Introduction

Both paternal and maternal care are critical to the survival as well as the physical and mental well-being of the offspring (He et al., 2019). Although we know a great deal about the neural mechanisms underlying maternal care, the neural substrates of paternal behavior remain elusive because of lacking ideal animal model of paternal care. Only in some monogamous rodents (e.g. prairie voles), canids and primates, do males assist and spend a great deal of energy caring for pups (Malcolm, 2015; Mendoza & Mason, 1986; Rosenfeld, Johnson, Ellersieck, & Roberts, 2013). However, some male rodents without reproductive experience also show paternal care toward alien pups, while some others ignore or even attack alien pups (Dai et al., 2022). These different pup-directed behavioral responses are based on their physiological and environmental states, and the killing of young of the conspecifics by sexually inexperienced mammals is a widespread phenomenon among different animal taxa (Lukas & Huchard, 2014). Infanticide is thought to benefit the infanticide by promoting their own reproduction (Hrdy, 1974). In laboratory mice, male mice without pairing experience typically display infanticide (Svare & Mann, 1981), but males are able to shift from infanticide to pup care when they have the opportunity to encounter their own offspring (Elwood, 1977). Contrary to our extensive understanding of the maternal circuit, little is known about the neural substrates underlying female infanticide. The virgin mandarin voles (Microtus mandarinus) naturally exhibit biparental care and infanticide in both female and male that provide ideal animal model to reveal mechanism underlying paternal care in male and infanticide in females.

OT is well-known as a key hormone for initiating and maintaining maternal care (Yoshihara, Numan, & Kuroda, 2018), which is primarily synthesized in the PVN and supraoptic nucleus (SON), among which the PVN plays an important role in initiating maternal care in rats (Munetomo, Ishii, Miyamoto, Sakuma, & Kondo, 2016). There is evidence to suggest that OT not only regulates maternal motivation, but also mediates paternal behavior (Bales, Kim, Lewis-Reese, & Sue Carter, 2004). When mouse fathers were exposed to their pups, OT neurons in the PVN were specifically activated and they also showed more aggression towards intruders to protect the own pups (Shabalova et al., 2020). Compared with rodents, similar neuropeptides and hormones are involved in paternal behavior in non-human primates (Woller et al., 2012). Similar to other mammals, paternal care only exists in a few primate species (Dulac, O’Connell, & Wu, 2014; Woller et al., 2012). A study on marmoset monkeys showed that fathers have higher levels of OT secretion in the hypothalamus than non-fathers (Woller et al., 2012), and intraventricular infusion of OT reduces the tendency of marmoset fathers to refuse to transfer food to their young offspring (Saito & Nakamura, 2011). However, where and how pathways of OT neurons regulates pup care and infanticide behavior remain largely unknown.

The mPFC is involved in attention switching, decision-making, behavioral flexibility and planning, making it potentially crucial for rapidly expressing pup care or infanticide behavior (Alsina-Llanes & Olazábal, 2021). It has been reported that the mPFC is highly activated in human mothers when they hear cry from their babies (Lorberbaum et al., 2002). The mPFC is also activated when rat mother contacts their pups first time (Fleming & Korsmit, 1996), while the damage of the mPFC disrupts pup retrieve and grooming behavior in rats (Afonso, Sison, Lovic, & Fleming, 2007). In rodents and humans, the mPFC was activated during the process of caring offspring (Hernández-González, Navarro-Meza, Prieto-Beracoechea, & Guevara, 2005; Seifritz et al., 2003). In addition, a large number of neurons in the mPFC express oxytocin receptors (OTRs) (Smeltzer, Curtis, Aragona, & Wang, 2006). OT in the circulation of mice can act on the medial prefrontal cortex (mPFC) to increase social interaction and maternal behavior (Munesue et al., 2021). Although there are some studies on the role of the mPFC in pup care, its involvement in infanticide has not yet been reported. Thus, we hypothesized that PVN OT neurons projections to mPFC may causally control paternal or infanticide behaviors.

Based on the potential antagonistic effects between pup care and infanticide behavior in neural mechanisms (Dulac et al., 2014; Kohl, Autry, & Dulac, 2017; Mei, Yan, Yin, Sullivan, & Lin, 2023), as well as the potential role of PVN-to-mPFC OT projections in pup care and infanticide. Here, a combination of methods, including immunohistochemistry, optogenetics, fiber photometry and intraperitoneal injection of OT were used to reveal neural mechanisms underlying paternal care and infanticide. We found that PVN OT neurons regulated the expression of pup care and infanticide, and further identified the involvement of PVN-to-mPFC OT projections in paternal care and infanticide. Collectively, these findings establish a regulatory role for PVN-to-mPFC OT neurons in the expression of pup-directed behaviors and suggest potential targets for the future development of intervention strategies against psychiatric disorders associated with infanticide.

Results

Pup care behavior activate more OT⁺ cells than infanticide in PVN

In order to observe the activated OT neurons in virgin voles during pup care and infanticide behaviors, we co-stained OT and c-Fos on brain slices from voles exhibiting different behaviors using immunofluorescence method (Fig. 1a). Histological analysis showed no difference in the number of OT or c-Fos positive cells between the pup care and infanticide groups of female (Fig. 1b,c, Fig. 1-source data 1) and male (Fig. 1e,f, Fig. 1-source data 1) voles. Approximately 11% (male) and 14% (female) of OT cells expressed c-Fos during pup caring, whereas only about 3% (males) and 7% (females) of OT neurons were labeled by c-Fos during infanticide (female: t (6) = 5.173, P < 0.01, Fig. 1d, Fig. 1-source data 1; male: t (6) =2.607, P < 0.05, Fig. 1g, Fig. 1-source data 1). In male and female voles, more OT neurons were activated during pup-caring than infanticide (Fig. 1d, g).

Figure 1:

Effects of optogenetic activation of PVN OT neurons on pup-directed behaviors

To reveal causal role of PVN OT neuron in regulation of pup-care and infanticide behaviors, effects of optogenetic activation of PVN OT neurons on pup-directed behaviors were investigated (Fig. 2a,b,c). Over 89% of CHR2 expression overlapped with OT neurons indicating high specificity of CHR2 virus (Fig. 2d, Figure 2-source data 1). 473 nm light simulation increase c-Fos expression in CHR2 virus infected brain region that validated the effect of optogenetic activation (Supplementary Data Fig.1a-c). We found that optogenetic activation of PVN OT cells significantly reduced latency to approach (CHR2: off vs on F (1, 7) = 11.374, P < 0.05, Fig. 2o, Figure 2-source data 1) and retrieve pups (CHR2: off vs on F (1, 4) = 14.755, P < 0.05, Fig. 2p, Figure 2-source data 1) and prolonged crouching time (CHR2: off vs on F (1, 7) = 60.585, P < 0.001, Fig. 2s, Figure 2-source data 1) in pup-care males, but had no effect on females (Fig. 2j-n, Figure 2-source data 1), nor in control virus group. Optogenetic activation of these neurons significantly reduced the latency to approach and attack pups in male (approach: CHR2: off vs on F (1, 5) = 185.509, P < 0.0001, Fig. 2h, Figure 2-source data 1; infanticide: CHR2: off vs on F (1, 5) = 59.877, P < 0.01, Fig. 2i, Figure 2-source data 1) and female voles (approach: CHR2: off vs on F (1, 6) = 64.810, P < 0.001, Fig. 2f, Figure 2-source data 1; infanticide: CHR2: off vs on F (1, 6) = 75.729, P < 0.001, Fig. 2g, Figure 2-source data 1) displaying infanticide behaviors whereas had no effect on control virus group (Fig. 2f-I, Figure 2-source data 1). These results suggest that activation of PVN OT neurons facilitate pup care behavior and significantly inhibit infanticide behavior.

Fig. 2

Effects of optogenetic inhibition of PVN OT neurons on pup-directed behaviors

To further verify the roles of PVN OT neurons on pup-induced behavior, we optogenetically inhibited OT cells by eNpHR virus and tested pup-directed behaviors (Fig. 3a,b,c). More than 90% of neurons expressing eNpHR overlapped with OT positive neurons indicating high specificity of eNpHR virus infection (Fig. 3d, Figure 3-source data 1). 589 nm light stimulation to eNpHR virus infected brain regions reduced c-fos expression verifying the effectiveness of opotogenetic inhibition via eNpHR virus (Supplementary Data Fig. 1d-f). Inhibition of PVN OT neurons showed no significant effect on pup care behavior in male and female voles that spontaneously exhibited pup caregiving behaviors (Fig. 3j-s, Figure 3-source data 1). For both male and female voles in the infanticide group, eNpHR inhibition significantly shortened the latency to approach (female: F (1, 5) = 1331.434, P < 0.0001, Fig. 3f, Figure 3-source data 1; male: F (1, 5) = 10.472, P < 0.05, Fig. 3h, Figure 3-source data 1) and attack pups (female: F (1, 5) = 291.606, P < 0.0001, Fig. 3g, Figure 3-source data 1; male: F (1, 5) = 46.901, P < 0.01, Fig. 3i, Figure 3-source data 1). These results suggest that inhibition of OT neurons in the PVN significantly facilitates infanticide behavior.

Fig. 3

Changes in OT release upon pup-directed behaviors

The results of the optogenetic manipulation demonstrated that PVN OT neurons regulated pup-induced behavior. We next detected the OT release in the mPFC during pup-induced behavior by OT1.0 sensor (Fig. 4a-c). Pup-caring female and male voles showed a significant increase (female: F (1.958, 13.708) = 45.042, P < 0.001; male: F (5, 35) = 24.057, P < 0.01) in the signal for OT1.0 sensors upon approaching (female: P < 0.01; male: P < 0.05) and retrieving (female: P < 0.01; male: P < 0.05), whereas there was no significant difference in the signal at the onset of other behaviors compared with the signal before the introduction of the pups (Fig. 4f-m, Figure 4-source data 1). In addition, we compared the signals of OT1.0 sensors in the pup-caring voles upon the first, second and third approaches to the pups (female: F (2, 14) = 10.917, P < 0.01; male: F (2, 14) = 13.351, P < 0.01), OT release peaked at the first approach and tended to decrease thereafter (Fig. 4d,e, Figure 4-source data 1). In infanticide female and male voles, OT release decreased upon attacking in infanticide males (F (1.117, 7.822) = 85.803, P < 0.001) and females (F (1.068, 7.479) =36.336, P < 0.001) (Fig. 4n-r, Figure 4-source data 1). In addition, no significant changes in signals were detected from the individuals with control AAV2/9-hSyn-OTmut during pup-directed behaviors (Supplementary Data Fig. 2). In addition, no changes in OT release were detected while subjects were exposed to object with similar size and shape as the pup (Supplementary Data Fig. 3). These results indicate that mPFC OT release significantly increased upon approaching and retrieving in pup-care voles, but decreased upon attacking pups in infanticide voles.

Fig. 4

Effects of optogenetic activation of PVN OT neurons terminals in the mPFC on pup-directed behaviors

Previous experiment found that OT release in the mPFC changed upon pup-directed behavior. To manipulate the neural circuit, we first verified oxytocin projections from PVN to the mPFC. We injected retrogradely labeled virus in the mPFC and observed the overlap of virus with OT in the PVN (Fig. 5). Then, we tested whether optogenetic activation of the PVN OT neuron projection terminals in the mPFC affects pup-induced behaviors (Fig. 6a-d). Similar to previous results, in the pup caring group, activation of the terminals facilitated approachng (F (1,5) = 23.915, P < 0.01) and retrieving (F (1,5) = 39.664, P < 0.01) in male voles (Fig. 6o-s, Figure 6-source data 1), but had no effect on females (Fig. 6j-n, Figure 6-source data 1). In male and female infanticide group voles, optogenetic activation of the PVN OT neuron projection terminals prolonged latency to approach (female: F (1,5) = 37.094, P < 0.01; male: F (1,5) = 74.718, P < 0.001) and attack pups (female: F (1,5) = 38.347, P < 0.01; male: F (1,5) = 61.589, P < 0.01) (Fig. 6f-i, Figure 6-source data 1). These results suggest that activation of the PVN OT neurons to mPFC projection promoted the onset of pup care behavior in pup-care male voles and inhibited the onset of infanticide behavior in infanticide voles.

Fig. 5

Fig. 6

Optogenetic inhibition of the PVN OT neuron projection terminals promoted infanticide

We then optogenetically suppressed the projection terminals from PVN OT neurons to mPFC and observed changes in pup-directed behaviors (Fig. 7a-d). Similar to the results of the PVN OT neurons inhibition, we found that optogenetic inhibition of the PVN OT neuron projection terminals promoted approach (F (1,5) = 119.093, P < 0.001) and infanticide (F (1,5) = 112.501, P < 0.001) in infanticide male (Fig. 7h,i, Figure 7-source data 1) and female voles (approach: F (1,5) = 280.031, P < 0.0001; retrieve: F (1,5) = 268.694, P < 0.0001) (Fig. 7f,g, Figure 7-source data 1). For pup-care male and female voles, inhibition of these terminals did not significantly affect their pup care behaviors (Fig. 7j-s, Figure 7-source data 1). These results suggest that optogenetic inhibition of the PVN OT neuron projection terminals promotes the onset of infanticide behavior in infanticide voles.

Fig. 7

Intraperitoneal injection of OT

For pre-clinic purpose, we tested the effect of peripheral administration of OT on pup-directed behavior (Fig. 8a). Delivery of OT promoted approach (t (4) = 3.737, P < 0.05) and retrieval (t (4) = 4.190, P < 0.05) in pup-care male voles (Fig. 8k,m), while it had no significant effect on the pup-directed behaviors in females (Fig. 8f-j). For infanticide voles, there was not significant prolongation of approach latency (Fig. 8d), but a significant extension of infanticide latency in males after delivery of OT (t (6) = -2.988, P < 0.05, Fig. 8e). In infanticide female voles, both the latency to approach (Z = -2.380, P < 0.05, Fig. 8b) and attack pups (t (7) = -3.626, P < 0.01, Fig. 8c) were significantly prolonged by the delivery of OT. In addition, we found that intraperitoneal injection of OT significantly reduced the number of infanticide female (χ² = 4.740, P < 0.05, Fig. 8p) and male voles (χ² = 3.039, P = 0.081, Fig. 8q). These results indicate that peripheral delivery of OT can promote the onset of pup care behavior in pup-care male voles and can significantly suppress the infanticide in both sexes. This provides a basis for the application of OT in clinic and wild life management.

Fig. 8

Discussion

In this study, we used monogamous, highly social mandarin voles to explore the role of PVN OT neurons and PVN^OT-mPFC projections in the modulation of pup care and infanticide behaviors. More OT neurons in the PVN were activated during pup care than infanticide behaviors. Optogenetic activation of the OT neurons in the PVN or OT neuron fibers in the mPFC promoted pup care and inhibited infanticide behavior, whereas inhibition of these neurons and their fibers in the mPFC promoted infanticide. In addition, intraperitoneal administration of OT promoted approach and retrieve in pup care male voles, and inhibited infanticide in both male and female voles. The present study revealed that the PVN to mPFC OT neural circuits regulate pup care and infanticide behavior in virgin mandarin voles.

Firstly, we found that more OT neurons in the PVN were activated during pup care than infanticide behaviors, which is consistent with the well-established prosocial role of OT and its ability to promote pup care behavior (Bosch & Young, 2018). Meanwhile, we found that activation of OT neurons in the PVN facilitated pup care behaviors such as approach, retrieval and crouching in pup-care male voles; whereas inhibition of OT had no effect on paternal behavior; activation and inhibition of OT neurons in the PVN had no significant effect on pup care behaviors in pup-care females; and activation of OT neurons in the PVN inhibited pup killing in infanticide voles, whereas the corresponding inhibition of OT neurons in the PVN facilitated infanticide. This finding is consistent with previous report that silencing OT neurons delayed the retrieval behavior in virgin mice (Carcea et al., 2021). Further study found that simply observing dams retrieve pups through a transparent barrier could increase retrieval behavior and PVN OT neuron activity of virgin females (Carcea et al., 2021). If OTR knockout mice are used, no pup retrieval occurs after observation, and these results suggest that input from visual signals activated OT neurons in the PVN of virgin mice allowing it to adapt to the needs of its pups and produce pup care behaviors (Carcea et al., 2021). In addition, this study further demonstrated that OT in the PVN facilitated the retrieval behavior by modulating the plasticity of the left auditory cortex and amplifying the response of mice to the pup’s call (Carcea et al., 2021). This studies also support finding from optogenetic activation of OT neurons in the PVN in the present study. The results of the present study are also supported by a recent study that chemogenetic activation OT neurons in the PVN increase pup care and reduce infanticide (Inada et al., 2022).

However, manipulation of OT neurons in the PVN produced no significant effect on pup caring in pup-care females. This may be due to the fact that female voles have inherently higher OT neural activity (Häussler, Jirikowski, & Caldwell, 1990), and female mice have more OT neurons and OT axon projections than males (Häussler et al., 1990), and that there are also significant differences in OTR expression between two sexes (Insel, Gelhard, & Shapiro, 1991; Uhl-Bronner, Waltisperger, Martínez-Lorenzana, Condes Lara, & Freund-Mercier, 2005) that possibly shows ceiling effects of the OT system. This sex difference has been reported in other study that activation of OT neurons in the PVN activated noradrenergic neurons in the locus coeruleus by co-releasing OT and glutamate, increased attention to novel objects in male rats, and that this neurotransmission was greater in males than in females (Wang, Escobar, & Mendelowitz, 2021). In the present study, we also observed that optogenetic activation of OT neurons in PVN increased crouching behavior in the pup-care male voles, but did not affect grooming time possibly via increase of OT release. This result is in line with a previous study that injection of an OTR antagonist into the MPOA of male voles significantly reduced the total duration of pup care behavior and increased the latency to approach pups and initiate paternal behavior in male voles (Yuan et al., 2019). This finding suggests that the results of peripherally administered OT and optogenetically activated OT of PVN in the present study may also have the involvement of OT-OTR interactions in MPOA. Parturition in experimental animals is accompanied by a decrease in infanticide and the emergence of pup-care behavior, and along with this process OTR expression increased not only on mammary contractile cells, but also in various regions of the brain, such as MPOA, VTA, and OB, which are all considered to be important brain regions related to the onset and maintenance of pup care behaviors (Bosch & Neumann, 2012; Shahrokh, Zhang, Diorio, Gratton, & Meaney, 2010; Yu, Kaba, Okutani, Takahashi, & Higuchi, 1996). For example, in a OB study, intra-OB injection of OT antagonist significantly delayed the onset of pup care behaviors such as retrieve, crouching, and nesting in female rats, whereas intra-OB injection of OT in virgin females induced 50% of females to show intact pup care behaviors (Yu et al., 1996). Therefore, the effects of activation of PVN OT neurons may be a result of actions on multiple brain regions involved in the expression of pup care behavior. Which brain regions that PVN OT neurons project to are involved in pups caring or infanticide needs further studies.

Optogenetic activation of the OT neural projection terminal from the PVN to the mPFC facilitated the onset of pup care behaviors, such as approach and retrieval, in pup-care male voles, whereas inhibition of this circuit had no effect; neither activation nor inhibition had a significant effect in pup-care females; and activation of this neural circuit inhibited infanticide in infanticide voles, whereas the corresponding inhibition of this circuit facilitated infanticide. In addition, we have demonstrated an increase in OT release in pup-care voles upon approaching and retrieving pups, and a decrease in OT release in infanticide voles when infanticide occurs, as recorded by OT sensors in the mPFC. There is some evidence indicating that the mPFC may be an important brain region for OT to exert its effects. In addition to expressing OTRs in the mPFC (Liu, Pappas, & Carter, 2005; Smeltzer et al., 2006), the mPFC contains OT-sensitive neurons (Ninan, 2011) and receives projections of OT neurons from the hypothalamus (Knobloch et al., 2012; Sofroniew, 1983). It has been further shown that blocking OTR in the mPFC of postnatal rats by an OTR antagonist delayed the retrieval of pups and reduced the number of pups retrieved by rats, impaired the care of pups, and decreased the latency to aggression towards intruders, increased the number of attacks, and increased anxiety in postnatal rats but had no effect on the level of anxiety in virgin rats (Sabihi, Dong, Durosko, & Leuner, 2014). These results further suggest that OT in mPFC is involved in the regulation of pup care behavior and support the results of this study in pup care voles and infanticide voles from an OTR perspective. Previous studies have shown that optical activation of the mPFC maintains aggression within an appropriate range, with activation of this brain region suppressing aggression between male mice and inhibition of the mPFC resulted in quantitative and qualitative escalation of aggression (Takahashi, Nagayasu, Nishitani, Kaneko, & Koide, 2014), which is similar to our findings with infanticide voles. For pup care behavior it has been reported that the mPFC brain region may be involved in the rapid initiation of pup care behavior in mice without pairing experience (Alsina-Llanes & Olazábal, 2020), which supported the experimental results of pup-care male voles in the present study, but the present study further suggested that OT neurons projecting to mPFC regulated pup-caring and infanticide behavior possibly via increase of OT release in the mPFC.

Another interesting finding is that peripheral OT administration promoted pup care behavior in male voles, such as approaching and retrieving pups, but had no significant effect on pup care behaviors in female voles. This result was supported by previous study that OT administered peripherally inhibited infanticide in pregnant and reproductively inexperienced females and promoted pup caring in pregnant females (McCarthy, Bare, & vom Saal, 1986). Further research has demonstrated that OT in the central nervous system also inhibited infanticide in female house mice (McCarthy, 1990). In addition, peripheral administration of OT inhibits infanticide in male mice without pairing experience (Nakahara et al., 2020). Moreover, the experience of pairing facilitated the retrieval of pups by increasing peripheral OT levels in both male and female mice (Lopatina et al., 2011). This is similar to the findings of pup-care males in the present study, whereas the absence of similar results in females may be due to differences in OT levels between the sexes (Tamborski, Mintz, & Caldwell, 2016). Pup-care female voles in this study already showed short latency to approach and retrieve pups before peripheral administration of OT. We also found that peripheral OT administration suppressed infanticide behavior in both male and female voles. It is consistent with previous study that infanticide in female house mice with no pairing experience and pregnant females with pre-existing infanticide was effectively suppressed by subcutaneous injection of OT, while injection of OT also helped pregnant females to show care for strange pups (McCarthy et al., 1986). Unpaired male mice’s aggression towards pups was accompanied by changes in the activity of the vomeronasal neurons, and as males were paired with females and lived together, the activity of these neurons decreased, accompanied by a shift from infanticide to pup care behavior (Nakahara et al., 2020). The OTR expresses throughout the vomeronasal epithelium that allows OT to potentially inhibit infanticide by reducing vomeronasal activity. The previous study found that intraperitoneal injection of OT reduced the activity of the vomeronasal, and further validated that OT modulates the activity of vomeronasal neurons by acting on the sensory epithelium to produce behavioral changes by intraperitoneal injection of an OTR antagonist that cannot cross the blood-brain barrier (Nakahara et al., 2020). It has recently been shown that peripheral OT was able to cross the blood-brain barrier into the central nervous system to act (Yamamoto et al., 2019), meaning that delivery of OT via the periphery may have increased central OT levels and thus exerted an effect. A recent study demonstrated that the secretion of OT in the brain of male mice with no pairing experience facilitated the performance of pup care behaviors and inhibited infanticide (Inada et al., 2022), which also supported the results of the present study. Similar to rodents and non-human primates, there is evidence to suggest that OT contributes to paternal care (Feldman & Bakermans-Kranenburg, 2017). Fathers with partners have higher plasma OT levels than non-fathers without partners (Mascaro, Hackett, & Rilling, 2014). Intranasal OT treatment increases fathers’ play, touch, and social interaction with their children (Weisman, Zagoory-Sharon, & Feldman, 2012). Our result provide possible application of OT in reduction of abnormality in pup-direct behavior associated with psychological diseases in human and increases of well being in wild life management.

In summary, these results indicate that the PVN to mPFC OT neural projection is involved in the regulation of pup care and infanticide behavior in virgin mandarin voles. These data provide new insights into the neural circuits underlying OT-mediated pup-directed behaviors.

Supplementary Data Fig. 1

Supplementary Data Fig. 2

Supplementary Data Fig. 3

Methods

Animals

All animal care and use were in compliance with the guidelines of the China’s Animal Welfare council and approved by Animal Welfare Committee of Shaanxi Normal University. The virgin Mandarin voles (Microtus mandarinus) used in this study were bred at Animal Center of Shaanxi Normal University and were kept at 24℃ under a 12h light-dark cycle (lights on at 8 a.m.) with food and water provided ad libitum. The stereotactic surgery was performed at the age of 8 weeks of voles. After surgery, voles were housed with their cage mates. Behavioral tests were carried out 3weeks after surgery for animal recovery and the viral infection, and 1-5 days old pups were from other breeders.

Viruses

AAV2/9-mOXT: Promoter-hCHR2(H134R)–mCherry-ER2-WPRE-pA (8.41×10¹² μg/ml) and AAV2/9-mOXT: Promoter-mCherry-pA (1.21×10¹³ μg/ml) were purchased from Shanghai Taitool Bioscience LTD. AAV2/9-mOXT-eNpHR3.0-mCherry-WPRE-hGH-pA (2.27×10¹² μg/ml) were purchased from BrainVTA (Wuhan, China) LTD. AAV2/9-hSyn-OT 1.0 (2.06×10¹² μg/ml), AAV2/9-hSyn-OTmut (2.10×10¹² μg/ml) and AAV11-EF1α-EGFP (5.00×10¹² μg/ml) were purchased from Brain Case Biotechnology LTD. All viruses were dispensed and stored at -80℃.

Immunohistochemistry

After behavioral tests, serial brain sections were harvested for histological analysis. Anesthetized voles were perfused with 40ml of PBS and 20ml of 4% paraformaldehyde. After perfusion, brains were excised and post-fixed by immersion in 4% paraformaldehyde overnight at 4℃. Brains were dehydrated in 20% and then 30% sucrose for 24h respectively before they were embedded in OCT and cryosectioned into 40 μm slices. Brain slices were rinsed with PBS (10 min) and PBST (PBS and 0.1% Triton X-100, 20 min), blocked in ready-to-use goat serum (Boster, AR0009) for 30 minutes at room temperature (RT), and then incubated overnight at 4℃ with primary antibody in PBST. The primary antibodies used were: mouse anti-OT (1:7000, Millipore, MAB5296), rabbit anti-c-Fos (1:1000, Abcam, ab190289). Following PBST washing (3×5 min), sections were incubated with secondary antibodies in PBST for 2h (RT) and stained with DAPI, and then washed once more with PBS (3×5 min). The secondary antibodies used were: goat anti-mouse Alexa Fluor 488 (1:200, Jackson ImmunoResearch, 115-545-062), goat anti-rabbit Alexa Fluor 488 (1:200, Jackson ImmunoResearch, 111-545-003), goat anti-rabbit TRITC(1:200, Jackson ImmunoResearch, 111-025-003) and ready-to-use DAPI staining solution (Boster, AR1177). Finally, the brain slices were sealed with an anti-fluorescent attenuation sealer.

Images were captured with a fluorescent microscope (Nikon) to confirm the viral expression, placements of optic fiber and viruses, and also the number of c-Fos, OT and virus positive cells. To analyze the activity of OT neurons (co-expression of c-Fos and OT) among different behaviors and the specificity of viral expression (co-expression of viruses and OT) in the PVN brain region, positive cells in the PVN were manually counted based on the Allen Mouse Brain Atlas.

Stereotactic Surgery

For optogenetic manipulation experiments, CHR2, eNpHR and mCherry-expressing control virus were stereoactically injected into the PVN (AP: -0.4 mm, ML: 0.2 mm, DV: 5.3mm) bilaterally through a Hamilton needle using nanoinjector (Reward Life Technology, KDS LEGATO 130) at 100 nl/min. For optogenetic manipulation, optic fibers (2.5 mm O.D., Reward Life Technology, China) with an appropriate exposed fiber length (PVN: 6 mm, mPFC: 3 mm) were implanted ∼100 μm above the PVN and mPFC (AP: 2.2 mm ML: 0.9, DV: 1.89 with a 20° angle lateral to middle) and secured with dental cement (Changshu ShangChi Dental Materials Co., Ltd.,202005). For fiber optometry experiment, optic fibers were inserted ∼100 μm above the mPFC (AP: 2.2 mm ML: 0.3, DV: 1.8) after injecting the OT1.0 sensor viruses. The stereotactic coordinates were determined from three-dimensional brain atlas (Chan, Kovacevíc, Ho, Henkelman, & Henderson, 2007) and the adjusted coordinates in our lab. (The stereotactic coordinates were determined by the Allen Mouse Brain Atlas and laboratory-corrected data used for voles.)

Optogenetic

Test animals were injected with 300 nL AAV2/9-mOXT: Promoter-hCHR2(H134R)–mCherry-ER2-WPRE-pA or rAAV-mOXT-eNpHR3.0-mCherry-WPRE-hGH-pA bilaterally into the PVN at 100 nl/min. Control animals were injected with AAV2/9-mOXT: Promoter-mCherry-pA in the same condition. Two weeks later, an optic fiber was implanted ∼100 μm above the PVN and mPFC bilaterally and was secured using dental cement. After surgery, animals were housed with their cage mates as before. Before the behavioral test, each implanted fiber was connected to a light laser (NewdoonInc., Aurora 300, 473 nm for activation, 589 nm for inhibition) with a 400 μm patch cord and then we introduced the test animal into center of test arena. After the test animal acclimated to the arena for at least 20 min and settled down, we placed a pup in the farthest corner away from the test animal and immediately began recording and applying light stimulation.

In pup-directed pup care behavioral test, light stimulation lasted for 11 min (3 mW, 20 Hz, 20 ms, 8 s ON and 2 s OFF) to cover the entire interaction. In the infanticide behavioral test, the stimulation lasted until the pup was removed. Each vole was tested twice successively, more than 30 min apart, once with the stimulation OFF, once ON.

To confirm that ChR2 or eNpHR3.0 stimulation indeed induced neural activation or inhibition, we used light to stimulate the brain through an optical fiber when voles were alone in their home cage, and subsequently determined neural activation or inhibition by c-Fos staining one and a half hours after light stimulation .

Fiber photometry

To record the fluorescence signals of the OT1.0 sensor during various pup-directed behaviors. Virgin voles were anesthetized with 1-2% isoflurane and immobilized on a stereotaxic device (RWD, China). Then, 300 nl AAV9-hSyn-OT1.0 or AAV9-hSyn-OTmut virus was injected into the left side of mPFC (AP: 2.2 mm ML: 0.3, DV: 1.8). After the injection, a 200 μm optical fiber was implanted ∼100 μm above the injection site and fixed with dental cement. After 2 weeks of recovery, the optic fiber was connected to the fiber photometry system (QAXK-FPS-LED, ThinkerTech, Nanjing, China) through a patch cable. To avoid bleaching the sensor, the 470 nm laser power at the tip of the fiber was adjusted to 50 μW. Voles were placed in test cages and allowed to move freely for at least 20 minutes to acclimate to the environment. Then, a pup was placed in the cage at a distance from the testing vole. If the vole retrieved the pup or ignored the pup completely or attacked the pup, gently removed the pup and introduced another pup ∼60 seconds later to stimulate more interaction. This process was repeated 3-6 times, and then the vole exhibiting pup care behavior was allowed to freely interact with the last introduced pup until the vole crouched over the pup for more than 10 s, after which the pups were removed for about 60 s and reintroduced. This latter process was repeated 3-4 times. After the pup test, we subsequently placed an object (a vole-sized plastic toy) into the cage and recorded 4 voles that investigated the object 3-6 times.

Videos were recorded with screen-recoding software to synchronize the OT1.0 fluorescence signals and pup-directed behaviors. Fluorescence signals were recorded into MATLAB mat files and analyzed with customized MATLAB code. Data were matched to a variety of behaviors towards pups based on individual trials. The change in signal was displayed as z-scored ΔF/F, which was measured by (V_signal - mean (V_basal))/std (V_basal). The V_signa and V_basal refer to the recorded values at each time point and the recorded values during the baseline period before the stimuli. The AUC (area under the curve) was calculated based on z-scored ΔF/F matching the duration of the behavior, and the AUC per second was used to compare the different fluorescence signals of behaviors and the baseline.

Behavioral paradigm and analysis

Animal behaviors in optogenetic experiments were recorded by a camera from the side of a transparent cage. ‘Approach pup’ was defined as the testing vole faced and walked right up to pup, and the latency of approach was the period from the time the pup was placed in the cage until the vole began to approach the pup. ‘Investigate pup’ was defined as the vole’s nose came into close contact with any part of the pup’s body. ‘Attack pup’ was defined as the vole attacked and bit a pup which can be recognized by the wound, and the latency to attack was the period from the time the pup was placed in the cage until the vole launched an attack. ‘Retrieve pup’ was defined as from the time a vole picked up a pup using its jaws to the time it dropped the pup at or around the nest, and the latency of retrieval was the time between the pup was put in the cage and the time the vole picked up the pup in its jaws. ‘Groom pup’ was defined as a vole combed the pup’s body surface with its muzzle, accompanied by a rhythmic up-and-down bobbing of the vole’s head and displacement of the pup. ‘Crouch’ was defined as the vole squatted quietly over the pup with no apparent movement. Pup-directed behaviors in optogenetic experiments were scored and analyzed by JWatcher.

OT treatments

Test virgin voles were acclimatized in their cages for 20 min before being injected intraperitoneally with 0.9% NaCl (1 ml/kg) and pups were introduced 30 min later to record from the side using a video camera. Thirty mins after the first record, voles were re-injected intraperitoneally OT 1 mg/kg (Leuner, Caponiti, & Gould, 2012) (Bachem, 50-56-6), and pups were introduced 30 min later to record behaviors. All behaviors were scored and analyzed using JWatcher.

Statistics

Parametric tests were used to analyze normally distributed data, and nonparametric tests were used for data that is not normally distributed according to Kolmogorov–Smirnov tests. Independent samples t-tests (two-tailed) were performed to assess number of OT, c-fos and merge rate of OT & c-fos during different pup-directed behaviors (Pup-care vs. Infanticide) and the number of c-fos-IR positive neurons (mCherry vs. CHR2; mCherry vs. eNpHR 3.0). The behavioral changes following optogenetic activation and inhibition (factors: treatment × stimulus) of PVN and mPFC were analyzed by two-way repeated-measures ANOVA. One way ANOVA were used to analyze the AUC changes recorded by the OT sensors in different behaviors. Paired-samples t-test (two-tailed) and Wilcoxin signed ranks test were used to analyze changes in pup-directed behaviors before and after intraperitoneal injection of OT. The Pearson chi-square test was used to compare the difference in the number of infanticide voles between the saline and OT groups. All data were presented as mean ± SEM, statistical analyses of data were performed using MATLAB and SPSS 22.0 software.

Conflict of interest

The authors declare no conflict of interest.

Fundings

This research was funded by the STI2030-Majior Projects grant number 2022ZD0205101, the National Natural Science Foundation of China grants numbers 32270510 and 31901082, Natural Science Foundation of Shaanxi Province, China grant number 2020JQ-412, China Postdoctoral Science Foundation grant number 2019M653534 and Fundamental Research Funds for Central University grants numbers GK202301012.