Oxytocin Receptor-Expressing Neurons in the Medial Preoptic Area Are Essential for Lactation, whereas Those in the Lateral Septum Are Not Critical for Maternal Behavior

Abstract

Introduction

In nurturing systems, the oxytocin (Oxt)-oxytocin receptor (Oxtr) system is important for parturition, and essential for lactation and parental behavior. Among the nerve nuclei that express Oxtr, the lateral septal nucleus (LS) and medial preoptic area (MPOA) are representative regions that control maternal behavior.

Methods

We investigated the role of Oxtr- and Oxtr-expressing neurons, located in the LS and MPOA, in regulating maternal behavior by regulating Oxtr expression in a region-specific manner using recombinant mice and adeno-associated viruses. We quantified the prolactin (Prl) concentrations in the pituitary gland and plasma when Oxtr expression in the MPOA was reduced.

Results

The endogenous Oxtr gene in the neurons of the LS did not seem to play an essential role in maternal behavior. Conversely, decreased Oxtr expression in the MPOA increased the frequency of pups being left outside the nest and reduced their survival rate. Deletion of Oxtr in MPOA neurons prevented elevation of Prl levels in plasma and pituitary at postpartum day 2.

Discussion/Conclusion

Oxtr-expressing neurons in the MPOA are involved in the postpartum production of Prl. We confirmed the essential functions of Oxtr-expressing neurons and the Oxtr gene itself in the MPOA for the sustainability of maternal behavior, which involved Oxtr-dependent induction of Prl.

Introduction

Parenting is essential for the survival of mammalian species, and its disruption leads to societal problems in humans. Maternal behavior is established by plasticity in the central nervous system (CNS), which is induced by steroids, peptide hormones, and their receptors. Changes from the virgin to maternal state are accompanied by dramatic alterations in hormone and neuromodulator concentraions and the expression of their receptors in the CNS, including oxytocin (Oxt) [1], estrogen [1], prolactin (Prl) [2], progesterone [1], dopamine [3], norepinephrine [4], serotonin [5], galanin [6], cortisol [7], and γ-aminobutyric acid (GABA) [8]. Moreover, hormone level changes during pregnancy and lactation may coordinate with the exertion of maternal behavior [9]. Estrogen, progesterone, and Prl levels continuously increase throughout pregnancy. Peripherally at delivery onset, estrogen and progesterone levels rapidly decrease, and a surge in Oxt levels occurs, initiating uterine contractions. During the postpartum period, Prl pulses stimulate milk production, whereas Oxt pulses lead to milk ejection in response to the physical suckling stimulus from the newborn offspring [10]. The levels of Oxt and Prl, as well as those of their respective receptors, also change dramatically in the brain in response to olfactory, visual, and auditory stimulation by the offspring during parenting [11–13]. Oxt administration induces maternal behavior in virgin female rats [14], and Prl also plays a role in the stimulation of maternal behavior in this species [15]. It has been reported that the effect of Prl secreted by the mother during the perinatal period in the expression of nurturing behaviors is mediated by the medial preoptic area (MPOA) [16, 17]. Some reports suggest that the secretion of Prl is regulated by Oxt [18, 19] and that Prl regulates Oxt secretion during lactation [20]. Gorden found that both Prl and Oxt were related to patterns of paternal care in human fathers [21]. These papers suggest that both Prl and Oxt are involved in parental behavior, but more research is needed to fully understand their interactions.

Oxt is a nonapeptide hormone produced predominantly in the paraventricular nucleus (PVN) and the supraoptic nucleus of the hypothalamus. In previous studies using Oxt^−/− mice, many research groups, including ours, showed that Oxt has a pleiotropic function in mammalian reproduction and plays an important role in milk ejection during nursing [22] and maternal behavior [23]. Oxt, which is known for its neuromodulatory and neurotransmitter actions [24], plays essential roles in peripheral organs for milk ejection and delivery, whereas central Oxt plays a complicated role in various types of social behaviors, including maternal behavior. A recent study reported that alloparenting could be promoted in nulliparous mice by activation of Oxt-expressing neurons in the PVN through observation of maternal behavior by dams [25].

Oxt exerts various physiological functions via the Oxt receptor (Oxtr), which is expressed in many peripheral tissues and in the brain [26, 27]. Oxtr-expressing neurons are abundant in brain regions that are important for maternal behavior [28], and Oxtr−/− mice exhibit severe disabilities in milk ejection and moderate impairments in maternal behavior [29]. Many reports have suggested the importance of the Oxt/Oxtr system in the regulation of maternal behavior. Mice lacking Cd38, whose genetic defects were presumed to have the closest relationship with those of the Oxt/Oxtr system, showed reduced Oxt secretion coupled with impairments in maternal behavior [30].

The lateral septal nucleus (LS) and the MPOA are nerve nuclei with a high density of Oxtr-expressing neurons [28] and are known to modulate maternal behavior [31–33]. In a radiolabeled ligand binding study, the LS and MPOA regions of high-licking/grooming (LG) rat mothers showed significantly higher densities of binding capability to an ¹²⁵I-labeled Oxtr antagonist than those of low-LG controls [34]. Oxt neurons in the PVN project to several forebrain nuclei, including the LS and MPOA [35]. The overlaps of altered gene expression in the postpartum CNS were particularly higher in the MPOA and LS [36]. Studies have implied that the LS is a critical region for controlling maternal behavior [37, 38], including the regulation of motivation and mood processes [39]. The LS is interconnected with many brain regions that affect emotion, including the hippocampus, amygdala, hypothalamus, and ventral striatum [39]. More than 90% of the neurons in the LS are GABA-positive [40], and our group and Menon et al. have reported that the Oxtr neurons in the LS are also mostly GABAergic [41, 42]. MPOA has long been known as a key regulatory region in maternal behavior [43]. Lesions in the MPOA disrupt parenting, whereas hormonal stimulation of the MPOA activates it [44]. Prl receptors (Prlrs), Oxtrs, estrogen receptors, and many other receptors for neuropeptide hormones are expressed in the MPOA, which makes it difficult to understand the mechanism controlling maternal behavior in this region. Therefore, in this study, we investigated whether Oxtr neurons, or the Oxtr itself, in the LS and MPOA were essential for maternal behavior during nurturing.

Materials and Methods

Maintenance of Mice

OxtrCre (Oxtr^Cre/+) [45], Oxtrfx (Oxtr^fx/fx) [46], Oxtr^V/+ (Oxtr^Venus/+) [20], Oxtr−/− [29], and WT (wild type) mice were maintained on a C57BL/6J genetic background. The mice were allowed ad libitum access to standard chow (CE-2; CLEA, Shizuoka, Japan) and water and were kept at 25°C in a room with a 12-h light/dark cycle (07:00–19:00). All experimental procedures and animal care were conducted in accordance with the relevant guidelines and regulations and were approved by the Fukushima Medical University Institute of Animal Care and Use Committee.

Preparation of Nurturing and Diestrus Mice

Nurturing mice were prepared as follows: adult virgin female mice were housed with male mice and checked daily for vaginal plugs. When the female exhibited a vaginal plug, the plug day was designated as d0.5 of pregnancy. At d12.5, the pregnant mice were singly housed until the parturition day (d19.5). After parturition, the pups were picked up gently when the dam was out of the nest. The details of tissue collection are listed in Tables 1 and 2. The genotype, manipulation, and number of mice used in the maternal behavior tests are shown in Table 3. As a control group, females in the diestrus stage were identified from a pool of virgin mice by a vaginal smear test, as described previously [47], and subjected to tissue collection (Tables 1, 2).

Table 1.

Sampling design with genotype, condition, and numbers

Figure	Abbreviation	Physiological condition of mice	Genotype	AAV using infection	Infected nucleus	Time point of sampling	Housing condition	Sampling numbers
Figure 1a–d	D	Virgin	Oxtrvenus/+(Oxtr-Venus knock-in)			Diestrus day	Housed with cage mates about 2–4	6
	N	Nurturing mother	Oxtrvenus/+(Oxtr-Venus knock-in)			The first day after parturition	Singly housed on pregnant day 12.5 until parturition day	6
Figure 2a, b	D WT	Virgin	WT			2–3 p.m. on the diestrus day	Housed with cage mates about 2–4	Figure 2a: 3, Figure 2b: 4
	N WT	Nurturing mother	WT			2–3 p.m. on the next day of parturition	Singly housed on pregnant day 12.5 until parturition day
	D Oxtr−/−	Virgin	Oxtr−/−			2–3 p.m. on the diestrus day	Housed with cage mates about 2–4
	N Oxtr−/−	Nurturing mother	Oxtr−/−			2–3 p.m. on the next day of parturition	Singly housed on pregnant day 12.5 until parturition day
	D OxtrCre+dtA/MPOA	Virgin	OxtrOxtrcDNA−IRES-Cre/+ (Oxtr-Cre knock - in)	AAV-mCherry-FLEX-dtA	MPOA	2–3 p.m. on the diestrus day	Housed with cage mates about 2–4
	N OxtrCre+dtA/MPOA	Nurturing mother	OxtrOxtrcDNA−IRES-Cre/+ (Oxtr-Cre knock - in)	AAV-mCherry-FLEX-dtA	MPOA	2–3 p.m. on the next day of parturition	Singly housed on pregnant day 12.5 until parturition day
	D Oxtrfx+Cre/MPOA	Virgin	Oxtrflox/flox	AAV-Cre	MPOA	2–3 p.m. on the diestrus day	Housed with cage mates about 2–4
	N Oxtr+fxCre/MPOA	Nurturing mother	Oxtrflox/flox	AAV-Cre	MPOA	2–3 p.m. on the next day of parturition	Singly housed on pregnant day 12.5 until parturition day
Figure 2c	Diestrus	Virgin	Oxtrvenus/+(Oxtr-Venus knock-in)			Diestrus day	Housed with cage mates about 2–4	9
	Nurturing	Nurturing mother	Oxtrvenus/+(Oxtr-Venus knock-in)			The first day after parturition	Singly housed on pregnant day 12.5 until parturition day	9
Online supplementary Figure 3F	WT/Diestrus	Virgin	WT			Diestrus day	Housed with cage mates about 2–4	2
	WT/Nurturing	Nurturing mother	WT			The next day of parturition	Singly housed on pregnant day 12.5 until parturition day	2
	Oxtr−/−/Nurturing	Nurturing mother	Oxtr−/−			The next day of parturition	Singly housed on pregnant day 12.5 until parturition day	2
	Oxtr-Cre+dtA/MPOA Nurturing	Nurturing mother	OxtrOxtrcDNA−IRES-Cre/+ (Oxtr-Cre knock - in)	AAV-mCherry-FLEX-dtA	MPOA	The next day of parturition	Singly housed on pregnant day 12.5 until parturition day	2
Online supplementary Figure 3G	Diestrus	Virgin	Oxtrvenus/+(Oxtr-Venus knock-in)			Diestrus day	Housed with cage mates about 2–4	3
	Nurturing	Nurturing mother	Oxtrvenus/+(Oxtr-Venus knock-in)			The first day after parturition	Singly housed on pregnant day 12.5 until parturition day	3

For all online suppl. material, see https://doi.org/10.1159/000535362.

Table 2.

Genotypes and sampling condition of virgin mice used in behavior test and immunohistochemical analysis with nurturing mice as control

Figure	Abbreviation	Genotype	Physical condition of mice	Age, weeks	Sampling time point	Housing condition	Sample numbers
Figure 1f	Oxtr−/−	Oxtr −/−	Virgin	9	The day after singly housing	Singly housed for 1 week	9
	WT	WT	Virgin	9			9
Figure 1g, h	Pup stimulation − WT	WT	Virgin	9	The day after singly housing	Singly housed for 1 week	4
	Pup stimulation − Oxtr−/−	Oxtr −/−	Virgin	9			4
	Pup stimulation + WT	WT	Virgin	9	Pup stimulation: the day after singly housing Perfusion: 90 min after pup stimulation		4
	Pup stimulation + Oxtr−/−	Oxtr −/−	Virgin	9			4
Online supplementary Figure 1	WT Diestrus	WT	Virgin	13–20	Diestrus day	Singly housed for 1 week	6
	WT nurturing	WT	Nurturing mother	13–20	The first day after parturition	Singly housed from pregnant day 12.5 until parturition day	6
	Oxtr−/− Diestrus	Oxtr −/−	Virgin	13–20	Diestrus day	Singly housed for 1 week	6
	Oxtr−/− Nurturing	Oxtr −/−	Nurturing mother	13–20	The first day after parturition	Singly housed from pregnant day 12.5 until parturition day	6

Table 3.

Genotype, manipulation, and numbers of mice used in maternal behavior test

Figure	Abbreviation	Genotype	Sample numbers	AAV using infection	Infected nucleus	Genotype of control (sample numbers)
Fig. 3d–g	OxtrCre + dtA/LS	OxtrOxtrcDNA−IRES-Cre/+ (Oxtr-Cre knock - in)	Fig. 3d–f (13) Fig. 3g (11)	AAV-mCherry-FLEX-dtA	LS	OxtrCre OxtrOxtrcDNA−IRES-Cre/+ (Oxtr-Cre knock - in) Fig. 3d–k (6)
	WT + dtA/LS	WT	Fig. 3d (9) Fig. 3e–g (10)
Fig. 3h–k	OxtrCre + dtA/MPOA	OxtrOxtrcDNA−IRES-Cre/+ (Oxtr-Cre knock - in)	Fig. 3h–k (10)	AAV-mCherry-FLEX-dtA	MPOA
	WT + dtA/MPOA	WT	Fig. 3h–k (9)
Fig. 4c–f	Oxtrfx + LacZ/LS	Oxtrflox/flox	Fig. 4c, d (8) Fig. 4e (7) Fig. 4f (9)	AAV-LacZ	LS	Oxtrfx Oxtrflox/flox Fig. 4c, e, f, g, i, j (9) Fig. 4d, h (10)
	Oxtrfx + Cre/LS	Oxtrflox/flox	Fig. 4c, d (8) Fig. 4e (7) Fig. 4f (12)	AAV-GFPCre
Fig. 4g–j	Oxtrfx + LacZ/MPOA	Oxtrflox/flox	Fig. 4g–i (9) Fig. 4j (7)	AAV-LacZ	MPOA
	Oxtrfx + Cre/MPOA	Oxtrflox/flox	Fig. 4g (12) Fig. 4h–j (8)	AAV-GFPCre
Online suppl. Fig. 3A–D	Oxtr−/−	Oxtr−/-	Online suppl. Fig. 3A–D (11)	–	LS	WT wild type Online suppl. Fig. 3A–D (11)
	Oxtr−/− + Oxtr/LS	Oxtr−/-	Online suppl. Fig. 3A–D (11)	AAV-Oxtr

Pup Stimulation of Virgin Mice

Virgin wild type (WT) and Oxtr−/− female mice (9 weeks old) were exposed to pups to observe the effect of the stimulation. The mice were single-housed for 1 week before stimulation. The test mice were then removed from their home cage, and three pups (1–3 days old) that had been delivered by a WT dam were individually placed in different corners of the cage, distant from the nest. The subject was then returned to the home cage, and its behavior was recorded for 5 min. Variables recorded included sniffing latency for each pup, time spent licking the pups, and number of pups retrieved to the nest. Ninety minutes after the behavioral stimulation, each subject was perfused with 4% paraformaldehyde under anesthesia. Non-stimulated mice were single-housed for 1 week and perfused in a similar manner (Table 2).

Measurement of Pup Body and Stomach Weight

9-11 week-old nulliparous OxtrCre and wild-type female mice that had been administered AAV-mCherry-flex-dtA in the MPOA (OxtrCre + dtA/MPOA, WT + dtA/MPOA) and OxtrCre were housed with WT C57BL6 male mice. After confirmation of the presence of a vaginal plug in the female, the male was removed from the cage. Six pregnant mice were divided into two groups of three animals each. The day of parturition was recorded as day 1 (D1). On the day of birth, the number of pups was adjusted to five pups per individual mother. The body weight of live pups was measured on D1. On the next day (D2), the pups from one of the groups were sacrificed by decapitation after body weight measurement, and their stomachs were removed and weighted. The body weight and stomach weight of the pups from the remaining group were measured on D3 (2 days after D1). The measurements were performed in all cases between 9:00 and 10:00 a.m.

Injection of Adeno-Associated Virus (AAV) Vectors

Viral vectors were injected bilaterally into the LS (bregma, +0.5 mm; lateral, +0.3 mm; depth, +2.75 mm) or MPOA (bregma, +0.26 mm; lateral, +0.3 mm; depth, +5.25 mm). The injection volume was 1.0 µL per side. Recombinant AAV serotype2 carrying mCherry-flex-dtA (AAV-mCherry-flex-dtA; Addgene, MA, USA), Oxtr-IRES-Venus (AAV-Oxtr), LacZ (AAV-LacZ), or GFP-Cre (AAV-GFP-Cre; Addgene, MA, USA) were used. AAV-mCherry-flex-dtA was selected as a tool to eliminate Oxtr-expressing neurons with minimal risk of Cre-independent dtA expression, which may lead to inaccurate results [48].

Recombinant AAV-Oxtr and AAV-LacZ were prepared as described in our previous report [49]. The recombinant vectors were provided by Dr. Mizukami at Jichi Medical University. All injections were administered at a rate of 100 nL/min. Genotype of mice, manipulation of AAV, and sample numbers used in maternal behavior tests are shown in Tables 1 and 3. One microliter of the AAV vector (3.0 × 10⁹ virus genomes per µL) was injected using a stereotaxic injector (Narishige Co., Tokyo, Japan) into adult mice aged 8–10 weeks anesthetized using a mixture of three types of anesthetic agents (5 mL/kg): an α2 adrenaline receptor agonist (medetomidine 0.03 mg/mL), a GABAα receptor agonist (midazolam 0.4 mg/mL), and an opioid κ receptor agonist (butorphanol 0.5 mg/mL). The needles used for injection were shaped from a glass capillary (G1; Narishige Co.) using a puller (PC10; Narishige Co.) to have a tip length of 1.8 mm and an outer diameter of 0.005 mm.

Maternal Behavior Test

Maternal behavior was tested in postpartum mice (13–20 weeks old), as described previously [29]. On the morning of the day of parturition, each mother was observed with minimal disturbance, and the percentage of scattered newborns was recorded. The pups were then removed, and after 1 h, each female was exposed to three 1–3-day-old foster pups from a WT dam, which were placed individually in different corners of the cage distant from dislocation the female’s nest. Since pups delivered by genetically manipulated mothers may become too weak due to hypothermia, low milk intake, or other reasons, we only used pups from healthy WT donors in the maternal behavior tests, in order to make the experimental conditions equal among the groups. During the next 30 min, each female was continuously observed, and the number of retrieved pups as well as the duration of crouching over all three pups in the nest was recorded. After the test, two newborn vigorous pups were added to the test cage, and the survival rate of the pups after 5 days was recorded. The group of OxtrCre and Oxtrfx mice not subjected to virus injection was used as a control in the experiments with LS-injected and MPOA-injected mice.

Immunohistochemistry

For c-fos immunohistochemical analyses of virgin mice stimulated with pups, virgin WT and Oxtr−/− female mice (9 weeks old) were singly housed for 1 week before stimulation and exposed to WT pups for 5 min. Ninety minutes after pup stimulation, the mice were perfused with 4% paraformaldehyde under anesthesia. Non-stimulated mice were singly housed for 1 week and perfused in a similar manner. Brain tissues were post-fixed for 24 h at 4°C. After post-fixation, the brains were transferred in 30% sucrose for 48 h at 4°C and stored at −30°C before embedding. Then, 40-μm-thick brain sections were cut using a cryostat (Leica, Wetzlar, Germany). Sections at 120-μm intervals +0.6 mm from the bregma for the LS and +0.2 mm from the bregma for MPOA were used for c-fos immunostaining. The sections were rinsed in phosphate-buffered saline (PBS: 0.01 M, pH 7.4) and incubated in 0.3% H₂O₂ for 15 min, which was repeated twice. After rinsing, the sections were incubated with PBS containing 0.3% Triton-X (T-PBS) for 30 min and incubated with T-PBS containing 2% normal goat serum and 2% bovine serum albumin. The sections were then incubated with PBS containing rabbit anti-c-fos antibody (#RPCA-c-fos AP; Encor Biotechnology Inc., FL. USA) at a 1:1,000 dilution for 18 h at 4°C for c-Fos immunohistochemical analysis. After rinsing in PBS, all sections were incubated for 40 min with biotinylated goat anti-rabbit IgG (BA-1000-1.5; Vector Laboratories Inc., CA, USA) diluted to 1:500. Next, the sections were incubated with an avidin-biotin complex solution (VECTASTAIN Elute ABC kit, PK-6100; Vector Laboratories Inc., CA. USA) for 1 h. Immunoreactivity was visualized by incubation in nickel-diaminobenzidine (DAB; 349-00903; Dojin Laboratories, Kumamoto Japan) solution (0.3% nickel ammonium sulfate, 0.02% DAB, and 0.0045% H₂O₂ in Tris-HCl buffer [0.05 M, pH 7.4]). Images were captured using a microscope (Ts2R; Nikon, Tokyo, Japan).

To visualize the distribution of Oxtr, brain sections from Oxtr^V/+ [28] mice were permeabilized with 0.5% Triton-X/PBS for 30 min and then blocked with 5% normal goat serum/T-PBS for 30 min. After blocking, the sections were incubated overnight at 4°C in chicken polyclonal anti-GFP antibody/PBS (1:1,000; ab13970; Abcam, Cambridge, UK) as the primary antibody. The sections were then washed three times with PBS and incubated with Alexa Fluor-488-conjugated goat anti-chicken antibody/PBS (1:1,000; A11039; Life Technologies, MA, USA) as the secondary antibody for 2 h at room temperature. To detect the distribution of Prl, rabbit monoclonal anti-Prl antibody (1:200; ab183967; Abcam, Cambridge, GB) and Alexa Fluor-594-conjugated donkey anti-rabbit IgG antibody (1:1,000 dilution; A21207; Invitrogen, MA, USA) were used as the primary and secondary antibodies, respectively. To detect the distribution of Prlr, rabbit polyclonal anti-Prlr antibody (1:250; ab214303; Abcam, Cambridge, GB) was used as the primary antibody and Alexa Fluor-594-conjugated donkey anti-rabbit IgG antibody (1:1,000 dilution; A21207; Invitrogen, MA, USA) was used as the secondary antibody. Images were captured using a confocal laser scanning microscope (LSM 800; Zeiss, Oberkochen, Germany). Immunoactivities were quantified using ImageJ software (National Institutes of Health, MD, USA) in three animals by using three consecutive sections per animal in both the left and right hemispheres.

Quantitative RT-PCR

The brain tissue of the female 11–13-week-old mice which were injected AAV vector before 3 weeks (AAV-mCherry flex dtA in OxtrCre or WT, AAV-LacZ or AAV-GFPCre in Oxtr-floxed, and AAV-Oxtr in Oxtr−/−) were collected after euthanized by cervical dislocation. As the control, the brain tissue of same-week-old female WT and OxtrCre mice was collected in the same way. Brain sections with a thickness of 300 μm at +0.6 mm to +0.3 mm from the bregma for the LS and at +0.3 mm–0 mm from the bregma for the MPOA were obtained using a Vibratome (VT1200S, LEICA, Wetzlar, Germany), and the LS or MPOA was excised by a disposable biopsy punch (Kai industries co., ltd. Gifu, Japan). Total RNA was extracted using TRIzol Reagent (15596018; Invitrogen, MA, USA) and converted into cDNA by using the Prime Script RT reagent kit (RR037A; Takara Bio. Inc., Shiga, Japan) according to the manufacturer’s instructions. Real-time quantitative PCR was performed using a Thermal Cycler Dice Real-Time System (Takara Bio. Inc., Shiga, Japan) and TB Green Premix Ex Taq II (RR820A; Takara Bio. Inc., Shiga Japan). For qPCR, the cDNA product was added to 10 µL of TB Green Premix Ex Taq II (RR420S, Takara Bio. Inc., Shiga, Japan) and 0.4 µm primers to a final volume of 20 µL. For each sample, a parallel reaction was set up with Oxtr and Gapdh as an endogenous control. The following primers were used: Oxtr, forward, 5′-TGC​TGC​AAC​CCA​TGG​ATC​TA-3′ and reverse, 5′-TAA​TGC​TCG​TCT​CTC​CAG​GC-3′; GAPDH, forward, 5ʹ-TGA​CGT​GCC​GCC​TGG​AGA​AA-3ʹ and reverse, 5ʹ-AGT​GTA​GCC​CAA​GAT​GCC​CTT​CAG-3ʹ.

Whole-Mount Staining of the Mammary Gland

To isolate the mammary glands, the mice were euthanized by cervical dislocation. The skin was then peeled back to make the mammary glands visible. The 2nd or 3rd gland was collected and spread out over a microscopy glass slide. The slides were then dried until the glands were stuck to the glass surface. The mammary glands were fixed for 2 h in 4% paraformaldehyde at 4°C. The tissue was rinsed in PBS and stained with carmine alum solution overnight at 20°C. The tissue was then washed in 70% ethanol for 1 h, 95% ethanol for 1 h, and finally in 100% ethanol. The tissue was cleared in xylene overnight before being mounted with Marinol (10781; Muto Pure Chemical Co. Tokyo, Japan). Details of the sampling are shown in Table 1.

Prl Measurement

Prl levels in plasma and pituitary gland were measured in WT and Oxtr−/− mice, as well as in OxtrCre mice infected with AAV-mCherry-flex-dtA in the MPOA and in Oxtrfx mice infected with AAV-GFPCre in the MPOA. The details of sampling are shown in Table 1. Plasma samples were collected between 2:00 and 3:00 p.m., on the day of diestrus or 1 day after parturition. Ten minutes after intraperitoneal administration of a mixture of three types of anesthetic agents, blood samples were collected by cardiocentesis. After ice-cooling the blood samples for 2 h, they were processed by centrifugation at 12,000 rpm, 4°C for 15 min, and stored at −20°C until further analysis. Prl plasma levels were measured using a commercially available ELISA kit (ab100736; Abcam, Cambridge, UK). The pituitary glands were collected immediately after euthanasia by cervical dislocation between 2:00 and 3:00 p.m., on the day of diestrus or 1 day after parturition. To avoid undesired effects of anesthesia induction on the results, different mice were used for the collection of pituitary glands and of blood samples. Pituitary glands were homogenized in 1m urea/RIPA lysis buffer (WSE-7420; ATTO Co. Tokyo, Japan) and separated using SDS-PAGE. Western blotting was performed with rabbit anti-Prl antibody (1:5,000; ab183967; Abcam, Cambridge, UK) and anti-GAPDH antibody (1:5,000; GTX100118; Genetex Co. CA, USA) as the primary antibody and anti-rabbit IgG-coupled HRP (1:5,000; VECTOR Co. CA, USA) as the secondary antibody. The antibody-antigen complexes were detected using an ECL system (183967 Thermo Scientific Co. MA. USA).

Statistical Analysis

GraphPad Prism (GraphPad Software Inc. CA, USA) was used to analyze the data. Data are expressed as the mean ± standard error of the mean (SEM). Two-way ANOVA was used to analyze the data shown on Figure 1g, h, Figure 2a, b, and online supplementary Figure 1 (for all online suppl. material, see https://doi.org/10.1159/000535362). Student’s t test was used to analyze the data shown on Figure 1a, c, f, Figure 2c, and online supplementary Figure 3E. One-way ANOVA and Dunnett’s multiple comparison test were used to analyze the data shown on Figure 3b, f, j, l, m, Figure 4b, e, i and online supplementary Figure 3C. The remaining data from the maternal behavior tests were analyzed using Dunn’s method with Kruskal-Wallis multiple comparisons (Fig. 3d, e, g, h, i, k, Fig. 4c, d, f, g, h, j, and online supplementary Fig. 3A, B, D). All the statistical results are presented in the figure legends. Significance levels were set at *p < 0.05, **p < 0.01, ***p < 0.005, ****p < 0.0001.

Fig. 1.

Changes in Oxtr expression and neuronal activation in the LS and MPOA in the presence of offspring. a Quantification of the number of Venus-positive cells in the LS during diestrus and nurturing (n = 6). b Representative photomicrographs of immunohistochemical labeling for Oxtr (green; Venus) in the LS of female OxtrV/+ mice during diestrus and nurturing. c Quantification of the number of Venus-positive cells in the MPOA during diestrus and nurturing (n = 6). d Representative photomicrographs of immunohistochemical labeling for Oxtr in the MPOA of female OxtrV/+ mice during diestrus and nurturing. e Temporal schedule of behavioral testing of virgin mice with pup stimulation. f Quantification of the sniffing latency of each pups (left) and licking time (right) of WT and Oxtr−/− mice (n = 9 each). g Quantification of the number of c-fos-positive cells in the LS without and with pup stimulation in WT and Oxtr−/− virgin mice, respectively (n = 4; left graph) and representative photomicrographs of immunohistochemical labeling for c-fos (black nuclear staining; right panel). h Quantification of the number of c-fos-positive cells in the MPOA without and with pup stimulation in WT and Oxtr−/− mice, respectively (n = 4; left panel) and representative photomicrographs of immunohistochemical labeling of c-fos (black nuclear staining) (right panel). Scale bar represents 100 µm. *p < 0.05. **p < 0.01. NS: not significant. Error bars indicate the standard error of the mean (SEM). LS, lateral septal nucleus; MPOA, medial preoptic area; Oxtr, oxytocin receptor.

Fig. 2.

a Oxtr expression in the MPOA correlates with Prl levels. Representative Western blot of lysate from the pituitary gland of a female in diestrus or during nurturing (left panel). WT, D: WT in diestrus. WT, N: WT during nurturing; Oxtr−/−, D: Oxtr−/− in diestrus; Oxtr−/−, N: Oxtr−/− during nurturing; OxtrCre+dtA/MPOA, D: OxtrCre+dtA/MPOA in diestrus; OxtrCre+dtA/MPOA, N: OxtrCre+dtA/MPOA during nurturing; Oxtrfx+Cre/MPOA, D: Oxtrfx+Cre/MPOA in diestrus; Oxtrfx+Cre/MPOA, N: Oxtrfx+Cre/MPOA during nurturing. The molecular weights from the Prl-specific band were 23 kDa and that for the GAPDH-specific band was 36 kDa. Quantitative densitometry analysis of Prl/GAPDH (n = 3, right panel). b Prl concentrations in the plasma of females in diestrus or during nurturing measured by ELISA (n = 4). The notation of the samples is the same as that for Western blotting. c Representative photomicrographs of Venus (Oxtr; green) and Prl (red) immunohistochemistry staining in the MPOA during nurturing and diestrus in Oxtrv/+ mice. Quantification of the number of Venus- and Prl-positive cells in the MPOA during diestrus and nurturing (n = 9). *p < 0.05, **p < 0.01, ***p < 0.005, ****p < 0.001. NS: not significant. Error bars indicate the standard error of the mean (SEM). GAPDH, glyceraldehyde 3-phosphate dehydrogenase; MPOA, medial preoptic area; Oxtr, oxytocin receptor; Prl, prolactin; +dtA, +AAV-mCherry-flex-dtA; +Cre, + AAV-GFP-Cre.

Fig. 3.

Elimination of Oxtr-expressing cells in the LS or MPOA elicit abnormal nurturing behavior. a Diagram of the AAV vector depicting the mCherry and diphtheria toxin A and OxtrCre knock-in cassette (top); Temporal schedule of maternal behavioral testing of primipara (bottom). b Representative photomicrographs of viral infected area in the LS (top) and MPOA (bottom), relative expression of Oxtr mRNA of the WT + dtA/LS group (n = 6), the OxtrCre+dtA/LS group (n = 9), and the OxtrCre group (n = 8) in the LS (top), and relative expression of Oxtr mRNA the WT + dtA/MPOA group (n = 4), the OxtrCre+dtA/MPOA group (n = 9), and the OxtrCre group (n = 9) in the MPOA (bottom). c Schematics of the maternal behavior tests on the morning of parturition. Pregnant mice were individually housed 1 week before parturition. Percentages of scattered newborns were recorded on the morning after their birth (1). The mother and her pups were removed 1 h before the test (2) Three one-three day-old pups from WT dams were placed in each corner of the cage distant from the nest (3) During the next 30 min, the behaviors of the test mice were recorded by a video system, and the number of crouching episodes over pups and the duration of crouching over all three pups in the nest was recorded (4,5) After the behavioral test, five foster pups were placed in the home cage of the female mouse being tested, and the survival rates were recorded on the fifth day. d–g Maternal behavior test of WT + dtA/LS, OxtrCre+dtA/LS, and OxtrCre mice. d Percentage of scattered pups [WT + dtA/LS (n = 9), OxtrCre+dtA/LS (n = 13), OxtrCre (n = 6)]. e Number of retrieved pups [WT + dtA/LS (n = 10), OxtrCre+dtA/LS (n = 13), OxtrCre (n = 7)]. f Duration of crouching over all pups [WT + dtA/LS (n = 10), OxtrCre+dtA/LS (n = 13), OxtrCre (n = 6)]. g Survival rate of pups [WT + dtA/LS (n = 10), OxtrCre+dtA/LS (n = 11), OxtrCre (n = 6)]. h–k Maternal behavior test of WT + dtA/MPOA (n = 9), OxtrCre+dtA/MPOA (n = 10), and OxtrCre mice (n = 6). h Percentage of scattered pups. i Number of retrieved pups. j Duration of crouching over all pups. k Survival rate of pups. l Rate of body weight gain (left panel) of two-day-old pups born from WT + dtA/MPOA mothers (n = 15), OxtrCre+dtA/MPOA mothers (n = 14), and OxtrCre mothers (n = 15); and stomach weight (right panel) of 2-day-old pups born from WT + dtA/MPOA mothers (n = 14), OxtrCre+dtA/MPOA mothers (n = 14), and OxtrCre mothers (n = 15). m Rate of body weight gain (left panel) of 3-day-old pups born from WT + dtA/MPOA mothers (n = 13), OxtrCre+dtA/MPOA mothers (n = 14), and OxtrCre mothers (n = 15); stomach weight (right panel) of 3-day-old pups born from WT + dtA/MPOA mothers (n = 13), OxtrCre+dtA/MPOA mothers (n = 14), and OxtrCre mothers (n = 13). *p < 0.05, **p < 0.01, ***p < 0.005, ****p < 0.001. NS: not significant. Error bars indicate the standard error of the mean (SEM). LS, lateral septal nucleus; MPOA, medial preoptic area; Oxtr, oxytocin receptor; +dtA, +AAV-mCherry-flex-dtA.

Results

Nurturing Mothers Showed Increased Expression of Oxtr in the LS and MPOA

Here, we analyzed changes in the number of Oxtr-expressing neurons in the LS and MPOA in relation to parturition by using heterozygous Oxtr-Venus knock-in mice (Oxtr^V/+). The Oxtr-Venus knock-in mouse model was established by Yoshida et al. [28] to visualize the distribution of neurons that express Oxtr in the CNS. The numbers of Venus-positive neurons in the LS and MPOA of Oxtr^V/+ mothers during nurturing were greater than those in the same regions of animals during diestrus (LS: p = 0.0052, Fig. 1a; MPOA: p < 0.0001, Fig. 1c, online suppl. Fig. 4). This result revealed that the increase in Oxtr expression in the LS and MPOA might be under the same control mechanism, including the induction of Oxt-dependent maternal behavior, and we hypothesized that the increase itself would be essential for the control of maternal behavior.

Pup stimulation increases neural activation in the LS and MPOA of virgin WT mice but not of Oxtr−/− mice. To characterize the role of Oxtr in maternal behavior, we compared the behavior of virgin WT and Oxtr−/− mice upon exposure to three neonates (Fig. 1e, f). The sniffing latency for the second and third pup, but not for the first pup, was longer in Oxtr−/− mice compared to wild-type mice (1st pup, p = 0.8702; 2nd pup, p = 0.016; 3rd pup, p = 0.0124). On the other hand, there was no difference in the time spent licking the pups (p = 0.1503; Fig. 1f, online suppl. Fig. 4). Next, we observed the neural activation in the LS and MPOA in response to pup stimulation (Fig. 1g and h, online suppl. Fig. 4). To verify the Oxtr-dependent neural activation in the LS with pup stimulation, we quantified the number of c-fos-positive cells in the LS of the virgin WT and Oxtr−/− virgin mice. After exposure to pups for 5 min, c-fos density increased significantly in the LS of WT virgin mice compared to Oxtr−/− virgin mice (p = 0.0023; Fig. 1g). These results suggest that Oxtr might be involved in pup cognition behavior and in the neural activation in the LS. In addition, we quantified the number of c-fos-positive cells in the MPOA following pup stimulation. The number of c-fos positive cells in the MPOA of Oxtr−/− mice in response to pup stimulation was not impaired by the disruption in Oxt-Oxtr signaling (exposed WT vs. exposed Oxtr−/−; p = 0.9364; Fig. 1h). Next, we quantified the number of c-fos-positive cells in the MPOA of the WT and Oxtr−/− mice during diestrus and during nurturing (online suppl. Fig. 1, 8). In the MPOA, c-fos induction at the postpartum stage was observed in WT, but not in Oxtr−/− mice (WT during nurturing vs. Oxtr−/− during nurturing, p = 0.0009). These results suggest that the expression of Oxtr may be associated with neural activation in the MPOA during maternal behavior after parturition.

Elimination of Oxtr-Expressing Cells in the LS or MPOA Elicit Abnormal Nurturing Behavior

To specifically ablate Oxtr-expressing neurons in the LS or MPOA, recombinant AAV expressing a Cre-dependent diphtheria toxin A fragment (designated as AAV-mcherry-flex-dtA, shown in Fig. 3a) was injected bilaterally into the LS or MPOA of OxtrCre mice, which were designated as OxtrCre+dtA/LS and OxtrCre+dtA/MPOA, respectively. As control groups, WT mice were injected with AAV-flex-dtA into the LS or MPOA (designated as WT + dtA/LS and WT + dtA/MPOA). OxtrCre mice (designated as OxtrCre) were also used as control group. The expression levels of Oxtr mRNA in the LS of mice from the OxtrCre+dtA/LS group were significantly reduced compared to the WT + dtA/LS group (p < 0.005) and the OxtrCre group (p < 0.005; Fig. 3b, online suppl. Fig. 4). The expression level of Oxtr mRNA in the MPOA of mice from the OxtrCre+dtA/MPOA group was significantly decreased compared with that of WT + dtA/MPOA (p < 0.01) and OxtrCre (p < 0.005) mice (Fig. 3b). To confirm disruption of Oxtr, OxtrVenus/Cre mice, with Venus and Cre sequences in the coding region of Oxtr, were generated by crossing OxtrVenus and OxtrCre mice, and received an AAV-mCherry-FLEX-dtA injection into the MPOA. Infecting the MPOA of OxtrVenus/Cre mice with AAV-mCherry-FLEX-dtA resulted in the distinct disappearance of the Venus signal and numerous mCherry signals remained in the infected area; however, there was no co-localization of Venus and mCherry signals within the same cells. We concluded the highly specific destruction of Oxtr-expressing cells caused by AAV-mCherry-FLEX-dtA in the infected area of OxtrCre+dtA mice, and even if there were side effects (low-frequency leak expression and/or integration of the AAV-mCherry-FLEX-dtA sequence into the infected cell genome), they did not significantly affect the experimental results (online suppl. Fig. 2).

On the morning of the day of parturition, the percentage of scattered newborns was recorded. No significant differences were observed between OxtrCre+dtA/LS and control groups in the percentages of scattered pups (vs. WT + dtA/LS, p = 0.9863; vs. OxtrCre, p = 0.7373; Fig. 3d, online suppl. Fig. 5). Next, the pups were removed and after 1 h, each female was exposed to three 1 to 3 days-old pups from a WT dam. OxtrCre+dtA/LS dams retrieved a significantly lower number of pups compared with dams from the WT + dtA/LS and OxtrCre groups (vs. WT + dtA/LS, p = 0.0146; OxtrCre, p = 0.0322; Fig. 3e, online suppl. Fig. 5) and spent less time crouching over the pups compared to WT + dtA/LS and OxtrCre dams (WT + dtA/LS, p = 0.0075; OxtrCre, p = 0.0343; Fig. 3f, online suppl. Fig. 5). After the test, each female was housed with five foster pups for 5 days, and the survival rates of the pups were recorded. No significant differences were observed in the survival rates of pups from OxtrCre+dtA/LS mice compared with those from WT + dtA/LS and OxtrCre mice (WT + dtA/LS, p = 0.9050; OxtrCre, p = 0.157; Fig. 3g, online suppl. Fig. 5). In contrast to the results from mice injected in the LS, the survival rate of pups cared by OxtrCre+dtA/MPOA dams decreased significantly when compared with those from both control groups (WT + dtA/MPOA, p = 0.0127; OxtrCre, p = 0.0007 Fig. 3k, online suppl. Fig. 5) despite the fact that pup retrieval by OxtrCre+dtA/MPOA dams was not seriously affected (number of pups retrieved: OxtrCre+dtA/MPOA vs. WT + dtA/MPOA, p = 0.2325; OxtrCre+dtA/MPOA vs. OxtrCre, p = 0.3245; Fig. 3i; duration of crouching over all pups: OxtrCre+dtA/MPOA vs. WT + dtA/MPOA, p = 0.1063; OxtrCre+dtA/MPOA vs. OxtrCre, p = 0.0852; Fig. 3j, online suppl. Fig. 5). Compared with the OxtrCre mice, OxtrCre+dtA/MPOA dams abandoned more pups outside the nest (OxtrCre+dtA/MPOA vs. OxtrCre, p = 0.0046; Fig. 3h, online suppl. Fig. 5).

To determine the cause of the decrease in the pup survival rate, we measured the body weight and stomach weight of the surviving pups. On the day of birth (D1), pups from the OxtrCre+dtA/MPOA, WT + dtA/MPOA, and OxtrCre groups had normal weight, with no significant differences among the experimental groups (data not shown). However, on D2 and D3, the rate of body weight gain in pups from the OxtrCre+dtA/MPOA group was significantly lower than in pups from the WT + dtA/MPOA (D2 and D3; p < 0.0001) and OxtrCre (D2 and D3; p < 0.0001) groups (Fig. 3l, m, online suppl. Fig. 6). On D2, the stomach weight of pups from the OxtrCre+dtA/MPOA group was significantly lower than that from the WT + dtA/MPOA (p = 0.0065) and OxtrCre (p < 0.0001) groups (Fig. 3l). On D3, the stomach weight of pups from the OxtrCre+dtA/MPOA group was significantly lower than that from the OxtrCre group (p = 0.0265) but not from the WT + dtA/MPOA group (p = 0.1452; Fig. 3m). Pups from OxtrCre+dtA/MPOA mothers on D2 and D3 were shown to be undernourished. Taken together, these results reveal that Oxtr-expressing neurons in the LS and MPOA are both important for maternal nurturing. However, the influence of Oxtr-expressing neurons on maternal behavior may differ between the LS and the MPOA.

Deletion of Oxtrs in the LS Leads to Normal Nurturing, whereas Their Deletion in the MPOA Elicits Abnormal Nurturing Behaviors

To identify the roles of Oxtrs in several nuclei in postpartum maternal behavior, we injected AAV-AcGFP-Cre (designated as AAV-Cre) into the LS or MPOA of Oxtrfx mice (designated as Oxtrfx+Cre/LS and Oxtrfx+Cre/MPOA; Fig. 4a, b). Oxtrfx mice injected with the AAV-β-galactosidase vector into the LS or MPOA (designated as Oxtrfx+LacZ/LS and Oxtrfx+LacZ/MPOA, respectively) and the same mutant mice without viral vector injection (designated as Oxtrfx) were used as control groups.

Fig. 4.

Deletion of Oxtrs in the LS leads to normal nurturing, whereas their deletion in the MPOA elicits abnormal nurturing behaviors. a Diagram of the AAV vector depicting the GFP-Cre and Oxtr-flox knock-in cassettes (top); Temporal schedule of maternal behavioral testing of primipara (bottom). b Representative photomicrographs of viral infected area in the LS (top) and MPOA (bottom), relative expression of the Oxtr mRNA in the LS of the Oxtrfx+LacZ/LS group (n = 5), the Oxtrfx+Cre/LS group (n = 5), and the Oxtrfx group (n = 5; top), and in the MPOA of the Oxtrfx+LacZ/MPOA group (n = 6), the Oxtrfx+Cre/MPOA group (n = 6), and the Oxtrfx group (n = 6; bottom). c–f Maternal behavior test of Oxtrfx+LacZ/LS, Oxtrfx+Cre/LS, and Oxtrfx mice. c Percentage of scattered pups [Oxtrfx+LacZ/LS (n = 8), Oxtrfx+Cre/LS (n = 8), Oxtrfx (n = 9)]. d Number of retrieved pups [Oxtrfx+LacZ/LS (n = 8), Oxtrfx+Cre/LS (n = 8), Oxtrfx (n = 8)]. e Duration of crouching over all pups [Oxtrfx+LacZ/LS (n = 7), Oxtrfx+Cre/LS (n = 8), Oxtrfx (n = 9). f Survival rate of pups [Oxtrfx+LacZ/LS (n = 9), Oxtrfx+Cre/LS (n = 12), Oxtrfx (n = 9)]. g–j Maternal behavior test of Oxtrfx+LacZ/MPOA, Oxtrfx+Cre/MPOA, and Oxtrfx mice. g Percentage of scattered pups [Oxtrfx+LacZ/MPOA (n = 9), Oxtrfx+Cre/MPOA (n = 12), and Oxtrfx (n = 9)]. h Number of retrieved pups [Oxtrfx+LacZ/MPOA (n = 9), Oxtrfx+Cre/MPOA (n = 8), Oxtrfx (n = 10)]. i Duration of crouching over all pups [Oxtrfx+LacZ/MPOA (n = 9), Oxtrfx+Cre/MPOA (n = 8), Oxtrfx (n = 9)]. j Survival rate of pups [Oxtrfx+LacZ/MPOA (n = 7), Oxtrfx+Cre/MPOA (n = 8), Oxtrfx (n = 9)]. *p < 0.05, ***p < 0.005, NS: not significant. Error bars indicate the standard error of the mean (SEM). AAV, adeno-associated virus; LS, lateral septal nucleus; MPOA, medial preoptic area; Oxtr, oxytocin receptor; +LacZ, + AAV-LacZ; +Cre, + AAV-GFP-Cre.

As a result, the expression level of Oxtr mRNA in the LS in the Oxtrfx+Cre/LS group was significantly reduced compared with those of the Oxtrfx+LacZ/LS group (p < 0.01; Fig. 4b, online suppl. Fig. 6) and the Oxtrfx group (p < 0.01; Fig. 4b), and the expression level of Oxtr mRNA in the MPOA of the Oxtrfx+Cre/MPOA group was significantly decreased compared with those of the Oxtrfx+LacZ/MPOA group (p < 0.01; Fig. 4b) and the Oxtrfx group (p < 0.01; Fig. 4b). Their maternal behaviors were examined on the first day of nurturing after parturition (Fig. 4c–j, online suppl. Fig. 4, 5).

The decreased Oxtr expression in the LS in the Oxtrfx+Cre/LS group had no adverse effects on the percentage of scattered pups compared with those of the Oxtrfx+LacZ/LS group (p > 0.9999) and the Oxtrfx group (p = 0.7589; Fig. 4c), the number of pups retrieved (Oxtrfx+Cre/LS vs. Oxtrfx+LacZ/LS, p = 0.4413; Oxtrfx+Cre/LS vs. Oxtrfx, p = 0.4413; Fig. 4d), duration of crouching over all pups (Oxtrfx+Cre/LS vs. Oxtrfx+LacZ/LS, p > 0.9999; Oxtrfx+Cre/LS vs. Oxtrfx, p > 0.9999; Fig. 4e) and survival rate of the pups (Oxtrfx+Cre/LS vs. Oxtrfx+LacZ/LS, p = 0.1912; Oxtrfx+Cre/LS vs. Oxtrfx, p = 0.1445; Fig. 4f).

These results indicate two possible mechanisms: (1) Oxtr in the LS did not contribute to maternal behavior; pup retrieval, and crouching over pups; and or (2) normal maternal behavior required the minimum expression of Oxtr in the LS, and Oxtr expression greater than in this case may impair maternal behavior. Based on the second hypothesized mechanism, we challenged the injection of the AAV-Oxtr vector [45] into the LS of Oxtr−/− mice to rescue expression of Oxtr in the LS region non-selectively (designated as Oxtr−/− + Oxtr/LS). The impaired pup-retrieving behavior observed in Oxtr−/− mice was rescued by the exogenous expression of Oxtr in the LS (Oxtr−/− + Oxtr/LS group) compared with the Oxtr−/− group (p = 0.0205; online suppl. Fig. 3B, 9), and the duration of crouching behavior over all pups was increased in the Oxtr−/− + Oxtr/LS group compared with the Oxtr−/− group (p = 0.0114; online suppl. Fig. 3C, 9). In contrast, no improvement by exogenous expression of Oxtr in the LS of Oxtr−/− mice was observed for the remaining two variables: percentage of scattered pups compared (Oxtr−/− + Oxtr/LS vs. Oxtr−/−; p > 0.9999) and survival rate of pups (Oxtr−/− + Oxtr/LS vs. Oxtr^−/−, p > 0.9999; Oxtr−/− + Oxtr/LS vs. WT, p < 0.0001; online suppl. Fig. 3D, 9). In the MPOA, decreased Oxtr expression (Oxtrfx+Cre/MPOA) increased the percentage of scattered pups compared with those of the Oxtrfx group (p = 0.0299) and the Oxtrfx+LacZ/MPOA group (p = 0.0619; Fig. 4g), and reduced the survival rate of pups in comparison with both control groups (Oxtrfx+Cre/MPOA vs. Oxtrfx+LacZ/MPOA, p = 0.0263; Oxtrfx+Cre/MPOA vs. Oxtrfx, p = 0.0003; Fig. 4j). However, the number of retrieved pups and the duration of crouching behavior over all pups displayed by the Oxtrfx+Cre/MPOA group were not different from those displayed by the Oxtrfx+LacZ/MPOA and Oxtrfx groups (Fig. 4h, i). Taken together, these results suggest that Oxtr in the MPOA contributes to lactation, but we could not find a correlation between Oxtr expression in the MPOA and maternal motivation.

Oxtr-Expressing Cells in the MPOA Regulate Prl Concentration after Parturition

The reduction in pup survival rates due to reductions in the number of Oxtr-expressing cells or Oxtr expression in the MPOA (Fig. 3k, 4j) suggested that Oxtr in the MPOA may regulate differentiation of the mammary glands or milk ejection. Using whole-mount staining of the mammary glands, we examined their differentiation during lactation in Oxtr−/−, OxtrCre+dtA/MPOA, and WT mice. All mice showed normal mammary gland development (online suppl. Fig. 3F). We had previously reported that female Oxtr−/− mice showed defects in maternal behavior and milk ejection and that their pups starved within 24 h [29]. Prl is a protein hormone that influences many physiological events [50] and regulates maternal behavior through Prlr [16, 51]. Many reports have suggested that Oxt is a Prl-releasing factor [18, 19, 52–54] and is secreted in response to suckling [50]. For these reasons, we quantified the Prl concentrations in the pituitary gland by Western blotting (Fig.2a, online suppl. Fig. 7) and in plasma by ELISA (Fig. 2b, online suppl. Fig. 8) in diestrus and 1 day after parturition. Plasma samples were collected under anesthesia. The plasma concentrations of Prl reportedly change more than 2 h after the administration of anesthesia in rats [55]. Therefore, we set out the conditions for blood collection following anesthesia, while minimizing any potential impact on alterations of plasma Prl concentration induced by anesthesia (as described in the Methods section). In addition, different mice from those employed for blood sample collection were used to obtain the pituitary gland specimens, in order to mitigate any potential confounding influences arising from anesthesia induction on the obtained results. One day postpartum, WT dams showed significantly elevated Prl concentration in the pituitary gland and plasma in comparison with the corresponding concentrations during diestrus (p = 0.0311; Fig. 2a; p < 0.0001, Fig. 2b). However, Oxtr−/−, OxtrCre+dtA/MPOA, and Oxtrfx+Cre/MPOA dams did not show any statistically significant increase in Prl concentrations upon parturition in comparison with the levels measured during diestrus in the pituitary gland (Oxtr−/−, p = 0.9877; OxtrCre+dtA/MPOA, p = 0.9933; Oxtrfx+Cre/MPOA, p = 0.9803; Fig. 2a) or plasma (Oxtr−/−, p = 0.9981; OxtrCre+dtA/MPOA, p = 0.1734; Oxtrfx+Cre/MPOA, p = 0.9394; Fig. 2b). These results suggest that Oxtr expression in the MPOA during nurturing correlate with Prl secretion. We also assessed the co-localization of Oxtr and Prl, as well as the number of cells expressing Oxtr and Prlr in the MPOA during nurturing by immunostaining brain sections from Oxtr^v/+ mice. After parturition, the MPOA showed a significant increase in Oxtr-positive cells and Prl-positive cells compared to the diestrus period (Oxtr; p = 0.0008, Prl; p = 0.0045). The number of cells expressing Oxtr and Prl also increased (p < 0.0001). The frequency of cells expressing Prl in cells expressing Oxtr also increased compared with diestrus (p = 0.0007; Fig. 2c, online suppl. Fig. 8). On the other hand, Oxtr and Prlr did not co-localize in the MPOA (online suppl. Fig. 3G). These results suggest that Prl-Prlr signal transduction is coupled with Oxt-Oxtr signaling in the MPOA.

Discussion

This study showed the contribution of Oxtr-expressing neuron for maternal behavior in the two nuclei, MPOA and LS, that control parental behavior [31–33]. Our experimental data strongly suggested that Oxtr expressed in MPOA was mainly important for the lactation process and was involved in the regulation of postnatal Prl secretion, without affecting retrieving and crouching over behavior just after mother-infant separation. Under those experimental condition, the Oxtr-expressing cells in LS influenced maternal behavior such as retrieving and crouching over, but the Oxtr expression itself was not essential. These results were significant to be able to reveal functions restricted to Oxtr in each of the two nuclei.

Role of Oxtr in Neural Activation Related to Maternal Behavior

In the LS and MPOA, Oxtr expression was increased after delivery (Fig. 1a and c). We then analyzed maternal behavior and neural activation of the LS and MPOA in WT and Oxtr−/− mice by measuring the level of c-fos expression in these nuclei when nulliparous mice had their first contact with pups. The purpose of this evaluation was to observe the behavior and the neural activation caused by pup stimulation as a simple trigger for the onset of maternal behavior, while avoiding the effect of changes in the endocrine system due to delivery. Sniffing latency for the second and third pup, but not for the first pup, was longer for Oxtr−/− than for WT mice. On the other hand, there was no difference in the time spent licking the pups (Fig. 1f). In WT mice, pup stimulation increased c-fos expression in both the LS and MPOA of nulliparous mice. In Oxtr−/− mice, the increase in c-fos expression was still suppressed in the LS, whereas an increase in c-fos expression was observed in the MPOA, similar that in the WT mice (Fig. 1g, h). When exposed to neonates for the first time, virgin mice are known to sniff them from a distance, examine them, and then exhibit a conflict between approach and avoidance, finally retrieving them to the nest [56]. In the case of the virgin Oxtr−/− females tested in our study, it took longer for them to sniff out compared to WT mice (Fig. 1f). However, there were no differences in the time spent licking the pups, which is a typical maternal behavior. These results suggest that Oxtrs were involved in the neural activation of the LS when recognizing neonates by their scent. Our previous study showed that experimentally activation of Oxtr-expressing neurons in the LS resulted in an increased contact time with novel adult mice, indicating that Oxtrs located in the LS are involved in the response to social novelty (i.e., social cognition) [41]. Therefore, it is possible that the activation of Oxtr-expressing neurons in the LS may not be specific to maternal behavior, but rather a neural activation that is common to all forms of social recognition.

On the other hand, Oxtr in the MPOA was unimportant for neuronal activation in nulliparous mice stimulated by offspring, indicating that other factors may compensate for its absence. Administration of estrogen, Prl, Oxt, or dopamine into the MPOA of virgin female rats facilitated maternal responsiveness [9, 57–59]. In contrast, neural activation in postpartum Oxtr−/− mother mice were suppressed in the MPOA. c-fos expression in the MPOA of WT maternal rats or mice is increased after delivery [59, 60]. After delivery, the concentrations of reproductive hormones such as Prl, estrogen, and Oxt in the brain are elevated, and their receptors are highly expressed in the MPOA [10, 61, 62]. The dramatic changes in the endocrine system and gene expression due to delivery, milk ejection, and nurturing, in addition to stimulation by offspring, cause elevation in c-fos expression in the MPOA. In the present study, we showed that Oxtr expression in the MPOA played an important role in postpartum neural activation through the endocrine system.

Role of Oxtr in LS for Maternal Responsiveness

Our group generated three lines of Oxtr gene-modified mice and infected the LS and MPOA of these mice with AAV viral vectors to specifically modify Oxtr-expressing neurons in these nuclei. By doing so, we were able to measure changes in maternal behaviors. We showed that reduction of Oxtr-expressing neurons in the LS of OxtrCre mice impaired maternal behavior after delivery (Fig. 3e, f). However, specific blockade of Oxtr expression in the LS of Oxtrfx mice did not affect maternal behavior after delivery (Fig. 4d, e). The results shown in Figure 4d and e suggest that: (1) the expression of Oxtr in the LS itself was not essential for the behavior of retrieving and crouching over the pups, or (2) the decrease in the levels of the Oxtr expression was insufficient, with an undetectable effect on at the behavioral level. However, the fact that most Oxtr-expressing cells in the LS are GABAergic [28, 41, 42] suggest that GABAergic neurons in the LS are important for postpartum maternal behavior. Moreover, it is known that in the postpartum LS, the concentration of GABA and the expression levels of GABA synthases (Gad65 and Gad67) are increased [40]. The results of this report suggest that Oxtr-expressing neurons in the LS are essential for postpartum maternal behavior regardless of Oxt-Oxtr signaling, at least in our experimental system. On the other hand, when the LS region of Oxtr−/− mice was infected with AAV-Oxtr to non-selectively express Oxtr in the LS region, the behavior of retrieving the pups and duration of crouching over all pups were rescued (online suppl. Fig. 3B, C). This indicated that some neurons distributed in the LS that did not originally express Oxtrs might have the potential to regulate maternal behavior with the expression of the Oxtr gene, which was exogenously added in the present study. The results of the maternal behavior test for Oxtr−/− mice were different from those presented in the study by Tsuneoka et al. [63]. There were some differences in the experimental conditions between the two studies. Tsuneoka et al. compared Oxtr+/− and Oxtr−/− mice. Oxtr+/− mice have been reported to exhibit impaired performance in social novelty tests compared to WT mice [64]. It is possible that the maternal behavior of Oxtr+/− mice may not be equal to that of WT mice. Differences in background may also subtly affect maternal behavior. The separation period between mother and offspring was set at 1 h before performing the maternal behavior test in our experimental protocol. In contrast, the behavioral tests reported by Tsuneoka et al. were conducted without a previous separation period. The forced separation can be stressful for both the mothers and the offspring, which may have influenced the test results. The behavioral observation time was 30 min in this study and 15 min in the study by Tsuneoka et al., which may have contributed to the differences in the crouching time results. Differences in the environment of the experimental facility may also have affected the results.

Role of Oxtr Expression in the MPOA for Maternal Responsiveness

The level of Oxtr expression in the MPOA was also increased in mice after delivery (Fig. 1c, d). We first injected AAV-flex-dtA into the MPOA of OxtrCre mice to remove Oxtr-expressing neurons specifically from the MPOA region (Fig. 3h–k). Second, we injected AAV-Cre into Oxtrfx mice to remove the Oxtr gene (Fig. 4g–j). In the morning of the day of parturition, the pups derived from OxtrCre+dtA/MPOA and Oxtrfx+Cre/MPOA mothers were scattered outside the nest at a higher rate (Fig. 3h, 4g). We suspected that the pups were debilitated due to reduced lactation and were abandoned by the mothers, leading to an increased percentage of pups left outside per the total number of pups. Next, we removed the pups born from the test mice and used three newborn WT pups to perform maternal behavior tests. Surprisingly, OxtrCre+dtA/MPOA and Oxtrfx+Cre/MPOA mothers showed normal behavior of retrieving and crouching over the pups (Fig. 3i, j, 4h, i). We suspected that the pups used for the maternal behavior test were vigorous because they had been transferred from a WT mice nest, and their stimuli with vocalization or body temperature to the mutant mother were normal. Therefore, it can be assumed that the retrieving and crouching over behaviors of OxtrCre+dtA/MPOA and Oxtrfx+Cre/MPOA mothers were normally induced. After that, two newborn vigorous pups were added to the cage, and the survival rate of the pups after 5 days was observed. The survival rate of the WT pups nurtured by OxtrCre+dtA/MPOA and Oxtrfx+Cre/MPOA mothers decreased (Fig. 3k, 4j).

The results showed that the survival rate of WT pups nurtured by OxtrCre+dtA/MPOA and Oxtrfx+Cre/MPOA mothers decreased, despite normal nurturing behavior. Moreover, pups from the OxtrCre+dtA/MPOA group also had lower body and stomach weight (Fig. 3l, m). These findings suggest that Oxtr in the MPOA plays an important role in lactation and sustained nurturing behavior. It is expected that the lack of Prl increase after parturition might lead to low milk supply in the early postnatal stages, resulting in a reduced survival rate. During lactation, prevention of Prl increase after parturition might inhibit induction of maternal behavior via the Prl-Prlr system. It is known that pStat5 (Prl)-induced signal and transduction factor5) and Prl increase in the hypothalamus during lactation. Brown et al. [16] reported that female mice lacking Prlr in the MPOA show normal parturition and initiate normal pup-directed behavior, but they do not continue nurturing their litters, resulting in the death of all pups. The observed deficits were expected to directly result from a lack of Prl input into the circuitry controlling maternal behavior [16]. This study found that Prl concentration in the serum and pituitary gland increased after delivery in WT mice, but not in Oxtr−/− mice. Reducing the expression of the Oxtr gene only in the MPOA also resulted in a suppression of the increase in Prl concentration normally observed after delivery. The lack of Prl increase after parturition may lead to low milk supply in early postnatal stages, leading to a reduced survival rate.

Role of Prl and Oxt in Lactation

Prl is hormone mainly synthesized by lactotrophs in the anterior pituitary gland, the hypothalamus, and the mammary gland. Prl secretion is triggered by external stimuli such as light stress, auditory stimuli, and olfaction, by estradiol secretion, by mating, and by lactation. Especially, suckling stimulation causes the release of large amounts of Prl [50]. Oxt is another hormone involved in lactation, primarily responsible for milk ejection. Oxt is synthesized in the hypothalamus and released into the bloodstream by the posterior pituitary gland. It stimulates the contraction of myoepithelial cells surrounding the alveoli of the mammary glands, leading to milk ejection. Several studies have described a significant interrelationship between Prl and Oxt [18–20]. Our results suggest that Oxtr in the MPOA may regulate the Prl concentration in the serum and pituitary gland after delivery. Increased Prl levels may be due to suckling stimulation by pups [65]. Disruption of the Oxt-Oxtr system in the MPOA may suppress the increase in Prl levels after parturition and consequently reduce lactation ability. Conversely, reduction of milk ejection by a reduction in the Oxtr-mediated signal may cause a decrease in suckling stimuli from pups and prevent the increase in Prl levels. The Oxt-mediated milk ejection reflex is thought to contribute to the establishment and maintenance of milk production [66, 67]. It may be difficult to determine if Oxtr directly affects milk production in vivo. This is because Oxtr−/− mice cannot eject milk, so milk will accumulate in the lobules, and the resulting feedback may affect the amount of milk produced. However, a study on isolated mammary epithelial cells found that Oxt accelerates milk production [68]. The study showed that mammary epithelial cells from lactating rats and rabbits express Oxtr. The addition of Oxt increased the number of vesicles carrying milk component in close contact with the apical membrane via Oxtr and induced an accumulation of milk constituents. To better understand the regulation of lactation, it is necessary to clarify the factors that may regulate the expression of Oxtr and Prl in the MPOA during lactation. For example, tuberoinfundibular peptide 39 (Tip39) and galanin could be considered as candidates. Tip39 is expressed in the posterior intralaminar complex of the thalamus, which is characterized as a relay center for suckling stimulation by pups. Neurons containing Tip39 are projected to the preoptic area and hypothalamus. Tip39 was induced in lactating rats by suckling stimulation, and Tip39 containing nerve terminals innervated Oxt neurons in the PVN and galanin neurons in the preoptic area [69].

Role of Oxtr in MPOA

The MPOA is a region of the brain that plays a critical role in regulating maternal behavior. Neurons in the MPOA are responsible for the behavioral changes associated with motherhood and express estrogen receptors, Prlr, and Oxtrs [10, 70]. Many reports have indicated that maternal behavior is hindered by nonselective lesions of the MPOA [44, 71, 72]. Oxtrs in the MPOA are expressed under the control of the estrogen signal, and the levels of Oxtr expression in the MPOA correlate with the maternal response [59, 73, 74]. In contrast to the findings of our present study, the administration of an Oxtr or vasopressin receptor antagonist into the MPOA or ventral tegmental area of rats during delivery impaired maternal behavior in a previous study [75]. However, the rats in that study received an Oxtr antagonist, whereas our results were based on the use of AAV vectors in mice with modified genes. In addition, the methods used to analyze maternal behavior were also different in the two studies. Further analyses will be necessary to determine whether the differences in the results are due to differences in species or experimental methods. In addition, Prl signaling in the MPOA is required for postpartum maternal behavior and offspring survival [16, 51]. In the present study, we showed that the number of nerve cells co-expressing Prl and Oxt increased in the MPOA after parturition (Fig. 2c). Even if the Oxtr-expressing neurons in the MPOA were removed, the development of the mammary gland was unaffected (online suppl. Fig. 3F); nevertheless, the survival rate of the offspring was reduced (Fig. 3k). We have previously shown that Oxtr−/− mice exhibited abnormal maternal behavior and that the survival rate of offspring 24 h after birth was 0% [29]. Our results suggest that Oxtr in the MPOA is responsible for lactation, but that the nerve system directly related to Oxtr-dependent maternal behavior, such as retrieving and crouching over pups, may not be in the MPOA.

Conclusion

Our findings demonstrate that Oxtrs in the LS and MPOA are not essential for maternal behavior, at least in our experimental system, and that Oxtrs in the MPOA may be involved in lactation by controlling Prl production. Previously, we reported strong Oxtr expression in the prefrontal cortex, nucleus accumbens, bed nucleus of the ventral part of the stria terminalis, and ventral pallidum, which are important for parental behaviors, as well as in the accessory olfactory bulb and medial amygdala, which are related to abandonment of care for pups and aggressive behavior against them [28, 33]. Oxtr expression was found in various nuclei in the brain, and it was expected that the control of maternal behavior via the Oxt-Oxtr system differed among these nuclei. Further analysis of the influence of Oxtrs in other Oxtr expressing nuclei on maternal behavior using our experimental system will provide a better understanding of Oxt-Oxtr signaling and their roles in maternal nurturing. The study results have implications for understanding the neural basis of maternal behavior and may have implications for developing therapies for disorders related to parenting.

Acknowledgments

We thank Dr. Larry J. Young and Dr. Kengo Horie for their critical appraisal of the manuscript and their valuable comments. The authors also wish to acknowledge the assistance provided by Dr. Jun Hidema, Dr. Yuko Cho, and Daichi Osada of Tohoku University, in interpreting the significance of the results of this study.

Statement of Ethics

This study protocol was reviewed and approved by the Fukushima Medical University Institute of Animal Care and Use Committee, approved number 2019092, 2019099, and 2019100.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This work was supported by grants from the Japan Society for the Promotion of Science (JSPS) and the Japan Agency for Medical Research and Development (AMED).

Author Contributions

S.H., K.S. (Keisuke Sato), Y.T., and K.N. conceived the idea of the study. All experiments were mainly carried out by S.H., and partly by K.S. and the other members. H.M. and Y.M. especially contributed in terms of technology and materials, to the experiment of viral vector infection, and to that of the immuno-histological staining, respectively. S.H. drafted the original manuscript. K.S. significantly contributed to maintaining and improving the research environment. S.H. and K.N. made a major contribution to the acquisition of research funding. K.N. supervised the conduct of this study. All authors reviewed the manuscript draft and revised it critically on intellectual content. All authors approved the final version of the manuscript to be published.

Funding Statement

This work was supported by grants from the Japan Society for the Promotion of Science (JSPS) and the Japan Agency for Medical Research and Development (AMED).

Data Availability Statement

All data generated or analyzed during this study are included in this article and its online supplementary Figures 4–9. Further inquiries can be directed to the corresponding author.