A NO way to BOLD?: Dietary nitrate alters the hemodynamic response to visual stimulation

Highlights

• We studied the effect of nitrate on the brain's hemodynamic responses with BOLD fMRI.

• Nitrate intake decreases the hemodynamic lag and the BOLD amplitude.

• Nitrate intake also decreases the voxel-wise variation of both these measures.

• We propose that the nitrate–nitrite–NO cycle plays a role in neurovascular coupling.

Abstract

Neurovascular coupling links neuronal activity to vasodilation. Nitric oxide (NO) is a potent vasodilator, and in neurovascular coupling NO production from NO synthases plays an important role. However, another pathway for NO production also exists, namely the nitrate–nitrite–NO pathway. On this basis, we hypothesized that dietary nitrate (NO₃⁻) could influence the brain's hemodynamic response to neuronal stimulation. In the present study, 20 healthy male participants were given either sodium nitrate (NaNO₃) or sodium chloride (NaCl) (saline placebo) in a crossover study and were shown visual stimuli based on the retinotopic characteristics of the visual cortex. Our primary measure of the hemodynamic response was the blood oxygenation level dependent (BOLD) response measured with high-resolution functional magnetic resonance imaging (0.64 × 0.64 × 1.8 mm) in the visual cortex. From this response, we made a direct estimate of key parameters characterizing the shape of the BOLD response (i.e. lag and amplitude). During elevated nitrate intake, corresponding to the nitrate content of a large plate of salad, both the hemodynamic lag and the BOLD amplitude decreased significantly (7.0 ± 2% and 7.9 ± 4%, respectively), and the variation across activated voxels of both measures decreased (12.3 ± 4% and 15.3 ± 7%, respectively). The baseline cerebral blood flow was not affected by nitrate.

Our experiments demonstrate, for the first time, that dietary nitrate may modulate the local cerebral hemodynamic response to stimuli. A faster and smaller BOLD response, with less variation across local cortex, is consistent with an enhanced hemodynamic coupling during elevated nitrate intake. These findings suggest that dietary patterns, via the nitrate–nitrite–NO pathway, may be a potential way to affect key properties of neurovascular coupling. This could have major clinical implications, which remain to be explored.

Introduction

Nitric oxide (NO) is a potent vasodilator playing an important role in establishing neurovascular coupling in the brain (Attwell et al., 2010). In this context, NO is normally considered to originate from NO synthases (NOSs), but the human body also possesses another pathway for NO production that has received much attention recently. This pathway is known as the nitrate–nitrite–NO pathway (Lundberg et al., 2008). However, the impact of this pathway on the brain's hemodynamic responses has so far attracted little attention. Nitrate is often considered a toxin (Avery, 1999), but via its reduction to nitrite and NO, it is known to subserve several important physiological functions (Kevil et al., 2011). E.g. nitrate intake can increase exercise performance (Larsen et al., 2007) and decrease blood pressure (Larsen et al., 2006), and increased levels of nitrite in the blood can help protect against ischemia reperfusion injury (Jung et al., 2006). Most of the nitrate consumed originate from green leafy vegetables (Alexander et al., 2008), and after ingestion, nitrate is reduced to nitrite via symbiotic salivary bacteria (Goaz and Biswell, 1961). Subsequently, nitrite can be converted into NO through an enzyme (van Faassen et al., 2009) and in a pH-dependent manner (Aamand et al., 2009, Li et al., 2008, Li et al., 2009, Zweier et al., 2010). As NO readily crosses cell membranes (Garthwaite, 2008), this means that the blood–brain–barrier is also readily crossed. Within the brain, the conversion of nitrite to NO could be facilitated by hemoglobin (Huang et al., 2005), xanthine oxidase (Zhang et al., 1997), endothelial NOS (Gautier et al., 2006), or carbonic anhydrase (CA) (Aamand et al., 2009). As an increase in pCO₂ leads to a decrease in pH by CA (Lindskog, 1997), the NO production from the nitrate–nitrite–NO pathway is inherently linked to the energy metabolism of a tissue. Since NO is a potent vasodilator (Ignarro et al., 1987), this could afford a direct coupling between neuronal activity and increases in blood flow.

To study whether the nitrate–nitrite–NO pathway is of physiological importance to the hemodynamic response to neuronal activation, we increased or minimized the dietary intake of nitrate in two groups of young healthy male participants in a randomized, double-blinded, placebo-controlled, crossover design. A 3-day intervention period was chosen as the effects of prolonged nitrate intake are not necessarily mimicked by acute interventions (Larsen et al., 2011). Participants ingested weight-adjusted dosages of sodium chloride (NaCl = saline placebo) or sodium nitrate (NaNO₃) corresponding to the amount of nitrate in 500 mL beetroot juice or a large plate of salad (Alexander et al., 2008, Larsen et al., 2007).

In human participants, the blood oxygenation level dependent (BOLD) response used in functional magnetic resonance imaging (fMRI) (Ogawa et al., 1990) is a highly sensitive way of studying the cerebral vasculature's response to neuronal activity. The BOLD response depends on local changes in the deoxyhemoglobin (deoxyHb) content modulated by cerebral blood flow (CBF), cerebral blood volume (CBV), and the cerebral O₂ consumption (CMRO₂). Upon neuronal stimulation, the concerted changes in these parameters cause the local deoxyHb content to decrease, and as a consequence the BOLD signal increases (Ogawa et al., 1990).

As neuronal activity causes the extracellular pH to decrease (Chesler, 2003), and as more NO is produced from nitrite at lower levels of pH, we expected that a higher intake of nitrate would cause more NO to be produced via the nitrate–nitrite–NO pathway during neuronal activity. This could cause the local vasculature to dilate more readily when the neuronal activity increases, with little or no impact globally. We thus hypothesized that a decrease in the hemodynamic lag, in this case a decreased lag of the BOLD response, could be expected upon increased intake of nitrate. With regards to the amplitude of the BOLD response we were less certain of what to expect. Following the same logic, a larger BOLD amplitude could be expected. However, nitrate also increases the oxidative phosphorylation efficiency (P/O ratio) of the mitochondria (Larsen et al., 2011), and thus potentially decreases the need for O₂ and glucose. Hence, if the BOLD amplitude has any relation to neuronal metabolism, one could expect a decrease in BOLD amplitude upon heightened nitrate intake.

Despite early acknowledgement of its importance (Bandettini et al., 1993, Henson et al., 2002, Menon et al., 1998), the hemodynamic lag is rarely explicitly estimated and studied in BOLD fMRI studies — perhaps because studies have to be specifically designed with this in mind for it to be feasible. An exception to this is the field of retinotopic mapping where the hemodynamic lag is estimated routinely in order to reliably delineate the borders of the visual cortex (Sereno et al., 1995). However, in this context the hemodynamic lag itself is rarely a topic of scrutiny, but rather a confounder, which needs to be estimated in order to construct reliable retinotopic maps. In the present study, we turned this priority around. Three features of the visual cortex makes it an optimal area of choice for estimating hemodynamic lag: 1) Its relatively large size allows for a robustly sized dataset; 2) the existence of ocular dominance columns makes it possible to minimize vascular smearing by using a visual stimulus matching the extent of such columns (Turner, 2002); and 3) it is possible, in an fMRI setting, to reverse the stimulation order due to the retinotopic organization of the visual cortex. Together these three features allow for a reliable and time-efficient estimation of the hemodynamic lag in a large number of voxels, with minimal contribution from large vessels. This is done by considering the difference between the phase angles obtained for opposite stimulus directions (i.e. contraction versus expansion) (Sereno et al., 1995). This also means that the BOLD response is easily parameterized in terms of phase, amplitude, and frequency.