We agree with the comments by Dr Underwood reinforcing that the mechanisms by which lithium may exert a neuroprotective effect in patients with amnestic mild cognitive impairment ^{Reference Forlenza, Diniz, Radanovic, Santos, Talib and Gattaz1} still must be clarified. The inhibition of glycogen synthase kinase-3 (GSK-3B) by lithium is a plausible effect, given its pivotal role in the pathogenesis of Alzheimer’s disease, but most likely not the only one. In addition to the prevailing mechanism of action involving the inhibition of inositol monophosphatase and downstream effects towards the up-regulation of autophagy, many other neurobiological effects have been attributed to lithium. These include the inhibition of apoptosis and the up-regulation of neurotrophic cascades. ^{Reference de Sousa, van de Bilt, Diniz, Ladeira, Portela and Souza2} The modification of these intracellular signalling systems by lithium has been shown to yield neurotrophic and/or neuroprotective effects, which have been consistently demonstrated in cell culture and animal models. These effects are probably unspecific and may be beneficial to patients with distinct psychiatric and neurodegenerative diseases, including bipolar disorder, ^{Reference Nunes, Forlenza and Gattaz3} amyotrophic lateral sclerosis ^{Reference Aggarwal, Zinman, Simpson, McKinley, Jackson and Pinto4} and Alzheimer’s disease. ^{Reference Forlenza, Diniz, Radanovic, Santos, Talib and Gattaz1}

We hypothesise that the inhibition of GSK-3B by lithium is more specific to processes that ultimately lead to the formation of neuritic plaques and neurofibrillary tangles. According to the ‘GSK3 hypothesis of Alzheimer’s disease’, overactive GSK-3B accounts for memory impairment, Tau hyperphosphorylation, increased beta-amyloid production and local plaque-associated inflammatory responses mediated by the microglia. ^{Reference Hooper, Killick and Lovestone5} The inhibition of GSK-3B is currently regarded as a candidate disease-modifying approach for the treatment and prevention of Alzheimer’s disease, and specific inhibitors such as NP-031112 are being tested in phase II clinical trials (www.clinicaltrials.gov). Therefore, lithium may contribute to the attenuation of the pathological process in Alzheimer’s disease through inhibition of GSK-3B, and deliver additional, unspecific benefits via modification of other signalling pathways that favour autophagy, preclude apoptosis and up-regulate the secretion of neurotrophic factors in the brain. Presumably, the interplay of complementary mechanisms is necessary to warrant clinically relevant benefits, which we were able to show in our study. ^{Reference Forlenza, Diniz, Radanovic, Santos, Talib and Gattaz1} We thus speculate that the effects of lithium on multiple homeostatic systems downstream from membrane receptor-based neurotransmission may in fact represent an advantage as a candidate drug for the treatment of complex neurobiological diseases. In our study, the doses of lithium used were very well tolerated. This, together with its wide availability and low cost, warrant the further investigation of the potential protective effect of lithium in Alzheimer’s disease.

Declaration of interest

Funding for the present work provided by Conselho Nacional de Pesquisa Científica (CNPq, Project ), Alzheimer’s Association (), FundaçãodeAmparoà Pesquisa do Estado de São Paulo (FAPESP, Project ) and Associação Beneficente Alzira Denise Hertzog da Silva (ABADHS).