The experience of social interaction during postnatal development and adolescence is fundamental for setting the basis for social life, and the deprivation of social experience (hereby defined social isolation) impacts the survival of all species, as suggested by the adverse effects of the massive social isolation imposed by the COVID-19 crisis on mental health (Pancani et al., 2021). Identifying the possible neural mechanisms underlying the negative consequences of social isolation may help prevent and treat mental disorders.

In rodents, depending on the duration of juvenile social isolation, increased or decreased sociability in adulthood has been reported, suggesting that adolescence is a sensitive period for establishing social behavior later in life (Yamamuro et al., 2020; Yamamuro et al., 2018; Makinodan et al., 2012; Rivera-Irizarry et al., 2020). Few studies have examined the acute effects of social isolation and the cellular and circuit mechanisms that regulate the long-lasting impacts of social isolation. For example, short-term isolation has increased social interaction in rats (Niesink and van Ree, 1982).

Social interaction is generally a rewarding experience with reinforcing properties. Recent studies have highlighted the necessity and sufficiency of dopamine (DA) neurons of the ventral tegmental area (VTA) to promote social interaction (Gunaydin et al., 2014; Solié et al., 2022). Brief periods of acute social isolation have been reported to activate midbrain regions in humans (Tomova et al., 2020) and increase DA neurons’ activity within the dorsal Raphe nucleus (DRN) in mice (Matthews et al., 2016). While in rodents, 24 hr of social isolation does not change synaptic strength onto DA neurons of the VTA (Matthews et al., 2016), in humans, the response of the VTA to social cues after brief isolation is increased (Tomova et al., 2020). These data suggest that changes within VTA DA neurons may be the substrate for social craving caused by acute social isolation. The neuronal mechanisms and long-term consequences remain uninvestigated.

DA neurons of the VTA contribute to reward-seeking behavior, motivation, and reinforcement learning. Their activity is controlled upstream by several brain structures (Tomova et al., 2020), each of which may contribute to distinct behavioral aspects. DA neuron activity is controlled by glutamatergic and GABAergic inputs and tightly regulated by neuromodulators that act on G protein-coupled receptors (GPCRs). Oxytocin is a neuropeptide released by neurons within the paraventricular nucleus (PVN) of the hypothalamus that directly projects to the VTA. Oxytocin in the VTA activates oxytocin receptors on DA neurons regulates their activity, suggesting that oxytocin in the VTA gates social reward (Hung et al., 2017; Tang et al., 2014; Xiao et al., 2017). Indeed, the presence of a conspecific activates the oxytocin system, which increases oxytocin release, activates DA neurons of the VTA, and promotes the initiation and maintenance of social interaction (Resendez et al., 2020; Oettl et al., 2016). Although it has been hypothesized that oxytocin senses changes in the environment and facilitates behavioral stability to better adapt to changes (Quintana and Guastella, 2020), the role of oxytocin neurons in the behavioral consequences of social isolation remains largely unknown.

We first characterized the acute consequences of short-term social isolation on social interaction during adolescence. Male mice were weaned at postnatal day (P) 21 and then isolated between P28 and P35 (Figure 1A). On the last day, we exposed the experimental mice to an unknown sex-matched juvenile conspecific or a novel object in a direct free-interaction task (Figure 1B). Isolated mice spent more time interacting with the conspecific than the grouped control mice (Figure 1C). Conversely, object exploration did not differ between the two groups, indicating that social isolation preferentially affects social exploration (Figure 1B, D). Increased social interaction was also found when a former cage mate was presented as a social stimulus (Figure 1E, F). At the same time, it was not observed after a brief 24-hr period of social isolation, as expected from previous work (Matthews et al., 2016; Figure 1—figure supplement 1A, B). These data suggest that the duration of the isolation is an essential factor in determining the behavioral consequences of social isolation.

Figure 1 with 3 supplements see all

Download asset Open asset

Figure 1—source data 1

Figure 1 - raw data and statistical output.

https://cdn.elifesciences.org/articles/73421/elife-73421-fig1-data1-v2.xlsx

Download elife-73421-fig1-data1-v2.xlsx

To further investigate the consequences of social isolation on different aspects of social behavior, we used a three-chamber interaction task to characterize sociability and social novelty preference (Figure 1G). We found that the socially isolated and grouped mice spent significantly more time investigating a juvenile conspecific over a novel object (Figure 1H and – supplement 1 D, E, F). In the second part of the test, in contrast to the control mice, the socially isolated mice spent the same amount of time interacting with familiar and unknown conspecifics (Figure 1I and – supplement 1 G, H), indicating that social isolation affects social novelty preference. After social isolation, mice presented deficits in habituation when exposed to the same juvenile conspecific for four consecutive days (Figure 1J–N). but not when exposed to the same juvenile conspecific for four consecutive trials on the same day (Figure 1—figure supplement 1I-M). Novel object recognition (Figure 1—figure supplement 2A-F) and behavior in the EPM (Figure 1—figure supplement 2A, H-L) did not differ between the socially isolated and control mice. Moreover, social isolation does not impact the reinforcing properties of the social interaction since both groups of mice preferred the social chamber when they were subjected to social conditioned place preference (sCPP—Figure 1—figure supplement 2M-R). To investigate whether social isolation-dependent behavioral deficits are age-dependent, we studied isolated mice during adulthood (7 days of isolation; P53–P60, Figure 1—figure supplement 3A). We found that isolated adult mice spent less time interacting with a conspecific than the control mice. At the same time, object exploration (Figure 1—figure supplement 3B-D), sociability, and social novelty preference (Figure 1—figure supplement 3E-L) were no different. Altogether, these data indicate that adolescence is a critical period for developing social behavioral skills and that social isolation during this period increases social interaction at the expense of an impaired social novelty preference and altered habituation to interact with a familiar conspecific.

In humans, acute social isolation has been reported to have a rebound effect on conspecific interaction, accompanied by an increase in the response of reward circuits and, in particular, the VTA in response to social cues (Tomova et al., 2020). In line with these discoveries and further investigating the related neural mechanisms, we measured the excitability of pDA in the VTA after social isolation. Mice were isolated or maintained in group housing from P28 to P35, and acute brain slices were subsequently prepared (Figure 2A). We observed increased excitability after social isolation without a change in the resting membrane potential (Figure 2B and D). Aiming to identify the upstream key brain regions responsible for regulating the activity of DA neurons, we performed cFos analysis of brain slices seven days after social isolation. We observed an increase in PVN neurons immunopositive for cFos, suggesting increased activity (Figure 3A, B, Figure 3—figure supplement 1B). Because of their role in social behavior and the regulation of DA neuron activity (Hung et al., 2017; Xiao et al., 2017), we then focused our analysis on oxytocin neurons within the PVN. By confocal imaging quantification, we observed an increase in the density of oxytocin neurons in the PVN (Figure 3A, C, Figure 3—figure supplement 1D) and an increase of OXT⁺/cFOS⁺ double-positive cells (Figure 3A and D) after social isolation. On the contrary, social isolation does not affect the number of vasopressin (AVP) positive cells nor the AVP⁺/cFOS⁺ or OXT⁺/AVP⁺ (Figure 3—figure supplement 1C, E, H), suggesting that social isolation induces an alteration of oxytocin rather than a modification of the cell identity.

Figure 2

Download asset Open asset

Figure 2—source data 1

Figure 2 - raw data and statistical output.

https://cdn.elifesciences.org/articles/73421/elife-73421-fig2-data1-v2.xlsx

Download elife-73421-fig2-data1-v2.xlsx

Figure 3 with 1 supplement see all

Download asset Open asset

Figure 3—source data 1

Figure 3 - raw data and statistical output.

https://cdn.elifesciences.org/articles/73421/elife-73421-fig3-data1-v2.xlsx

Download elife-73421-fig3-data1-v2.xlsx

PVN neurons have been shown to regulate the activity of DA neurons in the VTA; therefore, we tested the hypothesis that these neurons are the master regulator of DA neuron activity during social isolation. To test our hypothesis, mice were first injected with CTB-488 in the VTA at P21 and then isolated between P28 and P35. After isolation, we performed patch-clamp recording from PVN neurons projecting to the VTA (PVN-VTA), and we observed increased excitability compared to that of the control group (Figure 3E–G).

To prove the causal link between oxytocin neurons and DA neuron activity during social isolation, we crossed R26-hM4Di/mCitrine mice with Oxytocin-Ires-Cre mice to express a designer inhibitor receptor-activated exclusively by designer drugs expressed under the Cre promoter in Oxt-positive cells (Figure 3H). Clozapine-N-oxide was dissolved in drinking water and administered during social isolation (or in parallel to grouped control mice). We obtained whole-cell patch-clamp recordings from pDA neurons in the VTA. We observed the rescue of excitability in cells recorded from isolated mice treated with CNO but not isolated mice treated with vehicle (Figure 3I-L, Figure 3—figure supplement 1I-L). Finally, to prove causality between neuronal hyperexcitability and behavior, we treated Oxt-hM4Di mice with CNO or vehicle. We compared the time spent interacting with a novel conspecific after social isolation (Figure 3M). We also used grouped mice treated with CNO or vehicle as controls. As expected, we found an increase in interaction time in untreated isolated versus grouped mice, while no difference was observed between mice treated with CNO independent of housing (Figure 3N). Altogether, the data presented here indicate that increased social interaction after social isolation is the consequence of the increased excitability of oxytocin neurons of the PVN and suggest that this effect is mediated by the increased activity of pDA neurons within the VTA.

We next investigated whether social isolation during adolescence has long-lasting consequences and the consequent neural mechanisms. After 7 days of social isolation from P28 to P35, the mice were regrouped, and social behavior was then tested during adulthood (Figure 4A). We still observed increased social interaction in mice isolated during adolescence compared to the control group (Figure 4B and C), but object exploration was not affected (Figure 4D). Sociability, social novelty preference (Figure 4—figure supplement 1A-I), and social habituation were similar between grouped and regrouped mice (Figure 4E, I), although during the first interaction, regrouped mice interacted more (Figure 4F). These behavioral data indicate that acute social isolation during adolescence leads to a long-lasting increase in free/unrestrained social interaction during adulthood. Remarkably, inhibition of oxytocin neuron activity during social isolation was sufficient to block the long-lasting consequences of social isolation and restore social behavior (Figure 4J–K). These data indicate that an isolation-dependent increase in social interaction occurs during adulthood and support the possible role of oxytocin neurons in regulating social craving after isolation. To investigate the related neural mechanisms underlying the long-lasting consequence of social isolation, we performed a whole-cell patch-clamp recording of pDA neurons in the VTA.

Figure 4 with 2 supplements see all

Download asset Open asset

Figure 4—source data 1

Figure 4 - raw data and statistical output.

https://cdn.elifesciences.org/articles/73421/elife-73421-fig4-data1-v2.xlsx

Download elife-73421-fig4-data1-v2.xlsx

Neuronal excitability in adulthood did not differ between isolated and control mice (Figure 4—figure supplement 2A-I), suggesting that the long-lasting behavioral consequences of social isolation, although induced by neuronal excitability, are exerted by a different neuronal mechanism. Neurons undergo other mechanisms of homeostatic adaptation to overall changes in neuronal activity, and many studies have reported in vivo scaling triggered by sensory manipulation and exerted by the regulation of calcium-permeable (CP)-AMPARs (Goel et al., 2011). We, therefore, characterized whether a long-lasting increase in free/unrestrained social interaction after social isolation during adolescence is accompanied by changes at the level of synaptic transmission at excitatory inputs onto pDA neurons in the VTA. We obtained whole-cell patch-clamp recordings from pDA neurons while pharmacologically isolating excitatory transmission. Considering the output-dependent heterogeneity of DA neurons and the previously identified neuronal type specificity in the form of experience-dependent synaptic plasticity (Bariselli et al., 2016a; Saal et al., 2003), we decided to characterize long-lasting, isolation-dependent effects on the synaptic plasticity of pDA neurons in the VTA in an output-specific manner. To that end, we injected choleratoxin in the prefrontal cortex (PFC) or nucleus accumbens (NAc). We then obtained whole-cell patch-clamp recordings from the identified neuronal population (Figure 5A, B and G). While we observed no difference in the ratio of AMPAR- and NMDAR-mediated currents between control and isolated mice (Figure 5C and D), we observed an increase in rectification index (RI) at excitatory inputs onto pDA neurons projecting to the PFC in isolated mice (Figure 5E and F). No change in the AMPA/NMDA ratio or the RI was observed at excitatory inputs onto pDA neurons projecting to the NAc (Figure 5H–K). These results led us to verify whether synaptic plasticity was already present after social isolation or if, on the contrary, it represented a specific adaptive mechanism occurring in adulthood. We injected CTB into the NAc or mPFC at P21, isolated the mice between P28 to P35, and performed whole-cell patch-clamp recordings at the end of the isolation (Figure 5—figure supplement 1A). While we observed no difference in the ratio of AMPAR- and NMDAR-mediated currents between control and isolated mice (Figure 5—figure supplement 1B-D), we observed that RI was higher in isolated compared to the control group at excitatory inputs onto pDA neurons projecting to the PFC (Figure 5—figure supplement 1E-F). No change in the AMPA/NMDA ratio or the RI was observed at excitatory inputs onto pDA neurons projecting to the NAc (Figure 5—figure supplement 1G-K). Moreover, no differences between groups were observed in the spontaneous inhibitory postsynaptic currents (sIPSCs) (Figure 5—figure supplement 1L-P), suggesting that synaptic scaling was not the consequence of changes in inhibitory transmission.

Figure 5 with 2 supplements see all

Download asset Open asset

Figure 5—source data 1

Figure 5 - raw data and statistical output.

https://cdn.elifesciences.org/articles/73421/elife-73421-fig5-data1-v2.xlsx

Download elife-73421-fig5-data1-v2.xlsx

We then tested whether the expression of CP-AMPARs on DA neurons projecting to the PFC in adulthood is the consequence of the increased excitability of oxytocin neurons in the PVN during social isolation. We isolated Oxt-hM4Di mice during adolescence, treated them with CNO or vehicle during isolation, regrouped them after 7 days until adulthood, and recorded excitatory transmission from pDA neurons projecting to the PFC from acute VTA slices (Figure 5L). Notably, we observed that RI was normalized when the activity of the oxytocin neurons was chemogenetically reduced during social isolation (Figure 5M and N). These data indicate that social isolation during adolescence leads to long-lasting effects in free/unrestrained social interaction accompanied by oxytocin neuron-dependent changes in synaptic transmission at excitatory inputs onto pDA neurons projecting to the PFC.

Finally, to assess the causal link between the presence of CP-AMPARs and isolation-dependent changes in social interaction, the VTA in each mouse was cannulated, and we injected a CP-AMPAR antagonist (NASPM) into the region 10 min before the direct interaction task (Figure 6A and B). While NASPM did not affect the interaction time in control mice, the inhibition of CP-AMPARs in the VTA was sufficient to normalize social interaction in the regrouped mice (Figure 6C). The data presented here show that the increased activity of oxytocin neurons during social interaction induces synaptic scaling exerted by the presence of CP-AMPARs and that these receptors are responsible for increased social interaction during adulthood.

Figure 6

Download asset Open asset

Figure 6—source data 1

Figure 6 - raw data and statistical output.

https://cdn.elifesciences.org/articles/73421/elife-73421-fig6-data1-v2.xlsx

Download elife-73421-fig6-data1-v2.xlsx

Social isolation is an adverse experience across social species with long-lasting behavioral and physiological consequences (Cacioppo et al., 2014). Although anxiety, depression, and obsessive-compulsive behavior generally emerge after prolonged social isolation (Teo et al., 2013; Pietrabissa and Simpson, 2020; Leigh-Hunt et al., 2017), it is evident that a short period of deprivation from conspecific interaction may also trigger aversive consequences. The adverse effects of social isolation are particularly apparent for adolescents (Almeida et al., 2021; Orben et al., 2020). Indeed, adolescence is a critical period for not only social interaction (Larsen and Luna, 2018; ) but also the emergence of psychiatric diseases (Orben et al., 2020; Marco et al., 2011), and social isolation during adolescence can exacerbate mental health problems (Larsen and Luna, 2018; Almeida et al., 2021; ; Graf et al., 2021; Andersen, 2021). Why is adolescence significantly affected, and what are the circuit mechanisms underlying the negative consequences of social isolation?

Social interaction is a basic need across species, and rodents are excellent models for investigating the neural bases of social interaction and social isolation. During adolescence, social and cognitive skills depend on peer-to-peer interactions (Vanderschuren et al., 1997; Baarendse et al., 2013), and acute social isolation in rodents is sufficient to cause behavioral abnormalities, including increased vulnerability to drug addiction and depressive-like behavior, later in life (Lampert et al., 2017; Takatsu-Coleman et al., 2013; Whitaker et al., 2013). Here, we show that 1 week of social isolation during adolescence is sufficient to generate a rebound increase in social interaction during the free-interaction task in male mice, independently whether the stimulus is a former cage mate or a novel stimulus. The increase in interaction is accompanied by deficits in social novelty preference and social habituation across days. However, social isolation does not affect short social habituation, suggesting that isolation does not affect short-term social memory (Figure 7). These data support the hypothesis that social interaction is a need and that the absence of social cues generates a craving response like that caused by food cues after fasting (Tomova et al., 2020). Indeed, we could speculate that social isolation during adolescence results in an intense urge to interact with whatever conspecific is present.

Figure 7

Download asset Open asset

Previous studies have indicated the role of DA neurons in the VTA in social behavior (Gunaydin et al., 2014; Bariselli et al., 2018; Bariselli et al., 2016b) and shown that chronic isolation alters DA levels in the NAc (Hall et al., 1998; ). Remarkably, acute social isolation during adulthood changes the synaptic strength at excitatory inputs onto DA neurons of the DRN in rodents, leaving synaptic transmission onto DA neurons of the VTA unaltered (Matthews et al., 2016). It has been shown that human exposure to social cues after acute isolation evokes activity in the VTA (Tomova et al., 2020), suggesting that while DA neurons within the DRN mediate the loneliness-like brain state induced by social isolation, DA neurons in the VTA mediate social craving. Our data indicate that 1 week of isolation during adolescence is sufficient to increase the overall excitability of neurons within the mesocorticolimbic system. We next investigated the underlying mechanisms and found that the activity of oxytocin neurons within the PVN is causally linked to neuronal excitability within the reward system and the behavioral consequences of social isolation. Previously, the release of oxytocin in the VTA was shown to increase DA neuron firing within the VTA (Xiao et al., 2017).

Furthermore, restoration of oxytocin signaling in the VTA in an autism spectrum disorder-related mouse model was sufficient to rescue social novelty responses (Hörnberg et al., 2020). Taken together, these data suggest that oxytocin neurons could represent the neural signature of social craving in male mice. Although sex differences in oxytocin and oxytocin receptor distribution have not been reported, sex differences in oxytocin’s behavioral effects can be found in rodents (Miller and Caldwell, 2015; Tamborski et al., 2016; Hammock and Levitt, 2013). Indeed, neonatal oxytocin treatment modifies the development of social behavior resulting in a reduction of alloparenting (Mogi et al., 2014) and treatment during puberty reduces social play behavior (Bredewold et al., 2014). Both these effects occur only in female. The lack of female in the present study poses a limit in our results since we do not know whether our results apply to both sexes. In the brain there are sex differences in the tissue structures and the dimorphic nuclei exhibiting morphological sex differences are considered as the structural basis for generating sex differences in brain functions (Ogawa et al., 2020). Gonadal steroids regulate Oxt activity, and among them, estrogen seems to have higher impact on the Oxt system (Richard and Zingg, 1990; Miller and Caldwell, 2015; Gimpl and Fahrenholz, 2001). Indeed, estrogens play a fundamental role in the sexually dimorphic formation of brain structures by modulating neural circuits that control sex-specific behavior, including social interaction (Ogawa et al., 2020). Future studies will aim to study the effects of juvenile social isolation in females and investigate the role of the oxytocin system in isolation-dependent changes in social behavior.

We next chose to investigate whether acute social isolation during adolescence has long-lasting consequences. Isolated mice spent more time interacting with a conspecific than control mice did 1 month after regrouping, indicating that the rebound increase in free/unrestrained social interaction in a long-lasting effect of isolation . Recently, it was reported that 2 weeks of social isolation after weaning altered the pathway from the medial PFC to the posterior paraventricular thalamus, leading to decreased sociability in adult mice (Yamamuro et al., 2020). Although the duration of social isolation can explain the differences at the behavioral level between the two studies, these data show that adolescence is a critical period for the establishment of social behavior and that isolation during this period alters sociability during adulthood.

During adulthood, increased social interaction is accompanied and causally linked to the increase in GluA2-lacking AMPARs at excitatory inputs onto pDA neurons projecting to the PFC. GluA2-lacking AMPARs are CP ionotropic receptors with larger single-channel conductance and can undergo voltage-dependent blockade by polyamines (Twomey et al., 2018; Man, 2011). These receptors are inserted at excitatory synapses after exposure to addictive drugs (Bellone and Lüscher, 2006) and have been proposed to mediate the incubation of drug craving (Conrad et al., 2008). Moreover, it has also been shown that stress induces the insertion of these receptors, and local blockade of GluA2-lacking AMPARs was shown to attenuate stress-induced behavioral changes in rodents (Kuniishi et al., 2020; Yi et al., 2017). Furthermore, the presence of these receptors changes the rules for the induction of plasticity (Mameli et al., 2011) and mediates in vivo synaptic scaling (Garcia-Bereguiain et al., 2013). Here, we show that prolonged activation of DA neurons induced by social isolation promoted the insertion of CP-AMPARs. In vivo blockade of these receptors was sufficient to rescue social isolation-induced behavioral deficits. Taken together, all these findings indicate that GluA2-lacking AMPARs are critical therapeutic targets for treating maladaptive motivated behavior and suggest that inhibitors of these receptors may counteract the negative consequences of social isolation.

By investigating the short-term and long-term consequences of social isolation during adolescence, our work contributes to our understanding of how social isolation impacts neural circuits and behavior. Since social isolation has a tremendous impact on mental health and increases vulnerability to psychiatric diseases, our work is relevant to identifying new therapeutic approaches to ameliorate maladaptive social behavior.

All data generated or analysed during this study and the statistical results are included in the manuscript and supporting files. Source data files have been provided for all the figures and figure supplements.

1. 1. Almeida I
   2. Rego JF
   3. Teixeira ACG
   4. Moreira MR

   (2021) Social isolation and its impact on child and adolescent development: a systematic review

   Revista Paulista de Pediatria 40:e2020385.

   https://doi.org/10.1590/1984-0462/2022/40/2020385

   • PubMed
   • Google Scholar
2. 1. Andersen SH

   (2021) Association of Youth Age at Exposure to Household Dysfunction With Outcomes in Early Adulthood

   JAMA Network Open 4:e2032769.

   https://doi.org/10.1001/jamanetworkopen.2020.32769

   • PubMed
   • Google Scholar
3. 1. Baarendse PJJ
   2. Counotte DS
   3. O’Donnell P
   4. Vanderschuren LJMJ

   (2013) Early social experience is critical for the development of cognitive control and dopamine modulation of prefrontal cortex function

   Neuropsychopharmacology 38:1485–1494.

   https://doi.org/10.1038/npp.2013.47

   • PubMed
   • Google Scholar
4. 1. Bariselli S
   2. Glangetas C
   3. Tzanoulinou S
   4. Bellone C

   (2016a) Ventral tegmental area subcircuits process rewarding and aversive experiences

   Journal of Neurochemistry 139:1071–1080.

   https://doi.org/10.1111/jnc.13779

   • PubMed
   • Google Scholar
5. 1. Bariselli S
   2. Tzanoulinou S
   3. Glangetas C
   4. Prévost-Solié C
   5. Pucci L
   6. Viguié J
   7. Bezzi P
   8. O’Connor EC
   9. Georges F
   10. Lüscher C
   11. Bellone C

   (2016b) SHANK3 controls maturation of social reward circuits in the VTA

   Nature Neuroscience 19:926–934.

   https://doi.org/10.1038/nn.4319

   • PubMed
   • Google Scholar
6. 1. Bariselli S
   2. Hörnberg H
   3. Prévost-Solié C
   4. Musardo S
   5. Hatstatt-Burklé L
   6. Scheiffele P
   7. Bellone C

   (2018) Role of VTA dopamine neurons and neuroligin 3 in sociability traits related to nonfamiliar conspecific interaction

   Nature Communications 9:3173.

   https://doi.org/10.1038/s41467-018-05382-3

   • PubMed
   • Google Scholar
7. 1. Bellone C
   2. Lüscher C

   (2006) Cocaine triggered AMPA receptor redistribution is reversed in vivo by mGluR-dependent long-term depression

   Nature Neuroscience 9:636–641.

   https://doi.org/10.1038/nn1682

   • PubMed
   • Google Scholar
8. 1. Bredewold R
   2. Smith CJW
   3. Dumais KM
   4. Veenema AH

   (2014) Sex-specific modulation of juvenile social play behavior by vasopressin and oxytocin depends on social context

   Frontiers in Behavioral Neuroscience 8:216.

   https://doi.org/10.3389/fnbeh.2014.00216

   • PubMed
   • Google Scholar
9. 1. Cacioppo S
   2. Capitanio JP
   3. Cacioppo JT

   (2014) Toward a neurology of loneliness

   Psychological Bulletin 140:1464–1504.

   https://doi.org/10.1037/a0037618

   • PubMed
   • Google Scholar
10. 1. Conrad KL
    2. Tseng KY
    3. Uejima JL
    4. Reimers JM
    5. Heng L-J
    6. Shaham Y
    7. Marinelli M
    8. Wolf ME

    (2008) Formation of accumbens GluR2-lacking AMPA receptors mediates incubation of cocaine craving

    Nature 454:118–121.

    https://doi.org/10.1038/nature06995

    • PubMed
    • Google Scholar
11. 1. Garcia-Bereguiain MA
    2. Gonzalez-Islas C
    3. Lindsly C
    4. Butler E
    5. Hill AW
    6. Wenner P

    (2013) In vivo synaptic scaling is mediated by GluA2-lacking AMPA receptors in the embryonic spinal cord

    The Journal of Neuroscience 33:6791–6799.

    https://doi.org/10.1523/JNEUROSCI.4025-12.2013

    • PubMed
    • Google Scholar
12. 1. Gimpl G
    2. Fahrenholz F

    (2001) The oxytocin receptor system: structure, function, and regulation

    Physiological Reviews 81:629–683.

    https://doi.org/10.1152/physrev.2001.81.2.629

    • PubMed
    • Google Scholar
13. 1. Goel A
    2. Xu LW
    3. Snyder KP
    4. Song L
    5. Goenaga-Vazquez Y
    6. Megill A
    7. Takamiya K
    8. Huganir RL
    9. Lee H-K
    10. Meuth S

    (2011) Phosphorylation of AMPA Receptors Is Required for Sensory Deprivation-Induced Homeostatic Synaptic Plasticity

    PLOS ONE 6:e18264.

    https://doi.org/10.1371/journal.pone.0018264

    • PubMed
    • Google Scholar
14. 1. Graf GHJ
    2. Biroli P
    3. Belsky DW

    (2021) Critical Periods in Child Development and the Transition to Adulthood

    JAMA Network Open 4:e2033359.

    https://doi.org/10.1001/jamanetworkopen.2020.33359

    • PubMed
    • Google Scholar
15. 1. Gunaydin LA
    2. Grosenick L
    3. Finkelstein JC
    4. Kauvar IV
    5. Fenno LE
    6. Adhikari A
    7. Lammel S
    8. Mirzabekov JJ
    9. Airan RD
    10. Zalocusky KA
    11. Tye KM
    12. Anikeeva P
    13. Malenka RC
    14. Deisseroth K

    (2014) Natural neural projection dynamics underlying social behavior

    Cell 157:1535–1551.

    https://doi.org/10.1016/j.cell.2014.05.017

    • PubMed
    • Google Scholar
16. 1. Hall FS
    2. Wilkinson LS
    3. Humby T
    4. Inglis W
    5. Kendall DA
    6. Marsden CA
    7. Robbins TW

    (1998) Isolation rearing in rats: pre- and postsynaptic changes in striatal dopaminergic systems

    Pharmacology, Biochemistry, and Behavior 59:859–872.

    https://doi.org/10.1016/s0091-3057(97)00510-8

    • PubMed
    • Google Scholar
17. 1. Hammock EAD
    2. Levitt P

    (2013) Oxytocin receptor ligand binding in embryonic tissue and postnatal brain development of the C57BL/6J mouse

    Frontiers in Behavioral Neuroscience 7:195.

    https://doi.org/10.3389/fnbeh.2013.00195

    • PubMed
    • Google Scholar
18. 1. Hörnberg H
    2. Pérez-Garci E
    3. Schreiner D
    4. Hatstatt-Burklé L
    5. Magara F
    6. Baudouin S
    7. Matter A
    8. Nacro K
    9. Pecho-Vrieseling E
    10. Scheiffele P

    (2020) Rescue of oxytocin response and social behaviour in a mouse model of autism

    Nature 584:252–256.

    https://doi.org/10.1038/s41586-020-2563-7

    • PubMed
    • Google Scholar
19. 1. Hung LW
    2. Neuner S
    3. Polepalli JS
    4. Beier KT
    5. Wright M
    6. Walsh JJ
    7. Lewis EM
    8. Luo L
    9. Deisseroth K
    10. Dölen G
    11. Malenka RC

    (2017) Gating of social reward by oxytocin in the ventral tegmental area

    Science (New York, N.Y.) 357:1406–1411.

    https://doi.org/10.1126/science.aan4994

    • PubMed
    • Google Scholar
20. 1. Kuniishi H
    2. Yamada D
    3. Wada K
    4. Yamada M
    5. Sekiguchi M

    (2020) Stress induces insertion of calcium-permeable AMPA receptors in the OFC-BLA synapse and modulates emotional behaviours in mice

    Translational Psychiatry 10:1–11.

    https://doi.org/10.1038/s41398-020-0837-3

    • PubMed
    • Google Scholar
21. 1. Lampert C
    2. Arcego DM
    3. de Sá Couto-Pereira N
    4. Dos Santos Vieira A
    5. Toniazzo AP
    6. Krolow R
    7. Garcia E
    8. Vendite DA
    9. Calcagnotto ME
    10. Dalmaz C

    (2017) Short post-weaning social isolation induces long-term changes in the dopaminergic system and increases susceptibility to psychostimulants in female rats

    International Journal of Developmental Neuroscience 61:21–30.

    https://doi.org/10.1016/j.ijdevneu.2017.05.003

    • PubMed
    • Google Scholar
22. 1. Larsen B
    2. Luna B

    (2018) Adolescence as a neurobiological critical period for the development of higher-order cognition

    Neuroscience and Biobehavioral Reviews 94:179–195.

    https://doi.org/10.1016/j.neubiorev.2018.09.005

    • PubMed
    • Google Scholar
23. 1. Leigh-Hunt N
    2. Bagguley D
    3. Bash K
    4. Turner V
    5. Turnbull S
    6. Valtorta N
    7. Caan W

    (2017) An overview of systematic reviews on the public health consequences of social isolation and loneliness

    Public Health 152:157–171.

    https://doi.org/10.1016/j.puhe.2017.07.035

    • PubMed
    • Google Scholar
24. 1. Makinodan M
    2. Rosen KM
    3. Ito S
    4. Corfas G

    (2012) A critical period for social experience-dependent oligodendrocyte maturation and myelination

    Science (New York, N.Y.) 337:1357–1360.

    https://doi.org/10.1126/science.1220845

    • PubMed
    • Google Scholar
25. 1. Mameli M
    2. Bellone C
    3. Brown MTC
    4. Lüscher C

    (2011) Cocaine inverts rules for synaptic plasticity of glutamate transmission in the ventral tegmental area

    Nature Neuroscience 14:414–416.

    https://doi.org/10.1038/nn.2763

    • PubMed
    • Google Scholar
26. 1. Man HY

    (2011) GluA2-lacking, calcium-permeable AMPA receptors--inducers of plasticity?

    Current Opinion in Neurobiology 21:291–298.

    https://doi.org/10.1016/j.conb.2011.01.001

    • PubMed
    • Google Scholar
27. 1. Marco EM
    2. Macrì S
    3. Laviola G

    (2011) Critical age windows for neurodevelopmental psychiatric disorders: evidence from animal models

    Neurotoxicity Research 19:286–307.

    https://doi.org/10.1007/s12640-010-9205-z

    • PubMed
    • Google Scholar
28. 1. Matthews GA
    2. Nieh EH
    3. Vander Weele CM
    4. Halbert SA
    5. Pradhan RV
    6. Yosafat AS
    7. Glober GF
    8. Izadmehr EM
    9. Thomas RE
    10. Lacy GD
    11. Wildes CP
    12. Ungless MA
    13. Tye KM

    (2016) Dorsal Raphe Dopamine Neurons Represent the Experience of Social Isolation

    Cell 164:617–631.

    https://doi.org/10.1016/j.cell.2015.12.040

    • PubMed
    • Google Scholar
29. 1. Miller TV
    2. Caldwell HK

    (2015) Oxytocin during Development: Possible Organizational Effects on Behavior

    Frontiers in Endocrinology 6:76.

    https://doi.org/10.3389/fendo.2015.00076

    • PubMed
    • Google Scholar
30. 1. Mogi K
    2. Ooyama R
    3. Nagasawa M
    4. Kikusui T

    (2014) Effects of neonatal oxytocin manipulation on development of social behaviors in mice

    Physiology & Behavior 133:68–75.

    https://doi.org/10.1016/j.physbeh.2014.05.010

    • PubMed
    • Google Scholar
31. 1. Motulsky HJ
    2. Brown RE

    (2006) Detecting outliers when fitting data with nonlinear regression - a new method based on robust nonlinear regression and the false discovery rate

    BMC Bioinformatics 7:123.

    https://doi.org/10.1186/1471-2105-7-123

    • PubMed
    • Google Scholar
32. 1. Niesink RJ
    2. van Ree JM

    (1982) Short-term isolation increases social interactions of male rats: A parametric analysis

    Physiology & Behavior 29:819–825.

    https://doi.org/10.1016/0031-9384(82)90331-6

    • PubMed
    • Google Scholar
33. 1. Oettl L-L
    2. Ravi N
    3. Schneider M
    4. Scheller MF
    5. Schneider P
    6. Mitre M
    7. da Silva Gouveia M
    8. Froemke RC
    9. Chao MV
    10. Young WS
    11. Meyer-Lindenberg A
    12. Grinevich V
    13. Shusterman R
    14. Kelsch W

    (2016) Oxytocin Enhances Social Recognition by Modulating Cortical Control of Early Olfactory Processing

    Neuron 90:609–621.

    https://doi.org/10.1016/j.neuron.2016.03.033

    • PubMed
    • Google Scholar
34. 1. Ogawa S
    2. Tsukahara S
    3. Choleris E
    4. Vasudevan N

    (2020) Estrogenic regulation of social behavior and sexually dimorphic brain formation

    Neuroscience and Biobehavioral Reviews 110:46–59.

    https://doi.org/10.1016/j.neubiorev.2018.10.012

    • PubMed
    • Google Scholar
35. 1. Orben A
    2. Tomova L
    3. Blakemore SJ

    (2020) The effects of social deprivation on adolescent development and mental health

    The Lancet. Child & Adolescent Health 4:634–640.

    https://doi.org/10.1016/S2352-4642(20)30186-3

    • PubMed
    • Google Scholar
36. 1. Pancani L
    2. Marinucci M
    3. Aureli N
    4. Riva P

    (2021) Forced Social Isolation and Mental Health: A Study on 1,006 Italians Under COVID-19 Lockdown

    Frontiers in Psychology 12:663799.

    https://doi.org/10.3389/fpsyg.2021.663799

    • PubMed
    • Google Scholar
37. 1. Pietrabissa G
    2. Simpson SG

    (2020) Psychological Consequences of Social Isolation During COVID-19 Outbreak

    Frontiers in Psychology 11:2201.

    https://doi.org/10.3389/fpsyg.2020.02201

    • PubMed
    • Google Scholar
38. 1. Quintana DS
    2. Guastella AJ

    (2020) An Allostatic Theory of Oxytocin

    Trends in Cognitive Sciences 24:515–528.

    https://doi.org/10.1016/j.tics.2020.03.008

    • PubMed
    • Google Scholar
39. 1. Resendez SL
    2. Namboodiri VMK
    3. Otis JM
    4. Eckman LEH
    5. Rodriguez-Romaguera J
    6. Ung RL
    7. Basiri ML
    8. Kosyk O
    9. Rossi MA
    10. Dichter GS
    11. Stuber GD

    (2020) Social Stimuli Induce Activation of Oxytocin Neurons Within the Paraventricular Nucleus of the Hypothalamus to Promote Social Behavior in Male Mice

    The Journal of Neuroscience 40:2282–2295.

    https://doi.org/10.1523/JNEUROSCI.1515-18.2020

    • PubMed
    • Google Scholar
40. 1. Richard S
    2. Zingg HH

    (1990) The human oxytocin gene promoter is regulated by estrogens

    The Journal of Biological Chemistry 265:6098–6103.

    https://doi.org/10.1016/S0021-9258(19)39297-X

    • PubMed
    • Google Scholar
41. 1. Rivera-Irizarry JK
    2. Skelly MJ
    3. Pleil KE

    (2020) Social Isolation Stress in Adolescence, but not Adulthood, Produces Hypersocial Behavior in Adult Male and Female C57BL/6J Mice

    Frontiers in Behavioral Neuroscience 14:129.

    https://doi.org/10.3389/fnbeh.2020.00129

    • PubMed
    • Google Scholar
42. 1. Saal D
    2. Dong Y
    3. Bonci A
    4. Malenka RC

    (2003) Drugs of abuse and stress trigger a common synaptic adaptation in dopamine neurons

    Neuron 37:577–582.

    https://doi.org/10.1016/s0896-6273(03)00021-7

    • PubMed
    • Google Scholar
43. 1. Solié C
    2. Girard B
    3. Righetti B
    4. Tapparel M
    5. Bellone C

    (2022) VTA dopamine neuron activity encodes social interaction and promotes reinforcement learning through social prediction error

    Nature Neuroscience 25:86–97.

    https://doi.org/10.1038/s41593-021-00972-9

    • PubMed
    • Google Scholar
44. 1. Takatsu-Coleman AL
    2. Patti CL
    3. Zanin KA
    4. Zager A
    5. Carvalho RC
    6. Borçoi AR
    7. Ceccon LMB
    8. Berro LF
    9. Tufik S
    10. Andersen ML
    11. Frussa-Filho R

    (2013) Short-term social isolation induces depressive-like behaviour and reinstates the retrieval of an aversive task: mood-congruent memory in male mice?

    Journal of Psychiatry & Neuroscience 38:259–268.

    https://doi.org/10.1503/jpn.120050

    • PubMed
    • Google Scholar
45. 1. Tamborski S
    2. Mintz EM
    3. Caldwell HK

    (2016) Sex Differences in the Embryonic Development of the Central Oxytocin System in Mice

    Journal of Neuroendocrinology 28:E364.

    https://doi.org/10.1111/jne.12364

    • PubMed
    • Google Scholar
46. 1. Tang Y
    2. Chen Z
    3. Tao H
    4. Li C
    5. Zhang X
    6. Tang A
    7. Liu Y

    (2014) Oxytocin activation of neurons in ventral tegmental area and interfascicular nucleus of mouse midbrain

    Neuropharmacology 77:277–284.

    https://doi.org/10.1016/j.neuropharm.2013.10.004

    • PubMed
    • Google Scholar
47. 1. Teo AR
    2. Lerrigo R
    3. Rogers MAM

    (2013) The role of social isolation in social anxiety disorder: A systematic review and meta-analysis

    Journal of Anxiety Disorders 27:353–364.

    https://doi.org/10.1016/j.janxdis.2013.03.010

    • PubMed
    • Google Scholar
48. 1. Tomova L
    2. Wang KL
    3. Thompson T
    4. Matthews GA
    5. Takahashi A
    6. Tye KM
    7. Saxe R

    (2020) Acute social isolation evokes midbrain craving responses similar to hunger

    Nature Neuroscience 23:1597–1605.

    https://doi.org/10.1038/s41593-020-00742-z

    • PubMed
    • Google Scholar
49. 1. Twomey EC
    2. Yelshanskaya MV
    3. Vassilevski AA
    4. Sobolevsky AI

    (2018) Mechanisms of Channel Block in Calcium-Permeable AMPA Receptors

    Neuron 99:956–968.

    https://doi.org/10.1016/j.neuron.2018.07.027

    • PubMed
    • Google Scholar
50. 1. Vanderschuren LJ
    2. Niesink RJ
    3. Van Ree JM

    (1997) The neurobiology of social play behavior in rats

    Neuroscience and Biobehavioral Reviews 21:309–326.

    https://doi.org/10.1016/s0149-7634(96)00020-6

    • PubMed
    • Google Scholar
51. 1. Whitaker LR
    2. Degoulet M
    3. Morikawa H

    (2013) Social deprivation enhances VTA synaptic plasticity and drug-induced contextual learning

    Neuron 77:335–345.

    https://doi.org/10.1016/j.neuron.2012.11.022

    • PubMed
    • Google Scholar
52. 1. Xiao L
    2. Priest MF
    3. Nasenbeny J
    4. Lu T
    5. Kozorovitskiy Y

    (2017) Biased Oxytocinergic Modulation of Midbrain Dopamine Systems

    Neuron 95:368–384.

    https://doi.org/10.1016/j.neuron.2017.06.003

    • PubMed
    • Google Scholar
53. 1. Yamamuro K
    2. Yoshino H
    3. Ogawa Y
    4. Makinodan M
    5. Toritsuka M
    6. Yamashita M
    7. Corfas G
    8. Kishimoto T

    (2018) Social Isolation During the Critical Period Reduces Synaptic and Intrinsic Excitability of a Subtype of Pyramidal Cell in Mouse Prefrontal Cortex

    Cerebral Cortex (New York, N.Y 28:998–1010.

    https://doi.org/10.1093/cercor/bhx010

    • PubMed
    • Google Scholar
54. 1. Yamamuro K
    2. Bicks LK
    3. Leventhal MB
    4. Kato D
    5. Im S
    6. Flanigan ME
    7. Garkun Y
    8. Norman KJ
    9. Caro K
    10. Sadahiro M
    11. Kullander K
    12. Akbarian S
    13. Russo SJ
    14. Morishita H

    (2020) A prefrontal-paraventricular thalamus circuit requires juvenile social experience to regulate adult sociability in mice

    Nature Neuroscience 23:1240–1252.

    https://doi.org/10.1038/s41593-020-0695-6

    • PubMed
    • Google Scholar
55. 1. Yi ES
    2. Oh S
    3. Lee JK
    4. Leem YH

    (2017) Chronic stress-induced dendritic reorganization and abundance of synaptosomal PKA-dependent CP-AMPA receptor in the basolateral amygdala in a mouse model of depression

    Biochemical and Biophysical Research Communications 486:671–678.

    https://doi.org/10.1016/j.bbrc.2017.03.093

    • PubMed
    • Google Scholar

Decision letter after peer review:

Thank you for submitting your article "Oxytocin neurons control the effects of social isolation via the mesocortical pathway" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Kate Wassum as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Michy Kelly (Reviewer #1).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission. In addition to addressing the concerns raised in the public review, please address each of the essential revisions provided below.

Essential revisions:

1) All reviewers noted that the lack of females profoundly lessen the impact of the findings. Ideally, all experiments would need to be conducted in females. However, the reviewers understand that this may not be feasible depending on the availability of animals and staffing levels. If this is the case, it is strongly recommended that the authors at the least acknowledge this significant shortcoming and add extensive discussion related to the effect of oxytocin in females (please see Caldwell, Curr. Op. Behavior. Sci, 2018).

2) There was consensus among the reviewers that the statistical analyses are not appropriate for the experimental design and this also needs to be addressed thoroughly. In particular:

a. Parametric statistics are based on assumptions of normality and equal variance, yet only a test of normality is mentioned and only for t-tests and one way ANOVAs. It is stated that normality was assumed for datasets requiring multifactorial ANOVAs. Please assess each dataset for both normality and equal variance before using parametric statistics. 2way ANOVAs frequently fail one or the other assumption, thus necessitating the employ of multiple non-parametric tests to assess the various factors (or, alternatively, a non-parametric multifactorial ANOVA if you have access to custom analysis software). This is a particular concern because in many datasets, variance appears to systematically increase with increasing mean (usually a sign of failed equal variance).

b. Throughout the manuscript, 2-way ANOVAs are reported for behavioral tasks where 2-way repeated measure ANOVAs should have been used (e.g., in the case of tasks tracking performance in the same subject across multiple compartments or objects). In the ephys data, 2 WAY RM ANOVA were used as appropriate; however, it is surprising there was not a significant interaction for many of those experiments (i.e., only main effects were reported) given the distinct overlap at "0". Was the "0" data not included in the RM ANOVA?

c. Throughout the manuscript, only main effects are reported when a significant interaction between effects would be required for the post hoc testing conducted and conclusions drawn. For example, in Figure 1G, the main effect of chamber that is reported says that across the experimental groups, novel > familiar. To statistically make a conclusion that the group vs isolated mice performed differently, you would need a significant interaction between chamber x housing condition to justify the post hoc tests examining the effect of chamber within group housed and chamber within isolated mice. In Figure 1J as well, you are trying to conclude the housing conditions perform differently across days. But to say that, the 2way rm would need to yield a significant interaction of housing condition x day to warrant a post hoc analysis of days within group housed and days within isolated. Again, this is not limited to Figure 1, I just use them here as examples.

d. Many 2-way ANOVAs have fractional df's reported. In the case of linear models, it is my understanding df's should be whole numbers, so it is not clear why dfs are presented in many 2way analyses as decimals (e.g., "2.027, 20.27). I believe in more complicated analyses, unequal variances can yield fractional df's, but in the case of unequal variance, a non-parametric statistic should be employed. Would you please clarify or revise?

e. The method of data presentation in Figure 1 S2 is not clear. Typically, "preference index" is a single number that incorporates relative preference for a novel object as a function of total time of exploration (=n-f/n+f). I do not believe it would be statistically appropriate to compare normalized data for ob1 vs ob2 because the combined value of both factors would be the same between groups (i.e., there could be no effect of group). The more typical way of analyzing these data is to compare the preference index for the novel object between the group vs. isolated mice to determine if there is a group effect, and then to conduct a one-sample t-test for each group's index vs.5 to determine if they exhibited significant memory (i.e., an index that differed significantly from chance). The authors may refer to the following paper for guidance:

https://www.sciencedirect.com/science/article/pii/S1074742714000070. If you were to analyze the raw seconds of exploration, then you could do a paired analyses of O1 vs O2, that said, the analysis should be a 2factor RM with housing condition vs object as factors. The same applies to the analyses of closed vs open arms.

f. Legend for Figure 3L is missing the post hoc analyses to say which group is different from which. Further, it seems equal variance could possibly fail given that variability increases in parallel with the increases mean (also with study in 3D).

g. In Figure 3 S1B, was there not a significant interaction? It seems there is an effect of increasing intensity within the no CNO group, but not the CNO group. Note that a significant interaction overrides main effects as it states that the effect of one factor depends on the level of the other factor.

3) It is not clear how oxytocin neurons were identified in Figure 3 S1. With current details provided, it appears that recordings of all PVN neurons (non-oxytocin neurons as well) were obtained in mice; please clarify. Moreover, cFOS+ and Oxt+ coexpression images are needed to relate the increase in the activity of PVN (via the IEG proxy) and the density of Oxt+ cells in PVN (figure 3 and page 4, line 11-17). Also, please clarify whether neurons are increasing Oxt production to become more efficiently labeled or changing identity of their synthesized neuropeptides?

4) Although it is reasonable to suggest that the lack of preference for the novel mouse and the failure to habituate to a mouse could reflect a reduced preference for novelty (and is that because they are less motivated to explore novel social stimuli or because they find social stimuli less rewarding?). This could also/alternatively reflect a social learning and/or memory deficit (they fail to form a memory of the first-presented mouse). To disentangle these seemingly opposite consequences, typically habituation sessions are conducted for a longer period of time and behavior examined across time within the same day (e.g., Hedge et al., 2016, Neuroscience https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5031549/).

5) The extensive and definitive use of the term "social craving" throughout is unjustified based on the data presented (e.g., there is no breakpoint data shown). Craving implies having experienced something that is rewarding and wanting more of it; however, these mice never experienced any social interactions with non-cagemates, so how can they crave it? To this end, it might be informative to test these behaviors using cage mates as stimulus mice.

6) Regarding the statement "Indeed, we could speculate that social isolation during adolescence results in an intense urge to interact with whatever conspecific is present." This can be tested in a straightforward fashion by using cage mates as stimulus mice, or measuring interactions in the home cage, instead of using novel mice.

7) It is important to clarify why stimulus mice were 1 week younger and whether this is this true for adolescent and adult studies? This is of particular relevance because in the adult study illustrations, a small (seemingly juvenile) mouse is drawn--which would suggest much more than 1 week younger in the case of testing adults. If adults were tested with juvenile mice, as the illustration implies, it complicates interpretation of the adult isolation study since the differential effect of isolation may not be related to the period of isolation but rather a fundamentally different preference between adolescent and adult mice for other juvenile mice.

8) The authors identified putative DA neurons based on position, morphology, and capacitance, while providing sparse details in the methods. This section should contain detailed information and/or a small post hoc immunohistochemistry validated dataset.

9) The authors have exclusively shown scaled-up excitatory responses by pharmacologically blocking GABAA receptors, determining synaptic scaling after social isolation without considering potential effects on inhibitory post synaptic responses. Please provide prediction for the latter. Furthermore, it is not clear if synaptic scaling reported in Figure 5 induced by adolescent isolation only emerges slowly in the adult, or it is already acutely happening at the end of isolation at p35. Please discuss in detail.

10) It is unclear how oxytocin neuron activity contributes to synaptic plasticity onto VTA-NAc projections. One possibility is that oxytocin expressing PVN neurons projecting to VTA also release glutamate and contribute to the synaptic plasticity. It would be informative to optogenetically examine the synaptic connectivity between PVN->VTA projections and VTA->NAc projection neurons.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Oxytocin neurons mediate the effect of social isolation via the VTA circuits" for further consideration by eLife. Your revised article has been evaluated by Kate Wassum (Senior Editor) and a Reviewing Editor.

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

1) Please acknowledge the lack of females as a shortcoming of the study and provide more discussion of this limitation; e.g., that differences in the system in males vs females and how these findings here may only apply to the male brain. Please also state the sex of the subjects in the abstract.

2) The post hoc analyses referred to in 2f of the letter are still missing. The statistical interaction was reported, but the post hoc tests were not, nor were significance markers added to the figure to indicate at which pA the groups differ.

3) Please make sure all statistics are reported. For instance, in some cases a 2 way ANOVA was used but only one main effect is reported.

https://reviewer.elifesciences.org/author-guide/full "Report exact p-values wherever possible alongside the summary statistics and 95% confidence intervals. These should be reported for all key questions and not only when the p-value is less than 0.05."

4) As isolated mice show increased social interaction during direct social interaction, but not during the 3 chamber sociability assay, it is important to more thoroughly explore why isolated mice show different outcomes between these two tests. To better understand the cause of increased direct social interaction (social craving vs aggression), it would be informative to conduct additional behavioral testings from isolated mice (e.g. social CPP, aggression assay). If new behavior testing data such as social CPP or aggression assay (resident intruder etc) is not possible, at the very least a more detailed behavioral characterization such as chasing during free social interaction test is needed to address this concern and better distinguish craving vs aggressive encounter. Overall, more careful consideration is required to interpret the nature social behavior in isolated mice captured by free social interaction and 3 chamber task.

5) The response to essential review 2c to statistically compare between groups was not fully addressed. [Results] page 1, lines 18-19 and Figure 1EF. To evaluate if the house condition modifies the sociability, it is essential to run these statistics for Figure 1HI. Related to this, the 3 chamber testing data should be presented and analyzed in a similar way across different sets of data. E.g., Figure 1 HI are missing center chamber time and the preference index data while all the rest of the 3 chamber data include these.

In general, more clarity on whether or not the 3-chamber social preference test results do or do not support the conclusion of social isolation leading to an increase in social interaction would help. E.g., by clarifying whether isolation leads to an increase in social interaction specifically in "free/unrestrained"social interactions in adulthood.

Essential revisions:

1) All reviewers noted that the lack of females profoundly lessen the impact of the findings. Ideally, all experiments would need to be conducted in females. However, the reviewers understand that this may not be feasible depending on the availability of animals and staffing levels. If this is the case, it is strongly recommended that the authors at the least acknowledge this significant shortcoming and add extensive discussion related to the effect of oxytocin in females (please see Caldwell, Curr. Op. Behavior. Sci, 2018).

We thank the reviewers for this comment. We agree that results from the female cohort would increase the impact of the findings; however, these experiments would take a considerable amount of time and be a matter of future study. As suggested by the reviewers, we have now added a comment about this point (p 9 lines 8 to 15)

2) There was consensus among the reviewers that the statistical analyses are not appropriate for the experimental design and this also needs to be addressed thoroughly. In particular:

We apologize for the inappropriate statistical analysis. In the new version of the paper we have revised all the statistical analysis and for each “Source data” file, we have now added the output of all statistical tests that were used.

a. Parametric statistics are based on assumptions of normality and equal variance, yet only a test of normality is mentioned and only for t-tests and one way ANOVAs. It is stated that normality was assumed for datasets requiring multifactorial ANOVAs. Please assess each dataset for both normality and equal variance before using parametric statistics. 2way ANOVAs frequently fail one or the other assumption, thus necessitating the employ of multiple non-parametric tests to assess the various factors (or, alternatively, a non-parametric multifactorial ANOVA if you have access to custom analysis software). This is a particular concern because in many datasets, variance appears to systematically increase with increasing mean (usually a sign of failed equal variance).

We thank the reviewer for the comment. We now used a parametric test only if both normality and equal variance were not violated in the revised statistical analysis. If the contrary, we used a non-parametric test. Regarding multifactorial ANOVAs, we first checked normality, and all the analysis were then performed assuming no sphericity, which implies Geisser-Greenhouse correction. We have now included these details in Material and methods, and all the stats have been upgraded in the figures and legends.

b. Throughout the manuscript, 2-way ANOVAs are reported for behavioral tasks where 2-way repeated measure ANOVAs should have been used (e.g., in the case of tasks tracking performance in the same subject across multiple compartments or objects). In the ephys data, 2 WAY RM ANOVA were used as appropriate; however, it is surprising there was not a significant interaction for many of those experiments (i.e., only main effects were reported) given the distinct overlap at "0". Was the "0" data not included in the RM ANOVA?

We apologize for the mistake, and we have now used the appropriate statistical test for each dataset. In all the figures we have now reported the F value for the “interaction” between variables, and as written before, each “Data source” file contains all statistical analysis outputs.

c. Throughout the manuscript, only main effects are reported when a significant interaction between effects would be required for the post hoc testing conducted and conclusions drawn. For example, in Figure 1G, the main effect of chamber that is reported says that across the experimental groups, novel > familiar. To statistically make a conclusion that the group vs isolated mice performed differently, you would need a significant interaction between chamber x housing condition to justify the post hoc tests examining the effect of chamber within group housed and chamber within isolated mice. In Figure 1J as well, you are trying to conclude the housing conditions perform differently across days. But to say that, the 2way rm would need to yield a significant interaction of housing condition x day to warrant a post hoc analysis of days within group housed and days within isolated. Again, this is not limited to Figure 1, I just use them here as examples.

We have now modified the statistical analysis according to what we have recently published (Mol Psychiatry. 2022 Jan 12. doi: 10.1038/s41380-021-01427-0; Eur J Neurosci. 2021 May;53(9):3199-3211. doi: 10.1111/ejn.15179). Since the aim of the 3-chamber task is to evaluate if a mouse prefers a social stimulus over an object and a novel social stimulus over a familiar one, we decided to perform a stand-alone analysis for each group of mice. In particular, we used paired t-test analysis for interaction time (new Figure1 H-I and new Figure1-Supplement 1E, G), One-way RM ANOVA for time in the chamber (new Figure1-Supplement 1D, F), and unpaired samples t-test for the distance moved (new Figure1-Supplement 1C). The 3-chamber task for Figure1-Supplement 2 and Figure4-Supplement 1 were analyzed in the same way.

Regarding the habituation task (new Figure1J-N, new Figure 4E-I, and new Figure1-Supplement 1H-L ), we now have included the interaction effect for the 2-way RM ANOVA and we added new panels comparing the interaction time between DAY1 (or TRIAL1) and DAY4 (or TRIAL4).

d. Many 2-way ANOVAs have fractional df's reported. In the case of linear models, it is my understanding df's should be whole numbers, so it is not clear why dfs are presented in many 2way analyses as decimals (e.g., "2.027, 20.27). I believe in more complicated analyses, unequal variances can yield fractional df's, but in the case of unequal variance, a non-parametric statistic should be employed. Would you please clarify or revise?

We thank the reviewer for this comment. The presence of fractional df’s is due to the Geisser-Greenhouse correction. We now added all the statistical details in the material and methods section.

e. The method of data presentation in Figure 1 S2 is not clear. Typically, "preference index" is a single number that incorporates relative preference for a novel object as a function of total time of exploration (=n-f/n+f). I do not believe it would be statistically appropriate to compare normalized data for ob1 vs ob2 because the combined value of both factors would be the same between groups (i.e., there could be no effect of group). The more typical way of analyzing these data is to compare the preference index for the novel object between the group vs. isolated mice to determine if there is a group effect, and then to conduct a one-sample t-test for each group's index vs.5 to determine if they exhibited significant memory (i.e., an index that differed significantly from chance). The authors may refer to the following paper for guidance:

https://www.sciencedirect.com/science/article/pii/S1074742714000070. If you were to analyze the raw seconds of exploration, then you could do a paired analyses of O1 vs O2, that said, the analysis should be a 2factor RM with housing condition vs object as factors. The same applies to the analyses of closed vs open arms.

We thank the reviewer for the comment; however we would like to point out that using “n/(n+f)” (the one that we use in our paper) or “(n-f)/(n+f)” as preference index will give the same statistically results. Indeed, the latter is a mathematical transposition of the first equation. However, since a mouse has a 50% of chance to explore one or the other object, we believe that our way to calculate preference index is more intuitive because if a mouse explores for the same amount of time the two objects, the preference index will be equal to 0.5, which correspond to the chance level. Our analysis aimed to demonstrate that both groups of mice were able to prefer a novel object over a familiar one and not to state that one group performed better compared to the other. Indeed, we first show no bias toward one of the two objects during the familiarization phase. Certainly, the paired t-test between the preference index for objects one and two is not significantly different. Our analysis mathematically gives the same result as calculating a one-sample t-test against the chance level. Since the social novelty was impaired in the three-chamber task in social isolated mice, our analysis of the NORT wants to determine if isolated mice discriminate a novel versus a familiar object and not compare performances between groups. We applied the same analysis for the EPM.

f. Legend for Figure 3L is missing the post hoc analyses to say which group is different from which. Further, it seems equal variance could possibly fail given that variability increases in parallel with the increases mean (also with study in 3D).

We revised the statistical analysis, and we decided to use a three-way RM ANOVA since we have three variables (current step, house condition, and CNO treatment). Moreover, we slightly increase the n for some groups to have a more homogeneous number.

g. In Figure 3 S1B, was there not a significant interaction? It seems there is an effect of increasing intensity within the no CNO group, but not the CNO group. Note that a significant interaction overrides main effects as it states that the effect of one factor depends on the level of the other factor.

We are sorry for the lack of information. We now added all the statistical analysis details.

3) It is not clear how oxytocin neurons were identified in Figure 3 S1. With current details provided, it appears that recordings of all PVN neurons (non-oxytocin neurons as well) were obtained in mice; please clarify. Moreover, cFOS+ and Oxt+ coexpression images are needed to relate the increase in the activity of PVN (via the IEG proxy) and the density of Oxt+ cells in PVN (figure 3 and page 4, line 11-17). Also, please clarify whether neurons are increasing Oxt production to become more efficiently labeled or changing identity of their synthesized neuropeptides?

We are sorry for the lack of information in the previous version of the manuscript. We have now added in the material and methods section the relative details. The OXT cells expressing hM4Di channel can be recognized thanks to the mCitrine reporter present on the hM4Di transgene Thus, in new Figure 3S1 J-L, we report excitability recorded from identified OXT neurons. Moreover, in the new figure 3-S1, we now added the immunohistochemical validation of the OXT-hM4Di mouse line (new figure 3-S1I).

For the cFOS/OXT analysis, we have now repeated the experiment, and we performed triple staining with three different antibodies against cFOS, OXT, and AVP, and analyzed the relative and absolute expression of the three proteins as well as the co-expression of cFOS-OXT, cFOS-AVP and OXT-AVP (new Figure 3A-D and new figure 3-S1A-G). Our new analysis shows an increase in the absolute number of OXT+ and cFOS + cells in isolated mice and an increase in the double-positive cells (OXT+ /cFOS + cells). Furthermore, since no differences were found in AVP expression or OXT-AVP colocalization, the results suggest that PVN neurons from isolated mice produced oxytocin more efficiently.

4) Although it is reasonable to suggest that the lack of preference for the novel mouse and the failure to habituate to a mouse could reflect a reduced preference for novelty (and is that because they are less motivated to explore novel social stimuli or because they find social stimuli less rewarding?). This could also/alternatively reflect a social learning and/or memory deficit (they fail to form a memory of the first-presented mouse). To disentangle these seemingly opposite consequences, typically habituation sessions are conducted for a longer period of time and behavior examined across time within the same day (e.g., Hedge et al., 2016, Neuroscience https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5031549/).

We thank the reviewer for the comment. To prove that social isolation does not induce social learning or memory deficits, we performed another habituation task presenting the same social stimulus for 4 consecutive trials for 2 minutes with an intertrial interval of 5 minutes. We scored the interaction time for each trial. We observed that both Grouped and Isolated mice habituate to the same social stimulus with these settings. Indeed, the interaction time on Trial 4 is statistically different from Trial 1 for both groups (new figure 1-S1I-M). These results demonstrate that isolated mice can form short-time memory of the social stimulus while long term memory is impaired (new figure 1I-M)

5) The extensive and definitive use of the term "social craving" throughout is unjustified based on the data presented (e.g., there is no breakpoint data shown). Craving implies having experienced something that is rewarding and wanting more of it; however, these mice never experienced any social interactions with non-cagemates, so how can they crave it? To this end, it might be informative to test these behaviors using cage mates as stimulus mice.

6) Regarding the statement "Indeed, we could speculate that social isolation during adolescence results in an intense urge to interact with whatever conspecific is present." This can be tested in a straightforward fashion by using cage mates as stimulus mice, or measuring interactions in the home cage, instead of using novel mice.

We thank the reviewer for the comment. We now performed free-social interaction after social isolation using former cage mates as stimuli as suggested by the reviewer (new figure 1E-F). We found that isolated mice interact more compared to grouped mice even when stimulus mice are former cage-mate.

7) It is important to clarify why stimulus mice were 1 week younger and whether this is this true for adolescent and adult studies? This is of particular relevance because in the adult study illustrations, a small (seemingly juvenile) mouse is drawn--which would suggest much more than 1 week younger in the case of testing adults. If adults were tested with juvenile mice, as the illustration implies, it complicates interpretation of the adult isolation study since the differential effect of isolation may not be related to the period of isolation but rather a fundamentally different preference between adolescent and adult mice for other juvenile mice.

We thank the reviewer for the comment. As stated in the material methods, stimuli mice are always 1-week younger compared to the experimental mice thus they are P28 for adolescent experimental mice, and P50-53 for adult experimental mice. We have now modified the illustration in figure 1-S2B to be more in line with the age of the mice.

8) The authors identified putative DA neurons based on position, morphology, and capacitance, while providing sparse details in the methods. This section should contain detailed information and/or a small post hoc immunohistochemistry validated dataset.

We are sorry for the lack of information. We have provided all the details now in the material and method session. Moreover, we now added a post-hoc immunohistochemistry validation (new figure 5 – Supplement 2)

9) The authors have exclusively shown scaled-up excitatory responses by pharmacologically blocking GABAA receptors, determining synaptic scaling after social isolation without considering potential effects on inhibitory post synaptic responses. Please provide prediction for the latter. Furthermore, it is not clear if synaptic scaling reported in Figure 5 induced by adolescent isolation only emerges slowly in the adult, or it is already acutely happening at the end of isolation at p35. Please discuss in detail.

We thank the reviewer for the comment. We have now performed new experiments and patched DA neurons projecting the PFC or NAc obtained from adolescent mice immediately after the social isolation. We found an increase in rectification index when recoding from DA neurons projecting to PFC at P35 (new figure 5-supplement 1A-K). Moreover, we now recorded sIPSCs from VTA DA neurons, and we found no differences in amplitude, frequency, rise, and decay time of the inhibitory postsynaptic currents, suggesting thus that inhibitory responses are not affected by adolescent social isolation.

10) It is unclear how oxytocin neuron activity contributes to synaptic plasticity onto VTA-NAc projections. One possibility is that oxytocin expressing PVN neurons projecting to VTA also release glutamate and contribute to the synaptic plasticity. It would be informative to optogenetically examine the synaptic connectivity between PVN->VTA projections and VTA->NAc projection neurons.

We thank the reviewer for the comment. As suggested by the reviewer, we injected ChR2 in the PVN at PND5 and CholeraToxin in the mPFC at PND21. We then isolated the mice for seven days (P28 and P35), and we recorded from DA neurons projecting to the PFC while optogenetically stimulating PVN inputs in slices. Unfortunately, the injection of ChR2 in the PVN at PND5 resulted very challenging and we were not able to measure any current due to the lack of infection of the PVN.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

1) Please acknowledge the lack of females as a shortcoming of the study and provide more discussion of this limitation; e.g., that differences in the system in males vs females and how these findings here may only apply to the male brain. Please also state the sex of the subjects in the abstract.

We thank the reviewer for the comment. We have now clearly stated in the abstract that we used only male, and we add a discussion about the limitation of the study in the discussion session (lines265-275).

2) The post hoc analyses referred to in 2f of the letter are still missing. The statistical interaction was reported, but the post hoc tests were not, nor were significance markers added to the figure to indicate at which pA the groups differ.

We are sorry for the lack of information. All the statistical parameters are now reported in the legend. Post-hoc analyses have now been reported. For figure 3K (former figure 3L) we reported in the figure the post-hoc analysis of isolated group versus all other groups while reported in the source file all the rest of the comparison.

3) Please make sure all statistics are reported. For instance, in some cases a 2 way ANOVA was used but only one main effect is reported. https://reviewer.elifesciences.org/author-guide/full "Report exact p-values wherever possible alongside the summary statistics and 95% confidence intervals. These should be reported for all key questions and not only when the p-value is less than 0.05."

We are sorry for the lack of statistical details that we have now added all the statistical information in the revised version of the manuscript. Moreover, all the statistical data are also reported in the source files.

4) As isolated mice show increased social interaction during direct social interaction, but not during the 3 chamber sociability assay, it is important to more thoroughly explore why isolated mice show different outcomes between these two tests. To better understand the cause of increased direct social interaction (social craving vs aggression), it would be informative to conduct additional behavioral testings from isolated mice (e.g. social CPP, aggression assay). If new behavior testing data such as social CPP or aggression assay (resident intruder etc) is not possible, at the very least a more detailed behavioral characterization such as chasing during free social interaction test is needed to address this concern and better distinguish craving vs aggressive encounter. Overall, more careful consideration is required to interpret the nature social behavior in isolated mice captured by free social interaction and 3 chamber task.

We thank the reviewer for the comment. As suggested, we performed the social conditioning place preference task after isolation. The new results (Figure 1 – Supplement 2 M-R) show that both Grouped and Isolated mice show preference for the compartment paired with social interaction. Concerning figure 1H, since the ANOVA does not show interaction for social preference, we are not allowed to perform multiple comparison. However, we would like to report that if we perform an unpaired t-test between interaction time stimulus 1 for grouped and isolated mice, we found a significant difference (t₍₃₀₎=2.167, p=0.0383).

5) The response to essential review 2c to statistically compare between groups was not fully addressed. [Results] page 1, lines 18-19 and Figure 1EF. To evaluate if the house condition modifies the sociability, it is essential to run these statistics for Figure 1HI. Related to this, the 3 chamber testing data should be presented and analyzed in a similar way across different sets of data. E.g., Figure 1 HI are missing center chamber time and the preference index data while all the rest of the 3 chamber data include these.

In general, more clarity on whether or not the 3-chamber social preference test results do or do not support the conclusion of social isolation leading to an increase in social interaction would help. E.g., by clarifying whether isolation leads to an increase in social interaction specifically in "free/unrestrained"social interactions in adulthood.

We thank the reviewer for the comment. The data regarding the time in chamber and preference index for figure 1 were already present in Figure 1 – supplementary 1. Since figure 1 has many panels, we decided to include in the main figure only the interaction time.

In regards of the statistical analysis for the three-chamber test throughout the manuscript, in this new version of the manuscript we revised all the statistic such as:

– For the interaction time, we first performed a paired t-test within single groups. This statistical test is important for the sociability test to verify that control group prefer stimulus over object. Indeed, the absence of sociability in this group would drop all conclusions. As now requested by the reviewer, we also performed RM 2-way ANOVA. Now all the statistical analysis have been reported in the figures, legends, and source data files. Multiple comparisons were made when interaction between variables was significant.

– For time in chamber, One-way RM ANOVA

– For preference index, One-sample t-test against chance level.

We now show that during the social novelty phase (figure 1I) there is an interaction between the house condition and the interaction time confirming that social isolation induces an alteration of the social novelty preference.

Moreover, as suggested, we added in the text “in free/unrestrained social interaction” to better clarify the effect of social isolation in adulthood.

Author details

1. Stefano Musardo

   Department of Basic Neuroscience, University of Geneva, Geneva, Switzerland

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Software, Validation, Visualization, Writing - original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0392-0905

2. Alessandro Contestabile

   Department of Basic Neuroscience, University of Geneva, Geneva, Switzerland

   Contribution

   Formal analysis, Investigation, Visualization, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-6528-1403

3. Marit Knoop

   Laboratory of Child Growth and Development, University of Geneva, Geneva, Switzerland

   Contribution

   Formal analysis, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

4. Olivier Baud

   1. Laboratory of Child Growth and Development, University of Geneva, Geneva, Switzerland
   2. Division of Neonatology and Pediatric Intensive Care, Children's University Hospital of Geneva, Geneva, Switzerland

   Contribution

   Formal analysis, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

5. Camilla Bellone

   Department of Basic Neuroscience, University of Geneva, Geneva, Switzerland

   Contribution

   Conceptualization, Funding acquisition, Methodology, Writing - original draft, Writing – review and editing

   For correspondence

   Camilla.Bellone@unige.ch

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-6774-6275

• 3,250

  Page views

• 664

  Downloads

• 9

  Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.