NLRP3 inflammasome inhibition with MCC950 improves diabetes-mediated cognitive impairment and vasoneuronal remodeling after ischemia

Abstract

Diabetes increases the risk and worsens the progression of cognitive impairment via the greater occurrence of small vessel disease and stroke. Yet, the underlying mechanisms are not fully understood. It is now accepted that cardiovascular health is critical for brain health and any neurorestorative approaches to prevent/delay cognitive deficits should target the conceptual neurovascular unit (NVU) rather than neurons alone. We have recently shown that there is augmented hippocampal NVU remodeling after a remote ischemic injury in diabetes. NLRP3 inflammasome signaling has been implicated in the development of diabetes and neurodegenerative diseases, but little is known about the impact of NLRP3 activation on functional and structural interaction within the NVU of hippocampus, a critical part of the brain that is involved in forming, organizing, and storing memories. Endothelial cells are at the center of the NVU and produce trophic factors such as brain derived neurotrophic factor (BDNF) contributing to neuronal survival, known as vasotrophic coupling. Therefore, the aims of this study focused on two hypotheses: 1) diabetes negatively impacts hippocampal NVU remodeling and worsens cognitive outcome after stroke, and 2) NLRP3 inhibition with MCC950 will improve NVU remodeling and cognitive outcome following stroke via vasotrophic (un)coupling between endothelial cells and hippocampal neurons. Stroke was induced through a 90-min transient middle cerebral artery occlusion (MCAO) in control and high-fat diet/streptozotocin-induced (HFD/STZ) diabetic male Wistar rats. Saline or MCC950 (3 mg/kg), an inhibitor of NLRP3, was injected at 1 and 3 h after reperfusion. Cognition was assessed over time and neuronal density, blood-brain barrier (BBB) permeability as well as NVU remodeling (aquaporin-4 [AQP4] polarity) was measured on day 14 after stroke. BDNF was measured in endothelial and hippocampal neuronal cultures under hypoxic and diabetes-mimicking condition with and without NLRP3 inhibition. Diabetes increased neuronal degeneration and BBB permeability, disrupted AQP4 polarity, impaired cognitive function and amplified NLRP3 activation after ischemia. Inhibition with MCC950 improved cognitive function and vascular integrity after stroke in diabetic animals and prevented hypoxia-mediated decrease in BDNF secretion. These results are the first to provide essential data showing MCC950 has the potential to become a therapeutic to prevent neurovascular remodeling and worsened cognitive decline in diabetic patients following stroke.

Keywords: Diabetes; Hippocampus; NLRP3 inflammasome; Neurovascular unit; Stroke; Vascular cognitive impairment and dementia (VCID).

Fig. 1.

Stroke exacerbates diabetes-mediated cognitive deficits. (A) Sensorimotor deficits were investigated by composite score. Diabetes had no impact on functional outcomes 2 weeks after MCAO. Novel object recognition (B), two-trial Y-maze (C) and novel social recognition (D) were used to examine cognitive function in control and diabetic rats. Diabetes impaired cognition even before stroke (B–D). Short-term learning and memory deficits were further exacerbated by MCAO as seen by NOR (n = 8; BL = baseline, 3–5 days prior to MCAO induction).

Fig. 2.

Diabetes reduced neuronal density in the DG region of hippocampus. Neuronal density was unchanged in the CA1 region after induction of diabetes and/or middle cerebral artery occlusion (MCAO) (A). Diabetes alone reduced number of NeuN-positive neurons in the DG, but stroke had no effect (B) (n = 6–8; scale bar = 100 μm).

Fig. 3.

Transient MCAO exacerbates hippocampal vascular injury in diabetic rats. Diabetes disrupted blood brain barrier integrity in the hippocampus as seen by increased IgG staining (A and C; p < 0.0001). Representative images from the CA1 and the dentate gyrus are shown in Panels B and D. Stroke further elevated IgG staining in the DG of diabetic rats. Representative images of FITC-filled blood vessels in the CA1 and DG are shown in Panels F and H, respectively. Diabetes increased branch density, vascular volume, and surface in sham but not stroked animals (E, G) (n = 6–8; scale bar = 100 μm).

Fig. 4.

Diabetes disrupts AQP4 polarity and is exacerbated after ischemia. Representative images of aquaporin 4 (AQP4; red), GFAP (blue) and FITC-filled vessels (green) in the CA1 and dentate gyrus (DG) of the hippocampus (A–B). Diabetes increases un-polarized AQP4 in sham animals which is exacerbated after stroke in the CA1 (C) and DG. Astrogliosis was increased after stroke in diabetic rats as seen by elevated surface area in the CA1 (E) and DG (F) of the hippocampus. (n = 8; scale bar = 50 μm) (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

Fig. 5.

Diabetes elevated NLRP3 expression and activation in the hippocampus after ischemic injury. NLRP3 inflammasome (A) and activated IL-1β (B) was up-regulated in the hippocampus after ischemia in diabetic rats. HT22 hippocampal neurons had increased NLRP3 activation as seen by elevated protein levels or NLRP3 scaffold protein (C) and IL-1β (D). (n = 3–4; magnification 63x; scale bar = 20 μm).

Fig. 6.

MCC950 improves post-stroke cognition in diabetic rats. All groups had lower composite scores 3 days after stroke but recovered by day 14 (D14) (A). Cognitive function was measured by novel object recognition (NOR; B), two-trial Y-maze (C), and novel social recognition (NSR; D). Diabetic animals had lower baseline (BL) scores compared to controls. Treatment with MCC950 prevented further decline observed in saline-treated diabetic animals. (Control n = 8; diabetes n = 5).

Fig. 7.

IL-1β and NLRP3 stained sections in the hippocampus. (A) Diabetes amplified IL-1β expression in the CA1 and DG of the hippocampus. IL-1β was mainly co-localized in the neurons as seen by NeuN staining. Treatment with MCC950 appeared to reduce IL-1β staining in both hippocampal regions (identical parameters were used). (B) Sections were co-stained for the scaffold protein in the NLRP3 inflammasome and NeuN. (Magnification 40x; scale bar = 50 μm).

Fig. 8.

MCC950 treatments improved blood brain barrier (BBB) integrity and reduced neurodegeneration in diabetic animals after stroke. MCC950 had no effect on control rats, but improved BBB integrity by reducing percent IgG area threshold (A–B.) Diabetes increased the number of TUNEL-positive neurons in the CA1 (C) and DG (D) after 90-min middle cerebral artery occlusion (MCAO). Treatment with MCC950 significantly reduced cell death in both control and diabetic animals. (n = 8 in control groups and 5 in diabetic groups; scale bar = 100 μm).

Fig. 9.

MCC950 reduced un-polarized aquaporin 4 (AQP4) and reactive microglia phenotype after stroke in diabetes. 90-min middle cerebral artery occlusion (MCAO) induces greater un-polarization of AQP4 in the CA1 (A) and DG (B) in diabetic rats which was ameliorated with MCC950 treatments. NLRP3 inhibition reduced markers of reactive microglia as seen by reduced cell body size (C), elevated number of endpoints (D), protrusions per cell (E) and total processes length (F) in Iba-1 positive microglia of the CA1 region after ischemia. (n = 8 for control; n = 5 for diabetes).

Fig. 10.

MCC950 treatment improved cell viability in HT22 hippocampal cells and brain microvascular endothelial cells (BMVECs) after hypoxia in control and high glucose/palmitate (HG/Pal) conditions. Neurotrophic mature brain derived neurotrophic factor (mBDNF) was measured by western blot in HT22 cells grown in control or HG + Pal in normoxic or oxygen glucose deprivation (OGD) conditions. HG + Pal reduced mBDNF expression in normoxic condition (A). In endothelial cells, BDNF expression was reduced in HG + Pal conditions during normoxia. mBDNF was increased after OGD, but MCC950 reduced expression after hypoxia (B). HT22 were treated with endothelial conditioned media (CM) as shown by the schematic (C). Diabetes reduced cell survival of HT22 cells under normal conditions. Hypoxia reduced cell viability in untreated cells grown in control or HG/Pal diabetic conditions compared to normoxia CM and/or MCC950 seemed to slightly improve cell viability after hypoxia in diabetic conditions (D). (n =. 3–4).