Researchers profile over 6000 genes in the two primary types of normal breast cells

May 1 2004

For several years, scientists have attempted to identify gene expression changes, using microarrays or ‘DNA chips’, in order to understand and predict breast cancer onset, progression, and clinical outcome. Although important insights into breast cancer have been drawn from determining the expression ‘profiles’ of thousands of genes in tumors, the interpretation of experimental results has been complicated by the absence of knowledge about the gene expression in ‘normal’, non-cancerous, breast cells. However in a paper published today in the journal Cancer Research, a team of scientists from The Breakthrough Toby Robins Breast Cancer Research Centre at The Institute of Cancer Research (ICR) in London, and the Breast Cancer Laboratory of Ludwig Institute for Cancer Research (LICR) and University College London report that they have now elucidated the expression profiles of over 6000 genes in the two primary types of normal breast cells.

The majority of breast cancers originate in an internal structure of the breast, the terminal duct-lobular unit, which is comprised of two different types of cell; the inner ‘luminal’ cells, potential milk-secreting cells, in which cancerous genetic changes occur; and the surrounding basal layer of contractile ‘myoepithelial’ cells. The LICR/Breakthrough team separated and purified the two cell types from material from breast reduction surgery, and showed that the cell types have distinct and quite different gene expression profiles.

“The problem is that the vast majority of breast cancer experiments have used tumor samples because that was usually the only material available,” explains LICR’s Dr. A. Munro Neville, one of the senior authors of the study. “But tumors actually have different mixtures of normal luminal cells, normal myoepithelial cells, and cancer cells. Now we not only know that these cells have very different gene expression changes, we actually know the base-line expression of genes in both the normal cell types. So we can go back through all the data from the experiments with tumor samples, and discriminate between the genetic changes due to cell type differences and genetic changes due to cancer.”

Another important finding from the study was the identification of differences between luminal cells taken from primary breast samples and luminal cells cultured in the laboratory to which, for many ethical and logistical reasons, scientists frequently have to resort when performing experiments. The LICR/Breakthrough study results sound a cautionary warning for interpreting microarray data from cultured cells, and may also be helpful in determining between real experimental observations and artefacts relating to in vitro cell culture.

The results will also help to determine smaller and more accurate classifiers of different types of tumor. Microarray analyses have previously been used to distinguish between sub-classes of breast tumors based on the differential expression of many hundreds of genes. For example, one sub-class of tumor was identified as ‘basal-like’ because the tumors expressed a series of genes thought to be expressed in myoepithelial, but not luminal cells. The basal-like tumors appeared to be more aggressive, and a retrospective study showed that those patients with basal-like tumors had a poorer prognosis than those with another sub-class. Applying the base-line dataset from the normal cells to the study that originally proposed the classification, allowed the LICR/Breakthrough team to identify a handful of critical ‘marker’ genes that may be better able to prospectively diagnose tumor sub-classes, and confer independent prognostic information. These markers might also indicate possible avenues of therapy.

Importantly, according to Dr. Neville, the generation of this more accurate dataset may also have a major impact on patient care. “In order to discriminate between different types of breast cancers in the pathology lab, the surgeon often has to take a sample large enough to incorporate the basal myoepithelial cell layer. If, one day, we could use a fine needle biopsy in conjunction with the novel myoepithelial markers we’ve identified, we could not only potentially improve diagnosis, but also our therapeutic approach.”