Lithium response in bipolar disorder correlates with improved cell viability of patient derived cell lines

Lithium is an effective, well-established treatment for bipolar disorder (BD). However, the mechanisms of its action, and reasons for variations in clinical response, are unclear. We used neural precursor cells (NPCs) and lymphoblastoid cell lines (LCLs), from BD patients characterized for clinical response to lithium (using the “Alda scale” and “NIMH Retrospective Life chart method”), to interrogate cellular phenotypes related to both disease and clinical lithium response. NPCs from two biologically related BD patients who differed in their clinical response to lithium were compared with healthy controls. RNA-Seq and analysis, mitochondrial membrane potential (MMP), cell viability, and cell proliferation parameters were assessed, with and without in vitro lithium. These parameters were also examined in LCLs from 25 BD patients (16 lithium responders and 9 non-responders), and 12 controls. MMP was lower in both NPCs and LCLs from BD; but it was reversed with in vitro lithium only in LCLs, and this was unrelated to clinical lithium response. The higher cell proliferation observed in BD was unaffected by in vitro lithium. Cell death was greater in BD. However, LCLs from clinical lithium responders could be rescued by addition of in vitro lithium. In vitro lithium also enhanced BCL2 and GSK3B expression in these cells. Our findings indicate cellular phenotypes related to the disease (MMP, cell proliferation) in both NPCs and LCLs; and those related to clinical lithium response (cell viability, BCL2/GSK3B expression) in LCLs.

PM brain cortex tissue from 15BD and controls Studied Hexokinase1 (HK1) attachment to outer mitochondrial membrane (OMM) in BD brain tissue Decreased HK1 attachment in BD compared to controls. HK1 attachment to OMM, a critical feature of brain energy metabolism and survival of neurons through prevention apoptosis. Gubert et al., 2013 [5] Plasma and PBMC samples from 12BD and 30HC.
Evaluated oxidative stress marker in plasma and ETC complex activities in PBMCs of BD No significant difference in oxidative stress markers and ETC complex activities between BD and HC.
de Sousa et al., 2015 [6] 24 HC and 25 BD; Patients treated with lithium for 6 weeks Evaluated leukocyte ETC complex activities in BD & effect of lithium on ETC No significant differences in mitochondrial ETC complex activities between BD and HC. Lithium treatment significantly increased mitochondrial complex I activities. Yoshimi et al., 2016 [7] CSF: 54 BD & 40 controls; PM brain tissue: 35BD & 34HC.
Evaluated the association of isocitrate with BD Iso-citrate levels significantly Increased in CSF from BD. mRNA levels of iso-citrate dehydrogenase in BD PM tissue was low compared to controls Scaini et al., 2017 [8] PBMC from 16BD and 16HC Analysed levels of mRNA, protein and activity of mitochondrial related factors in PBMCs of BD and HC.
Levels of anti-apoptotic proteins and citrate synthase activity were significantly lower, while caspase activity was higher in PBMC of BD. Levels of mRNA, protein related to mitochondria fusion were lower & related to fission were higher in PBMC of BD. Showed mitochondrial dynamics & cell death pathway activation in BD, supporting link between mitochondria & pathophysiology of BD. Bosetti et al., 2002 [9] Rat-Lithium treatment for 7 days (acute) or 42 days (chronic).
Gene expression analysis of lithium treated rat brain Chronic treatment at therapeutic concentration altered expression of several genes regulating mitochondrial enzymes in rat brain. Lai  VPA treatment improves cell viability and reduces cell apoptosis in cell exposed to TG. ER stress induced apoptosis response protein were inhibited by VPA treatment. VPA upregulated the ratio of BCL2/Bax proteins in SH-SY5Y cells. VPA promotes cell proliferation through PI3K, AKT, GSK3B pathways. Breen et al., 2016 [42] LCLs from 23 Caucasian individual (8 BD lithium responders, 8 BD lithium nonresponders, 7 HC) Exploring the effect of lithium 1mM (7 days) on transcriptome levels in LCLS from BD lithium response patients Differential gene expression in apoptosis signalling system, defence response, protein processing pathways and response to ER stress pathways were discovered on treatment with lithium.

McCurdy et al., 2006 [27]
Biopsies of olfactory mucosa from 8 BD & 10 HC Explored the cell proliferation rates in BD compared to HC No significant differences in mitosis between the BD and HC, however 11 genes involved in cell proliferation & 4 in neurogenesis were differentially expressed in BD. Cell death was significantly more in BD compared to HC. Hippocampus tissues from PM brain of 7 BD & 7 controls Gene expression profiling of brain tissues from BD and controls GAD67 (glutamate decarboxylase 67) gene expression was significantly decreased in BD than controls and CCND2 (Cyclin D2) is known to regulate GAD67. CCND2 was significantly downregulated in CA2/3 region of brain tissues in BD. Exploring the effect of lithium 1mM (7 days) on transcriptome levels in LCLS from BD lithium response patients Gene markers related to cell cycle and nucleotide excision repair were found to be differential in response to lithium between BD lithium responders and non-responders.    [13,14] Cell migration, proliferation [15,16], oxidative stress [17] NIPBL c.A4496C Novel 4 BD1/BD2 Notch pathway [18], Cohesion [19], Wnt and PI3K-AKT pathway [18] Cell migration, proliferation, apoptosis [20,21] SCUBE3 c.C1996T Novel 3 BD1/BD2 FGF Pathway [22] and hedgehog signaling [23] Cell proliferation [24] ANLN c.C128T rs575071809 3 BD1/BD2 PI3K pathway [25] Cell migration, Cell cycle [26,27] Supplementary Case-control study and antimanic treatment response (protein level) 1) 30 medication free BD manic subjects were compared with 30 healthy controls from Beijing Anding Hospital, China. 2) 47 BD (4 weeks) and 28 BD (8 weeks) subjects were analyzed for pre and post treatment with lithium, valproate and atypical antipsychotics To test the regulation of GSK3B in BD patients with manic episode and in response to treatmentexamined the protein level and the inhibitory serine phosphorylation of GSK3B in PBMCs of patients compared with healthy controls 1) The total protein levels of GSK3B was significantly higher in BD manic subjects than in healthy controls. 2) Phospho-Ser9-GSK3B was reported to be trend toward lower in bipolar manic subjects than in healthy controls.
3) Significant increase in phospho-Ser9-GSK3B was reported in 28 BD subjects post 4 weeks and 8 weeks treatment. 4) Total GSK3B among the 47 subjects was not significantly different at treatment.  Case-control study (protein level) 1) 21 BD patients and 21 HC from Chicago were investigated for GSK3B level in platelet, 2) The patients compared for GSK3B level before and after treatment (lithium or valproate or antipsychotics) for 8 weeks.
To explore the role of GSK3B in BD and in response to treatment 1) GSK3B protein level in cytosol and membrane fraction of platelets from BD were decreased in comparison to controls.
2) The protein level after 8 weeks of treatment was increased.
To compare the levels of GSK3B protein in brain tissues of BD and HCs Evaluated change in serine phosphorylation of GSK3B in PBMCs of BD and HC subjects at baseline and in vitro treatment with lithium (20mMol/L) for 1 hour 1) Basal level of phospho-Ser9-GSK3B was reported to be lowest in HC followed by 3fold increase in lithium free BD and highest in lithium treated BD subjects. 2) In vitro lithium treatment was also associated with elevation of phospho-Ser9-GSK3B level. 3 To test the association of the SNP with BD risk and therapeutic response to lithium treatment. To test the association of SNP with lithium treatment response.
No significant difference was reported in genotype and allele frequency of the SNP between lithium responders and nonresponders. However, haplotype blocks T-A and C-A (with another SNP [-1727A/T]) was reported to be associated with higher lithium response and lower lithium response respectively. Mitjans et al., 2015 [40] Pharmacogenetic association study Total 131 BD patients from Barcelona which included 26 excellent responders (ER); 62 partial responders (PR) and 43 non-responders (NR) based on lithium response.
To test the association of SNP with lithium treatment response.
No significant difference was reported in genotype and allele distribution between the lithium response groups. However haplotype rs1732170-rs11921360-rs34558 was associated with lithium response. The C-C-A haploblock was significantly less frequent in group of lithium partial and non-responders than excellent responders. NR1D1 Yang