|Home | About | Journals | Submit | Contact Us | Français|
Age-related dementia is increasingly recognized as having a mixed pathology, with contributions from both cerebrovascular factors and pathogenic factors associated with Alzheimer’s disease (AD). Furthermore, there is accumulating evidence that vascular risk factors in midlife, e.g., obesity, diabetes, and hypertension, increase the risk of developing late-life dementia. Since obesity and changes in body weight/adiposity often drive diabetes and hypertension, understanding the relationship between adiposity and age-related dementia may reveal common underlying mechanisms. Here we offer a brief appraisal of how changes in body weight and adiposity are related to both AD and dementia on vascular basis, and examine the involvement of two key adipocyte-derived hormones: leptin and adiponectin. The evidence suggests that in midlife increased body weight/adiposity and subsequent changes in adipocyte-derived hormones may increase the long-term susceptibility to dementia. On the other hand, later in life, decreases in body weight/adiposity and related hormonal changes are early manifestations of disease that precede the onset of dementia and may promote AD and vascular pathology. Understanding the contribution of adiposity to age-related dementia may help identify the underlying pathological mechanisms common to both vascular dementia and AD, and provide new putative targets for early diagnosis and therapy.
Age-related dementia is an incurable condition characterized by progressive cognitive decline in the elderly. This devastating disease affects an estimated 35.6 million people worldwide and is expected to become a major health epidemic if no effective therapy is developed . The underlying etiology is complex and multi-factorial, but Alzheimer’s disease (AD) and cognitive impairment on vascular bases (vascular cognitive impairment and dementia) constitute the vast majority of the age-related dementias. Alzheimer’s disease is a progressive neurodegenerative disorder that is defined by the neuropathological hallmarks of amyloid plaques and neurofibrillary tangles comprised of abnormal accumulation of amyloid-beta (Aβ) peptides and hyperphosphorylated tau, respectively . On the other hand, vascular dementia is a heterogeneous group of brain disorders where cognitive decline is attributable to a wide range of cerebrovascular pathologies .
In the past, AD and vascular dementia were considered distinct disorders; however, accumulating evidence suggests that there is significant overlap between these two conditions . Dementia due to vascular causes is likely to represent one end of a spectrum of age-related dementia, while dementia due to AD pathology may represent another end of the spectrum, with many cases having contributions from both [4–6]. Epidemiological studies have supported this notion by demonstrating that traditional vascular risk factors such as obesity, hypertension, and diabetes can all increase the risk of age-related cognitive decline and dementia [7–9]. Since disorders in body weight, obesity in particular, can increase the risk for cardiovascular disease  and often drive the hypertension and insulin resistance [11,12], factors that modulate body weight and adiposity may be significant contributors to age-related dementia caused by either vascular or AD pathology. In addition, dysfunction of the hypothalamus, a region of the brain that is critical for the central regulation of body weight and systemic metabolism, are commonly seen in AD and related dementias . Therefore, understanding the mechanisms through which changes in body weight and adiposity contribute to impaired cognition may enable us to identify common pathways for age-related dementia.
Here, we review the evidence supporting the notion that alterations in body weight, adiposity in particular, can influence the risk of developing age-related dementia through both vascular factors and pathological mechanisms associated with AD. We first examine the epidemiological evidence linking changes in body weight and adiposity to age-related dementia and brain pathologies. We then focus on two adipocyte-derived hormones, leptin and adiponectin, as examples for how factors that regulate and are regulated by adiposity can interact with the brain resulting in neurovascular changes, worsening AD pathology, and eventually cognitive decline and dementia.
During the early 1900s Alois Alzheimer first identified dementia associated with plaques and tangles as a distinct pathological condition that was different from common forms of cognitive decline, at the time attributed to cerebrovascular pathology (“hardening of the arteries”) . However, in the 1990s, the molecular identification of Aβ peptides as the major component of the plaques and the subsequent identification of mutations in the amyloid precursor protein (APP) in familial forms of AD rapidly shifted the focus from cerebrovascular causes of dementia to those caused by AD pathology. A dichotomous framework was thus developed over time where age-related dementia was defined as vascular dementia when the cognitive decline was attributable to cerebrovascular causes and AD when the cognitive decline was attributable to amyloid and tau pathology. While AD has received the bulk of attention in regards to drug discovery efforts for age-related dementia , it has become increasingly recognized that many patients have a mixed pathological picture with contributions from both vascular and AD pathologies [4–6]. This is not surprising since both cerebrovascular disease and AD pathology significantly increase in prevalence with aging. Furthermore, epidemiological studies have increasingly demonstrated that traditional vascular risk factors, such as obesity, diabetes and hypertension, have important etiological roles not only in vascular dementia but also in AD [7–9]. Due to the increased recognition of a significant overlap between these conditions, there is a need to examine vascular dementia and AD beyond a simple dichotomy, but as part of a continuous spectrum of age-related dementia where factors contributing to either vascular dementia or AD are considered together .
Since obesity, hypertension, and diabetes are now believed to be involved in the pathogenesis of age-related dementia, there is an urgent need to explore how these risk factors relate to the dementia on a mechanistic level. Here, we will focus on obesity and body weight disorders because changes in adiposity can often drive the pathological mechanisms underlying both hypertension and diabetes [11,12]. While there may be overlap among adiposity, hypertension, diabetes and their contributions to age-related dementia, adiposity has been found to be an independent risk factor for age-related dementia , suggesting that there are also potentially independent contributing factors related to adiposity. Therefore, understanding how adiposity affects brain structure and function could help identify common molecular and cellular mechanisms underlying both vascular and AD pathologies in age-related dementia.
In normal healthy adults, the body composition changes with aging. Thus, adiposity generally increases throughout midlife until late-life around the seventh or eighth decade of life [16,17]. The increase in adiposity is most evident in visceral adipose tissue, while there may be decreases in subcutaneous adipose tissue. Once late-life is reached, body weight tends to decrease as part of normal aging, which results from a combination of loss of lean tissue mass or sacropenia and loss of adipose mass. This age-associated change in body composition, where adiposity increases throughout midlife but decreases once late-life is reached, has been consistently found in several different ethnic groups and may be influenced by dietary intake [18–20]. Furthermore, an epidemiological study investigating the influence of the recent obesity epidemic on the age-related changes in body composition found that later birth cohorts had higher percent body fat compared to earlier birth cohorts in both men and women, but the overall pattern of initially increasing percent body fat with age then a leveling off later in life was the same amongst the different birth cohorts . These studies strongly suggest that there are fundamental mechanisms that regulate changes in body composition with aging, which, however, remain unclear. One possibility is that the age-related changes in body composition may be due to changes in the homeostatic response to hormonal and signaling factors that regulate body weight and adiposity . Additionally, other factors such as changes in circadian rhythm, which are commonly seen with aging and aging-related dementia, may impact systemic metabolism that lead to age-related alterations in body weight and body composition [23,24].
Several epidemiological studies have consistently demonstrated that midlife obesity is a risk factor for developing cognitive decline and dementia that is independent of other known vascular risk factors, such as hypertension and diabetes [7,25–27]. Similarly, diets high in saturated fatty acids, a major cause of obesity, have been associated with increased risk of developing AD and exacerbation of AD pathology in several animal and human studies [28–30]. This positive association between body weight and dementia has been consistently reported in several different countries and ethnic groups [7,25–27,31–33]. However, a recent epidemiological study from the UK found that midlife obesity reduced the risk of developing dementia . Although the reasons for this discrepancy are not clear, one contributing factor could be that using body weight or body mass index (BMI) as a measure of body composition and adiposity does not distinguish between changes that are due to alterations in lean mass from adipose mass. Future studies using more accurate measurements of body composition (e.g., imaging adipose tissue using computed tomography or similar imaging modalities) may provide a more coherent view of the relationship between adiposity and dementia . These potential methodological confounders aside, the majority of studies thus far support a link between midlife obesity and dementia (see Table for summary of studies).
Late-life body weight changes are also commonly seen once the dementia manifests. Weight loss was originally listed in the 1984 report by the National Institute of Neurological and Communicative Disorders and Stroke (NINCDS) and the Alzheimer’s Disease and Related Disorders Association (ADRDA) work group as one of the clinical features consistent with a diagnosis of probable AD . Importantly, the weight loss in AD and related dementias correlates with disease morbidity and mortality since individuals who lose body weight have more rapid disease progression and increased risk of death [37,38]. In addition, several population studies have consistently demonstrated that low body weight in late-life is associated with an increased risk of developing cognitive decline and dementia [39–45]. The appreciation that AD pathology develops decades prior to the initial cognitive symptoms suggest that the late-life changes in body weight may reflect the earliest stage of the disease process [40,46]. Supporting this notion, low BMI is associated with worsening AD pathology in postmortem brains [47,48] as well as worsening CSF biomarkers (tau and Aβ1–42) . Collectively, these studies suggest that increased adiposity during midlife contribute or influence the risk of developing age-related dementia, while reduced adiposity in late-life is an early consequence of the disease and may contribute to its progression. The underlying mechanisms for the weight changes in AD have not been elucidated. Interestingly, alterations in feeding behavior secondary to reduced mental capacity may not play a role since the weight loss often precedes the cognitive decline [40,44]. Another hypothesis is that AD pathology involving hypothalamic centers regulating body weight homeostasis leads to uncoupling between food intake and energy expenditure, as seen in mouse models of AD . However, while there is accumulating evidence of hypothalamic dysfunction in AD  further studies are needed to provide insight into the impact of AD pathology on central mechanisms regulating body weight homeostasis.
The white adipose tissue (WAT) was classically believed to be simply a fat storage depot tissue. However, the identification of the adipocyte hormone leptin has revolutionized the field to now view the adipose tissue as a major endocrine organ releasing hundreds of peptides, proteins, metabolites and inflammatory signals to modulate a wide variety of physiological functions [52,53]. WAT is dynamic and changes in adiposity will change the amount and nature of the factors that are secreted from it . While an exhaustive appraisal of all the signaling molecules produced and regulated by WAT is beyond the scope of this review, we will focus on leptin and adiponectin as examples of adipocyte derived factors that affect vascular and AD pathology in age-related dementia.
Leptin is a 16kDa adipocyte-derived hormone that is secreted in proportion to the adipose stores with high circulating plasma levels in obesity and low circulating plasma levels in states of starvation [54–56]. This positive correlation between body adiposity and circulating leptin levels persists throughout aging and in older adults [20,57,58]. Leptin has both central and peripheral actions by binding to its receptors, which are part of the class I cytokine family, and activating several key downstream intracellular signaling pathways . The hypothalamus is a major site of action of leptin in the brain, where leptin acts as a negative feedback signal in the afferent loop that maintains body weight homeostasis by regulating feeding behavior, energy expenditure, thermogenesis, autonomic function, and systemic metabolism including insulin sensitivity and facilitating lipolysis . Therefore, high plasma leptin due to increased adiposity would reflect a state of energy excess and activate the hypothalamic circuit to lower adipose mass by coordinating a complex physiological response including decreasing food intake, increasing energy expenditure, and increasing sympathetic tone . Leptin can also act by central and peripheral mechanisms to improve glucose metabolism by increasing insulin sensitivity and by insulin-independent mechanisms [60,61]. Recently, leptin has been demonstrated to mediate the increase in blood pressure associated with obesity through hypothalamic circuits ; however, whether leptin can directly modulate blood pressure in humans remains controversial . In addition to its metabolic effects, leptin has also been demonstrated to have pleotropic effects including immune regulation, reproduction, and bone metabolism .
Rare autosomal recessive mutations in the leptin gene result in morbid obesity associated with hyperphagia and severe neuroendocrine abnormalities that are corrected with exogenous leptin replacement therapy . However, most forms of obesity are not from a lack of leptin but are associated with high circulating leptin due to the increased adiposity, resulting ultimately in leptin resistance . Therefore, it is not surprising that exogenous leptin administration has little effect in obesity associated with high plasma leptin [67,68]. Despite the initial disappointment in the lack of efficacy of leptin in common obesity, leptin treatment using a recombinant modified form of leptin (metreleptin) was recently approved by the US Federal Drug Agency for the treatment of congenital or acquired generalized lipodystrophy and appears to be a promising therapy for other disorders associated with leptin deficiency or low circulating leptin .
Leptin has a strong neuroprotective effect under a variety of pathologies in experimental models. In particular, leptin treatment ameliorates ischemic damage in models of oxygen-glucose deprivation in vitro and in animal models of focal cerebral ischemia produced by occlusion of the middle cerebral artery [69,70]. Furthermore, exogenous leptin can augment the cerebral hemodynamic reserve in a three-vessel occlusion cerebral ischemic model . Conversely, larger infarct volume after focal cerebral ischemia is seen in leptin deficient ob/ob mice when compared to wild-type mice . The beneficial effects of leptin is mediated in part through JAK-STAT3 signaling pathways that lead to the induction of the anti-apoptotic protein Bcl-xL, stabilization of mitochondrial membrane potential, and reduction in mitochondrial oxidative stress . Despite the strong evidence for leptin having a protective effect against cerebral ischemia in preclinical studies, several epidemiological studies have found that high circulating leptin levels are associated with ischemic cerebrovascular disease [74–76]. While a pathogenic role of high circulating leptin cannot be entirely excluded, the association with ischemic stroke seen in these studies is likely to reflect leptin resistance leading to a decrease in the neuroprotective effects of leptin . Overall, these studies support a protective role of leptin signaling against cerebrovascular pathologies, but the mechanisms remain unclear.
Leptin has also been demonstrated to have a neuroprotective role against AD pathology and dementia. Preclinical studies have found that leptin can decrease Aβ levels by reducing key enzymes required for Aβ production, increasing Aβ clearance and degradation, and inhibiting the aggregation of Aβ peptides [78–82]. Furthermore, exogenous leptin reduces whereas leptin deficiency exacerbates AD pathology and cognitive function in transgenic mouse models of AD [83–85]. With notable exceptions [86,87], most epidemiological studies also support the notion that leptin may be protective in dementia, since low plasma leptin levels in late-life are associated with worsening cognitive decline and increased dementia risk in longitudinal studies [88,89]. Furthermore, in cross-sectional studies plasma leptin levels were found to be lower in AD subjects compared to normal controls independent of body weight/adiposity [90–92]. Experimental studies also implicated leptin in AD pathology. While there are significant body weight differences among mouse strains , several mouse models of AD exhibit low body weight compared to wild-type controls [50,94–97]. In one transgenic mouse model overexpressing a mutated form of APP (Tg2576), there was a decrease in body weight and low plasma leptin prior to the development of amyloid pathology and cognitive deficits . These changes in the Tg2576 mice were attributed to Aβ-mediated inhibition of leptin-responsive hypothalamic neurons. As the hypothalamus is affected in AD, these observations suggest that Aβ could directly cause weight loss and low plasma leptin levels by inhibiting neural circuits involved in the regulation of body weight [13,50]. Collectively, these findings open up the possibility that Aβ can be a driver of the weight loss to further decrease plasma leptin and its neuroprotective effects, which would lead to a downward spiral of worsening AD pathology and further weight loss.
Adiponectin is a 30kDa adipocyte derived polypeptide that is structurally similar to the complement factor C1q and is comprised of a globular C-terminus domain and a collagenous N-terminus domain with the collagenous domain leading to formation of oligomeric and higher order structures [98,99]. Adiponectin exerts its effects by binding to two related seven transmembrane receptors adipoR1 and adipoR2, which are functionally and topologically distinct from G-protein-coupled receptor [99,100]. Adiponectin circulates in the blood and is relatively abundant representing up to 0.05% of total plasma proteins ; however, unlike many adipocyte derived factors, circulating adiponectin levels decrease with increasing adiposity . Low circulating adiponectin levels have been associated with heart failure, coronary artery disease, hepatic steatosis, dyslipidemia and type 2 diabetes, suggesting that adiponectin may have a protective role against obesity-associated disorders [53,99].
Exogenously administering adiponectin has been demonstrated to have multiple effects on body weight and systemic metabolism [53,99]. Adiponectin is an effective insulin sensitizer and can also induce weight loss by acting in the brain to increase energy expenditure [53,99,102]. More recently, a synthetic small molecule adiponectin agonist was reported to significantly improve insulin sensitivity and glucose tolerance leading to increased longevity in a mouse model of obesity, further strengthening the therapeutic potential for adiponectin in treating obesity-associated disorders . Adiponectin-deficient mice exhibit severe diet-induced insulin resistance associated with high levels of tumor necrosis factor-α in adipose tissue and plasma . Furthermore, adiponectin can promote the polarization of tissue macrophages towards an M2 or anti-inflammatory phenotype, suggesting that the insulin sensitizing effects may be mediated by its anti-inflammatory properties .
Similar to the systemic obesity-associated disorders, low circulating adiponectin levels have been consistently associated with ischemic cerebrovascular disease . In a case-control study, subjects with ischemic cerebrovascular disease had significantly lower adiponectin levels compared to controls . Another study compared serum adiponectin levels in patients with cerebral infarction with other atherosclerotic disorders and found that basal adiponectin levels were lower in subjects with cerebral infarction compared to healthy controls, but there was no difference between subjects with cerebral infarction and those with other atherosclerotic disorders, such as ischemic heart disease . However, during the acute phase of ischemic stroke, circulating adiponectin levels were temporarily reduced followed by recovery to the low basal levels, suggesting that adiponectin levels can dynamically and acutely change during cerebrovascular insults . Low circulating adiponectin levels have been associated with increased mortality after ischemic stroke and may even help to predict neurological severity and functional outcome [108,109]. In a focal cerebral ischemia model in mice, the overexpression of adiponectin ameliorated the ischemia-induced brain atrophy and improved overall neurological function, effects associated with enhanced angiogenesis . There is also evidence that adiponectin can have protective effects against ischemic insults through endothelial nitric oxide synthase-dependent hemodynamic mechanisms and/or anti-inflammatory effects through the NF-κB pathway [111,112]. Collectively, these studies are consistent with adiponectin having neuroprotective effects against ischemic stroke. Thus, the positive association between obesity and cerebrovascular disease could be explained at least in part by obesity leading to a state of low circulating adiponectin.
The role for adiponectin in age-related dementia is less well established. Exogenous adiponectin was reported to be protective against Aβ-induced neurotoxicity under oxidative stress conditions in a cell model . In epidemiological studies, two independent crosssectional studies found increased plasma and/or CSF adiponectin levels in AD subjects compared to normal controls [91,114], which correlated with worsening cognitive impairment in one of the studies . In addition, a prospective study as part of the Dementia-free Framingham Study found that increased plasma adiponectin levels were an independent risk factor for developing all-cause dementia and AD in elderly women . The association of increased plasma adiponectin levels with dementia found in these studies may be initially surprising since low adiponectin levels are often found in pathological states most commonly seen with obesity-associated disorders. However, these findings are in congruence if one considers that these studies were conducted in late-life with AD, where weight loss may be present (see above), thus leading to higher circulating adiponectin levels. An additional question that would need to be addressed is: if adiponectin generally has protective properties why would high circulating adiponectin levels not improve or decrease the dementia risk? One possibility is that the high circulating adiponectin levels leads to subsequent resistance to adiponectin in a similar fashion to insulin resistance and leptin resistance [116,117]. While showing an association between high plasma adiponectin and dementia, these findings are still controversial. For example, two additional cross-sectional studies found no significant differences in plasma adiponectin levels [90,118], while other studies have found significantly lower circulating adiponectin levels in mild cognitive impairment or AD subjects compared to controls [35,119]. Therefore, the relationship between adiponectin and dementia has not been completely elucidated and additional studies are required to clarify these discrepancies.
In addition to leptin and adiponectin, there are several other adipocyte-derived factors that could have an impact on age-related dementia such as plasminogen activator inhibitor 1 (PAI-1) and retinol binding protein 4 (RBP4) . Since obesity can be considered a state of inflammation both in brain and in adipose tissue, neuroinflammatory factors associated with obesity could also play a significant role in the development of dementia . In addition, more recently, the gut microbiota transferred from mice on a high-fat diet was found to induce neurobehavioral changes in mice that were on a normal diet, suggesting that changes in the gut microbiome could play a significant role in the cognitive deficits associated with obesity . However, the mediators of this effect remain to be established and it is unclear whether the effects attributed to the gut microbiota are independent of adipocyte-derived hormones such as leptin and adiponectin.
The previous sections have summarized recent findings indicating that adiposity has a complex and significant interaction with the age-related dementia. Midlife obesity seems to contribute and increase the risk for developing dementia, while late-life weight loss portends impending dementia. We propose the following model for how these changes in body weight/adiposity can contribute to age-related dementia (Figure). During midlife, obesity would lead to increased plasma leptin and subsequently leptin resistance, while plasma adiponectin levels will decrease. The net effect of the obesity related leptin resistance and low plasma adiponectin levels would be an increased risk of cerebrovascular damage, decreased resistance to AD pathology, and worsening systemic metabolic deficits, such as insulin resistance, which may independently worsen the cognitive decline . During late-life, AD pathology may cause the weight loss by potentially inhibiting hypothalamic circuits involved in regulating body weight . The weight loss in AD would lead to low plasma leptin, reduced leptin signaling, and high plasma adiponectin, resulting in adiponectin resistance. The net effect of reduced leptin signaling and adiponectin resistance would be similar to the midlife obesity changes, leading to increased cerebrovascular risk, decreased protection against AD pathology and worsening systemic metabolic deficits.
This model raises several important questions that need to be addressed. First of all, the high plasma leptin seen in midlife obesity and the high plasma adiponectin seen with AD are presumed to lead to leptin and adiponectin resistance respectively. While there is some evidence to suggest that leptin resistance due to obesity can contribute to worsening AD pathology and cognitive function , studies investigating adiponectin resistance in dementia are lacking. In fact, it has not been conclusively established if adiponectin has protective effects against dementia and AD pathology, much less if central adiponectin resistance occurs with age-related dementias. Therefore, careful studies investigating the role of adiponectin in AD are clearly needed. Furthermore, while the focus has been exclusively on leptin and adiponectin, as mentioned above there are other adipose-derived factor that may also be involved. Therefore, a systematic analysis of adipocyte function and adipocyte-derived factors associated with age-related dementias could potentially yield important new candidates. Importantly, due to the inherent metabolic differences between mice and men, findings in mouse models will need to be validated in human studies.
Investigating the effects of body weight and adiposity on age-related dementias also opens the possibility of developing new diagnostic tools and identifying novel therapeutic targets. Both leptin and adiponectin are potential new biomarker candidates for AD. As weight loss often precedes the cognitive symptoms in AD , identifying changes in adipocyte-derived factors either in isolation or as an adjunct to other biomarkers may further enhance the diagnosis of AD and related dementias. This could enable earlier diagnosis of the dementia leading to an increased window of opportunity to treat the disease. Additionally, assessing adiposity and adipocyte-derived factors could prove to be useful in therapeutic trials particularly in presymptomatic subjects without cognitive impairment, since it could provide additional information regarding the effectiveness of the drug early in the course of the trial.
Leptin and adiponectin remain active areas of pharmaceutical research, and both have the potential to be used as a therapeutic agent against age-related dementia. Preclinical studies in animal models of AD have demonstrated that leptin treatment can improve both AD pathology and cognitive function [83,84]. Therefore, leptin and its signaling pathways are attractive drug targets against dementia. Furthermore, since leptin treatment is most efficacious against conditions associated with low plasma leptin as opposed to conditions associated with leptin resistance , dementia patients who are losing body weight and have low plasma leptin levels might be good candidates for leptin treatment. Due to the paucity of preclinical studies on the effects of adiponectin in AD much less is known about the potential therapeutic benefit of adiponectin, but the recent development of adiponectin receptor agonists is likely to facilitate future drug development based on the adiponectin pathways .
We have briefly examined the accumulating evidence supporting the strong relationship between body weight/adiposity and age-related dementia. We highlighted two major adipocyte-derived hormones, leptin and adiponectin, as examples of how changes in adiposity can affect both cerebrovascular and AD pathology to contribute to age-related dementias. Furthermore, we proposed a model where midlife changes in body weight/adiposity leads to alterations in adipocyte-derived hormones to make the brain more susceptible to dementia, whereas late-life changes are one of the earliest signs of impending dementia and may contribute to the overall disease process. While significant advances have been made particularly in identifying the association between body weight/adiposity and dementia, our understanding of this interaction on a molecular and cellular level is still in its infancy leaving significant challenges as well as opportunities for the development of new diagnostic tools and novel drug targets against age-related dementias.
The authors gratefully acknowledge research support from the BrightFocus Foundation (MI), and NS37853 (CI). MI is the recipient of a Paul B. Beeson Award from the National Institute on Aging, the American Federation for Aging Research, The John A. Hartford Foundation and The Atlantic Philanthropies under award K08AG051179. The content is solely the responsibility of the authors and does not necessarily represent the official views of any of the funding agencies.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.