Tight connections between presynaptic bouton and postsynaptic density by cell adhesion molecules (CAMs) are critical to form stable and functional synapses [1–6]. Two types of CAMs, Ca²⁺-dependent and Ca²⁺-independent, work as synaptic adhesion molecules in the central nervous system (CNS). Cadherin and selectin are dependent on Ca²⁺ ion for binding, while members of the immunoglobulin superfamily (IgSF) contribute to establishment and remodeling of neuronal synapses [7–9] in a Ca²⁺-independent manner. Among IgSF, immunoglobulin LON (IgLON) family proteins function as synaptic adhesion molecules that play important roles in axonal extension, dendritic arborization, and synaptogenesis during brain development [10–13]. Molecules of the IgLON family including LAMP, OBCAM, neurotrimin, GP55, and AvGP50 have three Ig-like domains and localize to the membrane raft of the Triton-insoluble low-density fraction via a glycosylphosphatidylinositol (GPI)-anchor [14–18]. Among the IgLONs, neuronal growth regulator 1 (Negr1), also named Kilon (a kindred of IgLON), is the most recently characterized IgLON subgroup member [19, 20].

Negr1 is a GPI-anchored membrane protein of 46 kDa with three immunoglobulin-like domains and six putative glycosylation sites (Fig. 1A) [19, 20]. Initial studies during mouse development show that Negr1 expression gradually increases during post-natal brain development and reaches a constant level in adulthood [20–22]. Both immunohistochemistry and western blot analysis confirm that the Negr1 protein is expressed in the cerebral cortex and hippocampus of the adult rat [19]. In situ hybridization studies revealed that Negr1 mRNA is expressed in the cerebral cortex, hippocampus, amygdala, hypothalamus, and olfactory bulb [23]. Moreover, electron microscopic analysis showed specific localization of Negr1 at postsynaptic regions of dendritic synapses in the cerebral cortex and hippocampus of the adult brain [22]. In another study, Negr1 expression was detected in reactive astrocytes, which contribute to neurite outgrowth of hippocampal neurons [21]. These data suggest that Negr1 has distinct functions depending on its cellular and subcellular localization in the brain.

The role of Negr1 in synaptogenesis and neurite outgrowth has been reported mainly in in vitro studies. Comparing cultured cortical neurons from wild type and Negr1 knockout (negr1 ^−/−) mice, negr1 ^−/− cortical neurons showed significantly reduced neurite numbers, length, and branches [23]. It was suggested that Negr1 is shed from the cortical neuronal membrane by metalloproteinase ADAM10 [10], and the soluble form of Negr1 can promote synaptogenesis and neurite outgrowth [24]. Pischedda and Piccoli [24] showed that membrane-released soluble Negr1 promoted neurite outgrowth by modulating ERK1/2 activation and the fibroblast growth factor receptor 2 (FGFR2) signaling pathway as an underlying mechanism. In cultured hippocampal neurons, Negr1 was detected mainly in the presynaptic axon terminals in the early culture stage (DIV 3–14); however, at late culture stage (DIV21–28), it was detected mainly in the dendritic postsynaptic spine of mature neurons [25]. This suggests that Negr1 expression and subcellular localization are differentially regulated depending on developmental stage of hippocampal neurons. Moreover, heterologous overexpression of negr1 in hippocampal neurons decreased dendritic synapse number at early culture stages (DIV 3–14), whereas it increased dendritic synapses in late culture stages (DIV 21–28) [25]. Together, these studies argue that Negr1 differentially modulates neural outgrowth depending on developmental stage [25]. The morphology of neuronal dendritic processes is also affected by Negr1 expression [26]. When Negr1 expression was knocked-down by miRNA in cultured cortical neurons at DIV 1, it altered the distribution of neurites in each branching order at a late culture stage (DIV 16) but not at an early culture stage (DIV 6) [26]. Taken together, these results indicate that Negr1 is an important regulator of not only synapse number, but also of maturation of dendritic processes; moreover, its regulatory effects are dependent on the developing stages of neurons. Direct control by Negr1 of synapse number and neurite outgrowth was also demonstrated in an in utero study [27]. When negr1 expression was silenced in mouse embryonic brain at E15.5 by in utero electroporation, dendritic length and number of neurite processes of layer II/III cortical neurons at P7 decreased significantly compared with their controls [27]. These results confirm that Negr1 deficiency prevents normal morphological development of pyramidal neurons in vivo.

Since the cloning of negr1, the function of this putative synaptic adhesion molecule has been investigated mainly in synaptogenesis and neurite outgrowth. Therefore, it was unexpected when human genetics studies showed Negr1 as a major risk factor for human obesity by genome-wide association studies (GWAS) [28–31]. These studies have been replicated in human subjects of various ethnic groups including Europeans [32–37], Asians [38–40], and African Americans [41, 42], suggesting that common genetic variants of negr1 increase the risk of obesity in diverse ethnic backgrounds. Moreover, one animal study demonstrated that increased fat mass, enlarged adipose cells, and decreased muscle mass were observed in negr1 -deficient mice [43], suggesting that Negr1 may serve as a potential drug target that could be exploited for treatment of human obesity. However, there also were some studies that failed to pinpoint Negr1 as an obesity-associated risk factor [44–46]. An obesity risk locus that was identified in obese adults was not related to weight gain of overweight children [44]. Moreover, SNPs of negr1 identified in a European population were not related to and had no significant effects in a Chinese population [45, 46]. Thus, although most reports indicate Negr1 as an obesity risk factor, further precise investigation is still needed.

Currently, the molecular mechanisms regarding how Negr1 serve as a risk factor for human obesity are not clear. Although an in vitro study showed a putative role of Negr1 in lipid metabolism [43], studies using animal models failed to show any mechanism but rather showed conflicting data. For example, one study showed significantly reduced body mass, food intake, and physical activity in negr1 ^−/− mice while energy expenditure remained unchanged [47]. In contrast, another study that targeted Negr1 expression in a specific brain region showed an opposite phenotype; blocking Negr1 expression in the periventricular hypothalamus area in vivo led to increase in body weight and food intake and reduced locomotion activity [48]. Therefore, it can be speculated that Negr1 expression in the hypothalamus may play a different role in body mass increase compared with that in other brain/body regions. Therefore, the brain region-specific effects of Negr1 polymorphisms on the pathogenesis of obesity need to be addressed.

Dennis et al. [49] recently reported that Negr1 has effects on brain structure independent of its effects on obesity. By comparing white matter integrity and SNPs of Negr1 in healthy young adults, they concluded that negr1 was closely associated with low white matter integrity. Considering that precise temporal and spatial expression of CAMs is crucial for development, construction, and maintenance of functional neural connectivity [5, 50, 51], it is anticipated that Negr1 may impact the pathogenesis of brain disorders in addition to obesity [52–54], and recent studies suggest that Negr1 is implicated in major depressive disorder (MDD).

Maccarrone et. al. [55] screened for a disease-specific protein bio-signature in the cerebrospinal fluid (CSF) of MDD patients and found a significant elevation of Negr1 in these patients compared to healthy individuals. The involvement of Negr1 in MDD was also supported by a study screening for putative genetic traits associated with treatment response of selective serotonin reuptake inhibitors (SSRIs) [56] in MDD patients. In this recent cross-trait meta-analysis study, negr1 was uncovered as a key genetic locus associated with SSRI responsiveness in MDD patients and was implicated in both obesity and SSRI treatment response [56]. This implies that the genetic variant of obesity is closely linked to SSRI treatment efficacy in MDD. In line with this, aberrant expression of synaptic adhesion molecules was proposed as an etiologic factor of depressive disorder [57, 58], and a strong association between SNPs of the limbic system-associated membrane protein (LSAMP) gene and human MDD has been reported [59]. These prior studies suggest that negr1 may also be involved in psychiatric disorders such as MDD.

To understand the relationship between Negr1 and psychiatric disorders, we used negr1 ^−/− mice to test whether Negr1 influences affective behaviors such as anxiety and depression [60]. We found that negr1 ^−/− mice exhibit anxiety- and depressive-like behaviors caused by impairment of synaptic transmission of granule cells and neurogenesis in the hippocampal dentate gyrus (DG) [60]. Interestingly, we found that expression of Lipocalin-2 (Lcn2), a 24-kD secretory cytokine [61, 62], is severely compromised in negr1 ^−/− mice, and this Lcn2-decrease is responsible for the cellular and behavioral abnormalities observed in negr1 ^−/− mice [60]. Specifically, heterologous Lcn2 expression in the DG region of negr1 ^−/− mice rescued the abnormal electrophysiological properties of granule cells, impaired hippocampal neurogenesis, and anxiety- and depressive-like behaviors in negr1 ^−/− mice [60]. Therefore, our data indicate that the abnormal affective behaviors observed in negr1 ^−/− mice are due to lack of expression of Lcn2, an immune mediator. In our effort to elucidate the molecular mechanisms underlying the functions of Negr1, we discovered that Negr1 directly interacts with leukemia inhibitory factor receptor (LIFR), a co-receptor for LIF [63–65], and thereby potentiates Lcn2 expression [60]. Taken together, our data suggest that Negr1 potentiates LIFR-induced Lcn2 expression and thereby affects hippocampal neurogenesis and affective behaviors (Fig. 1B). In previous reports, reduced hippocampal neurogenesis was observed in Lcn2-deficient mice [66], and these mice showed both anxiety and depressive behaviors [61, 62], which is in line with our findings. Thus far, both cytokines and synaptic adhesion molecules have been shown to play pivotal roles in MDD and anxiety [67, 68]; however, the cross-talk between these two depression/anxiety-mediating pathways has not been reported. Consequently, Negr1’s function via regulation of the expression of immune mediators is unique compared to other synaptic adhesion molecule functions.

In addition to MDD, Negr1 has been implicated in other psychiatric disorders and neurological diseases. Several studies have provided evidence that Negr1 is involved in Alzheimer disease (AD) [69, 70], autism spectrum disorder (ASD) [27], and schizophrenia (SCZ) [71]. A GWAS study using patients with comorbid MDD and AD identified negr1 as a genetic risk factor that affects AD development [70]. These results suggest a possibility that negr1 serves as a common risk factor for both MDD and AD pathologies. Furthermore, downregulation of Negr1 in mice resulted in impaired behaviors that are similar to those of individuals with ASD [27]. In that study, the authors proposed a putative mechanism by which Negr1-deficiency leads to ASD behavior. According to their proposed mechanism, Negr1 directly interacts with FGFR2 and cooperatively regulates cortical development. Thus, Negr1-deficiency led to impaired cortical development resulting in the ASD phenotype. Moreover, a clinical study with schizophrenic patients showed increased expression levels of Negr1 transcript in the dorsolateral prefrontal cortex of these patients [71]. Taken together, these studies strongly support that the altered expression of Negr1 may be associated with abnormal behaviors of various psychiatric disorders (Table 1).

To address the breadth of Negr1 influence in psychiatric disorders, we subjected our negr1 ^−/− mice to a series of cognitive tasks assessing spatial memory, recognition memory, and context fear memory. In the spatial version task of Morris water maze for measuring spatial memory, our negr1 ^−/− mice performed worse than wild-type mice in locating the hidden platform (Fig. 2A) while exhibiting no differences in the swimming speed (Fig. 2B). On the probe trial, the negr1 ^−/− mice swam much farther from the hidden platform location than the wild-type mice, as measured by proximity (Fig. 2C). These results indicate hippocampal dysfunction of negr1 ^−/− mice because the hippocampus is necessarily recruited for performing the hidden platform version of Morris water maze [72]. Meanwhile, in the novel object recognition, the novel object location, and the contextual fear conditioning task, no differences between negr1 ^−/− and wild-type mice were observed, suggesting that Negr1 expression is not critical for recognition memory or contextual fear memory (Fig. 3). It is interesting that, although negr1 ^−/− mice exhibited impaired spatial learning and memory, they showed no significant differences in the recognition task or contextual fear conditioning task (Fig. 3). One plausible explanation for these distinct behavioral phenotypes is that Negr1 may differentially regulate each neural circuit. Behavioral performance in the Morris water maze depends primarily on the hippocampus region in which Negr1 is highly expressed. However, novel object recognition and location tests are affected by both the cerebral cortex and hippocampus [73], and neural circuits between the cerebral cortex and amygdala areas are critical regions for performance in the contextual fear conditioning test [74–76]. Brain regional differences in expression levels of Negr1 might have caused different cognitive behavioral performances of negr1 ^−/− mice (spatial learning vs. fear learning; Table 2). Therefore, future studies using optogenetics or conditional knockout mice dissecting the brain regionspecific roles of Negr1 are necessary to elucidate the roles of Negr1 in other psychiatric disorders.

In this short review, we summarized and discussed recent studies on the role of Negr1 in psychiatric disorders. Accumulating evidence based on both human and animal studies support critical roles of Negr1 in psychiatric disorders such as MDD, schizophrenia, and ASD. However, the molecular and cellular mechanisms underlying Negr1’s role in these psychiatric disorders remain elusive. Considering the diverse brain expression pattern of Negr1 including hippocampus, sensory cortex, and prefrontal cortex region, future studies using brain region/cell type-specific negr1 conditional knockout mice will be instrumental to elucidate the pathophyisological mechanisms of Negr1 in psychiatric disorders. Concerted effort in the fields of genetics, molecular biology, neuroscience, and clinical psychiatric medicine are also needed in the future to dissect the exact pathophysiologic function of this molecule. Although the mechanisms need to be elucidated, studies have proposed Negr1 as a novel target for treatment of psychiatric disorders. Considering the high comorbidity between depression and obesity, Negr1 may act as a central hub bridging the two diseases. If so, Negr1 can be a critical common target for treatment of individuals with MDD comorbid with obesity.

This work was supported by the Samsung Science & Technology Foundation (SSTF-BA1502-13).