This story originally ran on May. 6.
By Tony Fong
Name: Momar Ndao
Position: Laboratory director, National Reference Center for Parasitology, 2002 to present; assistant professor, McGill University Faculty of Medicine, division of infectious diseases, 2008 to present
Background: Associate professor, University Cheikh Anta Diop, CES Parasitology, 2005; research associate member, Centre for Host-Parasite Interactions, McGill University, 2005
Using SELDI-TOF/TOF, a multinational team of researchers has uncovered serum-based biomarkers that it hopes will lead to a more effective manner to diagnose Chagas disease.
More than 10 million people, primarily in the Americas, are thought to have Chagas disease, caused by the protozoan parasite Trypanosoma cruzi. Most are asymptomatic, but untreated, the chronic disease can affect the host's nervous system, digestive system, and heart, and in some cases lead to death.
In late 2006, Ortho-Clinical Diagnostics' ORTHO T. cruzi ELISA Test System became the first screen for Chagas disease to be approved by the US Food and Drug Administration. Last week, the agency approved the second such test, Abbott Laboratories' Abbott Prism Chagas test.
In a study published last month in the Journal of Clinical Microbiology, researchers reported a set of biomarkers for the disease that they found in sera. They also reported that assays using three to five of these biomarkers achieved up to 100 percent sensitivity and 98 percent specificity for the disease.
Among the biomarkers that they said have the most diagnostic potential are M1P1 alpha, C3a anaphylatoxin, unusually truncated forms of fibronectin, apolipoprotein A1, and C3.
"There is no gold standard assay for CD, but tests in current use include ELISAs, IFA, PCR, and even xenodiagnosis," the authors of the study wrote. "Better tests are needed to protect transfusion and transplant recipients worldwide and to address the needs of frontline physicians in regions where CD is endemic."
In addition to using SELDI-TOF/TOF, the researchers used Western blotting to confirm their original data. "Efforts are now underway to generate monoclonal reagents that will discriminate between the full-length and truncated host proteins for use in traditional immunoassay formats" such as solid- and liquid-phase enzyme immunoassays and Western blotting, they added.
This week, ProteoMonitor spoke with Momar Ndao, laboratory director of the National Reference Center for Parasitology of the Research Institute of the McGill University Health Center, and the first author on the JCM study.
Below is an edited version of the interview.
In your study you acknowledge that some of the criticism that has been directed at the SELDI-TOF instrument is justified. Why did you use it then for your work, anyway?
I think the bottom line is [no] technology that is mass spec[-based] … will give you everything that you are looking for. We have a MALDI-TOF/TOF here … but I strongly recommend people to use SELDI if they know how to use it.
If you take any kind of sample, even a cell sample, and you do profiling, I can guarantee you [that] sometimes you are not able to see some protein you are looking for, even spiked proteins on the MALDI while you can see the peak on the SELDI.
Why did we use the SELDI? First of all, at that time, to be honest, that was the technology that we had on hand. And since we were working on parasitic disease, we took 44 samples, [disease] and controls, and just said, 'Let's see if you can see a difference.'
[ pagebreak ]
Seeing that difference, we said, 'Hmm, that is interesting.'
[After we took] more samples, eight by eight, 16 by 16 … [and saw] these differences for the same profiling, seeing the same peaks … we said, 'Now, we think it's time to run a lot of samples.'
So we contacted [Ciphergen, now called Vermillion] and said 'We're interested in this technology.' And we had an agreement and got the machine; then I went down [to its research facility] with some samples. At that time, we [were working on Chagas disease] but I threw in some other disease samples.
[Editor's note: Bio-Rad Laboratories acquired the SELDI business in 2006 (PM 08/17/06).]
They had some big trouble with those samples … but I was laughing inside because I knew what I did there but they didn't know. When I took my data back … I processed a huge number of samples just to do a validation of those biomarkers.
So this paper is out now, but, in fact, my proteomic work [began] six years ago with Chagas disease.
Lots of people, what they do is criticized because they run just a few samples and then they want to publish their results and try to get more funding. It's very difficult if you have funding and you don't publish a paper. Unfortunately, this is what is happening in academia. That is why some people sometimes publish a mistake.
And I think … with the SELDI-TOF, the biggest mistake [was] people oversold the technology in the beginning. Secondly, the users didn't know how to use it. They need a kind of [standard operating procedure]. If you have a machine, normally every week you should do a calibration to make sure that this machine works well.
The huge mass range is the difference between the SELDI and the MALDI; the sensitivity is [greater]. The resolution is lower because you can't have both at the same time. The peak might be ugly because the problem is that the SELDI can pick up any protein or peptide ranging from 2 kDa to 180 kDa.
Normally, the machine should be [calibrated] using the insulin chips with the low mass range, and IgG with the high mass range to make sure that everything is OK. Also, we have our own internal SOP.
Also, [the SELDI] is high throughput. We could run at that time [when we began our research], because we were using an auto-loader, 96 times four samples in a day, which is really a big deal for us.
So you had experience with the SELDI before you started using it for this study that was just published?
Oh yes, oh yes, definitely.
You also used alternative assays for this discovery work. Was that in recognition that there are problems with the SELDI, and you wanted to make sure that the data you were getting off the machine were true?
Exactly. I think that anything that is generated by mass spec — I don't want to say that you have to take it with a grain of salt, I don't want to be critical of all mass spectrometry — but from my experience, if you cut a gel and send it off to three different people for identification, you might get three different answers.
But people have to use it as a tool to interpret the data. That's why when we did this, we set up an in vivo assay. … Honestly, when we had this data, I wasn't very happy because I was expecting to see a lot of biomarkers from the parasite.
[ pagebreak ]
After doing all this [and] getting the same stuff, I said, 'What is going on?' And we even went to a cardiologist at Royal Victoria Hospital — he's the head of cardiology at Royal Victoria Hospital — and when he saw this data, he jumped and said, 'My goodness, do you know what you guys are seeing here?'
And I said, 'No,' because this was expressed in HDL.
And what we did was first of all confirm it and we ordered the antibodies instead of in vitro assays. … Then we also infected mice to look for the cell in vivo. At the same time, we also cloned and expressed this protein.
It was a lot of work because by cloning a protein, all those proteins are expressing. And we injected it into a rabbit after developing the Western blot. This is another way of using another platform to validate the biomarker we found.
There are already FDA-approved tests for Chagas disease, as you point out in your study.
Before [Ortho-Clinical's test was approved in 2006] … there was no test on the market here in North America.
How is your research going to improve on what's already available then for screening Chagas?
We are developing a monoclonal antibody. I think that this is extremely important. We got some funding from the government.
As you know, making an antibody takes a lot of time and money. Unfortunately, last year, the paper was not ready — it was ready for patent [application] but we held it. … But I am developing some monoclonal antibodies that people can use … that would be more specific for screening.
Is there data that you've generated since you submitted your paper that you can share?
There is a study now that I want to publish, that was a blind study [but] I don't know if I am able to give it to you because I have to write a paper with the American Red Cross. But they sent out some samples something like one-and-a-half years ago or two years [ago] … and I was able to tell them this, this, and this based on only one biomarker … and I think we did well based on only one biomarker, not all of the biomarkers.
It's not published, but in the coming weeks, I think I will write this paper.
How many markers have you identified that you think are viable candidate markers?
For Chagas disease, we have 27 potential biomarkers. There is an algorithm tree … and we used something called BPS [Biomarker Pattern Software from Vermillion] that helped us to say that if you combine this and this and this, [you can see this].
So we are focusing firstly on those biomarkers, but there are [more] in our drawer that we're going to take. Recently, I was talking to a professor who is working in lipids. … But we're working in other infectious diseases, let's say Dengue, malaria, [and] toxoplasmosis. We are looking for biomarkers also.
Twenty-seven biomarkers is a lot to develop a diagnostic test with. Are you going to whittle that down to something that will be more manageable?
I am focusing on those biomarkers that were identified in the [Journal of Clinical Microbiology] paper. Otherwise, making all those monoclonal antibodies will be a headache. That's why I'm focusing on those and leaving the rest in my drawer.
[ pagebreak ]
But in sharing this with other colleagues, they just jump and say, 'Why don't you check on this also?' But honestly, I'm just focusing on what we have put in that paper right now.
There is a big list there.
But within those 27, are there a few that you think are better candidates, or is 27 the number that you think is optimal for developing a diagnostic?
Those are optimal because there are some that can predict even the cardiac form of [Chagas] disease.
Were there any steps that you took that you think were particularly novel that other researchers in proteomics can learn from?
I think if you look at our methodology, it's extremely clear. I remember a comment from a reviewer saying the materials and methods [section] is so long. I said, 'No, I'm not going to cut it because I want to leave all information there for people who want to repeat it.'
I think the approach we took, by first of all, doing our discovery stage and validating, using a lot of samples on Chagas disease, and [doing] another validation looking for other protozoan parasites to make sure those biomarkers are specific for what we are looking for, and afterward trying to do another validation using immunoassays, cloning and expressing those proteins, I think we just worked through from A to [Z] and just closed the loop.
I don't want to say that this is the best [way], but I think this is the most appropriate way to complete research.
Do you have any plans to replicate these results using another mass spec platform?
We have the MALDI-TOF/TOF right now in house … and we are seeing the same things.
Every technology has its strength and weakness. For the MALDI we are not able to see sometimes high, high, high molecular weight, but the way I know it after is when I take, let's say insulin or something I know, and I spike it.
But … since I've had the new edition of the MALDI-TOF/TOF, the 4800 [from AB Sciex] … I've never done a comparison where I could not pick up the protein. To make the story short, yes, we are using another mass spec technology to do another cross validation.
Are you seeing more data, possibly additional biomarkers, on the MALDI platform?
Yes, when you switch [platforms], even when you switch from the PBS IIc, the earlier version of the SELDI-TOF/TOF, to the PCS 4000, this is more sensitive [and] you discover some more biomarkers.
What are your plans now in terms of turning your results into a diagnostic?
Right now, we're almost [seeing the light at the end of the tunnel] regarding generating the monoclonal antibody because it's actually the biggest next step. Any small biotech company can take it, ELISA it, and use it.
So we see the light at the end of the tunnel by developing monoclonal antibodies for these discovery biomarkers.
We have something on hand. I am a bit paranoid. I have to double-check, verify … to make sure what we are using is fine. …The timeline is probably in eight, nine months [before] I think we can release the reagent.
Do you have a commercial partner who will be helping you in launching this or will you be making it available directly from your lab?
I am an academic researcher, I don't know how to do business. We can talk to these people, let's say people from technology transfer, or whatever, who will get involved. Honestly, I can talk about science but not business.