|
 I
think that ends my presentation.
Discussion
DR. SAXMAN:
Questions or discussion?
DR. BUNN: I
think we should get the skeptics here - which is me - to roll up
our sleeves and discuss this. Proof of principle for p53, if you
take Velban and inject it into a tumor 100 percent of the time,
the tumor will regress. It almost doesn't matter what the tumor
type is.
If you take
p53 and inject it into a tumor, lung cancer, 13 percent of the time
the tumor regresses. That is proof of principle, and it tells us
a lot. Now, I don't actually mind systemic delivery to find out
how to deliver genes, but if you get a 13 percent response when
you inject it directly into the tumor, other than finding out if
you can deliver it, I am not sure what you are going to learn.
Now, FHIT,
perhaps works better in animals. I think it is actually reasonable
to inject FHIT into the tumor in 20 patients and see how often the
tumor shrinks. But it is a little hard for me to believe that systemic
treatment is going to have a profound effect when direct injection
causes a response in 13 percent of the people. A totally different
paradigm from what you discussed, David, but I think it is reasonable
to take FHIT and some of these other things, inject them into peoples'
tumors and see if they ever shrink.
You can also
see what the gene expression is and determine if there is any correlation
between gene expression and whether it shrinks or not.
I don't know
what other people think, but to me if we are going to talk about
genes, inject them into tumors, see if the gene gets expressed and
see if the tumor shrinks. When somebody else solves delivery, then
we will know which genes we ought to be delivering systemically.
DR. MINNA:
First of all, I don't think the experiment with Velban has been
done.
DR. BUNN: I
do it all the time in my patients that have got subcutaneous nodules
that are bothering them. We know a dose, and we know it works. It
works every time.
DR. MINNA:
Okay, I think that as far as I know, there has been no treatment
of any small cell lung cancer nodule or patient with a direct injection
of p53, and so I think the response rates in small cell we don't
necessarily know.
DR. BUNN: We
should find out.
DR. GANDARA:
So Paul, I guess I don't understand your issue. Is it the cytostatic
issue, in other words that if an agent doesn't cause tumor shrinkage,
it is not going to be clinically meaningful?
DR. BUNN: First
of all, you can inject these genes in tumor nodules, find out if
the gene is expressed after you inject it, and you can find out
if the tumor nodule shrinks after you inject it. If you inject it,
you get expression, and the tumor doesn't shrink, then I would say
that that is probably not going to be a very good drug to give systemically.
DR. GANDARA:
So tumor growth delay would not be a reasonable objective for some
of these new agents?
DR. BUNN: I
don't think so for p53, FHIT, and bcl-2, probably not.
DR. MINNA:
I think for FHIT and p53 you would be relying on induction of programmed
cell death and bystander effects, too. I mean you are not talking
about a static effect, and obviously the other issue is the combination
of these agents with chemotherapy. The amazing thing is that small
cell lung cancers present de novo having a p53 mutation, and 70
or 80 percent of them will have a beneficial response to chemotherapy.
That must exist via p53 independent pathway. The question is, "If
you could have p53 plus chemotherapy would you get several extra
logs of cell kill as well?"
DR. ELIAS:
I am just supporting Paul in his devil's advocate role here. One
thing that we don't know very well is, in pre-clinical models, "What
is the mechanism of escape of cells that are in fact transfected
with active wild-type p53 or any other gene target?"
DR. MINNA:
We actually know that, because when you do that experiment you get
the cells growing out. They have re-mutated to p53, and that may
happen even systemically. I think Paul's most valid point is the
question of systemic delivery. If there are methods for systemic
delivery, I think that would make these gene therapies much more
attractive. Unless of course, we could deliver them as aerosols
as early treatment for very early lesions.
DR. BUNN: Part
of my point is that there are lots of genes. So how do we pick the
best gene therapy?
You know it
is sort of interesting. If p53 only causes the tumor to regress
13 percent of the time maybe we could find some genes that would
do it 50 percent of the time. Then if you had a systemic delivery
you would know which gene that you want to deliver systemically.
Look at all the genes that we have to try for gene therapy. It would
be fascinating for me to know whether if you give FHIT directly
into a tumor, it regresses more or less often than 13 percent of
the time, and if it regressed 80 percent of the time, I would be
pretty excited about that.
DR. MABRY:
I think the point is actually delivery because, compared to molecules
that usually get into cells, p53 is a big old molecule. Cells, I
think through evolution, have tried to develop mechanisms to get
rid of large molecules, and organisms that are multicellular do
the same thing.
So I think
that the real question about gene therapy or the treatment of cancers,
small cell lung cancer and others, with large molecules has to do
with delivery, and I think that is the rate-limiting step. It has
to be looked at primarily. I don't hear people talking about that.
I think that is the first step. We have plenty of genes. In transfection
experiments, you are washing the cells or you are flooding the cells
with plasma, and so that doesn't necessarily compute to the delivery
in humans.
DR. MINNA:
I think now that Paul has thrown down the gauntlet, we probably
need to do this. My guess is that we could take 20 small cell lung
cancer lines, each with a p53 mutation of different types and with
various types of drug resistance, and if we try to force the expression
of wild-type p53 in those cells it works probably in virtually every
case.
Now, there
is a different delivery problem. With a 13 percent response rate
with an injection of p53 into a non-small cell lung cancer, and
from the stuff that we have done with FHIT, the vast majority of
cell lines, I guess it is 8 or 10 different ones we have done so
far, nearly all of them have been sensitive to apoptosis induction
by FHIT. So these are as good as any chemotherapy agents that we
have.
DR. BUNN: My
point, exactly. So FHIT might be better. In vitro you say, maybe
it is better. Let us inject FHIT into some tumor nodules in small
cell patients, give chemo before or after like David was saying,
but find out whether their tumor nodules shrink. If they shrink
all the time, I will get really excited.
DR. MABRY:
The thing is that FHIT or p53 in terms of gene therapy are good
in vitro model systems and are proof of principle. My question would
be, "Would the next step be looking for FHIT or p53 or other members
of that class, i.e., that mimic small molecules?"
DR. MINNA:
I agree.
DR. D. JOHNSON:
The sessions obviously have a lot of commonality in terms of obstacles
and questions about how to test these new compounds. As sort of
a pragmatic issue, I am interested in the comments about functional
imaging because that is something that theoretically we could do
right away. I am not an expert on PET scanning, for example, and
I know there are other imaging technologies that perhaps the group
here at the NCI who are the diagnostic branch people could give
us some help here. Do we have any mouse models of functional imaging?
Has anyone looked to see how much change in the metabolic activity
of a tumor needs to take place before current PET technology can,
in fact, detect that? How large does the tumor have to be in order
to detect it? Is it something even worth doing?
We can do all
that in humans at a huge cost, but it seems to me we can do it in
mice if someone will just throw a few mice under a PET scan. Do
we have any of that data and, if so, where would it be? I don't
know, and I am sure someone else does.
DR. TEICHER:
In terms of PET scanning, it is very possible. In terms of limited
resolution, limited resolution is about the size of the mouse. So
mice and PET scanning are not too compatible. However, if you get
a good size tumor in a rat you can PET scan it, but again, cost
has been a barrier. I have tried many times, and I haven't yet successfully
done PET scanning studies.
DR. MINNA:
If you get a tumor in a dog you can do a PET scan.
DR. D. JOHNSON:
We have done it in rats, and we have tried it in mice and it doesn't
work very well, at least with our scanner. But in rats you can do
it. At least, with one compound I know we have done that. We are
still sort of struggling with how to image this, but this was a
question that we discussed briefly yesterday in our session. I just
wondered if there is other technology that gives you the same type
of information that a PET might give that could be applied to a
mouse.
DR. GROCHOW:
There appears to be no one here from diagnostic imaging, but in
fact, there are many funding mechanisms to try to address that.
As you know, there are several small animal imaging centers that
have been funded around the country for PET scanning. That will
help some of the new initiatives, will help to produce PET materials
for such studies for rats with big tumors.
MRI of rats
is certainly feasible, and I am aware that there are studies currently
going on with angiogenic agents and MRI to look at blood flow in
murine models.
DR. GANDARA:
I think there is a clear clinical perspective on the use of some
sort of functional imaging that is very pertinent. For instance,
in limited small cell lung cancer, where CT scan is the typical
way that we follow patients but is an inaccurate way of gauging
tumor response after chemoradiation (as shown in Dr. Turrisi's long-term
follow-up of his recent trial with chemoradiation). The Southwest
Oncology Group also demonstrated this in our trial, where we have
followed chemoradiation with surgery in locally advanced non-small
cell lung cancer. We had 26 patients in whom by CT scan there was
no evidence of antitumor activity, meaning tumor shrinkage, and
yet when those patients went to surgery and the tumor was taken
out, half of those patients had a complete or near complete pathologic
response. So I think again as would apply to these new agents some
sort of better means of assessing tumor response by tumor metabolism
or functional imaging would be of great assistance because otherwise
we may assume that some agents are much less active than they really
are. We are not going to have the luxury of taking out tumors very
often afterwards to have that sort of pathologic correlation.
DR. D. JOHNSON:
I wonder if Bev might address an issue. In the models that you have
done, are there any techniques that you have used in the mice that
allow you to gauge blood flow? How do you do that, and if you do
it is it a useful surrogate end point? Is it worth doing?
DR. TEICHER:
Of course, we have measured oxygen in tumors with microelectrodes
quite a lot, and that can be a very sensitive measure.
In terms of
blood flow, I think there has been quite a lot of nice work done
in dog tumors, and that is a useful measure, but I cannot really
definitively speak to that.
DR. D. JOHNSON:
PET can be done with oxygen, too.
DR. TEICHER:
Yes, oxygen-15, yes.
DR. D. JOHNSON:
That would be maybe a better way of assessing the blood flow approach.
DR. TEICHER:
I think that the fluorodeoxyglucose method though is going to be
really tough to tell which tumors are alive and which are dead by
that method of PET. I think that is going to be a very, very difficult
way to go.
DR. JOHNSON:
So oxygen 15 might be a better end point for assessing the antiangiogenic
agents.
DR. TEICHER:
Yes, I think so.
DR. VALDIVIESO:
Valdivieso, Southwestern in Dallas. I think that from a clinical
perspective we need to be reminded that we are still at the very
earliest stages of this technology. So to try to expect too much
in terms of complete response and cure, I think that we need to
be very fair, and I tend to agree.
Let us identify
the one or two markers we must move forward in a Phase I fashion,
identify the schedules, the various pharmacology, the pharmacokinetics,
the basic strategies and don't expect too much.
The second
comment I would like to make is this: clinicians in the real world
are hard pressed to deliver appropriate, cost-effective care. That
includes staging. Even though we very nicely discussed the best
way that we can use our PET scans and MRIs in our academic institutions,
the fact is that most patients will not come to us but will stay
outside.
So there was
a study that I had done when I was in Houston that I think should
be a refreshing experience. Five hundred patients all evaluated
by nine board-certified medical oncologists, all of them doing lung
cancer, nothing else, were asked to stage the patient based on a
good clinical, physical exam, blood chemistries and chest x-ray.
The outcome? Thirty-five percent of the time we were wrong in staging
people just between limited and extensive disease, and it was only
when we had the CAT scans added - this is I remind you after the
normal blood chemistry finding abnormalities in the liver and the
bones or in the brain.
So what I am
saying is that it would be very nice as part of this effort perhaps
to identify the one or two molecular markers that we must do in
the tumors early to predict invasiveness, so that when we do the
staging of patients, even in the community, we can provide tissues
to the investigators to try to eventually come out with the algorithm
that may help us to better recognize the patient who must have the
CAT scan and the patient who won't have anything at all but surgery.
So I would
like to invite you to think beyond really the therapeutic molecular
trial and rather help us out to intervene early in patients who
will not come to academic institutions but we can influence perhaps
at the appropriate staging. Otherwise, you may be dealing with so-called
"limited" disease patients, a stage 1 who may not be a stage 1 because
they were not properly staged.
DR. MINNA:
One of the things I didn't mention was Ed Gabrielson's presentation,
and this was one of the questions for high throughput technology
in terms of microarray analysis. I mean there is obviously a variety
of markers that are possible. But one of the things I see coming
out of microarray analysis is that it needs to be tested as to whether
or not a tumor initially has an expression signature that tells
you whether it is going to respond to therapy or not and whether
it is going to be highly metastatic or not.
It may be that
all primary small cell lung tumors still have the expression signature
of primary small cell lung tumors, and the metastatic cells have
acquired different genetic abnormalities. So they would have two
different expression profiles, but we will come to learn that. But
by the same token, we have been talking a lot with our imaging people,
and I think one of the things that is going to come out of the expression
arrays - and we are going to have to see this under different hypoxic
conditions and things like that - are some new insights to make
new probes to use in functional imaging. For instance: Is there
a signature for a small cell lung cancer that is going to undergo
programmed cell death drom chemotherapy? Could that be generated
into a probe that then could be given and scanned after a cycle
of therapy to see how much tumor cell kill there is? The Director's
challenge is obviously, well, Rick Klausner was thinking first in
terms of the pathology diagnosis, "Is there a signature that one
could use in functional imaging to tell you the histologic type
and response to therapy?" So the major hypothesis is: Is the expression
profile of a tumor cell looking at 5, 10, 25 thousand genes, 50
thousand, whatever it is, going to tell you how it is going to behave?
I think it would be very important to know the answer to those questions.
The other question
is, and maybe Jim Jett could answer, "Are there any advances in
MRI that would obviously give you much better resolution? Are there
ways to couple the functional imaging with MRI or does it all have
to be through a PET-based mechanism?"
DR. JETT: I
don't have any information on that.
DR. D. JOHNSON:
You can certainly fuse CT scans and PET. I would assume one could
do the same with MRIs, and a few institutions now have these so-called
"fusion scan" capabilities. Actually the nice thing about it is
that the PET portion of it can be done, some would argue not as
well as a regular PET scan, but it could be done with the gamma
camera as well. So the cost is coming down on that, and they actually
make for very nice scans. I don't know how much data we actually
get, but you get the PET activity within the CT abnormality. It
is the very issue that David alluded to, whether the mass is active
or not, and it is on the same image that you are able to see this.
DR. MINNA:
By the way, if Paul and I and Tony Elias were being too crusty with
one another here in terms of this, I was trying to focus on the
issue of our session which was molecular abnormalities in tumors.
What are they, what are the common ones and then, when is therapy
going to be specifically directed towards those and relate towards
those? That was the nature of the issue, and I think Mack is right
that, maybe with combinatorial chemistry or new design methods for
drugs, it would obviously be better if you could have small molecules
that would replace these functions. There is no question about that.
DR. SMITH:
One of the problems though, and we sort of touched on this about
microarrays, is that the small cell system really doesn't give you
material to work with for microarrays. We were in a breakout session
headed by John, and clearly his orientation is going to be that
the cell lines will be a viable model for the primary tumors. I
agree that you could probably do your initial interrogations, but
sooner or later you are going to have to get down to primary tumors
with a subset of genes - maybe 100 to 500 - but we don't have any
samples to even do this analysis on.
DR. MINNA:
I think that became very clear. Just as a technical aspect, the
way I see this technology being exported out into clinical practice
is through immunohistochemistry. For instance, the Stanford group
is making a major effort in breast cancer. So all of the genes that
go up or down that they are discovering, they are really developing
antibodies against. That would be, obviously if you had to do 100
antibodies, that is too many, but let us say that there are 5 or
10. I think what we need to validate these things is the tissue.
DR. JETT: I
don't think that that is a problem if you require, as Joan and Everett
said in their presentation, an adequate tissue sample to get on
a study. You can easily do a mediastinoscopy which is not all that
an invasive procedure as part of the initial requirement to get
on a study, if you are using some of these new drugs.
DR. MINNA:
I think that would be fabulous, but I can see that not getting reimbursed.
I think that if we had patient care dollars for grant supplements
that would be extremely valuable.
DR. GANDARA:
Don't they have a new intergroup trial in limited stage small cell
lung cancer, and an ancillary tissue acquisition study? Ancillary,
I think, rather than part of the trial, would be a good way to establish
a tumor bank for small cell lung cancer and would require CTEP to
provide some additional funding and incentive for us to do that.
But small cell is not the easiest of all the tumors that we are
trying to treat to get particularly fresh tissue. Dave, I think
with your support and ECOG's we probably could try to build that
in as an ancillary, but doing that in non-small cell lung cancer.
DR. D. JOHNSON:
I am not sure I understood Jim's point. I mean,"Are you suggesting
you can do that in small cell, mediastinoscope the patient?"
DR. JETT: Yes,
small cell.
DR. BUNN: We
did that at the NCI-Navy, because we were looking at new drugs,
but it was free. When you get in the real world, a mediastinoscopy
is a very expensive procedure.
DR. JOHNSON:
Not only that, you can do it in non-small cell lung cancer, but
I would be stunned, even with financial support, if one could do
this in small cell lung cancer patients. I would certainly do everything
in my power, personally. We are very committed to it at our institution.
DR. MINNA:
Part of the problem is in terms of the ethics. It may be important
to have the design of the trial and the treatment based on the results
from getting these tissues. That may be more reasonable.
DR. JETT: That
is not different from what the HER-2/neu stuff that is going on.
You require tissue to show that you express that.
DR. D. JOHNSON:
Yes, but you could use the original material removed at the time
of the operation to do that, and you are not going back for a re-biopsy.
DR. JETT: Yes,
a mediastinoscopy, well, is a $3,000 procedure, roughly.
DR. D. JOHNSON:
And it has a mortality associated with it, albeit small.
DR. JETT: Zero
mortality. It has a little morbidity associated with it.
DR. D. JOHNSON:
At Mayo Clinic, it has a mortality of zero.
DR. B. JOHNSON:
One thing I want to point out in attempting to do mediastinoscopies;
in our single institutional experience we put 54 patients on study.
Of those, one-half of them we didn't think would tolerate it because
of either obstructive atelectasis, superior vena cava syndrome or
a cardiac condition where you didn't think it was safe. So that
got rid of half of them.
Approximately
another 10 or 20 percent refused the procedure. We were left with
being able, in a single institution, to do 18 out of the 54 which
we took to mediastinoscopy. Of those 18, two of them we were unable
to obtain tumor tissue. Of the rest of those 16, three of them we
got a very poor specimen, and then we were able to grow eight out
of the 18 people we took to mediastinoscopy. So you had enough that
you had viable tissue. Your denominator is going to fall off pretty
fast if you are hoping to be able to do something meaningful with
this group, and that is in a single institution.
DR. GANDARA:
I think the issue I was thinking about was really in those situations
where the tumor tissue was more easily accessible. For instance,
a lymph node or something like that because there won't be as many
opportunities to also have the patient demographics very carefully
documented as part of the trial to be able to compare with the tissue
results.
DR. DENNIS:
Regarding the array issue, wouldn't it be more important to get
array analysis on what remains after treatment, because we know
that most of the tumor cells are going to undergo apoptosis in a
p53 independent manner and most of the tissue responds as we wish?
But there is a small amount that remains, for which I think the
arrays would be more important because perhaps the cells that actually
survive have wild-type p53. Maybe we want to put mutant p53 into
the cells that actually survive. I think that the issue of trying
to get circulating tumor cells and cells that remain after therapy
might be important.
DR. BUNN: Perhaps
an illustration: Last Sunday morning Mark Jeraci brought me a chest
x-ray from a patient in whom basically one lung was gone, and in
his other lung his bronchus was about 3 millimeters, and he said,
"This patient is kind of sick."
I said, "He
looks pretty sick from his chest x-ray." I said, "Is this a young
guy with lymphoma or an old guy who smokes?"
He said, "This
guy smokes like a chimney." I said, "You had better get some tissue."
So a fine needle aspirate was obtained from his supraclavicular
node which showed small cell lung cancer. Well, there was no way
in hell that this guy was going to have any other procedure, and
a day and one-half later after he got treated with chemo he left
the hospital without a palpable supraclavicular node, and you are
not going to get any more tissue on that guy. I mean he had his
FNA, and that is it. That is what you got and you are not getting
anything else on that guy in the beginning, and you aren't getting
anything else on that guy later. So that is the practical reality
of the situation.
DR. MINNA:
To speak on Dr. Bunn's favorite subject, one of the topics that
didn't get discussed here was the whole aspect of nicotine addiction
and the new evidence developing on the genetics and genetic epidemiology
of nicotine addiction. I think Adi's results spoke to some of the
major damage that goes on. I would say if you adjusted cigarette
for cigarette, even if you compared squamous with small cell lung
cancer patients, there is a difference there in the normal epithelium
in the amount of genetic damage. I think there are at least several
different markers, that there are polymorphisms that vary in the
population that predict an odds ratio for whether or not you are
likely to be addicted to nicotine and there is probably an inherited
component. In twin studies, there is at least 50 percent in both
men and women for both initiation and persistence of smoking. There
is a heritable component as well. This obviously has impact for
all types of lung cancer, but it might be interesting to know if
there is a difference between the genetic susceptibility to nicotine
addiction in small cell versus non-small cell lung cancer patients.
DR. SMITH:
As long as you raised that point, this is something else that we
didn't discuss at all, but the capabilities of using SNP-based chips
to try to determine which individuals that do smoke are going to
develop lung cancer is a very important area. Then you could more
aggressively just attack those individuals who have the wrong genotype.
DR. SIEGFRIED:
Since you mentioned nicotine, John, and also since you brought up
some of my work a little bit earlier, I actually didn't present
that data in the other breakout session because it is embargoed
right now, but I feel like I can talk about it a little bit since
you brought it up.
First of all,
we have been able to show that bronchial epithelial cells do have
what look like nicotinic acetylcholine receptors. So there actually
are probably biological responses to nicotine in the airway cells
themselves. One of the genes for which the transcription is altered
by nicotine is the GRP receptor. So it appears that the mechanism
that may be behind my observation that the longer an individual
smokes the more likely they are to express the GRP receptor in the
bronchial epithelium could be driven by nicotine. You know we also
have been able to show a sex difference in the frequency of expression
so that women are much more likely to express the gene in the airway
than men and can even express it with minimal or no smoking.
DR. MINNA:
This is the GRP receptor?
DR. SIEGFRIED:
Right, and we believe that at least in part may be caused by the
fact that the gene is located on the X chromosome and escapes X
inactivation, so that women have two copies that are able to be
transcribed, whereas men only have one.
So I will just
leave it at that.
DR. MINNA:
We discussed that finding, and I would just point out that we had
known for several years that lung cancer cells of all histologic
types express several different, high-affinity nicotine receptors,
which are nicotinic acetylcholine receptors and I can refer you
to several papers which characterize those by binding assays now
that the molecular genetics are known. The other fact that we found
really striking was that nicotine inhibits programmed cell death.
So if you put a programmed cell death signal in, nicotine will block
that. So I have often thought, that since nicotine can be converted
to a carcinogenic derivative as well like NNK or things like that
it may be both a carcinogen and a promoter, and I think this would
really fit. Then Dr. Resnick Shuler has data that it can stimulate
proliferation. So it may be that nicotine may be coming in, generating
the secretion of a variety of substances, altering the transcription
of GRP receptor which augments this, and then blocking programmed
cell death as well. I think that that would be really interesting.
DR. SIEGFRIED:
We have evidence that GRP, transcription and release of GRP is promoted
by nicotine, also. So the whole autocrine loop may be modulated
by nicotine exposure.
DR. MINNA:
Enrique, do you have any data? Obviously there have been lots of
studies in adrenal chromatin cells with nicotine.
DR. ROZENGURT:
No, we don't have any data on GRP receptor expression by nicotine.
DR. MINNA:
Dr. Bunn, do you want to discuss the old evidence about whether
smoking cessation helps with survival if you are treated for small
cell lung cancer?
DR. BUNN: It
does. Somebody who has a decent performance status and life expectancy
of more than a few months, certainly if you stop smoking in regard
to both small cell and non-small cell you are likely to A) live
longer and B) have a lower chance of developing second primary tumors.
DR. GANDARA:
John, you just mentioned that nicotine blocks apoptosis. So maybe
this is in the therapeutic realm as well.
TOP
|
The
next breakout session to be discussed is the molecular genetics
breakout group. The Chairpersons of this group are Dr. John Minna,
who is the Director of the Hayman Center for Therapeutic Oncology
Research at the University of Texas Southwestern Medical Center
in Dallas, and Dr. David Gandara, who is Director for Clinical Research
at the University of California Davis Cancer Center in Sacramento.
Dr. Gandara is also Chairperson for the Lung Cancer Committee in
the Southwest Oncology Group.
One
of the first questions was, "What are the molecular abnormalities
that are most relevant for small cell lung cancer and how good is
this evidence?" While this may not be a complete list, I think that
mutations in RB, p53 and the FHIT, which all are in the 75 to 100
percent range, are clearly key examples of this. 
Some
of the new observations that were discussed during our breakout
session - I will just go through them in bullet statement, and then
if there are questions, we can direct it to the people that are
here in the audience. Kay Huebner presented very exciting new data
on the knockout mouse, which in their heterozygous state can get
tumors after carcinogen exposure. These are so far predominantly
esophageal and gastric cancer. She had plans in collaboration for
using urethane for a lung cancer specific induction in these mice.
Dr.
Ball from Johns Hopkins presented some data on the hASH-1 knockout
mice, and this is why part of my comments are about targeting this
as a particular therapeutic as well as a diagnostic marker. In these
mice, they actually lose chromogranin expression in the lung. So
you knock out this important transcription factor, I believe, Dr.
Ball, it is a transcription factor. You lose these cells in the
lung; presumably these precursor cells are knocked out. Just as
interesting was the flip side of these experiments, where they forced
the expression of, I guess this was the human, it was hASH-1 as
opposed to mASH for the mouse version transgene in which they saw
proliferation of neuroendocrine cells. These were under promoters
that would force the expression in peripheral, in Clara cells. Is
that correct? 
DR.
MINNA: I was particularly very excited with that. Dr. Laird-Offringa,
as she was discussing earlier, talked about the neuronal RNA binding
proteins, HuD et al as a family of them that give these paraneoplastic
antigens. In addition to their use for diagnostic purposes or this
recent discussion about immunotherapy, these RNA binding proteins
interact with messenger RNA and can actually alter the half-life
of key molecules. So, for example, if you express HuD, if I interpret
your results correctly, the half-life of myc and fos messenger RNA
increases. So this would be another mechanism of getting over-expression
of myc. 
DR.
MINNA: Moving into the diagnostic realm, York Miller presented some
very interesting data on bombesin-like peptides in urine. Obviously,
in the signal transduction session, we are going to hear a lot about
autocrine and paracrine growth factor. York did a very careful study
of people that couldn't be taking drugs, couldn't be smoking; their
parents couldn't smoke. It turns out that in people's urine you
can either be a high producer of bombesin-like peptides, intermediate
or low. Actually he thinks he has fallen to three genotypes - LL,
HL and HH - with a gene frequency of about 86 percent for the low
and 13 percent for the high. From the calculations, this was about
80 percent heritable. So I think that it may be that people may
vary in terms of their production of bombesin-like peptides. Obviously
this would interact with cigarette smoking, and I am sure we will
probably hear later about Jill Siegfried's, observation of bombesin
receptors in the epithelium of smokers and the generation of bombesin-like
peptides, and I would put those two observations together.
Ethan
Dmitrovsky has developed a very exciting new model system using
immortalized bronchial epithelial cells, BEAS 2B, which I believe
Curt Harris derived, in which he treats them with carcinogen for
just 24 hours and then after that he will treat with agents like
retinoids. With just the carcinogen treatment alone, the cells would
become transformed and fully malignant such that they would form
tumors in athymic nude mice. The carcinogens he used were NNK and
cigarette smoke condensate. 
So
I thought that was very interesting. One of the questions that the
NCI staff posed here was, "What are the obstacles?" Now there are
several different obstacles. I am just going to tell you one of
them that I think needs to be resolved, and that is the use of small
cell lung cancer cell lines, both for in vitro tests and in vivo
tests as xenografts. I think it is very important for us to know
how reliable the pre-clinical models of these lines are.
Another
question that came up you see with small cell versus non-small cell,
common versus distinct, and I am just focusing on the genetic abnormalities.
Most of us would agree p53,3p, FHIT, myc, cyclin D1, bombesin-like
peptides in urine, as far as we know now, and bcl-2 are common to
both small cell and non-small cell. Distinct to small cell, would
be the differences in the RBP16 pathway and the HuD and hASH1s,
and there may be others. Obviously, anything that targeted these
common areas would work for non-small cell lung cancer, as well
as small cell. I think if you had something that was specifically
based on RB or HuD or hASH-1, that would be specific for small cell
lung cancer.
What
are the relevant agents for targeting abnormalities? P53 is a target.
Yes, we already have one agent, at least for proof of principle.
It is gene therapy, and I think the need there is going to be for
systemic forms of delivery. FHIT is a target. Yes, there are already
pre-clinical models for gene therapy that look pretty dramatic.
bcl-2, yes, and it is a target and yes, there is therapy in the
form of antisense that several companies are developing.
DR.
MINNA: And then for appropriate therapeutic approaches, I think
for right now p53 gene therapy and FHIT gene therapy we need to
explore systemic delivery. I think Esther Chang has worked out some
new methods that could be done for that. bcl-2 is antisense, and
we need to have a proof of principle in humans that it works. For
p53, there is a proof of principle, I think, in humans.
RB
null cells, and UCN01, I think are going to be very interesting.
Then this mouse metastatic model to test new agents, and, in the
case of new therapeutic approaches for hASH1, to develop inhibitors.
I would add on the HuD for vaccine. I would have to think about
how you would develop those for safety, because once you start doing
those vaccines, the patient is committed. 
DR.
GANDARA: The second part of our deliberations was in regard to how
to take these novel molecularly targeted agents into clinical trials
and what would be the differences in clinical trial design and methods
of evaluating the agents compared to classic chemotherapeutics.
Some of these really mimic the discussions of the first breakout
group, for antiangiogenic agents in particular. So some of the considerations
for taking molecular target agents into early clinical trials we
thought would be agent-specific, but there would be certain generalities.
One would be: Has there been a biologic effect demonstrated of the
agent on its molecular target in the pre-clinical trials, is this
something that should be incorporated in a clinical trial design,
and could it be measured? 
So
in terms of these agents or any sort of novel agent versus the classic
chemotherapeutic, early trials - which would ordinarily be designed
to determine a maximal tolerated dose - may instead be designed
to determine a threshold dose for biologic effect, again, with the
classic chemotherapeutic objective response being examined with
these new agents. Again, would there be a molecular or biochemical
end point or a clinical end point, such as time to progression,
that would be most applicable to determining efficacy of the agent?
Compared to a classic chemotherapeutic, where we would do a trial
in small cell lung cancer or breast cancer? In this case, would
we pick a study population, for example, with abnormalities in p53
as the target population?
We
had discussions about alternative study end points. I think our
discussions reflect those that were presented in the earlier session
as well - biologic effects on a molecular target and either some
sort of surrogate tissue or, for instance, circulating blood cells
or tumor tissue. The difficulty, in particular for small cell lung
cancer - is obtaining fresh tissue or serial tissue. 
One
of the questions was in regard to identifying patient subgroups
who were more likely to respond to a targeted therapy, and here
the hypothesis would be that response and/or survival would be influenced
by the underlying molecular profile in an individual patient's tumor
or perhaps in a specific tumor type, such as small cell lung cancer.
For example, for RB for instance, the data that Paul Gumerlock presented
about the RB null phenotype. If so, then this information could
be exploited to optimize the therapeutic approach. I think our consensus
was that although this is a hypothesis, it would need to be tested
for each of these molecular target agents and may or may not hold
true. An example of this is the FTIs and whether they really target
what we think or what we initially thought they did.
In
terms of assessing the clinical utility of a molecular targeting
agent, one issue is whether these agents will be used as single
agent therapy - an issue very similar to the anti-angiogenesis drugs.
The likelihood is that they will not have their final or ultimate
use as single agents, but rather that they will be used in combination
with whatever is the standard therapy for that stage and type of
disease, i.e., chemotherapy, radiotherapy or surgery.
So,
for establishing proof of principle in the early clinical trials,
there could be two designs, and this is exactly what Paul Bunn brought
up earlier. The first, in advanced stage, using the ECOG model of
a lead-in of a molecular target agent prior to active chemotherapy.
The second in earlier stage disease, this perhaps not being appropriate
to small cell lung cancer, but the window of opportunity where the
agent is administered after a biopsy and then there is definitive
surgery. This way you have pre-treatment and post-treatment tissue
to look for biologic effect.
So it may be that establishing proof of principle, particularly
for a cytostatic agent that has a molecular target, would mean
that Phase I trials would be done to establish safety, the biologic
end point. Then you would skip Phase II and perhaps go directly
toward Phase III testing, which would be required as a definitive
test because of the risk of false negatives in the Phase II setting.
This is actually the design that is being done, particularly with
some of the MMPIs, where there is really very little Phase II
data but going right to the Phase III trial. 
So
if these are to be combined with chemotherapy, one possibility would
be chemotherapy first and then a randomization to the molecular
target agent, particularly again if the design is such that it would
be given as long-term maintenance therapy versus a placebo or observation.
Another possibility would be testing as part of a direct combination,
for example, with chemotherapy.
For
earlier stage disease, again not applicable necessarily for small
cell surgical resection and then randomization to the molecular
target agent or observation/placebo, looking for a time to progression
or survival difference.