A Cancer Risk Assessment of Di (2-Ethylhexyl) Phthalate: Application of The New U.S. EPA Risk Assessment Guidelines

Regulatory Toxicology and Pharmacology 29, 327357 (1999)
Article ID rtph.1999.1296, available online at http://www.idealibrary.com on
A Cancer Risk Assessment of Di(2-ethylhexyl)phthalate: Application

of the New U.S. EPA Risk Assessment Guidelines
John Doull,* Russell Cattley, Cliff Elcombe, Brian G. Lake, James Swenberg,
Christopher Wilkinson,\ Gary Williams,** and Marcia van Gemert ,1
*University of Kansas Medical Center, Kansas City, Kansas; Chemical Industry Institute of Toxicology, Research Triangle Park,
North Carolina; Biochemical Research Centre, University of Dundee, Dundee, United Kingdom; BIBRA International;
University of North Carolina, Chapel Hill, North Carolina; \Jellinek, Schwartz & Connolly, Inc., Arlington, Virginia;
**American Health Foundation, Valhalla, New York; and van Gemert & Hauswirth,
L.L.C., Charlotte Hall, Maryland 20622
Received October 20, 1998
The current United States Environmental Protection Agency (EPA) classification of di(2-ethylhexyl)phthalate (DEHP) as a B2 probable human carcinogen is based on outdated information. New toxicology
data and a considerable amount of new mechanistic
evidence were used to reconsider the cancer classification of DEHP under EPAs proposed new cancer risk
assessment guidelines. The total weight-of-evidence
clearly indicates that DEHP is not genotoxic. In vivo
administration of DEHP to rats and mice results in
peroxisome proliferation in the liver, and there is
strong evidence and scientific consensus that, in rodents, peroxisome proliferation is directly associated
with the onset of liver cancer. Peroxisome proliferation is a transcription-mediated process that involves
activation by the peroxisome proliferator of a nuclear
receptor in rodent liver called the peroxisome proliferator-activated receptor (PPARa). The critical role of
PPARa in peroxisomal proliferation and carcinogenicity in mice is clearly established by the lack of either
response in mice genetically modified to remove the
PPARa. Several mechanisms have been proposed to
explain how, in rodents, peroxisome proliferation can
lead to the formation of hepatocellular tumors. The
general consensus of scientific opinion is that PPARainduced mitogenesis and cell proliferation are probably the major mechanisms responsible for peroxisome
proliferator-induced hepatocarcinogenesis in rodents.
Oxidative stress appears to play a significant role in
this increased cell proliferation. It triggers the release
of TNFa by Kupffer cells, which in turn acts as a potent mitogen in hepatocytes. Rats and mice are
uniquely responsive to the morphological, biochemical, and chronic carcinogenic effects of peroxisome
proliferators, while guinea pigs, dogs, nonhuman primates, and humans are essentially nonresponsive or
refractory; Syrian hamsters exhibit intermediate re1
To whom correspondence should be addressed at 12521 Old

Homeplace Dr., Charlotte Hall, MD 20622.
sponsiveness. These differences are explained, in part,

by marked interspecies variations in the expression of
PPARa, with levels of expression in humans being
only 110% of the levels found in rat and mouse liver.
Recent studies of DEHP clearly indicate a nonlinear
doseresponse curve that strongly suggests the existence of a dose threshold below which tumors in rodents are not induced. Thus, the hepatocarcinogenic
effects of DEHP in rodents result directly from the
receptor-mediated, threshold-based mechanism of
peroxisome proliferation, a well-understood process
associated uniquely with rodents. Since humans are
quite refractory to peroxisomal proliferation, even following exposure to potent proliferators such as hypolipidemic drugs, it is concluded that the hepatocarcinogenic response of rodents to DEHP is not relevant to
human cancer risk at any anticipated exposure level.
DEHP should be classified an unlikely human carcinogen with a margin of exposure (MOE) approach to
risk assessment. The most appropriate and conservative point of reference for assessing MOEs should be 20
mg/kg/day, which is the mouse NOEL for peroxisome
proliferation and increased liver weight. Exposure of
the general human population to DEHP is approximately 30 mg/kg body wt/day, the major source being
from residues in food. Higher exposures occur occupationally [up to about 700 mg/kg body wt/day (mainly by
inhalation) based on current workplace standards]
and through use of certain medical devices [e.g., up to
457 mg/kg body wt/day for hemodialysis patients (intravenous)], although these have little relevance because the routes of exposure bypass critical activation
enzymes in the gastrointestinal tract. 1999 Academic Press
INTRODUCTION
The United States Environmental Protection Agency

(EPA) characterizes the cancer hazard of di(2-ethylhexyl)phthalate (DEHP) as a B2 carcinogen (probable
327
0273-2300/99 $30.00
Copyright 1999 by Academic Press
All rights of reproduction in any form reserved.
328
DOULL ET AL.
human carcinogen), based solely on studies conducted

by the National Toxicology Program (NTP) in the early
1980s using F344 rats and B6C3F1 mice where liver
cancer was identified in both species (NTP, 1982a). In
addition, EPA has not assumed a threshold for carcinogenic effects and has used the linear multistage model
to characterize the risk to humans from these NTP
studies. Prior to the 1982 NTP studies, three long-term
feeding studies showed no carcinogenic effects.
Since the NTP studies were completed, considerable
information has been developed on the comparative
pharmacokinetics, metabolism, and hepatic effects of
DEHP in various species. In addition, DEHP has been
shown in numerous recent studies to be non-genotoxic.
Two new chronic dietary studies administering DEHP
to F344 rats and B6C3F1 mice have shown DEHP to
cause an increased incidence of hepatic adenomas and
carcinomas in both species. These two new studies
suggest a nonlinear doseresponse curve.
In light of the new draft Cancer Risk Assessment
Guidelines, the newly developed carcinogenicity studies, and new mechanistic information on both DEHP
and other peroxisome proliferators, we believe that
there is now an opportunity to reexamine DEHPs classification as a potential carcinogen under the old guidelines and to use the threshold-based margin of exposure (MOE) approach to extrapolate carcinogenic risk
from animal studies to humans.
SCIENTIFIC RATIONALE FOR RECLASSIFICATION
DEHP is one of many substances that increase the

numbers and size of peroxisomes in the liver of rodents
(Lake et al., 1975; Lundgren et al., 1987; Moody and
Reddy, 1978; Osumi and Hashimoto, 1978). There is
increasingly strong scientific evidence that peroxisome
proliferation is of central importance in the hepatocarcinogenesis of DEHP. Although a complete understanding of the precise mechanism of rodent hepatocarcinogenesis of DEHP is not yet at hand, the carcinogenic
potential of other peroxisome proliferators has been
investigated in primates and humans, and the currently available data indicate that DEHP is unlikely to
pose a cancer hazard to humans. Primarily, the effects
of peroxisome proliferation concern lipid metabolism,
to which rat and mouse livers appear to be particularly
sensitive. Peroxisomes are single membrane-limited
cytoplasmic organelles that can increase in number
and size and are present in cells of animals, plants,
fungi, and protozoa. They are characterized by their
content of catalase and a number of fatty acid-oxidizing
enzymes, one of which, acyl-CoA oxidase, generates
hydrogen peroxide (Lazarow and deDuve, 1976; Reddy
and Lalwani, 1983). Peroxisomes have a number of
important functions in intermediary metabolism (Mannaerts and Van Veldhoven, 1993; Reddy and Mannaerts, 1994).
The characteristics of peroxisome proliferation in rats

and mice have been extensively reviewed (Reddy and
Lalwani, 1983; Bentley et al., 1993; Ashby et al., 1994;
Lake, 1995a,b; Lake and Lewis, 1996). From a quantitative point of view, the intensity of all peroxisome
proliferator effects is highly dependent on the individual compound and, of course, on the dose and duration
of treatment (Huber et al., 1996). Liver enlargement is
due to both hepatocyte hyperplasia and hypertrophy.
Morphological examination reveals an increase both in
the number and size of peroxisomes and in the amount
of smooth endoplasmic reticulum. The major biochemical alterations consist of induction of both peroxisomal
and microsomal fatty acid-oxidizing enzyme activities.
The activity of the peroxisomal fatty acid v-oxidation
cycle is normally determined either by measuring the
overall activity [e.g., as cyanide-insensitive palmitoylCoA oxidation (POX)] or by assaying the first ratelimiting enzyme of the cycle, namely acyl-CoA oxidase
[e.g., as palmitoyl-CoA oxidase (ACO)]. Importantly,
there is a differential induction of peroxisomal enzymes,
in that while the v-oxidation cycle enzymes are markedly increased, much smaller changes are observed in
the activities of other enzymes such as catalase and
D-amino acid oxidase. The stimulation of microsomal
fatty acid-oxidizing enzymes, normally measured as
lauric acid 12-hydroxylase (i.e., b-hydroxylation), is due
to induction of cytochrome P450 isoenzymes in the
CYP4A subfamily. Good correlations have been reported
between the induction of palmitoyl-CoA oxidation and
either lauric acid 12-hydroxylase or changes in peroxisome morphometry (Lake, 1995a; Lake and Lewis,
1996). The good correlations observed between enzyme
activity and changes in organelle morphometry permit
palmitoyl-CoA oxidation to be employed as a specific
biochemical marker of peroxisome proliferation. Carnitine acetyltransferase (CAT) activity can be markedly
stimulated by rodent peroxisome proliferators; this enzyme has also been employed as a marker for hepatic
peroxisome proliferation. It should be noted that CAT
is located in peroxisomal, mitochondrial, and microsomal fractions. Enzyme induction may thus be due to
effects in more than one subcellular compartment and
hence CAT is not considered to be a specific marker of
peroxisome proliferation.
The intimate linkage of effects of peroxisome proliferators on hepatic peroxisomes and hepatocyte growth
is attributable to the activation of PPARa, a nuclear
receptor (Green, 1995). This receptor binds selectively
to DNA to activate transcription of genes for several
peroxisomal proteins. In a recently developed knockout
mouse model that lacks functional PPARa, adaptive
effects elicited by peroxisome proliferators in the liver
were abolished. These adaptive effects that were eliminated include the increased peroxisomal volume density, peroxisomal enzyme induction, and hepatomegaly
(Lee et al., 1995), as well as the hepatocyte prolifera-
329
DEHP RISK ASSESSMENT
tion (Peters et al., 1997) observed in normal mice. The

relevance of these findings indicated the need for carcinogenicity studies of peroxisome proliferators in PPARa
knockout mice. A PPARa knockout mouse study, to be
covered in more detail in a later section on mechanisms
of carcinogenicity, reveals that expression of functional
PPARa is necessary for carcinogenesis by peroxisome
proliferators (Peters et al., 1997).
The genotoxicity of DEHP and mono-(2-ethylhexyl)phthalate (MEHP) (the primary metabolite of DEHP)
will not be discussed in detail in this paper. The total
weight-of-evidence for the mutagenicity studies on DEHP
clearly indicates that it is not genotoxic as previously
reviewed by Budroe and Williams (1993). This conclusion is based on negative findings in a wide range of
both in vivo and in vitro studies which are summarized
here.
Following the cessation of dosing and metabolic clearance, parameters associated with peroxisome proliferation return relatively rapidly to control levels. This
reversibility of the process indicates that the continued
presence of the peroxisome proliferator is an absolute
requirement to maintain the increased numbers of peroxisomes and increased peroxisomal enzyme activities.
COMPARISON OF DEHP WITH OTHER
PEROXISOME PROLIFERATORS
A wide variety of chemicals have been shown to

produce liver enlargement and enhanced peroxisome
proliferation in the rat and mouse (Reddy and Lalwani,
1983; Bentley et al., 1993; Lake and Lewis, 1993; Ashby
et al., 1994; Lake, 1995a). Classes of chemicals that
produce this effect include hypolipidemic drugs and
other therapeutic agents, herbicides, plasticizers, industrial solvents, natural products, and food flavors.
With respect to plasticizers, apart from DEHP, peroxisome proliferation in rodent liver has been observed
following treatment with a number of adipate, phthalate, and trimellitate esters (Bentley et al., 1993; Lake
and Lewis, 1993; Ashby et al., 1994). Although the vast
TABLE 1
Comparison of DEHP with Other Peroxisome Proliferators: Relative Strength of POX a and CAT b Induction
a
b
Enzyme induction
Peroxisome proliferators
Very weak
Weak
Intermediate
Strong
Very strong
Acetylsalicyclic acid
DEHP
Clofibrate
Nafenopin
WY-14,643
Ciprofibrate
Palmitoyl-CoA oxidation.
Carnitine acetyltransferase.
TABLE 2
Comparison of DEHP with Other Peroxisome
Proliferators: Induction of Relative Liver Weight
Compound
Dose
(mg/kg body wt/day)
Relative liver weight

(ratio to control)
DEHA
DEHP
Clofibrate
Nafenopin
WY-14,643
Ciprofibrate
1250
600
250
50
50
12.5
1.36
1.52
1.79
1.97
1.94
1.95
Note. Values from a single study by Takagi et al. (1992) with doses
readjusted to mg/kg body wt/day (Huber et al., 1996).
majority of rodent peroxisome proliferators are nongenotoxic agents, a number of these chemicals have
been shown to increase the incidence of liver tumors in
the rat and/or mouse (Reddy and Lalwani, 1983; Bentley et al., 1993; Ashby et al., 1994).
Because the hepatic morphological and biochemical
effects of DEHP in rats and mice are essentially identical to those produced by a wide range of other rodent
peroxisome proliferators, it is appropriate to utilize all
information on rodent peroxisome proliferators per se
for assessing the risk of DEHP to humans. DEHPinduced hepatic peroxisome proliferation is not due to
the parent diester but to DEHP metabolites. The primary metabolites of DEHP are MEHP and 2-ethylhexanol (2-EH). With MEHP, additional studies have
demonstrated that the proximate peroxisome proliferators are the (v-1)-hydroxy and keto metabolites,
whereas with 2-EH the proximate peroxisome proliferator is EHA (Elcombe and Mitchell, 1986; Lake and
Lewis, 1993; Lhuguenot and Cornu, 1993; Mitchell et
al., 1985b).
Rodent peroxisome proliferators exhibit marked potency differences (Bentley et al., 1993; Lake and Lewis,
1993; Ashby et al., 1994; Lake, 1995a; Huber et al.,
1996). For example, the hypolipidemic agent ciprofibrate is several orders of magnitude more potent than
DEHP, which is somewhat more potent than di-(2ethylhexyl)adipate (DEHA) (Reddy et al., 1986). Calculations from literature data suggest that DEHP and
DEHA are 15 and 5 times more potent, respectively,
than acetylsalicylic acid (Barber et al., 1987). Thus,
based on potency to induce enzyme activities, such as
the peroxisomal fatty acid b-oxidation cycle (e.g., palmitoyl-CoA oxidation) and CAT, DEHP may be considered a relatively weak proliferator (Table 1). The more
potent peroxisome proliferators, such as clofibrate and
nafenopin, have been reported to exhibit lower steeper
doseresponse curves than DEHP (Huber et al., 1996).
In addition, the more potent agents also produce greater
increases in relative liver weight at lower dose levels
(Table 2).
All peroxisome proliferators produce an initial in-
330
DOULL ET AL.
TABLE 3
Comparison of the Hepatocarcinogenic Potential of DEHP
with That of Other Peroxisome Proliferators in Rats
Chemical
% in diet
mg/kg
body wt /day
Duration of study
(months)
% Animals
with tumors
24
010
Control
DEHA
DEHP LOEL
DEHP medium dose
DEHP highest dose
Clofibrate
2.5
0.6
1.2
2.0
0.5
1250
300
600
1000
250
24
24
24
24
1630
4
12
24
6079
4091
Nafenopin
Nafenopin
WY-14,643
Ciprofibrate
0.1
0.2
0.1
0.02
50
100
50
10
1825
13
1416
14
73
65
100
100
Reference(s)
Goodman et al. (1979), NTP (1982a),
Rao et al. (1990)
NTP (1982b)
NTP (1982a)
NTP (1982a)
Rao et al. (1987, 1990)
Svoboda and Azarnoff (1979), Reddy and
Qureshi (1979)
Reddy and Reddy (1977)
Kraupp-Grasl et al. (1990)
Reddy et al. (1979), Rao et al. (1984b)
Rao et al. (1984a)
Note. All doses were calculated from Huber et al. (1996).
crease in cell replication in rodent hepatocytes during

the first few days of treatment and with continued
administration, the more potent agents, depending on
the dose administered, may produce a sustained stimulation of replicative DNA synthesis greater than that
of the weaker peroxisome proliferators (Marsman et
al., 1988; Lake, 1995a). Rodent peroxisome proliferators demonstrate a clear tissue specificity and at least
for some compounds structureactivity relationships
(many are carboxylic acids or esters) have been observed (Small et al., 1982; Reddy and Lalwani, 1983;
Lake and Lewis, 1993; Huber et al., 1996). Many studies have demonstrated clear dose thresholds for peroxisome proliferation in rodent liver (Lake et al., 1984,
1989; Tomaszewski et al., 1987; Budroe et al., 1992;
Huber et al., 1996).
A number of rodent peroxisome proliferators have
been shown to increase the incidence of liver tumors in
rats and/or mice. From an examination of the literature, Ashby et al. (1994) reported a strong concordance
between peroxisome proliferation and hepatocarcinogenesis in rats and mice. On the basis of a more limited
data base, a similar concordance was observed between
hepatocellular proliferation and liver tumor formation.
Peroxisome proliferators such as ciprofibrate,
WY-14,643, clofibrate, BR-931, gemfibrozil, methylclofenapate, nafenopin, tibric acid, benzbromarone, and
di- and trichloroacetic acids differ in their carcinogenic
potency (Stott, 1988; Marsman et al., 1988; Parzefall et
al., 1990; Kraupp-Grasl et al., 1991; Huber et al., 1996).
Table 3 gives a summary of the relative hepatocarcinogenicity of the most typical representatives of peroxisome proliferators in rats, and it also contains data on
three doses of DEHP. As already recognized for the
induction of indicators of peroxisome proliferation, ciprofibrate and WY-14,643 are among the most carcinogenic peroxisome proliferators, whereas DEHP is a less
efficient carcinogen. Ciprofibrate and WY-14,643 dis-
play a tumor incidence of 100% at doses at or below

0.1% in the diet for about 1 year, a result that was not
even approximated by 2% DEHP for 2 years, that is, 20
times the dose for twice the duration.
PHARMACOKINETICS OF DEHP
Humans are exposed to DEHP via oral, inhalation,

dermal, and intravenous routes (Huber et al., 1996).
The highest potential exposures, mainly via inhalation, are occupational, although high exposures can
also occur from certain intravenous medical procedures
such as hemodialysis. Exposures of the general population to DEHP occur mainly through ingestion of residues in food with much lower exposures from air and
water.
In most species, the first and most important step in
the metabolic activation of DEHP is the hydrolytic
cleavage of one of the two ethylhexyl moieties to yield
MEHP and 2-EH. Hydrolytic cleavage of both ethylhexyl groups of DEHP to yield free phthalic acid occurs
only to a very small extent (,3% in the urine of rats,
monkeys, and humans and about 12% in mice). Since
pancreatic lipase is the most effective enzyme hydrolyzing DEHP, MEHP formation occurs mainly in the
gut following oral ingestion. Only low titers of DEHPhydrolyzing enzymes are present in the liver and other
tissues (Albro, 1986; Huber et al., 1996). As a result,
the largest doses of MEHP result from the oral ingestion of DEHP since other routes of exposure (inhalation, iv, dermal) avoid contact with the lipase present
in the gut. This has important toxicokinetic implications because MEHP not only constitutes an active
peroxisome proliferator but, relative to DEHP, it also is
more readily absorbed into the systemic circulation.
There are also important species differences in the rate
of MEHP formation in the intestine, hydrolytic activity
being highest in the mouse, followed by the rat, guinea
331
pig, and hamster (Lake et al., 1984; Albro, 1989). In the

marmoset, a primate, DEHP hydrolysis is considerably
lower than in the rat (Rhodes et al., 1986) and humans
appear to metabolize DEHP similarly to primates (Albro et al., 1982). Following iv administration, the rate
of hydrolysis of DEHP to MEHP is lower in humans
than in rodents so that circulating levels of MEHP and
other active metabolites are very low in humans undergoing transfusion or dialysis procedures (Rock et
al., 1978; Chu et al., 1978).
In the liver, the ethylhexyl chain of MEHP (not of
DEHP) is subject to extensive oxidative metabolism by
cytochrome P450-mediated enzymes (Albro et al., 1973,
1982; Albro, 1986; Astill, 1989; Huber et al., 1996).
Reactions occurring include:
1. Hydroxylation of the terminal carbon atoms (voxidation) of both the hexyl and ethyl side chains and
the penultimate carbon (v-1 oxidation) of the hexyl
chain;
2. Conversion of these hydroxy groups to either a
carboxylic acid (v-oxidation) or a ketone (v-1 oxidation); and
3. Further metabolism (shortening of the carbon
chains) of the dicarboxylic acid products of v-oxidation
by a- or b-oxidation.
The major metabolites responsible for peroxisome
proliferation activity are MEHP and the metabolites IX
and VI, respectively, the hydroxy and keto metabolites
resulting from v-1 oxidation.
Following oral exposure of humans to DEHP, MEHP
constitutes approximately 25% of the total urinary metabolites. Metabolite IX resulting from v-1 oxidation of
MEHP represents a further 37% of the human urinary
metabolites. In contrast, in rat urine, MEHP is present
only in trace amounts; about 75% of all urinary metabolites consist of dicarboxylic metabolites, mainly metabolite V. No parent DEHP is present in urine regardless of the route of exposure (Huber et al., 1996). The
appearance of unchanged DEHP in feces following oral
ingestion can be attributed to incomplete hydrolysis
and absorption.
Different metabolites are also subject to varying degrees of conjugation with glucuronic acid (Peck et al.,
1979). The total rate of glucuronidation is strongly
species dependent. Primates glucuronidate the oxidative metabolites of MEHP more completely, while oxidizing metabolites less effectively, and rats, unlike other
species, do not glucuronidate DEHP metabolites (Albro
et al., 1979; Rhodes et al., 1986).
DEHP has an elimination half-life of about 10 to 18 h
(Sjoberg et al., 1985; Pollack et al., 1985a). The elimination of DEHP depends largely on its metabolism
(Tanaka et al., 1975; Huber et al., 1996). According to
Short et al. (1987), rats eliminated 87% of an oral dose
of 100 mg/kg body wt radiolabeled DEHP after 4 days,

while in monkeys the percentage excreted was 79%.
Evaluation
Species differences have been reported in the absorption, distribution, metabolism, and excretion of DEHP
(Huber et al., 1996). However, it is clear that species
differences in DEHP metabolism (esterase activity, metabolite oxidation and conjugation) cannot explain species differences in DEHP-induced hepatic peroxisome
proliferation. Rather, additional factors are involved in
determining species susceptibility to DEHP and other
rodent peroxisome proliferators.
CARCINOGENICITY OF DEHP
A large number of chronic/carcinogenicity studies

have been conducted with DEHP in rodents and other
species. Some of the earlier studies used doses that
were too low to reveal any evidence of tumorigenic
potential. In addition, some of the other studies had
experimental design flaws. For example, in the mouse
study by Tobe et al. (1984) all the animals fed 2500
mg/kg/day DEHP in their diet died before 38 weeks of
treatment. Of the remaining rodent carcinogenicity
studies with DEHP, five demonstrated a significant
increase in liver tumors. Two of these studies were
conducted by the NTP (1982) in male and female rats
and mice given dietary concentrations of 0.6 and 1.2%
(rat) and 0.3 and 0.6% (mouse). These studies, using
higher doses than had been used previously, revealed a
DEHP-related induction of hepatic tumors in all groups
except in the male rats at the 0.6% dose level (300
mg/kg/day).
The NTP results were essentially replicated and extended by Cattley et al. (1987) who dosed rats at 0.03,
0.1, or 1.2% DEHP in conventional diets. At the high
dose of 1.2% (600 mg/kg body wt/day) there was a 30%
increase of liver tumors compared with 0% in the controls. Rao et al. (1987, 1998) found a clear increase in
liver tumor incidence in rats fed diets containing 2%
DEHP (1000 mg/kg body wt/day).
The most recent rat and mouse carcinogenicity studies conducted by Covance provide further doseresponse
information. In the rat study (David et al., 1996), DEHP
was administered to F344 rats for 104 weeks at dietary
concentrations of 0, 100, 500, 2500, and 12,500 ppm (0,
5.8, 28.9, 146.6, and 789.0 mg/kg/day, respectively, for
males, and 0, 7.3, 36.1, 181.7, and 938.5 mg/kg/day,
respectively, for females). In addition, a sixth group
(the recovery period group) was administered a diet
containing 12,500 ppm DEHP for 78 weeks, followed by
a 26-week recovery period during which a basal diet
was provided. All relevant parameters for a chronic/
carcinogenicity study were evaluated. In addition, at
weeks 1, 2, 13, 7, and 9 and at study termination
332
DOULL ET AL.
analyses were conducted to evaluate DEHP-induced

cell proliferation and peroxisome proliferation in the
livers of dosed animals.
In the rat study, a dietary concentration of 500 ppm
(group 3) was the no-observed-effect level (NOEL) and
the low-observed-effect level (LOEL) was 2500 ppm
(group 4) in males and females for systemic effects. No
increase in serum levels of liver enzymes was found at
any concentration. In fact, serum GGT was reduced,
which corresponds to the report of Maruyama et al.
(1990) that DEHP suppresses liver GGT activity.
There was no evidence of carcinogenic potential at
500 ppm, whereas the concentrations of 2500 ppm
(group 4) and 12,500 ppm (group 5) induced gross and
microscopic evidence of liver effects in males and females, with an increased incidence of liver neoplasms
in males and an increased incidence of mononuclear
cell leukemia in males. In group 6, administration of a
dietary concentration of 12,500 ppm for 78 weeks, followed by a 26-week recovery period, indicated that
effects of DEHP in the liver, including the induction of
neoplasms, were reversible. However, DEHP effects in
the kidney, testis, and pituitary were not reversible,
and the incidence of mononuclear cell leukemia in
group 6 males was not decreased compared to group 5
males. The relationship of mononuclear cell leukemia
to DEHP is questionable due to the unusually low
incidence in the male controls compared to a recent
compilation of NTP data (NTP, 1996) and the lack of
reproducibility of this effect in the NTP bioassay of
DEHP in F344 rats.
In the most recent mouse study (David et al., 1997),
DEHP was administered to B6C3F1 mice in the diet for
104 weeks at concentrations of 0, 100, 500, 1500, and
6000 ppm (groups 15, respectively). These doses
equated to 0, 19.2, 98.5, 292.2, and 1266.1 mg/kg/day
for males and 23.8, 116.8, 354.2, and 1458.2 mg/kg/day
for females, respectively, of DEHP. Mice in group 6
were administered 6000 ppm for 78 weeks followed by
a 26-week recovery period in which they received basal
diet alone. In group 6 the average daily consumed dose
during weeks 1 to 78 was 1157.1 and 1253.5 mg/kg/day
in males and females, respectively.
The NOEL and LOEL values for systemic effects
were 100 and 500 ppm, respectively, with no evidence
of carcinogenicity. The DEHP dietary concentrations of
1500 ppm (group 4) and 6000 ppm (group 5) induced
gross and microscopic evidence of liver effects, including an increased incidence of liver neoplasia in males
and females. In group 6, administration of a dietary
concentration of 6000 ppm for 78 weeks followed by a
26-week recovery period indicated that effects of DEHP
in the liver, including the induction of neoplasms, and
in the kidney and testis were at least partially reversible following cessation of DEHP exposure.
Most of the tumorigenic effects seen in these two
carcinogenicity studies were in the high-dose animals
only. In these studies, there was a dose-related effect,

especially in the liver with peroxisome proliferation,
adenomas, and hepatocellular carcinomas. At lower
doses (20 mg/kg/day) in the mouse study, where no
peroxisome proliferation was seen, there were also no
increases in tumors, so there was a clear no-effect level
for this effect.
Many of these liver lesions were reversible after 79
weeks of DEHP administration. While the experimental design does not clearly allow demonstrating reversibility of neoplasia, the data are highly supportive of
this conclusion. In nearly all cases, the incidence of
adenomas and carcinomas was reduced in both mice
and rats. This interpretation is supported by the study
of Marsman and Popp (1994) who elegantly demonstrated that basophilic foci and adenomas were reversible in a stop study using WY-14,643. Such reversibility of adenomas provides evidence that peroxisome
proliferators such as DEHP and WY-14,643 exert tumor-promoting activity which is required for development of some lesions. Also, the fact that induced species and tumor increases were found in strains with
significant background incidences could reflect an enhancement of spontaneous tumorigenesis through promotion. As a promotor, the tumor response is most
likely described mathematically by a threshold rather
than a linear doseresponse model.
In addition, a very recent subchronic marmoset study
revealed no indication of peroxisome proliferation or
associated changes in these animals (Kurata et al.,
1998).
MODE OF RODENT HEPATOCARCINOGENIC ACTIVITY
OF PEROXISOME PROLIFERATORS, INCLUDING DEHP
As described elsewhere, DEHP and other peroxisome

proliferators induce profound morphological and biochemical changes in the livers of exposed rodents, characterized by an increase in both the number and volume of peroxisomes in the hepatocyte and the selective
induction of peroxisomal enzymes, as well as other
changes. These responses are accompanied by liver
enlargement and hyperplasia of hepatocytes. All hepatic effects have been attributed to activation of a
receptor, PPARa. Since, as with other peroxisome proliferators (Budroe and Williams, 1993), DEHP is not
genotoxic (vide infra), there is strong scientific consensus that its hepatocarcinogenic action in rodents is
associated with peroxisome proliferation and increased
hepatocyte replication.
Several hypotheses have been proposed to account
for the mode of action of non-genotoxic peroxisome
proliferators in producing liver tumors in rats and mice
(Ashby et al., 1994; Lake, 1995a,b). A recent workshop
organized by the International Life Sciences Institute
(Cattley et al., 1998) concluded the following:
1. Liver tumors in rodents resulting from exposure to
peroxisome proliferators are not produced through a

direct genotoxic mechanism (but rather through a secondary mechanism);
2. The role of enhanced cell proliferation is critical to
peroxisome proliferator-mediated hepatocarcinogenesis;
3. Oxidative stress is unlikely to play a major role in
peroxisome proliferator-mediated hepatocarcinogenesis,
but probably plays a contributory role;
4. There are marked species differences in response
to rodent peroxisome proliferators, with the rat and
mouse being extremely sensitive and certain other species, including humans, being relatively insensitive or
unresponsive;
5. Both peroxisome proliferation and enhanced cell
replication are good interspecies markers for peroxisome proliferator-induced hepatocarcinogenesis;
6. Peroxisome proliferation in the rodent is a consequence of PPARa activation;
7. Preliminary evidence indicates that PPARa activation is involved in peroxisome proliferator-induced
liver growth;
8. It is clear that the various PPAR subtypes differ
between rodents and humans, with respect to abundance in liver and activation by peroxisome proliferators;
9. The interspecies differences in PPAR are consistent with the lack of peroxisomal enzyme induction as
measured in human hepatocytes;
10. Enzyme induction (as measured by acyl-CoA oxidase and/or palmitoyl-CoA oxidation) and cell proliferation (as measured by DNA synthesis) are important
adjuncts in the characterization of the doseresponse
curve for peroxisome proliferator-induced liver tumors;
and
11. The shape of the doseresponse curve and the
quantitative species differences in response should be
considered in the risk assessment process.
The following sections discuss the main hypotheses
that have been proposed.
Oxidative Stress
As peroxisomes are the primary site for generation of
H 2O 2 within hepatocytes, increased oxidative stress
has been considered to have some involvement in the
mode of carcinogenic activity of peroxisome proliferators in rodent livers (Rao and Reddy, 1991).
Studies with DEHP and other peroxisome proliferators have yielded inconsistent data for effects on markers of oxidative stress such as lipid peroxidation, lipofuscin deposition, and hepatic antioxidants (Lake, 1995a).
The possibility of oxidative DNA damage by DEHP was
examined by Takagi et al. (1990a), who measured the
induction of 8-hydroxydeoxyguanosine (8-OH-dG) adducts in liver and kidney DNA from male rats fed a diet
containing 1.2% DEHP for 1 or 2 weeks. After exposure, hepatic DNA was extracted, hydrolyzed, and an-
333
alyzed by high-performance liquid chromatography and

electrochemical detection. DEHP caused statistically
significant increases in 8-OH-dG levels in liver but not
kidney DNA after both 1 and 2 weeks of exposure. The
observed increase in liver, however, was less than twofold greater than control. Subsequent studies have indicated that these minimal increases are not sustained
with chronic DEHP treatment (Takagi et al., 1990b;
Cattley and Glover, 1993). The increase in 8-OH-dG
caused by peroxisome proliferators has been somewhat
controversial, due to the usually small increases demonstrated for most agents administered. Moreover, a
paper by Sausen et al. (1995) on rats dosed with WY14,643 suggested that such increases might be due in
part to increases in mitochondrial DNA. These studies
are flawed, however, in that antioxidants were not
present during DNA and mitochondrial isolation, leading to high amounts of 8-OH-dG. Also, evaluation of
8-OH-dG levels in liver may represent an incomplete
approach to address the oxidative stress mechanism.
The evaluation of 8-OH-dG as the major measurement related to oxidative stress is likely to represent
an incomplete approach to address the oxidative stress
mechanism. Oxidative stress causes many additional
types of DNA damage that are not measured by 8-OHdG. This includes oxidized pyrimidines such as 5-OHcytosine and thymine glycol. It has also been shown
that lipid peroxidation leads to the formation of numerous cyclic DNA adducts including M 1G, a malondialdehyde adduct, propano DNA adducts, arising from crotonaldehyde and acrolien, and the etheno DNA adducts.
Furthermore, some of the oxidized bases rapidly depurinate/depyriminate, yielding abasic sites in the DNA.
To date, none of these adducts has been evaluated in
peroxisome proliferators. Recent studies have demonstrated that subchronic exposure to WY-14,643 also
caused induction of N-methyl purineDNA-glycosylase
(MPG), one of the DNA repair pathways for oxidative
DNA damage (Roussyn et al., 1997). In this study,
MPG expression was 250 300% greater in animals
exposed to the peroxisome proliferator. These investigators also found evidence for increased urinary excretion of ethenoadenine, an endogenous DNA adduct
formed as a result of lipid peroxidation and repair by
MPG. Thus, measurements of 8-OH-dG in DNA only
provide part of the relevant data related to oxidative
stress, since repair of these adducts by MPG removes
the altered bases, but leaves an abasic site in the DNA.
Such abasic sites are promutagenic.
Thus, oxidative stress appears to play a significant
role in peroxisome proliferator hepatocarcinogenesis.
It should be pointed out, however, that some oxidative
stress is present in all cells. It is well controlled by
normal homeostatic mechanisms unless these are overwhelmed by high chronic exposures to peroxisome proliferators which leads to a nonlinear increase in detrimental effects of oxidative stress when vitamin E and
334
DOULL ET AL.
glutathione S-peroxidase are absent. DEHP has been

shown to lack initiating activity in rat liver, an effect
which would be expected of an agent that gives rise to
alteration in DNA.
Cell Proliferation
Peroxisome proliferators induce changes in hepatocyte size and proliferation (both hypertrophy and hyperplasia) as well as effects on apoptosis and intercellular communication (Popp et al., 1994; Roberts et al.,
1995; Tsuda et al., 1995; Cattley et al., 1998). The
magnitude of sustained increases in cell proliferation
induced by peroxisome proliferators is highly indicative of eventual tumor yield (Cattley et al., 1998). In
this respect, DEHP is a weak inducer of cell proliferation (Marsman et al., 1988).
The induction of cell proliferation by peroxisome proliferators is considered to be a mitogenic response mediated by activation of PPARa. Thus, there is a linkage
between peroxisome proliferation and cell proliferation. The mitogenic mode of action of peroxisome proliferators distinguishes them from agents that produce
hepatocellular injury and compensatory regeneration.
This is confirmed for DEHP by the most recent rat and
mouse studies (David et al., 1996, 1997) of no hepatotoxicity as evidenced by unchanged or decreased liver
function tests with exposures that produced hepatocellular hypertrophy, hyperplasia, and increased liver
weight.
Cell proliferation is well established to lead to increases in tumor development through a variety of
mechanisms (Butterworth et al., 1989; Williams, 1992)
including tumor promotion, as discussed in the next
section.
Initiation and Promotion Studies
Peroxisome proliferators have not been shown to
exhibit initiating activity in two-stage models of rodent
hepatocarcinogenesis (Popp and Cattley, 1993), which
is consistent with their lack of genotoxicity. In particular, DEHP administered to rats at the carcinogenic
dose for up to 12 weeks did not produce a carcinogenic
response when followed by phenobarbital as a promoting agent (Williams et al., 1987; Garvey et al., 1987).
In studies of promoting activity, results in rats have
been mixed (Popp et al., 1985; Williams et al., 1987). In
contrast, in mice, studies have generally revealed promoting activity (Ward et al., 1983). Therefore, DEHP
has been considered to have liver tumor-promoting
activity (Huber et al., 1996).
The tumor-promoting activity of DEHP may be related to its enhancement of cellular proliferation since
DEHP differs in its biochemical effects on liver from
the classic liver tumor promotor phenobarbital (Maruyama et al., 1990). Enhanced cell proliferation would
serve to liberate initiated cells from the homeostatic
TABLE 4
Table of Abbreviations
Abbreviation
Written
PPARa
MOE
NOEL
LOEL
DEHP
EPA
POX
CAT
MEHP
2-EH
DEHA
8-OH-dG
MPG
RXR
mPPARa
IARC
WHO
TGF
PPRE
UDS
SSDB
PEL
STEL
ACGIH
Peroxisome proliferator-activated receptor a

Margin of exposure
No-effect level
Lowest-effect level
Di(2-ethylhexyl)phthalate
Environmental Protection Agency
Palmitoyl-CoA oxidation
Carnitine acetyltransferase
Mono-(2-ethylhexyl)phthalate
2-Ethylhexanol
Di-(2-ethylhexyl)adipate
8-Hydroxydeoxyguanosine
N-Methyl purineDNA-glycosylase
Retinoid X receptor
Mouse PPARa
International Agency for Research on Cancer
World Health Organization
Transforming growth factor
Peroxisome proliferator response element
Unscheduled DNA synthesis
Single-strand DNA breaks
Permissible exposure limit (OSHA)
Short-term exposure level
American Conference of Environmental
Industrial Hygienists
Maximum workplace concentration (Germany)
Lifetime average daily dose
MAK
LADD
controls that regulate cell turnover of the normal tissue.

Receptor Mediation
Induction of peroxisome proliferation in rodent hepatocytes is mediated through peroxisome proliferatoractivated receptors (PPARs) (Table 4) (Reddy and Mannaerts, 1994; Ashby et al., 1994; Lake, 1995a). PPARs
have been identified in several species, including rats,
mice, and humans, and are members of the steroid
hormone receptor superfamily. Four different PPAR
subtypes (alpha, beta, gamma, and delta) have been
identified in various species, of which the alpha-form
(PPARa) is directly involved in the induction of peroxisome proliferation in rodent liver. PPARs bind to DNA
as a heterodimer with the retinoid X receptor (RXR),
and peroxisome proliferator response elements have
been found in genes for peroxisomal and microsomal
fatty acid-oxidizing enzymes. In in vitro expression
systems, PPARs can be activated by both peroxisome
proliferators and certain fatty acids. Some PPARs may
be dominant repressors of other forms. Multiple PPARs,
together with activating (e.g., RXR) and repressing
receptors, may account for the known tissue and species differences in response to rodent peroxisome proliferators. The importance of the mouse PPARa (mP-
PARa) in the induction of peroxisome proliferation has

been recently demonstrated (Lee et al., 1995). In these
studies a mPPARa-deficient mouse was developed by
gene targeting and, unlike the wild-type mouse, did not
respond with peroxisome proliferation to the administration of peroxisome proliferators. A subsequent study
with these mice demonstrated that mPPARa was also
necessary for induction of cell replication by peroxisome proliferators (Peters et al., 1997).
A recently completed chronic feeding study of the
potent peroxisome proliferator WY-14,643 compared
the development of liver tumors in the knockout mice
lacking functional PPARa to that of wild-type mice
having normal functional PPARa (Peters et al., 1997).
WY-14,643 was fed at 0.1% of diet for 11 months. In
wild-type mice, treatment resulted in 100% incidence
of livers with multiple tumors. In PPARa knockout
mice, all livers were completely devoid of tumors. Using the same wild-type and null mice, Lee et al. (1995)
demonstrated increased hepatic cell proliferation and
peroxisome proliferation in wild-type mice exposed to
WY-14,463, but not in similarly exposed null mice.
These unequivocal results confirmed the requirement
for PPARa in peroxisome proliferation, cell proliferation, and liver tumor formation in this species. The
relevance of these findings for DEHP is clearly indicated by the in vitro data showing that its metabolites,
MEHP and 2-ethylhexanoic acid, are activators of PPARa
(Lapinskas and Corton, 1997).
Recently, Ward et al. (1998) demonstrated that
PPARa wild-type mice developed typical upregulation
of mRNA expression for peroxisomal and P450 enzymes in liver and kidney following exposure to diets
containing 12,000ppm DEHP, whereas null mice were
not different from control wild-type or null mice.
Whereas all hepatic toxicity was confined to the wildtype mice exposed to DEHP, some non-receptor-mediated toxicity also was noted in the kidneys and testes of
null mice (Ward et al., 1998). The severity of these
lesions in null mice was markedly decreased compared
to those in wild-type mice, with most of the toxicity
only being evident after all wild-type DEHP-exposed
mice had died. Thus, the hepatic toxicity was solely due
to PPARa activation, while both the renal and testicular toxicities had a receptor- and a non-receptor-mediated response.
Mice lacking peroxisomal fatty acyl-CoA oxidase develop severe peroxisomal proliferation and increased
mRNA levels of genes regulated by PPARa, suggesting
that acyl-CoA and other substrates of acyl-CoA oxidase
may be the natural ligands for PPARa (Fan et al.,
1998). The knockout acyl-CoA oxidase mice developed
hepatocellular adenomas and carcinomas by 15 months
of age. This model represents the equivalent of continuous occupancy of PPARa (Fan et al., 1998).
335
Evaluation
The effects of peroxisome proliferators on hepatic
peroxisomes and on hepatocyte growth are intimately
linked, and it is not possible to distinguish the relative
contribution of these two responses in the mode of
carcinogenic activity. However, it is generally accepted
that these two effects, when taken together, can completely account for the mode of carcinogenic activity of
peroxisome proliferators in rodent liver (Lake, 1995a,b;
Cattley et al., 1998). These adaptive effects, and ultimately carcinogenicity, are dependent on activation of
gene expression via a nuclear receptor, PPARa. Therefore, it becomes necessary to consider species differences in cellular responses and PPARa to determine if
the mode of action would be reasonably anticipated to
operate in humans.
SPECIES DIFFERENCES IN HEPATIC PEROXISOME
PROLIFERATION
Many studies have investigated species differences

in hepatic peroxisome proliferation (Rodricks and
Turnbull, 1987; Bentley et al., 1993; Ashby et al., 1994;
Lake, 1995a,b; Cattley et al., 1998). As discussed elsewhere in this paper, peroxisome proliferation in rodent
liver is mediated through PPARa. Important biomarkers of PPARa-mediated responses include peroxisome
proliferation (i.e., organelle proliferation and effects on
associated marker enzyme activities) and cell replication. Such biomarkers have been employed to assess
species difference in the effects of DEHP, DEHP metabolites, and other rodent peroxisome proliferators.
These studies have been conducted either in vivo or in
vitro in primary hepatocyte cultures and are reviewed
below.
In Vivo Studies in Experimental Animals
Rats and mice are clearly the most responsive species to rodent peroxisome proliferators. While some
strain differences in the response of rats and mice to
peroxisome proliferators have been reported, these are
minor in comparison with the magnitude of species
differences in response (Butler et al., 1988; Budroe et
al., 1992; Huber et al., 1996). Based on the activities of
marker enzymes (e.g., palmitoyl-CoA oxidation, lauric
acid 12-hydroxylase, CAT) and ultrastructural examination, the Syrian hamster appears to exhibit a somewhat lower response than rodents, whereas the guinea
pig and dog are either unresponsive or refractory. Some
examples of species differences in the relative potency
of DEHP and other rodent peroxisome proliferators are
shown in Table 5.
The ability of rodent peroxisome proliferators to produce hepatic peroxisome proliferation in nonhuman
primates has been evaluated in both in vivo and in
vitro studies (Williams and Perrone, 1996). The in vivo
336
DOULL ET AL.
TABLE 5
Examples of Species Differences in Hepatic Peroxisome Proliferation
Species examined
Compound
Responsive
Intermediate
Nonresponsive
Bezafibrate
Rat, mouse
Ciprofibrate
Rat, mouse
Dog, rabbit, Rhesus

monkey
Guinea pig, marmoset
Globuzarit
Clofibrate
Di(2-ethylhexyl)phthalate
Rat, mouse
Rat, mouse
Rat, mouse
Syrian hamster,
guinea pig
Syrian hamster,
rabbit
Syrian hamster
Syrian hamster
Syrian hamster
Dehydroepiandrosterone
LY 171883
Rat, mouse
Rat, mouse
Syrian hamster
Syrian hamster
Nafenopin
Oxadiazon
Rat
Rat, mouse
Syrian hamster
Guinea pig
Guinea pig, dog, Rhesus
monkey
Guinea pig, marmoset
Dog
Dog, marmoset
Marmoset, Rhesus monkey
Guinea pig, Cynomolgus
monkey, marmoset
Reference(s)
Watanabe et al. (1989)
Makowska et al. (1992), Graham
et al. (1994)
Orton et al. (1984)
Holloway et al. (1982)
Osumi and Hashimoto (1978),
Lake et al. (1984), Rhodes et al.
(1986), Short et al. (1987)
Sakuma et al. (1992)
Eacho et al. (1986)
Lake et al. (1989)
Richert et al. (1996)
Note. Peroxisome proliferation was assessed by ultrastructural examination and/or measurement of marker enzyme activities. Intermediate species are less responsive than the rat and mouse, whereas nonresponsive species either are refractory or exhibit only a small response
at high dosage. For further details see individual references.
studies include chronic investigations where peroxisome proliferators have been administered for extended
periods without any evidence of either significant peroxisome proliferation or tumor formation (see below).
The majority of studies performed in primates suggests that rodent peroxisome proliferators have little
or no effect in either Old World (e.g., Cynomolgus and
Rhesus monkey) or New World (e.g., marmoset) monkeys (Tucker and Orton, 1988, 1993, 1995; Bentley et
al., 1993; Lake, 1995a,b). However, albeit at high doses,
positive responses with ciprofibrate and DL-040 have
been reported in the Cynomolgus and/or Rhesus monkey (Reddy et al., 1984; Lalwani et al., 1985). The lack
of effect of rodent peroxisome proliferators on primate
liver in in vivo studies is supported by the results of in
vitro studies with cultured primate hepatocytes.
Some primate studies have been conducted with DEHP
and DEHP metabolites. No increase in peroxisome
numbers or associated enzyme activities were observed
in marmosets exposed by oral gavage to DEHP at up to
2000 mg/kg/day for 14 days (Rhodes et al., 1986). More
recently, marmosets exposed for 13 weeks to DEHP at
doses of up to 2500 mg/kg/day showed no treatmentrelated effects on hepatic peroxisomes or peroxisomal
enzymes (Kurata et al., 1998). Similarly, DEHP at
doses of 100 and 500 mg/kg/day for 21 days did not
produce hepatic peroxisome proliferation in the Cynomolgus monkey (Short et al., 1987). In agreement with
the lack of effect of DEHP on primate liver after in vivo
treatment, a number of DEHP metabolites have been
shown not to produce peroxisome proliferation in hepatocyte cultures from marmosets and Cynomolgus
monkeys.
In assessing species differences in hepatic peroxi-
some proliferation, a number of factors must be considered. These include the metabolism, disposition,
and dose of the test compound, sex differences, and
intrahepatic differences in response. Toxicokinetic
studies using clobuzarit, di-n-butylphthalate, 2,4-dichlorophenoxyacetic acid, gemfibrozil, and WY-14,643
have demonstrated that species differences in peroxisome proliferation and/or cell replication do not result
solely from differences in blood concentrations (Cattley
et al., 1998).
Although many studies have demonstrated that the
chronic administration of peroxisome proliferators may
produce liver tumors in the rat and/or mouse, few chronic
studies have been performed in other species. Clobuzarit
has been shown to produce liver tumors in the mouse and
both nafenopin and WY-14,643 produced tumors in both
rats and mice (Reddy and Lalwani, 1983; Tucker and
Orton, 1995), whereas in the Syrian hamster clobuzarit
did not produce tumors after 2 years (Tucker and Orton,
1995) and neither nafenopin nor WY-14,643 produced
tumors after 80 weeks of treatment (Lake et al., 1993a).
In the marmoset, clofibrate (which produces liver tumors
in the rat) did not produce any tumors in a 6 1/2-year
study, which is around half the expected life span of this
species (Tucker and Orton, 1993). Other studies in New
or Old World primates of ;6 month duration have been
conducted with several peroxisome proliferators, including ciprofibrate, clobuzarit, clofibrate, and fenofibrate.
While some increases in liver weight were reported, there
was no evidence of significant peroxisome proliferation or
peroxisome proliferator-induced liver lesions (Blane and
Pinaroli, 1980; Graham et al., 1994; Tucker and Orton,
1993, 1995; Williams and Perrone, 1996).
337
TABLE 6
Peroxisomal Effects of Hypolipidemic Drugs in Human Liver
Compound
Number of
patients
Ciprofibrate
Clofibrate
7
16
Fenofibrate
Fenofibrate
Fenofibrate
Gemfibrozil
8
38
12
9
Effect on peroxisomes a
Increase (30%) in volume density of peroxisomes
Increase (50%) in peroxisome numbers, but not in volume density of
peroxisomes
No change (qualitative)
No effect on either volume density or number of peroxisomes
Reference
Bentley et al. (1993)
Hanefeld et al. (1983)
Blane and Pinaroli (1980)
Blumcke et al. (1983)
Gariot et al. (1987)
De la Iglesia et al. (1982)
Effect on peroxisomes assessed by either subjective (qualitative) or morphometric (quantitative) analysis.
Human Studies
Studies in humans have been conducted in patients
treated with several hypolipidemic agents (all rodent
peroxisome proliferators), including ciprofibrate, clofibrate, fenofibrate, and gemfibrozil [reviewed in Bentley
et al., 1993; Ashby et al., 1994; International Agency
for Research on Cancer (IARC), 1995, 1996; Williams,
and Perrone, 1996; Cattley et al., 1998]. While most
studies with fenofibrate and gemfibrozil (Blane and
Pinaroli, 1980; De la Iglesia et al., 1982; Blumcke et al.,
1983; Gariot et al., 1987) failed to detect any changes in
peroxisomes in human hepatocytes, some morphological effect was noted in one study with clofibrate and in
one study with ciprofibrate (Table 6). In the clofibrate
study, morphometric analysis revealed a small increase in the number of peroxisomes, but no significant
change in the volume density of peroxisomes (Hanefield et al., 1983). As volume density is a more sensitive
indicator of peroxisome proliferation than peroxisome
number, the significance of this study is questionable
(IARC, 1995; Huber et al., 1996). In a study with ciprofibrate, a small 30% increase in the proportion of
hepatocyte cytoplasm occupied by peroxisomes was observed (reviewed in Bentley et al., 1993). As in other
studies, considerable interindividual variation in peroxisome morphometrics was observed, which together
with cell-to-cell and lobular variations makes it difficult to attach any clear biological significance to these
findings. Thus, humans, like nonhuman primates, are
refractory to peroxisome proliferation. This is in accord
with the low expression in humans of PPARa, which
mediates peroxisome proliferation by peroxisome proliferators.
Several epidemiological studies have examined the
association of cancer with the use of hypolipidemic
drugs which are peroxisome proliferators in rodents.
None has reported a clear association. In 1978, the
World Health Organization (WHO) reported the results of a randomized trial to determine whether clofibrate treatment would lower the incidence of ischemic
heart disease in men (WHO, 1978). This study reported
a nonsignificant excess of deaths from cancer in treated
subjects. The original WHO report stated that crude

mortality rates were higher for clofibrate-treated men,
but the age-adjusted mortality rates did not differ. This
could be accounted for by the fact that the treated
group was protected from ischemic heart disease, compared to the controls, so the treated men live longer
than the cohort to which they are being compared. One
entity that appeared to increase in the treated group
was gall stones. This may be related to an increased
secretion of bile acid produced by clofibrate. The increased intestinal symptoms also could be due to increased bile acids reaching the colon. In this report, the
incidence of liver tumors in treated patients was not
significantly different from controls, but when several
cancer sites were combined yielded the nonsignificant
excess of cancer deaths. All of these studies were reviewed by the IARC in 1996 and no evidence for carcinogenicity of clofibrate was found (IARC, 1996).
Subsequently the association between clofibrate and
cancer risk was examined in three randomized trials
and a small case control study. A further 4-year follow-up of the WHO trial showed no difference in the agestandardized death rates from malignant neoplasms.
In two other trials, there was also no difference in
cancer deaths between clofibrate-treated patients and
a placebo-treated group. A metaanalysis of results from
these trials also found no excess cancer mortality due
to use of clofibrate as a cholesterol-lowering drug (IARC,
1996).
In a trial with the hypolipidemic drug gemfibrozil,
also a rodent peroxisome proliferator, that was similar
to the WHO clofibrate trial, men were followed for 5
years. Cancer mortality was not affected by treatment,
although a borderline statistically significant increase
in basal cell carcinomas of the skin was found (Frick et
al., 1987). This study was reviewed by the IARC (1996)
with the evaluation of inadequate evidence in humans for the carcinogenicity of gemfibrozil.
With respect to DEHP, the International Program
for Chemical Safety of WHO has concluded that currently there is not sufficient evidence to suggest that
DEHP is a potential human carcinogen (IPCS, 1992).
338
DOULL ET AL.
TABLE 7a
Some Examples of Species Differences in the Effects of Rodent Peroxisome
Proliferators in Primary Hepatocyte Cultures a
Species
Compound
Beclobric acid
Benzbromarone
Bezafibrate
Ciprofibrate
Clofibrate
Clofibric acid
2-Ethylhexanol
2-Ethylhexanoic acid
Fenofibric acid
Fomesafen
LY171883
MEHP
MEHP metabolite VI
MEHP metabolite IX
Methylclofenapate
Mono-(2-ethylhexyl)
adipate and derivatives
Monoisodecyl and
monoisononyl phthalates
Nafenopin
Oxadiazon
Propaquizafop and free
acid derivative
Trichloroacetic acid
Wy-14,643
Rat
Mouse
Guinea pig
Marmoset
Cynomolgus
monkey
Rhesus
monkey
Human
Reference
Yes b
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
ND
ND
ND
Yes
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Yes
Yes
ND
ND
ND
Yes
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
No
ND
ND
ND
ND
ND
ND
ND
No
ND
ND
No
No
No
No
No
No
ND
ND
ND
No
No
No
No
No
No
No
ND
ND
ND
No
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
No
No
ND
ND
ND
No
ND
ND
ND
ND
No
No
ND
ND
ND
ND
No
No
ND
ND
ND
No
ND
No
ND
No
ND
ND
No
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
No
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
No
ND
No
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
No
ND
ND
ND
ND
ND
ND
ND
ND
ND
No (3) c
ND
No (3)
ND
ND
No (1)
ND
No (1)
No (3)
No (3)
No (1)
No (3)
ND
No (6)
ND
ND
No (3)
ND
No (3)
No (3)
ND
No (3)
No (2)
ND
No (3)
No (3)
No (3)
ND
No
No (3)
Blaauboer et al. (1990)

Mennes et al. (1994)
Bichet et al. (1990)
Bouis et al. (1993)
Foxworthy et al. (1990)
Allen et al. (1987)
Allen et al. (1987)
Blaauboer et al. (1990)
Butterworth et al. (1989)
Elcombe et al. (1996)
Mennes et al. (1994)
Cornu et al. (1992)
Cornu et al. (1992)
Cornu-Chagnon et al. (1995)
Smith and Elcombe (1989)
Dirven et al. (1993)
Elcombe and Mitchell (1986)
Elcombe and Mitchell (1986)
Mitchell et al. (1985b)
Elcombe and Styles (1989)
Yes
Yes
No
No
ND
ND
ND
Yes
Yes
Yes
ND
ND
ND
ND
ND
ND
No
No
ND
ND
ND
ND
ND
ND
ND
ND
ND
No (6)
Benford et al. (1986)

Bieri et al. (1988)
Yes
Yes
Yes
Yes
Yes
Yes
ND
ND
No
ND
No
ND
No
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
No (2)
No (3)
No (2)
Bouis et al. (1993)

Elcombe (1985)
Cornu et al. (1992)
a
Effects of test compounds were assessed by changes in enzyme activities (e.g., palmitoyl-CoA oxidation, lauric acid 12-hydroxylase) and/or
organelle proliferation.
b
Species response defined as yes, significant concentration-related response; no, no effect/little effect; ND, not determined.
c
When given in the reference, the figures in parentheses represent the numbers of different human hepatocyte preparations examined.
DEHP is leached in significant amounts from the

PVC tubing used in transfusion and dialysis equipment. Patients undergoing renal dialysis were studied
by Ganning et al. (1984, 1987) for evidence of liver
peroxisome proliferation. Based on a subjective ultrastructural evaluation, these authors concluded that
while no effect on liver peroxisomes was observed after
1 month of dialysis, increased numbers of peroxisomes
were observed in a liver biopsy from one subject after
12 months of renal dialysis. Ganning et al. (1984, 1987)

suggested that such effects may be due to exposure to
phthalate esters. However, such data should be viewed
with caution (Huber et al., 1996), as only a small number of liver biopsies were examined, no quantitation
was performed, and appropriate controls were not
studied.
In summary, the available data from in vivo studies
suggest that rodent peroxisome proliferators do not
339
TABLE 7b
Some Examples of Species Differences in the Effects of Rodent Peroxisome
Proliferators in Primary Hepatocyte Cultures a
Species
Compound
Rat
Mouse
Syrian
hamster
Guinea
pig
Marmoset
Human
Reference
Ciprofibrate
2-Ethylhexanoic acid
Fomesafen
Methylclofenapate
Yes b
Yes
Yes
Yes
Yes
Yes
ND
Yes
ND
ND
ND
ND
ND
ND
ND
ND
Yes
ND
ND
ND
ND
ND
ND
ND
ND
No
ND
ND
ND
ND
No
ND
ND
ND
No
ND
ND
ND
ND
ND
ND
ND
Yes/no d
ND
ND
No (4) c
No (3)
No (3)
No
No (3)
ND
ND
ND
No (5)
Perrone et al. (1998)

Elcombe and Styles (1989)
Bieri et al. (1984, 1990)
Bieri et al. (1988)
James and Roberts (1994, 1996)
Parzefall et al. (1991)
Nafenopin
Effects of test compounds were assessed by measurement of the hepatocyte labeling index or incorporation of [ 3H]thymidine into DNA.
Species response defined as yes, significant response; no, no effect/little effect, ND, not determined.
c
When given in the reference, the figures in parentheses represent the numbers of different human hepatocyte preparations examined.
d
Replicative DNA synthesis was increased in the absence but not in the presence of serum. However, no effect on peroxisome proliferation
was noted.
b
produce any significant peroxisome proliferation in human hepatocytes. This is consistent with the small
number of PPARa found in human livers. A large number of human subjects have been treated for relatively
long periods of time with hypolipidemic drugs which
are known potent rodent peroxisome proliferators. The
potential carcinogenic risk of hypolipidemic therapy
with clofibrate and gemfibrozil has been evaluated in
two, albeit limited, clinical trials (reviewed in Bentley
et al., 1993; IARC, 1995, 1996). Both these studies and
other human studies with lipid-lowering drugs showed
no indication of any increase in cancer associated with
long-term exposure to hypolipidemic peroxisome proliferators. Indeed, when recently evaluated by IARC
(1996), both clofibrate and gemfibrozil were placed in
group 3 (i.e., the agent is not classifiable as to its
carcinogenicity to humans).
In Vitro Studies in Cultured Hepatocytes
Studies in various laboratories have demonstrated
that primary hepatocyte cultures represent a validated
in vitro model for studying various aspects of hepatic
peroxisome proliferation including compound potency
comparisons and evaluation of species differences in
response (Lake and Lewis, 1993; Lake et al., 1993b;
Bentley et al., 1993; Bieri, 1993; Foxworthy and Eacho,
1994; Ashby et al., 1994). Many of the known characteristics of peroxisome proliferation in vivo, including
increased number and size of peroxisomes, differential
induction of peroxisomal enzyme activities, and stimulation of replicative DNA synthesis, have been observed in cultured rat and mouse hepatocytes treated
with peroxisome proliferators (Gray et al., 1983; Bieri
et al., 1984; Elcombe, 1985; Elcombe and Mitchell, 1986;
Lake et al., 1986; Bieri, 1993; Foxworthy and Eacho,

1994).
Generally, data from in vitro studies with primary
hepatocyte cultures from rats, mice, Syrian hamsters,
guinea pigs, and primates reflect the quantitative differences in peroxisome proliferation observed in in vivo
studies in these species (Bentley et al., 1993; Ashby et
al., 1994; Lake et al., 1995a,b; Cattley et al., 1998).
Some examples of compounds examined for their ability to produce peroxisome proliferation in vitro are
shown in Tables 7a and 7b. All of these compounds
produce peroxisome proliferation (i.e., organelle proliferation and/or stimulation of peroxisomal and/or microsomal fatty acid-oxidizing enzyme activities) in cultured rat hepatocytes. Where tested, such compounds
also produce peroxisome proliferation in mouse hepatocytes, but have little or no effect in guinea pig hepatocytes. In keeping with in vivo studies, cultured Syrian hamster hepatocytes exhibit an intermediate
response to rodent peroxisome proliferators (Lake et
al., 1986).
The effects of rodent peroxisome proliferators have also
been studied in cultured hepatocytes from New (e.g.,
marmoset) and Old World (e.g., Cynomolgus monkey,
rhesus monkey) primates. As shown in Tables 7a and
7b essentially no response has been reported in studies
with 16 different rodent peroxisome proliferators and
derivatives. Studies with human hepatocytes and a
range of 13 different rodent peroxisome proliferators
have failed to show any significant effects on either
organelle proliferation or peroxisomal (e.g., palmitoylCoA oxidation) or microsomal (e.g., CYP4A-dependent
lauric acid 12-hydroxylase) fatty acid oxidation. The
data in Tables 7a and 7b demonstrate that many of
340
DOULL ET AL.
TABLE 8
The Distribution of PPARa in Rat and Human Tissues within Species, Not across Species e
Species
Method a
Liver
Brain
Lung
Heart
Muscle
Kidney
Pancreas
Adipose
Reference b
Rat
Rat
Human
Human
Human
A
B
C
C
D
Low/high
High
High
Low
18 f
None
Low
Low
None
ND
ND d
Low
Low
Low
ND
None/mod c
High
High
None
ND
ND
High
High
High
5
High
High
High
Mod
5
Low
ND
High
None
ND
None/mod
ND
ND
ND
1
A
B
B
C
D
Note. In this table comparisons of relative PPAR expression levels are only valid for tissues within a species and not a cross species.
a
A, in situ hybridization/immunocytochemistry; B, RNAase protection; C, Northern blotting; D, RT-PCR.
b
A, Braissant et al. (1996); B, Mukhejee et al. (1994); C, Tugwood et al. (1996); D, Auboeuf et al. (1997).
c
Moderate.
d
Not determined.
e
Data also reported in Lemberger et al. (1996).
f
Attomole PPAR RNA/mg total RNA.
these compounds have been examined in a number of

human hepatocyte preparations in different laboratories. While no or little response to rodent peroxisome
proliferators was reported in the studies listed in Tables 7a and 7b, Perrone et al. (1998) observed a 1.8- and
2.7-fold induction of acyl-CoA oxidase activity in cultured human hepatocytes treated with clofibric acid
and ciprofibrate, respectively. These increases are very
small compared with the up to 40-fold increases often
observed in rodents.
The rodent peroxisome proliferators listed in Tables
7a and 7b include a number of active metabolites of
DEHP including MEHP, MEHP metabolites VI and IX,
2-ethylhexanol, and 2-ethylhexanoic acid. While all
these metabolites produce peroxisome proliferation in
cultured rat hepatocytes and where examined also in
mouse hepatocytes, they are essentially inactive when
tested in guinea pig, marmoset, and Cynomolgus monkey hepatocytes. Moreover, MEHP, MEHP metabolite
VI, and 2-ethylhexanoic acid (i.e., MEHP and two proximate peroxisome proliferators derived from DEHP)
have been shown not to produce peroxisome proliferation in a number of human hepatocyte preparations.
The clear species differences in peroxisome proliferation described in Tables 7a and 7b are not attributable
to a lack of functional viability of the hepatocyte preparations. In many instances where a negative response
to rodent peroxisome proliferators was reported, the
functional viability of the hepatocyte preparations used
was assessed in parallel experiments with other end
points. For example, although nonresponsive to rodent
peroxisome proliferators, human hepatocytes (Bichet
et al., 1990; Richert et al., 1996) and guinea pig hepatocytes (Mitchell et al., 1985b) were responsive to known
inducers of hepatic xenobiotic metabolism. While rhesus monkey hepatocytes did not respond to treatment
with three rodent peroxisome proliferators, dexamethasone was shown to induce tyrosine aminotransferase
activity (Foxworthy et al., 1990).
Another biomarker of the effect of rodent peroxisome
proliferators is the stimulation of cell replication, which

is observed in the livers of rats and mice after in vivo
administration. Several studies have examined the effects of a range of rodent peroxisome proliferators on
replicative DNA synthesis in cultured rat heptocytes
(Bieri et al., 1984, 1990; Elcombe and Styles, 1989;
Parzefall et al., 1991; Marsman et al., 1993; Elcombe et
al., 1996; James and Roberts, 1996; Perrone et al.,
1998; Plant et al., 1998). While some variation in the
magnitude of response has been observed between different laboratories, significant increases in replicative
DNA synthesis have been noted in the majority of
studies conducted.
Table 8 lists five rodent peroxisome proliferators,
including 2-ethylhexanoic acid (a proximate peroxisome proliferator derived from DEHP), which stimulate replicative DNA synthesis in cultured rat hepatocytes, but not in cultured human hepatocytes. These
compounds have been tested at various concentrations
in a number of human hepatocyte preparations, with
high concentrations resulting in an inhibition of replicative DNA synthesis (Elcombe et al., 1996; Perrone et
al., 1998). While rodent peroxisome proliferators do not
appear to induce replicative DNA synthesis in human
hepatocytes, increased replicative DNA synthesis was
observed in cultures treated with epidermal growth
factor, thus confirming the functional viability of the
human hepatocyte preparations used for these studies
(Parzefall et al., 1991; Elcombe et al., 1996; Perrone et
al., 1998).
Using a soft agar cloning assay, James and Roberts
(1994, 1995) observed that nafenopin could synergize
with epidermal growth factor to cause the clonal outgrowth of rat and mouse hepatocytes. In contrast, no
such effects were observed with human hepatocytes or
with Syrian hamster and guinea pig hepatocytes
(James and Roberts, 1995).
As in rat hepatocyte monolayer cultures, nafenopin
is also mitogenic in mouse hepatocytes (James and
Roberts, 1996). In contrast, nafenopin is not mitogenic
in Syrian hamster hepatocytes and neither methylclofenapate nor nafenopin stimulates replicative DNA
synthesis in guinea pig hepatocytes (Elcombe and Styles,
1989; James and Roberts, 1996). The effect of rodent
peroxisome proliferators on replicative DNA synthesis
in primate hepatocytes does not appear to have been
extensively investigated. In one study nafenopin was
reported to increase replicative DNA synthesis in marmoset hepatocytes when cultured in the absence, but
not in the presence, of serum (Bieri et al., 1988). The
significance of this finding is uncertain as these workers reported that nafenopin did not produce peroxisome proliferation (as assessed by a lack of effect on
palmitoyl-CoA oxidation activity and organelle numbers) in cultured marmoset hepatocytes.
A number of investigations have demonstrated that
peroxisome proliferators may inhibit the rates of basal
and/or transforming growth factor-b 1 (TGF-b 1)-induced
apoptosis in cultured rodent hepatocytes (James and
Roberts, 1996; Roberts, 1996; Perrone et al., 1998). By
using the terminal transferase dUTP biotin nick endlabeling technique, Perrone et al. (1998) demonstrated
that ciprofibrate and clofibrate had no effect on TGFb 1-induced apoptosis in human hepatocytes. In contrast, both peroxisome proliferators significantly increased the basal rate of apoptosis in cultured human
hepatocytes. These authors concluded that there were
differences between rat and human hepatocytes in the
effects of peroxisome proliferators on DNA synthesis
(see above) and apoptosis. As such, their data support
the hypothesis that human liver cells are refractory to
peroxisome proliferator-induced hepatocarcinogenesis
(Perrone et al., 1998). Unlike rodent hepatocytes, where
apoptosis may be suppressed, human hepatocytes respond to peroxisome proliferators with an increase in
apoptosis, which should be anticarcinogenic.
In conclusion, studies of in vitro responses to peroxisome proliferators have been consistent with reported
species differences in PPARa expression and in vivo
responses to these agents. The induction of key biomarkers of peroxisome proliferation, including fatty acidoxidizing enzyme activities (e.g., palmitoyl-CoA oxidation, acyl-CoA oxidase, lauric acid 12-hydroxylase),
organelle proliferation, and replicative DNA synthesis
has been clearly demonstrated in rodent hepatocytes
following in vitro treatment with a range of peroxisome
proliferators. The overall induction of fatty acid-oxidizing enzyme activities and/or organelle proliferation in
human hepatocytes has been reported to be weak in
one study and essentially nonexistent in other studies
with a range of 13 rodent peroxisome proliferators.
Moreover, the induction of replicative DNA synthesis
in human hepatocytes has been consistently negative
in studies with five rodent peroxisome proliferators.
Such agents also appear to exhibit different effects on
apoptosis in rodent and human hepatocytes. These findings clearly strengthen the conclusion that humans are
341
much less responsive than rodents to peroxisome proliferators including DEHP.

Species Differences in PPARa Expression
Marked interspecies variations in the expression of
PPARa mRNA have been demonstrated between rodent and human liver, the latter expressing 110% of
the levels found in mouse or rat liver (Palmer et al.,
1996, 1998; Tugwood et al., 1996).
Based on the knowledge that rodent PPAR is upregulated in different physiological states such as high
fat diets, glucocorticoid treatment, and diabetes, it has
been suggested that such pathophysiological changes
in humans could lead to increased PPARa levels. However, examination of liver samples from diabetics, alcoholics, and others presenting fatty livers, conditions
that may be expected to increase PPAR levels, has
demonstrated no increase in PPARa (Palmer et al.,
1998). Additionally, in vitro experiments using human
hepatocytes have been performed in the presence of
glucocorticoids and no evidence for peroxisome proliferation and related effects has been consistently reported.
Using a sensitive and specific immuno/DNA binding
assay, Palmer et al. (1998) have shown that active
PPARa protein is expressed at variable concentrations
in human livers. This study compared 20 different
human livers and found that the highest PPAR protein
expressors contained less than 10% of the mouse activity. Most of the samples, 13/20, contained no detectable
PPAR activity, but did exhibit peroxisome proliferator
response element (PPRE) binding activity that did not
correspond to PPARa. These data support a threshold
through competition concept. In most human samples
studied it was found that PPREs are mainly bound by
other competing proteins that may block peroxisome
proliferator responsiveness (Palmer et al., 1998).
In this study the levels of PPARa protein were lower
than anticipated by RNA analysis and this was explained by the finding that a significant fraction of
PPARa mRNA is misspliced in human liver (Palmer et
al., 1998).
It is possible that the low level of PPARa detected in
human liver may explain the apparent lack of responsiveness of human heptocytes to peroxisome proliferators. In support of this hypothesis, Palmer et al. (1998)
have shown that transcriptional responses to peroxisome proliferators are not evident in the Huh7 cell line,
which contains similar levels of PPARa and RXR to
those found in intact human liver. However, increasing
PPARa levels by heterologous expression in the Huh7
cell line conferred responsiveness to peroxisome proliferators, thus suggesting that the low level of endogenous PPARa is the major factor in the nonresponsiveness of these cells.
Peroxisome proliferator responsiveness is unlikely to
342
DOULL ET AL.
be a linear function of PPAR expression levels. It is

likely that there will be threshold levels where a certain amount of PPAR is required before an effect is
seen. This may be due to competition for binding to
PPREs by other transcription factors such as HNF4,
ARP-1, and TR (Palmer et al., 1994; Miyamoto et al.,
1997), binding of PPARa to other transcriptional factors (e.g., LXRa) leading to formation of nonfunctional
heterodimers (Miyata et al., 1996), competition of PPARa
with other hormone receptors (e.g., TR) for dimerization with RXR (Juge-Aubry et al., 1995), and lack of
coactivating proteins such as SRC-1 and PBP (Zhu et
al., 1997). Other PPARs may compete with PPARa; for
example, PPARD (hNUC1) can repress the activation
of hPPARa (Jow and Mukherjee, 1995).
The insensitivity toward peroxisome proliferators
exhibited by human hepatocytes is reflected in guinea
pigs which also express similar low levels of PPARa
(Tugwood et al., 1998; Bell et al., 1998).
Although significant responses such as peroxisome
proliferation, induction of acyl-CoA oxidase, and stimulation of replicative DNA synthesis (phenomena believed to be associated with the hepatocarcinogenic
process) are not observed in humans and guinea pigs,
these species exhibit hypolipidemia when exposed to
some peroxisome proliferators (Bell et al., 1998). The
hypolipidemic response has been attributed to regulation of lipoprotein lipase and various lipoproteins. These
effects are weak and are not associated with the hypertrophic and hyperplastic responses in guinea pig
and human studies.
This delineation may be due to differences in the
mechanism of action of PPAR on the hypolipidemic
genes, which may alter the threshold of activation
and require lower concentrations of receptor, due to
differences in binding affinities for different PPREs,
than the genes involved in the hypertrophic and hyperplastic responses. It is also plausible that the hypolipidemia genes bind competitors less efficiently than they
bind PPAR. Differences in the activation properties for
different promoters have been described; for example,
the acyl-CoA oxidase and CYP4A6 promoters show different degrees of intrinsic activation by PPARa and
display different EC 50s for activation by peroxisome
proliferators (Hsu et al., 1995).
three receptors cloned to date were not activatable by

peroxisome proliferators (Tugwood et al., 1996).
Thus, at present, an argument for a subpopulation of
genetically sensitive individuals is without any sound
mechanistic or experimental basis.
Interindividual Variation in PPARa Expression and

Activity
Evaluation
There is no good evidence for genetic polymorphisms in

the human PPARa. Although different DNA sequences
have been reported, neither their frequency nor even
their existence in the human population has yet been
determined (Mukherjee et al., 1994; Tugwood et al., 1996)
and they may be merely a result of the introduction of
artifacts in the PCR-based cloning process. Two of the
Tissue Differences in PPARa Expression

As yet, there are no reported sites of robust PPARa
expression in humans (Table 8). Reports of high levels
(the same or more than liver) of PPARa in human
skeletal muscle (Mukherjee et al., 1994; Tugwood et al.,
1996) must be regarded with suspicion, as the same
commercially available Northern blot showed a similar
distribution of signal using a b-actin normalization
probe. This suggests that loading differences may play
a large part in determining the signal intensity obtained. In addition, the two studies were contradictory
regarding the relative levels of expression of PPARa in
the heart. The possibility of experimental error is supported by other work, using competitive RT-PCR (a
more quantitative technique), showing PPARa expression in human skeletal muscle to be only 25% of that
found in liver (Auboeuf et al., 1997).
Any role for PPARa in mediating carcinogenesis in
nonhepatic tissue is doubtful, since rat skeletal muscle,
heart, and kidney express similar high levels of receptor as the liver (Mukhejee et al., 1994; Braissant et al.,
1996) and peroxisome proliferators have not been implicated in carcinogenesis in these extrahepatic organs.
The extremely low level of PPARa expression in human tissues, both hepatic and nonhepatic, argues against
carcinogenesis mediated via the PPARa activation
pathway.
In conclusion, the well-documented species differences in peroxisome proliferation are probably largely
due to low-level expression of PPARa in the nonsensitive species. Additionally, competitive factors for PPRE
and RXR binding may contribute to the species differences in transcriptional activation observed. The low
level of PPARa expression in human tissues argues
against the potential carcinogenesis of peroxisome proliferators in humans. As yet, no evidence has been
found for the existence of a sensitive, responsive human population.
There is clear evidence for species differences in xenobiotic-induced hepatic peroxisome proliferation and
cell replication. Peroxisome proliferators produce marked
biochemical and morphological changes in rat and mouse
hepatocytes and following prolonged exposures may
also produce liver tumors. Species other than the rat
and mouse are clearly less responsive or refractory to
peroxisome proliferators. The available data suggest
343
that peroxisome proliferators do not produce liver tumors in such species.

The most consistent and well-documented species
differences in response to peroxisome proliferators are
in the enzyme as well as the enzyme receptor studies.
These differences provide an elegant basis for clearly
distinguishing between the susceptible species (rats
and mice) and less susceptible and/or refractory species
(guinea pigs, monkeys, and humans). These in vivo
differences have been confirmed by a number of in vitro
studies. Attempts to increase the in vitro dose to effective levels in primary hepatocyte cultures can result in
excessive cytotoxicity without evidence of peroxisome
proliferation. The peroxisome proliferation, cell proliferation, and eventual tumor formation observed in mice
and rats are dependent on the robust expression of
PPARa, which is not observed in human liver (Palmer
et al., 1996, 1998; Tugwood et al., 1996). Studies with
transgenic mice (knockout mice) support these conclusions (Peters et al., 1997).
The hepatic effects of DEHP in the rat and mouse
are characteristic of a relatively weak peroxisome
proliferator. Moreover, in keeping with other rodent
peroxisome proliferators, there are marked species
differences in DEHP-induced hepatic peroxisome
proliferation. While DEHP is more extensively absorbed after oral administration in the rat than
in the marmoset (Rhodes et al., 1986), the observed
species differences in vivo are undoubtedly explained
by the observation that the metabolites of DEHP
that produce peroxisome proliferation in rat hepatocytes in vitro have no significant effect on marmoset
hepatocytes. Similarly, the same DEHP active
metabolites have no effect on cultured human
hepatocytes (Elcombe and Mitchell, 1986; Butterworth et al., 1989; Bichet et al., 1990; Elcombe et al.,
1996).
Although rodent peroxisome proliferators do not appear to produce significant peroxisome proliferation in
human hepatocytes, the hypolipidemic effects of these
agents are observed in several species including rodents and humans (Huber et al., 1996). Studies in
rodents have demonstrated that a hypolipidemic effect
is not necessarily associated with hepatic peroxisome
proliferation (reviewed in Lake and Lewis, 1993). Thus,
an observed hypolipidemic effect should not be construed as evidence of peroxisome proliferation in human liver.
Data both from in vivo studies and from in vitro studies
with cultured hepatocytes demonstrate that rodent peroxisome proliferators do not produce any marked effects
in human liver. The far greater sensitivity of the rat and
mouse to DEHP and other peroxisome proliferators demonstrates that any calculation of human risk based on
rodent data and doses will result in a marked overestimation of the actual risk. Epidemiology studies have
failed to demonstrate any increase in tumor incidence in
TABLE 9
Summary Listing of Genotoxicity Data
for DEHP and MEHP
Assay
Prokaryotic mutation assays
Salmonella
Other bacteria
Yeast/fungus genotoxicity assays
Gene mutation
Mitotic recombination
Aneuploidy
Drosophila mutation assays
Mammalian DNA damage assays
Binding of radiolabeled chemical
32
P-postlabeling assay
Single-strand DNA breaks
Unscheduled DNA synthesis
Mammalian mutation assays
Chinese hamster ovary cells
L5178Y mouse lymphoma cells
Syrian hamster embryo cells
Hamster V79 cells
Human lymphoblast cells
Mammalian chromosome damage assays
Rodent in vitro chromosomal damge
Aberrations
Aneuploidy
Human in vitro chromosomal damage
Rodent in vivo chromosomal aberrations
Rodent sister chromatid exchange
Human sister chromatid exchange
In vitro rodent micronucleus
In vivo rodent micronucleus
In vivo co-recombinogenesis
Mammalian cell transformation
C3H1OT1/2 cell
SHE cell
BALB/c-3T3 cells
Mammalian germ cell mutation assays
Dominant lethal
DEHP
MEHP
2
2
1/2, 2
1/2, 2
1/2, 2
2
1?
1/2, 2
ND
2
2
2
2
2
ND
ND
2
2
2
?
2
2
ND
2
?
?
ND
2?
1
2
2?
2?
2
2
2
1
1
ND
ND
2?
1?
ND
ND
ND
1/2
1
2
ND
ND
ND
1?
ND
Note. 1, positive; 1/2, weakly positive; 1 ?, potentially positive; 2,

negative; 2 ?, potentially negative; ?, inconclusive; ND, no data.
humans exposed daily to therapeutic doses of more potent rodent peroxisome proliferators for extended periods
of time (IARC, 1995, 1996).
GENOTOXICITY
In this assessment, genotoxicity includes molecular

changes in DNA, such as adduct formation or oxidative
damage, and biological responses that can be attributed to molecular changes in DNA. Most assessments
of the overall data on DEHP and its metabolite MEHP
have concluded that there is no clear genotoxic potential (International Programme on Chemical Safety, 1992;
Budroe and Williams, 1993; Dybing et al., 1995; Huber
et al., 1996). The genetic toxicology studies are summarized in Table 9.
344
DOULL ET AL.
Overall Assessment on Genotoxicity of DEHP

More data are available on DEHP than on MEHP, in
part because DEHP was studied in the International
Programme on Chemical Safety Collaborative Study
(Ashby et al., 1985). As summarized in Table 9, DEHP
and MEHP clearly did not induce gene mutation in Salmonella and other bacteria, yeast, and fungi or Drosophila. They also generally lacked activity in mammalian
gene mutation assays, did not covalently bind to DNA or
induce unscheduled DNA synthesis (UDS), and did not
induce single-strand DNA breaks (SSDB), which is consistent with the absence of sites of potential electrophilicity in their structures. DEHP did not induce chromosomal aberrations in most in vitro systems or in vivo in
rodents and did not induce micronuclei or sister chromatid exchanges in rodent cells or increased chromosomal
aberrations or sister chromatid exchange frequencies in
human cells exposed in vitro. However, data suggest that
DEHP may induce aneuploidy in yeast and two studies
reported chromosome effects in rodent cells in vitro.
MEHP produced chromosome aberrations in cultured rodent cells. Extrapolation of these in vitro effects is uncertain. DEHP induced transformation in some cell systems
and was co-recombinogenic in mice, all responses that
can be produced by non-genotoxic agents. Data on the
genotoxicity of DEHP to rodent germ cells in the dominant lethal assay were inconclusive. Overall, the data
base supports the conclusion that DEHP and MEHP are
non-genotoxic. Accordingly, they appear not to possess
the essential feature of a carcinogen that acts as an
initiator.
rigid application of the linear multistage model, the

new guidelines allow methodology more appropriate to
chemicals acting through threshold-based, non-genotoxic mechanisms when, indeed, sufficiently strong
supportive evidence exists.
The nonlinear approach advocated in the proposed
guidelines is the MOE, which is defined in the Guidelines as the LED 10 (or other point of departure, such as
the NOEL) divided by the environmental exposure of
interest (the latter being actual or projected future
levels). This approach is consistent with those currently used for assessing the risks of noncarcinogens.
The proposed guidelines state that If in a particular
case, the evidence indicates a threshold, as in the case
of carcinogenicity being secondary to another toxicity
that has a threshold, the margin of exposure analysis
for the toxicity is the same as is done for a non-cancer
endpoint. As will be discussed later, the size of the
MOE that is considered acceptable will depend on the
results of a MOE analysis that takes into account the
overall strength of the data, knowledge about the
mechanism of action and its relevance to humans, and
a number of other factors that require individual judgment decisions.
DEHP is an excellent candidate for reconsideration
under the new Guidelines for Carcinogen Risk Assessment and there is strong scientific justification for using
the MOE approach for assessing its potential human
risks. This document has provided detailed evidence that
Alternative Approach
1. DEHP is a hepatocarcinogen in F344 rats and

B6C3F1 mice. The most recent studies clearly indicate
a nonlinear doseresponse curve that indicates the
existence of a dose threshold below which tumors are
not induced;
2. The total weight-of-evidence indicates that DEHP
is not genotoxic;
3. DEHP exerts its tumorigenic effect in rodent liver
through its action as a peroxisome proliferator. This
action is initiated by the activation of PPARa, a nuclear receptor that initiates the transcription of genes
for increased peroxisome proliferation and cell proliferation; and
4. The hepatocarcinogenic response observed with
DEHP is secondary to the interaction of DEHP (or
more accurately its activated metabolite, MEHP) with
the PPARa receptor. The latter is a reversible, receptor-mediated process that exhibits a clearly nonlinear,
threshold-based doseresponse curve.
EPAs Proposed Guidelines for Carcinogen Risk Assessment (EPA, 1996) allow consideration of several
alternative methods for analyzing carcinogen dose
response data and for extrapolating the effects observed at relatively high doses in laboratory animals to
predict those that might occur at much lower doses
relevant to human exposure. Thus, in contrast to the
Based on the established association of DEHP-induced rodent liver cancer with a receptor-mediated
mechanism and general scientific consensus that receptor mechanisms involve dose thresholds (Poland,
1997), it is concluded that risk assessment with DEHP
can most appropriately be conducted using the MOE
approach.
Current Risk Assessment of DEHP

Currently DEHP is classified by EPA as a B2 probable human carcinogen, based on the NTP bioassays
conducted in 1982 with F344 rats and B6C3F1 mice
(NTP, 1982). Using the mouse liver data from the NTP
bioassay, EPA calculated a cancer potency factor (q*) of
1.41 3 10 22 (mg/kg/day) 21 (EPA, 1987, 1989). In arriving at this conclusion, the EPA assumed a linear, nothreshold dose response. In the past, this assumption
has dictated that risk assessment be conducted using
the linear multistage model, which was required under
EPAs 1986 Cancer Risk Guidelines.
345
TABLE 10
Morphological, Biochemical, and Tumorigenic Observations from DEHP Bioassays
with Rats and Mice (David et al., 1996, 1997)
F344 Rats
Group/dose
100 ppm
7.3 mg/kg/day (F)
5.8 mg/kg/day (M)
500 ppm
36.1 mg/kg/day (F)
28.9 mg/kg/day (M)
2500 ppm
181.7 mg/kg/day (F)
146.6 mg/kg/day (M)
12500 ppm
938.5 mg/kg/day (F)
789.0 mg/kg/day (M)
B6C3F1 Mice
Observation
No
No
No
No
Group/dose
treatment-related effects
organ weight changes
histopathological changes
detectable peroxisome prolifer
100 ppm
No treatment-related effects
No organ weight changes
No histopathological changes
No detectable peroxisome prolifer
NOEL: Peroxisome proliferation
Increased peroxisome proliferation
Increased liver weight
Increased incidence of hepatocellular
adenomas (males only)
NOEL: Carcinogenic potential
Decreased survival
adenomas and carcinomas
500 ppm
It is further concluded that the point of departure

from which the MOE should be calculated is the mouse
NOEL for peroxisome proliferation. This is a highly
conservative point of departure for regulatory assessment because
1. The mouse is the most sensitive species;
2. Peroxisome proliferation is a biochemical precursor effect and its NOEL is about fivefold lower than the
NOEL for tumorigenesis (see below); and
3. The approach assumes that the peroxisome proliferation data in mice are relevant to humans, whereas,
as discussed extensively in this document, there is
strong scientific evidence demonstrating that humans
are nonresponsive to peroxisome proliferators.
The data from the latest rodent bioassays with DEHP
(David et al., 1996, 1997) show the B6C3F1 mouse to be
the most sensitive species with NOELs for both peroxisome proliferation and increased liver weight of 19.2
mg/kg/day for males and 23.8 mg/kg/day for females
(average of about 20 mg/kg/day) (Table 10). This indicates that both cell proliferation (reflected by liver
weight) and peroxisome proliferation are occurring at
similar doses, the LOEL being 98.5 mg/kg/day in males
and 116.8 mg/kg/day in females. In F344 rats, the
NOEL values for peroxisome proliferation are 28.9 mg/
kg/day for males and 36.1 mg/kg/day for females with
corresponding LOELs of 146.6 and 181.7 mg/kg/day,
respectively (Table 10). The NOEL of 20 mg/kg/day
from the David et al. (1997) mouse study is in agree-
23.8 mg/kg/day (F)

19.2 mg/kg/day (M)
116.8 mg/kg/day (F)

98.5 mg/kg/day (M)
1500 ppm
354.2 mg/kg/day (F)
292.2 mg/kg/day (M)
6000 ppm
1458.2 mg/kg/day (F)
1266.1 mg/kg/day (M)
Observation
No treatment-related effects
No organ weight changes
No detectable peroxisome prolifer
NOEL: Peroxisome proliferation
Increased liver weight (males only)
NOEL: Carcinogenic potential
Decreased survival
ment with data from a number of other studies. Based

on an evaluation of the literature, Huber et al. (1996)
concluded that, for risk assessment purposes, the
NOEL and LOEL for DEHP-induced hepatic peroxisome proliferation in the rat are 15 and 50 mg/kg/day,
respectively. The NOEL of 20 mg/kg/day from the
David et al. (1997) mouse study is about fivefold lower
than the NOEL for cancer (both adenomas and carcinomas) observed in the same study (116.8 mg/kg/day
for females and 98.6 mg/kg/day for males) (Table 10).
Human Exposure Assessment
A complete characterization of human exposures to
DEHP as outlined in the U.S. EPA Guidelines for Exposure Assessment (U.S. EPA, 1992) is outside the
scope of this document. Instead, a brief summary of
major sources and routes of potential human exposure
to DEHP will be provided and MOEs will be calculated
for several worst-case exposure scenarios.
Potential human exposures to DEHP may be occupational (mainly inhalation) or may occur in the general
population through ingestion of residues in food or water
or inhalation of ambient air. Special exposures also occur
through a variety of medical treatments (e.g., transfusion
and hemodialysis) (Huber et al., 1996; ATSDR, 1993;
IPCS, 1992) and since these represent some of the highest nonoccupational human exposures they will be given
particular attention in this discussion.
346
DOULL ET AL.
GENERAL POPULATION
The major source of exposure for the general population is via food residues arising from migration of
DEHP into food from plastic containers. In general,
food residues do not exceed 1 mg/kg (1000 mg/kg) of
food, although considerably higher residues may be
associated with fatty foods such as oils, milk, cheese,
meat, and fish. The magnitude of the residues will
depend on the DEHP content of the container/wrapper
material and the duration and conditions (temperature) of storage. The U.S. FDA has established a limit
of 3% DEHP content for all materials used to store
foodstuffs. It is generally estimated that the total daily
intake of DEHP in food products ranges from about
0.27 mg/day (3.8 mg/kg/day) to a high of about 2.0
mg/day (30 mg/kg/day) (Huber et al., 1996).
In comparison with possible intakes from food, DEHP
exposures from ambient air and drinking water are very
small (Huber et al., 1996) and make an insignificant
contribution to the total daily intake of the general population.
OCCUPATIONAL EXPOSURES
Occupational exposures may be significant during

the manufacture of a wide variety of DEHP-containing
PVC products; most occupational exposures occur via
inhalation (ATSDR, 1993; Huber et al., 1996). The National Institute of Occupational Safety and Health
(NIOSH) has estimated that about 340,000 workers
were potentially exposed to DEHP in the early 1980s
(ATSDR, 1993) and workplace levels ranging up to
about 4 mg/m 3 were reported at about this time. It is
likely that actual worker exposures are considerably
below the OSHA permissible exposure limit (PEL). A
recent worker-monitoring study conducted by Aristech
in a plasticizer plant found that 20 of 24 person samples were below the detection limit and the highest
value indicated an exposure of only 0.04 mg/m 3 during
an 8-h working day; this is more than 100 times lower
than the PEL. The current OSHA PEL time-weighted
average for DEHP for an 8-h work day is 5 mg/m 3 and
the short-term exposure level (STEL) is 10 mg/m 3
(OSHA, 1989). American Conference of Environmental
Industrial Hygienists (ACGIH) threshold limit values
(TLVs) are the same as the OSHA PELs and the corresponding German MAK value (maximum workplace
concentration) is 10 mg/m 3 (Huber et al., 1996).
Assuming that a worker inhales about 10 m 3 of air
during a typical working day, currently permissible
workplace intakes of DEHP range from 50 mg/person/
day (OSHA PEL and ACGIH TLV) to 100 mg/person/
day (German MAK); assuming a 70-kg person, these
are equivalent to intakes of 0.71.4 mg/kg body wt/day.
If we further assume a 5-day working week and a
40-year working life, the lifetime average daily dose

(LADD) of DEHP will be 0.28 0.57 mg/kg body wt/day.
EXPOSURE FROM MEDICAL DEVICES
A multitude of medical devices constructed from PVC

are used regularly in the United States and around the
world. These devices range from simple adhesive bandages to blood transfusion and dialysis equipment, implants such as pacemakers, nasogastric tubes, etc. DEHP
(up to 30% by weight) is commonly incorporated into
many of these products to impart the desired properties
of softness and flexibility.
In view of the well-established ability of DEHP to migrate from the surface of PVC materials, it is perhaps not
surprising that when receiving intravenous transfusions
of fluids, such as whole blood or blood components that
have been stored in PVC containers (e.g., blood bags),
patients coming into contact with these devices may be
exposed to significant levels of DEHP. PVC tubing used
for administration of these fluids or during hemodialysis
or autopheresis potentially may also contribute to patient
exposures. Since medical devices undoubtedly represent
a major nonoccupational route of human exposure to
DEHP, they will be discussed in some detail.
Obtaining realistic estimates of human exposure to
DEHP from medical devices depends, in part, on a thorough
understanding of the nature of these devices and the
standard clinical practices under which they are used.
Examples of generic-use profiles for selected medical devices are shown in Table 11. The following discussion
provides estimates of DEHP exposures associated with
typical use of the medical devices indicated in Table 11.
EXPOSURE THROUGH USE OF INTRAVENOUS SETS
FOR TRANSFUSION OF PLATELETS
On a weight basis, DEHP may constitute 30 40% of a

typical blood bag (Rubin and Schiffer, 1976). Jaeger and
Rubin (1972) were one of the first research teams to
report the leaching of DEHP from PVC blood bags into
stored blood components; their data suggest a leaching
rate of 0.25 mg DEHP/100 ml/day for whole blood stored
at 4C. Jacobson et al. (1977) conducted a study of pediatric aplastic anemia and leukemia patients that quantified the number of units of platelets received by the
patients from PVC blood bags over the course of 1 year of
treatment. Rubin and Schiffer (1976) reported the leaching of DEHP into platelet concentrate (PC) stored in
blood bags and quantified the levels in the stored platelets. It was estimated that each patient received a total of
26.4 to 82.4 mg DEHP, with an arithmetic mean of 60.3
mg DEHP and a standard deviation of 20.6 mg for 5 bags.
The mean of Rubin and Schiffer (1976) is considerably higher than the value of 14.7 mg DEHP received
from PC stored for 72 h at 22C, when the data of Rock
et al. (1978) are used. The data of Labow et al. (1986)
347
TABLE 11
Use Information for Medical Devices Containing DEHP
Use information
Product
category
Selected use
scenario
Duration per event
No.
units/day
Replacement
time
Treatment regimen
Comments
iv blood sets
Transfusion of
platelets in
leukemia
patients
2060 min (average 30

min) (Rubin and
Schiffer, 1976)
1 per session
70 times/year for 2
years (Jacobson et
al., 1977)
Hemodialysis
sets
Short-term and
long-term
hemodialysis
Typically 5 h (Flaminio
et al., 1988; Lewis et
al., 1978)
1 per session
Autopheresis
sets
Donation of plasma
and return of
RBCs
60 min (assumption)
1 per session
Nasogastric
tubes
Enteral feeding
tubes
2 days
(assumed)
33/week for 2 months

(short-term trauma
victim) and 15
years (long-term
patient)
Once/month for 2
months (short-term
donor) or 5 years
(long-term donor)
12 days for trauma
patient and 1 year
for long-term use
24 h
suggest an average value of 13 mg DEHP for PC for five

different bag products involving storage for 5 days at
22C. 2 Lower, but significant amounts of DEHP (4.45
to 10.2 mg) have been extracted from sets of PVC iv
tubing as a result of 5 h of perfusion of plasma through
the tubing (Easterling et al., 1974). Because infusion of
platelets requires typically 30 min, and assuming a
linear leaching rate of DEHP with time, the tubing
involved in administering platelets might contribute at
most (10.2 mg) 3 (30 min)/(300 min) 5 1.0 mg. Ignoring this minor contribution of tubing to DEHP exposures, and based on the combined data of Rock et al.
(1978) and Labow et al. (1986), a weighted mean of 13
mg DEHP/unit is obtained. Based on this value, a
LADD of 71 mg/person/day is obtained.
This calculation of DEHP exposures from transfusion
of platelets can be arrived at using the weighted mean of
13 mg DEHP/unit (see above); the exposures over each
year of treatment (E T) and over a lifetime (LADD), assuming a life span of 70 years, can be calculated as
Exposure per treatment ~mg!
3 ~No. treatments/year! 3 ~1000 mg/mg!
ET 5
~365 days/year!
Treatment period exposure
3 ~2 years treatment!
LADD 5
.
70 Years
2
The clear outlier of 649.5 mg for one bag type was discarded from
the original data set.
Maximum storage time

for platelets is 3 days
in DEHP-PVC bags
(Myrhe et al., 1987;
Gulliksson et al., 1991)
Assumptions on
treatment regimen are
consistent with current
clinical practice
Assumptions on
treatment regimen are
based on anticipated
donor behavior
Tubing estimated to be 1
oz (28 g) in weight
Thus, for leukemia patients, the treatment period exposure (E T) and LADD are
13 mg DEHP
3 ~70 treatments/year! 3 ~1000 mg/mg!
ET 5
~365 days/year!
5 2,500 m g/person/day
LADD 5
2500 mg/person/day 3 ~2 years!

~70 years!
5 71 mg/person/day.
EXPOSURE THROUGH USE OF HEMODIALYSIS

EQUIPMENT
Because of the rapid clearance of DEHP from the blood,

the circulating blood concentration of DEHP in patients
during hemodialysis does not reflect the total body burden
of the chemical. Thus, circulating blood levels may underestimate the total delivered dose of DEHP unless adjusted
for metabolism and time. Several studies have been conducted in which the concentration gradient across the dialysis unit (i.e., difference in DEHP levels at the outflow and
inflow ports) has been used to successfully quantify the
time-integrated dose of DEHP delivered to the patient during a 4- to 5-h dialysis session. In a study by Flaminio et al.
(1988) using this mass-balance approach, on the order of
46 mg DEHP was estimated to be leached during a 4-h
dialysis session. The 12 patients studied had experienced
prior dialysis treatment over periods of 4 months to 12
348
DOULL ET AL.
years. Gibson et al. (1976) followed 5 patients through the

length of 5-h hemodialysis sessions and measured DEHP
exposures that ranged from 9 to 150 mg, with an arithmetic mean of 80.2 mg and a standard deviation of 66.3
mg. Lewis et al. (1978) conducted an in vitro perfusion
study by circulating blood through a dialysis unit, leaching 6.1 mg DEHP into the blood from tubing, bags, and
the dialysis unit itself over a 1-h period. Assuming leaching of DEHP to be linear over time, as shown by Jaeger
and Rubin (1972) for PVC blood bags, this translates to
an exposure of 30.5 mg over a 5-h dialysis period.
Pollack et al. (1985a) conducted an inflow/outflow
study on 11 patients undergoing maintenance dialysis
for treatment of renal failure. The highest degree of
patient variability was observed in the Pollack et al.
(1985b) study. DEHP exposures ranging from 23.8 to
360 mg DEHP were measured for the 4-h dialysis sessions, with a mean of 105 mg and a standard deviation
of 113 mg. 3 There was no statistically significant correlation between the amount of DEHP exposure incurred and the number of years of prior dialysis (patients ranged from 0.02 to 12 years of prior dialysis).
Considering that these four studies span a period of 17
years and four different research teams, the degree of
agreement between the studies is extremely good. For the
purpose of obtaining a central value, a mean of the study
means (weighted by number of patients) yields a value of
74 mg. Based on this value and the assumptions noted in
Table 11, the LADD was calculated to be 12 mg/person/
day for short-term dialysis patients and 6900 mg/person/
day for long-term dialysis patients.
The estimation of DEHP exposures associated with
hemodialysis can be arrived at using the weighted
mean exposure of 74 mg per hemodialysis session; the
exposures over each year of treatment (E T) and over a
lifetime (LADD) for short-term (2 months) and longterm hemodialysis patients (15 years), assuming a life
span of 70 years, are calculated as
ET 5
~365 days/year!
3 ~No. years treatment!
LADD 5
.
70 years
Thus, for short-term hemodialysis patients, the use
period exposure (E T) and LADD are
3
Most of the DEHP exposure values from the Pollack et al. (1985) study
clustered in the 23.8 to 98.9 mg range; if the two extreme high values in
the data set (289 and 360 mg) are excluded as statistical outliers, a
mean of 56.6 mg is obtained, with a standard deviation of 33 mg.
74 mg DEHP/treatment
ET 5
~365 days/year!
5 4900 mg/person/day
LADD 5 ~4900 mg/person/day!
3 ~0.167 years/70 years!
5 12 mg/person/day
and for long-term dialysis patients
74 mg DEHP/treatment
ET 5
~365 days/year!
5 32,000 mg/person/day
LADD 5 ~32,000 mg/person/day! 3 ~15 years/70 years!
5 6900 mg/person/day.
AUTOPHERESIS SETS
Autopheresis involves the donation of blood, during

which the donors red blood cells are returned (thus,
the donor effectively contributes plasma). No data are
available on the amount of DEHP leached into blood
and returned to the donor during autopheresis. If it is
assumed that the available data on hemodialysis
bound the possible exposures for autopheresis and that
the leaching is linear over brief time periods as shown
by Jaeger and Rubin (1972) for PVC blood bags, the
weighted mean DEHP exposure of 74 mg for hemodialysis can be scaled downward based on the shorter
contact time: 74 mg 3 (1 h)/(5 h) 5 14.8 mg. Based on
this value and the assumptions noted in Table 11, a
LADD of 1.2 mg/person/day was obtained for the shortterm donor and a LADD of 35 mg/person/day was obtained for the long-term donor.
The estimation of DEHP exposures associated with
autopheresis was arrived at using the scaled mean
exposure of 14.8 mg per autopheresis session; the exposures over each year of treatment (E T) and over a
lifetime (LADD) for a short-term donor (once/month for
2 months) and long-term donor (once/month for 5 years),
assuming a life span of 70 years, are calculated as
~ET! 5
Use period ~days/year!
3 ~No. years treatment!
.
LADD 5
70 years
349
Thus, for the short-term donor, the use period exposure

(E T) and LADD are
14.8 mg DEHP/treatment
ET 5
~61 days/year!
5 485 mg/person/day
LADD 5 ~485 mg/person/day! 3 0.167 years/70 years
5 1.2 mg/person/day.
For the long-term donor, the use period exposure (E T)
and LADD are
14.8 mg DEHP/treatment
ET 5
~365 days/year!
the density of water, i.e., 1 g/cm 3, is assumed), and

assuming a linear leaching rate over time, the oral dose
of DEHP over 2 days of product use would be (0.25
mg/day) 3 (1500 ml)/(100 ml) 3 2 days 5 7.5 mg/day.
In a 28-g device that is 15% DEHP; this represents
leaching of (7.5 mg/day)/(28,000 mg 3 0.15) 3 100% 5
0.18%/day of the DEHP in the product over the assumed 2-day replacement period. For this assessment,
we have rounded this number to 0.2%/day. Using the
above assumptions, the calculated LADD for shortterm feeding tube use is 2.0 mg/person/day, and the
calculated LADD for long-term use is 60 mg/person/
day.
These calculations were obtained as follows.
The average daily exposure during the use period
(E T) is
ET 5
0.002 3 28,000 mg 3 0.15 3 ~1000 mg/mg!

2 days
5 487 mg/person/day
LADD 5 ~487 mg/person/day! 3 ~5 years!/~70 years!
5 35 mg/person/day.
NASOGASTRIC FEEDING TUBES
Nasogastric feeding tubes are sold in a variety of

diameters and lengths, ranging from 5 French gauge
16 in. for use in infants to 14 18 French gauge 9 in. for
adults. The tubing portion of the nasogastric feeding
tubes ranges in weight from approximately 0.1 to 1 oz.
PVC-containing feeding tubes are used primarily for
short-term applications ranging from 1 or several days
to less than 2 weeks at an extreme. If a patient is
anticipated to need a feeding tube for several weeks to
a month, often a polyurethane feeding tube (which
contains no significant levels of DEHP) is placed into
the patient. If a patient is anticipated to require a
feeding tube for more than 1 month, often a silicone
feeding tube (which also does not contain significant
levels of DEHP) is used.
No data are available on the degree to which DEHP
is leached from a nasogastric feeding tube. For the
purpose of this assessment, it is assumed that a feeding tube is approximately 1 oz. (28 g) in weight, that it
contains 15% DEHP by weight, and that a certain
portion of the DEHP contained in the tubing leaches
out and becomes available to the patient over the approximate 2-day replacement period. The percentage of
DEHP leaching over 2 days can be conservatively
bounded by the use of data on leaching of DEHP from
blood bags into blood products (a more lipophilic matrix
than typical nutrient homogenate) per the data of Jaeger and Rubin (1972). Using the leaching rate of 0.25
mg DEHP/100 ml/day from Jaeger and Rubin (1972),
assuming a food intake of 1.5 kg/day (1.5 liters/day if
5 4200 mg/day.
The LADD for a trauma patient using the device for 12
days is
LADD 5
4200 mg/day 3 ~12 days! 3 ~year/365 days!

70 years
5 2.0 mg/person/day.
It is likely that no long-term use of PVC-containing

nasogastric feeding tubes occurs, but that other types
of tubing are used for long-term patients (e.g., polyurethane or silicone tubing). However, if it is assumed that
a long-term patient (e.g., in the case of an esophageal
cancer patient) utilizes PVC feeding tubes for 1 year,
and assuming a life span of 70 years, the LADD would
be
LADD 5
4200 mg/day 3 ~1 year!

5 60 mg/person/day.
70 years
Summary of Human Exposures to DEHP

As might be anticipated, maximum human exposures to DEHP are those encountered occupationally
where, based on currently accepted workplace standards (U.S. OSHA and ACGIH and German MAK),
workers may be exposed to 700 1400 mg/kg body wt/
day mainly by inhalation. When calculated as a LADD
based on an anticipated 40-year working life and a
70-year life span, this is equivalent to 300 600 mg/kg
body wt/day. As discussed earlier, however, calculations of occupational exposure based on the OSHA PEL
for DEHP are clearly overestimates since actual worker
exposures are undoubtedly 10- to 100-fold lower.
In general, exposures of the general population to
350
DOULL ET AL.
DEHP are much lower than those of workers, the major source of exposure, food, leading to oral intakes in
the range of 330 mg/kg body wt/day. Ingestion of water
and inhalation of ambient air constitute negligible
sources of exposure to DEHP.
In contrast, nonoccupational exposures from medical
devices are significant and in some cases may approach
those of workers. Of several devices evaluated, maximum exposures are encountered by long-term hemodialysis patients. These individuals receive up to about
457 mg/kg body wt/day of DEHP over the course of a
typical 1-year treatment period and this equates to a
LADD of about 99 mg/kg body wt/day when estimated
on the basis of a possible 15-year dialysis period in a
70-year life span. Hemodialysis patients constitute a
relatively small patient subpopulation and more common exposures from medical devices result, for example, from the use of transfusion equipment (iv sets and
blood bags). Even with fairly frequent use of transfusions, however (e.g., 70 treatments/year for leukemia
patients), the exposures are much lower and do not
exceed about 36 mg/kg body wt/day (LADD of about 1
mg/kg body wt/day for a 2-year transfusion period in a
70 year life span).
In considering the potential health relevance of human exposures to DEHP it is important to take into
account the route as well as the level of exposure. As
discussed in the pharmacokinetics section of this article, the enzymatic breakdown of DEHP to MEHP is of
critical importance because, in rodents, MEHP and
several of its metabolites are the active peroxisome
proliferators; DEHP is not, itself, an active peroxisome
proliferator. Since the conversion of DEHP to MEHP
occurs almost exclusively in the gastrointestinal tract
(the lipase responsible is secreted into the small intestine from the pancreas), the activation of DEHP is
limited mainly to exposures via the oral route. Only
low titers of DEHP-hydrolyzing enzymes are present in
the liver and other tissues (Albro, 1986; Huber et al.,
1996) so that DEHP entering the systemic circulation
by the inhalation, dermal, or intravenous routes avoids
contact with the lipase present in the gut and is not
converted to an active peroxisome proliferator. Clearly,
this has important toxicokinetic consequences because,
even if humans were sensitive to MEHP (which they
are not), circulating levels of MEHP and other active
metabolites would be extremely low in humans undergoing transfusion or dialysis procedures (Rock et al.,
1978; Chu et al., 1978). The only possible exception to
this is if, as is apparently the case under certain conditions, DEHP is converted to MEHP to any significant
extent during storage of blood or blood products.
MOE Approach
MOE values for the various exposure scenarios discussed and based on the mouse NOEL for peroxisome
TABLE 12
MOEs for Selected Routes of Human Exposure
Exposure population/route
General population/oral
Daily ingestion in food
Occupational/inhalation
OSHA, ACGIH (PEL, TLV)
LADD (40-year work-life)
Medical devices/iv and other
Transfusion
Intensive (12 years)
LADD (2 years/lifetime)
Hemodialysis
Over 1 year
LADD (15 years/lifetime)
Level of exposure a
(mg/kg body wt/day)
MOE b
Up to 30
667
700
280
29
71
36
1
556
20,000
457
99
44
202
The exposure values given represent worst-case scenarios. Actual exposures are likely to be very much lower than these values.
b
The MOE (margin of exposure) values are obtained with reference to the mouse NOEL of 20 mg/kg/day for peroxisome proliferation.
proliferation (20.0 mg/kg body wt/day) are shown in

Table 12. This indicates that the lowest MOE of 29 is
associated with the OSHA and ACGIH occupational
standards (PEL and TLV 5 5 mg/m 3); the occupational
LADD is associated with a MOE of 71. The lowest
nonoccupational MOE (44) was that for hemodialysis
patients while the MOE for the general population,
based on DEHP exposure to food residues, was almost
670. The Proposed Guidelines for Carcinogen Risk Assessment address the issue of what determines an acceptable MOE for regulatory purposes and several criteria are identified that a risk manager should consider
in reaching a decision regarding the magnitude of an
acceptable MOE. The decision-making process is
termed a MOE analysis. Important points to be considered in the analysis include
1. The slope of the observed doseresponse relationship;
2. The nature of the response used for the dose
response assessment;
3. Human sensitivity to the phenomenon compared
with experimental animals;
4. The nature and extent of human variability in
sensitivity to the phenomenon involved;
5. The persistence of the agent in the body; and
6. Other factors.
In the absence of data showing otherwise, a commonly used default MOE of 100 is often considered
acceptable. This is based on a 103 uncertainty factor to
account for the possibility that humans are more sensitive than the test species to the biochemical/toxicological end point used and an additional 103 factor to
account for possible differences in sensitivity within
the human population. The proposed guidelines specify
that if information concerning human variability or

interspecies differences is available, it may be used
instead of the default or to modify it as appropriate.
To determine acceptable human exposures to DEHP,
a MOE analysis was conducted taking into account the
several criteria listed above.
Slope of the doseresponse relationship. The biochemical end point on which it is proposed to base
DEHP risk assessment (hepatic peroxisome proliferation) exhibits clearly defined rodent NOEL and LOEL
values indicative of a fairly steep doseresponse relationship. Consequently, as the dose is further reduced
below the measured NOEL (the point of departure),
there is an even greater confidence in the probability
that peroxisome proliferation will not occur. The facts
that the point of departure is the NOEL (rather than
the ED 10 or LED 10) and that the doseresponse relationship is relatively steep support the acceptance of a
lower MOE.
Nature of response used for doseresponse assessment. The response used for the doseresponse assessment, hepatic peroxisome proliferation in rodents,
is a reversible precursor effect that is not per se an
adverse effect. There is strong experimental evidence
and general scientific consensus that peroxisome proliferation is an early, sensitive biochemical marker of
potential liver cancer in rodents. Importantly, peroxisome proliferation, increased liver weight, and hepatic
foci and tumors are reversible when the peroxisome
proliferator is removed. All anticipated human exposures to DEHP are intermittent in nature and would
not be expected to cause cumulative effects. The NOEL
for peroxisome proliferation is about fivefold lower
than the NOEL for the tumorigenic response. The nature of the response used supports the acceptance of a
lower MOE.
Human sensitivity to the phenomenon relative to experimental animals. As discussed throughout this article, there is strong scientific support for the fact that,
compared with rodents, humans are remarkably nonresponsive both to DEHP-mediated peroxisome proliferation and to the resulting hepatocarcinogenic response. Primary cultures of human hepatocytes treated
in vitro with high concentrations of potent rodent peroxisome proliferators fail to exhibit a peroxisome proliferation or mitogenic response. Furthermore, epidemiological studies with humans exposed for months or
even years to high daily doses (up to about 2 g) of
hypolipidemic drugs that, in rodents, are more potent
peroxisome proliferators than DEHP show no evidence
of adverse liver effects and no increase in any form of
cancer.
As previously discussed, a default factor of 10 is
typically incorporated into the MOE for interspecies
extrapolation based on the assumption that humans
351
are more sensitive to the adverse effects of the chemical than the test animals. This assumption is clearly
quite contrary to all available data on the action of
DEHP as a liver carcinogen. Consequently, it is concluded that the acceptable MOE for DEHP can justifiably be reduced by a factor of 10 from the usual default
value.
Human variability in sensitivity to the phenomenon
involved. There is no evidence to suggest any significant variation in sensitivity to DEHP or other peroxisome proliferators within the human population (ATSDR,
1997). As indicated above, humans are extremely nonresponsive to the biochemical, morphological, and
other effects of DEHP and other peroxisome proliferators. It seems likely that, if they exist, some evidence of
genetic polymorphism or ethnic variation would have
been uncovered in the clinical and epidemiological
studies that have been conducted with hypolipidemic
drugs; to date, this has not been the case. While this
does not provide unequivocal proof that such polymorphic variation does not exist, it suggests a low probability of occurrence and an insignificant difference in
phenotype expression. It must be concluded that currently available data are inadequate to justify moving
away from the default factor of 10 associated with
intraspecies variation.
Persistence of DEHP in the body. DEHP is rapidly
metabolized and excreted from the body (plasma halflife for DEHP is 10 18 h, while that for MEHP is 3 6
h) and there is no evidence of persistence or tissue
accumulation following repeated daily exposures.
Other factors. Because of the critical importance of
the route of exposure in the metabolic activation of
DEHP, it is concluded that, even though relatively
high, human occupational inhalation exposures and
medical intravenous exposures have little or no relevance to any potential effects that might be mediated
through peroxisome proliferation. Exposure by any
route other than ingestion will not be expected to have
an effect because it bypasses the gastrointestinal conversion of DEHP to MEHP that is required to generate
an active peroxisome proliferator.
This clearly indicates that the only relevant DEHP
exposures likely to affect the general human population are residues in food. As discussed, these are small,
about 30 mg/kg body wt/day, and are associated with a
MOE of almost 700.
Based on the total weight-of-evidence available, it is
concluded that 10 constitutes an acceptable and highly
conservative MOE for purposes of assessing the potential human risks associated with exposure to DEHP.
This conclusion has been reached by procedures that
are consonant with those outlined in EPAs Proposed
Guidelines for Carcinogen Risk Assessment.
352
DOULL ET AL.
CONCLUSIONS
1. DEHP should be classified as unlikely to be a

human carcinogen under any known conditions of human exposure.
2. DEHP is a hepatocarcinogen in F344 rats and
B6C3F1 mice. There is clear evidence that DEHP exerts its tumorigenic response through the process of
peroxisome proliferation, probably through increased
cell proliferation and hepatocellular hyperplasia. The
most recent studies clearly indicate a nonlinear dose
response curve below which tumors are not induced.
3. The total weight-of-evidence from mutagenicity
studies clearly indicates that DEHP is not genotoxic.
This conclusion is based on negative findings in a wide
range of test systems to determine in vitro bacterial,
yeast, insect, and mammalian cell mutagenesis, DNA
interactions, and chromosomal aberrations both in
vitro and in vivo.
4. Peroxisome proliferation is a sensitive early biochemical marker for DEHP-induced hepatocarcinogenesis in rodents. It exhibits a clear dose threshold (NOEL
of 20 mg/kg/day in mice) that is about fivefold lower
than the NOEL for a tumorigenic response.
5. Peroxisome proliferation is a reversible, receptormediated process that is maintained only in the continued presence of the causative agent. When DEHP is
removed from rodent diets there is a regression in
biochemical and morphological changes associated
with peroxisome proliferation and in the number of
adenomas seen.
6. Rats and mice are extremely sensitive to DEHP
and other peroxisome proliferators, due to their expression of the PPARa, while nonhuman primates and
humans are almost completely nonresponsive or refractory. Therefore, rats and mice are inappropriate for
human risk assessment purposes.
7. In view of the lack of genotoxicity and the clearly
nonlinear tumorigenic dose response in rodents, it is
inappropriate to conduct human risk assessment for
DEHP using a linear multistage model. Instead, human risk assessment should be conducted using the
MOE approach outlined in EPAs Proposed Guidelines
for Carcinogen Risk Assessment. MOEs should be calculated in relation to the NOEL of 20 mg/kg/day for
peroxisome proliferation in B6C3F1 mice.
8. Exposure of the general human population to
DEHP is approximately 30 mg/kg body wt/day, the major source being from residues in food. Higher exposures occur occupationally (up to about 700 mg/kg body
wt/day, mainly inhalation) and through use of certain
medical devices (e.g., up to 457 mg/kg body wt/day for
hemodialysis patients, intravenous), although these
have little relevance because the routes of exposure
bypass the critical activation enzymes in the gastrointestinal tract.
9. The MOE for exposures of the general population
is almost 700. This exceeds by a factor of 70 an acceptable MOE obtained by a MOE analysis conducted according to the Proposed Guidelines for Carcinogen Risk
Assessment.
REFERENCES
Albro, P. W., and Levenhar, S. R. (1989). Metabolism of di(2-ethylhexyl)phthalate. Drug Metab. Rev. 21(1), 1334.
Albro, P. W. (1986). Absorption, metabolism and excretion of di(2ethylhexyl) phthalate by rats and mice. Environ. Health Perspect.
65, 293298.
Albro, P. W., Hass, J. R., Peck, C. C., et al. (1979). Identification of
the metabolites of di(2-ethylhexyl)phthalate in urine from the
African green monkey. Drug Metab. Dispos. 9(3), 223225.
Albro, P. W., Thomas, R., and Fishbein, L. (1973). Metabolism of
di-ethylhexylphthalate by rats. Isolation and characterization of
the urinary metabolites. J. Chromatogr. 76, 321330.
Albro, P. W., Corbett, J. T., Schroeder, J. L., Jordan, S., and Matthews, H. B. (1982). Pharmacokinetics, interactions with macromolecules and species differences in metabolism of DEHP. Environ. Health Perspect. 45, 19 25.
Allen, K. L., Green, C. E., and Tyson, C. A. (1987). Comparative
studies on peroxisomal enzyme induction in hepatocytes from rat,
Cynomolgus monkey and human by hypolipidemic drugs. Toxicologist 7, Abstract No. 253.
Ashby, J., de Serres, F. J., Draper, M., et al. (1985). Evaluation of
Short-Term Tests for Carcinogen, pp. 1752. Elsevier Science Publishers, Amsterdam.
Ashby, J., Brady, A., Elcombe, C. R., et al. (1994). Mechanisticallybased human hazard assessment of peroxisome proliferator-induced hepatocarcinogenesis. Hum. Exp. Toxicol. 13(Suppl. 2), S1
S117.
Auboeuf, D., Rieusset, J., Fajas, L., Vallier, P., Frering, V., Riou, J.,
Staels, B., Auwerx, J., Laville, M., and Vidal, H. (1997). Tissue
distribution and quantification of the expression of mRNAs of
peroxisome proliferator-activated receptors and liver X receptor-a
in humans. Diabetes 46, 1319 1327.
Bell, A. R., Savory, R., Horely, N. J., Choudhury, A. I., Dickins, M.,
Gray, T. J. B., Salter, A. M., and Bell, D. R. (1998). Molecular basis
of non-responsiveness to peroxisome proliferators: The guinea pig
PPARa is functional and mediates peroxisome proliferator-induced hypolipidaemia. Biochem. J. 332, 689 693.
Benford, D. J., Patel, S., Reavy, H. J., Mitchell, A., and Sarginson,
N. J. (1986). Species differences in the response of cultured hepatocytes to phthalate esters. Food Chem. Toxicol. 24, 799 800.
Bentley, P., Calder, I., Elcombe, C., Grasso, P., Stringer, D., and
Wiegand, H.-J. (1993). Hepatic peroxisome proliferation in rodents
and its significance for humans. Food Chem. Toxicol. 31, 857907.
Bichet, N., Cahard, D., Fabre, G., Remandet, B., Gouy, D., and Cano,
J.-P. (1990). Toxicological studies on a benzofuran derivative. III.
Comparison of peroxisome proliferation in rat and human hepatocytes in primary culture. Toxicol. Appl. Pharmacol. 106, 509 517.
Bieri, F. (1993). Cultured hepatocytes: A useful in vitro system to
investigate effects induced by peroxisome proliferators and their
species specificity. In Peroxisomes: Biology and Importance in Toxicology and Medicine (G. Gibson and B. Lake, Eds.), pp. 299 311.
Taylor and Francis, London.
Bieri, F., Muakkassah-Kelly, S., Waechter, F., Sagelsdorff, P.,
Staubli, W., and Bentley, P. (1990). The significance of in vitro
studies on peroxisome proliferation. Toxicol. In Vitro 4, 428 431.
Bieri, F., Bentley, P., Waechter, F., and Staubli, W. (1984). Use of
primary cultures of adult rat hepatocytes to investigate mechanisms of action of nafenopin, a hepatocarcinogenic peroxisome
proliferator. Carcinogenesis 5, 10331039.

Bieri, F., Staubli, W., Waechter, F., Muakkassah-Kelly, S., and Bentley, P. (1988). Stimulation of DNA synthesis but not of peroxisomal
b-oxidation by nafenopin in primary cultures of marmoset hepatocytes. Cell Biol. Int. Rep. 12, 10771087.
Blaauboer, B. J., van Holsteijn, C. W. M., Bleumink, R., Mennes,
W. C., van Pelt, F. N. A. M., Yap, S. H., van Pelt, J. F., van Iersel,
A. A. J., Timmerman, A., and Schmid, B. P. (1990). The effect of
beclobric acid and clofibric acid on peroxisomal b-oxidation and
peroxisome proliferation in primary cultures of rat, monkey and
human hepatocytes. Biochem. Pharmacol. 40, 521528.
tudes de toxicoloBlane, G. F., and Pinaroli, F. (1980). Fenofibrate: E
gie animal en rapport avec les effets secondaires chez les malades.
Presse Med. 9, 37373746.
Blumcke, S., Schwartzkopff, W., Lobeck, H., et al. (1983). Influence of
fenofibrate on cellular and subcellular liver structure in hyperlipidemic patients. Atherosclerosis 46, 105116.
Bojes, H. K., and Thurman, R. G. (1996). Peroxisome proliferators
activate Kupffer cells in vivo. Cancer Res. 56, 1 4.
Bouis, P., Bieri, F., Lang, B., Thomas, H., and Waechter, F. (1993).
Effect of propaquizafop and its free-acid derivative on lauric acid
hydroxylation and peroxisomal b-oxidation in primary cultured
rat, mouse, guinea pig and marmoset hepatocytes. Toxicol. In Vitro
7, 427 431.
Braissant, O., Foufelle, F., Scotto, C., Dauca, M., and Wahli, W.
(1996). Differential expression of peroxisome proliferator-activated receptors (PPARs): Tissue distribution of PPARa, -b, and -w
in the adult rat. Endocrinology 137, 354 366.
Budroe, J. D., Umemura, T., Angeloff, K., and Williams, G. M. (1992).
Doseresponse relationships of hepatic acyl-CoA oxidase and catalase activity and liver mitogenesis induced by the peroxisome
proliferator ciprofibrate in C57BL/6N and BALB/c mice. Toxicol.
Appl. Pharmacol. 113, 192198.
Budroe, J. D., and Williams, G. M. (1993). Genotoxicity studies of
peroxisome proliferators. In Peroxisome: Biology and Importance
in Toxicology and Medicine (G. G. Gibson and B. G. Lake, Eds.), pp.
525568. Taylor and Francis, London.
Butler, E. G., England, P. J., and Williams, G. M. (1988). Genetic
differences in enzymes associated with peroxisome proliferation
and hydrogen peroxide metabolism in inbred mouse strains. Carcinogenesis 9, 1459 1463.
Butterworth, B. E., Smith-Oliver, T., Earle, L., Loury, D. J., White,
R. D., Doolittle, D. J., Working, P. K., Cattley, R. C., Jirtle, R.,
Michalopoulos, G., and Strom, S. (1989). Use of primary cultures of
human hepatocytes in toxicology studies. Cancer Res. 49, 1075
1084.
Carpenter, C. P., Weil, C. S., and Smith, H. F., Jr. (1953). Chronic
oral toxicity of di(2-ethylhexyl)phthalate for rats, guinea pigs and
dogs. Am. Med. Assoc. Arch. Ind. Hyg. Occup. Med. 8, 219 226.
Cattley, R. C., Conway, J. G., and Popp, J. A. (1987). Association of
persistent peroxisome proliferation and oxidative injury with
hepatocarcinogenicity in female F-344 rats fed di(2-ethylhexyl)phthalate for two years. Cancer Lett. 38, 1522.
Cattley, R. C., and Glover, S. E. (1993). Elevated 8-hydroxydeoxyguanosine in hepatic DNA of rats following exposure to peroxisome
proliferators: Relationship to carcinogenesis and nuclear localization. Carcinogenesis 14(12), 24952499.
Cattley, R., DeLuca, J., Elcombe, C., Fenner-Crisp, P., Lake, B.,
Marsman, D., Pastoor, T., Popp, J. A., Robinsoh, D. E., Schwetz, B.,
Tugwood, J., and Wahli, W. (1998). Do peroxisome proliferating
compounds pose a hepatocarcinogenic hazard to humans? Regul.
Toxicol. Pharmacol. 27, 47 60.
Chu, I., Villeneuve, D. C., Secours, V., et al. (1978). Metabolism and
tissue distribution of mono-2-ethylhexyl phthalate in the rat. Drug
Metab. Dispos. 6, 146 149.
353
Cornu, M. C., Lhuguenot, J. C., Brady, A. M., Moore, R., and Elcombe, C. R. (1992). Identification of the proximate peroxisome
proliferator(s) derived from di(2-ethylhexyl)adipate and species
differences in response. Biochem. Pharmacol. 43, 2129 2134.
Cornu-Chagnon, M.-C., Dupont, H., and Edgar, A. (1995). Fenofibrate: Metabolism and species differences for peroxisome proliferation in cultured hepatocytes. Fundam. Appl. Toxicol. 26, 6374.
David, R. M., Moore, M. R., Cifone, M. A., Finney, D. C., and Guest,
D. (1996). Further investigations of the oncogenicity of di(2-ethylhexyl) phthalate in rats. Fundam. Appl. Toxicol. 30(Suppl.), 204.
David, R. M., Moore, M. R., Cifone, M. A., Finney, D. C., and Guest,
D. (1997). Correlation of peroxisome proliferation and oncogenicity
of di(2-ethylhexyl) phthalate in mice. Fundam. Appl. Toxicol.
36(Suppl.), 173.
De La Iglesia, F. A., Lewis, J. E., Buchanan, R. A., et al. (1982). Light
and electron microscopy of liver in hyperlipoproteinemic patients
under long-term gemfibrozil treatment. Atherosclerosis 43, 19 37.
Dirven, H. A. A. M., Van Den Broek, P. H. H., Peeters, M. C. E.,
Peters, J. G. P., Mennes, W. C., Blaauboer, B. J., Noordhoek, J.,
and Jongeneelen, F. J. (1993). Effects of the peroxisome proliferator mono(2-ethylhexyl)phthalate in primary hepatocyte cultures
derived from rat, guinea pig, rabbit and monkey. Relationship
between interspecies differences in biotransformation and peroxisome proliferating potencies. Biochem. Pharmacol. 45, 24252434.
Dybing, E., Mikalsen, S-O., Huttunen, J. K., and Sanner, T. (1995).
Peroxisome proliferation genotoxicity and carcinogenicity. In Peroxisome Proliferation and its Role in Carcinogenesis. IARC Technical Report 24, p. 64. Lyon, France.
Eacho, P. I., Foxworthy, P. S., Johnson. W. D., Hoover, D. M., and
White, S. L. (1986). Hepatic peroxisomal changes induced by a
tetrazole-substituted alkoxyacetophenone in rats and comparison
with other species. Toxicol. Appl. Pharmacol. 83, 430 437.
Easterling, R. E., Johnson, E., and Napier, E. A. (1974). Plasma
extraction of plasticizers from medical grade polyvinylchloride
tubing. Proc. Soc. Exp. Biol. Med. 147, 572574.
Elcombe, C. R. (1985). Species differences in carcinogenicity and
peroxisome proliferation due to trichloroethylene: A biochemical
human hazard assessment. Arch. Toxicol. Suppl. 8, 6 17.
Elcombe, C. R., and Mitchell, A. M. (1986). Peroxisome proliferation
due to di(2-ethylhexyl)phthalate (DEHP): Species differences and
possible mechanisms. Environ. Health Perspect. 70, 211219.
Elcombe, C. R., and Styles, J. A. (1989). Species differences in peroxisome proliferation. Toxicologist 9, 63.
Elcombe, C. R., Bell, D. R., Elias, E., Hasmall, S. C., and Plant, N. J.
(1996). Peroxisome proliferators: Species differences in response of
primary hepatocyte cultures. Ann. N.Y. Acad. Sci. 804, 628 635.
EPA (U.S. Environmental Protection Agency) (1987). Health Effects
Assessment for Selected Phthalic Acid Esters. EPA-600/8-88-053.
NTIS PB88-178934, Washington, D.C.
EPA (U.S. Environmental Protection Agency) (1989). Ambient Water
Criteria Document Addendum for Phthalate Esters, Final Draft,
PB91-161653. Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office, Cincinnati, OH.
EPA (U.S. Environmental Protection Agency) (1996). Proposed
Guidelines for Carcinogen Risk Assessment. Office of Research
and Development, U.S. Environmental Protection Agency. Fed.
Reg. 61(79), 17960 18011.
Fan, C. Y., Pan, J., Usuda, N., Yeidaudi, A. V., Rao, M. S., and Reddy,
J. K. (1998). Steatohepatitis, spontaneous peroxisome proliferation and liver tumors in mice lacking peroxisomal fatty acyl-CoAoxidase. J. Biol. Chem. 273, 15639 15645.
Flaminio, L. M., Bergia, R., De Angelis, L., Ferazza, M., Marinovich,
M., Galli, G., and Galli, C. L. (1988). The fate of leached di-(2ethylhexyl)-phthalate (DEHP) in patients on chronic hemodialysis. Int. J. Artif. Organs 11(6), 428 435.
354
DOULL ET AL.
Foxworthy, P. S., White, S. L., Hoover, D. M., and Eacho, P. I. (1990).

Effect of ciprofibrate, bezafibrate, and LY 171883 on peroxisomal
b-oxidation in cultured rat, dog and Rhesus monkey hepatocytes.
Toxicol. Appl. Pharmacol. 104, 386 394.
Foxworthy, P. S., and Eacho, P. I. (1994). Cultured hepatocytes for
studies of peroxisome proliferation: Methods and applications.
J. Pharmacol. Toxicol. Methods 31, 2130.
Frick, H., Elo, O., Haapa, K., et al. (1987). Helsinki Heart Study:
Primary-prevention trial with gemfibrozil in middle aged men
with dyslipidemia, safety of treatment, changes in risk factors and
incidence of coronary heart disease. N. Engl. J. Med. 317, 1235
1247.
Furukawa, K., Numoto, S., Furuya, K., Furukawa, N. T., and Williams, G. M. (1985). Effect of hepatocarcinogen nafenopin, a peroxisome proliferator, on the activities of rat liver glutathionerequiring enzymes and catalase in comparison to the action of
phenobarbital. Cancer Res. 45, 50115019.
Ganning, A. E. (1987). Effects of Di(2-ethylhexyl)phthalate on the
Structure and Function of the Rat Hepatocyte. Dissertation thesis,
Univ. Stockholm, Sweden.
Ganning, A. E., Brunk, U., and Dallner, G. (1984). Phthalate esters
and their effect on the liver. Hepatology 4, 541547.
Ganning, A. E., Brunk, U., Edlund, C., et al. (1987). Effects of
prolonged administration of phthalate ester on the liver. Environ.
Health Perspect. 73, 251258.
Gariot, P., Barrat, E., Drouin, P., Genton, P., et al. (1987). Morphometric study of human hepatic cell modifications induced by fenofibrate. Metabolism 36, 203210.
Garvey, L. K., Swenberg, J. A., Hamm, T. E., Jr., and Popp, J. A.
(1987). Di(2-ethylhexyl)phthalate: Lack of initiating activity in the
liver of female F-344 rats. Carcinogenesis 8, 285290.
Gibson, T. P., Briggs, W. A., and Boone, B. J. (1976). Delivery of
di-2-ethylhexyl phthalate to patients during hemodialysis. J. Lab.
Clin. Med. 87(3), 519 524.
Goodman, D. G., Ward, J. M., Squire, R. A., and Chu, K. C. (1979).
Neoplastic and nonneoplastic lesions in aging F344 rats. Toxicol.
Appl. Pharmacol. 48, 237248.
Graham, M. J., Wilson, S. A., Winham, M. A., et al. (1994). Lack of
peroxisome proliferation in marmoset liver following treatment
with ciprofibrate for 3 years. Fundam. Appl. Toxicol. 22, 58 64.
Gray, T. J. B., Lake, B. G., Beamand, J. A., Foster, J. R., and
Gangolli, S. D. (1983). Peroxisome proliferation in primary cultures of rat hepatocytes. Toxicol. Appl. Pharmacol. 67, 1525.
Green, S. (1995). PPAR: A mediator of peroxisome proliferator action. Mutat. Res. 333, 101109.
Hanefeld, M., Kemmer, C., Leonhardt, W., et al. (1980). Effects of
p-chlorophenoxy-isobutyric acid (CPIB) on the human liver. Atherosclerosis 36, 159 172.
Hanefeld, M., Kemper, C., and Kadner, E. (1983). Relationship between morphological changes and lipid-lowering action of p-chlorophenoxyisobutyric acid (CPIB) on hepatic mitochondria and peroxisomes in man. Atherosclerosis 46, 239 246.
Harris, R. S., Hodge, H. C., Maynard, E. A., and Blanchet, H. J. J.
(1956). Chronic oral toxicity of di(2-ethylhexyl)phthalate in rats
and dogs. Arch. Ind. Health 13, 259 264.
Holloway, B. R., Thorp, J. M., Smith, G. D., and Peters, T. J. (1982).
Analytical subcellular fractionation and enzymatic analysis of
liver homogenates from control and clofibrate-treated rats, mice,
and monkeys with reference to the fatty acid oxidizing enzymes.
Ann. N.Y. Acad. Sci. 386, 453 455.
Huber, W. W., Grasl-Kraupp, B., and Schulte-Hermann, R. (1996).
Hepatocarcinogenic potential of di(2-ethylhexyl)phthalate in rodents and its implications on human risk. Crit. Rev. Toxicol. 26(4),
365 481.
Hsu, M.-H., Palmer, C. A. N., Griffin, K. J., and Johnson, E. F.

(1995). A single amino acid change in the mouse peroxisome proliferator-activated receptor a alters transcriptional responses to
peroxisome proliferators. Mol. Pharmacol. 48, 559 567.
IARC (International Agency for Research in Cancer) (1995). Peroxisome Proliferation and Its Role in Carcinogenesis, IARC Technical
Report No. 24, pp. 3 85. World Health Organization, Geneva.
IARC (International Agency for Research in Cancer) (1982). Di(2ethylhexyl)phthalate. In Some Industrial Chemicals and Dyestuffs, IARC Technical Report No. 29, pp. 269 294.
IARC (International Agency for Research in Cancer) (1996). Hypolipidaemic Drugs, Vol. 66, pp. 389 485. Lyon.
International Programme on Chemical Safety (IPCS) (1992). Environmental Health Criteria Document No. 131, Diethylhexyl Phthalate. World Health Organization, Geneva.
Jacobson, M. S., Kevy, S. V., and Grand, R. J. (1977). Effects of a
plasticizer leached from polyvinyl chloride on the subhuman primate: A consequence of chronic transfusion therapy. J. Lab. Clin.
Med. 89(5), 1066 1079.
Jaeger, R. J., and Rubin, R. J. (1972). Migration of a phthalate ester
plasticizer from polyvinyl chloride blood bags into stored human
blood and its localization in human tissues. N. Engl. J. Med.
287(22), 1114 1118.
James, N. H., and Roberts, R. A. (1994). The peroxisome proliferator
class of non-genotoxic hepatocarcinogens synergizes with epidermal growth factor to promote clonal expansion of initiated rat
hepatocytes. Carcinogenesis 15, 26872694.
James, N. H., and Roberts, R. A. (1995). Species differences in the
clonal expansion of hepatocytes in response to the coaction of
epidermal growth factor and nafenopin, a rodent hepatocarcinogenic peroxisome proliferator. Fundam. Appl. Toxicol. 26, 43149.
James, N. H., and Roberts, R. A. (1996). Specific differences in
response to peroxisome proliferators correlate in vitro with induction of DNA synthesis rather than suppression of apoptosis. Carcinogenesis 17, 16231632.
Jow, L., and Mukherjee, R. (1995). The human peroxisome proliferator-activated receptor (PPAR) subtype NUC1 represses the activation of hPPARa and thyroid hormone receptors. J. Biol. Chem. 270,
3836 3840.
Juge-Aubry, C. E., Gorla-Bajszczak, A., Pernin, A., Lemberger, T.,
Wahli, W., Burger, A. G., and Meier, C. A. (1995). Peroxisome
proliferator-activated receptor mediates cross talk with thyroid
hormone receptor by competition for retinoid X receptor. J. Biol.
Chem. 270, 1811718122.
Kraupp-Grasl, B., Huber, W., Taper, H., et al. (1991). Increased
susceptibility of aged rats to hepatocarcinogenesis by the peroxisome proliferator nafenopin and the possible involvement of altered liver foci occurring spontaneously. Cancer Res. 51, 666 671.
Kraupp-Grasl, B., Huber, W., Putz, B., et al. (1990). Tumor promotion by the peroxisome proliferator nafenopin involving a specific
subtype of altered foci in rat liver. Cancer Res. 50, 37013708.
Kurata, Y., Kidachi, F., Yokoyama, M., et al. (1998). Subchronic
toxicity of di(2-ethylhexyl)phthalate in common marmosets: Lack
of hepatic peroxisome proliferation, testicular atrophy, or pancreatic acinar cell hyperplasia. Toxicol. Sci. 42, 49 56.
Labow, R. S., Tocchi, M., and Rock, G. (1986). Contamination of
platelet storage bags by phthalate esters. J. Toxicol. Environ.
Health 19, 591598.
Lake, B. G., and Lewis, D. F. V. (1993). Structureactivity relationships for chemically induced peroxisome proliferation in mammalian liver. In Peroxisomes: Biology and Importance in Toxicology
and Medicine (G. Gibson and B. Lake, Eds.), pp. 313342. Taylor
and Francis, London.
Lake, B. G., Evans, J. G., Cunningham, M. E., and Price, R. J.
(1993a). Comparison of the hepatic effects of nafenopin and WY-

14,643 on peroxisome proliferation and cell replication in the rat
and Syrian hamster. Environ. Health Perspect. 101, 241248.
Lake, B. G. (1995a). Mechanisms of hepatocarcinogenicity of peroxisome-proliferating drugs and chemicals. Annu. Rev. Pharmacol.
Toxicol. 35, 483507.
Lake, B. G. (1995b). Peroxisome proliferation: Current mechanisms
relating to non-genotoxic carcinogenesis. Toxicol. Lett. 82/83, 673
681.
Lake, B. G., Gangolli, S. D., Grasso, P., and Lloyd, A. G. (1975).
Studies on the hepatic effects of orally administered di-(2-ethylhexyl)phthalate in the rat. Toxicol. Appl. Pharmacol. 32, 355367.
Lake, B. G., Gray, T. J. B., Foster, J. R., Stubberfield, C. R., and
Gangolli, S. D. (1984). Comparative studies on di(2-ethylhexyl)phthalate-induced hepatic peroxisome proliferation in the rat and
hamster. Toxicol. Appl. Pharmacol. 72, 46 60.
Lake, B. G., Gray, T. J. B., and Gangolli, S. D. (1986). Hepatic effects
of phthalate esters and related compoundsIn vivo and in vitro
correlations. Environ. Health Perspect. 67, 283290.
Lake, B. G., Evans, J. G., Gray, T. J. B., et al. (1989). Comparative
studies on nafenopin-induced hepatic peroxisome proliferation in
the rat, Syrian hamster, guinea pig, and marmoset. Toxicol. Appl.
Pharmacol. 99, 148 160.
Lake, B. G., and Lewis, D. F. V. (1996). The CYP4 family. In Cytochromes P450: Metabolic and Toxicological Aspects (C. Ioannides,
Ed.), pp. 271297. CRC Press, Boca Raton.
Lake, B. G., Lewis, D. F. V., Gray, T. J. B., and Beamand, J. A.
(1993b). Structureactivity relationships for induction of peroxisomal enzyme activities in primary rat hepatocyte cultures. Toxicol.
In Vitro 7, 605 614.
Lalwani, N. D., Reddy, M. K., Ghosh, S., Barnard, S. D., Molello, J. A.,
and Reddy, J. K. (1985). Induction of fatty acid b-oxidation and
peroxisome proliferation in the liver of Rhesus monkeys by DL-040, a
new hypolipidemic agent. Biochem. Pharmacol. 34, 34733482.
Lapinskas, P. J., and Corton, J. C. (1997). Phthalate ester plasticizers differentially activate the peroxisome proliferator-activated
receptors. Fundam. Appl. Toxicol. (Suppl. Toxicologist) 36, 144.
Lazarow, P. B., and deDuve, C. (1976). A fatty acyl-CoA oxidizing
system in rat liver peroxisomes: Enhancement by clofibrate, a
hypolipidemic drug. Proc. Natl. Acad. Sci. USA 73, 20432046.
Lee, S. S. T., Pineau, T., Drago, J., lee, E. J., Owens, J. W., Kroetz,
D. L., et al. (1995). Targeted disruption of the alpha isoform of the
peroxisome proliferator-activated receptor gene in mice results in
abolishment of the pleiotropic effects of peroxisome proliferators.
Mol. Cell. Biol. 15(6), 30123022.
Lewis, L. M., Flechtner, T. W., Kerkay, J., Pearson, K. H., and
Nakamoto, S. (1978). Bis (2-ethylhexyl) phthalate concentrations
in the serum of hemodialysis patients. Clin. Chem. 24(5), 741746.
Lhuguenot, J. C., and Cornu, M. C. (1993). Metabolism of di-(2-ethylhexyl)-phthalate (DEHP) and di-(2-ethylhexyl)-adipate (DEHA) and
their relationship to peroxisome proliferation in different species. In
Peroxisomes: Biology and Importance in Toxicology and Medicine (G.
Gibson and B. Lake, Eds.), pp. 465 485. Taylor and Francis, London.
Lundgren, B., Meijer, J., and DePierre, J. W. (1987). Examination of
the structural requirements for proliferation of peroxisomes and
mitochondria in mouse liver by hypolipidemic agents, with special
emphasis on structural analogues of 2-ethylhexanoic acid. Eur.
J. Biochem. 163, 423 431.
Makowska, J. M., Gibson, G. G., and Bonner, F. W. (1992). Species
differences in ciprofibrate induction of hepatic cytochrome P450
4A1 and peroxisome proliferation. J. Biochem. Toxicol. 7, 183191.
Mannaerts, G. P., and Van Veldhoven, P. P. (1993). Metabolic role of
mammalian peroxisomes. In Peroxisomes: Biology and Importance
in Toxicology and Medicine (G. Gibson and B. Lake, Eds.), pp.
19 62. Taylor and Francis, London.
355
Marsman, D. S., and Popp, J. A. (1994). Biological potential of basophilic hepatocellular foci and hepatic adenoma induced by the
peroxisome proliferator, WY-14,643. Carcinogenesis 15(1), 111
117.
Marsman, D. S., Cattley, R. C., Conway, J. G., and Popp, J. A. (1988).
Relationship of hepatic peroxisome proliferation and replicative
DNA synthesis to the hepatocarcinogenicity of the peroxisome
proliferators di(2-ethylhexyl)phthalate and [4-chloro-6-(2,3-xylidino)2-pyrimidinylthio]acetic acid (WY-14,643) in rats. Cancer Res. 48,
6739 6744.
Marsman, D. S., Swanson-Pfeiffer, C. L., and Popp, J. A. (1993). Lack
of comitogenicity by the peroxisome proliferator hepatocarcinogens, Wy-14,643 and clofibric acid. Toxicol. Appl. Pharmacol. 122,
1 6.
Maruyama, H., Tanaka, T., and Williams, G. M. (1990). Effects of the
peroxisome proliferator di(2-ethylhexyl)phthalate on enzymes in
rat liver and on carcinogen-induced liver altered foci in comparison
to the promoter phenobarbital. Toxicol. Pathol. 18(2), 257267.
Mennes, W. C., Wortelboer, H. M., Hassing, G. A. M., van Sandwijk,
K., Timmerman, A., Schmid, B. P., Jahn, U., and Blaauboer, B. J.
(1994). Effects of clofibric and beclobric acid in rat and monkey
hepatocyte primary culture: Influence on peroxisomal and mitochondrial b-oxidation and the activity of catalase, glutathione Stransferase and glutathione peroxidase. Arch. Toxicol. 68, 506
511.
Mitchell, F. E., Price, S. C., Hinton, R. H., Grasso, P., and Bridges,
J. W. (1985a). Time and doseresponse study of the effects on rats
of the plasticizer di(2-ethylhexyl)phthalate. Toxicol. Appl. Pharmacol. 81, 371392.
Mitchell, A. M., Lhuguenot, J.-C., Bridges, J. W., and Elcombe, C. R.
(1985b). Identification of the proximate peroxisome proliferator(s)
derived from di(2-ethylhexyl)phthalate. Toxicol. Appl. Pharmacol.
80, 2332.
Miyamoto, T., Kaneko, A., Kakizawa, T., Yajima, H., Kamijo, K.,
Sekine, R., Hiramatsu, K., Nishim, Y., Hashimoto, T., and Hashizume, K. (1997). Inhibition of peroxisome proliferator signaling
pathways by thyroid hormone receptor. J. Biol. Chem. 272, 7752
7758.
Miyata, K. S., McCaw, S. E., Patel, H. V., Rachubinski, R. A., and
Capone, J. P. (1996). The orphan nuclear hormone receptor LXRa
interacts with the peroxisome proliferator-activated receptor and
inhibits peroxisome proliferator signaling. J. Biol. Chem. 271,
9189 9192.
Moody, D. E., and Reddy, J. K. (1978). Hepatic peroxisome (microbody) proliferation in rats fed plasticizers and related compounds.
Toxicol. Appl. Pharmacol. 45, 497504.
Mukherjee, R., Jow, L., Noonan, D., and McDonnell, D. P. (1994).
Human and rat peroxisome proliferator activated receptors
(PPARs) demonstrate similar tissue distribution but different responsiveness to PPAR activators. J. Steroid Biochem. Mol. Biol.
51, 157166.
National Toxicology Program (NTP) (1982a). Carcinogenesis bioassay of di-(2-ethylhexyl)phthalate (CAS No. 117-81-7) in F344 rats
and B6C3F1 mice (feed study). Tech. Rep. Ser. 217.
National Toxicology Program (NTP) (1982b). Carcinogenesis bioassay of di(2-ethylhexyl)adipate (CAS No. 103-23-1) in F344 rats and
B6C3F1 mice (feed study). Tech. Rep. Ser. 212.
National Toxicology Program (NTP) (1996). Tumor Incidence in Control Animals by Route and Vehicle of AdministrationF344 Rats.
Contract No. N01-ES-65401, Analytical Sciences, Inc.
Orton, T. C., Adam, H. K., Bentley, M., Holloway, B., and Tucker,
M. J. (1984). Clobuzarit: Species differences in the morphological
and biochemical response of the liver following chronic administration. Toxicol. Appl. Pharmacol. 73, 138 151.
356
DOULL ET AL.
Osumi, T., and Hashimoto, T. (1978). Enhancement of fatty acyl-CoA

oxidizing activity in rat liver peroxisomes by di(2-ethylhexyl)phthalate. J. Biochem. 83, 13611365.
Palmer, C. N. A., Hsu, M.-H., Griffith, K. J., and Johnson, E. F.
(1996). Species differences in PPARa expression levels may determine species specific responsiveness to peroxisome proliferators.
Abstract, 11th International Symposium on Microsomes and Drug
Oxidations, Los Angeles, CA, July 1996.
Palmer, C. A. N., Hsu, M.-H., Muerhoff, A. S., Griffin, K. J., and
Johnson, E. F. (1994). Interaction of the peroxisome proliferatoractivated receptor a with the retenoid X receptor a unmasks a
cryptic peroxisome proliferator response element that overlaps an
ARP-1-binding site in the CYP4A6 promoter. J. Biol. Chem. 269,
1808318089.
Palmer, C. A. N., Hsu, M.-H., Griffin, K. J., Raucy, J. L., and Johnson, E. F. (1998). Peroxisome proliferator activated receptor-a expression in human liver. Mol. Pharmacol. 53, 14 22.
Parzefall, W., Schuppler, J., Barthel, G., et al. (1990). Toxicological
studies on a benzofurane derivative. I. A comparative study with
phenobarbital on rat liver. Toxicol. Appl. Pharmacol. 106, 482
499.
Parzefall, W., Erber, E., Sedivy, R., and Schulte-Hermann, R. (1991).
Testing for induction of DNA synthesis in human hepatocyte primary cultures by rat tumor promoters. Cancer Res. 51, 11431147.
Peck, C. C., Odorn, D. G., Friedman, H. T., et al. (1979). Di-(2ethylhexyl)phthalate (DEHP) and mono-(2-ethylhexyl)phthalate
(MEHP) accumulation in whole blood and red cell concentrates.
Transfusion 19, 137146.
Perrone, C. E., Shao, L., and Williams, G. M. (1998). Effect of rodent
hepatocarcinogenic peroxisome proliferators on fatty acyl-CoA oxidase, DNA synthesis, and apoptosis in cultured human and rat
hepatocytes. Toxicol. Appl. Pharmacol. 150, 277286.
Peters, J. M., Cattley, R. C., and Gonzalez, F. J. (1997). Role of
PPARa in the mechanism of action of the nongenotoxic carcinogen
and peroxisome proliferator WY-14,643. Carcinogenesis 18, 2029
2033.
Plant, N. J., Horley, N. J., Savory, R. L., Elcombe, C. R., Gray,
T. J. B., and Bell, D. R. (1998). The peroxisome proliferators are
hepatocyte mitogens in chemically-defined media: Glucocorticoidinduced PPARa is linked to peroxisome proliferator mitogenesis.
Carcinogenesis 19, 925931.
Poland, A. (1997). Reflections on risk assessment of receptor-acting
xenobiotics. Regul. Toxicol. Pharmacol. 26, 41 43.
Pollack, G. M., Li, R. C. K., Ermer, J. C., and Shen, D. D. (1985a).
Effects of route of administration and repetitive dosing on the
disposition kinetics of di(2-ethylhexyl)phthalate and its mono-diesterified metabolite in rats. Toxicol. Appl. Pharmacol. 79, 246
256.
Pollack, G. M., Buchanan, J. F., Slaughter, R. L., et al. (1985b).
Circulating concentration of di(2-ethylhexyl)phthalate and its deesterified phthalic acid products following plasticizer exposure in
patients receiving hemodialysis. Toxicol. Appl. Pharmacol. 79,
257267.
Popp, J. A., Garvey, L. K., Hamm, J. T. E., and Swenberg, J. A.
(1985). Lack of hepatic promotional activity by the peroxisomal
proliferating hepatocarcinogen di(2-ethylhexyl)phthalate. Carcinogenesis 6, 141144.
Popp, J. A., and Cattley, R. C. (1993). Peroxisome proliferators as
initiators and promoters of rodent hepatocarcinogenesis. In Peroxisomes: Biology and Importance in Toxicology and Medicine (G.
Gibson and B. Lake, Eds.), pp. 653 665. Taylor and Francis,
London.
Rao, M. S., Lalwani, N. D., Watanabe, T. K., and Reddy, J. K.
(1984a). Inhibitory effect of the antioxidants ethoxiquin and 2(3)tert-butyl-4-hydroxyanisole on hepatic tumorigenesis in rats fed
ciprofibrate, a peroxisomal proliferator. Cancer Res. 44, 1072

1076.
Rao, M. S., Usuda, N., Subbarao, V., and Reddy, J. K. (1987). Absence of gamma-glutamyl transpeptidase activity in neoplastic
lesions induced in the liver of male F-344 rats by di-(2-ethylhexyl)phthalate, a peroxisome proliferator. Carcinogenesis 8, 13471350.
Rao, M. S., and Reddy, J. K. (1991). An overview of peroxisome
proliferator-induced hepatocarcinogenesis. Environ. Health Perspect. 93, 205209.
Rao, M. S., Yeldandi, A. V., and Subbarao, V. (1990). Quantitative
analysis of hepatocellular lesions induced by di(2-ethylhexyl)phthalate in F-344 rats. J. Toxicol. Environ. Health 30, 85 89.
Rao, M. S., Lalwani, N. D., and Reddy, J. K. (1984b). Sequential
histologic study of rat liver during peroxisome proliferator (4chloro-6-(2,3-xylidino)-2-pyrimidinylthio)acetic acid (WY-14,643)induced carcinogenesis. J. Natl. Cancer Inst. 73, 983990.
Reddy, J. K., and Qureshi, S. A. (1979). Tumorigenicity of the hypolipidemic peroxisome proliferator ethyl-a-p-chlorophenoxyisobutyrate (clofibrate) in rats. Br. J. Cancer 40, 476 482.
Reddy, J. K., and Mannaerts, G. P. (1994). Peroxisomal lipid metabolism. Annu. Rev. Nutr. 14, 343370.
Reddy, J. K., and Reddy, M. S. (1977). Malignant tumors in rats fed
nafenopin, a hepatic peroxisome proliferator. J. Natl. Cancer Inst.
59, 16451650.
Reddy, J. K., and Lalwani, N. D. (1983). Carcinogenesis by hepatic
peroxisome proliferators: Evaluation of the risk of hypolipidemic
drugs and industrial plasticizers to humans. Crit. Rev. Toxicol. 12,
158.
Reddy, J. K., Lalwani, N. D., Qureshi, S. A., Reddy, M. K., and
Moehle, C. M. (1984). Induction of hepatic peroxisome proliferation
in non-rodent species, including primates. Am. J. Pathol. 114,
171183.
Reddy, J. K., Azarnoff, D. L., and Sell, S. (1979). Mitogenic and carcinogenic effects of a hypolipidemic peroxisome proliferation, (4-chloro6-(2,3-xylidino)-2-pyrimidinylthio)acetic acid (WY-14,643), in rat and
mouse liver. Cancer Res. 39, 152161.
Rhodes, C., Orton, T. C., Pratt, I. S., et. al. (1986). Comparative
pharmacokinetics and subacute toxicity of di(2-ethylhexyl)phthalate (DEHP) in rats and marmosets: Extrapolation of effects in
rodents to man. Environ. Health Perspect. 65, 299 308.
Richert, L., Price, S., Chesne, C., Maita, K., and Carmichael, N.
(1996). Comparison of the induction of hepatic peroxisome proliferation by herbicide oxadiazon in vivo in rats, mice, and dogs and
in vitro in rat and human hepatocytes. Toxicol. Appl. Pharmacol.
141, 35 43.
Roberts, R. A. (1996). Non-genotoxic hepatocarcinogenesis: Suppression of apoptosis by peroxisome proliferators. Ann. N.Y. Acad. Sci.
804, 588 611.
Rock, G., Secours, V. E., Franklin, C. A., Chu, I., and Villeneuve,
D. C. (1978). The accumulation of mono-2-ethylhexylphthalate
(MEHP) during storage of whole blood and plasma. Transfusion
18, 553558.
Rodricks, J. V., and Turnbull, D. (1987). Interspecies differences in
peroxisomes and peroxisome proliferation. Toxicol. Ind. Health 3,
197212.
Rose, M. L., Germolec, D. R., Schoonhoven, R., and Thurman, R. G.
(1997). Kupffer cells are causally responsible for the mitogenic
effect of peroxisome proliferators. Carcinogenesis 18, 14531456.
Rusyn, I., Tsukamoto, H., and Thurman, R. G. (1998) WY-14-463
rapidly activates nuclear factor GB in Kupffer cells before hepatocytes. Carcinogenesis 19, 12171222.
Roussyn, I., Yen, T. Y., Wong, V. A., Cattley, R. C., Cunningham,
M. L., and Swenberg, J. A. (1997). WY-14,463, a potent peroxisome
proliferator, induces expression of N-methylpurineDNA-glycosylase (MPG) in rat liver and increases urinary excretion of 1,N 6-

ethenoadenine (e-A). NATO Conference on DNA damage and repair, Nov. 1997.
Rubin, R. J., and Schiffer, C. A. (1976). Fate in humans of the
plasticizer, di-2-ethylhexyl phthalate, arising from transfusion of
platelets stored in vinyl plastic bags. Transfusion 16(4), 330 335.
Sakuma, M., Yamada, J., and Suga, T. (1992). Comparison of the
inducing effect of dehydroepiandrosterone on hepatic peroxisome
proliferation-associated enzymes in several rodent species. A
short-term administration study. Biochem. Pharmacol. 43, 1269
1273.
Sausen, P. J., Lee, D. C., Rose, M. L., and Cattley, R. C. (1995).
Elevated 8-hydroxydeoxy-guanosine in hepatic DNA of rats following exposure to peroxisome proliferators: Relationship to mitochondrial alterations. Carcinogenesis 16, 17951801.
Short, R. D., Robinson, E. C., Lington, A. W., and Chin, A. E. (1987).
Metabolic and peroxisome proliferation studies with di(2-ethylhexyl)phthalate in rats and monkeys. Toxicol. Ind. Health 3, 185195.
Sjoberg, P., Bondesson, U., and Hammarlund, M. (1985). Non-linearities in the pharmacokinetics of di-(2-ethylhexyl) phthalate and
metabolites in male rats. Arch. Toxicol. 58, 7277.
Small, G. M., Burdett, K., and Connock, N. V. (1982). Clofibrateinduced changes in enzyme activities in liver, kidneys and small
intestine of male mice. Ann. N.Y. Acad. Sci. 386, 460 463.
Smith, L. L., and Elcombe, C. R. (1989). Mechanistic studies: Their
role in the toxicological evaluation of pesticides. Food Addit. Contam. 6(Suppl. 1), S57S65.
Stott, W. T. (1988). Chemically induced proliferation of peroxisomes.
Regul. Toxicol. Pharmacol. 8, 125129.
Takagi, A., Sai, K., Umemura, T., Hasegawa, R., and Kurokawa, Y.
(1990a). Significant increase of 8-hydroxydeoxyguanosine in liver
DNA of rats following short-term exposure to the peroxisome proliferators di(2-ethylhexyl)phthalate and di(2-ethylhexyl)adipate.
Jpn. J. Cancer Res. (Gann) 81, 213215.
Takagi, A., Sai, K., Umemura, T., Hasegawa, R., and Kurokawa, Y.
(1990b). Relationship between hepatic peroxisome proliferation
and 8-hydroxydeoxyguanosine formation in liver DNA of rats following long-term exposure to three peroxisome proliferators:
di(2ethylhexyl) phthalate, aluminum clofibrate and simfibrate.
Cancer Lett. 53, 3338.
Tamura, H., Iida, T., Watanabe, T., and Suga, T. (1991). Lack of
induction of hepatic DNA-damage on long-term administration of
peroxisome proliferators in male F-344 rats. Toxicology 69, 55 62.
Tanaka, A., Adachi, T., Takahashi, T., and Yamaha, T. (1975). Biochemical studies on phthalic esters. I. Elimination, distribution and metabolism of di-(2-ethylhexyl)phthalate in rats. Toxicology 4, 253264.
Tobe, M., Suzuki, Y., Ikeda, Y., and Naito, K. (1984). A Study on
Toxicity of Phthalates and Adipate. Chronic Toxicity Test in Mice,
FDA FAP No. 1T 3573: 010623 010676.
357
Tomaszewski, K. E., Derks, M. C., and Melnick, R. L. (1987). Acyl

CoA oxidase is the most suitable marker for hepatic peroxisomal
changes caused by treatment of F344 rats with di(2-ethylhexyl)phthalate. Toxicol. Lett. 37, 203212.
Tucker, M. J., Topham, J. C., and Marsden, A. M. (1988). Long term
effects of clofibrate treatment in the primate (Callithrix Jacchus).
Toxicologist 8(1), Abstract 435.
Tucker, M. J., and Orton, T. C. (1993). Toxicological studies in
primates with three fibrates. In Peroxisomes: Biology and Importance in Toxicology and Medicine (G. Gibson and B. Lake, Eds.),
pp. 425 447. Taylor and Francis, London.
Tucker, M. J., and Orton, T. C. (1995). Comparative Toxicology of
Hypolipidaemic Fibrates. Taylor and Francis, London.
Tugwood, J. D., Aldridge, T. C., Lambe, K. G., McDonald, N., and
Woodyatt, N. J. (1996). Peroxisome proliferator-activated receptors: Structures and functions. Ann. N.Y. Acad. Sci. 804, 252
265.
Tugwood, J. D., Holden, P. R., James, N. H., Prince, R. A., and
Roberts, R. A. (1998). Isolation and characterization of a functional
peroxisome proliferator-activated receptor-alpha from guinea pig.
Arch. Toxicol. 72, 169 177.
Ward, J. M., Rice, J. M., Creasia, D., et al. (1983). Dissimilar patterns
of promotion by di-(2-ethylhexyl)phthalate and phenobarbital of
hepatocellular neoplasia initiated by diethylnitrosamine in
B6C3F1 mice. Carcinogenesis 4, 10211029.
Ward, J. N., Peters, J. M., Perella, C. M., and Gonzalez, F. J. (1998).
Receptor and nonreceptor-mediated organ-specific toxicity of di(2ethylhexyl)phthalate (DEHP) in peroxisome proliferator-activated
receptor a-null mice. Toxicol. Pathol. 26, 240 246.
Watanabe, T., Horie, S., Yamada, J., Isaji, M., Nishigaki, T., Naito,
J., and Suga, T. (1989). Species differences in the effects of bezafibrate, a hypolipidemic agent, on hepatic peroxisome-associated
enzymes. Biochem. Pharmacol. 38, 367371.
Williams, G. M. (1992). DNA reactive and epigenetic carcinogens.
Exp. Toxicol. Pathol. 44, 457 464.
Williams, G. M., Maruyama, H., and Tanaka, T. (1987). Lack of rapid
initiating, promoting or sequential syncarcinogenic effects of di(2ethylhexyl)phthalate in rat liver carcinogenesis. Carcinogenesis 8,
875 880.
Williams, G. M., and Perrone, C. (1996). Mechanism-based risk assessment of peroxisome proliferating rodent hepatocarcinogens.
Ann. N.Y. Acad. Sci. 804, 554 572.
Zhu, Y., Qi, C., Jain, S., Rao, M. S., and Reddy, J. K. (1997). Isolation
and characterization of PBP, a protein that interacts with peroxisome proliferator-activated receptor-alpha. J. Biol. Chem. 272,
25500 25506.

A Cancer Risk Assessment of Di (2-Ethylhexyl) Phthalate: Application of The New U.S. EPA Risk Assessment Guidelines

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Cancer Risk Assessment of Di (2-Ethylhexyl) Phthalate: Application of The New U.S. EPA Risk Assessment Guidelines

Uploaded by

Copyright:

Available Formats

Regulatory Toxicology and Pharmacology 29, 327357 (1999)

Article ID rtph.1999.1296, available online at http://www.idealibrary.com on

A Cancer Risk Assessment of Di(2-ethylhexyl)phthalate: Application

To whom correspondence should be addressed at 12521 Old

sponsiveness. These differences are explained, in part,

The United States Environmental Protection Agency

human carcinogen), based solely on studies conducted

DEHP is one of many substances that increase the

The characteristics of peroxisome proliferation in rats

DEHP RISK ASSESSMENT

tion (Peters et al., 1997) observed in normal mice. The

A wide variety of chemicals have been shown to

Relative liver weight

Note. All doses were calculated from Huber et al. (1996).

crease in cell replication in rodent hepatocytes during

play a tumor incidence of 100% at doses at or below

Humans are exposed to DEHP via oral, inhalation,

DEHP RISK ASSESSMENT

pig, and hamster (Lake et al., 1984; Albro, 1989). In the

of 100 mg/kg body wt radiolabeled DEHP after 4 days,

A large number of chronic/carcinogenicity studies

analyses were conducted to evaluate DEHP-induced

only. In these studies, there was a dose-related effect,

As described elsewhere, DEHP and other peroxisome

DEHP RISK ASSESSMENT

peroxisome proliferators are not produced through a

alyzed by high-performance liquid chromatography and

glutathione S-peroxidase are absent. DEHP has been

Peroxisome proliferator-activated receptor a

controls that regulate cell turnover of the normal tissue.

DEHP RISK ASSESSMENT

PARa) in the induction of peroxisome proliferation has

Many studies have investigated species differences

Dog, rabbit, Rhesus

DEHP RISK ASSESSMENT

Effect on peroxisomes assessed by either subjective (qualitative) or morphometric (quantitative) analysis.

subjects. The original WHO report stated that crude

Blaauboer et al. (1990)

Benford et al. (1986)

Bouis et al. (1993)

Cornu et al. (1992)

DEHP is leached in significant amounts from the

12 months of renal dialysis. Ganning et al. (1984, 1987)

DEHP RISK ASSESSMENT

Perrone et al. (1998)

Lake et al., 1986; Bieri, 1993; Foxworthy and Eacho,

these compounds have been examined in a number of

proliferators is the stimulation of cell replication, which

DEHP RISK ASSESSMENT

much less responsive than rodents to peroxisome proliferators including DEHP.

be a linear function of PPAR expression levels. It is

three receptors cloned to date were not activatable by

Interindividual Variation in PPARa Expression and

There is no good evidence for genetic polymorphisms in

Tissue Differences in PPARa Expression

DEHP RISK ASSESSMENT

that peroxisome proliferators do not produce liver tumors in such species.

Note. 1, positive; 1/2, weakly positive; 1 ?, potentially positive; 2,

In this assessment, genotoxicity includes molecular

Overall Assessment on Genotoxicity of DEHP

rigid application of the linear multistage model, the

1. DEHP is a hepatocarcinogen in F344 rats and

DEHP RISK ASSESSMENT