Engineering an in vitro model of demyelinated spinal cord tissue

This study presents a robust, reproducible *in vitro* model of spinal cord demyelination using engineered human neural stem cell-derived tissues on piezoelectric scaffolds, which successfully recapitulates key pathological features and serves as a versatile platform for investigating disease mechanisms and therapeutic interventions.

The Gist

Imagine your nervous system as a massive, high-speed internet network. The "cables" are your nerve fibers (axons), and the "insulation" wrapping around them is called myelin. Just like the plastic coating on an electrical wire prevents energy from leaking out, myelin ensures your brain's signals zip along quickly and efficiently.

When diseases like Multiple Sclerosis (MS) strike, they strip away this insulation. The signals get slow, garbled, or stop completely. While scientists have built many models to study this "insulation stripping" in the brain, the spinal cord (the main cable running down your back) has been largely ignored. Why? Because the spinal cord is long, thin, and hard to study in a petri dish.

This paper is about building a miniature, high-tech spinal cord in a lab to finally understand how to fix these broken cables.

Here is the story of how they did it, broken down into simple steps:

1. The Blueprint: Building a "Train Track" for Nerves

In the brain, nerve cells grow in all directions, like a tangled ball of yarn. But in the spinal cord, they are like a straight train track, running long distances in one direction.

To mimic this, the researchers didn't just dump cells into a dish. They used a clever trick:

• The Microwells: Imagine a cookie cutter with two holes separated by a long, narrow bridge. They used a mold to force nerve cells to grow in two distinct "towns" (colonies) with a gap in between.
• The Piezoelectric Scaffold: They laid down a special fabric made of tiny fibers that generate a tiny electric spark when you squeeze or bend them (like a lighter that sparks when you press it).
• The "Gym" for Nerves: They placed this fabric in a machine that gently vibrated it up and down. This created Mechano-Electrical Stimulation (MES). Think of it as a personal trainer for the cells. The vibration and electric sparks told the cells: "Don't just sit there! Grow long, straight, and strong!"

The Result: They successfully grew human nerve cells that stretched across a 2,000-micrometer gap (about the width of a human hair, but huge for a cell culture) and became covered in myelin. They even grew the right mix of "wiring" (neurons) and "insulation" (glial cells).

2. The Stress Test: Breaking the Model on Purpose

Now that they had a working "mini-spinal cord," they needed to break it to see how it fails, just like a crash test dummy. They used two different "weapons" to strip the insulation, hoping to mimic different types of damage:

• Weapon A: The "Cuprizone Cocktail"
  • What it is: A mix of a toxin (cuprizone) and inflammatory chemicals (cytokines).
  • The Effect: This was like a nuclear bomb. It didn't just strip the insulation; it also blew up the cables themselves. The nerve fibers broke, and the signal transmission died completely. This mimics severe cases where both the wire and the insulation are destroyed.
• Weapon B: The "LPC" (Lysophosphatidylcholine)
  • What it is: A chemical found in the spinal fluid of MS patients.
  • The Effect: This was like a precision laser. It stripped the insulation off the wires but left the wires themselves mostly intact. The cables were still there, but they were exposed and slow.

3. The Diagnosis: Listening to the Signals

How did they know the model was working? They hooked the mini-spinal cord up to a Microelectrode Array (MEA), which is basically a high-tech stethoscope with 60 tiny microphones.

• Healthy Model: When they sent a signal in, it zipped across the gap at high speed with a loud, clear "voice" (strong electrical signal).
• Cuprizone Model: The signal barely made it across. It was weak, slow, and often died out. The "cables" were broken.
• LPC Model: The signal made it across, but it was sluggish and weak. The "insulation" was gone, so the electricity leaked out, slowing everything down.

Why This Matters

Think of this new model as a flight simulator for spinal cord diseases.

Before this, scientists mostly studied brain models (which are like tangled yarn) or animal models (which don't always match human biology). This new "mini-spinal cord" is the first to accurately recreate the long, straight, insulated highways of the human spinal cord.

The Big Takeaway:
Because they can now grow these tissues and break them in specific ways, they can test new drugs.

• If a drug fixes the "LPC" model, it might be great at remyelinating (re-insulating) nerves.
• If a drug fixes the "Cuprizone" model, it might be great at protecting the nerve fibers from dying.

In short, the researchers built a tiny, human spinal cord in a dish, broke it in two different ways, and proved they can measure exactly how it fails. This gives them a powerful new tool to find cures for diseases that leave people paralyzed or in pain.

Technical Summary

1. Problem Statement

Demyelinating diseases, such as Multiple Sclerosis (MS), involve the destruction of the myelin sheath, leading to impaired nerve signal conduction and potential axonal degeneration. While approximately 80–90% of MS patients develop lesions in the spinal cord (SC), existing research models predominantly focus on brain pathology. Current limitations include:

• Anatomical Discrepancy: The SC possesses unique features, such as long, uniaxially aligned axons, which are difficult to replicate in standard brain organoids or spheroids.
• Translational Gaps: Animal models fail to fully recapitulate human pathophysiology, and existing in vitro models often lack the structural complexity (extended axons, specific cellular diversity) required to study long-distance signal transmission defects characteristic of SC demyelination.
• Imaging Challenges: In vivo imaging of the SC is hindered by small size, anatomical variability, and motion artifacts, necessitating robust in vitro alternatives.

2. Methodology

The authors developed a bioengineered in vitro platform combining microwell technology, piezoelectric scaffolds, and mechano-electrical stimulation (MES) to create human neural stem cell (hNSC)-derived nerve tissues.

• Scaffold Fabrication:

  • Material: Electrospun poly(vinylidene-trifluoroethylene) (P(VDF-TrFE)) nanofibrous scaffolds (500 nm diameter, ~200 µm thick) were synthesized.
  • Processing: Scaffolds were annealed at 90°C to enhance crystalline phase transition and piezoelectricity.
  • Function: The scaffolds generate electrical stimuli via the piezoelectric effect when mechanically deflected, promoting neural differentiation.

• Microwell System & Cell Seeding:

  • Design: PDMS microwell molds (1200 µm × 400 µm × 500 µm) were 3D-printed and used to spatially control hNSC inoculation.
  • Configuration: Two colonies of cells were seeded with varying gap distances (500–2000 µm) to force axonal bridging.
  • Stimulation: A vertical translation stage applied periodic mechanical motion to the scaffold, inducing MES.

• Biochemical & Temporal Control:

  • Phase 1 (Days 1–14): Media supplemented with Nerve Growth Factor (NGF) and a ROCK inhibitor (Y-27632) to promote axonal elongation while suppressing premature glial differentiation.
  • Phase 2 (Weeks 3–5): Withdrawal of NGF/ROCK inhibitor to allow oligodendrocyte differentiation and myelination.
  • Duration: Cultures were maintained for 5 weeks to achieve mature, myelinated tissues.

• Demyelination Induction:
  Two distinct chemical protocols were tested to model different pathological severities:

  1. Cuprizone Cocktail: 1000 µM cuprizone + 50 ng/mL TNF-α + 50 ng/mL IFN-γ (24 hours). This targets oligodendrocyte toxicity and inflammation.
  2. Lysophosphatidylcholine (LPC): 30 µg/mL (24 hours). This directly disrupts lipid membranes to induce demyelination.

• Characterization:

  • Immunofluorescence: Staining for axonal markers (β3-tubulin), myelin markers (MBP, MOG, GalC), and synaptic markers (VGLUT1/PSD95, VGAT/Gephyrin).
  • Electrophysiology: Multielectrode Array (MEA) recordings to measure action potential amplitude, conduction velocity, and network connectivity.

3. Key Contributions

• First Reproducible SC In Vitro Model: Successfully engineered human nerve tissues with extended, uniaxially aligned axons up to 2000 µm in length, mimicking the anatomical structure of the spinal cord.
• Dual-Pathology Modeling: Demonstrated the ability to induce two distinct pathological states:
  • Cuprizone Cocktail: Induced both myelin loss and axonal degeneration (severe pathology).
  • LPC: Induced selective demyelination while largely preserving axonal integrity (milder pathology).
• Integration of MES: Validated the use of piezoelectric scaffolds combined with microwell geometry to guide complex tissue morphogenesis and functional maturation.

4. Results

• Structural Development:

  • Axons successfully bridged gaps up to 2000 µm with a mean orientation angle of 90.21° (Gaussian distribution), confirming uniaxial alignment.
  • By Week 3, oligodendrocyte differentiation (GalC+) and mature myelination (MBP+, MOG+) were observed.
  • Functional synapses (both excitatory and inhibitory) formed within colonies and along the extended axons, indicating a mature neural network.

• Demyelination Outcomes:

  • Control: Robust myelin sheath (MOG+) and continuous axonal networks.
  • Cuprizone Cocktail: Significant reduction in MOG expression and severe disruption of the axonal network (disconnected β3-tubulin) in both colonies and gaps, indicating axonal death.
  • LPC: Selective loss of myelin (reduced MOG) primarily in the inter-colony gaps, while the axonal network (β3-tubulin) remained largely intact.

• Electrophysiological Validation:

  • Control: High signal amplitude (1000 µV) and conduction velocity (1.2 m/s).
  • Demyelinated Groups: Both groups showed significantly reduced amplitude (200 µV) and velocity (0.4 m/s), consistent with impaired saltatory conduction.
  • Differentiation: The Cuprizone group showed the most severe signal loss (localized near stimulus), correlating with axonal degeneration. The LPC group retained signal propagation over longer distances than the Cuprizone group, correlating with preserved axonal structure despite demyelination.

5. Significance

• Pathophysiological Insight: This platform provides a human-relevant tool to dissect the specific mechanisms of SC demyelination, distinguishing between pure myelin loss and axonal degeneration, which is critical for understanding MS progression.
• Therapeutic Screening: The model offers a versatile, high-throughput in vitro system for testing potential disease-modifying therapies (DMTs) that target remyelination or axonal protection separately.
• Overcoming Limitations: By replicating the long-distance, uniaxial architecture of the spinal cord, this model addresses the gap in current research that focuses heavily on brain-centric models, potentially accelerating the development of treatments for spinal cord-specific neurological disorders.