Changes in Corticotropin-Releasing Factor Receptor Type 1, Co-Expression with Tyrosine Hydroxylase and Oxytocin Neurons, and Anxiety-Like Behaviors across the Postpartum Period in Mice

Corticotropin-releasing factor (CRF) and CRF receptor type 1 (CRFR1) have been implicated as a link between stress and mood regulation [1, 5]. Numerous studies have suggested that stress-induced elevations in CRF and chronic/excessive activation of CRFR1 contribute to psychiatric diseases, including anxiety [6, 13]. The CRF/CRFR1 system has also been shown to regulate behaviors during the postpartum period and has been associated with disorders, including postpartum depression and anxiety [14, 20].

There are dramatic shifts in stress-related behaviors and the HPA axis in postpartum rodents, which are regulated in part by CRF and CRFR1 [21, 26]. For example, intracerebroventricular administration of CRF decreases maternal care and increases anxiety-like behaviors in postpartum rats, while antagonism of CRF receptors attenuates these responses to CRF [19]. Furthermore, CRFR1 blockade in select brain regions of postpartum rats rescues deficits in maternal behaviors induced by acute stress, including nursing and aggression, while activation of CRFR1 decreases maternal behaviors and promotes an anxiogenic response [18, 20]. While studies involving postpartum stress-related functions have primarily been performed in rats, one study suggests anxiety-, despair-, and anhedonia-like behaviors are elevated at specific postpartum time points in mice [25]. In relation to the CRF system, activation of CRF neurons in postpartum mice increases postpartum depression-like behaviors and decreases interaction with pups [27]. A global transcription analysis study comparing postpartum and nulliparous mouse brains has also implicated CRFR1 as a key candidate gene that may underlie postpartum changes in stress-related behaviors [28].

Recent studies from our laboratory demonstrated dynamic changes in CRFR1 levels in mice during the mid-lactation period (postpartum day 14, P14) that may contribute to stress-related behavioral changes in maternal mice. Specifically, we found changes within three brain regions associated with stress regulation and maternal care: the rostral anteroventral periventricular nucleus (AVPV/PeN), paraventricular hypothalamus (PVN), and medial preoptic area (MPOA) [26]. We also observed changes in CRFR1 co-localization with tyrosine hydroxylase (TH) within select brain regions, particularly the AVPV/PeN [26]. More recently, we reported that hypothalamic oxytocin neurons express CRFR1 only in postpartum female mice [29].

Although our previous study showed differences in CRFR1 between P14 and nulliparous mice [26], this study sought to determine the specific timing of changes in CRFR1 during the postpartum period. We mapped out the changes in CRFR1 across five postpartum time points utilizing a validated bacterial artificial chromosome identified by green fluorescence protein mouse line [30]. Collectively, monitoring the changes of CRFR1 throughout the postpartum period can aid in understanding the potential sensitive time points that might leave postpartum females vulnerable to stress-related behavioral impairments. Since there are limited data regarding changes in stress-related behaviors in postpartum mice, we also assessed changes in anxiety-like behaviors across the postpartum period. This information can be used to understand the timing of changes in stress-related behaviors relative to postpartum hormone fluctuations and various stress-regulating genes, including CRFR1.

We used a validated CRFR1-GFP reporter mouse line [26, 30, 31]. Prior to the experiment, mice were genotyped to determine GFP+ mice using polymerase chain reaction. Forty-eight sexually inexperienced female mice were housed with one or two same-sex litter mates in clear polycarbonate cages (7.25″ W × 11.5″ D × 5″ H) with microfilter top and Aspen chip bedding, and each cage included 1 polycarbonate mouse igloo. Mice were maintained under a 12 h light: 12 h dark cycle (lights on at 7:00 a.m.) and had access to pellets (LabDiet 5P76 Irradiated ProLab IsoPro RMH 3000) and water ad libitum. For breeding, female mice were co-housed with sexually experienced breeding males until pregnancy. Females were monitored at 2 weeks after mating and were singly housed prior to parturition. The removal of the male mouse was essential to prevent impregnating the female mouse after parturition and potentially inducing hormonal and behavioral fluctuations. Day of parturition was considered postpartum day 0 (P0) and pups remained with dams throughout postpartum. We compared postpartum mice at five time points (P1, P7, P14, P21, P28; n = 8 per group) and nulliparous mice (n = 8). All procedures were conducted according to the National Institutes of Health Guide for the Care and Use of Animals. This research protocol was reviewed and approved by the Institutional Animal Care and Use Committee at University at Albany, approval number 19-004.

The open field test (OFT) apparatus was an opaque Plexiglas cube (16″ × 16″ × 16″) with the top side open as previously described [32]. Testing was performed between 10:00 a.m. and 12:00 p.m. and the OFT was cleaned with Virex^® and dried between animals. Specifically, mice were transferred from the vivarium into the OFT room. Each mouse was placed into a standard corner of the apparatus and was recorded for 10 min by a camera mounted on the ceiling. After 10 min, mice were returned to the vivarium for 90 min after OFT initiation until blood collection. We used ANY-Maze software (Stoelting Co.) to superimpose inner (8″ × 8″) and outer areas of the apparatus [32, 33]. We quantified distance traveled, center entries, latency to enter the center, time spent in the center, mean speed, and freezing time. OFT methods and analysis are identical to our previous study [32].

Trunk blood samples were collected 90 min after OFT initiation, kept on ice, and subsequently centrifuged at 5500 RCF for 10 min at 4°C. Plasma samples were stored at −80°C until corticosterone (CORT) levels were determined using a double antibody 125I radioimmunoassay kit (MP Biomedicals, Solon, USA) [26, 34]. Samples ran in duplicate according to the manufacturer’s instructions. The intra-assay coefficients of variation were less than 5%.

Ninety minutes after OFT initiation and immediately after trunk blood collection, brains were extracted, post-fixed in 4% paraformaldehyde overnight at 4°C, and cryoprotected in 30% sucrose solution containing sodium azide at 4°C until brain sectioning. Brains were embedded in Optimal Cutting Temperature (OCT) Compound (Tissue-Tek^®), coronally sectioned at 30 µm on a cryostat (Microm HM505E, MICROM International GmbH, Waldorf, Germany), separated into four vials containing cryopreserve, and stored at 4°C until immunohistochemistry.

We visualized CRFR1-GFP-ir via immunohistochemistry as previously described [26, 35, 36]. Briefly, cryopreserved free-floating tissue sections were transferred into Netwell™ inserts, rinsed with 0.1 m phosphate-buffered saline (PBS; pH 7.6) 5 times (5 min each time) to wash off cryopreserve, incubated in 0.3% Triton-X in PBS (0.3% PBS-TX) and 1% hydrogen peroxide for 10 min, rinsed in 0.1 m PBS 5 times (5 min each time), blocked in 4% normal goat serum and 0.3% PBS-TX for 60 min, and incubated in primary green fluorescent protein (GFP) antisera (rabbit, Life Technologies, RRID: AB221570, 1:5,000) at room temperature overnight. On the following day, tissue sections were rinsed in 0.1 m PBS 3 times (5 min each time), incubated in biotinylated antisera solution (goat anti-rabbit; Vector Laboratories; 1:500) for 60 min, rinsed in 0.1 m PBS, incubated in avidin-biotin peroxidase complex (ABC Elite kit, Vector Laboratories; 1:1,000) for 60 min, rinsed in tris-buffered saline, developed in diaminobenzidine for 10 min, and rinsed in 0.1 m PBS 4 times (5 min each time). Sections were mounted onto gel-coated slides, processed with ethanol and xylene, and coverslipped with permount.

Cryopreserved free-floating tissue sections were rinsed with 0.1 m PBS (pH 7.6) 5 times (5 min each time), blocked in 4% normal donkey serum and 0.3% PBS-TX for 60 min, and incubated in primary antisera including GFP (chicken; Abcam, RRID: AB300798, 1:1,500), TH (rabbit; Millipore AB152, RRID: AB390204, 1:500), oxytocin (OT; rabbit; Peninsula Laboratories, RRID:AB_518522; 1:500), and c-Fos (goat; Santa Cruz sc-52-G, RRID:AB_2629503; 1:250) at room temperature overnight. On the following day, tissue sections were rinsed with 0.1 m PBS (pH 7.6) 3 times (5 min each time), incubated in various secondary antisera (anti-rabbit alexa 594 [1:500, Invitrogen]; anti-rabbit alexa 350 [1:250, Invitrogen], anti-chicken alexa 488 [1:1,000, Jackson Laboratory], anti-goat alexa 594 [1:250, Invitrogen]) in 4% normal donkey serum and 0.3% PBS-TX for 2.5 h. Sections were mounted onto gel-coated slides and coverslipped with antifade mounting medium (VECTASHIELD^® HardSet™, Vector Laboratories). Antibody characterizations for GFP, TH, OT, and c-Fos were previously described [29, 32, 35, 37].

Brain regions of interest were identified using Allen Institute mouse brain coronal reference atlas (https://mouse.brain-map.org/static/atlas) and were captured on a Nikon 80i Eclipse microscope at ×20 magnification. Quantitative assessment of CRFR1-GFP neurons was performed using published regions of interest as previously described [26]: AVPV/PeN (Allen Atlas plate numbers 53-54), PVN (plates 62-63), MPOA (plates 53-54), arcuate nucleus (ARN; plates 72-73), supraoptic nucleus (SON; plates 60-61), and ventral tegmental area (VTA; plates 82-83). We used ImageJ (1.51j8 Wayne Rasband, NIH, USA) to quantify the number of CRFR1-GFP, TH, OT, c-Fos, and co-labeled cells. Each brain region was captured bilaterally from 2 sections. In some instances, an animal was omitted from analysis due to tissue tearing or overlapping. For chromagen-labeled tissue sections in the AVPV/PeN, PVN, MPOA, and ARN, we assigned individual CRFR1-GFP cells as having dark or light label based on a subjective rating as previously described for assessing CRFR1 changes in postpartum mice [26]. We applied this method to assess relative GFP content within a cell, which is reflective of CRFR1 gene expression as the GFP is driven by the CRFR1 promoter [30]. Thus, a darkly labeled cell reflects greater CRFR1 levels than a lightly labeled cell. This method allows us to assess relative changes in CRFR1 as in instances where CRFR1 might decrease within a neuron but not fully disappear or conversely in instances where CRFR1 levels increase within a neuron that previously may have had low levels. Number of cells shown in figures indicate average cell number per unilateral brain section.

Statistical analyses were performed using GraphPad (v5.01, GraphPad Software, San Diego, CA, USA). All data were analyzed using one-way univariate analysis of variance (ANOVA) with postpartum time points as between-subjects factor. Statistically significant ANOVA effects were further analyzed using Tukey’s post hoc test. Data were reported as mean +/− standard error of the mean with statistical significance level set at p ≤ 0.05.

We used nonfluorescence microscopy to assess the number of CRFR1-GFP-ir neurons within the AVPV/PeN, MPOA, PVN, and ARN. In the AVPV/PeN (Fig. 1), there were statistically significant differences in the number of dark (F[5, 42] = 8.257, p < 0.0001), light (F[5, 42] = 16.12, p < 0.0001), and total number (F[5, 42] = 3.599, p < 0.0085) of CRFR1-GFP-ir neurons. Post hoc analyses indicate a greater number of dark CRFR1-GFP-ir neurons in postpartum (P1 through P21) mice compared to nulliparous mice (ps < 0.05) with a peak at P1. Light CRFR1-GFP-ir neurons were lower in all postpartum mice compared to nulliparous mice (ps < 0.05), with the lowest levels found in P1 mice. Total CRFR1-GFP-ir neurons were higher at P1 compared to all other groups (ps < 0.05). In the MPOA (Fig. 1), there were significant differences in the number of dark (F[5, 42] = 5.841, p < 0.0004), light (F[5, 42] = 3.544, p < 0.0092), and total number (F[5, 42] = 4.992, p < 0.0011) of CRFR1-GFP-ir neurons. Post hoc analyses indicate a decrease in dark CRFR1-GFP-ir neurons at P1, P21, and P28 compared to nulliparous mice (ps < 0.05). Light and total CRFR1-GFP-ir neurons also showed a decrease in postpartum mice with significant decreases found at all postpartum time points in comparison to nulliparous mice (ps < 0.05). In the PVN (Fig. 2), there were again significant differences in the number of dark (F[5, 42] = 4.449, p < 0.0024), light (F[5, 42] = 2.497, p < 0.046), and total number (F[5, 42] = 3.548, p < 0.0091) of CRFR1-GFP-ir neurons. Post hoc analyses indicate a decrease in dark CRFR1-GFP neurons at P1 through P14 compared to nulliparous mice (ps < 0.05). Light CRFR1-GFP neurons tended to be elevated at P21 although post hoc analyses revealed no specific group differences beyond the overall ANOVA effect. Total PVN CRFR1-GFP neurons were decreased at P7 compared to nulliparous and P21 mice (ps < 0.05). In the ARN (Fig. 2), there were no significant differences in the number of dark, light, or total number of CRFR1-GFP-ir neurons.

We used fluorescence microscopy to assess the number of TH, CRFR1-GFP, and co-localized neurons within regions known to regulate stress and/or maternal functions, including the AVPV/PeN, VTA, PVN, and ARN. In the AVPV/PeN (Fig. 3), there were again significant differences in the number of CRFR1-GFP cells (F[5, 42] = 7.764, p < 0.0001), with all postpartum groups showing an increase compared to nulliparous mice (ps < 0.05). An effect of TH neurons was also found (F[5, 42] = 8.813, p < 0.0001) with P7 and P14 mice showing a significant decrease in TH compared to nulliparous mice. CRFR1-GFP/TH+ neurons also differed by group (F[5, 42] = 10.72, p < 0.0001). Post hoc analyses indicated decreases in CRFR1-GFP/TH co-localized neurons in P7 and P14 mice compared to nulliparous and P21 mice (ps < 0.05). P21 mice also had a greater number of co-labeled neurons compared to nulliparous mice (p < 0.05). Significant effects were also found for the percentage of CRFR1-GFP neurons that co-express TH (F[5, 42] = 9.92, p < 0.0001) and the percentage of TH neurons that co-express CRFR1-GFP (F[5, 42] = 7.187, p < 0.0001). Post hoc analyses indicate a decreased percentage of CRFR1-GFP neurons that express TH at P7 and P14 (ps < 0.001) and an increased percentage of TH neurons that express CRFR1-GFP at P21 (p < 0.05), relative to nulliparous mice.

In the PVN (Fig. 4), there were again significant differences in the number of CRFR1-GFP cells (F[5, 42] = 4.542, p < 0.01) with P7 mice showing a decrease compared to nulliparous and P28 mice (ps < 0.05). ANOVA also indicated a difference in TH neurons (F[5, 42] = 3.255, p < 0.05), although post hoc analyses indicated that no groups statistically differed from nulliparous. There were no significant differences CRFR1-GFP/TH co-expressing neurons in the PVN. There were also no significant differences in the number of CRFR1, TH, and CRFR1-GFP/TH co-expressing neurons in the ARN or VTA (Fig. 4).

Fluorescence microscopy was used to assess OT and CRFR1 co-localization within the PVN and SON. We previously reported substantial restraint stress-induced activation of CRFR1/OT co-expressing neurons in the PVN (assessed by co-localization with c-Fos) [29]. Therefore, in this study, we also examined how activation of CRFR1/OT co-expressing neurons might change across the postpartum period in mice exposed to an open field stress. In the PVN (Fig. 5), there were again significant differences in the number of CRFR1-GFP neurons (F[5, 42] = 3.785, p < 0.01) with P7 mice showing a decrease compared to nulliparous mice. Furthermore, there were significant differences in the number of CRFR1-GFP/OT co-expressing neurons (F[5, 42] = 16.74, p < 0.001) and percentage of OT neurons expressing CRFR1 (F[5, 42] = 20.21, p < 0.001), which indicated increases in CRFR1-GFP/OT co-localized neurons in P1 through P28 compared to nulliparous mice (ps < 0.05). C-Fos levels also differed across groups (F[5, 42] = 4.47, p < 0.01), with P1 mice showing decreased levels relative to nulliparous and P7 mice (ps < 0.05). Significant effects were also found for the number of CRFR1-GFP/OT/c-Fos co-expressing cells (F[5, 42 = 10.03, p < 0.001) and the percentage of CRFR1-GFP/OT neurons that co-express c-Fos (F[5, 42] = 5.20, p < 0.001), with neurons showing increasing co-expression beginning at P1.

In the SON (Fig. 6), there were significant differences in the number of CRFR1-GFP (F[5, 42 = 12.15, p < 0.01) and CRFR1-GFP/OT co-expressing (F[5, 42] = 11.86, p < 0.001) neurons with a slight increase in co-localized cells seen at P1 and a rise found at later postpartum time points. C-Fos levels also differed across groups in the SON (F[5, 42] = 2.99, p < 0.05), with P1 and P21 mice showing decreased levels relative to nulliparous and P7 mice (ps < 0.05). Significant differences were also found for the number of CRFR1-GFP/OT/c-Fos co-expressing cells (F[5, 42] = 2.92, p < 0.05) with OT neurons showing increased co-expression at P28 compared to nulliparous mice. However, it should be noted that there was generally very little co-expression of c-Fos within SON CRFR1-GFP/OT neurons. No significant differences were found for the number of OT neurons in either the PVN or SON.

There was a statistical difference in total time spent in the center area (F[5, 41] = 3.494, p < 0.01; Figure 7a). Post hoc analysis indicates P7 mice spent less time in the center area (indicating increased anxiety-like behavior) compared to nulliparous and P14 mice (ps < 0.05). There were no significant differences in total distance traveled (Fig. 7b), latency to center, total center entries, mean speed, and freezing time (data not shown). There was a significant difference in CORT levels measured at 90 min after open field exposure (F[5,42] = 7.299, p < 0.0001; Figure 7c). Post hoc analysis indicates significant elevations in CORT in P1 mice compared to all other groups (ps < 0.05).

There was a significant difference in body mass (g) (F[5, 45] = 3.494, p < 0.0001). Post hoc analysis indicates nulliparous mice weigh less than P7, P14, and P21 mice (ps < 0.001; Table 1). Litter characteristics (sex ratio and number of pups) did not differ significantly between postpartum mouse time points (Table 1).

Overall, this study demonstrated dynamic changes in CRFR1 in various brain regions (AVPV/PeN, MPOA, PVN, SON) across the postpartum period in mice. We also demonstrate changes in CRFR1 co-localization with TH and OT in brain regions involved in maternal and stress circuitries. We further observed an increase in anxiety-like behaviors mid-postpartum (P7), in the open field test. Together, these changes in CRFR1 and co-expression patterns may potentially influence stress-related behaviors (including changes in anxiety-like behaviors observed in this study) and maternal behaviors in postpartum mice.

Our previous study demonstrated postpartum mice to have higher levels of CRFR1 in the AVPV/PeN compared with nulliparous mice [26]. This study further investigated whether AVPV/PeN CRFR1 levels peaked mid-postpartum (P14) or early- or late-postpartum. We found that P1 mice had the highest levels of CRFR1, and levels remained elevated throughout the postpartum period, although at lower levels than the P1 time point. The timing of this peak in CRFR1 suggests a potential regulation by glucocorticoids since levels spike around parturition [38], which is further evidenced by elevated CORT levels at P1 in mice in this study. Approximately 75% of AVPV/PeN CRFR1 neurons in female mice co-express glucocorticoid receptors [37], suggesting that increased CRFR1 levels could be due to the surge in CORT and CORT binding to glucocorticoid receptors. Furthermore, female mice exposed to 9 days of chronic variable stressors (which is associated with persistent elevations in CORT) had twice as many AVPV/PeN CRFR1 neurons compared with nonstressed females [32]. Functionally, these increased levels of CRFR1 in the postpartum AVPV/PeN might be related to stress-related behavioral changes. A previous study indicated a correlation between anxiety-like behaviors and elevated CRFR1 levels in the AVPV/PeN [32].

Previously, we demonstrated the MPOA of P14 mice had significantly lower levels of CRFR1 compared to nulliparous mice [26]. Our current study revealed that all postpartum mice (P1–P28) had suppressed levels of CRFR1 compared with nulliparous mice, thus replicating and extending our previous results. Because the MPOA is critical for the onset and maintenance of maternal care and motivation [39], naturally suppressing MPOA CRFR1 levels might be key in reducing negative effects of CRFR1 binding on maternal behavior. In support of this idea, pharmacological activation of CRFR1 in the MPOA of postpartum rats decreases nursing and increases anxiety-like behaviors [20]. Therefore, the decrease in MPOA CRFR1 levels might be mediating the suppression of an avoidance/defense circuit to elicit proper maternal behaviors. Further, exposure to offspring throughout the postpartum period, independent of nursing demand, might also be contributing to the suppressed state of CRFR1 levels. For both the MPOA and AVPV/PeN, it is also possible that changes in CRFR1 initiate prior to parturition, although further studies are needed to assess this.

In the PVN, we demonstrated a time-dependent decrease in CRFR1 levels with lowest levels at P7 followed by a return to levels similar to nulliparous mice at P21. PVN CRFR1 neurons have been previously shown to regulate stress-related behaviors and HPA axis responses to stress in male mice. Specifically, PVN CRFR1-expressing neurons have been shown to function as an inhibitory mechanism that controls the HPA axis response to stressors [40]. That study revealed a novel PVN microcircuit between CRF and CRFR1 neurons: CRF neurons excite CRFR1+ neurons via glutamatergic synapses and CRFR1+ neurons inhibit CRF neurons via GABAergic synapses to limit further CRF release and activation of the HPA axis [40]. Therefore, lower levels of CRFR1 might be maladaptive and could potentially lead to increased stress responses and increased anxiety-like behaviors, as seen in P7 mice in the current study. Unlike other brain regions examined, CRFR1 levels in the ARN were not significantly altered during the postpartum period, suggesting a decreased sensitivity to modification in postpartum mice.

Dopaminergic neurons have previously been demonstrated to play a key role in regulating maternal behaviors by facilitating dam-pup interactions [41, 43]. Of note, TH is the rate-limiting enzyme involved in the production of dopamine and norepinephrine. However, all populations assessed in our current study have been characterized as dopaminergic [44, 45]. We expanded on our previous finding of changes in CRFR1 co-localization in dopaminergic neurons at P14 in select brain regions, most notably in the AVPV/PeN [26]. In the AVPV, levels of TH and TH/CRFR1-GFP co-localization in postpartum mice differed in a time-dependent manner, where P7 and P14 mice had the lowest levels of TH/CRFR1-GFP co-labeled neurons and TH neurons compared with nulliparous and other postpartum mice. These findings are consistent with previous reports of decreases in TH at P7 and P14 and TH/CRFR1 co-localization at P14 [26, 46]. TH-expressing neurons in the AVPV/PeN have been shown to regulate maternal behaviors through a circuit involving oxytocin neurons in the PVN [43]. Therefore, decreases in TH/CRFR1 co-localization in the AVPV/PeN might indicate that stress and CRF have diminished effects on these neurons during the mid-postpartum period, potentially impacting how stress alters maternal behaviors. P7 and P14 are time points of highest nursing demand in mice and rats [47]; therefore, these changes in CRFR1/TH co-localization may be driven by nipple stimulation, maternal experiences, or hormones associated with lactation, such as prolactin. TH neurons in the AVPV/PeN express prolactin receptors, suggesting this may be a potential mechanism for regulating these changes [46]. Furthermore, it is possible that increased levels of CRFR1 in the AVPV/PeN may result in increased sensitivity of TH neurons to CRF, which may then drive decreases in TH. Further studies are needed to determine the factors that drive changes in TH/CRFR1 co-localization and the functional significance of these changes. Unlike the AVPV/PeN, co-localization of TH with CRFR1 was generally low and unchanged during the postpartum period in mice within the PVN and ARN. In the VTA, co-localization of TH and CRFR1 was high as previously reported [26, 48], although no significant differences were observed across the postpartum period. Together, these findings suggest the AVPV TH cell group shows the greatest modification during the postpartum period and may therefore be involved in dopamine-related behavioral changes that occur in maternal mice.

As previously described using a CRFR1-Cre^tdTomato mouse line [29], the direct communication between CRF and OT+ neurons (via CRFR1) appears to be a unique feature in the postpartum mouse brain. In this study, we replicated these previous findings now using a CRFR1-GFP mouse line, demonstrating that OT neurons express CRFR1 during the postpartum period with a peak near P21 when mice are commonly weaned. We further demonstrated high levels of OT/CRFR1 co-localized neurons that express c-Fos in mice exposed to an open field stressor, with the greatest levels at P21. This suggests OT/CRFR1 neurons are stress activated and may contribute to stress-related behavioral adaptations during the maternal period. However, it is important to note that c-Fos expression in OT/CRFR1 neurons may not be exclusively due to exposure to the open field stressor. Neural activation in the PVN can also occur in response to pup care and nursing [49]; however, we have observed only very low expression of c-Fos in PVN CRFR1+OT neurons in mice that were not exposed to a stressor prior to euthanasia (Zuloaga lab, unpublished observation).

In our previous study, we demonstrated that mice that have produced several litters co-express CRFR1 in OT neurons of the SON. In this study, we describe the timing of CRFR1 initiation in SON OT neurons in primiparous mice. We report low levels of co-localization at P1 and P7, with an increase at P14 and a peak near P21–P28. Unlike the PVN, there is little to no CRFR1 in the SON of nulliparous mice. CRFR1 initiates expression in the SON during the postpartum period and this occurs almost exclusively within OT neurons. Of note, CRFR1 expression has been observed in OT neurons in rats but only after osmotic stress [50]. However, osmotic stress fails to induce CRFR1 in OT neurons of mice [51] and the potential hormones, experiences, or other factors that induce this phenomenon remain unknown. Overall, this expression of CRFR1 in OT neurons in maternal mice may be a mechanism that allows stress and CRF to regulate OT secretion to facilitate, or suppress, maternal behaviors.

Postpartum day 7 (P7) mice spent less time in the center of the open field, suggesting higher anxiety-like behavior. In rats, various changes in stress-related behaviors have been reported, although fewer studies have investigated postpartum changes in mice [22, 25]. However, the current findings are in line with a previous report in inbred BALB/c mice that indicated increased anxiety-like behavior in the elevated plus maze at P5 and further demonstrated increased passive coping strategy in the forced swim test at P6-P7 [25]. Together, these findings suggest a change in stress coping strategies at this postpartum time point. Of note, anxiety-like behaviors have previously been reported to decrease in outbred Swiss mice exposed to the light/dark box at P7 [22]. This suggests that genetic background is an important factor for regulating changes in stress-related behaviors in postpartum mice. As previously mentioned, there are relatively few studies that have assessed stress-related behaviors in postpartum mice. It will be important for future studies to quantify mouse behaviors in various stress-evoking assays and at various time points to pinpoint the precise phases during which these behaviors are altered.

Transient increases in anxiety-like behavior in postpartum mice may be associated with changes in CRFR1 reported in this study given the role of CRFR1 in regulating behavioral stress responses [8, 12, 20, 27, 52]. Specifically, global knockout and anterior forebrain/limbic brain structure knockout of CRFR1 have been shown to decrease anxiety-like behaviors and increase stress hormone release [8, 12]. Similar decreases in stress-related behaviors and stress hormone release are reported following administration of the CRFR1 antagonist, antalarmin [52]. However, it should be noted that CRF acting through CRFR1 can have differing effects on behavioral stress responses depending on brain region. For example, CRFR1 binding in the globus pallidus decreases anxiety-like behaviors [53] which is opposite to global effects of CRFR1 knockout and the role of CRFR1 in various other brain areas, which are associated with anxiogenic responses [8, 53, 54]. CRF and CRFR1 have also previously been shown to mediate anxiety-like behaviors in postpartum rats. For example, intracerebroventricular infusion of CRF increased anxiety-like behavior and a CRF receptor antagonist decreased anxiety-like behavior in lactating rats [55]. Furthermore, infusion of a CRFR1 agonist into select brain regions, including the MPOA, increases anxiety-like behaviors in postpartum rats [20]. This suggests that the transient increase in anxiety-like behavior reported here might be associated with changes in MPOA CRFR1 that occur during the postpartum period. However, the decrease in MPOA CRFR1 would be predicted to be stress reducing, based on aforementioned studies in rats. Furthermore, CRFR1 levels in the MPOA were suppressed at all postpartum time points with no notable fluctuation near the P7 time point. One change in CRFR1 that did coincide temporally with elevated anxiety-like behavior was a decrease in PVN CRFR1, which was most prominent at P7. CRFR1 in the PVN has been shown to regulate behavioral and neuroendocrine stress responses; thus, a decrease in CRFR1 may be a contributing factor [40, 56]. However, we also reported various other changes in CRFR1, including changes in co-expression of CRFR1 with dopaminergic and oxytocinergic neurons which might also contribute to postpartum changes in stress-related behaviors. It is quite possible that a combination of changes in the CRFR1 system, across a variety of brain regions and cell types, ultimately impact stress-related behaviors in postpartum mice. Furthermore, several other changes in stress-related genes have been reported during the postpartum period are also potential factors that modify postpartum changes in stress functions [57, 58]. Ultimately, it is not clear how these postpartum changes in CRFR1 are directly related to anxiety-like behavior changes that also occur during this period. Future studies that selectively knockdown or overexpress CRFR1 in the AVPV/PeN, MPOA, and PVN of postpartum mice are needed to clarify this relationship.

Elevated anxiety-like behavior (P7) in postpartum mice coincides with a period of estrogen withdrawal [59]. Rodent models that mimic this postpartum decline in 17β-estradiol show increases in anxiety-like and despair-like behaviors [60, 63]. High levels of prolactin and oxytocin are also present during this period of lactation, each of which have also been shown to mediate stress-related behaviors [64, 66]. Therefore, it will be interesting to determine in future studies whether these hormones might alter CRFR1 as a potential mechanism for regulating changes in stress-related behaviors in postpartum mice.

CORT levels assessed 90 min after exposure to the open field were elevated in P1 mice, although it is unclear whether elevated CORT levels are specifically due to the open field stress. High levels of CORT are present at the time of parturition [38] and CORT levels may still be elevated at P1 as a result of this surge. Furthermore, parturition-related stress or the initial demand for pup rearing may elevate CORT at P1. In our previous study, CORT levels were elevated in P14 relative to nulliparous mice exposed to a restraint stressor [26]. No differences were seen at P14 in the current study, which may be due to the open field stress not eliciting the same intensity of psychological stress as the restraint stress.

In this study, we mapped changes in CRFR1 and CRFR1 co-localization with oxytocin- and dopamine-expressing neurons across the postpartum period in mice. We report robust increases in CRFR1 in the AVPV/PeN and decreases in the MPOA that persist at least until 4 weeks after parturition. We also found transient decreases in CRFR1 in the PVN which occurred simultaneously to an increased expression of CRFR1 within oxytocin neurons in the PVN and SON. Coupling of CRFR1 and TH in the AVPV/PeN also showed a transient decrease in the mid-postpartum period. Several of the changes were temporally associated with increased anxiety-like behavior at P7. However, further studies are needed to determine which distinct changes in CRFR1 regulate changes in stress-related behaviors and potentially mediate how stress affects maternal behaviors. Future studies will also aim to determine the hormones and/or maternal experiences that initiate these changes in CRFR1. Together, this work will enhance our understanding of the neurobiology that mediates postpartum changes in behaviors associated with stress.

We gratefully acknowledge Timothy Quinn, Binoy Thomas, and Kristine Klein for expert animal care and James Dias and Barbara Weaver for assistance with and use of their equipment for the radioimmunoassay.

All procedures were conducted according to the National Institutes of Health Guide for the Care and Use of Animals. This research protocol was reviewed and approved by the Institutional Animal Care and Use Committee at University at Albany, approval number 19-004.

The authors have no conflicts of interest to declare.

This research was supported by University at Albany Research Initiation Funds (DZ), R15-MH118692 (DZ), R01-MH112768 (NJ), R01-NS110749 (KLZ), and U01-AG072464 (KLZ).

R.M.D., N.J.J., K.L.Z., and D.G.Z. designed research; R.M.D., Z.J.R., J.S.J., and D.G.Z. performed research; R.M.D., K.A.R., C.A.T., A.L.C., K.L.S., M.S., D.G.Z., K.E.P., and M.S.A. quantified and analyzed data; and R.M.D. and D.G.Z. wrote the manuscript, with approval of all authors.

Data produced in this study are included in this article. Data can also be obtained from the corresponding author upon request.