187 OVARIAN CYCLE LIPID DYNAMICS REVEALED BY DESI-MS IMAGING AND MORPHOLOGICALLY-DRIVEN MULTIVARIATE STATISTICS

Abstract

Understanding the role of lipid metabolism in ovarian physiology is crucial for the progression of reproductive biotechnology. The aim in this work was to explore the lipid composition and dynamics of ovarian tissue, specifically the stroma, follicles, and corpora lutea. Desorption electrospray ionization–mass spectrometry (DESI-MS), an ambient ionization technique, was applied in this investigation, acquiring chemical and spatial information simultaneously. A morphologically-friendly solvent, dimethylformamide-acetonitrile (1 : 1), was used for DESI-MS imaging which allowed for ovarian lipid characterisation and subsequent staining (hematoxylin and eosin) providing morphological information. By this approach, regions-of-interest (ROI) were selected from bovine (n = 8), swine (n = 3), and mice (n = 5) ovaries (including pre-pubescent and cycling adults) based on the stained morphological structures. ROI for stroma (n = 54), follicles (n = 89), and corpora lutea (n = 61) were selected and chemically profiled. Tissue sections (20 μm) were thaw mounted onto glass microscope slides and stored at –80°C until analysis. A linear ion trap mass spectrometer equipped with a custom DESI-MS imaging stage was operated in the negative ion mode (m/z 200 to 1000). A 300 × 300 µm pixel size was used in DESI-MS imaging of ovarian tissue. Hyperspectral DESI images were reconstructed and processed by principal component analysis (PCA) that allowed visualisation of relationships among spatial (i.e. morphology) and chemical features. Ions indicated by PCA were analysed using univariate analysis (ANOVA), supporting the significance of particular lipids between morphological structures, e.g. adrenic acid (P = 1.7 × 10^–8) and m/z 836 (P = 8.9 × 10^–9) between corpora lutea and follicles. All morphological structures could be differentiated by multivariate statistics (>90% prediction rate) independent of the species, indicating conserved lipid constitution. Smaller differences in the lipid profiles were noted between species, poly-ovulatory and mono-ovulatory species, and reproductive maturation. A large variety and abundance of lipids was observed in corpora lutea and follicles, where steroidogenesis is a prominent physiological activity. Additional insight into ovarian physiology was gained with the detection of arachidonic and adrenic acid. The spatial relationship of arachidonic and adrenic acid with the corpora lutea – the former is a known prostaglandin precursor and key signalling molecule in steroidogenesis regulation and the latter is metabolized in the prostaglandin pathway by the same enzymes – suggests the latter may also have a role in steroidogenesis regulation, previously unseen in ovarian physiology. DESI-MS imaging with morphologically-driven statistical analysis proved efficient in relating and interpreting the chemical and morphological features. This methodology can by further applied to unravel complex ovarian-related physiological mechanisms and to other physiological and physiopathological models.