Proposed Mode of Action for In Utero Effects of Some Phthalate Esters on the Developing Male Reproductive Tract

Abstract

Phthalate esters containing a straight-chain backbone of 4–6 carbons have demonstrated testicular toxicity and infertility in adult and pre-adolescent rats, mice, hamsters, and ferrets. In recent years, these same phthalates have been shown to interfere with the normal development of the male reproductive tract in rodents and rabbits. The review presented here summarizes studies that provide evidence of a mode of action for these effects. The data indicate that C4–C6 phthalate esters inhibit processes in the Leydig cell, such as the synthesis of testosterone (T) and production of insulin-like factor 3 (insl3), both of which are required for normal development of male genitalia. A proposed secondary effect of reduced androgen production is on Sertoli cells, resulting in failure to proliferate and interference with cell-cell communication (gap-junction intracellular communication) leading to the development of large multinucleate gonocytes. The possibility that phthalates act directly on the Sertoli cells to interfere with intracellular communication is not excluded. The strength, consistency, and plausibility of the proposed mode of action and alternate modes of action are discussed.

Introduction

Almost 10 years ago, Wine et al. (1997) reported malformations of the male reproductive organs following gestational exposure to di-n-butyl phthalate (DBP). Previous to that time it was known that some phthalate esters (PEs) affect adult testes and reduce sperm production in rodents (Peters and Cook, 1973; Cater et al., 1977). For the most part, effects were observed only at relatively high dose levels (>250 mg/kg) and involved only phthalate esters of straight-chain C4–C6 alcohols, di-(2-ethylhexyl) phthalate (DEHP—a branched C6 alcohol), and butyl benzyl phthalate (BBP) [BBP functions like DBP through the common metabolite, mono-butyl phthalate (MBP)] (Gray and Butterworth, 1980; Gangolli, 1982). The effects observed included loss of germ cells with subsequent decreases in sperm production, increased apoptosis of Sertoli cells, and, ultimately, reductions in gross testicular weight. Studies in vitro indicated that the diesters were inactive on the testes but that the monoesters were testicular toxicants (Gray and Beamand, 1984; Sjöberg et al., 1986); other metabolites were not as potent testicular toxicants (Foster et al., 1981; Gray and Beamand, 1984; Sjöberg et al., 1986).

The purposes of this review are to identify key events that lead to the malformations and effects observed in rodents and to perform a mode of action (MOA) analysis. Data that provide evidence of the key events or the overall MOA in rats and mice are included. For additional information, a critical review of all the developmental and reproductive toxicity of DEHP was recently compiled by the National Toxicology Program Center for Risks to Human Reproduction 〈http://cerhr.niehs.nih.gov/〉. To better understand the MOA for in utero effects, it may be helpful to provide a short review of some of the effects in adults because some events may be in common. Otherwise, the focus of this review will be studies using in utero or neonatal exposures.

Spermatogenesis is a hormone-dependent phenomenon. Consequently, early investigations evaluated the effect of C4–C6 phthalates on circulating hormone levels as a key to understanding the mechanism of toxicity. The results showed that circulating hormone levels do not consistently change up or down after exposure. Some of the inconsistencies may be due to experimental design [i.e., methodology and timing for testosterone (T) measurements] or source of biological sample (serum versus testes). For example, following treatment of animals with either DBP, BBP, or DEHP (>200 mg/kg), Oishi and Hiraga (1979) and Agarwal et al. (1985, 1986) reported decreased serum levels of testosterone (T), and increased testicular T levels (Oishi and Hiraga, 1980a,1980b,1980c). On the other hand, serum T in rats has also been shown to increase at lower dose levels (<6250 ppm in the diet; equivalent to ~400 mg/kg bw/day) of BBP but decrease at higher levels (Agarwal et al., 1985). Decreases in testicular T have also been reported following treatment with monoesters (Oishi and Hiraga, 1980a,1980b). Similar inconsistent results have been observed for other hormones involved in spermatogenesis. Serum luteinizing hormone (LH) was higher following treatment with high doses of DEHP, whereas follicle stimulating hormone (FSH) was apparently unaffected by treatment with phthalates (Agarwal et al., 1985, 1986). Decreases in steroidogenesis and steroid metabolism were observed (Bell et al., 1978; Oishi, 1982; Foster et al., 1983; Parmar et al., 1988; Fukuoka et al., 1990) and maybe a common feature of substances known as peroxisome proliferators (Liu et al., 1996). Recent studies by Akingbemi (2001, 2004) have evaluated the effect of DEHP exposure on peripubertal rats and the changes in T and LH production.

Interference with cellular processes has been shown, in rodents (Lloyd and Foster, 1988; Heindel and Chapin, 1989; Treinen et al., 1990) exposed to phthalates, which indicate that Sertoli cells did not respond to the FSH trigger for spermatogenesis (Lloyd and Foster, 1988; Heindel and Chapin, 1989; Treinen et al., 1990). Furthermore, communications between Sertoli cells and spermatogonia are disturbed, leading to sloughing of spermatids (Creasy et al., 1983; Richburg et al., 2002), and an increase in apoptosis has been associated with testicular toxicity following exposure to DEHP (Richburg, 2000).

Data generated in the past 10 years indicate that late-gestational exposure of pregnant rats to these same C4–C6 phthalate esters can adversely affect the developing male reproductive tract (fetal-onset testicular toxicity or developmental toxicity). One of the first reports of in utero effects indicated that DBP administered to pregnant rats resulted in some offspring that were infertile and/orr had malformations of the reproductive organs (Wine et al., 1997). Reports of the other C4–C6 phthalates causing such effects followed (Gray et al., 2000; Parks et al., 2000), with a report of DINP also showing increased areola retention and small testes (Gray et al., 2000) and reduced testosterone synthesis (Borch et al., 2004).1 The dose levels at which the effects on male reproductive development have been observed were, at first, as high as those for which adult-exposure toxicity was seen (≥250 mg/kg). However, more subtle effects, such are areola retention and reduced anogenital distance, were noted at lower dose levels of DBP (≤50 mg/kg) suggesting interference with hormonally triggered development (Mylchreest et al., 2000). These effects and dose-response observed following in utero exposure to C4–C6 phthalate esters are summarized in Table 1.

The gross effects observed in male offspring include dysgenesis of the gubernacular ligament leading to cryptorchidism and reduced T production leading to hypospadias (Mylchreest et al., 1998, 1999; Parks et al., 2000; Gray et al., 2000; Wilson et al., 2004); both of these outcomes are associated with Leydig cell function. In histological evaluations of testes, Leydig cells have also been observed in clusters or clumps (Barlow and Foster, 2003; Fisher et al., 2003), and there are reports of clumps of multinucleated gonocytes and clumps of immature Sertoli cells (Barlow and Foster, 2003; Fisher et al., 2003). Although the individual effects observed are not specific to phthalate esters, the aggregate of effects may be specific and the effects in aggregate have been termed “testicular dysgenesis syndrome” (Skakkabaek et al., 2001). To date, these effects have been observed in studies of C4–C6 phthalates in rodents (mice and rats) and in one study of DBP with rabbits (Higuchi et al., 2003).2

The underlying mechanism for these effects has been the focus of much research and discussion. Initial mechanistic arguments centered on phthalate esters acting as environmental estrogens (Colborn et al., 1993) and indeed, several phthalates were found to bind to the estrogen receptor (Jobling et al., 1995; Harris et al., 1997). However, subsequent studies showed that although some of the phthalate diesters bind to estrogen receptors, the phthalate monoesters do not. Given phthalates are absorbed as monoesters or rapidly metabolized to monoesters, this lack of binding may account for the lack of estrogenic effects under in vivo conditions (Coldham et al., 1997; Milligan et al., 1998; Zacharewski et al., 1998). Research was then directed at the ability of phthalates to interact with the androgen receptor because some in utero effects were similar to those known to be caused by anti-androgens, such as flutamide.

Studies of direct binding to the androgen receptor have shown that some phthalate diesters (unmetabolized parent compound) bind to the androgen receptor (Sohoni and Sumpter, 1998; Fang et al., 2003; Takeuchi et al., 2005), but that the monoester metabolites are not agonists or antagonists based on in vitro receptor studies (Mylchreest et al., 1999; Parks et al., 2000; Foster et al., 2001; McKee et al., 2004; Takeuchi et al., 2005). Phthalate esters known to affect reproductive development in rodents in studies involving in utero exposures (DBP, BBP, and DEHP) have been tested in in vivo systems for androgenic or anti-androgenic effects. Ashby and LeFevre (2000a) described inconsistent responses in the Hershberger assay using DBP and no response with BBP. Castrated male rats treated with 0.4 mg/kg/d T and 500 mg/kg DBP had significantly lower weights of the bulbo cavernosus/levator ani muscle complex, Cowper’s gland, seminal vescicles, and prostate glands compared with the group given T alone. If the dose of T was increased to 1 mg/kg/d, a dose of DBP increased to 1000 mg/kg/d was required before any significant changes in the weights of accessory organs were seen and only in the seminal vesicles and prostate gland. The 500 mg/kg DBP group and 1.0 mg/kg T had no differences in organ weight compared with the T treatment alonesignificantly lower.

Ashby and LeFevre (2000b) and O’Connor et al. (2002) were able to demonstrate more consistent effects with DBP using an intact peripubertal male, but BBP was again without effect (Ashby and LeFevre, 2000a). This result is intriguing since BBP and DBP are thought to act via the same active metabolite, MBP). In addition, data from in vivo studies did provide evidence of reduced testosterone synthesis following treatment with DBP (O’Connor et al., 2002). Recently, Stroheker et al. (2005) reported antiandrogenic effects with DEHP in a modified Hershberger assay (lower weights of the bulbo cavernosus/levator ani muscle complex at 100 mg/kg, prostate weight at 200 mg/kg, and all male accessory organs at >400 mg/kg). Nonetheless, the debate of anti-androgenicity as a mechanism continues with some suggesting an androgen receptor-related effect for the diester (Fang et al., 2003) and others suggesting that C4–C6 phthalate esters are not androgen antagonists (Mylchreest et al., 1999; Gray et al., 2000; Parks et al., 2000; Foster et al., 2001).

In summary, the reproductive toxicity of some phthalate esters may be divided into two parts: one affecting adult males, in which the Sertoli cell appears to be the target, and one affecting the developing male, in which the Leydig and Sertoli cells may be the targets. The fact that in utero exposure appears to have a different target than postnatal exposure suggests that there could be 2 different modes of action, but the latter hypothesis is not universally accepted. On the other hand, the possibility exists that cell-cell interaction between the Leydig and Sertoli cell in the developing male contributes to the adverse effects on Sertoli cells which then influence the developing germ cell. Thus, like the MOA of phthalate esters in the adult, the Sertoli cell may also be a target for phthalate ester effects on in utero. The focus of this assessment of phthalate ester MOA will be on in utero exposure, with additional evidence from neonatal exposure if it provides information on key events.

To understand the MOA of C4–C6 phthalate esters on the developing male reproductive tract in rodents, a review of some of the steps in sexual differentiation as known is warranted. Development of any organ system is complex, and not all the steps are understood. However, development of the male reproductive tract is clearly hormonally driven. Early in development, precursor cells and tissues are equipotent in their ability to develop into male or female structures; it is only after secretion of key hormones from the Sertoli and Leydig cells that fetal male tissues develop and the corresponding precursor female tissues regress. Anti-Müllerian Hormone (AMH; also known as Müllerian Inhibitory Substance, MIS) secreted from the Sertoli cells promotes the regression of the Müllerian ducts; T and its metabolite dihydroxytestosterone secreted from the Leydig cell promote the development of the Wolffian ducts into the epididymis, vas deferens, and seminal vescicles; and insulin-like factor 3 (insl3) secreted from the Leydig cells promotes the development of the gubernacular ligament to guide testicular descent (Sharpe, 2001). Failure of the proper hormone(s) to be produced in sufficient amounts at the proper time and/or blockage of the receptor(s) will result in feminization of the male.

The process of development in rodents and probably other mammals begins with gene expression when the gonads are mere genital ridges. Recent reviews (Haider, 2004; Park and Jameson, 2005) have discussed the sequence of gene expression in greater detail than can be presented here, but an outline is provided (Table 2). Expression of the same gene in different tissues may trigger a different sequence of events, for example, Sf1 (the gene that transcribes for the orphan nuclear receptor Steroidogenic Factor 1 or SF1) is expressed during early gestation (gestation day (GD) 10) in the precursor cells and again later (GD14) in Sertoli and Leydig cells. SF1 promotes production of key hormones, such as AMH in the Sertoli cell, while also promoting production of key enzymes for steroidogenesis in the Leydig cell a day or so later in development. Other genes are expressed and, based on knockout mouse studies, are key to development. However, the direct downstream events are not always clear. For example, Sry (Sex determining region of the Y chromosome) is required for cell proliferation and differentiation and, in conjunction with Sox9 and perhaps Sox8, triggers transcription of SF1 but does not seem to trigger Sox gene expression itself.

Development of the male reproductive tract in the rat appears to begin on GD 9 with expression of the genes Wt1, GATA4, and Sf1 in precursor cells. Following expression of these genes, Sry is expressed in the precursor Sertoli cells at GD10, thereby triggering cell proliferation (Schmahl et al., 2000) and segregation of SF1-positive cells from SF1-negative cells. This gene also acts synergistically with Sox9 and Sox8 at roughly GD14 to trigger expression of Sf1 in both Sertoli and Leydig cells. In the Sertoli cell, the nuclear receptor SF1 induces production of AMH (through expression of the gene Amh) at around GD 16. The androgen receptor is also present at this time, but its relationship to SF1 is not clear.

In the Leydig cell, Sf1 (and the nuclear receptor SF1) is expressed around GD14 (the receptor at GD15), probably in response to SOX9 as in the Sertoli cell. SF1 from the Leydig cell triggers transcription for the enzymes and transport proteins for T synthesis as well as the production of insl3 required for testicular descent. At this time of gestation, Inhibin can be detected, as can 3βHSD activity. Up-regulation of Fgf9 and FgfR2, at around GD13, and up-regulation of Dhh, lead to differentiation of Leydig and Sertoli cells. The timing and the level of gene expression are critical to the proper development of the reproductive tract (Haider, 2004; Park and Jameson, 2005; Small et al., 2005). Exposure to phthalate esters has been shown to affect the expression of several of these genes (Barlow et al., 2003; Lehmann et al., 2004; Bowman et al., 2005; Liu et al., 2005).

Proposed MOA

The MOA for effects of C4–C6 phthalates on the developing testes is likely to be complex, and some of the key events may be parallel to those seen for carcinogenesis (Klaunig et al., 2003). For example, phthalates must undergo hydrolysis to be biologically active, monoesters interfere with GJIC and affect cell proliferation. Likewise for testicular effects, exposure to C4–C6 phthalate esters either in utero or perinatal is associated with characteristic biochemical and cellular changes. Although not all C4–C6 phthalate esters have been studied in utero, DBP, BBP, and DEHP have been extensively studied and have shown similar results in offspring from treated pregnant rats (Table 1). Because the observed responses are similar, these phthalates are believed to act via the same MOA (potencies may vary, however), and evidence for key events may be taken from any of these phthalates known to cause deformities of the male reproductive tract.

The proposed MOA for any effects of phthalate esters has as the key first step the hydrolysis of the diester to the monoester (Albro and Thomas, 1973; White et al., 1980, 1983). This step is believed to be required prior to absorption from the gastrointestinal tract and is also believed to be a requisite to biological activity based on responses of cultured adult testes. Gray and Beamand (1984) demonstrated that the monoester was more active than the diester on adult Sertoli cells in culture. Incubation of primary testicular cells with monoester resulted in detachment of germ cells within 24 hours, whereas no detachment was seen following incubation with the diester. Similar results were obtained by Sjöberg et al. (1986) who treated adult animals and cultured testicular cells from adult animals with DEHP, MEHP and a number of oxidative metabolites. MEHP was a more potent testicular toxicant than DEHP or downstream metabolites V (2-ethyl-5-carboxypentyl phthalate), VI (2-ethyl-5-oxyhexyl phthalate), or IX (2-ethyl-5-hydroxy-hexyl phthalate) (based on the system of Albro et al., 1973).

While these studies focused testicular toxicity in adults, it is likely that the in utero toxicity on the developing male reproductive tract occurs via the same active metabolite(s). Ema (2002) reviewed in vivo studies conducted in his laboratory with DBP and MBP showing that pregnant rats treated with MBP had male pups with the same increased anogenital distance and incidence of cryptorchidism as dams treated with DBP (Ema and Miyawaki, 2001). Li and Kim (2003) found that cultured fetal or neonatal rat Sertoli cells incubated with MEHP had lower levels of AMH and GATA-4. Li et al. (1998) also showed that cultured rat neonatal Sertoli cells incubated with MEHP had lower levels of proliferation, and gonocytes detached from the culture media whereas these effects were not seen when DEHP was added to the culture medium.

On the other hand, a recent study suggested that in a human breast cell line, metabolites VI and IX were more potent at decreasing DHT activity than was MEHP (Stroheker et al., 2005); however, T production was not altered when metabolites were incubated with fetal testes as has been observed in vivo suggesting that the results in human breast cell lines may not reflect testicular cell responses. Thus, the initial key event is hydrolysis of the diesters to the monoesters, which may be the ultimate toxicants.

Once absorbed and hydrolyzed by the dam, the monoesters have been shown to cross the placenta and reach the fetus (Kihlstrom, 1985; Saillenfait et al., 1998; Fennell et al., 2004; Kessler et al., 2004; Calafat et al., 2006). Recent data have demonstrated that free MBP is present in the amniotic fluid on gestation day 20, just 4 hours after treatment of pregnant rats (Fennell et al., 2004), and MBP or MEHP present on GD 18 after repeated dosing (Calafat et al., 2006). Once the monoester reaches the fetus it can interact with the target cells. Two targets have been proposed: the primordial Sertoli cell, and the primordial Leydig cell (Corton and Lapinskas, 2005). The effects on each cell may be quite different, and it has not been determined if one cell type is the primary target and other a secondary target. Leydig cells are influenced by Sertoli cells via exocrine secretions (Huleihel and Lunenfeld, 2004), and gonocytes are influenced by Sertoli cells probably via GJIC (Kleymenova et al., 2005; Yu et al., 2005). Thus, while a primary effect may be observed in one cell type, the ultimate effect may be on another cell as suggested by Mylchreest et al. (2002) and Fisher et al. (2003). Therefore, the key events that follow reflect independent end results that may be parallel or interrelated depending on the target cell and final observation.

Once they reach the fetus, the monoesters interfere with steroidogenic processes and insl3 production in the Leydig cell that were triggered by SF1. There are no data indicating that SF1 is a direct target of phthalate exposure, but certain key enzymes triggered by SF1 have reduced activities resulting in lower production of T, dihydrotestosterone, and insl3. Indeed, evidence suggests that SF1 in the Sertoli cell is unaffected at least with respect to downstream phenomena associated with SF1 in the Sertoli cell such as AMH (Kai et al., 2005; Mahood et al., 2005) and the disappearance of the Müllerian ducts. On the other hand, Sertoli cell communication with the germ cells is disrupted leading to fewer and disorganized gonocytes (Kleymenova et al., 2005), which could be the consequence of a direct effect on the Sertoli cell. Thus, phthalate monoesters act on the Leydig cell to interfere with production of T (and subsequently lowering the availability of its metabolite, DHT) to produce hypospadias and other T-dependent phenomena such as areolas and anogential distance. The low T levels also inhibit Sertoli cells from proliferating, as reflected in lower cyclin D expression in neonatal cells (Li et al., 2000). The fewer Sertoli cells, perhaps in combination with disrupted cytoskeleton that leads to decreased GJIC from Sertoli cells to gonocytes, results in a mass of poorly differentiated gonocytes with multinuclei (Figure 1).

Key Events

The key events in the MOA (Table 3) begin with the hydrolysis of the diester to the monoester, which facilitates absorption from the gastrointestinal tract of the dam as indicated by data obtained from studies with adult intestine (White et al., 1980, 1983). The monoester then is transported across the placenta to reach the fetus. Once present, it can act on primordial Leydig cells and possibly Sertoli cells. At this point, several concurrent pathways are triggered, each having a distinct consequence.

In the first pathway (pathway “a” in Table 3), phthalate exposure lowers expression of genes for cholesterol transport and steroidogenic enzyme activities in the Leydig cell. This key event leads to a decrease in testicular T production. As a consequence, androgen-dependent tissue differentiation is adversely affected. The physical manifestation of this event is hypospadias, underdeveloped secondary male sex organs, and likely areolas and decreased anogenital distance. As with peroxisome proliferation and carcinogenesis, there is a cascade of events that occurs following exposure including effects on gene expression. Shultz et al. (2001) showed that several genes were up-regulated (TRPM-2, IGFBP-3, PDGF-associated protein, carboxypeptidase D precursor, ERK-2, PCNA, gelatinase) and others down-regulated (including StAR, P450_scc, P450c17, SR-B1 for steroidogenesis, LH receptor) in fetal testes of male rats from dams exposed to 500 mg/kg DBP from GD12-GD21. Barlow et al. (2003) confirmed these results using the same dose level of DBP with the addition that the 3β-HSD, FSH receptor, c-kit, and stem cell factor (SCF) genes were also down-regulated. Lehman et al. (2004) demonstrated further that dose-response changes in gene expression with 3β-HSD and c-kit down-regulated at 0.1 mg/kg, and SR-B1, StAR, P450scc down-regulated at 50 mg/kg at which dose fetal T was decreased. A comparison of the results of gene expression (Lehman et al., 2004) and pathological studies (Mylchreest et al., 1999, 2000; Barlow and Foster, 2003) revealed that while steroidogenesis genes were down regulated at 50 mg/kg, this was associated with retained nipples and/or areolas that were present at 5 and 50 mg/kg (Mylchreest et al., 2000). Similar effects on StAR gene expression (mRNA) have been observed in cultures of adult Leydig cells exposed to MEHP (Zhu et al., 2005).

The effects of altered gene expression are to decrease cholesterol transport into the Leydig cells (Shultz et al., 2001; Gazouli et al., 2002; Barlow et al., 2003) and to decrease T production through lower androgen biosynthesis enzyme activity (Parks et al., 2000; Akingbemi et al., 2001; Zhu et al., 2005). Interruption of T synthesis results in a failure of androgen-dependent tissues to develop. Mylchreest et al. (1998) and Parks et al. (2000) showed that animals treated with either DBP or DEHP had hypospadias. The effects on altered expression of steroidogenesis genes can be correlated with intratesticular T levels. Lehmann et al. (2004) showed that decreases in T levels occurred at dose levels of 50 mg/kg for DBP, and Akingbemi et al. (2001) showed decreased T synthesis at 10 mg/kg for DEHP. Mylchreest et al. (2000) reported no malformations at 50 mg/kg although there was an increased incidence of retained nipples and/or areolas at that dose, and Akingbemi reported no morphologic changes and no histopathologic changes in the testes of rats exposed in utero to 100 mg/kg DEHP. The biological significance of retained nipples and areolas is unclear, although Barlow et al. (2004) showed that they are retained for up to 180 days postnatal and that a high number of retained nipples or areolas were correlated with reproductive lesions. An NTP report of a 2-generation study of DEHP reported small testes in a few animals at 4.8–7.9 and/or 14–23 mg/kg (NTP, 2004).

In the second pathway (pathway “b” in Table 3), exposure lowers gene expression for insl3. This sequence is labeled in Table 2 as path ‘b.’ Decreased levels of insl3 (Wilson et al., 2004) result in malformations of the gubernacular ligament following exposure to C4–C6 phthalates, and this ligament is necessary for testicular decent into the scrotal sac (Nef and Parada, 1999; Zimmermann et al., 1999). Liu et al. (2005) showed that the gene that translates for insl3 protein is down-regulated by C4–C6 phthalates.

Parallel to the effects on the Leydig cell are effects on the Sertoli cell and the gonocyte. Sertoli cells stop proliferating and begin to differentiate; in addition, multinucleated gonocytes are observed in the seminiferous tubules (path ‘c’ in Table 2). The key events leading to these observations are not clear and may interplay. Atanassova et al. (2005) recently showed, for example, that lower levels of androgen and other hormones such as FSH can lower the number of Sertoli cells. Using DES to suppress T production, Atanassova et al. (2005) demonstrated that neonatal rats had lower numbers of Sertoli cells and that the nuclear volume was decreased. Thus, decreased T production caused by exposure to C4–C6 phthalate esters could reduce the number of Sertoli cells. But it is also possible that phthalate esters have direct effects on the Sertoli cell. Li and co workers (2000) demonstrated that DEHP exposure of neonatal Sertoli cells in culture decreased cyclin D2 expression and decreased Sertoli cell proliferation. Thus, there are two potential, but not exclusive pathways that lead to reduced numbers of Sertoli cells.

Lower Sertoli cell numbers, in turn, decrease germ cell volume (Yu et al., 2005). It is plausible that a reduced number of Sertoli cells could adversely affect the developing gonocyte such that replicating gonocytes might lack the intercellular signals to properly divide (Meachem et al., 2001). Because Sertoli cells communicate with and are required for proliferation of gonocytes, a reduction in the number of Sertoli cells leads to fewer gonocytes (Atanassova et al., 2005; Kleymenova et al., 2005). In fact, a decrease in GJIC following treatment of Sertoli cells with DEHP has been reported (Kang et al., 2002). This decrease in intracellular communication can lead to dysgenesis of gonocytes (Yu et al., 2005) similar to that observed following treatment with C4–C6 phthalates (Mylchreest et al., 2002; Barlow et al., 2003; Fisher et al., 2003). Kleymenova and coworkers (2005) recently demonstrated that exposure to DBP alters the cytoskeleton in Sertoli cells and interferes with communication with gonocytes. Inhibition of GJIC is a common associative event with peroxisome proliferators (Klaunig et al., 2003), and could explain the observation of Leydig cell adenomas observed for many peroxisome proliferators (Klaunig et al., 2003). Thus, the impaired communication between Sertoli cells and gonocytes might lead to the inability of dividing gonocytes to properly segregate.

Strength, Consistency, Specificity of Association of Response with Key Events/Dose-Response Relationship/Temporal Association

The data are consistent and sufficient to demonstrate that C4–C6 phthalates affect the steroidogenesis of androgens in fetal testes. Bell (1978), Oishi and Hiraga (1982), and Oishi (1982) showed that exposure to phthalate esters decreased the steroidogenesis in adult rat testes. This result has been confirmed in more recent studies of fetal and prepubertal testes (Parks et al., 2000; Akingbemi et al., 2001; Borch et al., 2004) and the consequences of decreased T levels for the developing male are well defined. Therefore, it is plausible that the inhibition of steroidogenesis in the fetal rat testes has a role in the alterations in male reproductive development. Furthermore, the effects on T synthesis have been observed for more than one phthalate ester: data for at least DEHP, DBP, and BBP (Mylchreest et al., 1998; Gray et al., 2000; Borch et al., 2004) have shown similar results on T synthesis in fetal testes leading to consistent malformations associated with lower T levels (hypospadias and lower anogenital distance) providing support for a consistent mode of action among these substances. In addition, there are data to show that DEP does not alter T synthesis in the testes (Gazouli et al., 2002), that DEP and DMP do not alter gene expression for steroidogenesis (Liu et al., 2005), and that DEP and DMP do not produce the malformations in rodents observed with the C4–C6 phthalates (Gray et al., 2000; Liu et al., 2005).

Dose-response data for DBP and DEHP indicate that significant changes in T production are observed at dose levels at which anatomical malformations or effects are produced (Mylchreest et al., 1998, 1999, 2002; Akingbemi et al., 2001; Lehmann et al., 2004). Mylchreest and others have observed frank malformations (hypospadia) at dose levels >50 mg DBP/kg/day. Lehmann et al. also observed significant decreases in T synthesis at >50 mg DBP/kg/d. On the other hand, Akingbemi et al. (2001) showed that decreases in T synthesis occurred at dose levels as low as 10 mg DEHP/kg/d, but histopathologic changes in the testes have not been observed at 100 mg/kg although the NTP two-generation reproductive toxicity study reported small testes at dose levels between 10 and 100 mg/kg (NTP, 2004). Table 1 presents dose-response information for the effects observed.

Data to support the other pathways in this mode of action are not as specific, but are robust. Cryptorchidism has been observed following exposure to DEHP and DBP. Cryptorchidism is associated with failure of the gubernaculums and ligament to develop, processes that are controlled by insl3 and androgen (Emmen et al., 2000; McKinnell et al., 2005). Wilson et al. (2004), Lehmann et al. (2004), and McKinnell et al. (2005) showed that exposure to DBP, DEHP, or BBP decreased insl3 expression and production in fetal males, plausibly associated with an increase in cryptorchidism, and the dose response data obtained by Lehmann et al. (2004) for DBP is consistent with malformations observed by Mylchreest et al. (1998). There are no data on insl3 expression or levels for DEP or DMP.

The temporal relationship of exposure to gestation day has been investigated using DBP demonstrating that the gestational timing of exposure is important to produce the critical effects (Ema et al., 2000; reviewed in Ema, 2002). The temporal relationship between alteration of gene expression and changes in T production has been investigated for DBP (Lehmann et al., 2004; Thompson et al., 2005). Initial increases in gene expression are followed by decreases in gene expression for genes associated with steroidogenesis. Thus, those key events of gene expression are temporally consistent with subsequent events. For the observed decreased steroidogenesis and decreased levels of insl3 protein, these are well established as precursors to anatomical changes in the developing male reproductive tract. Furthermore, Li and Kim (2003) demonstrated that fetal testes from GD18 are more sensitive to incubation with MEHP for these effects than are fetal testes from GD13. AMH levels, GATA-4 expression, and the number of gonocytes from GD13 testes were unchanged after incubation with 50–200 μM MEHP, while AMH and GATA-4 and Sertoli cell proliferation from GD18 testes were reduced at 100 and 200 μM MEHP. This supports the hypothesis that the morphological changes in male reproductive development are due to time-sensitive functions of the fetal testes rather than a consequence of a general cytotoxic effect.

Path c leading to decreased Sertoli cell proliferation and multinucleated gonocytes is plausible based on recent information concerning cell-cell interactions and the direct effect(s) of DBP on Sertoli cell cytoskeleton. The data supporting an effect of lower T levels on Sertoli cells are not sufficient to strongly support this effect as a key event in the MOA. However, decreases in Sertoli cell numbers, because of lower T production, could certainly influence the numbers of gonocytes. Furthermore, impairment of the cytoskeleton leading to inhibition of GJIC is associated with many peroxisome proliferators, and the relationship between the numbers of Sertoli cells and the number of gonocytes is assumed to be finite, i.e., there is a fixed number of gonocytes that can be supported by each Sertoli cell. Therefore, aspects of this pathway are biologically plausible.

Reversibility

There are currently no data on reversibility of morphological effects using pharmacologic doses of T or other hormones that might be involved in the MOA following in utero exposure of phthalates to evaluate the effect of replacement. So the value of this criterion to evaluate the strength of the MOA is limited. There are data to demonstrate that spermatogenesis has some capacity to recover based on sperm count per gram of tissue (Oishi, 1985; Dostal et al., 1988; Cammack et al., 2003), but this rebound is not indicative of any key event or MOA.

Biological Plausibility and Coherence of the Data

The key events leading to a decrease in steroidogenesis are plausible and are consistent with the effects observed. There is coherence in the response of T synthesis for the C4–C6 phthalates that has been tested and for which malformations have been observed. In addition, the lack of effect on T synthesis for DEP or gene expression for T synthesis following DMP, phthalate esters that have not shown developmental toxicity on the male reproductive tract, is consistent with the MOA. Cryptorchidism observed following exposure to C4–C6 phthalate esters can be explained by decreased levels of insl3, a control protein for gubernaculum development and testis descent. At least three phthalate esters have been shown to decrease insl3 levels in the Leydig cell. The plausibility of effects directly on the Sertoli cell that cascade to multinucleated gonocytes is predicated on the cell-cell communication between these two cell types. A role for intracellular communication is plausible as the data support the need for the 2 cell types to interact. The possibility that phthalate esters decrease intercellular communication has already been demonstrated in the liver and testes. Thus, this pathway is plausible. Whether the Sertoli cell is also targeted is not clear. The effect of lower T levels on Sertoli cells is to interfere with proper development and proliferation, which could, in turn, adversely effect the interaction of Sertoli cells and gonocytes.

Alternative MOAs

A role for PPARα in testicular toxicity, either through direct or indirect action, has been suggested by Corton and Lapinskas (2005). Their hypothesis is supported by the observation that circulating T levels in PPARα-null mice were increased following treatment with DEHP compared with a decrease in T levels in wild-type mice suggesting that PPAR has a role in postnatal testicular toxicity. However, there are no data to support or reject the hypothesis that PPARα has a role in the dysgenesis of the male reproductive tract following in utero exposure. PPARα activation may have some role in the developmental toxicity of nonreproductive organs (Lampen et al., 2003), but the data to support PPARα-activation as a key event in in utero effects on the reproductive tract are insufficient. Furthermore, not all C4–C6 phthalate esters that are toxic to the developing male reproductive tract are potent activators of PPARα (Bility et al., 2004).

Other possible key events involve a direct effect on the Sertoli cell. Barlow et al. (2003) reported that DBP lowers expression of genes for the FSH receptor, c-kit, and SCF which the authors felt could lead to decreased cell proliferation in Sertoli cells. Lower levels of c-kit and SCF, in turn, might lead to lower levels of germ cells. The ligand SCF and its receptor c-kit are required for migration of primordial germ cells and the differentiation and proliferation of germ cells (Ohta et al., 2003). Thus, that this series of events are key is plausible, but relies on a direct effect on the Sertoli cell. However, other data suggest that the Sertoli cell is not a primary target: Kai et al. (2005) did not observe changes in AMH levels, typically secreted by the Sertoli cell, following exposure to DBP. If the Sertoli cell were a direct target, it is reasonable to assume that all processes associated with the Sertoli cell would be affected, but this does not appear to be the case. Thus, effects on the Sertoli cell are either indirect, or do not result in decreased exocrine secretions. These observations may be time dependent. Li and Kim (2003) demonstrated that AMH was unaffected in cultured testes from GD13, but was decreased in a dose-dependent manner in cultured testes from GD18. The decrease was associated with impaired Sertoli cell proliferation in postnatal testes. Thus, evidence that the Sertoli cell is the target for in utero exposure is not compelling, but the apparent inconsistencies may be related to the timing of exposure and observation.

Liu et al. (2005) reported that decreased levels of inhibin α from the Leydig cell altered Sertoli cell proliferation. They reported that Inha, the gene that expresses inhibin alpha, is down-regulated by phthalate exposure. This event might complement the events already proposed for indirect effects on the Sertoli cell. There may be other paracrine events involved in the MOA that have not yet been investigated.

Conclusions-assessment of MOA, Statement of Confidence

The data supporting an effect on steroidogenesis and insl3 protein levels are biologically plausible and are supported by data for more than one phthalate. Furthermore, gene expression data for DMP and DEP, which are not known to cause the testicular dysgenesis observed with C4–C6 phthalates, failed to show changes in steroidogenic gene expression characteristic of the C4–C6 phthalates. Decreased steroidogenesis is associated with dysgenesis of male reproductive organs based on data for anti-androgenic substances (Mylchreest et al., 1998), and there are sufficient data to demonstrate the key events for this pathway. In addition, the data supporting the pathway to cryptorchidism are sufficient to support this MOA.

The key events leading to differentiation of Sertoli cells and multinucleated gonocytes are not supported by sufficient data to have confidence that these key events are correct, although a number of the events have been demonstrated with C4–C6 phthalates and recent data suggest that these genes associated with Sertoli cell communications are not down-regulated with DMP or DEP (Liu et al., 2005). Such data should be supported by dose-response information to provide certainty. Furthermore, it is also possible that inhibin α or other paracrine communications between the Leydig cell and Sertoli cell play a role in this pathway. These phenomena need further investigation. Once the events for effects in animals are defined, these events can be investigated in humans and primates to establish concordance.