1. Clinical Trials: Challenges and Opportunities
Erika Augustine ,MD
University of Rochester Medical Center
October 31, 2012
Good afternoon. Thank you to the organizers for the invitation to participate in today’s NDC symposium. You have heard fantastic talks today about our current clinical and scientific understanding of the NCLs, and in this talk, I hope to present a forward-looking view to lead us into the final panel of the day – future directions. I will be talking about challenges and opportunities for clinical trials.

2. Current state of treatment in NCLs
No FDA approved disease-modifying therapy for any NCL
Symptomatic treatments
At least 5 recent or ongoing clinical trials
Goals: stabilization, improvement, or reversal of symptoms
Before we look forward – let’s recap where are we now with respect to treatment in the NCLs. It is dismaying that nearly 200 years after Stengel described the first patients, we are still completely lacking in established disease-modifying therapies.
This does not mean that we have nothing to offer families – we are able to target specific symptoms like seizures and behavioral challenges, albeit in a manner that is relatively independent of the actual NCL diagnosis..
Yet, encouragingly, there have been at least five completed or ongoing clinical trials in the NCLs in recent years. And I do believe we’re entering a new era of therapeutics.
What do we hope for new therapeutics? Simply put we strive for a cure, we strive to change lives – but even the term cure can have multiple meanings and multiple steps along the way, Along the spectrum of possibilities – we could halt disease in it’s tracks, we could improve symptoms, and in a perfect world perhaps even reverse disease. Understanding the goal is important along every aspect of drug discovery.
Rare/orphan disease definition – low prevalence and the idea that current therapies are inadequate
Challenges in developing therapies for rare disorders
Novel therapeutic strategies
Innovative approaches to CNS delivery

3. Clinical trials in rare diseases
<200,000 persons in US
>8,000 rare diseases affecting 6-10% of US population
I’d like to place this a broader context. The NCLs are rare diseases, which the Orphan Drug Act defines as disorders affecting fewer than 200,000 individuals in the US. At times it seems our focus on the NCLs is an extremely small population . . Yet when taken collectively, the over 8,000 recognized rare diseases affect almost 30million individuals and their families. Many of the concepts addressed here are common to rare diseases, and so our biggest opportunity we have is to learn from each other.
RC Griggs, et al. Molec Genet Metab 2009; 96:

4. Drug Development Timeline
We would like to have new options now. Drug development can take 10, 15 years, or longer once preclinical testing, and the traditional sequence of clinical trials take place. Phase I (first in human), phase II (dosing, safety, and preliminary efficacy), and phase III (pivotal efficacy studies). These are followed by New Drug Application and FDA review. We want to shorten this timeline , but without compromise in safety, quality, or efficacy. – true in common and rare diseases alike.
H Liu and M Schmid. J Comm Biotech 2009; 15: 199–214.

5. Challenges to Therapeutic Development in Rare Diseases
Proof-Of-Concept
Trial Execution
There are challenges in developing therapies for rare disorders. These challenges can be organized into two broad but related categories:
The first I think of as factors that relate to Proof-of-concept: How do we find new agents to study?
The second category are factors that relate to Trial execution: We have an idea with reasonable likelihood of success (so we think) – but how do we actually prove whether or not it’s effective?

6. Let’s take each of those concepts in turn.
For infantile, late infantile, and juvenile NCL, we are in a fortunate position of having identified the causative gene for each, yet our full understanding of pathophysiology remains incomplete.
When there are pathophysiologic questions, how do begin to develop therapeutic targets? Do we target the gene defect itself or downstream pathophysiology, both? How do we assess whether the target is appropriate?
In neurodegenerative disorders, there is the additional issue of CNS delivery or crossing the blood-brain barrier.
No animal model is perfect. So the question becomes how good are they? How well do the preclinical models predict human disease and human drug activity. And how accurate are these models for the target in question, mode of delivery, and estimation of anticipated clinical benefits.
In animal models, we may be able to assess target engagement post-mortem. But how do we achieve that in vivo, in humans? For some targets that may not be possible.
Overall - When have we learned enough to launch a therapy development program?
I won’t speak about each of these, but I would like to elaborate on opportunities for one . . .
What is the level of unmet need (clearly it is high in the NCLs)?
What is the risk/benefit ratio – particularly with respect to novel targets
Challenge 1: Science – understanding disease, developing targets

7. Proof-Of-Concept: Opportunities
Therapeutics for Rare and Neglected Diseases (TRND program)
These are challenges facing all diseases that are magnified across all rare diseases, where the absolute number of researchers are fewer and the funding dollars are less. As NCL researcher, as rare disease researchers, it may not be feasible from a cost or resource allocation standpoint to develop all that is needed across the translational spectrum at a single academic center, or set of centers. So it becomes critical that we make use of existing resources.
NCATS – the NIH National Center for Advancing Translational Sciences, and specifically within NCATS, the TRND program – Therapeutics for Rare and Neglected Diseases – aim to stimulate drug discovery and speed development of new and repurposed drugs. A unique aspect of these programs is access not to funding dollars, but to pre-clinical study resources needed to enable future IND applications. This is an opportunity for bench and clinical researchers to come together and work to accelerate development of new interventions for NCL.

8. The second set of challenges relate to Clinical Trial Execution – or the process of human study.
Finding individuals with interest or experience to design trials for rare disorders and who understand the rigors of experimental therapeutics.
Appropriate trial design to answer the question
Appropriate measurements to complement the design
Selecting the right sample and recruiting them to participate in an ethical manner.
Finding the funds to complete the study and having knowledge and resources to take on the many regulatory issues that may arise.
These challenges are not unique to rare diseases, but are magnified by the limitations that studying a small population bring.

9. Gold Standard – blinded RCT
Gold standard for establishing cause and effect
Random allocation of subjects - minimizes bias (selection bias) - balances known and unknown confounders - facilitates blinding
Large samples, long follow-up, high cost
We consider the double blind randomized controlled trial to be the gold standard of establishing efficacy in a trial setting.. This design minimizes bias selection bias, should distribute confounders equally between study groups, and facilitates blinding, which may improve participant compliance and retention, and limits investigator bias in outcomes assessment. Blinded RCTs looking for small effects can be costly and time consuming and often enroll a large number of patients – these aspects are less feasible in the NCLs.
Schulz, et al. Lancet 2002; 359: 696–700

10. Trial Execution: Study Design
CHALLENGES
Purpose
Selection of subjects
Selection of controls, placebo
Measures to minimize bias - randomization
Statistical analysis - Sample size = Power - Individual Variability
Understanding the disease and the goals of any one study directly inform study design. While the RCT is the gold standard, a small uncontrolled trial may be used when the disease is homogenous, well understood, and the anticipated effect size is large.
Yet, in the NCLs, even as single gene disorders, there is considerable variability, and decline that is superimposed on childhood development, making the clinical course anything but homogenous, even within a single family. There may be a desire to study patients with early disease or in contrast advanced disease. But, where patients are few, it is not always feasible to significantly narrow the entry criteria.
Trials are further strengthened through use of a comparator, but in a fatal degenerative disorder, ethical questions about use of placebo may arise. And again this becomes an issue of sample size. Randomization doesn’t apply in a single-group study, so other means are needed to minimize bias. And as it relates to analysis, we are at a dual disadvantage. By necessity, these are small studies, but there is also individual variability in clinical course, both of which diminish a study’s power. We need design and statistical techniques that maximize data from small numbers of subjects. When a compound fails, we need to truly understand why. It is ideal to avoid missing an effect due to study design and sample size, and really have a sense that there is a true lack of biological effect.
Dickson, et al. Mol Genet Metab March ; 102(3): 326–338

11. Trial Execution: Study Design
CHALLENGES
OPPORTUNITIES
Purpose
Selection of subjects
Selection of controls, placebo
Measures to minimize bias - randomization
Statistical analysis - Sample size = Power - Individual Variability
Historical controls
Self controls
Alternative trial design
Science of Small Clinical Trials
We need to think creatively about controls. Although there are ongoing trials, not every child with NCL is enrolled. We have an opportunity now to prospectively gather data for future use. Historical controls present their own challenges, but a registry of specific, prospective data may introduce efficiency into future studies.
We have the opportunity to employ alternative trial designs.
Existing resources for strengthening trial design include the Office of Orphan Products Development Science of Small Clinical Trials annual course. Direct interactions with the FDA itself regarding study design and outcomes, and the NINDS Network for Excellence in Neuroscience Clinical Trials where protocol working groups are formed for each study being considered for execution through the network. There is a breakfast talk on NeuroNEXT later this week during CNS.

12. Opportunity: Alternative designs
Parallel group
Cross-over
Factorial
Three-stage
Historical controls
Randomized withdrawal
n-of-1
Group sequential
Selection studies
Dose-response
Adaptive
We have the opportunity to think beyond traditional parallel group or open-label studies to add design features that allow us to potentially increase knowledge gain from a single study, or introduce efficiency in sample size.

13. Trial Execution: Outcomes
CHALLENGES
Well-defined, reliable outcome measures and endpoints
Clinical Measures
Biomarkers
Patient-Centered Outcomes
Natural History
In rare diseases, commonly, validated measures of disease activity or disease progression are lacking. We have some clinical measures, but these need refinement, and there is a role for biomarker development which can help support proof-of-concept or biological plausibility of interventions in early phase studies. This is especially needed in JNCL where enzyme assay is not relevant and the course of disease is slow. We may not yet fully understand what is important to our patients, and we are still learning about natural history.

14. Trial Execution: Outcomes
CHALLENGES
OPPORTUNITIES
Well-defined, reliable outcome measures and endpoints
Clinical Measures
Biomarkers
Patient-Centered Outcomes
Natural History
Rating Scale refinement
Patient-engagement
Biomarker development
Consensus – preclinical and clinical outcomes
Outcomes conference
There are opportunities to refine our clinical rating scales to better quantify these multifaceted diseases, to increase their precision related to small changes, and to address ceiling and floor effects. We have an opportunity to engage families about outcomes meaningful to them. This is the time to establish consensus regarding common use of outcomes, both preclinical and clinical. Heather Adams from the University of Rochester is in the process of organizing a meeting to do just that.

15. Trial Execution: Recruitment, Ethics
CHALLENGES
Participant Accrual in a timely manner
Early, Accurate Diagnosis
Geography
Controls, Placebo
Therapeutic Misconception
Consent
Risk/benefit
To recruit a sufficient sample, and to do so in a timely manner is a challenge due to the small number of affected, eligible individuals. Diagnosis requires recognition, and a patient with NCL may come along once in a career for most neurologists, so there may be delays in diagnosis, and thus more advanced stage when considering trials. Geographic dispersion requires multi-center or even multi-national collaboration. The idea of placebo may be unacceptable to some families. There is likely therapeutic misconception, which is exaggerated in the context of a fatal neurodegenerative disorder where there is a strong desire to participate in treatment trials.

16. Trial Execution: Recruitment, Ethics
CHALLENGES
OPPORTUNITIES
Participant Accrual in a timely manner
Early, Accurate Diagnosis
Geography
Controls, Placebo
Therapeutic Misconception
Consent
Risk/benefit
Patient Organization
Registry development
Education - NDC Symposium
Telemedicine
Networks – NORD, LDN, RDCRN, Centers of Excellence
Education
Again, we have opportunities to partner with patient organizations, to develop registries of individuals interested in clinical trial participation (ideally with natural history data), an opportunity to try to decrease time to diagnosis – symposia like today). Opportunities to take advantage of technologies like telemedicine in our assessments. And to make use of existing networks. I listed twice, the first time for clinicians and investigators, the second for families.

17. Trial Execution: Funding, Resource Allocation
Small populations = small $ = small interest
Orphan Drug Act 1983
Orphan Product Grant Program
Academic-Industry Partnership
Funding mandates for rare disease
Consensus
How do we fund studies. Therapeutic development is expensive, and in this economic climate we are all aware that there is only so much to go around. Traditional thinking is that common diseases have greater economic impact – rare disease represents a very small market, and thus a small return on investment.
The orphan drug act developed incentives to stimulate drug development for rare disease. These incentives include protocol assistance from the FDA, tax credits, and 7 years of market exclusivity.
This is a time to develop consensus regarding priorities as new therapies will enter the pipeline, competing to recruit the same population.
The ODA is considered to be highly successful: in the 1970s there were only 10 approved drugs for rare diseases but since 1983 more than 2,500 small molecules and biologics have been designated as orphan drugs, with more than 390 achieving marketing approval (FIG. 1). Currently, there are 460 medicines in clinical trials for a variety of rare diseases (FIG. 2). Rare oncology indications represent the largest segment and account for over 30% of the pipeline and all orphan drug approvals

18. The orphan drug act has been successful in increasing interest from pharma. Since 1983, there has been a steady increase in orphan drug designations in red. Approvals in blue have been relatively stable over time in terms of absolute numbers, but represent an increasing proportion of annual FDA approvals. Clearly, interest is on the rise.
Figure 1 | Orphan drug designations and unique orphan drug approvals (1984–2011). In the United States, the number of orphan drug designations has doubled during the past decade from an average of 63 per year in the 1990s to 126 per year during 2001–2010. This can be attributed to the growing interest by biopharmaceutical companies in developing products to treat rare diseases. The number of orphan drug approvals remained relatively constant between 1984 and 2010, with an average of 10 per year. However, as the number of total US Food and Drug Administration (FDA) approvals has shrunk from its peak in the 1990s, the proportion of orphan drugs receiving marketing authorization has increased from 17% in the 1990s to over 35% during 2008–2010. This trend is expected to continue in the future. Source: FDA application.
Melnikova, et al. Nature Reviews Drug Discovery 11, (April 2012)

19. Trial Execution: Clinical Trialists
CHALLENGES
Dwindling number of clinician scientists
A final challenge to discuss today is a dwindling number of clinician scientists to conduct clinical trials.
Statland, et al. Neurology 2012; 79: e

20. Trial Execution: Clinical Trialists
CHALLENGES
OPPORTUNITIES
Dwindling number of clinician scientists
Experimental Therapeutics Fellowships
NINDS Clinical Trial Methods Course
ASENT Training for Neurotherapeutics Discovery
NeuroNEXT
ASENT – American Society for Experimental NeuroTherapeutics
Statland, et al. Neurology 2012; 79: e

21. INCL – HuCNS-SC Design: open-label, single group Sample size = 3
Recruitment: siblings with INCL
Outcomes: developmental assessment, EEG, ERG, VEPs, Tc-HMPAO SPECT, MRI, PPT1 enzyme activity
Rectal biopsy for storage assessment
PPT1 enzyme activity in leukocytes and CSF
Hypothesis: correction of deficient PPT activity
Lonnqvist, et al. Hematopoietic stem cell transplantation in infantile neuronal ceroid lipofuscinosis. Neurology 2001;57:1411–1416

22. INCL - Cystagon Design: open-label, single group Sample size = 10
Inclusion restriction: patients with specific PPT1 mutations; restricted age range
Centrally conducted – NIH
Outcomes: MRI, EEG, ERG, VEP, skin biopsy
Inclusion criterion later expanded

23. LINCL – AAV2CUhCLN2 Design: open-label case series, sequential
Sample size = 11 - Group A severe (5), Group B moderate (6)
Outcomes: safety and efficacy - LINCL clinical rating scale, MRI/MRS
Gene transfer study. Subsequent studies involve a new delivery system
This study
Dose questions – initial dose, # of sites, repeated administrations?
High risk, rare disease – need to combine phases 1-3
Want to understand efficacy of gene transfer. Clinical outcome measures are important, but may not be the most critical take away piece of information.
What will the inclusion/exclusion be – advanced patients who may be at higher complication risk and less benefit?

24. INCL/LINCL – HuCNS-SC
Preclinical data: ↑ PPT and TPP1 production in (PPT1 deficient) mice following HuCNS-SC, ↓ storage, ↓ neuronal death
Study design: open-label, single group, phase I – dose escalation study. First investigational use of purified human neural stem cells in clinical testing
Goals: evaluate safety of surgical technique, immunosuppressive regimen, cells
Method: delivered directly to cerebral hemispheres and lateral ventricles
Sample size: N = 6 (moderately to severely affected children)
1o outcome: safety – no unexpected AEs, no definitely attrib AEs
2o outcome: preliminary efficacy – one died 11 mos post transplant with evidence of CNS engraftment
Follow-up: 13months
Limitations: open label, small cohort, advanced disease state at transplantation – limits interpretations of efficacy – first-in-human type trial

25. JNCL - Mycophenolate Design: Phase II crossover study Sample size = 30
Recruitment: BDSRA, contact registry
Measures: UBDRS
1o outcome: tolerability
2o outcomes: safety, preliminary efficacy

26. Heather Adams, PhD Jonathan Mink, MD, PhD Frederick Marshall, MD
Amy Vierhile, NP
Elisabeth de Blieck, MPA
Jennifer Kwon, MD, MPH
Paul Rothberg, PhD
Alyssa Thatcher
Sara Defendorf
Chris Beck, PhD
Laurie Seltzer, MD
Jennifer Cialone, MD