Fourier transform infrared spectroscopy coupled with machine learning classification for identification of oxidative damage in freeze-dried heart valves.

Freeze-drying can be used to ensure off-the-shelf availability of decellularized heart valves for cardiovascular surgery. In this study, decellularized porcine aortic heart valves were analyzed by nitroblue tetrazolium (NBT) staining and Fourier transform infrared spectroscopy (FTIR) to identify oxidative damage during freeze-drying and subsequent storage as well as after treatment with H2O2 and FeCl3. NBT staining revealed that sucrose at a concentration of at least 40% (w/v) is needed to prevent oxidative damage during freeze-drying. Dried specimens that were stored at 4 °C depict little to no oxidative damage during storage for up to 2 months. FTIR analysis shows that fresh control, freeze-dried and stored heart valve specimens cannot be distinguished from one another, whereas H2O2- and FeCl3-treated samples could be distinguished in some tissue section. A feed forward artificial neural network model could accurately classify H2O2 and FeCl3 treated samples. However, fresh control, freeze-dried and stored samples could not be distinguished from one another, which implies that these groups are very similar in terms of their biomolecular fingerprints. Taken together, we conclude that sucrose can minimize oxidative damage caused by freeze-drying, and that subsequent dried storage has little effects on the overall biochemical composition of heart valve scaffolds.

Spectroscopic assessment of oxidative damage in biomolecules and tissues.

Oxidative damage is one of the main causes of cryopreservation injury compromising the use of cryopreserved biospecimens. The aim of this study was to evaluate the use of Fourier transform infrared spectroscopy (FTIR) as a non-invasive method to assess changes in biomolecular composition and structure, associated with oxidative stress in isolated biomolecules, acellular heart valve tissues, and ovarian cortex tissues. FTIR spectra of these specimens subjected to various treatments (H2O2- and Fenton-treatment or elevated temperatures) were vector normalized and selected spectral regions were analyzed by principal component analysis (PCA). Control and damaged biomolecules can easily be separated using PCA score plots. Acellular heart valve tissues that were subjected to different levels of oxidative damage formed separate cluster in PCA score plots. In hydrated ovarian tissue, large variation of the principal components was observed. Drying the ovarian tissues samples resulted in improved cluster separation of treatment groups. However, early signs of oxidative damage under mild stress conditions could not be detected by PCA of FTIR spectra. For the ovarian tissue samples, the standardly used nitro blue tetrazolium chloride (NBT) assay was used to monitor the amount of formazan production, reflecting reactive oxygen species (ROS) production at various temperatures. At 37 °C, formazan staining rapidly increased during the first 30 min, and then slowly reached a saturation level, but also at lower temperatures (i.e. 4 °C) formazan production was observed. In summary, we conclude that ATR-FTIR combined with PCA can be used to study oxidative damage in biomolecules as well as in tissues. In tissues, however, sample heterogeneity makes it difficult to detect early signs of oxidative damage.