With a $13 million grant from the US National Cancer Institute, a new integrated-biology program at Lawrence Berkeley National Laboratory and other institutions aims to incorporate genomics, proteomics, RNAi, and cell-based assays to study cancer, and may be making strides toward identifying molecular predictors of Herceptin response.
“The selling point of that kind of grant was to say, ‘Well, there’s a big gap between the number of successful anti-cancer agents in preclinical screens [in cancer drugs], and the number of drugs actually reaching the clinic [because of the large number that fail due to toxicity or efficacy in Phase 1 and Phase 2 clinical trials],’” Rich Neve, an LBL biologist, told Pharmacogenomics Reporter this week, citing this FDA white paper. “You’re talking several billion dollars for each drug.”
The program concentrates on high-content screens of breast and ovarian cancer cell lines to piece together cancer-cell molecular signaling pathways, and their response to therapeutics, Neve said. So far, the program has established approximately 60 breast-cancer cell lines and between 20 and 30 ovarian cell lines, using multiple parameters — expression arrays, protein analysis, and post-translational analysis — to find a “lead pathway” that generates cancer, he said.
The integrative biology program; largely supported by an NCI P50 grant, said Joe Gray, associate laboratory director for biosciences at LBL. The program will be spread among NCI; the University of California, in San Francisco, Cancer Center; Strategic Research International; and the Netherlands Cancer Institute; with an automated cell and molecular analysis system at LBL, said Gray.
Corporate partners Cellomics and Affymetrix donated “mostly equipment,” the Avon Foundation provided some funding through a $12 million grant to the UCSF Cancer Center, while LBL contributed an undisclosed amount, said Gray.
Neve plans to publish his breast-cancer research in two academic papers over the next month or two, while Joe Gray’s group is nearly finished with two papers on ovarian cancer.
Both sides are trying to identify genomic abnormalities linked with poor clinical response, identify the related genes, and manipulate gene expression using the automated system with siRNAs and “other techniques” to look for knockdowns that result in a reversal of cancer phenotypes, he said.
While Gray’s research has examined the entire genome of the cells to identify targets and markers related to anti-cancer phenotypes, such as reduction of cell proliferation and increase in cell death, Neve’s is more specific.
Partly supported by a Genentech grant, Neve and colleagues are treating breast-cancer cell lines with known therapeutics, using them to predict drug response in certain tumors, he said. A similar approach might be used to identify patient responders, he added.
About 10 of the 60 breast-cancer cell lines are HER2-amplified, “so they become our little model for testing drugs. We’ve got drugs, such as Herceptin and a couple of [tyrosine kinase inhibitors], which target HER2. That’s the data I’m looking at now — Iressa, Tarceva, [and a research compound called] AG1478,” said Neve.
In the search for predictors of response, the group has had the most luck with proteomics, using mass spectrometry, protein arrays, and other methods through collaborators. One protein “actually does predict response” in three of the 10 cell lines of HER2-amplified cells, roughly the same proportion of response in real tumors, he added.
Neve’s group also uses RNAi to identify potential oncogenes and tumor suppressors, to piece together cancer-cell signaling pathways, and to identify genes mediating drug sensitivity. Already there are “a number” of drug-resistance gene candidates, he said.
As a next step, his group is trying to get pre- and post-Herceptin-treated cancer patient biopsies to evaluate the protein’s use as a predictive marker, he added.