ESR1 amplification in breast cancer – A new predictive marker?

Breast cancer is the leading malignancy in women, accounting for more than 350.000 deaths per year worldwide. Several molecular pathways are known to play a role in breast cancer development and progression. Perhaps the most important pathway involves estrogen receptor alpha protein (ER). Binding of estrogen to ER causes phosphorylation and dimerization of the receptor that acts as a transcriptional promoter conferring a growth signal to breast epithelial cells. This makes ER one of the most important therapeutic targets in breast cancer. Gene copy number increase (amplification) is a major mechanism for cancer cells to boost the expression of gene products that provide them with a growth or survival advantage. Numerous genes have been found amplified in a fraction of breast cancers including HER2at 17q21, CMYC at 8q24, EGFR at 7p12, AIB1at 20q13, or CCND1at 11q13. The example of HER2 also shows that amplified genes may be particularly suited as therapeutic targets. Trastuzumab (Herceptin) is highly effective in the treatment of HER2 amplified/overexpressing breast cancers. Several lines of evidence suggest that various genes that can give a growth advantage to breast epithelial cells through amplification are still not discovered. First, amplicons usually span several megabases of genomic distance and harbor numerous genes. This makes it difficult to determine which gene(s) drive amplification. Second, amplifications may be very small and therefore difficult to detect by classical amplification screening methods. In this project we used a three-step approach to search for previously undetected, clinically relevant gene amplifications in breast cancer. First, we used a DNA microarray approach (Affymetrix 10k mapping array) to screen for gene copy number changes in 22 high-grade breast cancers. Second, we specifically filtered these DNA copy number data for small size amplifications of potentially relevant gene classes. Then we used fluorescence in situ hybridization (FISH) to validate identified amplifications and to screen for their clinical relevance on a tissue microarray (TMA) containing samples from more than 2,000 breast cancer patients with clinical follow up data.

Using the Affymetrix 10k SNP array, we found a 6q25.1 amplification in two of 22 cancers that involved only 8 SNPs, all located within one gene: Estrogen receptor alpha (ESR1). The subsequent analysis of more than 2000 clinical breast cancer samples revealed ESR1 amplification in 20.6% and lower level ESR1 copy number increases (ESR1 gains) in 15% of cancers. ESR1 amplified cancers showed estrogen receptor protein overexpression in 99% as compared to 66.6% of non-amplified cancers (p<0.0001). Within the group of estrogen receptor positive breast cancers, ESR1 amplification was a predictor of response to adjuvant anti-hormonal therapy. In a subset of 261 patients who had received adjuvant tamoxifen monotherapy, survival in women with ESR1 amplified cancers was significantly longer than in women with non-amplified estrogen receptor expressing cancers (p=0.023). Remarkably, ESR1 amplification was also found in a multitude of benign and precancerous breast diseases. Our data not only suggest that ESR1 amplification may be a very early genetic alteration in breast cancer, but also highlight the possibility that ESR1 amplification may be a useful diagnostic marker for the selection of breast cancer patients who might benefit specifically well from anti-ER therapies.

However, studies using different methods for detection of ESR1 amplification in clinical cancer specimens reported highly variable data, ranging from as little as 1% amplification up to 27% (Nembrot et al, 1990) or even 36% (Holst et al, 2007) ESR1 copy number increase in breast cancer. These discrepancies are not only attributable to the use of different techniques, because considerable variation is also found between different studies using the same methodology. These data suggest that detection of ESR1 amplification for diagnostic purposes might be more difficult as compared, for example, to HER2.