Intercellular adhesion molecule 5 is the first dendrite-specific cell adhesion molecule to be identified and the transmembrane protein discovered to form and maintain filamentous dendritic spines, inhibiting dendritic spine maturation. Unlike other ICAM family members, ICAM5 is exclusively expressed in excitatory neurons of the forebrain, also known as telencephalin (Tian et al., 2007; Gahmberg et al., 2008, 2014; Paetau et al., 2017). ICAM5 possesses nine extracellular Ig domains, making it the largest member of the ICAM subfamily (Yang, 2012). It mediates homophilic binding between neurons (Tian et al., 2000). ICAM5 subcellularly localizes to the postsynaptic neuron soma and dendritic membrane but is not expressed in the axon (Yang, 2012). During the early formation of synapses, the immunoreactivity of ICAM5 gradually increases in filamentous dendritic spines and thin dendritic protrusions, aiding in the formation and maintenance of their filamentous morphology. However, its expression decreases or disappears in mature dendritic spines, promoting spine maturation (Raemaekers et al., 2012; Kelly et al., 2014; Pei et al., 2020). Studies have also shown that ICAM5 loss in neurons increases dendritic spine maturation, while overexpression impedes maturation and increases the dendritic protrusion count (Matsuno et al., 2006; Kelly et al., 2014). Furthermore, research suggests a close correlation between the expression level and functional status of ICAM5 and the transition from dendritic filopodia to mature dendritic spines. This correlation is linked to dendritic spine development, synaptic plasticity, neural circuit formation, and even learning and memory (Yang, 2012; Cheng et al., 2019).

Darnell et al. (2011) identified ICAM5 as a FMRP target gene. Research by Zeng et al. in FXS animal models, specifically Fmr1 gene knockout (KO) mice, validated ICAM5 as an mRNA target of FMRP, highlighting its crucial role in dendritic spine maturation and cognitive dysfunction associated with FXS (Pei et al., 2020). They found an abnormal upregulation of ICAM5 protein expression during critical periods of synaptic development in Fmr1 KO mice, providing insights into the molecular mechanisms underlying dendritic spine maturation impairments in FXS. Pei et al. also observed increased ICAM5 expression in various brain regions of Fmr1 KO mice, including the hippocampus, frontal cortex, and amygdala. Furthermore, they revealed that CLSTN1, another target of FMRP, plays a crucial role in mediating the redistribution of ICAM5 in the postsynaptic membrane (Cheng et al., 2019; Pei et al., 2020) (Figure 3B).

Identifying FMRP target mRNAs is crucial for understanding the pathogenesis of FXS. However, obtaining a complete understanding of the specific responses of FMRP and identifying its targets is challenging. Accurately determining FMRP targets remains a major challenge, and each newly identified target represents a significant step in exploring FMRP functions (Fernández et al., 2013), providing a direction for future research.

Pei et al. (2020) alleviated the behavioral deficits in Fmr1 KO mice through genetic ICAM5 intervention, which may provide therapeutic benefits for the treatment of FXS cognitive impairment and other neurodevelopmental disorders. In the future, this molecule could be used as an important drug target for the treatment of behavioral defects in FXS patients.

Additionally, Tian et al. (2007) validated ICAM5 as a substrate for matrix metalloproteinases-2 (MMP-2) or MMP-9 using various experimental approaches and concluded an important role of MMP-mediated ICAM5 proteolytic cleavage in the regulation of dendritic spine development. Conant et al. (2010) found that N-methyl-D-aspartic acid (NMDA) can stimulate rapid shedding of ICAM5 from cortical neurons in dissociated cell cultures. Such shedding is diminished by the pretreatment of cultures with inhibitors that target MMP-9. MMP-2 and MMP-9 are the most abundantly expressed in the developing brain. MMP-2 is mainly found in astrocytes, while MMP-9 is highly expressed in neuronal cell bodies and dendrites (Ayoub et al., 2005; Tian et al., 2007). The expression and activity of MMP-9 have been shown to depend on NMDA receptor activation and long-term potentiation (Meighan et al., 2006; Nagy et al., 2006; Wright et al., 2007). Growing data also suggest the association of MMP-9 (Meighan et al., 2006; Nagy et al., 2006; Huntley, 2012) with dendritic spine remodeling, synaptic plasticity, learning, and memory formation. A mechanism by which MMPs may rapidly modulate synaptic structure and function is through their ability to cleave specific synaptic cell adhesion molecules (Conant et al., 2010). MMPs cleave ICAM5 in a rapid, neuronal activity-dependent manner. Moreover, MMP-mediated proteolysis is associated with LTP (Conant et al., 2010). To better understand the relationship between ICAM5 and MMP-9, Kelly et al. (2014) also explored ICAM5 expression in MMP-9 null animals. Recent studies have shown that the synaptic translation of MMP-9 is regulated by FMRP (Dziembowska and Wlodarczyk, 2012; Janusz et al., 2013; Gkogkas et al., 2014; Lepeta et al., 2017; Aishworiya et al., 2023). Interestingly, aberrations in dendritic spines that are observed in FXS patients (Rudelli et al., 1985) and Fmr1 KO mice (Comery et al., 1997) have been linked to elevated synaptic levels of MMP-9 (Dziembowska and Wlodarczyk, 2012; Janusz et al., 2013; Lepeta et al., 2017). Clinical trials have reported that minocycline, a broad-spectrum tetracycline antibiotic (Yau et al., 2018), improves cognition and aberrant social behaviors in FXS subjects (Sidhu et al., 2014). In the Fmr1 KO mice, an abnormally elevated expression of MMP-9 in the brain was pharmacologically downregulated after treatment with minocycline (Dziembowska et al., 2013), while genetic removal of MMP-9 rescued the symptoms of FXS (Sidhu et al., 2014). These data suggest that targeting MMP-9, even in late development, may reduce FXS symptoms. However, it remains to be explored whether alleviation of FXS symptoms by genetic removal of MMP-9 is associated with MMP9-mediated ICAM5 elimination. The specific association between the molecular mechanisms of MMP-9 and ICAM5 may reveal new avenues for individualized treatment of neurodevelopmental disorders, especially FXS, in the future.

L1 neural adhesion molecule is a transmembrane protein encoded by the first X-linked gene identified in the IgSF (Djabali et al., 1990). The L1-CAM protein consists of six Ig-like domains, five FNIII-like repeats, a transmembrane, and an intracellular domain (Crossin and Krushel, 2000; Sytnyk et al., 2017; Stoyanova and Lutz, 2022). L1-CAM is expressed in the central nervous system by subpopulations of neurons and on glial cells in the peripheral nervous system (Rathjen and Schachner, 1984; Seilheimer and Schachner, 1987). L1-CAM is involved in adhesion between neurons, the formation of neural fiber bundles, and the growth of nerve processes (Moos et al., 1988; Wang et al., 2012; Congiu et al., 2022; Loers et al., 2023). It mediates homophilic and heterophilic interactions between neurons (Kadmon et al., 1990a; Sytnyk et al., 2017; Stoyanova and Lutz, 2022) and cooperates with NCAM through a mechanism termed assisted homophilic adhesion (Kadmon et al., 1990b). As evidenced in murine model systems, L1-CAM also facilitates axon repulsion and growth cone collapse in response to secreted class 3 Semaphorins (Sema3) (Mohan et al., 2019b; Duncan et al., 2021b), which prune distinct populations of dendritic spines during development and homeostatic scaling (Mohan et al., 2019a,b; Murphy et al., 2023a). Sema3 dimers bind to heterotrimeric receptors comprising L1-CAM, neuropilins (Npn1/2), and plexins A (PlexA1-4), activating intracellular signaling through PlexA Ras-GAP activity that results in spine pruning (Duncan et al., 2021a; Murphy et al., 2023a). Genetic knockouts of L1-CAM in mice lead to increased density of immature dendritic spines on the apical dendrites of cortical pyramidal neurons (Murphy et al., 2023b).

Djabali et al. (1990) first determined that the L1-CAM gene is located in a conserved region of the X chromosome and considered this protein a typical X-linked NCAM. Notably, several genes linked to neuromuscular diseases are also located in this region adjacent to the fragile site connected with intellectual disability (FRAXA), suggesting a possible association between neuromuscular diseases and intellectual disability. Using pulse-field gel electrophoresis, they confirmed the physical connection between L1, and other genes located on Xq28, such as genes encoding eye pigment and glucose-6-phosphate dehydrogenase (G6PD). These locations are consistent with those of the X-linked neuromuscular disease mapping region of the L1 molecule (Djabali et al., 1990).

According to the findings of Loers et al. (2023), L1 siRNA has an inhibitory effect on the expression of long-chain autism genes neurexin 1 (NRXN1) and neuroligin 1 (NLGN1) and mitochondrial-encoding genes such as NADH ubiquinone oxidoreductase core subunit 2 (ND2). Additionally, Lai et al. (2016) found that FMRP binds to NLGN1 and NLGN3 mRNA, whereas Dahlhaus and El-Husseini (2010) revealed an association between the core symptoms of FXS and the neurexin-neuroligin network. Other studies have shown that myelin basic protein cleaves L1 and promotes neurite outgrowth and neuronal survival (Lutz et al., 2014), confirming an indirect association between L1-CAM and FMRP. However, to date, there have been no studies showing a direct connection between L1-CAM and FXS or FMRP. Future research should focus on the functional role of the L1 gene in the neurexin-neuroligin network and explore its implications for dendritic spine abnormalities in FXS.

A review of pertinent studies has revealed that L1-CAM binds Ankyrin B (AnkB) at a conserved cytoplasmic domain motif (FIGQY), an actin-spectrin adaptor encoded by Ankyrin2, a gene with high confidence in relation to ASDs (Bennett and Healy, 2009; Murphy et al., 2023a). Additionally, L1 knock-in mouse mutants harboring a point mutation at the L1 ankyrin binding site demonstrated augmented spine density in the prefrontal cortex (PFC) (Murphy et al., 2023b). These findings indicate that AnkB may play a vital role in regulating the pruning of dendritic spines in vivo. Meanwhile, various studies have indicated that patients with ASD or FXS display elevated spine density of pyramidal neurons in PFC, where essential circuits contribute to social behavior and cognition (Martínez-Cerdeño, 2017; Murphy et al., 2023b). Using mouse models deficient in L1 family members, Murphy et al. investigated the role of L1 and its interaction with AnkB in dendritic spine regulation in L1-null mice. They found that deletion of L1 or mutation of the FIGQY Ankyrin binding site in the cytoplasmic domain of L1 increased the density of spines on apical dendrites of pyramidal neurons in the mouse neocortex (Murphy et al., 2023a). Moreover, they rescued cortical neurons with impaired dendritic spine development by re-expression of the 220 kDa AnkB isoform in a new inducible mouse model (Nex1Cre-ERT 2: Ank2flox: RCE) (Murphy et al., 2023a). The findings of L1 and its interaction with AnkB in dendritic spine regulation provides a new research direction for the association between L1-CAM and neurodevelopmental diseases such as ASD and FXS and reveal potential pharmacological targets.

Down syndrome cell adhesion molecule (DSCAM) genes are emerging risk genes for ASDs (Varghese et al., 2017; Chen P. et al., 2022). DSCAM is a transmembrane protein belonging to the IgSF class and is classified as a homophilic cell adhesion molecule (Yamakawa et al., 1998; Agarwala et al., 2000; Guo et al., 2021). The DSCAM protein is expressed in the developing nervous system, where it intervenes in various stages of neuronal development. Such effects range from functions in early development (generation, migration, and differentiation) to plasticity and the formation of neuronal networks (Pérez-Núñez et al., 2016). DSCAM is characterized by a large extracellular region, comprising 10 Ig and 6 FN III domains. Additionally, the intracellular domain lacks identifiable motifs (Ly et al., 2008). DSCAM can mediate cell adhesion by forming homophilic dimers between cells and plays a crucial role in neural development by participating in several processes such as axon collateral guidance, dendritic branching, and targeted synaptic formation (Garrett et al., 2012; He et al., 2014; Li et al., 2015; Mitsogiannis et al., 2020; Guo et al., 2021). Recent reports have identified neuroligin 1 (NLGN1) as a novel heterophilic partner that interacts with the extracellular domain of DSCAM (Chen P. et al., 2022). DSCAM on Purkinje cell membranes interact in a heterophilic manner with the glutamate transporter GLAST in astrocytes (Hizawa et al., 2023; Dewa et al., 2024). In Drosophila, DSCAM exhibits remarkable genetic diversity, with tens of thousands of splicing isoforms that modulate the specificity of neuronal wiring. Notably, this splice variant diversity of DSCAM is absent in vertebrates (Hizawa et al., 2023).

A growing body of evidence supports that mutations in the DSCAM gene (Brown et al., 2001; Darnell et al., 2011; Varghese et al., 2017; Mitsogiannis et al., 2020) and increased DSCAM protein expression are associated with FXS pathogenic mechanisms (Sterne et al., 2015; Montesinos, 2017). The FMRP protein binds to DSCAM mRNA (Brown et al., 2001) and inhibits translation of the DSCAM gene in Drosophila and mammalian brain neurons (Darnell et al., 2011). FMRP regulates DSCAM isoform splicing in various cell types to achieve diverse functions (Brown et al., 2001). In Drosophila fragile X mutants, the absence of FMRP leads to increased levels of DSCAM protein owing to translational inhibition, impairing precise synaptic targeting and neural circuit function (Cvetkovska et al., 2013). Abnormal axon targeting caused the degradation of sensory circuit function to an extent that affects Drosophila perception (Cvetkovska et al., 2013). By reducing DSCAM levels in fragile X mutants, scientists observed a reduction in targeting errors and rescued corresponding behavioral responses. Moreover, dysregulation of DSCAM protein expression promotes abnormal dendritic spine development in FXS (Nimchinsky et al., 2001; Cvetkovska et al., 2013). In the mammalian brain, DSCAM is a target gene of FMRP (Darnell et al., 2011). Jain and Welshhans (2016) supported this finding and suggested that FMRP plays a role in regulating the translation of DSCAM mRNA during hippocampal synapse development. They also found that DSCAM mRNA localized to the axons of mouse hippocampal neurons and was dynamically regulated by the axon-guidance molecule Netrin-1 (Figure 3C). Two RNA-binding proteins, FMRP and cytoplasmic polyadenylation element-binding protein, colocalize with DSCAM mRNA and regulate its stability and local translation. Taken together, netrin-1 increases DSCAM protein in growing axons, and overexpression of DSCAM delays axonal growth and branching in mouse cortical neurons, suggesting that netrin-1-induced local translation of DSCAM mRNA is an important mechanism of axonal growth regulation and nervous system development, with increased expression of DSCAM protein leading to structural changes associated with synaptic development (Jain and Welshhans, 2016).

Kim et al. (2013) later showed that DSCAM expression levels are critical in the regulation of presynaptic dendritic development; they detected an association between Drosophila FMRP (dFMRP) and DSCAM mRNA in larval brain lysates using RNA immunoprecipitation and concluded that dFMRP binds DSCAM mRNA and regulates DSCAM expression to inhibit presynaptic dendritic spine development. Sterne et al. (2015) used genetic and pharmacological methods in neurons overexpressing DSCAM and in a Drosophila FXS model to demonstrate the reversal of cell defects caused by imbalanced DSCAM levels in response to Abelson kinase (Abl) inhibition. Abl is a well-established target for treating chronic myeloid leukemia, and multiple Abl inhibitors are approved by the US Food and Drug Administration (FDA) (Speck-Planche et al., 2012). Furthermore, studies indicate the potential for a genetic interaction between DSCAM and Abl in the development of neurites in the brain of Drosophila embryos (Andrews et al., 2008; Yu et al., 2009; Sterne et al., 2015).The investigation showed that DSCAM must interact with Abl to influence presynaptic terminal growth. Besides, the larger presynaptic terminals seen in fruit fly larvae, which produce too much DSCAM, are a result of the DSCAM protein over activating Abl. These findings raise the interesting possibility that targeting Abl might be a viable therapy for brain disorders caused by increased DSCAM expression. Studies have attempted to rescue the developmental defects caused by DSCAM overexpression using Abl inhibitors (Sterne et al., 2015). Sterne et al. (2015) used two Abl kinase inhibitors to treat fruit fly larvae, and found that this reversed the detrimental effects of extra DSCAM on the larvae’s neural circuit. Furthermore, the drugs repaired neural defects in a fruit fly model designed to reproduce FXS symptoms.

Mitsogiannis et al. (2020) found that DSCAM and DSCAM-like 1 (DSCAML1) are highly expressed in neural populations in the embryonic mouse cortex. Moreover, using animal FXS models, researchers have discovered that regulating DSCAM expression can significantly alleviate signs of disease. Other studies have revealed that the spontaneous loss of mouse DSCAM alleles can lead to motor coordination disorders and seizures, with behavioral manifestations similar to those of FXS (Laflamme et al., 2019). Increased occurrence of seizures can be caused by the loss of FMRP function (Musumeci et al., 1999; Kim et al., 2013; Hagerman and Hagerman, 2021), further highlighting the functional significance of dysregulated DSCAM expression in neuronal development.

A study has demonstrated that DSCAM regulates neuronal delamination by exerting local suppression of the RapGEF2-Rap1-N-cadherin cascade at the apical endfeet in the dorsal midbrain. DSCAM is associated with RapGEF2 to inactivate Rap1, whose activity is required for membrane localization of N-cadherin (CDH2). Among them, RapGEF2 (also known as PDZ-GEF1/RA-GEF1), a Rap1-specific guanine nucleotide exchange factor (GEF), was identified as a DSCAM-interacting protein. These findings shed light on the molecular mechanism by which DSCAM regulates a critical step in early neuronal development (Arimura et al., 2020). Together, the previously outlined findings provide possible targets for the treatment of FXS, suggesting that interventions in biological processes related to DSCAM may improve the symptoms of patients with FXS in the future.

Neural cell adhesion molecule (NCAM), an IgSF member, has been identified as a protein target of FMRP (Darnell et al., 2011). NCAM consists of five Ig domains and two FN III domains and is associated with various aspects of synaptic development and function (Sytnyk et al., 2002). NCAM plays a crucial role in the development and maintenance of the nervous system through homophilic and heterophilic interactions (Chu et al., 2018). The absence of NCAM leads to abnormal synaptic differentiation, which not only disrupts synaptic development but also affects synaptic plasticity (Kochlamazashvili et al., 2012). NCAM also possesses signal transduction capabilities associated with neural growth responses mediated by neural cadherin (CDH2) and interacts with L1-CAM, playing a critical role in neurons (Wiertz et al., 2011; Colombo and Meldolesi, 2015). NCAM is crucial not only for the development of the nervous system, but also for maintaining high cognitive functions of the adult brain (Stoyanova and Lutz, 2022).

While there is currently no literature confirming a direct association between NCAM and FMRP, identifying interactions between NCAM and other adhesion molecules, as well as their functions in neural system development, may provide new avenues to explore the relationship between NCAM and FMRP.

In summary, the results not only reveal biological associations between IgSF members and FMRP and their association with FXS but also offer new insights into therapeutic strategies for FXS and FMRP-related disorders. By studying members of this family, we can gain a deeper understanding of the pathogenic mechanisms underlying FXS and those associated with FMRP, laying the foundation for the development of effective treatments and breakthroughs in clinical therapy.

Classical cadherins of the calcium adhesion protein family include E-cadherin (CDH1) and neural cadherin (CDH2). CDH1 is a type I transmembrane glycoprotein located in the adhesive junctions and in the basolateral membrane of epithelial cells. It consists of a large extracellular domain, a transmembrane segment, and a conserved cytoplasmic domain (van Roy and Berx, 2008; Wilkerson et al., 2023). There are relatively rare studies on these proteins related to FXS.

Neural(n)-cadherin, first discovered at synapses, is a calcium-dependent single-pass transmembrane glycoprotein that is mainly expressed on postsynaptic membranes (Obst-Pernberg and Redies, 1999). CDH2 plays a crucial role in both homophilic and heterophilic adhesions at synapses and significantly influences neural system development and functional regulation. Its structure comprises a hydrophobic transmembrane region, an extracellular region, and a highly conserved intracellular C-terminus. The intracellular domain associates with the actin cytoskeleton via p120-catenin, α-catenin, and β-catenin, forming the adherens junction (Angst et al., 2001; Marie et al., 2014; Halperin et al., 2021). Since β-catenin is an effector factor in the canonical Wnt/β-catenin pathway, CDH2 can also modulate signal transduction via Wnt/β-catenin in multiple ways (Marie et al., 2014; Yang et al., 2022). CDH2 is involved in multiple aspects of axon development and morphogenesis, including axon extension, fasciculation, and target selection (Jontes, 2018). It helps establish neuronal polarity in the developing cortex and initiates axon outgrowth (Xu et al., 2015).

La Fata et al. (2014) suggested that CDH2 mRNA is a key target of FMRP during early development and that its reduction leads to delayed development of cortical neurons in patients with FXS. CDH2 also plays a crucial role in the multipolar-to-bipolar neuronal transition during brain development. Additionally, studies using diffusion tensor imaging and magnetic resonance imaging have shown abnormal structural connections in the brains of young patients with FXS. In FXS mouse models, researchers found that FMRP regulates the positioning of cortical plate neurons during embryonic development, thereby affecting the multipolar-to-bipolar transition in these neurons. Correcting this abnormality is possible by reintroducing FMRP or CDH2 during embryonic development. Stan et al. (2010) also discovered that CDH2 interacts with the scaffolding molecule S-SCAM to control the accumulation of vesicles during synaptic development by binding to NLGN1. The precise molecular mechanism underlying the association between CDH2 and FXS remains to be elucidated. However, this represents a promising avenue for future investigation.

A study has indicated that FMRP coordinates Wnt/β-catenin signaling during corticogenesis (Casingal et al., 2020), and CDH2, as one of the targets, simultaneously participates in the Wnt/β-catenin pathway (Yang et al., 2022). In addition, some results suggest that PPARγ agonists such as pioglitazone, rosiglitazone, and the synthetic agonist GW1929, are used as therapeutic agents in neurological disorders. These compounds interact with intracellular transduction signals (e.g., GSK3β, PI3K/Akt, Wnt/β-catenin, Rac1, and MMP-9). It appears that interaction with these pathways may improve memory recognition in FXS animal models (Farshbaf et al., 2014). Taken together, these associations may provide new research directions to explore the role of CDH2 in mediating the Wnt/β-catenin signaling pathway in FXS. The Wnt/β-catenin signaling pathway may be considered a new target for FXS treatment.

Protocadherins (PCDHs) are predominantly expressed in the nervous system and constitute the largest subgroup within the calcium adhesion protein superfamily, comprising more than 80 genes, including 60 genes in the α-, β-, and γ-PCDH gene clusters and non-clustered δ-PCDH genes (Keeler et al., 2015). PCDHs and other atypical cadherins have been shown to play roles in dendrite development and branching and regulation of dendritic spines. They function through homophilic adhesion between neurons (Hoshino et al., 2023).

Here, we focus on PCDH10, also known as OL-protocadherin (Morishita and Yagi, 2007), which contains six extracellular cadherin repeats in the ectodomain, a transmembrane domain, and a unique cytoplasmic domain (Hirano et al., 1999; Kim et al., 2011; Zhen et al., 2023), and is a protein target of FMRP (Darnell et al., 2011). Tsai et al. (2012) reported that PCDH10 is an ASD gene necessary for activity-dependent elimination of excitatory synapses in mice. Myocyte enhancer factor 2 (MEF2) and FMRP collaboratively regulate PCDH10 expression in dendrites. MEF2-induced synapse elimination requires FMRP (Pfeiffer et al., 2010). MEF2 induces PCDH10-dependent degradation of PSD-95 by transferring ubiquitinated PSD-95 to the proteasome. Thus, without MEF2 activation, PSD-95 degradation is not expected to occur. Tsai et al. (2017) discovered that, in FMRP-deficient neurons, increased levels of eukaryotic translation elongation factor 1α (EF1α) prevented mouse double minute 2 (MDM2) ubiquitination of PSD-95 after MEF2 activation, blocking MEF2-induced PSD-95 degradation and synapse elimination (Figure 3D).

Now that PCDH10 is known to target FMRP and MEF2, respectively, perhaps we can further study the relationship between these three in the mouse model of FXS, which may become a potential target for treating FXS.

The CLSTN family of atypical cadherins (Südhof, 2021; Liu et al., 2022) includes calsyntenin 1 (CLSTN1), calsyntenin 2 (CLSTN2), and calsyntenin 3 (CLSTN3). All three CLSTN proteins are expressed in the postsynaptic membranes of neurons (Vogt et al., 2001; Hintsch et al., 2002; Um et al., 2014). CLSTN1 is a type I transmembrane protein with an extracellular domain containing two cadherin repeat sequences and a laminin-alpha/neurexin/sex hormone-binding globulin (LNS) domain (Um et al., 2014). It plays a crucial role in mediating dendritic spine development, synaptic plasticity, and neural circuit formation (Alther et al., 2016; Cheng et al., 2019).

A recent study suggest that CLSTN1 is an important target of FMRP (Darnell et al., 2011). In animal models of FXS, Cheng et al. (2019) first confirmed interactions between CLSTN1 and ICAM5 in the regulation of dendritic spine maturation, demonstrating a key role for CLSTN1 in the development of dendritic spines in Fmr1 KO mice. They further revealed that CLSTN1 is a target of FMRP and that CLSTN1 expression was reduced in multiple brain regions in Fmr1 KO mice, including the cerebellum, resulting in impaired protein transport function. This phenomenon ultimately leads to the accumulation of ICAM5 on cell membranes, further impeding the development and maturation of dendritic spines and synapses and causing abnormal spatial and social learning behavior in Fmr1 KO mice (Cheng et al., 2019; Pei et al., 2020).

Furthermore, other studies suggest that CLSTN3 can form a synaptic adhesion complex with α-NRXNs to induce presynaptic differentiation in developing neurons, thereby participating in synapse formation, regulating synaptic function, and affecting neuron development (Pettem et al., 2013; Um et al., 2014; Gomez et al., 2021; Liu et al., 2022). Currently, there is no clear evidence of a direct association between CLSTN2, CLSTN3, and FMRP. However, their cooperative interactions with other synaptic adhesion molecules that affect the development of synapses make these molecules worth exploring in FXS.

Taken together with the link between CLSTN1 and ICAM5 above, we may try to genetically intervene in CLSTN1, which provides us with a new research idea to further explore the potential pharmacological targets of FXS.