Developmental exposure to the DE-71 mixture of polybrominated diphenyl ether (PBDE) flame retardants induce a complex pattern of endocrine disrupting effects in rats

Chemicals

The test compound was a commercial mixture of PBDEs, DE-71 (pentabrominated diphenyl ethers (pentaBDE) from the Great Lakes Chemical Corporation of West Lafayette, IN, CAS No. 32534-81-9, lot 7550OK20A), a kind gift from Dr. Kevin Crofton at the U.S. Environmental Protection Agency. The manufacturer reports DE-71 to contain 50–62% pentaBDE, 24–38% TetraBDE, 4–12% hexaBDE, and 0–1% triBDE (amounting to a possible total of 154 congeners) (ENVIRON International Corporation, 2003). Corn oil (Sigma-Aldrich) was used as control compound and vehicle for all treatments.

Animals and treatment

We conducted two rat toxicity studies (Fig. 1). Study 1 (range-finding) aimed at determining the highest doses of DE-71 that could be given to pregnant and lactating rat dams without causing overt systemic toxicity in either dams or offspring. Study 2 (main study) aimed at assessing a series of endocrine sensitive endpoints to clarify the full potential endocrine effect pattern of DE-71 in a developmental toxicity study. We also assessed effects on the thyroid hormone system and on neurodevelopment (L. Ramhøj & M. Axelstad, in preparation). In both studies, time-mated nulliparous young adult Wistar rats (HanTac: WH, SPF, Taconic Europe, Ejby, Denmark) were received on gestation day (GD) 3 of pregnancy (day of plug-detection designated GD1), randomly distributed for pairwise housing and on GD4 pseudo-randomly divided into groups with similar weight distribution.

The expected day of delivery, GD23, was designated PD1 irrespective of the actual day of delivery.

In Study 1, dams were divided into four groups (n = 10 per group) and exposed to vehicle control or 20, 40 or 60 mg/kg body weight (bw)/day DE-71. Study 2 was performed in two balanced blocks with 3 groups (n = 22 per group, in order to have sufficient sample size for mammary gland examinations) exposed to vehicle, 40 or 60 mg/kg bw/day DE-71. Dosing of the dams was performed daily by gavage in the morning from GD7 to GD22, and again from postnatal day (PD)2 –PD14 in Study 1 (study termination) and PD2-16 in Study 2. Dosing was performed with a metal gastric tube with either vehicle (corn oil) or experimental solutions at a constant volume of 2 ml/kg/day, with the individual doses based on the body weight of the animal prior to dosing.

Dams were housed pairwise until GD17 and individually thereafter. We used semitransparent plastic cages (15  × 27  × 43 cm) with aspen wood hides and with aspen bedding (Tapvei, Gentofte, Denmark) placed on racks (balanced for treatment group) in an animal room with controlled environmental conditions: reversed light/dark cycles of 12 h (light from 9 pm-9 am, light intensity 500 lux), temperature 22 ± 1 °C, humidity 55 ± 5%, and ventilation changing air 10 times per hour. All animals were fed ad libitum on a standard diet, Altromin 1314 (soy- and alfalfa-free; Altromin GMbH, Lage, Germany), and acidified tap water was provided ad libitum in PSU bottles (84-ACBT0702SU Techniplast).

All endpoint examination procedures were carried out with treatment groups represented in a random order (i.e., not examining all controls first, then treated animals). Examinations and outcome measures of live animals were conducted blinded to exposure group while data analysis was conducted unblinded.

Animals were observed twice daily for health and signs of overt toxicity (such as hunched posture and raised fur). Criteria followed for euthanasia of animals: if it was considered irresponsible or unethical to let the animal live or if they were in a severe condition (expected to die within 24hrs). Accordingly, some litters and pups in Study 1 (see Table 1) were euthanized, as the dams did not take care of them and they were without milk in their stomachs. This phenomenon only took place in Study 1 and was unrelated to exposure, it also includes some dams that cannibalized their pups. In Study 2, one dam was euthanized on GD23 as she appeared unwell and incapable of giving birth (necropsy did not reveal any explanations for the dystocia).

Animal experiments were carried out at the DTU National Food Institute (Mørkhøj, Denmark) facilities. Ethical approval was obtained from the Danish Animal Experiments Inspectorate, with authorization number 2012-15-2934-00089 C4. The experiments were overseen by the National Food Institute’s in-house Animal Welfare Committee for animal care and use. All methods in the study were performed in accordance with relevant guidelines and regulations.

Birth and postnatal development

On the morning after overnight birth, dam and pup body weights were registered, and the pups were sexed and checked for macroscopic anomalies. Body weights of offspring were recorded on PD6, PD10 and PD14 in Study 1 and on PD6, PD14 and PD27 in Study 2.

Endocrine sensitive endpoints were investigated in Study 2. Anogenital distance (AGD), the distance between the anus and the genital papilla was measured in all live offspring on PD1, using a stereomicroscope with a micrometer eyepiece. The AGD index (AGDi) was calculated by dividing the AGD with the cube root of the body weight. On PD14 all offspring had their areolas/nipples counted. Nipple retention (NR) of male pups was defined as the number of areolas/nipples (a dark focal area with or without a nipple bud) visible where nipples are usually located (along the milk lines) in female pups (Schwartz et al., 2021). Both AGD and NR were assessed by the same technician, blinded to exposure group.

On PD27, one male pup from each litter was weaned and housed pairwise with another male from the same group (when no other males were available, housing was with a littermate).

Necropsy

Before necropsy was performed, all animals were weighed, anesthetized with CO₂/O₂ and killed by decapitation. Selected reproductive organs were examined in offspring killed on PD16, PD27 and PD300 in Study 2. On PD16 one male and one female from each litter were killed and ovaries and ventral prostate were excised and weighed. Mammary glands were collected as whole-mounts from one male and one female on PD27. On PD300, one male offspring from each litter was terminated and prostate together with seminal vesicle was excised and weighed. The ventral prostate was subsequently dissected and weighed alone.

All dams not giving birth (i.e., with resorbed embryos or never pregnant) were humanely killed on PD3, those with litters were humanely killed, along with remaining pups, on PD14 in Study 1 and on PD27 in Study 2. Implantation scars in uteri were counted to determine pregnancy rates and resorptions.

Whole-mounts

Mammary glands were excised and spread onto a glass slide, covered with parafilm and placed under pressure for 2 h. The whole-mounts were subsequently fixed in 10% neutral buffered formalin, stained with alum carmine, dehydrated in alcohol and cleared with xylene. The slides were scanned in a flatbed scanner at 4800 dpi and images of the 4th gland evaluated as described in Mandrup et al., (2015) for outer area (defined as the smallest polygon enclosing the gland), transverse growth (as defined by Mandrup et al. (2012)), longitudinal growth, distance to lymph node, distance to 5th gland (n = 13–17, except distance to the 5th gland where n = 11, 4 and 10 in control, PBDE-40 and PBDE-60, respectively) and number of terminal end buds (TEB) (TEB was defined as tear-drop shaped buds with a diameter >100 µm in zone C, as defined by (Russo & Russo, 1996) (n = 17, 11 and 14 in control, PBDE-40 and PBDE-60, respectively). Distance to lymph node was measured and scored (scores 1–3, with a score of 1 representing mammary tissue not reaching the lymph node, score 2 representing mammary tissue reaching the lymph node and score 3 was given to mammary glands where the tissue reaches beyond the lymph node). All data were assessed for males and females separately and blinded to exposure. In addition, for area measurements, longitudinal and transverse growth data from both sexes were pooled for analysis with increased statistical power. Images were analyzed with Image Pro Plus 7.0 (Media Cybernetics, Bethesda, MD, USA) and calibration performed for each picture before measurements were made.

Statistical analysis

The alpha level for statistical significance was always set to 5% and all samples were included in the analysis. Data with normal distribution and homogeneity of variance were analyzed for treatment-related effect differences relative to the control by analysis of variance (ANOVA) followed by Dunnet’s correction for multiple comparisons. Data were transformed if these conditions were not met and data not fulfilling the criteria were analyzed using a non-parametric statistical test (Kruskal–Wallis with Dunn’s multiple comparison test). When relevant, body weight was included as a covariate in the analysis (ANCOVA), e.g., for terminal organ weights and whole-mount measurements. Litter effects were accounted for by only analyzing one pup per litter, using litter means (offspring growth) or by including the litter as a random effect variable (AGD, NR and pooled whole mounts). Distance to lymph node was evaluated as the distance and as scores. The scores were analyzed using Kruskal–Wallis and using a two-sided Fishers exact test to compare the number of animals with glands reaching past the lymph node with the number not reaching past. The number of nipples/areolas (NR) was assumed to follow a binomial-distribution with a response range between 0 and 12 (12 assumed to reflect the biologically possible maximal number of nipples in rats). Litter effects on NR and over-dispersion in the data were accounted for by using Generalized Estimating Equations (GEE) as reported in (Christiansen et al., 2012).

SAS Enterprise guide 4.3 (2010) (SAS Institute Inc, Cary, NC, USA) and GraphPad Prism 5 (Graphpad Software, San Diego, CA; USA) was used for statistical analysis.

Effects on pregnancy, postnatal growth and general toxicity

We administered DE-71 at doses of 20, 40 and 60 mg/kg to the pregnant and lactating dams from GD7 to PD14/16. Pregnancy and litter data from Study 1 and Study 2 are listed in Tables 1 and 2. In Study 1, there were signs of decreased maternal weight gain at 40 and 60 mg/kg. Pups from the 60 mg/kg bw/day exposure group weighed 8% less than control pups at birth and gained 20% less during the first postnatal week. However, none of these findings were statistically significant. Thus, 60 mg/kg was chosen as the highest dose in Study 2 as it did not induce excessive systemic toxicity in dams and pups.

In Study 2, dam weight gain during pregnancy was significantly decreased at 60 mg/kg (Table 2). The high dose pups had decreased body weight on PD6 (Table 2) and body weight remained non-significant reduced throughout the postnatal period (Table 2) with non-significant reductions also on PD27 (∼7% reduction). Gestation lengths, litter size and pup mortality was not affected by exposure.

Effects on markers of early-life endocrine disruption

A shorter anogenital distance (AGD) and retained nipples in male offspring are considered sensitive markers of disrupted androgen action during development (Schwartz et al., 2019). DE-71 exposure induced a significantly shorter AGD and smaller AGDindex (AGDi, AGD adjusted for pup body size) at both 40 and 60 mg/kg. However, the dose–response curve was ‘flat’ in that reductions in AGD/AGDi were comparable between the two dose groups (10/11% reduced in the low dose group and 8/6% in the high dose group). The effect on AGD was also comparable between both sexes (Fig. 2A and Table 2). Nipple retention was increased in males from the high dose group (p = 0.0249). Notably, the mean number of 0.25 nipple in these high dose males is considered very small (Fig. 2B) and falls well within the range of our historic control data (Schwartz et al., 2021).

Early- and late-life effects on reproductive tissue weights

Weights of selected reproductive organs were assessed both early in life and in adulthood. DE-71 at a dose of 60 mg/kg reduced absolute and relative male PD16 prostate weights with 37% and 42%, respectively (Fig. 3 and Table 3). In female offspring, ovary weights were assessed on PD16 and showed no significant differences between controls and exposed animals.

On PD300, changes to prostate weights in exposed offspring were no longer apparent, nor were effects on seminal vesicle or ventral prostate weights (Table 3).

Effects on mammary gland development

We assessed mammary gland development by whole-mounts in both male and female offspring at PD27 and found that mammary gland outgrowth was stunted by perinatal exposure to DE-71. In male offspring, the transverse growth and mammary gland areas were statistically significantly reduced in the highest exposure group (Fig. 4). Females appeared to display a similar trend, albeit not statistically significant. When analyzing pooled data from both sexes (litter effects were accounted for by including the litter as a random nested factor in the statistical analysis) there was an effect on area in both dose groups and on the transverse growth in the high dose group (Fig. 4). The observed effects are likely not due to reduced body weight of the offspring in the high dose as the effects were observed both on absolute numbers and when accounting for the body weight of the animals in the statistical analysis.