Pat Brown needs no introduction in the microarray community. Most academic labs have either consulted with his Stanford lab directly, tried to build an arraying robot based on the model he developed along with Joe Derisi and Dari Shalon, or at least logged on to his lab’s web site to get protocols and advice.
But last Thursday at Rockefeller University, where Brown was invited to give the black tie Harvey Society lecture, a monthly inter-institutional talk given by a world class researcher to New York’s scientific elite, he received quite an introduction. Before the tuxedo-clad flock of society members, Cold Spring Harbor Laboratory director Bruce Stillman introduced Brown as the genius inventor of the two-color probe hybridization method, as well as an example of what the physician-scientist should be. Later, former NIH director and Nobel laureate Harold Varmus would add to these accolades the proclamation that Brown is not only a “visionary,” but “one of the great prophets of what we’ve been calling — over Francis Collins’ objections — the post-genomic era.”
Brown, who looked a little out of his California element in a black dinner jacket, began his lecture by launching into what appeared to be a variation on the standard “genome” talk that has been given lately by the lecture-circuit leaders: First the slide of Wilt Chamberlain next to the lilliputian jockey Willey Shoemaker, to illustrate the vast variation in the human genome; next a genome cartoon, followed by the comparison of the genome to a dictionary.
Then Brown (fortunately) shifted gears: “The book I’d really like to read is the book that the genome is writing every day in every cell of our bodies,” he said. And thanks to Brown’s work, microarrays make this process of reading the transcribed genome, in his words, “trivial and simple.”
To illustrate the rich fabric of data that microarrays weave, Brown showed several slides of color-coded clusters, noting how the “richness in variation of expression patterns extends to the single-gene level” and how, in clusters showing different tissues in the body at different points in time, microarray data can allow researchers to look at “the temporal choreography of gene expression during the cell cycle.” This sort of choreography as well as the systematic understanding of gene expression across different tissues that microarray data provides, he said, is especially important in examining cancer at the molecular level.
In breast cancers, early systematic studies show the “remarkable diversity in gene expression patterns,” Brown said, which “might provide the basis for sorting them into some sort of groups.” These groups might not necessarily be limited to existing subtypes, as the discovery of new subtypes of breast cancer has already been suggested by preliminary microarray research. Developing these molecular clustering patterns of different breast cancers, then using them as a way to forge early-stage diagnostic tools could prove “a hell of a lot more useful than developing the next generation of Taxol,” Brown said.
Protein arrays also promise to provide molecular signatures of disease, according to Brown. The major obstacle, he acknowledged, is trying to get the specific antibodies to spot down on the chip, but he expressed optimism that this would be accomplished soon.
Brown closed the lecture with an intriguing idea of how gene expression profiling could be used as the basis for an innovative treatment for a tumor. This process, which he dubbed “molecular stereotaxis,” involved a similar concept to that of stereotactic radiosurgery.
In stereotactic radiosurgery, which is designed to zap tumors in difficult-to-reach areas of the brain, eight different non-coplanar beams are aimed at the tumor from different angles, avoiding the eyes and previously irradiated normal tissues. While the toxic effect of a single beam on any area is not enough to do extensive damage along its path, the confluence of all eight beams at the tumor area provides sufficient radiation to kill the tumor.
Similarly, he said, “a repository of agents that takes all the features that distinguish a tumor from a normal cell and selectively knock out the tumor cells” could be developed. Each agent would knock out cells with a particular gene expressed at a certain level. While other cells that express these genes expressed would each take a hit, just as the normal tissue that lay along the line of the radiation beam does in stereotactic radiosurgery, the collective impact of the treatment would only fall upon the tumor cells, as they would be the only cells to be attacked with all of the agents.
“Maybe it’s stupid, but I have been excited about it enough that I am starting to do experiments about” it, Brown said.