MECHANISTIC CONSIDERATIONS FOR HUMAN RELEVANCE OF CANCER HAZARD OF DI(2-ETHYLHEXYL) PHTHALATE

Abstract

Di(2-ethylhexyl) phthalate (DEHP) is a peroxisome proliferator agent that is widely used as a plasticizer to soften polyvinylchloride plastics and non-polymers. Both occupational (e.g., by inhalation during its manufacture and use as a plasticizer of polyvinylchloride) and environmental (medical devices, contamination of food, or intake from air, water and soil) routes of exposure to DEHP are of concern for human health. There is sufficient evidence for carcinogenicity of DEHP in the liver in both rats and mice; however, there is little epidemiological evidence on possible associations between exposure to DEHP and liver cancer in humans. Data are available to suggest that liver is not the only target tissue for DEHP-associated toxicity and carcinogenicity in both humans and rodents. The debate regarding human relevance of the findings in rats or mice has been informed by studies on the mechanisms of carcinogenesis of the peroxisome proliferator class of chemicals, including DEHP. Important additional mechanistic information became available in the past decade, including, but not limited to, sub-acute, sub-chronic and chronic studies with DEHP in peroxisome proliferator-activated receptor (PPAR) α-null mice, as well as experiments utilizing several transgenic mouse lines. Activation of PPARα and the subsequent downstream events mediated by this transcription factor represent an important mechanism of action for DEHP in rats and mice. However, additional data from animal models and studies in humans exposed to DEHP from the environment suggest that multiple molecular signals and pathways in several cell types in the liver, rather than a single molecular event, contribute to the cancer in rats and mice. In addition, the toxic and carcinogenic effects of DEHP are not limited to liver. The International Agency for Research on Cancer working group concluded that the human relevance of the molecular events leading to cancer elicited by DEHP in several target tissues (e.g., liver and testis) in rats and mice can not be ruled out and DEHP was classified as possibly carcinogenic to humans (Group 2B).

1. General overview

Exposure of the general population to DEHP is ubiquitous [1, 2] and occurs due to the use of consumer products, medical devices (blood bags and medical tubing), and intake from food and the environment (air, water and soil). Human data on cancer hazard of DEHP are largely inconclusive, because most studies lack appropriate exposure assessment and only indirect evidence for associations between various cancers and exposure to DEHP has been ascertained. The human database includes studies that infer possible associations of exposure to DEHP or other chemicals present in polyvinyl chloride-containing products with excess mortality from pancreatic [3, 4], testicular [5], and respiratory tract [6, 7] cancers, excess incidence of multiple myeloma [8], as well as increased risk of breast cancer [9] and pediatric hepatoblastoma [10–12].

The carcinogenicity database in animals conclusively shows that DEHP causes cancer of the liver in male and female mice and rats [13–16]. Additional potential target sites for DEHP carcinogenesis have also been reported. In a study in rats, a statistically significant incidence of pancreatic acinar cell adenomas was found in high-dose males [15]. In another study in rats, the incidence of Leydig cell tumors was increased, and was dose-related with early onset [16]. These findings are of potential importance as tumors of the exocrine pancreas are rare spontaneous tumors in experimental animals.

While the primary human health concern of exposure to phthalates is placed on early life exposures to DEHP and associated risks of reproductive and developmental effects, especially in males [17], this review’s focus is on the cancer hazard of DEHP in liver and other tissues. We consider evidence from studies in humans, animals and in vitro experimental systems. Both cancer and relevant non-cancer toxic effects are described, and, where available, studies on the molecular mechanisms of such are emphasized. Even though the scientific debate regarding cancer hazard of DEHP is dominated by the data on liver effects, we include information on other potential target sites in both humans and animals.

2. Toxic effects of DEHP in the liver

Human studies

Several studies assessed potential hepatotoxicity of DEHP in humans. A study of 28 term infants with respiratory failure, 18 of whom received extracorporeal membrane oxygenation, a clinical procedure that results in a high intravenous dose of DEHP, and 10 untreated infants evaluated various clinical parameters of liver, pulmonary and cardiac function. Even though DEHP exposure was estimated to be as high as 2 mg/kg bw over 3–10 days (mean peak plasma concentration, 8 μg/mL), all clinical parameters were found to be unaffected in treated infants [18]. However, a study that compared cholestasis in premature and newborn infants receiving parenteral nutrition through PVC-containing or – free infusion systems found that the use of polyvinylchloride tubing correlated strongly (P = 0.0004) with the development of cholestasis and that the incidence of cholestasis decreased from 50% to 13% after PVC-containing infusion systems were discontinued [19].

The liver cancer hazard of DEHP in humans may be inferred from the studies that suggest that a risk of hepatoblastoma, a rare childhood cancer of the liver, is considerably elevated in children who are born with very low birth weight [10]. Several additional publications have suggested that the duration of neonatal intensive care, which may have involved extensive use of PVC-containing medical devices, is significantly associated with both incidence [12] and severity (i.e., tumor stage) of hepatoblastoma [11].

Due to a very limited database on liver cancer hazard of DEHP in humans, evidence from studies with other chemicals that may act through similar molecular mechanisms may be considered as relevant. On the one hand, fibric acid analogues clofibrate, gemfibrozil and fenofibrate, drugs that act via activation of PPARα, have been used extensively in patients with dislipidemia and risk factors for coronary heart disease [20]. All of the published reports on the health outcomes of chronic administration of these pharmaceuticals in large human cohorts found no elevated risk of mortality from liver cancer [21–23]. In one of the cohorts [24, 25], the excess mortality due to a higher incidence of the malignant neoplasms of the “liver, gallbladder, and intestines” was reported in clofibrate-treated subjects; however, death rates among the clofibrate-treated group for cancer were similar to the official mortality statistics for individuals from the same area, the number of observed cases of gastrointestinal cancers was very small, and there was no difference among groups in a follow up analysis of the mortality trends in this cohort [23]. On the other hand, several ubiquitous environmental contaminants (e.g., chlorinated solvent trichloroethylene) are rapidly metabolized in rodents and humans to molecules that can act as PPARα ligands [26] and exposure to these compounds have been associated with liver toxicity and cancer. In rodents, sufficient evidence exists for hepatotoxicity [27, 28] and hepatocarcinogenicity of trichloroethylene [29]. While some evidence exists of liver toxicity of trichloroethylene in humans, as assessed from liver function and other tests, the data are considered to be inadequate for making conclusions regarding causality due to great difficulty with attributing exposure to a particular environmental chemical to disease [reviewed in detail in [30]]. The evidence is also limited for trichloroethylene and liver and biliary tract cancer [31] mainly because only cohort studies are available; most of these studies have small numbers of cases resulting in wide confidence intervals on the relative risk estimates; and difficulties in establishing exposure-response relationships [detailed descriptions of the individual studies and the meta-analysis may be found in the Environmental Protection Agency’s toxicological review of tricholorethylene publicly available at http://www.epa.gov/iris/subst/0199.htm].

Primate studies

Subchronic liver toxicity has been reported in non-human primates who have been subject to chronic transfusions through DEHP-containing PVC tubing. Abnormal liver function (e.g., bromosulfophthalein clearance) and cholestasis have been reported in rhesus monkeys in chronic experiments that mimicked conditions of patients undergoing repeated blood or platelet transfusions through PVC-containing tubing, an effect absent using polyethylene containers [32]. The average cumulative amount of DEHP infused in 1 year was 69.3 mg, or 21.3 mg/kg/day which the authors found to be comparable or even lower than that in humans on chronic transfusion therapy. A subsequent study in rhesus monkeys by the same group evaluated hepatic function and liver histology up to 26 months after cessation of transfusions [33]. It was reported that abnormal liver function tests and histologic abnormalities (e.g., disturbances of hepatic architecture, the presence of round-cell infiltration, and multinucleated giant cells) in liver biopsies persisted throughout the transfusion study and the follow-up period. While these studies implicate DEHP as a potential toxic ingredient of plastic medical devices, these associations need to be further verified. Similar liver toxic effects of DEHP appear not to be observed in rodents and may be a factor of exposure route (intravenous vs dietary) or comprise another set of important species differences. DEHP (500 mg/kg/day) exposure in young adult male cynomolgus monkeys revealed no distinctive treatment-related effects in the liver [34].

Rodent studies

Liver toxicity (defined here as tissue damage; hepatomegaly and other effects are discussed extensively in other sections below) of DEHP has been reported in some, but not all studies in rats and mice. Dietary administration of DEHP (2%) to mice for 10 days resulted in liver dysfunction as evidenced from histological observations and serum analysis [35]. In a 104-week dietary exposure study to DEHP in B6C3F1 mice [14], an increase in a number of signs of liver damage have been reported. Hepatocyte pigmentation, increased cytoplasmic eosinophilia and chronic inflammation were observed in all male and female mice in the 6000 ppm group (no signs of these histopathological changes were found in control and lower dose groups). In a 104-week dietary feeding study in F344 rats, significant increases in the incidence of Kupffer/hepatocyte pigmentation was reported in male and female rats at 12500 ppm dose, and of spongiosis hepatis in male rats at 2500 and 12500 ppm [15]. Several other studies reported no signs of liver injury [36–38]. Furthermore, even though there are no studies evaluating cholestasis due to DEHP in rats or mice for comparison with human and primate studies, other peroxisome proliferators are known to have a protective effect against cholestasis in rodents. For example, cholestasis was observed in bezafibrate-fed PPARα-null, but not wild type mice [39].

Consistent findings of liver carcinogenicity of DEHP were reported from several independent chronic studies in mice and rats of both sexes [40, 41]. The incidence of liver tumors is known to be dependent both on dose of DEHP and duration of exposure. A study which compared the carcinogenic potency of DEHP in wild type and PPARα-null mice [38] reported that after 24 months of feeding the diet containing 0.05% DEHP, a greater number of adenomas of the liver was observed in PPARα-null mice. Also, one hepatocellular carcinoma and one cholangiocellular carcinoma were observed in PPARα-null, but not wild type mice for this dose group.

3. Hepatocyte-specific effects of DEHP in the liver

The landmark paper by Klaunig et al [42] defined the “PPARα mode of action” and concluded that it exclusively mediates the carcinogenicity of DEHP and other related compounds. A number of reviews have considered the evidence for receptor-mediated and –independent molecular events in the development of liver and other tissue tumors induced by peroxisome proliferators, including DEHP, as well as their relevance to human health assessment [43–45]. These various perspectives have contributed greatly to the understanding of the issues pertinent to the interpretation of rodent data on DEHP and related chemicals to characterize the potential for human cancer hazard.

3.1. Effects of DEHP on PPAR activation

Phthalates, including DEHP, are activators of nuclear receptors PPARα, β and γ [46–49].

Human studies

No studies could be found which showed evidence that DEHP activates PPARα in human liver. In vitro, trans-activation assays were used to assess the activation potential of DEHP, MEHP and 2-ethylhexanoic acid for either the full length PPAR subtype [50, 51] or hybrid transcription factors consisting of the PPAR ligand binding domain cloned to the glucocorticoid receptor [52] or GAL4 [53] DNA binding domains. DEHP was not able to activate human PPARα or human PPARγ [50]; however, several studies showed that all three human PPAR subtypes were activated by MEHP [51–53]. In addition, both human PPARγ isoforms γ1 and γ2 were activated by MEHP [54]. The DEHP metabolite 2-ethylhexanoic acid weakly activated human PPARα but not human PPARγ [50]. No studies have examined the activation of human PPARβ by DEHP or 2-EH.

Lapinskas et al. [49] determined whether phthalate esters interact directly with human PPARα or γ using the scintillation proximity assay. The K_i for the MEHP to bind to the α and γ subtypes were 15 uM and 12 uM, respectively. DEHP was negative in this assay (i.e., K_i > 150 uM). The PPARβ subtype was not examined. In a recent study [55] DEHP was examined for the ability to activate several response elements (cis assays) in a multiplex human response element transactivation assay in the human HepG2 cell line. In addition, a modification of the approach was used to generate the human ligand-binding domain of nuclear receptors expressed as a chimera with the yeast GAL4 DNA-binding domain that activated a 5X-UAS_G-TATA promoter linked to a reporter sequence (trans assay) [56]. DEHP was shown to significantly activate PPARα and PPARγ: PPARγ_TRANS (EC₅₀=46 μM), PPRE_CIS (EC₅₀=48 μM), PPARα_TRANS (EC₅₀=50 μM). PPARβ TRANS assay was negative for DEHP.

Experimental systems

In vivo, DEHP has little effect on mRNA levels of PPARα in rats or mice [57]; however, the increase in PPAR transactivation activity by DEHP and its metabolites is well established. The DEHP metabolite MEHP activates mouse PPARα, mouse PPARγ and to a lesser extent, mouse PPARβ in trans-activation assays [50–52, 54]. In addition, both mouse PPARγ isoforms γ1 and γ2 were activated by MEHP [54]. 2-ethylhexanoic acid (2-EH) weakly activated mouse PPARα, β and γ [49, 52]. In one study 2-EH failed to activate mouse PPARγ [50]. DEHP itself is a weak activator of PPARα and PPARγ although it cannot be ruled out that the active species was a metabolite of DEHP [49, 52]. DEHP did not activate mouse PPARβ [49, 52].

Species comparisons

Important species differences in expression and molecular signalling for PPARα have been reported. Mice and rats express PPARα at high levels in liver, whereas PPARα is expressed at a lower level in human liver [58]; however, expression of hPPARα has not been determined in a sufficient number of samples to unequivocally conclude that all populations express less PPARα than responsive rodents. In one study the expression of hPPARα protein in one of six humans appeared to approach the levels found in mouse livers [59], yet the hPPARα protein expression in this study was not normalized to housekeeping or loading controls that would help to evaluate whether the proteins in the sample were intact. In a side-by-side comparison of expression in liver, mice had ~10-fold more PPARα mRNA expression than guinea-pigs and ~3-fold more than Syrian hamsters [60]. No differences in constitutive PPARα mRNA expression was reported between CD-1 mice and Sprague-Dawley rats, but a lower level of expression (~25% the level of mouse) was found in marmosets (Calithrix jacchus) [57].

There is evidence that human and rodent PPARα differ in their ability to be activated by PPARα agonists, as would be expected given that the amino acid sequences within the ligand binding domains differ between species. The mouse and rat PPARα ligand binding domains are 94% similar to that of the human PPARα [48, 61]. Higher concentrations of MEHP are required for activation of human PPARs to the same levels as the corresponding mouse receptor [50, 51, 53].

A truncated human PPARα variant has been reported, and it was suggested that the truncated form may be responsible for differential responses between species. This truncated form, identified in several laboratories and called hPPARα-8/14 [62], hPPAR_SV [58], PPARα_tr [63], and PPARα2 [64], lacks exon 6 due to alternative splicing, resulting in an altered hPPARα lacking the ligand binding domain. In in vitro transactivation assays, this form acts as a dominant negative, inhibiting the ability of the wild-type receptor to activate transcription, possibly by titrating out limiting amounts of co-activators such as C/EBP binding protein/p300 [63]. The level of the truncated hPPARα mRNA ranges from 10% to 50% of full-length hPPARα mRNA [58, 63, 64]. One study concluded that the level of the truncated form does not correlate with responsiveness to PPARα agonists [64]; however, this study only measured primary human hepatocyte mRNA levels of the ACO gene, not an optimal biomarker of hPPARα activity given that this gene does not appear to be regulated in the same manner as the rodent gene by PPARα [180].

3.2. Effects of DEHP on markers of PPAR activation

PPARα activation in the liver may also be characterized indirectly by assessing: (1) increases in the size and/or numbers of peroxisomes in cells; (2) increases in acyl-CoA oxidase (ACO) expression, protein or activity; (3) increased expression, protein levels or activity of CYP4A protein, an ω-lauric acid hydroxylase (LAH); and/or (4) increases in the levels of carnitine acyl-CoA transferase (CAT). Some of these markers have been shown to be activated even in the absence of PPARα, i.e., in PPARα-null mice exposed to peroxisome proliferators [65]; thus, the results from these indirect assays should be interpreted with caution if no direct evidence for changes in PPARα activity is available.

Human studies

No study assessed markers of PPARα activation after administration of DEHP to humans; however, studies are available from people who may have been exposed to DEHP and other agents leaching from plastics used in medical devices. Dialysis patients were evaluated for evidence of liver peroxisome proliferation in biopsy samples [66, 67]. Based on a subjective ultrastructural evaluation of one subject, no effect was seen after one month of dialysis. However, in a liver biopsy from one subject after 12 months of dialysis, an increased number of peroxisomes was reported. Others have suggested that more cautious evaluation, including objective measurements, increased numbers of biopsy intervals, and appropriate controls, would be needed to determine conclusively whether peroxisome proliferation due to DEHP occurs in dialysis patients [68].

Experimental systems

In young adult male cynomolgus monkeys exposed to DEHP (500 mg/kg/day) by intragastric intubation for 14 consecutive days, no effect on body weight, liver weights, histological observations in the liver, kidney, or testes, or markers for peroxisomal proliferation were observed [34]. When male and female marmosets (5–6 per group) were exposed to DEHP (0, 100, 500, or 2500 mg/kg) by daily oral gavage for 65 weeks from weaning (3 months age) to sexual maturity (18 months), no changes were observed in cyanide-insensitive palmitoyl-CoA β oxidation in males [69]. However, large variability in activity of some peroxisomal enzymes was observed in treated females, especially at the 500 mg/kg dose. For lauric acid ω-1-hydrolase activity, females exhibited dose-related increases in comparison to controls that were statistically significant at 500 mg/kg.

In the rat, DEHP increases relative liver weights, markedly induces activity of hepatic peroxisomal enzymes, and increases microsomal cytochrome P450 content [70, 71]. Ultrastructural examination showed marked peroxisome proliferation and a dilation of the smooth and rough endoplasmic reticulum. DEHP is a hypolipidemic agent, as serum triglyceride levels were reduced. In mice, DEHP had effects similar to those found in rats, albeit slightly weaker in magnitude [71].

The effects of DEHP in the rat and mouse were dose-dependent [41, 72]. Male and female Fischer 344 rats and B6C3F₁ mice were fed DEHP for up to 13 weeks [41]. In rats fed 12500 ppm DEHP, there was an increase in hepatic peroxisomal β-oxidation activity after one, two and 13 weeks administration. In mice fed 10000 and 17500 ppm DEHP, there was an increase in hepatic peroxisomal β-oxidation activity after one, two and 13 weeks administration. In mice fed 1000 ppm DEHP, there was no statistically significant increase in hepatic peroxisomal β-oxidation activity after one, two or 13 weeks administration. Peroxisomal enzyme induction was also observed in the kidney after exposure to DEHP [73].

A correlation between the enzymatic marker of the peroxisomal fatty acid β-oxidation cycle and changes in peroxisome morphometry is a well-established phenomenon, demonstrating that peroxisomal cyanide-insensitive palmitoyl-CoA oxidation is a good marker for peroxisome proliferation in rodent liver [72]. However, DEHP has been shown to affect expression of genes other than those involved in peroxisomal or microsomal metabolism. Microarray analysis of male C57BL/6 mice treated with DEHP (1.0% in the diet) for 13 weeks identified 51 DEHP-regulated genes in the liver involved in peroxisome proliferation, xenobiotic detoxification, oxidative stress response, immune function, steroid hormone metabolism, testis development, and pheromone transport [74]. In Currie et al. [75], B6C3F1 male mice were treated with DEHP (1150 mg/kg/day, daily oral gavage) and gene expression levels were measured 2, 8, 24, and 72 hours after the initial dose. A robust increase in Cyp4a10 expression was observed from the earliest time point sampled. A coordinate induction of genes involved in fatty acid metabolism was confirmed by the overrepresentation of Gene Ontology terms and PANTHER and GenMAPP pathways involved in metabolism of lipids (e.g., acyl-CoA metabolism, fatty acid β-oxidation, pantothenate and CoA biosynthesis, or coenzyme metabolism).

Dostal et al. [76] compared effects of DEHP in neonatal, adolescent and adult rats. Groups of male Sprague-Dawley rats were given 0, 10, 100, 1000 or 2000 mg/kg/day DEHP by gavage for 5 days, beginning at the age of 6, 14–16, 21, 42 or 86 days. After two doses of 2000 mg/kg/bw per day, virtually all pups in the three youngest age groups died, whereas six- and 12-week-old rats showed significantly decreased body weight but no fatalities. Five daily doses of 1000 mg/kg/bw per day caused significant decreases in body weight gain in one-, two- and three-week-old rats. Absolute and relative liver weights were significantly increased at 100 mg/kg/day in all age groups (except for 1-week-old rats) and in all age groups at higher dose levels. Morphological examination revealed increased peroxisome proliferation in neonatal as well as in adult rats. The activities of cyanide-insensitive palmitoyl-CoA oxidase and carnitine acetyltransferase were increased in a dose-dependent manner in all age groups. The activities of these enzymes were similar in control rats of all ages.

Transfer of DEHP through milk was evaluated in female Sprague-Dawley rats given a total of five oral doses of 2000 mg/kg/day by gavage on days 2–6, 6–10 or 14–18 of lactation [77]. The body weights of lactating rats and of their suckling pups were significantly reduced at all time points. Food consumption was reduced in the dams dosed on days 14–18. Relative liver weights were increased in the lactating dams at all three stages of lactation but not in the suckling pups. The hepatic peroxisomal enzyme activities (cyanide-insensitive palmitoyl-CoA oxidase and carnitine acetyltransferase) were increased five- to eight-fold in treated dams at all three stages of lactation. Two-fold increases in these enzyme activities were also observed in pups suckling the treated dams. Hypolipidaemia was observed in treated lactating rats at all 3 stages of lactation. In a study that used Fischer 344 rats, exposure of lactating females to DEHP (0, 0.5, 1, 2.5 or 5 g/kg per day) by oral gavage from birth through lactation day 21, similar effects in the pups and dams were observed [78]. No pup survived exposure to doses of 2.5 g/kg per day. Pup growth was impaired at the two lowest doses. In the liver, cyanide-insensitive palmitoyl-CoA oxidase activity was increased to a similar extent in both pups and adult females.

A 13-week study with DEHP (100, 500 or 2500 mg/kg/day, gavage) in male and female marmosets significant suppression was observed of body weight gain in male marmosets given 2500 mg/kg/day [79]. No effect was found on relative liver weight, morphological features in the livers, hepatic cyanide-insensitive palmitoyl-CoA oxidation, or carnitine acetyltransferase activity. Ultrastructural examination of the liver did reveal small increases in mean peroxisomal volume in male marmosets given 500 or 2500 mg/kg DEHP.

Induction of peroxisomal enzymes by DEHP metabolites was also observed in vitro with isolated cultured rat and mouse hepatocytes. MEHP and EH increased carnitine acetyltransferase activity several fold in hepatocytes from male Wistar or Sprague-Dawley rats [80, 81]. Another DEHP metabolite, mono(2-ethyl-5-oxohexyl)phthalate, was shown in vitro to inhibit fatty acid oxidation at the site of transport across the mitochondrial inner membrane with a marked species difference. This finding supports the idea that induction of peroxisome proliferation could be due to an initial biochemical lesion of fatty acid metabolism [82].

Species comparisons

A number of studies conducted a direct species comparison of the effects of DEHP on markers of PPARα activation. A comparative study of rats, mice and guinea-pigs that were fed DEHP (2%) for two weeks found that hepatic cyanide-insensitive palmitoyl-CoA oxidation activity was significantly increased in rats and mice, but not in guinea-pigs [83]. Similar observations of non-responsiveness of guinea pigs as compared to rats were reported in a 4-day study with DEHP (950 mg/kg/day, gavage) [84]. Significant increases in liver weights and hepatic β-oxidation activity were observed in the rats but not the guinea-pigs. Administration of DEHP (25–1000 mg/kg/day by gavage) for 14 days to male Sprague-Dawley rats and male Syrian hamsters produced dose-related increases in relative liver weights and activites of both hepatic cyanide-insensitive palmitoyl-CoA oxidation and carnitine acetyltransferase in both rats and hamsters, yet the hamsters were much less responsive [85]. In the same study, rats and hamsters were treated with MEHP (500 mg/kg bw/day). Similar to the difference observed with DEHP, MEHP produced a greater increase in relative liver weight and a greater stimulation of enzyme activities in rats compared to hamsters.

There are several published reports comparing in vivo responses to DEHP in rats or mice with non-human primates. DEHP (2000 mg/kg bw/day, gavage) was administered for 14 days to male and female Wistar rats and male and female marmosets [86]. Marmosets were also treated with DEHP (1000 mg/kg bw/day) by daily intraperitoneal injection for 14 days. In rats, there were increases in relative liver weight, ω-lauric acid hydroxylation and hepatic peroxisome proliferation, as demonstrated by ultrastructural peroxisome examination and increased cyanide-insensitive palmitoyl-CoA oxidation and. No such effects were observed in marmosets given DEHP by either oral or intraperitoneal administration. In a recent study, male mice (CD-1), rats (Sprague-Dawley), and marmosets (Callithrix jacchus) were treated with DEHP by gavage [57]. Mice and rats were treated with 0, 1.25 and 2.5 mmol/kg DEHP for 2 weeks, while marmosets were exposed to 0, 0.25, 1.25 and 6.25 mmol/kg DEHP for 15 months. Liver to body weight ratios increased in rats and mice in both treatment groups. No effect on relative liver weight was observed in marmosets. Liver concentrations of MEHP (measured at the end of the treatment period) increased in a dose-dependent manner in all species tested, yet it was ~3-fold lower in the marmoset compared to rats and mice when similar dose-groups were compared. Induction of liver peroxisomal enzymes was detected in DEHP-treated rats and mice, but not marmosets. A study that compared male Fischer 344 rats and cynomolgus monkeys treated with DEHP (rats: diet with 100–25000 ppm for 21 days; monkeys: daily gavage with 100 or 500 mg/kg bw for 25 days) reported a dose-related increase in relative liver weights, and enzymatic markers and ultrastructural evidence (subjective evaluation) of hepatic peroxisome proliferation in rats [87]. No effect on relative liver weights, activities of cyanide-insensitive palmitoyl-CoA oxidation, carnitine acetyltransferase and lauric acid 12-hydroxylase, or abnormalities in light or electron microscopic examination of liver sections were observed in cynomolgus monkeys.

In cultured hepatocytes, effects of DEHP metabolites were also compared across different species. While rat hepatocytes respond to treatment with MEHP by concentration-dependent induction of cyanide-insensitive palmitoyl-CoA oxidation, no such effect was observed in guinea-pig [88–90], Syrian hamster [89], New Zealand rabbit [90], or cynomolgus monkey [90] hepatocytes. Treatment with either 2-ethylhexanol or 2-ethylhexanoic acid for 72 hours produced a concentration-dependent induction of cyanide-insensitive palmitoyl-CoA oxidation in rat and mouse, but not guinea-pig or marmoset hepatocytes [91]. Goll et al. [92] examined the effects of various peroxisome proliferators including DEHP on peroxisomal enzyme activities, apoptosis and DNA synthesis in rat FaO and human HepG2 hepatoma cell lines. Both growing and confluent cultures were treated for 48 or 72 hours. DEHP increased peroxisomal enzyme activity in rat FaO cells but not in human HepG2 cells. In Hasmall et al. [84], rat hepatocytes in vitro, treated with MEHP (250, 500 and 750 μM) exhibited increased peroxisomal β-oxidation. In contrast, there was no response of human hepatocytes at these concentrations.

3.3. Effects of DEHP on PPAR-independent receptor-mediated events

Human studies

All of the available studies have been performed using human cells in vitro.

The estrogenic activity of DEHP was investigated in competitive ligand-binding assays, yeast and mammalian gene expression assays, and in a uterotrophic assay. DEHP did not compete for estrogen receptors (ER) or induce luciferase activity in transfected MCF-7 cells or stably transfected HeLa cells [93]. Ghisari and Bonefeld-Jorgensen [94] investigated in vitro estrogenic activities of a range of widely used plasticizers and phenols, including DEHP. The estrogenic activities of the compounds were assessed in human breast cancer MVLN cells, stably transfected with an estrogen receptor element luciferase reporter vector. Furthermore, the combined effect of a multicomponent mixture of six plasticizers was evaluated for its estrogenic activities. The chemicals were tested in an ER trans-activation assay alone and upon co-treatment with 25 pM E2. DEHP antagonized the 17β-estradiol induced ER function (EC₅₀=84 μM). No significant activity was observed for DEHP in the ER trans-activation assay. Takeuchi et al. [95] characterized the activities of human estrogen receptor α (hERα) and human estrogen receptor β (hERβ) in the presence of 22 phthalates, including DEHP and MEHP, using reporter gene assays. DEHP, but not MEHP activated hERα. Neither DEHP nor MEHP activated hERβ. DEHP but not MEHP had antagonistic effects on ERβ. Kang & Lee [96] reported increased expression of ERβ after DEHP exposure (3.0 μg/ml) of the human breast cancer cell line MCF-7.

DEHP and its metabolite MEHP do not bind to the human androgen receptor (AR) [97]. It was also reported that neither DEHP nor MEHP possessed human AR antagonism [95]. DEHP did not affect human AR activation using chemically activated luciferase gene expression bioassays in transiently transfected Chinese Hamster Ovary (CHO-K1) cells [98].

DEHP was shown to activate human pregnane X receptor (PXR)-mediated transcription [99]. Transfection assays were performed with a human PXR expression plasmid and a reporter plasmid containing the XREs in the CYP3A4 gene promoter in HepG2 cells. This study indicates that DEHP may be an inducer of the CYP3A4 gene through PXR, and may influence the metabolism of endogenous steroids, drugs, and other xenobiotics. The same group also tested the hypothesis that leaching of DEHP during parenteral chemotherapy for cancer patients may facilitate PXR mediated MDR1 expression in various tissues, including cancer cells, promoting drug resistance [100]. The effect of DEHP on PXR-mediated transcription of the MDR1 gene was studied in the human colon adenocarcinoma-derived cell line, LS174T, which endogenously express PXR. DEHP increased PXR-mediated transcription of the MDR1 gene in luciferase-reporter assays. The induction by DEHP was abrogated when a reporter plasmid containing a mutated DR+4 motif in the XRE was used. In a mammalian two-hybrid assay, DEHP recruited steroid receptor co-activator-1 to the ligand-binding domain of PXR. Using real-time reverse transcriptase-PCR, DEHP increased MDR1 gene expression in a dose-dependent manner. The data support activation of the MDR1 gene by DEHP through PXR. The ability of DEHP and MEHP to induce PXR-mediated transcription of the CYP3A4 promoter in a dose-dependent fashion was also shown using HepG2 cells [101]. Co-exposure to either MEHP or DEHP and dexamethasone (Dex) resulted in enhanced CYP3A4 promoter activity. This induction was abrogated by both the GR antagonist RU486 and GR small interfering ribonucleic acid. CYP3A4 protein was highly inducible by Dex and DEHP co-administration in human primary hepatocyte cultures. Enhanced 6β-hydroxytestosterone formation in Dex and DEHP or MEHP co-treated human primary hepatocytes confirmed CYP3A4 enzyme induction. DEHP was shown to activate human PXR (EC₅₀=2.5 μM) in stably transfected cells, HGPXR cells, derived from HeLa cells and expressing luciferase under the control of a chimeric hPXR [102]. DEHP was also able to induce CYP3A4 expression in two preparations of primary cultured human hepatocytes at 10 μM (8- and 38-fold).

The ability of DEHP and/or MEHP to activate human constitutive androstane receptor (CAR) and induce expression of its target genes has been shown in a number of studies. DEHP was able to induce CYP2B6 expression, an indication of CAR activation in two preparations of primary cultured human hepatocytes at 10 μM (2- and 4-fold). In humans, a novel splice variant of CAR (CAR2) results in a 4 amino acid insertion in the ligand binding domain. The CAR2 comprises approximately 30% of the total CAR transcript levels in human hepatocytes [103]. The CAR2 transcripts are not generated in mice, rats, and marmosets. CAR2 was shown [104] to be potently activated in cells transfected with CAR2 by DEHP (0.098μg/ml) and weakly by MEHP (2.8 μg/ml). In the same study, primary human hepatocyte cultures from 3 donors, cultured in such a way to preserve differentiation, had very different responses to CAR-responsive induction of CYP2B6 and CYP3A4. The DEHP concentrations that induced these enzymes (in the absence of cytotoxicity) was 3.9 ×10⁻³ μg/ml, 19.5 μg/ml, and 0.39 μg/ml for the 3 donors. In addition, induction of CYP2B6 gene expression indicative of CAR activation in primary human hepatocyte cultures (culture conditions not cited) was increased in cultures from two patients exposed to 19.5 μg/ml DEHP [105]. In the multiplex transcription factor assay described above [56], DEHP significantly activated human PXR (PXRE_CIS (EC₅₀=37 μM), PXR_TRANS (EC₅₀=38 μM)) and human CAR (CAR_TRANS (EC₅₀=50 μM)). In a study that used a pooled human hepatocyte sample from multiple donors, DEHP was tested for its ability to increase transcription of several genes associated with activation of PPARα, CAR, PXR, FXR and AhR [55]. DEHP significantly increased the expression of only the CAR responsive gene CYP2B6 at the 3 time points (6, 24 and 48 hrs) studied. The interaction of alternatively spliced human CARs and PXR with a range of suspected endocrine disruptors, including DEHP, was studied using COS-1 cells [106]. DEHP had EC₅₀ values for activation of CAR2 and PXR of 0.1 and 3.8 μM. DEHP did not activate another CAR isoform (CAR3) which possesses a 5 aa insertion in the ligand binding domain. Mutation analysis of CAR2, in silico modeling, and ligand docking studies suggested that the SPTV amino acid insertion of CAR2 creates a unique ligand-binding pocket that selectively recognizes phthalates.

DEHP was shown to weakly induce human aryl hydrocarbon receptor (AhR) activity, reaching 1.75-fold above the solvent control at the highest concentration tested (100 μM) using chemically activated luciferase gene expression bioassays in recombinant mouse Hepa1.12cR cells [98]. The ability of MEHP to induce CYP1A1, a known AhR target gene, was studied in human cell lines CaCo-2 (derived from colon adenocarcinoma), HepG2 (hepatoma), A549 (lung adenocarcinoma), as well as primary human keratinocytes [107]. The induction specifically involved PPARα and required 2 PPRE sites that were located within the CYP1A1 promoter. Whether this also occurs in normal human hepatocytes is yet to be established. CYP1A1 is responsible for bioactivation of several environmental pro-carcinogens, and it was reported that DEHP and other peroxisome proliferators enhance DNA adduct formation in rat liver when co-administered with benzo[a]pyrene [108]. Thus, the fact that MEHP is capable of regulating CYP1A1 in PPARα–mediated manner in human cells may have significant implications; however, others have shown that another peroxisome proliferator, clofibrate, inhibits expression of CYP1A1 and CYP1A2 in rat liver cells [109].

Experimental systems

Tomonari et al. [69] treated male and female marmosets daily with 0, 100, 500, or 2500 mg/kg DEHP by oral gavage for 65 weeks from weaning (3 months age) to sexual maturity (18 months) and examined serum hormone levels and several enzyme activities. Although the authors note no obvious treatment related changes in testosterone levels in any treatment groups, there were large variations in individual values (serum testosterone in males varied over 100-fold). Significantly higher estradiol levels were recorded in all female marmosets by week 65 in the 500 mg/kg group. For females, increased ovarian and uterine weights were observed in higher dose groups (500 and 2500 mg/kg body weight). For testis, the authors report that there were no treatment related differences in GSH content, or differences in the activities of SDH (sorbitol dehydrogenase), γ-GT (γ-glutamyl transpeptidase), and GSH-Px (glutathione peroxidase). There was a decrease of SDH activity of ~15–20% for the dosed groups that was not statistically significant but the low number of animals limited the power of the study to detect a change. There was also a similar magnitude of increased γ-GT that was not statistically significant and a ~30% decrease in total glutathione (GSH and GSSG) was also observed in all groups that was not statistically significant. However, GST activity and zinc content in the 100 and 500 mg/kg treated groups were significantly reduced. For testosterone 6β-hydrolase activity (CYP3A) in the liver, females were reported to have increased levels at 500 and 2500 mg/kg exposure levels but only the changes observed at the 2500 mg/kg dose group were statistically significant. This activity was increased in all dose groups for males but was not statistically significant (~85% at 100 mg/kg dose for nmol/g liver/min).

An effect of DEHP on estrogen metabolism has been reported [110]. Male Fischer 344 rats fed diets containing 1.2% DEHP for 4, 8 or 16 weeks had significantly increased serum estradiol levels. This was explained by the observation that these rats showed significant loss of hepatic activity of a major male estrogen-metabolizing enzyme, estrogen 2-hydroxylase, and a male-specific estrogen-sequestering protein.

Wong et al. [74] reported a treatment-related suppression of hormone metabolizing 3β-hydroxysteroid dehydrogenase V (HSD3b5), in the livers of male C57B6 mice exposed to DEHP. Northern blots showed HSD3b5 mRNA levels decreased dramatically upon one day exposure to 2.0% dietary DEHP and were nearly undetectable after one week of treatment. Another mouse 3β-hydroxysteroid dehydrogenase (HSD3b4), predominantly expressed in kidneys was also down-regulated by DEHP in the liver. After 4 weeks of exposure to DEHP, HSD3b5 and HSD3b4 mRNAs were reduced in a dose-dependent manner. DEHP treatment had a greater effect on HSD3b5 levels than HSD3b4.

Recent in vivo studies in PPARα-null, PPARα humanized transgenic, and hepatocyte-specific constitutively activated PPARα transgenic mice suggest that several PPARα-independent molecular signals and multiple pathways may be activated by peroxisome proliferators, including DEHP.

Transgenic mice with hepatocyte-specific constitutively active PPARα were developed and observed for up to 11 months of age [111]. In the absence of treatment, these transgenic mice exhibited various responses that mimic wild-type mice treated with peroxisome proliferators. These effects included a significant decrease in serum fatty acids, and numerous liver effects: hepatomegaly, hepatocyte hypertrophy, increased rate of cell proliferation, marked induction of PPARα target genes encoding fatty acid oxidation enzymes and increased accumulation of triglycerides. Even though these phenotypic changes were similar (and of comparable magnitude) to those induced in wild-type mice by a peroxisome proliferator WY-14,643 (0.1% w/w diet), no liver tumors were detected in untreated transgenic mice by 11 months of age even though all WY-14,643-treated wild-type mice developed liver tumors.

Another study used hepatocyte transplantation to generate chimeric livers composed of PPARα-null and -positive hepatocytes in PPARα-null or -positive mice [112]. The authors examined the relationship between PPARα status and the hepatocyte’s ability to proliferate in response to WY-14,643 in vivo. When treated with WY-14,643 for 7 days, both PPARα-null and -positive hepatocytes in chimeric livers displayed elevated DNA synthesis regardless of host receptor status, as long as at least some hepatocytes contained the receptor. These findings suggest that the mitogenic response to peroxisome proliferators does not require the presence of active PPARα in all hepatocytes.

Two studies used a transgenic mouse line that overexpresses human PPARα in a PPARα−null mouse [113] also called PPARα humanized mice. These mice have been used for subchronic studies with WY-14,643 and fenofibrate [113], and a chronic feeding study with Wy-14, 643 [114]. In these studies, PPARα-humanized mice did not exhibit hepatocellular proliferation, or liver tumors when treated with peroxisome proliferators; however, induction of typical markers of fatty acid β-oxidation were observed. These mouse models have not been evaluated with DEHP.

PPARα-null mice and corresponding wild-type mice (C57BL/6J strain) were treated with DEHP (0, 20 or 200 mg/kg day) for 21 days by gavage [105]. While this study did not report on liver weight or liver histopathology, gene expression profiling was performed on liver tissues and the authors reported that several prototypic CAR target genes were induced by DEHP in PPARα-null mice. In addition, Cyp1a1, a marker gene for AhR activation, was increased in PPARα-null mice but not wild-type mice.

Fan et al. [115] reported that in male SV129 wild type and PPARα-null mice fed DEHP (0.6%) in diet for 3 weeks, the expression of mouse 6α-testosterone hydroxylase Cyp3a11 gene was increased in the liver. The increase by DEHP was PPARα-independent and did not match the effects of WY-14,643 in the same study.

A 2-year dietary feeding study with DEHP (0.01 and 0.05%) in wild type and PPARα-null mice reported that DEHP induced liver tumors independent of PPARα status [38]. Wild- type and PPARα-null mice did not exhibit equivalent levels of tumor induction. There were no statistically significant increases in liver tumors in the wild-type mice under these exposure conditions, indicating that the biological effects of exposure may have not been equivalent in these two strains. The authors reported that inflammation and oxidative stress markers were affected differently in wild type and PPARα-null mice, suggesting a possibility that different pathways were operating in two strains. The authors indicate that the increase in liver tumors could be related to the higher incidence of spontaneous liver tumors in control PPARα-null mice compared to control wild-type mice that was observed in an earlier study [116].

A follow up study [117] conducted transcript profiling to examine gene expression in the DEHP-induced tumors from the wild-type or PPARα-null mice. Microarray analysis was performed using what appeared to be only one tumor or normal untreated tissue from each strain precluding the opportunity to perform routine statistical analysis. The authors report the gene expression by RT–PCR of 5 genes involved in carcinogenesis. They compared the expression between normal untreated liver (n = 7) and liver tumors (n = 3). The liver tumors could be from both control or treated groups. They found significant increases in the expression of tumor versus controls in wild-type mice for Gadd45a and Apaf1 and significant increases in tumor versus controls in PPARa-null mice for Cyclin B2 and Mcl1. However, these differences are difficult to interpret because they could be due to treatment with DEHP or to differences between the tumors and surrounding tissue or both.

Eveillard et al. [105] examined whole liver gene expression in wild-type and PPARα-null C57Bl/6 mice exposed to 0, 20, or 200 mg/kg DEHP for 21 days (n = 10/group). Using a nuclear receptor dedicated cDNA macroarray 284 genes with measurable expression levels were initially retained for further analysis. About one-third (88 genes), displayed a significant regulation in at least one of the experimental conditions (genotype or treatment effects). Fatty acid homeostasis and xenobiotic metabolism were the most represented pathways. 56 transcripts were differentially expressed between wild-type and PPARα-null control mice. DEHP (high dose) altered the expression of 49 transcripts in wild-type mice. 16 genes were altered in PPARα-null livers. Four genes were similarly altered in both strains including acyl-Co A oxidase 1 (Acox1), aldehyde dehydrogenase family 1 subfamily 1a1 (Aldh1a1), aminolevulinic acid synthase 1 (Alas1) and cytochrome P450 2c29 (Cyp2c29). Thus, most of the genes were altered by DEHP in wild-type but not PPARα-null mice.

A meta-analysis of transcript profiles of livers from rats treated with nuclear receptor activators identified PPARα and CAR as targets of DEHP [118]. This study also reported on a transcriptional analysis in wild-type mice and mice nullizygous for these two nuclear receptors. Microarray analysis showed an overlap in the profiles of DEHP, valproic acid and clofibrate treated rats with a classical activator of CAR, and to a lesser extent, an activator of PXR. The overlapping genes included CYP gene families that are often considered signature genes for nuclear receptor activation. In the same study, groups of wild-type, PPARα-null and CAR-null mice were treated with 0, 200, or 1150 mg/kg DEHP in corn oil by daily gavage for 4 days. A microarray gene expression comparison of DEHP-treated wild-type and PPARα-null mice revealed that PPARα is required for ~94% of all transcriptional changes in wild-type mice. The remaining 6% of the genes were dominated by those involved in xenobiotic metabolism and are known target genes of CAR or PXR, and cholesterol biosynthesis which are regulated by several transcription factors including RXR. Cyp2b10, Cyp3a11 and Cyp3a41a, as well as Mt1 were induced by DEHP partially or completely dependent on CAR, but not PPARα as determined by comparison of effects in wild-type mice and mice nullizygous for PPARα or CAR. The expression of the CAR gene itself was increased by DEHP in PPARα-null but not in wild-type mice. Several putative CAR and PXR targets exhibited PPARα- and CAR-independent induction, including Cyp8b1, Gstm4, and Gstm7. It was concluded that DEHP requires CAR for induction of a small subset of genes (compared to PPARα) and that some liver transcriptional effects may be PPARα-independent.

Kim et al. [119] examined the effects of DEHP on nuclear receptor expression and phospholipase D (PLD), an enzyme that catalyses the hydrolysis of phosphatidyl choline (PC) to generate phosphatidic acid (PA) and choline. PLD is believed to play an important role in cell proliferation, survival signalling, cell transformation, and tumor progression. DEHP (500 mg/kg/d), was orally administered, daily to prepubertal rats (4 wks of age) for 1, 7, or 28 days. In this study, protein expression levels of PLD1/2, peroxisome proliferator-activated receptor α (PPARα), and cytochrome P-450 were determined by Western blot analysis using specific antibodies. Liver weight was significantly increased in the DEHP treatment groups. A significant rise in PLD1/2 expression was observed in the livers of DEHP-exposed rats after 7 days. The authors state that PPARα, CAR, PXR, and CYP2B1 protein expression levels were markedly elevated in DEHP-treated groups.

Several in vitro studies examined the ability of DEHP or MEHP to activate rat or mouse nuclear receptors other than PPARα. CAR trans-activation assays using mouse CAR in HepG2 cells showed that MEHP increased CAR activity greater than 3.4-fold at 100μM [120]. Effects of DEHP on rat androgen, estrogen and thyroid receptors were evaluated in reporter assays [121]. DEHP not only exhibited antiandrogenic activity (IC₅₀>0.1 mM), but also showed androgenic activity (EC₅₀>0.1 mM), thyroid receptor antagonist activity (IC₅₀>0.1 mM) but not thyroid receptor agonist activity. In the estrogen receptor-mediated reporter gene assay, DEHP showed no agonist activity. The thyroid hormone disrupting potential of DEHP was studied by the effect on the thyroid hormone-dependent rat pituitary GH3 cell proliferation [94]. DEHP significantly affected the GH3 cell proliferation below levels that were cytotoxic.

Species comparisons

In Parmar et al. [122] administration of 2000 mg/kg of DEHP for a period of 7 or 15 days in rats, mice, guinea-pigs and rabbits produced a differential effect as judged by alterations in body weight gain, liver weight and activities of mixed function oxidases. DEHP exposure for 7 days caused an increase in the activity of aniline hydroxylase (CYP2E1), arylhydrocarbon hydroxylase (CYP1A) and ethylmorphine N-demethylase (CYP3A) in rats, mice and guinea-pigs, but a decrease in that of rabbits. However, exposure for 15 days produced smaller increases in the activity of these enzymes in rats and mice and produced a decrease in the activity of these enzymes in guinea-pigs.

Hurst and Waxman [123] investigated species-specific effects of phthalates on PXR using a HepG2 cell reporter assay following transfection with mouse PXR (mPXR), human PXR (hPXR), or the hPXR allelic variants V140M, D163G, and A370T. MEHP increased the transcriptional activity of both mPXR and hPXR (5- and 15-fold, respectively) with EC₅₀ values of 7–8 μM. hPXR-V140M and hPXR-A370T exhibited patterns of phthalate responses similar to the wild-type receptor. By contrast, hPXR-D163G was unresponsive to all phthalate monoesters tested.

3.4. Cell proliferation

Human studies

No information about DEHP effects on cell proliferation in the human liver is available. Studies with cultured human hepatocytes failed to produce evidence of increased proliferation even though they were generally conducted in parallel with positive rat hepatocyte assays and at comparable doses [92, 124]. It should be noted that only a limited number of human donors were sampled which may have affected the significance levels of these studies due to presumably higher genetic variability in the human population. In addition, the condition of the livers used for cell isolations was unknown and may also be considered as confounding variables.

Experimental systems

The effects of DEHP were evaluated in young adult male cynomolgus monkeys with emphasis on detecting hepatic and other effects seen in rats and mice after treatment with high doses of phthalates [34]. Groups of 4 monkeys received DEHP (500 mg/kg/day) or vehicle (0.5% methyl cellulose, 10 ml/kg) by intragastric intubation for 14 consecutive days. DEHP had no effect on body weights or liver weights. Histopathological examination of tissues revealed no distinctive treatment-related effects in the liver, kidney, or testes. There were also no changes in any of the hepatic markers for peroxisome proliferation, including peroxisomal β-oxidation or replicative DNA synthesis. None of the test substances produced any toxicologically important changes in urinalysis, hematology, or clinical chemistry.

Rats and mice are species that are most susceptible to the induction of liver cell proliferation by DEHP. Hepatomegaly is the most pronounced early gross pathological change that occurs after continuous oral exposure to DEHP in rodents [70, 71]. This effect is the result of both hypertrophy (increase in the size of each cell) and hyperplasia (increase in the number of cells) of liver parenchymal cells. A burst of proliferation is observed in DEHP-treated rats and mice. Elevation in the rate of mitosis of hepatocytes was observed as early as 24 hrs after the initial dose [125], gradually increased until about 1 week of exposure, and was not sustained over longer time courses in either rats, or mice [125–127].

Cell-cycle genes appear not to be under direct transcriptional control of PPARα. Instead, several indirect mechanisms have been proposed, including involvement of miRNAs [128], activation of p38 mitogen-activated kinase [129], activation of NF-κB [130], or activation of Ras-like proto-oncogenes via post-translational modification [131]. None of these mechanisms has been investigated with regards to DEHP in susceptible species, and their relevance to human hazard remains to be elucidated.

DEHP and other peroxisome proliferators have not been shown to exhibit a marked effect on in vitro purified hepatocyte cultures, even in rodents. The increase in DNA synthesis, a marker of cell proliferation in rat and mouse hepatocytes in vitro has been reported to be only 120–200% [92, 124]. The magnitude of the induction of replicative DNA synthesis by DEHP or its metabolite MEHP is much lower than that seen in rat and mouse liver in vivo and it has been hypothesized that other cells in liver (e.g., Kupffer cells) may play an important role by potentiating the proliferative response of the hepatocytes by producing mitogenic cytokines. It is also noted that human liver may be more refractory than rodent liver to mitogenic stimuli.

Species comparisons

Male Fischer 344 rats and male Dunkin-Hartley guinea-pigs were given 950 mg/kg/day DEHP by gavage for four days [84]. Significant increases in liver weight, hepatic β-oxidation activity and hepatocyte DNA replication, and decreases in hepatocyte apoptosis were observed in the rats but not the guinea-pigs.

3.5. Apoptosis

Suppression of apoptosis in the liver has been suggested to act as an additional mechanism through which normal balance of cell turnover may be impaired to create conditions promoting neoplastic growth [132].

Human studies

No information about DEHP effects on apoptosis in human liver is available. A report provides evidence that cultured human hepatocytes are non-responsive to anti-apoptotic effects of MEHP [124]. In this study MEHP caused no induction of β-oxidation, stimulation of DNA synthesis or suppression of apoptosis in human hepatocytes cultured from three separate donors. The lack of an effect was consistent for both spontaneous and cytokine-induced (TGFβ1, or TNFα/amanitine) apoptosis. It should be noted that human hepatocyte cultures used in these experiments may have contained up to 5% non-parenchymal cells [92]. Viability of human monocytic leukaemia U937 cells was decreased by 20 hours of exposure to 42 μg/ml MEHP as well as concurrent increases in DNA fragmentation [133]. Caspase-3 activity was increased by treatment with 56μg/ml MEHP. This increased activity was diminished by 50% with pretreatment by selective PPARγ antagonists but increased 2-fold by pretreatment with a PPARγ ligand. A PPARα antagonist also increased MEHP-induced caspase-3 activity. The mRNA for Bcl-2, and Bcl-2/Bax ratios were decreased by 84μg/ml MEHP with Bax mRNA increased. In human lymphoblast cells, DEHP (up to 10 uM) and MEHP (up to 100 uM) had no significant effect on the activity of caspase-3/7 after 24 hrs of exposure [134].

Experimental systems

In both mice and rats treated with DEHP (mice at 1150 mg/kg; rats at 950 mg/kg) by gavage for two days, suppression of apoptosis and induction of hepatocyte DNA synthesis was observed [135]. In the same study, the authors showed that MEHP was able to suppress apoptosis (both spontaneous and transforming growth factor beta1-induced) and induce hepatocyte DNA synthesis in both mouse and rat hepatocytes. Kim et al. [136] reported that DEHP exposure increased cell proliferation in MCF-7 cells but not MDA-MB-231 human breast adenocarcinoma cells at the same concentration. DEHP mimicked estrogen in the inhibition of tamoxifen-induced apoptosis in MCF-7 cells (1/3 reduction of tamoxifen-induced cell death at 3.9 μg/ml). Apoptosis measured by the TUNEL assay decreased in parallel with increasing Bcl-2/Bax ratios by DEHP exposure (3.9 μg/ml). The survival of human cultured keratinocytes (NCDTC 2554) was decreased by DEHP-induced necrosis (not apoptosis) at 97μg/ml at 4 hours exposure and 20μg/ml at 24 or 48 hours of exposure [137]. Necrosis (induced by 195 μg/ml DEHP) was inhibited by the addition of antisense oligonucleotide against PPARβ. Western blot showed decreased pErk1, Erk2, c-myc, and increased PPARβ at 97 μg/ml, and increased PPARα at 20 μg/ml but decreased PPARα expression at 195 μg/ml DEHP.

3.6. Oxidative stress

It is widely accepted that in vivo exposure of rats or mice to DEHP leads to an increased oxidative stress in liver [138]. Induction of peroxisomal and microsomal enzymes, a pathway largely dependent on activation of PPARα, contributes to an increase in reactive oxygen species formed in the hepatocytes.

Experimental systems

Several in vivo studies have been carried out in rats in which DEHP was given and endpoints indicative of oxidative damage in the liver were measured. These include increases in lipofuscin [139–142] and malondialdehyde [143].

Takagi et al. [144, 145] investigated the relationship between hepatic peroxisome proliferation and levels of 8-hydroxydeoxyguanosine (8-OHdG) in hepatic DNA. Male Fischer 344 rats were fed 1.2% DEHP in the diet for periods of 1–12 months. Treatment with DEHP resulted in sustained stimulation of cyanide-insensitive palmitoyl-CoA activity and produced up to a two-fold increase in levels of 8-hydroxydeoxyguanosine in hepatic DNA.

A spin-trapping technique and electron spin resonance spectroscopy were used to provide direct evidence for DEHP-induced oxidative stress in liver [146]. Rats were given DEHP in the presence of the spin trapping agent α-(4-pyridyl-1-oxide)-N-tert-butylnitrone (POBN) and bile samples were collected for 4 hours of DEHP exposure. Under these conditions, the intensity of the six-line radical adduct signal increased to a maximum value of 2.5-fold 2 hours after administration of DEHP, before peroxisomal oxidases were induced. Furthermore, DEHP given with [(13)C(2)]dimethyl sulfoxide produced a 12-line electron spin resonance spectrum, providing evidence that DEHP stimulates (*)OH radical formation in vivo. Furthermore, when rats were pretreated with dietary glycine, which inactivates Kupffer cells, DEHP did not increase radical signals. Moreover, similar treatments were performed in knockout mice deficient in NADPH oxidase (p47(phox) subunit). Importantly, DEHP increased oxidant production in wild-type but not in NADPH oxidase-deficient mice. These data provide evidence for the hypothesis that the molecular source of free radicals induced by peroxisome proliferators is NADPH oxidase in Kupffer cells. Radical adduct formation was not affected in PPARα-null mice. These observations provide in vivo evidence that DEHP increases free radicals in liver before peroxisomal oxidases are induced.

In a follow up study, Woods et al. [147] hypothesized that continuous treatment with peroxisome proliferators will cause a sustained formation in POBN radical adducts in liver. Mice were fed diets containing either WY-14,643 (0.05% w/w) or DEHP (0.6% w/w) for up to 3 weeks. Liver-derived radical production was assessed in bile samples by measuring POBN radical adducts using electron spin resonance. WY-14,643 caused a sustained increase in POBN radical adducts in mouse liver and this effect was greater than that of DEHP. Free radical production after DEHP administration occured at 2 hours and 3 weeks but not at 3 days or 1 week of exposure. To understand the molecular source of these radical species, p47^phox-null and PPARα-null mice were examined after treatment with WY-14,643. No increases in radicals were observed in PPARα-null mice that were treated with WY-14,643 for 3 weeks, while the response in p47^phox-nulls was similar to that of wild-type mice. These results show that PPARα, not NADPH oxidase, is critical for a sustained increase in POBN radical production caused by peroxisome proliferators in rodent liver. Therefore, peroxisome proliferator-induced POBN radical production in Kupffer cells may be limited to an acute response to these compounds in mouse liver.

In Seo et al. [143], rats were treated with DEHP (50, 200, 1000 mg/kg) for 14 days and the activities of metabolizing enzymes and peroxisomal enzymes were investigated. Oxidative damage was measured using 8-OHdG in the DNA and malondialdehyde (MDA) in the livers. DEHP significantly increased relative liver weights, palmitoyl-CoA oxidation and activity of carnitine acyltransferase. A dramatic and dose-dependent increase in hepatic MDA levels was observed with DEHP (≥50 mg/kg). The 8-OHdG in hepatic DNA was increased only in the highest DEHP exposure (1000 mg/kg) group.

Chronic treatment with DEHP or other peroxisome proliferators elicits an adaptive response to oxidative DNA damage [148], including increased DNA repair, in both rat and mouse liver [149]. The degree of induction of DNA repair gene expression correlated with the dose and carcinogenic potency of the peroxisome proliferator compounds. Furthermore, it was shown that oxidative DNA damage is a PPARα-dependent event [150].

In a chronic (22 months) dietary feeding study with DEHP in wild-type (SV/129 strain) and PPARα-null mice, treatment-related induction of oxidative DNA damage and inflammatory marker expression was observed in both strains, albeit these changes were more pronounced in PPARα-null mice exposed to DEHP [38].

Primary cultures of hepatocytes isolated from male F344 rats were incubated in medium containing MEHP and the induction of peroxisomal acyl-CoA oxidase activity and lipid peroxidation was examined [151]. The latter was determined by measuring levels of conjugated dienes in lipid fractions extracted from harvested cells. While MEHP increased the relative specific activity of peroxisomal acyl-CoA oxidase, the level of conjugated dienes was not affected by up to 200 μM MEHP.

3.7. Gap-junctional communications

Experimental systems

The effects of DEHP or its metabolites on gap-junctional intercellular communication (GJIC) have been examined ex vivo and in vitro.

In cynomolgus monkeys, Pugh et al. [34] investigated the effects of DEHP (500 mg/kg/day), or vehicle (0.5% methyl cellulose, 10 ml/kg) administered by intragastric intubation for 14 consecutive days. In situ dye transfer studies, a marker for functional GJIC, using fresh liver slices revealed that DEHP had no effect on GJIC.

Isenberg et al. [152] reported inhibition of dye transfer in situ for liver strips from male B6C3F1 mice treated with DEHP (500 ppm at 2 weeks and 6000 ppm at 4 weeks) and male F344 rats treated with DEHP (6000 ppm at 1 and 6 weeks). Food consumption was not recorded and the rats were fed ad libitum so that a mg/kg/day dose could not be calculated. In a follow up study, Isenberg et al. [153] showed that dietary administration of 20000 mg/kg DEHP to male rats for 2 weeks decreased intercellular communication (67% of control) and enhanced replicative DNA synthesis (4.8-fold over control). Elevation of the relative liver weight and the induction of peroxisomal β oxidation were also observed following treatment with 20000 mg/kg DEHP for 2 weeks. Following DEHP administration at a dose of 6000 mg/kg for 18 months, inhibition of GJIC persisted, and the relative liver weight and induction of peroxisomal β-oxidation remained elevated in both rats and male B6C3F1 mice. In recovery studies in which DEHP was administered to male F344 rats for 2 weeks and then withdrawn, the relative liver weights, rate of peroxisomal β-oxidation, increase in replicative DNA synthesis, and inhibition of GJIC returned to control values within 2 to 4 weeks after DEHP treatment ceased. Recovery studies with phenobarbital produced similar results. MEHP was detected in the livers of animals treated with DEHP for greater than 2 weeks. However, it could not be detected after removal of DEHP from the diet for 2 weeks. This study demonstrated that inhibition of GJIC, along with indicators of peroxisomal proliferation, including increased relative liver weight and enhanced peroxisomal β oxidation, persist while DEHP treatment continues but reverses when treatment is stopped.

Inhibition of GJIC in Chinese hamster V79 cells was reported after in vitro exposure to DEHP. The lowest effective concentration detected in these cells for inhibition of GJIC was as low as 1 μg/ml [154]. Inhibition of GJIC was also studied in SHE cells exposed to DEHP. Mikalsen & Sanner [155] reported that normal SHE cells exposed to DEHP (14 μg/ml) for 24 hours were positive in the assay and that 5 days of exposure of DEHP (30 μg/ml) to transformed SHE cells gave a positive response. Cruciani et al. [156] reported transient inhibition of dye transfer in SHE cells after DEHP exposure at even lower concentrations (<10 μg/ml).

The only available study of MEHP for GJIC is that of Cruciani et al. [156] who reported inhibition of metabolic cooperation and inhibition of dye transfer in Chinese hamster V79 cells exposed to MEHP. The concentrations tested that inhibited metabolic cooperation (28 μg/ml) also decreased cell survival which was already at 60%. Cell survival was not noted in the study of inhibition of dye transfer of MEHP but 112 μg/ml MEHP was reported as the dose causing this effect in V79 cells. In the same study, inhibition of dye transfer in SHE cells was reported at 7μg/ml MEHP.

Species comparisons

GJIC was measured in vitro in liver cells from a variety of species 4 and 24 h after treatment with MEHP by lucifer yellow dye coupling [157]. GJIC was inhibited in rat and mouse hepatocytes by MEHP in a concentration-dependent manner (significantly reduced even at the lowest concentrations tested, 50 pM). Inhibition of GJIC in rodent cells was substantially reversed within 24 h of MEHP removal. In contrast, cell-to-cell communication was not inhibited in hamster, cynomolgus, or human hepatocytes, or in a human liver cell line (HLEC-4 established from hepatocytes isolated from the liver of a 35-yr-old male and immortalized with a SV40 T antigen) at any concentration examined.

4. Non-parenchymal cell-specific effects of DEHP in the liver

Experimental systems

Rose et al. [158] tested the hypothesis that Kupffer cells are directly activated by peroxisome proliferators, including DEHP and MEHP. Kupffer cell superoxide production was measured following treatment in vitro. WY-14,643 increased superoxide production in a dose-dependent manner (0.1 and 50 μM) with half-maximal stimulation at 2.5 μM. DEHP and 2-ethylhexanol did not increase superoxide production even at doses 50 times higher than WY-14,643; however, MEHP activated superoxide production as effectively as WY-14,643 with half-maximal stimulation at 5 μM. Treatment of rats with WY-14,643 but not DEHP for 21 days caused a 2-fold increase in Kupffer cell superoxide production. Pretreatment of Kupffer cells in vitro with staurosporine (0.01–10 pM) completely blocked generation of superoxide, demonstrating that protein kinase C is required. Moreover, WY-14,643 increased Kupffer cell protein kinase C activity 3-fold. Pretreatment of Kupffer cells with the amino acid glycine (0.01–3 mM), which blunts calcium signalling, inhibited WY-14,643-stimulated superoxide production and increased protein kinase C activity completely. The authors state that these data are consistent with the hypothesis that potent peroxisome proliferators (WY-14,643 and MEHP) directly activate Kupffer cell production of oxidants via mechanisms involving protein kinase C [159].

An alternative PPARα-independent mechanism for increased oxidative stress has been shown to involve activation of Kupffer cells [160]. Evidence for Kupffer cell-mediated increased oxidant production in vivo after treatment with DEHP was obtained using a spin-trapping technique and electron spin resonance spectroscopy [146]. Specifically, when rats were given DEHP acutely for 2 hrs, a radical adduct signal was detected. No increase in the radical signal due to DEHP was observed when Kupffer cells were inactivated in vivo with glycine pre-treatment, or in p47^phox-null mice. The rapid DEHP-induced free radical production in vivo occurred long before H₂O₂-generating enzymes in peroxisomes were induced, and it was not dependent on PPARα status.

A gene expression profiling study examined transcriptional changes induced by DEHP in mouse liver [75]. In addition to many genes that have been traditionally associated with hepatocyte-specific responses to PPs, a number of known components of the TNF/IL-1 signaling pathways, including Irak2, Myd88, Ikbkg and others were induced very early (2 hrs) and declined at later times (24 hrs) after acute treatment with DEHP, consistent with other studies showing a time-course of Kupffer cell activation.

5. Susceptible populations

5.1. Genetic polymorphisms and enzyme induction

The human PPARα (hPPARα) is highly homologous to the rodent PPARα in overall sequence and structure [48, 61, 62], but a number of allelic variants of hPPARα have been isolated which possess properties different from the original cloned hPPARα. The L162V variant containing an amino acid change in the DNA-binding domain is found at an allelic frequency of ~0.025–0.073 in ethnically diverse populations [161–163]. In North Indians, this allele is found with high frequency (0.745) [164]. The hPPARα L162V variant exhibits no response to low doses of WY but greater ligand-induced activity (up to ~4-fold) at higher doses compared to the wild-type receptor [161, 164]. Humans carrying this variant exhibit greater decreases in total serum cholesterol to the hypolipidemic agent, bezafibrate [161]. Three different Asian populations carry a hPPARα variant (V227A) within the hinge region between the DNA binding and ligand binding domains at frequencies of 0.003–0.051 [165, 166]. This allele has been associated with decreases in serum cholesterol and triglycerides in a Japanese population [165] and in Chinese women [166]. Because of increased interactions with a co-repressor, nuclear receptor corepressor (NCoR), this variant exhibits decreased responsiveness to PPARα activators [167].

The hPPARα-6/29 variant containing four amino acid substitutions is a dominant negative that binds to a PPRE but cannot be activated by PPARα activators [168]. The hPPARα-6/29 variant is likely very rare, as it was not detected in any of the 173 human subjects from two studies [164]. Overall, some PPARα allelic heterogeneity exists in human populations, but no variants have been identified that are more sensitive to low, environmentally-relevant doses of PPARα activators than the “wild-type” human receptor. The field would benefit from a side-by-side comparison of wild type and hPPARα variants in trans-activation assays to determine dose-response relationships of PPARα activators.

5.2. Identification of groups or subpopulations with an enhanced susceptibility with a focus on fetal and neonatal responses

A substantive database exists on the susceptibility of male reproductive system to exposures to DEHP and related chemicals during early life [169, 170]. The overall weight of evidence for these non-cancer effects have led to public health actions in Europe and United States concerning early life exposures to DEHP-containing medical devices, toys and related consumer products [171].

Most studies conducted in rodents that examined liver effects of DEHP during gestational, lactational or post-natal exposures conclude that there is little evidence of age-specific differences in toxicity. Cimini et al. [78] treated F344 rat dams with 1 g/kg/day of DEHP by gavage for up to 21 days from day of delivery through lactation. Pups were sacrificed on day 14, day 21, or day 35 following 14 days of recovery. Relative liver weights increased 1.65-fold in the dams at weaning and 1.47-fold in 14- and 21-day pups. At day 21, palmitoyl-CoA oxidase increased 9.3-fold in dams, while it increased 6-fold in the nursing pups at 14 days and 4.85-fold at 21 days. However, palmitoyl-CoA oxidase activity was substantially less in the pups than in the dams treated with DEHP (pups, 1.2 mU/mg protein at 14 days; dams, 34.4 mU/mg protein at 21 days). Dihydroxyacetone-phosphate acyltransferase (DHAP-AT) was increased about two-fold in 14- and 21-day neonates, but DHAP-AT levels were unaffected in DEHP-treated dams. Catalase activity was increased about two-fold in 14-day and 21-day neonates and adults. Following 14 days of recovery, most enzyme levels returned to normal in the dams and pups, although catalase activity remained slightly higher in both the dams and pups.

In a separate study, pregnant lactating F344 rat dams were administered by gavage 1 g/kg/day DEHP for 21 days beginning at the day of delivery, and the nursed pups were sacrificed after two or three weeks or following a 14-day recovery period [172]. The numerical density or volume density of peroxisomes was increased marginally (less than twofold), relative to controls, in both pup groups. Dams treated for 21 days with DEHP showed a more pronounced increase in the volume density of peroxisomes (about two-fold), but the numerical density of peroxisomes was increased in the dams to the same degree as the two- or three-week pups. The increases in volume density or numerical density of peroxisomes did not decline to control levels in the three-week pups after a 14-day recovery period. Volume density of peroxisomes apparently declined to close to control levels after a recovery period of eight days in dams treated for three weeks, but there was no apparent decline in the numerical density of peroxisomes. Relative liver weights were increased about equally in two- and three-week-old neonates and adults (1.5 to 1.6-fold).

Two studies investigated the effects of DEHP on neonatal rats of different ages [76, 77]. Male Sprague-Dawley rats 6, 14, 16, 21, 42, or 86 days of age were administered (by gavage) daily doses of DEHP for five days, and 24 hours after the last dose the activities of hepatic peroxisomal enzymes, palmitoyl-CoA oxidase, and carnitine acetyltranferase were determined. The doses administered were 0, 10, 100, 1000, or 2000 mg/kg/day. Administration of 1000 mg/kg/day caused significant decreases in body weight and mortality (66% to 70%) in pups 14 to 18 days of age, and administration of 2000 mg/kg/day caused mortality in virtually all pups of these ages. At a non-lethal dose (100 mg/kg/day), absolute liver weight increases relative to those in the controls were 0%, 17%, 3%, 10%, and 14% for 6–10-, 14–18-, 21–25-, 42–46-, and 86–89-day old pups and adults, respectively. At 100 mg/kg/day, measurements of palmitoyl-CoA activity showed that there was a greater increase only in the 14–18-day pups when compared with 86–90-day adults (6.9-fold increase versus a 3.98-fold increase). A greater increase in carnitine acetyl transferase was also shown at this dose level only for 14–18-day pups when compared with 86–90-day adults (7.8-fold increase versus a 4.4-fold increase). The data on increased liver weights and peroxisomal enzyme activities from this study indicate that neonatal or young adult rats differ little from adult rats in their response to treatment with DEHP, although the palmitoyl-CoA and carnitine acetyl transferase activities were higher in the 14–18-day pups compared with adults.

A recent study in CD-1 mice investigated effects of in utero (e.g., gestational days 11–19) DEHP exposure (100 mg/kg/day dissolved in corn oil by gavage) on liver development [173]. The data show that in utero DEHP exposure altered post-natal liver development in weanling (21 day old) mice causing significant dose-related hepatosteatosis, impaired glycogen storage, and increased beta-catenin and alpha fetal protein expression. At puberty (35 day old), a significant decrease in glycogen storage in hepatocytes was still observed in male mice. The authors concluded that DEHP alters post-natal liver development, as evidenced by the increased expression of onco-fetal genes, and delayed programming of glycogen metabolism.

In addition, several human studies have suggested that the risk of hepatoblastoma, a rare childhood cancer of the liver, is high in pre-term newborns [10–12]. While it is difficult to draw any definitive conclusions from these studies regarding the potential role of DEHP and/or PVC-containing medical devices that may have been used in neonatal care and risk of hepatoblastoma, it is well known that exposure to DEHP is high in pre-term neonates and infants [174].

6. Toxic effects of DEHP in the lung

Human studies

A significant increase in morbidity from both total and respiratory tract cancers was observed in a cohort of 2,031 male workers at a PVC processing plant [6]. In addition, cases of asthma and/or respiratory symptoms that may have been associated with exposure from PVC-containing plastic fumes have been reported in the literature [175]; however, in most of these studies no linkage can be made exclusively to DEHP. A number of epidemiologic studies report a high prevalence of work-related upper and lower respiratory tract symptoms among meat wrappers where the risk is substantially higher among the exposed, but the effects on lung function were reported to be inconsistent. In children, the meta-analysis [175] of 5 epidemiological studies showed an association between PVC surface materials in the home and the risk of asthma [fixed-effects model: summary odds ratio (OR), 1.55; 95% confidence interval (CI), 1.18–2.05; four studies] and allergies [OR: 1.32; 95% CI: 1.09–1.60; three studies]. A cross-sectional study found no association between phthalate exposure (urinary phthalate metabolite levels) and pulmonary function parameters among the 240 adult participants in the National Health and Nutrition Examination Survey (NHANES) participants [176].

Animal studies

There are only few rodent studies that reported adverse effects of DEHP on the lung. A 28-day inhalation toxicity study of DEHP (head-nose inhalation of aerosols) in Wistar male and female rats [177] reported a statistically significant increase in relative lung weights, accompanied by increased foam-cell proliferation and thickening of the alveolar septi, at the highest exposure groups (230 mg/kg/day in males and 360 mg/kg/day in females). In a 104-week dietary exposure study to DEHP in B6C3F1 mice [14], an increase in mean relative lung weight in male mice at the highest dose group (6000 ppm) was reported. A similar effect was observed in male F344 rats in a 104-week dietary feeding study at the doses of 2500 and 12500 ppm [15]. A study using rat tracheal strips found that mono(2-ethylhexyl) phthalate (MEHP, 0.1 mmol/L) induced hypersensitivity to methacholine-induced contraction of rat tracheal muscle, an effect that the authors suggested may result in clinical bronchial hyperreactivity [178].

7. Toxic effects of DEHP in the reproductive system

Human studies

In a case-control study, occupational exposures to PVC were found to be associated with increased risk of testicular cancer [OR=6.6; 95% confidence interval= 1.4–32] [5]. In humans, the database of the reproductive and developmental effects of phthalates has been growing rapidly in the past decade and numerous studies reported data suggestive of an association between exposure to DEHP and/or other phthalates and adverse effects on both male [179–183] and female [184–187] reproduction. A recent case-control study of female breast cancer study explored possible linkages to environmental exposure to phthalates [9]. Phthalate metabolites were measured in urine and an increased risk of breast cancer was identified for one of the four DEHP metabolites evaluated, mono(2-ethyl-5-carboxypentyl) phthalate. Significant negative associations were observed for monobenzyl phthalate and mono(3-carboxypropyl) phthalate. No association was found for the other DEHP metabolites: mono(2-ethylhexyl) phthalate, mono(2-ethyl-5-hydroxyhexyl) phthalate and mono(2-ethyl-5-oxohexyl) phthalate.

Animal studies

In a lifetime dietary DEHP feeding study in rats, the incidence of Leydig cell tumors was found to be increased, and was dose-related with early onset [16]. Animal evidence that DEHP is teratogenic and a reproductive toxicant has been reviewed [68, 169]. Testicular toxicity of DEHP is independent of the status of the nuclear receptor PPARα, because identical, albeit slightly delayed, effects were observed in PPARα-null mice [36]. Effects on testicular development in rats following exposure to DEHP prenatally and during suckling or during adolescence at dose levels below those associated with peroxisome proliferation have been reported [188]. The molecular events associated with reproductive and developmental effects of DEHP and other phthalates are not well characterized, but many studies suggest that phthalate’s effects on metabolism and other cellular functions leads to disruption of steroidogenesis [189], increased oxidative stress [190], increased apoptosis [191], and other events [192]. Data also suggest that both Sertoli [193] and Leydig [194] cells are targets for toxicity.

8. Toxic effects of DEHP in the kidney

Human studies

No data on potential toxicity or carcinogenicity of DEHP in human kidneys was identified.

Animal studies

In young adult male cynomolgus monkeys treated for 14 days with DEHP (500 mg/kg/day), or vehicle (0.5% methyl cellulose, 10 ml/kg) by intragastric intubation, no distinctive treatment-related effects were observed in the kidney [34]. Several studies in mice have reported dose- and time-dependent kidney lesions following exposure to DEHP [127, 195, 196]. David et al [14] also found that DEHP is nephrotoxic and the authors suggested that because chronic progressive nephropathy is a common finding in ageing mice, DEHP may exacerbate this pathological change. The association of nephropathy with the status of PPARα was evaluated using knockout and wild-type mice [36]. Nephropathy was observed in both wild type and PPARα-null mice, albeit it was more severe in wild type animals. DEHP also causes nephropathy in rats [15]. It was suggested that DEHP-induced nephropathy in rodents may be related to peroxisome proliferation that has been observed in the renal proximal tubules [78, 197, 198]. While kidney tumors were not observed in chronic studies with DEHP, a related chemical, di(isononyl) phthalate was shown to cause kidney tumors in male, but not female rats [199].

9. Toxic effects of DEHP in the immune system

Human studies

Excess incidence of multiple myeloma among subjects exposed to phthalates and PVC at the workplace was observed in a large population-based case-control study in Denmark [8]. The possible effects of DEHP on immune function were studied in vitro. The human lung carcinoma cell line A549 was exposed to 15.6–2000 ug/ml of MEHP and concentrations of the proinflammatory cytokines IL-6 and IL-8 were measured in the cell culture supernatant [200]. At low (100–200 ug/ml) concentrations, a concentration-dependent increase in cytokine production was observed; however, at higher (~1000 ug/ml) concentrations, the cytokine production was suppressed. When the human monocytic cell line (THP-1) or peripheral blood mono-nucleated cells obtained from allergic patients and non-allergic controls were cultured in presence of MEHP (0.2, 2.0, 20, and 200 μg/mL), no effect on cytokine production was observed [201]. A study in human peripheral blood mononuclear cells (containing 0.1–1% basophils) showed that MEHP and DEHP (5–500 uM) had an effect on IgE and IgG release [202].

Animal studies

Several studies have been performed to assess whether various phthalates, including DEHP and MEHP, may act as sensitizers. Subcutaneous injection, not a likely route of exposure in humans yet an important mode of administration in studies of immune function, was used in several studies. MEHP was shown to have an immuno-suppressive effect, defined as a statistically significant reduction in IgE or IgG1 antibody production (indicative of the stimulation of the T_H2 pathway, which is predominant in type I allergies) at the dose of 1000 ug/ml, and an adjuvant effect, defined as a statistically significant increase in IgE or IgG1 antibody level at the dose of 10 ug/ml [203]. In a study which tested whether PPARα may be involved in the adjuvant effect of DEHP, it was shown that DEHP induced highly increased levels of ovalbumin-specific IgG1 and IgG2a, both in wild-type and in PPARα-null mice, indicating that it is a mixed Th1/Th2 adjuvant and that the adjuvant mechanism is mediated through a PPARα-independent mechanism [204]. The findings that DEHP can act as an adjuvant have been challenged in a study [205] where topical administration of DEHP was without impact on antibody responses, regardless of whether it was applied locally or distant to the site of ovalbumin immunization. Inhalational exposure to MEHP (60 minutes; 0.3–43.6 mg/m³) in mice led to a concentration-dependent decrease in tidal volume and an increased number of alveolar macrophages, but no change in numbers of neutrophils, lymphocytes, eosinophils, or epithelial cells in broncho-alveolar lavage fluid [206]. An in vitro study with rabbit alveolar macrophages showed that phagocytosis, a primary function of alveolar macrophages, was significantly enhanced by DEHP (up to 2% in media) in a dose-dependent manner [207]. The effects of DEHP (50–500 uM) on antigen-induced degranulation of rat basophilic leukaemia cells (RBL-2H3) was demonstrated upon antigen stimulation [208].

10. Conclusions

Effects of DEHP on the liver, a primary target organ for the pleiotropic effects of DEHP and other peroxisome proliferators in rodents, have been the focus of scientific debate for the last three decades [42, 138, 209]. In the liver of rodents, parenchymal cells, also called hepatocytes, are a major cell type that are responsive to DEHP; however, other cells, such as resident hepatic macrophages, called Kupffer cells, may also play an important role. The increase in the number and size of peroxisomes in hepatocytes, so called ‘peroxisome proliferation’ that results in elevation of fatty acid metabolism, is a hallmark response to DEHP and other peroxisome proliferator compounds in the liver of susceptible species [210]. A link between peroxisome proliferation and liver tumor response has been a predominant theory to explain the cause of a hepatocarcinogenic effect, even though the experimental support is not unequivocal [211]. Other molecular events, such as induction of cell proliferation, decreased apoptosis, oxidative DNA damage, and selective clonal expansion of the initiated cells, have also been proposed to be critically involved [42].

Overall, it is believed that the events that occur relative to DEHP-induced liver carcinogenesis in rodents involve the combination of the molecular signals and multiple pathways, rather than a single hallmark event (such as activation of PPARα, peroxisome proliferation, or cell proliferation), that contribute to the formation of tumors: (i) rapid metabolism of the parental compound to primary and secondary bioactive metabolites that are readily absorbed and distributed throughout the body; (ii) receptor-independent activation of hepatic macrophages and production of oxidants; (iii) activation of PPARα in hepatocytes and sustained increases in expression of peroxisomal and non-peroxisomal metabolism-related genes; (iv) enlargement of many hepatocellular organelles (peroxisomes, mitochondria, etc.); (v) rapid, but transient increases in cell proliferation and decreases in apoptosis; (vi) sustained hepatomegaly; (vii) chronic low-level oxidative stress and accumulation of DNA damage; (viii) selective clonal expansion of initiated cells; (ix) appearance of pre-neoplastic nodules; (x) development of adenomas and carcinomas.

Despite the wide use of phthalates, including DEHP, which leads to appreciable exposures in the general population, only a limited amount of human mechanistic data is available. The majority of experimental human data comes from in vitro studies in cultured human liver cells; however, the data from these studies suggest that human cells do not give responses that parallel those observed in cultured rodent liver cells in response to DEHP or its metabolites. At the same time, even rodent liver cells do not recapitulate many events observed in the rodent liver in vivo.

Major differences in the molecular signalling events elicited by DEHP in the liver, such as the activation of the nuclear receptor peroxisome proliferator-activated receptor (PPAR) α, have been observed between species. In addition, it is important to note that phthalate metabolite liver concentrations in larger mammals (i.e., marmosets) are several fold lower than that of rodents in similar dose groups which suggests that differences in metabolism between small and large mammals exist and thus may further compound species comparisons [45, 138]. In 2000, an IARC review panel [212] considered DEHP and concluded that in rodents, peroxisome proliferators exercise their pleiotropic effects in liver due to activation of PPARα and that this process is essential for liver hypertrophy and hyperplasia and eventual hepatocarcinogenesis. This conclusion was based on the data from a variety of studies that considered, among many, the molecular biology of PPARα signalling in different species, transactivation potency of PPARα from different species, in vivo studies in non-human primates, and studies in genetically modified mice. It was concluded that DEHP cancer classification should be downgraded from possibly carcinogenic to humans (Group 2B) to not classifiable as to its carcinogenicity to humans (Group 3).

One of the key pieces of evidence which led to a change in DEHP cancer classification in 2000 [212] was the chronic feeding study with the potent peroxisome proliferator WY-14,643 in PPARα wild-type and –null mice that showed that the nulls were completely refractory to liver carcinogenesis [213]. Although this study made one of the most significant contributions to mechanistic research on peroxisome proliferators, the large dose of the agent used, shorter than 2-yr exposure, and relatively small numbers of animals evaluated, somewhat limit the broad interpretation of the findings. More recent studies that used genetically engineered mice provide important additional data key for consideration of the relevance of the PPARα mode of action to rodent and human liver carcinogenesis. These include, but are not limited to, studies in PPARα-null mice [38, 105, 117, 118], PPARα humanized transgenic mice [114], and hepatocyte-specific constitutively activated PPARα transgenic mice [111]. The data from these animal models suggest that even though activation of PPARα and the subsequent downstream events mediated by this transcription factor represent a key mechanism of action [42], it is evident that several additional molecular signals and multiple pathways in several cell types in the liver, rather than a single molecular event, contribute to the formation of liver tumors in rats and mice.

In addition, many human and rodent studies suggest that the liver is not the only target tissue for DEHP-associated carcinogenesis. Pancreas, respiratory tract, breast, testis and other tissues have been linked through either human studies of occupational, medicinal, or environmental exposure to materials that contain DEHP, or studies in rodents where exposure is known. Equally important are studies in animals and humans chronically exposed to chemicals that are thought to act via pathways similar to DEHP, such as lipid lowering drugs (e.g., fibric acid analogues), or chlorinated solvents (e.g., trichloroethylene). While there is little evidence for increased risk for liver cancer in subjects treated with fibrates, the epidemiology evidence for a causal association between exposure to trichloroethylene and liver cancer is positive but rather limited. Thus, the overall body of evidence on human cancer hazard of DEHP remains inconclusive.

Furthermore, it should be noted that even though important species differences in activation of PPARα or its signalling network by peroxisome proliferators exist, human cells express PPARα and are not devoid of trans-activation response to many peroxisome proliferators, including MEHP [55]. Important inter-individual differences in PPARα expression have been reported, suggesting that the differences in expression between species may need to be verified using a larger sample of both humans and animal strains. Thus, although quantitative differences between species may very well exist, qualitative similarities cannot be ignored, especially because DEHP and other PPARα activators are known to induce molecular responses independent of PPARα activation. It remains possible that these pathways contribute to human risk in ways somewhat different from those postulated for rat and mouse liver cancer. Overall, the mechanisms for the induction of cancer by DEHP are not entirely established and are certainly complex.

In summary, additional data from animal models and studies in humans exposed to DEHP from the environment suggest that multiple molecular signals and pathways in several cell types in the liver, rather than a single molecular event, contribute to the cancer in rats and mice. Importantly, the toxic and carcinogenic effects of DEHP are not limited to liver. The International Agency for Research on Cancer monograph 101 working group concluded [214] that the human relevance of the molecular events leading to cancer elicited by DEHP in several target tissues (e.g., liver and testis) in rats and mice can not be ruled out and DEHP was classified as possibly carcinogenic to humans (Group 2B).