Pharmacokinetics, interactions with macromolecules and species differences in metabolism of DEHP.

The recent long-term carcinogenesis bioassay of di(2-ethylhexyl) phthalate (DEHP) in rats and mice reported by the National Toxicology Program was the first such bioassay to implicate DEHP as a hepatocarcinogen. At the levels of DEHP fed (up to 1.2% of the diet for two years), the livers of the rats would have been exposed to unhydrolyzed diester; this would not have been the case at lower dosages. Extrapolation to lower dosages is therefore questionable. We do not have sufficient pharmacokinetic data in mice to evaluate the dose relationships as yet. Rodents differ conspicuously from primates in their manner of metabolizing DEHP, both in terms of the demand made on the oxidation potential of the liver and in the chemical properties of the major metabolites. The relevance of these differences must be determined before rodent species can be considered models for the effects of DEHP in humans. Radioactivity from carbonyl-labeled DEHP did not associate with purified protein, RNA or DNA from rat liver in vivo. Label from 2-ethyl-(1-14C)-hexyl-labeled DEHP or mono(2-ethylhexyl) phthalate (MEHP) did appear to associate strongly with purified DNA, but label from free 14C-labeled 2-ethylhexanol did not. The apparent binding from DEHP and MEHP was not exchangeable, but was not proven to be covalent. This phenomenon needs additional study.


Introduction
The studies to be presented in this report were undertaken in response to the recent announcement by the National Toxicology Program (NTP) that di(2-ethylhexyl) phthalate (DEHP), fed at high levels in the diet to rats and mice for two years, was a hepatocarcinogen. We wanted to know something about the possible mechanism of carcinogenesis in rodent liver and also to explore the relevance of the extremely high doses used to produce it.

Dose Dependence of Distribution/ Elimination
The first study relates to the effect of the dosageof DEHP on its metabolism, distribution and excre-*National Institute of Environmental Health Sciences, P. 0 tion. Preliminary experiments with both Sprague-Dawley (CD) and Fischer 344 rats revealed that the maximum amount of DEHP that could be given as a single oral dose without significant excretion of unabsorbed DEHP in the feces was 200 mg/kg body weight. Pharmacokinetic studies using 14C-labeled DEHP at doses above this level would rapidly become dominated by unabsorbed DEHP and could not be directly compared to rates of elimination at lower doses.
We chose to give DEHP by gavage in cottonseed oil rather than incorporated into the solid diet for several reasons. First, we wished to avoid the problem of determining the amount of food actually eaten, and the problem of achieving a comparable dose among replicates on an ad libitum schedule. Second, we wished to avoid the possibility of contamination of urine and feces with radioactivity from food crumbs. Most importantly, doctoring feed with high levels of substances that can be smelled alters feeding habits. Feed containing low levels of DEHP is consumed over a shorter time interval than food containing a high level of DEHP. In order 20 to have DEHP dosage be the dominant variable, then, it was necessary to administer the compound by gavage.
To lessen the discrepancy between the NTP bioassay feeding conditions and the present study, the DEHP was given each day at 5 P.M., shortly before the beginning of the dark cycle. The rats usually began feeding shortly thereafter, so absorption occurred in the presence of the normal dietary nutrients.
The first study involved 36 male Fischer 344 rats, 12 in each of three groups (A, B and C). The rats received NIH 31 chow and water ad libitum. Feces and urine were collected daily at 4:30 P.M. and stored frozen. Group A rats received 1.8 mg of [7-14C]-DEHP (carbonyl label) per kilogram body weight per day. Group B received 18 mg/kg per day, and group C received 180 mg/kg per day. All received the same amount of radioactivity, with different amounts of nonradioactive DEHP diluent.
Rats received a total of 10 daily doses of DEHP.
Three rats from each group were sacrificed 1, 3, 10 or 12 days after receiving their first dose of DEHP. Various tissues and feces were radioassayed using a tissue oxidizer. Urine was radioassayed directly in liquid scintillation fluid. The profiles of radioactive metabolites in urine were determined by HPLC as described previously (1). Lipid-soluble metabolites were extracted from liver (2) and chromatographed in Florisil with 15% diethyl ether in hexane. Unhydrolyzed DEHP was measured by radio-TLC of the eluate from Florisil on silica gel GF in hexane:diethyl ether:acetic acid, 120:30:1.5

Absorption Thresholds
In a separate experiment rats and mice were, in triplicate, given single oral doses of DEHP in cottonseed oil ranging from 1.8 mg/kg body weight to 1 g/kg. The animals were sacrificed 6 hr later, as this was the time after dosing at which the maximum concentration of 14C was found in the livers. The livers were assayed for intact DEHP as described above. The results of this experiment were compared to the level of DEHP found in the livers of rats fed a diet containing 1% DEHP until the level reached a steady state (-6 days).

Species Dependence
The distributions of DEHP-derived metabolites in urine of Fischer 344 rats, CD mice, Syrian golden hamsters, and Hartley albino guinea pigs were determined following single oral doses of DEHP in cottonseed oil by techniques applied previously to urine from CD rats (1, 3, 4,), African Green monkeys (5) and humans (6). In general, the distributions were independent of dose up to the amount of DEHP that caused intact 14C-DEHP to appear in the feces.

Labeling of Macromolecules
In the last type of experiment to be reported here, male Fischer 344 rats were fed diets containing either 0.01% or 1% DEHP, unlabeled, ad libtum for 11 days. The rats were given a tracer dose of labeled DEHP or one of its metabolites on day 1, Pellet (Protein) day 6 or day 10. Rats given labeled DEHP on either day 1 or day 10 were sacrificed on day 11; those given labeled DEHP on day 6 were sacrificed on day 7 (24 hr later). Livers were homogenized, after which protein, RNA and DNA (7) were isolated in highly purified form. The purification procedures are briefly diagrammed in Figures 1 and 2. The fractions were tested for cross contamination by spectrophotometric assays (7-9), after which they were radioassayed, as were aliquots of the original homogenate, by digestion in NCS and dispersion in Aquasol.

Dose Effects
The cumulative excretion of 14C expressed as a percentage of the cumulative dose is diagrammed in Figure 3. After an "adaptation period" of about 4 days, excretion became quite independent of the dose. On an absolute basis, rats in group C were eliminating 14C at a rate 100 times that of the group A rats. Up to a dose rate of 180 mg/kg per day, then, there was no indication of even beginning to saturate the overall elimination mechanism.
Looking more specifically at the liver, we saw ( Table 1) that the percentage of the dose of 14C retained in the livers tended to decrease with time and also with increasing dose. The mean rates of dropoff in radioactivity after dosing was stopped were not statistically significantly different in the three groups. There was no evidence, then, for accumulation in this organ.
Essentially the same observations applied to the testes as were noted for liver ( Table 2 Table 3. These compounds are grouped according to functionality, or roughly according to the extent of oxidative metabolism needed to produce them, in Table 4. This table indicates that the only statistically significant effect of the dosage of DEHP on metabolite distribution in rat urine was a slight increase in free phthalic acid with increasing DEHP dosage. This probably reflects a high Km for mono-(2-ethylhexyl) phthalate (MEHP) of the lipase that produces phthalic acid. A previous study of lipases and esterases in various rat tissues (10) indicated that the only detectable MEHP-hydrolase was located in liver microsomes. It is also possible that this enzyme was induced at high levels of DEHP, but there is no direct evidence at the present time.

Absorption Threshold
As the size of a single oral dose of DEHP was increased, a threshold was reached above which there was a steady climb in the amount of unhydrolyzed DEHP reaching the liver of rats. This is diagrammed for Fischer rats in thereby ingested as though it were a single dose, the concentration of DEHP in liver predicted in Figure 4 was in excellent agreement with what was found on analysis (123 pug/g vs 121 Rug/g found).
Making the approximation that rats consume feed to the extent of 10% of their body weight daily, then intact DEHP will reach their livers whenever its concentration exceeds 0.43% of the diet. Levels above and below the threshold are therefore qualitatively different relative to what chemical compounds reach the liver. However, it must be kept in mind that when the diet contains more than about 0.5% DEHP, rats tend to reduce their food consumption.
In contrast to the results seen in rats, we were unable to detect an absorption threshold in either CD-1 or C3B6F1 mice for doses up to 1 g of DEHP/kg body weight. This may reflect the higher level of DEHP-hydrolase in the intestines of mice than in the intestines of rats (10). Association with Macromolecules Table 5 shows the amounts of radioactivity associated with liver homogenates, purified protein, RNA and DNA when Fischer 344 rats were given a tracer dose of 14C-DEHP or its primary hydrolysis products. The rats had been eating NIH 31 chow fortified with unlabeled DEHP (1% by weight) for 6 days to build up the pools of metabolites and intermediates to a steady-state level; the 14C tracer was given, and the rats sacrificed 24 hr later. When 100 ,uCi of carbonyl-14C (7-14C) DEHP was given, no detectable radioactivity was found associated with purified macromolecules. However, when the 14C was in the ethylhexyl moiety of DEHP (1'-14C), all three classes of macromolecules became labeled. If the [1'-14C]-DEHP was saponified to phthalic acid plus two equivalents of [1'-'4C]-2-ethylhexanol before administration, protein was still labeled, but no detectable 14C became associated with DNA. If [1'-14C]-MEHP was given, DNA was again labeled, If the label associated with DNA was covalently bound, which cannot be proven without isolation and characterization ofthe adduct(s), it would appear that C-1 (at least) of 2-ethylhexanol is incorporated into DNA only when esterified ethylhexanol is given. In any event, it seems clear that the phthalate moiety of DEHP is not bound to either DNA, RNA or protein in rat liver under these conditions. Chemical analysis of the DNA isolated according to the procedure described detected less than 0.1% RNA and less than 0.01% protein. Table 6 shows the decrease with time in the amount of radioactivity associated with macromolecules labeled from [1'-14C]-DEHP. The decrease rate for protein was consistent with the average turnover rate of protein in rat liver (11) and much slower than would be expected for simple exchange with the unlabeled DEHP from the diet. Labeling of DNA was slightly higher for rats on the 1% DEHP diet than for those on the 0.01% DEHP diet, but there is insufficient information on pool sizes and turnover rates to speculate on this phenome-23 non. The decrease in labeling of DNA with time was faster for rats on the 0.01% DEHP diet than would be expected if the decrease was simply due to normal DNA turnover. This suggests that some sort of repair mechanism could be involved that was less effective at reducing DNA radioactivity in rats on the 1% DEHP diet. However, much more information would be needed to establish that repair was occurring. At best, this experiment suggests that further studies of the interaction of DEHP with DNA might be worthwhile.

Species Differences in Metabolism
The data in Table 7 are collected from several sources. Urine was collected at intervals after single oral (species other than primates) or IV [African Green monkeyTs (5) and human leukemia patients (6)] doses of [7 1 C]-DEHP. Urine containing 90% of the total radioactivity excreted was pooled for analysis. Table 7 represents the distributions of radioactive metabolites in pools of urine from three animals except monkeys (two) and humans (two, nonradioactive DEHP). Urine was hydrolyzed with P-glucuronidase for the data in Table 7. A compari- Table 5. "Bound" radioactivity in macromolecules.  son of extracts from unhydrolyzed and glucuronidase-hydrolyzed portions of urine gave the results shown in Table 8. No conjugates other than glucuronides have been detected in any of the species tested; glycine, taurine and sulfate conjugates have been sought and found absent, in contrast to results reported for substituted benzoic and phenylacetic acid metabolites (12).
Most of our investigations of urinary metabolites have involved rats. The metabolite distributions in rats, examined over the past 8 years, have been surprisingly reproducible and independent of factors such as dose of DEHP and age of animal. Rats do not excrete conjugates of DEHP metabolites, an observation that has been confirmed in at least three strains (5,13). In contrast, each of the other five species examined excretes glucuronide conjugates (Table 6). In most cases, conjugates are the major urinary metabolites.
Whereas primates excrete predominately glucuronides of MEHP, whose formation does not require oxidative metabolism, and metabolites with hydroxyl side chains which require only one round ofoxidative metabolism, rats excrete predominately metabolites having carboxyl groups on the side chain (metabolites I-V). These diacids require from three to six oxidative steps for their formation. Carrying oxidative metabolism all the way to the highly water-soluble diacids may be necessary to compensate, in rats, for not making glucuronides.
If formation of hydroxyl side chains involves, by analogy with fatty acid w-oxidation, a mixed function oxidase reaction as postulated earlier (3), one would expect a net conversion of NAD(P)H to NAD(P). The additional steps, from alcohol to aldehyde (or ketone) and from aldehyde to acid, as well as the apparent a-and p-oxidations needed to produce metabolites I-III, would all be associated with net conversion of NAD(P) to NAD(P)H. Thus the overall demand on the oxidation potential of the liver when high doses of DEHP are given would be in opposite directions for rat and primate. To the extent that metabolism of DEHP is involved in its biological activity, then, one must question seriously whether rats can be accepted as a model for man.

Summary
The recent long-term carcinogenesis bioassay in rats and mice as reported at this conference by the NTP is the first such bioassay to report carcinogenesis by DEHP. It is also the first such bioassay to have involved a level of DEHP in the diet high enough to exceed the absorption threshold for unhydrolyzed DEHP. The livers of the rats in the NTP study would have been exposed to moderate levels of intact diester, whereas at levels half those tested or below this would not have occurred. This qualitative difference makes any attempts to extrapolate the hepatocarcinogenesis observed in rats to lower dosage levels questionable. We do not have enough pharmacokinetic data from mice to evaluate the dose relationships as yet.
While we were unable to obtain any evidence for 24 ALBRO ET AL.
FATE OF DEHP 25 covalent binding of the phthalate moiety of DEHP to macromolecules in rat liver, and there was also no apparent binding of radioactivity from [1'-14C]-2-ethylhexanol to DNA in vivo, the labeled carbon atom from either di-([1 '-14C]-2-ethylhexyl) phthalate or the corresponding monoester did appear to bind to DNA in rat liver in vivo. The diester was a more efficient labeling precursor than the monoester. Disappearance of 4C from the labeled DNA had a time course consistent with the operation of a repair mechanism. If so, repair was less effective when there was a high level of DEHP in the diet. That the apparent binding to DNA was truly covalent has not been proven; however, it was not exchangeable. We found no saturation of DEHP metabolic pathways in rats over the 100-fold range of dosages between 1.8 mg DEHP/kg body weight per day andi 180 mg/kg per day. There was no apparent reason to suspect metabolic saturation in the biological activity of high DEHP dosages. Levels of metabolites in liver and testes fluctuated for about four days on a regimen of daily dosing, then slowly but smoothly declined. This hinted at the operation of an adaptation mechanism, and argued against tissue accumulation.
Rats were unique among six species (including man) tested, in that they did not excrete glucuronide conjugates of DEHP metabolites in urine. Rodent species in general carried oxidative metabolism of DEHP much farther than did primates; the former should experience a net decrease and the latter a net increase in the ratio of NAD(P) to NAD(P)H in liver during exposure to high levels of DEHP. Whereas the major metabolites of DEHP in rat urine are diacids, the major metabolites in urine of primates are glucuronides of hydroxyacids. These species differences in the metabolism of DEHP cast doubts upon the validity of using rats as a model for the effects of DEHP in man.