Challenges

‍

• Neurotoxicity is a leading cause of drug development failures, with poor predictive power from current preclinical models (animal studies and traditional in vitro assays).

• Existing methods struggle to detect subtle or delayed neurotoxic effects, leading to false positives and negatives.

• Specificity in preclinical animal models remains low, causing overestimation of neurotoxicity risks and resource wastage.

• The lack of physiologically relevant, high-throughput in vitro models hampers early detection and de-risking in drug discovery pipelines.

‍

Methods

‍

• Model Development: Human-induced pluripotent stem cell (iPSC)-derived 3D neural spheroids (“microBrains”) were used to mimic neural networks with both neurons and astrocytes.

• Assay Design: Drug-induced neurotoxicity was assessed using:
  ‍
  1.  Calcium Oscillation Profiles: Evaluated through six parameters, including peak count and amplitude.
     ‍
  2. Cellular ATP Levels: Measured to determine cell viability.

• Compounds Tested: A library of 84 pharmaceuticals (clinical and preclinical compounds) with diverse neurotoxic liabilities.

• Statistical Analysis: Logistic regression was employed to calculate a neurotoxicity (NT) score, determining sensitivity, specificity, and predictive values.

‍

Results

• Predictive Performance:
  ‍
  • The neural spheroid model achieved high specificity (93.33%) but moderate sensitivity (53.49%), outperforming nonclinical animal models in avoiding false positives.
    ‍
  • The NT score successfully differentiated neurotoxic from non-neurotoxic compounds.
    ‍
• Comparison with Traditional Models:
  ‍
  • Animal studies showed higher sensitivity (75%) but poor specificity (30.4%), leading to over-prediction of neurotoxicity.
    ‍
  • Neural spheroids demonstrated superior likelihood ratios (LR+ and iLR-), indicating better reliability for clinical neurotoxicity predictions.
    ‍
• Utility:
  ‍
  • The model effectively detected compounds associated with seizures, convulsions, and neurodegeneration.

‍

‍

Key Takeaways

• Advancement in Predictive Accuracy: The iPSC-derived neural spheroids offer a promising alternative to traditional methods, enabling early identification of neurotoxic compounds.

• Reduced False Positives: High specificity minimizes unnecessary resource allocation to compounds with false neurotoxicity concerns.

• Translatability to Clinical Outcomes: The NT score aligns closely with clinical data, enhancing decision-making in drug discovery.

• Potential for Integration: This model can complement existing preclinical workflows, bridging the gap between in vitro findings and clinical safety profiles.

‍

The study concludes that incorporating the neural spheroid model into pharmaceutical safety assessments can revolutionize neurotoxicity screening, fostering safer and more efficient drug development processes.