Interview: Roger Barker

Roger Barker says that he was inspired to become a clinician researcher after meeting the great neuroscientist David Marsden. As Barker puts it, “the minute I met David Marsden I wanted to be David Marsden, a person equally at home in the clinic and the laboratory”. When I visited the 51-year-old Parkinson’s and Huntington’s researcher during a recent visit to the UK, he seemed to be well on his way to accomplishing that goal.

Barker appears to speak with a faux Australian accent. He is in fact solidly British. After getting a First at Oxford, he attended St Thomas’s Hospital, did a PhD in Cambridge (on neural grafting), received neurological training at Queens Square, before settling at The John van Geest Centre for Brain Repair in Cambridge. The CBR is tucked away in a corner of the Addenbrooke's Hospital campus. Inside the modern two-storey building, posters line the walls announcing scientific speakers from elite institutions like Harvard and MIT. Bright young researchers from all over world congregate in the coffee lounge to socialize in Italian, Japanese, Portuguese and French. This is what a hot center of neuroscience looks like. This is Roger Barker’s world.

Barker is a provocative figure. A leading player in the TRANSEURO neural grafting project, he is known for his critiques of current clinical research methods, in particular the hegemony of placebo-controlled trials. He had just co-authored a commentary in the Journal of Clinical Investigation about the “repurposing” of an injectable diabetes drug called Exenatide for Parkinson’s disease. Its efficacy was studied using an open label trial design.

Jon Palfreman: Roger, I know you’re a fan of repurposing FDA approved drugs to see if they help Parkinson’s patients. But what did we learn about Exenatide? The authors report that in a 12-month follow up (with a two month washout), the treatment group reduced (i.e. improved) its UPDRS motor score by 2.7 points while the control group’s score increased (i.e. got worse) by 2.2 points. Given it’s not a placebo controlled study, what makes you think the 4.9 spread might be important?

Roger Barker: I think it's interesting and encouraging (but not necessarily definitive) for two reasons. First, the clinical exams were videotaped and rated blind by neurologists not involved in the trial. Second, the treatment group continued to improve compared to the control group after the two-month washout period… after the injections were discontinued. If there had been a substantial placebo effect it would most likely have waned during the two-month washout period, when the patients knew that they no longer were getting active treatment.

Jon Palfreman: Why not use a placebo control?

Roger Barker: Cost. If you need an injectable device for a placebo control, someone has to make it for you. The drug company isn't going to make the injectors for you, and to get someone else to do it costs between half a million and one million pounds. So you’re stuck, do you try this or not? Sometimes it's not possible to do what you want to do.

Jon Palfreman: Placebo-controlled trials are viewed as the gold standard of clinical trials, but you think they can sometimes be counter productive, don’t you?

Roger Barker: It's not that you shouldn't ever do double blind placebo-controlled trials, but if you do them too early in the development of a therapy people think that you will get a definite answer. And if you haven't optimized the treatment, you're going to get a result that is hard to explain.

Jon Palfreman: And your point is that if we haven’t sorted out the dose, frequency, duration etc we might prematurely reject a therapy that could work?

Roger Barker: Absolutely. If I took 100 people with a chest infection, and I gave them 100 mg of ampicillin, once a day, for 3 days, and then looked at the patients two days later, what would I find? Probably, the majority of them will still have a chest infection. So the conclusion would be that ampicillin, which is an antibiotic, is not useful for the treatment of infection — no more helpful than a placebo. And of course, the problem is that I've used the wrong dose, for the wrong amount of time, in a mixed group of patients, and the outcome led to a wrong conclusion.

Jon Palfreman: So the fact that a trial fails doesn’t mean the therapy is necessarily a hopeless failure?

Roger Barker: That’s right. The history of experimental therapeutics shows that new therapies rarely work first time. Recall what happened with heart transplants — people died only a few days after receiving the transplant. Based on our modern criteria for clinical testing, we would say that it's not working, let's just forget about it. But look what happened. Scientists studied the reason for failure — the rejection process — and once they mastered what was going on heart transplants slowly became routine therapy. It was an iterative process. We learnt as we went along.

Jon Palfreman: This is true for Parkinson’s too. In the first controlled double blind trial of the drug in 1966, for example, Swedish neurologist Claus Fehling found that levodopa not only had no effect on Parkinson’s symptoms, it also caused high blood pressure and nausea in a third of patients.

Roger Barker: Right. It was only through the persistent efforts of researchers like George Cotzias, that scientists learned that much larger doses — thousands of times bigger — were required. It’s a powerful example. Imagine how different things would be without L-dopa.

Jon Palfreman: On the other hand, we can be fooled into thinking something is real when it isn’t, can’t we? Open label studies of neural grafting appeared to work but placebo-controlled trials couldn’t demonstrate efficacy. Open label studies of neurotrophic factors appeared to work. But a whole series of placebo-controlled trials have failed.

Roger Barker: I think the placebo effect is a vastly exaggerated phenomenon. I’m not saying that it's not real… but that randomized controlled trials to get round the placebo effect have made it difficult to prove new therapies actually work.

Jon Palfreman: So what’s your view today of neural grafting as a therapy?

Roger Barker: I have no doubt that it works in some case — several patients live off all dopaminergic medication more than 15 years after surgery — the issue is can we get it to work more consistently and is there another source of tissue apart from fetal tissue? In a recent article in the Lancet (Lancet Neurology), Anders Björklund and I explored likely reasons for the inconsistency of results. They include patient selection (i.e. age, type, stage), graft preparation, graft placement, immunotherapy, inappropriate outcome measures, insufficient follow up periods etc. There’s so much that you can get wrong.

Jon Palfreman: In the TRANSEURO project you will not use placebo-controlled trials like they did in the 2001 Freed trial. Many US neuroscientists are critical of this decision.

Roger Barker: I know. We have selected 150 people with early-stage Parkinson's disease from which we will select randomly 20 cases for fetal dopamine grafting. The patients are not blinded (and may be susceptible to a placebo effect) but they’ll be videotaped so independent movement disorder specialists can rate them blindly. And they’ll be followed up for years. Placebo effects don’t last for years.

Jon Palfreman: Is the story of neurotrophic factors another example where the failed trials don’t necessarily imply it’s a failed therapy? This is the appealing theory that growth factors (e.g. glial-derived neurotrophic factor (GDNF), brain-derived neurotrophic factor (BDNF) and neurturin (NRTN)) might rescue weakened neurons and halt Parkinson’s disease.

Roger Barker: It’s a classic example where we may be prematurely closing the book on a potentially useful therapy.

Jon Palfreman: But look at the litany of failures. I count two failed Amgen direct GDNF infusion trials, two failed Ceregene NTN gene therapy trials, and the failure of Cogane, an oral medicine designed to stimulate production of GDNF and BDNF. That’s a lot of failure.

Roger Barker: True. But you have to explain why the open label studies worked for Steve Gill (the British neurosurgeon) and for the Kentucky group, and why in postmortem cases they found neuronal sprouting. And you have to explain why in imaging studies they found dopamine uptake. And you have to explain why there were clinical benefits in some patients. And all this despite the fact that some patients had advanced stage PD and in some cases the growth factor delivery was far from optimal. So that's hard to just dismiss all this as irrelevant really.

Jon Palfreman: I accept that it’s possible that when they optimize the therapy — say, by delivering to the right location in the right dose etc. — that the therapy could prove efficacy even in a placebo-controlled trial. But there’s another possible explanation for the failure. It may be that the damage to the study subjects’ brains is too advanced, that it’s too late to make a difference. Pathologists like Jeff Kordower and Tom Beach report that the damage to nigral axons is devastating even 5 years into the disease.

Roger Barker: A fair point. Currently, there are two competing theories of what’s going on. There’s Jeff Kordower’s finding that most tyrosine hydroxylase (TH) fibers in the striatum have disappeared after 5 years. Tyrosine hydroxylase is the enzyme that converts tyrosine into L-dopa. So that implies that there may be nothing left to save with neurotrophic factors. That’s one theory. The second theory is that put forward by Anders Björklund who had a paper in 2012 in Science Translational Medicine. He showed in an animal model that alpha synuclein switched off the GDNF receptor-signaling pathway. The pathways could be restored, in theory, with nuclear receptor related 1 protein or NURR 1. This indicates that if you could just put the receptors back online then GDNF could work.  At the moment, we don’t know whether either or both of these theories is the true story.

Jon Palfreman: Do we keep trying with neurotrophic factors?

Roger Barker: I think we keep going, but the important thing is to go back and learn from what you've done. It may be that the damage is too great and we need to do de novo newly diagnosed patients. I think that one trial that’s doing it right is the ProSavin study.

Jon Palfreman: In the ProSavin story, a British firm, Oxford BioMedica, has taken a gutted virus, inserted three genes, and is delivering multiples copies of it to the striatum. The three genes together (tyrosine hydroxylase (the enzyme that converts tyrosine into L-dopa), AADC (the enzyme that turns L-dopa into dopamine) and GTP-cyclohydylase 1) can collectively manufacture dopamine.

Roger Barker: That’s right. This gene therapy improves the efficiency with which you make dopamine — either your own dopamine, or the levodopa dopamine you take by mouth. ProSavin, is involved in a Phase One open label study under the direction of Professor Stéphane Palfi of the Henri Mondor Hospital in Paris. Here in Cambridge we have three of the trial patients.

Jon Palfreman: So what’s good about this study?

Roger Barker: In this case they didn't know how to do it optimally at first, so they started the first patients on low doses to see what would happen and then increased the dose with subsequent patients, observing when effects kicked in.

Jon Palfreman: I understand one of your patients has done pretty well.

Roger Barker: Yes, she’s been on Sky News. Before, when she was off she couldn't do anything… she was stuck in her chair, she couldn't talk, she couldn't move. Today, thanks to ProSavin she has no off periods at all, more consistent control of her PD with lower doses of drugs that makes her thinking much clearer as well. But she still has some dyskinesias.

Jon Palfreman: But this will need proving in a controlled study, surely?

Roger Barker: At some point they will need to do a larger controlled trial but then rather than matching against a placebo they might match it up against duodopa, which sends in a continuous 24-hour smooth dose of L-dopa through the intestine.

Jon Palfreman: Another interest of yours is the identification of clinical subtypes. This is important both medically but also in relation to running clinical trials.

Roger Barker: I have been interested in the early identification of Parkinson's disease dementia patients. If you're over 72, can't produce a list of 20 animals in under 90 seconds, and can’t draw interlocking pentagons, the facts are pretty clear: your chance of dementia after 5 years is 88 fold greater than younger people who can do these tasks. Such rapidly dementing cases are going to do badly. They will have Lewy bodies everywhere. Clinical studies looking at dopamine therapies need to avoid this group, but they may be ideal for disease modifying therapy trials.

Jon Palfreman: You like to say, that Parkinson’s disease is one disease but it runs with different kinetics. What can affect the kinetics?

Roger Barker: Age, environment, genes… and more. Genes are very interesting. In some studies we and others have done, not only differences in tau haplotypes, but also mutations in the Gaucher’s disease seem to indicate worse progression. The patient who inherits a Gaucher’s mutation doesn't look any different from your regular PD patient at presentation, but every one of them is demented within 8 years in our study. You don’t want such cases in experimental trials for neural grafts or neurotrophic factors.

Jon Palfreman: The tau haplotype is predictive of whether you get Parkinson's disease, yet tau pathology isn't a feature of Parkinson's disease (PD is a synucleinopathy), so how is it relevant?

Roger Barker: Well it's unclear but it may be changing something to do with the axonal transport. So for example, it's like taking a three-lane motorway and turning it into a two-lane motorway. And for the vast majority of life that's fine. But as you get a little bit older, and the motorway starts to crumble and someone has to do a bit of roadwork here and there, the system doesn't do as well. And so, once the disease starts, everything starts to go wrong much more quickly.

Jon Palfreman: And the Gaucher’s mutation?

Roger Barker: The Gaucher’s mutation must be affecting some garbage-clearing pathway (for alpha synuclein aggregates). So this mutation perhaps make garbage trucks come every two weeks instead of every week. Okay initially, that's fine, but as other thing start to go wrong, the system reaches a tipping point. And then all of the factors drive the disease process.

Jon Palfreman: What does this suggest for new treatments?

Roger Barker: Well it looks good in some ways. If you can simply get one of these players to work more efficiently, it would probably have a profound effect. You probably don't have to return everything to normal, improving things just by a little bit maybe all you need.

Jon Palfreman: What role does chance play in scientific discovery?

Roger Barker: I'm a huge believer in serendipity. The real sign of a great scientist is that he/she notices something isn't right and then decides it is worth pursuing. Most of us would say “oh my I don't know what went wrong, but let's forget it and get on with what we were planning to do”. A great scientist says instead, “there may be something in that… it may be important”. That’s what Bill Langston did noticing the frozen addicts (which led to MPTP), what Larry Golbe did discovering the Contursi kindred (which led to alpha synuclein) and what Alim Louis Benabid did with deep brain stimulation.