REVIEW
Mechanism-Based Cancer Prevention Approaches:
Targets, Examples, and the Use of Transgenic Mice
Stephen D. Hursting, Thomas J. Slaga, Susan M. Fischer, John DiGiovanni,
James M. Phang
Humans are exposed to a wide variety of carcinogenic insults, including endogenous and man-made chemicals, radiation, physical agents, and viruses. The ultimate goal of carcinogenesis research is to elucidate the processes involved in
the induction of human cancer so that interventions may be
developed to prevent the disease, either in the general population or in susceptible subpopulations. Progress to date in
the carcinogenesis field, particularly regarding the mechanisms of chemically induced cancer, has revealed several
points along the carcinogenesis pathway that may be amenable to mechanism-based prevention strategies. The purpose of this review is to examine the basic mechanisms and
stages of chemical carcinogenesis, with an emphasis on ways
in which preventive interventions can modify those processes. Possible ways of interfering with tumor initiation
events include the following: i) modifying carcinogen activation by inhibiting enzymes responsible for that activation or
by direct scavenging of DNA-reactive electrophiles and free
radicals; ii) enhancing carcinogen detoxification processes
by altering the activity of the detoxifying enzymes; and iii)
modulating certain DNA repair processes. Possible ways of
blocking the processes involved in the promotion and progression stages of carcinogenesis include the following: i)
scavenging of reactive oxygen species; ii) altering the expression of genes involved in cell signaling, particularly those
regulating cell proliferation, apoptosis, and differentiation;
and iii) decreasing inflammation. In addition, the utility for
mechanism-based cancer prevention research of new animal
models that are based on the overexpression or inactivation
of specific cancer-related genes is examined. [J Natl Cancer
Inst 1999;91:215–25]
The major stages of carcinogenesis were deduced over the
past 50 years, primarily from animal model studies (particularly
in the mouse skin); these stages are termed initiation, promotion,
and progression (1,2) and are shown in Fig. 1. Tumor initiation
begins when DNA in a cell or population of cells is damaged by
exposure to exogenous or endogenous carcinogens. If this damage is not repaired, it can lead to genetic mutations. The responsiveness of the mutated cells to their microenvironment can be
altered and may give them a growth advantage relative to normal
cells. In the classic two-stage carcinogenesis system in the
mouse skin, a low dose of 7,12-dimethylbenz[a]anthracene
(DMBA) causes permanent DNA damage—the initiating
event—but does not give rise to tumors over the lifespan of the
mouse unless a tumor promoter, such as 12–O-tetradecanoylphorbol-13-acetate (TPA), is repeatedly applied (3). The tumorpromotion stage is characterized by selective clonal expansion
of the initiated cells, a result of the altered expression of genes
whose products are associated with hyperproliferation, tissue
remodeling, and inflammation (2). During tumor progression,
preneoplastic cells develop into tumors through a process of
clonal expansion that is facilitated by progressive genomic instability and altered gene expression (4).
The classic view of experimental carcinogenesis, in which
tumor initiation is followed by tumor promotion and progression
in a sequential fashion, remains conceptually important to experimental carcinogenesis research. However, while the processes involved in each stage of experimental carcinogenesis
also appear to be involved in human carcinogenesis (5), the
temporal nature of initiation, promotion, and progression events
in the natural world is complex. For instance, we know from the
work of Fearon and Vogelstein (6) and of Sugimura (7) that
multiple mutational events are involved in the formation of a
tumor. Humans are generally exposed to mixtures of agents that
can simultaneously act at different stages of the carcinogenesis
process, and it has become clear that promotional events—which
frequently increase cellular proliferation or decrease apoptosis—
can influence subsequent initiation events. It is also increasingly
apparent that an individual’s genetic background can dramatically affect his or her susceptibility to a carcinogen (8–10).
Thus, rather than occurring in three discrete stages in a predictable order, human carcinogenesis is best characterized as an
accumulation of alterations in genes regulating cellular homeostasis, such as oncogenes, tumor suppressor genes, apoptosisregulating genes, and DNA repair genes (11). Most animal models used in carcinogenesis research were developed before the
identification of the major cancer-related genes, the recognition
of the importance of host susceptibility to a carcinogenic insult,
or the realization that mitogenesis and apoptosis together regulate cell number. Nonetheless, these animal models have contributed significantly to our current understanding of carcino-
Affiliations of authors: S. D. Hursting, Department of Epidemiology, The
University of Texas M. D. Anderson Cancer Center, Houston, and Department
of Carcinogenesis, The University of Texas M. D. Anderson Cancer Center,
Science Park-Research Division, Smithville; S. M. Fischer, J. DiGiovanni, Department of Carcinogenesis, The University of Texas M. D. Anderson Cancer
Center, Science Park-Research Division, Smithville; T. J. Slaga, Center for
Cancer Causation and Prevention, The AMC Cancer Research Center, Denver.
CO; J. M. Phang, Laboratory of Nutritional and Molecular Regulation, National
Cancer Institute-Frederick Cancer Research and Development Center, Frederick, MD.
Correspondence to: Stephen D. Hursting, Ph.D., M.P.H., Department of Epidemiology, Box 189, The University of Texas M. D. Anderson Cancer Center,
Houston, TX 77030 (e-mail: [email protected]).
See “Note” following “References.”
© Oxford University Press
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
REVIEW 215
Fig. 1. Multistage carcinogenesis in the mouse skin. Schematic presentation of the stages in
experimental carcinogenesis, using the mouse skin as an example. The initiation stage is characterized by the generation of a cell that is genetically altered through covalent binding of a carcinogen to DNA and subsequent fixation of a mutation. The epidermis retains a normal appearance at
this stage. The promotion stage involves the expansion of initiated stem cells through epigenetic
mechanisms, accompanied by alterations in the expression of critical genes that regulate proliferation, inflammation, differentiation, and other processes. Also pictured is the development of clonal
outgrowths that result in benign papillomas. The progression stage is characterized first by dysplasia and ultimately by invasion and metastasis, due to additional genetic alterations (such as loss
of tumor suppressor function) and progressive genomic instability.
genesis and the ways to interfere with that
process. Given the multistage nature of carcinogenesis and our advancing knowledge of the critical mechanisms involved at each stage, strategies
for the prevention of cancer should use mechanism-based approaches to block the carcinogenesis process at all stages along the pathway.
It is beyond the scope of this paper to comprehensively review the field of carcinogenesis.
The carcinogenesis literature is extensive and
many reviews (1–5,7,10,11) have critically evaluated much of it. Rather, the objective of this review is to use our current knowledge about carcinogenesis to suggest opportunities for
preventive interventions. In the following sections, each of the potential targets for the nutritional modulation or chemoprevention of carcinogenesis shown in Figs. 1 and 2 will be
examined in greater detail, and examples of how
they can be modulated will be presented. The
information sources for this review include the
MEDLINE (from January 1, 1993, through January 1, 1998) and CANCERLIT (from January 1,
1983, through January 1, 1998) databases, which
were searched with the key words “carcinogens,
environmental”; “nutrition and neoplasms”; and
“chemoprevention and neoplasms.” Reviews,
editorials, and primary journal articles identified
by this search, along with chapters from textbooks on cancer etiology and prevention available at The University of Texas M. D. Anderson
Cancer Center were used to summarize our current knowledge of carcinogenesis and the effects
of nutrition or chemopreventive agents on that
process. Table 1 presents a summary of anticarcinogenic dietary factors and chemopreventive
agents that we found in this search. We also present a comprehensive review of the literature on
the nutritional modulation and chemoprevention
of cancer in transgenic and knockout mice, based
on searches of the MEDLINE and CANCERLIT
databases (for the same periods as above), using
keywords “nutrition and transgenic mice,” “chemoprevention and transgenic mice,” “nutrition
and knockout mice,” and “chemoprevention and
knockout mice.”
TARGETS FOR ANTI-INITIATION
STRATEGIES
Carcinogen Activation
Fig. 2. Multistage carcinogenesis: processes and prevention strategies. A schematic presentation of
stage-specific prevention strategies. The initiation stage is characterized by the conversion of a
normal cell to an initiated cell in response to DNA-damaging agents, with the genetic damage
indicated by an X. The promotion stage is characterized by the transformation of an initiated cell
into a population of preneoplastic cells. due to alterations in gene expression and cell proliferation.
The progression stage is characterized by the transformation of the preneoplastic cells to a neoplastic cell population as a result of additional genetic alterations (indicated by additional Xs).
Intervention points for strategies to prevent these processes are indicated by brick walls along the
pathway. ROS ⳱ reactive oxygen species.
216 REVIEW
Most chemicals are not carcinogenic, but a
wide variety of chemicals and chemical classes
can cause cancer in animals and humans (1).
Most chemical carcinogens are genotoxic, causing DNA damage by reacting with DNA bases.
The carcinogens form covalent adducts with
DNA in the nucleus and mitochondria. Endogenous carcinogens—which are often reactive
oxygen species (ROS) generated as part of normal oxidative metabolism or as a result of the
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
Table 1. Examples of dietary factors or chemopreventive agents that target
specific stages of carcinogenesis
Prevention strategy
Tumor initiation
Inhibit carcinogen activation
Scavenge electrophiles
Enhance carcinogen
detoxification
Enhance DNA repair
Tumor promotion/progression
Scavenge reactive oxygen
species
Alter gene expression
Decrease inflammation
Suppress proliferation
Induce differentiation
Encourage apoptosis
Examples
Epigallocatechin gallate (EGCG),
selenium, phenyl-isothiocyanate
(PEITC), indole-3-carbinol, coumarins,
ellagic acid, resveratrol, genistein
Ellagic acid, EGCG
Oltipraz, diallyl sulfide, PEITC, EGCG,
N-acetylcysteine, resveratrol
Calorie restriction, EGCG, selenium
Antioxidants (carotenoids, ␣-tocopherol,
ascorbic acid, EGCG), selenium,
calorie restriction
Retinoids (all-trans retinoic acid,
fenretinide), calorie restriction,
monoterpenes (i.e., d-limonene),
dehydroepiandrosterone (DHEA),
fluasterone, genistein
Nonsteroidal anti-inflammatory drugs,
calorie restriction, DHEA, fluasterone,
antihistamines
Calorie restriction,
difluoromethylornithine, selenium,
tamoxifen, DHEA, fluasterone,
genistein, retinoids
Retinoids, calcium, sodium butyrate
DHEA, fluasterone, fenretinide, sodium
butyrate
metabolism of xenobiotic compounds, as well as by ultraviolet
radiation and gamma radiation—can also cause extensive DNA
damage. For instance, proto-oncogenes and tumor suppressor
genes are normal cellular genes that can be mutated to cause
uncontrolled cell growth or other characteristics that increase the
probability of neoplastic transformation (11–13).
Metabolic activation of procarcinogens (i.e., carcinogens requiring enzymatic conversion to DNA-reactive intermediates)
is generally catalyzed by cytochrome P450 enzymes
through oxidation. More than 100 distinct mammalian P450 enzymes have been identified (14). In addition, there are other
enzyme systems involved in carcinogen activation such as peroxidases (including the cyclooxygenases, which will be discussed in more detail below) and certain transferases such as
N-acetyltransferase and sulfotransferase (15,16). Each of these
enzymes provides a potential target for modulating carcinogen
activation.
One common feature of the metabolic activation of all procarcinogens is that their ultimate DNA-reactive carcinogenic
species are electrophilic (17). In addition, many direct-acting
carcinogens damage DNA through electrophilic intermediates
(18). Thus, the electrophilicity of the ultimate carcinogenic species serves as a shared intervention target for most chemical
carcinogens. The electrophilic metabolites may themselves be
ROS and interact as such with DNA (19). Oxygen-free radicals
may also be involved in a step required for activation of a
procarcinogen, and thus the reactions involved in metabolic activation of carcinogens may release ROS that can in turn attack
DNA (19). Thus, directly scavenging DNA-reactive intermediates with antioxidants or other agents that can scavenge electophiles constitutes a plausible strategy for modulating this early
stage of carcinogenesis.
Carcinogen Detoxification
In addition to the carcinogen-activating enzymes, a series of
enzymes (the so-called phase II enzymes) detoxify activated
carcinogens, thus preventing their binding to DNA. The induction of the glutathione S-transferases (GSTs) is an important
response for the detoxification of xenobiotics (20). This class of
enzymes couples a number of diverse substrates to glutathione to
excrete them from the body. GSTs are segregated into three
classes based on their sequence homology and specificity for
substrates (21). Other detoxification enzymes include uridine
diphosphate-glucuronosyl transferase, quinone reductase, and
the epoxide hydrolases (22,23). The efficiency with which these
and other enzymes detoxify carcinogens is a critical factor in
determining the carcinogenicity of a particular xenobiotic.
DNA Repair Processes
The generation of DNA-reactive intermediates by most
chemical carcinogens leads to the production of DNA adducts or
other types of damage. As reviewed by Mitchell et al. (24),
normal mammalian cells can efficiently remove DNA damage
induced by carcinogens. Cells use different strategies to repair
DNA damage, depending on the structure of the damage and its
location in the genome. For example, small lesions (such as
alkylated DNA bases) can be repaired by a mechanism termed
base excision repair (25). This process involves removal of the
damaged base followed by a “small cut-and-patch” repair involving removal of a few nucleotides. When methylation occurs
at either the O6 or O4 positions of guanine or thymine, the
modified bases can be repaired by the direct transfer of the
methyl group to a methyl transferase (26). Bulky carcinogeninduced DNA adducts and ultraviolet light photodimers can be
repaired through a “large cut-and-patch” mechanism involving a
region of approximately 27–29 nucleotides that includes the
damaged bases; this is termed nucleotide excision repair (27).
The integrity of the genetic information is threatened not only by
various environmental exposures but also by errors produced
during normal DNA replication, for example, non-Watson–
Crick base-pairing and slippage during DNA replication. To
correct the errors resulting from such mis-replication, cells have
also developed a mismatch repair mechanism (28).
NUTRITIONAL MODULATION AND CHEMOPREVENTION
OF TUMOR INITIATION PROCESSES: EXAMPLES
Inhibiting Carcinogen Activation
Fruits, vegetables, herbs, and other foodstuffs (as well as
inedible plants) contain numerous chemical constituents known
to affect the metabolic activation of chemical carcinogens. Examples of food sources containing agents that decrease carcinogen activation are the cruciferous vegetables, such as cauliflower, broccoli, and cabbage. The crucifers are sources of
isothiocyanates, which are known to interfere with the metabolism of nitrosamines. Studies by Chung et al. (29), Hecht (30),
Stoner and Mukhtar (31), and Morse et al. (32,33) have shown
conclusively that the metabolism and carcinogenicity of the tobacco carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone are decreased by the administration of phenylisothiocyanate. Extensive structure-activity studies by these
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
REVIEW 217
investigators have also shown that the carbon chain length of the
isothiocyanate moiety is associated with tumor prevention activity, with a longer chain isothiocyanate apparently more suitable for insertion into the cell due to increased lipophilicity (33).
Also, diallyl sulfide, a common volatile in garlic, has been
shown to be a potent inhibitor of cytochrome P450 2E1 (34,35).
This cytochrome P450 metabolizes ethanol, acetone, several
known chemical carcinogens (including several nitrosamines
that target the nasal tissues, oral cavity, liver, and esophagus),
and dimethylhydrazine and its metabolites [which induce colon
tumors in rodents (36)].
Members of several other classes of plant compounds have
demonstrated anti-initiation activity, including the flavonoids,
isoflavonoids, and coumarins. For example, earlier studies had
suggested that coumarins, which are widely distributed in nature
and are found in all parts of plants (37), could modulate drugmetabolizing enzymes and cytochrome P450s (38). Cai et al.
(39) showed that several naturally occurring coumarins can
block skin tumor initiation by polycyclic aromatic hydrocarbons,
such as benzo[a]pyrene and DMBA, by inhibiting cytochrome
P450s involved in the metabolic activation of these carcinogens
(39).
As reviewed by MacLeod and Slaga (17), several compounds
have been proposed as scavengers for the ultimate electrophilic
metabolites of carcinogens, such as benzo[a]pyrene diol epoxide
(BPDE), a metabolite of benzo[a]pyrene. Earlier studies (40)
showed that several sulfhydryl compounds, including cysteine
and 2-mercaptoethanol, were effective as nucleophilic traps for
BPDE. In addition, riboflavin has been shown to promote the
detoxification of BPDE by enhancing its hydrolysis (41). Also,
a group of plant phenols, notably ellagic acid, has been identified that react facilely with BPDE and thereby block the mutagenicity of BPDE in vitro (42). Ellagic acid has been shown to
be an anticarcinogen in vivo by demonstration of a protective
activity against topically applied BPDE in the mouse skin model
(43). In addition, a major polyphenolic antioxidant found in
green tea, epigallocatechin-3-gallate (EGCG), has been shown
to have strong anticarcinogenic effects in several models, including the mouse skin, lung, forestomach, esophagus, duodenum, pancreas, liver, breast, and colon models (44). EGCG reportedly can trap the activated metabolites of several
procarcinogens (31).
Enhancing Carcinogen Detoxification
As mentioned above, the process by which potentially dangerous xenobiotics are conjugated to an endogenous cellular
nucleophile and then excreted from the body is also an important
target for the prevention of tumor initiation. Detoxification of
chemical carcinogens by enzymes such as GSTs and uridine
diphosphate-glucuronyl transferase is enhanced by several constituents of garlic and onions, cruciferous vegetables, and certain
spices. For example, in animals, so-called phase II detoxification
enzymes are induced by oral exposure to diallyl sulfide and
s-allyl-cysteine, which are found in garlic; both compounds enhance GST levels in liver and colon (45). As previously discussed, these organosulfur compounds are also known to inhibit
the activity of several cytochrome P450s (34). Thus, the tumorinhibiting effects of diallyl sulfide and related compounds may
be due to the dual effect of decreased carcinogen activation and
enhanced carcinogen detoxification (46). The isothiocyanates
also have a dual role: they suppress carcinogen activation as well
218 REVIEW
as enhance carcinogen detoxification by increasing GST activity
(30). The antischistosomal drug oltipraz (47) and the antioxidative agent N-acetylcysteine (29) are also highly effective inducers of GSTs and are potent inhibitors of induced colon, lung, and
bladder carcinogenesis. The phytoalexin resveratrol, found in
grapes and other food products and an inhibitor of the two-stage
mouse skin carcinogenesis model, also has been shown to induce
phase II enzymes such as quinone reductase (48).
Enhancing DNA Repair
Although the gene products and general mechanisms of DNA
repair in prokaryotes are fairly well characterized, mammalian
repair systems have only recently been elucidated, and little is
known about the influence of dietary factors on these processes.
Calorie restriction (49,50), epigallocatechin galate (51), and selenium (52) enhance unscheduled DNA synthesis and other
measures of repair capacity. Calorie restriction has also been
shown to increase apoptotic cell death in heavily damaged cells,
thereby accelerating the elimination of cells with irreparable
DNA damage (53).
TARGETS FOR ANTIPROMOTION
STRATEGIES
AND
ANTIPROGRESSION
Epigenetic Changes in Cell Signaling
The tumor promotion phase of multistage carcinogenesis involves the clonal expansion of initiated cells. Tumor-promoting
agents are not mutagenic as are carcinogens, but rather alter the
expression of genes whose products are associated with hyperproliferation, apoptosis, tissue remodeling, and inflammation. At
some point, the developing tumor constitutively expresses these
genes and thus becomes tumor-promoter independent. The identification of the mechanisms by which tumor promoters alter
gene expression has been a major goal during the past decade,
particularly because determining the critical events will reveal
targets for the development of new prevention strategies. It has
also become clear in the past few years that apoptosis and mitogenesis are equally important in the homeostasis of cell number, and that the growth advantage manifested by initiated cells
during promotion is usually the net effect of increased proliferation and decreased apoptosis. Thus, in addition to cell proliferation, apoptosis has emerged as a critical target for prevention (54).
As described earlier (Fig. 1), the mouse skin model of multistage carcinogenesis is an excellent system for studying molecular alterations associated with the various stages of tumor
development and so will be the primary focus of the following
discussion on mechanisms involved in tumor promotion. Among
the most potent mouse skin tumor promoters are the phorbol
esters; TPA has been the prototype for many years (2). However,
a wide variety of compounds are tumor promoters and bring
about both biologic and molecular changes similar to those elicited by TPA (55). Changes in gene expression as a consequence
of external tumor promoter stimuli usually activate (but sometimes inactivate) specific signal transduction pathways. The nature of the initial interaction of tumor promoters with the cell
depends on the type of promoter used. For example, TPA interacts with specific receptors that are isoforms of protein kinase C
(PKC) (56). The identification of PKC as the major target for
phorbol esters and other promoters, such as mezerein, indole
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
alkaloids, and polyacetates, suggests that activation of PKC is a
critical event in carcinogenesis. By activating PKC, phorbol esters and related tumor promoters appear to bypass the normal
cellular mechanisms for regulating cell proliferation. Several
oncogenes (particularly ras), hormones, growth factors, and cytokines are also known to activate this signaling pathway. Other
promoters such as okadaic acid are potent inhibitors of phosphatases and increase the level of phosphorylated proteins,
which often have an activating effect equivalent to the activation
of kinases (57). However, regardless of the disparity in the tumor
promoters’ initial signaling events, the key biologic and molecular changes they elicit, such as increased DNA synthesis, induction of ornithine decarboxylase, induction of growth factors and
cytokines, and increased production of eicosanoids, are all similar. This overall alteration in signal transduction and gene expression contributes significantly to the selection and growth of
the initiated cell population.
As discussed by Fischer and DiGiovanni (57), not all tumor
promoters work through receptor-mediated mechanisms.
Organic peroxides, such as benzoyl peroxide and hydroperoxides, are examples. Unlike the phorbol esters, the peroxides
require metabolic activation. The molecular targets of benzoyl
peroxide have not been elucidated, although it has been shown
to produce macromolecular damage, particularly covalent adducts with proteins (19). Altering signaling by nonreceptormediated mechanisms would require considerably higher doses
of promoter than required by receptor-dependent modulators,
e.g., a 20-mg dose of benzoyl peroxide is needed for tumor
promotion, whereas only microgram amounts of TPA are required (57).
Despite its importance, activation of PKC alone does not
appear to be sufficient for mediating phorbol ester-induced hyperproliferation in vivo, in part because a major consequence of
PKC activation in keratinocytes is the induction of terminal
differentiation. The regulation of keratinocyte proliferation and
differentiation is a complex process and probably involves interaction of different cell types in the epidermis and dermis, as
well as multiple signaling pathways within the keratinocytes
themselves. Several receptor tyrosine kinases and their ligands
appear to be linked to keratinocyte proliferation. The four main
receptors are the epidermal growth factor receptor (EGFR), insulin-like growth factor-1 receptor, basic fibroblast growth factor receptor, and hepatocyte growth factor receptor (57). All four
of these receptors are expressed on the surfaces of keratinocytes;
however, with the exception of transforming growth factor-␣
(TGF-␣), which is the ligand for EGFR, the ligands for the
receptors are produced by dermal fibroblasts or inflammatory
cells and act in a paracrine manner.
Several lines of evidence suggest that tumor promoters generally increase the expression of a number of growth factors and
cytokines. TPA induces TGF-␣, TGF-␤, tumor necrosis factor␣, granulocyte-macrophage stimulating factor, and interleukins
1 and 6 (58). The profile of growth factor induction is different
for promoters with different initial mechanisms of action, although most seem to induce TGF-␣ messenger RNA (mRNA)
expression. TPA treatment also increases expression of EGFR,
possibly as a consequence of activating c-Ha-ras. Alternatively,
high TGF-␣ levels may lead to autoinduction of EGFR (57).
Regardless of how it occurs, elevated levels of EGFR and its
principal ligand, TGF-␣, are strongly associated with the development of neoplasias.
Inflammation
In addition to inducing changes in gene expression by activating specific signaling pathways, tumor promoters can elicit
the production of proinflammatory cytokines (such as tumor
necrosis factor and several interleukins) and nonprotein factors
(such as nitric oxide) involved in inflammation and carcinogenesis (57). Of critical importance to the promotion process is the
release of arachidonic acid and its metabolism to eicosanoids
(59). Eicosanoids, which include the prostaglandins and hydroperoxy forms of arachidonic acid, are involved in such processes as inflammation, the immune response, tissue repair, and
cell proliferation.
Prostaglandin synthesis is regulated by COX gene expression. Two separate gene products, COX-1 and -2, have similar
cyclooxygenase and peroxidase activities, although they are differentially regulated (59,60). While a variety of factors, including serum, growth factors, and phorbol esters, can increase the
mRNA levels of both cyclooxygenases, the COX-2 gene generally responds in a much more dramatic fashion and thus has been
referred to as a phorbol ester-inducible immediate early gene
product (61).
In the mouse two-stage skin carcinogenesis model, tumor
promotion is a distinct, rate-limiting step that determines the
formation of premalignant tumors. As discussed above, the role
of tumor promoters in human cancer is more complex because
human exposure tends to involve sporadic low doses of complex
mixtures of carcinogens, co-carcinogens, and tumor-promoting
agents. Nonetheless, studies of rodent tumor models of liver,
bladder, colon, and breast cancers—and analyses of human tumor formation—suggest that processes analogous to tumor promotion by TPA on the mouse skin are a common feature of
carcinogenesis (1). Thus, epigenetic changes in cell signaling,
such as altered growth factor production and receptor expression, and elevated synthesis of inflammatory and mitogenic factors, such as cytokines and eicosanoids, are key targets for inhibiting tumor promotion.
Tumor Progression
As noted earlier, tumor progression involves the accumulation of additional genetic alterations in an initiated cell clone and
generally gives a growth advantage to the progressing clone. In
progression, a focal lesion consisting of a population of initiated
and promoted cells ultimately becomes an invasive malignant
tumor. One frequently observed genetic alteration that appears to
contribute to malignant progression is mutation in the p53 [also
known as TP53] tumor suppressor gene (62). The p53 gene
product is a transcription factor that regulates the expression of
a number of DNA-damage and cell cycle-regulatory genes and
genes regulating apoptosis. By enhancing transcription of these
critical genes, p53 regulates the cellular response to DNA damage (63). p53 also plays a role in maintaining genomic stability
(64). Genomic instability, a hallmark of spontaneous malignant
progression, is characterized by sequential chromosomal aberrations, such as duplications, deletions, and loss of heterozygosity, which lead to rapid accumulation of unfavorable genetic
alterations and eventually to malignant cell growth. Cell numbers, normally maintained by a balance of genes regulating cell
proliferation and apoptosis, are altered. DNA hypomethylation,
frequently observed in malignant tumors, may also contribute to
malignant progression (65). Thus, p53, other cell cycle and ap-
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
REVIEW 219
optosis regulators, and other genes regulating genomic instability and DNA methylation are critical targets for cancer prevention at later stages of carcinogenesis.
NUTRITIONAL MODULATION OF TUMOR PROMOTION
AND PROGRESSION: EXAMPLES
Scavenging Reactive Oxygen Species
ROS play an important role in a variety of normal processes
within the body, including the immune response against pathogens, intracellular signaling, and vascular permeability. However, the accumulation of ROS as byproducts of normal energy
metabolism and in response to inflammatory conditions or ROSgenerating environmental exposures (i.e., to particulates in tobacco smoke) has been associated with the pathogenesis of cancer in rodents and humans (19,66). Experimental studies (19,66)
have shown that ROS can act as both initiators and promoters of
tumors by damaging critical cellular macromolecules such as
DNA, proteins, and lipids, and by acting as cell-signaling molecules (as nitric oxide does). Antioxidants, including ascorbic
acid, ␣-tocopherol, selenium, and several polyphenolic compounds found in green tea, spices, fruits, and vegetables have
been shown to effectively inhibit TPA promotion in the mouse
skin (19). Calorie restriction, which is one of the best documented and most effective experimental manipulations for decreasing rodent tumor development (67)—including TPAinduced skin carcinogenesis (68), may exert its antitumor effects
largely by decreasing ROS production and enhancing antioxidant defenses. Calorie restriction decreases the rate of accumulation of oxidized DNA and protein that accompanies aging in
rodents (69). In addition, a number of intracellular antioxidant
defense systems, including superoxide dismutase, catalase, and
glutathione peroxidase, are reportedly enhanced by calorie restriction (70). Thus, evidence is mounting that calorie restriction
may decrease oxidative stress by decreasing oxidant production
and enhancing antioxidant capacity, although the exact mechanisms involved have yet to be fully established.
Altering the Expression of Genes Regulating Cell
Proliferation, Apoptosis, and Differentiation
Many studies using the two-stage carcinogenesis model in
mouse skin have identified dietary components that act through
diverse mechanisms to alter tumor promotion and progression. A
number of retinoids, particularly all-trans retinoic acid, are specific inhibitors of TPA-induced tumor promotion in mouse skin
(71,72). Although the retinoids’ mechanism of action is not fully
understood, data indicate that they affect epithelial differentiation and also reduce elevated polyamine levels by inhibiting the
induction of epidermal ornithine decarboxylase (73). Several
reports (73,74) indicate that polyamines are involved in regulating cellular differentiation and growth. Retinoids bring about
many of their effects by interacting with nuclear receptors.
These nuclear receptors are trans-activating factors that can
regulate the expression of specific genes involved in differentiation, proliferation, and apoptosis (75,76). The synthetic retinoid fenretinide, which has shown promising chemopreventive
activity against several cancers, appears to exert its antitumor
effects primarily by inducing apoptosis in damaged cells (77).
As mentioned earlier, it has become clear in the past few years
that apoptosis and mitogenesis are equally important in main220 REVIEW
taining cell number homeostasis; thus, in addition to cell proliferation, apoptosis has emerged as a critical target for prevention.
Perhaps the clearest example of dietary modulation of skin
carcinogenesis through alteration of the PKC pathway comes
from the laboratory of Birt et al. (68) that showed that calorie
restriction, which inhibits skin tumor promotion by TPA, inhibits PKC activity, and decreases the concentrations of different
PKC isoforms (particularly PKC␣ and PKC␨). Birt et al. (78)
have also shown that feeding diets high in corn oil increases
PKC activity in epidermal cells, apparently by influencing intracellular lipid metabolism rather than altering the distribution
of PKC isoforms.
Decreasing Inflammation
A number of prostaglandin synthesis inhibitors are effective
in counteracting skin tumor promotion and carcinogenesis.
Compounds such as anti-inflammatory steroids (i.e., glucocorticoids) are potent inhibitors of mouse skin tumor promotion by
phorbol esters (58). These compounds are effective phospholipase A2 inhibitors, which may explain their ability to decrease
the amount of arachidonic acid available for metabolism to important proinflammatory endproducts. Inhibitors of the cyclooxygenase pathway, such as indomethacin and flurbiprofen, have
been best studied as colon cancer chemopreventives (79) and
also inhibit skin tumor promotion in most mouse strains (80).
The cyclooxygenase pathway is a major prevention target, primarily because these enzymes (particularly COX-2) play a role
in inflammation as well as in apoptosis and cellular adhesion in
some cells (81). In addition, several safe, effective, and inexpensive agents that can perturb the pathway, such as the nonsteroidal anti-inflammatory drugs (NSAIDs), are readily available.
FUTURE PROSPECTS: TRANSGENIC ANIMAL MODELS
FOR NUTRITION AND CANCER CHEMOPREVENTION
STUDIES
Future progress in nutrition and cancer prevention research
may be facilitated by the use of animals with specific genetic
susceptibilities for tumor development. As mentioned previously, most chemically induced tumor models used in carcinogenesis research involve high doses of a single genotoxic carcinogen that can induce severe genetic damage, often randomly.
The main goal for many of these models, which were generally
developed before the identification of cancer-related genes, was
to create animals that quickly developed neoplasias to provide
investigators with sufficient material in a timely fashion for
studies of tumor formation. Although some of the molecular
alterations in the commonly used models have been identified,
the types of alterations caused by high-dose chemical exposure
do not generally reflect the gene–environment interactions underlying carcinogenesis in humans. Furthermore, it can be difficult to interpret the activity of preventive compounds in these
models, since the effects of an agent on the metabolic activation
or detoxification of a high dose of a single chemical may not be
mechanistically relevant to typical chronic, low-level human exposures to mixtures of exogenous or endogenous carcinogens.
The recent development of mouse strains with overexpressed
or inactivated carcinogenesis-related genes provides investigators with new models for studying carcinogenesis and for testing
strategies to offset specific genetic susceptibilities to cancer.
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
Thus far, prevention studies in genetically altered mice have
predominantly focused on alterations in two specific genes implicated in human cancer: the adenomatous polyposis coli (Apc)
gene and the p53 tumor suppressor gene (82).
Min and Apc Knockout Mice
As reviewed by Dove et al. (83), mutation of the Apc gene
may be a common denominator in all human colon cancer, including polypoidal and nonpolypoidal familial colon cancers, as
well as sporadic colon cancers. Thus, an animal model with an
alteration in this Apc gene should be very useful for testing
prevention strategies. The multiple intestinal neoplasia (Min)
mouse, which carries a fully penetrant dominant mutation converting codon 850 of the murine Apc gene from a leucine to a
stop codon, was first described in 1990 (84). The Min mutation
was discovered by phenotypic screening after random germline
mutagenesis with ethylnitrosourea, a chemical known to induce
point mutations. Min/Min homozygous mice are not viable, but
heterozygotes live and develop scores of large adenomas
throughout the small intestine (but only rarely in the colon) by
1–3 months of age (84). Interestingly, the phenotypic expression
of Apc mutations in the Min mouse is somewhat different from
that in humans with familial adenomatous polyposis, whose adenomas occur exclusively in the colon and duodenum (83). Despite this potential limitation, the Min mouse model should facilitate studies of carcinogenesis and anticarcinogenesis related
to Apc in intestinal neoplasia. Three recent papers (85–87) report that NSAIDs reduce spontaneous adenoma formation in
Min mice, further enhancing the interest in NSAIDs as colon
cancer chemopreventives (79).
Several Apc knockout mice strains have also been developed
(88,89). Oshima et al. (88) used gene targeting techniques to
introduce a truncation mutation in codon 716 of the mouse Apc
gene. Like homozygous Min mice, the homozygous Apc⌬716
knockout mice are not viable, but the heterozygotes live and
develop numerous intestinal polyps at an early age. Oshima et al.
(90) reported that Apc⌬716 knockout mice lacking COX-2 (created by crossing Apc⌬716 knockout mice with COX-2 knockout
mice) had dramatically fewer and smaller intestinal polyps due
to inactivation of the Apc gene. Furthermore, treating the
Apc⌬716 mice with the NSAIDs sulindac or MF Tricyclic (a
specific COX-2 inhibitor) reduced the polyp number relative to
the untreated controls. Furthermore, MF Tricyclic reduced the
polyp number more significantly than sulindac (which inhibits
both COX-1 and COX-2). This study using the Apc⌬716 mice
provides the strongest evidence to date that COX-2 plays a key
role in Apc-related tumorigenesis. These findings, along with
data indicating that daily aspirin use reduces colon cancer risk in
humans (79) suggest that COX-2 inhibitors have great promise
as colon cancer chemopreventives.
p53 Knockout Mice
Mutation of the p53 tumor suppressor gene is the most frequently observed genetic lesion in human cancer; more than
50% of all human tumors examined to date have identifiable p53
gene point mutations or deletions (91). Donehower et al. (92)
first reported in 1992 that homozygous p53 knockout (p53−/−)
mice were viable, but highly susceptible to spontaneous tumorigenesis (particularly lymphomagenesis) at an early age. p53−/−
mice have been useful tools for studying the role of p53 in
carcinogenesis. For example, in response to the two-stage skin
carcinogenesis protocol, p53−/− mice have the same frequency
of benign papillomas as do wild-type (p53+/+) mice, but the
papillomas progressed to malignant carcinomas much faster
(93). Furthermore, the carcinomas in the p53−/− mice were histopathologically more malignant, further confirming the importance of p53 loss in the acceleration of tumor progression. These
mice are also an attractive and relevant tumorigenesis model for
studying cancer prevention strategies, given the frequency of
p53 mutations in human tumors and the rapidity with which
spontaneous tumors develop.
We have evaluated (94–97) the ability of several dietary and
chemopreventive interventions to offset the increased susceptibility of p53−/− mice to spontaneous tumorigenesis, including
calorie restriction. Calorie restriction reduces cell proliferation
in most tissues studied (67,70). This decreased rate of DNA
replication in normal cells in response to calorie restriction may
make those cells less susceptible to mutations induced by carcinogens and may suppress the progression of preneoplastic
cells. In p53−/− mice, a 40% restriction of carbohydrate calories
results in a 75% delay in the development of spontaneous tumors
(which are mostly lymphomas in these mice) and a slowing of
the traverse of lymphocytes through cell cycle (94). The time to
tumor onset in these mice is p53 dependent: most p53−/− mice
develop and die from spontaneous tumors by approximately 6
months of age compared with nearly 2 years for p53+/+ mice.
However, the highly statistically significant tumor-delaying effect of calorie restriction, relative to ad libitum feeding, is similar in both p53−/− and wild-type mice, indicating that the
mechanisms underlying calorie restriction may be independent
of p53 (95).
In addition, the chemopreventive steroid dehydroepiandrosterone (DHEA; 0.3% in the diet) significantly delays spontaneous tumorigenesis (particularly lymphomas) in p53−/− mice
(96). The DHEA analogue 16-␣-fluoro-5-androsten-17-one (fluasterone; 0.15% in the diet) also suppresses spontaneous lymphoma development and lengthens survival in p53−/− mice (97).
These findings clearly demonstrate that the increased susceptibility to cancer as a result of a genetic lesion, such as loss of p53
tumor suppressor function, can be offset at least in part by preventive approaches.
p53−/− mice have also been useful for elucidating how calorie restriction and administration of chemopreventive steroids
inhibit tumorigenesis. For example, the antitumor effect of
DHEA in p53−/− mice is independent of its effects on food
intake or on nucleotide pool levels (97), as was previously suggested in studies using other animal models (98). We showed
(99) that both calorie restriction and DHEA decreased the rate of
thymocyte proliferation, whereas DHEA, but not calorie restriction, increased the rate of apoptosis. The apoptosis-inducing
effects of the chemopreventive steroids appear to be mediated by
decreased Bcl-2 gene expression. In addition, we showed (100)
that calorie restriction, DHEA, and fluasterone each reduce nitric oxide levels and decrease nitric oxide synthetase expression.
Heterozygous p53 knockout mice (p53+/−), having only one
p53 allele inactivated, may be analogous to humans susceptible
to heritable forms of cancer due to decreased p53 gene dosage,
such as individuals with Li–Fraumeni syndrome (101). While
p53+/− mice have a low rate of spontaneous tumorigenesis up to
12 months of age (102), they are more susceptible to chemically
induced tumor development than are wild-type mice. Experimentally induced cancers—including p-cresidine-induced blad-
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
REVIEW 221
der tumors (101), dimethylnitrosamine-induced liver tumors
(102), and radiation-induced lymphomas and sarcomas (103)—
all appear significantly earlier in p53+/− mice than in similarly
treated p53+/+ mice. As mentioned previously, malignant progression of DMBA-initiated, TPA-promoted skin papillomas occurs much faster in p53+/− mice than in p53+/+ mice (93).
These findings suggest that p53+/− mice are more sensitive to
mutagenic carcinogens than are p53+/+ mice and appear to be
susceptible to at least some low-dose, chronic carcinogen regimens that closely mimic human exposures. Thus, these mice
have tremendous potential as models for studying gene–
environment interactions relevant to human cancer prevention.
For example, we recently found (104) that p53+/− mice develop
approximately twice as many azoxymethane-induced aberrant
crypt foci and three times as many colon tumors as p53+/+ mice.
Furthermore, the NSAID sulindac offsets the increased susceptibility of these mice to azoxymethane-induced aberrant crypt
foci and colon tumors.
Table 3. Examples of knockout mouse strains with DNA repair
genes inactivated
Gene
Spontaneous tumors/other
phenotypes
Cellular role
XRCC1
Strand break repair
PARP
MSH1
Strand break repair
Mismatch repair
MSH2
XPA
Mismatch repair
Nucleotide excision repair
XPC
Nucleotide excision repair
ERCC1
Nucleotide excision repair
MGMT
O-6-methylguanine-DNA
methyltransferase
−/− embryonic lethal, +/− runted,
repair deficient
Skin tumors
No phenotype, cells show
microsatellite instability
Lymphomas
Skin cancer susceptibility
(ultraviolet radiation and
chemicals)
Skin cancer susceptibility
(ultraviolet radiation)
−/− embryonic lethal, +/− runted,
repair deficient but no increased
sensitivity to chemicals
Susceptible to
nitrosomethylurea-induced tumors
Other Potential Models for Chemoprevention Studies
Numerous mouse models with inactivated or overexpressed
cancer-related genes other than Apc or p53 have been developed
in recent years (105–108). Examples of mice with inactivated
oncogenes or tumor suppressor genes are provided in Table 2,
inactivated DNA repair genes in Table 3, and inactivated inflammation-related genes in Table 4. In addition, as reviewed
(105), several mouse strains involving targeted disruption of
various genes, such as the peroxisome proliferator activator receptor (which is thought to mediate the biologic effects of peroxisome proliferators, such as hyperlipidemic drugs, plasticisers,
herbicides, and pesticides), the aryl hydrocarbon receptor (which
binds dioxin), and various P450 genes involved in carcinogen
activation have also been developed. Furthermore, a large and
rapidly increasing number of transgenic mouse strains have been
developed that overexpress many of the cancer-related genes,
often in a tissue-specific fashion depending on the promoter. The
number of transgenic strains is too great to list here in a meaningful way; however, there are several reviews in the literature
(105–107), as well as online databases of transgenic and knockout mice, such as the Induced Mutation Registry database mainTable 2. Examples of knockout mouse strains with tumor suppressor genes
or oncogenes inactivated
Gene
APC⌬716
p53*
Cellular role
p21
Cytoskeletal regulation
DNA damage response, cell
cycle control
Cell cycle control
NF1
Ras regulatory protein
BRCA1
Unknown–possibly DNA
repair, proliferation
Unknown–possibly DNA
repair
Cell cycle control, DNA
damage checkpoint
Cell cycle control
BRCA2
ATM
Rb
*Also known as TP53.
222 REVIEW
Spontaneous tumors/other
phenotypes
Intestinal tumors
Lymphomas, sarcomas, others
No spontaneous cancers, cells
defective in G1 arrest after
damage
−/− lethal, +/− develop
pheochromocytomas and
myelogenous leukemia
−/− lethal, +/− no tumor
phenotype
−/− lethal, +/− no tumor
phenotype
Lymphomas
−/− lethal, +/− develop
pituitary tumors
Table 4. Examples of knockout mouse strains with inflammation-related
genes inactivated
Gene
COX-1/COX-2
Alox5
Alox12
IL1
IL1r
IL2
IL2r (␣ and ␤)
Ifn␥
Nos2
TNF
Gene product
Cyclooxygenase 1 and 2
Arachidonate 5-lipoxygenase
Arachidonate 12lipoxygenase
Interleukin 1
Interleukin 1 receptor
Interleukin 2
Interleukin 2 receptor,
(alpha and beta
chains)
Interferon gamma
Nitric oxide synthetase 2
Tumor necrosis factor
Function
Arachidonic acid metabolism
Arachidonic acid metabolism
Arachidonic acid metabolism
Immune- and
inflammation-regulating
cytokine
Cytokine receptor
Immune-regulating cytokine
Cytokine receptors
Inflammatory cytokine
Generation of nitric oxide
Inflammatory cytokine
tained by the Jackson Laboratory (http://www.jax.org/resources/
documents/imr).
As reviewed (105), many of these knockout and transgenic
mouse models have been used effectively in toxicology and
carcinogenesis studies. In addition, some models have already
been used in chemoprevention studies (109). For example, ethylnitrosourea-induced T-cell lymphomas in pim-1-transgenic
mice, which overexpress the pim-1 oncogene in their lymphoid
compartments and are susceptible to genotoxic carcinogens that
target the lymphoid system, are delayed by the synthetic retinoid
fenretinide (110). In addition, spontaneous neurofibromatoses in
transgenic mice overexpressing part of the human Tlymphotrophic virus genome are delayed by a diet supplemented
with red wine solids rich in catechin, a polyphenolic compound
with potent antioxidant activity (111). These studies further support the utility of genetic susceptibility models for cancer prevention studies.
Two different transgenic mouse models of prostate cancer
were developed that are based on prostatic overexpression of the
simian virus 40 large T antigen (which can bind to and inactivate
the p53 tumor suppressor protein and the retinoblastoma tumor
suppressor protein) (112,113). Because progress in prostate cancer chemoprevention has been hampered by the lack of relevant
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
animal models and because both of these models have pathologic and molecular characteristics similar to those observed in
human prostate tumor progression, these models are currently
being evaluated for their usefulness for testing prostate cancer
chemoprevention strategies.
As discussed above, alterations in genes encoding carcinogen-metabolizing or -detoxifying enzymes, DNA repair enzymes, or factors involved in the epigenetic events associated
with tumor promotion and progression can greatly influence
tumor development. Most mice carrying such genetic alterations
do not develop tumors without additional treatment and hence
are not generally considered tumorigenesis models reflective of
human carcinogenesis. However, such animals may be extremely useful for studying specific gene–environment interactions on their own or when used in combination with other
models of genetic susceptibility. For example, in an ongoing
study, we evaluated (114) several dietary interventions and chemopreventive agents in p53-deficient mice crossed with Wnt-1
transgenic mice, which spontaneously develop mammary tumors
at an early age relative to Wnt-1 transgenic mice without p53
deficiency. Another example of the usefulness of such models is
the finding that Min mice crossed with MSH2 knockout mice,
which are deficient in a critical mismatch repair gene, develop
more adenomas at a much earlier age than do Min mice that have
normal MSH2 function (115). This type of bi-transgenic approach
should prove to be very useful for studying gene–gene and gene–
environment interactions relevant to cancer prevention.
SUMMARY
As discussed above, strategies for the nutritional modulation
and chemoprevention of cancer should focus on stopping carcinogenesis at the earliest possible point in the pathway. The purpose of this review is to examine the specific points along the
carcinogenesis pathway that may be amenable to mechanismbased prevention strategies. Ways of interfering with tumor initiation include modifying carcinogen activation by inhibiting the
enzymes responsible for that activation or by directly scavenging DNA-reactive electrophiles, enhancing carcinogen detoxification by altering the activity of detoxifying enzymes, and enhancing DNA repair processes. Ways of blocking promotion and
progression include altering the expression of genes involved in
cell signaling (particularly those regulating cell proliferation,
apoptosis, and differentiation) and decreasing inflammation. Examples of dietary factors and chemopreventive agents that affect
each of these processes were provided. Experimental models of
carcinogen-induced cancer have been crucial to advancing our
understanding of the influence of diet and chemopreventives on
chemical carcinogenesis, and future progress may be facilitated
by the integration of genetically altered animals into cancer prevention studies. An additional promising area for future studies
may also be the use of multiple agents, particularly combinations
of agents that work on different stages (e.g., anti-initiators in
combination with antipromoters) or on different mechanisms
within the same stage (e.g., anti-inflammatory agents in combination with inhibitors of proliferation).
REFERENCES
(1) Yuspa SH, Shields PG. Etiology of cancer: chemical factors. In: DeVita
VT Jr, Hellman SH, Rosenberg SA, editors. Cancer: principles and practice of oncology. 5th ed. Philadelphia (PA): Lippincott-Raven; 1997. p.
185–202.
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
(2) Slaga T J, Fischer SM, Weeks CE, Klein-Szanto AJP. Multistage carcinogenesis. In: Seije M, Bernstein IA, editors. Biochemistry of normal and
abnormal epidermal differentiation. Tokyo: University of Tokyo Press;
1980. p. 19–38.
(3) Slaga TJ. Mechanisms involved in two-stage carcinogenesis in mouse
skin. In: Slaga TJ, editor. Mechanisms of tumor promotion. Boca Raton
(FL): CRC Press; 1984. p. 1–93.
(4) Pitot HC. Progression: the terminal stage in carcinogenesis. Jpn J Cancer
Res 1989;80:599–607.
(5) Yuspa SH, Poirier MC. Chemical carcinogenesis: from animal models to
molecular models in one decade. Adv Cancer Res 1988;50:25–70.
(6) Fearon ER, Vogelstein B. A genetic model of human colorectal tumorigenesis. Cell 1990;61:759–67.
(7) Sugimura T. Multistep carcinogenesis: a 1992 perspective. Science 1992;
258:603–7.
(8) Spitz MR, Bondy ML. Genetic susceptibility to cancer. Cancer 1993;72(3
Suppl):991–5.
(9) Gonzalez FJ. Genetic polymorphism and cancer susceptibility: fourteenth
Sapporo Cancer Seminar. Cancer Res 1995;55:710–5.
(10) Drinkwater NR, Bennett LM. Genetic control of carcinogenesis in experimental animals. Prog Exp Tumor Res 1991;33:1–20.
(11) Stanley LA. Molecular aspects of chemical carcinogenesis: the roles of
oncogenes and tumour suppressor genes. Toxicology 1995;96:173–94.
(12) Weinberg RA. Oncogenes and tumor suppressor genes. CA Cancer J Clin
1994;44:160–70
(13) Weinstein IB. The origins of human cancer: molecular mechanisms of
carcinogenesis and their implications for cancer prevention and treatment—twenty-seventh G.H.A. Clowes memorial award lecture. Cancer
Res 1988;48:4135–43.
(14) Nebert DW, Nelson DR, Coon MJ, Estabrook RW, Feyereisen R, FujiiKuriyama Y, et al. The P450 superfamily: update on new sequences, gene
mapping, and recommended nomenclature [published erratum appears in
DNA Cell Biol 1991;10:397–8]. DNA Cell Biol 1991;10:1–14.
(15) Guengerich FP. Metabolic activation of carcinogens. Pharmacol Ther
1992;54:17–61.
(16) Eling TE, Thompson DC, Foureman GL, Curtis JF, Hughes MF. Prostaglandin H synthase and xenobiotic oxidation. Annu Rev Pharmacol Toxicol 1990;30:1–45.
(17) MacLeod MC, Slaga TJ. Multiple strategies for the inhibition of cancer
induction. Cancer Bull 1995;47:492–8.
(18) Miller EC, Miller JA. Searches for ultimate chemical carcinogens and
their reactions with cellular macromolecules. Cancer 1981;47:2327–45.
(19) Perchellet JP, Perchellet EM, Gali HU, Gao XM. Oxidant stress and
multistage carcinogenesis. In: Mukhtar H, editor. Skin cancer: mechanisms and human relevance. Boca Raton (FL): CRC Press; 1995. p.
145–80.
(20) Mantle TJ, Pickett CB, Hayes JD, editors. Glutathione S-transferases and
carcinogenesis. Philadelphia (PA): Taylor and Francis; 1987. p. 1–121.
(21) Ketterer B. Protective role of glutathione and glutathione transferases in
mutagenesis and carcinogenesis. Mutat Res 1988;202:343–61.
(22) Oesch F, Doehmer J, Friedberg T, Glatt H, Oesch-Bartlomowicz B, Platt
KL, et al. Control of ultimate mutagenic species by diverse enzymes. Prog
Clin Biol Res 1990;340B:49–65.
(23) James MO, Schell JD, Boyle SM, Altman AH, Cromer EA. Southern
flounder hepatic and intestinal metabolism and DNA binding of benzo[a]pyrene (BaP) metabolites following dietary administration of low doses of
BaP, BaP-7,8-dihydrodiol, or a BAP metabolite mixture. Chem Biol Interact 1991;79:305–21.
(24) Mitchell DL, Adair GM, MacLeod MC, Tang MS, Nairn RS. DNA damage and repair in the initiation phase of carcinogenesis. Cancer Bull 1995;
47:449–55.
(25) Friedberg EC, Walker GC, Siede W. DNA repair and mutagenesis. Washington (DC): American Society of Microbiology Press; 1995. p. 1–79.
(26) Sancar A. Mechanisms of DNA repair. Science 1995;266:1954–6.
(27) Sancar A, Tang MS. Nucleotide excision repair. Photochem Photobiol
1993;57:905–21.
(28) Modrich P. Mismatch repair, genetic stability, and cancer. Science 1994;
266:1959–60.
(29) Chung FL, Kelloff G, Steele V, Pittman B, Zang E, Jiao D, et al. Chemopreventive efficacy of arylalkyl isothiocyanates and N-acetylcysteine
REVIEW 223
(30)
(31)
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(39)
(40)
(41)
(42)
(43)
(44)
(45)
(46)
(47)
(48)
(49)
(50)
(51)
for lung tumorigenesis in Fischer rats. Cancer Res 1996;56:
772–8.
Hecht SS. Chemoprevention by isothiocyanates. J Cell Biochem Suppl
1995; 22:195–209.
Stoner GD, Mukhtar H. Polyphenols as cancer chemopreventive agents. J
Cell Biochem Suppl 1995;22:169–80.
Morse MA, Wang CX, Stoner GD, Mandal S, Conran PB, Amin SG, et al.
Inhibition of 4-(methylnitrosoamino)-1-(3-pyridyl)-1-butanone-induced
DNA adduct formation and tumorigenicity in the lung of the F344 rat by
dietary phenethyl isothiocyanate. Cancer Res 1989;49:549–53.
Morse MA, Elkind KI, Amin SG, Hecht SS, Chung FL. Effects of alkyl
chain length on the inhibition of NNK-induced lung neoplasia in A/J mice
by arylalkyl isothiocyantes. Carcinogenesis 1989;10:1757–9.
Brady JF, Wang MH, Hong JY, Xiao F, Li Y, Yoo JS, et al. Modulation
of rat hepatic microsomal monooxygenase enzymes and cytotoxicity by
diallyl sulfide. Toxicol Appl Pharmacol 1991;108:342–54.
Brady JF, Ishizaki H, Fukuto JM, Lin MC, Fadel A, Gapac JM, et al.
Inhibition of cytochrome P-450 2E1 by diallyl sulfide and its metabolites.
Chem Res Toxicol 1991;4:642–7.
Chen L, Lee M, Hong JY, Huang W, Wang E, Yang CS. Relationship
between cytochrome P450 2E1 and acetone catabolism in rats as studied
with diallyl sulfide as an inhibitor. Biochem Pharmacol 1994;48:
2199–205.
Murray RD, Mendez J, Brown SA, editors. The naturally occurring coumarins: occurrence, chemistry and biochemistry. New York (NY): John
Wiley & Sons; 1982. p. 97–111.
Bickers DR, Mukhtar H, Molica SJ Jr, Pathak MA. The effect of psoralens
on hepatic and cutaneous drug metabolizing enzymes and cytochrome
P-450. J Invest Dermatol 1982;79:201–5.
Cai Y, Kleiner H, Johnston D, Dubowski A, Bostic S, Ivie W, et al. Effect
of naturally occurring coumarins on the formation of mouse epidermal
DNA adducts and skin tumors induced by benzo[a]pyrene and 7,12dimethylbenz[a]anthracene. Carcinogenesis 1997;18:1521–7.
Kootstra A, Haas BL, Slaga TJ. Reactions of benzo[a]pyrene diolepoxides with DNA and nucleosomes in aqueous solution. Biochem Biophys Res Commun 1980;94:1432–8.
Wood AW, Sayer JM, Newmark HL, Yagi H, Michaud DP, Jerina DM,
et al. Mechanism of the inhibition of mutagenicity of a benzo[a]pyrene
7,8-diol 9,10-epoxide by riboflavin 5⬘-phosphate. Proc Natl Acad Sci
U S A 1982;79:5122–6.
Wood AW, Huang MT, Chang RL, Newmark HL, Lehr RE, Yagi H, et al.
Inhibition of the mutagenicity of bay-region diol epoxides of polycyclic
aromatic hydrocarbons by naturally occurring plant phenols: exceptional
activity of ellagic acid. Proc Natl Acad Sci U S A 1982;79:5513–7.
Lesca P. Protective effects of ellagic acid and other plant phenols on
benzo[a]pyrene-induced neoplasia in mice. Carcinogenesis 1983;4:
1651–3.
Katiyer SK, Mukhtar H. Tea in chemoprevention of cancer. Epidemiologic and experimental studies. Int J Oncol 1996;8:221–38.
Sparnins VL, Barany G, Wattenberg LW. Effects of organosulfur compounds from garlic and onions on benzo[a]pyrene-induced neoplasia and
glutathione S-transferase activity in the mouse. Carcinogenesis 1988;9:
131–4.
Sumiyoshi H, Wargovich MJ. Chemoprevention of 1,2-dimethylhydrazine-induced colon cancer in mice by naturally occurring organosulfur
compounds. Cancer Res 1990;50:5084–7.
Rao CV, Tokomo K, Kelloff G, Reddy BS. Inhibition of dietary oltipraz
of experimental intestinal carcinogenesis induced by azoxymethane in
male F344 rats. Carcinogenesis 1991;12:1051–5.
Jang M, Cai L, Udeani GO, Slowing KV, Thomas CF, Beecher CW, et al.
Cancer chemopreventive activity of resveratrol, a natural product derived
from grapes. Science 1997;275:218–20.
Srivastava VK, Busbee DL. Decreased fidelity of DNA polymerases and
decreased excision repair in aging mice: effects of caloric restriction.
Biochem Biophys Res Commun 1992;182:712–21.
Haley-Zitlin V, Richardson A. Effect of dietary restriction on DNA repair
and DNA damage. Mutat Res 1993;295:237–45.
Weisburger JH, Rivenson A, Kingston DG, Wilkins TD, Van Tassell RL,
Nagao M, et al. Dietary modulation of the carcinogenicity of the heterocyclic amines. Princess Takamatsu Symp 1995;23:240–50.
224 REVIEW
(52) Birt DF, Julius AD, Runice CE, White LT, Lawson T, Pour PM. Enhancement of BOP-induced pancreatic carcinogenesis in selenium-fed
Syrian golden hamsters under specific dietary conditions. Nutr Cancer
1988;11:21–33.
(53) James SJ, Muskhelishvili L. Rates of apoptosis and proliferation vary with
caloric intake and may influence incidence of spontaneous hepatoma in
C57BL6 × C3H F1 mice. Cancer Res 1994;54:5508–10.
(54) Thompson HJ, Strange R, Schedin PJ. Apoptosis in the genesis and prevention of cancer. Cancer Epidemiol Biomarkers Prev 1992;1:597–602.
(55) Slaga TJ, Klein-Szanto AJ, Fischer SM, Weeks CE, Nelson K, Major S.
Studies on the mechanism of action of anti-tumor promoting agents: their
specificity in two-stage promotion. Proc Natl Acad Sci U S A 1980;77:
2251–4.
(56) Blumberg PM. Protein kinase C as the receptor for the phorbol ester tumor
promoters: sixth Rhoads memorial award lecture. Cancer Res 1988;48:
1–8.
(57) Fischer SM, DiGiovanni J. Mechanisms of tumor promotion: epigenetic
changes in cell signaling. Cancer Bull 1995;47:456–63.
(58) DiGiovanni J. Multistage carcinogenesis in mouse skin. Pharmacol Ther
1992;54:63–128.
(59) Furstenberger G, Marks F. Prostaglandins, epidermal hyperplasia and skin
tumor promotion. In: Honn KV, Marnett LJ, editors. Prostaglandins,
leukotrienes and cancer. Boston (MA): Martinus-Nijhoff; 1985.
p. 22–37.
(60) Sigal E. The molecular biology of mammalian arachidonic acid metabolism. Am J Physiol 1991;260(2 Pt 1):L13–28.
(61) Simonson MS, Wolfe JA, Dunn MJ. Regulation of prostaglandin synthesis by differential expression of the gene encoding prostaglandin endoperoxide synthase. In: Sammuelsson B, editor. Advances in prostaglandin,
thromboxane and leukotiene research. Vol. 21. New York (NY): Raven
Press; 1990. p. 69–82.
(62) Harris CC. p53: at the crossroads of molecular carcinogenesis and risk
assessment. Science 1993;262:1980–1.
(63) Smith ML, Chen IT, Zhan Q, O’Connor PM, Fornace AJ Jr. Involvement
of the p53 tumor suppressor in repair of u.v.-type DNA damage. Oncogene 1995;10:105–9.
(64) Livingstone LR, White A, Sprouse J, Livanos E, Jacks T, Tlsty TD.
Altered cell cycle arrest and gene amplification potential accompany loss
of wild-type p53. Cell 1992;70:923–35.
(65) Guinn BA, Mills KI. p53 mutations, methylation and genomic instability
in the progression of chronic myeloid leukaemia. Leuk Lymphoma 1997;
26:211–26.
(66) Trush MA, Kensler TW. An overview of the relationship between oxidative stress and chemical carcinogenesis. Free Radic Biol Med 1991;10:
201–9.
(67) Hursting SD, Kari FW. Inhibition of tumorigenesis by calorie restriction.
Cancer Bull 1995;47:505–10.
(68) Birt DF, Kris ES, Luthra R. Modification of murine skin tumor promotion
by dietary energy and fat. In: Mukhtar H, editor. Skin cancer: mechanisms
and human relevance. Boca Raton (FL): CRC Press; 1995. p. 371–81.
(69) Ames BN, Shigenaga MK, Hagen TM. Oxidants, antioxidants, and the
degenerative diseases of aging. Proc Natl Acad Sci U S A 1993;90:
7915–22.
(70) Weindruch R, Walford RL, editors. The retardation of aging and disease
by dietary restriction. Springfield (IL): C. Thomas; 1988. p. 1–291.
(71) Weeks CE, Slaga TJ, Hennings H, Gleason GL, Bracken WM. Inhibition
of phorbol ester-induced tumor promotion in mice by vitamin A analog
and anti-inflammatory steroid. J Natl Cancer Inst 1979;63:401–6.
(72) Gensler HL, Sim DA, Bowden GT. Influence of the duration of topical
13-cis-retinoic acid treatment on inhibition of mouse skin tumor promotion. Cancer Res 1986;46:2767–70.
(73) Verma AK, Shapas BG, Rice HM, Boutwell RK. Correlation of the inhibition by retinoids of tumor promoter-induced mouse epidermal ornithine decarboxylase activity and of skin tumor promotion. Cancer Res
1978;39(2 Pt 1):419–25.
(74) Pegg AE. Polyamine metabolism and its importance in neoplastic growth
and a target for chemotherapy. Cancer Res 1988;48:759–74.
(75) Petkovich M, Brand NJ, Krust A, Chambon P. A human retinoic acid
receptor which belongs to the family of nuclear receptors. Nature 1987;
330:444–50.
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
(76) Lotan R. Retinoids in cancer chemoprevention. FASEB J 1996;10:
1031–9.
(77) Lotan R. Retinoids and apoptosis: implications for cancer chemoprevention and therapy [editorial]. J Natl Cancer Inst 1995;87:1655–7.
(78) Birt DF, White LT, Choi B, Pelling JC. Dietary fat effects on the initiation
and promotion of two-stage skin tumorigenesis in the SENCAR mouse.
Cancer Res 1989;49:4170–4.
(79) Sandler RS. Aspirin and other nonsteroidal anti-inflammatory agents in
the prevention of colorectal cancer. Important Adv Oncol 1996;96:
123–37.
(80) Tsujii M, DuBois RN. Alterations in cellular adhesion and apoptosis in
epithelial cells overexpressing prostaglandin endoperoxide synthase 2.
Cell 1995;83:493–501.
(81) Slaga TJ, DiGiovanni J. Inhibition of chemical carcinogenesis. In: Searle
CE, editor. American Chemical Society Monograph 182. Vol 2. Washington (DC): American Chemical Society; 1984. p. 1279–94.
(82) Hursting SD. Experimental models of gene-environment interaction for
cancer chemoprevention studies. Curr Opin Oncol 1997;9:487–91.
(83) Dove WF, Gould KA, Luongo C, Moser AR, Shoemaker AR. Emergent
issues in the genetics of intestinal neoplasia. Cancer Surv 1995;25:
335–55.
(84) Moser AR, Pitot HC, Dove WF. A dominant mutation that predisposes to
multiple intestinal neoplasia in the mouse. Science 1990;247:322–4.
(85) Jacoby RF, Marshall DJ, Newton MA, Novakovic K, Tutsch K, Cole CE,
et al. Chemoprevention of spontaneous intestinal adenomas in the Apc
Min mouse model by the nonsteroidal anti-inflammatory drug piroxicam.
Cancer Res 1996;56:710–4.
(86) Boolbol SK, Dannenberg AJ, Chadburn A, Martucci C, Guo XJ, Ramonetti JT, et al. Cyclooxygenase-2 overexpression and tumor formation are
blocked by sulindac in a murine model of familial adenomatous polyposis.
Cancer Res 1996;56:2556–60.
(87) Beazer-Barclay Y, Levy DB, Moser AR, Dove WF, Hamilton SR, Vogelstein B, et al. Sulindac suppresses tumorigenesis in the Min mouse.
Carcinogenesis 1996;17:1757–60.
(88) Oshima M, Takahashi M, Oshima H, Tsutsumi M, Yazawa K, Sugimura
T, et al. Effects of docosahexaenoic acid (DHA) on intestinal polyp development in Apc knockout mice. Carcinogenesis 1995;16:2605–7.
(89) Fodde R, Edelmann W, Yang K, van Leeuwen C, Carlson C, Renault B,
et al. A targeted chain-termination mutation in the mouse Apc gene results
in multiple intestinal tumors. Proc Natl Acad Sci U S A 1994;91:8969–73.
(90) Oshima M, Dinchuk JE, Kargman SL, Oshima H, Hancock B, Kwong E,
et al. Suppression of intestinal polyposis in Apc delta716 knockout mice
by inhibition of cyclooxygenase 2 (COX-2). Cell 1996;87:803–9.
(91) Hollstein M, Sidransky D, Vogelstein B, Harris CC. p53 mutations in
human cancers. Science 1991;253:49–53.
(92) Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CA
Jr, Butel JS, et al. Mice deficient for p53 are developmentally normal but
susceptible to spontaneous tumors. Nature 1992;356:215–21.
(93) Kemp CJ, Donehower LA, Bradley A, Balmain A. Reduction of p53 gene
dosage does not increase initiation or promotion but enhances malignant
progression of chemically induced skin tumors. Cell 1993;74:813–22.
(94) Hursting SD, Perkins SN, Phang JM. Calorie restriction delays spontaneous tumorigenesis in p53-knockout transgenic mice. Proc Natl Acad Sci
U S A 1994;91:7036–40.
(95) Hursting SD, Perkins SN, Brown CC, Haines DC, Phang JM. Calorie
restriction induces a p53-independent delay of spontaneous carcinogenesis in p53-deficient and wild-type mice. Cancer Res 1997;57:2843–6.
(96) Hursting SD, Perkins SN, Haines DC, Ward JM, Phang JM. Chemoprevention of spontaneous tumorigenesis in p53-knockout mice. Cancer Res
1995;55:3949–53.
(97) Perkins SN, Hursting SD, Haines DC, James SJ, Miller BJ, Phang JM.
Chemoprevention of spontaneous tumorigenesis in nullizygous p53deficient mice by dehydroepiandrosterone and its analog, 16-␣-fluoro-5androsten-17-one. Carcinogenesis 1997;18:989–94.
(98) Schwartz AG, Pashko LL. Mechanism of cancer preventive action of
DHEA. Role of glucose-6-phosphate dehydrogenase. Ann N Y Acad Sci
1995;774:180–6.
(99) Wang TT, Hursting SD, Perkins SN, Phang JM. Effects of dehydroepiandrosterone and calorie restriction on the Bcl-2/Bax-mediated apoptotic
pathway in p53-deficient mice. Cancer Lett 1997;116:61–9.
(100) Mei JM, Hursting SD, Perkins SN, Phang JM. p53-independent inhibition
of nitric oxide generation by cancer preventive interventions in ex vivo
mouse peritoneal macrophages. Cancer Lett 1998;129:191–7.
(101) Tennant RW, French JE, Spalding JW. Identifying chemical carcinogens
and assessing potential risk in short-term bioassays using transgenic
mouse models. Environ Health Perspect 1995;103:942–50.
(102) Harvey M, McArthur MJ, Montgomery CA Jr, Butel JS, Bradley A,
Donehower LA. Spontaneous and carcinogen-induced tumorigenesis in
p53-deficient mice. Nat Genet 1993;4:225–9.
(103) Kemp CJ, Wheldon T, Balmain A. p53-deficient mice are extremely
susceptible to radiation-induced tumorigenesis. Nat Genet 1994;8:66–9.
(104) Hursting S, Purewal M, Woods C, Velasco M, Wargovich M. Chemoprevention of azoxymethane-induced aberrant crypt foci and colon tumors
in p53-deficient mice by sulindac. Cancer Epidemiol Biomarkers Prev
1998;7:176.
(105) Gonzalez FJ. Use of transgenic animals in carcinogenesis studies. Mol
Carcinog 1996;16:63–7.
(106) Allemand I, Angulo JF. Transgenic and knockout models for studying
DNA repair. Biochimie 1995;77:826–32.
(107) Palapattu GS, Bao S, Kumar TR, Matzuk MM. Transgenic mouse models
for tumor suppressor genes. Cancer Detect Prev 1998;22:75–86.
(108) Friedberg EC, Meira LB, Cheo DL. Database of mouse strains carrying
targeted mutations in genes affecting cellular responses to DNA damage.
Mutat Res 1997;383:183–8.
(109) Hursting SD. Experimental models of gene-environment interaction for
cancer chemoprevention studies. Curr Opin Oncol 1997;9:487–91.
(110) McCormick DL, Johnson WD, Rao KV, Bowman-Gram T, Steele V,
Lubet RA, et al. Comparative activity of N-(4-hydroxyphenyl)-all-transretinamide and alpha-ifluoromethylornithine as inhibitors of lymphoma
induction in PIM transgenic mice. Carcinogenesis 1996;17:2513–7.
(111) Clifford AJ, Ebeler SE, Ebeler JD, Bills ND, Hinrichs SH, Teissedre PP,
et al. Delayed tumor onset in transgenic mice fed an amino acid-based diet
supplemented with red wine solids. Am J Clin Nutr 1996;64:748–56.
(112) Shibata M, Ward JM, Devor DE, Liu M, Green JE. Progression of prostatic intraepithelial neoplasia to invasive carcinoma in C3(1)/SV40 large
T antigen transgenic mice: histopathological and molecular biological
alterations. Cancer Res 1996;56:4894–903.
(113) Gingrich JR, Greenberg NM. A transgenic mouse prostate cancer model.
Toxicol Pathol 1996;24:502–4.
(114) Donehower LA, Godley LA, Aldaz CM, Pyle R, Shi YP, Pinkel D, et al.
Deficiency of p53 accelerates mammary tumorigenesis in Wnt-1 transgenic mice and promotes chromosomal instability. Genes Dev 1995;9:
882–95.
(115) Reitmair AH, Cai JC, Bjerknes M, Redston MT, Cheng H, Pind MT, et al.
MSH2 deficiency contributes to accelerated APC-mediated intestinal tumorigenesis. Cancer Res 1996;56:2922–6.
NOTE
Manuscript received March 4, 1998; revised November 18, 1998; accepted
December 11, 1998.
Journal of the National Cancer Institute, Vol. 91, No. 3, February 3, 1999
REVIEW 225