenon, but can you also potentially inhibit metastases? And that is an important point to remember in terms of therapeutics, especially for non-small cell lung cancer.

The other thing we should remember and realize is that we can up regulate and down regulate these various tyrosine kinase oncogenes, as well as phosphatases, and study their behavior. For example, I took the same bcr/abl transformed baf-3 cells, and over expressed it with PI3 kinase. So this is a constitutively active PI3 kinase. You would have thought that CML cells have their maximum mobilization. But you can actually increase and enhance mobilization by adding more constituently active tyrosine kinases or other kinases.

Slide 26:

Our lab has been able to show with all this biologically relevant data that would be very useful for various lung cancers, as well as other models of leukemias and lymphomas, and especially for non-small cell lung cancer. We wondered why we had some sensitive cell lines, and some resistant cell lines for STI571. Knowing the data in the context that abl was not activated, PDGF receptors were not expressed in these small cells, we looked at c-kit expression. This is a Western blot against an anti c-kit with the various whole cell lysates of MO7E, which is used as a positive control, since this is a megakaryocytic cell line, which has a lot of c-kit expression.

We noticed indeed that the sensitive cell lines, that is 896, H146 and H209, had positive c-kit expression; H82 and H249 had negative c-kit expression. The ones that were positive for c-kit expression actually also responded beautifully to stem cell factor as well in terms of tyrosine phosphorylation.

Slide 27:

Then we went on to show indeed, as an example, that c-kit phosphorylation by stem cell factor was decreased by STI571. Here is an example. In this panel we used the various dosages of STI571. Stem cell factor was used at 50 nanogram per ml, and stimulation was done with respect to MO7 cells, as well as 869 cells. You could see that phosphorylation of c-kit was abrogated between 0.1 and 1 micromolar.

Slide 28:

At this point we are trying to figure out if this makes sense in non-small cell lung cancers as well. It has been published from multiple groups that non-small cell lung cancers, up to 30-40% may have expression of c-kit, and they might have an autocrine or a paracrine stimulation with respect to stem cell factor. But here are various cell lines. Basically you can see there are several non-small cell lung cancer cell lines which have a high expression of c-kit.

Slide 29:

So I think that would be a quite useful target.

What would be also important, as mentioned earlier, and I think we're going to mention it throughout this talk, is how does this relate in terms of hitting it with chemotherapy? How does this relate in terms of radiation therapy? Those are very important questions as well.

Finally, to end the presentation, I would like to talk about another receptor tyrosine kinase which we have been interested in, and which might play an important role in non-small cell lung cancer, and that is c-MET. C-MET is a transmembrane receptor tyrosine kinase that is expressed in a variety of epithelial cells. It is 190 KD heterodimer, alpha-beta chain. And HGF is the natural ligand or hepatocyte growth factor or scatter factor, and causes multiple biological effects.

It may be over expressed in approximately 30-40% of non-small cell lung cancers.

Slide 30:

Expression of c-MET is much more common than c-kit. The same panel was re-stripped and probed for c-MET, and you can actually see approximately 7 out of 11 non-small cell lines have an expression of c-MET.

Slide 31:

This is actually a functional c-MET.

What we believe is that c-MET actually works in a paracrine fashion. That is HGF is produced by the stromal cells, and c-MET is expressed in non-small cell lung cancers, and it might be quite important in terms of scattering or metastasis, mobilization, as well as other biological effects. But one of the effects that we have been studying from a biochemical perspective is looking at signal transduction events with respect to HGF.

And one can see in a dose responsive fashion when you add more HGF to the non-small cell lung cancer cells, you can actually appreciate that there is increased tyrosine phosphorylation as dose progresses.

Slide 32:

In summary, there are multiple pathways that are targeted in non-small cell lung cancers. Novel therapeutics are designed against these pathways, and they would be quite useful. And combinations with conventional therapy and novel therapy would be useful to consider.

Slide 33:

Finally, I would like to acknowledge my laboratory with Dr. Wenlan Wang, Dr. Gautam Maulik, and Mary Ellen Healy, who did most of the experiments you saw. Martin Sattler, as well as Jim Griffin for their useful insights in the bcr/abl world, and Arthur Skarin for his incredible insight into the clinical world for non-small cell lung cancer, and Bruce Johnson for his leadership.

Thank you very much.

[Applause.]

Slide 34:

DR. SAXMAN: We have a few minutes for questions for Dr. Salgia.

DR. GANDARA: Are there any clinical trials in lung cancer yet of the STI571?

DR. SALGIA: Not that I know of. Maybe David can talk about that.

DR. PARKINSON: No, not yet. The reason relates to supply issues. There are plans in a number of different tumors including lung cancer to do proof of concept trials, and they are being put in place right now. So as soon as supplies allow, and that may be within the next two or three months, we expect to start limited studies. Then by the end of year or January, the supply issue will no longer be a problem, and we'll move into much broader studies. That's why we are working together with the NCI on this, to expand this way beyond the registration direction and proof of concept.

DR. SAXMAN: Ravi, I have a question about the models that you are using. They are very elegant. Is there any way to look at some of the effects of radiation with these models, or are these models going to be purely useful in terms of looking at pathway inhibitors?

DR. SALGIA: No, actually I think we can look at it from the perspective of radiation and/or chemotherapeutics, and in combination as well. That is actually being done as we speak with some of the cell lines. What we would like to do is to look at cell lines and comparable cells lines which are radiation-resistant, and compare and contrast the cell motility, cell migration. That would actually be very useful. In the context of these novel therapeutics, can you combine radiation plus the novel therapeutics and look at the cell motility? I think that would be useful as well, and that is being planned.

The other thing we are planning is actually trying to model a lymph node surface or a bone marrow surface, to look at the metastatic potential of non-small cell and small cell. Can we study that? What we have done is lay down a stromal layer, and then put on these non-small cells or small cells, and then study the effects of inhibition in terms of paracrine, as well as autocrine fashion.

DR. GUMERLOCK: That was wonderful video microscopy. The one thing I didn't see was any cell division going on. One question I had was, with all that motility and ruffling, does that influence the replication of the cells?

DR. SALGIA: Absolutely. These cells actually divide very frequently. Let's take the CML cell as an example. They will actually stop ruffling. They will stop mobilizing. They will stop migrating. They will adhere to cell surface, whatever surface that might be, for example, fibronectin. Then they will replicate. They will divide. They split at 120 degrees, which has been classically described about 30 years ago by [unintelligible], and then they will start moving very quickly again. You didn't see that because it's only a one minute tape. But certainly if you are interested, or if anybody else is interested, we can present all the tapes to show cell division, which is quite elegant as well.

DR. SCHMIDT-ULLRICH: I want to come back to the radiation-related question. How many parameters do you measure and quantify in the cell mobility experiments? And how soon do they occur? Because with radiation you have the unique challenge that you have a very brief exposure, and then you have cells reacting to that. That may occur over minutes and hours. And so to quantify these effects and link them to specific genetic events or expression of certain genes would be very critical. And I was wondering how flexible that system is, or how much you can carve out these individual motility parameters for those studies?

DR. SALGIA: That's the beauty of more imaging and technology. There are a couple of things. We can actually constantly observe these cells over 72 hours. So I think that question about how can you observe and can you quantify the difference between 1 hour versus 24 hours versus 48 or 72, we can actually do that experimentally.

Quantification is a very tough question in terms of cell motility. There are two ways people do it. One way is a semi-quantitative fashion where we can say, well, this cell is ruffling for so much, and we grade it from 0 to 3 and become pathologists at the time. And then we can say it has formation of filapodia, formation of lamellapodia at this time. And we have actually published with respect to that.

But there is now data available as well as tools available that you can actually do quantitative measurements. As an example of uses, there is the NIH image analysis program, which is available on the Internet. That actually is quite elegant in terms of quantifying exact distances moved and speed. We can also figure out not only the speed, but from a factorial based velocity, as well as how much time the cell spent in dividing, how much time the cell spent in ruffling. So I think the quantitation is coming. As we get more and more technologically adept, we should be able to do that very elegantly.

DR. MABRY: I have two questions. The first question is on your video microscopy. Before that you showed that some small cell lung cancer cell lines are sensitive to STI571, others are not, and then showed ruffling with 869, which is sensitive. Does that mean that H82 cells, for example, which are insensitive or relatively so, do not ruffle natively, or STI571 does not prevent ruffling?

DR. SALGIA: H82 and 869 cells ruffle approximately the same. They also travel in a cluster in a very similar fashion. The only cell line that we have investigated under time-lapse video microscopy that is very different looking is H209. These are very big clusters, big spheroids basically, and they tend to have endomitosis as far as I can tell you. But H82 and H69 from the time-lapse video microscopy on various surfaces such as uncoated, fibronectin coated, collagen-1, collagen-4 are very similar in terms of membrane ruffling. But the 869 cells are the only responsive ones that have decreased ruffling in the context of STI, but H82 do not have decreased ruffling.

DR. MABRY: Then my second question, which you may have alluded to, do you have any proposed mechanisms with relationship to c-kit phosphorylation and ruffling?

DR. SALGIA: I think so. in the sense that it has to be a downstream signaling event. And the most common scenario would be either raf, rho, or PI3 kinase.