Translation Cliff

The persistent failure of preclinical biology to predict human outcomes — the single most common reason promising drugs die. Discussed extensively at Session IV by all three Nobel laureates.

The Problem

• Beautiful preclinical biology routinely fails to be human biology
• The animals lie — politely, but they lie (Bertozzi)
• Sometimes PK divergence between NHPs and humans; sometimes the underlying biology is just different
• Often fixable — if you can survive the runway; in current funding climate, you usually can’t

Disease-Specific Failures

• Neurodegeneration: animal models are a “persistent embarrassment” (Schekman); newer alpha-synuclein knock-in mice are only now recapitulating Parkinson’s progression
• Alzheimer’s: the amyloid plaque removal → clinical benefit hypothesis consumed decades and billions; plaques are removable but cognitive decline only slows by ~27%
• IL-17: mouse shows no IL-17F signature; the human peripheral blood pattern was hiding the answer (Henry, Session I)

Proposed Solutions

• Better organoids with real organ-level complexity
• Humanized transgenics (a “big white space” in glycoscience — Bertozzi)
• Lower regulatory barrier for small Phase 0 human studies before full clinical commitment
• New Approach Methods (NAMs)
• iPSC-derived human neurons for neurological targets
• Human pathobiology-first approaches

Cross-References

• Human pathobiology — the alternative philosophy
• New Approach Methods (NAMs) — the tools
• Parkinson’s disease — case study in animal model failure