Background: Lithium remains one of the most effective treatment options in individuals with bipolar disorder (BPD). However, the clinical response to lithium is highly variable, with some patients experiencing significant benefits while others showing little to no improvement. Despite advances in understanding the genetic determinants of lithium response through genome-wide association studies (GWAS), the downstream cellular mechanisms remain incompletely understood. Recent research has highlighted immune-mediated processes and immune cells as key players and potential biological markers in the pathophysiology of BPD and treatment response, where both genetic and environmental factors likely play important roles. To address this, this study focused on a comprehensive investigation of immune cell behaviour in lithium treated peripheral blood mononuclear cells (PBMCs) of individuals with BPD.
Methods: In a first step, FUMA was used to understand the intersection of genomic loci highlighted by a GWAS of lithium response in BPD (Hou et al., 2016). In a second step, this study capitalizes on a whole transcriptome dataset from PBMCs of individuals with BPD. PBMCs were collected from 149 BPD subjects in whom the BPD diagnosis was made in line with DSM-IV criteria and response to treatment with lithium as a lifetime measure was judged using the Alda-Scale. 21 individuals with the most extreme response profiles (n=10 non-responders and n=11 responders) were selected and PBMCs were cultivated with and without lithium. Bulk RNA-Seq was performed on the Illumina platform. Transcriptomic data were normalized, log-transformed, and differentially expressed genes (DEGs) were calculated using the DESeq2 package in R. DEGs between lithium responders and non-responders were identified and gene set enrichment analysis (GSEA) was performed. SingleR was employed to deconvolute the transcriptomic profiles and to identify immune cell populations.
Results: Among the top 10 FUMA-nominated genes, 4 genes (SLIT2, AEN, PHF20, and SCAND1) have reported roles in immune function, and Reactome-based infection phenotypes and histone modification were significantly enriched. DEGs of lithium responders vs non-responders intersected with a list of immune-related genes (n=3714, Calvano et al., 2005) revealed 15 common genes, 13 of which were significantly downregulated (ATP5I, CR1, DISC1, DTX4, FLT3LG, HIST2H2BE, NDUFA3, NPHP3, ORM2, RAB3D, RSC1A1, THRA, and TIAF1) in lithium responders with BPD. Moreover, GSEA revealed enrichment in inflammatory pathways, such as inflammatory response, cytokine receptor binding, interleukin-1 beta production and its regulation, regulation of interleukin-1 production, and regulation of inflammatory response. Furthermore, deconvolution analysis of DEGs identified 36 cell types, predominantly representing immune cell populations. Interestingly, we noted a significant downregulation of identified immune cell populations in lithium responders.
Discussion: This study explored the transcriptional dynamics within PBMCs in patients with BPD in the context of lithium response. Through comprehensive transcriptomics analysis and deconvolution techniques, we have identified distinct gene expression pattern linked to immune-specific cellular responses in lithium response. We noted a significant downregulation of immune-related genes, pathways, and cell types in lithium responders, suggesting a potential immune modulatory mechanism upon lithium treatment.