Summary




SLIDES & TRANSCRIPTS
Wednesday, June 14, 2000

Mechanisms of Resistance to Radiotherapy in Non-Small Cell Lung Cancer:  How Might We Overcome Them?
W. Gillies McKenna, M.D., Ph.D.
Slide 1:

DR. MC KENNA:  I'd like to thank the organizers for inviting me to this meeting.  The topic I was given was mechanisms of resistance in non-small cell lung cancer.  But before I get into that topic,


TOP
Slide 2:

I want to point out that not all failures of patients treated with radiotherapy in non-small cell lung cancer is in fact due to resistance of the tumor to radiation.  A lot of it is simply the fact that there is only so much radiation that you can give any given patient, and only so many clonogens that you can kill with that dose of radiation.  In lung cancer there is a very high tumor burden in many of the patients at the time that you initiate treatment.  So this is not truly a problem of resistance to treatment.  This is a tumor burden problem.

We also know that tumors can proliferate during treatment itself, since radiation is given over six or seven weeks.  And in fact, that proliferation can accelerate during treatment, and this also becomes a tumor burden problem, because this increases the number of clonogens that you are trying to sterilize.  But in addition to this problem, there are problems in the radiosensitivity of lung cancer, due to both physiological effects and genetic effects.  It's these problems that we are concentrating on at this meeting. 

TOP
Slide 3:

But if you look at the field of radiation oncology as a whole, we have until recently been a relatively empirical therapy, deriving our treatments just from experience.  But there is now a divergence in terms of development of targeted therapies.

We are interested in this arm.  Can we develop biologically targeted therapies for the purpose of this meeting?  But I would have to say that in the field of radiation oncology as a whole, a lot of effort is going on to physical targeting.  Attempting to get more dose into the tumor to overcome the tumor burden problem.  And a lot of those developments are being driven by imaging.

TOP
Slide 4:

So that we really have three ways of improving radiotherapy outcomes in lung cancer.  One would be improved physical targeting of the tumor, and then molecular and physiological targeting to overcome mechanisms of resistance.

TOP
Slide 5:

Physical targeting

TOP
Slide 6:

is largely being driven by the improved imaging.  We can now visualize as three-dimensional structures, tumors much more accurately relative to the normal tissue that surrounds them.  And that is encouraging radiotherapists to develop techniques to try and spare the normal tissues, and escalate the dose to overcome the tumor burden problem.

TOP
Slide 7:

So in lung cancer you are seeing conformal therapy and intensity modulated therapy developing.  We also have improved methods of dosing calculations, allowing us to account for the non-homogeneity within the lung, which is a major problem in radiotherapy.  And there is some interest developing in some centers around the country in the possibility of developing proton therapy for improved dose localization. 

TOP
Slide 8:

This is likely to accelerate, because we're now going to get improved both functional and anatomical imaging.  We can now do image correlation using multiple modalities.  We have much better volume visualization, and we are now doing a lot of dynamic and serial imaging to look at organ motion, which again, many radiotherapists believe is a significant problem in the treatment of lung cancer as obviously the volume of the tumor and the lung that is in the field vary continuously as the patient breathes during treatment.

TOP
Slide 9:

So one of the major developments in radiotherapy technique for overcoming both tumor burden and tumor resistance will probably be in the development of intensity modulated therapy, both for x-rays, and potentially even for protons.

TOP
Slide 10:

The main purpose of this meeting is not to talk about these developments, but to talk about molecular targeting and other mechanisms of targeting. 

TOP
Slide 11:

One of the points that I think we need to make about resistance to radiotherapy, unlike resistance to chemotherapy,  is that we cannot in fact select resistance to radiotherapy in tissue culture or in animal models.  It does not appear to arise in response to therapy in the patients.

TOP
Slide 12:

So if you look at the distribution of radiation sensitivities of various kinds of tumors, they are relatively predictable.  And they are intrinsic properties of the tumors themselves.  But we also know if we look at adeno and squamous cells that there are variations in sensitivity within these tumor types.

TOP
Slide 13:

We also believe that these variations in sensitivity are a pre-existing condition in the tumors prior to the initiation of treatment, and we have some evidence that supports that view. 

TOP
Slide 14:

For example, this is a study of squamous cell carcinoma of the cervix done in England in patients treated with radiation alone, where the radiosensitivity of the cells or the tumor was determined prior to any treatment being initiated.  Then it is compared to the survival of the patients months after treatment.  You can see that there is a highly statistically significant correlation between patient survival and radiosensitivity of the tumor prior to any treatment being delivered.

TOP
Slide 15:

So if you believe that differences in radiation sensitivity are intrinsic properties of the tumor cells that exist prior to any treatment being initiated, one hypothesis that I would like to suggest is that since the tumor can't possibly know that it is going to be exposed to radiation, so it can't have been preselected to be resistant to radiation, the properties of resistance to radiation must arise from some of the basic, fundamental properties of the tumor cells themselves.

And this is from Hanahan and Weinberg's article in the millennial issue of Cell, where they suggested that you could reduce the basic capabilities of cancer to six.  This was their list of six, and I think you could debate this list.  But nevertheless, they are saying that all of the genetic changes that you see in cancer cells can be reduced to a certain limited number of categories.   I would also like to suggest that it should follow that some these changes in these acquired capabilities should also result in alterations in radiosensitivity of tumors, and what we see clinically as radioresistance.

TOP
Slide 16:

The first data that suggested that might be true came from Jim Mitchell's lab at the NIH during the small cell studies in the early 1980s, where Des Carney looked at the large cell variants of small cell compared to classic small cell, and showed that the large cell variants were much more resistant to killing by ionizing radiation when studied in tissue culture.   He also found a genetic change that correlated with this phenotype in that they showed amplification of c-myc relative to the classic small cells.  We know from Drew Turrisi's subsequent study of small cell in combination with chemotherapy that essentially all of the local failure that he saw were in the class of patients that had this large cell variant.  So I think this supports the view that there might be underlying pre-existing genetic changes in the tumor cells that alter their radiosensitivity and lead to a clinically important radioresistant phenotype.

TOP
Slide 17:

I have been interested in studying this for a number of years, also since about the early 1980s, when we noted that you could pretty reliably, in tissue culture, make cells radioresistant by transfecting them with the ras oncogene.  The ras oncogene would lead to a radioresistant phenotype.

TOP
Slide 18:

This is of course of interest in that ras is expressed at pretty high levels in many human cancers, including lung cancer.  It's not the most frequent alteration that we see in lung cancer, but it is present in a high percentage of the patients.

TOP
Slide 19:

Ras, as has already been alluded to, is an interesting therapeutic target, because we now have potential ways of attacking it, even in the presence of a ras mutation, in that we know that even mutant ras, in order to be active, has to be inserted into the inner surface of the cell membrane.  And this insertion requires a post-translational modification of the protein.  It requires farnesylation of the C terminal of the protein.

TOP
Slide 20:

So we collaborated with Said Sebti and Andy Hamilton, looking at some of their non-peptide mimetics of the ras CAAX farnesyl transferase recognition site to see whether we could use these drugs to alter the radiosensitivity of cells that were expressed in mutant ras. 

TOP

Slide 21:

And indeed we could.  When we took these rat embryo cells that we had made radioresistant by transfecting them with ras, and now treated with one of their farnesyltransferase inhibitors, we could show that we could completely overcome the radioresistance that ras had induced.

TOP
Slide 22:

In fact, we could do the same thing in human tumor cells that were expressing naturally occurring ras mutations.  Here in the T24 bladder carcinoma, which expresses an H-ras mutation, when we treated this cell with the farnesyltransferase inhibitor we could also show that these cells would become more sensitive to radiation therapy in tissue culture.

TOP
Slide 23:

This led us to a series of animal experiments to see if this property would hold up in tumors grown as xenographs in nude mice.  This particular figure is a drug from Merck, L744832.  But you see the same effects with Sebti=s and Hamilton's drug, where we could show that treatment of the tumors with the drug alone or with radiation alone would result in a growth delay.  However, you got a highly statistically significant synergistic effect when you looked at the combination of the drug plus radiation.  

TOP
Slide 24:

I have shown you data so far for H-ras, which is of course the gene that is most commonly used in tissue culture experiments.  But as has previously been pointed out, the most common mutations of human tumors are in K- and N-ras. 

TOP
Slide 25:

So as proof of principle that K-ras and N-ras could also contribute to radioresistance, we collaborated with Eric Standbridge and made a series of knockout experiments, taking tumor cells that were expressing naturally occurring K-ras mutations, and knocking out the mutant K-ras allele and only the mutant K-ras allele.  The wild type K-ras allele remains in these cells.  And when you knock out the mutant allele, the cells become more sensitive.

TOP
Slide 26:

In this slide, in HT1080, which expresses an N-ras allele, if you knock out the N-ras allele, the cells become more sensitive.  If you put N-ras back in by transfection, they become more resistant again.  So we believe that this property of H-ras was generalizable to other members of the ras family.

TOP
Slide 27:

So we initiated a Phase I clinical trial, a multicenter trial which is still underway, not completed.  So I'm only going to show you some anecdotal data from patients that we actually treated at Penn.  This is not a definitive report of this trial.

TOP
Slide 28:

What we were attempting to do was determine the maximally tolerated dose of L744123, farnesyltransferase inhibitor manufactured by Merck in combination with radiotherapy.

TOP
Slide 29:

We had two cohorts of patients that we treated.  The cohort that is of interest today is the cohort with non-small cell lung cancer, but we also included with that patients with head and neck cancers, since we thought they would have similar toxicities.  We also had a second cohort of pancreatic cancer, but the drug was escalated independently in these two cohorts.

TOP
Slide 30:

We had two dose levels, starting at 280 milligrams per meter squared, and escalating to 560 milligrams per meter squared.  We were attempting to sensitize as many weeks of radiotherapy as possible.  Because if you are going to give a drug that sensitizes to radiation, then it stands to reason that the drug has to be present for the most part when radiation is actually delivered.  This type of radiosensitizer would have to be present throughout the entire course of treatment, or at least throughout much of it.  In this trial in dose level two, we were achieving drug levels in five out of the seven weeks of treatment. 

TOP
Slide 31:

We gave what we thought was standard radiotherapy in these patients, 65 Gray in the lung patients, 70 Gray in the head and neck patients, and 60 Gray in the pancreas patients.  All the patients were simulated and CT treatment planned.

TOP
Slide 32:

Most of the patients that we recruited for this trial were patients with pretty far advanced disease that had failed other therapies.  This is a woman in her mid-forties who had been treated with platinum and VP-16, and had in fact progressed through this treatment, and was in pretty much extremis when she came to us. 

TOP
Slide 33:

Within one month of treatment she had achieved not quite a complete response, but at least an exceedingly good partial response.  She subsequently failed with distant metastasis, first in the adrenal gland, and then elsewhere, but remained locally controlled out to 20 months after treatment.

TOP
Slide 34:

This is another patient that we treated, again, with a very, very mediastinal mass, and already some blockage of the airways.  Again, not quite a complete response, but a pretty dramatic response.

TOP
Slide 35:

So if you look at the group of patients that we treated at Penn in the lung and head and neck patients, we saw no toxicity after treatment in any of these patients.  All of the patients achieved at least a partial response.  The head and neck patient had a complete response.  This patient who had a good partial response at four months actually converted to a complete response at six months.  So far the follow-up is limited.  None of the patients have progressed locally. 

TOP
Slide 36:

In the pancreas patients, as you might expect, the outcome was a little different.  We did see some grade three toxicity in these patients, although it is not clear how related it was to therapy itself.

This patient had a major GI bleed from a stomach ulcer outside the field of radiation, although we counted this as a significant toxicity.  We saw some patients who developed stable disease, although many of them progressed with metastasis either during or shortly after treatment.  And we have one patient who has a CR, but the follow-up is only two months in this patient.

TOP
Slide 37:

Although we entered patients in this trial regardless of ras mutational status, and we have seen responses in patients who did not did not have ras mutations, in the future we want to look at the mechanism of action of FTI’s in the absence of ras mutations, perhaps in other either upstream effector of ras like RB or EGFR, or looking at potential cell cycle effects of these drugs in the tumor cells, because of some interesting laboratory studies that we did.

TOP
Slide 38:

I have given you one example as a proof of principle of targeting a signal transduction member as a mechanism of altering radiosensitivity.  Another interesting example under intensive study in a number of labs including Dr. Schmidt-Ullrich's, who will be leading this session this afternoon, is the EGFR family of receptors, because of the availability of inhibitors of this family of receptor.

TOP
Slide 39:

These are data from [unintelligible] lab in Wisconsin, looking at established squamous cell tumors treated with C225, the monoclonal antibody against EGFR plus radiotherapy, where you can again see a growth delay with either radiation or the drug.  But you see animals being cured by the combination treatment.

TOP

Slide 40:

This is the control animal -- x-rays alone, drug alone -- and this is an animal treated with the combination of C225 plus radiation.

TOP
Slide 41:

This has been looked at in a small series of head and neck patients by Jim Bonner at the University of Alabama, where he saw very favorable interactions of radiation with EGFR blockage.  This is now going on to a Phase III trial led by Jim Bonner and Paul Harrari. 

TOP
Slide 42:

So in addition to potentially attacking ras, there is now rationale and preliminary data that would support the use of EGFR blockade in combination with radiation therapy in EGFR-rich epithelial tumors, which would certainly include many lung cancers.

TOP
Slide 43:

So from these two examples, I think one of the things that is happening in the field currently is that we are beginning to think about radiosensitivity -- alterations in radiosensitivity as being secondary to alterations in signal transduction pathways.  We have begun to look at the earliest members of those pathways, EGFR and ras, but there are now many drugs that allow us to look also at downstream members of signal transduction pathways as potential modifiers of the radiation response.

[unintelligible], here in Washington, has some very interesting data looking at anti-sense to raf as a potential modifier of radiation sensitivity.  I think one of the developments that you will see in the field in the next few years is an increasing focus in looking at signal transduction modifiers as potential modifiers of radiation sensitivity.

TOP
Slide 44:

However, the other problem that we potentially have in alterations in radiation sensitivity or in radiation resistance we have to think about in terms of physiological targeting.  We know that cells in hypoxic environments are resistant to killing by radiation regardless of their underlying genetic status.

TOP
Slide 45:

One interesting aside is that when we looked at the bladder tumors that we treated in our FTI experiments.  Here we are staining these tumors with EF5, which is a 2-nitrylimidazole drug that stains viable cells in hypoxic environments.  As you can see looking at the red stain here, the green is a counter-stain for vessels, these tumors were highly hypoxic in the control animals.  However, after as little as one to three days of FTI treatment, this hypoxia essentially completely disappeared.  So it's possible that signal transduction modifiers may also alter the physiological status of tumors.

TOP
Slide 46:

If you looked at cells that did not express ras mutations, the drugs had no effect whatsoever.

TOP
Slide 47:

However, in addition to potentially modifying the environment of tumor cells, there are other attempts under way in the field to actually exploit this hypoxic environment by looking at drugs that are selectively cytotoxic to hypoxic cells.  One of the drugs that has been most closely studied is the drug developed by Martin Brown and his colleagues at Stanford, tirapazamine, which in hypoxic environments is highly toxic to cells secondary to DNA damage.

TOP
Slide 48:

This has also been studied in combination with radiation.  These are data from Lester Peters in Melbourne, Australia, looking again at a Phase I trial, but in patients with very far advanced head and neck malignancies, where you would typically expect only a very small percentage of the patients to be locally controlled, never mind survive radiation treatment.  He was able, in this small group of patients, to achieve a local control in almost 90% of the patients, and survival in 75% of them with 2.5 years of follow-up.  This is now also being expanded to a Phase III trial, and is also potentially applicable to lung cancer patients.

TOP
Slide 49:

We have some data from a multicenter trial of tirapazamine in combination with radiation, where again the combination of tirapazamine plus cisplatin gave an improvement in median survival, and in overall response that was highly statistically significant, although as I'm sure many of you know, there are other trials that have failed to see this effect.

TOP
Slide 50:

In conclusion, I think there are multiple potential ways that we can think of in modifying the radiation response, both by modifications in signal transduction pathways, and in potentially modifying the physiological environment of tumors.  I would like to thank the contributors to our Phase I trial,

TOP
Slide 51:

and also the people that have worked with me in my lab in doing the preclinical studies:  Ruth Muschel and Eric Bernhard have worked with me for a number of years; Said Sebti and Andy Hamilton first collaborated with us on the FTI trials, and we have subsequently collaborated with the group at Merck.

TOP
Slide 52:

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

DR. SEBTI:  In terms of the anti-tumor activity of FTI, it appears that ras may not be the only target, and other farnesylated proteins may be.  Is this also the case for the radiosensitizer activity of FTI, especially in terms of if the tumors that you have showed responded or were radiosensitized, and they have K-ras mutations, K-ras becomes geranyl geranylated as you inhibit it.  So is the ras the target, in terms of this ability of FTI's?

DR. MC KENNA: This is something that we have debated for a long time.  In the preclinical laboratory studies, with very few exceptions, the radiosensitization tracks extremely closely with inhibition of ras activity.  We have one in a number of cell lines we have studied, we have only ever seen radiosensitization in one cell line that did not have a ras mutation, and there was a lot of cytotoxicity from the drug in that cell line directly.  So it was very hard to know if you were truly seeing a synergistic interaction.

We have failed to see radiosensitization in some cells lines that were expressing ras mutations.  But typically in those cell lines it was because we could not inhibit ras farnesylation effect with the drug.  So in the lab the correlation is very tight.  In the clinical trial, the correlation is not tight at all.  Most of those patients in the Phase I trial were not ras-positive.

That's why we are extremely interested now in looking to see whether this is some target totally independent of ras, or whether by inhibiting ras you are altering a signal transduction pathway that is active in those patients, but due to some other member of the pathway, but I don't know the answer.

DR. BUNN:  Gillies, why don't you just study some small cell lines?  They never have ras mutations.  They are sensitive to the FTI inhibitors, and you can see if you increase their radiation sensitivity?  Because if they do, then you know it isn't ras.  That's an important question, but it's answerable by just studying the small cells.

DR. MC KENNA:  As I said, in the lab we don't see radiosensitization, and we have looked at some.  Clearly the drug is having effects in the clinic that are quite distinct from anything we predicted from the lab studies.  We did not predict that treatment with FTI’s would drastically alter the vascularity of the tumors or their oxygen supply.  That is not predictable from a petri dish, and yet it was clearly true in the animal studies.

DR. GANDARA:   If I understand you right,  the hypoxia reversal by FTI’s was only in tumors that were ras mutant?

DR. MC KENNA:  We only saw that -- but again, it's in a limited number of cell lines that we studied in animal.

DR. GANDARA: What would be the presumed mechanism for hypoxia reversal?

DR. MC KENNA:  I don't know.  There are a number of potential ways that it could occur.  We know that in cells that express ras mutations, you see increased oxygen consumption, because ras stimulates metabolism, and you see increased oxygen consumption by the tumor cells.  John [unintelligible] and I showed that a number of years ago.  Potentially you could improve oxygenation by decreasing oxygen consumption, thereby allowing oxygen to diffuse more adequately throughout the tumor.  Other possibilities, however, ras is known to affect interstitial pressure in tumors.  Ras is known to affect the VEGF expression in tumors.  There are a variety of mechanisms that could be resulting in altered oxygen supply.  We see the effect in as little as 24 hours after treatment.  So it seems to me that that is too short of a time for it to be secondary to vessel proliferation, but it might be about the time you would expect if it was affecting a metabolic property.  But that's just speculation.  I really don't know the answer.

DR. CHOY:  I understand you are measuring the serum level of FTI in your phase I study.  Were you able to achieve a sufficient amount of FTI in the serum that resulted in radiosensitization in in vivo and in vitro studies?

DR. MC KENNA:  Yes, we achieved levels of about 5-7 micromolar in the patients and in the animals, which is sufficient, from our lab studies to see the radiosensitizing effect.

DR. SEBTI:  Just a comment about a possible mechanism by which the FTI’s may work.  Even though a tumor, say a lung tumor, does not have a ras mutation -- and we are talking about K-ras actually, that's the problem.  H-ras, as you have shown, is sensitive to FTI’s.  But it possible that a tumor also requires H-ras for its growth or malignancy or what not.  And H-ras itself may not be mutated.  But H-ras in that tumor is actually driving the survival of that tumor.  We have recently shown that in the H-ras, which actually drives the PI3 kinase pathway, that is really hit by the FTI’s predominantly, rather than the K-ras that actually goes through raf-myc-erk pathway, and that is more of the proliferation.  So there is a balance there between H-ras driving survival through PI3 kinase and K-ras driving the proliferation aspect through the raf-myc-erk.  And we know that in all these cells H-ras is hit, whether it is mutated or not.  So that could be a possibility.

DR. MC KENNA:  Yes, that's an interesting hypothesis.

DR. SAXMAN:  The final speaker this morning is Dr. Norman Coleman.  Dr. Coleman is the director of the radiation oncology sciences program at the National Cancer Institute.  He is going to speak with us this morning about integrating radiotherapy and new therapeutic agents for the treatment of locally advanced non-small cell lung cancer, what do we know, and importantly I think, what do we need to know.

TOP