Rifampicin potentiation of aminoglycoside activity against cystic fibrosis isolates of Pseudomonas aeruginosa

Abstract

Objectives

Rifampicin potentiates the activity of aminoglycosides (AGs) versus Pseudomonas aeruginosa by targeting the AmgRS two-component system. In this study we examine the impact of rifampicin on the AG susceptibility of cystic fibrosis (CF) lung isolates of P. aeruginosa and the contribution of AmgRS to AG resistance in these isolates.

Methods

amgR deletion derivatives of clinical isolates were constructed using standard gene replacement technology. Susceptibility to AGs ± rifampicin (at ½ MIC) was assessed using a serial 2-fold dilution assay.

Results

Rifampicin showed a variable ability to potentiate AG activity versus the CF isolates, enhancing AG susceptibility between 2- and 128-fold. Most strains showed potentiation for at least two AGs, with only a few strains showing no AG potentiation by rifampicin. Notably, loss of amgR increased AG susceptibility although rifampicin potentiation of AG activity was still observed in the ΔamgR derivatives.

Conclusions

AmgRS contributes to AG resistance in CF isolates of P. aeruginosa and rifampicin shows a variable ability to potentiate AG activity against these, highlighting the complexity of AG resistance in such isolates.

Introduction

Pseudomonas aeruginosa is a common lung pathogen and a major cause of morbidity and mortality in patients with cystic fibrosis (CF).¹ Aminoglycosides (AGs) are commonly employed in the management of P. aeruginosa lung infections in CF² although their use is complicated by the toxicity of these agents³ and development of resistance in the infecting P. aeruginosa.⁴ To address these issues, AG-potentiating agents have previously been sought, with a focus on those that target AG resistance mechanisms.⁵ One AG-potentiating agent, rifampicin, was found to target the envelope stress-responsive AmgRS two-component system (TCS),^6,7 a demonstrated determinant of AG resistance in laboratory⁶ and clinical⁸ isolates of P. aeruginosa that also mediates the AG induction of the AG resistance-promoting MexXY multidrug efflux system.⁷ Interestingly, rifampicin only reduced resistance to a subset of AGs, the 4,5-linked AGs neomycin and paromomycin, in a WT laboratory strain and some clinical isolates, although it did increase susceptibility to all AGs, including the more clinically relevant 4,6-linked AGs tobramycin, amikacin and gentamicin, in a clinical isolate in which resistance was attributable to AmgRS-dependent up-regulation of MexXY.⁵ To better understand its utility in potentiating AG activity versus clinical strains of P. aeruginosa, the impact of rifampicin on the AG susceptibility of a large collection of AG-resistant CF lung isolates was assessed, as was the contribution of AmgRS to AG resistance in these isolates.

Materials and methods

Bacterial strains

Strains K3518–K3691 are tobramycin-resistant CF lung isolates of P. aeruginosa, either internationally collected epidemic strains or strains obtained (with consent) from patients attending the Calgary Adult Cystic Fibrosis Clinic (see Table S1, available as Supplementary data at JAC Online). Permission from the Conjoint Health Region Ethics Board of the University of Calgary was granted for the collection and analysis of these strains (REB 15-0854). Strains K3716, K3710, K3712, K3715, K3713 and K3714 are ΔamgR derivatives of CF strains K3519, K3524, K3527, K3528, K3530 and K3533, respectively.

Construction of ΔamgR derivatives of P. aeruginosa

Despite repeated attempts, construction of amgR deletion derivatives of several tobramycin-resistant CF isolates of P. aeruginosa using the previously described pEX18Tc::ΔamgR suicide vector⁶ was unsuccessful. Difficulties in using pEX18Tc to generate deletions in clinical strains has been noted previously⁹ and likely stems from the need to simultaneously select for conjugation of the vector into the strain being mutated and its integration into the chromosome (pEX18Tc lacks a P. aeruginosa origin of replication), either of which may be less efficient in clinical strains. To address this issue, a pEX18Tc::ΔamgR derivative carrying a temperature-sensitive P. aeruginosa origin of replication was constructed, permitting vector conjugation and integration to be carried out separately and sequentially, the former at the permissive temperature and the latter at the non-permissive temperature. Thus, a 1.9 bp PstI fragment carrying a temperature-sensitive origin of replication, mSF^ts1, was excised from plasmid pSS255¹⁰ and inserted into the PstI site of pEX18Tc::ΔamgR, yielding pEX18Tc::ΔamgR-Ts. The vector was introduced into Escherichia coli strain S17-1 (selected on 10 mg/L tetracycline) for subsequent mobilization into the clinical strains using a previously described protocol,⁸ with modifications. Briefly, aliquots of overnight cultures of S17-1 carrying pEX18Tc::ΔamgR-Ts (700 μL) and individual clinical isolates (300–600 μL) were mixed and spotted onto L-agar plates and incubated for 6 h at 30 °C. Following resuspension of the mating mixtures, transconjugants were selected at 30 °C on L-agar plates containing tetracycline (20–100 mg/L) and chloramphenicol (5 mg/L; to counter-select E. coli), the exact tetracycline concentration being determined empirically for each strain being mutated. Transconjugants showing the expected temperature-sensitive phenotype (i.e. growth in tetracycline-containing medium at 30 °C, but not at 42 °C) were cultured overnight in tetracycline-containing L-broth at 30 °C, plated on tetracycline-containing L-agar plates and then cultured for 1–2 days at 42 °C to select for plasmid integration into the chromosome. Tetracycline-resistant colonies were streaked onto sucrose-containing [10% (v/v)] L-agar plates, incubated at 42 °C for 1–2 days and sucrose-resistant colonies were then screened for deletion of amgR using colony PCR as described previously.⁸

Antibiotic susceptibility testing

The susceptibility of P. aeruginosa to AGs was assessed using the 2-fold serial microtitre broth dilution method described previously.¹¹ MICs were recorded as the lowest concentration of antibiotic inhibiting visible growth after 18 h of incubation at 37 °C. Where indicated, rifampicin was included at ½ MIC for the strains being tested.

Results and discussion

Rifampicin potentiates AG activity versus clinical strains of P. aeruginosa

In a previous study, rifampicin was shown to potentiate the activity of 4,5-linked AGs versus the few clinical strains that were examined, with additional potentiation of 4,6-linked AGs seen in a single isolate.⁵ To get a better sense of rifampicin’s utility in potentiating AG activity versus clinical P. aeruginosa strains, the impact of rifampicin on AG activity versus 45 CF lung isolates provisionally identified as tobramycin resistant was assessed. This included several internationally recognized epidemic strains of P. aeruginosa and multiple isolates (median 2/patient) from 12 patients with genotypically distinct infections followed by the Calgary Adult Cystic Fibrosis Clinic.¹² Initially, the susceptibility of these strains to AGs, including tobramycin, was determined in the absence of rifampicin, generally confirming the original tobramycin resistance and, indeed, revealing elevated MICs of multiple AGs (Table 1). At ½ MIC, rifampicin had a variable impact on AG susceptibility that varied with strain and AG (Table 1). As expected, rifampicin had the greatest impact on neomycin, a representative 4,5-linked AG, both in terms of the number of strains for which rifampicin yielded a ≥4-fold decline in MIC (36/45) and the magnitude of the fold decrease (20/45 showed decreases ≥16-fold, with decreases of up to 128-fold seen). Indeed, in several instances it was the only AG whose activity was potentiated ≥4-fold (9/45), reflecting what was seen earlier in the WT strain K767.⁵ Instances of rifampicin-promoted reductions in AG MICs ≥4-fold were less common though still substantial for the 4,6-linked AGs (tobramycin, 22/45 strains; amikacin, 22/45; gentamicin, 24/45). The magnitude of the 4,6-linked AG MIC decrease prompted by rifampicin was also generally lower than that seen for neomycin (typically <16-fold; Table 1). For most strains, there was ≥4-fold potentiation of at least two AGs (32/45) and a few examples where no AG potentiation was seen (4/45), with instances of rifampicin potentiation ≥4-fold for all of the 4,6-linked AGs less common (12/45) and for all AGs even less so (10/45). While rifampicin appears to have a substantial impact on AG susceptibility in the CF isolates, with a marked impact on all clinically relevant 4,6-linked AGs in about a quarter of the isolates examined, the wide variability in magnitude and AG impacted in each strain doubtless speaks to the multiplicity of resistance mechanisms that can contribute to AG resistance in P. aeruginosa¹³ and their presumably variable susceptibility to rifampicin action.

Table 1.

Impact of rifampicin on AG susceptibility of CF lung isolates of P. aeruginosa^a

Strainb	MIC (mg/L)c
	TOB		AMK		GEN		NEO
	−RIF	+RIFd	−RIF	+RIF	−RIF	+RIF	−RIF	+RIF
K3518	32	8 (4)	128	64 (2)	256	64 (4)	512	256 (2)
K3519	32	2 (16)	128	8 (16)	128	4 (32)	256	64 (4)
K3520	16	8 (2)	32	64 (0.5)	64	64 (1)	256	64 (4)
K3522	16	4 (4)	32	16 (2)	32	8 (4)	1024	8 (128)
K3524	128	16 (8)	512	256 (2)	512	128 (4)	1024	64 (16)
K3525	16	16 (1)	1	2 (0.5)	64	64 (1)	32	16 (2)
K3526	1	2 (0.5)	8	8 (1)	8	8 (1)	512	64 (8)
K3527	4	2 (2)	64	8 (8)	64	16 (4)	512	64 (8)
K3528	4	1 (4)	16	4 (4)	16	2 (8)	256	16 (16)
K3529	32	4 (8)	4	4 (1)	512	32 (16)	32	1 (32)
K3530	16	16 (1)	64	64 (1)	64	64 (1)	1024	16 (64)
K3531	32	8 (4)	64	32 (2)	128	32 (4)	1024	128 (8)
K3532	16	16 (1)	32	32 (1)	16	32 (0.5)	256	64 (4)
K3533	16	16 (1)	128	128 (1)	64	64 (1)	1024	8 (128)
K3534	4	4 (1)	32	8 (4)	16	8 (2)	512	8 (64)
K3535	16	16 (1)	64	32 (2)	128	32 (4)	128	32 (4)
K3536	32	32 (1)	256	256 (1)	128	64 (2)	1024	64 (16)
K3537	16	8 (2)	32	16 (2)	64	32 (2)	512	64 (8)
K3538	64	8 (8)	512	128 (4)	512	64 (8)	512	4 (128)
K3539	16	8 (2)	128	32 (4)	256	64 (4)	2048	128 (16)
K3540	8	8 (1)	64	64 (1)	64	32 (2)	1024	16 (64)
K3641	32	16 (2)	32	32 (1)	64	64 (1)	128	128 (1)
K3642	64	16 (4)	256	64 (4)	128	32 (4)	256	128 (2)
K3645	16	4 (4)	64	16 (4)	64	16 (4)	256	2 (128)
K3647	16	8 (2)	64	16 (4)	64	32 (2)	32	8 (4)
K3649	16	8 (2)	64	32 (2)	64	8 (8)	512	32 (16)
K3651	16	8 (2)	64	64 (1)	64	32 (2)	128	64 (2)
K3653	32	8 (4)	128	32 (4)	64	32 (2)	256	64 (4)
K3656	64	32 (2)	128	64 (2)	128	64 (2)	512	128 (4)
K3657	64	32 (2)	256	64 (4)	256	64 (4)	512	256 (2)
K3659	128	32 (4)	512	64 (8)	256	32 (8)	512	64 (8)
K3660	128	64 (2)	512	256 (2)	256	256 (1)	512	256 (2)
K3665	16	4 (4)	128	32 (4)	128	16 (8)	1024	8 (128)
K3667	16	4 (4)	64	32 (2)	64	64 (1)	512	32 (16)
K3669	32	8 (4)	128	64 (2)	128	64 (2)	256	8 (32)
K3673	512	256 (2)	4096	512 (8)	≥4096	256 (≥16)	512	512 (1)
K3675	64	16 (4)	256	128 (2)	256	64 (4)	2048	256 (8)
K3677	8	4 (2)	32	8 (4)	32	16 (2)	256	16 (16)
K3679	32	4 (8)	128	16 (8)	128	16 (8)	512	8 (64)
K3681	128	8 (16)	256	16 (16)	256	16 (16)	512	4 (128)
K3683	128	16 (8)	256	32 (8)	256	32 (8)	1024	128 (8)
K3685	1024	128 (8)	1024	128 (8)	2048	256 (8)	512	256 (2)
K3687	128	8 (16)	512	32 (16)	512	16 (32)	1024	128 (8)
K3689	16	4 (4)	128	16 (8)	16	8 (2)	256	16 (16)
K3691	16	8 (2)	128	32 (4)	64	32 (2)	256	64 (4)

TOB, tobramycin; AMK, amikacin; GEN, gentamicin; NEO, neomycin; RIF, rifampicin.

^aSusceptibility of P. aeruginosa CF isolates to the indicated AGs was assessed in the absence (−) and presence (+) of rifampicin at ½ MIC.

^bStrains where rifampicin enhanced susceptibility to all tested AGs ≥4-fold are underlined. Strains where rifampicin enhanced susceptibility to all AGs except neomycin ≥4-fold are italicized.

^cFold change in MIC of the indicated AG in the presence versus absence of rifampicin is indicated in parentheses. MIC values ≥4-fold lower in the presence of rifampicin are in bold.

^dRifampicin was included in AG susceptibility assays at 2 mg/L (K3526), 4 mg/L (K3528, K3532, K3536, K3641, K3656, K3667), 8 mg/L (K3518, K3519, K3522, K3525, K3529, K3530, K3531, K3533, K3534, K3535, K3537, K3538, K3539, K3540, K3651, K3653, K3657, K3660, K3669, K3675, K3677, K3679, K3685, K3687, K3691) or 16 mg/L (K3520, K3524, K3527, K3642, K3645, K3647, K3649, K3659, K3665, K3673, K3681, K3683, K3689).

AmgRS contributes to AG resistance in clinical strains of P. aeruginosa

Rifampicin potentiation of AG activity was previously shown to be dependent on the presence of AmgRS,⁵ an indication that it was directly or indirectly targeting this TCS. To gain some insights into the nature of the rifampicin potentiation of AGs seen in the current study, as well as to assess the contribution of AmgRS to AG resistance in CF isolates generally, attempts were made to delete amgRS in the CF isolates and to determine the impact on both AG resistance and rifampicin potentiation of AG activity. In focusing initially on strains K3518–K3540, and despite repeated attempts, amgR deletions were achieved in only six isolates (Table 2). This reflects an ongoing problem with deletion construction in clinical isolates and may be explained by conjugation and/or recombination deficiencies in such isolates, the latter potentially related to DNA sequence variations between the plasmid-borne deletion construct and the corresponding genomic region of the clinical isolates. Nonetheless, loss of amgR had a substantive impact on AG resistance in the CF isolates for which deletions could be engineered, with MIC decreases ≥4-fold seen for at least one AG in 6/6 strains and for all 4,6-linked AGs in 3/6 strains (Table 2), an indication that AmgRS is contributing to AG resistance in these isolates. In one case in particular, K3710, the MICs of all four AGs tested decreased a minimum of 16-fold upon loss of amgR (Table 2). Surprisingly, however, given the earlier link between rifampicin potentiation and AmgRS, in most cases and for most AGs, rifampicin still reduced AG MICs for the deletion derivatives, in several instances more so than for the AmgR⁺ parental strains. Thus, rifampicin can clearly impact AG susceptibility independent of AmgRS. The observation, too, that in some instances rifampicin only modestly potentiated AG activity in an AmgR⁺ strain despite a major contribution of AmgRS to AG resistance [e.g. strain K3524 where rifampicin increased susceptibility to amikacin and gentamicin 2- and 4-fold, respectively, while loss of amgR in this strain (see K3710) reduced MICs of these AGs 16- and 32-fold] indicates that rifampicin is unable to reverse all examples of AmgRS-mediated AG resistance. Thus, while rifampicin does have a broad ability to increase susceptibility to individual AGs in clinical strains of P. aeruginosa, its mode(s) of action appear to be varied and, for the most part, undefined.

Table 2.

Impact of loss of amgR on AG susceptibility of CF lung isolates of P. aeruginosa^a

Strainb	AmgR	MIC (mg/L)c
		TOB		AMK		GEN		NEO
		−RIF	+RIFd	−RIF	+RIF	−RIF	+RIF	−RIF	+RIF
K3519	+	32	2	128	8	128	4	256	64
K3716	−	4 (8)	1	16 (8)	16	32 (4)	4	256 (1)	64
K3524	+	128	16	512	256	512	128	1024	64
K3710	−	8 (16)	4	32 (16)	16	16 (32)	8	16 (64)	8
K3527	+	4	2	64	8	64	16	512	64
K3712	−	4 (1)	0.5	16 (4)	1	16 (4)	2	64 (8)	8
K3528	+	4	1	16	4	16	2	256	16
K3715	−	1 (4)	0.25	4 (4)	1	4 (4)	1	16 (16)	8
K3530	+	16	16	64	64	64	64	1024	16
K3713	−	8 (2)	4	32 (2)	8	32 (2)	8	256 (4)	16
K3533	+	16	16	128	128	64	64	1024	8
K3714	−	16 (1)	4	64 (2)	16	32 (2)	16	256 (4)	8

TOB, tobramycin; AMK, amikacin; GEN, gentamicin; NEO, neomycin; RIF, rifampicin.

^aSusceptibility of amgR deletion derivatives of P. aeruginosa CF isolates to the indicated AGs was assessed in the absence (−) and presence (+) of rifampicin at ½ MIC. Results for the AmgR⁺ parent strains are shown for comparison purposes.

^bK3716, K3519 ΔamgR; K3710, K3524 ΔamgR; K3712, K3527 ΔamgR; K3715, K3528 ΔamgR; K3713, K3530 ΔamgR; K3714, K3533 ΔamgR. Strains where loss of amgR enhanced susceptibility to all tested AGs ≥4-fold are underlined. Strains where loss of amgR enhanced susceptibility to all AGs except neomycin ≥4-fold are italicized.

^cFold change in MIC of the indicated AG in the absence versus presence of amgR is indicated in parentheses. MIC values 4-fold lower in the absence of amgR are in bold.

^dRifampicin was included in AG susceptibility assays at 2 mg/L (K3715), 4 mg/L (K3528, K3716), 8 mg/L (K3519, K3530, K3533, K3712, K3713, K3714) or 16 mg/L (K3524, K3527, K3710).