Rats were treated with lithium administered either via the food or by intraperitoneal injection. Lithium administration via the food results in a rather stable serum lithium concentration, whereas lithium injection results in a varying serum lithium concentration whereby a sharp increase shortly after the injection is followed by an exponential decline until the next injection (Plenge et al. 1981). After 5 months of lithium treatment the 5HT transport protein, the beta-adrenergic receptor and several other brain proteins were determined. The 5HT transport protein, labelled with [3H]paroxetine, was found to be decreased in the lithium-injected rats (Bmax = 347 fmol/mg protein) but was unchanged in the lithium-fed rats (Bmax = 389 fmol/mg protein), as compared with control rats (Bmax = 396 fmol/mg protein), and therefore probably is a specific effect only seen with varying lithium concentration. In contrast, the neuronal membrane marker protein D3 was decreased in the lithium-fed rats (88% of the control value), and showed a trend towards decrease in the lithium-injected rats. The decrease in D3 in the lithium-fed rats may indicate some neuronal damage due to the continuous presence of lithium. This damage may be more pronounced than in rats, where periods of low lithium concentration enable repair to take place. The beta-adrenergic receptor and the neural cell adhesion molecule NCAM were unaffected by the different lithium treatment regimens. Lithium has been reported to inhibit the 5HT1B receptor (the serotonin autoreceptor).(ABSTRACT TRUNCATED AT 250 WORDS)