Metformin, Rapamycin, or Nicotinamide Mononucleotide Pretreatment Attenuate Cognitive Impairment After Cerebral Hypoperfusion by Inhibiting Microglial Phagocytosis

Abstract

Vascular cognitive impairment (VCI) is the second leading form of dementia after Alzheimer's disease (AD) plaguing the elder population. Despite the enormous prevalence of VCI, the biological basis of this disease has been much less well-studied than that of AD, with no specific therapy currently existing to prevent or treat VCI. As VCI mainly occurs in the elderly, the role of anti-aging drugs including metformin, rapamycin, and nicotinamide mono nucleotide (NMN), and the underlying mechanism remain uncertain. Here, we examined the role of metformin, rapamycin, and NMN in cognitive function, white matter integrity, microglial response, and phagocytosis in a rat model of VCI by bilateral common carotid artery occlusion (BCCAO). BCCAO-induced chronic cerebral hypoperfusion could cause spatial working memory deficits and white matter lesions (WMLs), along with increasing microglial activation and phagocytosis compared to sham-operated rats. We found the cognitive impairment was significantly improved in BCCAO rats pretreated with these three drugs for 14 days before BCCAO compared with the vehicle group by the analysis of the Morris water maze and new object recognition tests. Pretreatment of metformin, rapamycin, or NMN also increased myelin basic protein (MBP, a marker for myelin) expression and reduced SMI32 (a marker for demyelinated axons) intensity and SMI32/MBP ratio compared with the vehicle group, suggesting that these drugs could ameliorate BCCAO-induced WMLs. The findings were confirmed by Luxol fast blue (LFB) stain, which is designed for staining myelin/myelinated axons. We further found that pretreatment of metformin, rapamycin, or NMN reduced microglial activation and the number of M1 microglia, but increased the number of M2 microglia compared to the vehicle group. Importantly, the number of MBP⁺/Iba1⁺/CD68⁺ microglia was significantly reduced in the BCCAO rats pretreated with these three drugs compared with the vehicle group, suggesting that these drugs suppress microglial phagocytosis. No significant difference was found between the groups pretreated with metformin, rapamycin, or NMN. Our data suggest that metformin, rapamycin, or NMN could protect or attenuate cognitive impairment and WMLs by modifying microglial polarization and inhibiting phagocytosis. The findings may open a new avenue for VCI treatment.

Keywords: metformin; microglia; nicotinamide mono-nucleotide; rapamycin; vascular cognitive impairment.

Figure 1

Metformin, rapamycin, or NMN pretreatment improves cognitive impairment in BCCAO rats. (A) Experimental protocol. Rats received either saline (Veh), metformin (Met), rapamycin (Rapa), or NMN (NMN) at 2 weeks before surgery via intraperitoneal (i.p.) injection daily for 14 consecutive days. The animals were scarified 32 days after surgery. (B) The body weight of rats from different groups (n = 20). The p-values were assessed by two-way repeated-measures ANOVA with Tukey's post hoc test. The escape latency (C) and swimming speed (D) during the training phase, and duration in the target quadrant (E), and the times crossing the area the platform placed in the training phase (F) during probe trial of MWM test (n = 18–19 per group). The p-values in C and D were assessed by two-way repeated-measures ANOVA with Tukey's post hoc test. The p-values in E and F were assessed by Kruskal–Wallis test with a Dunn post-test. (G) The discrimination ratio in novel object recognition test (n = 14–16 per group). The p-values were assessed by one-way ANOVA with a Tukey's test. (H) Representative trajectories in swim path in sham-operated rats and BCCAO rats treated with Veh, Met, Rap, or NMN. All data were shown as mean ± SEM. 2VO, two-vessel occlusion; NOR, novel objective recognition; MWM, Morris water maze; LFB, Luxol fast blue staining; IF, immunofluorescence; WB, Western blot; and NS, not significant.

Figure 2

Metformin, rapamycin, or NMN pretreatment ameliorates WMLs in BCCAO rats. (A) Representative images of Luxol fast blue (LFB) staining in the corpus callosum (CC), internal capsule (IC), and striatum (Str) in BCCAO or sham rats pretreated with vehicle (Veh), metformin (Met), rapamycin (Rapa), or NMN (NMN) at 2 weeks before surgery. (B–D) Quantitative analysis of WMLs in BCCAO or sham-operated rats after treatment. The extent of WMLs was graded as normal (Grade 0), disarrangement of nerve fibers (Grade 1), formation of marked vacuoles (Grade 2), and loss of myelinated fibers (Grade 3). All data were shown as mean ± SEM, n = 6 per group. The p-values were assessed by Kruskal–Wallis test with a Dunn post-test. 2VO, two-vessel occlusion.

Figure 3

Metformin, rapamycin, or NMN pretreatment improves axonal and oligodendrocyte damage induced after BCCAO. (A) Representative images of dephosphorylated neurofilament protein (SMI32, red) and MBP (green) in striatum 32 days after BCCAO in rats. (B) Quantification of the ratio of SMI32 to MBP fluorescence intensity in the striatum, illustrated as fold-change compared with the sham average value. n = 6 per group. The p-values were assessed by one-way ANOVA with a Tukey's test. (C) Western blot analysis of MBP protein from the striatum in sham and BCCAO rats after treatment. (D) Quantitative analysis of MBP protein from the striatum in each group. MBP protein was normalized to the β-actin level of the same sample (n = 4 per group). The p-values were assessed by one-way ANOVA with a Tukey's test. All data were shown as mean ± SEM. Met, metformin; Rapa, rapamycin; NMN, nicotinamide mononucleotide; Veh, vehicle; 2VO, two-vessel occlusion.

Figure 4

Metformin, rapamycin, or NMN pretreatment suppresses BCCAO-induced microglial response. (A) Representative images of immunostaining of Iba1⁺ microglia (red, top panel), CD68⁺ (green, middle panel) and CD206⁺ (green, bottom panel) in striatum of sham or BCCAO rats pretreated with vehicle (Veh), metformin (Met), rapamycin (Rapa), or NMN (NMN) at 2 weeks before surgery. Quantitative analysis of Iba1⁺ cells (B), Iba1⁺CD68⁺ cells (C) and Iba1⁺CD206⁺ cells (D) in striatum in sham or BCCAO rats after pretreatment. The p-values in A were assessed by one-way ANOVA with a Tukey's test. The p-values in B and C were assessed by Kruskal–Wallis test with a Dunn post-test. All data were shown as mean ± SEM, n = 6 per group. 2VO, two-vessel occlusion; NS, not significant.

Figure 5

Metformin, rapamycin, or NMN pretreatment alleviates microglial phagocytosis of myelin in the striatum after BCCAO. (A) Photomicrographs of triple immunofluorescent staining for Iba1 (red), CD68 (green), and MBP (blue). White arrows indicate Iba1⁺ microglia. Yellow arrows indicate Iba1⁺CD68⁺ microglia. Green arrows indicate Iba1⁺CD68⁺ MBP⁺ microglia. (B) Quantification of CD68 (green), Iba1 (red), and MBP (blue)-positive cells in the striatum of sham-operated and BCCAO rats pretreated with vehicle (Veh), metformin (Met), rapamycin (Rapa), or NMN (NMN) for 2 weeks before surgery. The p-values were assessed by the Kruskal–Wallis test with a Dunn post-test. All data were shown as mean ± SEM. N = 6 per group. 2VO, two-vessel occlusion.

Figure 6

Potential mechanisms underlying neuroprotection of metformin, rapamycin, and NMN after BCCAO in rats. In chronic cerebral hypoperfusion (CCH) rats, activated microglia (M1) increase and destroy white matter integrity by releasing pro-inflammatory factors (like IL-6, IL-1β) and inhibiting oligodendrogenesis and myelination via microglial phagocytosis. Metformin and NMN could activate AMP-activated protein kinase (AMPK) and Sirtuin 1 (SIRT1), respectively. Rapamycin inhibits mTOR to increase M2 microglial polarization. AMPK and SIRT1 could also inhibit M1 polarization by decreasing the expression of NF-κB, as well as promoting M2 polarization. The activation of AMPK and the inhibition of mTOR could promote phagocytosis via the phosphorylation of ULK1, while SIRT1 could promote phagocytosis through deacetylation of transcription factor EB (TFEB). There is a complex interplay among AMPK, mTOR, and SIRT1.