Tau Modulates Neurovascular Coupling

The microtubule-associated protein tau was identified in the neurofibrillary tangles (NFTs) of Alzheimer’s disease (AD) more than 30 years ago, and its mutation can directly cause neurodegeneration. As implied by its purification from microtubules, tau was thought to stabilize microtubules and therefore to be of crucial importance in the brain. However, the physiological function of tau is still in debate since there is no detectable deficit in tau-knockout mice until they are 12 months old [1]. Recently, Park and colleagues [2] have discovered that tau binds to PSD-95, in competition with neuronal nitric oxidase synthase (nNOS). Decoupling of nNOS from PSD-95 leads to reduced production of nitric oxidase (NO), and therefore impairs cerebrovascular coupling. Suppression of the mutant tau (P301L and P301S) restores NO production and improves cerebrovascular and cognitive functions. These results provide further evidence for the toxic-gain-of-function hypothesis of tau, suggesting that neurovascular coupling may be therapeutically targeted.

Neurovascular coupling (NVC) is measured as the spatiotemporal distribution of blood flow, regulated by cerebral blood flow (CBF) within local areas, and Park et al. examined two tau transgenic mouse models (PS19 and rTg4510 mice) that overexpress the disease-related tau mutants P301S and P301L, respectively, for NVC function. They found that the increases of CBF and dilation of the arteriolar network in the whisker barrel cortex upon whisker stimulation were attenuated in both PS19 and rTg4510 mice compared with WT mice, indicating that NVC might be impaired by the overexpression of mutant tau. The vasodilatation of intracerebral arterioles is therefore impaired, which leads to the NVC dysfunction. This dysfunction occurs at the age of 2–3 months, which is earlier than the neuronal or microvascular damage in these mice. By examining the frequency distribution of the electrocorticogram and the amplitude of field potentials, Park and colleagues concluded that neural activity upon whisker stimulation was not affected in the mice. Similarly, NMDAR activation increased the intracellular Ca²⁺ levels equally in both PS19 and control mice, suggesting that the functional hyperemia is not due to changes in neural activity. By suppressing the tau mutation expression in rTg4510 mice, Park et al. found that the NVC deficits, as well as cognitive function, are restored without affecting astrogliosis or the formation of NFTs. These results collectively suggest that overexpression of mutant tau disrupts NVC through inhibiting the vasodilation of intracerebral arterioles (Fig. 1).

Fig. 1.

Postsynaptic neurovascular coupling and neurovascular dysfunction induced by mutated tau. Under normal conditions (left), neurovascular coupling (NVC) is achieved by a series of postsynaptic reactions in which local cerebral blood flow (CBF) sensitively responds to glutamatergic synaptic activity after facial whisker stimulation. Park et al. found that tau directly binds to PSD-95, causes the dissociation of PSD-95 and nNOS, and impedes the synthesis and release of NO (right). Local vasodilatation and CBF are therefore inhibited, leading to neurovascular dysfunction.

Park and colleagues then performed a series of experiments to explore the mechanism of tau-induced NVC impairment. Previous studies had found that nNOS is bound to NMDARs coupling with the scaffolding protein PSD-95, and activation of NMDARs stimulates NO production to dilate blood vessels [3, 4] Park and colleagues found that injection of both the NMDAR antagonist MK-801 and the NOS inhibitor l-nitroarginine (L-NNA) prevented the CBF elevation induced by whisker stimulation or NMDA stimulation in wild-type (WT) mice. PS19 mice were not affected by either drug in these studies. However, PS19 mice were responsive to acetylcholine-induced CBF increase, a process mediated by endothelial NOS, indicating that neuronal NO may be involved. By directly measuring NO production in neocortical neurons isolated from WT or mutant tau mice (both PS19 and rTg4510), the authors further found that pathological tau overexpression affected neuronal NO production. Both tau and its mutations were found in cell culture (but not tested in mouse brains) to competitively bind to PSD-95, blocking the binding of nNOS to PSD-95, which leads to a decrease in NO production. The reduced production of NO may, therefore, impair the NVC, causing cognitive dysfunction, as shown in the study.

Notably, Park and colleagues concluded that the NVC impairment in their study was independent of cognitive impairment and pathological changes in the mice. The NVC dysfunction here also occurred in 2–3 month-old PS19 and rTg4510 mice, and found no changes in cognitive function at the same age. However, other studies have shown that rTg4510 mice can be cognitive impaired as early as 2.5 months [5], and impaired synaptic function has been reported in 3-month-old PS19 mice [6]. Different behavioral tests can be a challenge to compare between laboratories, while other factors such as housing conditions, food, or slightly different genetic background may also affect the results [7]. Besides, Park et al. showed that suppression of P301L tau expression decreased AT8 immunostaining, but not Thioflavin S-positive staining. Considering these stains represent different stages of tau aggregation, further investigations are needed to support the claim.

Tau mis-sorting to dendrites, triggered by Aβ, was previously shown to transfer fyn to form NMDAR/PSD-95/fyn complex, enhancing the binding of PSD-95 to NMDARs, thereby causing downstream neurotoxicity (Fig. 2) [8]. Here, Park and colleagues found that tau at the synapse can bind to PSD-95 directly and prevent the physiological binding between PSD-95 and nNOS, which triggers the NVC impairment. Therefore, the mis-sorting of tau (either WT or mutant) may act as a prime event to trigger both synaptic toxicity and neurovascular dysfunction (figure 2). Interestingly, tau has been shown to be age-dependently related to brain ischemia-reperfusion-induced cognitive impairment through NMDAR-mediated intracellular toxicity, where it inhibits the binding of SynGAP1 to a PSD-95 protein complex [9, 10]. It would be of interest to investigate the connections between these results and determine how these events collectively contribute to neurodegeneration.

Fig. 2.

Tau functions in the synapse. Park et al. found that tau blocks the postsynaptic formation of nNOS/PSD-95 complexes and reduces the synthesis and release of NO, leading to CBF reduction (left) [2]. In the presence of Aβ, tau also recruits fyn to form fyn/PSD-95 complexes and activate downstream cell death pathways (right) [8]. Whether these processes independently or collectively cause neurodegeneration is unknown.

Tau research in the past 30 years has mainly focused on its role in NFT formation. However, by showing that pathological tau protein in the early stage of disease may impede neuronal NO production, which leads to neurovascular dysregulation, Park et al. have provided a fresh insight linking pathological tau protein and neurovascular events, which may contribute to the pathogenesis of neurodegeneration. They have highlighted the critical roles of vascular function in neurodegeneration, and the neuronal NO signaling pathway may provide new directions for early intervention for AD.

Conflict of interest

The authors declare no conflict of interest.