Abstract P2-05-21: Molecular subtypes of invasive breast cancers show differential expression of the proliferation marker Aurora Kinase A (AURKA)

Abstract

Abstract Background: Invasive breast cancer (IBC) has been classified into four major subtypes based on gene expression profiling. The luminal A subtype (LA) has the best prognosis, when compared to luminal B (LB), HER2+, and basal-like (Basal). Ki67 by gene expression or immunohistochemistry (IHC) is commonly used as a proliferation index. The function of Ki67 in proliferation remains unknown. AURKA (STK15) is known to play an important role in mitosis, and is a component of the 21-gene recurrence score of the Oncotype Dx. With multiple platforms of molecular data available from hundreds of IBC tissues in The Cancer Genome Atlas project (TCGA), we sought to study the association of AURKA with different IBC subtypes and explore its use as a proliferation marker in IBCs. Methods: Gene expression (Agilent, log2 transformed), relative DNA copy number (CN, Affymetrix SNP 6.0), and exome sequence mutation (Illumina) data for 459 IBC cases were downloaded from the TCGA data portal. PAM50 classification results of all samples were obtained from the TCGA breast cancer AWG group and included 203 LA, 113 LB, 51 HER2+, 84 Basal-like, and 8 Normal-like which were not used in this study due to the low numbers. Kruskal-Wallis tests were used to evaluate the differences among four subtypes on AURKA expression and CN, followed by Wilcoxon Mann-Whitney test with Bonferroni adjustment for pairwise analyses. Pearson's Correlation Coefficient was used for correlation analyses. All statistical analyses were performed using SAS and R, and two-sided, p values <.05 were considered statistically significant. Results: There was a significant difference among IBC subtypes, in gene expression as well as in CN (p values < 0.0001). AURKA mRNA levels were significantly lower in LA (mean±SD, −2.61±0.63) compared to LB (−1.45±0.78), HER2+ (−1.38±0.61), and Basal (−1.26±0.62) subtypes (p values all < 0.0001). No significant difference was detected between other subtype pairs. In CN analysis, Basal (0.09±0.22) was lower than HER2+ (0.32±0.308, p < 0.0002) and LB (0.33±0.41, p < 0.0001), and LA (0.14±0.28) is lower than HER2 (p < 0.0016) and LB (p < 0.0001), but no other significant CN difference between the subtypes were found. The means and SDs are provided for reference only. No correlation of p53 mutation status and AURKA expression were observed. However, AURKA gene expression level is correlated with MKI67 gene expression (R = 0.69, p < 2.2e−16), and its correlation with PAM50 proliferation score is even higher (R = 0.80, p < 2.2e−16). Discussion: Using the TCGA data we observed that the mean gene expression level of AURKA is significantly lower in LA than the other IBC subtypes, by more than 50% (note the log2 transformation). This differential expression is not completely due to CN changes (especially for the Basal subtype). There is a strong association with other tumor cell proliferation markers such as the MKI67 gene and the PAM50 proliferation score. We are using computational and laboratorial studies to better understand the role of AURKA in the etiology of invasive breast cancers. The views expressed in this article are those of the author and do not reflect the official policy of the Department of Defense, or U.S. Government. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P2-05-21.