25 March 2011. It’s the rare occasion that scientists get unvarnished advice from their colleagues at the drug approval authorities at public conferences. The thinking that goes on at these underfunded, much-maligned, yet all-important government agencies can seem opaque to most working scientists, especially those toiling at the benches. So when the 10th International Conference on Alzheimer’s and Parkinson’s Diseases (AD/PD), held 9-13 March 2011 in Barcelona, put up Cristina Sampaio of the European Medicines Agency (EMA) to share her perspective on where clinical trials in Parkinson’s disease stand at the moment, your roving reporter whipped out her pen and paper. Sampaio is a Parkinson’s expert at the University of Lisbon who brings a penchant for frank language to her service on the EMA’s Committee for Medicinal Products for Human Use (CHMP). Previously, Alzforum covered Sampaio’s take on Alzheimer’s disease trial failures (see ARF Geneva story), and on planned preclinical trials in genetic populations (see ARF London DIAN story; ARF DC API story). Curious about the similarities and differences between AD and PD? Below are extended excerpts of Sampaio’s talk, followed by excerpts of a subsequent panel discussion.—Gabrielle Strobel.
“I take an optimistic view on the prospects for clinical trials in PD,” Sampaio began. “In this disease we have plenty of effective treatments. We have drugs and surgery, which are highly effective and can be used with wisdom to improve our patients’ quality of life. [Editors note: DBS; see ARF series.] The challenge in PD is to find a further drug that makes a difference. That is much different from AD, where the drugs available have made little impact on the disease so far.
“In PD, people want to find a disease-modifying (DM) drug that will change the course of the disease. In that sense, the prospects are not different from AD. But the background of already having effective drugs changes the way we look to disease modification in PD.
“Unlike in AD, the pipeline for PD is commonly said to be empty. Few drugs are coming out, and the ones that do are not particularly novel. That is true in the sense that most drugs in the last phases of development are targeting symptomatic treatment, and most drugs that you can expect to see approved in 2012, 2013 are new formulations of well-known drugs.
“If you read about early-stage compounds, you will count 181 in the pipeline across all stages. Most of these are in very early stages of development. Most also appear in the pipeline of many neurodegenerative disorders, so are not specific for PD. For example, glutamatergic compounds are in the pipeline for AD, PD, and Huntington’s disease; sirtuin antagonists are being tried for several indications. Companies are trying a broad approach to certain compounds, and so this 181 number of compounds that can be potentially developed for PD is somewhat artificial. The fact is, in the next few years we will see mostly symptomatic treatments.
“The hottest issue in the regulatory field has been trying to obtain a claim in disease modification. One problem is that if someone develops a drug that might have DM effects but has also symptomatic effects, how do we disentangle these two effects and obtain a claim that the regulatory authorities can recognize as disease modifying? Consider, for example, the ADAGIO study of rasagiline. It has been famously debated in the literature. It was conducted under orders of the FDA, and the company followed strictly what the FDA told them to do. The EMA never told the company that a claim for DM could be obtained in this way. We always said: To prove DM, you need to prove that the drug interferes with the pathogenesis of the disease. This was difficult to do just by trial design; you need biomarkers for that. So this was mostly a battle on the U.S. side of the Atlantic. The issue ended up being settled because the higher dose did not reach the primary outcome, so the DM claim was not granted. [Editor’s note: for context, see Sampaio and Ferreira, 2010; Olanow et al., 2009; Pagonabarraga and Kulisevsky, 2010; discussion thread on PD Online Research.]
“I am nevertheless optimistic about future prospects for disease modification, because an understanding of disease pathophysiology is now being revealed. The advances in genetics are very important for that. We are starting from scratch, with α-synuclein and the GWAS results. The field needs time to mature and produce the right targets and drugs.
“The most important unmet medical need in this field is targeting the non-motor features of PD; the depression, the psychosis, the cognitive deficits. Everyone now recognizes this. The number of scientific papers on this is growing, but at the same time, only 44 randomized controlled trials were published addressing different non-motor indications in PD populations. Unfortunately, in the last decade, the trials were of poor methodological quality, and there are still problems in defining the populations. People know how to diagnose PD, but they are unsure how to define the non-motor problems. A further hindrance is that most of the drugs that have been tried are off patent.
“Gene therapy for PD is a very active field. [Editor’s note: see ARF related news story for latest Phase 2 results.] A number of gene therapy trials are ongoing with relatively good success, at least in early phases. So there are at least different approaches, which in most cases try to replace the production of dopamine or deliver growth factors.
“There is a serious hint that there are clinical subtypes of PD. A recent paper clearly demonstrates them (Van Rooden et al., 2011). This confirms what many people have said in many papers, and what we knew from empirical clinical observation. These subtypes must be identified and trials should then be made type-specific.
“The PD field is behind AD because we cannot detect the pre-motor stage. We have a good research concept for it based on enriched populations, but we cannot apply it in clinical trials yet. We have to improve pre-motor detection and biomarker use before we can do better clinical trials. The preclinical window of opportunity is critical. We need a pragmatic way to detect the pre-motor stage.
“In closing, it is urgent that there be an immediate methodological improvement in trials targeting non-motor aspects of PD. We need more trials in PD. And success in disease modification will come as you advance translational science and increase knowledge about disease subtypes.”
Jaime Kulisevsky, Hospital de la Santa Creu i Sant Pau, Barcelona: Representing the PD field and non-motor symptoms of PD here, I agree with everything Cristina said. In PD, we are entering a change of paradigm. All past effort has been put on the motor side. Our success there, with good drugs and surgery, has only reinforced our recognition that the disease is more than a motor problem. The cognitive problem emerges as the main unsolved issue. The more drugs we have to treat the motor symptoms, the more it becomes evident that the patients are becoming demented. In PD, I see cognition being the problem of the next few years. We are still defining what is the cognitive deterioration in PD, and this is crucial for good trials. We must share the effort of the AD side to make this happen.
For example, one of the main markers of dementia in PD can be certain haplotypes of the tau gene. [Editor’s note: see MAPT holding place 2 on PDGene Top Results]. We know that, but from a trials point of view, we are still wrestling with defining the cognitive deterioration in PD. In these areas, we can collaborate with our AD colleagues.
Raphael Blesa, Hospital de la Santa Creu i Sant Pau, Barcelona: As a clinician, a problem I have is to translate new diagnostic research criteria into normal clinical practice. So far, we have been relying on the opinion of the families/caregivers. Now, we are moving to base the diagnosis in biomarkers. That is interesting because research specialists will implement changes in the clinic in order to learn how to make the diagnosis of prodromal AD work. But for now, 99 percent of the normal doctors will keep depending on the opinion of relatives in order to establish dementia. That transition is a challenge.
Jean Marc Orgogozo, Université Victor Segalen, Bordeaux, France: I believe the problem is the drugs. Our tools are not that bad. We can distinguish disease-modifying from symptomatic treatments. For example, in migraine we have drugs for both, and we can distinguish their effects quite clearly. That is also true for depression. We can separate the two, even if it is intrinsically difficult to do that for a drug that has both effects, for example, selegiline. Even if we build long-term trials extremely well, they will fail if the drugs do not work.
Bruno Dubois, Salpetrière Hospital, Paris: Clearly, we need better drugs. But we also need to select patients better. In the 1990s, the problem was having the MCI group and trying to determine among those who has AD. We can do that now. Please use the test that can help you identify the specific AD pattern. In this sense, I disagree with the paper by Lon Schneider (Schneider et al., 2010). Even if the biomarkers did not have so much diagnostic accuracy in the ADNI cohort, they have very high negative predictive value. At least for anti-amyloid drug trials, we need to be 100 percent sure that the patients do have AD. The biomarkers can give us that certainty. It is very important to have that homogeneous population.
Christoph Hock, University of Zurich, Neurimmune, Inc., Switzerland: This is a time of great opportunity. We have achieved removing amyloid from the brains of living people. That is a tool we can use now. We can stratify the population we want to treat with anti-amyloid compounds. These are two major achievements, and we should acknowledge them.
There are problems, also. The timing: Can we go in early enough before synapses have died, and how do we do that? The mixed pathologies: amyloid, tau, α-synuclein. Lowering Aβ is not sufficient for all these cases, but it is a start. The right target: What form of amyloid should we remove? That is what academia and industry are working on.
Khalid Iqbal, New York State Institute for Basic Research In Developmental Disabilities, Staten Island: The prodromal phase is very long. Are you thinking of giving preventive drugs to children?
Hock: We have not resolved the timing issue yet.
Iqbal: Another problem I see is that we treat a heterogeneous group of patients with one drug.
Hock: We have to improve diagnostic stratification. We have to measure the tau load, the α-synuclein load, and the Aβ load in patients, and then treat accordingly the pathology we know they do have.
Reisa Sperling, Brigham and Women’s Hospital, Boston: I have been very struck at this conference how the challenges in PD mirror those in AD. In both diseases, we have the imaging agents to detect early. But if we treat early, how can we have outcome measures to know if the treatment works? How do we link the biomarker to the behavior without waiting 10 years? That is the greatest challenge.
Sampaio: That is exactly what we have to learn.