|
The Role of Proteinomics in the Prevention of Prematurity - transcript
From 23rd Annual Society for Maternal-Fetal
Medicine (SMFM) Conference held in San Francisco, California - February
2003
The Role of Proteinomics in the Prevention of Prematurity
Irina Buhimschi, MD interviewed by Hans van der Slikke, MD, PhD
Hans van der Slikke, MD, PhD:
“It’s February the 6th. We are in San Francisco at the meeting of the Society for Maternal Fetal Medicine and next to me, Dr Buhimschi, Irina Buhimschi; welcome!”
Irina Buhimschi, MD: “Thank you.”
Hans van der Slikke, MD, PhD: “You did a wonderful talk about the role proteinomics can play in prevention of pre-term labour. Congratulations on your presentation. First, a few key words about proteinomics, what is it?”
Irina Buhimschi, MD: “Well, it’s a science that enables you to look at all the proteins at the same time. Now the challenge is to extract the proteins that are meaningful for a certain disease or a certain state of what the organism is in. Also, the proteinomics tools that separate and the proteinomics tools that identify proteins. Sometimes people believe that it is not necessary to know which are the proteins. To me, it makes more sense that you do know. We were very impressed with what, just by using this technology in very, very little time and with very small amount of samples, kind of information you can gather.”
Hans van der Slikke, MD, PhD: “So, you collected the amniotic fluid and how were your two groups compared, let’s talk about that first.”
Irina Buhimschi, MD: “First I had a bank of samples of amniotic fluid that Dr. Romero provided for the pre-term delivery group and my challenge was to look at some of the clinical characteristics of these samples and look how to group the patients so that I got a sense of what the disease group is and the non-disease group is. When you deal with proteinomics, you have to make sure that you know the disease that you are looking for and for pre-term delivery in particular, that is very difficult because there are several pathways that lead to pre-term delivery. Inflammation is one of them, but there’s hypoxia and other probably genetic and so other causes.”
Hans van der Slikke, MD, PhD: “Other syndromes that you can study.”
Irina Buhimschi, MD: “Yes. If one were to take samples from all these women and would aim to look at proteinomic pattern would not be able to extract any difference because it’s not the same disease.
So I looked a long time at these clinical characteristics to really get a feeling of what the information is that I have available there to say these patients are really, really, really sick and these patients are not. I used the white blood count in amniotic fluid and I used the positive amniotic fluid culture result together for the disease group, and the non-disease group were those that did not have any of those and also delivered at term, and then that’s what I compared and what was very interesting, I didn’t have time to talk in my talk, was that there were some out-layers at the end when I looked at the entire population of samples and basically those out-layers were actually misdiagnoses in the first place; clinical misdiagnoses. A couple were contamination of the sample with bacteria from the hands or something like that, so you find those only after you actually have the golden standard.”
Hans van der Slikke, MD, PhD: “So you had two extreme groups: very sick and very healthy?”
Irina Buhimschi, MD: “Yes, and then you go back and look at your whole thing: what you would see at different times and then you would see the disease that you’re actually looking for.”
Hans van der Slikke, MD, PhD:
“So how did you select these proteins; from the amniotic fluid?”
Irina Buhimschi, MD: “Right. So I looked at the profiles visually for a long time and I went to Switzerland and I went and I visited the Swiss protein bio-group and that was in March last year and just by listening to their talks and the way they were doing the analysis of the electrophoresis and talking to the people, I mean: it is probably the most wonderful group of proteinomics, I believe. And just by being there I already was running the samples.
I knew I had differences because in the year 2000-2001, I did an experiment where I observed that I had an amniotic fluid sample and I was supposed to do an experiment to see if fetal membranes get degraded in that amniotic fluid sample and I also was doing something on another type of experiment, so I thought, you know, if one gets degraded, if the amniotic fluid would degrade the membranes, I would be able to see at least a product of degradation. So that’s what I did in Switzerland and I saw that some of the fluid looked different than the other and immediately I said: "you know, I think this is not an infection", so I called my colleague and I said, "you know, I think you have a mistake here", and he said, "why?" So I said, "go check your data" and then he went and said, "yes, I think you are right". So then I knew I had something so I needed to analyse it. I said, "that’s something". So then I said, "okay, but now I need numbers. I need to transform this into numbers". So I went to Switzerland, and I realised that when they do the electrophoresis, they actually have, they do the same thing, diseased and non-diseased and then they make a master gel of each and that’s what I tried to do this analysis, and I simply wrote down what my logic was in my mind on what I wanted to do.
So those billion criteria, the first is that I’m only looking for peaks that show up, so I’m not looking for peaks that go down or just changing intensity, so I’m looking at absolute new peaks. The second criteria is that it has to be at least a big difference in height in that peak because the automatic software transforms in numbers these peaks, but when you use the automatic detection, it detects everywhere, so you’re not really able to see, okay, that is a peak or not. Then I compared the diseased with the non-diseased and I said, I don’t want peaks in the disease to overlap the peaks from the non-diseased, but not even in the wake, so not even in the wake of the peak, and to do that I compared the non-diseased with just a dilutant of the amniotic fluid and if there was a difference here, then it meant that in the non-diseased, there is a peak. So I never took the peak above.
Anyway, so it was a complicated analysis that the end, after the four criteria gave me thirteen peaks, and then I thought that from those, I have to restrict to only the peaks that would give me enough sensitivity, the same for sensitivity and specificity. So I came up with the four peaks to which this larger profile was restricted. These two peaks come from two protein chips that are done in the same time, so experimentally; you would do in duplicate your experimental procedure. But in one, you put one kind of matrix and in the other one, you put the other kind of matrix, so it’s not really a duplicate, you get a little bit of different information, although you can analyse this chip and see the chip with low monochromatic and see the high monochromatic as well, but not as well. But you still have a duplication of your experiment.
Then I saw that that reference peak is very useful because you would orient your samples, all samples there, then you know that you’re looking for two peaks, here, and here. So it’s very easy visually to see which woman is diseased or not. It can be done very easily. My husband has been collecting samples prospectively from the women undergoing amniocentesis at Wayne State and that has also been very important for us because we could see that most of the time, I had the results immediately after, you know, the amniocentesis is done, not for a long time, you know.”
Hans van der Slikke, MD, PhD:
“It is a question of minutes?”
Irina Buhimschi, MD: “Well, it’s ninety minutes in the procedure that I presented, because I did a one-hour incubation, but that was the design of the experiment to be comprehensive. Now, because these peaks are high, you probably can cut down the incubation time to half an hour, so at least to detect these four.”
Hans van der Slikke, MD, PhD:
“So this was the next step after your experiments and finding out the four proteins. Then you did a small prospective study in this way and there you found out that, indeed, it was highly significant?”
Irina Buhimschi, MD: “I did not analyse the data yet because some of the patients didn’t have very long pregnancies, so I cannot analyse for outcome but, with the amniotic fluid analysis, it’s always perfect, fit the result of the culture that comes later, the result also would come.”
Hans van der Slikke, MD, PhD: “That’s what I would say. So are these chips already commercially available?”
Irina Buhimschi, MD: “Yes, the Cyphergen Biosystems is the company who commercialises them and they’re available for scientific use.”
Hans van der Slikke, MD, PhD: “So in this way, it’s a very fast and a very rapid way to see if there is really amniotic infection and, in this way, this could help us to prevent pre-term delivery?”
Irina Buhimschi, MD: “Yes. We have in the last eight months in my laboratory, been involved in trying to monitor the intra-uterine environment non-invasively in patients with ruptured membranes. This category of patients is at very high risk for intra-uterine infection because of the ascending infection after the membranes are ruptured, so these women stay in bed and just wait for the baby to be delivered and they don’t know what happens inside the uterus. Some centres, some hospitals monitor their temperature and they tell them when you have fever, just come to the hospital. So in this, we are able from the leakage that comes from the vagina to look and see, just from the fact that that profile changes from day to day, and to monitor the state of the intra-uterine environment. So I think that is very exciting for us.”
Hans van der Slikke, MD, PhD: “Is that a way of monitoring monophasally?”
Irina Buhimschi, MD: “What was interesting was that during this experiment, the wife of my collaborator ruptured her membranes. So poor Chris, unfortunately.”
Hans van der Slikke, MD, PhD: “So I guess it was a personal touch?”
Irina Buhimschi, MD: “It was a personal touch, yes, it was.”
Hans van der Slikke, MD, PhD: “And my last question: do you think that it will appear in the blood, as well? These same proteins?”
Irina Buhimschi, MD: “I do not think at this time. I haven’t done the experiment, but we have three cases of twin pregnancies where we looked. My husband was the one who identified these cases and thought to do this, and they had amniocentesis from both sacs and one sac had inflammation and the other one didn't, so I think that the disease is actually localised. It probably fully involves the whole maternal organism but by then I guess clinically, yes, later. So if you probably would like to have a diagnosis of sub-clinical inflammation, then you probably would, I mean I don’t think you would get these in blood, but I have to see. I mean these cases with the twins told me that actually it is a local process.”
Hans van der Slikke, MD, PhD: “Thank you very much.”
Irina Buhimschi, MD: “Thank you.”
|
|
