Biology of lithium response in bipolar disorder : genetic mechanisms and telomeres

Background: Bipolar disorder is a common, chronic and severe mental illness, causing suffering and large costs. Lithium treatment is the golden standard and works in 2/3 of patients, of which 50% are called lithium responders. There is strong evidence that both bipolar disorder and the degree of lithium response are highly heritable, although many mechanisms are unknown. Short telomere length has been found in both somatic and psychiatric disorders, but little is known about telomeres in bipolar disorder and nothing about telomeres and lithium treatment. A few studies limited in size have reported an increased risk of kidney cancer in lithium-treated patients. These findings have led to warnings and changes in policies for lithium treatment, although clinicians and researchers have disputed it. There is thus a clinical need for large-scale studies and reliable evaluations of the cancer incidence in lithium-treated bipolar patients. Aims: The overall aim of this thesis is to increase the knowledge of the genetic mechanisms in bipolar disorder and in lithium response. Specifically, it aims to find genetic associations to lithium response and to investigate how telomere length is related to bipolar disorder and lithium treatment. Additionally, it explores the overall and site-specific cancer incidence in bipolar disorder and lithium treatment. Methods: Study I is a GWAS of lithium response. Study II is a retrospective study of telomere length in lithium-treated bipolar patients. Study III is an association study of a genetic variant in the hTERT gene (previously associated with short telomeres) and depression, and a retrospective cohort study of telomere length in depression. Study IV is a longitudinal study of telomere length in lithium-treated bipolar patients. Study V is a register study of overall and specific cancer incidence in lithium-treated bipolar patients compared to the general population. Results: I) A single locus with four linked common gene variants on chromosome 21 coding for long, non-coding RNAs, which might be important for brain gene regulation, was associated with lithium response. II) Leukocyte telomeres were 35% longer in bipolar patients compared to healthy controls and correlated positively with length of lithium treatment in patients who had had lithium for more than 2.5 years. Lithium responders had 10% longer telomeres than non-responders. Short telomeres were associated with a larger number of depressive episodes. III) A genetic variant in the hTERT gene was associated with the number of depressions in bipolar type 1 patients responding to lithium and with unipolar depression. Telomere length was shorter in depressed patients without previous childhood trauma. IV) The total reduction of leukocyte telomere length between tests was, for bipolar patients, an average of 2.1% per year versus 3.7% per year in healthy controls. Age at baseline had a positive effect while total time on lithium between tests had a negative effect on the leukocyte telomere length reduction independent of other confounders. In bipolar patients there was no association between the total number of leukocytes or leukocyte subtypes and leukocyte telomere length at follow-up. V) There was no increase in the overall or site-specific cancer incidence in lithium-treated patients in the age span of 50-84 years (N=2 393) compared to the general population (N=2 593 011). Conclusions: Association of lithium response to a genomic region containing long, noncoding RNA with potential importance for gene regulation in the brain adds a new piece of knowledge to the heritability of lithium response in bipolar disorder. Results must be replicated and translated into a biological context. The new finding of a decelerating effect of lithium treatment on telomere shortening suggests that lithium might have operative effects on telomere biology, which also has potential importance for lithium response and should be investigated further. An important clinical implication of the lack of difference in cancer incidence between lithium-treated bipolar patients and the general population is that recently added warnings for renal cancer in patients with long-term lithium treatment is unnecessary and ought to be omitted from the current policies