Repositioning of Memantine as a Potential Novel Therapeutic Agent against Meningitic E. coli-Induced Pathogenicities through Disease-Associated Alpha7 Cholinergic Pathway and RNA Sequencing-Based Transcriptome Analysis of Host Inflammatory Responses

Abstract

Neonatal sepsis and meningitis (NSM) remains a leading cause worldwide of mortality and morbidity in newborn infants despite the availability of antibiotics over the last several decades. E. coli is the most common gram-negative pathogen causing NSM. Our previous studies show that α7 nicotinic receptor (α7 nAChR), an essential regulator of inflammation, plays a detrimental role in the host defense against NSM. Despite notable successes, there still exists an unmet need for new effective therapeutic approaches to treat this disease. Using the in vitro/in vivo models of the blood-brain barrier (BBB) and RNA-seq, we undertook a drug repositioning study to identify unknown antimicrobial activities for known drugs. We have demonstrated for the first time that memantine (MEM), a FDA-approved drug for treatment of Alzheimer's disease, could very efficiently block E. coli-caused bacteremia and meningitis in a mouse model of NSM in a manner dependent on α7 nAChR. MEM was able to synergistically enhance the antibacterial activity of ampicillin in HBMEC infected with E. coli K1 (E44) and in neonatal mice with E44-caused bacteremia and meningitis. Differential gene expression analysis of RNA-Seq data from mouse BMEC infected with E. coli K1 showed that several E44-increased inflammatory factors, including IL33, IL18rap, MMP10 and Irs1, were significantly reduced by MEM compared to the infected cells without drug treatment. MEM could also significantly up-regulate anti-inflammatory factors, including Tnfaip3, CISH, Ptgds and Zfp36. Most interestingly, these factors may positively and negatively contribute to regulation of NF-κB, which is a hallmark feature of bacterial meningitis. Furthermore, we have demonstrated that circulating BMEC (cBMEC) are the potential novel biomarkers for NSM. MEM could significantly reduce E44-increased blood level of cBMEC in mice. Taken together, our data suggest that memantine can efficiently block host inflammatory responses to bacterial infection through modulation of both inflammatory and anti-inflammatory pathways.

Fig 1. Effects of MEM on bacterial intracellular survivals of HBMEC and extracellular bacterial growth.

HBMECs were incubated with various concentrations of MEM at 12 h (A) and 1 h (B) before infection (BI), 0 h during infection (C) (DI), or 1 h postinfection (D) (PI). The numbers of surviving intracellular bacteria were determined. All values represent the means of triplicate determinations. Error bars indicate standard deviations. (E) Effect of MEM on the growth of E. coli K1 E44 in BHI broth at various concentrations of MEM. Bacterial growth was monitored by measuring the absorbance of liquid cultures at 600 nm (A600). Similar results were obtained with the bacterial cultures grown in RPMI 1640 containing 10% FBS.

Fig 2. Effects of MEM-mediated blockages of α7 nAChR on nicotine-enhanced E44 invasion and leukocyte transmigration.

E44 invasion (A) and leukocyte (PMN) transmigration (B-C) across HBMEC after exposure to nicotine (NT). (A-B) Effects of different doses of MEM (1 h incubation) on E44 invasion (A) and PMN transmigration (B) across HBMEC treated with (10 μM NT for 48 h) and without NT. (C) Effect of MEM treatment of either HBMEC or PMN on NT-enhanced PMN transmigration. HBMEC were pre-exposed to 10 μM NT for 48 h, and then HBMEC and PMN were treated with MEM for 1 hr prior to the leukocyte transmigration assay. In all treatments, HBMEC without any treatment was taken as a control. All results are expressed as relative invasion and PMN transmigration compared to the corresponding controls without treatments. All invasion and PMN transmigration assays were performed in triplicate wells. Bar graphs show the means ± SD of triplicate samples. Significant differences with regard to the controls are marked by asterisks (*P<0.05; ***P<0.001).

Fig 3. Comparative analysis of the effect of MEM, NMDA and two NMDAR antagonists (DM and Kyn) on bacterial intracellular survival.

HBMECs were incubated with various concentrations of DM (A) and Kyn (B) 24 h before adding bacteria. (C) Effect of NMDA (10 μM) on bacterial intracellular survivals of HBMEC. All values are presented as relative invasion %. All invasion assays were performed in triplicate wells. Bar graphs show the means ± SD of triplicate samples. Significant differences between the treatment and the control groups are marked by asterisks (*P<0.05; ***P<0.001).

Fig 4. Dose-dependent blockage of bacteremia and meningitis.

C57BL/6J mice (10days) were injected (i.p.) twice with various concentrations (0–20 mg/kg) of MEM at 12 h before and at the time of bacteria inoculation (2×10⁵ CFU of E44). Bacteremia and meningitis in mice were evaluated at 18h after infection. The numbers of surviving bacteria in blood (A) and CSF (B) were determined. Neonatal mice were divided into 5 groups (5 pups/group). All values represent the means of determinations. Each experiment was performed three times. *P<0.05, **P<0.01, ***P<0.001.

Fig 5. Blocking effects of MEM on BBB injury and inflammatory responses.

(A) Peripheral blood concentration of cBMEC in mice treated with or without MEM. (B) Magnitude of NF-κB (p65) activation in mice treated with or without MEM. (C) Cerebrospinal fluid (CSF) concentration of MMP-9 in mice treated with or without MEM. Neonatal mice were divided into 4 groups (7 pups/group). Each experiment was performed three times. *P<0.05, **P<0.01, ***P<0.001.

Fig 6. MEM potentiates intracellular killing of E. coli K1 in HBMEC with a combination of Amp.

A: Five combination settings of HBMEC were tested with the same concentration (10 μM) of MEM and different concentrations of Amp (0, 5, 25, 50 and 100 μg/ml). B: Five combination settings of cells were treated with the same amount (50 μg/ml) of Amp and different concentrations of MEM (0, 1, 5, 10, and 15 μM). The intracellular killing activity with a combination of drugs was significantly higher (**P<0.001) than that of the treatment with one drug alone.

Fig 7. MEM and Amp synergistically block the magnitude of bacteremia and meningitis.

C57BL/6J mice (10days) were injected (i.p.) with MEM (20 mg/ kg) and Amp (20 mg/ kg) alone, or MEM in combination with Amp. The treatment was started at 6 h after bacterial infection (2×10⁵ CFU of E44). Bacteremia and meningitis in mice were evaluated at 14h after drug treatment. The bacterial loads in blood (A) and CSF (B) were determined. Neonatal mice were divided into 5 groups (n = 6). All values represent the means of determinations. Each experiment was performed three times. **P<0.01, ***P<0.001.

Fig 8. Transcriptome profile differences between the MEM-treated (MEM; E44+MEM) and untreated control (CON; E44) groups.

The heat map diagram shows the patterns of differentially expressed genes (DEG) within and between biological replicates. MEM-1, MEM-2, MEM-3, CON-1, CON-2 and CON-3 represent biological replicates of uninfected MBMEC, while E44-1, E44-2, E44-3, MEM-E44-1, MEM-E44-2 and MEM-E44-3 represent biological replicates of E44-infected MBMEC. The color red on the heat map indicates upregulation during infection, while the color green indicates downregulation during infection.

Fig 9. RNA-seq analysis of BMEC infected with E. coli and treated with MEM.

Mouse BMECs (triplet) were treated with E44 (MOI: 10:1 for 1.5 h), PBS (Con), MEM (50 μM 12 h) and E44/MEM. Twelve RNA samples were prepared and sent to BGI America for RNA sequencing. Four inflammatory (IL-33, IL-18rap, MMP10 and Irs1 and anti-inflammatory (A20, CISH, Ptgds and ZFP36) genes are significantly down- and up-regulated, respectively, in E44-infected cells treated with MEM when compared to the infected BMEC without treatment.