&EPA
United States Office of Health and
Environmental Protection Environmental Assessment
Agency Office of Regulatory
Support and Scientific Analysis
Washington, DC 20460
EPA/600/9-87/013
June 1987
Research and Development
Report of the EPA
Workshop on the
Development of Risk
Assessment
Methodologies for
Tumor Promoters
-------
-------
EPA/600/9-87/013
June 1987
REPORT OF THE EPA WORKSHOP ON THE
DEVELOPMENT OF RISK ASSESSMENT METHODOLOGIES
FOR TUMOR PROMOTERS
Prepared by:
Jan Connery
Eastern Research Group, Inc,
6 Whittemore Street
Arlington, MA 02174
EPA Contract 68-03-3234
EPA Project Officer
Hugh L. Spitzer
OFFICE OF HEALTH AND ENVIRONMENTAL ASSESSMENT
AND
OFFICE OF REGULATORY SUPPORT AND SCIENTIFIC ANALYSIS
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
-------
ACKNOWLEDGEMENTS
This document was prepared by Jan Connery, Eastern Research
Group, Inc., Arlington, Massachusetts, for Hugh L. Spitzer, EPA
Office of Research and Development, based on transcripts of the
workshop, it was reviewed by all workshop chairmen and
panelists. Their time and contributions are gratefully
acknowledged.
NOTICE
^ document has been reviewed in accordance with U S
Environmental Protection Agency policy and approved for
publication.. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use
-11-
-------
TABLE OF CONTENTS
Page
PREFACE Vi
EXECUTIVE SUMMARY Vii
1. INTRODUCTION ••• 1
2. DEFINITIONS « 3
Summary 3
Discussion '• 3
3. MECHANISMS OF INITIATION 5
Summary 5
Discussion • 5
4. MECHANISMS OF PROMOTION 8
Summary 8
Characteristics of the Lesions Produced
by Promotion 8
Receptor Mechanism 9
Cell-Cell Communication • 9
Altered Differentiation 10
Oncogenes 10
Behavior in In Vitro Systems 11
Stages of Promotion 12
Spontaneous initiation and Promotion... 12
Memory for Promotion 12
Structure-Activity Relationships • 13
Pure Agents 13
Cytotoxicity • • i4
Promoters as irritants. .. • 15
Promoters as Anti-initiators 15
Chow Diet as a Promoter 15
Promotion by Saline? « 16
5. CHARACTERISTICS OF PROMOTERS 17
Summary • - 17
Criteria for chemicals that Can Only Promote,, 17
Are Promoters carcinogens? « 18
Do Promoters Cause Genetic Damage?... 19
Are Promoters Initiator-Specific? 19
• ill-
-------
TABLE OF CONTENTS (CONT.)
Page
••- a—,—
6. REVERSIBILITY . ;>> 2Q
Summary t 2Q
Existing Data ! *.'!*! *.'****'**'**"!* 20
Mechanisms of Regression „ ..**!*! 21
Demonstrating the Existence of a Threshold."!**!'*' 22
7. THE TWO-STAGE BIRTH-DEATH-MUTATION MODEL...,.., 23
Summary o .. . . 23
Description of the Model I!!!!!!!!!!!!""!!!"! 23
Validating' the Model.. ».....!!!!!!!!! 27
Data Requirements „ !!!!!!!!!!!! 27
Research !!!!!!!!!!!! 28
8. THE CELL-CELL COMMUNICATION MODEL 32
Summary ... „. 32
Model Description !!!!!!!!!!!!!!! 32
Scrape Loading Assay .....!!!!'** 33
Current Data , !!!!!!!!!!." 34
Future Research !!!!!!!!!!!!!!!!! 35
9. QUANTIFICATION IN THE LIVER 36
Summary t ^ ^Q
System Description *'!''*'!'**** 36
Quantification ! '*'*'.''*"** 36
10. SYNBRGISM 38
Summary 3g
Existing Data .....!!!!!!!!!!!!!!!!!!! ss
!!• SPECIES DIFFERENCES/HUMAN STUDIES 40
Summary 40
Species Differences ."**!*' °."!'° ******* 40
Human Studies !*"******'.' 41
12. PROGRESSORS ; 46
Summary .^ > ^g
Mechanisms of Progression...."""""."'!"'""]""'"]""", 45
Promoters as Progressors ******* 48
-iv-
-------
TABLE OF CONTENTS (CONT.)
13. RESEARCH RECOMMENDATIONS 49
Summary 49
Organ Sy-stems « • • • • 50
Chemicals •• • 5^
Pure Promoters and Nonpromoters 52
Human Data 53
Animal Models 53
Species and Strain Differences , 53
in Vitro Systems 54
Mechanisms • 55
Reversibility 55
Receptor Binding 55
Cell Differentiation 56
Oncogenes • 56
Phorbol Esters 56
Models • 57
The Two-Stage Birth-Death-Mutation Model 57
Intercellular communication. 57
Other Models • 58
Quantification 58
Expansion of the NTP Bioassay ;... 58
Additional Research Recommendations 59
14. REFERENCES ........... 60
APPENDIX A
LIST OF PANELISTS AND OBSERVERS
APPENDIX B AGENDA
APPENDIX C PANELIST PREMEETING COMMENTS
APPENDIX D EDITED TRANSCRIPTS OF THE WORKSHOP
-V-
-------
PREFACE
At present, relatively little is known about the mechanisms
of promotion and the identity of promoters. However, available
data suggest that promoters may have very different
implications for risk assessment than carcinogens. On February
3-5, 1987, the U.S Environmental Protection Agency (EPA) Office
of Research and Development sponsored a workshop in Bethesda,
Maryland, on "Development of Risk Assessment Methodologies for
Tumor Promoters." The purpose of this workshop was to identify
and prioritize research to provide data that could be used in
risk assessment of tumor promoters. During the two and
one-half days of the workshop, thirteen expert panelists
discussed the current state of the art in tumor promotion and
developed specific research recommendations. Several observers
were present to witness and join the discussion. This report
summarizes the proceedings of the workshop.
-VI-
-------
EXECUTIVE SUMMARY
At a workshop sponsored by the EPA Office of Research and
Development in February 1987, thirteen expert panelists
discussed research needed to support the development of risk
assessment methodologies for tumor promoters. During the two
and one-half days of the workshop, the panelists exchanged
current data on promotion, identified data gaps, and formulated
general and specific research recommendations.
The panelists agreed that available data suggest that there
are probably at least three stages of carcinogenesis -
initiation, promotion and progression - and that there are
agents that are associated predominantly with these three
stages. Initiation was described as a sudden change probably
involving DNA that is irreversible over a long period of time.
There is a growing body of data suggesting that the initiation
stage is relatively common and involves nonspecific damage to
DNA. There is also evidence that there may be a spectrum of
initiated cells that vary in their degrees of initiation and
thus in their susceptibility to promotion. Promotion was
defined as "the reversible selective clonal expansion of
initiated cells and the reversible alteration of gene
expression." A list of criteria for chemicals that can only
promote was developed. Progression was defined by a majority
of panelists as "an irreversible change in DNA towards
malignancy."
The panelists agreed that the mechanism of promotion is not
currently understood and they suggested that there may be
several different mechanisms of promotion. Available data
suggest that promotion is substantially different from
initiation, and that traditional risk assessment models for
carcinogens are not appropriate for promoters.
-vii-
-------
Promoters appear to show more extreme differences in
species and strain responses than carcinogens. The panelists
agreed that much more work needs to be done to understand these
differences from a mechanistic standpoint. Epidemiological
studies should be conducted to obtain human data, and existing
epidemiological data on promotion should be examined as a
potential source of information on human promoters. Although
no agents have been unequivocally classified as human
promoters, data indicate that several chemicals may be working
as human promoters.
Available data suggest that promotion is reversible in the
liver and skin, but currently there are not enough data to
ascertain whether reversibility is characteristic of all
promoters in all systems. There was concern that there may be
synergism among promoters. Research is needed to study this
phenomenon and to identify the kinds of promoters that are
likely to interact.
There is a need to develop and validate statistical models
for promotion and to develop data to test the models. The
two-stage birth-death-mutation model, developed by Moolgavkar,
Venzon and Knudson, was discussed at the workshop. The
panelists agreed that it appears to provide a good theoretical
framework from which to propose and interpret studies on
promotion. Various approaches to validating the model were
discussed, including an initiation/promotion/initiation
protocol (Potter, 1981) using multiple doses of both the
initiating and promoting agents.
The panelists agreed that not enough data are currently
available to assess the risks of promoters, and that
substantial research is needed in several areas, including:
-Vlll-
-------
« Mechanisms of initiation, promotion and progression,
particularly data on dose-response and frequency of
response.
« The behavior of promoters in humans. Epidemiological
studies of promoters in humans are a high priority. The
panelists suggested several populations for
epidemiological studies.
« Development and validation of statistical models for
initiation/promotion systems.
• The behavior of promoters in organs other than the skin
and the liver.
o Interspecies differences in promotion.
• Expansion of the chemical data base for known and
potential promoters. The panelists offered several
suggestions of chemicals to study.
9 Synergism among promoters.
« Development and validation of in vitro screening models
for known experimental promoters. If successful, the in
vitro approach should expedite the selection of
chemicals for in vivo study.
-IX-
-------
-------
1. INTRODUCTION
In recent years, there has been a growing recognition that
risk assessment of tumor promoters is important but is
precluded by a lack of data. In 1982, the EPA Office of Toxic
Substances held a workshop to examine how information on
promoter activity could be incorporated into risk assessment.
Participants agreed that such information should be
incorporated into risk assessment but could not offer the
Agency guidance on how to do this. Recently, both the Science
Advisory Board in its review of perchloroethylene and the EPA
Office of Pesticides and Toxic Substances' panel on. dioxin
recommended that the EPA consider integrating promotional
activity into the traditional risk assessment.
With regard to promoters, the current EPA Guidelines for
Carcinogen Risk Assessment (Federal Register, 1986) state:
Agents that are positive in long-term animal experiments
and also show evidence of promoting or cocarcinogenic
activity in specialized tests should be considered as
complete carcinogens unless there is evidence to the
contrary because it is, at present, difficult to determine
whether an agent is only a promoting or cocarcinogenic
agent. Agents that show positive results in special tests
for irritation, promotion or cocarcinogenicity and no
indication of tumor response in well conducted and well
designed long-term animal studies should be dealt with on
an individual basis.
While this approach was not felt to be wholly satisfactory,
there was not enough consensus to develop an alternative
approach in terms of either a qualitative judgement of how
likely an agent is to be a promoter, or, quantitatively, of how
great a cancer risk a promoter might pose for given levels of
exposure.
As a first step towards risk assessment for tumor
promoters, the EPA Office of Research and Development convened
-1-
-------
a workshop on "Development of Risk Assessment Methodologies for
Tumor Promoters" on February 3-5, 1987, in Bethesda, Maryland.
The workshop provided an opportunity for expert scientists to
pool th-eir knowledge and set research goals to improve the
scientific bases for risk assessment of promoters. The group
was asked not to address specific chemicals, but rather to
identify research concerning promoters as a class of
substances, and to prioritize this research according to its
impact and utility for risk assessment. The workshop was
chaired by Dr. Albert (University of Cincinnati Medical
Center), Dr. Langenbach (National Institute of Environmental
Health Science), and Dr. Parland (EPA Carcinogen Assessment
Group).
This report summarizes the discussion at the workshop. The
first day of discussion focussed on current knowledge of
promotion. Panelists exchanged data and identified data gaps.
On the second and third days, general and specific research
needs were identified.
The report is organized into 13 sections that reflect the
major themes of discussion:at the workshop. Each section has
been synthesized from many different parts of the discussion
that pertain to the topic. .A list of the panelists and
observers can be found in Appendix A.. The agenda is provided
in App'endix B, and premeeting comments prepared by the
panelists can be found in Appendix c.
The reader should bear in mind that this report is based
solely on the workshop discussion and panelist comments. As
such, it -reflects the opinions and data of a limited number of
participants exchanged over a brief period of time, and
therefore does not provide a comprehensive treatment of the
various subject areas. The amount of information provided on a
particular topic in this report does not. indicate its relative
importance, and there may be important aspects of tumor
promotion that are not touched on in this report.
— 2 —
-------
2. DEFINITIONS
Summary
The group agreed that there are probably at least three
stages of carcinogenesis: initiation, promotion and
progression. Panelists offered various definitions of these
stages in their premeeting comments (see Appendix C). Some of
these definitions were presented and discussed at the
workshop. Discussion focussed on the definition of promotion.
Discussion
The group initially defined promotion as "the reversible
expansion of initiated cells." Dr. Huberman questioned this
definition. In the in vitro hamster embryo cell transformation
system, treatment of initiated cells with tumor promoters such
as phorbol esters produces transformed colonies. The majority
of these colonies revert to a normal phenotype when the
promoter is removed„ Dr. Huberman argued that a promoter can
directly convert an initiated cell from a normal to a tumor
phenotype. Other panelists also thought that promoters may
cause changes in genetic expression, so the group agreed to
define promotion as "the reversible selective clonal expansion
of initiated cells and the reversible alteration of gene
expression."
Dr. Slaga cautioned that the definition must pertain to all
organ systems. He felt that the interval between initiation
and promotion is important in an operational definition. He
suggested that the definition include a requirement that the
promoter must still be effective after a reasonable time (at
-3-
-------
least several months) after initiator application,
proposal was not discussed by the group.
This
Some panelists presented their definitions of initiation
and progression. The group did not develop a definition of
initiation, but a majority of panelists defined progression as
"an irreversible change in DNA towards malignancy." Dr. Magee
pointed out that this definition differs from the original
usage of the term "progression" by Dr. Leslie Foulds (Foulds,
1975).
-4-
-------
3. MECHANISMS OF INITIATION
Summary
There was a general discussion of possible mechanisms of
initiation and characteristics of initiated cells. Several
panelists offered different opinions and data. The mechanism
of initiation was not clear. Initiation has been described as
a sudden change that is irreversible for a long period of
time. The commonly held view is that initiation is a
relatively rare event in vivo that is best explained by a
inutational event. However, recent evidence from several
laboratories (Fernandez et al., 1980; Ethier and Ullrich, 1982;
Clifton et al., 1984; Gould, 1984; Kennedy, 1985a; Terzaghi and
Nettesheim, 1979; Stenback et al., 1981) suggesting that the
initiating event is a common one challenges the original
concept of an initiated cell as one that is mutated. "An
alternative mechanism suggested was that initiation may involve
some irreversible differentiation in cells. How this can be
brought about and what it really means is not clear.
Conventionally, initiation is thought to be linked to genotoxic
agents. It is not clear what kind of genotoxic event would
result in an irreversible change in differentiation. The
frequency of the initiation event would affect priorities for
research. If initiation is common, then promotion and
progression would be rate-limiting steps, suggesting that they
should be given priority in research.
Discussion
Both Dr. Hennings and Dr. Slaga said they thought that
initiation produces a whole spectrum of initiated cells that
vary in their degrees of initiation. Some are more easily
-5-
-------
promoted than others. One panelist suggested that initiation
may have more than one stage. Another panelist thought that
the dose has an effect on the type of initiating event that
occurs.
Dr. Kennedy questioned the notion that initiation was
caused by a single base mutation. She described research by
Mulcahy and others at the University of Wisconsin (Mulcahy et
al., 1984; Gould, 1984) who found that as few as twenty
carcinogen-treated cells were sufficient to give rise to a
cancer in a large percentage of the exposed animals. She also
mentioned research by Terzaghi and Nettesheim (1979), Ethier
and Ullrich (1982) and Stenback et al. (1981) who gave
different doses to mouse skin over orders of magnitude and got
approximately the same final tumor incidence when promoters
were applied to all the animals. it has been concluded by many
researchers that initiation must be a common event, even at low
doses of carcinogen. If the initiating event is common, then
it cannot be a single base mutation in DNA since this occurs at
a very low frequency and would not be expected to occur in a
high proportion of carcinogen-treated cells.
Dr. Pitot described research by Japanese investigators who
claim they can identify initiated cells by an immunohisto-
chemical marker, glutathione transferase-p. Only 1 in 10 or
1 in 100 of the cells identified as initiated expand if a
promoter is applied, so Dr. Pitot questioned the researchers'
assertion that a single change in a marker indicates initiated
cells. He thought that one important characteristic of
initiated cells is the ability to expand in the presence of
promoter, i.e., the promoter selects for the expansion of
initiated cells.
-6-
-------
Dr. Kennedy said her studies at Harvard University's School
of Public Health have indicated that initiation in vitro is
potentially, though not usually, reversible. When protease
inhibitors are given after radiation exposure and'are then
removed, no transformation develops at a later time. This
suggests that some agents such as protease inhibitors can
completely revert cells to a noninitiated state. Dr. Slaga
said that protease inhibitors have very little effect on the
initiation of skin tumors if given at the same time as the
carcinogen exposure. The question of whether they can reverse
initiation if given later in time has not been looked at.
Another issue that has not been addressed is whether any of
the noncarcinogenic mutagens are pure initiators.
Oncogene activation was thought to play a role in
initiation, in the sense that an amplification of a.mutated
c-Ha-ras protooncogene has been demonstrated to push cells
toward malignancy. Dr. Slaga mentioned studies showing that
activated c-Ha-ras will lead to papillomas in the skin if
applied by skin scraping and followed by tumor promotion.
Balmain's (Quintanilla et al., 1986) and the Millers' work
(Wiseman, 1986) suggest that mutation of c-Ha-ras probably
occurs during initiation with some chemicals (including DMBA).
-7-
-------
4. MECHANISMS OF PROMOTION
Summary
There was considerable discussion of potential mechanisms
of promotion. A number of questions were raised. Panelists
offered data from their own experience. The group agreed that
there may be several different mechanisms of promotion, and
that the same promoter may have different mechanisms of action
in different tissues and species. Not only the mechanism, but
the type of action may vary. Some agents may act as a promoter
in one model, and as a complete carcinogen in another.
Promotion in some organ systems may have more than one stage.
There were some data to suggest a structure-activity
relationship for promotion by phorbol esters. Some panelists
concluded, based on limited studies, that promotion might not
involve the activation of some known oncogenes, although
examination of other oncogenes should be undertaken.
Characteristics of the Lesions Produced by Promotion
The consensus was that the majority of the initial lesions
that promoters induce in the skin and the liver are benign
tumors. However, research suggests that there may be a big
difference in papillomas, i.e., that the promoter brings out a
spectrum of transformations ranging from benign tumors to those
that have characteristics of carcinomas such as aneuploidy.
Dr. Hennings offered data to suggest that different papillomas
have very different abilities to progress to cancer. In
studies at the National Cancer Institute (Hennings et al.,
1985), promotion for only 5 weeks (DMBA initiation, TPA
promotion) produced one-fourth as many papillomas as did
promotion for 10, 20 or 40 weeks; however, the number of
-8-
-------
carcinomas was the same regardless of the duration.of
promotion. Every carcinoma apparently arose from a papilloma.
Dr. Slaga mentioned a recent study (in press, PRAS) that he 'and
Drs. Aldaz and Conti had performed which examined benign tumors
induced by promotion. They found that early in tumor
promotion, most benign tumors are diploid. However, after
about 40 weeks of promotion, with treatment twice a week, every
benign tumor was aneuploid with areas that could be called
carcinoma - in situ. Thus, even benign tumors could have
characteristics of carcinomas (e.g., aneuploidy) if they are
analyzed in.detail.
Receptor Mechanism
According to Dr. Pitot, in the skin and the liver the major
known promoters act through a receptor mechanism. There was a
discussion of how one could demonstrate that a promoter is
working by receptor binding (which implies the existence of a
threshold). Since this mechanism is reversible and does not
necessarily involve the tumorigenic process, the ability to
demonstrate a receptor mechanism would have an enormous impact
on the risk assessment of these agents, because it would imply
the use of a completely different extrapolation model than
low-dose linear extrapolation. A receptor .mechanism would
explain why some promoters are tissue-specific. The panelists
offered suggestions about how to study receptor mechanisms (see
Research Recommendations). Further discussion of this topic
can be found in Section 6, Reversibility.
Cell-Cell Communication
Dr. Trosko described data suggesting the promotion may
occur by blockage of intercellular communication (see Section
8, The Cell-Cell Communication Model).
—9 —
-------
Altered Differentiation
Dr. Huberman suggested modulation of differentiation by
tumor promoters as a possible mechanism of action of tumor
promoters. Studies at the National Cancer Institute (Hennings,
Yuspa) showed that treatment of normal epidermal cells in
culture with TPA induced terminal differentiation in about half
the cells. The other half appeared to be unaffected and could
then proliferate. If the initiated cells are among the
unaffected population, this could be how TPA works (Yuspa et
al., 1982). There are several "initiated" cell lines that give
papillomas when put on an animal, none of which give a terminal
differentiation response to TPA (Hennings et al., 1987a; Yuspa
et al., 1986). These cell lines could provide an opportunity
to study particular cell groups that respond differently from
other cells.
Oncogenes
The panelists concluded,that the data indicate that
promotion does not involve activation of protooncogenes by
mutation or transcription,' although evidence in the skin is
incomplete since oncogenes are activated in the promoted
lesions but not in the skin itself. Work by Balmain in the
skin (Quintanilla et al., 1986) and the Millers and others in
the liver suggests that mutation of the c-Ha-ras gene probably
occurs during initiation (Roop et al., 1986). Work in mouse
liver is also inconclusive. Researchers at the University of
Texas System Cancer Center (UTSCC) Science Park did not find
any evidence of expression of several different oncogenes by
promoters in mouse skin in vivo. Data suggest that oncogenes
become activated during progression.
Dr. Pitot reported that studies at McArdle (Beer et al.,
1986) did not show any transcriptional activation of the
-10-
-------
protooncogenes c-mycy c-Ha-ras or Ki-ras in foci or nodules;
however, such activation has been seen in carcinomas, with one
exception (Wogan et al., n.d.), mutational activation of
protooncogenes in rat hepatocarcinogenesis has been either
nonexistent or occasionally transient. But mutational
activation of the c-Ha-ras gene does occur in carcinomas of the
mouse liver. It has also been shown in some mouse adenomas.
.Recently, fairly consistent transcriptional activation of the
c-raf gene has been shown both in nodules and carcinomas in the
rat liver. Some, but not all, foci exhibit transcriptional
activation by in situ hybridization. Some foci show a lowering
or absence of the gap junction protein by the
.immunohistochemical technology; others show normal levels. One
question to be answered is whether the foci that are expressing
c-r.af also have low levels of the gap junction protein (D.
Beer, M. Nevev and H.C. Pitot, unpublished observations).
Dr. Slaga reported that U.TSCC Science Park researchers
looked at the expression of several different oncogenes by
promoters in mouse skin in vivo and did not find any change
except from benign papillomas and carcinomas. -He did not know
of any studies that suggest that the oncogenes are .involved in
promotion in vivo.
Behavior in In Vitro Systems
....- Dr. Kennedy mentioned that in vitro systems provide
dramatic evidence of the presence of a promoter through the
shape of their dose-response curve. The curve is essentially a
quadratic or linear quadratic in the presence of an initiator,
but..becomes linear in the presence of a promoter. She said
that TPA and other agents can promote transformation in cells
thirteen generations after initiation.
-11-
-------
Stages of Promotion
The group discussed whether promotion may have more than
one stage. This was felt to be a possibility in the skin, but
Dr. Pitot indicated there was no evidence for it in the liver.
(For discussion, see Hennings and Yuspa [1985].)
Spontaneous Initiation and Promotion
There was a brief discussion of evidence for spontaneous
initiation and promotion. There is evidence that both
phenomena occur, and these are factors that may impact risk
assessment. In the absence of an initiating agent, the risk
from a promoter will be a function of the background of
spontaneous initiation or the initiating and promoting actions
of the promoter. Dr. Hennings mentioned NCI data suggesting
that spontaneous promotion was occurring. Papillomas were
found to appear 2 to 3 weeks earlier if promotion was delayed
for 5 to 20 weeks following initiation (Hennings and Yuspa,
1985). Dr. Pitot said that studies at McArdle Laboratory
(Pitot et al., 1985) suggest that spontaneous initiation in the
liver occurs up to 6 to 12 -weeks of age, but not from that
point up to a year of age. The number of spontaneous foci is
three or four orders of magnitude lower than the number of foci
induced by an agent. Spontaneous promotion also appears to be
occurring, with a few initiated liver cells expanding in the
absence of an exogenous promoter. The possibility of studying
spontaneous initiation rates in human liver was dismissed
because of the need to serial section the liver.
Memory for Promotion
Dr. Langenbach mentioned German studies (Furstenberger et
al., 1983 and 1985) in which pretreatment with TPA followed by
-12-
-------
an initiator some weeks later and then a stage II promoter
increased the number of tumors. This may mean that there is a
memory for TPA treatment. In similar experiments, Dr. Slaga
found that such memory does not appear to have anything to do
with cell proliferation since the pretreatment time is longer
than the proliferative response by TPA (Slaga, unpublished
results).
Structure-Activity Relationships
Dr. Rosenkranz mentioned a recently completed study at Case
Western Reserve University School of Medicine that suggested
some structure among the PAHs that appeared to correlate with
promoting ability. Thus, there appear to be structural
determinants that contribute to promoting ability. He said a
study of the relationship between structure and promoting
activity would require at least 50 or 60 chemicals,, and
recommended the list of chemicals compiled by Upton et al.
(1984) as a starting point. However, the lack of established
negative chemicals is a problem.
Pure Agents
The panelists discussed whether there are any known agents
that act purely as a promoter or an initiator. Such agents
would be extremely useful for risk assessment-related
research. The group agreed that it is theoretically possible
to determine experimentally whether something is acting as a
initiator or promoter, but some panelists felt this might be
difficult without pure initiators or promoters.
In the liver, phenobarbital and dioxin may be pure
promoters since they have shown no evidence of initiation.
However, dioxin has a very long half-life, which makes it
difficult to study because a single dose is effectively a
continuous dose.
-13-
-------
In the skin, stage-specific agents can be identified.
Urethane is an initiator in the skin, but is a complete
carcinogen in other tissues. Chrysarobin and benzoyl peroxide
are fairly pure promoters in the skin. Likewise the diol
epoxide of benzopyrene can be considered a pure skin tumor
initiator.
The group was unable to identify compounds that were pure
initiators or promoters in all organ systems in which they had
been tested. Likewise, it is difficult to identify
nonpromoting chemicals due to organ/species differences. Given
the present lack of knowledge about promoters, there is a risk
that a nonpromoting chemical in one organ/species may be active
in another.
Cytotoxicity
The relationship between cytotoxicity and promotion was
discussed. While acknowledging that some promoters probably
act without cell killing, Dr. Trosko thought that anything that
was cytotoxic in the liver would also .be a promoter at high
enough doses. He pointed out that any agent that is a mutagen
not only damages DNA but also kills cells at the appropriate
dose, i.e., is cytotoxic, and any agent that can induce
cytotoxicity, which would then force compensatory hyperplasia,
can act as a promoter. So he thought it was important to
acknowledge cytotoxicity regardless of the mechanism by which
an agent kills cells.
There was some question about chemicals that promote
because they are highly cytotoxic. The cytotoxic effects might
occur before all initiated cells have been expressed, thus the
maximal effect may not occur following initiation.
Dr. Hennings said that, in the skin, virtually all
promoters produce hyperplasia, perhaps as a result of
-14-
-------
cytotoxicity. Some promoters may work by a selective
cytotoxicity. But there is no cytotoxicity in the liver with
phenobarbital and dioxin, and there probably is no direct
cytotoxicity by these compounds in other systems. So
cytotoxicity may be promoter-specific.
The effect of substituting a tissue-damaging agent such as
turpentine for a promoter in an initiation/promotion/
progression protocol has not been studied.
Promoters as irritants
The panelists discussed the fact that all known skin tumor
promoters were found as irritants. They agreed that
nonirritant promoters should be identified. Dr. Albert pointed
out that irritation is not a characteristic of liver
promoters.
Promoters as Anti-initiators
Dr. Trosko mentioned that some of the best studied
promoters - PCBs, PBBs, DDT, BHT and phenobarbital - can, in
some circumstances, act as anti-initiators also. If given
before the carcinogen, they protect the animal? if given after,
they promote (Williams and Weisburger, 1986).
Ch^vMDiet as a Promoter
Scientists at McArdle have found that the normal chow diet
is an effective promoting agent in the liver. They speculate
that this may be due to the plant estrogens in the diet, which
vary with the time of year. Semisynthetic diets eliminate much
-15-
-------
of this problem (S. Hendrich and B.C. Pitot, submitted for
publication).
Promotion by Saline?
I
The panel discussed whether saline is a promoting agent.
Dr. Kennedy provided data to suggest that saline instillations
could lead to promotion in the lung (Little and Kennedy, 1982;
Shami et al., 1982 - see also Human Studies in Section 11,
Species Differences/Human Studies). Dr. Pitot did not think
saline could be considered to be a promoter. He thought that
the apparent promoting activity of saline was due to an
alteration of the hormonal environment which leads to an
alteration of gene expression. He said there are many agents
that are not considered to be promoters, but that change the
internal environment in a way that creates effects very similar
to promotion. Dr. Homburger pointed out that the instillation
of saline or anything else into the hamster lung cannot be
compared with human response. Unlike humans, the hamster takes
saline in without any general response - no adrenal
enlargement, hormonal change or struggle. He said that it is
important not to automatically label something as a promoter
just because there is an increased tumor incidence when the
substance is administered following exposure to a carcinogen.
For example, in recent experiments with guinea pigs (McFadden
et al., 1986), the retention of small particles of asbestos was
increased by the inhalation of cigarette smoke. This could
potentially enhance the carcinogenic effect, but it would not
be promotion.
-16-
-------
5. CHARACTERISTICS OF PROMOTERS
Summary
The panelists discussed two questions:
1. What data would allow us to determine that a chemical
has the ability to promote?
2. What data would allow us to determine that a chemical
has the ability only to promote?
Based on this discussion, the panelists developed a list of
criteria for chemicals that can only promote. These criteria
would constitute the weight of evidence for a finding that a
chemical essentially only promotes.
Criteria for Chemicals that Can Only Promote
1. The maximal effect follows initiation.
2. There is an experimentally measurable threshold.
3. The effects at both the cellular and gene level are
reversible at early stages.
4. There is no covalent binding to DNA.
5. In many cases, a receptor mechanism mediates the effect.
6. Promotion may occur in the absence of cytotoxicity, but
certain forms of promotion may involve a cytotoxic
mechanism.
7. There is selective clonal expansion of transformed
cells.
Another suggested criterion to add to the list was "Causes
decreased gap junction function characteristics." There was
concern as to whether the definition for promoters eliminates
-17-
-------
other types Of epigenetic carcinogens as defined by G. Williams
and J. Weisburger, i.e./ chemicals that may cause gene ;.
amplification or gene rearrangement that also cause cancer but
are not really initiators although they may act as
progressors.
Dr. Albert argued that another criterion for promoters is
the induction of benign tumors that take a long time to go to
carcinomas. Dr. Pitot objected on the basis that it is
difficult to define a "benign tumor" morphologically.
Morphology cannot distinguish between a lesion which is still
reversible and one that is permanent.
Are Promoters Carcinogens?
The group discussed whether promoters are carcinogens. Dr.
Pitofe argued that all known promoters are carcinogens because
they cause an age-specific increase in neoplasms. He said that
any agent that results in a neoplasm following application
likely has promoting action. Dr. Albert disagreed. He said
that promoters serve to expand the cell population at any early
stage of transformation before the cells are malignant, and it
is in that expanded cell population that progression toward
malignancy occurs. Dr. Langenbach said that the consensus at
an NIEHS meeting in the fall of 1986 was that with the
information currently available promoters should be considered
as a class of carcinogens. Dr. Trosko said it was a question
of whether initiation, promotion and progression are discrete
stages, but not enough was known about mechanisms to determine
this.
Dr. Slaga pointed out that all promoters that have been
studied in detail show some carcinogenic activity; however,
they do not generally show a dose-response. This suggests that
a finite number of spontaneously initiated cells will saturate
-18-
-------
at fairly low doses. Dr. Pitot pointed out that TCDD,
phenobarbital and saccharin do show a dose-response over a
relatively narrow rangve of doses.
Do Promoters Cause Genetic Damage?
The panelists discussed whether promoters cause genetic
damage. Dr. Krewski thought that the notion of a threshold was
inconsistent with a genetic component associated with
promotion. However, some panelists thought that the
possibility that some promoters can cause genetic damage could
not be ruled but unequivocally. Dr. Huberman pointed out that
Dr. Peter Cerutti and others still assume that genetic damage
is an important component of tumor promotion. Dr. Pitot said
that data suggest that the known liver promoters do tend to
increase DNA synthesis, at least transiently. The panelists
agreed that promoters may cause the reversible alteration of
gene expression and included this in the definition of
promotion (see Section 2, Definitions).
Are Promoters Initiator-Specific?
The question of whether promoters are initiator-specific
was left open. Data suggest that promoters are not
initiator-specific in the liver and skin, but are in the lung
adenoma model using BHT as the promoter (Witschi and Lock,
1978).
-19-
-------
6. REVERSIBILITY
Summary
The panelists discussed whether reversibility - which
implies a threshold - is inherent in the definition of
promotion. This concept has major implications for risk
assessment. If promoters have thresholds, then a no-effect
level could theoretically be demonstrated. The panelists
agreed that, while some promotion appears to be reversible,
there are not enough dose-response data to ascertain whether
reversibility is characteristic of all promoters in all
systems. It is possible that reversibility may be complete in
the liver, but not in the skin, based on the recurrence of
promoted lesions.
Existing Data
The panelists discussed the degree to which reversibility
had been studied in various organ systems, in the liver,
reversibility has not been tested thoroughly with, numerous
compounds, it has been tested with phenobarbital and the
choline-deficient diet, and AAF as a selecting agent. In the
skin, reversibility is more difficult to study than in the
liver because initiated cells cannot be detected in the skin,
which is important for quantification. Of the various skin
models, only the GDI mouse strain has been extensively studied
for reversibility. Initially there is complete reversibility;
however, if promotion is stopped after about 4 to 8 weeks, some
of the initiated cells do not revert but continue to develop
into carcinomas (Verma and Boutwell, 1980). Other studies,
described by Dr. Slaga, also indicate there may be a residual
effect. In these studies, promotion was stopped after about
-20-
-------
4 to 6 applications and then restarted 6 months later. Tumors
appeared much more quickly following the second promotion,
although it was not possible to say whether these were from the
same foci. Thus, reversibility may not be complete in the
skin. Other organ systems, such as the bladder or the breast,
have not been adequately characterized to determine whether
promotion entails reversibility.
Mechanisms of Regression
The panelists discussed possible mechanisms of regression.
One panelist suggested that perhaps papillomas regress to
micropapillomas rather than to initiated cells. Dr. Slaga said
that, in his progression studies, about 30% of the benign
tumors regressed. After promotion was stopped and a progressor
applied, papillomas decreased about 50% in size before becoming
squamous cell carcinomas.
In the liver, the mechanism may vary from one system to
another. In the Solt-Parber liver model, hyperplastic nodules
are produced at the expense of the normal liver, with
remodelling and loss of cells when promotion is stopped. In
other systems (Pitot's, Shulte-Herrmann's), it appears that
cells in the promoted foci die in the absence of the promoter.
Thresholds in the liver do not appear to be pharmacokinetically
determined, since 80 to 90% of the ingested material reaches
the liver.
Dr. Trosko suggested a model for promotion that could
account for reversibility. In this model, promotion is not a
one-hit event in which the promoter blocks the gap junction,
preventing intercellular communication.. In the absence of a
promoter, normal cells suppress the phenotype of premalignant
lesions. When promoters block the gap junction, the
-21-
-------
interactive suppressing effect is also blocked, resulting in
clonal expansion. (The cell-cell communication model is
discussed further in Section 8.) Dr. Trosko suggested that
models be developed that consider intercellular phenomena.
Demonstrating the Existence of a Threshold
There was doubt about whether the presence or absence of a
threshold can be ascertained experimentally for promoters.
Theoretically, if promotion is occurring by a threshold
mechanism and the equilibrium constant of the ligand is known,
the actual concentration of the threshold level could be
predicted. Reversibility might be more easily studied. For
example, the biochemical effect on gene expression of the
compound is reversible and could be measured.
There was some discussion of how in vitro systems could be
used to describe reversibility. Promoters do show a
dose-response and reversibility in vitro. In Dr. Huberman's
experience, about 80 to 90% of the phenotypic changes observed
following promotion in vitro are reversible.
-22-
-------
7. THE TWO-STAGE BIRTH-DEATH-MUTATION MODEL
Summary
The panelists discussed a two-stage, birth-death-mutation
model, developed by Moolgavkar, Venzon and Knudson (M-V-K),
that incorporates the concepts of initiation, promotion and
progression {Moolgavkar and Venzon, 1979; Moolgavkar and
Knudson, 1981; Moolgavkar, 1986). This model can potentially
be used to provide quantitative predictions of risk at various
doses of initiators and promoters. However, the model must be
validated before it can be used in risk assessment.
Specifically, data are needed on normal cell growth,, cell
kinetics at the proliferation stage, and mutation rates and
tumor occurrence as a function of dose. At present, the model
provides a theoretical framework with which to propose and
interpret experiments. The panelists agreed that it would be
worthwhile to perform studies to validate the model and test
the biological notions of initiation, promotion and progression
that it incorporates. Several research suggestions were offered
Description of the Model
The model was presented by Dr. Krewski. It is a stochastic
birth-death-mutation model that involves only two stages
relating to mutational events. Unlike the multistage model,
the M-V-K model explicitly incorporates information on the
kinetics of tissue growth and differentiation. It seems to be
consistent with much of the experimental and epideroiological
data that are currently available on carcinogenesis.
The model assumes three possible fates for normal stem
cells: death, division into normal progeny, or mutation
resulting in one normal daughter cell and an intermediate or
-23-
-------
initiated cell. Likewise, the population of initiated cells
can either divide, die, or undergo a second mutation to produce
a fully transformed malignant tumor cell along with another
intermediate cell (see figure).
Under this model, the age-specific incidence (I) for
cancerous lesions at time (t) is:
Kt) = u1U2 * x(s)e(a2-b2)(t-s)ds
v°
where:
x(s) is the number of normal cells in the tissue at time s.
Ui and U2 are mutation or DNA damage rates for normal
and intermediate cells, respectively.
32 and 02 are the birth and death rates, respectively,
for intermediate cells.
(The birth and death rates of normal cells [a. and b
respectively] do not appear in l(t) since the number of normal
cells [X(s)] is assumed to be sufficiently large so as to
constitute a deterministic process.)
The product of the mutational rates determines overall
tumor incidence. The factors following the integration sign
involve tissue growth, cell proliferation, and cell necrosis;
they determine the shape of the curve.
In terms of this model, initiator, promoter and progressor
can be defined as follows:
An initiator is a substance that increases the rate at
which the first mutation occurs, i.e., it increases ui«
If it is a genetic lesion, it may be reasonable to assume
that anything that increases ui may also increase u2,
although the second rate of increase may be less than the
first. If this is so, then prolonged application with an
initiator may result in complete carcinogenesis.
-24-
-------
cell
tissue proliferation
t » growth s~~A N
determines
overall
incidence
determines shape
of incidence curve
Figure 1. Moolgavkar-Venzon-Knudson Model.
-25-
-------
A promoter is a substance that increases the pool of
intermediate cells available for subsequent malignant
transformation, i.e., it increases the birth rate of the
intermediate cells (32) and/or decreases the death rate
(b2) so that (a2~b2) is positive. Promotion is
assumed to involve a reversible nongenetic mechanism such
as recurrent cytotoxicity or stimulation of cell
proliferation or, possibly, the preclusion of terminal
differentiation (i.e., the death rate of intermediate cells
would be reduced). Thus, a2 and b2 are probably not
linear and have thresholds. The group agreed that the
intermediate or initiated cell in this model should not be
considered as a neoplastic lesion. They did not feel it
appropriate to refer to an expanded colony of such
intermediate cells as a neoplastic change; however, the
group agreed that this expanded colony was not "normal."
The panelists agreed that, in this model, a progressor is a
substance that increases U2«
A complete carcinogen is a substance that increases the
rates of both the first and second mutations. The group
agreed that complete carcinogenesis does not necessarily
involve promotion. Thus, a complete carcinogen could
involve either initiation and progression alone or all
three stages.
In this model, if the mutation rate per initiated cell
division (u2/[u2+a2 ] ) is considered to be a constant,
then an agent that increases the proliferation of the
intermediate cells must also increase the mutation rate for the
second stage. If this mutation rate is a constant, then
substances that possess promotional activity within the context
of this model may also demonstrate some potential for
progression.
The age-incidence curve does not involve the birth and
death rates of unaltered cells (&1 and b1) because the
original tissue mass can be thought of as being sufficiently
'large to be described as a deterministic rather than a
stochastic process. All that enters in is the number of cells
in that tissue as a function of time.
-26-
-------
Validating the Model
Data Requirements
The panelists discussed what data are necessary to validate
the model and how they could be obtained. The first data
requirement is the growth rate of the normal cells [x(s)l in
.the tissue of interest as a function of age. This can be
obtained fairly readily from studies that are separate from the
bioassay. Second, information is needed on the birth and death
rate (a, and b2) of the intermediate (or initiated) cells.
This could be obtained from laboratory assays for promotion.
Third, information is needed on the transformation rates for
initiation and progression (i^ and u2).- This, in
conjunction with the previous data, could be obtained from a
2-year rodent bioassay. Possibly, an IPI protocol might be
needed to factor out the progression step. To fully examine
how mutation rates vary with dose, a bioassay would have to be
conducted with various levels of exposure to the initiator.
More work needs to be done by statisticians to determine
exactly what kind of data are needed to estimate specific
parameters in the model and, in particular, how to separate
u, and u~.
I(t) is the time rate of appearance of lesions in the
bioassay. These data could be obtained by (1) counting lesions
in the skin, (2) using serial sacrifices in the liver, (3)
assuming that the lesion of interest is rapidly fatal so that
the survival time of the animal serves as the proxy for the
actual time to tumor induction, or (4) assuming that death as a
result of tumor occurrence in a bioassay is independent of
death from competing causes, in which case time to tumor could
be separated out statistically.
-27-
-------
If all the other parameters of the model are known, then
the rate of spontaneous initiation can be determined from the
rate of spontaneous tumor formation in the unexposed controls.
There was some discussion of how partial hepatectomy in the
liver would affect the variables in the model. This was not
resolved. Dr. Krewski thought that data on all variables would
have to be obtained under similar conditions. This could be a
problem since partial hepatectomies are not normally performed
in rodent bioassays. Dr. Pitot thought that partial
hepatectomy would change only u;L in his system since many
studies have shown that initiation in the liver and other,
tissues will not occur in the absence of cell division. Dr.
Travis argued that, since partial hepatectomy only increases
cell division, it would affect the birth rate (a, and,
probably, a2) and x(s) but not u,.
Research
The group discussed past, present and future studies that
may help to validate the model. One panelist said that several
initiation/promotion/initiation (IPI) studies have shown a
dramatic increase in the crop of malignant lesions observed at
the end of the study. This tends to confirm the idea that the
second initiator enhances the mutation rate, u , of the
promoted pool of initiated cells. Dr. Krewski mentioned that
the effects of changes in the dosing pattern .over time on
carcinogenic risks are currently being investigated.
. Dr. Travis described recent research performed at Oak Ridge
National Laboratory to validate the model, in this study,
background liver cancer rates in ratSias a function of age
[I(t}] were obtained from NTP data down to about one-tenth,of a
percent. The growth rate of the liver as a function of age
-28-
-------
[x(s)J was found from the literature.
estimated from mitotic rates in the literature
The term (a^b^) was
The terms a.
and u2 were both assumed to be equal and constant with age.
The model was run using these data, and it exactly reproduced
the age-specific liver cancer rates in rats. Dr. Travis said
the similar studies would be done for mice and humans.
He described plans for a study with tetrachloroethylene,
which has been shown to be carcinogenic in mice. Data on the
dose-dependency of the increase in cell turnover rate are
available from the animal bioassay. Scientists at Dow Chemical
Company have measured the increase in mitotic rate as a
function of applied dose to the liver. Dr. Travis has
estimated the increase in mitotic rate as a function of
effective dose to the liver and found that it increases
linearly above a no-effect-level threshold. Dr. Travis plans
to enter these data into the model. He will assume that
tetrachloroethylene has no genotoxic effects, thus u;L and
u will be equal to the transition probabilities associated
with the background cancer rate. Using these data, the model
should predict the age-specific incidence of cancer from the
tetrachloroethylene bioassay. If so, this would suggest that
tetrachloroethylene is working solely through a promotional
mechanism.
Dr. Travis recommended that similar research be conducted
to validate the model, i.e., obtain (from the literature or
studies) background cancer rates, mitotic rates as a function
of age, and increased cell turnover rates as a function of
dose; run the model with these data to see if it predicts the
observed cancer bioassay rates. (Cell turnover rates provide a
measure of the turnover rates of normal cells and not foci;
this approach assumes that these rates are the same.) This
model approach could be used, for example, with the hamster
lung. Moolgavkar has used this approach for the breast,
-29-
-------
although more work could be done on this tissue. Dr. Travis
pointed out that the model assumes that the growth rates of
foci (a2-b2) are time-independent. He felt that the rate
of foci growth would increase with time, and suggested that
experiments be done to look at livers at different times to see
if their volumes increase at a different rate.
Dr. Trosko proposed an IPI protocol (Potter, 1981) to
validate the model. According to the model, the pool of
promoted cells should vary depending on the duration of
promotion. One way to validate it would be to conduct an IPI
study in which groups of animals are initiated with the same
dose of initiator, promoted for various lengths of time, and
then exposed to the same dose of a new initiator. Dr. Trosko
suggested using x-rays as the second initiator because this
would eliminate complications of metabolism or selective
mutagenicity of the cells. Dr. Hennings said this research had
been performed at the National Cancer Institute but the data
have not yet been published. NCI scientists initiated with
DMBA, promoted for 5, 10 or 20 weeks, and then injected i.p.
with urethane. Response to urethane was best with the shorter
promotion (Hennings et al., 1987b). Similar results were
obtained when 4-nitroquinoline-N-oxide was applied topically to
papilloma-bearing mice.
Dr. Krewski suggested the following protocol to model dose
dependency. Initiation in the chronic bioassay with three
doses (e.g., 0, 0.5 and 1, where 1 is the maximum tolerated
dose) in a single application, followed by chronic exposure to
the promoter at three doses (e.g., 0, 0.5 and 1). There are
thus nine possible exposure combinations of initiator and
promoter. This would provide dose-response data for the
initiator and promoter.
-30-
-------
If u« and U2 are the same, an IP protocol would be
sufficient to estimate the product of the two rates and hence
their common value. If ^l and u2 are not the same, a
second initiator (or progressor) would be essential to separate
out the rates at which the two mutations occur. The IP
protocol described above could be expanded by including a
single exposure to varying levels of a second initiator
following promotion. The second initiator may or may not be
the same as the first.
Smaller experiments could also be conducted. For example,
if spontaneous initiation is occurring at a sufficient rate for
the promoter to be effective, then the initiation step could be
omitted.
Saccharin in the bladder was mentioned as a system to study
to validate the model. Cohen and Ellwein have been doing work
with the saccharin data base. A critical review of their work
might help define further research in this area.
-31-
-------
8« THE CELL-CELL COMMUNICATION MODEL
Summary
Dr. Trosko presented a model for promotion involving
cell-cell communication. He also described a new technology -
the scrape loading/dye transfer assay - that has potential for
testing mechanisms of promotion at the cellular level (El-Fouly
et al., 1987).
Model Description
In the cell-cell communication model, promotion (i.e:.,
selective clonal expansion of initiated cells) occurs as a
result of removal of the suppressive contact-inhibiting effects
of the normal neighboring cells (Trosko et al., 1983). Removal
can be effected in a number of ways, including wounding,
surgery, physical irritation, and placement of a solid, such as
plastic or metal, next to initiated cells. The cell-cell
communication model introduces a higher order of biology into
the M-V-K two-stage model because the phenotype and future of '
the initiated cell depend totally on the communication
properties of the normal neighbors.
In this model, promoters may also cause normal cells to
proliferate. However, normal cells go into terminal
differentiation after proliferation. With initiated cells,
once the critical mass of the initiated cells gets large
enough, the suppressing effects of the normal cell neighbors
are diluted out.
Intercellular communication of molecules below 1,000
daltons is mediated by gap junctions. .These structures are
-32-
-------
found in virtually all normal cells in every organ. All the
critical .molecules and ions below 1,000 daltons are in
equilibration for cells coupled by gap junctions. Gap
junctions are modulated by drugs, food additives, nutrients,
endogenous growth factors, biological toxins, pollutants,
neurotransmitters, hormones, heavy metals and several oncogenes
(Trosko et al., in press).
Scrape Loading Assay
Dr. Trosko described a new technology - the scrape
loading/dye transfer technique ~ which he felt had potential
for testing mechanisms such as thresholds, reversibility,
synergisms and antagonisms that have been speculated in the
animal promotion model. Scrape loading is an extremely simple
technology that can be used with any animal or human cell.
Human cells and approximately 75 different cell strains and
lines have been tested with this technology.
in this technology, cells ace grown to confluence, which
mimics the normal situation in solid tissues. Then two dyes -
lucifer yellow and rodamine red dextran - are applied to the
cells. The yellow dye is a small molecular weight dye that can
easily pass through gap junctions once it penetrates the cell.
The red dye is too large to pass through gap junctions.
Normally, the dyes will not pass through the cell membrane.
However, the next step is to scrape the cells with a toothpick,
which simulates wounding. The dye enters the cells along the
wound line where the membranes are temporarily disrupted. The
membrane heals within milliseconds, trapping the dye in the
cells along the edge. The cells are then washed, immediately
put under a fluorescent microscope with two filters, and
photographed. At this stage, both dyes can be seen at the
edge. The cells are then put back in the incubator with and
-33-
-------
without a presumptive modulator of cell-cell communication.
Five minutes later the cells are photographed again, if the
cells have good gap junction function, the yellow dye will have
diffused away from the edge but the red will not. in
dose-response studies with this technique, a clear no-effect
level can be seen, in cells that communicate well, the rate of
communication (i.e., dye diffusion) can be guantitated with a
laser machine, it varies between different types of cells.
All studies are done at noncytotoxic doses. At cytotoxic
doses, dye goes in all the cells.
Current Data
In vitro data indicate that there are at least three
classes of promoters: those that have receptors and work at
nanogram levels (hormones, TPA, TCDD, etc.); those that do not
seem to need receptors (DDT, PBB, etc.) but diffuse into the
membrane because they are lipophilic (these agents usually work
at microgram levels); and those that do not need receptors but
are not lipophilic (e.g., saccharin - these usually work at
milligram levels).
Four intercellular chemical messagers seem to be
responsible for modulating gap junctions: PKC, calcium, PH and
cyclic AMP. The first three close gap junctions. Cyclic AMP
has been shown to increase gap junction communication in
certain cells (Spray and Bennett, 1985).
There is now direct evidence that many known growth factors
work by blocking contact inhibition. A few growth factors have
been shown to have promoting properties. Over 100 chemicals'
have been tested at Michigan State University. For several
chemicals that were tested in vitro and in vivo, the in vitro
results predicted the in vivo promoting ability of the agent.
-34-
-------
Drs« Lowenstein and Borek pointed out about 20 years ago
that most cancer cells seem to have defects in their gap
junction communication (Kanno, 1985). The scrape loading
technique corroborates this. Some tumor cells do not seem to
communicate at all. Several oncogenes seem to block cell-cell
communication when they are expressed in the appropriate cell.
At Michigan State University, studies were conducted to
compare the metabolic cooperation assay using V79 and rat liver
WB cells with the scrape loading assay. Over 100 chemicals
were tested in the metabolic cooperation assay. It takes at
least 3 days in the metabolic cooperation assay before the
cooperative donor cell can die or the recipient cell can be
rescued. In the scrape loading assay, TPA blocks communication
in liver cells, but only for an hour or two. However, in the
metabolic cooperation assay, TPA does not appear to block
cell-cell communication, because cell communication inhibition
was transient and the cooperating cell dies. So these assays
measure two different responses: a transient response and a
more long-term response. Also, the metabolic cooperation
systems use serum which contains growth factors so it may not
reflect in vivo conditions. The scrape assay appears to be a
better mimic of in vivo conditions because it can use
serum-free media.
Dr. Trosko also described research suggesting synergism can
occur between promoters that modulate gap junctions (see
Section 10, Synergism).
Future Research
The panelists agreed that cell-cell communication should be
studied further. Suggestions for future research are provided
in Section 13, Research Recommendations.
-35-
-------
9. QUANTIFICATION IN THE LIVER
Summary
Dr. Pitot described the liver system studied in his
laboratory and how it can be used to quantify the potency of a
single chemical for the three stages ofcarcinogenesis.
System Description
The liver system used at the McArdle Laboratory for Cancer
Research is analogous to the skin with one exception:
initiation must take place during cell proliferation. Cell
proliferation is stimulated using a partial hepatectomy. Then
diethylnitrosamine (or another agent) is administered, followed
by the promoter usually continuously in the diet, in drinking
water or by gavage. Lesions are identified using three
different histochemical markers. Oncogene expression is
examined using in situ hybridization.
Quantification
The foci can be quantitated using computers. A computer
plot is obtained for each of three serial sections stained for
three different markers. These are overlaid to determine the
phenotype of each focus and the number and volume (or area) of
the foci.
The potency of initiation and promotion can be quantified
by developing an initiating index and a promoting index. The
initiating index is the log of the number of foci (corrected
for the background level) per liver per millimole of the
-36-
-------
compound given in a single dose. The initiation index for TCDD
and phenobarbital is zero. The promoting index is the volume
occupied by the foci in the liver in the presence of the
promoter divided by the volume of the foci in the absence of
the promoter per millimole per week. This index measures the
ability of the promoter to expand the population of the progeny
of initiated cells. It is dependent on time. The effect of
the promoter on the initiation index and of the initiator on
the promotion index has not yet been studied extensively in
this system.
In studies at McArdle, researchers have
foci. Generally, these are morphologically
involve only a small part of the population
.focus. However, in an IPI experiment, the
increase by at least an order of magnitude.
interpreted as a transition from promotion
Quantification of the foci within foci may
measuring the transition from promotion to
found foci within
carcinomas and
of the original
foci within foci
This hois been
to progression.
provide a means of
progression.
-37-
-------
10. SYNERGISM
Summary
The issue of synergism of promoters was discussed, with the
conclusion that very little is known about this possible type
of interaction. Synergism has not been considered in risk
assessment before and may be an important factor. Research is
needed to elucidate the mechanism of synergism among promoters
and to identify the kinds of promoters that are likely to
interact.
Existing Data
Dr. Trosko described research on potential synergism
between two promoters - DDT and TPA - that was recently
performed at Michigan state University. These studies have
been submitted for publication (Aylsworth et al., submitted).
The research was conducted to investigate whether synergism
could occur by modulating gap junctions. The researchers
postulated that the action of some promoters is mediated by PKC
(a phospholipid, calcium-dependent enzyme), if so, then a
promoting agent that stimulates the phospholipid component of
PKC and a promoting agent that modulates calcium should react
synergistically. The study investigated the interaction of
DDT, which blocks the efflux of calcium through the membrane,
and TPA. Either chemical alone produced a clear dose-response
curve, when TPA was held constant and DDT was added over the
same dose range, the agents showed synergistic rather than
additive effects. Synergism was also found between unsaturated
fatty acids and DDT, whereas DDT and aldrin show additive
effects. Quercetin, which is an inhibitor of PKC, was found to
completely block the TPA effect on cell-cell communication
-38-
-------
(unpublished results). Dr. Trosko suggested that synergism of
DDT and TPA be investigated in the liver and skin.
Dr. Pitot has found that the lab chow diet appears to have
a synergistic effect with phenobarbital as a promoter in the
initiation/promotion system in the liver. This could
potentially confound studies to identify promoters.
One panelist questioned why cis-retinoic acid and TPA show
synergism in the metabolic cooperation system, but cis-retinoic
acid is antagonistic to TPA in mouse skin.
-39-
-------
11. SPECIES DIFFERENCES/HUMAN STUDIES
Summary
The panelists discussed species differences in response to
initiation, promotion and progression. Strong carcinogens are
notable for attacking multiple strains and species; however,
this is not necessarily the case with some of the promoters
that have been studied. Thus, it appears that promoters may
show more extreme differences in species and strain responses
than carcinogens. This would make risk assessment for
promoters even more difficult than for carcinogens. The
panelists agreed that much more work needs to be done to
determine whether there are species differences and to
understand these differences from a mechanistic standpoint.
The panelists proposed several ideas for human studies.
Species Differences
t
Phorbol esters promote in some mouse species or strains but
not in others. TPA has different effects in different
species. Repetitive treatment with TPA results in sustained
hyperplasia in the mouse, but not in the rat or hamster.
Phenobarbital may have different effects in rats and
humans. Phenobarbital is known to cause liver tumors in rats.
However, there was no evidence of increase in any type of
neoplasm in 25,000 patients who were given PB as an
anticonvulsant for several years (Clemmesen, 1977)„ Dr. Pitot
pointed out that the human dose in the Clemmesen study was
comparable to the threshold for neoplastic effects in rats, so
it may be that the human dose was simply not high enough to see
an effect. •••••
-40-
-------
DDT was mentioned as an example of an agent that has been
clearly shown to be a promoter in the liver in rodents, but
does not seem to be carcinogenic in humans in at least 30
epidemiological studies. One panelist suggested that maybe the
human dose was not high enough to see promotion.
Human Studies
Several panelists emphasized the importance of doing
epidemiological studies to elucidate how promoters affect
humans. A suggestion was made to study the correlation between
human and animal data for known or potential promoters for
which epidemiological data are currently available, e.g.,
cigarette smoke, arsenic and dioxin. Another suggestion was
made that animal studies for promoters should focus on agents
for which there is human exposure and thus the opportunity to
collect epidemiological data.
Dr. Kennedy described several human populations that could
be studied for promotional effects (Kennedy, 1985b). These
populations have been exposed to an initiator and,
subsequently, to a potential or known promoter.
Two such populations are individuals who have received
occupational exposure to uranium mine dust or asbestos and are
now having their lungs gavaged with saline at regular intervals
to remove the material. . In these populations, asbestos or
alpha radiation is the initiator and saline instillations could
lead to promotion. Studies conducted at the Harvard University
School of Public Health several years ago suggest that saline
instillations can result in promotion (Little et al., 1978;
Shami et al., 1982). In these studies, a relatively low dose
of polonium 210 to hamster lungs produced few lung tumors,
whereas treatment with seven instillations of saline 5 months
-41-
-------
later resulted in lung tumors in 22 to 44% of the exposed
animals. Saline instillations alone did not lead to cancer.
Other panelists questioned whether saline could be considered a
promoter (see Section 4, Mechanisms of Promotion).
Uranium miners who smoke versus those who do not smoke are
another population that could be studied. The curve for the
induction of lung cancer in white uranium miners who smoke is
linear (Committee on Biological Effects of ionizing Radiation,
1972). The nonwhite miners (American Indians) who do not smoke
(or who smoke very little) had a nonsignificant incidence of
cancer (Lundin et al., 1971), although the incidence of cancer
has been increasing recently (Archer et al., 1976; Gottlieb and
Husen, 1982; Samet.et al., 1984). Occupants of many houses and
other structures in the United States are exposed to levels of
radon comparable to those which are known to exist in uranium
mines. This could be another population to study.
Other human populations that could be studied for promotion
with radiation as the initiating agent include individuals who
were exposed to radium occupationally or were given radium for
medical problems and who are now at high risk for the
development of bone cancer.' Many people were exposed to x-rays
in the 1940's and 1950's for various benign disorders such as
eczema, acne and thymus enlargement. Dr. Kennedy suggested
that, in the X-irradiated population, individuals who have
contracted thyroid cancer might be very appropriate to study.
Approximately 20,000 cases of thyroid cancer are expected to
result from therapeutic x-radiation treatments in the United
States. Women, individuals with a Jewish ethnic background,
and emigrants from Tunisia and Morocco appear to be at higher
risk of developing cancer (Committee on Biological Effects 'of
Ionizing Radiation, 1980; Ron and Modan, 1982); these unusual
risk groups suggest that promotion plays a role in the
development of this disease. Mortality from thyroid cancer is
-42-
-------
extremely.low (1 to 3%), so that affected individuals are
available for interviewing. At many hospitals, thyroid cancer
is diagnosed by scanning with iodine 131 at approximately
yearly intervals. This gives a dose of approximately 200 rads
to the adult thyroid, which is considered the optimal dose for
the induction of cancer in some systems and could be an
excellent progressor., ..
. Many of the different cancers resulting from radiation
exposure show a dose-response relationship. The thyroid and
the female breasts have the greatest sensitivity to
radiation-induced cancer, and the dose-response relationship
for both types of cancer is linear (Committee on the Biological
Effects ,of Ionizing Radiation, 1980; U.N. Report, 1977; Maxbn
et al.,,1977). Both the breast and the thyroid are under
strict hormononal controls and these hormones mayvact as
built-in promotional agents (Troll,. 1976). Both thyroid
hormones and those affecting the breast act as promoters or
^carcinogens in several systems: in vivo (Berenblum, 1974;
..Foster,,. 1975; Doniach, 1974; Hall, 1948; Suss et .al.., 1973) and
in vitro (Blumberg, 1980,and 1981; Weinstein et al., 1979;
Guernsey et al., 1980; Fisher et al., 1983; Borek- et al.,
.1983). As promotion in the laboratory results in linear
c,uryes, the linear dose-response relationships for
radiation-induced breast and thyroid cancers suggest that
promoting agents may be important in the genesis of both
thyroid and breast cancers in human populations.
~ One potential source of data on the levels of potential
promoters in tissues from human populations could come from
samples of breast tissue from reduction mammoplasties and
mastectomies.. Samples from reduction mammoplasties are
currently maintained in liquid nitrogen by Dr. Michael Gould at
the Department of Human. Oncology, Wisconsin Clinical Center,
Madison, Wisconsin (608-263-6615). A study could be performed
-43-
-------
to compare levels of potential promoters in the breast tissue
samples with known breast cancer rates in the areas of the
country from which the tissue originated.
The Moolgavkar-Venzon-Knudson model was developed based on
patients with retinoblastoma. This human model may offer a
opportunity to study progression. Tumors (which also appear in
many sites other than the eyes) are caused by an inherited gene
mutation, where all the somatic cells in the embryo are
initiated. Presumably promotion occurs in the eye because of
differentiation of the eye tissue. Chemotherapy is inducing
very high cancer rates in other tissues in the survivors. The
survivors of therapy for retinoblastoma could be studied for
progression. The advantages of this model are that the
mechanism is fairly well understood and it seems to fit well
with the M-V-K model. If retinoblastoma proves to be a model
of IPI, then animals that are genetically susceptible to organ
site cancers could also be studied to determine whether they
have inherited an I state or a P state.
Another human model mentioned was xeroderma pigmentosum.
This recessive disease predisposes the individual to initiation
in the skin by UV light. Most, but not all, tumors form in the
skin. Dr. Trosko mentioned that Dr. Kraemer at the National
Cancer Institute had been studying nonskin tumors in this model
as evidence of progression {Kraemer, 1980). Since UV cannot
penetrate internally, these tumors must result from exposure to
chemical initiators. Thus, this model could potentially be
used to study IPI. Nevertheless, it is a good example of a
mutant that might have an unusually high background initiated
state.
Two other human models are patients who receive PUV-A
(psoralen plus near-UV light [360 nm]) therapy and then X-ray
therapy later; these individuals rapidly get tumors.
-44-
-------
Individuals with psoriasis, having high rates of cell
proliferation, might also be a possible population to study.
Epidemiological data should be obtained for chemicals that
are known to be promoters in animals. For example, saturated
fatty acids are associated with breast tumors in rats. The
change in the U.S. diet in the last few years may offer an
opportunity to see whether there is a similar association in
humans. In addition, existing epidemiological data should be
examined as a potential source of data on promoters.
-45-
-------
12. PROGRESSORS
Summary
The concepts of progression and progressors were
introduced. There was general agreement that progression may
be a distinct stage in carcinogenesis (Hennings et al., 1983)
and that there are probably agents that act predominantly as
progressors. It is likely that promotion and progression can
be separated experimentally in the liver as well as in the
skin. (The liver has the advantage that single initiated cells
and their early clonal progeny can be identified to provide a
quantitative measurement of the effectiveness of the
progressor.) The mechanism for progression was not understood;
however, a majority of the group agreed that progression is
probably related to additions in DNA. A proposed definition
for progression was "an irreversible change in DNA towards
malignancy." The issue of whether progression can occur by
other mechanisms was left open.
Mechanisms of Progression
The group agreed that there was some evidence suggesting
that progression is related to DNA damage, and they considered
other possible mechanisms. Dr. Hennings suggested that
progression indicates all the changes that occur after the
development of a benign lesion, and might be divided into at
least two stages: "malignant conversion" (i.e, the conversion
of the papilloma to a squamous cell carcinoma), and metastasis
(the spread of this malignant tumor to other organs) (Nowell,
1986). Dr. Pitot proposed that oxygen radical effects (which
are indirect) may take a cell in the reversible stage of
promotion and place it into progression. He said that benzoyl
•46-
-------
peroxide and hydrogen peroxide are probably the best current
examples of progressors. He also thought that gene
amplification is involved in progression.
Oncogene activation was thought to play a role in
progression. Dr. Hennings said that studies in progress at the
National Cancer Institute indicate that c-Ha-ras activation can
be either the initiating step (Roop et al., 1986) or the
progressing step. Dr. Pitot said that most studies
demonstrating oncogene activation indicate the activation of
oncogenes during progression (Nicolson, 1987; Nowell, 1986;
Klein and Klein, 1985).
The relationship between cytotoxicity and progression was
discussed. Studies at UTSCC Science Park indicate that, in
some cases, progression appears to be a process that selects
more aggressive cells through cytotoxicity and leads to
cancer. Initial studies by Slaga and coworkers suggested that
there is a fairly strong association between progression and
cytotoxicity. However, Dr. Slaga expressed some doubt that
cytoxicity was the mechanism by which all progressors worked,
even though all the compounds are cytotoxic. Several
experiments were performed at UTSCC Science Park in which the
dose of TPA was raised to the point of cytoxicity. Even at
these doses, it did not act as a progressor. On the other
hand, the antidiol-epoxide of benzo(a)pyrene, which is a
noninitiator in the skin, is extremely potent as a progressor
when applied to benign tumors.
In the skin, application of an initiator followed by a
progressor (i.e., no promoter application) gives different
results depending on the agent. Urethane gives no malignant
tumors and 4NQO gives very few (Hennings et al., 1986), whereas
complete carcinogens such as MNNG do give tumors.
-47-
-------
Many of the progressors in the skin also progress in the
liver. Dr. Pitot mentioned a study by Scherer et al. (1984)
that showed that ENU acted as a progressor in the liver
following initiation with diethylnitrosamine and promotion with
phenobarbital. Research in the liver has focussed on classical
initiating agents - alkylating agents - all of which are
clastogenic. Other types of agents need to be studied.
Promoters as Progressors
The ability of promoters to act as progressors was
discussed. Several panelists offered data from their studies.
It appears that the ability of promoters to cause progression
varies with different agents. Dr. Hennings mentioned studies
where application of TPA following initiation and promotion
caused no increase in the progression of papillomas to
carcinomas (Hennings, 1983).
Studies at UTSCC Science Park with promoters (including
benzoyl peroxide and chrysarobin) and cytotoxic.agents (such as
acetic acid and hydrogen peroxide) suggest that some promoters
can act as progressors; however, cytotoxic agents can be just
as effective.
-48-
-------
13. RESEARCH RECOMMENDATIONS
Summary
The nature of promoters has important implications for risk
assessment. Data suggest that promoters may have very
different characteristics from complete carcinogens. However,
the panelists agreed that not enough data are currently
available to perform risk assessment for promoters. Probably
less than 30 to 40 tumor promoters have been studied. Most
tumor promotion studies have not been designed to consider risk
assessment.
The panelists identified many different general and
specific areas of research and made specific suggestions for
studies. Basic research is needed to elucidate the identity
and mechanism(s) of promotion, initiation and progression. .
More chemicals need to be studied for their promoting
capability, and much more data are needed on the behavior of
promoters in organs other than the skin and the liver.
Epidemiological studies should be conducted to provide human
data on promotion. Models to quantitate initiation and
promotion should be developed and validated. In vitro
screening models for promoters should be developed. Several
questions of particular significance for risk assessment are:
Is promotion reversible and does it have a threshold? How does
the action of promoters vary from one organ to another and from
one species to another? To what degree do agents that have
been identified as promoters also possess initiating and
progressing abilities? How can the potencies for these actions
be quantified?
Many scientists are concerned that promoters may pose as
great or greater an environmental hazard than complete
-49-
-------
carcinogens. The panelists agreed that long-term research on
promotion was needed and is important. However, they felt
that, even with a relatively high level of funding, it may take
5 to 10 years to generate sufficient data to formulate a risk
assessment policy for promoters.
Organ Systems
The skin is less attractive for purposes of modelling
because skin lesions are less of a human health concern than
lesions in some other organs. (Although skin cancers are the
most numerous of human cancers, they are not as
life-threatening as some other malignancies.) The liver has
the disadvantage that a partial hepatectomy is required' in the
adult (but not in the neonate [Peraino model]). This might
make it difficult to get in vitro cell transformation,data
under comparable circumstances.
The panelists agreed that a major priority is to find out
whether the promoting characteristics that are true of the two
most studied systems (the skin and the liver) are also true in
other systems. Panelists proposed and discussed several other
systems to study. The bladder would be a good system to study,
although it may be complicated by the fact that urine itself
acts as a promoting agent. Saccharin was proposed as a
chemical to study in this system. It has a strong promoting
effect and there are several good initiators. The colon is- .
another system to study.
Dr. Pitot thought it important to develop a good system
with a hormonal background. The thyroid is one possibility..
It has the advantage that it is cellularly homogeneous. The
breast may be more complicated. The kidney may not be a good
system to develop because it is cellularly heterogeneous so
-50-
-------
tumors may derive from many different cell types, which may
have different responses to promoters. The lung is a difficult
system to study but important because it is a major route of
exposure. Unlike the skin and the liver, the respiratory
system of animal models produces lesions that are
histologically identical to human lung lesions. In addition,
two alternative systems that may hold promise for study are:
the Klein-Szanto and Nettesheim tracheal system (which involves
a denuded trachea that can be repopulated with human cells to
study human tissue) and Craighead and Mossman's system of
hamster tracheal explants (Mossman and Craighead, 1978).
Panelists also discussed the possibility of testing
promoters in several systems or organs in the whole animal.
This would be possible if a universal agent were identified
that initiated in a number of organ or tissue systems.
Alternatively, a cocktail of chemicals could be used to
initiate every organ. This approach would be difficult if
promoters are shown to be initiator-specific.
Chemicals
The panelists discussed which agents should be studied
further. Several agents were mentioned. Panelists differed in
their opinions about which should receive priority. Some
panelists thought that studies should focus on chemicals that
\
pose the greatest potential human health hazards. Others felt
it was important to enhance the data base for substances that
have already been well studied in order to understand how they
act in the model systems. The chemicals proposed for study
r
were:
• TPA. This is a good promoter but its promoting ability
is limited to skin of various strains of mice. This may
be because the ester groups are removed, eliminating its
promoting ability.
-51-
-------
• Teleocidin may be a better chemical than TPA for
promotion studies because it is not metabolized.
• TCDD is a good candidate for study. One area for study
is the receptor-TCDD interaction. TCDD actions are
believed to be mediated through a receptor, but the
affinity of TCDD for the receptor is not related to its
toxicology. One explanation may be that TCDD actions
are mediated by the affinity of the receptor-TCDD
complex for DNA. This needs to be studied.
• Chrysarobin is probably also a good chemical to study,
since it promotes tumors at low doses in the skin.
• Sodium phenobarbital.
• PBB.
• Alcohol acts as a promoter in epidemiological studies.
However, it generally has not been shown to act as a
promoter in the liver in experimental studies. Alcohol
changes the total number of altered cells and thus the
volume of foci that are present after initiation. It
does not increase the number of foci (S. Hendrich, T.
Glauer and H.C. Pitot, unpublished observations). This
is an interesting mechanism of action if ethanol is a
promoter.
• Pesticides. .
_____^___ .,
• Solvents.
Pure Promoters and Nonpromoters
The lack of negative chemicals, i.e., agents that have no
promoting ability, was discussed. The lack of negative
chemicals has been a problem in testing tumor promoters in in
vitro assays. Thus, one needed area of research is the
identification of chemicals that act primarily as initiators or
promoters.
The panelists agreed that there are ways of determining
experimentally whether something is acting as a initiator or
promoter; however, the lack of chemicals that have been
-52-
-------
identified as pure promoters and initiators makes this
difficult.
Human Data
The panelists agreed that obtaining data on promotion in
humans and identifying human promoters is a high priority.
Several specific suggestions were made for obtaining data on
promotion in humans. These are described in Section 11,
Species Difference/Human Studies.
Animal Models
The panelists agreed that short-term and long-term animal
models should be developed. A variety of animal models were
mentioned: (1) a fish model that produces a pathological
lesion that looks like a retinoblastoma when exposed to
chlorinated aliphatics; (2) a fish model that involves crossing
a molly and a swordtail that produces melanoma; and (3) a model
for Mendelian-inherited kidney adenoma in rats (Eker and
Mossige, 1961). Another model mentioned was Balmain's system
in mice whose epidermis has been infected with the Harvey
sarcoma virus. Treatment with a promoter induces papillomas, a
certain portion of which progress to carcinomas. However, in
this system, the presence of the whole virus may complicate the
results.
Species and Strain Differences
The panelists agreed that species and strain differences
were an important area for both in vitro and in vivo research
They discussed possible ways to investigate species
-53-
-------
differences. One panelist proposed that, since TCDD is a good
promoter in the rat liver, a model for TCDD using the human
liver should be developed so that TCDD action could be compared
in both systems.
In Vitro Systems
The panelists discussed the need to improve in vitro
screening models to detect promoters, and to develop in vitro
models for studying the mechanism of promotion. They agreed
this would be a difficult undertaking, not only to select the
most appropriate in vitro tests but also setting up parallel
studies in animals. The ability of in vitro data to predict in
vivo phenomena should be studied, and cell culture systems that
are best able to reproduce in vivo conditions should be given
priority for study. Serum may be a problem in in vitro testing
for'promoters; the action of many promoters depends on which
lot of serum is used, so serum lots must be carefully
screened.
Dr. Huberman stressed the importance of testing promoters
in in vitro cell transformation systems, especially in human
cell transformation systems. He recommended that the human
myeloid leukemia cell differentiation test be studied because
it is a relatively simple assay .that can detect changes in the
expression of various genes including a series of known
oncogenes. Dr. Slaga mentioned that no one has yet shown a
requirement for a promoter to get cell transformation in an in
vitro epithelial system.
-54-
-------
Mechanisms
Reversibility 4
Because of its implications for risk assessment, the
question of reversibility of promotion is an important area for
study. If promoters have thresholds, then presumably exposure
to pure promoters at levels below the threshold will not induce
tumors. One panelist raised the following question: If it can
be shown that promoters have thresholds, is it worthwhile
continuing to model IPI phenomena, since greater knowledge in
this area may not lead to different regulatory actions for
promoters. The sentiment was that modelling would be
worthwhile and efforts should continue in this direction. A
validated theory of initiation/promotion/progression would
enable quantification of potency at each stage.
One suggested area of research was for statisticians to
determine at what level of uncertainty experimental data can be
used to describe the behavior of the receptor as a basis for
estimating the threshold model. The EPA Carcinogen Assessment
Group is currently investigating this.
Receptor Binding
Two approaches to studying receptor mechanisms were
suggested: (1) in vitro or in vivo mutation studies and (2)
competitive binding studies. Dr. Trosko suggested looking for
a cell line that has an EGF-receptor mutation, in which the
receptor would bind, but the signal would not be transduced.
If the parent and mutant lines gave different responses in the
presence of a promoter, that would imply a receptor mechanism.
He mentioned a TPA-resistant strain used by Yamasaki et al.
(1985) to investigate mechanisms and the role of cell-cell
-55-
-------
communication in transformation. This mutant strain is not
promoted by TPA, whereas TPA does promote transformation in the
parental line. This is one example where a mutant for a
receptor-mediated response of TPA in in vitro promotion could
be used as a model.
Dr. Pitot suggested-locking the receptor up with an
irreversible inhibitor. He said that data suggest that TCDD
acts as a promoter through a genetic receptor mechanism
(although the fact that removal of the thyroid decreases TCDD
toxicity may cast some doubt on this theory).
Cell Differentiation
Modulation of cell differentiation studies are important,
Certain promoters like phorbol esters and TCDD are extremely
effective in modulating differentiation in some cell systems.
Oncogenes
The correlation of oncogenes with tumor promotion should be
researched. Cells in which oncogenes are activated may be more
sensitive to tumor promoters, e.g., £-Ha-_ras can initiate in
mouse skin (Balmain).
Phorbol Esters
Studies should be done to find out why phorbol esters
promote in some mouse strains but not others. If this
mechanism can be elucidated then much of the background data on
TPA may be applicable to other systems.
-56-
-------
Models
The Two-Stage Birth-Death-Mutation Model
There was support for conducting studies to validate this
model. Several research suggestions were made. These are
described in Section 7.
Intercellular Communication
The role of intercellular communication in linking tumor
promoters to a mechanism of action in vivo should be
investigated. Many technologies are available to measure
cell-cell communication and its role in growth control and
differentiation. Mutants are becoming available for gap
junctions, antibodies exist for gap junctions, and the gene for
the gap junction has been cloned. Genetic, molecular and
cellular experiments with normal cells are needed, both in
vitro and in vivo.
The
Gap junction technology may offer an opportunity to
investigate mechanisms of promotion at the cellular level.
ability to apply gap junction technology to liver and skin
could be investigated. The scrape loading assay could be
adapted to use cells (such as primary human keratinocytes) that
metabolize agents in order to study the role of metabolites in
promotion.
Areas for future research include the question of how big a
clone must be before it is free of inhibition by surrounding
normal cells via cell-cell communication. This maiy depend on
the type of cell since not all cells have the same number or
size of gap junctions. Identification of the
diffusion-suppressing molecule(s) is another possible research
•57-
-------
area. Another research area is to perform the scrape technique
using whole tissue.
Other Models
The prooxidative model (Cerutti, 1985) should be tested
versus the PKC model (Nishizuka, 1986). Data suggest that
these two models may not be mutually exclusive, i.e., one may
affect the other. These two models should be correlated with
the cellular and genetic models that have been proposed, e.g.,
the cell-cell communication model and the recombination model.
A method of testing the cell-cell communication model was
proposed: Correlate (1) sustained hyperplasia after TPA
treatment with the total absence of gap junctions, and (2)
nonsustained hyperplasia in the Syrian hamster with the
presence of gap junctions.
Quantification
It appears that the liver can be used to quantitate
initiation, promotion and progression. The ability to
quantitate these stages in other organs should be developed.
Promotion could probably be quantitated in the bladder, skin,
colon. It may be difficult to quantitate initiation in the
skin unless ways to detect initiated cells can be developed.
Expansion of the NTP Bioassay
One panelist proposed the idea of expanding the NTP
bioassay to study the ability of agents to promote. In
addition to the standard protocol for carcinogenicity, each
agent would be administered to a group of animals that had been
-58-
-------
previously exposed to a universal initiator. The expanded
protocol could also include studies in which treatments with
the agent would be stopped to determine the ability of tumors
to regress.
Additional Research Recommendations
The panel made no specific research recommendations but
mentioned several other potential areas for research. These
include:
• Distribution, metabolism and pharmacokinetics of tumor
promoters in various systems.
• The influence of the sequence of administration of
agents on the action of promoters.
• Synergism among promoters.
• Spontaneous initiation and promotion.
• The relationship between cytotoxicity and promotion.
• conversion of benign tumors to malignant tumors.
• Progressors and progression.
-59-
-------
14. REFERENCES
Aylsworth, C P., J.E. Trosko, C.C. Chang, and K. Benjamin.
Submitted. Synergistic inhibition of metabolic cooperation
by oleic acid on TPA and DDT in Chinese hamster cells-
implication of a role for protein kinase C in the
regulation of gap junctional intercellular communication.
Bell, G.I. 1976. Models of carcinogenesis as an escape from
mitotic inhibitors. Science 192:569-572.
Berenblum, I. 1974. Carcinogenesis as a Biological Problem.
North Holland Publishing Company, Amsterdam, pp. 133-141.
Blumberg, P.M. 1980, 1981. in vitro studies on the mode of
action of the phorbol esters, potent tumor promoters-
Parts 1 and 2. CRC Critical Rev. Toxicol., 8:153-197,
J. .7 i? """
Borek, C., D.L. Guernsey, A. Ong, and I.S. Edelman. 1983
Critical role played by thyroid hormone in induction of
neoplastic transformation by chemical carcinogenesis in
tissue culture. Proc. Natl. Acad. Sci. 5:749-752.
Cerutti, P. A. 1985. Prooxidant states and tumor promotion.
Science 227:375-381.
Clemmesen. 1977. Statistical studies in the aetiology of
malignant neoplasms. Acta Pathologica et Microbiologica
Scandinavica, Supplement 261, pp. 38-50. - -
Clifton, K.H., K. Kamiya, R.T. Mulcahy, and M.N. Gould.
1984. Radiogenic neoplasia in the thyroid and mammary
clonogens: progress, problems and possibilities. in-
Symposium Proceedings, Estimation of Risk from Low D^ses of
Radiation and Chemicals: A Critical Overview. Brookhaven
National Laboratory, May 20--23.
Committee on Biological Effects of Ionizing Radiation. 1972
Committee Report I, The Effects on Populations of Exposure
to Low Levels of Ionizing Radiation. National Academy
Press, Washington, DC, p. 154.
Committee on Biological Effects of Ionizing Radiation. 1980
Committee Report III. The Effects on Populations of
Exposure to Low Levels of Ionizing Radiation. National
Academy Press, Washington, DC.
-60-
-------
Doniach, I. 1974. Brit, j. Cancer 30:487-495.
Eker, R. and J. Mossige. 1961. A dominant gene for renal
adenomas in the rat. Nature (London) 189:858-859.
El-Fouly, M.H., J.E. Trosko, and C.C. Chang. 1987. Scrape
loading and dye transfer: a rapid and simple technique to
study gap junction intercellular communication. Exp. Cell
Res. 168:422-430.
Ethier, S.P. and R.L. Ullrich. 1982. Detection of ductal
dysplasia in mammary outgrowths derived from
carcinogen-treated virgin female BALB/c mice. Cane. Res.
42:1753-1760.
Federal Register. 1986. Gudelines for Carcinogenic Risk
Assessment. Federal Register 51(185);33992-34003.
Fernandez", A., S. Mondel and C. Heidelberger. 1980.
Probabilistic view of the transformation of cultured
C3H/10T-1/2 mouse embryo fibroblasts by
3-methylcholanthrene. Proc. Natl. Acad. Sci. 77:7272-7276.
Fisher, P.B., D.L. Guernsey, I.B. Weinstein, and i.S. Edelman.
1983. Modulation of adenovirus transformation by thyroid
hormone. Proc. Natl. Acad. Sci. 80:196-200.
Foster, R.S., Jr. 1975. Amer. J. Surg. 130:608-611.
Foulds, L.J. 1975. Neoplastic Development, Volume 2 (p. 729).
Academic Press, London.
purstenberger, G., B. Sorg, and F. Marks. 1983. Science
220:89-91.
Furstenberger, G., V. Kinzel, M. Schwarz, and F. Marks. 1985.
Science 230:76-78.
Gottlieb, L.S. and L.A. Husen. 1982. Che^t 81:449-452.
Gould, M.N. 1984. Radiation initiation of carcinogenesis in
vivo- a rare or common cellular event. In; J.D. Boice,
jr. and J.F. Fraumeni, Jr., eds. Radiation Carcinogenesis:
Epidemiology and Biological Significance. Raven Press, New
York, pp. 347-358.
Guernsey, D.L., A. Ong, and C. Borek. 1980. Thyroid hormone
modulation of X-ray induced in vitro neoplastic
transformation. Nature 288:591-592.
Hall, W.H. 1948. Brit. J. Cancer 2;273-280.
-61-
-------
« H: ^ a1' 1!83' Malignant conversion of mouse skin
tumors is increased by tumor initiators and unaffected by
tumor promoters. Nature 304:67-69.
Hennings, H. et al. 1985. Induction of papillomas with a
high probability of conversion to malignancy.
Carcinogenesis 6:1607-1610.
Hennings, H. et al. 1986. Malignant conversion and metastasis
of mouse skin tumors: a comparison of SENCAR and CD-I
mice. Environ. Health Perspec. 68:69-74.
Hennings, H. et al. 1987a. Response of carcinogen-altered
mouse epidermal cells to phorbol ester tumor promoters and
calcium. J. Invest. Dermatol. 88:60-65.
Hennings, H., -R. Shores, M. Balaschak, and S.H. Yuspa. 1987b.
Papillomas with a high frequency of spontaneous conversion
to carcinomas are also sensitive to malignant conversion by
4-nitroquinoline-N-oxide or urethane. Proc. Amer. Assn.
HO L C* ci I\C 61 C ~~ ~"
Hennings, H. and S.H. Yuspa. 1985. Two-stage tumor promotion
in mouse skin: an alternative explanation, j. Natl
Cancer Inst. 74:735-740. -^Ji^ii
Herschman, H.R. and D.w. Brankow. 1986. Ultraviolet radiation
transforms C3H10T-1/2 cells to a unique suppressable
phenotype. Science 234:1385-1388.
Kennedy, A.R. 1985a. Evidence that the first step leading to
carcinogen-induced malignant transformation is a
high-frequency, common event. in_: j.c. Barrett and R w
Tennant, eds. Carcinogenesis, A Comprehensive Survey, Vol
9: Mammalian Cell Transformation (Mechanisms of
Carcinogenesis and Assays for Carcinogens) Raven Press, New
xotK, pp. Jbo— 364,
Kennedy, A.R. 1985b. Relevance of tumor promotion to
oa^^n?fenesis in human Populations. Carcinogenesis
8:431-435. - •* -
Klein, G. and E. Klein. 1985. Evolution of tumours and the
impact of molecular oncology. Nature; 315:190-195.
Kraemer, K.H. 1980. Oculo-cutaneous and internal neoplasms
in xeroderma pigmentosum: implications for theories of
Carcinogenesis. in: B. Pullman, P.O. P. T'so, and H.
Gelboin, eds. Carcinogenesis: Fundamental Mechanisms and
Environmental Effects. D. Reidel Pub. Company, pp. 503-507
-62-
-------
Little, J.B., R.B. McGandy, and A.R. Kennedy. 1978. Cance£
Res. 38:1929-1935.
Little, J.B. and A.R. Kennedy. 1982. In.: E. Hecker, N.E.
Fusenig, W. Kunz, F. Marks, and H.W. Thielmann,_eds.
Carcinogenesis, Vol. 7: Cocarcinogenesis and Biological
Effects of Tumor Promoters. Raven Press, New York, pp.
243-257.
Lundin, F.E., J.K. Wagoner, and V.E. Archer. 1971. Radon
Daughter Exposure and Respiratory Cancer: Quantitative and
Temporal Aspects. NIOSH and NIEHS Joint Monograph No. 1.
National Technical Information Service. Springfield, VA.
Maxon, H.R., S.R. Thomas, E.L. Saenger, C.R. Buncher, and J.G.
Kereiakes. 1977. Amer. J. Med. 53:967-978.
McFadden, D., J.L. Wright, B. Wiggs et al. 1986. Bracking
inhibits asbestos clearance. Am. Rev. Resp. Disease
133(3):372-374.
Moolgavkar, S.H. and D.J. Venzon. 1979. Two-event models for
carcinogenesis: Incidence curves for childhood and adult
tumors. Mathematical Biosciences 47:55-77.
Moolgavkar, S.H. and A.G. Knudson, Jr. 1981. Mutation and
cancer: A model for human carcinogenesis. J. Natl. cane.
inst. 66:1037-1052.
Moolgavkar, S.J. 1986. Carcinogenesis modelling: From
molecular biology to epidemiology. Ann. Rev. Publ. Health
7:151-169.
Mossman, B.T. and J.E. Craighead. 1978, induction of
neoplasms in hamster tracheal grafts with
3-methylcholanthrene-coated Licra Fibres. Cane. Res.
38:3717-3722.
Mulcahy, R.T., M.N. Gould, and K.H. Clifton. 1984. Int. J.
Radiat. Biol. 45:419-426.
Nicolson, G.L. 1987. Tumor cell instability, diversification,
and progression to the metastatic phenotype: From oncogene
to oncofetal expression. Cane. Res. 47:1473-1487.
Nishizuka, Y. 1986. Studies and perspectives of protein
kinase C. Science 233:305-312.
Novell, P.C. 1986. Mechanisms of tumor progression. Cane.
Res. 46:2203-2207.
-63-
-------
n - -A new Protoc°l and its rationale for the
study of initiation and promotion of carcinogenesis in rat
liver. Carcinogenesis 2:1375-1379.
Quintanilla, M., K. Brown, M. Ramsden, and A. Balmain. 1986
Carcinogen-specific mutation and amplification of Ha-ras
during mouse skin carcinogenesis. Nature 322:78. --
Rivedal, E., T. Sanner, T. Enomoto, and H. Yamasaki. 1985.
iS i £°? °^ interceHular communication and enhancement
of morphological transformation of Syrian hamster embryo
cells by TPA: Use of TPA-sensitive and TPA-resistant cell
lines. Carcinogenesis 6:899-902.
Ron, E. and B. Modan. 1982. in: Radiation Carcinogenesis:
Epidemiology and Biological Significance, edited by j D
Boice and J.F. Fraumeni. Raven Press, New York.
Roop, D., D.R. Lowy, P.E. Tambourin, j. Strickland, J.R.
?o«?er/ M* Palaschak, E.F. Spangler, and S.H. Yuspa.
J.y86. An activated Harvey ras oncogene produces
benign tumours on mouse epidermal tissue. Nature
— 824. — — — _
Samet, J.M., D.M. Kutvirt, R.J. Waxweiler, and c.R. Key. 1984
New Engl. j. Med. 310:1481-1484.
Scherer, E., A.W. Feringa, and P. Emmelot. 1984. Initiation-
promotion-initiation. induction of neoplastic foci within
islands of precancerous liver cells in the rat. in- M
Borzsonyi, K. Lapis, N.E. Day and H. Yamasaki, edFT*
Models, Mechanisms and Etiology of Tumour Promotion
international Agency for Research on Cancer (IARC)
Scientific Publication No. 56. Lyons, France, pp. 57-66.
" L'A* Tnibodeau, A.R. Kennedy, and J.B. Little.
Cancer Res. 42:1405-1411.
Spray, D.C., and M.V.L. Bennett. 1985. Physiology and
°f gap ^unctions. Ann. Rev. Physiol.
Stenback, F , R. Peto, and P. Shubik. 1981. Br. J. Cancer
44:24—34. - — — -
Suss, R., v. Kinzel, and J.D. Scribner. 1973. Cancer
Experiments and Concepts. Springer-Verlag, New York.
Nettesheim- 1979. Dynamics of neoplastic
tracheal ""•'cosa. Cane.
-64-
-------
Troll, W. 1976. in: P.N. Magee, S. Takayama, T. Sugimura,
and T. Matsushima, eds. Fundamentals in Cancer
Prevention. University Park Press, Baltimore, Maryland,
pp. 41-55.
Trosko, J.E., C.C. Chang, and A. Medcalf. 1983. Mechanisms
of tumor promotion: Potential role of intercellular
communication. Cancer Invest. 1:511-522.
Trosko, J.E., C.C. Chang, and B.V. Madhukar. In press.
Chemical and oncogene modulation of intercellular
communication in tumor promotion. In: .H.A. Milman and E.
Elmore, eds. The Importance of Biochemical Mechanisms of
Cell-Cell communication in Toxicology. Princeton
Scientific Publish, Co., Princeton, NJ.
U.N. Report, united States Scientific committee on the Effects
of Atomic Radiation. 1977. Sources and Effects of.
ionizing Radiation. Report to the General Assembly, with
annexes.
Upton, A.C., D.G. Clayson, J.D. Jansen, H.S. Rosenkranz, and
G.M. Williams. 1984. Report of the international
Commission for Protection Against Environmental Mutagens
and carcinogens. Task Group on the Differences Between
Genotoxic and Nongenotoxic Carcinogens. Mut. Res. uj:±-
Verma, A. and R.K. Boutwell. 1980. Effects of dose and
duration with the tumor-promoting agent,
12-0-tetradecanoylphorbol-13-acetate on mouse skin
carcinogenesis. Carcinogenesis 1:271-276.
Weinstein, I.B., H. Yamasaki, M. Wigler, I. Lih-Syng, P.B.
Fisher, A. Jeffrey, and D. Grunberger, 1979. In. A.C.
Griffin and C.A. Shaw, eds. carcinogens: Identification
and Mechanisms of Action. Raven Press, New York, pp.
399-418.
Williams, G.M. and J.H. Weisburger. 1986. Chemical
carcinogens. In: C.D. Klausen, M.O. Amdur, J. Doull,
eds. Toxicology, 3rd edition. Macmillan publishing Co.,
New York, pp. 99-173.
Wiseman, R.W., S.J. Stowers, E.G. Miller, M.W. Anderson, and
j.A. Miller. 1986. Activating mutations of the c-Ha-r&s
protooncogene in chemically induced hepatomas of the male
B6C FI mouse. Proc. Natl. Acad. Sci. 83:5825-5829.
Witschi and Lock. 1978. Mechanisms of tumor promotion and
cocarcinogenesis. In: T.J. Slaga, A. Swak, and R.K.
Boutwell, eds. carcinogenesis: A comprehensive Survey,
Volume II. Raven Press.
-65-
-------
Wogan et al. n.d. Article concerning activation of the
' published in Proc.'Natl. Acad. sci. in
Diver9ent responses in epidermal
« tumor Prom°ter 12-0-tetradecanoyl-
phorbol-13-acetate. Cancer Res. 42:2344-2349.
;' ®fc ^; 1986* Cultivation and characterization of
«< d?rived fr01? mouse skin papillomas induced by an
initiation promotion protocol. Carcinogenesis 7:949-958.
-66-
-------
APPENDIX A
LIST OF PANELISTS AND OBSERVERS
-------
-------
FIRST EPA WORKSHOP ON RESEARCH PLANNING FOR RISK ASSESSMENT
Research Planning for the Development
of Risk Assessment Methodologies for
Tumor Promoters
r.TST OF PANELISTS
Roy E. Albert
Chairman and Director
Department of Environmental
Health
University of Cincinnati
Medical Center
3223 Eden Avenue
Cincinnati, OH 45221
513-872-5701
Eula Bingham
Office of Vice President
University Dean for Graduate
Studies and Research
University of Cincinnati
Cincinnati, OH 45221
513-475-4532
Henry Hennings
National Cancer Institute
Building 37, Room 3B26
Bethesda, MD 20892
301-496-3248
Freddy Homburger
Bio-Research institute, Inc.
380 Green Street
Cambridge, MA 02139
617-864-8735
Eliezer Huberman
Argonne National Laboratory
9700 South.Cass Avenue
Argonne, IL 60439
312-972-2000
Anne Kennedy
School of Public Health
Harvard University
Department of Cancer Biology
665 Huntington Avenue
Boston, MA 02115
617-732-1184
Daniel Krewski
Health and Welfare Canada
Room 115
Environmental Health Center
Tunney's Pasture
Ottawa, Ontario K2A 062
Canada
613-954-0164
Robert Langenbach
NIEHS
Cellular and Genetic
Toxicology Division
. Building 101 (Mail Drop
E4-05)
P.O. BOX 12233
Research Triangle Park, >NC
27709
919-541-7558
Peter Magee
Pels Research institute
Temple University School of
Medicine
3420 N. Broad Street
Philadelphia, PA 19140 ,
215-221-4311
Henry Pitot
McArdle Laboratory for
Cancer Research
University of Wisconsin
Medical School
Madison, WI 53706
608-262-3247
A-l
-------
I
Herbert s. Rosenkranz
Professor and Chairman
Department of Environmental
Health Sciences
Case Western Reserve
University School of Medicine
Cleveland, OH 44106
216-368-5961
Thomas Slaga
UTSCC-Science Park
P.O. Box 389
Park Road 1C
Smithville, TX 78957
512-237-2403
James E. Trosko
B236B Life Science Building
Department of Pediatrics and
Human Development
Michigan State University
East Lansing, Ml 48824
517-353-6346
A-2
-------
FIRST EPA WORKSHOP ON RESEARCH PLANNING FOR RISK ASSESSMENT
Research Planning for the Development
of Risk Assessment Methodologies for
Tumor Promoters
LIST OF OBSERVERS
Karl Baetcke
Office of Toxic Substances
U.S. Environmental
protection Agency
401 M Street, S.W.
Washington, DC 20460
Herbert Blumenthal
Division of Toxicology
CFSAN
FDA (HFF 150)
200 C Street, S.W.
Washington, DC 20204
Gary Burin
TS-769C
Office of Toxic Substances
U.S. Environmental
Protection Agency
401 M Street, S.W.
Washington, DC 20460
Chao Chen
Office of Health and
Environmental Assessment
RD-689
U.S. Environmental
Protection Agency
401 M Street, S.W.
Washington, DC 20460
Margaret Chu
Office of Health and
Environmental Assessment
RD-689
U.S. Environmental
protection Agency
401 M Street, S.W.
Washington, DC 20460
Ila Cote
OAQPS (MD-12)
U.S. Environmental
Protection Agency
Research Triangle Park, NC
27711
Bernard Daniels
Health Effects Research
Laboratory
U.S. Environmental
Protection Agency (Room 635)
26 West St. Clair Street
Cincinnati, OH 45268
William H. Farland
Director, Carcinogen
Assessment Group, ORD
U.S. Environmental
Protection Agency
Washington, DC 20460
Herman Gibb
U.S. Environmental
Protection Agency
Carcinogen Assessment Group
Washington, DC 20460
Sara Henry
Division of Toxicology
CFSAN
FDA (HFF 160)
200 C Street,- S.W.
Washington, DC 2,0204
Charlie Hireman
U.S. Environmental
Protection Agency
Carcinogen Assessment Group
Washington, DC 20460
James Holder
Office of Health and
Environmental Assessment
RD-689
U.S. Environmental
Protection Agency
401 M Street, S.W.
Washington, DC 20460
A-3
-------
Donald Hughes
American Industrial Health
Council
1330 Constitution Avenue.
N.W.
Washington, DC 20036
Robin Killman
Oak Ridge National Laboratory
Health and Safety Research
Division
Office of Risk Analysis
Building 4500 South,
Mail Stop 109
P.O. Box X
Oak Ridge, TN 37831
Lark Lambert
Division of Toxicology
PDA (HFP 164)
200 C Street, S.W.
Washington, DC 20204
Chiu S. Lin
Division of Toxicology
CFSAN
FDA (HFF 160)
200 C Street, S.W.
Washington, DC 20204
Ronald Lorentzen
Division of Toxicology
CFSAN
FDA (HFF 100)
200 C Street, 'S.W.
Washington, DC 20204
James McDermott
Proctor and Gamble Company
Miami Valley Laboratories
P.O. Box 39175
Cincinnati, OH 45247
Robert G. McGaughy
U.S. EPA
OHEA
401 M Street, S.W.
Washington, DC 20460
Bruce Means
TS-796
Office of Toxic Substances
U.S. EPA
401 M Street, s.w.
Washington, DC 20460
Philip Merker
Richardson Vicks
One Far Mill Road
Shelton, CN
Edwin A. Miraud
Rosewell park Memorial
Institute
Buffalo, NY 14163
Tim Mohin
U.S. Environmental
Protection Agency
(MD-12)
Research Triangle Park, NC
27711
Stephen Nesnow
Health Effects Research
Laboratory (MD-68)
U.S. Environmental
Protection Agency
Research Triangle Park, NC
27711
Carol Scott
CCEHRP
Room 14-101 Parklawn
5600 Fishers Lane
Rockville, MD 20857
Hugh Spitzer
Office of Regulatory Support
U.S. EPA (RD-672)
401 M Street, s.w.
Washington, DC 20460
Catherine St. Hilaire
Risk Science Institute
Suite 111
1126 16th Street, N.W.
Washington, DC 20036
Curtis Travis
Oak Ridge National Laboratory
Health and Safety Research
Division
Office of Risk Analysis
Building 4500 South,
Mail Stop 109
P.O. Box X
Oak Ridge, TN 37831
A-4
-------
APPENDIX B
AGENDA
-------
-------
FIRST EPA WORKSHOP OH RESEARCH PLANNING FOR RISK ASSESSMENT
Research Planning for the Development
of Risk Assessment Methodologies for
Tumor Promoters
AGENDA
Tuesday, February 3, 1987
8:00-8:30 a.m. Registration
Chairmen:
8:30 a.m.
8:40 a.m.
8:55 a.m.
9:15 a.m.
10:30 a.m.
10:45 a.m.
12:15 p.m.
1:45 p.m.
2:30 p.m.
3:30 p.m.
Roy Albert
University of Cincinnati Medical Center,
Cincinnati, OH
Hugh Spitzer
Office of Research and Development, U.S.
EPA, Washington, DC
Announcements
Welcoming Remarks
Vaun Newill, Assistant Administrator for
Research and Development, U.S. EPA,
Washington, DC
Introduction and Overview
Roy Albert and Hugh Spitzer
Panelists' Presentations of Pre-meeting Comment
Summaries
COFFEE BREAK
Panelists' Presentations of Pre-meeting Comment
Summaries (cont.)
LUNCH BREAK
Panelists' Discussion of Areas of Consensus or
Lack Thereof
1. Testable Hypotheses - Biological
a. Mechanisms of Action
i. What we know now
ii. What we need to know
COFFEE BREAK
B-l
-------
Tuesday, February 3, 1987 (cont.)
3:45 p.m.
5:15 p.m.
5:30 p.m.
6:00 p.m.
2. Testable Hypotheses - Modeling
a. Possible approaches for integrating
promoter activity into risk assessment
i. Mechanisms of action
^ii. Biological half-life
iii. Considerations of species
differences
Closing Discussion
Adjourn
Workshop Dinner
*****
Wednesday/ February 4, 1987
Chairman:
8:30 a.m.
9:00 a.m.
10:30 a.m.
10:45 a.m.
12:00 p.m.
1:30 p.m.
3:30 p.m.
Robert Langenbach
NIEHS, Cellular and Genetic Toxicology
Division, Research Triangle Park, NC
Announcements and Overview of Preceding Day's
Issues
The Biology of Tumor Promotion
1. Discussion of possible appropriate
chemicals that can serve as surrogates for
classes' of chemicals
2. Need for developing new methodologies
3. Short- and long-term research proposals
COFFEE BREAK
The Biology of Tumor Promotion (cont.)
LUNCH BREAK
Modeling
1. Can we integrate promotional activity into
current risk assessment methodologies
a. what assumptions are required
b. identify uncertainties encountered
COFFEE BREAK
B-2
-------
Wednesday, February 4, 1987 (cont.)
3:45 p.m.
5:30 p.m.
2. Alternative models/approaches
a. Data base required for testing
b. Biological considerations
c. What assumptions are required
d. Identify uncertainties encountered
Adjourn
* * * * *
Thursday, February 5, 1987
Chairman:
8:30 a.m.
9:30 a.m.
10:30 a.m.
10:45 a.m.
12:15 p.m.
12:30 p.m.
William Farland
Director, Carcinogen Assessment Group* U.S,
EPA, Washington, DC
Summary of Discussions
Roy Albert and Robert Langenbach
Setting Research Priorities for Planning Risk
Assessment Methodologies
1. Biological research
a. Long-term
b. Short-term
2. Model development
a. Long-term
b. Short-term
COFFEE BREAK
Setting Research Priorites (cont.)
Closing Remarks
Adjourn
0887Y
B-3
-------
-------
APPENDIX C
PANELIST PREMEETING COMMENTS
-------
-------
United States Environmental Protection Agency
Office of Regulatory Support
Office of Research and Development
First EPA Workshop on Research
Planning for Risk Assessment
PRE-MEETING COMMENTS
FOR THE WORKSHOP ON THE
DEVELOPMENT OF RISK
ASSESSMENT METHODOLOGIES
FOR TUMOR PROMOTORS
February 3-5, 1987
Bethesda Hyatt Regency
Bethesda, MD
-------
PRE-MEETING COMMENTS FOR WORKSHOP
ON THE DEVELOPMENT OF RISK ASSESSMENT
METHODOLOGIES FOR TUMOR PROMOTORS
February 3-5, 1987
TABLE OF CONTENTS
Page
List of Pre-Meeting Questions ]_
Eula Bingham ^ ^ 3
Henry Hennings 9
Fred Homburger ^
Eliezer Huberman 19
Anne Kennedy 21
Daniel Krewski 27
Robert Langenbach 33
Peter Magee 3 5
Henry Pitot 41
Herbert Rosenkranz 45
Thomas Slaga 4 57
James E. Trosko,
61
0912Y
-------
PANELISTS' PRE-MEETING ASSIGNMENT
Topics for Pre-meeting Comments
In an effort to stimulate discussion and to develop
consensus or_ a spectrum of views on key topics, panelists are
requested to bring to the workshop written statements on each
of the following topics:
1.
2.
3.
4.
5.
6.
How do you define tumor promotion? What is the basis
for your definition? What are the limitations of the
definition?
Is it possible to quantitatively separate the
promotional activity from the initiating activity of a
chemical in assessing the carcinogen risk using the
available data set?
How would you approach incorporating the qualitative
characterization of promotional activity into a risk
assessment?
What data would you require before accepting the
conclusion that a chemical is only a promotor?
What chemicals, do you think, would be good candidates
for use in developing risk assessment methodologies
for tumor promoters?
What additional research, do you think, is required on
these chemicals to better understand their interaction
with biological systems?
What generic considerations should be given to
interspecies extrapolation of the risk associated with
exposure to promoters (body wt. vs. surface area;
liquid soluble vs. water soluble)?
What chemical specific considerations should be given
in interspecies extrapolation of the risk associated
with exposure to promoters?
0841Y
C-l
-------
-------
RISK ESTIMATION - TUMOR PROMOTION
Question 1.
Promotion is a term that has been defined by the design of carcinogenesis
experiments. Originally the term co-carcinogenesis was used and later the
specific experimental regimen delinated the term promoter. Tumor promotion is
one of a series of steps whereby normal somatic cells become neoplasic lesions.
The carcinogenic process begins with the initiation step in which an agent
(initiator) damages a critical cellular target, presumably DNA. Although
necessary, the initiation process is not sufficient for neoplastic
transformation of a cell. A promotion step is required in which exposure to a
second agent (a promoter) results in cellular changes that commit initiated
cells to a progressive process that ultimately results in full neoplasitic
expression ~ .
The definitions of initiation and promotion are based on studies of
sequential exposures to agents that produce tumors. Early experiments involved
a single dermal application of agents such as 7,12-dimethylbenz[a] anthrecene
(DMBA) that did not produce tumors in the population of exposed mice or only an
occasional tumor. It was observed that repeated topical application of a
second agent such as croton oil, which does not produce tumors alone , caused
a high incidence of malignant tumors. Initiation and promotion are defined by a
temporal sequence of events that is necessary for tumor production: exposure to
an initiator followed by exposure to a promoter. At high doses or repeated
exposures, some initiating agents such DMBA act both as initiators and
promoters and are therefore called complete carcinogens.
* Promoters may induce a very low incidence of tumors but this is usually
attributed mechanistically to previous undetected initiation (background).
C-3
-------
In addition to the temporal relationship of exposures necessary to produce
tumors, initiators and promoters are characterized by the temporal nature of
their effects. Initiating events are believed to occur in a relative short
period of time (minutes to hours?). The effects of initiators are persistant,
perhaps irreversible, and cumulative at all doses and frequencies of exposure.
The promotion phase of carcinogenesis occurs over a longer period of time
(weeks to years?). At low doses or low frequencies of exposure the effects of
certain promoters appear to be reversible and non-cumulative. Sustained
exposures to sufficient doses of promoters are necessary for initiated cells to
"evolve" into neoplastic tissue.
A major limitation of the above definition of tumor promotion is that it
is an operational definition based on experimental protocols used to produce
tumors in amnimals. The biochemical mechanisms underlying tumor promotion are
far from clear and different promoters may have vastly different modes of
action.
C-4
-------
Question 2.*
It is difficult to separate quantitatively promotional arid initiating
activities of chemicals because most of the available data set does not
specifically deal with initiation and promotion as defined in answer one. It
has been generally assumed that initiation is a direct result of DMA damage.
Most data come from mutation assays or estimating adduct formation that are
supposed to reflect underlying DKA damage/binding. Agents that test positive
in these assays are assumed to be at least initiators. Agents that do not test
positive in these assays but induce tumors are often assumed to be promoters 5.
The relationship of these assays to the exact nature of underlying DNA
damage/binding and the relationship of that damage to carcinogenesis are far
from clear (see review by Perera 6 and references therein). Until a much
clearer picture of the biochemical/physical mechanisms of initiation and
promotion are defined, we will have to rely on classical bioassays to
differentiate initiators from promoters. In addition, classical bioassays
still give us the most relevant data to use in making our best guesses about
the relative risks of carcinogenic agents.
What available data set?
C-5
-------
Question 3.
Because of the necessity to repeatedly apply a promoter in early
experiments and the fact that stopping the exposures causes a less than 100%
incidence of tumors, a kind of reversibility is assumed and or a threshold
event (total accumulated dose) is considered to be operating. For this reason
there has been some discussion about regulating promoters differently from
initiators 5. There may be some justification for this but it must be
emphasized that, as pointed out above, the original definition of promoters was
based on the production of tumors in long-term bioassays under a specific set
of circumstances. However carcinogens are commonly classified as promoters
based upon their mechanisms of action and assumptions abou the biological
significance of in vitro assays. The present state of scientific knowledge
does not allow clear categorization of carcinogens based on their mechanisms of
action, in addition, the activity of a promoter in producing tumors depends on
the type and degree of initiation involved 7. Setting a no-effect level for a
promoter would require knowledge about qualitative and quantitative
relationships between initiators and promoters, and the normal level of
background initiation and ambient initiators. Even in the absence of obvious
initiators, promoters increase the background incidence of tumors in
experimental animals.
C-6
-------
Question 4.
To be just a promoter, a substance must have been tested experimentally
and fit the definition of a promoter given in answer one. Operationally a
chemical may be a promoter but are the mechanisms of action all the same? Data
still need to be developed.
Question 5.
Good candidates for studying promotion are chemicals such as arsenic which
has been demonstrated to be a carcinogen in human populations but is difficult
to demonstrate to be a carcinogen in animal studies. Particular attention
should be paid to substances where the potential human exposure is high.
Phorbol esters are excellent experimental promoters but how relevant are they
to the human situation? Another very intriguing compound that has been
repoorted to be a promoter is TCDD. It should be evaluated further as a
promoter to determine whether or not the experimental animal data and the human
epidimiological data fit.
What about asbestos and/or cigarette smoke?
Question 6.
It is difficult to make generalizations about interspecies differences in
dose-response relationships of promoters because very little research has been
done in this area 6. Not only is there a void in terms of species differences
but there is a paucity of experimental data for all except skin and perhaps
liver. The lung seems to be such an important area for experimental work based
upon clues from epidemiology but alas who supports such experimental research.
C-7
-------
References
1. Shear, M.J. (1938): Studies on carcinogenesis, V. Mehtyl-derivatives of
1,2-benzathracene, American Journal of Cancer, 33:499-537.
2. Berenblum, I. (1941): The cocarcinogenic action of croton resin. Cancer
Research, 1:44-48.
3. Berenblum, I. and Shubik, P. (1947): A new quantitative approach to the
study of the steps of chemical carcinogenesis in the mouse skin. British
Journal of Cancer, 1:703-708.
4. Boutwell, R.K. (1978): Biochemical mechanism of tumor promotion. Slaga,
T.J., A. Sivak, and R.K. Boutwell (eds). Carcinogenesis, Vol. 2.
Mechanisms of Tumor Promotion and Cocarcinogenesis. Raven Press, New
York, pp. 49-58.
5. Weisburger, J.H. and Williams, G.H. (1983): The distinct health risk
analyses required for genotoxic carcinogens and promoting agents.
Environmental Health Perspectives, 50:233-245. \
6. Perara, F.P. (1984): The genotoxic/epigenetic distinction: relevance to
cancer policy. Environmental Research, 34:175-191.
7.
Burns, F., Albert, R., Altshuler, B., and Morris, E. (1983): Approach to
risk assessment for genotoxic carcinogens based on data from the mouse
skin initiation-promotion model. Environmental Health Perspectives,
50:309-320, ;
C-8
-------
Comments for the EPA Workshop on Risk Assessment
Methodologies for Tumor Promoters~
Henry Hennings
Epidermal carcinogenesis, epidermal cell culture
National Cancer Institute, NIH
Building 37, Room 3B26
Bethesda, MD 20878
1.) Tumor promotion is defined based on the initiation-promotion protocol ,
developed in the mouse skin model system. After a single application of an
initiator (or an initiating dose of a complete carcinogen), which by itself
does not produce skin tumors, repeated applications of a promoting agent
are required to produce skin tumors. Most of these tumors are benign
papillomas, a small percentage of which may progress to carcinomas. Promoter
treatment by itself induces very few tumors.
The end point of promotion is the benign papilloma. Papillomas are
heterogeneous in their potential for progression to malignancy, which I believe
reflects the heterogeneity of initiated cells. Progression from the papilloma
stage proceeds in 2 further stages, which are distinct from promotion:
malignant conversion is the progression to malignancy (end point is a
squamous cell carcinoma); metastatic conversion is the progression of the
malignant tumor to metastasize (end point is a metastasizing squamous cell
carcinoma).
This definition is a starting point from which modifications can be
made for other tissues or for man. For a process to be called "promotion"
(so that mechanistic studies from skin, liver or other animal models are
relevant), the process must l)follow initiation, 2)require repeated exposure
and 3)by itself produce few (or no) tumors. Initiation could be the result
of repeated, as well as single, exposures, but by itself should produce few
tumors. A distinction should be made between promoters and co-carcinogens. -."
C-9
-------
2.) The relative initiating and promoting abilities of a chemical can be
assessed (Cancer Res. 43, 2034-2041, 1983) by testing at various dose levels
in a 2-stage model utilizing optimal doses of known initiator and promoter.
Few chemicals have been tested in this way.
3.) If a chemical is a pure promoter in an in_ vi_vo animal model (producing
benign tumors with little potential to progress to malignancy), the likely
risk to man would be small. The risk would be much greater for an agent
active in the malignant conversion stage or co-carcinogenic (by simultaneous
treatment) to produce malignant tumors.
4.) For a chemical to be only a promoter, it must be tested for activity
as an initiator, a promoter, a malignant converting agent, and a co-
carcinogen and be found negative for all stages other than promotion.
5.) Based on the mouse skin model, comparisons should be made between
Dphorbol esters and other promoters which apparently act through protein
kinase C, 2)benzoyl peroxide and related compounds such as hydrogen peroxide,
3)hydrocarbon derivatives such as bromomethylbenzanthracene, and 4)anthrones.
There are apparently differences between the mechanisms by which these agents.
act. They also have different potencies when tested for malignant conversion.
Shouldn't radiation be considered, or are we only interested in chemicals?
The biological effects of phorbol esters on various aspects of control of
epidermal proliferation and differentiation are being well-characterized.
Similar efforts should be undertaken on the other classes of promoters.
What is the role of free radicals in promotion? Do promoters all
necessarily act via a similar mechanism?
6.) Hasn't EPA already dealt with these questions from considerations of
carcinogen risk?
C-10
-------
comments for the EPA Workshop on Risk Assessment
Methodologies for Tumor Promotors
Freddy Homburger, M.D.
Experimental Pathology, Toxicology
Bio-Research Institute, Inc.
380 Green St.
Cambridge, MA 02139
1. Definition of Tumor Promotion:
The classical definition of "promoter" is as follows:
"The enhancement of the carcinogenicity of an agent by a
second agent not carcinogenic by itself under the test
conditions acting after exposure to the first has
ended." This classical definition, as conceived by Rous
and Berenblum, is quoted from Williams and Weisburger in
A Guide to General Toxicology (Homburger, Hayes, and
Pelikan, eds.), published by S. Karger, A.G. Basel/New
York, 1983, p. 220.
The two-stage concept of carcinogenicity, postulating
a first phase of "initiation" followed by a second stage
of "promotion," assumes the existence of an "initiated"
or "dormant" cancer cell which can be stimulated into
C-ll
-------
neoplastic growth by promoters.
The concept of "promoter" must be clearly separated
from that of "co-carcinogen," which is any substance
enhancing the carcinogenicity of a chemical simul-
taneously administered and not itself carcinogenic under
the test conditions.
Clearly separate from the two above is the concept of
"synergism" between carcinogens, i.e.,, the mixture of two
carcinogens may be more carcinogenic than each of the
carcinogens alone (J.p. Greenstein, The Biochemistry of
Cancer, Academic Press, 1954, p. 79). To the best of my
knowledge, these concepts and definitions are still
valid.
2' Separation of Promoting Activity from Initiating
Activity;
By definition, a promoter never has initiating
activity. The few tumors caused by some promoters in
mouse skin are assumed to be due to spontaneously
occurring latent tumor cells activated by the promoter
(Berenblum, Carcinoqenesis as a Biological Problem.
North-Holland Publishing Co., Amsterdam/Oxford, 1974).
c-12
-------
3. incorporation of qualitative characterization of
Promotional Activity into a Risk Assessment:
If this question means "how will the presence of a
promotor affect a risk assessment?" my answer would be "I
do not know." Each experimental data set would have to
be judged on its own merits.
4. Data Required to Decide that a Chemical Is only a
Promotor;
By definition, a chemical is either a promotor or an
initiator. It cannot be both.
5. candidate Chemicals for Study of Risk Assessment
Methodology for Promotors;
Most studies of the mechanism of promotion have been
done with croton oil, and later, with its active
ingredients, the phorbol esters, and particularly, with
Becker's compound Aj. (Cancer Res. 28:2338, 1968), or
TPA (12-o-teradecanoyl-phorbol-13-acetate). Most
recently, a number of diverse compounds have been
C-13
-------
described as promoters in different experimental settings
(Proceedings of the American Association for Cancer
Research, March 27, 1986), for example, orotic acid,
promoting liver tumors and intestinal carcinogenesis in
rats (Rao, et al, abs 561), polychlorinated biphenyls
promoting lung and liver tumors induced in infant mice by
N-nitrosodimethylamine (Anderson, et al, abs 560),
benzodiazepine tranquilizers promoting hepatocellular
neoplasms in mice (Diwan, et al, abs 559), long-acting
barbiturates promoting tumors in rat liver (Diwan, et al,
abs 558), cyclosporine promoting induction of thymic
lymphoma in mice by N-methyl-N-nitrosourea (Shihozuka, et
al, abs 546), aspirin promoting urinary bladder cancer
induced in rats (Sakata, et al, abs 490), peroxides as
promoters in the 2-stage mouse skin model (Rotstein, et
al, abs 567). These are merely cited as examples o.f the
most recent vintage. There are many other substances
claimed to be promoters, some of them of practical
importance, such as saccharine and cigarette smoke
suggested as promoting bladder cancer (in rats), and
asbestos inducing lung cancer (in humans), respectively.
Which of these many possible chemicals to use as models
for further study of promotion is debatable.
c-14
-------
The type of additional research retired to better
understand the interaction, of promotors with^biological,
systems requires a great deal of thought. I doubt that
the scientific basis exists for a strictly rational
formulation of hypotheses that could be verified by
experiments. In some cases, it is found that what may
have been considered a case of promotion turns out. to be
the result of entirely different mechanisms, for example,
the case of the interaction of asbestos inhalation and
smoking resulting in lung cancer. It appears that
cigarette smoking impedes asbestos clearance from the
lungs by increasing retention of short fibers (in Hartley
strain guinea pigs), the same mechanism which might
contribute to the rate of disease seen in asbestos
workers who smoke (McFadden, et al, Am. Rev. Resp. Dis.
131(3) -.372-374, 1986).
in light of recent advances in immunology, some
alleged promotors must be re-investigated to determine
whether their apparent promoting effect might not be
mediated through irnraunosuppression, which would release
dormant cancer cells from immunosurveillance. Cyclo-
sporin might be a paradigm for such situations.
C-15
-------
6* Generalizations on Extrapolation from One Species t
Another;
Extrapolation from one species to another is always a
risky task. It would be a worthwhile exercise in risk
assessment to quantify the risk involved in any such
procedure.
While it is reasonably safe to extrapolate from acute
toxicity tests in animals to humans (with some notable
exceptions, such as atropine, which is perfectly safe for
rabbits, but fatal to infants of comparable weight), the
problem is far more complex when it comes to the transfer
of information from animal carcinogenesis tests to the
human epidemiological situation. A listing of human
carcinogens detected by animal tests (Homburger, in
A Guide to Toxicology, s. Karger, A.G. Basel/New York,
1983, pp. 205-208) shows that with proper safeguards,
such tests have validity and are usable. However, this
is only true in a qualitative way. Expression of
numerical risks extrapolated from animal assays are, in
most cases, useless, and may be misleading.
In the case of promoters (as opposed to most known
initiators), the scientific basis for the planning of
C-16
-------
protocols for-animal testing is still weak, and muciv more
fundamental research is needed before experiments can be
designed that are sure to provide information useful in
regulatory decisions. For"lack of a thorough knowledge
of their mechanism of action we may see ourselves
compelled to adopt an arbitrary regulatory stance
somewhat akin to the Delaney clause, and regulate that
any substance active as a promotor of carcinogenesis in
animals must be assumed to be a potential promotor of
carcinogenesis for humans.
c-17
-------
Conclusion
I conclude with my statement at the end of the
chapter "Carcinogenesis - Concepts," in the above-quoted
book (p. 211):
"Whatever is known about chemical carcinogens
suggests that there are not only the generally recognized
classes of initiators, which may induce latent or overt
cancer cells, but a multitude of promoters of widely
varying chemical structures which may activate latent
tumor cells into an overtly cancerous state. Intensive
efforts to better understand the carcinogenic process are
therefore needed if regulation of carcinogens is ever to
be placed on a sound scientific basis."
C-18
-------
Premeeting comments
on risk assessment methodologies
for tumor formation
by E. Huberman
1. Tumor promotion is an operational term that characterizes one component of
the multistage and multifactorial process of cancer causation. The term
can be defined as the process by which (usually) protracted exposure of
animals (humans) to noncarcinogenic or weakly carcinogenic agents
(promoters) following and separated by time from the administration of a
nontumorigenie dose of a carcinogen (initiator) results in a tumor(s)
(individual) or in an increased tumor incidence (population).
This definition is limited because it is not based on a mechanistic
understanding of the process of tumor promotion. Furthermore, all
currently known tumor initiators are potent carcinogens, and a major
fraction of the promoters are capable of eliciting a tumorigenic response
by themselves.
2. Optimal experimental protocols that can discriminate unequivocally between
phenomena associated with initiation and promotion have not yet been fully
developed. Reliable quantitative estimations of relative initiating and
promoting activities of a given chemical are therefore yet not possible.
3.
4. To characterize a chemical as being a promoter only, it should have tumor-
promoting activity and, be unable itself to initiate tumorigenic activity.
C-19
-------
5. The chemicals to be used in developing risk assessment methodologies for
tumor promoters should be these that may alter tumor incidence in
humans. Until such agents are identified, one should use representatives
of currently known classes of tumor promoters. Additional research should
therefore be directed tow
-------
COMMENTS FOR EPA WORKSHOP ON RISK ASSESSMENT METHODOLOGIES FOR TUMOR PROMOTERS
Ann R. Kennedy, Experimental Carcinogenesis (Radiation Biology)
Harvard University, School of Public Health, 665 Huntmgton Avenue, Boston, MA 021.L 5
1. How do you define tumor promotion? What is the basis for your definition? What are the
limitations of the definition?
At this point, there are no generally agreed upon definitions of tumor promotion, as many of
the classical concepts have been challenged recently and are still quite controversial. In a recent
debate on this issue which was published (1), the Chemical Pathology NIH study section could only
agree that initiation starts off the carcinogenic process, while promotion finishes the process.
Even this definition is at present controversial, as data have now been presented to indicate that a
single dose of a promoting agent before initiation can be partially effective as a "promoter .
I define a compound as a promoter if it causes no, or a low level of, tumor formation (or
transformation foci in vitro etc.) by itself, but synergistically enhances the yield of tumors or foci
when given as repeated exposures after initiation has occurred. My definition would include, for
example, saline instillations to the lung as a promotional stimulus. In our studies on lung
carcinogenesis, we observed that small amounts of saline (given as seven weekly intratracneal
instillations beginning 5 months after the animals were exposed to a single dose of ^luPo) could
markedly enhance the incidence of lung cancer in hamsters exposed to low doses of alpha
radiation from 210po (210po is a radionuclide found in cigarette smoke and cigarette smokers
lungs: 210p0 aipha radiation is similar to that which arises from Radon gas, plutonium (a
byproduct of nuclear power), etc.(2,3). While saline is not a classical promoting agent such as
TPA, its effect appears to result in tumor promotion. Thus, many of the classical definitions
derived for promoting agents such as TPA are inappropriate for agents such as saline and yet,
tumor promotion can clearly be achieved with saline.
As my definition is for the process of tumor formation, it does not cover many of the
classical characteristics attributed to promoters, as summarized recently by Weinstem et al.W as
follows:
A Comparison of Biologic Properties of Initiating Agents & Promoting Agents
Initiating Agents
Promoting Agents
1. Carcinogenic by themselves -
"solitary carcinogens"
2. Must be given before promoting agent
3. Single exposure is sufficient
4. Action is irreversible and additive
5. No apparent threshold
6. Yield electrophiles that bind covalently
to cell macromolecules
7. Mutagenic
1. Not carcinogenic alone
2. Must be given after the initiating agent
3. Required prolonged exposure ^
f. Action is reversible (at early stage^)
and not additive
5. Probable threshold
6. No evidence of covalent binding
7. Not mutagenic
C-21
-------
Personally, I still believe that many of these characteristics for promoters are still reasonable
an«,?Uf- aS the re.yersible nature of changes induced (at early times), the necessity of repeated
applications, etc. The most questionable of these characteristics is the lack of mutasenicity of
promoting agents, as even classical promoting agents such as TPA have recently been shown to cause
chromosome-type mutations (reviewed by Marx (5)).
2. Is it possible to quantitatively separate the promotional activity from the initiating activity of
a chemical in assessing the carcinogen risk using the available data set?
• -v y the available data set include two-stage in vivo carcinogenesis experiments in animals, then
SSt ?£* XT0,!!0"™*? be distinquished relatively easily-as they have been in the extensive
hterature in this field. If the "data set" refers to human exposures to chemicals, I do not believe that it
will be possible to separate the initiating and promoting activity of the agents being studied. I do
believe that these activities could be separated with radiation as the initiating agent, however, as
described below. ,
3.
How would you approach incorporating the qualitative characterization of promotional activity
into a risk assessment?
One approach to determining whether in fact a chemical is a promoter in human populations is to
study whether specific agents have the characteristics of promoters in people. From many different
i«£?f«? — ^^ T1 ^^ and human epidemiologic studies, it is now clear that radiation can serve as an
initiating agent. Irradiated human populations would be very appropriate as "initiated" groups for
detailed studies, as there are now many people who have been exposed to radiation in discrete (and
•^ ,,n™WhlCh ^ have much information about. (Many of these populations are discussed in detail
in the BEIR" report (6) as well as several other documents). The individuals in such irradiated
populations could be questioned about exposures to agents we suspect as being promoting agents.
From previous studies of irradiated populations, the thyroid and the female breast have emerged
as the most sensitive tissues for the induction of radiation induced cancer (6). Approximately 20,000
cases of thyroid cancer (in the > 200,000 people irradiated) are expected to occur in this country in
people whose thyroid happened to be in the irradiated field when they were irradiated for the treatment
?„•%? e?S*.ine, disease/problem- (for examples, thymus enlargement, acne, eczema, etc.- were treated
with relatively high doses of x-irradiation in the 1940's and 1950's). As the form of thyroid cancer that
is induced by radiation has a very low mortality rate (~3%, although mortality estimates have varied
irom 1710%;, most patients who have or have had thyroid carcinoma will be alive to discuss the effects
ot possible promoting agents in their lives. In fact, there are already "risk" groups that have been
defined in this irradiated population. In human populations, it has been observed that the risk of
SwthPltLr-adiattn"ind^ thy!;0id CfCer iS considerably higher in: 1) females (6), 2) those having a
Jewish ethnic background (6), and 3) those who have emigrated from Morocco or Tunisia (7); these data
suggest that factors other than the radiation exposure play a very large role in the genesis of this
disease. It is thought that the sexual difference is "related to the flunctuating hormonal status in
temales, with significantly greater variations in the pituitary-thyroid axis and in secretion of thyroid-
stimulating hormone than in males" (6). The greater relative risk among the Jewish population could be
due to genetic susceptibility; however, all of the increased risk factors considered together suggest that
promotional factors, some of which may be present in the diet of "emigrating" populations, may be the
most important determinants of whether cancer will result from the random distribution of energy by
C-22
-------
In
ionizing radiation. Clearly, diet is one of the major determinants of the cancer incidence in human
populations (8), and many promotional factors are known to be present in the human diet. Many of
these irradiated patients in this population have already been "recalled" to determine whether they have
a thyroid cancer; thus, they could be an already assembled population available for questioning about
their exposures to potential promoting agents.
There are several other populations which have been irradiated which could be studied for possible
prompting effects. For example, people treated with 1*31 (5 mCi by mouth) for thyrotoxicosis receive
the following doses:
"A patient with a 30 g gland with 60% uptake and who is treated with 5mCi of 1*31 by mouth will
receive the following average tissue doses: 5 rads - whole body, 9000-10,000 rads -thyroid gland,
2-5 rads - hemopoietic tissue, approximately 1 rad to the testes and 2 rads to the ovary" (9).
(In vitro transformation studies suggest that even very low doses (10 rads) of x-radiation are capable
of Initiating cells (10); thus even the organs receiving the relatively low doses of radiation indicated
above are like to have initiated cells following the 1311 treatment). A study of these irradiated patients
could have "controls" built into the study, as some of the patients will have had part of their thyroids
removed surgically, some will be taking a thyroid replacement hormone (and some will not), etc.
Certain hormones are likely to be potent promoting agents for some types of human cancer (for
example, thyroid hormones have already been shown to have such promoting activity in vivo (11-13)).
the many populations of irradiated individuals, it is likely that many of the female patients will have
been exposed to sex hormones (other likely promoting agents (11, 14-16)) for birth control, to treat
symptoms of menopause, etc. Exposure to these hormones could easily affect cancer development in
irradiated organs such as the breast, uterus, ovaries etc. Specific hormone use has already been
associated with a promotion- like response in animals (17) and humans (18). Like the reversible actions
of tumor promoting agents in vivo and in vitro, many tumors in animals, such as mouse mammary
tumors, are under hormonal controls in that removal of a necessary hormone can cause tumor regression
(17). For human cancer, it has been reported that women who take estrogen as a therapy for
postmenopausal problems have a greatly increased risk of developing cancer and that there is a rapid
decline in risk following the discontinuation of estrogen use (18). Again, such a reversible effect is
what would be expected of a promoting agent operating in human carcinogenesis. Endogenous hormonal
promotion in the breast could easily account for the fact that the female breast is the most sensitive
human organ to the induction of cancer by ionizing radiation (6, 19).
Many of the solid cancers expected to develop in the atom bomb survivors should be detected over
the next decade (there are approximately 30,000 people remaining in the currently ongoing study (begun
by the Atomic Bomb Casualty Commission) to determine the cancer incidence occurring in individuals
irradiated when the bombs were dropped in Hiroshima and Nagasaki). These studies could be extended
to search for promoting factors present in the environments/diets etc. of the Hiroshima and Nagasaki
survivors. To determine whether a promotional response has been observed in these studies, the shapes
of the dose-response curves, with and without the suspected promoting agent, could be compared.
The presence of a promoting agent in radiation carcinogenesis often results in a linear curve, in
both in vivo (20) and in vitro (21) experimental studies, while the curve expected for radiation
treatmenTa'lone (in the systems cited above) is a quadratic or linear-quadratic curve. The conclusion
reached by the most recent report from the Committee on Biological Effects of Ionizing Radiation 16) is
that for low linear energy transfer radiation, the dose-response curve for most radiation-induced human
C-23
-------
cancer is best represented by a linear quadratic form. Radiation-induced breast cancer in females is a
highly notable exception, both for human (6, 19) and animal (22) data, due to its linear dose response-
this linear dose response could be due to the endogenous promotion by hormones in breast tissue. '
The only other human organ with a radiosensitivity comparable to that of the female breast for
the induction of cancer is the thyroid (6, 19). Like the dose response for radiation induced human breast
cancer in females, the dose-respanse curve for radiation-induced human thyroid cancer is also linear (6,
I . Y- breast, the presence of endogenous hormonal promotion could account for the linear
relationship. Not only are thyroid hormones known to enhance the cancer incidence in animals (11-13),
as discussed above, but it has also been shown that increased levels of thyroid-stimulating hormone
result in an increase in the incidence of human thyroid neoplasia (2*).
f. What data would you require before accepting the conclusion that a chemical is only a promoter?
The data that it caused no or few tumors in animals by itself but caused an enhancement in tumur
formation when given as sequential treatments after an initiating carcinogen.
5. What chemicals, do you think, would be good candidates for use in developing risk assessment
methodologies for tumor promotors?
What additional research, do you think, is required on these chemicals to better understand their
interaction with biological systems?
Those for which some data exist which suggest that the agent acts as a possible "promoter" and
for which human exposure levels can be determined. For examples:
1. drugs such as-
a) vallum-inhibits metaboHc cooperation (25) as do many other promoters (25) and medical
records could give accurate exposure histories, and b) phenobarbitai- a promoter for liver
carcinogenesis (reviewed in ref. 26); medical records could give accurrate exposure histories.
2. Compounds widely present in the environment, such as pesticides, etc.
a) dieldrin- an example of a potential promoter which, when tested as a carcinogen in animal
studies, formed "compound" dependent tumors in vivo (showing the reversibility characteristic of
promoter dependent tumors in vivo- in that tumors only remained in the tissue as long as the "promoting
agent" was present, and disappeared when the "promoting agent" was removed)
b) DDT, polychlorinated biphenyls, dioxins etc. (levels can be measured in body fat- from
mastectomies, etc.) (There is already in vitro data suggesting that such compounds can act as promoting
agents, for example see reference 25).
(5b) What additional research do you think is required on these chemicals to better understand
their interaction with biological systems?
C-24
-------
Further research is needed for those compounds suspected as promoters and for which there is
known to be widespread human exposure.
6. a. What generic considerations should be given to interspecies extrapolation of the risk associated
with exposure to promoters (body wt. vs. surface area; liquid soluble vs. water soluble)?
a. As promoters are known to be species and organ specific, and we don't know whether any
human promoting agents truly exist, there is no clear-cut answer to this question.
b. What chemical specific considerations should be given in interspecies extrapolation of the risk
associated with exposure to promoters?
b. As with the question above, we will be able to answer this question only after we have solid
human data documenting the fact that there are specific "human" promoters (which have been
documented by epidemiologic studies).
C-25
-------
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
15.
16.
17.
18.
19.
20.
21.
22.
23.
2*.
25.
26.
Copeland, E.5. Free Radicals in Promotion - A Chemical Pathology Study Section Workshoo
Cancer Res. 43: 5631-5637, 1983. v
Little, a.B., McGandy, R.B., and Kennedy, A.R. \1978): Cancer Res., 38s 1929-1935.
Snami, S.G., Thibodeau, L.A., Kennedy, A.R., and Little, J.8. (1982): Cancer Res., 42:1405-1411.
Weinstein, I.B., Wigler, M., Fisher, P., Sisskin, E., and Pietropaolo, C. (1978): Cell culture studies
on the biologic effects of tumor promoters. In: Carcinogenesis, Vol. 2, Mechanisms of Tumor
Promotion and Cocarcinogenesis, T.J. Slaga, A. Sivak, and R.Ko Boutwell, eds., Raven Press, New
York, pp. 313-333.
Marx, J. Do tumor promoters affect DNA after all? Science 219, 158-159, 1983.
Biological Effects of Ionizing Radiation - Committee Report III. (1980): The Effects on
Populations of Exposure to Low Levels of Ionizing Radiation. National Academy Press
Washington, DC.
Ron, E., and Modan, B. (1982): IN: Radiation Carcinogenesis: Epidemiology and Biologic
Significance, edited by J.D. Boice and 3.F. Fraumeni. Raven Press, New York.
Doll, R., and Peto, R. (1981): J. Natl. Cancer Inst., 77:1192-1308.
Dalrymple, G.V., Gaulden, M.E., Kollmorgen, G.M., and Vogel, H.H. Medical Radiation Biology,
W.B. Saunders Co., Philadelphia, 1973, p.74. 5*>
Kennedy, A.R., S. Mondal, C. Heidelberger and J.B. Little. Cancer Res. 38, 439-443, 1978.
Berenblum, I. (1974): Carcinogenesis as a Biological Problem, pp. 133-141, Norht Holland
Publishing Co., Amsterdam.
Doniach, I. (1974)* Brit. J. Cancer, 30:487-495.
Hall, W.H. (1948): Brit. 3. Cancer, 2:273-280.
Troll, W. (1976): In: Fundamentals in Cancer Prevention, edited by P.N. Magee, S. Takayama,
T. Sugimura, and T. Matsushima, pp. 41-55. University Park Press, Baltimore, MD.
Blumberg, P.M. (1980,1981): In Vitro Studies on the Mode of Action of the Phorbol Esters, Potent
Tumor Promoters: parts 1 and 2. CRC Critical Rev. Toxicol., 8:153-197, 199-234.
^ei/\SQTm' LB., Yamasaki, H., Wigler, M., Lih-Syng, L., Fisher, P.B., Jeffrey, A., and Grunberger,
D. U979): In: Carcinogens: Identification and Mechanisms of Action, edited by A.C. Griffin and
C.A. Shaw, pp. 399-418. Raven Press, New York.
Suss, R., Kinzel, V., and Scribner, J.D. (1973): Cancer, Experiments and Concepts. Springer-
Verlag, New York. . °
no^H\,Wat=ins,' ^V ?"?,£» J
-------
C.. .>ents for the EPA Workshop on Risk Assessment Methodologies for
Tumor Promoters
Daniel Krewski
Biostatisties
Carleton University
Ottawa, Ontario
CANADA K1S 5B6
1. How do you define tumor promotion? What is the basis for your
definition? What are the limitations of the definition?
A tumor promoter is a substance which, when administered over
an extended period of time, may, through nongenetic toxic
mechanisms, appreciably increase the rate of occurrence of
neoplastic lesions which have been previously initiated.
This definition is based on the notion that a tumor may be
initiated following the occurrence of genetic damage within one or
more cells within a specific tissue. Such initiated cells may then
undergo malignant transformation to give rise to a histologically
or clinically detectable cancerous lesion, with the rate of
occurrence of such lesions depending on the potency of the
carcinogen. The rate of occurrence of these lesions may be
increased by subsequent exposure to a promoter, which serves to
increase the pool of initiated cells through nongenetic toxic
mechanisms such as cellular proliferation. In order to demonstrate
a detectable increase in tumor occurrence rates, it may be
necessary to employ moderate to high doses of the promoter for
prolonged periods of time in order to induce toxic effects
sufficiently great so as to promote the development of the lesions
induced by the initiator.
C-27
-------
This definition of promotion is based on the concept of
expanding the pool of initiated cells within a given tissue. Even
without promotion, some initiated cells may undergo malignant
transformation as a result of further exposure to the initiator.
It is thus difficult to distinguish between a substance which
demonstrates promotional activity as defined here and one which
enhances the development of fully differentiated cancerous lesions.
It is also possible that substances with promotional properties may
also possess initiating activity, thereby precluding the
classification of some compounds as either pure initiators or pure
promoters.
2. Is it possible to quantitatively separate the promotional activity
from the initiating activity of a chemical in assessing the
carcinogen risk using the available data set?
Separation of initiation and promotional effects within an
initiation/promotion system requires special bioassay protocols.
At a minimum, treatment groups involving (i) short-term low-level
exposure to the initiator, (ii) long-term high-level exposure to
the promoter, (iii) a combination of these two regimens (with the
promoter being administered following the initiator) and (iv)
unexposed controls. This would allow direct measurement of the
effects of the initiator and promoter separately as well as the
effect of the initiator/promoter pair, and the corresponding
calculation of a quantitative measure of carcinogenic potency for
these three cases.
C-28
-------
Assuming initiation and promotion to be based on genetic and
nongenetic mechanisms, it may be possible to separate these two
effects through the application of a series of short-term tests
designed to measure genotoxicity and nongenetic effects such as
changes in cell kinetics and metabolism. Without direct bioassay
information on tumor occurrence rates, however, it may be difficult
to obtain good quantitative information on initiation and
promotional activity for use in carcinogenic risk assessment.
The Moolgavkar-Knudson two-stage model of carcinogenesis
suggests another possible approach to quantitatively separating
initiation and promotional activity. This stochastic
birth-death-mutation model assumes that two mutations, each
occurring at the time of cell division, are necessary for a normal
cell to become malignant. Initiating activity may be quantified in
terms of the rate of occurrence of the first mutation, which
transforms a normal cell to an intermediate or initiated cell. The
second mutation then transforms intermediate cells to cancerous
cells. In this model, promotional activity is quantified in terms
of the difference between the birth and death rates of intermediate
cells. Application of this approach in practice requires both
bioassay data on tumour occurrence and supplementary in_ vitro data
on cell kinetics in order to estimate all of the unknown model
parameters required to guage initiation and promotional activity.
C-29
-------
3. How would you approach the qualitative characterization of
promotional activity into a risk assessment?
In analogy with general toxicity other than carcinogen!city,
it may not be unreasonable to postulate the existence of a
no-effect level below which nongenotoxic promotional effects may
occur. In this case, a suitable safety or uncertainty factor be
applied to the experimentally observed no-effect level may
represent a viable method of risk assessment. Since the existence
of a no-effect level for genotoxic effects is less well
accepted, this approach should not be considered with initiators
which may be effective even at very low doses.
4. What-data would you require before accepting the conclusion that a
chemical is only a promoter?
In order to establish that a chemical acts only as a promoter
and not as an initiator, it would be necessary to demonstrate
negative results in bioassays designed to rule out the possibility
that the substance is either a complete carcinogen or an initiator.
Assuming that initiators act through interaction with genetic
material, evidence against the presence of initiating activity
would also be provided by the observation, of negative results in a
suitable battery of short-term tests for genotoxicity. In
r
practice, this may be the only way to demonstrate a lack of
initiating activity, since it would not be feasible to carry out a
series of bioassays for initiating activity using a variety of
known promoters.
C-30
-------
5. What chemicals, do you think, would be good candidates for use in
developing risk assessment methodologies for tumour promotors?
Since the phenomenon of initiation/promotion is complex and
not fully understood, it seems desirable to select compounds for
which a useful body of data on carcinogenic potential (including
information on both initiating and promoting activity) and
genotoxicity already exists. Saccharin, for example, is known to
promote urinary bladder lesions initiated by FANFT and BBN, and has
been subjected to extensive toxicological testing.
What additional research, do you think, is required on these
chemicals to better understand their interaction with biological
systems?
In order to better understand the biological interaction of
promotors with biological systems, it is important to improve our
understanding of the initiation/promotion mechanism itself. This
may be addressed using specially designed i£ vitro and in_ vivo
studies intended to provide further information in this regard.
6. What generic considerations should be given to interspecies
extrapolation of the risk associated with exposure to promotors
(body wt. vs. surface area; lipid soluble vs. water soluble!?
There is little empirical evidence to support the use of gross
indicators such as body weight or surface area in extrapolating
between species with carcinogens generally, including promotors.
C-31
-------
In the absence of such evidence, it is -difficult to recommend a
generic method of species conversion. In the case of body weight
vs. surface area, simplicity would suggest the use of body weight
for species extrapolation although prudence may dictate to the use
of more conservative surface area conversions.
What chemical specific considerations should be given in
interspecies extrapolation of the risk associated with exposure to
promoters?
Since many substances must undergo some form of metabolic
activation in order to exert their toxic effects, it is of interest
to determine the dose of the test compound delivered to the target
in addition to the dose administered exogeneously. This will be of
particular interest when the delivered dose is not directly
proportional to the administered dose, as will occur when one or
more steps in the metabolic activation process are saturable.
Prediction of the delivered dose in the target species may be
possible using physiologic pharmacokinetic models, but must be done
on a compound specific basis.
Since promoters may act through cellular proliferation, it may
also be useful to study their effects on cell kinetics in different
species.
C-32
-------
Robert Langenbach, NIEHS
1. Define Tumor Promotion
Tumor promotion is a process manifested by an increase in the number of
tumors (relative to controls) in uninitiated or chemically initiated ani-
mals due to subsequent repetitive treatment with a promoter. The defini-
tion is a generalization covering most model systems used to study the
phenomenon and which may also be applicable to the human situation. The
limitations of the definition are'that it does not describe any model pre-
cisely, and assumes spontaneous as well as carcinogen-induced tumor cells
are promotable.
2. Quantitative Separation of Initiation and Promotion Activities for
Carcinogen Risk
With our present understanding of initiation and promotion, and the
model systems available, it is not now possible to quantitatively separate
these activites and thus not possible to assess risk due to each indivi-
dually. For certain model systems (skin and liver) it may be possible to
state that a limited number of chemicals (i.e. TCDD, TPA, teliociden)
behave primarily as promoters. But evidence that this is not system-
specific (i.e. species, organ specific) is lacking. In addition, for TPA,
the most thoroughly studied promoter known, there is still a debate as to
whether it is a weak carcinogen. However, this does not mean some relative
separation of activities is unfeasible.
3. Qualitative incorporation of promotional activity into risk assessment
First, I would obtain ample usable qualitative data. If specific pro-
perties of a chemical, such as promotional activity are needed,, a more
comprehensive testing approach is also needed. However, I am not certain
currently available tests for promoters are entirely adequate; but I
believe, with system modification and/or improvement greater insight into a
chemical's promotional activity can be obtained. In a sense, each chemical
would require some basic research based on the nature of the chemical and
the information sought.
4. Data needed for a chemical to be only a promoter
It is not now possible to show that a chemical is only a promoter. The
reason for this is that promotion of "spontaneously" initiated cells cannot
be separated from cells which could be initiated and then promoted by the
presumed promoter. Until we can differentiate between spontaneously ini-
tiated and chemically initiated cells there will always be some level of
uncertainty. Furthermore, the absence of genotoxic activity in short-term
tests by a rodent carcinogen may suggest promotional activity; but the
limitation(s) of present-day short-term tests makes such conclusions
equivocal. In summary more knowledge about the carcinogenic process itself
is needed.
c-33
-------
5. Chemicals useful for developing risk assessment methodologies and addi-
tional research needed.
Chemicals which are believed to act via a nearly entirely promotional
mechanism would be good candidates to start. Examples of such chemicals
would include TCDD, PBB, teliociden and TPA. Eventually, chemicals with
both initiation and promotional activities, and chemicals with only ini-
tating activities (if any can be found) should be studied and all three
resultant risk assessment models compared.
In addition, I suggest that chemicals be studied for which there is (or
will be) human exposure data (phenobarbital, valium, etc.) and which can
also be studied in model systems for promotional activity. With this
approach, model systems and mathematical extrapolations can be developed
and validated relative to effects in humans.
Additional research is needed to understand mechanistic differences
between initiation, promotion and progression at the tissue, cellular and
genetic level. Understanding the causes of organ and species differences
would contribute greatly to practical (risk extrapolation) and mechanistic
knowledge. As the mechanisms of promotion for different chemical classes
are probably varied, research considering different mechanisms, rather than
a unifying mechanism should be conducted.
6. Generic Consideration and interspecies extrapolation
Too few interspecies studies have been done to answer this question.
There are species differences in response to promoters as evidenced by
different mouse, rat and hamster skin responses to TPA. However, many more
studies are needed before extrapolations could be conducted with con-
fidence.
c-34
-------
Comments for the EPA Workshop on Risk Assessment
Methodologies for Tumor Promotors
Peter N. Magee
Experimental Pathology
Fels Research Institute
Temple University School of Medicine
3420 N. Broad Street
Philadelphia, PA 19140
1. The concept of tumor promotion has been
derived from the early experimental work of Rous,
Mottram, Berenblum, Shubik and others in which it was
found that a single application of coal tar or a poly-
cyclic hydrocarbon to the skin of rabbits or mice in
subcarcinogenic amounts could result in the induction
of skin tumors if it was followed by wounding the skin
or by repeated application of croton oil, a powerful
irritant. In these classical experiments the carcinogen
was described as the initiator and the croton oil as
the promoter. Subsequent work by many investigators
confirmed and extended these findings and led to the
conclusion that cancer arises in a series of stages and
similar conclusions were drawn from studies of human
cancer epidemiology. The original criteria for a tumor
promoter in the mouse skin model have been recently
restated (Hicks, 1983) as follows:
The operational criteria originally defined for a
promoter in the mouse skin model were;
c-35
-------
(i) that it should not be carcinogenic per ^e_;
Cii) that it should not increase tumour yield if
administered before the initiating carcinogen;
(iii) that when applied after an initiating, sub-
carcinogenic dose of the carcinogen, it should
accelerate the rate of development of tumours
and thus increase the total, time-related tumour
incidence;
(iv) that the total yield of tumours produced should
be dose-related to the initiator not to the pro-
moter, providing the promoter is used in excess
of the minimum amount required to promote all
initiated cells;
(v) that unlike initiation which can take place
rapidly during a single exposure to the initiator
and which is a permanent event, promotion requires
long exposure to the promoter before the changes
induced become irreversible.
These criteria have served as guidelines for the
application of the concept of promotion to tumor induc-
tion in other organs including liver, bladder, colon.
Although there is wide acceptance that the concept of
promotion can be extended beyond the mouse skin system,
in most cases all of the criteria have not been met.
C-36
-------
Much recent work has been done on the biochemistry
and molecular biology of tumor promotion but a very
large part of this has involved the use of only one
compound, the diterpene ester 12-0-tetradecanoylphorbol-
13-acetate (TPA). it is probably premature to incor-
porate the findings of these studies into the criteria
of a tumor promoter and those listed above, derived
from the classical whole-animal experiments, should be
used to define tumor promotion.
With these considerations in mind, the definition
of a tumor promoter provided by 1.8. Weinstein, as
follows, is recommended?
Tumor promoters can be defined as compounds that
have weak or no carcinogenic activity when tested alone
but result in markedly enhanced tumor yield when
applied repeatedly following a low or suboptional dose
of a carcinogen (initiator). A possible extension of
this definition may be considered, again according to
Weinstein, as follows:
At the biochemical level, it appears that the
major difference between initiators and promoters is
that initiators (or their metabolites) bind covalently
to cellular DNAj but this is not the case for tumor
promoters.
C-37
-------
References:
Hicks, R.M. Carcinogenesis 4^ 1209-1214 (1983).
Weinstein, I.B. In: Host Factors in Human
Carcinogenesis (Bartsch, H. and Armstrong, 8., eds.)
pp. 9-25. International Agency for Research on Cancer,
Lyon, 1982.
2. Presumably this question refers to complete
carcinogens which induce tumors after single or
multiple doses without additional treatment with other
agents. If the chemical in question is known to be an
initiator and if it is accepted that initiation
involves genotoxic activity and interaction with DNA it
is difficult to see how any promotional activity it
might have, e.g. interaction with membranes, could be
separated from additional genotoxic activity. Such a
separation would not be helpful from a regulatory
standpoint.
3. Important qualitative characteristics of pro-
motional activity for risk assessment are that it is
claimed to be reversible and non-additive and that, in
contrast to initiation, threshold levels of activity
can be demonstrated. If these claims are true, a com-
pound with only promotional activity could be regulated
in the same way as one for which there is no evidence
of carcinogenic activity.
C-38
-------
4. In principle, data would be required to show
that the chemical had no carcinogenic activity by
itself. However, this criterion has never been met,
even with TPA or phenobarbital, the two most widely
investigated promoters. In practice, the decision
would depend on minimal or apparent total lack of car-
cinogenic action ir± vivo and _in_ vitro and minimal or
total lack of mutagenic activity in a variety of tests.
Absence of genotoxic action in its wider sense presu-
mably could not be required since TPA, the promoter par
excellence, has been shown to induce DNA damage in
several ways.
5. The literature on tumor promotion has been
dominated by TPA, first in the classical mouse skin
model and then in the various In vitro cell culture
systems in which sequential changes comparable to in
vivo promotion have been described. More recently
phenobarbital has been increasingly used as a promoter
in the several initiation-promotion systems involving
rat liver. Although other chemicals have been
described as tumor promoters in various systems they
have not been sufficiently studied to be good can-
didates for use in developing risk assessment methodo-
logies. Until more information is available on the
C-39
-------
mechanisms for action of the other known promoters, TPA
and phenobarbital are probably the best examples for
assessment methodologies.
A major field of additional research with these
agents would be to investigate their biological actions
for points.of similarity which might throw light on
possible common molecular mechanisms underlying their
common biological property of acting as a tumor promoter,
6. Interspaces extrapolations of the risk associated
with exposure to promoters poses an extremely difficult
problem because most of the work on TPA has been with the
mouse and that on phenofaar&ital has been mainly in the rat.
There may not be enough data currently available to allow
meaningful interspecies extrapolation.
C-40
-------
Comments for the EPA Workshop on Risk Assessment
Methodologies for Tumor Promoters
Henry C. Pitot
Carcinogenesis/Multistage
McArdle Laboratory for Cancer Research
University of Wisconsin
Madison, Wisconsin 53706
1. The stage of promotion in the natural history of
neoplastic development is operationally defined as
the reversible expansion of the initiated cell
population. Mechanistically this reversible expan-
sion is the result of altered gene expression
induced by the presence of a promoting agent.
The basis for this definition is the fact
that the carefully defined instances of tumor pro-
motion, skin and liver, demonstrate reversibility
of this stage, and the action of virtually all
promoting agents induce reversible effects in cells
Furthermore, all known promoting agents alter gene
expression in one way or another.
The limitations of this definition are to our
advantage in the understanding the natural history
C-41
-------
of neoplastic development in that previous interpre-
tations of "tumor promotion" involved all changes
beyond initiation to include malignancy. Again,
in those well-defined systems demonstrating multi-
stage carcinogenesis, the reversible stage of promo-
tion p.recedes the irreversible stage of progression
when malignancy appears.
2.
In multistage hepatocarcinogenesis it is possible
to quantitate both promotional and initiating
activity of a single chemical and to use such values
in assessing the carcinogenic risk of a particular
agent. In the present chronic bioassay systems
utilized it is not possible to quantitate or
distinguish initiating, promoting, or complete
carcinogenic activities.
3. If an agent demonstrates no DNA-damaging activity
but exhibits promotional activity, then the charac-
teristics of tumor promoters (see 4) must be taken
into account in risk assessment. These include
the reversibility of its effects and its concomitant
threshold leading to permissible exposures for
intermittent periods.
C-42
-------
6. Information on interspecies extrapolation of risk
associated with exposure to promoting agents should
include the following;
Pharmacokinetics and disposition based on
body weight, surface area, body lipid content,
etc.
Effect of intermittent exposure and "stop"
experiments.
Presence of receptors specific for promoting
agent involved.
Chemical considerations in interspecies extrapola-
tion should take the following into consideration:
Metabolism of the promoting agent with respect
to metabolites formed and possible DNA damag-
ing intermediates formed in one species and
not in another.
Action of the agent in producing an alteration
in gene expression, especially with respect
to its interaction with specific receptor
molecules.
C-43
-------
4. The following data would be required in order to
characterize a chemical exclusively as a promoting
agent:
a. induces the reversible alteration of genetic
expression in cells.
b. lacks DNA-damaging activity either as a simple
mutagen, a clastogen, or an agent altering
the structure of DNA by indirect means.
c. induces the reversible expansion of one or
another histogenetic initiated cell population
in vivo.
5. Estrogens or other steroid or polypeptide hormones,
phenobarbital, carbon tetrachloride, butylated
hydroxyanisole, ethanol, nicotinaraide, and lead
acetate.
Careful determination as to the DNA-damaging
action of these chemicals should be carried out
including effects _in vitro, clastogenic activity,
and potential DNA damage resulting from indirect
effects of these chemicals. The reversibility of
their effects in expanding initiating cell popula-
tions should be better understood in those tissues
where this effect occurs.
C-44
-------
Comments for the EPA Workshop
on Risk Assessment Methodologies for Tumor Promoters
Herbert S. Rosenkranz
Area of Expertises Strategy for the Deployment of Surrogate Tests
Department of Environmental Health Sciences
Case Western Reserve University
School of Medicine
Cleveland, Ohio 44106
This written contribution will be restricted to discussing strategies
for developing batteries of tests to identify promoters and to differ-
entiate these from complete carcinogens, non-genotoxic carcinogens, non-
carcinogens and to possibly refine the methodologies further to differ-
entiate between potent, moderate and weak promoters and finally, to
explore the possibility of differentiating between promoters acting at
different stages of the tumor progression process.
This approach is strongly influenced by our experience of the
analysis of data bases relating genotoxicants/mutagens to carcinogens.
Moreover, it is our hope that this could result in the development of cost-
effective as well as highly predictive batteries of tests for promoters
while avoiding the costly mistakes that were made in the development of
short-term tests for genotoxic carcinogens.
Before we delve into tumor promoters, let us first briefly sum-
marize the situation as it exists with respect to genotoxic carcinogens:
1. Approximately 17,000 chemicals have been tested in one or more
short-term tests. This does not include the results generated by industrial
and contract laboratories.
C-45
-------
2. There are more than 100 different short-term tests. As a matter of
fact, a recent IARC study lists in excess of 150 of these.
3. As a result of this unorganized effort, the data base is best
characterized as incomplete, i-e., very few chemicals have been tested in
a broad spectrum of tests. The few carcinogens that have been widely
tested are primarily alkylating agents or other strong electrophilic
species.
4. There are only about 1,000 chemicals that have been adequately
tested for carcinogenicity. Among these, there is a predominance (about
85%) of carcinogens. This does not reflect the chemical universe and
poses problems when devising algorithms to predict carcinogenicity based
upon the results of short-term tests.
5. In order to use these incomplete data predictively, carcinogenicity
bioassay results had to be pooled with respect to species and organ
specificity. Because of the paucity of data on specific tumors, it is
almost impossible to analyze the results of short-term tests with respect
to their predictivity vis-a-vis organ and species specificity.
6. No plans for systematic testing were developed even after it was
realized that the performance of individual tests, or combinations of
these, were not predictive.
7. There is no universal concensus as to what constitutes a positive or
a negative result in a specific assay. Thus classifications are frequently
incompatible. This is further aggravated by the fact that there is no
C-46
-------
agreement on how batteries of tests containing mixed results are to be
interpreted.
In view of the abovementioned situation, it is but now that system-
atic approaches are being developed to use the wealth of available
information to devise methodologies that would allow one to use the data
in a predictive fashion as well as in the design of strategies for the
assembly of predictive batteries for genotoxic carcinogens.
It should be pointed out that in developing such strategies we have
heretofore paid very little attention to the philosophical goals of the
testing program (the goals of a regulatory agency being different from
those of a manufacturer of a widely used over-the-counter medicinal
agent). Nor have we addressed the societal cost of the misclassification
of carcinogens on the basis of the results of short-term tests. Such
misclassifications will occur unless we take into consideration the fact
that there is no perfectly predictive test, each assay is characterized by a
different spectrum of false positive and false negative responses.
Accordingly, to devise successful strategies for the deployment of short-
term tests to identify tumor promoters, a concerted effort will be
required. The model could be of the type that our colleagues at the
National Cancer Research Center in Japan, under the leadership of
Dr. Sugimura, have undertaken, namely a targeted approach.
Some of the requirements for the development of adequate data
bases of tumor promoters should include the following:
C-47
-------
1. We need one or several standard animal systems to identify tumor
promoters against which other assays will be calibrated. This requires
agreement on the choice of one or several systems such as the two-stage
mouse skin assay, the two-stage urinary bladder assay, the induction of
hepatic foci, the two-stage forestomach assay, etc. Additionally, in each
instance a single protocol must be agreed upon as well as the a priori
interpretation of the assay
2. The choice of a panel of chemicals is also very important. It is
essential that tumor promoters as well as non-tumor promoters be
included. Promoters must not predominate in the data base as presumably
they do not predominate in the universe either. For future structure
activity considerations, a broad spectrum of chemical classes of
promoters should be included .
The choice of the short-term assays is very critical. We do not want
to repeat the mistakes that were made in the development of genotoxic
screening assays, i.e., a large number of tests, but each tested with a
restricted number of chemicals. Initially, because tumor promoters may
be organ-specific, a battery consisting of ten short-term tests should be
selected and tested with all of the test chemicals. If this approach is
chosen, the choice of the endpoints is most important. We do not want to
include a variety of tests that all measure the same endpoint, i.e.,
inhibition of cell to cell communication or a variety of cell transform-
ation assays (Table 1). Rather it is essential that initially each of the
major endpoints (Table 2) be represented by a single system using an
C-48
-------
agreed protocol, e.g., altered foci formation (Pitot), inhibition of cell to
cell communication, and enhancement of cell transformation. Moreover,
before such studies are undertaken, the protocol must define as to what
constitutes positive and negative responses and moreover when a positive
response is obtained, what constitutes a weak, moderate or strong
response.
During the early developmental phase , genotoxicity tests should be
included. This could be accomplished by the parallel testing of the
candidate chemicals for their ability to induce gene mutations,
chromosomal aberrations, cell transformations. Alternatively, the chosen
chemicals can include agents that have already been so tested.
The next step involves the establishment of criteria on how to
analyze the results especially with respect to the analysis of mixed
results. The goal being to reduce the 10 initial tests to a battery
consisting of 3 or 4 highly predictive tests that ideally should be able to
predict a variety of tumor promoters differing in species and organ-
specificity (however, see below). Our experience indicates that Bayes'
theorem is a very powerful tool for making such predictions even when
mixed results are obtained as it provides a numbered index of risk (from 0
to 100%). Additionally, it is also very useful in the batteries selection
process.
In developing the algorithms for selecting batteries, we have to bear
in mind that unlike the unique target (i.e.. DNA) involved in genotoxic
carcinogenesis, we may find that with promoters each target organ may
C-49
-------
be characterized by a different spectrum of predictive tests, i.e.. we may
have a different battery for liver as compared to bladder promoters.
In the selection of batteries we have to be able to differentiate
between tumor promoters, complete carcinogens, non-genotoxic carcin-
ogens and non-carcinogens. This may require the use of so-called mixed
batteries, for example, a complete carcinogen might be described as one
that is positive in a battery consisting of a subset of genotoxicity tests
and a subset of tests designed to identify tumor promoters. We then have
to determine by appropriate analysis and modelling, a priori, the level of
required response in each subset. Finally we must be able to differentiate
between non-genotoxic carcinogens and promoters, this also might be
accomplished by a different mixed battery of tests in which for example
we may require negativity in a battery of genotoxicity assays and
positivity in a battery for tumor assays. Again, the level of response in
each subset of tests will have to be predetermined by appropriate analysis
and modelling.
If we are fortunate enough to develop a data base and to identify a
number of tests, possibly five or six assays which are predictive of tumor
promotion, then we can set about to devise batteries which reflect certain
philosophical or scientific requirements; such as (1) a battery to differ-
entiate between weak, moderate and potent promoters, (2) risk averse or
(3) cost averse batteries. Finally, we might attempt to take into
consideration the societal cost of misclassification of tumor promoters
and to develop models to estimate the various scenarios and to construct
C-50
-------
batteries to satisfy these requirements. Thus, for genotoxic carcinogens
it has been suggested that the societal cost of false negatives is ten times
that of a false positive and batteries to meet these criteria have been /
constructed.
With respect to additional methods for predicting tumor promoting
activity, structure activity relationships present another target, Two
main approaches can be taken. One of these involves the application of
the artificial intelligence system (such as the CASE program developed at
Case Western Reserve University). Such an approach requires a data base
composed of active as well as inactive chemicals as well as a measure of
their potency, or lack thereof. CASE requires a learning set of approxi-
mately 30 chemicals divided among positive and negative responses for
each of the data sets, i.e., endpoints under consideration, e.gy bladder
tumor promoters. CASE allows the merging and comparison of data bases
to investigate, for example, whether the same mechanism underlies liver
and skin tumor promotion. The advantage of such an approach is that it
can then be used in a predictive mode to predict not only activity, or lack
thereof, but also the expected potency, and it is totally independent of
operator biases. .
Another approach involves the use of computer graphics to match
regions on the various molecules that may have the same configuration.
This approach has been taken by Drs. I.B. Weinstein and T. Sugimura.
However, in order to be effective and to restrict the number of possible
models, x-ray crystallographic patterns of some of the key molecules are
C-51
-------
needed. This greatly simplified the analysis by restricting the number of
possible solutions.
When undertaking such studies, it is evident that a number of
structural determinants have to be taken into consideration. For
example, the active phorbol esters all appear to share a single (or family
of) receptor site(s) which is (are) structurally determined. However, this
is not necessarily a stumbling block in the application of the CASE
program or a combination of the CASE program and computer graphics.
Thus, in a recent study on the structural basis of the activity of inhibitors
of bacterial DNA gyrases, it was found that there were two types of
structural determinants that were crucial; (a) a species-dependent deter-
minant controlling entry into the cell (ue.. permeability), and (b) a unique
determinant that involved the inhibition of the intraceUular DNA gyrase
and was independent of species. There is a similarity between DNA
gyrase and the tumor promoters with respect to the fact that we already
know that there may be specific receptor sites (e.g., for TPA-like
promoters) that are promoter/species specific and may affect entry or be
related to effects other than tumor promotion. Additionally, however,
there may be unique structural determinants which are related to the
intrinsic tumor-promoting activity proper. Obviously, there could be a
family of such determinants thay may be organ-specific. The data base to
be developed should resolve these possibilities.
In summary, it would appear that techniques are available for
identifying and predicting the activity of tumor promoters as well as for
C-52
-------
providing mechanistic information. However, to assure sucess will require
a concerted effort using standard protocols, an agreed upon group of
chemicals, an a priori agreement on the interpretation of test results and
the application of recently developed computer-based methods.
C-53
-------
Table 1
Inhibition of Intercellular Communication
Inhibition of metabolic cooperation - V79
- HPC
- Human teratocarcinoma
H-J-Uridine exchange in C3H/10T*
Citrulline incorporation - V79
- Human fibroblast
Dye transfer - Balb/c 3T3
- V79
- Syrian hamster cells
Permeability of tight junction - MDCK cells
etc.
Transformation and Enhancement of Transformations
C3H/10Ti - focus formation
Syrian hamster embryo cells - Morphological transformation
Balb 3T3
Rat embryo fibroblasts
Mouse epidermal cell line JB-6
Enhancement of SV40 ts - Swiss 3T3
C-54
-------
Table 2
Some Tests Used to Identify Tumor Promoters
Activation of EB-virus
Inhibition of TPA-induced activation of EB-virus
Release of lipid metabolism
Stimulation of arachidonic acid metabolism
Reduction of nitroblue tetrazolium
Skin irritation
Ornithone decarboxylase in vitro and in vivo
Modulation of cell differentiation
Aggregation: Lymphocytes
Adhesion of promyelocytic cells
Human promyelocytic cells +
Friend erythroleukemia cells
Hamster cells -
Sea urchin - etc.
DNA damage
Stimulation of DNA synthesis
Agglutinability by concanavalin A
Aneuploidy in yeast
Co-recombinogenicity in yeast
Inhibition of neurite formation
Gene amplification
Transformation assays
Inhibition of intercellular communication
Altered foci formation
etc.
C-55
-------
-------
Comments for the EPA Workshop on Risk Assessment
Methodologies for Tumor promoters
Thomas J. Slaga, Ph.D.
Biochemist and Professor of Carcinogenesis
The University of Texas System Cancer Center
Science Park - Research Division
P.O. Box 389
Smithville, Texas 78957
1. Tumor promotion can be defined both operationally
and mechanistically. The operational definition
of tumor promotion is the process that leads to
the induction of tumors when a weak or noncarcinogenic
agent is given repetitively after a subthreshold
or approaching a subthreshold dose of a carcinogen
(initiation). Tumor promotion must be defined in
relationship to tumor initiation. Mechanistically,
tumor promotion is defined as the process by
which an agent brings about the selective expansion
of initiated cells which increases the probability
of malignant transformation. Initiation may be
defined as a change in a target tissue which
induces an essentially irreversible alteration
such that subsequent treatment with a tumor
promoter expresses this event.
In the absence of a method to measure a"d
identify initiated cells, it is impossible to
057
-------
rigorously classify compounds as pure .initiators
or pure promoters.
Yes, it is possible to quantitatively separate
the promotional activity from the initiating activity
of a chemical in assessing the carcinogenic risk.
This can be achieved by testing a given chemical
at several dose levels as an initiator using a
standard promoter and as a promoter using a
standard initiator. However, in attempting to
classify agents quantitatively as either a promoter
or an initiator, it is generally recognized that
most agents appear to exhibit both activities to
a variable degree.
In order to incorporate the qualitative charac-
terization of promotional activity into a risk
assessment, it would be necessary for an unknown
compound to meet the operational definition of a
tumor promoter in a given tissue in several different
species. For example, it would be difficult to
accept an agent as a human skin promoter if it
was a recognized promoter only in mouse skin.
This is a very difficult question to answer since
to my knowledge, there are no known pure promoters.
i
This is especially notable in extensively studied
C-58
-------
compounds. In order to designate a promoter as
pure, it would be necessary to secure extensive
dose-response data on a given chemical following
testing as a complete carcinogen, a tumor initiator,
and a tumor promoter in a large number of possible
target tissues from several species. If these
conditions were met then one could be reasonably
certain that the tested compound was only a
promoter (pure promoter). This is highly unlikely,
however, due to the distinct possibility of
spontaneously initiated eel 1s.
5. Since major observable differences exist in tumor
promoters from various species and organs, it would
be necessary to select a number of chemicals in
order that all major target tissues would be
represented. The following chemicals would be
useful :
Teleocidin
TPA
Chrysarobin
Benzoyl peroxide
Phenobarbital
Saccharin
C-59
-------
Although others might be considered, these
compounds are especially useful due to the available
data base. However, extensive dose-response data
is also needed on these compounds in both tumor
studies as well as a number of short-term parameters
In addition, the effects of these compounds on
human tissue in culture are desperately needed.
6. This is also a very difficult question to answer
since a number of the generic considerations for
interspecies extrapolation depends to a certain
degree on the target tissue of the tumor promoting
agent. For example, a skin tumor promoter may
have different considerations than the liver in
terms of using body weight or surface area.
Penetration into the skin is difficult for highly
water soluble compounds, whereas lipid soluble
compounds can penetrate quite easily. Consequently,
the route of exposure is very important. The
possible metabolism and pharmacokinetics of a
chemical in various species are very important in
Interspecies extrapolation.
C-60
-------
Connnents for the EPA Workshop on Risk Assessment
Methodologies for Tumor Promoters
James E. Trosko, Ph.D.
Geneticist/In Vitro Assessment of Mammalian Mutagenesis
and Intercellular Communication
Department of Pediatrics/Human Development
Michigan State University
East"Lansing, Michigan 48824
517/353-6346 '••'-•
C-61
-------
1• Definition of Tumor Promotion
The concepts of initiation and promotion are operational, in that they
describe distinct phenomenological observations in experimental animal
model systems demonstrating that carcinogenesis in many organ systems
of several species consists of multi-steps.
I accept the Berenblum definition, and the Boutwel1-type of animal
protocols of initiation and promotion. That is, tumors are produced
experimentally when an animal is exposed to a single subthreshold dose
of a "carcinogen" (1), followed by a repetitive or continuous exposure
to a non-carcinogenic stimuli. I also accept the idea that initiation
and promotion involve distinctly different molecular/cellular mechan-
isms. In addition, I feel promotion is the r^ate-limiting step of
carcinogenesis (2).
The basis of my definition is the result of the historical whole
animal studies, plus hundreds of molecular, biochemical and cellular
studies on agents/conditions, _in vivo and in vitro, which can influ-
ence either one or both phases. In addition, the definition is,' in
part, re-inforced by my own research experience studying agents which
have been shown to be "carcinogens, initiators and promoters".
The major limitations to the definition are: (a) The concepts do not
imply any molecular or cellular mechanisms for the two stages, other
than the fact that initiation appears to induce an irreversible alter-
ation in a cell's genome, and that promotion must involve the clonal
expansion of the initiated cell; (b) also, physical objects such as
implanted solids, or wounding, burns, .etc. can also "promote"; and (c)
C-62
-------
the practical problem of distingusihing what really constitutes "car-
cinogens" or non-carcinogens? (3)
2. Since I'm not sure what the question is asking, I'm going to answer on
the basis of what I believe are the distinctions between carcinogens
and initiators and between initiators and promoters.
By definition, a"carcinogen" is an agent which can accomplish both
mechanisms underlying initiation and promotion. When a "carcinogen"
is a carcinogen, it can initiate and promote. When a carcinogen is
given at "subthreshold doses", it can initiate but can not promote.
Therefore the difference between carcinogen and initiator is basically
the dose level. In practical terms, this dose level seems to be the
threshold, in tissues, between no detectable tissue damage and de-
tected necrosis. (4)
Promotional activity on the other hand, in the case of mouse skin
papillomas and rat liver enzyme altered foci, seems to involve, among
other things, the clonal expansion of the initiated cell. Therefore,
without mitogenic activity, clonal expansion, by definition, can not
occur.
Clearly, agents which can mutate cells, via any mechanism (i.e.,
error-prone replication off of normal DNA; or error-prone repair off
of damaged DNA), have been shown to be good initiators. At low doses,
rare mutations can occur without much cell death, thus being
consistent with the concept of "initiator". At high dose levels, both
mutations and cell death would occur. The cell death, caused by a
C-63
-------
high dose of an mutagen, would induce regenerative hyperplasia, thus
allowing one to refer to this high dose of a mutagen as a
"carcinogen".
Promotion can be induced by an agent or condition which would allow a
single initiated cell to clonally expand, either by non-cytotoxic,
mitogenic stimuli (i.e., growth factors, 2',4',5',2,4,5-HBB) or cyto-
toxic-induced hyperplasia (wounding, high dose of "carcinogen" or non-
genotoxic cytotoxins). (5)
3. First, from my experience, there are promoters and then there are
promoters! The literature makes it quite clear that promoters can be
classified many ways (i.e., TPA-type which bind to receptors and act
as hormones at ngm levels; DDT or phenobarbital-types not needing
receptors and working at ugm levels; saccharin or NTA-types, which
work at mgm levels). Also, there are promoters which can be metabol-
ized (i.e., TPA) or excreted (i.e., saccharin), and others which are
not biodegraded nor excreted (i.e., PBB's).
Therefore, these factors of mechanism of action and of biodegradation-
excretion must be accounted for.
M. Aside from my own personal bias related to my hypothesis on the cellu-
lar mechanism of tumor promotion (i.e., inhibition of intercellular
communication) (6), I would have to know that the agent can not induce
point mutations, using only a few mutation markers. I do not accept
"""^™^ ™"^*™ /
..- ' (
any bacterial mutation assay data as relevant to the mammalian situa-
C-64
-------
tion. I also do not accept SCE's, UDS, or alkaline elution data as
equivalent to "genotoxicity". Moreover, the TK~ or TGr markers must
be accompanied by ouabain-resistance before I would accept any data on
a chemical's presumptive mutagenicity (7).
5. Based on my previous comments in (3), I believe there are at least
three or four distinct biochemical mechanisms of promotion. TPA is a
classic model for what I believe to be a relatively rare type of
environmental promoter. To me, it is a good model for endogenous
growth factor or hormone-types of promoters (i.e., those needing
receptors). I believe, PBB, Dieldrin, and Phenobarbital to be excel-
lent models for those typical environmental promoters which, being
lipophilic, need no membrane receptor to trigger their effects. In
addition, NTA or saccharin represent another distinct class. Finally,
TCDD seems to be in a class, quite distinct from all the rest, which
might again represent a rare, but important environmental promoter.
In general, since I feel promoters must be, among other things, mito-
gens, research on the ways these model compounds can be mitogens must
be studied (8).
6. In my personal opinion, one major area which has been ignored, except
in a few laboratories, is that of synergisms/antagonisms between
environmental promoters and endogenous factors (growth factors, such
as EGF, or hormones). This might explain sex, developmental stage,
and tissue differences.
While I understand the historic role of the whole animal model to test
and study promoters and promotion, I believe we are almost at the
065
-------
stage of _in vitro modeling with various normal human cell systems
(keratinocytes, kidney epithelium, hepatocytes). Therefore, I do not
hold any hope for whole animal to whole human extrapolation, since I
believe there are too many intervening factors.
In answering the last question on the nature of chemical specific
considerations needed for extrapolation of risks, based on my experi-
ence working with different kinds of promoters on different kinds of
cells from different organisms, including human fibroblasts and epi-
thelial cells (9), I'm not sure one can generalize for all types of
promoters (i.e., TPA versus PBB versus NTA). However, I have found
that if a given promoting chemical has a measureable effect, in vitro,
on one cell type of one organism, it has the same cellular effect on
the similar cell type of another organism. The chemical can have
dramatically different effect on a different cell type of the same
organism. Any difference between the chemicals effect on the whole
animal on one species and the potential risk to the human probably is
due to the indeterminable intervening physiological/immunological
factors unique to the human individual.
1.
One of the most important considerations, in my view, must be, "Is the
chemical metabolized or excreted in the humans." If not, it has the
potential to accummulate and reach critical mass levels needed for
promotion to occur.
References
J.E. Trosko, C. Jone, and C.C. Chang. "The role of tumor promoters on
phenotypic alterations affecting intercellular communication and
tumorigenesis." Ann. N.Y. Acad. Sci. 407:316-327, 1983.
C-66
-------
2. J.E. Trosko and C.C. Chang. "Potential role of intercellular communication
in the rate-limiting step in carcinogenesis." In: Cancer and the Environ-
ment (J.A. Cimino et al, eds.). pp. 5-22, Mary Ann Liebert, Inc., New York.
3. J.E. Trosko and C.C. Chang. "Role of tumor promotion in affecting the
multi-hit nature of carcinogenesis." In: Assessment of Risk From Low
Level Exposure to Radiation and Chemicals (A.D. Woodhead, C. Shellabarger,'
V. Pond, A. Hollaender, eds.), pp. 261-284, Plenum Press, N.Y. 1985.
4. J.E. Trosko and C.C. Chang. "Implications for Risk Assessment of Genotoxic
and Non-Genotoxic Mechanisms in Carcinogenesis. In: Methods for Estimat-
ing Risk of Chemical Injury; Human and Non-human Biota and Ecosystems
(V.B. Vouk, G.C. Butler, D.G. Hoel and D.B. Peakall, eds.) pp. 181-201, J.
Wiley and Sons, 1985.
5. J.E. Trosko and C.C. Chang. "Role of intercellular communication in
modifying the consequences of mutations in somatic cells. "In:
Antimutagenesis and Anticarcinogenesis Mechanisms (D.M. Shankel, P.E.
Hartman, T. Kada, A. Hollender, eds.) pp. 439-456, Plenum Press, New York
1986.
6. J.E. Trosko, C.C. Chang and A. Medcalf. "Mechanisms of tumor promotion:
Potential role of intercellular communication." Cancer Invest. 1:511-526.
7. J.E. Trosko. "A new paradigm is needed in toxicology evaluation."
Environ. Mutag. 6:767-769, 1984.
8. J.E. Trosko and C.C. Chang. "Chemical and oncogene modulation of gap
junctional communication." In: Tumor Promoters; Biological Approaches
for Mechanistic Studies and Assay Systems (R. Langenbach, E. Elmore and
J.C. Barrett, eds.), Raven Press, N.Y., in press.
9. Chang, C.C., R. Gibson-D'Ambrosio, J.E. Trosko and S.M. D'Ambrosio,
"Growth-promoting effect of TPA on cultured normal human fetal kidney
epithelial cells. Cancer Res. 46:6360-6363* 1986.
C-67
-------
-------
APPENDIX D
EDITED TRANSCRIPTS OF TOMOR PROMOTERS WORKSHOP
(Feb. 3-5, 1987)
-------
-------
EDITED TRANSCRIPTS OF TUMOR PROMOTERS WORKSHOP1
(Feb. 3-5,
FEBRUARY 3
TAPE 1
Vaun Newill: The purpose of this workshop is to seek your
advice in identifying research that will help to reduce the
uncertainty in one aspect of the risk assessment procedure. In
1982, the EPA Office of Toxic Substances held a workshop on
tumor promoters. Scientists were asked to address the issue of
how to incorporate information on promoter activity into risk
assessment. The participants acknowledged the need to
incorporate such information into risk assessment but were not
able to offer the Agency guidance on how to address the issue.
Recently, both the Science Advisory Board in its review of
perchloroethylene and the OPTS panel on dioxin recommended that
the EPA consider integrating promotional activity into the
traditional risk assessment. We have therefore convened
scientists familiar with the issue to help set research goals
to improve the scientific bases for addressing promotional
activity in risk assessment. We are not seeking guidance for
specific chemicals. We are seeking guidance on identifying
research to fill knowledge gaps in assessing the potential
hazard associated with exposure to promoters as a class of
substances, and on how to prioritize this research according to
its impact and utility for risk assessment.
Hugh Spitzer: We are seeking guidance for prioritizing
research on issues of tumor promotion.
Roy Albert: In the EPA guidelines for carcinogen risk
assessment that were adopted in I think September 1985, the
whole issue of promoters was waffled by saying that in the
absence of any evidence to the contrary, promoters should be
considered as complete carcinogens for risk assessment.
Everyone felt this was unsatisfactory, but there was not enough
consensus to develop an alternative approach in terms of either
a qualitative judgement of how likely an agent is to be a
promoter, or quantitatively, how great a cancer risk it might
pose for given levels of exposure. So for practical purposes,
EPA currently regards agents that show signs of being promoters
as complete carcinogens. This has been done for dioxin. so
I'll present one view of assessing the risk posed by
.promoters. Some key questions one might ask to identify a
promoter include:
Is there reversibility with respect to dose, i-e-
interrupt the dose, is the response more markedly
if Y°u
Icomplete citations for references in this appendix are
provided in Section 14, References, of the main report.
D-l
-------
diminished than with the total dose? (This is an accepted
pattern of behavior in the skin. I don't know if it's the
case in the liver.)
Is the initial response one of benign tumors?
interval fco cancer (e.g., 3 to 6 months in
There is some evidence of dose-response for proliferation
as related to promoting activity.
Short-term tests, cell-to-cell communication, activated
oxygen radical, SCE, etc.
I don't think that the issue of dose with respect to malignancy
has been well established (i.e., does the malignancy of tumors
diminish as dose decreases?). The assumption with carcinogens
is that as dose diminishes the incidence of tumors diminishes
but not the severity of malignancy. This is very different for
noncarcinogens.
I offer this list as a strawman as an example of a way to
evaluate promoters. Not all promoters will show all these
characteristics.
Henry Pitot: in systems where the multistage phenomenon has
been well characterized, I think that reversibility is almost
an absolute characteristic of tumor promotion, in the skin and
£h! itV^ thlS if certainly true, in other systems, such as
the bladder or the breast, where the stages have not been
adequately characterized, I don't think you can yet make a
statement such as this. I would argue that, for this
discussion, we should stick to systems that are well defined.
Roy Albert: Do you agree that the first lesions of promoters
are oenign?
Tom Slaga: That's obviously true in the skin and liver.
Benign tumors at least come before malignancy.
Henry Hennings: When we promoted for only 5 weeks (DMBA
initiation, DBA promotion), we only got one-fourth as many
papillomas as we did with 10, 20 or 40 weeks promotion, but we
go.t 3ust as many carcinomas. This suggests that there is a big
difference in papillomas with regard to whether they will
ornrtrooo »-„ „=,,„< Every carcinoma apparently arose from
Tom Slaga: I'd like to emphasize that different promoters
behave differently, some substances give few papillomas but
^anL?arC^n0inaS'. Dr' Saldez' Dr- Conti (sp?) and I have done
an extensive study currently in press at the NAS where we
D-2
-------
looked at all the benign tumors. Even early after tumor
promotion, most benign tumors are diploid. AS treatment
progresses, around 40 weeks (with treatment twice a week), all
benign tumors are aneuploid with sections that could be called
carcinomas. So even benign tumors could have a^lot of
characteristics as carcinomas if they are analyzed in detail.
ROY Albert: I think the point of what you are saying is that
the regressible characteristic of promoted lesions is the
characteristic of one component of the population. The
promoter brings out a spectrum of transformations in terms of
graded degrees of malignancy which range from very low
?ransformation to those that are well on their way to cancer.
I wonder whether that is not a property of the initiator, i.e.,
that the initiator is producing a spectrum of transformations
and the regressiblity is a reflection of that.
Henry Hennings: I think the initiator gives a spectrum of
initiated cells and the first one that are promoted by TPA for
example are those that are furthest along. The papiilomas that
Ire promoted for only 5 weeks with TPA don't/egress, With
longer promoter treatment, you hit a peak and then there is
regression (in Sencar mice).
TAPE 2
ROV Albert: We.could add another characteristic to the
list: A promoter is a poor progressor in advancing papiilomas
to carcinomas.
Henry Hennings: I agree with this but it depends on the
agent. In our studies, after treatment with the initiator and
promoter we got many papiilomas; then we treated with a third
stage agent to look for the progression of papiilomas to
carcinomas. When we continued TPA treatment, we found no
increase in the progression from papiilomas to carcinomas.
With urethane or 4-NQO, we found an increase in the number o£
carcinomas.
Tom Slaga: We looked at promoters such as benzyl peroxide
and chrysarobin and cytotoxic agents such as acetic acid and
hydrogen peroxide. Though certain promoters can act as
progressors, we find as potent agents that are nust as
effective. So we are not sure of the mechanism for bringing
about what we think is a second selection process through
cytotoxicity which selects out more aggressive cells and leads
to cancer.
Roy Albert: Is there a strong association between agents
that are progressors and genotoxicity?
Tom Slaga: Initial studies by Hennings and Yuspa suggested
that. But we're not sure it's all genotoxicity, even though
D-3
-------
all the compounds have a cytotoxic base. We did
nofacte^Vhere W6 kSPt raiSing the a°ae Vw
not act as a progressor. if you raise the dose of TPA to the
point where it becomes very cytotoxic, it at ill does not act as
a progressor. So it's not as simple as just cytotoxiSitv
There is something we don't understand. cycotoxicity.
We tried several isomers of the diol (sp?) epoxide of
b!n?anYfn.e ** pro^ressors "here we applied them once when the
benign tumors were present. The minus antidiol epoxide which
^nf-*°£initiat?r in the Skin is the most Potent of in the
agents we have looked at as a progressor to the point wherl all
our mice had multiple carcinomas. So there arl agents that aJe
extremely potent, we just don't understand what is going on?
Henry Pitot: In the liver, thus far, there is a clear
ff «i ?nltn ^tween a Progressor and a promoter. And I think
if hrui w°iS8eiy "K ^ Skin We wil1 find that the sa<"e thing
iff 5 f Phenobarbital in a carefully defined system you
get very few carcinomas that take a long time to appear . YOU
could argue that if the progressor effect is genotoxic in the
iJ««f' f fche incidence of the conversion of a cell in the
hh^n %hf ?r???ti?n to Progression is probably as low or lower'
than the initiation stage itself. That is, you can get
r»rS?S2 °fa,SelJ? initiated' b"t only one or two that develop
carcinoma. The liver has the advantage that you can identify
single initiated cells and their early clonal progeny I would
argue that we should separate promotion from progression! I
think we can do it in the skin and the liver. _ *>"°n. i
Roy Albert: What are examples of liver progressors?
o u Same as in the skin' One example was
Scherer et al- (1984). They initiated with
Promofced with Phenobarbital, and gave the
meffect n ?aSS an a^lati^ a^nt - END) and got the
same effect, in the liver, you can follow the lesions
resisting the second dose from very early on. You get the
beginning of a microcarcinoma within a focus or nodule. We now
think we have methods to quantitate the foci which provide! a
agent! 6 measuremenfc of the effectiveness of a progresso?
be^
^^°' Z ,think what is happening is that we have
Hkn „ dl?secfc the Sta9es much better. Initiation can be
likened to a point mutation which alters a cell but which in nn
way makes it malignant. A second alteration is probity a much
more complicated mutation such as a clastogenic effect. We
D-4
-------
have been thus far using classical initiating agents -
alkylating agents - all of which are clastogenic. We don t
have a system yet in which we have tried other things. It may
well be that agents that are purely clastogenic may turn out to
be processors! I think that we are clarifying the Picture.- I
think we must distinguish the reversible stage of promotion
from something beyond that which we have recently been calling
progression.
wov Albert- That could be the first landmark of this
wo?ksJop! Does everyone agree that we are talking ab^ three
stages - initiation, promotion and progression - and that there
are" Igents that are associated predominantly with these three
steps?
Peter Magee: How does the development of liver tumors after
application of DEN, phenobarbital and ENU differ from
application of DEN only by repeated dose?
Henry Pitot: By using Scherer's technique you can dissect
the stages. If you just give a complete carcinogen
continuously you can't distinguish the stages. Scherer's
system is a model system that allows you to dissect, quantitate
and determine something about each of the stages.
ROY Albert: Is it fair to say that all initiators are
progressors although not all progressors are initiators?
Henry Pitot: I think benzoyl peroxide may be an example of a
progressor that is nongenotoxic.
Jim Trosko: Anything that damages DNA can be a mutagen.
furthermore, agents that damage DNA not only act as nmtagens
but can also kill cells at appropriate doses, i.e., they are
cytotoxic. However, there are many cytotoxic agents that^do
not mutate cells. The example of acetic acid as a progressing
agent is probably associated with its cytotoxicity rather than
any mutagenic action whereas a mutagen as a progressor may also
be a oroaressor by its cytotoxic nature rather than
mutagenicity? FrSm what I understand, anything that can induce
cytotoxicity which then would force compensatory hyperplasia
can act as a promoter. So I am saying that it is important for
us to acknowledge cytotoxicity regardless of the mechanism.
Roy Albert: Cytotoxicity is a mushy term.
Bob Langenbach: Henry and Tom, in the systems you have
described, if you omit the promoter from this approach and
apply the initiator and then the progressor, do you get
malignant tumors?
Henry Hennings: It depends on the agent. Urethane gave no
malignant tumors. 4-NQO gave a few malignant tumors.
D-5
-------
Bob Langenbach: How about if you substitute a
tissue-damaging agent such as turpentine for a promoter in an
initiation/promotion/progression study? promoter in an
This experiment has not been done, if you
h ?n Waifc and give a Progressor, the majority of the
5« f? 6n "S6d (MNNG' EMU) wil1 give tumors because
Complete carcinogens, if you do those studies with
acid von rti lni^iafcf' wait ^d give acetone, then acetic
acid), you do not get any significant level of tumors, if you
JSJ SS ?i-Wlth TPA t0 9et benign tumors 'followed by promotion?
you don't express any additional benign tumors. notion,
Jre. Proinoters really of hazard relative to
agents are
trge.
m°del system' the promoter is just
P°P n °f initiated cells so ^u have a much
ir ™ P fr°m a Promoter ^ a function of the
background of spontaneous initiation, i.e., how many initiated
nro^011 ^ave^hen You st*rt. How can we'define
promoter when there is this background?
Henry Pitot: There is also spontaneous promotion. We have
SaoLPf°m0tecS bUilt int° °ur own organism and we eat many
wo?r?^ f A S° J W°Uld argUS that in humans we shou^ be more
worried about promotion than initiation.
ISnlt the Problem that we don't have pure
Progressors? Are there any
the other stages of carcinogenesis. But dioxin has peculiar
pharmacokinetics because it has a very long half-Ufe so "
?ne?eeif no e?idenc:llY ?H C°?tinuous dose- For'thJfl'cKmic.ls
cnere is no evidence in the liver of initiation.
Eliezer Huberman: Shouldn't we concentrate on these agents?
identi??^ Inrt!?H Ski?' sta?e-sP^if ic agents can be
identified. Urethane is an initiator in the skin, but can
oeroxidHr" jn-°!rher tissues- Chrysarobin and Senzoyl
peroxide are fairly pure promoters in the skin. Likewise the
iiitia?or benzoPyrene can be considered a pure
D-6
-------
[BRIEF DISCUSSION OF EXPERIENCE WITH INDIVIDUAL CHEMICALS]
Roy Albert: But my understanding is that agents may have
different properties in different tissues...
TAPE 3
Roy Albert: so it is difficult to extrapolate from
animals to humans. How do you know if the agent will behave
the same way? I hear that nongenotoxic agents can be
progressors.
Henry Pitot: My definitions are initiation of a cell>
produces some irreversible change in that cell which is
inherited by its progeny which means the cell could potentially
become malignant. Promotion is the reversible expansion of the
initiated cell population. Progression is an irreversible
change in a cell that likely involves a major structural change
in the DNA - either a clastogenic event, a major deletion, or a
translocation - so the cell is malignant.
Roy Albert: This workshop has been the most vigorous
exposition of progression that I have heard. So I think_that a
consideration of both promotion and progression are within the
scope of this workshop - and even initiation as a stage in ,
carcinogenesis.
Eliezer Huberraan: I consider initiation to be a genotypic
change that does not necessarily result in malignancy;
promotion, on the other hand, I consider to be the stage that
involves the conversion of an initiated cell to a tumor cell
and the clonal expansion of these cells.
Henry Pftot: In the liver, we can demonstrate that in the
presence of a promoter, the initiated cell population expands.
Once the population achieves independence of the presence ot
the promoter, we would argue that it is in the stages of
progression. If you remove the promoter, the majority of cells
in the altered foci disappear. We have isolated these cells
and put them back into an animal. They only survive as long as
the promoter is there.
Eliezer Huberman: The expanded cell population in tumor
promotion represents simply a quantitative difference between
tumor initiation and promotion.
Henry Pitot: Yes.
Roy Albert: But in the skin a promoter can produce both
reversible and irreversible lesions.
D-7
-------
Henry Pxtot: The same thing can happen in the liver. All I
am saying is that once you have gone to irreversibility, you
are likely in the stage of progression.
Peter Magee: How do you distinguish an initiated cell from
one that simply reacted with the initiator?
Henry Pitot: Japanese investigators have recently claimed
that they can demonstrate 1 in 1,000 cells initiated in the
liver using a different marker. I am not ready to accept
that, if you apply a promoter, only about 1 in 10 or 1 in 100
of these cells expand to a colony. So many of the marked cells
don t act like they have been initiated. So I would say that
the initiated cell population is only that population that can
expand in the presence of promoter. Only a few cells will
expand in the absence of promotion, but very slowly. One could
argue that this is spontaneous promotion.
Henry Hennings: We found that if we initiate and then wait
for 5 to 20 weeks before promotion, we get papillomas earlier.
The simplest way to explain this is that the clone of initiated
cells has increased in that time. We typically get papillomas
2 to 3 weeks earlier.
Roy Albert: Do we have a consensus on these definitions:
that progression is an irreversible change in DNA towards
malignancy and promotion is the reversible expansion of
initiated cells? NO COMMENT ON AGREEMENT OR DISAGREEMENT FROM
THE PANELISTS.
Dan Krewski: I was wondering if we should look at the
two-stage birth-death-mutation model that Moolgavkar developed
because that leads to very precise notions for initiation,
promotion and progression, if those precise definitions
correspond to what you have just indicated, then I would
agree.
Roy Albert: I think the striking thing here is the notion
that initiated cells are so common.
Anne Kennedy: work in Wisconsin by Clifton et al. (1984) has
shown that if you carcinogen-treat cells and then put them into
an animal, you only need a few cells to give rise to a cancer
If you can go down to 20 cells treated with the carcinogen and
it causes cancer in animals, it is not a single base mutation
that is the initiating event. Single base mutations are not
going to be in that high a proportion of cells. So if the
initiating event is common, we need to look at things quite
differently.
Henry Pitot: if you treat with a DNA-damaging agent at
sufficient levels, you can skip promotion. So we are dealing
D-8
-------
with a cell that is both initiated and progressed .
simultaneously. There is a lot of evidence in liver that high
doses almost immediately create cells that are malignant from
the start. Even Williams has claimed that when you feed AAF
continuously one can identify carcinomas that arise de novo.
So there may be doses of chemicals where you skip promotion and
go right to progression.
Anne Kennedy: Similar experiments in other labs have shown
that even low doses of carcinogens can lead to foci, formation.
Henry Pitot: I wouldn't agree with your conclusions for some
of those experiments. But I agree that mutation is quite
common. But I question the Japanese assumption that just a
change in a marker indicates an initiated cell.
Anne Kennedy: Stenback, peto 'and Shubik (1981) gave
different doses to mouse skin over orders of magnitude and got
the same final tumor incidence when all the animals_were
promoted. Their experiments indicated that initiation was a
common event even at low doses.
Jim Trosko: When a mutagen is applied, millions of cells are
exposed, but not all DNA lesions will lead to a mutation. Not
all cells are in the same cell stage, some are stem, committed
or differentiated. Another question is: Where does the
mutation occur? Not all mutations will affect the potential
for promotion. When we refer to initiated cells, we are
referring to a select number of cells in which a few events
occur in a select number of genes, even though all the cells
were exposed to the initiator. ,
Roy Albert: From a risk assessment standpoint, it might be
useful to try to characterize an agent in terms of the
individual actions and its potency with respect to these
individual actions. For example, if an agent were purely a
promoter, then it would probably be less dangerous at low doses
than another agent that had an irreversible component to its
action. Perhaps ultimately we may be able to dissect an agent
according to the balance between initiation, promotion and
progression. ,
COFFEE BREAK
Eliezer Buberman: The concept that promotion is simply an
expansion of initiated cells that are already transformed to
tumor cells does not agree with studies in which the hamster
embryo colony cell assay was used. In this,cell transformation
system, we treat single cells seeded for colony formation with
a low dose of a carcinogen (ice., an initiator) and then a
couple of days later incubate the cultures with a tumor
promoter (e.g., a phorbol diester). Thus, the frequency of
D-9
-------
intiated cells in the control and tumor-promoter treated dishes
is the same, yet, the frequency of transformed colonies in
cultures treated with this tumor, promoter is higher than in
cultures of_cells that had been treated only with the low dose
of the carcinogen, if it was simply a question of expansion,
you wouldn't be able to see the effect of tumor promoters in
the hamster colony assay. Furthermore, removal of the phorbol
ester results in the reversion of most transformed colonies to
a normal phenotype.
Roy Albert: No, it is not a question of expanding the number
of>initiated cells, but rather the clonal outgrowth of
initiated cells. So the number of initiated cells doesn't
change but it makes the initiated cells proliferate rapidly to
form a clone.
Eliezer Huberman: But even that is a problem, because if you
remove the promoter from the dishes, more than 80% of the
transformed colonies will revert to the normal phenotype.
Jim Trosko: in a recent paper in Science (Herschman and
Brankow, 1986), UV was used as an initiator in a cell system,
and only after TPA did they find transformed colonies, when
they plucked these colonies out and grew them in the absence of
TPA, they looked normal, but when they cocultivated those pure
initiated clones with normal cells they didn't see any foci
except in the presence of TPA. So I would suggest tha€ TPA
removed a suppression.
Eliezer Huberman: I'm arguing that a tumor promoter can
directly alter the phenotype of the cell, but this alteration
requires continuous presence of the tumor promotor. I would
therefore, suggest that the definition of tumor promotion be
modified.
Henry Pitot: If you remove the TPA and then reintroduce it,
are the cells phenotypically different, or are you dealing with
a phenotype that is set within the initiated cell and you are
getting selective growth of those cells in the presence of
promoting agents? -
Eliezer Huberman: in our case, we are dealing with a colony
from a single cell. *
TAPE 4
Eliezer Huberman: We must inject into the definition of
tumor promotion that the promoter induces the initiated cells
to convert from a normal phenctype to a tumor phenotype.
[SOME MORE DISCUSSION OF HUBERMAN'S EXPT.
SIMILAR RESULTS IN THE LIVER]
HENRY PI,TOT HAS SEEN
D-10
-------
Henry Pitot: Suppose we alter the definition to say "the
reversible expansion and the reversible alteration of gene
expression in the initiated cell population"?
Eliezer Huberman: Yes, and I would alter the order.
Roy Albert: So does everyone agree that "promotion is the
reversible alteration of gene expression and the reversible
expansion of initiated cells." [NO ONE DISAGREES]
Freddy Horaburger: What you are talking about works in the
Sencar mouse but may not work in another strain or species. We
should make an effort to study these things in different
species to see if there are significant species differences.
Then we could conclude with more certainty whether we could
extrapolate to humans. "
[ROY ALBERT AND TOM SLAGA AGREE]
Anne Kennedy: I don't think even an exhaustive study of _
species will help us know what might affect humans. I think we
must do human studies to get data on human risks. We already
know that promoters are extraordinarily species and organ
specific.
[HENRY PITOT TALKS ABOUT PHENOBARBITAL EFFECTS IN HUMANS VS.
ANIMALS]
PRESENTATION OF PANELIST PREMEETING COMMENTS (the name of the
panelist presenting the comments is underlined):
Eula Bingham: I have trouble translating animal data to
humans. I would like to see us talk about human experience
with promoters and then back up and see how that fits the
experimental animal data. Cigarette smoke is one possible
candidate for this. Arsenic and dioxin are another.
Henry Pitot: DDT is clearly a promoting agent for the liver
in rodents. There have been at least 30 epidemiological
studies for DDT, none of which indicate any carcinogenic effect
in humans. So you have a relatively high exposure of a
compound which, by our definition, is producing reversible
effects yet you see nothing in humans to indicate its (DDT)
carcinogenicity.
[DISCUSSION OF DDT]
Eula Bingham: I was suggesting starting with a chemical that
does have an effect in humans.
Henry Pitot: I would argue the other way. It is important
that we find out what a known promoter does in the human.
D-ll
-------
Bob Langenbach: I agree, if we are going to do animal
experiments, we might as well do them on chemicals for which we
have or will someday have human data.
[MORE DISCUSSION OF DDT AND HOW VALID THE DATA ARE]
Bob Langenbach: Maybe the dose was not high enough in humans
to see promotion.
Jim Trosko: Two papers are coming out (Aylsworth et al.,
SJ£? ?} that show synergism between two promoters (TPA and
DDT). Although these two substances seem to act very
differently, there is a tremendous synergism. This may pose
another element of complexity. There will be other confounding
tactors that we won't measure, especially in human studies.
Synergism is another potential area for study.
Henry Henningsr I think that progression can be divided into
oJSv2r«°°re«8tS?eS'. Weuhave called the first stage "malignant
conversion," which is the conversion of a papilloma to a
squamous cell carcinoma. The next stage is metastasis.
Promotion must come after initiation. I think we need to be
careful about our definitions, in terms of human exposure, I
would be much more concerned about progressors than promoters.
John Scribner in a 1983 paper in Cancer Research looked at
bromomethylbenzanthracene. ~~
TAPE 5
....He used the approach of testing in stages to determine
whether something is a promoter, initiator or progressor. Just
take a defined model system, .and test the chemical separately
with a known initiator and a known promoter. Few chemicals
have been tested in this way Some agents are better than
others at giving carcinomas versus papillomas.
[TOM SLAGA TALKS ABOUT B(a)P MECHANISM OF ACTION]
Freddy Homburger; [Presents his premeeting comments (see
Appendix C).]Knowledge of the mechanism of cancer is too
fragmentary to permit any conclusive risk assessment, except
from human epidemiology data. Risk assessment for promoters is
even more complex.
Eula Bingham: What is there about TPA or anything else that
takes a papilloma to a carcinoma? what is the significant
biochemical event?
Henry Hennings: In our experiments, TPA is not good at
converting papillomas to carcinomas. I think it is something
fc?? initiated cell, i think you have a spectrum of
cells, some of which are predestined to be carcinomas
D-12
-------
from the beginning and you can promote those cells more easily
than the others.
Tom Slaga: I believe like Henry that you have a whole
spectrum of initiated cells with different degrees of
initiation, and that is what makes the whole process so
complicated.
Does anyone know the mechanism of regression of
This is an area of research that we could
Roy Albert:
papillomas?
recommend.
Eliezer Huberman: At the present time we don't know the
mechanisms of any of the stages in tumorigenesis, namely, tumor
initiation, promotion and progression. We have discussed the
possibility of clonal expansion and inductive processes, and
there is also the possibility that immune surveillance is
involved. I think we have to analyze not only the mechanism of
tumor promotion but also that of tumor intiation and
progression since each step may involve a different mechanism.
I also have a problem with the fact that most tumor promoters
that we use were originally isolated as irritants. So a priori
we have selected a group of tumor promoters that are
irritants. We may be missing other classes of tumor promoters.
Roy Albert: But irritation is not a characteristic of liver
promoters. Does phenobarbital in the liver induce cell
proliferation?
Henry Pitot: It depends on what reference one cites. [He
mentions specific studies.] I think that in the liver, the
evidence argues more that the known liver promoters do tend to
increase DNA synthesis, at least transiently.
[PANELISTS DISCUSS NANCY COLBERT'S STUDIES WITH GENES]
Anne Kennedy; We do not yet have one human promoter. I
think a top priority is to establish what data could establish
promotion in humans. My comments concern what data we could
look at to establish promotion in humans. Many years ago we
promoted with saline instillations to the lung. As an
initiator, we used polonium 210, which is an alpha-emitting
radionuclide that is in cigarette smoke and is like the
radiation we are exposed to in radon gas and its daughter
products, plutonium from the nuclear power industry. It is a
relatively widespread environmental contaminant. A relatively
low dose of polonium 210 to hamster lungs produced few lung
tumors. However, promotion with 7 installations of saline 5
months later results in lung tumors in 22-44% of exposed
animals. Saline instillations themselves do not cause cancer.
At present, some uranium mine workers who have been exposed to
alpha radiation from the radon daughter products are having
D-13
-------
their lungs washed with saline to remove the material. This is
also being done for people exposed to asbestos. This has
happened in Boston, Texas and Colorado. So there are some two
populations that could be studied for promotional effects from
bronchial lavage we get peripheral lung
adenocarcinomas. The uranium miners get primarily bronchoaenic
carcinomas....
TAPE 6
.....I want to challenge the concept of irreversibility of the
initiating lesion. Our studies indicate that the first event
is potentially, though not usually, reversible, if protease
inhibitors are given sometime after the radiation exposure,
there is no transformation. We think that some agents such as
protease inhibitors can completely revert cells to a
noninitiated state whereas other agents cannot.
Several agents are highly antipromotional in in vitro systems.
When we do look at humans we could look at the shape of the
curve to determine if promotion is occurring, in in vitro
transformation systems the curve is the same (essentially a
quadratic or linear quadratic) no matter what the initiator
is. The presence of a promoter makes the curve linear. This
is a dramatic effect.
....TPA and other agents can promote transformation in cells 13
generations after initiation.
There are many human populations that could be studied for
promotion with radiation as the initiating agent. These are
the radium dial painters that were exposed to alpha-emitting
radionuclides that led to bone cancer. Also, people who were
given radium treatments in the '40s and '50s for various
medical problems.
The studies that I think would be most appropriate are thyroid
Ca2C?cn' From irradiations of people carried out in the '40s
and 50s, we are expecting about 20,000 cases of thyroid cancer
in this country. These are people who were irradiated for
IJ5XOUmubenign disorders - eczema, acne, thymus enlargement,
etc. They have found that in these groups it is highlr in
women, Jewish people, people who have emigrated from Tunisia or
Morocco, people who have immigrated to Hawaii. They have
defined many risk groups. The people are alive because
mortality is extremely low (1-3%) for thyroid cancer, so they
are around for purposes of interviewing. The way they tell if
you have thyroid cancer is to give you a scan with an isotope
like iodine 131 that gives a dose of 200 rads to the adult
thyroid. That is the optimal dose for the induction of cancer
in many different systems. This would be an excellent
progressor - a yearly dose of radiation.
D-14
-------
There is a dose-response relationship for many of the several
different kinds of cancers that have developed from radiation.
The thyroid and the female breasts have the highest
sensitivity. These have linear curves. The breast is a tissue
that is under strict hormonal controls and has perhaps a
built-in promotional agent and maybe this is why the curve is
linear. The same is true for the thyroid. We know that
thyroid hormones act as promoters in several in vivo and in
vitro systems. The linear response suggests that a promoting
agent may be important in the genesis of those kinds of
cancers.
I think the uranium miner studies are the best potential data
for human promotion. Many houses in Massachusetts have
exposure to levels of radiation like those in uranium mines.
Radiation is an environmental agent that is capable of
initiating cells. The curve for induction of lung cancer in
white uranium miners who smoke is linear. When the first study
of the Colorado plateau uranium miners was performed, the
nonwhite miners (American Indians) who didn't smoke had a
nonsignificant incidence of cancer. But recently Indians have
been getting cancer at an elevated rate/ and the cancer
incidence curve is now significantly above control levels.
Here are in vivo/in vitro data from three labs through 1984.
These data allow investigators to know how many cells have been
exposed to carcinogens and give rise to tumors. Clifton et al.
(1984) used fairly high doses of radiation to get the effects.
They found that at 20 cells per graft site, approximately half
the animals got tumors. It is hard to envision a single base
mutation in a large proportion of those cells which could.lead
to the tumors. Dr. Pitot, don't you assume that the initiating
event is the same, whether the dose of carcinogen is high or
low?
Henry Pitot: The dose does matter, because with a high dose,
you are getting initiation and progression almost
simultaneously. I would argue that if you want to look at
initiation you should isolate the event. If you telescope it,
you are looking at complete carcinogenesis.
Anne Kennedy: If the initiating event is common and there
are many initiated cells around, then the rate-limiting steps
in carcinogenesis are the later ones. If the first event is
common, then it is not a single base mutation in DNA which
occurs with a frequency of about one in a million. The
frequency of the initiation event makes a difference in how you
go about studying it. If it is a common event, adduct removal
studies, for example, to my thinking are meaningless. If we
think it is common, then perhaps we should be focussing on
promotion and progression.
D-15
-------
Henry Pitot: YOU cannot say that saline is a promoting
agent, it starts a chain of events in the animal that alters
the hormonal environment which alters gene expression. In the
whole animal, you can give many things which we would never
think of as promoters, but they change the internal environment
in such a way that the end result of many factors in tumor
promotion.
Freddy Hoinburger: The instillation of saline or anything
else into the hamster lung cannot be compared with human
response. The hamster takes it without any general response -
no adrenal enlargement, hormonal change, or struggle. I think
we are a little too quick to label something as a promoter just
because we see an increased tumor incidence when it is
administered following exposure to a carcinogen, in recent
experiments with guinea pigs (McFadden et al., 1986), it was
shown that the retention of small particles of asbestos is
increased by the inhalation of cigarette smoke. Therefore
their effect could be enhanced, but this would not be
promotion. With saline, you may just be changing the dynamics
of the disposal of polonium particles.
Roy Albert: Freddy Byrnes (sp?) did an experiment with
radiation on the back of the rat using doses that were
marginally tumorigenic. We did skin stripping which produces a
brisk proliferative response and we plucked hairs once a month,
and nothing happened. So maybe cell proliferation work's in
some places but it doesn't in others.
Henry Pitot: That points out that Eliezer Huberman's
addendum to the definition may be very important.
LUNCH
Dan Krewski: [Reads his definition of promoter from his
premeeting comments. See Appendix c.]
TAPE 7 .
Tne multistage model assumes that a cancerous lesion occurs
following the completion of k distinct stages which are usually
thought of as being some kind of mutational event. This model
does a good job of describing the age-incidence curves for
human cancer which are generally related to some power of age.
One problem is that you may require a model with five or six
stages in order to adequately describe some human cancer data.
This may not be biologically reasonable.
An alternative model is the Moolgavkar, Venzon and Knudsen
(M-V-K) model which is a stochastic birth-death-mutation model
that involves only two stages. The advantage of this model is
that it incorporates explicitly information on the kinetics of •
D-16
-------
tissue growth and differentiation. It involves only two stages
relating to mutational events and it seems to be consistent
with much of the experimental and epidemiological data that we
have in the area of carcinogenesis. The model assumes that you
begin with normal stem cells that can either divide into normal
progeny or be killed or you can have mutation resulting in one
normal daughter cell and an intermediate or initiated cell.
The same things can happen with the population of initiated
cells. They can divide, die, or undergo a second mutation to
produce a fully transformed malignant tumor cell along with
another intermediate cell. [SEE FIGURE 1 IN REPORT] It
follows that the age-specific incidence for cancerous lesions
under this model at time t is:
o
This involves the two mutation rates u^ and u2 and the
number of normal cells x(s) in the tissue at time s and another
term that takes into account the birth (a2) and death (b2)
rates for intermediate cells. The mutation rates determine
what the overall level of tumor incidence is going to be under
the model, and the second term, which involves normal tissue
growth and the birth and death rate of initiated cells,
determines the shape of the curve. If we define initiator,
promoter and progressor in terms of this model we get the
following definitions:
An initiator is a substance that increases the rate at which
the first mutation occurs, i.e., it increases m. If it is a
genetic lesion, it may be reasonable to assume that anything
that increases u^ may also increase u2 although the
magnitude of the two changes may differ. If this is so, then
prolonged application of an initiator may result in the
induction of a neoplastic lesion.
A promotor is substance that increases the pool of intermediate
cells available for subsequent malignant transformation. This
can happen either by increasing the birth rate of the
intermediate cells - a2 - or decreasing the death rate - b2
- or both. So it is the difference (a2 - b2> that is
important.
I would like to define a progressor as something that increases
U2. [THE GROUP AGREES WITH THIS]
Do we want to consider an intermediate or initiated cell as a
neoplastic lesion? [THE GROUP RESPONDS NO]
Do we want to consider an expanded colony of such intermediate
cells as a neoplastic change? [THE GROUP AGREES THAT IT IS NOT
NORMAL]
D-17
-------
A complete carcinogen is something that would increase the
rates of occurrence of both the first and second mutations, is
it also necessary to have promotion occurring in order for a
complete carcinogen to exist? I would say no. So a complete
carcinogen could involve either initiation and progression
alone or all three stages. [THE PANELISTS AGREE]
What would we call an agent that increases the rate at which a
malignant lesion develops? [THE PANELISTS AGREE THIS WOULD BE
CALLED A GROWTH FACTOR]
Roy Albert: This model is a simplification of reality.
There is probably a spectrum of lesions of graded malignancy
all the way across. For instance, there are papillomas that
regress and those that don't.
Dan Krewski;- Promotion was suggested as being largely a
reversible phenomenon, but how would you reverse the expansion
of a pool of cells? Why would cells suddenly start to die when
you remove the promoter?
Jim Trosko: I think it is because all these studies have
ignored intercellular phenomena. Traditionally we have thought
that cancer lies within a single cell, but people such as
Potter'(1981) argue that cancer involves relationships between
cells. This kind of a model does not capture that
interaction. So the interaction between cells could be
responsible for the reversibility of the hyperplastic effects.
Normal cells can suppress the phenotype of this premalignant
lesion if there is normal interaction, but if promoters block
that suppressing effect, then not only does it cause clonal
expansion but it prevents the phenotypic alteration which I and
others believe is prevented by some intercellular signal. We
must develop models with interactive features.
Dan Krewski: That sounds very reasonable.
Tom Slaga: The way it stands, the initiated cell would
continue to gain number all the way through until you promote.
It may change slightly, but it would not be double or triple
the tumor response.
Jim Trosko: Bell (1976) in 1973 or so presented a
mathematical model where he suggested that the premalignant
cells could become autonomous when they reached a critical
mass. So that implies that once the pool of initiated cells
gets large enough, it negates the interactive effect.
Henry Hennings: We know that there are papillomas that are
TPA-dependent and there are also autonomous ones.
Roy Albert: It is not clear why papillomas regress.
D-18
-------
Dan Krewski: I don't think anyone has mentioned the
initiation/promotion/initiation (IPI) type of study which is
relevant for separating promotional and progression effects,.
If we agree with the model that we are increasing the pool of
initiated cells, if we a second initiator to enhance the second
mutation rate, u2, we should get a dramatic increase in the
crop of malignant lesions that you observe at the end of the
study. There are several studies where that is the case.
Substances may possess initiating, promoting and progressing
activity. In this two-stage model, if the mutation rate per
intermediate cell division (u2/l^2+a2^ is considered to
be a constant, then an agent that increases the proliferation
of the intermediate cells must also increase the mutation rate
for the second stage. So if it is a constant, then substances
that possess promotional activity within the context of this
model may also demonstrate some potential for progression as
well.
Roy Albert: That relates to the question of whether simply
increasing cell turnover would accomplish promotion and
progression. That would tend to support your formulation. I
think the model fits what we have been talking about. What
does this model do for us in terms of risk assessment?
Dan Krewski: That is a difficult question. To apply the
model you will need data on normal cell growth, on cell
kinetics at the proliferation stage, and on tumor occurrence as
a function of dose. So more data will be needed betore the
model can be applied. We need data on how the mutation rates
vary with dose. The model has been applied to both
experimental and epidemiological data and it seems to fit well.
Curtis Travis: I think the model can be used because it
provides a theoretical framework with which to interpret
experiments or to propose experiments. We can't use it for
risk assessment until we know more about mechanisms. One
research suggestion is to start with background cancer rates in
mice, rats and humans, and measure regular mitotic rates as a
function of age. Then the only parameter that you are missing
for this model is the mutation rate. That is supposed to be a
constant. So there, is only one unknown constant, and you want
to see if you can reproduce the age-specific cancer Lates in
mice or rat livers. We have done this, where we found the
background mitotic rates for liver as a function of age,
assumed a mutation rate of 10~8 and were able to reproduce
the age-specific cancer rates in rats just from the model. We
are trying to do it for mice now. It could also be done for
humans.
Another suggestion involves tetrachloroethylene, for which
there are long-term cancer bioassay data for mice. Increased
D-19
-------
measured ?5 £ ** aS ? function °f ^ose in mice have been
measured. If we can determine the background mutation rates in
mice from the background cancer rates in -mice, you can combine"
??^f "* actual measured increased cell turnover rates in
liver from TCE application, use the model to predict the
ihSeaf.!Lin SUT°r ra^S and comPare ^ with 'the actual data.
^™X model provides a conceptual framework with which to
propose and interpret experiments. YOU could take a look at
the increased cell turnover rates in the mouse liver, apply the
model and make predictions as to the enhancement you should
have seen in the background tumor rates. Moolgavkar did it for
breast tumors, although this work could be improved. Almost
every parameter in this model is obtainable. The model
provides a strong theoretical framework and driving force for
experimental work in this area.
(Potter, 1981)
after niM- One PrediCti°n Of this framework is that
after initiating groups of animals with the same dose of
initiator, but promoting for different periods of time, the
1 ia 8d ^11S W°Uld Vary dePending on the duration of
T?2n' de?endin9 on the Period of promotion, you
°ellS t0 the Same level of a new initiator.
This would be a way to test the model. I don't think this
?22SJenS ?aS been done: initiate, (2) promote for various
lengths of time, and (3) then initiate at those different
promoting times.
*ennin9S: We did this, but Che data are not published
J*e. Prom°ted for 5, 10 or 20 weeks after initiation by
??;vin9 urethane or 4-NQO in the third stage. We found
that the response to urethane or 4-NQO was best with the
shorter promotion. This study will be published in the
Abstracts of the American Association for Cancer Research in
j.yo I .
von S?9aJ -V? haV6 d°ne similar experiments. I think once
level ??ncnn??h% ?Snry Ptt0tr in YOUr rat stales, you see a
levelling off of the number of what you believe to be
thSt 2S!?U^fc?S0irS in.the animals that y°u Promote, which means
that most of the spontaneous tumors probably occurred during
X8 ?• neonatal life« Do vo^ think that spontaneous
continues to occur during the animal's lifetime?
TAPE 8
°Ur uata su^est that spontaneous initiation in
mn^Ko i1Ver rea2hes a maximum somewhere between 6 weeks and 3
months of age. After that time up to almost a year of age
there is no significant change in the total number of initiated
D-20
-------
cells in the liver. We have interpreted that to mean that the
process of fixation of whatever the requirement for cell liver
replication in the initiation process is lost just because the
liver cell doesn't replicate significantly after that time.
Dan Krewski: Why don't those initiated cells revert?
Henry Pitot: We think that the promotion of these
spontaneously initiated cells is due to endogenous and dietary
factors. We have found that the normal chow diet is an
effective promoting agent for the liver. We think it might re
the plant estrogens that are present which vary with the time
of year. Semisynthetic diets eliminate much of this effect. I
don't think the skin people have that problem.
Roy Albert: Can you compare the response to phenobarbital in
rats with different background occurrences of liver tumors so
that you can see whether or not the logic of the system holds
up - i e., if you have a higher incidence of spontaneous
tumors, it might suggest a higher amount of initiation and you
would expect a larger yield with a given dose of
phenobarbital?
Henry Pitot: The number of spontaneous foci is 3 or 4 orders
of magnitude lower than what you would get when you initiate
with an agent. So it doesn't really contribute anything when
you do the experiment. It looks as if the Fischer rat has a
higher incidence of spontaneous initiation than the Sprague
Dawley rat.
Roy Albert: The whole tenor of what we are saying is that
promoters act on spontaneous initiation when you don't give an
initiator and that is why you get tumors. So it ought to
follow that the more spontaneous initiation you have, the more
tumors you have.
Henry Pitot:
initiators?
Can you use the model for single doses of
Dan Krewski: Yes. One thing we are looking at is what are
the effects of changes in the dosing pattern over time on
carcinogenic risks.
Bob Langenbach: At a recent meeting at NIEHS, Kinzel
reported that~pretreatment with TPA followed by an initiator
some weeks later and then a second-stage promoter increased the
number of tumors (Furstenberger et al., 1985). This may mean
that there is a memory for the TPA treatment.
Tom Slaga: We have repeated these experiments up to 10
weeks. The memory has nothing to do with cell proliferation.
By all indications, the skin is back to normal.
D-21
-------
Bob Langenbach: This memory for promoter treatment could be
an exception to our definition of promotion. Secondly, our
definition assumes that carcinogen-induced initiated cells and
spontaneously induced initiated cells are both promotable. Are
all initiated cells the same and do they all respond in the
same way to a given tumor promoter? Even among very potent
!2T™£!?0te™J ^!re are P°ssiblY verY different mechanisms.
For example, TPA interacts with protein kinase c and also
inhibits metabolic cooperation, but TCDD does not do well for
?Jo heF.one- .;"s° assays for tumor promoters (in addition to
the skin and liver systems), which take into account all
possible mechanisms, are needed. Within the NTP bioassay, it
may be possible to use a universal initiator where the test
SrSSSli/0" ld also be.used with a group of animals that were
previously initiated with a universal initiator.
Alternatively, stop studies may be useful to see if tumors in
test-chemical-treated animals can regress ..... [Presents his
premeeting comments. See Appendix c.3 However, the
identification of nonpromoting chemicals is difficult, which
may be a reflection of our level of understanding of
promotion, in recent studies with NIOSH, we attempted to
identify chemicals that did not have promoting activity to be
Sh^n £oVailda^ing Trasko's V79 metabolic cooperation assay.
When we took into account results from the skin, liver,
mammary, lung, and colon systems, we were unable to come up
with a chemical that was not a tumor promoter in at least one
of those systems. For example, phorbol - the standard
SS2?f°m0t!r *? fcSe ?kin s*stem - is a promoter in a mammary
system and a leukemic system, so, I don't think looking at
Dust one system is sufficient to determine whether a chemical
is active or inactive as a promoter.
Tom Slaga: I think there are ways of determining
experimentally whether something is acting as a initiator or
Du?e°in^^LieaSt theoreticallY- But if we don't have known
pure initiators or promoters, this would be difficult to do.
Bob Langenbach: it seems to me that, if nothing else, the
spontaneous level of tumors, which TPA will promote, will
always interfere with a final statement of unequivocal as an
in=!ia^°5' ^S Z undersfcarid risk assessment, there is a need to
quantitatively say that there is no initiating activity and
this may not be possible at present.
™ . T°Sether' all the studies suggest that in general
TPA will give you some papillomas and even fewer carcinomas.
[DISCUSSION OF IVERSON/HUBERSON DATA]
Roy Albert: Do you find initiated foci in human liver?
D-22
-------
Henry Pitot: In patients that have either hemosiderosis or a
hemachromatosis, you can see focal areas where the hepatocytes
have very little iron pigment, but these foci don t look nice
what we find in rats.
Roy Albert: Is this good enough to use from an
epidemiological standpoint?
Henry Pitot; You'd have to take multiple sections of the
liver which might be difficult to impossible.
Peter Magee: [Presents his premeeting comments. See
Appendix C.3
TAPE 9
Roy Albert: Has anyone tried treating mice with antipane
(sp?) to see if you can eliminate the TPA tumors.
Tom Slaga: Other protease inhibitors have very little effect
on this initiation if given at the time of initiation. But
whether they reverse initiation later in time has not been
looked at.
Roy Albert: Is the nature of the lesions produced after PB
application the same for initiated and noninitiated cells?
Henry Pitot: The foci look the same.
Henry Pitot; I will talk about how one might quantitate the
three stages in the liver system. You can in the liver
quantitate the potency of a single chemical for the different
stages. The system we use is analogous to the skin with one
exception: initiation in the liver must take place during cell
proliferation. We stimulate cell proliferation using a partial
hepatectomy. Then we administer diethylnitrosamine (you can
use many other agents), followed by the promoter usually
continuously in the diet or drinking water or gavage. We look
at lesions by using histochemical markers and in situ
hybridization looking at oncogene expression.
Using computers, you can quantitate the foci. We use three
different markers. We have found foci within foci (very
occasionally). If you do the IPI experiment, the foci within
foci (where one marker clearly only involves a small part or
the population of the focus) increases by at least an order of
magnitude. We interpret this as a transition from promotion to
progression because these are morphologically carcinomas,
whereas the focus is quite different. In the future we hope to
quantitate the foci within foci, which should allow one to
determine in a quantitative manner the transition from
promotion to progression.
D-23
-------
In the liver system you can develop an initiating index and a
K^fiS9 Jndr WhiCh is a Way of ^antitatiSg the potencj of
these two factors. The initiating index is the log of the
number of foci (corrected for the background level? per liver
per milliinole of the compound given in a single dose! The
initiation index for TCDD and phenobarbital is zero. [He
iJdexSfS%Se IilndeX for.other chemicals.] The promoting
index is the volume occupied by the foci in the liver in the
presence of the promoter divided by the volume of the foci in
SSSir 2nCe °f thS Pfomoter Per millimole of the agent per
?h! ? a'6 measurinS the ability of the promoter to expand
the population, it is relatively independent of time. [He
e J-inde? f°r Vari°US agents and answe^ ques?!ons
,indlces;l You have to take the threshold into
u Ifutil\rate of promoter administration does not
indx threshold level, one cannot calculate a promotion
Freddy Horaburger: HOW are the foci counted?
Henry Pitot: - This is a computer plot of three serial
hM^n* *tained fof three different markers. YOU can overlay
and thS vo?^minVhK Phen°tyPe °f each focus and the number
™ J ?u lumu* We have shown that ifc is invalid to simply
count the number of foci when, as in most cases in our studies,
they have unequal diameters. YOU must look at the &cuaies'
foci/ volume. Area and volume occupied by foci are equal so
you can get the index from the area. The number If foci
depends somewhat on shape. Most are spherical. A few are
ellipsoid ..... Different phenotypes grow at different rates.
TAPE 10
Skin is that you can"ot identify the
cells, m the skin, you could get a
promotion index based on the papillomas.
looked at the effect of the
Sn^flf ^wtT! extenslve1*- W. speculates on
COFFEE BREAK
Herb Rosenkranz: I will describe how you would go about
setting up a test battery to identify promoters. [He presents
his premeeting comments, see Appendix c.] presents
initiators?
°f
noncarcin°9enic mutagens pure
D-24
-------
Herb Rosenkranz: This was looked at initially but it was
said that they hadn't been tested adequately We
systematically went through all mutagenic and genotoxic
potencies using the TD50 values accomplished by Gold et al.
(1984- A carcinogenic potency data base of standardized results
of animal bioassays; Environ. Health Persp. 58:9-319) about a
year ago on 800 chemicals. We found no correlation between
carcinogenic potency and short-term test potency. There is a
correlation between carcinogenic potency as expressed by the
TD50 and the +/- response in some of the short-term tests. For
example, UDS is only responsive to potent or moderate
carcinogens. Some tests respond to weak and moderate
carcinogens and also noncarcinogens (Ennever and Rosenkranz,
1987; Mutagenesis 2:39-44).
Roy Albert; Has there been any_attempt to relate
carcinogenic potency to initiation?
Herb Rosenkranz: Not as far as I know..
TAPE 11
Bob Langenbach: As I said earlier, one problem we have had
in selecting chemicals to put in in vitro assays for tumor
promoters is identifying chemicals that have been adequately
studied to show that they are not tumor promoters.
Herb Rosenkranz: ...We combined the data base of PAHs and
some standard tumor promoters and we did find some structure
among the PAHs (and it is not the bay (sp?) region) which
appeared to be correlated with promoting ability. So there are
structural determinants which contribute to promoting ability.
Jim Trosko: There are many different mechanisms for
mutations (e.g., aneuploidy vs. point mutations).
Roy Albert: How many chemicals does it take to build up an
understanding of structure and promoting activity relationship?
Herb Rosenkranz: If you use the computered automated
structure evaluation (CASE) system and if you are dealing with
a single endpoint that is measuring a finite biological
relationship, then a data base of about 15 active and 15
inactive chemicals is sufficient to give you high reliability.
If you are dealing with a multifunctional event such as a
mutagen or carcinogen, you need 50 to 60 chemicals and even
then the probability of being correct is not as high. ...we
have the positive chemicals, we don't have the negative ones.
Jim Trosko: I have said that everything will end up being a
tumor promoter under the right conditions. [He talks_about two
similar chemicals that behave differently: 2,4,5-2,4,5
D-25
-------
nnnh VS; 3 ' 4 ' 5-3 ' * ' 5-hex . ] if you test in high
enough doses in a cytotoxic protocol, you find that chemicals
that do not otherwise show activity then show promoting
activity. _....! have no reason to doubt that anything that is
cytotoxic in the liver is a promoter. That is my gut feeling.
FEBRUARY 4
MORNING
Tom Sla
-------
study. Chrysarobin is probably also good, since you can induce
tumors with low doses in the skin. Sodium phenobarbital would
also be good. We need to look at these compounds and other
promoters in different tissues, as well as distribution,
metabolism and pharmacokinetics.
We don't have any pure promoters. The greater the potency of
tumor promoter, the greater the possibility that you will get
some tumors.
Curtis Travis: The question is what data do you need to do a
risk assessment.
[DISCUSSION OF TCDD]
Roy Albert: For risk assessment purposes, if something is a
promoter, then there may be threshold below which it may be
safe. The implications of what we are talking about are that
for compounds that have more than one type of activity, we can
take apart the various components and consider them separately.
Curtis Travis: In risk assessment, we are interested in
prevalent environmental chemicals. Those are the chemicals
that we should be looking at. We now think that every chemical
has some initiation and some promotion properties. We must
find a way of separating those properties for risk assessment.
I think a chemical like TCE is acting mostly through promotion,
so that at low doses most of its activity would discippear. If
it's an initiator, then there is no threshold.
Tom Slaga: Unless you have a real data base and extensive
dose-response then you can't make judgments about other
chemicals.
Curtis Travis: Let us study environmental chemicals and get
a data base for them.
Eliezer Huberman: Is it true that promoters have thresholds?
Tom Slaga: This assumption is based on a limited amount of
data. We need more dose-response data.
Roy Albert: The argument is made on the basis of
reversibility. Presumably reversibility is due to the fact
that the effects of the agent are countered at some rate. As
the dose gets low, the rate of producing the effect is lower
than the rate of counteraction, so you get no effect.
TAPE 13
Henry Pitot: The basis for the no-threshold in the DNA
damaging effects is also purely theoretical.
D-27
-------
Dan Krewski;
promotion?
Could we look at Dr. Slaga's definition of
Tom Slaga: We have added "growth in differentiated related
genes."
Henry Pitot: it leaves out the point of reversibility.
Jim Trosko: Modulation implies an epigenetic mechanism that
occurs all the time in normal cells, e.g., during
differentiation of the cell cycle, it's a fact documented in
the literature that those chemicals identified in promotion
systems do have that property to turn on and off genes through
nongenetic mechanisms, it is modulatable and reversible.
Dan Krewski: if we are going to try to associate the notion
of threshold with promotion, we don't want to have a definition
that implies that promotion can occur through genetic
mechanisms. I got a sense from yesterday's discussion that
there may well be a genetic component associated with
promotion.
Anne Kennedy: Six of the tests in Dr. Rosenkranz's Table 2
(in his premeeting comments) that are used to identify
promoters are based on a genetic endpoint.
Roy Albert: This give promoters a little initiating
ability. Does the issue of cytotoxicity fit in with your
definition?
Tom Slaga: Yes it does. You can bring about selection in
different ways, e.g., direct effect on the initiated cell,
increase or decrease in differentiation, and selective
cytotoxicity.
Roy Albert: Selective toxicity allows the outgrowth of cells
that are resistant, but that doesn't fit the definition of
reversible gene expression.
Tom Slaga:
toxicity.
The modulation of growth can be by selective
Jim Trosko: When cells die, they release compounds that
trigger the surviving cells to go into wound healing.
Tom Slaga: We can still modify my definition to take into
account reversibility.
Bob Langenbach: I think we are trying to put some of these
things into categories for which we don't really have the
information. There are also cytotoxic chemicals that aren't
tumor promoters so cytotoxicity alone is not sufficient.
D-28
-------
Henry Hennings: "Selective" is the important factor.
Bob Langenbach: I don't think we can rule out unequivocally
that some promoters may cause genetic damage.
Eliezer Huberman: Bob raised an important point. Peter
Cerutti and others assume that genetic damage is a critical
component of tumor promotion.
Henry Pitot: The oxygen radical effects are most' likely
working at the interface between promotion and progression.
Progression involves DNA damage/alterations in a major way.
Oxygen radical effects are clearly indirect. Whenever you have
an indirect effect, you will also have a threshold. So one
could argue that the oxygen radical effect will show the same
sort of thing. I think this effect is most likely working to
take a cell in the reversible stage of promotion and place it
into progression. Benzoyl peroxide and hydrogen peroxide are
probably the best current examples of progressors....The
concept of stages of promotion is unique to the skin. In
multistage carcinogenesis in rat liver no such stages have been
identified.
Roy Albert: It's my impression that reversibility has not
been tested on a scale that would permit it to be evaluated
adequately. For instance, when you stop promoting with TPA in
the skin, you get regression of the papillomas, but there is
something left there, and when you start TPA again, the lesions
pop right out. In the liver, has the reversibility been tested
thoroughly?
Henry Pitot: Not with numerous compounds. Yes with
phenobarbital, and the choline-deficient diet, and AAF as a
selecting agent. In the liver, as I think in the skin, the
stage of promotion in all models of the two-stage phenomenon
can be shown to be reversible.
Eula Bingham: What are the data on reversibility in the skin?
Tom Slaga: This is difficult to answer because so many
different mouse models have been used. Only GDI has been
extensively studied for reversibility. Initially you get
complete reversibility, but if you stop promotion after about 4
to 8 weeks, there is a little bit that is irreversible because
* you have set some of the initiated cells to the point where
they will go all the way through. A few studies have been done
in which promotion was stopped after say 4 to 6 applications
and then restarted 6 months later to see if you can recoup the
effect. Those tumors appear much more quickly so there appears
to be a residual effect.
D-29
-------
Henry Hennings: I'm not sure we would expect a papilloma to
regress all the way back to a single initiated cell, it
probably regresses back to a micropapilloma.
Eula Bingham: We don't have much data on regression.
Torn Slaga: The skin is at a disadvantage over the liver in
that we can't detect initiated cells which is the only way to
really quantitate promotion and its reversibility.
Eliezer Huberman: in the second step of promotion, do you
see the same foci as those that appeared the first time?
Tom Slaga: We can't tell this in the skin.
Bob Langenbach: Could you explain how regression occurs from
a full-blown papilloma?
Tom Slaga: I don't know any study that has done the
extensive histology necessary to determine this. ...When you
do a progression experiment, where we induce a fixed number of
benign tumors, the chances are that about 30% might regress.
When you give a progressor, in every case, all tumors that
become squamous cell carcinomas from the papillomas decrease in
size before they become , and that is why we think that
cytotoxicity is involved there. You decrease the tumor size by
greater than 50%. When it reaches the smaller size, then it
looms up into a carcinoma. (The promoter is not still being
given when you give the progressor.)
Roy Albert: Isn't there a major difference in reversibility
in the liver and the skin? In the liver, you get hyperplastic
nodules at the expense of the normal liver and when you stop
promotion there is remodelling and the loss of cells.
Henry Pitot: That's the picture in the Solt-Farber liver
model. In our system and Shulte-Herrmann's system and probably
Peraino's system, it appears that the cells in the promoted
foci die in the absence of the promoter.
Jim Trosko: I think we should discuss some of the three to
five current models to explain the mechanism of promotion. I
have identified at least 12 areas that I consider gaps that we
should discuss to deal with promotion. I would like to
introduce three brand new technologies that I think will give
insight into testing the three to five models.
One hypothesis is the activation of the protein kinase C
molecule by a variety of tumor promoters. Another is the
Troll-Cerutti model - the prooxidative oxygen radical model.
These are biochemical and molecular models. A third model, a
cellular model, that we have presented...is the inhibition of
D-30
-------
gap junctional intercellular communication by chemical tumor
promoters. It is possible that the molecular/biochemical
models can be integrated with the cellular model, since there
is evidence linking PKC with cell-cell communication. There
are also genetic models, e.g., the genetic recombination
model. I would like to present the cell-cell communication
model.
TAPE 14
The crux of that model is that once the initiated cell is
formed, it is surrounded by and communicating with its normal
neighbors. This introduces a higher order of biology into the
Moolgavkar two-stage model because the phenotype and future of
the initiated cell will depend totally on the communication
properties of the normal neighbors. In the literature, there
are a variety of mechanisms by which one can clonally expand
the initiated cell, simply by removing the suppressive,
contact-inhibiting effects of the normal neighbors, by
wounding, surgery, physical irritation, etc.
We now have direct evidence that many known growth factors work
by blocking contact inhibition. A few of the growth factors
have been shown to have promoting properties. We have tested
over 100 chemicals. We tested several chemicals in vitro and
later in vivo and we were able to predict that these would be
promoters.
You can also clonally expand an initiated stem cell by putting
a solid (e.g., plastic or metal) next to an initiated cell.
That solid is not communicating so expansion is not inhibited.
Promoters may also cause normal cells to proliferate. The
difference is that the normal cell after proliferation goes
into terminal differentiation. Initiated cells can't do this.
Once the critical mass of the initiated cells gets large
enough, the suppressing effects of the normal neighbors will be
diluted out. This is the essence of the cell-cell
communication model. We have developed a short-term test to
test this model. This assay has more applications than just
tumor promotion.
The process of intercellular communication is mediated by a
structure called the gap junction found in virtually all normal
cells in every organ. Small molecules below 1,000 daltons are
transferred between these gap junctions. Cells in tissue that
are coupled by gap junction have all their critical molecules
and ions below 1,000 daltons in equilibration. Gap junctions
are modulated by drugs, food additives, nutrients, endogenous
growth factors, biological toxins, pollutants,
neurotransmitters, hormones, heavy metals, etc.
D-31
-------
It appears to us using in vitro models to test cell-cell
communication that there are at least three classes of
promoters: those that have receptors and work at nanogram
levels (hormones, TPA, TCDD, etc); those that don't seem to -
need receptors (DDT, PBB, etc.) but just melt into the membrane
because they are lipophilic - they usually work.at microgram
levels; and those that don't need receptors but are not
lipophilic (saccharin) - they usually work at milligram
levels.
How does this link to gap junctions? Four intercellular
'^second" messengers seem to be responsible for modulating gap
junctions. One is PKC. One is calcium. One is pH. One is
cyclic AMP. The first three close gap junctions. Cyclic AMP
increases gap junction function in certain cells.
There is a brand new technology which I think can be used to
test thresholds, reversibility, synergisms, antagonisms that
have been speculated in the animal promotion model. This
technology enables us to bypass rodent cells. We can use human
cells - any human cell. We call this technology the scrape
loading/dye transfer technique, it is extremely simple. You
grow cells to confluence, which we feel mimics the normal
situation in solid tissues. We take two dyes - lucifer yellow
and rhodamine red dextran - and put these on living cells.
These won't go in the membranes of living cells. The yellow
dye is a small molecular weight dye so that when it does get
into the cell it can easily go through the gap junction. The
red dye is too large. NOW we scrape the cells, which is just
like wounding. The dye will go into the cells along the wound
line where the membranes are temporarily disrupted. The
membrane heals within milliseconds, trapping the dye in the
cells along the edge. You dump the dye out, wash the cells,
and immediately put it under a fluorescent microscope with two
filters, take a picture and you will see both dyes at the
edge. You then put the cells back in the incubator with and
without a presumptive modulator of cell-cell communication.
Five minutes later you take another picture of the cells. If
the cells have good gap junction function, the yellow dye will
diffuse away from the edge but the red will not.
We can use this to study dose-response. You can see a clear
no-effect level.
Drs. Lowenstein and Borek pointed out about 20 years ago that
most cancer cells seem to have defects in their gap junction
communication. This technique corroborates this. Tumor cells
don t seem to communicate at all. some oncogenes seem to block
cell-cell communication when they are expressed in the
appropriate cell.
D-32
-------
There aren't many synergisra studies in the cancer field. There
are even fewer studies of synergism in tumor promotion. There
are studies on antagonism trying to block TPA action. We
believe that some promoters have their action mediated by PKC
(a protein kinase enzyme/ a phospholipid, calcium-dependent
enzyme). Therefore, if you have two chemicals - one of which
will stimulate the phospholipid component of the PKC and the
other which modulates calcium - you will have a much more
effective activator of PKC than either one alone. We conducted
an experiment to see whether synergisms could occur in
modulating gap junctions. We used phospholipid-activating
chemicals and calcium-modulating chemicals. DDT blocks the
efflux of calcium through the membrane. We then added TPA
alone and got a nice dose-response curve. Same with DDT. We
then held TPA constant and added DDT over the same dose range.
They showed synergistic rather than additive effects. We also
found synergism between unsaturated fatty acids and DDT. I
suggest that a study of synergism of promotion in the liver and
the skin be done. PKC is present in both. I would say, do DDT
and TPA in the skin, and I would bet you would get art effect
level The other two assays are photobleaching and a
biological assay to measure communication. We have tested the
metabolic cooperation assay for V79 in over 100 chemicals and
we have found that it takes at least 3 days in V79 before the
cooperative donor cell can die or be rescued. We have found
that, in the scrape loading/dye transfer assay, TPA will block
communication in liver cells, but only for an hour or two. in
V79, it takes days. So these assays measure two different
responses: a transient response and a more long-term
response. Also, the V79 system uses serum which contains
growth factors so we think that the assay may not reflect in
vivo conditions. We think the scrape assay is a better mimic
of in vivo conditions because we can use serum-free media.
Roy Albert: How big does a clone have to be before it's free
of cell-cell communication?
Jim Trosko: That is an important question but I don"t think
anyone can answer it at this time. Not all cells have the same
number or size of gap junctions. We don't know what the
diffusion-suppressing molecule(s) is or are. In cells that
communicate well, we can determine the rate of communication
(i.e., dye diffusion) which varies between different types of
cells. We can quantitate all of this with a laser machine.
Henry Pitot: In the initiation/promotion system in the
liver, the lab chow diet appeared to have a synergistic effect
with promoters. It is possible that these diets may be
synergistic with many compounds tested in chronic assays.
We always test compounds in their crude state because that
is the way they are in nature. The diet may be "natural but
it is not like the human diet. Secondly, you are not answering
D-33
-------
the question of whether a compound is truly a promoter or has
to act with something in the diet. This is a question that the
regulatory agencies will have to deal with.
Jim Trosko: We have tested synergism with DDT and aldrin and
found that they were additive. We also added quercetin, which
is an inhibitor of PKC and we completely blocked the TPA effect
on cell communication.
Tom Slaga: How long do you wait after you scrape with the
toothpick? Do all cells take an equal amount of time to repair
this damage?
Jim Trosko: We haven't studied the latter. The loading time
is extremely rapid because cell membrane healing is a very fast
process. Not all toothpicks will work the same. Some cells
lay down collagen and fibernectin. So if you use the blunt end
of the toothpick you may lift up the collagen but you won't
affect the membrane. ....All of these studies are done at
noncytotoxic doses. You know if you have reached the cytotoxic
when dye goes into all the cells.
Bob Langenbach: Have you done this with mouse dermal cells?
Jim Trosko: To my knowledge at least 75 different cell
strains and lines have been used. But we are just using human
cells.
Bob Langenbach: I think it would be a good opportunity to
test whether the system is doing what you think to test mouse,
rat and hamster dermal cells and see if the blockage is related
to promotion because you have sensitive and resistant species
there. It would be interesting to test mouse, rat/ and hamster
dermal cells.
Jim Trosko: We have done that I challenge whole animal
people to do it in whole tissue.
Bob Langenbach: How do you limit your doses?
Jim Trosko: There is a built-in visual demonstration of the
cytotoxic level.
TAPE 15
[DISCUSSION OF c-Ha-ras ONCOGENS)
Henry Pitot: We have not seen any transcriptional activation
of the protooncogenes - £-my_£, c-Ha-ras or Ki-ras - in foci or
nodules. We have seen it in carcinomas. With one exception,
mutational activation of protooncogenes in rat
hepatocarcinogenesis has been either nonexistent or only
D-34
-------
occasionally seen. But you can get muta'tional activation of
the c-Ha-ras gene in carcinomas of the mouse liver. It has
also~~b¥en shown in mouse adenomas. Recently, we have found
fairly consistent transcriptional activation of the c-raf gene,
both in nodules and in carcinomas in the rat liver. One may
demonstrate that some but not all foci exhibit transcriptional
activation of the c-raf protooncogene by in situ
hybridization. We are interested in what the phenotype of
these foci is. Some foci show a lowering or absence of the gap
junction protein by the immunohistochemical technology,
although some show normal levels. The question is, if we do
the overlays, do those that express c-raf also have a low level
of the gap junction protein?
Tom Slaga; We have looked at the expression of several
different oncogenes by promoters in mouse skin in vivo and do
not find any change except from benign papillomas. There are
studies that activated £-Ha-_ras will lead to papillomas if you
get it in by skin scraping and then give a tumor promoter, but
I don't know of any studies that suggest that the oncogenes are
involved in vivo in terms of promotion....There has not been
increased expression of c-Ha-ras in epidermal cells before
tumor formation.
Henry Heimings: Our studies indicate that c-Ha-r_as_ can be
the initiating step or the malignant conversion (progression).
[MORE DISCUSSION OF C-Ha-ras]
Henry Pitot: I think the evidence in the oncogene work is
preponderantly that activation of oncogenes by mutation or
transcription does not occur during promotion. However, the
work in the skin by Balmain leaves it open as to exactly when
you get mutational activation of at least the c-Ha-ras gene
during multistage carcinogenesis in the skin. There also may
be some question in the mouse liver. But by and large, the
data indicate that promotion does not involve activation of
protooncogenes Studies suggest that most become activated
during progression. Balmain's and the Miller's work suggest
that there is a mutation of .c-Ha-ras. that probably occurs
during initiation. But is that the direct cause of the whole
process? I don't think that has been shown.
COFFEE BREAK
ROY ALBERT SUMMARIZES
i
1. The concepts of progression and progressors were
introduced. There was general agreement on them. This is the
first time that I am aware of that there has been such general
agreement on the use of the term and its consideration for risk
assessment.
D-35
-------
2. There was consensus on the definition of promoter - that it
is a reversible alteration of gene expresion and the reversible
expansion of initiated cells. We didn't define initiation or
progression. Problems with the concept of mechanisms of
initiation may make that definition difficult.
3. Another important point that was brought out was that one
can quantitate initiation and promotion and possibly
progression. Possible every agent could have properties of
each of the three functions. Different agents would have
different balances of these activities. This could be an
important consideration in risk assessment in terms of the
characterization of agents, particularly if their activities
could be quantitated, so one might express potency or relative
potency of the three types of agents.
4. We discussed whether reversibility is inherent in the
definition of promotion. This implies a threshold. There was
some concern that reversibility may not be complete in all
systems. There was a thought that reversibility may be
complete in the liver but not in the skin based on the
recurrence of promoted lesions.
5. We discussed synergism of promoters. This has not been
considered in risk assessment before, it may be an important
factor.
6. There was considerable discussion of the mechanisms of
promotion. We concluded that promotion doesn't involve the
activation of oncogenes although the evidence in the skin is
incomplete since you see activation of oncogenes in promoted
lesions but not in the skin itself. The presumption is that
the initiated cells had activation of oncogenes which are
demonstrable by clonal expansion but that is not really
testable unless you can get to the cellular level. Oncogene
activation does play a role in initiation and progression, at
least in the sense that a double dose of c-Ha-ras has been
demonstrated to push the cells towards maTign~anTy. Cell-cell
communication as a mechanism of promotion is also being linked
to oncogene activation.
°f initiation is not clear. Initiation has
^ a SUdden irreversible change which is
* *°?g period of time' However, there is
evidence that the initiated state is so common that the
ro»??eP£-J -an initiated cell as one that is mutated
^in terms Of the expected frequency of
«» ^ern.ative that has been suggested is that it
n ES h «S1£Je differe"tiation change in the cells. How
"ht?b°Ut and What ifc rea11* means is nofc
nnv-i « ly, initiation is thought to be linked to
genotoxic agents and yet the evidence doesn't support the
D-36
-------
notion that it is a mutation. If it is a genotoxic change that
is linked with a irreversible change in differentiation/ what
sort of genotoxic effect are we talking about?
8. There is now a mathematical model - Moolgavkar's - that
incorporates the concept of both initiation and promotion and
progression. It might be useful to test some of our biological
notions to see if the model expresses the biology.
BOB LANGENBACH SOLICITS COMMENTS FROM THE PANEL ON RESEARCH
NEEDS AND RECOMMENDATIONS. BASED ON THEIR COMMENTS, HE DRAWS
UP THE FOLLOWING LIST:
1. Species and strain differences
2. Other organs, colon, bladder, etc.
3. In vitro systems - human
4. Chemicals to use?
5. Differentiation, organ, chemical variation
6. Mechanisms, 02/ PKC, etc.
7. Animal models, short- and long-term compare to data from in
vitro
8. Oncogenes - role? antagonisms?
9. Chemicals for human hazard - promoters
10. Potencies: initiators, promoters and progressors
11. Benign: malignant, reversibility
12. Sequence of administration
13. Progressors
Freddy Homburger: We need more data on species and strain
differences in in vivo and in vitro....using inbred animals as
well as first generation hybrids.
Eliezer Huberman: (1) We need more work in I/P protocols for
organs other than the skin and the liver, e.g., the colon and
breast and other organs that are relevant to cancer induction
in man. (2) We also need to expand the number of chemicals
that have been shown to act as tumor promoters. (3) I suggest
that more emphasis be given to testing promoters in in vitro
cell transformation systems, especially those using human
cells. (4) Studies on modulation of cell differentiation
should also be included because there is evidence that certain
tumor promoters like phorbol esters and TCDD are effective in
modulating differentiation processes.
Tom Slaga: Since a lot of the data are based on phorbol
esters, I think studies should be done to find out why these
compounds don't work in other species or mouse strains. It may
be related to removing the ester groups. This is worth knowing
because then maybe much of the background data on TPA could
spill over into other systems.
D-37
-------
Jira Trosko: (1) We need to test the available hypotheses
that we have. The prooxidative model should be tested versus
the PKC model. There is evidence now that these two models may
not be mutually exclusive, i.e., one may affect the other. And
we should try to correlate those two models with the cellular
and genetic models that have been proposed, namely the
cell-cell communication model versus the recombination model
for example. (2) The role of intercellular in linking tumor
promoters to a mechanism of action in vivo should be
investigated, i.e., comparing the well-demonstrated in vitro
phenomenon with in vivo. (3) We should research the
correlation of oncogenes with tumor promotion.
Bob Langenbach: Certain oncogene-infected cells may be more
sensitive to tumor promoters. it is a good suggestion. Maybe
it is a kind of testing approach that we could utilize to see
if a chemical is really a promoter? Henry, in your earlier
papers you propose utilizing several systems in addition to the
animal bioassay for identifying tumor promoters and how much of
the chemical's activity may be due to promotion compared to
initiation.
Henry Pitot: I showed a slide of recent work based primarily
on the liver system. I think you potentially could get these
quantitative relationships from that system. You could get
systems, particularly in solid organs to do the same thing.
Certainly for promotion, you could do it in the bladder, skin,
colon. For surface organs it may be more difficult to
quantitate the initiation unless you develop spreads. Many
years ago, Roy Albert showed some beautiful pictures of spreads
with microlesions in them. These were animals that were
irradiated as I recall. I have often wondered if something
like that couldn't be used for bladder and colon where you
could see the early lesions and really quantitate initiation.
I think however that the chronic bioassay is essential to see
which tissues are involved, so I think the two have to be done
in concert.
Jira Trosko: if you could correlate, for example in the
Sencar mouse, sustained hyperplasia after TPA treatment with
the total absence of gap junctions as opposed to the
nonsustained hyperplasia in the Syrian hamster with the
presence of gap junctions, that is a test of a model. No one
has done that.
Peter Magee: Maybe we should focus our research on chemicals
that are of greater concern as human hazards because of their
higher exposure.
Dan Krewski: Maybe we should expand on the data base that
Dr. Pitot presented yesterday where we had a measure of potency
for the initiating and promoting activity of about a half dozen
agents. ...Get these values for more chemicals.
D-38
-------
Eula Bingham: We need to research the mechanism of
reversibility and the threshold question. Is there a threshold?
Bob Langenbach:
conversion.
We should also look at benign to malignant
Jim Trosko: Some of the best studied promoters - PCBs, PBBs,
phenobarbital, BHT and TPA - depending on the circumstances may
act as anti-initiators also. If given before the carcinogen,
they protect the animal. If given after, they promote.
Roy Albert: The issue of what is spontaneous initiation is a
good one. The relationship of cytotoxicity and promotion is
important from a risk assessment standpoint. Synergism in
promoters is another area for study.
Bob Langenbach: Let's begin discussion of item #1. What
kind of studies are needed to understand the differences
between the rat and the mouse?
Roy Albert: TCDD is a good promoter in the rat liver. Could
we develop a model using the human liver so we could compare
TCDD action in both systems?
Henry Pitot: We need to learn more about it. The most
reasonable way to do this would be to look at receptor-TCDD
interaction. The current argument is that all TCDD actions are
mediated through the receptor. But the affinity of TCDD is in
no way related to its toxicology. One can argue that maybe it
is the affinity of the receptor-TCDD complex for DNA. So a lot
of work is needed in this area.
TAPE 16
Tom Slaga: We need to look at why you don't get sustained
hyperplasia after repetitive treatment with TPA in rat,
hamster, etc. Differences from mouse.
Jim Trosko: I challenge whole animal people to take the gap
junction technology and see if it can be adapted to liver and
skin.
Tom Slaga: You have to be careful with cell culture systems
because they don't maintain what is going on in vivo. You can
put the liver cells into the spleen, the breast cells, etc.
Maybe there are systems like that where you maintain more of
the tissue relationship. Those systems should be studied
more.
Jim Trosko: Another line of investigation is the
epidemiological data on chemicals known to be promoters in
animals, such as the linkage of known saturated fatty acids
D-39
-------
h« !:urnors in rats and the change in the American diet
in the last few years, in the comparative models, we should be
studying chemicals that could be studied epidemiologically.
Henry Pitot: Alcohol is an interesting compound.
Epidemiologically it acts as a promoter. But only one or two
experimental studies have shown it to act as a promoter in the
rodent liver, it does not change the number of foci after
initiation. It changes their volume, it increases the number
of cells in the foci that are there. So this is an interesting
mechanism of action if ethanol is a promoter. This may be a
compound to study.
LUNCH
Hugh Spitzer: Are the dose-response and frequency issues
that were brought up yesterday the issues that we need to
address to begin to get the data we need for extrapolation
between species and to humans?
[DR. PITOT EXPLAINS TO DR. KREWSKI WHAT IS MEANT BY
RIvIRrST?BIIl]CTY' THERE IS FURTHER DISCUSSION OP TCDD. DR. MAGEE
ASKS WHAT IS MEANT BY GENOTOXIC. A DISCUSSION OF THIS
FOLLOWS.]
TAPE 17
Bob Langenbach: What other systems are available for study
besides the skin and liver? Is the bladder ready as a
screening system?
Henry Pitot: The bladder is complicated because the urine
itself is a promoting agent.
Bob Langenbach: Another problem is that the initiator also
induces tumors sometimes.
Tom Slaga: I think the bladder or maybe the colon is the
best third system to study after the skin and liver.
Henry Pitot: The thyroid may be a good system. There hasn't
been a good system developed that has a hormonal background.
Since many hormones may be promoters, an endocrine system might
be important to develop. The breast is possible but may be
complicated. The thyroid is nice because it is cellularly
homogeneous. The kidney is a possibility but it is cellularly
heterogeneous so the kidney tumors can be derived from many
different cell types, if each cell type has a different
response to a promoting agent, it could get complicated.
Roy Albert: The lung is notable for being bad, but it is an
important system.
D-40
-------
Tom siaga: The Nettesheim tracheal system is a nice one
because they have a denuded trachea which they can repopulate
with human cells and get human tissue growing there.
Herb Rosenkranz: I think studies should be conducted in the
lung because this is an important route of exposure for humans.
Freddy Homburger: I would suggest craighead and Mossman's
system of hamster trachea explants.
Hugh Spitzer: How will these systems help us for risk
assessment?
Henry Pitot: We need to find out if promoting
characteristics in the two systems are true in others. If not,
then we must rethink the whole thing. You can't base risk
assessment on two systems.
Herb Rosenkranz: The respiratory systems produce lesions
that are historically identical to human lung lesions. This
is not the case in the skin and the liver. So there is a much
closer morphological association.
Peter Magee: Methylnitrosourea might be considered as a
universal initiator. I can think of about five organs where it
produces tumors.
Bob Langenbach: If we had a universal initiator that worked
in several systems, we could test suspect promoters in these
systems or organs in the whole animal using the same
initiator.
Roy Albert: You could also use a cocktail of nitrosamines
that was designed to initiate every organ,
Dan Krewski: Aren't promoters initiator-specific?
Bob Langenbach: I think that is a fundamental question and
we don't have enough data to answer it.
Tom Slaga: I think the only place I know of whether they are
initiator-specific is the lung adenoma model. In the liver and
skin, you can change the initiator and the promoter still works.
Bob Langenbach: I think the consensus is that there is merit
in looking at the various organ systems to look at the
universality of the phenomenon in testing. ..I would like to
ask Drs. Huberman and Kennedy what they think the in vitro _
transformation systems might offer for identifying and studying
promoters.
D-41
-------
Anne Kennedy: I think the problem with testing for promoters
using in vitro systems is that promotion in vitro is highly
serum dependent. The action of many promoters depends on which
lot of serum is used. So lots must be carefully screened.
Tom Slaga: No one has shown a requirement for a promoter to
get cell transformation in an in vitro epithelial system,
especially the skin. This is puzzling.
Jim Trosko: YOU may be promoting just by the way you set up
the experiment in an in vitro system.
Bob Langenbach;
studied.
Let's talk about which chemicals should be
Roy Albert: Chlorinated solvents and pesticides.
TAPE 18
Tom Slaga: We need to test chemicals to develop the model
systems first, and then look at the ones of concern to humans.
For model development we would want to focus on chemicals for
which there is a data base and which are known to be promoters.
Herb Rosenkranz: We should look at the list of 50 or so
chemicals drawn up by Upton et al. (1984).
Bob Langenbach: From-that discussion, I ended up with TPA,
PB, PBB and TCDD.
Tom Slaga: I would add teleocidin and maybe chrysarobin.
Eliezer Huberman: I believe that one way in which some
chemical promote tumors is by altering cellular differentiation
processes. I would recomm'end that more work be done in this
area. The human myeloid leukemia cell systems are useful for
such studies because they provide simple assays for testing the
ability of tumor promoters to induce differentiation
processes.
Bob Langenbach: I am concerned about the possible role or
need for metabolism to manifest promotion.
Jim Trosko: RUSS Malcolm's study showed that there are
several compounds that don't block cell-cell communication, but
their metabolites do. My question is how relevant are the
metabolites. Under normal circumstances (where there isn't
cytotoxicity), I think the metabolites have as their target the
membrane or some cytosolic fraction. We could adapt the scrape
loading assay to cells that can metabolize agents like the
primary human keratinocytes.
D-42
-------
Do we have any promoter that has to be
Bob kangenbach;
activated?
Tom Slaga: I don't know of any. Unless you have a complete
carcinogen, that may have both an initiator and a promoter
among its metabolites. It could be instructive to study this.
Dan Krewski: A model that can lead to quantitative
predictions of risk at various doses would be a useful tool for
risk assessment for initiators and promoters. We talked about
normal cells that change to become initiated cells. The
initiated cells could replicate to form an expanded pool of
such cells. An initiated cell could undergo further
transformation to lead to a malignant tumor cell.
I was happy until the end of yesterday with thinking of the
first change as involving some sort of genetic lesion in which
there was mutation or DNA damage. I wanted to make that same
assumption for the second transformation. I wanted to think of
the process of expanding the population of intermediate cells
as being a nongenetic mechanism which would involve for example
recurrent cytotoxicity to stimulate cell proliferation or
perhaps the preclusion of terminal differentiation in order to
reduce the death rate of the intermediate cells.
First, I would like to ask whether the assumptions of which are
genetic and nongenetic are correct.
Roy Albert: Experimentally there is evidence that initiation
is linear.
Dan Krewski: We have to model the dose dependency of this
first and second transition rates. Most people who model
carcinogenesis would probably like to assume that they are
linear functions of dose. We also have to model the dose
dependency of the birth and death rates of the initiated
cells. I don't think they should be linear functions of dose
because they may have thresholds since we are talking about
nongenetic mechanisms.
How would we get at all the parameters in this model? First,
the age-incidence curve does not involve the birth and death
rates of unaltered cells (aL and b^ simply because we can
think of the original tissue mass as being sufficiently large
that it can be described as a deterministic rather than a
stochastic process. All that enters in is the number of cells
in that tissue as a function of time. So what we do need in
the way of data is information on the growth of the tissue of
interest as a function of age.
D-43
-------
Henry Pitot: if you have a fixed number of initiated cells,
does ui drop out? CJ.AO,
Dan Krewski: These are just constants that scale the
position of the age-incidence curve and they will be
dose-dependent. The higher the dose of the initiator, the
greater the incidence of lesions at the end of the process So
you would leave this constant in.
Jim Trosko: YOU are trying to set this up for the IPI?
Dan Krewski: That is a good question. I wanted to mention
another version of this model developed by Leon Ellwein which
essentially does the same thing but in much more detail, m
his version, the various transformation rates can be
time-dependenf.
TAPE 19
Dan Krewski: in the standard IP protocol, the first data
requirement is the growth rate of the normal cells within your
tissue. This is something you could get fairly readily
Second, you need information on the birth and death rate of the
intermediate or initiated cells. For that, you would have to
go the laboratory assays for promotion that we have been
discussing in detail. Third, you would need information on the
transformation rates for initiation and progression (UT and
h?,LuJ!ere ?°£ W°Uld ?ave t0 go to a 2-ygar' long-term rodent
bioassay. I haven't looked at whether you could get these
rates from a single standard bioassay or whether you would have
to use some kind of IPI protocol to factor out the progression
seep.
I(t) is the time rate of appearance of lesions in the
bioassay. You could (1) use skin and count them; (2) use
serial sacrifices; (3) assume that the lesion of interest is
rapidly fatal, in which case the survival time of the animal
would serve as the proxy for the actual time to tumor
induction; or (4) assume that death as a result of tumor
occurrence in a bioassay is independent of death from competinq
causes in which case, we could statistically separate out time
to tumor. These are fairly classical problems and we do have
approaches for them.
The idea of this model is that you get data on tissue growth
and cell proliferation from studies separate from the bioassay,
and then you factor in bioassay data. Knowing these
parameters, you try to estimate transformation rates and
possibly use an IPI to separate u^ and u2.
In the bioassay, unexposed controls would allow you to measure
spontaneous initiation. The spontaneous rate of tumor
D-44
-------
formation in the controls is not a measure of the spontaneous
rate of initiation, but we would get a measure of this
transformation rate from the tumor occurrence rate in the
control animals if we knew all of the other parameters in the
model from these separate experiments.
Little s is a dummy variable that indexes time going from zero
to t.
Henry Hennings: How do you take care of spontaneous
progression?
Roy Albert; In the presence of constant internal level of
promotion, the number of benign lesions that one might get
would be a measure of ulr but the rate of progression of
these benign lesions to malignant lesions would be a measure of
U2-
Dan Krewski: That is correct.
Roy Albert: For instance in the lung adenoma system, you get
a lot of benign tumors and few carcinomas. From that you could
get an independent measure of U]_ and u2.
Dan Krewski: I think the statistical community needs to
figure out exactly what kind of data we need to estimate the
specific parameters in the model - to separate U]_ and u2.
Herb Rosenkranz: Would the partial hepatectomy in the liver
system be a problem? Can you correct for that?
Dan Krewski: Yes, we would have to have information on 33
and b2 that would be relevant to the conditions under which
the bioassay would run. We don't do partial hepatectomy in
rodent bioassays.
Henry Pitots Partial hepatectomy will change only u^ in
our system. You can get around that by initiating during the
neonatal period for ui and if you do an IPI experiment, then
do a partial hepatectomy for U2. We administer a single dose
of the initiator 24 hours after hepatectomy. Then you can wait
a year to give the promoter. In Farber's selection protocol,
you are forcing cell replication in the presence of the
selecting agent so you are allowing the altered cells to grow
in the presence of an inhibitor of end cells.
Curtis Travis: I don't think the partial hepatectomy would
affect ui at all because that is the mutation rate which
should be constant per cell division. All the hepatectomy does
is increase the cell division which affects the birth rate and
x(s)...As I said yesterday, we took background cancer rates
[I(t)] in rats as a function of age. We found x(s) - the
D-45
-------
growth rate of the liver as a function of age - from the
literature. We also assumed that the initiated cells would be
growing at the same rate (a2 - b2) as the background cells,
estimates of which were obtained from the literature. We then
had all the parameters except ux and u2. We assumed that
these were the same and that they were 1Q-8. we ran the
model and exactly reproduced the age-specific incidence of
cancer rates in rats, we used the NTP data so we knew the
background rate down to about one-tenth of a percent, we did
humans°r ^ liver in rats' We wil1 be doin9 it in mice and
pan Krewski: To model dose dependency, we would then have to
have control and exposed animals.
Curtis Travis: For TCE, which was shown to be carcinogenic
in mice, we already have a dose-dependence for increase in cell
turnover rate, it has been measured at the levels used in the
animal bioassay. DOW Chemical Company measured the increase in
mitotic rate as a function of dose, it is a threshold
phenomenon. AS effective dose to the liver, it is zero for a
while, then you reach the threshold and it starts increasing
linearly with effective dose to the liver. I want to put these
data into this model. I will assume that TCE has no genotoxic
effects so ux and u2 would be the same as background, we
know the x(s) and I will put in (a2 - b2) from the data.
We should be able to predict the age-specific incidence of
cancer from the TCE bioassay. if we can, then that will prove
in my mind that TCE is working solely through a promotional
mechanism. I feel that we need just this kind of research in
tnis area - studies on particular chemicals, determine
increased cell turnover rates as a function of dose, plug them
into these models and see if we can predict the cancer bioassay
rates that have been observed.
Dan Krewski: if we wanted to do a full examination of
initiation/promotion we would have to do a bioassay involvinq
various levels of exposure to the initiator in order to
establish how these mutation rates vary with dose.
Curtis Travis: Several refinements need to be made, when
you are measuring increased cell turnover rates you are reallv
measuring increased turnover rates of the normal cells and not
the foci. I have assumed the turnover rates are the same.
Also, this model assumes that these rates are time-
independent. I think that the rate of foci growth will
increase with time. We need to do experiments where we look at
the livers at different times to see if those volumes are
increasing at a different rate.
Roy Albert: Does it make any difference that you get a
spectrum of initiation at least in the skin? Small doses or
D-46
-------
short-term application of promoters gives a papilloma that
shows a lower rate of regression than with larger or longer or
stronger doses. Does it matter that on the malignant side,
there is no objective measure of malignancy such as growth
rate? There is also a spectrum of malignant response.
Henry Pitot: If you use a sufficiently small dose of _
initiating agent, you won't get the spectrum. The same is true
in the liver. At high doses, you telescope the whole thing and
get a spectrum. If you want to model initiation and promotion,
you must make sure that when you initiate, that is all you do -
that you do not get any progression or promotion.
Bob Langenbach: in the models, do you really measure bi
and b9 and if so how? I think u2 is really progression and
not promotion. I'm not sure_that (a2 - b2) is an adequate
representation of the promotion.
Jim Trosko: [Draws diagram.] This is what I would refer to
as promotion (a2): the nonmutagenic event that clonally
expands all the initiated cells. On further promotion these
will give more and more initiated cells. Depending on where
you put mutagen 2 you should increase the target size of the
number of initiated cells. Now, presumably this occurs for
spontaneous tumors also. This model is a test of whether you
need two hits.
Dan Krewski: I agree completely with Jim. His expanded
model is the same as mine. ... By doing IPI you will estimate
U]_ and u2. The two mutagens don't have to be the same.
Jiin Trosko: The point Tom made yesterday is that the same
mutagen at ui may be very effective as an initiator but may
not be as effective as u2 because of the pharmacodynamics
after the clone gets large. That is why I suggested using
X-rays because you don't have to worry about metabolism or
selective mutagenicity of the cells.
Dan Krewski: If the same agent affects ux and u2 and w«
knew this for sure, I think that we could estimate the product
of those two from the IP protocol.
What kind of studies would I do? We would use doses of an
initiator of 0, 0.5 and 1 (think of 1 as the MTD) in a single
application. You would have chronic exposure of a promoter at
three doses - say 0, 0.5 and 1. This allows you to study
dose-response for the promoter and initiator and you could add
'a second initiator which may or may not be the same as the
first and this would be single administration in IPI protocol
at various times after promotion. You could have various time
lags. You could administer the initiator at various points in
the animal's lifetime. Basically, you would need several doses
D-47
-------
SL I initiator and of the promoter in order to model dose
dependency within the model, if you wanted to separate out the
rates at which the two mutations occur you would hJve to add in
the second initiating application.
occnrn' If spontaneous initiation is
occurring at a sufficient rate for your promoter to be
effective, then you don't have to administer an initiator...
TAPE 20
Freddy Homburger: what is the purpose of this model?
Dan Krewski
To predict effects at low exposures. But the
before it can be used for
Roy Albert: IPI experiments are expensive. Could they be
2S??,ii2 tissue culture? Are the clones produced in the tissue
culture promotion assay malignant?
Eliezer Huberman: it depends on the assay, in the hamster
embryo cell system, only a small fraction of cells became
malignant .
Roy Albert: if you further treat these colonies with a
carcinogen, can you make them malignant?
experiment."6
trled
Peter Magee: If the IPI were done, what agents would we use?
Roy Albert: I would use a direct-acting agent in both cases
because you cannot be sure of the metabolism. cases
I think ifc should be done systemically as an
to the normal NTP bioassay.
an ™*o i.Jn the S!?in' you could use MNNG as the' initiator
and TPA as the promoter and urethane as the second
initiator.
m SLARIFY THE MEANING OF THE TERMS IN THE MODEL
AND WHAT WOULD HAPPEN TO THOSE TERMS AND WHAT YOU MIGHT qpp TP
EXPER?MENT?]PES °P AGENTS ARE USED °N ^E THREE STAGES OF AN
Dan Krewski: We could call the IPI study IPP or IPC for
progressor or converter at the third stage.
END FEBRUARY 4
D-48
-------
FEBRUARY 5
MORNING - CONCLUSION
TAPE 20
ROY ALBERT SUMMARIZES
We talked mostly about long-term research that dealt with
mechanisms of action. In the area of initiation, we talked
about the issue of uncertainties in mechanisms>with respect to
irreversible differentiation versus the induction of mutation.
in the area of promotion, we discussed approaches in terms of
receptor binding with reversibility and the role ofc promoters
in terms of gap junctions and cell-to-cell communication.
We didn't talk much about the mechanisms of action of
progressors, except to say that progression is probably related
to DNA damage. The issue of whether there are agents that
don't damage DNA that can cause progression was left open.
Such agents would be particularly sinister because they
wouldn't necessarily show up.on bioassays.
We alluded to the importance of species differences in response
to initiation, promotion and progression. We need to
understand these differences from a mechanistic standpoint.
TAPE 21
be able to develop some basis for predicting whether humans
are going to respond to promoters. It seems that promoters
have more extreme differences in species and strain responses
than carcinogens, because strong carcinogens are notable for
attackinq multiple strains and species. This is not
nlcelsarlly the case with some of the promoters that have been
studied. This makes risk assessment even more difficult
because of the difficulty of extrapolating animal data to
humans.
The issue of synergism of promoters was discussed, with some
evidence for DDT and TPA. This opens up a new area, both in
terms of mechanisms and identification of the kinds of
promoters that are likely to interact.
The need for more extensive models capable of demonstrating
promotion was expressed, particularly models that are more
relevant to the major human cancers, i.e., colon and lung, to
enhance extrapolation of animal data to human data.
We discussed mathematical modelling of a two-stage promotion
model and ways and means of validating the model. Also, we
talked about whether tissue culture could be used for this
purpose, primarily by initiation/promotion and second
initiation studies.
D-49
-------
We discussed the need to improve in vitro screening models we
slllct fchf mot1S W°Uld be/ formidabl* undertaking nit only to
select the most appropriate in vitro tests but to set up
parallel studies in animals. However, many people think that
promoters can be as great or greater invirSnSental hazard than
carcinogens. So there was a feeling that this enterprise is
needed, it would probably be undertaken by the NTP rather than
We discussed how to quantitate initiation, promotion and
progression. This has been done in the liver. There was
^hiUenl: thfl: thi? should be extended at least to the skin and
?££ £ ? tissues. There is a fair amount of agreement
that agents can be promoters as well as initiators
Eventually, we may reach a point where we assess these
characteristics separately for each chemical.
We have essentially been talking about an NIH long-term
research program, if these areas were well funded, one could
expect results in 5 to 10 years that may or may not Ce
applicable, it leaves open the issue of what kinds of
short-term studies might be useful to EPA for risk assessment.
o£?MnY ih3?6 1S 1?w-dose extrapolation. There is evidence and
opinion that promoters are reversible with thresholds. The
question is how do you demonstrate this. The other issue is
whether an agent that shows behavior as a promoter in one
tissue also demonstrates that behavior in another tissue, even
in the same animal, and then whether there are differences
among species and strains. These are the gut issues of the
controversy of agents of current regulatory concern.
Bob Langenbach: We need more studies in rodents and human
epidemiology to help to better understand species differences
£™r* J^0^8. lunation of thresholds was very informative - .
how f a threshold for a promoter may differ from a threshold for
an initiator. We could use further discussion on how to
de,termin\such a threshold, and how such a threshold
! in risk assessment. we discussed mechanisms of
dit^nerS%S8peSially "«Pt«-mediated mechanisms, we
didn t discuss other types of promoters. Research is needed in
t0 i? V*tr° screening, I personally believe that
genotoxic systems we have today are not doing the
they were designed and promise'd to do. if the field of
T?erm *e*tS *S t0 confcrib*te to carcinogenesis screening,
will need in vitro and short-term models to identify
chemicals acting by a promotional mechanism that are not picked
up in the genetic toxicology systems.
We developed a list of chemicals to study: teleocidin,
phenobarbital, PBB, TCD, chrysarobin and, with some
D-50
-------
qualifications, TPA. We should also consider weaker promoters
such as saccharin. Chemicals to which humans are exposed and
for which there will eventually be epidemiology data should
also be studied.
Oncogene involvement and the relative potencies of promoters
versus initiators, either within the same chemical or in
different chemicals were not well addressed.
Another point to consider is how useful are models that only
give papillomas or foci in risk assessment? Do we need
carcinogenesis as a final endpoint in these systems?
I'm beginning to think that progression may be as important as
promotion. We need to further consider the
promotion/progression interaction and overlap in doing risk
assessment.
Bill Parland: I'd like the panel's responses to two
questions:
1) What data would allow us to determine that a chemical has
the ability to promote?
2) What data would allow us to determine that a chemical has
the ability only to promote?
The first information we need is that a chemical has the
ability to produce tumors in an animal system.
v,
Henry Pitot: Promoters are carcinogens, i.e., they cause an
age_Specific increase in neoplasms. The only question is the
mechanism by which carcinogenesis occurs. In answer to the
first question, anything that results in a neoplasm may have
promoting action.
Roy Albert: I disagree that promoters are carcinogens.
Promoters serve to expand the cell population at any early
stage of transformation before they are malignant, and it's in
that expanded cell population that you get progression toward
malignancy, so they heighten the likelihood of developing
cancer. If you have an agent that has been clearly
demonstrated to be a promoter with initiation/promotion
studies, and you think it's working by receptor binding, which
implies reversibility and the existence of a threshold, how do
you demonstrate this in a persuasive way? Since this is
reversible and doesn't include the tumorigenic process, except
by inference, if we had a decisive way to demonstrate this, it
would have an enormous impact on the risk assessment approach
to the evaluation of these agents, because it would imply the
use of a completely different extrapolation model than low-dose
linear extrapolation.
D-51
-------
Jim Trosko: If you could find a cell line that has an
EGF-receptive mutation, where it may bind but the signal is not
transduced, and you have the parent line where the receptor is
there and the EGF binds to it, in one case you should get a
response and not in the other. You have a specific molecule
needing a specific receptor for a biological effect, and if you
remove the promoting activity you would be proving the case
That would be better than using a drug that would interfere
with signal transduction because drugs usually have multiple
Herb Rosenkranz: Just looking for a binding site may mislead
you, since the site may have nothing to do with the activity.
Henry Pitot: What Roy was saying is how do we demonstrate
that that receptor binding is related to the response we are
talking about. I think TCDD has been demonstrated to work
through a receptor, with the possible exception of the thyroid
effect (if you remove the thyroid you decrease the effect) if
you eliminate the receptor it doesn't work. The receptor is
genetic. Some animals have it, others don't.
Roy Albert: can you take the animal with the receptor and
inactivate the receptor?
Henry Pitot: You can lock the receptor up with an
irreversible inhibitor.
Roy Albert: I think this is an important area for study.
TAPE 22
Tom Slaga: All promoters that have been looked at in detail
do have some carcinogenic activity. But I think that in most
cases where they have been extensively studied, they do not
show a dose response, if it does not show a dose-response,
then if you have spontaneously initiated cells, you express
those relatively easily so you don't have a dose-response by
the so-called carcinogen as a promoter.
Henry Pitot: TCDD, phenobarbital and saccharin have
dose-response.
Tom Slaga: If you have a finite number of spontaneously
initiated cells, those are going to saturate fairly easily.
Henry Pitot: At the maximally tolerated dose, I agree.
uu Weinsfcein defines tumor promoters as compounds
that have weak or no carcinogenic activity when tested alone,
but result in markedly enhanced tumor yield when applied
repeatedly following a low or suboptimal dose of a carcinogen
D-52
-------
initiator. Is everyone happy say that promoters are
carcinogens?
Jim Troskos If we all agree that carcinogenesis is due to
initiation, promotion and progression, then we must decide
whether the mechanism for initiation is discrete from the
mechanisms underlying promotion and progression. But we don't
know the mechanisms yet so we can't say if they are discrete.
Henry Pitot: I don't see any problem with promoters as
carcinogens because I can't think of any known promoter that is
not carcinogenic by the definition that it increases the
age_Specific incidence of neoplasms in a set strain of
animals. Let's not destroy the original definition but rather
dissect it into it's original components..
Bob Langenbach; At an NIEHS meeting 3 or 4 months ago, the
consensus was that promoters are a class of carcinogens.
Implicit in the defintion we agreed to earlier at this_workshop
was that promoters were carcinogens because it was an increase
in the number of tumors in an animal which is very nearly the
definition of a carcinogen.
Roy Albert; The more important question is how do promoters
behave, in terms of whether you can extrapolate from animals to
humans, and whether or not one should use a threshold model.
Can an agent that acts as a promoter in one organ act as a
whole carcinogen in another?
Eliezer Hubermans We need to focus on what are the
mechanisms underly the promotional event.
t
Bill Farland: We have said that one of the mechanisms is a
receptor mechanism. There were two suggestions that one could
use mutation studies in vitro or in vivo and competition
studies - competitive binding studies - to look at mechanisms.
Are there any other suggestions?
Henry Pitot; To say that something is a promoter, these
characteristics should be present: 1) There is no direct DNA
damaging or altering effect. 2) In many cases a receptor
mechanism mediates the effect of the agent. 3) There is»a
maximal effect of the promoting agent in producing tumors
following initiation in the absence of toxicity (this is very
questionable but has been shown in the two major systems). 4)
There is an experimentally measurable threshold. 3) The
effects at both the cell and gene level are reversible. 6) The
effects of the promoting agent are modulated by environmental
means (e.g., aging, alteration of the hormonal environment of
the animal).
Roy Albert; I agree with this list.
D-53
-------
Bill Farland: This type of list provides a weight of
evidence for saying that something is a promoter.
Henry Pitot: Prom a practical standpoint, we must first
determine whether a substance is carcinogenic in a long-term
bioassay. Then you ask whether it is an initiator or promoter,
etc. And, if you want to be pure, you do it for each orqan
system. ^
Bill Farland: in different tissues, you may be looking at
different mechanisms of promotion.
Roy Albert: Radiation is an example of an agent that acts as
a complete carcinogen in some tissues and as a promoter in
others.
Eliezer Huberman: Hormones are examples of substances that
have receptors in various parts of the body yet may affect
these different parts in different ways.
Roy Albert: How do we show the existence of threshold?
Henry Pitot: in the liver you have a baseline level of
foci. YOU can go to a level of no change in phenobarbital and,
possibly, dioxin.
Jim Trosko: Cell-cell communication may play a role in tumpr
promotion with regard to thresholds. Cells in a 3-D tissue are
surrounded by other cells and communicating with them by
hundreds of gap junctions, if we assume that knocking out a
gap ^unction by one molecule - a promoter - is a one-hit event,
then we have an explanation for thresholds.
Dan Krewski:
threshold?
Shouldn't we use the term NOEL instead of
Bill Farland: Are we able to say that threshold in the liver
is not a pharmacokinetic situation - that you are not dealing
with dose to the liver - in terms of demonstrating the
threshold? Are we talking about the dose to the liver or to
the animal as being a threshold?
Henry Pitot: Eighty to ninety percent of ingested material
gets to the liver.
Tom Slaga: Regarding the promoter-only list, I would rank
the "maximum effect follows initiation in the absence of
toxicity" as one of the most important and put the concept of
threshold and "effects are reversible" as the key things.
Receptors are less important. 1, 2 and 3 are the key things.
Peter Magee: what is the criterion of direct DNA damage?
D-54
-------
Henry Pitot: That the compound itself chemically reacts with
DNA, intercolates with DNA or directly causes a cision of the
chain. You are dealing with one chemical reaction versus two
or more.
Bill Farland: If one wants to build characteristics for the
weight of evidence of something being only a promoter, the
evidence of no direct DNA binding is the strongest piece of
evidence you can have. But this is not sufficient to
demonstrate that an agent is a promoter.
Jim Troskoj What about chemicals that don't initiate but are
very cytotoxic and because of cytotoxicity also act as
promotersr e.g., 3,4,5-3,4,5-hexabromobiphenyl?
Henry Pitot: You probably can't use highly cytotoxic
compounds because cytotoxic effects will occur before you
express all the initiated cells. If you get into that hedge,
you have to eliminate maximal effect and go to the others.
Roy Albert: Why do you call it maximal? Isn't it because
you get a bigger response with an initiator than without?
Henry Pitot: No, you get a dose-effect with dioxin and
phenobarbital, but that dose goes up to a point and then
quits. I don't think that's true in the absence of toxicity of
a complete carcinogen. You keep initiating and getting tumors
until you get toxicity.
Dan Krewski: Shouldn't there be something in this list to
rule out the possibility of progression, perhaps a criterion
that would suggest the occurence of benign lesions rather than
malignant?
Henry Pitot: That's why I said no direct DNA damage. I
argue that indirect DNA damage produced by agents that have no
initiating activity is more closely related to progression. So
you will have a problem if you let indirect DNA damage be part
of the definition since the agent may have progresses activity.
Bob Langenbach: I'm not sure our definition for promoters
eliminates other types of epigenetic carcinogens as defined by
Gary Williams - chemicals that may cause gene amplification or
gene rearrangement and such events that also cause cancer but
aren't really initiators.
Henry Pitot: Maybe we should say no direct DNA structural
alteration. I would argue that gene amplification is involved
in progression.
Herb Rosenkranz: Hydroxyurea does cause direct DNA damage
but it also induces free radicals. So I think it fits in.
D-55
-------
TAPE 23A
Peter Magee: How would you measure direct DNA damage?
Henry Pitot: Types of direct DNA structural alteration are:
alkylation (adduct formation), intercalation (demonstrate by
spectral techniques), scission of the chain.
Bill Farland: The chemistry is the information that you
have. Do we have experiments that allow us to identify the
threshold phenomenon? The classic animal experiment is not the
one to pursue to demonstrate a threshold.
Henry Pitot: The demonstration of a threshold is always open
to question. You can demonstrate an experimental threshold
with many complete carcinogens, but nobody believes it's there,
for good reasons. With promoters, you can demonstrate there is
a threshold and argue that it is there. The liver system
allows you do this, and the skin and, I think, the lung (at
least with BHA). Certainly, in the bladder you get a
measurable threshold.
Bill Farland: How about the demonstration of maximal effect
in the absence of toxicity?
Henry Pitot: It has been demonstrated in the liver and
skin. I don't know if it has been demonstrated in othe'r
tissues.
Roy Albert: I think that another criterion for promoters is
the induction of benign tumors which take a long time to go to
carcinomas.
Henry Pitot: AS a pathologist, I object to this. What is
meant by benign tumor - a callus in the bone, a neoplastic
nodule in the liver, a papilloma in the s'kin, or an adenoma in
the liver?
Roy Albert: I am talking about a' benigr. precursor lesion
that goes to a malignancy.
Henry Pitot: I would not buy that, because once you have an
adenoma in the liver, for practical purposes it is a
carcinoma. Once you have a papilloma that has dysplasia in the
skin, you will also get carcinoma. You need to distinctly
define benign tumor, both with respect to morphology and to
natural history. A reversible benign mass of promoted cells
would be fine as a definition, but I don't think you can do
that on a morphologic basis. Morphology cannot distinguish
between a lesion which is still reversible and one which is
permanent.
•D-56
-------
Bill Farland: What are some other characteristics that one
would look for to describe reversibility?
Henry Pitot: One characteristic implied in that definition
is that the biochemical effect on gene expression of that
compound is reversible. That can be easily measured.
Bill Farland: What about in vitro systems?
contribute to this type of definition?
How could they
Anne Kennedy: I think they could be very useful....Isn't it
true that some promoters will be working by a nonreceptor
mechanism but you just need more of them. How will that
distinguish a promoting agent? We have already demonstrated
most of the other things in vitro for the classical promoters
and we could do it for other agents whose mechanism is unknown,
so I think it would be easy to study these things in vitro.
Eliezer Huberman: I agree with Anne. We do not, however,
see in vitro a threshold of promotion of cell transformation.
From my experience with tumor promoters like teleocidin and
TPA, one can establish a reasonable dose response. We also can
see reversibility of the transformed phenotype in hamster
embryo colony assay. I would say that 80 to 90% of the
phenotypic changes- are reversible.
Henry Pitot: I think the question of the presence or absence
of a threshold cannot be ascertained experimentally for
complete carcinogens or promoters. But for promoters, one
should be able to demonstrate them experimentally.
Eliezer Huberman: I agree, but where is the threshold?
Henry Pitot: That depends on the strength of the chemical.
It can vary dramatically. If you are dealing with a receptor
mechanism and you know.the equilibrium constant of the ligand,
you should be able to predict the actual concentration of the
threshold level. You can do this with dioxin. It turns out
that the threshold for promotion may be as much as an order of
magnitude higher than what the receptor data predict in
practical terms.
Anne Kennedy: How could you tell if something was directly
causing DNA structural alterations or not? For example, with
estrogen, we know that the estrogen receptor complex is in the
nucleus and doing things. When you measure aneuploidy, which
strikes me as a structural alteration, how do you know whether
that is indirect or direct for an agent?
Henry Pitot: If you don't get an alkyl group and you don't
get direct intercalation in a cell-free system, you could think
of all sorts of ideas. But the weight of evidence for the
hormones is that you are dealing with indirect effects.
D-57
-------
Bill Farland: We are trying to get too fine.
Henry Hennings: Should we include anything about
cytotoxicity on the list? in the skin at least, with promotion
you have a balance between cytotoxicity and promotional
effects. All skin promoters that I know of give you
hyperplasia in the skin that may be regenerative as a result of
cytotoxicity. I think that some promoters may act by a
selective cytotoxicity.
Henry Pitot: if we include it, we would have to have a hedge
because there is no cytotoxicity in the liver with
phenobarbital and dioxin. I don't think there is cytotoxicity
in the mammary gland with prolactin as the promoter. In the
thyroid and the kidney there is no cytotoxicity.
Henry Hennings: So we might want to say that cytotoxicity is
tissue- and compound-specific. Not all the things on the list
are necessarily criteria since not all promoters act in that
way. I think the three that Tom s>laga mentioned are the most
important. I would limit the list t*o these three and also
include no direct DNA structural alteration.
Bill Parland:
criterion.
So DNA interaction becomes.the fourth
Henry Pitot: I would still argue that the receptor is
important, certainly in the skin and the liver the major known
promoters act through a receptor mechanism.
Roy Albert: I agree for a different reason, which is that
the force of calling something a promoter based on these
criteria markedly affects the characterization of the low-level
dose-response. If you limit it to those that have demonstrable
receptor mechanisms, you have something to talk about,
otherwise you are just waving your hands. So the inclusion
puts you in a much better position to make statements about
dose-response.
Bob Langenbach: But I don't think we can assume that all
promoters act by receptor mechanisms.
Henry Pitot: One thing about receptors that is important is
that it is very clear that some of the best known promoters are
tissue-specific. I don't understand how that is possible
unless there is some other mediating mechanism; receptors seem
the best candidate that we have at the moment.
Roy Albert: I am concerned that the list does not contain
the essence of promotion which is clonal expansion of
transformed cells. We can quibble about whether benign is
benign, but I think the list should contain this property.
D-58
-------
Eliezer Huberman: I think that the critical point is
selective clonal expansion.
~~~~~~~~~~~~~~~~ • "•'
Bill Parland: The list would be titled "Criteria for
Chemicals that Can Only Promote." These elements would
constitute the weight of evidence for one, to make a finding
that a chemical is essentially only promoting.
Anne Kennedy: It seems that me that most people still think
of progression as an outgrowth of a more aggressive variant
from a malignant tumor. I like this definition better. But we
should recognize that it is not the classical definition.
Bill Farland: At a meeting this fall at NIEHS on promoters,
there was much discussion that the definition of progression
was changing to represent the types of things we have been
talking about.
Eliezer Huberman: We are back to the original definition of
progression. The previous definition dealt with conversion of
the tumor cell to its final stage of metastasis, invasiveness,
etc. In a way we didn't change much, but we are now defining
where it starts.
Bill Farland: Let's get back to the question of: What are
the types of information that allow us to identify or
characterize promotional activity? What types of experiments
would we want to do? We looked at mechanisms of promotion.
The first thing we discussed was receptors. We talked about
how we could examine receptors using mutants and competitive
studies. What are other mechanisms and data that we should
examine. What about cytoxicity?
Eliezer Huberman: I propose a working hypothesis of
carcinogenesis, especially of its promotion step. I suggest
that tumor formation may, in principle, result from continuous
expression of growth facilitating genes which, as a result of
some types of genetic changes during tumor initiation, were
placed under the control of genes that are expressed during
normal cell differentiation. Therefore, some chemicals may
promote tumor formation by inducing cell differentiation
processes in initiated cells.
Peter Magees Wouldn't DMSO be a promoter on mouse skin?
Henry Hennings: It isn't a promoter on mouse skin.
Anne Kennedy: DMSO is active at very low concentrations as
an inhibitor of transformation.
[MORE DISCUSSION OF DMSO AND ITS SUITABILITY FOR STUDY]
D-59
-------
Jim Trosko: In a recent study, Rivedal et al. (1985) used
TPA-resistant cells to investigate mechanisms and the role of
cell-cell communication in transformation. Here is a mutant
that doesn't respond to TPA, is not promoted by TPA, and for
which communication is not blocked by TPA, whereas in the
parental line TPA does promote transformation., So here is one
examplewhere a mutant for a receptor-mediated response of TPA
in in vitro promotion could be used as a model.
Bill Farland: So that would be another suggestion for the
use of mutants in the receptor category.
TAPE 23B
Jim Trosko: There are all kinds of technologies to measure
cell-cell communication and its role in growth control and
differentiation. Mutants are becoming available for gap
junctions. Antibodies exist for gap junctions. They have
cloned the gene for the gap junction. I see both genetic
molecular and cellular experiments with normal cells not only
in vitro using cell-cell communication as an endpoint but
cell-cell communication in transformation assays and in vivo
such as the kind of studies that Henry is doing in the liver.
There is a_hypothesis to be tested and technologies to measure
this endpoint of cell-cell communication. There are mutants
and antibodies available. The biochemistry to link the
receptor PKC to that endpoint is out there, we now need to
formulate good experiments to test the hypothesis.
Roy Albert: We may want to consider adding this to the
laundry list for promoter-only chemicals - namely that they
show decreased gap junction characteristics.
Henry Hennings: With regard' to the induction of
differentiation, we have shown that with normal epidermal cells
in culture, treatment with TPA induces terminal differentiation
in about half the cells. The other half appear to be
unaffected. They can then proliferate, if the initiated cells
were among that population that is not induced to terminally
differentiate, then this would give a selective mechanism for
how TPA works. We now have "initiated" cell lines which give
papillomas when put on an animal. These cell lines do not
respond to TPA by a terminal differentiation as normal cells
do. So this is a reasonably good possibility for explaining
how TPA works.
Bill Farland: This could enable us to sequester particular
cell groups that respond differently from other cells.
Dan Krewski: I would like to suggest moving the last two
lines on the list - cytotoxicity and selective clonal expansion
- to the top half of the slide since those are properties of
D-60
-------
promoters that work in the direction of cell proliferation and
effects on cell kinetics.
Bill Farland: I think the suggestion is that if one has
information on cell kinetics that allows one to start to
describe promotional activity. If one can go as far as to say
that the cell kinetics point toward a clonal expansion of
initiated cells, then you can go further and talk about
promotional activity only. As long as you include mechanistic
studies in general, but in terms of your specific definition
you focus on clonal expansion of initiated cells,' then I think
we have captured that.
Roy Albert: The advantage of this type of laundry list is
that it permits you to set up criteria for each item in terms
of the kind of evidence needed.
Bill Farland: What types of data do we want to develop with
regard to the models that we will be exploring, e.g., the
Moolgavkar model which is a leading candidate for modelling
multiple activities and responses at low doses? Roy has
indicated that some of these studies may be very expensive. If
we are going to collect those data, we need to think carefully
up front what those data are, what they will tell us and how
they will fit into the model. One suggestion I heard was that
we should start to get a handle on rates of cell division in
particular organs. Let's discuss the feasibility of this. Do
the data exist in the literature? What studies are needed?
Can we use in vitro systems to collect those data?
Jim Trosko: The Moolgavkar-Knudsen model was developed based
on patients with retinoblastoma. This human model may be an
excellent candidate for study. We are finding that tumors are
appearing in many other sites than the eyes. This is- a case of
an inherited gene mutation, where all the somatic cells in the
embryo will be initiated. Presumably promotion has occured in
the eye because of the differentiation of the eye tissue.
Chemotherapy is inducing very high cancers in other tissues in
the survivors. Closer examination of the survivors of therapy
for retinoblastoma may give risk assessors some assessment of
the progressor step. If retinoblastoma can be used as a model
of IPI, then we should be able to look at animals that are
genetically susceptible to organ site cancers and determine
whether they have inherited an I state or a P state.
Bill Farland: I heard of a fish model that produces a
pathological lesion that looks like a retinoblastoma when
exposed to chlorinated aliphatics. The questions is whether
this is a promotional event. So there are animal systems that
could be characterized as models for some specific issues.
Roy Albert: There is an old fish model that involves
crossing a molly and a swordtail that produces melanoma.
D-61
-------
Bill Parland: That is a system that could be used to
evaluate that carcinogenic response.
Peter Hagee: There is a model for kidney tumor in rats,
first described by Eker and Mossige in 1961. I know someone
who has these rats. This is a Mendelian inherited tumor.
Dan Krewski: The two most important areas that I think are
relevant to risk assessment are the question of thresholds for
promoters and the modelling of initiation/promotion phenomena
using something like the two-stage birth-death mutation model.
If we look at thresholds, then we are back into the arena of of
general toxicology where we establish a no-effect level by a
suitable uncertainty factor in order to arrive at an acceptable
level of exposure, if that is satisfactory, then maybe that is
as far as we need to go, and we should focus our efforts into
elucidating that (1) we do have initiation/promotion phenomena
and (2) the criteria for "only promotion" are satisfied. Then
the rest of the story from the regulatory point of view would
be fairly straightforward. If you wanted to go beyond that and
do modelling, the advantages I see would be that it would
perhaps allow you to validate a theory of initiation/promotion
in quantitative terms, it would be nice to see the concepts
embodied in the two-stage birth-death mutation model supported
by experimental data, it would also allow one to calculate a
measure of potency for initiation, promotion and progression
from one large unified data set. But it may not lead to any
different regulatory actions if we accept a hypothesis that
promotion is a threshold phenomenon. So we need to ask: what
are we going to gain in going beyond identifying something as a
promoter to developing a mathematical model to describe the
process?
Bill Farland: if we are going to fully characterize the
data, then we may find ourselves at a point where it doesn't
matter how we treat the data mathematically. We may get to the
same point, but we have done a service by characterizing that
information carefully.
Dan Krewski: My experience has been that the margin of
safety approach will often lead to a substantially different
result than the mathematical modelling approach to setting safe
levels of exposure.
Bill Farland: I think we still want to pursue data to model
initiation/promotion phenomena. The Moolgavkar model is the
best model we now have explain the current thinking about the
multistage model for carcinogenicity. The question is can we
get reasonable data to flesh out this model and validate it.
Dan Krewski: If we want to -start working with the model, I
think we could take the example of retinoblastoma and follow up
D-62
-------
the cases that have been treated for that lesion. That would
be a worthwhile data base against which to fit the model.
Animal species - fish or rodents - that would allow you to
induce retinoblastoma in the laboratory would be another
fruitful approach to testing the model. In general, if we
could get a good initiation/promotion system, where we knew a
lot about the mechanism of promotion, e.g., we knew that
cytoxicity was responsible for proliferation of the initiated
cells, and we could quantitate the rate of cell proliferation
using various in vitro or in vivo studies in the laboratory,
and then couple that with bioassay data to fit the whole model,
this would be very worthwhile. We need a large bioassay with
multiple doses of initiator and multiple doses of promoter.
Maybe we could even factor in the progressor. I think that
would be the ultimate data base that we could generate for
purposes of model validation.
Peter Hagee: Could we use the kidney model that I mentioned?
Dan Krewski: It doesn't have to be retinoblastoma, but I
like it because we know quite a bit about the mechanism and it
seems to fit nicely with the two-stage birth-death-mutation
model. The idea of doing it in several species is very
attractive. Other tumors that are genetically determined would
be good candidates for study.
Jim Trosko: The classic that led us into this mechanistic
era of cancer is xeroderma pigmentosum. This is a recessive
disease that predisposes the individual to the initiation
action of UV light in the skin, where most of the tumors are
formed in the skin, but not all. I think that Ken Kraemer
(1980) at NCI has been using this model to look for nonskin
tumors as evidence of the progressor. Since UV can't penetrate
internally, these tumors must be the result of exposure to
chemical initiators. This could be used to test IPI.
Bob Langenbach: Another possible system to look into is
Balmain's system where he has infected the skin of the backs of
mice with a virus containing the V or C c-Ha-jCjas. With
treatment with a promoter, he then gets papillomas, a certain
fraction of which progress to carcinomas. In this model, you
may have eliminated the need to calculate the u^ term.
Anne Kennedy: In that system, the whole virus is put in,
which brings in other functions. Many of the studies that have
concluded that the activation of c-Ha-ras is the initiating
event have been done with Richard Mulligan's shuttle vector
which brings in functions attributable to my_£. So you can't
conclude that c-Ha-ras activated by a single base mutation is
the initiating event.
D-63
-------
Dan Krewski: The question of which initiation/promotion
system is relevant for applications of the birth-death-mutation
model is quite important because ...
TAPE 24
...I understand that the skin and liver work best. For
purposes of modeling, I am not attracted to the skin because
many of the lesions that we are concerned with from the
regulatory point of view are not skin lesions. The liver would
be much more advantageous if a partial hepatectomy is not
required as part of the protocol because then the bioassay and
in vitro cell proliferation could be done under comparable
circumstances. I think that the bladder with saccharin might
be a good thing to study because we do have several good
initiators, and there is a strong promoting effect. Saccharin
has some initiating activity on its own, albeit it very weak.
Maybe an elaborate study of that compound with varying doses of
the initiator and saccharin as the promoter would provide good
information on how well the model works, it may even provide
indices of potency for saccharin as both a promoter and an
initiator.
Bill Farland: , Cohen and Ellwein at Nebraska have been doing
that work with the saccharin data base. Hopefully a critical
look at their work might lead us to make some suggestions about
what additional data might help to tie the loose ends together.
Roy Albert: Maybe statisticians should look at receptor
behavior and how one would characterize it in terms of deciding
what is a threshold level.
Dan Rrewski: Statisticians can determine a level at which
there is no statistically significant increase in the response
over background, but that is not necessarily a good estimate of
a true threshold level corresponding to no elevation in risk.
Roy Albert: That is not what I was driving at. There is a
theoretical description of receptor behavior. Then you look at
a specific receptor and from that you attempt to describe the
theoretical behavior but this has uncertainty in terms of the
experimental data, if you want to use the observed behavior to
describe the behavior of the receptor as a basis for estimating
what could be threshold model, this involves statistical
considerations.
Bill Farland: We are doing some of that right now. Steve
Bayard is working on it-in our group. They are looking at TCDD.
Bob Langenbach: Maybe the skin would be a good system to
look at because of the tremendous data base.
D-64
-------
Roy Albert: Would Peter describe the genetically based
kidney tumor system?
Peter Mageej They are adenomas. Half of the progeny get
them, half don't. The data were published by R. Eker (1961) in
Nature. The strain is delicate.
Herb Rosenkranz: There is the human model of PUV-A therapy.
Those that have received it and then get X-ray therapy rapidly
get tumors. Would that fit into the model?
Bill Parland: You must also remember that these are
individuals who have 'psoriasis which might be a means for
proliferation of cells.
END WORKSHOP.
*U.S. GOVERNMENT PRINTING OFFICE: 1988.5*8168/87015
D-65
-------
-------
-------
Jl
Si
CD
0)
w
o
o
m
o
o
10
i
00
o
t—•
co
3T O O
MI
;r O -l
2 o to
^ < o
II
o o
= s
?p
i° 3
ill
OT - —
-i (D
O (A
c S
3 <
8 I
5' n,
> m c
CO 3 3
I I'S.'
"* I
------- |