Synergistic Killing of Polymyxin B in Combination With the Antineoplastic Drug Mitotane Against Polymyxin-Susceptible and -Resistant Acinetobacter baumannii: A Metabolomic Study

Abstract

Polymyxins are currently used as the last-resort antibiotics against multidrug-resistant Acinetobacter baumannii. As resistance to polymyxins emerges in A. baumannii with monotherapy, combination therapy is often the only remaining treatment option. A novel approach is to employ the combination of polymyxin B with non-antibiotic drugs. In the present study, we employed metabolomics to investigate the synergistic mechanism of polymyxin B in combination with the antineoplastic drug mitotane against polymyxin-susceptible and -resistant A. baumannii. The metabolomes of four A. baumannii strains were analyzed following treatment with polymyxin B, mitotane and the combination. Polymyxin B monotherapy induced significant perturbation in glycerophospholipid (GPL) metabolism and histidine degradation pathways in polymyxin-susceptible strains, and minimal perturbation in polymyxin-resistant strains. Mitotane monotherapy induced minimal perturbation in the polymyxin-susceptible strains, but caused significant perturbation in GPL metabolism, pentose phosphate pathway and histidine degradation in the LPS-deficient polymyxin-resistant strain (FADDI-AB065). The polymyxin B - mitotane combination induced significant perturbation in all strains except the lipid A modified polymyxin-resistant FADDI-AB225 strain. For the polymyxin-susceptible strains, the combination therapy significantly perturbed GPL metabolism, pentose phosphate pathway, citric acid cycle, pyrimidine ribonucleotide biogenesis, guanine ribonucleotide biogenesis, and histidine degradation. Against FADDI-AB065, the combination significantly perturbed GPL metabolism, pentose phosphate pathway, citric acid cycle, and pyrimidine ribonucleotide biogenesis. Overall, these novel findings demonstrate that the disruption of the citric acid cycle and inhibition of nucleotide biogenesis are the key metabolic features associated with synergistic bacterial killing by the combination against polymyxin-susceptible and -resistant A. baumannii.

Keywords: combination therapy; metabolomics; mitotane; multidrug-resistance; polymyxin.

FIGURE 1

(A) PCA score plots showing metabolomic variance for untreated samples (blue) and samples treated with polymyxin B (red), mitotane (green) and combination (purple) for each A. baumannii strain along principal components 1 and 2. The degree of separation shows the level of variance between these groups. Principal component 1 axis is responsible for higher level of variance. (B) Venn diagrams showing the number of statistically significant metabolites affected by different treatments (one-way ANOVA, FDR ≤ 0.05; Fisher’s LSD, p ≤ 0.05) in each A. baumannii strain (PMB, polymyxin B; MIT, mitotane).

FIGURE 2

Bar graphs showing the number of significantly perturbed metabolites (ANOVA, FDR ≤ 0.05; Fisher’s LSD, p ≤ 0.05) in different metabolite classes following treatment with polymyxin B, mitotane, and the combination for polymyxin-susceptible A. baumannii ATCC 17978, polymyxin-resistant A. baumannii FADDI-AB225, polymyxin-susceptible A. baumannii ATCC 19606, and polymyxin-resistant A. baumannii FADDI-AB065. The class designated as ‘Others’ includes cofactors and vitamins, glycan, secondary metabolites and metabolites that could not be mapped into pathways based on existing bacterial metabolite databases.

FIGURE 3

Glycerophospholipids in A. baumannii ATCC 17978, ATCC 19606, FADDI-AB065, and FADDI-AB225 following treatment with polymyxin B, mitotane and the combination. (^∗one-way ANOVA, FDR ≤ 0.05; Fisher’s LSD, p ≤ 0.05, log₂ fold-change ≥ |1| thresholds are indicated by vertical dotted lines).

FIGURE 4

Metabolites involved in glycerophospholipid metabolism in A. baumannii significantly impacted by polymyxin B, mitotane, and the combination. Red boxes indicate statistically significant metabolites (Mean ± SD; ^∗one-way ANOVA, FDR ≤ 0.05; Fisher’s LSD, p ≤ 0.05; ^∗log₂ fold-change ≥|1|).

FIGURE 5

Putative fatty acyls from fatty acid metabolism in A. baumannii significantly impacted by polymyxin B, mitotane, and combination treatment (Mean ± SD; ^∗one-way ANOVA, FDR ≤ 0.05; Fisher’s LSD, p ≤ 0.05; ^∗log₂ fold-change ≥ |1|).

FIGURE 6

Metabolites in pentose phosphate pathway in A. baumannii significantly impacted by polymyxin B, mitotane, and the combination. Red boxes indicate statistically significant metabolites (Mean ± SD; ^∗one-way ANOVA, FDR ≤ 0.05; Fisher’s LSD, p ≤ 0.05; ^∗log₂ fold-change ≥|1|).

FIGURE 7

Metabolites in citric acid cycle in A. baumannii significantly impacted by polymyxin B, mitotane, and the combination. Red boxes indicate statistically significant metabolites (Mean ± SD; ^∗one-way ANOVA, FDR ≤ 0.05; Fisher’s LSD, p ≤ 0.05; ^∗log₂ fold-change ≥|1|).

FIGURE 8

Metabolites in pyrimidine and guanine ribonucleotide biogenesis in A. baumannii significantly impacted by polymyxin B, mitotane, and the combination. Red boxes indicate statistically significant metabolites (Mean ± SD; ^∗one-way ANOVA, FDR ≤ 0.05; Fisher’s LSD, p ≤ 0.05; ^∗log₂ fold-change ≥|1|).

FIGURE 9

Metabolites in the histidine degradation pathway in A. baumannii significantly impacted by polymyxin B, mitotane, and the combination. Red boxes indicate statistically significant metabolites (Mean ± SD; ^∗one-way ANOVA, FDR ≤ 0.05; Fisher’s LSD, p ≤ 0.05; ^∗log₂ fold-change ≥|1|).