Reproductive Lesions in Female Harlan Sprague-Dawley Rats Following Two-Year Oral Treatment with Dioxin and Dioxin-like Compounds

Abstract

Results from previously published animal studies suggest that prenatal and postnatal exposure to dioxin and dioxin-like compounds (DLCs) may profoundly affect the reproductive system of both sexes via endocrine disruption. In the present work, we evaluate the toxicity and carcinogenicity of various DLCs, with an emphasis on their effect on the reproductive organs, induced by chronic exposure of female adult Harlan Sprague-Dawley rats. This investigation represents part of an initiative of the National Toxicology Program to determine the relative potency of chronic toxicity and carcinogenicity of polychlorinated dioxins, furans, and biphenyls. For fourteen, thirty-one, or fifty-three weeks or for two years, animals were administered by gavage 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD); 3,3′,4,4′,5-pentachlorobiphenyl (PCB126); 2,3,4,7,8-pentachlorodibenzofuran (PeCDF); 2,2′,4,4′,5,5′-hexachlorobiphenyl (PCB153); 2,3′,4,4′,5-pentachlorobiphenyl (PCB118); a tertiary mixture of TCDD, PCB126, and PeCDF; a binary mixture of PCB126 and 153; or a binary mixture of PCB126 and PCB118.

The ranges of treatment-related changes in the reproductive system included chronic active inflammation in the ovary that occurred in the 1,000 and 3,000 μg/kg core groups (two-year exposure) of PCB153 and in the 300 ng/3,000 μg/kg core group of binary mixture of PCB126 and PCB153. Increases in the incidence of acute and/or chronic active inflammation of the uterus were observed in all dosed groups, including the stop-exposure group (withdrawal after thirty-week exposure) of PeCDF and the 1,000 μg/kg and/or higher group dosed with PCB153. The incidence of cystic endometrial hyperplasia was marginally increased in the 92 PeCDF ng/kg group at two years. The incidence of squamous metaplasia was significantly increased in the 44 ng/kg and higher dose group, including the stop-exposure group. The incidence of uterine squamous cell carcinoma was significantly or marginally increased in the 6 ng/kg core and 100 ng/kg stop-exposure groups of TCDD and in the 300 ng/300 μg/kg core group that received the binary mixture of PCB126 and 153. The incidence of uterine carcinoma was marginally increased in the 92 ng/kg PeCDF group at two years and clearly increased in the 1,000 and 4,600 μg/kg PCB118 core group and the 4,600 μg/kg stop group. In the studies of PCB 126, the tertiary mixture, and the binary mixture of PCB126 and PCB118, no increased incidence of any change occurred in the reproductive systems. The range of changes seen with the different compounds suggests that more than one mechanism may have been involved in promoting the female reproductive pathology.

Introduction

2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) (commonly referred to as dioxin), certain polychlorinated dibenzodioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and coplanar polychlorinated biphenyls (PCBs) possessing the ability to bind to the aryl hydrocarbon receptor (AhR) and exhibit biologic actions similar to those of TCDD, are termed dioxin-like compounds (DLCs). They may induce developmental, endocrine, and immunological toxicity; chloracne and hyperpigmentation; and multi-organ carcinogenicity in animals and/or humans (Agency for Toxic Substances and Disease Registry [ATSDR] 1998 Agency for Toxic Substances and Disease Registry [ATSDR] 2000; Bertazzi et al. 2001; Kociba et al. 1978; Schecter et al. 2006; Steenland, Deddens, and Piacitelli 2001). Incidence of cancer has been evaluated in several analyses of human populations exposed to elevated amounts of dioxin and DLCs. The most recent studies indicate that exposure is associated with an increase in all cancers combined and several specific cancers including rectal cancer, lung cancer, Hodgkin’s disease, non-Hodgkin’s lymphoma, and myeloid leukemia (Bertazzi et al. 2001). 2,3,7,8-Tetrachlorodibenzo-p-dioxin was listed as a Group 1 carcinogen (nongenotoxic, carcinogenic to humans) by the International Agency for Research on Cancer in 1997 (IARC carcinogen classification on the Web: http://www.iarc.fr/ENG/Databases/index.php, IARC 1997), whereas PCBs have been categorized as a Group 2A carcinogen (probably carcinogenic to humans). Reproductive cancer risks have not been reported clearly following the “Seveso Accident” in 1976, which caused exposure to TCDD of a large residential population at a chemical plant in Seveso, Italy (ATSDR 1998; Bertazzi et al. 2001; Pesatori et al. 2003). Information highlighting health effects of PCBs has been derived from various studies of people exposed in the work-place, consumption of contaminated rice oil in the Japanese “Yusho Accident,” ingestion of contaminated fish in the Taiwanese “Yu-Cheng Accident” (Masuda et al. 2007), consumption of contaminated fish and other food products of animal origin, and general environmental exposure. Evidence suggests that PCBs can produce a shorter menstrual cycle, reproductive abnormalities, and reduced fecundity among couples attempting pregnancy, but not with increased risk of conception delay (ATSDR 2000; Schecter et al. 2006).

The National Toxicology Program (NTP) conducted 8 two-year bioassays in female rats to evaluate the chronic pathology and carcinogenicity induced by DLCs and structurally related PCBs, and mixtures of these compounds, including TCDD; 3,3′,4,4′,5-pentachlorobiphenyl (PCB126); 2,3,4,7,8-pentachloro-dibenzifyran (PeCDF); 2,2′,4,4′,5,5′-hexachlorobiphenyl (PCB153); 2,3′,4,4′,5-pentachlorobiphenyl (PCB118); the toxic equivalency factor (TEF) tertiary mixture of TCDD, PCB126, and PeCDF; the binary mixture of PCB126 and PCB153; and the binary mixture of PCB126 and PCB118 (NTP 2004a–g). In these studies, an increase occurred in the incidence of squamous cell tumors, such as squamous cell carcinoma (SCC) of the oral cavity and/or uterus, and cystic keratinizing epithelioma of the lung (NTP 2004a, 2004g; Walker et al. 2005; Walker et al. 2007; Yoshizawa, Walker et al. 2005). Moreover, administration of these DLCs caused increased incidence of nonneoplastic lesions in multiple organs, including the reproductive system (Brix et al. 2004; Jokinen et al. 2003; Nyska et al. 2004; Nyska et al. 2005; Tani et al. 2004; Yoshizawa, Marsh et al. 2005; Yoshizawa, Walker et al. 2005; Yoshizawa et al. 2007). Additionally, in the NTP studies, TCDD, PCB126, and/or PeCDF induced neoplastic lesions (hepatocellular and/or cholangiolar, pancreatic, thyroidal, and pulmonary) and non-neoplastic effects (hepatocellular and/or cholangiolar lesions in the liver, squamous hyperplasia in the oral cavity, cardiovascular damage, thyroidal hypertrophy, and immunosuppression) similar to those reported in humans (Yoshizawa et al. 2007).

This article, one of a series highlighting specific findings from the NTP dioxin TEF evaluations, focuses on the incidence and morphologic aspects of reproductive lesions across these investigations. In addition, we discuss literature regarding DLC-related reproductive pathology and potential mechanisms of these lesions.

Materials and Methods

Study Design and Animals

These studies of female Harlan Sprague-Dawley (SD) rats (Harlan Laboratories, Indianapolis, IN, USA) were performed in accordance with the Good Laboratory Practices (Food and Drug Administration) at Battelle Columbus Laboratories in Columbus, Ohio, USA. The rats were approximately six weeks of age upon receipt, underwent health screening during a quarantine period of two weeks, and were released for study when they were about eight weeks old. The animals were randomly assigned to their respective experimental groups and permanently identified by tail tattoo. Housing was five per cage in solid-bottom polycarbonate cages (Lab Products, Inc., Maywood, NJ) suspended on stainless steel racks. Filtered room air was supplied at the rate of at least ten room changes per hour. The rats were maintained at 69°F–75°F, in relative humidity of 35% to 65%, and with a twelve-hour light/dark cycle. They were watered and fed ad libitum. The food was irradiated NTP-2000 pellets from Zeigler Bros., Inc. (Gardners, PA, USA). All rats were observed twice a day for morbidity and once a month for formal clinical signs of toxicity. Moribund animals were sacrificed and necropsied. Health monitoring via sentinel animals showed no evidence of any significant rodent pathogens. Animal husbandry was done in accordance with the guidelines of the National Institutes of Health (Institute of Laboratory Animal Resources).

Groups of female SD rats were treated by gavage with several doses of TCDD, PCB126, PeCDF, PCB153; a “TEF mixture” of TCDD, PCB126, and PeCDF; a binary mixture of PCB126 and PCB153; and a binary mixture of PCB126 and PCB118 (Table 1). The rats were dosed once daily for five days per week for fourteen, thirty-one, or fifty-three weeks or two years (core group). Additionally, the stop groups (withdrawal after thirty-week exposure) of some studies were included and examined at two years. As control groups, the rats were dosed with a corn oil vehicle in each study. Doses used were based on values of the TEF selected by the World Health Organization (Van den Berg et al. 1998; Van den Berg et al. 2006). The rationale for dose selection in each study was summarized in the original NTP study reports (NTP 2004a–g).

Chemicals

Dose formulations of TCDD (The IIT Research Institute, Chicago, IL, USA), PCB126 (AccuStandard, Inc., New Haven, CT), PeCDF (Cambridge Isotope Laboratories, Cambridge, MA), PCB153 (Radian International LLC, Austin, TX, USA), and PCB118 (Radian International) were prepared by mixing the test articles in a corn oil vehicle containing 1% USP-grade acetone. Dose formulations were prepared fresh on a monthly basis; the mixtures were stable for that period.

Pathology

Moribund rats and all rats scheduled to be sacrificed were euthanized by using carbon dioxide. All study rats were necropsied according to standard procedures; complete tissue sets of systemic organs, including all gross lesions and masses, were fixed in 10% neutral-buffered formalin. The reproductive systems examined were bilateral ovaries (including oviducts), uterus (including cervix), and vagina. All collected tissues were trimmed after fixation, processed, and embedded in paraffin. The tissues were then sectioned at a thickness of approximately 5 μm, stained with hematoxylin and eosin, and examined via light microscopy. The severity of lesions was graded on a four-point scale of 1 = minimal, 2 = mild, 3 = moderate, and 4 = marked. The pathology findings underwent comprehensive NTP peer reviews and Pathology Working Group reviews (Boorman et al. 2002).

Statistical Analysis

The probability of survival was estimated by the Kaplan-Meier product-limit procedure (Kaplan and Meier 1958). Incidence of lesions was evaluated statistically by the poly-k test, which makes adjustments for survival differences among groups. This test is a survival-adjusted quantal-response procedure that modifies the Cochran-Armitage linear trend test to take survival differences into account. More specifically, this method modifies the denominator in the quantal estimate of lesion incidence to approximate more closely the total number of animal years at risk. For analysis of a given site, each animal is assigned a risk weight. This value is 1 if the animal had a lesion at that site or if it survived until terminal sacrifice; if the animal died prior to terminal sacrifice and did not have a lesion at that site, its risk weight is the fraction of the entire study time that it survived, raised to the kth power. In the present studies, the value of k was 3. For animals in the two-year studies, the total-lesion incidence, including findings from animals that survived until study termination and early-death animals, were included in the analysis.

Results

All individual animal data for these studies, found on the NTP Web site (http://ntp-apps.niehs.nih.gov/ntp_tox/index.cfm), are summarized in individual technical reports on the evaluation of carcinogenicity. Treatment-related lesions of the selective reproductive systems were commonly observed in this series of investigations and are presented by study in Tables 2 to 9.

Incidence of Ovarian Lesions

Ovarian Atrophy:

At fourteen weeks in the study of the TEF mixture of TCDD, PCB126, and PeCDF, there was increased incidence of moderate to marked ovarian atrophy at 100 ng TEQ/kg group (Table 5). Incidence of ovarian atrophy was increased without statistical significance in the 1,000 ng/kg PCB126 fourteen-week interim group. In contrast, the incidence was decreased without statistical significance in the 100 ng/kg two-year group in the TCDD study, 1,000 ng/kg thirty-one–week interim and two-year groups in the PCB126 study, and 216 and/or 360 ng TEQ/kg thirty-one–week and longer groups in the mixture study of PCB126 and PCB118. Atrophy was characterized not only by overall reduction in size as well as absence of ovarian structures, primarily corpora lutea, but also by lack of follicles in some cases.

Chronic Active Inflammation:

At two years, in the study of the TEF mixture of TCDD, PCB126, and PeCDF, increased incidence of moderate to marked chronic active inflammation occurred with atrophic ovary and were significantly increased in the 100 ng TEQ/kg group (Table 5). At two years in the PCB153 study, significantly increased incidence of chronic active inflammation in the ovary occurred in the 1,000 and 3,000 μg/kg core study groups (Table 6). At two years in the binary mixture study of PCB126 and PCB153, a positive trend occurred in 300 ng PCB126/3,000 mg PCB153/kg (Table 8). Those ovaries affected by the inflammation were nearly or completely replaced by large aggregates of neutrophils and debris surrounded by a layer of fibrous tissue that was infiltrated with macrophages. Chronic active inflammation of the ovary essentially consisted of an encapsulated abscess (Figure 1A, 1B). In the oviduct, this inflammation exhibited a similar microscopic appearance, suggesting that it may have been an extension of the ovarian lesion (data not shown).

Incidence of Uterine Lesions

Squamous Cell Carcinoma:

The incidence (n = 5) of SCC in the 46 ng TCCD/kg group at two years in the TCDD study was significantly greater than that in the vehicle controls; two SCCs were found in the 100 ng TCDD/kg stop-exposure group (Table 2). The historical incidence of SCC is 2/473 female Harlan SD rats in the NTP studies (data not shown). Squamous cell carcinoma occurred on the endometrial surface, often caused dilatation of the uterus, and was characterized by irregular cords and clusters of atypical stratified squamous epithelial cells that invaded the underlying myometrium (Figure 2A, 2B). One SCC appeared in the uterus of a 92 ng PeCDF/kg rat in the PeCDF study (Table 4). At two years in the TEF mixture study, a single incidence of SCC occurred in the 22 ng TEQ/kg group (Table 5). The incidence of squamous cell carcinoma of the uterus was as follows: zero for the controls and the 100 and 4,600 μg PCB118/kg dose groups; three cases in the 220 μg PCB118/kg-dosed group; and a single case in each of the 460 and the 1,000 μg/kg PCB118-dosed two-year groups (Table 7). A single case of this tumor was noted in the 4,600 μg/kg stop-study group. The incidence in the 220 μg PCB118/kg group was higher than the historical control incidence. At two years in the mixture study of PCB126 and PCB153, the incidence of uterine SCC was increased in the 300 ng PCB126/300 mg PCB153/kg group (4/53); single incidences of this lesion occurred in the vehicle control, 10 ng PCB126/10 μg PCB153/kg, and 100 ng PCB126/100 μg PCB153/kg groups (Table 8). The incidence of uterine SCC in the 300 ng PCB126/300 μg PCB153/kg group exceeded the historical vehicle control range.

Cystic Endometrial Hyperplasia:

The incidence of cystic endometrial hyperplasia (CEH) was significantly lower in the 100 ng TCDD/kg core study group compared to vehicle controls and were significantly less than those in the 100 ng TCDD/kg stop-exposure group (Table 2). The incidence of CEH was significantly decreased in the 1,000 ng PCB126/kg core study group, compared to vehicle control and the 100 ng PCB126/kg stop-exposure groups (Table 3). In the thirty-one–week interim group in the PeCDF study, the incidence and severity of CEH in the 200 ng PeCDF/kg animals were increased (Table 4, Figure 3A, 3B). Cystic endometrial hyperplasia, a common aging change in female rats, was characterized by varying degrees of hyperplasia of the endometrial epithelium with concurrent dilatation of endometrial glands. At two years in the PeCDF study, increases in the incidence of squamous metaplasia (SM) of the uterus occurred in the 44, 92, and 200 ng PeCDF/kg core study groups, and increased incidence of CEH was documented in the 92 and 200 ng PeCDF/kg groups. These lesions were detected in the 200 ng PeCDF/kg stop group with a similar tendency to that in the core group. In the fifty-three–week interim group in the TEF mixture study, the incidence of CEH in all treatment groups was decreased compared with control groups. A decreased incidence of CEH was noted in the 100 ng TEQ/kg two-year group of the TEF mixture study (Table 5). In the fifty-three–week group of the PCB153 study, the incidence of CEH was decreased in the 1,000 μg PCB153/kg group; however, there was no statistical significance with the incidence of CEH among the groups in the two-year PCB153 study (Table 6). The incidence and/or severity of CEH were decreased in the group of rats treated for two years with 4,600 μg/kg PCB118 (Table 7). In the fifty-three–week group of the mixture study of PCB126 and PCB153, the incidence of CEH was decreased in all treatment groups (Table 8). Incidence of CEH was decreased in the 1,000 ng PCB126/1000 μg PCB153/kg two-year group of the mixture study of PCB126 and PCB153 compared to the control and other treated groups. In the thirty-one–week, fifty-three–week, and two-year groups of the mixture study of PCB126 and PCB118, the incidence of CEH was decreased in the groups that received the highest doses (Table 9).

Squamous Metaplasia:

The incidence of SM was significantly lower in the 100 ng/kg TCDD core study group compared to vehicle controls and was significantly less than that in the 100 ng/kg stop-exposure group (Table 2). The incidence of SM was significantly decreased in the 1,000 ng PCB126/kg core study group, compared to vehicle control and the 100 ng PCB126/kg stop-exposure groups (Table 3). After two years in the PeCDF study, increases in the incidence of SM occurred in the 44, 92, and 200 ng PeCDF/kg core study groups, and these lesions were detected in the 200 ng PeCDF/kg stop group with a similar tendency to that in the core group (Table 4). Compared to controls, an increase was found in the incidence of SM in all two-year treatment groups of the TEF mixture study (Table 5). The incidence and/or severity of SM were decreased in the group of rats treated for two years with 4,600 μg/kg PCB118 (Table 7). Compared to the control and other treated groups, incidence of SM was decreased in the 1,000 ng PCB126/1,000 μg PCB153/kg two-year group of the mixture study of PCB126 and PCB153 (Table 8). In the thirty-one-week, fifty-three-week, and two-year groups of the mixture study of PCB126 and PCB118, the incidence of SM was decreased in the groups that received the highest doses (Table 9).

Suppurative and Chronic Inflammation:

The incidence of suppurative inflammation was significantly lower in the 550 and 1,000 ng PCB126/kg core study groups (Table 3). Chronic active inflammation was detected in all two-year treatment groups in the PeCDF study, including the 200 ng PeCDF/kg stop group (Table 4). In the two-year group of the PCB153 study, incidence of suppurative inflammation in the 1,000 μg PCB153/kg group and chronic active inflammation in the 3,000 μg PCB153/kg core study group was significantly greater than those in the vehicle control groups (Table 6). These lesions consisted of an accumulation of small to moderate numbers of neutrophils, usually mixed with varying amounts of keratin, within the uterine lumen, accompanied by a diffuse infiltrate of small numbers of eosinophils within the myometrium (Table 6, Figure 5A, 5B). The microscopic appearance of the change was not similar to that seen normally during estrus, but it suggested the possibility of an endocrine influence of the chemical on the uterus. Chronic active inflammation consisted of an accumulation of a mixture of varying numbers of neutrophils and macrophages within the uterine lumen and extending into the underlying myometrium (Table 6, Figure 5B). Decreased incidence of suppurative inflammation and chronic active inflammation was also noted in the higher-dosed two-year groups of the PCB126 and PCB118 study (Table 9).

Uterine Carcinoma:

At two years in the PeCDF study, the incidence of uterine carcinoma in the 92 ng PeCDF/kg group was marginally greater than that in the vehicle controls, and two carcinomas occurred in the 200 ng PeCDF/kg group (Table 4). This incidence of carcinoma exceeded the historical vehicle control range; the historical incidence for two-year gavage studies with Harlan SD vehicle control groups is 6/473, in the range of 0% to 4%. Carcinoma consisted of sheets of poorly differentiated ovoid to spindloid epithelial cells, often with invasion into the underlying myometrium. Squamous cell carcinoma occurred on the endometrial surface, caused dilatation of the uterus, and was characterized by irregular cords and clusters of atypical stratified squamous epithelial cells that invaded the underlying myometrium (Figure 4A, 4B). Proliferations of neoplastic glands and infiltrated neoplastic cells were present in the adipose tissue and mesometrium surrounding the uterus. Extensive inflammation, predominantly neutrophilic, appeared associated with the neoplasms. Increased incidence of uterine carcinoma also occurred following treatment with PeCDF at 92 ng PeCDF/kg–treated group.

In the two-year group of the PCB118 study, the incidence of carcinoma of the uterus was two for the controls; one to three for the 100, 220, and 460 μg PCB118/kg groups (Table 7, Figure 6A, 6B); three to four for the 1,000 μg PCB118/kg and 4,600 μg PCB118/kg groups; and eleven for the 4,600 μg PCB118/kg stop-study group. The incidence in the 4,600 μg PCB118/kg stop group was higher than the historical control incidence (NTP historical data 6/473).

Discussion

Aryl hydrocarbon receptor (AhR), a ligand-activated transcription factor that mediates the biological and toxicological effects of TCDD and related DLCs (Denison and Nagy 2003; Nebert et al. 2004; Poland and Knutson 1982; Schmidt and Bradfield 1996), is expressed ubiquitously in human and rodent systemic organs, particularly in the reproductive system (Chaffin, Trewin, and Hutz 2000; Pohjanvirta and Tuomisto 1994; Valdez and Petroff 2004). Most organs with AhR expression are thought to be more susceptible to TCDD-induced toxicity, as evidenced by AhR-deficient mice, which are resistant to some kinds of TCDD-induced toxicity, including uterine toxicity (Fernandez-Salguero et al. 1996; Gonzalez and Fernandez-Salguero 1998). The most well-studied response to TCDD and DLCs is induction of the 1A1 classes of cytochrome P450 (CYP) (Denison and Nagy 2003; Schmidt and Bradfield 1996) regulated by the AhR (Nebert et al. 2004; Vrzal, Ulrichova, and Dvorak 2004). 2,3,7,8-Tetrachlorodibenzo-p-dioxin elicits a number of AhR-dependent antiestrogenic responses in the female reproductive tract, including inhibition of estrogen-induced increases in uterine wet weight, DNA synthesis, and gene-expression responses (Boverhof et al. 2006; Klinge et al. 1999; Safe and Wormke 2003). The endocrine disruptive actions of PCBs can be a result of the parent PCB compounds, the enzymes induced by the parent PCB, or an action of metabolites in vivo (Carpenter 2006; Garner et al. 1999). The coplanar and AhR-activating PCBs, for example, are antiestrogenic because they induce P450s 1A1 and 1B1. Dioxin-like compounds have the ability to disrupt estrogen signally by several mechanisms including induction of CYP1 family P450s that have the ability to alter estrogen metabolism, by altering the expression of estrogen responsive genes and through altered estrogen receptor signaling (Rahman et al. 2006; Safe et al. 1998).

The incidence of ovarian atrophy was increased in fourteen-week interim groups of PCB126 and TEF mixture studies. A diminished number of corpora lutea were noticed after a single exposure to 10 to 32 μg/kg and thirteen weeks after continuous daily exposure of SD rats to 1 μg/kg TCDD (Kociba et al. 1976: Pohjanvirta and Tuomisto 1994; Roby 2001). In a two-year study (Kociba et al. 1976), cytoplasmic foaminess and nuclear hyperchromatism in the interstitial gland cells of the ovarian stroma were reported at the same dose levels. These lesions suggest that the development and maturation of follicles and the blockage of ovulation are disrupted, probably followed by suppression of the estrous cycle (Gao et al. 1999; Poland and Knutson 1982; Roby 2001). Ovarian atrophy can result from reduced gonadotropin secretion following stress, reduction of food intake, or a stimulatory/inhibitory effect on the central nervous system by chemical exposure; impairment of follicular growth resulting from direct or indirect damage of oocyte/granulosa cells; or enzymatic interference, leading to a reduction of synthesis of sex steroidal hormones caused by chemicals that have estrogenic, progesteronic, or combined activities (Yuan and Foley 2002). Estrogen and genistein can induce ovarian atrophic changes in rodents (Biegel et al. 1998; Hart 1990; McClain et al. 2006). In general, other hormone-sensitive tissues are also affected (e.g., uterus, vagina) when there is sufficient hormonal dysregulation to affect the ovary (Hernandez-Ochoa et al. 2009). However, in the early interim groups of the present PCB126 and TEF mixture studies, except in the ovaries, no changes in histopathology were detected in hormonal organs.

Chronic active inflammation in the ovary occurred in the PCB153 study and binary-mixture study of PCB126 and PCB153. Increases in the incidence of acute and/or chronic active inflammation of the uterus were observed in the PeCDF and PCB153 studies. These effects have not been reported in previous studies of PCB153. Similar effects, such as inflammation of the ovary and uterine pyometra, may be estrogenic effects and have been observed following exposure to the synthetic estrogen diethylstilboesterol (DES) and estrogen (Durr 1975; Hart 1990; McLachlan, Newbold, and Bullock 1980; Pandey et al. 2005). Estradiol alters the nature of the endometrial epithelium, in which estrogen receptors are located in the rodent and human uterus; the induction of purulent endometritis is dependent on ovarian steroid hormones, such as estrogen (Durr 1975). Several studies suggest that PCB153 possesses estrogenic properties and competes with estrogen at the estrogen receptor (Bonefeld-Jorgensen et al. 2001; Desaulniers et al. 1999; Li, Zhao, and Hansen 1994; Wojtowicz et al. 2000). The estrogenic potential of PCB153 and PeCDF may have contributed to the induction of these effects in the present NTP studies.

In the PeCDF study, the incidence of squamous metaplasia was significantly increased in the 44 ng/kg and higher dose groups, including the stop-exposure group. The incidence of SCC of the uterus were increased in the TCDD study and the binary-mixture study of PCB126 and PCB153. In the TCDD study, a significantly increased incidence of SCC of the uterus occurred in animals administered 46 ng/kg only. In addition, two animals exhibited SCC in the stop-exposure group. Although there were no neoplasms in the 100 ng/kg core study group, the observed neoplasms were considered to be related to TCDD treatment. The cause of the absence of SCC in the high-dose group is unknown but may be the result of a suppressive effect of reduced gain in body weight seen at this dose. In previous rat studies, in which the dose and route of administration were similar (Kociba et al. 1978; Walker et al. 2006), no effect on uterine SCC induction was observed.

2,3,7,8-Tetrachlorodibenzo-p-dioxin induces several AhR-mediated changes in gene expression and tissue-/species-specific toxicities and is a potent inhibitor of estrogen-mediated activity; both tumorgenic and anticarcinogenic responses occur, including inhibition of estrogendependent uterine and mammary tumor formation and growth via inhibitory AhR–estrogen receptor cross-talk (Safe and McDougal 2002; Safe et al. 1998; Safe and Wormke 2003; Takemoto et al. 2004). 2,3,7,8-Tetrachlorodibenzo-p-dioxin and other AhR ligands suppress estradiol-induced responses in the rodent uterus, mammary tumors, and human breast cancer cells (Wormke et al. 2003). That TCDD and DLCs disrupt retinoid homeostasis and cause vitamin A deficiency (Fattore, Trossvik, and Hakansson 2000; Ray and Swanson 2003; Schmidt et al. 2003; Van Birgelen et al. 1995) in the systemic organs (Fiorella, Olson, and Napoli 1995; Fletcher, Hanberg, and Hakasson 2001; Hakansson et al. 1991) must also be considered. Hepatic vitamin A is reduced in rats following dietary exposure to TCDD for thirteen weeks (Hakansson et al. 1994; Van Birgelen et al. 1995). A characteristic of retinoid deficiency is abnormal epithelial differentiation to a keratinized squamous phenotype (Everts, Sundberg, and Ong 2005; Lancillotti et al. 1992; Lotan 1994). The action of DLCs might, therefore, be a disruption of retinoid action, leading to altered growth and differentiation of the endometrial epithelium resulting in SM and SCC. We reported that, in the same studies, TCDD and DLCs were documented to induce epithelial SM and SCC in other organs, including gingival squamous-cell hyperplasia and/or SCC, SM, and/or SCC in the lung, and squamous hyperplasia in the forestomach (NTP 2004a–g; Yoshizawa, Walker et al. 2005).

The incidence of uterine carcinoma was increased in the PeCDF and PCB118 studies. Uterine tumors occur in mice treated neonatally with endocrine-disrupting chemicals, such as diethylstilboesterol (DES) and genistein, a phytoestrogen in soybeans (Newbold et al. 2001). A potent nonsteroidal estrogen, DES is able to interact with the estrogen receptor and mimic the broad spectrum of estradiol actions in the reproductive system of female rodents (Marselos and Tomatis 1993). Exposure to DES has been documented to be associated with abnormal differentiation and organization of the uterus (Couse and Korach 2004). Estrogen induces uterine carcinoma in mice, rabbits, and menopausal and postmenopausal women (Drill 1979). Exposure to estrogens in the fetal or newborn period can also induce uterine tumors in mice and rats (Kitamura et al. 1999; Newbold, Bullock, and McLachlan 1990). Genistein, DES, and estrogen also induce a high incidence of uterine SM in rodents (Biegel et al. 1998; Couse and Korach 2004; Drill 1979; Newbold et al. 2001; Wordinger and Morrill 1985). Squamous cell carcinoma also appears subsequent to chronic exposure of mice to estrogen (Hart 1990). The incidence of spontaneous uterine cancers in rodents is relatively low, except for a few reported cases, such as in F344 rats (Nyska et al. 1994) and the high incidence of spontaneous endometrial adenocarcinoma reported in Donryu rats (Yoshida et al. 2002). The development in the Donryu strain is thought to be caused by an age-related imbalance of ovarian hormone, that is, an elevated estrogen:progesterone ratio.

There were no significant increases in the incidence of uterine carcinoma in the continuous-exposure groups in the study of PCB118. However, in that study, the highest incidence of uterine carcinoma was in the 4,600 μg PCB118/kg stop-exposure group. The increase in uterine carcinoma in the stop-exposure group was clearly significantly increased over controls. Uterine carcinomas have been seen in 6/473 control animals (1.3%) such that the incidence of 11/50 animals seen in the current study is considered to be a result of PCB118 exposure. However, the mechanism by which this lesion is induced by exposure, and why the increased incidence was seen in the stop-exposure group alone, remains uncertain. The increase in uterine carcinomas in the stop-exposure group may be related to endocrine changes during the study as a result of PCB118 exposure. The decrease in body weight gain seen in the higher-dose groups clearly led to a suppression of development of endocrine-related neoplasms in the mammary and pituitary glands (data not shown). The association between body weight and the incidence of mammary and pituitary tumors in rats has been reported (Haseman et al. 1997). Moreover, the cessation of exposure led to a recovery of normal body weight gain, which may have resulted in the reestablishment of “normal” endocrine stimulation of the mammary and pituitary glands. Evidence is provided by the observation that the incidence of mammary fibroadenoma in the 4,600 μg/kg stop-exposure group was significantly greater than in the 4,600 μg/kg core study group, but not significantly different from vehicle controls (data not shown). It is speculated that it is the exposure to high doses of PCB118 early in the study that led to the development of hormonally responsive uterine carcinoma and that the cessation of exposure reestablishes a hormonal milieu that promotes the development of these neoplasms that would otherwise have been suppressed had exposure been maintained.

Partial estrogen agonists such as tamoxifen can induce uterine carcinoma in the mouse (Newbold et al. 1997). In addition, PCB118 can be metabolized to hydroxylated metabolites, and it is known that some hydroxylated PCBs do have partial estrogenic activity in vivo (Martinez et al. 2005). It has been reported that 4-hydroxy-2,3,3N,4N,5-pentachlorobiphenyl can be formed from PCB118 and that in utero exposure to 4-hydroxy-2,3,3N,4N,5-pentachlorobiphenyl can have endocrine-disrupting effects, especially in female offspring (Meerts et al. 2004). Consequently, it is speculated that the induction of uterine carcinoma by PCB118 may be mediated in part by an estrogenic action of hydroxylated PCB118 metabolites. The metabolite profile of PCB118 was not measured in the present study.

In addition to uterine carcinoma, sporadic incidence of uterine SCC were seen with three animals affected in the 200 μg/kg group. None of the incidences was statistically different from controls. However, this neoplasm (SCC) is relatively rare and was seen in only 2/473 (0.4%) animals in the NTP historical controls database. Given the observations of other effects in the uterus, the known ability of DLCs to alter squamous differentiation in multiple tissues and an incidence higher than that in historical controls, it is concluded that this increase may be related to treatment.

The cause of decreased incidence of ovarian or uterine inflammation, SM, CEH, and/or SCC noted in some of the highest- and lowest-dosed groups exposed to TCDD and DLCs for two years is unknown, but the effects of decreased body weight in these groups cannot be excluded.

In conclusion, exposure of rats to TCDD; PCB126; PeCDF; and the TEF mixture of TCDD, PCB126, and PeCDF induced chronic active inflammation in the ovary (PCB153, binary mixture of PCB126 and PCB153), acute and/or chronic active inflammation of the uterus (PeCDF, PCB153), cystic endometrial hyperplasia (PeCDF), SM (PeCDF), SCC (TCDD, binary mixture of PCB126 and PCB153), and uterine carcinoma (PeCDF, PCB118). These effects were possibly via antiestrogenic mechanisms, endocrine disruption of the reproductive organs, or a local retinoid deficiency pathway resulting in abnormal circumstances or epithelial differentiation. Our data documenting effects on these tissues in the rat do not always correlate with those seen in the human—for example, endometriosis in particular—but the present studies did not result from in utero or postnatal exposure. Although we have begun to show such correlations, more extensive epidemiological investigations of potentially cancerous effects in the reproductive system must be performed to provide understanding of the extrapolations from rats to humans of dioxin-induced reproductive lesions.