The effects of early life adversity on growth, maturation, and steroid hormones in male and female rats

Abstract

Early life adversity is a risk factor for psychiatric disorders, yet the mechanisms by which adversity increases this risk are still being delineated. Here we used a limited bedding and nesting (LBN) manipulation in rats that models a low resource environment to examine effects on growth, developmental milestones, and endocrine endpoints. In LBN, dams and pups, from pups’ postnatal days 2–9, are exposed to a limited resource environment where dams lack proper materials to build a nest. This manipulation is compared to control housing conditions, where rat dams have access to ample nesting materials and enrichment throughout pups’ development. We found that the LBN condition altered maternal care, increasing pup-directed behaviors while reducing self-care. This, perhaps compensatory, increase in nursing and attention to pups did not mitigate against changes in metabolism, as LBN reduced weight gain in both sexes and this effect persisted into adulthood. Although adult stress hormone levels in both sexes and vaginal opening and estrous cycle length in females were not disrupted, there was other evidence of endocrine dysregulation. Compared to controls, LBN rats of both sexes had shorted anogenital distances, indicating reduced androgen exposure. LBN males also had higher plasma estradiol levels in adulthood. This combination of results suggests that LBN causes a demasculinizing effect in males that could contribute to lasting changes in the brain and behavior. Importantly, alterations in metabolic and endocrine systems due to early life adversity could be one mechanism by which stress early in life increases risk for later disease.

Introduction

The experience of early life adversity can increase risk for developing a range of psychiatric disorders, including anxiety, depression, schizophrenia, and substance abuse (Scheller-Gilkey et al., 2002; Scheller-Gilkey et al., 2004; Anda et al., 2006; Danese et al., 2009; Heim et al., 2010; Enoch, 2011; Heim & Binder, 2012). The mechanisms by which stressor exposure early in life increases disease risk are still being elucidated and much research focuses on how stress directly impacts the developing brain. Studies on stress effects on brain development report alterations to dendritic morphology (Brunson et al., 2005; Ivy et al., 2010; Wang et al., 2011a; Yang et al., 2014; Farrell et al., 2016) and neurogenesis and neural proliferation in rodent brains (Van den Hove et al., 2006; Naninck et al., 2015; Bath et al., 2016), as well as brain hippocampal-cortical connectivity (Silvers et al., 2016) and resting state cortical activity in humans (Demir-Lira et al., 2016). However, early life adversity also can affect endocrine systems, changing gonadal hormones and hypothalamic pituitary adrenal (HPA) axis responses to stress (Moffitt et al., 1992; Bhatnagar & Meaney, 1995; Mishra et al., 2009; Morgan & Bale, 2011; Murgatroyd & Spengler, 2011; Boynton-Jarrett et al., 2013; Dismukes et al., 2015; Ruttle et al., 2015). Alterations in gonadal hormones and the HPA axis have been linked to increased risk for depression, anxiety disorders, schizophrenia, and substance abuse (Payne, 2003; Altemus, 2006; Evans et al., 2012; Fernández-Guasti et al., 2012; Lenz et al., 2012; Quinones-Jenab & Jenab, 2012; Sugranyes et al., 2012; Sinha & Jastreboff, 2013; Misiak et al., 2018; Soria et al., 2018; Du & Hill, 2019; Gogos et al., 2019). Gonadal hormones can affect brain function (Conrad & Bimonte-Nelson, 2010; Wingenfeld & Wolf, 2015; Kinner et al., 2016). It is therefore critical to investigate how early life adversity leads to lasting alterations in endocrine systems in order to understand the etiology of and develop better interventions for psychiatric diseases.

In humans, early life adversity can take a wide variety of forms, but one very prevalent form of stress is growing up in a low socioeconomic status (SES) household, which affects millions of people worldwide (Bank, 2015). Low SES can be associated with limited access to resources (Bradley & Corwyn, 2002). To study how this type of adversity impacts development, the Baram laboratory developed a rodent model, the limited bedding/nesting model (LBN), which mimics a low resource environment (Ivy et al., 2008; Rice et al., 2008; Walker et al., 2017). In LBN, mouse or rat dams and neonatal pups are housed in an environment where access to nesting materials is restricted (Gilles et al., 1996; Brunson et al., 2005; Ivy et al., 2008). This model induces stress in the pups by altering maternal care (Rice et al., 2008; Baram et al., 2012; Molet et al., 2016). LBN has been shown to affect both the quality of maternal care (e.g., licking, grooming, nursing), as well as the pattern of care, such that it is often more fragmented (Ivy et al., 2008; Baram et al., 2012; Goodwill et al., 2019).

The LBN manipulation is typically used to study lasting effects on cognition, anxiety-related behavior, and the brain-gut axis (Brunson et al., 2005; Ivy et al., 2008; Moriceau et al., 2009; Wang et al., 2011b; Dalle Molle et al., 2012; Wang et al., 2012; Machado et al., 2013; Malter Cohen et al., 2013; Yang et al., 2014; Moloney et al., 2015; Naninck et al., 2015; Prusator & Greenwood-Van Meerveld, 2015; Moussaoui et al., 2016). Fewer studies have examined how this model affects gonadal hormones, although recent reports found that LBN in mice and LBN plus a Western diet in rats leads to earlier vaginal opening in females, a sign of precocious puberty (Manzano Nieves et al., 2019; Strzelewicz et al., 2019). More research has been done on how LBN affects HPA axis measures in pups immediately following cessation of the model. Typically, baseline levels of the glucocorticoid, corticosterone, are increased in pups immediately following LBN exposure (Gilles et al., 1996; Avishai-Eliner et al., 2001; Brunson et al., 2005) but this is not always reported (Moussaoui et al., 2017). One study measured basal corticosterone levels in weaning-aged rats (PND21) and found elevated levels in LBN animals compared to controls, with a greater elevation in females than in males (Moussaoui et al., 2017), but these baseline corticosterone seem to disappear in aged (12 month old) rats (Gilles et al., 1996). Another measure of HPA axis reactivity is a change in corticosterone release in response to a heterotypic stressor, and pups tested immediately after LBN exposure have decreased corticosterone responses to cold stress and immobilization stress (Gilles et al., 1996; McLaughlin et al., 2016). However, it is unknown whether changes in HPA axis reactivity persist into adulthood.

Collectively, these studies have focused on characterizing early developmental effects of LBN on the endocrine system. However, relatively few laboratories have investigated the lasting endocrine effects of this model in adult animals. Additionally, despite known sex differences in stress reactivity (Kudielka & Kirschbaum, 2005; Goel et al., 2014; Panagiotakopoulos & Neigh, 2014; Bangasser et al., 2019), few studies have directly compared the effects of LBN in male versus female offspring. The current experiments begin to address these gaps in the literature by investigating the effects of LBN on certain indices of reproductive development, as well as adult HPA axis reactivity and gonadal hormones in both male and female rats.

Methods

Animals

Adult male and female Long-Evans rat breeders were obtained from Charles River. All breeders were used only once. Rats were maintained in standard laboratory housing conditions (ample cob bedding, ad libitum access to rat chow and water, plastic hut for enrichment) except for rats undergoing the LBN manipulation as described in the following section. All experimentation took place under red light during the rats’ dark cycle unless otherwise noted. Pregnant rat dams were housed individually in the week leading up to parturition. All animals were kept on a 12-hour light/dark cycle (lights off at 11am) during breeding and up until weaning on PND21. Post-weaning, pups from both the LBN and control exposed conditions were moved to a separate colony room which was kept on a 12–12 light/dark cycle (lights off at 8:30am). It should be noted that this 2.5h shift in light cycle could be an additional peri-weaning stressor. However, relative to the stress of the abrupt weaning and rehousing that occurs at the timepoint (Cook, 1999), the impact of the light cycle shift would be expected to be relatively minimal. The day of parturition was designated postnatal day (PND) 0. For organ weights, gonadal hormones, corticosterone assay, and estrous cycle duration experiments, a maximum of 3 rats per sex per litter were used. For anogenital distance (AGD), body weight, vaginal openings, and eye opening measures, data were collected from all pups in a given litter, with the number of litters used for each measure as follows: AGD, 16 litters; PND10 body weight, 16 litters; PND18 body weight, 21 litters; post-weaning body weights, 8 litters; vaginal openings, 7 litters; eye openings, 7 litters. A total of 42 rat dams and 391 pups were used in all studies combined. All studies were conducted in accordance with the National Institutes of Health guidelines and were approved by Temple University's Institutional Animal Use and Care Committee.

LBN Manipulation

Pregnant dams were randomly assigned to either control or LBN conditions. On PND2, all litters were culled to 10 pups with an even representation of the sexes when possible. Control animals then remained in standard laboratory housing conditions and were given access to cotton nesting material. From PND2 through PND9, LBN animals were removed from the standard housing conditions and kept in the experimental housing conditions. LBN housing consisted of a custom stainless-steel metal platform (39.37cm × 18 cm) that separated the rat dam and pups from accessing bedding and a single paper towel to use as nesting material. Platforms included .95cm diameter round holes staggered by .635cm to allow for urine and feces to pass through and had four bolts attached to elevate the platform 1.27cm from the cage floor. Litters were left undisturbed until PND9, when LBN animals were moved back to standard laboratory housing. Pups from both control and LBN conditions were weaned on PND21 and pair housed with same-sex siblings.

Maternal Care Observations

We observed the quality of maternal care by evaluating behavior twice per day from PND3 to PND11: once during the last two hours of the light cycle from 9:00 am to 11:00 am, and once during the dark cycle from 1:00 pm to 4:00 pm (i.e., between 2–5 hours after lights off). For each observation, a trained experimenter observed the litters in the colony room for 1 hour. Every 5 min during that hour, a “snapshot” of maternal care behaviors was taken, with the behaviors being exhibited at that moment being recorded. Thus, each hour of observations consisted of 13 individual time points which were then summed to generate a single data point for each observed behavior. Behaviors recorded via this method of observation were categorized as either pup-directed behaviors or self-care behaviors. Pup-directed behaviors included arched back nursing, blanket/passive nursing, and licking/grooming pups. Self-care behaviors included self-grooming, eating/drinking, and resting outside of the nest. Each of these behaviors was analyzed individually for each day. Additionally, the sum of all pup-directed behaviors and the sum of all self-care behaviors were analyzed to determine the overall effect of LBN on these behavioral categories.

Developmental Measures

Pup body weights were recorded on PND10, 18, 22, 28-30, 36-38, 50-52, 58, and 65. Pups’ individual identities were not tracked until post-weaning. Therefore, body weight data from pre-weaning aged animals was analyzed as between subjects, while data from post-weaning aged animals was analyzed as within subjects.

Pups’ eyes were checked daily from PND14 to PND 16. Eye openings were checked during the dark cycle between 1pm and 4pm each day. PND 15 was used for analysis because that was the day that enough animals had their eyes open to do a meaningful comparison.

Reproductive and Gonadal Hormones Measures

In males and females, anogenital distance (AGD), defined as the distance in millimeters between the anus and the genitals, was measured between PND28–30. In females, the day of vaginal opening was recorded as an indication of the onset of puberty. Adult females were lavaged daily in order to evaluate estrus cycle phase as previously described (Bangasser & Shors, 2008). The average duration of estrous cycle (from estrus to estrus) was measured in LBN and control females in adulthood (after PND60). The duration of at least 3 full estrous cycles were recorded and averaged for each animal.

Plasma was extracted from trunk blood of adult male and female rats following rapid decapitation. Plasma 17b-Estradiol was measured using an MPBiomedicals 17b-Estradiol double antibody kit (0713810-CF) with an I-125 tracer. This kit detects plasma 17b-Estradiol in the 0-3,000 pg/ml range. This 17B-estradiol antibody cross-reacts 100% with Estradiol-17b, 20% with Estrone, 1.51% with Estriol, and .68% with 17a-Estradiol. The Intra-assay variation is 4.7 and the Inter-assay variation is 9.1 (% C.V.). Plasma Testosterone levels were measured using MPBiomedicals Testosterone double antibody kit (0718910-CF) with an I-125 tracer. This kit detects plasma testosterone in the 0-10ng/ml range. This testosterone antibody cross-reacts 100% with testosterone, 3.4% with 5a-Dihydrotestosterone, 2.2% 5a-Androstane-3b, 17b-diol, .95% 11-Oxotestosterone, .71% 5b-Androstane-3b, 17b-diol, .63%5b-Dihydrotestosterone, .56% Androstenedione, .2% Epiandrosterone. The Intra-assay variation is 9.1-11.1 and the Inter-assay variation is 10-12(%C.V.). Testosterone levels were not detectable in our female rats, therefore only plasma from male rats was analyzed for testosterone.

Organ weights and corticosterone ELISA

Adult LBN and control animals were sacrificed by rapid live decapitation for organ measures. Thymus, spleen, and adrenals were dissected and organ weight as a percentage of body weight was recorded.

Corticosterone response curves to acute restraint stress were generated. Adult rats were placed in restraint tubes for 1 hour. Blood samples were collected at baseline before restraint (Time 0), immediately after cessation of restraint (Time 60), and 30 minutes after cessation of restraint (Time 90). Blood was collected from the saphenous vein for Time 0 and Time 60 in an awake animal. At Time 90, blood was collected from the saphenous vein for a subset of animals and from trunk blood following rapid decapitation from another subset of animals. All blood samples for corticosterone analysis were collected between 8-11 hours into the animals’ dark cycle. Plasma corticosterone was measured using an Enzo corticosterone ELISA kit (ADI-900-97). The kit detects plasma corticosterone in the 32 - 20,000 pg/ml range. The corticosterone cross-reacts 100% Corticosterone, 28.6% Deoxycorticosterone, 1.7% Progesterone, .13% Testosterone, .28% tetrahydrocorticosterone, .18% Aldosterone, and .046% Cortisol. The intra-assay variation is 6.6-8.0 and the inter-assay variation is 7.8-13.1 (%C.V.)

Statistical Analysis

Maternal care behaviors, estrous cycle length, vaginal opening day, and plasma testosterone levels were analyzed with independent samples t-tests. Pre-weaning body weights, AGDs, estradiol levels, and organ weights were analyzed with 2×2 ANOVAs with sex and housing condition as factors with two levels. Post-weaning body weights and corticosterone levels were analyzed with mixed factors ANOVAs with sex and housing condition as between-subjects factors and time as a within-subjects factor. Significant interactions and simple main effects were followed up with LSD posthocs. A Chi-squared analysis was used to determine the proportion of rats with their eyes open from the LBN and control housing conditions on PND 15. The time to vaginal opening was also assessed, using the Log-rank (Mantel Cox) test, and plotted as a survival curve. Values that exceeded 2 SDs above or below their respective group mean for the dependent variables assessed were considered outliers and dropped. Results were considered statistically significant at p < 0.05.

Results

LBN increases pup-directed behaviors and decreases self-care behaviors in rat dams

To investigate the effects of the LBN manipulation on the quality of maternal care provided by rat dams, pup-directed behaviors (arched back nursing, blanket/passive nursing, and licking/grooming pups) and self-care behaviors (eating/drinking, resting outside of the nest, and self-grooming) were quantified. We score any aggressive behaviors towards pups, but because they were not observed except in 1 case where a control mom cannibalized her pups and these data were dropped from the study. LBN dams exhibited significantly more pup-directed behaviors [t(26)=−3.00, p=.006, d=1.12] and less self-care behaviors [t(26)=2.45, p=.021, d=0.95] overall than control dams during the dark cycle on PND5. LBN dams also exhibited significantly more pup-directed behaviors overall than control dams [t(22)=−2.30, p=.032, d=0.93] during the dark cycle on PND7. LBN dams exhibited more overall pup-directed behaviors [t(26)=−2.76, p=.010, d=1.02] than control dams on PND9 during the light cycle. Levene’s test for equality of variances was significant for overall self-care behaviors during the light cycle on PND9 [F(1,14)=13.76, p=.002] and Welch’s t-test revealed that LBN dams engaged in less self-care behaviors than control dams at this time point [t(12.69)=2.813, p=.015, d=1.16]. There was no effect of LBN on overall pup-directed behaviors or overall self-care behaviors at any other time point during the dark cycle nor during the light cycle. All dark cycle maternal care behavior data can be found in Figure 1a–j. A table of all maternal care measure outcomes taken during the light cycle is provided in Supporting Information.

Figure 1.

The effects of LBN on pup-directed behaviors and self-care behaviors during the dark cycle on PND 3, 5, 7, 9, and 11. (A-E) Stacked bars depict the frequency of pup-directed behaviors (arched back nursing, blanket/passive nursing, licking/grooming pups). LBN increased overall pup-directed behaviors on PND5 and PND7 and specifically increased arched back nursing and licking/grooming pups on PND5. (F-J) Stacked bars depict the frequency of self-care behaviors (self-grooming, resting outside the nest, eating/drinking). LBN decreased overall self-care behaviors on PND5 and specifically decreased time spent resting outside of the nest on PND7. Asterisks indicate p<.05 for an effect of LBN.

Individual pup-directed and self-care behaviors were also analyzed. Levene’s test for equality of variances was significant for arched back nursing [F(1,26)=16.65, p<.001] and resting outside of the nest[F(1,26)=4.86, p=.036] during the dark cycle on PND5, and resting outside of the nest during the light cycle on PND7 [F(1,14)=10.99, p=.005] and PND9 [F(1,14)=8.73, p=.010], so Welch’s t-test values are reported for those measures. LBN dams showed a trend toward increased arched back nursing during the light on PND3 [t(22)=−1.86, p=.076, d=0.80]. During the dark cycle on PND5, LBN dams exhibited significantly more arched back nursing [t(13.73)=−2.41, p=.030, d=0.94], more licking/grooming pups [t(26)=−2.22, p=.036, d=0.83], and less time resting outside of the nest [t(23.30)=2.05, p=.052, d=0.76] than control dams. During the dark cycle on PND7, LBN dams spent significantly more time resting outside of the nest than control dams [t(22)=2.95, p=.007, d=1.23] and showed a trend toward increased passive/blanket nursing [t(22)=−1.78, p=.089, d=0.48]. During the light cycle on PND7, LBN dams spent significantly more time licking/grooming pups than control dams [t(21)=−2.31, p=.031, d=0.95] and showed trends toward increased self-grooming [t(21)=−1.83, p=.082, d=0.76] and decreased time resting outside of the nest [t(12.86)=2.01, p=.066, d=0.82]. During the light cycle on PND9, LBN dams spent less time eating/drinking than control dams [t(26)=2.61, p=.015, d=0.99] and showed a trend toward decreased time resting outside the nest [t(11.613)=2.00, p=.069, d=0.84]. There was no significant effect of LBN on any other individual maternal care measures. For the dark cycle, PND3 control n=10, LBN n=14; PND5 control n=15, LBN n=13; PND7 control n=13, LBN n=11; PND9 control n=11, LBN n=16; PND11 control n=13, LBN n=9; for the light cycle, PND3 control n=10, LBN n=14; PND5 control n=16, LBN N=14; PND7 control n=12, LBN n=11; PND9 control n=11, LBN n=17; PND11 control n=11, LBN n=10.

LBN slows pup development in terms of body weight gain but not eye opening

In pre-weaning animals, on both PND10 [F(1, 266)=10.53, p=.001, ${\mathrm{\eta }}_{partial}^2=.038$]; (male control n=64; male LBN n=81; female control n=52; female LBN n =73)] and PND18 [F(1, 178)=17.20, p<.001, ${\mathrm{\eta }}_{partial}^2=.088$]; (male control n=42; male LBN n=51; female control n=38; female LBN n=51), LBN pups had significantly lower body weights than control animals (Fig. 2a). As expected, females weighed significantly less than males at both of these time points [PND10, F(1, 266)=5.35, p=.022, ${\mathrm{\eta }}_{partial}^2=.020$; PND18, F(1, 178)=5.66, p=.018, ${\mathrm{\eta }}_{partial}^2=.031$]. There was no significant LBN by sex interaction effect on PND10 body weights [F(1, 266)=1.18, p=.279] or PND18 body weights [F(1, 178)=1.404, p=.238].

Figure 2.

LBN decreased body weight gain. (A) On PND10 and PND18, LBN males and females weighed significantly less than sex-matched controls. (B) LBN males and females continue to weigh less and showed slower weight gain over time than sex-matched controls into adulthood. Asterisks indicate p<.05 for an effect of LBN.

We analyzed the effect of sex and housing condition on post-weaning weight gain. Mauchly’s test of sphericity was significant [χ²(20) = 226.14, p < .001], therefore degrees of freedom were corrected using Greenhouse-Geisser estimates of sphericity (ε = 0.31). All rats gained weight over time [F(1.89, 90.53)=2739.1, p<.001, ${\mathrm{\eta }}_{partial}^2=.983$]; (male control n=14; male LBN n=12; female control n=15; female LBN n=11). As expected, there was a sex by week interaction, such that females gained less weight over time than males [F(1.89, 90.53)=110.1 p<.001, ${\mathrm{\eta }}_{partial}^2=.696$]. There was also a housing condition by week interaction such that LBN exposed rats gained less weight over time [F(1.89, 90.53)=10.75, p<.001, ${\mathrm{\eta }}_{partial}^2=.183$]. There was no significant sex by condition by week interaction [F(1.89, 90.53)=1.56, p=.217]. There was also a sex by condition interaction [F(1,48)=4.80, p=.033, ${\mathrm{\eta }}_{partial}^2=.091$] and post-hoc tests revealed that LBN males weighed less than control males (p<.001), and LBN females weighed less than control females (p<.001) (Fig. 2b).

The proportion of rats with their eyes open on PND15 was assessed for a subset of litters and 19 out of 26 rats in the control housing condition and 24 out of 29 rats in the LBN condition had their eyes open. A chi-square test of independence was performed and revealed no significant difference in the proportion of rats with their eyes open between the two groups [χ² (1, N = 55) = 0.75, p =.385]. Data are not depicted.

LBN does not affect adult adrenal, thymus, or spleen weights

LBN did not affect adrenal weights as a percentage of body weights in adult rats [F(1, 37)=.63, p=.433]; (male control n=12; male LBN n=8; female control n=11; female LBN n=10). Adrenal weights accounted for a significantly larger percentage of body weights in adult females than in adult males [F(1, 37)=39.11, p<.001, ${\mathrm{\eta }}_{partial}^2=.514$] (Fig 3a). There was no significant LBN by sex interaction on percent adrenal weights [F(1, 37)=.12, p=.732]. Neither LBN [F(1, 35)=.10, p=.757] nor sex [F(1, 35)=2.63, p=.114] affected thymus weights as a percentage of body weights in adult rats and there was no significant LBN by sex interaction on thymus weights [F(1, 35)=1.93, p=.174] (male control n=10; male LBN n=8; female control n=11; female LBN n=10)]. (Fig. 3b). LBN did not affect spleen weights as a percentage of body weights in adult rats [F(1, 35)=.13, p=.719]; (male control n=10; male LBN n=8; female control n=11; female LBN n=10)]. Spleen weights accounted for a significantly larger percentage of body weights in adult females than in adult males [F(1, 35)=19.90, p<.001, ${\mathrm{\eta }}_{partial}^2=.362$]. There was no significant LBN by sex interaction on percent spleen weights [F(1, 35)=.16, p=.695] (Fig. 3c).

Figure 3.

The effects of LBN on organ weights in adulthood. LBN had no effect on the weights of adrenals (A), thymus (B), or spleen (C). The corticosterone response to 1-hour restraint stress was assessed (black bars indicate stressor duration), but LBN did not alter corticosterone levels in males (D) or females (E). Asterisks indicate p<.05 for a main effect of group.

The effect of LBN on stress-induced corticosterone release

To test whether LBN exposure altered the HPA axis, we assessed the corticosterone response to 1-hour restraint stress in adult rats exposed to LBN or the control nesting environment (male control n=8; male LBN n=7; female control n=8; female LBN n=8). There was a main effect of time [F(2, 54)=35.0, p<.001, ${\mathrm{\eta }}_{partial}^2=.564$] with corticosterone levels at 60 min being significantly higher than corticosterone levels at baseline (time 0 min, p<.001) and at 90 min (p=.005) (Fig. 3d–f). This finding indicates that we observed the expected corticosterone peak response to stress followed by recovery. There was also a main effect of sex [F(1, 27)=9.75, p=.004, ${\mathrm{\eta }}_{partial}^2=.265$], such that females had higher levels of corticosterone than males. However, there was neither an LBN by time interaction [F(2, 54)=.20, p=.821] nor an LBN by sex by time interaction [F(2, 54)=.76, p=.474]. These results suggest that exposure to LBN early in development does not reprogram the HPA axis response to acute stress.

LBN alters certain reproductive and gonadal hormone measures

We measured AGD between at a time point just before adolescence, between PND28–30. As expected, males had significantly longer AGD than females [F(1, 102)=269.25, p<.001, ${\mathrm{\eta }}_{partial}^2=.725$]; (male control n=25; male LBN n=32; female control n=22; female LBN n=28). PND28–30 body weight covaried with PND28–30 AGD [F(1, 102)=17.75, p<.001, ${\mathrm{\eta }}_{partial}^2=.148$] but there was still a significant effect of housing condition with LBN rats having smaller AGD than controls [F(1, 102)=8.13, p=.005, ${\mathrm{\eta }}_{partial}^2=.074$] (Fig. 4a). There was no significant LBN by sex interaction effect on AGD [F(1, 102)=1.39, p=.242].

Figure 4.

LBN altered measures of reproductive and gonadal hormones. (A) On PND28-30, as expected females had smaller AGDs than males, but LBN reduced AGD in both sexes. (B) LBN did not affect the timing of vaginal opening in adolescent females. (C) Survival curves show that the percentage of rats with vaginal opening over development did not differ between LBN and control rats. (D) LBN did not affect the average duration of the estrous cycle in adult females. (E) Adult females had higher levels of estradiol than adult males. LBN increased plasma estradiol levels and this effect was driven by males. (F) LBN had no effect on plasma testosterone levels in adult males. Asterisks indicate p<.05 for a main effect of group. ^ indicates a significant difference from the same sex controls.

The average day of vaginal opening in adolescent females did not differ between LBN (n= 22) and control (n =17) groups (Fig. 4b). Levene’s test for homogeneity of variance was significant [F(1,37) = 10.7, p=.002], so we ran a Welch’s t-test [t(23.1)=−.939, p=.357]. The time of vaginal opening was plotted with survival curves and evaluated using the Long-rank (Mantel-Cox) test but no differences were found [X²(1)=0.03, p=.874] (Fig. 4c). Rats from all treatment groups displayed vaginal openings by P35. The average duration of the estrous cycle in adult females also did not differ between LBN (n=12) and control (n=12) groups [t(22)=−.53, p=.602] (Fig. 4c).

As expected, plasma estradiol levels were significantly higher in adult females than in adult males [F(1, 31)=16.16, p<.001, ${\mathrm{\eta }}_{partial}^2=.343$] (Fig. 4e). Rats exposed to LBN had significantly higher plasma estradiol levels than controls [F(1, 31)=7.78, p=.009, ${\mathrm{\eta }}_{partial}^2=.201$]; (male control n=9; male LBN n=9; female control n=8; female LBN n=9)]. There was no LBN by sex interaction on plasma estradiol levels [F(1, 31)=.33, p=.569]. However, cycle stage data was not collected on all female samples, so it could be possible that the effect of LBN in females was attributable to a greater number of LBN females being cycle phases with higher estradiol levels. Therefore, we conducted a separate independent samples t-tests on estradiol levels for females and males. There was no significant difference in plasma estradiol between LBN and control females [t(15)=−1.451, p=.167]. However, LBN increased plasma estradiol levels in males [t(16)=−2.58, p=.020, d=1.21]. These results reveal that LBN specifically affects estradiol levels of males and that the main effect of LBN on estradiol levels is not likely due to potential cycle stage differences between the LBN and control group. Plasma testosterone levels in adult males were not significantly different between LBN (n=8) and control groups (n=9) [t(15)=−.20, p=.846] (Fig 4f).

Discussion

These studies examined the effect of the LBN model on maternal care, pup development, and endocrine changes in adulthood. A feature of the design was the comparison of male and female offspring, which is different than most studies using this model that either focus on one sex or do not disambiguate data between the sexes (e.g., (Avishai-Eliner et al., 2001; McLaughlin et al., 2016; Walker et al., 2017; Manzano Nieves et al., 2019). We found that dams in the LBN condition spent less time engaged in self-care, but more time engaged in pup-directed care, including an increase in nursing. However, this, perhaps compensatory, change in care to deal with the low resource environment did not prevent all adverse outcomes for pups. For example, male and female rats exposed to the LBN manipulation gained less weight than those in the standard housing condition. There was also evidence of endocrine disruption, as pups of both sexes exposed to LBN had smaller AGDs than controls when measured prior to puberty, indicative of a reduction in perinatal androgen exposure. We also wanted to assess the lasting effects of LBN on the HPA axis and other steroid hormones. There were no differences in the weight of the spleen, thymus, and adrenals in adult rats exposed to LBN or control housing. Consistent with no effect of LBN on adrenal weight, the corticosterone response to 1-hour restraint stress was comparable between LBN and control rats. LBN exposure did not alter the estrous cycle in females or testosterone levels in adult males. However, it did increase plasma estradiol in males. This result, considered together with the smaller observed AGD, suggests that LBN may have a demasculinizing effect in males, meaning that their brains show less masculinization due to insufficient androgen exposure. This demasculinizing effect could contribute to changes in the adult brain and behavior.

The LBN model alters maternal care

The hallmark of the LBN model is that it alters maternal care. One fairly consistent finding in rats and mice is that dams in the LBN condition make more nest departures causing maternal care to be more fragmented than in standard housing conditions (Ivy et al., 2008; Rice et al., 2008; Baram et al., 2012; Goodwill et al., 2019) but this effect is not always reported (McLaughlin et al., 2016). Behavior on the nest has also been assessed in rats and there are differing findings about the quality of maternal care. An early report suggested that LBN dams spend less time licking and grooming their pups and less time engaging in arched-back nursing than standard housed dams (Ivy et al., 2008). Yet others reported that LBN does not alter licking and grooming and, in some cases, increases nursing (McLaughlin et al., 2016; Moussaoui et al., 2016). These discrepancies could reflect strain differences or other differences in environmental conditions (e.g., handling, shipping dams while pregnant or not, amount of enrichment in the standard housing condition, cage changes during the model, method of scoring care, timing of scoring care, etc.) (Tractenberg et al., 2016; Murthy & Gould, 2018).

We counted a variety of maternal care behaviors and classified them as pup-directed or self-care. In the dark cycle towards the middle period of the LBN exposure (PND5 and PND7), we found a significant increase in pup-directed behaviors, including more arched-back nursing on PND5 in the LBN dams. In contrast, LBN dams spent less time engaging self-care behaviors. For example, LBN dams were rarely found resting outside the nest, an effect that reached significance on PND7. The pattern of altered maternal care in the LBN condition (i.e., more pup-directed behaviors and less self-care) was also observed on PND3 and PND9, although it did not reach significance. Fewer significant effects were observed in the light cycle, which is consistent with other reports (e.g., (McLaughlin et al., 2016; Strzelewicz et al., 2019)) and highlights the utility of dark cycle observations for assessing maternal care. It should be noted that we did not do continuous monitoring of maternal care, so neither the sequence of care nor the degree of fragmentation (as assessed by nest entries and exits) were evaluated. However, a recent analysis using continuous video monitoring in mice found that, compare to control dams, LBN dams make more nest entries and exits, while also spending more time on their nests (Gallo et al., 2019). It is likely that a similar pattern of behavior would be observed in our rats if we were able to conduct in-depth, continuous monitoring. Interestingly, this increased nest time and pup-directed care in the LBN condition may be an attempt by LBN dams to compensate for the limited resources provided.

Increased maternal care can improve certain pup outcomes (Francis et al., 1999; Liu et al., 2000; Lacagnina et al., 2017). However, excess time on the nest engaged in care does not reflect the typical pattern of maternal care in a more naturalistic setting. To better mimic the living conditions of rats in the wild, researchers have developed enriched environments, which typically consist of large multilevel spaces with many objects to interact with (Bradshaw & Poling, 1991; Makowska & Weary, 2016). Rat dams in an enriched environment spend less time on the nest than dams in standard housing conditions (Connors et al., 2015; Strzelewicz et al., 2019). However, when on the nest, dams in the enriched environment have an increase in arched-back nursing relative to dams in standard housing. Thus, in a naturalistic setting, dams spend much of their time engaged in activities that do not involve pups, but when they do provide care, it is of a high quality. When compared to dams and pups exposed to enriched environments, dams raising pups in standard laboratory housing conditions have been equated to “helicopter parents” due to the large amount of time spent with their pups and the fact that their offspring are less adaptable to potentially threatening environments (Connors et al., 2015). Relative to standard housing conditions, we found that the LBN model increased dams’ pup-directed care, while reducing their time spent engaged in self-care behaviors. It is unclear what drives the altered care, but perhaps the stress of the limited resources causes a high arousal state in the dams resulting in hypervigilance towards the pups. Unfortunately, the increased nursing and care by LBN dams does not appear to mitigate against all of the negative effects of early exposure to a low resource environment, because their offspring gain less weight and have altered endocrine profiles compared to pups raised in a standard housing environment as detailed below.

The LBN model affects weight gain, but not other measures of development

Consistent with previous reports, we found that LBN pups weighed less than those raised in the standard housing condition when assessed immediately following exposure to the LBN environment (Brunson et al., 2005; McLaughlin et al., 2016; Moussaoui et al., 2017). Even after being returned to standard housing conditions, LBN exposed pups gained less weight and this effect persisted into adulthood in our hands. There were no sex differences in the effect of LBN on weight gain, although females weighed less than males, which is an expected finding. This delayed growth seems specific to metabolic processes because other markers of development, such as eye opening and vaginal opening, were unaffected by the LBN manipulation. It is important to note that we found that LBN dams spent more time nursing their pups than dams in standard housing conditions. Thus, these results suggest that the stress of the LBN environment changes the quality, not the quantity, of nutrition provided by the dams, causing lasting metabolic changes in the offspring. In support of this, prior studies found that LBN exposed pups have reduced levels of certain micronutrients, glucose, white adipose tissue, and plasma leptin, and these changes could contribute to their decreased weight (Moussaoui et al., 2017; Naninck et al., 2017; Yam et al., 2017). Interestingly, dietary supplementation to dams restored micronutrient levels in pups and rescued LBN-induced cognitive deficits, implicating changes in the micronutrients in aspects of brain programming by LBN exposure (Naninck et al., 2017).

LBN does not affect organ weights and HPA axis reactivity in adulthood

Dating back to the pioneering work of Dr. Selye, chronic stress has been known to alter the weight of organs critical for immune and endocrine responses to stress, including the thymus, spleen, and the adrenal glands (Selye, 1937). We assessed the weight of these organs in adult rats exposed to LBN and control nesting. To our knowledge, we are the first group to test whether the weight of the thymus and spleen was altered by LBN exposure in adult rats, though studies have reported a LBN-induced reduction in thymus weight in PND10 mouse pups (Naninck et al., 2016), but no effect of LBN on PND10 rat pups (Guadagno et al., 2018) nor thymus weights in adult mice (Naninck et al., 2015). Similarly, we found no effect of LBN on adult thymus weights in rats. LBN also had no effect on spleen weights. However, consistent with previous reports, the spleen per body weight was higher in females than in males (Webster et al., 1947; Piao et al., 2013). These results may suggest that LBN exposure has no effect on these organs. However, taking organ weight is a gross measure, so it remains possible that a molecular assessment of these tissues would reveal an effect of LBN, and such endpoints could be tested in future studies.

Consistent with reported findings, we found that adult female rats had heavier adrenal glands per body weight than male rats (Trieb et al., 1976; Westenbroek et al., 2003a; Westenbroek et al., 2003b), but there was no effect of LBN on adrenal weights in either sex. Although prior studies did not evaluate females, our findings are consistent with male data revealing that LBN does not induce adrenal hypertrophy in adulthood (Brunson et al., 2005). In contrast, adrenal hypertrophy is typically found in pups after cessation of the LBN manipulation (Avishai-Eliner et al., 2001; Brunson et al., 2005), suggesting a transient effect of LBN on adrenals that recovers in adulthood.

Adrenal hypertrophy is associated with excess glucocorticoid secretion and, as noted, baseline corticosterone levels in pups are often higher in those exposed to LBN than standard housing (Gilles et al., 1996; Avishai-Eliner et al., 2001; Brunson et al., 2005). Although one study found a decrease in adrenal weights and basal corticosterone levels in LBN exposed pups, a discrepancy that was attributed to strain differences (Moussaoui et al., 2016). Chronic stress not only alters basal corticosterone levels, but it can also alter HPA axis reactivity to a heterotypic stressor (Dallman et al., 1992). As noted, prior studies have found that immediately following LBN exposure there is a decrease HPA axis hormones in response to a heterotypic stress (Gilles et al., 1996; McLaughlin et al., 2016). Collectively, these studies reveal that upon completion of the LBN stressor exposure, there are observable changes in the HPA axis of pups. However, it remained unclear whether the LBN condition caused a lasting increase in HPA axis reactivity to a heterotypic stressor that persisted into adulthood. We tested this here by exposing adult male and female rats in the LBN or control manipulation to 1-hour restraint stress and assessing the time course of corticosterone release. Consistent with prior studies (Kitay, 1961; Weinstock et al., 1998), we found that female rats had higher levels of corticosterone than male rats. However, LBN exposure did not alter the corticosterone response to stress in adulthood in either sex. When considered with adrenal weight data, which was also unaffected by LBN, these findings indicate that LBN exposure does not reprogram the HPA axis. Thus, effects of LBN on metabolism and other endpoints cannot be attributed to alterations in the endocrine response to stress.

Effect of LBN on reproductive development and gonadal hormones

LBN exposure reduced AGD in both prepubertal male and female rats, even when controlling for body weight. To our knowledge, this is the first report of a postnatal stressor altering AGD in developing rats. Prenatal stressors either decrease or increase AGD, depending on the stressor and timing of exposure (Morgan & Bale, 2011; Barrett et al., 2013; Ashworth et al., 2016; Desaulniers et al., 2016). There is also evidence that postnatal stressors can alter the onset of puberty in both male and female rodents (Biagini & Pich, 2002; Bodensteiner et al., 2014; Cowan & Richardson, 2019; Knop et al., 2019). Yet, fewer studies have investigated the impact of postnatal stressors on AGD. The studies that have addressed this question report no effect of maternal separation stress (Biagini & Pich, 2002; Mesquita et al., 2007) nor neonatal corticosterone administration on AGD (Biagini & Pich, 2002). The results presented here are therefore novel in demonstrating that AGD development can still be sensitive to disruption by stressors as late as the first week of postnatal life.

A reduction in AGD is associated with lower prenatal androgen exposure in both sexes (Clemens et al., 1978; McDermott et al., 1978; Hotchkiss et al., 2007; MacLeod et al., 2010; van den Driesche et al., 2011). Interestingly, the LBN manipulation happens postnatally after the perinatal surge in testosterone observed in males (Corbier et al., 1992). Thus, the LBN condition could not alter the perinatal testosterone surge itself in males. Rather something about the sensory experience of the pups in the LBN condition is likely affecting hormone exposure. Prior work has found that variations in maternal care can alter the masculinization of behavior estrogen receptor (ER)α gene expression in the brains of offspring via methylation of the ERα gene (Champagne et al., 2006; Kurian et al., 2010; Edelmann & Auger, 2011). Yet, the methylation pattern and direction of ERα expression differ between reports, and these differences are thought to be attributed to the brain region specificity, offspring age, and/or type of maternal care manipulation (i.e., natural variations in care vs. experimenter simulated care) (Champagne et al., 2006; Kurian et al., 2010; Edelmann & Auger, 2011). However, given that testosterone is aromatized into estradiol to masculinize the brain, an LBN-induced downregulation of ERα could reduce that efficacy of androgens on brain masculinization. Future studies are needed to examine this idea. Another untested possibility is that LBN affects androgen receptor expression via an epigenetic process.

Despite evidence for changes in early androgen exposure, plasma testosterone levels in adult males were not affected by LBN. This result could indicate that androgens are not permanently altered throughout development. In contrast, plasma estradiol levels in males were increased by LBN exposure. In males, plasma estradiol is primarily derived from testosterone by aromatase in fat and, in fact, estradiol levels in males correlate with fat mass (Vermeulen et al., 2002). In our model, LBN exposed males weighed less but had higher levels of plasma estradiol, so the change in hormones cannot be attributed to the change in weight. Given that LBN exposure in males does not increase testosterone but it does increase estradiol, it is likely that LBN increases aromatase activity causing a more efficient conversion of testosterone to estradiol.

As noted, LBN reduces AGD but does not affect plasma testosterone levels in adult males, suggesting that LBN causes a transient change in androgen exposure. However, even a transient change in hormone exposure early in life can have lasting consequences on brain development. In rats, testosterone exposure (via its conversion to estradiol) during the sensitive period, between embryonic day 18 to PND 7, can masculinize the brain causing changes that persist throughout the lifespan (Rhees et al., 1990; Rochellys Diaz et al., 1995). The LBN manipulation, which occurs from PND 2–9, overlaps with a portion of this sensitive period. We found that the LBN manipulation caused a demasculinizing effect on the body by reducing AGD in males. It is therefore likely that this manipulation is also causing a demasculinizing effect on the brain. Consistent with this idea, maternal separation stress during the sensitive period alters male sex behavior, increasing latencies to mount and intromit, as well as reducing ejaculations (Rhees et al., 2001). Future studies will examine whether LBN similarly disrupts male reproductive behavior.

We assessed aspects of female reproductive development and found that LBN exposure neither altered the timing of vaginal opening during puberty nor the duration of estrous cycle in adult females. Consistent with this, maternal separation in rats from PND 2–9 also had no effect on these endpoints (Rhees et al., 2001). In mice, the estrous cycle in adulthood was also unaffected by LBN exposure between PND 4–11, but this manipulation did delay vaginal opening, a discrepancy that may be due to a species difference or slight differences in the timing of the model (Manzano Nieves et al., 2019). In humans, abuse early in life is associated with precocious puberty (Power et al., 2014; Noll et al., 2017). It is possible that the LBN model in rats is not traumatic enough to cause these changes. Another factor to consider, however, is diet. Many people in developed countries eat a Western diet, characterized by high levels of fat. In our study, rats were fed standard chow that is relatively low in fat. However, female rats exposed to both LBN and a Western diet high in fat had earlier vaginal opening and alterations in the hypothalamic kisspeptin system (Strzelewicz et al., 2019). Thus, early life stress and a Western diet may interact to accelerate puberty. In support of this idea, girls adopted from developing countries into Western countries go through puberty earlier than their peers that remain in the developing country, an effect that has been attributed to the change in diet but also could be exacerbated by the stress of adoption into a different culture (Virdis et al., 1998).

Conclusions

The studies here add to the literature on how early life adversity affects dams and their offspring. We found that LBN exposure caused lasting metabolic and endocrine effects but did not alter all aspects of development. Although female reproductive behavior seemed unaffected, the smaller AGD and higher plasma estradiol observed in LBN males suggests that this manipulation may have altered steroid hormone exposure in males. These findings highlight a need to consider the effect of LBN on reproductive hormones and reproductive function in males. Moreover, given that estradiol regulates a host of other behaviors, such as memory and anxiety, LBN-induced changes in steroid hormones may contribute to the alterations in cognition and affect observed by others using this model (Walker et al., 2017).

Abbreviations

AGD

anogenital distance

ER

estrogen receptor

HPA

hypothalamic pituitary adrenal

LBN

limited bedding and nesting

LSD

least significant difference

PND

postnatal day

SES

low socioeconomic status

SD

standard deviation