Slitrks as emerging candidate genes involved in neuropsychiatric disorders

Abstract

Slitrks are a family of structurally-related transmembrane proteins belonging to the leucine-rich repeat (LRR) superfamily. Six family members exist (Slitrk1–Slitrk6), and all are highly expressed in the central nervous system (CNS). Slitrks have been implicated in mediating basic neuronal processes ranging from neurite outgrowth and dendritic elaboration to neuronal survival. Recent studies in humans and genetic mouse models have led to the identification of Slitrks as candidate genes that may be involved in the development of neuropsychiatric conditions such as obsessive compulsive spectrum disorders and schizophrenia. While these system level approaches have suggested that Slitrks play prominent roles in CNS development, key questions remain regarding the molecular mechanisms through which Slitrks mediate neuronal signaling and connectivity.

SLITRK FAMILY

The Slitrk gene family is composed of six members (Slitrk1 - Slitrk6), which are highly expressed in the central nervous system (CNS) [1,2]. This family was identified in a screen for genes that were differentially expressed in mice with neural tube defects [1]. Slitrk5 had been previously discovered as a gene expressed in early hematopoietic progenitors but not in mature hematopoietic cells [3]. All Slitrks form single pass (type I) transmembrane proteins with an intracellular domain that varies in length [1]. At the extracellular domain, Slitrks contain two leucine-rich repeat (LRR) domains, which are each composed of 13 to 17 LRRs, flanked by cysteine rich domains (Glossary) [1]. Sequence analysis has revealed that the extracellular LRR domains of Slitrks resembled Slit proteins, and a conserved region in their intracellular carboxyl terminus (C-terminus) has a high degree of consensus with the last 16 amino acids of the neurotrophin receptor, tropomyosin-related kinase (Trk) [1]. Based on their similarity with Slits and Trks, these proteins were named Slitrks [1].

Slitrks are classified as members of the LRR superfamily, due to the presence of the LRR domains. The LRR superfamily is a class of molecules containing structural motifs of 20–29 amino acids with a defining sequence LxxLxLxxN/GxL (x being any amino acid) [4,5]. LRR domains are composed of tandem repeats of this LRR motif, and are usually involved in protein-protein interactions [3–5]. LRR-containing proteins are emerging as key regulators of nervous system development and function, playing important roles in processes such as neurite outgrowth, neuronal survival, synapse formation and dendritic morphogenesis (see Table 1 and [5–7]). Genes encoding human Slitrks are present at three different loci on chromosomes 3 (Slitrk3), 13 (Slitrk1, 5 and 6), and X (Slitrk2 and 4) [2]. This chromosomal organization is conserved in the mouse where Slitrks are located on chromosomes 3, 14 and X. All Slitrk genes have a single protein coding exon [2]. As we will discuss in this Review, a recent series of human and mouse genetic studies have identified Slitrks as candidate genes involved in the development of neuropsychiatric conditions. Thus, understanding their various functional roles in the CNS will have important implications for understanding not only normal CNS development, but may also provide valuable insights into mechanisms that go awry during the development of neuropsychiatric disorders.

Table 1.

Summary of LRR-containing proteins and their functions

Protein	Expression	Functions	Refs
AMIGO 1/2	Enriched in CNS	- TM protein that promotes prominent neurite extension in cultured hippocampal neurons	[6,99]
	Amigo 2	- Exhibits homophilic and heterophilic binding activity
	ubiquitous	- Modulates depolarization and NMDA-dependent survival of cerebellar granule neurons
FLRT3	Ubiquitous	- TM protein identified in a screen for genes upregulated after nerve injury	[100]
		- Promotes neurite outgrowth in cultured DRG neurons
Lingo-1	CNS	- TM protein that acts as co-receptor for Nogo receptor	[101–104]
		- Over-expression enhances responsiveness to myelin-associated inhibitors
		- Dominant negative form protects midbrain dopaminergic neurons against degeneration
		- Pathologically upregulated in SN of PD patients
		- Also interacts with EGFR and decreases its protein levels
Linx	DRG neurons	- TM protein that interacts with Trk and Ret receptors and augments neurotrophin and   GDNF/Ret signaling	[31]
		- Knocking out Linx partially photocopies axonal projection defects in mice lacking NGF,   TrkA, or Ret
Lrig1	Ubiquitous;	- TM protein that interacts with ErbB receptor family members (EGFR, ErbB2-4) and   recruits cytoplasmic E3 ubiquitin ligases	[105–108]
	Ubiquitous	- Attenuates GDNF/Ret signaling in neuronal cells through its physical interaction with Ret
Lrig3		- TM protein that interacts with Xenopus FGFR through extracellular domain	[109]
		- Inhibits FGF-dependent Erk phosphorylation
		- Negatively regulates FGF expression levels
LRRTM2	CNS, post- synaptic	- TM post-synaptic partner for α and β Neurexins	[29,32–34]
		- Promotes excitatory synapse formation
		- Interacts with PSD-95 and regulates surface expression of AMPARs
NGL-1	CNS: ST, CT, TCA, dendritic	- TM protein that binds netrin G-1 through LRRs	[110,111]
		- Stimulates outgrowth of embryonic thalamic axons
NGL-2	CNS, HC, CT, post-synaptic	- TM protein that induces presynaptic differentiation and clustering of postsynaptic   proteins	[35,36,110]
		- Increases the number of excitatory synapses
NGL-3	Brain	- TM protein that binds LAR through LRRs	[35]
		- Induces presynaptic differentiation
	CNS	- Promotes excitatory synapse formation
Omgp		- GPI-anchored protein that is expressed by neurons and oligodendrocytes	[112]
		- Mediates growth cone collapse and inhibition of neurite outgrowth
SALM 1–3	CNS	- TM proteins that form homo and heteromeric complexes	[113–116]
		- Promotes neurite outgrowth in young neurons (DIV 4–6) but not in older neurons (DIV 14–   16)
		- SALM1-3 contain a PDZ-binding motif
		- SALM1 induces the dendritic clustering of NMDA receptors
		- SALM2 increases the number of excitatory synapses and dendritic spines
		- SALM3 & 5 induces excitatory and inhibitory pre-synaptic differentiation
Trks	CNS	- TM receptors that bind the neurotrophin family of growth factors	[14,15]
		- Influences the proliferation, differentiation, plasticity and survival of neurons

Abbreviations: Amigo, amphoterin-induced protein; FLRT3, fibronectin leucine rich transmembrane protein 3; Lingo-1, leucine rich repeat and Ig domain containing 1; Linx, immunoglobulin superfamily containing leucine-rich repeat 2; Lrig, leucine-rich repeats and immunoglobulin-like domains; NGL, Netrin G ligand; Omgp, oligodendrocyte myelin glycoprotein; SALM, Synaptic adhesion like molecule; TCA, thalamocortical axons; ST, striatum; CT, Cortex; PD, Parkinson’s disease; DA, dopamine; TM, transmembrane; DRG, dorsal root ganglia; GDNF, glial cell line-derived neurotrophic factor; SN, substantia nigra; TM, transmembrane; HC, hippocampus, AMPAR, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; FGF, fibroblast growth factor; ErbB/EGFR, epidermal growth factor receptor; LAR, leukocyte antigen-related; GPI, glycosylphosphatidylinisotol; DIV, developmental day in vitro; PSD-95, synapse associated protein 95 kDa

SLIT AND TRK PROTEINS

Slits are one of the most well known LRR-containing protein families [8,9], functioning as soluble ligands for Robo receptors [9]. They are involved in axonal guidance and repulsion, tangential neuronal migration, cytoskeletal dynamics, and modifying cell adhesion properties [10,11]. Despite the high degree of sequence homology between Slit and Slitrk LRR domains, recent findings indicate that these protein families are likely to serve divergent functions in the CNS. Firstly, a similar degree of sequence and structural similarity exists in the LRR domains of Slitrk and many other LRR family members. Secondly, Slits do not have any particular homology outside the LRR domains with Slitrks. Thirdly, Slit is a secreted protein, whereas Slitrks are transmembrane proteins with intracellular domains, suggestive of a functional role in regulating intracellular signaling events [1,10]. Lastly, Slitrks do not interact with Robo, suggesting that they likely signal via different receptor-ligand pairings [1]. Although a recent study reported that Slitrk1 can undergo extra-cellular domain shedding (by α-secretase activity) in vivo, it is unknown if the secreted domain plays a functional role [12]. Further information regarding the exact site of α-secretase cleavage within Slitrk1 will elucidate if such a mechanism is conserved amongst other Slitrk family members.

Trk receptors are a family of receptor tyrosine kinases that selectively bind the neurotrophin family of growth factors (eg. brain derived neurotrophic factor, BDNF, and nerve growth factor, NGF, are two neurotrophins from this family) [13,14]. Trk signaling influences the proliferation, differentiation, plasticity and survival of neurons, and has been implicated in higher-order cognitive activities such as learning and memory [13–15]. Neurotrophins are secreted ligands that bind to Trks leading to the phosphorylation and activation of their catalytic tyrosine kinase domains, which rapidly initiate intracellular signaling cascades (Figure 1) [16]. All Trk receptors exhibit high conservation in their intracellular domains, including the catalytic tyrosine kinase domain, the juxtamembrane NPxY motif, as well as the C-terminus domain, in particular, tyrosine 791 (Y791 in human TrkA) [13–15]. Upon phosphorylation, the NPxY motif serves as a binding site for adaptor proteins such as Shc (that initiates Ras and phosphoinositide 3 kinase downstream signaling) [16–18]. Phosphorylation of the tyrosine residue within this particular motif can also serve as signal for receptor endocytosis [19].

Figure 1. Possible models for the functions of Slitrks.

Slitrk5 was selected as the prototype of the Slitrk family and protein domains are shown according to the SMART database (http://smart.embl-heidelberg.de/). a) Synapse formation hypothesis for Slitrk function. According to this hypothesis, Slitrks could localize to the postsynaptic membrane and bind pre-synaptic membrane proteins. The identity of such binding partners remains unknown. This hypothesis is based on studies in culture whereby overexpression of Slitrk2 in fibroblasts induced excitatory synaptic differentiation in contacting neurons [29], as well as immunofluorescence studies demonstrating that Slitrk5 is localized at synaptic sites [28]. Other LRR-containing proteins that are known to play such a function in synapse formation are depicted. Netrin G-ligand (NGL) localizes to the post-synaptic cell, where it binds Netrin G through its LRR domains. LRRTM2 also localizes to the post-synaptic cell, where it binds both α and β Neurexins (NRXN) through its LRR domains. β-NRXN1 is shown as an example. b) Soluble ligand hypothesis. Slitrk1, 5 and 6 have been shown to play trophic roles in neurons [26–28]. Slitrk6 has been demonstrated to modulate neurotrophin signaling [26]. In PC12 cells, Slitrks modulate neurite outgrowth upon NGF treatment [1]. Thus, it is likely that Slitrks modulate RTK signaling. In this model, two possible scenarios are presented: one in which Slitrks act as receptors for soluble ligands that lead to signaling transduction events related to the neurotrophin system (ie. similar to how Trks and their neurotrophins interact). In a second scenario, Slitrks could physically interact with Trk receptors (and/or other RTKs) and directly modulate their function. The depicted Trk amino acid numbers correspond to human TrkA but similar residues and signaling components are found in TrkB and TrkC. Neurotrophin binding to Trks leads to receptor dimerization and subsequent activation and phosphorylation of their catalytic tyrosine kinase domains as well as phosphorylation of intracellular tyrosines such as the juxtamembrane NPXY domain and Y791 (in human TrkA). Y791 recruits PLC-γ. This enzyme subsequently hydrolyses phosphatidyl inositol to generate inositol tris-phosphate (IP3) and diacylglycerol (DAG), which ultimately leads to the release of Ca²⁺ from internal stores and protein kinase C (PKC) activation, respectively [13,16,20]. For simplicity, a single TrkA molecule is shown instead of the active dimmer. Y791 in human TrkA (and analogous tyrosines in TrkB and TrkC) show some degree of homology with the cytoplasmic tail of Slitrks [1], namely Y945 in human Slitrk5, raising the question if Slitrks are also implicated in PLC-γ signaling. Moreover, Slitrk2, 3 and 5 also contain an NPXY motif at the juxtamembrane region. Phosphorylated NPXY serves as a docking site for phosphotyrosine bindig proteins such as Shc that links the activated neurotrophin receptor to two separate intracellular signaling pathways: mitogen activated protein kinase (MAP kinase) and PI-3 kinase [17]. Slitrks might mediate similar signaling events through their conserved NPXY motif. Finally, it was found that Y833 could be phosphorylated in Slitrk5, however its function is still unknown [21]. Abbreviations: TM, transmembrane domain; Ig, immunoglobulin-like domain; LRR, leucine rich repeat; EGF, epidermal growth factor-like domain; LamNT, laminin N-terminal domain; LamG, laminin G domain.

Slitrk2, 3 and 5 share similarities with Trk receptors in that they contain an NPxY motif near their intracellular juxtamembrane region. The stretch of amino acids that spans the NPxY motif in human Slitrk5 shares 25% amino acid identity with the TrkA NPxY region, and higher than 60% with Slitrk2 and 3 suggesting they might also recruit Shc and other adaptor proteins that could initiate intracellular signaling cascades. Furthermore, Slitrk2–6 all contain a common C-terminal region and a conserved tyrosine with significant homology to Y791 in human TrkA [1]. With respect to the TrkA receptor, it is known that phosphorylation of Y791 leads to the recruitment and activation of the γ isoform of phospholipase C (PLC), resulting in the release of Ca²⁺ from internal stores and protein kinase C (PKC) activation [13,15,20]. It is currently unknown if this conserved tyrosine residue is phosphorylated in Slitrks, however, it is tempting to speculate that Slitrks may couple to PLC-γ signaling cascades, based on this domain similarity with the TrkA receptor. Slitrk5 has been shown to be phosphorylated at another intracellular tyrosine residue (Y833) in HeLa cells, but the function and conservation of this site in other Slitrks still needs to be clarified [21].

SLITRK EXPRESSION IN THE CNS

Slitrks are highly and broadly expressed in the mammalian CNS [1,2,22–24] and also in hematopoietic and leukemic cells [25]. Initially it was shown by Northern blot analysis that Slitrk1, 2 and 4 can be detected as early as embryonic day 11.5/12.5 (E11.5/12.5) [1]. Despite some overlapping expression, several key brain regions express different combinations of Slitrks suggesting that each member might play a particular role in the maturation of selective populations of neurons across different stages of development. Slitrk6 is the member that differs mostly from the rest of the family, with expression being restricted to a few areas of the nervous system (such as the thalamus, lateral geniculate nucleus and auditory and vestibular system of the ear) [1,22,26]. Slitrk6 is, however, also expressed in other organs, including the epidermis, limb bud, cochlea and tongue of mouse embryos [23].

Similarly as in the mouse, the expression of Slitrk mRNAs has been demonstrated to be enriched in human brain tissue [3]. In human fetal tissue, Slitrk1–5 were predominantly expressed in the brain, whereas Slitrk6 was also present in the lung and liver [2]. In the adult, Slitrk1 mRNA expression was highest in the frontal lobe and Slitrk5 was expressed in both occipital and frontal lobes [2]. Slitrk2 and 5 were the only members significantly expressed in the spinal cord and medulla [2]. Slitrk6 expression differs greatly from the other members, with no expression in the cortex but high levels of expression in the putamen [2].

A comparative study that analyzed Slitrk1 expression in mouse, rhesus monkey and humans revealed that regional and developmental expression is conserved amongst these species [24]. Slitrk1 was found in projecting neurons of the corticostriatal-thalamocortical circuits (CTC) compartmentalized in striosomes. Subcellularly, it is enriched in the somato-dendritic compartment and cytoplasmic vesicles of cortical pyramidal neurons [24]. In the adult striatum, the expression is confined to cholinergic interneurons [24].

Besides neurons, Slitrks can be found in brain tumors, embryonic stem cells, subsets of endothelial cells, hematopoietic stem cells, as well as leukemic and lymphoma cells [2,25]. Regarding brain tumors, Slitrk5 had the broadest expression amongst the various tumors analyzed, whereas gliomas mainly express Slitrk1–3 [2]. Thus, besides their roles in normal development, Slitrks might also be implicated in malignancy.

SLITRK FUNCTIONS IN THE CNS

Slitrks have been implicated in the modulation of neuronal process outgrowth [1,27,28] promoting neuronal survival [26,28], and in synapse formation [29]. Initial studies performed by over-expressing each of the Slitrk members in PC12 cells revealed that exposure to NGF resulted in a decreased number of neurites per cell in the case of all Slitrk members [1]. Furthermore, the neurite length was decreased in all cases, with the exception of Slitrk1 and 4 [1]. Subsequently, the neurite outgrowth function of Slitrk1 has been analyzed in mouse cortical neurons [12,27]. Cortical neurons cultured from mouse embryos previously electroporated with Slitrk1 cDNA showed increased dendritic length [27], suggesting that this Slitrk member plays an important role in promoting neurite outgrowth. More recently, it has been demonstrated that phosphorylation of Slitrk1 on Ser695 by casein kinase II is critical for the interaction of Slitrk1 with 14-3-3 proteins [12], which are ubiquitously expressed phosphorylation-binding proteins that regulate a number of important cellular processes including cell proliferation, neuronal migration and membrane excitability [12,30]. Mutation of this serine residue to an alanine residue abolished the interaction of Slitrk1 with 14-3-3 proteins as well as blocking the induction of neurite outgrowth in cultured mouse cortical neurons [12].

Recently, a Slitrk5 knockout mouse was generated that displayed reduced total brain volume, which was most prominent in the striatum [28]. In addition, there was a marked decrease in dendritic complexity of medium spiny neurons in the striatum of adult Slitrk5 null mice, as determined by Golgi tracing analyses [28].

Considering the more restricted distribution of Slitrk6 in the brain, with selectively high expression in the inner ear, a recent study performed a detailed analysis of this organ in a mouse model that ubiquitously lacks Slitrk6 expression [26]. Results from this study demonstrated that Slitrk6 promotes the survival and neurite outgrowth of sensory neurons of the inner ear [26]. Histological examinations revealed that vestibular innervation was markedly decreased and sometimes misguided in the Slitrk6 null mice. The mutant mice also showed significant cell death in the spiral and vestibular ganglia. Furthermore, cochlear sensory epithelia taken from these mice and co-cultured with wild-type spiral ganglion neurons were less effective in promoting neurite outgrowth of the spiral gangion neurons compared to co-cultures in which both types of cells were taken from wild-type mice [26]. Evidence for a trophic role of Slitrk6 was further strengthened by the observation that both BDNF and neurotrophin 3 (NT-3) mRNA levels, as well as their corresponding receptor protein levels, TrkB and TrkC respectively, were down-regulated in the inner ear of Slitrk6 null mice. Interestingly, this was observed at an early developmental stage of the cochlea, when cell death had not occurred [26]. This study suggests that Slitrk6 may exert trophic actions, in part, by modulation of the neurotrophin system. Several LRR containing proteins have been shown to interact and modulate receptor tyrosine kinase (RTK) signaling (summarized in Table 2), for example Linx, a leucine-rich repeat and immunoglobulin (LIG) family protein, physically interacts with both Trk and Ret receptors, resulting in increased neurotrophin and GDNF signaling, respectively [31]. Besides Slitrk6, it would be interesting to investigate if other Slitrks could also modulate Trk receptor signaling.

Table 2.

Summary of human genetic studies assessing associations between Slitrks and neuropsychiatric disorders OCD-spectrum disorders

Slitrk gene	Disorder	Main conclusions of the studies		Refs
Slitrk1	GTS	•	Identified 3 patients (out of 174 unrelated individuals with GTS) carrying mutations in the Slitrk1 gene in 174; one patient had the varCDfs variant, while the other two expressed var321. These mutations were absent in more than 3600 control subjects.	[27]
		•	No mutations found in 6 patients belonging to a large Dutch family with GTS	[62]
		•	No association of varCDfs or var321 was reported in a cohort of 812 GTS patients from a Costa Rican and AJ population, however, over- representation of var321 in AJ population was found.	[60]
		•	Five subjects with GTS were identified in the AJ population who contained the var321, two of these did not transmit the mutation to affected child. One individual was found with var321 but did not have GTS.
		•
		•	var321 nor varCDfs were found in 82 Caucasian patients with GTS; a new mutation was found that neither changed amino acid nor altered mRNA splicing.	[58]
		•
		•	Did not find var321 nor varCDfs in 208 affected children, however, found significant association with 3 other novel SNPs in the Slitrk1 gene	[64]
		•
		•	No var321 or varCDfs found in 92 Austrian patients with GTS. A new 3’UTR variant was found in one patient and parents with GTS and was absent in 192 control individuals.	[66]
	GTS, CT	•	Large sequencing study did not find association between Slitrk1 and GTS or CT. Out of 1048 individuals with GTS or CT one individual carried var321; a second carrier of var321 had OCD. Neither transmitted the mutation to GTS offspring. Both of these carriers were of AJ descent.	[61]
	GTS, CT, OCD	•	Provided additional information regarding the ethnicity of the two initially identified variants that are of European descent and found 4 additional affected individuals carrying var321 of Spanish, Eastern Europe and Caucasian. In addition, one AJ unaffected individual with var321 was identified.	[63]
	OCD	•	Did not find var321 nor varCDfs in 322 subjects with OCD	[59]
	TTM	•	Two novel non-synonymous mutations were found in a cohort of 44 patients with TTM and which were absent in 3000 healthy controls. var321 and varCDfs were not found in this population.	[53]
Slitrk2	SCZ	•	Systematic re-sequencing of X-chromosome genes found two novel missense variants in Slitrk2 gene in patients with SCZ and in their affected siblings	[97]

Abbreviations: CT, chronic tics; AJ, Ashkenazi Jew; SCZ, Schizophrenia;

Taken together, the results obtained with Slitrk1 overexpression in neurons [12,27] and the analyses of the Slitrk5 and 6 null mice [26,28], suggest that multiple Slitrk members play key roles in promoting neurite outgrowth. Even though when initially described, Slitrks were shown to play a negative role on neurite outgrowth when over-expressed in PC12 cells [1], it is likely that crucial binding partners involved in neurite outgrowth that are endogenously expressed in neurons are lacking in PC12 cells.

Slitrks have also been implicated in synapse formation. Slitrk5 has been demonstrated by immunocytochemical studies to localize to synaptic sites in cultured neurons [28]. A recent study that performed an expression screen for synaptogenic proteins revealed the abundance of LRR containing proteins capable of inducing synapse formation [29]. Slitrk2 was one of the LRR proteins identified in this screen, and it was shown to induce excitatory pre-synaptic neuronal differentiation in a cellular co-culture system, suggesting that it could be a postsynaptic protein [28]. Another LRR protein that was identified in the screen was LRRTM2 [29], a member of the brain enriched Leucine-rich repeat transmembrane neuronal protein (LRRTM) family [32]. LRRTM2 acts as a ligand for α and β Neurexin I and this binding is crucial for trans-synaptic cell adhesion, mediating excitatory synapse formation [33,34]. Besides LRRTM2, other LRR-containing proteins bind pre-synaptic membrane partners through their LRR domains [7], for example, Netrin G ligand 2 and 3 (NGL-2 and NGL-3) (Table 1) [35–37]. It is possible that Slitrks could interact with a pre-synaptic partner in a similar manner via their extracellular LRR domains and could recruit intracellular post-synaptic proteins (Figure 1). Identifying the intracellular binding partners for Slitrks would contribute significantly to understanding the molecular basis of Slitrk functions in the nervous system.

SLITRKS AND POTENTIAL LINKS TO NEUROPSYCHIATRIC DISORDERS

Considering the significant heritability of many psychiatric disorders, recent efforts have focused on finding susceptibility genes for these disorders [38–40]. Genome-wide association studies (GWAS) have allowed for the screening of known single nucleotide polymorphisms (SNPs) spanning the entire genome in large population samples [41,42]. However, most common variants found so far show low penetrance, thus do not contribute significantly towards disease risk [43]. In contrast to common polymorphisms, the effects of rare variants, whether SNP’s or copy number variants (CNV’s), on disease risk cannot be identified through GWAS approaches. The main reasons for this are twofold: firstly, such rare variants are not well tagged by the common SNP’s tested on whole-genome panels, and secondly, the statistical power of association analysis drops off sharply for rare variants [43]. Rather, the genetic effects of rare variants are generally identified through re-sequencing of candidate genes/loci in large subject and control samples and are often analyzed in aggregate to improve statistical power. In contrast to common polymorphisms, rare variants usually have high penetrance and therefore their detection greatly contributes to understanding the etiology of a disease. Furthermore, they are valuable as biomarkers for preventive strategies. The rare variant hypothesis has been strengthened by the recent findings that de novo mutations in synaptic genes such as Neuroligin 3 and 4 (NLGN3, NLGN4X) and SH3 and Multiple Ankyrin Repeat Domains 3 (SHANK3) are associated with autistic spectrum disorders in a small number of families [44–48]. Genetic studies have recently identified rare mutations in Slitrk genes as being associated with Obsessive-Compulsive Spectrum Disorders and other neuropsychiatric disorders, as discussed in more detail below.

Slitrk1 and OCD-spectrum disorders

Gilles de la Tourette's syndrome (GTS) is a common psychiatric disorder, present in ~1% of children between 5 and 16 years in USA and Europe [49,50]. It is characterized by persistent involuntary vocal and motor tics and it has comorbid occurrence with Obsessive-Compulsive disorder (OCD) and attention deficit hyperactivity disorder (ADHD) [49–53]. Family studies have suggested that genetic factors are involved in the manifestation of GTS, however finding susceptible genes for this disorder has been a daunting task [52,54]. The cortico-striatal thalamocortical circuit (CSTC) has been implicated in the pathogenesis of both GTS and OCD [55–57], however, the molecular and genetic mechanisms underlying these disorders are less clear. In a recent study, a de novo inversion of chromosome 13 was identified in one patient that suffered from GTS and ADHD but had no family history of any OCD-spectrum disorders [27] (Table 2). The 13q33.1 breakpoint mapped approximately 350kb from Slitrk1, and this gene was selected for re-sequencing in a cohort of 174 individuals with GTS. Two additional mutations were discovered in this screen: a single nucleotide deletion that led to a frame-shift and a prematurely truncated protein (varCDf) was found in one individual; and a missense mutation in the 3’ untranslated region (3’UTR) named variant 321 (var321) was detected in two unrelated individuals [27]. In two of these patients, the mothers were also carriers of the mutation and were affected. The truncated Slitrk1 human protein failed to promote dendritic growth when expressed in cultured mouse cortical neurons, in contrast to expression of wild type human Slitrk1 protein [27]. The var321 of Slitrk1 altered binding to the microRNA hsa-miR-189 and it negatively modulated Slitrk1 mRNA expression. These mutations were absent in over 3600 control samples, indicating that they were rare variants [27].

Subsequent studies were unable to find associations between these two reported variants in larger GTS populations [58–62], arguing against a segregation of Slitrk1 and GTS. Furthermore, it was reported that var321 was over-represented in the Ashkenazi Jew population, raising the possibility that population stratification might have led to a false positive result [60,61]. The largest screen for Slitrk1 and GTS performed so far, which involved sequencing 1048 samples from the Tourette Syndrome Association International Consortium for Genetics, did not find association between var321 and GTS [61]. Only two individuals were found who carried var321: one was diagnosed with GTS, OCD and Trichotillomania (TTM), and the other with OCD but not GTS. None transmitted the mutation to their affected offspring [61]. Additional studies provided more information regarding the ethnicity of the initially reported cases, and five additional subjects carrying var321 were reported [63]. Fine mapping of all seven carriers of the Slitrk1 region argued against population stratification confounding the original data [63].

In addition to these two variants, a recent study testing for association of common tag SNP’s spanning SIitrk1 and GTS identified an association to one specific tag SNP (rs959383) as well as two three-marker haplotypes, suggesting there may be a common GTS risk factor of low penetrance in linkage disequilibrium (LD) with the associated marker/haplotypes [64]. Furthermore, novel mutations in the Slitrk1 gene have been found that co-segregate with OCD-spectrum disorders [64–66]. Two non-synonymous mutations in the Slitrk1 extracellular region were discovered in two independent individuals of European descent, in a set of 44 families with TTM, and were absent in a group of almost 3000 healthy controls [65]. TTM is an OCD-spectrum disorder thought to be genetically associated with GTS [67,68]. Moreover, a mutation in the 3’UTR of Slitrk1 gene was found in a group of 92 Austrian patients, that segregated in two additional family members with tic disorders, and which was absent in 192 control subjects [66] (Table 2).

Slitrk1 null mice have been generated and studied with the aim of better clarifying the potential role that Slitrk1 may play in neuropsychiatric disorders [69]. It has been demonstrated that these mutant mice display increased anxiety-like behavior (as measured by the elevated plus maze test; see Box 1) as well as depressive-like behaviors (as assessed in a forced swim task). Neurochemical analysis revealed that Slitrk1 null mice had increased levels of norepinephrine in the prefrontal cortex, striatum, and nucleus accumbens [69] which is consistent with the pathophysiology of GTS since patients with this disorder have been reported to have high concentrations of norepinephrine in their cerebrospinal fluid [70]. Although these mice did not recapitulate the hallmark motor characteristics of human GTS, anxiety and depressive disorders are highly comorbid with GTS [51,53], and administration of clonidine (an α2-adrenergic agonist commonly used to treat GTS patients [71–73]) was able to rescue the anxiety-like behavior of Slitrk1 null mice [69]. These findings add support to a likely role of Slitrk1 in neuropsychiatric disorders. The generation of future mouse models for Slitrk1, especially those which conditionally knock-down the gene in a spatially and/or temporally-specific manner, will help to further delineate the exact roles that this protein plays, and will help to avoid potential developmental compensation from other Slitrk family members that may occur with traditional knockout mouse models. Furthermore, it would be interesting to directly assess the consequences of the introduction of human Slitrk1 variants (eg. var321 or varCDfs) in future knock-in mouse models, and to assess these mice for their subsequent neuroanatomical and behavioral phenotypes. Such studies will help to directly assess the in vivo consequences of specific mutants in the Slitrk1 gene.

Box1. Behavioral tasks commonly used to study endophenotypes of psychiatric disorders in mice

Elevated plus maze

A behavioral test to assess anxiety-like behaviors in rodents. The experimental apparatus consists of a plus-shaped platform with two open and two closed arms that is elevated above the floor. This test is based on the aversion of open space in rodents. Anxiety–like behavior is indicated by a decrease in the proportion of time spent and entries into the open arms.

Open field test

Assesses locomotor activity and anxiety-like behavior in rodents. Subjects are allowed to explore a novel and bright arena; distance travelled as well as percent time spent in the center of the arena is recorded. Generally rodents avoid bright and aversive areas such as the center of the arena; anxious rodents have this behavior accentuated.

Marble burying test

A task that measures repetitive, stereotypical and anxiety-related behaviors in rodents. The experimental setting consists of a novel cage with multiple marbles spaced out regularly on the bedding. The number of marbles buried is measured. Anxious rodents tend to bury more marbles compared to control animals.

Fear conditioning test

Measures associative learning in rodents and it can have two components: contextual and cued-fear learning. A rodent is placed in a novel environment, where an unconditioned stimulus, usually a foot shock, is paired with a conditioned stimulus, usually a tone. Rodents can learn to associate the context/tone with the foot shock, and exhibit freezing behavior when exposed to the same environment (contextual fear) or the tone (cued fear). Contextual fear is hippocampal-dependent, whereas cued-fear is amygdala-dependent.

Forced swim test

A behavioral despair paradigm that reveals depressive-like behavior in rodents. Animals are subjected to trials in which they are forced to swim in a glass cylinder filled with water. The time that the animal spends immobile in the water is recorded. Mice with depressive-like behaviors spend more time immobile.

Tail suspension test

Similar to the forced swimming test, the tail suspension test measures depressive-like behavior in rodents. In this test, a mouse is suspended by its tail, and movements of the mouse are recorded. Mice with depressive-like behaviors spend more time immobile.

Slitrk5 and OCD-spectrum disorders

A recent study of a mouse model in which the coding region of the Slitrk5 gene was replaced by the β-galactosidase gene (lacZ) reporter gene has revealed further evidence in support of an association between Slitrk family members and the development of neuropsychiatric disorders [28]. Slitrk5 null mice initially developed increased anxiety-like behaviors (as assessed by the open field test), and subsequently displayed repetitive and excessive self-grooming behaviors, which led to the formation of severe facial skin lesions and hair loss [28]. Treatment with chronic fluoxetine, a selective serotonin reuptake inhibitor (SSRI) that is one of the major pharmacological agents used to treat depression as well as OCD [74,75], alleviated excessive grooming behavior in the mutant mice. These mice also showed selective over-activation of the orbitofrontal cortex [26]. Activation levels in this brain region relative to other areas of the brain was determined by measuring the protein levels of FosB, a transcription factor that is routinely used to assess neuronal activity in a variety of paradigms [28,76]. Overactivation of the orbitofrontal-subcortical circuits has been observed in functional imaging studies of human subjects with OCD, and is posited to be due to an imbalance in the basal ganglia pathways [77–79]. Thus, the Slitrk5 null mouse recapitulates an important aspect of the human disease.

Slitrk5 null mice also exhibit anatomical deficits in the striatum, such as reduced striatal volume and decreased dendritic complexity, as well as neurophysiological alterations in cortico-striatal transmission, which may have resulted from alterations in glutamate receptor expression levels in the striatum [28]. In this context, it has recently been postulated that striatal dysfunction, in the presence of orbitofrontal cortex overactivation, could lead to deficits in thalamic filtering, or an imbalance in the direct and indirect pathways of the basal ganglia [80,81]. Taken together, the Slitkr5 mouse model recapitulates several key aspects of OCD: (i) increased anxiety-like behavior that is followed by (ii) repetitive behaviors that are ameliorated by administration of a SSRI, as well as (iii) orbitofrontal cortical overactivation, and 4) alterations in basal ganglia circuitry and function.

A question that arises from this study is, given the wide distribution of Slitrk5 throughout the brain [1,22,28], why do these mutant mice reflect a predominantly basal ganglia-dependent phenotype? On the one hand, it is currently unknown if these mice also exhibit higher order cognitive defects, such as impaired learning and memory, since the results of such experiments have not yet been reported. On the other hand, it may be the case that Slitrk5 is the most highly expressed Slitrk family member in the corticostriatal circuitry. Studies comparing the protein expression levels between all members of the Slitrk family have not yet been performed, however, a recent report that compared levels of Slitrk1 protein in different murine brain regions showed very little expression of Slitrk1 in the striatum, as compared with the amygdala, frontal cortex or hippocampus [69]. Therefore, lack of developmental compensation by another Slitrk protein in the striatum of Slitrk5 null mice might be one possible explanation for the observed phenotype involving the basal ganglia circuitry. Considering the particular expression pattern of Slitrk5 within the striatum, together with the behavioral phenotype of these mice, it would be of interest to sequence the Slitrk5 gene within human subjects, to assess whether rare mutations or variants within this gene are associated with OCD and/or OCD-spectrum disorders.

It is striking that many of the behavioral and circuit phenotypes observed in the Slitrk5 null mice have been reported in a previously described genetic mouse model, which lacks the synapse-associated protein 90/ postsynaptic density-95-associated protein 3 (SAPAP3) [82]. SAPAP3 is an intracellular scaffolding molecule that localizes to the post-synaptic density (PSD) of excitatory synapses, and is crucial for maintenance of synaptic structure [83,84]. Similar to Slitrk5 null mice, SAPAP3 null mice displayed excessive grooming that led to the formation of facial lesions, increased anxiety-like behaviors, defects in cortico-striatal transmission and altered expression of glutamate receptors in the striatum [82]. One of the main differences observed between these two models is that the behavioral defects are also observed in the haploinsufficient mouse in the case for Slitrk5, but not in the case of SAPAP3 [26, 73]. It is interesting to speculate whether Slitrk5 and SAPAP3 form part of a common signaling complex that is involved in synaptic maintenance/function in the cortico-striatal circuit. The discovery of Slitrk5 binding partners would help to elucidate these mechanisms. With respect to SAPAP3, recent genetic analyses have led to the identification of a number of rare non-synonymous mutations that have been associated with OCD, GTS and TTM, however larger population analysis will be necessary to understand the precise involvement of SAPAP3 mutations in these disorders [85–87].

In addition to the SAPAP3 null mouse, there are a number of other genetic mouse models which display repetitive behaviors similar to the Slitrk5 null mouse (for recent reviews, see [88,89]). One such model which is of direct relevance to the Slitrk5 null mouse is the Hoxb8 null mouse [90–93] (Table 3). The Hoxb8 gene is a member of the mammalian Homeo-box containing complex (Hox), an important family of transcription factors. Hoxb8 null mice were initially described to develop large skin lesions on their bodies [93]; excessive grooming was later found as the cause of the lesions [92]. This phenotype had 100% penetrance in the offspring of Hoxb8 null mice and it was present in two different genetic backgrounds [92].

Table 3.

Summary of the Slitrk5, SAPAP3 and Hoxb8 genetic mouse models for OCD

Genetic Model	Behavioral Phenotype	Circuit Phenotype	Rescue (of overgrooming behavior)	Refs
Slitrk5−/−		ObF cortical activity		[28]
	Repetitive grooming (head)
		Striatal volume	Fluoxetine (21 days)
	Anxiety-like behaviors
		Corticostriatal transmission
Sapap3−/−			Fluoxetine (6 days)	[82]
	Repetitive grooming (head)	Corticostriatal transmission
	Anxiety-like behaviors		Genetic (overexpression of Sapap3 in ST)
Hoxb8−/−	Repetitive grooming (body)		Bone marrow transplant	[91]
	Repetitive grooming (cagemates)		Subcutaneous lidocaine (acute)	[90]

Abbreviations: ObF, Orbitofrontal cortex; ST, striatum.

In a subsequent study, a Hoxb8 cell lineage mouse model was developed where expression of Hoxb8 could be analyzed by expression of yellow fluorescent protein (YFP) [91]. Hoxb8 mRNA was determined to be broadly expressed in the brain, including in the corticostriatal circuit [92], however, this mouse model unexpectedly revealed that Hoxb8 was not expressed in neurons, but rather in a subpopulation of microglia cells [91]. Consistent with the hypothesis that microglia cells derive from the bone marrow where Hoxb8 is strongly expressed [94,95], a set of bone marrow transplants were performed to evaluate if the over-grooming phenotype could be rescued. Interestingly, transplanting wild type mouse bone marrow into lethally irradiated Hoxb8 mutant mice partially rescued the excessive grooming behavior in a subset of mice (~60%), whereas transplantation of bone marrow from mutant Hoxb8 mice into lethally irradiated wild-type mice led to the development of overgrooming in ~20% of the mice [91]. The excessive grooming phenotype of Hoxb8 mutant mice was reported in a parallel study to be rescued by subcutaneous application of the local anesthetic, lidocaine [90], however, these experiments were not replicated in the subsequent report [91]. It was recently reported that microglia cells derive from primitive (embryonic) macrophages and that in adulthood these cells are maintained independently of definitive haematopoiesis [96]. In light of these conflicting studies, the association of Hoxb8 with OCD-spectrum disorders remains questionable and future studies are necessary to address key questions remaining. For example, it is currently unknown if brain regions that are affected in OCD are anatomically and functionally normal in the Hoxb8 null mice. It will also be of importance to determine if pharmacological treatments (such as SSRIs) can rescue the overgrooming phenotype of the Hoxb8 mutant mouse, as was the case with the behavioral phenotypes exhibited by the Slitrk1, Slitrk5 and SAPAP3 knockout mice [28,69,82]. Furthermore, it would be interesting to investigate whether there is any functional overlap between Hoxb8 and Slitrks. Slitrk5 is highly expressed in immature hematopoietic cells, and it is downregulated upon differentiation [25], as is the case for Hoxb8. Finally, determining if patients with OCD-spectrum disorders carry mutations in Hoxb8 gene will be an important line of investigation in linking this gene to disease processes.

Slitrk2 and psychiatric disorders

Slitrk2 is another member of the Slitrk family that has recently been associated with a neuropsychiatic disorder, namely schizophrenia [97]. A list of synaptic genes located on the X-chromosome was used as candidates for a genetic screen of subjects with schizophrenia [97]. Several new synonymous and non-synonymous mutations were found in genes that have not been previously associated with this disorder, including an amino acid change (V89M) in the second LRR domain of Slitrk2 [97]. Once the function of Slitrk2 is elucidated, it would be interesting to analyze if this point mutation disrupts normal Slitrk2 function.

While preliminary, this study further supports the notion that Slitrks play crucial roles in the CNS and that mutations in these genes can contribute to enhanced susceptibility to neuropsychiatric disorders. As shown by the examples of the Slitrk1, 5 and 6 knockout mice studies [26,28,69], generation of mutant mouse models for Slitrk2 will likely be highly informative for the human genetic studies, in terms of validating the specific biological impact of Slitrk2 on CNS function and in relation to neuropsychiatric disorders.

CONCLUSIONS

Slitrks represent a relatively newly discovered family of structurally-related transmembrane proteins that have a predominant expression in the CNS. Studies so far have shown that Slitrks play trophic roles in the CNS promoting neurite outgrowth and neuronal survival [12,26–28]. Human genetic variants and mouse genetic knockout models have suggested an association of these genes with neuropsychiatric disorders, in particular, OCD-spectrum disorders [27,28,64,65,98]. However, additional large population studies are needed to clarify these findings. Furthermore, the detailed cellular and molecular mechanisms by which Slitrks transmit intracellular signals are largely unknown. An important question to be addressed in future studies will be to define the main functions of the Slitrk family (Box 2). Furthering our understanding of the role of Slitrks in terms of intracellular signaling as well as neuronal connectivity will not only be informative for understanding their roles in normal CNS development but also will provide the foundation for subsequent vertically integrated studies to link Slitrks to the development of neuropsychiatric disorders.

Box2. Outstanding Questions

1. What are the main intra- and extra- cellular binding partners for Slitrks?

   1. Do all Slitrk members function in a similar way or are several scenarios and unique binding partners possible based on divergent intra- and/or extra- cellular amino acid sequences?

   2. Do Slitrks play structural roles at the synapse by binding other pre-synaptic transmembrane proteins, similar to neuroligins and neurexins?

   3. Do Slitrks act as a receptor for an unknown secreted ligand and trigger signaling cascades with their intracellular domains?

   4. Do Slitrks modulate RTK signaling directly by interacting with these membrane receptors or is the modulation indirect?

2. Are there specific mutations in the Slitrk5 gene that contribute to OCD-spectrum disorders in humans?

3. Is Slitrk1 broadly associated with OCD-spectrum disorders and not specifically with GTS as initially described?

4. Are other Slitrk members associated with neuropsychiatric conditions?

5. Besides the brain, are Slitrks playing key roles in the hematopoietic system or cancers such as leukemia?

Glossary

Direct and indirect pathways

two opposing circuit loops in the basal ganglia. The direct pathway originates in the frontal cortex and projects to the striatum. Medium spiny neurons in the striatum, expressing D1 dopamine receptors, project to the globus pallidus interna/substantia nigra pars reticulata (GPi/SNr) complex, which projects to the thalamus, that in turn has reciprocal, excitatory projections to and from the cortical site of origin. The indirect pathway also originates in the frontal cortex and projects to medium spiny neurons expressing D2 dopamine receptors in the striatum, then projects to the globus pallidus externa (GPe), then to the subthalamic nucleus, then back to GPi/SNr, before returning to the thalamus and finally back to frontal cortex. The direct pathway contains two excitatory and two inhibitory connections, resulting in a net positive circuit back to the cortex. The indirect pathway has three inhibitory connections resulting in a net negative feedback loop to the cortex

Genome-wide association study (GWAS)

involves sequencing and comparing the genomes of individuals to search for variations and mutations that associate with a disease of interest

Linkage disequilibrium

in a population, co-occurrence of a specific DNA marker and a disease at a higher frequency than would be predicted by random chance

LRR

Leucine-rich repeat is a protein structural motif that is 20–29 amino acids in length and contains a signature sequence, LxxLxLxxN/GxL, with x being any amino acid. An individual LRR motif has a β strand-turn-α helix structure. Many LRR motifs often assemble into a functional domain called a LRR domain which typically has a horseshoe shape and is amenable for protein-protein interactions

Obsessive-compulsive disorder (OCD)

is a psychiatric disorder characterized by intrusive, persistent thoughts (obsessions) and/or repetitive, intentional behaviors (compulsions) that significantly affect normal life. OCD is usually a chronic anxiety disorder and affects ~2–3% of the population.

Penetrance

the probability that a given genotype manifests itself in a given phenotype. Conditions with high penetrance imply that most individuals carrying the genetic variant in question reflect the associated phenotype. These conditions are rare. Low penetrance conditions are more common and imply that the proportion of individuals with a specific genotype that manifest in an associated phenotype is low

Gilles de la Tourette’s syndrome (GTS)

is a familial OCD-spectrum disorder with childhood-onset characterized by involuntary motor and vocal tics. It affects ~1% of the children between 5 and 16 years and its symptoms usually improve by adulthood. Approximately 75% of individuals with GTS have comorbid psychiatric conditions such as OCD ADHD and TTM

Trichotillomania (TTM)

is an OCD-spectrum chronic disorder characterized by a compulsive urge to pull out one’s hair leading to noticeable hair loss. It can be triggered by stress and has high comorbidity with anxiety disorders as well as GTS