Observed Antagonistic Effect of Linezolid on Daptomycin or Vancomycin Activity against Biofilm-Forming Methicillin-Resistant Staphylococcus aureus in an In Vitro Pharmacodynamic Model

Abstract

Pharmacodynamic activity in antibiotic combinations of daptomycin, vancomycin, and linezolid was investigated in a 48-h in vitro pharmacodynamic model. Using human-simulated free drug concentrations, activity against clinical biofilm-forming methicillin-resistant Staphylococcus aureus isolates was evaluated. Linezolid antagonized vancomycin activity at 24 and 48 h. Linezolid antagonized daptomycin at 24 and 48 h depending on dose and strain. Adding daptomycin increased vancomycin activity at 48 h (P < 0.03). These results may be strain dependent and require further clinical investigation.

TEXT

There is recent increased interest in the activity of protein synthesis inhibitors in combination with cell wall-active agents. Some combination regimens are being used clinically but lack data to support their combined use (1). High-dose daptomycin and linezolid were recommended for use as a combination therapy in the Infectious Diseases Society of America 2011 methicillin-resistant Staphylococcus aureus (MRSA) treatment guidelines for persistent bacteremia or vancomycin failure (2). However, other in vitro studies have demonstrated antagonism with combinations of linezolid and vancomycin (3, 4). To date, there have been limited investigations of daptomycin and linezolid in combination (5, 6). The combined use of these agents prompted an investigation into pharmacokinetic and pharmacodynamic activity and possible interactions when using combinations of bactericidal and bacteriostatic antimicrobials as previously described (7, 8).

(A portion of these results were presented as a poster at the 53rd annual Interscience Conference on Antimicrobial Agents and Chemotherapy [ICAAC], Denver, CO, 10 September 2013.)

Two randomly selected clinical MRSA blood isolates (L31 and L328) from the LaPlante Laboratory at the Providence Veterans Affairs Medical Center were selected for analysis. The two isolates are known biofilm-producing strains, previously isolated from patients with catheter-related bloodstream infections (9). Biofilm formation was determined as previously described (9, 10). Daptomycin (lot CDC271; Cubist Pharmaceuticals, Inc., Lexington, MA), linezolid (lot 11C10U10, 13F05U09; Pfizer, New York, NY), and vancomycin (lot 12070DD, 382553A; Hospira, Lake Forest, IL) were tested. Mueller-Hinton broth (MHB) (Becton Dickinson, Sparks, MD, USA) supplemented with calcium and adjusted to 25 mg/liter calcium chloride (for daptomycin studies, 50 mg/ml of calcium chloride, ionized Ca, and 1.03 to 1.23 mmol/liter) and 12.5 mg/liter magnesium was used for all assays for MICs, minimum bactericidal concentrations (MBCs), and in vitro pharmacodynamic (IVPD) infection models (11–13). Colony counts were determined using tryptic soy agar (TSA) (BD Difco).

A previously described IVPD model was used to evaluate several antibiotic regimens against MRSA (7). Briefly, a 0.5 McFarland standard of planktonic bacteria from overnight growth on TSA was diluted in a one-compartment model (250-ml working volume) to a starting inoculum of ∼10⁶ CFU/ml. Free concentrations of antimicrobials were evaluated. Daptomycin was administered to simulate a 6 mg/kg of body weight dose (half-life [t_1/2], 8 h; maximum concentration of drug in serum [C_max], 98.6 μg/ml; protein binding, 92%; free, unbound fraction of drug in serum [fC_max], 7.9 μg/ml) or a 10 mg/kg dose (t_1/2, 8 h; C_max, 140 μg/ml; protein binding, 92%; fC_max, 11.2 μg/ml) every 24 h (q24h) (14); linezolid at 600 mg q12h (t_1/2, 6 h; C_max, 21 μg/ml; protein binding, 31%; fC_max, 14.5 μg/ml) (15); and vancomycin at 1.25 g q12h (t_1/2, 6 h; C_max, 45 μg/ml; C_min, 15 to 20 μg/ml; protein binding, 55%; fC_max, 20.3 μg/ml) (16). Antibiotics were given as boluses into the compartment, and peristaltic pumps were used to achieve the desired half-lives and replace the medium with fresh MHB. All model experiments were performed in duplicate or triplicate to ensure reproducibility. In addition, simulations in the absence of antibiotics were performed to ensure adequate growth of organisms in the model. Samples were removed from each model at each 0-, 4-, 8-, 24-, 32-, and 48-h time point. Once removed, samples were immediately diluted, plated on TSA, and incubated at 37°C for 24 h before colony enumeration. The limit of detection for this method is 2.0 log₁₀ CFU/ml (17). Antimicrobial carryover was minimized by serial dilution (1:10 to 1:10,000) of plated samples in conjunction with vacuum filtration, if needed, as previously described (12).

MICs and MBCs of study antimicrobial agents were determined by Etest methodology and broth microdilution according to Clinical and Laboratory Standards Institute (CLSI) guidelines (18, 19). All samples were incubated at 37°C in ambient air for 24 h. Etests were used to assess changes in MICs at 24 and 48 h to detect resistance. Plates were examined for growth after 24 h of incubation at 37°C. Changes in MICs were confirmed with broth microdilution MICs. Samples were evaluated directly from the model to prevent changes in MICs from removing antibiotic pressure and to optimize the detection of MIC changes.

Time-kill curves were plotted to determine the reduction in log₁₀ CFU per milliliter over 48 h. Bactericidal activity (99.9% kill) was defined as a ≥3 log₁₀ CFU/ml reduction, and bacteriostatic activity was defined as a <3 log₁₀ CFU/ml change in colony count from the initial inoculum (20). The time to kill 99.9% of the bacteria present was determined by nonlinear regression (using a minimum of 4 data points) if r² was ≥0.95 or by visual inspection. Enhancement of activity was defined as an increase in kill of ≥2 log₁₀ CFU/ml by a combination of antimicrobials versus that of the most active single agent of that combination (7). Improvement was defined as a 1 to 2 log₁₀ CFU/ml increase in kill compared to that of the most active single agent, while combinations that resulted in a ≥1 log₁₀ increase in bacterial growth compared to that of the most active single agent were considered to represent antagonism (7). The terms improvement and enhancement were used because our simulations involve therapeutically obtained serum drug concentrations, and this does not permit the mathematical modeling necessary to consider the standard terms additivity and synergy (7, 21). Indifference was defined as a <1 log₁₀ CFU/ml change in activity.

Samples for pharmacokinetic analyses were obtained through the injection port at 0, 0.5, 1, 2, 4, 6, 8, and 24 h for verification of target antibiotic concentrations. All samples were stored at −80°C until analysis. Daptomycin concentrations were determined by a previously described and validated high-pressure liquid chromatography (HPLC) method (Center for Anti-Infective Research and Development, Hartford, CT) (11). Vancomycin concentrations were determined by a homogeneous particle-enhanced turbidimetric immunoassay (PETIA) (Architect Multigent; Abbott Diagnostics, Abbott Park, IL, USA) at the Providence Veteran Affairs Medical Center (11). The vancomycin assay has a detection range of 0.5 to 80.0 μg/ml and a between day sample precision and coefficient of variation (CV) of 1.6% and <5.0%, respectively. Linezolid concentrations were evaluated using HPLC (Infectious Disease Pharmacokinetics Laboratory) as previously described (7). The half-life, area under the concentration-time curve (AUC), C_max, and minimum concentration (C_min) of the antibiotics were determined by the trapezoidal method utilizing PK Analyst software (Version 1.10; MicroMath Scientific Software, Salt Lake City, UT). Maximum concentration (C_max) to MIC ratios, the percentage of time above the MIC (%T > MIC), and the area under the concentration-time curve from 0 to 24 h (AUC_0–24) to MIC ratios were calculated for each antibiotic and were compared to literature values (22–25).

Changes in bacterial growth (log₁₀ CFU/ml) at 4, 8, 24, and 48 h and time to 99.9% kill were compared by analysis of variance with Tukey's post hoc test. A P value of <0.05 was considered significant (7, 11). All statistical analyses were performed using SPSS statistical software (release 20; SPSS, Inc., Chicago, IL).

The MIC results are shown with the MBCs and pharmacodynamic parameters obtained in Table 1. Pharmacokinetic values obtained were within 8% of targeted values. The results of the IVPD models are demonstrated in Fig. 1 and Table 2.

TABLE 1.

MIC, MBC, and pharmacodynamic parameters obtained from IVPD experiments using free drug concentrations

Isolate and regimen	MIC (μg/ml)	MBC (μg/ml)	fCmax/MIC	%T > MIC	fAUC/MICa	Estimated total AUC/MIC
MRSA (L31)
Daptomycin (6 mg/kg)	0.5	1	17.13 ± 0.61	100	170–181	2,121–2,262
Linezolid	1	>64	14.49 ± 0.66	100	213	309
Vancomycin	2	2	10.77 ± 1.23	100	181–185	402–411
MRSA (L328)
Daptomycin (6 mg/kg)	0.25	0.25	34.26 ± 1.22	100	339–361	4,243–4,524
Linezolid	2	>64	7.24 ± 0.33	100	107	155
Vancomycin	1	1	21.55 ± 2.45	100	362–370	804–823

^afAUC, area under the concentration-time curve for the free, unbound fraction of a drug.

FIG 1.

Activity of daptomycin and linezolid (A and C) or vancomycin and linezolid (B and D) combinations on planktonic MRSA L31 and L328 over 48 h. GC, growth control; DAP6, daptomycin at 6 mg/kg; DAP10, daptomycin at 10 mg/kg; VAN, vancomycin; LZD, linezolid.

TABLE 2.

Activity of each antibiotic alone and in combination in an IVPD model at 24 and 48h

Regimen	MRSA strain	Changea in log10 CFU/ml relative to 0 h at:
		24 h	48 h
Growth control	L31	+2.52 ± 0.11	+2.37 ± 0.27
	L328	+2.46 ± 0.23	+3.29 ± 0.29
Daptomycin at 6 mg/kg	L31	−3.51 ± 0.08	−3.03 ± 0.68
	L328	−3.11 ± 0.32	−3.15 ± 0.28
Daptomycin at 10 mg/kg	L31	−3.54 ± 0.03	−3.48 ± 0.09
	L328	−3.45 ± 0.11	−3.24 ± 0.56
Linezolid	L31	−2.90 ± 0.47	−0.84 ± 0.43
	L328	−2.82 ± 0.69	−1.51 ± 0.54
Vancomycin	L31	−2.85 ± 0.15	−2.02 ± 0.15
	L328	−3.08 ± 0.52	−1.39 ± 0.57
Daptomycin at 6 mg/kg + linezolid	L31	−2.62 ± 0.80 (inhibited 0.81 log CFU/ml, indifference)	−1.14 ± 0.68 (inhibited 1.82 log CFU/ml, antagonism)
	L328	−2.05 ± 0.35 (inhibited 1.04 log CFU/ml, antagonismb)	−1.62 ± 0.89 (inhibited 1.52 log CFU/ml, antagonism)
Daptomycin at 10 mg/kg + linezolid	L31	−2.55 ± 0.58 (inhibited 1.14 log CFU/ml, antagonism)	−1.21 ± 0.66 (inhibited 2.43 log CFU/ml, antagonismb)
	L328	−2.40 ± 0.18 (inhibited 1.01 log CFU/ml, antagonismb)	−2.35 ± 0.83 (inhibited 0.85 log CFU/ml, indifference)
Linezolid + vancomycin	L31	−1.88 ± 0.98 (inhibited 1.00 log CFU/ml, antagonism)	−0.60 ± 0.55 (inhibited 1.36 log CFU/ml, antagonism)
	L328	−1.43 ± 0.17 (inhibited 1.67 log CFU/ml, antagonismb)	−0.14 ± 0.17 (inhibited 1.27 log CFU/ml, antagonismb)
Vancomycin + daptomycin at 6 mg/kg	L31	−3.57 ± 0.08 (no change, indifference)	−3.57 ± 0.08 (enhanced 0.48 log CFU/ml, indifference)
	L328	−3.51 ± 0.10 (enhanced 0.43 log CFU/ml, indifference)	−3.51 ± 0.10 (enhanced 0.39 log CFU/ml, indifference)

^aImprovement, 1 to 2 log₁₀ CFU/ml increase in kill over the most active component; enhancement, >2 log₁₀ CFU/ml increase in kill over the most active component; antagonism, ≥1 log₁₀ CFU/ml increase in growth over the most active component; indifference, <1 log₁₀ CFU/ml change in activity from the most active component.

^bSignificant antagonism from the most active component of the regimen (P < 0.05).

Against the two biofilm-forming isolates, all regimens, including monotherapy and combination, demonstrated statistically significant kill (decrease in CFU per milliliter) by 8 h compared to growth control (P < 0.001). Linezolid demonstrated initial kill until 24 h with regrowth until 48 h. Vancomycin demonstrated bacteriostatic activity at 24 h against L31 but bactericidal activity against L328 at 24 h. Vancomycin was bacteriostatic at 48 h against the two isolates. No increases in MICs were found at 24 or 48 h in any of the experiments.

For the two isolates, daptomycin at 6 mg/kg and 10 mg/kg demonstrated bactericidal activity by 24 h. Daptomycin and vancomycin plus daptomycin were the only regimens to demonstrate sustained bactericidal activity from 24 to 48 h. Daptomycin alone was significantly more active than linezolid at 48 h (mean differences in log CFU per milliliter were 1.78 to 2.73; P < 0.04).

In combination studies, at 24 h, vancomycin plus daptomycin 6 mg/kg and daptomycin 6 mg/kg or 10 mg/kg plus linezolid were not statistically significantly different from their most active components. This is despite meeting the definition for antagonism against the two isolates for daptomycin 10 mg/kg plus linezolid and against L328 for daptomycin 6 mg/kg plus linezolid. Linezolid plus vancomycin was the least active regimen. Linezolid plus vancomycin met the definition for antagonism at 24 h for the two isolates but was significantly different only for L328 (mean difference, 1.67; 95% confidence interval [CI], 0.76 to 2.59; P < 0.01).

Linezolid plus daptomycin 6 mg/kg met the definition for antagonism at 24 h for one isolate and 48 h for both isolates, while the higher dose of daptomycin plus linezolid demonstrated antagonism at 24 h for both isolates and 48 h for one. Against L31, the activity of daptomycin 6 mg/kg or 10 mg/kg alone was significantly greater than daptomycin (either dose) plus linezolid at 48 h (mean difference, 1.82 to 2.43; P < 0.01). The differences in activity between linezolid-containing regimens (linezolid alone, linezolid plus vancomycin, daptomycin plus linezolid) were not statistically significant at 48 h for the two isolates, but linezolid alone was less active than either dose of daptomycin alone (mean differences, 1.78 to 2.73; P < 0.04). Adding daptomycin 6 mg/kg improved the activity of vancomycin at 48 h (mean difference, 1.65 to 2.20; P < 0.03) but was not significantly different compared to daptomycin alone.

Despite common concomitant clinical use of linezolid with bactericidal antibiotics (1), we have demonstrated in vitro antagonism at 24 and 48 h using combinations of linezolid plus vancomycin and linezolid plus daptomycin. Support for the use of these combinations of antibiotics is lacking in both in vitro and clinical outcomes data. Combinations of two active antibiotics are frequently excluded or not analyzed in clinical trials where single agents are the main focus due to small numbers of patients (1, 26). Notably, a landmark study by Lepper et al. demonstrated an increase in mortality in meningitis patients receiving tetracycline-penicillin combination therapy over that in patients receiving the same penicillin dose alone (27). The stasis produced by protein synthesis inhibitors, including linezolid, likely inhibits the activity of cell wall-active antibiotics, which work best on actively dividing bacteria. Antagonism has been demonstrated in previous time-kill studies using static concentrations of combinations of vancomycin and linezolid (3–6). Linezolid has also demonstrated attenuation of activity of aztreonam or ceftazidime against Escherichia coli isolates in an in vitro pharmacodynamic model (7). This highlights the importance of pharmacodynamic interactions with combination therapy, even for antibiotics with a completely different spectrum of activity. Of interest, one study demonstrated the activity of daptomycin and linezolid in combination against MRSA, but in contrast to our study, this study tested formed biofilms on coupons (28).

In our study, regrowth was noted between 24 and 48 h for the two strains, though no increases in MIC were noted using Etests. This may be due to biofilm formation of these planktonic strains after 24 h increasing growth without susceptibility changes since biofilms can withstand 10 to 1,000 times the concentrations of antibiotics compared to planktonic bacteria. According to research by our group, approximately 50% of MRSA isolates from our institution form biofilms (29). Biofilm-forming isolates are known to cause persistent, difficult-to-treat infections where combination therapy may be considered. The strains used in this study previously tested positive for biofilm formation as noted above, using the same temperature and inoculum, with similar media to this IVPD model. Over the 48-h period tested, biofilm growth may seed susceptible bacteria into the model during sampling, which would appear as regrowth (9). A previous study demonstrated a reduction in biofilm biomass but no reduction in cell viability using combinations of linezolid and vancomycin against formed MRSA biofilms (30).

Despite reaching the target of the estimated total AUC/MIC ratio of >400 for vancomycin and with an estimated total vancomycin trough concentration of 15.5 μg/ml, vancomycin did not achieve bactericidal activity against L31 during the 48-h period. This indicates that for an isolate with a vancomycin MIC of 2 mg/liter, this regimen may not be adequate.

Regarding limitations, we evaluated two strains and recognize that these observations may be isolate-specific or dependent on the MICs of the isolates for each antibiotic.

In these daptomycin-, linezolid-, and vancomycin-susceptible strains of biofilm-forming MRSA, regimens containing daptomycin were more active than those containing linezolid. Linezolid antagonized the activity of vancomycin and daptomycin 6 mg/kg and 10 mg/kg at 24 and 48 h. Adding linezolid to daptomycin 6 mg/kg or 10 mg/kg significantly decreased activity at 48 h against L31 versus that of daptomycin alone. The combination of vancomycin plus daptomycin 6 mg/kg or daptomycin 6 mg/kg or 10 mg/kg alone demonstrated sustained bactericidal activity through the 48-h period. Based on these data, combinations of linezolid with either daptomycin at 6 mg/kg or 10 mg/kg or vancomycin should be investigated for the clinical implications of in vitro antagonism.