Targeting selectively oxytocin receptor signalling efficiently improves social interaction in Fmr1 KO mice

Targeting selectively oxytocin receptor signalling efficiently improves social interaction in Fmr1 KO mice | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 23 December 2025 V1 Latest version Share on Targeting selectively oxytocin receptor signalling efficiently improves social interaction in Fmr1 KO mice Authors : Caroline GORA , Nicolas AZZOPARDI , Emilia CAIRE , Lucile DROBECQ , Emmanuel Pecnard , Patrick SCHNIDER , Pascale David-Pierson , Romain Yvinec , Christophe Grundschober , and Lucie P. Pellissier 0000-0001-7085-3242 [email protected] Authors Info & Affiliations https://doi.org/10.22541/au.176650769.91664930/v1 292 views 50 downloads Contents Abstract INTRODUCTION DISCUSSION AUTHOR CONTRIBUTIONS Supplementary Material References Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Background and Purpose : No drugs targeting the core social features of autism spectrum disorder (ASD) have been approved. Although clinical trials with oxytocin (OT) and vasopressin (AVP) have yielded mixed results, targeting their receptors remains one of the most promising pharmacological strategies for addressing social impairments in ASD. This study aims to identify which receptors and signalling pathways within this family can sustainably improve social impairments. Experimental Approach : We used dose-response and kinetic analyses, along with mathematical modelling, to evaluate OT, AVP, their homologs, and novel synthetic ligands on G protein coupling, β-arrestins recruitment, and internalisation of mouse oxytocin (OTR) and vasopressin (V 1A , V 1B , V 2 ) receptors in Neuro-2a cells. We tested acute and subchronic administration of OTR agonists and the novel V 1A receptor antagonist, alongside OT and AVP, for their effects on social interaction in Fmr1 KO mice, a model exhibiting ASD-like features. Key Results : While OT, AVP and most compounds were non-selective across the four receptors, the OTR agonists TGOT or RO6958375 and the V 1A antagonist RO6893074 were selective. OT, AVP, or RO6893074 exhibited limited efficacy in Fmr1 KO mice. In contrast, RO6958375 and TGOT enhanced social interactions in Fmr1 KO mice, while showing limited side effects in wild-type mice. Conclusion and Implications: Selective OTR agonists, unlike OT and AVP, improved social impairments in Fmr1 KO mice. These findings highlight the necessity for developing highly selective OTR agonists to achieve clinical outcomes in ASD. INTRODUCTION Autism spectrum disorder (ASD) is a neurodevelopmental condition characterised by impairments in social communication and interaction, along with stereotyped or repetitive behaviours (American Psychiatric Association, 2013). It affects approximately 1% of the global population (Zeidan et al., 2022). Recent genome-wide association studies have identified hundreds of candidate genes (SFARI Gene), implicating convergent neurobiological mechanisms such as gene expression regulation and neuronal communication, signalling or plasticity (De Rubeis et al., 2014; Satterstrom et al., 2020; Pintacuda et al., 2023). Among them, loss of the Fragile X Mental Retardation 1 ( FMR1 ) gene expression causes the Fragile X syndrome, a monogenic disorder frequently associated with ASD, anxiety, hyperactivity and intellectual disability (Pieretti et al., 1991; Verheij et al., 1993). The Fmr1 knockout (KO) mice reproduce these core features—showing social impairments and hyperactivity accompanied by synaptic plasticity defects and oxytocin family deficiency—thus meeting construct and face validity of the human condition (Mientjes et al., 2006; Kat et al., 2022; Giua et al., 2024; Gora et al., 2024). Consequently, Fmr1 KO mice serve as a well-validated preclinical model for studying ASD-related social deficits and testing oxytocin-derived therapeutic strategies. To date, no effective pharmacological options targeting core social features have been approved for individuals with ASD or Fragile X syndrome. Placebo effects, lack of efficacy, and the heterogeneity of patients within the spectrum have prevented success in phase 3 clinical trials. G protein-coupled receptors (GPCRs), are targeted by around 30% of approved drugs (Hauser et al., 2017; Alexander et al., 2021) and are key regulators of downstream signalling that are dysregulated in ASD (De Rubeis et al., 2014; Satterstrom et al., 2020; Pintacuda et al., 2023). Over 200 pathogenic variants in 26 GPCR genes have been associated with ASD (Annamneedi et al., 2023). Thus, GPCRs have emerged as robust and specific therapeutic targets for ASD. Among them, oxytocin (OT) or vasopressin (AVP) bind to oxytocin (OTR) and vasopressin (V 1A and V 1B ) receptors in the brain, modulating social behaviours (Busnelli et al., 2013; Quintana et al., 2019; Rigney et al., 2022; Theofanopoulou et al., 2022). These receptors are primarily coupled to Gα q/11 proteins, leading to the activation of the IP 3 -Ca 2+ signalling pathway and recruit β-arrestins upon ligand binding (Busnelli et al., 2013). Although results have been inconsistent, some clinical trials have reported improvements in social scales and repetitive behaviours following intranasal administration of OT (Bernaerts et al., 2020; Annamneedi et al., 2023; Daniels et al., 2023; Guastella et al., 2023). Notably, while plasma and brain concentrations of OT are not correlated (Valstad et al., 2017; Gora et al., 2024), individuals with lower plasma OT levels tend to respond better to exogenous OT administration (Parker et al., 2017). Furthermore, recent studies suggest that combining OT administration with behavioural interventions enhances sociability more than OT alone (Ford and Young, 2022; Daniels et al., 2023; Pantouli et al., 2024). Similarly intranasal AVP administration was well tolerated in early trials and improved social responsiveness, particularly in children with ASD and low plasma AVP concentrations (Parker et al., 2019). While administration of the V 1A antagonists balovaptan or RG7713 enhanced social skills in adults with ASD in phase 2 or exploratory clinical studies (Umbricht et al., 2017; Bolognani et al., 2019), balovaptan did not show improvement over placebo in phase 3 trials (Hollander et al., 2022; Jacob et al., 2022). Despite this, OTR and V 1A receptors remain the most promising targets to improve sociability in individuals with ASD (Annamneedi et al., 2023). Further research is needed to identify the most suitable ligands, targets and downstream signalling pathways. In this study, we aim to enhance our understanding of why compounds targeting OTR and V 1A receptors have underperformed in clinical trials. We used a systematic pharmacological approach to assess both existing and novel ligands, determining which ligand-receptor pairings are most promising in vitro and in Fmr1 KO mice. Revisiting oxytocin and vasopressin on mouse vasopressin or oxytocin receptors in murine Neuro-2a cells To address the inconsistent results from OT and AVP clinical trials (Annamneedi et al., 2023), we investigated their distinct pharmacological kinetic and concentration-response relationship profiles on OTR, V 1A , V 1B and V 2 receptors, over ten signalling outputs in the murine Neuro-2a neuroblastoma cell line ( Figures 1, S1, Table S1 ). Spider plots illustrated that all receptors recruited miniGq, leading to intracellular Ca 2+ mobilisation, and both β-arrestin-1 and -2 and internalised (using CAAX and FYVE sensors) following OT or AVP stimulation ( Figures 1A-B, S1 ). As expected, V 2 also recruited miniGs and induced cAMP production. Conversely, none of the receptors recruited miniGi or inhibited forskolin-induced cAMP production ( Figures 1A, S1 ). Overall, AVP acted as a partial agonist on OTR compared to OT, while OT was a partial agonist on V 1A receptors across these assays (except for Ca 2+ mobilisation). OT acted as a full agonist with lower potency than AVP on V 1B and V 2 receptors, although it was less efficient at recruiting β-arrestin on V 2 receptors ( Figure 1B ). The pharmacological profiles of OT and AVP at the four human receptors in human HEK293A cells were comparable to those observed for the murine receptors ( Figure S2, Table S2 ). Overall, miniGq and β-arrestin-2 recruitment served as the most robust BRET1 assays to identify potential ligand bias ( Figure 1B ). Next, we evaluated OT and AVP bias following GPCR ligand bias guidelines (Kolb et al., 2022). Bias plots represented miniGq recruitment (y-axis) versus β-arrestin-2 recruitment (x-axis) across molar concentrations from the respective dose–response curves ( Figure 1B ). The diagonal lines (y = x) display equal activation of both pathways (no bias), while points above indicating miniGq bias, and points below indicating β-arrestin-2 bias. Bias was calculated from statistical differences between miniGq and β-arrestin-2 transduction coefficients, ∆log(τ/K A ), that reflects the EC 50 potency and E max efficacy of each ligand-receptor, relative to the reference ligand ( Table 1 ). Considering both the bias plots and ∆log(τ/K A ), we found that OT was biased toward miniGq at V 2 compared with AVP ( Figure 1B , Table 1 ). To estimate the potential role of receptor signalling kinetics and internalisation between OT and AVP in the lack of efficacy resulting from repeated administration during clinical trials, we performed mathematical modelling incorporating miniGq and β-arrestin recruitment, and subsequent receptor internalisation kinetic data for the four receptors ( Figures 1C, S3 , Material and Methods , Table S3 ). The parameter identifiability analysis produced robust, normalised parameter estimates with small confidence intervals for predicted affinity (1/K D ) and for the kinetic parameters, Gq recruitment (k τ2 ), Gq inactivation (k des ), β-arrestin recruitment (k τ ), receptor internalisation (k int ) and recycling (k rec , Table S3 ). We identified significant receptor recycling parameters (except for AVP on OTR and OT on V 1A that were negligible), which were slower than internalisation within the 30-minute period ( Table S3 ). Radar plots highlighted key kinetic differences, showing that OT at OTR, AVP at V 1A and V 2 receptors have greater efficacy for Gq and β-arrestin recruitment than AVP or OT at these receptors ( Figure 1C, S3 ). Furthermore, OT at OTR and AVP at V 1A have higher internalization and recycling rates than AVP or OT at these receptors. OT and AVP displayed very similar kinetic at V 1B receptor, consistent with their comparable pharmacological profile ( Figure 1A ). In conclusion, although OT and AVP activated all four receptors, they display distinct potency, efficacy and kinetics that may contribute to the high variability observed in clinical outcomes with these peptides. The RO6893074, a related chemical compound to balovaptan used in clinical trials, is a selective V 1A antagonist The antagonist RO6893074 is a chemical compound related to balovaptan, which has been evaluated in clinical trials (Schnider et al., 2020). RO6893074 showed no partial agonist effect at any of the four oxytocin or vasopressin receptors when tested alone ( Figure 2A, Table S1 ). It displayed high antagonist selectivity for the mouse V 1A receptor and did not block mouse OTR, V 1B and V 2 receptors ( Figure 2B ). In human HEK293A cells, its pharmacological profile at the four human receptors was equivalent to those at murine receptors ( Table S4 ). Furthermore, in an exploratory study, RO6893074 did not inhibit a selection of 47 human receptors, channels, transporters and enzymes when tested at 3 µM, except for a mild antagonist effect on MT3, κOR, 5-HT 2B and SST4 with not enough off-target occupancy for these receptors in vivo ( Table S5 ). The pharmacokinetics of RO6893074 following a single intraperitoneal dose of 5 mg.kg −1 and 50 mg.kg −1 in mice, indicated an average maximum plasma concentration (Cmax) of approximately 220 ng.mL -1 and 7000 ng.mL -1 , with a time to reach Cmax (Tmax) of 0.5 hours and 0.84 hours, respectively. The area under the plasma concentration-time curve from time 0 to infinity (AUC₀-∞) was 581 and 16400 ng.h.mL -1 , respectively. The elimination clearance (CL = Dose/AUC) was dose-dependent. Based on free plasma exposure of 6.3% (considering the compound is not a substrate for mouse P-glycoprotein and has good permeability) and in vitro V 1A binding, the corresponding brain V 1A receptor occupancy was estimated to be approximately 49% and 96% at doses of 5 mg.kg −1 and 50 mg.kg −1 , respectively. Administration at doses of 10 mg.kg −1 and 100 mg.kg −1 is expected to yield approximately 65% and 98% occupancy, respectively. Biological effects of V 1A blockade are expected to be detectable between 70 to 90% receptor occupancy (Grimwood and Hartig, 2009). Higher exposure may lead to inhibition of off-target receptors, inducing non-specific side effects. In conclusion, RO6893074 is a selective antagonist for human and mouse V 1A receptors in vitro and in vivo . OT, AVP or V 1A antagonist did not robustly enhance social interactions in Fmr1 KO mice To better understand the mixed outcomes reported in clinical trials, we compared the effects of acute (Day 0) and subchronic (Day 7) administration of OT and AVP (intranasally at 0, 20 and 40 µg.kg −1 ) and V 1A antagonist RO6893074 (intraperitoneally at 0, 25, 50 and 100 mg.kg −1 ), as well as their delayed effects four (Day 11) and seven (Day 14) days after the last administration, on social interaction in Fmr1 KO mice ( Figure 3, Table S6 ). Fmr1 KO mice are well-established models for studying Fragile X syndrome and ASD associated with oxytocin deficiency (Lewis et al., 2020; Kat et al., 2022; Giua et al., 2024; Gora et al., 2024). We used the Live Mouse Tracker (de Chaumont et al., 2019), which automatically quantifies multiple parameters of social interaction between two unfamiliar sex-, treatment- and genotype-matched mice. This system has successfully detected social impairments and hyperactivity in Fmr1 KO mice (Giua et al., 2024), findings that we confirmed in this study ( Figure S4 ). Neither acute nor subchronic intranasal administration of OT or AVP, at any dose, improved nose contacts or huddling, two parameters of social exploration, in Fmr1 KO mice ( Figures 3, S5 ). Acute OT at 20 µg.kg −1 tended to increase social motivation, indicated as ‘move in contact’, whereas acute OT at 40 µg.kg −1 or AVP at 20 µg.kg −1 reduced social motivation. Notably, OT induced higher inter-individual variability than AVP, at all doses ( Figure S5 ). Acute administration of the V 1A antagonist RO6893074, aimed to redirect endogenous OT effects toward OTR by blocking V 1A receptors, increased nose contact, huddling and ‘move in contact’, as well as ‘move in contact and social approach seven days post-treatment at 50 mg.kg −1 in Fmr1 KO mice ( Figures 3, S5, Table S6 ). Acute 100 mg.kg −1 reduced all social parameters in Fmr1 KO mice. Overall, RO6893074 induced higher inter-individual variability than OT ( Figure S5 ). We also tested potential side effects in WT mice ( Figure S6, Table S6 ). Subchronic OT at 40 µg.kg −1 reduced stretch-attend postures (SAP), an index of anxious-like behaviours. Similar to Fmr1 KO mice, acute RO6893074 at 100 mg.kg −1 reduced all parameters in WT mice ( Figure S6, Table S6 ). In conclusion, neither OT, AVP nor V 1A antagonist RO6893074 effectively improved social deficits in Fmr1 KO mice, while high-dose RO6893074 displayed side effects in both WT and Fmr1 KO mice. Identifying existing and novel agonists targeting selectively OTR receptors in Neuro-2a cells To identify the most specific and potentially biased OTR agonists, we compared natural and synthetic peptide agonists targeting oxytocin and vasopressin receptors on miniGq and β-arrestin-2 recruitment in Neuro-2a cells ( Figure 4 , Table S1 ). The bird and fish OT homologues mesotocin and isotocin recruited both miniGq and β-arrestin-2 at all four receptors, except for isotocin at V 1A ( Figure 4A ). Considering both the bias plots and ∆log(τ/K A ), mesotocin and isotocin, similar to OT, were biased toward miniGq (points above the diagonal) at the V 2 receptor ( Figures 4A, Table 1 ). The bird AVP homologue vasotocin activated all receptors and showed a slight β-arrestin-2 bias (points below the diagonal) at V 2 compared with AVP ( Figures 4A, Table 1 ). Among the synthetic peptide OTR agonists, RO6958375 (Janz et al., 2023) and TGOT (Elands et al., 1988) activated OTR, with RO6958375 being less potent and efficient than TGOT ( Figure 4B ). RO6958375 also activated V 2 to a lesser extent, whereas neither agonist activated V 1A and V 1B , in contrast to OT. TGOT exhibited a slight miniGq bias at OTR, while RO6958375 was biased toward miniGq at V 2 ( Figure 4B, Table 1 ). In contrast, kB7-OT (Koehbach et al., 2013) and carbetocin previously described as a Gα q -biased OTR agonist (Passoni et al., 2016), did not activate any of the receptors in our assays, except V 2 , where carbetocin exhibit a miniGq bias ( Figure 4B, Table 1 ). In conclusion, both TGOT and RO6958375 are selective for OTR agonists, with TGOT acting as a partial miniGq-biased ligand. Their use in vivo will therefore isolate the effect of OTR activation on social interaction, unlike OT or AVP, which activate all four receptors. Targeting selectively oxytocin receptors improves social interactions in Fmr1 KO mice To determine whether selective OTR activation could improve social interactions, we administered the selective OTR agonists TGOT (intranasally, at 0, 20 and 40 µg.kg −1 ) and RO6958375 (subcutaneously, at 0, 0.03, 0.06 and 0.12 mg.kg −1 ) in Fmr1 KO mice ( Figures 5, S5, Table S6 ). Acute TGOT at 20 µg.kg −1 enhanced huddling and nose contact, with nose contact remaining elevated three days after treatment at 40 µg.kg −1 . Acute RO6958375 at 0.03 and 0.06 mg.kg −1 increased nose contact and ‘move in contact’ in Fmr1 KO mice, with enhanced nose contact (tendency, P value = 0.022) remaining three days post-treatment at 0.03 mg.kg −1 . Inter-individual variability for both agonists was comparable to OT ( Figure S5 ). However, acute and subchronic TGOT at 40 µg.kg −1 reduced social interaction in WT mice. RO6958375 at 0.06 mg.kg −1 enhanced huddling and nose contact, whereas acute and subchronic administration at 0.03 mg.kg −1 enhanced SAP and subchronic administration at 0.12 mg.kg −1 reduced on social motivation (‘move in contact’ and ‘social approach’) and nose contact in WT animals ( Figure S7 ). These effects were reversed after treatment. To better understand the differences between the OTR agonists RO6958375 and TGOT, we performed mathematical modelling of miniGq, β-arrestin and FYVE internalisation kinetics in Neuro-2A cells ( Figure 6 , Table S3 ). We identified distinct miniGq recruitment kinetics for the two selective OTR agonists compared to OT, while their predicted affinities (1/K D ) and β-arrestin-2 recruitment kinetics (k τ ) were relatively similar to OT. TGOT displayed an internalisation profile comparable to OT (k int ), but with faster miniGq recruitment (k τ2 ), and recycling (k rec ), consistent with its biased signalling profile. In contrast, RO6958375 showed slower Gq recruitment and internalisation than OT and negligible recycling. In conclusion, both selective OTR agonists improved social impairments in Fmr1 KO mice. Particularly, RO6958375 was effective at low and moderate doses, with minimal side effects in WT animals, likely due to reduced desensitisation and prolonged cell-surface signalling in vivo . DISCUSSION In our study, we evaluated the efficacy, potency and selectivity of various ligands targeting oxytocin and vasopressin receptors in neuronal heterologous cells. Despite similar binding affinities (Busnelli et al., 2013), OT and AVP exhibited distinct efficacy and potency across their receptors. They primarily recruit miniGq and β-arrestins at OTR, V 1A , and V 1B receptors, while the V 2 receptor mainly recruits miniGs and also activates miniGq. These findings are consistent with previous studies (Birnbaumer, 2000; Busnelli et al., 2012, 2013; Heydenreich et al., 2022). In our assays, none of the receptors coupled with miniGi proteins nor mediated cAMP inhibition, which contrasts with previous reports for OTR in other cell lines (Strakova and Soloff, 1997; Busnelli et al., 2012). The lower affinity of the chimeric miniG protein may explain these differences and the variations observed in the calcium assay. However, we did not detect any cAMP inhibition resulting from forskolin-induced adenylyl cyclase activity. Previous studies identified an association between G⍺ i3 and OTR in rat myometrium and CHO cells (Strakova and Soloff, 1997; Strakova et al., 1998) and described indirect conformational changes between G⍺ i/o and Gᵦᵧ subunits or OTR fused to YFP and Gᵧ upon OT stimulation in HEK293 cells (Busnelli et al., 2012). These findings imply that OTR could signal through Gᵦᵧ release, as previously shown (Camps et al., 1992), or G⍺ i/o recruitment depend of the cellular context. While β-arrestin-1 and -2 recruitment was comparable across receptors and ligands, we cannot exclude the possibility of distinct conformational changes between the two β-arrestins (Haider et al., 2022). Isotocin, mesotocin or vasotocin homologs activated all four receptors similarly to OT or AVP, except isotocin on V 1A . As expected, isotocin and mesotocin orthologs displayed pharmacological or bias profiles comparable to OT, while vasotocin mirrored AVP. Surprisingly, kB7-OT and carbetocin did not activate OTR in our assays, except for carbetocin at V 2 , contrasting with previous studies (Koehbach et al., 2013; Passoni et al., 2016). This difference may be attributed to distinct receptor expression (Neuro-2a vs. HEK293A cells) or signalling assays (miniGq vs. G ⍺ -Gᵧ BRET assays). We identified TGOT, but not RO6958375, as partial miniGq-biased OTR agonist, a profile previously reported (Busnelli et al., 2013), whereas RO6958375 was biased toward miniGq at V 2 . Since V 2 is not expressed in neurons or glia and does not contribute to social behaviours (Ostrowski et al., 1992; Kato et al., 1995), in vivo administration of these agonists primarily targets OTR, regulating social circuits in the brain. Intranasal administration of OT and AVP showed limited efficacy and high inter-individual variability (particularly for OT) in improving social impairments in Fmr1 KO mice. These findings are consistent with previous preclinical studies (Lindenmaier et al., 2022; Pan et al., 2022) and clinical trial outcomes, where OT or AVP failed to produce consistent therapeutic effects and showed high variability in individuals with ASD (Guastella et al., 2010; Bernaerts et al., 2020; Annamneedi et al., 2023; Daniels et al., 2023). The therapeutic potential of OT may be limited by vasopressin receptor activation, despite OT partially activates V 1A in vitro . Hetero-oligomerisation between OTR and V 1A (Terrillon et al., 2003) or V 1B receptors may also impact in vivo effects. Additionally, OT has poor blood-brain barrier penetration, even through the intranasal route, resulting in low brain exposure and high inter-individual variability (Leng et al., 2022). Although variable, acute selective blockade of V 1A with RO6893074 at one dose improved social interaction in Fmr1 KO mice, whereas subchronic administration did not. Interestingly, seven days post-treatment increased social interaction, indicating potential rearrangements in OTR or V 1A receptor levels or OT/AVP circuits. The 100 mg.kg -1 dose affected locomotion and sociability in both Fmr1 KO and WT mice, suggesting a sedative effect likely due to high exposure and off-target inhibition. These findings suggest that blocking V 1A receptors alone may be insufficient to redirect endogenous OT toward OTR activation. Consequently, combining RO6893074 with exogenous OT administration could potentially enhance therapeutic effects on sociability. Conversely, selective targeting of OTR with RO6958375 and TGOT agonists improved social interactions in Fmr1 KO mice at low to moderate doses. This supports the hypothesis that non-selective activation of vasopressin receptors by endogenous OT may counteract the pro-social effects of OTR activation. Combining OT administration with behavioural interventions is more efficient in improving social deficits (Ford and Young, 2022; Daniels et al., 2023; Pantouli et al., 2024). In our study, exposing mice to repeated social interactions with unfamiliar mice may have enhanced the efficacy of these agonists, as suggested by the persistent post-treatment therapeutic effects. TGOT and RO6958375 showed distinct effects in WT mice. While both agonists reduced social interaction at the highest dose (reversed post-treatment), RO6958375 improved social interactions in WT mice at low and moderate dose. The difference in compound effects may reflect the administration route (intranasal vs. subcutaneous), influencing brain exposure and inter-individual variability. For instance, RO6893074 (intraperitoneal) produced the highest inter-individual variability, AVP showed the lowest, while TGOT (intranasal) and RO6958375 (subcutaneous) showed similar variability than OT (intranasal). This suggests that variability is driven by compound-specific mechanisms rather than by the administration route. The two OTR agonists are characterised by reduced miniGq and β-arrestin recruitment efficiency and potency in vitro compared to OT. However, mathematical modelling suggests that differences between these agonists stem from their distinct kinetics. RO6958375 provide sustained OTR signalling with limited internalisation, compared to OT or TGOT, offering a crucial therapeutic advantage for chronic treatment. Indeed, the overall beneficial effect of RO6958375 in vivo , in both Fmr1 KO and WT mice, suggest that it display a more favourable pharmacological profile of RO6958375 than TGOT. Targeting OTR with selective agonists may provide new opportunities of targeted therapies for ASD. Although challenging, the development of OTR positive allosteric modulators enhancing OTR signalling in brain regions where OT is endogenously released holds also therapeutic promise. Additionally, ligands selectively biased towards G protein or β-arrestin pathways have been shown to optimise therapeutic effects in other neurological disorders while minimising side effects of other signalling pathways (Park et al., 2016; Lee et al., 2021; Kolb et al., 2022). Full Gα q/11 - or β-arrestin-biased OTR agonists, which induce signalling selectivity through specific pathways, may represent effective therapies for ASD, as TGOT display only a partial bias toward Gα q/11 . Understanding the precise contributions of each receptor to social interaction, as well as their expression patterns and the roles of endogenous peptides, remains crucial for the development of targeted interventions. Our study offers new insights into OTR activation and V 1A inhibition on sociability. Future research should also explore the effects of selective V 1A agonists (Andrés et al., 2002; Marir et al., 2013) or OTR antagonists (Goodwin et al., 1994) in combination with OTR agonists, as well as the role of V 1B receptors, which have more restricted expression in the brain (Young et al., 2006; Stevenson and Caldwell, 2012), to clarify their contributions. A limitation of the present study is the lack of direct measurements of mouse brain exposure for the tested compounds. Such data would allow more accurate estimates of receptor occupancy and the relative contributions of G protein signalling versus β-arrestin recruitment. While our study focused on social features, assessing effects on stereotyped behaviours, or cognitive flexibility would provide a more comprehensive evaluation of treatment efficacy. Testing OTR agonists in mouse models using a molecular stratification approach based on specific molecular markers across models mimicking the autism spectrum (Gora et al., 2024), may enhance translatability and facilitate the development of personalised medicine for individuals with ASD. MATERIALS AND METHODS Compounds For in vitro assays, drug powders of oxytocin (OT; 1910, Tocris Bioscience™, UK; ≥95% purity (HPLC)), (Arg⁸)-vasopressin (AVP; 2935, Tocris Bioscience™, UK), (Butyryl¹,Tyr(Me)²)-1-Carbaoxytocin (carbetocin; 4040269, Bachem, Switzerland), (Ser⁴,Ile⁸)-oxytocin (isotocin; 4030890, Bachem, Switzerland), (Ile⁸)-oxytocin (mesotocin; 4030888, Bachem, Switzerland), (Arg⁸)-vasotocin (vasotocin; 4100576, Bachem, Switzerland), (Thr⁴,Gly⁷)-oxytocin (TGOT; 4013837, Bachem, Switzerland), RWJ22164 (atosiban; 4065438, Bachem, Switzerland), (Gly⁵,Thr⁷,Ser⁹)-oxytocin (kB7-OT; 4144391, Bachem, Switzerland), L-371,257 (2410, Tocris Bioscience™, UK), RO6958375 OTR peptide agonist (c[Gly-Tyr-Ile-GIn-Asn-Glu]- trans -4-fluoro-Pro-Leu-Gly-NH 2 , Roche pharmaceutical, Switzerland)(Janz et al., 2023) and RO6893074 V 1A chemical antagonist ( trans -2-(4-(4-(4-chlorophenyl)-5-methyl-4 H -1,2,4-triazol-3-yl)cyclohexyloxy)pyridine, Roche pharmaceutical, Switzerland) were dissolved at 10 -2 M in 100% DMSO (20-139, Sigma-Aldrich, USA), aliquoted and stored at -20°C until further use. The structure of RO6893074 is represented here and its synthesis has been described in the patent number WO2022018121 (example 3 in ’Cyclohexyl substituted triazoles as vasopressin receptor V 1A antagonists). Binding and functional characteristics of RO6893074 for the hV 1A receptor and exploratory (n = 1-4) for the other receptors and species in HEK293 are presented in the Table S4 . On the testing day, all compounds were diluted in PBS 1X + 10 mM HEPES (CSTHEP00-0P, Eurobio, France) at 4°C and tested alone. In addition, for their antagonist effects, atosiban, L-371,257 and RO6893074 V 1A antagonists were diluted at EC 80 of each reference ligand (concentration reaching 80% of maximal efficacy; 10 -6 M of OT for OTR or in 10 -6 M of AVP for V 1A and V 1B or 10 -8 M for V 2 ). Coelenterazine luciferase substrate (R3078C, Interchim, France) was protected from light exposure and stored at 1 mM in 100% ethanol (64-17-5, Carlo Erba Reagents, France) and then diluted to 5 µM final in PBS 1X. Cell culture Neuro-2a mouse neuroblastoma (Neuro-2a; CCL-131™, RRID:CVCL_0470, ATCC, USA) cells were regularly tested for negative mycoplasma contamination were cultured in Eagle’s Minimal Essential Medium (EMEM; CM1MEM40-01, Eurobio, France) supplemented with 10% (v/v) foetal bovine serum (CVFSVF00-01, Eurobio, France), 1% penicillin/streptomycin (100 U.mL −1 , 100 µg.mL −1 , respectively; 15140-122, Eurobio, France) and 1% L-glutamine (2 mM; CSTGLU00-0U, Eurobio, France), and maintained at 37°C with 5% CO 2 . When reaching 90% confluency, cells were washed with PBS 1X (CS1PBS01K-BP, Eurobio, France) followed by trypsin (0.5 g.L −1 , L-EDTA 0.2 g.L −1 , without phenol red; CEZTDA00-0U, Eurobio, France) treatment. After cell counting, cells were transiently transfected in suspension using the Metafectene Pro transfection reagent (T040-5.0, BioNTex Laboratories, Germany) according to the manufacturer’s protocol. G protein, 𝛽-arrestin recruitment and internalisation using Bioluminescence Resonance Energy Transfer (BRET) assays Each assay (sensor/receptor/drug) was replicated in n=5 independent experiments. In 96-well plates (30196, SPL Life Sciences, Korea), around 40,000 Neuro-2a cells per well were transiently co-transfected with the mouse oxytocin receptor cDNA Oxtr or mouse vasopressin Avpr1a, Avpr1b subcloned using HindIII and XhoI in pcDNA3 plasmid or with the Avpr2 receptors in pcDNA3.1(+) plasmid, fused in their C-terminus to the BRET donor Renilla reniformis luciferase (RLuc8) at 50 ng/well, along with BRET acceptor sensors containing a yellow fluorescent proteins (YFP, Venus or YPET) at 50 ng/well each. For the miniG protein recruitment, we used the pNluc–C1 vector encoding the miniG proteins fused in their N-terminus to the Nuclear Export Signal (NES) to avoid their transport to the nucleus; NES-Venus-mG s , NES-Venus-mG i or NES-Venus-mG q (provided by Pr Nevin A. Lambert, Augusta, GA, USA)(Wan et al., 2018), namely miniGs, miniGi and miniGq, respectively. MiniGs proteins are engineered G⍺ subunits with modifications such as the exclusion of the α-helical domain, a shortened N-terminus which removes both membrane anchors and the Gᵦᵧ-binding surface, specific mutations to enhance in vitro protein stability and a mutation in the C-terminal 5-helix to stabilise receptor–miniG complexes even in the presence of guanine nucleotides. For miniGq and miniGi, the five C-terminal amino acids of miniGs, which form the main receptor binding site and are key determinants of coupling specificity, were replaced with those of Gα q , necessitating three mutations. For miniGi, 9 mutations were done into the Gα5 helix to change the coupling specificity from Gα s to Gα i (Nehmé et al., 2017; Wan et al., 2018). For β-arrestin recruitment, we used YPET-β-arrestin-1 or YPET-β-arrestin-2 (provided by Dr M.G. Scott, Paris, France)(Kamal et al., 2009) encoded in pcDNA3.1 plasmid. For internalisation experiments, the yellow fluorescent protein YPET was fused to the CAAX box of KRas protein and the FYVE domain of endofin, according to (Lee et al., 2005; Namkung et al., 2016a), namely YPET-CAAX addressed at the plasma membrane or YPET-FYVE addressed at the endosome membrane, respectively. Their sequences were synthesised at Twist Bioscience and subcloned in pcDNA3.1 plasmid by recombination between HindIII and XhoI restriction sites. For cAMP assay, pcDNA3.1(+) encoding untagged murine receptors and the CAMYEL (provided by Dr Lily Jiang, Texas, USA)(Jiang et al., 2007), a cAMP EPAC sensor (YFP-Epac-RLuc8) were co-transfected at 50 ng/well. 48 hours after transfection, Neuro-2a cells were starved for 4 hours in phenol red-free DMEM (21063-029, Gibco, France) at 37°C in 5% CO 2 . Cells incubated with coelenterazine substrate (5 µM) were first measured for 5 minutes at 480 ± 20 nm and 530 ± 25 nm (Mithras2 LB 943 with the Mikrowin 2010 software, Berthold Technologies GmbH & Co., Germany). Then, cells were rapidly stimulated with the different agonists at 3.10 -5 M, 10 -5 M, 10 -6 M, 10 -7 M, 10 -8 M, 10 -9 M, 10 -10 M diluted in PBS 1X and HEPES 10 mM final, or PBS 1X and HEPES 10 mM alone for the baseline, then 96-well plates were recorded for 30 minutes. When testing the effect of the antagonists, they were tested at the same concentration as agonists and in presence of OT at EC 80 for OTR or AVP at EC 80 for the V 1A , V 1B or V 2 receptors. For cAMP production and inhibition BRET assays, compounds were tested in absence or presence of forskolin at 5.10 -6 M (F6886, Sigma-Aldrich, USA), respectively. Each condition was performed in triplicate within the same 96-well plate, and all experiments were independently repeated at least three times. Based on Rluc8 luciferase signals, expression levels across receptors were similar, except for a slightly lower mV 2 in Neuro-2a. Using RNA-seq studies, OTR was expressed only five times more than endogenous GPCRs, and expression of endogenous GPCRs remained unaffected by OTR expression. Calcium mobilisation assays using luminescence Each assay (sensor/receptor/drug) was replicated in n = 5 independent experiments. In 6-well plates (353046, Corning, USA), around 400,000 Neuro-2a cells per well were plated 24 hours at 37°C with 5% CO 2 before transfection. Then, cells were transiently transfected with the untagged mouse oxytocin ( Oxtr ) or vasopressin ( Avpr1a, Avpr1b or Avpr2 ) receptor cDNAs subcloned using HindIII and XhoI in the pcDNA3.1(+) plasmid at 1 µg/well and pPD16 vector encoding the calcium-sensitive bioluminescent protein aequorin from the jellyfish Aequorea victoria jellyfish (GFP-aequorin provided by Dr Bertrand Lambolez, Paris, France) at 2 µg per well (Drobac et al., 2010). 48 hours after transfection, Neuro-2a cells were Ca 2+ -deprived for 4 hours at 37°C with 5% CO 2 using the Hank’s Balanced Salt Solution (CS1SSH22-0U HBSS; Eurobio, France) with no calcium or phenol red and containing 5 μM coelenterazine in the dark. Subsequently, cells were harvested and centrifuged at 500g for 5 minutes. Cell pellets were resuspended in around 2.10 6 cells.mL -1 of HBSS with 1.26 mM of calcium (CS1SSH23-0U; Eurobio, France) and 5 μM coelenterazine, and incubated for one hour at 37°C in the dark. OT and AVP ligands were diluted in HBSS with 1.26 mM of calcium across at 10 -5 M, 10 -6 M, 10 -7 M, 10 -8 M, 10 -9 M, 10 -10 M, 10 -11 M, 10 -12 M final concentrations on ice, and HBSS with 1.26 mM of calcium alone for baseline. Ten µL per well of each concentration of ligand (5X) was loaded in duplicate into a 384-well plate and 40 µL of cells (around 40,000 cells per well) were injected into each well. Luminescence of calcium-bound aequorin signals was continuously measured for 25 seconds. In vitro data modelling and analysis BRET and calcium analyses The 530 nm/480 nm BRET ratios were normalised to their initial unstimulated BRET ratio and then computed into “Induced BRET” values by subtracting the control (PBS) normalised ratio at each measurement time point. Then, the area under curves (AUC) were estimated over the 30-minute induced BRET measurements or directly for calcium assays over the 25-second measurements for each combination of receptor, ligand and concentration to study the concentration–response relationship. Concentration-response curves were modelled with a four-parameters log-logistic equation, with E min (minimal response), E max (maximal response, equal to 1 for the reference ligand), EC 50 (concentration to reach E max /2) and sigmoid coefficient. Concentration-response curve values were then normalised to the E max value of the reference ligand for each receptor (i.e. OT for OTR and AVP for V 1A , V 1B and V 2 receptors). Ligand bias calculation and visualisation A transduction coefficient log(τ/K A ) for induced miniGq and β-arrestin-2 recruitment (normalised to the reference ligand) was calculated for each normalised dose-response curve, using the Monolix Suite 2024R1 (Lixoft, Antony, France) defined by the following equation (Kenakin et al., 2012): Where 𝛽 1 = ln(K A ) and 𝛽 2 = ln(𝜏) and log(𝜏/K A ) = (𝛽 2 - 𝛽 1 )/ln(10). The transduction coefficient (τ/KA), described in the GPCR ligand bias guidelines (Kolb et al., 2022), is a composite parameter that reflects the EC 50 potency and E max efficacy of a ligand-receptor couple, thereby providing a comprehensive measure of receptor activation of miniGq and β-arrestin-2 recruitment assay (Kenakin et al., 2012). The difference between transduction coefficient of each downstream signalling, miniGq and β-arrestin-2 recruitments ( Table 1 ), provides a measure of signalling bias for a given receptor that is, the preferential engagement of one signalling pathway over another (Kenakin et al., 2012; Namkung et al., 2016b; De Pascali et al., 2021). Bias plots , following GPCR ligand bias guidelines (Kolb et al., 2022), represent miniGq recruitment (y-axis) versus β-arrestin-2 recruitment (x-axis) at each intermediate molar concentration from their respective concentration–response curves, normalised to the reference ligand (OT for OTR and AVP for V 1A , V 1B and V 2 ). The diagonal line (y = x) represents homogeneous activation of both pathways (no bias). Points above the diagonal indicate ligands biased toward miniGq recruitment, whereas points below the diagonal indicate bias toward β-arrestin-2 recruitment. This representation facilitates the visualisation of ligand bias. A ligand is considered “biased” only when both the bias plots and the bias calculations consistently indicate pathway selectivity. Mathematical modelling of Gq, β-arrestin signalling and internalisation In this section, we detail the kinetic analysis on respectively reversible miniGq activation and recruitment, β-arrestin-2, FYVE and CAAX induced BRET signals. This methodology was independently applied for both endogenous ligand OT and AVP on each murine receptor OTR, V 1A , V 1B and V 2 , as well as OT, AVP, RO6958375 and TGOT for OTR using miniGq, β-arrestin-2 and FYVE sensors. I) Kinetic model We adapted and extended the methodology presented in (Hoare et al., 2020). More precisely, we used the following kinetic reaction network model: (1) R→RAB→RI→R; ∅ → Gq → ∅ where R denotes the unbound membrane receptor, RAB denotes the receptor-agonist-β-arrestin-2 complexes; RI denotes the internalised receptor and Gq the activated Gq protein. We assume fast binding kinetics between the agonist and the receptor and that the agonist is in excess compared to the receptor (Hoare et al., 2020). We further assume that: (i) the free β-arrestin-2 is in excess compared to the receptor-agonist complexes; thus, its depletion can be neglected; (ii) Gq kinetics is faster than receptor kinetics, thus receptor internalisation kinetics can be neglected for Gq activation. As a result, the model (1) is a first-order kinetic reaction network whose dynamics is represented by the linear four-dimensional ordinary differential equation (ODE): (2) where A (M) is the agonist concentration, K D (M) is the binding affinity of the agonist and the receptor, k τ (min -1 ) is β-arrestin-2 recruitment rate, k int (min -1 ) is the internalisation rate, k rec (min -1 ) is the recycling rate, k τ2 (min -1 ) is the Gq activation rate, and k des (min -1 ) is the Gq inactivation rate. The initial condition associated with system (2) is: R (0)= R tot ,RAB (0)=0, RI (0)=0, where R tot (M) is the total quantity of receptors. II) Model fitting and parameter estimation The CAAX induced BRET ratio was taken as proportional to the loss of receptors at plasma membrane, R-R tot , the β-arrestin-2 induced BRET ratio was taken as proportional to the quantity of receptor-agonist-β-arrestin-2 complex RAB , the FYVE-induced BRET ratio is taken proportional to the quantity of internalised receptor RI , and the miniGq induced BRET ratio was taken as proportional to the quantity of activated Gq protein, with standard gaussian measurement errors with unknown variance, as follows: (3) CAAXinducedBRETratio = y 1 = k bret,1 ⋅ (R-R tot ) + ε 0,σ1 Barr 2 inducedBRETratio = y 2 = k bret,2 ⋅ RAB + ε 0,σ2 FyveinducedBRETratio = y 1 = k bret,3 ⋅ RI + ε 0,σ3 GqinducedBRETratio = y 4 = k bret,4 ⋅ Gq + ε 0,σ4 where k bret,1 , k bret,2 , k bret,3 nd k bret,4 are proportional constant, and ε 0,σ1 , ε 0,σ2 , ε 0,σ3 and ε 0,σ4 are standard centred standard gaussian distribution of variances σ 1 , σ 2 σ 3 and σ 4 respectively. As detailed in (Raue et al., 2013), model adimentionalisation and reparameterisation is a prerequisite for a successful and meaningful estimation. To that, we divided each variable by R tot and transformed the system (2) into the following equivalent system: (4) and In the normalised ODE system (4) , k int and k des have (min -1 ) unit, A is the (known) agonist concentration, K D has molar unit, and all other constants are unitless. This model has a total of 13 unknown parameters. In order to improve parameter identifiability, we used a global fitting approach for all ligand on a given receptor. All observable parameters are assumed to be only receptor dependant and shared across ligands on the same receptor, leading to seven parameters by receptor: , while the remaining kinetic and affinity parameters, are both ligand and receptor dependant. Given that the Gq kinetic profile is very fast, we further impose that the desensitisation constant k des (which dictates Gq timescale) is maximal for the reference ligand (OT on OTR, AVP on V 1A V 1B and V 2 ) . Parameter search intervals are given in Table S3 . We solved the system of ODE (4) using the AMICI package (Fröhlich et al., 2021) and performed parameter estimation following the maximum likelihood approach. We specified the parameter estimation problem following the Petab (Schmiester et al., 2021) format and solved it using the pyPESTO toolbox (Schälte et al., 2023), in Python. Parameter confidence intervals were determined using the profile likelihood approach, implemented in pyPESTO. In addition, we verified that the parameter estimation problem is theoretically well-posed using the StructuralIdentifiability package in Julia (Dong et al., 2023), which shows that all parameters of model (4) are structurally identifiable given data on y 1 , y 2 , y 3 for at least two agonist doses. We report in Table S3 maximum likelihood estimates with their profile-based confidence intervals. Whenever meaningful (finite and symmetric confidence intervals), Wald test was used to test significant parameter difference using the maximum likelihood estimate as the mean value, and the width of the 95% confidence intervals as 1.96 times the standard error on the mean. Animals All mouse breeding, care and experimental procedures were in accordance with the European and French Directives and approved by the local ethical committee CEEA Val de Loire N°19 and the French ministry of teaching, research and innovation (APAFIS #18035-2018121213436249). Following heterozygous breeding scheme, independent cohorts of Fmr1 KO (provided by Rob Willensem (Mientjes et al., 2006)) and WT mice were generated from a minimum of 3 different homozygous non-inbred couples maintained on a mixed 50%-50% C57BL/6J;129S2 background, recapitulating better human genetic heterogeneity (Dudas et al., 2025), in the same breeding room of the animal facility. Mice were outcrossed with fresh mixed backgrounds every 5-10 generations, and between them, with WT mice from other lines, to prevent inbreeding and ensure consistency across independent batches. Fmr1 KO mice replicates aspects of the human Fragile X syndrome (Kat et al., 2022; Giua et al., 2024; Gora et al., 2024), such as deletion of Fmr1 expression and robust social impairments, providing clinical relevance of the study. Two-month-old naïve males and females (7-8 weeks old at the start of the experiments, with balanced female/male sex ratio reported in Table S7 ) from Fmr1 KO and WT animals (20-30g) were raised in groups of 2-4 animals in type 2 cages on a 12-hour light/dark regular cycle, with food and water ad libitum and controlled temperature (21°C) and humidity (50%) in conventional health housing status, exempt from any monitored viral, bacterial, mycoplasma, fungi, parasites or pathological lesions, except mouse norovirus and helicobacter spp . Power analysis of animal sample size calculated the use of minimum 8 mice per group (power of 80%, alpha level of 5% (p = 0.05), differences in means of 41%). The number of animals used in the behavioural experiments was 8-16 animals per group (8–12 animals/compound group, and 16 animals/vehicle groups; unequal and small difference in group sizes were accounted by the use of the non-parametric Kruskal-Wallis tests) for the following reasons: 1) all animals housed together (2–4 per cage) were treated with the same ligand and dose and tested to avoid confounding factors related to differential treatment effects on social interactions or dominance hierarchies, 2) to avoid loss of statistical power in this 4-week longitudinal study in case of animal death or non-recognition by the LMT, and, 3) vehicle-treated animals were included in every cohort as internal controls. Cages of WT and Fmr1 KO mice were randomly allocated to treatment and dose. All animal criteria have been reported in agreement with the ARRIVE guidelines (Kilkenny et al., 2010) . The conducted research was not preregistered with an analysis plan in an independent, institutional registry. After experiment, animals were humanely euthanised using a combined O₂ and CO₂ approach, following veterinary best practice guidelines for mouse group euthanasia. Mouse intraperitoneal single dose pharmacokinetics with RO6893074 and V 1A Receptor occupancy calculation An exploratory single-dose pharmacokinetic study in three C57BL/6J male mice (Charles River) was conducted with RO6893074 V 1A small molecule antagonist following intraperitoneal administration at the doses of 5 mg.kg -1 and 50 mg.kg -1 . The compound was formulated as a microsuspension in HPMC / DOSS (1.25% / 0.1%) / Parabens, with a pH of 6. Blood samples were collected at 6 time points: 0.25, 0.5, 1, 3, 5, 7 hours after administration and the compound was detected using mass spectrometry. Pharmacokinetic parameters were derived from the plasma concentration versus time profile. Using in vitro plasma free fraction (fup = 6.3%) and in vitro V 1A binding data (mouse V 1A Ki = 39 nM, n = 3), the following Emax equation was used to calculate receptor occupancy at Cmax (assuming competitive inhibition): where Bmax is the maximum binding, assumed to be 100%, Cp is the plasma concentration of the compound, fup is the free fraction in plasma, and Ki represents in vitro binding to the receptor. Drug administration For in vivo assays, OT, AVP and TGOT were diluted at 20 and 40 µg.kg −1 in NaCl 0.9% (190/12604022/1021, Braun, Germany), corresponding to low and moderate doses (0.3 and 0.6 IU, respectively, according to OT conversion to UI (Uvnäs-Moberg et al., 2019)) at 4°C and stored in aliquot for daily use at −20°C (vehicle: NaCl 0.9% saline (SAL)). RO6958375 and RO6893074 compounds were formulated at Roche Pharmaceutical at 0.03, 0.06 and 0.12 mg.kg −1 in PBS 1X for subcutaneous administration of RO6958375 OTR peptide agonist (vehicle: PBS 1X), and at 25, 50 and 100 mg.kg −1 in 0.1% polysorbate Tween-80, 0.21% citric acid monohydrate, 0.8% sodium chloride, sodium hydroxide 1N at pH 7 diluted in water for intraperitoneal injection of RO6893074 V 1A small molecule antagonist (vehicle: 0.1% polysorbate Tween-80, 0.21% citric acid monohydrate, 0.8% sodium chloride, sodium hydroxide 1N at pH 7) and stored in aliquot for daily use at 4°C until further use. WT and Fmr1 KO mice were administered every morning for 8 consecutive days from day 0 (acute) to day 7 (subchronic) with OT, AVP or TGOT (20 or 40 µg.kg −1 , 0.2 mL.kg −1 , intranasally, 15 minutes prior to behavioural tests), RO6958375 OTR agonist (0.03, 0.06 or 0.12 mg.kg −1 , 10 mL.kg −1 , subcutaneously, 30 minutes prior to behavioural tests) or RO6893074 V 1A antagonist (25, 50 or 100 mg.kg −1 , 10 mL.kg −1 , intraperitoneally, 30 minutes prior to behavioural tests). Timing and doses were chosen based on previous studies (Bales et al., 2013; Meziane et al., 2015; Peñagarikano et al., 2015; Lindenmaier et al., 2022; Janz et al., 2023) and Roche recommendations, along with their respective vehicles (for OT, AVP or TGOT: NaCl 0.9% (SAL), intranasally, 0.2 mL.kg −1 ; for RO6958375: PBS 1x (PBS), intraperitoneally, 10 mL.kg −1 ; for RO6893074: 0.1% polysorbate Tween-80, 0.21% citric acid monohydrate, 0.8% sodium chloride, sodium hydroxide 1N at pH 7 (VEH), subcutaneously, 10 mL.kg −1 ). For intranasal administration, each drop was placed alternately to each nostril until the animal aspirated the drop into its nasal cavity. As both TGOT and RO6958375 OTR agonists can activate the V 2 receptor, we carefully monitored the mouse weight, which remained stable, for potential diuretic effects of V 2 receptor activation. Social interactions using the Live Mouse Tracker One week before the behavioural experiments, RFID chips (APT12 PIT Tag, Biomark, USA) were inserted under the skin on the lateral and ventral side of the abdomen in Fmr1 KO and WT mice under Isoflurane gas anaesthesia (FR/V/4397332 3/2021, Isoflurin®; Axience SAS, France) and a Procaine local anaesthetic (FR/V/6012689 8/2013, Procamidor®; Axience SAS, France) at 40 µg/10g -1 subcutaneously 10 min at the insertion site. All behavioural tests were carried out, when possible, in the morning, to avoid any circadian cycle effect, in a dedicated quiet room and a dim light intensity of 15 lux. Males were always tested before females to prevent any sexual olfactory bias. The Live Mouse Tracker (LMT; Rodent Phenotyping Toolkit, France (de Chaumont et al., 2019)) is a reproducible and automatic tracking device that identifies in real time over 30 behavioural parameters of reciprocal social interaction using constant RFID detection and an infrared–depth sensing camera. Two unfamiliar mice matched by sex, age, genotype and treatment that had never met before were placed in the LMT for 10 min at 15 lux in a transparent red Plexiglass open field (50 x 25 x 30 cm). Floors were covered with a thin layer of fresh wood litter to favour normal locomotion, well-being and reduced anxious-like behaviours. Social interaction tests were performed 3 days before the treatment for habituation, the first day (D0) and the last day (D7) of treatment (acute vs. subchronic effects), as well as four (D11) and seven days (D14) post-treatment. For each mouse line, mice injected with vehicles (SAL, PBS and VEH) were used as control of those injected with the different drug compounds at D0 and D7. The SAL, PBS and VEH WT group consisted of independent batches of WT mice, serving as the controls for each mouse line or treatment condition and batch. Experimenters were not blinded to the conditions during the tests, as the same person performed compound administration to animals where genotypes were mentioned on the cage label and RFID reading to prevent any exchange of mice while placing the mice according to experimental design (ensuring exposure only to unfamiliar conspecifics across trials) and returned them carefully to their correct home cages. Scoring was conducted automatically for reciprocal social interaction by the Live Mouse Tracker. SQLIte databases containing the coordinates of each mouse and movies were generated for each run of interaction. All individual behavioural parameters were extracted using Python scripts (github.com/fdechaumont/lmt-analysis version v1.0.6). Multi-behavioural parameters were analysed. Description of parameters were extracted from (de Chaumont et al., 2019). Individual events provide motor activity and individual behaviours ‘stop isolated’ refers to ‘the mouse is stopped (speed lower than 5 pixels per frame) and not in contact with any other mouse’, ‘stretched attend posture (SAP)’ as ‘the mouse is in stretched attend posture, i.e., the mouse is moving at a speed lower than 5 pixels per frame, its body length longer than the mean body length plus one standard deviation of body length and the height of the centre of mass is lower than the median height of the centre of mass’ as a proxy for anxious-like behaviours. Dyadic dynamic events as a proxy for social motivation involve ‘move in contact’ described as ‘the mouse is moving at a speed higher than 5 pixels per frame and in contact with another mouse’ and ‘social approach’ as ‘the mouse is approaching another one, i.e. the distance between the two animals shortens, the speed of the mouse is higher than the speed of the approached mouse, and the distance between the two animals is shorter than two mean body lengths of the approached animal (this approach does not necessarily lead to a contact)’. Dyadic state events involve ‘nose contact’ is the sum of ‘contact nose-nose’ for ‘the mouse is sniffing the oral region of another mouse, i.e. the two nose points are less than 15 pixels (26 mm) from one another’ and ‘contact nose-anogenital’ for ‘the mouse is sniffing the ano-genital region of another mouse, i.e. the nose point of the first mouse is within 15 pixels (26 mm) from the tail point of the other mouse’ and ‘huddling’, is ‘the side of a mouse is within 30 pixels (52 mm) from the side of another mouse with the two animals oriented in the same or opposite direction’, the two latter parameters indicated social exploration. Locomotion ‘Move’ has been calculated from the sum of total time ‘move in contact’ and ‘move isolated’ over the total duration of the test. The parameter ‘Nose contact over move’ has been calculated as the percentage of ‘Nose contact’ divided by ‘Move’ parameter. Data, statistic and analysis For in vitro cell assays, each group was constituted of n = 5 independent experiments. Computation and subsequent statistical analysis were performed using R software (version 4.5.2). Raw data were corrected to the baseline values and area under the curve (AUC) were determined over the 30 minutes of the duration of the test while concentrations were transformed in log. AUC values were normalised to the E max AUC of each reference ligand, e.g., OT for OTR or AVP for V 1A , V 1B and V 2 , and represented using dose-response curves. Parameter values E max and log(EC 50 ) of the dose-response curves were determined by the function nlsLM of the R package minpack.lm (version 1.2-4)(Timur V. Elzhov, Katharine M. Mullen, Andrej-Nikolai Spiess, Ben Bolker, 2022). Dose-response curve parameter log(τ/Ka) was determined by a nonlinear mixed-effect model (population approach) using Monolix Suite 2024R1 (Lixoft, Antony, France) for the five experimental independent replicates, normalised to the reference ligand for each receptor and compared between the two effector recruitments with a Wald test as previously used (Namkung et al., 2016b), accounting for the mean and standard error mean. For the Live Mouse Tracker in vivo data analysis, non-parametric Kruskal-Wallis tests, which account for the varying number of animals across groups (8-16 animals per group) and are dedicated for small sample size (Hollander et al., 2014), were conducted using the rstatix package (Alboukadel Kassambara, 2023)(R package version 0.7.2), followed by Dunn’s post hoc tests for multiple comparisons with Holm corrections to adjust P value to prevent false-discovery. P adjusted values < 5% (p < 0.05) were considered significant. No animal outliers were removed from the analysis. Regarding the small sample size, sex was not included as a variable in the study design and analysis. Both the size and colour intensity of P adjusted were scaled according to the dose. All in vitro and in vivo N, sex ratio, means, standard deviations and statistics are available in the supplementary Tables. In vivo raw data, codes and configuration files are available at https://doi.org/10.5281/zenodo.17953414. Movies of behavioural experiments or in vitro raw data are available upon request. ABBREVIATIONS ASD autism spectrum disorders AVP Arginine vasopressin Avpr1A mouse vasopressin V 1A receptor gene Avpr1B mouse vasopressin V 1B receptor gene Avpr2 mouse vasopressin V 2 receptor gene BRET bioluminescence resonance energy transfer DMEM dulbecco’s modified eagle medium EMEM eagle’s minimal essential medium FMR1 Fragile mental retardation 1 gene HBSS hank’s balanced salt solution kB7-OT [Gly⁵,Thr⁷,Ser⁹]-Oxytocin KO knockout LMT live mouse tracker Neuro-2a mouse neuroblastoma Neuro-2a cell line OT oxytocin OTR oxytocin receptor Oxtr mouse oxytocin receptor gene RLuc8 sea pansy renilla reniformis luciferase SAP stretch-attend postures TGOT [Thr4,Gly7]OT WT wild-type CONFLICTS OF INTEREST PDP, PS and ChG are or were employed by F. Hoffmann-La Roche AG, Basel, Switzerland. The other authors declare no conflict of interest. ACKNOWLEDGEMENTS Mouse breeding and care was performed at INRAE Animal Physiology Facility (https://doi.org/10.15454/1.5573896321728955E12). This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 851231). We thank Dr Pascale Crépieux and Dr Eric Reiter for their advice on the manuscript. LPP and CG acknowledge the LabEx MabImprove (grant ANR-10-LABX-53-01) for the financial support in co-financing CG’s PhD fellowship. We used ChatGPT, developed by OpenAI, for assistance with English editing. AUTHOR CONTRIBUTIONS PS and ChG provided materials and an unpublished tool compound; CG, ChG and LPP designed the experiments; CG, EC, LD, EP and LPP performed the experiments and contributed to the data collection; CG, EC, LD, ChG, RY, NA and LPP contributed to the interpretation of data; RY and NA performed all data integration, modelling and statistical analysis; PDP performed the brain receptor occupancy modelling; CG and LPP wrote the original drafts; CG, LD, EP, PS, PDP, ChG, RY, NA and LPP reviewed and edited the manuscript; LPP contributed to the funding acquisition, project conceptualisation and supervision. FIGURE LEGENDS Figure 1 Pharmacological systemic analysis of OT and AVP effects at murine oxytocin and vasopressin receptors in Neuro-2a cells ( A ) schematic representation of OT (red) and AVP (blue) on miniGq, miniGi, miniGs, ꞵ-arrestin-1 and ꞵ-arrestin-2 recruitment, intracellular Ca 2+ mobilisation, cAMP production, forskolin-induced cAMP inhibition, and membrane CAAX and endosome FYVE internalisation at mouse OTR, V 1A , V 1B , V 2 receptors in the murine Neuro-2a cell line. Radar plots illustrate the relative effect of OT and AVP on the ten intracellular inputs of these four receptors (n = 5 per sensor; Table S1 ). ( B ) Relative dose-response of OT and AVP on miniGq and ꞵ-arrestin-2 recruitments for the four receptors. Bias plot representation of the percentage of miniGq (y-axis) vs . β-arrestin-2 recruitment (x-axis) of OT and AVP respective dose–response curves for the four receptors. Points above the diagonal indicate that OT was biased toward miniGq at V 2 . ( C ) Mathematical model used for equation fitting with miniGq and β-arrestin-2 recruitments, and internalisation profiles of the four receptors. Radar plots summarise the robust predicted affinity (1/K D ) and kinetic parameters of Gq recruitment (k des , k τ2 ), β-arrestin-2 recruitment (k τ ), internalisation (k int ) and recycling (k rec ) for each receptor and endogenous ligand ( Figure S3, Table S3 ). A, agonist; AVP, arginine vasopressin; ext, extracellular space; int, intracellular compartment; OT, oxytocin; R, Receptor; RAB, Receptor-Agonist-β-arrestin-2 complex; RI, internalised receptor. Figure 2 Effect of V 1A antagonist at murine oxytocin and vasopressin receptors in Neuro-2a cells ( A ) Dose-response curves of the V 1A antagonist RO6893074 (green) alone, as well as OT (red) and AVP (blue) on miniGq, and ꞵ-arrestin-2 recruitment profiles at mouse OTR, V 1A , V 1B , V 2 receptors in murine Neuro-2A cell lines. ( B ) Dose-response curves of the V 1A antagonist RO6893074 in the presence of OT and AVP at their EC 80 concentrations on miniGq, and ꞵ-arrestin-2 recruitment profiles at the four receptors. The dose-response curve represents the mean of the biological replicates (n = 5 per sensor; Table S1 ). AVP, arginine vasopressin; OT, oxytocin. Figure 3 Effect of OT, AVP and R06893074 V 1A antagonist on social interaction in Fmr1 KO mice In the Live Mouse Tracker, the cumulative time in ‘nose contact’, ‘huddling’, ‘move in contact’ and ‘social approach’ over a 10 min period were compared following 15 minutes of acute (day 0) or subchronic (day 7) administration of OT, AVP (intranasal) and RO6893074 V 1A antagonist (intraperitoneal) at different doses and after treatment (day 11 and day 14) in Fmr1 KO mice (SAL = NaCl 0.9% (grey); OT 20 µg.kg -1 (pink); OT 40 µg.kg -1 (red); AVP 20 µg.kg -1 (light blue); AVP 40 µg.kg -1 (dark blue); VEH = 0.1% Tween-80 vehicle for RO6893074 (grey); RO6893074 25 mg.kg −1 (light green); RO6893074 50 mg.kg −1 (green); RO6893074 100 mg.kg −1 (dark green)). Data are presented as mean ± sd (n = 8-16 animals/ligand/dose, mean and statistics in Table S6 ). Statistical analysis was conducted using Kruskal-Wallis tests followed by Dunn post hoc tests with asterisks indicating treatment effect ( p = P adjusted with Holm correction, size and colour intensity of p were scaled according to the dose): * p <0.05, ** p <0.01, *** p <0.001, **** p <0.0001. AVP, vasopressin; OT, oxytocin; KO, Fmr1 KO mice; SAL, saline; SAP, Stretch-attend posture; VEH; vehicle; WT, wild-type. Figure 4 Effect of orthologs and synthetic agonists at murine oxytocin and vasopressin receptors in Neuro-2a cells ( A ) Relative dose-response of the orthologs isotocin (slate), mesotocin (khaki), and vasotocin (mauve), as well as OT (red) and AVP (blue) reference ligands for OTR and vasopressin receptors, respectively, on miniGq, and ꞵ-arrestin-2 recruitment profiles at mouse OTR, V 1A , V 1B , V 2 receptors in murine Neuro-2A cell lines. Bias plots of the orthologs on miniGq and β-arrestin-2 recruitments revealed that isotocin and mesotocin were biased toward miniGq at the V 2 (points fall above the diagonal) and vasotocin was biased toward β-arrestin-2 at the V 2 (points fall below the diagonal). ( B ) Relative dose-response of the synthetic agonists TGOT (mustard), RO6958375 (purple), carbetocin (light blue) and KB7 (light brown) on miniGq, and ꞵ-arrestin-2 recruitments at the four receptors. Bias plot of TGOT, RO6958375, carbetocin and kB7 showed that TGOT was biased toward miniGq at the OTR, and RO6958375 and carbetocin were biased toward miniGq at the V 2 . The dose-response curve represents the mean of the biological replicates (n = 5 per sensor; Table S1 ). Figure 5 Effect of TGOT and RO6958375 OTR agonists on social interaction in Fmr1 KO mice In the Live Mouse Tracker, the cumulative time in ‘nose contact’, ‘huddling’, ‘move in contact’ and ‘social approach’ over a 10 min period were compared following 15 minutes of acute (day 0) or subchronic (day 7) administration of TGOT (intranasal) and RO6958375 (subcutaneous) OTR agonists at different doses and after treatment (day 11 and day 14) in Fmr1 KO mice (SAL= NaCl 0.9% for TGOT (grey); TGOT 20 µg.kg -1 (light brown); TGOT 40 µg.kg -1 (brown); PBS = PBS 1X vehicle for RO6958375 (grey); RO6958375 0.03 mg.kg −1 (light violet); RO6958375 0.06 mg.kg −1 (violet); RO6958375 0.12 mg.kg −1 (dark violet)). Data are presented as mean ± sd (n = 8-16 animals/ligand/dose, mean and statistics in Table S6 ). Statistical analysis was conducted using Kruskal-Wallis tests followed by Dunn post hoc tests with asterisks indicating treatment effect (p = P adjusted with Holm correction, size and colour intensity of p were scaled according to the dose): *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. KO, Fmr1 KO mice; OTR, oxytocin receptor; PBS, Phosphate-buffered saline; SAL, saline; TGOT, (Thr⁴,Gly⁷)-oxytocin; WT, wild-type. Figure 6: Kinetic mathematical modelling of TGOT and RO6958375 OTR agonists at murine OTR in Neuro-2a cells Fitting ( A ), model ( B ), radar plot ( C ) and kinetics parameters ( D ) of TGOT (mustard), RO6958375 (purple), OT (red) and AVP (blue) on miniGq and ꞵ-arrestin-2 recruitments and FYVE internalisation profiles at mouse OTR in murine Neuro-2A cells (n = 5 per sensor; Table S3 ). AVP, vasopressin; nd, not determined; ns, not significant; OT, oxytocin; TGOT, (Thr⁴,Gly⁷)-oxytocin. TABLE LEGENDS Table 1: Bias factors of the ligands on murine oxytocin and vasopressin receptors in Neuro-2a Data are presented as mean ± s.e.m. (n = 5 independent experiments per sensor; Table S1 ). Statistical analysis was conducted using Wald test with asterisks indicating bias effect ( p = P value): *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. §, ligand of reference, AVP, vasopressin; OT, oxytocin; TGOT, (Thr⁴,Gly⁷)-oxytocin. Supplementary Material File (gora_ra2_table_1_bias_r2.xlsx) Download 11.36 KB References 1. Alboukadel Kassambara (2023). rstatix: Pipe-Friendly Framework for Basic Statistical Tests.Alexander, S.P., Christopoulos, A., Davenport, A.P., Kelly, E., Mathie, A., Peters, J.A., et al. (2021). THE CONCISE GUIDE TO PHARMACOLOGY 2021/22: G protein-coupled receptors. Br J Pharmacol 178 Suppl 1 : S27–S156.American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association).Andrés, M., Trueba, M., and Guillon, G. (2002). Pharmacological characterization of F-180: a selective human V(1a) vasopressin receptor agonist of high affinity. Br J Pharmacol 135 : 1828–1836.Annamneedi, A., Gora, C., Dudas, A., Leray, X., Bozon, V., Crépieux, P., et al. (2023). Towards the convergent therapeutic potential of G protein‐coupled receptors in autism spectrum disorders. British J Pharmacology bph.16216.Bales, K.L., Perkeybile, A.M., Conley, O.G., Lee, M.H., Guoynes, C.D., Downing, G.M., et al. (2013). Chronic intranasal oxytocin causes long-term impairments in partner preference formation in male prairie voles. Biol Psychiatry 74 : 180–188.Bernaerts, S., Boets, B., Bosmans, G., Steyaert, J., and Alaerts, K. (2020). Behavioral effects of multiple-dose oxytocin treatment in autism: a randomized, placebo-controlled trial with long-term follow-up. Mol Autism 11 : 6.Birnbaumer, M. (2000). Vasopressin receptors. Trends Endocrinol Metab 11 : 406–410.Bolognani, F., Del Valle Rubido, M., Squassante, L., Wandel, C., Derks, M., Murtagh, L., et al. (2019). A phase 2 clinical trial of a vasopressin V1a receptor antagonist shows improved adaptive behaviors in men with autism spectrum disorder. Sci Transl Med 11 : eaat7838.Busnelli, M., Bulgheroni, E., Manning, M., Kleinau, G., and Chini, B. (2013). Selective and potent agonists and antagonists for investigating the role of mouse oxytocin receptors. J Pharmacol Exp Ther 346 : 318–327.Busnelli, M., Saulière, A., Manning, M., Bouvier, M., Galés, C., and Chini, B. (2012). Functional selective oxytocin-derived agonists discriminate between individual G protein family subtypes. J Biol Chem 287 : 3617–3629.Camps, M., Hou, C., Sidiropoulos, D., Stock, J.B., Jakobs, K.H., and Gierschik, P. (1992). Stimulation of phospholipase C by guanine-nucleotide-binding protein beta gamma subunits. Eur J Biochem 206 : 821–831.Chaumont, F. de, Ey, E., Torquet, N., Lagache, T., Dallongeville, S., Imbert, A., et al. (2019). Real-time analysis of the behaviour of groups of mice via a depth-sensing camera and machine learning. Nat Biomed Eng 3 : 930–942.Daniels, N., Moerkerke, M., Steyaert, J., Bamps, A., Debbaut, E., Prinsen, J., et al. (2023). Effects of multiple-dose intranasal oxytocin administration on social responsiveness in children with autism: a randomized, placebo-controlled trial. Mol Autism 14 : 16.De Pascali, F., Ayoub, M.A., Benevelli, R., Sposini, S., Lehoux, J., Gallay, N., et al. (2021). Pharmacological Characterization of Low Molecular Weight Biased Agonists at the Follicle Stimulating Hormone Receptor. Int J Mol Sci 22 : 9850.De Rubeis, S., He, X., Goldberg, A.P., Poultney, C.S., Samocha, K., Cicek, A.E., et al. (2014). Synaptic, transcriptional and chromatin genes disrupted in autism. Nature 515 : 209–215.Dong, R., Goodbrake, C., Harrington, H.A., and Pogudin, G. (2023). Differential Elimination for Dynamical Models via Projections with Applications to Structural Identifiability. SIAM J. Appl. Algebra Geometry 7 : 194–235.Drobac, E., Tricoire, L., Chaffotte, A.-F., Guiot, E., and Lambolez, B. (2010). Calcium imaging in single neurons from brain slices using bioluminescent reporters. J Neurosci Res 88 : 695–711.Dudas, A., Novak, A., Gora, C., Pecnard, E., Azzopardi, N., Morisset-Lopez, S., et al. (2025). Mixed genetic background better recapitulates developmental and psychiatric phenotypes and heterogeneity than inbred C57BL/6J mice. Sci Rep 15 : 39130.Elands, J., Barberis, C., and Jard, S. (1988). [3H]-[Thr4,Gly7]OT: a highly selective ligand for central and peripheral OT receptors. Am J Physiol 254 : E31-38.Ford, C.L., and Young, L.J. (2022). Refining oxytocin therapy for autism: context is key. Nat Rev Neurol 18 : 67–68.Fröhlich, F., Weindl, D., Schälte, Y., Pathirana, D., Paszkowski, Ł., Lines, G.T., et al. (2021). AMICI: high-performance sensitivity analysis for large ordinary differential equation models. Bioinformatics 37 : 3676–3677.Giua, G., Strauss, B., Lassalle, O., Chavis, P., and Manzoni, O.J. (2024). Adaptive group behavior of Fragile X mice in unfamiliar environments. Prog Neuropsychopharmacol Biol Psychiatry 135 : 111111.Goodwin, T.M., Paul, R., Silver, H., Spellacy, W., Parsons, M., Chez, R., et al. (1994). The effect of the oxytocin antagonist atosiban on preterm uterine activity in the human. Am J Obstet Gynecol 170 : 474–478.Gora, C., Dudas, A., Vaugrente, O., Drobecq, L., Pecnard, E., Lefort, G., et al. (2024). Deciphering autism heterogeneity: a molecular stratification approach in four mouse models. Transl Psychiatry 14 : 416.Grimwood, S., and Hartig, P.R. (2009). Target site occupancy: emerging generalizations from clinical and preclinical studies. Pharmacol Ther 122 : 281–301.Guastella, A.J., Boulton, K.A., Whitehouse, A.J.O., Song, Y.J., Thapa, R., Gregory, S.G., et al. (2023). The effect of oxytocin nasal spray on social interaction in young children with autism: a randomized clinical trial. Mol Psychiatry 28 : 834–842.Guastella, A.J., Einfeld, S.L., Gray, K.M., Rinehart, N.J., Tonge, B.J., Lambert, T.J., et al. (2010). Intranasal oxytocin improves emotion recognition for youth with autism spectrum disorders. Biol Psychiatry 67 : 692–694.Haider, R.S., Matthees, E.S.F., Drube, J., Reichel, M., Zabel, U., Inoue, A., et al. (2022). β-arrestin1 and 2 exhibit distinct phosphorylation-dependent conformations when coupling to the same GPCR in living cells. Nat Commun 13 : 5638.Hauser, A.S., Attwood, M.M., Rask-Andersen, M., Schiöth, H.B., and Gloriam, D.E. (2017). Trends in GPCR drug discovery: new agents, targets and indications. Nat Rev Drug Discov 16 : 829–842.Heydenreich, F.M., Plouffe, B., Rizk, A., Milić, D., Zhou, J., Breton, B., et al. (2022). Michaelis-Menten Quantification of Ligand Signaling Bias Applied to the Promiscuous Vasopressin V2 Receptor. Mol Pharmacol 102 : 139–149.Hoare, S.R.J., Tewson, P.H., Quinn, A.M., Hughes, T.E., and Bridge, L.J. (2020). Analyzing kinetic signaling data for G-protein-coupled receptors. Sci Rep 10 : 12263.Hollander, E., Jacob, S., Jou, R., McNamara, N., Sikich, L., Tobe, R., et al. (2022). Balovaptan vs Placebo for Social Communication in Childhood Autism Spectrum Disorder: A Randomized Clinical Trial. JAMA Psychiatry 79 : 760–769.Hollander, M., Chicken, E., and Wolfe, D.A. (2014). Nonparametric statistical methods (Hoboken, New Jersey: John Wiley & Sons, Inc).Jacob, S., Veenstra-VanderWeele, J., Murphy, D., McCracken, J., Smith, J., Sanders, K., et al. (2022). Efficacy and safety of balovaptan for socialisation and communication difficulties in autistic adults in North America and Europe: a phase 3, randomised, placebo-controlled trial. Lancet Psychiatry 9 : 199–210.Janz, P., Knoflach, F., Bleicher, K., Belli, S., Biemans, B., Schnider, P., et al. (2023). Selective oxytocin receptor activation prevents prefrontal circuit dysfunction and social behavioral alterations in response to chronic prefrontal cortex activation in male rats. Front Cell Neurosci 17 : 1286552.Jiang, L.I., Collins, J., Davis, R., Lin, K.-M., DeCamp, D., Roach, T., et al. (2007). Use of a cAMP BRET sensor to characterize a novel regulation of cAMP by the sphingosine 1-phosphate/G13 pathway. J Biol Chem 282 : 10576–10584.Kamal, M., Marquez, M., Vauthier, V., Leloire, A., Froguel, P., Jockers, R., et al. (2009). Improved donor/acceptor BRET couples for monitoring beta-arrestin recruitment to G protein-coupled receptors. Biotechnol J 4 : 1337–1344.Kat, R., Arroyo-Araujo, M., Vries, R.B.M. de, Koopmans, M.A., Boer, S.F. de, and Kas, M.J.H. (2022). Translational validity and methodological underreporting in animal research: A systematic review and meta-analysis of the Fragile X syndrome (Fmr1 KO) rodent model. Neurosci Biobehav Rev 139 : 104722.Kato, Y., Igarashi, N., Hirasawa, A., Tsujimoto, G., and Kobayashi, M. (1995). Distribution and developmental changes in vasopressin V2 receptor mRNA in rat brain. Differentiation 59 : 163–169.Kenakin, T., Watson, C., Muniz-Medina, V., Christopoulos, A., and Novick, S. (2012). A simple method for quantifying functional selectivity and agonist bias. ACS Chem Neurosci 3 : 193–203.Kilkenny, C., Browne, W.J., Cuthill, I.C., Emerson, M., and Altman, D.G. (2010). Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLoS Biol 8 : e1000412.Koehbach, J., O’Brien, M., Muttenthaler, M., Miazzo, M., Akcan, M., Elliott, A.G., et al. (2013). Oxytocic plant cyclotides as templates for peptide G protein-coupled receptor ligand design. Proc Natl Acad Sci U S A 110 : 21183–21188.Kolb, P., Kenakin, T., Alexander, S.P.H., Bermudez, M., Bohn, L.M., Breinholt, C.S., et al. (2022). Community guidelines for GPCR ligand bias: IUPHAR review 32. Br J Pharmacol 179 : 3651–3674.Lee, J., Kwag, R., Lee, S., Kim, D., Woo, J., Cho, Y., et al. (2021). Discovery of G Protein-Biased Ligands against 5-HT7R. J Med Chem 64 : 7453–7467.Lee, S.A., Eyeson, R., Cheever, M.L., Geng, J., Verkhusha, V.V., Burd, C., et al. (2005). Targeting of the FYVE domain to endosomal membranes is regulated by a histidine switch. Proc Natl Acad Sci U S A 102 : 13052–13057.Leng, G., Leng, R.I., and Ludwig, M. (2022). Oxytocin-a social peptide? Deconstructing the evidence. Philos Trans R Soc Lond B Biol Sci 377 : 20210055.Lewis, E.M., Stein-O’Brien, G.L., Patino, A.V., Nardou, R., Grossman, C.D., Brown, M., et al. (2020). Parallel Social Information Processing Circuits Are Differentially Impacted in Autism. Neuron 108 : 659-675.e6.Lindenmaier, Z., Ellegood, J., Stuive, M., Easson, K., Yee, Y., Fernandes, D., et al. (2022). Examining the effect of chronic intranasal oxytocin administration on the neuroanatomy and behavior of three autism-related mouse models. Neuroimage 257 : 119243.Marir, R., Virsolvy, A., Wisniewski, K., Mion, J., Haddou, D., Galibert, E., et al. (2013). Pharmacological characterization of FE 201874, the first selective high affinity rat V1A vasopressin receptor agonist. Br J Pharmacol 170 : 278–292.Meziane, H., Schaller, F., Bauer, S., Villard, C., Matarazzo, V., Riet, F., et al. (2015). An Early Postnatal Oxytocin Treatment Prevents Social and Learning Deficits in Adult Mice Deficient for Magel2, a Gene Involved in Prader-Willi Syndrome and Autism. Biol Psychiatry 78 : 85–94.Mientjes, E.J., Nieuwenhuizen, I., Kirkpatrick, L., Zu, T., Hoogeveen-Westerveld, M., Severijnen, L., et al. (2006). The generation of a conditional Fmr1 knock out mouse model to study Fmrp function in vivo. Neurobiol Dis 21 : 549–555.Namkung, Y., Le Gouill, C., Lukashova, V., Kobayashi, H., Hogue, M., Khoury, E., et al. (2016a). Monitoring G protein-coupled receptor and β-arrestin trafficking in live cells using enhanced bystander BRET. Nat Commun 7 : 12178.Namkung, Y., Radresa, O., Armando, S., Devost, D., Beautrait, A., Le Gouill, C., et al. (2016b). Quantifying biased signaling in GPCRs using BRET-based biosensors. Methods 92 : 5–10.Nehmé, R., Carpenter, B., Singhal, A., Strege, A., Edwards, P.C., White, C.F., et al. (2017). Mini-G proteins: Novel tools for studying GPCRs in their active conformation. PLoS One 12 : e0175642.Ostrowski, N.L., Lolait, S.J., Bradley, D.J., O’Carroll, A.M., Brownstein, M.J., and Young, W.S. (1992). Distribution of V1a and V2 vasopressin receptor messenger ribonucleic acids in rat liver, kidney, pituitary and brain. Endocrinology 131 : 533–535.Pan, L., Zheng, L., Wu, X., Zhu, Z., Wang, S., Lu, Y., et al. (2022). A short period of early life oxytocin treatment rescues social behavior dysfunction via suppression of hippocampal hyperactivity in male mice. Mol Psychiatry 27 : 4157–4171.Pantouli, F., Pujol, C.N., Derieux, C., Fonteneau, M., Pellissier, L.P., Marsol, C., et al. (2024). Acute, chronic and conditioned effects of intranasal oxytocin in the mu-opioid receptor knockout mouse model of autism: Social context matters. Neuropsychopharmacology.Park, S.M., Chen, M., Schmerberg, C.M., Dulman, R.S., Rodriguiz, R.M., Caron, M.G., et al. (2016). Effects of β-Arrestin-Biased Dopamine D2 Receptor Ligands on Schizophrenia-Like Behavior in Hypoglutamatergic Mice. Neuropsychopharmacology 41 : 704–715.Parker, K.J., Oztan, O., Libove, R.A., Mohsin, N., Karhson, D.S., Sumiyoshi, R.D., et al. (2019). A randomized placebo-controlled pilot trial shows that intranasal vasopressin improves social deficits in children with autism. Sci Transl Med 11 : eaau7356.Parker, K.J., Oztan, O., Libove, R.A., Sumiyoshi, R.D., Jackson, L.P., Karhson, D.S., et al. (2017). Intranasal oxytocin treatment for social deficits and biomarkers of response in children with autism. Proc Natl Acad Sci U S A 114 : 8119–8124.Passoni, I., Leonzino, M., Gigliucci, V., Chini, B., and Busnelli, M. (2016). Carbetocin is a Functional Selective Gq Agonist That Does Not Promote Oxytocin Receptor Recycling After Inducing β-Arrestin-Independent Internalisation. J Neuroendocrinol 28 : n/a.Peñagarikano, O., Lázaro, M.T., Lu, X.-H., Gordon, A., Dong, H., Lam, H.A., et al. (2015). Exogenous and evoked oxytocin restores social behavior in the Cntnap2 mouse model of autism. Sci Transl Med 7 : 271ra8.Pieretti, M., Zhang, F.P., Fu, Y.H., Warren, S.T., Oostra, B.A., Caskey, C.T., et al. (1991). Absence of expression of the FMR-1 gene in fragile X syndrome. Cell 66 : 817–822.Pintacuda, G., Hsu, Y.-H.H., Tsafou, K., Li, K.W., Martín, J.M., Riseman, J., et al. (2023). Protein interaction studies in human induced neurons indicate convergent biology underlying autism spectrum disorders. Cell Genomics 100250.Quintana, D.S., Rokicki, J., Meer, D. van der, Alnæs, D., Kaufmann, T., Córdova-Palomera, A., et al. (2019). Oxytocin pathway gene networks in the human brain. Nat Commun 10 : 668.Raue, A., Schilling, M., Bachmann, J., Matteson, A., Schelker, M., Kaschek, D., et al. (2013). Lessons learned from quantitative dynamical modeling in systems biology. PLoS One 8 : e74335.Rigney, N., Vries, G.J. de, Petrulis, A., and Young, L.J. (2022). Oxytocin, Vasopressin, and Social Behavior: From Neural Circuits to Clinical Opportunities. Endocrinology 163 : bqac111.Satterstrom, F.K., Kosmicki, J.A., Wang, J., Breen, M.S., De Rubeis, S., An, J.-Y., et al. (2020). Large-Scale Exome Sequencing Study Implicates Both Developmental and Functional Changes in the Neurobiology of Autism. Cell 180 : 568-584.e23.Schälte, Y., Fröhlich, F., Jost, P.J., Vanhoefer, J., Pathirana, D., Stapor, P., et al. (2023). pyPESTO: a modular and scalable tool for parameter estimation for dynamic models. Bioinformatics 39 : btad711.Schmiester, L., Schälte, Y., Bergmann, F.T., Camba, T., Dudkin, E., Egert, J., et al. (2021). PEtab-Interoperable specification of parameter estimation problems in systems biology. PLoS Comput Biol 17 : e1008646.Schnider, P., Bissantz, C., Bruns, A., Dolente, C., Goetschi, E., Jakob-Roetne, R., et al. (2020). Discovery of Balovaptan, a Vasopressin 1a Receptor Antagonist for the Treatment of Autism Spectrum Disorder. J Med Chem 63 : 1511–1525.Stevenson, E.L., and Caldwell, H.K. (2012). The vasopressin 1b receptor and the neural regulation of social behavior. Horm Behav 61 : 277–282.Strakova, Z., Copland, J.A., Lolait, S.J., and Soloff, M.S. (1998). ERK2 mediates oxytocin-stimulated PGE2 synthesis. Am J Physiol 274 : E634-641.Strakova, Z., and Soloff, M.S. (1997). Coupling of oxytocin receptor to G proteins in rat myometrium during labor: Gi receptor interaction. Am J Physiol 272 : E870-876.Terrillon, S., Durroux, T., Mouillac, B., Breit, A., Ayoub, M.A., Taulan, M., et al. (2003). Oxytocin and vasopressin V1a and V2 receptors form constitutive homo- and heterodimers during biosynthesis. Mol Endocrinol 17 : 677–691.Theofanopoulou, C., Andirkó, A., Boeckx, C., and Jarvis, E.D. (2022). Oxytocin and vasotocin receptor variation and the evolution of human prosociality. Compr Psychoneuroendocrinol 11 : 100139.Timur V. Elzhov, Katharine M. Mullen, Andrej-Nikolai Spiess, Ben Bolker (2022). minpack.lm: R Interface to the Levenberg-Marquardt Nonlinear Least-Squares Algorithm Found in MINPACK, Plus Support for Bounds.Umbricht, D., Del Valle Rubido, M., Hollander, E., McCracken, J.T., Shic, F., Scahill, L., et al. (2017). A Single Dose, Randomized, Controlled Proof-Of-Mechanism Study of a Novel Vasopressin 1a Receptor Antagonist (RG7713) in High-Functioning Adults with Autism Spectrum Disorder. Neuropsychopharmacology 42 : 1914–1923.Uvnäs-Moberg, K., Ekström-Bergström, A., Berg, M., Buckley, S., Pajalic, Z., Hadjigeorgiou, E., et al. (2019). Maternal plasma levels of oxytocin during physiological childbirth - a systematic review with implications for uterine contractions and central actions of oxytocin. BMC Pregnancy Childbirth 19 : 285.Valstad, M., Alvares, G.A., Egknud, M., Matziorinis, A.M., Andreassen, O.A., Westlye, L.T., et al. (2017). The correlation between central and peripheral oxytocin concentrations: A systematic review and meta-analysis. Neurosci Biobehav Rev 78 : 117–124.Verheij, C., Bakker, C.E., Graaff, E. de, Keulemans, J., Willemsen, R., Verkerk, A.J., et al. (1993). Characterization and localization of the FMR-1 gene product associated with fragile X syndrome. Nature 363 : 722–724.Wan, Q., Okashah, N., Inoue, A., Nehmé, R., Carpenter, B., Tate, C.G., et al. (2018). Mini G protein probes for active G protein-coupled receptors (GPCRs) in live cells. J Biol Chem 293 : 7466–7473.Young, W.S., Li, J., Wersinger, S.R., and Palkovits, M. (2006). The vasopressin 1b receptor is prominent in the hippocampal area CA2 where it is unaffected by restraint stress or adrenalectomy. Neuroscience 143 : 1031–1039.Zeidan, J., Fombonne, E., Scorah, J., Ibrahim, A., Durkin, M.S., Saxena, S., et al. (2022). Global prevalence of autism: A systematic review update. Autism Res 15 : 778–790. Google Scholar Information & Authors Information Version history V1 Version 1 23 December 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Authors Affiliations Caroline GORA INRAE Transfert SAS View all articles by this author Nicolas AZZOPARDI INRAE Transfert SAS View all articles by this author Emilia CAIRE INRAE Transfert SAS View all articles by this author Lucile DROBECQ INRAE Transfert SAS View all articles by this author Emmanuel Pecnard INRAE Transfert SAS View all articles by this author Patrick SCHNIDER F Hoffmann-La Roche AG Research and Development Division View all articles by this author Pascale David-Pierson F Hoffmann-La Roche AG Research and Development Division View all articles by this author Romain Yvinec INRAE Transfert SAS View all articles by this author Christophe Grundschober F Hoffmann-La Roche AG Research and Development Division View all articles by this author Lucie P. Pellissier 0000-0001-7085-3242 [email protected] INRAE Transfert SAS View all articles by this author Metrics & Citations Metrics Article Usage 292 views 50 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Caroline GORA, Nicolas AZZOPARDI, Emilia CAIRE, et al. Targeting selectively oxytocin receptor signalling efficiently improves social interaction in Fmr1 KO mice. Authorea . 23 December 2025. DOI: https://doi.org/10.22541/au.176650769.91664930/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text {"doi":"10.22541/au.176650769.91664930/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9ffeacf01d8f8650',t:'MTc3OTQ4Mjg0MQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();