Effects of phthalate esters on lipid metabolism in various tissues, cells and organelles in mammals.

The effect of phthalate ester plasticizers on a variety of enzyme systems was studied in rats, rabbits and pigs. The feeding of di(2-ethylhexyl) phthalate (DEHP) to animals at levels from 0.1% to 1.0% in the diet resulted in diverse biochemical effects, such as inhibition of cholesterologenesis in liver, testes, and adrenal gland; inhibition of cholesterologenesis in brain and liver of fetal rats from DEHP-fed dams; decreased plasma cholesterol levels; decreased synthesis of hepatic phospholipid and triglyceride; increased fatty acid oxidation in isolated liver mitochondria; and a transient decrease in fatty acid oxidation in isolated heart mitochondria. The addition of DEHP to preparations of rat liver in vitro resulted in inhibition of cholesterologenesis, and its addition to isolated mitochondria from rat heart produced an inhibition of adenine nucleotide translocase. DEHP-feeding to rats and rabbits, however, did not affect platelet function as judged by collagen- and ADP-induced aggregation. The studies presented here indicate that the exposure of animals to phthalate esters can result in a significant perturbation of normal metabolism in liver, heart, testes, adrenal gland and brain and can affect blood lipid levels.

Phthalates were evaluated in the 1940s and 1950s by classical toxicological criteria and judged to have a low order of acute and chronic toxicity (27)(28)(29)(30). It was not until the 1960s that the safety of the phthalates came into serious question. During the *Diabetes and Atherosclerosis Research, The Upjohn Company, Kalamazoo, Michigan 49001.
What was surprising to me, however, was that as late as 1975, the knowledge that had accrued concerning phthalate toxicity was devoid of studies on the effects of phthalate esters on biochemical pathways and events, which by necessity would underlie the qualitative toxic responses being observed. It was for this reason, in 1975, that efforts in my laboratory were directed towards investigating the possible effects of phthalate esters on lipid metabolism in various species. The observations presented in this paper summarize results of studies which were initiated in 1975 and which have continued up to the present time. The studies were conducted in rats, rabbits and pigs. Depending upon the nature of the study, phthalate esters were either fed as an admixture to the diets or added, in vitro, to tissue preparations. Details of the techniques and methods used throughout the studies have been described elsewhere and will be referenced in the table footnotes and figure legends. Statistical analyses were performed using Student's independent t-test.

Effects of Phthalate Esters on Sterol Metabolism
The reports that phthalate esters possess teratogenic and mutagenic potential (41)(42)(43)(44) suggested the possibility of their interference with sterologenesis which is a key process in mammalian development. This, in fact, proved to be the case, and the evidence was first published in 1976 (51,52).
Liver (Male Rats). An inhibition of cholesterol synthesis from '4C-acetate is apparent within 48 hr after feeding 0.5% DEHP to male rats ( Table 1). The inhibition is progressive and rises from 14% at 2 days.to 70% by 11 days. Inhibition of labeled mevalonate incorporation into cholesterol by 0.5% DEHP feeding also occurs but is slower to develop; inhibition is not apparent until after 6 days of DEHP exposure. By 11 (52). After incubation, the samples were hydrolyzed (saponified) with alcoholic KOH, and then the neutral sterols were extracted with n-hexane and subsequently fractionated by thin-layer chromatography (52). Values are means of two or three DEHP-fed animals expressed relative to control values obtained in parallel experiments.
sis (56) (Fig. 1). Inhibition at this site would first affect the flow-through of acetate to cholesterol as demonstrated in Table 1 (and diagrammed in Fig.  1) and permit the inhibition to be circumvented initially by use of exogenous mevalonate as a substrate. The inhibition of sterologenesis from labeled mevalonate which develops later (after 6 days, Table  1) may be explained if one or more of the enzymes between mevalonate and cholesterol are substrateinduced and thereby decline in activity as mevalonate production diminishes. In all studies, the decreased synthesis of cholesterol was paralleled by decreases in synthesis of lanosterol and squalene (52), as would be anticipated from Figure 1. Additionally, the cholesterol content of livers from rats fed 0.5% DEHP for 6 days was not different from that of control rats (52) (60). After incubation, the samples were hydrolyzed with alcoholic KOH (52). The nonsaponifiable lipids were recovered by extraction with n-hexane (52) and the 313-H sterols (principally cholesterol) were precipitated with digitonin and assayed for radioactivity (52,59). Values are means ± SEM of the number of animals designated by n.
thus indicating that DEHP-induced inhibition of hepatic cholesterologenesis is unlikely to be attributed to end product inhibition. In fact, feeding of phthalates for longer periods (21 days) tends to reduce hepatic cholesterol (55). End product inhibition is also inconsistent with the observation that DEHP inhibits 14C-acetate incorporation into sterols when added directly to normal male rat liver minces ( Table 2). Liver (Female Rats). Studies in female rats fed 1.0% DEHP for 13 days (61) indicated that inhibition of hepatic sterologenesis is not sex-specific (Table 3).
Liver (Male Rabbits). Inhibition of sterologen-  esis in liver from rabbits fed DEHP (59) also dismissed the possibility that the effect of DEHP is species-specific for the rat (Table 4). Adrenal and Testes. Inhibition of sterologenesis in DEHP-fed animals is not restricted to the liver.
The results of Table 5 show that adrenal glands and testes are similarly affected (59,61), with inhibition reaching 20-25% of normal synthesis. Impaired sterologenesis in adrenal and testicular tissue may, in part, be responsible for the developmental abnormalities (62) reported to occur in fetuses from phthalate-treated dams (41,42,46) and for testicular atrophy in young and adult animals (63)(64)(65)(66)(67). Developing Fetuses and Suckling Offspring. Placental transfer of DEHP has been reported in rats (48). The exposure of the developing fetus to DEHP by this route may have important implications in the development of the fetus. The results of Table 6 indicate that fetuses taken by cesarean section from pregnant rats fed DEHP beginning 5 to 10 days after conception have an impairment of sterologenesis in brain and liver (68). Furthermore, if dams exposed to DEHP during gestation, and during the postnatal period as well, are permitted to nurse their naturally born offspring, hepatic sterologenesis in the pups is reduced below control values (Table 7). In 8-day suckling rats, hepatic sterologenesis was reduced 20% below control values and was accompanied by increased liver size and decreased body weight ( Table 7). The decreased body weight seen in the 8-day-old pups was not apparent in litters permitted to reach weaning age (21 days) ( Table 8). DEHP feeding to the dams resulted in a significant lowering of plasma cholesterol (p < 0.02) but did not affect plasma cholesterol levels in the pups (Table 8). 'Sprague-Dawley rats of both sexes were fed Purina Laboratory Chow or the chow diet supplemented with DEHP. Adrenal glands were removed from females (body wt 251 ± 3 g, n = 12) and males (body wt 320 ± 20 g, n = 10), dissected free of adhering tissue and capsular material, and then incubated for 3 hr at 37°C in 1.52 ml of Krebs-Ringer-bicarbonate buffer, pH 7.4, containing 1.0 ,uCi DL-mevalonic-2-14C acid (DBED salt, 40.8 Ci/mole). Testes (right side) from rats weighing 225-250 g were decapsulated and incubated as the whole gland for 3 hr at 37°C in 5.0 ml Krebs-Ringer-bicarbonate buffer, pH 7.4, containing 2.5 RCi DL-mevalonic-2-'4C acid (DBED salt, 40.1 Ci/mole). After incubation, the tissues were treated with alcoholic KOH as described in the footnote to Table 1. The neutral sterols and squalene were extracted with n-hexane and subsequently fractionated by thin-layer chromatography. Values are means ± SEM (male adrenals, n = 5 per group; all others, n = 6 per group). beginning on the 5th day of gestation. Fetuses n = (7) n = (14) were taken by cesarean section op the 18th day DEHP 11510 ± 520 5675 ± 220 of gestation. Brains were obtained from seven n = (7) n = (13) fetuses from seven different dams. Livers were p < 0.05 p < 0.02 obtained from fetuses from up to eight dams; not more than two fetuses per dam were used. 'Livers and brains were incubated as the entire tissue for 3 hr at 37°C in 3.5 ml Krebs-Ringer-bicarbonate buffer, pH 7.4. Radioactive precursors (acetate-1-14C, sodium salt, SA 59 Ci/mole, DL-mevalonic acid-2-14C, DBED salt, SA 46 Ci/mole, and DL-mevalonic acid-5-3H, DBED salt, SA 5 Ci/mmole) were present at a level of 2,uCi. Plasma Cholesterol. Certain phthalate esters evoke a hypocholesterolemic response in rats of either sex (Tables 8 and 9). At levels of 0.5% and 1.0% in the diet, DEHP reduces plasma cholesterol about 20% in the male rat (Table 9) (55,59,69); this reduction appears to represent a maximum response inasmuch as 0.5% and 1.0% DEHP appear to have equivalent effects (55,59,69). The maximum cholesterol-lowering effect occurs within the first week of DEHP feeding (55). Since a hypocholesterolemic effect of phthalates has also been observed in mice (70) but not in rabbits (59,69), this response appears to have some degree of species specificity.
Comparison ofDimethyl Phthalate (DMP), Din-butyl Phthalate (DBP) and Di-2-ethylhexyl Phthalate (DEHP). In rats fed DMP, DBP or DEHP at a level of 25 mmole/kg diet (equivalent to 1.0% DEHP) for 21 days, DEHP and DBP reduced plasma cholesterol levels 16% and 2, respectively, and inhibited hepatic cholesterologenesis from labeled mevalonate by 67% and 30%, respectively (Table  9). It is possible that the hypocholesterolemic effect of DBP and DEHP is a direct consequence of the inhibition of hepatic sterologenesis, although other mechanisms could be possible. Since inhibition of hepatic sterologenesis from labeled mevalonate (Table   44 F. P. BELL 18.8 ± 0.7 3.01 ± 0.06 3107 ± 208 p < 0.01 p < 0.01 p < 0.05 'Pregnant rats were fed 0.5% DEHP in the diet (w/w) beginning on the 5th day of gestation and were permitted to deliver and to nurse the pups for 8 days; the DEHP diet was continued throughout the entire postnatal period. Liver minces were prepared from each pup and incubated 3 hr at 37°C in 3.5 ml of Krebs-Ringer-bicarbonate buffer, pH 7.4, which contained 2,uCi DL-mevalonic-5'H acid, (DBED salt, SA 5 Ci/mmole). Control pups (n = 10) were from 1 dam; pups in the DEHP group (n = 12) came in equal numbers from two dams. Pups 26 48 ± 1 102 ± 4 'Female rats were fed Purina Laboratory Chow or the chow supplemented with 0.5% DEHP (w/w) from the first day of gestation until the pups were weaned at 21 days. Cholesterol was measured as described (59). Values are means ± SEM of the number of animals designated by n. bSignificant, p < 0.02. 33 ± 8c 'Male Sprague-Dawley rats were fed Purina Chow (control) or the chow containing dimethyl phthalate (DMP), di-n-butyl phthalate (DBP), or di(2-ethylhexyl) phthalate (DEHP); the phthalates were present at a level of 2.5 mmole/100g feed. The diets were fed for 21 days. Details of the analyses have been presented (55). All values are means ± SEM of 5-10 observations. bp < 0.05, Student's independent t-test. Cp < 0.001, Student's independent t-test. 9) and labeled acetate (55) is greater with DEHP than DBP and since DMP does not have this action (Table 9) (55), the presence of a two-carbon alkyl chain appears to be the minimum requirement for inhibition of sterologenesis.

EFFECTS OF PHTHALATE ESTERS ON LIPID METABOLISM
DMP differed from DBP and DEHP in two other respects; it reduced liver total lipid and liver total cholesterol (Table 9), whereas DBP and DEHP showed no significant effect on these parameters.
Additionally, DMP did not produce hepatomegaly as did DBP and DEHP ( Table 9). The increased liver size in the DBP-and DEHP-fed animals appeared to represent a true increase in tissue mass rather than an edematous change since the ratio of lipid-free dry weight to wet weight of liver was similar to control values (55).

Effects of DEHP Feeding on Biosynthesis of Fatty Acids and Glyceryl Lipids in the Rat
Feeding DEHP to rats affects not only sterol synthesis as mentioned previously, but synthesis of other lipids as well. Interestingly, not all tissues appear to be affected to the same extent or in the same manner. The incorporation of 1-14C-acetate into fatty acids, phospholipids and triglycerides was not statistically significantly altered, in vitro, in heart and aorta of male rats fed 0.5% DEHP for 10 days (Table 10). By contrast, 1-_4C-acetate incorporation into these same lipids was reduced about 40% in liver; total 14C-labeled lipid synthesis (includes sterols) was also reduced about 48% ( Table  Table 10. Effect of DEHP feeding (0.5%) on the incorporation of "C-acetate into fatty acids and glyceryl lipids in various tissues of the rat, in vitro.a "C-Acetate incorporation, dpm/g wet weight 10). In kidney, 14C-acetate incorporation in triglycerides was reduced 33%, whereas synthesis of the fatty acids and phospholipids was not significantly altered by DEHP feeding ( Table 10). Fractionation of the hepatic phospholipids synthesized from 14Cacetate in rats fed 0.5% DEHP showed a significant reduction in the distribution of 14C-acetate into the phosphatidylcholine fraction ofthe liver phospholipids (53). In addition, feeding 0.5% DEHP to male rats for 9 days resulted in 37-44% reduction in the esterification of 3H-oleate into diglycerides and triglycerides in liver minces (55). The above studies suggest that DEHP can, in certain tissues, inhibit de novo fatty acid synthesis, decrease fatty acid esterification, and selectively affect the synthesis of specific phospholipid species; the latter has been confirmed chemically as well (17).

Effects of DEHP on Mitochondrial Function
The ability of DEHP to accumulate in heart tissue of animals (23,24) and man (21) and, more specifically, to accumulate in the mitochondrial fraction of heart tissue (23, 24) suggested a need to investigate possible biochemical effects of DEHP on mitochondrial function.
Fatty Acid Oxidation in Heart and Liver Mito-chondriafromRat,RabbitandPig. DEHPfeeding to rats, rabbits and pigs resulted in an increased capacity oftheir liver mitochondria to oxidize palmitic acid (72). In suitably prepared suspensions of isolated mitochondria, the oxidation of 14C-palmitoyl CoA was enhanced two-to threefold after several days' exposure to the DEHP (Table 11). The enhanced capacity to oxidize 14C-palmitoyl CoA substrate was undiminished throughout the longest periods of exposure studied (i.e., 12 days in rabbits and 35 days in rats). In the rat, the maximum response appeared at a level of 0.1% DEHP in the diet (Table  11). The response of heart mitochondria to DEHP feeding was markedly different from that shown in liver. Ingestion of 0.1% DEHP for 2 days resulted in a 25% reduction in fatty acid oxidation by heart mitochondria. The inhibition increased to about 50% in 3 to 4 days ( Table 11). The inhibition, however, appears to be a transient effect, since full oxidizing capacity is restored in 8 or 9 days even though DEHP-feeding is continued. The addition of DLcarnitine and cofactors to heart mitochondria from rats fed DEHP for 5 days resulted in a stimulation of 14C-palmitoyl CoA oxidation (Table 12). This suggests that the inhibition is not a result of functional change in the fatty acid oxidation system of the mitochondria.
Adenine nucleotide translocase is one of the mitochondrial anion translocators. It is located in the inner mitochondrial membrane and is responsible for affecting a mole:mole exchange of extramitochondrial ADP for intramitochondrial ATP. It is via this specific mechanism that ATP generated by oxidative phosphorylation can be transported into the cytoplasma to support cytosolic energy-requiring reactions. The addition of DEHP to suspensions of isolated heart mitochondria results in an inhibition of adenine nucleotide translocase. Activity of the transporter was reduced 37% by 300 ,uM DEHP (Fig. 2). Such an effect of DEHP may be the mechanism underlying the death of myocardial cells cultured in the presence of DEHP (45,74) and may 46 F. P. BELL Table 11. Effect of oral ingestion of DEHP on the oxidation of "C-palmitoyl CoA in isolated heart and liver mitochondria.'  aMale Sprague-Dawley rats (225-250 g) were fed Purina Laboratory Chow containing 1.0% DEHP (w/w) for 6 days. Mitochondria were isolated from pools of two hearts, and all assays were performed in triplicate (see footnote to Table 11) in a total volume of 7.0 ml (72). When present, the levels of the various cofactors in the incubation mixtures were: ATP, 15 mg; CoA-SH, 4 mg; NaF, 0.8; MgCl6H20, 3 mg. DL-Carnitine *HCl was added to a final concentration of 2.6 mM. also be related to the decrease in spontaneous heart rate that is observed in rat hearts perfused with DEHP-containing medium (75,76). Despite the fact that DEHP accumulates in the hearts of DEHP-fed rats (50), we did not observe a change in adenine nucleotide translocase activity in mitochondria isolated from hearts of rats fed 0.5% DEHP for up to 10 days (73). The fact that DEHP inhibited adenine nucleotide translocase when added directly to mitochondria, but not when fed to the rats, may indicate that the level of DEHP accumulating in the hearts during the 10-day feeding trials was insufficient to significantly affect the enzyme.

Effect of DEHP Feeding on Platelet Lipid Synthesis and Function in Rats and Rabbits
The well-known fact that blood and blood products become contaminated with DEHP during stor- Rats (150-160 g at beginning of study) received 0.5% DEHP for 32 days and rabbits received 1.0% DEHP for 28 days. All methods used are detailed elsewhere (69). Values are means ± SEM of six animals in each control group, and seven animals in each treated (DEHP) group.
age in PVC plastic storage bags has caused great concern because ofthe uncertainty as to what DEHP may do to blood cell physiology (77). For example, it has been reported that DEHP may cause increased platelet aggregation in blood stored in PVC storage bags (45), increased platelet adhesion to artificial surfaces (78) and decreased survival of platelets stored in the frozen state in PVC bags (79). In view of the above, we studied the function of platelets exposed to DEHP in vivo through a regimen of DEHP feeding. DEHP was fed in the diets of rats (0.5%) and rabbits (1.0%) for 32 and 28 days, respectively. At the end of the feeding period, platelet function was studied. Synthesis of phospholipids and sterols by the platelets was also studied because of the importance of these two lipid classes to membrane structure and stability (80), particularly since platelet aggregation is a membrane-dependent phenomenon. DEHP feeding did not affect phospholipid synthesis from acetate-1-14C, or sterol synthesis from mevalonic-5-3H acid, in platelets from young (6-week-old) or old (5-month-old) rats (69). Additionally, the feeding of DEHP did not significantly (p > 0.05) alter platelet number, collageninduced platelet aggregation or ADP-induced aggregation in either the rabbit or rat (Table 13). Prothrombin time was also essentially unchanged (Table  13). The results suggest that the increased sensitivity of platelets exposed to DEHP in vitro (45,78,79) does not accurately reflect their response to DEHP exposure in vivo.

Conclusions and Suggestions
The data presented here make it abundantly clear that the exposure of animals to phthalate esters can result in a significant perturbation of normal metabolic patterns in liver, heart, testes, adrenal gland and brain and can affect blood lipids (51-53, 55, 59-61, 68, 69, 72, 73, 77). The ability of phthalates to alter lipid metabolism in the developing fetus and in suckling animals is also apparent and becomes a point of particular concern. At the time of this writing, no published biochemical studies similar to those reported here have been conducted in tissues from subhuman primates or man; such studies are urgently needed in order to assess more accurately the potential hazard ofthe phthalates to human health. At the present time, the concern of the medical profession over the exposure of patients to phthalates while undergoing medical treatment has attracted widespread attention in government and industry. Hopefully, the concern over the use of phthalate-containing devices in the field of medicine will attract greater attention to the more important issue of what potential health hazards mankind is exposed to as a result of involuntary chronic exposure to phthalates in the environment. Evidence that chronic exposure of man to DEHP may result in neurological disorders has been reported (81). A further point to be made is a suggestion that pharmacokinetic studies with phthalates should be less preoccupied with tissue distribution and more concerned with subcellular distribution; the reasoning follows. Much of the work described here has been concerned with biochemical systems dependent upon the activity of membrane-bound (particulate) enzymes-e.g., sterol synthesis (HMG CoA reductase), adenine nucleotide translocase, and fatty acid oxidation. Membranebound enzymes, in general, are sensitive to changes in membrane fluidity which reflects alterations in the physical state of the membrane-lipids (82,83). The possibility exists that the effects of phthalates (inhibitory or stimulatory) on various membranebound enzymes is an indirect effect of their ability to modify membrane fluidity. Being lipophilic, the phthalates can partition across organellar membranes and perturb the composition and physical state of the lipid microenvironment (domain) surrounding certain particulate enzymes. Since the response of specific membrane-bound enzymes to changes in membrane fluidity cannot be determined a priori, it is not inconsistent that phthalate exposure should inhibit some enzymes while stimulating others. Con- 48 F. P. BELL sequently, phannacoldnetic studies would yield information of greater potential value if distribution data were obtained for specific cell fractions (endoplasmic reticulum, mitochondria, cytosol, etc.) rather than for total tissue. The growing knowledge of factors regulating particulate enzymes makes it clear that relatively low levels of a chemical agent in tissue may be misleading in terms of toxicity potential, since levels considered to be inconsequential on a tissue or organ basis may be disastrously high on an organellar basis.