Simplified schematic overview of lithium effects on GSK3β, eIF2B, and β-catenin. Lithium inhibits GSK3β activity. GSK3β inhibition lowers eIF2B phosphorylation (yellow) levels resulting in increased eIF2B activity and presumably downregulated expression of ATF4 and the ATF4-regulated transcriptome. GSK3β inhibition also stabilizes β-catenin by impeding its phosphorylation (yellow) at Ser33, Ser37, and Thr41. Phosphorylated β-catenin is highly unstable and was therefore not quantified. Non-phosphorylated β-catenin moves to the nucleus and joins with a member of the T-cell factor/lymphoid enhancer factor (TCF/LEF) family of transcription factors. This joining upregulates the expression of several genes, including EphB, Sox9, and Axin2. ➔ indicates stimulation; --| indicates inhibition. The image was created with BioRender.com.

Lithium improves motor deficits in mutant eif2b5^zc103/zc103 zebrafish. WT (circles) and eif2b5^zc103/zc103 (VWM; squares) zebrafish were treated with placebo (black; WT: n = 48, VWM: n = 38) or 1 mM lithium (blue; WT: n = 12, VWM: n = 26) and assessed on distance moved, moving time, and velocity at 7 dpf. The graphs show individual and mean values, +/− SEM. The shown motor behavior parameters differ significantly in placebo-treated WT versus placebo-treated eif2b5^zc103/zc103 zebrafish (p < 0.0001; not indicated, Supplementary Data Sheet S1). Treatment effects were statistically analyzed using an unpaired t-test or Mann–Whitney test (VWM, distance moved; Supplementary Data Sheet S1). *p < 0.05 and **p < 0.01.

Lithium induces more pronounced polydipsia in WT than in 2b4^he2b5^ho mutant mice. WT (circles) and 2b4^he2b5^ho mutant (VWM; squares) mice were i.p. injected with placebo (black) or 200 mg/kg lithium (blue) (n = 8 per treatment, per genotype). Water intake (ml/g) is shown per cage housing two mice (individual and mean values, +/− SEM). Statistical analysis was performed with a mixed-effects model with post-hoc Tukey’s multiple comparison test (Supplementary Data Sheet S1). ****p < 0.0001; placebo-lithium differences: ^____, WT-VWM difference: ----. Raw data are in Supplementary Data Sheet S2.

Lithium does not increase the expression and activity of non-phosphorylated β-catenin. WT (circles) and 2b4^he2b5^ho mutant (VWM; squares) mice (n = 8 per treatment, per genotype) were i.p. injected daily with placebo (black) or 200 mg/kg lithium (blue) for 1 week. Mice were terminated approximately 8 h after the final injection. Cerebellar RNA and protein samples were derived from the same lysate. Levels of non-phosphorylated β-catenin were determined using the Western blot (n = 3–4). Lithium treatment effects on indicated β-catenin-regulated markers were determined with qPCR (n = 6–8) (Sorcs1 and Nrxn as reference). Lithium differences were analyzed with two-way ANOVAs (Supplementary Data Sheet S1). The graphs show individual and mean values, +/− SEM. Raw data are available in Supplementary Data Sheet S2.

Lithium subtly increases the expression of the ATF4-regulated transcriptome in the mouse brain. WT (circles) and 2b4^he2b5^ho mutant (VWM; squares) mice were i.p. injected daily with 200 mg/kg lithium (blue) or placebo (black) for 1 week (n = 8 per treatment, per genotype). Mice were terminated approximately 8 h after injection. Cerebellar RNA and protein samples were derived from the same lysate. Lithium effects on indicated ISR markers were determined using the Western blot for eIF2α phosphorylation (n = 7–8) and qPCR for mRNA expression (Sorcs1 and Nrxn as reference, n = 8). The graphs show individual and mean values, +/− SEM. The shown ISR markers differ significantly in placebo-treated WT versus placebo-treated 2b4^he2b5^ho mice (p < 0.05; not indicated, Supplementary Data Sheet S1). Treatment effects were statistically analyzed using an unpaired t-test or Mann–Whitney test (VWM: Ddit3/Chop, Gadd34, Nupr; Supplementary Data Sheet S1). Raw data are available in Supplementary Data Sheet S2. *p < 0.05 and **p < 0.01.