At A Glance
Position: Professor of biochemistry and molecular biology, Mayo Clinic College of Medicine, since 2002. Director and founder of W.M. Keck FT-ICR Mass Spectrometry Laboratory Research, Mayo Proteomics Research Center.
Background:Faculty, founder of the Mass Spectrometry Center for the Study of Biocomplexity, Virginia Commonwealth University, 1997-2002.
Postdoc, Richard Smith's lab, Pacific Northwest National Laboratory, 1995-1997.
PhD in analytical chemistry, David Hercules' lab, University of Pittsburgh, 1990-1995.
At this week's Association of Biomolecular Resource Facilities conference, David Muddiman gave a talk on using LC-FT-ICR-MS to search for biomarkers predictive of early stage ovarian and pancreatic cancer. ProteoMonitor caught up with Muddiman before his talk to ask him about his background in FTICR and proteomics.
When did you get into proteomics?
I went to Virginia Commonwealth University in Richmond, Va., and that's where I started my academic career. I was just relatively still new to FTICR, but it really fascinated me. So I went into FTICR and went to VCU. I kept playing around with DNA, but I had a few students that were interested in peptides and proteins, so we just kind of got started with that in the late '90's. John Fenn was in my department - he's the Nobel laureate who invented the electrospray. His former student Matthias Mann — he came by and we all went out for dinner. He gave a talk in the department and hung out with me for a couple of days and said, 'You've got to start doing proteomics.' I said, 'You know Matthias I'd love to, but you've got lots of money and we don't.' So Matthias actually sent me a small amount of money as a strong encouragement to start thinking about proteomics. And it was no strings attached — it was just to get me to start thinking about proteomics.
So I did — we started thinking more and more about proteomics and started to make some progress in doing more about that area. There weren’t many resources within the university to do proteomics, but I met this guy on a study section at the NIH. His name is Stan Hefta. He's at Bristol Myers Squibb in Princeton, New Jersey, and he runs the proteomics effort there, and he's a very clever guy. So I said, 'Can I come up and visit you sometime to see what your lab looks like, because I'm trying to do this in my lab?' He said, 'Sure!' So I went up there and he gave me all the protocols and all the tricks of the trade. So I got all these things between Matthias encouraging me, and Stan giving me all these this help.
Around that time Mayo called me up and asked me if I wanted to become head of this new proteomics center. I was a little hesitant because I was in a true academic institution - a university versus a medical school — and also because I was just getting my feet wet with proteomics. But then I figured hey, I'm still young, I should just jump right into it.
Once you got to Mayo, did you start immediately working on clinical applications?
Yes, that's right. This is very much a clinical proteomics focus since it's a medical school. We have people within the group who spend time with me developing new clinical tests that we convert into the clinic. Converting over, or translation, means translating from a basic research lab experiment into the clinical laboratory, which is something that has a lot of quality controls and reproducibility studies that have to be done before they can actually call it a test.
What is the first test you converted over to the clinic?
That was for the familial form of amyloidosis. It's a protein called transthyretin. That's actually a proteomic/genomic test. It's first a quick proteomic measurement. From the time the sample is obtained to the time we get the answer is 20 minutes. And then if it's positive what we do is we reconfirm it with DNA sequencing.
How did you end up honing in on transthyretin?
Well papers have been published out there — it's actually a pretty well-known protein. The thing is it's very different when you go from publishing a paper in a mass spectrometry journal to actually converting it to a clinical test — all the quality control, the integration with a genomic test took time.
Are you working on developing other tests?
We have four other tests right now that we're developing. They range from hypertension to hormonal diseases — thyroid disease — that's about all I can tell you because these are undergoing patenting. But the two I can tell you about, these are very high occurrence diseases.
Are those also proteomic tests?
Yes, they're all proteomic tests. It looks like they won't be followed up with a genomic test because of the size of the gene. See, transthyretin is pretty easy because it's a very small gene, small protein. But a lot of the other genes we're looking at are enormous, so following it up with DNA sequencing would be very labor intensive. Just to sequence transthyretin, it's about four to five hundred dollars to do that in a clinical setting. And that's a really small gene. If you look at some of these other genes, that would be probably two or three thousand dollars. You can't charge somebody's Medicare $3,000. You can't spend $3,000 to rule out diseases. And that's where proteomics has a lot of potential. Because mass spectrometry proteomic-based techniques are very cheap. Once you have the instrument, you can just run lots and lots of samples.
Are you working with any kinds of chips at all?
No chips. We're in Minnesota, not Las Vegas.
No, we did play with a few protein chips and stuff, but we weren't overly impressed. Just trying to find markers and up and down regulation is what protein chips are good for. We're not opposed to them, but I think that the clinic is pretty geared around mass spectrometry. And clinical testing raises the level of quality control to a level that I wasn't even quite appreciative of until I got here. And protein arrays just aren't there yet.
What do you think about the reproducibility of mass spectrometry tests?
It takes time. You have to become the devil's advocate very quickly — you have to feel yourself around in the three-dimensional space and figure out what not to do. We're right down to what pipette tips to use, what vials to use, which solvents to use, where you get the solvents from, how long they sit on the shelf. Quality control in a clinical lab, you just work through those things. Once it's worked out, usually it has nothing to do with the mass spectrometry, it has to do with all the handling up front. Mass spectrometry tests in our hands anyway is always dead on. So it's just the sample handling up front — if your protein degrades or something - the mass spectrometer will expose that, but then you won't know if it's a mutation or degredation.
What kind of work are you doing in terms of biomarkers?
Ovarian and pancreatic are two main biomarker discovery diseases right now. We're going to be looking at colorectal cancer and prostate cancer here in the future. Right now, all the biomarker discovery is all based on FTICR. Our philosophy is that for discovery - to find markers - you need a very complex technology. But we're not advocating that once you find the markers you keep using that technology. Once you find the marker, then you reduce it down to a very simple mass spectrometer. It's discovery proteomics versus targeted proteomics. For discovery, you want to look at everything and figure out what's different between a disease population and a control, or two different disease populations. And for targeted, you say, 'OK, this peak right here - what is it?' And you measure that a bunch of times, and quickly.
What's the advantage of using FTICR over other simpler mass spectrometers?
One advantage is peak capacity, which is how many components you can differentiate in a single mass spectrum. That's because the resolving power is so high. Your peaks are very narrow, so you can see in a conventional mass spectrometer a peak, but if you put it in an FTICR, you'll see two. And they're real.
The second thing is mass accuracy. If you want to take a disease population and a control population, you're detecting thousands of components. How do you know it's the same thing in the disease, versus control? We can say look, down to four decimal places, this peptide is in the sample. And down to four decimal samples in the same exact mass is in this other sample. We use mass accuracy as a way to compare.
The third thing is dynamic range. In serum and plasma, you're looking at ten orders of magnitude of dynamic range. FTICR offers a pretty broad dynamic range in a single measurement.
What are you looking to work on in the future? Are you looking to do any kind of validation studies?
Yes. We'll do that in our lab too. We have 12 mass spectrometers in our group and three are FTICRs. So we quickly will push [the biomarkers we have] down into a different instrument platform and validate those markers.
Then of course we're doing a lot of different other research collaborations with a lot of other people, for example in bone disease. These are just really focused collaborations with biochemists. It's funny — everybody thinks we just do FTICR, but we also still do a lot of 2D gel electrophoresis. And we're doing a lot of phosphorylation, glycosylation — a lot of post-translational modification work.