Application of protein expression profiling to screen chemicals for androgenic activity.

Protein expression changes can be used for detection of biomarkers that can be applied diagnostically to screen chemicals for endocrine modifying activity. In this study, surface enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF-MS) coupled with a short term fish assay was used to investigate changes in plasma protein expression as a means to screen chemicals for androgenic activity. Adult gravid female sheepshead minnows (Cyprinodon variegatus) were placed into separate aquaria for seawater control, ethanol solvent control, and the following androgen agonist treatments at 5.0μg/L: dihydrotestosterone (DHT), methyldihydrotestosterone (MDHT), testosterone (T), methyltestosterone (MT) and trenbolone (TB). Treatments of 0.6μg/L endosulfan and 40μg/L chlorpyrifos (CP) served as non-androgenic negative stressor controls. Test concentrations were maintained using an intermittent flow-through dosing apparatus supplying exposure water at 20L/h. Fish were sampled at 7 days, the plasma diluted, processed on weak cation exchange CM10 ProteinChip arrays and analyzed. Spectral processing resulted in 249 individual m/z peak clusters for the androgen exposed fish. Partial least squares-discriminant analysis was used to develop an androgen-responsive model using sample spectra from exposures with DHT and unexposed solvent control fish as the training set. The androgen classification model performed with ≥79% specificity (% true negative) and ≥70% sensitivity (% true positive) for non-aromatizable androgens. The aromatizable androgens T and MT were classified as androgenic with specificities of 42 and 79%, respectively. The reduction in sensitivity observed with T is thought to be caused by its metabolic conversion to an estrogen by aromatase. The results of these studies show diagnostic plasma protein expression models can correctly classify chemicals by their androgenic activity using a combination of high throughput mass spectrometry and multivariate approaches.