CNS Drug Safety Science
Neurotoxicity Screening and CNS Safety Testing
A scientific overview of in vitro neurotoxicity screening, why standard assays miss functional CNS risk, and how human brain microphysiological systems with MEA electrophysiology improve predictive accuracy.
On this page
Definition
What is Drug-Induced Neurotoxicity?
Drug-induced neurotoxicity refers to structural or functional damage to the central or peripheral nervous system caused by a pharmaceutical compound, its metabolites, or reactive species generated during neural metabolism. Neurotoxicity is a significant cause of CNS drug attrition, clinical trial failure, and post-approval safety withdrawals.
Neurotoxicity is broadly classified into two categories based on mechanism:
Structural neurotoxicity
Direct damage to neuronal cell bodies, axons, or dendrites. Detectable via imaging-based viability assays, neurite morphology analysis, and cell death markers such as LDH or live/dead staining.
Functional neurotoxicity
Disruption of neuronal network activity without immediate cell death. Only detectable by electrophysiological methods such as MEA — entirely missed by standard viability assays at therapeutic concentrations.
For background on microphysiological brain models, see the Brain MPS page and the Glossary.
Model Limitations
Why Traditional Models Fail to Predict Neurotoxicity
Simple neuronal cultures (2D monoculture)
Standard 2D cultures of primary neurons or cell lines (e.g., SH-SY5Y) are the most widely used neurotoxicity models, but they have fundamental limitations:
- Lack of supporting glial cells (astrocytes, oligodendrocytes, microglia) which mediate the majority of neuroinflammatory and protective responses
- No functional network connectivity — spiking activity and burst firing that characterise human neuronal networks are absent or non-physiological
- Rapid loss of maturation markers; cultures typically do not reflect adult human CNS biology
- Poor predictors of sub-lethal functional disruption — compounds that alter network synchrony at therapeutic concentrations are undetectable by LDH or ATP assays alone
Animal Models
Rodent and non-human primate in vivo studies are standard for CNS safety assessment, but carry significant limitations for predicting human neurotoxicity:
- Substantial species differences in neuronal composition, CNS pharmacokinetics, and blood–brain barrier permeability
- Behavioural endpoints do not translate directly to molecular CNS mechanisms
- Ethical, cost, and throughput constraints make animal models incompatible with early-stage compound screening
- Regulatory frameworks (FDA Modernization Act 2.0, EU NAM strategies) increasingly encourage human-relevant alternative approaches
Model Requirements
What Makes a Good In Vitro Brain Model for Neurotoxicity Screening?
A high-quality brain model for neurotoxicity screening must capture both the structural and functional dimensions of human neuronal biology:
Key requirements for a predictive brain model
- Established functional neuronal networks with measurable spontaneous electrophysiological activity
- Inclusion of astrocytes and optionally microglia — the primary mediators of neuroinflammatory responses
- Long-term culture stability (14+ days) to enable chronic and repeated-dose studies
- MEA compatibility for non-destructive, real-time monitoring of network activity over time
- Reproducible cell composition and network activity baselines for quantitative compound comparison
- Sensitivity to known functional neurotoxicants at clinically relevant concentrations
NANOSTACKS™ Approach
How NANOSTACKS™ Supports Neurotoxicity Screening
Revivocell's Brain MPS platform uses NANOSTACKS™ inserts to build multicellular neural models with physiological architectural organisation. Dynamic culture conditions generated by orbital shaking provide mechanical stimulation without external pumps, maintaining network maturation and stability over weeks.
Two validated configurations are available, each suited to different safety assessment objectives:
Imaging-Based Brain MPS
Neurons + Astrocytes
Structural neurotoxicity assessment. Quantification of neurite length, branching, cell viability, and morphological changes via high-content imaging.
MEA-Compatible Brain MPS
Neurons + Astrocytes (MEA substrate)
Functional neurotoxicity and seizure liability. Real-time electrophysiological recording of spike rate, burst frequency, and network synchrony.
For compounds with suspected CNS liability, both configurations can be run in parallel to provide a complete structural and functional neurotoxicity profile.
Cell Biology
Cell Types and Their Roles
| Cell Type | Primary Function | Role in Toxicity Detection |
|---|---|---|
| Excitatory neurons | Generate action potentials and form synaptic networks | Primary targets of structural and functional neurotoxicity; produce the spike and burst activity detected by MEA |
| Astrocytes | Metabolic support, neurotransmitter recycling, synapse modulation | Mediate neuroinflammatory responses; protect or sensitise neurons to toxic insult depending on compound mechanism |
| Microglia (optional) | CNS resident immune cells; phagocytosis and cytokine secretion | Key mediators of immune-driven neuroinflammation; activate in response to damage-associated molecular patterns |
Analytical Endpoints
Key Readouts and Endpoints
The following analytical endpoints are available with NANOSTACKS™ Brain MPS:
Functional (MEA) endpoints
- Mean spike rate per electrode
- Burst frequency and duration
- Network synchrony index
- Inter-burst interval
- Compound effect on baseline activity (% inhibition / activation)
Structural and viability endpoints
- Neurite length and branching (high-content imaging)
- LDH release (membrane integrity)
- ATP content (metabolic viability)
- Live/dead staining (fluorescence imaging)
- Axonal integrity (β-III tubulin immunostaining)
Neuroinflammatory markers
- Pro-inflammatory cytokines (IL-1β, IL-6, TNF-α) by ELISA
- GFAP upregulation (astrocyte reactivity)
- Complement activation markers
Advanced and AI-ready outputs
- Time-series electrophysiology (minutes to weeks)
- Multiparametric feature extraction for machine learning
- High-content imaging Z-stacks
- Dose–response curve generation
Drug Discovery Use Cases
Applications in Drug Discovery
Early-stage CNS safety screening
Rank-order neurotoxicity risk of compound series during lead optimisation using MEA-based functional readouts to deprioritise CNS-active liabilities before resource-intensive in vivo studies.
Seizure liability assessment
Detect pro-convulsant compound activity by analysing changes in burst frequency, network synchrony, and inter-burst interval in response to compound exposure — a critical ICH S7B-relevant endpoint for CNS and non-CNS drugs alike.
Chronic and repeated-dose neurotoxicity
Assess progressive neuronal network degradation over 14–28 day dosing regimens, relevant to neurodegenerative conditions, oncology CNS compounds, and drugs requiring prolonged therapeutic exposure.
Mechanistic neurotoxicity investigation
Distinguish between excitotoxicity, mitochondrial dysfunction, neuroinflammatory mechanisms, and axonal transport disruption to inform structural modification and de-risking strategies.
New Approach Methodology (NAM) packages
Generate human-relevant CNS safety data under the FDA Modernization Act 2.0 and EU NAM strategy frameworks, as a complement or alternative to in vivo CNS safety pharmacology studies.
Discuss a neurotoxicity study
Our team can advise on model selection, study design, and interpretation of CNS safety data for your compounds.
