Corticotropin-releasing factor infusion in the bed nucleus of the stria terminalis of lactating mice alters maternal care and induces behavioural phenotypes in offspring

The peripartum period is accompanied by numerous physiological and behavioural adaptations organised by the maternal brain. These changes are essential for adequate expression of maternal behaviour, thereby ensuring proper development of the offspring. The corticotropin-releasing factor (CRF) plays a key role in a variety of behaviours accompanying stress, anxiety, and depression. There is also evidence that CRF contributes to maladaptations during the peripartum period. We investigated the effects of CRF in the bed nucleus of the stria terminalis (BNST) of lactating mice during maternal care and analysed locomotor activity and anxiety-like behaviour in the offspring. The BNST has been implicated in anxiety behaviour and regulation of the stress response. The effects of intra-BNST CRF administration were compared with those induced by the limited bedding (LB) procedure, a model that produces altered maternal behaviour. BALB/cJ dams were exposed to five infusions of CRF or saline into the BNST in the first weeks after birth while the LB dams were exposed to limited nesting material from postnatal days (P) 2–9. Maternal behaviour was recorded in intercalated days, from P1-9. Offspring anxiety-like behaviour was assessed during adulthood using the open-field, elevated plus-maze, and light/dark tests. Both intra-BNST CRF and LB exposure produced altered maternal care, represented by decreased arched-back nursing and increased frequency of exits from the nest. These changes in maternal care resulted in robust sex-based differences in the offspring’s behavioural responses during adulthood. Females raised by CRF-infused dams exhibited increased anxiety-like behaviour, whereas males presented a significant decrease in anxiety. On the other hand, both males and females raised by dams exposed to LB showed higher locomotor activity. Our study demonstrates that maternal care is impaired by intra-BNST CRF administrations, and these maladaptations are similar to exposure to adverse early environments. These procedures, however, produce distinct phenotypes in mice during young adulthood and suggest sex-based differences in the susceptibility to poor maternal care.


Results
Maternal behaviour. Regarding ABN behaviour, repeated-measures ANOVA showed a significant treatment effect, in which CRF treated dams showed a reduction in ABN compared to saline solution treated dams (SAL) [F (1, 10) = 21.385, p = 0.001; Fig. 3a]. Moreover, LB dams also presented a decrease in ABN compared to undisturbed control dams (CT) [F (1, 10) = 16.281, p = 0.002; Fig. 3b]. There was a significant treatment effect in relation to the number of exits from the nest, where CRF dams had an increased number of exits from the nest [F (1, 10) = 20.514, p = 0.001; Fig. 4a]. This pattern was also seen in LB dams compared to the pattern seen in CT dams [F (1, 10) = 218.788, p = 0.001; Fig. 4b]. Moreover, when comparing CT (non-infused) and SALtreated groups, the two-way repeated-measures ANOVA showed no significant treatment or interaction effects [F (1,12) = 0.568, p = 0.465; F (1,12) = 0.91, p = 0.768, respectively]. These results indicate the absence of a saline infusion effect. No significant differences were detected in the other behaviours (Supplementary Figs. S1, S2 and S3).
Offspring body weight. There were no significant differences in body weight between SAL and CRF mice on P9 or P21 (Fig. 5a,b, respectively). However, LB animals had reduced body weight at P9 compared to CT animals [t (14) = 2.391, p = 0.0314; Fig. 5d], but this difference was not detected on P21 (Fig. 5e). We identified a sex effect [F (1, 42) = 218.74, p < 0.0001; Fig. 5c; SAL × CRF] [F (1, 45) = 500.01, p < 0.0001; Fig. 5f; CT × LB] on P58; as expected, independent of the treatment, females presented a lower body weight than males.   Fig. 7e,f, respectively]. No significant differences were observed in the number of head dipping behaviour and in self-grooming behaviour in CRF and LB animals compared to SAL and CT animals, respectively (Supplementary Table S1). Moreover, no significant differences were observed between LB and CT animals on the time spent in the open arms (Fig. 7d).
There was a significant treatment effect in the LD test, in which animals of the CRF group took longer time to enter the dark zone [F (1,40) Table S1).

Discussion
We demonstrated that CRF infusion in the BNST of dams during the first weeks after birth altered maternal care by decreasing ABN and increasing the number of exits from the nest. These changes in maternal care resulted in increased anxiety-like behaviour among female offspring, while it was decreased among male offspring. Higher locomotor activity was observed in both male and female offspring of CRF-and LB-exposed dams. Although this was a sex-specific effect, the present study suggests that altered maternal care is an important factor when accounting for the effects of rodent models of early life stress. Additionally, we have shown that these behavioural alterations resemble those observed in offspring raised in a stressful environment elicited by LB exposure. Anxiety-like behaviour can be assessed using different protocols. In this study, we opted to use the classic ones: the OF, EPM, and LD tests 35 . Standardized measures such as the time spent in the open arms of the EPM or in the centre of the OF are most often reported when investigating anxiety-like behaviour. However, ethological measures are also important to fully comprehend how an intervention can impact a rodent behaviour 36 . With that in mind, we also analysed risk assessment, rearing, stretching and self-grooming behaviour as ethological measures. In the OF test, CRF and LB animals increased their total distance travelled compared to SAL and CT animals, respectively. This is a well-known behavioural measure of locomotor activity and exploration, and our results are congruent with those found by Aya-Ramos et al. 37 and Wang et al. 38 , who found that measures of locomotor activity increase significantly in animals subjected to postnatal maternal separation. Correspondingly, CRF and LB animals increased the number of rearing behaviours compared to its respective controls, which is associated with locomotion and exploration 39 . Nevertheless, we observed no change in the time spent in the centre zone, which is a classical measure of anxiety-like behaviour in this test.  40 . Intriguingly, we observed a significant interaction effect, where CRF males spent more time in the open arms, whereas CRF females spent less time compared to SAL animals, but no difference in this regard was observed between CT and LB animals. This interaction is in agreement with previous studies that showed an anxiolytic effect of maternal separation in male offspring 41 . In addition, it also agrees with previous studies showing that female animals are more vulnerable to the effects of early life stress in a wide range of behavioural and neurobiological outcomes 42,43 . According to Keller et al. 44 , the increased vulnerability in females may be related to impaired maternal care. These authors showed that in a maltreatment condition, female offspring received higher adverse care from the dams than male offspring. Additionally, Bowers et al. 45 presented evidence that male pups, when separated from their mothers, emitted more ultrasonic vocalisations than female pups, and the dams were more likely to retrieve the male pups first. With this data, we suppose that due to the increased vocalisation, males receive more maternal care than females and, therefore, could be better protected against early life stress effects, thereby presenting less anxiety-like behaviour phenotypes later in life. In the LD test, we showed that CRF and LB animals took longer to enter the dark zone and increased the number of head pokes into the transition from the light to the dark zone than SAL and CT animals. These behaviours suggest a stronger link to risk assessment and exploratory actions, corroborating the OF test results. However, we could not observe differences in the time spent in each chamber and the total number of transitions, which are considered classic measures of the LD test. www.nature.com/scientificreports/ Behavioural alterations observed in the offspring could be elicited by activation of the CRF-BNST signalling pathway in the limbic system of dams. As mentioned, CRF expression activates the HPA axis, increasing the release of glucocorticoids 28 . For this reason, CRF infusion in the BNST has been associated with different behavioural outcomes such as anxiogenic-like effects and involvement in the fear response [46][47][48][49] . CRF in the BNST also plays a role in social-defeat and anhedonia 29,50 . Although there is growing evidence suggesting that the CRF signalling in the BNST has an important role in maternal care 51 there are other regions strongly related to this behaviour, such as the hypothalamic medial preoptic area (mPOA). To be noted, the mPOA and its adjacent projections to the BNST form what is known as a "super-core" brain hub that regulates maternal care 52 . In this sense, a recent study from Klampfl et al. 53 presented the effects of CRFR1 agonist infusion in the mPOA on maternal care and maternal anxiety-like behaviour. The study evidenced that the infusion of CRFR1 in the mPOA leads to a decrease in ABN, and also to an increase in off-nest behaviours and anxiety-like phenotypes. Infused dams also presented increased expression of oxytocin in the mPOA. The oxytocin is a neuropeptide highly expressed during lactation; its receptor is present in different regions including the BNST and the mPOA 54 . This increase in oxytocin expression may be a tentative mechanism triggered to counteract the effects of CRF overexpression. However, even though the oxytocin is extremely important to the onset of maternal care it does not enhance ongoing maternal care 55 . Although the assessment of oxytocin goes beyond the scope of this study, one possibility is that the BNST and the mPOA act together in the regulation of those behaviours through CRF and oxytocinergic signalling. It also should be noted that the BNST is highly interconnected with the central amygdala, which mediates key behavioural changes triggered by stress exposure 56 . As mentioned, CRF injection into the BNST may elevate general anxiety levels, thereby altering maternal behaviour and anxiety-like behaviour in the offspring. However, this might indicate an indirect involvement of this nucleus in mediating specific maternal behaviours. Therefore, our findings should be carefully interpreted, particularly considering that maternal behaviours have been reported to be more dependent on ventral area (below the anterior commissure) of the BNST sub-nuclei 57 .
Several studies that applied the LB protocol reported that stressed LB dams increased the number of exits from the nest compared to CT dams 18,20,58-61 which correspond with our results, as CRF and LB dams increased their nest exits compared to SAL and CT dams, respectively. This augmentation in the frequency of exits from the nest alters the natural pattern of maternal behaviour, resulting in fragmented maternal care, due to the increased www.nature.com/scientificreports/ number of times that the dam interrupts nurturing behaviour. We also observed a decrease in the frequency of ABN in CRF dams compared to that in SAL dams. However, our ABN results do not totally correspond to those of other studies with early life stress protocols. Most studies that used maternal separation report an increase in this behaviour after reunion in the mothers [62][63][64][65] , probably as an attempt to compensate for the time away from the pups. Unfortunately, there is a lack of information about maternal behaviour before and after a longer period before reunion. Although these maternal separation studies reported an increased ABN, they also reported anxiety-like behaviour in the offspring of these stressed dams, so we suppose that the compensatory nursing behaviour post-reunion is not effective in protecting pups from further consequences. Certain limitations should be considered when interpreting the present study's findings. First, we did not measure specific anxiety and depression-like behavior in the dams during the infusion period. Moreover, the same offspring cohort of animals was used for the behaviour battery. However, in order to address the cumulative stress issue, we performed the tests from less to more stressful, and all groups went through the same protocol. Next, this study tested only one dose of CRF. This dose was chosen due to our intent to replicate previously presented results 23 . Finally, the technique used for the infusions did not allow us to measure the degree of CRF spread to the surrounding regions of the BNST. Even though we cannot exclude this possibility, it is unlikely that the CRF spread to subjacent areas since we used a slow injection rate (0.1 ul/min) for a small amount of drug (1 μg/0.5 μL).
Therefore, we provided additional evidence to the hypothesis that poor maternal care during early life anticipates altered behavioural phenotypes in rodents during adolescence and adulthood, including changes in exploratory and anxiety-like behaviours. Although the exact mechanisms underlying these outcomes and the sex-specific effect remain unknown, it is important to note that sex is a key factor to alterations in exploratory and anxiety-like behaviours elicited by poor mother-infant attachment.  www.nature.com/scientificreports/ Stereotaxic surgery and infusions. Adult females (~ P60) from the SAL and CRF groups were implanted with a dual-cannula system (P1 Technologies; Roanoke, VA) targeting the BNST under intraperitoneal anaesthesia using a combination of ketamine (100 mg/kg) and xylazine (10 mg/kg). As per the mouse brain stereotaxic coordinate atlas 66 , the coordinates were + 0.3 mm posterior to the bregma, 1.1 mm lateral to the midline, and 4.3 mm ventral to the dura. CRF (1 μg/0.5 μL, CRF human/rat, Tocris Bioscience; Bristol, United Kingdom) and SAL (0.5 μL, 0.9% NaCl) were infused at 0.1 μL/min using an infusion pump (Accu-Chek Insight, Roche; Basel, Switzerland). Doses were chosen based on previous studies 23,34,67 . To perform the infusions, dams were removed from their home cage and transported into a clean cage to the infusion room, where they were quickly immobilised. Injectors were inserted into the cannula, and the dam was able to freely move during the infusion.
Infusions were repeated five times on P1, P3, P5, P7, and P9. To verify the correct placement of the cannula, the dams were perfused on the day of weaning. The brains were cut into 30-μm coronal sections, mounted on slides, and stained with haematoxylin and eosin (Fig. 2a,b).
LB protocol. The LB protocol was adapted from Rice et al. 18 . From P2 to P9, CT dams and offspring were placed in cages with standard amounts of wood shavings, while LB dams and offspring were placed in cages with 1 g of cotton and an aluminium mesh above the wood shavings. After P9, the LB dams and litters were placed in new cages prepared according to CT conditions.

Maternal behaviour observations. The maternal behaviour observation protocol was adapted from
Klampfl et al. 23 . The protocol was conducted on P1, P3, P5, P7, and P9 (on days of SAL/CRF infusions). Observations were conducted in seven blocks of 30 min each: before the infusion (~ 08:00-09:00, blocks 1 and 2), 30 min after infusion (~ 09:30-11:00; blocks 3, 4, and 5), and the last 5 h after infusion (~ 14:00-15:00, blocks 6 and 7). The same days and hours were used to observe the maternal behaviour of CT and LB dams. One observation was made every 2 min for 105 observations per day. The frequency of the following behavioural categories was evaluated: ABN, blanket or passive nursing (N), licking and grooming (L), off nest (X), self-grooming (S), eating or drinking (E), and the number of exits from the nest (EN). Maternal behaviour was recorded by two independent observers. Results are presented as the mean of all days from each block.
Anxiety-like behaviour tests. OF, EPM, and LD tests were performed on P58, P59, and P60, respectively.
All tests were video-recorded and analysed using the ANY-maze software (Stoelting; Wood Dale, IL). All tests were performed during the light phase. The apparatus was cleaned with 70% ethanol before testing the next animal.
OF test. The OF apparatus consisted of a clear Plexiglas box (33 cm × 33 cm), and the test was performed under 140-150 lx illumination. Mice were placed in the middle of the apparatus and allowed to explore for 10 min. The box was divided into 16 squares: the central four squares composed the centre zone, and the remaining squares composed the periphery zone. We evaluated the time spent in the centre and periphery zones, the total distance travelled and the number of instances of rearing, stretching and self-grooming behaviour. LD test. The apparatus used for the LD test was constructed with Plexiglas (21 cm × 42 cm × 20 cm), divided in two identical-sized chambers connected by a door. One side was illuminated at 390 lx, and the other was completely dark. The animal was placed in the bright chamber and allowed to explore for 10 min. The total time spent in each chamber, latency to first entry into the dark chamber, number of transitions, head pokes (risk assessment measure, when the animal lean its head through the connection door) the animal did to enter or exit the dark zone and the number of instances of rearing and self-grooming behaviour were analysed.
Statistical analysis. All statistical analyses were performed using SPSS version 23.0 (IBM; Armonk, NY), and graphs were constructed using GraphPad Prism 8 (GraphPad Software; San Diego, CA). Data are presented as means standard error of the mean (SEM). Statistical significance was considered when p < 0.05. Body weight (P9 and P21) was analysed using Student's t-test. Repeated-measures analysis of variance (ANOVA) and multivariate analysis of variance (MANOVA) were used to analyse maternal behaviour. Two-way ANOVA was used for behavioural tests (OF, EPM, LD) and body weight (P58), with a sex factor (male × female) and a treatment or condition factor (SAL × CRF or CT × LB, respectively). ANOVAs were followed by Tukey's post-hoc tests.