FW: tech: Hormesis, toxins and cancer

[Guest guest]

Hi All,

The below is a review for you to read. I recommend the whole nine yards.

The PDF is available. It is on hormesis, “toxins” and cancer. It gives me

an idea of how important hormesis is and how toxin doses impinges on the

importance of the effect on cancer.

Hormesis is defined as: " An effect where a toxic substance acts like a

stimulant in small doses, but it is an inhibitor in large doses " .

It seems that all things bad a high levels are good for you at low doses.

Saccharine prevents cancer – now there is a switch. Caffeic acid in

coffee has the same result. The studies on cells were complemented by other

studies in animals and measured cancer stages at the beginning, middle and

end. Many toxins have this hormesis effect. It seems like so many of us

CRONies are trying to get 0 levels of toxins that maybe we provide ourselves

with lower than optimal levels of good things for us.

I added tech: to the subject and the size of the review makes me wonder if

anyone has problems with it. It was a lot of work, but I think a very good

review. Personal or other feedback to those with difficulties with this are

requested.

I tried adding the text directly and hope the archives store it better than

it hopefully does not come out in email.

Cheers, Al.

Calabrese EJ, Baldwin LA.

Can the concept of hormesis be generalized to carcinogenesis?

Regul Toxicol Pharmacol. 1998 Dec;28(3):230-41.

PMID: 10049795 [PubMed - indexed for MEDLINE]

“The concept of hormesis (i.e., low-dose stimulation/

high-dose inhibition) has been shown to be widely gen-eralizable

with respect to chemical class, animal model,

gender, and biological end point. The public health im-plication

of this lack of linearity in the low-dose area of

the dose–response curve raises the question of whether

low doses of carcinogens will reduce cancer risk. Arti-cles

relating to the process of carcinogenesis (i.e., initi-ation,

promotion, tumor development, and progression)

were obtained from a recently developed chemical

hormesis database and evaluated for their evidence of

hormesis. Numerous examples in well-designed studies

indicate that U- or J-shaped dose–response relation-ships

exist with respect to various biomarkers of carci-nogenesis

in different animal models of both sexes. Ex-amples

of such J-shaped dose–response relationships in

each stage of the process of carcinogenesis were selected

for detailed toxicological examination. These results

have important implications for both the hazard assess-ment

of carcinogens and cancer risk assessment proce-dures.

INTRODUCTION

A question commonly asked concerning the public

health implications of hormesis is whether low doses of

carcinogens will reduce cancer risk. The concept of

hormesis (i.e., low-dose stimulation/high-dose inhibition)

is counter to the cancer risk assessment practices by U.S.

regulatory agencies such as the EPA, FDA, and OSHA

which assume that cancer risk is linear in the low-dose

area (Fig. 1). Nonetheless, the recognition that hormetic

responses are widely generalizable with respect to chem-ical

class, animal model, gender, and biological end

point (Calabrese and Baldwin, 1997a, suggests that

the process of carcinogenesis should likewise be an end

point where hormetic responses could be anticipated.

However, this question is not as readily approached as

would be expected since most cancer bioassays involve

only a very limited number of relatively high doses. In

addition to such challenges confronting the assessment

of hormesis with respect to cancer in animal models,

the issue is no less challenging for epidemiological

investigations. Even in the case of positive cancer stud-ies

increases in relative risk of less than two- to three-fold

are very difficult to resolve with adequate confi-dence.

This would also be very difficult to resolve if one

were to expect that a hormetic response may reduce

human cancer incidence by 30–60%.

FIG. 1. (A) The most common form of the hormetic dose–re-sponse

curve depicting low-dose stimulatory and high-dose inhibi-tory

responses, the b- or inverted U-shaped curve. Examples of end

points demonstrating the b- or inverted U-shaped dose–response

curve include growth, longevity, fecundity, and weight gain. ( The

hormetic dose–response curve depicting low-dose reduction and

high-dose enhancement of adverse effects, the J- or U-shaped curve.

Examples of end points demonstrating the J- or U-shaped dose–

response curves include mutations, cancer incidence, and birth-de-fects

incidence.

As a result of limitations in study design (e.g., num-ber

and range of doses and selection of end point), the

hormetic hypothesis has been much more extensively

evaluated in other less expensive, though highly com-plex

models, such as bacteria, fungi, plants, inverte-brates,

and fish. Nevertheless, there are a number of

well-conducted toxicological investigations that have

addressed the issue of carcinogenesis relevant to the

hormetic hypothesis. However, with very few excep-tions

the linkage of hormesis to the cancer bioassay

was only introduced at the time of interpretation of the

toxicological data rather than at the time of hypothesis

development and study design formulation. This rep-resents

an important consideration given the critical

study design requirements for assessing the hormesis

hypothesis (Calabrese and Baldwin, 1997b).

Despite the above concerns this paper assesses

whether the hormetic hypothesis can be generalized to

include the process of carcinogenesis. The hormesis da-tabase

was assessed for articles relating to the process of

carcinogenesis; relevant articles were then assigned to

one of the three stage categories [i.e., early (initiation),

middle (promotion), or late (tumor development and pro-gression)].

The number of relevant articles was modest

(i.e., approaching 20), but distributed with good balance

across the various stages of carcinogenesis. Each of the

articles was evaluated by application of the a priori

hormesis database criteria and ranked for their evidence

of hormesis (Calabrese and Baldwin, 1997b). Each stage

of the process of carcinogenesis was evaluated separately

and then integrated in an overall discussion.

EARLY STAGE (INITIATION)

Early-stage processes of carcinogenesis would be ex-pected

to include the initiation of DNA damage, at-

tempts to repair damage, factors affecting the perma-nence

of the genetic alteration, and production of

altered proteins. To assess the hormetic hypothesis it

is necessary that the control cells have a sufficiently

elevated background damage incidence that could be

lessened by a low-level exposure. If there was negligi-ble

background damage it would be impossible within

the framework of a U-shaped dose–response assump-tion

to test hypotheses relating to hormesis.

1. Response of Human Keratinocytes to Very Low

Concentrations of N-Methyl-N9-nitro-N-nitrosoguanidine (MNNG)

In 1996 Kleczkowska and Althaus assessed the re-sponse

of human keratinocytes to low concentrations of

the well-known methylating agent MNNG. This eval-uation

included an assessment of the effects of MNNG

on DNA integrity, poly(ADP)-ribose metabolism, clono-genic

survival, and DNA synthesis. These researchers

assessed the effects of the MNNG over a 500,000-fold

concentration range. At high doses DNA unwinding

was more extensive than in the untreated controls as

expected. However, as the concentration was reduced

to a 50 to 0.05 nM range, DNA unwinding was signif-icantly

reduced relative to the untreated controls (Fig. 2). In fact, cells treated

with these low concentrations

had fewer DNA strand breaks than untreated controls.

FIG. 2. Dose dependence of the MNNG effect on DNA integrity

in human keratinocytes. Confluent cells (5–8 days after subculture)

were exposed to the indicated concentrations of MNNG for 60 min.

The percentage of dsDNA (double-stranded DNA) remaining after

alkaline unwinding in cells not exposed to MNNG represents 100.

The points show mean values of triplicate samples of one (0.05 nM)

and nine (2500 nM) experiments. Error bars, SEM (source: Klecz-kowska and

Althaus, 1996).

Time-course experimentation revealed that at low con-centrations

the MNNG induced DNA alterations (e.g.,

strand breaks) that were subsequently repaired in a

manner consistent with the overcompensation theory

of hormetic effects proposed by Stebbing (1997). The

mechanism by which strand breaks were repaired was

related to the enhanced activity of poly(ADP)-ribose by

the low-dose concentrations of MNNG. Furthermore,

colony formation of human keratinocytes in the low-dose

range was demonstrated to have a higher survival

rate than the untreated controls by nearly 50%.

The authors concluded with the statement that the

“effect of MNNG on human cells at extremely low, but

environmentally relevant doses seems to be quite dif-ferent

from what would be extrapolated from toxicity

testing at higher doses.”

2. Effects of Mercury on DNA Repair Enzyme Activity

in Human Buccal Cells

O 6 -Methylguanine–DNA methyltransferase (MGMT)

repairs premutagenic O 6 -methylguanine lesions in-duced

in DNA by alkylating agents. MGMT serves as

both a transferase and an acceptor for the alkyl group

by forming S-alkylcysteine at the catalytic site of the

protein. This activity results in the restoration of intact

DNA along with the functional inactivation of MGMT.

It is believed that the level of MGMT expression plays

a critical role in affecting the degree of protection

against toxic, mutagenic, and carcinogenic effects of

alkylating agents. In 1997 Liu et al. assessed the ef-fects

of Hg 21 on MGMT activity of human buccal fibro-blasts.

Exposure of the cells to seven doses of Hg 21 over

a 300-fold range revealed a typical b-curve with the

highest three doses effecting a dose-dependent de-crease

in MGMT activity, while at the lower doses

increased activity was observed (Fig. 3).

FIG. 3. Cell survival and MGMT activity of human buccal fibro-blasts

exposed to Hg 21 . The results were expressed as means 1 SEM

and derived from three separate experiments with duplicate dishes

in each experiment (source: Liu et al., 1997).

3. Effects of Styrene on Chromosomal Aberrations in Humans

An assessment of the potential for occupational ex-posure

to styrene to cause chromosomal aberrations

was assessed by Camurri et al. (1983). Assuming that

the lowest exposure group would serve as the control,

the authors assessed the impact of a 14-fold range in

airborne concentrations of styrene. A dose-dependent

decrease in chromosomal aberrations was reported for

the three lowest treatment groups with a 20–46% de-crease

in aberrations reported.

4. Effects of Mutagens on Rat Liver DNA

Kitchin and Brown (1994, 1996) published two ex-tensive

papers on the effects of various mutagens on

the occurrence of rat liver DNA damage over a large

number of doses and an extremely broad dosage range.

DNA damage was experimentally assessed in female

Sprague–Dawley rat liver to obtain insight on the na-ture

of dose–response curves for chemical carcinogen-esis.

DNA damage was selected as the measurement

end point since all agents observed to damage hepatic

DNA were also rodent carcinogens. Dose–response

curves for rat hepatic DNA damage were reported over

an unusually wide dose range of up to six orders of

magnitude. With limited exception, the lower doses

selected were usually 1/10, 1/100, 1/1000, or 1/10,000 of

the initial dose (i.e., usually 1/5 of the LD50 ). Of 49

rodent liver carcinogens initially selected for study, 12

produced DNA damage: 1,2-dibromoethane, 1,2-dibromo-3-

chloropropane, 1,2-dichoroethane, 1,4-dioxane, methyl-ene

chloride, auramine O, Michler’s ketone, selenium

sulfide, 1,3-dichloropropene, 1,2-dimethylhydrazine, N-nitroso-

piperidine, and butylated hydroxytoluene (BHT).

Eleven dose–response curves (i.e., with the exception of

BHT) fit a linear model well (r 2 5 0.886) but a quadratic

model better (r 2 5 0.947). Of the 11 chemical data sets,

the quadratic regression analysis yielded a negative lin-ear

slope for all agents when plotted against logged but

not unlogged dose. The authors concluded that the data

“happened to have random variation around the control

values which makes some treated values lower than con-trol

values and causes the overall logged regression curve

to dip below zero at low doses.” They further state that

the shape of the dose–response curve was “exactly the

opposite of what would be predicted on the basis of bio-chemical

and pharmacological theories based on the law

of mass action.” Because of the simplicity of the linear

model with no y-axis intercept and the fairly high r 2 of

0.886 for 11 different DNA-damaging chemicals, the au-thors

favored the simpler linear regression model using

untransformed dose. The authors did not believe ex-tremely

low doses of chemical carcinogens actually de-crease

the degree of DNA damage found in treated ani-mals

or improve the animal’s health in any way (Kitchin

and Brown, 1994).

The interpretation of these findings was extensively

debated in an issue of the BELLE Newsletter (1995)

which ultimately demonstrated how difficult it is to

establish unequivocal support for the hormetic hypoth-esis

unless there are sufficient doses in the low-dose

range and that statistical power is adequate.

Assessment of the Mutagenicity Data

The four examples selected for review included two

human cell models, one occupational epidemiological

investigation, and a massive investigation with 49

chemical agents in Sprague–Dawley rat liver. The find-ings

involved the measurement of a variety of cellular

effects including DNA damage, DNA strand breaks,

colony survival, DNA repair capacity, and chromo-somal

aberrations. Despite the consistency of response

across studies it should be emphasized that the rank-ings

in the hormesis database were generally not im-pressive

as evidence supporting hormesis. For exam-ple,

all 12 chemicals assessed in the initial Kitchin and

Brown (1994) report received rankings of low evidence

of hormesis. The DNA repair reported by Liu et al.

(1997) also received a low ranking, as well as in the

very impressive report of Kleczkowska and Althaus (1996).

The reasons for these low rankings are principally a

function of the evaluation criteria. This ranking sys-

tem is designed to reward studies with a large number

of doses below the NOAEL, where statistical signifi-cance

is achieved for low-dose treatments and where

reproducibility of the findings is established. While

each study provided data to support the hormetic hy-pothesis,

each lacked some component that limited its

capacity to receive a high ranking in the present sys-tem.

MIDDLE STAGE (PROMOTION)

The promotional phase of the process of carcinogen-esis

has been explored in a limited fashion in the low-dose

range. Examples selected for assessment included

cell turnover of caffeic acid in the rat forestomach and

kidney, the effects of the tumor promoters [i.e., pheno-barbital

and dioxin (TCDD)] on altered hepatic foci

formation in diethylnitrosamine (DEN)-pretreated

partially hepatectomized rats, and urinary bladder hy-perplasia

in saccharin-treated rats. As will be seen,

these specific cases offer highly suggestive evidence

consistent with the hormetic hypothesis. These partic-ular

investigations will be seen to have additional sig-nificance

since in each instance these studies have

been linked to other research either extending the orig-inal

observations or using the findings to account for

what appears to be protective effects on the agent at

low doses in cancer bioassays.

1. The Effects of Caffeic Acid on the Stimulation of

Cell Division in Forestomach and Kidney of the Male Rat

Caffeic acid (3,4-dihydroxycinnamic acid) is a natu-ral

phenolic antioxidant that is broadly distributed in

vegetables, fruits, and beverages. Toxicological studies

have revealed that when administered at 2% in the

diet in a lifetime rat study caffeic acid induced hyper-plasia

and tumors in the forestomach and kidney of

F344 rats and B6C3F1 mice (Hagiwara et al., 1991). It

also promotes carcinogenesis following initiation by

the genotoxic carcinogens MNNG (Hirose et al., 1991,

1992) and 7,12-DMBA (Hirose et al., 1988). However,

caffeic acid was also shown to inhibit the formation of

squamous epithelial carcinomas of the rat tongue in-duced

by 4-nitroquinoline-1-oxide (Tanaka et al., 1993)

and of mouse forestomach tumors induced by BP

(Wattenberg et al., 1980). The dosages employed in the

cancer bioassays revealed that caffeic acid was carci-nogenic

at 2%, tumor promoting at 0.5–1%, and anti-carcinogenic

at 0.05–0.5% (Lutz et al., 1997). Such

dosage dependence with respect to tumor outcome sug-gested

that linear low-dose default extrapolation pro-cedures

may not be appropriate for application to caf-feic acid.

Caffeic acid is believed to enhance the process of

carcinogenesis via a non-genotoxic process since it did

not induce a mutagenic response in bacteria assays.

However, it did induce forward mutations in cultured

mouse lymphoma L5178Y cells and chromosomal ab-errations

in cultured Chinese hamster ovary cells

(Hanham et al., 1983). Such findings lead IARC (1993)

to suggest that caffeic acid might act via a reactive

oxygen species or tumor-promoting activity, possibly

via enhancement of cell division. This hypothesis was

subsequently evaluated by Lutz et al. (1997) in which

male F344 rats were fed caffeic acid at five different

dietary concentrations including the controls (i.e., 0,

0.05, 0.14, 0.40, and 1.64%) for 4 weeks. The total

number of epithelial cells per unit comparison in the

forestomach was increased nearly 2.5-fold at the two

highest concentrations, while no treatment effects

were observed at 0.05%, the lowest dose studied. How-ever,

at the 0.14% concentration there was a decrease

in these cellular responses by approximately one-third.

Observations in the kidney closely mirrored those seen

in the forestomach with the lowest concentration

(0.05%) having no treatment effect, the 0.14% group

responses were about one-third that of the controls,

and the highest doses were progressively greater than

the controls. These U-shaped dose–response relation-ships

in these two critical target organs display en-hanced

cellular division at the higher doses. Lutz et al.

(1997) suggested that the low-dose response of a de-layed

cell division may account to some extent for the

cancer-protective effect seen in previous bioassays as

noted above.

With respect to human exposure, the lowest dietary

concentration that produced a slight increase in the

labeling index was 0.4%, which would be approxi-mately

270 mg/kg body wt/day. The decrease of the

labeling index responsible for the U shape of the dose–

response relationship was equivalent to 35 mg/kg/day.

Therefore, even for strong coffee drinkers, who would

consume up to 10 mg caffeic acid/kg body wt/day, the

exposure would be in the range observed to be poten-tially

protective (i.e., in the hormetic zone) with respect

to effects on the rate of cell division.

2. The Effect of Phenobarbital and TCDD on DEN-Induced

Liver Foci Formation

Several published reports have quantified the rela-tive

initiating and promoting potencies of hepatocarci-nogenic

agents using altered hepatic foci. Of particular

interest were experiments that assessed the capacity of

a broad dose range of phenobarbital (PB), a known

promoter of hepatocarcinogenesis, to affect DEN-in-duced

precancerous changes in the liver of partially

hepatectomized rats (Pitot et al., 1987). In this exper-iment

foci formation was estimated by three markers:

GGT (glutamyl transpeptidase), adenosine triphospha-tase

(ATPase), and glucose 6-phosphatase (G6Pase).

The shape of the dose–response curves for the number

of altered hepatic foci and their volume as percentage

of liver revealed U-shaped appearances (Fig. 4). This

diminished response at the lowest dose rate of PB, both

with respect to numbers of foci and with their volume

percentages compared with control rats receiving no

PB, was similar to that reported by these authors in a

previous study (Goldsworthy et al., 1984) with the

Sprague–Dawley rat model. According to the authors,

these findings suggested a dose-dependent protective effect.

FIG. 4. Difference between (A) the altered hepatic foci (AHF)/liver and (

the volume percentage of AHF in the liver as a function of the percentage of

phenobarbital fed in the NIH-07 diet of Fischer 344 female rats for a

6-month period following initiation or not with DEN (10

mg/kg) (source: Pitot et al., 1987).

Similar investigations were undertaken using TCDD

as the tumor promoter in place of PB in the Pitot et al.

(1987) investigations. Striking U-shaped dose–re-sponse

relationships were also reported. While the au-thors

reported the apparent “protective” effect seen at

the lower doses, it was later noted that the TCDD

control groups were sacrificed at 8 rather than 6

months as in the treatment groups. This has recently

been interpreted to mean that the reported U-shaped

curves, especially that for the volume fraction, are

probably falsely enhanced compared to that which

would have been seen had the control been assessed

with the TCDD-exposed groups at 6 months. Subse-quent

evaluation of the 8-month control time point and

back extrapolating from 8 to 6 months suggested that

there is no U-shaped response behavior for the volume

fraction. However, the model predicted that the num-ber

of altered foci per liver would change little between

6 and 8 months, suggesting that the U-shaped re-sponse

for this end point may be a reproducible obser-

vation (Fig. 5). Teeguarden et al. (1996), in a prelimi-nary

report on a replication of the experiment of Pitot

et al. (1987), did not observe U-shaped behavior by

volume fraction and only an equivocal U-shaped re-sponse

for number of altered foci per liver. It should be

emphasized that the methodological and reproducibil-ity

issues relating to the TCDD investigations do not

apply to the above-noted findings of the DEN/PB ex-perimentation

(H. C. Pitot, personal communication, 1997).

FIG. 5. Difference between the altered hepatic foci (AHF)/liver in

animals initiated with DEN (10 mg/kg) or not and administered

TCDD at the dose levels noted. TCDD was injected intramuscularly

in corn oil biweekly at concentrations which resulted in the daily

dose shown. In this graph the x-axis is logarithmic (source: Pitot et

al., 1987).

Recently Conolly and Andersen (1997) have em-ployed

a quantitative, stochastic initiation–promotion

model to assess results for the 1987 study of Pitot et al.

using a negative-selection model of tumor promotion.

In their approach two types of initiated cells (A and

are assumed to be produced by DEN initiation. Excel-lent

agreement was shown between model predictions

and the Pitot et al. data when dose–response effects on

cell division and death rate of the two assumed altered

cell types were considered. For A cells, both cell divi-sion

and apoptosis rates increased, while the difference

between cell division and apoptosis rates declined. As

for B cells, the difference between division and apopto-sis

rates increases, principally because of a decrease in

the apoptosis rate. The authors also linked these

changes in cell kinetics to a pharmacokinetic model for

TCDD which incorporated a five-subcompartment

model of the liver acinus with induction of CYP1A1 and

1A2 genes. Changes in A cell kinetics were associated

with effects of TCDD in the region most sensitive to

induction (subcompartment 5, centrilobular region),

while B cell dynamics were associated with induction

in subcompartments 3–5 (centrilobular and midzonal

regions). The results of this modeling procedure indi-cated

that the two-cell model reproduced the data of

Pitot et al. (1987) and that induction of CYP1A1/1A2 in

different regions of the hepatic acinus can be employed

as a general correlate of this assumed alteration in cell

growth kinetics.

That TCDD may alter other aspects of the tumor

promotional process was suggested in the work of Fan

et al. (1996) who reported a striking U-shaped dose–

response relationship with respect to changes in cell-mediated

immunological responses in the Sprague–

Dawley rat. These authors cited the earlier work of

Chandra (1991) that a slight excess intake of certain

nutrients may be associated with enhanced immune

responses and that all nutrients given in quantities

greater than a certain threshold will reduce immune

responses. They suggested that the U-shaped response

may in fact be a much more common phenomenon than

is general assumed.

3. The Effects of Saccharin on Hyperplasia of the Urinary Bladder

In 1973 the FDA published the results of a multigen-eration

cancer bioassay concerning the artificial sweet-ener

sodium saccharin in which histology was conducted

only with F1 generation (OTA, 1977). Of relevance to the

concept of tumor promotion was the incidence of urinary

bladder hyperplasia data. In this investigation urinary

hyperplasia was evaluated over five doses as well as

concurrent controls for male and female rats. Six groups

of Sprague–Dawley rats were fed the saccharin-treated

diets. The data revealed an apparent U-shaped dose re-sponse

for both sexes for urinary bladder hyperplasia

(Table 1). Although both sexes displayed the U-shaped

dose response, the optimal range of the apparent de-crease

in the incidence of the hyperplasia is somewhat

different, being upshifted in the male. More specifically,

the optimal zone for the males was from 0.1 to 1.0% of the

diet, while for the females the optimal range was from

0.01 to 0.1%. Whether the male would have displayed a

further decrease between these two boundary values was

not explored. The data for the males were more striking

since the background incidence of the urinary hyperpla-sia

was considerably higher and thereby permitted a

greater opportunity to assess the possibility of a hormetic

response.

TABLE 1

Incidence of Urinary Bladder Hyperplasia for Male

and Female Sprague–Dawley Rats in the FDA 1973

Saccharin Study (Source: Downs and owski,

1982)

Dose (% No. hyperplasias (%)

of diet) Male Female

0 10/73 (14) 3/85 (4)

0.01 6/71 (8) 0/81 (0)

0.1 4/81 (5) 0/81 (0)

1.0 4/76 (5) 3/90 (3)

5.0 6/64 (9) 5/88 (6)

7.5 19/62 (31) 10/76 (13)

Discussion

The four examples of possible hormetic processes

affecting tumor promotion occurred in the rat model

with widely differing chemical agents (i.e., caffeic acid,

PB, TCDD, and saccharin) affecting different target

organs. To some extent these investigations have been

employed to explain reported nonlinear and in fact

U-shaped dose–response relationships in cancer bioas-says

for caffeic acid, TCDD, and saccharin. Such at-tempts

to link mechanistic investigations with cancer

bioassay findings offer the opportunity to assess not

only the significance of promotion in the process of

carcinogenesis of these specific agents but also the

significance of hormesis in critical studies of shorter

duration than traditional chronic bioassays. The

rather striking and consistent findings of these four

investigations and their predictive relationship to the

outcome of cancer bioassays in the same rodent strain

may provide a cost-effective means to assess the role of

hormesis with respect to the process of carcinogenesis.

LATE STAGE (TUMOR DEVELOPMENT AND

PROGRESSION)

The selection of studies for this section on the relation-ship

of hormesis to tumor development and progression

was affected by a number of factors. The control group

required a reasonably high tumor background incidence

and the experimental design necessitated an adequate

number of doses especially in the low-dose range. How-ever,

it was also considered desirable to include a range of

animal models, both sexes, range of affected organs, and

carcinogenic agents. In addition, it was hoped that the

information in the late-stage section could be linked with

independent findings in the previous stages, especially

those investigations involving tumor promotion. While

all goals were not fully achieved, considerable progress

was made in most of these areas, thereby permitting a

more robust assessment of the generalizability of the

hormetic hypothesis to the process of carcinogenesis.

1. The Effects of Saccharin on Bladder Tumor Development

The data on the effects of saccharin on the incidence of

urinary bladder hyperplasia suggested that cancer occur-rence

may be that of a U-shaped or hormetic dose–

response relationship (OTA, 1977). However, with re-spect

to the urinary bladder cancer end point, the control

rats used in the FDA study had such a low urinary

bladder tumor incidence (i.e., 4% in males and 0.0% in

females) that it precluded assessing the hormetic hypoth-esis

for this cancer end point.

The 1974 WARF (Wisconsin Alumni Research Foun-dation)

study was similarly designed as the 1973 FDA

sodium saccharin bioassay but the number of rats in

each treatment group was smaller and the number of

treatment groups was reduced from five to three (OTA,

1977). In this cancer bioassay, groups of 40 Sprague–

Dawley rats (20 male, 20 female) were employed with

sodium saccharin added to the diets at 0, 0.05, 0.5, or

5.0%. Rats that became moribund or died during the

study were necropsied and survivors were necropsied

at 100 weeks. The investigators reported that the over-all

tumor incidence was increased in the 5% males

compared to controls. However, at the lowest two treat-ment

groups, the females displayed notable decreases

in total tumor response (Table 2). The males displayed

a similar trend (Table 2), but since their background

total tumor incidence was low it was not possible to

explore adequately a hormetic hypothesis. In addition,

since the control urinary bladder tumor incidence for

males and females was 0.0% it was not possible to

relate the saccharin-induced decrease in urinary blad-der

hyperplasia at low dosages in the 1973 FDA study

in the same strain of rats to the urinary bladder tumor

end point in the WARF study.

TABLE 2

Total Number of Tumors for Male and Female Spra-gue–Dawley Rats in the 1974

WARF Study (Source: Downs and owski, 1982)

No.tumors (%)

Male Female

0 3/20 (15) 12/20 (60)

0.05 2/20 (10) 6/20 (30)

0.5 2/20 (10) 9/20 (45)

5.0 14/20 (70) 18/20 (90)

At least four one-generation saccharin-feeding ex-periments

have been conducted in mice (OTA, 1977).

Each of these studies is considered to be negative with

respect to saccharin-induced carcinogenicity. However,

one of the experiments may be relevant to the present

assessment since the spontaneous tumor incidence in

the concurrent control group was substantial [i.e., 38%

(19/50)]. In this investigation genetically homogeneous

mice of the dde strain were given saccharin for 21

months. Each group was composed of 100 mice (50

males, 50 females) and the saccharin was fed at 0, 0.2,

1.0, and 5.0% of the diet. The saccharin had no effect on

animal mortality and there were no treatment-related

effects on body weight. However, the incidence of total

tumors in the low-dose group of male mice in the Jap-anese

study displayed only an 8% (4/50) incidence com-pared

to the 38% incidence of the control (OTA, 1977).

The U-shaped dose response was consistent with that

reported for the rats in the WARF study. However, the

direct relevance of this study to the urinary bladder

hyperplasia hormetic response as seen in the FDA

study is unknown. These collective investigations with

sodium saccharin have been used to explore cancer risk

assessment models that incorporate the concept of tis-sue

repair (Downs and owski, 1982). Given the

inconsistency of the dose–response relationships with

the linear low-dose model paradigm, the approach of

Downs and owski (1982) enhances a more biolog-ically

plausible perspective to this analysis. However,

the intent of the present report is to explore not only

whether the above findings are inconsistent with the

linear low-dose model for cancer risk assessment, but

also whether these studies provide support for a

hormetic interpretation of the data. In the case of the

1973 FDA investigation the urinary bladder hyperpla-sia

findings are highly consistent with a hormetic in-terpretation.

This experiment is very strong given the

large sample size and the four dosages below the

LOAEL. Further supporting the hormetic hypothesis

for this end point are the data from the females, which

closely mimic the response of the male for the urinary

hyperplasia, though not so convincingly due to the

lower background response in the controls. Nonethe-less,

these two consistent responses clearly add collec-tive

weight to support the hormetic interpretation for

this end point. However, the low urinary bladder can-cer

background incidence data in the Sprague–Dawley

rats of both the FDA and WARF studies did not provide

an opportunity to relate the promotional findings in the

low-dose area to the urinary bladder cancer data.

2. Chemically Induced Pulmonary Tumors

One of the most common cancer bioassays employed

in a screening evaluation mode has been the pulmo-nary

tumor incidence in strain A and Swiss mice. Since

these mouse strains display very high spontaneous

pulmonary tumor incidence, they have the opportunity

to be utilized to assess the hormetic hypothesis for this

end point. To this end, a number of studies have been

found in which multiple doses have been employed that

would enable an initial evaluation of the hormetic hy-pothesis

(Nesnow et al., 1994; O’Gara et al., 1965; Pra-halad

et al., 1997). These investigations have typically

assessed the response to various carcinogenic hydro-carbons.

While these studies tend to vary to some ex-tent

with respect to study design, number of doses,

dose range, and statistical power, they are remarkably

consistent in displaying U-shaped responses. Tables 3

and 4, which highlight two of these studies (O’Gara et

al., 1965, and Prahalad et al., 1997, respectively),

clearly illustrate the high background pulmonary inci-dence

and the U-shaped nature of the dose–response

relationship.

TABLE 3

Pulmonary Tumors Present at 56–79 Weeks of Age in Female Mice Given a Single

Subcutaneous Injection of 3-Methylcholanthrene (MCA) as Newborns (Source: O’

Gara et al., 1965)

MCA Tumors/mice Tumors (%)

0 15/34 44.1

0.005 1/18 5.6

0.015 5/19 26.3

0.046 7/18 38.9

0.137 6/20 30.0

0.4 12/24 50.0

1.2 8/11 72.7

3.7 10/10 100

11.1 11/11 100

TABLE 4

Lung Tumors in Male Strain A/J Mice Treated with Dibenzo[a,l]pyrene (Source:

Prahalad et al., 1997)

Mg/kg Tumor % Tumors/mouse

0 50 0.67 6

0.3 43 0.42 6

1.5 97 4. 30 6 *

3 100 7.50*

6 100 16.10*

Note. Animals were administered dibenzo[a,l]pyrene ip in trica-prylin at 6–8

weeks of age and tumors counted at 8 months following injection.

* Statistically different from tricaprylin-treated control at P, 0.001 by

Mann–Whitney rank-sum test.

EPA has based its cancer risk assessments on the

dose-dependent increase in tumor responses at the

highest two doses. However, Cook (1994) has recently

emphasized that when total tumors are concerned and

normalized per 100 animals, striking U-shaped dose–

response relationships were evident for both males and

females (Table 5). The dioxin-treated rats displayed

substantial decreases in tumors of the adrenals and

pancreas and, more modestly, the liver. As for the

females, the decrease in tumor incidence was princi-pally

accounted for by the changes in tumor incidence

of the uterus, mammary glands, and pituitary. In fact,

even at the lowest dose the females displayed a modest

decrease in liver tumors, the critical target organ for

the EPA cancer risk assessment.

TABLE 5

Tumor Frequency in Male and Female Rats Exposed to 2,3,7,8-TCDD (Source:

Cook, 1994)

Male rats

Mg/kg/day 0 0.001 0.01 0.1

No. of rats 85 50 50 50

Tumors/rat

Total 162 80 98 120

Liver 9.4 0.0 6 6

Pulmonary 2.4 .0 0 4.0

Testes 2.4 4 0 0

Mammary .4 0 0 2

Pancreas 34.1 20 16 10

Female rats

Dose ( mg/kg/day) 0 0.001 0.01 0.1

No. of rats 86 50 50 49

No. Rate No. Rate No. Rate No. Rate

Total tumors 230 267.4 96 192.0 102 204.0 120 244.9

Liver 9 10.5 3 6.0 20 40.0 34 69.4

Pulmonary 0 0.0 0 0.0 1 2.0 7 14.3

Ovary 3 3.5 1 2.0 1 2.0 0 0.0

Uterus 36 41.9 14 28.0 14 28.0 11 22.4

Cervix/vagina 2 2.3 0 0.0 1 2.0 0 0.0

Mammary 81 94.2 39 78.0 40 80.0 24 49.0

Pituitary 49 57.0 18 36.0 14 28.0 14 28.6

Pancreas 5 5.8 4 8.0 1 2.0 1 2.0

Adrenal 16 18.6 8 16.0 3 6.0 8 16.3

Note. Rates are in number of tumors per 100 animals.

4. Testicular Cancer

One of the strongest cancer bioassays that provide

evidence consistent with the hormetic hypothesis was

reported by Waalkes et al. (1988) concerning the effects

of cadmium chloride on testicular cancer in Wistar

rats. This study involved the administration of six

doses of cadmium over a 40-fold dose range (1–40

mmol/kg). The control displayed a 17.8% (8/45) back-ground

incidence of testicular cancer. The treatments

demonstrated a striking U-shaped dose–response rela-tionship

with the lowest dose having only 1 of 30 rats

(3.3%) with testicular cancer. This study was particu-larly

impressive since it had four doses below the

LOAEL thereby making its possible to better explore

the nature of the dose–response relationship in the

lower dose region (Table 6).

TABLE 6

Incidence of Testicular Tumors in Rats Treated with

Cadmium Chloride (Source: Waalkes et al., 1988)

Dose

( mmol/kg) No. of rats

No. of tumors

(%)

0 45 8 (17.8)

1.0 30 1 (3.3)

2.5 29 3 (10.3)

5.0 30 3 (10.0)

10.0 30 4 (13.3)

20.0 29 21 (72.4)

40.0 29 24 (82.8)

5. Radiation and Mammary Cancer

A series of bioassays has been conducted by Broerse

et al. to assess the mammary cancer risks for low-dose

radiation exposures, such as those based on mammog-raphy,

in several rat strains (Broerse et al., 1978, 1982,

1987). The rats were irradiated at 8 weeks of age with

single doses of either X-rays or monoenergetic fast

neutrons. The animals were permitted to live out their

natural lifespans and were sacrificed when moribund.

A complete necropsy was performed and representa-tive

sections from all tissues were assessed histologi-cally

with particular attention given to the number and

types of mammary tumors as well as nonneoplastic

mammary gland lesions. Special exposure arrange-

ments were developed to permit bilateral or multilat-eral

irradiations with the goal of achieving uniform

dose distribution over the mammary gland tissue. In

addition, the animals were irradiated on a fixture ro-tated

coaxially with the ion beam axis. The procedure

resulted in proper tissue distribution of the neutron

exposure in the mammary tissue.

With these types of experimental settings the au-thors

conducted a series of cancer bioassays in several

rat strains using a variety of exposures to different

types of radiation, dose levels, and biological treatment

subgroups (e.g., with or without estrogen supplemen-tation).

Of relevance to the issue of possible hormetic

responses, all experimental settings involved three or

four dose levels with a three- to fourfold dose separa-tion

factor.

Suggested evidence of hormetic relationships was

reported in replicated studies with Sprague–Dawley

rats in response to 0.5 MeV of neutrons doses of 2, 8,

and 32 rads. Both experiments displayed a U-shaped

dose–response relationship for mammary tumor inci-dence

(Table 7). The control groups in both experi-

ments displayed high tumor background incidences of

29 and 30%, respectively. Likewise, in both experi-ments

the lowest treatment group displayed a 50–60%

reduction in mammary tumor incidence. At the higher

doses the tumor responses progressively increased, ex-ceeding

the controls. Similar findings were also re-ported

in the Wag/Rij rats as well (Table 7) (Broerse et

al., 1987).

TABLE 7

The Influence of Neutrons on Mammary Tumors in

Female Sprague–Dawley and Wag/Rij Rats (Broerse et

al., 1978, 1982, 1987)

Sprague–Dawley

Dose (rads)

Mammary tumor incidence (%)

Experiment 1 Experiment 2

0 2930

2 1215

8 5053

32 59 63

WAG/RIJ

Dose (rads)

Mammary tumor incidence (%)

Experiment 1 Experiment 2 Experiment 3

0 27 27 18.9

0.05 20 5.7

2 15 33 13.3

8 34 53 29.4

32 56

DISCUSSION

This paper addressed the issue of whether toxicolog-ical

investigations exist in the area of carcinogenesis

relevant to assessing the hormetic hypothesis. It is

important to note that no articles were obtained in

which the investigators explicitly set forth to evaluate

whether the dose–response relationship for various

end points of carcinogenesis including initiation, pro-motion,

and progression displayed a hormetic dose–

response relationship. The hormetic explanation/hy-pothesis

was invariably proposed only after the

investigations had reached the interpretational phase.

This fact serves to establish a lack of bias on behalf of

the investigators since it is obvious that they were not

interested in proving whether hormesis exits or not. In

fact, with but few exceptions, most of the references

cited in this chapter never mention the term hormesis

when interpreting their U-shaped dose–response rela-tionships.

They simply report their findings even when

the dose–response relationship is inconsistent with the

prior expectation of the linear and/or threshold (i.e.,

hockey stick) phenomenon. Most of the authors cited in

this chapter failed to even attempt to explain the U-shaped

data as well. Rather, they tended to exclusively

emphasize only the high-dose responses. In fact, this

was the case with every example observed with respect

to tumor end point. This was also the case for mutage-nicity.

Although the second paper published by Kitchin

and Brown (1996) addressed the issue of hormesis, this

occurred only after these authors were asked to become

directly involved with debating the hormesis hypothe-sis

with respect to their initial data (Kitchin and

Brown, 1995). However, more engaging discussions

and interpretations were evident with respect to pro-motional

end points such as alterations in cell-medi-ated

immunity, cell cycle changes, and other biomark-ers

associated with promotional changes.

The lack of explicit consideration of the hormetic

hypothesis in the conceptualization of the studies cited

in this chapter has important implications for subse-quent

evaluation of hormesis. There are extraordinary

study design demands necessary to assess whether

biologically relevant changes can occur below the tox-icological

NOAEL (Calabrese and Baldwin, 1997a,.

Not only does the NOAEL need to be established, but a

number of properly spaced doses below the NOAEL

need to be incorporated into the study design. Since

most investigators are concerned with toxic effects and

their underlying mechanisms, it becomes clear that

occasions where hormesis can be established by typical

cancer bioassay evaluation are the rare exception. Fur-ther

stacking the odds against finding a hormetic re-sponse

is that the control group needs to have a rather

high spontaneous tumor background. This creates a

situation which most investigators tend to avoid since

it places undesirable demands on the study to establish

treatment effects with adequate statistical power.

Thus, the typical ideal testing situation is where there

is a low background tumor incidence, just the opposite

of what would be necessary to evaluate the hormetic

hypothesis.

Despite these substantial obstacles, a number of ex-cellent

investigations have been published that in ret-rospect

provided substantial capacity to evaluate var-ious

aspects of the hormetic hypothesis. As expected,

such opportunities have presented themselves within

the context of investigations in which the background

end point incidence was high. This is clearly seen in the

case of pulmonary tumors in strain A and Swiss mice,

mammary tumors in Sprague–Dawley and Wag rats,

urinary bladder hyperplasia in Sprague–Dawley rats,

and testicular cancer in Wistar rats. That hormetic

responses occurred with such a wide range of cancer

end points argues that the phenomenon is highly gen-eralizable.

The fact that hormetic effects can be best assessed

when the control has a relatively high background

incidence should not be interpreted to mean that

hormesis cannot occur when the control incidence is

negligible or very low. If hormesis were to exist in such

a situation, it is likely that the dose–response relation-ship

would be manifest, not with a U-shaped dose–

response relationship but with the more traditional

threshold or hockey stick-shaped relationship. This

point has never been explicitly addressed within the

context of describing the hormetic dose–response rela-tionships.

However, if hormesis were seen within this

context it would considerably broaden the available

studies for evaluation and open up a consideration of

hormesis within the more traditional toxicological par-adigm

of threshold dose responses and their underly-ing

mechanisms.

The hormetic paradigm may not be inherently asso-ciated

with low-dose protective responses even though

these seem to be predominant in the toxicological lit-erature.

However, it is possible that low doses of agents

may also enhance cellular promotional processes in

particular instances. This has been suggested in sev-eral

recent reports on the mechanism of arsenic-in-duced

skin cancer (Germolec et al. 1996, 1997).

Further efforts will be required to develop general

understandings of how animal models and humans

respond to low-level exposures to genotoxic/carcino-genic

agents. The present assessment provides a num-ber

of examples of reliable studies where the dose–

response relationships are not only inconsistent with

the linear paradigm in the low-dose zone but are also

strongly supportive of the hormetic perspective. The

consistency of responses across model, chemical, and

end point is comparable to those reports for the spec-trum

of noncancerous responses as well. Such findings

should provide a strong incentive for further develop-ment

of this area of inquiry given its potential to en-hance

understanding of responses at realistic levels of

exposures.

REFERENCES

BELLE Newslett. (1995). Dose–response studies of genotoxic rodent

carcinogens: Thresholds, hockey sticks, hormesis or straight lines?

3, 1–19.

Broerse, J. J., Knaan, S., van Bekkum, D. W., Hollander, C. F.,

Noteboom, A. L., and van Zwieten, M. J. (1978). Mammary carci-nogenesis

in rats after X- and neutron irradiation and hormone

administration. Late Biol. Effects Ionizing Radiat. II, 13–27.

Broerse, J. J., Hennen, L. A., van Zwieten, M. J., and Hollander, C. F.

(1982). Mammary carcinogenesis in different rat strains after sin-gle

and fractionated irradiations. Commission of the European

Communities Report. Commission of the European Communities,

Directorate General Information Market and Innovation, Luxem-bourg,

pp. 155–168.

Broerse, J. J., Hennen, L. A., Klapwijk, W. M., and Solleveld, H. A.

(1987). Mammary carcinogenesis in different rat strains after ir-radiation

and hormone administration. Int. J. Radiat. Biol. 51,

1091–1100.

Calabrese, E. J., and Baldwin, L. A. (1997a). The dose determines the

stimulation (and poison): Development of a chemical hormesis

database. Int. J. Toxicol. 16, 545–559.

Calabrese, E. J., and Baldwin, L. A. (1997b). A quantitatively-based

methodology for the evaluation of chemical hormesis. Hum. Ecol.

Risk Assess. 3, 545–554.

Camurri, L., Codeluppi, S., Scarduelli, L., and Candela, S. (1983).

Sister chromatid exchange in workers exposed to low doses of

styrene. In Sister Chromatid Exchanges, Part B (R. R. Tice and A.

Hollaender, Eds.), pp. 957–963. Plenum Press, New York.

Chandra, R. K. (1991). 1990 McCollum award lecture: Nutrition and

immunity—Lessons from the past and new insights into the fu-ture.

Am. J. Clin. Nutr. 53, 1087–1101.

Conolly, R. B., and Andersen, M. E. (1997). Hepatic foci in rats after

diethylnitrosamine initiation and 2,3,7,8-tetrachlorodibenzo-p-di-oxin

promotion: Evaluation of a quantitative two-cell model and of

CYP 1A1/1A2 as a dosimeter. Toxicol. Appl. Pharmacol. 146, 281–

293.

Cook, R. R. (1994). Responses in humans to low level exposures. In

Biological Effects of Low Level Exposures: Dose–Response Relation-ships

(E. J. Calabrese, Ed.), pp. 99–109. Publishers, Boca

Raton, FL.

Downs, T. D., and owski, R. F. (1982). Influence of repair

processes on dose–response models. Drug Metab. Rev. 13, 839–

852.

Fan, F., Wierda, D., and R

ozman, K. K. (1996). Effects of 2,3,7,8-

tetrachlorodibenzo-p-dioxin on humoral and cell-mediated immu-nity

in Sprague–Dawley rats. Toxicology 106, 221–228.

Germolec, D. R., Spalding, J., Boorman, G. A., Wilmer, J. L., Yo-shida,

T., Simeonova, P. P., Bruccoleri, A., Kayama, F., Gaido, K.,

Tennant, R., Burleson, F., Dong, W., Lange, R. W., and Luster,

M. I. (1997). Arsenic can mediate skin neoplasia by chronic stim-ulation

of keratinocyte-derived growth factors. Mutat. Res. 386,

209–218.

Germolec, D. R., Yoshida, T., Gaido, K., Wilmer, J. L., Simeonova,

P. P., Kayama, F., Tennant, R., Burleson, F., Dong, W., Lange,

R. W., and Luster, M. I. (1996). Arsenic induces overexpression of

growth factors in human keratinocytes. Toxicol. Appl. Pharmacol.

141, 308–318.

Goldsworthy, T., , H. A., and Pitot, H. C. (1984). The nat-ural

history and dose–response characteristics of enzyme-altered

foci in rat liver following phenobarbital and diethylnitrosamine

administration. Carcinogenesis 5, 67–71.

Hagiwara, H., Hirose, M., Takahashi, S., Ogawa, K., Shirai, T., and

Ito, N. (1991). Forestomach and kidney carcinogenicity of caffeic

acid in F344 rats and C57BL/6N 3 C3H/HeN F1 mice. Cancer Res.

51, 5655–5560.

Hanham, A. F., Dunn, B. P., and Stich, H. F. (1983). Clastogenic

activity of caffeic acid and its relationship to hydrogen peroxide

generated during autooxidation. Mutat. Res. 116, 333–339.

Hirose, M., Kawabe, M., Shibata, M., Takahashi, S., Okazaki, S., and

Ito, N. (1992). Influence of caffeic acid and other o-dihydroxyben-zene

derivatives on N-methyl-N9-nitro-N-nitrosoguanidine-initi-ated

rat forestomach carcinogenesis. Carcinogenesis 13,

1825–1828.

Hirose, M., Masuda, A., Fukushima, S., and Ito, N. (1988). Effects of

subsequent antioxidant treatment on 7,12-dimethylbenz(a)an-thracene-

initiated carcinogenesis of the mammary gland, ear duct

and forestomach in Sprague–Dawley rats. Carcinogenesis 9, 101–104.

Hirose, M., Mutai, M., Takahashi, S., Yamada, M., Fukushima, S.,

and Ito, N. (1991). Effects of phenolic antioxidants in low dose

combination on forestomach carcinogenesis in rats pretreated with

N-methyl-N9-nitro-N-nitrosoguanidine. Cancer Res. 51, 824–827.

IARC (1993). Some naturally occurring substances: Food items and

constituents, heterocyclic aromatic amines and mycotoxins. IARC

Monogr. 56, 114–134.

Kitchin, K. T., and Brown, J. L. (1994). Dose–response relationship

for rat liver DNA damage caused by 49 rodent carcinogens. Toxi-cology

88, 31–49.

Kitchin, K. T., and Brown, J. L. (1995). Response to: Dose–response

studies of genotoxic rodent carcinogens: Thresholds, hockey sticks,

hormesis or straight lines? BELLE Newslett. 3, 16–19.

Kitchin, K. T., and Brown, J. L. (1996). Dose–response relationship

for rat liver DNA damage caused by 1,2-dimethylhydrazine. Tox-icology

114, 113–124.

Kleczkowska, H. E., and Althaus, F. R. (1996). Response of human

keratinocytes to extremely low concentrations of N-methyl-N9-nitro-

N-nitrosoguanidine. Mutat. Res. 367, 151–159.

Kociba, R. J., Keyes, D. G., Beyer, J. E., Carreon, R. M., Wade, C. E.,

Ditenber, D. A., Kalnins, R. P., Frauson, L. E., Park, C. N., Bar-nard,

S. D., Hummel, R. A., and Humiston, C. G. (1978). Results of

a 2-year chronic toxicity and oncogenicity study of

2,3,7,8-tetra-chlorodibenzo-

p-dioxin in rats. Toxicol. Appl. Pharmacol. 46, 279–303.

Liu, Y., Egyhazi, S., Hansson, J., Bhide, S. V., Kulkarni, P. S., and

Grafstrom, R. C. (1997). O 6 -Methylguanine–DNA methyltrans-ferase

activity in human buccal mucosal tissue and cell cultures.

Complex mixtures related to habitual use of tobacco and betel quid

inhibit the activity in vitro. Carcinogenesis 18, 1889–1895.

Lutz, U., Lugli, S., Bitsch, A., Schlatter, J., and Lutz, W. K. (1997).

Dose response for the stimulation of cell division by caffeic acid in

forestomach and kidney of the male F344 rat. Fundam. Appl.

Toxicol. 39, 131–137.

Nesnow, S., Ross, J., , G., , K., Roop, B., Jeffers, A.,

Galati, A., Stoner, G., Sangaiah, R., and Gold, A. (1994). Cyclopen-ta[

c,d]pyrene-induced tumorigenicity, Ki-ras codon 12 mutations

and DNA adducts in strain A/J mouse lung. Carcinogenesis 15,

601–606.

O’Gara, R. W., , M. G., Brown, J., and Mantel, N. (1965).

Induction of tumors in mice given a minute single dose of dibenz-[

a,h]anthracene or 3-methylcholanthrene as newborns. A dose–

response study. J. Natl. Cancer Inst. 35, 1027–1042.

Office of Technology Assessment (OTA) (1977). Cancer Testing Tech-nology

and Saccharin. U.S. Government Printing Office, Washing-ton,

DC.

Pitot, H. C., Goldsworthy, T. L., Moran, S., Kennan, W., Glauert,

H. P., Maronpot, R. R., and , H. A. (1987). A method to

quantitate the relative initiating and promoting potencies of

hepa-tocarcinogenic

agents in their dose–response relationships to al-tered

hepatic foci. Carcinogenesis 8, 1491–1499.

Prahalad, A. K., Ross, J. A., , G. B., Roop, B. C., King, L. C.,

Nesnow, S., and Mass, M. J. (1997). Dibenzo[a,l]pyrene-induced

DNA adduction, tumorigenicity, and Ki-ras oncogene mutations in

strain A/J mouse lung. Carcinogenesis 18, 1955–1963.

Stebbing, A. R. D. (1997). A theory for growth hormesis. BELLE

Newslett. 6, 1–11.

Tanaka, T., Kojima, T., Kawamori, T., Wang, A., Suzui, M., Oka-moto,

K., and Mori, H. (1993). Inhibition of 4-nitroquinoline-1-

oxide-induced rat and tongue carcinogenesis by the naturally oc-curring

plant phenolics caffeic, ellagic, chlorogenic and ferulic

acids. Carcinogenesis 14, 1321–1325.

Teeguarden, J. G., Dragan, Y. P., Singh, J., Vaughan, J., Goldswor-thy,

T. L., and Pitot, H. C. (1996). Time and dose dependence of

altered hepatic foci by 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol-ogist

30(Part 2), 163.

Waalkes, M. P., Rham, S., Riggs, C. W., Bare, R. M., Devor, D. E.,

Poirier, L. A., Wenk, M. L., Henneman, J. R., and Balaschak, M. S.

(1988). Cadmium carcinogenesis in male Wistar [Crl:(WI)BR] rats:

Dose–response analysis of tumor induction in the prostate and

testes and at the injection site. Cancer Res. 48, 4656–4663.

Wattenberg, L. W., Coccia, J. B., and Lam, L. K. T. (1980). Inhibitory

effects of phenolic compounds on benzo(a)pyrene-induced neopla-sia.

Cancer Res. 40, 2820–2823.

Alan Pater, Ph.D.; Faculty of Medicine; Memorial University; St. 's, NF

A1B 3V6 Canada; Tel. No.: (709) 777-6488; Fax No.: (709) 777-7010; email:

apater@...