Analytical challenges for measuring steroid responses to stress, neurodegeneration and injury in the central nervous system.

Levels of steroids in the adult central nervous system (CNS) show marked changes in response to stress, degenerative disorders and injury. However, their analysis in complex matrices such as fatty brain and spinal cord tissues, and even in plasma, requires accurate and precise analytical methods. Radioimmunoassays (RIA) and enzyme-linked immunosorbent assays, even with prepurification steps, do not provide sufficient specificity, and they are at the origin of many inconsistent results in the literature. The analysis of steroids by mass spectrometric methods has become the gold standard for accurate and sensitive steroid analysis. However, these technologies involve multiple purification steps prone to errors, and they only provide accurate reference values when combined with careful sample workup. In addition, the interpretation of changes in CNS steroid levels is not an easy task because of their multiple sources: the endocrine glands and the local synthesis by neural cells. In the CNS, decreased steroid levels may reflect alterations of their biosynthesis, as observed in the case of chronic stress, post-traumatic stress disorders or depressive episodes. In such cases, return to normalization by administering exogenous hormones or by stimulating their endogenous production may have beneficial effects. On the other hand, increases in CNS steroids in response to acute stress, degenerative processes or injury may be part of endogenous protective or rescue programs, contributing to the resistance of neural cells to stress and insults. The aim of this review is to encourage a more critical reading of the literature reporting steroid measures, and to draw attention to the absolute need for well-validated methods. We discuss reported findings concerning changing steroid levels in the nervous system by insisting on methodological issues. An important message is that even recent mass spectrometric methods have their limits, and they only become reliable tools if combined with careful sample preparation.