-
MR SPECTROSCOPY UNMASKS SURPRISES IN NF-1
Martha Bridge Denckla, M.D.
Kennedy Krieger Institute
Baltimore, Maryland
-
Introduction: We have known Dr. Denckla for a very long time, dating back to the time she was still here at NIH in the Neurological Disorders and Strokes Institute. Many of our children, when they were young, participated in a pilot study on cognitive differences in children with NF and their unaffected siblings. That was before there even was a NF-1 and a NF-2 distinction. She has spoken to our group many, many times on her longitudinal study of NF and cognitive deficits. Currently she is the director of the Department of Developmental Cognitive Neurology, also the director of the Mental Retardation/Mental Disability Research Center at the Kennedy Krieger Institute in Baltimore. She was the 1996 NF, Inc. Scholar. For some time she has served on our Mid-Atlantic Chapter Professional Advisory Board which means that when we have questions in her field, we call and ask her help. We appreciate your bringing three of your colleagues who work together with you this evening. Dr. Michele Mazzocco, who is studying NF-1 and math skills. Her published article is available this evening. Chris Koth is the coordinator who has been on all the studies in the center. Dr. Laurie Cutting is progressing from postdoctoral fellow into a Hopkins faculty member in about a month or so. She has been doing a lot of work on NF-1.
I am going to be masquerading as a basic scientist. For those who are doing research this is an interesting example of how you have these nice, neat hypotheses that you put into your grant proposal to the NIH, and then what you find is much more exciting, but it is not what you were looking for. So that's what I want to tell you about, how techniques and the exact methods that are available to us are very important in what we find out. In other words, serendipity or stumbling over findings still happens in research, -- which is kind of fun!
I am not going to talk very much about the background except to say why we did this. You all know what NF-1 is, so we'll go on.
Many of you know that I've been on one side of a long-standing controversy about the significance of the UBOs in the brains of children who have uncomplicated NF-1 (meaning they don't have seizures and they don't have any kind of tumors) but they have those fuzzy white spots on magnetic resonance image (MRI) that we call "unidentified bright objects," more formally called T2-weighted hyperintensities. By using a method that is called a "multiple regression prediction" and looking at the cognitive function of children with NF-1, we have been able to show that the number of these locations in the brain occupied by these unidentified bright hyperintensities on MRI really accounts for a lot of the lowering of the cognitive ability compared to the siblings. And I won't go into all the controversy; but, as you probably know, a lot of people don’t use sibling controls in their studies, and there has been a lot of controversy about whether the UBOs have any effect on this or not. So our predictive term is lowering rather than the absolute level of cognition.
For example, there have been some people who have participated in our studies, some of them in this room, whose family IQ is pretty high. And the child with NF-1 is still quite robustly normal, but there has been some effect of having the NF-1 compared to the rest of the family. This gets masked, we believe, when you simply say, "Here's a bunch of people with UBOs and here's a bunch of people without." You do not know whether by accident you have a child from a 140 IQ family who comes out with a 100 IQ in this group versus the reverse happening in the other group when you do not use a sibling control.
So we were quite convinced about the significance of UBOs. The reason I am belaboring this is that this is the rationale why we decided to study a little bit more using spectroscopy.
You know what the UBOs look like on MRI. These fuzzy things on MRI are typically in the basal ganglia, and the second most common site is in the cerebellum. There is somebody here on this slide what we are showing in four different time zones: 1) just before coming into our study; 2) a few months after coming in, with a big basal ganglia UBO; 3) by 1993 it's fading; and 4) by 1995, at the age of 15 (in this one person whom we followed over this period), it is really breaking up and the UBOs are small. We wanted to ask. "Maybe when they break up and they go away, there's still some chemistry that's left behind that's going to tell us what they were." That was what we were interested in.
So we designed a study to look at the UBOs using spectroscopy. Spectroscopy is a way of essentially getting living chemistry. There are two chemicals, NAA, n-acetyl aspartate, which is a marker for nerve cells. (Some people think maybe also for glial cells but mainly for nerve cells.) Then there is choline, which is mainly a marker for myelin, and breakdown of myelin in particular, part of the chemicals that go into making the white matter.
So it's sort of chemicals for gray matter and white matter, as a sort of a simplification. We wanted to see whether the chemical abnormalities would persist the brain tissue where there had previously been UBOs. These are very visible MRI abnormalities. Because nobody knows what these MRI abnormalities are, since thank God, most of them go away and the people live many, many years later (so nobody has much in the way of autopsies to find out what are these things in the brain), we wanted to study them over time.
We wanted to know more about NF-1 brains, about the variability. We all know that there are some folks with NF-1 who even in childhood do not have any UBOs. So we want to know if the chemicals are there in the absence of any UBOs, and we also want to know more about it if a UBO is there. Maybe its chemistry will bespeak itself --we'll know what the chemistry is. So there are three purposes: (1) Is there microscopic chemical something-or-other where there used to be a UBO? (2) Is there microscopic chemical abnormality where there has never been a UBO? (3) What about when there is currently a UBO?
We had nine boys who were controls ranging in age from 6 to 19 that we studied. We have 9 boys now (we have not analyzed all of them), two of whom have UBOs, typically in the elementary school age group; two (currently 12 and 16 years) who had never had them. Since this a longitudinal study, we were fortunate enough to have known them since they were little, thanks to many of you in this room who have helped us, so we knew they had never had UBOs at all. And two did have fading UBOs. In addition we had one person who developed an optic glioma and one with a past-only UBO. So we had a sort of sampler of these different NF-1-related MRI states.
We happen to have a technique available to us which gave us whole brain section choline readings. We found different abnormalities in the globus and the thalamus. The surprise was that the chemical abnormalities did not match where the UBOs were. None of these people uniformly had UBOs of much volume in the thalamus. (A few of them had a tiny little UBO in the thalamus.) But we did get most of our spectroscopic chemical findings in the thalamus, which is a location not particularly common to have the UBOs. The globus pallidus is the most common place for UBOs. We then focused our qualitative analysis on the thalami (which is the plural of thalamus).
Back to spectroscopy in the thalamus. You get these scientific-looking charts with all these little readouts - your NAA peak, your choline peak, and you have a reference peak here which is creatinine. (The control is an eight-year-old child.)
The much brighter stuff that you see is in the little boy with NF-1. The choline peak is very much elevated in the thalamus. The interesting thing in the 16-year-old is that instead of there being such a big choline peak, there is a diminished NAA peak. There is a hint of a developmental change.
We use something in statistics which is called the "robust rank order statistic." We found that the NAA was lower in the NF-1 group. But the big finding was the ranking of the ratios of the NAA/choline; there was no overlap. If you listed the NAA/choline ratios, NAA over choline, remember it's a neuroma marker divided by a marker of something happening to the white matter. The NF-1 group ranked ratios and are on one side the control group is at the opposite end. This is not the kind of data that you have to worry very much about statistics. It happens very rarely in people doing human research that they get that kind of clear separation.
Some hints of a trend for a developmental dynamic difference are that the choline level is sky-high in the youngsters who are pre-pubertal. However, the choline level does not look like such a big difference from normal in the older children, but a huge difference in the older ones.
So we have more of a mystery story. I guess all of us who go into medical research are people who like mystery stories. Sometimes what you were looking for and what you find are in different directions. So we have a whole new thing to think about now. What's going on?
The young NF-1 research participants who do not necessarily have UBOs in the thalamus tend to have this elevated choline; the older ones with faded or absolutely no UBOs tend to have reduced NAA. NAA and choline are markers for different compartments of brain tissue. So what do we think is going on?
The possible interpretation is that these UBOs may represent intramyelinic edema (that's swelling), which triggers compensation. There is some attempt to make new myelin sheaths, an increased production of myelin intermediates. This phosphocholine is a very important component that will diffuse out of the initial area of abnormality.
As the compensation fails, then you may have axonal injury (that is, injury to a part of the nerve cells), secondary to this event in the myelin (that is, edema). We see the choline when this process of compensation reaches a limit so that no more effective compensation can occur; we then get some injury, possibly, to the axon. So then again, for therapeutic reasons we may say, "Is there some way we can intervene in this early edema phase if that's the mechanism?"
Our pathologist who looked over the entire set of imaging sections felt that the intramyelinic edema is widespread and is not just all confined in the UBOs. The selective vulnerability of the globus pallidus, and the thalamus may be due to their similar myelin bundle packing. It is good to have a pathologist who knows that there's something that represents a basic similarity in certain brain regions.
We made this a sort of post-experimental prediction that will carry on to later research participants: that we'll see an early choline increase and a later NAA disease. This time, instead of having a set of different people (that is, a cross-sectional study), we would like to do it on the same person to see if this is going on longitudinally in the same individual.
The whole idea is that you will have in the early stages the choline problem, and that in the later stages you will have the reduced NAA. In the cortical white matter, we get more of an abnormality down in thalamus and globus pallidus. It happens that we can visibly see the UBO in the much larger region of globus pallidus, but in a different section. First the problems are in the white matter and later on there will be axonal problems.
I just want to say thanks to my team. I want you all to know how many people are working on this. Chris Koth, Paul Wong (The young Fellow who actually did the analysis of the spectroscopy, but who has gone back to Toronto), Walter Kaufmann, who is our pathologist, and Dr. Barker, who is the head of spectroscopy. Without his method (that he developed) to do whole brain spectroscopy, we could never have begun to find these things. Some other of our team are Mike Kraut, in charge of the neuroimaging corps, and Mike Abrams, who runs our whole image laboratory from day to day. Chris Cox, helped us find the robust rank order statistic.
The Learning Disability Research Center has the two grants that are funding this NF research. In the past, you have heard me talk about spatial ability, verbal ability, so I want to end up on something for which I don't have a slide.
The thalamus makes me excited for my more traditional role as a behavioral neurologist. In NF-1 we have repeatedly found that there is essentially a biparietal syndrome. What I mean by parietal is that the individuals with NF-1 have spatial problems that will be associated with the right parietal and they have language, especially expressive language naming problems, that will be associated with the left parietal lobe. The thalamus, if we really get down and look at it anatomically, is mainly affected in the pulvinar, and that's a "dancing partner" of the parietal lobes. So coming full circle, lest you think I've given up any interest in cognition just because I got so excited about the basic science, spectroscopy also relates back to making a lot of sense out of what we see as an actual picture of children and young people with NF-1; namely, their two big problems are really biparietal, and they have, it appears, regardless of UBOs, something in the thalamus that is indicative of a process that is perhaps interfering with the thalamocortical connections to the parietal lobes specifically. So we have some more directions in terms of understanding the evolution of the learning disability.
We also think maybe now that we have more basic directions that may give us some therapeutic handles. If this is truly some kind of edema, maybe there is something we can do about it and interfere with the whole process, some way we can intervene. I was so excited that I absolutely was so rude as to ask to come tonight and present this to you.
This information has only been presented in abstract form at a neuroradiology meeting. Hopefully it will be presented at a more clinical type of a meeting soon. But I really wanted to share with you the direction that this sort of basic science has really taken.
Questions:
Q: My daughter has UBOs and a learning disability. An MRI showed that she had a mass. We also found out that the mass six months later had reduced in size. My first question is, does a mass or tumor shrink?
A: That is one of the most mysterious occurrences; we have had this happen in our research study. Those are atypically behaving UBOs. There is a big confusion as to whether they really are tumors, or whether they might be just part of this process that we think may be some kind of edema.
Q: So the edema would be confused with tumors? We decided not to go to chemotherapy.
A: Yes. Especially -- was it in the brainstem? Brainstem was the involved region in our one case. But that's unusual. So, thank God, with serial MRIs you can follow it and see that it's actually not what you thought it was initially. That's a very alarming situation, but blessedly it's not too common.
Q: Wouldn’t the thalamus problem also have to do with a problem such as motor skill problems like a posterior thalamus, like the tremors and fine motor problems that cause problems with such as writing?
A: We do, of course, see a lot of motor coordination problems in our friends with neurofibromatosis-1. They get better as they get older, though. Our experience is that that is a part of their problem, which does not seem to stay consistently at the same relative level. But then, you know, again, we have these other structures like the globus pallidus and the cerebellum involved, and those are much motorically related, It doesn’t appear that the anterior thalamus is involved, but the thalamus is a small place, I grant you that, a critical place. It is a place everything goes through. You could explain anything you want on the basis of the thalamus. That is the only problem about the thalamus to a systems neuroscientist. But it doesn't appear to affect the motoric front end of the thalamus as much as it does what we call the pulvinar, which is the back end of the thalamus, the one that related so the back end, the parietal end, of the cortex. But as I said, we have the globus pallidus and the cerebellum which is so often involved. (Tremor and cerebellum is certainly a very familiar neurological correlation.) So I would say that, again, there is such a variable expression in the different people. I think the motor things are much more related to those other, more conventional motor structures than to what we are finding in the thalamus.
Q: I recently attended a meeting where a presentation was made saying that UBOs had been found in some other children with disabilities like autism. I'm wondering if you have looked at any in your work of children with autism to find out if the same things are going on?
A: News to me. Are you sure they're not children with tuberous sclerosis? They have them in tuberous sclerosis--this is the problem. Autism is not a disorder on the same level of analysis. It's a neurobehavioral syndrome that has multiple causes. So that's the problem. I don't know of it. They have a lot of learning disabilities as well, obviously. They have much worse then neurofibromatosis, believe me. They should be so lucky as to just have learning disabilities. I don't know of anything in the medical literature on this topic.
Q: Tuberous sclerosis, are you finding the same things? With the edema?
A: I haven't had the opportunity to study anybody with a tuberous sclerosis, so I don't know. Actually very few people have this particular methodology for doing whole brain slice spectroscopy. A lot of places are limited to only placing a particular magnetic coil at a particular place, like if you were going ice fishing, they could only go right there. But Dr. Barker, and I believe one person at the NIH is the only other person locally that has this whole brain spectroscopy available. So I don't think people have studied much of anything clinically with it. It would be interesting to see if the UBOs are the same.
You see UBOs in multiple sclerosis all the time. You see them in adrenoleukodystrophy. I always bet on them being white matter. Because every condition I know of where they were very typical always involved white matter. I'm not surprised that the spectroscopy is pointing toward the thing that is clinically important. I think this casts a lot of doubt on the radiology habit of reporting the UBOs in NF-1 as "hamartomas," which makes them sound like they're all some kind of rumor. I think we should radically change and not report them out with that term, because the moment you say something like "hamartoma," everyone says, "oh, my God, that's a tumor." I'm glad you brought that up. That's important.
END
|