Exaggerated hypoxic vascular breakdown in aged brain due to reduced microglial vasculo-protection

Abstract

In a recent study of young mice, we showed that chronic mild hypoxia (CMH, 8% O₂ ) triggers transient blood-brain barrier (BBB) disruption, and that microglia play an important vasculo-protective function in maintaining BBB integrity. As hypoxia is a common component of many age-related diseases, here we extended these studies to aged mice and found that hypoxia-induced vascular leak was greatly amplified (5-fold to 10-fold) in aged mice, being particularly high in the olfactory bulb and midbrain. While aged mice showed no obvious difference in the early stages of hypoxic angiogenic remodeling, the compensatory increase in vascularity and vessel maturation was significantly delayed. Compared with young brain, microglia in the normoxic aged brain were markedly activated, and this was further increased under hypoxic conditions, but paradoxically, this correlated with reduced vasculo-protection. Microglial depletion studies showed that microglial still play an important vasculo-protective role in aged brain, but interestingly, partial attenuation of microglial activation with minocycline resulted in fewer vascular leaks and reduced loss of endothelial tight junction proteins. Taken together, these findings suggest that increased BBB disruption in hypoxic aged mice can be explained both by a delayed vascular remodeling response and reduced microglial vasculo-protection. Importantly, they show that overly activated microglia in the aged brain are less effective at maintaining vascular integrity, though this can be improved by reducing microglial activation with minocycline, suggesting therapeutic potential for enhancing BBB integrity in the hypoxia-predisposed elderly population.

Keywords: aging; angiogenesis; blood vessels; blood-brain barrier integrity; brain; chronic mild hypoxia; endothelial proliferation; fibrinogen; microglia.

FIGURE 1

Chronic mild hypoxia (CMH)‐induced vascular leak is much greater in aged brain. (a) Frozen brain sections taken from young (8–10 weeks) or aged (20 months) mice exposed to normoxia or 7‐day hypoxia (8% O₂) were stained for the endothelial marker CD31 (AlexaFluor‐488) and fibrinogen (Cy‐3). Images were captured in the olfactory bulb and the midbrain. Scale bar = 200 μm. (b) High power image of olfactory bulb shown in A. Scale bar = 50 μm. (c) Comparison of young and aged brains at time of sectioning. Scale bar = 2 mm. (d) Quantification of the number of vascular leaks (fibrinogen‐positive)/FOV in different brain regions after 7‐day hypoxia. OB, olfactory bulb; CX, cerebral cortex; ST, striatum; CC, corpus callosum; HC, hippocampus; TH, thalamus; MB, midbrain; PO, pons; MO, medulla oblongata; CB, cerebellum. (e,f) Quantification of the number of vascular leaks in the olfactory bulb and midbrain after 0‐, 4‐, 7‐, and 14‐day hypoxia. All results are expressed as the mean ± SEM (n = 4–8 mice/group). **p < 0.01, ***p < 0.001 vs. normoxic conditions. Note that vascular leak was much greater in aged brain and that the highest number of vascular leaks occurred in the olfactory bulb and midbrain, peaking after 7‐day hypoxia but declined by Day 14. (g) Frozen brain sections taken from mice exposed to 7‐day hypoxia (8% O₂) were dual‐labeled for tyrosine hydroxylase‐1 (TH‐1) (AlexaFluor‐488) and fibrinogen (Cy‐3). Scale bar = 200 μm except for high power (HP) image on the right, where scale bar = 50 μm. Note that hypoxia‐induced vascular leak occurred in the TH‐1+ substantia nigra (SN) and in the surrounding red nucleus.

FIGURE 2

Evaluation of cerebrovascular remodeling in young and aged mice exposed to hypoxia. (a) Frozen brain sections (images show midbrain) taken from young (8–10 weeks) or aged (20 months) mice exposed to hypoxia (8% O₂) for 4 days were stained for CD31 (AlexaFluor‐488) and the proliferation marker Ki67 (Cy‐3). Scale bar = 100 μm. High power (HP) images on the right highlight CD31/Ki67 co‐localization. (b) CD31/MECA‐32 dual‐IF of frozen brain sections (midbrain shown) taken from mice exposed to normoxia or 4‐day hypoxia. Scale bar = 100 μm. (c–e) Quantification of the number of proliferating endothelial cells (CD31⁺/Ki67⁺ cells)/FOV (c), number of blood vessels/FOV (d), and number of MECA‐32+ vessels/FOV (e) after 0‐, 4‐, 7‐, and 14‐day hypoxia. Results are expressed as the mean ± SEM (n = 4–8 mice/group). *p < 0.05, **p < 0.01. Note that the aged brain showed a faster drop‐off of hypoxia‐induced endothelial proliferation, attenuation of increased vascularity response, and a greater number of MECA‐32+ vessels at all time‐points.

FIGURE 3

Microglia in aged brain show much greater activation. Frozen brain sections taken from young (8–10 weeks) or aged (20 months) mice exposed to normoxia or hypoxia (8% O₂) for 4 days were stained for Mac‐1 (AlexaFluor‐488) (a) or Mac‐1 (AlexaFluor‐488) and the proliferation marker Ki67 (Cy‐3) (b). Images were captured in the midbrain. Scale bars = 50 μm (a) or 100 μm (b). Quantification of the number of morphologically activated microglia/FOV (c), total Mac‐1 area/FOV (d) and number of Ki67+/Mac‐1+ cells/FOV (e) after 0‐, 4‐, 7‐, and 14‐day hypoxia. Results are expressed as the mean ± SEM (n = 4–6 mice/group). *p < 0.05, **p < 0.01, ***p < 0.001. Note that microglia in the aged brain are morphologically more active, express higher levels of Mac‐1 and show higher rates of proliferation at all time‐points. In addition, while young microglia show a mute response to hypoxia, those in aged brain exhibit a strong response.

FIGURE 4

Microglia in aged brain show higher levels of CD68 expression but reduced aggregation around leaky blood vessels. Frozen brain sections taken from young and aged mice exposed to normoxia or hypoxia (8% O₂) for 4 days were stained for CD68 (AlexaFluor‐488) (a) or CD31 (AlexaFluor‐488), fibrinogen (Cy‐5), CD68 (Cy‐3), and DAPI (blue). (b). Images were captured in the midbrain. Scale bars = 100 μm. Quantification of the number of CD68⁺ cells/FOV (c), total CD68 area/FOV (d), number of vascular leaks with aggregation of CD68⁺ cells (e), or number of CD68⁺ microglial cells per fibrinogen+ area (f) after 0‐ or 4‐day hypoxia. Results are expressed as the mean ± SEM (n = 5–6 mice/group). ** p <0.01, ***p < 0.001. Note that microglia in aged brain express higher levels of CD68 both under normoxic and hypoxic conditions but show a markedly reduced aggregation around fibrinogen+ vascular leaks.

FIGURE 5

Microglial depletion in aged mice results in greater CMH‐induced cerebrovascular leak. (a) Frozen brain sections taken from aged mice fed normal chow or PLX5622‐containing chow and maintained under normoxic conditions for 21 days were stained for the microglial marker Mac‐1 (AlexaFluor‐488) and DAPI (blue). Scale bar = 200 μm. (b) Images of the olfactory bulb and midbrain taken from aged mice fed normal chow or PLX5622‐containing chow for 21 days before being maintained under hypoxic conditions for 4 days were stained for CD31 (AlexaFluor‐488) and fibrinogen (Cy‐3). Scale bar = 200 μm. (c,d) Quantification of microglial depletion after 21‐day PLX5622 in the olfactory bulb (OB) and midbrain (MB) (c) or the number of vascular leaks/FOV in the olfactory bulb (OB) and midbrain (MB) in aged mice fed normal chow or PLX5622‐containing chow and maintained under hypoxic conditions for 4 days (d). All results are expressed as the mean ± SEM (n = 6–7 mice/group). **p < 0.01. ***p < 0.001. Note that 21‐day PLX5622 reduced microglial density in both brain regions to less than 15% of untreated controls and that both brain regions in PLX5622‐treated mice showed a much higher number of vascular leaks. (e) CD31/fibrinogen dual‐IF images of the hippocampus (HC)/corpus callosum (CC)/cerebral cortex (CTX) regions taken from control chow or PLX5622‐treated mice maintained under normoxic control conditions. Scale bar = 100 μm. Note the presence of vascular leaks within the corpus callosum of normoxic PLX5622‐treated mice but not control mice.

FIGURE 6

Minocycline reduces microglial activation in the aged brain, resulting in less vascular leak. Frozen brain sections from vehicle or minocycline‐treated aged mice exposed to hypoxia for 4 days were stained for Mac‐1 (AlexaFluor‐488) (a), CD68 (AlexaFluor‐488) (c), Mac‐1 (AlexaFluor‐488) and Ki67 (Cy‐3) (e), or CD31 (AlexaFluor‐488) and fibrinogen (Cy‐3) (g). Images were captured in the midbrain. Scale bars = 100 μm (a,c,e) or 200 μm (g). Proliferating microglia are denoted by arrows. Quantification of Mac‐1 area (b), CD68 area (d), number of Mac‐1+/Ki67+ cells/FOV (f), or number of vascular leaks (h). Results are expressed as the mean ± SEM (n = 8–12 mice/group). ***p < 0.001. Note that minocycline markedly reduced microglial activation markers and proliferation, resulting in fewer vascular leaks.

FIGURE 7

Model proposing a biphasic relationship between microglial activation vasculo‐protective function. According to this model, microglia in young normoxic mice (green arrow) are resting but upon hypoxic‐induced vascular leak, they become more activated and display enhanced vasculo‐protective function. In contrast, microglia in aged normoxic mice (red arrow) occupy a higher baseline activation state, so that when stimulated by hypoxia, they become overly activated and less vasculo‐protective.