Susan G Komen  
I've Been Diagnosed With Breast Cancer Someone I Know Was Diagnosed Share Your Story Join Us And Stay Informed Donate To End Breast Cancer
    Home > Research & Grants > Grants Program > Research Grants > Research Grants Awarded > Abstract

    Research Grants Awarded

    Integrating Computational And Experimental Biology To Discover Novel Biomarkers: Activating Mutations In The Breast Cancer Oncogene Pik3Ca

    Grant Mechanism:
    Investigator Initiated Research

    Scientific Abstract:
    Background: Prior to the genomic era, serum levels of antibodies and secreted proteins were state-of-the-art biomarkers. Today, increasingly sophisticated biomarkers are being introduced, including complex profiles of mRNA and protein expression in tumor samples. Among this new generation of biomarkers are somatic mutations discovered in tumor DNA by genomic resequencing projects. Because they contribute directly to tumorigenesis, they are the most informative biomarkers. In breast tumors, the catalytic subunit (p110) of phosphatidylinositol (PI) 3-kinase (PIK3CA) has been found to have a high frequency of somatic missense mutations. Remarkably, 30% of breast tumors studied have such mutations in PIK3CA (over 70,000 patients per year in the U.S.), emphasizing its importance as a breast cancer oncogene. The pharmaceutical industry has considerable interest in developing isoform-specific PIK3 inhibitors. This is because the different Class 1 PI3K isoforms have distinct functions; thus a drug that inhibits all isoforms can have undesirable side effects. Even drugs that inhibit both wild-type and mutant PIK3CA can have such side effects, because a patient whose breast tumor has PIK3CA mutations depends on proper functioning of healthy cells that express wild-type PIK3CA. The development of personalized breast cancer treatments in which drugs are targeted at particular mutant PIK3CA proteins requires that we know in advance how to identify how patients with particular "mutation profiles" will respond to a drug. Given the importance of hyperactive PIK3CA in breast tumors, learning how different PIK3CA mutations impact drug response holds much promise for breast cancer treatment. Objective/Hypothesis: Our central objective is to learn how to model the molecular mechanisms that induce gain of function in PIK3CA and to systematically identify missense mutants in PIK3CA that drive oncogenesis in breast tissue. This will allow us to develop an initial set of mutation-specific knock-in cell lines for high-throughput inhibitor testing. The results will contribute to design of mutation-specific PIK3CA kinase inhibitors and help physicians to identify patients who will benefit from these inhibitors. Specific Aims: Aim 1) Use computational modeling to identify structural properties that are correlated with malignant transformation and combine these properties in a statistical learning algorithm to predict mutations that contribute to neoplastic growth; (Aim 2) validate and improve the algorithm by testing PIK3CA mutant genes with in vitro and in vivo experiments in vertebrates; (Aim 3) Combine results into a public database of annotated PIK3CA mutations; (Aim 4) Use recombinant knock-in technology in MCF10A cells to develop an initial set of mutant cell lines that can be used for high-throughput drug screening. Study Design: We will combine computation and experiment to predict and validate the subset of PIK3CA mutations found in breast tumors that drive cancer development. We will initially analyze mutation-induced conformational changes in PIK3CA with computational protein structure modeling and molecular dynamics to identify signature properties of mutations that have been characterized as benign or oncogenic in published studies. We will combine these properties in a statistical learning algorithm to select 50 PIK3CA mutants for vertebrate testing with MCF10A breast epithelial cell kinase and growth assays and chicken chorioallantoic membrane (CAM) tumorigenicity assays. The empirical data gathered from vertebrate testing will provide feedback about the validity of our algorithm and allow us to refine our computational approach. In a second iteration, we will use our improved algorithm to select an additional 50 mutants for vertebrate testing. The nine mutations with most potent oncogenic properties will be further tested for nude mouse tumorigenicity. Our work will result in a published database of PIK3CA mutations that includes all computational models and features relevant to PIK3CA hyperactivation, and detailed results of the vertebrate tests. We will also provide reagents for isoform and mutation specific inhibitor design: knock-in cell lines of three mutations found to be positive in cell growth, kinase and CAM assays. Funding for drug discovery will be obtained through other mechanisms. Relevance: In this work, we will identify PIK3CA mutations that cause dominant hyperactivation of the enzyme in breast cancer, regardless of the frequency of mutation in the patient population. It has been suggested that the most common PIK3CA mutations are the most oncogenic, but insufficient evidence exists to definitively answer this important question. Not knowing the impact of BRCA1 and 2 mutations has left scores of families with a knowledge that mutation exists but is of "unknown significance". Patients who carry oncogenic PIK3CA mutations will directly benefit from the development of PIK3CA mutation-specific inhibitors. Such improvements in treatment will significantly contribute to reduction of breast cancer mortality over the next 10 years. Patients without PIK3CA mutations and those whose PIK3CA mutations that are not oncogenic will also benefit as physicians will be informed that the new PIK3CA inhibitors are not a good choice for their treatment. This will prevent them from receiving costly therapies that are of no benefit. Finally, the methods developed here can be adapted and applied generally to other oncogenes important in breast carcinogenesis.

    Lay Abstract:
    In the not too distant past, breast cancer was thought to be one disease. Today we know that there are many sub-types of breast cancer, and that choice of therapy and chance of recurrence depend on characteristics of each subtype (estrogen-receptor positive or negative, progesterone-receptor positive or negative etc). The future of breast cancer treatment will be personalized therapies, targeted at biomarkers discovered in each individual cancer. Advances in biomarker discovery have the potential to significantly reduce mortality and morbidity of breast cancer worldwide in the next decade. A recent discovery in breast cancer research is that a gene known as PIK3CA is frequently mutated. Approximately 30% of breast cancers harbor mutations in the gene, resulting in the production of an abnormal protein, which plays an important role in tumor growth. This percentage (30%) represents over 70,000 women diagnosed per year in the U.S. alone. The mutations are now believed to play an active role in the development of breast cancer. Normal PIK3CA generates a protein ?signal? that is turned on some of the time and turned off some of the time. Some mutant PIK3CA generates a signal that is never turned off and which converts a normal cell into a cancer cell. However, not all mutations in PIK3CA cause cancer. Patients with different mutations are likely to have different responses to the same drug, as has been demonstrated in studies of the drug Iressa for non-small cell lung cancer patients. A member of our team has developed a system in which specific PIK3CA mutations can be inserted into breast epithelial cells so that potential cancer drugs can be tested, but the process is currently expensive and time consuming, limiting the number of mutations that can be explored in this way. To most intelligently use this system and enable breast cancer patients to maximally benefit from new drugs, we want to prioritize PIK3CA mutations systematically by their potential to drive breast cancer growth. Using the tools of computational biology, we first propose to predict which PIK3CA mutations produce an unregulated, overactive protein that can cause cancerous changes and to discover how the shape and movement of the protein molecule change when a mutation occurs. A cancer biologist will test these candidate mutant PIK3CA genes in cell culture and in mice to see if these mutations can convert normal breast cells into breast cancer cells. By prioritizing PIK3CA mutations by their carcinogenicity, we will select three mutations to be inserted into normal breast epithelial cells for high-throughput drug testing. Drug companies working on PIK3CA inhibitors for breast cancer treatment will be able to use the cell lines we develop to test the effects of candidate drugs on PIK3CA mutants and to discover which patients will respond best to a particular drug. We will seek additional funding to produce cell lines for additional high priority mutants. A recent study showed that certain PIK3CA mutations are linked with cancer recurrence and length of survival. From a patient?s perspective, it is important to know whether a PIK3CA mutation found in their tumor is harmful or harmless. Lack of this knowledge with respect to BRCA1 and 2 has left patients with the unsatisfactory result that genetic testing finds a mutation, but it is labeled ?of unknown significance?. By understanding the connection between PIK3CA mutation and breast cancer, we will be able to inform physicians who will benefit most when prescribed drugs targeted at PIK3CA. This will then prevent patients from receiving costly therapies that are of no benefit. This work will translate directly to breast cancer clinical practice and therapy, improving the quality of life of breast cancer patients and prolonging the lives of those with PIK3CA mutations sensitive to the new generation of drugs.