The intensity of the total choline (tCho) signal in magnetic resonance spectroscopic imaging (MRSI) of tumors is heterogeneous. Magnetic resonance spectroscopy (MRS) studies have shown an elevation of phosphocholine (PCho) and total choline-containing metabolites (tCho) in breast cancer cells and tumors. In this study, we investigated the protein changes that occur in tumor regions with tCho in a breast tumor model by combining in vivo MRSI with ex vivo mass spectrometric imaging (MSI).

MDA-MB-231-HRE-tdTomato breast tumor xenografts, which express the tdTomato fluorescent protein under the control of hypoxia response elements, were grown in mice. tCho was detected by H1 MRSI in the tumors in vivo. Each tumor was cryo-sectioned into fiducially marked slices to perform MSI. On-tissue tryptic digest followed by matrix-assisted laser desorption ionization (MALDI) MSI was performed to detect the tryptic peptides. MRSI, MSI images were fused by using the fiducial markers. The tCho volume was segmented into high-tCho-containing and low-tCho-containing areas. 2499 high-tCho-containing voxels and 2500 low-tCho-containing voxels were selected randomly, followed by principal component analysis-linear discriminant analysis (PCA-DA) to classify high- and low-tCho-containing voxels through the MALDI MSI data. The MALDI MSI data between m/z 1000 and 3000 was analyzed and the loading spectrum of discriminant function (DF) was sorted to select candidate m/z peaks, which mostly contributed to the differentiation between high- and low-tCho-containing voxels. The molecular m/z peak candidates were identified by searching a peptide database with the resulting m/z values as keywords.

It was possible to separate low and high tCho spectra by this analysis. Peptides from hypoxia up-regulated protein 1 (HYOU1, m/z 1047.4) and peptides from ribosomal proteins such as m/z 1066.5, 1429.7 and 2824.4 displayed the highest peaks in the DF loading spectrum obtained from MALDI MSI data. tCho was previously found increased in the hypoxic regions of breast tumor xenografts. However, the HYOU1 peak at m/z 1047.4 was negatively correlated with tCho in MRSI data. HYOU1 is up-regulated by hypoxia and associated with poor prognosis in breast cancer. The relative decrease of HYOU1 protein in regions of high tCho is currently undergoing investigation in our lab. The molecular ions at m/z 1066.5, 1429.7 and 2824.4 were identified as ribosomal proteins that positively correlated with the tCho map in MRSI. This finding suggests that increased biosynthesis of ribosomal proteins may happen in high tCho regions in breast tumors. All other identified peptides are currently undergoing further validation by ion fragmentation using MSI-based MS/MS methods. By combining MRSI with MALDI MSI, we identified for the first time some specific proteins that are differentially expressed in breast tumor regions that contain high tCho. This work was supported by NIH R01 CA134695.

Citation Format: Lu Jiang, Kamila Chughtai, Tiffany Greenwood, Gert Eijkel, Ron Heeren, Kristine Glunde. Revealing protein biomarkers in breast tumor models by combining MRSI and MSI. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 2668. doi:10.1158/1538-7445.AM2013-2668