Adult leukemia strikes about 3,500 Canadians every year—often with deadly results. For them, the disease is pervasive and ever present. But for the physicians who treat these patients, adult acute leukemia can be frustratingly elusive.
Part of that elusiveness can be attributed to the highly unpredictable nature of the disease, and to how it reacts to treatment. After witnessing the effects of treatments like chemotherapy, physicians and researchers have determined there are three different types of patients: the 20 to 30 percent whose lives are saved by chemotherapy; those who appear to respond initially, but then relapse and succumb to the disease; and those who don't respond at all to treatment.
Researchers grappling with the question of why chemotherapy and other drugs help some patients, while others don't respond to at all, have turned to genomics for answers. By examining and comparing patient samples that contain thousands of gene sequences, RNA (ribonucleic acid), and proteins made by RNA, they hope to discover which genes are switched on inappropriately in cancer, and which genes are switched off inappropriately.
At the University of Saskatchewan, a team of researchers is tackling the problem and hoping to solve the mystery of the "on/off" switch. Hematopathologist John DeCoteau and his colleagues at the university's College of Medicine have teamed up with Anthony Kusalik, a computer science bioinfomatician. DeCoteau diagnoses diseases of the blood, primarily involving cancer, while Kusalik is the director of the university's new Bioinformatics and Computational Biology Research Laboratory.
With a contribution from the Canada Foundation for Innovation, and with deep discounts on computer equipment and software provided by IBM Canada, the new bioinformatics laboratory is enabling Kusalik and his colleagues to further investigate previously unsolvable problems posed by biology and life sciences. One of those problems is the comparison of gene sequencing data. DeCoteau and his team will rely on this comparison in order to answer questions like why adult patients with acute leukemia react differently to chemotherapy.
To get the job done, the new lab is stocked with state-of-the-art computers, high-speed data networks, and storage facilities that facilitate the field of bioinformatics.
"Ten years ago, or even five years ago, if you wanted to start hammering away at which genes seem to be the most important in either being inappropriately turned on or off, you could look at only five or 10 genes at a time," says DeCoteau. "Now, you can look at thousands of genes at a time, and compare among patients the pattern of (gene) expression."
Using microarray DNA technology, researchers can now examine 19,000 spots of DNA on a single slide. As a result, they are profiling thousands of genes in different patients, comparing them with samples from other patients, and are then searching for recurring patterns of gene expression. Those patterns will help DeCoteau and his team spot the reasons why patients have different responses to particular drugs. They may also help to confirm DeCoteau's hypothesis that the reason patients respond differently is that they are actually suffering from different types of cancer—genetically different diseases that are all currently classified as leukemia. "The fundamental hypothesis is what appears to be the same (disease) is genetically very different," says DeCoteau.
Once Kusalik and his colleagues find possible patterns in their analysis of data, they turn that information back over to DeCoteau and his team to develop experiments that will confirm the veracity of the patterns.
Without the computer computations, "it's a needle-in-a-haystack search if they just try to guess what the patterns are," says Kusalik. "With our tools, we make the search practical. They know where to look. Instead of searching the entire haystack, we tell them the needle is in this particular fork full of hay."
The supercomputers in the bioinformatics lab are essential tools for this work because of the staggering amount of data involved in the analysis, Kusalik says. "One of the key features of this bioinformatics lab is a very large data repository," he explains.
Solving the leukemia conundrum is just one of the many applications the bioinformatics lab is tackling. Other problems they're dealing with include the development of a vaccine to combat a virus that attacks cattle, and creating genetically enhanced canola and wheat.
"We're getting a lot of interaction and a lot of seeding of ideas, which come from the interaction of people from these diverse disciplines," says Kusalik. "The fact that you have a biologist and a chemist and a mathematician all sitting and chatting about an interesting problem means you get these diverse views on things."
At the University of Saskatchewan, a collaboration among researchers in medicine, life sciences, and bioinformatics is offering hope for adult patients with acute leukemia. The hope that, one day, doctors will be able to specifically target chemotherapy—taking into consideration the individual response that each of these patients has to treatment and drugs.
Because patients with acute leukemia currently respond differently to conventional treatment, doctors, patients, and their families are often frustrated when medicine that works for one patient has no effect on another. By using the Bioinformatics and Computational Biology Research Laboratory to compare genetic sequences in samples from about 100 patients initially, John DeCoteau and other researchers hope to pinpoint the genetic differences, in patients or in their cancers, that explain why not all patients respond to chemotherapy in the same way.
By identifying patterns in gene sequences, DeCoteau hopes to one day be able to predict which patients will respond well to particular drugs, and which ones will not. That information would ultimately allow doctors to screen patients at the time of diagnosis, and then determine if convention chemotherapy will work for them. Alternatively, if a doctor knows in advance that a certain treatment won't work, the doctor might decide to move immediately to more aggressive treatment that might lead to a better outcome.
Since chemotherapy often carries with it toxic side effects, it's important to precisely identify which patients would respond well to standard treatment, says DeCoteau. If it's clear, in advance, that a patient won't respond to chemotherapy, there's no sense risking those side effects. "For a lot of blood-based cancers, the treatment can be very aggressive," he adds.
Finding patterns in gene sequences could also lead to the development of new drugs or new therapeutic targets. "Say we find a group of patients who don't respond at all to the treatments—their leukemia cells seem to be completely resistant to the drugs they are given," says DeCoteau. "You might find there is a pattern of genes critical to that whole process. You could potentially identify the key players causing the resistance to that drug and develop targeted or directed therapies—new approaches to try to overcome it."
The bioinformatics lab is critical to helping discover a solution to the leukemia puzzle because the lab's supercomputers can analyze huge amounts of data. "If you look at 19,000 genes at a time in 100 patients, you've got 1.9 million pieces of data to try to sort out, to find recurring patterns that are meaningful," says DeCoteau. "In my opinion, the work cannot be done without having excellent bioinformatics resources.
When Professor Anthony Kusalik was assembling a state-of-the-art Bioinformatics laboratory at the University of Saskatchewan, he knew the institution couldn't afford all the necessary computer hardware required to perform the vast and complex calculations that the discipline involves. So he turned to IBM Canada Inc. for help.
IBM supports research at major Canadian universities and has been investing heavily in the life sciences for several years, says Stephen Perelgut, manager of university relations for IBM. The University of Saskatchewan project was particularly interesting to IBM because of its cross-border potential. One of the collaborators Kusalik works with is Isidore Rigoutsos, a scientist at IBM's research lab in Yorktown Heights, New York.
Rigoutsos is interested in matching gene sequence patterns compared through microarray technology. Microarray technology is a technique that examines the entire genome to identify gene sequences or to identify individual genes. Finding the patterns in different genetic samples from cancer patients is one of the problems Kusalik is tackling at the lab—in collaboration with his colleague John DeCoteau, a hematopathologist at the university.
The marriage between IBM and the University of Saskatchewan seemed designed to serve both interests—private and public. IBM provided deep discounts on the computer hardware Kusalik needed to set up the lab. And Rigoutsos is working with Kusalik to refine the software tools necessary to analyze the gene sequencing data that DeCoteau and his team are gathering from leukemia patients.
Although IBM recognizes the project's private-sector potential, the company says it's pleased to be contributing to a project that could make a huge difference to leukemia patients. "IBM has found that raising the river floats all the boats," says Perelgut. "We like the end result of these things too."