LRFN2

LRFN2 (SALM1) is a leucine-rich repeat transmembrane adhesion molecule that organizes postsynaptic signaling complexes to control excitatory and inhibitory synapse development, synaptic plasticity, and NMDA receptor-mediated calcium signaling in both neuronal and non-neuronal contexts. LRFN2 directly binds PSD-95 via its PDZ-binding motif and NMDA receptor subunits (NR1, GRIN2B), and its interaction with PSD-95 is required for synaptic surface expression of AMPA receptors (GluA1), normal postsynaptic density size, and spine maturation; knockout mice exhibit reduced AMPA/NMDA ratio, enhanced LTP, decreased inhibitory synapse density, and behavioral abnormalities (PMID:28604739, PMID:29798891). Intracellular trafficking of LRFN2 is governed by a dileucine ER retention motif that restricts surface expression and by the endosomal adaptor SNX27, which sorts LRFN2 to the surface and thereby indirectly controls AMPA receptor-mediated synaptic transmission (PMID:22174418, PMID:34251337). Beyond the central nervous system, LRFN2 modulates NMDA receptor-dependent calcium influx in erythroblasts to regulate erythropoiesis, is a component of the cone photoreceptor–depolarizing bipolar cell signaling complex required for normal photopic vision, and suppresses Wnt/β-catenin and NF-κB signaling through GRIN2B–GSK3β interaction in cancer cells (PMID:33481887, PMID:38408870, PMID:35879265).

1. 2007 Medium

   Establishing that LRFN2 functions outside the nervous system, gain-of-function experiments showed LRFN2 expression skews hematopoietic differentiation toward erythropoiesis and cooperates with Myc to cause erythroblastosis, revealing an unexpected role in cell fate decisions.

   Evidence Colony-forming assays with LRFN2 overexpression and transgenic mouse crosses

   PMID:17577922

   Open questions at the time

   • Mechanism linking LRFN2 to erythroid lineage commitment not identified
   • Whether NMDAR signaling mediates this effect was not tested
   • Single laboratory finding
2. 2011 High

   Addressing how LRFN2 surface expression is controlled, mutagenesis revealed a dileucine ER retention motif that restricts LRFN2 to the ER and a PDZ-binding motif required for dendritic (but not axonal) surface delivery, establishing a dual-signal trafficking mechanism that gates LRFN2's synaptic availability.

   Evidence Serial deletion mutagenesis, endoglycosidase H assays, heterologous cell and hippocampal neuron transfection, electron microscopy

   PMID:22174418

   Open questions at the time

   • Identity of PDZ-domain proteins that release ER retention was not determined
   • Whether trafficking regulation is activity-dependent was not tested
3. 2015 Medium

   Ultrastructural localization confirmed that LRFN2 resides at synapses and associates with the NR1 subunit of NMDA receptors, establishing LRFN2 as a bona fide synaptic NMDAR-associated protein.

   Evidence Electron microscopy immunogold labeling in rat cerebellar and hippocampal neurons

   PMID:26486473

   Open questions at the time

   • Single-lab finding without reciprocal biochemical validation of NR1 binding
   • Functional consequences of the LRFN2–NR1 association were not tested
4. 2017 High

   Knockout studies established that LRFN2's direct interaction with PSD-95 is required for synaptic delivery of AMPA receptors (GluA1), normal postsynaptic density structure, and spine maturation, and that loss of LRFN2 reduces the AMPA/NMDA ratio while paradoxically enhancing LTP.

   Evidence Lrfn2 KO mice with in vitro binding assays, electrophysiology, and morphological synapse analysis

   PMID:28604739

   Open questions at the time

   • Whether LRFN2 directly scaffolds AMPA receptors or acts solely through PSD-95 was not resolved
   • Mechanism of enhanced LTP in KO remained unclear
5. 2018 High

   Independent KO studies extended LRFN2's role to inhibitory synapse development, showing that Lrfn2-/- CA1 neurons have decreased inhibitory synapse density and reduced spontaneous inhibitory transmission, establishing LRFN2 as a bidirectional organizer of excitatory–inhibitory circuit balance.

   Evidence Lrfn2 KO mice, whole-cell patch clamp, immunofluorescence, behavioral assays

   PMID:29798891

   Open questions at the time

   • Whether the inhibitory synapse phenotype is cell-autonomous or circuit-level was not distinguished
   • Molecular partners mediating the inhibitory synapse role were not identified
6. 2021 High

   Identification of SNX27 as an endosomal sorting partner for LRFN2 revealed the mechanism by which LRFN2 surface levels are regulated post-ER, and demonstrated that LRFN2 indirectly controls AMPA receptor surface expression and hippocampal LTP through this trafficking pathway.

   Evidence Proteomics in rat primary neurons, Co-IP/pulldown, LRFN2 knockdown, surface biotinylation, mEPSC and LTP recordings

   PMID:34251337

   Open questions at the time

   • Whether SNX27 and PSD-95 compete for the same PDZ-binding motif on LRFN2 was not tested
   • Recycling kinetics and activity dependence of LRFN2 endosomal sorting were not measured
7. 2021 Medium

   KO mice demonstrated that LRFN2 regulates NMDA receptor-mediated calcium influx in late erythroblasts, connecting the synaptic adhesion molecule to erythropoiesis control through NMDAR function in hematopoietic cells.

   Evidence Lrfn2 KO mice, flow cytometry, peripheral blood analysis, CFU assay, calcium influx assay with MK801

   PMID:33481887

   Open questions at the time

   • Single-lab finding; independent replication needed
   • Direct physical interaction between LRFN2 and NMDAR in erythroblasts was not demonstrated biochemically
   • Downstream signaling from NMDAR calcium influx to EPO receptor regulation not mapped
8. 2022 Medium

   Co-immunoprecipitation established a direct LRFN2–GRIN2B interaction in cancer cells, and overexpression experiments showed LRFN2 suppresses Wnt/β-catenin and NF-κB signaling through a GRIN2B–GSK3β axis, demonstrating a tumor-suppressive signaling function outside the nervous system.

   Evidence Co-IP in ESCC cells, overexpression/western blot, pharmacological antagonism with NMDA-IN-1, in vivo tumor growth assay

   PMID:35879265

   Open questions at the time

   • Single-lab finding in one cancer type
   • Whether endogenous LRFN2 levels are sufficient to engage this pathway in normal tissue unknown
   • Structural basis of LRFN2–GRIN2B interaction not resolved
9. 2023 Medium

   Knockdown experiments in bladder cancer revealed that tumor-intrinsic LRFN2 suppresses pro-inflammatory cytokine/chemokine secretion, inhibiting CD8+ T cell recruitment and enabling immune evasion, identifying LRFN2 as a regulator of the tumor immune microenvironment.

   Evidence LRFN2 knockdown in bladder cancer cells, multiplex immunoassay, in vivo models, single-cell RNA-seq, spatial tissue analysis

   PMID:37802603

   Open questions at the time

   • Molecular mechanism connecting LRFN2 to cytokine/chemokine transcription or secretion machinery not identified
   • Whether NMDAR signaling mediates the immune-evasion phenotype was not tested
10. 2024 Medium

    Proteomics and KO studies identified LRFN2 as a component of the cone photoreceptor–depolarizing bipolar cell signaling complex, demonstrating that LRFN2 is required for normal cone-mediated synaptic transmission in the retina.

    Evidence Unbiased proteomics, immunolocalization at cone terminals, Lrfn2 KO mice, photopic ERG

    PMID:38408870

    Open questions at the time

    • Single-lab finding; specific binding partners at cone terminals not identified
    • Whether LRFN2 functions as a trans-synaptic adhesion molecule at ribbon synapses was not tested

• The trans-synaptic binding partners of LRFN2, the structural basis of its interactions with NMDAR subunits and PSD-95, and the molecular mechanism by which LRFN2 controls cytokine secretion and immune evasion in tumors remain unresolved.
• No trans-synaptic ligand identified for LRFN2
• No crystal/cryo-EM structure of LRFN2 or its complexes
• Mechanism linking LRFN2 to cytokine/chemokine regulation in cancer is unknown
• Whether neuronal and non-neuronal functions share a unified NMDAR-dependent mechanism is untested