Clinically Relevant Epithelial Lining Fluid Concentrations of Meropenem with Ciprofloxacin Provide Synergistic Killing and Resistance Suppression of Hypermutable Pseudomonas aeruginosa in a Dynamic Biofilm Model

Treatment of exacerbations of chronic Pseudomonas aeruginosa infections in patients with cystic fibrosis (CF) is highly challenging due to hypermutability, biofilm formation, and an increased risk of resistance emergence. We evaluated the impact of ciprofloxacin and meropenem as monotherapy and in combination in the dynamic in vitro CDC biofilm reactor (CBR). Two hypermutable P. aeruginosa strains, PAOΔmutS (MIC of ciprofloxacin [MIC_{ciprofloxacin}], 0.

ABSTRACT

Treatment of exacerbations of chronic Pseudomonas aeruginosa infections in patients with cystic fibrosis (CF) is highly challenging due to hypermutability, biofilm formation, and an increased risk of resistance emergence. We evaluated the impact of ciprofloxacin and meropenem as monotherapy and in combination in the dynamic in vitro CDC biofilm reactor (CBR). Two hypermutable P. aeruginosa strains, PAOΔmutS (MIC of ciprofloxacin [MIC_{ciprofloxacin}], 0.25 mg/liter; MIC_meropenem, 2 mg/liter) and CW44 (MIC_{ciprofloxacin}, 0.5 mg/liter; MIC_meropenem, 4 mg/liter), were investigated for 120 h. Concentration-time profiles achievable in epithelial lining fluid (ELF) following FDA-approved doses were simulated in the CBR. Treatments were ciprofloxacin at 0.4 g every 8 h as 1-h infusions (80% ELF penetration), meropenem at 6 g/day as a continuous infusion (CI) (30% and 60% ELF penetration), and their combinations. Counts of total and less-susceptible planktonic and biofilm bacteria and MICs were determined. Antibiotic concentrations were quantified by an ultrahigh-performance liquid chromatography photodiode array (UHPLC-PDA) assay. For both strains, all monotherapies failed, with substantial regrowth and resistance of planktonic (≥8 log₁₀ CFU/ml) and biofilm (>8 log₁₀ CFU/cm²) bacteria at 120 h (MIC_{ciprofloxacin}, up to 8 mg/liter; MIC_meropenem, up to 64 mg/liter). Both combination treatments demonstrated synergistic bacterial killing of planktonic and biofilm bacteria of both strains from ∼48 h onwards and suppressed regrowth to ≤4 log₁₀ CFU/ml and ≤6 log₁₀ CFU/cm² at 120 h. Overall, both combination treatments suppressed the amplification of resistance of planktonic bacteria for both strains and of biofilm bacteria for CW44. The combination with meropenem at 60% ELF penetration also suppressed the amplification of resistance of biofilm bacteria for PAOΔmutS. Thus, combination treatment demonstrated synergistic bacterial killing and resistance suppression against difficult-to-treat hypermutable P. aeruginosa strains.

TEXT

Respiratory tract infections in patients with cystic fibrosis (CF) present a serious medical challenge, and Pseudomonas aeruginosa, a difficult-to-treat pathogen, has a great impact on this group of patients (1). CF is a complex genetic disease caused by defective function of the CF transmembrane conductance regulator, resulting in altered sputum viscosity, disrupted airway anatomy, and impaired mucociliary clearance (2). These pathological conditions predispose to a repetitive cycle of acute infective exacerbations (AIEs) of chronic P. aeruginosa infections, causing worsened disease progression and increased early death in patients with CF (3, 4).

P. aeruginosa has the exceptional capacity to evade virtually all antimicrobials when used alone, resulting in treatment failure due to the selection of resistant mutants (5). Hypermutation (up to a 1,000-fold-increased mutation rate) occurs in up to ∼65% of P. aeruginosa strains from patients with CF (6, 7) and is highly correlated with the establishment of chronic P. aeruginosa infection through biofilm formation (8). The important hallmark of biofilm-related infections is the increased secretion of the bacterial extracellular matrix, which limits the access of antimicrobials to the infecting pathogens and harbors phenotypic diversity (9). The presence of hypermutable P. aeruginosa in association with biofilm formation renders the treatment of AIE difficult and often results in persistence of infection and multidrug resistance (10, 11).

A combination of two or more antibiotics is currently recommended for the treatment of P. aeruginosa early exacerbations (12–14); however, information about the rational dosing of antibiotic combinations in CF is limited. Recently, we demonstrated that combining ciprofloxacin with meropenem, antibiotics with different mechanisms of action and resistance, combats hypermutable P. aeruginosa in the dynamic hollow-fiber infection model (HFIM) (15). However, the antibacterial activity of this combination against hypermutable P. aeruginosa embedded in biofilms has not been explored. Therefore, the dynamic CDC biofilm reactor (CBR) model was used in the present study. The CBR is a well-accepted, state-of-the-art pharmacokinetic/pharmacodynamic (PK/PD) model that allows the simulation of clinically relevant epithelial lining fluid (ELF) concentration-time profiles as seen in patients and also enables the examination of antibacterial effects on both planktonic and biofilm bacteria simultaneously.

The main objective of the present study was to systemically evaluate the impact of ciprofloxacin and meropenem as monotherapy and in combination against hypermutable P. aeruginosa strains in the dynamic in vitro CBR model. We examined the time course of bacterial killing and resistance suppression of both planktonic and biofilm bacteria over 120 h by simulating clinically relevant ELF concentrations of ciprofloxacin and meropenem alone and in combination in the CBR.

RESULTS

PK validation and microbiological response.

The observed ciprofloxacin and meropenem concentrations in the CBR were on average within 5% of the targeted concentrations (Table 1). Viable-cell-count profiles for total populations of planktonic and biofilm bacteria (n = 2 measurements) for both strains are shown in Fig. 1, while the corresponding profiles for less-susceptible populations of PAOΔmutS and CW44 are shown in Fig. 2 and 3, respectively. Log changes in viable-cell counts of total bacteria, mutant frequencies, and MICs are shown in Tables 2 to 4, respectively.

TABLE 1.

Clinically representative ELF concentrations, exposures, and pharmacokinetic/pharmacodynamic indices for ciprofloxacin and/or meropenem against PAOΔmutS and CW44 in the CBR^a

Isolate	Treatment	fCmax/fCmin or fCss (mg/liter)	fAUC24 (mg · h/liter)	fCmax/MIC or fCss/MIC	fT>MIC (%)	fAUC24/MIC
PAOΔmutS	CIP at 0.4 g every 8 h	2.64/0.50	31.4	10.5		125.7
	MER at 6 g/day as CI	4.51	108	2.26	100
	MER at 6 g/day as CI (60%)	9.02	216	4.51	100
CW44	CIP at 0.4 g every 8 h	2.64/0.50	31.4	5.27		62.9
	MER at 6 g/day as CI	4.51	108	1.13	100
	MER at 6 g/day as CI (60%)	9.02	216	2.25	100

^aThe simulated half-life was 2.9 h for ciprofloxacin (CIP). The meropenem dosage regimen (6 g/day as a continuous infusion [CI]) was started at an unbound average steady-state concentration (fC_ss) of either 4.51 mg/liter (30% ELF penetration) or 9.02 mg/liter (60% ELF penetration). The simulated ELF penetration was 30% for meropenem (MER), unless specified as 60%, and was 80% for ciprofloxacin. fC_max, unbound peak concentration; fC_min, unbound minimum concentration before the next dose; fAUC₂₄, area under the unbound concentration-time curve over 24 h; fC_max/MIC, ratio of fC_max to MIC; fT_>MIC, cumulative percentage of a 24-h period that unbound concentrations exceeded the MIC; fAUC₂₄/MIC, ratio of fAUC₂₄ to MIC.

FIG 1.

Total viable counts for growth controls and treatments with ciprofloxacin (CIP) and/or meropenem (MER) with clinically relevant ELF concentration-time profiles. The simulated ELF penetration was 30% for meropenem, unless specified as 60%, and was 80% for ciprofloxacin. Samples were obtained from the medium within the reactor, i.e., planktonic bacteria, and from coupons, i.e., biofilm bacteria. The y axis starts from the limit of counting.

FIG 2.

Effect of each dosage regimen on the counts of PAOΔmutS bacteria able to grow on agar plates containing 0.57 or 1.25 mg/liter of ciprofloxacin or 5 or 10 mg/liter of meropenem. To differentiate less-susceptible subpopulations from the predominant population, the antibiotic concentrations in agar were based upon Etest MICs, which were 0.50 mg/liter for meropenem and 0.064 mg/liter for ciprofloxacin (19).

FIG 3.

Effect of each dosage regimen on the CW44 counts able to grow on agar plates containing 5 or 10 mg/liter of meropenem or 0.57 or 1.25 mg/liter of ciprofloxacin. To differentiate less-susceptible subpopulations from the predominant population, the antibiotic concentrations in agar were based upon Etest MICs, which were 0.5 mg/liter for meropenem and 0.19 mg/liter for ciprofloxacin (19).

TABLE 2.

Log changes in viable-cell counts of total bacteria at various time points with clinically relevant ELF concentration exposures of meropenem and/or ciprofloxacin^a

^aThe green background indicates synergy (a ≥2-log₁₀ decrease in the CFU per milliliter or CFU per square centimeter with the combination compared to its most active component) (for planktonic bacteria, a ≥2-log₁₀ decrease in the CFU per milliliter compared to the initial inoculum was also required for synergy); the blue background indicates a 1.0- to <2-log₁₀ decrease in the number of CFU per milliliter or CFU per square centimeter with the combination compared to its most active component. MER, meropenem; CIP, ciprofloxacin.

TABLE 3.

Log₁₀ mutant frequencies at 5 mg/liter and 10 mg/liter meropenem and at 0.57 mg/liter and 1.25 mg/liter ciprofloxacin^a

^aThe red background indicates a high mutant frequency, i.e., a large proportion of less-susceptible bacteria being present in the total population; the green background indicates a low mutant frequency, i.e., a small proportion of less-susceptible bacteria being present in the total population. CIP, ciprofloxacin; MER, meropenem.

TABLE 4.

MIC values for colonies retrieved from antibiotic-containing agar plates (1.25 mg/liter ciprofloxacin and 10 mg/liter meropenem) at 0 and 120 h for each dosage regimen^a

Isolate	Arm	Time (h)	MIC (mg/liter)
			Ciprofloxacin at 1.25 mg/liter		Meropenem at 10 mg/liter
			Planktonic bacteria	Biofilm bacteria	Planktonic bacteria	Biofilm bacteria
PAOΔmutS	Control	0	2	2	NC	NC
		120	4	2	4	NC
	CIP at 0.4 g every 8 h	120	8	4	−	−
	MER at 6 g/day as CI	120	−	−	64	16
	MER at 6 g/day as CI (60% ELF)	120	−	−	32	16
	MER at 6 g/day as CI + CIP at 0.4 g every 8 h	120	4	2	16	8
	MER at 6 g/day as CI (60% ELF) + CIP at 0.4 g every 8 h	120	NC	NC	NC	NC
CW44	Control	0	2	2	8	4
		120	4	4	16	16
	CIP at 0.4 g every 8 h	120	8	8	−	−
	MER at 6 g/day as CI	120	−	−	128	64
	MER at 6 g/day as CI (60% ELF)	120	−	−	64	32
	MER at 6 g/day as CI + CIP at 0.4 g every 8 h	120	4	2	32	8
	MER at 6 g/day as CI (60% ELF) + CIP at 0.4 g every 8 h	120	NC	NC	NC	NC

^aNC, no colonies grew on the antibiotic-containing plates; −, not tested.

Planktonic bacteria.

The starting inocula (means ± standard deviations [SD]) in all arms were 7.5 ± 0.18 log₁₀ CFU/ml (n = 6) for PAOΔmutS and 7.0 ± 0.094 log₁₀ CFU/ml (n = 6) for CW44. PAOΔmutS grew rapidly in growth control chambers and plateaued at ∼8.8 log₁₀ CFU/ml by 24 h (Fig. 1A). Less-susceptible populations plateaued at ∼5.7 and 5.2 log₁₀ CFU/ml on agar containing 0.57 mg/liter and 1.25 mg/liter of ciprofloxacin and grew to ∼3.4 and 1.9 log₁₀ CFU/ml on agar containing 5 mg/liter and 10 mg/liter of meropenem at 120 h, respectively (Fig. 2A, C, E, and G). CW44 grew to ∼8.3 log₁₀ CFU/ml by 24 h and plateaued at ∼8.7 log₁₀ CFU/ml from 72 h (Fig. 1C). Less-susceptible populations increased approximately in proportion to the growth of the total bacterial population (Fig. 3A, C, E, and G and Table 3).

With PAOΔmutS, ciprofloxacin monotherapy produced an initial killing of ∼2.6 log₁₀ CFU/ml at 7 h, followed by slow regrowth close to the values of the growth control by 96 h (Fig. 1A and Table 2). Amplification of ciprofloxacin-less-susceptible populations was observed, such that a large proportion of the entire population grew on plates containing 0.57 mg/liter and 1.25 mg/liter of ciprofloxacin at 120 h (Fig. 2A and C and Table 3). Emergence of resistance was observed with an ∼3-log increase of the ciprofloxacin-resistant population compared to the growth control at 120 h. The MIC of the colonies isolated from agar plates containing 1.25 mg/liter of ciprofloxacin was 8 mg/liter at 120 h (Table 4). For CW44, ciprofloxacin monotherapy achieved an initial killing of ∼1.8 log₁₀ CFU/ml at 7 h, followed by steady regrowth approaching control values by 48 h (Fig. 1C and Table 2). Growth of ∼8.1 log₁₀ CFU/ml on agar plates containing 0.57 mg/liter and 1.25 mg/liter of ciprofloxacin was obtained at 120 h (Fig. 3A and C). Emergence of resistance was observed with a >4-log increase of ciprofloxacin-resistant bacteria at 120 h on both sets of drug plates compared to the growth control. The MIC at 120 h was 8 mg/liter (Table 4).

For PAOΔmutS, meropenem monotherapy representing 30% ELF penetration produced ∼2.4-log₁₀ CFU/ml bacterial killing at 24 h, followed by substantial regrowth close to the values for the growth control; with 60% ELF penetration, an additional ∼0.5-log₁₀ CFU/ml killing was achieved at 24 h (Fig. 1A and Table 2). Less-susceptible populations of ≤7.8 log₁₀ CFU/ml on plates containing 5 mg/liter of meropenem and of ≤7.6 log₁₀ CFU/ml on plates containing 10 mg/liter of meropenem were obtained for both levels of ELF penetration (Fig. 2E and G). Emergence of resistance was observed with both meropenem monotherapies, with up to a 5.7-log CFU/ml increase of the meropenem-resistant bacteria compared to the growth control at 120 h. The MICs of colonies recovered from plates containing 10 mg/liter meropenem were 64 mg/liter for 30% ELF penetration and 32 mg/liter for 60% ELF penetration at 120 h (Table 4). For CW44, meropenem monotherapy simulating 30% ELF penetration produced ∼2.3 log₁₀ CFU/ml over the first 24 h, followed by regrowth close to the value for the growth control at 120 h, whereas with meropenem at 60% ELF penetration, ∼1.2 log₁₀ CFU/ml more killing was achieved at 24 h, and regrowth stayed ∼1 log below the control values (Fig. 3C and Table 2). Emergence of meropenem resistance was observed for both levels of ELF penetration in comparison to the growth control at 120 h (Fig. 3E and G and Table 3). The MICs were 128 mg/liter for 30% ELF penetration and 64 mg/liter for 60% ELF penetration at 120 h (Table 4).

For PAOΔmutS, the combination of ciprofloxacin with meropenem representing 30% ELF penetration produced rapid initial killing of ∼3.2 log₁₀ CFU/ml at 3 h, which increased to ∼3.9 log₁₀ CFU/ml by 28 h; the combination simulating meropenem at 60% ELF penetration achieved an additional ∼0.5-log₁₀ CFU/ml killing at 28 h (Fig. 1A and Table 2). With both levels of meropenem ELF penetration, enhanced killing by the combination (reduction of ≥1 to <2 log₁₀ CFU/ml compared to the most active corresponding monotherapy) was observed from the first few hours, and synergy occurred from 48 h and 72 h for the low and high levels of meropenem ELF penetration, respectively (Table 2). Growth on agar plates containing ciprofloxacin at 1.25 mg/liter was ∼2.6 log₁₀ CFU/ml at 120 h, growth on those containing 10 mg/liter meropenem was ∼1.6 log₁₀ CFU/ml for 30% ELF penetration (Fig. 2C and G), and the MIC was 4 mg/liter for ciprofloxacin and 16 mg/liter for meropenem at 120 h (Table 4). With the combination simulating meropenem at 60% ELF penetration, at 120 h, ∼0.6 log₁₀ CFU/ml were observed on agar containing 1.25 mg/liter of ciprofloxacin, but no colonies were present on meropenem-containing agar.

For CW44, bacterial killing of ≥4 log₁₀ CFU/ml was achieved with the combination for both meropenem ELF penetration levels (Fig. 1C and Table 2), with greater killing for the regimen representing 60% ELF penetration. Enhanced bacterial killing by the combination compared with the most active monotherapy occurred within the first day, and synergy was observed from 48 h and 72 h, respectively (Table 2). Growth of ∼2.3 log₁₀ CFU/ml was obtained on plates containing 10 mg/liter meropenem with the combination simulating 30% ELF penetration, which was ∼1.5 log₁₀ CFU/ml below the control values; no colonies were detected from 24 h onwards with the combination simulating 60% ELF penetration (Fig. 3G). No ciprofloxacin-resistant colonies were observed with the combination simulating meropenem at 60% ELF penetration. For the combination with meropenem at 30% ELF penetration, ∼2 log₁₀ CFU/ml were retrieved from plates containing 1.25 mg/liter ciprofloxacin, which was ∼0.5 log₁₀ CFU/ml below the growth control counts (Fig. 3A and C and Table 3). For this combination regimen, at 120 h, the MICs were 4 mg/liter for ciprofloxacin and 32 mg/liter for meropenem (Table 4).

Biofilm-embedded bacteria.

The starting inocula (means ± SD) in all arms were 8.3 ± 0.12 log₁₀ CFU/cm² (n = 2 coupons × 6) and 7.3 ± 0.07 log₁₀ CFU/cm² (n = 2 × 6) for PAOΔmutS and CW44, respectively. By 24 h, the growth controls for both isolates grew to ∼9 log₁₀ CFU/cm² and ∼8.4 log₁₀ CFU/cm², respectively, and plateaued until 120 h (Fig. 1B and D). Less-susceptible populations for ciprofloxacin (0.57 and 1.25 mg/liter on agar) and meropenem (5 and 10 mg/liter on agar) grew to ∼1.7 and 5.5 log₁₀ CFU/cm², respectively, at 120 h (Fig. 2 and Fig. 3B, D, F, and H).

With PAOΔmutS, the monotherapy treatments produced ≤0.9-log₁₀ CFU/cm² killing at 7 h, followed by regrowth close to the growth control; with meropenem at 60% ELF penetration, ∼1.6-log₁₀ CFU/cm² bacterial killing was achieved at 48 h, and regrowth remained ∼1 log₁₀ CFU/cm² below the control values at 120 h (Fig. 1B and Table 2). Substantial increases in less-susceptible populations occurred with ciprofloxacin treatment and both levels of ELF penetration of meropenem (Fig. 2B, D, F, and H). At 120 h, the MIC was 4 mg/liter for ciprofloxacin; the MIC for meropenem at 30% and 60% ELF penetration was 16 mg/liter (Table 4). In contrast, ciprofloxacin in combination with meropenem produced enhanced bacterial killing, with regrowth suppressed from 48 to 120 h. Growth plateaued at ∼6 log₁₀ CFU/cm², which was ∼3.5 log₁₀ CFU/cm² below the control values. The combinations with both levels of meropenem ELF penetration were synergistic from 48 h onwards (Fig. 1B and Table 2). After 120 h of treatment with combinations containing meropenem at either level of ELF penetration, growth on agar containing meropenem at 10 mg/liter or ciprofloxacin at 1.25 mg/liter was >5 log₁₀ units lower than that with the corresponding monotherapy (Fig. 2B, D, F, and H). At 120 h, with the combination containing the lower level of meropenem ELF penetration, the MIC of ciprofloxacin (MIC_CIP) was 2 mg/liter, and that of meropenem was 8 mg/liter (Table 4). A negligible number (∼0.6 log₁₀ CFU/cm²) of colonies were retrieved from plates containing ciprofloxacin at 1.25 mg/liter, and no colonies were observed on agar containing meropenem at 5 or 10 mg/liter (Fig. 2B, D, F, and H).

With CW44, all monotherapies produced bacterial killing that closely matched that of PAOΔmutS at 24 h, followed by slow regrowth to values close to those of the growth control by 72 to 120 h (Fig. 1D and Table 2). Substantial increases in less-susceptible populations for both ciprofloxacin (0.57 and 1.25 mg/liter on agar) and meropenem (5 and 10 mg/liter on agar) were observed (Fig. 3B, D, F, and H and Table 3). Emergence of resistance was observed with a ∼3.5- to 6-log₁₀ CFU/cm² increase of the ciprofloxacin- and meropenem-resistant bacteria compared to the growth control. MICs at 120 h were 8 mg/liter for ciprofloxacin, 64 mg/liter for meropenem at 30% ELF penetration, and 32 mg/liter for meropenem at 60% ELF penetration (Table 4). The combinations simulating meropenem at 30% or 60% ELF penetration achieved a substantially greater antibacterial effect. With meropenem at 30% ELF penetration, synergy was observed from 48 h onwards, and with meropenem at 60% ELF penetration, an increasing level of synergy occurred from 24 h; regrowth remained suppressed below ∼6 log₁₀ CFU/cm² at 120 h (Fig. 1D and Table 2). Less-susceptible populations for both ciprofloxacin and meropenem were observed at 120 h; these were ∼1 to 2 log₁₀ CFU/cm² below the control values, except on agar containing 5 mg/liter meropenem, where they were slightly above the values of the growth control (Fig. 3B, D, F, and H and Table 3). At the end of treatment with the combination containing meropenem at 30% ELF penetration, the MIC of ciprofloxacin was 2 mg/liter, and that of meropenem was 8 mg/liter at this time point; no colonies were observed on antibiotic-containing agar plates with the combination with meropenem at 60% ELF penetration (Table 4).

DISCUSSION

Ciprofloxacin and meropenem have been widely used to treat respiratory infections caused by P. aeruginosa, including in patients with CF (16, 17). However, P. aeruginosa can readily acquire resistance to these antibiotics in monotherapy via selection of mutations (18). Rates of resistance to these antibiotics of >30% have been reported for isolates from respiratory infections in patients with CF, with hypermutable strains especially having high resistance rates (19, 20). Current guidelines endorse the use of antipseudomonal agents in combination for the treatment of acute exacerbations of chronic respiratory infections in patients with CF (12–14). However, the antibacterial effect of ciprofloxacin with meropenem against P. aeruginosa isolates has never been explored in a dynamic biofilm model. This study systematically investigated the impact of ciprofloxacin and meropenem, in monotherapies and in combination, on bacterial killing and resistance emergence of hypermutable P. aeruginosa isolates in the CBR. The simulated pharmacokinetic profiles were representative of antibiotic concentrations expected to be achieved in the ELF of patients with CF following intravenous (i.v.) administration of approved daily doses.

For quinolones such as ciprofloxacin, antibacterial activity has been correlated with the ratio of the unbound (free) area under the plasma concentration-time curve (fAUC) to the MIC (fAUC/MIC) and the ratio of the free plasma peak concentration (fC_max) to the MIC (fC_max/MIC) (21–23). In acutely ill patients with bacterial infections, an AUC/MIC of 125 (corresponding to an fAUC/MIC of 87.5) and a C_max/MIC of ≥8 (fC_max of ≥5.6) have been proposed for clinical cure (24–26). However, these targets were often not reached in hospitalized patients infected with a strain having an MIC of ≥0.5 mg/liter (27). It is very likely that the isolates used in establishing the above-mentioned targets were not hypermutable. In the present study, the corresponding values for fAUC/MIC and fC_max/MIC were 125.7 and 10.5 for PAOΔmutS (MIC_CIP = 0.25 mg/liter) and 62.9 and 5.27 for CW44 (MIC_CIP = 0.5 mg/liter) (Table 1), exceeding both PK/PD targets described above for PAOΔmutS but not for CW44. However, suppression of regrowth and resistance was not achieved with ciprofloxacin monotherapy against planktonic and biofilm bacteria of either strain. This result was in agreement with our previous HFIM study where ciprofloxacin monotherapy was ineffective in suppressing the regrowth of planktonic CW44 (15). Similarly, in another HFIM study, extensive emergence of resistance was observed after 48 h with a regimen generating a ciprofloxacin AUC/MIC of 180 against P. aeruginosa strains in planktonic growth, although it was not stated whether the isolates were hypermutable (28). The emergence of resistance in both studies (15, 28) was consistent with the inverted U relationship and mutant selection window (29, 30). In the present study, the effect of ciprofloxacin monotherapy on biofilm bacteria was attenuated, compared to that on planktonic bacteria. There are two factors that may have contributed to the poorer effect against biofilm bacteria. First, although quinolones are considered to diffuse readily through the biofilm matrix, low oxygen concentrations within the biofilm decrease their antibacterial effect due to the insufficient formation of reactive oxygen species (31, 32). Second, following exposure to a subinhibitory ciprofloxacin concentration in a static system, mutations in certain efflux pump regulators were more frequently found in biofilm growth than in planktonic bacteria (33).

The antibacterial activity of β-lactams, including meropenem, has been traditionally correlated with the fraction of the dosing interval for which the unbound concentration remains above some multiple of the MIC of the infecting pathogen (fT_>MIC) (21–23, 34). Traditionally, an fT_>MIC of ∼40% was considered to be necessary for the optimal bactericidal activity of meropenem (21, 34). More recent studies suggested a value of 100% fT_{>4–5× MIC} to be necessary for resistance suppression (25). In our previous HFIM studies, meropenem monotherapy administered as intermittent (short or prolonged) infusions at a daily dose of 3 g or 6 g or as a continuous infusion (CI) at 3 g/day was ineffective in suppressing the regrowth of planktonic PAOΔmutS (35) and CW44 (15), for fT_>MIC values of 61% and 69 to 88%, and extensive emergence of resistance occurred. In the present study, in the CBR model, 6 g/day meropenem was delivered as a CI to maximize fT_>MIC. Since different levels of ELF penetration of meropenem have been reported in the literature, two ELF penetration levels (30% and 60%) were examined (36–39). When meropenem was administered as monotherapy at both levels of ELF penetration, the concentrations remained above the MIC at all times. Indeed, for PAOΔmutS, the unbound meropenem concentrations remained at >2× MIC and >4× MIC across the entire study duration with ELF penetrations of 30% and 60%, respectively; the corresponding values for CW44 were >1× MIC and >2× MIC. Nevertheless, while some bacterial killing was observed initially with planktonic bacteria and to a lesser extent with biofilm bacteria, extensive regrowth occurred, which was associated with the amplification of meropenem-resistant cells. This failure of meropenem CI, representing an extreme mode of administration, raises questions about the use of monotherapy, especially against hypermutable strains, because such strains can readily develop resistance and become multidrug resistant (MDR) due to the amplification of resistant mutant subpopulations (40).

Combination therapy could be a viable option for the treatment of P. aeruginosa infections involving hypermutable strains. Increases in minimum bactericidal concentrations observed with single agents can be minimized by combining two antipseudomonal agents (40). In the present study, we considered not only synergy but also enhanced bacterial killing (reduction of ≥1 to <2 log₁₀ CFU compared to the most active monotherapy), since even a relatively small increase in activity with combination therapy may be beneficial for patient care. In the CBR studies with combination regimens containing meropenem at either level of penetration against both PAOΔmutS and CW44, enhanced activity was evident from the early stages of treatment, particularly for planktonic bacteria, and synergy against biofilm bacteria occurred across the last 3 days. The synergy observed was notable, given that the isolates are strong hypermutators.

The enhanced and synergistic bacterial killing observed with the combination regimens in the present study may be due to differences in the mechanism of action and resistance of each antibiotic. Combining antibiotics with different mechanisms requires separate and independent mutations for resistance development and may help to minimize the chances of positive selection of resistant mutants (40). Resistance to ciprofloxacin occurs primarily via target-site mutation and overexpression of efflux pumps, whereas resistance mechanisms against meropenem include enzymatic inactivation via carbapenemases, AmpC β-lactamase overexpression, and reduced expression of the gene for the outer membrane porin OprD, which decreases access to the periplasmic space and the penicillin-binding proteins (PBPs) located there (41, 42). In addition, ciprofloxacin has been shown to increase the permeability of the outer membrane of P. aeruginosa and thereby may increase meropenem concentrations in the periplasm (43). The differences in the mechanism of action and resistance of each antibiotic in the combination, together with possibly higher meropenem concentrations in the periplasm, may have contributed to the enhanced and synergistic bacterial killing and resistance suppression against hypermutable strains in the present study. Moreover, strain CW44 was particularly challenging since, in addition to the mutator phenotype, it already contained relevant meropenem resistance mechanisms, including OprD inactivation, PBP3 modification, and AmpC overexpression (19). This study suggests that the above-mentioned types of mechanistic synergy may apply not only to planktonic bacteria but also to biofilm bacteria, although given the complex nature of biofilm matrices, other mechanisms may also be operative for this growth form.

To our knowledge, this is the first study examining the antibacterial activity of meropenem and ciprofloxacin in combination against hypermutable P. aeruginosa isolates in planktonic and biofilm growth in a dynamic biofilm model. Previously, we evaluated this combination in an HFIM study simulating ELF pharmacokinetics (15). However, that study employed only one clinical isolate (CW44) and investigated only planktonic growth. In 72-h static-concentration time-kill (SCTK) studies, we previously explored the activity of this combination against a range of P. aeruginosa isolates, but the use of static concentrations and the examination of only planktonic growth were limitations (15). Some other studies also investigated this combination in vitro via checkboard methods and 24-h SCTK studies against P. aeruginosa isolates from different patient groups and reported synergistic outcomes (44–47). However, those studies did not quantify the time course of bacterial killing in a biofilm, nor did they examine the emergence of resistance. Importantly, the latter studies did not employ a dynamic system, nor did they include hypermutable isolates, and many of the antibiotic concentrations in the SCTK study were higher (reflecting those in plasma) than those that are clinically achievable in ELF after intravenous exposure.

The present study has a number of particular strengths. First, this is the only study to examine the impact of the ciprofloxacin-meropenem combination against biofilm growth and newly shed bacteria from a biofilm. Second, this is the only in vitro biofilm study exploring the effects of concentration-time profiles representative of those in ELF for this combination against both modes of growth of hypermutable P. aeruginosa. To encompass different levels of meropenem ELF exposure, we examined both 30% and 60% penetration, reflecting literature reports (36–39). Both planktonic and biofilm bacteria were enumerated over 5 days, and emergence of resistance was examined. In addition, two hypermutable P. aeruginosa isolates (1 reference and 1 clinical) were employed to inform future studies. Conversely, this study has some limitations. In the future, remaining questions need to be addressed. First, genomic analysis of emergent resistant populations would assist in the confirmation of mechanisms involved in adaptation and resistance amplification, which commonly occur in hypermutable isolates. Furthermore, confocal imaging would elucidate changes in biofilm structure. This information will assist in developing next-generation mechanism-based mathematical models to explore antibiotic effects on biofilm bacteria. In addition, although the CBR is an ideal in vitro dynamic model for examining antibacterial effects, it does not allow the assessment of immune system effects on residual populations, and therefore, in vivo studies are required.

In summary, this study showed that neither ciprofloxacin nor meropenem was effective as monotherapy against planktonic and biofilm bacteria of hypermutable P. aeruginosa strains. However, both antibiotics in combination regimens demonstrated promising results when simulating ELF pharmacokinetic profiles achievable with FDA-approved daily doses in patients with CF. The combination regimens exhibited enhanced bacterial killing and resistance suppression against both isolates. Thus, this promising combination warrants further evaluation.

MATERIALS AND METHODS

Antibiotics, media, bacterial isolates, and susceptibility testing.

Stock solutions of ciprofloxacin (Sigma-Aldrich, Sydney, Australia) and meropenem (Kabi, Melbourne, Australia) were prepared in Milli-Q water as described previously (15). All experiments used cation-adjusted Mueller-Hinton broth (CAMHB) with 1% tryptic soy broth (TSB; BD, Sparks, MD, USA) containing 25 mg/liter Ca²⁺ and 12.5 mg/liter Mg²⁺. Viable-cell counting was carried out on cation-adjusted Mueller-Hinton agar (CAMHA; BD, Sparks, MD, USA). Drug-containing agar plates were prepared on the day of the experiment by adding appropriate volumes of antibiotic stock solutions to CAMHA.

The hypermutable P. aeruginosa strain PAOΔmutS and a clinical hypermutable P. aeruginosa isolate (CW44) were examined. PAOΔmutS is the isogenic hypermutable strain of the P. aeruginosa wild-type reference strain PAO1, constructed by Mena et al. via mutS gene deletion (48). Clinical isolate CW44 was obtained from an adult patient with CF and a respiratory infection and was documented to be deficient in mutL (19). MICs for PAOΔmutS and CW44 were determined in triplicate on each of three separate days using agar dilution according to Clinical and Laboratory Standards Institute (CLSI) guidelines (49); the respective MICs were 0.25 mg/liter and 0.5 mg/liter for ciprofloxacin and 2 mg/liter and 4 mg/liter for meropenem. CLSI guidelines were used to define susceptibility and resistance criteria, with MICs of ≤0.5 mg/liter and ≥2 mg/liter for ciprofloxacin and ≤2 mg/liter and ≥8 mg/liter for meropenem (49). Thus, PAOΔmutS was susceptible to both antibiotics, whereas CW44 was susceptible to ciprofloxacin and intermediate to meropenem. The mechanisms underlying increased meropenem MICs have been previously studied and included the inactivation of the porin OprD (insertion of 1 nucleotide [nt] at position 1200), the modification of the β-lactam target PBP3 (F533L mutation, known to be involved in β-lactam resistance), and the overexpression of the β-lactamase AmpC (deletion of nt 104 in mpl, a negative regulator of ampC) (19). Hypermutability was defined as at least a 20-fold increase in the mutant frequency on rifampin-containing agar compared to that obtained for the control strain PAO1; the mutant frequencies of PAOΔmutS and CW44 were 1,052-fold and 123-fold higher than that of PAO1, respectively (7, 19). The crystal violet assay (50) was carried out to confirm the biofilm formation capacity of both isolates.

In vitro dynamic biofilm model and antibiotic dosing schemes.

The CBR model (Bio Surface Technologies, Bozeman, MT, USA) was used to explore the microbiological response and emergence of resistance of planktonic and biofilm bacteria to the ciprofloxacin and meropenem treatments, in monotherapy and combinations, over 120 h, as described previously (51).

In brief, the CBR model consisted of three components connected in series: a 10-liter carboy containing sterile drug-free CAMHB–1% TSB, a 1-liter central glass reactor, and a carboy for waste collection. A peristaltic pump delivered the broth medium to the central reactor, where a magnetic stir bar operating at 130 rpm provided mixing. The system was maintained at 36°C. Biofilm formation occurred on removable polycarbonate coupons (diameter, 12.7 mm) located in eight polypropylene coupon holders suspended from the reactor lid (three coupons per holder); the total surface area of each coupon was 2.53 cm². Prior to each experiment, isolates were subcultured onto CAMHA and incubated at 36°C for 24 to 48 h (depending on the growth kinetics). Random colonies (2 to 3) were selected and grown overnight in 10 ml TSB, from which early-log-phase growth was obtained. At the commencement of the experiment, 1 ml of this early-log-phase bacterial suspension was inoculated into each reactor containing 350 ml TSB, and the flow of the system was halted for a 28-h conditioning phase to allow the bacteria to grow to form a biofilm. Conditioning involved the removal of all the broth from the reactor at 24 h to expel all planktonic bacteria. Subsequently, CAMHB–1%TSB was passed through the reactor for 4 h (flow rate of 11.67 ml/min) to ensure that planktonic bacteria present at the start of antibiotic treatment (i.e., 0 h) were those newly shed from the biofilm. The flow rate was changed to 1.39 ml/min at 0 h for all treatments, to represent a ciprofloxacin elimination half-life (t_1/2) of 2.9 h, reflecting patients with CF (52).

FDA-recommended daily doses of 1.2 g for ciprofloxacin and 6 g for meropenem were selected for administration. Ciprofloxacin was delivered as a 60-min i.v. infusion every 8 h via syringe drivers. The meropenem regimen was initiated with an appropriate loading dose at 0 h directly into the reactor, and thereafter, it was delivered as a continuous infusion (CI) by spiking the meropenem stock solution into the medium bottle so that all the media flowing through the system contained a constant concentration (unbound average steady-state concentration [fC_ss]) (Table 1) of meropenem. The medium bottles were stored in the fridge and changed every 24 h to avoid thermal degradation. Antibiotic concentrations in ELF are considered to be more relevant than those in plasma for pulmonary infections such as acute exacerbations in patients with CF. Thus, the ELF concentration-time profiles simulated in the CBR were determined based on population PK models from clinical studies in patients with CF and the ELF/plasma penetration ratios of ciprofloxacin and meropenem (52–57). The expected antibiotic plasma concentration-time profiles were simulated in silico using Berkeley Madonna (version 8.3.18) (58). The extents of penetration of ciprofloxacin (80%) (59, 60) and meropenem (30% and 60%) (36–39) into ELF were derived from multiple published studies using different groups of patients. Growth controls for both isolates were also included. Syringe drivers and pumps were tested and calibrated prior to the experiment, and the flow rate through the CBR was monitored during the experiment to ensure the optimal function of the system.

Quantification of bacterial killing and emergence of resistance.

For viable-cell counting, 1-ml broth samples were collected from the reactor vessel at 0, 1, 2, 3, 5, 7, 24, 28, 48, 72, 96, and 120 h for planktonic bacteria. For biofilm bacteria, coupons were aseptically removed at 0, 3, 7, 24, 48, 72, 96, and 120 h and replaced with a blank holder. The coupons were carefully detached from the holders, washed twice in 10 ml of phosphate-buffered saline (PBS) (pH 7.4) to remove planktonic cells, and then stored in tubes containing 10 ml sterile PBS. Biofilm bacteria were extracted by three alternating 1-min cycles of vortex mixing and sonication at 43 kHz, followed by a final 1-min vortexing step (51). The total bacterial population was enumerated by manual plating of 100 μl of an appropriately diluted bacterial suspension onto drug-free CAMHA, followed by incubation for 24 h for PAOΔmutS and for 48 h due to the slow growth of the hypermutable CW44 isolate at 36°C. Planktonic bacteria were expressed as log₁₀ CFU per milliliter, and the number of bacteria recovered from coupons was expressed as log₁₀ CFU per square centimeter. Less-susceptible subpopulations were quantified for planktonic and biofilm bacteria at 0 h (pretreatment) and 24, 72, and 120 h following the start of treatment, and 200 μl of an appropriately diluted sample was plated onto CAMHA supplemented with meropenem at 5 mg/liter or 10 mg/liter or with ciprofloxacin at 0.57 mg/liter or 1.25 mg/liter. The plates were incubated for 48 h. MICs were determined at 0 and 120 h by the agar dilution method for colonies isolated from antibiotic-containing plates to verify phenotypically the presence of stable resistance.

PK validation.

Samples (1 ml) were serially collected from the growth control and treatment arms at multiple time points and immediately stored at −80°C. Meropenem and ciprofloxacin concentrations in CAMHB–1% TSB were measured using validated ultrahigh-performance liquid chromatography photodiode array (UHPLC-PDA) assays on a Shimadzu (Kyoto, Japan) Nexera UHPLC system coupled to a Shimadzu photodiode array detector. Test samples were assayed in a run order alongside matrix-matched calibrators (standards) and quality controls (QCs). A 5-μl aliquot of the sample was injected onto the UHPLC-PDA. For meropenem, an Onyx monolithic C₁₈, 50- by 2.0-mm analytical column (Phenomenex, Torrance, CA, USA) was used at room temperature. The binary mobile phase consisted of 0.1% formic acid in water (mobile phase A) and 0.1% formic acid in methanol (mobile phase B). The gradient was programmed as 100% mobile phase A for 2.0 min that changed over 1.5 min to a mobile phase A/B ratio of 10:90, which was held for 0.5 min, followed by reequilibration to 100% mobile phase A over 1.0 min. The mobile phase flow rate was 1.00 ml/min. For ciprofloxacin, a Shim-pack XR-ODS III, 2.0- by 50-mm (1.6-μm) analytical column (Shimadzu, Kyoto, Japan) was used, preceded by a Security Guard Ultra C₁₈ guard cartridge (Phenomenex, Torrance, CA, USA) held at 30°C. The mobile phase was 87% phosphate buffer (0.1 M; pH 7) with 13% methanol and was delivered isocratically at 0.25 ml/min. The PDA monitored the UV spectrum from 250 to 340 nm. Approximate retention times and wavelengths for quantitation for each analyte peak were 2.7 min at 310 nm for meropenem and 8 min at 330 nm for ciprofloxacin. Samples were quantified against a calibration curve generated from the batch calibrators, and assay performance was ensured by batch acceptance criteria (61). Precisions were 1.4, 0.4, and 0.2% and accuracies were −0.1, −1.3, and −1.8% at 1.6, 16, and 160 mg/liter of meropenem in CAMHB–1% TSB, respectively. Precisions were 5.8, 1.3, 0.5, and 0.4% and accuracies were 3.6, −1.1, 3.5, and 5.4% at 0.2 mg/liter (lower limit of quantification [LLOQ]), 1.6 mg/liter, 16 mg/liter, and 160 mg/liter of ciprofloxacin in CAMHB–1% TSB, respectively.

Data analysis to describe bacterial killing and emergence of resistance.

Synergy was defined as ≥2-log₁₀ CFU/ml or CFU/cm² killing for the combination relative to the most active corresponding monotherapy at a specified time and ≥2 log₁₀ CFU/ml below the initial inoculum (62). Combination regimens achieving enhanced killing with a reduction of ≥1 to <2 log₁₀ CFU/ml or CFU/cm² compared to the most active corresponding monotherapy were also noted, as such an increase in killing may be important clinically. The log change in viable-cell counts was calculated as the difference of log₁₀ CFU at each sample collection time during treatment from the log₁₀ CFU at time zero. Mutant frequencies were calculated as the difference between the log₁₀ CFU per milliliter (log₁₀ CFU per square centimeter) on antibiotic-containing agar and the log₁₀ CFU per milliliter (log₁₀ CFU per square centimeter) on antibiotic-free agar at the same time point. Emergence of resistance was evaluated by comparing the bacterial counts on antibiotic-containing plates for the different treatments to those observed for the growth control.

Data availability.

The figures and tables include the data from the reported studies.