By Aaron J. Sender
When 30-year National Institute of Mental Health veteran Michael Brownstein first submitted his proposal to study gene expression in mice brains to the institute’s Animal Care and Use Committee, “there was a gasp in the room,” he says. To get just a first pass, low-resolution picture, “I would probably need to kill many hundreds if not thousands of animals,” says Brownstein. “And people said, ‘Gee, isn’t there something you can do about that?’”
He did. Conventional microarray experiments require at least 20 micrograms of RNA. But Brownstein and colleagues have come up with an inexpensive way to make high-quality probes from as little as five percent of that amount — without the hassle or expense of RNA or signal amplification.
First he persuaded Charlie Xiang, who had left NIH to first join Stratagene and then run UCLA’s microarray core facility, to return to his Bethesda, Md., roots. Together they began to tackle the problem. They started by looking at the amount of dye label normally introduced into probes. If they could increase the signal, maybe they could get away with less starting material.
No luck. The ratio of labeled bases to unlabeled bases had already been set. “You certainly can incorporate a lot more fluorescent dye into the probe, but then it doesn’t hybridize to your targets anymore,” says Brownstein. “So then we said, ‘OK, well where else can we gain some ground?’”
The first step was to squeeze more out of the RNA as it is converted to the double-stranded cDNA, which is then spotted on the array. “We said, ‘Gee, if you want to make a lot more cDNA, maybe 10 times as much, then you should use random priming,’” says Brownstein. Then to increase the signal, they added fluorescent dye to the end of each primer. “It’s on the very end of the primers, so it doesn’t affect binding of the probe to the target. It can just dangle around up in the water and not affect anything,” says Brownstein.
To save some work, they skipped the step of isolating mRNA, a tiny fraction of the total RNA, out of the sample. “We decided, just as a matter of convenience, why not try random priming total RNA?” says Brownstein. “People said, ‘What a dumb idea, 99 percent of what you’re labeling is junk.’” But not only does it save hours of work, argues Brownstein, it is also more efficient: “You lose product when you isolate.”
Using this method, they began to see how little RNA they could get away with. The answer: a mere microgram.
But that wasn’t enough. “Where we decided we want to be now is single cells,” says Brownstein. Brain tissue is just too heterogeneous to study as one entity. “By profiling single populations of nerve cells in the brain, like the dopaminergic cells that die in Parkinsonism or the motor neurons that die in ALS or the serotonergic neurons that are the target of antidepressants, you might be able to discover new rational targets for drug development.” Using the Eberwine method as a starting point, “we’ve now developed RNA amplification methods to do quantitative array experiments with anywhere from one to 10 cells,” says Brownstein.
Several companies have expressed interest in licensing the method, according to Brownstein, but no deal yet.
Brownstein, who turns 60 this month, jokes that he’s stayed at NIH for his entire career because it’s been a perfect place “to maintain my downward mobility.” It has also provided fodder for the novel he’s writing in his spare time on scientific corruption at the NIH. “It’s fiction,” he says. “Loosely veiled fiction.” In it, the main character Walter Russell, or the Walrus as his colleagues call him, publishes a paper based on fraudulent data collected by a postdoc with whom he becomes sexually entangled.
Literary inspiration aside, “I enjoy the scientific freedom that we have here,” says Brownstein. Even at this stage in his career, he’s not just a cheerleader but still likes to get his hands wet. “It’s one of the few places where senior scientists can still work at the bench,” he says. Getting down to the RNA of a single cell seems like the limit. But it’s just the beginning, says Brownstein. “The limit ultimately is the questions that you ask. So what drives most of this is discovering that you turn the page and, gee, here’s another question.”