We have shown previously that, both in human Alzheimer’s disease (AD) and in mice mimicking AD, pericytes constrict capillaries, thus reducing cerebral blood flow, while arteriole and venule diameter is unaffected in the AD mice (Nortley et al., 2019). The decrease of CBF is an early event in AD, and a decrease of brain energy supply is known to upregulate production of amyloid beta, suggesting that maintaining CBF by preventing pericyte constriction might prevent some of the symptoms of AD.

To assess this, we used 2-photon microscopy to image capillaries in vivo (through a cranial window over somatosensory cortex) in wild-type or AD [APP(NL-G-F) knock-in] NG2-dsRed mice in which pericytes fluoresce red. FITC dextran (70 kDa) was adminstered intravenously to visualise blood flow. Pericyte calcium concentration was assessed using NG2-Cre^ERT2 mice crossed with floxed GCaMP5g mice, which were administered tamoxifen. After assessing normality of each data distribution, statistical analysis employed (2-tailed) paired or unpaired t-tests or Mann-Whitney tests, as appropriate.

In 4 month old AD mice, a  femoral vein infusion of the blood-brain barrier permeable voltage-gated calcium channel blocker nimodipine (220 microg/kg total, over 10 minutes, using a 60 microg/ml solution) decreased the [Ca²⁺]_i by 16±4% (mean±s.e.m., p=0.006) in 14 1st-3rd branch order pericytes, and increased capillary diameter at the pericyte somata by 17±4% (p=0.003). These changes, together with a similar relaxation of arteriolar smooth muscle cells, increased cortical CBF measured by laser Doppler by 50±6% in 13 AD animals (p=0.0003). Blockage of capillaries (probably by circulating neutrophils or other blood cells) occurred in ~1% of capillaries in 13 wild-type mice, but in 13 AD mice near AD plaques ~20% of capillaries were blocked, which was reduced to ~5% by nimodipine (p<0.0001 for both differences).

Giving nimodipine in the drinking water for 1.5 months, from 2.5 months of age, to mimic clinical prophylaxis for AD, similarly increased the diameter (at pericytes) of 1st-3rd branch order capillaries from 5.2±0.3 microns (n=16 capillaries) to 6.9±0.3 microns (n=38, p=0.004). In area CA1 of the hippocampus of 9 WT animals, hypoxia (assessed with pimonidizole) was seen on average in 2 neurons and glial cells per image stack, while in 9 AD mice this rose to 10.5 per stack (p=0.003), but in 8 AD mice given nimodipine this was reduced to 4.8 per stack (p=0.03).

In brain slices made from neurosurgically-derived human cortical tissue, 75 nM amyloid beta constricts capillaries at pericytes (Nortley et al., 2019). Nimodipine (3 microM) reversed this constriction to the extent that the diameter was not significantly different from that in the absence of amyloid beta, suggesting that the actions of nimodipine on mouse pericytes that are described above are also likely to occur for human pericytes.

These data suggest that prophylactic use of agents targeted at pericytes to reduce capillary constriction may preserve CBF and brain oxygenation in the early stages of Alzheimer’s disease.