Activity of Tigecycline or Colistin in Combination with Zidovudine against Escherichia coli Harboring tet(X) and mcr-1

Alternative therapeutic options are urgently needed against multidrug-resistant Escherichia coli infections, especially in situations of preexisting tigecycline and colistin resistance. Here, we investigated synergistic activity of the antiretroviral drug zidovudine in combination with tigecycline or colistin against E. coli harboring tet(X) and mcr-1 in vitro and in a murine thigh infection model. Zidovudine and tigecycline/colistin combinations achieved synergistic killing and significantly decreased bacterial burdens by >2.

ABSTRACT

Alternative therapeutic options are urgently needed against multidrug-resistant Escherichia coli infections, especially in situations of preexisting tigecycline and colistin resistance. Here, we investigated synergistic activity of the antiretroviral drug zidovudine in combination with tigecycline or colistin against E. coli harboring tet(X) and mcr-1 in vitro and in a murine thigh infection model. Zidovudine and tigecycline/colistin combinations achieved synergistic killing and significantly decreased bacterial burdens by >2.5-log₁₀ CFU/g in thigh tissues compared to each monotherapy.

INTRODUCTION

Due to the paucity of novel antibiotics, tigecycline and colistin have been increasingly used for treating multidrug-resistant (MDR) bacterial infections (1). However, the rapid spread of the mobile colistin resistance gene mcr-1 poses a significant threat to our already shrinking antibiotic reservoir (2). Recently, a unique mobile tigecycline resistance gene, tet(X) [formerly named tet(X4)], was identified in Enterobacteriaceae (3, 4). The tet(X) gene encodes a flavin-dependent monooxygenase that confers high-level resistance to tigecycline (3, 4). In fact, MDR Escherichia coli that are resistant to both colistin and tigecycline have emerged (4). Therefore, there is an urgent need to evaluate alternative therapeutic options against such resistant superbugs. Repurposing FDA-approved nonantibiotic drugs in combination with existing antibiotics is one potential strategy to fill the void of therapies.

Zidovudine (azidothymidine) is an antiretroviral drug for the treatment of HIV infections (5). Previous investigators have reported increased activity of tigecycline in combination with zidovudine against carbapenem-resistant E. coli (6). More recently, Liu et al. (7) showed that the addition of zidovudine suppressed tet(X)-mediated resistance enzyme and prevented the evolution of tigecycline resistance. Additional synergistic activity of the zidovudine and colistin combination has also been reported against MDR Klebsiella pneumoniae (8, 9). This synergy is related to the increased bacterial outer membrane permeability by colistin (8). Zidovudine is a thymidine analogue that inhibits bacterial DNA replication by chain termination (9), while tigecycline inhibits bacterial protein translation by binding to the 30S ribosomal subunit (10). A possible synergistic mechanism of this combination can plausibly be ascribed to interference with both bacterial DNA replication and protein synthesis. Therefore, in this study, we sought to extend the synergistic activity of zidovudine in combination with tigecycline or colistin against MDR E. coli harboring both tet(X) and mcr-1.

Nine well-characterized E. coli strains were utilized in this study (4). PCR and sequencing assays confirmed that all strains harbored tet(X) and mcr-1 (4). MICs were determined by the broth microdilution method in accordance with CLSI guidelines (11). All media were freshly prepared in order to minimize the oxidative degradation of tigecycline (12). The MICs for zidovudine ranged from 0.5 to 4 mg/liter (Table 1). As expected, all E. coli strains were resistant to both tigecycline and colistin. Even with the FDA newly approved eravacycline, the study E. coli exhibited significantly higher MICs (2 to 8 mg/liter; median, 4; Table 1) than the previously reported MIC₉₀ (0.5 mg/liter; n = 2,866) (13). However, in the presence of the clinically achievable concentration of 0.5 mg/liter zidovudine (14), tigecycline was highly potentiated, as characterized by a 4- to 64-fold reduction in MIC (Table 1).

TABLE 1.

Susceptibility profiles of E. coli strains harboring tet(X) and mcr-1 and checkerboard testing for zidovudine in combination with tigecycline or colistin

E. coli strain	MLST	MIC (mg/liter) ofa:											TGC+ZDV FIC indexc	CST+ZDV FIC indexc
		CTX	GEN	CIP	CHL	CST	CSTb	Tetracyclines				ZDV
								TET	TGC	TGCb	ERA
LHM10-1	ST515	4	64	8	>256	8	1	256	8	0.13	4	1	0.25	0.417
DDX6-3	ST515	4	64	8	>256	8	1	256	16	4	2	2	0.5	0.375
DDBR1-3	ST515	4	64	8	>256	4	<0.13	256	16	<0.06	2	0.5	0.313	0.813
LHS1-1	ST515	8	128	8	>256	8	0.5	256	16	1	2	2	0.271	0.438
G3X16-2	ST1196	8	16	>256	>256	8	1	>256	16	2	4	2	0.354	0.656
G3P11-1	ST1196	8	8	>256	>256	8	2	256	16	4	2	2	0.563	0.75
G3X12-2	ST1196	16	16	>256	>256	8	4	256	32	8	8	4	0.75	0.875
G3P8-2	ST1196	8	2	>256	>256	8	1	256	16	2	4	4	0.25	0.458
XFYP18-1	ST8338	0.06	16	0.5	>256	4	0.5	256	32	1	4	1	0.375	0.5

^aCTX, cefotaxime; GEN, gentamicin; CIP, ciprofloxacin; CHL, chloramphenicol; CST, colistin; TET, tetracycline; TGC, tigecycline; ERA, eravacycline; ZDV, zidovudine. MICs were determined for at least three biological replicates.

^bMIC of colistin or tigecycline in the presence of 0.5 mg/liter zidovudine; P < 0.01 for colistin and P < 0.005 for tigecycline (paired t test).

^cInterpreted as synergy (FIC index ≤ 0.5), no interaction (0.5 < FIC index < 4), or antagonism (FIC index > 4) (18).

Checkerboard assays were conducted with the combination of zidovudine and tigecycline/colistin serially diluted at eight concentrations to create an 8 by 8 matrix. Meropenem was used as a nonsynergistic control drug in combination with tigecycline (Fig. S1 in the supplemental material). Tigecycline and zidovudine together were synergistic with a fractional inhibitory concentration (FIC) index of ≤0.5 against 78% of strains tested (Fig. 1 and Fig. S2). The synergistic tigecycline concentrations (1 to 8 mg/liter) in checkerboard assays were comparable to those observed in the gallbladder and colon of tigecycline-treated patients (15). In addition, prior safety evaluation studies showed negligible cytotoxicity to Chinese hamster ovary cells in the presence of the zidovudine (0.5 mg/liter) and tigecycline (<128 mg/liter) combination (7). This suggested that concomitant exposure to synergistic concentrations of tigecycline and zidovudine can be feasible in intraabdominal infections. Similarly, a significant synergistic effect was also observed between zidovudine and colistin against 56% of strains tested (Fig. 1 and Fig. S3). Of note, zidovudine exhibited high clinical potential as a combinatorial adjuvant with colistin, as it is potentiated to active concentrations below the corresponding clinical breakpoint (2 mg/liter; Table 1) for nearly all strains tested.

FIG 1.

Synergistic effects between tigecycline/colistin and zidovudine. Checkerboard assays showing dose-dependent potentiation of tigecycline (a) and colistin (b) by zidovudine against the representative E. coli harboring tet(X) and mcr-1. Dark red and blue regions represent higher bacterial cell densities. The FIC index and optical density at 600 nm (OD_600nm) represent the mean of three biological replicates.

Time-kill experiments were performed with an initial inoculum of ∼10⁶ CFU/ml log-phase E. coli cells in the presence of zidovudine alone or in combination tigecycline-colistin. Bacterial densities were determined by serial viable counts collected at 0, 3, 6, 9, and 24 h of incubation. To control for strain variation in MIC, the subinhibitory concentration (one-fourth MIC) of each drug was used according to the checkerboard assay as a synergistic combination. For the representative strains having different multilocus sequence types (MLSTs), sub-MICs of tigecycline and colistin were inactive and produced a growth pattern similar to that of controls. Zidovudine alone showed bacteriostatic activity for 9 h but was followed by a >2.0-log₁₀ CFU/ml regrowth at 24 h. However, when tigecycline or colistin were combined with zidovudine, significant killing was achieved at 24 h with >3.0-log₁₀ CFU/ml reduction compared with each drug alone (Fig. 2a to c). Similar synergistic killing was also observed in six other E. coli strains carrying tet(X) and mcr-1 (Fig. S4). For the E. coli strain G3P8-2, a dose-dependent killing effect was seen when zidovudine at 1 to 4 mg/liter was combined with tigecycline (4 to 8 mg/liter; Fig. S5) or colistin (2 to 4 mg/liter; Fig. S6).

FIG 2.

Tigecycline/colistin and zidovudine combination therapies are efficacious in vitro and in a murine thigh infection model. (a to c) Time-kill curves showing in vitro synergistic activity of tigecycline-colistin in combination with zidovudine against E. coli harboring tet(X) and mcr-1. Bacterial counts were carried out at different time points after exposure to tigecycline and colistin alone or in combination with zidovudine. Subinhibitory concentrations (one-fourth MIC) of each drug were chosen according to the checkerboard assay to control for strain variation in MICs. The limit of detection was 1.6-log₁₀ CFU/ml. (d to f) Efficacy of zidovudine, tigecycline, and colistin mono- and combination therapies for 24 h against E. coli harboring tet(X) and mcr-1 in a murine thigh infection model. Data are expressed as the mean ± SD from three mice (i.e., six thighs). Bacterial burdens in thighs of untreated mice were determined at 0 h (baseline) and 24 h (control), respectively. The Student's t test was employed to compare bacterial densities in thighs between different groups. P values of <0.05 were considered statistically significant.

A well-characterized murine thigh infection model was used for in vivo studies (16). Six-week-old, specific-pathogen-free female ICR mice (25 to 27 g) were obtained from the Guangdong Medical Lab Animal Center (Guangzhou, China). The Animal Research and Ethics Committee (IACUC) of South China Agricultural University approved these animal studies (approval no. 2019095). Mice were rendered neutropenic by injecting cyclophosphamide intraperitoneally 4 days (150 mg/kg) and 1 day (100 mg/kg) prior to infection. Thigh infections with each E. coli were produced by injecting 0.1 ml of 10^6.5 to 10^7.0 CFU/ml suspension intramuscularly. At 2 h postinfection, mice received either no therapy (control), zidovudine at 10 mg/kg intraperitoneally (i.p.) once daily (q.d.), tigecycline at 2 mg/kg subcutaneously (s.c.) twice daily (b.i.d.), colistin at 5 mg/kg s.c. three times daily (t.i.d.), or their combinations. The drug doses in mice were chosen to mimic pharmacokinetic values similar to those achieved by the clinical dosing of humans, i.e., 600 mg of zidovudine per os (p.o.) q.d., 50 mg of tigecycline intravenously (i.v.) b.i.d., and 3 million IU of colistin i.v. t.i.d. (Table S1). After 24 h of therapy, mice were sacrificed, and thighs were sterilely harvested, homogenized, and quantitatively cultured to determine the mean log₁₀ CFU/g tissue (± standard deviation [SD]; n = 6).

Tigecycline or colistin monotherapy in vivo resulted in increased bacterial burdens in thighs after 24 h by 1.13- to 2.33-log₁₀ CFU/g compared to the baseline. Zidovudine monotherapy inhibited bacterial growth or led to >1.0-log₁₀ CFU/g increase in bacterial burdens. Importantly, when combined with zidovudine, tigecycline or colistin exhibited significantly enhanced therapeutic efficacy, with 1.2- to 2.3-log₁₀ CFU/g killing compared to the baseline or >2.5-log₁₀ CFU/g killing compared to each monotherapy at 24 h (P < 0.0005; Fig. 2d to f). These results are in agreement with previous observations in the murine peritoneal infection model (9). Of note, combinations of antibiotics with other adjuvants should be administered cautiously, bearing in mind the potential additive toxicity (17). In fact, mice survived over 14 days even with combination therapy of high doses of tigecycline (100 mg/kg) and zidovudine (10 mg/kg) (7). Therefore, zidovudine is a safe combinatorial partner for recovering tigecycline or colistin efficacy.

In conclusion, zidovudine can be repurposed as a promising combinatorial partner with tigecycline or colistin against MDR E. coli infections, especially in situations of preexisting colistin and tigecycline resistance.

Data availability.

The raw data supporting the conclusion of the manuscript will be made available by the authors, without undue reservation, to any qualified researcher.