Postpartum changes in affect-related behavior and VTA dopamine neuron activity in rats

Abstract

The onset of motherhood is accompanied by alterations in emotional and affective behaviors. Many new mothers experience transient and mild depressive symptoms that typically resolve spontaneously (i.e. postpartum blues) but increase the risk for postpartum depression (PPD). There is little data regarding the neural adaptations occurring in response to parturition and shortly after birth that may be associated with these affective changes. Although the dopamine (DA) system is involved in affect, maternal motivation and PPD, little is known about postpartum DA function. We compared affective behavior in virgin and postpartum adult female rats at early and late time points. In vivo extracellular recordings of VTA DA neurons were performed to evaluate 3 parameters: number of active DA neurons (i.e. population activity), firing rate, and firing pattern. Compared with virgins, postpartum rats exhibited increased anxiety-like behavior in the elevated plus maze at 1-day postpartum; reduced social motivation at 1- and 3-days postpartum, reduced anxiety-like behavior in the novelty suppressed feeding test throughout the first week postpartum and increased forced swim test immobility at 1-day postpartum. 1- and 3-day postpartum females exhibited attenuated VTA population activity without changes in firing rate or pattern. None of these effects were observed in late postpartum females when compared with virgins. These data suggest that parturition induces time-dependent changes in a subset of affect-related behaviors and DA function during the postpartum period in rodents, with early postpartum females exhibiting depression-related phenotypes (i.e. low social motivation, higher immobility, blunted DA activity).

1. Introduction

The period after childbirth (i.e. postpartum period) is a time of elevated risk for the development of mood and affective disorders^1-3. The highest rates of anxiety and depression in women occur following childbirth⁴: as many as up to 75% of new mothers experience a transient mood disturbance (i.e. postpartum blues) 1-2 weeks postpartum^{2, 5,6}, and 10-15% go on to develop postpartum depression (PPD)^{3, 7}. Although postpartum blues is a common phenomenon and considered a normal physiologic response to the hormonal events from childbirth^{2, 6, 7}, it is a risk factor for PPD^{8, 9}, thereby underscoring the need to understand the normative neural changes involved in this transient state. Yet, there is little data regarding the neural adaptations in response to parturition (i.e. giving birth) and during the early postpartum, so the neural substrates that may be associated with this increased affective dysregulation remain unknown.

Alterations in maternal mood and affect have been partially attributed to rapid and robust changes in ovarian and stress hormones associated with the onset of motherhood in both humans and rodents^10-13, which can influence the activity of the mesolimbic dopamine (DA) system^14-16. The mesolimbic DA system originates in the ventral tegmental area (VTA) and plays a pivotal role in affective processes, maternal motivation as well as the pathophysiology of depression and PPD^{14, 16-19}. Importantly, mesolimbic DA dysregulation has been observed in women with PPD^{20, 21} and animal models of PPD^{14, 22}; and, preclinical studies in rodents have highlighted a causal link between DA dysregulation (i.e. decreased VTA DA neuron activity) and stress-induced depression-related behaviors (i.e. anhedonia, despair)^23-25. Moreover, early/mid postpartum females exhibit increased depressive-like behavior²⁶, and changes in reward-related-behaviors^{27, 28} compared with late postpartum rats, which may reflect time-dependent alterations in dopaminergic activity across the postpartum period. However, few studies have directly examined normative changes in behavior and VTA activity during the early postpartum, and whether a negative affect state occurs in early postpartum rodents is unknown. Thus, the primary purpose of this paper is to determine what changes in affect-related behaviors and VTA activity occur in the postpartum rat, with a focus on the early postpartum period.

We assessed parity-driven changes in affect-related behavior and DA activity by comparing virgin and postpartum female rats across distinct timepoints (early: 1-day, 3-days, 1-week; late: 21-23days). Animals underwent a behavioral test battery consisting of: elevated plus maze (EPM), three-chambered social approach test (SAT), novelty suppressed feeding (NSF), and the forced swim test (FST). In vivo single-unit extracellular recordings were conducted within the VTA of a subset of animals tested for behavior (EPM, SAT) to measure 3 parameters of DA neuron activity: number of spontaneously active DA neurons per electrode track (population activity), basal firing rate and firing pattern (percentage of spikes firing in bursts). We hypothesized that, like humans, early postpartum female rats would exhibit increased negative affect, although these effects may be limited to depression-related phenotypes and likely vary throughout the postpartum period. Moreover, given that negative affect behaviors are often associated with DA dysregulation (i.e. downregulation) in rodent models of depression^{25, 29}, we hypothesized that early postpartum rodents would exhibit alterations in VTA DA activity.

2. Methods and Materials

2.1. Animals

Virgin or timed-pregnant (gestational day 13) adult (and age-matched) female Sprague Dawley rats were shipped overnight (Envigo, Indianapolis, IN) and arrived in our facility (~6:00-7:00AM) the next day. Rats were housed in a temperature-controlled room on a 12-h light/dark cycle (lights on 7AM-7PM) with food (Lab Diet, Rodent Diet #5001) and water available ad libitum. The day each litter was born was designated as postpartum day 0. Dams that gave birth to small litters (< 6 pups) were not included in the study and sacrificed; dams used in the study had a minimum of 8 pups and a maximum of 12 pups. Virgin rats were co-housed in pairs and dams were co-housed with their litter. Animals were kept in their home cage and undisturbed except for weekly routine animal care done by the experimenter and when undergoing behavioral testing. All postpartum females, including late postpartum females, were kept with their litter to ensure all dams had exposure to offspring during the experiment and to avoid possible disruption due to the removal of offspring just before testing³⁰. The estrous cycle was not monitored in virgin rats because animals would have to be swabbed daily for a prolonged period of time and this would introduce a confound, as virgin females would be exposed to differential handling procedures and additional stress compared with postpartum dams, in which estrous cycle is not monitored since this reproductive stage (i.e. postpartum) is characterized by persistent diestrus and cessation of ovarian cycling ^{31, 32}. Although we are aware of literature indicating a decrease in anxiety-like behavior during proestrus and estrus in rats^{33, 34}, social approach/preference for same-sex conspecifics in female rats tested in the three-chamber social approach apparatus has been shown to be independent of the estrous cycle ³⁵. With regards to VTA DA neuron activity, a previous in vivo electrophysiological study conducted in female rats indicated no change in VTA population activity (i.e. the same measure used here) across the estrous cycle³⁶. All experiments were performed in accordance with the guidelines outlined in the National Institutes of Health Guide for Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh. For timeline and experimental design, please see Figure 1.

Figure 1. Timeline and experimental design.

A) Virgin or early postpartum female rats at different timepoints (1d-PP, 3d-PP, 1wk-PP) were tested for anxiety-like and social behavior in the elevated plus maze (EPM) and the three-chambered social approach test (SAT), respectively (n=9-16 per group). In vivo extracellular recordings of VTA DA neurons were conducted 1-day post-behavioral testing using an acute anesthetized preparation in the same cohort of rats used for behavior to evaluate: number of spontaneously active DA cells per track (i.e. population activity), firing rate and burst firing (n=8-12 per group). B) A separate cohort of virgin and early postpartum female rats was tested for spontaneous locomotor activity (SLA), novelty suppressed feeding (NSF) and in the forced swim test (FST) (n=6-9 per group). C)To determine whether any observed effects were specific to the early postpartum period, a separate cohort of animals consisting of virgins and late postpartum females (21-23d-PP), which were still kept with their litter, were tested in the EPM and SAT, followed by VTA recordings 1-2 days post-behavioral testing (n=8-12 per group).

2.2. Behavioral Testing

Behaviors were recorded and scored by an experimenter blind to experimental condition. All behaviors, except the social approach test, were conducted during the light cycle, similar to previously published work in early postpartum female rats^{37, 38} and social behavior studies conducted by our group^{39, 40}.

2.2.1. Elevated Plus Maze

The EPM was positioned 50cm above the floor and consisted of a plus-shaped apparatus composed of 2 opposite open arms (each 50cm long × 10cm wide) crossed at a right angle by 2 arms enclosed by 40cm high opaque walls⁴¹. Animals were habituated to the testing room within the home cage for approximately 1-hour prior to testing. Each rat was placed on the central platform and its movement was recorded for 5-minutes with a camera positioned overhead³⁹. The percentage of open arm entries (open arm entries/total × 100) and open arm time (time in open arms/total × 100), defined as front two paws and head in the arm, were used as indices of anxiety-like behavior. The total number of entries was used as an index of locomotor activity.

2.2.3. Social Approach Test

A three-chambered apparatus (total: 92cm × 45cm × 44.5 cm; side chambers: 36cm × 45cm ×44.5cm; center chamber: 20cm × 45cm × 44.5cm) was used to assess social approach (motivation), similar to previously described³⁹. Animals were habituated to the testing room within the home cage for approximately 1-hour prior to testing. Rats were placed in a smaller center chamber adjacent to two other chambers, each containing a wire cage that allows the test rat to see and smell its content but prevents physical interactions⁴². After a 5-min habituation period, an unfamiliar, younger (<200g), same-sex rat (to prevent aggression and/or mating behaviors)^{39, 43} that had previously been habituated to the wire cage was enclosed inside it and placed in a side chamber. Social stimulus animals were used a maximum of 3 times. An inanimate object (i.e. stuffed toy rat) was placed inside the other wire cage as a novel object control. The experimental rat was then allowed to explore the entire apparatus and time spent sniffing the receptacle containing the social stimulus was recorded for 10 minutes.

2.2.4. Spontaneous Locomotor Activity

Animals were placed into an open field arena (approx. 42cm × 42cm × 39cm; Coulbourn Instruments), and spontaneous locomotor activity was monitored (under red light) for 10-minutes by beam breaks with TruScan software and indexed as total distance traveled (cm) and total time spent moving (seconds)²⁹.

2.2.5. Novelty Suppressed Feeding

In this test, food-deprived animals are placed in a situation (i.e. brightly lit open field) that provokes conflict between the drive to eat and the fear of novel and open spaces⁴⁴. Following an 18-hour home cage food deprivation, rats were placed in a rectangular open field (70 × 40 × 30 cm) baited with a chow pellet during a 10-minute trial. Latency to begin eating was recorded. This measure is thought to reflect how the animal copes with a behavioral conflict (i.e. drive to eat vs fear of novel open spaces) and is used as an index of anxiety-like behavior^{44, 45}.

2.2.6. Forced Swim Test

The FST took place in a clear Plexiglas cylinder (50cm high, 20cm diameter) filled with water (25±1 °C) up to 38-40cm high to prevent escape or the animal’s tail touching the bottom, similar to that described previously²⁹. Animals received a 15-minute pre-exposure (day 1; habituation) followed by a 5-minute swim session the next day (day 2; test) in which immobility behavior, defined as passive floating in which the animal is making only minor necessary movements to maintain head above water⁴⁶, was measured²⁹. Water was changed between animals on each day. Rats were dried off after each session before being placed back in the home cage. While the interpretation of FST immobility has been subject to controversy over the last couple of years^{47, 48}, we included this test because it is still used as a standard for depression-related behavior and particularly for assaying antidepressant activity, its use is actually increasing in preclinical research, and the vast majority of researchers qualify the rodent’s floating response as depressive-like behavior, although there is an increasing trend for the interpretation of immobility as a passive coping strategy⁴⁹.

2.3. In vivo Electrophysiological Recordings

2.3.1. Surgery.

Single-unit extracellular recordings were performed using an acute preparation in a subset of rats the day after behavioral testing (EPM, SAT). Prior work from our group suggests no impact of these 2 tests on VTA DA neuron activity ^{40, 50}. Recordings were not conducted in animals that underwent the FST because this introduces a confound in VTA activity and is sufficient to induce stress-induced downregulation of VTA DA neuron activity²⁹. Rats were anesthetized with 8% chloral hydrate (400mg/kg, intraperitoneally), placed in a stereotaxic frame (Kopf, Tujunga, CA), and maintained at 37°C using a temperature-controlled heating pad (Fine Science Tools, Foster City, CA). Anesthesia level was monitored periodically by assessing the footpinch reflex and adjusted by intraperitoneal administration of additional chloral hydrate (0.3ml supplements) as needed. After clearing the skull of skin and fascia, a burr hole was drilled in the region overlying the VTA [from bregma, anteroposterior (AP): −5.4mm, mediolateral (ML): +0.6mm] on the right side of the brain.

2.3.2. VTA sampling.

Single barrel electrodes were constructed from 2mm diameter borosilicate capillary tubes (World Precision Instruments, Sarasota, Florida) using a vertical electrode puller (Narishige, Tokyo, Japan) and broken back under microscopic control. Glass electrodes were filled with 2% Chicago Sky Blue (Sigma-Aldrich, St. Louis, MO) dissolved in 2M saline and lowered into the VTA using a hydraulic microdrive (Kopf). DA neurons were sampled by making 6-9 vertical electrode passes (tracks), each separated by 0.2mm, in a predetermined pattern spanning the antero-posterior and medio-lateral extent of the VTA [AP: 5.4-5.7mm; ML: 0.6-1.0mm from bregma, and dorsoventral (DV): 6.5-9.0mm from dura]. This procedure has been used by our group to sample DA neurons with a variety of different projection targets⁵¹ in multiple studies^{52, 53}. Signal was acquired using a preamplifier (Dagan, Minneapolis, MN) and displayed on an oscilloscope (B&K Precision, Yorba Linda, CA) with a signal fed to a computer running Lab Chart 7 (AD Instruments, San Diego, CA).

2.3.3. Dopamine neuron identification.

DA neurons were identified with open filter settings (50Hz low cutoff, 16kHZ cutoff) using well-established electrophysiological criteria: location, slow, irregular firing pattern, long duration, variable shape biphasic action potential waveform (>2.2ms), half width (>1.1ms), temporary cessation of firing during tail/foot pinch^54-56. Once identified, DA neurons were recorded for 3 minutes (1-minute minimum) when signal to noise ratio exceeded 3:1. Three parameters of DA neuron firing were measured: number of spontaneously active DA cells per electrode track (population activity, all active DA neurons in each rat divided by the number of tracks)⁵⁷, basal firing rate, and proportion of spikes occurring in bursts, with burst initiation defined as the occurrence of two spikes with an interspike interval of ≤ 80ms and burst termination defined as the occurrence of an interspike interval of > 160 ms⁵⁸. This procedure allows us to measure both the tonic (i.e. population activity) and phasic (i.e. burst firing) state of encountered VTA DA neurons.

2.3.4. Placement verification.

Electrode placement was marked by electrophoretic ejection of Chicago Sky Blue dye at the final recording site. Rats were then overdosed with additional chloral hydrate, decapitated, their brains removed and fixed in 8% paraformaldehyde for at least 48 hours. Brains were then transferred to 25% sucrose solution for cryoprotection, sectioned using a cryostat (Cryostar NX50, ThermoScientific, Waltham, MA) into 60μm coronal slides, mounted on to gelatin-chromalum-coated glass slides, and stained with cresyl violet and neutral red to check recording electrode placements. Animals were required to have a minimum of 6 tracks within 0.4mm of target coordinates to be included in the study.

2.4. Statistical Analyses

For behavioral data, comparisons between 3 or more groups were analyzed using one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison post hoc test when appropriate. Pairwise comparisons were analyzed using unpaired two-tailed t-tests. Electrophysiological data was collected with Powerlab Lab Chart (AD Instruments) to identify spike time courses and exported to Neuroexplorer (NEX Technologies, NexTech Systems) software to calculate firing rate and burst firing for each DA neuron. Track location data were analyzed by repeated measures (RM) 2-way ANOVA. Normally distributed electrophysiological data were analyzed using one-way ANOVA when comparing 3 or more groups or two-tailed unpaired t-tests for pairwise comparisons. Data sets deviating from the normal distribution were analyzed with one-way ANOVA on ranks (Kruskal-Wallis H-test) followed by Dunn’s post hoc test as appropriate or Mann-Whitney U-test, respectively. All statistics were calculated using GraphPad Prism 7.0; differences were considered significant when p < 0.05.

3. Results

3.1. Changes in affect-related behaviors during the early postpartum period

Changes in anxiety-like behavior and social motivation during the early postpartum period were assessed by testing virgin (n=16) and 1-day (n=9-10), 3-day (n=9-10) or 1-week (n=11) postpartum rats in the EPM and SAT (Figure 1A). At 1-day postpartum, females made fewer open arm entries (one-way ANOVA: F_3,41=5.114, p<0.01) and spent less time in the open arm (one-way ANOVA: F_3,41=7.346, p<0.01) compared with virgin or 1-week postpartum rats (Figure 2A,B). 1d-PP and 3d-PP females made fewer total entries (one-way ANOVA: F_3,41=21.05, p<0.0001) compared with virgins (Tukey’s, p<0.0001) and 1-week PP females (Tukey’s, p<0.01) (Figure 2C). 1d-PP and 3d-PP females exhibited reduced social motivation, as indexed by reduced social sniff time (one-way ANOVA: F_3,47=4.462, p<0.01) and fewer crossings into the social chamber (one-way ANOVA: F_3,47=4.801, p<0.01) compared with virgins (Tukey’s, p<0.05) (Figure 2D,E). Changes in the total number of chamber crossings were also found (one-way ANOVA; F_3,47=23.24, p<0.0001) (Figure 2F). 1d-PP and 3d-PP females made fewer total crossings compared with virgins (Tukey’s; p<0.05). Differences were also found between 1d-PP and 3d-PP females (Tukey’s; p<0.001) and between 1d-PP and 1week-PP females (Tukey’s; p<0.0001), suggesting changes in locomotor activity throughout the first week postpartum.

Figure 2. Time-dependent changes in anxiety-like and social behavior during the first week postpartum.

A,B) Females tested at 1-day postpartum exhibited increased anxiety-like behavior in the EPM with (A) fewer open arm entries (p < 0.01) and (B) less time spent in the open arm (p < 0.05) compared with virgins or 1-week postpartum females. C) 1-day and 3-day postpartum females made fewer total entries compared with virgin and 1-week postpartum females (p < 0.05). D,E) Compared with virgins, 1- and 3-day postpartum females showed reduced social motivation, as indexed by D) lower levels of social cage sniff time (p < 0.05) and E) decreased number of crossings into the social chamber (p < 0.05). F) . F) 1-day postpartum females made fewer total number of crossings compared to all other groups (p < 0.001); 3-day postpartum females made fewer total chamber crossings than virgins (p<0.05) but more total chamber crossings than 1-day postpartum females (p <0.001). *p < 0.05, ** p < 0.01, *** p < 0.001, ****p <0.0001. Error bars represent mean ± SEM. Gray bars represent virgins (n=16) and red bars represent postpartum females (n=9-11 per group).

1d-PP and 3d-PP females exhibited reduced number of total entries in the EPM and total number of crossings in the SAT, suggesting altered locomotor activity. For this reason, we tested spontaneous locomotor activity as well as additional behavioral measures relevant to anxiety and depression-related phenotypes (i.e. NSF, FST) in a separate cohort of animals (n=9-16 per group; Figure 1B). No differences were found for total distance traveled (one-way ANOVA: F_3,36=1.184, p=0.329) or time spent moving (one-way ANOVA: F_3,36=0.610, p=0.613) between early postpartum (1d-PP, 3d-PP, 1week-PP) and virgin rats (Figure 3A,B) suggesting comparable baseline locomotor activity. Early postpartum females exhibited reduced latency to feed in a novel environment compared with virgins (one-way ANOVA: F_3,35=25.78, p<0.0001; Tukey’s; p<0.00001) (Figure 3C). An effect on FST immobility was detected in 1d-PP females (one-way ANOVA: F_3,35=6.364, p<0.05). Post-hoc analysis revealed increased immobility in 1d-PP females compared with virgins and 1-week PP females (Tukey’s; p<0.05) (Figure 3D). This cohort was used exclusively for behavior and did not undergo VTA recordings given our results showing that FST exposure downregulates VTA activity in female rats ²⁹.

Figure 3. Reduced latency to feed and enhanced FST immobility in early postpartum females.

A,B) Early postpartum rats had comparable spontaneous locomotor activity (SLA) to virgins, as they exhibited A) similar distance traveled (p=0.32) and B) time spent moving (p=0.61) compared with virgins throughout the first week postpartum. C) Early postpartum females exhibited reduced latency to consume food in a novel environment across the first week postpartum compared with virgins (p<0.0001). D) 1-day postpartum females showed increased immobility in the FST compared with virgins and 1wk-PP females. Error bars represent mean ± SEM. Gray bars represent virgins, red bars represent postpartum females (n=6-9 per group). *p < 0.05, ** p < 0.01, *** p < 0.001.

3.2. Attenuated VTA population activity during the first 3 days postpartum

Extracellular recordings of VTA DA neurons were conducted in the same animals subjected to behavioral testing (EPM, SAT) 1-day post-test (virgins: n=8, 43 neurons, 3-8 neurons per rat; 1d-PP : n=9, 28 neurons, 2-7 neurons per rat; 3d-PP: n=9, 33 neurons, 2-7 neurons per rat; 1-week PP: n=12; 64 neurons, 1-9 neurons per rat; Figure 1A). Compared with virgin rats (0.86 ± 0.24 CPT), postpartum females exhibited lower numbers of spontaneously active DA cells per track (i.e. population activity) (one-way ANOVA: F_3,34=5.86; p<0.01) at 1-day (0.41 ± 0.19 CPT; Tukey’s; p<0.01) and 3-days postpartum (0.52 ± 0.19 CPT; Tukey’s; p<0.05), but not at 1-week postpartum (0.69 ± 0.29 CPT) (Figure 3A). We also observed a trend towards a significant difference between 1d-PP and 1-week PP females (0.69 ± 0.29 CPT; Tukey’s; p=0.06). Given the evidence that the VTA is functionally segregated ^{51, 59}, population activity data for all groups were analyzed according to location in the medial, central, or lateral VTA. For a rat to be included in this analysis, data must have been available from all three VTA locations (i.e. medial, central, lateral). Population activity was selectively reduced in the reward-related medial VTA of postpartum females (n=6-11 rats per group) when compared to virgins (n=5) (RM 2-way ANOVA; F_3,26=4.73; p<0.01) (Figure 3C). No differences in firing rate (one-way ANOVA on ranks: H=0.33; p=0.95) were observed between virgin (3.90 ± 2.33 Hz) and postpartum females (1d-PP: 3.65 ± 2.17 Hz; 3d-PP: 3.55 ± 2.08 Hz; 1week-PP: 3.61 ± 2.28 Hz) (Figure 3D). No differences in the percentage of spikes fired in burst (one way ANOVA on ranks: H=6.61; p=0.08) were observed between virgin (29.22 ± 28.64 %SIB) and postpartum females (1d-PP: 13.69 ± 22.04 %SIB; 1d-PP: 20.13 ± 26.09 %SIB; 1week-PP: 21.16 ± 25.45 %SIB) (Figures 3E).

3.3. Comparable behavior and VTA DA neuron activity between virgins and late postpartum female rats

To determine whether similar behavioral and electrophysiological effects were observed in the late postpartum period, a separate cohort of rats were tested during postpartum days 21-23 (virgin: n=8; 55 neurons, 4-10 neurons per rat; late PP: n=8; 57 neurons, 4-10 neurons per rat; Figure 1C). No differences in anxiety-like behavior, as indexed by comparable percentages of open arm entries (t₁₅=1.33; p=0.20) and open arm time (t₁₅=1.25; p=0.23), were found (Figure 5A-C). No between group differences were found in the total number of arm entries (t₁₅=0.87; p=0.40). Similarly, no between group differences were found for social sniff time (t₁₈=0.73; p=0.48), crossings into the social chamber (t₁₈=0.70; p=0.49), or total number of chamber crossings (t₁₈=0.60; p=0.56) (Figure 5D-F). Virgin and late postpartum female rats exhibited comparable numbers of spontaneously active DA cells per electrode track (virgins: 0.95 ± 0.29 CPT, late postpartum: 1.14 ± 0.37 CPT) and were not significantly different from each other (t₁₄=1.17; p=0.26) (Figure 5G). No between group differences in firing rate (virgins: 3.99 ± 2.18 Hz, late postpartum: 3.50 ± 2.12 Hz; t₁₁₁=1.21; p=0.23) or burst firing (virgins: 17.96 ± 21.83 %SIB, late postpartum: 16.54 ± 21.59 %SIB; U=1565; p=0.23) were found (Figure 5H-I).

Figure 5. Comparable anxiety-like behavior, social behavior and VTA activity in virgin and late postpartum female rats.

A-C) No differences were found between virgin and late postpartum females in A) open arm entries (p=0.20), B) open arm time (p=0.23), or C) total number of entries in the EPM (p=0.40). D-F) No differences between virgin and late postpartum females in D) social sniff time (p=0.48), E) number of crossings into the social chamber (p=0.49) or F) total number of chamber crossings (p=0.56). G-I) Virgins and late postpartum females exhibit comparable VTA activity. No differences were found in G) population activity (p=0.26), H) firing rate (p=0.23) or I) burst firing (p=0.23). Error bars represent mean ± SEM. Gray bars represent virgins (behavior: n=8, recordings: n=8), pink bars represent late (i.e. 22-24 days) postpartum females (behavior: n=9-12 per group, recordings: n=8 per group).

4. Discussion

This is the first study to assess affect-related behaviors and the electrophysiological activity of VTA DA neurons during the early and late postpartum period in rodents. We identified time-dependent changes in a subset of affective behaviors that coexisted with an attenuation of VTA population activity, with effects being limited to early postpartum females. Late postpartum females and virgin rats exhibited comparable levels of anxiety-like behavior, social motivation and VTA activity. Collectively, these data suggest that normative changes in behavior and DA activity are limited to the first 3 days postpartum.

Behavioral changes during the early postpartum period in rodents

1-day postpartum female rats exhibited increased anxiety-like behavior (i.e. reduced percentages of open arm entries and open arm time) in the EPM compared to virgins and 1-week postpartum females, suggesting a temporary increase in anxiety-like behavior shortly after parturition. 1-day postpartum females also exhibited a reduction in the total number of arm entries, suggesting decreased locomotion in this group. These findings are in contrast with a prior study conducted in Long Evans rats indicating reduced anxiety-like behavior and increased total arm entries at 1-day postpartum³⁷. Methodological differences including testing room habituation periods, illumination conditions, estrous cycle stage, composition of control groups, and test duration could have contributed to this discrepancy⁶⁰. For example, our virgin control group was not stratified by estrous cycle and this may have contributed to our result, since anxiety-like behavior in the EPM varies throughout the estrous cycle^{33, 34}.

A separate cohort of rats was tested for spontaneous locomotor activity (SLA) at baseline and in a different assay of anxiety-like behavior also based on approach-avoidance conflict (i.e. NSF). Early (i.e. 1d, 3d and 1-week) postpartum females and virgins did not differ in terms of total distance traveled and total time spent moving, indicating comparable baseline locomotor activity. This may suggest that the reduction in total arm entries in the EPM and SAT reflects a test and/or apparatus-dependent change in behavior. Early postpartum females tested in the NSF exhibited decreased latency to feed in a novel environment compared with virgin animals. These data are consistent with a prior study indicating that 1-day postpartum females consume more food in a novel environment than virgins⁶¹ and that early lactating females typically show a large increase in food consumption^{62, 63}. Thus, we acknowledge that there are significant differences between females and early postpartum rats in food intake (i.e. increased food intake in early postpartum) and energy requirements (i.e. decreased energy expenditure in early postpartum)^{62, 64}, which are associated with lactation and maternal behavior, that may contribute to the observed effects in the NSF. However, we also acknowledge that this is a confound that will be present in any study comparing virgin and postpartum rodents given that these differences are characteristic of the postpartum period. The finding of reduced latency to feed in the NSF is likely influenced by between group differences in food intake and energy expenditure and may reflect increased hunger due to metabolic needs rather than reduced anxiety-like behavior. Collectively, these data suggest alterations in anxiety-like behavior during the first week postpartum, although this effect may depend on the assay used and be influenced by task-dependent changes in locomotor activity.

Most studies assessing social behavior in postpartum female rodents have focused on aggression^{65, 66}, sexual behavior^{67, 68}, and maternal behaviors directed at pups⁶⁹. However, there are many distinct aspects of social behavior and little is known about social approach/motivation to a novel conspecific in postpartum rats. We identified changes in social motivation to a novel conspecific during the first 3 days postpartum: 1d-PP and 3d-PP dams exhibited reduced social cage sniff time and fewer crossings into the chamber containing the younger female rat (i.e. social stimulus animal) compared with virgins, suggesting attenuated social motivation. Notably, reductions in social behavior are also observed following acute or chronic stress exposure and in rodent models relevant to depressive-like phenotypes, including females⁷⁰, in which changes in glucocorticoid levels trigger social dysregulation^{42, 71}. In the current study, our finding of reduced social motivation to a novel conspecific in 1d-PP and 3d-PP females is consistent with studies indicating elevated levels of total basal corticosterone (CORT) on postpartum day 1 and 3 in Sprague Dawley rats^72-74, which may suggest that increased basal CORT levels contribute to the social behavior dysregulation observed in early postpartum females. Alternatively, reduced social approach in 1d-PP and 3d-PP dams may reflect an adaptive shift in social salience in favor of pup-related stimuli during the early stages of motherhood, as pup interaction is one of the most highly motivating behaviors in maternal mammals and can induce conditioned place preference (CPP) in postpartum rats^{28, 75}. Since this is the first study to compare social approach in virgin and postpartum female rats, this remains to be more thoroughly explored. Nonetheless, 1d-PP and 3d-PP rats exhibiting decreased social motivation to a novel conspecific also had lower numbers of active VTA DA cells. This is significant given the role of VTA DA neurons in modulating social behavior: chemogenetic inhibition of VTA DA neurons attenuates social exploration of a nonfamiliar conspecific (i.e. social stimulus animal) in mice⁷⁶ and stimulation of VTA DA neurons increases social approach/motivation in female mice⁷⁷. Thus, it seems likely that attenuated social motivation in 1d-PP and 3d-PP reflects reduced dopaminergic tone within the VTA.

Finally, 1d-PP females exhibited increased FST immobility compared with virgins and 1week-PP rats. This is consistent with a prior report indicating no difference between virgins and 1-week postpartum females⁷⁸, but extends it by identifying increased FST immobility at 1-day postpartum. Although this measure has been recently proposed to reflect a switch from active to passive coping in the face of an acute stressor^{47, 48}, increased immobility is also a common outcome in rodent models relevant to depression^{24, 71}, including females ²⁹. Optogenetic activation of VTA DA neurons reverses increased immobility duration, increased kick frequency and promoted escape-related behaviors in stressed rats²³, indicating a causal relationship between VTA DA neuron activity and FST immobility behavior regardless of whether this behavior reflects a switch in coping strategy or behavioral despair.

In sum, early postpartum females exhibit a transient behavioral dysregulation limited to a subset of tests (i.e. SAT, FST) during the first 3 days postpartum that overlaps with outcomes in rodent models of depression associated with alterations in DA function (i.e. attenuated VTA population activity). This is consistent with human studies indicating increased depressive-like symptomatology and poor social adjustment during late pregnancy and the early postpartum compared with non-childbearing women⁷⁹.

Changes in VTA activity during the early postpartum period in rodents

1-day and 3-day postpartum females exhibited DA downregulation, as demonstrated by reduced numbers of spontaneously active VTA DA neurons (i.e. attenuated population activity) compared with virgin females, which was more pronounced in reward-related medial aspects of the VTA. These findings overlap with our prior reports of stress effects on the DA system in animal models of depression (i.e. learned helplessness, chronic mild stress)^{24, 53}, in which DA downregulation (i.e. attenuation of VTA population activity) is observed²⁵, with females showing greater effects²⁹. The maternal HPA-axis undergoes dramatic changes during pregnancy and the postpartum^{12, 13, 80, 81}. In rats, CORT levels peak during parturition and slowly fall over the postpartum period so that the early postpartum period is characterized by sustained high flattened levels of glucocorticoids^{11, 72, 82}, which is a similar hormonal profile observed in depressed patients and chronically stressed female rats^{80, 83}. Thus, in the context of our current study, elevated basal levels of CORT at 1 and 3 days postpartum^72-74 may contribute to the attenuation in VTA population activity observed in 1-day and 3-day postpartum females.

Importantly, DA neurons must be spontaneously active to respond to behaviorally relevant, phasic inputs^{84, 85}. Therefore, a change in population activity, or the number of spontaneously active DA cells, is thought to the reflect the responsivity of the DA system, which is altered in several disease states⁸⁶. Our finding of reduced DA population activity specifically within the medial tracks of the VTA is consistent with prior studies by our group showing a preferential reduction of population activity within the medial and central VTA regions in rats exposed to chronic mild stress or learned helplessness^{52, 53}. Because the majority of the DA neurons in medial VTA project to the ventromedial accumbens, reducing the active number would be expected to decrease the response of ventromedial accumbens projecting DA neurons to reward-related stimuli⁸⁶, as only spontaneously active DA neurons are capable of responding to signals with burst firing⁸⁷. Thus, a decrease in DA neuron population activity would attenuate stimulus-driven DA neuron responses, leading to diminished activation of the system. Prior reports have shown reduced basal DA release in the accumbens of 1d-PP rats compared with virgin rats⁸⁸, which is consistent with reduced tonic activity of DA neurons within the medial VTA. Moreover, accumbens DA release is diminished in response to food in 1d-PP dams, but robustly stimulated by pup exposure⁸⁸, suggesting a shift in sensitivity of phasic DA system responses towards offspring-related stimuli and away from stimuli not related to the offspring. This suggests that, in addition to changes in tonic VTA activity, we would likely find time-dependent changes in VTA activity in response to pup exposure, although this is beyond the scope of the current study. Since this is our group’s initial step in accounting for sex and reproductive status by incorporating postpartum females into our work, at this point we cannot determine whether the changes in VTA activity are driving the behavioral changes (i.e. SAT, FST) observed- only that the DA changes observed here (i.e. attenuation of population activity within the medial track of the VTA) are consistent with those we have previously seen in animal models of depression^{25, 29, 86}. In addition, although both virgin and timed-pregnant rats were transported to our facility under the same conditions, it is possible that pregnant females may have experienced this stressor differently than virgin animals, which could be an indicator of increased stressed susceptibility.

Lastly, our results suggest that alterations in a subset of negative affect behaviors and DA activity in postpartum females may be limited to the first 3 days postpartum. Late postpartum females exhibited levels of anxiety-like and social behavior comparable to virgins, and no changes were observed in any parameter of VTA function assessed (population activity, firing rate, burst firing).

5. Conclusion

Taken together, these findings suggest normative changes in a subset of affect-related behavior and VTA activity that are confined to the first 3 days postpartum. Our findings of increased depression-related phenotypes and attenuated VTA DA activity in early postpartum female rats suggest that parity and the onset of motherhood comes with an unexpected cost involving negative (albeit transient) effects on DA system function.

Figure 4. Reduced VTA DA population activity during the first 3 days postpartum.

A) 1- and 3-day postpartum females exhibited a reduction in the number of spontaneously active DA cells in the VTA (i.e. population activity) compared with virgins (p < 0.05). B) Early postpartum females exhibited a selective attenuation of DA neuron activity in the medial aspect of the VTA (p <0.01). C,D) No changes between virgins and early postpartum females were observed in C) firing rate (p=0.95) or D) percentage of spikes firing in bursts (p=0.08). *p < 0.05, ** p < 0.01, #p=0.06, Error bars represent mean ± SEM. Gray bars represent virgins (n=8) and red bars represent postpartum females (n=9-12 per group).

Highlights:

• 1- and 3-day postpartum females exhibit reduced social approach compared to virgins

• 1-day postpartum females exhibit task-dependent changes in anxiety-like behavior

• 1-day postpartum females exhibit increased FST immobility compared to virgins

• 1- and 3-day postpartum females exhibit reduced VTA DA activity compared to virgins