Telomere Shortening and Psychiatric Disorders: A Systematic Review

Telomeres are aging biomarkers, as they shorten while cells undergo mitosis. The aim of this study was to evaluate whether psychiatric disorders marked by psychological distress lead to alterations to telomere length (TL), corroborating the hypothesis that mental disorders might have a deeper impact on our physiology and aging than it was previously thought. A systematic search of the literature using MeSH descriptors of psychological distress (“Traumatic Stress Disorder” or “Anxiety Disorder” or “depression”) and telomere length (“cellular senescence”, “oxidative stress” and “telomere”) was conducted on PubMed, Cochrane Library and ScienceDirect databases. A total of 56 studies (113,699 patients) measured the TL from individuals diagnosed with anxiety, depression and posttraumatic disorders and compared them with those from healthy subjects. Overall, TL negatively associates with distress-related mental disorders. The possible underlying molecular mechanisms that underly psychiatric diseases to telomere shortening include oxidative stress, inflammation and mitochondrial dysfunction linking. It is still unclear whether psychological distress is either a cause or a consequence of telomere shortening.

The length of the telomeres varies with the type of cell. Nevertheless, the TL among different types of cells often shows a positive association, which means that a reduction in leukocyte telomere length usually means that other types of cells in the same body probably also have a related reduced value. This is the reason why white blood cells are used as biomarkers in clinical studies. Surprisingly, race must be taken into account, as people of African ancestry have been found to possess longer telomeres lengths [5].
During cell replication, the enzymes that duplicate genetic material are unable to perform their activity all the way to the end of the DNA [6]. Since chromosomes get shorter after each cell division, the main function of telomeres is to allow the cells to divide without losing their coding DNA regions [6]. When telomeres get too short, cells usually undergo

Inclusion and Exclusion Criteria
Study inclusion criteria were as follows: (i) Original papers; (ii) Written in English; (iii) Conducted on human subjects with proper control groups; (iv) Clearly evaluated TL on patients diagnosed with depression, anxiety or PTSD; (v) Published before 3 December 2020. Studies without a control group or those using animal models were excluded.

Study Selection
After an initial abstract/title screening, two independent authors (PAP and RMS) selected potential articles on the PubMed database and two independent authors (PHMM and TSCM) selected articles from the Cochrane Library and ScienceDirect databases. These initially screened studies were fully analyzed independently by the four authors (PAP, RMS, PHMM and TSCM) and any discrepancies related to the final decision were discussed among them. If a consensus was not achieved, then we referred to a fifth author (ACSS). Our study selection process counted with assistance from the Rayyan website and a mobile application [13].

Data Extraction
PAP, PHMM and RMS individually extracted data from the selected studies and an independent author (BM) reviewed all selected studies and the table to avoid any errors in data extraction. 1. Study identification: first author, publication date, country and type of study; 2. Participants: age, gender and sample size; 3. Variables: main comorbidities and lifestyle factors associated, race/ethnicity, psychiatric diseases (including severity and methods for diagnosis), medication and telomere measurement method.

Quality Assessment
All the included studies in this systematic review were evaluated according to the Newcastle-Ottawa Quality Assessment Scale (NOS) [14]. Studies were categorized into case-control studies, cohort studies and cross-sectional studies. Interestingly, the NOS guidelines assess only case-control studies and cohort studies-which was unfortunate, considering the high number of cross-sectional studies that we retrieved. Therefore, the authors designed an adapted protocol for this category using a NOS-adapted scale as a reference (Appendices A-C) [15]. All studies were scored by a "star-system", in which the higher the number of stars, the better the quality. Each category of evaluation has a pre-established maximum quantity of possible stars: (i) Selection quality may sum up to four stars total; (ii) Comparability totalizes two stars only; (iii) Exposure and outcome may sum to three stars each. Our full quality assessments are displayed in Tables S1-S3.
TL measurement unit was mean TL (mTL) in 20 studies, relative telomere length (RTL) in other 20 studies and salivary telomere length (STL) in one study. One study did not specify the units. Source tissues were: (i) Leukocytes (35 studies); (ii) Saliva samples (3 studies); Blood (1 study); White matter oligodendrocytes (1 study); Brain tissue (1 study); Tissue not specified (1 study). These findings are summarized in Table 1.

Telomere Lenght
Garcia-Rizo et al. [       Conflicting results around the influence of the duration MDD on TL were also reported. In general, TL was associated with major depression disorder (MDD). There were some conflicts as to whether the duration of MDD had an important role on TL. While some studies suggested [23,24,40,41] that shorter TL was a consequence of long-term exposure to MDD, other studies reported that adolescents with MDD had shorter TL [50] and that only MDD younger adults had shorter TL [31][32][33][34][35][36].

TL and PTSD
Thirteen studies assessed TL on 5237 PTSD patients [22,54,[57][58][59][60][61][62][63][64][65][66][67] (Table 2). Nine studies were cross-sectional articles, two studies were case-control articles and two studies were longitudinal articles. Six studies (3980 patients) established a negative association between TL and PTSD [60,[62][63][64][65]67], three studies (649 patients) did not observe association of TL and PTSD [22,54,57], one study (128 patients) found a positive association between TL and PTSD [61] and three studies (480 patients) showed either conflictual results or provided association with other variables [58,59,66]. The TL measurement unit was R in eight studies and mTL in three studies. Source tissues were leukocytes in 12 studies and peripheral blood in one study.    Subjects with PTSD showed shorter telomere length and larger amygdala volume than those without PTSD among veterans exposed to high trauma, while there was no significant group difference in these parameters among those exposed to low trauma.   Adjustment for health behaviors and medical conditions did not attenuate this association. Trauma type was not associated with TL and did not account for the association of PTSD with TL.       List of abbreviations: mTL-mean telomere length; RTL-relative telomere length; PCR-Polymerase Chain Reaction; qPCR-Quantitative PCR; RT-PCR-real-time PCR; rQ-PCR or RT-qPCR-real-time quantitative PCR; TL-telomere length; LTL-leukocyte telomere length; PBMC-peripheral blood mononuclear cell; BMI-body mass index, mean (SD); TC/HDL-C-total cholesterol/ high-density lipoprotein-cholesterol; NSAID-nonsteroidal anti-inflammatory drug; MDD-major depressive disorder; G.E.D-General Educational Development. The studies that did not find any association between TL and PTSD were conducted in military personnel [22,54,57,61]. The association between trauma severity with or without PTSD diagnosis, hostility, early trauma, global psychopathological severity and TL were also observed [22,[57][58][59]66].

Quality Assessment
A total of 19 studies were assessed by the guidelines of the NOS scale for case-control studies [14]. Regarding the selection quality assessment, only two studies scored four stars ( * * * * ), eight studies scored three stars ( * * * ), six studies scored two stars ( * * ) and three studies scored only one star ( * ). Regarding our comparability quality assessment, 13 studies scored two stars ( * * ) and six studies scored one star ( * ). According to our outcome quality assessment, six studies scored three stars ( * * * ), 11 studies scored two stars ( * * ) and two studies scored one star ( * ). Most of these studies did not discuss the representativeness of the presented cases and did not include information regarding the recruitment of controls. All case-control studies scored at least one one-star score.
The NOS scale for cohort studies was applied to assess the quality of nine longitudinal studies [14]. In the selection quality assessment, three studies scored four stars ( * * * * ), five studies scored three stars ( * * * ) and one study obtained two stars ( * * ). Our comparability quality assessment classified six studies as two stars ( * * ) and three studies as one star ( * ). Next, about our outcome quality assessment, four studies scored three stars ( * * * ), four studies scored two stars ( * * ) and one study scored one star ( * ). Again, the lack of a discussion about the representativeness of the psychiatric groups was one of the main issues. Additionally, many studies failed at reporting the TL at the beginning of the studies. All longitudinal studies scored at least one one-star score across the categories.
Twenty-five cross-sectional studies were assessed for quality by our adapted version of the NOS scale for cross-sectional studies [15]. Regarding the selection quality assessment, seven studies scored four stars ( * * * * ), seven studies obtained three stars ( * * * ), seven studies scored two stars ( * * ) and four studies scored only one star ( * ). Our comparability quality assessment classified 20 studies as two stars ( * * ) and five studies as one star ( * ). Lastly, our outcome quality assessment classified eight studies with three stars ( * * * ), and 17 studies with two stars ( * * ). Several studies had small sample sizes and did not provide a proper description of non-respondents. The lack of confidence interval analyses was also another limitation. All cross-sectional studies scored at least one one-star score across the categories. Tables S1-S3 present our full quality assessment.

Discussion
In the present study, we reviewed whether TL is associated with distress-related psychiatric disorders (depressive and anxiety disorders and PTSD). The potential association between anxiety, depression and PTSD with telomere shortening should consider which features and mechanisms were common among these disorders to result in telomere erosion. The common aspect of these disorders is psychological distress, which includes the experience of suffering, and behavioral, emotional and thought problems, including the cultural varieties of the psychiatric diagnosis in diverse cultures. Under these assumptions, we discuss the common features and the mechanisms potentially impacting telomere length.

Risk Factors and Telomere Length
Many studies identified an association between distress and telomere shortening. Distress during pregnancy catalyzes telomere erosion in the fetus-and predisposes telomere erosion during childhood and adulthood [71]. Additional risk factors for telomere erosion include socioeconomic vulnerability [72], obesity [73] and smoking [74]. Interestingly, adulthood cellular senescence-and subsequent telomere erosion-was associated with early childhood trauma (<5 years of age) [75][76][77]. Early telomere shortening, due to distress, possibly leads not only to the emergence of chronic diseases but also to mental illnesses [78,79].
The cumulative effect of risks explains why telomere shortening can be connected with early life stress vulnerability through affecting telomerase activity, which can possibly influence the later emergence of chronic disease [78] and mental illness, which is, in turn, associated with accelerated cellular senescence [79]. In this context, active and healthy lifestyles, including healthy diets [80], adequate sleep [81] and meditation [82] may attenuate age-related TL shortening [83,84]. Accordingly, longer leukocyte telomeres are associated with longer lifespans [85].

Psychological Distress-Related Diseases and TL
TL has been associated with psychiatric disorders [10,[90][91][92][93][94]. Darrow and colleagues observed a robust association between psychiatric diseases and shorter telomeres in their meta-analysis, including significant size effects for PTSD, depression and anxiety [90]. Three additional meta-analyses reported an association between short telomeres and depression [91][92][93]. Finally, one meta-analysis observed an association between PTSD and shortened TL while another meta-analysis reported an association between anxiety and shorter telomeres [10,94].
Puterman and Etel [95] proposed a lifespan model of stress-induced cell aging. Stressors mediate the behavioral and psychosocial traits of individuals, establishing two features: multisystem vulnerability, which leads to faster aging and shorter cellular lifespan, and multisystem resilience, which results in slower aging and longer lifespan [95].
Even though the relationship between psychiatric disorders and telomere shortening is well-established in the literature, the studies neither assessed the nature of this association nor investigated variables. In this regard, the present review aimed not only at investigating whether telomere length is associated with psychological distress, but also at identifying which variables (such as molecular mechanisms) might underpin this association.

MDD and TL
MDD is currently described as persistent sadness and a lack of interest or pleasure in previously rewarding and enjoyable activities. It is estimated that this disorder affects over 322 million people worldwide, considered one of the leading causes of disability and increasing in prevalence [96,97]. Depression and other mental health disorders are raising awareness worldwide due to increasing prevalence rates, as the total estimated number of people living with depression increased by 18.4% between 2005 and 2015 [97].
MDD has a complex etiology associated with a multidimensional nature including emotional, behavioral, physical and cognitive aspects [98]. MDD has clinically heterogeneous patterns and associations with other disorders, causing methodological limitations to the consistency, coherence, validity and utility of research findings [99]. Therefore, several articles found in our systematic search showed conflicting results and the main implications and limitations will be later discussed.
While clinical information provides paradoxical and heterogeneous aspects of MDD, cellular and molecular mechanisms of the disease are advancing towards a unified model for understanding depression [100]. In patients with MDD, the corticolimbic system is deregulated and molecular pathways related to inflammatory mediators, oxidative stress, DNA damage, DNA repair and mitochondrial dysfunction might interact in the pathophysiology of the disorder [10,100]. In this sense, the study of Dean and Keshavan [101] suggests that depression has several endophenotypes, associated with distinct pathophysiological mechanisms, which can explain complex and variable clinical features. However, the authors state that these mechanisms have reciprocal interactions [101]. Therefore, molec-ular mechanisms of depression represent an important aspect of the disease, acting not only as causative agents of the disease but also as biomarkers of depression.
Patients with MDD patients are at greater risk for developing aging-related somatic conditions [25,102]. As mentioned, TL is a biomarker of aging, being associated with various aging-related somatic diseases [7,103]. Hence, it is reasonable to hypothesize that the shortening of TL might be associated with MDD. Our systematic search found several articles that assessed the association between TL and MDD symptoms. Despite most articles and even a meta-analysis supported the association between shorter telomeres and MDD, the heterogeneity of the methodology among the studies precludes a conclusion [90][91][92][93].
There are very few studies on the nature of this relationship. Findings from the Heart and Soul Study [39] suggest that multisystem resilience, stronger social connections, greater physical activity and better sleep quality modulate the association between telomere shortening and depressive disorders. Hence, MDD-related biological outcomes should be analyzed under a psychosocial-behavioral context, as this context shapes the nature of the direct relationship. A large cohort study conducted over 10 years showed that severe depressive symptomatology is associated with TL shortening, but shortening rates do not follow disease progression [25]. In this sense, Verhoeven and colleagues proposed two hypotheses: (i) telomere shortening is a consequence of depressive disorders, acting as a long-term cellular scar, (ii) and shorter TL is a risk factor for depression development, and longer TL attrition rates do not increase after disease establishment [25,26].

PTSD and TL
PTSD is a condition that results from traumatic events. The symptoms of PTSD include intrusive memories, hypervigilance, mood disorders and emotional withdrawal [105]. The majority of individuals who experienced traumatic events do not develop PTSD. The number of exposures and the severity of the traumatic events are associated with both the development and the severity of PTSD [105]. PTSD individuals display central noradrenergic hyperactivation and abnormalities in the hypothalamic-pituitary-adrenal (HPA) axis-thus resulting in a dysregulated stress response via unbalanced cortisol release [105,106].
The disease burden of PTSD is very high. PTSD has been linked to earlier mortality and aging-related comorbidities, such as cardiovascular disease, type 2 diabetes mellitus and dementia [107]. The accelerated aging and the mechanisms related to telomere shortening in PTSD, such as inflammation and oxidative stress, made plausible the relationship between PTSD and TL.
We analyzed 13 articles to assess the relationship between PTSD and telomere length. The findings were somewhat confusing and the studies methodologically heterogeneous, with variability in the technique of TL measurement and PTSD diagnosis. Most studies were cross-sectional and conducted on the military population, in which the overall prevalence of PTSD ranged from 6 to 13% [105,[108][109][110].
The studies presented mixed findings in regard to the relationship between TL and PTSD [22,54,57,61]. Studies with military personnel did not find any association between TL and PTSD [22,54,57,61]. However, many of these studies had small sample sizes and lacked uniformity in regard to experimental design [22,54,61,64,65]. Studies that were conducted on non-military populations reported significant associations between TL shortening and PTSD [60,62,63].
In summary, although mixed findings were found, there seems to be an association between PTSD and the shortening of TL. However, the nature of this relationship remains unclear and future research is needed to determine whether mechanisms associated with the pathophysiology of PTSD promote telomere shortening and if the shortening of TL is a risk factor for PTSD or if both PTSD and telomere shortening are consequences associated with trauma [60,63,67].

Anxiety and TL
Anxiety is defined as a feeling of inner anguish [111]. Anxiety may lead to obsessive behaviors such as nail-biting, leg bouncing (restless legs syndrome) or even stress-induced gastritis [112,113]. What differs anxiety disorders apart from benign anxiety is that anxiety disorders persist excessively and may cause a wide range of symptoms that may reduce the quality of life [111].
Anxiety disorders have been associated with dysfunctional processing of fear and dread by the amygdala, two aggregates of neurons located within the temporal lobe [111]. The limbic system, of which the amygdala is part, triggers autonomic responses through the hypothalamus, which results in cortisol release [114][115][116]. Although we observed many conflicting findings, most studies demonstrate a negative association between TL and anxiety [21,26,31,34,35]. One interesting study suggested that shorter TL could be a biomarker of later onset of anxiety [26].
Two studies did not report any association between anxiety disorders and TL [35,70]. However, one of these studies only included one anxiety disorder: panic syndrome [70]. The studies did not take into account other types of anxiety disorder including Generalized Anxiety Disorder, Agoraphobia or Social Anxiety. Likewise, the remaining study only included elderly participants [35]. Due to this reason, their findings are restricted to just one age group.
In addition, the nature of the association between TL and anxiety requires further investigation, as only one study assessed this issue [26]. Verhoeven et al. [26] suggest that it is not anxiety that causes faster telomere shortening, but instead, shorter TL could be a risk factor for the future onset of anxiety. In a longitudinal evaluation after 6 years, the authors did not find that anxiety accelerated the natural process of telomere shortening. On the other hand, individuals who already possessed shorter TLs were those most likely to suffer from an anxiety disorder.

Mechanisms Underlying TL Shortening and Psychological Distress-Related Diseases
Stress is defined as a homeostatic imbalance, which results in behavioral, chemical and psychiatric alterations. In an effort to restore physiological conditions, the human body responds with behavioral, chemical and psychiatric reactions [117]. Under stress, neuroendocrine, cardiovascular and psychological mechanisms are triggered, causing adrenergic hormones release, tachycardia, anxiety and mood swings [114,[118][119][120][121][122]. Psychiatric disorders typically affect cognition, humor and behavior, and share common distresses [122].
Biological agents and environmental conditions that cause stress are called stressors. Stressors have an impact on the release and uptake of neurotransmitters, activating or inhibiting certain brain areas. The HPA axis plays a central role in the stress response, including cortisol release. After the arrival of stressful stimuli, the axis is activated by the release of adrenocorticotropic hormone (ACTH) secretagogues, corticotropin-releasing factor (CRF) and arginine-vasopressin (AVP) by neurosecretory neurons in the parvocellular component of the paraventricular nucleus of the hypothalamus (PVN) [123]. The long-term HPA activation and chronic cortisol release cause neurostructural changes in several brain regions, including the hippocampus, the prefrontal cortex (PFC) and the amygdala-which may even contribute to stress maladaptation faced by several individuals. This process may have a role in the perpetuation of the hypercortisolism state that characterizes chronic stress. It also contributes to the maladaptation by some individuals to stress [124,125]. However, these responses also vary according to the duration of the stress exposure, the long-and short-term consequences of the phenomenon and the related neuronal circuitries [115,116,125].
Facing stress not only impacts the overall maintenance of the HPA axis but also weakens our immune system by downregulating glucocorticoid receptors in leukocytes [126][127][128][129]. It has been demonstrated that the overactivity of the sympathetic nervous system may chronically affect the immune system. The reason for this effect remains obscure, although it has been suggested that the downregulation of glucocorticoid receptors in leukocytes may be responsible [129]. Some studies have shown that chronic stress also inhibits neurogenesis and is related to a higher incidence of neuropsychological disorders, such as depression, anxiety and Alzheimer's disease [130][131][132].
Due to stress, depression and anxiety, the release of stress hormones, such as glucocorticoids, can promote hippocampal lesions by disrupting cellular metabolism and altering synaptic structures in specific areas. These mechanisms enhance levels of cell damage in chronic stress-depression-anxiety syndromes, elevating oxidative stress load and promoting hippocampal lesions [124].
The HPA axis relation to psychological distress-related diseases seems to differ from one condition to another. While depression is associated with hyperactivity of the HPA axis, anxiety disorders present subtle and less consistent alterations, displaying different patterns of HPA activity. Differences in HPA axis activation were also found in PTSD [123].

Oxidative Stress
Oxidative stress is characterized by an imbalance between reactive oxygen/nitrogen species (ROS/RNS) and antioxidants [133]. ROS are essential second messengers in many intracellular signaling cascades [134]. When endogenous systems fail to neutralize these species, key biomolecules become a target of oxidative attack-even resulting in cell death [133,134]. Oxidative imbalance is involved in the pathophysiology of over 100 diseases [133,135,136].
Glucocorticoid release, such as cortisol, affects the expression of antioxidant genes that play a role in the glutathione (GSH) redox-cycle: a mechanism to detoxify cells and prevent the formation of free-radicals, which may damage telomeres [137,138]. Psychological distress-induced glucocorticoid release increases oxidative stress, which contributes to telomere shortening. Additionally, the brain is known to be the most vulnerable organ to oxidative damage, which may partially explain the association of degenerative and psychiatric diseases to oxidative stress [138,139]. For example, pyramidal cells from the hippocampal dentate gyrus-the brain region associated with the etiology of depression and memory decline-are damaged by oxidative stress [138].
Comparative studies [136] in vertebrates suggest that glucocorticoids may influence the expression of antioxidant genes involved in the glutathione (GSH) redox cycle and in regulating the antioxidant enzyme GPX (glutathione peroxidase). The glutathione redox cycle is used to detoxify cells from the accumulation of hydrogen peroxide and organic hydroperoxides, whose cleavage induced by metal ions can generate very reactive free radicals that damage telomeres [136]. Moreover, glucocorticoids may also produce nongenomic effects on glutathione metabolism by increasing the metabolism and production of ROS [137]. Therefore, the release of glucocorticoids during psychological distress may increase oxidative stress and this could be one of the factors responsible for telomere shortening in psychiatric diseases. The brain is the most vulnerable organ to oxidative damage, due to its high oxygen consumption and lipid-rich content [138]. A total of 83 degenerative syndromes and psychiatric disorders were already associated with oxidative stress, including MDD, anxiety and PTSD [139]. The hippocampus, amygdala and prefrontal cortex are related to behavioral and cognitive deficits [138]. These brain regions are also targets to oxidative stress [138]. Pyramidal cells of dentate gyrus-cornu ammonis CA3 and granule cells of the dentate gyrus (DG) are prone to suffer from damage by oxidative stress, leading to depression, and dysfunction in learning and memory [138]. The amygdala can be hyperactivated and the PFC can suffer dendritic shrinking, which can disrupt the hippocampus-amygdala projections [138]. At last, extracellular sites of glutamatergic N-methyl-D-aspartate receptors can be oxidized by free radicals leading to the attenuation of long-term potentiation (LTP) and synaptic neurotransmission [138]. In this regard, Ahmed and Lingner [140] suggest that telomeres are especially susceptible to oxidative damage [140]. ROS-induced damage may not be accurately repaired as double-stranded telomere DNA sequences are more susceptible to cleavage by ROS [140].
Vaváková et al. reviewed many markers of oxidative stress in MDD patients related to neuroprogression, signaling, DNA damage, prooxidant enzymes, antioxidant enzymes, antioxidants, micronutrients, inflammation, immune reaction and lipid peroxidation [145]. Patients with mood disorders had increased 8-oxo-dG and decreased gene expression levels of OGG1 during depressive episodes and these changes might be reversed by the resolution of depressive symptoms [147]. Miller and colleagues reviewed cross-sectional studies that found significant differences in blood antioxidant enzyme concentrations and OXS-related gene expression between PTSD patients and controls [147].
As both telomeres and the brain are susceptible to oxidative damage, it seems reasonable to postulate that telomeres would be shortened in nervous tissues after chronic psychological distress. However, the role of oxidative stress in distress-related psychiatric disorders cannot be inferred from alterations of peripheral parameters [148]. Only two studies of our systematic search measured TL of MDD patients in brain tissues [45,56], while the majority of the studies measured TL in leukocytes. Post-mortem analysis in the brains of MDD donors revealed shorter TL in white matter oligodendrocytes [149]. Additionally, Szebeni and colleagues' [45] post-mortem analyses showed shorter relative telomere lengths in white matter oligodendrocytes, but not astrocytes, in MDD donors as compared to matched control donors. Gene expression levels of oxidative defense enzymes superoxide dismutases (SOD1 and SOD2), catalase (CAT) and glutathione peroxidase (GPX1) were significantly lower in the post-mortem brain tissue of MDD donors as compared to control donors [149]. Mamdani et al. observed a significant decrease in TL and altered mRNA levels of genes involved in neuroprotection during stress response (FKBP5, CRH) in the hippocampus of MDD subjects [56]. Distress-induced hypercortisolism possibly promotes oxidative stress, which results in telomere shortening (Figure 3). Whether telomere shortening leads to psychological diseases or vice versa still remains unclear [150,151]. This finding suggests the presence of hippocampal stress-mediated accelerated cellular aging in depression. Enhanced synaptic connectivity in the basolateral amygdala (BLA) is associated with distress-related psychiatric disorders [152]. TL is positively associated with increased activations of the amygdala and cuneus as well as increased connectivity from the posterior facial region to the ventral prefrontal cortex [153]. The mechanisms that make up the association between TL and brain functional activity remain unknown [154].
Several studies also associated oxidative stress with anxiety disorders by evaluating the expression of antioxidant genes, activity of antioxidant proteins, lipid peroxidation markers and direct and indirect intracellular ROS levels [153]. One of the pathophysiological hypotheses is that a hypercortisolism presentation in depression, anxiety and PTSD Enhanced synaptic connectivity in the basolateral amygdala (BLA) is associated with distress-related psychiatric disorders [152]. TL is positively associated with increased activations of the amygdala and cuneus as well as increased connectivity from the posterior facial region to the ventral prefrontal cortex [153]. The mechanisms that make up the association between TL and brain functional activity remain unknown [154].
Several studies also associated oxidative stress with anxiety disorders by evaluating the expression of antioxidant genes, activity of antioxidant proteins, lipid peroxidation markers and direct and indirect intracellular ROS levels [153]. One of the pathophysiological hypotheses is that a hypercortisolism presentation in depression, anxiety and PTSD would produce oxidative stress, which, in turn, might promote telomere shortening ( Figure 3). However, it is not clear if oxidative stress causes telomere erosion in CNS cells, nor is it known if ROS-induced telomere shortening in neurons and glia are a causal or contributing factor in neurological diseases [151]. Accumulating evidence suggests that by enhancing synaptic connectivity in the basolateral amygdala (BLA), symptoms of chronic anxiety and disorders such as MDD and PTSD are facilitated by enhanced availability of postsynaptic dendritic surface and synaptic inputs on principal neurons of the BLA during structural encoding of aversive experiences [152]. One study revealed that TL was positively associated with increased activation of the amygdala and cuneus, as well as increased connectivity from posterior regions of the face network to the ventral prefrontal cortex. This finding suggests that TL and genetic loading for shorter telomeres influence the function of brain regions known to be involved in emotional processing. However, precise neural mechanisms contributing to the association between TL and functional activity are still unknown [153].

Inflammation
Inflammation is a protective physiological response involving immune cells and molecular mediators triggered by external or internal factors, such as microbial infection and ischemia, respectively [154][155][156]. Peripheral inflammation results in the production of several cytokines [155]. For example, activated macrophages stimulate produce proinflammatory cytokines, including TNF-α and interleukin-6 (IL-6) [157]. The association between psychiatric disorders, inflammation and telomere length is not clearly understood, but the production of pro-inflammatory cytokines seems to be the intersection between these three components.
Pro-inflammatory cytokines are associated with telomere shortening. One possible explanation is that inflammation increases the normal human aging process and the consequent telomere shortening [165]. Cells undergoing senesce are linked to the overexpression of inflammatory cytokines, including TNF-α and IL-6 in circulating macrophages [166]. This may explain why LTL has also been associated with TNF-α, IL-6, IL-1β and CRPs [167]. Therefore, more studies are required in order to determine if depression causes higher levels of pro-inflammatory cytokines, leading to telomere shortening, or whether the shortening of the telomere provokes higher levels of these inflammatory molecules that are involved in the pathophysiology of depression.
Moreover, inflammaging is a mechanism of aging-related alteration in intercellular communication, leading to a pro-inflammatory phenotype that can accompany aging in humans [168]. This process can be provoked by proinflammatory tissue damage, affecting the immune system and host cells. These changes lead to the activation of the NLRP3 inflammasome, ROS production and cathepsin B release from unstable lysosomes and aggregated proteins. These processes are followed by an increase in IL-1β and TNF production [168], which might participate in telomere alterations due to their direct involvement in aging. This could explain why inflammaging is essentially connected to somatic cellular senescence-associated phenotype (SASP), which has been linked not only to age-related diseases in peripheral tissues but also to neurodegenerative diseases, such as Alzheimer's [169].
Although psychological stress increases plasma levels of pro-inflammatory cytokines, the mechanism by which psychological stress is detected by the innate immune system is not known [170]. It is hypothesized that the inflammasome has a pivotal role in linking psychological and physical stressors to the development of depression. It is suggested that these stressors would act as danger signals that are detected via pattern-recognition receptors (PRRs), such as NLRP3 and TLRs. The PRRs prepare defensive responses, inducing the synthesis of pro-IL-1β and the secretion of IL-1β [170]. In addition, NLRP3 is activated by several factors, such as the release of DNA damage into the cytosol [99], mitochondrial dysfunction and ROS [2]. Therefore, this could explain the interplay between inflammation and oxidative stress and how they would be involved in the relationship between telomere length and psychological distress.
Another intersection between psychiatric disorders, inflammation and telomere length is the involvement of cyclo-oxygenase 2 (COX 2). COX-2 is responsible for the formation of prostaglandins (lipophilic molecules produced for prolonging acute inflammatory response) [171,172]. COX-2 action is induced by proinflammatory cytokines, such as IL-1β and IL-6, that have been identified at increased levels in depressed patients [173]. Interestingly, COX 2 inhibitors have shown efficiency in enhancing the response of antidepressants. COX 2 is also continuously upregulated during aging and telomere shortening [174,175].
Inflammation is also able to regulate telomerase activity. The nuclear factor kappa B (NF-κB) regulates telomerase activity via the shelterin protein RAP1, which is involved in telomere homeostasis. Additionally, ROS displays a positive association with NF-κB during chronic inflammation [99]. This relationship seems to be tissue-specific, as telomerase activity is upregulated in immune cells and downregulated in non-immune cells. Inflammatory mechanisms are associated with leukocyte proliferation, which enhances DNA loss through successive mitosis and promotes TL shortening [176,177] (Figure 3). This idea could partially explain some of the heterogeneous results found in this systematic review since the majority of studies assessed TL from blood leukocytes.
In conclusion, the association between psychiatric disorders, inflammation and telomere length still remains to be elucidated. High stress levels might result in antigenic stimulation that imposes a demand for increased immune cell activity. Reciprocally, inflammatory mechanisms are associated with leukocyte proliferation, and consequently, enhancing DNA loss through successive mitosis [176]. Additionally, chronic inflammation may lead to oxidative stress and consequently promote telomere shortening [177] ( Figure 3).

Mitochondrial Dysfunction
Mitochondria are the main site of ROS synthesis in the cell. Unbalanced redox signaling may cause DNA damage and telomere shortening [178,179]. Mitochondrial treatment with a mitochondria-specific antioxidant, MitoQ, not only reduced telomere shortening but also increased cell lifespan under oxidative stress [180]. On the other hand, treatment with carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP), a compound that causes mitochondrial depolarization, increased ROS synthesis and TL shortening [179,181]. Redox signaling is important for homeostasis, but its overproduction can be pathogenic, causing damage to the cell, organelles proteins and DNA. [178] This imbalance in the production of reactive species could cause oxidative and redox stress and, consequently, telomere shortening [179].
Mitochondrial dysfunction could be one of the mechanisms related to telomere shortening in psychiatric disorders. There is evidence suggesting that brain metabolism, mitochondrial functions and redox balance are impaired in psychiatric disorders [181]. Kim et al. propose that the combination of genetic and environmental factors could lead to mitochondrial damage/dysfunction and oxidative stress, resulting in psychiatric disorders [181]. A direct link between telomere dysfunction and decrease of mitochondrial mass and energy production was also proposed. Reciprocally, there is a decrease of mitochondrial DNA (mtDNA) copies in psychiatric diseases. In fact, evidence suggests a decrease in mtDNA copy number with illness progression and aging for schizophrenia, bipolar disorder, MDD and possibly autism [181].
Besides that, varied polymorphisms in the mtDNA can cause different mitochondrial oxidative phosphorylation activities. Many variants have been described as risk factors for the development of various mental illnesses. Despite different studies indicating mtDNA mutations in psychiatric disorders over the last 25 years, it is still not clear if the mutations participate in the pathophysiology of these conditions [181].
The relationship between oxidative stress and mitochondrial dysfunction results in a vicious cycle. When telomeres have reached a critical point or when their stability is threatened, classic DNA damage responses (DDR) are activated and p53 is released. A study suggests that p53 suppresses AMPK, which is responsible for activating the peroxisome proliferator-activated receptor γ (PPARγ) and co-activator (PGC-1α), the master regulator of mitochondria biogenesis [178,182]. Sahin et al. suggest that p53 binds to PGC-1α/PGC-1β promoters and represses them [183]. In both scenarios, p53 s release causes the suppression of PGC-1α and PGC-1β and their downstream targets, resulting in disturbed oxidative phosphorylation, increased ROS and diminished ATP production and gluconeogenesis [178,183]. The increased ROS can further damage telomeres and can activate the DDR, forming the vicious cycle ( Figure 3) [178]. Finally, the suppression of PGC1α/β, which is the master regulator of mitochondrial biogenesis, decreases the renewal of mitochondria and contributes to the accumulation of damaged mitochondria-further damaging telomeres [178,184].
On the other hand, other studies reported that p53 increases mitochondrial function and oxidative defense mechanisms. The causes for these differences still need to be further elucidated, but the switch can be associated with ROS levels and tissue needs [178]. An association between mitochondrial dysfunction and psychiatric disorders is likely. Further research is needed to better establish this relationship and its implications in the pathogenesis of telomere shortening in individuals with mental illnesses.

Conclusions
Throughout life, we are under exposure to stressors. An unbalanced response to distress may lead to faster cellular aging. A growing body of evidence suggests that psychological distress, including depression, anxiety and PTSD, may be related to telomere shortening. Our study systematically reviewed 56 articles on the topic. Although studies were heterogeneous, most of them support the association between PTSD, depression and anxiety and shorter TL. These results were in agreement with previous reports. However, the lack of longitudinal analyses remains a big limitation for the establishment of causality.
Oxidative stress, inflammation and mitochondrial dysfunction in psychological distressrelated diseases and telomere shortening are likely to mediate this relationship. Finally, these processes remain hypothetical and future investigations are still necessary to understand the effect of psychological distress on TL.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/cells10061423/s1, Table S1: Quality assessment of case-control studies by the Newcastle-Ottawa Scale (NOS); Table S2: Quality assessment of cohort studies by the Newcastle-Ottawa Scale (NOS); Table S3: Quality assessment of cross-sectional studies by the adapted version of the Newcastle-Ottawa Scale (NOS)-see Appendices A-C.  Data Availability Statement: Data will be available upon request.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A. Newcastle-Ottawa Quality Assessment Scale Case-Control Studies
Note: A study can be awarded a maximum of one star for each numbered item within the Selection and Exposure categories. A maximum of two stars can be given for Comparability. Selection Data Availability Statement: Data will be available upon request.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A. Newcastle-Ottawa Quality Assessment Scale Case-Control Studies
Note: A study can be awarded a maximum of one star for each numbered item within the Selection and Exposure categories. A maximum of two stars can be given for Comparability. Selection Data Availability Statement: Data will be available upon request.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A. Newcastle-Ottawa Quality Assessment Scale Case-Contro
Note: A study can be awarded a maximum of one star for each numbered Selection and Exposure categories. A maximum of two stars can be given f ity.

Selection
(1) Is the case definition adequate?
(a) yes, with independent validation  (b) yes, e.g., record linkage or based on self-reports (c) no description (2) Representativeness of the cases (a) consecutive or obviously representative series of cases  (b) potential for selection biases or not stated (3) Selection of Controls (a) community controls  (b) hospital controls (c) no description (4) Definition of Controls (a) no history of disease (endpoint)  (b) no description of source

Comparability
(1) Comparability of cases and controls on the basis of the design or anal (a) study controls for age  (b) study controls for sex, co-morbidities and lifestyle factors 

Exposure
(1) Ascertainment of exposure (a) secure record (e.g., surgical records)  (b) structured interview where blind to case/control status  (c) interview not blinded to case/control status (d) written self-report or medical record only (e) no description (2) Same method of ascertainment for cases and controls Data Availability Statement: Data will be available upon request.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A. Newcastle-Ottawa Quality Assessment Scale Case-Control Studies
Note: A study can be awarded a maximum of one star for each numbered item within the Selection and Exposure categories. A maximum of two stars can be given for Comparability.

Selection
(1) Is the case definition adequate?
(a) yes, with independent validation  (b) yes, e.g., record linkage or based on self-reports (c) no description Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: Data will be available upon request.

Conflicts of Interest:
The authors declare no conflict of interest.