3.1. Telomeres and Bipolar Disorder
The majority of studies published so far explored peripheral TL in BD, using genomic DNA extracted from leukocytes [47
], peripheral blood mononuclear cell (PBMC) [51
] or buccal smears [53
]. Overall, findings suggest that BD patients have a significant decrease of peripheral TL, compared with healthy controls [47
]. This effect, also confirmed by a recent meta-analysis [55
], seems to be independent from mood state, as shorter telomeres have been reported in BD patients in euthymic, depressed, and manic states [55
]. Shorter TL has also been associated with familial risk for BD [53
To date, only one study reported increased TL in patients with BD, compared with healthy individuals [48
]. However, in the study by Martinsson and coworkers, the large majority of patients were chronically treated with lithium at the time of recruitment. Several studies suggest that lithium may counteract telomere shortening, and as such, the discrepancy in findings might be explained by the effect of chronic exposure to lithium. The correlation between pharmacological treatments and TL will be further discussed in Section 4.1
of this review.
Several studies have also explored the correlation between TL and the clinical features of BD. Two recent studies compared early- or late-stage BD patients. The study by Kose Cinar and colleagues [54
], showed shorter LTL in late-stage BD patients compared to early-stage and healthy controls (p
< 0.001). Early-stage patients were defined as those with less than five episodes, while late-stage were those with more than 10 episodes. In another study, Barbè-Tuana et al. [56
] measured LTL in BD patients where the disease stage was characterized as I and IV, based on Kapczinski’s criteria [57
]. Findings showed that telomere shortening was already present in patients in the early stages of disease, compared to the controls, and reported no significant difference in LTL between early vs late stage. The meta-analysis of these two studies conducted by Huang et al. [55
] showed that although both groups of patients had significantly shorter LTL compared to the controls, late-stage patients had a further reduction of LTL compared to early-stage patients. However, these results should be carefully interpreted, due to the different methods used by the two studies to define the stage of disease, as well as due to the limited sample size.
Other clinical variables, such as the duration of illness [51
] and the number of lifetime hypomanic or manic episodes [47
] have not been found to correlate with TL in BD. However, contrasting findings have been reported regarding the number of depressive episodes. Some studies reported a negative correlation between LTL and number of depressive episodes [48
], while others found no association [47
]. Furthermore, no study reported an association between LTL and the number of suicide attempts [47
To the best of our knowledge, the only study investigating the relationship between BD and genetic variants in genes involved in telomere biology was carried out by Wei and coworkers [58
]. In this study, the minor allele of the human TERT (hTERT) polymorphism rs2736100 was associated with the number of depressive episodes in patients with BD type 1 who were good responders to lithium, but not in non-responders [58
]. On the other hand, polygenic risk scores for BD, MDD, or schizophrenia were not associated with peripheral TL in a cohort of healthy individuals [59
]. Further studies will be needed to understand if cumulative evaluations of the genetic risk for psychiatric disorders may be of help in identifying patients who are at a higher risk of accelerated aging.
While the large majority of published studies measured peripheral TL, few studies explored the putative association between BD and shorter TL in postmortem brains. However, findings from these studies so far do not support an association between BD and TL in cerebellar gray matter [60
], or in different brain regions, including the dorsolateral prefrontal cortex, hippocampus, amygdala, nucleus accumbens, and substantia nigra [61
]. Overall, the limited number of postmortem data on TL greatly limits our interpretation of the putative correlation between peripheral and brain TL in BD.
Interestingly, a recent study by Powell and colleagues [53
] suggested that peripheral TL might predict brain volume. In this study, TL measured on buccal DNA explained a substantial variance in hippocampal volume measured with magnetic resonance in a sample of bipolar patients, first-degree relatives, and unrelated healthy controls. This result supports the hypothesis that TL might represent a marker of hippocampal vulnerability, as previously suggested [62
]. A subsequent study further explored the association between peripheral TL and functional brain activation and connectivity, in a sample comprising patients with BD and first-degree relatives, as well as healthy volunteers [63
]. TL was positively associated with increased face-related activation in the amygdala, during a task in which participants were asked to identify facial emotions. This association was observed, regardless of the diagnosis status. Furthermore, a polygenic risk score for TL was positively associated with medial prefrontal cortex activation [63
]. These results support the existence of a link between TL and emotional brain activity.
3.2. Telomeres and Major Depressive Disorder
As in the case of BD, the majority of studies reported shorter TL in MDD. The most recent meta-analysis pooling results from 38 studies found that depression, as well as depression severity, were significantly associated with shorter TL (p
< 0.00001 and p
= 0.03, respectively) [64
]. Furthermore, Vance and coworkers [65
] conducted a prospective longitudinal study to evaluate the association between depression at baseline, and a change in LTL over two years [65
]. A diagnosis of MDD was found to prospectively predict LTL shortening after correcting for age, sex, and BMI, in a sample of 67 well-characterized MDD patients, and 50 healthy controls [65
]. However, LTL was not associated with a change of MDD symptoms severity or duration during the follow-up. A larger study published by Révész and coworkers in 2016, including 2750 participants of the Netherlands Study of Depression and Anxiety, showed that the relationship between psychopathology and LTL was modulated by levels of inflammatory markers (CRP and IL-6) as well as by waist circumference, triglycerides, high-density lipoprotein cholesterol, and cigarette smoking [66
]. Remarkably, this study provides significant support for the hypothesis of an interaction between factors involved in inflammation and telomere dynamics in mood disorders.
Few studies have investigated the role of genetic variants located in genes that are relevant for telomere biology. Wei and coworkers [58
] showed an association between the intronic hTERT rs2736100 single nucleotide polymorphism (SNP) and the diagnosis of depression among subjects with no experience of childhood adversity. Michalek and coworkers [67
] used a different approach and applied Mendelian randomization to investigate the association of TL and risk for recurrent MDD in a large UK sample, including 1628 MDD cases and 1140 controls. The authors found that the T allele of rs10936599, which is located upstream of the human telomerase RNA component (hTERC) gene, and which is associated with shorter LTL, increased the risk for childhood-onset MDD relative to controls or to adult-onset MDD. These results suggest that the genetic risk for shortened LTL is able to increase the risk for MDD, and specifically for childhood-onset MDD [67
]. Conversely, a recent study did not find any association between a polygenic risk score, including nine SNPs that were previously associated with inter-individual variation in TL [17
] and the risk of lifetime depression in a large sample of 17,693 female participants of European ancestry from the Nurses' Health Study (NHS) [69
]. Potential explanations of the discrepancies observed among the studies might be found in the different definitions used to define MDD, as well as in our still-limited knowledge of genetic variants that are able to explain inter-individual variability in TL.
As in the case of BD, few studies have investigated the interplay between peripheral TL and either TL in brain or brain volume. A recent study conducted by Wolkowitz and coworkers [62
] found a positive association between hippocampal volume and PBMC telomerase activity, but not TL, in a sample including 25 un-medicated patients with MDD [62
]. In a postmortem brain study, Mamdani and coworkers [61
] showed a significant telomere shortening in the hippocampus of MDD subjects compared with controls [61
]. Importantly, single-cell populations in the brain might be differentially vulnerable to telomere shortening, as suggested by a study showing reduced TL in oligodendrocytes, but in not astrocytes derived from post-mortem prefrontal cortex brain samples from patients with MDD, compared to healthy controls [70
]. Thus, future investigations on telomere shortening in the brain should consider accounting for the effects of different cell types in their experiments.