| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Neuroendocrinology; |
The Institute for Neuroscience,3 Division of Pharmacology & Toxicology,4 College of Pharmacy, Section of Integrative Biology,5 and the Institute for Cell and Molecular Biology,6 The University of Texas at Austin, Austin, Texas 78712
Department of Biology,7 University of Wisconsin, Whitewater, Wisconsin 53190
ABSTRACT
Perinatal exposures to endocrine-disrupting chemicals, such as polychlorinated biphenyls (PCBs), can cause latent effects on reproductive function. Here, we tested whether PCBs administered during late pregnancy would compromise reproductive physiology in both the fetally exposed female offspring (F1 generation), as well as in their female offspring (F2 generation). Pregnant Sprague-Dawley rats were treated with the PCB mixture, Aroclor 1221 (A1221; 0, 0.1, 1, or 10 mg/kg), on Embryonic Days 16 and 18. Somatic and reproductive development of F1 and their F2 female offspring were monitored, including ages of eye opening, pubertal landmarks, and serum reproductive hormones. The results showed that low doses of A1221 given during this critical period of neuroendocrine development caused differential effects of A1221 on F1 and F2 female rats. In both generations, litter sex ratio was skewed toward females. In the F1 generation, additional effects were found, including a significant alteration of serum LH in the 1 mg/kg A1221 group. The F2 generation showed more profound alterations, particularly with respect to fluctuations in hormones and reproductive tract tissues across the estrous cycle. On proestrus, the day of the preovulatory GnRH/gonadotropin surge, F2 females whose mothers had been exposed perinatally to A1221 exhibited substantially suppressed LH and progesterone concentrations, and correspondingly smaller uterine and ovarian weights on estrus, compared with F2 descendants of control rats. These latter changes suggest a dysregulation of reproductive physiology. Thus, low levels of exposure to PCBs during late fetal development cause significant effects on the maturation and physiology of two generations of female offspring. These findings have implications for reproductive health and fertility of wildlife and humans.
Aroclor 1221, endocrine disruption, environment, estradiol, luteinizing hormone, PCBs, progesterone, reproduction, transgenerational effects
Polychlorinated biphenyls (PCBs) were used as nonflammable lubricants and insulators in industry in the United States beginning in 1929 until they were banned in 1977. Because of their lipophilic structures, PCBs are easily absorbed from the environment into the food chain, rendering human and animal exposure ubiquitous and persistent. As a result, a greater understanding of the many negative outcomes of exposure continues to be essential to human health and for addressing declining wildlife fecundity and viability.
Polychlorinated biphenyl exposure has been linked with a broad spectrum of effects, both in vivo and in vitro, which vary depending on method/age of exposure, sex of the individual, and dose/duration of exposure. Fetal and early developmental exposures to PCBs are particularly devastating, and can have different outcomes from adult exposure [1]. Latent effects of early exposures include, but are not limited to, depressed circulating thyroid hormone and abnormal thyroid cytology [2–6], delayed cognitive development [7, 8], altered sensory and motor abilities [9–11], reproductive impairment [12–15], and compromised neural function [16–19].
Not only do PCBs and other environmental endocrine-disrupting chemicals directly affect the exposed individual, but they can also exert effects on subsequent generations that may differ from those associated with primary exposure [20, 21]. One compelling mechanism for multigenerational effects of PCBs is via the hypothalamic-pituitary-gonadal reproductive axis. Exposure of the first generational animals can result in latent aberrant effects in adulthood on reproductive physiology and behavior, including improper steroid hormone production during pregnancy and deficiencies in parturition, lactation, and maternal behavior [22, 23]. Such alterations can potentially be transmitted to a second generation of offspring through improper hormonal exposure while in utero, and/or altered parental care postnatally [24, 25]. Thus, PCBs, particularly developmental exposures, can affect the immature organism, adult, and subsequent generations of offspring, although relatively little is known about the nature and mechanisms of these outcomes.
The current study investigated the effects of perinatal exposure to PCBs on sexual and somatic development of two generations, employing low doses of PCBs intended to approximate a range of exposures and body burdens in humans and wildlife [26–29]. Aroclor 1221 (A1221), a commercial PCB mixture composed of lightly chlorinated isomers, was chosen because of its previously reported disruption of the neuroendocrine system [30–36]. To our knowledge, the specific effects of fetal exposure of the F1 females on the reproductive physiology of their adult female F2 offspring has never been examined in this context, and our results support neuroendocrine effects of fetal endocrine disruption on the exposed generation as well as their descendants.
Sprague-Dawley rats were fed low-phytoestrogen Harlan-Teklad 2019 Global Diet ad libitum and were housed in a 12:12 partially reversed light cycle (lights on 2300 h, lights off 1100 h). All animal procedures were conducted in compliance with protocols approved by the Institutional Animal Care and Use Committee at the University of Texas at Austin. One day timed-pregnant Sprague-Dawley rats (F0 generation) were purchased from the Animal Resources Center at the University of Texas at Austin. Dams (n = 11–13 per treatment; Table 1) were injected intraperitoneally with A1221 (0.1, 1, or 10 mg/kg; Lot 072-202; C-221N-50MG; AccuStandard; all at 0.1-ml volume) or vehicle (0.1 ml dimethyl sulfoxide 99.5%; D4540; Lot 122K0027; Sigma) on Gestational Days 16 and 18, during the period of embryonic brain sexual differentiation [37]. Intraperitoneal injection has been used in other studies investigating the effects of PCB exposure [34, 38, 39], and here it was chosen to avoid stress of gavage or possible variability in consumption or absorption of infused food pellets. Our laboratory and others have previously shown significant effects of intraperitoneal A1221 on endpoints relevant to neuroendocrinology, brain, and behavior [31, 34–36], and we chose this route of treatment for continuity with previous reports. It should be noted that A1221 exposure to the F1 offspring occurs both during the late embryonic period and continues postnatally through lactation from mother to pups.
|
Polychlorinated biphenyls are only partially transferred from dam to offspring. To estimate actual exposure of each pup to the dose administered the dam, we followed the methods of Takagi et al. [40] and estimated that individual pups in each of the three dosed groups were exposed to approximately 0.2, 2, or 20 µg/kg PCBs, respectively, based on the prediction that amounts transferred to each pup are approximately 500-fold lower than the amount to which the dam is exposed [36, 40]. Because A1221 is a lightly chlorinated PCB mixture, it is believed to be relatively volatile [41], with a short half-life of several days [42], and its low degree of chlorination renders it difficult to detect [43]. We were thus unable to directly confirm body burden of experimental animals following dosage of the pregnant dam. Nevertheless, we estimate that our chosen doses approximate human and wildlife exposures. Average PCB serum levels of full-term babies in Germany have been measured at 0.5 µg/l [27], and from this figure adipose PCB levels can be estimated at 50 µg/kg [28]. Polychlorinated biphenyl body burdens in wildlife are also in the range of those predicted for the current study [29]. Notably, we were not trying to replicate identical body burdens of humans or wildlife exposures—both initial exposures as well as the timing and nature of metabolc inactivation of endocrine-disrupting chemicals vary by species. Thus, our goal was to approximate a range of low-dose exposures to PCB mixtures.
The day of parturition, determined by the birth of at least one pup before lights out at 1100 h, was termed Postnatal Day 0 (P0). On P1, live and dead pups were counted, sex ratio was determined, and litters were culled to four females (average litter size = 12; Table 1) to minimize effects of litter composition on body size, physiology, and behavior. This cohort is referred to as the F1 generation. At least one adult F1 female per litter was later allowed to become impregnated and produce the F2 generation. F2 litters were culled to six females when possible. Although treatment groups were known to the experimenter in the first cohort of F1 and F2 animals (14 litters) for the developmental measures of body weight and anogenital distance, experimenters were blind to treatment for these endpoints in subsequent cohorts (32 litters). No differences were found in any of these endpoints whether the investigator was blind or not, and data are shown combined. For all hormonal and postmortem measures, the investigators were blind to treatments for all 46 litters. In addition, each cohort had representation from each of the four groups to avoid any cohort effect, and at the time of killing there were always representative animals from each group.
Litter size, sex ratios, number of dead pups, anogenital distance (a marker of masculinization/defeminization [44]), and body weight were recorded for F1 and F2 pups on P1, prior to culling. After culling, the remaining females' body weights and anogenital distance were recorded two to three times per week until P30 or vaginal opening (VO). Anogenital distance was measured with digital calipers and was independently confirmed by a second investigator. Anogenital distance values were normalized to the cubed root of body weight for analysis [45]. The day of eye opening was recorded. On P22, pups were weaned to three to four littermates per cage. After P30, body weights were recorded once per week, and again on the day of killing. Timing of puberty was quantified as age at VO, first estrus (FE), and first diestrus (FD) [31]. For both the F1 and F2 generations, daily vaginal smears were conducted from VO until killing, excepting pregnant or lactating F1 rats.
Intact F1 females were mated on the closest age following P50 at which proestrus occurred. Mating tests took place on the afternoon of proestrus (1600–2100 h), when female rats in our colony are highly receptive, to generate a sperm-positive female for producing the second generation and in conjunction with another experiment that quantified the mating behaviors [36]. Some F1 females from the paced mating trial [36] were killed by decapitation approximately 16 h after mating, 1–2 h before lights out, between 0900 and 1000 h, to provide the samples assayed herein. This time of killing was chosen to avoid variability caused by differences in time of day. Sisters of those F1 females from the mating trials were allowed to carry a litter to term. Their F2 offspring were subject to the same measures and were killed at the same time of day as the F1 generation, with three exceptions: 1) F2 females were sexually naive, 2) they were killed on 
P42, and 3) because they were unmated, they were killed on random days of the estrous cycle, with vaginal cytology noted. For both generations, ovaries and uteri were dissected out, and wet weights were recorded. Trunk bloods were collected and were centrifuged at 6000 x g for 5 min to separate serum. Serum was stored at –80°C until hormone analysis.
Serum LH for each animal was assayed in duplicate samples (100 µl) in a single radioimmunoassay (RIA) with the specific rat LH RP-3 standard, iodinate, and antibody. The assay was performed in the laboratory of Dr. Michael J. Woller, University of Wisconsin–Whitewater, using previously published methods [46–49]. Reagents were obtained from the National Hormone and Pituitary Program of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), courtesy of Dr. A.F. Parlow. The sensitivity of this LH assay was 1 ng/ml (0.05 ng/tube) at 90% binding. Intraassay coefficient of variation (CV) was 2.36%.
Serum estradiol was measured in duplicate (200-µl) samples via the Diagnostic Systems Laboratories (Webster, TX) DSL-4800 ultrasensitive estradiol RIA kit. Five assays were performed according to the kit protocols. Assay sensitivity was 5 pg/ml. Intraassay CVs for each of the five estradiol RIAs were 3.53%, 9.54%, 4.33%, 7.63%, and 5.42%, and interassay CV was 9.43%.
Serum progesterone was measured in duplicate samples of 25 µl via the DSL-3900 progesterone RIA kit in accordance with the manufacturer's instructions. The assay sensitivity was 0.12 ng/ml. Samples were run in a total of eight assays, for which intraassay variabilities were 3.89%, 1.82%, 0.69%, 2.76%, 8.09%, 1.13%, 3.06%, and 1.52%, respectively, and interassay CV was 14.50%.
Except where noted (sex ratio), the F1 and F2 generations were analyzed independently due to methodological differences between the two experimental generations. Repeated-measures ANOVA was used to assess significant differences between treatment groups for serially recorded body weights and anogenital distances using Statview 5.0 software. For these endpoints, the unit of statistical analysis was the mean of each litter. The remaining endpoints (ovarian and uterine weights and circulating hormone levels), which exhibited considerably more variability, were analyzed using a linear mixed-model ANOVA approach with Proc Mixed in SAS [50, 51]. In these latter cases, each individual was included in analyses, but the "dam effect" was considered as a random covariate to control for individual litter effects. We fit models that included a fixed treatment effect, a random dam (treatment) effect, and a fixed categorical covariate representing estrous cycle stage of the appropriate category (estrus, diestrus 1, diestrus 2, or proestrus). In addition, our analyses tested for significant interactions between the treatment effect and the estrous cycle stage. If significant interactions were detected, we completed subsequent analyses, splitting the data by cycle stage to more fully explore the biological basis of the interaction. Models were estimated by restricted maximum likelihood, and significance was determined by a z score or F value for random and fixed terms, respectively. Several of the endpoints (LH, estradiol, uterine weight, body weight) were nonnormally distributed, and therefore we used a nonparametric permutation testing approach to determine significance. Here, phenotypes were randomized with respect to the experimental effects 1000 times, and analyses were then carried out as above [50, 51]. The test statistics from the randomized analyses were then used to determine the distribution of the test statistic under the null hypothesis and to empirically obtain an alpha level of 0.05. Where significant findings were observed, we employed a series of posthoc t-tests for all combinations of treatments, controlling for multiple tests with a Tukey-Kramer adjustment.
A1221 did not affect litter sizes or sex ratios in the F1 or F2 generation (Fig. 1). Because profiles of the two generations were similar for sex ratio, we analyzed the F1 and F2 generations together and found a significant effect of treatment (P < 0.05). Posthoc analysis revealed no specific significant interactions between treatment groups, although the control versus 0.1 mg/kg groups for the combined generations had a nonsignificant trend for a difference (P < 0.09).
|
F1 Body Weight and Anogenital Distance on P1
Body weight and anogenital distance were measured for P1 male and female pups, as shown in Table 2. For females, anogenital distance tended to be largest in the 0.1 mg/kg A1221 group, although this did not attain significance (P = 0.054 for 0.1 mg/kg vs. 1 mg/kg; P < 0.1 vs. 10 mg/kg; Table 2), and no group was significantly different from control. A1221 did not affect anogenital distance in male pups on P1, nor did it affect birth weight in pups of either sex (Table 2).
|
F1 Body Weight and Anogenital Distance During Development
After litters were culled, body weights and anogenital distance were recorded over time for each remaining experimental female (Fig. 2). A significant effect of treatment (P < 0.01) and a treatment x age interaction (P < 0.0001) were detected. Posthoc analysis revealed that the interaction was attributable to the P34 age, at which point the 1 mg/kg group was significantly heavier than all other groups (P < 0.005 for all).
|
Anogenital distance was also affected in the F1 females (Fig. 2). Although treatment alone did not cause any effects, age was associated with significant differences in anogenital distance (P < 0.0001). A nonsignificant trend for an interaction of age with treatment was also found (P = 0.056), with the 0.1 mg/kg group in general having the largest anogenital distance, and the 10 mg/kg A1221 group the smallest anogenital distance.
As shown in Table 3, there were no significant effects of A1221 treatment on the developmental markers of eye opening, VO, FE, and FD.
|
F1 Serum Hormone Levels and Estrous Cycles
Serum LH, progesterone, and estradiol levels were assayed. No significant effects of A1221 treatment were detected for serum progesterone and estradiol (Fig. 3, A and B, respectively). However, serum LH concentrations were significantly altered (P < 0.05; Fig. 3C). Posthoc analyses showed that the 1 mg/kg group was significantly different from the 0.1 and 10 mg/kg groups (P < 0.05) and tended to be different from the control group (p = 0.0516). Rat estrous cyclicity was monitored by daily vaginal smears beginning at VO, and statistical analysis showed that there were no differences in cycle length or percentage of days in each cycle stage.
|
The uterus and paired ovaries were carefully dissected and weighed after killing. There was no effect of A1221 treatment on F1 uterine or ovarian weights (Fig. 4).
|
F2 Litter Size and Composition on P1
Litter sizes of the F2 generation are shown in Figure 1. No significant differences were found. For sex ratio, no significant effects were observed for the F2 generation alone, although combining both generations revealed a treatment effect of P < 0.05 (see results for F1 litter size and composition on P1; Fig. 1).
F2 Body Weight and Anogenital Distance on P1
On P1, body weights in F2 generation females were smallest in the 1 mg/kg and largest in the 10 mg/kg A1221 group. These two groups were significantly different from each other (P < 0.05; Table 2), although they did not differ from control. Several significant differences were found for body weight in F2 male rats on P1, in which overall the 1 mg/kg group had the lowest body weights and the 10 mg/kg group had the highest body weights, similar to the females. The 10 mg/kg males had significantly larger body weights than both the 1 mg/kg and the control groups (P < 0.01 for both; Table 2). Anogenital distances were equivalent across all groups for both males and females (Table 2).
F2 Body Weight and Anogenital Distance During Development
F2 female body weights were significantly affected by age (P < 0.0001) but not A1221 treatment, although a significant interaction of age with treatment was detected (P < 0.02; Fig. 2C). Overall, body weights were smallest in the 1 mg/kg group and largest in the 10 mg/kg group. Posthoc analysis showed significantly larger body weights for the 10 mg/kg group versus the 1 mg/kg group on P1 and P3 (P < 0.05) and P5 (P < 0.01). On P40–P42, the 1 mg/kg group was significantly lighter than the control group or 0.1 mg/kg group (P < 0.05). F2 female anogenital distance was similar across all groups, as no main or interaction effects were detectable (Fig. 2D).
F2 Postnatal Maturational Markers
There were no significant differences between treatment groups for developmental markers of EO, VO, FE, or FD (Table 3).
F2 Serum Hormone Concentrations and Estrous Cycles
Serum progesterone, estradiol, and LH concentrations were assayed (Fig. 5). Because the F2 rats were randomly cycling at the time of killing, data are presented according to different days of the estrous cycle, across which serum hormone and gonadal/uterine weights exhibit natural fluctuations. There were no differences among F2 rats in estrous cycle length or percentage of days in any cycle stage, so only hormone data are reported in detail below.
|
Progesterone. Circulating progesterone concentrations were analyzed on each of four estrous cycle stages: proestrus, estrus, diestrus 1, and diestrus 2 (Fig. 5A). Although there was no main effect of treatment, there was a highly significant effect of estrous cycle stage (P < 0.0001) and a significant interaction of treatment with estrous cycle stage (P < 0.005). Analyses of interactions between cycle stage and treatment on proestrus showed that the control group was different from the 0.1 mg/kg group (P < 0.005) and from the 1 and 10 mg/kg A1221 groups (P < 0.05). On diestrus 2, significant differences were found for the 0.1 and 1 mg/kg groups versus the 10 mg/kg group (P < 0.05). The treatment x cycle stage interaction was further investigated to detect differences between circulating progesterone levels across the estrous cycle within each treatment group. All treatments showed significant effects of cycle stage (control and 1 mg/kg groups: P < 0.05; 0.1 and 10 mg/kg groups: P < 0.001). Posthoc analyses for the control group revealed significant differences between estrus versus diestrus 1 (P < 0.01) and between diestrus 1 and proestrus (P < 0.05). In comparison, only the 10 mg/kg group exhibited significant differences between estrus and diestrus 1 (P < 0.001), and only the 0.1 mg/kg group showed significant differences between diestrus 1 and proestrus (P < 0.01). In addition, all three treatment groups had significantly different progesterone levels on proestrus versus estrus (P < 0.01 for the 0.1 mg/kg group and P < 0.001 for the 1 and 10 mg/kg groups). The 10 mg/kg group also exhibited differences between estrus and diestrus 2 (P < 0.001).
Estradiol. Although a significant effect of cycle stage (P < 0.001) was detected by two-way ANOVA, neither effect of treatment nor any interaction of treatment with cycle stage was detected (Fig. 5B).
Luteinizing hormone. Luteinizing hormone levels were nonnormally distributed, probably due to pulsatile release and the presence/absence of the LH surge on different days of the estrous cycle. Indeed, LH levels varied significantly by estrous cycle stage (P = 0.01; Fig. 5C). A significant interaction of treatment with cycle stage was also found (P < 0.05). On proestrus, there were significant differences between control and all three treatment groups (control vs. 0.01 mg/kg and 10 mg/kg: P < 0.05; control vs. 1 mg/kg: P < 0.01). Because a significant treatment x estrous stage effect was observed, a further ANOVA was performed to determine possible differences between LH levels across the estrous cycle of each treatment group. Only the control and 10 mg/kg groups showed significantly different LH levels across the cycle (control: P < 0.05; 10 mg/kg: P < 0.01). Control animals had significantly higher LH levels on proestrus than on diestrus 1 or 2 (P < 0.05 for each). The 10 mg/kg group also had significant fluctuations across the estrous cycle (P < 0.01) and had significantly lower LH levels on proestrus than on estrus (P < 0.01).
F2 Ovarian and Uterine Weights
Ovarian and uterine wet weights were measured in randomly cycling F2 rats on P42, and results herein are divided by estrous cycle stage and are presented as a percentage of body weight (Fig. 6). For the ovary (Fig. 6A), two-way ANOVA showed a significant effect of treatment (P < 0.05) and cycle stage (P < 0.0001), but no interaction. Posthoc analyses of treatment showed a significant difference for control versus 0.1 mg/kg (P < 0.05), and a nonsignificant trend between control versus 1 mg/kg (P = 0.0514). With respect to estrous cycle stage, ovarian weight on estrus was significantly different from that at all other estrous cycle stages (P < 0.05 for all comparisons).
|
Analyses of uterine weight (normalized to body weight) showed no significant effect of treatment, a significant effect of estrous stage (P < 0.0001), but no interaction between treatment and estrous stage (Fig. 6B). Posthoc analyses of estrous stage showed that uterine weight was significantly larger on estrus than on all of the other days of the cycle (P < 0.05 for all comparisons).
The current study was designed to test the effects of exposures to PCBs on female development, reproductive physiology, and fertility. Although aspects of such work have previously been undertaken, our study brings together three important features. First, we used low-dose concentrations of PCBs [26, 29, 52] and employed a dose-response approach. The latter is particularly important, because hormonally active compounds can act via nontraditional mechanisms, resulting in U-shaped or inverted-U-shaped dose-response curves, and it is important to span a range of dosages to reveal these dose-response relationships [33, 53, 54]. Second, we examined effects of fetal exposure of the F1 individuals on adult endpoints. Understanding the fetal/developmental basis of adult phenotype is essential, because the developing organism is particularly vulnerable to exposures to endocrine disruptors during periods when organ and neural systems are being formed and organized [55]. Indeed, low-dose exposures to the fetus often have no discernable effects at birth, but are manifested much later in life [36, 56]. Third, our experiments extended analyses to the F2 offspring in order to assess possible transgenerational effects, which have been reported for several species [57–59], with other environmental toxicants [56, 60], and for the pharmaceutical, diethylstilbestrol (DES). The DES story has been particularly informative in this regard, as published data show second generational effects on tumor susceptibility [61–64], although this phenomenon has not, to our knowledge, been addressed specifically from the neuroendocrine perspective in mammalian species. Thus, our current results showing modest effects of fetal A1221 exposure on F1 females and more potent effects on the F2 generation are consistent with a body of evidence for multigenerational effects of endocrine-disrupting chemicals in other systems.
A1221 Had No Effect on Litter Size, and Modest Effects on Sex Ratio, in the F1 and F2 Generations
For the current study, we chose dosages of A1221 that mimic environmental exposures [26], anticipating that these low-level exposures would not have gross morphological effects. Rather, we sought to investigate more subtle and long-lasting outcomes of A1221 treatment. As predicted, A1221 had no significant effect on litter size in the fetally exposed F1 litters or their F2 offspring, indicating no overt toxicity of the treatments. Although there was no significant difference in sex ratio for the individual F1 and F2 generations, we noticed a nonsignificant trend for a female bias in the 0.1 and 1 mg/kg A1221 rats. We speculate that with greater statistical power (our current sample sizes were 11–13 litters per treatment), there may be a significant effect of treatment. Therefore, we combined F1 and F2 generations for statistical analysis, and overall there was a significant effect of PCB exposure, with a bias toward females. In rats, the sex ratio at birth is normally male biased (consistent with both F1 and F2 control litters in the current study and a previous report [65]). Polychlorinated biphenyls are associated with effects on sex determination in amphibians, reptiles, mollusks, and humans, an effect that may be attributable to altered fertilization, embryogenesis, implantation, or embryonic survival [66–70]. Studies on humans exposed to PCBs suggest that paternal or maternal exposure is related to altered birth sex ratios that may favor either males or females, depending upon the nature of the exposure [71, 72]. The estrogenic chemical DES has also been linked with altered sex ratios in mice [73]. Although there is a need to interpret these data cautiously due to relatively low statistical power, we hypothesize that male F1 embryos may have been more vulnerable to fetal exposure to A1221, resulting in loss of implantation or fetal death. A plausible explanation of changes in the F2 sex ratio includes epigenetic effects in germline stem cells of the F1 generation that decrease the likelihood of male F2 embryonic survival. We will investigate these possibilities in future studies, using larger sample sizes.
No Gross Morphological Effects of A1221 Were Observed in F1 and F2 Rats
Overall, we found few effects of the A1221 treatment on developmental parameters, indices of health, and anogenital distance. Body weight early in life is a useful index of maternal lactation or nursing behavior and/or the robustness of suckling by pups. Previous PCB studies have shown decreased birth weight and growth rate in exposed animals [74–76] and humans [77]. Later in life, endocrine-disrupting chemicals are associated with differences in body weight (reviewed in Newbold et al. [78]). Nevertheless, in our study, perinatal A1221 had no robust effects on body weight in either F1 or F2 rats at birth and across development, suggesting that pups were generally healthy and that the A1221 treatment did not affect this endpoint. Similarly, anogenital distance in the neonate, which is primarily determined by prenatal androgens [79, 80], was not profoundly affected by A1221 treatment in either generation of rats. This result is not surprising, as A1221 is considered to be an estrogenic mix of PCBs [81], and anogenital distance is largely an androgen-dependent event [80]. Finally, the developmental landmarks of eye opening, VO, FE, and FD were unaffected by PCB exposure in either generation. Again, we anticipated this result based on the relatively low doses employed in the current study and based on other published reports [31].
A1221 Significantly Alters Serum Hormones and Uterine/Ovarian Weight in F2 Adult Females but Has Few Effects in the F1 Generation
In the F1 generation, two of the three hormones measured (progesterone and estradiol) were unaffected by A1221, whereas serum LH concentrations were highest in the 1 mg/kg group. These data need to be interpreted cautiously: whereas the 1 mg/kg group differed significantly from the other two treatments, it did not differ significantly from the control group (P = 0.0516). Nevertheless, the observation that the effects of an intermediate PCB dose were different from other doses is similar to reports of inverted-U-shaped or U-shaped dose-response curves for endocrine-disrupting chemicals [53, 82, 83]. The current study did not find any effects on ovary or uterus weights by treatment in the F1 females, similar to a previous report [31].
In the F2 generation, serum hormone concentrations, specifically progesterone and LH, were profoundly altered by A1221, indicative of an uncoupling of normal hormonal profiles across the estrous cycle. On proestrus, the day of the preovulatory GnRH/gonadotropin surge, none of the three A1221 F2 treatment groups exhibited the increased progesterone and LH concentrations characteristic of a surge, contrary to our control rats, whose hormonal profiles are typical of female rats [84]. In fact, when considered across the estrous cycle, the hormone profiles exhibit a dampening and/or a shift in the timing and pattern of circulating hormone levels. Thus, it is possible that the preovulatory GnRH/LH surge is blocked or delayed in F2 females, similar to effects of pentobarbital, which causes 24-h delay in the preovulatory surge [84]. Moreover, low levels of LH on proestrus in the absence of altered estradiol concentrations suggest a difference in the positive feedback effects of estradiol on that day and are supportive of a hypothalamic and/or pituitary impairment.
The results on ovarian and uterine wet weights across the estrous cycle are consistent with our hormonal profiles. Ovarian and uterine weights fluctuated across the estrous cycle, peaking on estrus, as normally occurs following the preovulatory increase in estradiol [85]. These fluctuations were dampened in the PCB groups compared with the control groups, significantly so for ovarian weight. As a whole, these results on sex hormone levels and reproductive tract weights indicate an effect of A1221 on the F2 generation. Although we cannot state whether these same mechanisms were in play in the F1 generation, because rats were killed the day after mating, we do know that the F1 generation was fertile and capable of producing the F2 generation. However, we found in other experiments that the F1 PCB-treated rats required additional paced mating trials to mate [36], suggesting an uncoupling of reproductive physiology and behavior. Although the design of the current study did not anticipate uncoupled hormonal profiles in the F2 generation and used F1 rats the day after a mating trial (as opposed to across the estrous cycle), in future studies we will characterize the full reproductive ovulatory profiles of rats across multiple generations.
It is not surprising that the current second-generational effects of A1221 were observed. A1221 impairs fertility and embryogenesis in a mouse in vitro fertilization model [86] and induces genotoxic effects via intrachromosomal recombination [87] and deletions [88] in situ and in vivo. We postulate three possible explanations for the transgenerational effects we observed: 1) A1221 exposure directly introduced genotoxic effects in the germline of F1 females; 2) A1221 altered the methylation pattern of gene promoters in the F1 embryos, which was then passed to the F2; or 3) perinatal A1221 exposure altered circulating hormone levels during pregnancy in the F1, which exposed developing embryos to an improper hormonal environment. Genotoxic effects on the germline may explain the trend for decreased numbers of male offspring in the F2 generation, because male fetuses are more susceptible to detrimental genomic errors [89]. Thus, altered circulating hormones and gonadal weights in second generation-exposed females are most likely due to epigenetic patterning, either by gene promoter alteration transmitted to F2 offspring and/or an improper embryonic hormonal milieu.
We have demonstrated that perinatal exposure to low levels of A1221, a lightly chlorinated PCB mixture and environmental toxicant, has complex effects on physiology, fertility, and fecundity spanning two generations. The most salient results from this study include aberrant ovulatory processes in the F2 generation that may represent diminished reproductive success. Thus, PCBs, while banned for decades, should continue to be scrutinized for their persistent and latent risks to reproductive health in humans and wildlife.
ACKNOWLEDGMENTS
We are grateful to the PhRMA Foundation for its generous support, to the National Hormone and Pituitary Program of NIDDK, and to Dr. A.F. Parlow for contributions to the LH assay.
FOOTNOTES
1Supported by the PhRMA Foundation (Predoctoral Fellowship to R.M.S.) and the National Institute of Environmental Health Services (ES12272 and ES07784 to A.C.G.). 
Correspondence: 2Andrea C. Gore, Division of Pharmacology & Toxicology, A1915 University of Texas at Austin, Austin, TX 78712. FAX: 512 471 5002; e-mail: andrea.gore{at}mail.utexas.edu
Received: 16 December 2007.
First decision: 11 January 2008.
Accepted: 20 February 2008.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  E. Diamanti-Kandarakis, J.-P. Bourguignon, L. C. Giudice, R. Hauser, G. S. Prins, A. M. Soto, R. T. Zoeller, and A. C. Gore Endocrine-Disrupting Chemicals: An Endocrine Society Scientific Statement Endocr. Rev., June 1, 2009; 30(4): 293 - 342. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
