1. Clinical Trials in Alzheimer’s Disease Paul S. Aisen, MD Professor, Department of Neurosciences, UCSD Director, Alzheimer’s Disease Cooperative Study ASENT Symposium Neuroprotection and Disease Modification: Successes, Failures and Lessons Learned

2. AD Therapeutics 1906 Alzheimer’s description 1906 Alzheimer’s description 1970’s Cholinergic hypothesis 1970’s Cholinergic hypothesis 1985 first THA trial 1985 first THA trial 1993 Tacrine approved 1993 Tacrine approved 1997 Vitamin E 1997 Vitamin E 1997 Donepezil 1997 Donepezil 2000 Rivastigmine 2000 Rivastigmine 2001 Galantamine 2001 Galantamine 2003 Memantine 2003 Memantine 2008-2015 Disease-modifying therapy 2008-2015 Disease-modifying therapy

3. FDA Guidelines for AD Trials Co-Primary outcome measures Co-Primary outcome measures Memory/cognition test, plus global or functional measure Memory/cognition test, plus global or functional measure

4. Therapeutic Strategies for AD Symptomatic treatment Symptomatic treatment Disease-modifying treatment Disease-modifying treatment

5. Symptomatic treatment Symptomatic treatment

6. *P<.001. † P<.05. ‡ P=.001. Intent-to-treat population. LS = least squares; SE = standard error; LOCF = last observation carried forward. Source: Winblad B, et al. Neurology. 2001;57:489-495. LS Mean Change in Score From Baseline (±SE) –2.5 –2.0 –1.5 –1.0 –0.5 0 0.5 1.0 Donepezil Placebo Week 52362412 † ‡ 0LOCF * * Donepezil: MMSE ChEI Monotherapy in Mild to Moderate AD: Cognition (N = 286) Improvement Decline

7. Memantine Monotherapy in Moderate to Severe AD: Cognition (N = 252) Difference in SIB Score -12 -10 -8 -6 -4 -2 0 2 41228 Week Memantine Placebo 0 End Point (LOCF) 126 126117 119 106 107 83 96124 123 P<.001 n = SIB SIB = Severe Impairment Battery. Source: Reisberg B, et al. N Engl J Med. 2003;348:1333-1341.

8. Disease-modification Most important need Most important need –AD treatment- beyond symptomatic –MCI, primary prevention Clear, specific targets Clear, specific targets Methodologic problems may be the limiting issue Methodologic problems may be the limiting issue

9. Disease-Modifying Strategies APPAβAβ Neuron death β -secretase γ -secretase inflammation oxidative stress excitotoxicity direct toxicity secretase modulators immunotherapy amyloid binders anti-inflammatories antioxidants neuroprotectants

10. Disease-Modifying Drug Development for AD May be no symptomatic effect May be no symptomatic effect Goal is usually slowing the rate of decline Goal is usually slowing the rate of decline Implications: Implications: –Long trial –Many subjects –Slope analysis? –Still need co-primary outcome measures If “disease-modifying” is to be specified in label: If “disease-modifying” is to be specified in label: –May need two randomization steps in pivotal trials –Alternatively, need convincing biomarker evidence of disease modification

11. Disease-Modifying RX: Phase II problem No short-term benefit expected, rather change in slope of decline No short-term benefit expected, rather change in slope of decline Placebo groups in mild AD studies don’t decline in 6 months; minimal in 12 months Placebo groups in mild AD studies don’t decline in 6 months; minimal in 12 months To see effect on slope, need hundreds or thousands of subjects (1500?-Myriad; 4000?- Elan) followed for 18 months To see effect on slope, need hundreds or thousands of subjects (1500?-Myriad; 4000?- Elan) followed for 18 months CANNOT SEE PROOF OF EFFICACY IN CANNOT SEE PROOF OF EFFICACY IN PHASE II-TYPE TRIAL

12. Phase II Aim for hints of clinical efficacy (Myriad) Aim for hints of clinical efficacy (Myriad) Focus on biomarkers (Alzhemed, IgIV) Focus on biomarkers (Alzhemed, IgIV) Or both Or both

13. Comparison between Tramiprosate and Tarenflurbil Phase II trials Design Element Tramiprosate (Neurochem) Tarenflurbil(Myriad) Target Population Mild-to-moderate (MMSE 15-26) Mild-to-moderate, Mild (MMSE 15-26) Trial Design and Duration 3-arm Double-blind, placebo control (3 months), with single dose open-label extension (41 months) 3-arm Double-blind, placebo control (12 months) with 2-arm cross- over follow-on (12 months) Sample Double-blind: n = 58 Open-label: n = 42 Double-blind: n = 207 Follow-on = 41 Outcome Measures Psychometric (ADAS-cog, MMSE, CDR- SB) Psychometric Biomarker CSF-A β and tau None Efficacy Endpoint Change from baseline Slopes

14. Tramiprosate: CSF-Aß ResultsTarenflurbil: Psychometric Results Source: www.myriad.com Comparison between Tramiprosate and Tarenflurbil Phase II trials

15. Disease-Modification: Trial Design Issues Regulatory issues Regulatory issues –Randomized start/withdrawal –Slope change –Use of biomarkers

16. Cognitive Decline in AD Treatment Trials Months 0369181512 Cog Symptomatic+dis mod Symptomatic Disease-modifying Placebo

17. Randomized Start Design: Symptomatic Months 0369181512 Cog 212427303336 Symptomatic Placebo

18. Randomized Start Design: Disease Modifying Drug Months 0369181512 Cog Disease-modifying Placebo 212427303336

19. Difficulties with long trials Cumulative informative drop-outs Cumulative informative drop-outs Site variance (Alzhemed) Site variance (Alzhemed) Changes to background therapy Changes to background therapy

20. Biomarkers Plasma amyloid Plasma amyloid CSF amyloid, tau CSF amyloid, tau Oxidative, inflammatory markers Oxidative, inflammatory markers Imaging: structural, functional, amyloid Imaging: structural, functional, amyloid

21. Establishing disease modification in AD: comments Standard measures of efficacy Standard measures of efficacy –cognitive plus global/functional –biomarkers will never be sufficient Trial duration: 18-24 months Trial duration: 18-24 months –mild AD is probably the target population –slow rate of cognitive/functional decline in mild AD (NS data) –even with 18 month duration, need very large trials to see 40% effect size Second randomization problematic Second randomization problematic –total trial duration may need to be 36 months  if it takes 18 months to show efficacy, it will take at least another 18 months to show that randomized start group does not catch up –this is not feasible- non-ignorable (informative) drop-outs may exceed 75%; no satisfactory method to manage drop-outs

22. In absence of symptomatic (ie, short- term) efficacy (no efficacy at 3-6 months), long term efficacy suggests disease modification In absence of symptomatic (ie, short- term) efficacy (no efficacy at 3-6 months), long term efficacy suggests disease modification –statistical evidence: growing group difference v. slope change by regression Supported by mechanism, animal model data Supported by mechanism, animal model data Supported by biomarkers related to proposed mechanism (eg, for anti-amyloid rx, CSF Abeta or PET PIB; for anti-tangle, CSF tau), esp. if biomarker validated in animal studies Supported by biomarkers related to proposed mechanism (eg, for anti-amyloid rx, CSF Abeta or PET PIB; for anti-tangle, CSF tau), esp. if biomarker validated in animal studies

23. Tramiprosate (Alzhemed TM ) Phase III Study Design Placebo (n=315) Alzhemed TM 150 mg BID (n=315) Alzhemed TM 100 mg BID (n=315) Alzhemed TM 150 mg BID 78 Week Double - Blind 78 Week Open – label Extension CL-758007 Study entry Titration Period Maintenance Dose Period Titration Period Maintenance Dose Period + AChEI +/- NMDA antagonist

24. Efficacy Endpoints Primary endpoint – Approval: Change from baseline to month 18 in both ADAS ‑ cog and CDR-SB scores – Disease Modification Claim: Rate of brain volume change as measured by MRI Secondary endpoints – Changes from baseline in the MMSE, CIBIC-plus, NPI and DAD scores – Changes from baseline in plasma & CSF A β, CSF tau

25. Tramiprosate Phase III Summary Primary analysis of North American Phase III trial is inconclusive with respect to tramiprosate treatment effect Some descriptive data show numerical differences on the primary clinical endpoints data in favor of tramiprosate. Some descriptive data show numerical differences between groups on the primary disease modification endpoint (MRI).

26. Prevention Trials MCI (secondary prevention?) MCI (secondary prevention?) –Subject selection –Operationalizing AD end-point –Significance of other end-points (cognition, function, global) Primary Prevention Primary Prevention –Enriching population; numbers required; duration –Operationalizing MCI/AD end-point –Significance of other end-points (cognition, function, global) –Simplifying trial design to allow large N Biomarker targets Biomarker targets –Treat plasma/CSF abeta? PIB? CSF tau?

27. AD Trials: current status FDA guidelines have remained consistent FDA guidelines have remained consistent ADAScog, SIB, CIBIC+ have performed adequately ADAScog, SIB, CIBIC+ have performed adequately 5 symptomatic drugs approved 5 symptomatic drugs approved But: But: –No successful MCI trials –No completed prevention trials –Disease-modification??

28. AD Disease-Modifying Trials in Progress Results soon: Results soon: –Myriad Flurizan Phase III –Elan AAB-001 Phase II Later: Later: –Phase III: AAB-001, IgIV, Dimebon, Lilly gamma secretase inhibitor –Phase II: ELND005, ACC-001, Pfizer RAGE inhibitor, DHA