The mechanisms of lithium action are not known in detail. First messengers, second messengers, and gene expression all appear to be involved: the wide breadth of targets makes the lithium therapeutic path difficult to disentangle. In the present paper, we focused on the most direct biochemical lithium targets at therapeutic concentration, for which some pharmacogenetic finding is present (i.e. inositol mono phosphatase (IMPase), inositol polyphosphate-1-phosphatase (IPPase) and glycogen sinthase kinase 3 beta (GSK-3 beta)). They are all inhibited by lithium at therapeutic concentrations and are representative of the inositol depletion and of the GSK-3 beta based theories of lithium action. Then we surveyed gene variants on those targets that have been associated also with bipolar disorder. On the basis of the molecular characteristics of these proteins, we suggest a set of critical genetic variations. IMPase2, IPPase and GSK-3 beta gene appear to be good candidates for the analysis of lithium prophylactic efficacy and bipolar disorder phenotypes but the genetic association analysis conducted so far reported negative or not unequivocal finding. This may be due to the incomplete coverage of gene mutations in most studies or to the several actions that lithium is thought to perform and trigger in cell machinery, including receptors, calcium equilibrium, gene expression, activation of neuroprotective paths and other yet undetected or less considered mechanisms.