Diabetes mellitus is a chronic disease leading to renal and cardiovascular complications associated with high personal and societal costs. Mouse models are routinely used to assess the mechanisms of disease and disease status, by monitoring biomarkers, including 8-isoprostane, which is a marker of oxidative stress. Here we report that transflection FTIR spectroscopy coupled with partial least squares regression is able to predict urinary 8-isoprostane concentrations in various murine models of type 1 and type 2 diabetes. Regression analysis produced linear correlations, with R2 values of correlation between 0.6 and 0.8, and RMSECV (%) values between 12.3 and 29.4%, between known (measured by enzyme-linked immunosorbent assay) and predicted urinary 8-isoprostane concentration, based on the fourth derivative spectra. This approach has significant advantages over the current methods utilised, namely it has higher throughput and requires minimal sample preparation, compared to enzyme-linked immunosorbent assay and gas- or liquid-chromatography coupled with mass spectrometry. Furthermore, it has the potential to be expanded into analysis of other biomarkers and human samples in a point-of-care context.
- Fourth derivative spectroscopy
- Partial Least Squares Regression
- Transflection Fourier Transform IR spectroscopy