NEW YORK (GenomeWeb) – The University of Texas MD Anderson Cancer Center has, along with Boston's Broad Institute, been selected by the National Cancer Institute as the site for one of its two new Genome Characterization Centers.
Headed by researchers Gordon Mills and Rehan Akbanie, the MD Anderson GCC will focus on proteomic analyses, using reverse phase protein arrays (RPPAs) to investigate tumor tissue from a variety of NCI projects including the Exceptional Responders Initiative, the ALCHEMIST precision medicine trials, the Cancer Driver Discovery Project, the Cancer Trials Support Unit, and the Cancer Therapy Evaluation Program.
The center is slated to run these analyses for the next four years and will be paid on a per-sample basis, Mills told GenomeWeb.
Data from the analyses will be added to The Cancer Proteome Atlas, a database containing RPPA-based proteomic characterizations of thousands of patient samples and cell lines. Launched by Mills and his colleagues in 2013, the database will, upon the launch next month of its latest version, 4.0, contain RPPA datasets for roughly 10,000 samples.
Mills' lab is one of the main centers for RPPA-based cancer proteome characterization in the world and has served as the primary proteomic resource for past NCI initiatives, most notably the Cancer Genome Atlas (TCGA) project for which he and his colleagues did all of the proteomic characterization.
In all, Mills said, the MD Anderson core has characterized around 103,000 samples using the technique.
Originally developed by George Mason University researchers Emanuel Petricoin and Lance Liotta — who are also major users of the technique — RPPA uses cell lysates spotted in array formats that can then be probed with antibodies to multiple proteins of interest.
While RPPA measures only a limited roster of known analytes — as opposed to shotgun mass spec, which enables broad discovery work — the method has proven popular for analysis of clinical tumor samples due to its high sensitivity and ability to work with limited amounts of tissue, which can be key when dealing with clinical specimens.
"The technology is ideal as a cost-effective, tissue-sparing approach to analysis of the key pathways involved in tumorigenesis and drug responsiveness," Mills said.
He noted that his lab has "vastly expanded and improved" their RPPA platform since its work on the TCGA project. During the TCGA work, the researchers were able to characterize around 190 targets, a number they have since upped to around 300. At the same time, Mills said, they have improved the platform's quality control measures and throughput. The MD Anderson core routinely runs RPPA assays on 1,000 samples a month and can do as many as 2,000 a month, he said.
The additions to the platform's content were aimed at filling gaps primarily in the areas of tumor-stromal interactions, immune response, and DNA damage repair, Mills said, noting that these were "three new areas of excitement in terms of therapeutic opportunities."
The platform's 300 antibodies are intended to "give us as broad and unbiased a view of the pathways involved in cancer and the targets of the therapeutic agents used in the [various NCI] trials," Mills said. He added that the center can also customize the platform to add assays to additional targets, "depending on, for instance, the drug being used in a clinical trial or the tissue that is being analyzed."
Mills said that the center could process as many as 25,000 NCI samples in its GCC role. The total number of the samples analyzed, he noted, would depend significantly on how many samples contained enough material for genomic and proteomic analysis.
"DNA and RNA analysis is going to be the first priority," with proteomic work coming after that, he said.
In the past, one worry regarding such combined analyses has been that samples collected with genomic studies in mind might not be suitable for some proteomic analyses. In particular, certain phosphoproteins are highly labile and so phosphoproteomic profiles — which RPPA is commonly used to interrogate — can change if, for instance, a sample ressected during surgery is left out at room temperature for too long.
Because this is not as much of an issue with genomics, there has been concern that projects like the NCI's Clinical Proteomic Tumor Analysis Consortium (CPTAC) that used samples originally collected for genomic work, might be looking at samples not collected and preserved properly for phosphoproteomic analysis.
Mills said, however, that because RNA is subject to similar issues, the samples should be collected under protocols that will preserve their original proteomic and phosphoproteomic profiles.
"Everything is going to be done to limit [this] challenge in terms of operating room time and ischemia time," he said. He added that most of the samples used will be fresh biopsies, which are "much easier to snap freeze."
"That is going be a massive advantage," he said. "So I think we will probably have much better control of tissue handling."
Data from the NCI initiatives will be added to the TCPA, which, Mills said, in addition to hosting MD Anderson's analysis, has become the primary repository for RPPA data in the field at large.
"We have a number of [RPPA] datasets that have come out in publications that have been submitted by the authors," he said.
The intention is to establish the TCPA as a resource for RPPA data similar to mass spec repositories like the European Bioinformatics Institute's PRIDE repository or the Institute for Systems Biology's PeptideAtlas.
"All data [in the TCPA] can be downloaded and there is no restriction on its use," Mills said, noting that while the MD Anderson RPPA core has run more than 100,000 samples, the repository is focused primarily on large, high-quality datasets that "are likely to be of use to the public."
"Many of the things we run are small sets for customers that are asking specific questions," he said. "That is not going to be as useful broadly."