Reported by Nancy Nelson and Jennifer Michalowski
March 11, 2002
The past two decades of biomedical research have yielded an enormous
amount of information about the molecular events that take place
during the development of cancer. With this added knowledge, scientists
are creating new drugs such as Herceptin and Gleevec, that target
the molecular alterations involved in the biological pathways important
in cancer. The hope is that by targeting specific alterations in
cancer cells, these innovative therapies will be more effective
in killing tumor cells and less harmful to normal cells. As a result,
they should also have a major impact on survival and quality of
life of the cancer patient.
Edward A. Sausville, M.D., Ph.D., and Louise B. Grochow, M.D.,
are National Cancer Institute (NCI) scientists who play major roles
in helping to develop new cancer drugs. Dr. Sausville, the associate
director of the Developmental Therapeutics Program of the Division
of Cancer Treatment and Diagnosis at NCI, is primarily involved
with pre-clinical evaluation of drugs, while Dr. Grochow's responsibilities
as chief of the Investigational Drug Branch at NCI are to develop,
monitor and implement clinical trials. BenchMarks interviewed them
prior to their presentations at the Science Writers' Seminar on
March 11, 2002, in Bethesda, Md., titled "Molecular Targets in Cancer
Therapy."
One of the purported advantages of molecularly targeted drugs
is that they are expected to be less toxic to normal cells than
standard chemotherapies. In the trials with targeted therapies so
far, is this turning out to be true?
Dr. Sausville: Early returns with Gleevec and Iressa are
encouraging because there have been positive clinical responses
without the common toxicities seen with usual anti-cancer agents.
However, it would be wrong to think that these agents are without
any toxicity. Clinical experience will likely define "agent specific"
toxicities, such as the skin rash commonly seen with molecules like
Iressa. While most patients seem to experience improvement in the
toxicity either spontaneously or with alterations of dose, some
do not. The long-term consequences of these newer molecules will
need to be considered carefully as the information emerges.
Dr. Grochow: Although, overall, the new agents are less
toxic than traditional chemotherapy, trastuzumab (Herceptin) is
now known to cause damage to the heart muscle in a few patients.
Some EGFR (epidermal growth factor receptor) agents can produce
diarrhea and a severe acne-like rash, and bleeding into the lung
has been seen in some patients with lung cancer treated with anti-VEGF
(vascular endothelial growth factor). None of these agents is as
simple to take as acetominophen (Tylenol), and all are much harder
than strategies for not getting cancer in the first place: eating
five fruits and vegetables a day and not smoking.
Does a molecular-targeted approach to cancer therapy have
any implications for restructuring clinical trials?
Dr. Sausville: Yes and no---if the drug can be aligned to
a particular target that is responsible for actually sustaining
the disease, such as Gleevec, then conventional trial designs will
probably suffice, as was the case with Gleevec. But if the drug
is affecting one of a number of aberrant pathways in a tumor cell,
then potentially we will need to assess the actual target pathway
to help decide the correct dose and schedule of the agent before
devising new trials to combine the drug with agents directed at
the other aberrant pathways.
Dr. Grochow: Many of the novel targeted agents differ from
older agents in two critical ways that alter the conduct of both
early and later clinical trials. In traditional dose-finding trials
in cancer, doses were increased until unacceptable toxicities occurred;
the more drug you could give, the more cancer cells would die. Some
newer agents may have dose-response curves that do not steadily
increase (like interferon, where particular desirable effects may
diminish at higher doses). Some may not produce any additional benefit
at higher doses. Some are not toxic at any plausible dose. So we
can't use conventional phase I dose escalation plans to select the
dose and schedule for subsequent studies. We have to invent new
ways to determine whether the drug is affecting its target in the
desired way, and whether giving higher doses produces additional
useful effects.
In standard phase II trials to estimate activity, older cancer
drugs produced tumor regressions. These regressions were the basis
for deciding to proceed with large efficacy studies to support widespread
clinical use and drug approval. Many of the novel agents don't produce
regressions in tumor models; they slow the growth of the cancer.
That's easy to see in uniform mouse models but impossible to determine
in individual patients with very variable clinical situations and
rates of disease progression. So in phase II trials, new designs
that compare the time for tumor progression, or measures of clinical
benefit (reduction in pain, increased ability to perform regular
activities), or that evaluate novel endpoints (metabolic activity
of the tumor using imaging studies, for example) may have to be
incorporated in early trials. For relatively quickly growing cancers,
actual patient survival may be the endpoint in a phase II trial
of some novel agents, but that won't work for more slowly progressive
disease states.
And then there are new agents like imatinib mesylate (Gleevec)
-- when it turns off the growth signal from Bcr-Abl, no new cells
divide and older chronic myelogenous leukemic (CML) cells die off
in days, so blood counts promptly returned to normal and the patient's
well-being was evident within the first month of treatment. Some
patients in early trials with the small molecule EGFR inhibitors
like Iressa and OSI-774 have had tumors shrink, even though the
pre-clinical models suggested that there would only be slowing of
tumor growth.
Dr. Sausville, you pointed out that the reason for Gleevec's
success is that it targets a unique protein in leukemia cells that
the body doesn't normally produce. This is in contrast to many cancer
drugs that target proteins that are either under- or overexpressed
in cancer cells, but are found in normal cells, and so, in effect,
are not foreign to the body. Is there any intentional attempt to
identify other unique proteins in tumors, such as Bcr-Abl protein
in CML patients?
Dr. Sausville: Yes. NCI's Cancer Genome Anatomy Program
(CGAP), and its allied efforts to isolate full length cDNAs from
tumors, are among the initiatives that will tell us about genes
encoding protein molecules that may be present in tumors but not
in normal cells. A more difficult question, though, is deciding
about the functional importance of the expressed protein, and whether
it actually contributes to the development of the tumor or is a
"bystander" reflecting where the tumor originated. The latter types
of targets, while not as unique as the Bcr-Abl example, still might
be the basis for useful therapies, indeed as the work developing
targeted toxins by many investigators including the NCI's own Drs.
Ira Pastan, David Fitzgerald, and Susanna Rybak has demonstrated.
In the new budget proposed for fiscal year 2003, "Molecular
Targets of Prevention and Treatment" is one of NCI's scientific
priorities. What kinds of programs have been initiated in the last
year or two to move forward with this approach to therapy?
Dr. Sausville: We started an extramural Molecular Targets
Drug Discovery Program where investigators can use NCI resources
-- people, in-house expertise at our Frederick campus -- to develop
a particular molecule as a drug target. Forty research groups are
currently supported by this program. (http://dtp.nci.nih.gov/branches/gcob/gcob_web9.html)
Another NCI program, the Rapid Access to Intervention Development
(RAID) program which began in 1998, is designed to speed up the
pre-clinical testing for promising drugs. This program targets academic
laboratories that have novel candidate compounds, but lack specific
resources or expertise to develop them further. (http://dtp.nci.nih.gov/docs/raid/raid%5Findex.html)
Dr. Grochow: To enhance the ability to incorporate translational
research into early clinical trials, the IRT-MTA (Interdisciplinary
Research Teams for Molecular Target Assessment) cooperative agreements
have been funded to support extramural teams with expertise in a
molecular target area. The funds will allow researchers to develop
clinically- usable probes (imaging tests, assays, etc.) to determine
whether new targeted agents actually are having the desired effect
on the planned target in pre-clinical models and in patients.
The early clinical trials program (early clinical trials contracts
and cooperative agreements) has been reorganized in several ways:
-
to increase annual accrual to evaluate more targeted agents;
-
to decrease the time from solicitation for trials to completion;
-
to simplify reporting procedures and integrate clinical trials
databases for enhancing safety reporting and patient safety;
-
to support the integration of translational endpoints in early
clinical trials to inform decisions regarding further development.
What are some unique resources that NCI has to offer to scientists
who are interested in molecular targets?
Dr. Sausville: Expertise in discovery (screening and evaluation
in animal models) and development (studies in animals and development
of dose forms and assays for use in humans); discovery resources
include collections of compounds and extracts from natural sources;
and databases---open and free to the public---of how candidate drugs
perform in screening assays. These databases may be linked to the
presence or action of particular molecular targets. (http://dtp.nci.nih.gov)
Is it true that the rate-limiting step in drug development
is often the translation from laboratory to the intact animal --
that is, once the drug has been shown to affect a particular target
by some assay, the difficult part is to achieve high enough levels
in the blood and tumor for the drug to be effective?
Dr. Grochow: Sometimes the rate-limiting step is drug synthesis,
sometimes it's just the time it takes to do the toxicology. More
often it's finding an optimal "drug." Frequently, the first proof
of principle compound in the laboratory isn't really suited to be
a medication -- it's not soluble, can't be given by mouth, has too
short a duration of action, or is too toxic to use in an animal,
let alone a patient.
Are there programs or funds available at NCI to help academic
scientists try to solve some of these problems?
Dr. Sausville: Yes. The RAID program for particular candidate
molecules and the new R*A*N*D (Rapid Access to NCI Discovery) program
for molecules at an earlier stage in their development are programs
to address these sorts of problems. These are described on our website:
http://dtp.nci.nih.gov.
Since many common tumors actually contain dozens of potential
targets, might it be possible to design a single agent that targets
more than one defective protein? Are there any efforts to create
such a molecule?
Dr. Sausville: Yes. Indeed, some of the most promising agents
of this type can target many different proteins. This is exemplified
by SU6668 from the Sugen Company, which targets at least three different
growth factor receptors; Millenium's PS341, which by affecting the
proteasome can alter the behavior of a number of important cell
cycle regulatory proteins; or NCI's 17-allylamino 17 demethoxygeldanamycin,
which can affect many molecules that talk to the drug target, heat
shock protein 90 (hsp90).
Are there any classes of molecules that appear to be particularly
effective as targeted drugs?
Dr. Sausville: I think recent history would point to certain
protein kinase antagonists, including Gleevec and Iressa, as well
as the proteasome inhibitor, PS341.
Dr. Grochow: Not at this time. Antibodies to ErbB-2 or Her2/neu
(trastuzumab or Herceptin) and to certain leukemia cell targets
are the first targeted agents to reach the market, but they have
been followed closely by small molecules like imatinib mesylate
(Gleevec). Because drugs like imatinib mesylate are used in patients
with a disease that has a critical target (making a treatment decision
based on the presence of the molecular target), they may help a
larger fraction of patients than when a treatment is used for all
patients whose tumors look alike under the microscope (making a
treatment decision based on the traditional histologic diagnosis).
### |