At A Glance
Name — Timothy Behrens, Professor of Medicine, University of Minnesota
Education: 1979 — BS, Biology, University of Illinois-Urbana
1982 — MD, Southern Illinois University, Springfield, Ill.
1986-1989 — Rheumatology / Immunology fellow, Medical College of Wisconsin, Milwaukee, Wisc.
1989-1992 — Metabolism Branch, National Cancer Institute, NIH, Bethesda, Md.
Research Interest — The lab has engaged in a large gene mapping project in human systemic lupus erythematosus (SLE); has enrolled over 250 SLE sib-pair families; and has identified the location of several chromosomal regions that harbor susceptibility loci for human SLE. A second focus of the laboratory is the use of mouse models to understand the mechanism by which tolerance to self is maintained in the immune system. The lab is applying gene expression microarray technology to study both patients with autoimmunity and various mouse models to identify novel molecules and pathways that are dysregulated in autoimmunity, so that improved therapies for patients with autoimmune disorders can be designed.
Timothy Behrens, a professor of medicine at the University of Minnesota, and a team of researchers at North Shore Long Island Jewish Research Institute, have identified a distinct gene expression signature for patients with the autoimmune disease lupus, an initial step towards the possibility of developing a diagnostic test for a disease that today affects 500,000 Americans, primarily women, and is tricky to diagnose because the disease evolves over time, and because there exists no simple test for its presence.
The results of the study, by Baechler, E., et al, “Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus,” were published Feb. 10 online in the Proceedings of the National Academy of Sciences.
The research project, conducted by gene expression analysis, yielded 14 genes, listed collectively as an interferon expression signature, which was linked to a subset of SLE patients with severe disease.
BioArray News recently caught up with Behrens to talk about this microarray research.
How important are microarrays in your research, particularly this lupus project?
Microarrays have made it possible, really. [The lupus study] was exploratory at the beginning, looking at peripheral blood cells of normal versus people with lupus.
Lupus is a systemic autoimmune disease and we know that there are abnormalities in the blood cells, as demonstrated by functional assays over the last 20 years, so it was a good disease to pick. But, in terms of what the nature of the defects were in blood cells, it was very much an open question. After we did [microarray analysis] on the first 10 patients and controls, we could see impressive differences in gene expression levels.
How did you go about the research?
We used the Affymetrix platform. Right away, we saw several hundred genes that were different between patients and controls. And, we did notice that there [was] a whole host of genes that were previously known to be regulated by interferon that were at quite a bit higher level in the patients compared to the normals. That was a couple of years ago. We spent the next year gathering more patients and confirming the findings that we saw in that initial pilot study. Everything held up, and we were able to generate hypotheses based on the initial few samples, and those held up in the subsequent validation stages.
What kind of chips did you use?
We used the [Affymetrix] U95As, just the A chip. We did 48 patients and 42 controls, one chip for each patient and control. We burned a few other chips on the side with variation studies. We wanted to see how much variation there was going to be in profiles of people’s blood cells because these populations are not static; they are moving all the time. We have done a number of studies to look at how reproducible expression profiles are in blood cells over time in normal people. We have gone back in these patients that either do or don’t have this interferon signature and looked at how robust those signatures are. We have done a lot of other studies following up on this.
What were the costs in the project?
We had an academic discount; all the chips were run at North Shore University Hospital at Long Island with Peter Gregerson. Our bottom line cost per chip, which we got through the New York State Medical Consortium [microarray resource center], was in the range of $700 for chips, probes, and hybridization.
Did you consider home-brew arrays?
There was no other option for us. I’m familiar with microarrays: I trained with Lou Staudt at NIH, who has been doing a lot of work with lympho chips and spotted cDNA arrays with Pat Brown. We specifically chose the [Affymetrix] platform because of the troubles I had heard through the grapevine with the spotted array systems. We didn’t have [a microarray facility] up and going here, so it really wasn’t an option anyway.
How are you funded?
We have two grants now, together with our colleague Peter Gregerson, head of the department of genetics and genomics at North Shore Long Island Jewish Research Institute. He is the principal investigator on some large NIH contracts to collect families with rheumatoid arthritis, and another to collect families with other autoimmune diseases. We recently had funded a large study to use microarrays in the clinical autoimmune setting, looking at rheumatoid arthritis and lupus. He is the principal investigator on that contract, which is for five years.
How did you conduct the collaboration?
For this study, we did all collecting of the lupus samples here [in Minnesota]. We were contacting the patients and collecting the controls. We had them come to the lab for blood draws. We make the RNA, and take it to the point of making double-stranded probes, or more recently, the biotinylated RNA probes. We just freeze them and send them [to Long Island]. They do the hybridizations. It has worked seamlessly.
As we move forward, we have a couple of other large studies following patients with lupus and rheumatoid arthritis, longitudinally. We do some work here now that we have a facility. We got it a year ago.
How do you handle the data?
We didn’t do anything too fancy. We used the Affymetrix software for average difference calls. For the purpose of this study, we more or less ignored the absent/present calls algorithm, unless it was absent across all samples. It turns out to be difficult with the Affymetrix platform, because of SNPs in DNA, you never know if something is not expressed in a tissue because SNPs are interfering with the signal or not. So, we took the average difference values and looked for transcripts across these large populations of patients that were different. We did simple T tests, and that kind of analysis, initially, to identify those genes that were different at a significant T value between the patient group and control group. Each chip gets normalized, based on global intensity based on: Affymetrix protocols, T tests, and the threshold 1.5-fold difference in the means of gene expression across the entire patient group, versus the control group. We looked at the data in a number of different ways. That seems to be the most representative. There was a threshold — there had to be a difference of a least a couple hundred expression units between the mean of control and the mean of patients — to get rid of the very low expressed transcripts.
What did this data tell you?
This tells us that there is a distinct signature of genes in this disease that we haven’t seen in any other disease we have looked at, and this could lead to a new diagnostic test that could be microarray-based. We have been in contact with a couple of diagnostic companies about where to take this. The university has patented the gene expression signatures.
There has been a little bit of evidence in mouse —there are inbred strains of mice that get lupus — that interferon may be playing a role in lupus. Now we have a clue that it may be important in human lupus as well. There is significant interest in targeting this pathway for new therapies and new drugs, blocking the interferon pathway in some way. So, we are also interested in further pursuing that. That’s what’s been exciting about the finding: It may help diagnose the disease better and design therapies and use the microarrays to even target the therapies in more of a designer-type fashion.