In vivo dissection of human NRXN1 isoforms reveals gain-of-function pathogenicity of schizophrenia-associated 3' deletions

This study utilizes *C. elegans* to demonstrate that specific schizophrenia-associated 3' deletions in the human *NRXN1* gene generate novel isoforms that actively cause pathogenic gain-of-function behavioral defects, revealing a mechanism distinct from simple loss-of-function.

The Gist

The Big Picture: A Glitch in the Brain's "Glue"

Imagine your brain is a massive, bustling city. For the city to function, the buildings (neurons) need to be connected by roads and bridges. Neurexin 1 (NRXN1) is like the super-strong glue or the specialized connector that holds these bridges together, ensuring signals travel smoothly from one building to the next.

When this glue is missing or broken, the city's traffic jams, leading to problems like schizophrenia. Scientists have known for a while that people with schizophrenia often have a "broken" NRXN1 gene. But there was a mystery: Does the broken gene just stop working (a "loss of function"), or does it actually create a new, toxic monster that actively makes things worse (a "gain of function")?

This paper uses tiny worms to solve that mystery.

---

The Experiment: Building a "Human-Worm" Hybrid

To figure this out, the researchers didn't just look at human cells in a petri dish (which is hard to watch). Instead, they used C. elegans, a tiny, transparent worm that is a classic model for studying the brain.

1. The Setup: They took worms that had their own version of the "glue" gene completely deleted. These worms were like cities with no bridges at all—they couldn't move or interact properly.
2. The Swap: They injected human NRXN1 genes into these worms. They tested eight different versions of the human gene:
   • 4 "Normal" versions: These represent the standard, healthy human glue.
   • 4 "Schizophrenia" versions: These are the specific broken versions found in patients with schizophrenia. These versions have a "3' deletion," which is like a chunk of the instruction manual being ripped out of the middle of the book.

The Discovery: It's Not Just Broken; It's Dangerous

The researchers watched how these worms behaved and where the human glue ended up inside the worm's body. Here is what they found, using some fun analogies:

1. The "Traffic Jam" in the Cell Body

In a healthy cell, the glue is supposed to be shipped out to the "bridges" (synapses) where it does its job.

• The Good News: Some of the normal human versions worked well in the worms, fixing the broken bridges. This proves that the basic instructions for making glue are so ancient that they work even in a worm!
• The Bad News: The "Schizophrenia" versions (the ones with the missing chunk) didn't ship out correctly. Instead of going to the bridges, they got stuck in the warehouse (the cell body) or formed clumps (puncta) in the wrong places.
  • Analogy: Imagine a delivery truck that is supposed to deliver bricks to a construction site. Instead, the truck gets stuck in the garage, and the bricks pile up in a messy heap, blocking the garage door.

2. The "Super-Bad" Behavior (Gain of Function)

This is the most surprising part. The researchers tested the worms on two specific tasks:

• Task A: How active are they when they are hungry?
• Task B: Do they like to hang out in groups while eating (social feeding)?

The Result:

• The worms with the "broken" human genes didn't just act like the worms with no glue at all.
• Two specific broken versions (SS77 and SS89) made the worms act worse than the worms with no glue at all.
  • Analogy: If the "no glue" worms are like a car with a flat tire (it won't move), these "broken human gene" worms are like a car with a flat tire and the driver is stomping on the gas pedal while spinning the wheels, causing the car to shake violently and break down even faster.

This is called a "Gain of Function." The broken gene isn't just silent; it's actively toxic. It's creating a new, harmful behavior that the brain doesn't know how to handle.

Why This Matters

For a long time, doctors and scientists thought that if you had a broken NRXN1 gene, the solution was simply to add more of the good gene to fix the problem.

This paper says: "Wait a minute!"

Because some of these broken genes are actively making things worse (the "Gain of Function"), simply adding more good glue might not be enough. It's like trying to fix a house fire by just adding more water, while the fire is being fueled by a new, dangerous chemical. You might need to stop the bad chemical (the broken gene) and add the good water.

The Takeaway

• The Gene: NRXN1 is the brain's connector.
• The Problem: In schizophrenia, a chunk of this gene is missing.
• The Surprise: This missing chunk doesn't just stop the gene from working; it creates a "glitch" that jams the cell and makes behavior worse than if the gene were just missing entirely.
• The Future: Treatments for these patients might need to be very specific. We can't just "fix" the gene; we might need to specifically silence the "bad" version while keeping the "good" one, or find a way to stop the "glitch" from jamming the cell.

By using these tiny worms, the scientists proved that the specific shape of the broken gene matters just as much as the fact that it is broken. This gives us a new roadmap for how to treat these complex brain disorders.

Technical Summary

1. Problem Statement

• Clinical Context: Heterozygous deletions in the NRXN1 gene (encoding Neurexin 1) are among the most frequent rare variants associated with schizophrenia and other neuropsychiatric disorders.
• The Knowledge Gap: While these deletions are known to cause loss-of-function (haploinsufficiency), patient sequencing reveals that 3' deletions often generate novel isoforms not present in the intact locus. It remains unclear whether these novel isoforms are:
  1. Non-functional (passive).
  2. Partially functional.
  3. Actively pathogenic via a gain-of-function (GoF) mechanism.
• Complexity: The NRXN1 gene produces over 1,000 potential splice variants. Understanding how specific patient-derived isoforms (particularly those with 3' deletions) alter protein localization, synaptic function, and behavior is difficult in complex mammalian systems.

2. Methodology

The authors utilized Caenorhabditis elegans (C. elegans) as an in vivo platform to dissect the functional consequences of specific human NRXN1 isoforms.

• Model System:
  • Used the nrx-1(wy778) null mutant strain (lacking the endogenous nrx-1 alpha and gamma isoforms).
  • Crossed into the npr-1(ad609) background to induce social feeding aggregation behavior.
• Isoform Selection:
  • Selected 8 human NRXN1 isoforms derived from schizophrenia patient cell lines (harboring 3' deletions) and control lines.
  • 4 Control Isoforms: Wild-type sequences (SS39, SS47, SS71, SS75).
  • 4 Deletion Isoforms: 3' deletion variants (SS73, SS77, SS84, SS89). All deletion variants lack exons 20 and 22 (specifically removing parts of the C-terminal extracellular domain).
• Experimental Design:
  • Cloning: Human NRXN1 cDNAs were cloned into a C. elegans expression vector (pMPH34) replacing the endogenous nrx-1 cDNA.
  • Promoter: Used the ric-19 promoter to drive pan-neuronal expression.
  • Tagging: All constructs included an N-terminal sfGFP tag for visualization.
  • Rescue Assay: Transgenes were injected into nrx-1(wy778) mutants to test if human isoforms could rescue the null phenotype.
• Assays:
  1. Localization: Confocal microscopy to assess protein distribution (nerve ring vs. cell body vs. puncta).
  2. Food Deprivation Response: Measured locomotion/activity levels in the absence of food using the WorMotel device (8-hour time course).
  3. Social Feeding: Quantified aggregation behavior (clustering of worms) in the npr-1(ad609) background.

3. Key Contributions

• Establishment of a Cross-Species Model: Demonstrated that human NRXN1 isoforms can be functionally expressed in C. elegans neurons, validating C. elegans as a tractable model for studying human neurexin variants despite ~700 million years of evolutionary divergence.
• Discovery of Gain-of-Function Pathogenicity: Provided the first in vivo evidence that specific 3' deletion isoforms are not merely loss-of-function but actively pathogenic via gain-of-function mechanisms, causing phenotypes more severe than the null mutant.
• Isoform-Specific Circuit Effects: Revealed that different isoforms have distinct effects on different behavioral circuits (food deprivation vs. social feeding), suggesting isoform-specific roles in neuronal circuitry.
• Structure-Function Correlation: Linked specific structural deletions (removal of exons 20/21) to aberrant protein localization (accumulation in cell bodies/puncta) and subsequent behavioral deficits.

4. Key Results

A. Protein Localization and Structure

• Control Isoforms: Most control isoforms (e.g., SS71, SS75) localized correctly to the nerve ring and neuropil, similar to endogenous NRX-1.
• Deletion Isoforms: Several 3' deletion variants (specifically SS73 and SS89) showed aberrant localization:
  • Accumulation in neuronal cell bodies.
  • Formation of puncta in the neuropil.
• Structural Insight: AlphaFold modeling predicted that the deletion of exons 20/21 removes a significant portion of the unstructured linker between the transmembrane domain and the extracellular globular domains. This likely impairs ectodomain folding or trafficking, preventing proper synaptic targeting.

B. Behavioral Phenotypes (Food Deprivation)

• Null Phenotype: nrx-1(wy778) mutants showed significantly reduced activity during food deprivation compared to wild-type.
• Partial Rescue: Control isoforms SS71 and SS75 partially rescued the activity deficit, confirming functional conservation.
• No Effect: Control isoforms SS39 and SS47 failed to rescue, indicating intrinsic functional differences among wild-type isoforms.
• Gain-of-Function (GoF): Deletion isoforms SS77 and SS89 caused activity levels to drop significantly lower than the nrx-1 null mutant. This indicates a toxic or dominant-negative effect.

C. Behavioral Phenotypes (Social Feeding)

• Null Phenotype: nrx-1 mutants in the npr-1 background showed reduced social aggregation.
• Rescue: Isoform SS73 (a deletion variant) fully rescued the aggregation phenotype to wild-type levels.
• Gain-of-Function: Isoform SS89 (a deletion variant) caused aggregation levels to drop below the null mutant levels.
• Discordance: Notably, SS73 rescued social feeding but failed to rescue food deprivation, while SS89 was pathogenic in both. This highlights that isoform function is circuit-specific.

D. Summary Table of Findings

Isoform Type	Specific Isoform	Localization	Food Deprivation	Social Feeding
Control	SS71, SS75	Normal (Nerve Ring)	Partial Rescue	Partial Rescue / Trend Up
Control	SS39, SS47	Normal / Soma	No Effect	No Effect
3' Deletion	SS73	Mostly Cell Soma	No Effect	Full Rescue
3' Deletion	SS77	Punctate	GoF (Severe)	Trend Up
3' Deletion	SS89	Punctate	GoF (Severe)	GoF (Severe)

5. Significance and Implications

• Mechanism of Pathogenicity: The study challenges the traditional view that NRXN1 deletions act solely through haploinsufficiency. It demonstrates that the novel isoforms generated by the deletion are actively toxic, likely through:
  1. Dominant-negative effects: Competing for binding partners but failing to signal.
  2. Neomorphic effects: Creating new, aberrant protein interactions.
  3. Proteotoxicity: Accumulation of misfolded proteins in cell bodies.
• Therapeutic Implications: Treatment strategies for NRXN1-associated disorders cannot simply aim to "restore" wild-type expression. Therapies must specifically eliminate the pathogenic gain-of-function isoforms while preserving or restoring functional wild-type isoforms.
• Precision Medicine: The findings underscore the necessity of isoform-level characterization in genetic diagnostics. The specific combination of splice variants expressed in a patient determines the clinical outcome, not just the presence of a deletion.
• Evolutionary Conservation: The ability of human neurexins to functionally interact with C. elegans circuits validates the use of simple model organisms for dissecting complex human neuropsychiatric genetics.

In conclusion, this paper provides a critical mechanistic link between specific NRXN1 structural variants, aberrant protein localization, and distinct behavioral pathologies, establishing a new paradigm for understanding the dual loss-and-gain-of-function nature of neurexin mutations in schizophrenia.