Specific profile of ultrasonic communication in a mouse model of neurodevelopmental disorders

Mice emit ultrasonic vocalizations (USVs) in different social conditions: pups maternal separation, juveniles play, adults mating and social investigation. The USVs measurement has become an important instrument for behavioural phenotyping in neurodevelopmental disorders (NDDs). Recently, we have demonstrated that the deletion of the NFκB1 gene, which encodes the p50 NF-κB subunit, causes NDDs phenotype in mice. In this study, we investigated the ultrasonic communication and the effects of an early social enrichment in mice lacking the NF-κB p50 subunit (p50 KO). In particular, USVs of wild-type (WT), p50 KO and KO exposed to early social enrichment (KO enriched) were recorded using an ultrasound sensitive microphone and analysed by Avisoft software. USVs analysis showed that p50 KO pups emit more and longer vocalizations compared to WT pups. On the contrary, in adulthood, p50 KO mice emit less USVs than WT mice. We also found significant qualitative differences in p50 KO mice USVs compared to WT mice; the changes specifically involved two USVs categories. Early social enrichment had no effect on USVs number, duration and type in p50 KO mice. Together, these data revealed social communication alterations in a mouse model of NDDs; these deficits were not recovered by early social enrichment, strengthening the fact that genetic background prevails on environmental enrichment.

Several studies have shown that environmental enrichment (EE) can revert and/or prevent pathological conditions linked to genetic alterations and environmental insults 30 . Given that the mother represents the main component of the environment of a mouse pup, enhancing maternal care constitutes the earliest source of EE. This stimulation of mother-pup interactions can be obtained by housing pups with an additional female from birth until weaning 31 . This early enrichment has been reported to have long-term beneficial effects on brain and behaviour in WT mice and more importantly in murine models of NDDs 31,32 .
Based on this context, we analysed ultrasonic communication in another mouse model of NDDs, p50 KO mice that present a deletion of the NFκB1 gene, coding for the NF-κB p50 subunit. p50 is a member of the NF-κB family. In response to numerous stimuli, NF-κB is rapidly activated and plays different roles in many cellular processes including regulation of neuronal survival and plasticity, inflammation, cell proliferation, neurogenesis, apoptosis, oncogenesis, learning and memory [33][34][35][36][37][38][39] . After six months of age p50 KO mice have a reduced body weight in comparison with WT and display a particular behaviour characterized by decreased anxiety-like responses, increased exploratory activity and reduced tendency to establish dominant-subordinate relationships among cage mates 40 . They also present altered hippocampal neurogenesis linked to a cognitive deficit in spatial short-term memory 41 . Furthermore, we have previously demonstrated that p50 KO mice have cortical structure abnormalities and social behaviour impairment typical of NDDs animal models 42,43 . Nonetheless, it is still unknown whether p50 KO mice have quantitative and/or qualitative impairments in communication. Hence, in this study we investigated the ultrasonic communication and evaluated the effects of early social enrichment in p50 KO mice on USV emission both in infancy and adulthood. Ultrasonic communication analyses were associated in infancy with measures of neurodevelopment (e.g. eyes opening) and of social interest (homing test), in adolescence with evaluation of sensory abilities (olfactory habituation/dishabituation) and at adulthood with measures of locomotion (open field) and social interaction (male-female social interaction).

Results
Reduced maternal care in p50 KO mice. Maternal care, such as nursing postures, non nursing postures, pups grooming and nest building, were analysed during the first postnatal week in WT, KO and KO enriched mice (Fig. 1). Maternal behaviour was different between genotypes [F (2, 14) = 15.48, p < 0.001; Fig. 1A]. KO pups received significantly lower nursing postures in comparison with WT pups. On the contrary, KO enriched pups got higher nursing postures than KO pups and similar to those of WT pups. In addition, early social enrichment increased non nursing postures [F (6, 42) = 15.71, p < 0.0001]. Furthermore, it was interesting to assess the individual contribution of the two females to the behaviours measured in social enrichment condition. For this reason, maternal activities of KO mothers and WT females were observed and analysed ( Supplementary  Fig. S1A). KO mothers displayed more nursing postures and less non nursing postures than WT females [F (3, 24) = 12.71, p < 0.0001]. These data highlight the fact that KO enriched females expressed more nursing postures than when they were alone [F (3, 27) = 2.99, p < 0.05; Supplementary Fig. S1B]. So early social enrichment increased maternal care, but the body weight of pups did not change. Indeed, body weight of KO enriched pups was similar to those of KO pups ( Supplementary Fig. S2).
These data consolidate improved maternal care also for p50 KO mice. As reported previously, this could influence neurodevelopment of pups. To verify this, we analysed ultrasonic communication and some neurodevelopment markers in KO and KO enriched compared to WT mice.
Eyes opening delay in p50 KO and p50 KO enriched pups. Eyes opening, a typical developmental parameter, was evaluated and significant differences between genotypes were found ( Fig. 1B)  On PND 14 social recognition was evaluated in the homing test. No significant genotype effects were found on the latency to reach the area covered by nest sawdust, on locomotor activity ( Supplementary Fig. S3) and on the time spent in all different areas (Fig. 1C). Interestingly, pups of all genotypes spent more time in the nest area than in the clean area [F (2, 182) = 19.42, p < 0.0001; Fig. 1C and Supplementary Fig. S3]. WT and KO enriched pups preferred to spend time in the nest area respect to all other areas; instead KO pups spent similar time in the nest and start area. This suggests a partial deficit of social recognition in KO pups, recovered by early social enrichment.
To investigate if results obtained in the homing test depend on any olfactory deficit, olfactory habituation/ dishabituation test was performed on adolescent mice (Fig. 1D). There were not significant differences in time spent sniffing odours between genotypes. Data obtained reveal that KO and KO enriched mice have not olfactory deficit.

Altered ultrasonic communication in p50 KO and p50 KO enriched pups. Quantitative analysis of
USVs. For the first time, ultrasonic communication was analysed in this model. WT, KO and KO enriched pups were evaluated on PND 4, 6, 8 and 10 for USVs in response to maternal separation. WT pups emitted a number of USVs that increased from PND 4 to PND 6, when it reached its maximum and then gradually decreased. This pattern was exacerbated in KO. USVs number of KO enriched pups increased from PND 4 to PND 8 and then decreased ( Fig. 2A).
Regarding the number of USVs, KO emitted significantly more calls than WT pups [F (2, 92) = 48.19 p < 0.0001; Fig. 2A]. Duration of USVs of KO was significantly longer that those of WT pups [F (2, 92) = 16.68 p < 0.0001; Fig. 2B]. Early social enrichment did not influence ultrasonic communication pattern; indeed USVs of KO enriched were similar to those of KO pups. Concerning other communication traits, such as peak amplitude www.nature.com/scientificreports www.nature.com/scientificreports/ max and mean, peak frequency max and mean of USVs, no significant differences were found between genotypes ( Fig. 2C-F).
These results show altered ultrasonic communication of KO pups, both at basal condition and with early social enrichment. In addition, a more detailed analysis was performed to investigate if there were differences Qualitative analysis of USVs. Considering that highest number of USVs was observed at PND 6 both in WT and KO pups, spectrograms of pups on this day were manually analysed and USVs were classified into 10 categories as previously reported by Scattoni 14 . In Supplementary Fig. S7, some examples of different USVs emitted by WT pups are reported. Also KO and KO enriched pups had the same pattern of classification.
As shown in Fig. 3A  To analyse social behaviour, male-female social interaction test was done. As shown in Fig. 4B, KO and KO enriched mice had social interaction deficit. Indeed, KO spent significantly less time in social activities and more time in non social activities than WT mice [F (2, 18) = 4.935, p < 0.05]. Early social enrichment did not ameliorate social deficit. In particular, a more detailed analysis of social and non social behaviours displayed a lower time spent doing sniffing behaviour for KO and KO enriched mice in comparison with WT mice [F behaviour (4, 72) = 114.6, p < 0.0001; Fig. 4C]. In addition, KO enriched spent higher time doing exploring than WT mice [F behaviour (1,18) = 209, p < 0.0001; Fig. 4C].
During this test, USVs emitted by adult male interacting with oestrous female were recorded. Mice not vocalizing were excluded (2 WT, 1 KO and 1 KO enriched). KO mice emitted a lower number of USVs than WT [F (2, 13) = 9.832, p < 0.01; Fig. 5A]. USVs number of KO enriched was similar to those of KO mice. No significant differences in other vocalization traits were found (Fig. 5B-F). On the contrary, a qualitative analysis of ultrasonic communication revealed important differences between genotypes. In particular KO adult mice emitted a lower number of two syllable and frequency steps calls [F (18, 117) = 3.183, p < 0.0001; Fig. 6A] with a shorter duration of this last category than WT mice [F (18, 117) = 3.264, p < 0.0001; Fig. 6B]. KO enriched mice had an ultrasonic communication pattern similar to those of KO mice. Finally, Fig. 6C displays the vocal repertoire of mice represented as the proportions of different call typologies for genotype.

Discussion
p50 KO is an inbred strain of mice that, as we previously demonstrated, presents cortical structural alterations, hyperactivity and social interaction deficit typical of NDDs animal models 42,43 . Now we analysed ultrasonic communication in p50 KO mice, both in pups and adults, and compared it with USVs of WT mice. Concerning maternal separation-induced USVs, we found that WT pups emitted an inverted U-shaped call emission pattern that followed a typical ontogenetic profile of USVs mouse pups 12,44 and it was exacerbated in p50 KO pups. p50 KO pups emitted significantly more and longer USVs than WT pups. A detailed analysis displayed that WT pups emitted a quite homogenous repertoire of calls; on the other hand p50 KO pups emitted a different repertoire of www.nature.com/scientificreports www.nature.com/scientificreports/ calls, which included higher number of two syllable and frequency steps with a longer duration of these calls than WT pups.
Different studies have suggested the translational value of pups USVs, that can be compared to the human babies cry in the context of NDDs and ASD. This is an extensively studied theme and Scattoni, for first, proposed this analogy for the acoustic and functional features of USVs; they can reflect the same functions of babies cry such as their capacity to elicit parents attention and care and for their important role in social communication between mother and infant 14 . Other studies confirmed the idea that pups USVs can present a social communication form similar to the nonverbal infants calls 13,29 . Understanding USVs role in mice pups is not only necessary to understand the mother-pups interactions, but also to better know communication in children and this can have implications with the knowledge and early identification of NDDs, in particular of ASD 45 .
To further support the translational value of murine USVs study, a similar analysis of pups-mother communication has been applied also in humans, studying the crying of newborn babies 46 . Abnormal features of baby crying emerged to be strictly related to specific neurological disorders 47 , including also autism spectrum disorders 48,49 .
After the USVs analysis in pups, we focused our attention on USVs of adult mice. During male-female social interaction test, adult KO mice emitted decreased number of USVs than WT mice associated with a reduced social interaction, in particular reduced sniffing behaviours. We hypothesize that this reduced number of USVs found in adult p50 KO mice can be linked to reduced communication and social interaction typical of adult human patients with NDDs. But what is very interesting, is that the same categories of calls altered in KO pups were reduced in KO adult mice. Therefore, alterations of ultrasonic communication found in p50 KO mice are not generalized but call-specific; indeed only few types of calls are different in comparison to those of WT mice. This could have an important role in the context of specific call meaning. Therefore, in p50 KO mice quantitative and qualitative alterations of ultrasonic communication were found. In other models of NDDs, similar altered calling patterns were detected. For instance, BTBR pups emitted more calls with a longer duration than their control mice 14 and during adulthood, number of USVs of BTBR mice decreased 15 , as p50 KO mice. From a qualitative point of view, alterations of USVs were different but always call-typologies specific. Indeed, BTBR pups emitted an unusual pattern with a high number of harmonics, two syllable and composite calls 14 . In adulthood, during male-female social interaction test, BTBR emitted fewer short and frequency steps calls than control adult mice 15 . Furthermore, other call-typologies specific deficits are present in the literature, such as increased number of frequency jump calls on PND 7 50 , and decreased number of downward calls on PND 8 in Fmr1-KO compared to their WT pups 18 . Also at adulthood, during courtship Fmr1-KO mice emitted a higher proportion of upward syllables 19 , complex, chevron and flat, and a reduced number of composite and frequency steps than control mice 51 . In addition, as ultrasonic communication specific deficit, in another model of NDDs, Shank1 −/− mice, a categorical shift in the proportions of two clusters of frequency (first cluster between 50 and 80 kHz and second cluster between 80 and 100 kHz) was observed 22,52 . Interestingly, heterozygotes mice for 16p11.2 deletion displayed call-types specific alterations. These adult mice uttered calls in a different proportion for the three-phase male-female social interaction test. In particular, a higher percentage of short calls was found during first exposure to novel oestrus female (phase 1) and a lower percentage of upward calls after the female was removed from the interaction cage (phase 2). Also the number of calls in each category was different resulting in fewer complex, two syllable, upward and frequency steps in comparison with control mice in phase 1 and two syllable and upward in phase 2 28 . Unfortunately, to date the meaning of mice USVs various typologies is unknown. In the future it will be compelling to understand the meaning, also because different mouse models of NDDs seem to have their own ultrasonic communication scheme.
We also wanted to investigate if ultrasonic communication of p50 KO mice could be influenced by maternal behaviour. One possible explanation of communication altered in p50 KO mice was linked to reduced maternal care observed in KO mothers. Indeed, KO mothers had a reduced frequency of maternal behaviours, such as nursing postures, in comparison with WT mothers.
In line with this hypothesis, the study of Kikusui and Hiroi suggested that poor maternal care can be considered a "self-generated environmental factor" of diseases like ASD, as it is induced, through atypical vocal sequences, by a genetic ASD risk carrier 53 . For instance, heterozygosity of Tbx1, an ASD risk gene, caused atypical pups USVs (less complex call types and less variable call sequences), which evoked less maternal behaviour 54 . Therefore, mutation of ASD risk genes can alter the neonatal call sequence, which renders pup's social communication with mothers ineffective and maternal care less efficient 29,53,54 .
In this context, early social enrichment was performed to better understand the link between USVs and maternal behaviour. In particular, KO pups were exposed to early social enrichment (a WT female flanking their natural KO mother) until weaning and their USVs were recorded and analysed. First, we observed that KO enriched pups received more maternal care than standard KO pups, similar to those of WT pups. In addition, we analysed the individual contribution of the two females that took care of the pups to understand if increased nursing behaviour derived only by added WT female, or by changes in KO mother behaviour. We found that early social enrichment increased nursing behaviour for KO mothers. We evaluated also the effect of early social enrichment on WT pups, by adding a virgin WT female to the WT mother. We found that WT enriched pups received maternal care similar to WT pups.
Concerning developmental traits of pups, early social enrichment did not influence body weight of pups. Indeed, body weight of KO enriched did not significantly change compared to those of KO pups. In addition, some marks of delayed development were found in KO pups and only partially rescued in KO enriched pups. For example, KO pups did not display a significant preference between nest area and start area in the homing test contrary to WT and KO enriched pups. This may suggest a less developed olfactory discrimination for KO pups at the age of testing partially rescued by early social enrichment. Furthermore, KO pups had a delayed eyes opening compared to WT; on the contrary, KO enriched pups presented a heterogeneous phenotype, since most of them opened their eyes as at day 13 (as KO pups), but almost 40% opened eyes at day 12 (as WT pups). To date, no studies have reported that early social enrichment allows a recovery of these delayed development signals. It is known that enhanced social environment can regulate gene expression through epigenetic modulations but if Different groups of mice were used in this study. First group of 17 adult primiparous mothers (4-6 month old) was evaluated for mother-pup interaction. Second group: offspring of other mothers kept behaviourally naïve, were tested for behavioural tests during infancy, adolescence and adulthood.
Animals were housed in standard cages in a 12 hours light/dark cycle (light phase: from 8:00 a.m. to 8:00 p.m.) with food and water available ad libitum. Temperature (22 °C) and humidity (50% ± 10) in the cage were automatically regulated by the Sealsafe Aero System by individually ventilated cages with EPA filters (Tecniplast Group, Italy). Mice were housed up to a maximum of 4 adults per cage or 1 mother with its litter or 1 female with 1 male (for mating). To reduce genetic variations, we used for all the experiments mice bred and tested in the same time lapse (4 months).
All experiments were performed in conformity with the European Communities Council Directive of 1986 (86/609/EEC), and approved by the Italian Ministry of Health, Animal care and use Committee of the University of Brescia.
Mother-pup interaction. Maternal behaviour was analysed for 17 litters derived by 6 WT, 6 KO and 5 KO enriched mothers of first group. From PND 1 to 7, mothers were observed twice day for 1 h (at 11:00 a.m. and at 4:00 p.m.) using a method described by Oddi and colleagues 32 . Maternal behaviours, such as nursing postures, non nursing postures, pups grooming and nest building, were analysed and expressed as frequencies 57 and 18 male) and 20 KO enriched (13 female and 7 male) pups of second group were assessed for social recognition in the homing test. After separation from the mother, each pup was transferred in a Plexiglas T-maze with three arms (10 × 8 cm), surrounded by a wall 15 cm high and 0.5 cm thick. Two arms of the apparatus named "start" and "clean" were covered by clean sawdust and another arm called "nest" with sawdust collected from the nest of the pup's cage. The pup placed in the start arm was free to explore the maze for 5 min. Images captured by a camera above the maze, were analysed manually by the operator, measuring the first latency to reach the nest, the time spent in the nest, in clean and in start arms and locomotor activity by square crossing.
Olfactory habituation/dishabituation. For the test of olfactory habituation/dishabituation adapted from 59 , 20 WT (10 female and 10 male), 20 KO (10 female and 10 male) and 17 KO enriched (10 female and 7 male) adolescents (6 weeks old) of second group were used. A sequential presentation of different odours was performed in three consecutive trials of 2 min duration each. Olfactory investigation such as nasal contact with the applicator (within a 2 cm distance) was evaluated (Supplementary Methods).
Spectrograms of each pup at PND 6 were inspected manually by the operator and ultrasonic vocalizations were classified into 10 categories, generally based on criteria previously described by Scattoni and colleagues 14 . Open field exploration test. Male WT, KO and KO enriched adult of the second group (4-6 month old, 6-7 mice for genotype) were introduced in the centre of the arena (40 × 40 cm in Plexiglas) and left free to explore for 5 minutes. A camera vertically montated 1.5 m above the arena recorded mice movements. Locomotor activity,

Male-female social interaction and adult ultrasonic vocalizations.
For male-female social interaction, 7 couples of WT, KO and KO enriched adult mice of the second group were used. Male were isolated in their cages for 5 days and then an unfamiliar female of the same genotype was introduced in the cage of the isolated male for 5 minutes. On the day of test, the vaginal oestrous phase of females was checked as previously described 15,60 . Only females in oestrous phase were tested. Social behaviours (e.g. sniffing, following, mounting, contact and wrestling) and non social behaviours (e.g. self grooming and exploring) were recorded with a camera placed facing the side of the cage and analysed with Observer XT (version 14.1, Noldus, The Netherlands).
An ultrasound microphone suspended 20 cm above the cage, recorded USVs emitted by the male resident during 5 minutes interaction with oestrous female. Male-female social interaction induced USVs were quantitative analysed and manually classified as previously described for pups USVs.
Statistical analysis. Statistical analysis was performed by GraphPad Prism 6 software (GraphPad, San Diego, California). For mother-pup interaction, homing test, olfactory habituation/dishabituation test, quantitative and qualitative USVs analysis of pups, qualitative USVs analysis of adults and detailed evaluation of social/ non social behaviors, Two-way analysis of variance (ANOVA) followed by Tukey or Sidak's multiple comparison test were used. For other behavioural tests, One-way ANOVA followed by Tukey or Sidak or Dunnet's multiple comparison test was used. Data are presented as the means ± S.E.M., with the statistical significance level set at p < 0.05 (WT versus KO mice and KO versus KO enriched).