Name: Norman Relkin
Position: Director, Memory Disorders Program, New York-Presbyterian Hospital/Weill Cornell Medical Center, 1993 to present; associate professor of clinical neurology and neuroscience, Weill Cornell Medical College, 1997 to present.
Background: PhD in neuroscience, Yeshiva University, 1987; MD, Albert Einstein College of Medicine; 1987; internship, internal medicine, New York Hospital-Cornell Medical College, 1987 to 1988; fellow, neurophysiology and behavioral neurology, New York Hospital-Cornell Medical College, 1991 to 1992; assistant professor, neurology and neuroscience, New York Hospital-Cornell Medical College, 1992 to 1997.
Last month, the Leon Levy Foundation awarded New York-Presbyterian Hospital/Weill Cornell Medical Center in New York $1.4 million to study two neurological disorders, including $700,000 to use proteomics tools to better understand normal pressure hydrocephalus, a buildup of fluid in the brain.
Because the symptoms of NPH are very similar to other neurodegenerative ailments such as Alzheimer’s disease, it is suspected that the disease is under-diagnosed.
There is no current well-defined diagnostic, and conventional methods have an accuracy of about 50 percent, according to Norman Relkin, director of the Memory Disorders Program at New York-Presbyterian Hospital/Weill Cornell Medical Center, and the leader of the three-year study into the disease.
ProteoMonitor spoke with Relkin this week about the work he will be conducting into NPH. Below is an edited version of the conversation.
What are you looking for? Protein biomarkers?
That’s more or less the nature of [our work]. Normal pressure hydrocephalus is a disorder that predominantly affects the elderly, and it causes progressive disability that renders people unable to walk, [they’re] demented, and usually they lose control of their bladders, as well, as a result of the disease.
Its prevalence isn’t known but it’s thought to be maybe 1 or 2 percent of all cases of dementia. And what makes it very noteworthy is that it’s a treatable and reversible disorder, so if it’s correctly identified, one can treat it with a neurosurgical procedure and essentially cure the person. And then they regain their intellectual function and the ability to walk, and so forth.
[It’s] a very dramatic and positive thing. The challenge, however, is identifying people who have it because its symptoms overlap diseases like Alzheimer’s, Parkinson’s, and a host of other conditions.
What are the mechanisms of the disease?
Hydrocephalus means water in the brain, and is actually a number of different diseases. This is an adult form of hydrocephalus. The one that many people are familiar with occurs in children, and it’s usually secondary in children to a blockage in the pathways that drain the spinal fluid.
In adults, we don’t know what causes it; it occurs idiopathically and it manifests from a physical standpoint with the enlargement of the ventricles, the fluid spaces in the brain. [It’s] a usually slow and insidious progression of symptoms over the space of months to years.
How is it diagnosed now?
It’s a combination of clinical findings and radiologic findings, so typically a CT scan or an MRI is required to verify that the ventricles have enlarged due to hydrocephalus. Here is one of the sources of error; the ventricles get bigger, [which also is symptomatic of] diseases like Alzheimer’s, during which the brain shrinks. But that’s not the same illness, so there are certainly misdiagnoses on the basis of radiographic findings.
Likewise, clinicians have been sort of trained to look for the combination of dementia, ataxia, inability to walk, and incontinence. But very often, [NPH] only manifests as one or two of those symptoms, and it may not be as full blown as to make it obvious what [a patient is] dealing with. So, it takes an astute clinician thinking of this disease, sending the person for a brain-imaging study, and then putting it all together.
Have you already started on developing a proteomics-based diagnostic for NPH?
I was involved in writing an international consensus guideline on the diagnosis and treatment of this disease in the early 2000s, and we worked on improving the clinical detection, and this is one of those efforts that we’re making to turn this into more of a science than an art.
I published with my collaborator, Kelvin Lee [now at the University of Delaware], … last year a set of 23 proteomic markers for Alzheimer’s disease [See PM 12/26/06], and basically in the course of that we collected a large number of neurologic controls that included a set of [cerebrospinal fluid samples] from patients with hydrocephalus.
So we actually have enough of those CSFs already in our bank, I think we have about 20, from well-characterized cases to do the exploratory component of the study, and that’s getting underway in February. We’re going to use the same sort of approaches that we used in the Alzheimer’s study and attempt to identify a set of markers that differentiate NPH from AD and other neurologic illnesses that are common in that age group.
And then we’re going to do a prospective validation study over the course of the next two to three years with newly acquired clinical cases and look at how well we can do in terms of classifying our cases on the basis of the markers that come out of the initial study.
Having done proteomics studies now with Alzheimer’s disease, from the standpoint of clinical use for these kinds of tests, clinicians don’t normally do spinal fluid studies in Alzheimer’s disease, so there’s a reluctance to do a spinal tap for what is largely deemed experimental purposes if there are no established biomarkers.
But there’s a big difference with NPH because lumbar punctures or spinal taps are part of the routine evaluation of patients with NPH, so when you want to figure out if someone has NPH, you do the MRI and you do the clinical evaluation.
Now, you want to know [if they are] going to respond to this neurological procedure, which is a placement of a shunt. And one way you can do that is to perform a spinal tap and take off 50 ccs of spinal fluid and see if they transiently get better. That’s kind of a miniature version of what the shunt would do on a continuous basis.
That component of the evaluation means that just about everybody who is suspected of having hydrocephalus ends up getting a spinal tap anyway, so it’s a natural, from that standpoint, for CSF proteomics.
The other thing is NPH is a disturbance of CSF dynamics to the extent that the protein pool in spinal fluid is in part a function of what the brain produces and what the meninges produce, its flow, and the passage of proteins across the blood-brain barrier. Alterations in CSF dynamics would be expected to change multiple proteins in CSF in terms of the level of expression, and that provides, I think, the rationale for applying proteomics to this particular disease.
In studies of other neurodegenerative diseases, the challenge seems to be trying to find markers that can differentiate one neurodegenerative disease from another, as much as finding any marker.
That’s a good point. In fact, how you do these studies really does determine what you get. If you go in and you say, ‘I’m going to find markers that associate with … Alzheimer’s disease,’ what you choose as your controls determines what you see. And likewise where in the disease the person may be, what stage they’re in, the demographics of the population, all those things will influence the markers that you get.
What was fairly unique about our [Alzheimer’s] study was that we did the exploratory study with autopsy-proven cases, so we had a high degree of certainty that the diagnosis was accurate. And we used not only age-matched normal controls, but age-matched neurologic controls. There, what we were doing is trying to identify multiplexed panels of markers that distinguished markers from everything else.
When I look at some of the other published studies, very frequently what people do is they’ll choose one disease like Parkinson’s disease and they’ll say ‘What’s different about Alzheimer’s from that disorder?’ Or they’ll just do age-matched normals.
For [our] study we were thinking more in terms of differential diagnosis, and here, I think, the same is true. Here, we know that there are a number of disorders that overlap NPH. What we’d like to be able to do is find a set of markers that really tells us this patient, they could have Alzheimer’s as well, but the markers point to them having hydrocephalus.
How are you going to do that? Alzheimer’s can be diagnosed through autopsy. Do you have similar control with NPH?
That’s a critical kind of thing. There’s no gold standard of autopsy in hydrocephalus, so hydrocephalus at this point is purely a clinical, radiographic diagnosis. But that’s the reason we felt this was a very good study for us to do. As we discussed earlier on, I wrote the diagnostic criteria, so I’m very familiar with the process of careful characterization of the cases.
We have a National Institutes of Mental Health grant for characterizing response to treatment in NPH, so patients are getting an exhaustive research-grade battery of tests — neuropsychological testing, advanced imaging techniques, etc. These are extremely well-characterized cases, and they’re also in the great majority of cases, followed forward longitudinally, so we can get a sense of either the natural history if they go untreated, or the response to surgery if they do get treated.
So these are clinical gold standards. They’re not autopsy gold standards, but they’re the best we can do. I have a program in my center where people come for specialty evaluation of suspected NPH. I can’t possibly do everybody in the country, but if there’s a biomarker or a combination of biomarker/brain imaging test that give the same level of certainty as I can get in my center, that would be of great value to patients and clinicians everywhere.
Is there any commercial interest or support for this work?
At the current time, no. But, again, that’s because it’s at a very nascent stage right now. With our 23-marker test [for Alzheimer’s], there has been some commercial interest. I think that these types of tests, once defined by academic groups like mine, have to go through a period of development where they’re translated into something which is commercially viable.