42 - 53: Depression, Diabetes and Dementia Open Access

Depression has been recognized by the World Health Organization (WHO) as one of the leading causes of disability worldwide, affecting an estimated 350 million people globally [1]. Depression has also been identified as a risk factor and poor prognostic indicator for several medical comorbidities including, but not limited to, metabolic disorders such as diabetes, metabolic syndrome and obesity [2]. According to WHO estimates, diabetes, which now affects more than 350 million people globally, will be the seventh leading cause of death by 2030 [3]. Moreover, depression and diabetes have a well-established link as numerous investigators have shown a bidirectional association between these disorders [4]. Both depression and diabetes have been independently associated as risk factors for the development of cognitive impairment and dementia [5,6]. Furthermore, depression and diabetes have been recognized as having a synergistic effect in increasing the risk for dementia above and beyond the effects that depression or diabetes would have on the risk of developing dementia independently. More specifically, comorbid type 2 diabetes mellitus and major depressive disorder have been documented to increase the incidence of Alzheimer's disease greatly later in life [7].

Alzheimer's disease is the leading cause of dementia, the sixth leading cause of mortality and the third most costly disease in the USA [8,9,10]. It is progressive in nature and is ultimately fatal [10]. Currently, available treatments are only palliative as there are no agents that have been shown to have disease-modifying properties [11]. Therefore, primary prevention represents a priority research vista for identifying and targeting modifiable risk factors of Alzheimer's disease. Modifiable risk factors that are addressed in this review are type 2 diabetes and major depressive disorder.

The overarching aim of this review is to review the potential pathophysiologic mechanisms which may account for this observed association between Alzheimer's disease, type 2 diabetes and major depressive disorder. There is also a brief discussion about preventative and therapeutic options with particular emphasis on a lifetime approach to dementia. Evidence for the association between diabetes, depression and dementia has been extensively reviewed and can be found elsewhere [7,12,13,14].

Evidence from preclinical and clinical studies has suggested several potential mechanisms connecting diabetes, depression and dementia. Essential to this discussion is the recognition that the neural networks implicated in cognitive and emotional function and dysfunction have significant overlap, as shown in figure 1[15], which is central to the observed interactions between cognition and mood. It has been amply documented that cognitive dysfunction and changes in mood are key symptoms that have been observed to appear together in individuals diagnosed with major depressive disorder [6,16]. Therefore, the observed symptoms of Alzheimer's disease and major depressive disorder may be associated with a spectrum of structural and functional changes of shared neural circuits of the frontal and subcortical regions [6,16]. Likewise, the epidemiological observation and proposal of major depressive disorder as a prodrome of Alzheimer's disease may be indicative of progressive damage to these neural circuits with repeated major depressive episodes [6,13,16].

In addition, central to the discussion of mechanisms involved in diabetes, depression and dementia is the bidirectional relationship of the body's metabolic milieu and neural circuitry [17]. Pathologic metabolic processes may damage these foregoing neural circuits. Evidence for such damage to neural circuits is shown through impaired cognition and altered brain connectivity in people with diabetes, independent of vascular pathology, as demonstrated through functional imaging techniques, such as electroencephalography, magnetoencephalography and functional magnetic resonance imaging [18,19]. The damage to frontosubcortical circuits by metabolic pathology, namely diabetes, insulin resistance, and obesity, may thus impair cognition and affect mood [18,20].

Insulin, produced by the pancreas, enters the blood stream allowing for its systemic circulation and systemic effects [21]. To enter the central nervous system (CNS), blood-borne insulin is transported across the blood-brain barrier via a saturable, receptor-mediated process [15]. Once in the CNS, insulin may exert a plethora of important effects, including, but not limited to, promoting glucose uptake in specific regions, neurogenesis, neuroplasticity, synaptic strengthening and preventing neurodegeneration [22,23]. CNS insulin and glucose have also been proposed to be an important modulator of the reward system and appetite [24].

The previous understanding that insulin had no effect on the brain has now been abolished with the discovery of insulin receptors in numerous brain regions [25]. Notably, insulin receptors have been identified in regions involved with cognition and emotion including the hypothalamus, olfactory bulb, cerebral cortex, substantia nigra, basal ganglia, hippocampus and amygdala [25]. As such, altered insulin levels and signaling, as seen in type 2 diabetes, may have important CNS consequences.

Type 2 diabetes is characterized by hyperglycemia, insulin resistance with compensatory hyperinsulinemia, and subsequent pancreatic decompensation, resulting in hypoinsulinemia at later stages, unless treated [21]. Preliminary evidence indicates that the blood-brain barrier insulin transporters may be downregulated in chronic hyperinsulinemia [15] and that, consequentially, CNS insulin levels may be decreased [26,27,28]. As previously discussed, CNS insulin is an important signaling molecule in numerous CNS processes. Therefore, in the setting of type 2 diabetes where CNS insulin has been decreased because of blood-brain barrier receptor downregulation, several deleterious effects may ensue [23,26].

The simplest deleterious effect to conceptualize is a decrease in intracellular glucose, secondary to decreased CNS insulin levels, as certain brain regions, notably the medial temporal lobe and hippocampus, have insulin-dependent glucose uptake [26]. GLUT4 receptors in the hippocampus have been shown to facilitate insulin-dependent glucose uptake [29]. In type 2 diabetes animal models of insulin resistance, there is decreased glucose uptake in the hippocampus [26]. Therefore, metabolically active CNS tissue important for cognition and emotion receive suboptimal levels of glucose and thus may have suboptimal function.

CNS insulin not only increases glucose uptake in particular regions, but also acts as a growth factor through its downstream effects. In brief, in vitro studies and animal models have shown that when insulin binds its receptor, the phosphoinositide-3 kinase pathways are activated, which promotes the production and release of several growth factors including brain-derived neurotrophic factor (BDNF) and vascular endothelial growth factor [30], thereby promoting neuronal survival, synaptogenesis and dendritic arborisation [23]. Similarly, long-term potentiation cascades, important for learning and memory, are promoted by insulin [26]. Indeed, animal models have shown a strong trophic effect of insulin on hippocampus size and benefits in cognition and memory [22,23,26,27,31]. Furthermore, in animal models, inhibiting the binding of insulin to hippocampal cells has a marked negative effect on hippocampus size and function (as manifested through impaired memory) [26,31].

Taken together, insulin acts as a growth factor and mediator of glucose uptake in the hippocampus and other areas of emotion and cognition, promoting the health of these neural circuits. Therefore, insulin resistance may lead to decreased CNS insulin levels, leading to decreased hippocampal glucose uptake and decreased growth factor signaling, which may lead to impaired neurogenesis and ultimately hippocampal atrophy. In keeping with this hypothesis, imaging and postmortem studies have found hippocampal atrophy in type 2 diabetes, major depressive disorder and Alzheimer's disease [32].

Numerous energetically ‘expensive' processes are continually occurring in the brain. These energetically intensive processes include, but are not limited to, maintenance of an electrochemical gradient in cells, neurogenesis, cellular depolarization and repolarization, synaptogenesis, and release and reuptake of neurotransmitters [33,34]. In addition, brain topology studies have shown that higher functioning neural circuits have a greater volume and length of connections, and as such, are even more energetically expensive [35].

To support these processes, the brain, while representing only 2% of total body mass, consumes 25% of the body's available glucose [34]. Furthermore, Peters et al. [24] hypothesized that the ‘selfish brain' ensures adequate energy for itself through modulating appetite, prioritizing glucose supply to the brain above other organs and decreasing energy demands. In support of this hypothesis, empirical data have shown that carbohydrate intake is increased in humans after a stressful intervention (a model shown to increase brain energy consumption by 10-15%) [36,37]. In addition to postintervention changes in food consumption, participants had increased blood glucose with a blunted insulin response. Since brain glucose uptake is largely insulin independent, this pattern is indicative of a preferential bias of brain glucose uptake relative to other organs when brain energy demands are increased [37].

Low CNS insulin and intracellular glucose levels, as seen in type 2 diabetes with downregulated blood-brain barrier insulin transport and insulin resistance, may lead to the perception of low energy availability and thus may trigger downstream mechanisms to conserve energy [24,38]. To conserve energy, neurogenesis is prevented as it requires great consumption of energy [24,29,39]. Animal models of type 2 diabetes have reported that N-methyl-D-aspartate (NMDA) signaling for the purpose of long-term potentiation is inhibited as it too is energetically expensive [40].

Ultimately, energy deficits perceived by the brain (e.g. lowered CNS insulin and glucose in the case of type 2 diabetes) may negatively impact neuroplasticity, optimal neural circuitry and thus optimal function [24,29,39]. In keeping with this view, when the frontosubcortical regions are forced to reduce energy expenditure by allocating resources to less demanding neural circuits, the domains of mood and cognition may be impaired to conserve energy in times of perceived energy depletion [17,24,38]. Furthermore, chronic perceived bioenergetic depletion may lead to chronic and progressive impairment of these circuits [17]. Taken together, comorbid mood and cognitive symptoms, commonly observed and reported in epidemiological studies as the co-occurence of major depressive disorder and Alzheimer's disease, may represent late-stage bioenergetic bias. More specifically, older individuals with type 2 diabetes experience years of brain insults resulting from glycemic dysregulation and aberrant insulin levels, affecting neuroplasticity and neural function, leading to bioenergetic bias towards energy conservation rather than higher functioning [17].

Another mechanism whereby hyperinsulinemia with insulin resistance and low CNS insulin may impact function is through the insulin-amyloid pathway. Amyloid-β protein oligomers have been strongly implicated in the pathoetiology of Alzheimer's disease [41]. What induces amyloid-β oligomer formation is not yet fully understood; however, a strong interaction between amyloid-β and insulin has been identified [41,42]. Notably, insulin and amyloid-β compete for the same degrading enzyme, namely, insulin degrading enzyme, thereby indirectly affecting each other's systemic and central concentration [43,44]. Furthermore, insulin degrading enzyme levels are reduced in type 2 diabetes animal models [14,18]. Moreover, insulin has been shown to modulate amyloid-β removal, reduce amyloid-β load and prevent pathogenic binding of amyloid-β [28,43].

The interaction between insulin and amyloid-β is bidirectional. More specifically, accumulating amyloid-β can bind to insulin receptors and downregulate insulin signaling, thus impairing neurogenesis and function through the previously discussed mechanisms [45]. The foregoing bidirectional interaction may perpetuate a deleterious positive feedback loop whereby insulin resistance promotes increased amyloid-β and amyloid-β induces further insulin resistance [26]. This loop may provide another reason for the high prevalence of Alzheimer's disease in people with type 2 diabetes [26].

Neuroplasticity is the brain's ability to change in response to environmental stimuli [46]. The inability of the brain to change at a molecular, cellular, structural and ultimately functional level is an important cause of impaired cognition and mood. Diabetes and depression have been shown to impair neuroplasticity through the inflammatory pathway, oxidative stress, depletion of neurotrophic factors and hypothalamic-pituitary-adrenal (HPA) axis derangement, leading to impaired neurogenesis, impaired long-term potentiation and neurotoxicity [47,48].

A proinflammatory state has been shown to be associated with stress (physical, emotional, psychological), inflammatory medical comorbidities (infection, obesity, diabetes, metabolic syndrome, autoimmune diseases, cardiovascular disease) and psychiatric comorbidities (mood disorders, anxiety disorders, psychotic disorders) [49]. The cytokines produced, most notably TNF-α, IL-6 and IL-1β, in this proinflammatory state have been associated with ‘sickness behavior' including depressive symptoms of lethargy, anhedonia and cognitive decline [49]. The downstream effects of the proinflammatory state include derangement of monoamine levels, pathologic microglial cell dysfunction, increased oxidative stress, HPA axis activation and structural changes of the subcortical regions, all of which may lead to decreased neuroplasticity and dysfunctional emotional and cognitive processing [50].

More specifically, proinflammatory cytokines have been shown to cause tryptophan depletion and increased serotonin turnover, both well-known causes of altered mood and cognition [51]. Increased oxidative stress also accompanies the elevated metabolic rate induced by a proinflammatory state, thus increasing cellular damage at the molecular level ultimately leading to neurotoxicity [51].

Also implicated in the inflammatory pathway are microglial cells, the macrophages of the CNS. Microglial cells are important for normal brain function and synaptic pruning; however, in a chronic inflammatory state, microglial cells may induce high levels of neural death through pathologic synaptic pruning leading to impairment of the cognitive and mood neural circuits [52,53].

The proinflammatory state is also associated with activation of the HPA axis leading to hypercortisolemia as well was impaired HPA negative feedback through decreased glucocorticoid receptor expression, translocation and signaling [54]. Of note, diabetes and depression as well as stress all have the ability to activate the HPA axis, inducing hypercortisolemia [54,55] which leads to further hyperglycemia and thus to poorer control of diabetes and increased insulin requirements [21]. Chronic exposure to high levels of cortisol can also cause neurotoxicity in the hippocampus, leading to impaired cognition and mood symptoms [56]. This neurotoxic effect is amplified in the presence of insulin resistance [26]. Notably, cortisol has also been implicated in altering metabolism of amyloids and may thus exhibit pro-Alzheimer's disease properties through the amyloid mechanism as well [57].

Low levels of BDNF are commonly observed in people with type 2 diabetes, impaired glucose tolerance, major depressive disorder and Alzheimer's disease [58]. This observation may suggest another common mechanism of pathogenesis of these diseases. For example, during a major depressive episode, hippocampal 5-HT2A receptors are upregulated, which may lead to decreased BDNF and a subsequent decrease in hippocampal volume [59]. Moreover, BDNF is one of the downstream targets of insulin receptor activation [30]. Therefore, in the setting of insulin resistance where signaling is diminished, BDNF levels are also lower [26]. The low levels of BDNF, as seen in both type 2 diabetes and major depressive disorder, may thus increase vulnerability to hippocampal dysfunction and atrophy secondary to inadequate growth factor stimulation [26,31,60].

The monoamine pathway has been the primary therapeutic target of major depressive disorder for decades. Causes of monoamine derangement are numerous, including the previously discussed inflammatory pathway [49] and CNS insulin [26]. Monoamines, notably serotonin, norepinephrine and dopamine, are now also being recognized as having significant roles in cognition [61]. For example, serotonin receptors are found abundantly in cognitive regions such as the hippocampus, prefrontal cortex and septum, where they play an important role in cognition, most notably creation of episodic memories as evidenced by tryptophan depletion studies [62]. Norepinephrine plays a significant role in alertness and arousal, thus affecting cognitive function in the domains of working memory, memory consolidation, attention and executive function [61]. Dopamine levels are also salient to cognition, motivation and the reward system [63]. In sum, the monoamine system can be altered by depression and type 2 diabetes and has large effects on both cognition and emotion. Therefore, it may provide another pathophysiologic nexus to account for the interplay between depression, diabetes and cognitive impairment [61].

Many decades ago, the vast and dynamic ecosystem of the gut microbiota was hypothesized to affect mental health; however, only recently has this idea been revisited [64]. Currently, there is only limited evidence for the bidirectional interaction between the gut and the CNS, but interest across the fields of psychiatry, neurology, gastroenterology and endocrinology is increasing [65]. Preclinical evidence suggests that this interaction may provide mechanisms whereby the gut microbiota may induce metabolic, behavioral, mood and cognitive changes [65]. This field is still in its infancy but is being implicated as an etiologic factor in dictating eating habits, BDNF levels, HPA axis activity, inflammation and oxidative stress, all of which are important factors in the mechanisms of affective and cognitive dysfunction, as previously discussed [64,65]. In brief, the gut may have bidirectional communication with the brain to induce behaviors that increase the risk of diabetes as well as may be the nidus of a chronic inflammatory state which may induce sickness behavior, hypercholesterolemia, hyperglycemia, hypercortisolemia and ultimately neural damage, affecting cognitive and emotional function [64,65].

Microvascular (nephropathy, neuropathy, retinopathy) and macrovascular (stroke, myocardial infarction) complications of diabetes are well established in the literature [21]. Moreover, it has now been well-established that the brain has increased susceptibility to microvascular disease, independent of the presence of large vessel disease, in people with diabetes [66]. Postmortem and imaging analyses has revealed these microangiopathic changes to be associated with the development of depression, cognitive impairment and dementia, particularly vascular dementia and Alzheimer's disease [67]. Microangiopathic changes may be linked with amyloid-β oligomer formation [68]. The proposed mechanism and evidence in support of the microangiopathy-cognitive-affective connection have been reviewed elsewhere [66,67,69].

A separate but not mutually exclusive explanation of the connection between diabetes, depression and dementia may be due to the psychosocial risk factors that are shared by these disease processes [17]. Epidemiological studies provide evidence for low socioeconomic status, childhood adversity, medical comorbidities, addiction and psychiatric comorbidities as being common risk factors for type 2 diabetes and major depressive disorder [70,71,72]. Therefore, one potential explanation for the observed association may be common etiologic factors [17]. As shown in figure 2, both common biological and psychosocial factors may be creating a pathological nexus between peripheral metabolism and brain function. Taken together, certain life circumstances may predispose individuals to the development of both type 2 diabetes and major depressive disorder, which could synergistically damage the neural circuits of emotion and cognition, thereby increasing the risk of cognitive impairment and dementia [17].

As shown, numerous mechanisms may be involved in the interaction of diabetes, depression and dementia, and may present new therapeutic targets. Therefore, in this section, the focus will be placed on selected new and novel therapeutic and preventative options that target components of the mechanisms previously discussed. Emphasis will be placed on (1) the benefit of early detection and treatment of major depressive disorder and diabetes, (2) the importance of lifestyle modifications, namely exercise and dietary changes, and (3) novel metabolic therapeutic options which hold promise for improving mood and cognition in the context of major depressive disorder, type 2 diabetes and Alzheimer's disease.

As previously discussed, no disease-modifying agents are currently identified for Alzheimer's disease [10]. Therefore, prevention through risk factor modification has been a focus of Alzheimer's disease research and public health efforts [10,73]. Understanding the pathophysiology of Alzheimer's disease as occurring over a lifetime, rather than only in old age when symptoms begin to manifest, is essential to this preventative approach. Epidemiologic and mechanistic data strongly suggest both depression and diabetes to be modifiable risk factors of Alzheimer's disease. Therefore, screening programs to allow for early detection and treatment of both major depressive disorder and type 2 diabetes may be extremely helpful in the prevention of Alzheimer's disease. Indeed, evidence suggests that an estimated 10-15% of Alzheimer's disease cases are attributable to depression and a 25% reduction in depression prevalence would result in approximately 830,000 fewer Alzheimer's disease cases worldwide [74]. Similarly for type 2 diabetes, appropriate therapy achieving adequate glycemic, lipid and blood pressure control has been shown to decrease the risk of cognitive impairment, Alzheimer's disease and vascular dementia [75,76]. Taken together, screening for major depressive disorder in people with type 2 diabetes and vice versa would thus have clear benefits for early detection and treatment for both conditions while reducing the risk of developing Alzheimer's disease [77].

From a public health perspective, exercise and a healthy diet have been promoted because of their positive impact on cardiovascular health [78]. Evidence consistently documents the importance of exercise and a healthy diet and their impact on diabetes, depression and dementia [14,79,80,81,82,83]. For example, exercise has been shown repeatedly to have a positive effect on mood and cognition in individuals with depression [32,83]. Likewise, among individuals with Alzheimer's disease, exercise has been shown to improve mood and cognition both in the short and long term [80,81]. In type 2 diabetes, lifestyle management has long been a part of first-line therapy with clear effects on cognition, mood and Alzheimer's disease prevention [10,84,85]. Therefore, lifestyle management should also be emphasized from a mental health perspective, rather than only for its cardiovascular benefits.

From a pharmacologic perspective, increasing interest has been developing for the use of diabetic medications for improvement of mood and cognition. For example, recent studies have been conducted investigating the role of insulin in mood and cognition [22]. As previously discussed, CNS insulin plays a key role in brain function, and CNS insulin levels are reduced in people with Alzheimer's disease [27]. Intravenous administration of insulin has been investigated for its effects on cognition and emotion. Several investigators have shown that in both human and animal models, intravenous insulin administration can increase hippocampal neural activity and improve mood and cognition (most reproducibly declarative memory) in the short and long term [86,87]. However, concerns of the systemic effects and risks, namely hypoglycemic events, were a limiting factor to their wider promotion [88,89]. Therefore, interest has shifted to intranasal insulin administration, whereby insulin directly enters the CNS, bypassing the blood-brain barrier, presumably via the olfactory and trigeminal nerves [88,89]. Intranasal administration has been shown to increase the level of insulin in the CNS while not affecting the systemic concentration of insulin or glucose [88,89,90]. Furthermore, in human and animal studies, intranasal insulin has been shown to improve mood and cognition reproducibly in healthy, obese and Alzheimer's disease subjects [91]. Moreover, intranasal insulin has benefited disparate measures of neurocognition in adults with bipolar disorder, a group highly susceptible to cognitive dysfunction [92].

Oral diabetic medication, particularly thiazolidinediones and incretins have also produced interesting results. Pioglitazone, a thiazolidinedione, is an insulin sensitizer that has been shown to also have effects on cognition [87,93]. Indeed, cognitive improvements as well as a reduction in the development of Alzheimer's disease have been observed in people with type 2 diabetes treated with pioglitazone [87,93]. Incretins, gastrointestinal hormones which delay gastric emptying and increase insulin release from the pancreas, have also shown great promise to treat metabolic, cognitive and affective symptoms simultaneously [22,94]. In a recent review by McIntyre et al. [94], the positive effects of incretins, specifically glucagon-like peptide-1, were discussed. In brief, glucagon-like peptide-1 may improve glycemic control, improve cognition and mood, and decrease the risk of Alzheimer's disease through its neuroprotective effects, ability to reduce amyloid-β load and through promotion of long-term potentiation and neuroplasticity [22,94].

Bariatric surgery produces vastly improved glycemic control and significant sustained reduction in weight far beyond the effects of medical therapy alone [95,96,97]. This significant improvement in diabetic control, in theory, may decrease the risk of depression and cognitive impairment later in life by preventing neural degeneration from the previously described mechanisms. However, to the authors' knowledge, there have been no studies to assess the effect of bariatric surgery on mood, cognition and risk of Alzheimer's disease development. Therefore, this effect may be an interesting topic for future research.

Several mechanisms have been proposed to account for the well-documented association between diabetes, depression and dementia. The metabolic-brain axis appears to be a key mediator connecting these conditions. Brain regions important for cognition and emotional regulation may be damaged by the effects of hyperglycemia and insulin resistance. Indeed, a decreased insulin effect and thus decreased intracellular glucose levels in frontal and subcortical regions results in neurotoxicity, decreased neuroplasticity, decreased signaling and decreased synaptic connectivity resulting in an overall effect of dysfunctional neural circuits. Major depressive disorder may further facilitate neural circuit damage through the inflammatory pathway, HPA axis deregulation, monoamine changes and lowering of central BDNF levels. Stress and psychosocial determinants of health may also be key mediators and etiologic factors in these interactions. Taken together, several mechanistic pathways may be involved, presenting new potential drug targets for the treatment and prevention of dementia using a lifetime approach. Systemic and intranasal insulin, insulin sensitizers, incretins, exercise, dietary changes, bariatric surgery, improved screening and early treatment practices of type 2 diabetes and major depressive disorder have all been implicated for potential use in the treatment and prevention of dementia. Further research is most definitely indicated as the literature reviewed here shows great promise for further studies of the metabolic-brain axis which could revolutionize the understanding, treatment and prevention of cognitive and affective disorders.