Molecular classification of early breast neoplasia

Institution: Stanford University
Investigator(s): Robert West, M.D., Ph.D. -
Award Cycle: 2011 (Cycle 17) Grant #: 17IB-0038 Award: $156,561
Award Type: IDEA
Research Priorities
Biology of the Breast Cell>Pathogenesis: understanding the disease



Initial Award Abstract (2011)

Early breast lesions that precede cancer (such as atypical ductal hyperplasia (ADH) andcolumnar cell lesions) represent a dilemma in clinical management. They are important for three reasons. First, these early lesions are often the only lesions discovered on a screening biopsy and it is unclear at the moment how to manage these patients. Second, these early lesions may contain the critical mutations that promote cancer which are likely not yet obscured by the tremendous genomic alterations that typically occurs in the clinically apparent cancers. Third, an understanding of these early lesions may lead to insights into the etiology of breastcancer. An understanding of these lesions may lead to novel interventions, fewer misdiagnoses, and improved patient outcomes. However, these breast abnormalities have not been analyzed well on a genomic (DNA or RNA level).

We have developed a technologic breakthrough that allows us to analyze the gene expression patterns of small, formalin-fixed tissue. We use “next generation sequencing”, a technique that allows for the simultaneous sequencing of millions of small fragments of DNA, to analyze (count) RNA and thereby measure gene expression for all approximately 20,000 genes in the human genome. Specifically, we designed a novel RNA-seq protocol, 3SEQ, to specifically expression profile poly A+RNA extracted from archival tissue, which does not generally lend itself to microarray analyses, since the RNA is hydrolyzed into small fragments. Unlike standard microarray technology, this method works very well in archived patient samples generated by clinical pathology laboratories, which provide the only source for obtaining these lesions. Thus,we can now for the first time profile these lesion-associated tissues, make comparisons to patient-matched normal breast tissue, and the breast cancer tumors that sometimes develop from early lesions.

The results of these studies are expected to provide novel and important information regarding the classification and biology of early breast neoplasias and the progression to carcinoma. Gene expression profiling studies have suggested that the events for metastasis occur early in the carcinoma and can be predicted by studying early stage cancers. We anticipate that we will identify deregulation of initiating and possibly causal pathways, and new cell types or activation states important to progression. This will allow for the identification of expression patterns that predict progression and gene pathways that could be targeted to prevent progression. The expression data set generated will be made publicly available and provide a rich public resource for future studies.




Final Report (2013)

The study of breast cancer molecular progression is important for two main reasons. First, early neoplasias present a dilemma in clinical management. The eventual proper molecular classification of such lesions into high risk and low risk can save thousands of lives as well as reduce patient stress and cost. Previous studies have suggested that the clinical phenotype of the cancer has taken place at the stage of primary invasive breast carcinoma or earlier, and therefore information on early neoplasias will be useful in predicting the likelihood of developing an aggressive cancer and determining treatment. Second, an understanding of molecular progression of breast cancer may lead to insights into its etiology and to development of preventive measures.

Over the course of this grant, we have pursued two specific goals. We have expressed profiled 24 sets of patient matched samples of breast cancer progression for a total of 72 samples. The expression profiling was performed using a RNA-seq technique that we developed, 3SEQ, to specifically expression-profile polyA+RNA extracted from formalin-fixed, paraffin-embedded archival tissue. This is necessary as the early breast neoplasias (non-obligate, morphologically-recognized precursor candidates including usual ductal hyperplasia, columnar cell lesion, and flat epithelial atypia) are small (~1-3mm) and are only recognized by microscopic examination of the fixed tissue. We have also developed an analytic pipeline for characterizing the expression profiles of known genes and also identifying novel expressed transcripts.

We were able to overcome a barrier of obtaining highly pure samples by changing our tissue sampling protocol to individually screen cores of lesional material and reject samples that did not meet our criteria for purity. We were also able to analytically correct for sample to sample sequencing depth variation.

We have generated a multi gene expression classifier for distinguishing normal breast tissue from early neoplasia from invasive or in situ carcinoma. We have identified genes, based on this classification, that are significantly differentially expressed in the three groups of lesions. These genes identify pathways that are activated and repressed in the groups. We have validated our findings for GATA3, FOXA1, and lncRNA 13741 using a tissue microarray cohort of 100 cases. We have begun to identify gene expression changes that correlate with aneuploidies at the early stages of breast cancer progression.

We will continue with our Aims, focusing on the relationship between aneuploidies and gene expression changes at the early stages on breast cancer progression.



Determination of Stromal Signatures in Breast Carcinoma
Periodical:PLoS Biology
Index Medicus: PLoS Biol
Authors: West RB, Nuyten DSA, Subramanian S, Nielsen TO, Corless CL, et al.
Yr: 2005 Vol: 3 Nbr: 6 Abs: Pg:187