What makes a clinical candidate

(William Z. Potter, Merck Research Laboratories)

Definitions and concepts:

  • Active – A “substance” that meets a threshold level of activity in a primary screen
    • Associated with a HTS
    • Activity level defined by precedent and screen performance
    • Structure and purity of screening substance not confirmed
  • Hit – Compound that provides a suggestive relationship of molecular structure and biological activity to the target
    • Confirmed intrinsic activity and characterization against secondary assays
    • Confirmed structural identity and purity
    • Meets pre-defined criteria for further assessment (usually of limited potency/selectivity)
    • Provides starting point for SAR/structural hypotheses to improve biological potency/selectivity and drug-properties
    • Rarely suitable for in vivo study or advanced mechanistic studies
  • Lead – Compound series derived from a hit that demonstrates a relationship between chemical structure and target-based activity in biochemical and cell-based models; possess physiochemical properties, potency and selectivity deemed appropriate for in vivo evaluation
    • Primary activity in nM range
    • Sub uM activity in cell based assays
    • 10x selectivity for related targets
    • Enhanced solubility, chemical and metabolic stability, evidence of in vivo exposure
  • Biomarker (drug development) – any measure of a drug action proximal to a clinical effect
  • Biomarker (CNS disease) - quantifiable measure in brain tissue (including CSF) associated with severity or course of clinical symptoms (plasma or urine can be utilized when established to reflect CNS)

Factors guiding successful candidate selection (after lead optimization):

  • Hypothesis driven identification of molecular target
    • This approach works best for “me better” drugs, but has a low success rate for compounds designed against novel targets
  • Druggability of target and compound
    • Distribution and density of the target in the CNS
    • Amount of compound reaching target in the CNS
    • Predictability of physiological/functional effects from experimental species to humans
  • Animal models validation
  • See FIG. 1 for overview

There is a great need for predictability from animal studies to humans. Factors to consider:

  • Homology of gene products (e.g. acetylcholinesterase is quite homologous between species, but gamma-secretase is likely less so)
  • Homology of behaviors (e.g. animal behavior tasks can be used as an inference, but not direct measurement of memory)
  • Homology of PD response (e.g. degree of receptor occupancy vs. a functional read-out)

The predictability of animal models remain unproven; there is a high potential for false positives.

  • To date, there is no validation that effects on one or more CNS targets in an animal model predict clinical effects in humans.
  • A hypothesis-driven approach may be a better alternative – drug targets are derived from known biological mechanisms.
  • See REF – ADDF animal models guidelines

Biomarkers are needed to test hypotheses, measure drug action and define relevant patient populations.

  • Biochemical biomarkers – cerebrospinal fluid analyses, magnetic resonance spectroscopy, PET imaging, etc.
  • Physiological biomarkers – EEG, fluorodeoxyglucose (FDG)-PET, functional MRI, arterial spin labeling (ASL), etc.
  • Behavioral biomarkers – responses to tasks
  • Different targets/classes of compounds require different biomarker approaches (See FIG. 2).
  • EXAMPLE: There are many known biological targets for Alzheimer’s disease but only a few may have translatable biomarkers (e.g. acetylcholinesterase - acetylcholine, beta and gamma-secretases – amyloid-beta peptides, dopamine uptake - dopamine, etc.); there are no tools to test hypotheses for the other 100+ targets.

If there is no ligand that can be used for a biomarker, then a novel mechanism-based lead requires a wide net.

  • POC studies can be done in an enriched population with a dose inferred to engage the target in the brain.
  • Current PD “markers” to set the dose: fRMI, EEG, FDG-PET

A good clinical candidate has these characteristics:

  • Hits the molecular target for which occupancy vs. effect can be studied in animal models and humans (rare).
  • Has a strong theoretical basis and solid biochemical readout in animal models and humans (e.g. CSF amyloid-beta).
  • Relatively low-cost POC possible in humans

PET receptor occupancy is the primary biomarker for proof of target engagement.

  • PET imaging provides unique access to the closed brain compartment.
  • PET receptor occupancy provides a direct and quantitative measure of drug interaction with its target in the brain and correlates with dose.
  • However, this method is not always possible in humans (e.g. due to very low receptor density or when the ligand development chemistry is not feasible).
  • Keep in mind that receptor occupancy does not always predict efficacy (especially for agonists and allosteric modulators).
  • See example reference (Bergström 2004)

EEG and CSF analyses have been major PD marker platforms and can be used as alternatives when receptor occupancy studies are not possible.

  • Electroencephalography (EEG) can identify patterns of change in the brain’s electrical activity across species.
    • Particularly useful for drugs that affect level of alertness (e.g. orexin).
    • Deep preclinical investment in obtaining and analyzing EEG in rodents, dogs, and non-human primates.
    • Maturation of drug vs. placebo power spectral analysis increases ability to show dose response relationships.
    • Automated telemetry setups for the measurement of EEG, EMG, and locomotor activity can greatly increase throughput.
  • Cerebrospinal fluid (CSF) sampling is particularly suitable for exploring biochemical changes within individuals.
    • Routine in rats, feasible in mice.
    • Cisterna magna ported non-human primates (with chronically implanted catheter and port system) provide a model for repeat sampling in the same individual
    • CSF Phase 1 studies can be employed in both healthy volunteers and patients.
Working Effectively with Your TTO: Technology Evaluation Process and IP Protection >