Metformin Restores Tetracyclines Susceptibility against Multidrug Resistant Bacteria

Abstract

Highly persistent incidence of multidrug resistant (MDR) bacterial pathogens constitutes a global burden for public health. An alternative strategy to alleviate such a crisis is to identify promising compounds to restore antibiotics activity against MDR bacteria. It is reported that the antidiabetic drug metformin exhibits the potentiation effect on tetracycline antibiotics, particularly doxycycline and minocycline, against MDR S. aureus, E. faecalis, E. coli, and S. enteritidis. Mechanistic studies demonstrate that metformin promotes intracellular accumulation of doxycycline in tetracycline-resistant E. coli. In addition, metformin boosts the immune response and alleviates the inflammatory responses in vitro. Last, metformin fully restores the activity of doxycycline in three animal infection models. Collectively, these results reveal the potential of metformin as a novel tetracyclines adjuvant to circumvent MDR bacterial pathogens and to improve the treatment outcome of recalcitrant infections.

Keywords: antibiotic adjuvant; metformin; multidrug resistant bacteria; tetracycline.

Figure 1

Metformin drastically potentiates doxycycline activity against various multidrug resistant bacteria. Chequerboard broth microdilution assays between metformin and doxycycline against MRSA T144, S. aureus 215, VRE A4, E. coli B2, and S. enteritidis H8. Dark‐blue regions represent higher cell density and lower inhibition rate of combinational treatment. Data represent the mean OD (600 nm) of two biological replicates. X‐ and Y‐axes of figures were presented as log2 scale. The FIC indices were calculated at one quarter of MICs of metformin (2.5 mg mL⁻¹ for MRSA T144 and S. aureus 215, 5 mg mL⁻¹ for VRE A4, E. coli B2, and S. enteritidis H8). Synergy is defined as an FIC index of ≤ 0.5.

Figure 2

Time‐dependent killing of pathogens by the combination of doxycycline and metformin. E. coli B2 were grown to A) early and B) late exponential phases in MHB broth, then treated with PBS, doxycycline (DOX, 2, 32, or 128 µg mL⁻¹) or metformin (MET, 5 mg mL⁻¹) alone or in combination (DOX + MET, 2 µg mL⁻¹ + 5 mg mL⁻¹ or 32 µg mL⁻¹ + 5 mg mL⁻¹). The bacterial CFUs per mL at different time points during 24 h were determined. All experiments were performed three times, and the mean ± SD is shown. C) The combination of doxycycline (128 µg mL⁻¹) and metformin (5 mg mL⁻¹) leads to bacterial lysis. D) The addition of metformin (2.5 mg mL⁻¹, one quarter of MIC) prevents the evolution of doxycycline resistance to E. coli ATCC 25922 in vitro. Resistance acquisition during serial passaging in the presence of 0.25 × MIC levels of doxycycline (0.25 µg mL⁻¹).

Figure 3

Synergistic mechanisms of doxycycline‐metformin combination. A) Metformin disrupts the outer membrane of E. coli B2 by measuring fluorescence intensity of 1‐N‐phenylnaphthylamine (NPN) after exposure to increasing concentrations of metformin for 1 h. B) Metformin dissipates membrane potential of E. coli B2. DiSC₃(5) dye was first injected at 10 min followed by self‐quenching and stabilization, then metformin and doxycycline were added at 40 min. The fluorescence units were monitored during 80 min. C) Disruption of proton motive force with increased metformin by monitoring the fluorescence intensity of BCECF‐AM‐probed E. coli cells. Glucose was recognized as positive control due to its ability to enhance PMF. D) Decreased production of intracellular ATP in E. coli cells treated with metformin, measured by a luciferin‐luciferase bioluminescence assay. E) Increased intracellular accumulation of doxycycline in E. coli B2 caused by metformin in a dose‐dependent manner, measured by LC‐MS/MS. Initial concentration of doxycycline was 32 µg mL⁻¹. F) Metformin inhibits the transcription of tet(A) in a dose‐dependent manner, determined by RT‐PCR analysis. All data are presented as mean ± SD and the significances were determined by nonparametric one‐way ANOVA (*p < 0.05, **p < 0.01, ***p < 0.001).

Figure 4

Transcriptome analysis of E. coli B2 after exposure to doxycycline alone or the combination of doxycycline plus metformin. A) Volcano plot and B) GO (gene ontology) annotation analysis of the differential expression genes (DEGs) in E. coli B2 after exposing doxycycline (16 µg mL⁻¹) or the combination of doxycycline (16 µg mL⁻¹) plus metformin (5 mg mL⁻¹) for 4 h. The x‐ and y‐axes in (A) represent the expression changes and corresponding statistically significant degree, respectively. An adjusted p‐value < 0.05 (Student's t‐test with Benjamini–Hochberg false discovery rate adjustment) and |log₂ Fold change| ≥1 were applied as the cutoff for significant DEGs. KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis of C) upregulated DEGs and D) downregulated DEGs. The 10 most significant enriched pathways are shown. E) Selected differential expression genes involved in ribosome, oxidative phosphorylation, ABC transporters and multidrug efflux pump. Data were presented as means of three biological replicates. DOX, doxycycline alone; DOX + M, the combination of doxycycline and metformin.

Figure 5

Immunomodulatory functions of metformin in the eradication of resistant pathogens. A) Supplement of metformin (5 mg mL⁻¹) in combination with doxycycline (0–256 µg mL⁻¹) for 6 h decreases the intracellular bacteria load of E. coli B2 in Vero cells, compared with the doxycycline alone group. B) Cytotoxicity of metformin in RAW264.7 cells by WST‐1 assay. RAW264.7 cells were stimulated with metformin (1–10 mg mL⁻¹). Absorbance at 450 nm were determined in culture supernatant after 24 h incubation. The y‐axis shows the percentage of cell viability to unstimulated cells. C) Effect of metformin on the release of various cytokines in RAW264.7 cells. The cytokines production including TNF‐α, IL‐1β, IL‐8, and IFN‐γ in supernatants after 24 h stimulation was determined by ELISA. Data are presented as mean ± SD from three independent experiments and the significances were determined by unpaired t‐test (n.s., not significant, *p < 0.05). D) Metformin alleviates inflammatory response induced by bacterial LPS (1 µg mL⁻¹). RAW264.7 cells were pretreated with metformin (5 mg mL⁻¹) for 30 min, then stimulated by LPS for 24 h. After incubation, the cytokines in culture samples were monitored by ELISA analysis. All data are presented as mean ± SD from three independent experiments and the significances were determined by nonparametric one‐way ANOVA (*p < 0.05, **p < 0.01, ***p < 0.001). E) Scheme of synergistic mechanisms of metformin in combination with doxycycline against tetracycline‐resistant pathogens. After destroying the outer membrane (only in Gram‐negative bacteria), metformin dissipates membrane potential of cytoplasmic membrane and decreases the proton motive force (in both Gram‐positive bacteria and Gram‐negative bacteria), which subsequently undermines the functions of PMF‐driven efflux pump. To counter this effect, bacteria increase the pH gradient and in turn aid the uptake of doxycycline. These combined actions promote the intracellular accumulation of doxycycline in resistant bacteria. Meanwhile, accumulated doxycycline inhibits the synthesis of bacterial proteins including TetA. In addition, metformin could moderately modulate the immune response by recruitment of neutrophils and control of inflammatory response. Multiple synergistic mechanisms make metformin able to restore tetracyclines activity against resistant pathogens.

Figure 6

Metformin rescues doxycycline activity in vivo. A) Survival rates of the G. mellonella larvae (n = 10 per group) infected by E. coli B2 (1.0 × 10⁶ CFUs) at the right posterior gastropoda with the treatments of doxycycline (50 mg kg⁻¹) or metformin (50 mg kg⁻¹) alone or in combination (50 + 50 mg kg⁻¹) at left posterior gastropoda (n = 10 per group). p‐values were determined by log‐rank (Mantel‐Cox) test. B) Survival rates of the female BALB/C mice (n = 8 per group) infected by a lethal dose of E. coli B2 (1.0 × 10⁸ CFUs) and treated with a single dose of doxycycline (50 mg kg⁻¹) or metformin (50 mg kg⁻¹) alone or a combination of doxycycline plus metformin (50 + 50 mg kg⁻¹), or PBS as vehicle by intraperitoneal injection. p‐values were determined by log‐rank (Mantel‐Cox) test. C) Bacterial load of infected thigh muscle in neutropenic mice (n = 8 per group) by a nonlethal dose of E. coli B2 or MRSA T144 decreased significantly after a single intraperitoneal combination therapy. p‐values were determined by Mann–Whitney U test.