Abstract
A recent study further investigated the potential effects of maternal thyroid function and morphology on fetal development upon maternal exposure to ammonium perchlorate during gestation and lactation. Female Sprague-Dawley rats (25/group) were given continual access to 0 (carrier), 0.01, 0.1, 1.0, and 30.0 mg/kg-day perchlorate in drinking water beginning 2 weeks prior to cohabitation through lactation day 10. Maternal, fetal, and pup serum thyroid hormone (thyroid-stimulating hormone [TSH], triiodo thyronine [T3], thyroxine [T4]) levels and thyroid histopathology were evaluated on gestation day 21, and lactation days 5, 10, and 22. No effects of exposure were observed on cesarean-sectioning, litter parameters, or fetal alterations. Reproductive parameters, including gestation length, number of implants, litter size, pup viability, and lactation indices, were comparable among all groups. Thyroid weights of dams sacrificed on gestation day 21, and lactation days 10 and 22 were significantly increased at 30.0 mg/kg-day. Increased thyroid weights were observed in male and female pups as early as postpartum days 5 and 10, respectively. Changes in maternal and neonatal thyroid histopathology were detectable at 1.0 mg/kg-day exposure. The maternal no-observable-effect level (NOEL) was 0.1 mg/kg-day (follicular cell hyperplasia was present at 1.0 and 30.0 mg/kg-day). The developmental NOEL was less than 0.01 mg/kg-day; thyroid weights of postpartum day 10 pups were increased at all exposures. Colloid depletion at 1.0 and 30.0 mg/kg-day exposures and changes of hormone levels at all exposures were considered an adaptive effect and appeared reversible in the rodent.
Considerable concern exists regarding the potential risks of thyroid dysfunction following exposure of sensitive populations to naturally occurring and man-made chemicals such as perchlorate. Perchlorate, a soluble anion commonly in the form of an ammonium salt, has been manufactured in the United States since the early 1940s for use in rocket and missile propellants, fireworks and explosives, matches, as well as fertilizers. Because of its reactivity in its solid state, ammonium perchlorate has a limited shelf life, thereby necessitating frequent disposal and replacement of industrial inventories. Unfortunately, disposal of large volumes of this chemical has lead to ground water contamination.
Perchlorate is an anion that is easily dissociated from salts such as ammonium perchlorate or sodium perchlorate. In 1997, perchlorate was detected in the drinking-water supplies in the southern regions of California and Nevada at levels of 4 to 16 μg/L (Lamm and Doemland 1999; Soldin, Braverman, and Lamm 2001). These are sites of production and storage of ammonium perchlorate, a strong oxidizer used by the Department of Defense in propellant systems such as those found in rockets and munitions. Ammonium perchlorate is readily soluble in water, and the long-term stability of the perchlorate ion has been demonstrated (Tsui, Mattie, and Narayanan 1998). Because of possible prolonged contamination in the drinking water supply, there is considerable potential for exposure in people working and living near facilities where the oxidant is manufactured and stored. One specific area of concern is the effect on processes controlling neurogenesis and synaptogenesis in developing fetuses of females exposed to perchlorate (Crump et al. 2000). To date, perchlorate releases have been reported and confirmed in at least 20 U.S. states (U.S. EPA 2002a), ranging from less than 4 ppb to more than 3700 ppm in some municipal water supplies(Clewell et al. 2003). Despite efforts by governmental agencies, states, water suppliers, and industry, the extent of perchlorate contamination has yet to be fully determined.
Normal functioning of the thyroid gland is part of a relatively simple feedback loop with the brain. Iodide (I−) is crucial to the normal functioning of the thyroid, and is an essential ion for normal production of the thyroid hormone thyroxine (T4). In turn, triiodothyronine (T3) is derived from T4 such that interference with T4 production subsequently reduces levels of T3. Production of T4 is stimulated when low levels of T3 in the brain signal increased release of thyroid-stimulating hormone (TSH). If the neural message for increased T4 synthesis is impeded, a condition of hypothyroidism, or goiter, results. This is the case when animals are exposed to the perchlorate anion. Mechanistically, perchlorate competitively inhibits the thyroid’s ability to concentrate iodine (Wolff 1998), an essential component for proper maintenance and function of the thyroid gland. The decrease in I− slows production of T4, thus stimulating the feedback cycle signaling for TSH. Decreased T4 is production over time, due to the inhibition of I−, results in hypothyroidism.
The thyroid gland plays a key role in regulating adult metabolism, as well as controlling growth and development in children. Although the maternal effects of perchlorate exposure tend to be transient and reversible, the effects of perchlorate on fetal and neonatal development are more profound and irreversible (Haddow et al. 1999; Morreale de Escobar, Obregon, and del Rey 2004). The effects of perchlorate exposure in children may cause sensory impairment (e.g., vision, speech and/or hearing) to thyroid gland tumors (Morreale de Escobar 2001). Additionally, congenital hypothyroidism is associated with varying degrees of mental retardation and occurs when both the maternal and fetal thyroid are incapable of supplying adequate thyroxine to the fetus (Boyages et al. 1989; Cao et al. 1994; Hollowell and Hannon 1997). Due to these effects, perchlorate is used to treat Graves’ disease, a condition of hyperthyroidism. However, uncontrolled exposure to perchlorate, such as through contaminated drinking water, is a health concern addressed in the present research.
Developmental neurotoxicity studies conducted in rodents (York 1998; York et al. 2004) have demonstrated (1) increases in the occurrence of thyroid follicular epithelium hypertrophy and hyperplasia; (2) decreases in follicular size in postpartum day 5 pups; and (3) increases in corpus callosum thickness in postpartum day 12 pups following maternal exposure to 10 mg/kg-day of ammonium perchlorate in drinking water. The developing fetus and neonate have been postulated to be the most sensitive subpopulations for human exposure to perchlorate as normal thyroid function is necessary for normal neurological development (Crump et al. 2000).
In response to recommendations made in 1998 regarding an external review of the Environmental Protection Agency (U.S. EPA) document entitled “Perchlorate Environmental Contamination: Toxicology Review and Risk Characterization” (http://www.epa.gov/ncea/perch.htm), the National Center for Environmental Assessment arranged a blinded peer review to be conducted on the thyroid glands from studies evaluating perchlorate exposure. To ensure consistency, this review of all sections of the thyroid, conducted by Pathology Working Group (PWG) (Experimental Pathology Laboratories Inc. 2000), employed grading criteria developed jointly by pathologists from the U.S. EPA, the National Toxicology Program and Experimental Pathology Laboratories (EPL), Inc. (Wolf 2000, 2001). Following reevaluation by the PWG, a consensus was reached where initial differences of opinion as to the presence or severity of a particular finding in the thyroid glands from perchlorate animal studies previously existed.
The purpose of this study, also called the “Effects Study” in the toxicology review of perchlorate by the U.S. EPA (U.S. EPA 2002a), was to further evaluate the effects of ammonium perchlorate in response to recommendations made during the 1999 external peer review. This study, with input from the U.S. EPA, was designed to evaluate and refine potential effects on the maternal and fetal thyroids, including histology and serum hormone assessments, following maternal exposure to ammonium perchlorate in drinking water during the gestation and lactation periods. The exposure levels selected were two of the same exposures as the previous work, plus one additional lower and one additional higher exposure levels (0.01, 0.1, 1.0, and 30.0 mg/kg-day).
Reproductive parameters, available from the cesarean-sectioning and littering portions of the study, were also assessed for comparative purposes. Evaluation of the neurodevelopmental consequences of ammonium perchlorate as it relates to the mammalian motor system is addressed in the companion paper that follows.
MATERIALS AND METHODS
Test Material
Ammonium perchlorate (CAS no. 7790-98-9), 99.8% purity, was obtained from Aldrich Chemical Company, Inc. (Milwaukee, Wisconsin; lot 03907LF). Test formulations of ammonium perchlorate in deionized water were prepared at least weekly and stored refrigerated; dosage solutions were allowed to reach room temperature before use. Test drinking solutions were adjusted to concentrations that yielded target doses of 0 (carrier), 0.01, 0.1, 1.0, and 30.0 mg/kg-day based on actual weekly water consumption. During the exposure period, rats were given continual access to either deionized water (carrier control group) or ammonium perchlorate in deionized water (test exposure groups) as drinking water. Dosing concentrations were monitored and confirmed on a regular basis using ion chromatography conducted at the Air Force Research Laboratory (AFRL).
Animals and Husbandry
Female Sprague Dawley [Crl:CD(SD)IGS BR VAF/Plus] rats were supplied by Charles River Laboratories, Inc. (Raleigh, North Carolina). Male rats of the same source and strain were used only as breeders. The female rats were approximately 72 days of age at arrival and weighed 187 to 249 g. Mated female rats (evidence of spermatozoa observed in a vaginal smear or a copulatory plug observed in situ) were considered to be day 0 of gestation (DG 0) and randomly assigned to one of five exposure groups, 23 rats per exposure group. The Sprague-Dawley rat was selected because it has been demonstrated to be sensitive to developmental toxins and is widely used throughout the industry for nonclinical studies of developmental toxicity, and had been used in the previous developmental neurotoxicology study (York et al. 2004).
Rats were provided ad libitum Certified Rodent Chow 5002 (PMI Nutritional International, St. Louis, Missouri) from individual feeders during the exposure period. Iodine in the feed from calcium iodate was 0.67 ppm (PMI Nutrition International).
All cage sizes and housing conditions were in compliance with the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources 1996). P generation rats were individually housed in stainless-steel wire-bottomed cages, except during the cohabitation and postpartum periods. Beginning no later than DG 20, P generation female rats were individually housed in nesting boxes with Bed-o’cobs (The Andersons Industrial Products Group [Maumee, Ohio]). Each dam and delivered litter was housed in a common nesting box during the postpartum period. The study rooms were maintained under conditions of positive airflow relative to a hallway and independently supplied with a minimum of 10 changes per hour of 100% fresh air that had been passed through 99.97% HEPA filters (Airo-Clean rooms). Room temperature and humidity targeted at 64°F (18°C) to 79°F (26°C) and 30% to 70%, respectively, were monitored constantly throughout the study. An automatically controlled fluorescent light cycle was maintained at 12-h light:12-h dark, with each dark period beginning at 1900 h EST.
Experimental Design
The in-life portion of this study was conducted at Argus Research Laboratories, Inc., in Horsham, Pennsylvania, USA. Because the purpose of this study was to better define effects observed in a previously conducted developmental neurotoxicology (DNT) study following in utero and postnatal exposure of rats to ammonium perchlorate, dams were given continual access to ammonium perchlorate in deionized drinking water or deionized water (carrier) beginning 2 weeks prior to cohabitation and continuing through location day (DL) 10. Rats were exposed for a longer period than recommended by guidelines (U.S. EPA 1996) to ensure a hypothyroid state. Target exposures were 0 (carrier), 0.01, 0.1, 1.0, and 30.0 mg/kg-day.
The animals in the study were divided into four subsets to facilitate study management and tissue collections. The thyroid and reproductive effects are reported in this article and the respective brain and behavior effects separately in the companion article that follows. For the first subset, dams were sacrificed on DG 21 and maternal blood, thyroid, and brain samples were collected. In the second subset, dams were sacrificed on DL 10, and maternal blood, thyroid, and brain samples were collected. Blood (pooled by litter), thyroid, and brain samples were collected from culled pups on postpartum day (PPD) 5 and from the remaining PPD 10 pups. For the third subset, dams were sacrificed on DL 22 and maternal blood and thyroid samples were collected. Blood (pooled by sex and litter), thyroid, and brain samples were collected from culled pups on PPD 5 and from the remaining PPD 22 pups. The last subset of the study was a standard rat developmental toxicity study that has been previously reported (York et al. 2003).
Reproductive Observations
The P generation female rats were observed at least twice daily for viability, and once daily during the exposure period for clinical signs of toxicity, abortions, premature deliveries, and deaths. Body weights and feed and water consumption were recorded daily throughout the exposure period. Feed and water consumption values were not tabulated after DL 14, when it was expected that pups would begin to consume the maternal feed and water.
On DG 21, P generation female rats were sacrificed, and maternal blood and tissue samples were collected as previously described. A gross necropsy of the thoracic, abdominal, and pelvic viscera was performed, and the number of corpora lutea in each ovary was recorded. The uterus of each rat was excised and examined for pregnancy status, number and distribution of implantations, early and late resorptions, and live and dead fetuses. Placentae were examined for abnormalities (i.e., size, color, or shape). Cesarean-delivered fetuses were individually identified, weighed, and examined for sex and gross external alterations.
Day 1 of lactation (DL 1) was defined as the day of birth and was also the first day on which all pups in a litter were individually weighed. Female rats were evaluated for adverse clinical signs observed during parturition, duration of gestation (day 0 of presumed gestation to the time the first pup is observed), litter size (defined as all pups delivered), and pup viability at birth. Maternal behavior was recorded on DLs 1, 5, 8, 11, 14, and 22. Pup clinical observations and viability were recorded daily, and pup weights were recorded PPDs 1, 5, 8, 10, 14, and 22. On PPD 5, were standardized to eight pups each (four males and four females, when possible).
Thyroid Histopathology
The thyroid gland from one fetus per sex per litter or one pup per sex per litter was removed and shipped to Experimental Pathology Laboratories (EPL; Sterling, VA) for histological evaluation. With the exception of the DG 21 fetuses and culled pups, the fixed thyroids were dissected from the attached tissue, weighed, processed by standard hematoxylin and eosin staining, and placed in paraffin blocks. In order to examine both thyroid lobes, the thyroids on the DG 21 fetuses and PPD 5 culled pups were left intact with the trachea, processed routinely, and each thyroid sectioned longitudinally. Thyroid glands from sacrificed pups on PPDs 5, 10, and 22 were weighed by EPL; thyroids could not be weighed for DG 21 fetuses. Colloid depletion, follicular cell hypertrophy, and follicular cell hyperplasia were each diagnosed separately as specified in the following text.
Colloid depletion was scored as grade 0 (no depletion in the amount of colloid), grade 1 (1% to 10% follicles affected), grade 2 (11% to 50% follicles affected), or grade 3 (51% to 100% follicles affected). The grade was assigned primarily on the basis of reduction in area of colloid in thyroid follicles, and secondarily on the tinctoral quality or staining characteristics of the colloid. Follicular cell hypertrophy was based on size of cell (cytoplasmic to nuclear ratio). For the grades used in this study, grade 0 was considered within normal limits (no hypertrophy). Grade 1 (minimal) consisted of follicular epithelium taller than the normal cubodial cell, approaching columnar; and individual cells exhibited increased height and width just above the limits of detection. Grade 2 (mild) consisted of follicular epithelium cells distinctly larger than normal and the follicular lumen was consistently obliberated by the hypertrophied cells. Follicular cell hyperplasia was only diagnosed when there was either stratification (multiple layers) or papillary infolding of single or multiple layers of follicular cells. Grade 1 (minimal) with 1% to 10% of follicles affected; grade 2 (mild) with 11% to 50% of follicles affected; and grade 3 (moderate to severe) with 51% to 100% of follicles affected.
Hormone Analyses
As previously described, blood samples from individual dams were collected from the inferior vena cava immediately after sacrifice on DG 21, DL 10, or DL 22 for evaluation of thyroid hormones (TSH, T3, and T4). On DG 21, blood samples were collected from the remaining fetuses not selected for evaluation of thyroid histopathology; fetal blood samples were collected via decapitation and pooled by litter due to the small volumes of blood available.
Pups not selected for continuation on study were sacrificed on PPD 5 by carbon dioxide asphyxiation; samples of blood (via cardiac puncture) and tissue were collected. The remaining pups were sacrificed in the same manner on PPD 10 for blood and tissue sample collection. Pups not selected for continuation in study were sacrificed in the same manner on PPD 5 for blood sample collection. Pups selected for continuation on study were sacrificed on PPD 22 for blood and tissue sample collection. Samples collected from all pups sacrificed on PPD 5 and PPD 10, including culled pups, were pooled by litter. Samples collected from pups sacrificed on PPD 22 were pooled by sex and by litter. The resulting serum was aliquotted into three vials for TSH, T3 and T4 determination, stored frozen (−70°C), and shipped (frozen on dry ice) to Air Force Research Laboratory (Wright Patterson Air Force Base, Dayton, Ohio) for analysis.
The radioimmunoassay (RIA) T3 assay kits using canine T3 antibody-coated tubes and RIA T4 assay kits using T4 antibody-coated tubes were purchased from Diagnostic Product Corp. (Los Angeles, California). RIA TSH assay kits were purchased from Amersham Corp. (Arlington Heights, Illinois) and TSH levels determined using lyophilized rabbit anti-rat TSH serum and Amerlex-M second antibody (donkey anti-rabbit serum coated magnetized polymer particles containing sodium azide). RIA kits used for all of the standard and unknown hormone measurements were of the same batch number and with the same expiration date. Tracer (125I) radioactivity was measured with a gamma counter (Packard Instruments Co., Meriden, Connecticut). All assays were performed in triplicate and according to the manufacturer’s recommended procedures.
Statistical Analyses
Clinical observations and other proportion data were analyzed using the variance test for homogeneity of the binomial distribution (Sokal and Rohlf 1969a). Continuous data (e.g., maternal body weights, body weight changes, feed consumption data, organ weights, duration of gestation, litter averages for pup body weights, percent male pups, pup viability, and cumulative survival) were analyzed using Bartlett’s test of homogeneity of variances (Sokal and Rohlf 1969a) and the analysis of variance (Snedecor and Cochran 1967), when appropriate (i.e., Bartlett’s test was not significant (p>.001) ]. If the analysis of variance was significant (p ≤ .05), Dunnett’s test (1955) was used to identify the statistical significance of the individual groups. If the analysis of variance was not appropriate (i.e., Bartlett’s test was significant [(p ≤ .001]), the Kruskal-Wallis test (Sokal and Rohlf 1969b) was used, when 75% or fewer ties were present; when more than 75% ties were present, Fisher’s Exact Test (Siegel 1956) was used. In cases where the Kruskal-Wallis test was statistically significant (p ≤ .05), Dunn’s method of multiple comparisons (1964) was used to identify the statistical significance of the individual groups. All other natural delivery data involving discrete data were evaluated using the Kruskal-Wallis test (Sokal and Rohlf 1969b) procedures previously described.
RESULTS
Consumed Dosages
The consumed dosages for the female rats for the precohabitation, gestation, and lactation exposure periods are shown in Table 1. The average actual consumed daily dosages (mg/kg-day) for the female rats for the precohabitation exposure period (study days 1 to 14) were 0.00, 0.01, 0.09, 0.77, and 23.80 mg/kg-day for the 0 (carrier), 0.01, 0.1, 1.0, and 30.0 mg/kg-day target dosage groups, respectively. The average actual consumed daily dosages for the female rats for the gestation exposure period (DGs 0 to 21) were 0.00, 0.01, 0.09, 0.97, and 28.34 mg/kg-day for the 0 (carrier), 0.01, 0.1, 1.0, and 30.0 mg/kg-day target dosage groups, respectively. The average actual consumed daily dosages for the female rats for the lactation exposure period (DLs 1 to 10) were 0.00, 0.01, 0.12, 1.26, and 38.32 mg/kg-day for the 0 (carrier), 0.01, 0.1, 1.0, and 30.0 mg/kg-day target dosage groups, respectively.
General Toxicity Measurements
There were no deaths, adverse clinical observations, or necropsy findings that were considered exposure-related during the precohabitation, gestational, and/or lactation periods. Average body weights and body weight changes during the precohabitation, gestation, and/or lactation periods were comparable among the five exposure groups and did not differ significantly (Figure 1). Average maternal absolute (g/day) and relative (g/kg/day) feed and water consumption values for female rats were also comparable among the five exposure groups during the precohabitation, gestation, and/or lactation periods (data not shown).
Reproductive Observations
Cesarean-sectioning observations were based on 16 pregnant dams on DG 21 in the five respective dosage groups. No cesarean-sectioning or litter parameters were affected by dosages of ammonium chloride up to 30.0 mg/kg-day (Table 2). There were no statistical differences for the average numbers of corpora lutea and implantations, or for the percent preimplantation loss and the averages for litter size, early or late resorptions, sex ratio, or fetal weights among the five exposure groups.
There were 16 pregnant dams that delivered and were selected for natural delivery observations in each of the 0 (carrier), 0.01, 0.1, 1.0, and 30.0 mg/kg-day target dosage groups, respectively. Natural delivery observations were unaffected by targeted dosages of ammonium perchlorate as high as 30.0 mg/kg-day (Tables 3a and 3b). There were no statistical differences for the average gestation length, implants, litter size, stillborn or pup viability, or lactation indices among the five exposure groups. Female pup weights were significantly increased in the 0.01, 1.0, and 30.0 mg/kg-day exposure group during the third week of lactation. These increases were considered unrelated to ammonium perchlorate exposure because fetal/pup weights were not significantly different across groups on DG 21, PPD 5, or PPD 10, and were most likely related to the decreased pup weights observed in the control group, which were low compared to the historical control values of the testing facility. All clinical and necropsy observations in the F1 generation pups were considered unrelated to the test substance.
Thyroid Weights
The absolute and relative (relative to terminal body weight) thyroid weights were significantly increased over the carrier group values in the 30.0 mg/kg-day dams sacrificed on DG 21, and DLs 10 and 22 (Tables 4 to 6). The thyroids were not weighed in the DG 21 fetuses; however, absolute thyroid weight of the PPD 5 male pups in the 30.0 mg/kg-day exposure group was significantly increased over the carrier group value (Table 7). There were no significant differences in absolute thyroid weights among the PPD 5 female pups. The absolute thyroid weights of the PPD 10 male pups in the 0.01, 0.1, 1.0, and 30.0 mg/kg-day exposure groups were significantly increased (p ≤ .05 or p ≤ .01) over the carrier group value (Table 8); absolute thyroid weight of the PPD 10 female pups was significantly increased over the carrier group value only in the 30.0 mg/kg-day exposure group (Table 8). The weight of the thyroid was significantly increased (p ≤ .05 or p ≤ .01) for the female pups on PPD 22 in the 1.0 and 30.0 mg/kg-day exposure groups ( Table 9b) and for male pups in only the 30.0 mg/kg-day dosage group (Table 9a). Although the absolute thyroid weight of PPD 22 male pups was significantly reduced (p ≤ .05) in the 1.0 mg/kg-day exposure group, this reduction was considered unrelated to treatment.
Thyroid Histopathology
Thyroid histopathology was evaluated using the scoring system developed for the Pathology Working Group (PWG) and was performed by one of the pathologists (KAF) who served on the PWG. A peer review of these slides was reported by Wolf (2000).
On DG 21, all dams in the 30.0 mg/kg-day exposure group had an increased incidence of decreased colloid in the thyroids (Table 4). A significantly increased (p ≤ .001) incidence of follicular cell hypertrophy was present in 30.0 mg/kg-day exposure group dams. An exposure-related increase in the incidence and severity of decreased colloid was present in male and female fetuses in the 1.0 and 30.0 mg/kg-day exposure groups (Table 10).
As shown in Table 7, an exposure-related increase (p ≤ .05 or p ≤ .01) in the incidence and severity of decreased colloid was observed in day 5 male and female pups in the 1.0 and 30.0 mg/kg-day exposure groups; all pups had this observation in the 30 mg/kg-day exposure group.
An exposure-related increase in the incidence (p ≤ .001) and severity of decreased colloid occurred in the 1.0 and 30.0 mg/kg-day exposure group dams (Table 5) sacrificed on DL 10; all dams in the 30.0 mg/kg-day exposure group had decreased colloid. An increased number of 30.0 mg/kg-day dams had follicular cell hyperplasia, and follicular cell hypertrophy occurred in all 30.0 mg/kg-day exposure group dams. Decreased colloid occurred in 100% of the PPD 10 male pups in the 30.0 mg/kg-day exposure group (Table 8). An exposure-related, significant increase (p ≤ .01 or p ≤ .001) in the incidence and severity of decreased colloid occurred in the PPD 10 female pups in the 1.0 and 30.0 mg/kg-day exposure groups; nearly all female pups in the 30.0 mg/kg-day exposure group had decreased colloid.
On DL 22, all dams in the 30.0 mg/kg-day exposure group (p ≤ .001) had decreased colloid (data not shown). An increased incidence of follicular cell hyperplasia was present in the 1.0 and 30.0 mg/kg-day dams, and an increased incidence (p ≤ .01) of follicular cell hypertrophy occurred in dams in the 30.0 mg/kg-day (Table 6). Male and female pups sacrificed on PPD 22 in the 30.0 mg/kg-day exposure group had an increased incidence (p ≤ .01 or p ≤ .001) of decreased colloid (Tables 9a and 9b, respectively).
The significantly increased (p ≤ .01) thyroid weights of dams sacrificed on DLs 10 and 22 in the 30.0 mg/kg-day exposure group correlated to the follicular cell hypertrophy observed at this exposure level (Tables 5 and 6, respectively). Follicular cell hyperplasia may also have contributed to the increased thyroid weights in the 30.0 mg/kg-day exposure group dams.
Thyroid Hormone Levels
Changes for all hormone levels occurred in an exposure-dependent manner, and were considered expected effects of ammonium perchlorate exposure. DG 21 maternal TSH levels (Table 4) were significantly increased (p ≤ .001) and T4 levels were significantly reduced (p ≤ .001) at exposure levels of 0.01 mg/kg-day and higher. At 30.0 mg/kg-day, the maternal T3 level was significantly reduced (p ≤ .01), compared to the carrier group (Table 4). Fetal TSH levels (Table 10) were significantly increased (p ≤ .01 or p ≤ .001) at exposures of 1.0 and 30.0 mg/kg-day, whereas significant reductions (p ≤ .01 or p ≤ .001) were observed for T4 and T3 levels (T4 at 30 mg/kg-day, and T3 at all perchlorate exposure levels).
Exposure-dependent, significantly increased (p ≤ .01 or p ≤ .001) DL 10 maternal TSH values occurred at all exposure levels (Table 5), compared to the carrier control group values. Maternal T4 levels were reduced with increasing exposure; significant reductions (p ≤ .05) occurred only at 30.0 mg/kg-day. A similar pattern was observed for maternal T3 levels; however, these changes were not statistically significant at any exposure level.
Evaluation of pooled serum TSH levels revealed exposure-dependent increases in PPD 5 pups. Conversely, pooled serum T4 levels decreased with increasing exposure, with significance (p ≤ .01) occurring in the 1.0 and 30.0 mg/kg-day exposure groups. A similar pattern occurred for pooled pup serum T3 levels, with statistically significant reductions (p ≤ .05) occurring only at the 30.0 mg/kg-day exposure level.
As shown in Table 11, pooled serum TSH levels for PPD 5 culled pups were significantly increased (p ≤ .01) in the 1.0 and 30.0 mg/kg-day groups, compared to the carrier control group values. Pooled serum T4 and T3 levels decreased with increasing exposure; statistically significant differences (p ≤ .001) were only observed for T4 levels. Serum TSH levels in PPD 10 pups (Table 8) were significantly increased (p ≤ .001) at only the 30.0 mg/kg-day exposure level. Although, the pooled serum T4 values decreased with increasing perchlorate exposure, these reductions were not statistically significant at any exposure level. Pooled serum T3 values for PPD 10 pups were significantly decreased (p ≤ .001) at the 1.0 and 30.0 mg/kg-day exposure levels. Hormone levels were similar between the PPD 5 and PPD 10 pups.
The 0.1, 1.0, and 30.0 mg/kg-day exposure levels caused exposure-dependent, significant increases (p ≤ .001) in DL 22 maternal serum TSH values (Table 6), compared to the carrier control group values. Conversely, serum T4 and T3 levels decreased with increasing exposure; however, these reductions were only significant (p ≤ .001) only at the 30.0 mg/kg-day exposure level for T4 hormone levels. Serum TSH was significantly increased (p ≤ .01 or p ≤ .001) in all PPD 22 male pups (Table 9a), compared with carrier control group values, whereas the same exposure levels significantly reduced (p ≤ .01 or p ≤ .001) serum T4 levels. Serum T3 levels for PPD 22 male pups were also significantly decreased (p ≤ .01) in the 1.0 and 30.0 mg/kg-day exposure groups, compared with carrier control group values. Serum TSH levels were increased in the PPD 22 female pups at all exposure levels, reaching statistical significance (p ≤ .001) in the 0.1, 1.0, and 30.0 mg/kg-day exposure levels. There was an exposure-dependent decrease in serum T4 levels, whereas serum T3 levels were significantly decreased (p ≤ .05) at 30.0 mg/kg-day.
DISCUSSION
Evaluation of the reproductive outcome of the study indicated that the perchlorate ion is not a selective developmental or reproductive toxicant in rats at exposures as high as 30 mg/kg-day. No effects of perchlorate exposure were observed on any cesarean-sectioning or litter parameters, and there were no fetal alterations observed at gross external examination. These findings are in agreement with the results of previously published rat and rabbit developmental toxicity studies (York et al. 2001b, 2003) using similar exposure parameters that concluded ammonium perchlorate should not be identified as a specific rat or rabbit developmental toxicant. Reproductive parameters of gestational length, number of implants, litter size, and pup viability and lactation indices for the rat were also comparable and did not differ significantly. These findings support the rat multigeneration study (York et al. 2001a) using similar exposure parameters that concluded ammonium perchlorate was not a reproductive toxicant in rats when administered in the drinking water at doses up to 30 mg/kg-day.
Research with pregnant female rats indicates increased thyroid size in both the dams and pups following maternal perchlorate exposure. As expected, enlargement of the thyroid gland directly correlated with reduced I− uptake due to increasing levels of perchlorate exposure in the dams, fetuses, and nursing pups (Brown-Grant 1966; Brown-Grant and Sherwood 1971). The most notable changes in thyroid parameters and presumably I−uptake in the fetuses were found in the final few days of gestation, the stage when the thyroid becomes active in the developing rat fetus (Sztanyik and Turai 1988). In the case of nursing pups, the evidence suggests perchlorate is not transferred through the dam’s milk; rather, the amount of I− in the milk is significantly reduced (Brown-Grant and Sherwood 1971; Zeghal et al. 1992). Perchlorate was measured in milk from lactating rats (Clewell et al. 2003). This low level of available I−has been correlated with decreased levels of both T3 and T4 in pups treated pre- and neonatally with perchlorate (Golstein et al. 1988), and with a higher concentration of T3 receptors found in the brain of hypothyroid pups on PPD 14 (Ishiguro et al. 1980). Such evidence suggests that any T3- and/or T4-dependent processes of normal brain development may be delayed or otherwise abnormal in the pups of dams treated with high levels of perchlorate. Human epidemiology and clinical studies have revealed that perchlorate doses that result in a 70% inhibition of iodine uptake has no apparent effect on human T4 levels (Strawson, Zhao, and Dourson 2004).
The absolute and relative (% body weight) thyroid weights for dams exposed to 30 mg/kg-day were significantly increased on DG 21 and DLs 10 and 22; thyroid weights were generally comparable for dams at these time points for exposures up to 1.0 mg/kg-day. On PPD 5, absolute and relative thyroid weights for male and female pups of dams exposed to 30 mg/kg-day were increased or significantly increased. By PPD 10, absolute and/or relative thyroid weights were significantly increased for male pups of dams at all exposure levels, whereas significance persisted only at the 30.0 mg/kg-day exposure level for female pups of the same age. Based on these findings, and the assumption that the pups would not have consumed the dams drinking water on PPDs 5 or 10, the effects noted in the thyroids presumably reflect secondary exposure during lactation. At weaning on PPD 22, absolute and relative thyroid weights for male and female pups were significantly increased in the 1.0 and 30.0 mg/kg-day exposure groups.
Perchlorate does not appear to exhibit cellular toxicity, as would be characterized by cell death or cell degeneration. The changes in thyroid histopathology observed, including primarily colloid depletion, hypertrophy and some hyperplasia, resulted from impaired thyroid function by perchlorate inhibition of iodide uptake.
Since iodide uptake is the key biochemical event that precedes all potential thyroid-mediated effects of perchlorate exposure, iodide uptake inhibition is not an adverse effect but a biochemical change (NRC, 2005). Based on these data, the maternal no-observable-effect level (NOEL) was 0.1 mg/kg-day (follicular cell hyperplasia was present at 1.0 and 30.0 mg/kg-day). The developmental NOEL was less than 0.01 mg/kg-day; thyroid weights of postpartum day 10 pups were increased at all exposures. Colloid depletion at 1.0 and 30.0 mg/kg-day exposures and changes of hormone levels at all exposures were considered an adaptive effect and appeared reversible in the rodent (USEPA 2002b).
There are currently no Federal or State drinking water standards for perchlorate. Several States have developed interim guidelines or action levels for perchlorate that range from 1 ug/L in Massachusetts to18 ug/L in Nevada and New York. The EPA reference dose and the State water guidelines all incorporate uncertainty factors that account for uncertainties inherent in the risk assessment process. In addition, the State water guidelines are adjusted downward to account for the proportion of perchlorate exposure that comes from drinking water, and are adjusted for the amount of drinking water consumed and body weight. As a result, these values are all much lower than any actual doses in either animal studies or human clinical or epidemiology studies and are lower than the lowest used in this study which was found to be a no-observable-adverse-effect level (NOAEL).
In February 2005, EPA established an official reference dose (RfD) of 0.0007 mg/kg-day of perchlorate. This RfD assumes total intake from both water and food sources and contains a full ten-fold uncertainty factor to protect the most sensitive population, the fetuses of pregnant women who might have hypothyroidism or iodide deficiency. This translates to a Drinking Water Equivalent Level (DWEL) of 24.5 ppb, which is almost 1000 times lower than the targeted levels of this study.
Footnotes
Figure and Tables
Acknowledgements
We gratefully acknowledge the services provided by Latha Narayanan, ManTech Environmental Technology, Inc. at AFRL/HE, for adult and pup T3, T4 and TSH hormone analyses. Funding for this research was by the Perchlorate Study Group, Michael Girard, Chairman.