Involvement of MCH-oxytocin neural relay within the hypothalamus in murine nursing behavior

Multiple sequential actions, performed during parental behaviors, are essential elements of reproduction in mammalian species. We showed that neurons expressing melanin concentrating hormone (MCH) in the lateral hypothalamic area (LHA) are more active in rodents of both sexes when exhibiting parental nursing behavior. Genetic ablation of the LHA-MCH neurons impaired maternal nursing. The post-birth survival rate was lower in pups born to female mice with congenitally ablated MCH neurons under control of tet-off system, exhibiting reduced crouching behavior. Virgin female and male mice with ablated MCH neurons were less interested in pups and maternal care. Chemogenetic and optogenetic stimulation of LHA-MCH neurons induced parental nursing in virgin female and male mice. LHA-MCH GABAergic neurons project fibres to the paraventricular hypothalamic nucleus (PVN) neurons. Optogenetic stimulation of PVN induces nursing crouching behavior along with increasing plasma oxytocin levels. The hypothalamic MCH neural relays play important functional roles in parental nursing behavior in female and male mice.

Regulation of the neural circuits that govern social behavior, including parental behavior, is a topic of immense interest in neuroscience. After parturition, females exhibit maternal care, and at the same time, maternal care exhibited by virgin females is easily activated by priming with exposure to pups 1 . However, virgin males that commit infanticide have inactive neural circuits governing parental behaviors, whereas males that have experienced mating with gestating females exhibit parental behavior 2,3 . In our previous studies conducted virgin animals, we observed that the intrinsic features of parental behavior characteristics were easily elicited in both sexes after social isolation 4 . It has been reported that the preoptic area [5][6][7][8][9] and the anteroventral periventricular nucleus (AVPV) 10 influence parental behavior, which was also observed in studies conducted using virgin animals. A subpopulation of galanin-expressing neurons was found to be activated during parenting episodes involving pup grooming and retrieving 6 . Neuronal connections from the AVPV to the paraventricular hypothalamic nucleus (PVN) were found to influence oxytocin secretion 10 . Oxytocin-secreting neurons play a crucial role in the onset and maintenance of maternal behavior in rodents [10][11][12] . Further evidence showed that PVN neurons project anatomically posterior to the lateral hypothalamic area (LHA) 13 to regulate the level of melanin concentrating hormone (MCH) 14 , a neuromodulator that integrates physiological functions [15][16][17][18][19][20][21][22][23] . Abolishing the expression of MCH will decrease maternal behaviors, especially in decreasing litter sizes 24 . Although these observations suggest that MCH neurons mediate or modulate parental nursing behaviors, further research is needed to determine how MCH neurons are projected through a detailed investigation on parental behavior regulation.
Increasing nursing couching behavior by chemogenetics and optogenetic activation of MCH neurons. Using chemogenetics, we further examined the involvement of MCH neurons, which are activated by the synthetic ligand clozapine-N-oxide (CNO) 26 in MCH-Cre mice injected with the AAV-hM3Dq-mCherry into the LHA (Fig. 3a). We found that 69.7% ± 1.9% of MCH neurons in females and 70.3% ± 6.8% in males expressed hM3Dq-mCherry (Fig. 3b,c) and c-fos (Fig. 3b,d). From the pup presentation test that began 90 min after the CNO injection, the crouching duration significantly increased in virgin MCH-Cre females and males compared with that of AAV-GFP controls (Fig. 3e). No significant difference in licking behavior (Fig. 3f) and locomotor activity ( Supplementary Fig. 4c) was observed. c-fos expression was 53.4% ± 3.3% of hM3Dq-expressing MCH neurons in females and 46.7% ± 7.3% of these in males (Fig. 3d). MCH neurons have been suggested to be important for controlling nursing crouching behavior in both virgin females and males. To confirm whether another method of MCH-neuron activation induces nursing behavior, we performed optogenetic simulation of MCH neurons (Fig. 4a). An in vitro study was performed to confirm the stimulation of channelrodopsin 2 (ChR2) expressing cells in our ( Supplementary Fig. 5a) and previous studies 23 . ChR2-enhanced yellow fluorescent protein (EYFP) was selectively expressed on MCH-producing neurons ( Fig. 4b and Supplementary Fig. 5a) in the LHA of virgin MCH-Cre females and males. Before the behavioral test, we bilaterally inserted optic fibres just above the LHA (Fig. 4a). Stimulation through the optic fibres was provided to ChR2 expressing mice and control GFP mice equally during the pup presentation test. Initially, blue light pulses (473 nm, 15 mW, 10 ms, 10 Hz) were applied. No parental behavior was observed, while the mice stayed still away from or next to the pups. This blue laser pulses (475 ± 17.5 nm, 2.5 mW,10 ms, 10 Hz at the tip of the fibre cannula) was similar to the condition of REM sleep induction which is previously reported in MCH-tTA; TetO ChR2 mice 18 . We Quantitative analysis of c-fos-ir in MCH neurons of female (left) and male (right) ddN mice that performed crouching (CR) or did not perform crouching (NCR). Mean ± S.E.M. t-test, ***p < 0.001, **p < 0.01. (c) Ablation of MCH neurons in MCH-tTA; TetO DTA bigenic mice. MCH neurons and orexin neurons in MCH-tTA; TetO DTA bigenic +/− (top) and +/+ (bottom) mice. Scale bar, 100 μm. (d) Survival rate of pups after parturition of MCH-tTA; TetO DTA bigenic +/+ and +/− female mice. Two-way ANOVA, day after parturition F(2, 72) = 5.58, **p = 0.006, Genotype F(1, 72) = 16.6, ***p < 0.0001, post-hoc test with Bonferroni, +/− vs +/+ , Day 1, p = 0.14, non-significance, Day 2, **p = 0.006, Day 3, **p = 0.008. (e) Percentage of parental behaviour or attack pups of virgin bigenic +/+ and +/− females (left) and males (right). The ratio of parental behaviour (crouching) in females was analysed by Chi-squared test, X-squared = 9.56, df = 1, p = 0.002 **. (f) Duration of crouching behaviour of virgin MCH-tTA; TetO DTA bigenic +/+ female mice when compared with +/− females. Mean ± S.E.M. t-test, ***p < 0.001. www.nature.com/scientificreports/ determined the appropriate blue laser pulses (473 nm, 10 ms, 0.5 Hz, 1 mW at the tip of the fibre cannula) to be applied through the optic fibre (COM2-DF2-500, LUCIR Inc.) during the test sessions. ChR2-EYFP was expressed in almost 40% of MCH-producing neurons ( Supplementary Fig. 5b). Pulse photo-stimulations of 1 mW (0.5 Hz) pulse to MCH neurons infused with ChR2 significantly increased crouching behavior ( Fig. 4c) but did not affect licking behavior (Fig. 4d) in females and males. We observed that the condition of optogenetic photo-stimulation required differed between males and females; photo-stimulations were started 5 min before pup presentation in males but not in females. This difference may be caused by unknown sex-dependent differences in neural circuits. During optogenetic stimulation of the LHA, which induces crouching behavior (Fig. 4c), c-fos-expressing cells were visualised in ChR2-EYFP-positive MCH neurons ( Supplementary Fig. 5a,c). Photo-stimulation did not affect locomotion in the LHA ( Supplementary Fig. 6). These results also indicate that the MCH neurons are cardinal in the regulation of parental nursing behavior.

MCH-oxytocin neuronal relay modulates nursing behavior.
Using Cre-dependent ChR2-EYFP, we observed fluorescently labelled fibres projecting from LHA-MCH to the PVN in both females and males (Fig. 5a). Projection analysis also showed that LHA-MCH neurons reached several brain regions in both sexes ( Supplementary Fig. 7) including the diagonal band broca, lateral septum and lateral preoptic area. The dense MCH neuronal EYFP fibres expressing ChR2 derived from LHA-MCH neurons projected close proximity to the oxytocin neurons of the PVN (Fig. 5a,b). MCH neurons synthesise and could utilise GABA as a neurotransmitter. GABA neurons innervate neurosecretory somata and dendrites 27,28 and as well as GABAergic synapses formed to oxytocinergic neurons 28 . We confirmed that the ChR2-expressing MCH neurons contained GABA (Fig. 5c). Moreover, we detected GAD 65-ir puncta in ChR2-EYFP fibres (Fig. 5d). GABAergic projection derived from MCH neuron could regulate the PVN-oxytocin neurons. We stereotacxically injected GABA agonist, musicimol into the PVN. We found significantly increased c-fos expressing oxytocin neurons in the socially isolated females and males of the PVN than in those of the group-housing female and male mice (Fig. 5e), indicating that the possible involvement of MCH neuronal innervation is excitatory. The PVN neurons secrete oxytocin, which contributes to maternal behavior in mice 11,12 . We then examined whether MCH neurons projecting from the LHA to the PVN were involved in the regulation of crouching  www.nature.com/scientificreports/ behavior. The optic fibres placed just above the PVN (Fig. 6a) were used to perform photo-stimulation using the same procedure of investigating LHA. During the pup presentation test, both females and males exhibited crouching behavior of longer duration than that of the controls ( Fig. 6b) but did not affect licking behavior ( Fig. 6c) or locomotion ( Supplementary Fig. 6). We observed significantly higher c-fos expression in the oxytocin neurons in the PVN of the photo-stimulation virgin female and male mice (Fig. 6d). We collected plasma samples immediately after pup presentation to measure oxytocin levels with PVN photo-stimulation, which were significantly higher in ChR2 than those in GFP females (Fig. 6e). Males showed the same trend, however, the differences between the ChR2 and GFP groups were not significant (p = 0.057).

Discussion
We elucidated that the tuned activity of MCH is entailed for nursing associated with social interactions. It also modulates MCH neural circuit characteristics onto oxytocin neurons in the PVN. The results demonstrated that the relay between LHA and PVN is important in the regulation of parenting nursing behavior in both sexes. Furthermore, increased plasma oxytocin levels after photo-simulation to the ChR2-expressed fibre in the PVN comprise the modulation of social nursing interaction. Although optogenetic conditions were virtually applied under exactly the same condition of induced REM sleep 18 , no parental behavior was observed. We determined the appropriate blue laser pulses that are of lower frequency than that of REM sleep induction. This appears to be an alternative usage of an identical neurotransmitter in different types of behavior, which argues the multiple functions of MCH neurons in the brain. The changes in neuronal excitability dependent on the behavior state 29 extend our results, prompting inherent behavior characteristics. The MCH-immunoreactive fibre (Fig. 5a,b) and its receptors are distributed in the PVN 30 [32][33][34][35] . Social isolation can be assumed as a favourable stressor in rodents 36 and prompts changing the synaptic organisation action in the brain 37 . Therefore, social isolation stress may change the mode of GABAergic excitation. The relay of this LHA neural projection to the PVN demonstrated in the present study and anatomical neural circuit from the PVN to LHA reported previously 14 , may contribute to the continued nursing crouching behavior. The LHA-PVN neural circuit is involved in the recurrent behavior 22 by possibly including continuous crouching for completing lactation. On one hand, the periaqueductal gray (PAG) in the midbrain is responsible for reproduction, i.e., lordosis 38 in sexual behavior and the maternal arched back crouching posture; on the other hand, no effect was observed on retrieval or in pup grooming in the female rat 39 . In contrast, the stimulation of galanin-expressing neurons in the medial preoptic area to the PAG projection increased the pup grooming behavior, but there was no effect on crouching in both female and male mice 8 . Alternatively, these maternal behaviors may be controlled and modulated by several neuronal inputs into the PAG from other brain regions. Although the MCH receptor in the PAG has not been confirmed 30 , it is necessary to define the possible involvement in neural circuits with neurotransmitter-derived LHA-MCH neurons projecting into the PAG in terms of motor control, which may prompt the maternal crouching behavior.
Oxytocin has been implicated in social reward 40 , and the involvement of PVN-LHA and LHA-PVN neural relay provides a rationale for explaining the continued nursing crouching behavior that sometimes persisted for a prolonged time period. Projections from the medial preoptic area to other functional brain areas have been reported to underlie parental behavior, excluding crouching behavior 8 . It is important to understand how the neural circuits that govern the social interaction of each behavior are induced or altered by social cues. Neural circuits involved in the heterogeneous components of this parental behavior have been extensively investigated at the levels of the brain area and neuromolecular materials 5,6,8,10,11,41 . It is possible that the centres governing distinct types of parental behaviors, e.g., retrieving and crouching are present in a variety of brain regions and are controlled independently. The implications of this diversity are not completely understood, thus necessitation further study to elucidate the neural integration and collaboration of each behavior involved in nursing and other parenting behaviors in mice. Each brain area and neuronal peptide must collaborate with and consolidate the multiple behavioral processes regulated by several components, resulting in forming parenting. The behavior www.nature.com/scientificreports/ centre presumed to be an integrative role in each social component of parenting behavior regulated by distinct brain regions, including diverse neuromolecular materials remains to be elucidated. For protecting their offspring, mice have several neuronal peptides that organise the neural network for assuring maternal care. The MCH receptor is expressed throughout the reward circuitry, including the nucleus accumbens, for their consistency with oxytocinergic projection of the mesolimbic system 42  Parental behavior. Each experimental female and male was placed in a clean cage with fresh bedding 2 days before testing followed by individual housing for approximately 1 week for females and 3 weeks for males. All behavioral tests were conducted during the day, 2-4 h after the lights were turned off. The observation began with the placement of three pups aged 4-7 days, in the corner opposite and farthest from the animals or the resident mouse's nest. Parental behavior was assessed as previously reported 4 . Parental nursing behavior was assessed by time spent crouching, which was defined as the limb extension and the assuming of nursing-like posture over the pups. Time spent licking was also recorded. ' Attack' was defined as biting a pup, often accompanied by observable wounds on the pup, and 'Ignore' was defined as showing no response to the pups. If the pups were attacked, they were quickly removed, and the test was terminated. Behavioral data of mice that attacked pups were excluded from analysis of maternal behavior. Mating behavior. Each female MCH-tTA; TetO DTA bigenic bigenic (+/+) and control (+/−) mouse was introduced to a WT male on the day of proestrus. The session lasted 30 min starting between 5 p.m. and 6:30 p.m. Sexually experienced WT males were randomly assigned to the bigenic females. When successful mating behavior was observed, the female was kept with the WT male after the session. If mating behavior did not occur during the session, another male was introduced on the next oestrus day.
Pup survival rate after birth. Pregnant MCH-tTA; TetO DTA bigenic females (+/+) and control mouse (+/−) were separated from the males and checked once a day for the birth of pups. The number of pups was counted once daily for the first week after birth.

Target cell activation.
Chemogenetics was employed with CNO as a synthetic ligand to activate MCH neurons in MCH-Cre mice. To selectively manipulate the MCH neurons, Gq-coupled receptor were used to modify G protein-coupled receptor hM3Dq in a Cre-recombinase-dependent manner. The adeno-associated virus vector, AAV-hSyn-FLEX-hM3Dq-mCherry (hM3-mCherry) (600 nL/injection, 3 × 10 12 copies/mL) was injected using a 27-gauge stainless steel tube and a push-pull For ChR2-mediated MCH cell photoactivation, we used recombinant AAV-EF1a-DIO-hChR2 (E123T/ T159C)-EYFP (ChR2-EYFP) vectors that were serotyped with AAV 10 coat proteins and packed by the viral vector core at Jichi Medical University. Cre-dependent ChR2-EYFP AAV vector (AAV-ChR2) (600 nL/injection, 3 × 10 12 copies/mL) was stereotactically injected into the LHA in the same manner and adjustment (Bregma: anteroposterior, − 1.5 mm; mediolateral, ± 0.9 mm; dorsoventral, − 5.0 mm) described above of virgin MCH-Cre female and male mice. All the surgical procedures were performed under anaesthesia using a ketamine-xylazine mixture (100 and 10 mg/kg i.p., respectively). The animals were injected AAV-ChR2 at 8 weeks of age and tested at 20 weeks. The female mice were housed alone starting 1 week before behavioral tests. The male mice were housed alone starting 3 weeks before the tests. The animals were fitted with optic fibre cannulas (diameter 400 μm, length 5 mm) (COME2-FTR/C-F5, LUCIR Inc., Tsukuba, Japan) placed above the LHA bilaterally (AP, L, depth from the brain surface, cannulas were angled 10° toward the outside) or singly for PVN stimulation (AP, L, depth from the brain surface) at least 4 days before the pup presentation test. The optic fibre cable was connected to an optical swivel (COME2-UFC, LUCIR, Inc.) to allow the animals to move around the test cage. Blue laser pulses (473 nm, 10 ms, 0.5 Hz, 1 mW at the tip of the fibre cannula, 400 μm diameter) were administered through the optic fibre during the test sessions. Optogenetic stimulation was performed for the same length of time as the pup presentation (15 min) for mice; for male mice, photo-stimulations were started 5 min before pup presentation. Their behavior was observed for 15 min. The same optogenetic stimulation was performed for GFPinjected females and males. After the pup presentation test, the animals were immediately anaesthetised using isoflurane to collect blood samples from the orbital sinus. The blood samples were drawn into tubes containing EDTA (1 mg/mL blood) and aprotinin (500 KIU/mL of blood, SIGMA). The samples were centrifuged and the supernatants of samples were stored at − 80 °C until hormone assays were performed. The animals were perfused 90 min after the pup-presentation test to obtain brain samples for immunohistochemistry (IHC).
Open-field assay. Each mouse was placed in the centre of a polypropylene white box (40 × 40 × 30 cm 3 ) with approximately 60 lx of illumination for 5 min under video recording. The video was automatically analysed using the Top scan L software (CleverSys, Inc., Reston, Virginia, USA) for the total distance moved.
Effect of GABA agonists. GABA agonist, muscimol (Sigma, St. Louis, MO) (10 ng) was injected using a 27-gauge stainless steel tube and a push-pull pump (Microprocessor-controlled syringe pumps) with an infusion rate of 0.04 ml/min (Bregma: anteroposterior, − 0.7 mm; mediolateral, ± 0.0 mm; dorsoventral, − 3.5 mm) using a stereotactic frame (Devid Kopf Instruments) into mice. Mice were housed in two groups; one for housing alone and another for continuous group-housing with the same sex siblings separated from litter mates at weaning 4. The animals were perfused 90 min after injection.
IHC for MCH, c-fos, oxytocin, orexin, GAD 67 and GABA. All the experimental animals were anaesthetised using an overdose of a pentobarbital sodium-xylazine mixture and perfused with 4% paraformaldehyde. Serial coronal sections (40 µm) were cut using a Microm HM 560 cryostat (Thermo Fisher Scientific, Waltham, MA, USA). For labelling, the sections were incubated overnight with anti-rabbit MCH antibodies Morphometry. Using the Keyence DC 100 digital imaging system and software (Keyence Microsystems, Osaka, Japan), we digitised images of 200 µm 2 . The number of cells was counted manually using Image J software (National Institutes of Health, Bethesda, Maryland, USA). The expression of AAV-EF1a-DIO-hChR2 (E123T/ T159C)-EYFP as an anterograde tracer in MCH neurons allowed us to trace the fibres reaching to the PVN (Fig. 5a) and other brain areas ( Supplementary Fig. 7). Images of fibres in the PVN were captured using confocal microscopy (Olympus FV 1200) and LSM 880 Airyscan (Carl Zeiss Co., Ltd.). The section with the largest sectional area was identified, and the area was compared between the animal groups.
RIA assay. Immediately after performing the behavioral test, blood samples were collected from the orbital sinus of mice anaesthetised with isoflurane. The plasma concentrations of oxytocin were determined using a radioimmunoassay with specific anti-oxytocin, as described previously 46 , after extraction with acetone and diethyl ether. The coefficients of intra-assay and inter-assay variations for oxytocin were 4% and 10%, respectively 47 . The minimum detection limit was 4 pg/mL for oxytocin.
Statistics. The sample sizes were chosen based on the common practice in animal behavior. Two trained observers blinded to the groups performed all the behavioral scorings. The degree of association between parametric factors was determined using Pearson's correlation. To compare two behavior groups or number data sets, we used a two-tailed Student's t-test. When comparing more than two groups, we used repeated two-way ANOVA with post-hoc test of Bonferroni's method (i.e., for the number of pups in Fig. 1d, the main factors were days after parturition and genotype of the subject). The ratios of behavioral responses in incidence of the categories of maternal behavior in the pup presentation tests were compared using the χ 2 test with Bonferroni post-hoc tests to examine significant differences. All the statistical analyses were performed using SPSS (version 23, IBM, Armonk, NY, USA). Statistical significance was presented as p < 0.05, p < 0.01 and p < 0.001.