Molecular subtyping of breast cancer is necessary for therapy selection and mandatory for all breast cancer patients. Metabolic alterations are considered a hallmark of cancer and several metabolic drugs are currently being investigated in clinical trials. However, the dependence of metabolic alterations on breast cancer subtypes has not been investigated on -omics scale. Thus, 204 estrogen receptor positive (ER+) and 67 estrogen receptor negative (ER−) breast cancer tissues were investigated using GC–TOFMS based metabolomics. 19 metabolites were detected as altered in a predefined training set (2/3 of tumors) and could be validated in a predefined validation set (1/3 of tumors). The metabolite changes included increases in beta-alanine, 2-hydroyglutarate, glutamate, xanthine and decreases in glutamine in the ER− subtype. Beta-alanine demonstrated the strongest change between ER− and ER+ breast cancer (fold change = 2.4, p = 1.5E − 20). In a correlation analysis with genome-wide expression data in a subcohort of 154 tumors, we found a strong negative correlation (Spearman R = − 0.62) between beta-alanine and 4-aminobutyrate aminotransferase (ABAT). Immunohistological analysis confirmed down-regulation of the ABAT protein in ER− breast cancer. In a Kaplan–Meier analysis of a large external expression data set, the ABAT transcript was demonstrated to be a positive prognostic marker for breast cancer (HR = 0.6, p=3.2E-15).

It is well-known for more than a decade that breast cancer exhibits distinct gene expression patterns depending on the molecular subtype defined by estrogen receptor (ER) and HER2 status. Here, we show that breast cancer exhibits distinct metabolomics patterns depending on ER status. Our observation supports the current view of ER+ breast cancer and ER− breast as different diseases requiring different treatment strategies.

Metabolic drugs for cancer including glutaminase inhibitors are currently under development and tested in clinical trials. We found glutamate enriched and glutamine reduced in ER− breast cancer compared to ER+ breast cancer and compared to normal breast tissues. Thus, metabolomics analysis highlights the ER− subtype as a preferential target for glutaminase inhibitors.

For the first time, we report on a regulation of beta-alanine catabolism in cancer. In breast cancer, ABAT transcript expression was variable and correlated with ER status. Low ABAT transcript expression was associated with low ABAT protein expression and high beta-alanine concentration. In a large external microarray cohort, low ABAT expression shortened recurrence-free survival in breast cancer, ER+ breast cancer and ER− breast cancer.