Pericytes, and control of brain energy supply
Although brain blood flow is often thought to be controlled by the smooth muscle around arterioles (Attwell et al., 2010), there are contractile cells (pericytes) at roughly 30 micron intervals along capillaries.
In the brain, unlike most other tissues, it turns out that most of the vascular resistance is located in capillaries rather than in arterioles, so the adjustment of capillary diameter by these cells can strongly affect blood flow. We have shown that capillary pericytes respond to messengers generated by neurotransmitter glutamate (such as prostaglandin E2) and that they respond more rapidly than arterioles to increases of neuronal activity (Peppiatt et al., 2006; Hall et al., 2014).
They also generate the majority of the blood flow rise evoked by neuronal activity, and thus may be the main driver of the BOLD fMRI signals that are used to non-invasively probe brain function by psychologists.

Interestingly, we have recently found that dilation of capillaries is evoked by a different signalling pathway from dilation of arterioles: capillary dilation reflects neuronal ATP release evoking a rise of calcium concentration in astrocytes which then triggers the release of prostaglandin E2, while arteriole dilation does not seem to involve ATP release or astrocyte calcium, and may be driven mainly by NO generated in interneurons (Mishra et al., 2016).
Movie showing that depolarisation of a pericyte constricts the retinal capillary on which it is located, and this is followed by constriction of the pericytes on each side of the stimulated cell.
Neurovascular coupling mediated by pericytes in whisker barrel somatosensory cortex of anaesthetised NG2-dsRed mouse (from Hall et al., 2014). Top left image shows a penetrating arteriole (0th order vessel, going down into the cortex) and capillaries coming off it, labelled green with FITC-dextran in the blood. Red cell is a pericyte on the 1st order capillary, at the junction with the 2nd order capillaries.
Top right movie shows that the capillary dilates before the arteriole when the whisker pad is stimulated at t=0, implying active relaxation by the pericyte in response to neuronal activity. Our recent work (Mishra et al., 2016) shows that this is mediated by a calcium-dependent release of messengers from astrocytes.
Pericytes play a key role in brain ischaemia. Work from the 1960s showed that after brain ischaemia the microvasculature is not properly reperfused even if the causative clot is removed from an artery to the brain. This restriction of energy supply will lead to ongoing damage to neurons. Imaging pericytes during ischaemia showed that they constrict capillaries (Peppiatt et al., 2006), presumably because the loss of ATP supply inhibits their ion pumping, leading to a rise of intracellular calcium concentration. Further work demonstrated that the reason that the decrease of blood flow is so long lasting is that pericytes die readily, in rigor, leaving the capillaries constricted even if the energy supply is restored (Hall et al., 2014). Similar events occur in the heart after cardiac ischaemia (O'Farrell et al., 2017).
