Search

The pathophysiology of bipolar disorder remains to be elucidated and there are no diagnostic or prognostic biomarkers for the condition. In this explorative proteomic study, we analyzed 201 proteins in cerebrospinal fluid (CSF) from mood stable bipolar disorder patients and control subjects sampled from two independent cohorts, amounting to a total of 204 patients and 144 controls. We used three Olink Multiplex panels, whereof one specifically targets immune biomarkers, to assess a broad set of CSF protein concentrations. After quality control and removal of proteins with a low detection rate, 105 proteins remained for analyses in relation to case–control status and clinical variables. Only case–control differences that replicated across cohorts were considered. Results adjusted for potential confounders showed that CSF concentrations of growth hormone were lower in bipolar disorder compared with controls in both cohorts. The effect size was larger when the analysis was restricted to bipolar disorder type 1 and controls. We found no indications of immune activation or other aberrations. Growth hormone exerts many effects in the central nervous system and our findings suggest that growth hormone might be implicated in the pathophysiology of bipolar disorder.

The etiopathology of bipolar disorder is largely unknown. We collected cerebrospinal fluid (CSF) samples from two independent case-control cohorts (total n = 351) to identify proteins associated with bipolar disorder. A panel of 92 proteins targeted towards central nervous system processes identified two proteins that replicated across the cohorts: the CSF concentrations of testican-1 were lower, and the CSF concentrations of C-type lectin domain family 1 member B (CLEC1B) were higher, in cases than controls. In a restricted subgroup analysis, we compared only bipolar type 1 with controls and identified two additional proteins that replicated in both cohorts: draxin and tumor necrosis factor receptor superfamily member 21 (TNFRSF21), both lower in cases than controls. This analysis additionally revealed several proteins significantly associated with bipolar type 1 in one cohort, falling just short of replicated statistical significance in the other (tenascin-R, disintegrin and metalloproteinase domain-containing protein 23, cell adhesion molecule 3, RGM domain family member B, plexin-B1, and brorin). Next, we conducted genome-wide association analyses of the case-control-associated proteins. In these analyses, we found associations with the voltage-gated calcium channel subunit CACNG4, and the lipid-droplet-associated gene PLIN5 with CSF concentrations of TNFRSF21 and CLEC1B, respectively. The reported proteins are involved in neuronal cell-cell and cell-matrix interactions, particularly in the developing brain, and in pathways of importance for lithium's mechanism of action. In summary, we report four novel CSF protein associations with bipolar disorder that replicated in two independent case-control cohorts, shedding new light on the central nervous system processes implicated in bipolar disorder.