The classical pathway for ^•NO synthesis involves a family of enzymes – nitric oxide synthase (NOS) – that catalyzes the oxidation of L-arginine to L-citrulline and ^•NO, provided that oxygen (O₂) and several other cofactors are available [nicotinamide adenine dinucleotide phosphate (NADPH), flavin mononucleotide (FMN), flavin adenine dinucleotide (FAD), heme and tetrahydrobiopterin (BH₄)]. For this to occur, the enzyme must be in a homodimeric form that results from the assembly of two monomers through the oxygenase domains and allows the electrons released by the NADPH in the reductase domain to be transferred through the FAD and FMN to the heme group of the opposite subunit. At this point, in the presence of the substrate L-arginine and the cofactor BH₄, the electrons enable the reduction of O₂ and the formation of ^•NO and L-citrulline. Under conditions of disrupted dimerization, ensured by different factors (e.g., BH₄ bioavailability), the enzyme catalyzes the uncoupled oxidation of NADPH with the consequent production of superoxide anion (O₂^–•) instead of ^•NO (Knowles and Moncada, 1994; Stuehr, 1999). There are three major members of the NOS family which may diverge in terms of the cellular/subcellular localization, regulation of their enzymatic activity, and physiological function: type I neuronal NOS (nNOS), type II inducible NOS (iNOS), and type III endothelial NOS (eNOS) (Stuehr, 1999). The nNOS and eNOS are constitutively expressed enzymes that rely on Ca²⁺-calmodulin binding for activation. The nNOS and eNOS activity is further regulated both at the transcriptional and post-translational levels and via protein-protein interactions (Forstermann and Sessa, 2012). While not exclusively, the nNOS is mainly expressed in neurons where it is intimately associated with glutamatergic neurotransmission. The dominant splice variant of this isoform (nNOSα) possesses an N-terminal PDZ motif that allows the enzyme to bind other PDZ-containing proteins, such as the synaptic density scaffold protein PSD-95. This allows the enzyme to anchor itself to the synaptic membrane by forming a supramolecular complex with the N-methyl-D-aspartate receptors (NMDAr), whose activation upon glutamate binding results in Ca²⁺ influx, and ultimately, ^•NO production. The eNOS isoform is mainly expressed at the endothelium and is critically involved in vascular homeostasis. In the endothelial cells, the eNOS is predominantly localized within the caveolae, forming a complex with caveolin-1 that inhibits its activity. The stretching of the vascular wall, induced by shear stress, results in the dissociation of this complex and allows the enzyme to be activated, either by Ca²⁺-calmodulin binding and/or by PI3K/Akt-mediated phosphorylation of specific serine residues (e.g., 1,177) (Forstermann and Sessa, 2012). Unlike the other two isoforms, iNOS does not rely on Ca²⁺ increases for activation but on the de novo synthesis, which occurs predominantly in glial cells following an immunological or inflammatory stimulation. Because iNOS has much lower Ca²⁺ requirements (calmodulin binds with very high affinity to the enzyme even at basal Ca²⁺ levels), it produces ^•NO for as long as the enzyme remains from being degraded (Knott and Bossy-Wetzel, 2009).

In recent years, studies have supported ^•NO production independent of NOS activity, through the stepwise reduction of nitrate (NO₃^–) and nitrite (NO₂^–) via the so-called nitrate-nitrite-nitric oxide pathway. Viewed as stable end products of ^•NO metabolism, both NO₃^– and NO₂^– are now recognized to be able to be recycled back into ^•NO, thereby acting as important ^•NO reservoirs in vivo. NO₃^– and NO₂^– can be consumed in the regular vegetable components of a diet, fueling the nitrate-nitrite-nitric oxide pathway (Rocha et al., 2011; Lundberg et al., 2018). NO₃^– can be reduced to NO₂^– by the commensal bacteria in the gastrointestinal tract and/or by the mammalian enzymes that can acquire a nitrate reductase activity under acidic and hypoxic environments. In turn, the reduction of NO₂^– to ^•NO can be achieved non-enzymatically via a redox interaction with one-electron reductants (e.g., ascorbate and polyphenols) or can be catalyzed by different enzymes (e.g., hemoglobin, xanthine oxidoreductase, and cytochrome P450 reductase). All these reactions are favored by low O₂ and decreased pH, thereby ensuring the generation of ^•NO under conditions of limited synthesis by the canonical NOS-mediated pathways which require O₂ as a substrate (Lundberg et al., 2008). It is also worth mentioning that S-nitrosothiols might serve as downstream ^•NO-carrying signaling molecules regulating protein expression/function (Chen et al., 2008).

The transduction of ^•NO signaling may involve several reactions that reflect, among other factors, the high diffusion of ^•NO, the relative spatial and temporal abundance of the targets, and the relative rate constants with the potential targets. Most of the physiological actions of ^•NO are promoted by the chemical modification of relevant proteins either via nitrosylation or nitrosation [reviewed in Picon-Pages et al. (2019)]. Nitrosylation refers to the reversible binding of ^•NO to inorganic protein moieties (e.g., iron in heme groups), while nitrosation involves the modification of organic moieties (e.g., thiol groups in cysteine residues), not directly, but intermediated by the species produced upon ^•NO autoxidation, namely N₂O₃. In addition, ^•NO can react with superoxide anion (O₂^–•), yielding peroxynitrite (ONOO^–), a potent oxidant and nitrating species that conveys the main deleterious actions associated with the ^•NO signaling (e.g., oxidation and/or nitration of proteins, lipids and nucleic acids) (Radi, 2018).

The best characterized molecular target for the physiological action of ^•NO is the soluble guanylate cyclase (sGC), a hemeprotein that is frequently and controversially tagged as the classical “^•NO receptor.” The activation of the sGC by ^•NO involves the nitrosylation of heme moiety of the enzyme that induces a conformational change, enabling it to catalyze the conversion of guanosine triphosphate (GTP) to the second messenger cyclic guanosine monophosphate (cGMP) (Martin et al., 2005). Nitric oxide may additionally regulate the catalytic activity of sGC by promoting its inhibition via nitrosation of critical cysteine residues (Beuve, 2017).

The conventional pathway for ^•NO- mediated NVC involves the activation of the glutamate-NMDAr-nNOS pathway in neurons. The binding of glutamate to the NMDAr stimulates the influx of [Ca²⁺] through the channel that, upon binding calmodulin, promotes the activation of nNOS and the synthesis of ^•NO. Being hydrophobic and highly diffusible, the ^•NO produced in neurons can diffuse intercellularly and reach the smooth muscle cells (SMC) of adjacent arterioles, there inducing the activation of sGC and promoting the formation of cGMP. The subsequent activation of the cGMP-dependent protein kinase (PKG) leads to a decrease [Ca²⁺] that results in the dephosphorylation of the myosin light chain and consequent SMC relaxation [reviewed by Iadecola (1993) and Lourenço et al. (2017a)]. Additionally, ^•NO may promote vasodilation via the stimulation of the sarco/endoplasmic reticulum calcium ATPase (SERCA), via activation of the Ca²⁺-dependent K⁺ channels, or via modulation of the synthesis of other vasoactive molecules [reviewed by Lourenço et al. (2017a)]. Specifically, the ability of ^•NO to regulate the activity of critical heme-containing enzymes involved in the metabolism of arachidonic acid to vasoactive compounds suggests the complementary role of ^•NO as a modulator of NVC via the modulation of the signaling pathways linked to mGLuR activation at the astrocytes. ^•NO has been demonstrated to play a permissive role in PGE₂-dependent vasodilation by regulating cyclooxygenase activity (Fujimoto et al., 2004) and eliciting ATP release from astrocytes (Bal-Price et al., 2002).

The notion of ^•NO as a key intermediate in NVC was initially grounded by a large set of studies describing the blunting of NVC responses by the pharmacological NOS inhibition under different experimental paradigms [reviewed (Lourenço et al., 2017a)]. A recent meta-analysis, covering studies on the modulation of different signaling pathways in NVC, found that a specific nNOS inhibition produced a larger blocking effect than any other individual target (e.g., prostanoids, purines, and K⁺). In particular, the nNOS inhibition promoted an average reduction of 2/3 in the NVC response (Hosford and Gourine, 2019). It is recognized that the dominance of the glutamate-NMDAr-NOS pathway in NVC likely reflects the specificities of the neuronal networks, particularly concerning the heterogenic pattern of nNOS expression/activity in the brain. Although nNOS is ubiquitously expressed in different brain areas, the pattern of nNOS immunoreactivity in the rodent telencephalon has been pointed to a predominant expression in the cerebellum, olfactory bulb, and hippocampus and scarcely in the cerebral cortex (Bredt et al., 1990; Lourenço et al., 2014a). Coherently, there is a prevalent consensus for the role of ^•NO as the direct mediator of the neuron-to-vessels signaling in the hippocampus and cerebellum. In the hippocampus of anesthetized rats, it was demonstrated that the ^•NO production and hemodynamic changes evoked by the glutamatergic activation in dentate gyrus (DG) are temporally correlated and both dependent on the glutamate-NMDAr-nNOS pathway (Lourenço et al., 2014b). The blockage of either the NMDAr or nNOS also showed to blunt the ^•NO production and vessels dilation to mossy fiber stimulation in the cerebellar slices (Mapelli et al., 2017).

In the cerebral cortex, ^•NO has been suggested to act as a modulator rather than a direct mediator of the NVC responses, but this view has been challenged in recent years. Emergent evidence from ex vivo approaches indicates that the regulation of vasodilation may diverge along the cerebrovascular tree: at the capillary level, vasodilation seems to be mainly controlled by pericytes via an ATP-dependent astrocytic pathway while at the arteriolar level it involves neuronal ^•NO-NMDAr signaling (Mishra et al., 2016).

Recent data support that the optogenetic stimulation of nNOS positive interneurons can promote central blood flow (CBF) changes in the somatosensory cortex comparable to those evoked by whiskers stimulation on awake and behaving rodents (Krawchuk et al., 2020; Lee et al., 2020). The implication of the GABAergic interneurons in NVC has been previously demonstrated, both in the cerebellum and somatosensory cortex (Cauli et al., 2004; Rancillac et al., 2006). Also, in the hippocampus, parvalbumin GABAergic interneurons are suggested to drive, via ^•NO signaling, the NVC response to hippocampus-engaged exploration activities with impact in the neurogenesis in the dentate gyrus (Shen et al., 2019). The involvement of GABAergic interneurons in neurovascular regulation is not unexpected as some of them have extended projections in close contact with arterial vessels and secrete diverse molecules with vasoactive properties which are able to modulate the vascular tone (e.g., ^•NO, vasopressin, and NPY) (Hamel, 2006). A novel and striking hypothesis suggest that nNOS-expressing neurons can control vasodilation independent of neural activities. The optogenetic activation of NOS-positive interneurons regulates CBF without detectable changes in the activity of other neurons (Echagarruga et al., 2020; Lee et al., 2020). The activation of GABAergic interneurons has further been shown to promote vasodilation while decreasing neuronal activity; this occurring independently of ionotropic glutamatergic or GABAergic synaptic transmission (Scott and Murphy, 2012; Anenberg et al., 2015). The hypothesis stating that evoked CBF is dynamically regulated by different subsets of neurons, some independently of neuronal activity, calls into question the linearity of the correlation between the net ongoing neuronal activity and CBF changes and raises concerns regarding the interpretation of functional MRI (fMRI) data.

As for the systemic vascular network, endothelial-derived ^•NO has also been implicated in the regulation of CBF. Endothelial cells are able to respond to diverse chemical and physical stimuli by producing, via Ca²⁺-dependent signaling pathways, a myriad of vasoactive compounds (e.g., ^•NO), thereby modulating the vascular tone. Additionally, Ca²⁺ may directly induce the hyperpolarization of the endothelial membrane and adjacent SMC through the activation of Ca²⁺-dependent K⁺ channels (Chen et al., 2014; Guerra et al., 2018). Despite this, the critical requirement of endothelium for the development of a full neurovascular response to neuronal activity only recently started to be valued. Specifically, endothelial-mediated signaling stands to be essential for the retrograde propagation of NVC-associated vasodilation. The discrete ablation of the endothelium was demonstrated to halt the retrograde dilation of pial arteries in response to hindpaw stimulation (Chen et al., 2014). Additionally, in the somatosensory cortex, NVC was shown to be regulated via eNOS upon the activation of the purinergic receptors at the endothelium in a mechanism involving a glio-endothelial coupling (Toth et al., 2015). Recent data further pointed to the ability of endothelial cells to directly sense neuronal activity via the NMDAr expressed in the basolateral endothelial membranes, thereby eliciting vasodilation via eNOS activation (Stobart et al., 2013; Hogan-Cann et al., 2019; Lu et al., 2019). While the precise mechanisms by which the eNOS-derived ^•NO shape NVC response is still to be defined, eNOS activation is suggested to contribute to the local but not to the conducted vasodilation, the latter being associated with K⁺-mediated hyperpolarization (Lu et al., 2019). Yet, it is proposed that ^•NO-dependent vasodilation may be also involved in a slower and shorter-range retrograde propagation cooperating with the faster and long-range propagation mediated by endothelial hyperpolarization (Chen et al., 2014; Tran et al., 2018). Of note, ^•NO can modulate the activity of connexins at the gap junctions to favor the propagation of the hyperpolarizing current upstream to the feeding vessels (Kovacs-Oller et al., 2020). Additionally, vascular-derived ^•NO has been pointed to facilitate Ca²⁺ astrocytic signal and was forwarded as an explanation for the late endfoot Ca²⁺ signaling (Tran et al., 2018).

Despite the extensive accumulated evidence for the involvement of ^•NO in the NVC in animal models, these studies have only been applied to humans recently. By addressing the hemodynamic response to visual stimulation, Hoiland and coworkers provided the first demonstration for the involvement of ^•NO in the NVC in humans via modulation by a systemic intravenous infusion of the non-selective competitive NOS inhibitor L-NMMA (Hoiland et al., 2020). The authors proposed a two-step signaling mechanism for the NVC in humans translated in a biphasic response with the first component being attributed to the NOS activation elicited by glutamatergic activation. They hypothesized that ^•NO may be further involved in the second component of the hemodynamic response via erythrocyte-mediated signaling (either by releasing ^•NO from nitrosated hemoglobin or by mediating NO₂^– reduction) (Hoiland et al., 2020).

The mechanisms underpinning the NVC dysfunction in AD and other pathologies are expectedly complex and likely enroll several intervenients through a myriad of pathways, that may reflect both the specificities of neuronal networks (as the NVC itself) and that of the neurodegenerative pathways. Yet, oxidative stress (nowadays conceptually denoted by Sies and Jones as oxidative distress) is recognized as an important and ubiquitous contributor to the dysfunctional cascades that culminate in the NVC deregulation in several neurodegenerative conditions (Hamel et al., 2008; Carvalho and Moreira, 2018). Oxidative distress is generated when the production of oxidants [traditionally referred to as reactive oxygen species (ROS)], outpace the control of the cellular antioxidant enzymes or molecules [e.g., superoxide dismutase (SOD), peroxidases, and catalase] reaching toxic steady-state concentrations (Sies and Jones, 2020). While ROS are assumed to be critical signaling molecules for maintaining brain homeostasis, an unbalanced redox environment toward oxidation is recognized to play a pivotal role in the development of cerebrovascular dysfunction in different pathologies. In the context of AD, Aβ has been demonstrated to induce excessive ROS production in the brain, this occurring earlier in the vasculature than in parenchyma (Park et al., 2004).

At the cerebral vasculature, ROS can be produced by different sources, including NADPH oxidase (NOX), mitochondria respiratory chain, uncoupled eNOS, and cyclooxygenase (COXs), among others. In this list, the NOX family has been reported to produce more ROS [essentially O₂^–• but also hydrogen peroxide (H₂O₂)] than any other enzyme. Interestingly, the NOX activity in the cerebral vasculature is much higher than in the peripheral arteries (Miller et al., 2006) and is further increased by aging, AD, and VCID (Choi and Lee, 2017; Ma et al., 2017). Also, both the NOX enzyme activity level and protein levels of the different subunits (p67phox, p47phox, and p40phox) were reported to be elevated in the brains of patients with AD (Ansari and Scheff, 2011) and AD transgenic mice in correlation with a cognitive decline (Park et al., 2008; Bruce-Keller et al., 2011; Han et al., 2015; Lin et al., 2016). As mentioned earlier, NOS enzymes may produce O₂^–• themselves in their uncoupled state, critically contributing to the decreased BH₄ bioavailability. Of note, the BH₄ metabolism is described to be deregulated in AD (Foxton et al., 2007).

The reaction of O₂^–• with ^•NO proceeds at diffusion-controlled rates and is favored by an increased steady-state concentration of O₂^–•, providing that ^•NO diffuses to the sites of O₂^–• formation. This radical-radical interaction has two important consequences for cerebrovascular dysfunction: (1) reduced ^•NO bioavailability and ensued diminished ^•NO- mediated vasodilation, and (2) increased peroxynitrite formation and fostering of oxidative distress (Figure 2).

The decrease of ^•NO bioavailability is translated into inefficient signaling to support the hemodynamic response to neuronal activation and might occur either due to increased scavenging (e.g., by O₂^–•) or limited ^•NO synthesis (e.g., in hypoxia). Thus, the fast reaction of ^•NO with O₂^–• might be a persuasive reaction at the origins of the ^•NO decreased bioavailability and NVC dysfunction.

The impact of O₂^–•-mediated ^•NO scavenging in the efficiency of NVC has been addressed in different animal models. In healthy young rats, we were able to mimic the impairment of ^•NO-mediated NVC observed in the hippocampus of 3xTg-AD mice and aged Fisher 344 rats by using 2,3-dimethoxy-1,4-naphthoquinone (DMNQ), a redox-cycle quinone that triggers the intracellular O₂^–• generation (Lourenço et al., 2017b, 2018). Park and coworkers demonstrated that the attenuation of the CBF changes to whisker stimulation, as observed in aged and AD mice, could be abrogated by both a SOD mimetic (MnTBAP) and a NOX inhibitor (peptide gp91ds-tat) (Park et al., 2007, 2008). In the Tg2576 mice, it was further shown that the genetic ablation of the NOX2 subunit of NADPH oxidase was able to prevent both the NVC dysfunction and spatial memory decline (Park et al., 2008). More recently, the mitochondria-targeted overexpression of catalase has been shown to hamper the age-related NVC dysfunction by preserving the ^•NO-mediated component of the hemodynamic response (Csiszar et al., 2019).

The ^•NO synthesis by the NOS enzymes involves the oxidation of L-arginine to L-citrulline, dependent on O₂. Under conditions of limited O₂ concentration (e.g., ischemic conditions) and going lower than the K_M for NOS, the synthesis of ^•NO by the canonical pathway became limited, and expectedly, the ^•NO concentration decreases (Adachi et al., 2000).

As mentioned earlier, the reaction of ^•NO with O₂^–•, yielding ONOO^–, conveys the major pathway underlying the deleterious actions of ^•NO, that eventually culminates into neurodegeneration (Radi, 2018). This pathway is largely fueled by the activity of iNOS, an isoform much less dependent on Ca²⁺ concentration and capable to sustain a continuous ^•NO production, thereby producing a much larger amount of ^•NO relative to the constitutive isoforms (Pautz et al., 2010). The ONOO^– formed can oxidize and nitrate several biomolecules, including proteins. Specifically, the nitration of the tyrosine residues of proteins, resulting in the formation of 3-nitrotyrosine (3-NT), may irreversibly impact signaling pathways (either by promoting a loss or a gain of function of the target protein) (Radi, 2018).

A large body of evidence supports the enhanced 3-NT immunoreactivity in the brains of AD patients and rodent models, as well as the nitration and oxidation of several relevant proteins [reviewed in Butterfield et al. (2011) and Butterfield and Boyd-Kimball (2019)]. Among them, the mitochondrial isoform of SOD (MnSOD) was reported to occur nitrated in AD (Aoyama et al., 2000), a modification associated with enzyme inactivation (Radi, 2004) and expected increased oxidative distress. Also, tau protein has been demonstrated to be a target for nitration, a modification linked to increased aggregation (Horiguchi et al., 2003). In the 3xTgAD mice with impaired NVC, we detected increased levels of 3-NT and iNOS of the hippocampus (Lourenço et al., 2017b). Peroxynitrite can further impair NVC by altering the mechanisms for vasodilation (e.g., oxidizing BH₄, inhibiting sGC expression/activity, inactivating prostacyclin) and by promoting structural alterations in the blood vessels [reviewed by Chrissobolis and Faraci (2008) and Lee and Griendling (2008)].

In the past two decades, both arginine (NOS substrate) and its precursor, citrulline, received a lot of attention for their potential to increase ^•NO-dependent signaling, being currently accepted that citrulline is more efficient than arginine as a precursor for ^•NO synthesis (Bahadoran et al., 2021). Citrulline, but not arginine, has been demonstrated to improve the acetylcholine-induced retinal vasodilation in diabetic rats (Kurauchi et al., 2017). It also has been shown to mitigate the cognitive dysfunction linked to transient brain ischemia in mice (Yabuki et al., 2013) and to improve the CBF response associated with spreading depolarization events in rats (Kurauchi et al., 2017). Yet, in patients with mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS) syndrome, oral treatment with L-Arginine was able to restore the NVC response to the visual cortex stimulus (Rodan et al., 2020).

More recently, much focus has been paid to both NO₂^– and NO₃^– anions as relevant biological precursors of ^•NO, capable to support ^•NO-dependent mechanisms, including vasodilation, via the NO₃^–-NO₂^–-^•NO pathway (Lundberg et al., 2008). In line with this, different pieces of evidence provide support for the benefits of NO₂^– and NO₃^–, particularly in the cardiovascular system (Lundberg et al., 2015). Additionally, acute nitrite was demonstrated to enhance basal CBF in rats (Rifkind et al., 2007) and to improve the hemodynamic response to somatosensory activation under conditions of NOS inhibition (Piknova et al., 2011). In essence, NO₂^– is now recognized as a regulator of hypoxic signaling in mammalian physiology through the formation of ^•NO (Rocha et al., 2011; Lundberg et al., 2018), and human studies support the application of NO₂^– to alleviate cardiovascular dysfunctions and as a supplement for healthy arterial aging (Lundberg et al., 2018). NO₃^–, which can be obtained from diets, salt supplementation, or endogenous ^•NO oxidation, is also proposed to increase both cerebral blood perfusion and cognitive performance in humans, but the overall evidence is still not conclusive (Clifford et al., 2019). This may be justified by the short duration of the interventions and the health status of the subjects. Magnetic resonance imaging studies in an elderly human population exposed to an increased intake of NO₃^–-rich foods showed that a 2-day intervention with dietary NO₃^– had no effect on the global CBF but enhanced the CBF in the deep white matter in the frontal lobe, suggested to be involved in executive functioning (Presley et al., 2011). Moreover, in healthy adults, a single dose of beetroot juice (containing 5.5 mmol NO₃^–) was demonstrated to modulate their CBF response to task performance in the frontal cortex assessed by near-infrared spectroscopy and to improve their performance to a serial 3 s subtraction task (Wightman et al., 2015). Also, intrathecal levels of NO₃^– in patients with vascular dementia are inversely correlated to the degree of intellectual impairment (Tarkowski et al., 2000). In the cardiovascular and renal systems, dietary NO₃^– was able to attenuate oxidative distress by inhibiting NOX expression/activity, conveying an additional mechanism by which NO₃^– supplementation could improve ^•NO bioavailability (Carlstrom and Montenegro, 2019).

Found predominantly in fruits and vegetables, polyphenols are another set of bioactive compounds recognized to improve ^•NO bioavailability, thereby providing positive health benefits (Oak et al., 2018). For decades, explored for their antioxidant capacity associated with the free radical scavenging, polyphenols are recognized to directly modulate the redox-sensitive processes in vivo (Fraga et al., 2019). Moreover, polyphenols are known to stimulate ^•NO production in the endothelial cells via the phosphorylation of eNOS mediated by both PI3K/Akt and p38 MAPK signaling pathways. Several polyphenols were also described to modulate the redox status of the cellular environment via the activation of the nuclear factor (erythroid-derived 2)-like 2 (Nrf2), a transcription factor needed for anti-oxidant response element gene (ARE) expression, leading to the expression of several antioxidant enzymes (e.g., SOD). Interestingly, they have also been demonstrated to downregulate the expression of NOX enzymes (Forte et al., 2016; Fraga et al., 2018). Furthermore, as mentioned above, polyphenols are also capable to promote the reduction of NO₂^– to ^•NO, particularly under acidic environments (Rocha et al., 2009). Polyphenols accompany NO₂^– and NO₃^– in the vegetables consumed in the human diet and mounting evidence from epidemiological and randomized controlled trials support their beneficial effects in preventing age-associated cognitive decline (Fraga et al., 2019). The ability of some polyphenols to improve NVC has also been demonstrated by some clinical and pre-clinical studies. For instance, resveratrol, a polyphenol found in grape seeds, was able to rescue the hemodynamic responses to neuronal and endothelial activation in aged mice, occurring along with downregulation of the NADPH oxidase and decreased 3-NT levels (Toth et al., 2014). In humans, a similar beneficial effect was demonstrated in type 2 diabetic patients and postmenopausal women. In diabetic patients, a single dose of resveratrol enhanced their performance on a multi-tasking test battery and the blood flow velocity in the middle cerebral artery to this cognitive stimulus (Wong et al., 2016). In postmenopausal women, a long-term treatment with resveratrol significantly increased their cognitive performance and NVC response as compared with their basal condition (Thaung Zaw et al., 2020). Cocoa flavonoids were also demonstrated to improve NVC in humans, increasing the BOLD fMRI response to cognitive tasks in healthy subjects (Decroix et al., 2016) and type 1 diabetic patients (Decroix et al., 2019).

Regular physical exercise is widely recognized to have a substantial neuroprotective impact on brain function, contributing to restore and maintain cognitive performance (Alkadhi, 2018), and amending several cardiovascular risk factors (e.g., diabetes, hypertension) (Kokkinos and Myers, 2010). In particular, exercise is suggested to promote the upregulation of eNOS expression and activity via activation of Akt-dependent signaling by temporally increasing shear stress (Chistiakov et al., 2017). Shear stress has also been shown to stimulate the expression of Nrf2 and thus, the upregulation of antioxidant defenses (Warabi et al., 2007). Also, exercise is suggested to enhance NO₂^– plasmatic levels similar to dietary NO₃^– (Lundberg et al., 2015). The evidence for the beneficial effect of exercise in the CBF and brain functioning is overwhelmingly convincing both in humans and experimental animal models (Ogoh and Ainslie, 2009; Smith and Ainslie, 2017; Trigiani and Hamel, 2017). Recently, physical exercise has been demonstrated to alter the white matter pathology and hemodynamic response to whiskers stimulation in a VCID mouse model (Trigiani et al., 2020). Also, in an AD mouse model, exercise training was demonstrated to reduce cerebrovascular dysfunction via enhancing the P2Y2-eNOS signaling (Hong et al., 2020).