By Business & Finance 
Eammon Conaghan describes how a small genomics lab is making discoveries that could save millions of lives, but have their hands tied when it comes to publishing their findings.
After eight years as a research partner to leading labs across the USA, two pioneers of the genomics industry believe scientists have ignored hundreds of biomarkers that could stop disease in its tracks. Armed with technology that can identify minute changes in gene expression, they intend to improve upon the research projects they’ve seen and set the pharma industry awash with new diagnostics and drug targets.
In an industrial park on the outskirts of Chicago, wedged between a mechanic’s garage and an accountant’s office, two grad school chums pore over microarray data from a recent cancer study. It’s hard to believe, given the modest surrounds of their converted workshop-lab, but a nationwide study has proven theirs to be one of the best genomics labs in the USA.
Biological chemist Scott Magnuson and physiologist Mike Falduto have spent the past eight years examining disease tissue for researchers around the country to help them identify which genes are ‘switched on or off’ when illness takes hold. They founded their lab, GenUs Biosystems, soon after they’d collaborated to develop one of the industry’s premier microarray platforms, CodeLink.
With Sofie the dog – known as the lab-rador – laying patiently on her beanbag next to the beaten-up kitchen-cum-conference table, the two agree the study they’re reviewing has just found six brand new biomarkers for breast cancer.
“Those biomarkers are definitely diagnostic and potentially drug targets,” explains Magnuson. “If a pharma company can develop a compound that acts on them in the right way, they could cure the disease.” His research and business partner, Falduto, nods in agreement. He has spent much of his career working with Abbott Laboratories and has no doubt the industry has the ability to turn these new biomarkers into successful treatments. “There are very talented chemists in the industry that can make medicines to change very specific gene functions. They just need to know what functions to change to stop a disease,” he says.
In other words, they’re crying out for targets like the six that Magnuson and Falduto are looking at right now. Just one of them could save thousands, if not millions, of lives.
Unnovation
Sofie perks up as a lost courier cruises past their office window, then yawns and stretches back out on her beanbag. Magnuson and Falduto don’t normally get many visitors at GenUs, but they’ve been mighty busy since trumping the Food and Drug Administration’s MicroArray Quality Control (MAQC) survey of 2006, which found their tests provided the most sensitive and reproducible insight into gene expression among leading genomics labs.
They’re regularly contracted to run microarrays for prominent academic and research institutes across the country, such as Stanford, Harvard, UCLA, the National Cancer Institute and John Hopkins University. In the process, they’ve analysed many thousands of samples, spanning hundreds of studies, and seen the country’s brightest researchers unearth hundreds of disease biomarkers. Never mind the six they’re looking at now.
Unfortunately, they’ve seen the same researchers ignore those biomarkers because of a crisis of confidence that has produced a culture of ‘unnovation’.
“It’s very frustrating,” Magnuson explains. “During the past five years we’ve almost certainly seen the biomarkers that could help cure breast cancer, yet the data remains buried in archive boxes around the country because researchers don’t take their outcomes seriously.”
He says the problem starts with sample sets, which are often too small to deliver meaningful insight. Even when patterns do emerge, researchers dismiss them because they understand the sample’s so small that it could be an anomaly.
As counterintuitive as it seems for researchers to conduct studies they themselves don’t trust, Falduto says it’s part of a self-defeating mentality that has become endemic. “In many cases research is not regarded as the means to an end but as an end in itself. Scientists will use ‘research-only’ analysis techniques that are thought too inaccurate to draw real-world conclusions, so the study ends up being little more than a technical exercise.
“It’s ok for a white paper or a conference talk, but the findings are really just ‘academic’. No one will stake their reputation on bold claims or risk entering the commercial pipeline based on findings they generated from a small sample set, using untrusted technologies.”
Magnifying the possibilities
However, Magnuson and Falduto say those researchers are doing themselves a disservice. Having provided microarray services for hundreds of studies across the country, the pair say it’s clear that scientists are indeed uncovering the secrets of disease.
From coast to coast, samples from animals and humans consistently show the same gene expression for the same diseases. The GenUs principles don’t even need to read the sample label to know what they’re looking at anymore because the biomarkers are so familiar.
“When microarray technology started out, it was a much grainier picture,” Magnuson explains. “Just a few years ago, the gold standard could only pick up a 1,000-fold range in gene expression. That meant you could only see exaggerated gene behavior that occurred when the disease was already well established. The findings were less revealing and the picture was murkier.”
In other words, it made sense to use microarrays as a ‘research-only’ tool. More recently, however, Magnuson and Falduto collaborated on what became GE Healthcare’s CodeLink microarray, which could detect a 10,000-fold range in gene expression.
They both freely admit that, today, even their baby is old news. The Agilent microarray they currently use at GenUs can pick up a 100,000-fold range in gene expression, for even more penetrating insight into gene function.
“When using modern microarrays properly, you can take a very high-resolution picture of what is happening at a genetic level during disease,” Falduto explains. “Genes might be expressing themselves more than usual as a result of disease, or their heightened expression could set off a chain reaction that leads to a disease state. If you can make drugs to act on them at that very early stage, you could stop diseases from progressing and keep symptoms from developing.”
Painstaking prep
To generate such a comprehensive picture of genetic machinations requires a painstaking devotion to the process of sample prep. Magnuson says that this may be another factor holding back progress in genomic research.
“During the early days, when microarrays were imprecise, people got into the habit of running quick and dirty tests. Nationwide studies have proven that even leading practitioners introduce significant margins of error because of the way they prepare samples. The culture surrounding microarray prep has not evolved with the growing accuracy of the technology,” he explains.
Unfortunately, it’s very uncommon to find labs that combine an accurate process with all the other factors needed to realise the full potential of microarray technology. GenUs comes close – with the nationwide MAQC study underlining their credentials and thousands of samples in their bank – but even they are hamstrung. While they’re perfectly placed to conduct a meta-analysis of all the projects that have used their services, the data is not theirs. The samples and results belong to the hundreds of researchers and institutions they have worked for. And those researchers work in isolation from the many others that GenUs can see are producing similar results.
“It’s frustrating, but the data isn’t ours so we have to play a passive role. All we can do is hand back the results and try to tell researchers what we think is significant based on our experience. Unfortunately, that advice isn’t proving very compelling to them,” Magnuson laments.
Having made no progress with that approach, he and Falduto are now looking at ways to design and perform their own studies. “We can’t use someone else’s results, but we can collect our own samples and do it in a way that will produce clinically useful data. That’s our only option at this stage. It’s either that or live with the frustration of seeing hundreds of biomarkers for a disease and then shrugging our shoulders when people ask to see the evidence. It will be costly to get off the ground but it’s something we have to do.”
*This article was originally published in the Life Sciences Review in January 2012.
