. Caveolae in CNS arterioles mediate neurovascular coupling. Nature. 2020 Mar;579(7797):106-110. Epub 2020 Feb 19 PubMed.


Please login to recommend the paper.


  1. Cerebral blood flow (CBF) and the increase in CBF evoked by neuronal activity are reduced early in Alzheimer’s disease (AD) (Iturria-Medina et al., 2016), with the decrease in CBF exceeding 40 percent in affected areas (Asllani et al., 2008). In humans developing AD, and in an AD mouse model, contractile pericytes constrict cortical capillaries sufficiently to generate these CBF decreases (Nortley et al., 2019), and in the AD mouse model there is no change in arteriole or venule diameter implicating pericyte constriction as the main cause of the CBF decrease (Nortley et al., 2019). Two new papers, one (Grubb et al., 2020) from the Lauritzen group in Copenhagen and one (Chow et al., 2020) from the Gu group at Harvard, now report novel mechanisms for the control of CBF, which may have implications for our understanding of the CBF fall that occurs in AD.

    CBF can be regulated by vascular smooth muscle cells around arterioles, or by contractile pericytes on the first four branching orders of capillaries coming off an arteriole (Hall et al., 2014). The Lauritzen group now report (Grubb et al., 2020) a specialized structure—termed a sphincter—present on 28 percent of the first-order capillaries branching off cortical penetrating arterioles. The sphincter consists of a narrowed region of the capillary wrapped by an NG2-expressing contractile cell, which conventionally would be called a pericyte (although the paper is careful to avoid getting involved in the debate about the nomenclature of the contractile cells on capillaries) (Attwell et al., 2016). 

    The authors found that some of these contractile cells were weakly labelled by the Nissl dye NeuroTrace 500/525. This was previously reported to only label non-contractile pericytes (Damisah et al., 2017), but our own unpublished data confirm labelling by NeuroTrace 500/525 of a substantial proportion of α-smooth muscle actin expressing pericytes (which contracted in response to a thromboxane A2 analogue).

    The sphincter has a smaller diameter than the adjacent part of the first-order capillary. The authors propose that this allows the sphincter to more significantly reduce the blood pressure being applied to downstream capillaries, and indeed sphincters are found preferentially in the capillaries coming off the penetrating arteriole near the cortical surface where the blood pressure is higher. Conceivably, pericytes or endothelial cells on capillaries downstream of the sphincter-containing vessel may sense neuronal activity and send a signal back to the sphincter to alter blood flow (Longden et al., 2017). It will be exciting to determine whether, in AD, constriction of the pericytes forming the sphincters has a larger effect on CBF than constriction of contractile pericytes that are not forming sphincters, or whether the sphincter-fed vessels are relatively protected from constriction. Detailed computer modelling, based on reconstructions of the vascular topology (Blinder et al., 2013; Kirts et al., 2020) will probably be needed to assess this. 

    The Gu group paper (Chow et al., 2020) proposes a completely novel mechanism regulating neurovascular coupling, based on the effect of knocking out the protein caveolin-1 from the endothelial cells forming arteries. Caveolin-1 is present in the membranes of caveolae—indentations of the plasma membrane that have been variously proposed to play a role in mediating transcytosis, regulating endothelial cell Ca2+ signalling, sorting membrane proteins, regulating G protein coupled receptors, anchoring GPI-linked proteins, modulating eNOS activity, mechanosignalling, regulating inflammatory pathways, regulating Akt and Erk signalling, and controlling cell proliferation.

    Gu’s group had earlier shown that suppression of caveolae-mediated transcytosis in endothelial cells by the lipid transporter Mfsd2a plays a key role in establishing the blood-brain barrier (Andreone, 2018). Surprisingly, they now report that, although this is true for the endothelial cells forming capillaries, which contain few caveolae, there are a large number of caveolae in the endothelial cells forming arterioles. This suggests that the blood-brain barrier may be weaker for arterioles than for downstream capillaries.

    When caveolin-1 was knocked out in arterial endothelial cells, caveolae were suppressed in these endothelial cells, and the 15 percent dilation of arterioles normally evoked by neuronal activity was roughly halved (their Fig. 2c), suggesting a major role for caveolin-1 in regulating neurovascular coupling at the arteriole level. A 15 percent increase of arteriole diameter would be expected (by simplistic use of Poiseuille’s law) to evoke a 1.154 = 1.75-fold decrease in arteriole resistance, and hence a 75 percent increase in flow if arterioles constituted the only significant resistance in the cortical vasculature. (In fact, arterioles comprise only around one-seventh of the total vascular resistance from the pial end of a penetrating arteriole to the pial end of a venul (Blinder et al., 2013), so the predicted increase would in fact be much smaller).

    Intriguingly, however, the authors in fact documented a fourfold (300 percent) increase in downstream capillary blood flow (their Fig. 2d), indicating that most of the increase in blood flow evoked by neuronal activity is presumably generated by relaxation of contractile pericytes on downstream capillaries. This idea could be tested by measuring capillary diameters near pericytes. This raises the question of whether the change of neurovascular coupling that they see on deleting caveolin-1 is not, as one might expect, occurring at the arteriole smooth muscle adjacent to the endothelial cells where the caveolae are being suppressed, but somehow instead at downstream capillary pericytes.

    The paper does not provide mechanistic investigations of how deletion of caveolin-1 alters neurovascular coupling (other than showing it is independent of eNOS activity). Whether the effects observed occur at the arteriole smooth muscle level or the capillary pericyte level, it will be fascinating to determine whether the caveolae-mediated effect is a novel neurovascular coupling mechanism independent of previously known pathways, as has been previously suggested to exist (Hosford and Gourine, 2019)  or whether caveolin-1 deletion somehow modulates the properties of an already known pathway.

    The role of the caveolin-regulated pathway in disorders such as AD will also be of great interest, since the caveolin-1 KO mouse has already been proposed (Dudau et al., 2016) as a model of dementia, including AD, and there is debate about the effect of changes in the levels of caveolin 1-3 in AD (Gaudreault et al., 2004; van Helmond et al., 2007; Alsaqati et al., 2017). 


    . Early role of vascular dysregulation on late-onset Alzheimer's disease based on multifactorial data-driven analysis. Nat Commun. 2016 Jun 21;7:11934. PubMed.

    . Multivariate and univariate analysis of continuous arterial spin labeling perfusion MRI in Alzheimer's disease. J Cereb Blood Flow Metab. 2008 Apr;28(4):725-36. PubMed.

    . Amyloid β oligomers constrict human capillaries in Alzheimer's disease via signaling to pericytes. Science. 2019 Jul 19;365(6450) Epub 2019 Jun 20 PubMed.

    . Precapillary sphincters maintain perfusion in the cerebral cortex. Nat Commun. 2020 Jan 20;11(1):395. PubMed.

    . Capillary pericytes regulate cerebral blood flow in health and disease. Nature. 2014 Apr 3;508(7494):55-60. Epub 2014 Mar 26 PubMed.

    . What is a pericyte?. J Cereb Blood Flow Metab. 2016 Feb;36(2):451-5. Epub 2015 Oct 14 PubMed.

    . A fluoro-Nissl dye identifies pericytes as distinct vascular mural cells during in vivo brain imaging. Nat Neurosci. 2017 Jul;20(7):1023-1032. Epub 2017 May 15 PubMed.

    . Capillary K+-sensing initiates retrograde hyperpolarization to increase local cerebral blood flow. Nat Neurosci. 2017 May;20(5):717-726. Epub 2017 Mar 20 PubMed.

    . The cortical angiome: an interconnected vascular network with noncolumnar patterns of blood flow. Nat Neurosci. 2013 Jun 9; PubMed.

    . Mapping the Fine-Scale Organization and Plasticity of the Brain Vasculature. Cell. 2020 Feb 8; PubMed.

    . Blood-Brain Barrier Permeability Is Regulated by Lipid Transport-Dependent Suppression of Caveolae-Mediated Transcytosis. Neuron. 2017 May 3;94(3):581-594.e5. Epub 2017 Apr 13 PubMed.

    . What is the key mediator of the neurovascular coupling response?. Neurosci Biobehav Rev. 2019 Jan;96:174-181. Epub 2018 Nov 24 PubMed.

    . Arguments for caveolin-1 knock-out mice as an Alzheimer’s disease model. Austin Alzheimers J Parkinsons Dis. June 2, 2016

    . Increased caveolin-1 expression in Alzheimer's disease brain. Neurobiol Aging. 2004 Jul;25(6):753-9. PubMed.

    . Caveolin-1 and -2 and their relationship to cerebral amyloid angiopathy in Alzheimer's disease. Neuropathol Appl Neurobiol. 2007 Jun;33(3):317-27. PubMed.

    . Proteins involved in endocytosis are upregulated by ageing in the normal human brain: implications for the development of Alzheimer's disease. J Gerontol A Biol Sci Med Sci. 2017 Jun 26; PubMed.

    View all comments by David Attwell
  2. Brian Chow, Vicente Nuñez, and colleagues from Chenghua Gu’s lab at Harvard reveal the molecular mechanisms for a fascinating and fundamental feature of the brain vasculature: How do blood vessels “know” when neurons and other brain cells need more fuel and blood supply—and how exactly is the process regulated? Do all vascular cells participate in this process or only some, and if so, which ones?

    Neuronal activity induces a rapid increase in regional blood flow, referred to as neurovascular coupling. The conventional model posits that this is mediated by neuron-derived vasodilatory factors which directly relax arterial smooth-muscle cells (SMCs). Yet from recent work, it seems brain endothelial cells (ECs) can sense neural activity as well. Perhaps then, vasodilatory signals act first on ECs before being relayed to SMCs.

    Chow et al. explore this potential EC-mediated neurovascular coupling by adopting a very elegant approach. They focus on the barrel cortex, where whisker stimulation reliably triggers neural activity, vessel dilation, and blood flow. There, they show that classic NO sensing in smooth muscle cells is insufficient for full neurovascular coupling. Instead, the caveolae enriched in arteriolar ECs are also necessary for effective coupling. Using various cell-type-specific and global gene knockout and overexpression mouse models, they confirm that caveolae in ECs—but not in SMCs—are required for neurovascular coupling. Thus, EC caveolae and general eNOS are necessary for a full coupling response.

    These discoveries inspire exciting questions relevant to understanding the biology of the brain vasculature in health, aging, and disease. First, what is the mechanism by which caveolae mediate neurovascular coupling? Are they facilitating transport of specific vasodilatory molecules and what are they? In physiological vasodilation, are there cues to preferentially engage EC caveolae versus the NO pathway—and does this change if not in the barrel cortex?

    Second, the authors have previously shown caveolae to mediate relatively nonspecific transport. Caveolae likely transport albumin, fibrinogen, and other possibly detrimental factors into the brain, such as in the context of acute brain trauma. Given the abundance of caveolae in arterioles, what special mechanisms prevent this presumable bulk uptake of blood factors?

    Third, how do the documented changes in EC gene expression with aging relate to neurovascular coupling? For example, recent papers document an increase in caveolin expression in aged capillaries. By analogy, can age-upregulated capillary caveolae provide vasodilatory signals to contractile pericytes?

    Finally, how does this model change with disease? For example, the vascular accumulation of β-amyloid in cerebral amyloid angiopathy has been correlated with a loss of SMCs. In short, the current study provides new impetus for studying the fascinating complexity of the brain vasculature and hopefully will pave the way towards a better understanding of how this structure degenerates with age and disease.

  3. This paper by Gu et al. shows that caveolae in the arterial endothelial cells are required for neurovascular coupling at the level of vascular smooth-muscle cells and arterioles. This is an important and novel finding. It provides new mechanistic insight into the mechanisms of neurovascular coupling.

    At the same time, it raises the question whether a similar mechanism operates at the level of brain capillaries, since some recent studies have suggested that capillary dilation precedes arteriolar dilation during neurovascular coupling.  This would be an important direction for future studies.

  4. Grubb et al. provide an excellent structural and functional description analysis of the precapillary sphincters. They express alpha smooth-muscle actin and basement membrane proteins, and their main role appears to be in reducing the perfusion pressure in the capillaries. Apart from preventing cerebrovascular complications related to high blood pressure, the sphincters may have a role in smooth vasomotion and the motive force for drainage of interstitial fluid from the brain (Carare et al., 2020). 

    Chow et al. demonstrate the presence of caveolae on arterial endothelium. They are essential for neurovascular coupling, whereas caveolae on smooth-muscle cells are not necessary for neurovascular coupling. The major facilitator superfamily domain-containing 2a (Mfsd2a) protein is absent from the caveolae of arterial endothelium, while it is associated with the blood-brain barrier and repair after injury such as subarachnoid hemorrhage.

    It will be exciting to decipher the full molecular pathways involved in neurovascular coupling and dysregulation, as they could provide valuable therapeutic targets for stroke or subarachnoid hemorrhage.


    . Vasomotion Drives Periarterial Drainage of Aβ from the Brain. Neuron. 2020 Feb 5;105(3):400-401. PubMed.

    View all comments by Roxana Carare
  5. The two papers by Chow et al. and Grubb et al. significantly advance our understanding of neurovascular physiology and blood flow control.

    Through a combination of elegant in vivo two-photon imaging, molecular manipulations, pharmacology, and electron microscopy, Chow et al. report that neurovascular coupling is mediated by arteriolar endothelium to smooth muscle cell (SMC) interactions via an endothelial caveolae-dependent mechanism.

    They show that deletion of Caveolin 1 exclusively in endothelium markedly reduces the amplitude of vasodilation following whisker stimulation. Additional deletion of endothelial nitric oxide synthase (eNOS) almost completely abolished neurovascular coupling. Ectopic expression of MFSD2A, a molecule they previously identified as a suppressor of caveolae formation (Ben-Zvi et al., 2014), also attenuates NVC.

    There are several interesting conclusions from this study:

    1) The arteriolar endothelium appears to be a central hub through which all other signals (possibly neuronal or astrocytic) need to go to achieve vasomotility. In the absence of caveolae or endothelial eNOS, NVC coupling is essentially abolished.

    2) Caveolae are markedly enriched in arteriolar endothelium, but not in capillaries. This is consistent with the observation that the overlying smooth muscle cells (SMCs) but not capillary pericytes are the contractile cells that modulate NVC (Hill et al., 2015; Grutzendler and Nedergaard, 2019; Jun 2015 news). 

    3) Their observations are consistent with previous studies, mainly in the peripheral vasculature, showing that loss of Caveolins induces endothelial dysfunction and reduces myogenic tone, likely through loss of eNOS signaling (Lian et al., 2019). 

    4) Their results are complementary to recent work from the Nelson lab showing a role of endothelium in potassium sensing through inward-rectifier K+ (KIR2.1) channels to produce a rapidly propagating retrograde hyperpolarization that causes upstream arteriolar dilation (Longden et al., 2017). 

    5) Their demonstration of structural and functional regional endothelial differences is consistent with the differential transcriptome of arteriolar and capillary endothelial cells as well as the overlying SMCs and pericytes as observed by single-cell RNA seq analysis (Vanlandewijck et al., 2018) and the differential labeling with the uniquely pericyte-specific (completely excluded from SMCs) Neurotrace 500/525 dye which we previously described (Damisah et al., 2017). 

    In Grubb et al., the authors implement in vivo two-photon imaging in mice that express a fluorescent reporter in all mural cells (NG2-DsRed) while studying aspects of neurovascular coupling. They report the unexpected finding that in addition to the normal-looking, ring-like SMCs around brain arterioles (Hill et al., 2015), there is a small subpopulation of SMCs that have the appearance of a sphincter with greater degree of contractility, and which are located at the transition between the penetrating arteriole and first-order branches.

    They provide some evidence, mostly by modeling, that these sphincter-like SMCs generate the largest changes in resistance of all brain vessels and might be important for protecting vasculature and brain tissue from potentially adverse pressure changes.

    This study raises several questions:

    1) What is unique about these cells from a cellular and molecular point of view compared to the immediately adjacent SMCs?

    2) Are these sphincter-like properties stable over time, or are many of these branchpoint SMCs capable of hypercontractility under certain physiological or pathological circumstances?

    3) The authors demonstrate the hypercontractile features of these cells during spreading depolarization, potentially a critical finding in the context of ischemia, migraine, and microvascular occlusions.

    In summary, this paper adds significantly to our understanding of the structural and functional diversity of mural cells in the brain.


    . Mfsd2a is critical for the formation and function of the blood-brain barrier. Nature. 2014 May 22;509(7501):507-11. Epub 2014 May 14 PubMed.

    . Regional Blood Flow in the Normal and Ischemic Brain Is Controlled by Arteriolar Smooth Muscle Cell Contractility and Not by Capillary Pericytes. Neuron. 2015 Jun 23; PubMed.

    . Cellular Control of Brain Capillary Blood Flow: In Vivo Imaging Veritas. Trends Neurosci. 2019 Aug;42(8):528-536. Epub 2019 Jun 26 PubMed.

    . Pathophysiological Role of Caveolae in Hypertension. Front Med (Lausanne). 2019;6:153. Epub 2019 Jul 10 PubMed.

    . Capillary K+-sensing initiates retrograde hyperpolarization to increase local cerebral blood flow. Nat Neurosci. 2017 May;20(5):717-726. Epub 2017 Mar 20 PubMed.

    . A fluoro-Nissl dye identifies pericytes as distinct vascular mural cells during in vivo brain imaging. Nat Neurosci. 2017 Jul;20(7):1023-1032. Epub 2017 May 15 PubMed.

    . A molecular atlas of cell types and zonation in the brain vasculature. Nature. 2018 Feb 14; PubMed.

    View all comments by Lei Tong

Make a Comment

To make a comment you must login or register.

This paper appears in the following:


  1. Ruffles and Sphincters Control the Spigot of Fresh Blood in the Brain