Extracellular neuroligin-ICAM5 coupling drives dendritic growth via actin remodeling

This study identifies ICAM5 as a novel extracellular binding partner for neuroligins that drives dendritic growth by sustaining PAK-Cofilin signaling and actin remodeling, thereby coupling extracellular recognition to intracellular structural development without directly contributing to synaptogenesis.

The Gist

Imagine the brain as a bustling, high-tech city where billions of neurons are the buildings, and they need to build roads (dendrites) to connect with one another. For these roads to grow and reach their destinations, the buildings need to talk to each other.

This paper discovers a new, crucial "conversation" between two specific proteins that helps build these roads. Here is the story in simple terms:

The Main Characters

1. Neuroligins (The Architects): These are like master architects sitting on the surface of a neuron. Their famous job is to help build the actual "offices" where neurons talk to each other (synapses). They are well-known for holding hands with a partner called Neurexin to lock two neurons together.
2. ICAM5 (The Construction Foreman): This is a protein found on the tips of the growing roads (dendritic filopodia). Think of it as a foreman standing at the construction site, holding a blueprint and directing the workers.
3. Actin (The Bricks): Inside the neuron, there is a scaffold made of tiny protein bricks called "actin." To build a road, the neuron needs to stack these bricks quickly and efficiently.

The Discovery: A New Handshake

Scientists already knew that Architects (Neuroligins) and Neurexins were the main team for building offices. But they wondered: Do Architects have other partners to help them build the roads themselves?

Using a high-tech "fishing" method (affinity proteomics), the researchers used a Neuroligin as a bait to see what else it grabbed onto in the brain. They found ICAM5!

It turns out that Neuroligins and ICAM5 shake hands directly. It's a bit of a loose handshake (it doesn't last very long), but that's actually perfect for a busy construction site where things need to change quickly.

What Does This Handshake Do?

The researchers tested what happens when these two proteins work together. They found two surprising things:

1. It's NOT for building offices: When they blocked this handshake, the neurons could still build their "offices" (synapses) just fine. So, this new partnership isn't about connecting neurons to each other yet.
2. It IS for building roads: When they blocked the handshake, the neurons stopped growing their roads (dendrites). The construction stalled.

The Analogy: Imagine the Architect (Neuroligin) is trying to direct a construction crew. If the Architect tries to shout orders alone, the crew gets confused. But if the Architect grabs the Foreman (ICAM5), the Foreman can immediately rally the workers (actin) to start stacking bricks. Without the Foreman, the Architect's orders don't translate into action.

The Mechanism: The "Power Switch"

How does this handshake actually make the bricks move?

The paper found that when Neuroligin grabs ICAM5, it flips a switch inside the cell. This switch activates a signaling pathway involving two proteins: PAK and Cofilin.

• PAK is like the foreman's megaphone, shouting "Build! Build!"
• Cofilin is the worker who actually arranges the bricks (actin).

In neurons missing ICAM5, the megaphone (PAK) goes silent, and the worker (Cofilin) stops working. The result? The road stops growing.

Why Does This Matter?

• Brain Development: This explains how neurons know how to grow long, complex branches to find their neighbors during early development.
• Autism and Neurodevelopment: We know that mutations in Neuroligins are linked to autism. This paper suggests that maybe some of the problems in autism aren't just about neurons failing to connect, but also about them failing to grow the roads to get there in the first place.
• New Therapeutic Targets: If we understand this "Architect-Foreman" team, we might be able to design drugs to help fix the road-building process in people with neurodevelopmental disorders.

The Bottom Line

This paper reveals that Neuroligins have a secret side job. While they are famous for connecting neurons, they also team up with ICAM5 to act as a construction manager. This team tells the cell's internal machinery to rearrange its skeleton (actin), allowing the neuron to grow its branches and explore the brain. Without this specific partnership, the brain's wiring diagram never gets built correctly.

Technical Summary

1. Problem Statement

Neuroligins (NLGNs) are well-established postsynaptic cell adhesion molecules that organize neuronal connectivity by binding to presynaptic neurexins (NRXNs). While the canonical NLGN–NRXN axis is known to drive synaptogenesis, the mechanisms by which extracellular recognition events couple to intracellular growth programs—specifically dendritic outgrowth and cytoskeletal remodeling—remain incompletely understood. Furthermore, NLGNs interact with other modulators (e.g., MDGAs, RPTPδ), but it is unclear if additional, non-canonical extracellular partners exist that regulate neuronal structural development distinct from synapse formation.

2. Methodology

The authors employed a multi-faceted approach combining proteomics, biophysics, cell biology, and mouse genetics:

• Affinity Proteomics: The researchers used purified recombinant Fc-fused NLGN3 ectodomain (NLGN3ECD-Fc) as bait to pull down interacting proteins from detergent-solubilized synaptosomes of postnatal rat brains. Reciprocal pull-downs using ICAM5ECD-Fc were also performed. Mass spectrometry (MS) was used to identify binding partners.
• Biophysical Characterization:
  • Surface Plasmon Resonance (SPR): Used to measure direct binding kinetics and equilibrium dissociation constants ($K_D$) between NLGNs and ICAM5, as well as other known partners (NRXN, MDGA, RPTPδ).
  • Cell-Surface Binding Assays: HEK293T cells expressing various ectodomains were incubated with Fc-fusion proteins to validate interactions.
• Functional Assays in Neuronal Cultures:
  • Dendritic Outgrowth: Cortical neurons from Wild-Type (WT), Nlgn3 Knockout (KO), and Icam5 KO mice were transfected with GFP and various constructs (NLGN3, ICAM5, NLGN4X/Y). Total dendritic length was quantified via MAP2 immunostaining.
  • Synaptogenesis Assays: Heterologous co-culture assays (HEK293T cells expressing NLGNs + cortical neurons) were used to assess presynaptic terminal recruitment (Synapsin-1) and excitatory synapse density (PSD-95).
• Signaling Pathway Analysis:
  • Western Blotting: Analyzed levels of phosphorylated and total proteins involved in mTOR, PI3K/Akt, and PAK–Cofilin signaling pathways in KO neurons.
  • SUnSET Assay: Measured global protein synthesis rates via puromycin incorporation.
  • Immunocytochemistry: Visualized F-actin (phalloidin) and phospho-PAK in dendritic growth cones.

3. Key Contributions

• Discovery of a Novel Interaction: Identified Intercellular Adhesion Molecule 5 (ICAM5/Telencephalin) as a direct, low-affinity binding partner for Neuroligin-3 (NLGN3) and, more broadly, all human neuroligin isoforms (NLGN1–4).
• Functional Decoupling: Demonstrated that the NLGN–ICAM5 axis regulates dendritic outgrowth but is dispensable for synaptogenesis, distinguishing it from the canonical NLGN–NRXN pathway.
• Mechanistic Insight: Defined a signaling cascade where ICAM5 links extracellular adhesion to the PAK–Cofilin pathway, sustaining F-actin organization in growth cones.
• Structural Specificity: Mapped the interaction to the N-terminal Ig1 domain of ICAM5, noting that while Ig1 is required, the full architecture is needed for optimal binding.

4. Key Results

A. Identification and Biophysics of the NLGN–ICAM5 Interaction

• Proteomics: ICAM5 was the most prominent candidate recovered in NLGN3 pull-downs (21 spectral counts), and NLGN3 was recovered in ICAM5 pull-downs (7 spectral counts).
• Direct Binding: SPR confirmed direct binding between NLGN3ECD and ICAM5ECD with a weak affinity ($K_D \approx 30 \mu M$). This affinity is comparable to or weaker than NLGN interactions with NRXN ($\sim 22 \mu M$) and MDGA ($\sim 15 \mu M$), suggesting a transient, dynamic interaction.
• Isoform Specificity: All tested neuroligin isoforms (NLGN1, 2, 3, 4X, 4Y) bind ICAM5. However, ICAM5 selectively binds neuroligins and does not bind other ICAM family members (ICAM1–3).
• Domain Mapping: The Ig1 domain of ICAM5 is essential for binding, but Ig1 alone is insufficient; the intact Ig1–9 architecture is required for full interaction strength.

B. Functional Role in Dendritic Outgrowth vs. Synaptogenesis

• Synaptogenesis: Overexpression of ICAM5 reduced PSD-95 intensity (a marker of excitatory synapses), consistent with its known role as a negative regulator of spine maturation. However, co-expression of ICAM5 did not inhibit the ability of NLGN1 or NLGN3 to recruit presynaptic terminals (Synapsin-1) in co-culture assays. Genetic deletion of Icam5 or Nlgn3 did not alter the synaptogenic effects of the other, indicating they regulate synapse density independently.
• Dendritic Outgrowth:
  • Overexpression of NLGN3 increased total dendritic length in WT neurons. This effect was significantly attenuated in Icam5 KO neurons.
  • Conversely, overexpression of ICAM5 increased dendritic length in both WT and Nlgn3 KO neurons, suggesting ICAM5 can drive growth independently but is required for the full effect of NLGNs.
  • Similar results were observed for NLGN4X (but not NLGN4Y), indicating a conserved mechanism for specific neuroligins.

C. Mechanism: PAK–Cofilin Signaling and Actin Remodeling

• Signaling Deficits in KO: Icam5 KO neurons showed reduced total levels of PAK (p21-activated kinase) and Cofilin, as well as reduced phosphorylation of PAK and Cofilin. In contrast, Nlgn3 KO neurons showed reduced PAK phosphorylation but increased Akt phosphorylation, suggesting distinct upstream regulation.
• Actin Cytoskeleton: Icam5 KO neurons exhibited reduced F-actin levels and phospho-PAK signals specifically within dendritic growth cones.
• Protein Synthesis: Both Nlgn3 and Icam5 KOs showed reduced global protein synthesis (SUnSET assay) and reduced mTOR signaling (phospho-4E-BP1), but the destabilization of the PAK–Cofilin axis was unique to Icam5 loss.
• Conclusion: ICAM5 acts as a downstream effector that sustains PAK–Cofilin signaling, thereby stabilizing F-actin to drive dendritic growth.

5. Significance

• Expansion of the Neuroligin Interactome: The study establishes ICAM5 as a major extracellular partner for neuroligins, expanding the "recognition code" beyond NRXNs, MDGAs, and RPTPδ.
• Separation of Functions: It provides a clear mechanistic separation between synaptogenesis (driven by NLGN–NRXN) and dendritic morphogenesis (driven by NLGN–ICAM5). This explains how neurons can regulate structural growth independently of synaptic connectivity.
• Developmental Context: The findings align with the developmental expression patterns of ICAM5 (high in early filopodia, low in mature spines) and neuroligins, suggesting this axis operates during a specific window of neuronal maturation to shape dendritic arborization.
• Pathological Relevance: Given the link between NLGN3 mutations and Autism Spectrum Disorder (ASD), and the role of ICAM5 in dendritic growth, this pathway may represent a critical node for understanding the structural deficits observed in neurodevelopmental disorders.
• Therapeutic Implications: Understanding how extracellular adhesion couples to intracellular actin remodeling via PAK–Cofilin offers new targets for modulating neuronal connectivity and plasticity.

In summary, the paper defines a NLGN–ICAM5 signaling axis that couples extracellular adhesion to intracellular actin remodeling via the PAK–Cofilin pathway, serving as a critical driver of dendritic outgrowth distinct from synapse formation.