Neural cell adhesion molecule Negr1 deficiency in mouse results in structural brain endophenotypes and behavioral deviations related to psychiatric disorders

Neuronal growth regulator 1 (NEGR1) belongs to the immunoglobulin (IgLON) superfamily of cell adhesion molecules involved in cortical layering. Recent functional and genomic studies implicate the role of NEGR1 in a wide spectrum of psychiatric disorders, such as major depression, schizophrenia and autism. Here, we investigated the impact of Negr1 deficiency on brain morphology, neuronal properties and social behavior of mice. In situ hybridization shows Negr1 expression in the brain nuclei which are central modulators of cortical-subcortical connectivity such as the island of Calleja and the reticular nucleus of thalamus. Brain morphological analysis revealed neuroanatomical abnormalities in Negr1−/− mice, including enlargement of ventricles and decrease in the volume of the whole brain, corpus callosum, globus pallidus and hippocampus. Furthermore, decreased number of parvalbumin-positive inhibitory interneurons was evident in Negr1−/− hippocampi. Behaviorally, Negr1−/− mice displayed hyperactivity in social interactions and impairments in social hierarchy. Finally, Negr1 deficiency resulted in disrupted neurite sprouting during neuritogenesis. Our results provide evidence that NEGR1 is required for balancing the ratio of excitatory/inhibitory neurons and proper formation of brain structures, which is prerequisite for adaptive behavioral profiles. Therefore, Negr1−/− mice have a high potential to provide new insights into the neural mechanisms of neuropsychiatric disorders.


Results
Altered brain anatomy in Negr1 −/− mice. First, in situ hybridisation was carried out to label the Negr1 expression in adult mouse brain. Negr1 is expressed extensively in the forebrain and cerebellum. Strong expression was observed in all cortical layers in different areas (somatomotor, somatosensory, parietal association area, visual area, retrosplenial area), in the limbic system (hippocampus:DG, CA1-3 subfields), entorhinal cortex, subiculum, amygdala, hypothalamus, islands of Calleja, olfactory bulb, olfactory tubercle, lateral geniculate complex and reticular nuclei of thalamus), globus pallidus, and granular layer of cerebellum along with caudate putamen (Fig. 1a,b). No Negr1 signal was detected in Negr1 −/− brain sections ( Supplementary Fig. S1). For the initial screening of sub-regional organisation in Negr1 −/− mouse brain we carried out an immunostaining detecting neurofilament. There were no obvious changes in the gross anatomy of Negr1 −/− brain except for remarkably enlarged lateral ventricles were observed. Cytoarchitecture and the fibre-tracts are illustrated in Fig. 1c,d.
To study neuroanatomical changes, we analysed the volume of the whole brain and selected brain regions in Negr1 −/− mice compared to their Wt littermates using high resolution MRI. No significant differences in body weight and brain weight of Negr1 −/− mice were observed compared to their Wt littermates ( Table 1). The analysed brain structures, including the whole brain, ventricular system, white matter tracts, and cortical and subcortical grey matter structures were depicted in Table 1. We detected a small but significant (5.7%) reduction in total brain volume in Negr1 −/− mice as compared to Wt controls (Fig. 1e,f,i). Also, Negr1 −/− mice had significantly enlarged lateral ventricles (64.6%), third ventricle (37%) and fourth ventricle (35.7%) (Fig. 1g,h,j,k). Regarding white matter areas, the corpus callosum was found to be significantly reduced (15.7%) (Fig. 1l), while other white matter tracts like the anterior commissure (anterior and posterior), internal capsule, fornix and fimbria remained unchanged. Enlargement of the ventricular system in Negr1 −/− mice might partly reflect the reduction in the volume of some cortical and subcortical areas. We found a significant reduction in the size of the globus pallidus (15.5%) and hippocampus (10%) (Fig. 1m,n). No changes were observed in the frontal cortex, olfactory system (olfactory bulb, olfactory tubercle, lateral olfactory tract and rhinocele), striatum, hypothalamus, medulla, midbrain, brain stem and cerebellum volume ( Table 1). Impact of Negr1 on hippocampal neuronal population. Since Negr1 is expressed in hippocampus and we have observed reduced hippocampal volume in Negr1 −/− mice, we examined whether Negr1 deficiency  sections (a,b). OB olfactory bulb, Pir piriform cortex, OT olfactory tubercle, iSI island of Calleja, Th thalamus, Amy amygdala, Sub subiculum, DG dentate gyrus, IGL inter geniculate lateral geniculate complex, CP caudate putamen, MO somatomotor cortex, SS somatosensory cortex, PTLP posterior parietal association area, VIS visual area, Epd endopiriform cortex, HA hypothalamus, Cb cerebellum, RsPD retrosplenial cortex, CLA claustrum, RT reticular nucleus of thalamus, LGD/V dorsal/ventral part of lateral geniculate complex, MG medial geniculate nucleus. Gross cytoarchitecture shown by neurofilament immunostaining in Wt (c), enlarged ventricles in Negr1 −/− mice brain (d), LV lateral ventricles. A 3D reconstruction of the Wt (e) and Negr1 −/− (f) mice brains and ventricles (green: lateral ventricles, yellow: third ventricle and purple: fourth ventricle) in Wt (g) and Negr1 −/− (h) brains. Ventricle enlargement is observed in the Negr1 −/− mice. Quantitative analysis of effects of Negr1 deletion in mice on the volume of total brain (i), lateral ventricles (j), third ventricles (k), corpus callosum (l), globus pallidus (m) and in hippocampus (n). Data represent mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, Mann-Whitney U test (Wilcoxon rank sum test).
affect the whole neuronal population in hippocampus or some specific neuronal subtypes. We found no significant difference in number of NeuN (pan-neuronal marker) positive neurons (total hippocampal region, p = 0.13; dentate gyrus, DG = 0.40; Cornu Ammonis, CA = 0.16) (Fig. 2a,b,e,f,k). However, the number of parvalbumin (PV) positive interneurons was found to be significantly reduced (p < 0.0001), in the CA (p < 0.0001) and in DG (p < 0.02) regions of the hippocampus (Fig. 2c,d,i,j,l,).
Lack of barbering behavior and selective deficits in social interaction in Negr1 −/− mice. After weaning, no whisker trimming or barbering behavior was observed in either genotype (the mice were group-housed by genotypes). At 8-9 weeks of age, most Wt mice were completely devoid of whiskers and had trimmed facial hair. In contrast to Wt littermates, all Negr1 −/− mice had full sets of whiskers and facial hair (χ 2 = 143.72, p < 0.00001; Fig. 3a). Similar differences in barbering behavior were evident at 20-21 weeks of age (χ 2 = 145.02, p < 0.00001; Fig. 3b). Notably, 100% of Negr1 −/− mice had a full set of whiskers and intact facial hair.
In our previous study, Negr1 −/− mice displayed impaired sociability and social dominance 40 . Here, we examined social interaction between two freely moving male mice (12-14 weeks old) of the same genotype as this test is considered to be more sensitive for studying social interactions in adult mice. During the 10 min direct social interaction test, aggressive behavior, anogenital sniffing, sniffing of other body parts, active contact, passive contact, rearings, digging and self-grooming were assessed (Fig. 3c-i). No attacks or aggressive behavior were registered during the interactions in either genotype. Mann-Whitney U test revealed that Negr1 −/− mice spent substantially less time in sniffing of other body parts than genitals (W = 82, p = 0.014) and in active contacts (W = 78, p = 0.035) and tended to have a shorter total social contact bout duration (W = 73, p = 0.08; Fig. 3j-l, p-r,v-x), whereas a higher number of bouts during total social contacts (W = 20.5, p = 0.027) was registered for Negr1 −/− mice as compared to Wt mice (Fig. 3l). Total social contact is a summarized measure reflecting the sum total of anogenital sniffing, sniffing of other body parts, passive contacts and active contacts ( Table 2). As for non-social activities, Negr1 −/− mice spent more time on rearings (W = 18, p = 0.014) and had a longer rearing bout duration (W = 8, p = 0.0007); also, Negr1 −/− mice tended to spend more time digging (W = 25, p = 0.06) and had a larger number of digging bouts (W = 13, p = 0.0057) (Fig. 3m,n,s,t,y,z). In contrast, Negr1 −/− mice spent less time on self-grooming (W = 81, p = 0.019) and had less self-grooming bouts (W = 80.5, p = 0.02) (Fig. 3o,u, z'). Assessment of marble burying and tail-suspension tests revealed no differences between Negr1 −/− and Wt mice ( Table 2).
Despite the fact that that Negr1 −/− mice showed no aggression, their social and non-social activities during the direct social interaction (DSI) test were nevertheless altered. We thus hypothesized that genotype-dependent www.nature.com/scientificreports www.nature.com/scientificreports/ differences in aggressive behavior could become evident in a social stress situation. We performed the resident-intruder (RI) test to assess aggressive behavior and social interaction. Neither Wt nor Negr1 −/− mice showed significant aggressive behavior, a surprising effect, possibly caused by the low propensity of the mice from the mixed (129S5/SvEvBrd × C57BL/6) genetic background to be aggressive. In the open field (OF) test the number of rearing bouts was reduced in Negr1 −/− mice (p = 0.049) (Fig. 4e). Total distance travelled was similar in both groups, whereas distance travelled in the centre was increased in Negr1 −/− mice compared with Wt mice (p = 0.0043) (Fig. 4f,g).
Deficits in social behavior correlate with changes in the brain structure of Negr1 −/− mice. Next, we investigated the correlations between the regional volumetric changes in the brain and social interaction measures. Here, we also included sociability measurements from 3-chamber test which has been described in our previous study 40 . Interestingly, most of the significant correlations were found in Negr1 −/− mice: reduced total brain volume was negatively correlated with sniffing (bouts) and active contact (bouts and time), and positively correlated with passive and total social contact (length) ( Table 3). Reduced hippocampal volume was positively correlated with passive contact (length) and the reduction in the volume of globus pallidus was negatively correlated with active contact (length) and positively with passive and total contact (length). Thinner corpus callosum was negatively correlated with sniffing (bouts), active contact (bouts and length) and 3-chamber sociability test. The size of the 3 rd ventricles was positively correlated with sniffing (bouts and length), active contact (bouts and length) and 3-chamber sociability test (Table 3). In Wt mice, non-social activity during direct social interactions showed positive correlations with the size of different brain regions, such as digging bouts with the volume of total brain and hippocampus. Rearing and self-grooming were positively correlated with the size of lateral and 3 rd ventricles (Supplementary Table S1). After pooling both genotypes, self-grooming was positively correlated with the total brain volume and hippocampal size, whereas self-grooming and rearings were negatively correlated with the volume of lateral ventricles (Supplementary Table S2).

Negr1 deficiency leads to impaired neuritogenesis in hippocampal neurons. To examine whether
Negr1 is involved in the initiation stage of the formation of neurites and their outgrowth, we prepared dissociated hippocampal neuronal culture from postnatal (P) day P0-1 of Negr1 −/− mice and the corresponding Wt mice. The cytoskeleton of neurite sprouting was examined at 6 hrs post plating with immunolabelling and scanning electron microscopic imaging. F-actin binding compound phalloidin was used to label growing actin filament aggregates www.nature.com/scientificreports www.nature.com/scientificreports/ through the spherical neuronal cells at the neurite growth initiation site 42 , and the neuronal marker MAP2 that labels microtubules of the neurons was used to distinguish neurons from glia. We observed that spherical Wt neurons started to develop lamellopodia with few filopodia protrusions ( Fig. 5a-d). In contrast, the Negr1 −/− neurons possess large F-actin rich protrutions that began to aggregate with diffused lamellopodia and a higher number of filopodia that develop faster compared to control neurons ( Fig. 5e-h). Quantification of F-actin intensity revealed a significant (p < 0.001) increase in neurite initiation sites in Negr1 −/− hippocampal neurons (Fig. 5i). Similar topographical features of profound filopodia protrusions on the surface and accelerated neurite sprouting in Negr1 −/− neurons were also visualised by scanning electron microscopic images ( Fig. 5j-m). Tracing of neuronal development was also done by transfecting the hippocampal culture at DIV2 with plasmid expressing RFP only in neurons under the synapsin promoter. Morphometric analysis of neurite outgrowth and branching in transfected neurons was examined 24 hrs after transfection (at DIV3). Our analysis showed that Negr1 deficiency led to a significant (p < 0.0001) increase in neurite number, neurite length and branching at DIV3 (  Table S3).

Discussion
The present study expands the phenotyping of Negr1 −/− mice and sheds light on the relationship between observed neuroanatomical and behavioral patterns. We show that Negr1 −/− mice possess neuroanatomical and behavioral endophenotypes which are related to the core diagnostic domains of several psychiatric disorders like SCZ, ASD and ADHD. NEGR1 has been implicated in normal brain development and susceptibility to a wide spectrum of psychiatric disorders and in AD pathology in humans. Therefore, it is important to ask the question of how the alterations of NEGR1 expression may underlie the neuropathology of psychiatric disorders. www.nature.com/scientificreports www.nature.com/scientificreports/ Negr1 expression was observed in cortical-subcortical brain areas that are known to be important for cognitive, affective and motivational behavior. Importantly, we observed intensive expression of Negr1 in thalamic reticular nucleus which is the functional hub for information flow in thalamo-cortical circuits. High Negr1 expression was also detected in the islands of Calleja which modulate dopamine signalling between PFC and temporal lobe, in the ventral striatopallidal system. These nuclei play a significant role in maintaining normal connectivity of brain and their alterations are related to the pathophysiology of psychiatric disorders 43,44 .
MRI-based volumetric analysis revealed neuroanatomical abnormalities in Negr1 −/− mice, including an enlargement of ventricles, and a decrease in the volume of the total brain, hippocampus, globus pallidus, and corpus callosum. These anomalies corroborate reports of animal models of several psychiatric disorders summarised in Supplementary Table S4. Smaller total brain and hippocampal volume in Negr1 −/− mice is in line with our previous study showing impaired learning and sociability in Negr1 −/− mice 40 . The hippocampus is essential for the formation of memory, spatial navigation, learning, emotional and social behavior through its widespread connections with the PFC, amygdala, thalamus, hypothalamus and basal ganglia 45 .
Reduced total brain and hippocampal volume has been observed in several animal models, such as a SCZ model 46 , a MDD model 47 , ASD model 48 , and a post-traumatic stress disorder (PTSD) model 49 . Smaller globus pallidus in Negr1 −/− mice is indicative of disrupted cortico-basal ganglia circuitry and/or connections of the limbic pallidum with the dopaminergic system 50 . The globus pallidus and striatum are components of the basal ganglia that make connections with the PFC and thalamus. In addition, they are involved in the reward prediction, memory, attention and movement planning 51,52 . Several studies indicate that the globus pallidus is also related to the pathophysiology of depressive disorders 53 , Tourette syndrome, obsessive compulsive disorder, ADHD and accompanying neuropsychiatric symptoms associated with PD and Huntington's disease 50 . A reduction in globus pallidus size has also been observed in ITGβ3 model of ASD 48 . These results corroborate that the reductions in the volume of brain areas present in Negr1 −/− mice are similar to brain endophenotypes of several neuropsychiatric disorders.
Reduced corpus callosum volume observed in Negr1 −/− mice is in line with a study 23 which showed that specific variants of the NEGR1 gene were linked to lower white matter integrity of the corpus callosum and fornix. A similar phenotype has also been found in several mouse models of SCZ such as MAP6 KO 63 , DISC1 tr 64 , Zic2 kd/+ 59 , in an ASD mouse model ITGβ3 48 and in an animal model of MDD 65,66 . Corpus callosum is the largest white matter tract mediating information flow between two cerebral hemispheres via excitatory and inhibitory neurotransmission 67 . The hypothesis of imbalance in the ratio between excitation (E) and inhibition (I), called the E/I balance due to reduced corpus callosum volume has been suspected in psychiatric disorders like autism, SCZ    www.nature.com/scientificreports www.nature.com/scientificreports/ and other overlapping phenotypes of mental disorders 68 . Structural changes in corpus callosum have also been linked with faulty hemispheric connectivity and are associated with impaired sensory motor, social, emotional and cognitive functions 69,70 . Furthermore, Negr1 −/− mice were found to exhibit increased susceptibility to pentylenetetrazol (PTZ)-induced seizures which may reflect E/I imbalance 40 . Therefore, we hypothesize that social impairments in Negr1 −/− mice might be influenced by E/I imbalance due to reduced corpus callosum volume.
Immunohistochemical analysis of hippocampus revealed significant reduction in PV positive interneurons with unchanged number of NeuN positive nuclei in Negr1 −/− mice as compared to Wt mice. Recent study showed significant decrease in adult hippocampal neurogenesis in the Negr1 −/− mice with no change in NeuN positive neurons and hippocampal neurites 71 . Our previous study describes the impaired entorhinal axonal growth and abnormal entorhinal fibre projections in the hippocampus of Negr1 −/− mice 40 . Defective neuronal migration in the somatosensory cortex and decrease in spine density were also reported in Negr1 −/− mice 41 . Taken together we suggest that Negr1 deficiency in Negr1 −/− mice does not affect the total number of neurons rather it is interchanging some specific subtypes of neurons or newly born neurons.
Decrease in GABAergic signalling is among the most likely pathophysiological mechanisms causing psychiatric disorders like SCZ, ASD, MDD, stress and anxiety [72][73][74][75][76] . Increasing evidence suggest that a decrease in the activity of parvalbumin-expressing inhibitory interneurons is due to the reduced excitability of neurons 75,77 . A recent www.nature.com/scientificreports www.nature.com/scientificreports/ study showed decreased long term potentiation (LTP) and miniature excitatory postsynaptic currents (EPSCs) in the hippocampus of Negr1 −/− mice 71 . Profound alterations in the distribution of functional neurotransmitter receptors in the hippocampus has been also shown in Negr1 −/− mice 40 . Therefore, the reduced number of PV positive interneurons could influence overall hippocampal activity by decreasing inhibitory signals, possibly leading to hyperactivated hippocampus. Alternatively, it could be a compensatory mechanism for reductions of excitatory tone due to reduced corpus callosum volume. Altogether, we suggest that Negr1 deficiency does not directly influence the number of neurons, comparatively it plays an important role in neuronal migration, axonal projection and circuit formation. Additionally, Negr1 is also essential for balancing the E/I ratio for proper synaptic transmission.
Abnormal social behavior is a common feature of several psychiatric disorders. Decline of social interactions and social withdrawal that is one of the negative symptoms of SCZ. One important feature of Negr1 −/− mice observed during this study was the lack of barbering behavior or whisker trimming, a behavior commonly seen in the Wt mice in our animal facility, which reflects cooperative social activity and cognition and is evident in several group-housed mouse strains 78,79 . Whisker trimming or barbering, associated with dominance, is commonly observed in the C57BL/6 strain as an index of social hierarchy 80,81 . Lack of whisker trimming has also been described as an indication of inability to establish social hierarchy and impaired social cognition 82 . Interestingly, a drastic decrease in barbering behavior was also present in mice with the deletional mutation of another IgLON family gene, Lsamp 83 . From the previous study, we know that the social stimuli are less attractive for Negr1 −/− mice. In the 3-chamber test, Wt mice clearly preferred the presence of a conspecific, compared to an empty room, whereas in their Negr1 −/− littermates no such effect was seen 40 . In the current study, we evaluated the aspects of social interaction of Negr1 −/− mice in more detail. We show that the lack of preference of social stimuli in the 3-chamber test could be due to deficits in social memory/recognition, as Negr1 −/− mice spent equal amount of time in all three chambers, showing no preference for the rooms with either familiar or unknown conspecific ( Supplementary Fig. S5). However, in our current test, the common preference of rodents to investigate a novel conspecific more than a familiar one was only evident as a tendency in Wt mice, therefore the possible deficits in social recognition in Negr1 −/− mice need to be clarified in future studies.
During the DSI test, the total time spent in social contact with an unknown partner was not different in Negr1 −/− mice compared to their Wt littermates. However, Negr1 −/− mice made more approaches towards their partner compared with Wt mice whereas the duration of each interaction bout was relatively shorter, indicating disoriented and confused social behavior of Negr1 −/− mice. We have described similar behavioral pattern of Negr1 −/− mice in the tube dominance test 40 where the winning time of Negr1 −/− mice was significantly shorter. All together, these results indicate that Negr1 −/− mice display hyperactivity in the social contacts engaging only in brief superficial bouts of social contact. Moreover, Negr1 −/− mice show reduced self-grooming, a measure of repetitive behavior that are highly stereotyped patterns of sequential movements 84 . The reduced self-grooming in Negr1 −/− mice could also be caused by increased digging and rearing time, exhibiting hyperactivity and behavioral perseverations. Decreased self-grooming has been also observed in mice lacking D1A dopamine receptors 85 , and in 16p11 +/− mice which is associated with ASD and other neurodevelopmental disorders 86 . We also performed the marble burying test, reflecting obsessive-compulsive behavior 87 . Negr1 −/− mice showed no difference in marble burying as compared with controls despite having a higher number of digging bouts in the DSI test. This indicates that in Negr1 −/− mice digging in the DSI test does not reflect repetitive obsessive-compulsive behavior, but is rather an effort to escape or find a shelter 88 .
Negr1 −/− mice performed less supported rearings in the RI and OF tests, but more supported rearings in the DSI test, showing that this behavioral parameter is heavily dependent on the test type. Similar to the DSI test, neither Negr1 −/− nor Wt mice displayed aggressive or attacking behavior in the RI test. It is possible that the absence of aggression in these animals is influenced by the mixed (129S5/SvEvBrd × C57BL/6) genetic background of the mice. In fact, reduced aggression level probably enables us to observe more subtle aspects of social behavior 89 .
In the present study, we draw correlations between brain structure and behavioral parameters. MRI imaging was done on the same set of mice used for the behavioral testing. Volumetric alterations in the ventricles, hippocampus, globus pallidus, and corpus callosum in Negr1 −/− mice were most significantly correlated with activities like sniffing bouts of other body parts, and the duration of active and passive contact during social interactions. Although the obtained correlative data need to be taken with caution since the number of mice used for the MRI experiment was limited, our findings provide initial evidence that the rate of alterations in the brain structures could be correlated with the behavioral changes present in the Negr1 −/− mice.
To validate the putative essential role of Negr1 in the brain development by regulating neuronal outgrowth and synapse formation, we investigated the early neurite sproutings in the developing hippocampal neurons derived from Negr1 −/− mice. Compared to later stages of neuronal development and function (dendritic and axonal development), the role of Negr1 in the early steps of neuritogenesis has not been well described. Proteins regulating actin, such as F-actin, mark the initial sprouting of the neurites at the neurite initiation site 90 . Here, we show for the first time defective F-actin accumulation and filopodia at neurite initiation stage of neuritogenesis in Negr1 −/− hippocampal neurons. Enhanced neurite outgrowth was also recorded at DIV3, indicating disruption in the initial dynamics of cytoskeleton formation during neuritogenesis. Previous reports show that downregulation of NEGR1 through siRNA approach impairs/lowers neuronal arborisation 39,91 . The possible explanation for this contrasting result could be the difference in time points and a different source of neuronal cells. Despite the incongruence in the results it is evident that Negr1 is a central factor regulating neuronal morphology at different developmental stages during neurite formation. IgLONs are known to form dimers through homophilic and heterophilic interactions and control neuronal growth and synaptogenesis 38,[91][92][93][94] . For example, it has been shown that Ntm mediates bifunctional effects on neurite outgrowth through attractive and repulsive mechanisms, which are cell type specific 95 . The combined effect of Lsamp and Ntm regulates neuritogenesis through complementary interaction, which was independent of their cell-cell adhesion functions 96 . Neuritogenesis involves multiple www.nature.com/scientificreports www.nature.com/scientificreports/ interactions between the developing neurites and the extracellular matrix. Constructural changes during neuritogenesis were related to abnormal neural circuit development in ASD and SCZ 97,98 . Taken together, our results indicate that Negr1 regulates the structural molecules of neurite initiation stage during neuritogenesis even before any connections with other neighbouring neurons are made.
In conclusion, the present study demonstrates the importance of Negr1 at brain structure and function level. We show that Negr1 deficient mice exhibit behavioral alterations and morphological abnormalities in the brain that are similar to the ones observed in psychiatric disorders such as SCZ, ASD and ADHD. The link between neuroanatomical and behavioral findings is significant for understanding the neuronal development and structural changes underlying neuronal connectivity problems related to the etiology of psychiatric disorders. Further research is required to elaborate the structural and functional connectivity of the neural network in this mouse model. This knockout mouse line can be a useful model to elucidate the key molecular targets for the development of new therapeutic strategies in neuropsychiatric research. In situ hybridisation and neurofilament immunostaining. Negr1 (650 bp) transcripts were cloned from a cDNA pool of a C57BL/6 mouse brain and inserted into pBluescript KS vector (Stratagene, La Jolla, CA) to create an in situ probe. RNA in situ hybridization on sagittal and coronal free-floating PFA-fixed 40 μm mouse brain cryosections using digoxigenin-UTP (Roche) labelled Negr1 antisense RNA probes was performed as described previously 100 . Neurofilament immunohistochemistry was carried out with mouse anti-2H3 antibody, (1:100; Developmental Studies Hybridoma Bank) following the peroxidase method as described previously 96 . Images were captured using inverted light microscope (Olympus BX61 microscope) equipped with Olympus DX70 CCD camera (Olympus, Hamburg, Germany).

Methods
Magnetic resonance imaging. At 7 months of age, mice were deeply anesthetized and perfused with 0.1 M PBS followed by 4% paraformaldehyde (4 °C). Brains were left in the skulls to preserve the anatomy and incubated in 4% PFA at 4 °C and then in PBS until 2 days prior to imaging. Skulls were then placed in 2 mM gadovist in PBS and incubated at 4 °C with rocking until imaging. A T2 RARE sequence was used for imaging using a 94/20 Bruker BioSpec small animal MRI with the following parameters: Tr, 900 ms; TE, 47.13 ms; imaging matrix, 512 × 360 × 80; spatial resolution, 0.0444 × 0.03 × 0.2 mm for an imaging time of approximately 3 h and 4 min. Volumes were segmented manually by an observer blinded to the genotype using ITK-SNAP (V3.2.0). The entire brain and the ventricles were manually delineated 101  Behavioral testing. Behavioral testing was performed between 9:00 A.M. to 17:00 P.M. under 30 lux light intensity. Behavioral testing started when the mice were 12-14 weeks old and the same mice were repeatedly used in the behavioral tests. The testing order was as follows: three -chamber test, direct social interaction, marble burying and tail-suspension. Resident intruder test and Open field test was performed with different set of mice. Before each experiment, mice were let to habituate to the experimental room and the lighting conditions therein for 1 h. The mice were allowed to rest 1-2 weeks between the tests. Barbering behavior. Barbering  www.nature.com/scientificreports www.nature.com/scientificreports/ Direct social interaction test. The social interaction test was carried out as described previously 83 , briefly: 10 pairs of two unfamiliar mice of same genotype were matched according to the body weight and their behavior was video-recorded for 10 min. The videos were later scored by a trained observer. Three measures (time spent in s, the number of bouts, and average bout duration) were registered for each mouse for the following parameters: (1) sniffing the body of the other mouse, (2) anogenital sniffing of the other mouse, (3) passive contact, (4) rearing, (5) digging, (6) aggressive attack, and (7)self-grooming. Parameters 1, 2 and 3 were also summarized to get an additional parameter of "total social contact time" for each animal. Altogether, 40 mice were tested (Wt: n = 20; Negr1 −/− : n = 20). Tail suspension test. Mice were suspended for 6 min from the edge of a shelf 58 cm above a table top by adhesive tape, placed approximately 1 cm from the tip of the tail. The duration of immobility was scored during the last 4 min from the recorded videos by an observer blind to the genotype. Mice were considered immobile only when they hung passively and completely motionless for at least 3 seconds. A total of 32 mice were tested (Wt: n = 16; Negr1 −/− : n = 16).

Resident-intruder test.
Previously group-housed males were separated and housed individually for 1-2 months before testing. A group-housed mouse of the same age and same genotype was used as an intruder mouse. The resident animal was placed from its home cage into a separate small cage and left alone for 30 minutes. After 30 minutes an intruder was introduced into the same cage. The test was stopped immediately after the first attack (an attack being defined as a bite) and lasted up to 5 min if no attack occurred. The interactions between the two animals were videotaped for 5 minutes from above and later scored for further analysis. The number of animals engaged in aggressive and non-aggressive social behavior (sniffing, anogenital sniffing rearings, digging and self-grooming) was recorded allowing the comparison of three parameters (time spent in s, the number of bouts, and average bout duration). Altogether, 40 mice were tested (Wt: n = 20; Negr1 −/− : n = 20).
Primary hippocampal culture and assessment of early neuritogenesis. Assessments of early neuritogenesis were performed as reported earlier 96 . To study the neurite initiation stage, 6 hr post culturing neurons were prepared for scanning electron microscopy; immunostaining and quantification of F-actin were done. For neurite outgrowth and branching analysis, pAAV-hSyn-RFP transfected primary hippocampal neurons were imaged on DIV3 and neurite number, neurite length and branches were analysed as reported earlier 96 . Statistical analysis. Data were analysed using Statistica V12 (Statsoft Inc., Oklahoma, USA) and graphs were plotted using Prism 5 (GraphPad, La Jolla, CA, USA). Differences in continuous variables were evaluated by unpaired t-test, one-way ANOVA or repeated measures ANOVA followed by Mann-Whitney U test as a post hoc test (Wilcoxon Rank Sum test) or two-way ANOVA followed by Newman-Keuls post hoc test. Chi-square one-sample analysis was used to analyse the results of whisker trimming. As the behavioral scores were not normally distributed, Spearman's rank-order method was used for the calculation of correlation coefficients. The differences were considered to be significant if the p-values were less than 0.05. All results are displayed as mean ± SEM (standard error of mean).
Ethical approval. All animal procedures in this study were performed in accordance with the European Communities Directive (2010/63/EU) and permit (No. 29, April 28, 2014) from the Estonian National Board of Animal Experiments. In addition, the use of mice was conducted in accordance to the regulations and guidelines approved by the Laboratory Animal Centre at the Institute of Biomedicine and Translational Medicine.
This article does not contain any studies with human participants or human samples.

Data Availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.