Prenatal EDCs Impair Mate and Odor Preference and Activation of the VMN in Male and Female Rats

Abstract

Environmental endocrine-disrupting chemicals (EDCs) disrupt hormone-dependent biological processes. We examined how prenatal exposure to EDCs act in a sex-specific manner to disrupt social and olfactory behaviors in adulthood and underlying neurobiological mechanisms. Pregnant rat dams were injected daily from embryonic day 8 to 18 with 1 mg/kg Aroclor 1221 (A1221), 1 mg/kg vinclozolin, or the vehicle (6% DMSO in sesame oil). A1221 is a mixture of polychlorinated biphenyls (weakly estrogenic) while vinclozolin is a fungicide (anti-androgenic). Adult male offspring exposed to A1221 or vinclozolin, and females exposed to A1221, had impaired mate preference behavior when given a choice between 2 opposite-sex rats that differed by hormone status. A similar pattern of impairment was observed in an odor preference test for urine-soaked filter paper from the same rat groups. A habituation/dishabituation test revealed that all rats had normal odor discrimination ability. Because of the importance of the ventrolateral portion of the ventromedial nucleus (VMNvl) in mate choice, expression of the immediate early gene product Fos was measured, along with its co-expression in estrogen receptor alpha (ERα) cells. A1221 females with impaired mate and odor preference behavior also had increased neuronal activation in the VMNvl, although not specific to ERα-expressing neurons. Interestingly, males exposed to EDCs had normal Fos expression in this region, suggesting that other neurons and/or brain regions mediate these effects. The high conservation of hormonal, olfactory, and behavioral traits necessary for reproductive success means that EDC contamination and its ability to alter these traits has widespread effects on wildlife and humans.

Endocrine-disrupting chemicals (EDCs) are a class of compounds capable of interfering with the hormone systems of the body (1). Contamination by EDCs is detectable in nearly all ecosystems on the planet, and humans and wildlife are exposed to a variety of EDCs throughout the lifespan. Much research on EDCs has focused on early life because of implications for the expression of sex-typical adult behaviors that are dependent on sex-typical hormone levels during the organizational period of brain development (2). Experimental manipulation of hormone levels during this time can abolish sex-typical behavior or even elicit behaviors typical of the opposite sex in rodents (3-6). Among EDCs are the ubiquitous environmental contaminant polychlorinated biphenyls (PCBs). When pregnant rats were treated with low, ecologically relevant doses of Aroclor 1221 (A1221), a weakly estrogenic PCB mix (reviewed in (7)), the resulting female offspring showed disrupted sex behavior in a paced mating paradigm (8), and males showed impairment in tests of sociosexual preference and social novelty (9, 10).

It is well established that sexually motivated rats are able to distinguish between high- and low-quality potential mates on the basis of hormone status and will behaviorally demonstrate their preference (11-16). When given the choice, experimental rats will show greater interest in a sexually active opposite-sex partner with appropriate levels of circulating sex-steroid hormones than in one with low or no hormones (11, 17, 18). The physiology underlying this complex behavioral task is not fully understood, although olfaction is a highly salient sensory modality for this type of conspecific interaction (19). It is a process that involves both olfactory and hypothalamic brain structures that transduce the olfactory signal into a neurobehavioral response (16, 20-25). It is unknown how exposure to EDCs may affect this innate preference. This is a pressing question due to the critical role that sexual selection and fertility play in shaping populations and the evolution of species. In fact, prior work showed that female rats will have much lower preference for a male partner who descended from ancestors given EDC exposure compared to those from ancestral vehicle exposure (26). This latter study utilized a different class of EDC, the commercial fungicide vinclozolin, which is used in agriculture and has anti-androgenic properties (27, 28). Despite this evidence for transgenerational effects of vinclozolin, its direct effects on an exposed offspring’s mate choice has not been studied.

In this study, we determined how prenatal exposure to the 2 selected classes of EDCs with different mechanisms (A1221, estrogenic; vinclozolin, anti-androgenic) affected mate preference and olfactory discrimination. The time frame of prenatal exposure included the birth and organization of neurons in all noncortical, nonhippocampal brain regions as well as the beginning of brain sexual differentiation (29, 30). We also examined patterns of neuronal activation (Fos protein expression) in a brain region known to be required for the manifestation of mate preference in females, the ventrolateral component of the ventromedial nucleus (VMN) of the hypothalamus (16).

Materials and Methods

Experimental design

All animal procedures were conducted in compliance with protocols approved by IACUC at the University of Texas at Austin. Sprague-Dawley rats were purchased from Harlan and Envigo (Houston, TX), and housed in a colony room with controlled temperature (22 °C) and light cycle (12:12 light:dark, lights on at 24:00). Virgin females were mated with sexually experienced males, and the day after mating was designated embryonic day 1 (E1).

Pregnant dams were exposed to 1 of 3 treatments (n = 20, 23, and 20 for dimethyl sulfoxide [DMSO], A1221, and vinclozolin, respectively) via daily intraperitoneal injection from E8-E18 (total 11 injections). The vehicle (6% DMSO in sesame oil) was given alone or used to deliver 1 mg/kg A1221 or 1 mg/kg vinclozolin (VIN). The timing of treatment, dosages, and route of exposure were selected to match prior work, based on ecological relevance, and to span the period of fetal gonadal development and the early stages of brain sexual differentiation (31-33). The day of parturition was assigned as postnatal day 0 (P0). On P1, litters were culled to 5 males and 5 females. The pups were weighed weekly and weaned at P21 into same-sex groups of 2 or 3. Following vaginal opening, daily vaginal smears were taken, and cell cytology was examined to determine estrous cycle stage. Beginning at P60, a cohort of rats were tested in a mate preference paradigm. Females with a vaginal smear indicating proestrus or estrus were placed with a sexually experienced male at least 2 hours after lights-off and observed for receptive behaviors. Females only advanced to mate preference testing after displaying lordosis in response to a mount. Mate preference testing took place within 2 hours of receptivity testing to ensure that females were still in behavioral proestrus or estrus. Up to 4 males and females per litter were behaviorally characterized. The majority of litters had 2 to 3 male and female rats included in the study. The sample sizes (n) are based on individual rats rather than litter. The large number of litters (20, 23, and 20 for DMSO, A1221, and vinclozolin, respectively) ensured that littermates made a relatively small contribution to any endpoint, and we used non-littermates whenever possible. One month after mate preference testing, the rats were euthanized, and the brains and blood samples were collected for later assay. A second cohort of rats was tested in an odor preference and odor discrimination paradigm. As with mate preference, females were only tested if they displayed sexual receptivity that day. These rats were perfused 90 minutes after completion of the odor preference test, and their brains were collected for immunohistochemistry.

To produce stimulus rats for the mate testing paradigm, virgin males were castrated (gonadectomized; GDX) and females ovariectomized (OVX) under isoflurane anesthesia. During the surgery, rats assigned to the hormone-replaced stimulus groups also received a 1.5-cm silastic capsule containing 100% testosterone (GDX + T males) or a 1.0 cm capsule containing 5% estradiol/95% cholesterol (OVX + E2 females) implanted subcutaneously in the nape of the neck. On the day of testing hormone-replaced stimulus females also received a subcutaneous injection of 0.6 mg progesterone (P4) in sesame oil to simulate proestrus. The no-hormone groups were GDX or OVX but not given a Silastic capsule during the surgery. Fifty stimulus animals were generated for each of the 4 groups (GDX, GDX + T, OVX, OVX + E2 + P4).

Mate preference

A 1 m × 1 m 3-chambered apparatus (Stoelting) was used as the testing arena (9, 34). Testing was conducted under dim red light. Stimulus rats were confined in cylindrical cages placed in 2 opposite corners of the apparatus on removable plates. Stimulus rats were always of the opposite sex from the experimental (EDC or vehicle) rat; each trial gave experimental rats a choice between 1 rat with and 1 rat without hormone replacement. The bars of the cylindrical cages allowed for visual, olfactory, auditory, and minimal tactile interaction between the confined stimulus rat and the free-moving experimental rat. Trials began with the stimulus rats already in place in their cylindrical cages. The experimental rat was placed into the center of the apparatus with the doors to the side chambers closed for a 5-minute habituation period. The dividers were then removed and the experimental rat was allowed to freely explore the entire apparatus for 10 minutes. The rats’ movements were recorded by overhead video and tracked using ANY-maze software (Stoelting) (9, 10, 35). The entire apparatus was cleaned using 70% ethanol between each test subject. Stimulus rats were used for no more than 3 rounds per day and were given at least 1 day of rest between testing days. Some male experimental rats failed to investigate one or both of the stimulus options during the allotted 10 minutes and were excluded from the analyses.

Odor preference

Odor preference testing was conducted in the same apparatus in a nearly identical manner to mate preference. However, during odor preference testing, the cylindrical cages were empty, and the odor stimulus was the removable plate on which a caged stimulus rat had sat for 15 minutes previously. At least 3 stimulus rats of the same sex and hormone category (with or without hormone replacement) were allowed to deposit urine onto the plate before odor preference testing to reduce the chances of bias for or against an individual rat’s scent affecting the overall preference. Again, the experimental rat was restricted to the center chamber for a 5-minute habituation period, and then allowed to explore the entire apparatus for 10 minutes. Due to the failure of some male rats to explore the whole apparatus during the mate preference testing in the previous cohort, all rats were pre-habituated to a completely empty apparatus for 10 minutes twice during the week preceding the odor preference testing. All rats explored both stimulus options during odor preference testing.

Odor discrimination

Odor discrimination was tested using a habituation-dishabituation paradigm. For this test, multiple types of odor stimuli were created. 50 μL of each odorant was pipetted onto identical pieces of filter paper (2.5 cm × 5 cm). There were 2 nonsalient odorants for all experimental rats (3% acetic acid, sesame oil), and 2 socially salient odorants per sex (for females: urine from castrated males without hormone replacement, urine from castrated males with T hormone replacement; for males: urine from OVX females without hormone replacement, and urine from OVX females with E2 and P4 hormone replacement). The urine samples were pooled from 9 individuals per sex and hormone category and thoroughly homogenized. Each experimental rat performed 3 rounds of the habituation-dishabituation test. The first round used urine from opposite-sex stimulus rats. The second round used the neutral odors, acetic acid and sesame oil. The third round used urine from opposite-sex stimulus rats but presented in the opposite order from the first round. Whether the experimental rat received hormone or no-hormone urine first was randomized between subjects. The combinations are presented in Table 1.

Table 1.

Order of Testing in the Habituation-Dishabituation Paradigm

Experimental Subject	Trial 1	Trial 2	Trial 3	Trial 4	Trial 5 (Switch)
Females, males	acetic acid	acetic acid	acetic acid	acetic acid	sesame oil
Females	GDX	GDX	GDX	GDX	GDX+T
Females	GDX + T	GDX + T	GDX + T	GDX + T	GDX
Males	OVX	OVX	OVX	OVX	OVX + E2 + P4
Males	OVX + E2 + P4	OVX + E2 + P4	OVX + E2 + P4	OVX + E2 + P4	OVX

The assignment of odor stimuli by trial number during the habituation-dishabituation assay is shown. Both females and males were tested with acetic acid versus sesame oil once. Females were tested with male urine twice to allow for urine from both hormone categories to be presented first. Males were also tested with female urine twice to allow for urine from both hormone categories to be presented first. Whether the rat received hormone or no-hormone urine first was randomized between subjects. The acetic acid versus sesame oil test was always given between the 2 urine tests.

Abbreviations: E2, estradiol; GDX, gonadectomized; OVX, ovariectomized; P4, progesterone; T, testosterone.

The experimental rat was placed into a clean, polystyrene cage with bedding and a filter lid and allowed to habituate for 3 minutes. After habituation, the lid was gently raised and an odor stimulus (filter paper) was dropped into the cage for 1 minute. The time that the rat spent in direct investigation of the stimulus was manually timed using a stopwatch. Direct investigation included any time that the rat spent with its snout in contact with the filter paper. Sniffing, licking, and chewing were all included. Time that the rat spent standing on or holding the filter paper without snout contact was not included. After the allotted minute, the stimulus was removed, and a 30-second intertrial interval (ITI) began. After the ITI, a new filter paper with the same odorant was placed into the cage and the second 1-minute trial began. This process was repeated until 4 trials were completed. After the fourth ITI, the odor stimulus was changed to the other odorant in the pair (acetic acid vs sesame oil, GDX vs GDX + T, OVX vs OVX + E2 + P4). Extreme care was taken to ensure that no changes or cues signaled the switch trial to the rat other than the new odorant.

Brain sectioning and immunohistochemistry

Perfused brains were sectioned on a Leica vibrating microtome at a thickness of 40 μm into phosphate-buffered saline (PBS). A 1:4 series was taken for each area of interest. Single labeling for Fos protein was conducted by the diaminobenzidine (DAB) immunohistochemistry method, as published (36, 37). All steps unless otherwise indicated occurred on a rotator at room temperature and were followed by three 5-minute washes in PBS unless otherwise stated. Endogenous peroxidase activity was eliminated by incubation in 3:1 methanol and 3% hydrogen peroxide for 20 minutes. Sections were blocked in 5% normal goat serum (Vector Labs, Burlingame, CA) with 0.5% Triton-X (Sigma, St. Louis, MO) in PBS for 1 hour. Sections were incubated in the polyclonal primary antibody (rabbit anti-Fos, 1:1000, Synaptic Systems, Goettingen, Germany, Cat. # 226003 (38)) for 18 hours at 4 °C. Sections were then incubated in a biotinylated goat–anti-rabbit secondary antibody (Vector Labs, Burlingame, CA, Cat. BA-1000) for 2 hours. This was followed by incubation with the ABC Kit (Vector Labs, Burlingame, CA, Cat. PK-6100) for 1 hour and DAB reaction in water (Vector Labs, Burlingame, CA, Cat. SK-4100) for 3 minutes on ice. Sections were washed and mounted on slides. They were counterstained with methyl green (Vector Labs, Burlingame, CA), dehydrated in ascending concentrations of ethanol, and coverslipped with DPX (Sigma, St. Louis, MO). The Fos primary antibody has previously been validated by Western blot (39), and our negative control wells with the Fos primary antibody omitted did not show any DAB staining above background levels.

Double labeling of Fos and estrogen receptor alpha (ERα) were performed by immunofluorescence, as published (40). All steps occurred on a rotator at room temperature and were followed by three 5-minute washes in PBS unless otherwise stated. Selected sections were blocked in 5% normal goat serum (Vector Labs, Burlingame, CA) with 0.5% Triton-X (Sigma, St. Louis, MO) in PBS for 1 hour. For double-label immunofluorescence, sections were incubated in the same rabbit-anti-Fos (1:1000, Synaptic Systems, Goettingen, Germany, Cat. #226003) together with the polyclonal sheep anti-human ERα (1:500, R&D Systems, Minneapolis, MN, Cat. #AF5715 (41)) in 1% normal goat serum and 0.5% Triton-X in PBS for 48 hours at 4 °C. Sections were then incubated in the dark with 1:300 donkey–anti-sheep AlexaFluor 568 (ThermoFisher, Waltham, MA, Cat. A-21099) and 1:300 goat–anti-rabbit AlexaFluor 488 (ThermoFisher, Waltham, MA, Cat. A-11008). Fluorescent sections were washed and mounted on slides in the dark with Vectashield mounting medium containing DAPI (Vector Labs, Burlingame, CA, Cat. H-1200). Negative control wells omitting the ERα primary antibody did not show fluorescence above background levels. For both DAB and immunofluorescence immunohistochemistry, treatment groups were equally represented in all processing batches.

Imaging and image analysis

DAB labeling of Fos was analyzed by stereological counting (40) in the anterior and posterior piriform cortices, regions in the main olfactory pathway involved in transmitting nonpheromonal olfactory information to the hypothalamus, and the ventrolateral division of the ventromedial nucleus of the hypothalamus (VMNvl), preoptic area (POA), and 2 subregions of the medial amygdala (MeA), the latter regions linked to the regulation of behavioral output in response to odor. Counting frames with an inclusion edge and exclusion edge were placed at a random starting point within the region of interest and used to sample at least 200 cells in each section. Fos+ and Nissl stained cells were counted at 40× magnification on an Olympus BX51 bright field microscope. Sections spanned from Bregma −0.8 to −1.3 and −2.8 to −3.3 for the anterior and posterior piriform cortices, respectively. To find a consistent level for analysis of the VMNvl and MeA, the compact part of the dorsomedial hypothalamic nucleus (DMC) was used to identify the section corresponding to Bregma −3.14. That section and a section 160 μm anterior to it were counted for the VMNvl, posteroventral medial amygdala (MePV), and posterodorsal medial amygdala (MePD). The POA was counted from ~Bregma −0.26 to −0.8.

Immunofluorescent labeling of Fos and ERα were analyzed using a Zeiss 710 Laser Scanning Confocal microscope. Consistent levels of the VMNvl were selected as described above. The VMNvl region of interest was identified via dense ERα labeling and other anatomical landmarks. Sections were scanned at 20× magnification and the entire VMNvl was captured in 1 field of view. Scans were taken in a z-stack at 1 image every 1 μm for 10 μm. For counting, the stacks were collapsed into a projection image of maximum intensity. Brightness and contrast were adjusted to bring each image to a similar threshold. All cells positively labeled for Fos within the boundary of the ERα-labeled VMNvl were counted, as were all cells double-labeled for Fos and ERα. Imaging and counting here and elsewhere were performed by an experimenter blind to the treatment groups.

Statistical analysis

A 2-way ANOVA was used to determine the difference in time spent near the 2 stimulus rat options in the mate preference and odor preference tasks between the treatment groups. Multiple comparisons were performed between the hormone-replaced and non–hormone-replaced stimulus rats for each treatment. Adjusted P values using Sidak’s correction were used to determine differences. For the odor discrimination test, a repeated measures 2-way ANOVA was used to determine whether the interest of the rats in the odor stimuli varied over the trials (main effect of trial) and differed between treatment groups. A main effect of trial was present for all sexes and treatment groups in all tests. Next, a Dunnett’s test of multiple comparisons was used to ensure that Trial 4 (the fourth and final presentation of the initial odor stimulus) was significantly decreased compared to Trial 1, indicating successful habituation. Trial 5 (Switch) needed to be significantly increased compared to Trial 4 to indicate dishabituation. Again, all groups met both of these criteria for all tests. The percentage of Nissl counterstained cells that were Fos+ was averaged across sections for each rat in the anterior and posterior piriform cortex, POA, MeA subregions, and VMNvl. The number of fluorescent Fos+ and Fos+ ERα+ double-labeled cells were averaged between the 2 VMNvl sections counted for each rat. A 1-way ANOVA was used to determine differences between the treatment groups within the sexes.

Results

Mate preference

The mate preference test was performed on 81 females (25 DMSO, 29 PCB, 27 VIN) and 79 males (26 DMSO, 26 PCB, and 27 VIN). Twenty-five male rats distributed across treatment groups failed to investigate both of the stimulus rat options during the allotted 10 minutes and therefore were excluded from analysis as they could not form a preference with no knowledge of one or both options. This resulted in a final n = 19 DMSO, 15 PCB, and 20 VIN males.

The time experimental rats spent in proximity to the opposite-sex stimulus rats with or without hormone replacement is shown for females and males (Fig. 1). There was a significant difference in the amount of time female rats explored the male stimulus rats with and without T replacement (main effect of hormone treatment in stimuli: F_{(1, 161)} = 33.39, P < 0.0001). There was no main effect of EDC treatment (F_{(2, 161)} = 0.12; P = 0.889). Within EDC treatments, a Sidak’s post hoc test revealed that females exposed prenatally to DMSO or VIN showed a significant preference to associate with the GDX + T over the GDX males (DMSO: t(_{1, 161)} = 3.4; P = 0.0025; VIN: t(_{1, 161)} = 4.7; P < 0.0001) (Fig. 1A). Prenatal exposure to PCB, however, abolished the females’ preference for the hormone-replaced males (PCB: t_{(1, 161)} = 1.8; P = 0.188). There was no effect of treatment on the total time females spent in close proximity to the 2 stimulus rats.

Figure 1.

Time spent near the stimulus rats during 10 minutes of exploration. (A) Females exposed to DMSO and VIN spent more time near GDX + T males over GDX males, whereas PCB females did not show a preference. (B) Males exposed to DMSO showed a preference for OVX + E2 + P4 females over OVX females; this was abolished in PCB and VIN males. *P < 0.05; **P < 0.01; ****P < 0.0001; ns, not significant.

Males showed a significant difference in the amount of time they explored the female stimulus rats with and without hormone replacement (main effect of hormone: F_{(1, 107)} = 4.6; P = 0.034) (Fig. 1B). There was no main effect of treatment (F_{(2, 107)} = 0.6; P = 0.573). Males prenatally exposed to DMSO vehicle spent significantly more time investigating the OVX + E2 + P4 females than OVX females (t_{(1, 107)} = 2.4; P = 0.021). Prenatal exposure to either of the EDC treatments, however, abolished this preference in males; PCB and VIN males spent equal time associating with the 2 female stimulus options (PCB: t_{(1, 107)} = 0.68; P = 0.500; VIN: (t_{(1, 107)} = 0.734; P = 0.463).

Odor preference

The odor preference test was performed on a separate group of 40 females (14 DMSO, 14 PCB, 12 VIN) and 44 males (15 DMSO, 15 PCB, 14 VIN). All rats explored both stimulus options and were included in the analysis. Results for females are shown in Fig. 2A. There was a significant difference in the amount of time females spent exploring the odor of urine from stimulus males with and without hormone replacement (F_{(1, 79)} = 42.7; P < 0.0001). There was also a significant difference in the amount of time females of different treatments spent exploring the stimulus options (F_{(2, 79)} = 4.2; P = 0.019). Similar to what was observed during the mate preference experiment, DMSO and VIN females preferred the odor of GDX + T males over that of GDX males (DMSO: t(_{1, 79)} = 3.4; P = 0.003; VIN: t(_{1, 79)} = 6.2; P < 0.0001), whereas PCB females showed no preference (PCB: t(_{1, 79)} = 1.6; P = 0.372).

Figure 2.

Time spent near the odor stimuli during 10 minutes of exploration. (A) Females exposed to DMSO and VIN spent more time near urinary odors from GDX + T males over GDX males, whereas PCB females did not show a preference. (B) Males exposed to DMSO showed a preference for odors from OVX + E2 + P4 females over OVX females; this was abolished in PCB and VIN males. *P < 0.05; **P < 0.01; ****P < 0.0001; ns, not significant.

There was a significant difference in the amount of time males spent exploring the odor of urine from stimulus females with and without hormone replacement (F_{(1, 87)} = 8.8; P = 0.004; Fig. 2B). There was also a significant difference in the amount of time males of different treatments spent exploring the stimulus options (F_{(2, 87)} = 3.6; P = 0.033). DMSO males spent more time investigating the odor of OVX + E2 + P4 females than OVX females (DMSO: t(_{1, 87)} = 2.9; P = 0.015). As observed during the mate preference experiment, males exposed to either PCB or VIN failed to show a preference for the odor of OVX + E2 + P4 females; PCB and VIN males spent equal time near the two stimulus options (PCB: t(_{1, 87)} = 1.6; P = 0.3186; VIN: t(_{1, 87)} = 0.72; P = 0.855).

Odor discrimination in the habituation-dishabituation paradigm

In order to determine whether any deficits in mate and odor preference induced by EDC treatments were due to an impairment in the ability to discriminate odors more generally, we performed a habituation-dishabituation assay. The same 40 females (14 DMSO, 14 PCB, 12 VIN) and 44 males (15 DMSO, 15 PCB, 14 VIN) described above were tested earlier on the same day as the odor preference test. In the habituation-dishabituation task, experimental rats were habituated to 4 repeated presentations of the same odor stimulus, followed by the dishabituation test with the presentation of a novel odor stimulus. Decreased interest in the odor stimulus across consecutive presentations indicates habituation, while the sudden recovery of investigation time indicates dishabituation. When presented with non-socially salient odors (acetic acid, sesame oil), both females and males readily habituated and dishabituated, indicating successful discrimination between the odors (Fig. 3A, females: main effect of trial (F_{(4, 194)} = 67.5; P < 0.0001; Fig. 3B, males: main effect of trial (F_{(4, 219)} = 81.0; P < 0.0001). There were no effects of treatment (females: F_{(2, 194)} = 0.46; P = 0.637; males: F_{(2, 219)} = 0.21; P = 0.811). Similarly, when presented with the urine of opposite-sex stimulus rats, both males and females of demonstrated habituation during repeated presentations, and dishabituation when the odor was changed (Fig. 3C, females: main effect of trial (F_{(4, 194)} = 52.2; P < 0.0001; Fig. 3D, males: main effect of trial (F_{(4, 219)} = 29.8; P < 0.0001). Importantly, the rats were able to distinguish between the urine types regardless of the order of presentation (Fig. 3E, females: main effect of trial (F_{(4, 194)} = 41.6; P < 0.0001; Fig. 3F, males: main effect of trial (F_{(4, 219)} = 84.4; P < 0.0001). Again, there were no effects of treatment (females: Fig. 3C F_{(2, 194)} = 0.33; P = 0.720; Fig. 3E: F_{(2, 194)} = 1.3; P = 0.281; males: Fig. 3D F_{(2, 219)} = 0.02; P = 0.983; Fig. 3F: F_{(2, 219)} = 1.4; P = 0.257). These data indicate that prenatal exposure to EDCs does not impair odor discrimination ability for neutral or sexually relevant odors in males or females.

Figure 3.

Habituation-dishabituation results. The type of odor stimulus is indicated in the box below the y-axis for trials 1-4 (presentation of first odor) and trial 5 (switch). Females of all treatments were able to discriminate between neutral odors (A) and sexually relevant odors (C, E). Males of all treatments were able to discriminate between neutral odors (B) and sexually relevant odors (D, F). Asterisks refer to significant differences between trial 4 and 5 (the latter when the stimulus was switched) to evaluate dishabituation; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Abbreviations: GDX, castrated; GDX + T, castrated with testosterone replacement; OVX, ovariectomized; OVX + E2 + P4, ovariectomized with estradiol and progesterone replacement.

Stereological counting of Fos

Rats were perfused 90 minutes after performing the odor preference test to evaluate Fos immunoreactivity in the anterior and posterior piriform cortices, POA, medial amygdala (MePV, MePD), and the VMNvl. We first performed a modified stereological analysis to identify whether regions involved in the transduction of olfactory signals in assessment of potential mates were affected. There were no differences in the percentage of cells that were Fos+ in the anterior piriform cortex, posterior piriform cortex, POA, MePV, or MePD due to treatment in females or males (Table 2), nor were there any differences in the VMNvl of males (Fig. 4B). In females, however, prenatal exposure to PCB increased the percentage of cells that were Fos+ in the VMNvl after the odor preference test (Fig. 4A) (F_{(2, 18)} = 8.517; P < 0.01). This result led us to conduct a more in-depth follow-up in the VMNvl.

Table 2.

Fos+ Cells in Piriform Cortex, Medial Amygdala, and Preoptic Area

	DMSO			PCB			VIN
	Mean	SEM	N	Mean	SEM	N	Mean	SEM	N
Anterior Piriform Cortex
Females	21.17	3.63	6	26.33	4.13	9	23.78	4.31	9
Males	13.25	2.86	8	15.42	3.41	9	16.61	3.13	9
Posterior Piriform Cortex
Females	22.33	1.17	6	24.33	1.68	9	21.50	2.38	8
Males	14.56	3.19	9	16.70	2.81	10	14.64	2.65	11
Preoptic Area
Females	10.20	1.07	5	8.33	0.69	9	8.67	1.04	9
Males	1.84	0.49	8	4.28	1.72	9	2.03	0.84	9
Medial Amygdala PV
Females	10.10	1.22	6	10.34	2.34	7	10.53	1.77	8
Males	5.58	0.88	9	7.42	1.27	9	5.87	1.53	9
Medial Amygdala PD
Females	8.97	0.98	6	7.81	1.28	7	9.76	1.66	8
Males	4.13	0.67	9	3.72	0.70	9	2.92	0.73	9

The percentage of Fos+ cells in the anterior piriform cortex, posterior piriform cortex, preoptic area, posteroventral (PV) medial amygdala, and the posterodorsal (PD) medial amygdala. There were no treatment differences in any of the regions investigated with in either females or males.

Abbreviations: DMSO, dimethyl sulfoxide; PCB, polychlorinated biphenyl; VIN, vinclozolin.

Figure 4.

The percent of Fos+ cells in the VMNvl (A, B) of females and males perfused 90 minutes after the odor preference test. (A) Prenatal exposure to PCB increased the percent of Fos+ cells in the VMNvl of females. (B) Fos+ cells did not differ by treatment in the VMNvl of males. Bars with different letters are significantly different within a sex. For females, n = 5 DMSO, 8 PCB, 8 VIN; for males: 7 DMSO, 6 PCB, 7 VIN.

Co-expression of Fos and ERα in the VMNvl

Because ERα activity is required in the VMNvl for mate preference behavior in female rats (16), we next determined whether the change in neuron activation in the VMNvl in PCB-exposed females was specific to cells expressing ERα. The VMNvl was double-labeled for Fos and ERα. Confocal microscopy was used to count single-labeled Fos+ cells, and those double-labeled with Fos and ERα, in rats that were perfused 90 minutes after completion of the odor preference test. Representative micrographs are shown in Fig. 5. We confirmed our initial finding of increased Fos expression in the VMNvl from the DAB study (Fig. 4): in the immunofluorescence study, females rats exposed to PCBs had increased numbers of Fos+ cells in the VMNvl compared with VIN or DMSO females (F_{(2, 15)} = 4.8, P < 0.05; Fig. 6A). In males, similar to our DAB results (Fig. 4) there was no effect of treatment on the number of Fos+ cells in the VMNvl (F_{(2, 13)} = 0.20, P = 0.82; Fig. 6B). Unexpectedly, Fos+ cells in both males and females were largely not double-labeled with ERα (Fig. 6C-D), with a majority of rats having little to no double-labeling. There were no significant treatment effects on the number of cells double-labeled with ERα and Fos.

Figure 5.

Representative images of Fos and ERα immunofluorescence in the VMNvl of females euthanized 90 minutes after odor preference behavior. Scale bar = 25 µm.

Figure 6.

The number of Fos+ cells in the VMNvl of females (A) and males (B). Females exposed to PCBs had significantly more Fos+ cells than DMSO and VIN females. Males of all treatments had similar numbers of Fos+ cells. The number of ERα+/Fos+ double-labeled cells in the VMNvl of females (C) and males (D). Females and males of all 3 treatments had no significant differences in double-labeled cells in the VMNvl, and overall numbers were very low. Bars with different letters are significantly different within a sex. For females, n = 6 DMSO, 6 PCB, 6 VIN. For males, n = 6 DMSO, 5 PCB, 5 VIN.

Discussion

Prenatal EDCs change mate and odor preference in a sex-specific manner

This study investigated the effects of prenatal exposure to 2 known EDCs—PCBs and vinclozolin—on the behavior of adult rats in a sociosexual assay and whether and how changes to odor processing may underlie those effects. Our overall finding was that each EDC treatment had sex-specific effects on the disruption of mate preference behavior, and that changes to odor preference at least partially explained the observed deficits. Moreover, while activation of the VMNvl was associated with the behavioral changes in females (but not males), this was independent of the ERα cell population in this region.

Sexually motivated female mammals prefer to associate with males with whom matings are most likely to lead to successful reproduction. Males with higher testosterone concentrations (16), and odors of a male with higher testosterone levels, are preferred over low- or no-testosterone counterparts (13, 42). The current results are novel in that they show that prenatal exposure to PCBs abolished the females’ preference for a male with testosterone on board, as well urinary odors from these males, compared with castrated males or their urine. This could potentially compromise mating success under more naturalistic conditions. Interestingly, vinclozolin treatment had no effect in the females; as vinclozolin is best studied as an anti-androgenic EDC, and the PCB mix used (A1221) is weakly estrogenic, this difference may be due to different hormonal mechanisms acting upon the developing female brain.

For the males, we made the somewhat unexpected observation that the vehicle-treated virgin males in this study showed a clear preference to associate with the hormone-primed stimulus females over the females without hormone replacement. Though this preference is well-established in male rats with sex experience (14, 18, 43, 44), to our knowledge there are no reports of virgin males displaying this preference. There is some evidence that virgin males prefer the odor of estrous females to sexually active males, but overall, the results are mixed (44-46). Therefore, our vehicle-treated Sprague-Dawley males add new information to this previous literature. Vis-à-vis the EDC exposures in this study, prenatal exposure to either PCB or VIN abolished the preference for the hormone-primed females observed in the DMSO-treated males. The data from the odor preference experiment lend further corroboration: DMSO-treated males preferred to investigate the odor of hormone-primed females over the odor of those without hormone replacement. Exposure to PCB or VIN abolished this preference, similar to the mate preference experiment. It is interesting that males were sensitive to both PCBs and VIN; during development the male rat brain develops under the influence of relatively high concentrations of both androgens and estrogens, perhaps conferring greater sensitivity to disruption of these pathways by VIN and PCBs, respectively. Consistent with this, other studies have reported sexually dimorphic responses to VIN treatment. Perinatal exposure to VIN increased play behavior in juvenile males but not females (47). Most studies on VIN, however, have focused exclusively on males because of this EDC’s known anti-androgenic effects.

EDCs do not impair olfactory function

Because of the effects of EDCs on odor preference, we determined whether rats prenatally exposed to EDCs might have impairments in olfaction, something that to our knowledge has not been investigated. Our habituation-dishabituation results clearly indicated that males and females of all treatments had no loss of function in distinguishing different types of odors, whether they were sexually salient or not. Taken together with our findings in the mate preference and odor preference experiments, these results strongly suggest that changes to odor processing rather than a sensory deficit may contribute to the loss of preference in the EDC-treated rats.

This type of observation is not without precedent: female ferrets will preferentially approach male odors over female odors (like a sexually receptive female rat), but electrolytic lesions of the VMN abolish this behavioral preference. Similar to our results, female ferrets who lacked the expected preference had no deficits in distinguishing male and female urine in a habituation-dishabituation task (48). These findings mean that a different neurobiological process other than olfaction per se is likely to be deficient.

A plausible possibility is that prenatal EDCs cause changes in how rewarding the different stimulus types were to the experimental rats (49). More time spent near one option over the other indicates a higher value placed on the preferred option; equal time spent with both suggests the options are perceived as equally rewarding. Alternatively, the observed deficits could be explained by an overall decrease in sexual motivation due to treatment. If this is the case, an experimental rat who is uninterested in sex would theoretically perceive both stimulus options as equally attractive or unattractive social companions regardless of hormone status.

Research on effects of several types of EDCs on sociosexual behavior in females and males supports this hypothesis, namely declining sexual interest and the likelihood for a diminution of rewarding aspects of sex (50). Previous work on prenatal A1221 showed that paced mating behavior was disrupted in female rats (8), and male rats showed decreased sociosexual preference (10). Prenatal and perinatal exposures to other PCBs have caused reduced sexual motivation and sexual receptivity in females and altered sexual performance in males (51-55). An extended dietary exposure to VIN (E14 to adulthood) induced a lack of sexual interest and deficits in sexual performance (reduced erections and ejaculations) in male rabbits (56, 57). Neonatal exposure to VIN decreased androgen-dependent play behavior in male rats to levels characteristic of females (58). Treatment with testosterone during the same neonatal time period increased play behavior in females (59). However, several other studies have found no effects of VIN on sexual behavior in female rats (60, 61).

The ventrolateral portion of the VMN is activated by PCBs in females, but not males

The VMNvl is part of the neural circuitry involved in the control of sexual behavior, and an intact VMNvl is required for female sexual receptivity (62, 63). Single-unit recordings in the VMNvl of female mice show that this region is robustly activated in the presence of a sexually active male mouse (more so than in response to a female), and that this activation is estrous cycle–dependent: activation in response to a male, but not a female, is increased when recording from a sexually receptive female (64). This region is also involved in the processing of sexually relevant olfactory information. VMN neurons receive input from the accessory olfactory bulb through the medial amygdala and are activated by exposure to innate or conditioned (65) sexual cues in females and males (25, 66, 67). In female mice, the VMNvl responds differentially even to nuanced sexual cues, such as the urine of dominant versus subordinate males (25). Estrogen signaling is implicated in the functional organization of this region; female aromatase knockout mice lack the strong activation of the VMN that wild-type females have in response to urine from a male mouse (66). The most compelling evidence for a role of VMN ERα neurons in the results we observed comes from a study in which the knockdown of ERα expression specifically in the VMN abolished the innate preference of female rats for an intact versus castrated stimulus male. ERα knockdown in the medial amygdala or cortex had no effect on behavior. The effect was not tested in males (16).

Because of the importance of VMNvl in the transduction of olfactory information into a behavioral outcome related to mate choice, we assessed expression of the immediate early gene product Fos as a marker of neuronal activation, and its co-expression in ERα neurons, in rats euthanized 90 minutes after completion of the olfactory preference test. We note that only animals exposed to odorants were used in the Fos study, enabling us to make comparisons between vehicle, PCB, and VIN animals, but a limitation of the study was a lack of comparison to no-odorant controls due to the large additional number of rats that would have been needed to do that work. Here, the only group that showed a difference in Fos protein expression in response to odorant was the females prenatally exposed to PCB, in which the number of Fos+ neurons in the VMNvl was significantly increased compared to the female vehicle group. Neither VIN females, nor any of the EDC-exposed males, differed in Fos expression compared with their respective controls. Regarding the PCB females, we made the unexpected finding that the Fos+ cells activated by the task were largely not ERα-positive. In fact, only very small number of ERα+ cells in the VMNvl of any group co-expressed Fos 90 minutes after the odor preference test, with a range of 0 to 11 double-labeled cells and the majority of individuals having 3 or fewer double-labeled cells. The identity of the non-ERα cells that were activated by the behavioral task remains unknown and is a subject of future investigation. The VMNvl contains a population of GABAergic cells that are sensitive to circulating hormones levels and are associated with sex behavior in both males and females (68). It is possible that disruptions to inhibition within the VMNvl could be related to the deficits observed during the mate and odor preference tasks in female PCB rats in which Fos is upregulated. Considerable future work is needed to establish the circuitry and the nature of cells involved in transmitting olfactory information into a behavioral outcome in EDC-exposed rats; the VMNvl was selected as a logical first step.

Summary and implications

Successful reproduction in rats requires that females are able to accurately sense and assess the characteristics of potential mates; while males are less choosy, they are a key player in this dyadic interaction (69). If PCB exposure impairs the ability of females to show a preference for a more fit over a less fit partner, this could lead to impaired mating success, a decrease in fertility, and over generations, could even result in eventual population collapse (70). In fact, a study of a closed population of fish exposed to a pharmaceutical disruptor (ethinyl estradiol) was a real-world example of exactly such a collapse (71). Because hormonal control of fertility and olfactory communication between conspecifics are both evolutionarily ancient and well-conserved processes, it is likely that the phenotypes demonstrated in this study have broad implications to other mammals, and indeed, all vertebrates exposed to EDCs.

Glossary

Abbreviations

A1221

Aroclor 1221

DAB

diaminobenzidine

DMSO

dimethyl sulfoxide

E1

embryonic day 1

E2

estradiol

EDC

endocrine-disrupting chemical

ERα

estrogen receptor alpha

GDX

gonadectomized

ITI

intertrial interval

MeA

medial amygdala

MePD

posterodorsal medial amygdala

MePV

posteroventral medial amygdala

OVX

ovariectomized

P0

postnatal day 0

P4

progesterone

PCB

polychlorinated biphenyl

POA

preoptic area

T

testosterone

VIN

vinclozolin

VMN

ventromedial nucleus

VMNvl

ventrolateral portion of the ventromedial nucleus

Additional Information

Disclosure Summary:  The authors have nothing to disclose.

Data Availability:  The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.