Imaging the breast cancer metabolome

Breast cancer has the highest incidence and mortality of all malignant diseases among women globally. The heterogeneous nature of breast cancer causes variable clinical presentation and outcome. Prediction of breast cancer treatment response remains highly inadequate. Drugs targeting oncogenic signaling in breast cancer are in the pipeline, and tools for stratification of patients to these treatments are needed. Magnetic resonance (MR) technology is increasingly used in oncology its versatility and translatability. The various MR techniques allow acquisition of anatomical, functional, metabolic and molecular information. Still, bridging molecular features and MR imaging is a challenge. We therefore aim to develop methods for translation of cancer-specific metabolic traits into imaging biomarkers or novel drug targets. Abnormal phospholipid metabolism is frequently seen in breast cancer. Using MR, malignant breast lesions can be discriminated from benign lesions based on the concentration of choline-containing compounds. In this project, we study phospholipid metabolism in great detail, in order to better understand the metabolic differences between cancer cells and normal cells, and to identify potential therapeutic strategies based on metabolic characteristics. In a panel of patient-derived xenograft models and in biopsies from a cohort of breast cancer patients (Grinde MT et al, Breast Cancer Research 2014), we found clear differences in phospholipid metabolism between different molecular subgroups of breast cancer. One of the most significant differences was the overexpression of cytosolic phospholipase A2 in basal-like breast cancer. In collaboration with the biotech company Avexxin AS, we have shown that the cPLA2 inhibitor AVX235 inhibit tumor growth in a basal-like PDX model, primarily through an antiangiogenic effect (Kim E, Tunset HM et al, BMC Cancer 2016). We are also developing technology for improved and more detailed studies of tumor metabolism. Using this technology, we demonstrated how metabolic profiling can be used to monitor response to everolimus in a panel of 13 different PDX models. Using 31P MR spectroscopy, we have demonstrated that it is possible to discriminate between different breast cancer subtypes in vivo using metabolic biomarkers (Esmaeili M et al, Journal of MRI 2014). Furthermore, we have shown that these metabolic biomarkers can be used to detect response to treatment with PI3K inhibitors (Esmaeili M et al, Magnetic Resonance in Medicine 2014). These results indicate that it may be feasible to develop clinical 31P MR protocols for therapy monitoring, and we continue our work towards this goal. In previous work, we have demonstrated that the most frequent breast cancer subtype - hormone receptor positive breast cancer - has a distinct metabolic profile. In a collaboration with Karolinska Institutet, we have studied how the oestrogen receptor activates several key metabolic genes and thereby shown that the protein CHPT1 plays an important role in translating hormone stimulation to increased consumption of sugar and fat in cancer cells. Inhibiting this protein reduced cancer cell growth in vitro, and in animal models we also found an antimetastatic effect. The project generates large amounts of clinical MR images, which we have used to develop new methods for image analysis. Here, our work has resulted in novel techniques for correction of geometric distortion in MR images (Teruel JR et al, Magnetic Resonance in Medicine 2014), and we have demonstrated how multivariate description of intratumoral heterogeneity in contrast-enhanced MRI has predictive value in patients receiving neoadjuvant chemotherapy (Teruel JR et al, NMR in Biomedicine 2014). Together with New York University School of Medicine, we have explored the value of diffusion tensor imaging in breast cancer (Teruel JR et al, JMRI 2016), and demonstrated how diffusion-weighted MRI provides diagnostic information linked to microvascular perfusion (Teruel JR et al, Radiology 2016). An important aspect of the project is large-scale metabolomics studies in breast cancer. Using data from large cohorts, we explore how genetic, proteomic and metabolomic characteristics can be integrated and used to develop new diagnostic tools. In 2016, we demonstrated how metabolic fingerprints can be used to subclassify breast cancer, which emphasizes the heterogeneity of the disease and indicates that metabolic characteristics must be taken into account in future personalised treatment regimens. We have also implemented NMR protocols for analysis of serum samples, and recent work proves that the serum metabolic profile is associated with central clinical characteristics of breast cancer.

Breast cancer has the highest incidence and mortality of all malignant diseases among women globally. The heterogeneous nature of breast cancer causes variable clinical presentation and outcome. Prediction of breast cancer treatment response remains highl y inadequate. Drugs targeting oncogenic signaling in breast cancer are in the pipeline, and tools for stratification of patients to these treatments are needed. The use of magnetic resonance (MR) in oncology is steadily increasing due to its versatility and translatability. The various MR techniques allow acquisition of anatomical, functional, metabolic and molecular information. Still, bridging molecular features and MR imaging is a challenge. We therefore aim to develop methods for translation of canc er-specific metabolic traits into imaging biomarkers or novel drug targets. The main objective of this proposal is to develop and implement functional and molecular MR to establish prognostic and predictive biomarkers for personalized diagnosis and targe ted treatment in breast cancer. Thus, more efficient disease management and decision-making can be achieved, combined with reduced risk of negative side effects and significant economic savings. The research infrastructure is excellent, with access to r elevant model systems and large human breast cancer tissue biobanks, high resolution MR spectrometers and preclinical and clinical MR scanners in addition to a hybrid MRI-PET scanner. The high quality core facilities at NTNU offer the necessary services f or gene expression analyses and immunohistochemistry. Nationally and internationally acknowledged research groups will actively contribute to the project, and research stays for both PhDs are scheduled. The collaboration with K.G. Jebsen Centre for Breast Cancer Research and the international partners are well established and will foster the progress of the project.