Maintaining Brain Health by Monitoring Inflammatory Processes: a Mechanism to Promote Successful Aging

Abstract

Maintaining brain health promotes successful aging. The main determinants of brain health are the preservation of cognitive function and remaining free from structural and metabolic abnormalities, including loss of neuronal synapses, atrophy, small vessel disease and focal amyloid deposits visible by neuroimaging. Promising studies indicate that these determinants are to some extent modifiable, even among adults seventy years and older. Converging animal and human evidence further suggests that inflammation is a shared mechanism, contributing to both cognitive decline and abnormalities in brain structure and metabolism. Thus, inflammation may provide a target for intervention. Specifically, circulating inflammatory markers have been associated with declines in cognitive function and worsening of brain structural and metabolic characteristics. Additionally, it has been proposed that older brains are characterized by a sensitization to neuroinflammatory responses, even in the absence of overt disease. This increased propensity to central inflammation may contribute to poor brain health and premature brain aging. Still unknown is whether and how peripheral inflammatory factors directly contribute to decline of brain health. Human research is limited by the challenges of directly measuring neuroinflammation in vivo. This review assesses the role that inflammation may play in the brain changes that often accompany aging, focusing on relationships between peripheral inflammatory markers and brain health among well-functioning, community-dwelling adults seventy years and older. We propose that monitoring and maintaining lower levels of systemic and central inflammation among older adults could help preserve brain health and support successful aging. Hence, we also identify plausible ways and novel experimental study designs of maintaining brain health late in age through interventions that target the immune system.

Maintaining brain health and preserving cognitive functioningas we age are critical for promoting autonomy and protecting against disability, dementia, and mortality [1, 2]. Brain health is affected by several modifiable factors, including cardiometabolic parameters and aspects of our lifestyles. Emerging evidence shows that these risk factors for brain health decline are also related to systemic levels of inflammation. Moreover, there is comparable evidence indicating that increases in levels of peripheral markers of inflammation are associated with age-related declines in brain health. Such evidence parallels longstanding findings from neurocognitive studies in populations of older adults living in the community, consistently indicating an association between higher inflammatory levels and lower cognitive levels and higher risk of cognitive impairment over time. (see Gorelick for a review [3])

Considering the above in aggregate, it is important to note that an association of heightened peripheral inflammatory markers with structural and functional brain changes in association with neurocognitive changes over the course of life is biologically plausible, because the CNS is intimately and continuously engaged in a two- way communication with the peripheral immune system (Figure 1).

Figure 1.

Communication between central nervous system and immune system. Triggers (cardiometabolic and lifestyle factors, or injuries of other nature) occurring anywhere in the body are communicated to the CNS via release of proinflammatory factors. An inflammatory response is initiated when monocytes/macrophages are activated by pathogens or tissue damage to release pro-inflammatory cytokines, including interleukin (IL)-6, IL-1β, and tumor necrosis factor (TNF)-α, and chemokines, such as IL-8. These chemical mediators coordinate a local inflammatory response, resulting in the recruitment and activation of leukocytes to the site of invasion/injury. They also enter peripheral circulation to stimulate a systemic response, which includes the synthesis and release of acute phase proteins, such as C-reactive protein and fibrinogen. The CNS responds to this information by initiating behaviors to adapt to such triggers (e.g., fever, reduced activity) and by releasing immune mediators to respond to the peripheral stimulus (green arrow). In normal circumstances, peripheral inflammatory responses are terminated quickly by the action of anti-inflammatory factors released locally and systemically to shut off the inflammatory response, leading to the adaptive restoration of a lower inflammatory state.

To elaborate, multiple behavioral and disease processes can lead to heightened levels of systemic inflammation. In the case of acute illness, the inflammatory response is tightly controlled by a diverse set of regulatory mechanisms to handle the extent of the precipitating infection or injury, resolve quickly and restore health. In contrast, in older age these regulatory processes seem to be impaired, resulting in augmented and/or longer inflammatory response with subsequent tissue damage. For example, aging is associated with a chronic increase in the level of systemic inflammation [4, 5], that is relatively stable over extended periods [6], and predicts risk for a range of age-associated diseases, including cardiovascular disease, type 2 diabetes, frailty and general functional decline [7–10].

Additionally, recent evidence shows that older brains mount exaggerated central inflammatory responses to peripheral inflammatory stimuli even in the absence of clinically overt neuropathology [11, 12]. Thus, prolonged exposure to inflammatory triggers, higher levels of peripheral pro-inflammatory mediators that communicate with the CNS, lower levels of anti-inflammatory factors, and the “priming” of immune target cells in the brain can all contribute to heightened or prolonged neuroinflammation (Figure 2).

Figure 2.

Biological rationale for a relationship between peripheral inflammatory factors and brain health. Proinflammatory cytokines cause greater endothelial permeability and adhesively of the blood brain barrier (BBB) and may reach the CNS through several pathways: 1 = through the post-inflammatory dysfunctional endothelium of the BBB; 2 = through binding to specific receptors on the BBB cell, and subsequently triggering mRNA production of pro-inflammatory factors; 3 = via diffusion through “nude” areas. 4. Via stimulation of the adrenergic system through the vagal nerve. Upon stimulation, the microglia in the CNS is also capable of producing cytokines directly, thus cytokines levels can increases in the CNS and can trigger brain inflammation even if they are not transferred from the peripheral blood into the brain parenchyma.

This review assesses evidence for, and potential mechanisms of, the contribution of peripheral inflammatory markers to brain health and the potential for prevention and intervention. Specifically, it remains to be determined whether and to what extent higher levels of peripheral inflammation reflect the presence and the severity of brain abnormalities and of brain inflammation in community-dwelling older adults. The answer to this question is critical to understand whether serum measurements of inflammatory markers could serve as biomarkers of underlying brain health. The potential use of these markers is very attractive because these measures present low risk, they are reasonably easy to be obtained, inexpensive, and reliable. We caution in this review, however, that researchers cannot assume that circulating levels of inflammatory mediators provide one-to-on measures of immune-derived inflammatory processes, as many cells other than immune cells produce these signaling proteins, including adipocytes and endothelial cells [13, 14]. Indeed, adipose tissue is a key source of peripheral IL-6, with adipocytes (particularly those residing in visceral adipose tissue) producing 10–35% of circulating levels [13]. Thus, this review will also consider the possible contribution of obesity to neuroinflammatory processes.

This review is organized into three sections: Section One describes and defines the determinants of brain health and proposes a model for potential inflammatory mechanisms; Section Two assesses evidence for the contribution of peripheral inflammatory markers to declining brain health; Section Three identifies opportunities to help maintain brain health late in age, through interventions targeting the immune system.

The complexity of the inflammatory and immune systems will not be addressed in detail in this review. The reader is invited to consult the many excellent reviews published on these topics [15–24].

1. Brain health in older age: measurements and potential inflammatory mechanisms

This section provides the background for a relationship between inflammation and brain health decline in older adults. We first summarize the main measurements of brain health that is brain function and structure using cognitive testing and neuroimaging methods, and brain indicators of inflammation (1.1.). We then briefly review the main cardiometabolic and lifestyle factors related to brain health (1.2.) and their relationship with inflammatory factors (1.3). Lastly, we describe potential mechanisms that may contribute to age-related declines in brain health (1.4), considering the role of cardiometabolic and lifestyle factors, and related systemic inflammation.

1.1. Overview of measurements of brain function and structure in older adults

Cognitive decline generally begins in the late 20s and progresses at a consistent rate across adulthood [25]. Some domains, including working memory, attention, and executive functioning, are more affected than others in older age, and age-related declines are predominant in processing speed [26, 27]. Compared with younger individuals, healthy elderly also perform less well on measures of delayed recall and recognition, and mental flexibility [28–30].

Overall, declines in brain function among older adults are associated with brain atrophy, hyperintensities in the white matter, and lacunae, as well as with amyloid deposits and metabolic changes [2, 31, 32]. However, these neuroimaging findings accompany numerous conditions and are not disease-specific. For example, global brain atrophy and hippocampal atrophy commonly accompany a variety of memory impairment syndromes, including Alzheimer’s, Vascular, and Parkinson dementias [33], as well as memory impairment in healthy elderly adults [34]. They are also frequently seen in those who have suffered severe hypoxia and traumatic brain injury [35]. Observational studies have also shown associations of higher prevalence of infarcts, white matter hyperintensities, and atrophy with inflammatory markers in older adults in the absence of overt clinical disease [36–43].

Applications of advanced neuroimaging methods that quantify micro-structure and metabolism with increased spatial resolution and regions-of interest approaches have begun to change our understanding of brain aging processes. For example, atrophy in older adults free from dementia primarily affects the prefrontal cortex and the subcortical regions, including the basal ganglia and the hippocampus [44–50]. Similarly, the very recent application of the Pittsburgh-Compound B to study amyloid deposition in cognitively normal older adults has shown focal amyloid deposition within fronto-parietal and posterior cingulated networks, known to be related to executive control function [51, 52]. Additionally, Diffusion Tensor Imaging (DTI) has uncovered the presence of microstructural abnormalities that remain otherwise invisible on conventional MRI [53, 54] Specifically, higher mean diffusivity and lower fractional anisotropy from DTI indicate loss of homogeneity of brain tissue and are observed in brain abnormalities that develop with aging [55] and are associated with slower processing speed [56]. Overall, these neuroimaging findings support the clinical observations that older adults who are non-demented primarily display executive and memory dysfunction.

Recent methodological developments in neuroimaging are also beginning to provide markers of brain inflammation in vivo. For example, MRI combined with Gadolinium as a contrast agent can provide images of brain vascular permeability, a correlate of blood-brain barrier alterations and thus a potential inflammatory marker. Notably, however, this newer method has a limited spatial resolution. Additionally, greater vascular permeability occurs in advanced stages of inflammation; thus, this method would not yet appear to capture subtle signs or earlier stages of inflammation. Ultra high-field neuroimaging combined with injections of iron oxide nanoparticles is emerging as a neuroimaging method to visualize inflammatory phenomena at the cellular level by marking activated microglia in early inflammatory stages. This method shows greater regional binding of contrast agent even in the absence of frank disruptions to blood-brain barrier permeability and with greater spatial resolution than Gadolinium contrast approaches (see references [57–60] for reviews). Magnetic resonance molecular imaging has been used in acute ischemic stroke to identify endothelial activation by targeting biomolecular agents [61]. Although very promising, to date these methods have been applied mostly in patient populations and in animal disease models [62–69].Moreover, compared with many of the other modalities reviewed here, these neuroimaging approaches are more invasive and less feasible for large-scale and longitudinal epidemiological studies.

1.2. Association of cardiometabolic and lifestyle factors with brain health

Although the biological bases for the selective spatial distribution of brain abnormalities accompanying aging remain unclear, growing evidence points to the vascular anatomy and hemorrheological characteristics of the flow within these regions as predisposing factors. Fronto-subcortical networks, deep striato-capsular and lenticular regions, and finally the deep white matter are localized within the watershed areas of the cerebral blood supply system, and thus have poor collateral vascularization as well as low cerebral perfusion pressure. These predisposing conditions make the fronto-parietal and subcortical regions more susceptible to disturbances of cerebral perfusion [70].

Several factors related to aging, in particular hypertension, arteriosclerosis and glucose dysregulation, contribute to disturbances of cerebral perfusion, and are associated with cognitive decline and related changes in brain structure and metabolism [71–73]. Among adults 75 years and older, hypertension is the most common risk factor longitudinally associated with increased white matter hyperintensities, along with arteriosclerosis, and hemodynamic dysregulation [44, 74–77]. A role of cerebral amyloid angiopathy in altering the blood brain barrier permeability and contributing to small vessel disease has also been proposed [78, 79]. Pulse pressure (the difference between systolic and diastolic blood pressure, which is associated with vessel stiffening and vascular aging) has also been cross-sectionally associated with white matter hyperintensities in older adults [80]. Glucose-related disturbances have also been associated with brain structural changes [81–83].

Among lifestyle factors, attention has recently focused on the role of diet and physical activity as well as to social and cognitive engagement in brain aging [84]. This review focuses on these modifiable factors because they may be amenable to cost-effective, non-pharmacological interventions that promote other aspects of whole-body and brain health, including lower cardiovascular risk and mood improvements [85]. Higher levels of adiposity are associated with higher circulating levels of inflammatory factors [86], and predict temporal lobe and global brain tissue atrophy, a reduced integrity of white matter tracts, cognitive decline, and even incident dementia [87–91]. Corroborating these human epidemiological findings, experimental and animal studies further demonstrate that obesity predicts deficits in hippocampal-dependent learning and memory [92–94]. We have shown [95, 96] that greater percent body fat is associated with poorer cognitive function, and lower hippocampal grey matter volume among relatively healthy mid-life adults, 30–54 years of age, suggesting that the cognitive decline associated with adiposity begins well before clinically significant deficits. It is plausible that long-term exposure to adiposity could presage accelerated brain and cognitive aging. In addition to adiposity, atherogenic diet has been associated with higher levels of brain inflammation in animal models [97].

Several lines of evidence suggest that level of social engagement and other social behaviors related to lifestyle predict cognitive aging, independently of other known risk factors (e.g., depressive symptoms, frailty, and cardiovascular and cerebrovascular health status). For example, higher scores on a social engagement scale (assessing membership in diverse social groups and frequency of social interactions) predicted less of a 12-year decline in global mental status among 2,812 community dwelling adults older than 65 yrs [98]. Such findings have been corroborated and extended in studies of social participation [99–101] and social-network size in particular, as well as in studies of social engagement and working memory and related executive control functions [102].

Recent studies indicate that interventions that decrease cardiometabolic risk in older adults [75, 103–108] and improve lifestyle parameters such as diet, social engagement, and exercise [109, 103, 110, 111]can delay cognitive decline and slow the progression of structural brain abnormalities, specifically signs of subclinical cerebral disease [75], white matter hyperintensities [112], and brain tissue atrophy. Critically, there is also evidence that human brains retain substantial plasticity, or the potential to change in function and structure, and several reserve capacities even in their late eighties [85]. Therefore, it is very important to understand the mechanisms underlying the relationship between these factors and brain health.

1.3. Association of cardiometabolic and lifestyle factors with inflammation

Many of the risk factors for brain health decline, including higher homocysteine, oxidized lipids, free radicals, and Angiotensin II are directly toxic to blood-brain barrier permeability, and can trigger vascular inflammatory events [19, 71]. Of relevance to the current review, activated immune cells involved in the inflammatory response are a primary source of oxidative stress [113, 114]. Indeed, evidence shows a positive association of systemic markers of inflammation with oxidative burden [115–117], which may play a role in the pathogenesis of neurodegeneration. Additionally, small vessel abnormalities, occlusions and/or lower blood flow can cause an inflammatory response in the vascular districts via poor oxygenation and lower glucose delivery to neurons, which in turn trigger subsequent cellular modifications and changes in beta protein metabolism. Specifically, higher levels of beta fragments, due either to a genetic predisposition (APO-e4) and/or to neurodegeneration following vascular insults, can function as “irritants” and trigger an inflammatory response [118, 119]. These findings suggest that a main mechanism linking cardiometabolic factors with inflammation may be localized to the vascular district, specifically, an induced state of vascular inflammation with release of cytokines and chemokines locally and systemically across the blood brain barrier.

It is well established that adipose tissue is a potent source of peripheral inflammatory mediators, with adipocytes producing 10–35% of circulating IL-6 levels [92]. Further, mononuclear cells of obese people are “primed” to express pro-inflammatory cytokines [92]. Thus, obesity can reasonably be viewed as an inflammatory condition, with circulating levels of IL-6 and other inflammatory mediators covarying positively with adiposity. Conversely, greater social engagement and related social factors (e.g., social support) are inversely related to expression of inflammatory markers [24]. Animal evidence indicates that manipulation of the social environment, particularly social isolation, increases pro-inflammatory cytokine expression (specifically IL-1beta) in the hippocampus, and decreases brain-derived neurotrophic factor (BDNF) expression [120]. Moreover, social isolation impairs hippocampal-dependent memory, and this effect is blocked by intra-hippocampal injection of an IL-1 receptor antagonist, which also prevents BDNF decreases [120]. However, we are unaware of studies linking social engagement with indicators of brain morphology and related aspects of cognition that co-vary with inflammation, as would be predicted by converging lines of epidemiological, animal, and neuroimaging work.

1.4. Potential mechanisms underlying decline in brain health

The pathways linking pro-inflammatory levels in the circulating blood with brain health have been examined in detail in many excellent reviews and are only briefly summarized here (Figure 2) [15–17, 21–24, 105, 121–125]. Direct pathways include: a) active passage across the blood brain barrier; and b) passive diffusion from the plexus choriodeus and other blood brain barrier-nude regions into other brain regions. Cytokines can also affect the CNS through indirect pathways that include: a) stimulation of the vagal nerve, and b) inflammation of the endothelial cells, which in turn produce inflammatory factors.

Regardless of the nature and location of the originating trigger, increased levels of inflammatory factors cause loss of vascular anti-adhesive and impermeability properties of the blood brain barrier — thus allowing adherence and entrance of inflammatory cells into the CNS parenchyma, oxidation of LDL and accumulation of an inflammatory locus within the vascular wall. Furthermore, inflammatory factors in the CNS and peripheral pro-inflammatory factors stimulate the production of central pro-inflammatory factors by microglial cells in discrete brain regions (Figure 2). The production and release of pro-inflammatory factors from the microglia ultimately contribute to neurotoxicity and brain parenchyma abnormalities. These in turn can sustain the inflammatory reaction, thus maintaining a vicious cycle.

It has been suggested that increases in low-grade systemic inflammation, such as those that accompany cardiometabolic risk and aging as described above, sensitize the brain response to inflammatory factors. Specifically, in vivo and in vitro animal studies, as well as post-mortem human studies, indicate that older brains are in a native heightened inflammatory state even in the absence of overt disease [16, 17, 22–24, 97, 121, 126]. In addition to the chronic exposure to low-grade systemic inflammation, this state has also been attributed to an age-related change in microglia to a phenotype that release inflammatory factors and respond excessively to stimulation. Another possible cause of this primed state is the life-time exposure to cellular debris from pruning and remodeling as well as neuronal senescence processes that act as local “irritants” and trigger inflammatory reactions. It has been shown that there is a relative increase in expression of pro-inflammatory genes in older mice brains [21]. Since peripheral pro-inflammatory cytokines stimulate the production of central pro-inflammatory cytokines by microglial cells, a sensitization of neuroinflammatory responses may play a role in the premature brain and cognitive aging that accompany increased systemic inflammation.

1.5. Implications for mechanisms linking inflammation and brain health

In summary, risk factors for declining brain health in older adults include numerous cardiometabolic conditions and lifestyle factors that each relate to inflammation. Prolonged exposure to these conditions and risk factors might trigger and then sustain a systemic inflammatory response as well as a focal response in the brain. The increasing levels of pro-inflammatory factors in the blood may in turn affect the inflammatory state of the CNS through direct and/or indirect pathways as described above. Since the CNS and the immune system are in constant communication, a heightened inflammatory state in the periphery can plausibly occur concurrently with a heightened inflammatory state in the brain and vice versa. For example, cardiometabolic risk factors may trigger vascular inflammation in multiple vascular districts concurrently, including the peripheral and cerebral districts. Once these factors cause vascular inflammation in the cerebro-vascular district, then endothelial cells of the blood brain barrier may produce and release cytokines within the CNS. The preferential genetic expression of pro- versus anti-inflammatory factors in older age might also facilitate either the start and/or the maintenance of an inflammatory state. In addition to the biological plausibility of an association between inflammation and brain health, there is recent consistent evidence of associations between peripheral cytokines and brain health as described in the next section.

2. Contribution of peripheral inflammatory markers to declining brain health

A number of peripheral blood markers of inflammation have been examined in studies of brain aging. Key markers include interleukin (IL)-6, IL-1β, and tumor necrosis factor (TNF)-α, and chemokines, such as IL-8 [127]. In contrast to IL-1β and TNF-α, which decay rapidly, levels of IL-6 and C-Reactive Protein can be reliably detected in peripheral circulation and are widely assumed to reflect systemic levels of inflammation. Consistent evidence from animal studies and disease models indicates that peripheral inflammatory mediators, such as interleukin (IL)-6, modulate central inflammatory processes that affect cognitive function [24, 128–130]. Notably, receptors for IL-6 are concentrated in the hippocampus and prefrontal cortex [130–132].We have recently shown that higher IL-6 levels are associated with smaller hippocampal and prefrontal gray matter volume and with lower executive function and memory performance in otherwise healthy middle-aged community dwelling adults [95, 96]. It is likely that other peripheral proinflammatory mediators, such as IL-1 beta and TNF-alpha, also play a role in the induction of central neuroinflammation; however it is harder to reliably assess these factors due to their low circulating levels and short half-lives. Research suggests that IL-1β, IL-6, and TNFα are involved in neurophysiological processes that subserve cognitive function and brain health [133]. Animal studies indicate that IL-1 affects long-term potentiation and possibly neurogenesis, IL-6 is an important regulator of neurogenesis, and TNFα influences cognitive function through direct effects on long term potentiation and synaptic scaling [133]. Experimental studies using animal models demonstrate that long-term exposure to IL-6, as seen in normal aging and certain neurodegenerative diseases, can interfere with neurocognitive functioning by impairing adult neurogenesis [21, 134].

2.1. Evidence from patients with primary neurodegenerative disorders

Studies in patients with ischemia and stroke indicate that peripheral levels of the inflammatory markers IL-6 and TNF-α provide markers of the severity of the underlying lesions as well as predict patient prognosis [135–140]. One study showed that patients with recurrent chronic brain infarctions had higher levels of monocyte chemoattractant proteins and C-reactive protein [141].

Studies of patients with Alzheimer’ and Parkinson’s diseases show an association between blood levels of inflammatory markers and severity of brain functional impairment [142–144]. Still unknown is whether higher levels of peripheral inflammatory markers indicate underlying declining brain health in community-dwelling older adults. The majority of longitudinal studies examining the relationship of inflammation with brain health in older adults have focused on dementia populations, with comparatively few examining community-dwelling older adults [145].

2.2. Evidence from patients with primary chronic peripheral inflammatory diseases

Studies of rheumatoid arthritis and lupus strongly support the presence of immune-to-brain communication pathways [146–148]. Pioneering neuroimaging studies in humans report brain metabolic alterations and volumetric reductions in patients relative to controls, suggestive of underlying brain degeneration [147, 149–151]. It is not clear whether these inflammatory diseases affect the brain via increased levels of peripheral inflammatory factors (e.g., via the mechanisms described in 1.3 and illustrated in Figure 2) or by pain-mediated mechanisms, which have also been suggested [152]. One recent study of patients with rheumatoid arthritis suggests an association between TNF alpha blocking agents and worse micro-structural brain characteristics [153].

A review of seven studies of patients with arthritis taking anti-inflammatory medications (NSAIDs and steroids) found strong evidence for decreased risk for Alzheimer disease [154]. However, a similar association was not observed in patients with rheumatoid arthritis taking corticosteroids [71], raising the possibility that the effects are specific to NSAIDs. In this regard, two reviews of observational studies of non-demented populations [155, 156] concluded that NSAIDs intake is associated with higher cognitive function, slower progression of cognitive decline and lower risk of developing Alzheimer’s disease. However, NSAID intake in demented adults appears to have no or little effect on progression of cognitive decline. These preliminary results should spur further research to test the hypothesis that NSAID intake is most beneficial early, rather than late in the course of the neurodegenerative process. The mechanisms underlying the effects of NSAIDs have not been examined. It is not clear whether NSAIDs’ neuroprotective effects result from the dampening of inflammatory processes or from the direct lowering of Aβ levels. Mediating effects on pain have also been suggested as potential mechanisms.

2.3. Evidence from experimental studies

Experimental studies inducing higher blood levels of pro-inflammatory markers via peripheral immune stimuli in young adults have shown a slowing of information processing and negative mood changes within 12 hours [157–159]. Related functional neuroimaging data obtained within this timeframe show concurrent changes in patterns of neuronal activation and heightened response to emotional-related stressors. This is consistent with the loop linking peripheral stressors with adaptive neuronal responses (Figure 1). Initial experimental studies using peripheral administration of cytokine-blocking agents also show rapid and short-term changes in the CNS [160–162]. The work by Hess et al [161], suggests that TNF neutralization in a mouse model of rheumatoid arthritis has rapid effects on neuronal activity in regions related to pain perception six hours after drug intake, without detectable changes in peripheral cytokines levels. Another study of rheumatoid arthritis patients taking naproxen [162] found a reduced hypothalamic-pituitary-adrenal (HPA) axis response in the short, though not long term. These initial findings suggest that inflammatory factors may affect the CNS directly.

While we take stock of the mounting evidence that cytokines are associated with brain function and structure, we also notice that most studies are cross-sectional, leaving open the debate as to whether raised peripheral inflammatory markers’ levels are the cause or consequence of brain health decline [163]. Initial evidence indicates that early interventions on factors that initiate inflammatory processes may also decrease risk for brain health decline. Future work is needed to understand the exact mechanisms of the relationship between peripheral inflammation and brain health decline (see Implications for future studies, below).

3. Opportunities for preventative and therapeutic intervention

The central notions that the aging brain becomes increasingly subject to systemic inflammation in later life, and retains a remarkable level of plasticity well into late life, each provide a basis for understanding how several lifestyle factors and interventions can have preventative and therapeutic benefits.

Perhaps the most salient lifestyle factor amenable to intervention is physical activity. It is well established that a sedentary lifestyle confers risk for several syndromes associated with heightened levels of inflammation—including obesity, diabetes, cardio-vascular disease, depression, and dementia. Even moderate levels of physical activity are associated with decreased levels of inflammation [84, 164–168], which could plausibly slow the progression of functional and cognitive decline in aging [169, 170]. For example, aerobic exercise regimens involving regular and brisk walking among older adults (approximately 30 minutes per day/most days of the week) have been associated with improvements in cognitive performance and with changes in patterns of brain activation in prefrontal and parietal cortices that more closely resemble those seen in younger individuals [171, 172]. Moreover, regular physical activity has been associated with an increase in the volume of gray matter and related cognitive functions [111, 173, 174]. An open research question is whether physical activity may improve or maintain aspects of brain health in aging by directly down-regulating inflammatory processes. The ongoing intervention trials SMART [175] and LIFE-Main [176] may provide answers to these questions because they will monitor longitudinal changes in physical activity and in brain function, and are also collecting and storing sera that might be employed to quantify levels of peripheral inflammation.

Weight gain in mid- and later life (possibly related to low levels of physical activity), is associated with reduced brain tissue volume and associated cognitive changes ([177], also see above). One recent study also reports that weight loss is associated with decreased levels of systemic inflammation and concurrent increases in brain activity within memory-related regions [178]. Most of these studies have focused on adults in their sixties or younger. Weight changes in very old adults follow differential patterns and are linked with frailty and other co-morbidities; their role in improving brain health requires further investigation.

It is noteworthy that particular dietary aspects of energy balance and food choices throughout life could plausibly affect brain aging and be targeted by intervention. In their recent review, Jang and Johnson [179] suggest that a diet rich in flavonoids- that is fruits and vegetables- might reduce the age-related priming of microglia and/or the magnitude of peripheral inflammatory reactions and consequently assuage the severity of cognitive decline. Additionally, there is emerging evidence that higher intake of key ω-3 fatty acids – eicosapentaenoic and docosahexaenoic acids (EPA, DHA), found primarily in cold water fatty fish and in fish oil supplements –is associated with a reduction in systemic inflammation, decreased risk for premature cognitive impairments, and higher cognitive function and regional brain tissue volumes. Interestingly, whereas EPA and DHA comprise about 4% of the fatty acids in plasma and red blood cells, 14–30% of the fatty acids in the phospholipids of human brain tissue, particularly gray matter tissue, are either EPA or DHA [180, 181]. Hence, the types and relative concentrations of fatty acids in the diet may, over time, affect brain tissue composition and integrity—specifically through known diffusion or active transport pathways [182, 183]. Finally, it is notable here that ω-3 and ω-6 polyunsaturated fatty acids are precursors to prostaglandins and leukotrienes. EPA is a substrate for anti-inflammatory prostaglandins and leukotrienes, whereas the principal ω-6 fatty acid, arachidonic acid, is a substrate for pro-inflammatory mediators [184]. Viewed as inflammatory precursors, it has been recommended that dietary intake of fatty acids be regarded as preventative strategy to decrease levels of systemic inflammation and thereby promote physical and mental health [185].

In addition to the functions described above, EPA and DHA have functional and morphological roles in the brain, being associated with complexity of dendritic branching, synaptic plasticity, and long-term potentiation within the hippocampus; ω-3 fatty acid deficiency also simplifies dendritic arborization in the cortex [186, 187]. Moreover, increasing ω-3 fatty acid intake by oral DHA or EPA administration increases dendritic spine expression [188, 189]. Finally, ω-3 fatty acid deprivation has been shown to reduce brain-derived neurotrophic factor and neuron size in the hippocampus, hypothalamus, and parietal cortex [186, 190–192]. Consistent with work in animals, we found that self-reported dietary intake of EPA and DHA was associated with greater gray matter volumes in the prefrontal and medial temporal lobes [193]. The recent OmegAD Study indicates that treatment with ω-3FAs in AD patients did not substantially modify inflammatory markers’ levels in the blood or spinal fluid, suggesting that fatty acids protective effects might be more evident among non-demented adults [194].

In the aggregate, promising evidence regarding habits of lifestyle amendable to intervention, namely physical activity, weight maintenance, and intake of flavonoids and ω-3 fatty acids, suggests that brain aging and brain health could be preserved and positively affected by preventative behavioral changes and programs. However, it remains unclear precisely how much flavonoids or fatty acid individuals should consume on a daily or weekly basis, over what period of time, in what form (e.g., from food or supplement), and beginning at what age to achieve improvements or maintenance of brain and cognitive health in later life. Also unclear is how EPA and DHA consumption interacts with other factors, such as physical activity or use of NSAIDs, in affecting inflammatory processes linked to aging. In this regard, epidemiological studies in ethnic groups with life-time exposure to high levels of fatty acids, specifically Japanese, Icelandic and Norwegian cohorts may provide insights into the effectiveness of diet in different genetic populations [195].

4. Implications for future studies

It is possible that low-grade chronic inflammatory and neurodegenerative conditions contribute to a recursive cycle of damage that may worsen brain health over time. The greatest challenge in efforts to maintain brain health is to identify how to interrupt this inflammatory cycle. While we have described several targets for low cost interventions, we propose that advancements in this field require the integrated application of novel methods, longitudinal designs and unique populations that move beyond maintenance of a favorable profile of cardiometabolic and lifestyle factors.

First, the rapid advances in neuroimaging technology can now provide direct measures of brain inflammation and blood brain barrier permeability which were not available until only a few years ago. (see references in Section one). Multimodal neuroimaging measures of the blood brain barrier, concurrent with brain micro-structure and quantification of amyloid deposition can quantify early abnormalities at the micro-structural and metabolic level. Addressing these abnormalities before they build into brain macro-structural damage, disease and overt inflammation may have a substantial impact on maintaining cognitive function in older persons. There is also the need to conduct studies in community dwelling older adults to relate markers inflammation in the blood to markers of inflammation in the cerebrospinal fluid (CSF). Values of inflammatory markers in the CSF of older adults in the absence of overt disease are unknown. To date, only few reports have documented levels of inflammatory factors concurrently in the blood and the CSF. Most of these studies have examined patients populations [194, 196–198] and report a significant correlation between cytokines’ levels in the blood and in the CSF. Overall, these results suggest that plasma levels of inflammatory factors might be candidate biomarkers of underlying neuroinflammatory processes in patients’ population. We are aware of one normative study in healthy volunteers [199] 20–90 years old, without history of acute or chronic inflammatory disease. The study showed that IL-6,IL-8, IL-10 and the soluble TNF receptor are detectable both in the blood and CSF, and that age did not significantly affect the plasma to CSF ratio of inflammatory markers. However, this study did not report the strength of the correlation between markers in the two compartments. Longitudinal study designs with repeated CSF and blood levels measurement of inflammatory factors are warranted to quantify the repeatability and reliability of blood levels of inflammatory markers in predicting CSF markers. These studies could also clarify whether inflammatory levels increase concurrently in the blood and spinal fluid or are independent of each other. Specifically, studies applying serial and integrated measurements of levels of inflammatory markers in these two compartments, together with multimodal neuroimaging and cognitive testing can help clarify whether there is a causal pathway between peripheral inflammatory factors and declining brain health and higher brain inflammation.

It is not clear whether inflammatory levels increase concurrently in the blood and spinal fluid or are independently of each other and values in the absence of overt disease are also unknown. Studies applying serial and integrated measurements of levels of inflammatory markers in these two compartments, together with multimodal neuroimaging and cognitive testing can help clarify whether there is a causal pathway between peripheral inflammatory factors and declining brain health and higher brain inflammation. Novel study designs applying international epidemiological collaborative approaches would be particularly helpful to understand the impact of ethnicity, lifestyle and ecological variations on the interplay between inflammation and brain health.

Secondly, there are new drugs with potent anti-inflammatory effects that appear to have an effect on the brain. The application of cytokine blocking agents has revolutionized the treatment of peripheral inflammatory diseases, specifically of patients with rheumatoid arthritis. However, with the exception of one study, the effect on the brain has not been directly quantified concurrently with symptomatology of the main disease. Ongoing trials to test TNFα receptor antagonists in patients with rheumatoid arthritis should include the measurements described above to quantify CNS changes related to drug intake. If these drugs do affect the brain and slow the progression of cognitive decline among patients with chronic inflammatory conditions, then there will be new avenues for exploration of their application in the prevention of brain declines that accompany normal aging.

A unique opportunity to study the course of inflammatory diseases and cognitive decline in aging adults is presented by the cohort of patients with multiple sclerosis and HIV who are now living into older age. Studies applying the measurements described above can clarify the interaction between disease- and age-related brain changes, and can provide biomarkers to quantify the effects of pharmacologic interventions on either and/or both processes. These findings can be particularly important among those who convert to dementia.

Further, the application of advanced neuroimaging methods can model personalized medicine approaches. Neuroimaging may assist in identifying the optimal window of time for administration of NSAIDs. Preliminary evidence suggests that NSAIDs may be most useful in delaying cognitive decline if taken before the onset of dementia. However, it is not clear how “early” would be too early. NSAIDs lowering effects of the inflammatory response even when clinically indicated may cause harm, because the inflammatory response is at its core an adaptive response aimed at restoring health. Specifically, preliminary studies indicate that NSAIDs might harm the inflammatory response to Aβ amyloid depositions [155]. Thus, neuroimaging markers of brain inflammation, structure, and metabolism may serve to identify the window of time to maximize the therapeutic effects of NSAIDs.

In conclusion, addressing the existing empirical gaps and pursuing the lines of basic and clinical inquiry discussed in this review will help to better characterize the complex interplay between the central nervous and immune systems in the context of brain and cognitive aging. Understanding the mechanisms and causal pathways linking peripheral inflammatory markers with brain health and neuroinflammation will help understand whether inflammatory markers can serve as surrogate markers of declining brain health. To date, the application of these markers to quantify risk of recurrent brain vascular events has been shown in patient populations. While it is biologically plausible that peripheral inflammatory markers reflect underlying brain health, the validity of peripheral inflammatory factors as prognostic/diagnostic markers of brain health will need to be tested in community-dwelling adults using advanced neuroimaging methods.