Parenting behaviors in mice: Olfactory mechanisms and features in models of autism spectrum disorders

Rodents, along with numerous other mammals, heavily depend on olfactory cues to navigate their social interactions. Processing of olfactory sensory inputs is mediated by conserved brain circuits that ultimately trigger social behaviors, such as social interactions and parental care. Although innate, parenting is influenced by internal states, social experience, genetics, and the environment, and any significant disruption of these factors can impact the social circuits. Here, we review the molecular mechanisms and social circuits from the olfactory epithelium to central processing that initiate parental behaviors and their dysregulations that may contribute to the social impairments in mouse models of autism spectrum disorders (ASD). We discuss recent advances of the crucial role of olfaction in parental care, its consequences for social interactions, and the reciprocal influence on social interaction impairments in mouse models of ASD .


Introduction
Most mammals live in complex social groups that ensure mutual care and survival.Members of a species instinctively express social behaviors in a stereotypical manner without obvious training via dedicated hardwired neural circuits (Wei et al., 2021).Social interactions start from recognizing the conspecifics by different sensory cues causing either social avoidance or social approach and investigation, ultimately leading to consummation, such as mating, aggression, and parenting (Wei et al., 2021).Although innate, social behaviors are also flexible.Social interactions can experience significant transformations influenced by internal state, reviewed in (Anderson, 2016), early social experience, (Stagkourakis et al., 2020b), and genetic (Bendesky et al., 2017) or environmental factors.Conversely, any substantial dysregulation of these pivotal internal and external factors can result in alterations affecting social circuits (Sato et al., 2023).A combination of genetic, neurological, and environmental factors during pregnancy and early childhood may ultimately develop into neurodevelopmental conditions, including autism spectrum disorders (ASD) (Masini et al., 2020;Modabbernia et al., 2017).The diagnosis of ASD hinges on two core behavioral traits: impairments in social interaction and communication, as well as the presence of stereotyped, restrained, or compulsive behaviors.These traits are often accompanied by co-occurring features, including atypical sensory processing (American Psychiatric Association, 2013).In the realm of essential social behaviors in mammals, parental behavior heavily relies on the accurate processing of sensory cues (McRae et al., 2023).The display of parental behavior requires the caregiver to perceive relevant cues from the offspring (and vice versa) and react accordingly to optimize survival.Consequently, sensory impairments may impact parental care, even among individuals with autism without affecting their children (Saito et al., 2022).Certainly, parental care demands a vast energetic commitment for parents to increase the fitness of their offspring.This investment increases the likelihood of survival and promotes optimal development through a range of stereotypic behaviors, such as nursing, grooming, nest building, transport and defense of young, as well as maintaining their body temperature (Numan and Insel, 2003).
The onset of parenthood involves changes in the parents' behavioral responses, which arise from alterations in their cognitive and affective responses to infant-related sensory cues (Kohl et al., 2017).These cues that activate parental behavior circuits are composed of multimodal sensory stimuli, spanning olfactory, auditory, visual, somatosensory, and gustatory modalities.These different sensory systems work in concert to facilitate parental care (McRae et al., 2023).However, it remains difficult to identify how these distinct sensory modalities integrate and concurrently contribute to such complex behavior.Among these sensory modalities, the olfactory sensory system exhibits high degree of plasticity in the context of parenthood, indicating a vital involvement in both innate and learned aspects of parental behavior (Tachikawa et al., 2013;Vinograd et al., 2017;Wilson et al., 2022).In rodents, olfaction is the primary sense through which individuals participate in social interactions, including parental behaviors (Kohl et al., 2017;Sanchez-Andrade and Kendrick, 2009).The detection of pup-derived cues spans relatively distant (volatiles) and close (non-volatiles) ranges of perception that mediate location and recognition.Furthermore, responses to pups seem to be balanced by the activity of two competing neural pathways: one that promotes parental motivation toward pups, and another that activates aggressive responses to them.These responses are modulated by plastic changes affecting the parents' ability to detect, and discriminate infant-related cues, ultimately suppressing or activating these two opposing pathways (Mei et al., 2023;Miyamichi, 2023;Stolzenberg and Mayer, 2019).Assessing parenting and other olfactory-directed behaviors in rodents is a relatively straightforward task within a laboratory setting, granting the possibility to study well-established genetic models, including those of ASD.In humans, olfaction has been suggested to influence parent-child social interaction (Pause, 2012), although its role remains a vastly underexplored area, particularly its contribution to social impairments in ASD.For these reasons, we focus on olfaction as sensory modality, limiting our discussion to rodents, the group in which the neurobiology of parenting has been studied most extensively.Nonetheless, it should also be kept in mind that most findings on rodent parental behavior come from a very small number of domesticated species, primarily mice, housed under laboratory conditions.These conditions can differ dramatically from the natural ecological, physical, social, and sensory environment, affecting sensory function and behavioral display (Chesler et al., 2002;Nigri et al., 2022).
Therefore, in this review, we provide an overview of the physiological processes involving olfaction that participate in the display of parental care in rodents, as well as the influence of altered social interaction skills in mouse models of ASD.In the first part, we highlight recent studies that have begun to unravel the neural substrates that process the olfactory cues involved in parent-offspring interactions.Specifically, we present a brief overview of the basic olfactory neural pathways involved and examine the detection of pup-derived molecules.We discuss the effects of parenthood on olfactory plasticity that eventually alter parental behaviors and neural responses to stimuli from infants.In the second part, we provide an overview of the olfactory and parental behavior impairments reported in different mouse models of ASD.We also discuss the consequence of these alterations of these neural circuits on core social features in these models.

From olfactory to central mechanisms of parenting behaviors
Odors from pups are largely detected by two related but independent pathways: the main and the accessory (or vomeronasal) olfactory systems.The sensory interfaces of both subsystems reside in the nasal cavity where specialized sensory neurons responsible for the detection of molecules are located.The main olfactory system primarily detects volatile odors by the main olfactory epithelium (MOE), whereas the accessory olfactory system mainly detects non-volatile odors by the vomeronasal organ (VNO) [Reviewed in (Munger et al., 2009)].The circuitry of these two subsystems is anatomically segregated, but overlaps functionally projecting directly or indirectly to forebrain regions implicated in the control of parental behaviors (Mohrhardt et al., 2018;Mori and Sakano, 2024) (Fig. 1).Both olfactory subsystems innervate Fig. 1.Olfactory neuronal circuit processing in parental behaviors.The schematic shows a sagittal view of a rodent brain illustrating the projections of sensory neurons that originate in the vomeronasal organ (VNO) and main olfactory epithelium (MOE).The brain areas innervated by these projections are involved in the regulation of aggressive (red) and parental (orange) behaviors, or both (grey).Solid lines indicate well-studied brain projections involved in social behaviors, whereas dashed lines indicate potential or indirect projections.These brain areas include the olfactory bulb (OB), the accessory olfactory bulb (AOB), the bed nucleus of the stria terminalis (BNST), the medial amygdala (MeA), the medial preoptic area (MPA), lateral septum (LS), and the hypothalamic paraventricular nucleus (PVN).Aggressive behaviors are mainly initiated by the accessory olfactory system (VNO), whereas parenting by the main olfactory system (MOE).Maternal aggression is dependent on both subsystems.Redundant projections indicate a complex interplay in higher order processing brain areas.
limbic system regions (Halpern and Martínez-Marcos, 2003;Martinez-Marcos, 2009;Pro-Sistiaga et al., 2007) that control the display of instinctive behaviors, such as social exploration, aggression and sexual behaviors, but these behavioral responses are also widely modulated by experience and the animal's internal and social states.Most studies in the parental brain have focused on these integrative regions involved in motivational, cognitive, and emotional components of behavior (Kohl and Dulac, 2018;Kuroda and Numan, 2014;Miyamichi, 2023;Stolzenberg and Mayer, 2019).However, less is known about the role of the sensory systems in facilitating parental care.

Detection and processing of social cues by the olfactory system
Chemosignals are secreted by animals to signal social status to conspecifics eliciting diverse physiological and behavioral responses such as aggression and nursing (Wyatt, 2014).Social odors are recognized by sensory neurons residing in the MOE and the VNO (Chamero et al., 2012;Liberles, 2014;Munger et al., 2009;Trotier and Doving, 1998).Lesion experiments and genetic invalidation of signaling pathways specific to these two epithelia have shown the importance of olfaction in parental behavior (Lévy et al., 2004).Olfactory cues from the pup detected by both the MOE and VNO have an essential role in the initiation and coordination of parental behavior in mammals.In female mice, disruption of the whole olfactory system by olfactory bulb (OB) removal abolished nest building and maternal care in lactating and in virgin mice (Gandelman et al., 1971;Vandenbergh, 1973).Other more targeted methods aimed to dissect between MOE or VNO function have also led to impairments in maternal behaviors in mice: gene deletion of adenylyl cyclase-3 (Adcy3), the cyclic nucleotide gated channel subunit alpha 2 (Cnga2), or the mutation of voltage-gated sodium channel Nav1.7 (Scn9a), which are essential parts of the MOE signal transduction in olfactory sensory neurons, resulted in deficient pup retrieval, reduced maternal aggression and deficient nest building (Fraser and Shah, 2014;Matsuo et al., 2015;Wang et al., 2011;Weiss et al., 2011).The role of the VNO in maternal behavior seems to be less critical, at least for the display of pup retrieval.Lactating female mice mutant for the VNO signal transduction channel Trpc2, and the G proteins Gαi2 and Gαo that control vomeronasal receptor activity, show intact pup retrieval behavior (Chamero et al., 2011;Fraser and Shah, 2014;Norlin et al., 2003;Trouillet et al., 2021).Conversely, odor-evoked activity in the VNO was shown to be essential for the display of maternal aggression in mice (Chamero et al., 2011;Del Punta et al., 2002;Fraser and Shah, 2014;Hasen and Gammie, 2009;Kimchi et al., 2007;Leypold et al., 2002;Norlin et al., 2003), indicating that both structures, MOE and VNO, are required to be functional for the display of maternal aggression.Indeed, the main olfactory and vomeronasal systems seem to be coactivated during interaction with pups in female mice during advanced pregnancy stages, both at the level of sensory epithelia and encoding brain structures, including the main and accessory olfactory bulbs (MOB and AOB), piriform cortex, medial amygdala (MeA), posteromedial cortical amygdala and bed nucleus of the stria terminalis (BNST) (Navarro-Moreno et al., 2020).MeA and BNST neurons further innervate specific parts of the ventromedial hypothalamus (VMH) and the medial preoptic area (MPA), regions that are highly implicated in social behavior (Numan and Sheehan, 1997).Specifically, the MPA is activated during parental behaviors and it is one of the most studied brain areas for the control of parental care.(Kohl et al., 2018(Kohl et al., , 2017;;Kohl and Dulac, 2018;Numan and Numan, 1997;Tsuneoka et al., 2015Tsuneoka et al., , 2013;;Wu et al., 2014).Lesions in the MPA of rat mothers resulted in impairment of maternal behavior, and bilateral lesions led to infanticide behavior in nulliparous virgin female mice (Kuroda and Numan, 2014;Numan, 1974;Numan and Sheehan, 1997;Tsuneoka et al., 2013).(Fig. 1).
On the opposing direction of parental care, the AOB, MeA and BNST are activated during pup-directed aggression, following a circuit pattern that broadly overlaps with the vomeronasal system (Mei et al., 2023;Miyamichi, 2023;Stolzenberg and Mayer, 2019;Tachikawa et al., 2013;Trouillet et al., 2019;Tsuneoka et al., 2015;Wu et al., 2014).The MPA is highly interconnected with these regions, including Esr1-expressing neurons in the BNST whose activity is modulated by social experience (Mei et al., 2023).Therefore, circuits promoting aggressive responses to pups seem to be balanced with the ones promoting affective responses towards them, and one or the other would be activated according to the reproductive status of the individual (Mei et al., 2023;Miyamichi, 2023;Stolzenberg and Mayer, 2019).
Pup-derived chemosensory cues stimulate or inhibit parental care according to the sex and physiological status of the receiver.Virgin male mice kept under lab conditions display infanticide of foreign pups, while recently mated males display parental behavior (Tachikawa et al., 2013;Wu et al., 2014).Infanticide is also shown by adult wild female mice captured in nature (McCarthy and Vom Saal, 1986).The vomeronasal pathway plays a key role in inhibiting parental behavior and driving pup-directed aggression in virgin animals.Genetic or surgical inactivation of the VNO suppress infanticide and promotes parental behavior in virgin males (Mennella and Moltz, 1988;Nakahara et al., 2016;Tachikawa et al., 2013;Trouillet et al., 2019;Wu et al., 2014).Virgin males with a genetic inactivation of the apical VNO become parental and show elevated c-Fos expression in the MPA after pup exposure, but also display increased levels of male-male aggression, when compared with controls (Trouillet et al., 2019).This suggests that infanticide involves the activation of a dedicated circuit different from intermale aggression.This is supported by recent studies identifying urocortin-3 (Ucn3)-expressing neurons in the hypothalamic perifornical area as important drivers of pup-directed aggression in male and female mice, but not of male-male aggression (Autry et al., 2021).
Overall, this literature indicates that the olfactory system plays a key role in parental behavior; the vomeronasal system involved in maternal aggression and pup-directed aggressive behaviors and the main olfactory system necessary for adaptive parental behaviors.However, the olfactory cues emitted by pups that facilitate parental care have not been identified.Recent studies using ultra-high pressure liquid chromatography combined with gas chromatography and mass spectrometry have identified 16 compounds enriched in the samples from mouse neonatal vs. older pups (Lacalle-Bergeron et al., 2023, 2021).These putative pup pheromones are both volatile and non-volatile, including amino acids, glycol ethers and sphingolipid derivates, but so far only one of these compounds, 1,2,4,5-tetramethylbenzene (or durene), has been identified as VNO activator (Lacalle-Bergeron et al., 2023, 2021;Sam et al., 2001).The potential role of these molecules in parental behavior is currently unknown.Unexpectedly, two compounds which are found in mouse pups but also in adults, the submandibular gland protein C and hemoglobin beta, have been described to induce infanticide in virgin male mice (Isogai et al., 2018).These compounds can be detected by sensory neurons in the VNO expressing vomeronasal receptors Vmn2r65 and Vmn2r88, respectively.Male mice knockout for these two receptor genes show reduced infanticidal behavior (Isogai et al., 2018).Interestingly, hemoglobin is found in pups after bleeding during parturition, and its detection by Vmn2r88-expressing VNO neurons activate the MeA and the VMH only in lactating female mice, enhancing digging and rearing behaviors (Osakada et al., 2022).These behaviors are associated with an elevated state of alert in lactating females, suggesting that they may be ultimately intended to provide protection to the pups.Other receptors in the VNO have been implicated in pup chemosignal detection and pup-directed aggression, including Vmn2r122/123 and members of the V1R vomeronasal receptor sub-family (Isogai et al., 2018;Trouillet et al., 2019).Notably, atypical VNO neurons that express one type of olfactory receptors (Olfr692) normally expressed by MOE neurons, are activated by pup odors preferentially in virgin males (Nakahara et al., 2016), suggesting a potential role of Olfr692 in pup-directed aggression.The causal relationship between activity of Olfr692-expressing neurons and pup-directed behavior, as well as the pup-emitted ligands for Olfr692 remain to be elucidated.
A. Dudas et al.

Central and peripheral plasticity induced by parenting
In many species, such as mice, virgin animals are infanticidal and undergo a dramatic behavioral change to become caregivers (Miyamichi, 2023;Stolzenberg and Mayer, 2019).Such conversion involves plastic changes at the sensory interface, as well as in the activity of several regions in the brain.The olfactory system shows some evidence of olfactory processing plasticity during pregnancy.Virgin female mice are typically not attracted to infants nor display maternal behavior until late in their first pregnancy.Changes in gene expression and in sensory processing activity of pup-derived chemosignals have been evidenced by immediate early genes analysis and in vivo two-photon calcium imaging, respectively, in the brain of mothers or late pregnant mice (Abellán-Álvaro et al. 2021;Navarro-Moreno et al., 2020;Vinograd et al., 2017).Smell perception changes are also reported by pregnant women (Nordin et al., 2004), although the underlying mechanisms of this effect are not well known.Neural plasticity can manifest at multiple levels of organization, including changes in morphology, number of synapses and electrophysiological properties of neurons.Pregnancy has been reported to increase neurogenesis in the mouse VNO (periphery) and subventricular zones in the brain that ultimately modulate OB organization (Chaker et al., 2023;Oboti et al., 2015;Shingo et al., 2003).This neurogenic process can also be modulated by pheromones and has been suggested to play a key role for normal postpartum behavioral responses in the mother, including maternal behavior (Larsen et al., 2008;Larsen and Grattan, 2010;Lévy et al., 2004;Oboti et al., 2011).Interestingly, virgin females acquire and improve maternal care after co-habitation with mothers and pups (Carcea et al., 2021;Rosenblatt, 1967;Stolzenberg and Rissman, 2011), implicating that experience can override the importance of gestation for the display of maternal behaviors.This type of acquired maternal behavior by social transmission seems to be facilitated by endogenous oxytocin in the brains of virgin female mice.Oxytocin neurons in the hypothalamic paraventricular nucleus (PVN) of virgin females are activated during observation of maternal care by experienced mothers, later promoting alloparenting (Carcea et al., 2021).Both mothers and virgin females rely heavily on sensory detection of pup signals, including olfactory cues, for pup identification and for the display maternal behavior (Carcea et al., 2021;Trouillet et al., 2021).
Similar to the neural plasticity observed in females after pregnancy and delivery, mating and fatherhood also modulate the male central nervous system influencing parental behavior.This plasticity typically involves changes in endocrine and neuropeptide signaling, including androgens, estrogen, progesterone, glucocorticoids, prolactin, vasopressin and oxytocin, acting in multiple central brain structures to orchestrate pup-directed behavior [extensively reviewed in (Horrell et al., 2018)].However, how these hormonal and neuropeptide changes affect sensory function, and particularly the olfactory system, is still poorly understood.One study pointed to a potential role of oxytocin in the modulation of olfactory sensory activity in male mice after social experience with females controlling the male behavioral transition from infanticide to parenting (Nakahara et al., 2020).This study found that VNO sensory activity in response to pups is strongly reduced after cohabitation with a female, which correlates with a sharp decrease in oxytocin receptor (Oxtr) expression in the VNO in the same period.Systemic oxytocin injection in virgin mice reduced both pup-induced vomeronasal activity and pup-directed aggression (Nakahara et al., 2020), suggesting that oxytocin can silence VNO sensory activity regulating this behavioral transition.Oxytocin has also been widely implicated in modulating central olfactory processing (Choe et al., 2015;Oettl et al., 2016), as well as other sensory systems such as the auditory and gustatory systems (Carcea et al., 2021;Marlin et al., 2015;Sinclair et al., 2015).The olfactory processing of pup cues changes dramatically according to the reproductive status of the individuals, both at the sensory detection interface and centrally.The aforementioned studies showed the involvement of oxytocin in this example of neuronal plasticity.
However, it is unlikely that alterations in olfactory activity by oxytocin fully explain the behavioral changes.These also require central modulation by hormones and neuropeptides in specific hypothalamic nodes, such as the MPA, BNST and arcuate nucleus (Horrell et al., 2018;Stagkourakis et al., 2020a;Tachikawa et al., 2013).An intriguing facet of this model is that the ablation of VNO sensory inputs or the disruption of the associated downstream circuitry results in not only a decrease in pup-directed aggression but also the complete expression of parental behavior (Nakahara et al., 2016;Trouillet et al., 2019;Wu et al., 2014).This suggests that in virgin males, VNO sensory inputs typically suppress parental circuits, thereby establishing a potential mechanism for promoting parental behavior by silencing the sensory input.
In sum, the olfactory system plays a vital role in mediating pupdirected behaviors: affiliative behaviors are often driven by the main olfactory system, whereas agonistic behaviors are largely guided by the vomeronasal system, each activating distinct brain circuits.The presence of redundant pup-detection systems indicates a complex interplay between the main and vomeronasal systems in higher order processing brain areas.A significant level of neural plasticity, affecting parental behavior in both males and females, is influenced by various internal and external factors, including pregnancy, delivery and mating.This plasticity encompasses alterations in endocrine and neuropeptide signaling, affecting both the olfactory periphery and central brain structures to regulate pup-directed behavior.Hence, neurodevelopmental conditions, which commonly impact sensory and cognitive functions, can significantly influence offspring detection and the expression of parental behaviors.

Olfaction and parenting in mouse models of ASD
The involvement of specific cortical and subcortical circuits in social behavior and ASD has been largely studied [recently reviewed in (Sato et al., 2023)].However, the importance of neuronal circuits governing olfactory-related and parental behaviors and their impact on social interaction impairments in mouse models of ASD has received less attention.In fact, atypical olfactory processing, altered parental behaviors, or their interplay might contribute to the altered social interaction observed in mouse models of ASD.While parental care does not cause social interaction impairments in children with autism, it can influence positively and negatively children development (Crowell et al., 2019).Olfactory or parental behavior phenotypes have been already reported in thirty-one genetic mouse models of ASD.These models include full or conditional knock-out (KO) or heterozygous, one chromosomal deletion syndrome model and three inbred mouse lines (Balb/c, BTBR and C58/J).Olfactory phenotypes have been observed in different mouse models of autism [reviewed in (Lyons-Warren et al., 2021)], although olfactory responses to both neutral and social odors in these models are seldom reported.It is crucial to assess whether the observed olfactory phenotypes are manifested in response to neutral or social odors, given that impairments in social interactions in autism models could affect the processing of social cues, including social odors.Furthermore, linking the olfactory phenotypes with potential molecular or cellular dysfunctions within the olfactory system is crucial to understand the mechanisms underlying olfactory impairments in these models.Thus, we present here an overview of the expression of ASD-associated genes and the molecular and/or cellular dysfunctions, within the olfactory system of autism mouse models, as well as their olfactory phenotype using neutral and/or social odors.This information is essential for evaluating the parental behavior phenotypes of these models that is often tested in olfactory-based assays.In Table 1 we indicate the expression of these genes in the olfactory and CNS tissues and the phenotypes on olfaction and parental behavior of the corresponding mouse model (Fig. 1, Table 1).We next discuss the potential contribution of these genes to social interaction impairments.

Table 1
Olfaction and parental behaviors in mouse models of ASD.The expression of ASD associated genes and/or proteins were determined in the Allen brain and Human Protein Atlas database (https://www.proteinatlas.org/) and Ibarra-Soria and colleagues (2014) (Ibarra-Soria et al., 2014;Oboti et al., 2015).Behavioral tests for food detection (food localization test), sensitivity (olfactory sensitivity test and flow olfactometer) and discrimination (operant conditioning, two-choice test and habituation/dishabituation test) habituation (habituation/dishabituation test), exploration (sniffing time of neutral and not social odors), detection of odors (olfactory investigation test) for parental behavior we discriminated pup-directed parental care (nursing, crouching, grooming, huddling and retrieval) or neglecting behavior (pup outside of nest, climbing/digging), environment-directed behaviors (nest building), self-maintenance (self-grooming, eating, drinking) and adult-directed aggression.When reported in the same publication, social interaction and social memory/novelty was included in the  (Ferguson et al., 2001(Ferguson et al., , 2000;;Johnson et al., 2023;Nishimori et al., 1996;Young et al., 1997) (continued on next page) A. Dudas et al.

Atypical olfactory processing
Olfaction influences social interactions also in humans (Barros and Soares, 2020;Loos et al., 2023;Pause, 2012).Olfactory dysfunctions in odor detection and identification contribute to the global atypical sensory processing in ASD (Boudjarane et al., 2017;Crow et al., 2020;Kumazaki et al., 2016).For instance, increased sniffing is one of the common examples of stereotyped behavior in children with ASD (Cunningham and Schreibman, 2008).Furthermore, it has been suggested that impaired social olfactory detection could be an early marker of ASD and that use of specific social odors could promote social interactions (Barros and Soares, 2020).In rodents, olfaction shapes social interaction and recognition in various social contexts (Chamero et al., 2012).Molecular, cellular and behavioral changes in olfaction have been reported in some mouse models of ASD.Although some differences exist between the VNO, MOE and OB, most ASD-associated genes are broadly expressed in these olfactory tissues, with the exception of Avrpr1b, Magel2, Oxt and Tph2 (Table 1).This suggests that altered olfactory performance in some models might directly originate from sensory neuron dysfunction in the VNO and MOE.Celf6, Cntnap2, Tbr1, Tshz1 and Ube3a knockout (KO) mice display an overall clear impairment in both neutral and social olfactory detection (Dougherty et al., 2013;Gordon et al., 2016;Huang et al., 2019;Koyavski et al., 2019;Ragancokova et al., 2014).Similarly, Adnp and Fmr1 KO mice, and mice with reduced levels of Grin1 also display impaired responses to neutral odors, although the responses to social odors remain to be investigated (Hacohen-Kleiman et al., 2018;Kuruppath et al., 2023;Moy et al., 2008;Nitenson et al., 2015).Furthermore, molecular and cellular dysfunctions are observed in the OB of Adnp and Fmr1 models (Table 1).Relatedly, the OB of Fmr1, Tbr1 and Tshz1 KO mice is not well developed showing an overall reduced size and shape or reduced number of synaptic components (Galvez et al., 2005;Huang et al., 2019;Ragancokova et al., 2014).In addition, models such as Adnp, Cntnap2 and Tbr1 KO mice show impaired OB activity and connectivity (Geramita et al., 2020;Hacohen-Kleiman et al., 2018;Huang et al., 2019).Ptena and Shank3 KO mice show general alterations in olfactory responses associated with mild abnormalities in the OB when exposed to neutral but not to social odors (Drapeau et al., 2018;Geramita et al., 2020;Yuan et al., 2019).Recent findings have shown that VNO neurons of Shank3 KO mice exhibited reduced responses to conspecific urine compared to WT animals (Gora et al., 2023).Additionally, Shank3 heterozygous mice display impaired odor discrimination within an odor-enriched background (Ryndych et al., 2023), supporting the notion that this model displays broad olfactory impairments.Additional research is needed to explore the implications of these findings in Ptena and other ASD models.Interestingly, Arhgap32, Avpr1b, Crmp4, Kirrel3, Nrxn1 and Tph2 KO mice show altered odor interest in response to social odors, but not to food or neutral odors (Grayton et al., 2013;Hisaoka et al., 2018;Huo et al., 2018;Nakamura et al., 2016;Rabaneda et al., 2014;Tsutiya et al., 2015;Wersinger et al., 2004), suggesting an effect perhaps limited to the vomeronasal system.Similarly, Balb/c, BTBR and C58/J inbred mouse lines also show impaired odor interest when exposed to social odors (Chourbaji et al., 2011;Moy et al., 2014;Yang et al., 2012) while their VNO, MOE, and/or OB morphology is normal, except for Kirrel3 KO mice (Prince et al., 2013).However, fewer neurons are activated in the OB after pup exposure in BTBR females (Tartaglione et al., 2021).Altered olfactory responses reported in other models may arise from atypical processing of social cues in central brain areas.For example, Magel2 KO mice display a decreased preference for both neutral and social odors (Mercer and Wevrick, 2009), even though Magel2 is not expressed in the olfactory system, suggesting dysfunctional olfactory processing in the brain.Social olfaction in Cd38, Oxt and Oxtr KO mice has not been yet investigated (Table 1), but we could speculate that they may exhibit impairments primarily to social odors, based on their social memory phenotype.
Altogether, olfactory phenotypes remain to be assessed in most mouse models of ASD.Importantly, most of these models, with the exception of Tbx1, Pcdh10, Itgb3, Shank1 and Shank2 KO mice, show altered olfactory responses (Table 1), suggesting a potential link of olfactory function with social interaction.It should be also noted that olfactory assays vary between studies and often rely on intrinsic motivation and locomotion of mice.Additionally, most studies do not report responses to both neutral and social odors which could, combined with the diverse olfactory assays used, explain some of the observed discrepancies in the phenotypes.
The number of models exhibiting altered olfaction is expected to increase with the use of odor discrimination tests in complex odor environments and behavior-independent context.These assays, which do not rely on animal interest and motivation, have been successfully tested in Shank3 and Fmr1 KO mice (Kuruppath et al., 2023;Ryndych et al., 2023).Considering the major impact of olfaction on social behaviors, further investigations are required to fully understand the contribution of olfactory dysfunction on the core social traits in mouse models of ASD, including Dhcr7, Pten, Slc6a4, Nlgn3 and Nrcam KO mice among others.Nevertheless, exploring mouse lines with conditional deletion of ASD-associated genes in the olfactory system would provide a better understanding of the role of olfaction in social interaction.

Reciprocal effect of parent-pup interaction in mouse models of ASD
Parental care tightly depends on the detection and processing of olfactory cues and may influence social interaction skills in mice (Chamero et al., 2012;Rymer, 2020).Thus, we review the potential impairments in parental behaviors observed in these mouse models of ASD.Parental care features, such as pup care, maternal aggression and nest building, have been poorly studied, even in the most used mouse models of ASD (Table 1).Maternal behavior is intact in Fmr1 heterozygous, Oxt and Shank3 KO and inbred C58/J females (Nishimori et al., 1996;Ryan et al., 2010;Wong et al., 2019;Young et al., 1997;Zupan et al., 2016).By contrast, Cd38, Tph2, Magel2, Shank2 KO and inbred Balb/c and BTBR mice show impairments in different aspects of maternal behavior, but also paternal behavior in Magel2 and Cd38 KO mice (Akther et al., 2013; Chourbaji et al., 2011;Da Prato et al., 2022;Grabrucker et al., 2021;Higgs et al., 2023;Jin et al., 2007;Tartaglione et al., 2021).In addition, Avpr1b but not Avpr1a KO females show reduced maternal aggression (Wersinger et al., 2007).Furthermore, the high rate of pup mortality in Pten KO suggests that maternal behavior could also be affected (Kwon et al., 2006).Reports of parental behavior in mouse models of ASD remain limited.Nonetheless, half of the models tested showed impairments in parental care.This is not surprising, as parental behaviors were frequently assessed in assays dependent on olfaction, which is impaired in most of the previously mentioned models.Further investigations of parental behaviors in mouse models of ASD, integrating auditory, visual and gustatory cues would aid in determining whether impairments in parental care originate from olfactory deficits (McRae et al., 2023).Furthermore, it remains unknown in these models whether impairments arise from the pup or the parent genotype.Indeed, parental behavior is a reciprocal parent-pup interaction where pups might influence the behavior of the parent (Bailey and Isogai, 2022;Kuroda et al., 2011).Tests involving crossing KO mothers with WT fathers may not be sufficient to clarify this point, as heterozygous pups could also manifest behavioral impairments.One possibility to determine the contribution of pups to the impairment in parental care could be to assess pup ultrasound vocalizations following maternal separation, perform the homing tests or distinguish between olfactory and auditory impairments using anesthetized pups (Iezzi et al., 2022;Servadio et al., 2016).Additional important aspects of parental behavior, including paternal care, nesting, pup retrieval and maternal aggression remains to be investigated in these models.

Consequence of altered olfaction and parenting on social interaction impairments in mouse models of ASD
Mouse models of ASD have greatly enhanced our understanding of this condition and the underlying mechanisms by studying the functional consequences of the deletion of candidate genes.However, olfactory and parental phenotypes are rarely reported for the different mouse models of ASD.Numerous ASD-associated genes are widely expressed in the murine olfactory system and most mouse models of ASD display altered olfactory performance, suggesting a potential contribution of olfactory dysfunction on the social traits (Table 1).Indeed, social interactions in mice often start with detection of conspecifics using mainly olfactory cues, which will trigger either social avoidance or social approach and investigation (Wei et al., 2021).Therefore, olfactory abilities are critical to interpret social skills in mice and olfactory impairment might contribute to impaired social interaction in mouse models of ASD.It is challenging to distinguish whether olfactory phenotypes observed in the models arise from impairments in olfactory processing or from a reduced motivation in mice to explore odors.Indeed, in experiments assessing motivation where mice were trained to press a lever to gain access to another rodent or food, mice from inbred BTBR line showed decreased social and food motivation (Martin et al., 2014).Similarly, reduced social motivation was also observed in Shank3 KO males (Maloney et al., 2023).Consequently, the reduced exploration of odors in these models may be partially attributed to their decreased intrinsic motivation.
Since most of the discussed mouse models of ASD exhibit various degrees of social interaction impairments, the reduced exploration of social odors observed in Arhgap32, Avpr1b, Crmp4, Kirrel3, Nrxn1 and Tph2 KO mice could stem from their reduced interest in social cues.Given that most of the genetic models of autism reviewed here are complete KOs, investigating the social phenotype of models with conditional deletions of ASD-associated genes in structures not directly implicated in olfactory primary processing would be beneficial.For example, mice with Oxtr deletion in forebrain neurons (Lee et al., 2008), or knockdown of Nlgn3 in dopaminergic neurons of the ventral tegmental area (Bariselli et al., 2018) still exhibited social recognition impairments.Though the response to social odors remains to be further investigated, the persistence of the social phenotypes in conditional KO lines suggests that the olfactory phenotypes might be caused by social recognition and motivation impairments.However, mice with conditional Shank3 deletion in forebrain, striatum or striatal D 1 and D 2 cells, as well as conditional deletion of Fmr1 in somatostatin-expressing neurons, did not display the sociability impairments observed in full KO lines (Bey et al., 2018;Kalinowska et al., 2022).This suggests that altered olfactory processing might account for the social phenotype observed in Shank3 and Fmr1 KO mice.Nonetheless, some full KO lines recapitulate human conditions, such as the complete Fmr1 KO mouse mirroring the loss of FMRP in individuals with fragile X syndrome (de Rubeis and Bagni, 2011).
Furthermore, social interaction skills also rely on early social experience, such as parental care.Similarly, parental behavior in mouse models of ASD is poorly documented or not thoroughly assessed.Therefore, it is difficult to estimate the extent of parental behavior impairments in mouse models of ASD and/or linked to olfactory impairments.Considering the significant impact of olfaction on parental behavior in mice (Chamero et al., 2012;Rymer, 2020), there is an urgent need to determine whether impairments in parental care observed in certain ASD models stem from an olfactory origin.Atypical olfactory processing in mouse models of ASD might lead to impaired parental care (Table 1), both potentially contributing to the social features in these models.This seems to be the case for Balb/c and BTBR females, which display impairment in maternal care accompanied with reduced number of activated neurons in the OB after pup exposure (Chourbaji et al., 2011;Tartaglione et al., 2021).Whether olfactory processing, parental behavior, their interplay, as well as social interaction are regularly affected in most mouse models of ASD remains unknown.This point is essential to obtain an appropriate interpretation of the specific dysfunctions of social circuits and connectivity in these models.Conversely, in certain cases impairment in parental behavior could be independent of olfactory function.Indeed, Shank2 deletion does not seem to disturb olfaction, but induces strong impairments in maternal behavior (Grabrucker et al., 2021), rather suggesting an impact on central social circuits of maternal care.
In conclusion, evidence suggests that some social interaction impairments in ASD models can be influenced by atypical olfactory processing, and impaired parental care.Although their etiology is different, comparison between mouse models that display impairments only in olfaction or parental behavior would allow to discriminate the impact of each function on social interaction skills.

Discussion and further directions
Here, we highlight research on well-characterized olfactory and parental behaviors described mostly in mice and rats.Nonetheless, it should also be kept in mind that those findings come from these domesticated species, housed under laboratory conditions.These conditions can differ dramatically from the natural ecological, physical, social, and sensory environment, affecting sensory function and behavioral display.Laboratory models show also limitations due to low genetic diversity and limited phenotypic variability.Strong differences in phenotypes of parental behavior exist between mouse strains, and inbred vs. outbred animals (Kopachev et al., 2023).Furthermore, other rodents like the monogamous prairie voles (Microtus ochrogaster) present enormous differences in parental care in males and females, making voles an interesting model to test the effect of ASD-associated gene deletion.
The use of standardized and robust tests to assess responses to social vs. neutral odors or methods to address which specific olfactory process is affected (e.g., sensory detection vs. central processing) will help to clarify the exact contribution of olfaction on the social core traits of ASD mouse models.Furthermore, the complexity of the odor environment, such as odors presented in an enriched and naturalistic environment rather than one single odor, may also help to discern olfactory phenotypes.Indeed, Cntnap2 KO, Fmr1 KO and Shank3 heterozygous mice were unable to discriminate a novel odor in such complex environment, even though their general olfaction impairments were relatively mild (Kuruppath et al., 2023;Li et al., 2023;Ryndych et al., 2023); Table 1).A similar phenotype has been observed in people with autism carrying a mutated variant of the TSHZ1 gene, who display reduced odor sensitivity and discrimination, but not odor identification (Ragancokova et al., 2014).However, a solid characterization of olfactory processing, parental behavior, their interplay and social interaction impairments still needs to be completed in most mouse models of ASD.Particularly, the use of conditional KO in the olfactory epithelia and/or forebrain neurons (e.g., OMP vs. CaMKIIα -Cre mouse lines) would decipher the main effect of each ASD-associated gene on olfaction, parental behavior and social interaction impairments and their interplay.This would allow to draw new hypothesis about atypical olfaction and therapeutic options using social cues in individuals diagnosed with autism (Barros and Soares, 2020;Boudjarane et al., 2017).Finally, four olfactory receptor genes have been associated with ASD.These receptors are expressed in the olfactory tissues, as well as in the brain (Annamneedi et al., 2023).Deletion of these olfactory receptor genes (among those with orthologs) in rodents may also contribute to our understanding of the interplay between olfaction, parenting and ASD-like phenotypes. table.