Citrus Polyphenols in Brain Health and Disease: Current Perspectives

In addition to essential micronutrients such as vitamin C, citrus fruits represent a considerably rich source of non-essential bioactive compounds, in particular flavanones which form a sub-set of the flavonoid group. Preclinical studies have demonstrated the neuroprotective potential of citrus flavonoids and have highlighted both the well-established (anti-inflammatory and anti-oxidative properties), and newly emerging (influence upon blood-brain barrier function/integrity) mechanistic actions by which these neurological effects are mediated. Encouragingly, results from human studies, although limited in number, appear to support this preclinical basis, with improvements in cognitive performance and disease risk observed across healthy and disease states. Therefore, citrus fruits – both as whole fruit and 100% juices – should be encouraged within the diet for their potential neurological benefit. In addition, there should be further exploration of citrus polyphenols to establish therapeutic efficacy, particularly in the context of well-designed human interventions.


INTRODUCTION
A varied and balanced diet, rich in plant-derived material (fruit and vegetables), is widely regarded as optimal for maintaining overall health and wellbeing. Unfortunately, typical modern dietary preferences are increasingly skewed in favor of nutrient poor, calorie dense, ultra-processed food choices. Combined with the background of an aging population, circumstances have produced a 'perfect storm' leading to a surge in metabolic disease burden, including rising incidence of neurodegenerative conditions which present significant societal and economic challenges.
Plant bioactives such as polyphenolic compounds have shown potential in negating the metabolic disease process (Chiva-Blanch and Badimon, 2017). Polyphenol describes a broad range of non-essential, usually naturally occurring compounds of which fruits and vegetables represent a significant source. The use of polyphenolic compounds in a therapeutic capacity is being increasingly considered, as we 'return to nature' to find sources of potential drug candidates (Xiao, 2017;Bellavite and Donzelli, 2020). Although currently lacking confirmation through rigorous long-term randomized control trial (RCT) evidence, the use of polyphenolic compounds in the prevention of cardiovascular disease appears promising (Parmenter et al., 2020). Indeed, an inverse relationship between polyphenolic compound consumption and cardiovascular disease mortality, risk and biomarkers are consistently highlighted in observational and preclinical studies. Given that cardiovascular health is a contributing factor to brain health, and cardiovascular disease is a significant risk factor for various forms of dementia, the benefits of polyphenols may also extend to brain functions.
Intriguingly, the mechanistic basis by which polyphenols may contribute to brain health and cognition extends further than those associated with cardiovascular health to include antiinflammatory and anti-oxidative capabilities (Vauzour et al., 2017;Flanagan et al., 2020). In addition, the developing role of the gut microbiota and its interplay with plant bioactives offers a novel route by which physiological effects can be exerted (Kumar Singh et al., 2019;Catalkaya et al., 2020). Continuing to reveal the underlying biochemical processes to which these compounds exert their effects is essential to utilizing their potential in brain health and disease, however, the importance of evaluating these compounds clinically cannot be underestimated. In this review, we will examine the current evidence on polyphenols and their effects on brain health, focusing specifically on citrus polyphenols since these are one of the main groups of flavonoids in the European diet.

PLANT POLYPHENOLS
Synthesized by plants to form either structural/functional tissue, or as a defense mechanism against pathogens and herbivores, polyphenol describes a compound with a chemical structure containing one or more phenolic rings (Frank et al., 2020). Further classifications can be made based upon phenolic ring numbers, and associated structures binding these rings together (e.g., oxygenated heterocycle) (Cutrim and Cortez, 2018). With an estimated 15,000 varieties currently identified, flavonoids represent a considerably diverse class of polyphenol (Xiao, 2017). The structural basis of the flavonoid is a 15-carbon atom structure, forming a C6-C3-C6 heterocyclic skeleton, consisting of two benzene rings, linked to a heterocyclic ring (flavon ring). Ubiquitous in plant-based diets, flavonoids are the main polyphenolic component of citrus fruits. Differing heterocyclic ring oxidization gives rise to further subclasses: flavanones, flavonols, flavones, and anthocyanins (present in blood oranges). Citrus fruits (Rutaceae) such as oranges, grapefruit, lemons and limes are a particularly rich source of flavanones and are therefore sometimes called citroflavonoids. Flavanones are also present in other food sources such as seasoning herbs (e.g., rosemary, oregano, peppermint), but in much lower concentrations (Rothwell et al., 2013).

MARKERS OF BRAIN HEALTH AND FUNCTION
Markers of brain health and disease are key to understanding and treating neurological conditions. Development of sophisticated neuroimaging techniques [e.g., Positron Emission topography (PET) and functional Magnetic Resonance Imaging (fMRI)] has enabled subtle differences in brain activity, metabolism and structure to be monitored and distinguished between healthy and disease states (Agosta et al., 2013). Future incorporation of artificial intelligence techniques, e.g., deep learning, will likely enhance the prospect and feasibility of early detection of neurological conditions (Noor et al., 2020). However, the cost, expertise, and scalability of neuroimaging limits widespread use. Therefore, biological markers, detectable in the periphery (commonly cerebral spinal fluid or blood) remain equally important, and better suited for screening purposes. Below we discuss some specific markers of brain health in more detail.

Neurotrophins
Neurotrophins are blood-brain barrier permeable signaling molecules with a crucial role in synaptic plasticity, neuronal cell survival and neurogenesis/synaptogenesis (Gibon and Barker, 2017). The presence of neurotrophins in the peripheral circulation makes them an ideal candidate for monitoring brain health and function. Of the neurotrophins, brain-derived neurotrophic factor (BDNF) has been the most extensively studied in the context of brain health, with peripheral levels associating with cognitive performance and neurological diseases (Nascimento et al., 2014;Küster et al., 2017;Palasz et al., 2020). BDNF levels have been reported to increase linearly to the age of 65, after which their levels markedly decrease (Neshatdoust et al., 2016). Interestingly, this predominantly corresponds to various MRI measures of brain health and cognition, although a high level of inter-study heterogeneity exists (McPhee et al., 2020).

Blood-Brain Barrier
Loss of blood-brain barrier (BBB) integrity/function is considered an early process of neurodegenerative disease. Thus, markers of BBB permeability and/or breakdown are being actively explored as measures of brain health and disease risk. Leakiness of BBB can be determined using MRI (dynamic contrast-enhanced MRI) or PET with various tracers such as gadolinium or 18F-fluoro-2-deoxyglucose (FDG) respectively (Sweeney et al., 2018). Additionally, MRI (T2 * -weighted and susceptibility-weighted imaging MRI) can detect a chronic accumulation of small blood products, 'microbleeds, ' thought to be linked to BBB breakdown. Peripheral markers of BBB function have also been established. CSF soluble platelet-derived FIGURE 1 | Polyphenol categorization -polyphenol describes a compound with a chemical structure containing one or more phenolic rings and can be classified into four main groups: phenolic acids, lignans, stilbenes, and flavonoids. Flavonoids represent the main polyphenolic component of citrus fruits and can be subdivided into six further subgroups. Flavanones are particularly abundant in citrus fruits and include hesperidin as one of the major flavanones. Figure created with BioRender.com. growth factor receptor β (sPDGFRβ), shed from pericytes (which unsheathe the endothelium of brain capillaries) in response to injury, correlates with neurodegenerative disease progression and BBB breakdown (Miners et al., 2019;Nation et al., 2019;Montagne et al., 2020). Similarly, the ratio of CSF albumin to serum albumin, is frequently used as an indicator of BBB breakdown (Musaeus et al., 2020), although conflicting results have been reported (Skillbäck et al., 2017). Less invasive methods including measuring circulating tight junction proteins (occludin, ZO-1 and CLDN5) have also been forwarded as BBB breakdown measures (Jiao et al., 2015;Zhu et al., 2017), whilst serum 14-kDa monomeric form of transthyretin (TTR) and serum concentrations of the astrocytic protein S100B, have been suggested as measures of BBB damage (Marchi et al., 2004).

Metabolic Function
Metabolic homeostasis is particularly important for brain health and function due to the energy-demanding nature of the brain. Brain energy metabolism declines subtly during the aging process and prior to neurological disease diagnosis (Zilberter and Zilberter, 2017), and accumulating evidence demonstrates how this impairment of energy metabolism can exacerbate neurodegenerative disease progression (Cunnane et al., 2016;Camandola and Mattson, 2017). Neuroimaging techniques can effectively indicate metabolic disturbances. For example, FDG-PET is frequently utilized to determine brain glucose metabolism rates across the brain, which subsequently correlate to synaptic activity and disease risk (Shivamurthy et al., 2015). A multitude of other positron-emitting isotope tracers exist, including [1-11 C]-DHA (Yassine et al., 2017), and the incorporation of such tracers can hone in on specific aspects of brain metabolism which may be disease-specific. Furthermore, the ratio of N-acetylaspartate (NAA) and myoinositol (MI), two abundant metabolites in the human brain, has been reported to be a good predictor of mild cognitive impairment in cognitively normal older adults as measured by 1 H MRS in the posterior cingulate cortex (PCC) (Kantarci et al., 2013). Finally, peripheral measures of metabolic function such as insulin resistance and blood glucose, are routinely determined in clinical trials in which cognitive health is a primary output, and have been found to correlate with cognitive decline (Wium-Andersen et al., 2019), and brain atrophy (Erus et al., 2015).

Inflammatory Markers
Although neuroinflammation is an essential part of the brain response to infection or injury (Glass et al., 2010), sustained neuroinflammatory processes may contribute to the progressive neuronal damage observed in aging (Barrientos et al., 2015) and aged-related cognitive disorders (Heppner et al., 2015;Herrero et al., 2015). As resident macrophages, activated microglial cells have the capacity to synthesize a wide range of pro-inflammatory and anti-inflammatory cytokines and molecular mediators which contribute to the systemic inflammatory milieu and to the progression of neurodegenerative disease (Perry and Holmes, 2014). As an integral component of numerous neurological diseases, effective monitoring of neuroinflammation would be highly advantageous. PET may represent a viable option to achieve this (Chandra et al., 2019;Kreisl et al., 2020), along with a range of markers for activated microglia such as Translocator protein (TSPO) radioligands showing promise both in a preclinical setting and in humans. Optimization of TSPO tracers and identification of more specific neuroinflammatory tracers is, however, warranted to improve upon accuracy (Kreisl et al., 2020). Similar to peripheral markers of metabolic function, although not a direct marker of brain inflammation, circulating levels of the proinflammatory (IL-1α, IL-1β, IL-6, TNF-α) and anti-inflammatory cytokines (IL-1ra and IL-10) indicate a chronic inflammatory environment and probably contribute to brain inflammation. Again, these measures frequently correlate with measures of brain health and disease such as cognitive performance (Magalhães et al., 2018;Contreras et al., 2020;Ribeiro-Santos et al., 2020).

Cognitive Testing and Sensory Parameters
As well as biological and neuroimaging techniques, brain health can be determined through cognitive assessment. For this, a wide range of tests can be employed, usually in combination as a battery, to address multiple aspects of cognition such as memory, language, executive functions and attention. Memory, a common target of numerous neurodegenerative disorders, can be separated into specific memory domains, enabling researchers to pinpoint disruptions to specific brain systems (e.g., episodic memory; hippocampaldiencephalic system, working memory; dorsolateral prefrontal cortex, semantic memory; anterior temporal lobe, executive functioning; frontal lobe and beyond) (Kipps and Hodges, 2005). Traditionally, such examination has required extensive clinical testing, performed by trained individuals (Kipps and Hodges, 2005), however, recent advancements in technology, e.g., smartphones and internet-based cognitive assessment tools may enable screening to take place outside of a clinical setting (Coutrot et al., 2019;Hays et al., 2019;Sternin et al., 2019).
Finally, it is worth briefly mentioning sensory parameters, particularly olfaction, which are potentially underrated, and certainly overlooked markers of brain health, despite consistent reports of their predictive capacity in cognitive decline (Brai et al., 2020).

BIOAVAILABILITY OF CITRUS FLAVONOIDS
The extent to which citrus flavanones may exert their biological action is strongly related to their bioavailability which can be affected by many factors including the structure of the compound, the food matrix, or host factors (age, sex, genetic polymorphism, gut microbiota etc.) (Kay et al., 2017;Morand and Tomás-Barberán, 2019). There is an apparent difference between the absorption of glycosides and aglycones with the glycosylation of flavanones increasing their hydrophilicity therefore abolishing passive diffusion and thus lowering their bioavailability (Najmanová et al., 2019). For example, following oral administration of hesperidin, plasma concentrations of flavanone conjugates (e.g., hesperetin-glucuronides and sulfoglucuronides) were detected after 3 h and reached a peak between 5 h and 7 h (Manach et al., 2003), highlighting that the main site for flavanone absorption is the small intestine and the colon. Such results were further confirmed by a more recent study where fasted participants, aged 51-69 years, received either orange juice or a hesperidin supplement (both providing 320 mg hesperidin) versus a control (all matched for sugars and vitamin C content). Total plasma flavanone metabolite concentrations were significantly higher 5 h after the orange juice intervention than after control with hesperidin-glucuronide and naringenin-7-O-glucuronide, largely contributing to the total plasma flavanone concentration. Unexpectedly, no significant concentration of hesperidin metabolites was observed at 5 h following the hesperidin supplement intake, which may highlight the importance of food matrix in the bioavailability of these compounds (Schär et al., 2015). In addition to the main phase II metabolites found in the systemic circulation, the intestinal microflora further degrades aglycones into smaller phenolics. For example, the main degradation product of hesperetin is 3-(3 -hydroxy-4 -methoxyphenyl)propionic acid, although many other compounds have also been reported (see Kay et al., 2017;Najmanová et al., 2019 for extensive reviews of flavanones metabolism).
Flavanone bioavailability can differ across citrus source and/or form. For example, the bioavailability of hesperidin, indirectly determined from excreted hesperetin (main hesperidin metabolite marker) is comparable for both whole orange fruit (1.5%) and orange juice (2.9%), despite the considerably higher flavanone content of the fruit (Aschoff et al., 2016), reflecting potential absorption and metabolic differences. Fresh and packaged orange juice, on the other hand, appear to have relatively similar metabolic kinetics in regards to flavanones, and therefore the higher flavanone content found in packaged juice (machine pressed) translates to a threefold greater flavanone status (Silveira et al., 2014).
In addition, a large inter-individual variability in the bioavailability of citrus flavanones has been reported with high, medium, and low flavanone metabolite excreters identified following intake of citrus flavanones and citrus juices (Nielsen et al., 2006;Brett et al., 2008;Vallejo et al., 2010;Aschoff et al., 2016). One of the key factors in this variability may reside in the host-microbiota diversity necessary for the conversion of the flavanone rutinoside into their aglycone forms (Stevens et al., 2019), although the impact of lifestyle factors such as exercise have also been reported . Despite the demonstrated high inter-individual variability in citrus flavanones metabolism, limited evidence is currently available regarding the role of the gut microbiota composition and no correlation with the effects on brain health biomarkers has been demonstrated.
When considering the bioavailability of citrus flavonoids from a neurological perspective, one must also consider the compounds ability to traverse the BBB. The extent to which specific citrus polyphenols cross the BBB remains to be fully categorized, however, evidence to date in the form of in vitro cellular models (e.g., RBE4, b.END5, and hCMEC/D3) (Youdim et al., 2003;Yang W. et al., 2014), and preclinical approaches (Peng et al., 1998;Tsai and Chen, 2000;Youdim et al., 2004) indicate that citrus flavonoids, namely; hesperetin, naringenin, as well as their relevant metabolites, are able to reach the brain. Whether these models fully translate to humans, particularly when ingested with other complex food sources is yet to be proven, however, pursuing this question will undoubtedly enhance our understanding of citrus polyphenols in the context of brain health and disease, elucidating the extent to which their bioactivity stems from direct interaction with the brain parenchyma.

MICROBIOTA: GUT: BRAIN AXIS
A healthy gastrointestinal tract is a nutrient-rich environment hosting approximately 100 trillion microbes (Lin and Zhang, 2017), that constitute the gut microbiota. Gut microbial composition is increasingly recognized as a central factor in health and disease, protecting the intestinal gut barrier and preventing the establishment of pathogenic microorganisms. The gut microbiota also provides a large repertoire of genes, antigens and metabolites that can regulate immune and metabolic functions. The gut-brain axis describes a bidirectional system that encompasses both neuro-immune (Teratani et al., 2020) and neuro-endocrine communication as well as a direct neuronal connection (vagus nerve) (Tan et al., 2020), with each mode of transmission receiving microbial modulation (Lv et al., 2019;Wang et al., 2019). The gut-brain axis has rapidly emerged in recent decades, with an influx of literature implicating gut health and microbial dysbiosis with specific neurological diseases/conditions, although the multifactorial nature and complexity of these diseases/conditions has left mechanistic validation somewhat lagging (Sherwin et al., 2018). Indeed, further elucidation of the mechanistic as well as the determination of overall therapeutic validity are still required (Peterson, 2020). An area in which mechanistic evidence is beginning to gain traction is the synthesis and/or regulation of metabolites by the gut-microbiota, with many derivative metabolites established as neurochemicals or possessing neuromodulatory properties. Short-chain fatty acids (SCFA) are a relatively well-characterized example of this (Dalile et al., 2019) and are produced via fermentation of complex carbohydrates (dietary fibers). SCFA levels are diminished in response to antibiotic treatment (Høverstad et al., 1986) and similarly low levels tend to associate with conditions such as depression, Alzheimer's (AD) and Parkinson's disease (Dalile et al., 2019). For example, the SCFA propionate was recently reported to inhibit pathways associated with nonspecific microbial infections via a CD14-dependent mechanism, to suppress the expression of LRP-1 and to protect the BBB from oxidative stress via NRF2 (NFE2L2) signaling (Hoyles et al., 2018). The influence of metabolites such as SCFA produced by our 'gut-biofactory' has been extensively reviewed by Skonieczna-Zydecka et al. (2020).
The considerable inter-individual variability of the gut microbiota combined with its apparent modifiable/dynamic nature has led to it being identified as a therapeutic target. The fact that beneficial shifts could be achieved through noninvasive, relatively safe interventions, e.g., prebiotics, make it an attractive target. The gut microbiota is both modulated by, and modulates, polyphenolic compounds (Koudoufio et al., 2020). Citrus polyphenols appear to be no exception, with neohesperidin recently shown to reverse high-fat-diet-induced intestinal microbiota dysbiosis by increasing general microbial diversity as well as specific strains including Bacteroidetes and Firmicutes (Lu et al., 2020). A similar experiment in which hesperidin was administered exerted similar prebiotic effects with treated mice displaying improved metabolic maker profile (Guirro et al., 2020). These beneficial hesperidin mediated effects are in agreement with an earlier experiment, which interestingly reported upon the concomitant immunomodulatory actions (Estruel-Amades et al., 2019). In healthy volunteers, continuous consumption of commercial pasteurized orange juice for 2 months improved blood biochemical parameters, such as lowdensity lipoprotein-cholesterol, glucose, and insulin sensitivity and positively modulated the composition and metabolic activity of the microbiota, increasing the population of fecal Bifidobacterium spp. and Lactobacillus spp. (Lima et al., 2019). In addition, daily consumption of 500 mL of two Brazilian orange juices (e.g., Cara Cara and Bahia juices) for 7 days increased the abundance of Mogibacteriaceae, Tissierellaceae, Veillonellaceae, Odoribacteraceae, and Ruminococcaceae families in healthy volunteers (Brasili et al., 2019). Furthermore, daily consumption of 300 ml of orange juice for 60 days affected the levels of Lactobacillus spp., Akkermansia spp., and Ruminococcus spp. and improved the glycemia and lipid profiles in 20-35 years old healthy female volunteers (Fidélix et al., 2020). These recent reports emphasize the prebiotic potential of citrus polyphenols, particularly in metabolic disease, with no specific information related to brain functions (e.g., gut-brain axis). With this being said, it must be mentioned that current evidence has been predominantly provided through preclinical and in particular, rodent experiments. Therefore, current perspectives derived from these studies should be cautiously interpreted until validated through robust clinical trials since fundamental human to mouse differences, such as microbial composition and gut physiology, may render direct translation inappropriate.

METHOD
A non-systematic search was performed in PubMed using the following keywords to retrieve preclinical and clinical data: [(Dementia OR depression OR Alzheimer OR Parkinson OR neurodegeneration OR neuroinflammation) AND (citrus polyphenol OR flavanone OR hesper * OR naring * ) AND (cognition OR brain)]. Articles in English or with English abstracts were retrieved. All evidence was read and chosen based on the Authors' evaluation of relevance. The search was concluded in August 2020 and updated in November 2020, and included all articles present in PubMed, without temporal limits.

PRECLINICAL EVIDENCE FOR CITRUS FLAVONOIDS AND BRAIN HEALTH/FUNCTION
Neurodegeneration describes a process of progressive cell dysfunction and eventual neuronal cell death. In diseases such as late-onset Alzheimer's Disease (LOAD) this process can span decades. Although usually categorized based upon distinct pathologies, the overall process of neurodegeneration is one that is multi-factorial, with many neurodegenerative diseases sharing common pathogenic mechanisms underpinning disease progression (Jellinger, 2010), potentially explaining the lack of efficacy of single target drugs. These common mechanisms associated with neurodegeneration include: neuroinflammatory/neuroimmune responses, oxidative stress, impaired bioenergetics/mitochondrial dysfunctions, dysfunction of neurotrophins and abnormal protein dynamics. The mechanistic understanding of citrus polyphenols in brain health stems largely from preclinical and in vitro studies. The focus of these preclinical studies often centers on mitigating specific disease pathologies. Indeed, a wealth of preclinical evidence has highlighted the multifaceted nature of citrus polyphenols neurologically. Interestingly, as highlighted in Table 1, the major citrus polyphenols share common mechanistic actions, overlapping considerably with the aforementioned deficits associated with neurodegeneration. Below we discuss these mechanisms in more details focusing on the key citrus flavonoids and in particular on the flavanones hesperidin and naringin along with their aglycone forms (i.e., hesperetin and naringenin) and the flavonol, kaempferol.

Anti-oxidative
Oxidative stress is a well-established contributing factor in neurological disorders (Michalska and León, 2020) with high metabolic activity combined with a lack of antioxidant defense capability leaving the brain particularly susceptible. Although citrus polyphenols reportedly demonstrate free radical scavenger capacity in vitro (Di Meo et al., 2013), the current literature search unanimously highlighted their ability to stimulate the endogenous antioxidant defense machinery, with superoxide dismutase (SOD), catalase (CAT), glutathione (GSH), glutathione S-transferases (GST), glutathione reductase (GR) and glutathione peroxidase (GPx) activation common across all the citrus polyphenols reviewed (Table 1). Effective dosage ranged from as little as 5 mg/kg up to 300 mg/kg, however, dose dependant effects were established in a number of studies, with greater doses usually eliciting greatest effects (Table 1). Activation coincided with subsequent reduction of reactive oxygen species (ROS) such as hydrogen peroxide (H 2 O 2 ), nitric oxide (NO) and other oxidative markers namely malondialdehyde (MDA) and thiobarbituric acid reactive substances (TBARS). Upregulation of the transcription factor NRF2 features in a number of these studies Sugumar et al., 2019;Welbat et al., 2020), and is likely central to this polyphenol mediated anti-oxidative system through the activation of the antioxidant response element (Vauzour, 2012). As alluded to, activation of the antioxidant defense machinery was generally consistent across the flavonoids with only a few discrepancies found, mainly confined to naringenin and hesperetin, where occasionally no effect was detected (Chtourou et al., 2014;Kheradmand et al., 2018;Mani et al., 2018). A few studies reported the opposite effect for naringenin with SOD and CAT reduced (Chtourou et al., 2014;Haider et al., 2020). This reduction associated with naringenin, which is believed to be able to cross the BBB (Youdim et al., 2003), may relate back to its free radical scavenging capabilities, or arise as a result of the differing disease models used. In addition to the more common antioxidant machinery influenced, hesperidin and hesperetin upregulated Haem-oxygenase (HO-1) and downregulated the superoxide radical generating enzyme Xanthine Oxidase (XO) respectively (Ashafaq et al., 2014).

Anti-inflammatory
The immunomodulatory capabilities of citrus polyphenols within the brain are similarly evident, and are in some cases coupled to anti-oxidative mechanisms (e.g., HMGB1/RAGE, Yu et al., 2015). Again, the molecular targets with which the reviewed citrus polyphenols interact appear to be consistent, with reduction of pro-inflammatory cytokines IL-1β, IL-2, IL-6, IFN-γ, and TNF-α, particularly prolific in the literature ( Table 1). This is likely mediated through the mitigation of hyperactive immune cells as is suggested by the reduction of GFAP, and NF-κB which governs chemokine and inflammatory mediator transcription (Fu et al., 2019;Zhou et al., 2020). In contrast to evidence on anti-oxidative effects, there was little indication to suggest lack of effect across any of the reviewed flavonoids, supporting the notion that neuro-inflammatory modulation is inherent across all citrus flavonoids. As with anti-oxidative effects, effective dosage had a considerable range of 20-300 mg/kg, with efficacy remaining even at lower doses 25-50 mg/kg. This latter dose equates to 145-290 mg human equivalent dose in a 70 Kg adult, if the animal is a mouse (Nair and Jacob, 2016), which is achievable through consumption of 0.5-0.75 L of orange juice (De Rycker et al., 2020).

Proteinopathy
Only a limited amount of preclinical evidence exists to support the beneficial effects of citrus on proteinopathies. This primarily relates to Alzheimer's disease, with reports revealing improvements in Tau phosphorylation (Yang W. et al., 2014;Okuyama et al., 2018;Zhou et al., 2020), and Aβ deposition (Yang W. et al., 2014;Li et al., 2015;Zhou et al., 2020), however no effect has also been reported (Wang et al., 2014). The mechanistic basis for these changes remains relatively weak, although the glycogen synthase kinase 3 beta (GSK3β) may be a plausible candidate (Yang W. et al., 2014;Zhou et al., 2020). A reduction of α-synuclein, the protein attributed to parkinopathy has also been described (Mani et al., 2018), potentially indicating that the effect extends further than Alzheimer's disease-associated proteins. Higher dosage appears crucial to establish these effects, with <100 mg/kg failing to have any impact. Finally, a report in which hesperetin was administered demonstrated restoration of brain proteolytic enzyme levels (Shagirtha et al., 2017), potentially linking citrus flavonoids to lysosomal degradation processes (Aufschnaiter et al., 2017) (Table 1).

Cognitive Outcomes
Unsurprisingly, given the established effects of flavonoids on key markers of brain health and functionality at molecular and metabolic levels, in studies where cognition was determined, flavonoid supplementation led to improvements in cognitive performance. Cognitive and behavioral assessment further supports/demonstrates the neuroprotective properties associated with citrus flavonoids ( Table 2). Anxiolytic and antidepressant actions were particularly prominent across the literature and appeared to be consistent across several disease models, suggesting modulation of a fundamental anxiety and depression related process. Similarly, citrus polyphenol supplementation improved deficits in learning and memory, specifically spatial and recognition memory, which may indicate protection of medial temporal lobe, particularly vulnerable to Alzheimer's like neurodegenerative diseases. Improvements in motor functions and locomotion were also apparent, and impressively remained even in most severe models such as middle cerebral artery occlusion (Raza et al., 2013), and traumatic brain injury (Chitturi et al., 2019).

HUMAN EVIDENCE FOR CITRUS POLYPHENOLS AND BRAIN HEALTH
Whilst a wealth of pre-clinical (reviewed above) and in vitro data exploring neurocognitive outcomes in relation to citrus flavonoids is currently available, the same cannot be said for human trials which remain considerably limited in number. In our literature search, we identified 10 human studies (five observational; five interventions) assessing the effects of citrus flavonoids on brain health and cognition in healthy adults, or in addition to other co-morbidities including depression, dementia, schizophrenia, and stroke.

Studies on Healthy Adults
The benefits of citrus fruit consumption in the context of healthy aging were highlighted by a cross-sectional study involving 2031 elderly (aged 70-74 years) Norwegian individuals which explored the impact of different plant foods on cognitive performance (Nurk et al., 2010). Study participants underwent extensive cognitive testing in addition to completing comprehensive food frequency questionnaires. After adjustment for multiple  Table 3).
Following on from these observations, Kean et al. (2015), examined the chronic consumption of flavanone-rich orange juice in relation to cognition in 37 healthy volunteers (60-81 years). The study was a double-blind RCT with a crossover design and involved consumption of a 500 mL high flavanone (daily serving of 305 mg -549 mg hesperidin/L and 60 mg narirutin/L) orange juice daily for 8 weeks. In support of the cross-sectional study mentioned above, global cognition increased in response to chronic consumption of the flavanonerich orange juice relative to control. In addition, high flavanone intake significantly improved recall and nominally increased executive function, although post hoc analysis found this to be non-significant (p = 0.06). These effects were independent of mood and blood pressure which both remained unchanged ( Table 3).
Acute neurological responses to citrus flavonoids have similarly been investigated (Alharbi et al., 2016;Lamport et al., 2016). Firstly, in a randomized, double-blind, placebo-controlled, crossover trial, Alharbi et al. (2016) explored the cognitive benefits associated with flavonoid-rich orange juice (272 mg -220.46 mg hesperidin, 34.54 mg narirutin, 17.14 mg other flavonoids) at 2 and 6 h post-consumption in 24 healthy middle-aged adults (30-65 years). From the cognitive battery performed, flavonoid rich orange juice consumption led to higher performance in Simple Finger Tapping (measure of psychomotor speed) and Continuous Performance Task (measure of attention and more broadly executive function) at 2 and 6 h respectively. A non-significant trend for higher global cognitive performance (all tests combined) was also observed, as well as an increase in subjective alertness. Interestingly, significant improvements observed in cognition and subjective alertness at 6 h coincide with an anticipated peak in flavanone metabolites at 5-7 h (Manach et al., 2003), although this was not physically measured.
As with the previous study, Lamport et al. (2016) assessed acute neurological response, this time to a commercially available high flavanone beverage (HF, 70.5 mg -42.15 mg hesperidin, 17.25 mg naringin, 6.75 mg narirutin, 4.3 mg caffeic acid) and utilizing an additional measure of cerebral blood flow (CBF). The study was single blind, randomized, cross over, by design involving healthy adult volunteers (aged 18-30 years). Participants underwent either cognitive testing 2 h post consumption (n = 28) or completed an fMRI assessment of CBF 2 and 5 h post consumption (n = 16). High flavanone beverage intake significantly increased cerebral perfusion in the inferior frontal and middle right frontal gyrus in the right hemisphere at 2 h. Similarly, at 2 h, improvement in digit symbol substitution test (a measure of executive function) was seen, correlating with the increased regional perfusion of the inferior frontal gyrus, known to be involved in executive function (Aron et al., 2004). Despite the extensive cognitive battery, no additional effects were found. However, as addressed by the authors, this study had a number of limitations. In particular, the fact MRI and cognitive test were performed on separate individuals, and that cognitive tests were only performed at 2 h limits comparative potential. Additionally, as stated by the authors, the fact that the participants were generally young and highly educated may have limited the response given they were likely to be optimally functioning (Table 3).

Studies on Depression
Depression is a complex mood disorder which can often be challenging to treat effectively. Preclinical studies have reported anti-depressant effects of flavonoids, usually attributable to their antioxidant and anti-inflammatory characteristics, and inhibition of monoamine oxidases (Hritcu et al., 2017). An important serological marker identified in depression is BDNF (involved in processes within the central nervous system) which has been observed to be significantly lower in patients with major depressive disorder compared to non-depressed control groups, and subsequently recovered antidepressant users (Chen et al., 2001). Interestingly, BDNF is often increased in response to flavonoid consumption (Neshatdoust et al., 2016).
In a prospective cohort study following 82,643 women with no previous diagnosis of depression from the Nurses' Health Study (NHS; aged 53-80 years) and Nurses' Health Study II (NHSII; aged 36-55 years) throughout a 10-year period, an inverse association between incident depression and citrus intake, with greatest flavanone intake (>64.2 mg/day) resulting in a significant 10% reduction in incident depression risk was observed. Furthermore, citrus fruit and juice intake of >2 servings/day had HR of 0.82 (Chang et al., 2016).
These anti-depressant effects of citrus flavonoids were recently put to the test by Park et al. (2020), in a single blind, randomized control study examining the effects of daily 380 ml flavonoid rich (FR; 600 ± 5.4 mg flavonoid content) orange juice consumption on depressive symptoms and gut microbiota for an 8-week period in young adults (aged 20-30 years). The results in relation to depression were by no means clear cut, with no apparent significant differences between high and low flavonoid groups after 8 weeks. However, compared with baseline the results suggested potential improvement in BDNF, and to some extent to the Center for Epidemiological Studies Depression Scale (CES-D), a psychiatric screening tool used to detect pre-existing mental disorders, although both treatments appeared to improve upon baseline CES-D scores ( Table 3).

Studies on Dementia
Despite the preclinical evidence backing the beneficial effect of citrus flavonoids in models of Alzheimer's and Parkinson's diseases, etc. there remains very limited evaluation at the human level.
In a retrospective cohort study , examined the association between daily citrus intake and dementia incidence in 13,373 participants (age ≥ 65 years). FFQs in combination with the Japanese Long-term Care Insurance database were used over a 5.7 years follow up period. Overall, an inverse dose-response relationship was established between weekly citrus fruit intake and incident dementia, with hazard ratios for citrus fruit consumption 3-4 times/week and every day 0.92 (95% CI 0.80, 1.07) and 0.77 (95% CI 0.73, 1.01) respectively. As dementia represents a wide range of neurodegenerative diseases, lack of further classification limits the extent to which we can attribute these effects to specific neurodegenerative diseases and should be considered in future studies (Table 3).

Studies on Schizophrenia
Schizophrenia is a complex psychotic condition affecting cognition, the cognitive hallmark of schizophrenia being poor learning and retention of verbal information (Bowie and Harvey, 2006). In an open label pilot study, 20 outpatients diagnosed with schizophrenia receiving second generation antipsychotic medication consumed a flavanone-rich bergamot polyphenolic fraction (BPF) daily at a dose of 1000 mg/day for 8 weeks. Daily BPF consumption significantly improved Wisconsin Card Sorting Test "perseverative errors" and semantic fluency test. With a trend for other cognitive variables evident (Bruno et al., 2017) (Table 3).

Studies on Stroke and Stroke Risk Factors
The protective effects of citrus flavonoids in cerebrovascular disease are well documented (Testai et al., 2017;Mahmoud et al., 2019) and appears to extend to stroke incidence. A prospective cohort study, assessing the association of dietary flavonoid intake toward stroke risk, followed 69,622 women (30-55 years) from the NHS Study throughout a 14-year period (Cassidy et al., 2012). Although flavonoid intake was not inversely associated with stroke risk, women with greatest intake of flavanones > 62.95 mg/day had a relative risk for experiencing an ischemic stroke of 0.81 compared to lowest intake < 13.72 mg/day. Additionally, an inverse trend was established between citrus fruit/juice intake and risk of ischemic stroke. These apparently protective effects did not influence haemorrhagic strokes (Scheffers et al., 2018;D'Elia et al., 2020;Zurbau et al., 2020).
A recent prospective cohort study (Goetz et al., 2016), utilizing 'The Reasons for Geographic and Racial Differences in Stroke' (REGARDS) database, assessed whether any association exists between flavonoid intake and incident ischemic stroke in a biracial cohort. 20,024 participants (>45 years) of whom flavonoid consumption was determined from FFQ, were followed for 6.5 years. After multivariable adjustment, the highest flavanone intake > 48 mg/day was inversely associated with incidences of ischemic stroke compared to lowest flavanone intake < 3.9 mg/day. With citrus fruit/juice consumption displaying similarly reduced risk (HR of 0.69). In agreement with the above-mentioned study of total flavonoids and other flavonoid subclasses did not show any statistically significant association with incident ischemic stroke, clearly showcasing the importance of the flavanone subclass ( Table 3).
Of particular relevance to the mitigation of stroke risk, and also neurodegenerative conditions, it is worth briefly mentioning the influence of citrus flavonoid intake on blood pressure, and vascular functions. RCTs in which flavonoids were administered in the form of orange juice have established significant reductions in blood pressure (Morand et al., 2011;Valls et al., 2020), pulse pressure (Valls et al., 2020), while also improving flow-mediated dilation (Li et al., 2020).

FUTURE PERSPECTIVES AND CONCLUSION
The preclinical literature search distinguished fundamental mechanisms central to citrus flavonoids, with protective effects linked with anti-oxidative and anti-inflammatory action particularly well established. Yet, there are a number of areas requiring further investigation. First, the overwhelming majority of studies to date involve relatively young male animals making it difficult to establish whether or not the observable effects are sex or age specific. Nor do we have an understanding of modulation by either factor. Second, the models employed tend to lean toward the severe side of the disease spectrum, thus translation to a healthy aging context, or milder conditions, remains to be fully determined, although some initial work provides supportive results (Liaquat et al., 2018). Third, despite recent attention on the relationship between gut microbiota and polyphenols, and the apparent ability of citrus flavonoids to mitigate LPS associated neuroinflammation/NDG, to our knowledge there are no studies which have explored the role of the microbiota in combination with citrus flavonoids in the context of brain health and disease. Finally, as can be seen in Table 1, the impact of citrus flavonoids upon BBB integrity/function appears to have emerged over the last couple of years (Cheng et al., 2018;Li et al., 2019;Yang et al., 2019;Lee et al., 2020), particularly for kaempferol, and therefore should be examined in further studies, especially as BBB integrity has similarly been associated with gut microbiota (Braniste et al., 2014).
From a human perspective, there is an obvious lack of human clinical studies which needs to be addressed if a robust assessment of therapeutic potential is to be made. Similarly, very few human studies have followed up on the mechanistic insights established in the preclinical setting. Future human studies should take note of the limitations arising from other human studies in which nutraceuticals were assessed in the context of brain health and disease, for example ensuring optimal participant targeting (in demographics where significant change is most likely to occur), dosage, timing, and duration of treatment for measurable effects to be established.
The complex mixtures of polyphenols present in citrus fruits and juices and their bioactive nuances likely convey greater benefit than one purified compound, accumulatively acting upon multiple targets, and producing synergistic effects. Given the multifactorial nature of neurodegenerative diseases, one would speculate that this complex form would therefore offer greater efficacy, but this is yet to be fully determined.
In conclusion, although significant work remains to fully establish the benefits of citrus polyphenols in brain health and disease, the accumulating in vitro and preclinical data combined with the support of steadily emerging human studies indicates future potential.

AUTHOR CONTRIBUTIONS
MP, MMM, and DV wrote the manuscript. MP, MMM, EC, MM, and DV contributed to the literature search and edited the manuscript. All the authors contributed to the article and approved the submitted version.